Masaryk University
Faculty of Science
Department of Geography
Response of biotic and abiotic environments on
climate variability in the region of the Antarctic
Peninsula and Svalbard Archipelago
Kamil Láska
Habilitation thesis
Brno 2016
Acknowledgments
First of all, I would like to express my sincere gratitude to Pavel Prošek, who has been an
excellent supervisor of my Ph.D. thesis and taught me a great deal about the atmospheric
sciences and meteorological instruments. Thanks to him, I was invited to attend one of the
first research projects on King George Island in the Antarctic Peninsula region, which
fundamentally changed my scientific career. Over the last two decades of my interest in
the polar regions, I had the unique opportunity of meeting and conducting joint research
with numerous colleagues: my special thanks go to Daniel Nývlt, Miloš Barták, Josef Elster,
Zbyněk Engel, Marie Budíková, Ladislav Budík, Filip Hrbáček, Ondřej Zvěřina, Boyan
Petkov, Vito Vitale, Gennadi Milinevsky, Zuzana Chládová, Jiří Hošek, Zdeněk Stachoň,
Peter Váczi, Linda Nedbalová, Kristián Brat, Jan Kavan, Pavel Kapler, Václav Pavel, Bedřich
Mlčoch, Jiří Komárek, Pavel Ševčík, Tomáš Hájek, Radek Vodrážka, Jana Kvíderová, Zděněk
Máčka, Denisa Witoszová, Olga Bohuslavová, Tomáš Jagoš and Tomáš Franta. Their
extensive help and support during the fieldwork activities are highly appreciated. There
are also many other co-authors with whom I published further research and joint
scientific papers. Additionally, I would like to thank all of the colleagues at the Department
of Geography, Faculty of Science, Masaryk University, who contributed to my scientific
education and current research activities. Funding for this work was provided by
numerous projects of several funding agencies, especially that of the Ministry of
Education, Youth and Sports of the Czech Republic and the Czech Science Foundation
GACR. Particular information about the financial support is acknowledged in the
individual papers listed below.
Last but not least, I would like to thank my family, namely my wife, Gabriela, and the
children, Ondřej and Kateřina, for their immense patience and strength while waiting for
me to return from the polar expeditions.
Abstract
In the thesis, consisting of ten scientific papers, the results of the long-term atmospheric
and environmental research in the selected areas of the Arctic and Antarctic are
summarized. The study investigates the responses of biotic and abiotic environments on
the climate variability that have taken place on different time scales, ranging from the last
few decades to individual years. Attention is paid mainly to the ice-free coastal areas of
the Antarctic Peninsula and the central part of the Arctic Svalbard Archipelago. Both of
these areas have experienced significant atmospheric warming followed by a cascade of
environmental changes and impacts on the terrestrial and aquatic ecosystems. The
Antarctic research has been carried out in the vicinity of the Ukrainian Vernadsky Station
on Galindéz Island and the Czech Johann Gregor Mendel Station, located in the northern
part of James Ross Island. The Arctic research has been focused on the coastal zone of
Petuniabukta, central Spitsbergen Island. The presented work provides the first
comprehensive information about climate conditions in regards to how the constituents of
the polar environments respond to climate variability. A close relationship between some
of the biotic and abiotic components was found to be related to the inter-annual, seasonal
and day-to-day variations of weather patterns and climate features in the studied areas.
For the evaluation of atmospheric and environmental factors and their effects on the geoecosystems,
a variety of statistical methods and models were applied. Among others, the
nonlinear regression model for the estimation of erythemally weighted solar radiation
was developed and successfully validated under various atmospheric conditions. In
addition, the capability of the mesoscale numerical weather prediction model to simulate
the surface wind field in the complex topography of the Svalbard Archipelago was
successfully tested. The remarkable regional differences in climatic conditions and their
effects on the abiotic components were distinguished and further discussed between the
western and eastern side of the Antarctic Peninsula and in the Arctic fjords.
Comprehensive and more detailed information on the individual topics can be found in
the enclosed original publications in the Supplements.
Contents
List of Papers .................................................................................................................................................................. 6
1 Introduction ................................................................................................................................................................ 7
2 Results .............................................................................................................................................................................. 9
2.1 Solar radiation variability in high-latitude locations .................................................. 13
2.1.1 Modelling of solar ultraviolet radiation in the Antarctic Peninsula region ........... 14
2.1.2 Effect of ozone and cloudiness variation on solar ultraviolet radiation
on the western side of the Antarctic Peninsula ................................................................. 15
2.1.3 Evaluation of solar ultraviolet radiation on the eastern side of the Antarctic
Peninsula – the influence of ozone, cloudiness and surface reflectivity ................. 17
2.1.4 Effects of atmospheric and environmental factors on solar radiation
variability at two contrasting Antarctic sites ..................................................................... 19
2.2 Impacts of climate variability on glacial, periglacial and aquatic
environments of northern James Ross Island, Antarctic Peninsula ................. 22
2.2.1 Changes of land-based glaciers on the Ulu Peninsula during the last
three decades ................................................................................................................................... 22
2.2.2 Changes of the permafrost active layer and its thermal properties on
the Ulu Peninsula ............................................................................................................................ 24
2.2.3 Effects of abiotic factors and climatic conditions on the cyanobacterialmicroalgal
community in two shallow lakes on the Ulu Peninsula .......................... 26
2.3 Climate variability and its impact on terrestrial ecosystems in the Arctic
Svalbard Archipelago .............................................................................................................................. 28
2.3.1 Climate conditions and weather patterns in central Spitsbergen ............................. 28
2.3.2 Modelling of the surface wind characteristics in the complex
topography of central Spitsbergen ......................................................................................... 31
2.3.3 Impacts of simulated warming on hummock tundra with cyanobacteria
mats in central Spitsbergen ........................................................................................................ 33
3 Summary and Outlook ...................................................................................................................................... 37
4 References .................................................................................................................................................................... 40
5 Supplements ............................................................................................................................................................... 49
6
List of Papers
Paper1 Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2009. Prediction of
erythemally effective UVB radiation by means of nonlinear regression model.
Environmetrics, 20, 633–646.
Paper2 Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2010. Estimation of solar UV
radiation in maritime Antarctica using a nonlinear model including cloud effects.
International Journal of Remote Sensing, 31, 831–849.
Paper3 Láska, K., Budík, L., Budíková, M., Prošek, P., 2011. Method of estimating solar UV
radiation in high-latitude locations based on satellite ozone retrieval with an improved
algorithm. International Journal of Remote Sensing, 32, 3165–3177.
Paper4 Petkov, B.H., Láska, K., Vitale, V., Lanconelli, C., Lupi, A., Mazzola, M., Budíková, M., 2016.
Variability in solar irradiance observed at two contrasting Antarctic sites. Atmospheric
Research, 172-173, 126–135.
Paper5 Engel, Z., Nývlt, D., Láska, K., 2012. Ice thickness, areal and volumetric changes of Davies
Dome and Whisky Glacier (James Ross Island, Antarctic Peninsula) in 1979-2006. Journal
of Glaciology, 58, 904–914.
Paper6 Hrbáček, F., Láska, K., Engel, Z., 2015. Effect of Snow Cover on the Active-Layer Thermal
Regime – A Case Study from James Ross Island, Antarctic Peninsula. Permafrost and
Periglacial Processes.
Paper7 Elster, J., Nedbalová, L., Vodrážka, R., Láska, K., Haloda, J., Komárek, J., 2016. Unusual
biogenic calcite structures in two shallow lakes, James Ross Island, Antarctica.
Biogeosciences, 13, 535–549.
Paper8 Láska, K., Witoszová, D., Prošek, P., 2012. Weather patterns of the coastal zone of
Petuniabukta, central Spitsbergen in the period 2008-2010. Polish Polar Research, 33,
297–318.
Paper9 Láska, K., Chládová, Z., Hošek, J., submmited. High-resolution numerical simulation of
summer wind field over complex topography of Svalbard archipelago. Quarterly Journal
of the Royal Meteorological Society.
Paper10 Elster, J., Kvíderová, J., Hájek, T., Láska, K., Šimek, M., 2012. Impact of warming on Nostoc
colonies (Cyanobacteria) in a wet hummock meadow, Spitsbergen. Polish Polar Research,
33, 395–420.
7
1 Introduction
The Intergovernmental Panel on Climate Change (IPCC), the leading international body for
the assessment of climate change, in its Fifth Assessment Report (IPCC, 2013) assumed a
gradual increase in global surface temperature of the Earth (Fig. 1), which will very
probably have serious impacts on the landscape and its components. The impacts of
climate change, however, are reflected in a different way in the functioning of natural
ecosystems and their adaptations in low, middle and high latitudes. In the case of the
polar regions, the substantial environmental impacts of climate change and climate
variability show significant regional differences both within and between the Arctic and
Antarctic regions, and enormous complexity in the interactions between the abiotic and
biotic components of the ecosystem (e.g.; Vincent, 2000; Convey, 2003; Comte et al.,
2007).
Fig. 1 Surface air-temperature trends in the period 1958–2008 based on the modified map
published by Mulvanay et al. (2012). Two insets show the area of interest described in the next
part of the thesis.
Both the Arctic and Antarctic play a very important role in many environmental
aspects and processes on the Earth. One of them is the impact on the global climate
system, which is driven by solar radiation and its redistribution through regional and
vertical energy fluxes and the poleward transport of heat by atmospheric and oceanic
circulation (Barreiro et al., 2011). The high latitude locations can receive a large amount
of solar radiation at the surface during summer, most of which is reflected back to space,
due to the high albedo (reflectivity) of the snow and ice surfaces. The cryosphere and its
particular components – ice sheets, ice shelves, glaciers, frozen ground, sea ice and snow –
therefore, play a crucial role in the global climate system. The areal extent, occurrence and
lifespan of each component of the cryosphere are very different and predispose their
significance and roles in the system. The East Antarctic Ice Sheet represents a huge and
8
relatively stable mass of ice with a very high albedo of 87% (Wang and Zender, 2011). The
Antarctic, therefore, has a specific atmospheric circulation, which also distinctly affects
the mid latitudes of the Southern Hemisphere. By contrast, significant albedo reduction
can be found in the coastal zone of the Antarctic, with a mix of ice floes with a typical
albedo of 70–80% and open water with an albedo of 10–15% (Kuipers Munneke et al.,
2011). One of the major difference between the Arctic and Antarctic is, therefore, related
to albedo and its temporal variation, as linked to changes in sea ice extent and thickness in
the Arctic Ocean and the Southern Ocean, or due to persistent winter snow cover in the
northern parts of Scandinavia, Greenland, North America and Russia. On the other hand,
the polar regions are closely coupled to the rest of the climate system through global
thermohaline circulation, with the oceanic system providing a direct link between the
Arctic and Antarctic (Gregory, 2005; Marshall et al., 2014). In contrast to the ocean, the
atmosphere is able to relatively quickly respond to changes in the high or low latitude
heating rates, with the mean flow changing in the order of days to years (Hansen et al.,
2011).
Over the last decades, the polar regions have experienced greater atmospheric
warming than the global average (e.g.; Turner et al., 2005; Comiso et al., 2008) and some
remarkable environmental changes, such as the ozone depletion above the Antarctic, the
large sea ice loss in the Arctic ocean, and the disintegration and collapse of floating ice
shelves around the Antarctic (Vaughan et al., 2003; Newman et al., 2004; West et al.,
2013). The thawing of permafrost, which has the potential to release large amounts of
carbon into the atmosphere (McGuire et al., 2009), is another important atmosphere–
cryosphere interaction influencing the global climate system. Moreover, there is strong
evidence of the ongoing impacts of climate change on species and communities of
terrestrial ecosystems (Vincent, 2000; Callaghan et al., 2004; Komárek et al., 2008).
However, the combination of ongoing processes in the Antarctic is rather different
from that of the Arctic, because of different responses to anthropogenic forces and
mechanisms linked with the positive snow and sea ice albedo feedback amplifying the
warming signal (Holland and Bitz, 2003; Serreze and Barry, 2011). The polar regions are
also the area for which the largest future warming over the next century is predicted if
greenhouse gas concentrations continue to rise (e.g.; Stoker et al. 2013; Hansen et al.,
2013). The impacts of climate change in the polar regions over the next 100 years will
exceed the impacts forecasted for many other regions, producing feedback that will have
globally significant consequences.
This study, consisting of ten scientific papers, investigates the responses of biotic and
abiotic environments on the climate variability that have taken place on different time
scales, from the last few decades to individual years, in selected areas of the Arctic and
Antarctic. We focus on the regions of the Antarctic Peninsula and central Spitsbergen in
the Arctic Svalbard Archipelago. Both of these areas have experienced significant
atmospheric warming over the last several decades (Turner et al., 2005; Nordli et al.,
2014) followed by a cascade of environmental changes and impacts on the terrestrial and
aquatic ecosystems. Hence, the evaluation of the impacts of climate conditions and climate
variability on geo-ecosystem functioning have become the other main objective of the
thesis.
9
2 Results
In the initial phase of our climate and environmental research in the region of the
Antarctic Peninsula and the Svalbard Archipelago, we had to solve an urgent need for data
collection, thus carrying out in situ observations and measurements using modern
instrumentation and techniques. These observations have begun to provide us with the
accurate measurements of atmospheric conditions and other abiotic factors in the areas,
whereas the early observations were sparse and insufficient in terms of spatial resolution,
accuracy, international standards and related approaches. We had to deal with several
other important tasks that had emerged from the former subjects of our scientific
activities and the first research projects. To achieve these requirements, we planned the
following:
1. to establish a long-term monitoring programme focusing on the main abiotic
components of the polar environments: atmosphere, glaciers, permafrost,
lacustrine geo- and eco-systems and their recent changes,
2. to provide quantitative data on local climate conditions and the parameters of a
variety of landscape types and environments at the selected representative sites,
3. to install data acquisition systems and devices adjusted for autonomous yearround
operation in harsh weather conditions,
4. to keep quality assurance, maintenance, methods and protocols for data gathering
in accordance with the international standards and procedures published by the
Commission for Instruments and Methods of Observations
(www.wmo.int/pages/prog/www/IMOP/) and Global Climate Observing System
(http://www.wmo.int/pages/prog/gcos) under the World Meteorological
Organization, the World Glacier Monitoring Service (www.wgms.ch), and the
International Permafrost Association (www.permafrost.org),
5. to gather representative data sets, which can be further used for comparative
studies and the evaluation of regional differences and impacts of climate
variability.
In order to study response of the constituents of polar environments on the climate
variability as well as the fulfilment of the above-mentioned objectives, we established the
long-term observational and monitoring network, consisting of an array of automatic
weather stations and measuring devices at five different localities: four of them are
situated in the region of the South Shetlands and north-eastern Antarctic Peninsula; the
last one is located in the central part of the Svalbard Archipelago.
At the beginning, meteorological observations and measurements were established
at two research stations in the Antarctic Peninsula region. In the period 1994–1998, the
monitoring of UV radiation, surface energy balance, snow distribution and the thermal
regime of the active layer was carried out at the coastal vegetation oasis, near the Polish
Arctowski Station (Fig. 2) on King George Island, South Shetland Islands (see e.g.; Caputa
et al., 1997; Kejna and Láska, 1999a, b; Prošek et al., 2000, 2001). As for the second site,
we selected the Ukrainian Vernadsky Station (formerly the British Faraday Station),
situated on Galindéz Island off the western coast of the Antarctic Peninsula (Fig. 2). The
10
atmospheric and climate research at Vernadsky Station was carried out in the period
2002–2006. Part of the long-term atmospheric and climatic monitoring programme is still
being carried out at Vernadsky Station in a framework of cooperation among the National
Antarctic Scientific Center of Ukraine in Kiev, the National Taras Shevchenko University of
Kyiv, Ukraine and Masaryk University, Brno, Czech Republic. The results of the solar
radiation measurements and comprehensive analysis of the relationships among the
stratospheric ozone, cloudiness and solar UV radiation were summarized [Paper 1–2].
Fig. 2 The location of the selected research stations (red dots) and the study site in the Antarctic
Peninsula region. The inset maps (a, b) show the position of the Antarctic Peninsula and James Ross
Island. The inset on the satellite map of James Ross Island (c) represents the area shown in detail in
Fig. 3.
Further prospects for continuing the atmospheric and environmental monitoring
programme emerged during the construction of the Czech Johann Gregor Mendel
(hereafter Mendel) Station in 2005–2006. The station was established on the northern
coast of James Ross Island (Fig. 2), which belongs to a transitory zone between the
maritime Antarctic and continental Antarctic regions. More than 80% of the island surface
is covered with ice and only the northernmost part of the island (Rabassa et al., 1982), the
Ulu Peninsula, is significantly deglaciated, representing one of the largest ice-free areas in
the Antarctic Peninsula region (Fig. 3). From this point of view, it provides a unique place
for studying many aspects of climate variability, land-atmosphere interaction and the
impacts of climatic change on terrestrial and aquatic ecosystems. Another important
moment related to the starting of the measurements was connected with two research
11
projects that were undertaken in the period 2007–2009 under the auspices of the Czech
Ministry of Education, Youth and Sports and the Czech Science Foundation.
However, the only available meteorological data and other records for the newly
investigated area of the Ulu Peninsula were being sourced from regular weather
observations at Esperanza and Marambio stations, situated 60 km and 80 km away,
respectively, from Mendel Station. Therefore, new monitoring systems consisting of
automatic weather stations and data acquisition systems were installed in the study area
Fig. 3 The location of the automatic weather stations, selected glaciers and other sampling sites on
the Ulu Peninsula, which are the subject of the interest and publications described in this thesis.
The topography of the Ulu Peninsula is based on the map of the Czech Geological Survey (2009).
12
during the first years of the station’s operation. Since 2007, an integrated
multidisciplinary study of glaciers, permafrost, terrestrial and aquatic ecosystems has
been carried out in the northern part of the Ulu Peninsula. Nowadays, the observation
network consists of eleven automatic weather stations and twenty other systems like
smart sensors for permafrost active layer or lacustrine monitoring. The results of the
atmospheric and climate research on the Ulu Peninsula represent the most important part
of the thesis and have been summarized [Paper 3–7].
Fig. 4 The location of the study site and automatic weather stations (blue triangles) on the western
coast of Petuniabukta in the northern part of Billefjorden, central Spitsbergen. The inset maps (a, b)
show the position of Svalbard Archipelago and the Billefjorden area. The automatic weather
stations mentioned in the text are indicated as follows: 1 – raised marine terrace, 2 – wet
hummock tundra, 3 – foreland of Hørbyebreen glacier and 4 – top of Mumien peak. The modified
map of Petuniabukta is based on Svalbardkartet data, Norwegian Polar Institute.
The Czech long-term monitoring programme in the Arctic was prepared under the
auspices of the International Polar Year (IPY 2007–2008) and three research projects
funded by the Czech Ministry of Education, Youth and Sports. The initial phase of the
monitoring programme was carried out over four summer seasons (2007–2010) by three
institutions: the University of South Bohemia in České Budějovice, Masaryk University in
Brno and the Institute of Botany of the Academy of Sciences of the Czech Republic. From
the beginning, the cooperation with the Adam Mickiewicz University in Poznań, the
Norwegian Polar Institute and the University Centre in Svalbard, which has investigated
the Svalbard Archipelago for many years, was very important from two viewpoints: the
scientific coordination and the field logistical support, both of which were highly
13
beneficial for the Czech research team. The project was proposed with the aim of
exploring the diversity of both climates and ecosystems at the landscape scale within the
Arctic region by applying new rigorous measurements of key biological and physical
variables and processes at multiple circumpolar Arctic observation sites (Elster and
Rachlewicz, 2012).
The area of Petuniabukta in the central part of Spitsbergen, Svalbard Archipelago
would become the most suitable place for the implementation of the complex monitoring
programme and fulfilment of individual project tasks. Petuniabukta is a northwardoriented
bay that closes the northern part of Billefjorden (Fig. 4). Due to the lack of yearround
meteorological and environmental data for Petuniabukta and the large distance
between Petuniabukta and the nearest weather station at Svalbard Lufthaven in
Longyearbyen, located 56 km away, the new monitoring network was established in the
period 2007–2009. The experience and technical solutions obtained during the previous
Antarctic expeditions proved the importance of carrying out comprehensive
measurements using autonomous measuring systems and weather stations with yearround
operation. Since 2014, the network has consisted of eight automatic weather
stations and ten other monitoring devices for gathering meteorological data and
important abiotic parameters from the terrestrial area of Petuniabukta. The summary of
Paper 8–10 is based on the meteorological data recorded in the area of Petuniabukta and
the evaluation of local climate conditions in relation to large-scale weather systems, the
complex topography, and the pronounced effect on biotic components of the Arctic
ecosystems.
2.1 Solar radiation variability in high-latitude locations
Solar radiation provides the Earth with energy that generates a variety of dynamic
processes and triggers numerous physical and chemical reactions in the atmosphere,
biosphere, hydrosphere and cryosphere. For that reason, the measurements and
modelling of solar radiation is considered an important issue in the natural sciences,
examining environmental problems. In the last few decades, interest in solar ultraviolet
(UV) radiation has increased within the scientific community as it has been found that
there is a broad variety of effects on biological processes, including harmful effects on
DNA, human skin and the overall immune system (e.g.; De Fabo et al., 1990; Schmalwieser
et al., 2002a; Weihs et al., 2013).
Solar UV radiation depends on many atmospheric factors. Among the most important
are ozone, clouds and aerosol load. The first observation of ozone depletion over
Antarctica was documented at the beginning of the 1980s (Farman et al., 1985). Over the
past three decades there has been a large-scale depletion of stratospheric ozone, except in
the year 2002, when an unprecedented sudden major stratospheric warming resulted in
an almost no-ozone-hole episode (Varotsos, 2004). Several studies indicate that both
atmospheric circulation and stratospheric temperature play an essential role in the
intensity of ozone losses over both hemispheres (Hofmann et al., 2009). Apart from that,
the consecutive deceleration of ozone losses over Antarctica during the last decade was
reported (Kravchenko et al., 2009). The most recent papers indicate that the Antarctic
14
ozone hole could return to 1980 levels by about the years 2050 to 2060 (Hassler et al.,
2011; Douglass et al., 2014; Chipperfield, 2015). As ozone is an important greenhouse gas,
changes to its distribution in the stratosphere can affect not only the climate of the
stratosphere but also the climate at the Earth's surface. Therefore, it is important to
understand the geographical distribution of stratospheric ozone over Antarctica, the
changes in solar UV radiation, and the processes and possible impacts on the Earth's
climate system.
2.1.1 Modelling of solar ultraviolet radiation in the Antarctic Peninsula region
[Paper 1]
In order to study and evaluate the effects of atmospheric factors on UV radiation, different
types of instrumentation and models were developed. Models of UV radiation prediction
can be divided into three main groups: (1) Radiative transfer models, including multiscattering
(e.g.; Stamnes et al., 1988; Madronich et al., 1998; Berk et al., 2005; Mayer and
Kylling, 2005), (2) Physical models with a simple parameterization for district
wavelengths (e.g., Schippnick and Green, 1982 or Schauberger et al., 1997) and (3)
Statistical models (Canadian model – Burrows et al., 1994; CHMI model – Vanicek, 1997;
ETH Zürich model – Renaud et al., 2000). Particularly the radiative transfer models
require extensive computation and many assumptions about the atmospheric parameters,
which are generally difficult to obtain in Antarctica. The simplest group of statistical
models uses a regression equation, which is obtained by fitting the observed UV radiation
to the selected file of input parameters. However, most of the UV radiation models were
designed and parameterized outside the Antarctic stratospheric depletion area and have
to be verified in terms of the treatment of ozone, clouds and their properties at the edge
region of the polar vortex.
Therefore, we decided to develop our own nonlinear regression model with a
hyperbolic transmissivity function [Paper 1], which takes into account both the character
of the available data and the specific conditions in the Antarctic Peninsula region. In the
initial stage of our work, we benefited from the experience and cooperation of the
Institute of Medical Physics and Biostatistics, University of Veterinary Medicine in Vienna,
which developed and operationally used the global UV forecast model (Schauberger et al.,
1997). The Austrian forecast model was a fast spectral model based on the measurements
of Bener (1972) at Davos (Switzerland) following a suggestion of Diffey (1977). The input
parameters were zenith angle, distance between the Earth and Sun, altitude as well as a
global total ozone column dataset prepared by a simple scheme (Schmalwieser et al.,
2002b), which was considered to be appropriate for the UV Index forecast. The validation
was done by comparing forecast values to ground-based measurements gathered by
Model 501 UV broadband radiometers (Solar Light, Inc., USA) at six sites on four
continents, ranging from 67°N to 60°S. One of the southernmost locations used for model
validation was the Polish Arctowski Station on King George Island (Fig. 2) using the
measurements taken by our team in the period 1996–1998 (Prošek et al., 2001;
Schmalwieser et al., 2002a). During the validation process, special attention was put on
the frequency of overestimation of the modelled irradiances. We found that the hit rate of
the forecast was 48.1%, while a large overestimation of the modelled UV Indexes reached
15
up to 50% during an ozone hole period and due to the significant effect of cloudiness
(Schmalwieser et al., 2002b).
The newly developed nonlinear regression model improved the way to estimating
erythemally weighted (EW) solar radiation incidence on the Earth's surface [Paper 1].
Instead of the widely used ozone transmissivity exponential function (e.g., Kondratyev
and Varotsos, 1996), we proposed a new approach based on a quantum transmission
model using the hyperbolic attenuation of the solar radiation. For the purpose of the
model, the ozone layer was divided into imaginary layers to solve the issue of absorption
by the ozone more precisely. In this approach, the radius of the Earth, the mean height of
the ozonosphere, the daily means of the solar elevation angle and total ozone column
were used for resolving the geometrical approximation of the actual sunlight path through
the spherical layer of atmosphere. The other input parameters were extraterrestrial
short-wave and EW radiation, and short-wave downwelling (SWD) radiation, which was
observed at the particular site [Paper 1].
The model was successfully validated at the Ukrainian Vernadsky Station for the
period 2002–2003. Systematic long-term measurements of the intensity of solar radiation
and important constituents of its spectrum were carried out using two types of
radiometers: the CM6B pyranometer (Kipp & Zonen, the Netherlands) and Model 501
version 3 UV-Biometer, manufactured by Solar Light, USA. The other input parameters
were retrieved from the ground-based ozone measurements of the Dobson
spectrophotometer and satellite-based measurements of the Total Ozone Mapping
Spectrometer [Paper 1]. The quality of the nonlinear model was examined using the mean
average prediction error (Von Storch and Zwiers, 2002), which ranged between 7.4% and
9.6%. These findings showed that the model outputs depend significantly on the seasonal
changes of the total ozone column as well as on the source of the ozone data, i.e. the
Dobson spectrophotometer or the Total Ozone Mapping Spectrometer. Although the
model differences according to ozone inputs did not exceed 3%, the prediction errors
occasionally reached 10%, due to differences in absorption and the scattering of UV
radiation given by the type of approximations used in our model [Paper 1].
Overall, our results were in agreement with the findings of Weihs and Webb (1997),
Koepke et al. (1998) and De Backer et al. (2001). They found differences between the
measurements and models for the middle-latitude regions in the range of 5 to 10% with
statistical uncertainty of 2–3%. Further improvement of the nonlinear model and several
case studies in the Antarctic Peninsula region are shown in the next sections.
2.1.2 Effect of ozone and cloudiness variation on solar ultraviolet radiation on the
western side of the Antarctic Peninsula [Paper 2]
The spatiotemporal variability of UV radiation and total ozone have become a high
priority topic in scientific research. In the high latitude locations of the Southern
Hemisphere, stratospheric ozone, solar elevation angle, clouds and surface albedo are the
main factors influencing solar UV radiation at ground levels (Nunez et al., 1997; Lubin and
Morrow, 2001; Labow et al., 2011). In contrast to the central part of continental
Antarctica, the short-term variability in the total ozone column in the coastal regions is
16
mainly related to atmospheric dynamics, while short-term variability in UV radiation is
mainly controlled by cloudiness (Bernhard, et al., 2010). However, cloudiness may often
mask the effect of stratospheric ozone changes on the incident UV radiation and could
either significantly reduce or even increase the UV flux reaching the ground (Calbó et al.,
2005).
The main aim of Paper 2 was therefore 1) to analyse the variability of cloudiness and
stratospheric ozone distribution at the edge of the polar vortex, 2) to evaluate the
influence of these factors on incident solar UV radiation, and 3) to test the performance of
the nonlinear model under various amounts of cloud cover. This work was a continuation
of our previous research accomplished in the region of the Antarctic Peninsula since 1995
(Prošek et al., 2001; Schmalwieser et al., 2002a, 2002b). For the purpose of this study, the
nonlinear regression model with a hyperbolic transmissivity function was used [Paper 1].
The measurements of SWD and EW irradiance were carried out at Vernadsky Station,
using a CM11 pyranometer (Kipp & Zonen, The Netherlands) and a broadband UVBiometer
Model 501A (Solar Light, USA) over a 3-year period. To improve the quality of
the total ozone column data, the exploitation of the ground-based ozone measurements
made by the Dobson spectrophotometer and satellite-based measurements of the Total
Ozone Mapping Spectrometer (TOMS) was performed. In addition, the daily mean
cloudiness based on 3-hours synoptic observations of total cloud amount at Vernadsky
Station was used for the purpose of further evaluation.
High seasonal variability of the SWD and EW irradiances and total ozone column at
Vernadsky Station were found in the period 2002–2005 [Paper 2]. Ozone column
variation, expressed by the variation coefficient, decreased significantly from spring
(~30%) to autumn (~8%), as a consequence of ozone depletion weakening. However, the
ozone depletion period (September–October) was not clearly pronounced on the EW
radiation due to a high level of cloudiness and the low solar elevation angle in this time.
The high occurrence of cloud cover between 7–8 oktas (64% of all cases) coincided with
cyclogenesis in the circumpolar trough near 60° S (Turner et al., 1998). The mean
cloudiness at Vernadsky reached 6.9 oktas with the maximum variation during summer
and minimum during autumn and winter. The seasonal fluctuation is consistent with the
occurrence of a layer of low stratiform clouds over the high-latitude of the Southern
Hemisphere and has the largest incidence during winter (Klein and Hartmann, 1993).
Results in Paper 2 demonstrated that while the variation in cloudiness determined shortterm
fluctuations of the radiation fluxes, ozone declines caused solar radiation changes of
prolonged periodicity in the second half of the year. Therefore, the EW radiation was the
greatest in November and December, when the low total ozone column coincided with
relatively large solar elevation angles. The largest EW daily doses of 5.2–5.4 kJ m-2 were
measured on 5 December 2003 and 25 November 2003. These values were lower than the
largest dose of 8.8 kJ m-2 recorded at Palmer station in the western Antarctic Peninsula on
10 November 1997 and 7 December 1998 (Bernhard et al., 2005). This illustrates the
emerging evidence that ozone depletion peaks were highly pronounced in the late 1990s
and the first decade of the 21st century (Newchurch et al., 2003; Eleftheratos et al., 2015).
17
In the last part of Paper 2, the modelled EW irradiance was evaluated against the
cloudiness. Generally, the modelled daily irradiance showed very good agreement with
the observed values (R2 = 99.2%). Two different statistical parameters were used for the
model validation: 1) residuals of the EW irradiance in kJ m-2, and 2) mean average
prediction error expressed in percents. The highest EW residuals (–0.18 and +0.24 kJ m-2)
occurred from September to March, independently of the ozone depletion period. The
largest variation of EW residuals, expressed by the standard deviations, occurred around
the summer solstice (November–January) in connection with a higher variation of
cloudiness. Nevertheless, the occurrence of EW residuals of higher than ±0.15 kJ m-2
represented only 1% of all cases. The mean average prediction error reached 4.4% over
the studied period. Interestingly, such small mean average prediction errors were
reported during high summer when they ranged from 2.6 to 3.5%. Differences in the
frequency distribution of the EW residuals between the individual seasons of 2002–2005
were found to be small and insignificant [Paper 2]. The largest EW residuals were found in
partly cloudy conditions (0–4.0 oktas) with high variability of cloud types, and their
thickness and brightness during a day. In this case, the mean average prediction error
rose up to 5.9%. This can be partly explained by the fact that model parameterisation was
done for all sky conditions, under which daily mean cloudiness reached 6.9 oktas.
Furthermore, the partly cloudy category included thin and moderately thick clouds in
which the strong Rayleigh scattering of short wavelengths and consequent wavelengthdependent
directional distribution of the sky radiance were well documented (Kylling et
al., 1997; Bernhard et al., 2004; Lindfors and Arola, 2008).
The results in Paper 2 were difficult to compare to other studies, since only few
Antarctic stations equipped with the same instruments carried out both the
measurements and estimates of the EW radiation intensities. Generally, the cloud effects
are considerably larger compared to the interior part of Antarctica (e.g. South Pole) due to
optically thicker clouds in the coastal areas (Bernhard et al., 2006; Bromwich et al., 2012).
The results are also in agreement with the findings of De Backer et al. (2001),
Schmalwieser et al. (2002b) and Cordero et al. (2013), showing the sensitivity of EW
irradiances to the transmission factors of the clouds, subsequently, cloud types and their
base heights.
2.1.3 Evaluation of solar ultraviolet radiation on the eastern side of the Antarctic
Peninsula – the influence of ozone, cloudiness and surface reflectivity [Paper 3]
The Antarctic Peninsula mountains are an important orographic barrier to the persistent
Southern Ocean westerlies (King and Turner, 1997). The western Antarctic Peninsula,
therefore, has a polar maritime climate, dominated by the relatively warm and ice-free
Bellingshausen Sea, whilst the eastern Antarctic Peninsula and James Ross Island have a
rather polar continental climate, influenced by the Weddell Sea, which is mostly ice-bound
for much of the year (Vaughan et al., 2003). The orographic influence causes spatial and
temporal differences in the cloudiness and other meteorological variables shaped by
synoptic- and meso-scale circulation systems on the eastern (leeward) side of the
Antarctic Peninsula (King and Comiso, 2003). In addition, the linkage between the positive
18
Southern Annular Mode (Marshall et al., 2006) and the increase of cloud cover was found
on the windward side, while there was a decrease on the northern leeward side of the
Antarctic Peninsula (Van Lipzig et al., 2008). However, the specific features of this region,
such as the seasonal variations in the total ozone column, cloudiness and sea ice extent,
may complicate the forecasting or the reconstruction of the incident solar UV radiation.
Paper 3 presents the results from the first three years of solar radiation
measurement at Mendel Station, for which a new version of the nonlinear regression
model with an improved algorithm for estimation of EW radiation was tested.
Furthermore, the model results are discussed based on satellite measurements of total
ozone column and effective surface reflectivity at 360 nm (hereafter, reflectivity) for the
study site. The measurement of solar radiation at Mendel station was performed with an
identical set of instruments (Kipp & Zonen CM11 pyranometer and Solar Light 501A UVBiometer)
as at Vernadsky station. To protect them from snow and ice accumulations, the
instruments were equipped with Kipp & Zonen CV2 heater ventilation units. The
radiometers were regularly calibrated at the Solar and Ozone Observatory of the Czech
Hydrometeorological Institute in Hradec Králové, Czech Republic. There are not any
ground-based ozone measurements available at Mendel station. Therefore, total ozone
column and reflectivity data were obtained by the Ozone Monitoring Instrument (OMI)
onboard the EOS-Aura spacecraft for the geographical coordinates of Mendel Station
(http://avdc.gsfc.nasa.gov/). As reported by many authors, the enhanced algorithms
applied to the OMI measurements should have a root-mean squared error of 1-2%,
depending on the solar zenith angle and aerosol amount (e.g.; Ahmad et al., 2004; Jaross
and Warner, 2008).
The analysis of SWD and EW irradiances in the period 2007-2009 confirmed a large
seasonal and day-to-day variability of both parameters in relation to the solar elevation
angle and cloudiness [Paper 3]. The highest daily SWD and EW irradiances occurred from
the second half of November to mid December, and reached up to 35 MJ m-2 and 5 kJ m-2,
respectively. The total ozone column ranged from 112 DU to 419 DU in spring 2007, while
in 2008 it varied from 135 DU to 345 DU, due to a weakening of the polar vortex and much
smaller area of ozone depletion (Eleftheratos et al., 2015). Similarly, the variation
coefficient of the total ozone decreased more slowly from spring (~26%) to autumn
(~8%). The influence of ozone depletion on EW radiation was overlaid by the effect of
cloudiness with a high optical depth. The seasonal occurrence of low stratiform clouds
with spring minimum and autumn maximum values on the eastern side of the Antarctic
Peninsula was also reported by Verlinden et al. (2011). Therefore, daily EW radiation
measured within the ozone depletion period (September–October) ranged between 3 and
4 kJ m-2 under clear sky conditions. These results are in agreement with Paper 2, while the
earlier observations carried out at Palmer station showed an almost twice as high an EW
irradiance level due to large ozone depletion (Bernhard et al., 2005). It should be noted,
however, that ozone, cloudiness and UV irradiance during the 1990–2011 period did not
show any significant long-term trends over the southern high latitudes (Eleftheratos et al.,
2015).
19
The effects of the atmospheric factors on the prediction of EW radiation were
evaluated using a nonlinear regression model [Paper 1]. Model parameterization and
additional computation were done for 539 days in the period 2007–2009. To examine
whether the sensitivity of input parameters affects the modelled outputs, the relative
prediction errors of EW radiation based on SWD irradiance and total ozone column input
data were calculated. We found that uncertainties in the satellite-based ozone
measurements influenced the estimation of EW irradiance four times more than the
uncertainty of the SWD measurements. Hence, the accuracy of the EW prediction in the
ozone depletion period may reach up to ±7% on average (±27% in the extreme case),
while outside of the ozone depletion period, the accuracy of prediction may significantly
improve for ±3.5% (±10% in extreme case).
Overall, the modelled daily EW irradiance based on the satellite-based data showed
very good agreement with the observed values (R2 = 98.6%). The EW residuals
(observation against the model) varied from –0.21 to +0.28 kJ m-2, with the highest
frequency of occurrence from October to February. The mean average prediction error
reached 6.0% over the studied period. This can also be compared with the previous
results of Paper 1, where both ground and satellite ozone measurements were applied.
For the satellite-based data, the mean average prediction error at Vernadsky station
ranged from 8.8% to 9.5%, depending on the season. The new modelled outputs were
further evaluated using satellite-based reflectivity and EUV residuals classified according
to the solar elevation angle and ozone depletion period. The highest reflectivity (85–
100%) and its smallest variation were closely connected to the maximum sea-ice extent
along the Antarctic Peninsula from August to October each year. The summer shrinkage of
the sea ice consequently caused the high variability of the surface reflectivity from
November to February. However, it has been proven that seasonal variation of the
satellite reflectivity had only a marginal effect on the predicted EW irradiance. The mean
average prediction errors from November to January were below 4.4%, while the largest
errors of about 7% occurred during spring ozone decline.
This confirms the findings by Bernhard et al. (2005), who reported that particularly
high UV fluxes and their large variation were observed at Palmer station during the years
when the polar vortex became unstable and air masses with low ozone concentration
moved toward the Antarctic Peninsula. In combination with the relatively high solar
elevations in October and November, it led to UV intensities exceeding summer levels in
the mid latitude regions (Bernhard et al., 2010). Nevertheless, the nonlinear regression
model has provided high-quality estimates of the EW irradiance at ground level [Paper 3],
as documented by the sensitivity analysis and fact that the EW residuals of higher than
±0.15 kJ m-2 represented less than 5% (26 days) of all cases.
2.1.4 Effects of atmospheric and environmental factors on solar radiation
variability at two contrasting Antarctic sites [Paper 4]
Besides the astronomical parameter, the atmospheric and environmental characteristics,
such as the atmospheric transmittance, mainly determined by the cloud cover features
and aerosol loadings, together with the surface reflectivity (albedo), significantly impact
20
the solar irradiance on the ground ([Paper 2]; Lee et al., 2015). Since the environmental
factors usually change significantly with the season, latitude and altitude, the distinct
difference between the transmittance characteristics of the atmosphere over the coastal
regions (Antarctic Peninsula) and interior Antarctic Plateau affect the spatiotemporal
variability of solar irradiance on the continent. It is worth noting that the atmosphere over
the Antarctic Plateau is slightly contaminated by aerosols and, as a result, their impact on
the solar irradiance variations in the deep continental zone is assumed to be of minor
importance (Tomasi et al., 2007).
In Paper 4, the features of solar radiation observed at two contrasting Antarctic sites
located in the coastal zone and highly-elevated interior plateau were analysed with the
aims of examining the role of the factors, impacting irradiance variability during the
daylight part of the year. In addition, the approximate quantitative estimates were
presented, outlining the contribution of these factors to the solar irradiance variations at
the different Antarctic regions. We compared the measurements of SWD and EW
irradiances measured at Mendel and Dome Concordia stations during 260 days of the
austral spring–summer months from 2007 to 2011. Concordia is situated on the Antarctic
Plateau in the opposite part of the continent at a distance of about 4600 km from Mendel
station. More details of the stations and measuring devices can be found in Paper 3 and
the study of Vitale et al. (2011). The surface reflectivity data for both stations were
provided by version 3 of OMAERUVG daily global data products (level 2G) retrieved from
the Ozone Monitoring Instrument (OMI, 2014). The cloud cover fraction was taken to
characterise the atmospheric transmittance and was derived from the Atmospheric
Infrared Sounder L3 cloud property retrievals (Platnick et al., 2003). The values of the
reflectivity and cloud cover fraction were extracted for an area of ± 1° in both latitude and
longitude around each of the stations. Since the ground-based measurements give the
variations in the surface solar irradiance caused by all the impacting factors, it was
considered that a realistic representation of the seasonal component in the observed
variations can be determined through model evaluations performed for clear sky
conditions. The Tropospheric Ultraviolet-Visible (TUV) radiative transfer model
(Madronich and Flocke, 1997) was therefore used to achieve satisfactory results within a
comparatively short computation time.
The surface reflectivity at Mendel station ranged between 12% in summer (January)
and 96% in the late winter (August). In contrast, the permanent snow-covered surface at
Concordia station was characterised by a high albedo varying from 88% to 100% during
the year. The mean cloud cover fraction reached a value of about 63% due to prevailing
overcast conditions at Mendel station, while the cloudless atmosphere and low aerosol
content over Concordia were represented by a mean cloud cover fraction of 4%.
Therefore, the SWD irradiance observed at Concordia was strongly correlated with the
maximum daily value of the solar elevation angle (R2 = 96%), while a similar strong link
for the EW irradiance was less remarkable (R2 = 74%), likely because of the additional
effect of the ozone column variations during the spring period. In contrast, both EW and
SWD components registered at Mendel station did not show any functional relationship
with the solar elevation, due to low-level cloud occurrence. This is consistent with the
21
studies by Verlinden et al. (2011) and Bromwich et al. (2012), analyzing the occurrence
and vertical distribution of the tropospheric clouds and their seasonal changes in the
Antarctic Peninsula region.
Overall, the seasonal irradiance variations, determined by the solar elevation trend,
were clearly seen at Concordia station, where this astronomical factor caused 94% of the
irradiance changes, while the cloud cover fluctuations reduced the solar radiation by 5%
on average. On the contrary, the seasonal trend observed at Mendel, which caused about
71% of irradiance variations, was significantly masked by the cloudiness, aerosols and
albedo changes that contributed an average decrease in the solar irradiance of 25% for
EW and 29% for SWD irradiance [Paper 4]. However, slightly different figures have been
reported for Palmer station situated on the western side of the Antarctic Peninsula.
Exploiting the ground-based measurements, Bernhard et al. (2005) reported higher
attenuation of UV irradiance at 345 nm due to clouds between 28% (October and
November) and 42% (February) compared to clear-sky levels. It also proves that
significant differences in the cloud cover fraction and cloud thickness can be found
between the western and eastern sides of the Antarctic Peninsula (Verlinden et al., 2011).
Moreover, the large east-west gradient in upper-level cloud incidence was reported in the
region of the upwind (western) side of the Antarctic Peninsula. Considering that the
seasonal trend of solar elevation is a strongly determined parameter, it can be concluded
that the radiation on the interior Antarctic Plateau should be much easier to predict than
in the coastal region, where the day-to-day variability in cloud cover, together with the
changes in aerosol concentrations and surface albedo have a higher weight [Paper 4].
The ozone column was found to be the subject of similar change patterns at both
stations over the studied period, reducing the EW irradiance by nearly 46% with respect
to the corresponding EW calculated for the minimum value of ozone column at each site. It
is worth mentioning that for certain sub-periods, Mendel station was out of the ozone
depletion area over the Antarctic continent during the spring time, due to its lower
latitude with respect to Concordia. For instance, a total ozone of 175 DU was registered at
Mendel station on 23 November 2007, while the corresponding value was 330 DU at
Concordia. A similar event was observed on 24 November 2008, when the ozone column
at Mendel station was 173 DU versus 335 DU at Concordia. Finally, the sensitivity of
incident UV irradiance to the variations in the ozone column was further quantified by the
radiation amplification factor (Madronich, 1993; [Paper 4]). The radiation amplification
factors for both the Mendel and Concordia stations were very close to each other, with the
values of 0.92 and 0.88, respectively. The performed analysis allows the conclusion that
the ratio between the daily doses of EW and SWD irradiances, considered a function of the
ozone column, is able to provide a realistic estimate of the radiation amplification factor,
which quantifies the impact of the ozone on surface UV irradiance. Therefore, it can
improve the method of the assessment of solar UV radiation data retrieved from different
types of ground-based instruments placed in high latitude areas.
22
2.2 Impacts of climate variability on glacial, periglacial and aquatic
environments of northern James Ross Island, Antarctic Peninsula
2.2.1 Changes of land-based glaciers on the Ulu Peninsula during the last three
decades [Paper 5]
The largest annual atmospheric warming on the Earth for the second half of the 20th
century was reported from the western side of the Antarctic Peninsula, where the mean
air temperatures have increased at a rate of 0.56 °C per decade over the whole year and
by 1.09 °C per decade during the winter (Turner et al., 2005). The warming has taken the
form of a reduction in the number of extreme cold winters at the Antarctic Peninsula
(Turner et al., 2015). Temperatures on the eastern side of the Antarctic Peninsula have
risen most during the summer and autumn at a rate of 0.39 °C per decade due to the
strengthening of the westerlies that took place as the Southern Hemisphere Annular Mode
(SAM) shifted into its positive phase (Marshall, 2002). Stronger westerly winds now bring
warm, maritime air masses across the Antarctic Peninsula to the low-lying ice shelves on
the eastern side and surrounding islands with the Foehn Effect adding to the warming of
the air masses coming over the Antarctic Peninsula (Marshall et al., 2006).
This warming has coincided with the disintegration of ice shelves and the retreat of
many land-terminating glaciers in the region (e.g.; Skvarca et al., 1998; Cook et al., 2005;
Glasser et al., 2011). Land-based glacier mass loss has been most prominent at the
northern tip of the Antarctic Peninsula due to the pronounced acceleration of glacier flow
(Rignot et al., 2008). Detailed analyses of satellite imagery and some isolated massbalance
measurements suggest that small land-terminating glaciers are especially
sensitive to climate change (Rau et al., 2004). Estimation of ice sheet and glacier mass
balances and their potential contribution to sea level rise are dependent on a good
understanding of spatiotemporal changes in snow accumulation, comprising the net result
of precipitation, sublimation, snow drift, and melt. However, there are no representative,
or accurate measurements of precipitation along the eastern coast of the Antarctic
Peninsula due to its solid form and high wind speed, making a standard rain gauge
measurement impossible (Zvěřina et al., 2014; Nývlt et al., 2016). Because snow
accumulation and mass balance are difficult parameters to measure, it is important to use,
or combine direct geodetic measurements with advanced geophysical methods, remote
sensing data and atmospheric modelling. Moreover, only a few glaciers in the Antarctic
Peninsula and surrounding islands are monitored by field surveys (Simões et al., 1999;
Molina et al., 2007; Rückamp et al., 2011; Navarro et al., 2013; Marinsek and Ermolin,
2015). Furthermore, the determination of glacier bed topography, ice thickness and mass
balance changes may be used to detect ice dynamics and sensitivity to climate change as
well as to predict future glacier behaviour in different climatic scenarios (Fountain et al.,
1999; Rippin et al., 2011; Bamber and Aspinall, 2013).
The aim of Paper 5 was to identify the impact of climate change on the glaciers in the
north-eastern Antarctic Peninsula and to describe its areal and volumetric changes over
the last three decades. In Paper 5, two land-terminating glaciers on the Ulu Peninsula,
James Ross Island have been studied: Davies Dome, an ice dome originating on the surface
23
of a volcanic mesa with one prominent outlet flowing down to Whisky Bay, while Whisky
Glacier is a north-facing land-terminating valley glacier surrounded by vast areas of icecored
terminal moraine (Fig. 3). The changes in elevation, area and geometry of both
glaciers between 1979 and 2006 were derived from digital elevation models (Czech
Geological Survey, 2009) based on aerial photogrammetric imagery. Ice thickness was
measured during a field campaign in January and February 2010 using groundpenetrating
radar and post-processed in accordance with numerous descriptions of radar
investigations of glacier thickness (e.g.; Plewes and Hubbard, 2001; Navarro and Eisen,
2009). Glacier volume and glacier bed elevations were derived from these data.
In 2006, Davies Dome had an area of 6.5 km2 and lay at an altitudinal range of 0–514
m a.s.l. Whisky Glacier is a cold-based valley glacier, covering an area of 2.4 km2 in an
altitudinal range of 215–520 m a.s.l. The mean ice thickness of Davies Dome and Whisky
Glacier is 32.4 ± 1.2 m and 99.6 ± 1.8 m, respectively. Between 1979 and 2006, the area of
the ice dome decreased from 6.23 ± 0.05 km2 to 4.94 ± 0.01 km2 (–20.7%), while the area
of the valley glacier was reduced from 2.69 ± 0.02 km2 to 2.40 ± 0.01 km2 (–10.6%). These
values are higher than the mean annual areal changes of ice caps on Livingston Island
(Calvet et al., 1999; Molina et al., 2007) and King George Island (Simões et al., 1999) over
the period 1956–2000. According to recent studies, the mean annual area decrease of
Davies Dome and Whisky Glacier is similar to the accelerated area loss reported for
glaciers on King George Island in the period 2000–2008 (Rückamp et al., 2011), while
deceleration of the mass losses on Livingston Island was reported over the period 2002–
2011 (Navarro et al., 2013).
The volume of the ice dome and valley glacier in the period 1979–2006 was reduced
from 0.23 ± 0.03 km3 to 0.16 ± 0.02 km3 (–30.4%) and from 0.27 ± 0.02 km3 to 0.24 ± 0.01
km3 (–10.6%), respectively. The mean surface elevation decreased by 8.5 ± 2.8 m and 10.1
± 2.8 m in the same period. The mean annual elevations of Davies Dome and Whisky
Glacier declined by 0.32 ± 0.10 m and 0.37 ± 0.10 m, respectively. These values agree well
with the mean surface lowering of 0.32 m w.e. a–1 observed over the period 1988–1997 at
Rothera Point on the south-western Antarctic Peninsula (Smith et al., 1998) and with the
mean lowering rate of 0.2 m a–1 reported from the summit of Bellingshausen Dome on
King George Island (Rückamp et al., 2011). By contrast, our estimates are lower than the
values reported for the Glaciar del Diablo located on the northern coast of Vega Island ca.
30 km northeast of our studied glaciers. The mean annual surface lowering of this glacier
was –1.00 m w.e. over the period of 1985–1998 (Skvarca et al., 2004), which is much
higher than the mean values of –0.22 m w.e. a–1 over the period 2001–2011 published for
the same glacier by Marinsek and Ermolin (2015).
Generally, the average areal (0.048–0.011 km2 a–1) and volumetric (0.003–0.001 km3
a–1) changes of Davies Dome and Whisky Glacier are higher than the majority of other
estimates from Antarctic Peninsula glaciers [Paper 5]. If the estimated rate of volume loss
continues, they could disappear within 62 ± 52 and 227 ± 220 years, respectively. This
approximated timing corresponds well with the value based on the mean rate of areal loss
for Whisky Glacier (228 ± 22 years to extinction); however, the volumetric change of
Davies Dome is more rapid than its areal change (104 ± 5 years). We assume that area
24
rather than volume is a more important factor in the future evolution of both glaciers
[Paper 5]. However, recent glaciological observations available for some glaciers on
Livingston Island suggest that greater retreat and ice volume loss over the last three
decades have been replaced by a slower decrease of mass loss (Navarro et al., 2013). The
authors attribute this deceleration to a combination of decreased melt and increased
accumulation (precipitation) in the South Shetland Islands, which is associated with the
deepening of the circumpolar pressure trough, while the melt decrease is associated with
lower summer surface temperatures in the period 2002–2011. Meteorological data from
our recent studies on the Ulu Peninsula, James Ross Island revealed that changes in the
surface elevation and mass balance of Davies Dome and Whisky Glacier over the last
decade are most likely to be a result of orographically enhanced accumulation via snow
drifting by prevailing westerly and southwesterly winds, which are typical for the
northern part of James Ross Island (Zvěřina et al. 2014; Láska et al., 2015). The other
meteorological parameters, such as air temperature and solar radiation, have become less
important with time.
2.2.2 Changes of the permafrost active layer and its thermal properties on the Ulu
Peninsula [Paper 6]
The response of permafrost to the recent climate warming represents one of the most
important topics in periglacial research and climate modelling (IPCC, 2013). Numerical
models indicate that permafrost may become the dominant contributor of CO2 and CH4 to
the atmosphere in the 21st century (Schaefer et al., 2011). It is therefore crucial to fully
understand the dynamics and climatic sensitivity of permafrost and active layer thickness,
particularly in the polar areas, which are undergoing rapid warming. The mean annual air
temperatures rose substantially along the Antarctic Peninsula, with accelerating glacier
retreat and permafrost temperature increase (e.g.; Strelin et al., 2006; Cook and Vaughan,
2010; Bockheim et al., 2013; Davies et al., 2013). At the same time, the positive trend in
precipitation and snow accumulation has been observed and modelled on the western
Antarctic Peninsula since 1950 as a result of changes in the atmospheric circulation and
its regional variability (Van den Broeke et al., 2006; Thomas et al., 2008). However, the
magnitude and pattern of atmospheric warming differ significantly between the western
and eastern side of the Antarctic Peninsula due to changes in sea surface temperatures,
sea ice cover and atmospheric circulation. Despite the increasing number of periglacial
studies focusing on the Antarctic Peninsula in the last decade (e.g.; Vieira et al., 2010;
Bockheim et al., 2013; Almeida et al., 2014; de Pablo et al., 2014), the thermal conditions
in the active-layer and their interaction with meteorological factors are not well known. In
particular, the influence of snow on active-layer thickness and thermal regime is relatively
poorly understood, despite being a major modulating factor to the atmosphere (Boike et
al., 2008). Furthermore, most of the permafrost and active layer monitoring sites are
located along the western coast of the Antarctic Peninsula and on the South Shetland
Islands; only a limited amount of the data originates from the eastern Antarctic Peninsula
(Bockheim et al., 2013). The northernmost ice-free part of James Ross Island (Fig. 2)
25
represents a unique place to study the sensitivity of permafrost and active layer in
relation to regional atmospheric warming.
The main aims of Paper 6 were, therefore, (a) a description of the current state and
thermal dynamics of the active layer on the Ulu Peninsula, northern James Ross Island in
2011–2013 and (b) the evaluation of the effect of air temperature and snow cover on
active-layer temperature and their seasonal dynamics. The study site was located on a
Holocene marine terrace, approximately 100 m south of the Mendel Station at 10 m a.s.l.
(Fig. 3). Temperature conditions in the active layer were measured at depths of 5, 10, 20,
30, 50 and 75 cm using Pt100/Class A resistance probes (EMS Brno), while an air
temperature probe (EMS 33) was placed at 2 m above the ground. The occurrence of snow
cover was estimated from surface albedo and measurements of incoming and reflected
shortwave radiation using EMS-11 (EMS Brno) and CM6-B (Kipp & Zonen) radiometers.
Additionally, snow depth was measured using an ultrasonic depth sensor (Judd
Communication). In order to evaluate the buffering effect of snow cover on ground
temperatures, the thawing and freezing degree days and the freezing n-factors were
calculated (Guglielmin et al., 2008).
The maximum active layer thickness was observed between the end of January and
the middle of February. The maximum thicknesses of 58 and 52 cm measured in 2012 and
2013, respectively, were significantly lower than the values reported from the lowelevation
sites of the South Shetland Islands, South Orkney Islands, or the western
Antarctic Peninsula coast (Guglielmin et al., 2012; Almeida et al., 2014; de Pablo et al.,
2014; Wilhelm et al., 2015). Regression analysis indicates a significant effect of air
temperature on the ground temperature regime. This effect was especially apparent
under snow-free conditions at 5 cm in depth with the correlation coefficients varying
between 0.74 (2011) and 0.91 (2012). Conversely, the influence of global radiation was
found to be less significant than air temperature forcing. The effect of snow depth on the
ground thermal regime at the Mendel Station was found negligible mostly due to thin and
irregularly distributed snow cover. The maximum snow depth did not exceed 30 cm in the
flat areas of James Ross Island. The minor effect of thin snow cover on the ground thermal
regime during the freezing season at the study site is also indicated by the high freezing nfactor
values (0.95 to 0.97). By contrast, lower n-factor values (0.2 to 0.7) have been
reported from Maritime Antarctica (Almeida et al., 2014; De Pablo et al., 2014), reflecting
a higher precipitation rate with thicker winter snow cover at the sites on Livingston Island
(De Pablo et al., 2014) and King George Island (Almeida et al., 2014). The effect of snow
cover on ground temperature amplitude vanishes at 20 cm depth. The snow cover showed
large day-to-day and seasonal variability where the the thickness was less than 10 cm. Our
in situ measurements [Paper 6] have recently been confirmed by both numerical models
(Seo et al., 2015; Van Wessem et al., 2016) and observational data (Zvěřina et al., 2014;
Nývlt et al., 2016), showing that high wind speeds cause irregular deposition of snow
cover and significant snow removal from the flat landscape of James Ross Island.
26
2.2.3 Effects of abiotic factors and climatic conditions on the cyanobacterialmicroalgal
community in two shallow lakes on the Ulu Peninsula [Paper 7]
Recent climate and environmental changes in the region of the Antarctic Peninsula and
James Ross Island have significantly affected local terrestrial and aquatic ecosystems and
their diversity (Laybourn-Parry and Pearce, 1997; Quayle et al., 2002; Nielsen et al., 2012;
Kopalová et al., 2013). The ice-free areas of James Ross Island are consequently colonized
by diverse organisms, such as bryophytes, lichens, cyanobacteria, green algae and diatoms
(Komárek and Elster, 2008; Láska et al., 2011). Antarctic lakes represent unique aquatic
ecosystems for studying the evolution of microorganisms, as well as a wide range of
environmental and climate changes occurring in the past and present. The origin of the
lakes on James Ross Island is related to the last deglaciations of the Antarctic Peninsula ice
sheet and the retreat of the James Ross Island ice cap during the late Pleistocene and
Holocene (Nedbalová et al., 2013; Nývlt et al., 2014). Interactions between volcanic
landforms and glacial processes during previous glacial-interglacial cycles, the Holocene
paraglacial and periglacial processes, and the relative sea level changes have resulted in
the complex present-day landscape of James Ross Island (Davies et al., 2013). All of these
processes have influenced the development of the lakes, which are found on the Ulu
Peninsula at altitudes from <20 m a.s.l. near the coast to 400 m a.s.l. in the mountain areas
(Nedbalová et al., 2013).
The aim of Paper 7 is to describe the climatic and limnological characteristics of the
two endorheic lakes (Lake 1, Lake 2) on the Ulu Peninsula, together with the chemical and
biological composition of the unusual calcareous organo-sedimentary structures, which
have been found on the lake floors. Furthermore, a hypothesis concerning the formation
of calcite spicules is presented. The lakes are located near Andreassen Point on the
eastern coast of the Ulu Peninsula (Fig. 3). They are shallow with maximum depths of 1.1
and 0.9 m, and mean depths of 0.5 and 0.3 m (Nedbalová et al., 2013). The water levels in
both lakes fluctuated dramatically due to the changes in the melting water supply from
the surrounding snowfields and water losses through evaporation from the ice-free water
surface in a few weeks of austral summer. The thermal regime of Lake 1, together with the
local climate conditions, represented by global solar radiation and air temperature, were
measured from 1 February 2009 to 30 November 2010. The physico-chemical
characteristics of the lake water of both lakes were sampled three times in the
observation period. The stones covered by photoautotrophic mats-biofilm collected in the
field were used for (a) phytobenthos community description, (b) thin section fixing and
isolation of dominant species, (c) scanning electron and optical microscopy and (d)
determination of the structure and chemical composition of calcium carbonate spicules.
The measured data show that monthly mean water temperatures in the lake ranged
from -10.4 °C (August 2009) to 5.8 °C (February 2010), while monthly mean air
temperatures were between -18.7 °C and 0.7 °C. The highest night-day air temperature
fluctuations reached up to 28 °C and were recorded during the winter months, while the
highest amplitudes of lake water temperature were recorded from November to February,
with typical values between 2 °C and 4 °C. The water temperature above 0 °C was
recorded from November to April (139 days on average). The number of days with
27
temperatures between 0 °C and 4 °C, during which the benthic littoral community can be
metabolically active, remains the same as for liquid water occurrence with small changes
towards the transition period (February–April and September–November, respectively).
The lakes are frozen down to the floor for 8–9 months per year, which can be seen as a
natural incubator moderating potential mechanical disturbances and stabilising the
thermal regimes of the lakes [Paper 7]. Lake water chemistry is characterized by low
conductivity and neutral to slightly alkaline pH. A mucilaginous black biofilm in the stony
littoral zone of the lakes is covered by green spots formed by a green alga and
macroscopic structures packed together with fine clastic material. The photosynthetic
microbial mat is composed of filamentous cyanobacteria (Calothrix elsteri, Hassallia
andreassenni, Hassallia antarctica) and green algae Hazenia broadyi that are considered to
be Antarctic endemic species (Komárek et al., 2012; [Paper 7]). Inorganic compounds of
biofilms are represented by (1) allochthonous mineral grains that are overgrown and
incorporated by biofilms and (2) calcareous spicules of different sizes ranging from 0.5
mm to 1 cm that are precipitated within the cyanobacterial-microalgal community, with
the presence of Hassallia andreassenni in particular. Based on the character of the rock
substrate and lake sediments, it is suggested that two of the main prerequisites for the
existence of this cyanobacterial-microalgal community producing unusual biogenic calcite
structures are (1) the flat and stable substrate in both lakes and (2) the low sedimentation
rate. The specific community in the lakes is well adapted to both low sedimentation rates
resulting from minor water input and seasonally elevated sedimentation rates coming
from frequent and intense winds (Turner et al., 2009; Zvěřina et al., 2014). During storms,
the wind carries a relatively large amount of small mineral grains and rock
microfragments, which however does not stop the growth of cyanobacterial-microalgal
biofilms, due to their ability of incorporating mineral grains within the living tissue
(Riding, 2011).
Paper 7 has shown that inorganic substances precipitated by microbial lithogenetic
processes are exclusively represented by spicules composed of calcium carbonate
monocrystals, which have a layered structure and worn surface, reflecting growth and
degradation processes. However, the biogenic calcite structures in both lakes are quite
different from any microbially mediated structures described in modern environments
(Kremer et al., 2008; Couradeau et al., 2011) and differ from structures formed by abiotic
precipitation (e.g., Vogt and Corte, 1996). Although there are many lakes with thick mats
and similar chemical characteristics on the Ulu Peninsula, the calcite spicules were found
exclusively in these two endorheic lakes. We believe that their formation is linked to the
specific photoautotrophic mats in the lakes, influenced by a set of abiotic factors and the
local climate conditions described above. Therefore, we hypothesise that the more rapid
photosynthesis rate of Hazenia, in comparison with cyanobacteria, may induce the
conditions necessary for carbonate precipitation in the lakes (Schneider and Le CampionAlsumard,
1999; Vincent, 2000). However, some role of abiotic precipitation of calcite is
also possible. From our observations, we cannot clearly decide if the winter abiotic calcite
precipitation accompanies microbial lithogenetic processes. The results of Paper 7 can
serve as baseline information for further studies in the Antarctic Peninsula region,
28
focusing on e.g. the interactions between benthic microbial communities and their
environments, or the fossilisation potential of microcrystalline calcite spicules in the
Quaternary lake sediments in the studied area.
2.3 Climate variability and its impact on terrestrial ecosystems in the
Arctic Svalbard archipelago
Similar to the Antarctic Peninsula region, the largest atmospheric warming in the Atlantic
region of the Arctic over the last 50 years has been reported from the Svalbard
archipelago (Przybylak, 2007; Nordli et al., 2014). In consequence, pronounced thinning
and retreat of the glaciers was identified in this region, with an average volume change
rate of -9.71 ± 0.55 km3 yr–1 (-0.36 ± 0.02 m yr–1 w.e.) corresponding to a global sea level
rise of about +0.026 mm yr–1 (Lemke et al., 2007; Nuth et al., 2010). Major environmental
changes have also been observed in relation to sea ice extension, permafrost distribution,
or altered diversity of plant species (e.g.; Hop et al., 2002; Kohler and Aanes, 2004;
Rachlewicz et al., 2007; Moreau et al., 2009; Westermann et al., 2009; Willmes et al., 2009;
Etzelmüller et al., 2011; Van Pelt et al., 2012). The magnitude of these changes and their
spatial patterns on the Svalbard archipelago differ significantly due to the local forcing
effects (e.g., complex topography, land-surface characteristics, or sea ice occurrence),
which modulate the climate signal on a fine scale (Vihma et al., 2014). Therefore, it is a
crucial to obtain further quantitative information about the annual and seasonal variation
of weather patterns and climate conditions in central Spitsbergen in order to understand
the interactions among the components of the climate systems and predict possible
consequences for Arctic ecosystems.
2.3.1 Climate conditions and weather patterns in central Spitsbergen [Paper 8]
Over the last few decades, the intensive atmospheric and climate research on Spitsbergen
has been carried out primarily in the vicinity of the largest settlements (Longyearbyen,
Ny-Ålesund), or permanent research stations situated along the western coasts of
Spitsbergen Island (Fig. 4). The first meteorological observations in the central part of
Spitsbergen and Petuniabukta in particular were carried out during the Polish expedition
organized by the Adam Mickiewicz University in 1985 (Kostrzewski et al., 1989).
Nevertheless, most of the climate studies in this region dealt with the analysis of shortterm
meteorological observations conducted during limited parts of summer seasons
(July–August).
The main objectives of Paper 8 were, therefore, to 1) analyze temporal variability of
weather conditions and local climate features in the coastal ice-free zone of Petuniabukta,
2) evaluate the relationship between large-scale weather systems and the selected nearsurface
meteorological variables, and 3) compare the level of spatiotemporal variability of
air temperatures among several locations within the fjord. We focused primarily on the
description of the annual and diurnal regimes of the basic meteorological parameters
measured in Petuniabukta in the period 2008−2010 and their assessment in the context of
the synoptic situations over the Svalbard archipelago. Atmospheric circulation within the
29
study period was characterized using a catalogue of circulation types developed by
Niedźwiedź (2012). In order to evaluate local climate conditions, an automatic weather
station was installed on a raised marine terrace at the altitude of 15 m a.s.l. (Fig. 4). The
surrounding areas of the station are covered by permanent tundra vegetation; the
distance between the study site and the coastal line is 500 m. The station was equipped
with an EMS33 air temperature and humidity sensor (EMS, Czech Republic), a MetOne
034B anemometer (MetOne, USA), and an Omega OS-36-2 near-infrared temperature
sensor (Omega, USA). The occurrence of snow cover was estimated from the surface
albedo based on the measurements of incoming and reflected shortwave radiation using
two identical Shenk 8101 starpyranometers (Ph. Shenk, Austria).
The meteorological data presented in Paper 8 proved that atmospheric circulation
was clearly pronounced in the variability of air temperature and surface wind speed
recorded in Petuniabukta. The highest daily mean air temperatures, ranging from 0.2 °C to
-1.9 °C, were primarily found during easterly and southerly winds. These winds were
predominantly associated with cyclonic circulations (Ec, SEc and NEc types). The
maximum air temperature over the study period reached 16.2 °C (28 July 2009) during
partly cloudy weather conditions. By contrast, the lowest air temperatures were observed
during a flow from the western quadrant with a wind speed of less than 3 m s−1. The
occurrence of relatively cold air in Petuniabukta was explained by a combination of
several factors: (1) a blocking effect of the mountains along the north-western coast of
Petuniabukta, affecting the occurrence of low wind speed, (2) the transformation of the
flow over these mountains, causing a decrease in cloudiness, and (3) strong radiative
cooling of the airflow from the snow and sea ice surfaces. This is in agreement with the
findings of Westermann et al. (2009) and Bednorz (2011), who reported both the
occurrence of winter air temperature extremes on the Svalbard archipelago and the role
of individual components of the surface energy budget on the snow surface temperatures
during polar nights.
The results of Paper 8 demonstrated that cloudiness and, consequently, also
incoming solar radiation are significantly controlled by the atmospheric circulation
arising from pressure patterns. High seasonal and day-to-day variability of incoming solar
radiation was found between spring and autumn. This was documented through the
increasing variation coefficients for the individual months: from 28% in May to 107% in
October. In spring and early summer, the anticyclonic situations brought low cloudiness
and extremes in daily mean irradiance (higher than 500 W m-2). By contrast, increasing
cyclonic activity in autumn significantly reduced the incoming solar radiation and its
short-term fluctuation. This is consistent with the findings of Ørbæk et al. (1999),
Przybylak and Araźny (2006) and Bednorz et al. (2016), who addressed the major role of
circulation patterns on cloudiness and solar radiation variation at Ny-Ålesund and
Longyearbyen, respectively. Moreover, solar radiation reaching the tundra vegetation in
summer strongly affected the diurnal amplitude of ground surface and air temperatures.
The highest surface temperature rose up to 26.5 °C on 11 July 2009 due to the high
intensity of solar radiation incidence on the dry tundra vegetation surface with an albedo
level of 0.14. The close relationship between surface and air temperatures was primarily
30
observed during summer (r = 0.93) rather than winter (r = 0.43). The seasonal changes of
the correlation coefficient clearly confirmed the essential role of snow cover occurrence.
The pronounced effect of snow cover on the variability of air temperatures was also
documented by Leszkiewicz and Caputa (2004) in the area of Hornsund, situated on the
south-western coast of Spitsbergen. In addition, the influence of large-scale weather
systems on winter air temperatures in Petuniabukta was clearly seen in the day-to-day
temperature variability, often exceeding 15 °C, with the highest standard deviation found
in January. In contrast to Paper 8, Przybylak et al. (2014) reported the much lower value
of 6−7 °C for this area in the period from December 2010 to February 2011.
Over the period of 2009−2010, the large-scale flow was often modified by
channelling and drainage effects accompanied by an increase in local wind speed and a
change in the frequency distribution of wind direction in Petuniabukta [Paper 8]. The
wind rose pattern clearly corresponded with the local terrain morphology with the
enhanced wind frequencies in the longitudinal axis of Billefjorden and Petuniabukta (Fig.
4). In addition, there were large seasonal differences in wind direction found between
summer and winter season. The southerly and north-easterly winds were most common
from June to August, while westerly winds occurred primarily from November to
February. The highest wind speed (>10 m s−1) was observed in Petuniabukta from the
northeast and east towards the Ragnarbreen and Ebbabreen glaciers. Such a wind
occurred mainly in autumn and winter, and was associated with Nc, NEc, Ec cyclonic
circulation types, or the Ea anticyclonic type. The numerical simulations made by
Kilpeläinen et al. (2011) and Esau and Repina (2012) confirmed the pronounced effects of
the complex topography on the local wind structure in the other fjords on Svalbard,
Isfjorden and Kongsfjorden in particular.
Finally, the evaluation of air temperature variability within the fjord was based on a
comparison of air temperatures and sea ice conditions between Petuniabukta and the
nearest all-year round operating meteorological station at Svalbard Lufthavn
(Longyearbyen). The Svalbard Lufthavn station is located in the middle part of Isfjorden at
a distance of 56 km from Petuniabukta. To evaluate the atmosphere−sea-ice interactions
in the fjord, the high resolution sea ice charts of the Svalbard area provided by the Sea Ice
Service of the Norwegian Meteorological Institute were used. Hereafter, the negative
differences corresponded to higher temperatures at Svalbard Lufthavn and vice versa
[Paper 8]. Generally, the Svalbard Lufthavn was about 0.9 °C warmer than Petuniabukta,
with the monthly mean differences ranging from -2.5 °C in January 2010 to 0.4 °C in July
2010. The mean daily air temperatures at Svalbard Lufthavn were higher than those in
Petuniabukta in 69% of the period. The highest negative daily mean difference was
observed in January (-7.0 °C), while the highest positive one was in March (+6.2 °C). The
negative thermal differences in winter occurred during light westerly winds and strong
radiative cooling from the snow surface, when the air temperatures in Petuniabukta
dropped below -22 °C. Furthermore, the negative thermal differences occurred when the
sea ice conditions significantly differed between both sites. This happened frequently
when fast ice occurred in Petuniabukta, while the open water (or open drift ice) was
observed in the middle part of Isfjorden near Svalbard Lufthavn station. The presence of
31
sea ice cover as a very important factor determining the specific conditions of the
atmospheric boundary layer over the Svalbard fjord was also reported by Mäkiranta et al.
(2011), Kilpeläinen et al. (2012) and Vihma et al. (2014).
Paper 8 presents a complex analysis and new information regarding the fundamental
meteorological parameters measured at Petuniabukta in the period 2008−2010. It has
been shown that local weather conditions in central Spitsbergen were significantly
influenced by several major factors: synoptic-scale weather systems, cloudiness,
topography, and local surface characteristics, snow cover in particular. Moreover, this
paper reported that the interactions between atmosphere and sea-ice in the fjord play an
essential role in forming the spatial differences in the near-surface variables and air
temperature extremes in particular.
2.3.2 Modelling of the surface wind characteristics in the complex topography of
central Spitsbergen [Paper 9]
The largest future temperature increase in the Arctic is predicted in the 21st century by a
large number of climate models (Collins et al., 2013). The quality of climate projections
and numerical model outputs is closely related to the input of observational data and
ability to simulate atmospheric processes at sufficient resolution in space and time
(Dethloff et al., 2001). Among the efficient and prospective tools for the evaluation of the
atmospheric circulation patterns and their coupling with land surface features, numerical
weather prediction (NWP) models represent a reasonable option (Tjernström et al., 2005;
Žagar et al., 2006). The current NWP models allow us to analyse the regional and local
circulation conditions and provide relatively accurate estimates of the physical properties
of the atmospheric boundary layer at high temporal (in minutes) and spatial resolution (in
kilometres). However, the atmospheric circulation and boundary layer processes over the
Svalbard fjords are strongly influenced by complex topography and sea ice occurrence
(Kilpeläinen et al., 2011; Esau and Repina, 2012; [Paper 8]). The modelling of near-surface
meteorological variables in the case of the Svalbard archipelago, therefore, requires
specific attention to be paid to model setup and simulation strategy (e.g.; Mayer et al.,
2012; Vihma et al., 2014). Many authors claim that the Weather Research and Forecasting
(WRF) mesoscale model can meet these requirements, as it provides dynamic downscaling
techniques with a variety of physical options, as well as boundary layer
parameterization schemes with high grid resolution (e.g.; Michalakes et al., 2004;
Skamarock et al., 2008; Hines and Bromwich, 2008).
Although the number of numerical simulations in Svalbard have increased over time,
most of the measuring campaigns and related WRF simulations have been carried out in
the spring months in the area of Ny-Ålesund and Longyearbyen. Information about the
wind field structure from the other part of the Svalbard archipelago is still insufficient.
Therefore, the aim of Paper 9 was to use detailed in situ summer measurements from
Petuniabukta in central Spitsbergen to 1) analyse the temporal and spatial variation in the
near-surface wind field, 2) evaluate the WRF model performance using different boundary
layer parameterization schemes, especially in relation to the local topography, and 3)
identify model weaknesses. For the purpose of the second point, large-scale forcing and
32
atmospheric circulation patterns leading to the best and worst simulations of surface
wind characteristics were also analysed. The 850-hPa geopotential heights and
corresponding geostrophic winds available from the ERA-Interim reanalysis project were
acquired (Dee et al., 2011). We used the Advanced Research WRF mesoscale model
version 3.5, developed at the National Center for Atmospheric Research (NCAR, USA).
More details of the model can be found in Skamarock et al. (2008). The WRF simulations
were conducted using three boundary layer parameterization schemes, the Yonsei
University (YSU), the Mellor-Yamada-Janjic (MYJ) and the Quasi-Normal Scale Elimination
(QNSE) schemes, with 1-km horizontal resolution of the inner domain centred in
Petuniabukta. The model variables were estimated at 61 levels on the vertical axis, of
which 18 were normally in the lowest 1000 m of the atmosphere; the lowest full model
level was approximately 7.4 m above ground. The selected parameterization schemes
have already been tested in the Arctic regions under a stable stratified atmosphere in
several spring experiments (Hines and Bromwich, 2008; Kilpeläinen et al., 2012; Mayer et
al., 2012; Roberts et al., 2015).
In order to investigate the local wind systems and evaluate the WRF model outputs,
three automatic weather stations (Terrace, Hørbye foreland, Mumien Peak) were set up in
Petuniabukta at sites of different elevations, local topography, and surface properties (Fig.
4). The lowest station, Terrace, was situated on a flat marine terrace, approximately 500
m northwest of the Petuniabukta coast. The second station was set up 300 m southeast of
the snout of Hørbyebreen glacier, wherein the valley narrows to 2 km. The third station
was situated at the top of Mumien Peak (773 m a.s.l.). At all three stations, the surface
wind speed and wind direction data were measured with the same set of MetOne 034B
Wind Sensor (MetOne, USA). The experimental measurements were carried out in the
period 9–22 July 2013, i.e. typical summer season in the study area. The WRF model
simulation was verified against the wind observations using Pearson’s correlation
coefficient, the root-mean-square error (RMSE) and a bias. The positive bias means
overestimated wind speed by the WRF model and vice versa.
In Paper 9 we reported that the WRF simulations agreed fairly well with the surface
wind observations taken at all the locations. For the surface wind speed, the mean
correlation coefficient between the modelled and observed data ranged from 0.56 to 0.67
at the significant level p <0.01. The physical options given by the three boundary layer
parameterization schemes had a rather small effect on the wind speed estimates. The best
results across all the stations were found for the QNSE scheme (bias <0.1 m s-1, RMSE = 2.3
m s-1), while the YSU and MYJ schemes showed a higher RMSE (i.e. 2.4 and 2.5 m s-1,
respectively). We found that the terrain morphology and its representation in the WRF
model affected the final accuracy of the wind speed estimates. In this regard, the largest
bias (0.6 m s-1) and RMSE (2.5 m s-1) were found at the Hørbye foreland station, at which
the fjord morphology was not well represented, at 1-km spatial resolution. On the other
hand, the wind speed bias was very small at the Terrace station (-0.1 m s-1) and Mumien
Peak (-0.3 m s-1). In comparison to Paper 9, there were slightly higher discrepancies in
wind simulation (larger bias, weaker correlation) identified during the Arctic and
Antarctic experiments by Kilpeläinen et al. (2012), Tastula and Vihma (2011) and Roberts
33
et al. (2015). In Paper 9 the model errors in wind direction were slightly higher than the
discrepancies in the wind speed. We found that all the parameterization schemes had
overestimated the frequency of the occurrence of northern or north-western wind
components, especially at the low-lying stations of Terrace and the Hørbye foreland
stations. This predominant direction can be associated with the channelling of drainage
flows occurring in the fjord bottom, which were also captured by the WRF simulation. It
confirms the findings from earlier studies in the Arctic fjords by Claremar et al. (2012)
and Mayer et al. (2012), who suggested the model setup for resolving the atmospheric
boundary layer phenomena.
Moreover, the WRF estimates proved to be highly sensitive to the large-scale forcing.
When the circulation pattern was related to a deep low pressure system and prevailing
northerly or north-westerly geostrophic wind above 10 m s-1, the observed and modelled
data agree quite well. By contrast, the surface wind representation in the model often
failed during less distinctive circulation patterns, e.g. a high-pressure ridge or baric col
associated with easterly or south-easterly flow over the Svalbard archipelago. For
example, significant underestimation of the modelled wind speed was found during the
anticyclonic circulations with weak geostrophic winds (<5 m s-1) and clear or partly
cloudy weather. Such conditions likely caused the occurrence of local circulation systems
between the fjord water and surrounding valleys, which were not captured by the WRF
model. The existence of land-sea breeze-driving mechanisms was observed on the
Svalbard archipelago (e.g. Sandvik and Furevick 2002; Kilpeläinen et al., 2011; Esau and
Repina (2012). The authors report that land-sea breeze circulation systems can be driven
by horizontal temperature differences between the open water of the fjord and the
glaciers. Moreover, some source of inaccuracies in the WRF simulations can be connected
with the model inputs based on the ERA-Interim reanalysis data which were derived from
a limited number of observations in the Arctic. In a similar way, Nawri et al. (2012) found
that underestimation of the wind speed in the WRF model in the interior of Iceland was
related to weak geostrophic winds.
Overall, the results in Paper 9 demonstrate that surface wind characteristics
observed during the 12-day summer experiment were fairly well reproduced by the WRF
simulation. The results confirmed that the WFR model capability to reproduce local winds
over the complex topography of the Svalbard archipelago is still limited, even using 1-km
spatial resolution. Most attention should be paid to the terrain representation in the
model and the boundary layer schemes in relation to less distinct synoptic situations.
Concurrently, the effect of local thermal circulation systems on the wind-field structure in
the fjords and surrounding valleys should be considered especially in the summer period.
2.3.3 Impacts of simulated warming on hummock tundra with cyanobacteria mats
in central Spitsbergen [Paper 10]
Climate change significantly influences current air temperatures and moisture conditions
in Arctic ecosystems. The predicted effects of a doubling of atmospheric CO2 have
dramatic scenarios. Mean annual temperatures in the Arctic may be 3–5 °C higher within
100 years (ACIA, 2004; Callaghan et al., 2004; Stocker et al., 2013). Most models also
34
predict that overall annual global precipitation will increase (IPCC, 2013). Temperature
and precipitation increases may have a large impact on the Arctic terrestrial ecosystems,
especially on microorganisms, of which cyanobacteria are the most abundant autotrophic
component of these systems (e.g.; Vincent, 2000; Elster, 2002). Cyanobacteria play a dual
role there. Besides being considerable primary producers, they are capable of fixing
atmospheric nitrogen, thus being an important source of nitrogen for other organisms
(e.g.; Liengen and Olsen, 1997; Walker et al., 2008). Therefore, even slight increases in
temperature and precipitation could lead to a deepening of permafrost in the active layer,
higher rates of chemical transformation within the soil profile, greater nutrient
availability and, ultimately, consequent changes in the cyanobacteria and microalgae
community structure (Shaver et al., 2000; Walker et al., 2008; Etzelmüller et al., 2011).
Many field studies have been conducted on the potential effects of global warming on
the growth and community structure of vegetation, and soil bacterial and invertebrate
community performance in various polar ecosystems (e.g.; Hollister and Webber, 2000;
Walker et al., 2008; Rinnan et al., 2009). By contrast, equivalent experimental studies on
cryptogams (mosses, lichens, cyanobacteria and microalgae) in the Arctic are much less
frequent and comparative studies across Arctic sites are missing. Moreover, it remains
unclear how climate warming will affect nitrogen fixation on the community and
ecosystem scale. Passive open-top chambers (OTC) are the most frequent manipulation
technique to mimic atmospheric warming in the vegetation cover of different types.
