Introduction
Cavernous weathering (tafoni formation) is a classic example of
a feature most likely attributable to multifactorial processes.
Tafoni (singular, tafone) occur in hot and arid conditions
(Campbell, 1999), coastal environments (Mottershead and
Pye, 1994), and even in Antarctica (Strini et al., 2008).
Turkington and Paradise (2005) proposed detailed conceptual
models of the evolution of tafoni in sandstones. They summarize
the distribution as follows: ‘[…] They occur under cold,
temperate, hot, humid and arid environments, and are found
on a variety of rock types.’ However, an arid period is generally
considered necessary for tafoni development (Brandmeier
et al., 2011). The governing processes might be physical or
chemical or a combination of the two (French and Guglielmin,
2000; Dorn et al., 2013); various mechanisms such as salt
weathering, dissolution, alternating wetting and drying, temperature
fluctuations and aeolian processes have been proposed
(Mol and Viles, 2012). Tafoni development can only be explained
by a complex combination of processes (Turkington
and Phillips, 2004), although many authors believe salt
weathering to be the most important single process in many
cases (Rodriguez-Navarro et al., 1999; Brandmeier et al.,
2011) and moisture is often emphasized as a main driving factor
(Mellor et al., 1997).
To understand the development of tafoni, more data need to
be collected on the environmental controls of the contributing
weathering processes (e.g. salt weathering and hydration), in
particular the temperature and moisture regimes and distribution
of salts within tafoni and the surrounding rock. Our aim is
to investigate the control parameters of tafoni formation in
Tafraoute, Morocco, using an integrative approach which combines
high-resolution temporal and spatial measurements of
temperature, moisture and salt distribution. We aim to elucidate
the small-scale interaction of temperature, rock moisture and
salt levels and fluctuations that contribute to cavernous
weathering in granite. Besides analyzing how, in the Tafraoute
area, moisture and salts contribute to tafoni development, the
field data can be used to underpin (or challenge) theoretical
and mathematical models of tafoni formation.
Theories and Causes of Tafoni Weathering
Temperature and moisture regimes and salt distributions have all
been proposed as important drivers of tafoni development. The
effectiveness of ’pure’ insolation or temperature weathering has
been discussed for decades and will not be reviewed here in detail.
In recent years, authors such as Hall and André (2001) and
Hall et al. (2007) have pointed out that insolation is an underrated
process and introduced the terms thermal fatigue and thermal
shock. Rapid temperature fluctuations of >2 K minÀ1
are
regarded to be particularly important in causing rock breakdown,
although the validity of this threshold has been recently challenged
by Boelhouwers and Jonsson (2013). Nonetheless, there
are still significant shortcomings in the availability of highresolution
temperature data (Sumner et al., 2009). Temperature
fluctuations in isolation have never been considered as the main
agent of tafoni formation; but they might trigger moisture and salt
movement. Two recent developments could contribute to novel
datasets of temperature fluctuations within and outside tafoni:
(1) the use of many cheap microsensors (iButtons); and (2) infrared
thermography, which has, to our knowledge, never been
reported on in a published paper on tafoni.
Moisture content in porous rock outcrops is extremely variable
in time and space; quantitative assessment of moisture
concentration is still challenging and there are few results from
field sites. Most earlier field results were derived from weighing
and drying samples (Hall, 1986, 1991). Matsukura and
Takahashi (2000) measured moisture of sandstone blocks using
an infrared moisture meter, while Eklund et al. (2013) tested
and demonstrated the potential use of handheld capacitance
and resitance-based moisture meters. For one or two decades
geophysical techniques, such as 2D-geoelectrical measurements
have been applied to natural rock (Sass, 2003, 2005;
adapted by Mol and Viles, 2010) and the approach has been
adapted to building stone using non-invasive adhesive electrodes
(Sass and Viles, 2010a, 2010b) for monitoring moisture
ingress into the stonework. Moisture distribution is commonly
regarded as one of the most important determinants of tafoni
formation (Mellor et al., 1997) and moisture gradients may be
crucial for the concentration of salts in certain topographic
conditions (Huinink et al., 2004). The application of borehole
humidity sensors in this present investigation is a novel
approach that has not been used before in this context.
Salts may derive from various sources (e.g. weathering of bedrock,
road salts, atmospheric aerosols, capillary rise from groundwater,
and coastal spray) and exert physical pressure on pore
walls by thermal expansion, hydration, and crystallization
(Goudie and Viles, 1997). They are a particularly important agent
of weathering in arid, semiarid and urban environments and are
considered highly important for tafoni evolution (RodriguezNavarro
et al., 1999). Salt weathering processes have been
reviewed by Doehne (2002) and intensively investigated in the
context of building stone deterioration and conservation
(Maurício et al., 2006; Dionisio et al., 2013). Avariety of different
salts is thought to be involved in stone decay (Turkington et al.,
2003). A very heterogeneous distribution of dissolved salts within
porous rock and stonework has to be considered in most environments.
Capillary uptake of groundwater and atmospheric humidity
are the main pathways by which salts enter rock and
stonework (Martinho et al., 2012). Thus, uptake of moisture and
the behaviour of salt mixtures in the pores are mutually interconnected
(Tsui et al., 2003; Franzen and Mirwald, 2009).
Study Sites
The Anti-Atlas in southern Morocco is a Cambrian mountain
belt running from NE to SW. Tectonic windows are
found in this mountain belt in which Proterozoic rock complexes
crop out. One of these windows is the Kerdous
Massif in the western Anti-Atlas in which lies the Tasrirt
Plateau at an altitude of 1000 m (Soulaimani and Piqué,
2004). Our study site is located on this plateau south of
the small town of Tafraoute (GPS: 30°09′70″N; 09°13′15″
W). The predominant rock in the area is the Tafraoute
Granite (Figure 1).
Figure 1. Overview of the study area. (a) location; (b) simplified geological map (taken from Service Géologique du Maroc, 1985); (c) location of the
investigated sites in the study area; (d): impression of the granite landscape (taken from site 1 looking north).
474
Copyright © 2015 John Wiley & Sons, Ltd.
In the field, the 550 million years old Tafraoute Granite is
reddish-brown and rather coarse. Feldspar crystals of mm to
cm size can be quite easily removed from the rough surfaces,
and the rather low resistance to drilling (we drilled up to 50
cm deep with comparatively little effort) demonstrates that the
rock is highly weathered in its entirety. This is confirmed by
our own laboratory measurements: while typical textbook
values for density of granite are between 2600 and 2800 kg
mÀ3
and 0.4 to 1.5% for porosity (Materialarchiv, 2014), samples
from the area showed a density of 2240 kg mÀ3
and a
porosity of 8%. Details of the structural, geochronological and
geochemical data on Tafraoute Granite can be found in
Hassenforder (1985, 1987); Barbey et al. (2004) and Benziane
(2007). In the granite landscape within an area of approximately
100 km2
, countless tafoni with different characteristics in terms
of their shape, size and exposure occur especially on a number
of tor-like outcrops. Pilot survey mapping revealed no specific
distribution in terms of aspect or elevation. We chose six sites
with tafoni of different sizes and shapes which are shown in
Figure 2.