Within the last few decades, several in situ warming experiments have been done to study
community responses (Chapin and Shaver, 1985; Marion et al., 1997; Hollister and
Webber, 2000; Convey et al., 2002). This technique is standardized in design and forms
part of a broader geographical network of sites (e.g. the International Tundra Experiment,
ITEX), allowing for a comparison of responses on a circumpolar scale (Molau and
Mølgaard, 1996).
In [Paper 10] we reported on the first two years of the in situ experiments, using the
OTC to simulate warming, and control treatments in a wet hummock meadow in
Petuniabukta, central Spitsbergen (Fig. 4). Hummocks were 15–20 cm high, about 20–30
cm in diameter and there were about 2–3 hummocks per 1 m2. The vegetation was
dominated by mosses and cyanobacterial colonies of Nostoc commune s.l. (Prach et al.,
2012). In summer 2009, three OTCs and three Control Cage−like Structures (CCSs) were
installed at the study site. The OTC was designed as a hexagonal chamber with a bottom
diameter of 140 cm (side length of 70 cm), a top diameter of 90 cm (side length of 45 cm),
and a height of 50 cm. The bottom area of each OTC was 1.27 m2. The tasks of this field
experiment were: 1) to analyze the microclimatic parameters (soil temperature and
volumetric water content in the vegetative seasons, non−vegetative seasons and
transition period) at the wet hummock meadow under natural conditions (CCSs) and in
situ warming simulation, 2) to evaluate the OTC induced warming, based on the
comparison of the selected parameters between OTC and CCS, and 3) to evaluate the
photochemical processes of photosynthesis and nitrogenase activity (nitrogen fixation) of
Nostoc communities in relation to different microclimatic conditions.
35
In each treatment, soil temperature and volumetric water content (VWC) were
monitored in the central part of OTC and CCS at the bases and tops of the hummocks. The
soil temperature and VWC were measured at the depth of 2 cm using Pt100/Class A
resistance probes (EMS, Czech Republic) and ECH2O EC-5 soil moisture probes (Decagon
Device Inc., USA), respectively. The photochemical processes of Nostoc commune colonies
were measured with a FluorCam 700MF fluorescence imaging camera (Photon Systems
Instruments, Czech Republic). Summer nitrogenase activity was measured using
acetylene−ethylene reduction assay (Stewart et al., 1967) in 100 mL glass olasks after 90
min incubation. The ethylene in the samples was quantified with a gas chromatograph;
nitrogenase activity was expressed in nmol/h C2H4 per g of fresh biomass (Kvíderová et
al., 2011). In addition, the water level and water physico-chemical parameters were
observed at the study site during the summer [Paper 10].
The results of Paper 10 and those of several recent studies (Rennermalm et al. 2005;
Sullivan et al. 2008) clearly demonstrate that the effect of the OTC in wet hummock
meadows clearly depends upon microtopography. The dry tops of hummocks have
completely different microclimatic conditions in comparison with the wet hummock
bases. Although the mean temperature differences between the OTCs and CCSs reached
only 0.3 °C, the summer temperatures (July–August) at the hummock tops were
significantly higher and reached a mean value of 1.6 °C. The highest summer temperatures
in the OTC hummock tops ranged within 18–20 °C and rarely reached 30 °C. In the CCS,
they ranged from 12–14 °C, rising to 25 °C occasionally. On the contrary, the warming
effect was much lower in the wet bases of the hummocks in the OTC treatments, with
mean summer differences of 0.3 °C. The summer temperatures in the hummock bases
reached only 8–10 °C in OTCs and 7–8 °C in CCSs. The temperature differences between
hummock tops and bases were 8–10 °C in OTCs and 4–6 °C in CCSs treatments. These
differences were the highest during the melting season, when the mean daily differences
reached 4.2 °C, while the hourly maximum was 11 °C. Similar results of spring
temperature differences between warmed and control sites were demonstrated in a moist
tussock tundra site near Toolik Lake, Alaska (Sullivan and Welker, 2005). Immediately
after snow melting in 2002, the differences in soil temperatures between warmed and
ambient plots reached 1.5 °C. Later, the daily mean temperatures differed by only 0.8 °C
(Sullivan and Welker, 2005).
Because of the faster melt inside of the OTC treatments, positive temperatures and
liquid water occurrence at a depth of 2 cm were recorded several days earlier in the OTCs
than in the CCSs. After the thawing period in May and full water saturation in June, a
drought period occurred in July. It is often followed by increasing VWC levels in August
and early September. In total, the vegetation season was prolonged in the OTCs for 10 and
34 days in the hummock tops and the bases, respectively, in comparison to the CCS
treatments. This time shift was also clearly documented in the annual VWC course and the
mean differences between hummock tops and bases ranged from 0.08 to 0.34 m3 m-3,
respectively. Drought periods, as defined in Paper 10, lasted from 3 (2009) to 21 days
(2010) in the OTCs, although, the differences between OTCs and CCSs were small, on the
scale of a few days.
36
The high spatial variability of the soil temperature and VWC was found at the
experimental sites and reflected the heterogeneity of the microenvironment of the
particular hummocks. Although the ITEX studies were used very frequently in various
Arctic ecosystems, OTC warming experiments were rather rare in the wetland type
habitats (Nordstroem et al., 2001; Rennermalm et al., 2005; Sullivan and Welker, 2005;
Sullivan et al. 2008). In NW Greenland wetlands (Sullivan et al. 2008), the
microtopography of the sites is a mosaic of hummocks, which extends above the water
table and hollows, lying under 5–15 cm of water during the short growing season. The air
temperature (20 cm above the surface) in OTC was about 2 °C higher than in the control
plot. The soil temperatures presented in Paper 10 were also consistent with the seasonal
and diurnal variations of thermal conditions in the Greenland wetlands or wet tundra
near Toolik Lake (Sullivan and Welker, 2005) and Barrow in Alaska (Hollister and
Webber, 2000).
Desiccation stress is one of the most severe factors causing injury to vegetation in the
polar regions (Davey, 1989). In situ experiments in Petuniabukta showed that if Nostoc
commune colony water loss was less than ca 40% of relative water content, no or only
minor desiccation-induced inhibition of photosynthetic processes was observed.
Nitrogenase activity declined slowly when water loss exceeded 70% and diminished
completely at a weight loss of ca 80%. The photochemical activity remained unaffected
until the colonies lost ca 80% of their original weight (Kvíderová et al., 2011). Despite the
fact that open-top chambers raised soil temperatures and slightly changed water
availability in experimental treatments, no statistical differences in photochemical and
nitrogenase activity were observed between OTCs and CCSs in the period 2009–2010.
There are two possible explanations. Hummocks were elevated above the water table
while cyanobacteria mats consisting of Nostoc communities were located at the hummock
bottoms, which were often saturated by melt water from snow accumulation or shortterm
rainfall events in summer. It is, therefore, not surprising that microclimate
conditions in these bases were more favorable for Nostoc communities in spite of the
enhanced warming by OTCs. Moreover, Novis et al. (2007) and Kvíderová et al. (2011)
have shown that even a partially hydrated Nostoc commune s.l. colony is capable of
photosynthesis and nitrogen fixation in polar hydro-terrestrial habitats.
On the basis of data gathered during this pilot study in Petuniabukta, we proposed
the following: 1) to continue in the OTC experiment and prolong the monitoring period of
microclimate conditions and ecophysiological parameter measurements of Nostoc
colonies, and 2) to redesign the OTC construction with the aim of increasing the
temperature by up to 3–5 °C inside the chamber. In conclusion, the results in Paper 10
demonstrate that cyanobacteria and microalgae communities in the Arctic wetland
environment are influenced by a variety of abiotic factors and feedback mechanisms. They
tolerate a wide range of temperature and partial dehydration. Therefore, their responses
to climate change remain challenging for future research.
37
3 Summary and Outlook
The Antarctic Peninsula and Arctic Svalbard Archipelago have been considered among the
most rapidly warming parts of our planet over the last 50 years. The largest future
warming over the next century is predicted for these regions if greenhouse gas
concentrations continue to rise. Moreover, there is strong evidence of ongoing climate
change impact on the functioning of polar geo- and ecosystems and their particular
components. However, the responses of the biotic and abiotic components of polar
environments to climate variability have significant regional differences both within and
between the Arctic and Antarctic regions. Although the number of observations and
monitoring stations has increased significantly during the last decades, there are still
areas for which knowledge about the constituents of polar environments, their changes
and adaptations on climate variability remains insufficient.
In the thesis, the results of the extensive atmospheric and environmental research in
the selected areas of the Arctic and Antarctic are summarized. Attention is paid mainly to
the ice-free coastal areas of the Antarctic Peninsula and the central part of the Svalbard
Archipelago. The Antarctic research has been carried out in the vicinity of the Ukrainian
Vernadsky Station on Galindéz Island and the Czech Johann Gregor Mendel Station,
located in the northern part of James Ross Island. The Arctic research has been focused on
the coastal zone of Petuniabukta, central Spitsbergen Island. In the papers selected for the
habilitation thesis, an evaluation of the important aspects of climate variability and
environmental changes on different time scales is also presented. The thesis deals with
interactions and responses of the atmospheric, glacial, periglacial and aquatic
environments in the following research topics: 1) the analysis and modelling of the effects
of the ozone, cloudiness and surface reflectivity on short-wave and ultraviolet solar
radiation in the coastal and interior parts of the Antarctic, 2) the assessment of the impact
of climate change on the areal and volumetric changes of the two glaciers in the northeastern
Antarctic Peninsula, 3) the evaluation of the effect of air temperature and snow
cover on permafrost active-layer temperatures and their seasonal dynamics in northern
James Ross Island, 4) the analysis of climatic and limnological characteristics of the two
endorheic lakes in northern James Ross Island and the related description of the unusual
calcareous organo-sedimentary structures found on the lake floors, 5) the evaluation and
modelling of spatiotemporal variability of the weather conditions and local climate
features in Petuniabukta in relation to the large-scale circulation systems and complex
topography of the Svalbard Archipelago, and 6) the assessment of the impacts of the open
top chambers on microclimate conditions and the photochemical and nitrogenase activity
of Nostoc communities at the wet hummock meadow.
The presented work provides the first comprehensive information about climate
conditions in regards to how the constituents of the polar environments respond to
climate variability in the selected ice-free areas of the Antarctic Peninsula and Svalbard
Archipelago. A close relationship between some of the biotic and abiotic components was
found to be related to the inter-annual, seasonal and day-to-day variations of weather
patterns and climate conditions in the studied areas. In addition, remarkable regional
38
differences in climatic conditions and their effects on the abiotic components were
distinguished in the Antarctic Peninsula region. In contrast to more humid and snowy
areas of the South Shetland Islands, the thin and irregularly distributed snow cover on the
Ulu Peninsula has only a negligible effect on the active layer thermal dynamics. Therefore,
air temperature was identified as the major driving factor affecting permafrost active
layer development on James Ross Island. This applies also to the response of the studied
glaciers, which showed a large retreat and ice volume loss in relation to the air
temperature increases over the last three decades. Areal and volumetric changes of Davies
Dome and Whisky Glacier were higher than the majority of other estimates from Antarctic
Peninsula glaciers, although the glacier mass loss for the whole continent is very small.
Another intriguing result is the occurrence of the unusual calcareous organo-sedimentary
structures found in two lakes on the Ulu Peninsula. Their formation is most likely linked
to the specific photoautotrophic mats in the lakes, influenced by a set of abiotic factors,
low sedimentation rates, radiation and temperature regimes in particular. The analysis of
physical, chemical and biological attributes of the Antarctic lakes can be used to
reconstruct past natural climate changes and to predict the future evolution of the lakes
under the influence of anthropogenic activities. Similarly, the assessment of glacial and
periglacial environments in relation to major climatic parameters can be used for both the
reconstruction and prediction of glacier mass balance evolution and permafrost active
layer development in the near future.
The importance of carrying out complex observations and measurements using
autonomous measuring systems and automatic weather stations was also proven. For the
evaluation of atmospheric and environmental factors and their effects on the geoecosystems,
a variety of statistical methods and models were applied. Among others, the
nonlinear regression model for the estimation of erythemally weighted solar radiation
was developed and successfully validated under various atmospheric and environmental
conditions. In addition, the usefulness of satellite-based data for the reconstruction and
prediction of solar irradiances at the ground level were evaluated. The assessment of the
influence of selected factors on solar radiation clearly confirmed that irradiance variation
in the central part of Antarctica was strongly correlated with solar elevation angle and the
total ozone column, while the significant reduction of irradiance levels in the Antarctic
Peninsula region was related to cloudiness, aerosol loads and surface albedo changes.
Furthermore, the capability of the mesoscale numerical weather prediction model to
simulate the surface wind field in the complex topography of the Svalbard Archipelago
was successfully tested. The effects of local circulation systems on the wind-field structure
in the fjords and surrounding valleys were thoroughly documented. The presence of sea
ice cover as a very important factor determining the specific local climate conditions in
the Svalbard fjords was also shown.
A comprehensive evaluation of the important aspects of local climate conditions and
climate variability in the presented work indicated that some features and links between
biotic and abiotic environments remain challenging for a full explanation. The enormous
complexity in the interactions between the individual components needs to be assessed
over a long-term period using appropriate methods and data acquisition systems adjusted
39
for year-round operation in harsh weather conditions. The ice-free coastal areas of
northern James Ross Island and central Spitsbergen offer a great opportunity to continue
the study of the processes and interactions among the atmosphere, cryosphere,
hydrosphere, biosphere and geosphere. From this point of view, the long-term monitoring
research programmes already established in these areas can provide extensive
environmental data and unique information about the polar landscape systems for further
studies. Future possible investigations would also include the comprehensive scientific
works and comparative studies across other localities where a large extent of atmospheric
warming and environmental impact has been reported.
40
4 References
Ahmad, Z., Bhartia, P.K., Krotkov, N., 2004. Spectral properties of backscattered UV radiation in cloudy
atmospheres. Journal of Geophysical Research, 109, D01201.
ACIA, 2004. Impacts of warming climate: arctic climate impact assessment. Cambridge University Press,
Cambridge, 1039 pp.
Almeida, I.C.C., Schaefer, C.E.G.R., Fernandes, R.B.A., Pereira, T.T.C., Nieuwendam, A., Pereira, A.B., 2014. Active
layer thermal regime at different vegetation covers at Lion Rumps, King George Island, Maritime
Antarctica. Geomorphology, 225, 36–46.
Bamber, J.L., Aspinall, W.P., 2013. An expert judgement assessment of future sea level rise from the ice sheets.
Nature Climate Change, 3, 424–427.
Barreiro, M., Cherchi, A., Masina, S., 2011. Climate sensitivity to changes in ocean heat transport. Journal of
Climate, 24, 5015–5030.
Bednorz, E., 2011. Occurrence of winter air temperature extremes in Central Spitsbergen. Theoretical and
Applied Climatology, 106, 547–556.
Bednorz, E., Kaczmarek, D., Dudlik, P., 2016. Atmospheric conditions governing anomalies of the summer and
winter cloudiness in Spitsbergen. Theoretical and Applied Climatology, 123, 1–10.
Bener, P., 1972. Approximate values of intensity of natural ultraviolet radiation for different amounts of
atmospheric ozone. Approximate Values of Intensity of Natural Radiation for Different Amounts of
Atmospheric Ozone. Final Technical Report, Davos, Switzerland, 69 p.
Berk, A., Anderson, G.P., Acharya, P.K., Bernstein, L.S., Muratov, L., Lee, J., Fox, M., Adler-Golden, S.M., Chetwynd,
J.H., Hoke, M.L., Lockwood, R.B., Gardner, J.A., Cooley, T.W., Borel, C.C., Lewis, P.E., 2005. MODTRANTM 5, a
reformulated atmospheric band model with auxiliary species and practical multiple scattering options:
Update, In: Proceedings of SPIE - The International Society for Optical Engineering, pp. 662–667.
Bernhard, G., Booth, C.R., Ehramjian, J.C., 2004. Version 2 data of the National Science Foundation’s Ultraviolet
Radiation Monitoring Network: South Pole. Journal of Geophysical Research D: Atmospheres, 109, D21207,
1–18.
Bernhard, G., Booth, C.R., Ehramjian, J.C., 2005. UV climatology at Palmer Station, Antarctica, based on version
2 NSF network data. In: Proceedings of SPIE - The International Society for Optical Engineering, pp.
588607–588619.
Bernhard, G., Booth, C.R., Ehramjian, J.C., 2010. Climatology of ultraviolet radiation at high latitudes derived
from measurements of the National Science Foundation’s Ultraviolet Spectral Irradiance Monitoring
Network, In: Gao, W., Slusser, J.R., Schmoldt, D.L. (Eds.), UV Radiation in Global Climate Change. BerlinHeidelberg:
Springer, pp. 48–72.
Bernhard, G., Booth, C.R., Ehramjian, J.C., Nichol, S.E., 2006. UV climatology at McMurdo station, Antarctica,
based on version 2 data of the National Science Foundation’s ultraviolet radiation monitoring network.
Journal of Geophysical Research: Atmospheres, 111, D1120.
Bockheim, J., Vieira, G., Ramos, M., Lopez-Martinez, J., Serrano, E., Guglielmin, M., Wilhelm, K., Nieuwendam, A..
2013. Climate warming and permafrost dynamics in the Antarctic Peninsula region. Global and Planetary
Change, 100, 215–223.
Boike, J., Hagedorn, B., Roth, K.. 2008. Heat and water transfer processes in permafrost affected soils: A review
of field-and modeling-based studies for the Arctic and Antarctic. In: Douglas, L.K., Kenneth, M.H. (Eds.),
Proceeding of the Ninth International Conference on Permafrost, University of Alaska: Fairbanks; Vol. 1,
pp. 149–154.
Bromwich, D.H., Nicolas, J.P., Hines, K.M., Kay, J.E., Key, E.L., Lazzara, M.A., Lubin, D., McFarquhar, G.M.,
Gorodetskaya, I. V, Grosvenor, D.P., Lachlan-Cope, T., Van Lipzig, N.P.M., 2012. Tropospheric clouds in
Antarctica. Reviews of Geophysics, 50, RG1004.
Burrows, W.R., Vallee, M., Wardle, D.I., Kerr, J.B., Wilson, L.J., Tarasick, D.W., 1994. The Canadian operational
procedure for forecasting total ozone and UV radiation. Meteorological Applications, 1, 247–265.
Calbó, J., Pagès, D., González, J.-A., 2005. Empirical studies of cloud effects on UV radiation: A review. Reviews of
Geophysics, 43, 1–28.
Callaghan, T. V, Björn, L.O., Chernov, Y., Chapin, T., Christensen, T.R., Huntley, B., Ims, R.A., Johansson, M., Jolly,
D., Jonasson, S., Matveyeva, N., Panikov, N., Oechel, W., Shaver, G., Elster, J., Henttonen, H., Laine, K.,
Taulavuori, K., Taulavuori, E., Zöckler, C., 2004. Biodiversity, distributions and adaptations of arctic
species in the context of environmental change. Ambio, 33, 404–417.
Calvet, J., Sellés, D.G., Corbera, J., 1999. Fluctuations in the length of the Livingston (South Shetland Islands) ice
cap between 1956 and 1996 . Acta Geologica Hispanica, 34, 365–374.
Caputa, Z., Kejna, M., Láska, K., 1997. Accumulation of snow on Ecology Glacier (King George Island, South
Shetland Islands, Antarctica) in 1996 [in Polish]. Problemy Klimatologii Polarnej, 7, 125–142.
Chapin III, F.S., Shaver, G.R., 1985. Individualistic growth response of tundra plant species to environmental
manipulations in the field. Ecology, 66, 564–576.
41
Chipperfield, M.P., 2015. Global Atmosphere – The Antarctic Ozone Hole. In: Still Only One Earth: Progress in
the 40 Years Since the First UN Conference on the Environment. Issues in Environmental Science and
Technology, 40, pp. 1–33.
Claremar, B., Obleitner, F., Reijmer, C., Pohjola, V., Waxegård, A., Karner, F., Rutgersson, A., 2012. Applying a
Mesoscale Atmospheric Model to Svalbard Glaciers. Advances in Meteorology, Article ID 321649, 1–22.
Collins, M., Knutti, R., Arblaster, J.M., Dufresne, J.-L., Fichefet, T., Friedlingstein, P., Gao, X., Gutowski, W.J., Johns,
T., Krinner, G., 2013. Long-term climate change: projections, commitments and irreversibility, In: Stocker,
T.F., Qin, D., Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M.
(Eds.), Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth
Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University
Press, pp. 1054–1057.
Comiso, J.C., Nishio, F., 2008. Trends in the sea ice cover using enhanced and compatible AMSR-E, SSM/I, and
SMMR data. Journal of Geophysical Research: Oceans, 113, C02S07.
Comte, K., Šabacká, M., Carré-Mlouka, A., Elster, J., Komárek, J., 2007. Relationships between the Arctic and the
Antarctic cyanobacteria; three Phormidium-like strains evaluated by a polyphasic approach. FEMS
Microbiology Ecology, 59, 366–376.
Convey, P., 2003. Maritime Antarctic Climate Change: Signals from Terrestrial Biology, In: Antarctic Peninsula
Climate Variability: Historical and Paleoenvironmental Perspectives. American Geophysical Union, pp.
145–158.
Convey, P., Pugh, P.J.A., Jackson, C., Murray, A.W., Ruhland, C.T., Xiong, F.S., Day, T.A., 2002. Response of
Antarctic terrestrial microarthropods to long-term climate manipulations. Ecology, 83, 3130–3140.
Cook, A.J., Fox, A.J., Vaughan, D.G., Ferrigno, J.G., 2005. Retreating glacier fronts on the Antarctic Peninsula over
the past half-century. Science, 308, 541–544.
Cook, A.J., Vaughan, D.G., 2010. Overview of areal changes of the ice shelves on the Antarctic Peninsula over
the past 50 years. The Cryosphere, 4, 77–98.
Cordero, R.R., Damiani, A., Seckmeyer, G., Riechelmann, S., Labbe, F., Laroze, D., Garate, F., 2013. Satellitederived
UV climatology at Escudero station, Antarctic Peninsula. Antarctic Science, 25, 791–803.
Couradeau, E., Benzerara, K., Moreira, D., Gerard, E., Kaźmierczak, J., Tavera, R., López-García, P., 2011.
Prokaryotic and eukaryotic community structure in field and cultured microbialites from the alkaline
Lake Alchichica (Mexico). PLoS One, 6, 1–15.
Czech Geological Survey, 2009. James Ross Island - northern part. Topographic map 1 : 25 000. First edition.
Praha, Czech Geological Survey. ISBN 978-80-7075-734-5.
Davey, M.C., 1989. The effects of freezing and desiccation on photosynthesis and survival of terrestrial
Antarctic algae and cyanobacteria. Polar Biology, 10, 29–36.
Davies, B.J., Glasser, N.F., Carrivick, J.L., Hambrey, M.J., Smellie, J.L., Nývlt, D., 2013. Landscape evolution and
ice-sheet behaviour in a semi-arid polar environment: James Ross Island, NE Antarctic Peninsula.
Geological Society Special Publication, 381, 353–395.
De Backer, H., Koepke, P., Bais, A., De Cabo, X., Frei, T., Gillotay, D., Haite, C., Heikkilä, A., Kazantzidis, A.,
Koskela, T., Kyrö, E., Lapeta, B., Lorente, J., Masson, K., Mayer, B., Plets, H., Redondas, A., Renaud, A.,
Schauberger, G., Schmalwieser, A., Schwander, H., Vanicek, K., 2001. Comparison of measured and
modelled uv indices for the assessment of health risks. Meteorological Applications, 8, 267–277.
Dee, D.P., Uppala, S.M., Simmons, A.J., Berrisford, P., Poli. P., Kobayashi, S., Andrae, U., Balmaseda, M.A.,
Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A.C.M., Van de Berg, L., Bidlot, J., Bormann, N., Delsol, C.,
Dragani, R., Fuentes, M., Geer, A.J., Haimberger, L., Healy, S.B., Hersbach, H., Hólm, E.V., Isaksen, L.,
Kållberg, P., Köhler, M., Matricardi, M., McNally, A.P., Monge-Sanz, B.M., Morcrette, J.J., Park, B.K., Peubey,
C., De Rosnay, P., Tavolato, C., Thépaut, J.N., Vitart, F., 2011. The ERA-Interim reanalysis: configuration
and performance of the data assimilation system. Q. J. Royal Meteorological Society, 137, 553–597.
De Fabo, E.C., Noonan, F.P., Frederick, J.E., 1990. Biologically effective doses of sunlight for immune
suppression at various latitudes and their relationship to changes in stratospheric ozone. Photochemistry
and Photobiology, 52, 811–817.
De Pablo, M.A., Ramos, M., Molina, A., 2014. Thermal characterization of the active layer at the Limnopolar
Lake CALM-S site on Byers Peninsula (Livingston Island), Antarctica. Solid Earth, 5, 721–739.
Dethloff, K., Abegg, C., Rinke, A., Hebestadt, I., Romanov, V.F., 2001. Sensitivity of Arctic climate simulations to
different boundary-layer parameterizations in a regional climate model. Tellus A, 53, 1–26.
Diffey, B.L., 1977. The calculation of the spectral distribution of natural ultraviolet radiation under clear day
conditions (for UV dosimeter correction). Physics in Medicine and Biology, 22, 309–316.
Douglass, A.R., Newman, P.A., Solomon, S., 2014. The Antarctic ozone hole: An update. Physics Today, 67, 42–
48.
Eleftheratos, K., Kazadzis, S., Zerefos, C.S., Tourpali, K., Meleti, C., Balis, D., Zyrichidou, I., Lakkala, K., Feister, U.,
Koskela, T., 2015. Ozone and spectroradiometric UV changes in the past 20 years over high latitudes.
Atmosphere-Ocean, 53, 117–125.
Elster, J., 2002. Ecological classification of terrestrial algal communities in polar environments. In: Beyer, L.,
Bötler, M. (Eds.), Geoecology of Antarctic ice-free coastal lanscapes. Berlin: Springer-Verlag, pp. 303–326.
42
Elster, J., Rachlewicz, G., 2012. Preface. Petuniabukta, Billefjorden in Svalbard: Czech-Polish long term
ecological and geographical research. Polish Polar Research, 33, 289–295.
Esau, I., Repina, I., 2012. Wind climate in Kongsfjorden, Svalbard, and attribution of leading wind driving
mechanisms through turbulence−resolving simulations. Advances in Meteorology, Article ID 568454, 1–
16.
Etzelmüller, B., Schuler, T. V, Isaksen, K., Christiansen, H.H., Farbrot, H., Benestad, R., 2011. Modeling the
temperature evolution of Svalbard permafrost during the 20th and 21st century. The Cryosphere, 5, 67–
79.
Farman, J.C., Gardiner, B.G., Shanklin, J.D., 1985. Large losses of total ozone in Antarctica reveal seasonal ClO
x/NOx interaction. Nature, 315, 207–210.
Fountain, A.G., Lewis, K.J., Doran, P.T., 1999. Spatial climatic variation and its control on glacier equilibrium
line altitude in Taylor Valley, Antarctica. Global and Planetary Change, 22, 1–10.
Frederick, J.E., Qu, Z., Booth, C.R., 1998. Ultraviolet Radiation at Sites on the Antarctic Coast. Photochemistry
and Photobiology, 68, 183–190.
Glasser, N.F., Scambos, T.A., Bohlander, J., Truffer, M., Pettic, E.C., Davies, B.J., 2011. From ice-shelf tributary to
tidewater glacier: continued rapid recession, acceleration and thinning of Röhss Glacier following the
1995 collapse of the Prince Gustav Ice Shelf, Antarctic Peninsula. Journal of Glaciology, 57, 397–406.
Gregory, J.M., 2005. A model intercomparison of changes in the Atlantic thermohaline circulation in response
to increasing atmospheric CO2 concentration. Geophysical Research Letters, 32, L12703.
Guglielmin, M., 2012. Advances in permafrost and periglacial research in Antarctica: a review. Geomorphology,
155–156, 1–6.
Guglielmin, M., Ellis Evans, C.J., Cannone, N., 2008. Active layer thermal regime under different vegetation
conditions in permafrost areas. A case study at Signy Island (Maritime Antarctica). Geoderma, 144, 73–
85.
Hansen, J., Sato, M., Kharecha, P., von Schuckmann, K., 2011. Earth’s energy imbalance and implications.
Atmospheric Chemistry and Physics, 11, 13421–13449.
Hansen, J., Sato, M., Russell, G., Kharecha, P., 2013. Climate sensitivity, sea level and atmospheric carbon
dioxide. Philosophical Transactions of the Royal Society of London A: Mathematical, Physical and
Engineering Sciences, 371, 20120294.
Hassler, B., Bodeker, G.E., Solomon, S., Young, P.J., 2011. Changes in the polar vortex: Effects on Antarctic total
ozone observations at various stations. Geophysical Research Letters, 38, L01805.
Hines, K.M., Bromwich, D.H., 2008. Development and testing of polar weather research and forecasting (WRF)
model. Part I: Greenland ice sheet meteorology. Monthly Weather Review, 136, 1971–1989.
Hofmann, D.J., Johnson, B.J., Oltmans, S.J., 2009. Twenty-two years of ozonesonde measurements at the South
Pole. International Journal of Remote Sensing, 30, 3995–4008.
Holland, M.M., Bitz, C.M., 2003. Polar amplification of climate change in coupled models. Climate Dynamics, 21,
221–232.
Hollister, R.D., Webber, P.J., 2000. Biotic validation of small open-top chambers in a tundra ecosystem. Global
Change Biology, 6, 835–842.
Hop, H., Pearson, T., Hegseth, E.N., Kovacs, K.M., Wiencke, C., Kwasniewski, S., Eiane, K., Mehlum, F., Gulliksen,
B., Wlodarska-Kowalczuk, M., Lydersen, C., Weslawski, J.M., Cochrane, S., Gabrielsen, G.W., Leakey, R.J.G.,
Lønne, O.J., Zajaczkowski, M., Falk-Petersen, S., Kendall, M., Wängberg, S.-Å., Bischof, K., Voronkov, A.Y.,
Kovaltchouk, N.A., Wiktor, J., Poltermann, M., Di Prisco, G., Papucci, C., Gerland, S., 2002. The marine
ecosystem of Kongsfjorden, Svalbard. Polar Research, 21, 167–208.
Hosking, J.S., Orr, A., Marshall, G.J., Turner, J., Phillips, T., 2013. The influence of the amundsen-bellingshausen
seas low on the climate of West Antarctica and its representation in coupled climate model simulations.
Journal of Climate, 26, 6633–6648.
IPCC, 2013. Climate Change 2013: The Physical Science Basis. In: Stocker, T.F., Qin, D., Plattner, G.-K., Tignor,
M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V., Midgley, P.M. (Eds.), Contribution of Working Group I
to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University
Press, Cambridge, United Kingdom and New York, NY, USA, 1535 pp.
Jaross, G., Warner, J., 2008. Use of Antarctica for validating reflected solar radiation measured by satellite
sensors. Journal of Geophysical Research, 113, D16S34.
Kejna, M., Láska, K., 1999a. Spatial differentiation of ground temperature in the region of Arctowski Station,
King George Island, Antarctica in 1996. Polish Polar Research, 20, 221–241.
Kejna, M., Láska, K., 1999b. Weather conditions at Arctowski Station, King George Island, South Shetland
Islands, Antarctica in 1996. Polish Polar Research, 20, 203–220.
Kilpeläinen, T., Wihma, T., Manninen, M., Sjölblom, A., Jakobson, E., Palo, T., Maturilli, M., 2012. Modelling the
vertical structure of the atmospheric boundary layer over Arctic fjords in Svalbard. Q. J. Royal
Meteorological Society, 138, 1867–1883.
Kilpeläinen, T., Vihma, T., Ólafsson, H., 2011. Modelling of spatial variability and topographic effects over
Arctic fjords in Svalbard. Tellus, Series A: Dynamic Meteorology and Oceanography, 63, 223–237.
43
King, J.C., Comiso, J.C., 2003. The spatial coherence of interannual temperature variations in the Antarctic
Peninsula. Geophysical Research Letters, 30, 11–12.
King, J.C., Turner, J., 1997. Antarctic Meteorology and Climatology, Cambridge: Cambridge University Press, UK,
409 pp.
Klein, S.A., Hartmann, D.L., 1993. The seasonal cycle of low stratiform clouds. Journal of Climate, 6, 1587–1606.
Koepke, P., Bais, A., Balis, D., Buchwitz, M., De Backer, H., De Cabo, X., Eckert, P., Eriksen, P., Gillotay, D.,
Heikkilä, A., Koskela, T., Lapeta, B., Litynska, Z., Lorente, J., Mayer, B., Renaud, A., Ruggaber, A.,
Schauberger, G., Seckmeyer, G., Seifert, P., Schmalwieser, A., Schwander, H., Vanicek, K., Weber, M., 1998.
Comparison of Models Used for UV Index Calculations. Photochemistry and Photobiology, 67, 657–662.
Kohler, J., Aanes, R., 2004. Effect of winter snow and ground-icing on a Svalbard reindeer population: Results
of a simple snowpack model. Arctic, Antarctic, and Alpine Research, 36, 333–341.
Komárek, J., Elster, J., Komárek, O., 2008. Diversity of the cyanobacterial microflora of the northern part of
James Ross Island, NW Weddell Sea, Antarctica. Polar Biology, 31, 853–865.
Komárek, J., Nedbalová, L., Hauer, T., 2012. Phylogenetic position and taxonomy of three heterocytous
cyanobacteria dominating the littoral of deglaciated lakes, James Ross Island, Antarctica. Polar Biology,
35, 759–774.
Kondratyev, K.Y., Varotsos, C.A., 1996. Global total ozone dynamics: Impact on surface solar ultraviolet
radiation variability and ecosystems. Environmental Science and Pollution Research, 3, 153–157.
Kopalová, K., Nedbalová, L., Nývlt, D., Elster, J., van de Vijver, B., 2013. Diversity, ecology and biogeography of
the freshwater diatom communities from Ulu Peninsula (James Ross Island, NE Antarctic Peninsula).
Polar Biology, 36, 933–948.
Kostrzewski, A., Kaniecki, A., Kapuscinski, J., Klimczak, R., Stach, A., Zwolinski, Z., 1989. The dynamics and rate
of denudation of glaciated and non-glaciated catchments, central Spitsbergen. Polish Polar Research, 10,
317–367.
Kravchenko, V., Evtushevsky, A., Grytsai, A., Milinevsky, G., Shanklin, J., 2009. Total ozone dependence of the
difference between the empirically corrected EP-TOMS and high-latitude station datasets. International
Journal of Remote Sensing, 30, 4283–4294.
Kremer, B., Kazmierczak, J., Stal, L.J., 2008. Calcium carbonate precipitation in cyanobacterial mats from sandy
tidal flats of the North Sea. Geobiology, 6, 46–56.
Kuipers Munneke, P., Van den Broeke, M.R., Lenaerts, J.T.M., Flanner, M.G., Gardner, A.S., Van de Berg, W.J.,
2011. A new albedo parameterization for use in climate models over the Antarctic ice sheet. Journal of
Geophysical Research: Atmospheres, 116, D05114.
Kylling, A., Albold, A., Seckmeyer, G., 1997. Transmittance of a cloud is wavelength-dependent in the UV-range:
Physical interpretation. Geophysical Research Letters, 24, 397–400.
Kvíderová, J., Elster, J., Šimek, M., 2011. In situ response of Nostoc commune s.l. colonies to desiccation in
Central Svalbard, Norwegian High Arctic. Fottea, 11, 87–97.
Labow, G.J., Herman, J.R., Huang, L.-K., Lloyd, S.A., DeLand, M.T., Qin, W., Mao, J., Larko, D.E., 2011. Diurnal
variation of 340 nm Lambertian equivalent reflectivity due to clouds and aerosols over land and oceans.
Journal of Geophysical Research Atmospheres, 116, D11202.
Láska, K., Barták, M., Hájek, J., Prošek, P., Bohuslavová, O., 2011. Climatic and ecological characteristics of
deglaciated area of James Ross Island, Antarctica, with a special respect to vegetation cover. Czech Polar
Reports, 1, 49–62.
Láska, K., Nývlt, D., Engel, Z., Stachoň, Z., 2015. Monitoring of land-based glaciers on James Ross Island,
Antarctic Peninsula. Geophysical Research Abstracts, 17, EGU2015–8546.
Laybourn-Parry, J., Pearce, D.A., 1997. The biodiversity and ecology of Antarctic lakes: models for evolution.
Philosophic Transaction of the Royal Society, B 362, 2273–2289.
Lee, Y. G., Koo, J.-H., Kim, J., 2015. Influence of cloud fraction and snow cover to the variation of surface UV
radiation at King Sejong station, Antarctica, Atmospheric Research, doi: 10.1016/j.atmosres.2015.04.020
Lemke, P., Ren, J., Alley, R.B., Allison, I., Carrasco, J., Flato, G., Fujii, Y., Kaser, G., Mote, P., Thomas, R.H., Zhang, T.,
2007. Observations: Changes in snow, ice and frozen ground. In: Climate Change 2007: The Physical
Science Basis, pp. 337–383.
Leszkiewicz, J., Caputa, Z., 2004. The thermal condition of the active layer in the permafrost at Hornsund,
Spitsbergen. Polish Polar Research, 25, 223–239.
Liengen, T., Olsen, R.A., 1997. Nitrogen fixation by free-living cyanobacteria from different coastal sites in a
high archc tundra, Spitsbergen. Arctic and Alpine Research, 29, 470–477.
Lindfors, A., Arola, A., 2008. On the wavelength-dependent attenuation of UV radiation by clouds. Geophysical
Research Letters, 35, L0580.
Lubin, D., Morrow, E., 2001. Ultraviolet radiation environment of Antarctica 1. Effect of sea ice on top-ofatmosphere
albedo and on satellite retrievals. Journal of Geophysical Research Atmospheres, 106, 33453–
33461.
Luccini, E., Cede, A., Piacentini, R.D., 2003. Effect of clouds on UV and total irradiance at Paradise Bay, Antarctic
Peninsula, from a summer 2000 campaign. Theoretical and Applied Climatology, 75, 105–116.
44
Madronich, S. 1993. UV radiation in the natural and perturbed atmosphere. In: Tevini, M., (Ed), Environmental
Effects of UV (Ultraviolet) Radiation. Lewis, Boca Raton, pp. 17–69.
Madronich, S., Flocke, S., 1997. Theoretical estimation of biologically effective UV radiation at the Earth’s
surface, In: Zerefos, C.S., Bais, A.F. (Eds.), Solar Ultraviolet Radiation - Modeling, Measurements and Effects,
NATO ASI Series I: Global Environmental Change. Berlin: Springer-Verlag, Vol. 52, pp. 23–48.
Madronich, S., McKenzie, R.L., Björn, L.O., Caldwell, M.M., 1998. Changes in biologically active ultraviolet
radiation reaching the Earth’s surface. Journal of Photochemistry and Photobiology B: Biology, 46, 5–19.
Mäkiranta, E., Vihma, T., Sjöblom, A.,Tastula, E.M., 2011. Observations and Modelling of the Atmospheric
Boundary Layer Over Sea-Ice in a Svalbard Fjord. Boundary Layer Meteorology, 140, 105–123.
Marinsek, S., Ermolin, E., 2015. 10 Year mass balance by glaciological and geodetic methods of Glaciar Bahía
del Diablo, Vega Island, Antarctic Peninsula. Annals of Glaciology, 56, 141–146.
Marion, G.M., Henry, G.H.R., Freckman, D.W., Johnstone, J., Jones, G., Jones, M.H., Lévesque, E., Molau, U.,
Mølgaard, P., Parsons, A.N., Svoboda, J., Virginia, R.A., 1997. Open-top designs for manipulating field
temperature in high-latitude ecosystems. Global Change Biology, 3, 20–32.
Marshall, G.J., 2002. Analysis of recent circulation and thermal advection change in the northern Antarctic
Peninsula. International Journal of Climatology, 22, 1557–1567.
Marshall, G.J., 2007. Half-century seasonal relationships between the Southern Annular Mode and Antarctic
temperatures. International Journal of Climatology, 27, 373–383.
Marshall, G.J., Orr, A., van Lipzig, N.P.M., King, J.C., 2006. The impact of a changing Southern Hemisphere
Annular Mode on Antarctic Peninsula summer temperatures. Journal of Climate, 19, 5388–5404.
Marshall, J., Scott, J.R., Armour, K.C., Campin, J.-M., Kelley, M., Romanou, A., 2014. The ocean’s role in the
transient response of climate to abrupt greenhouse gas forcing. Climate Dynamics, 44, 2287–2299.
Mayer, B., Kylling, A., 2005. Technical note: The libRadtran software package for radiative transfer calculations
- Description and examples of use. Atmospheric Chemistry and Physics, 5, 1855–1877.
Mayer, S., Jonassen, M.O., Sandvik, A., Reuder, J., 2012. Profiling the Arctic Stable Boundary Layer in Advent
Valley, Svalbard: Measurements and Simulations. Boundary-Layer Meteorology, 143, 507–526.
McGuire, A.D., Anderson, L.G., Christensen, T.R., Scott, D., Laodong, G., Hayes, D.J., Martin, H., Lorenson, T.D.,
Macdonald, R.W., Nigel, R., 2009. Sensitivity of the carbon cycle in the Arctic to climate change. Ecological
Monographs, 79, 523–555.
Michalakes, J., Dudhia, J., Gill, D., Henderson, T., Klemp, J., Skamarock, W., Wang, W., 2004. The Weather
Research and Forecast Model: Software Architecture and Performance. 13 pp.
Molau, U., Mølgaard, P., 1996. ITEX manual (2nd ed.). Copenhagen: Danish Polar Centre, 85 pp.
Molina, C., Navarro, F.J., Calver, J., García-Sellés, D., Lapazaran, J.J., 2007. Hurd Peninsula glaciers, Livingston
Island, Antarctica, as indicators of regional warming: ice-volume changes during the period 1956–2000.
Annals of Glaciology, 46, 43–49.
Moreau, M., Laffly, D., Brossard, T., 2009. Recent spatial development of Svalbard strandflat vegetation over a
period of 31 years. Polar Research, 28, 364–375.
Mulvaney, R., Abram, N.J., Hindmarsh, R.C.A., Arrowsmith, C., Fleet, L., Triest, J., Sime, L.C., Alemany, O., Foord,
S., 2012. Recent Antarctic Peninsula warming relative to Holocene climate and ice-shelf history. Nature,
489, 141–144.
Navarro, F.J., Eisen, O., 2009. Ground-penetrating radar in glaciological applications. In: Pellika, P., Gareth Rees,
W. (Eds.), Remote Sensing of Glaciers: Techniques for Topographic, Spatial and Thematic Mapping of
Glaciers. London: Taylor & Francis, pp. 195–229.
Navarro, F.J., Jonsell, U.Y., Corcuera, M.I., Martín-Español, A., 2013. Decelerated mass loss of Hurd and Johnsons
Glaciers, Livingston Island, Antarctic Peninsula. Journal of Glaciology, 59, 115–128.
Nawri, N., Björnsson, H., Jónasson, K., Petersen, G.N., 2012. Evaluation of WRF mesoscale model simulations of
surface wind over Iceland. Report VÍ, 2012-010, 44 pp.
Newchurch, M.J., Yang, E., Cunnold, D.M., Reinsel, G.C., Zawodny, J.M., Russell, J.M., 2003. Evidence for
slowdown in stratospheric ozone loss: First stage of ozone recovery. Journal of Geophysical Research:
Atmospheres, 108, 4507.
Nedbalová, L., Nývlt, D., Kopáček, J., Šobr, M., Elster, J., 2013. Freshwater lakes of Ulu Peninsula, James Ross
Island, north-east Antarctic Peninsula: Origin, geomorphology and physical and chemical limnology.
Antarctic Science, 25, 358–372.
Newman, P.A., Kawa, S.R., Nash, E.R., 2004. On the size of the Antarctic ozone hole. Geophysical Research
Letters, 31, L21104.
Niedźwiedź, T., 2012. The calendar of atmospheric circulation types for Spitsbergen. A computer file,
University of Silesia, Department of Climatology, Sosnowiec, Poland.
http://klimat.wnoz.us.edu.pl/osoby/tn/Spitsbergen.zip
Nielsen, U.N., Wall, D.H., Adams, B.J., Virginia, R.A., 2011. Antarctic nematode communities: observed and
predicted responses to climate change. Polar Biology, 34, 1701–1711.
Nordli, Ø., Przybylak, R., Ogilvie, A.E.J., Isaksen, K., 2014. Long-term temperature trends and variability on
Spitsbergen: the extended Svalbard Airport temperature series, 1898–2012. Polar Research, 33, 21349.
45
Nordstroem, C., Soegaard, H., Christensen, T.R., Friborg, T., Hansen, B.U., 2001. Seasonal carbon dioxide
balance and respiration of a high-arctic fen ecosystem in NE-Greenland. Theoretical and Applied
Climatology, 70, 149–166.
Novis, P.M., Whitehead, D., Gregorich, E.G., Hunt, J.E., Sparrow, A.D., Hopkins, D.W., Elberling, B., Greenfield,
L.G., 2007. Annual carbon fixation in terrestrial populations of Nostoc commune (Cyanobacteria) from an
Antarctic dry valley is driven by temperature regime. Global Change Biology, 13, 1224–1237.
Nunez, M., Michael, K., Turner, D., Wall, M., Nilsson, C., 1997. A satellite-based climatology of UV-B irradiance
for Antarctic coastal regions. International Journal of Climatology, 17, 1029–1054.
Nuth, C., Moholdt, G., Kohler, J., Hagen, J.O., Kääb, A., 2010. Svalbard glacier elevation changes and contribution
to sea level rise. Journal of Geophysical Research: Earth Surface, 115.
Nývlt, D., Braucher, R., Engel, Z., Mlčoch, B., 2014. Timing of the Northern Prince Gustav Ice Stream retreat and
the deglaciation of northern James Ross Island, Antarctic Peninsula during the last glacial–interglacial
transition. Quaternary Research, 82, 441–449.
Nývlt, D., Fišáková, M.N., Barták, M., Stachoň, Z., Pavel, V., Mlčoch, B., Láska, K., 2016. Death age, seasonality,
taphonomy and colonization of seal carcasses from Ulu Peninsula, James Ross Island, Antarctic Peninsula.
Antarctic Science, 28, 3–16.
OMI (Ozone Monitoring Instrument), 2014, http://www.nasa.gov/mission_pages/aura/spacecraft/omi.html.
Ørbæk, J.B., Hisdal, V., Svaasand, L.E., 1999. Radiation climate variability in Svalbard: surface and satellite
observations. Polar Research, 18, 127–134.
Platnick, S., King, M.D., Ackerman, S. A., Menzel, W. P., Baum, B. A., Riedi, J. C., Frey, R. A. 2003. The MODIS cloud
products: Algorithms and examples from Terra. IEEE Trans. Geoscience Remote Sensing, 41, 459–473.
Plewes, L.A., Hubbard, B., 2001. A review of the use of radio-echo sounding in glaciology. Progress in Physical
Geography, 25, 203–236.
Prach, K., Klimesova, J., Kosnar, J., Redcenko, O., Hais, M., 2012. Variability of contemporary vegetation around
Petuniabukta, central Spitsbergen. Polish Polar Research, 33, 383–394.
Prošek, P., Janouch, M., Láska, K., 2000. Components of the energy balance of the ground surface and their
effect on the thermics of the substrata of the vegetation oasis at Henryk Arctowski Station, King George
Island, South Shetland Islands. Polar Record, 36, 3–18.
Prošek, P., Láska, K., Janouch, M., 2001. Ultraviolet radiation intensity at the H. Arctowski Base (South
Shetlands, King George Island) during the period from December 1994-December 1996. Folia Facultatis
Scientiarum Naturalium Universitatis Masarykianae Brunensis. Geographia, 25, 35–48.
Przybylak, R., 2007. Recent air-temperature changes in the Arctic. Annals of Glaciology, 46, 316–324.
Przybylak, R., Araźny, A., 2006. Climatic conditions of the north-western part of Oscar II Land (Spitsbergen) in
the period between 1975 and 2000. Polish Polar Research, 27, 133–152.
Przybylak, R., Araźny, A., Nordli, Ø., Finkelnburg, R., Kejna, M., Budzik, T., Migała, K., Sikora, S., Puczko, D.,
Rymer, K., Rachlewicz, G., 2014. Spatial distribution of air temperature on Svalbard during 1 year with
campaign measurements. International Journal of Climatology, 34, 3702–3719.
Quayle, W.C., Peck, L.S., Peat, H., Ellis-Evans, J.C., Harrigan, P.R., 2002. Extreme responses to climate change in
Antarctic lakes. Science, 295, 645.
Rabassa, J., Skvarca, P., Bertani, L., Mazzoni, E., 1982. Glacier inventory of James Ross and Vega Islands,
Antarctic Peninsula. Annals of Glaciology, 3, 260–264.
Rachlewicz, G., Szczuciński, W. And Ewertowski, M., 2007. Post-“Little Ice Age” retreat rates of glaciers around
Billefjorden in central Spitsbergen, Svalbard. Polish Polar Research, 28, 159–186.
Rau, F., Mauz, F., De Angelis, H., Jaña, R., Neto, J.A., Skvarca, P., Vogt, S., Surer, H., Gossmann, H., 2004. Variations
of glacier frontal positions on the northern Antarctic Peninsula. Annals of Glaciology, 39, 525–530.
Renaud, A., Staehelin, J., Fröhlich, C., Philipona, R., Heimo, A., 2000. Influence of snow and clouds on erythemal
UV radiation: Analysis of Swiss measurements and comparison with models. Journal of Geophysical
Research Atmospheres, 105, 4961–4969.
Rennermalm, A.K., Soegaard, H., Nordstroem, C., 2005. Interannual variability in carbon dioxide exchange from
a high arctic fen estimated by measurements and modeling. Arctic, Antarctic, and Alpine Research, 37,
545–556.
Riding, R., 2011. Microbialities, stromatolites, and thrombolites. In: Reitner, J., Thiel, V. (Eds.), Encyclopedia of
Geobiology - Encyclopedia of Earth Science Series. Berlin-Heidelberg: Springer, pp. 635–654.
Rignot, E., Bamber, J., Van den Broeke, M.R., Davis, C., Li, Y., Van de Berg, W.J., Van Meijgaard, E., 2008. Recent
Antarctic ice mass loss from radar interferometry and regional climate modelling. Nature Geoscience, 1,
106–110.
Rinnan, R., Rousk, J., Yergeau, E., Kowalchuk, G.A., Bååth, E., 2009. Temperature adaptation of soil bacterial
communities along an Antarctic climate gradient: Predicting responses to climate warming. Global
Change Biology, 15, 2615–2625.
Rippin, D.M., Vaughan, D.G., Corr, H.F.J., 2011a. The basal roughness of Pine Island Glacier, West Antarctica.
Journal of Glaciology, 57, 67–76.
46
Roberts, T.J., Dütsch, M., Hole, L.R., Voss, P.B., 2015. Controlled meteorological (CMET) balloon profiling of the
Arctic atmospheric boundary layer around Spitsbergen compared to a mesoscale model. Atmospheric
Chemistry and Physics Discussions, 15, 27539–27573.
Rückamp, M., Braun, M., Suckro, S., Blindow, N., 2011. Observed glacial changes on the King George Island ice
cap, Antarctica, in the last decade. Global and Planetary Change, 79, 99–109.
Sandvik, A.D., Furevik, B.R., 2002. Case study of a coastal jet at Spitsbergen - Comparison of SAR- and modelestimated
wind. Monthly Weather Review, 130, 1040–1051.
Schaefer. K., Zhang. T., Brujwiler, L., Barrett, A.P., 2011. Amount and timing of permafrost carbon release in
response to climate warming. Tellus, 63, 165–180.
Schauberger, G., Schmalwieser, A.W., Rubel, F., Wang, Y., Keck, G., 1997. UV-Index: Operationelle Prognose der
solaren, biologisch-effektiven Ultraviolett-Strahlung in Österreich. Zeitschrift Für Medizinische Physik, 7,
153–160.
Schippnick, P.F., Green, A.E., 1982. Analytical characterization of spectral actinic flux and spectral irradiance in
the middle ultraviolet. Photochemistry and Photobiology, 35, 89–101.
Schmalwieser, A.W., Schauberger, G., Janouch, M., Nunez, M., Koskela, T., Berger, D., Karamanlan, G., Prošek, P.,
Láska, K., 2002a. Global validation of a forecast model for irradiance of solar, erythemally effective
ultraviolet radiation. Optical Engineering, 41, 3040–3050.
Schmalwieser, A.W., Schauberger, G., Simic, S., Weihs, P., Janouch, M., Vanicek, K., 2002b. Total ozone content
as input parameter for the prediction of the biologically effective uv radiation: Analysis of the temporal
and spatial variability over Austria, In: Proceedings of SPIE - The International Society for Optical
Engineering, pp. 432–435.
Schneider, J., Le Campion-Alsumard, T., 1999. Construction and destruction of carbonates by marine and
freshwater cyanobacteria. European Journal of Phycology, 34, 417–426.
Seo, K.-W., Wilson, C.R., Scambos, T., Kim, B.-M., Waliser, D.E., Tian, B., Kim, B.-H., Eom, J., 2015. Surface mass
balance contributions to acceleration of Antarctic ice mass loss during 2003–2013. Journal of Geophysical
Research: Solid Earth, 120, 3617–3627.
Shaver, G.R., Canadell, J., Chapin III, F.S., Gurevitch, J., Harte, J., Henry, G., Ineson, P., Jonasson, S., Melillo, J.,
Pitelka, L., Rustad, L., 2000. Global warming and terrestrial ecosystems: A conceptual framework for
analysis. BioScience, 50, 871–882.
Serreze, M.C., Barry, R.G., 2011. Processes and impacts of Arctic amplification: A research synthesis. Global and
Planetary Change, 77, 85–96.
Simões, J.C., Bremer, U.F., Aquino, F.E., Ferron, F.A., 1999. Morphology and variations of glacial drainage basins
in the King George ice field, Antarctica. Annals of Glaciology, 29, 220–224.
Skamarock, W.C., Klemp, J.B., Gill, D.O., Barker, D.M., Wang, W., Powers, J.G., 2008. A Description of the
Advanced Research WRF Version 3. Mesoscale and Microscale Meteorology Division, National Centre for
Atmospheric Research, 125 pp.
Skvarca, P., De Angelis, H., Ermolin, E., 2004. Mass balance of ‘Glaciar Bahía del Diablo’, Vega Island, Antarctic
Peninsula. Annals of Glaciology, 39, 209–213.
Skvarca, P., Rack, W., Rott, H., Donángelo, T., 1998. Evidence of recent climatic warming on the eastern
Antarctic Peninsula. Annals of Glaciology, 27, 628–632.
Smith, A.M., Vaughan, D.G., Doake, C.S.M., Johnson, A.C., 1998. Surface lowering of the ice ramp at Rothera
Point, Antarctic Peninsula, in response to regional climate change. Annals of Glaciology, 27, 113–118.
Stamnes, K., Tsay, S.-C., Wiscombe, W., Jayaweera, K., 1988. Numerically stable algorithm for discrete-ordinatemethod
radiative transfer in multiple scattering and emitting layered media. Applied Optics, 27, 2502–
2509.
Stocker, B.D., Roth, R., Joos, F., Spahni, R., Steinacher, M., Zaehle, S., Bouwman, L., Prentice, I.C., 2013. Multiple
greenhouse-gas feedbacks from the land biosphere under future climate change scenarios. Nature
Climate Change, 3, 666–672.
Strelin, J.A., Sone, T., Mori, J., Torielli, C.A., Nakamura, T., 2006. New Data Related to Holocene Landform
Development and Climatic Change from James Ross Island, Antarctic Peninsula. In: Fütterer, D.K.,
Damaske, D., Kleinschmidt, G., Miller, H., Tessensohn, F. (Eds.), Antarctica: Contributions to Global Earth
Sciences. Berlin-Heidelberg: Springer, pp. 455–459.
Sullivan, P.F., Arens, S.J.T., Chimner, R.A., Welker, J.M., 2008. Temperature and microtopography interact to
control carbon cycling in a high arctic fen. Ecosystems, 11, 61–76.
Sullivan, P.F., Welker, J.M., 2005. Warming chambers stimulate early season growth of an arctic sedge: Results
of a minirhizotron field study. Oecologia, 142, 616–626.
Tastula, E.M., Vihma, T., 2011. WRF model experiments on the Antarctic Atmosphere in Winter. Monthly
Weather Review, 139, 1279–1291.
Thomas, E.R., Marshall, G.J., McConnell, J.R., 2008. A doubling in snow accumulation in the western Antarctic
Peninsula since 1850. Geophysical Research Letters, 35, L01706.
Tjernström, M., Žagar, M., Svensson, G., Cassano, J.J., Pfeifer, S., Rinke, A., Wyser, K., Dethloff, K., Jones, C.,
Semmler, T., Shaw, M., 2005. Modelling the Arctic boundary layer: An evaluation of six ARCMIP regionalscale
models using data from the SHEBA project. Boundary-Layer Meteorology, 117, 337–381.
47
Tomasi, C., Vitale, V., Lupi, A., Di Carmine, C., Campanelli, M., Herber, A., Treffeisen, R., Stone, R. S., Andrews, E.,
Sharma, S., Radionov, V., von Hoyningen-Huene, W., Stebel, K., Hansen, G. H., Myhre, C. L., Wehrli, C.,
Aaltonen, V., Lihavainen, H., Virkkula, A., Hillamo, R., Ström, J., Toledano, C., Cachorro, V. E., Ortiz, P., de
Frutos, A. M., Blindheim, S., Frioud, M., Gausa, M., Zielinski, T., Petelski, T., Yamanouchi, T., 2007. Aerosols
in polar regions: A historical overview based on optical depth and in situ observations, Journal of
Geophysical Research Atmospheres, 112, D16205.
Turner, J., Barrand, N.E., Bracegirdle, T.J., Convey, P., Hodgson, D.A., Jarvis, M., Jenkins, A., Marshall, G.,
Meredith, M.P., Roscoe, H., Shanklin, J., French, J., Goosse, H., Guglielmin, M., Gutt, J., Jacobs, S., Kennicutt II,
M.C., Masson-Delmotte, V., Mayewski, P., Navarro, F., Robinson, S., Scambos, T., Sparrow, M.,
Summerhayes, C., Speer, K., Klepikov, A., 2014. Antarctic climate change and the environment: An update.
Polar Record, 50, 237–259.
Turner, J., Chenoli, S.N., abu Samah, A., Marshall, G., Phillips, T., Orr, A., 2009. Strong wind events in the
Antarctic. Journal of Geophysical Research, 114, D18103.
Turner, J., Colwell, S.R., Marshall, G.J., Lachlan-Cope, T.A., Carleton, A.M., Jones, P.D., Lagun, V., Reid, P.A.,
Iagovkina, S., 2005. Antarctic climate change during the last 50 years. International Journal of
Climatology, 25, 279–294.
Turner, J., Marshall, G.J., Lachlan-Cope, T.A., 1998. Analysis of synoptic-scale low pressure systems within the
Antarctic Peninsula sector of the circumpolar trough. International Journal of Climatology, 18, 253–280.
Van den Broeke, M., Van de Berg, W.J., Van Meijgaard, E., 2006. Snowfall in coastal West Antarctica much
greater than previously assumed. Geophysical Research Letters, 33, L02505.
Van Lipzig, N.P.M., Marshall, G.J., Orr, A., King, J.C., 2008. The relationship between the Southern Hemisphere
annular mode and antarctic Peninsula summer temperatures: Analysis of a high-resolution model
climatology. Journal of Climate, 21, 1649–1668.
Van Pelt, W.J.J., Oerlemans, J., Reijmer, C.H., Pohjola, V.A., Pettersson, R. And Van Angelen, J.H., 2012. Simulating
melt, runoff and refreezing on Nordenskiöldbreen, Svalbard, using a coupled snow and energy balance
model. The Cryosphere, 6, 641–659.
Van Wessem, J.M., Ligtenberg, S.R.M., Reijmer, C.H., van de Berg, W.J., van den Broeke, M.R., Barrand, N.E.,
Thomas, E.R., Turner, J., Wuite, J., Scambos, T.A., van Meijgaard, E., 2016. The modelled surface mass
balance of the Antarctic Peninsula at 5.5 km horizontal resolution. The Cryosphere, 10, 271–285.
Vanicek, K., 1997. Relation between total ozone and erythemal solar radiation on clear days in Hradec Kralove.
In: Proceedings of the Workshop on Monitoring of UVB Radiation on Total Ozone, Poprad: Slovak
Hydrometeorological Institute, pp. 105–107.
Varotsos, C., 2004. The extraordinary events of the major, sudden stratospheric warming, the diminutive
antarctic ozone hole, and its split in 2002. Environmental Science and Pollution Research, 11, 405–411.
Vaughan, D.G., Marshall, G.J., Connolley, W.M., Parkinson, C., Mulvaney, R., Hodgson, D.A., King, J.C., Pudsey, C.J.,
Turner, J., 2003. Recent rapid regional climate warming on the Antarctic Peninsula. Climatic Change, 60,
243–274.
Verlinden, K.L., Thompson, D.W.J., Stephens, G.L., 2011. The three-dimensional distribution of clouds over the
Southern Hemisphere high latitudes. Journal of Climate, 24, 5799–5811.
Vieira, G., Bockheim, J., Guglielmin, M., Balks, M., Abramov, A.A., Boelhouwers, J., Cannone, N., Ganzert, L.,
Gilichinsky, D.A., Goryachkin, S., López-Martínez, J., Meiklejohn, I., Raffi, R., Ramos, M., Schaefer, C.,
Serrano, E., Simas, F., Sletten. R., Wagner, D., 2010. Thermal State of permafrost and active-layer
monitoring in the Antarctic: advances during the International Polar Year 2007-09. Permafrost and
Periglacial Processes, 21, 182–197.
Vihma, T., Pirazzini, R., Fer, I., Renfrew, I.A., Sedlar, J., Tjernstrom, M., Lupkes, C., Nygard, T., Notz, D., Weiss, J.,
Marsan, D., Cheng, B., Birnbaum, G., Gerland, S., Chechin, D., Gascard, J.C., 2014. Advances in
understanding and parameterization of small-scale physical processes in the marine Arctic climate
system: a review. Atmospheric Chemistry and Physics, 14, 9403–9450.
Vincent, W.F., 2000. Evolutionary origins of Antarctic microbiota: invasion, selection and endemism. Antarctic
Science, 12, 374–385.
Vitale, V., Petkov, B., Goutail, F., Lanconelli, C., Lupi, A., Mazzola, M., Busetto, M., Pazmino, A., Schioppo, R.,
Genoni, L., Tomasi, C., 2011. Variations of UV irradiance at Antarctic station Concordia during the springs
of 2008 and 2009. Antarctic Science, 23, 389–398.
Von Storch, H., Zwiers, F.W., 2002. Statistical Analysis in Climate Research. Cambridge: Cambridge University
Press, UK, 496 pp.
Walker, J.K.M., Egger, K.N., Henry, G.H.R., 2008. Long-term experimental warming alters nitrogen-cycling
communities but site factors remain the primary drivers of community structure in high arctic tundra
soils. ISME Journal, 2, 982–995.
Wang, X., Zender, C.S., 2011. Arctic and Antarctic diurnal and seasonal variations of snow albedo from
multiyear Baseline Surface Radiation Network measurements. Journal of Geophysical Research: Earth
Surface, 116, F03008.
48
Weihs, P., Schmalwieser, A., Reinisch, C., Meraner, E., Walisch, S., Harald, M., 2013. Measurements of personal
uv exposure on different parts of the body during various activities. Photochemistry and Photobiology, 89,
1004–1007.
Weihs, P., Webb, A.R., 1997. Accuracy of spectral UV model calculations 2. Comparison of UV calculations with
measurements. Journal of Geophysical Research Atmospheres, 102, 1551–1560.
West, A.E., Keen, A.B., Hewitt, H.T., 2013. Mechanisms causing reduced Arctic sea ice loss in a coupled climate
model. The Cryosphere, 7, 555–567.
Westermann, S., Lüers, J., Langer, M., Piel, K., Boike, J., 2009. The annual surface energy budget of a high-arctic
permafrost site on Svalbard, Norway. Cryosphere, 3, 245–263.
Wilhelm, K.R., Bockheim, J.G., Kung, S., 2015. Active Layer Thickness Prediction on the Western Antarctic
Peninsula. Permafrost and Periglacial Processes, 26, 188–199.
Willmes, S., Bareiss, J., Haas, C., Nicolaus, M., 2009. Observing snowmelt dynamics on fast ice in Kongsfjorden,
Svalbard, with NOAA/AVHRR data and field measurements. Polar Research, 28, 203–213.
Zvěřina, O., Láska, K., Červenka, R., Kuta, J., Coufalík, P., Komárek, J., 2014. Analysis of mercury and other heavy
metals accumulated in lichen Usnea antarctica from James Ross Island, Antarctica. Environmental
Monitoring and Assessment, 186, 9089–9100.
Žagar, N., Žagar, M., Cedilnik, J., Gregorič, G., Rakovec, J., 2006. Validation of mesoscale low level winds
obtained by dynamical downscaling of ERA40 over complex terrain. Tellus A, 58, 445–455.
49
5 Supplements
In the list of publications, the role of the applicant in the design of experiments, collection
of observational data, writing and editing of manuscript is specified.
Paper1 Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2009. Prediction of erythemally
effective UVB radiation by means of nonlinear regression model. Environmetrics, 20, 633–646.
Author’s contribution (ca. 80%): research design, acquisition and interpretation of observational
data, model validation, writing and editing the manuscript
Paper2 Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2010. Estimation of solar UV
radiation in maritime Antarctica using a nonlinear model including cloud effects. International
Journal of Remote Sensing, 31, 831–849.
Author’s contribution (ca. 80%): research design, acquisition and interpretation of observational
data, model validation, writing and editing the manuscript
Paper3 Láska, K., Budík, L., Budíková, M., Prošek, P., 2011. Method of estimating solar UV radiation in
high-latitude locations based on satellite ozone retrieval with an improved algorithm.
International Journal of Remote Sensing, 32, 3165–3177.
Author’s contribution (ca. 85%): research design, acquisition and interpretation of observational
data, evaluation of satellite-based data, model validation, writing and editing the manuscript
Paper4 Petkov, B.H., Láska, K., Vitale, V., Lanconelli, C., Lupi, A., Mazzola, M., Budíková, M., 2016.
Variability in solar irradiance observed at two contrasting Antarctic sites. Atmospheric
Research, 172-173, 126–135.
Author’s contribution (ca. 40%): motivation for the project, acquisition and interpretation of
observational data from Mendel Station, participation on writing and editing the manuscript
Paper5 Engel, Z., Nývlt, D., Láska, K., 2012. Ice thickness, areal and volumetric changes of Davies
Dome and Whisky Glacier (James Ross Island, Antarctic Peninsula) in 1979-2006. Journal of
Glaciology, 58, 904–914.
Author’s contribution (ca. 33%): part of the fieldwork, interpretation of observational data ,
participation on writing and editing the manuscript
Paper6 Hrbáček, F., Láska, K., Engel, Z., 2015. Effect of Snow Cover on the Active-Layer Thermal
Regime – A Case Study from James Ross Island, Antarctic Peninsula. Permafrost and
Periglacial Processes.
Author’s contribution (ca. 40%): acquisition and interpretation of observational data,
participation on writing and editing the manuscript
Paper7 Elster, J., Nedbalová, L., Vodrážka, R., Láska, K., Haloda, J., Komárek, J., 2016. Unusual biogenic
calcite structures in two shallow lakes, James Ross Island, Antarctica. Biogeosciences, 13,
535–549.
Author’s contribution (ca. 25%): acquisition and interpretation of meteorological data,
participation on writing and editing the manuscript
Paper8 Láska, K., Witoszová, D., Prošek, P., 2012. Weather patterns of the coastal zone of
Petuniabukta, central Spitsbergen in the period 2008-2010. Polish Polar Research, 33, 297–
318.
Author’s contribution (ca. 85%): research design, acquisition and interpretation of observational
data, writing and editing the manuscript
Paper9 Láska, K., Chládová, Z., Hošek, J., submmited. High-resolution numerical simulation of
summer wind field over complex topography of Svalbard archipelago. Quarterly Journal of the
Royal Meteorological Society.
Author’s contribution (ca. 70%): research design, acquisition and interpretation of observational
and modelled data, writing and editing the manuscript
Paper10 Elster, J., Kvíderová, J., Hájek, T., Láska, K., Šimek, M., 2012. Impact of warming on Nostoc
colonies (Cyanobacteria) in a wet hummock meadow, Spitsbergen. Polish Polar Research, 33,
395–420.
Author’s contribution (ca. 25%): acquisition and interpretation of meteorological data,
statistical analysis, participation on writing and editing the manuscript
50
List of other important papers and book chapters which I have published as the author or
co-author within the frame of atmospheric and environmental research in the polar
regions.
Budík, L., Budíková, M., Láska, K., Prošek, P., 2009. Application of UVB radiation regression model
for data of J. G. Mendel Station in Antarctica. In: Kovacova, M. (Ed.), APLIMAT 2009: 8th
International Conference, Proceedings. pp. 749–756.
Caputa, Z., Kejna, M., Láska, K., 1997. Accumulation of snow on Ecology Glacier (King George
Island, South Shetland Islands, Antarctica) in 1996. Problemy Klimatologii Polarnej, 7, 125–142.
[in Polish]
Engel, Z., Kopačková, V., Láska, K., Nývlt, D., 2013. Glaciologický výzkum ostrova Jamese Rosse. In:
Prošek, P. (Ed.), ANTARKTIDA. Academia, Praha, pp. 294–302. [in Czech]
Gajdošová, D., Pokorná, L., Prošek, P., Láska, K., Havel, J., 2001. Are there humic acids in Antarctica?
In: Ghabbour, E.A., Davies, G. (Eds.), Humic Substances: Structures, Models and Functions, Royal
Society of Chemistry Special Publications. pp. 121–131.
Kejna, M., Láska, K., 1999. Spatial differentiation of ground temperature in the region of Arctowski
Station, King George Island, Antarctica in 1996. Polish Polar Research, 20, 221–241.
Kejna, M., Láska, K., 1999. Weather conditions at Arctowski Station, King George Island, South
Shetland Islands, Antarctica in 1996. Polish Polar Research, 20, 203–220.
Kejna, M., Láska, K., 1999. Hydrological characteristics of the Arctowski Lake, King George Island,
Antarctica, in 1996. Problemy Klimatologii Polarnej, 9, 143–151. [in Polish]
Kejna, M., Láska, K., 1997. Dependence of weather conditions on wind direction at the Arctowski
Station (South Shetland Islands, Antarctica) in 1996. In: Glowacki, P. (Ed.), Polish Polar Studies,
24th Polar Symposium, Warsaw, 1997. Polish Polar Studies: 40th Anniversary of the Polish
Polar Station Hornsund-Spitsbergen 77°00’N 15°33'E. Warsaw, Institute of Geophysics, Polish
Academy of Sciences, pp. 153–157.
Kejna, M., Láska, K., Caputa, Z., 1998. Recession of the Ecology Glacier (King George Island) in the
period 1961-1996. In: Glowacki, P., Bednarek, J. (Eds.), Polish Polar Studies, 25th International
Polar Symposium, Warsaw, 1998. Polish Polar Studies: The 100th Anniversary of Prof. Henryk
Arctowski’s and Prof. Antoni Boleslaw Dobrowolski's Participation in the Belgica Expedition to
the Antarctic in 1887-1889. Warsaw: Institute of Geophysics, Polish Academy of Sciences, pp.
121–128.
Klánová, J., Matykiewiczová, N., Máčka, Z., Prošek, P., Láska, K., Klán, P., 2008. Persistent organic
pollutants in soils and sediments from James Ross Island, Antarctica. Environmental Pollution,
152, 416–423.
Láska, K., Barták, M., Hájek, J., Prošek, P., Bohuslavová, O., 2011. Climatic and ecological
characteristics of deglaciated area of James Ross Island, Antarctica, with a special respect to
vegetation cover. Czech Polar Reports, 1, 49–62.
Láska, K., Chládová, Z., Ambrožová, K., Husák, J., 2013. Cloudiness and weather variation in central
Svalbard in July 2013 as related to atmospheric circulation. Czech Polar Reports, 3, 184–195.
Láska, K., Prošek, P., 2013. Klima a klimatický výzkum ostrova Jamese Rosse. In: Prošek, P. (Ed.),
ANTARKTIDA. Academia, Praha, pp. 284–293. [in Czech]
Láska, K., Prošek, P., 2013. Ozónová anomálie. In: Prošek, P. (Ed.), ANTARKTIDA. Academia, Praha,
pp. 63–69. [in Czech]
Nývlt, D., Fišáková, M.N., Barták, M., Stachoň, Z., Pavel, V., Mlčoch, B., Láska, K., 2016. Death age,
seasonality, taphonomy and colonization of seal carcasses from Ulu Peninsula, James Ross
Island, Antarctic Peninsula. Antarctic Science, 28, 3–16.
Prošek, P., Barták, M., Láska, K., Suchánek, A., Hájek, J., Kapler, P., 2013. Facilities of J. G. Mendel
Antarctic station: Technical and technological solutions with a special respect to energy
sources. Czech Polar Reports, 3, 38–57.
Prošek, P., Janouch, M., Láska, K., 2000. Components of the energy balance of the ground surface
and their effect on the thermics of the substrata of the vegetation oasis at Henryk Arctowski
Station, King George Island, South Shetland Islands. Polar Record, 36, 3–18.
51
Prošek, P., Láska, K., Budíková, M., Milinevsky, G., 2004. The Regime of Total and Biological
Effective Ultraviolet Radiation at Vernadsky Station (Argentine Islands, Antarctica) and the
Impact of Ozone and Cloudiness in 2002 and 2003. In: Drbohlav, D., Kalvoda, J., Voženílek, V.
(Eds.), Czech Geography at the Dawn of the Millenium. Olomouc: Czech Geographic Society,
Palacky University in Olomouc, pp. 211–223.
Prošek, P., Láska, K., Janouch, M., 2001. Ultraviolet radiation intensity at the H. Arctowski Base
(South Shetlands, King George Island) during the period from December 1994-December 1996.
Folia Facultatis Scientiarum Naturalium Universitatis Masarykianae Brunensis. Geographia, 25,
35–48.
Schmalwieser, A.W., Schauberger, G., Janouch, M., Nunez, M., Koskela, T., Berger, D., Karamanian, G.,
Prošek, P., Láska, K., 2002. World-wide forecast of the biologically effective UV radiation: UVindex
and daily dose. In: Proceedings of SPIE - The International Society for Optical Engineering.
pp. 259–264.
Schmalwieser, A.W., Schauberger, G., Janouch, M., Nunez, M., Koskela, T., Berger, D., Karamanlan, G.,
Prošek, P., Láska, K., 2002. Global validation of a forecast model for irradiance of solar,
erythemally effective ultraviolet radiation. Optical Engineering, 41, 3040–3050.
Skácelová, K., Hrbáček, F., Chattová, B., Láska, K., Barták, M., 2015. Biodiversity of freshwater
autotrophs in selected wet places in northern coastal ecosystems of James Ross Island. Czech
Polar Reports, 5, 12–26.
Stuchlík, R., Stachoň, Z., Láska, K., Kubíček, P., 2015. Unmanned Aerial Vehicle – Efficient mapping
tool available for recent research in polar regions. Czech Polar Reports, 5, 210–221.
Witoszová, D., Láska, K., 2011. Klimatické poměry zátoky Petunia, Billefjorden (Špicberky) v
období 2008-2010. In: Středová, H., Rožnovský, J., Litschmann, T. (Eds.), Mikroklima a
Mezoklima Krajinných Struktur a Antropogenních Prostředí. Brno: Česká bioklimatologická
společnost, pp. 1-11. [in Czech]
Witoszová, D., Láska, K., 2012. Spatial distribution of air temperature in central part of Svalbard in
the period 2008-2010. Czech Polar Reports, 2, 117–122.
Zvěřina, O., Láska, K., Červenka, R., Kuta, J., Coufalík, P., Komárek, J., 2014. Analysis of mercury and
other heavy metals accumulated in lichen Usnea antarctica from James Ross Island, Antarctica.
Environmental Monitoring and Assessment, 186, 9089–9100.
2
Paper 1
Prediction of erythemally effective UVB radiation by means of
nonlinear regression model
Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2009,
Environmetrics, 20, pp. 633–646
Prediction of erythemally effective UVB radiation by means
of nonlinear regression model
Kamil La´ska1*,y
, Pavel Prosˇek1
, Ladislav Budı´k2
, Marie Budı´kova´3
and Gennadi Milinevsky4
1
Department of Geography, Faculty of Science, Masaryk University, Brno 611 37, Czech Republic
2
Czech Hydrometeorological Institute, Brno Regional Office, Brno 616 67, Czech Republic
3
Department of Mathematics and Statistics, Faculty of Science, Masaryk University, Brno 611 37, Czech Republic
4
Space Physics Department, Physics Faculty, National Taras Shevchenko University of Kyiv, Kyiv, 03680, Ukraine
SUMMARY
One of the research programs carried out within the Czech-Ukrainian scientific co-operation is the monitoring of
global solar and ultraviolet radiation at the Vernadsky Station (formerly the British Faraday Station), Antarctica.
Radiation measurements have been made since 2002. Recently, a special attention is devoted to the measurements
of the erythemally effective UVB radiation using a broadband Robertson Berger 501 UV-Biometer (Solar Light
Co. Inc., USA). This paper brings some results from modelling the daily sums of erythemally effective UVB
radiation intensity in relation to the total ozone content (TOC) in atmosphere and surface intensity of the global
solar radiation. Differences between the satellite- and ground-based measurements of the TOC at the Vernadsky
Station are taken into consideration. The modelled erythemally effective UVB radiation differed slightly
depending on the seasons and sources of the TOC. The model relative prediction error for ground- and
satellite-based measurements varied between 9.5% and 9.6% in the period of 2002–2003, while it ranged from
7.4% to 8.8% in the period of 2003–2004. Copyright # 2009 John Wiley & Sons, Ltd.
key words: Antarctica; UV radiation; ozone depletion; cloudiness; nonlinear regression model
1. INTRODUCTION
Antarctica plays a very significant role in many environmental aspects and processes of the Earth. One
of them is the impact on the global climate system (considerable influence on the planet’s albedo, long
‘‘climate memory’’ of the huge mass of ice, important factor of the meridional exchange of air masses)
and on the water circulation in the oceans. Many aspects of the Antarctic climate, such as seasonal
changes of sea ice in the Southern Ocean, temperature stability of the atmosphere, orographic and
katabatic winds, are affected by the annual regime of energy balance and solar radiation subsequently.
Specific feature of solar radiation in the Antarctic region is a seasonal increase of the ultraviolet (UV)
radiation intensity, due to the decrease of the stratospheric ozone concentration.
ENVIRONMETRICS
Environmetrics 2009; 20: 633–646
Published online 25 March 2009 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/env.968
*Correspondence to: K. La´ska, Department of Geography, Faculty of Science, Masaryk University, Kotla´rˇska´ 2, Brno 611 37,
Czech Republic.
y
E-mail: laska@sci.muni.cz
Copyright # 2009 John Wiley & Sons, Ltd.
Received 18 September 2008
Accepted 1 October 2008
The extraterrestrial solar radiation spectrum contains ultraviolet-C (UVC) radiation in the spectral
range of 100–280 nm, ultraviolet-B (UVB) radiation in the spectral range of 280–320 nm, and
ultraviolet-A (UVA) radiation in the spectral range of 320–400 nm. The UVC radiation is completely
absorbed by oxygen and ozone in the stratosphere. Due to the wavelength dependence of the ozone absorption,
a decrease in the ozone amount determines a greater increase in the UVB radiation than in the UVA.
Surface UVirradiance is affected by extraterrestrial factors (Earth’s distance from the Sun, and solar
elevation angle, which determines the optical path length), as well as by terrestrial factors, such as
aerosol concentration in the atmosphere including cloudiness and cloud types, ozone concentration, its
vertical distribution and integrated column amount, particularly the surface geometry and the ground
albedo. Model calculations also indicate that the UVB radiation increases by about 5%/km in altitude
(Bodeker, 1997).
The spatiotemporal variation of incident solar UV radiation increases from the central part
(Antarctic plateau) to the coastline (maritime Antarctica). Apart from ozone concentration, the
cloudiness is the main factor. The mean annual cloudiness in coastal regions of Antarctica near 608S is
about 80–90% (result of the frontal activity). Near 708S, the surface observations show a total cloud
cover of about 45–50% with little variability through the year and only a small decrease during the
winter months (King and Turner, 1997). In contrast to oceanic and coastal areas, the amount of clouds
near the South Pole shows large seasonal variations—from about 35% in autumn and winter to nearly
55% in spring and summer. Latitudinal changes in cloud types exhibit a certain influence, too; a large
amount of stratus occurs close to 608S, cirrus and altostratus prevail at 708S. Over the interior of
Antarctica, cirrus is the most commonly reported type of cloud (King and Turner, 1997).
The stratospheric ozone depletion over Antarctica was discovered at the beginning of the 1980s
(Farman et al., 1985). The lowest total ozone content (TOCs) in Antarctica are usually reached in the
first week of October, which, despite the rather low solar elevation in the autumn season, produce high
intensities of UVB radiation (Hermann et al., 1995). The formation of ozone depletion begins
approximately in the second half of August, culminates in the first half of October, and dissolves in
November. During the ozone depletion development, the average ozone concentration has varied at the
time of its culmination in October from the original value reaching over 300 Dobson Unit (DU) in the
1950s and 1960s to a level between 100 and 150 DU in 1990–2000. The area of ozone depletion has
been growing, although with some anomalies, in parallel with the increasing ozonosphere destruction.
Satellite measurements showed that in the early 1980s, the average October surface area with ozone
concentration below 220 DU did not exceed 7.5 million km2
, while at the end of the 1990s it was over
16 million km2
. At the beginning of the new millennium, the ozone depletion development can be
described as considerably irregular. With the exception of the year 2001 (15.3 million km2
), its surface
area did not surmount 10 million km2
(Antarctic Ozone Bulletin, 2002, 2003; Newman et al., 2004).
2. MATERIALS AND METHODS
Monitoring and forecasting of the UV radiation has become very important due to its harmful effects on
the biota as well as because of human health risks. Excessive exposure to UVB radiation can be harmful
to the skin and can result in damage of DNA and formation of skin cancers. Based on the cooperation
contract between the National Antarctic Scientific Center of Ukraine (NASCU) and the Faculty of
Science, Masaryk University, Brno, experts from the Department of Geography at the Faculty launched
a team work with the Ukrainian colleagues at the beginning of 2002 at the Ukrainian Vernadsky Station
in Antarctica (formerly the British Faraday Station) with geographical coordinates w ¼ 658010
4500
S;
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
634 K. LA´ SKA ET AL.
l ¼ 648150
2300
W (see Figure 1). The cooperation included systematic measurements of the intensity of
solar radiation and important constituents of its spectrum, among other things also:
– Global solar radiation—CM-6B Kipp&Zonen pyranometer (in service since 12 February 2002),
– Erythemally effective UVB radiation—broadband Robertson Berger 501 Version 3 UVB-Biometer,
Solar Light Co. Inc. (in service since 12 February 2002).
To protect them from snow and ice accumulations, all instruments were equipped with a heater
ventilation system (CV2, Kipp&Zonen). Solar and Ozone Observatory of the Czech Hydrometeorological
Institute in Hradec Kra´love´ performed calibration checks of all sensors thoroughly every
2 years. The sensors were installed on the Vernadsky Station building roof, while the original British
equipment for radiation measurement continued to work, allowing—among other things—to link the
original time series measurements with a new series of measurements and their integration without
disturbing the homogeneity of the entire series.
The installation of the above-mentioned equipment and the subsequent data analysis are to continue
in the earlier activities of the British Antarctic Survey and NASCU, which were focused on
atmospheric ozone and UV radiation issues and associated with the Vernadsky Station. Also, they link
up with the first Czech measurements of erythemally effective UV radiation carried out between
December 1994 and November 1998 at the Polish Arctowski Station in Antarctica (see Prosˇek and
Janouch, 1996; Prosˇek et al., 2001).
The measurement of the UV radiation intensity with the above-mentioned equipment was limited to
the UVB interval. With regard to the construction of the device, it is necessary to point out that the UVB
measurements using the UVB-Biometer were made in MED (Minimum Erythema Dose) units. One
Figure 1. The location of the Vernadsky Station in the northern part of the Antarctic Peninsula
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 635
MED hÀ1
is equivalent to UVB radiation intensity required for a minimum irritation (i.e., reddening,
skin erythema) in the skin of average pigmentation after 1 h of exposure. Conversion into physical units
is accomplished by means of the relationship: 1 MED hÀ1
¼ 0.0583 W mÀ2
, or 210 kJ mÀ2
hÀ1
. In view
of the units of measurement and their relationship to the effects on human skin, the sensors are
calibrated using the knowledge about the erythemal effects of individual wavelengths. These
wavelengths are described by the McKinlay–Diffey Erythemal Action Spectrum, which quantifies the
contribution made by the particular spectral intensities within the defined band to the appearance of
erythema (CIE, 1987; McKinlay and Differy, 1987). From the above it follows that the results of the
measurements presented in this paper, whether our own or those of other stations, do not represent direct
daily measurements of total UVB intensities in the physical sense, but rather their biological impact.
This paper presents results from the analysis of two periods (seasons) of measuring the intensity of
erythemally effective UVB radiation: Season 1—from 23 July 2002 to 28 February 2003, and Season
2—from 23 July 2003 to 28 February 2004. Both analyzed periods differ significantly, particularly in
the ozone anomaly area with the ozone concentration below 220 DU (Newman et al., 2004). In 2002,
the October average amounted to mere 3.0 million km2
and its development was very slow from
August. At the beginning of October, it split into two parts, which reunited again and then became extinct
relatively rapidly. The average surface area in October 2003 amounted already to 9.8 million km2
with the
ozone depletion formed as early as in mid-August and retained without any significant change until
mid-October. It became rapidly extinct in the second half of November (Stolarski et al., 2005).
Data about the ozone concentration can be obtained either from ground measurements predominantly
taken with Dobson and Brewer ozone spectrophotometers or from satellite measurements.
TOC is a parameter that is frequently used to determine ozone content of the atmosphere. Recently,
many authors have assessed the uncertainties in the TOC from satellite measurements as well as differences
between ground and satellite observations (e.g., Labow et al., 2004; Vanicek, 2006; Schmalwieser et al.,
2007). In this study, both TOC data sources were used and compared due to uncertainties raised in the
measured ozone. The first data series was derived from space-based measurements of the Total Ozone
Mapping Spectrometer (TOMS) aboard the NASA’s Earth Probe Satellite (EPTOMS). The EPTOMS
Version 8 provided both single retrievals and daily gridded ozone data from 1996 to 2004 (McPeters
et al., 1998; Labow et al., 2004). The daily gridded TOC values were acquired for the geographical
coordinates of the Vernadsky Station (http://toms.gsfc.nasa.gov/ftpdata.html). The second TOC series
based on the Dobson No 031 spectrophotometer measurements at Vernadsky Station was applied here.
Figures 2 and 3 primarily show the high diurnal and seasonal variability and the logical similarity of
TOC, daily sums of global solar radiation (Ig) and erythemally effective UVB regimes.
Basic extraterrestrial factors affecting these regimes are the Earth–Sun distance and the solar
elevation angle; terrestrial factors—in addition to ozone—are cloudiness, presence of aerosols and
atmospheric gases, altitude and the Earth’s surface albedo. Of these, attention is paid at this place to the
effects of cloudiness and ozone, which logically show most particularly in the UVB regime as it is most
significantly affected by ozone content. At the same time, the two terrestrial factors apparently operate
in different time regimes. While the variable cloudiness determines shorter-term fluctuations of the
radiation fluxes, ozone declines cause radiation changes of prolonged periodicity in the second half of
the year. These impacts are furthermore overlaid by the effect of cloudiness. In 2002, the effects were
clearly demonstrated for example in the period of ozone hole occurrence between 20 September and
22 October, and in the period from 23 October to 10 November. In 2003, a significant increase of UVB
between 25 September and approximately 12 October was related to the decline of ozone while the
subsequent increase of the ozone concentration from 13 October suppressed the natural increasing
trend of UVB until late November. This is why the absolute UVB maxima did not relate—unlike in the
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
636 K. LA´ SKA ET AL.
Figure 2. Daily sums of global solar radiation (Ig), erythemally effective UVB radiation (UVB), and total ozone content (TOC)
at the Vernadsky Station in the period of 2002–2003. The TOC values were measured by the Dobson spectrophotometer
Figure 3. As in Figure 2, but for 2003–2004
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 637
previous year—to the ozone minima, but rather occurred in early December, close to the summer
solstice in the Southern Hemisphere. This fact is well documented also by the timing and frequency of
occurrence of daily maximal UVB intensities and corresponding TOC values (Table 1). Only four
events of UVB intensities higher than 4 kJ mÀ2
dÀ1
were recorded in Season 1, while the number of the
events reached 10 in Season 2.
3. MODEL DESCRIPTION
Models of UV radiation prediction can be divided mainly into three groups: (1) radiative transfer
models including multi-scattering (e.g., Stamnes et al., 1988), (2) physical models with a simple
parameterization (e.g., Schippnick and Green, 1982 or Schauberger et al., 1997), and (3) statistical
models (Canadian model—Burrows et al., 1994; CHMI model—Vanicek, 1997; ETH Zu¨rich
model—Renaud et al., 2000). Particularly the radiative transfer models require extensive computation
and many assumptions about the atmospheric parameters, which are generally difficult to obtain. The
simplest group of statistical models uses a regression equation, which is obtained by fitting the
observed UV radiation to the selected file of input parameters. These models, mostly developed and
used outside Antarctica, have to be verified as to their sensitivity to the input data and their effect on
model outputs for the maritime Antarctica. This is why we came up with the development of our own
model, corresponding to both the character of available data and the specific conditions of Antarctica.
The model is to explain the regime of erythemally effective UVB radiation in dependence on
extraterrestrial (solar height) and terrestrial (atmospheric) factors.
Characteristics of the atmosphere can be expressed by the ratio of extraterrestrial global radiation
intensity and intensity of this radiation incident on the Earth’s surface. The intensity of the UVB
radiation can be then calculated on the basis of known atmospheric ozone content, changes of global
solar radiation when passing through the atmosphere, and the UVB radiation intensity at the top of the
Table 1. Maximal daily sums of erythemally effective UVB radiation (UVB) and daily means of total ozone
content (TOC) recorded at Vernadsky Station within the period of 2002–2004
Date UVB (kJ mÀ2
dÀ1
) TOC (DU)
Season 1
1.11.02 4.064 210
24.12.02 4.204 289
29.12.02 4.048 271
4.1.03 4.122 286
Season 2
25.11.03 5.478 230
28.11.03 4.635 255
4.12.03 4.390 271
5.12.03 5.157 248
6.12.03 4.423 279
7.12.03 4.332 284
8.12.03 4.117 296
25.12.03 4.126 302
28.12.03 4.067 300
2.1.04 4.217 287
The TOC values were measured by the Dobson spectrophotometer.
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
638 K. LA´ SKA ET AL.
atmosphere (TOA). Global radiation intensity at the TOA can be estimated from theory; global
radiation on the Earth’s surface and erythemally effective UVB radiation are measured directly by a
pyranometer and UV-Biometer, respectively. TOC in the atmosphere is measured by spectrophotometers
either from the Earth’s surface or from satellites, which normally carry two
spectrophotometers, one of which is oriented toward the Earth’s surface and the other one to the
outer space. With respect to the position of the Sun, directly measured TOC have to be converted to
values of a so-called effective total ozone content (TOCeff) for the purposes of the model. This is to take
account of the sunlight through the atmosphere path to the Earth’s surface, which makes their actual
path through the ozonosphere longer than toward the zenith.
We obtain the resulting daily mean TOCeff (in DU) by resolving the geometrical problem of sunlight
path through spherical layers of atmosphere (simpler solution in the plane gives higher concentration at
low solar heights)
TOCeff ¼
À2r Á sin a þ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð2r Á sin aÞ2
þ 4½ðr þ hÞ2
À r2Š Á ð1 þ sin2
aÞ
q
2ð1 þ sin2
aÞ
Á
TOC
h
(1)
where r and h are the radius of the Earth and mean height of the ozonosphere in kilometers, a is the
daily mean solar elevation angle in radians, and TOC is the daily mean total ozone content in DU.
The newly developed nonlinear regression model applies a hyperbolic attenuation of the
erythemally effective UVB radiation
UVB ¼
Ig
Igex
 b0
UVBex
1 þ b1TOC
b2
eff
(2)
where Ig is the daily sum of global solar radiation intensity incident on the Earth’s surface
(MJ mÀ2
dÀ1
), Igex is the daily extraterrestrial sum of global solar radiation intensity (MJ mÀ2
dÀ1
),
UVBex is the daily extraterrestrial sum of erythemally effective UVB radiation (kJ mÀ2
dÀ1
), and b0, b1,
b2 are the regression parameters. b0 approximates the dissimilar absorption of global radiation and
UVB radiation in the atmosphere while b1 and b2 express the ozonosphere effectiveness in the
absorption and diffusion of UVB radiation. The absorption of UVB radiation increases with increasing
ozone content in the atmosphere and with increasing length of solar ray paths through the atmosphere.
This situation particularly occurs at low heights of the Sun.
Despite the usually used ozone transmissivity exponential function (see e.g., Kondratyev and
Varotsos, 2000), we propose a new approach based on quantum transmission model of UVB radiation.
Such type of transmission is described by a hyperbolic function (Equation (2)). In the equation,
attenuation of solar UVB radiation related to optical phenomena such as absorption, reflection,
refraction, or scattering in the atmosphere is expressed by the element Ig
Igex
 b0
. The other part of the
equation, that is, the ozone absorption, is described by the element
1
1 þ b1TOC
b2
eff
(3)
For the purpose of the model, the ozone layer was divided into artificial layers. Each of them
contained such amount of ozone so that intensity of UVB radiation that passed the layer was reduced to
one half of an initial intensity of radiation at the upper surface of the layer. Such consideration means
that there is twice as much of ozone in each consecutive layer. Since the intensity of UVB radiation
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 639
does not decrease in jumps but smoothly, this consideration does not fully reflect reality. Therefore, we
multiplied TOCeff in real atmosphere by a coefficient b1 that related to the UVB attenuation due to
ozone amount. However, there are additional processes connected with the effective diameter of ozone
molecules in the atmosphere. It is reported (Kondratyev and Varotsos, 2000) that the effective diameter
is usually higher than the geometric one which might alter absorption effectiveness of UVB. Then, the
higher TOCeff, the higher effectiveness of UVB quantum absorption by ozone molecules. That was why
we used coefficient b2 in exponential part of TOCeff.
A model used in this way does not fulfill a boundary condition. The condition is a changeless UVB
radiation intensity if there is no ozone in layer. If number 1 is added to the denominator in the element (3),
the condition is fulfilled. For real TOC in the atmosphere, addition of 1 into the denominator has only
negligible effect on UVB attenuation and the range of b1 value (see the Section 5 below).
4. MODEL RESULTS
We explored the model properties for the two analyzed seasons using both ground and satellite ozone
measurements. The b0, b1, b2 parameters of the nonlinear hyperbolic regression model were estimated
by using the STATISTICA system and Levenberg–Marquardt’s method through least squares with the
initial approximations of all three parameters equal to 0.1.
As seen in Table 2, we can reject hypotheses about nonsignificance of regression parameters at the
0.05 significance level as all the 95% confidence intervals do not include 0. The variability of the
erythemally effective UVB radiation is explained by means of the model at 98.65% for Season 1 and at
98.6% for Season 2.
The hypotheses about the nonsignificance of the regression parameters at the 0.05 significance level
can be rejected also in the processing of satellite measurements of the ozone content (Table 3) since all
the 95% confidence intervals of these parameters do not include 0. The variability of the dependent
variable UVB is explained by means of the model at 98.5% for Season 1 and at 98.6% for Season 2.
Modeling of UVB radiation with the use of ground ozone measurement provides lower estimates of
the b0 and b1 parameters, and—in contrast to the satellite-based measurements—higher estimates
Table 2. Estimates of parameters and their 95% confidence intervals for ground ozone measurements
Estimates of parameters Season 1 Season 2
^b0
0.6704 Æ 0.0245 0.6794 Æ 0.0270
^b1
0.1055 Æ 0.0437 0.0724 Æ 0.0276
^b2
1.0920 Æ 0.0653 1.1492 Æ 0.0598
Table 3. Estimates of parameters and their 95% confidence intervals for satellite ozone measurements
Estimates of parameters Season 1 Season 2
^b0
0.7191 Æ 0.0247 0.7472 Æ 0.0273
^b1
0.1182 Æ 0.0501 0.0964 Æ 0.0347
^b2
1.0662 Æ 0.0662 1.0950 Æ 0.0558
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
640 K. LA´ SKA ET AL.
of the b2 parameter. Errors of estimates expressed by means of the 95% confidence intervals in the case
of using the ground ozone measurement are generally lower. This can be attributed to the fact that the
distribution of systematic errors in measuring the ozone content with respect to the area of cloudiness
and solar elevation angle differs in the ground- and satellite-based measurements. This discrepancy
comes from the fact that the ground measurements were made several times a day, while the satellite
ones took place only when passing over the concerned territory (usually once a day), and can affect the
accuracy. In this respect, the ground measurements are considered more accurate.
Using the model, we predicted UVB radiation, parameters b0, b1, and b2 respectively, for two
seasons (Table 4). Relative estimate error of the regression parameters was calculated as the half
width of 95% confidence interval to estimate value, expressed in percents. Estimates of the b0 and
Table 4. Relative estimate errors calculated on 95% confidence intervals for ground and satellite ozone
measurements
Estimates of parameters Ground ozone Satellite ozone
Season 1 Season 2 Season 1 Season 2
^b0
3.65% 3.97% 3.93% 3.65%
^b1
41.42% 38.12% 42.38% 36.00%
^b2
5.97% 5.20% 6.21% 5.10%
Figure 4. Scatterplot of daily sums of erythemally effective UVB radiation (UVB) and predicted daily sums of erythemally
effective UVB radiation (predicted UVB) at the Vernadsky station in the period of 2002–2003. The results are shown for both
satellite and ground ozone measurements used in the model
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 641
b2 parameters exhibited a considerable stability in both the cases and the relative estimate errors were
also very small. Differences in the estimated b1 values for Seasons 1 and 2 (see Tables 2 and 3) as well
as the large relative errors of the estimates were apparently caused by the moderate development of the
ozone anomaly over Antarctica in Season 1, while its influence in Season 2 was of much greater
significance (for details see Section 2).
5. DISCUSSION AND CONCLUSIONS
Figures 4 and 5 show the correlation between measured and predicted daily sums of erythemally
effective UVB radiation for Season 1 (2002–2003) and Season 2 (2003–2004). The results are shown
for both satellite and ground ozone measurements used in the model. The model quality for the two
considered seasons and both ozone estimation methods is comparable. High values of the indexes of
determination were reached. The mean average prediction errors ranged between 9.6 and 9.5% for
Season 1, and 7.4 and 8.8% for Season 2, when ground and satellite ozone measurements were applied.
Our results are in agreement with findings of Weihs and Webb (1997), Koepke et al. (1998), or De
Backer et al. (2001). They found differences between measurements and models in the range of 5–10%
with statistical uncertainty of 2–3%. To assess the differences between empirical and theoretical
distribution of residuals, the normal probability plots were used (Figures 6 and 7). The slightly worse
Figure 5. Normal probability plot of the ordered erythemally effective UVB residuals from the nonlinear regression model
against the ordered quantiles of the standard normal distribution. The results are shown for both satellite and ground ozone
measurements at the Vernadsky Station in the period of 2002–2003
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
642 K. LA´ SKA ET AL.
Figure 6. Histogram of the erythemally effective UVB residuals (UVB) at the Vernadsky Station in the period of 2002–2003.
The results are shown for both satellite and ground ozone measurements used in the model
Figure 7. As in Figure 4, but for 2003–2004
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 643
Figure 8. As in Figure 5, but for 2003–2004
Figure 9. As in Figure 6, but for 2003–2004
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
644 K. LA´ SKA ET AL.
results for Season 2 can be partly explained also by the fact that extreme values occurred in the
measurement of UVB radiation in November and December, for example, 25 November 2003 (see
Table 1). In the case of great change of ozone content, the results may be unfavorably affected by daily
average or daily sum values used at its application. The consequence may be that the model either
underestimates or overestimates the UVB radiation value at an expressively uneven solar radiation
regime within the day period (particularly at a different radiation intensity in the noon hours as
compared with the morning and evening ones). This fact is likely to affect also the magnitude of the
errors for the parameters b1 and b2, which express general ozone content, its role in UVB absorption,
and scattering.
The quality of the regression model is also documented by the distribution of residuals, which
should be ideally near the normal distribution (Figures 8 and 9). Hence the accumulation of UVB
residuals occurs with values close to zero. Therefore, the resulting distribution has high kurtosis, which
is essentially greater than in the case of normal distribution.
A model-based prediction of erythemally effective UVB radiation toward real values could be
possibly improved through the transformation of UVB radiation values. Such modification would
achieve stabilization of the variance. Searching for an appropriate transformation will be subject to
further research. We also assume that the quality of the model can be further enhanced by using values
of the studied variables measured simultaneously with the atmospheric ozone content—both ground
and satellite. This should at least eliminate problems with the informative value of daily averages and
sums with respect to the UVB radiation absorption model and its properties.
ACKNOWLEDGEMENTS
The authors express gratitude to all employees of the National Antarctic Scientific Center of Ukrainein Kyiv and to
members of crews wintering on the Vernadsky Station Alexander Kolesnik and Igor Gvozdovsky for the faultless
operation of radiation instruments and regular mailing of measured data. This study was handled as a part of
research Grant No. 205/07/1377 The effects of atmospheric factors on the regime of UV radiation in the region of
Antarctic Peninsula funded by Grant Agency of the Czech Republic.
REFERENCES
Antarctic Ozone Bulletin. 2002. http://www.wmo.ch/pages/prog/arep/gawozobull02_en.html Accessed on 28 November 2007.
Antarctic Ozone Bulletin. 2003. http://www.wmo.ch/pages/prog/arep/gawozobull03_en.html Accessed on 28 November 2007.
Bodeker G. 1997. UV radiation in polar regions. In Ecosystem Processes in Antarctic Ice-Free Landscapes, Lyons WB, HowardWilliams
C, Hawes I (eds). Balkema: Rotterdam; 23–42.
Burrows WR, Vallee´ M, Wardls DI, Kerr JB, Wilson LJ, Tarasick DW. 1994. The Canadian operational procedure for forecasting
total ozone and UV radiation. Meteorological Applications 1(3): 247–265.
CIE. 1987. A reference action spectrum for ultraviolet induced erythema in human skin. In Commission Internationale de l’ E´
clairage, CIE Publication No. 17.4, Vienna; 17–22.
De Backer H, Koepke P, Bais A, De Cabo X, Frei T, Gillotay D, Haite Ch, Heikkila¨ A, Kazantzidis A, Koskela T, Kyro¨ E, Lapeta
B, Lorente J, Masson K, Mayer B, Plets H, Redondas A, Renaud A, Schauberger G, Schmalwiesser A, Schwander H, Vanicek
K. 2001. Comparison of measured and modelled UV indices for the assessment of health risks. Meteorological Applications
8(3): 267–277.
Farman JC, Gardiner BG, Shanklin JD. 1985. Large losses of total ozone in Antarctica reveal seasonal CClOx/NOx interaction.
Nature 315: 207–210.
Hermann JR, Newman PA, Larko D. 1995. Meteor-3/TOMS observations of the 1994 ozone hole. Geophysical Research Letters
22(23): 3227–3229.
King JC, Turner J. 1997. Antarctic meteorology and climatology. Cambridge University Press: Cambridge.
Koepke P, Bais A, Balis D, Buchwitz M, De Backer H, de Cabo X, Eckert P, Eriksen P, Gillotay D, Koskela T, Lapeta B, Litynska
Z, Lorente J, Mayer B, Renaud A, Ruggaber A, Schauberger G, Seckmeyer G, Seifert P, Schmalwieser A, Schwander H,
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
UVB RADIATION MODELLING 645
Vanicek K, Weber M. 1998. Comparison of models used for UV index-calculations. Photochemistry and Photobiology 67(6):
657–662.
Kondratyev KY, Varotsos CA. 2000. Atmospheric Ozone Variability. Springer and Praxis Publishing: Chichester, UK.
Labow GJ, McPeters RD, Bhartia PK. 2004. A comparison of TOMS & SBUV Version 8 total column ozone data with data from
ground stations. In Proccedings of the XX Quadrennial Ozone Symposium, Zerefos CS (ed.). University of Athens: Kos,
Greece; 123–124.
McKinlay A, Differy BL. 1987. A reference action spectrum for ultraviolet induced erythema in human skin. In Human Exposure
to Ultraviolet Radiation: Risks and Regulations, Passchler WR, Bosnajokovic BFM (eds). Elsevier: Amsterdam; 83–87.
McPeters RD, Bhartia PK, Krueger AJ, Herman JR. 1998. Earth Probe Total Ozone Mapping Spectrometer (TOMS) Data
Products User’s Guide. NASA Technical Publication 1998-206985. NASA Goddard Space Flights Center: Greenbelt, MD.
Newman PA, Kawa SR, Nash ER. 2004. On the size of the Antarctic ozone hole. Geophysical Research Letters 31: L21104. DOI:
10.1029/2004GL020596
Prosˇek P, Janouch M. 1996. The measurement of ultraviolet radiation at the Polish Henryk Arctowski Station (South Shetlands,
Antarctica) in the Summer of 1994/95. Problemy Klimatologii Polarnej 6: 139–148.
Prosˇek P, La´ska K, Janouch M. 2001. Ultraviolet radiation intensity at the H. Arctowski Base (South Shetlands, King George
Island) during the period from December 1994–December 1996. In Ecology of the Antarctic Coastal Oasis, Prosˇek P, La´ska K
(eds). Folia Fac. Sci. Nat. Univ. Masarykianae Brunensis, Geographia 25: Brno; 35–48.
Renaud A, Staehelin J, Fro¨hlich C, Philipona R, Heimo A. 2000. Influence of snow and clouds on erythemal UV radiation:
analysis of Swiss measurements and comparison with models. Journal of Geophysical Research 105(D4): 4961–4969.
Schauberger G, Schmalwieser AW, Rubel F, Wang Y, Keck G. 1997. UV-index: Operationelle Prognose der solaren,
biologischeffektiven Ultraviolett-Strahlung in O¨ sterreich. Zeitschrift fuer Medizinische Physik 7: 153–160.
Schippnick PF, Green AES. 1982. Analytical characterisation of spectral actinic flux and spectral irradiance in the middle
ultraviolet. Photochemistry and Photobiology 35(1): 89–101.
Schmalwieser AW, Schauberger G, Erbertseder T, Janouch M, Coetzee JRC, Weihs P. 2007. Sensitivity of erythemally effective
UV irradiance and daily exposure to uncertainties in measured total ozone. Photochemistry and Photobiology 83(2): 433–443.
Stamnes K, Tsay SC, Wiscombe W, Jayaweera K. 1988. Numerically stable algorithm for discrete ordinate method radiative
transfer in multiple scattering emitting layered media. Applied Optics 27(12): 2502–2509.
STATISTICA software. http://www.statsoft.com/products/analytics.htm and http://www.statsoft.com/textbook/stnonlin.html
Accessed on 28 November 2007.
Stolarski RS, McPeters RD, Newman PA. 2005. The ozone hole of 2002 as measured by TOMS. Journal of Atmospheric Sciences
62(3): 716–720.
Vanicek K. 1997. Relation between total ozone and erythemal solar radiation on clear days in Hradec Kralove. In Proceedings of
the Workshop on Monitoring of UVB Radiation on Total Ozone. Slovak Hydrometeorological Institute: Poprad; 105–107.
Vanicek K. 2006. Differences between ground Dobson, Brewer and satellite TOMS-8, GOME-WFDOAS total ozone
observations at Hradec Kralove, Czech. Atmospheric Chemistry and Physics 6: 5163–5171.
Weihs P, Webb AR. 1997. Accuracy of spectral UV model calculations: 2. Comparison of UV calculations with measurements.
Journal of Geophysical Research 102(D1): 1551–1560.
Copyright # 2009 John Wiley & Sons, Ltd. Environmetrics 2009; 20: 633–646
DOI: 10.1002/env
646 K. LA´ SKA ET AL.
3
Paper 2
Estimation of solar UV radiation in maritime Antarctica using a
nonlinear model including cloud effects
Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2010,
International Journal of Remote Sensing, 31, pp. 831–849
Estimation of solar UV radiation in maritime Antarctica using
a nonlinear model including cloud effects
K. LA´ SKA*†, P. PROSˇEK†, L. BUDI´K‡, M. BUDI´KOVA´ §
and G. MILINEVSKY¶
†Department of Geography, Faculty of Science, Masaryk University, Brno 611 37,
Czech Republic
‡Czech Hydrometeorological Institute, Brno Regional Office, Brno 616 67,
Czech Republic
§Department of Mathematics and Statistics, Faculty of Science, Masaryk University,
Brno 611 37, Czech Republic
¶Space Physics Department, Physics Faculty, National Taras Shevchenko University
of Kyiv, Kyiv 03680, Ukraine
(Received 15 October 2008; in final form 15 March 2009)
A new approach to the estimation of erythemally effective ultraviolet (EUV)
radiation for all sky conditions that occur in maritime Antarctica is reported.
The spatial variability of the total ozone content (TOC) and attenuation of the
EUV radiation in the atmosphere are taken into consideration. The proposed
nonlinear regression model of EUV radiation is described by a hyperbolic transmission
function. The first results and the model validation for Vernadsky Station
(formerly the British Faraday Station) during the period 2002–2005 show very
good agreement with the measured values (R2
¼ 99.2). The developed model was
evaluated using daily doses of EUV radiation with respect to solar elevation angle
and cloudiness. The mean average prediction error (MAPE) for cloudy (4.1–7.0
oktas) and overcast skies (7.1–8.0 oktas) varied between 4.0% and 4.3%, while for
partly cloudy days (0–4.0 oktas) with high variability of cloud types during a day,
MAPE reached 5.9%.
1. Introduction
The amount of UV radiation reaching the Earth’s surface is affected by several
physical factors, including stratospheric ozone, solar elevation angle, clouds, surface
albedo and aerosols. The large-scale depletion of stratospheric ozone over the
Antarctic during the past two decades has led to an intensive investigation of both
physical and chemical changes in the atmosphere and the effects of increasing solar
ultraviolet (UV) radiation on polar ecosystems. The extraterrestrial solar radiation
spectrum contains ultraviolet-C (UVC) radiation in the spectral range 100–280 nm,
UVB radiation in the spectral range 280–320 nm, and UVA radiation in the spectral
range 320–400 nm. UVC radiation is completely absorbed by oxygen and ozone in the
stratosphere. Because of the wavelength dependence of the ozone absorption, a
decrease in the amount of stratospheric ozone results in a greater increase in UVB
radiation than in UVA.
*Corresponding author. Email: laska@sci.muni.cz
International Journal of Remote Sensing
ISSN 0143-1161 print/ISSN 1366-5901 online # 2010 Taylor & Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/01431160902897866
International Journal of Remote Sensing
Vol. 31, No. 4, 20 February 2010, 831–849
Ozone depletion over the Antarctic was discovered at the beginning of the 1980s
(Farman et al. 1985). The lowest total ozone columns in the Antarctic are usually
reached in the first week of October, and despite the fairly low solar elevation in the
spring season, they produce high intensities of UVB radiation incident on the surface
(Hermann et al. 1995). Ozone depletion begins approximately in the second half of
August, culminating in the first half of October and ending in November. During the
ozone depletion development, the average ozone concentration varied at the time of its
culmination in October in the 1950s and 1960s in the original value, reaching over 300
Dobson Units (DU). In 1990–2000 the ozone concentration declined to 100–150
DU. The area of ozone depletion grew, although with some anomalies, in parallel with
the increasing ozone layer destruction. Satellite measurements showed that, in the early
1980s, the average surface area in October with an ozone concentration below 220 DU
did not exceed 7.5 million km2
, while at the end of the 1990s it was over 20 million km2
.
At the beginning of the new millennium, the development of the ozone depletion can be
described as considerably irregular. With the exception of the year 2002 (the lowest
ozone hole area due to a sudden major stratospheric warming), its surface area exceeds
20 million km2
, with a maximum of 26 million km2
in 2006 (Varotsos 2002, WMO 2002,
2003, 2004, Newman et al. 2004). According to recent data, ozone loss at the high
southern latitudes has levelled off since the mid-1990s and the mean global and midlatitude
ozone losses for 2002–2005 were close to those for 1998–2001 (WMO 2008). The
weakening of the negative global trend in the total ozone content (TOC) was attributed
to stabilization in the level of ozone-depleting substances in the stratosphere due to the
Montreal Protocol international regulations. Model simulations suggest that global
ozone levels will recover to 1980 levels in the period 2060–2075 (WMO 2006).
In Antarctica, apart from the ozone concentration, ozone variability at the edge of
ozone depletion and the solar elevation angle, cloudiness is the main factor causing high
variability of solar UV radiation. As shown by several authors, the amount of clouds,
the cloud position relative to the sun, cloud type and its optical properties strongly
modulate the course of incident UV radiation (e.g. Frederick et al. 1998, Luccini et al.
2003). The temporal and spatial variation of the UV radiation increases from the central
part (Antarctic plateau) to the coastline (maritime Antarctica). The mean annual
cloudiness in the coastal regions of Antarctica near 60
S is about 80–90% (the result
of the cyclogenesis on the polar front). At around 70
S, surface observations show a
total cloud cover of about 45–50% with little variability throughout the year and only a
small decrease during the winter months (King and Turner 1997). In contrast to the
coastal areas, the amount of cloud near the South Pole shows large seasonal variations:
from about 35% in autumn and winter to nearly 55% in spring and summer. Latitudinal
changes in cloud types also have some influence; a large amount of stratus cloud occurs
close to 60
S, and cirrus and altostratus prevail at 70
S. Over the interior of Antarctica,
cirrus is the most commonly reported cloud type (Warren et al. 1986).
The biologically effective UV radiation plays a very important role in many biological
processes including harmful effects of DNA damage to human skin and the immune
system overall (De Fabo et al. 1990). The biological effectiveness of solar UV radiation
depends on the wavelength and can be measured or estimated by various theoretical
approaches. Models of biologically effective UV radiation can be divided mainly into
three groups: (1) radiative transfer models including multi-scattering (e.g. Stamnes et al.
1988, Schwander et al. 2001), (2) physical models with a simple parameterization for
discrete wavelengths (e.g. Diffey 1977, Schippnick and Green 1982, Schauberger et al.
1997) and (3) statistical models (Canadian model – Burrows et al. 1994, CHMI
832 K. La´ska et al.
model – Vanicek 1997, ETH Zu¨rich model – Renaud et al. 2000, Polish IMGW model –
Krzyscin et al. 2001). The radiative transfer models in particular require extensive
computation, many assumptions about the atmospheric parameters (ozone, temperature,
aerosols) and their vertical distributions, which are generally difficult to obtain. The
simplest group of statistical models uses regression equations, which are obtained by
fitting the observed UV radiation to the selected file of input parameters (solar zenith
angle, TOC). These models can incorporate a number of fixed input parameters such as
aerosol amount, trace gases or regional surface albedo (snow or snowless ground). To
consider attenuation of the solar UV radiation by the clouds, many of the models use a
cloud modification factor (CMF). The CMF for global solar and UV radiation is
defined as the ratio of all-sky to clear-sky irradiance (e.g. Calbo´ et al. 2005).
Most of the UV radiation models were designed and parameterized outside
Antarctica. The specific feature of the Antarctic continent and its atmosphere, such
as high surface albedo and the highest temporal variation of TOC and cloudiness at
50–70
S, made it difficult to simply transfer the models without previous validation.
Only a few models have been validated at a global scale and used for UV forecasting
or reconstruction of UV radiation in the past (e.g. Schmalwieser et al. 2002). We
therefore developed our own model (La´ska et al. 2009), corresponding to both the
character of the available data and the specific conditions of Antarctica.
This paper presents a comprehensive analysis of stratospheric ozone, cloudiness
and solar UV radiation measured at Vernadsky Station over a 3-year investigation. A
new approach to the estimation of solar UV radiation based on a nonlinear regression
model is presented. This work is a continuation of our previous research on stratospheric
ozone and UV radiation issues carried out in maritime Antarctica since 1995
(e.g. Prosˇek and Janouch 1996, Prosˇek et al. 2001, La´ska et al. 2009). The model
validation is carried out with respect to the temporal variation of TOC and the
different categories of cloudiness. To evaluate the model features and document its
usefulness, normal-probability plots and other statistical methods are used.
2. Materials and methods
The measurements in this paper were taken at the Ukrainian Vernadsky Station in
Antarctica (formerly the British Faraday Station) with geographical coordinates f ¼ 65
150
S; l ¼ 64
150
W; h ¼ 11 m a.s.l. Vernadsky Station is located on Galinde´z Island in the
Argentinean Islands off the western coast of the Antarctic Peninsula (see figure 1). Global
solar radiation was measured by a Kipp&Zonen CM-6B pyranometer. The measurement
was carried out at sampling interval of 10 s from which 10-min average values were
recorded to the MiniCube VV/VX Data Logger (EMS, Czech Republic). Then, daily
sums of global solar radiation were evaluated as the 10-min average values for each day.
Erythemally effective ultraviolet (EUV) radiation, according to the McKinlay and
Diffey (1987) Erythemal Action Spectrum, was measured with a Solar Light broadband
UV-Biometer 501 Version 3. The instrument was set to keep the temperature stable at
25
C and store data as 10-min integral values. In this study, only daily sums (doses) of
EUV radiation calculated from the corresponding 10-min values were used for the
model parameterization. The spectral response of the UV-Biometer ranges from 280
to 400 nm while its spectral sensitivity is similar to the sensitivity of the human skin to the
acute effect of erythema. With regard to the construction of this instrument, it should be
pointed out that the EUV radiation intensity was measured in minimum erythema dose
per hour (MED h–1
). One MED h-1
is equivalent to the UV radiation intensity required
UV radiation in maritime Antarctica 833
for minimum irritation (i.e. reddening) of the skin of average pigmentation after 1 h
exposure. Conversion into physical units is accomplished by means of the relationship: 1
MED h-1
¼ 0.0583 W m-2
, or 210 kJ m-2
h-1
. The solar radiation instruments were set
up on the roof of Vernadsky Station building on 12 February 2002. To protect them
from snow and ice accumulation, the instruments were equipped with a Kipp&Zonen
CV2 heater ventilation unit. Both radiometers were sent to the Solar and Ozone
Observatory of the Czech Hydrometeorological Institute in Hradec Kra´love´ for calibration
checks using a Brewer MKIV spectrophotometer and the Czech national standard
UV-Biometer and CM-11 pyranometer every 2 years.
Ozone concentration data can be obtained either from ground measurements
predominantly taken with Dobson and Brewer ozone spectrophotometers or from
satellite measurements. TOC is a parameter that is used frequently to determine the
ozone content of the atmosphere. Recently, many authors have assessed the uncertainties
in the TOC from satellite measurements as well as differences between ground
and satellite observations (e.g. Labow et al. 2004, Vanicek 2006, Schmalwieser et al.
2007). In this study, TOC data from both ground and satellite measurements were
applied to eliminate the uncertainties in measured ozone concentration. The first TOC
series based on the Dobson No. 031 and Dobson No. 123 (with modern electronics,
since 27 March 2005) spectrophotometer measurements at Vernadsky Station was
applied. The second set of TOC data was derived from the Total Ozone Mapping
Spectrophotometer (TOMS) onboard NASA’s Earth Probe Satellite (EPTOMS). The
TOC values were acquired for geographical coordinates of Vernadsky Station from
the daily gridded ozone of EPTOMS Version 8 (http://toms.gsfc.nasa.gov/ftpdata.html).
When the ground-based measurements were inaccurate (e.g. poor weather
conditions, low solar elevation angle), satellite data were applied. In this case, the
relative air mass of the ozone layer (m) was a criterion of the data quality (Kasten and
Young 1989). In addition, the daily mean cloudiness based on 3-h synoptic
Figure 1. Northern part of the Antarctic Peninsula showing the location of Vernadsky Station.
834 K. La´ska et al.
observations of total cloud amount (in oktas) at Vernadsky Station was used for the
purpose of model validation.
To estimate the EUV radiation reaching the Earth’s surface, a new approach based
on a nonlinear regression model with hyperbolic transmissivity function was developed
(La´ska et al. 2009). The daily sum of EUV radiation was calculated as
EUV ¼
Ig
Igex
 b0
EUVex
1 þ b1TOC
b2
eff
(1)
where Ig is the daily sum of global solar radiation intensity incident on the Earth’s
surface (MJ m-2
day-1
), Igex is the daily extraterrestrial sum of global solar radiation
intensity (MJ m-2
day-1
), EUVex is the daily extraterrestrial sum of erythemally
effective UV radiation (kJ m-2
day-1
), TOCeff is the effective total ozone content
(DU), and b0, b1 and b2 are regression parameters.
The characteristics of the atmosphere can be expressed by the ratio of the extraterrestrial
global radiation intensity (Igex) and the intensity of this radiation incident
on the Earth’s surface. Both Igex and EUVex at the top of the atmosphere (TOA) can
be easily estimated from theory for any geographical locality and time. Ig and EUV
radiations are measured directly by a pyranometer and UV-Biometer, respectively. b0
approximates the dissimilar absorption of global radiation and UV radiation in the
atmosphere, while b1 and b2 express the ozonosphere effectiveness in the absorption
and diffusion of UVB radiation. The absorption of UVB radiation increases with
increasing ozone content in the atmosphere and with increasing length of the solar ray
path through the atmosphere. This situation occurs particularly at low heights of the
Sun. In Antarctica, the amount of the aerosols is usually very low, and therefore it was
not taken into consideration. Moreover, it was assumed that the ratio of Ig to Igex
partly takes into consideration the seasonal variation in surface albedo. To eliminate
the heteroscedasticity of EUV and EUVex radiation, both parameters were adjusted
by the square root transformation.
Instead of using the usual ozone transmissivity exponential function (see
e.g. Kondratyev and Varotsos 1995, 1996), we have proposed a new approach
based on a quantum transmission model of UVB radiation. This type of transmission
is described by a hyperbolic function (La´ska et al. 2009). To compare the differences
between the exponential and hyperbolic ozone transmissivity functions, theoretical
TOC values in the atmosphere were used for calculation of the ozone transmissivity
values (figure 2). It can be seen that both ozone transmissivity functions are identical
up to 200 DU, while the exponential function overestimates TOC values (attenuates
the UVB radiation) more than the hyperbolic distribution at the lower TOC. In
equation (1), attenuation of solar UVB radiation related to optical phenomena such
as absorption (except the ozone absorption effect), reflection, refraction or scattering
in the atmosphere is expressed by the element (Ig/Igex)b0
. The other part of the
equation, that is the ozone absorption, is described by the element 1/(1 þ b1TOCeff
b2
).
TOC in the atmosphere is measured by spectrophotometers either from the Earth’s
surface or from satellites, which normally carry two spectrophotometers, one of
which is oriented towards the Earth’s surface and the other to outer space. With
respect to the position of the Sun, directly measured TOC has to be converted to
values of a so-called effective total ozone content (TOCeff) for the purposes of the
model (La´ska et al. 2009). This is to take account of the sunlight path through the
atmosphere to the Earth’s surface, which makes its actual path through the
UV radiation in maritime Antarctica 835
ozonosphere longer than towards the zenith. TOCeff is obtained by resolving the
geometrical problem of the sunlight path through spherical layers of the atmosphere:
TOCeff ¼
À2r sin a þ
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ð2r sin aÞ2
þ 4½ðr þ hÞ2
À r2Š ð1 þ sin2
aÞ
q
2ð1 þ sin2
aÞ
TOC
h
(2)
where r and h are, respectively, the radius of the Earth and the mean height of the
ozonosphere in kilometres, a is the daily mean solar elevation angle in radians, and
TOC is the daily mean total ozone column in DU.
For the purpose of the model, the ozone layer was divided into artificial layers. Each of
them contained an amount of ozone such that the intensity of UVB radiation that passed
the layer was reduced to one half of the initial intensity of radiation at the upper surface
of the layer. This means that there is twice as much ozone in each consecutive layer
towards the Earth. As the intensity of UVB radiation does not decrease in jumps but
smoothly, this consideration does not fully reflect the reality. Therefore, we multiplied
TOCeff in the real atmosphere by a coefficient b1, which is related to the UVB attenuation
due to the amount (van der Waals diameter) of ozone molecules in the atmosphere. It has
been reported (e.g. Feynman et al. 1989) that the effective diameter is usually higher than
the geometric one, which might alter the absorption effectiveness of the UVB. Then, the
higher the TOCeff, the higher the effectiveness of UVB quantum absorption by ozone
molecules. That was why we used the coefficient b2 in the exponential part of TOCeff.
3. Results and discussion
The seasonal variability of global solar (Ig) and EUV radiation intensities, TOC and
cloudiness at Vernadsky Station within the period from 1 August 2002 to 30 April
Figure 2. Values of ozone transmissivity functions based on hyperbolic and exponential
distribution with respect to the theoretical total ozone content in the atmosphere.
836 K. La´ska et al.
2005 is shown in figure 3. A logical similarity between cloudiness and high diurnal
variation in both solar radiation regimes is evident from this figure. The ozone
depletion period (September–October) is not clearly pronounced on the EUV radiation
intensities due to low solar elevation and short daytime. Figure 4 shows the
frequencies of the TOC values and daily mean cloudiness (in oktas) observed at
Vernadsky Station. Daily mean cloudiness reached 6.9 oktas in the period from 1
August 2002 to 30 April 2005. The high occurrence of total cloud amount from 7 to 8
oktas coincided with cyclogenesis in the circumpolar trough near 60
S. Figure 3 also
shows that TOC data were not available from May to July due to low solar elevation
angle. The weather conditions made it difficult to provide good quality observations
either on the zenith sky or against the Sun disc.
TOC values over Antarctica differ significantly in the analysed period, especially in
the ozone anomaly area with the ozone concentration below 220 DU (Newman et al.
2004). In 2002, the October average amounted to only 3.0 million km2
and its
development was very slow from August. At the end of September, it split into two
parts, which became extinct relatively rapidly on 1 October 2002. The average surface
area in October 2003 had already amounted to 9.8 million km2
, with the ozone
Figure 3. Daily sums of global solar radiation (Ig), erythemally effective UV (EUV) radiation,
daily means of cloudiness and total ozone content (TOC) at Vernadsky Station in the period
2002–2005. The TOC values were measured by the Dobson spectrophotometer.
UV radiation in maritime Antarctica 837
depletion formed as early as mid-August and retained without any significant change
until mid-October. It became rapidly extinct in the second half of November
(Stolarski et al. 2005). In 2004, the ozone hole started to grow during midSeptember.
Then in the next 3 weeks, the ozone hole increased in area by more than
23 million km2
. This slow increase was followed by a rapid decrease in area that ended
in mid-November, when the ozone hole disappeared. The ozone hole in 2004 was
much smaller than the average size over the past decade, and it was almost similar in
mid-October to 2002 when the ozone hole had split into two parts (WMO 2004).
While the variation in cloudiness determined shorter-term fluctuations of the
radiation fluxes, ozone decline caused solar radiation changes of prolonged periodicity
in the second half of the year. These impacts were furthermore overlaid by the
effect of cloud cover and cloud optical depth. In 2002, these effects were clearly
demonstrated; for example, in the period of ozone hole occurrence between 20
September and 22 October, and in the period from 23 October to 10 November. In
2003, a significant increase in EUV radiation between 25 September and approximately
12 October was related to the decline in ozone, while the subsequent increase in
ozone concentration from 13 October suppressed the natural increasing trend of EUV
radiation until late November. In 2004, a particularly significant ozone decline was
observed in two periods: 15–29 September and 4–14 October. However, in only the
second period was the EUV radiation increase pronounced. Maximum values of the
EUV radiation did not relate to the ozone minima (figure 3), but rather occurred in
mid-December, close to the summer solstice in the Southern Hemisphere. This fact
was well documented by the timing and frequency of occurrence of daily maximum
EUV radiation intensities and corresponding TOC values (see table 1). In 2002–2003
and 2004–2005, up to five events of EUV radiation intensities higher than 4 kJ m-2
were recorded at Vernadsky Station. By contrast, 10 days with maximum EUV
radiation were reported during partly cloudy and clear-sky conditions around the
summer solstice of 2003–2004.
Using the above-described nonlinear hyperbolic regression model, daily sums of the
EUV radiation intensities at Vernadsky Station were calculated. For the purpose of the
model parameterization, we selected 808 days from the original data set (2002–2005)
where both TOC and global solar radiation were available. The b0, b1 and b2
Figure 4. Relative frequency distribution of daily means of total ozone content (TOC) and
cloudiness at Vernadsky Station in the period 2002–2005. The TOC values were measured by
the Dobson spectrophotometer.
838 K. La´ska et al.
parameters of this model were estimated by the STATISTICA system and
Levenberg–Marquardt’s method using the least squares method with the initial approximations
of b0 ¼ 0.6, b1 ¼ 0.1, b2 ¼ 1.0. As seen in table 2, we could reject the hypotheses
about the nonsignificance of the regression parameters at the 0.05 significance level as
all of the 95% confidence intervals did not include 0. The relative estimates error of the
regression parameters was calculated as the half width of the 95% confidence interval to
estimate value, expressed as a percentage (see table 2). Estimates of the regression
parameters differed slightly from the previous study of La´ska et al. (2009) because of
the different period considered in this paper (2002–2004). However, the relative estimated
errors of the new regresssion parameters decreased significantly after the new
parameterization to half of the original value. This was probably caused by the square
root transformation applied to the EUV radiation data.
The model results for the period 2002–2005 are shown in figure 5 as the scatterplot
of measured versus predicted daily sums of EUV radiation intensities. The variability
in EUV radiation was explained by means of the model at 99.21% in the period
studied. As in our previous study (La´ska et al. 2009), the indexes of determination
reached high values. Two different statistical parameters were used for the model
Table 1. Maximum values of daily sums of global solar (Ig) and erythemally effective UV
(EUV) radiation, daily means of total ozone content (TOC), cloudiness (CL), and solar
elevation angle (a) at noontime recorded at Vernadsky Station within the period 2002–2005.
Date Ig (MJ m-2
) EUV (kJ m-2
) TOC (DU) CL (oktas) a (
)
1 November 2002 26.337 4.064 209 4.0 39.2
24 December 2002 32.652 4.204 288 5.9 48.2
29 December 2002 27.713 4.048 270 7.0 48.0
4 January 2003 33.910 4.122 285 5.3 47.5
25 November 2003 25.739 5.478 230 7.0 45.5
28 November 2003 32.464 4.635 257 2.5 46.1
4 December 2003 30.455 4.390 272 3.3 47.0
5 December 2003 34.286 5.157 249 3.6 47.1
6 December 2003 34.490 4.423 280 1.0 47.3
7 December 2003 34.907 4.332 285 1.1 47.4
8 December 2003 34.346 4.117 298 2.0 47.5
25 December 2003 34.501 4.126 303 5.3 48.2
28 December 2003 31.638 4.067 301 2.5 48.0
2 January 2004 34.171 4.217 287 6.6 47.7
29 November 2004 32.222 4.186 296 4.0 46.4
10 December 2004 31.736 4.161 307 5.0 47.8
11 December 2004 35.010 4.228 306 1.3 47.8
21 December 2004 35.455 4.524 280 0.8 48.2
30 December 2004 29.512 4.178 294 5.6 47.9
Table 2. Estimates of parameters and relative estimated errors calculated
with 95% confidence intervals for ground ozone measurements in the period
2002–2005.
Parameter Estimate Relative estimated error (%)
bb0 0.70 Æ 0.01 1.9
bb1 0.08 Æ 0.01 17.2
bb2 1.14 Æ 0.03 2.4
UV radiation in maritime Antarctica 839
validation: (1) residuals of the EUV radiation intensity in kJ m-2
and (2) mean average
prediction error (MAPE) expressed as a percentage.
The temporal variability of the EUV residuals in the period 2002-2005 is shown in
figure 6. It is apparent that the smallest EUV residuals (up to 0.08 kJ m-2
) occurred
only at low solar elevation (April, August). The highest EUV residuals occurred from
September to March independently of the ozone depletion. In addition, a notable
variation in the EUV residuals throughout the year can be seen in figure 6. The largest
absolute values of the EUV residuals reached -0.18 and þ0.24 kJ m-2
, respectively. As
seen in figure 7, the relative distribution of the largest EUV residuals (more than
Æ0.15 kJ m-2
) represented only 1% of all cases.
The MAPE of our model reached 4.4% over the period studied. This can be
compared with our previous results (La´ska et al. 2009), when both ground and
satellite ozone measurements were applied. In the latter case, MAPE values ranged
from 7.4% to 9.6% depending on the season and the method of ozone measurement.
The results were difficult to compare to other studies as only a few Antarctic stations
Figure 5. Scatterplot of daily sums of predicted erythemally effective UV radiation (predicted
EUV) and EUV radiation measured at Vernadsky Station in the period 2002–2005. The results
are shown for all types of cloudiness.
Figure 6. Residuals of the erythemally effective UV (EUV) radiation for all types of cloudiness
at Vernadsky Station in the period 2002–2005.
840 K. La´ska et al.
equipped with the same instruments carried out both measurements and estimations
of the EUV radiation intensities similarly. In general, our results are in agreement
with the findings of Weihs and Webb (1997), Koepke et al. (1998) and De Backer et al.
(2001). They found differences between measurements and models in the range 5–10%
with a statistical uncertainty of 2–3%.
The second part of our study was to examine the predicted EUV radiation intensities
and investigate how the variation in TOC and cloudiness affected the EUV
residuals in different seasons. EUV residuals were separated according to the solar
elevation angle (a) and the time interval of the ozone depletion occurrence as follows:
August–October (low a, the largest ozone depletion), November–January (high a,
ozone depletion retreat), and February–April (low a, no ozone depletion).
Furthermore, the EUV residuals were distinguished according to daily mean cloudiness
into three categories: partly cloudy (0–4.0 oktas), cloudy (4.1–7.0 oktas), and
overcast conditions (7.1–8.0 oktas).
Seasonal variability of TOC and cloudiness was calculated as the coefficients of
variation for the individual seasons of 2002–2005 (table 3). Then, model quality
expressed by the standard deviation of EUV residuals and MAPE was evaluated.
As shown in table 3, TOC variation decreased rapidly from spring to autumn (as a
consequence of the ozone depletion disappearing). However, the highest variation in
cloudiness occurred around the summer solstice (November–January). It is also
evident that the behaviour of the standard deviations of the EUV residuals and
MAPE did not relate to TOC but to the cloudiness. Hence, the smallest standard
deviations (up to 0.031 kJ m-2
) occurred in the same period as the lowest variability of
the cloudiness was observed (August–October and February–April). That was also
related to small a around both equinoxes. A significant dependence of the EUV
residuals on cloudiness was documented by MAPE. Contrary to the standard deviations,
the smallest values of MAPE were reported during high summer when they
ranged from 2.6% to 3.5%. The relative frequency distributions of the EUV residuals
confirmed slight differences between individual seasons of 2002–2005 (figure 8). It is
obvious that the largest dispersion and the lowest kurtosis of the EUV residuals
occurred annually within the summer months and was the effect of cloudiness variation.
On the contrary, high kurtosis and small dispersion occurred in both spring and
autumn months.
The quality of the regression model was also documented by the relative frequency
distributions of the EUV residuals according to different categories of cloudiness
(figure 7). A histogram of the EUV residuals for all cases of cloudiness was very
similar to the normal distribution. The largest EUV residuals were found in partly
cloudy conditions (0–4.0 oktas) with high variability of cloud types, their thickness
and brightness during a day (data not shown). In this case, MAPE increased to 5.9%.
On the contrary, the smallest EUV residuals documented by their large accumulation
close to zero were reported for cloudy and overcast days. Therefore, the resulting
distribution had high kurtosis, which was essentially greater than in the case of
normal distribution. Slightly better results were documented by means of MAPE,
ranging from 4.0% to 4.3%.
To assess the differences between empirical and theoretical distribution of the EUV
residuals, normal probability plots were used (figure 9). Similarly to previous outcomes,
slightly worse results were reached for partly cloudy conditions. This can be
partly explained by the fact that the model parameterization was undertaken for all
sky conditions, under which daily mean cloudiness reached 6.9 oktas. In addition, the
UV radiation in maritime Antarctica 841
Table3.Seasonalvariabilityoftotalozonecontent(TOC)andcloudinessexpressedbythecoefficientofvariationatVernadskyStationintheperiod
2002–2005.ModelqualityispresentedbythestandarddeviationoferythemallyeffectiveUV(EUV)residualsandthemeanaverageprediction
error(MAPE).
Coefficientofvariation(%)
Standarddeviation
PeriodTOCCloudinessEUVresiduals(kJm-2
)MAPE(%)
August2002–October200227.5019.140.0314.8
November2002–January200310.9822.100.0563.5
February2003–April20038.6912.110.0536.4
August2003–October200330.4616.430.0414.0
November2003–January20048.5226.410.0623.0
February2004–April20047.5118.480.0324.1
August2004–October200424.2225.460.0495.6
November2004–January20057.8422.120.0502.6
February2005–April20057.5721.450.0385.5
842 K. La´ska et al.
Figure 7. Relative frequency distribution of the residuals of erythemally effective UV (EUV)
radiation at Vernadsky Station in the period 2002–2005. The results are shown for different
categories of cloudiness (see §2).
Figure 8. Relative frequency distribution of the residuals of erythemally effective UV (EUV)
radiation for all types of cloudiness at Vernadsky Station. The results are shown for three consecutive
seasons: August–October, November–January and February–April in the period 2002–2005.
UV radiation in maritime Antarctica 843
partly cloudy conditions represented only 11% of all cases. Furthermore, the partly
cloudy category included thin and moderately thick clouds in which the strong
Rayleigh scattering of short wavelengths and consequent wavelength-dependent
directional distribution of the sky radiance were well documented (Kylling et al.
1997, Bernhard et al. 2004, Lindfors and Arola 2008). Figure 10 shows the relative
frequency distributions of the positive and negative EUV residuals according to the
cloudiness categories. Nearly identical shares of the positive and negative EUV
Figure 9. Normal probability plot of the ordered residuals of erythemally effective UV (EUV)
radiation from the nonlinear regression model against the ordered quantiles of the standard
normal distribution. The results are shown for different categories of cloudiness at Vernadsky
Station in the period 2002–2005.
Figure 10. Relative frequency distribution of the positive and negative residuals of erythemally
effective UV (EUV) radiation at Vernadsky Station in the period 2002–2005. The results
are shown for different categories of cloudiness (see §3).
844 K. La´ska et al.
residuals were found in the categories of 5.1–6.0 and 6.1–7.0 oktas, while in the
category of up to 4.0 oktas, negative EUV residuals dominated. On clear-sky days
(0–1.0 oktas), observed seven times within the period 2002–2005, only negative EUV
residuals were reported (see table 4). Moreover, overcast days (7.1–8.0 oktas), which
occurred 533 times at Vernadsky Station, provided the best results, among others,
with the mean EUV residuals of 0.007 kJ m-2
. However, the range of the EUV
residuals increased progressively from clear sky (0.123 kJ m-2
) to overcast conditions
(up to 0.388 kJ m-2
). The consequences may be that our model either underestimates
or overestimates the EUV radiation intensities at an expressively uneven regime of
solar radiation within the day period (particularly at a different radiation intensity
around noontime as compared with the morning and evening hours). This is also
likely to affect the magnitude of the errors for the parameters b1 and b2, which express
the general ozone content, its role in UVB absorption, and scattering.
4. Conclusions
In this paper we have demonstrated a new approach to the estimation of EUV
radiation using a nonlinear regression model with a hyperbolic transmissivity
function. Parameterization of the model was performed on daily sums and
daily mean values of the selected input parameters such as solar elevation
angle, global solar radiation and TOC measured at Vernadsky Station during
2002–2005.
Model validation showed that there were no significant differences for the cloudy
(4.1–7.0 oktas) and overcast conditions (7.1–8.0 oktas). MAPE for the two categories
ranged between 4.0% and 4.3%, while for partly cloudy days (0–4.0 oktas)
with high variability of cloud types, their MAPE reached 5.9%. These results were
mainly affected by the distribution of cloud cover within the analysed period.
Cloudiness in the category of 4.1–8.0 oktas represented 89% of all cases.
Therefore, estimates of the regression parameters were highly affected by days
with a large amount of clouds. Under clear-sky or partly cloudy conditions, the
differences between the absorption of global solar and UV radiation varied significantly,
therefore the model results based on daily mean values were slightly worse
than for overcast days. This problem could be reduced by using a different type of
absorption, which will be the subject of further study. In addition, an improvement
to the surface albedo approximation and the estimation of hourly EUV radiation
doses in our model will be addressed.
Acknowledgements
We thank all employees of the National Antarctic Scientific Centre of Ukraine in
Kyiv and members of the crews wintering on Vernadsky Station, and Andrey
Zalizowski, Alexander Kolesnik and Igor Gvozdovsky for their assistance and faultless
operation of the radiation instruments. We also appreciate the valuable discussions
with Milosˇ Barta´k, who helped to improve the manuscript. This study was
undertaken as a part of the research programme ‘The effects of atmospheric factors
on the regime of UV radiation of Antarctic Peninsula’ (Grant No. 205/07/1377)
funded by the Grant Agency of the Czech Republic.
UV radiation in maritime Antarctica 845
Table4.SelectedstatisticalcharacteristicsoftheresidualsoferythemallyeffectiveUV(EUV)radiationcalculatedforparticularcategoriesofcloudinessat
VernadskyStationintheperiod2002–2005.
Cloudinesscategory
0–1.01.1–2.02.1–3.03.1–4.04.1–5.05.1–6.06.1–7.07.1–8.0
Numberofdays7119223772117533
RelativefrequencyofEUVresiduals,0100.090.977.863.648.772.252.139.8
RelativefrequencyofEUVresiduals.00.09.122.236.451.327.847.960.2
MeanEUVresiduals(kJm–2
)-0.076-0.079-0.040-0.025-0.010-0.016-0.0020.007
MinimumEUVresiduals(kJm–2
)-0.149-0.175-0.129-0.120-0.132-0.141-0.152-0.177
MaximumEUVresiduals(kJm–2
)–0.0250.0040.0240.0970.0620.0850.2360.169
RangeofEUVresiduals(kJm–2
)0.1230.1780.1530.2180.1940.2270.3880.346
846 K. La´ska et al.
References
BERNHARD, G., BOOTH, C.R. and EHRAMJIAN, J.C., 2004, Version 2 data of the National Science
Foundation’s Ultraviolet Radiation Monitoring Network: South Pole. Journal of
Geophysical Research, 109, D21207, doi:10.1029/2004JD004937.
BURROWS, W.R., VALLE´ E, M., WARDLS, D.I., KERR, J.B., WILSON, L.J. and TARASICK, D.W.,
1994, The Canadian operational procedure for forecasting total ozone and UV radiation.
Meteorological Applications, 1, pp. 247–265.
CALBO´ , J., PAGS, D. and GONZALEZ, J.-A., 2005, Empirical studies of cloud effects on UV
radiation: a review. Reviews of Geophysics, 43, RG2002, doi:10.1029/2004RG000155.
DE BACKER, H., KOEPKE, P., BAIS, A., DE CABO, X., FREI, T., GILLOTAY, D., HAITE, CH.,
HEIKKILA¨ , A., KAZANTZIDIS, A., KOSKELA, T., KYRO¨ , E., LAPETA, B., LORENTE, J.,
MASSON, K., MAYER, B., PLETS, H., REDONDAS, A., RENAUD, A., SCHAUBERGER, G.,
SCHMALWIESER, A., SCHWANDER, H. and VANICEK, K., 2001, Comparison of measured
and modelled UV indices for the assessment of health risks. Meteorological
Applications, 8, pp. 267–277.
DE FABO, E.C., NOONAN, F.P. and FREDERICK, J.E., 1990, Biologically effective doses of sunlight
for immune suppression at various latitudes and their relationship to changes in
stratospheric ozone. Photochemistry and Photobiology, 52, pp. 811–817.
DIFFEY, B.L., 1977, The calculation of the spectral distribution of natural ultraviolet radiation
under clear day conditions. Physics in Medicine and Biology, 22, pp. 309–316.
FARMAN, J.C., GARDINER, B.G. and SHANKLIN, J.D., 1985, Large losses of total ozone in
Antarctica reveal seasonal CClOx/NOx interaction. Nature, 315, pp. 207–210.
FEYNMAN, R.P., LEIGHTON, R.B. and SANDS, M., 1989, The Feynman Lectures on Physics, 3 vols
(Redwood City, CA: Addison-Wesley).
FREDERICK, J.E., QU, Z. and BOOTH, C.R., 1998, Ultraviolet radiation at sites on the Antarctic
coast. Photochemistry and Photobiology, 68, pp. 183–190.
HERMANN, J.R., NEWMAN, P.A. and LARKO, D., 1995, Meteor-3/TOMS observations of the
1994 ozone hole. Geophysical Research Letters, 22, pp. 3227–3229.
KASTEN, F. and YOUNG, A. T., 1989, Revised optical air mass tables and approximation
formula. Applied Optics, 28, pp. 4735–4738.
KING, J.C. and TURNER, J., 1997, Antarctic Meteorology and Climatology (Cambridge, UK:
Cambridge University Press).
KOEPKE, P., BAIS, A., BALIS, D., BUCHWITZ, M., DE BACKER, H., DE CABO, X., ECKERT, P.,
ERIKSEN, P., GILLOTAY, D., KOSKELA, T., LAPETA, B., LITYNSKA, Z., LORENTE, J., MAYER,
B., RENAUD, A., RUGGABER, A., SCHAUBERGER, G., SECKMEYER, G., SEIFERT, P.,
SCHMALWIESER, A., SCHWANDER, H., VANICEK, K. and WEBER, M., 1998, Comparison
of models used for UV index-calculations. Photochemistry and Photobiology, 67, pp.
657–662.
KONDRATYEV, K.Y. and VAROTSOS, C.A., 1995, Atmospheric ozone variability in the context of
global change. International Journal of Remote Sensing, 16, pp. 1851–1881.
KONDRATYEV, K.Y. and VAROTSOS, C.A., 1996, Global total ozone dynamics – impact on surface
solar ultraviolet radiation variability and ecosystems. 1. Global ozone dynamics and
environmental safety. Environmental Science and Pollution Research, 3, pp. 153–157.
KRZYSCIN, J.W., JAROSLAWSKI, J. and SOBOLEWSKI, P., 2001, On an improvement of UV Index
forecast: UV index diagnosis and forecast for Belsk, Poland, in spring/summer 1999.
Journal of Atmospheric and Solar-Terrestrial Physics, 63, pp. 1593–1600.
KYLLING, A., ALBOLD, A. and SECKMEYER, G., 1997, Transmittance of a cloud is wavelengthdependent
in the UV-range: physical interpretation. Geophysical Research Letters, 24,
pp. 397–400.
LABOW, G.J., MCPETERS, R.D. and BHARTIA, P.K., 2004, A comparison of TOMS & SBUV
Version 8 total column ozone data with data from ground stations. In Proceedings of
the 20th Quadrennial Ozone Symposium, C.S. Zerefos (Ed.), pp. 123–124 (Kos, Greece:
University of Athens).
UV radiation in maritime Antarctica 847
LA´ SKA, K., PROSˇEK, P., BUDI´K, L., BUDI´KOVA´ , M. and MILINEVSKY, G., 2009, Prediction of
erythemally effective UV radiation by means of nonlinear regression model.
Environmetrics, 20, pp. 633–646.
LINDFORS, A. and AROLA, A., 2008, On the wavelength-dependent attenuation of UV radiation
by clouds. Geophysical Research Letters, 35, L05806, doi:10.1029/2007GL032571.
LUCCINI, E., CEDE, A. and PIACENTINI, R.D., 2003, Effect of clouds on UV and total irradiance
at Paradise Bay, Antarctic Peninsula, from a summer 2000 campaign. Theoretical and
Applied Climatology, 75, pp. 105–116.
MCKINLAY, A. and DIFFERY, B.L., 1987, A reference action spectrum for ultraviolet induced
erythema in human skin. In Human Exposure to Ultraviolet Radiation: Risks and
Regulations, W.R. Passchler and B.F.M. Bosnajokovic (Eds), pp. 83–87 (Amsterdam:
Elsevier).
NEWMAN, P.A., KAWA, S.R. and NASH, E.R., 2004, On the size of the Antarctic ozone hole.
Geophysical Research Letters, 31, L21104, doi:10.1029/2004GL020596.
PROSˇEK, P. and JANOUCH, M., 1996, The measurement of ultraviolet radiation at the Polish
Henryk Arctowski Station (South Shetlands, Antarctica) in the summer of 1994/95.
Problemy Klimatologii Polarnej, 6, pp. 139–148.
PROSˇEK, P., LA´ SKA, K. and JANOUCH, M., 2001, Ultraviolet radiation intensity at the
H. Arctowski Base (South Shetlands, King George Island) during the period from
December 1994–December 1996, Brno, Czech Republic. In Ecology of the Antarctic
Coastal Oasis, P. Prosˇek and K. La´ska (Eds), pp. 35–48 (Brno, Czech Republic: Folia
Facultatis Scientarum Naturalium Universitatis Masarykianae Brunensis, Geographia 25).
RENAUD, A., STAEHELIN, J., FRO¨ HLICH, C., PHILIPONA, R. and HEIMO, A., 2000, Influence of snow
and clouds on erythemal UV radiation: analysis of Swiss measurements and comparison
with models. Journal of Geophysical Research, 105, pp. 4961–4969.
SCHAUBERGER, G., SCHMALWIESER, A.W., RUBEL, F., WANG, Y. and KECK, G., 1997, UV-Index:
Operational forecast of the solar, biologically effective ultraviolet radiation in Austria
[in German]. Zeitschrift fuer Medizinische Physik, 7, pp. 153–160.
SCHIPPNICK, P.F. and GREEN, A.E.S., 1982, Analytical characterisation of spectral actinic flux and
spectral irradiance in the middle ultraviolet. Photochemistry and Photobiology, 35, pp.
89–101.
SCHMALWIESER, A.W., SCHAUBERGER, G., ERBERTSEDER, T., JANOUCH, M., COETZEE, J.R.C. and
WEIHS, P., 2007, Sensitivity of erythemally effective UV irradiance and daily exposure to
uncertainties in measured total ozone. Photochemistry and Photobiology, 83, pp. 433–443.
SCHMALWIESER, A.W., SCHAUBERGER, G., JANOUCH, M., NUNEZ, M., KOSKELA, T., BERGER, D.,
KARAMANIAN, G., PROSEK, P. and LA´ SKA, K., 2002, Global validation of a forecast
model for irradiance of solar, erythemally effective ultraviolet radiation. Optical
Engineering, 41, pp. 3040–3050.
SCHWANDER, H., KAIFEL, A., RUGGABER, A. and KOEPKE, P., 2001, Spectral radiative-transfer
modeling with minimized computation time by use of a neural-network technique.
Applied Optics, 40, pp. 331–335.
STAMNES, K., TSAY, S.C., WISCOMBE, W. and JAYAWEERA, K., 1988, Numerically stable algorithm
for discrete-ordinate-method radiative transfer in multiple scattering and emitting
layered media. Applied Optics, 27, pp. 2502–2509.
STATISTICA, 2008, STATISTICA software. Available online at: http://www.statsoft.com/
products/analytics.htm and http://www.statsoft.com/textbook/stnonlin.html (accessed
28 August 2008).
STOLARSKI, R.S., MCPETERS, R.D. and NEWMAN, P.A., 2005, The ozone hole of 2002 as
measured by TOMS. Journal of Atmospheric Sciences, 62, pp. 716–720.
VANICEK, K., 1997, Relation between total ozone and erythemal solar radiation on clear days in
Hradec Kralove. In Proceedings of the Workshop on Monitoring of UVB Radiation on
Total Ozone pp. 105–107 (Poprad, Slovak Republic: Slovak Hydrometeorological
Institute).
848 K. La´ska et al.
VANICEK, K., 2006, Differences between ground Dobson, Brewer and satellite TOMS-8,
GOME-WFDOAS total ozone observations at Hradec Kralove, Czech. Atmospheric
Chemistry and Physics, 6, pp. 5163–5171.
VAROTSOS, K., 2002, The southern hemisphere ozone hole split in 2002. Environmental Science
and Pollution Research, 9, pp. 375–376.
WARREN, S.G., HAHN, C.J., LONDON, J., CHERVIN, R.M. and JENNE, R.L., 1986, Global
Distribution of Total Cloud and Cloud Type Amounts over Land. NCAR Technical
Note TN-273þSTR/DOE Technical Report ER/60085-HI (Boulder, CO: NCAR).
WEIHS, P. and WEBB, A.R., 1997, Accuracy of spectral UV model calculations. 2. Comparison
of UV calculations with measurements. Journal of Geophysical Research, 102, pp.
1551–1560.
WMO, 2002, Antarctic Ozone Bulletin 2002. Available online at: http://www.wmo.ch/pages/
prog/arep/gawozobull02_en.html (accessed 28 November 2007).
WMO, 2003, Antarctic Ozone Bulletin 2003. Available online at: http://www.wmo.ch/pages/
prog/arep/gawozobull03_en.html (accessed 28 November 2007).
WMO, 2004, Antarctic Ozone Bulletin No. 7/2004. Available online at: http://www.wmo.int/
pages/prog/arep/04/bulletin_7_2004.pdf (accessed 23 November 2004).
WMO, 2006, Scientific Assessment of Ozone Depletion: 2006. Global Ozone Research and
Monitoring Project, Report No. 50 (Geneva, Switzerland: World Meteorological
Organization).
WMO, 2008, Antarctic Ozone Bulletin No. 1/2008. Available online at: http://www.wmo.int/
pages/prog/arep/gaw/ozone/index.html (accessed 28 August 2008).
UV radiation in maritime Antarctica 849
4
Paper 3
Method of estimating solar UV radiation in high-latitude
locations based on satellite ozone retrieval with an improved
algorithm
Láska, K., Budík, L., Budíková, M., Prošek, P., 2011,
International Journal of Remote Sensing, 31, pp. 831–849
International Journal of Remote Sensing
Vol. 32, No. 11, 10 June 2011, 3165–3177
Method of estimating solar UV radiation in high-latitude locations based
on satellite ozone retrieval with an improved algorithm
K. LÁSKA*†, L. BUDÍK‡, M. BUDÍKOVÁ§ and P. PROŠEK†
†Department of Geography, Faculty of Science, Masaryk University, Brno, 611 37,
Czech Republic
‡Czech Hydrometeorological Institute, Brno Regional Office, Brno, 616 67,
Czech Republic
§Department of Mathematics and Statistics, Faculty of Science, Masaryk University,
Brno, 611 37, Czech Republic
The effects of the cloudiness and satellite-based ozone measurements on erythemally
effective ultraviolet (EUV) radiation were examined using a non-linear
regression model. Instead of the widely used ozone transmissivity exponential
function, we proposed a new approach based on a quantum transmission model
using hyperbolic attenuation of the EUV radiation. The radiation data were collected
at the Czech Johann Gregor Mendel Station, James Ross Island, Antarctica
(63◦
48 S, 57◦
53 W), between 14 March 2007 and 3 March 2009. The total ozone
content and effective surface reflectivity at 360 nm were obtained from the Ozone
Monitoring Instrument on board the EOS-Aura spacecraft for the geographical
coordinates of the J. G. Mendel Station. The model predicted 98.6% variability of
the EUV radiation. The residuals between the measured and predicted EUV radiation
intensities were evaluated separately for the ranges of solar elevation angle,
total ozone content and surface reflectivity. The results of this study were compared
to previous findings where the influence of ground-based and satellite-based ozone
measurements and model usefulness was discussed.
1. Introduction
Over the last two decades, there has been a large-scale depletion of the stratospheric
ozone in Antarctica, except in 2002, when an unprecedented major sudden stratospheric
warming resulted in an almost no-ozone-hole episode (Varotsos 2002, 2003,
2004). Therefore, an intensive investigation has been carried out there into both
the physical and chemical changes in the atmosphere and the effects of rising solar
ultraviolet (UV) radiation on polar ecosystems. Recent studies indicate that both
atmospheric circulation and stratospheric temperature play an essential role in the
intensity of ozone losses over both hemispheres (e.g. Varotsos and Cracknell 1993,
Kondratyev et al. 1994, Varotsos and Cracknell 1994, Chandra and Varotsos 1995,
Gernandt et al. 1995, Varotsos et al. 1995, Varotsos et al. 2000, Efstathiou et al. 2003,
Balis et al. 2009, Hofmann et al. 2009). Apart from that, the consecutive deceleration
of ozone losses over Antarctica during the last decade was reported. It was supported
both by the ground-based measurements from the ozonometric stations and satellitebased
data (Kravchenko et al. 2009). In high latitude locations, the cloudiness and
*Corresponding author. Email: laska@sci.muni.cz
International Journal of Remote Sensing
ISSN 0143-1161 print/ISSN 1366-5901 online © 2011 Taylor & Francis
http://www.tandf.co.uk/journals
DOI: 10.1080/01431161.2010.541513
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3166 K. Láska et al.
total ozone content in the atmosphere are the most important factors affecting the
incident UV radiation. As shown by several authors, the amount of cloud, cloud position
relative to the sun, cloud type and its optical properties strongly modulate the
course of incident UV radiation (e.g. Varotsos 1994, Kondratyev et al. 1995, Varotsos
et al. 1995, Efstathiou et al. 1998, Frederick et al. 1998, Ziemke et al. 2000, Varotsos
et al. 2001, Luccini et al. 2003, Weihs et al. 2008).
Therefore, it is important to understand both the geographical distribution of
stratospheric ozone over Antarctica and the changes in solar UV radiation during
the last decade. Solar UV radiation plays a very significant role in many biological
processes, including harmful effects on DNA, human skin and the overall immune system
(De Fabo et al. 1990). The biological effectiveness of UV radiation depends on the
wavelength and can be either measured or estimated by various theoretical approaches.
However, the specific features of the Antarctic, such as high surface albedo, the seasonal
variations in the extent of sea ice, the atmosphere dynamics affecting the highest
temporal variation of total ozone content (TOC) and cloudiness may complicate
forecasting or the reconstruction of solar UV radiation. Moreover, most of the UV
radiation models were designed and parameterized outside the Antarctic vortex. Only
a few models have been validated on a global scale and used to forecast UV radiation
intensity (e.g. Schmalwieser et al. 2002, Varotsos et al. 2004, Cracknell and Varotsos
2007).
In order to study and evaluate the effects of atmospheric factors on UV radiation,
we proposed a new statistical model (Láska et al. 2009) taking into consideration
both the structure of the available data and the specific conditions of Antarctica.
This paper presents the results from the first three years of solar radiation measurement
at the Czech Johann Gregor Mendel Station, for which a new version of the
non-linear regression model with an improved algorithm for estimating erythemally
effective UV radiation is tested. Furthermore, recent validation results based on satellite
measurements of the total ozone content and effective surface reflectivity in high
latitude locations are discussed. This work is a continuation of the previous research
on stratospheric ozone and UV radiation issues carried out in maritime Antarctica
since 1995 (e.g. Prošek and Janouch 1996, Prošek et al. 2001, Láska et al. 2009, 2010).
2. Materials and methods
Measurements of solar radiation and the other meteorological observations were performed
at the Czech Johann Gregor Mendel Station (hereafter, the Mendel Station)
situated on the northern tip of James Ross Island, Antarctica (ϕ = 63◦
48 S;
λ = 57◦
53 W; 7 m above sea level). James Ross Island is a relatively large island
(2500 km2
) located off the south-east coast of the Antarctic Peninsula (see figure 1).
The Peninsula ridge, approximately 1000 m high, is 20 km north of the Mendel Station
across the Prince Gustav Channel. The Mendel Station was established in 2006 on the
marine terrace about 200 m southward from the coastal line. The recent climatological
program is focused mainly on the following objectives: (1) to study atmospheric circulation
and climate change along the Antarctic Peninsula, (2) to evaluate the impact of
regional atmospheric warming on deglaciation processes and (3) to estimate the effect
of atmospheric factors on solar UV radiation. Therefore, special attention is devoted
to monitoring solar radiation and its components, photosynthetically active radiation,
UVA, UVB (ultraviolet-A and ultraviolet-B), erythemally effective UVB and global
UV radiation in particular.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3167
Figure 1. Location of the Czech Johann Gregor Mendel Station in the northern part of James
Ross Island.
In this study, global solar radiation (Ig) and erythemally effective ultraviolet (EUV)
radiation was analysed between 14 March 2007 and 3 March 2009. Ig radiation was
measured with a CM-11 pyranometer (Kipp & Zonen, The Netherlands) with a spectral
interval between 305 nm and 2800 nm. EUV radiation, according to the McKinlay
and Diffey (1987) Erythemal Action Spectrum, was measured with a broadband UVBiometer
501A Version 3 (Solar Light, Glenside, PA, USA). The construction of the
UV-Biometer was based on experience gained during the Robertson–Berger meter
design (Berger 1976), thus its spectral intervals ranged between 280 nm and 400 nm.
The nominal temperature stabilization was set to and maintained at 25◦
C. With regard
to the construction of the UV-Biometer, it should be noted that the EUV radiation was
measured at the minimum erythema dose per hour (MED h−1
). It was then converted
into physical units afterwards by the equivalent: 1 MED h−1
= 0.0583 W m−2
or 210
kJ m−2
h−1
.
Both instruments were installed at a height of 4 m on a special platform on the
top of a technical container (figure 2). To protect them from snow and ice, the
instruments were equipped with Kipp & Zonen CV2 heater ventilation units. The
radiometers were thoroughly calibrated at the Solar and Ozone Observatory of the
Czech Hydrometeorological Institute in Hradec Králové using a Brewer MKIV spectrophotometer
(Kipp & Zonen, Canada), the Czech national standard UV-Biometer
and a CM-11 pyranometer every year. The measurements were taken at a sampling
interval of 10 s. From this basic data set, 10-min average values were calculated and
recorded on the MiniCube VV/VX Data Logger (EMS, Czech Republic). Then, the
daily sums of Ig and EUV radiation were calculated from the 10-min average values
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3168 K. Láska et al.
Figure 2. The radiometers equipped with the Kipp & Zonen CV2 ventilation units installed
on a special platform at the Mendel Station.
for each day. The annual course of Ig and EUV radiation intensities at the Mendel
Station between 2007 and 2009 is shown in figure 3. The maximum daily sums of Ig
and EUV radiation are ∼35 MJ m−2
and 5 kJ m−2
respectively. The high diurnal and
seasonal variability of both parameters is affected mainly by cloudiness and cloud
genera, together with cyclogenesis over the Antarctic Peninsula.
In order to estimate the solar EUV radiation, a non-linear regression model with a
hyperbolic transmissivity function was used. Instead of the widely used ozone transmissivity
exponential function (see, e.g., Kondratyev and Varotsos (1995, 1996)), we
proposed a new approach based on a quantum transmission model using hyperbolic
attenuation of the EUV radiation (Láska et al. 2009, 2010). To run this model, it is
necessary to know the information on the date, time, latitude and longitude, solar elevation
angle, TOC and Ig radiation. In this paper, the estimation of the daily dose of
EUV radiation with the improved algorithm is described as
EUV =
EUVex
1 + β0
Igex −Ig
Ig
1
1 + β1TOCeff
, (1)
where Ig is the daily sum of global solar radiation that incidents on the Earth’s surface
(MJ m−2
d−1
), Igex
is the daily extraterrestrial sum of global solar radiation intensity
(MJ m−2
d−1
), EUVex is the daily extraterrestrial sum of erythemally effective UV
radiation (kJ m−2
d−1
), TOCeff is the effective total ozone content (DU) and β0 and
β1 are the regression parameters. Both Igex
and EUVex at the top of the atmosphere can
be easily estimated for any geographical location, date and time. Ig and EUV radiation
are measured directly using a pyranometer and UV-Biometer respectively. To eliminate
the heteroscedasticity of EUV and EUVex radiation, the square-root transformation
was applied on both sides of the regression equation (1).
The previous version of our model (Láska et al. 2009) used three regression parameters
instead of two. However, the performance of the model and the results achieved
by the recent modification of the model were identical. Therefore, β0 is the ratio of
the absorption coefficient of EUV radiation in the atmosphere without ozone and
the absorption coefficient of Ig radiation in the atmosphere, while β1 expresses the
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3169
1 Mar 07 1 Jul 07 1 Nov 07 1 Mar 08 1 Jul 08 1 Nov 08 1 Mar 09
1 Mar 07 1 Jul 07 1 Nov 07 1 Mar 08 1 Jul 08 1 Nov 08 1 Mar 09
1 Mar 07 1 Jul 07 1 Nov 07 1 Mar 08 1 Jul 08 1 Nov 08 1 Mar 09
Figure 3. Annual course of daily sums of global solar radiation (Ig), erythemally effective UV
radiation (EUV) and daily mean total ozone content (TOC) at the Mendel Station between 2007
and 2009.
effectiveness of the ozonosphere in absorbing and diffusing EUV radiation. In this
approach, the quantum transmission model of EUV and Ig radiation is described by
a hyperbolic function (Láska et al. 2010). The first fraction in the model describes the
transmission of EUV radiation in atmosphere without ozone, as computed from the
measured Ig transmission. The transmission of Igex
to Ig radiation is calculated as
Ig =
Igex
1 + α1
. (2)
The transmission of EUVex to EUV radiation without ozone effects is calculated as
EUV =
EUVex
1 + α2
(3)
where α1 and α2 are the transmission coefficients and is the optical thickness of
the atmosphere. In order to resolve the first part of our model, from equation (2)
was computed and inserted into equation (3). Then α2/α1 can be expressed as β0. The
second fraction of the model is the transmission of EUVex radiation including ozone
effects but without the influence of the other parts of the atmosphere. Moreover, the
amount of aerosol emissions and their seasonal variation in Antarctica is usually very
low and so is not taken into consideration. Apart from that, the seasonal variation of
the surface albedo is assumed to be partly included in the Ig to Igex
ratio.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3170 K. Láska et al.
The effective total ozone content (TOCeff) is calculated from the TOC values with
respect to the path of sunlight through the ozonosphere to the Earth’s surface (Láska
et al. 2010). In this approach, the radius of the Earth, mean height of the ozonosphere,
the daily mean solar elevation angles and TOC are used to resolve the geometrical
problem of the actual path of sunlight through the spherical layer of the atmosphere.