Owing to logistical constraints we focused most data collection
on site 1 and in particular on the west-facing tafoni at this
site. Owing to the completeness of the dataset obtained here,
most of the results presented are derived from this site. Selected
results from other sites are mentioned either to underpin the
results of site 1 or to demonstrate deviations between sites.
There is no meteorological station directly in Tafraoute and so
the climate parameters had to be estimated from different internet
sources (Figure 3). As none of these sources give precise details on
the origin of the data we only calculated the mean value for temperature
and precipitation. Based on this calculation, Tafraoute
has a mean annual air temperature of approx. 16°C. Precipitation
is c. 250 mm yrÀ1
with a pronounced winter maximum (= BSk
climate according to Köppen-Geiger). Frost may occur occasionally
in winter nights. These figures broadly correspond to largescale
climate maps of Morocco. During the period of investigation
Figure 2. Representation of the different study sites; the black bar denotes 2 m.
Figure 3. Climate in Tafraoute as derived from different internet sources (http://www.weatherbase.com/, http://www.worldweatheronline.com, ttp://
en.climate-data.org).
475EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.
(April 2014), air temperatures (measured with an iButton under an
improvised radiation shield) ranged between 22.5 and 9°C.
Methodology
Fieldwork was carried out in February 2014 (during cool conditions
in late winter). The investigations focused on the assumed
basic parameters influencing tafoni formation: temperature,
moisture and salts. The range of techniques used to investigate
these three parameters and the measurements actually carried
out at the six sites are compiled in Table I.
Infrared thermography
We used a VarioCam HR infrared camera with a thermal resolution
of 0.03 K, an accuracy of ±1.5°C and a spectral range from
7.5 to 14 μm. The detector size is 640×480 pixels and can be
increased to 1280×960 pixels by resolution enhancement
technology. The spatial resolution at 5 m distance is 3 mm.
For the basics of infrared thermography we refer the reader to
Gaussorgues and Chomet (1993); Schuster and Kolobrodov
(2000); Budzier and Gerlach (2010) and others. Applications
of IR thermography to studies of rock walls are described by
Teza et al. (2011). Emissivity is a key variable for thermography
measurements and thus was determined in the field. To determine
emissivity, one iButton measured air temperature and
relative humidity and a second one the temperature of the
rockwall within the sensor field of the IR camera. The air temperature
data was then used to correct the evaluation software
of the thermal images and the thermography data compared
with the measurements from the iButton on the rock wall. The
difference was compensated for by modifying the emissivity factor.
Various measurements showed very similar results with a
calculated emissivity value of 0.95. Similar results have been
published by Gruner (2004); Shannon et al. (2005) and Testo
(2006).
Surface temperatures were measured on three consecutive
days for 24 h using a measurement interval of 10 min across
the entire surface of three sides of the site 1 rock tor (E, W and
S-facing). The aim of using this method was to derive daily average
values of different micro-sites, temperature fluctuations and
the possible impact on moisture transfer processes. The
parameters were modified according to the calculated emissivity
of 0.95.
Moisture meter
A handheld moisture meter (Voltcraft MF-100) was used to
determine the moisture in the near-surface area both within
and adjacent to selected tafoni. The capacitance sensor is commonly
used for detecting water damage in buildings; its electrical
field penetrates some 2–4 cm into the rock. The sensor can
be used in geomorphology to assess semi-quantitative moisture
distributions on reasonably smooth and sound surfaces (Eklund
et al., 2013). Detailed descriptions on the technical background
can be found in James (1963) and Skaar (1988). The
measurement points were arranged in the form of a grid with
a mesh size of 25 cm at the respective sites; the total area of
the grid depended on the micro-topographic conditions at each
site (usually 10 × 10 measurements). The experimental design
was to measure pairs of grids – one inside and one outside each
tafoni – but lack of smooth and round surfaces outside sometimes
precluded this. To reduce uncertainties, each point was
measured three times. The final value was the mean of the three
measurements. The aim was to get an overview of relative
moisture distribution across the near-surface zone.
I-buttons
Miniature loggers (iButton Hygrochron™
DS1923) with 17.35
mm diameter were used to measure temperature and relative
humidity (accuracy 0.5°C and 0.5%). At site 1, two sensors were
attached to the surface inside the tafoni (one in the upper third,
one in the lower third) and a reference sensor was positioned on
the outside rock wall. In addition, the temperature conditions
inside of the rock were determined by Thermochron DS1922L
iButtons inserted into 3, 25, and 50 cm deep bore holes, each
18 mm diameter. Humidity was measured only at 3 cm and
50 cm with the exception of site 1 (east) where moisture was
also determined at 25 cm depth for a period of four days (20/
02/14 to 24/02/14). The Hygrochron buttons measure air
humidity through a small hole at the top of the metallic housing.
For that reason, a second borehole (6 mm in diameter and 5–8
mm longer) was drilled into the lower end of each hole (see
Figure 4). This was to provide a confined air space to measure
borehole humidity, in equilibrium with the surrounding rock
moisture. The principle is based on the sorption isotherm
(Franzen and Mirwald, 2009), the line of equilibrium moisture
between air and rock in porous bodies at a certain temperature
(see also Krus, 1995 and Plagge et al., 2006). The iButtons were
Table I. Summary of measurements performed at the investigation sites.
iButtons
(surface)
iButtons
(temperature profiles,
3 depths)
iButtons
(moisture profiles,
2 depths)
IR Cam
24-hr
monitor.
moisture
meter
grids
ERT
profiles
Sampling
drillholes
site1W 1 1 1 1 2 10* 4
site 1E — 1 1 1 1 1 1
site 1S 1 1 1 1 — 1 —
site 1N — 1 1 — — — —
site 2W — 1 — — 1 1 —
site 2E — — — — — 1 1
site 3 — — — — 2* 3** 2
site 4 — — — — 1 2** —
site 5 — 1 1 — — — —
site 6 — 1 1 — — — —
At site1E measurements were also made at 25 cm for a short period (20/2/2014–24/2/2014).
Bold numbers: measurements presented in this paper.