In this paper, TOC data were obtained by the Ozone Monitoring Instrument (OMI)
on board the EOS-Aura spacecraft for the geographical coordinates of the Mendel
Station (http://avdc.gsfc.nasa.gov/). As reported by many authors, the Total Ozone
Mapping Spectrometer (TOMS) Version 8.5 algorithm applied to OMI measurements
should have a root-mean-squared error of 1–2%, depending on the solar zenith angle,
aerosol amount and cloud cover (Ahmad et al. 2004, Ialongo et al. 2008, Jaross and
Warner 2008, Tzanis and Varotsos 2008). To evaluate the quality of satellite-based
TOC with respect to the prediction errors of EUV radiation model, the effective surface
reflectivity at 360 nm (hereafter, reflectivity) and the distance between the Mendel
Station and the centre of the ground pixel (OMI field of view) were additionally
acquired. Figures 3 and 4 show the annual course of the daily mean TOC and reflectivity
for the Mendel Station between 2007 and 2009. In contrast to the central part of
the Antarctic plateau, there is high seasonal variation of TOC, reflectivity (due to seaice
concentration) and cloudiness in the coastal regions of the continent. Therefore,
the maximum values of the EUV radiation are not related to the ozone minima, but
instead occur at the end of November or mid-December, close to the summer solstice
in the Southern Hemisphere (Láska et al. 2010).
The previous version of our model was validated for the Ukrainian Vernadsky
Station in Antarctica (ϕ = 65◦
15 S; λ = 64◦
15 W; 11 m above sea level) from
2002 to 2005 (Láska et al. 2010). Apart from that, the sensitivity of EUV radiation to
uncertainties of TOC data from both ground- and satellite-based measurements was
studied (Láska et al. 2010). To examine whether the sensitivity of input parameters
affects recent model results, the relative prediction errors of EUV radiation based on
Ig radiation and TOC values were calculated (figure 5). In the first case, we supposed
1 Mar 07 1 Jul 07 1 Nov 07 1 Mar 08 1 Jul 08 1 Nov 08 1 Mar 09
1 Mar 07 1 Jul 07 1 Nov 07 1 Mar 08 1 Jul 08 1 Nov 08 1 Mar 09
Figure 4. Annual course of daily mean effective surface reflectivity at 360 nm (Reflectivity)
and residuals of the erythemally effective UV radiation (Residuals EUV) at the Mendel Station
in the period 2007–2009.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3171
the mean error of Ig measurement to be around ±5%, while the ozone measurement
error was hypothetically zero. It can be seen from figure 5(a) that the largest relative
prediction error (RPE) was up to 5% at the lowest intensities of Ig radiation.
Contrary to that, the RPE descended below 2% when Ig radiation increased. This
error was almost symmetrical, which was also reported by deviations of symmetry
<0.1%. In the second case, ozone measurement errors ±5 and ±20 DU were tested,
while the Ig measurement error was hypothetically zero. As shown in figure 5(b), the
relative prediction errors of EUV radiation were more asymmetrical at ±20 DU than
at ±5 DU. The largest RPE (up to 25%) occurred at the lowest TOC values and vice
versa. Based on these results, it can be concluded that uncertainties in TOC measurement
significantly affect the estimation of EUV radiation. Moreover, timing of the
ozone depletion over Antarctica coincides with the very low solar elevation angle in
the spring season (August–October). Hence, the aforementioned factors during ozone
decline may cause inaccuracies of EUV predictions of about 7% on average and as
high as 27% in extremes. Out of the ozone depletion period, the inaccuracy of EUV
predictions may reach 3.5% and 10% respectively.
3. Results and discussion
The effects of the atmospheric factor on the predicted daily sums of EUV radiation
were evaluated by means of a non-linear regression model with a hyperbolic
transmissivity function. Model parameterization and additional computation were
performed for 539 days from the original data set (2007–2009), in which both TOC and
Ig radiation were available. The β0 and β1 parameters of the model described were estimated
using the STATISTICA (2009) software package and Levenberg–Marquardt’s
method through the least squares method with the initial approximations of β0 = 0.5
and β1 = 0.2. Consequently, the relative estimate error of the regression parameters
was calculated as a half width of 95% confidence interval to estimate value, expressed
as a percentage. The estimates of the parameters and relative estimated errors were
0.57 for β0 (6.6%) and 0.19 (2.2%) for β1 respectively. The relative estimated errors
of the new regression parameters were slightly better than errors estimated in the
previous version of the model (Láska et al. 2010).
The model results for the period 2007–2009 are shown in figure 6(a) as the scatterplot
of measured versus predicted daily sums of EUV radiation intensities. In the
Figure 5. Relative prediction errors of EUV radiation estimated by means of nonlinear regression
model. Theoretical distribution of (a) Ig radiation and (b) total ozone content (TOC) at 5%
and 20% levels of uncertainty were considered.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3172 K. Láska et al.
Figure 6. (a) Scatterplot of daily sums of predicted erythemally effective UV radiation
(Predicted EUV) and erythemally effective UV radiation (EUV) measured at the Mendel
Station between 2007 and 2009. (b) Scatterplot of residuals of the erythemally effective UV
radiation (Residuals EUV) and effective surface reflectivity at 360 nm (Reflectivity).
studied period, the variability of EUV radiation was explained by means of the model
at 98.61%. However, the coefficient of determination decreased slightly in comparison
to the previous study (Láska et al. 2010). This was probably caused by uncertainties
of satellite-based TOC measurement applied as the input parameter. The mean average
prediction error (MAPE) and residuals of the EUV radiation intensity were used
to evaluate the quality of the model. The MAPE value for the recent version of our
model was 6.0% over the studied period. This can be also compared with the previous
results of Láska et al. (2009), when both ground and satellite ozone measurements
were applied. In the latter case, MAPE values ranged from 9.6% to 7.4% (ground
ozone) and from 9.5% to 8.8% (satellite ozone) depending on the season. Since only a
few stations in high-latitude locations performed both estimations and measurements
of EUV radiation using the same instruments and methods, the results obtained were
difficult to compare with the other relevant data. In spite of this, our results concur
with the studies of Weihs and Webb (1997), Koepke et al. (1998) and De Backer et al.
(2001). Moreover, we intend to compare results from our model with similar statistical
models with limited input parameters used for UV prediction in the future. A forthcoming
paper will be devoted to the intercomparison of models and their sensitivity
analysis for various input data.
Figure 4 shows the annual course of the satellite-based reflectivity and EUV residuals
between 2007 and 2009. Seasonal changes in sea-ice concentration, sea-ice age and
cloudiness in the vicinity of the Mendel Station had some consequences in the temporal
variability of reflectivity and EUV residuals. The largest reflectivity and its small
variation were closely linked with the maximum sea ice present along the Antarctic
Peninsula from August to November each year. Summer shrinkage of the sea ice consequently
caused the high variability of reflectivity from December to April. Therefore,
the performance of the model was strongly influenced by the daily mean reflectivity
(figure 6(b)). It was only when the reflectivity values were <45% (11.9% of all cases
– 60 days) that the EUV residuals ranged between –0.12 and +0.07 kJ m−2
. From
figure 4, it is apparent that the smallest EUV residuals (up to 0.1 kJ m−2
) can be
found only at low solar elevation (August–September and March–April). The highest
EUV residuals occurred frequently from October to February independently of the
ozone depletion. The largest values of EUV residuals reached –0.21 and +0.28 kJ m−2
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3173
Figure 7. Relative frequency distribution of the residuals of erythemally effective UV radiation
(Residuals EUV) at the Mendel Station. The results are shown for three consecutive seasons:
August–October, November–January and February–April between 2007 and 2009.
respectively. On the other hand, the largest EUV residuals (more than ±0.15 kJ m−2
)
represented only 4.8% of all cases (26 days).
The last part of our study was to evaluate how the variation in TOC affected
the EUV residuals in different seasons. Therefore, EUV residuals were analysed
separately according to the solar elevation angle (hS) and the time interval of the
ozone depletion. On the basis of the above-mentioned criteria, three periods were
distinguished: August–October (low hS, the largest ozone depletion), November–
January (high hS, ozone depletion retreat) and February–April (low hS, no ozone
depletion). From figure 7, it is apparent that the relative frequency distributions
of the EUV residuals in 2007–2009 were mainly dependent on the season and
hS respectively. Then, the lower the hS, the smaller the residuals of EUV radiation
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3174 K. Láska et al.
Table 1. Seasonal variability of total ozone content (TOC), global solar radiation (Ig) and effective
surface reflectivity at 360 nm (Reflectivity) expressed by the variation coefficient at the
Mendel Station between 2007 and 2009. Model quality is presented by the standard deviation
of erythemally effective UV residuals (EUV) and mean average prediction error (MAPE).
Coefficient of Variation (%)
Period TOC Ig Reflectivity
Standard
deviation EUV
residuals
(kJ m−2
)
MAPE
(%)
August 2007–October 2007 25.87 67.75 10.67 0.058 7.3
November 2007–January 2008 13.33 30.33 20.21 0.096 4.4
February 2008–April 2008 7.95 70.26 25.94 0.064 7.8
August 2008–October 2008 21.99 64.41 18.46 0.070 5.1
November 2008–January 2009 10.65 32.47 30.83 0.077 3.6
and model underestimation/overestimation. Therefore, high kurtosis and small dispersion
occurred in both spring and autumn months at low hS (August or April).
On the contrary, normal distribution or the lowest kurtosis of the EUV residuals
were found within the summer months at high hS. The highest EUV residuals
(–0.3 to +0.3 kJ m−2
) occurred in November or February. These results concur
with a previous study (Láska et al. 2010) in which a similar frequency analysis was
performed.
To supplement this validation, seasonal variability of TOC, Ig radiation and
reflectivity were calculated as the variation coefficients for the individual seasons of
2007–2009 (table 1). The model quality expressed by the standard deviation of EUV
residuals and MAPE was then examined. As seen in table 1, TOC variation rapidly
decreased from spring to autumn (as a consequence of the disappearing ozone depletion).
On the other hand, the highest variation of Ig radiation affected by cloudiness
was found from February to April 2008. The significant dependence of the EUV
residuals on both TOC and Ig radiation was documented by MAPE. The seasonal
variability of satellite reflectivity had only a marginal effect on MAPE and the estimation
of EUV radiation. Contrary to standard deviations, the smallest values of
MAPE were reported during summer when variation coefficients of TOC and Ig were
at their lowest. Therefore, MAPE reached only 4.4% between November 2007 and
January 2008 and 3.6% between November 2008 and January 2009. This fact supported
the aforementioned analysis of the sensitivity of input parameters (§2), where
the uncertainties in measured TOC and Ig values significantly affect the estimation of
EUV radiation. Therefore, the largest model-prediction error may occur during ozone
decline and reach about 7% on average.
4. Conclusions
In this study, EUV radiation was estimated using a new version of a non-linear regression
model with a hyperbolic transmissivity function applied. The parameterization
of the model was performed on the daily mean values of the selected input parameters
such as solar elevation angle, global solar radiation and total ozone content measured
at the Mendel Station from 2007 to 2009. The model results concur with previous
studies (Láska et al. 2009, 2010). The validation results imply improvements in the
accuracy of the model prediction thanks to the improved algorithm with transmission
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3175
function of both Ig and EUV radiation. The model results were mainly affected by
the temporal variation of Ig radiation and the uncertainties of satellite-based TOC
measurements.
Acknowledgements
We are grateful to the science team of the Aura Validation Data Center for providing
OMI–TOMS Version 8.5 overpass data for the Mendel Station location. We also
thank Miloš Barták from the Masaryk University, Brno, Czech Republic, for valuable
discussions and his comments on this work. This study was performed within the
research programme ‘The effects of atmospheric factors on the regime of UV radiation
of Antarctic Peninsula’ (Grant No. 205/07/1377) funded by the Grant Agency of
the Czech Republic.
References
AHMAD, Z., BHARTIA, P.K. and KROTKOV, N., 2004, Spectral properties of backscattered UV
radiation in cloudy atmospheres. Journal of Geophysical Research, 109, p. D01201.
BALIS, D., BOJKOV, R., TOURPALI, K. and ZEREFOS, C., 2009, Characteristics of the
ozone decline over both hemispheres. International Journal of Remote Sensing, 30,
pp. 3887–3895.
BERGER, D., 1976, The sunburning ultraviolet meter: design and performance. Photochemistry
and Photobiology, 24, pp. 587–593.
CHANDRA, S. and VAROTSOS, C.A., 1995, Recent trends of the total column ozone: implications
for the Mediterranean region. International Journal of Remote Sensing, 16,
pp. 1765–1769.
CRACKNELL, A.P. and VAROTSOS, C.A., 2007, The IPCC Fourth Assessment Report and
the fiftieth anniversary of Sputnik. Environmental Science and Pollution Research, 14,
pp. 384–387.
DE BACKER, H., KOEPKE, P., BAIS, A., DE CABO, X., FREI, T., GILLOTAY, D., HAITE, CH.,
HEIKKILÄ, A., KAZANTZIDIS, A., KOSKELA, T., KYRÖ, E., LAPETA, B., LORENTE, J.,
MASSON, K., MAYER, B., PLETS, H., REDONDAS, A., RENAUD, A., SCHAUBERGER,
G., SCHMALWIESSER, A., SCHWANDER, H. and VANICEK, K., 2001, Comparison of
measured and modelled UV indices for the assessment of health risks. Meteorological
Applications, 8, pp. 267–277.
DE FABO, E.C., NOONAN, F.P. and FREDERICK, J.E., 1990, Biologically effective doses of sunlight
for immune supression at various latitudes and their relationship to changes in
stratospheric ozone. Photochemistry and Photobiology, 52, pp. 811–817.
EFSTATHIOU, M.N., VAROTSOS, C.A. and KONDRATYEV, K.Y., 1998, An estimation of the surface
ultraviolet irradiance during an extreme total ozone minimum. Meteorology and
Atmospheric Physics, 68, pp. 171–176.
EFSTATHIOU, M.N., VAROTSOS, C.A., SINGH, R.P., CRACNELL, A.P. and TZANIS, C., 2003,
On the longitude dependence of total ozone trends over middle-latitudes. International
Journal of Remote Sensing, 24, pp. 1361–1367.
FREDERICK, J.E., QU, Z. and BOOTH, C.R., 1998, Ultraviolet radiation at sites on the Antarctic
Coast. Photochemistry and Photobiology, 68, pp. 183–190.
GERNANDT, H., GOERSDORF, U., CLAUDE, H. and VAROTSOS, C.A., 1995, Possible impact of
polar stratospheric processes on mid-latitude vertical ozone distributions. International
Journal of Remote Sensing, 16, pp. 1839–1850.
HOFMANN, D.J., JOHNSON, B.J. and OLTMANS, S.J., 2009, Twenty-two years of ozonesonde
measurements at the South Pole. International Journal of Remote Sensing, 30,
pp. 3995–4008.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
3176 K. Láska et al.
IALONGO, I., CASALE, G.R. and SIANI, A.M., 2008, Comparison of total ozone and erythemal
UV data from OMI with ground-based measurements at Rome station. Atmospheric
Chemistry and Physics Discussions, 8, pp. 2381–2401.
JAROSS, G. and WARNER, J., 2008, Use of Antarctica for validating reflected solar radiation
measured by satellite sensors. Journal of Geophysical Research, 113, p. D16S34.
KOEPKE, P., BAIS, A., BALIS, D., BUCHWITZ, M., DE BACKER, H., DE CABO, X., ECKERT,
P., ERIKSEN, P., GILLOTAY, D., KOSKELA, T., LAPETA, B., LITYNSKA, Z., LORENTE,
J., MAYER, B., RENAUD, A., RUGGABER, A., SCHAUBERGER, G., SECKMEYER, G.,
SEIFERT, P., SCHMALWIESER, A., SCHWANDER, H., VANICEK, K. and WEBER, M.,
1998, Comparison of models used for UV index-calculations. Photochemistry and
Photobiology, 67, pp. 657–662.
KONDRATYEV, K.Y., POKROVSKY, O.M. and VAROTSOS, C.A., 1995, Atmospheric ozone
trends and other factors of surface ultraviolet radiation variability. Environmental
Conservation, 22, pp. 259–261.
KONDRATYEV, K.Y. and VAROTSOS, C.A., 1995, Atmospheric ozone variability in the context
of global change. International Journal of Remote Sensing, 16, pp. 1851–1881.
KONDRATYEV, K.Y. and VAROTSOS, C.A., 1996, Global total ozone dynamics – Impact
on surface solar ultraviolet radiation variability and ecosystems. 1. Global ozone
dynamics and environmental safety. Environmental Science and Pollution Research, 3,
pp. 153–157.
KONDRATYEV, K.Y., VAROTSOS, C.A. and CRACKNELL, A.P., 1994, Total ozone amount trend at
St-Petersburg as deduced from NIMBUS-7 TOMS observations. International Journal
of Remote Sensing, 15, pp. 2669–2677.
KRAVCHENKO, V., EVTUSHEVSKY, A., GRYTSAI, A., MILINEVSKY, G. and SHANKLIN, J.,
2009, Total ozone dependence of the difference between the empirically corrected
EP-TOMS and high-latitude station datasets. International Journal of Remote Sensing,
30, pp. 4283–4294.
LÁSKA, K., PROŠEK, P., BUDÍK, L., BUDÍKOVÁ, M. and MILINEVSKY, G., 2009, Prediction
of erythemally effective UV radiation by means of nonlinear regression model.
Environmetrics, 20, pp. 633–646.
LÁSKA, K., PROŠEK, P., BUDÍK, L., BUDÍKOVÁ, M. and MILINEVSKY, G., 2010, Estimation
of solar UV radiation in maritime Antarctica using nonlinear model including cloud
effects. International Journal of Remote Sensing, 31, pp. 831–849.
LUCCINI, E., CEDE, A. and PIACENTINI, R.D., 2003, Effect of clouds on UV and total irradiance
at Paradise Bay, Antarctic Peninsula, from a summer 2000 campaign. Theoretical and
Applied Climatology, 75, pp. 105–116.
MCKINLAY, A. and DIFFEY, B.L., 1987, A reference action spectrum for ultraviolet induced
erythema in human skin. In Human Exposure to Ultraviolet Radiation: Risks and
Regulations, W.R. Passchler and B.F.M. Bosnajokovic (Eds.), pp. 83–87 (Amsterdam:
Elsevier).
PROŠEK, P. and JANOUCH, M., 1996, The measurement of ultraviolet radiation at the Polish
Henryk Arctowski Station (South Shetlands, Antarctica) in the Summer of 1994/95.
Problemy Klimatologii Polarnej, 6, pp. 139–148.
PROŠEK, P., LÁSKA, K. and JANOUCH, M., 2001, Ultraviolet radiation intensity at the
H. Arctowski Base (South Shetlands, King George Island) during the period from
December 1994–December 1996, Brno, Czech Republic. In Ecology of the Antarctic
Coastal Oasis, P. Prošek and K. Láska (Eds.), pp. 35–48 (Brno, Czech Republic:
Folia Facultatis Scientarum Naturalium Universitatis Masarykianae Brunensis,
Geographia 25).
SCHMALWIESER, A.W., SCHAUBERGER, G., JANOUCH, M., NUNEZ, M., KOSKELA, T., BERGER,
D., KARAMANIAN, G., PROSEK, P. and LÁSKA, K., 2002, Global validation of a forecast
model for irradiance of solar, erythemally effective ultraviolet radiation. Optical
Enginering, 41, pp. 3040–3050.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
Atmospheric studies by optical methods 3177
STATISTICA, 2009, STATISTICA software. Available online at: www.statsoft.com/products/
analytics.htm and www.statsoft.com/textbook/stnonlin.html (accessed 19 November
2009).
TZANIS, C. and VAROTSOS, C.A., 2008, Tropospheric aerosol forcing of climate: a case study for
the greater area of Greece. International Journal of Remote Sensing, 29, pp. 2507–2517.
VAROTSOS, C., 1994, Solar ultraviolet radiation and total ozone, as derived from
satellite and ground-based instrumentation. Geophysical Research Letters, 21,
pp. 1787–1790.
VAROTSOS, C., 2002, The southern hemisphere ozone hole split in 2002. Environmental Science
and Pollution Research, 9, pp. 375–376.
VAROTSOS, C., 2003, What is the lesson from the unprecedented event over Antarctica in 2002?
Environmental Science and Pollution Research, 10, pp. 80–81.
VAROTSOS, C., 2004, The extraordinary events of the major, sudden stratospheric warming,
the diminutive Antarctic ozone hole, and its split in 2002. Environmental Science and
Pollution Research, 11, pp. 405–411.
VAROTSOS, C., ALEXANDRIS, D., CHRONOPOULOS, G. and TZANIS, C., 2001, Aircraft observations
of the solar ultraviolet irradiance throughout the troposphere. Journal of
Geophysical Research-Atmospheres, 106, pp. 14843–14854.
VAROTSOS, C., CARTALIS, C., VLAMAKIS, A., TZANIS, C. and KERAMITSOGLOU, I., 2004,
The long-term coupling between column ozone and tropopause properties. Journal of
Climate, 17, pp. 3843–3854.
VAROTSOS, C.A. and CRACKNELL, A.P., 1993, Ozone depletion over Greece as deduced
from Nimbus-7 TOMS measurements. International Journal of Remote Sensing, 14,
pp. 2053–2059.
VAROTSOS, C.A. and CRACKNELL, A.P., 1994, Three years of total ozone measurements over
Athens obtained using the remote sensing technique of a Dobson spectrophotometer.
International Journal of Remote Sensing, 15, pp. 1519–1524.
VAROTSOS, C.A., KONDRATYEV, K.Y. and CRACKNELL, A.P., 2000, New evidence for
ozone depletion over Athens, Greece. International Journal of Remote Sensing, 21,
pp. 2951–2955.
VAROTSOS, C., KONDRATYEV, K.Y. and KATSIKIS, S., 1995, On the relationship between total
ozone and solar ultraviolet radiation at St Petersburg, Russia. Geophysical Research
Letters, 22, pp. 3481–3484.
WEIHS, P., BLUMTHALER, M., RIEDER, H.E., KREUTER, A., SIMIC, S., LAUBE, W.,
SCHMALWIESER, A.W., WAGNER, J.E. and TANSKANEN, A., 2008, Measurements of UV
irradiance within the area of one satellite pixel. Atmospheric Chemistry and Physics, 8,
pp. 5615–5626.
WEIHS, P. and WEBB, A.R., 1997, Accuracy of spectral UV model calculations: 2. Comparison
of UV calculations with measurements. Journal of Geophysical Research, 102,
pp. 1551–1560.
ZIEMKE, J.R., CHANDRA, S., HERMAN, J. and VAROTSOS, C., 2000, Erythemally weighted UV
trends over northern latitudes derived from Nimbus-7 TOMS measurements. Journal of
Geophysical Research-Atmospheres, 105, pp. 7373–7382.
Downloadedby[MasarykovaUniverzitavBrne],[K.Láska]at00:5619July2011
5
Paper 4
Variability in solar irradiance observed at two contrasting
Antarctic sites
Petkov, B.H., Láska, K., Vitale, V., Lanconelli, C., Lupi, A., Mazzola, M.,
Budíková, M., 2016,
International Journal of Remote Sensing, 31, pp. 831–849
Variability in solar irradiance observed at two contrasting Antarctic sites
Boyan H. Petkov a,
⁎, Kamil Láska b
, Vito Vitale a
, Christian Lanconelli a
, Angelo Lupi a
,
Mauro Mazzola a
, Marie Budíková c
a
Institute of Atmospheric Sciences and Climate (ISAC) of the Italian National Research Council (CNR), Bologna, Italy
b
Department of Geography, Faculty of Science, Masaryk University, Brno Czech Republic
c
Department of Mathematics and Statistics, Faculty of Science, Masaryk University, Brno, Czech Republic
a b s t r a c ta r t i c l e i n f o
Article history:
Received 20 May 2015
Received in revised form 3 December 2015
Accepted 8 January 2016
Available online 15 January 2016
The features of erythemally weighted (EW) and short-wave downwelling (SWD) solar irradiances, observed during
the spring–summer months of 2007–2011 at Johann Gregor Mendel (63°48′S, 57°53′W, 7 m a.s.l.) and Dome
Concordia (75°06′S, 123°21′E, 3233 m a.s.l.) stations, placed at the Antarctic coastal region and on the interior
plateau respectively, have been analysed and compared to each other. The EW and SWD spectral components
have been presented by the corresponding daily integrated values and were examined taking into account the
different geographic positions and different environmental conditions at both sites. The results indicate that at
Mendel station the surface solar irradiance is strongly affected by the changes in the cloud cover, aerosols and albedo
that cause a decrease in EW between 20% and 35%, and from 0% to 50% in SWD component, which contributions
are slightly lower than the seasonal SWD variations evaluated to be about 71%. On the contrary, the
changes in the cloud cover features at Concordia station produce only a 5% reduction of the solar irradiance,
whilst the seasonal oscillations of 94% turn out to be the predominant mode. The present analysis leads to the
conclusion that the variations in the ozone column cause an average decrease of about 46% in EW irradiance
with respect to the value found in the case of minimum ozone content at each of the stations. In addition, the
ratio between EW and SWD spectral components can be used to achieve a realistic assessment of the radiation
amplification factor that quantifies the relationship between the atmospheric ozone and the surface UV
irradiance.
© 2016 Elsevier B.V. All rights reserved.
Keywords:
Solar irradiance in Antarctica
Environmental effect on solar radiation
Ozone column
Radiation amplification factor
1. Introduction
The solar radiation provides the Earth with energy that generates
variety of dynamical processes and triggers numerous chemical reactions
in the atmosphere, biosphere and hydrosphere. For that reason,
the study of variability in solar radiation is considered an important
issue in natural sciences examining the environmental problems.
Antarctica is an interesting subject for such investigations because of
the significant contrast in the atmospheric and environmental features
characterising the costal and continental zones, on the one hand and the
specific atmospheric processes such as polar vortex creating favourable
conditions of intensive photochemical reactions in the stratosphere, on
the other (Farman et al., 1985; Molina and Rowland, 1974). Another important
feature determining Antarctica as a specific geographical region
is the lower, with respect to the mid-latitude areas, extent of the human
activity that more clearly outlines the effects of the natural factors on
the solar irradiance. These circumstances enhanced the interest of the
scientific community in solar radiation over Antarctica and led to
intensive studies in the past years (Lubin and Frederick, 1991;
Stamnes et al., 1992; Orsini et al., 2002; Bernhard et al., 2004; Láska
et al., 2009; Láska et al., 2010; Vitale et al., 2011; Lee et al., 2015).
The analyses of the solar irradiance observed at the Earth's surface
highlighted several important factors that determine its level and fluctuations
(McKenzie et al., 2007; Petkov et al., 2012). The Sun elevation that
closely depends on the geographical position, is considered the main factor
affecting the solar radiation at the ground and it is usually expressed
either by the angle α determining the position of the Sun over the horizon
or by the solar zenith angle Z defined relatively the zenith direction,
so that Z=90∘
−α (Iqbal, 1983). Beside of this astronomical parameter,
the environmental characteristics, such as the atmospheric transmittance,
mainly determined by the cloud cover features and aerosol loadings,
together with the surface reflectivity (albedo) significantly impact
the solar irradiance at the ground (Lee et al., 2015). In addition, the altitude
above mean sea level of the irradiated surface is an important factor
for the magnitude of solar radiation reaching the observational point. On
the other hand, the total ozone content in the atmosphere strongly affects
the fraction of the solar ultraviolet B (UV-B) irradiance that reaches
the ground (~0.295–0.315 μm) and acts upon photo-chemical reactions
in the biosphere. It is worth noting that the atmosphere over Antarctic
Plateau is slightly contaminated by aerosols (Tomasi et al., 2007) and,
Atmospheric Research 172–173 (2016) 126–135
⁎ Corresponding author.
E-mail address: b.petkov@isac.cnr.it (B.H. Petkov).
http://dx.doi.org/10.1016/j.atmosres.2016.01.005
0169-8095/© 2016 Elsevier B.V. All rights reserved.
Contents lists available at ScienceDirect
Atmospheric Research
journal homepage: www.elsevier.com/locate/atmosres
as a result their impact on the solar irradiance variations at deep continental
zone was assumed to be of minor importance.
Since the environmental factors usually change significantly with
the season, latitude and altitude, the distinct difference between the
transmittance characteristics of the atmosphere over the coastal regions
(Antarctic Peninsula) and interior Antarctic Plateau affects the spatiotemporal
variability of solar irradiance on the continent. This study analyses
the features of solar radiation observed at two contrasting Antarctic
sites placed at the coastal zone and plateau, respectively and aims to examine
the role of the factors, impacting radiation variability during the
daylight part of the year. As a result, approximate quantitative estimations
outlining the contribution of these factors to the solar irradiance
variations that take place at the different Antarctic regions are achieved
and discussed below.
2. Observational sites and instrumentation
The present study examines the solar radiation, simultaneously
measured at the Johann Gregor Mendel and Dome Concordia Antarctic
stations (hereinafter referred to as Mendel and Concordia) during
260 days of the austral spring–summer months from 2007 to 2011.
Mendel station is located in the northern part of James Ross Island,
close to the eastern coast of the Antarctic Peninsula, whilst Concordia
is situated on the plateau in the opposite part of the continent at a distance
of about 4600 km far from Mendel (see Fig. 1(a) and Table 1).
The two stations are characterised by quite different surroundings and
meteorological conditions that assumes quite different features of the
solar irradiance reaching these sites. In respect to limited in-situ measurements,
space-born observations performed by passive remote sensing
instruments were used to give evidence about contrasting optical
properties of the corresponding environments. Table 1 exhibits a basic
statistics of the surface reflectivity (albedo) variations during the year,
provided by the version 3 of OMAERUVG daily global data products
(level 2G) retrieved from the Ozone Monitoring Instrument (OMI
2014). Table 1 presents also a similar statistics of the cloud cover fraction
(CCF) that was taken to characterise the atmospheric transmittance and
was derived from the Atmospheric Infrared Sounder (AIRS) L3 cloud
property retrievals (Platnick et al., 2003), which give the daily average
cloud fraction with a spatial resolution of 1° (Susskind et al., 2006). The
values of the albedo and CCF were extracted for an area of ±1° in both
latitude and longitude around each of the sites and the averages of their
minimum, mean and maximum values are presented in Table 1 together
with the corresponding standard deviations. As can be seen, in the coastal
Antarctic zone, where the Mendel station is situated, the albedo is a subject
of large seasonal variations (Hrbáček et al., 2016), likely due to the
changes in the snow and sea–ice occurrences, which cause variability in
the surface reflectivity ranging between 12% in summer and 96% in the
late winter. In contrast, the permanent snow-covered surface of the plateau
(Concordia station) is characterised by a high albedo slightly varying
from 88% to 100% during the year that creates favourable conditions for
multiple reflections between the surface and the scattering atmosphere
(Wuttke et al., 2006; Cordero et al., 2014). The prevailing overcast conditions
at Mendel station, where the mean CCF reaches a value of about
63%, are able to reduce significantly the surface solar irradiance
(Bernhard et al., 2005; Láska et al., 2011), whilst the cloudless atmosphere
and low aerosol content over Concordia contribute to a higher
solar radiation level (Vitale et al., 2011).
An UV Biometer 501 A version 3 manufactured by Solar Light
(Glenside, PA, USA) provided the erythemally weighted (EW) irradiance
for the present analysis at Mendel station, whilst the short-wave
downwelling (SWD) irradiance was measured using a CM11
pyranometer, produced by Kipp & Zonen, the Netherlands (Láska et al.,
2011). At Concordia, the corresponding radiations were observed by narrowband
filter radiometer UV-RAD (EW) and Kipp & Zonen CM22 (SWD)
pyranometer (Petkov et al., 2006; Vitale et al., 2011), operating as a part of
Baseline Surface Radiation Network (BSRN) (Ohmura et al., 1998). UVRAD
is equipped with 7 spectral channels centred at peak wavelengths
ranging between 0.300 μm and 0.364 μm, each channel being defined
by a set of pass-band filters determining an overall bandwidth of about
1 nm and transmittance varying between 12% and 27%. The irradiances,
measured by the UV-RAD channels allow the extraction of the ozone column
and the solar UV spectrum, which is used to calculate the EW irradiance
by means of Eq. (1). The spectral response of UV Biometer 501 A is
very close to the erythemal action spectrum, whilst CM11 and CM22 provide
the total irradiance within 0.305–2.800 μm and 0.290–3.600 μm
spectral ranges, respectively. In addition, the ozone amounts given by
the OMI satellite instrument on the one hand (Levelt et al., 2006) and extracted
from UV-RAD measurements on the other, were also used in the
present analysis.
3. Data sets
For the purposes of the present study, the EW and SWD irradiances,
both observed at the two stations and computed by a radiative transfer
model have been examined.
3.1. Measurements
The EW(t) spectral component, measured at time t, represents the
solar irradiance I(λ,t) at wavelength λ, weighted by the erythemal action
spectrum E(λ) (McKinlay and Diffey, 1987):
EW tð Þ ¼
Z0:400
0:290
I λ; tð Þ E λð Þ dλ : ð1Þ
Fig. 1. (a) Location of the Mendel and Concordia stations on the Antarctic continent.
(b) Time-patterns of the variable cos(90∘
−αm), where αm is the daily maximum of the
solar elevation α at Mendel and Concordia, respectively. It is considered that this
variable determines the level and the seasonal trend of the solar irradiance at the Earth's
surface.
127B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
The SWD(t) irradiance measured by the CM11 and CM22
pyranometers covers the UV, visible and near infrared (IR) bands:
SWD tð Þ ¼
Z4:000
0:290
I λ; tð Þ S λð Þ dλ ; ð2Þ
where S(λ) is the relative spectral response of the corresponding instrument.
Taking into account that UV irradiance is less than 5% of the solar
radiation reaching the ground, SWD(t) is thought to represent the
behaviour of the visible and near infrared spectral ranges.
In order to remove the short-term diurnal oscillations in the measured
irradiance that allows a better representation of the variation features
observed during the comparatively long period considered here,
the study focussed on the daily irradiance doses DEW(G) and DSWD(G)
evaluated as:
DWW Gð Þ ¼
Z
ΔTG
WW tð Þ dt ; ð3Þ
where WW denotes either EW or SWD and ΔTG is the period from
00:00 h to 24:00 h local time of the day G of the year.
The action spectrum E(λ) in Eq. (1) determines a maximum of EW(t)
at about 0.305 μm and gives a low weight of the irradiance at
wavelengths longer than 0.325 μm. However, when the Sun is close to
the horizon, the surface irradiance at λ N 0.325 μm enhances significantly
with respect to the shorter wavelengths and its contribution to EW
correspondingly increases. As can be concluded from Fig 1(b), which
is discussed in Section 4.1, the maximum daily solar elevation αm is
higher than 17° during the days of the period considered. Model estimations
show that about 70% of the EW irradiance falls within 0.295–
0.330 μm spectral range for α≈17o
, whilst for α≈25o
this percentage
corresponds to 0.295–0.323 μm interval. Thus, for the period analysed
here, it can be concluded that the irradiance at wavelength no longer
than 0.330 μm mainly contributes the dose DEW(G), which is the spectral
band representing the ozone depending UV radiation.
3.2. Model estimations
Since the field measurements give the variations in the surface solar
irradiance caused by all the impacting factors it was considered that a
realistic representation of the seasonal component in the observed variations
can be determined through model evaluations performed for
clear sky conditions. MODTRAN (Berk et al., 2011) is one of the classical
computer codes providing precise estimates of the solar radiation in the
UV, visible and IR spectral ranges. However, the computations performed
by MODTRAN take a long time to obtain the doses given by
Eqs. (1)–(3). On the other hand, the free available Tropospheric
Ultraviolet-Visible (TUV, http://gcmd.nasa.gov/records/UCAR_TUV.
html) radiative transfer model (Madronich and Flocke, 1997) allows a
simple configuration of the calculations and achievement of satisfactory
results within comparatively short computation time. The TUV model
works in the UV and visible bands, so that, for the purposes of the present
analysis, it has been extended to the IR diapason by using the
MODTRAN code as discussed below. Similarly, Perrina et al. (2005)
made an extension of the TUV model to the IR diapason by applying
another approach, which gave results, similar to those presented here.
The calculations of both EW and SWD components were performed
through the TUV model by inserting the following inputs:
(i) the atmosphere temperature and density profiles constructed
by Tomasi et al. (2015) for Concordia and Neumayer Antarctic station
(70°39′S, 08°15′W, 40 m a.m.s.l.); the second station was considered
to represent the atmospheric conditions of the coastal zone;
(ii) null clouds; (iii) aerosol atmospheric thicknesses equal to 0.06
and 0.019 at Mendel and Concordia respectively, with the corresponding
values of Ångström exponent of 0.64 and 1.61 (Tomasi
et al., 2007); and (iv) surface albedo equal to 0.6 at Mendel and
0.9 at Concordia (see Table 1). The assumed input values of the
ozone column are discussed latter, in this section.
TUV model estimates the surface solar irradiance for the wavelengths
up to 0.700 μm, taking into account the scattering and absorption processes
in the atmosphere. However, only the first kind of these processes represents
the irradiance attenuation in the IR spectral range, excluding the
absorption bands of O2, H2O and CO2. Fig. 2(a) illustrates the IR spectra
of the surface solar irradiance calculated by TUV and MODTRAN models
for Mendel station on 15 November 2008. As can be seen, TUV gives a
smooth curve, whilst MODTRAN depicts a different spectral pattern. To
take into account the absorption bands in the TUV computations, the atmospheric
transmittance within 0.700 b λ b4.000 μm spectral range
was calculated through MODTRAN for clear sky conditions, no aerosol
loadings and different solar elevations. The calculations were performed
assuming values of perceptible water equal to 2 mm at Mendel and
0.2 mm at Concordia, which values correspond to the minimum water
amounts at costal and plateau areas (Tomasi et al., 2011; Sarti et al.,
2013), typical for clear sky conditions. The estimated transmittances
were used after that to correct the IR part of the TUV spectrum and calculate
SWD(t) irradiance and doses DSWDðGÞ through Eqs. (2) and (3), respectively.
Fig. 2(b) and (c) demonstrate that the results are consistent
with the irradiances, measured during days with predominantly clear
sky conditions.
The time patterns of DSWDðGÞ, shown in the lower panel of Fig. 3 represent
upper envelopes of the measurements and can be considered reference
curves for extracting the effects of environmental factors. Since
the EW irradiance is an ozone depredating component, the minimum
ozone column values of 135 Dobson Units (DU) at Mendel and 150
DU at Concordia were taken as inputs for the model calculations. Such
an assumption gives reference seasonal EW trends (see upper panel of
Fig. 3), analogously to the SWD case, which highlight the irradiance decrease
caused by the environmental factors and the ozone column when
the latter varies from its minimum to its maximum.
4. Results and discussion
The solar radiation observed at Mendel and Concordia stations within
the chosen period have been examined bearing in mind the main factors
affecting the surface irradiance, as discussed in Section 1. The analysis
was performed making an attempt to separate the effects of these factors
Table 1
Geographic location of the two selected stations together with their environmental characteristics presented by the surface reflectivity (albedo) and cloud cover fraction (CCF) determined
from space-born measurements over an area of ±1° in latitude and longitude, around each of the stations. The albedo values were derived from the OMAERUVG (level 2G) data provided
by OMI (2014), whilst the CCF was extracted from the AIRS L3 daily data products (both available at http://giovanni.gsfc.nasa.gov).
Station Location
Albedo (%) CCF (%)
Minimum Mean Maximum Minimum Mean Maximum
Mendel
63°48′S, 57°53′W
7 m a.s.l.
12 ± 12 62 ± 10 96 ± 3 22 ± 7 63 ± 3 94 ± 2
Concordia
75°06′S, 123°21′E
3233 m a.s.l.
88 ± 4 94 ± 2 100 0 4 ± 3 31 ± 19
128 B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
and to evaluate the contribution of each of them to the solar irradiance
variations at both sites.
4.1. Effect of the geographic position on seasonal irradiance variations
The solar irradiance reaching the Earth's surface strongly depends on
the solar elevation α behaving as cos(90∘
−α) or cos(Z) (Iqbal, 1983).
Fig. 1(b) shows the time patterns of the solar elevations at Mendel and
Concordia stations, expressed by cos(90∘
−αm) and assuming that the
maximum daily solar elevation αm realistically represents the seasonal
changes of the Sun's position on the sky. Hence, due to the different geographic
locations of the two stations it can be expected that both EW and
SWD irradiances should follow a seasonal trend, similar to that given in
Fig. 1(b), so that the irradiance at Mendel station turns out be higher by
20–60% from that observed at Concordia. However, except for DEW ðGÞ,
the observed and modelled irradiance time patterns shown in Fig. 3 indicate
that the irradiance at Mendel slightly exceeds the irradiance at
Concordia only in October and early November, whilst for the rest of
the period the relationship between the irradiances at the two stations
is predominantly opposite to the expected one. The exception concerning
the DEW ðGÞtrends is due to the different ozone columns taken as input for
the model evaluations at both stations.
Fig. 3 reveals another important distinction between the time series
provided by the field measurements at the two stations. It is clearly seen
that the seasonal component of the observed variations, expressed by a
minimum in October and maximum in December according to the
model estimates and Fig. 1(b), is almost masked by sharp fluctuations
at Mendel station, whereas this seasonal trend is easy to recognise at
Concordia. To examine these differences we resorted to the conclusions
achieved by the theory of nonlinear dynamical systems, where it is considered
that the probability distribution of a variable gives an idea about
the character of the processes affecting this variable. Usually, the Gaussian
distribution is attributed to a stochastic process, in other words, the
analysed parameter turns out to be impacted by numerous random factors,
whilst the deviation from the Gaussian distribution is a sign for deterministic
behaviour (Wernik and Yeh, 1996; Hannachi and O'Neill,
2001). In the second case, the variable is mainly affected by a few factors
connected through well determined relationships of each other. Applying
such a test to the parameters DEW(G) and DSWD(G), Fig. 4 shows that
Fig. 2. Panel (a) shows an example of the IR solar spectra calculated by TUV and MODTRAN
computer codes, respectively for Mendel station. The lower two panels demonstrate the
diurnal variations of the SWD irradiance observed at Mendel (b) and Concordia
(c) stations together with the corresponding estimations performed by the TUV model
after its extension to the IR spectral band as described in Section 3.2.
Fig. 3. Time patterns of the doses DEW(G) and DSWD(G), which represent the variations in the erythemally weighted (EW) and short-wave downwelling (SWD) irradiances, observed
during the spring–summer months of 2007–2011 at Mendel and Concordia stations. Time patterns of the corresponding computed by TUV values DEW ðGÞ and DSWDðGÞ that are
considered a realistic representation of the seasonal variations are exhibited by dashed curves. The doses DEW ðGÞ were evaluated for the minimum ozone column registered to be 135
DU at Mendel and 150 DU at Concordia for the period under examination.
129B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
the variations in the solar irradiance at Mendel station are characterised
by almost Gaussian distributions, whereas at Concordia the histograms
present completely different features. This result leads to the
assumption about random behaviour of the solar irradiance at the first
site, when at the second site the observed variations could be accounted
for a prevailing impact of deterministic circumstances (Wernik and Yeh,
1996). Generally speaking, amongst the factors, listed in Section 1, only
the solar elevation follows a well-determined trend (see Fig. 1(b)),
whilst the environmental variables, especially the cloud cover, are subjects
of sharp causal variations over different time scales. Bearing in
mind the ranges of variability in CCF and albedo given in Table 1, it is
logical to conclude that these factors are able to introduce random features
in the solar variations observed at Mendel station. In fact, Fig. 5
demonstrates that the SWD irradiance observed at Concordia is strongly
correlated with the maximum daily value of the solar elevation αm,
whereas a similar strong link between EW irradiance and αm is less
marked likely because of the additional effect of the ozone column
variations that enlarge the scatter of the data points with respect to
the SWD radiation case. In contrast, both EW and SWD components registered
at Mendel station do not show any functional relationship with
the solar elevation. As can be seen from Fig. 5, due to the frequent occurrences
of dense clouds at Mendel station, many days show considerably
lower than the seasonal trend DSWDðGÞ irradiances that disturbs a relationship
like in the Concordia case. Thus, it can be concluded that, except
for the solar elevation and ozone column, the other factors have a
minor importance for the solar irradiance variations taken place at
Concordia station, whilst at Mendel, the environmental factors seem
to play a prevailing role.
Since the seasonal variations at both stations, are characterised by
non-Gaussian distributions (see the right column of Fig. 4), the
corresponding medians were assumed to represent the average levels
D
aver
SWD of the daily SWD doses at the two stations. Table 2 provides
these averages together with the maximum values D
max
SWD and the relative
Fig. 4. Histograms of the daily doses DEW(G), DSWD(G) and DSWDðGÞ presented in Fig. 3, constructed for Mendel and Concordia stations. The corresponding means (black vertical lines) and
medians (grey dashed vertical lines) are also shown. The closeness amongst the mean, median and the peak of the histogram is considered an indicator for the Gaussian distribution.
Fig. 5. Scatter plots showing the dependence of DEW(G) and DSWD(G) on the solar elevation αm at Mendel (up) and Concordia (down) stations. The grey circles present the measured
values, whilst the best fit polynoms, approximating the dependences at Concordia are shown graphically by solid black curves and analytically by the corresponding formulas given
in the lower two panels together with the coefficients of determination R2
. The black dashed curve in the upper right panel indicates the dependence of the modelled
seasonal variations DSWDðGÞ on αm at Mendel station.
130 B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
seasonal variations ΔDSWD, evaluated with respect to D
aver
SWD. The high
seasonal and year-to-year variability in the ozone column (see
Fig. 8(a)) does not allow for the correct estimation of the seasonal
variations in the EW irradiance for the analysed period, so that the corresponding
relative variations in DSWDðGÞ can be considered a first approximation
for the EW component. The values given in Table 2
indicate that the average daily SWD dose registered at Concordia station
is higher than that at Mendel by about 9% in contrast to the expected
tendency as discussed at the beginning of this subsection. Such an occurrence
can be accounted for the significant differences in the environmental
characteristics and the large discrepancy between altitudes of
the two stations (see Table 1) on the one hand, and to the longer daylight
hours at Concordia on the other.
4.2. Effects of the environmental factors on the short-term irradiance
variations
The short-term variations in DEW(G) and DSWD(G) modulating their
seasonal cycle (see Fig. 3), can be attributed to the changes in cloud
cover, aerosols, albedo, and the ozone column. To analyse the contribution
of these factors, the seasonal trends DEW ðGÞ and DSWDðGÞ were
removed from the corresponding data sets that eliminates the effect of
different geographic and altitude positions of the stations. Thus, the relative
short-term variations ΔDEW(G) and ΔDSWD(G) can be defined as:
ΔDWW Gð Þ ¼
DWW Gð Þ−DWW Gð Þ
DWW Gð Þ
; ð5Þ
where WW denotes either EW or SWD. Since DEW ðGÞ and DSWDðGÞ are
upper envelopes of the measurements, the parameters ΔDEW(G) and
ΔDSWD(G) are positive values indicating the relative decrease in the
observed irradiance with respect to the seasonal trend. Fig. 6 exhibits
the time patterns of variables ΔDEW(G) and ΔDSWD(G), whilst their
probability distributions are shown as histograms in Fig. 7. Despite the
closeness amongst the means, medians and histogram peaks found for
Mendel, all the distributions cannot be identified as normal ones. Such
deviations from the expected Gaussian shapes can be attributed to the
different cloud occurrences during the different times within the period
examined in the present study. For instance, as can be seen from the
lower part of Fig. 6, the behaviour of ΔDSWD(G) determined for Mendel
in October is different than the behaviour in January.
Fig. 7(d) shows that the average decrease in the SWD irradiance at
Concordia, determined by the median value is about 5% with respect
to the seasonal trend and this decrease varies between 0% and 10%,
where more than 80% of the cases fall. Since the contribution of the
UV band to the SWD irradiance is negligible, it can be assumed that
the latter is slightly affected by the ozone and consequently, its variations
are caused predominantly by the cloud cover, aerosols and albedo
changes. On the other hand, the aerosol effect over the plateau area can
be neglected as pointed out in Section 1, and considering also that the
surface albedo slightly varied during the studied period (see Table 1) a
prevalent role of the cloud cover features in SWD variations can be expected.
Hence, the cloud occurrences at Concordia station present a
minor importance (only 5 ± 5%) for the irradiance fluctuations that
agrees with the conclusion made in Section 4.1. The cloud effect on
the UV irradiance was assessed to be 15%–45% lower than the effect
on SWD (Calbó et al., 2005) so that the decrease percentage obtained
for SWD component at Concordia can be assumed as an upper bound
for the cloud effect on EW irradiance. Assuming also that the ozone
and clouds affect EW irradiance independently of each other, the contribution
of the ozone column to the average decrease in DEW(G) with respect
to the seasonal trend, evaluated for the minimum ozone amounts
at each of the stations, should be approximately equal to the difference
Table 2
Characteristics of the seasonal trends DSWDðGÞ of the SWD irradiance, observed at Mendel and Concordia stations, represented by the corresponding medians D
aver
SWD, maximum values D
max
SWD
and relative seasonal variations ΔDSWD calculated with respect to the median values (fourth column). The average contributions of the main environmental factors causing a decrease in
the surface solar irradiance with respect to the seasonal trend are presented in the last two columns together with the intervals of their variations in parentheses. It should be pointed out
that the reference seasonal trend DEW ðGÞ, used to assess the impact of the ozone column was evaluated taking into account the minimum ozone values registered at each of the stations
within the studied period.
Station
Characteristics of seasonal trend DSWDðGÞ Effect of the environmental factors (%)
D
aver
SWD (MJ m−2
) D
max
SWD (MJ m−2
) ΔDSWD ¼ D
max
SWD −D
min
SWD
D
aver
SWD
100 (%) Cloud cover, aerosols and albedo Ozone column (concerns only DEW(G))
Mendel 31.2 36.2 71
25 (20–35) (EW)
29 (0–50) (SWD)
46 (31–56)
Concordia 34.0 41.9 94 5 (0–10) (only clouds, on both EW and SWD) 46 (31–56)
Fig. 6. Time patterns of the relative irradiance variations ΔDEW(G) and ΔDSWD(G) estimated for the Mendel and Concordia stations (solid curves) through Eq. (5). The horizontal dashed
lines present the average decreases in the irradiances expressed by the medians of the corresponding distributions given in Fig.7.
131B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
between the medians of the histograms in Fig. 7(c) and (d), i.e. 46%
(0.51–0.05). Fig. 7(c) indicates that the reduction in EW can vary mainly
from −20% to 15% around the median (more than 80% of cases), and assuming
that ±5% of these variations are due to the clouds it can be concluded
that the ozone column causes from −15% to 10% variations
around the median. Hence, the changes in the ozone column from 150
DU to 350 DU (see Fig. 8(a)) at Concordia, reduce the EW irradiance
by about 46% on average and this reduction can vary from 31% to 56%.
Assuming that the ozone column at both stations presents a similar
variability pattern (see Fig. 8(a)) it can be concluded that the cloud
cover contribution to the average EW decrease at Mendel station should
be equal to the difference between the median of the histogram in
Fig. 7(a), which is equal to 0.71 and the average decrease of 0.46 caused
by the ozone. However, this reduction of 25% cannot be attributed to the
clouds only because the effect of seasonal changes in albedo at Mendel
station cannot be neglected as in the case of Concordia (see Table 1).
In addition, the aerosols play an important role in the propagation of
solar radiation through the atmosphere over the Antarctic coastal zone
in contrast to the plateau areas (Tomasi et al., 2007; Láska et al., 2011),
so that the above contribution of 25% should be considered an additive
Fig. 7. Probability distributions of the relative variations ΔDEW(G) and ΔDSWD(G), shown in Fig. 6, both obtained through Eq. (5) for Mendel (a, b) and Concordia (c, d) stations. The vertical
lines represent the mean values (solid lines) and medians (grey dashed lines).
Fig. 8. (a) Time patterns of the ozone column at Mendel and Concordia stations determined by the satellite-based OMI (2014) instrument and the ground-based UV-RAD radiometer
during the studied period. (b) The daily ozone column distributions over the Southern Hemisphere, provided by the National Aeronautic and Space Administration (NASA, USA, http://
ozonewatch.gsfc.nasa.gov/monthly/SH.html) that outline the areas of the ozone depletion (in blue/black) for 4 days of the period under consideration. These days are indicated at the
top of the corresponding maps and the locations of the two stations are shown by red/black (Mendel) and yellow/white (Concordia) circles, respectively. (The additional indications of
the colours are given for the black–white version of the figure).
132 B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
effect of the clouds, albedo and aerosols. These factors determine the behaviour
of SWD irradiance at Mendel, causing a slightly stronger average
decrease of 29% as Figs. 6 and 7(b) indicate that can be due to the different
impact of these factors on EW and SWD spectral bands. At Mendel station,
the variations of ΔDEW(G) around the median are about ±20% (see
Fig. 7(a)) and assuming the contribution of the ozone column between
−15% and 10% as in the Concordia case, the other factors should introduce
variations from −5% to 10% or 15% around the median. This percentage
is much lower than that characterising the corresponding ΔDSWD(G)
variations estimated to be from 0% to 50% at Mendel (see Fig 7(b)).
The above estimates, summarised in Table 2, show that the seasonal
variations in SWD irradiance observed at Mendel station are comparable
with those caused by the environmental factors, whilst the corresponding
seasonal variations at Concordia significantly predominate over the
environmental factors. At Mendel station, the EW component decreases
until 91% due to the combined effects of clouds, albedo, aerosols and
ozone column that is appreciably higher than the seasonal variations,
whilst the similar decrease at Concordia is no more than 66%, which is
lower than seasonal variations. These findings allow the conclusion
that the solar irradiance observed at Concordia is much easier to
forecast.
4.3. Effect of the ozone column on EW irradiance
The ozone is an atmospheric component that determines to high extent
the level of UV-B irradiance at the Earth's surface. The specific dynamical
processes characterising the Antarctic atmosphere, create
favourable conditions for chemical ozone destruction in the stratosphere
during the spring months (Farman et al., 1985; Molina and
Rowland, 1974; Brasseur and Solomon, 2005) that leads to the formation
of a large ozone depletion area over the continent, where the
total ozone drops below 220 DU. Since the effect of such an event is important
for the Antarctic environment, the relationship between the
daily erythemal dose DEW(G) and the ozone column has been analysed
in the present subsection.
Fig. 8(a) presents the variations of the ozone column observed at
Mendel and Concordia stations by OMI (2014) over the studied period.
The ozone data provided by UV-RAD radiometer that are also shown in
Fig. 8(a) differ from the satellite measurements by about ±3% on average.
It can be seen that the ozone column variations follow a similar patterns
at both stations except for December 2007 and October–
November 2010, when the discrepancy between the ozone levels at
the two stations can be attributed to their different position with respect
to the ozone depletion area as it is shown in Fig. 8(b). During
these months the Mendel station was out of the area, whilst for the
rest of the period both stations were on the opposite sides of the
ozone depletion area as the pattern of 20 November 2008 shows.
Bernhard et al. (2010) observed sharp enhancements of the UV-B irradiance
at the Antarctic Peninsula region during November and early December
that were accounted for the combined effects of comparatively
high solar elevation at this time and a transport of air masses with low
ozone concentration towards the Peninsula due to instabilities in the
polar vortex. Similar episodes can be identified in the results discussed
here comparing the data on solar irradiance and ozone column given in
Figs. 3 and 8. In fact, the upper panel of Fig. 3 exhibits values of the EW
irradiance, measured at Mendel station in November 2007 and 2008
that exceeded those registered at Concordia by about or more than
1 kJ m−2
, despite that the corresponding SWD irradiances (see the
lower panel of Fig. 3) indicate a rise in the atmospheric attenuation at
Mendel. Such an occurrence can be attributed to sharp depletion episodes
in the ozone column over the Mendel station with respect to
Concordia during November 2007 and 2008. For instance, total ozone
of 175 DU was registered at Mendel on 23 November 2007, whilst the
corresponding value was 330 DU at Concordia. A similar event was observed
on 24 November 2008, when the ozone column at Mendel was
173 DU versus 335 DU at Concordia.
The sensitivity of the UV irradiance IUV(λ) reaching the ground to the
variations in the ozone column Q is quantified by the radiation amplification
factor (RAF), introduced through the relationship (Madronich,
1993):
IUV λð Þ ∝ Q−RAF λð Þ
: ð6Þ
The RAF is usually estimated by examining the results of observations
performed at fixed solar elevation α and clear-sky conditions
(Booth and Madronich, 1994; Blumthaler et al., 1995; Bodhaine et al.,
1996). Fig. 9 shows the dependence of the EW spectral component
and irradiance at λ = 0.306 μm, measured at Concordia station under
clear sky conditions and different solar elevations, on the OMI ozone
column. Fitting the presented dependences by power curves according
to Eq. (6) we obtain the RAF equal to 0.90 and 0.62 for EW irradiance at
solar elevations of 15° and 5°, respectively, and RAF = 1.55 for IUV(λ=
0.306 μm) and α = 5°. These values are consistent with those reported
by Vitale et al. (2011), who analysed the observations made at
Concordia station during 2008–2009 and found a good agreement between
the evaluated by them RAFs and those, given by previous
estimations.
As mentioned above, the cloud effect on the EW component is 15%–
45% lower than on the SWD irradiance (Calbó et al., 2005) that is confirmed
by the 16% difference between the average impacts of the environmental
factors on EW and SWD irradiances found at Mendel station
(see Table 2). On the other hand, the ozone column changes registered
Fig. 9. (a) EW irradiance measured at Concordia station for solar elevation α=15°, plotted
versus mean daily ozone column Q. (b) The same for α=5°. (c) The same for IUV(λ=
0.306 μm) at α=5°. In each of the plots the grey circles present the measured
irradiances, whilst the black curves indicate the corresponding best fit approximations
performed by power functions according to Eq. (6). The parameter RAF together with
the coefficient of determination R2
are given in each of the plots.
133B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
at the studied sites produce variations in EW component higher than the
other environmental factors (Table 2) that leads us to assume the ratio
RD Qð Þ ¼
DEW Gð Þ
DSWD Gð Þ
ð7Þ
as a function predominantly of the mean daily ozone column Q. In fact,
as Fig. 10 indicates, this ratio is satisfactorily correlated with Q and, following
the same approach as in determining of the RAF, the fitting of the
data shown in Fig. 10 by power functions (see Eq. (6)), defines an analogous
parameter RAF′. As can be seen, the values of RAF′, given in Fig. 10
for both Mendel and Concordia stations are very close to each other and
are consistent with the value of RAF, found for EW irradiance observed at
Concordia for α = 15° (see Fig. 9(a)) that assumes the ratio RD determined
by Eq. (7) as being able to give a realistic estimate of the radiation
amplification factor RAF. Such a conclusion makes possible the assessment
of the RAF at stations, where only broadband observations of the
solar radiation are carried out that offers further opportunities for improving
our knowledge on solar UV-B irradiance and its spatial distribution
over high latitude areas.
5. Conclusions
The variations in erythemally weighted (EW) and short-wave
downwelling (SWD) solar irradiances observed during the daylight periods
of 2007–2011 at Mendel and Concordia Antarctic stations, which
are placed at the coastal region and on the Eastern Antarctic Plateau, respectively,
have been analysed in the present study. The adopted approach
allows a separation of the effects produced by the factors,
predominantly affecting the surface solar radiation. The seasonal irradiance
variations, determined by the solar elevation trend are clearly seen
at Concordia station, where this astronomical factor causes 94% of the irradiance
changes, whilst the cloud cover fluctuations reduce the solar
radiation by 5% on average. On the contrary, the seasonal trend observed
at Mendel, which causes about 71% of irradiance variations, is
significantly masked by the clouds, aerosols and albedo changes that
lead to an average decrease in the solar irradiance of about 25% for EW
and 29% for SWD. Bearing in mind that the seasonal trend of the solar elevation
is a strongly determined parameter, it can be concluded that the
radiation on the interior Antarctic Plateau should be much easier to predict
than at coastal region where the day-to-day variability in cloud
cover, together with the changes in aerosol concentrations and surface
albedo has a higher weight.
The ozone column was found to be a subject of similar change patterns
at both stations over the studied period, reducing by nearly 46%
the EW irradiance with respect to the corresponding EW, calculated
for the minimum ozone value at each site. It is worth mentioning that
for certain times, the Mendel station was out of the ozone depletion
area occupying the Antarctic continent during the spring time, due to
its lower latitude with respect to Concordia. The performed analysis allows
the conclusion that the ratio RD between daily doses of EW and
SWD irradiances, considered a function of the ozone column Q, is able
to provide a realistic estimate of the radiation amplification factor,
which quantifies the impact of the ozone on surface UV irradiance.
Acknowledgements
This research was supported by the Italian programme PNRA
(Progetto Nazionale di Ricerche in Antartide) and developed as a part
of the subproject 2004/2.04 “Implementation of the BSRN station at
Dome Concordia”. The contribution of Kamil Láska has been supported
by the project of Masaryk University MUNI/A/1370/2014 “Global environmental
changes in time and space (GlobST)”. The authors would
like to thank the Czech Antarctic Research Infrastructure and the crew
of the Johann Gregor Mendel and Dome Concordia stations between
2007 and 2011 for the extensive technical support and field assistance.
We also thank the OMI science team for providing the ozone column
data used in this study, which are available from NASA GSFC FTP
website ftp://jwocky.gsfc.nasa.gov/pub/omi/data/overpass/. The OMI
surface reflectivity and AIRS cloud data products were obtained from
the NASA Goddard Earth Sciences Data and Information Services Center
(at http://giovanni.gsfc.nasa.gov).
Appendix A. Supplementary data
Supplementary data associated with this article can be found in the
online version, at http://dx.doi.org/10.1016/j.atmosres.2016.01.005.
These data include the Google map of the most important areas described
in this article.
References
Berk, A., Anderson, G.P., Acharya, P.K., Shettle, E.P., 2011. MODTRAN 5.2.1 User's Manual.
Spectral Sciences, Inc., 4 Fourth Ave., Burlington, MA 01803-3304 & Air Force Research
Laboratory, Space Vehicles Directorate, Air Force Materiel Command,
Hanscom AFB, MA 01731-3010 (May 2011, 69 pp.).
Bernhard, G., Booth, C.R., Ehramjian, J.C., 2004. Version 2 data of the National Science
Foundation's Ultraviolet Radiation Monitoring Network: South Pole. J. Geophys. Res.
109, D21207.
Bernhard, G., Booth, C.R., Ehramjian, J.C., 2005. UV climatology at Palmer Station,
Antarctica. In: Bernhard., G., Slusser, J.R., Herman, J.R., Gao, W. (Eds.), Ultraviolet
Ground- and Space-Based Measurements, Models, and Effects V. Proceedings of
SPIE International Society of Optical Engineering 5886, pp. 51–62.
Bernhard, G., Booth, C.R. & Ehramjian, J. C. 2010. Climatology of ultraviolet radiation at
high latitudes derived from measurements of the National Science Foundation's Ultraviolet
Spectral Irradiance Monitoring Network, in: Gao, W., Schmoldt, D.L., Slusser,
J.R, Herman, J.R, (Eds) UV Radiation in Global Climate Change: Measurements,
Modeling and Effects on Ecosystems. Tsinghua University Press, Beijing and Springer,
New York, 48–72.
Blumthaler, M., Salzgeber, M., Ambach, W., 1995. Ozone and ultraviolet-B irradiance: experimental
determination of the radiation amplification factor. Photochem. Photobiol.
61, 159–162.
Bodhaine, B.A., McKenzie, R.L., Johnston, P.V., Hofmann, D.J., Dutton, E.G., Schnell, R.C.,
Barnes, J.E., Ryan, S.C., Kotkamp, M., 1996. New ultraviolet spectroradiometer at
Mauna Loa Observatory. Geophys. Res. Lett. 23, 2121–2124.
Booth, C.R., Madronich, S., 1994. Radiation amplification factors: improved formulation
accounts for large increases in ultraviolet radiation associated with Antarctic ozone
depletion. In: Weiler, C.S., Penhale, P.A. (Eds.), Ultraviolet Radiation in Antarctica:
Fig. 10. Scatter plots of the ratio RD(Q) evaluated through Eq. (7) versus the mean daily
ozone column Q (OMI, 2014) for the considered stations. The grey circles present the
values derived from the measurements, whilst the black curves indicate the best fits made
through power functions, analogously to the cases presented in Fig. 9. The corresponding
coefficients of determination R2
and parameters RAF′ are also given in the plots.
134 B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
Measurements and Biological EffectsAGU Antarctic Research Series 62. American
Geophysical Union, Washington, DC, pp. 39–42.
Brasseur, G.P., Solomon, S., 2005. Aeronomy of the Middle Atmosphere. Springer,
pp. 443–531.
Calbó, J., Pagès, D., González, J.-A., 2005. Empirical studies of cloud effects on UV radiation:
a review. Rev. Geophys. 43, RG2002. http://dx.doi.org/10.1029/2004RG000155.
Cordero, R.R., Damiani, A., Ferrer, J., Jorquera, J., Tobar, M., Labbe, F., Carrasco, J., Laroze, D.,
2014. UV irradiance and albedo at Union Glacier Camp (Antarctica): a case study.
PLoS One 9, e90705.
Farman, J.C., Gardiner, B.G., Shanklin, J.D., 1985. Large losses of total ozone in Antarctica
reveal seasonal C1Ox/NOx interaction. Nature 315, 207–210.
Hannachi, A., O'Neill, A., 2001. Atmospheric multiple equilibria and non-Gaussian behaviour
in model simulations. Q. J. R. Meteorol. Soc. 127, 939–958.
Hrbáček, F., Láska, K., Engel, Z., 2016. Effect of snow cover on active-layer thermal regime
— a case study from James Ross Island, Antarctic Peninsula. Permafr. Periglac. Process.
http://dx.doi.org/10.1002/ppp.1871.
Iqbal, M., 1983. An Introduction to Solar Radiation. Academic Press, Toronto (390 pp.).
Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2009. Prediction of erythemally
effective UVB radiation by means of nonlinear regression model. Environmetrics 20,
633–646.
Láska, K., Prošek, P., Budík, L., Budíková, M., Milinevsky, G., 2010. Estimation of solar UV
radiation in maritime Antarctica using nonlinear model including cloud effects. Int.
J. Remote Sens. 31, 831–849.
Láska, K., Budík, L., Budíková, M., Prošek, P., 2011. Method of estimating solar UV radiation
in high-latitude locations based on satellite ozone retrieval with an improved algorithm.
Int. J. Remote Sens. 32, 3165–3177.
Lee, Y.G., Koo, J.-H., Kim, J., 2015. Influence of cloud fraction and snow cover to the variation
of surface UV radiation at King Sejong Station, Antarctica. Atmos. Res. 164–165,
99–109.
Levelt, P.F., van den Oord, G.H.J., Dobber, M.R., Claas, J., Visser, H., de Vries, J., Stammes, P.,
Lundell, J.O.V., Saari, H., 2006. The Ozone Monitoring Instrument. IEEE Trans. Geosci.
Remote Sens. 4, 1093–1101.
Lubin, D., Frederick, J.E., 1991. The ultraviolet radiation environment of the Antarctic Peninsula:
the roles of ozone and cloud cover. J. Appl. Meteorol. 30, 478–493.
Madronich, S., 1993. UV radiation in the natural and perturbed atmosphere. In: Tevini, M.
(Ed.), Environmental Effects of UV (Ultraviolet) Radiation. Lewis, Boca Raton,
pp. 17–69.
Madronich, S., Flocke, S., 1997. Theoretical estimation of biologically effective UV radiation
at the Earth's surface. In: Zerefos, C.S., Bais, A.F. (Eds.), Solar Ultraviolet Radiation
— Modeling, Measurements and EffectsNATO ASI Series I: Global Environmental
Change vol. 52. Springer-Verlag, Berlin, pp. 23–48.
McKenzie, R.L., Aucamp, P.J., Bais, A.F., Björn, L.O., Ilyas, M., 2007. Changes in biologicallyactive
ultraviolet radiation reaching the Earth's surface. Photochem. Photobiol. Sci. 6,
218–231.
McKinlay, A.F., Diffey, B.L., 1987. A reference action spectrum for ultraviolet induced erythema
in human skin. CIE J. 6, 17–22.
Molina, M.J., Rowland, F.S., 1974. Stratospheric sink for chlorofluoromethanes: chlorine
atom catalysed destruction of ozone. Nature 249, 810–814.
Ohmura, A., Dutton, E.G., Forgan, B., Frohlich, C., Gilgen, H., Hegner, H., Heimo, A., KonigLanglo,
G., McArthur, B., Muller, G., Philipona, R., Pinker, R., Whitlock, C.H., Dehne, K.,
Wild, M., 1998. Baseline Surface Radiation Network (BSRN/WCRP): new precision radiometry
for climate research. Bull. Am. Meteorol. Soc. 79, 2115–2136.
OMI (Ozone Monitoring Instrument), 2014. http://www.nasa.gov/mission_pages/aura/
spacecraft/omi.html.
Orsini, A., Tomasi, C., Calzolari, F., Nardino, M., Cacciari, A., Georgiadis, T., 2002. Cloud
cover classification through simultaneous ground-based measurements of solar and
infrared radiation. Atmos. Res. 61, 251–275.
Perrina, J.-M., Thuillierb, G., Fehrenbacha, M., Huppert, F., 2005. A comparison between radiative
transfer calculation and pyranometer data gathered at Observatoire de Haute
Provence. J. Atmos. Sol. Terr. Phys. 67, 449–463.
Petkov, B., Vitale, V., Tomasi, C., Bonafé´, U., Scaglione, S., Flori, D., Santaguida, R., Gausa, M.,
Hansen, G., Colombo, T., 2006. Narrow-band filter radiometer for ground-based measurements
of global UV solar irradiance and total ozone. Appl. Opt. 45, 4383–4395.
Petkov, B., Vitale, V., Gröbner, J., Hülsen, G., De Simone, S., Gallo, V., Tomasi, C., Busetto, M.,
Barth, V.L., Lanconelli, C., Mazzola, M., 2012. Short-term variations in surface UV-B irradiance
and total ozone column at Ny-Ålesund during the QAARC campaign. Atmos.
Res. 108, 9–18.
Platnick, S., King, M.D., Ackerman, S.A., Menzel, W.P., Baum, B.A., Riedi, J.C., Frey, R.A.,
2003. The MODIS cloud products: algorithms and examples from Terra. IEEE Trans.
Geosci. Remote Sens. 41, 459–473.
Sarti, P., Negusini, M., Tomasi, C., Petkov, B.H., Capra, A., 2013. Thirteen years of integrated
precipitable water derived by GPS at Mario Zucchelli Station, Antartica. Ann.
Geophys. 56, R0221.
Stamnes, K., Jin, Z., Slusser, J., 1992. Several-fold enhancement of biologically effective ultraviolet
radiation levels at McMurdo Station Antarctica during the 1990 ozone
“hole”. Geophys. Res. Lett. 19, 1013–1016.
Susskind, J., Barnet, C., Blaisdell, J., Iredell, L., Keita, F., Kouvaris, L., Molnar, G., Chahine, M.,
2006. Accuracy of geophysical parameters derived from Atmospheric Infrared Sounder/Advanced
Microwave Sounding Unit as a function of fractional cloud cover.
J. Geophys. Res. 111, D09S17.
Tomasi, C., Vitale, V., Lupi, A., Di Carmine, C., Campanelli, M., Herber, A., Treffeisen, R.,
Stone, R.S., Andrews, E., Sharma, S., Radionov, V., von Hoyningen-Huene, W., Stebel,
K., Hansen, G.H., Myhre, C.L., Wehrli, C., Aaltonen, V., Lihavainen, H., Virkkula, A.,
Hillamo, R., Ström, J., Toledano, C., Cachorro, V.E., Ortiz, P., de Frutos, A.M., Blindheim,
S., Frioud, M., Gausa, M., Zielinski, T., Petelski, T., Yamanouchi, T., 2007. Aerosols in
polar regions: a historical overview based on optical depth and in situ observations.
J. Geophys. Res. 112, D16205.
Tomasi, C., Petkov, B., Benedetti, E., Valenziano, L., Vitale, V., 2011. Analysis of a 4 year radiosonde
data set at Dome C for characterizing temperature and moisture conditions
of the Antarctic atmosphere. J. Geophys. Res. 116, D15304.
Tomasi, C., Petkov, B.H., Mazzola, M., Ritter, C., Di Sarra, A.G., Di Iorio, T., Del Guasta, M.,
2015. Seasonal variations of the relative optical air mass function for background
aerosol and thin cirrus clouds at Arctic and Antarctic sites. Remote Sens. 2015 (7),
7157–7180.
Vitale, V., Petkov, B., Goutail, G., Lanconelli, C., Lupi, A., Mazzola, M., Busetto, M., Pazmino,
A., Schioppo, R., Genoni, L., Tomasi, C., 2011. Variations of UV irradiance at Antarctic
Concordia station during the austral springs of 2008 and 2009. Antarct. Sci. 23,
389–398.
Wernik, A.W., Yeh, K.C., 1996. A note on chaotic vs. stochastic behaviour of the highlatitude
ionospheric plasma density fluctuations. Nonlinear Process. Geophys. 3,
47–57.
Wuttke, S., Seckmeyer, G., Koenig-Langlo, G., 2006. Measurements of spectral snow albedo
at Neumayer, Antarctica. Ann. Geophys. 24, 7–21.
135B.H. Petkov et al. / Atmospheric Research 172–173 (2016) 126–135
6
Paper 5
Ice thickness, areal and volumetric changes of Davies Dome and
Whisky Glacier (James Ross Island, Antarctic Peninsula) in
1979-2006.
Engel, Z., Nývlt, D., Láska, K., 2012,
Journal of Glaciology, 58, pp. 904–914
Ice thickness, areal and volumetric changes of Davies Dome and
Whisky Glacier (James Ross Island, Antarctic Peninsula)
in 1979–2006
Zbyne˘k ENGEL,1,2
Daniel NY´VLT,2
Kamil LA´ SKA3
1
Department of Physical Geography and Geoecology, Charles University in Prague, Praha, Czech Republic
E-mail: engel@natur.cuni.cz
2
Czech Geological Survey, Brno, Czech Republic
3
Department of Geography, Masaryk University, Brno, Czech Republic
ABSTRACT. This study calculates area, volume and elevation changes of two glaciers on James Ross
Island, Antarctica, during the period 1979–2006. Davies Dome is a small ice cap. Whisky Glacier is a
valley glacier. Ground-penetrating radar surveys indicate ice thickness, which was used for calculations
of the bed topography and volume of both glaciers. Maximum measured ice thicknesses of Davies Dome
and Whisky Glacier are 83 Æ 2 and 157 Æ 2 m, respectively. Between 1979 and 2006, the area of the ice
cap decreased from 6.23 Æ 0.05 km2
to 4.94 Æ 0.01 km2
(–20.7%), while the area of the valley glacier
reduced from 2.69 Æ 0.02 km2
to 2.40 Æ 0.01 km2
(–10.6%). Over the same period the volume of the ice
cap and valley glacier reduced from 0.23 Æ 0.03 km3
to 0.16 Æ 0.02 km3
(–30.4%) and from
0.27 Æ 0.02 km3
to 0.24 Æ 0.01 km3
(–10.6%), respectively. The mean surface elevation decreased by
8.5 Æ 2.8 and 10.1 Æ 2.8 m. The average areal ($0.048–0.011 km2
a–1
) and volumetric ($0.003–
0.001 km3
a–1
) changes are higher than the majority of other estimates from Antarctic Peninsula glaciers.
INTRODUCTION
The mean annual temperature in the Antarctic Peninsula
(AP) region has increased by 1.58C to 3.08C since the 1950s
(e.g. King, 1994; Vaughan and others, 2001, 2003; Turner
and others, 2005). This warming has coincided with the
disintegration of ice shelves and the retreat of many landterminating
glaciers in the region (e.g. Skvarca and others,
1998; Cook and others, 2005; Davies and others, 2011;
Glasser and others, 2011). Land-based glacial mass loss has
been most prominent at the northern tip of the AP due to the
pronounced glacier acceleration (Rignot and others, 2008).
Detailed analyses of satellite imagery and some isolated
mass-balance measurements together suggest that small
land-terminating glaciers are especially sensitive to climate
change (Rau and others, 2004; Skvarca and others, 2004).
The sensitivity of small land-terminating glaciers to climate
is related to a short response time of variations in mass
balance (Knap and others, 1996).
Along with direct mass-balance measurements, changes
in surface area and volume of glaciers are important
indicators of climate changes. Volumetric characteristics
are necessary for improved understanding of: (1) the
responses of these sensitive glacier systems to climate
change; (2) the present contribution of AP glaciers to sealevel
rise; and (3) future predictions of AP glacier behaviour.
Although changes in glacier mass over long time periods can
be analysed by volumetric methods, these do not replace the
field-based techniques, which provide detailed measures of
glacier change in space and time. Direct methods are timeconsuming
and expensive, however, and typically only
cover short time periods (Gudmundsson and Bauder, 1999).
Glacier bed topography and ice thickness are important for
understanding ice dynamics (e.g. Rippin and others, 2011a)
and for assessing basal pressure and thermal regimes (e.g.
Rippin and others, 2011b). Determination of glacial mass
changes may be used to detect shifts in climate and to
predict glacial changes associated with future climate
conditions (e.g. Paterson, 1994; Fountain and others, 1999).
In this study, we present the volumetric data for Davies
Dome ice cap and Whisky Glacier on James Ross Island
(JRI). The name Whisky Glacier is used here for IJR-45
glacier as given in the inventory by Rabassa and others
(1982). The name Whisky Glacier was referred to by Chinn
and Dillon (1987) and is given on the Czech Geological
Survey (2009) map. It should not be confused with the
tidewater Whisky Glacier located in Whisky Bay, as named
by the British Antarctic Survey (2010). The changes in
elevation, area and geometry between 1979 and 2006 were
derived from digital elevation models (DEMs) and aerial
photographs. Ice thickness was measured during a field
campaign in January and February 2010 using groundpenetrating
radar (GPR). Glacier volume and glacier bed
elevations were then derived from these data.
STUDY AREA
Approximately 81% of JRI is covered with 138 glaciers,
which are concentrated in the central and southern parts of
the island (Rabassa and others, 1982). The variety of existing
glacier systems and fluctuations of glacier front positions on
JRI have been presented by Rabassa and others (1982),
Skvarca and others (1995) and Rau and others (2004). Massbalance
changes have been studied at Glaciar Bahı´a del
Diablo on Vega Island (Skvarca and others, 2004) and at
Davies Dome and Whisky Glacier on JRI (Ny´vlt and others,
2010; La´ska and others, 2012). However, these existing
glaciological studies only describe elevation and surface
area; data on ice thickness and hence on ice volume are
hitherto unknown.
The investigated glaciers are located on the Ulu
Peninsula, northern JRI (Fig. 1). This predominantly
glacier-free area extends northwards from the central part
Journal of Glaciology, Vol. 58, No. 211, 2012 doi: 10.3189/2012JoG11J156904
of the island, which is dominated by the JRI ice field, an
area of ice covering 587 km2
(Rabassa and others, 1982).
The climate of the Ulu Peninsula is influenced by the
regional-scale atmospheric circulation formed by the AP,
which generally provides a barrier to westerly winds
associated with cyclonic systems centred in the circumpolar
trough (King and others, 2003). High interdiurnal and
seasonal variability of cloudiness, incoming solar radiation
and near-surface air temperature can be found along the
eastern coast of the AP (Marshall, 2002; La´ska and others,
2011) although the long-term trend appears to involve
increased frequency of warm airflows advected over the AP
as described by Marshall and others (2006).
Contemporary glaciation of JRI is characterized by ice
caps, outlet glaciers and different types of alpine glaciers
(Rabassa and others, 1982). The JRI ice field covers the
southern part of the island where only small masses of rock
protrude to the surface. By contrast, most ice has retreated
from the northern part of the Ulu Peninsula. Small glaciers
have persisted in higher locations with snow accumulation
areas above $460 m a.s.l. (Ny´vlt and others, 2010). The
glacial features include an ice cap on Davies Dome (514 m)
and on Lachman Crags (641 m), cirque, apron and piedmont
glaciers around these plateaux, and valley glaciers carved
into Smellie Peak (704 m) and Stickle Ridge (767 m). These
glaciers are relics of an ice cap that covered the whole island
in the last glacial period (Rabassa, 1983; Ingo´lfsson and
others, 1992). In the northern part of JRI, this ice cap merged
with the northern Prince Gustav Ice Stream, which drained
the northeastern AP (Rabassa, 1983). During the Holocene,
the ice cap was linked to the Ro¨hss Bay Ice Shelf, until it
disintegrated in 1995 (Glasser and others, 2011).
Two contrasting types of glacier from the Ulu Peninsula
were selected to assess recent volume changes. Davies Dome
(DD) is an ice dome (63852’–63854’ S, 58801–58806’ W)
which originates on the surface of a flat volcanic mesa at
>400 m a.s.l. and terminates as a single 700 m wide outlet in
Whisky Bay. In 2006, DD had an area of $6.5 km2
and lay in
the altitude range 0–514 m a.s.l. (Table 1). Whisky Glacier
(WG) is a cold-based land-terminating valley glacier (63855’–
63857’ S, 57856’–57858’ W), which is surrounded by an
extensive area of debris-covered ice (Chinn and Dillon,
1987). In 2006, the glacier covered an area of $2.4 km2
and
ranged from 520 to 215 m a.s.l. Mean annual near-surface air
temperatures of –7.88C and –8.88C have been recorded since
2009 by automatic weather stations (AWSs) on WG (356 m)
and DD (514 m), respectively (Fig. 1). Summer temperatures
(November–February) vary from –98C to +108C on WG and
from –128C to +78C on DD. The snowmelt period generally
starts at the beginning of November and lasts until the end of
February. The meteorological data were collected as part of
this study using AWSs on glaciers (Fig. 1).
DATA AND METHODS
DEMs and aerial photographs of DD and WG were used to
extract topographic information for glacier extent and
geometry. Two separate DEMs produced by the GEODIS
company (Czech Republic) were stereoplotted from aerial
photographs obtained by the British Antarctic Survey (BAS)
in 1979 and 2006 (Czech Geological Survey, 2009).
Ground-control points for the photogrammetric processing
were obtained in the deglaciated landscape in 2008 using a
dual-frequency differential GPS receiver. The 1979 aerial
photos were registered with root-mean-square (rms) errors of
2.0 m in both the horizontal and vertical. Rms errors of the
2006 dataset are 0.7 m and 0.8 m in the horizontal and
vertical direction, respectively. Meixner (2009) provides a
more complete description of the DEM generation method.
The GPR survey was conducted in January and February
2010 in accordance with numerous descriptions of radar
investigations of glacier thickness (e.g. Plewes and Hubbard,
2001; Navarro and Eisen, 2010). GPR surveys were applied
on DD (except over a steep and crevassed outlet glacier) and
across the whole surface of WG. GPR data were collected
along six radial profiles on DD (Fig. 1). On WG, GPR was
carried out along the central longitudinal profile and five
transverse profiles (Fig. 1). A longitudinal profiling was
carried out from the glacier terminus to the upper reaches of
the ablation zone (226–444 m a.s.l.). Three transverse profiles
were undertaken in the lower part of the snout
delimited by lateral moraines and two profiles in the upper
zone enclosed by the left lateral moraine and a valley slope
facing to the west. GPR profiling was carried out using an
unshielded 50 MHz Rough Terrain Antenna and RAMAC
CU-II control unit (MALA˚ GeoScience, 2005). The signal
acquisition time was set to 2040 ns, and scan spacing ranged
from 0.2 to 0.3 m. GPR position was given by GPS: a Garmin
GPSMAP 60CSx receiver (Garmin International, 2009) was
used to determine coordinates for the corresponding trace
numbers with the horizontal rms position accuracy <4 m.
The GPR profiling equipment was man-hauled at a mean
speed of $0.5 m s–1
.
GPR data were processed and interpreted using the
REFLEXW software version 5.0 (Sandmeier, 2008). The raw
GPR profiles were filtered using an automatic gain function
(AGC-Gain), background removal and static correction. The
AGC-Gain was used to amplify deep signals reduced due to
pulse dispersion and attenuation. Strong direct wave signal
Fig. 1. Location of GPR transects (labelled as D-n and W-n, n being
an integer) on Davies Dome and Whisky Glacier (dark grey areas) on
the Ulu Peninsula, northern JRI. Other ice bodies are shown in
medium grey. Triangles mark the position of meteorological stations.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers 905
and noise reflections were eliminated using the background
removal filter. Finally, the radargrams were corrected for
topographic elevation. For this purpose, elevations from the
collected GPS data were used. Depth was calculated
assuming a wave velocity for cold glacier ice, 0.168 m ns–1
(Narod and Clarke, 1994). The velocity model used to
convert two-way travel times to depths affects the accuracy
of the ice thickness determination. As signal velocities in
glacier ice range from 0.167 to 0.169 m ns–1
(Kovacs and
others, 1995; Pellikka and Rees, 2009), the velocity used
was accurate to within 0.002 m ns–1
. Thus, the estimated
uncertainty of conversion of travel time to depth is $1%.
Moreover, the accuracy in depth measurement is restricted
by the ability to determine the first arrival of the signal (Jol,
2009). This uncertainty was approximated by a quarter of the
signal’s wavelength in ice, which corresponds to 0.84 m for
the 50 MHz antenna used.
Ice-thickness values converted from two-way travel times
along GPR profiles were used to calculate glacier volumes.
Because the accuracy of the calculation depends on the
algorithm applied we investigated and selected the most
appropriate interpolation first. We applied five techniques
(inverse distance weighing, kriging, minimum curvature,
spatial neighbour and triangulation) to the complete icethickness
datasets to create grids of glaciers. We then
removed the ice-thickness information along one complete
section from each of the two datasets. We selected the
section D-1, which represents both the upper part of an ice
dome on a flat volcanic plateau and the lower termination
on mountain slopes. For WG, we chose the section W-4 as it
crosses the middle part of the snout. The reduced datasets
and the same interpolation techniques were used to generate
a new series of grids from which ice-thickness values were
assigned for verification points along sections D-1 and W-4.
The interpolated values were compared with the measured
values and the difference was expressed using standard
deviation values. Results of interpolation techniques were
also evaluated based on glacier volumes calculated from a
complete and reduced dataset. A comparison of the calculated
values shows that the minimum curvature technique
most accurately represents the ice-thickness data (Table 2).
This method was selected for further processing of the data.
The volume of glaciers was calculated using a combination
of 2006 surface elevation data and GPR-based icethickness
data taken in 2010. The elevation of GPS-derived
locations along GPR profiles was taken from the 2006 DEM
because the accuracy of an altitude measurement was low
during the GPR survey. The vertical error of GPS-based
coordinates ($8–18 m) is larger than the vertical error of the
DEM and elevation changes of glacier surfaces between
2006 and 2010. According to mass-balance measurements
on DD and WG the elevation differences ranged from –0.4
to +1.1 m during this period (Ny´vlt and others, 2012). The
volume of glaciers for 1979 was calculated based on the
2006 volumes and the difference between 1979 and 2006
surface DEMs. The uncertainty associated with the glacier
volume determination was calculated as the product of the
glacier surface area and vertical uncertainty. The vertical
uncertainty for 2006 is related to the accuracy of ice
thickness and to surface elevation changes in 2006–10. The
uncertainty for 1979 was calculated as the sum of the glacier
surface topography-related uncertainty for 1979 plus the
uncertainty of the glacier volume for 2006. The mean ice
thickness of DD and WG is 32.4 and 99.6 m, respectively,
and average uncertainties of conversion of travel time to ice
thickness are 0.3 and 1.0 m, respectively. The total
uncertainty of the ice volume per glacier area in 2006 is
3.5 m3
m–2
for DD and 4.1 m3
m–2
for WG. Uncertainties of
5.5 and 6.1 m3
m–2
have been obtained for the 1979 volume
of DD and WG, respectively.
Glacier bed topography maps were generated based on
the 2006 DEM and the 2010 GPR measurements. The bed
elevation was calculated for GPR survey locations by
subtracting the ice thickness from the surface elevation. At
the margin of DD, the ice thickness was set to zero except
for the upper reaches for the outlet where three bounding
points were placed. For these points, depth values were
approximated from the nearest points of GPR measurements
and from the position of a small rock outcrop that protrudes
to the surface of the eastern part of the outlet. The bed
elevation at the WG limit was determined from marginal
measurement points at GPR cross-sections that represent the
crest of lateral moraines. The minimum curvature interpolation,
which generates a smoothed surface with low
standard deviation, was then used to construct an 8 m
gridcell DEM of glacier beds. From the DEMs, contour lines
at 10 m spacing were constructed for bed topography.
RESULTS
Glacier bed topography
The boundary between the glaciers and underlying bedrock
and/or till is recorded as continuous basal reflection in all of
our radargrams (Figs 2 and 3). A strong reflection at the
glacier bed is clearly visible in DD profiles where the slope
of the presumed bedrock is small (Fig. 2a and c). By contrast,
the amplitude of reflections is halved in the area below the
subglacial eastern flank of the plateau as shown in transect
D-3 (Fig. 2b) and in the eastern sections of transects D-1,
Table 1. Morphological characteristics of DD and WG for 2006
Name Type Elevation range Mean elevation Mean thickness Maximum thickness Length* Area Volume
m a.s.l. m a.s.l. m m km km2
km3
Davies Dome Dome 124–515 368.5 Æ 0.8 32.4 Æ 1.2 83 Æ 2 3.43{
4.94 Æ 0.01 0.16 Æ 0.02
6.49 Æ 0.01{
Whisky Glacier Valley 215–520 340.6 Æ 0.8 99.6 Æ 1.8 158 Æ 2 3.19 2.40 Æ 0.01 0.24 Æ 0.01
*The length was measured along the central axis of the DD outlet and WG snout.
{
The value for the whole of DD including the outlet glacier.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers906
D-2 and D-6. Beneath WG, the strong and continuous
reflection was recorded in the upper half of profile W-1
(Fig. 3a). There are only a few minor diffractions visible
along the glacier bed in this section. The signal strength
decreases slightly in the middle section of the longitudinal
profile $1.8 km from the glacier terminus. The decreased
amplitude prevails in the lower part of the snout as seen
along the lower part of the longitudinal profile and in crossprofiles
W-4, W-5 and W-6 (Fig. 3c).
The subglacial topography of DD is relatively complex.
The uppermost part of DD is located on a near-horizontal
surface. GPR data indicate a slightly undulating glacier bed
with an elevation range of <50 m (Fig. 2a and c). Most of the
near-flat bed is inclined to the north, and only $7% of this
zone faces southeast or south. An undulating glacier bed
extends to half the length of the northwest–southeast axis of
DD. Here the reflection boundary descends from 400–
430 m a.s.l. by 70–90 m to more dissected bed topography
(Fig. 2b). The fall line is oriented towards the southeast and
represents an eastern flank of the volcanic plateau (Fig. 4a).
Steeper slopes and concave wide channel-like depressions
are indicated in the radargrams of two sections of the flank.
The southernmost of these coincides with the main outlet
glacier that drains DD into Whisky Bay. The second slope
depression is carved into the central part of the bedrock step
$1 km from DD summit to the east.
WG bed interface has a simple geometry. The data
obtained along the centre flowline show an even longitudinal
profile without any subglacial steps or depressions
(Fig. 3a). The bed topography is inclined to the north with
low average slope angle of $58. The mean slope of the
basal contact in the lower part of the profile is <28. The
transverse profiles W-2 and W-3 across the central line of
WG show an asymmetric bottom of a trough with steeper
slopes rising towards the east, where the glacier is
constricted by a steep slope of Smellie Peak (Fig. 3b). The
western part of the bed topography is characterized by a
gentle slope angle, which is within 58 of the glacier surface.
As a result, the glacier bed is located up to $125 m below
the western margin of the glacier surface (Fig. 5). The
pattern of the bed topography suggests that ice extends as a
debris-covered glacier towards the west. Similar conditions
are recorded in transverse profiles W-4, W-5 and W-6,
which were undertaken in the lower part of WG. These
Table 2. Results of glacier volume calculations
Gridding method Davies Dome Whisky Glacier
Volume SD* ÁV{
Volume SD* ÁV{
km3
% km3
%
Inverse distance weighing 0.153 10.2 2.0 0.236 33.8 0.5
Kriging 0.153 10.5 1.9 0.236 33.5 0.6
Minimum curvature 0.160 8.5 1.1 0.239 21.0 0.2
Spatial neighbour 0.152 10.7 2.3 0.232 29.7 0.6
Triangulation 0.154 9.4 2.4 0.232 20.7 0.5
*Standard deviation of interpolated ice thickness along sections D-1 and W-4.
{
Difference in glacier volume interpolated based on complete dataset and on data without sections D-1 and W-4.
Fig. 2. Selected GPR transects across DD. (a) Profile D-2 covers the dome glacier in the northwest–southeast direction showing the
subglacial surface of the DD mesa and its transition to more undulated bed topography in the eastern part of DD. (b) Profile D-3 shows
the lower part of the glacier below the eastern margin of the DD mesa. (c) Profile D-4 shows the upper flat-bottomed part of the glacier.
Inverted triangles indicate points of crossover of other profiles.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers 907
profiles suggest that the glacier ice also extends beyond the
eastern margin of WG forming a core of a right lateral
moraine (Fig. 3c). The minimum extent of debris-covered
ice is $0.5 km2
. However, if the whole area of lateral and
terminal moraines is taken into account then the maximum
extent of ice-covered moraines is $3 km2
.
Surface area, elevation and volume change
The area of DD in 1979 was 8.33 Æ 0.06 km2
and in 2006
was 6.49 Æ 0.01 km2
. The dome area decreased from
6.23 Æ 0.05 to 4.94 Æ 0.01 km2
, losing 20.7% of its area
(Table 3). Assuming a constant rate of change, the ice cap
therefore lost 0.048 Æ 0.002 km2
a–1
. Almost 70% of measured
areal loss occurred in the southeast periphery of the
dome at 120–370 m a.s.l.
Over the same time period, WG lost 10.6% of its nondebris-covered
area, decreasing from 2.69 Æ 0.02 km2
to
2.40 Æ 0.01 km2
. The mean glacier loss rate was
0.011 Æ 0.001 km2
a–1
. The form of WG remained almost
the same during this period. In 1979 its terminus descended
to 210 m a.s.l. and terminated behind a complex of
ice-cored terminal moraines. By 2006 the snout had
retreated by $60 m, with a mean retreat rate of 2.2 m a–1
terminating at 215 m a.s.l.
DEM comparisons show surface elevation changes
between 1979 and 2006 for each glacier. Surface lowering
occurred over 95% of the surface of DD, which lowered on
average by 8.5 Æ 2.8 m. Most of the loss occurred in the
southern and western part of the dome (Fig. 6). The site that
experienced the greatest reduction in surface elevation
($35 Æ 3 m) was occupied on 25 January 2010 by a
supraglacial lake (our observation during GPR survey).
Surface lowering greater than 25 m occurred on the whole
southern part of the dome, where DD glacier descends over
rugged relief at lower elevations (300–330 m a.s.l.). The
dome glacier surface decreased by $20 Æ 3 m around its
western margin, which is located on Davies Mesa at 400–
460 m a.s.l. The upper parts of the northwestern slope of the
dome above 450 m a.s.l. experienced lesser changes in
surface elevation (<10 m). The surface area located to the
north from the highest part of the dome remained either
unaltered or increased by <10 m.
Fig. 3. Longitudinal transect W-1 (a) and cross-sectional transects W-3 (b) and W-4 (c) for WG. The position of the glacier bed along the
western margin of WG indicates that glacier ice continues to the west (b, c). The ice thickness on the eastern side of profile W-4 suggests that
the right lateral moraine delimiting the snout is also ice-cored (c). Inverted triangles indicate points of crossover of other profiles.
Fig. 4. (a) Glacier bed topography and (b) ice thickness at DD based on the 2006 DEM and GPR measurements. The investigated dome
glacier is marked in dark grey and the outlet glacier is shown in medium grey.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers908
Surface elevations of WG decreased on average by
10.1 Æ 2.8 m over the period 1979–2006. The whole glacier
experienced lowering, with the greatest reduction in surface
elevations (>25 m) in the upper snow accumulation area
(Fig. 7). On average, surface elevations of WG decreased at
a rate of 0.37 Æ 0.11 m a–1
. This value is slightly higher than
the mean surface lowering for DD (0.32 Æ 0.11 m a–1
).
GPR data obtained between 25 January and 17 February
2010 reveal the ice thickness of the two glaciers (Table 1;
Figs 4b and 5b). At DD, the maximum thickness of 83 Æ 2 m
was recorded on its highest part near the crossing of the D-2
and D-4 profiles. The second area of greater thickness was
discovered in the eastern part of the dome (Fig. 4b). The
glacier ice along the subglacial step is $50–75 m thick.
The mean thickness of WG is greater than on DD,
99.6 Æ 1.8 m compared with 32.4 Æ 1.2 m. The greatest
thickness was observed in the central part of this valley
glacier, where the glacier ice was as much as 158 Æ 2 m
thick. The ice thickness of WG decreases towards the
eastern glacier margin, which abuts a steep valley side. In
contrast with the eastern margin, greater thickness values
(63–125 m) were recorded along the western margin of WG
(Fig. 5b). The basal reflection remains well below the lateral
moraine surface, indicating that glacier ice extends beyond
the current visible glacier limit.
The volume of DD decreased from 0.23 Æ 0.03 km3
to
0.16 Æ 0.02 km3
between 1979 and 2006. The glacier lost
30.4% of its volume at a mean annual rate of 0.003 Æ 0.002
km3
a–1
. For WG, the values of the total volume changed
from 0.27 Æ 0.02 km3
to 0.24 Æ 0.01 km3
. The volume loss
was 10.6% and the mean annual change was 0.001 Æ 0.001
km3
a–1
over the period 1979–2006. The volume change
calculated for the dome was almost three times as high as for
the valley glacier.
INTERPRETATION AND DISCUSSION
Spatial variability in bed topography and surface
lowering
The strong reflection observed in the GPR data from the flatbottomed
part of DD is interpreted as the transition from
glacier ice to solid bedrock. A possible interpretation of
weaker reflectors in the eastern flank of the plateau is that
sediments rather than solid bedrock form the glacier bed in
this area. A channel-like depression incised in the southeast
subglacial margin of the plateau suggests that the bedrock
has been deepened and widened by glacial erosion of an
outlet glacier (Figs 2a and 4a). The incised terrain in the
northeast margin of the plateau (Fig. 4a) indicates that
another outlet glacier has probably at some time in the past
drained the northeastern part of DD towards the Abernethy
Flats. The timing of this outlet glacier advance is unresolved
by our data, but it might have been synchronous with a midHolocene
advance of WG (Rabassa and others, 1982) dated
by Hjort and others (1997) at $4600 14
C BP. The interpolated
flatter subglacial relief in the northern part of DD (Fig. 4a)
leads us to suggest that DD has not expanded to inundate
cirques on the northern margin of the DD mesa. The fact that
the steep cirque headwalls have sharp upper edges is also
evidence that these rock lips have not been subjected to basal
glacier erosion. This contrasts with the southeastern edge of
the DD mesa beneath the outlet flowing towards Whisky Bay.
The strong reflection beneath the upper reaches of WG
and the straight long profile of the bed suggest that the bed
there is solid rock. By contrast, in the lower part of the
glacier the reflection from the assumed glacier bed is weak,
tentatively indicating a transition from ice to subglacial till.
Thick ice at the western periphery of WG and debriscovered
ice in adjacent moraine ridges together suggest that
WG is more extensive than the present bare-ice surface
would otherwise suggest. A large ice thickness along the
western margin of the snout supports the hypothesis that a
Fig. 5. (a) Bed topography and (b) ice thickness at WG based on the
2006 DEM and GPR measurements. The investigated valley glacier
is marked in dark grey, and the maximum possible extent of debriscored
glacier area is shown by dots.
Table 3. Changes of surface area, elevation and volume of DD and WG between 1979 and 2006
1979 2006 Change Mean annual change
Davies Dome (without outlet)
Mean elevation (m a.s.l.) 377.0 Æ 2.0 368.5 Æ 0.8 –8.5 Æ 2.8 (–2.3%) –0.32 Æ 0.10 (–0.1%)
Surface area (km2
) 6.23 Æ 0.05 4.94 Æ 0.01 –1.29 Æ 0.06 (–20.7%) –0.048 Æ 0.002 (–0.8%)
Volume (km3
) 0.23 Æ 0.03 0.16 Æ 0.02 –0.07 Æ 0.05 (–30.4%) –0.003 Æ 0.002 (–1.1%)
Whisky Glacier
Mean elevation (m a.s.l.) 350.7 Æ 2.0 340.6 Æ 0.8 –10.1 Æ 2.8 (–2.9%) –0.37 Æ 0.10 (–0.1%)
Surface area (km2
) 2.69 Æ 0.02 2.40 Æ 0.01 –0.28 Æ 0.03 (–10.6%) –0.011 Æ 0.001 (–0.4%)
Volume (km3
) 0.27 Æ 0.02 0.24 Æ 0.01 –0.03 Æ 0.03 (–10.6%) –0.001 Æ 0.001 (–0.4%)
Engel and others: Area, volume and elevation changes of James Ross Island glaciers 909
broad zone of the lateral moraine is an internal debriscovered
part of WG (Chinn and Dillon, 1987).
The spatial distribution of surface lowering indicates that
the two glaciers are evolving differently. At DD, ice surface
reduced mainly at lower altitudes, while the accumulation
zone experienced little or slightly positive change. By
contrast, the whole surface area of WG experienced
lowering, with the greatest magnitude in the upper reaches
of the glacier. The local differences in surface lowering
probably result from higher air temperatures on WG and
from different flow patterns and surface mass balances of the
two glaciers. The assumption of the different flow rate is
consistent with ice velocities observed at mass-balance
measurement sites on the two glaciers. The surface velocities
on WG were nearly twice as high as on DD during the
period 2009–11 (Ny´vlt and others, 2012). As the ice flow in
the flat-bottomed upper part of DD is limited, the increased
melting in lower parts of the dome was not compensated by
transport of ice mass between the upper and lower parts of
the glacier. By contrast, WG is a well-delimited valley
glacier with high flow velocities (Ny´vlt and others, 2012)
and the spatial distribution of ice volume is controlled by
enhanced transport of ice mass. The higher total volume
change calculated for DD probably results from lesser
thickness of its margins and from the position of the dome,
which is more exposed to air masses than is WG.
Comparison to other studies in the AP region
Although the AP ice sheet, island ice caps and valley glaciers
cover $80% of the AP region (Rau and Braun, 2002),
volumetric characteristics are known only for a minority of
glaciers (e.g. Davies and others, 2011). Apart from the AP ice
sheet, data on ice thickness have been collected only on
Alexander Island (Wager, 1982), Adelaide Island (Rivera and
others, 2005), Anvers Island (Dewart, 1971), the South
Shetland Islands (e.g. Macheret and others, 2009; Navarro
and others, 2009), JRI (Rabassa and others, 1982) and
Dolleman Island (Peel and others, 1988). At JRI, until recently
the ice thickness was known only for Mount Haddington ice
cap. According to radio-echo sounding, the ice dome is
estimated to be $300 m thick (Aristarain, 1980). This is a
typical value for island ice caps along the AP as shown in
Table 4. On DD the maximum ice thickness of $83 m
is about three times thinner than Mount Haddington, which
is not unexpected given the smaller surface area of the dome
compared with the extent of the JRI ice cap (587 km2
). The
ice thickness of WG is similar to the thickness of Spartan
Glacier at Alexander Island, which is the only valley glacier
in the AP region with known thickness (Wager, 1982).
Areal changes observed on DD and WG confirm that the
glacier retreat on JRI has increased in recent decades.
According to Rau and others (2004), ice masses on JRI lost
3.9% in area during the period 1975–2002 and, of these,
land-terminating glaciers were subject to the smallest retreat
rates. By contrast, we find that the DD glacier area
decreased by 22.1% and WG by 10.6% over the period
1979–2006. These values are in agreement with retreat rates
of DD reported by Davies and others (2011) for the period
1988–2009. The difference in retreat rate of the two glaciers
may be attributed to the faster downwasting of DD. The
tidewater front of the DD outlet in Whisky Bay shows more
prominent changes than the grounded glacier tongue of
WG. During the last 30 years the surface area of DD and
WG decreased at a mean rate of 0.8% a–1
and 0.4% a–1
,
Fig. 6. Surface topography and elevation change at DD. Surface
topography in (a) 1979 and (b) 2006 and (c) surface lowering
between 1979 and 2006. The investigated dome glacier is marked
in dark grey, the outlet glacier is shown in medium grey and the
dotted line represents the 1979 ice outline (b). Isolines within the
investigated glacier area in (c) represent surface change.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers910
respectively. These values are higher than mean annual areal
changes of ice caps on Livingston Island (Calvet and others,
1999; Molina and others, 2007) and King George Island
(Simo˜es and others, 1999) over the period 1956–2000
(Table 5). By contrast, the mean annual area decrease of DD
and WG is an order of magnitude less than the mean annual
area loss of cirque glaciers on King George Island (Simo˜es
and others, 2004). Overall, the increase in glacier retreat
observed on JRI is similar to the accelerated area loss
reported for glaciers on King George Island in the period
2000–08 (Ru¨ckamp and others, 2011).
The mean elevations of DD and WG decreased by
0.3 Æ 0.1 and 0.4 Æ 0.1 m a–1
, respectively. These values
correspond well with the mean surface lowering of 0.32 m
w.e. a–1
observed over the period 1988–97 at Rothera Point
on the southwestern AP (Smith and others, 1998) and with
the mean lowering rate of 0.2 m a–1
reported from the
summit of Bellingshausen Dome on King George Island
(Ru¨ckamp and others, 2011). By contrast, our estimates are
lower than values reported from Vega Island, Livingston
Island and the eastern part of King George Island. According
to recent observations on Vega Island (Skvarca and De
Angelis, 2003; Skvarca and others, 2004), mean annual
surface lowering of Glaciar Bahı´a del Diablo ranged from
1.0 to 1.6 m a–1
over the period 1985–2003. Pritchard and
others (2009) and Ru¨ckamp and others (2011) reported a
Table 4. Comparison of the maximum ice thickness of island glaciers along the Antarctic Peninsula
Island Glacier Maximum thickness Method Source
Name Type* m
Adelaide Fuchs Ice Piedmont P 300 RES Rivera and others (2005)
Alexander Spartan Glacier V 220 RES Wager (1982)
Anvers Anvers IC 600 Gravimetry Dewart (1971)
Dolleman Dolleman IC 460 ? Peel and others (1988)
James Ross Mount Haddington IC 300 RES Rabassa and others (1982)
Davies Dome IC 83 GPR This study
Whisky Glacier V 157 GPR This study
King George King George IC 395 RES Pfender (1999)
Lange Glacier O 308 RES Macheret and Moskalevsky (1999)
Fourcade Glacier IC 79 GPR Kim and others (2010)
Bellingshausen Dome IC 120 GPR Sobiech (2009)
Livingston Livingston ice cap IC 200 RES Macheret and others (2009)
Bowles Plateau IC 500 RES Macheret and others (2009)
Johnsons Glacier TW 196 RES Macheret and others (2009)
Hurd Glacier IC 205 RES Macheret and others (2009)
Nelson Nelson IC 169 RES Ren and others (1995)
*IC: ice cap; V: valley; O: outlet; P: piedmont; TW: tidewater.
Fig. 7. Surface topography and elevation change at WG. Surface topography in (a) 1979 and (b) 2006 and (c) surface lowering between 1979
and 2006. The investigated valley glacier is marked in dark grey, and the dotted line represents the 1979 ice outline (b). Isolines within the
investigated glacier area in (c) represent surface change.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers 911
similar value of $1.5 m a–1
on the South Shetland Islands for
the periods 2003–07 and 2000–08, respectively. Ximenis
and others (1999) obtained even higher lowering rates
(1–5 m a–1
) for glaciers on Livingston Island during the
period 1995–97. Comparison of the 1979 and 2006 DEMs
revealed a mean annual volume loss of 1.1% a–1
at DD and
0.4% a–1
at WG. The value for WG corresponds with mean
annual glacier volume loss (0.2–0.3% a–1
) reported for the
South Shetland Islands for the period 1956–2000 (Molina
and others, 2007). The value obtained at DD is three to four
times higher than the mean annual volume loss of Hurd
(0.3% a–1
) and Johnsons Glaciers (0.2% a–1
) according to
volume change described by Molina and others (2007).
A comparison of our spatial data with earlier glaciological
observations suggests that DD and WG were subject
to greater retreat, surface lowering and ice volume loss than
other island ice caps in the AP region. If the recent rate of
volume loss continues, DD and WG could disappear within
62 Æ 52 and 227 Æ 220 years, respectively. This approximated
timing corresponds well with the value based on
mean rate of areal loss for WG (228 Æ 22 years to extinction),
but the volumetric change of DD is more rapid than its
areal change (104 Æ 5 years). However, we assume that area
rather than volume has the largest control on the future
development of both glaciers. The estimated melting time of
WG based on both areal and volumetric changes is
consistent with the estimated future evolution of Bellingshausen
Dome on King George Island, which is predicted to
completely disappear after 285 years if the present warming
trend persists (Ru¨ckamp and others, 2011). A total disappearance
of cirque glaciers will probably occur much faster,
as was shown by Simo˜es and others (2004) for Ferguson and
Flagstaff Glaciers on King George Island.
SUMMARY AND CONCLUDING REMARKS
GPR data acquired in 2010 indicate that the subglacial
topography of an ice dome and of a valley glacier in the
northern part of JRI differ significantly. A near-flat plateau
forms the glacier bed of the upper accumulation area of DD,
whereas a steep plateau edge and dissected topography
underlie the lower elevated eastern part of the dome. In
contrast, almost uniform glacier bed topography exists
beneath WG. The maximum ice thicknesses as measured
by GPR at DD and WG are 83 Æ 2 and 158 Æ 2 m,
respectively. The average surface elevation of the dome
decreased by 8.5 Æ 2.8 m between 1979 and 2006, whereas
the valley glacier experienced average surface lowering of
10.1 Æ 2.8 m. DD lost 22.1% of its area between 1979 and
2006, decreasing from 8.33 Æ 0.06 km2
to 6.49 Æ 0.01 km2
.
A dome area of the glacier decreased from 6.23 Æ 0.05 km2
to 4.94 Æ 0.01 km2
(–20.7%) and its volume changed from
0.26 Æ 0.03 km3
to 0.16 Æ 0.02 km3
(–30.4%). The surface
area of WG decreased from 2.69 Æ 0.02 km2
to
2.40 Æ 0.01 km2
(–10.6%), resulting in the glacier mass loss
from 0.27 Æ 0.02 km3
to 0.24 Æ 0.01 km3
(–10.6%). Between
1979 and 2006, glacial area loss ranged from an average of
0.048 Æ 0.002 km2
a–1
(DD) to 0.011 Æ 0.001 km2
a–1
(WG).
Average volume losses of DD and WG were 0.003 Æ 0.001
and 0.001 Æ 0.001 km3
a–1
, respectively. The observed
changes are higher than areal and volumetric losses reported
from the AP region between 628 and 688 S. If the recent
trend in air temperature continues, DD and WG may
disappear in 104 Æ 5 and 228 Æ 22 years, respectively. However,
model projections reported by Carril and others (2005)
show that an increase in temperature trends is expected over
the 21st century.
ACKNOWLEDGEMENTS
This work was funded by the Czech Science Foundation
(project No. 205/09/1876) and by the Ministry of Environments
of the Czech Republic (VaV SP II 1a9/23/07).
Topographic data were provided by the Czech Geological
Survey. The study was undertaken in January and February
2010 during a fieldwork stay at the Johann Gregor Mendel
Base. We thank P. Va´czi, J. Pobor˘il and P. Pa´lka for field and
logistic support. Constructive comments and language
editing by Jonathan L. Carrivick were highly appreciated.
Helpful comments and suggestions from the scientific editor,
David M. Rippin, and two anonymous reviewers are also
acknowledged.
Table 5. Mean annual area changes of island glaciers along the northern part of the Antarctic Peninsula
Island Glacier Period Mean area change Source
Name Type* km2
a–1
%
James Ross 39 glaciers 1975–88 1.84 0.1 Skvarca and others (1995)
1988–93 2.08 0.1
58 glaciers 1988–2001 3.79 0.2 Rau and others (2004)
Davies Dome IC 1979–2006 0.048 0.8 This study
Whisky Glacier V 1979–2006 0.011 0.4 This study
Livingston Hurd Glacier IC 1956–2000 0.0116 0.2 Molina and others (2007)
Johnsons Glacier TW 1956–2000 0.0002 0.004 Molina and others (2007)
Livingston ice cap IC 1956–96 0.7902 0.1 Calvet and others (1999)
King George King George Island
ice field
IF 1956–95 2.28 0.2 Simo˜es and others (1999)
2000–08 2.56 0.2 Ru¨ckamp and others (2011)
Babylon Glacier H 1979–2000 0.0046 2.7 Simo˜es and others (2004)
Ferguson Glacier C 1979–2001 0.0078 4.0 Simo˜es and others (2004)
Flagstaff Glacier C 1979–2002 0.0053 3.1 Simo˜es and others (2004)
Lange Glacier O 1956–2000 0.0455 0.2 Barboza and others (2004)
Noble Glacier V 1979–2003 0.0049 2.1 Simo˜es and others (2004)
Stenhouse Glacier IC 1956–2000 0.0100 0.1 Simo˜es and others (2004)
*C: cirque; H: hanging; IC: ice cap; IF: ice field; V: valley; O: outlet; TW: tidewater.
Engel and others: Area, volume and elevation changes of James Ross Island glaciers912
REFERENCES
Aristarain AJ (1980) E´tude glaciologique de la calotte polaire de l’ıˆle
James Ross (Pe´ninsule Antarctique). Universite´ Grenoble,
Grenoble (Centre National de la Recherche Scientifique
Publication 322)
Barboza HHC, de Bortoli A´ L, Simo˜es JC, da Cunha RD and Braun
M (2004) Bidimensional numerical simulation of the Lange
Glacier, King George Island, Antarctica: preliminary results.
Pesqui. Anta´rt. Brasil., 4, 67–76
British Antarctic Survey (2010) Northern Antarctic Peninsula, 1:250
000 scale. Map No. BAS/UKAH, Sheet 3. British Antarctic
Survey, Cambridge
Calvet J, Garcı´a-Selle´s D and Corbera J (1999) Fluctuaciones de la
extensio´n del casquete glacial de la Isla Livingston (Shetland
del Sur) desde 1956 hasta 1996. Acta Geol. Hispa´n., 34(4),
365–374
Carril AF, Mene´ndez CG and Navarra A (2005) Climate response
associated with the Southern Annular Mode in the surroundings
of Antarctic Peninsula: a multimodel ensemble
analysis. Geophys. Res. Lett., 32(16), L16713 (doi: 10.1029/
2005GL023581)
Chinn TJH and Dillon A (1987) Observations on a debris-covered
polar glacier ‘Whisky Glacier’, James Ross Island, Antarctic
Peninsula, Antarctica. J. Glaciol., 33(115), 300–310
Cook AJ, Fox AJ, Vaughan DG and Ferrigno JG (2005) Retreating
glacier fronts on the Antarctic Peninsula over the past
half-century. Science, 308(5721), 541–544 (doi: 10.1126/
science.1104235)
Czech Geological Survey (2009) James Ross Island – northern part.
Czech Geological Survey, Praha
Davies BJ, Carrivick JL, Glasser NF, Hambrey MJ and Smellie JL
(2011) A new glacier inventory for 2009 reveals spatial and
temporal variability in glacier response to atmospheric warming
in the Northern Antarctic Peninsula, 1988–2009. Cryos.
Discuss., 5(6), 3541–3595 (doi: 10.5194/tcd-5-3541-2011)
Dewart G (1971) Gravimeter observations on Anvers Island and
vicinity. In Crary AP ed. Antarctic snow and ice studies II.
American Geophysical Union, Washington, DC, 179–190
(Antarctic Research Series 16)
Fountain AG, Lewis KJ and Doran PT (1999) Spatial climatic
variation and its control on glacier equilibrium line altitude in
Taylor Valley, Antarctica. Global Planet. Change, 22(1–4), 1–10
Garmin International (2009) GPSMAP 60CSx with sensors and
maps: owner’s manual. Garmin International Inc., Olathe, KS.
http://aspdf.com/ebook/gpsmap-60csx-with-sensors-and-mapsowner’s-manual-pdf.html
(accessed 27 April 2012)
Glasser NF, Scambos TA, Bohlander J, Truffer M, Pettit EC and
Davies BJ (2011) From ice-shelf tributary to tidewater glacier:
continued rapid recession, acceleration and thinning of Ro¨hss
Glacier following the 1995 collapse of the Prince Gustav Ice
Shelf, Antarctic Peninsula. J. Glaciol., 57(203), 397–406 (doi:
10.3189/002214311796905578)
Gudmundsson GH and Bauder A (1999) Towards an indirect
determination of the mass-balance distribution of glaciers using
the kinematic boundary condition. Geogr. Ann., 81A(4), 575–583
Hjort C, Ingo´lfsson O´ , Mo¨ller P and Lirio JM (1997) Holocene
glacial history and sea level changes on James Ross Island,
Antarctic Peninsula. J. Quat. Sci., 12(4), 259–273
Ingo´lfsson O´ , Hjort C, Bjo¨rck S and Smith RIL (1992) Late
Pleistocene and Holocene glacial history of James Ross Island,
Antarctic Peninsula. Boreas, 21(3), 209–222
Jol HM (2009) Ground penetrating radar theory and applications.
Elsevier Science, Amsterdam
Kim KY, Lee J, Hong MH, Hong JK and Shon H (2010) Helicopterborne
and ground-towed radar surveys of the Fourcade Glacier
on King George Island, Antarctica. Expl. Geophys., 41(1), 51–60
(doi: 10.1071/EG09052)
King JC (1994) Recent climate variability in the vicinity of the
Antarctic Peninsula. Int. J. Climatol., 14(4), 357–369
King JC, Turner J, Marshall GJ, Connolley WM and Lachlan-Cope
TA (2003) Antarctic Peninsula climate variability and its causes
as revealed by analysis of instrumental records. In Domack EW,
Burnett A, Leventer A, Conley P, Kirby M and Bindschadler R
eds. Antarctic Peninsula climate variability: a historical and
paleoenvironmental perspective. American Geophysical Union,
Washington, DC, 17–30 (Antarctic Research Series 79)
Knap WH, Oerlemans J and Cade´e M (1996) Climate sensitivity of
the ice cap of King George Island, South Shetland Islands,
Antarctica. Ann. Glaciol., 23, 154–159
Kovacs A, Gow AJ and Morey RM (1995) The in-situ dielectric
constant of polar firn revisited. Cold Reg. Sci. Technol., 23(3),
245–256
La´ska K, Budı´k L, Budikova´ M and Prosˇek P (2011) Method of
estimating solar UV radiation in high-latitude locations based
on satellite ozone retrieval with an improved algorithm. Int.
J. Remote Sens., 32(11), 3165–3177 (doi: 10.1080/01431161.
2010.541513)
La´ska K, Ny´vlt D, Engel Z and Budı´k L (2012) Seasonal variation of
meteorological variables and recent surface ablation/accumulation
rates on Davies Dome and Whisky Glacier, James Ross
Island, Antarctica. Geophys. Res. Abstr., 14 (EGU2012-5545)
Macheret YuYa and Moskalevsky MYu (1999) Study of Lange
Glacier on King George Island, Antarctica. Ann. Glaciol., 29,
202–206 (doi: 10.3189/172756499781820941)
Macheret YuYa and 6 others (2009) Ice thickness, internal structure
and subglacial topography of Bowles Plateau ice cap and the
main ice divides of Livingston Island, Antarctica, by groundbased
radio-echo sounding. Ann. Glaciol., 50(51), 49–56 (doi:
10.3189/172756409789097478)
MALA˚ GeoScience (2005) Ramac GPR. Hardware manual. MALA˚
GeoScience, Mala˚
Marshall GJ (2002) Analysis of recent circulation and thermal
advection change in the northern Antarctic Peninsula. Int. J.
Climatol., 22(12), 1557–1567 (doi: 10.1002/joc.814)
Marshall GJ, Orr A and Van Lipzig NPM (2006) The impact of a
changing Southern Hemisphere annular mode on Antarctic
Peninsula summer temperatures. J. Climate, 19(20), 5388–5404
Meixner P (2009) Mapping in Antarctica. GEODIS News 8, 8–9
http://www.geodis.cz/uploads/dokumenty/pdf_casopis/
GEODIS_NEWS_english_2009.pdf (accessed 27 April 2012)
Molina C, Navarro FJ, Calver J, Garcı´a-Selle´s D and Lapazaran JJ
(2007) Hurd Peninsula glaciers, Livingston Island, Antarctica, as
indicators of regional warming: ice-volume changes during the
period 1956–2000. Ann. Glaciol., 46, 43–49 (doi: 10.3189/
172756407782871765)
Narod BB and Clarke GKC (1994) Miniature high-power impulse
transmitter for radio-echo sounding. J. Glaciol., 40(134),
190–194
Navarro FJ and Eisen O (2010) Ground-penetrating radar in
glaciological applications. In Pellikka P and Reese WG eds.
Remote sensing of glaciers: techniques for topographic, spatial
and thematic mapping of glaciers. Taylor & Francis, London,
195–229
Navarro FJ and 6 others (2009) Radioglaciological studies on Hurd
Peninsula glaciers, Livingston Island, Antarctica. Ann. Glaciol.,
50(51), 17–24 (doi: 10.3189/172756409789097603)
Ny´vlt D, Kopacˇkova´ V, La´ska K and Engel Z (2010) Recent changes
detected on two glaciers at the northern part of James Ross
Island, Antarctica. Geophys. Res. Abstr., 12 (EGU2010-8102)
Ny´vlt D and 6 others (2012) Glacial history of James Ross Island
since the Late Glacial. In Blahu˘t J, Klimesˇ J, Ste˘panc˘ikova´ P and
Hartvich F eds Geomofologicky´ sbornik 10. U´ stav struktury a
mechaniky hornin, Praha, 35–36
Paterson WSB (1994) The physics of glaciers, 3rd edn. Elsevier,
Oxford
Peel DA, Mulvaney R and Davison BM (1988) Stable-isotope/airtemperature
relationships in ice cores from Dolleman Island and
the Palmer Land plateau, Antarctic Peninsula. Ann. Glaciol., 10,
130–136
Engel and others: Area, volume and elevation changes of James Ross Island glaciers 913
Pellikka P and Rees WG (2009) Remote sensing of glaciers:
techniques for topographic, spatial and thematic mapping of
glaciers. Taylor and Francis, London
Pfender M (1999) Topographie und Glazialhydrologie von
King George Island, Antarktis. (Master’s thesis, Westfa¨lische
Wilhelms-Universita¨t Mu¨nster)
Plewes LA and Hubbard B (2001) A review of the use of radio-echo
sounding in glaciology. Progr. Phys. Geogr., 25(2), 203–236 (doi:
10.1177/030913330102500203)
Pritchard HD, Arthern RJ, Vaughan DG and Edwards LA (2009)
Extensive dynamic thinning on the margins of the Greenland
and Antarctic ice sheets. Nature, 461(7266), 971–975 (doi:
10.1038/nature08471)
Rabassa J (1983) Stratigraphy of the glacigenic deposits in northern
James Ross Island, Antarctic Peninsula. In Evenson EB, Schlu¨chter
C and Rabassa J eds. Tills and related deposits: genesis/petrology/
application/stratigraphy. A.A. Balkema, Rotterdam, 329–340
Rabassa J, Skvarca P, Bertani L and Mazzoni E (1982) Glacier
inventory of James Ross and Vega Islands, Antarctic Peninsula.
Ann. Glaciol., 3, 260–264
Rau F and Braun M (2002) The regional distribution of the dry-snow
zone on the Antarctic Peninsula north of 708 S. Ann. Glaciol.,
34, 95–100 (doi: 10.3189/172756402781817914)
Rau F and 8 others (2004) Variations of glacier frontal positions on
the northern Antarctic Peninsula. Ann. Glaciol., 39, 525–530
(doi: 10.3189/172756404781814212)
Ren J and 8 others (1995) Glaciological studies on Nelson Island,
South Shetland Islands, Antarctica. J. Glaciol., 41(138), 408–412
Rignot E and 6 others (2008) Recent Antarctic ice mass loss from
radar interferometry and regional climate modelling. Nature
Geosci., 1(2), 106–110 (doi: 10.1038/ngeo102)
Rippin D, Vaughan DG and Corr HFJ (2011a) The basal roughness
of Pine Island Glacier, West Antarctica. J. Glaciol., 57(201),
67–76 (doi: 10.3189/002214311795306574)
Rippin DM, Carrivick JL and Williams C (2011b) Evidence towards
a thermal lag in the response of Ka˚rsaglacia¨ren, northern
Sweden, to climate change. J. Glaciol., 57(205), 895–903 (doi:
10.3189/002214311798043672)
Rivera A and 7 others (2005) Glacier wastage on southern Adelaide
Island, Antarctica, and its impact on snow runway operations.
Ann. Glaciol., 41, 57–62 (doi: 10.3189/172756405781813401)
Ru¨ckamp M, Braun M, Suckro S and Blindow N (2011) Observed
glacial changes on the King George Island ice cap, Antarctica, in
the last decade. Global Planet. Change, 79(1–2), 99–109 (doi:
10.1016/j.gloplacha.2011.06.009)
Sandmeier KJ (2008) REFLEXW: processing program for seismic,
acoustic and electromagnetic reflection, refraction and transmission
data, Version 5.0. Sandmeier Scientific Software, Karlsruhe
Simo˜es JC, Bremer UF, Aquino FE and Ferron FA (1999) Morphology
and variations of glacial drainage basins in the King
George Island ice field, Antarctica. Ann. Glaciol., 29, 220–224
(doi: 10.3189/172756499781821085)
Simo˜es JC, Dani N, Bremer UF, Aquino FE and Arigony-Neto J
(2004) Small cirque glaciers retreat on Keller Peninsula,
Admiralty Bay, King George Island, Antarctica. Pesqui. Anta´rt.
Brasil., 4, 49–56
Skvarca P and De Angelis H (2003) Impact assessment of regional
climate warming on glaciers and ice shelves of the northeastern
Antarctic Peninsula. In Domack EW, Burnett A, Leventer A,
Conley P, Kirby M and Bindschadler R eds. Antarctic Peninsula
climate variability: a historical and paleoenvironmental perspective.
American Geophysical Union, Washington, DC, 69–
78 (Antarctic Research Series 79)
Skvarca P, Rott H and Nagler T (1995) Satellite imagery, a baseline
for glacier variation study on James Ross Island, Antarctica. Ann.
Glaciol., 21, 291–296
Skvarca P, Rack W, Rott H and Dona´ngelo T (1998) Evidence of
recent climatic warming on the eastern Antarctic Peninsula.
Ann. Glaciol., 27, 628–632
Skvarca P, De Angelis H and Ermolin E (2004) Mass balance
of ‘Glaciar Bahı´a del Diablo’, Vega Island, Antarctic
Peninsula. Ann. Glaciol., 39, 209–213 (doi: 10.3189/
172756404781814672)
Smith AM, Vaughan DG, Doake CSM and Johnson AC (1998)
Surface lowering of the ice ramp at Rothera Point, Antarctic
Peninsula, in response to regional climate change. Ann.
Glaciol., 27, 113–118
Sobiech J (2009) Geometry and glacial hydrology of Bellingshausen
Dome, King George Island, Antarctica: results from GPRmeasurement.
(Diploma thesis, Westfa¨lische Wilhelms-Universita¨t
Mu¨nster)
Turner J and 8 others (2005) Antarctic climate change during the
last 50 years. Int. J. Climatol., 25, 279–294
Vaughan DG, Marshall GJ, Connolley WM, King JC and Mulvaney
R (2001) Climate change: devil in the detail. Science,
293(5536), 1777–1779 (doi: 10.1126/science.1065116)
Vaughan DG and 8 others (2003) Recent rapid regional climate
warming on the Antarctic Peninsula. Climatic Change, 60(3),
243–274
Wager AC (1982) Mapping the depth of a valley glacier by radioecho
sounding. Br. Antarct. Surv. Bull. 51, 111–123
Ximenis L, Calvet J, Enrique J, Corbera J, de Gamboa CF and
Furdada G (1999) The measurement of ice velocity, mass
balance and thinning-rate on Johnsons Glacier, Livingston
Island, South Shetland Islands, Antarctica. Acta Geol. Hispa´n.,
34(4), 403–409
MS received 21 July 2011 and accepted in revised form 12 April 2012
Engel and others: Area, volume and elevation changes of James Ross Island glaciers914
7
Paper 6
Effect of Snow Cover on the Active-Layer Thermal Regime –
A Case Study from James Ross Island, Antarctic Peninsula
Hrbáček, F., Láska, K., Engel, Z., 2015,
Permafrost and Periglacial Processes, DOI: 10.1002/ppp.1871
Short Communication
Effect of Snow Cover on the Active-Layer Thermal Regime – A Case Study from
James Ross Island, Antarctic Peninsula
Filip Hrbáček,1
* Kamil Láska1
and Zbyněk Engel2
1
Masaryk University, Faculty of Science, Department of Geography, Brno, Czech Republic
2
Charles University in Prague, Department of Physical Geography and Geoecology, Praha, Czech Republic
ABSTRACT
The response of active-layer thickness and the ground thermal regime to climatic conditions on the Ulu Peninsula
(James Ross Island, northeastern Antarctic Peninsula) in 2011–13 is presented. The mean air temperature over this
period was –8.0°C and ground temperature at 5 cm depth varied from –6.4°C (2011–12) to –6.7°C (2012–13). The
active-layer thickness ranged between 58 cm (January 2012) and 52 cm (February 2013). Correlation analyses indicate
that air temperature affects ground temperature more significantly on snow-free days (R2
=0.82) than on snow
cover days (R2
= 0.53). Although the effect of snow cover on the daily amplitude of ground temperature was observable
to 20 cm depth, the overall influence of snow depth on ground temperature was negligible (freezing n-factor of
0.95–0.97). Copyright © 2015 John Wiley & Sons, Ltd.
KEY WORDS: active-layer; ground temperature; snow cover; air temperature; Antarctic Peninsula; active layer thickness
INTRODUCTION
The Antarctic Peninsula (AP) has experienced the largest
atmospheric warming of all regions on Earth over the last
50 years (Turner et al., 2002), with the temperature increase
accelerating downwasting of ice sheets on the AP (Vaughan,
2006) and causing the collapse of ice shelves along its eastern
coast (Cook and Vaughan, 2010). The response of regional
permafrost to this warming remains unknown and represents
one of the most important topics in climate modelling,
because numerical models suggest that permafrost may
become the dominant contributor of CO2 and CH4 into the
atmosphere in the 21st
century (Schaefer et al., 2011).
Despite the increasing number of periglacial studies focusing
on the AP in the last decade (Vieira et al., 2010; Guglielmin
et al., 2014; Bockheim et al., 2013; De Pablo et al., 2014;
Almeida et al., 2014; Goyanes et al., 2014), thermal conditions
in the active layer and its interaction with meteorological
factors are not well known. In particular, the influence of
snow on active-layer thickness (ALT) and the thermal regime
is relatively poorly understood, despite being a major modulating
factor to the atmosphere (Boike et al., 2008; Vieira
et al., 2014). In this paper, we evaluate the effect of air temperature
and snow cover on active-layer temperature in the
northern part of James Ross Island (JRI) from March 2011
to April 2013, one of the largest permafrost regions in the
northeastern AP.
REGIONAL SETTINGS
The study site (63°48′S 57°52′W) is located in the Ulu
Peninsula, northern JRI (Figure 1), approximately 100m
south of the Johann Gregor Mendel Station at 10 m asl.
Glaciers started to retreat from the Ulu Peninsula before
12.9 ka (Nývlt et al., 2014), leaving low-lying areas ice-free
at the beginning of the Holocene. At present, small glaciers
persist only on high-altitude volcanic plateaus and in valley
heads (Engel et al., 2012). Permafrost in the northern part of
JRI can approach 95m in thickness and the ALT is highly variable,
ranging from 22 to 150cm (Borzotta and Trombotto,
2004; Engel et al., 2010; Bockheim et al., 2013). The study
site is located on a Holocene marine terrace (Figure 2)
formed by beach deposits composed of gravelly sand
(Stachoň et al., 2014).
The climate of JRI is dominated by the advection of air
masses, which are strongly influenced by the position of
the AP relative to the circumpolar trough of low pressure
*Correspondence to: F. Hrbáček, Masaryk University, Faculty of
Science, Department of Geography, Kotlářská 2, 611 37 Brno, Czech
Republic. E-mail: hrbacekfilip@gmail.com
PERMAFROST AND PERIGLACIAL PROCESSES
Permafrost and Periglac. Process. (2015)
Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/ppp.1871
Copyright © 2015 John Wiley & Sons, Ltd.
Received 11 September 2014
Revised 31 July 2015
Accepted 5 August 2015
(Domack et al., 2003). A complex orography causes frequent
variation between two main advection patterns: (1)
cold and dry southerly winds blowing along the eastern
coast of the AP, and (2) westerly winds bringing relatively
warm maritime air masses across the peninsula to northern
JRI (King et al., 2003; Zvěřina et al., 2014). The mean annual
air temperature (MAAT) at Mendel Station is –6.8°C
(2006–11) and the extremes of mean daily air temperatures
vary between around 8°C in January and –30°C in
July/August (Láska et al., 2012). Mean daily temperatures
above 0°C typically occur only for 2months each summer
(December–January), with hourly maximum and minimum
values of 10°C and –5°C, respectively (Láska et al., 2011).
According to data from Esperanza, the nearest station making
long-term observations on the northern AP, the temperature
was 0.2°C colder in 2011–13 than over the reference period
of 1961–2000, with a MAAT of –5.2°C (Turner et al.,
2004). Precipitation is mostly snow and estimated to range
from 300 to 500 mm water equivalent per year (Van Lipzig
et al., 2004).
METHODS
Temperature in the active layer was measured at depths of 5,
10, 20, 30, 50 and 75 cm using Pt100/Class A platinum resistance
thermometers (EMS, Brno, Czech Republic). Air
temperature was measured 2 m above ground level using
Figure 1 Location of the study site in the northern part of James Ross Island, close to the eastern coast of the Antarctic Peninsula.
Figure 2 (A) Detailed view of the study site and (B, C) its geomorphological position on the northern coast of the Ulu Peninsula. The red arrow marks the
study site near Mendel Station. This figure is available in colour online at wileyonlinelibrary.com/journal/ppp
F. Hrbáček et al.
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
an EMS33 sensor (EMS Brno) with a Pt100/Class A platinum
resistance thermometer placed inside a solar radiation
shield. Both ground and air temperatures were measured
with an accuracy of ± 0.15°C and data were recorded at
30 min intervals with an EdgeBox V12 datalogger (EMS
Brno). We then calculated mean daily air temperatures, the
cumulative sum of mean daily air temperatures above 0°C
(the thawing-degree days – TDDa) and the cumulative
sum of mean daily air temperatures below 0°C (the
freezing-degree days – FDDa), according to Guglielmin
et al. (2008) and De Pablo et al. (2014). Incoming and
reflected shortwave radiation (used to estimate albedo) were
measured using EMS-11 (EMS Brno) and CM6B (Kipp &
Zonen, Delft, The Netherlands) pyranometers, respectively,
at 10s time intervals and stored as 30min average values.
Snow depth was recorded every 2h using an ultrasonic depth
sensor (Judd Communication, Salt Lake City, UT, USA)
with an accuracy of±1cm. All meteorological parameters
and ground temperature data were analysed during the period
from 1 March 2011 to 30 April 2013. MAAT and mean
annual ground temperature (MAGT) were also calculated
for a period of 2years from March 2011 to February 2013,
referred to in the text as the 2011–12 and 2012–13 periods.
The ground thermal regime for the period of 2011–13 was
evaluated in accordance with recent studies investigating
Maritime Antarctic (Guglielmin et al., 2008; Michel et al.,
2012; De Pablo et al., 2014), using the following parameters:
(1) mean annual and monthly ground temperatures; (2) the cumulative
sum of mean daily ground temperatures above 0°C
(the thawing-degree days – TDDg); (3) the cumulative sum
of mean daily ground temperatures below 0°C (the freezingdegree
days – FDDg); and (4) the ALT, interpolated as the
0°C isotherm depth and derived from contours of the daily
mean ground temperature interpolated using kriging algorithms
in the Surfer® software program (Golden Software,
Golden, CO, USA).
Freezing and thawing n-factors (Karunaratne and Burn, 2003)
were calculated in order to evaluate the buffering effects of the
snow layer on heat transmission between air and the ground surface
(De Pablo et al., 2014), with the effect of air temperature on
the ground analysed at 5cm depth (e.g. Zhang et al., 1997).
Snow cover duration was estimated using a combination
of ultrasonic depth sensors and the radiometric albedo.
The criteria for snow occurrence were a depth > 2 cm (from
the ultrasonic sensors) and albedo > 0.4 (which corresponds
to old snow; Warner, 2004). Snow depth records were available
for the period from 1 March 2011 to 11 June 2012,
while radiometric albedo data were available for the periods
from 20 February to 16 May 2011 and 5 March 2012 to 30
April 2013. Despite the presence of gaps in the records due
to sensor malfunctions, we were able to analyse an insulation
effect of snow cover on the ground temperature regime
at 5 cm depth using (1) correlation analysis between mean
daily air and ground temperatures during the surface
snow-free and snow cover periods, and (2) comparison of
the snow cover records with ground temperature daily amplitudes
(Zhang et al., 1997).
RESULTS
Air and Ground Temperatures
The MAAT over the whole 2 year study period was –8.0°C;
although the individual MAAT values for the 2years were
equal, the temperature range differed from 44.0°C (2011–
12) to 42.3°C (2012–13). The larger temperature extreme
in 2011–12 was documented by a lower mean temperature
of the coldest month (July 2011, –18.5°C) and a higher
mean temperature of the warmest month (December 2011,
2.0°C) compared to those for 2012–13 (–15.0°C and 0.2°
C, respectively). The maximum air temperature recorded
during the entire study period was 11.6°C on 23 February
2013; the minimum recorded was –34.1°C on 26 July 2011.
The MAGT at 5 cm depth was –6.4°C in 2011–12 and –
6.7°C in 2012–13. The mean monthly ground temperature
at 5 cm depth varied between –16.3°C (July 2011) and
6.1°C (December 2011). Minimum (–26.3°C) and maximum
(16.0°C) 5cm ground temperatures over the study
period were recorded on 1 August 2011 and 18 December
2012, respectively. The MAGT at 50 cm depth, which represents
active-layer conditions close to the permafrost table,
ranged between –6.1°C in 2011–12 and –6.0°C in 2012–13.
Minimum (–16.3°C) and maximum (1.3°C) ground temperatures
at 50 cm were recorded on 5 August 2011 and 26
January 2012, respectively. The MAGT at 75 cm depth,
which represents the uppermost part of the permafrost zone,
reached –5.8°C, while maximum and minimum temperatures
for the greatest depth were –1.0°C and –14.4°C,
respectively.
Thawing and Freezing Seasons
The duration of the thawing season as defined by the thermal
regime at 5cm depth differed significantly between
the two periods (Table 1). In 2011–12, the thawing season
started on 9 October 2011 and terminated on 27 March
2012, lasting for 170days. In 2012–13, the thawing season
both started and ended later (13 December 2012 and 20
April 2013, respectively) and its duration was considerably
shorter (128 days). The mean ground temperature at 5 cm
depth was lower during the longer thawing season of
2011–2012 (2.3°C) than during the shorter thawing season
of 2012–13 (4.3°C). TDDg calculated for the thawing season,
however, was much higher in 2011–12 (496.1°C day)
than in 2012–13 (358.3°C day). The active layer thawed
slowly in 2011–12, reaching its maximum on 26 January
2012 (58 cm). In contrast, active-layer thaw was more rapid
in 2012–13, reaching its maximum on 13 February 2013
(52 cm).
The freezing season at 5 cm depth lasted for 203days in
2011 and 259days in 2012. The mean ground temperature
at 5 cm varied from –13.4°C in the freezing season of
2011 to –10.6°C in that of 2012. Despite the lower mean
temperature recorded in 2011, a small difference in total
FDDg was observed between the two freezing seasons, at
Active Layer Monitoring on James Ross Island
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
–2735.5°C day in 2011 and –2748.5°C day in 2012. Although
positive air temperatures were recorded on several
days during both the freezing seasons (Table 1), there were
no signs of ground thaw at 5 cm depth.
Effect of Snow Cover on Ground Temperature
The temporal distribution and duration of snow cover on
JRI varied significantly over the study period (Figure 3).
In 2011–12, snow covered the study site for 167days,
with the period of continuous snow cover lasting from
22 March to 20 September (67 days). The period of maximum
snow depth (21 to 34 cm) persisted from 29 June to
23 July. In 2012–13, continuous snow cover occurred
between 8 May and 12 December. The maximum snow
depth of 15 cm was observed on 8 May, with relatively
thick snow cover (up to 10 cm) lasting until at least 6 June.
The continuous period of fresh snow cover was registered
between 21 June and 10 December, based on an
albedo > 0.80 (Warner, 2004).
Figure 4 shows the relationship between air temperature
and ground temperature at 5cm depth for all days
(Figure 4A, D), snow-free days (Figure 4B, E) and snow
cover days (Figure 4C, F) during 2011–12 (Figure 4A–C)
and 2012–13 (Figure 4D–F). The correlation patterns reveal
a significant relationship between the air and ground temperatures
(coefficient of determination R2
= 0.80) in 2011–12.
Moreover, a closer correlation between these temperatures
was found for snow-free days (R2
=0.82) than for snow
cover days (R2
= 0.53). In contrast, a less significant relationship
(R2
=0.67) between air and ground temperatures
at 5cm depth, as well as a very small difference between
snow-free days (R2
= 0.56) and snow cover days
(R2
= 0.52), was observed in 2012–13.
The effect of snow cover on the active-layer thermal
regime was also indicated by a reduction in the daily amplitude
of ground temperature. Although the influence of snow
cover on amplitude values was detected to a depth of 20 cm,
the most significant temperature changes were recorded at
5 cm depth (Figure 5). Mean daily ground temperature
amplitudes during snow-free days in thawing seasons
ranged between 5.8°C at 5 cm and 1.5°C at 20 cm. The maximum
daily ground temperature amplitude was recorded
between 6 and 9 March 2012, when values ranged from
15°C at 5cm to 5°C at 20 cm. The effects of snow cover
on daily ground temperature amplitudes were more significant
during freezing seasons, with the longer duration of
snow cover in 2012 resulting in lower amplitude values
(1.9°C at 5cm and 0.9°C at 20 cm) than those recorded in
2011 (3.7°C and 1.2°C). Diurnal amplitudes of ground temperature
rarely decreased to 0.1°C during freezing seasons.
The longest period of very low diurnal amplitudes at depths
from 5 to 20 cm (0.1 to 0.4°C) was observed between 1 and
11 November 2012.
The overall influence of snow on the ground thermal
regime at 5 cm depth during the freezing seasons can be
seen in the obtained values of the freezing n-factors (Table 1;
Figure 6). Total freezing n-factors varied between 0.97
(2011) and 0.95 (2012), with values ranging from 0.85 to
0.95 on snow cover days. Differences in freezing n-factor
development were observed between the respective freezing
seasons of 2011 and 2012. The slower increase in the n-factor
during the period from the end of March 2011 to the end
of May 2011 indicates a more significant effect of snow
cover during this early winter than in 2012. The observed
freezing n-factor regime indicates periods with thicker snow
cover, which caused a slight decrease from values >0.95 to
approx. 0.90 in both study years.
Table 1 Quantitative characteristics of freezing and thawing seasons at Mendel Station during the period 2011–13.
Freezing season Thawing season
2011 2012 2011–12 2012–13
Period 19/3/2011 28/3/2012 9/10/2011 13/12/2012
8/10/2011 12/12/2012 27/3/2012 20/4/2013
Duration (days) 203 259 170 128
Mean air temperature (°C) À13.7 À11.0 À1.0 À0.5
Mean GT5 cm (°C) À13.4 À10.6 2.3 4.3
Min air temperature (°C) À34.1 À30.7 À17.2 À14.6
Max air temperature (°C) 5.7 7.1 9.9 11.6
Min GT5 cm (°C) À26.0 À23.5 À12.3 À11.4
Max GT5 cm (°C) À0.4 À0.3 15.3 16.0
Min GT75 cm (°C) À14.4 À12.0 À9.0 À3.3
Max GT75 cm (°C) À1.2 À1.0 À1.0 À1.1
FDDa (°C day) À2814.1 À2878.9 À330.4 À195.8
FDDg (°C day) À2735.5 À2748.5 À100.6 À77.6
n-factor 0.97 0.95 3.03 2.27
TDDa (°C day) 22.0 28.9 163.8 157.7
TDDg (°C day) 0.0 0.0 496.1 358.3
GT5 cm = Ground temperature at 5 cm depth. See text for other abbreviations.
F. Hrbáček et al.
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
DISCUSSION
Active-Layer Conditions
The annual variability of air and ground temperatures at Mendel
Station, compared to similar data from the South Shetland
Islands and South Orkney Islands, suggests that considerably
colder climatic conditions occur on JRI (see Table 2). This
is also indicated by the significantly lower FDDg values
(-2735.5 to –2748.5°C day) calculated for Mendel Station
compared to those of –500 and –900 to –1200°C day reported
from the South Shetland Islands and South Orkney
Islands, respectively (Michel et al., 2012; Guglielmin
et al., 2012; Almeida et al., 2014; De Pablo et al., 2014).
Although winters are generally much colder on JRI than
in the northern AP, we detected that subsurface conditions
can potentially be warmer during thawing seasons on JRI
than on South Shetland or South Orkney. TDDg values calculated
for JRI (496.1°C day in 2011–12 and 358.3°C day in
2012–13) are significantly higher than those for Livingston
Island (98 to 257°C day; De Pablo et al., 2014) and Signy
Island (2cm TDDg ranged between 260.5 and 378.0°C day
for the period 2006–09; Guglielmin et al., 2012). An even
higher TDDg value of 618°Cday has been recorded at
Rothera Point, Adelaide Island (Guglielmin et al., 2014).
The active layer is generally thinner and less variable on
JRI (22 to 150 cm according to Borzotta and Trombotto,
2004; Engel et al., 2010) than on the South Orkney Islands
Figure 3 (A) Daily mean air temperature, (B) albedo and snow depth, (C) ground temperature (GT) at 5, 10, 20, 30, 50 and 75 cm depths and (D) GT isopleths
at the study site for the period between 1 March 2011 and 30 April 2013. This figure is available in colour online at wileyonlinelibrary.com/journal/ppp
Active Layer Monitoring on James Ross Island
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
Figure 4 Relationship between air and ground temperatures at 5 cm depth for the periods (A–C) 2011–12 and (D–F) 2012–13 for (A, D) all days, (B, E) under
snow-free conditions and (C, F) under snow cover. Coefficients of determination R
2
for a simple linear regression model are given in the plots.
Figure 5 Variation in ground temperature (GT) daily amplitude, snow depth and albedo for the period from 1 March 2011 to 30 April 2013.
Figure 6 Cumulative sums of n-factors and freezing-degree days for air temperature (FDDa) and ground temperature at 5 cm depth (FDDg) during the freezing
seasons of (A) 2011 and (B) 2012.
F. Hrbáček et al.
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
(80 to 220 cm according to Guglielmin et al., 2008, 2012;
Bockheim et al., 2013) and South Shetland Islands (30
to> 500cm; Vieira et al., 2010; Michel et al., 2012;
Bockheim et al., 2013; De Pablo et al., 2014; Almeida
et al., 2014; Goyanes et al., 2014). The ALT of 52 to
58 cm recorded at Mendel Station during the period 2011–
13 is significantly lower than that observed in other lowelevation
parts of JRI and the AP region over the last decade
(Borzotta and Trombotto, 2004; Engel et al., 2010;
Bockheim et al., 2013).
Effects of Air Temperature and Snow Cover on the
Ground Thermal Regime
The data obtained in northern JRI between March 2011 and
April 2013 reveal the impacts of air temperature and snow
cover on the ground thermal regime. The correlation
between air and ground temperatures indicates a close relationship
between these two variables during snow-free days,
with diurnal ground temperature changes on these days
directly reflecting air temperature variation (Figure 3A, C).
Such changes in air temperature also control the duration
of the freezing and thawing seasons, as indicated by the seasonal
differences observed between 2011–12 and 2012–13.
Mean monthly air temperatures from October to December
2011 (–5.3 to 2.0°C) were significantly higher than those
in 2012 (–8.8 to –2.7°C), resulting in an extended thawing
period in 2011–12. The colder climatic conditions at the
end of the 2012 freezing season also prolonged the presence
of snow cover (see below), which in turn prevented the
ground from thawing on days with high solar radiation.
A number of studies have suggested that the effect of air
temperature on ground temperature is highly variable in the
AP region (Cannone et al., 2006; Guglielmin et al., 2014).
Whereas a relatively weak correlation (r= 0.58 to 0.82)
between air and ground temperatures was reported from
King George Island and Signy Island (Cannone et al.,
2006) in areas covered by different types of vegetation, a
stronger relationship (r >0.90) was described in areas of
bare ground at Rothera Point (Guglielmin et al., 2014). This
difference in effect may be attributed to the variable ground
surface conditions at the various study sites. The ground
surface at the site on JRI investigated in the present study
is free of vegetation, which is, by contrast, well developed
on Signy Island, where a significant influence of vegetation
cover on ground temperatures was found (Cannone et al.,
2006). We also observed greater variability in ground temperature
at 5cm depth during winters with thinner snow
cover, in contrast to the South Shetland Islands where a
snow depth greater than 40cm had a significant effect on
the active-layer thermal regime (De Pablo et al., 2014).
The minor effect of a thin snow cover on the ground thermal
regime during the freezing season at the study site is
also indicated by the high freezing n-factor values (0.95
to 0.97). Lower n-factor values (0.2 to 0.7) have been
reported from Maritime Antarctica (De Pablo et al., 2014;
Almeida et al., 2014), reflecting the deeper winter snow
cover at sites on Livingston Island (De Pablo et al., 2014)
and King George Island (Almeida et al., 2014).
Our conclusions regarding the effect of snow cover on
active-layer conditions, however, should be treated with caution,
as snow cover depth and distribution are highly variable
and our field data cover only a brief time period. The irregular
pattern of snow cover results from a combination of local
climatic conditions, wind structure and orographic effects on
precipitation (e.g. Turner et al., 2002). Moreover, Van Lipzig
et al. (2004) confirmed an orographic effect of the AP on the
spatial distribution and irregular accumulation of snow along
the AP coast. On the Ulu Peninsula, the relationship between
snow drift and prevailing wind direction was reported by
Table 2 Climatic conditions and active-layer depths at reported sites in the northern Antarctic Peninsula.
Region Study area Elevation
(m asl)
Period MAAT
(°C)
MAGT
(°C)
Ground
temperature
measurement depth
(cm)
ALT
(cm)
Reference
South Orkney Islands Signy
Island
80 1/06–12/09 À3.7 À2.4 2 112–158 Guglielmin
et al. (2012)À2.0a
2 101
South Shetland Islands King
George
Island
65–70 2/08–1/09 NA À0.8 8.5 89–92 Michel
et al. (2012)
86 4/09–1/11 À3.0 À1.4 10 120–147 Almeida
et al. (2014)
Livingston
Island
105 2/09–2/13 À2.5 À0.8 2.5 99–105 De Pablo
et al. (2014)
Deception
Island
130 1/10–12/10 À2.3 À1.5 2 46–67 Goyanes
et al. (2014)
Trinity Peninsula James
Ross
Island
10 3/11–2/13 À8.0 À6.6 5 52–58 This study
a
Data for site covered by vegetation. NA = Data not available. See text for other abbreviations.
Active Layer Monitoring on James Ross Island
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
Zvěřina et al. (2014), suggesting a redistribution of snow
deposits in areas of low elevation in the Abernethy Flats
(6km south of Mendel Station). Clearly, more work is needed
regarding the snow cover distribution across JRI, using both
ground-based and remote sensing observations.
CONCLUSIONS
Based on 2 years of meteorological observations and activelayer
ground temperature monitoring at Mendel Station on
JRI, we draw the following conclusions:
1. ALTs were observed between the end of January and the
middle of February. Thicknesses of 58 and 52 cm measured
in 2012 and 2013, respectively, are significantly
lower than the ALT reported from low-elevation sites
on the South Shetland Islands and South Orkney Islands.
2. Correlation analysis indicates a significant effect of air
temperature on the ground thermal regime. This effect is
especially apparent under snow-free conditions and during
the advection of relatively warm air masses that cause
large day-to-day changes in air temperature during winter.
3. The effect of snow cover depth on the ground thermal
regime is indicated not only by the reduction in the daily
ground temperature amplitude, but also by the total
freezing n-factor values that were higher than 0.9. The
overall influence of snow depth on ground temperature
seems therefore to be negligible regarding the high
values of the n-factor, associated with a thin and irregularly
distributed snow cover.
ACKNOWLEDGEMENTS
We thank the personnel at the Johann Gregor Mendel Station
for their hospitality and logistical support. We are also
very grateful to the members of the summer expeditions in
2011–14 for their field assistance on JRI. The work of F.
Hrbáček and K. Láska was supported by the Masaryk University
project MUNI/A/0952/2013: Analysis, evaluation
and visualisation of global environmental changes in the
landscape sphere. Finally, we would like to thank two anonymous
reviewers for their comments and suggestions which
improved the manuscript.
REFERENCES
Almeida ICC, Schaefer CEGR, Fernandes
RBA, Pereira TTC, Nieuwendam A, Pereira
AB. 2014. Active layer thermal regime at
different vegetation covers at Lion Rumps,
King George Island, Maritime Antarctica.
Geomorphology 225: 36–46.
Bockheim J, Vieira G, Ramos M, Lopez-Martinez
J, Serrano E, Guglielmin M, Wilhelm
K, Nieuwendam A. 2013. Climate warming
and permafrost dynamics in the Antarctic
Peninsula region. Global and Planetary
Change 100: 215–223.
Boike J, Hagedorn B, Roth K. 2008. Heat and
water transfer processes in permafrost
affected soils: A review of field-and
modeling-based studies for the Arctic and
Antarctic. In Proceeding of the Ninth International
Conference on Permafrost, Douglas
LK, Kenneth MH (eds). University of
Alaska: Fairbanks; Vol. 1:149–154.
Borzotta E, Trombotto D. 2004. Correlation
between frozen ground thickness measured
in Antarctica and permafrost thickness estimated
on the basis of the heat flow obtained
from magnetotelluric soundings. Cold
Region Science and Technology 40: 81–96.
Cannone N, Evans JCE, Strachan R,
Guglielmin M. 2006. Interactions between
climate, vegetation and the active layer at
two Maritime Antarctica sites. Antarctic
Science 18: 323–333.
Cook AJ, Vaughan DG. 2010. Overview of areal
changes of the ice shelves on the Antarctic
Peninsula over the past 50years. The
Cryosphere 4: 77–98.
De Pablo MA, Ramos M, Molina A. 2014.
Thermal characterization of the active layer
at the Limnopolar Lake CALM-S site on
Byers Peninsula (Livingston Island), Antarctica.
Solid Earth 5: 721–739.
Domack EW, Burnett A, Leventer A. 2003. Environmental
setting of the Antarctic Peninsula.
In Antarctic Peninsula Climate Variability:
Historical and Paleoenvironmental Perspectives,
Domack E, Leventer A, Burnett A,
Bindshadler R, Convey P, Kirby M (eds).
American Geophysical Union: Washington,
DC; Vol. 79: 1–13.
Engel Z, Láska K, Franta T, Máčka Z,
Marvánek O. 2010. Recent changes of permafrost
active layer on the James Ross Island,
Maritime Antarctic. In Abstracts from the
Third European Conference on Permafrost,
Mertes JR, Christiansen HH, Etzelmüller B
(eds). UNIS: Longyearbyen; 129.
Engel Z, Nývlt D, Láska K. 2012. Ice thickness,
bed topography and glacier volume changes
on James Ross Island, Antarctic Peninsula.
Journal of Glaciology 58: 904–914.
Goyanes G, Vieira G, Caselli A, Cardoso M,
Marmy A, Bernardo I, Hauck C. 2014. Geothermal
anomalies, permafrost and geomorphological
dynamics (Deception Island,
Antarctica). Geomorphology 225: 57–68.
Guglielmin M, Evans CJE, Cannone N. 2008.
Active layer thermal regime under different
vegetation conditions in permafrost areas. A
case study at Signy Island (Maritime Antarctica).
Geoderma 144: 73–85.
Guglielmin M, Worland MR, Cannone N. 2012.
Spatial and temporal variability of ground surface
temperature and active layer thickness at
the margin of maritime Antarctica, Signy
Island. Geomorphology 155–156: 20–33.
Guglielmin M, Worland MR, Baio F, Convey P.
2014. Permafrost and snow monitoring at
Rothera Point (Adelaide Island, Maritime
Antarctica): Implications for rock weathering
in cryotic conditions. Geomorphology 225:
47–56.
Karunaratne KC, Burn CR. 2003. Freezing
n-factors in discontinuous permafrost terrain,
Takhini River, Yukon Territory, Canada. In
Proceedings of the 8th International Conference
on Permafrost, Phillips M, Springman
SM, Arenson LU (eds). University of Alaska:
Fairbanks; 519–524.
King JC, Turner J, Marshall GJ, Connelly WM,
Lachlan-Cope TA. 2003. Antarctic Peninsula
climate variability and its causes as revealed
by analysis of instrumental records. In Antarctic
Peninsula Climate Variability: Historical
and Palaeoenvironmental Perspectives,
Domack EW, Leventer A, Burnett A, Bindschadler
R, Convey P, Kirby M (eds). American
Geophysical Union: Washington DC;
Vol. 79: 17–30.
Láska K, Barták M, Hájek J, Prošek P,
Bohuslavová O. 2011. Climatic and ecological
characteristics of deglaciated area of
James Ross Island, Antarctica, with a special
respect to vegetation cover. Czech
Polar Reports 1: 49–62.
Láska K, Nývlt D, Engel Z, Budík L. 2012. Seasonal
variation of meteorological variables
F. Hrbáček et al.
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
and recent surface ablation / accumulation
rates on Davies Dome and Whisky Glacier,
James Ross Island, Antarctica. Geophysical
Research Abstracts 14: EGU2012–5545.
Michel RFM, Schaefer CEGR, Poelking EL,
Simas FNB, Filho EIF, Bockheim JG.
2012. Active layer temperature in two
Cryosols from King George Island, Maritime
Antarctica. Geomorphology 155-156:
12–19.
Nývlt D, Braucher R, Engel Z, Mlčoch B, ASTER
team. 2014. Timing of the Northern
Prince Gustav Ice Stream retreat and the deglaciation
of northern James Ross Island,
Antarctic Peninsula during the last glacial–
interglacial transition. Quaternary Research
82: 441–449.
Schaefer K, Zhang T, Brujwiler L, Barrett AP.
2011. Amount and timing of permafrost
carbon release in response to climate
warming. Tellus 63: 165–180.
Stachoň Z, Russnák J, Nývlt D, Hrbáček F.
2014. Stabilization of geodetic points in
the surrounding of Johann Gregor Mendel
Station, James Ross Island, Antarctica.
Czech Polar Reports 4(1): 80–89.
Turner J, Lachlan-Cope TA, Marshall GJ,
Morris EM, Mulvaney R, Winter W. 2002.
Spatial variability of Antarctic Peninsula
net surface mass balance. Journal of Geophysical
Research 107: 4173.
Turner J, Colwell SR, Marshall GJ, LachlanCope
TA, Carleton AM, Jones PD, Lagun
V, Reid PA, Iagovkina S. 2004. The SCAR
READER project: towards a high-quality
database of mean Antarctic meteorological
observations. Journal of Climate 17:
2890–2898.
Van Lipzig NPM, King JC, Lachlan-Cope TA,
van der Broeke MR. 2004. Precipitation,
sublimation and snow drift in the Antarctic
Peninsula region from a regional atmospheric
model. Journal of Geophysical Research
109: D24106.
Vaughan DG. 2006. Recent trends in melting
conditions on the Antarctic Peninsula and
their implications for ice-sheet mass balance.
Arctic, Antarctic, and Alpine Research
38: 147–152.
Vieira G, Bockheim J, Guglielmin M, Balks
M, Abramov AA, Boelhouwers J, Cannone
N, Ganzert L, Gilichinsky DA, Goryachkin
S, López-Martínez J, Meiklejohn I, Raffi
R, Ramos M, Schaefer C, Serrano E, Simas
F, Sletten R, Wagner D. 2010. Thermal
State of permafrost and active-layer monitoring
in the Antarctic: advances during
the International Polar Year 2007–09. Permafrost
and Periglacial Processes 21:
182–197. DOI: 10.1002/ppp.685.
Vieira G, Mora C, Pina P, Schaefer C. 2014. A
proxy for snow cover and winter ground
surface cooling: mapping Usnea sp communities
using high resolution remote sensing
imagery (Maritime Antarctica). Geomorphology
225(15): 69–75.
Warner TT. 2004. Desert Meteorology. Cambridge
University Press: Cambridge, UK.
Zhang T, Osterkamp TE, Stamnes K. 1997.
Effects of climate on the active layer and
permafrost on the North Slope of Alaska,
U.S.A. Permafrost and Periglacial Processes
8: 45–67. 10.1002/(SICI)1099-1530
(199701)8:<45::AID-PPP240>3.0.CO;2-K.
Zvěřina O, Láska K, Červenka R, Kuta J,
Coufalík P, Komárek J. 2014. Analysis of
mercury and other heavy metals accumulated
in lichen Usnea antarctica from James Ross
Island, Antarctica. Environmental Monitoring
and Assessment 186: 9089–9100.
Active Layer Monitoring on James Ross Island
Copyright © 2015 John Wiley & Sons, Ltd. Permafrost and Periglac. Process., (2015)
8
Paper 7
Unusual biogenic calcite structures in two shallow lakes, James
Ross Island, Antarctica
Elster, J., Nedbalová, L., Vodrážka, R., Láska, K., Haloda, J., Komárek, J., 2016,
Biogeosciences, 13, pp. 535–549
Biogeosciences, 13, 535–549, 2016
www.biogeosciences.net/13/535/2016/
doi:10.5194/bg-13-535-2016
© Author(s) 2016. CC Attribution 3.0 License.
Unusual biogenic calcite structures in two shallow lakes,
James Ross Island, Antarctica
J. Elster1,2, L. Nedbalová2,3, R. Vodrážka4, K. Láska5, J. Haloda4, and J. Komárek1,2
1Centre for Polar Ecology, Faculty of Science, University of South Bohemia, Na Zlaté Stoce 3, 37005 ˇCeské Budˇejovice,
Czech Republic
2Institute of Botany, Academy of Sciences of the Czech Republic, Dukelská 135, 37982 Tˇreboˇn, Czech Republic
3Department of Ecology, Faculty of Science, Charles University in Prague, Albertov 6, 12843 Prague, Czech Republic
4Czech Geological Survey, Klárov 3, 11821 Prague, Czech Republic
5Department of Geography, Faculty of Science, Masaryk University, Kotláˇrská 2, 61137 Brno, Czech Republic
Correspondence to: J. Elster (jelster@prf.jcu.cz)
Received: 29 June 2015 – Published in Biogeosciences Discuss.: 21 August 2015
Revised: 11 January 2016 – Accepted: 15 January 2016 – Published: 28 January 2016
Abstract. The floors of two shallow endorheic lakes, located
on volcanic surfaces on James Ross Island, are covered
with calcareous organosedimentary structures. Their biological
and chemical composition, lake water characteristics,
and seasonal variability of the thermal regime are introduced.
The lakes are frozen down to the bottom for 8–
9 months a year and their water chemistry is characterised
by low conductivity and neutral to slightly alkaline pH.
The photosynthetic microbial mat is composed of filamentous
cyanobacteria and microalgae that are considered to be
Antarctic endemic species. The mucilaginous black biofilm
is covered by green spots formed by a green microalga and
the macroscopic structures are packed together with fine material.
Thin sections consist of rock substrate, soft biofilm,
calcite spicules and mineral grains originating from different
sources. The morphology of the spicules is typical of
calcium carbonate monocrystals having a layered structure
and specific surface texture, which reflect growth and degradation
processes. The spicules’ chemical composition and
structure correspond to pure calcite. The lakes’ age, altitude,
morphometry, geomorphological and hydrological stability,
including low sedimentation rates, together with thermal
regime predispose the existence of this community. We
hypothesise that the precipitation of calcite is connected with
the photosynthetic activity of the green microalgae that were
not recorded in any other lake in the region. This study has
shown that the unique community producing biogenic calcite
spicules is quite different to any yet described.
1 Introduction
The floors of most Antarctic lakes are covered with photosynthetic
microbial mats (Vincent and Laybourn-Parry,
2008). However, the degree of disturbance plays a key role
in the development of microbial mats. When growing in
low-disturbance habitats, interactions between benthic microbial
communities and their environments can produce
complex emergent structures. Such structures are best developed
in extreme environments, including benthic communities
of deep, perennially ice-covered Antarctic lakes,
where physical and chemical conditions and/or geographical
isolation preclude the development of larger organisms
that could otherwise disrupt organised microbial structures
(Wharton, 1994; Andersen et al., 2011). Many organosedimentary
structures that emerge in these conditions are laminated
and accrete through episodic trapping of sediments or
grains and precipitation of minerals within a growing biogenic
matrix (e.g. Arp et al., 2001; Reid et al., 2003). In
perennially ice-covered lakes, the seasonality of growth imposed
by the summer–winter light–dark conditions can induce
annual growth laminations (Hawes et al., 2001), reinforced
by calcite precipitation during growth and sediment
diagenesis (Wharton et al., 1982; Wharton, 1994; Sutherland
and Hawes, 2009). Calcite precipitation is not, however,
a prerequisite for laminated, stromatolite-like communities
(Walter, 1976; Schieber, 1999; Yamamoto et al., 2009). A diversity
of micro- to nanostructured CaCO3 associated with
extracellular polymeric substances and prokaryotes was dePublished
by Copernicus Publications on behalf of the European Geosciences Union.
536 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
scribed from the sediments of an East Antarctic lake (Lepot
et al, 2014). There is also a growing experimental evidence
that some carbonate precipitates are only produced in the
presence of organic matter (Cölfen and Antonietti, 1998;
Pedley et al., 2009).
The precipitation of calcite by expulsion (segregation) is
also a common process in nature related to the freezing of
common low ionic strength Ca2+-HCO−
3 waters. Calcite precipitation
related to water freezing was observed and described
also from various polar-alpine settings, e.g. from lake
bottoms of Dry Valleys in Antarctica (Nakai et al., 1975)
or as a result of aufeis (icing, naled) formation in northern
Canada (Clark and Lauriol, 1997). Crystalline precipitates
that form subglacially on bedrock were reported from numerous
locations (Ng and Hallet, 2002), for example finegrained
calcite powders were observed in subglacial deposits
and in aufeis formations, Svalbard (Wadham et al., 2000)
or in basal ice and subglacial clastic deposits of continental
glaciers of Switzerland (Fairchild et al., 1993). Calcite
pendants occurred beneath coarse clasts in well-drained sediments
on Svalbard (Courty et al., 1994) and calcite coatings
were found in cavities in cold-climate Pleistocene deposits
of Western Transbaikalia, Russia, and in modern surface deposits
at Seymour Island, Antarctica (Vogt and Corte, 1996).