*Only two profiles displayed;
**Only one profile displayed;
476 H. SCHNEPFLEITNER ET AL.
Copyright © 2015 John Wiley & Sons, Ltd.
glued to a wooden rod to insert them into the boreholes and the
narrow space between rod and rock was sealed at the surface
with silicone paste. In fact, the rods fit so tightly into the
drillholes that one of them could not be recovered and we thus
assume that humidity transport along the rod is negligible. A
similar setup of one Thermochron (25 cm depth) and two
Hygrochrons at 3 cm and 50 cm depth was also installed at
three further sites (additional to the four different orientations
at site 1). Measurements were recorded at intervals of three
min over the period between 14/02/14 and 24/02/14. Data
loss occurred due to vandalism and theft, therefore, no data
exist for some of the sensors from 19/02/14 to 21/02/14. At
the main site 1 (west), 7 days of continuous measurement
are available (14/02–19/02 and 21/02–24/02). Rock moisture
values were calculated from the humidity data based on the
sorption isotherm (Figure 5), which was determined in a
climate chamber.
Electrical resistivity tomography
Electrical resistivity tomography (ERT) is a widely used geophysical
tool for subsurface surveys and has been adapted
for small-scale investigations on rock walls and building stone
facades over the last decade (Sass, 2005; Mol and Viles,
2012). ERT is a geophysical profiling technique which detects
zones of different electrical conductivity, usually along 2Dsections
(for the principles we refer readers to Schrott and
Sass, 2008). In the environment of weathered rock, different
electrical properties are generally caused by heterogeneous
water content and/or salt concentration. However, without
the aid of further techniques, it is currently not possible to differentiate
whether zones of high resistivity are caused by high
water content, soluble salts, or a combination of both. We
used a Geotom device equipped with 50 electrodes. On the
relatively smooth surfaces within the tafoni, self-adhesive
medical ECG-electrodes were used to establish electrical contact
to the rock. Outside of the tafoni the surfaces were generally
rougher and more friable and adhesive electrodes did not
work properly; thus, we used metal screws temporarily
inserted in small boreholes. On the backwall of the large
west-facing tafoni at site 1, two profiles were measured: a vertical
one from the bottom to the roof and a horizontal one
about half way up. Electrode spacing was 6 cm resulting in
total profile lengths of 2.94 m. Both profiles were measured
every 6 h over a 24 h period (19/2/14–20/2/14). ERT profiles
of similar configuration were measured at most of the other
sites as well, but solely single measurements without 24 h
monitoring.
2D-resistivity sections were calculated using the software
package Res2Dinv (Geotomo Software, Penang, Malaysia).
We tested the standard least-square inversion as well as robust
inversion and compared both in order to verify the shape of
detected anomalies. The measurement setups used were
Wenner and dipole–dipole (Kearey et al., 2002) which allow
good resolution of surface-parallel layers (Wenner) and of lateral
inhomogeneities (dipole–dipole), respectively. Both arrays
were combined in Res2Dinv to benefit from the advantages of
both measurements. We checked the results of the calculation
with the depth of investigation (DOI) method (Oldenburg and
Li, 1999; Marescot et al., 2003). The DOI is an empirical
method for determining the depth down to which the results
are considered trustworthy, and model regions with DOI index
values >0.2 are usually considered unreliable in terms of their
modelled absolute resistivity values (Hilbich et al., 2009;
Angelopoulos et al., 2013).
Drilling and salt sampling
Sampling for salts was carried out by incrementally collecting
the ample drilling dust from the installation of the iButtons and
from further drillholes. This approach made it possible to sample
at a range of depths and thus, to assess the concentration
and the types of salts at different depths. At site 1, four drill holes
were sampled; three inside the west-facing tafone (top, mid,
bottom) and one from the open rock face beside the tafone.
The samples were taken from the following six depth
Figure 5. Sorption isotherm of the Tafraoute granite as derived from lab measurements.
Figure 4. Setup for temperature and relative humidity measurements
at different depths. Hatching: wooden rod, gray: iButton.
477EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.
increments: 0–2.5 / 2.5–5 / 5–10 / 10–15 / 15–20 / 20–25 cm.
Analyses were performed in the Oxford Rock Breakdown Laboratory
with the Ion Chromatography System Dionex IC DX500.
In addition to ion chromatography, a simple conductivity measurement
was performed for each sample (1 g of drill dust
dissolved in 5 mL of deionized water). The aim of this was to
provide a quick assessment of overall salt concentration. The
applied instrument was a bench-top conductivity meter (HI2315).
Results
Presentation of the results focuses on the west-facing tafone at
site 1 (shown in Figure 2) because of the completeness of the
datasets collected there. All of the following results refer to
the inside of this tafone if not specified otherwise.
Infrared thermography
Data were collected every 10 min over an extended period
(24 h), each measurement covering several m2
with a spatial
resolution of 3 mm. In Figure 6(a), the result of the averaging
of all measurements is shown. It is obvious that the inside of
the tafone is warmer on average than the surrounding rock
surface. Furthermore, a significant temperature difference
occurs between the bottom and the top of the tafone, whereas
the surrounding rock surface shows more homogenous conditions.
Figure 6(b) shows a horizontal and a vertical profile
(along the lines indicated in Figure 6(a)) to illustrate this point.
While there is almost no temperature difference in the
horizontal-profile (blue line), a significant temperature
decrease towards the bottom can be detected in the vertical
profile (red line). The temperature difference is almost 3°C.
In Figure 6(c) the diurnal temperature cycles at defined positions
(as marked in Figure 6(a)) are displayed. The selected
positions can be divided into three areas: area one includes
positions 1 to 8 (outside the tafone, green lines), area two
includes positions 9 to 15 (upper inside area of the tafone,
red lines) and area three includes positions 16 to 24 (lower
inside area of the tafone, blue lines). The respective mean
values for each area were calculated (Figure 6(d)). Area one
(outside the tafone) shows the biggest amplitude and therefore
the lowest (~3°C) and the highest (~20°C) values during the
day. In the curves of areas two and three the maximum and
minimum temperatures are subdued. Interestingly, the two
curves run almost parallel to each other with the lower tafone
being constantly cooler than the upper tafone. This effect cannot
be caused by stronger emission at night (otherwise, the
amplitudes would be higher); the most probable explanation
is higher moisture content towards the base (see discussion).
On the east side of site 1 the tafoni are much shallower and further
away from the ground; accordingly, similar patterns were
not recorded at this site (not displayed).