James Ross Island belongs to a transitory zone between
the maritime and continental Antarctic regions (Øvstedal and
Lewis Smith, 2001). Air temperature records indicate progressive
warming trends from 1.5 to 3.0 ◦C over the Antarctic
Peninsula during the past 50 years (Turner et al., 2014). More
than 80 % of the island surface is covered with ice (Rabassa
et al., 1982). Only the northernmost part of the island, the Ulu
Peninsula, is significantly deglaciated and represents one of
the largest ice-free areas in the northern part of the Antarctic
Peninsula. The origin of the lakes on James Ross Island is
related to the last glaciations of the Antarctic Peninsula ice
sheet and retreat of the James Ross Island ice cap during the
late Pleistocene and Holocene (Nývlt et al., 2011; Nedbalová
et al., 2013). Interactions between volcanic landforms and
glacial geomorphology during previous glacial–interglacial
cycles, the Holocene paraglacial and periglacial processes
and relative sea level change have resulted in the complex
present-day landscape of James Ross Island (Davies et al.,
2013). All of these processes have influenced the development
of the lakes which are found on the Ulu Peninsula at
altitudes from < 20 m above sea level (a.s.l.) near the coast to
400 m a.s.l. in the mountain areas (Nedbalová et al., 2013).
During two Czech research expeditions (2008 and 2009)
to James Ross Island, lake ecosystems of the Ulu Peninsula
were studied in respect to their origin, morphometry,
physical, chemical and biological characteristics (Nedbalová
et al., 2013), together with detailed cyanobacterial and microalgal
diversity descriptions (Komárek and Elster, 2008;
Komárek et al., 2012, 2015; Kopalová et al., 2013; Škaloud
et al., 2013). As part of this study, we encountered 1–5 mmscale
calcareous organosedimentary structures on the floor
GREENII
GREEN I
Johann Gregor Mendel
Base (CZE)
lliH
yrreB
Brandy Bay
Andreassen Point
a
Solorin
Valley
a
Solorin
Valley
Lac
h
m
a
nC
r
a
g
s
LAKE 1
LAKE 2
AWS
stalFyhtenrebA
stalFyhtenrebA
63°51'
63°55'
63°49'
GR
57°46 W'57°50'57°54'
63°47 S'
63°53'
0 1 2 3 km
Coastline
River, creek
Lake
Flood plain
Glacier, snowfield
Automatic weather station (AWS)
N
80
07
60
70
80
90 W
1000 km
lennahCvatsuGecnirP
James Ross I.
Vega I.
Snow Hill I.
Seymour I.
Antar
tic
Peni
sula
c
n
64°S
Dundee I.
63°S
.IellivnioJ
56°W57°58°59°
Cape
Lachman
a
Solorin
Valley
Figure 1. Location of lakes 1 and 2 and air temperature measurements
(AWS) in the Solorina Valley. The upper inset map shows the
position of the study site and James Ross Island in the northeastern
part of Antarctic Peninsula.
of two endorheic lakes, 1 and 2, which are quite different
to any microbially mediated structures yet described from
modern environments. These shallow lakes on higher-lying
levelled surfaces originated after the deglaciation of volcanic
mesas which became ice-free some 6.5–8 ka ago (Johnson et
al., 2011) and are considered among the oldest in the region.
However, a later appearance of these lakes is also possible,
as we have no exact dates from their sediments (Nedbalová
et al., 2013).
The aim of this paper is to describe in detail the chemical
and biological composition of the organosedimentary structures
together with the limnological characteristics of the
two lakes. A hypothesis concerning the formation of calcite
spicules is also presented. The results of this study can serve
as a baseline for understanding the microbial behaviours in
forming these organosedimentary structures, which will proBiogeosciences,
13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 537
0.9 m
.0 6 m
8 m
0.
.m
04
0.2 m
20 m20 m
1.1m
.1
0
m
5
0.
m
N N
niflow
inowfl
water level 22.2.2008
water level 12.1.2009
black mats
temperature sensor
(a) (b)
Figure 2. Bathymetric parameters of lakes 1 (a) and 2 (b) together with marked lines of water level and maximum extent of the photosynthetic
microbial mat littoral belt in lake 1.
vide insight into the interpretation of fossil forms from early
Earth.
2 Materials and methods
2.1 Study site
Endorheic lake 1 (63◦54 11.7 S, 57◦46 49.9 W, altitude
65 m a.s.l.) and endorheic lake 2 (63◦53 54.6 S,
57◦46 33.8 W, altitude 40 m a.s.l.) are shallow lakes located
near Andreassen Point on the east coast of the deglaciated
Ulu Peninsula, in the northern part of James Ross Island,
NE Antarctic Peninsula (Fig. 1). They are shallow with maximum
depth of 1.1 and 0.9 m, and mean depth of 0.5 and
0.3 m. Their catchment areas are 0.340 and 0.369 km2, lake
area 4220 and 2970 m2 and water volume 2183 and 1037 m3,
respectively (Nedbalová et al., 2013). Melt waters from the
surrounding snowfields feed the lakes for a few weeks during
the austral summer. The water level in both lakes fluctuated
dramatically. Water is mainly lost through evaporation from
the ice-free water surface. During this period, intense evaporation
in both lakes is coupled with macroscopic changes
in the littoral belt. The extent of water level fluctuation was
documented for lake 1 (Fig. 2).
Climate conditions of the Ulu Peninsula are characterised
by mean annual air temperatures around −7 ◦C and mean
summer temperatures above 0 ◦C for up to 4 months (Láska
et al., 2011a). The mean global solar radiation is around
250 W m−2 in summer (December–February), with large
day-to-day variation affected by extended cyclonic activity
in the circumpolar trough and orographic effects over the
Antarctic Peninsula (Láska et al., 2011b). The bedrock is
composed of two main geological units, namely Cretaceous
back-arc basin sediments and mostly subglacial Neogene to
Quaternary volcanic rocks (Olivero et al., 1986). The terrestrial
vegetation is limited to non-vascular plants and composed
predominantly of lichen and bryophyte tundra. A large
number of lakes can be found in this area, formed by glacial
www.biogeosciences.net/13/535/2016/ Biogeosciences, 13, 535–549, 2016
538 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
(a)
(b)
(c)
Figure 3. (a) Annual variation of daily mean water temperature in lake 1 (L1 water) and annual variation of daily mean air temperature in
the Solorina Valley (SV air), (b) diurnal temperature amplitudes in lake 1 (L1 water) and diurnal air temperature amplitudes in the Solorina
Valley (SV air), respectively. (c) Daily mean global radiation at Mendel Station. All parameters were measured from February 2009 to
November 2010.
erosion and deposition, followed by glacier retreat during the
Holocene (Nedbalová et al., 2013).
2.2 Sampling procedures
Lake 1 was sampled on 22 February 2008. In 2009, lake 1
was sampled on 5 January, lake 2 on 12 January. Air temperature
at 2 m above ground was measured by an automatic
weather station (AWS) located nearby (Fig. 1). Incident
global solar radiation was monitored with a LI-200 pyranometer
(LI-COR, USA) at Mendel Station, located 11 km
northwest of the study site (Fig. 1). The LI-200 spectral response
curve covers wavelengths from 400 to 1100 nm with
absolute error typically of ±3 % under natural daylight conditions.
Global radiation was measured at a 10 s time interval
and stored as 30 min average values, while air temperature
was recorded at 1 h intervals from 1 February 2009
to 30 November 2010. In lake 1, water temperature was
monitored from 10 February 2009 to 30 November 2010 at
1 h intervals using a platinum resistance thermometer with
Minikin T data logger (EMS Brno, Czech Republic) installed
on the lake bottom.
Conductivity, pH, temperature and dissolved oxygen were
measured in situ with a portable meter (YSI 600) at the
time the lakes were ice free. Water samples were collected
from the surface layer, immediately filtered through a 200 µm
polyamide sieve to remove zooplankton and coarse particles.
Chlorophyll a was extracted from particles retained on Whatman
GF/F glass microfibre filters according to Pechar (1987).
After centrifugation, chlorophyll a was measured by a Turner
TD-700 fluorometer equipped with a non-acidification optical
kit. The remaining water was kept frozen until analysed at
the Institute of Hydrobiology (Czech Republic). The chemical
analytical methods are given in Nedbalová et al. (2013).
The stones covered by photoautotrophic mats–biofilm collected
in the field were transported to the Czech Republic in
a frozen and/or dry state, documented with stereomicroscope
(Bresser, HG 424018) and imaging fluorometer (FluorCam,
PSI) and used for (a) phytobenthos community description
and isolation of dominant species, (b) fix for thin section
Biogeosciences, 13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 539
Figure 4. Dominant species in the photoautotrophic mats. (a) Calothrix elsteri, (b) Hazenia broadyi.
analyses, (c) scanning electron and optical microscopy and
(d) determination of the structure and chemical composition
of calcium carbonate spicules.
2.3 Thin-section analyses
Thin-section analyses were made to observe both rock substrate
and inorganic particles within biofilms. Dry microbial
mat were saturated with epoxy resins in vacuum, subsequently
cut perpendicularly and saturated again with epoxy
resin. The sample was cemented to a glass slide after grinding
and polishing, and a thin section was prepared by final
sectioning, grinding and polishing to a desired thickness of
50–55 µm. Thin sections of rocks were studied in transmitted
(PPL) and polarised (XPL) light (Olympus BX-51M) and
documented in transmitted light of a Nikon SMZ-645 optical
microscope using NIS-Elements software.
2.4 Biofilm scanning electron and optical microscopy
The morphology of photoautotrophic mats and calcareous
spicules was studied using standard methods of scanning
electron microscopy (SEM) using back-scattered electrons
(BSE) (Jeol JSM-6380, Faculty of Science, Charles University)
and optical microscopy (Nikon SMZ-645 using NISElements
software). Calcareous spicules were collected directly
from the surface of biofilms. Samples studied in SEM
were completely dried for 5 months at room temperature,
then mounted on stubs with carbon paste and coated with
gold prior to photomicrographing.
2.5 Structure and chemical composition of calcium
carbonate spicules – EDS and EBSD analyses
The chemical composition of the analysed spicules was measured
by using the Link ISIS 300 system with 10 mm2 Si–
Li EDS detector on a CamScan 3200 scanning electron microscope
(Czech Geological Survey, Prague). Analyses were
performed using an accelerating voltage of 15 kV, 2 nA beam
current, 1 µm beam size and ZAF correction procedures. Natural
carbonate standards (calcite, magnesite, rhodochrosite,
siderite and smithsonite) were used for standardisation. Subsequent
structural identification was confirmed by electron
backscattered diffraction (EBSD). Identification data and
crystallographic orientation measurements were performed
on the same scanning electron microscope using an Oxford
Instruments Nordlys S EBSD detector. The thick sections
used for EBSD applications were prepared by the process
of chemo-mechanical polishing using colloidal silica suspension.
The acquired EBSD patterns were indexed within
Channel 5 EBSD software (Schmidt and Olensen, 1989) applying
calcite and aragonite crystallographic models (Effenberger
et al., 1981; Caspi et al., 2005). Orientation contrast
images were collected from a four-diode forescatter electron
detector (FSD) integrated into the Nordlys S camera. EBSD
pattern acquisition was carried out at 20 kV acceleration voltage,
3 nA beam current, 33 mm working distance and 70◦
sample tilt.
3 Results
3.1 General description of the lakes and water
chemistry
Pictures and detailed bathymetric parameters of both lakes
together with marked lines of water level and the maximum
extent of the photosynthetic microbial mat littoral belt in lake
1 are presented in Fig. 2.
The physico-chemical characteristics of the lake water for
both lakes are given in Table 1. The sampling of lake 1
(pH 7.4–7.9, saturation of oxygen 98.9 %) was performed
during cloudy days. Oxygen supersaturation (128 %) together
with a relatively high pH (8.6) was observed in lake 2
during a sunny day. Conductivity was below 100 µS cm−1 in
www.biogeosciences.net/13/535/2016/ Biogeosciences, 13, 535–549, 2016
540 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
(a) (b)
(c) (d)
(e) (f)
(g) (h)
Figure 5. Photoautotrophic mats in lakes 1 and 2. (a, b) Rapid development of the mat in January 2009 (lake 1). The two photos show the
mat at a 1-week interval; note the growth of gelatinous clusters of densely agglomerated filaments of the green alga Hazenia broadyi. (c,
d) Fully developed mats with mosaic-like structures on the surfaces of stones in the littoral zone of lake 2; (e, f) detail; (g) drying of the mat
in the littoral zone leaves a characteristic structure on the surface of stones; (h) calcium carbonate spicule in situ (arrowed).
Biogeosciences, 13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 541
Table 1. Physico-chemical characteristics and chlorophyll a concentrations
in lake water. Samples were collected from surface of
lakes. ND – not determined, ANC – acid neutralisation capacity,
PN – particulate nitrogen, DP – dissolved phosphorus, PP – particulate
phosphorus, SRP – dissolved reactive phosphorus, DOC –
dissolved organic carbon, PC – particulate carbon, ∗ – laboratory
values.
Parameter Lake 1 Lake 2
Date 22/2/2008 5/1/2009 12/1/2009
Temperature ◦C 3.5 ND 12.3
O2 mg L−1 13.1 ND 13.7
O2 saturation % 98.7 ND 128.0
pH 7.9 7.4∗ 8.6
Conductivity µS cm−1 54 48∗ 97
(25 ◦C)
ANC mmol L−1 236 246 455
Na+ mg L−1 4.7 5.9 12.5
K+ mg L−1 0.24 0.29 0.60
Ca2+ mg L−1 2.12 1.26 2.32
Mg2+ mg L−1 1.24 0.77 1.65
SO2−
4 mg L−1 1.74 1.33 2.60
Cl− mg L−1 5.3 5.1 10.6
NO3-N µg L−1 < 5 11 < 5
NO2-N µg L−1 0.6 0.2 0.1
NH4-N µg L−1 6 < 5 < 5
PN µg L−1 20 50 73
DP µg L−1 7.8 20.2 30.4
PP µg L−1 4.6 5.9 11.7
SRP µg L−1 4.0 11.6 19.3
DOC mg L−1 1.25 1.13 2.17
PC mg L−1 0.13 0.41 1.33
Si mg L−1 1.45 0.87 2.85
chl a µg L−1 0.9 ND 6.0
both lakes. The concentrations of dissolved inorganic nitrogen
forms were low, whereas the concentration of dissolved
reactive phosphorus (SRP) was 19.3 µg L−1 in lake 2. Relatively
high concentrations of dissolved organic carbon, particulate
nutrients and chlorophyll a were also characteristic
for lake 2 (Table 1). Low autotrophic biomass in open water
was mostly formed by detached benthic species; no substantial
phytoplankton or floating mats occurred in the lakes. The
comparison of the two sampling dates available for lake 1
suggested high fluctuations of dissolved nutrient concentra-
tions.
3.2 Thermal regime
Figure 3a shows the annual variation of daily mean water
temperature in lake 1 and of daily mean air temperature in the
Solorina Valley (locations of temperature sensors are marked
in Figs. 1 and 2). Lake 1 was frozen to the bottom from the
end of March to the end of October or beginning of November.
Air temperatures were frequently lower than water temperatures.
Minimum daily mean temperatures on the bottom
of the lake were about −12 and −10 ◦C for 2009 and 2010,
respectively. Minimum daily mean air temperatures in the
same period were between −32 and −25 ◦C. Mean monthly
water temperatures in the lake ranged from −10.4 ◦C (August
2009) to 5.8 ◦C (February 2010), while monthly mean
air temperatures were between −18.7 and 0.7 ◦C. The differences
were greater at the beginning of the winter season
(June–July), due to a rapid drop of air temperature.
The highest night–day air temperature fluctuations (up to
28 ◦C) were recorded during the winter months, while the
lowest occurred in summer. In contrast, the highest night–day
amplitudes of lake water temperature were recorded from
November to February, with typical values between 2 and
4 ◦C (Fig. 3b).
The course of global solar radiation (Fig. 3c) was smooth,
with the maximum daily mean of 385 W m−2 during clearsky
conditions around the summer solstice. Global radiation
reached the bottom of both lakes during the ice-free period.
The relative frequency of hourly values of lake 1 water and
air temperature is shown in Fig. S1 in the Supplement. Water
temperature fluctuation was narrow, ranging from −16 to
8 ◦C. The bottom of the lake was frozen for most of the year,
and the growing season, with water at temperatures from 2
to 8 ◦C, covered only 2–3 months (Fig. S1a). In contrast to
lake water thermal regime, air temperature fluctuations were
much wider (from −38 to 8 ◦C) (Fig. S1b).
The temperature at the lake bottom was permanently below
−4 ◦C only during the coldest 2–3 months per year
(Fig. S2). The water temperature above 0 ◦C (liquid phase)
was recorded from November to April (139 days in average).
The number of days with temperature between 0 and
−4 ◦C remains the same as for liquid water occurrence with
small changes in the start and end dates towards to the transition
period (February–June and September–November, respectively).
In such thermal conditions, the benthic littoral
community can be metabolically active.
3.3 Littoral phytobenthos – biofilm community
description
The littoral benthic community in lakes 1 and 2 are dominated
by the heterocytous cyanobacterium Calothrix elsteri
(Komárek et al., 2012) (Fig. 4a), which forms a flat black
biofilm on the upper surface of bottom stones (Fig. 5), followed
by Hassallia andreassenni and Hassallia antarctica
(Komárek et al., 2012). Hassallia andreassenni is associated
with calcium precipitation, as described later. Hassallia
antarctica was found in stone crevices, being only loosely
attached to the substrate. Littoral benthic mats–biofilms on
stones (Fig. 6) are co-dominated on the surface of the blackish
cyanobacterial biofilm by the green filamentous and
richly branched alga Hazenia broadyi (Škaloud et al., 2013)
(Ulotrichales, Chlorophyceae) (Fig. 4b). Hazenia broadyi
grew in macroscopic colonies producing green spots (Fig. 5b,
d). Later in the summer season, the green spots connected miwww.biogeosciences.net/13/535/2016/
Biogeosciences, 13, 535–549, 2016
542 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
1 cm
(a)
(b)
Figure 6. Photoautotrophic mat covering a stone visualised using
imaging fluorometry. (a) Upper view, (b) lateral view.
cro fortified mucilaginous lines (Fig. 5c, d). Figure 5a shows
the community in early spring whereas Fig. 5b, d originated
from later summer when the littoral benthic community was
already well developed with a dense coverage of Hazenia
broadyi green spots. More detailed pictures (Fig. 5e, f) document
the structure of the black leather-like biofilm with
mucilaginous marble on its surface covered by green spots.
When the biofilm gets dry, the net of precipitated micro fortified
mucilage mixed with organic matter, mineral clasts and
crystals of calcium carbonate is visible (Fig. 5g, h).
Scanning electron micrographs document the structure of
the biofilm (Fig. 7). Figure 7a shows a lateral view (cross section)
of a biofilm with cyanobacterial filaments (Calothrix
elsteri and Hassallia andreassenni). A biofilm upper view
(Fig. 7b, d) shows the structure of the cyanobacterialmicroalgae
community producing the mucilaginous micro
fortified net of filaments with spots on its surface.
3.4 Inorganic compounds of biofilms
Thin sections, showing both dry biofilms and rock substrate
(Fig. 8), provided information on various inorganic compounds
associated with the soft tissue of the cyanobacterial–
microalgal community. These inorganic compounds are represented
by (1) allochthonous mineral grains that are overgrown
and incorporated by biofilms and (2) calcareous
spicules of different sizes ranging from 0.5 mm to 1 cm that
are precipitated within the cyanobacterial-microalgal com-
munity.
The rock substrate of biofilms is formed by subangular
to subrounded pebbles to boulders of basaltic rock, which
is dark grey in colour, compact and usually with a microcrystalline
porhyric texture. The rock is not homogeneous,
but contains numerous ball-like empty voids, which are often
partly filled with feldspathoids (Fig. 8a). Crystals of plagioclase
(feldspar group) and augite (pyroxene group) are easily
recognisable in thin sections (Fig. 8a–c).
Biofilms are often partly covered with various mineral
grains and rock fragments, but all specimens studied also
contain these particles incorporated directly within soft
cyanobacterial–microalgal biofilm (Fig. 8a–c).
Mineral grains embedded within biofilms close to the
basaltic rock surface are mainly angular to subangular crystal
fragments of plagioclase and augite (Fig. 8b, c), i.e. the
main mineral components of the basaltic rock substrate described
above. In the upper part of biofilms, however, partly
or fully incorporated grains of quartz occur, being typically
rounded or partly rounded (Fig. 8a, b). One of the thin sections
shows a calcareous spicule in situ and mineral grains
within the biofilm (Fig. 8c, d).
The structure and morphology of calcareous spicules was
studied on SEM (Fig. 9). Crystal facets on the surface and
cleavage (crystallographic structural planes) in the interior of
the spicules (Fig. 9a, b) are typical characteristics of calcium
carbonate monocrystals.
The superficial layer of microcrystalline calcite (e.g.
Fig. 9b) shows the structure of parallel needle-like calcite microcrystals
(Fig. 9d–f). Partial dissolution of spicules show
distinct layering of these needle-like microcrystals (Fig. 9d).
The layered structure of the spicules is confirmed in the ringlike
structures with a possible cyanobacterial filament in the
centre (Fig. 10).
The chemical composition of the studied calcareous
spicules determined by FSD corresponds to pure CaCO3.
Following chemical composition, calcite and aragonite structural
models were applied for the EBSD study focused on
structural identification of the crystals forming the spicule.
Structural identification of the studied specimen especially
prepared for the EBSD study confirmed the absolute agreement
between the recorded EBSD patterns and modelled patterns
for calcite. The presence of aragonite was not confirmed.
FSD images acquired for chemical and orientation
contrasts (Fig. 10) show a layered structure especially visible
in orientation contrast. This feature reflects continual growing
processes on layers with very similar crystallographic
orientation. Absolute angular differences between individual
layers are below 0.8◦.
Biogeosciences, 13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 543
Figure 7. SEM macrographs showing the structure of the dried mat in the lakes: (a) transversal section of the mat with visible cyanobacterial
filaments; (b) surface structure of the mat, the position of cyanobacterial filaments incorporated within mucilaginous matrix is indicated by
arrows; (c) general view of the surface structure of the mat with the net formed by mucilage (compare with Fig. 7d); (d) detail of the same
mucilaginous structure.
4 Discussion
4.1 Environmental properties
The lakes under study are characterised by a low content of
major ions due to their volcanic bedrock and lower marine influence.
In comparison with other lakes of this area, the two
lakes show no specific lake water chemistry characteristics,
with moderate SRP and nitrate concentrations frequently below
the detection limit (Nedbalová et al., 2013). High pH together
with oxygen supersaturation recorded in lake 2 could
be associated with high photosynthetic activity of the mats at
the time of sampling.
Because water in either liquid or solid form has a large heat
storage capacity, it acts as an important buffer to temperature
change. Local climatic conditions of shallow freshwater
lakes is the principal external factor controlling their ecological
functionality. Lake 1 is frozen to the bottom approximately
8–9 months per year. For most of the year, however,
the temperature of the littoral and lake bottom is only from
−2 to −4 ◦C. In such conditions, a thin layer of water probably
covers the surface of the littoral benthic community that
can be metabolically active (Davey et al., 1992). The growing
season, with liquid water at temperatures between 2 to
4 ◦C, covers only 2–3 months.
In regards to heat balance, the studied shallow lakes are
pond (wetlands) environments which freeze solid during the
winter. This inevitability is a strong habitat-defining characteristic,
which places considerable stress on resident organisms
(Hawes et al., 1992; Elster, 2002). In summer, they
must withstand drying in large parts of the littoral zone due
to a considerable drop in water level. In freezing and desiccation
resistance studies of freshwater phytobenthos in shallow
Antarctic lakes, several ecological measurements have
recorded seasonal, diurnal and year-round temperature fluctuations
and changes in water state transitions (e.g. Davey,
1989; Hawes et al., 1992, 1999). In localities with steady
moisture and nutrient supplies, the abundance and species
diversity of algae is relatively high. However, as the severity
and instability of living conditions increases (mainly due to
changes in mechanical disturbances, desiccation–rehydration
and subsequent changes in salinity), algal abundance and
species diversity decreases (Elster and Benson, 2004). The
speed at which the water state can change between liquid, ice
and complete dryness is one of the most important ecological
and physiological factors of these lakes. Studies based
on field or laboratory experiments have shown that some
cyanobacteria and algae are able to tolerate prolonged periods
of desiccation (Pichrtová et al., 2014; Tashyreva and
Elster, 2015). It is also obvious that there are strain/specieswww.biogeosciences.net/13/535/2016/
Biogeosciences, 13, 535–549, 2016
544 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
(c)
(a) (b)
(d)
Figure 8. Perpendicular thin sections of rock substrate covered by dry biofilms (recorded under cross-polarised light. Note that biofilms are
partly detached from the surface of the rock due to complete drying of the sample. (a) Conspicuous U-shaped empty void (arrowed “2”)
near the surface of basaltic rock partly infilled with crystals of feldspathoids (tectosilicate minerals, arrowed “1”); empty void is bridged by
biofilm (arrowed “3”) with partly incorporated mineral clasts, represented by semi-rounded quartz grains (arrowed “4”). (b) Rather thick
biofilm with numerous incorporated mineral grains; note that close to the rock substrate the angular grains of plagioclase (feldspar group)
and augite (pyroxene group) dominate, being derived from basaltoids, whereas close to the surface rounded grains of quartz occur (arrowed).
(c) In situ calcium carbonate spicule penetrating biofilm and surrounded by incorporated grains of feldspars (two arrows on the left) and
pyroxene (arrow on the right). (d) Close-up of the same calcium carbonate spicule with a possible cyanobacterial filament in its centre.
specific differences in the overwintering strategies, and also
between strains/species inhabiting different habitats (Davey,
1989; Hawes et al., 1992; Jacob et al., 1992; Šabacká and Elster,
2006; Elster et al., 2008). The ice and snow which cover
the lakes for about 8–9 months per year serve as a natural incubator,
which moderates potential mechanical disturbances
and stabilises the thermal regimes of the lakes.
4.2 Biodiversity
Patterns of endemism and alien establishment in Antarctica
are very different across taxa and habitat types (terrestrial,
freshwater or marine) (Barnes et al., 2006). Environmental
conditions, as well as dispersal abilities, are important in
limiting alien establishment (Barnes et al., 2006). Antarctic
microbial (cyanobacteria, algae) diversity is still poorly
known, although recent molecular and ecophysiological evidence
support a high level of endemism and speciation/taxon
distinctness (Taton et al., 2003; Rybalka et al., 2009; De
Wever et al., 2009; Komárek et al., 2012; Strunecký et al.,
2012; Škaloud et al., 2013).
The floors of the studied lakes are covered with photosynthetic
microbial mats composed of previously described
species of heterocytous cyanobacteria, mostly Calothrix elsteri
followed by Hassallia andreassenni and Hassallia
antarctica (Komárek et al., 2012). They are co-dominated by
a newly described species of green filamentous and richly
branched algae Hazenia broadyi (Ulotrichales, Chlorophyceae)
(Škaloud et al., 2013). All the previously mentioned
recently described species have special taxonomic positions
together with special ecology and are considered at present
as Antarctic endemic species.
Biogeosciences, 13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 545
(c)
(a) (b)
(d)
(e) (f)
Figure 9. SEM macrographs showing the morphology of calcium carbonate spicules. Spicules were washed away from the living tissue
and collected directly from the surface of biofilms, although residence time on the bottom cannot be determined. (a) Calcareous spicule
showing specific surface texture (“worn surface”). The spicule shows crystal facets on the surface and cleavage (crystallographic structural
planes) in the interior (arrowed) – i.e. typical characteristics of calcium carbonate monocrystal. (b) Detail of previous image; two parallel
systems of deep furrows on the surface are crystallographic structural planes of calcite monocrystal; remnants of a superficial layer of
microcrystalline calcite are, however, preserved in places on the surface of the crystal (arrowed). (c–f) Another type of spicule, formed
mainly by microcrystalline calcite: (c) lateral view of the spicule, (d) detail of the surface showing needle-like calcite microcrystals with
distinct layering, (e) parallel needle-like calcite microcrystals on the surface of the central part of the spicule, (f) tops of parallel needle-like
calcite microcrystals on the surface of the terminal part of the spicule; the view is perpendicular with respect to the previous macrograph.
The black leather-like biofilm with mucilaginous marble
on its surface is covered by green spots. These macroscopic
structures form mats a few millimetres thick consisting of the
above-mentioned species packed in mucilage glued together
with fine material. The regular leather biofilm structure with
distinct cyanobacterial–microalgal composition and incorporated
mineral grains is to our knowledge unique. During the
limnological survey of the whole Ulu Peninsula (Nedbalová
et al., 2013), this specific biofilm structure was observed only
in the two endorheic lakes, although lakes with very similar
morphometric and chemical characteristics are found in the
area. The mat structure is thus apparently tightly linked to the
species composition (Andersen et al., 2011).
The low abundance of benthic diatoms in the lakes is
unusual, but not unprecedented, as there are other areas in
Antarctica where diatoms are scarce or absent (Broady, 1996,
www.biogeosciences.net/13/535/2016/ Biogeosciences, 13, 535–549, 2016
546 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
(a)
(b)
Figure 10. FSD image of a transversely sectioned, partly recrystallised
calcite spicule acquired in (a) chemical contrast, (b) orientation
contrast.
Wagner et al., 2004). The reason underlying the absence of
diatoms is not immediately obvious, because diatoms are
quite a common and frequently dominant component of microbial
communities in most freshwater habitats of the Ulu
Peninsula, James Ross Island (Kopalová et al., 2013). Local
geographical separation of lakes 1 and 2 together with
founder effect may have precluded successful colonisation
by the subset of diatoms that are common in the surrounding
freshwater habitats. Although it has long been held that
diatoms are dispersed widely, some recent reports document
very small scale microbial distributions and endemism
(Kopalová et al., 2012, 2013).
4.3 Inorganic compounds of biofilms
Based on the character of the rock substrate and lake sediments,
it is suggested that two of the main prerequisites for
existence of this cyanobacterial–microalgal community producing
unusual biogenic calcite structures are (1) the flat and
stable substrate in both lakes and (2) the low sedimentation
rate.
The substrate for biofilms is composed of boulders and
pebbles of the stony littoral zone, petrographically corresponding
to compact and massive basaltoids (Smellie et al.,
2008; Svojtka et al., 2009). Rounded or subrounded quartz
grains that are incorporated (“trapped”) within biofilms cannot
originate from basaltic volcanic rocks forming the bottom
of both lakes and substrate of the studied biofilms. This
is evidenced by the petrographic character of the basaltoids,
which do not contain any quartz. The presence of abraded
quartz grains in lakes 1 and 2 can be easily explained by wind
transport (e.g. Shao, 2008).
The specific cyanobacterial–microalgal community described
above can prosper in the two shallow endorheic lakes,
because of low sedimentation rates resulting from minor
water input. Low sedimentary input is the main necessary
ecological parameter which facilitates the existence of this
special microbial community. The community is, however,
well adapted to seasonally elevated sedimentation rates coming
from frequent and intense winds. During wind storms,
the wind carries a relatively large amount of small mineral
grains and rock microfragments (intense eolic erosion; e.g.
Shao (2008) and references therein). These grains and particles
are usually derived from erosion of the rocks either in
the very close vicinity of the locality (weathering of basaltic
rocks), but mainly come from remote locations where especially
Upper Cretaceous marine sedimentary sequences
are outcropping (Smellie et al., 2008; Svojtka et al., 2009).
The amount of mineral grains transported into the lake by
wind does not stop the growth of cyanobacterial–microalgae
biofilms, due to their ability of incorporating and “trapping”
mineral grains within the living tissue (Riding, 2011).
This study has shown that inorganic substances precipitated
by microbial lithogenetic processes are exclusively represented
by calcite spicules. Precipitation of carbonate outside
of microorganisms during photosynthesis as a mechanism
of carbonate construction was described for many filamentous
cyanobacterial species (Schneider and Le CampionAlsumard,
1999). However, the biogenic calcite structures in
both lakes are quite different to any microbially mediated
structures yet described from modern environments (Kremer
et al., 2008; Couradeau et al., 2011) and also to structures
formed by abiotic precipitation (e.g. Vogt and Corte, 1996).
Although there are many lakes with thick mats and similar
chemical characteristics on the Ulu Peninsula, the calcite
spicules were found exclusively in the two endorheic
lakes. We believe that their formation is linked to the specific
photoautotrophic mats present in the lakes. From Fig. 6g
and h it is clearly visible that the calcareous organosedimentary
structures keep the contours of a viable photosynthetic
microbial mat after desiccation or calcite spicules precipitation.
More specifically, the co-dominance of a green microalga
is unique since mats in Antarctic lakes are most frequently
formed by filamentous cyanobacteria (Vincent and
Laybourn-Parry, 2008). Therefore, we hypothesise that the
more rapid photosynthesis rate of Hazenia in comparison
with cyanobacteria may induce conditions necessary for carbonate
precipitation in the lakes (Schneider and Le CampionAlsumard,
1999; Vincent, 2000). However, some role of abiotic
precipitation of calcite is also possible. From our obserBiogeosciences,
13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 547
vations we cannot clearly decide if the winter abiotic calcite
precipitation accompanies microbial lithogenetic processes.
Although we interpret the tubular hollow observed in the
centre of some spicules as the result of the presence of
cyanobacterial filament during the process of crystallisation,
such structures may form also as the result of abiotic precipitation
of calcite (Vogt and Corte, 1996; Fan and Wang,
2005).
It is striking that some calcite spicules probably exhibit
recrystallisation, forming spicules with the structure of calcite
monocrystals. However, these spicules could be also
interpreted as primary structures: mesostructured carbonate
crystals formed through highly oriented growth of micro/nanocrystals
and characterised by a specific surface texture
(Fig. 9a–b). There is already evidence that some biominerals
including calcite are mesocrystals (Cölfen and Antonietti,
2005) and the importance of extracellular polymeric substances
for the formation of some types of nanostructured
carbonate precipitates was documented (Pedley et al., 2009).
Determining the structure and material of precipitated inorganic
substances brought another relevant question: “Do
calcite spicules have fossilisation potential”? Microcrystalline
calcite forming the recrystallised spicule is a typical
material of calcite shells of fossil invertebrates (e.g. Vodrážka,
2009). Although calcite fossils may be partly or completely
dissolved during diagenetical processes in the fossil
record (e.g. Schneider et al., 2011; Švábenická et al., 2012),
their preservation potential is relatively high. Therefore, we
expect to find fossil and/or subfossil calcite spicules from the
Quaternary lake sediments of the studied area.
The Supplement related to this article is available online
at doi:10.5194/bg-13-535-2016-supplement.
Acknowledgements. This study was conducted during two Czech
Antarctic research expeditions of the authors (J. Elster., L. Nedbalová,
R. Vodrážka, K. Láska, J. Komárek) to the J. G. Mendel
station in 2008 and 2009 (headed by Miloš Barták). We are
indebted particularly to the staff and scientific infrastructure of the
station. The study was supported by the Ministry of Education,
Youth and Sports of the Czech Republic (CzechPolar LM2010009,
KONTAKT ME 945 and RVO67985939). R. Vodrážka has been
funded through a Research and Development Project of the Ministry
of Environment of the Czech Republic No. SPII 1a9/23/07 and
project GAˇCR GP14-31662P. K. Láska was supported by a project
of Masaryk University MUNI/A/1370/2014 “Global environmental
changes in time and space”. The authors gratefully acknowledge
the comments received from Kevin Lepot and two anonymous
reviewers. The technical work in laboratories was performed by
Jana Šnokhousová and Dana Švehlová.
Edited by: S. W. A. Naqvi
References
Andersen, D. T., Sumner, D. Y., Hawes, I., Webster-Brown,
J., and McKay, C. P.: Discovery of large conical stromatolites
in Lake Untersee, Antarctica Geobiology, 9, 280–293,
doi:10.1111/j.1472-4669.2011.00279.x, 2011.
Arp, G., Reimer, A., and Reitner, J.: Photosynthesis-induced biofilm
calcification and calcium concentrations in Phanerozoic oceans,
Science, 392, 1701, doi:10.1126/science.1057204, 2001.
Barnes, D. K. A., Hodgson, D. A., Convey, P., Allen, C. S.,
and Clarke, A.: Incursion and excursion of Antarctic biota:
past, present and future, Global Ecol. Biogeogr., 15, 121–142,
doi:10.1111/j.1466-822x.2006.00216.x, 2006.
Broady, P. A.: Diversity, distribution and dispersal of Antarctic
terrestrial algae, Biodiv. Conserv., 5, 1307–1335,
doi:10.1007/BF00051981, 1996.
Caspi, E. N., Pokroy, B., Lee, P. L., Quintana, J. P., and Zolotoyabko,
E.: On the structure of aragonite, Acta Crystallogr., 61,
129–132, doi:10.1107/S0108768105005240, 2005.
Clark, I. D. and Lauriol, B.: Aufeis of the Firth River basin, Northern
Yukon Canada: insights to permafrost hydrology and karst,
Arct. Alp. Res., 29, 240–252, doi:10.2307/1552053, 1997.
Couradeau, E., Benzerara, K., Moreira, D., Gérard, E., Ka´zmierczak,
J., Tavera, R., and López-García, P.: Prokaryotic and eukaryotic
community structure in field and cultured microbialites
from the alkaline lake Alchichica (Mexico), PloS ONE, 6, 1–15,
doi:10.1371/journal.pone.0028767, 2011.
Courty, M. A., Marlin, C., Dever, L., Tremblay, P., and Vachier, P.:
The properties, genesis and environmental significance of calcite
pendents from the high Arctic (Spitsbergen), Geoderma, 61, 71–
102, doi:10.1016/0016-7061(94)90012-4, 1994.
Cölfen, H. and Antonietti, M.: Crystal design of calcium carbonate
microparticles using double-hydrophilic block copolymers,
Langmuir, 14, 582–589, doi:10.1021/la970765t, 1998.
Cölfen, H. and Antonietti, M.: Mesocrystals: inorganic superstructures
made by highly parallel crystallization and controlled
alignment, Angew. Chem. Int. Edit., 44, 5576–5591,
doi:10.1002/anie.200500496, 2005.
Davies, B. J., Glasser, N. F., Carrivick, J. L., Hambrey, M. J.,
Smellie, J. L., and Nývlt, D.: Landscape evolution and ice-sheet
behavior in a semi-arid polar environment: James Ross Island,
NE Antarctic Peninsula, Special publication, Geological Society
of London, doi:10.1144/SP381.1, 2013.
Davey, M. C.: The effect of freezing and desiccation on photosynthesis
and survival of terrestrial Antarctic algae and cyanobacteria,
Polar Biol., 10, 29–36, 1989.
Davey, M. C., Pickup, J., and Block, W.: Temperature variation
and its biological significance in fellfield habitats on a maritime
Antarctic island, Antarct Sci., 4, 383–388, 1992.
De Wever, A., Leliaert, F., Verleyen, E., Vanormelingen, P., Van der
Gucht, K., Hodgson, D. A., Sabbe, K., and Vyverman, W.: Hidden
levels of phylodiversity in Antarctic green algae: further evidence
for the existence of glacial refugee, P. Roy. Soc. B-Biol.
Sci., 276, 3591–3599, doi:10.1098/rspb.2009.0994, 2009.
Effenberger, H., Mereiter, K., and Zemann, J.: Crystal structure refinements
of magnesite, calcite, rhodochrosite, siderite, smithsonite
and dolomite, with discussion of some aspects of the stereochemistry
of calcite-type carbonates, Z. Kristallogr., 156, 233–
243, 1981.
www.biogeosciences.net/13/535/2016/ Biogeosciences, 13, 535–549, 2016
548 J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes
Elster, J.: Ecological classification of terrestrial algal communities
of polar environment, in: GeoEcology of terrestrial oases, edited
by: Beyer, L. and Boelter, M., Ecological Studies, SpringerVerlag,
Berlin, Heidelberg, 303–319, 2002.
Elster, J. and Benson, E. E.: Life in the polar terrestrial environment
with a focus on algae and cyanobacteria, in: Life in the
frozen state, edited by: Fuller, B., Lane, N., and Benson E. E.,
Taylor and Francis, London, doi:10.1201/9780203647073.ch3,
111–149, 2004.
Elster, J., Degma, P., Kováˇcik, L’., Valentová, L., Šrámková, K., and
Pereira, A. B.: Freezing and desiccation injury resistance in the
filamentous green alga Klebsormidium from the Antarctic, Arctic
and Slovakia, Biologia, 63, 839–847, doi:10.2478/s11756-
008-0111-2, 2008.
Fairchild, I. J., Bradby, B., and Spiro, B.: Carbonate diagenesis
in ice, Geology, 21, 901–904, doi:10.1130/0091-
7613(1993)021<0901:CDII>2.3.CO;2, 1993.
Fan, Y. W. and Wang, R. Z.: Submicrometer-sized vaterite tubes
formed through nanobubble-templated crystal growth, Advanced
Materials, 17, 2384–2388, doi:10.1002/adma.200500755, 2005.
Jacob, A., Wiencke, C., Lehmann, H., and Krist, G. O.: Physiology
and ultrastructure of desiccation in the green alga
Prasiola crispa from Antarctica, Bot. Mar., 35, 297–303,
doi:10.1515/botm.1992.35.4.297, 1992.
Johnson, J. S., Bentley, M. J., Roberts, S. J., Binnie, S. A.,
and Freeman, S. P. H. T.: Holocene deglacial history of the
northeast Antarctic Peninsula – A review and new chronological
constraints, Quaternary Sci. Rev., 30, 3791–3802,
doi:10.1016/j.quascirev.2011.10.011, 2011.
Hawes, I., Howard-Williams, C., and Vincent, W. F.: Desiccation
and recovery of Antarctic cyanobacterial mats, Polar Biol., 12,
587–594, 1992.
Hawes, I., Smith, R., Howard-Williams, C., and Schwarz, A. M.:
Environmental conditions during freezing, and response of microbial
mats in ponds of the McMurdo Ice Shelf, Antarctica,
Antarct. Sci., 11, 198–208, 1999.
Hawes, I., Moorhead, D., Sutherland, D., Schmeling, J., and
Schwarz, A. M.: Benthic primary production in two perennially
ice-covered Antarctic lakes: pattern of biomass accumulation
with a model of community metabolism, Antarct Sci., 13,
18–27, 2001.
Komárek, J. and Elster, J.: Ecological background of cyanobacterial
assemblages of the northern part of James Ross Island, NW
Weddell Sea, Antarctica, Pol. Polar Res., 29, 17–32, 2008.
Komárek, J., Nedbalová, L., and Hauer, T.: Phylogenetic position
and taxonomy of three heterocytous cyanobacteria dominating
the littoral of deglaciated lakes, James Ross Island, Antarctica,
Polar Biol., 35, 759–774, doi:10.1007/s00300-011-1123-x, 2012.
Komárek, J., Bonaldo, G. D., Fatima, F. M., and Elster, J.: Heterocytous
cyanobacteria of the Ulu Peninsula, James Ross Island,
Antarctica, Polar Biol., 38, 475–492, doi:10.1007/s00300-014-
1609-4, 2015.
Kopalová, K., Veselá, J., Elster, J., Nedbalová, L., Komárek,
J., and Van de Vijver, B.: Benthic diatoms (Bacillariophyta)
from seepages and streams on James Ross Island
(NW Weddell Sea, Antarctica), Plant Ecol. Evol., 145, 1–19,
doi:10.5091/plecevo.2012.639, 2012.
Kopalová, K., Nedbalová, L., Nývlt, D., Elster, J., and Van de
Vijver, B.: Diversity, ecology and biogeography of the freshwater
diatom communities from Ulu Peninsula (James Ross
Island, NE Antarctic Peninsula), Polar Biol., 36, 933–948,
doi:10.1007/s00300-013-1317-5, 2013.
Kremer, B., Ka´zmierczak, J., and Stal, L. J.: Calcium carbonate
precipitation in cyanobacteria mats from sandy tidal flats
of the North Sea, Geobiology, 6, 46–56, doi:10.1111/j.1472-
4669.2007.00128.x, 2008.
Láska, K., Barták, M., Hájek, J., Prošek, P., and Bohuslavová, O.:
Climatic and ecological characteristics of deglaciated area of
James Ross Island, Antarctica, with a special respect to vegetation
cover, Czech Polar Reports, 1, 49–62, 2011a.
Láska, K., Budík, L., Budíková, M., and Prošek, P.: Method of estimating
of solar UV radiation in high-latitude locations based on
satellite ozone retrieval with improved algorithm, Int. J. Remote
Sens., 32, 3165–3177, doi:10.1080/01431161.2010.541513,
2011b.
Lepot, K., Compère, P., Gérard, E., Namsaraev, Z., Verleyen,
E., Tavernier, I., Hodgson, D. A., Vyverman, W., Gilbert, B.,
Wilmotte, A., and Javaux, E. J.: Organic and mineral imprints in
fossil photosynthetic mats of an East Antarctic lake, Geobiology,
12, 424–450, doi:10.1111/gbi.12096, 2014.
Nakai, N., Wada, H., Kiyoshu, Y., and Takimoto, M.: Stable isotope
studies on the origin and geological history of water and salts in
the Lake Vanda area, Antarctica, Geochem. J., 9, 7–24, 1975.
Nedbalová, L., Nývlt, D., Kopáˇcek, J., Šobr, M., and Elster,
J.: Freshwater lakes of Ulu Peninsula, James Ross Island,
north-east Antarctic Peninsula: origin, geomorphology, and
physical and chemical limnology, Antarct. Sci., 25, 358–372,
doi:10.1017/S0954102012000934, 2013.
Ng, F. and Hallet, B.: Patterning mechanisms in subglacial carbonate
dissolution and deposition, J. Glaciol., 48, 386–400, 2002.
Nývlt, D., Košler, J., Mlˇcoch, B., Mixa, P., Lisá, L., Bubík,
M., and Hendriks, B. W. H.: The Mendel Formation: evidence
for late Miocene climatic cyclicity at the northern tip of
the Antarctic Peninsula, Palaeogeogr. Palaeocl., 299, 363–394,
doi:10.1016/j.palaeo.2010.11.017, 2011.
Olivero, E. B., Scasso, R. A., and Rinaldi, C. A.: Revision of the
Marambio Group, James Ross Island, Antarctica. Instituto Antartico
Argentino, Contribución, 331, 1–28, 1986.
Øvstedal, D. O. and Lewis Smith, R. I.: Lichens of Antarctica and
South Georgia: A guide to their identification and ecology, Studies
in Polar Research, Cambridge University Press, Cambridge,
411 pp., 2001.
Pechar, L.: Use of acetone:methanol mixture for the extraction
and spectrophotometric determination of chlorophyll a in phytoplankton,
Arch. Hydrobiol./Supplement, Algological Studies,
46, 99–117, 1987.
Pedley, M., Rogerson, M., and Middleton, R.: Freshwater calcite
precipitates from in vitro mesocosm flume experiments:
a case for biomediation of tufas, Sedimentology 56, 511–527,
doi:10.1111/j.1365-3091.2008.00983.x, 2009.
Pichrtová, M., Hájek, T., and Elster, J.: Osmotic stress and recovery
in field populations of Zygnema sp. Zygnematophyceae, Streptophyta)
on Svalbard (High Arctic) subjected to natural dessication,
FEMS Microbiol. Ecol., 89, 270–280, doi:10.1111/1574-
6941.12288, 2014.
Rabassa, J., Skvarca, P., Bertani, L., and Mazzoni, E.: Glacier inventory
of James Ross and Vega Islands, Antarctic Peninsula, Ann.
Glaciol., 3, 260–264, 1982.
Biogeosciences, 13, 535–549, 2016 www.biogeosciences.net/13/535/2016/
J. Elster et al.: Unusual biogenic calcite structures in two shallow lakes 549
Reid, P., Dupraz, C., Visscher, P., and Sumner, D.: Microbial
processes forming marine stromatolites, in: Fossil and recent
biofilms – a natural history of life on Earth, edited by: Krumbein,
W. E., Peterson, D. M., and Zavarzin, G. A., Kluwer Academic
Publishers, London, 103–118, 2003.
Riding, R.: Microbialities, stromatolites, and thrombolites, in: Encyclopedia
of Geobiology, edited by: Reitner, J. and Thiel, V.,
Encyclopedia of Earth Science Series, Springer, Heidelberg,
635–654, 2011.
Rybalka, N., Andersen, R. A., Kostikov, I., Mohr, K. I., Massalski,
A., Olech, M., and Friedl, T.: Testing for endemism,
genotypic diversity and species concepts in Antarctic terrestrial
microalgae of the Tribonemataceae (Stramenophiles, Xanthophyceae),
Environ. Microbiol., 11, 554–565, doi:10.1111/j.1462-
2920.2008.01787.x, 2009.
Šabacká, M. and Elster, J.: Response of cyanobacteria and algae
from Antarctic wetland habitats to freezing and desiccation
stress, Polar Biol., 30, 31–37, doi:10.1007/s00300-006-0156-z,
2006.
Schieber, J.: Microbial mats in terrigenous clastics: the challenge
of identification in the rock record, Palaios, 14, 3–12,
doi:10.2307/3515357, 1999.
Schmidt, N. H. and Olensen, N. O.: Computer-aided determination
of crystal-lattice orientation from electron-channeling patterns in
the SEM, Can. Mineral., 28, 15–22, 1989.
Schneider, S. and Le Campion-Alsumard, T.: Construction and destruction
of carbonates by marine and freshwater cyanobacteria,
Eur. J. Phycol., 34, 417–426, doi:10.1017/S0967026299002280,
1999.
Schneider, J., Niebuhr, B., Wilmsen, M., and Vodrážka, R.: Between
the Alb and the Alps – The fauna of the Upper Cretaceous
Sandbach Formation (Passau region, southeast Germany), Bull.
Geosci., 86, 785–816, doi:10.3140/bull.geosci.1279, 2011.
Shao, Y.: Physics and Modelling of Wind Erosion, Atmospheric and
Oceeanographic Sciences Library 37, 2nd Edn., Springer, Heidelberg,
456 pp., 2008.
Škaloud, P., Nedbalová, L., Elster, J., and Komárek, J.: A curious
occurrence of Hazenia broadyi spec. nova in Antarctica and the
review of the genus Hazenia (Ulotrichales, Chlorophyceae), Polar
Biol., 36, 1281–1291, doi:10.1007/s00300-013-1347-z, 2013.
Smellie, J. L., Johnson, J. S., McIntosh, W. C., Esser, R., Gudmundsson,
M. T., Hambrey, M. J., and Van Wyk de Vries,
B.: Six million years of glacial history recorded in volcanic
lithofacies of the James Ross Island Volcanic Group,
Antarctic Peninsula, Palaeogeogr. Palaeocl., 260, 122–148,
doi:10.1016/j.palaeo.2007.08.011, 2008.
Strunecký, O., Elster, J., and Komárek, J.: Molecular clock evidence
for survival of Antarctic cyanobacteria (Oscillatoriales, Phormidium
autumnale) from Paleozoic times, FEMS Microbiol. Ecol.,
82, 482–490, doi:10.1111/j.1574-6941.2012.01426.x, 2012.
Sutherland, D. and Hawes, I.: Annual growth layers as proxies for
past growth conditions for benthic microbial mats in a perennially
ice-covered Antartic lake, FEMS Microbiol. Ecol., 67, 279–
292, doi:10.1111/j.1574-6941.2008.00621.x, 2009.
Švábenická, L., Vodrážka, R., and Nývlt, D.: Calcareous nannofossils
from the Upper Cretaceous of northern James Ross Island,
Antarctica, Geol. Q., 56, 765–772, doi:10.7306/gq.1053, 2012.
Svojtka, M., Nývlt, D., Muramaki, M., Vávrová, J., Filip, J., and
Mixa, P.: Provenance and post-depositional low-temperature evolution
of the James Ross Basin sedimentary rocks (Antarctic
Peninsula) based on fission track analysis, Antarct. Sci., 21, 593–
607, doi:10.1017/S0954102009990241, 2009.
Tashyreva, D. and Elster, J.: The limits of desiccation tolerance of
Arctic Microcoleus strains (Cyanobacteria) and environmental
factors inducing desiccation tolerance, Front. Microbiol., 6, 278,
doi:10.3389/fmicb.2015.00278, 2015.
Taton, A., Grubisic, S., Brambilla, E., de Wit, R., and Wilmotte,
A.: Cyanobacterial diversity in natural and artificial microbial
mats of Lake Fryxell (McMurdo Dry Valleys, Antarctica): a morphological
and molecular approach, Appl. Environ. Microb., 69,
5157–5169, doi:10.1128/AEM.69.9.5157-5169.2003, 2003.
Turner, J., Barrand, N. E., Bracegirdle, T. J., Convey, P., Hodgson,
D. A., Jarvis, M., Jenkins, A., Marshall, G., Meredith, M.
P., Roscoe, H., Shanklin, J., French, J., Goosse, H., Guglielmin,
M., Gutt, J., Jacobs, S., Kennicutt II, M. C., Masson-Delmotte,
V., Mayewski, P., Navarro, F., Robinson, S., Scambos, T., Sparrow,
M., Summerhayes, C., Speer, K., and Klepikov, A.: Antarctic
climate change and the environment: an update, Polar Rec.,
50, 237–259, doi:10.1017/S0032247413000296, 2014.
Vincent, W. F.: Cyanobacterial dominance in polar regions, in: The
Ecology of Cyanobacteria, edited by: Whitton B. A. and Potts,
M., Kluwer Academic Publishers, the Netherlands, 321–340,
2000.
Vincent, W. F. and Laybourn-Parry, J. (Eds.): Polar lakes
and rivers, Oxford University Press, Oxford, 346 pp.,
doi:10.1093/acprof:oso/9780199213887.001.0001, 2008.
Vodrážka, R.: A new method for the extraction of macrofossils from
calcareous rocks using sulphuric acid, Palaeontology, 52, 187–
192, doi:10.1111/j.1475-4983.2008.00829.x, 2009.
Vogt, T. and Corte, A. E.: Secondary precipitates in Pleistocene
and present cryogenic environments (Mendoza Precordillera,
Argentina, Transbaikalia, Siberia, and Seymour Island
Antarctica), Sedimentology, 43, 53–64, doi:10.1111/j.1365-
3091.1996.tb01459.x, 1996.
Wadham, J. L., Tranter, M., and Dowdeswell, J. A.: Hydrochemistry
of meltwaters draining a polythermal-based, high-Arctic glacier,
south Svalbard: II. Winter and early spring, Hydrol. Process., 14,
1767–1786, 2000.
Wagner, B., Cremer, H., Hulzsch, N., Gore, D., and Melles, M.: Late
Pleistocene and Holocene history of Lake Terrasovoje, Amery
Oasis, East Antarctica, and its climatic and environmental implications,
J. Paleolimnol., 32, 321–339, doi:10.1007/s10933-004-
0143-8, 2004.
Walter, M.: Stromatolites, Elsevier Science Ltd., Amsterdam, the
Nederlands, 790 pp., 1976.
Wharton, R.: Stromatolitic mats in Antarctic lakes, in: Phanerozoic
Stromatolites, II, edited by: Bertrand-Safari, J. and Monty, C.,
Springer, New York, USA, doi:10.1007/978-94-011-1124-9_3,
53–70, 1994.
Wharton, R. A., Parker, B. C., Simmons, G. M., and Love, F. G.:
Biogenic calcite structures forming in Lake Fryxell, Antarctica,
Nature, 295, 403–405, doi:10.1038/295403a0, 1982.
Yamamoto, A., Tanabe, K., and Isozaki, Y.: Lower Cretaceous
freshwater stromatolites from northern Kyushu, Japan, Paleontol.
Res., 13, 139–149, doi:10.2517/1342-8144-13.2.139, 2009.
www.biogeosciences.net/13/535/2016/ Biogeosciences, 13, 535–549, 2016
Supplement of Biogeosciences, 13, 535–549, 2016
http://www.biogeosciences.net/13/535/2016/
doi:10.5194/bg-13-535-2016-supplement
© Author(s) 2016. CC Attribution 3.0 License.
Supplement of
Unusual biogenic calcite structures in two shallow lakes,
James Ross Island, Antarctica
J. Elster et al.
Correspondence to: J. Elster (jelster@prf.jcu.cz)
The copyright of individual parts of the supplement might differ from the CC-BY 3.0 licence.
Supplements
Figure S1. Relative frequency of hourly values of water temperature in lake 1 (a) and air
temperature in the Solorina Valley (b) from February 2009 to November 2010.
Figure S2. Occurrence of days with lake 1 water temperature (L1 water) higher than –4 °C
complemented with monthly mean global solar radiation at Mendel Station both from
February 2009 to November 2010.
9
Paper 8
Weather patterns of the coastal zone of Petuniabukta, central
Spitsbergen in the period 2008-2010
Láska, K., Witoszová, D., Prošek, P., 2012,
Polish Polar Research, 33, pp. 297–318
Weather patterns of the coastal zone of Petuniabukta,
central Spitsbergen in the period 2008–2010
Kamil LÁSKA, Denisa WITOSZOVÁ and Pavel PROŠEK
Department of Geography, Faculty of Science, Masaryk University
Kotlářská 2, CZ−61137 Brno, Czech Republic
<laska@sci.muni.cz> <dratenicek@centrum.cz> <prosek@sci.muni.cz>
Abstract: This paper presents the first results of measurements of global solar radiation,
albedo, ground surface and 2−m air temperature, relative humidity, and wind speed and di−
rection carried out in the central part of Spitsbergen Island in the period 2008–2010. The
study site was located on the coastal ice−free zone of Petuniabukta (north−western branch of
Billefjorden), which was strongly affected by local topography, character of the ground sur−
face, and sea ice extent. Temporal analysis of the selected meteorological parameters shows
both strong seasonal and inter−diurnal variation affected by synoptic−scale weather systems,
channelling and drainage effects of the fjords and surrounding glaciers. The prevailing pat−
tern of atmospheric circulation primarily determined the variation in global solar radiation,
wind speed, ground surface and 2−m air temperatures. Furthermore, it was found that ther−
mal differences between Petuniabukta and the nearest meteorological station (Svalbard
Lufthavn) differ significantly due to differences in sea ice concentrations and ice types in
the fjords during the winter and spring months.
Key words: Arctic, Svalbard, climate, temperature, solar radiation.
Introduction
In recent studies, many authors indicated that near−surface air temperature in
the Arctic has increased about twice as fast as temperatures in lower latitudes dur−
ing the past few decades (Przybylak 2000; Knutson et al. 2006; Miller et al. 2010).
The Svalbard archipelago has been also the subject of many studies that confirmed
enormous climate warming, reductions in permafrost thickness, glacier retreat,
and changes in vegetation and terrestrial animal habitats (Ziaja 2005; Rachlewicz
and Szczuciński 2008; Moreau et al. 2009; Prach et al. 2012 this issue).
It is well known that regional climate change directly affects the local climate
on a temporal scale, and, therefore, has a pronounced effect on many aspects of
both the abiotic and biotic components of Arctic ecosystems. A local climate
Pol. Polar Res. 33 (4): 297–318, 2012
vol. 33, no. 4, pp. 297–318, 2012 doi: 10.2478/v10183−012−0025−0
(topoclimate) is defined as the climate of a relatively small region with predomi−
nantly uniform natural conditions (Barry 2008). A local climate is determined by
several environmental factors, e.g. topography, active ground surface (soil, vege−
tation, and glacier), which directly affect water, biota, and snow cover distribution.
There are pronounced spatial differences in these features over distances of 100 m
to 1–10 km. Topoclimatic effects usually extend to a height of 100 to 1000 m
above the ground depending on local topography and its elevation (Barry 2008).
In the case of the Svalbard archipelago, climate differentiation of ice−free
zones is mainly influenced by local topography, character of ground surface (e.g.
patterned ground, tundra zone, and wetland), and distance from an oceanic and/or
glacial margin. It is apparent that each of these features is related to the effects of
weather systems (atmospheric circulation mode) and microclimatic conditions at
the Earth's surface. However, the local topography (slope orientation, slope angle,
and other relief characteristics) plays an essential role in modifying airflow and
weather systems reaching the Svalbard archipelago. As reported by many authors,
airflow transformation and mountain effects lead to a local circulation system and
local wind occurrence (e.g. Hanssen−Bauer et al. 1990; Serreze et al. 1993;
Førland and Hanssen−Bauer 2003; Niedźwiedź 2007; Kilpeläinen et al. 2011;
Mäkiranta et al. 2011).
One of the first topoclimatic observations on Svalbard was carried out in the re−
gion of Treurenberg Bay and Massif Olimp (NE Spitsbergen) in 1899–1900 (Przy−
bylak and Dzierżawski 2004). The second oldest topoclimatic investigations were
done by the team led by Kosiba at Werenskiold Glacier during the International
Geophysical Year 1957–58 and the years 1959–60 (Kosiba 1960). The spatio−
temporal variation of topoclimatic conditions have been furthermore investigated in
the region of Hornsund, Calypsobyen (West Spitsbergen), Kaffiøyra (Northwestern
Spitsbergen), and Petuniabukta (Central Spitsbergen) by researchers from Wrocław
University, Masaryk University (Brno), Maria Curie Skłodowska University
(Lublin), Nicolaus Copernicus University (Toruń), and Adam Mickiewicz Univer−
sity (Poznań) during several summer expeditions (e.g. Baranowski and Głowicki
1975; Brázdil et al. 1988, 1991; Pereyma and Piasecki 1988; Przybylak 1992; Kejna
and Dzieniszewski 1993; Rachlewicz 2003). Recently, comprehensive studies of the
influence of atmospheric circulation on the local climate have been conducted in
Spitsbergen (Przybylak and Arazny 2006; Marsz 2007; Rachlewicz and Styszyńska
2007; Migała et al. 2008; Bednorz and Kolendowicz 2010).
Atmospheric and climate research have been carried out primarily in the vicinity
of large settlements (Longyearbyen, Ny−Ålesund, Barentsburg, Sveagruva) or re−
search stations (e.g. Polish Polar Station “Hornsund”, Nicolaus Copernicus Univer−
sity Polar Station “Kaffiøyra”) situated along the western and south−western coasts
of Spitsbergen. The first meteorological observations in the region of Petuniabukta
(central part of Spitsbergen) were carried out during the Polish expedition from
Adam Mickiewicz University in 1985 (Kostrzewski et al. 1989). Nevertheless, most
298 Kamil Láska et al.
of the climate studies at Petuniabukta dealt with the analysis of short−term meteoro−
logical measurements and observations conducted during single summer seasons
only (June–August). It is therefore crucial to obtain further quantitative information
about the annual and seasonal variation of weather conditions in the central part of
Spitsbergen in order to understand interactions among the components of the cli−
mate systems and possible consequences for Arctic ecosystems.
This paper provides the first results of comprehensive meteorological observa−
tions conducted in Petuniabukta (Billefjorden, Central Spitsbergen) in the period
2008–2010. Our objectives were, therefore, to evaluate temporal variability of
weather conditions of the coastal ice−free zone of Petuniabukta, and evaluate the
air temperature differences between the study site and the nearest meteorological
station (Svalbard Lufthavn) with regular all−year round observations (Fig. 1). We
focused primarily on describing the diurnal and annual regimes of basic meteoro−
Weather patterns of Petuniabukta, central Spitsbergen 299
Fig. 1. Location of the study site in Petuniabukta (Billefjorden), central Spitsbergen. Abbreviations:
AWS – Automatic Weather Station of the Czech Team, LYR – Svalbard Lufthavn Weather Station
(Norwegian Meteorological Institute). Map source: Svalbardkartet, Norwegian Polar Institute.
logical parameters measured at such a remote site as the Petuniabukta, and their as−
sessment in the context of synoptic situations over the Svalbard archipelago. How−
ever, the preliminary character of the presented results arises from their general
analysis and description on which the authors focused at the beginning. This was
mainly due to the quality of the meteorological datasets and partly to a number of
missing parameters (air pressure and wind direction). Hence, detailed analyses, in
particular evaluation of atmospheric processes (e.g. dependence of the selected
meteorological parameters on local atmospheric circulation by objective meth−
ods), will be the subject of further study.
Study area
The investigated area was located on the western coast of Petuniabukta in the
central part of Spitsbergen, the largest island of the Svalbard archipelago (Fig. 1).
The Petuniabukta is part of the northern branch of Billefjorden, which is an exten−
sion of Isfjorden, the second longest fjord in Spitsbergen. The automatic weather
station (AWS) was situated on the flat marine terrace at the geographical coordi−
nates 78°42.11’N, 16°27.64’E and the altitude of 15 m a.s.l. From the southwest−
ern to western sector, it is surrounded by the Pyramiden (935 m a.s.l.), Mumien
(770 m a.s.l.), and Svenbrehøgda (620 m a.s.l.) mountain ranges and numerous
valley glaciers (Ferdinandbreen, Svenbreen, Hørbyebreen, Ragnarbreen, and
Ebbabreen). From the northeastern to southeastern sector, it is close to the hum−
mock tundra, which is regularly flooded with thawing water from snow patches
and several glacier rivers. Another important feature of Petuniabukta is the pres−
ence of seasonal sea ice cover (mainly fast−ice and open drift ice) usually occurring
from December to May (Nilsen et al. 2008). As the AWS was located 500 metres
from the coastline, winds from the southeast to east represented the airflow from
the fjord. An ice−cement type of the permafrost occurs predominantly in the lower
part of the coastal zone. During summer, the permafrost table is located at a depth
of 40–120 cm depending on e.g. altitude, slope and relief exposition (Láska et al.
2010). The vicinity of the AWS is covered by permanent tundra vegetation with
several dominant species: Cassiope tetragona, Dryas octopetala, Carex rupestris,
Salix polaris, and Saxifraga opossitifolia (J. Gloser, personal communication).
Methods
The first part of the AWS was installed at Petuniabukta in October 2007 and
has been operated all−year round since June 2010. Shortwave incoming and re−
flected solar radiations were measured using two identical Shenk 8101 starpyrano−
meters (Ph. Shenk, Austria), with an accuracy of <5% of the nominal value. Addi−
tional errors of the starpyranometers due to ice formation and snow deposition in
300 Kamil Láska et al.
winter and spring may temporarily reduce the daily mean intensity of solar radia−
tion by up to 30%. Ground surface temperature was monitored by an Omega
OS−36−2 near−infrared temperature sensor (Omega, USA), with accuracy ±2% of
the nominal range (−18 to 85°C) and ±5% outside of the temperature range. The in−
struments were located at a height of 1.5 m above ground. As the OS−36−2 sensor
is sensitive to the emissivity of the measured surface, emissivity values were
changed during the year in accordance with the OS36 Operator’s Manual (Omega
Corp. USA, http://www.omega.com). In most cases, surface emissivity was set in
the range of 0.993 (tundra vegetation) to 0.995 (snow) according to the actual
ground surface conditions (M. Bartak, personal communication). Air temperature
and relative humidity were measured by an EMS33 probe (EMS Brno, Czech Re−
public) at the height of 2 m above ground. Accuracy of the temperature probe was
±0.15°C, while humidity probe accuracy was 3% of the nominal value. Wind
speed and direction were recorded by a MetOne 034B anemometer (MetOne,
USA) at a height of 6 m above ground (Fig. 2). Wind speed was monitored with an
accuracy of ±0.1 ms−1 (starting threshold at 0.4 ms−1), while wind direction accu−
racy was ±4°. Solar radiation measurements were taken every 10 s and data were
stored as 30 min averages in a MiniCube VV/VX Data Logger (EMS Brno, Czech
Republic). The other sensors were sampled at 30 min intervals only. High power
consumption of the AWS did not allow measurements of precipitation due to its
Weather patterns of Petuniabukta, central Spitsbergen 301
Fig. 2. General view of the automatic weather station located on the marine terrace in the western
coast of Petuniabukta with the Ebba Valley and Ebbabreen in the background.
prevailing solid state and the fact that the whole measuring system was supplied by
small solar panels. All sensors were thoroughly calibrated before installation at the
study site and then regularly every year at the calibration laboratory of the Czech
Hydrometerological Institute, EMS Brno (CZ), and during summer expeditions by
comparison with a CM−11 pyranometer (Kipp & Zonen, Netherlands).
In this study, we use the term albedo when we are referring to the hemispheri−
cal reflection of solar radiation from the Earth’s surface, with wavelengths be−
tween 300 and 3000 nm. Albedo was evaluated from reflected and incoming solar
radiation to determine the occurrence of snow cover at the study site. The albedo
(a) was calculated for each 30−min values as
a = RR / GR
where GR is the intensity of global shortwave radiation incident on the Earth’s sur−
face (Wm−2) and RR the intensity of shortwave radiation intensity reflected from
the surface back to the atmosphere (Wm−2). The albedo was not calculated during
low solar elevation angle, typically being less than 5° (November–February).
The 30−min, daily and monthly means and extreme values from the AWS were
analysed using STATISTICA (2011) and standard statistical methods. This also
includes the standard deviation, correlation coefficient, and variation coefficient
determined for the selected meteorological parameters according to Storch and
Zwiers (1999). Only complete time series of daily mean and extreme values of the
selected meteorological parameters corresponding to the period from August 1,
2008 to June 30, 2010 (698 days) were used for the following analyses. Therefore,
the wind speed and direction data, which included several malfunctions of the ane−
mometer, were reduced to the period from August 1 to September 30, 2008 and the
period from July 1, 2009 to June 30, 2010.
In order to evaluate the weather conditions within the study period, changes in
the atmospheric circulation patterns were analysed and compared with the multiyear
data (1970–2000). The description of atmospheric circulation was based on the cal−
endar of circulation types for the Spitsbergen area, provided by Niedźwiedź (2012).
The principles of the classification are synoptic maps from which direction of the
geostrophic wind and the kind of pressure pattern (cyclonic or anticyclonic) is deter−
mined (Niedźwiedź 2007). The author of the classification distinguished 21 circula−
tion types according to the common directions of advection, adding the symbol “a”
for anticyclonic (high pressure) and “c” for cyclonic (low pressure) systems. Other
distinguished types were the anticyclonic centre over Spitsbergen (Ca), anticyclonic
wedge (Ka), cyclonic centre over Spitsbergen (Cc), cyclonic trough (Bc), and baric
col or synoptic situations which was impossible to classify (X).
The second objective of the analysis was to evaluate air temperature differ−
ences between the study site and the nearest all−year round operating the weather
station at Svalbard Lufthavn (Fig. 1). The meteorological data for Svalbard
Lufthavn were provided by the Norwegian Meteorological Institute in Oslo. The
302 Kamil Láska et al.
station is situated near the mouth of Adventfjorden, approximately 600 m from the
coastline, at an elevation of 28 m a.s.l. Its location in the middle part of Isfjorden
(42 km from the open ocean) contributes to different climate conditions in compar−
ison to meteorological stations along the western and south−western coast of
Spitsbergen, e.g. Ny−Ålesund or Horsund (Przybylak and Araźny 2006). The dis−
tance between the Svalbard Lufthavn station and the study site in Petuniabukta is
56 km (Fig. 1). Due to the fact that sea ice occurrence plays an important role in
shaping the local climate conditions (Kilpeläinen et al. 2011), part of the analysis
was devoted to evaluation of the effects of sea ice occurrence on temperature varia−
tions between both sites. Sea ice conditions were evaluated according to the high
resolution sea ice charts of the Svalbard area provided by the Sea Ice Service of the
Norwegian Meteorological Institute (http://polarview.met.no/).
Results and discussion
Atmospheric circulation. — In order to evaluate the effect of atmospheric
circulation on local weather conditions in the observed period, the classification of
circulation pattern provided by Niedźwiedź (2012) for the Svalbard archipelago
was used. We compared the relative frequency of occurrence of circulation types
in the period from August 2008 to June 2010 against the long−term average values
from 1970 to 2000 (Fig. 3). In the study period (2008–2010), cyclonic activity pre−
vailed on 58.9% of the days, while anticyclonic situations occurred in only 38.1%.
The most common circulation types were from the northeastern sector, with fre−
quency of occurrence being 9.6% (Ec type), 9.0% (Nc type), 8.3% (SEc type) and
7.2% (NEc type). Considerably high frequencies were found also for the anticy−
clonic wedge (Ka type – 9.0%), trough of low pressure (Bc type – 8.6%), and anti−
cyclonic situations from the northeastern sector, Na (5.7%), NEa (5.3%), and Ea
(6.3%) in particular. The highest activity of the cyclonic circulation was found in
Weather patterns of Petuniabukta, central Spitsbergen 303
Fig. 3. Relative frequency of occurrence of circulation types over Spitsbergen area in the period from
August 2008 to June 2010, as compared with the long−term average from 1970 to 2000.
autumn (88.2%) and winter (86.6%). Conversely, the anticyclonic types were
most common in summer (35.3%) and spring (19.0%). A very similar distribution
of baric patterns was found from 1970 to 2000, where the cyclonic situations pre−
vailed in 59.4%, while the anticyclonic situations occurred in 37.3% of the days.
The remaining types of circulation (X – baric col or other unclassified situations)
were recorded at frequencies of 3.0% (2008–2010) and 3.3% (1970–2000) respec−
tively. The highest differences in circulation pattern between the study period and
the 1970–2000 period were found during cyclonic situations (Fig. 3). The circula−
tion types Nc, SEc and Bc (troughs of low pressure) showed considerable in−
creases in frequency by +3.3%, +1.5% and +2.3%, compared to the long−term av−
erage values. On the other hand, circulation types NEc and SWc, with differences
of −2.0% and −2.1%, occurred less frequently compared to the long−term averages.
The anticyclonic situations were most frequently represented by type Na, with a
difference of +1.9%, while type Ka decreased in frequency by −1.4% compared to
the long−term average values (Fig. 3). A similar pattern of atmospheric circulation
over the Svalbard area, as well as similar seasonal variation of circulation types,
were found by Førland and Hanssen−Bauer (2003), Przybylak and Araźny (2006),
Niedźwiedź (2007), and Migała et al. (2008).
Surface Wind Field. — The relative frequency of wind direction measured at
the 6−m mast was estimated from 30−min data. The effect of local topography on
the wind pattern in Petuniabukta is evident in Fig. 4. From July 2009 to June 2010,
the prevailing wind directions were observed from the south (9.6%), northeast
(8.2%), and northwest (7.8%), which clearly corresponded with the local topogra−
phy and longitudinal axis of the Billefjorden, Ragnardalen, and Ferdinanddalen
valleys. Such a channeling effect in Petuniabukta was confirmed by the studies of
Rachlewicz and Styszyńska (2007), and Bednorz and Kolendowicz (2010). On the
other hand, northerly winds (2.4%) had the lowest frequency of wind direction.
Seasonal variation in the frequency of main drainage flows was apparent between
cold and warm periods of the year. Southerly and northeasterly winds were most
common in summer (June–August), with a frequency of occurrence representing
16.1% and 14.5% of all cases. Conversely, these winds rarely appeared from Octo−
ber to February, which had frequencies between 6.6% and 5.0% in all cases.
Mean wind speed during the study period was estimated at 3.9 ms−1. The stron−
gest winds were primarily connected to the northeasterly flow from the Ragnar−
dalen and Ebbadalen valleys (Ragnarbreen and Ebbabreen glaciers). In this case,
mean wind speed varied between 5.4 ms−1 for ENE and 6.4 ms−1 for NNE directions
Fig. 4. The daily maximal wind speed from these directions exceeded 11.9 ms−1.
Due to the blocking effects of the Pyramiden, Mumien, and Svenbrehøgda moun−
tain ranges, the lowest wind speed was observed from the western sector, with
mean values ranging from 2.5 ms−1 (WSW) to 2.8 ms−1 (WNW). Therefore, daily
maximal wind speed from western sector was 8 ms−1 only. Generally, the westerly
winds occurred primarily in autumn and winter (October–February), with a wind
304 Kamil Láska et al.
frequency of 8.1% for all cases. The highest daily mean wind speed exceeded 12.2
ms−1 on April 12, 2010. During this event, a low pressure system passed over
Spitsbergen, which was accompanied by changes in atmospheric circulation from
types Bc (April 11, 2010), Cc, and Nc to SEc (April 14, 2010). This was also re−
flected in sudden changes in surface wind direction, turning from westerly to
northeasterly and southeasterly flows. A maximum 30−min average wind speed of
23.6 ms−1 was measured from the northeast on January 27, 2010. In this situation,
circulation type Ec stayed over Spitsbergen until 28 January 2010, while the direc−
tion of the geostrophic wind (estimated from an ECMWF analysis) approached
more to the southeast. On the other hand, daily mean wind speed was smaller than
1 ms−1 for 15.7% of the time between July 2009 and June 2010. The lowest
monthly mean wind speed was observed in December 2009 (2.9 ms−1), while the
highest monthly wind speed was reached in March 2010 (5.1 ms−1). These results,
however, differ from the previous study of Kruszewski (2006), who estimated sea−
sonal variation of geostrophic winds over the Spitsbergen area from 1981–2005,
with the highest monthly U−wind values in June and July, while the lowest values
occurred in February and March.
In general, the local winds in Petuniabukta were affected mainly by large−scale
circulation. As in most fjords in Svalbard, channeling effects and katabatic flows oc−
cur in Petuniabukta as well. It is evident that local topography, together with the ma−
jor driving mechanism of the land−sea breeze circulation driven by horizontal tem−
perature differences between the open water of the fjords and the glaciers, may lead
to significant differences in the surface wind pattern (Esau and Repina 2012). There−
fore, along−fjord (southern) winds in Petuniabukta were the most frequent in the
study period. Similar features were found during the northeasterly drainage flow
from the Ragnarbreen glacier, from which the highest wind speed was recorded.
Global solar radiation. — Temporal variability of the selected meteorological
elements measured at Petuniabukta from August 1, 2008 to June 30, 2010 is shown
in Fig. 5 and Table 1. The seasonal variation of incoming solar radiation intensity
ranged from 0 to about 380 Wm−2, which can be expected from the geographic loca−
Weather patterns of Petuniabukta, central Spitsbergen 305
Fig. 4. Relative frequency of wind direction, mean wind speed, and 2−m air temperature, according to
the airflow direction at Petuniabukta in the period from July 2009 to June 2010.
tion of the study site, and changes in solar elevation angle and the Earth–Sun dis−
tance through the year. The annual course of solar radiation intensity was modified
by the phenomena of the polar day and the polar night. The polar day at Petuniabukta
306 Kamil Láska et al.
Fig. 5. Time series of daily means of global solar radiation, albedo, ground surface temperature, 2−m
air temperature, and relative humidity at Petuniabukta in the period 2008–2010.
lasts from April 17 to August 25 (131 days) and the polar night from October 26 to
February 16 (114 days). Zero radiation intensity due to the actual atmospheric con−
ditions and local topography occurred from October 31, 2008 to February 9, 2009
(101 days) and from October 30, 2009 to February 10, 2010 (103 days) respectively.
The annual mean solar radiation in 2009 was 84.5 Wm−2, while the maximum
monthly means were recorded in July 2009 (247.2 Wm−2) and June 2010 (240.7
Wm−2). The maximum daily means of 381.3 Wm−2 and 369.0 Wm−2 were measured
during clear sky conditions on June 21, 2009 and June 28, 2010 respectively.
Apart from the above−mentioned factors, cloudiness and its optical properties
strongly modulate the intensity of solar radiation on the Earth’s surface. The diurnal
course of solar radiation intensity within two summer days with varied cloudiness is
Weather patterns of Petuniabukta, central Spitsbergen 307
Table 1
Seasonal variation of monthly means and extremes of the selected meteorological parame−
ters at Petuniabukta in the period 2008–2010.
Month
−Year
Global
radiation
Albe−
do
Ground surface
temperature
2−m air temperature
Rela−
tive
humi−
dity
Wind
speed
Max
wind
speedMean Max Min Mean Max Min
[Wm−2] [°C] [°C] [°C] [°C] [°C] [°C] [%] [ms−1] [ms−1]
Aug−2008 108.0 0.14 6.6 18.2 −1.7 4.7 10.9 0.2 86.0 3.0 13.9
Sep−2008 31.3 0.12 2.4 9.5 −8.0 2.5 9.8 −6.2 86.6 3.8 14.5
Oct−2008 5.6 0.47 −7.3 1.5 −20.0 −6.3 4.1 −16.2 78.5 – –
Nov−2008 0.0 – −12.3 0.5 −25.8 −9.6 3.4 −20.6 76.7 – –
Dec−2008 0.0 – −11.6 0.9 −30.6 −9.6 2.9 −24.2 81.5 – –
Jan−2009 0.0 – −16.8 0.7 −36.2 −14.6 1.7 −33.3 78.9 – –
Feb−2009 1.5 – −15.8 0.7 −33.4 −12.9 3.0 −27.0 76.8 – –
Mar−2009 33.2 0.80 −13.9 −1.8 −34.6 −12.3 0.1 −26.3 82.8 – –
Apr−2009 150.8 0.80 −19.5 −1.8 −33.7 −17.0 −0.6 −26.9 75.1 – –
May−2009 218.1 0.79 −3.2 1.0 −19.6 −1.4 9.1 −15.0 77.8 – –
Jun−2009 247.2 0.38 3.5 21.0 −5.9 2.7 10.0 −4.0 79.6 – –
Jul−2009 199.4 0.14 10.8 26.5 1.6 7.2 16.2 1.2 76.9 4.0 14.6
Aug−2009 114.5 0.14 8.2 23.5 −3.7 6.2 12.6 −0.2 86.1 3.2 13.6
Sep−2009 39.7 0.14 2.1 14.5 −5.6 1.9 9.7 −4.9 83.5 3.4 14.2
Oct−2009 4.8 0.74 −5.6 1.0 −19.5 −4.3 2.7 −13.4 84.7 3.2 14.6
Nov−2009 0.0 – −4.0 1.0 −18.7 −3.0 4.1 −16.6 85.9 3.4 15.6
Dec−2009 0.0 – −9.2 0.9 −24.0 −6.7 2.6 −18.3 80.7 2.8 16.0
Jan−2010 0.0 – −11.4 1.0 −31.1 −9.8 3.5 −24.6 79.9 4.6 23.6
Feb−2010 1.7 – −15.9 −2.4 −31.3 −12.2 −0.6 −26.6 72.8 3.2 15.7
Mar−2010 39.9 0.72 −18.2 −2.4 −30.6 −15.9 −1.5 −26.2 78.1 5.0 19.5
Apr−2010 148.3 0.74 −11.7 −1.5 −24.8 −9.5 −0.4 −19.6 78.0 4.3 16.4
May−2010 207.0 0.63 −1.2 3.2 −17.8 −0.1 7.0 −12.3 80.9 3.1 12.0
Jun−2010 240.7 0.14 7.9 23.1 −0.7 3.3 9.6 −1.0 80.4 4.0 14.2
shown in Fig. 6. Clear sky conditions on July 6, 2009 were apparent in the high in−
tensity of incoming radiation, with a 30−min mean maximum of 541 Wm−2, while an
overcast sky on July 8, 2009 was pronounced in the low diurnal course with a
30−min mean maximum of 271 Wm−2. High inter−diurnal and diurnal variations of
solar radiation were found in the summer period (June–August). The remoteness of
Petuniabukta from the nearest settlements, as well as the monitoring systems of the
AWS, did not allow year−round observations of cloudiness. Therefore, cloudiness
was approximately evaluated using global radiation and its attenuation by the cloud
cover. As seen in Figs 5 and 6, there were inter−diurnal changes in the incoming radi−
ative flux, which fluctuated between 90–200 Wm−2 per 24 hours depending on
cloudiness. The effect of reduced cloudiness on solar radiation was also apparent
within the spring periods, with the monthly mean values ranging between 150.8 and
218.1 Wm−2 (April–May 2009) and 148.3 and 207.0 Wm−2 (April–May 2010).
The significant dependence of solar radiation on cloudiness and prevailing
weather systems (air mass origin) was also confirmed by variation coefficients calcu−
lated for April–September according to Storch and Zwiers (1999). As seen in Table 2,
the variation coefficient of solar radiation gradually increased from spring to autumn
2008 and/or 2009, due to increasing cyclonic activity, cloudiness and its fast changes.
Therefore, the highest coefficients of variation were 66.2% (September 2008),
122.7% (October 2008) and 107% (October 2009). During these months, there was
an increase in cloudiness due to cyclonic situations represented mainly in higher fre−
quency of the types Nc (25.8% – October 2009), NEc (12.9% – October 2008), and
Ec (29.0% – October 2008). This was also confirmed by long−term average values
(1970–2000), where the circulation type Ec was the most frequent in September
(10.4%) and October (12.4%), while type NEc was dominant in October (10.8%).
Moreover, considerably high coefficients in March 2009 (70.6%) and 2010 (61.9%)
308 Kamil Láska et al.
Fig. 6. Diurnal course of global solar radiation intensity (grey bars), 2−m air temperature (Ta) and
ground surface temperature (Ts) at Petuniabukta on a clear sky (6 July 2009) and overcast conditions
(8 July 2009).
were probably caused by the small solar elevation angle and biases of solar radiation
measurements. On the other hand, sunny weather, with the smallest changes in cloud−
iness, was affected by anticyclonic situations, which prevailed in May 2009 and
2010. In these cases, the variation coefficients ranged only from 27.7 to 28.3% due to
anticyclonic situations, with the highest occurrence of types Ea (16.1% – May 2010),
Ka (16.1% – May 2010), and Wa (19.4% – May 2009). It is supported by the
long−term average values (1970–2000), where type Ka (anticyclonic wedge), with a
frequency of 17.6%, was one of the most frequent types in May. This is consistent
with the findings of previous studies on solar radiation, cloudiness and the surface en−
ergy budget in the Ny−Ålesund region (Ørbæk et al. 1999; Przybylak and Araźny
2006). In general, the maritime climate of the Svalbard archipelago is formed by a
high amount of clouds, large short−term variation in cloudiness, and the high fre−
Weather patterns of Petuniabukta, central Spitsbergen 309
Table 2
Seasonal variation of 2−m air temperature, ground surface temperature, global solar radia−
tion intensity, albedo, and relative humidity expressed by the standard deviation and coef−
ficient of variation at Petuniabukta in the period 2008–2010.
Month−Year
Standard deviation [°C] Coefficient of variation [%]
Air
temperature
Ground surface
temperature
Global solar
radiation
Albedo
Relative
humidity
Aug−2008 1.5 1.8 44.3 8.8 8.7
Sep−2008 2.9 3.0 66.2 15.3 10.9
Oct−2008 3.6 3.9 122.7 41.9 9.6
Nov−2008 5.1 5.9 – – 12.3
Dec−2008 5.5 6.2 – – 10.4
Jan−2009 10.7 11.3 – – 18.2
Feb−2009 6.6 8.6 – – 18.8
Mar−2009 6.2 7.0 70.6 7.2 15.6
Apr−2009 4.5 5.3 32.3 6.4 10.9
May−2009 3.0 3.5 27.7 3.1 13.5
Jun−2009 2.3 4.3 31.0 73.8 5.9
Jul−2009 2.8 3.2 51.1 10.2 16.2
Aug−2009 2.0 3.1 50.5 10.2 9.5
Sep−2009 2.5 2.6 45.4 20.7 10.2
Oct−2009 3.0 3.5 107.0 25.5 8.7
Nov−2009 4.2 4.4 – – 10.3
Dec−2009 5.1 6.0 – – 12.3
Jan−2010 7.6 8.7 – – 20.1
Feb−2010 6.5 7.7 – – 17.2
Mar−2010 3.8 4.5 61.9 8.8 13.8
Apr−2010 3.9 4.4 34.2 7.4 7.1
May−2010 2.5 2.7 28.3 22.9 10.4
Jun−2010 2.4 3.3 36.4 31.1 11.2
quency of Arctic sea fog, formed by the advection of relatively warm air over colder
surfaces, predominantly in summer (Svendsen et al. 2002).
Albedo and snow cover occurrence. — Albedo from reflected and incoming
solar radiation intensities was calculated to determine the occurrence of snow
cover at the study site. Mean albedo was 0.50 in the period from August 1, 2008 to
June 30, 2010. Seasonal variation of albedo was affected by the character of the
ground surface (Fig. 5). Therefore, monthly mean albedo varied from 0.12 over
tundra (September 2008) to 0.80 during snow accumulation (March–April 2009).
Albedo over tundra vegetation in both dry and wet conditions typically ranged
from 0.11 to 0.16. On the other hand, albedo of fresh snow in March and April
reached up to 0.87. Analysis of the albedo was used to estimate snow cover occur−
rence near the study site. The first day with snow cover was frequently observed at
the beginning of October, while the last day occurred in the first half of June. Per−
manent snow cover lasted from October 16, 2008 to June 16, 2009 (243 days) and
from October 3, 2009 to June 5, 2010 (245 days). The effect of direct solar radia−
tion and thermal advection on snow cover thawing is shown in Fig. 5. From the
albedo variation and its spring decline, it was estimated that the last snow cover
melted within different time periods: June 13–16, 2009 (4 days) and May 27 – June
16, 2010 (20 days). Apart from a low summer albedo (0.14), high inter−diurnal
variation was found in September and October due to the first snowfalls and fast
thawing of snow cover (Table 2). Due to this fact, daily mean albedo in autumn
varied between 0.11 and 0.94. The snow−free period in Petuniabukta typically
lasted from the first half of June to the beginning of October (120 days).
Surface ground temperature. — Surface ground temperature of the tundra
vegetation was measured close to the AWS. The annual mean temperature in 2009
was −5.2°C, while monthly mean temperature ranged from −19.5°C (April 2009) to
10.8°C (July 2009). Daily mean temperature varied between −32.6°C to 15.6°C
(Fig. 5 and Table 1). The lowest surface temperature of −36.2°C was observed on
January 7, 2009 on the snow surface. The highest surface temperature rose to
26.5°C on July 11, 2009 due to the high intensity of incoming radiation that inci−
dents on the dry tundra vegetation surface. These conditions occurred during an
eastern advection (circulation type Ea) characterized by a clear sky and wind speed
of less than 2 ms−1. The diurnal course of surface temperature during two different
days with different weather conditions in July 2009 is shown in Fig. 6. The typical
diurnal sinusoidal behavior was found under clear sky conditions (July 6, 2009)
with high intensity of solar radiation and a noon−time surface temperature reaching
22.3°C. On the contrary, a rather suppressed diurnal course, with noon−time tem−
perature of 16°C, was caused by an overcast sky and relatively low incoming radi−
ation (July 8, 2009). Differences in the diurnal course were mainly affected by
changes in circulation conditions over the Spitsbergen area, where the type NEa
(July 6, 2009) was swiftly replaced by the NWc type (July 8, 2009).
310 Kamil Láska et al.
Short−term variability of the daily mean surface temperature often exceeded
20°C during winter and spring, with the highest standard deviation between
11.3°C (January 2009) and 8.7°C (January 2010). As seen in Table 2, the lowest
deviation was found in August 2008 (1.9°C) and September 2009 (2.6°C). The
rapid increase of surface temperature above 0°C was observed in the first half of
June after intensive snow melting. The first day with a negative surface tempera−
ture occurred on September 4, 2008 and August 16, 2009, while permanent freez−
ing of the ground surface was measured on October 10, 2008 and October 2, 2009.
The dependence of surface ground temperature on the snow cover was clearly
documented in positive surface temperatures, which followed after the last day
with snow cover by a delay of 1–3 days.
Relative air humidity. — In the study period, mean relative humidity at 2 m
above the ground surface exceeded 80%. As seen in Table 1, seasonal variation of
relative humidity was relatively small, and therefore, monthly mean humidity
ranged from 72.8% (February 2010) to 86.6% (September 2008). Daily mean rela−
tive humidity varied between 29.6% and 99.9% (Fig. 5). Daily mean humidity was
higher than 70% in 84.1% of the study period. The highest variation of humidity
and daily minimum values (less than 50%) were typically observed from the be−
ginning of January to mid April (Table 2). On the other hand, the lowest variation
occurred from June to August due to weak cyclonic activity and small inter−diurnal
variation of the other meteorological parameters such as global solar radiation and
2−m air temperature. This suggests that the high frequency of Arctic sea fog
formed by the advection of relatively warm (maritime) air masses over the colder
surfaces of the fjords and glaciers significantly influenced the high values of rela−
tive humidity also in Petuniabukta.
Air temperature. — The monthly means and extremes of air temperature at 2
m above the ground surface are shown in Table 1. The annual mean temperature in
2009 reached −4.5°C, while monthly mean air temperature ranged from −17.0°C
(April 2009) to 7.2°C (July 2009). The lowest air temperature of −33.3°C was ob−
served on January 12, 2009. The highest air temperature rose to 16.2°C on July 28,
2009 during partly cloudy weather with a wind speed less than 4 ms−1. As seen in
Fig. 5, daily mean air temperature varied between −32.6°C and 12.2°C. The daily
mean temperature was higher than 5°C for 51% of the time between June and Au−
gust 2009. By contrast to the ground surface temperature, the diurnal course of 2−m
air temperature did not tightly follow solar radiation intensity (Fig. 6). Due to this
fact, the highest air temperature under clear sky conditions on July 6, 2009 was ob−
served at 18 local time. Short−term variability of daily mean air temperature was
found in winter and spring, when it often exceeded 15°C. Therefore, the highest
standard deviation of air temperature occurred in January 2009 (10.7°C) and Janu−
ary 2010 (7.6°C). Conversely, the lowest deviation was found in August 2008
(1.5°C) and August 2009 (2.0°C).
Weather patterns of Petuniabukta, central Spitsbergen 311
The positive surface energy balance due to the force of the solar radiation
reaching the tundra vegetation strongly affected the diurnal amplitude of both
ground surface and 2−m air temperatures (Fig. 6). However, the dependence of
daily mean surface temperature on daily mean solar radiation was apparently
weaker than between ground surface and 2−m air temperatures. The correlation co−
efficient varied between 0.433 (surface temperature as a function of radiation) and
0.983 (2−m air temperature as a function of surface temperature) at the significance
level of 0.05. The above results were calculated for the snow−free period, from
June to August in particular. Moreover, both relationships varied over time, with
the highest correlations of 0.543 and 0.931 in July 2009. The seasonal variation of
the correlation coefficient, therefore, clearly confirmed the essential roles of
large−scale thermal advection and snow cover occurrence on air temperature varia−
tion in Petuniabukta. These results are in agreement with the previous studies of
Winther et al. (2002), and Leszkiewicz and Caputa (2004), who carried out re−
search in the regions of Ny−Ålesund and Hornsund, respectively.
The role of local wind at 6 m above the ground surface on the 2−m air tempera−
ture was studied in the period from July 2009 to June 2010 (Fig. 4). The highest air
temperatures were primarily found in the air flow from the southeast sector. In
these cases, mean air temperature varied between 0.2°C (south), −1.1°C (east), and
−1.6°C (southeast). Furthermore, relatively warm air flow exceeding −1.9°C was
documented from the northeast. This feature suggests a significant influence of
large−scale circulation, through which mild maritime air masses come along the
longitudinal axis of the Billefjorden, Ragnardalen, and Ebbadalen valleys. These
conditions frequently occurred during cyclonic circulation linked to types Ec, SEc
and NEc. Considerably cold air flow was recorded from the southwest to north−
west sectors. The lowest mean temperature (−8.4°C) was found during a north−
312 Kamil Láska et al.
Fig. 7. A. Scatter plot of daily mean 2−m air temperature at Petuniabukta versus Svalbard Lufthavn in the
period 2008–2010. Abbreviations on the graph: AT – air temperature, r – correlation coefficient, p – sig−
nificance level. B. Relative frequency of differences of daily mean 2−m air temperatures between
Petuniabukta and Svalbard Lufthavn, with the fitting of normal distribution function. Positive differ−
ences correspond to higher temperatures at Petuniabukta than in Svalbard Lufthavn and vice versa.
westerly wind, while winds from the southwest to west caused the temperature to
drop to −4.7°C and −7.6°C, respectively. This is also consistent with a gradual in−
crease in geostrophic winds from the northern and western sectors and high fre−
quency of the circulation types Nc (9.7% – January 2009), NWc (9.7% – January
2009), and SWc (22.6% – January 2010) during winter.
Comparison of air temperature between Petuniabukta and Svalbard
Lufthavn. — In order to evaluate the differences in weather conditions between
Petuniabukta and the nearest all−year round operating AWS (Svalbard Lufthavn),
2−m air temperature differences were calculated for the study period. Hereafter,
positive differences correspond to higher temperatures at Petuniabukta, while neg−
ative differences indicate a higher temperature at the Svalbard Lufthavn station.
As seen in Fig. 7 and Table 3, thermal differences confirmed the similar features of
Weather patterns of Petuniabukta, central Spitsbergen 313
Table 3
Seasonal variation of monthly means and extremes of 2−m air temperature differences be−
tween Petuniabukta and Svalbard Lufthavn in the period 2008–2010.
Month−Year
Air temperature difference Standard
deviation
Monthly
correlation
coefficient
Mean Min Max
[°C] [°C] [°C] [°C]
Aug−2008 −0.3 −3.0 1.5 1.2 0.67
Sep−2008 −0.8 −2.7 0.7 0.8 0.97
Oct−2008 −0.7 −3.5 2.0 1.3 0.94
Nov−2008 −1.0 −3.2 1.3 1.2 0.97
Dec−2008 −1.5 −5.0 2.3 1.7 0.95
Jan−2009 −1.9 −7.0 5.8 3.0 0.96
Feb−2009 −1.6 −4.7 1.1 1.5 0.97
Mar−2009 −0.7 −4.8 6.2 2.2 0.94
Apr−2009 −1.2 −4.0 2.9 1.6 0.94
May−2009 −0.3 −2.6 1.7 0.8 0.96
Jun−2009 −0.3 −2.6 1.9 1.1 0.88
Jul−2009 −0.5 −2.4 1.7 1.2 0.90
Aug−2009 −0.1 −2.8 2.6 1.0 0.86
Sep−2009 0.2 −2.0 1.7 0.8 0.94
Oct−2009 −0.7 −4.0 2.7 1.6 0.84
Nov−2009 −1.1 −4.7 1.2 1.3 0.95
Dec−2009 −1.2 −3.8 1.1 1.0 0.98
Jan−2010 −2.5 −6.7 0.3 1.7 0.98
Feb−2010 −1.8 −6.5 1.8 1.9 0.97
Mar−2010 −0.1 −4.6 4.9 2.6 0.77
Apr−2010 −1.8 −4.5 0.6 1.3 0.94
May−2010 −0.1 −2.0 1.1 0.8 0.94
Jun−2010 −0.3 −2.2 2.2 0.9 0.93
the local weather conditions between both sites. The mean temperature difference
was −0.9°C from August 1, 2008 to June 30, 2010. The monthly mean differences
ranged from −2.5 (January 2010) to 0.4°C (July 2010). The occurrence of positive
daily mean differences was significantly smaller than negative ones (Fig. 7). Due
to this fact, daily mean temperature at Svalbard Lufthavn was higher than that in
Petuniabukta (negative thermal differences) in 69.2% of the period. Conversely,
positive thermal differences, ranging from 0–1°C, occurred in 22.1% of the period.
The highest negative daily mean difference was observed in January, while the
highest positive ones in March. Seasonal variation of the thermal differences was
expressed by standard deviation and the correlation coefficient between 2−m air
temperatures at Petuniabukta and Svalbard Lufthavn. From Table 3, it is apparent
that the closest relationship between both time series occurred in winter and spring
(November–May), with correlation coefficients of 0.96 (2009) and 0.93 (2010) at
the significance level of 0.05. On the other hand, the highest negative differences
and standard deviations were found in the same period. The highest standard devi−
ations were reported in January 2009 (3.0°C) and March 2010 (2.6°C). A slightly
weaker connection was found in summer, with a correlation coefficient of 0.88
(June–August 2009). In spite of this fact, small standard deviations (interval of
1.0–1.2°C) and thermal differences, ranging from −0.3 to −0.1°C, were observed in
the same period. Similar findings were previously reported by Rachlewicz and
Styszyńska (2007) who compared air temperatures between Petuniabukta and
Svalbard Lufthavn in the period from July 7, 2001 to August 13, 2003.
Seasonal variation of the thermal differences could be partly explained by both
the effects of large−scale advection and the effects of local topography (e.g. blocking
effect of the mountains, drainage effect of the fjords), character of the ground surface
and their consequences on cloudiness variation and surface energy fluxes. This sug−
gests that the local wind pattern, which is controlled by thermodynamical processes
during advection of different air masses (large−scale weather systems), may also lead
to thermal differences at the ground surface. Therefore, the highest negative thermal
differences occurred during light winds from the western quadrant and strong radia−
tive cooling of the snow/ice surface. Less often, the highest negative differences
were observed for the northern and eastern sectors. Hence, the highest negative dif−
ferences (less than −5.0°C) occurred only in January and February, when 2−m air
temperatures at Petuniabukta dropped below −22°C. In these cases, local weather
conditions were affected by the northerly and northwesterly geostrophic flows
linked to circulation types NWa, with a thermal difference of −6.8°C (January 12,
2010), Nc and NWc, with differences of −6.3 and −6.5°C, respectively (January
12–14, 2009). Conversely, positive differences were observed during northeasterly
drainage flow from the Ragnarbreen and Ebbabreen glaciers and southerly flow
along the longitudinal axis of Billefjorden and Petuniabukta in particular. In these
situations, the local weather conditions at both sites were affected by circulation type
Ec, with a positive thermal difference of 6.3°C (March 24, 2009) and type NEa with
314 Kamil Láska et al.
the difference reaching up to 6.0°C (January 2, 2009). During these events, 2−m air
temperatures at Svalbard Lufthavn dropped below −20°C, probably due to the differ−
ent orientation of the fjord axis and gradual deceleration of the movement of the cold
air mass into the inner parts of the fjords and the coastal zones.
Apart from the above−mentioned results, we found that in many cases the tem−
perature differences between both sites can be affected by the sea ice conditions in
Isfjorden and Petuniabukta in particular. High resolution sea ice charts of the
Svalbard area were used for the comparison and evaluation of these effects. Data
on sea ice extent indicate that Petuniabukta was mainly covered with fast ice or
open drift ice (ice concentration between 1/10 and 4/10), which usually begin to
form from October to December, but due to strong winds and high waves was bro−
ken up and often disappeared. The inner parts of Billefjorden and Petuniabukta
were completely covered by sea ice mostly from January to June. In the study pe−
riod, sea ice formation (open drift ice) started in Petuniabukta on November 3,
2008 (October 2, 2009), while the middle part of Isfjorden (near the Svalbard
Lufthavn station) was covered by sea ice from December 12, 2008 (November, 24,
2009). Petuniabukta was completely covered with fast ice or closed drift ice (ice
concentration more than 7/10) from January 2, 2009 (January 10, 2010) to June 11,
2009 (June 22, 2010). In the same period, open drift ice mostly occurred in the
middle part of Isfjorden with a high variable of ice concentration between 1/10 and
7/10. In addition, ice−free situations appeared in this area, e.g. on January 19–22,
2009, January 12, 2010 and from March 12 to April 6, 2010. Hence, the negative
thermal differences (higher temperature at Svalbard Lufthavn station) from −1.5°C
to −6.8°C frequently occurred when the sea ice conditions significantly differed be−
tween both sites. This was also the situation on January 12–14, 2009 and January
12, 2010, with fast ice in Petuniabukta and open water (or open drift ice) in the
middle part of Isfjorden. On the other hand, thermal differences higher than −1.4°C
were affected by both the particular sea ice extent and the effects of large−scale
circulations.
Conclusions
In this paper, we have provided the first analysis and results of the fundamental
meteorological parameters measured at Petuniabukta in the period from August
2008 to June 2010. It has been shown that weather conditions in the central part of
Spitsbergen were significantly influenced by atmospheric circulation, cloudiness,
local topography, and sea ice extent. The results must be treated with caution due to
the fact that several factors, such as air pressure variation and precipitation, were not
considered. In general, our results confirm that synoptic−scale weather systems,
character of the ground surface and air−sea temperature differences play major roles
in forming local weather patterns at Petuniabukta. The influence of atmospheric cir−
Weather patterns of Petuniabukta, central Spitsbergen 315
culation and local wind conditions were clearly pronounced in the variations of air
temperature and surface wind speed. The highest air temperatures, ranging from
0.2°C to −1.9°C, were primarily found when winds were from the eastern and south−
ern quadrants, associated with predominantly cyclonic circulations (Ec, SEc and
NEc types). These large−scale flows were often modified by channelling and drain−
age effects accompanied by an increase in local wind speed. Therefore, strong kata−
batic winds in Petuniabukta were documented from the northeast and east towards
the Ragnarbreen and Ebbabreen glaciers. Such a wind mainly occurred in autumn
and winter, associated with circulation types Nc, NEc, Ec and Ea. In spite of this fact,
the lowest air temperature was observed during a flow from the western quadrant
with a mean wind speed of less than 3 ms−1. The occurrence of relatively cold air in
Petuniabukta can be explained by a combination of several factors: (1) a blocking ef−
fect of the mountains along the north−western coast of Petuniabukta, (2) strong trans−
formation and a long fetch of the flow over these mountains, and (3) strong radiative
cooling of the airflow from the snow and sea ice surface. A considerable close rela−
tionship between the ground surface and 2−m air temperatures was primarily re−
ported during summer than winter. Furthermore, thermal differences between
Petuniabukta and Svalbard Lufthavn confirmed that local weather conditions differ
significantly between these sites, with sea ice cover in the fjords, advection of air
masses from different directions, and the area of origin playing essential roles. The
highest thermal differences between the sites were mainly affected by differences in
sea ice concentrations and ice types in Petuniabukta in contrast to the middle part of
Isfjorden close to the Svalbard Lufthavn station.
Acknowledgements. — This work was funded and supported by the project “Biological
and climate diversity of the central part of the Svalbard Arctic archipelago” funded by the Min−
istry of Education, Youth and Sport of the Czech Republic, INGO−LA 341, project LM2010009
CzechPolar, and project of the EHP No. B/CZ0046/3/0013 “Effects of climate change on tundra
ecosystem: Interaction of vegetation with extreme environments”. The work of Denisa Wito−
szová was supported by the project of Masaryk University (MUNI/A/0966/2009) “Expression
of Global Environmental Change in Component Earth's Spheres” (PROGLEZ). The authors are
very grateful to members of the Czech Arctic Expeditions for their field assistance and Øyvind
Nordli from the Norwegian Meteorological Institute for providing meteorological data for the
Svalbard Lufthavn station. We would also like to thank Keith Edwards for language revision
and three anonymous reviewers for their helpful comments and suggestions to improve the
manuscript.
References
BARANOWSKI S. and GŁOWICKI B. 1975. Meteorological and hydrological investigations in the
Hornsund region made in 1970. Acta Universitatis Wratislaviensis 251: 35–39.
BARRY R.B. 2008. Mountain Weather and Climate. Cambridge University Press, Cambridge: 506 pp.
BEDNORZ E. and KOLENDOWICZ L. 2010. Summer 2009 thermal and bioclimatic conditions in Ebba
Valley, central Spitsbergen. Polish Polar Research 31 (4): 327–348.
316 Kamil Láska et al.
BRÁZDIL R. 1988. Variation of air temperature and atmospheric precipitation in the region of
Svalbard. In: R. Brázdil (eds) Results of Investigations of the Geographical Research Expedition
Spitsbergen 1985. J.E. Purkyne University, Brno: 187–210.
BRÁZDIL R., PROŠEK P., PACZOS S. and SIWEK K. 1991. Comparison of meteorological conditions in
Calypsobyen and Reindalen in summer 1990. Wyprawy Geograficzne na Spitsbergen. UMCS,
Lublin: 57–76.
ESAU I. and REPINA I. 2012. Wind climate in Kongsfjorden, Svalbard, and attribution of leading
wind driving mechanisms through turbulence−resolving simulations. Advances in Meteorology
2012, Article ID 568454: 1–16.
FØRLAND E.J. and HANSSEN−BAUER I. 2003. Past and future climate variations in the Norwegian
Arctic: overview and novel analyses. Polar Research 22: 113–124.
HANSSEN−BAUER I., SOLÚS M.K. and STEFFENSEN E.L. 1990. The Climate of Spitsbergen. DNMI
Rep. 39/90 Klima: 40 pp.
KEJNA M. and DZIENISZEWSKI M. 1993. Meteorological conditions on the Kaffiöyra Plain (NW
Spitsbergen) from 26th June to 31st August 1985. Acta Universitatis Nicolai Copernici. Geografia
24: 43–54 (in Polish).
KILPELÄINEN T., VIHMA T. and ÓLAFSSON H. 2011. Modelling of spatial variability and topo−
graphic effects over Arctic fjords in Svalbard. Tellus A 63 (2): 223–237.
KNUTSON T.R., DELWORTH T.L., DIXON K.W., HELD I.M., LU J., RAMASWAMY V., SCHWARZKOPF
M.D., STENCHIKOV G. and STOUFFER R.J. 2006. Assessment of twentieth−century regional surface
temperature trends using the GFDL CM2 coupled models. Journal of Climate 19 (9): 1624–1651.
KOSIBA A. 1960. Some of results of glaciological investigations in SW−Spitsbergen carried out dur−
ing the Polish I.G.Y. Spitsbergen expeditions in 1957, 1958 and 1959. Zeszyty Naukowe, Nauki
Przyrodnicze, Uniwersytet Wrocławski im. Bolesława Bieruta, Ser. B 4: 29 pp.
KOSTRZEWSKI A., KANIECKI A., KAPUŚCIŃSKI J., KLIMCZAK R., STACH A. and ZWOLIŃSKI Z.
1989. The dynamics and rate of denudation of glaciated and non−glaciated catchments, central
Spitsbergen. Polish Polar Research 10 (3): 317–367.
KRUSZEWSKI G. 2006. The changes of zonal wind speed component (U−wind) at the West Spits−
bergen area (1981–2005). Problemy Klimatologii Polarnej 16: 107–114 (in Polish).
LÁSKA K., WITOSZOVÁ D. and PROŠEK P. 2010. Climate Conditions of the Permafrost Active Layer
Development in Petuniabukta, Billefjorden, Spitsbergen. In: J.R. Mertes, H.H. Christiansen and
B. Etzelmüller (eds.) Thermal State of Frozen Ground in a Changing Climate During the IPY,
Longyearbyen: 128–128.
LESZKIEWICZ J. and CAPUTA Z. 2004. The thermal condition of the active layer in the permafrost at
Hornsund, Spitsbergen. Polish Polar Research 25 (3–4): 223–239.
MARSZ A. 2007. Air temperature. In: A. Marsz, A. Styszyńska (eds) Klimat rejonu Polskiej Stacji
Polarnej w Hornsundzie. Wydawnictwo Akademii Morskiej w Gdyni: 131–174 (in Polish).
MÄKIRANTA E., VIHMA T., SJÖBLOM A. and TASTULA E.M. 2011. Observations and modelling of
the atmospheric boundary layer over sea−ice in a Svalbard fjord. Boundary−Layer Meteorology
140 (1): 105–123.
MIGAŁA K., NASIÓŁKOWSKI T. and PEREYMA J. 2008. Topoclimatic conditions in the Hornsund
area (SW Spitsbergen) during the ablation season 2005. Polish Polar Research 29 (1): 73–91.
MILLER G.H., BRIGHAM−GRETTE J., ALLEY R.B., ANDERSON L., BAUCH H.A., DOUGLAS M.S.V.,
EDWARDS M.E., ELIAS S.A., FINNEY B.P., FITZPATRICK J.J., FUNDER S.V., HERBERT T.D.,
HINZMAN L.D., KAUFMAN D.S., MACDONALD G.M., POLYAK L., ROBOCK A., SERREZE M.C.,
SMOL J.P., SPIELHAGEN R., WHITE J.W.C., WOLFE A.P. and WOLFF E.W. 2010. Temperature
and precipitation history of the Arctic. Quaternary Science Reviews 29 (15–16): 1679–1715.
MOREAU M., LAFFLY D. and BROSSARD T. 2009. Recent spatial development of Svalbard strandflat
vegetation over a period of 31 year. Polar Research 28: 364–375.
Weather patterns of Petuniabukta, central Spitsbergen 317
NIEDŹWIEDŹ T. 2007. Atmospheric circulation. In: A. Marsz and A. Styszyńska (eds) Klimat rejonu
Polskiej Stacji Polarnej w Hornsundzie. Wydawnictwo Akademii Morskiej w Gdyni: 45–64 (in
Polish).
NIEDŹWIEDŹ T. 2012. The calendar of atmospheric circulation types for Spitsbergen, a computer file,
University of Silesia, Department of Climatology, Sosnowiec, Poland. http://klimat.wnoz.us.
edu.pl/osoby/tn/Spitsbergen.zip
NILSEN F., COTTIER F.R., SKOGSETH R. and MATTSSON S. 2008. Fjord−shelf exchanges controlled
by ice and brine production: The interannual variation of Atlantic Water in Isfjorden, Svalbard.
Continental Shelf Research 28 (14): 1838–1853.
ØRBÆK J.B., HISDAL V. and SVAASAND L.E. 1999. Radiation climate variability in Svalbard: sur−
face and satellite observations. Polar Research 18 (2): 127–134.
PEREYMA J. and PIASECKI J. 1988. Topoclimatic and hydrological conditions in the Werenskiold Gla−
cier region Spitsbergen in the summer−autumn season of 1983. In: Wyprawy Polarne Uniwersytetu
Śląskiego 1980–1984. Prace Naukowe Uniwersytetu Śląskiego 910: 107–122 (in Polish).
PRACH K., KLIMEŠOVÁ J., KOŠNAR J., REDCHENKO O. and HAIS M. 2012. Variability of contemporary
vegetation around Petuniabukta, central Spitsbergen. Polish Polar Research 33 (4): 383–394.
PRZYBYLAK R. 1992. Spatial differentiation of air temperature and relative humidity on the western
coast of Spitsbergen in 1979–1983. Polish Polar Research 13 (2): 113–130.
PRZYBYLAK R. 2000. Temporal and spatial variation of surface air temperature over the period of in−
strumental observations in the Arctic. International Journal of Climatology 20: 587–614.
PRZYBYLAK R. and ARAŹNY A. 2006. Climatic conditions of the north−western part of Oscar II Land
(Spitsbergen) in the period between 1975 and 2000. Polish Polar Research 27 (2): 133–152.
PRZYBYLAK R. and DZIERŻAWSKI J. 2004. Thermal and humidity relations in Treurenberg Bay and
Massif Olimp (NE Spitsbergen) From 1st August 1899 to 15th August 1900. Problemy Klimato−
logii Polarnej 14: 133–147 (in Polish).
RACHLEWICZ G. 2003. Meteorological conditions in the Petunia Bay (central Spitsbergen) in sum−
mer seasons 2000–2001. Problemy Klimatologii Polarnej 13: 127–138 (in Polish).
RACHLEWICZ G. and STYSZYŃSKA A. 2007. Comparison of the air temperature course in Petunia
Bukta and Svalbard−Lufthavn (Isfjord, West Spitsbergen). Problemy Klimatologii Polarnej 17:
121–134 (in Polish).
RACHLEWICZ G. and SZCZUCIŃSKI W. 2008. Changes in thermal structure of permafrost active layer
in a dry polar climate, Petuniabukta, Svalbard. Polish Polar Research 29 (3): 261–278.
SERREZE M.C., BOX R.G., BARRY R.G. and WALSH J.E. 1993. Characteristics of Arctic synoptic ac−
tivity. Meteorology and Atmospheric Physics 51 (3): 147–164.
STATISTICA 2011. STATISTICA software. Available online at: http://www.statsoft.com/products/
statistica−product−catalog/ and http://www.statsoft.com/textbook/ (accessed 25 August 2011).
STORCH H. and ZWIERS F. 1999. Statistical Analysis in Climate Research. Cambridge University
Press: 496 pp.
SVENDSEN H., BESZCZYNSKA−MØLLER A, HAGEN J.O., LEFAUCONNIER B., TVERBERG V., GERLAND
S., ORBÆK J.B., BISCHOF K., PAPUCCI C., ZAJĄCZKOWSKI M., AZZOLINI R., BRULAND O.,
WIENCKE CH., WINTHER J.G. and DALLMANN W. 2002. The physical environment of Kongs−
fjorden–Krossfjorden, an Arctic fjord system in Svalbard. Polar Research 21 (1): 133–166.
WINTHER J.G., GODTLIEBSEN F., GERLAND S. and ISACHSEN P.E. 2002. Surface albedo in Ny−
Ålesund, Svalbard: variability and trends during 1981–1997. Global and Planety Change 32 (2):
127–139.
ZIAJA W. 2005. Response of the Nordenskiöld Land (Spitsbergen) glaciers Grumantbreen, Håbergbreen
and Dryadbreen to the climate warming after the Little Ice Age. Annals of Glaciology 42 (1):
189–195.
Received 18 October 2011
Accepted 23 October 2012
318 Kamil Láska et al.
10
Paper 9
High-resolution numerical simulation of summer wind field
over complex topography of Svalbard archipelago
Láska, K., Chládová, Z., Hošek, J., submmited,
Quarterly Journal of the Royal Meteorological Society
1
High-resolution numerical simulation of summer wind field over
complex topography of Svalbard archipelago
Kamil Láska1,2
, Zuzana Chládová2,3
, Jiří Hošek3
1
Department of Geography, Faculty of Science, Masaryk University, Kotlářská 2, 611 37,
Brno, Czech Republic
2
Centre for Polar Ecology, Department of Ecosystem Biology, Faculty of Science, University
of South Bohemia, Na Zlaté stoce 3, 370 05, České Budějovice, Czech Republic
3
Institute of Atmospheric Physics, Academy of Sciences of the Czech Republic, Boční
II/1401, Praha, Czech Republic
Corresponding author:
Kamil Láska (laska@sci.muni.cz)
Running head:
Summer wind field simulation on Svalbard
Keywords:
Svalbard; surface wind field; model evaluation; topographic effect; circulation pattern
2
Abstract
The Weather Research and Forecasting (WRF) mesoscale model has been run in three different
configurations over Svalbard archipelago and compared with the 12-day summer measurements of
surface wind characteristics at three sites along the western coast of Petuniabukta, central Spitsbergen.
For studying wind pattern over complex topography we chose the sites differing in terrain elevation
and local surface characteristics as follows: raised marine terrace (15 m a.s.l.), foreland of
Hørbyebreen glacier (67 m a.s.l.) and top of Mumien Peak (773 m a.s.l.). The WRF simulations were
conducted using three boundary layer parameterization schemes, the Yonsei University (YSU), the
Mellor-Yamada-Janjic (MYJ) and the Quasi-Normal Scale Elimination (QNSE) schemes with 1-km
horizontal resolution of the inner domain. The WRF simulations agreed fairly well with the surface
wind observations taken at all the sites. For the wind speed, the mean correlation coefficients between
the modelled and observed data ranged from 0.56 to 0.67. The best results across all stations were
found for the QNSE parameterization scheme with the bias of ~0.1 m s-1
. The wind speed simulations
were sensitive to geographical location and elevation of the stations. All parameterization schemes had
difficulties in capturing of surface wind field in the narrow valley with the Hørbye foreland station,
while satisfactory estimates were found at the top of the Mumien Peak and the Terrace station located
in a wide part of the fjord. The WRF estimates proved to be highly sensitive to the large-scale forcing,
as documented for the cyclonic circulation patterns with a strong northerly wind. It led to an
overestimation of the modelled wind speed on the leeward slopes of the highest peaks in the study
area. The model results tended to be underestimated during the anticyclonic situations with the
induced local circulation systems between the fjord water and surrounding valleys.
3
1. Introduction
Over the last few decades, an interest in the Arctic region has increased as it has been variety
of environmental impacts caused by climate change and positive snow and sea ice albedo
feedbacks amplifying the warming signal (Holland and Bitz, 2003; Serreze and Barry, 2011;
West et al., 2013). The largest atmospheric warming in the Atlantic region of the Arctic over
the last 50 years was reported from the Svalbard archipelago (Przybylak, 2007; Nordli et al.,
2014). Moreover, the largest future temperature increase in the Arctic is predicted in the 21st
century by a large number of climate models (Collins et al., 2013). The quality of climate
projections and numerical model outputs is closely related to the input of observational data
and ability to simulate atmospheric processes at sufficient resolution in space and time.
Among efficient and prospective tools in evaluation of the atmospheric circulation patterns
and their coupling with the land surface features, numerical weather prediction (NWP)
models represents reasonable option. The current NWP models allow to analyse the regional
and local circulation conditions and provide relatively accurate estimation of physical
properties of the atmospheric boundary layer (ABL) at high temporal (in order of minutes)
and spatial resolution (in order of kilometres).
In contrast to mid latitudes, the Arctic observations and measurements are often
insufficient as they do not cover whole region uniformly without data gaps. Additionally, they
require advanced computations or data transformation (Przybylak, 2003; Serreze and Barry,
2009). From this point of view, the Arctic Svalbard archipelago represents one of suitable and
easily accessible locations that enable to carry out both advanced measuring campaigns and
validate a performance of the NWP models under various boundary layer (BL) conditions.
However, the ABL processes over Svalbard fjords are strongly influenced by complex
topography and sea ice occurrence (Vihma et al., 2014). Due to specific vertical structure of
the Arctic atmosphere (a prevailing stable stratification), non-linear effects on flows over
  4
topography of the Svalbard archipelago are comparable to the non-linear effects on flows
across much higher topography elsewhere in the world (Skeie and Grønås, 2000). The
topographic effects over Svalbard include commonly known processes such as barrier winds,
foehn winds, drainage flows and mountain waves (Sandvick and Furevick, 2002; Mäkiranta et
al., 2011; Kilpeläinen et al., 2011; Láska et al., 2012; Vihma et al., 2014). Modelling of these
features in case of complex topography requires specific attention paid to model setup and
simulation strategy (Vihma et al., 2014). Another requirement is a selection of the correct
ABL scheme which fits the specific conditions of the high latitude location and allows for the
use of high grid resolution. Many authors claim that the Weather Research and Forecasting
(WRF) mesoscale model can provide dynamical downscaling techniques with a variety of
physical options and the BL parameterization schemes as well as accurate enough
representation of the terrain properties (e.g., Michalakes et al., 2004; Skamarock et al., 2008;
Hines and Bromwich, 2008; Mayer et al., 2012).
Numerous measuring campaigns associated with the WRF modelling have been
carried out over the Svalbard archipelago during last few years. For example, Mäkiranta et al.
(2011) compared WRF simulations with sonic anemometer and profile mast measurements in
Wahlenbergfjorden (Nordaustlandet Island) during few weeks in May 2006 and April 2007.
The estimates for 2-m air temperature and friction velocity were in a good agreement with
reality, but the model failed to reproduce the spatial variability in wind direction between
measurement sites 3 km apart. Kilpeläinen et al. (2011) compared model results from the
standard WRF model to tower observations and radiosoundings in Isfjorden and
Kongsfjorden (Spitsbergen Island) for 10 selected days in winter and spring 2008. The
comparison of model data with tower observations and radiosoundings showed fairly good
agreement, although a systematic warm bias and slightly overestimated wind speeds close to
the surface were found. Livik (2011) evaluated WRF and compared with measurements from
  5
the automatic weather stations placed at several locations along Kongsfjorden-Kongsvegen
valley in spring 2010. The author focused on evaluation of differences in the wind
characteristics in topographically deflected flows and the pure katabatic flows. Kilpeläinen et
al. (2012) compared both standard and polar version of the WRF model to tethered balloon
soundings and mast observations taken in Isfjorden and Kongsfjorden during two weeks in
March and April 2009. The difference in performance of the standard WRF and Polar WRF
was marginal, and the biases in modelled wind speed profiles were closely related to lowlevel
jets. The modelled low-level jets were stronger and deeper than observed ones.
Moreover, modelled jets were located at higher altitudes than the observed jets. Similarly to
this results, Mayer et al. (2012) evaluated the three BL parameterization schemes embedded
in the WRF against the tethered balloon soundings and the Small Unmanned Meteorological
Observer – SUMO (Reuder et al., 2009) aircraft flights in Longyerbyen area in March–April
2009. The WRF wind profiles showed good agreement with the SUMO measurements with
the bias about 1 m s−1
from 200 to 800 m a.s.l. Roberts et al. (2015) compared three BL
schemes in the WRF with observations made by controlled meteorological balloons launched
from Ny-Ålesund (Kongsfjorden) in the period 5–12 May 2011. The largest errors were found
in the meteorological parameters over sea ice in comparison to the coastal area close to NyÅlesund.
The model in general overestimated the wind speed and showed inability to capture
temperature inversions and low-level jets. As can be seen from the introductory section, in
spite of the progress in ABL parameterization for specific or marginal conditions, model
simulations in the Arctic are still not satisfactory. This is true particularly for some
atmospheric processes and variables, near-surface wind characteristics in particular.
Although a number of numerical simulations on the Svalbard archipelago increased in
time, most of the measuring campaigns and related WRF simulations were carried out in
spring months in the area of Ny-Ålesund and Longyearbyen. Information about the wind field
6
structure from the other part of the Svalbard archipelago are still insufficient. Therefore, the
aim of our study was to use detailed in situ summer measurements from Petuniabukta in
central Spitsbergen to 1) analyse the temporal and spatial variation in the near-surface wind
field, 2) evaluate the WRF model performance using different BL parameterization schemes,
especially in relation to the local topography and terrain properties, and 3) identify model
weaknesses. For the latter purpose, large-scale forcing and atmospheric circulation patterns
leading to the best and the worst simulations of surface wind characteristics were analysed.
Hence, the WRF simulation results were further evaluated using local meteorological data
acquired from the three automatic weather stations situated within a small-scale area. The
850-hPa geopotential heights and corresponding geostrophic winds available from the
reanalysis datasets were used as well.
2. Material and Methods
2.1. Study site and observation
The field measurements were carried out in the central part of Spitsbergen Island, which is a
part of the Svalbard Arctic archipelago (Figure 1(a)). We focused on the coastal zone of
Petuniabukta, where complex topography and local surface characteristics play a major role in
surface wind speed and direction variability (Láska et al., 2012; Láska et al., 2013).
Petuniabukta is a northward oriented bay that closes the northern part of Billefjorden (Figure
1(b)). At the study site, the terrain complexity is formed by numerous mountain ridges and
peaks reaching an altitude of 600-1000 m (Pyramiden, Mumien, Svenbrehøgda, Sfinksen,
Lovehøvden), glaciers (Bertilbreen, Ferdinandbreen, Svenbreen, Hørbyebreen, Ragnarbreen),
their glacial forelands and valleys ending into the Petuniabukta. The coastal zone of the bay is
mainly formed by glacial landsystems, comprising moraines, esker, debris ridges, bare soils
(cryosols/gelisols type), shallow pools, seepages, and sporadic tundra vegetation (Rachlewicz,
  7
2010; Evans et al., 2012; Prach et al. 2012; Hanáček et al., 2013; Ewertowski and Tomczyk,
2015). The study area belongs to the maritime high Arctic climate region (Nordli et al., 2014).
The area is mainly influenced by katabatic drainage flows caused by the local topography and
land-sea breeze circulation (Kilpeläinen et al., 2011; Esau and Repina, 2012). The
meteorological observations from Petuniabukta (2010-2012) showed that the prevailing wind
direction at the coast was from the southern and northern quadrant, while the strongest winds
were connected to northeasterly flow from the Ragnarbreen and Ebbabreen glaciers (Láska et
al., 2012).
In order to investigate local wind systems and evaluate the numerical model outputs,
three automatic weather stations (Terrace, Hørbye foreland, Mumien Peak) were set up at
sites with different elevation, local topography, and surface properties. The lowest station
Terrace was placed on the flat marine terrace, approximately 500 m northwest of the
Petuniabukta coast and 1100 m from the foothill of the Mumien mountain (Figure 1(c), Table
1). At the Terrace site, the fjord is 4 km wide and oriented North-South. The second station
was located at the foreland of Hørbyebreen glacier, which is composed by ice-cored laterfrontal
moraines with surge-related landforms (Evans et al., 2012). The observation site was
300 m southeast from the glacier snout wherein the valley narrows to 2 km. The third station
was situated at the top of Mumien Peak (773 m a.s.l.) which is the second highest summit in
the study area after Pyramiden (937 m a.s.l.). The top part of Mumien Peak has a round shape
slightly elongated towards the Pyramiden mountain located 3 km south. The geographical
position of the stations and their surface properties are given in Table 1.
At all three stations, the measurement of wind speed and wind direction was recorded
with the same set of a MetOne 034B Wind Sensor (MetOne, USA) with starting threshold at
0.4 m s-1
for both output parameters. Sampling and recording time intervals were set
identically at every 30 min and data were stored to an EdgeBox V12 datalogger (EMS, Czech
  8
Republic). The heights of the wind sensors, used in this study, were 6.0 m above ground for
the Terrace station, and 3.0 m for the Hørbye foreland, and Mumien Peak stations,
respectively. The experimental measurements were carried out in the period of 9 to 22 July
2013, i.e. typical summer season in the study area. However, the occurrence of icing in the
wind speed data was reported at the highest station at Mumien Peak; therefore, such records
have been removed or replaced by cross correlation using data measured concurrently at
Petuniabukta.
2.2. Modelling approach and simulation design
The Advanced Research WRF mesoscale model (ARW) version 3.5 was used for the
simulations in this study. The ARW is a state-of-the-art mesoscale non-hydrostatic numerical
weather prediction model developed at the National Center for Atmospheric Research
(NCAR) intended to be used in research tasks. It employs a fully compressible dynamical
core and terrain-following hydrostatic-pressure vertical coordinates. More details of the
model can be found in Skamarock et al. (2008). The polar version of WRF has not been
included in the simulation design because of the timing of the wind observations conducted
only in summer, without the presence of sea ice, stable temperature stratification or other
wintertime conditions in the fjords, where the polar WRF can benefits from the more
sophisticated treatment of the sea ice zone and related processes (Tatsula and Vihma, 2011).
In this simulation, the WRF model was configured with three one-way nested domains
in polar stereographic projection. The domains have a horizontal grid spacing of 9, 3 and 1
km from the outermost to the innermost domain (Figure 1(b)). All three domains contain
94×94 grid points and are centered at 78.71°N and 16.46°E. The outer domain D1 covers a
large area ranging from the northeast corner of Greenland in the west to Franz Josef Land in
the east. The middle domain D2 covers the large part of the Svalbard archipelago. The inner
  9
domain D3 covers the smaller fjords, Billefjorden, Dicksonfjorden and part of Austfjorden,
situated on the north coast of Spitsbergen, the largest island of the Svalbard archipelago and
captures the local complex topography at 1-km resolution. In all domains, the model variables
are estimated at 61 levels on vertical axis, of which 18 are normally in the lowest 1000 m of
the atmosphere; the lowest full model level is approximately at 7.4 m above ground. The
model top is set to 10 hPa. The model is initialized with ERA-Interim reanalysis data issued
by the European Centre for Medium-Range Weather Forecasts (ECMWF). The input
meteorological data are defined at standard pressure levels and have 0.5° horizontal
resolution. The lateral boundary conditions are updated at 6-hour intervals during the
simulation. The model runs with time steps of 60 s for the outer domain D1, 20 s for the
domain D2 and 4 s for the inner domain D3. During the simulations, Billefjorden and other
fjords were free of the sea ice. The sea surface temperature was set according to boundary
conditions from the ERA-Interim reanalysis data.
The aim of this study was to investigate the performance of the different BL schemes
in the WRF model when compared to near-surface measurements from three meteorological
stations in a small-scale area of Petuniabukta. Therefore the BL parameterization was crucial
and corresponded to the combinations normally tested in the Arctic regions (e.g. Kilpeläinen
et al., 2012; Hines and Bromwich, 2008). The two most common ABL parameterizations
appearing in the studies are the Yonsei University (YSU) scheme and the Mellor-YamadaJanjic
(MYJ) scheme. These are completed by the Quasi-Normal Scale Elimination (QNSE)
scheme applied for example in Mayer et al. (2012) and Kilpeläinen et al. (2012) on Svalbard
fjords. The surface fluxes over land and sea ice are provided by the Noah Land Surface Model
(LSM) (Chen and Dudhia, 2001), which calculates soil temperature and moisture in four
layers. It can be used for prediction of snow cover, which can be fractional, and moisture in
10
vegetation cover. The scheme employs frozen soil physics. The skin temperature is
diagnostic.
The YSU scheme (Skamarock et al., 2008) is a first order approach based on non-local
K-mixing in a dry convective boundary layer. The depth of the ABL is a function of the
temperature profile. The entrainment layer is treated explicitly. In the recent versions of the
WRF, a treatment of the stable boundary layer was significantly updated. In arctic regions,
however, it is reported to perform poorer than other options (Kilpeläinen et al., 2012). The
MYJ scheme (Janjic 1996; Janjic 2002) has a 1.5-order and 2.5-level closure of turbulence. It
predicts turbulent kinetic energy (TKE) in one dimension through local vertical mixing. The
scheme is normally used in the North American Mesoscale Forecast System developed in the
operational version of WRF. It is often considered the best option in the Arctic regions (Hines
and Bromwich, 2008). The QNSE scheme (Sukoriansky et al., 2005) follows a similar
approach to MYJ. It also uses 1.5-order closure and local TKE-based vertical mixing. The
main difference is a new theory that is used for the simulation of turbulence with stable and
weakly unstable stratified atmosphere. At present, the QNSE scheme has become popular and
shows promising results also in the Arctic region (Kilpeläinen et al., 2012; Roberts et al.,
2015). The parameterization of other processes affects the surface wind conditions
considerably less, especially in summer season when the sea ice is not present. They were,
therefore, not considered of primary importance. Long-wave and short-wave radiation is
parameterized with the updated version of the Rapid Radiative Transfer Model (RRTMG)
(Mlawer et al., 1997). The new version of Thompson scheme is used for cloud microphysics
(Thompson et al., 2008). The Kain-Fritsch cumulus scheme (Kain, 2004) is applied in the
outermost domain, in the other domains, convection is resolved explicitly.
Finally, it is important to state that our work was primarily focused on the evaluation
of the WRF performance with the three most widely used BL schemes in the Arctic region
  11
∑ 2
  
and wind estimation errors occurring during different circulation patterns at the selected
stations. Other surface variables, their spatiotemporal distribution and associated errors of the
WRF model were not investigated in this study.
2.3. Statistical evaluation
The 12-days model simulation was verified based on experimental observations from three
meteorological stations and different statistics. For wind speed we have used the Pearson’s
correlation coefficient (r), bias and root-mean-square error (RMSE). The bias quantifies the
overall bias of the observational wind speed and it can be used for the detection of a
systematic under or overestimation by the model:
∑
, 
(1)
where m and o are simulated and observed wind speeds and N is the number of records.
The RMSE represents the sample standard deviation:
. (2)
The wind estimates were calculated for the grid point nearest to each meteorological station
which represented the best agreement with actual terrain altitude and land cover properties.
Bias and RMSE values, i.e. the observations subtracted from the modelled data were
calculated at each station for the whole period, and then averaged over the main wind
directions. The periods with data missing for one of the locations were excluded from the
error calculations. For verification of wind direction data, the directional accuracy parameter
according to Santos-Alamillos et al. (2013) was used. The directional accuracy of circular
distance (DACD) is based on the circular distance defined as the smaller of the two arc
12
lengths between two points along the circumference. For two angles α and β a circular
distance is defined as follows (Jammalamadaka and Sengupta, 2001):
∆ , min | |, 360 | | .
(3)
DACD accounts the percentage of times for which the circular distance between the observed
and modelled data is less than a threshold, chosen as 30° (Santos-Alamillos et al., 2013), then:
. (4)
For the purpose of comparison of the wind characteristics, the wind direction data were
classified with a precision of 45° resulting in eight main sectors for which the wind rose
charts were constructed.
3. Results and Discussion
3.1. Atmospheric circulation and weather patterns
The synoptic situations and prevailing circulation patterns were identified according to 6-hour
geopotential height data extracted from the 850-hPa level of the ERA-Interim reanalysis
project (Dee et al., 2011). At the beginning of the studied period (9–11 July), an extensive
cyclone had its centre east of the Svalbard archipelago (Figure 2(a)) and moved firstly
northward to the Nordaustlandet, and then southeast (Figure 2(b)) towards the Novaya
Zemlya while weakening. The situation caused northwesterly and later northeasterly winds in
central Svalbard at 850-hPa ranging from 5 to 10 m s-1
, which carried cold air from the north.
On 11 July the wind gradually weakened to 4 m s-1
(Figure 2(b)), followed by anticyclonic
circulation that occurred at the 850-hPa level for the next 3 days. On 12 July, the ridge
extending from the high pressure above the north Scandinavia spread over the Svalbard
archipelago. It slowly weakened and was replaced by a high-pressure system from the North
13
Pole on 13 July (Figure 2(c)). In central Svalbard the northerly or northeasterly flow prevailed
with wind speeds below 3 m s-1
. In the third part of the studied period (15–17 July), Svalbard
weather was influenced by intense cyclonic circulation caused by a deep cyclone over the
Novaya Zemlya moving towards the Svalbard archipelago (Figure 2(d)). It later changed the
position and moved towards the Franz Joseph Land (16 July). This weather situation caused
relatively strong pressure gradient in the study area and moderate northwesterly or northerly
winds at 850-hPa between 8 m s-1
(17 July) and 11 m s-1
(15 July). On July 17 the lowpressure
system moved towards the North Pole and slowly disappeared. Instead, another
anticyclonic situation formed over the Svalbard archipelago in the period of 18–19 July. The
high-pressure ridge from the Novaya Zemlya approached the study area from the east (Figure
2(e)). It gradually weakened during the next day and was replaced by a large low pressure
system developing over the Scandinavian Peninsula. The wind speed in the Svalbard region
was very low (<3 m s-1
) and the wind direction ranged from east to southeast, however, the
wind speed gradually increased in the afternoon of 19 July. In the last part of the studied
period (20–22 July) the low pressure system moved from north Scandinavia towards Svalbard
and further to the North Pole. It lead to strong pressure gradients at the 850-hPa level in the
studied area causing an easterly to southeasterly flow with the highest wind speed (12 m s-1
)
on 20 July. Afterwards the wind direction changed towards west as the wind prevailed
southwestern and western varying between 9 to 11 m s-1
(Figure 2(f)).
3.2. Spatial variability of observed surface winds
Figure 3(a) shows large variability of surface wind speed during experimental observations
between all stations. Mean wind speed during the studied period reached 4.3, 3.2 and 3.6 m s-1
for Terrace, Hørbye foreland and Mumien Peak, respectively. The maximal wind speed at the
given stations ranged between 9.8 and 13.5 m s-1
. The strongest winds at the low-situated
  14
locations (Terrace and Hørbye foreland) can be generally accounted to the northerly flow
(Figure 3(b)) from the Hørbyebreen and Ragnarbreen glaciers. In our case, maximal wind
speed was observed on 11 July 2013 recorded as 9.8 m s-1
at Hørbye foreland and 10.2 m s-1
at Terrace. On this day a strong pressure gradient influenced the studied area. It resulted from
a large low-pressure system with the centre east of the Svalbard archipelago. A similar
circulation pattern with strong winds was also observed at all stations on 15 July 2013. Due to
deep low-pressure system and northwesterly flow at 850-hPa level, the surface wind speed
reached high maximal values of 9.3, 7.8 and 10.7 m s-1
at Terrace, Hørbye foreland and the
Mumien Peak, respectively (Figure 3(a)). The highest overall wind speed during the campaign
was recorded at the Mumien Peak on 22 July 2013, exceeding 13.5 m s-1
. During this event, a
low-pressure system with strong pressure gradients passed between the Svalbard archipelago
and Greenland towards the North Pole, while southerly surface wind was observed at the
Hørbye foreland station and southeasterly winds at the other stations (Figure 3(b)). Lower
wind speeds (<4 m s-1
) were observed mainly during three periods, namely 9–10 July, 12–13
July and 17–19 July. Although the low wind speed at the first period was influenced by a
weakened cyclone located northeast of the Svalbard archipelago, the next two cases were
caused by anticyclonic circulation in connection with a high-pressure ridge spreading over
Svalbard (Figure 2(c), 2(e)). The wind direction highly fluctuated between southern,
northwestern, and northern directions at all stations (Figure 3(b)). This confirmed high
sensitivity of surface wind field to the anticyclonic circulation patterns.
The strong influence of local topography on the surface wind pattern in Petuniabukta
is also apparent in Figure 4. The relative frequency of wind direction in the 12-days period
differed significantly between all stations. The prevailing wind directions were observed from
the northeast (30%) and north (23%) at Terrace station, northwest (24%) and north (21%) at
Hørbye foreland, and north (30%) and northeast (18%) at the Mumien Peak. On the other
  15
hand, westerly and easterly winds rarely occurred at the Terrace and Hørbye foreland stations
with a frequency of 15% (west) and 5% (east) at given stations. Morphology of the
surrounding topography clearly corresponded with the wind rose pattern at the Hørbye
foreland as well as Terrace station, increasing the frequency of the wind direction fitting the
longitudinal axis of the Billefjorden with Petuniabukta (see Figures 1(c) and 4) and the side
valleys with the Hørbyebreen and Ragnarbreen glaciers (in case of the Terrace station).
Conversely, the wind direction distribution and occurrence of strong winds at the Mumien
Peak confirmed that the air flow at the exposed and elevated sites is less disturbed by terrain
in comparison to the other sites affected by channelling effect of the valley and drainage flow.
3.3. Spatial and temporal variation in modelled wind speed
For the purpose of the WRF output evaluation, the time series of modelled surface level (10
m) wind from the inner domain was extracted at the grid point nearest to the position of each
meteorological station in Petuniabukta. Differences between actual and modelled terrain
elevation for all stations are given in Table 1. The highest elevation differences were found
for the Mumien Peak and the Hørbye foreland stations, resulting from the digital elevation
model (DEM) with a 1-km spatial resolution. The wind speed data were adjusted according to
the actual height of wind instruments located below the standard height of 10 m. The neutral
logarithmic wind profile was applied on the lowest model level data (ca 7.4 m above ground)
with a surface roughness length of z0=0.015 m identical for each station. The chosen value of
roughness parameter is typical for tundra and moor, based on data reviewed by Wiernga
(1993).
The time series of the surface wind speed simulated by WRF as well as the
corresponding observations at three studied sites are presented in Figure 5. In general, the
WRF simulations with all ABL schemes captured the temporal pattern of surface wind
  16
observations reasonably well. However, the magnitude of modelled wind speed at the
individual stations was estimated with less success in the observation period. One reason of
the model errors can be related to significant synoptic scale variability in the Arctic, which
caused fast changes in the near-surface variables with the typical period of 1 to 3 days
(Mokhov et al., 2007; Nawri and Steward, 2009). Significant changes in the pattern of
atmospheric circulation and related surface wind variability were also observed at
Petuniabukta during the 12-days summer experiment. We found that observed and modelled
data agree well when the circulation pattern belongs to cyclonic type driven by deep lowpressure
system with the prevailing northerly winds over Svalbard archipelago. Such
situations occurred in the period of 10–11 July, 14–15 July and 20 July 2013. During these
events, however, the overestimation of wind speed of about 3–4 m s-1
was observed at Hørbye
foreland and Terrace stations. This can be explained by the limits of the individual BL
schemes in the WRF model during specific synoptic conditions (Vihma et al., 2014). In these
days, strong northern and northwestern advections in connection with deep cyclones centred
east of the Svalbard archipelago were dominating at the 850-hPa geopotential height maps.
Example of strong overestimation of the modelled wind speed is shown for 0300 UTC on 15
July 2013 (Figure 6). The predominant northern advection was clearly reproduced by the
WRF model, while the wind speed overestimation was found using all ABL schemes behind
the steeper leeward slopes in the northern part of Petuniabukta. The areas of higher WRF
wind speed extended from the bottom of leeward slopes to the crest, e.g. behind the
Cheopsfjellet, Chephrenfjelletas, and partly also Pyramiden and Fortet peaks (Figure1(c)).
However, the WRF-simulated wind fields during this event differed marginally for the three
used BL parameterization schemes. The reasons for such a spatial bias in the WRF
simulations can be associated with 1) the relatively high geostrophic wind speed in
connection with large-scale forcing over the Svalbard archipelago, and 2) the dynamical
  17
effects of the flow over and around complex terrain, which are still not explicitly
parameterised. This result is comparable with the study of Nawri et al. (2012) where the WRF
simulations was used in the evaluation of surface wind characteristics over Iceland. They
demonstrated both significant effects of geostrophic winds on modelled surface winds and
large differences in the wind field structure on the windward and leeward slopes in
comparison to the WRF estimates.
Obviously our simulation could not explain all aspects of the large variability in the
wind field pattern caused by the complex topography of the Svalbard archipelago. As found
by Kilpeläinen et al. (2011) and Esau and Repina (2012), there are specific driving
mechanisms on the Svalbard archipelago, which can revert the background atmospheric wind
resulting in differences in the vertical structure of the wind profiles, e.g. opposite direction of
geostrophic and near-surface winds. The heterogeneity of modelled wind characteristics and a
significant underestimation of wind speed were found at the Terrace station on 12 July 2013
and during 19 July 2013 at all stations. In both cases the underestimation was likely caused by
synoptic weather conditions related to a high-pressure system during which the WRF model
failed in capturing local circulation systems as well as the surface wind characteristics
recorded in Petuniabukta. The right panels in Figure 6 present the modelled wind patterns for
the selected BL parameterization schemes on 19 July 2013 at 1200 UTC. This period was
characterized by anticyclonic circulation with the prevailing eastern and southeastern
advection, clear sky or partly cloudy conditions affecting the temperature lapse rate and
convective processes in the ABL (Láska et al., 2013). We found that observed wind speed at
Terrace was about 4 m s-1
higher than modelled values, which could indicate either the
adiabatic subsidence with relating mechanical drainage in the larger part of Petuniabukta or
land-sea breeze driving mechanisms also documented on Svalbard archipelago (e.g. Sandvik
and Furevick 2002; Kilpeläinen et al., 2011; Láska et al. 2012). The authors report that land-
  18
sea breeze circulation systems can be driven by horizontal temperature differences between
the open water of the fjord and the glaciers. Moreover, Esau and Repina (2012) suggested that
temperature gradient may induce the background winds depending on the fjord topography
and depth. In case of Petuniabukta and Terrace station, this likely occurs due to combination
of land-sea breeze mechanism and recirculation of the warmer marine air on the short distance
(less than 5 km) between the fjord water and surrounding valleys. Comparison of the
modelled wind fields among all BL schemes shows that WRF simulations could not have
captured such local processes. Therefore, wind estimates can be significantly underestimated
in the Petuniabukta (see Figure 6) or similar narrow fjords on the Svalbard archipelago. Some
source of inaccuracies in the WRF simulations can be connected with the model inputs based
on the ERA-Interim reanalysis data which are derived from the limited number of
observations in the Arctic. In a similar way, Nawri et al. (2012) found in the interior of
Iceland that the negative wind speed bias in the WRF model was related to the weak
geostrophic winds. To sum up, these results illustrate that in the absence of significant largescale
forcing, e.g. in the high-pressure system overlying the domain boundaries, the WRF
model can be missing an internal forcing over the land area or due to misrepresentation of the
local topographic features.
3.4. Statistical evaluation of surface wind simulations
Table 2 and Figure 7 provide summary statistics of average errors and correlation coefficients
between observed and modelled surface wind for different BL parameterization schemes
during the 12-days measurements. In general, the WRF estimates showed a reasonable
potential for simulation of the surface wind speed at all stations. It can be documented by
correlation coefficients (r) that varied between 0.56 and 0.67 at the significant level p < 0.01.
The highest correlation was found at the Hørbye foreland station (r = 0.67) but it had also the
  19
largest mean bias of wind speed (0.5 m s-1
) and root-mean-square error (RMSE = 2.4 m s-1
).
The mean wind speed bias was very small at the Terrace station (-0.1 m s-1
) and Mumien Peak
(-0.3 m s-1
), while the RMSE was quite large (Figure 7(c)), ranging from 2.3 to 2.5 m s-1
for
Mumien Peak and Terrace, respectively.
The partial results of the wind speed simulation differed only marginally for the three
used BL parameterization schemes (Figure 7). The comparisons indicate that all schemes tend
to overestimate or underestimate the wind speed with the mean difference less than 0.1 m s-1
.
The best results given by the smallest bias and the lowest RMSE were found for the QNSE
scheme across all stations. The YSU and MYJ produced moderately stronger wind with a
larger RMSE at the Hørbye foreland station in comparison to QNSE parameterization. It is
also confirmed by the relative frequency distribution of differences between observed and
modelled wind speed (Figure 8) and specifics in appearance of the particular histograms. It is
obvious that the largest differences and overestimation of modelled wind speed appeared at
the Hørbye foreland and Terrace stations. It is documented by the lowest kurtosis and left
skewness occurring at all schemes, with the second highest frequency at the interval of -1.5 m
s-1
(Figure 8(a), 8(b)). It could indicate the model faults induced by a local topography,
channelling or convergence of airflow descending from the higher part of the valleys
Hørbyedalen and Ragnardalen to Petuniabukta (see Figure 1(c)). Recent studies of
Kilpeläinen et al. (2011), Esau and Repina (2012) and Roberts et al. (2015) addressing the
numerical simulations of local wind structure reported similar processes and mechanisms in
the other fjords of the Svalbard archipelago, e.g. Isfjorden and Kongsfjorden. By contrast, the
smallest differences between observed and modelled wind speed and were found at the
Mumien Peak (Figure 8(c)). The high kurtosis and nearly symmetric distribution of the wind
speed residuals at this station occurred for the MYJ and QNSE scheme, while the YSU
scheme showed a slight left skewness. Therefore, it can be concluded that the best simulation
  20
given by combining the high correlation, low bias and low RMSE were obtained for the
Mumien Peak and then for Terrace station. The large overestimation of the surface wind
speed at the Hørbye foreland station resulted from topographic features around the station.
The morphology of narrow valley Hørbyedalen was not well represented in the DEM and
influenced the differences between actual and modelled terrain elevation. Hence, it would
affect the sensitivity and differences in the simulated surface level winds among all the BL
schemes. It is necessary to point out that problem with wind speed simulation concerning a
large bias and weak correlation in the Arctic and Antarctic was also identified by Kilpeläinen
et al. (2012), Tatsula and Vihma (2011) or Roberts et al. (2015).
The ability of three boundary-layer schemes to simulate the wind direction was
evaluated on the basis of the DACD parameter (Figure 7(d)) and by comparison with the
observed wind distribution (Figure 9). The mean DACD values (see Eq. 4) showed
considerable differences between all stations. The highest DACD values for all schemes
(above 35%) were found at the Hørbye foreland and Mumien Peak stations. On the other
hand, low DACD values (18–21%) and week dependence of wind direction on WRF spatial
resolution for all schemes were documented for Terrace station. In general, it seems that the
DACD parameter was less sensitive to the selection of BL schemes in comparison to RMSE
and bias values. The DACD differences among the BL scheme ranged between 2.5% and
4.9% and were probably caused by a misrepresentation of the local topography and elevation
(Figure 1(c) and Table 1). This resulted in unrealisticly strong northerly and northwestrely
winds estimated at the Hørbye foreland and Terrace stations (Figure 9). On the other hand,
northeasterly and southerly winds at Terrace station as well as northeasterly and southeasterly
winds in case of the Mumien Peak and the Hørbye forland stations respectively were not
captured by the WRF simulations. In these circumstances the physics options given by the BL
schemes had a minor impact on the wind direction distribution (Figure 9). Nevertheless, the
  21
highest score given by the mean values of DACD was found at the Mumien Peak for the YSU
and scheme. Interestingly, at Terrace station the YSU and MYJ schemes had the largest errors
in the DACD values due to the overestimation of northwesterly and northerly winds by more
than 10% (Figure 9). Our results agreed that the WFR model capability to reproduce local
winds over complex topography is still limited, even using 1-km spatial resolution. It
confirms the findings from earlier studies in the Arctic fjords by Claremar et al. (2012) and
Mayer et al. (2012) who suggested the model setup for resolving ABL phenomena. Moreover,
it confirmed difficulties in capturing small-scale turbulence in the strongly stratified Arctic
atmosphere even with a much higher resolution used in the WRF model (Esau and Repina,
2012).
Further statistical analysis was made to the investigate influence of the surrounding
topography on the spatial variability of modelled wind speed and differences occurring in the
eight main directions. The error calculations based on correlation coefficients, bias and
RMSE, and the DACD parameter for the three observation sites and individual boundarylayer
schemes are shown in Table 3. In order to evaluate the wind characteristics, only the
situations with a representative number of WRF estimates at each direction, defined by
occurrence higher than 50 cases for Terrace and the Hørbye foreland stations and 25 cases for
the Mumien Peak, were chosen. We found considerable but very inconsistent differences
among the particular wind directions confirming the statistically significant results with p
<0.01. Table 3 shows that the best results (high correlation, low bias and RMSE) between
modelled and observed wind speed were found for the air flow from the northern,
exceptionally also from the southern sector. The wind speed bias <0.5 m s-1
was documented
for northerly wind at the Mumien Peak and northeasterly wind at Terrace and the Mumien
Peak stations, while the bias for southerly, southeasterly and southwesterly winds at these
stations often exceeded 1.0 m s-1
. Moreover, a large overestimation was found at Hørbye
  22
foreland station, with the bias between 0.8 and 1.7 m s-1
during northwesterly and northerly
flows. On the other hand, the mean correlation coefficient for the air flow from the northern
sector varied from 0.43 to 0.69 among the all stations.
In general, the simulations had weak sensitivity to the BL parameterization schemes
used in running the WRF model for wind assessment. The differences in the wind speed bias
among all schemes were typically between 0.1 and 0.9 m s-1
, depending on the prevailing
wind direction and topography surrounding each station. Generally, the QNSE scheme
performed best for all stations independent of the wind direction (mean bias <0.1 m s-1
, mean
RMSE = 2.1 m s-1
), while the YSU and MYJ schemes showed in the same directions slightly
higher bias (0.2 m s-1
) and RMSE (2.3 and 2.5 m s-1
). On the other hand, the YSU has
generated the largest bias and RMSE during advection from the northern sector, especially at
the Hørbye foreland stations (bias = 1.7 m s-1
, RMSE = 3.2 m s-1
). The comparison between
the observed and modelled wind distribution showed that the DACD parameters varied more
significantly among the main wind directions and studied sites than the wind speed estimates
(Table 3).
The sensitivity of the DACD parameter to the BL schemes was generally low and
showed a strong dependence on the frequency of individual wind directions and local
topography. The wind direction distribution was quite well reproduced at the Mumien Peak
and the Hørbye foreland stations during northerly winds, with DACD values between 61%
(QNSE) and 82% (YSU), respectively. The considerable lower DACD values were occurred
during southerly or southeasterly winds at the all stations. Relatively high topographic
complexity around the low-lying stations, which is precondition of the increased wind
turbulence or occurrence of local circulation systems (Pielke, 2002), affected the WRF
performance and higher differences in the DACD parameters among the chosen ABL
schemes. As mentioned above, the Terrace and Hørbye foreland stations are situated in the
  23
narrow part of the fjord, enhancing higher frequency of the winds parallel with the
longitudinal axis of Petuniabukta and two other valleys Hørbyedalen and Ragnardalen. As a
result, DACD values during westerly winds at the Hørbye foreland station varied significantly
from 10% (MYJ) to 26% (YSU). Furthermore, the large discrepancy between modelled and
observed wind distribution was found during northeasterly wind at Terrace station, where the
DACD did not exceed 4% among the BL schemes. This is also evident from the modelled
wind roses wherein northeasterly wind was not correctly reproduced and instead the northerly
and northwesterly winds were enlarged (Figure 9). As it has been described earlier, the high
occurrence of the northeasterly wind was also reported by Láska et al. (2012) analysing the
surface wind characteristics in the period July 2009 to June 2010. In this study, we assumed
that reason for such large differences during the northeasterly wind can be attributed to a
misrepresentation of some terrain features in the DEM, or more likely can be related to the
local circulation systems between the fjord water and Ragnardalen valley. As found by
Kilpeläinen et al. (2011), Esau and Repina (2012) and Vihma et al. (2014), the land-sea
breeze circulations in addition to the katabatic and channelled wind mechanisms may occur in
the Svalbard fjords with a typical vertical structure and recirculation of the warmer marine air
in the valley. Unfortunately, there are no relevant wind profile data and turbulence
measurements in Petuniabukta to verify both the WRF capability and the causes of the
increased frequency of the northeasterly wind at Terrace station.
4. Conclusions
The wind observations and numerical model simulation were performed at the coastal ice-free
zone of Petuniabukta in central Spitsbergen during a 12-day summer experiment. The
evaluation was based on the near-surface wind measurements taken at the three
meteorological stations placed in different geographical location and altitude. In the WRF
  24
numerical simulation we have focused on the three most widely used parameterization
scheme of the BL in the Arctic region and an evaluation of the wind characteristics from the
1-km horizontal resolution runs of the inner domain. The main conclusions drawn from this
study are the following:
‐ The surface wind characteristics were fairly well reproduced by the WRF simulation,
and they were comparable with similar studies performed in the Arctic. Physical
options given by three BL schemes had a rather small effect on the wind speed
estimates. The mean correlation coefficients between the surface wind speed
simulations and the observations varied from 0.56 to 0.67. The best results across all
stations were found for the QNSE parameterization scheme (bias <0.1 m s-1
, RMSE =
2.3 m s-1
), while the YSU and MYJ schemes showed a higher RMSE of 2.4 and 2.5 m
s-1
, respectively.
‐ The wind speed simulations were sensitive to geographical position and elevation of
the stations. Therefore, terrain morphology and its representation in the WRF model
affected a final accuracy of the wind speed estimates. In this regard, the largest bias
(0.6 m s-1
) and RMSE (2.5 m s-1
) were found at the Hørbye foreland station in which
the fjord morphology was not well represented at 1-km spatial resolution. On the other
hand, the wind speed bias was very small at the Terrace station (-0.1 m s-1
) and the
Mumien Peak (-0.3 m s-1
), although the RMSE were large, ranging from 2.3 to 2.5 m
s-1
.
‐ The predominant northerly wind direction was clearly reproduced by the model
although there was large overestimation of wind speed documented by the mean bias
between 0.8 and 1.7 m s-1
depending on the BL parameterization schemes. The model
errors in wind direction were significantly higher than discrepancies in the wind
speed. We found that all BL schemes had overestimated the frequency of occurrence
  25
of northern or northwestern components with high wind speed especially at the lowlying
stations Terrace and the Hørbye foreland stations. This predominant direction
can be associated with the channelling of drainage flows occurring in the fjord bottom
and captured also by the WRF simulation. However, the wind estimates were turned
by 15–25° clockwise at Hørbye foreland and 25–30° counterclockwise at Terrace
stations. The large discrepancy between modelled and observed wind distribution was
found during northeasterly wind at Terrace station, which is probably related with the
local circulation systems and an occurrence of breeze circulation between the fjord
water and surrounding valleys.
‐ The accuracy of the wind speed estimates was also significantly dependent on the
synoptic weather conditions. When the circulation pattern was related to a deep low
pressure system and prevailing northerly or northwesterly flows, then observed and
modelled data agree quite well. The significant large-scale forcing during these events
was observed over the Svalbard archipelago in connection with the 850-hPa
geostrophic wind exceeding 10 m s-1
. On the other hand, surface wind speed and wind
direction representation in the model often failed during less distinctive circulation
pattern, e.g. high-pressure ridge or baric col associated with easterly or southeasterly
flow over the Svalbard archipelago. During the anticyclonic circulations, the weak
geostrophic winds (<5 m s-1
), clear or partly cloudy conditions accompanied by the
convective boundary-layer processes and local wind driving mechanisms caused the
significant underestimation of the modelled wind speed against the observations.
Although the comparisons showed in general a satisfactory agreement between the modelled
and observed wind characteristics, further research should be carried out. Most attention
should be paid to boundary layer parameterization in relation to weak situations of large-scale
  26
forcing, the dynamical effects of the flow around and over complex topography or the effect
of local thermal circulation systems on the wind structure in the Arctic fjords and surrounding
valleys. Another possibility to improve the capability of the WRF model relates to availability
of appropriate data sets and therefore the needs for increasing number of observations at the
near-surface and the lower part of troposphere by using a dense network of automatic weather
stations, tower measurements or utilization of meteorological balloons and autonomous
unmanned aerial systems.
Acknowledgments
The authors would like to thank the members of Czech summer expedition 2013 to
Petuniabukta for their field assistance and companionship. This research was supported by the
project of Masaryk University MUNI/A/1370/2014 “Global environmental changes in time
and space“ and the project CzechPolar LM2010009 funded by the Ministry of Education,
Youth and Sports of the Czech Republic.
References
Aas KS, Berntsen TK, Boike J, Etzelmüller B, Kristjánsson JE, Maturilli M, Schuler TV,
Stordal F, Westermann S. 2015. A Comparison between Simulated and Observed
Surface Energy Balance at the Svalbard Archipelago. J. App. Meteorol. Climatol. 54:
1102–1119.
Chen F, Dudhia J. 2001. Coupling an Advanced Land Surface-Hydrology Model with the Pee
State-NCAR MM5 Modeling System. Part I: Model Implementation and Sensitivity.
Mon. Weather Rev. 129: 569–585.
  27
Claremar B, Obleitner F, Reijmer C, Pohjola V, Waxegård A, Karner F, Rutgersson A. 2012.
Applying a Mesoscale Atmospheric Model to Svalbard Glaciers. Advances in
Meteorology, Article ID 321649: 1–22.
Collins M, Knutti R, Arblaster J.M, Dufresne JL, Fichefet T, Friedlingstein P, Gao X,
Gutowski WJ, Johns T, Krinner G. 2013. Long-term climate change: projections,
commitments and irreversibility, in: Stocker TF, Qin D, Plattner GK, Tignor M, Allen
SK, Boschung J, Nauels A, Xia Y, Bex V, Midgley PM (Eds.), Climate Chang., 2013.
The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge, UK: Cambridge
University Press: 1054–1057.
Czech Geological Survey. 2009. James Ross Island - northern part. Topographic map 1 : 25
000. First edition. Praha, Czech Geological Survey. ISBN 978-80-7075-734-5.
Dee DP, Uppala SM, Simmons AJ, Berrisford P, Poli P, Kobayashi S, Andrae U, Balmaseda
MA, Balsamo G, Bauer P, Bechtold P, Beljaars ACM, van de Berg L, Bidlot J,
Bormann N, Delsol C, Dragani R, Fuentes M, Geer AJ, Haimberger L, Healy SB,
Hersbach H, Hólm EV, Isaksen L, Kållberg P, Köhler M, Matricardi M, McNally AP,
Monge-Sanz BM, Morcrette JJ, Park BK, Peubey C, de Rosnay P, Tavolato C, Thépaut
JN, Vitart F. 2011. The ERA-Interim reanalysis: configuration and performance of the
data assimilation system. Q. J. R. Meteorol. Soc 137: 553–597.
Esau I, Repina I. 2012. Wind climate in Kongsfjorden, Svalbard, and attribution of leading
wind driving mechanisms through turbulence−resolving simulations. Advances in
Meteorology, Article ID 568454: 1–16.
Evans DJA, Strzelecki M, Milledge DG, Orton C. 2012. Hørbyebreen polythermal glacial
landsystem, Svalbard. J. Maps 8: 146–156.
  28
Ewertowski MW, Tomczyk AM, 2015. Quantification of the ice-cored moraines' short-term
dynamics in the high-Arctic glaciers Ebbabreen and Ragnarbreen, Petuniabukta,
Svalbard. Geomorphology 234: 211–227.
Hanáček M, Nývlt D, Flašar J, Stacke V, Lehejček J, Tóthová G, Břežný M, Procházková B,
Uxa T, Křenovská I. 2013. New methods to reconstruct clast transport history in
different glacial sedimentary environments: Case study for Old Red sandstone clasts
from polythermal Hørbyebreen and Bertilbreen valley glaciers, Central Svalbard. Czech
Polar Reports 3: 107–129.
Hines KM, Bromwich DH. 2008. Development and testing of Polar Weather Research and
Forecasting (WRF) model. Part I: Greenland ice sheet meteorology. Mon. Weather Rev.
136: 1971–1989.
Holland MM, Bitz CM. 2003. Polar amplification of climate change in coupled models. Clim.
Dyn. 21: 221–232.
Jammalamadaka S, Sengupta A. 2001. Topics in circular statistics. World Scientific Publ. Co.
Ins. 328 pp.
Janjic ZI. 1996. The surface layer in the NCEP Eta model. Eleventh Conference on
Numerical Weather Prediction, Norfolf, VA, 19-23 August, Amer. Meteor. Soc, Boston,
MA: 354–355.
Janjic ZI. 2002. Nonsingular implementation of the Mellor-Yamada level 2.5 scheme in the
NCEP Meso model. NECEP Office Note, 437. 61 pp.
Kain JS. 2004. The Kain–Fritsch Convective Parameterization: An Update. J. Appl. Meteor.
43: 170–181.
Kilpeläinen T, Wihma T, Ólafsson H. 2011. Modelling of spatial variability and topographic
effects over Arctic fjords in Svalbard. Tellus 63A: 223–237.
  29
Kilpeläinen T, Wihma T, Manninen M, Sjölblom A, Jakobson E, Palo T, Maturilli M. 2012.
Modelling the vertical structure of the atmospheric boundary layer over Arctic fjords in
Svalbard. Q. J. Roy. Meteorol. Soc. 138: 1867–1883.
Láska K, Chládová Z, Ambrožová K, Husák J. 2013. Cloudiness and weather variation in
central Svalbard in July 2013 as related to atmospheric circulation. Czech Polar Reports
3: 184–195.
Láska K, Witoszová D, Prošek P. 2012. Weather patterns of the coastal zone of Petuniabukta
(Central Spitsbergen) in the period 2008–2010. Polish Polar Research 33: 297–318.
Livik G. 2011. An observational and numerical study of local winds in Kongsfjorden,
Spitsbergen. Master Thesis of University of Bergen. 110 pp.
Mäkiranta E, Vihma T, Sjöblom A,Tastula, EM. 2011. Observations and Modelling of the
Atmospheric Boundary Layer Over Sea-Ice in a Svalbard Fjord. Boundary Layer
Meteorol. 140: 105–123.
Mayer S, Jonassen MO, Sandvik A, Reuder J. 2012. Profiling the Arctic Stable Boundary
Layer in Advent Valley, Svalbard: Measurements and Simulations. Boundary Layer
Meteorol. 143: 507–526.
Michalakes J, Dudhia J, Gill D, Henderson T, Klemp J, Skamarock W, Wang W. 2004. The
Weather Research and Forecast Model: Software Architecture and Performance. 13 pp.
Mlawer EJ, Taubman SJ, Brown PD, Iacono MJ, Clough SA. 1997. Radiative transfer for
inhomogeneous atmospheres: RRTM, a validated correlated-k model for the longwave.
J. Geophys. Res. 102: 16663–16682.
Mokhov II, Akperov MG, Lagun VE, Lutsenko EI. 2007. Intense Arctic mesocyclones.
Izvestiya, Atmospheric and Oceanic Physics 43: 259–265.
Nawri N, Björnsson H, Jónasson K, Petersen GN. 2012. Evaluation of WRF mesoscale model
simulations of surface wind over Iceland. Report Vedurstofa Íslands 2012-010. 44 pp.