Moisture meter
As pointed out in the methodology section, this method can
only be used to infer relative moisture distributions within the
near-surface zone (0–4 cm) as the reliability for quantitative
results is rather low (Eklund et al., 2013). At the east-facing side
of site 1, it was found that small, decimeter-scale tafoni are consistently
wetter than the unweathered surrounding rock
(Figure 7(c)). On the backwall of the large west-facing tafone
Figure 6. Results of the infrared measurments. (a) Average of the 24 h measurement (dots mark the positions for which the temperature fluctuations
were analysed); (b) horizontal and vertical temperature profile (24 h mean); (c) daily temperature cycle at defined points; (d) averaged temperature
cycles for certain areas (see text). This figure is available in colour online at wileyonlinelibrary.com/journal/espl
478 H. SCHNEPFLEITNER ET AL.
Copyright © 2015 John Wiley & Sons, Ltd.
at site 1, a distinct trend was found with a clear decrease in the
relative moisture values towards the top (Figure 7(a)). Thus, the
lower area of the tafone seems to be moister than the upper part,
which corresponds to the findings of the infrared monitoring.
Similar patterns were found at site 3 (Figure 7(d)) even if the
decrease of moisture with elevation is smaller at this site. Outside
tafoni the rock was generally much drier and no differences
in vertical direction were found (Figure 7(b)).
I-buttons
Similar to the IR results, the iButtons show a reasonably constant
temperature difference between top and bottom of the
tafone, and the largest amplitudes outside of the tafone (site 1,
west-facing). The humidity data show a reverse diurnal variation
with consistently higher relative humidity values at the bottom
in comparison with the top (Figure 8(c)). The reason for this
is the difference in temperature, as cooler air can absorb less
moisture. Thus, to understand the process of water vapor diffusion
relative humidity data are of limited benefit. The more significant
water vapour pressure was computed from the
temperature and relative humidity according to the Magnus formula
(Defant and Defant, 1958). In terms of absolute vapour
pressure it becomes clear that there are no humidity differences
close to the surface inside (top or bottom) or outside the tafone
(Figure 8(a)).
A pronounced diurnal cycle of temperature was detected only
at depths of 3 and 20 cm while variations at 50 cm were just about
0.5°C per day (Figure 8(e)). Similarly, there is a clear diurnal cycle
in both relative humidity and absolute vapour pressure at 3 cm
depth while both parameters are constantly high at 50 cm depth.
Special attention was paid to the increase of vapour pressure
with depth because vapour pressure differences trigger vapour
diffusion. Between 50 and 3 cm, and from 3 cm to the surface,
a vapour pressure gradient from the inner parts towards the surface
was detected over the whole measurement period. Both
curves depend on the time of the day: in the afternoon, the
highest vapor deficit is located between 3 cm and the surface
while in the morning hours, the maximum is between 50 cm
and 3 cm, but there is no reversal at any time. This finding confirms
that moisture is constantly transported from the interior to
the surface, even if the transport in the gaseous phase is not
able to move salts through the rock.
Relative humidity values from the iButtons can be converted
into degree of (liquid) pore saturation via the laboratorymeasured
sorption isotherm (Figure 5). The average pore saturation
in our measurement period was ~80% at 50 cm depth
and ~10% at 3 cm.
Electrical resistivity tomography
Figure 9 displays the vertical and horizontal ERT profiles at site
1, site 3 and site 4 (Wenner and dipole–dipole measurements
were combined in all cases). All extend to a depth of 30 cm
to 40 cm inside the rock. The limits of resistivity for granite vary,
depending on the textbook literature, between 103
and 105
ohm m (lower values for wet and higher values for dry rock).
A similar range of values is found in all displayed profiles; the
rather high maximum resistivity values can probably be attributed
to the unusually high porosity (8%) of the Tafraoute
granite. Because the structure of the rock appears to be
Figure 7. Results of the moisture meter measurements. (a) site 1
(west), inside the tafone; (b) site 1 (west), outside of the tafone; (c) site
1 (east), yellow-green patches correspond to small, shallow tafone; (d)
inside the tafone at site 3. Relative readings from 0 (dry) to 100 (wet);
the black bar denotes 2 m. This figure is available in colour online at
wileyonlinelibrary.com/journal/espl
Figure 8. Results of iButtons measurement at site1 (west). This figure is available in colour online at wileyonlinelibrary.com/journal/espl
479EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.
homogeneous and fractures are absent, it is likely that the
measurements represent areas of different moisture and/or salt
contents. Two general patterns can be seen in the vertical profiles
(Figure 9(a), (b)). (1) Lower resistivity in the lower half of
the profiles. For example, in the vertical profile of site 1, high
values (>105
Ohmm) are found in the upper area and relatively
low values (<103
ohm m) near the bottom (Figure 9(a)). As
these resistivity patterns correspond well with the infrared and
the moisure meter results, we assume that these low-resistivity
zones represent wetter areas, possibly amplified by salts near
the surface. (2) Higher resistivity deeper inside the rock, particularly
in the upper half of the profiles. This contradicts the very
clear results of the iButtons which indicate wetter conditions at
depth. Thus, it is probable that the high conductivity/low resistivity
recorded near the surface using ERT results from higher
salt concentration not higher moisture levels.
Although the contrasts are smaller in the horizontal profile
from site 1 (Figure 9(c)), it is evident that the edge of the tafone
is slightly drier than the middle, and that there is a slight
increase in moisture towards the inside of the rock. Similar
patterns were also found in the horizontal profile at site 4
(Figure 9(d)) even if the resistivities are much lower. All measurements
have a rather low RMS (root mean square; for the
exact values see Figure 9) error and the DOI values (not
displayed) do not exceed the threshold of 0.2 which indicates
that the results reflect the real situation very well.
Drilling and sampling
Figure 10 displays the salt concentrations found in the samples
of the four drill holes at site 1. Large quantities of sodium and
sulfate are apparent particularly near the surface, along with
other ions. At all measurement sites, except for the outer drill
hole, salt concentrations increase towards the surface. Certain
general patterns can be detected: first, a greater amount of salts
are found close to the surface at all three sites within the tafone;
second, concentrations of sulfate and sodium (to a lesser degree
of chloride, bromide and magnesium) increase from the top
towards the bottom of the tafone; and third, the increase of salt
concentrations near the surface is not found in the outer drill
hole (Figure 10(a)–(d)). These general findings are confirmed
by the results of the drill dust conductivity measurements where
maximum values are located at the surface and in the
lower/moister area of the tafone (Figure 11). Similar patterns
were also found within the tafoni borehole at site 4, but in this
Figure 9. Results of ERT measurement at site 1 west (a), (c), site 3 (b) and site 4 (d). Each profile is combined from Wenner and dipole–dipole measurements;
unit electrode spacing is 6 cm. This figure is available in colour online at wileyonlinelibrary.com/journal/espl
Figure 10. Concentrations of major ions in rock powder from different depths collected from the drillholes of site 1 (west). This figure is available in
colour online at wileyonlinelibrary.com/journal/espl
480 H. SCHNEPFLEITNER ET AL.
Copyright © 2015 John Wiley & Sons, Ltd.
case the drill profile outside of the tafoni on the top of the boulder
(which was vertical and not horizontal) is noticeably different
as there is a significant increase of the values with increasing
depth.