  30
Nawri N, Steward RE. 2009. Short-term temporal variability of atmospheric surface pressure
and wind speed in the Canadian Arctic. Theor. Appl. Climatol. 98: 151–170.
Nordli Ø, Przybylak R, Ogilvie AEJ, Isaksen K. 2014. Long-term temperature trends and
variability on Spitsbergen: the extended Svalbard Airport temperature series, 1898–
2012. Polar Research 33: 21349.
Pielke RA. 2002. Mesoscale Meteorological Modeling, International Geophysics Series.
Academic Press, Amsterdam. 676 pp.
Prach K, Klimešov, J, Košnar J, Redchenko O, Hais M. 2012. Variability of contemporary
vegetation around Petuniabukta, central Svalbard. Polish Polar Research 33: 383–394.
Przybylak R. 2003. The Climate of the Arctic. Atmospheric and Oceanographic Sciences
Library 26, Dordrecht, Netherlands: Springer. 272 pp.
Przybylak R. 2007. Recent air-temperature changes in the Arctic. Annals of Glaciology 46:
316–324.
Rachlewicz G. 2010. Paraglacial modifications of glacial sediments over millennial to decadal
time-scales in the high Arctic (Billefjorden, central Spitsbergen, Svalbard). Quaest.
Geogr. 29: 59–67.
Reuder J, Brisset P, Jonassen M, Müller M, Mayer S. 2009. The Small Unmanned
Meteorological Observer SUMO: A new tool for atmospheric boundary layer research.
Meteorol. Z. 18: 141–147.
Roberts TJ, Dütsch M, Hole LR, Voss PB. 2015. Controlled meteorological (CMET) balloon
profiling of the Arctic atmospheric boundary layer around Spitsbergen compared to a
mesoscale model. Atmos. Chem. Phys. Discussions 15: 27539–27573.
Sandvick AD, Furevick BR. 2002. Case study of coastal jet at Spitsbergen – comparison of
SAR- and estimated wind. Mon. Weather Rev. 130: 1040–1051.
  31
Santos-Alamillos FJ, Pozo-Vázquez D, Ruiz-Arias JA, Lara-Fanego V, Tovar-Pascador J.
2013. Analysis of WRF model wind estimate sensitivity to parameterization choice and
terrain representation in Andalusia (Southern Spain). J. Appl. Meteorol. Climatol. 52:
1592–1609.
Serreze MC, Barry, RG. 2009. The Arctic climate system. Cambridge, UK: Cambridge
University Press. 404 pp.
Serreze MC, Barry RG. 2011. Processes and impacts of Arctic amplification: A research
synthesis. Global and Planetary Change 77: 85–96.
Skamarock WC, Klemp JB, Gill DO, Barker DM, Wang W, Powers JG. 2008. A Description
of the Advanced Research WRF Version 3. Mesoscale and Microscale Meteorology
Division, National Centre for Atmospheric Research. 125 pp.
Skeie P, Grønås S. 2000. Strongly stratified easterly flows across Spitsbergen. Tellus 52A:
473–486.
Sukoriansky S, Galperin B, Perov V. 2005. Application of a new spectral theory of stably
stratified turbulence to the atmospheric boundary layer over sea ice. Boundary Layer
Meteorol. 117: 231–257.
Tastula EM, Vihma T. 2011. WRF model experiments on the Antarctic Atmosphere in
Winter. Mon. Weather Rev. 139: 1279–1291.
Tastula EM, Vihma T, Andreas EL. 2012. Evaluation of polar WRF from modelling the
atmospheric boundary layer over Antarctic sea ice in autumn and winter. Mon. Weather
Rev. 140: 3919–3935.
Thompson G, Field PR, Rasmussen RM, Hall WD. 2008. Explicit forecasts of winter
precipitation usin an improved bulk microphysics scheme. Part II: Implementation of a
new snow parametrization. Mon. Weather Rev. 136: 5095–5115.
  32
Vihma T, Pirazzini R, Fer I, Renfrew IA, Sedlar J, Tjernstrom M, Lupkes C, Nygard T, Notz
D, Weiss J, Marsan D, Cheng B, Birnbaum G, Gerland S, Chechin D, Gascard JC. 2014.
Advances in understanding and parameterization of small-scale physical processes in
the marine Arctic climate system: a review. Atmos. Chem. Phys. 14: 9403–9450.
Wiernga J. 1993. Representative roughness parameters for homogeneous terrain. Boundary
Layer Meteorol. 63: 323–363.
  33
Table 1. The geographical characteristics of the sites around the meteorological stations and
the elevation of the WRF model grid box closest to the station position.
Characteristics / Site Terrace Hørbye foreland Mumien Peak
Type of surface
marine terrace, tundra
vegetation
ice-cored moraine,
debris material
rock outcrops, regolith
material
Latitude (° N) 78.71060 78.74809 78.70021
Longitude (° E) 16.46059 16.40084 16.31638
Elevation (m a.s.l.) 15 67 773
Elevation in WRF (m a.s.l.) 14 138 593
  34
Table 2. Comparison of the observed and modelled hourly wind speed and wind direction in
the WRF model given for different ABL schemes and the selected meteorological stations.
Statistical characteristics for wind speed evaluation are represented by the correlation
coefficient (r), bias (WRF – observations), root-mean-square error (RMSE), while the wind
directions by a directional accuracy of circular distance (DACD). All correlation coefficient
are significant at the 0.01 level. Abbreviations YSU, MYJ and QNSE refer to the boundary
layer schemes used in the WRF model. The number of data pairs used for evaluation is given
in the column NUM. The validation period is from 9 July 0000 UTC to 22 July 2013
1800 UTC.
Site Terrace Hørbye foreland Mumien Peak
ABL
scheme YSU MYJ QNSE NUM YSU MYJ QNSE NUM YSU MYJ QNSE NUM
r 0.63 0.64 0.62 336 0.66 0.67 0.67 336 0.56 0.57 0.58 237
bias
(m s-1
) -0.13 -0.14 -0.17 336 0.53 0.57 0.34 336 -0.29 -0.32 -0.38 237
RMSE
(m s-1
) 2.71 2.43 2.40 336 2.54 2.41 2.13 336 2.36 2.21 2.22 237
DACD
(%) 19.01 20.83 18.45 336 38.39 37.80 39.00 336 40.08 38.40 35.02 237
  35
Table 3. Evaluation of the observed and modelled hourly wind speed for different ABL
schemes and meteorological stations categorized for the eight main wind directions
(WindDir). Statistical characteristics for wind evaluation are represented by the correlation
coefficient (r), bias (WRF – observations), root-mean-square error (RMSE) and directional
accuracy of circular distance (DACD). Abbreviations YSU, MYJ and QNSE refer to the
boundary layer schemes used in the WRF model. The number of data pairs used for
evaluation is given in the column NUM.
Terrace r bias (m s-1
) RMSE (m s-1
) DACD (%)
WindDir YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE
NUM
N 0.67 0.69 0.65 1.03 0.95 0.47 3.03 2.71 2.77 44.74 52.63 43.42 76
NE 0.50 0.50 0.51 0.28 0.06 0.09 2.99 2.71 2.68 2.00 2.00 4.00 100
E 0.77 0.71 0.71 -0.36 -0.31 -0.39 1.76 1.74 1.77 10.00 15.00 15.00 20
SE 0.06 -0.39 0.12 -0.59 -0.56 0.03 2.39 1.94 2.38 12.50 0.00 6.25 16
S 0.35 0.41 0.39 -1.35 -0.91 -0.76 2.27 1.89 1.82 16.90 14.09 11.27 71
SW 0.49 0.60 0.53 -1.42 -1.81 -1.27 2.24 2.27 2.04 10.00 13.33 0.00 30
W 0.32 0.15 0.41 -1.75 -0.91 -1.41 2.90 3.08 2.55 0.00 10.00 10.00 10
NW 0.73 0.86 0.81 1.86 1.40 1.02 3.04 2.21 2.10 69.23 76.92 92.31 13
Hørbye
foreland
r bias (m s-1
) RMSE (m s-1
) DACD (%)
WindDir YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE
NUM
N 0.55 0.62 0.53 1.68 1.33 0.82 3.21 2.91 2.84 81.69 80.28 78.87 71
NE 0.35 0.37 0.43 1.62 1.41 0.93 2.99 3.26 2.37 51.28 64.10 51.28 39
E 0.00 -0.02 -0.17 -0.25 -0.21 -0.15 1.53 1.08 1.03 10.00 20.00 10.00 10
SE 0.46 0.58 0.64 -1.03 -0.70 -0.62 1.85 1.45 1.32 26.00 30.00 30.00 50
S -0.22 -0.36 -0.12 -0.93 -0.83 -0.75 2.53 2.44 2.07 10.71 3.57 3.57 28
SW 0.58 0.41 0.84 0.12 0.03 0.18 1.12 1.33 0.80 0.00 0.00 14.29 7
W 0.32 0.37 0.31 0.14 0.34 0.21 2.20 2.01 1.97 25.49 9.80 23.53 51
NW 0.84 0.80 0.84 0.86 1.07 0.76 2.39 2.33 1.97 26.25 27.50 31.25 80
Mumien
Peak
r bias (m s-1
) RMSE (m s-1
) DACD (%)
WindDir YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE YSU MYJ QNSE
NUM
N 0.64 0.64 0.66 0.32 0.34 0.27 2.36 2.23 2.19 61.11 59.72 61.11 72
NE 0.44 0.43 0.48 -0.03 -0.12 -0.08 2.06 1.93 1.86 24.44 22.22 15.56 45
E 0.18 0.37 0.26 -0.28 -0.13 -0.33 2.23 1.90 2.02 16.67 11.11 5.56 18
SE 0.06 0.28 0.30 -1.93 -1.80 -1.92 2.89 2.64 2.71 14.82 7.41 3.70 27
S 0.83 0.85 0.89 -1.12 -1.11 -1.17 2.30 2.39 2.32 19.05 9.52 9.52 21
SW 0.63 0.74 0.67 -1.10 -1.37 -1.41 2.67 2.49 2.67 39.27 42.86 42.86 28
W 0.53 0.56 0.60 0.44 0.01 0.13 1.81 1.46 1.50 75.00 75.00 58.33 12
NW 0.74 0.72 0.77 1.11 1.22 1.09 2.19 2.05 1.93 64.29 78.57 64.29 14
  36
Figure 1. (a) Atlantic sector of the Arctic with the Svalbard archipelago where the wind
observations took place in July 2013. (b) Domains D1, D2 and D3 used in the WRF
simulation with 9, 3 and 1 km spatial resolutions. (c) The inset map of the northern part of
Billefjorden and Petuniabukta area with the location of automatic weather stations (blue
triangles): 1 – Terrace, 2 – foreland of Hørbyebreen glacier, 3 – top of Mumien Peak. The
modified map of Petuniabukta is based on the Svalbardkartet data, Norwegian Polar Institute.
  37
Figure 2. Synoptic situations and prevailing circulation patterns over the Svalbard
archipelago based on the 850-hPa geopotential height in meters (blue line) and winds in m s-1
(arrow) from the ERA-Interim reanalysis data over the observation period.
  38
Figure 3. Time series of 30-min wind speed (a) and wind direction (b) measured at the three
meteorological stations in Petuniabukta in central Spitsbergen during the 12-days experiment.
  39
Figure 4. Wind roses based on observed surface winds at the three meteorological stations in
Petuniabukta in central Spitsbergen during the 12-days experiment. The values in the circle
indicate the percentage of observation and colour scales show the wind speed categories (m s-
1
).
  40
Figure 5. Time series of hourly simulated and observed wind speeds at (a) Terrace, (b)
Hørbye foreland and (c) Mumien Peak stations during the 12-days experiment. Abbreviations
YSU, MYJ and QNSE refer to the boundary layer schemes used in the WRF model. Black
arrows indicate two events with a high overestimation or underestimation of the wind speed
estimates discussed in the text and Figure 6.
  41
Figure 6. Example of two different results of surface wind simulation in the area of
Billefjorden, central Svalbard: slightly overestimated surface wind speed on 15 July 2013 (left
column) and underestimated wind speed on 19 July 2013 (right column). Abbreviations YSU,
MYJ and QNSE refer to the boundary layer schemes used in the WRF model. The colors
show wind speed (m s-1
), while the vectors show wind direction. The black line represents the
surface elevation (100 m) based on the Svalbardkartet data, Norwegian Polar Institute. The
coastal line is shown by white line.
  42
Figure 7. Summary of wind speed and wind direction evaluation at the selected station in
Petuniabukta represented by the (a) correlation coefficient, (b) bias (WRF – observations), (c)
root-mean-square error (RMSE) and (d) directional accuracy of circular distance (DACD).
Abbreviations YSU, MYJ and QNSE refer to the boundary layer schemes used in the WRF
model.
  43
Figure 8. Relative frequency distributions of the differences between modelled and observed
wind speed at (a) Terrace, (b) Hørbye foreland and (c) Mumien Peak stations for the 12-days
experiment. Abbreviations YSU, MYJ and QNSE refer to the boundary layer schemes used in
the WRF model.
  44
Figure 9. Comparison of observed and modelled wind roses for the three boundary layer
schemes (YSU, MYJ and QNSE) used in the WRF model for the whole study period. The
values in the circle indicate the percentage of observation and colour scales show the wind
speed categories (m s-1
).
11
Paper 10
Impact of warming on Nostoc colonies (Cyanobacteria) in a wet
hummock meadow, Spitsbergen
Elster, J., Kvíderová, J., Hájek, T., Láska, K., Šimek, M., 2012,
Polish Polar Research, 33, pp. 395–420
Impact of warming on Nostoc colonies (Cyanobacteria)
in a wet hummock meadow, Spitsbergen
Josef ELSTER1,2*, Jana KVÍDEROVÁ1,2*, Tomáš HÁJEK1,2, Kamil LÁSKA1,3
and Miloslav ŠIMEK1,4
1 University of South Bohemia, Faculty of Science,
Branišovská 31, 370 05 České Budějovice, Czech Republic
2 Institue of Botany AS CR, Dukelská 135, 379 82 Třeboň, Czech Republic
3 Masaryk University, Faculty of Science, Kotlářská 2, 611 37 Brno, Czech Republic
4 Institute of Soil Biology, Biology Centre AS CR,
Na Sádkách 7, 370 05 České Budějovice, Czech Republic
* <jelster@butbn.cas.cz> <kviderova@butbn.cas.cz>
Abstract: In order to simulate the warming effects on Arctic wetlands, three passive
open−top chambers (OTCs) and three control cage−like structures (CCSs) equipped with
soil temperature and soil volumetric water content (VWC) probes for continuous micro−
climatic measurements were installed in a wet hummock meadow, Petuniabukta, Bille−
fjorden, central Spitsbergen, in 2009. The warming effects on primary productivity were
investigated during summer seasons 2009 and 2010 in cyanobacterial colonies of Nostoc
commune s.l., which plays an important role in the local carbon and nitrogen cycles. The
microclimatic data indicated that the effect of OTCs was dependent on microtopography.
During winter, two short−term snow−thaw episodes occurred, so that liquid water was
available for the Nostoc communities. Because of the warming, the OTC hummock bases
remained unfrozen three weeks longer in comparison to the CCSs and, in spring, the OTC
hummock tops and bases exceeded 0°C several days earlier than CCS ones. Mean summer
temperature differences were 1.6°C in OTC and CCS hummock tops, and 0.3°C in the
OTC and CCS hummock bases. The hummock tops were drier than their bases; however
the VWC difference between the OTCs and CCSs was small. Due to the only minor differ−
ences in the microclimate of OTC and CCS hummock bases, where the Nostoc colonies
were located, no differences in ecophysiological characteristics of Nostoc colonies ex−
pressed as photochemistry parameters and nitrogenase activities were detected after two
years exposition. Long−term monitoring of Nostoc ecophysiology in a manipulated envi−
ronment is necessary for understanding their development under climate warming.
Key words: Arctic, Svalbard, Cyanobacteria, nitrogenase activity, photosynthesis.
Introduction
In Arctic hydro−terrestrial environments, considerable, i.e. visible, amounts of
microalgal and cyanobacterial biomass can accumulate over a long time period
Pol. Polar Res. 33 (4): 395–420, 2012
vol. 33, no. 4, pp. 395–420, 2012 doi: 10.2478/v10183−012−0021−4
Unauthenticated
Download Date | 2/16/16 5:42 PM
(Vincent 2000; Elster 2002). Cyanobacteria play a dual role there. Besides being
considerable primary producers, they are able to fix atmospheric nitrogen, thus be−
ing an important source of nitrogen for other organisms. Nitrogen often limits pri−
mary production in the Arctic (Henry and Svoboda 1986; Davey and Rothery
1992; Liengen and Olsen 1997; Walker et al. 2008; etc.).
The climate change significantly influences current temperature and moisture
conditions in Arctic ecosystems. The predicted effects of a doubling of the atmo−
spheric CO2 will be dramatic; mean annual temperatures in the Arctic may be
3–5°C higher within 100 years (Maxwell 1992; ACIA 2004; Callaghan et al.
2004a; Callaghan et al. 2004b). Most models also predict that the overall annual
global precipitation will increase. Temperature and precipitation increases will
have a large impact on cyanobacteria and microalgae community structure, and
also on their in situ ecophysiological performance (photochemical processes and
nitrogen fixation; Elster et al. 2001; Callaghan et al. 2004a, b), as water and tem−
perature are the major environmental predictors of species survival in the Arctic.
Even slight increases in temperature and precipitation could lead to a development
of a deeper active layer, higher rates of chemical transformation and, ultimately,
greater nutrient availability (Shaver et al. 2000; Rolph 2003; Walker et al. 2008).
It remains unclear how climate warming will affect nitrogen fixation. Some
authors predict, that future Arctic environments will experience increased nitrogen
fixation due to heightened enzymatic activity and increased concentration of car−
bon dioxide, while others predict that nitrogen fixation will be inhibited by in−
creased available nitrogen due to increased mineralization acting as a negative
feedback upon this process (Paul and Clark 1996; Walker et al. 2008).
Many field studies have been conducted on the potential effects of global
warming on the growth and community structure of vegetation, soil bacterial and
invertebrate community performances in various polar/alpine ecosystems (Marion
et al. 1997; Hollister and Webber 2000; Walker et al. 2008; Rinnan et al. 2009b).
By contrast, equivalent experimental studies on cryptogams (mosses, lichens,
cyanobacteria and microalgae) in the Arctic are much sparser and comparative
studies across sites are missing. Since cryptogams are the main primary producers
in high Arctic wetlands, any change in their growth or community structure caused
by climate change will affect the local ecosystem.
Passive open−top chambers (OTC) are the most frequent manipulation tech−
nique in these in situ warming experiments (Chapin and Shaver 1985; Strathdee
and Bale 1993; Kennedy 1995; Marion et al. 1997; Day et al. 1999; Hollister and
Webber 2000; Convey et al. 2002). This technique is standardized in design and
forms part of a broader geographical network of sites (e.g. The International Tun−
dra Experiment, ITEX), which allows for a comparison of responses on a circum−
polar scale (Molau and Mølgaard 1996).
Our experimental locality in Petuniabukta, Billefjorden, Central Svalbard, the
Norwegian Arctic, is situated in a wet hummock meadow. The meadow is satu−
396 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM
rated with water at the beginning of the vegetation season due to extensive
snowfield melting, however it gradually dries during summer and early fall
(Kvíderová et al. 2011). The desiccation process can be interrupted by short rain
events that could supply the tundra with water again. Tundra vegetation can be
subjected to repeated cycles of drying and re−hydration. In this hydro−terrestrial
ecosystem, the N. commune colonies are the most significant nitrogen source
(Liengen and Olsen 1997; Vincent 2000; Kvíderová et al. 2011). These colonies
are very abundant and produce high biomass, which significantly contributes to
the local carbon and nitrogen cycles (Kvíderová et al. 2011).
In this paper, we report on the first two years (2009 and 2010) of in situ experi−
ments, using passive open−top chambers to simulate warming, and control treat−
ments in a wet hummock meadow, Petuniabukta, Billefjorden, Central Svalbard.
The tasks of this field experiments are:
• To describe the microclimatic environment (surface soil layer temperature and
volumetric water content in the vegetative seasons, non−vegetative seasons and
transition period) including water chemistry of the wet hummock meadow un−
der natural (control) and in situ warming simulation.
• To compare the microclimatic parameters (temperature and volumetric water
content in hummock tops and bases) in the OTCs and control treatments in or−
der to estimate the OTC efficiency of warming.
• To evaluate photochemical performance and nitrogenase activity (nitrogen fix−
ation) of N. commune colonies in the wet hummock tundra in manipulated and
non−manipulated environments.
• To relate the microclimatic data with physiological performance of N. commune
in the wet hummock tundra in manipulated and non−manipulated environments.
Methods
Site description and experimental design
The experimental site is located in a wet hummock meadow at Petuniabukta,
Billefjorden, Central Svalbard (N 78°43’49” E 16°26’41”, 15 m a.s.l.) where the
vegetation is dominated by mosses (see vegetation type unit “Vegetation domi−
nated by mosses” description in Prach et al., this issue). Hummocks are 15–20 cm
high, about 20–30 cm in diameter and there are about 2–3 hummocks per 1m2. In
summer 2009, three Open−top Chambers (OTC) and three Control Cage−like
Structures (CCSs) were installed at the experimental locality (Fig. 1A). The OTC
was designed as a hexagonal chamber with a bottom diameter of 140 cm (side
length of 70 cm), top diameter of 90 cm (side length of 45 cm), and height of 50
cm. The area of each OTC was 1.27 m2. The chambers had inwardly inclined sides
(60° with respect to horizontal), which improved transmittance of solar radiation
and helped to trap heat. The chambers were constructed of Perspex (Quinn XT FL
Impact of warming in a wet hummock meadow 397
Unauthenticated
Download Date | 2/16/16 5:42 PM
5 mm thick, Quinn Plastics, Ireland) designed specifically for solar applications.
This material had high solar transmittance in the visible wavelengths (90%) and
low transmittance in the infrared (heat) range (<5%). The tops of OTCs were cov−
ered by fencing over winter. The control treatments CCSs were designed as a cage
of equal length and width of 90 cm, and height of 40 cm. The area of each CCS was
0.81 m2. The cages were welded together with a 0.8 cm diameter iron pole stand
and covered by a wire screen. The entire experimental area is surrounded by a
wooden ring fence. The screen of the CCS and tops of the OTCs protected these
structures against damages e.g. from the Arctic fox, while the wooden ring fence
prevented a reindeer grazing. Construction of the OTC device is described in more
detail in the ITEX manual (Molau and Mølgaard 1996).
Water level and water physico−chemical parameters
In addition to the manipulation experiment, water level and water physico−
chemical parameters in the wet hummock meadow were monitored. A transect of
4 parallel rows of holes 30 cm deep were drilled into the permafrost across the exper−
imental plots. A polyethylene tube, 5cm diameter, with a densely perforated wall,
was installed inside each hole (Fig. 1B), in which the water level was measured by a
builder’s tape. Simultaneously, water from the tubes was sucked up by a syrette with
tubes, filtered through a pre−rinsed Whatman filter paper GF/C in the field and trans−
ferred to two acid−washed polythene bottles (100 ml). The first bottle was frozen for
transport to the Czech Republic, while the second was used in situ for measurements
of pH and conductivity (Kombibox WTE, Weilheim, CB 570).
Inorganic nitrogen and phosphorus concentrations were determined using a
Flow Injection Analyser (FIA, Tecator, Sweden; Růžička and Hansen 1981). Dis−
solved reactive phosphorus (DRP; PO4−P) was analysed by reaction with ammo−
nium molybdate and reduction by stannous chloride to phosphomolybdenum blue
(Proctor and Hood 1954, Application note AN60−83 Tecator). The detection limit
for DRP was 5 μg l−1. Nitrate− (NO3−N) and nitrite− (NO2−N) nitrogen were ana−
lysed by reaction with sulphonamide (Application note ASN 62−01−83) and am−
monium−nitrogen (NH4−N) by the gas diffusion method (Karlberg and Tweng−
strom 1983, Application note ASN 50−0187 Tecator). The detection limit for
NH4−N was 10 and for NO3−N was 3 μg l−1. For total nitrogen (TN) and total
phosphrus (TP) determination, the samples were treated by persulfate mineraliza−
tion at 151°C for 30 min. The concentration of Cl ions was determined by reaction
with mercury thiocyanate and ferric ions (Růžička et al. 1976, Application note
AN 63/83 Tecator, detection limit 5 μg l−1). The SO4−S content was analyzed by
FIA (Madsen and Murphy 1981, Application note ASTN42/86 Tecator, detection
limit 10 μg l−1). The dissolved substances (dry weight, DW) content was deter−
mined by evaporation and drying at 105°C (Howard 1933; Sokoloff 1933). Cation
contents (Na, K, Ca, Mg) were analyzed by a Varian Spectra AA 640 polarized
atomic absorption spectrophotometer (Techtron, Australia). A 300 mg sample was
398 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM
weighed into a thick−wall test tube and concentrated HNO3 of volume 0.75 ml and
concentrated HCl of volume 2.2 ml were added. The tube was closed with a loose
stopper. The sample was mixed and let to stand at ambient temperature for 16 hrs.
After that time, the sample was boiled at 120°C for 2 hrs. Then the sample was
cooled and filtered quantitatively into a 50ml volumetric flask. The 5% HNO3 was
added to get the final sample volume of 50 ml. For determination of cations con−
centration by atomic absorption spectrophotometry, the acetylele−air flame analy−
sis method was used for Na, K and Mg, and N2O−acetylene flame analysis of Ca.
Microclimatic parameters
In each treatment soil temperature and volumetric water content (VWC) were
permanently monitored approximately in the central part of each OTC and/or CCS
at the bases and tops of the hummocks. The soil capacitance method was used to
measure volumetric water content at a depth of ~2 cm below the ground. In total,
ten ECH2O EC−5 soil moisture probes (Decagon Device Inc., USA) and ten
Pt100/8 thermometers (EMS, Czech Republic) were inserted completely into the
substrate within the OTCs and CCSs. The ECH2O EC−5 probe dimensions were
5 × 2 × 0.1 cm which allowed for measurement of the apparent dielectric constant
of soil within a 2 cm zone on both flat sides of the probe. An EdgeBox V12 multi−
channel datalogger (EMS, Czech Republic) was used to excite and measure the
output from the ECH2O EC−5 soil moisture probes. The datalogger supplied 2.5V
to each ECH2O EC−5 probes and the outputs were converted by the calibration
equation for mineral soils (Parsons and Bandaranayake 2009). The output from all
probes was measured and stored at 60 min intervals.
Impact of warming in a wet hummock meadow 399
Fig. 1. The experimental site in the wet hummock meadow in Petuniabukta. A. Control cage−like
structure (left) and passive open−top chamber (right). B. Tube for water sampling. The distance be−
tween the holes is 5 cm.
Unauthenticated
Download Date | 2/16/16 5:42 PM
Ecophysiological parameters of Nostoc commune
In summer 2009 three plastic Petri dishes (9 cm diameter), each with one col−
ony of N. commune, were installed to each OTC and CCS. A dense net of holes was
prepared by boring across the bottom and lid of each Petri dish. Filter paper was in−
serted on to the bottom of each Petri dish (Fig. 2A). The closed Petri dishes were
kept in contact with the wet tundra surface by needles. In both experimental sea−
sons (2009 and 2010), Petri dishes with N. commune colonies were regularly col−
lected, put into a cool box and transported to the field laboratory for measurements
of photochemical processes and nitrogenase activity.
Special Exposition Chambers (EC) were prepared in summer 2010, because of
an expected greenhouse effect in the closed Petri dishes. These ECs were made from
perplex glass tubes with a diameter of 12.5 cm; the same material was used for OTC
construction. The top and bottom of each tube were covered by plastic nets (square
holes with length of 0.5 cm), which were fixed by narrow plastic rings (Fig. 2B). In
each experimental treatment (OTC and CCS) three ECs, with one N. commune col−
ony in each, were in free contact with the moss tundra surface. In 2010, the N. com−
mune colonies enclosed in the ECs were regularly measured like the Petri dishes. Af−
ter measurements, all Petri dishes and ECs were returned exactly to the same place.
Weight, photochemical processes and nitrogenase activity were evaluated in
each colony. The colony was weighed on a digital scale (Denver Instruments, Ger−
many). Photochemical processes were measured by a FluorCam 700MF fluores−
cence imaging camera (Photon Systems Instruments, Czech Republic) using the
quenching protocol. The Nostoc colony was dark−adapted for 4 hours before the
measurement. Red measurement pulses of less than 3 μmol m−2 s−1 and lasting
33.3 μs were applied to obtain minimum fluorescence in the dark (F0). A saturation
pulse of white light of 2100 μmol m−2 s−1 lasting 800 ms was used to determine
maximum fluorescence in the dark (FM). The maximum quantum yield (FV/FM)
was measured 10 s after the start of the measurement in the dark adaptation state
and was followed by 40 s of dark relaxation (refer to Kvíderová et al. 2011 for pro−
tocol details). The photochemical parameter, maximum quantum yield during dark
400 Josef Elster et al.
Fig. 2. Two types of exposition of Nostoc commune s.l. A. Petri dish. B. Exposition chamber.
Unauthenticated
Download Date | 2/16/16 5:42 PM
adaptation (FV/FM), was calculated using the FluorCam 7 software (Photon Sys−
tems Instruments, Czech Republic) as according to Roháček and Barták (1999)
and Maxwell and Johnson (2000). The equation used and detailed equation param−
eter descriptions are summarized in Kvíderová et al. (2011).
Nitrogenase activity (NA) was measured by acetylene−ethylene reduction as−
say (Stewart et al. 1967) in 100 mL glass flasks after 90 min incubation. Ethylene
in the samples was quantified with a gas chromatograph and nitrogenase activity
was expressed in nmol/h C2H4 per g fresh biomass. The details of the nitrogenase
activity evaluation are given in Kvíderová et al. (2011).
Statistical evaluation
All statistical evaluations were performed using Statistica 9.0 (StatSoft, USA);
differences were considered significant at P <0.05 (probability of Type I Error: null
hypothesis is rejected when it is true). All data were tested for normality before the
statistical evaluation. The non−parametric Kruskal−Wallis test was used for evalua−
tion of differences in physico−chemical parameters at the site below the terrace and
the experimental site (n = 10 in each sampling). The homologous groups determined
by the ANOVA/HSD test for unequal n at a significance level of 0.05 were used for
evaluating the differences in temperature and soil water content at the bases and tops
of each OTC and CCS for the whole experiment and vegetative season only (n = 3
for OTC and n = 2 for CCS in each data record; total of n = 13210 for all temperature
data from one probe and n = 10199 for all VWC data from one probe were used for
key data determination). Paired t−test was used to evaluate the inter−seasonal differ−
ence in number of “drought” days as an indicator of the difference in water availabil−
ity at the experimental locality (n = 10 for each year). The homologous groups deter−
mined by the ANOVA/Tukey HSD test at a significance level of 0.05 were used for
evaluating warming effects on Nostoc colonies (n = 9 in each treatment). Correlation
analysis was performed to evaluate the relationship between instantaneous microcli−
mate conditions (temperature, VWC) and ecophysiological parameters (weight,
FV/FM and nitrogenase activity) of the Nostoc colonies (Petri dishes: n = 36 for tem−
perature and n = 63 for VWC in OTCs, n = 24 for temperature and n = 42 for VWC in
CCSs; ECs: n = 27 in OTC and n = 18 in CCS for both variables), and the relation−
ship between water level depth and Nostoc colonies ecophysiology (Petri dishes: n =
63 in OTC, n = 42 in CCS; ECs: n = 27 in OTC, n = 18 in CCS).
Results
Water level and water physico−chemical parameters
Water levels in the wet hummock meadow greatly differed in the 2009 and 2010
summer seasons (Fig. 3). In 2009, the water level was well balanced through the
whole season, while in 2010 the water level decreased in the first half of the season
Impact of warming in a wet hummock meadow 401
Unauthenticated
Download Date | 2/16/16 5:42 PM
due to the much smaller snow field in the upper part of the wet meadow, but came
back during the second half of the season. However, the ranges of the standard devi−
ation bars indicated that water level in all 9 measuring points differed due to
microhabitat diversity. In particular, microhabitat water level can be several centi−
metres below the surface, while a shallow surface pool could occur just a few meters
away.
Water physico−chemical parameters are given in Table 1. Because of the drier
second season (see below), the TP and DW values were slightly higher in 2010
than in 2009. On the contrary, NH4−N was slightly higher in 2009. In 2009, the re−
duced form of ammonium−nitrogen in the wet environment could not be oxidised
into nitrate and nitrite forms. In 2009 at the upper part of the experimental site, be−
low the terrace from where water was flowing into the meadow, NO2−N, NO3−N,
TN, PO4−P and TP concentrations were higher than in the meadow, since most of
the mineral nutrients were in the inflowing water. The decrease of pH in the experi−
mental field could be caused by calcium, magnesium and sulphate precipitation. In
the drier 2010 season, the physical and chemical parameters of samples from be−
low the terrace and the experimental field were comparable with the exception of
lower NO3−N at the OTC area. Such a decrease was observed also in 2009. This
could indicate nitrate consumption by the meadow vegetation.
Microclimatic parameters
Temperature. — The microclimate monitoring was initiated at the hummock
wet meadow in July 2009 and continued through the winter season till September
2010. However, soil temperature data acquired from July to December 2009 were
not included in the analysis and subsequent evaluation due to bad quality and the
occurrence of missing values in the data set. The courses of soil temperature at the
402 Josef Elster et al.
Fig. 3. Water levels at the experimental locality (mean ± standard deviation, n = 10 in each sampling
date) in both the 2009 and 2010 growing seasons.
Unauthenticated
Download Date | 2/16/16 5:42 PM
tops and bases of hummocks in the OTCs and CCSs were therefore processed from
December 2009 till September 2010 (Fig. 4). Periods of temperature and volumet−
ric water content (introduced later) measurements were divided into; (i) vegetative
season 2009, (ii) non−vegetative season 2009−2010 including warming period in
spring, (iii) transition period 2010, and (iv) vegetation season 2010 (Figs 4, 5 and
Table 2). In both cases (OTC, CCS), temperatures fluctuated from 0°C down to −7
or −8°C from the start of measurements up to mid−June, 2010. There was a period
with temperatures around 0°C resulting in a partial snow thawing in the second
half of February 2010. During winter (non−vegetative season 2009/2010), the tem−
peratures in both OTCs and CCSs were very similar, however, Tmax was a little bit
lower in the OTCs (about 0.2°C; Table 2). The surface of the soil in all experimen−
tal treatments warmed up to temperatures close to 0°C from the beginning of May
(the non−vegetative season 2009/2010 – warming period in Table 2). The tempera−
ture reached −1°C in mid−May (transition period start date; Table 2) and, after this
period, the temperature remained stable around 0°C (transition period in Table 2).
The temperature of the OTC hummock tops reached positive values (vegetation
season 2010) for several days earlier in comparison to the CCSs; a similar pattern
was observed in the OTC and CCS hummock bases (Fig. 4). The duration of the
transition period was the shortest in the OTC hummock tops, and the CCS hum−
mock tops and bases, and was the longest in the OTC hummock bases (Table 2).
Impact of warming in a wet hummock meadow 403
Table 1
The physico−chemical parameters (mean ± standard deviation) of the site below the terrace and
the experimental field during the 2009 and 2010 seasons; n – number of samples, statistical sig−
nificance (non−parametric Kruskal−Wallis test) * – P <0.05, ** – P <0.01, *** – P <0.001.
2009 2010
all samples below terrace
experimental
field
all samples below terrace
experimental
field
n = 23 n = 19 n = 4 n = 24 n = 5 n = 21
pH 7.61 ± 0.45 8.30 ± 0.44 7.47 ± 0.36** 7.52 ± 0.23 7.77 ± 0.04 7.48 ± 0.25
Conductivity μS cm−1 689 ± 223 831 ± 656 666 ± 95 814 ± 217 626 ± 73 841 ± 218
NH4−N μg l−1
140 ± 77 249 ± 126 121 ± 56 55.0 ± 53.5* 67.4 ± 16.5 53.2 ± 56.9
NO2−N μg l−1 4.09 ± 7.92 15.1 ± 21.8 2.24 ± 0.47* 3.18 ± 1.41*** 3.49 ± 0.50 3.13 ± 1.50
NO3−N μg l−1 82.3 ± 77.8 249 ± 82 49.9 ± 13.2** 133 ± 270 150 ± 66 131 ± 288*
TN μg l−1 539 ± 175 768 ± 206 491 ± 138* 620 ± 322 687 ± 85 611 ± 343
PO4−P μg l−1 22.7 ± 16.0 52.4 ± 28.2 17.4 ± 5.0* 14.9 ± 13.9 18.4 ± 0.9 14.4 ± 14.9
TP μg l−1 42.5 ± 17.7 73.8 ± 37.8 37.5 ± 5.9* 58.2 ± 7.1*** 62.1 ± 8.7 57.7 ± 6.9
Cl mg l−1 19.8 ± 58.6 99.5 ± 163.7 7.25 ± 2.32 11.5 ± 6.9** 10.7 ± 2.1 11.7 ± 7.3
DW g l−1
0.036 ± 0.024 0.021 ± 0.010 0.038 ± 0.0260.065 ± 0.033*0.050 ± 0.004 0.067 ± 0.035
Na mg l−1 15.0 ± 37.8 66.5 ± 105.4 6.90 ± 1.23 6.88 ± 3.16 5.69 ± 0.95 7.05 ± 3.33
K mg l−1 1.42 ± 2.23 4.41 ± 6.17 0.91 ± 0.17 1.46 ± 1.41 1.32 ± 0.38 1.48 ± 1.51
Ca mg l−1 134 ± 26 91.2 ± 11.7 141 ± 21 171 ± 55 126 ± 11.2 178 ± 56
Mg mg l−1 18.0 ± 4.0 19.8 ± 10.9 17.5 ± 2.3 21.0 ± 7.0 14.6 ± 3.9 21.9 ± 6.9
SO4−S mg l−1 214 ± 107 205± 130 219 ± 109 246 ± 46 226 ± 27 249 ± 48
Unauthenticated
Download Date | 2/16/16 5:42 PM
404 Josef Elster et al.
Fig. 4. Soil temperature in OTC and CCS hummock tops and bases during the experiment. A. OTC
base. B. OTC top. C. CCS base. D. CCS top. Lines: black, mean (n = 3 for OTCs, n = 2 for CCSs);
light gray, mean, standard deviation; dark gray, mean + standard deviation.
Unauthenticated
Download Date | 2/16/16 5:42 PM
Above−zero temperatures were reached for the first time (0°C date; Table 2) in the
OTC and CCS hummock tops as early as in mid− or late May 2010 (OTC hummock
tops with the exception of OTC1 on 12.6.2010). Positive (above−zero) tempera−
tures were reached in the OTC and CCS hummock bases almost one month later, in
Impact of warming in a wet hummock meadow 405
Fig. 5. Volumetric water content (VWC) in OTC and CCS hummock tops and bases during the exper−
iment. A. OTC base. B. OTC top. C. CCS base. D. CCS top. Lines: black, mean (n = 3 for OTCs, n = 2
for CCSs), light gray, mean, standard deviation, dark gray, mean + standard deviation.
Unauthenticated
Download Date | 2/16/16 5:42 PM
406 Josef Elster et al.
Table 2
The seasonal characteristics (mean ± standard deviation; n = 13210 for all temperature data
from one probe, n = 10199 for all VWC data from one probe) of the OTCs and CCSs. The
letter indicates homologous groups as identified by an ANOVA/HSD test for unequal n at
p = 0.05. Point definitions: Vegetative season end date = date on which the VWC dropped
to 0 for the first time; VWCrefline = steady−state VWC; drought = days when the VWC is
lower than VWCrefline – 10%; flood = days when the RWV is higher than VWCrefline + 10%;
non−vegetative season end date = date on which the temperature exceeded 0°C for the last
time; warming period start date = date on which the temperature exceeded the warming
threshold for the last time; warming threshold = non−vegetative Tmn – s.d.; warming period
duration = number of days between the warming period start date and transition period start
date; transition period start date = date on which the temperature exceeded −1.0°C for the
first time; transition period duration = number of days between the transition period start
date and vegetative season start date; 0°C date = date on which the temperature exceeded
0°C for the first time; wetting/thawing start date = date on which VWC exceeded 0 m3 m−3
for the first time; wetting/thawing end date = date on which the VWC reached VWCrefline for
the first time; wetting/thawing period duration = number of days between wetting/thawing
start date and wetting/thawing end date; vegetative period start = date on which the temper−
ature remained above 0°C for whole day; vegetative period duration = number of days
since the start of the vegetation period till 31.7.2010.
OTC CCS
top base top base
Vegetative season 2009
End date (VWC) [days] 15.9.2009 ± 1 2.11.2009 ± 48 15.9.2009 ± 0 30.9.2009 ± 1
VWCmin [m3 m−3] 0.01 ± 0.01 0.00 ± 0.00 0.05 ± 0.07 0.00 ± 0.00
VWCmn [m3 m−3] 0.07 ± 0.01a 0.31 ± 0.11b 0.08 ± 0.00a 0.34 ± 0.06 b
VWCmax [m3 m−3] 0.09 ± 0.01a 0.40 ± 0.03b 0.10 ± 0.00a 0.41 ± 0.04b
VWCrefline [m3 m−3] 0.07 ± 0.01a 0.38 ± 0.03b 0.08 ± 0.01s 0.40 ± 0.04b
Drought [days] 5 ± 4 34 ± 46 1 ± 0 18 ± 1
Flood [days] 3 ± 4 0 ± 0 7 ± 2 0 ± 0
Non−vegetative season 2009/10
End date [days] 27.5.2010 ± 15 9.6.2010 ± 4 18.5.2010 ± 1 13.6.2010 ± 3
Tmin [°C] −7.7 ± 1.0 −7.1 ± 0.8 −7.9 ± 0.6 −7.8 ± 0.1
Tmn [°C] −4.2 ± 0.8 −3.5 ± 0.5 −3.9 ± 0.07 −3.8 ± 0.6
Tmax [°C] 0.0 ± 0.0a 0.0 ± 0.1a 0.1 ± 0.1ab 0.2 ± 0.01b
Warming period start date [days] 29.4.2010 ± 4 3.5.2010 ± 4 1.5.2010 ± 4 1.5.2010 ± 5
Warming threshold −6.5 ± 1.0 −5.9 ± 0.7 −6.5 ± 0.7 −6.3 ± 0.4
Warming period duration [days] 15 ± 3 12 ± 2 16 ± 4 16 ± 5
Transition period 2010
Start [days] 15.5.2010 ± 6 15.5.2010 ± 6 16.5.2010 ± 1 17.5.2010 ± 0
Duration [days] 25 ± 1a 32 ± 2b 27 ± 4ab 32 ± 2ab
0°C date [days] 27.5.2010 ± 14 10.6.2010 ± 4 17.5.2010 ± 1 13.6.2010 ± 3
Wetting/thawing start date [days] 20.5.2010 ± 3 30.5.2010 ± 8 18.5.2010 ± 1 28.5.2010 ± 15
Wetting/thawing end date [days] 23.5.2010 ± 4a 17.6.2010 ± 4b 1.6.2010 ± 9ab 18.6.2010 ± 1b
Wetting/thawing period duration [days] 3 ± 1 18 ± 10 14 ± 8 21 ± 16
Unauthenticated
Download Date | 2/16/16 5:42 PM
mid−June 2010 (Table 2). A few days later, the temperature remained above 0°C
for a whole day in the hummock tops and bases in both OTCs and CCSs (vegeta−
tive season start date; Table 2). In spring 2010, hummock tops had a longer vegeta−
tion season for about one week in comparison with bases, probably due to the sig−
nificantly shorter transition period (Table 2). Summer season temperatures in the
OTC hummock tops commonly reached 18–20°C (Fig. 6) and rarely even 30°C.
CCS hummock top temperatures were lower, ranging from 12 to 14°C, and rarely
reached the maximum of 25°C. Mean maximum and mean summer season temper−
atures were higher by 7.6°C and 1.6°C, respectively, in the OTC hummock tops
compared to the CCS ones (Table 2, Fig. 6). Although the temperature data proved
warming effects in the hummock tops in the OTCs, the warming effects were small
in the hummock bases. The mean summer seasonal differences between OTC and
CCS bases were 0.3°C (OTCs > CCSs). Summer−vegetation season temperatures
in the OTC hummock bases reached 8–10°C, while they were slightly lower in
CCS bases, being around 7–8°C (Fig. 6). There were no significant differences in
mean temperature and mean maximum temperature between OTC and CCS hum−
mock bases (Table 2, Fig. 6). Unfortunately, temperature measurements from the
end of summer 2009 and 2010 are not available; however we estimate that temper−
atures had fallen below 0°C in the second half of October.
Volumetric water content (VWC). — The courses of volumetric water con−
tent (VWC) in hummock tops and bases in the OTCs and CCSs, were measured
from July 2009 to September 2010 (Fig. 5). At the end of August and beginning of
September 2009, the water content dropped slightly in both (OTC and CCS) hum−
mock bases. Both OTC and CCS hummock tops were dry at that time. On Septem−
ber 15–16, 2009, the first frost came and water content in both (OTCs and CCSs)
hummock top localities decreased (vegetative season 2009 end date; Table 2). The
Impact of warming in a wet hummock meadow 407
OTC CCS
top base top base
Vegetative season 2010
Start date (T) [days] 9.6.2010 ± 7 16.6.2010 ± 4 12.6.2010 ± 4 17.6.2010 ± 2
Vegetative season duration [days] 52 ± 7 45 ± 4 49 ± 4 44 ± 2
Tmin [°C] −1.3 ± 0.5a
0.0 ± 0.1b
−0.8 ± 0.1ab
−0.2 ± 0.4b
Tmn [°C] 8.4 ± 0.5a 5.3 ± 0.2b 6.8 ± 0.3c 5.0 ± 0.1b
Tmax [°C] 28.2 ± 1.5a 12.7 ± 1.5b 20.6 ± 4.0c 13.3 ± 0.4b
VWCmin [m3
m−3
] 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00
VWCmn [m3 m−3] 0.08 ± 0.01a 0.30 ± 0.01b 0.06 ± 0.02a 0.31 ± 0.06b
VWCmax [m3 m−3] 0.27 ± 0.12 0.41 ± 0.02 0.27 ± 0.19 0.42 ± 0.03
VWCrefline [m3
m−3
] 0.06 ± 0.00a
0.39 ± 0.02b
0.06 ± 0.01a
0.40 ± 0.04b
Drought [days] 27 ± 8 35 ± 7 51 ± 48 45 ± 11
Flood [days] 44 ± 4a
0 ± 0b
33 ± 7a
0 ± 0b
Table 2 – continued.
Unauthenticated
Download Date | 2/16/16 5:42 PM
CCS hummock bases froze at the end of September 2009. At that time, the OTC
hummock bases still remained unfrozen and froze on October 6 to 8, 2009, with the
exception of OTC 1 which froze on December 28, 2009; perhaps the microclimate
was kept warmer due to the insulation effect of snow cover. The vegetative season
of OTC hummock bases was prolonged by three weeks in comparison with OTC
and CCS hummock tops and by one week compared to CCS hummock bases. Dur−
ing winter there were two episodes (mid December 2009 and mid January 2010)
when water content in all hummock localities increased because of snow thaw.
Snow thaw started in all experimental treatments in the second half of May 2010
(wetting/thawing start date; Table 2). Approximately three days later (wetting/
thawing period duration; Table 2), OTC hummock tops were saturated by water
(so−called wetting/thawing end date; Table 2).The CCS hummock tops were also
408 Josef Elster et al.
Fig. 6. Mean daily soil temperature and volumetric water content during the first part of the 2010 veg−
etative season (n = 3 for OTCs, n = 2 for CCSs).
Unauthenticated
Download Date | 2/16/16 5:42 PM
saturated by water in the beginning of June 2010. The thaw started about two
weeks later, in mid−June 2010, in the OTC and CCS hummock bases. The maxi−
mum saturation levels were reached in the second half of June. A period of drought
occurred in July when water content had decreased in all experimental treatments.
From that time, the hummock tops (in both OTCs and CCSs) were completely dry
till the end of the vegetative season. The hummock bases (in both OTC and CCS)
were wetted again at the end of July and water content remained at the same level
for the rest of the summer (Fig. 6). Generally, with the exception of the mean maxi−
mum VWC reflecting the thaw period in early summer 2010, VWC was always
lower in the hummock tops in both OTC and CCSs during both summer seasons
(Table 2, Fig. 6). Mean summer season relative water contents in the OTCs and
CCSs hummock bases were higher by 0.3 and 0.2 m3 m−3 than in the hummock tops
in 2009 and 2010, respectively (Table 2). Affectivity of OTC (warming/in−
crease−decrease relative water content/duration of summer season) is also well
documented in detail in the summer curves (Fig. 6).
The annual VWC course seems to be similar in both seasons. After the thaw
period and full water saturation in June, a drought period occurred in July and was
followed by water re−saturation in August and early September. The high variabil−
ity reflects the heterogeneity of the microenvironment at the experimental site. De−
spite this variability, the data indicate that summer 2009 was significantly wetter
than that for the same period in 2010. The drought periods, as defined in Table 2,
lasted 3 ± 4 days and 21 ± 16 days in 2009 and 2010, respectively (paired t−test, n =
10 for each year, t = −3.364, P = 0.005; only data between July 14, 12:00, and Sep−
tember 12, 10:00 were considered). The large inter−seasonal differences reflect
different water availabilities in individual summer seasons.
Ecophysiological parameters of Nostoc commune s.l.
Changes in fresh weight of the experimental Nostoc commune s.l. colonies dur−
ing both summer seasons are shown in Fig. 7A, B. The first datapoint in each chart
indicates the weight of the colonies in the fully−hydrated state. The colonies were
kept in plastic Petri dishes (installed on July 14, 2009) and ECs (installed on July 2,
2010) at hummock bases in the OTC and CCS treatments. At the beginning of sum−
mer 2009, shortly after installation of the OTCs and CCSs, the weight of the colonies
decreased due to the transfer of the colonies from a well wetted meadow to the Petri
dishes, and further due to desiccation of the hummock meadow (see previous chap−
ter). In the first half of August 2009, the weight slightly increased, however it did not
reach the weight in the fully−hydrated state as at the beginning of the experiment.
The observation in summer 2010 confirmed that the weights of the colonies, in both
the Petri dishes and Exposition Chambers, had followed the water availability in the
wet meadow (Table 3). No statistical differences were observed between OTCs and
CCSs at each date of measurement (Table 3). However, the date on which the mea−
surements were performed, corresponding to the actual conditions at the locality and
Impact of warming in a wet hummock meadow 409
Unauthenticated
Download Date | 2/16/16 5:42 PM
the history of these conditions, affected the weight of the colonies (two−way
ANOVA, n = 9 for each treatment and exposition type, F(7,122) = 10.9, P <0.001 for
Petri dishes and F(3, 64) = 75.9, P <0.001 for ECs, Fig. 7A,B).
Also FV/FM values of Nostoc commune indicate the maximum possible quan−
tum yield of the photosystem II, and thus physiological status of the colonies, fol−
lowed the water availability in the wet meadow (two−way ANOVA, n = 9 for each
treatment and exposition type, F(7,122) = 3.95, P = 0.001 for Petri dishes and F(3,
64) = 18.4, P <0.001 for ECs, Fig. 8A, B, Table 3). The maximum Fv/Fm in the
fully−hydrated state of the colonies is indicated by the first datapoint in both charts.
There were only small differences between OTC and CCS treatments reflecting
the similar conditions at hummocks bases in both OTCs and CCSs. However, the
FV/FM values in the drier summer season of 2010 were slightly higher in compari−
son with the wetter 2009 season (ANOVA/HSD test for unequal n, n = 54 for 2009
and n = 72 for 2010, F(1,122) = 19.800, P <0.001, Table 3). Decrease of FV/FM due
to meadow desiccation observed in the ECs in 2010 was not confirmed in the Petri
dishes, probably due to the experimental setting. The filter paper used for facilita−
410 Josef Elster et al.
Fig. 7. Changes in Nostoc commune s.l. colony weight (mean ± 95% confidence interval) during the
vegetation season(s). A. Colonies in Petri dishes. B. Colonies in Exposition Chambers (n = 9 for each
treatment).
Unauthenticated
Download Date | 2/16/16 5:42 PM
tion of water transfer between the dish and the outside environment probably
served as a temporary additional water supply to the colonies. Also, the enclosure
of the colonies in the Petri dishes reduced evaporation loses compared to the ECs.
The seasonal curves of N. commune nitrogenase activity in colonies enclosed
in the Petri dishes revealed that this was also related to the water availability in the
meadow (two−way ANOVA, n = 9 in each treatment, F(6,112) = 4.90, P <0.001,
Fig. 9A, B, Table 3). However, great variability of nitrogenase activity among the
colonies did not allow detection of any seasonal pattern in the ECs (see the 95%
confidential interval of the first datapoint in each chart indicating the values of
nitrogenase activity in fully−hydrated colonies). As in the cases of weight and
Impact of warming in a wet hummock meadow 411
Table 3
Physiological chatacteristics (mean ± s.d.) of the Nostoc colonies during the 2009 and 2010
growing seasons. The letter indicates homologous groups recognized by Tukey HSD test
for P = 0.05. n – number of samples in given treatment, FV/FM – maximum quantum yield,
NA – nitrogenase activity.
Weight [g] FV/FM NA [nmol g−1 hr−1]
OTC CCS OTC CCS OTC CCS
n = 9 n = 9 n = 9 n = 9 n = 9 n = 9
Petri dishes
14.7.2009 0.76 ± 0.47c 0.62 ± 0.36bc 3.41 ± 10.66 7.54 ± 12.61
22.7.2009 0.14 ± 0.13a 0.20 ± 0.27ab 0.19 ± 0.10 0.17 ± 0.14 0.00 ± 0.00 0.00 ± 0.00
1.8.2009 0.19 ± 0.11a 0.28 ± 0.16ab 0.14 ± 0.12 0.21 ± 0.13 −3.15 ± 51.07 23.73 ± 22.85
11.8.2009 0.27 ± 0.12ab 0.37 ± 0.20abc 0.25 ± 0.13 0.29 ± 0.12 24.05 ± 35.93 31.42 ± 59.40
2.7.2010 0.41 ± 0.14abc
0.47 ± 0.22abc
0.35 ± 0.09 0.29 ± 0.24 42.37 ± 31.07 24.45 ± 31.86
11.7.2010 0.23 ± 0.08ab 0.37 ± 0.22a 0.33 ± 0.15 0.33 ± 0.19 25.85 ± 32.34 27.54 ± 41.91
27.8.2010 0.19 ± 0.05a 0.22 ± 0.14a 0.33 ± 0.08 0.32 ± 0.21 5.67 ± 9.37 3.66 ± 10.79
10.9.2010 0.21 ± 0.10a 0.21 ± 0.13a 0.31 ± 0.08 0.32 ± 0.14 1.06 ± 9.18 1.28 ± 6.29
Exposition Chambers
29.6.2010 7.94 ± 2.54 6.10 ± 1.91 0.42 ± 0.06 0.49 ± 0.10 20.81 ± 10.08 18.64 ± 9.14
10.7.2010 0.38 ± 0.13 0.29 ± 0.10 0.13 ± 0.06 0.08 ± 0.16 106.4 ± 285.4 6.63 ± 7.78
28.8.2010 1.27 ± 1.56 1.27 ± 0.90 0.37 ± 0.23 0.42 ± 0.20 13.83 ± 12.44 14.07 ± 11.47
11.9.2010 1.60 ± 1.67 1.33 ± 0.71 0.38 ± 0.10 0.35 ± 0.12 9.39 ± 9.15 6.51 ± 4.72
Table 4
Correlation between environmental variables and Nostoc colonies ecophysiological pa−
rameters in ECs. The statistically significant correlations are marked by bold font.
OTC CCS
Weight FV/FM NA Weight FV/FM NA
Temperature
−0.334 −0.556 0.198 −0.594 −0.654 −0.199
p = 0.088 p = 0.003 p = 0.321 p = 0.009 p = 0.003 p = 0.430
VWC
0.313 0.447 0.049 0.470 0.228 0.282
p = 0.111 p = 0.019 p = 0.807 p = 0.049 p = 0.362 p = 0.257
Water level
−0.363 −0.5749 0.274 −0.591 −0.596 −0.258
p = 0.063 p = 0.002 p = 0.167 p = 0.010 p = 0.009 p = 0.301
Unauthenticated
Download Date | 2/16/16 5:42 PM
FV/FM, only negligible differences in nitrogenase activity between OTC and CCS
treatments were observed.
There was no significant correlation between the microclimate parameters and
ecophysiological features of the Nostoc colonies kept in Petri dishes in both OTCs
and CCSs.
Table 4 shows the correlations between the environmental variables (tempera−
ture, VWC and water level depth) and ecophysiological parameters (colony weight,
FV/FM and NA) of the Nostoc colonies kept in ECs, OTCs and CCSs. The nitro−
genase activity (NA) was not significantly affected by microclimate parameters for
colonies enclosed in ECs or water level depth for colonies enclosed in both OTCs
and ECs. However, in OTCs, the FV/FM was negatively correlated with temperature
and water level depth, but positively with VWC. This was also the case for CCSs in
general. However, for the colonies enclosed in the ECs in the CCSs, FV/FM was neg−
atively correlated with temperature and colony weight, but positively with VWC.
412 Josef Elster et al.
Fig. 8. Changes in Nostoc commune s.l. colony maximum quantum yield (FV
/FM
; mean ± 95% confi−
dence interval) during the vegetation season(s). A. Colonies in Petri dishes. B. Colonies in Exposition
Chambers (n = 9 for each treatment).
Unauthenticated
Download Date | 2/16/16 5:42 PM
These data indicate that the colonies enclosed in Petri dishes were protected
from the microclimatic effects, probably due to limited contact between the inte−
rior of the Petri dish and its surrounding environment. There may also be a green−
house effect inside the dish. This type of exposition should not be used in future
experiments. The data from the ECs indicate that the colonies were negatively af−
fected by temperature and water deficiency in the meadow; this type of exposition
should be preferred in future experiments.
Discussion
Arctic hydro−terrestrial habitats (wetlands, including hummock wet meadows)
are important ecosystems in an arid and cold tundra environment (Elster 2002;
ACIA 2004). They cover large areas of the Earth's surface and play a fundamental
role in the global climate system (Wieder 2001; ACIA 2004). Wet hummock
meadows are highly productive plant ecosystems where carbon uptake prevails
Impact of warming in a wet hummock meadow 413
Fig. 9. Changes in Nostoc commune s.l. colony nitrogenase activities (NA; mean ± 95% confidence
interval) during the vegetation season(s). A. Colonies in Petri dishes. B. Colonies in Exposition
Chambers (n = 9 for each treatment).
Unauthenticated
Download Date | 2/16/16 5:42 PM
above carbon loss from the soil (Rinnan et al. 2007; Biasi et al. 2008). In addition,
the high Arctic hummock wet meadows are a prevalent and highly sensitive type
of polar hydro−terrestrial ecosystem affected by climate change (ACIA 2004).
Water level and water physico−chemical parameters
The seasonal courses of climatic and microclimatic conditions of Petuniabukta
are discussed in detail in this issue (Láska et. al., this issue). However, from our
measurements it is obvious that the range of winter snow accumulation and its
melting at the beginning of summer in the upper part of the studied locality is the
principal factor influencing water availability in wet meadow hummock tundra.
On the contrary, water availability is better balanced at the end of summer proba−
bly because of the active layer and melting of the upper part of the permafrost. A
similar pattern of seasonal water availability was demonstrated in small tundra
lakes in the eastern part of Petuniabukta (Zwolinski et al. 2007).
The wet hummock meadow lies on a steplike system of raised marine terrace
which is characteristic for the post−Pleistocene shoreline layout of Svalbard
(Salvigsen 1984). Geologically, the Petuniabukta area is composed of Middle and
Upper carboniferous carbonates, anhydrites, gypsum, sandstone and conglomer−
ates (Dallmann et al. 1994). The lower wet hummock meadow terrace is com−
prised of sandy and gravel sediment originating from local waste material. In pre−
vious studies of water physico−chemical parameters of the Petuniabukta freshwa−
ter ecosystem (Zwolinski et al. 2007; Paluszkiewicz et al. 2008; Zwolinski et al.
2008) it was shown that marine aerosols, predominantly Na and Cl ions, are com−
monly present in higher concentrations in the wetland. In our samples, higher Na
and Cl contents were recorded only in water inflowing to the meadow from snow
melt at the beginning of summer 2009. Yellow−brown precipitates were recorded
in both the eastern and western parts of Petuniabukta bay, in the wet meadows and
even in shallow pools and lake bottoms, and on the surfaces of mosses, vascular
plants and even Nostoc commune s.l. mats (Zwolinski et al. 2007; Paluszkiewicz et
al. 2008; Zwolinski et al. 2008). Higher concentrations of calcium, magnesium
and sulphate in our samples from the experimental field site (compared to inflow
water) confirmed their precipitation in the wet hummock meadow. Concentrations
of the biogenic elements (N, P) in the inflow water were higher than in the
meadow. Hummock wet meadow serves as a biological filter of biogenic elements
including CO2 and serves as a sink for these biogenic elements in the high Arctic
ecosystem (Johansson et al. 2006; Oberbauer et al. 2007).
Microclimatic parameters
Our results and those of several recent studies (Nordstroem et al. 2001; Renner−
malm et al. 2005; Sullivan and Welker 2005; Sullivan et al. 2008) clearly demon−
strate that, in the hydro−terrestrial environment of wet hummock meadows, the ef−
414 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM
fect of passive open−top warming chambers clearly depends upon microtopography.
The dry tops of hummocks have completely different microclimatic conditions in
comparison with wet hummock bases. In addition, there was also a time shift in the
warming of dry tops and wet bases of hummocks. In situ warming experiments sim−
ulated by the OTCs even extended these microclimatic differences.
Because of faster melt inside of the OTC hummock tops, positive temperatures
were measured at a depth of ~2 cm below the surface (reached already in
mid−May) for several days earlier in the OTCs in comparison with the CCSs. In ad−
dition, during melt time the temperature differences between hummock tops in
OTC and CCS were as great as 11°C; these differences were not so dramatic later
in the season. Similar results of spring temperature differences between warmed
and control sites were demonstrated in a moist, acid, tussock tundra site near
Toolik Lake, Alaska. Immediately following snow melt in 2002, the differences in
soil surface temperatures between warmed and ambient plots approached a maxi−
mum of 1.5°C. Later, daily mean temperatures differed by only 0.84°C (Sullivan
and Welker 2005).
Mean summer temperatures differed by 1.6°C between OTC and CCS hum−
mock tops. During most of the summer temperatures in the OTC hummock tops
were at 18–20°C and rarely reached 30°C, while in the control boxes they were
12–14°C and rarely 25°C. Mean maximum temperatures were higher by 7.6°C in
OTC hummock tops than in control boxes. On the contrary, the warming effect was
much lower in the wet bases of hummock in the OTC treatments with mean summer
differences between hummock bases in OTCs and CCSs were 0.3°C. Summer tem−
peratures in hummock bases reached 8–10°C in OTCs and 7–8°C in CCSs. Temper−
ature differences between hummock tops and bases reached 8–10°C in OTCs and
4–6°C in CCSs treatments. Although the ITEX warming open top chambers were
used very commonly in various Arctic, Antarctic and alpine terrestrial ecosystems
(Marion et al. 1997; Hollister and Webber 2000; Kudo and Suzuki 2002; Wada et al.
2002; Bokhorst et al. 2008; Walker et al. 2008; Rinnan et al. 2009a; Rinnan et al.
2009b; Simmons et al. 2009), the OTC warming experiments are more rare in vari−
ous shallow wetland habitat types (Nordstroem et al. 2001; Rennermalm et al. 2005;
Sullivan and Welker 2005; Sullivan et al. 2008). In NW Greenland wetlands
(Sullivan et al. 2008), the microtopography of the sites is a mosaic of hummocks,
which extend above the water table, and hollows, which lie beneath 5–15 cm of wa−
ter during the short growing season. Air temperature (20 cm above surface) in OTC
was about 2°C higher than in the control. Hummock soil temperature was higher
than air temperature. Hollows temperature was colder than air temperature. Midday
soil temperatures were generally warmer in hummocks than in hollows. These re−
sults principally follow our data and characterize the warming effects of OTC in a
wetland ecosystem (high Arctic wet meadows).
A wide spectrum of ecological and physiological studies have shown that, in
terms of environmental priorities, demands for moisture (water in a liquid form)
Impact of warming in a wet hummock meadow 415
Unauthenticated
Download Date | 2/16/16 5:42 PM
precede demands for nutrients, which in turn, precede demands for high tempera−
tures (Svoboda and Henry 1987; Kennedy 1993; Elster 2002; Elster and Benson
2004). The courses of volumetric water content showed that OTC hummock bases
remained unfrozen until October or even December. The vegetative season of
OTC hummock bases was prolonged by one and/or three weeks in comparison
with OTC and CCS hummock tops, and CCS hummock bases. Episodes with snow
thaw can occur even during winter, meaning that liquid water can be available for
poikilohydric biota (ecological opportunists) living in wet meadows.
Ecophysiological parameters of Nostoc commune s.l.
Desiccation stress represent the most severe injury in the polar regions (Davey
1989). Cyanobacterial primary production and nitrogen fixation have to follow the
moisture conditions (Liengen and Olsen 1997; Vincent 2000; Novis et al. 2007;
Kvíderová et al. 2011). However, Davey (1989), Hawes et al. (1992), Novis et al.
(2007) and Kvíderová et al. (2011) have shown that even a partially hydrated
Nostoc commune s.l. colony is capable of photosynthesis and nitrogen fixation in
Arctic hydroterrestrial habitats. This can provide a big advantage for Nostoc
commune s.l. itself and also for the ecological functioning of the hummock mead−
ows when water supply fluctuates widely. In situ experiments at the same locality
showed that if colony water loss was less than ca 40% of its fully hydrated weight,
no or only minor desiccation damages were observed. However, when weight loss
exceeded ca 40%, the colony starts to shrink and changes in the mucilage structure
begins. Nitrogenase activity declined slowly when water loss exceeded 70% and
diminished completely at a weight loss of ca 80%. The photochemical activity re−
mained unaffected till the colonies lost ca 80% of their original weight and was de−
tectable till weight loss of ca 90% (Kvíderová et al. 2011). However, measure−
ments in summer 2010 confirmed that weights of Nostoc commune s.l. colonies,
their maximum quantum yield of photosystem II, and thus physiological status of
the colonies and their nitrogenase activity in both Petri dishes and Exposition
Chambers followed the water availability in the wet meadow. No statistical differ−
ences were observed between OTCs and CCSs at each date of measurement, prob−
ably due to only small differences in the microclimatological parameters at hum−
mock bases in the OTCs and CCSs. Our in situ one and/or two seasons warming
experiments consisting of passive open−top chambers, which raise ambient tem−
perature and slightly changed water availability in experimental treatments of wet
hummock meadow tundra, did not influence the photochemical processes and
nitrogenase activity of Nostoc commune s.l. colonies.
Proposition for further research
Ecosystem manipulation of the wet meadow where Nostoc commune s.l. colo−
nies produce high biomass in the wet bases of hummocks resulted in mean summer
416 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM
temperature being only 0.3°C higher in OTCs than control treatments. However,
short−term snow−thaw episodes occurred in the OTCs during winter. Earlier snow
melt in the spring and later freezing over in fall also significantly prolong the vege−
tation season (by several days/weeks) in the manipulated environment of the
OTCs. Liquid water is available for the growth of Nostoc colonies during this time.
However, we do not have photochemistry and nitrogenase activity measurements
in the non−vegetation season. On the basis of our results and experiences of this pi−
lot study of Nostoc commune s.l. ecophysiological reaction to simulated warming,
we propose the following research:
• Continuous whole year measurements of the ecophysiological parameters of
Nostoc colonies, which may show growth of Nostoc colonies during the non−
vegetation season.
• Continue with OTC manipulation using the same technique and measure the
ecophysiological response of Nostoc colonies to warming over a long period of
time (5 to 10 years experiment).
• Technical adjustment of OTCs with the aim of increasing the temperature by up
to 3 to 5°C at the hummock bottom. This can be done by covering the OTC’s
(with the same material by which the OTCs are constructed), with creation of
holes. The number and size of holes should be prepared according to an OTC’s
temperatures.
• Replacement of the Petri dishes by more open exposition chambers like our ECs
or direct fixation of a colony to the substrate by a wire mesh. The Petri dishes,
even with drilled holes, are not suitable for such experiments, since the transport
of water and heat between the interior of the Petri dish and surrounding environ−
ment is restricted.
Acknowledgement. — We are indebted to Mrs. Jana Šnokhousová for her technical assis−
tance and Keith Edwards for language revision. Our study was supported by grants from the
Czech Ministry of Education, Youth and Sport of the Czech Republic (INGO LA341,
LM2010009 CzechPolar and KONTAKT ME934). We also thank two reviewers: Professor
Josef Svoboda, University of Toronto and Associate Professor Falk Huettmann, University of
Alaska−Fairbanks for useful comments.
References
ACIA. 2004. Impacts of warming climate: Arctic climate impact assessment. Cambridge University
Press, Cambridge: 1039 pp.
BIASI C., MEYER H., RUSALIMOVA O., HÄMMERLE R., KAISER C., BARANYI C., H. D., LASH−
CHINSKY N., BARSUKOV P. and RICHTER A. 2008. Initial effect of experimental warming on
carbon exchange rates, plant growth and microbial dynamics of a lichen−rich dwarf shrub tundra
in Siberia. Plant and Soil 307: 191–205.
BOKHORST S., HUISKES A., CONVEY P., VAN BODEGOM P.M. and AERTS R. 2008. Climate change
effects on soil arthropod communities from the Falkland Islands and the Maritime Antarctic. Soil
Biology & Biochemistry 40: 1547–1556.
Impact of warming in a wet hummock meadow 417
Unauthenticated
Download Date | 2/16/16 5:42 PM
CALLAGHAN T.V., BJÖRN L.O., CHERNOV Y., CHAPIN T., CHRISTENSEN T.R., HUNTLEY B., IMS
R.A., JOHANSSON M., JOLLY D., JONASSON S., MATVEYEVA N., PANIKOV N., OECHEL W.,
SHAVER G., ELSTER J., HENTTONEN H., LAINE K., TAULAVUORI K., TAULAVUORI E. and
ZÖCKLER C. 2004a. Biodiversity, distributions and adaptations of Arctic species in the context
of environmental change. Ambio 33: 404–417.
CALLAGHAN T.V., BJÖRN L.O., CHERNOV Y., CHAPIN T., CHRISTENSEN T.R., HUNTLEY B., IMS
R.A., JOHANSSON M., JOLLY D., JONASSON S., MATVEYEVA N., PANIKOV N., OECHEL W.,
SHAVER G., ELSTER J., JONSDOTTIR I.S., LAINE K., TAULAVUORI K., TAULAVUORI E. and
ZÖCKLER C. 2004b. Responses to projected changes in climate and UV−B at the species level.
Ambio 33: 418–435.
CHAPIN F.S.I. and SHAVER G.R. 1985. Individualistic growth response of tundra plant species to en−
vironmental manipulation in the field. Ecology 66: 564–576.
CONVEY P., PUGH P.J.A., JACKSON C., MURRAY A.W., RUHLAND C.T., XIONG F.S. and DAY T.A.
2002. Response of Antarctic terrestrial microarthropods to long−term climate manipulation.
Ecology 83: 3130–3140.
DALLMANN W.K., OHTA Y., BIRJIKOV A.S., KARMOUSHENKO E.P. and SIROTKIN A.N. 1994. Geo−
logical map of Svalbard 1:100000, sheet C7G Dicksonfjorden. Norsk Polarinstitut.
DAVEY M.C. 1989. The effects of freezing and desiccation on photosynthesis and survival of terres−
trial Antarctic algae and cyanobacteria. Polar Biology 10: 29–36.
DAVEY M.C. and ROTHERY P. 1992. Factors causing the limitation of growth of terrestrial algae in
maritime Antarctic during late summer. Polar Biology 12: 595–601.
DAY T.A., RUHLAND C.T., GROBE C.W. and XIONG F. 1999. Growth and reproduction of Antarctic vas−
cular plants in response to warming and UV radiation reductions in the field. Oecologia 119: 24–35.
ELSTER J. 2002. Ecological classification of terrestrial algal communities in polar environments. In:
L. Beyer and M. Bötler (eds) Geoecology of Antarctic ice−free coastal lanscapes. Springer−
Verlag, Berlin, Heidelberg: 303–326.
ELSTER J. and BENSON E.E. 2004. Life in the polar terrestrial environment with a focus on algae and
cyanobacteria. In: B.J. Fuller, N. Lane and E.E. Benson (eds) Life in the frozen state. CRC Press,
Boca Raton, London, New York, Washington, D.C.: 111–150.
ELSTER J., SVOBODA J. and KANDA H. 2001. Controlled environmental platform used in temperature
manipulation study of a stream periphyton in the Ny−Ålesund, Svalbard. In: J. Elster, J.
Seckbach, W.F. Vincent and O. Lhotský (eds) Algae and extreme environments. Cramer,
Stuttgart: 63–75.
HAWES I., HOWARD−WILLIAMS C. and VINCENT W. 1992. Desiccation and recovery of Antarctic
cyanobacterial mats. Polar Biology 12: 587–594.
HENRY G.H.R. and SVOBODA J. 1986. Dinitrogen fixation (acetylene reduction) in high Arctic sedge
meadow communities. Arctic and Alpine Research 18: 181–187.
HOLLISTER R.D. and WEBBER P.J. 2000. Biotic validation of small open−top chambers in a tundra
ecosystem. Global Change Biology 6: 835–842.
HOWARD C.S. 1933. Determination of total dissolved solids in water analysis. Industrial & Engi−
neering Chemistry Analytical Edition 5: 4–6.
JOHANSSON T., MALMER N., CRILL P.M., FRIBORG T., AKERMAN J.H., MASTEPANOV M. and
CHRISTENSEN T.R. 2006. Decadal vegetation changes in a northern peatland, greenhouse gas
fluxes and net radiative forcing. Global Change Biology 12: 2352–2369.
KARLBERG B. and TWENGSTROM S. 1983. Application based on gas diffusion and flow injection
analysis. Focus 6: 14.
KENNEDY A.D. 1993. Water as a limiting factor in the Antarctic terrestrial environment: A biogeo−
graphical synthesis. Arctic, Antarctic and Alpine Research 25: 308–315.
KENNEDY A.D. 1995. Antarctic terrestrial ecosystem response to global environmental change. An−
nual Review of Ecology and Systematics 26: 683–704.
418 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM
KUDO G. and SUZUKI S. 2002. Warming effects on growth, production, and vegetation structure of
alpine shrubs: a five−year experiment in northern Japan. Oecologia 135: 280–287.
KVÍDEROVÁ J., ELSTER J. and ŠIMEK M. 2011. In situ response of Nostoc commune s.l. colonies to
desiccation in Central Svalbard, Norwegian High Arctic. Fottea 11: 87–97.
LIENGEN T. and OLSEN R.A. 1997. Nitrogen fixation by free−living cyanobacteria from different
coastal sites in a High Arctic tundra, Spitzbergen. Arctic and Alpine Research 29: 470–477.
MADSEN B.C. and MURPHY R.J. 1981. Flow injection and photometric determination of sulfate in
rainwater with methylthymol blue. Analytical Chemistry 53: 1924–1926.
MARION G.M., HENRY G.H.R., FRECKMAN D.W., JOHNSTONE J., JONES G., JONES M.H., LÉVESQUE
E., MOLAU U., MØLGAARD P., PARSONS A.N., SVOBODA J. and VIRGINIA R.A. 1997. Open−top
designs for manipulating field temperature in high−latitude ecosystems. Global Change Biology 3:
20–32.
MAXWELL B. 1992. Arctic climate: Potential for change under global warming. In: F.S.I. Chapin,
R.L. Jeffries, J.F. Reynold, G.R. Shaver and J. Svoboda (eds) Arctic esosystems in a changing
climate. Academic Press, San Diego: 11–34.
MAXWELL K. and JOHNSON G.N. 2000. Chlorophyll fluorescence – a practical guide. Journal of Ex−
perimental Botany 51: 659–668.
MOLAU U. and MØLGAARD P. (eds) 1996. ITEX manual (2nd. edition). Danish Polar Centre, Copen−
hagen: 85 pp.
NORDSTROEM C., SOEGAARD H., CHRISTENSEN T.R., FRIBORG T. and HANSEN B.U. 2001. Sea−
sonal carbon dioxide balance and respiration of a high−arctic fen ecosystem in NE−Greenland.
Theoretical and Applied Climatology 70: 149–166.
NOVIS P.M., WHITEHEAD D., GREGORICH E.G., HUNT J.E., SPARROW A.D., HOPKINS D.W.,
ELBERLING B. and GREENFIELD L.G. 2007. Annual terrestrial fixation in terrestrial populations
of Nostoc commune (Cyanobacteria) from an Antarctic dry valley is driven by temperature.
Global Change Biology 13: 1224–1237.
OBERBAUER S.F., TWEEDIE C.E., WELKER J.M., FAHNESTOCK J.T., HENRY G.H.R., WEBBER P.J.,
HOLLISTER R.D., M.D. W., KUCHY A., ELMORE E. and STARR G. 2007. Tundra CO2 fluxes in
response to experimental warming across latitudinal and moisture gradient. Ecological Mono−
graphs 72: 221–238.
PALUSZKIEWICZ R., MAZUREK M., RACHLEWICZ G. and ZWOLIŃSKI Z. 2008. Hydrogeochemia
zagłębień tundrowych na podniesionych terasach morskich, Petuniabukta, Spitsbergen Środkowy
[Hydrogeochemistry of tundra basins on the raised marine terraces, Petuniabukta, Central Spits−
bergen]. In: A. Kowalska and J. Pereyma (eds) Środowisko przyrodnicze obszarów polarnych.
Uniwersytet im. Adama Mickiewicza, Poznań: 141–149.
PARSONS L.R. and BANDARANAYAKE W.M. 2009. Performance of a new capacitance soil moisture
probe in a sandy soil. Soil Science Society of America Journal 73: 1378–1385.
PAUL E.A. and CLARK F.E. 1996. Soil microbiology and chemistry. Academic Press, Toronto: 340 pp.
PROCTOR C.M. and HOOD A.R. 1954. Determination of phosphate in sea water by an iso−butanol ex−
traction procedure. Journal of Marine Research 13: 112–132.
RENNERMALM A.K., SOEGAARD H. and NORDSTROEM C. 2005. Interannual variability in carbon di−
oxide exchange from a high Arctic fen estimated by measurements and modeling. Arctic, Ant−
arctic and Alpine Research 37: 545–556.
RINNAN R., MICHELSEN A., BÅÅTH E. and JONASSON S. 2007. Fifteen years of climate change ma−
nipulations alter soil microbial communities in a subarctic heath ecosystem. Global Change Bi−
ology 13: 28–39.
RINNAN R., ROUSK J., YERGEAZ E., G.A. K. and BÅÅTH E. 2009a. Temperature adaptation of soil
bacterial communities along an Antarctic climate gradient: predicting responses to climate
warming. Global Change Biology 15: 2615–2625.
Impact of warming in a wet hummock meadow 419
Unauthenticated
Download Date | 2/16/16 5:42 PM
RINNAN R., STARK S. and TOLVANEN A. 2009b. Responses of vegetation and soil microbial commu−
nities to warming and simulated herbivory in a subarctic heath. Journal of Ecology 97: 788–800.
ROHÁČEK K. and BARTÁK M. 1999. Technique of the modulated chlorophyll fluorescence: basic
concepts, useful parameters, and some applications. Photosynthetica 37: 339–363.
ROLPH S.G. 2003. Effects on a ten−year climate warming experiment on nitrogen cycling in the high
arctic tundra. MSc. thesis, University of British Columbia, Vancouver: 61 pp.
RŮŽIČKA J. and HANSEN E.H. 1981. Flow injection analysis. John Wiley, New York: 528 pp.
RŮŽIČKA J., STEWART J.W.B. and ZAGATTO E.A. 1976. Flow injection analysis: Part IV. Stream
sample splitting and its application to the continuous spectrophotometric determination of chlo−
ride in brackish waters. Analytica Chimica Acta 81: 387–396.
SALVIGSEN O. 1984. Occurrence of pumice on raised beaches and Halocene shoreline displacement
in the inner Isfjorden area, Svalbard. Polar Research 2: 107–113.
SHAVER G.R., CANADELL J., CHAPIN F.S.I., GUREVITCH J., HARTE J., HENRY G.H.R., INESON P.,
JONASSON S., MELILLO J., PITELKA L. and RUSTAD L. 2000. Global warming and terrestrial
ecosystems: A conceptual framework for analysis. Bioscience 50: 871–882.
SIMMONS B.L., WALL D.H., ADAMS B.J., AYRES E., BARRETT J.E. and VIRGINIA R.A. 2009. Long−
term experimental warming reduces soil nematode populations in the McMurdo Dry Valleys,
Antarctica. Soil Biology & Biochemistry 41: 2052–2060.
SOKOLOFF V.P. 1933. Water of crystallization in total solids of water analysis. Industrial & Engi−
neering Chemistry Analytical Edition 5: 336–337.
STEWART W.D.P., FITZGERALD G.P. and BURRIS R.H. 1967. In situ studies of N2 fixation using the
acetylene reduction technique. Proceedings of the National Academy of Sciences of the United
States of America 58: 2071–2073.
STRATHDEE A.T. and BALE J.S. 1993. A new cloche design for evaluating temperature in polar ter−
restrial ecosystem. Polar Biology 13: 577–580.
SULLIVAN P., ARENS S., CHIMNER R. and WELKER J. 2008. Temperature and microtopography
interact to control carbon cycling in a High Arctic fen. Ecosystems 11: 61–76.
SULLIVAN P.F. and WELKER J.M. 2005. Warming chambers stimulate early season growth of an
Arctic sedge: results of a minirhizotron field study. Oecologia 142: 616–626.
SVOBODA J. and HENRY G.H.R. 1987. Succession in marginal Arctic environments. Arctic, Antarctic
and Alpine Research 4: 373–384.
VINCENT W.F. 2000. Cyanobacterial dominance in the Polar Regios. In: B.A. Whitton and M. Potts
(eds) The ecology of cyanobacteria. Kluwer Academic Publishers, Dordrecht: 321–340.
WADA N., SHIMONO M., M. M. and KOJINA S. 2002. Warming effects on shoot development growth
and biomass production in sympatric evergreen alpine dwarf shrub Empetrum nigrum and
Loiseleuria procumbens. Ecological Research 17: 125–132.
WALKER J.K.M., EGGER K.N. and HENRY G.H.R. 2008. Long−term experimental warming alters ni−
trogen−cycling communities but site factors remain the primary drivers of community structure
in high Arctic tundra soil. The ISME Journal 2: 982–995.
WIEDER R.K. 2001. Past, present and future peatland carbon balance: an empirical model based on
210Pb−dated cores. Ecological Applications 11: 327–342.
ZWOLINSKI Z., MAZUREK M., PALUSZKIEWICZ R. and RACHLEWICZ G. 2008. The matter fluxes in
the geoecosystem of small tundra lakes, Petuniabukta coast, Billefjorden, Central Spitsbergen.
Zeitschrift für Geomorphologie (Suppl. 1) 52: 79–101.
ZWOLINSKI Z., RACHLEWICZ G., MAZUREK M. and PALUSZKIEWICZ R. 2007. The geoecological
model for small tundra lakes, Spitzbergen. Landform Analysis 5: 113–118.
Received: 19 July 2011
Accepted: 20 March 2012
420 Josef Elster et al.
Unauthenticated
Download Date | 2/16/16 5:42 PM