Because of the very similar distribution of sodium and sulfate
it can be assumed with high probability that the dominant salt
is sodium sulfate (as thenardite, the heptahydrite form or
mirabilite; see Figure 13). According to Goudie (1977) and
Goudie and Viles (1997), this salt is one of the most important
agents of salt weathering especially in arid environments
because of its high solubility in water, its ability to hydrate at
humidities and temperatures frequently encountered in nature,
and its long crystal habit.
Discussion
Utility of the techniques applied
The applied multi-method approach, combining relatively
simple methods (drill dust sampling, hand held moisture
meters) with novel and more sophisticated techniques (borehole
humidity and temperature, ERT, thermography) has been
demonstrated to provide useful data with which to gain insight
into temperature, moisture and salt distribution. The data
complement and support each other. Based on the measured
data the following results can be summarized:
○ Temperature amplitudes are subdued within tafoni
compared with those on the surrounding rock faces (derived
from: infrared, iButtons).
○ Areas of elevated moisture content are found in the
lower areas of the tafoni (ERT, infrared, iButtons, moisture
meter); the rock inside the tafoni is moister than that
outside (moisture meter). The humidity/absolute water
vapour pressure increases with increasing rock depth
(iButtons).
○ Inside, but not outside, tafoni there is an accumulation
of salts close to the rock surface (drill dust sampling).
Salt concentrations increase towards the base inside
tafoni (drill dust sampling).
Drill dust sampling turned out to be particularly helpful to
clarify the salt concentration along depth profiles. This simple
technique has rarely been carried out before (Viles and Goudie,
1992). Due to the high amount of material collected for
analysis at each depth interval (drill diameter 18 mm) the results
are likely to be highly representative. In many studies,
only small amounts of weathered material from the surfaces
within tafoni were collected (Mustoe, 1983; McBride and
Picard, 2000). The drilling results were well worth the logistic
effort because they give very clear evidence of the vertical
and depth distribution of salts. Of course this method is not
an option in many protected areas or conservation sites. The
information can be used to cross-check the ERT results: in the
current example it shows that high salt contents are a factor
affecting conductivity/resistivity values only close to the surface
(where counteracted by insufficient moisture) while resistivity
patterns in the deeper subsurface must be caused by varying
pore water content alone, as the salt levels are consistently low.
The measurements of borehole humidity have not been carried
out before to our knowledge in comparable settings and
they also delivered novel results, proving clearly that rock moisture
(at the time of our measurements) increases with depth. As
with drill dust sampling, the method depends on the availability
of a battery-driven hammer drill and is likely to be more problematic
on harder, less weathered granites. A longer measurement
programme (months or years rather than days) would
have been desirable and is planned for the future; it was not possible
in the current field campaign for logistical reasons and because
of vandalism. The cheap and easy-to-use hand held
moisture meters were helpful to underpin the results of the ERT
and IR measurements from which we infer spatial moisture distributions.
However, the scatter of the measurements is very
high and they are sometimes hardly reproducible; accordingly,
anything more than a very rough, semi-quantitative orientation
is not possible.
Finally, IR thermography (24 h monitoring) proved to be very
helpful to gain high-resolution information on spatial and temporal
patterning of surface temperatures and to delimit moister
and drier areas. The consistent temperature difference between
the upper and lower area within the west-facing tafone at site 1
clearly points to evaporative cooling and thus, to higher
amounts of moisture towards the base. This process seems to
be so strong that even the influence of direct radiation (westfacing,
approx. from 16:00 to 17:30) is suppressed. An effect
of higher emission at night can be ruled out because of the constant
temperature difference throughout the 24 h period. As
these assumptions are clearly confirmed by the results of other
methods, IR thermography was shown to be a valid method to
gain data on temperature and moisture distribution. However,
short measurement times (or single shot images) are of very
limited value.
Figure 11. Conductivity values from rock powder from different depths collected from the drillholes in site 1 and site 3. This figure is available in
colour online at wileyonlinelibrary.com/journal/espl
481EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.
Origin and distribution of salts
The results of our measurements show that moister areas near
the surface also have a higher salt content, a correlation found
in many publications on tafoni (Bradley et al., 1978). The reason
is that moisture is the agent of transport for dissolved salts.
Possible sources of water that merit consideration include dew,
fog, rain and groundwater, and two processes are responsible
for the movement of moisture in the rock: capillary transport
(which is also the main process of salt transport), and vapor diffusion
which is responsible for the dehydration of rock. As dehydration
is, in turn, responsible for the crystallization of salts
the process of vapor diffusion has a great influence on the position
of salts in the rock.
Huinink et al. (2004) tried to simulate water and salt movement
within a developing tafone in a mathematical model.
They regarded two different cases: a short drying period and a
long one (t=2000Δt) and their results show that a long drying
period is very important for the development of tafoni because
there is ample time for salts to migrate to the slow-drying areas,
following gradients of concentration. This mechanism is also
held responsible for the tendency of tafoni to grow upward
because ‘most salt accumulates at the more sheltered (from
the sun) part of the rock surface, having the lowest drying rates,
given that the drying period in a wetting/drying cycle is long
enough. Therefore, these parts weather faster than the exterior
part and cavernous structures, e.g. Tafoni, develop.’ (Huinink
et al., 2004, p.1231). This mechanism explains the tendency
of tafoni to grow upward. From our own data (e.g. moisture meter)
we are able to confirm that the interiors of tafoni dry more
slowly than external rock faces and that salts are concentrated
in these regions of slow drying (see salt sampling results). In
contrast to Huinink et al. (2004), in our study site this area is
located at the base and not in the upper parts of the tafoni. This
might be due to the specific topographic situation at Tafraoute.
In this case, we assume that under the rock tor at our main
study site a lens of high soil moisture develops (Figure 12), a situation
which may be different under isolated blocks or in
higher positions of a rock wall.
The model outlined in Figure 12 is based on our findings
from Tafraoute. One of the core questions is do water and
salts, respectively, derive from the outside (fog, rain, dew,
aeolian salt deposition) or from the inside (groundwater body)
of the rock. We assume two possible pathways of salt enrichment.
(1) Water (with salts) infiltrates from the top of the tor
(red arrow) and from the relatively flat surrounding (blue
arrow) and forms a body of +/À saturated rock under and
within the tor. Due to evaporation and vapor pressure gradients,
capillary water is sucked to the surface and salts are precipitated
near the surface. The main area of salt enrichment at
site 1 is the basal area of the tafone due to the proximity to the
saturated zone (primary path). (2) Salts may be also deposited
with aeolian dust. On smooth rock faces without tafoni, large
portions of these salts are probably washed off, while inside
the tafoni they remain at the rock surface, or are dissolved during
dewfall events and infiltrate into the outermost centimeters
of rock (secondary path). Both outlined mechanisms lead to
salt concentration in sheltered places and near the base of
the rock. This model contradicts the work of Bradley et al.
(1978) and the model of Huinink et al. (2004) both of which
hypothesized that salts (as well as moisture) should be primarily
found in the upper, protected areas of tafoni while the
lower parts of the tafoni are probably more exposed to wind,
rain and radiation. A reason for the unexpectedly high salt
and moisture content in the lower parts of the investigated
tafoni at Tafraoute could be that these areas receive a higher
input of atmospheric salts. However, if this was the case a
markedly higher amount of salt would have to be expected
outside of the tafoni which is not the case (with the notable
exception of the roof of tafone 3, see below). Thus, for the
inside of the tafoni it is more likely that the salts originate from
a water body (or area of higher humidity) inside the rock from
where they are transported by capillary transport towards the
surface. This would mean that tafoni in comparable settings
mainly develop near the base of boulders and tors and up to
the assumed level of capillary rise which is, in fact, often the
case.
The salt concentration at the top of the boulder at site 3
shows a significant increase in salt concentration with depth
(Figure 11, dashed line) and thus differs from the profiles inside
the tafoni at sites 1 and 3. A reason for this could be the combination
of rain and rapid dehydration caused by the high
direct radiation. Due to fast drying after rain, atmospheric salts
crystallize at a certain depth. During subsequent rain events the
salts are successively transported to greater depths (Figure 12,
red arrow). This pathway of salt displacement would explain
the development of Tafoni in higher areas of tors and their
growth upwards. This assumption is, however, supported only
by one measurement and is therefore speculative.
Implications for salt weathering
Textbook concepts of salt weathering assume that damage is
induced by: (i) crystallization pressure (Correns, 1949; Scherer,
1999); (ii) stresses induced by crystal hydration (Mortensen,
1933; Winkler and Wilhelm, 1970); and (iii) a differential thermal
expansion of salts (Cooke and Smalley, 1968; Charola,
2000). The borehole humidity data derived from the iButtons
can be used to underpin these assumptions. With high probability,
the dominant salt at our study sites is sodium sulfate which
occurs in nature in two varieties: the water-free thenardite and
the hydrated variety mirabilite. A third metastable variety which
might crystallise on cooling to around 10°C was reported on by
Figure 12. (a) Salt concentrations and moisture gradients according to the measurements; (b) supposed distribution of saturated rock; (c) supposed
migration paths of salts. This figure is available in colour online at wileyonlinelibrary.com/journal/espl
482 H. SCHNEPFLEITNER ET AL.
Copyright © 2015 John Wiley & Sons, Ltd.
Hall and Hamilton (2008). Assuming that crystallization
pressure is the primary source of damage, Rodriguez-Navarro
et al. (2000) argued that precipitation of thenardite rather than
mirabilite may be responsible for damage from sodium sulfate,
particularly in situations of constant capillary rise.
As Figure 13 shows, the results of our micro-climatic measurements
suggest that due to higher humidities found in the
pores, sodium sulfate at a depth of 50 cm should occur in
dissolved form (or supersaturated with respect to mirabilite).
Just below the surface (3 cm), however, the anhydritic
thenardite should prevail. Due to the short measurement
period the conclusions are speculative; yet it can be assumed
that the mirabilite–thenardite transition frequently occurs at
this depth range, either by the capillary transport towards
the surface and crystallization, or because of changing
weather conditions when the surface is wetted or the drying
front propagates deeper into the rock in summer. These
patterns were also found at the east side of site 1 and at site
5 and 6 (not displayed).
Conclusions
To sum up, the following conclusions can be drawn from the
measurements presented here:
• The combination of high-resolution temperature, moisture
and salt measurements enables valuable insights into tafoni
formation at our study site in Tafraoute, Morocco. The results
derived from very different techniques (IR, T and H sensors,
ERT, hand held moisture meters, drill dust sampling)
mutually support one another.
• Our results confirm that moisture and salt concentration are
elevated inside tafoni compared with the surrounding rock.
• A clear correlation was found between moisture and salt
contents. Within a tafone, areas of higher humidity also display
increased salt concentration near the surface. ERT
reacts to both parameters; the least resistive areas are those
with high levels of both moisture and salts.
• Our measurements clearly show a significant accumulation
of salts close to the rock surface in tafoni, but not on the
surrounding rock surfaces. These findings appear to be
consistent with the model of Huinink et al. (2004) which predicts
enhanced salt deposition in zones of low drying rates.
• Salts and moisture at our sites were concentrated near the
base of tafoni which has, to our knowledge, not been reported
in the literature before. We assume that the reason
is a saturated pore water body around the base of rock tors
which might be common in most places other than highly
free-standing boulders.
• The measured T/H-values indicate that salts (in this case sodium
sulfate) occur in various phases which implies frequent
phase changes from thenardite to mirabilite and vice versa.
• Two pathways of salt transport in and around tafoni are assumed
based on the data: infiltration with rainfall on the
top and around tors and boulders, and capillary rise from
saturated pore water bodies to the surface. Direct dry deposition
is probably of subordinate importance.
References
Angelopoulos MC, Pollard WH, Couture NJ. 2013. The application of
CCR and GPR to characterize ground ice conditions at Parsons Lake,
Northwest Territories. Cold Regions Science and Technology 85:
22–33.
Barbey P, Oberli F, Burg J-P, Nachit H, Pons J, Meier M. 2004. The
Palaeoproterozoic in western Anti-Atlas (Morocco): a clarification.
Journal of African Earth Sciences 39: 239–245.
Benziane F. 2007. Lithostratigraphie et évolution géodynamique de
l’anti-Atlas (Maroc) du paléoprotérozoïque au néoprotérozoïque:
exemples de la boutonnière de Tagragra Tata et du Jebel Saghro.
Doctoral thesis, Université Savoie, France.
Bradley WC, Hutton JT, Twidale CR. 1978. Role of salts in development
of granitic tafoni, South Australia. The Journal of Geology 86(5):
647–654.
Brandmeier M, Kuhlemann J, Krumrei I, Kappler A, Kubik PW. 2011.
New challenges for tafoni research: a new approach to understand
processes and weathering rates. Earth Surface Processes and Landforms
36: 839–852.
Boelhouwers J, Jonsson M. 2013. Critical Assessment of the 2°C min
À1
threshold for thermal stress weathering. Geografiska Annaler: Series
A, Physical Geography 95(4): 285–293.
Budzier H, Gerlach G. 2010. Thermische Infrarot Sensoren. Weinheim:
Wiley-VCH.
Campbell SW. 1999. Chemical weathering associated with tafoni at
Papago Park, Central Arizona. Earth Surface Processes and Landforms
24: 271–278.
Charola AE. 2000. Salts in the deterioration of porous materials: an
overview. Journal of the American Institute for Conservation 39:
327–343.
Cooke RU, Smalley IJ. 1968. Salt Weathering in Deserts. Nature 220:
1226–1227.
Correns CW. 1949. Einführung in die Mineralogie (Kristallographie und
Petrologie). Springer-Verlag: Berlin-Göttingen-Heidelberg 414 S., 405
Abb., 1 Tafel. Geh. DM 38.—, geb. DM 41.60.
Defant A, Defant F. 1958. Physikalische Dynamik der Atmosphäre.
Frankfurt: Akademische Verlagsgesellschaft.
Dionisio A, Martinho E, Grangeia C, Almedia F. 2013. Examples of the
use of non-invasive techniques for the evaluation of stone decay in
Portugal. Key Engineering Materials 548: 239–246.
Doehne E. 2002. Salt weathering: a selective review. In natural stone,
weathering phenomena, conservation strategies and case studies.
Geological Society Special Publications 205: 51–64.
Figure 13. Phase diagram for sodium sulfate. The continuous lines
indicate the limits between the stable phases. The discontinuous line
corresponds to a solution in metastabile eqilibrium with respect to
thenardite and supersaturated with respect to mirabilite (Tsui et al.,
2003). The grey rectangles (site1 west) indicate the measured T/H data
at 50 and 3 cm depth from 21/2/14 to 23/2/14; the rectangles (site 1
east) indicate the measured T/H data at 50, 20 and 3 cm depth from
14/2/14 to 24/2/14, respectively.
483EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.
Dorn RI, Gordon SJ, Krinsley D, Langworthy K. 2013. Nanoscale:
Mineral Weathering Boundary. In Treatise on Geomorphology.
Academic: Oxford.
Eklund JA, Zhang H, Viles HA, Curteis T. 2013. Using handheld moisture
meters on limestone: factors affecting performance and guidelines
for best practice. International Journal of Architectural
Heritage 7: 207–224.
Franzen C, Mirwald PW. 2009. Moisture sorption behaviour of salt mixtures
in porous stone. Geochemistry 69: 91–98.
French HM, Guglielmin M. 2000. Cryogenic weathering of granite,
northern Victoria land, Antarctica. Permafrost and Periglacial Processes
11: 305–314.
Goudie AS. 1977. Sodium sulphate weathering and the disintegration
of Mohenjo-Daro, Pakistan. Earth Surface Processes and Landforms
2(1): 75–86.
Goudie AS, Viles HA. 1997. Salt Weathering Hazards. Wiley: Chichester/
New York.
Gaussorgues G, Chomet S. 1993. Infrared Thermography. Springer
Verlag. ISBN:0412479001.
Gruner K. 2004. Grundlagen der berührungslosen Temperaturmessung
„Strahlungsthermometrie“. Revision B1. 12/2004. Raytek Europazentrale:
RaytekGmbH, Berlin.
Hall C, Hamilton A. 2008. The heptahydrate of sodium sulfate: does it
have a role in terrestrial and planetary geochemistry? Icarus 198:
277–279.
Hall K. 1986. Rock moisture content in the field and the laboratory and
its relationship to mechanical weathering studies. Earth Surface Processes
and Landforms 11: 131–142.
Hall K. 1991. Rock moisture data from the Juneau Icefield (Alaska) and
its significance for mechanical weathering studies. Permafrost and
Periglacial Processes 2: 321–330.
Hall K, André MF. 2001. New insights into rock weathering from highfrequency
rock temperature data: an Antarctic study of weathering
by thermal stress. Geomorphology 41: 23–35.
Hall K, Meiklejohn I, Arocena J. 2007. The thermal responses of rock
art pigments: implications for rock art weathering in southern Africa.
Geomorphology 91: 132–145.
Hassenforder B. 1985. Les mylonites de la zone de faille ductile panafricaine
des Ameln (Kerdous, Anti-Atlas occidental, Maroc). Une
analyse pétrostructurale de la déformation. Sciences géologique
Strasbourg 38: 215–226.
Hassenforder B. 1987. La tectonique panafricaine et varisque de l’AntiAtlas
dans le massif du Kerdous (Maroc). Doctoral Thesis, University
of Strasbourg.
Hilbich C, Marescot L, Hauck C, Loke MH, Mäusbacher R. 2009.
Applicability of electrical resistivity tomography monitoring to coarse
blocky and ice-rich permafrost landforms. Permafrost and Periglacial
Processes 20(3): 269–284.
Huinink HP, Pel L, Kopinga K. 2004. Simulating the growth of Tafoni.
Earth Surface Processes and Landforms 29: 1225–1233.
James WL. 1963. Electric moisture meters for wood. US Forest Service:
Madison, WI.
Kearey P, Brooks M, Hill I. 2002. An Introduction to Geophysical Exploration.
Blackwell Science: Malden.
Krus M. 1995. Feuchtetransport- und Speicherkoeffizienten poröser
mineralischer Baustoffe. Theoretische Grundlagen und neue
Messtechniken. Doctoral thesis, Universität Stuttgart.
Marescot L, Loke MH, Chapellier D, Delaloye R, Lambiel C, Reynard E.
2003. Assessing reliability of 2D resistivity imaging in mountain
permafrost studies using the depth of investigation index method.
Near Surface Geophysics 1(2): 57–67.
Martinho E, Alegria F, Dionisio A, Grangeia C, Almeida F. 2012.
3D-resistivity imaging and distribution of water soluble salts in
Portuguese Renaissance stone bas-reliefs. Engineering Geology
141–142: 33–44.
Materialarchiv. 2014. http://www.materialarchiv.ch/detail/43#/detail/
43/granit (accessed 05/03/2015)
Matsukura Y, Takahashi K. 2000. A new technique for rapid and nondestructive
measurement of rock-surface moisture content; preliminary
application to weathering studies of sandstone blocks. Engineering
Geology 55: 113–120.
Maurício A, Figueiredo C, Aires-Barros L. 2006. Monitoring salt systems
in monument stones: comparing two electrochemical methods. In
Proceedings of the International Conference on Heritage, Weathering
and Conservation (HWC-2006). Taylor & Francis: London.
McBride EF, Picard MD. 2000. Origin and development of tafoni in
Tunnel Spring Tuff, Crystal Peak, Utah, USA. Earth Surface Processes
and Landforms 25(8): 869–879.
Mellor A, Short J, Kirkby SJ. 1997. Tafoni in the El Chorro area,
Andalucia, southern Spain. Earth Surface Processes and Landforms
22: 817–833.
Mol L, Viles HA. 2010. Geoelectric investigations into sandstone moisture
regimes: Implications for rock weathering and the deterioration
of San Rock Art in the Golden Gate Reserve, South Africa. Geomorphology
118: 280–287.
Mol L, Viles HA. 2012. The role of rock surface hardness and internal
moisture in tafoni development in sandstone. Earth Surface Processes
and Landforms 37: 301–314.
Mortensen H. 1933. Die ‘Salzsprengung’ und ihre Bedeutung für die
regional klimatische Gliederung der Wüsten. Petermanns Geographische
Mitteilungen 79: 130–135.
Mottershead DN, Pye K. 1994. Tafoni on Coastal Slopes, South Devon,
UK. Earth Surfaces Processes and Landforms 19: 543–563.
Mustoe GE. 1983. Cavernous weathering in the Capitol Reef Desert,
Utah. Earth Surface Processes and Landforms 8: 517–526.
Oldenburg DW, Li Y. 1999. Estimating depth of investigation in DC
resistivity and IP surveys. Geophysics 64(2): 403–416.
Plagge R, Funk M, Scheffler G, Grunewald J. 2006. Experimentelle
Bestimmung der hygrischen Sorptionsisotherme und des Feuchtetransportes
unter instationären Bedingungen. Bauphysik 28(2): 81–87.
Rodriguez-Navarro C, Doehne E, Sebastian E. 1999. Origins of honeycomb
weathering: the role of salts and wind. Geological Society of
America Bulletin 111: 1250–1255.
Rodriguez-Navarro C, Doehne E, Sebastian E. 2000. How does sodium
sulfate crystallize? Implications for the decay and testing of building
materials. Cement and Concrete Research 30(10): 1527–1534.
Sass O. 2003. Moisture distribution in rockwalls derived from 2Dresistivity
measurements. Geophysical applications in geomorphology,
Zeitschrift für Geomorphologie Suppl. 132: 51–69.
Sass O. 2005. Rock moisture measurements: techniques, results, and
implications for weathering. Earth Surface Processes and Landforms
30: 359–374.
Sass O, Viles HA. 2010a. Wetting and drying of masonry walls: 2Dresistivity
monitoring of driving rain experiments on historic stonework
in Oxford, UK. Journal of Applied Geophysics 70: 72–83.
Sass O, Viles HA. 2010b. 2D resistivity surveys of the moisture contents
of historic limestone walls in Oxford, UK: implications for understanding
catastrophic stone deterioration. In Limestone in the Built
Environment: Present Day Challenges for Preservation of the Past,
Smith BJ, Gomez Heras M, Viles HA, Cassar J (eds). Geological
Society of London, Special Publication 331: 237–249.
Scherer G. 1999. Crystallization in pores: Princeton University,
CEE/PMI, Eng. Quad. E-319, Princeton, NJ 08544, USA. Cement
and Concrete Research 29: 1347–1358.
Schrott L, Sass O. 2008. Application of field geophysics in geomorphology:
Advances and limitations exemplified by case studies. Geomorphology
93: 55–73.
Schuster N, Kolobrodov VG. 2000. Infrarotthermographie. Weinheim:
WILEY-VCH.
Service Géologique du Maroc. 1985. Carte géologique du Maroc.
Échelle: 1/1000000, Rabat.
Shannon HR, Sigda JM, Van Dam RL, Hendrickx JMH, McLemore VT.
2005. Thermal Camera Imaging of Rock Piles at the Questa Molybdenum
Mine, Questa, New Mexico. In Proceedings of a Joint Conference
of American Society of Mining and Reclamation (ASMR) 22nd
Annual National Conference.
Skaar C. 1988. Wood–Water Relations. Springer-Verlag: Berlin.
Soulaimani A, Piqué A. 2004. The Tasrirt structure (Kerdous inlier,
Western Anti-Atlas, Morocco): a late Pan-African transtensive dome.
Journal of African Earth Sciences 39(3): 247–255.
Strini A, Guglielmin M, Hall K. 2008. Tafoni development in a cryotic
environment: an example from Northern Victoria Land, Antarctica.
Earth Surface Processes and Landforms 33: 1502–1519.
Sumner PD, Hall KJ, van Rooy JL, Meiklejohn KI. 2009. Rock
weathering on the eastern mountains of Africa: Review and insight
from case studies. Journal of African Earth Sciences 55: 236–244.
484
Copyright © 2015 John Wiley & Sons, Ltd.
Testo AG (ed). 2006. Leitfaden zur Infrarot-Messtechnik. http://www.
testo.com (accessed 06/2011).
Teza G, Marcato G, Castelli E, Galgaro A. 2011. IRTROCK: a MATLAB
toolbox for contactless recognition of surface and shallow weakness
of a rock cliff by infrared thermography. Computer & Geosciences
45: 109–118.
Tsui N, Flatt RJ, Scherer GW. 2003. Crystallization damage by sodium
sulphate. Journal of Cultural Heritage 4: 109–115.
Turkington AV, Martin E, Viles HA, Smith BJ. 2003. Surface change and
decay of sandstone samples exposed to a polluted urban atmosphere
over a six-year period: Belfast, Northern Ireland. Building and Environment
38: 1205–1216.
Turkington AV, Paradise TR. 2005. Sandstone weathering: a century of
research and innovation. Geomorphology 67(1): 229–253.
Turkington AV, Phillips JD. 2004. Cavernous weathering, dynamical
instability and self-organization. Earth Surface Processes and Landforms
29: 665–675.
Viles HA, Goudie AS. 1992. Weathering of limestone columns from
the Weymouth seafront, England. In Proceedings of the 7th International
Congress on Deterioration and Conservation of Stone, Lisbon,
June 1992.
Winkler EM, Wilhelm EJ. 1970. Salt burst by hydration pressures in
architectural stone in urban atmosphere. Geological Society of
America Bulletin 81(2): 567–572.
485EARTH SURFACE PROCESSES AND LANDFORMS
Copyright © 2015 John Wiley & Sons, Ltd.