Past hydrological events reflected in the Holocene fluvial record of Europe
M.G. Macklin a,*, G. Benito b
, K.J. Gregory c
, E. Johnstone a
, J. Lewin a
, D.J. Michczyn´ska d
,
R. Soja e
, L. Starkel e
, V.R. Thorndycraft b
a
River Basin Dynamics and Hydrology Research Group, Institute of Geography and Earth Sciences, University of Wales, Aberystwyth, Ceredigion, SY23 3DB,
UK
b
Consejo Superior de Investigaciones Cientı´ficas, Centro de Ciencias Medioambientales, C/Serrano, 115-bis, 28006 Madrid, Spain
c
Department of Geography, University of Southampton, Southampton, SO17 1BJ, UK
d
Radiocarbon Laboratory, Institute of Physics, Silesian University of Technology, Krzywoustego 2, 44-100 Gliwice, Poland
e
Institute of Geography and Spatial Organisation, Polish Academy of Sciences, Krako´w, Poland
Abstract
A comprehensive database of radiocarbon dated fluvial units in Great Britain, Poland and Spain has been compiled to investigate the
relationship between environmental change, flooding and Holocene river dynamics. Following the methodology recently developed by
Macklin and Lewin [Macklin, M.G., Lewin, J., 2003. River sediments, great floods and centennial-scale Holocene climate change. Journal of
Quaternary Science 18, 101–105], radiocarbon dates in fluvial sequences that coincide with a modification in sedimentation rate, or style,
have been highlighted, allowing geomorphologically significant changes in Holocene river activity to be identified. Data analysis has been
undertaken at both national and sub-national scales, and on catchments of different size, type and land-use history. Multiple phases of higher
flood frequency, characterized by accelerated erosion and sediment deposition on floodplains, are recognized and compared with a range of
climate proxies. The relative and varying roles of climate and land-use on river dynamics are considered and the value of the database for
reconstructing past hydrological events, as well as for predicting river response to future environmental change, is assessed.
D 2005 Elsevier B.V. All rights reserved.
Keywords: River floods; 14
C dating; Fluvial sediments; Environmental change; Europe
1. Introduction
Since the end of the 1950s, the number of investigations
in Europe of river system response to Holocene
environmental change has grown considerably. Presently,
Holocene fluvial sedimentary sequences and river terraces
of most major European river basins have been studied, to
some degree, although the resolution of dating control and
environmental reconstruction varies considerably between
catchments. From this research, the influence of both
human activity (primarily in terms of changes in land-use
and cover, and its impact on runoff and sediment supply)
and climate change (reflected principally by variations in
the occurrence of extreme hydrological events, notably
major floods and droughts) on Holocene river behaviour
have been documented. These factors have been found to
vary significantly both over time and geographically, and
have given rise to highly diverse and complex fluvial
sedimentary records. However, outside of a few relatively
well studied regions (e.g. Great Britain– Macklin and
Lewin, 1993, 2003; the Netherlands– Berendsen and
Stouthamer, 2001; Poland– Starkel, 1991; Spain– Benito
et al., 2003), Holocene river sequences in Europe have not
been analysed and compared systematically in a manner
likely to reveal an underlying structure, or pattern, within
the fluvial record. The International Council for Science
funded project ‘‘Past hydrological events related to
understanding global change’’ provided a timely collaboratory
framework to consider these important issues, and to
critically evaluate Holocene fluvial records across selected
0341-8162/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.catena.2005.07.015
* Corresponding author. Tel.: +44 1970 622656.
E-mail address: mvm@aber.ac.uk (M.G. Macklin).
Catena 66 (2006) 145 – 154
www.elsevier.com/locate/catena
parts of Europe and, from this, highlight gaps in our
understanding of longer term river dynamics and identify
future research priorities.
In this paper, recently compiled databases of 14
C dated
Holocene fluvial units from Great Britain (Macklin and
Lewin, 2003; Lewin et al., 2005; Johnstone et al., 2006-this
volume; Macklin et al., 2005), Poland (Starkel et al., 2006this
volume) and Spain (Thorndycraft and Benito, 2006-this
volume), in catchments unaffected by sea-level change, are
compared (Fig. 1). In order to identify possible large-scale
hydroclimatic teleconnections, correlations are also made
between Holocene riverine flooding episodes, the North
Atlantic ice rafting debris (IRD) record (Bond et al., 2001)
and phases of higher and lower lake levels in middle (Magny
et al., 2003) and southern Europe (Carrio´n, 2002). However,
one of the primary aims of adopting what might be termed a
Fmetadata_ approach in reviewing the three data sets was to
try and draw out some observations of more general
relevance, which would not have emerged from considering
each of the databases in isolation. This is certainly the first
time in Europe and, to the authors’ knowledge, probably
anywhere in the world that such a procedure has been used to
investigate Holocene river dynamics at approaching a
continental scale. The value of assembling and interrogating
large radiometrically dated fluvial databases for reconstructing
past hydrological events and for forecasting possible
river basin response to future environmental change is also
considered.
Fig. 1. Map of Europe showing the locations of Great Britain, Poland and Spain where databases of 14
C dated Holocene fluvial units have recently been
compiled.
M.G. Macklin et al. / Catena 66 (2006) 145–154146
2. Study regions and methodology
The history, nature and focus of Holocene river system
research over the past 30 to 40 years have differed markedly
in Great Britain, Poland and Spain. Research started in
Poland in the late 1950s with Starkel (1960) and his
colleagues, most notably Klimek and Starkel (1974), and
Kozarski and Rotnicki (1977), who pioneered the study of
late glacial and Holocene palaeohydrology in Europe,
particularly through the widespread use of 14
C dating of
peat and sub-fossil wood incorporated within fluvial
deposits. Much of the work during the 1980s and 1990s
centred on detailed investigations of longer-term river
channel planform development, especially palaeomeander
stratigraphies (e.g. Kozarski, 1983; Rotnicki, 1983). From
these studies, Starkel and his co-workers identified a large
number of relatively brief (c. 200–500 years) episodes
during the Holocene that were characterized by the more
frequent occurrence of major floods and linked to periods of
wetter and cooler climate (Starkel et al., 1996).
In Spain, research on Holocene river sequences began
somewhat later, principally through the studies of Vita-Finzi
(1969, 1976) who relied primarily on dating fluvial
sediments using incorporated or associated, archaeological
material and sites. He identified a unitary, pan-Mediterranean
post-Roman fluvial unit which he termed the ‘‘Younger
Fill’’, the deposition of which was dated to the late Roman
and medieval periods and attributed to climate change.
During the 1970s and 1980s, this view was challenged by
the findings of several archaeologically focused river valley
surveys, most notably in Greece (Pope and van Andel,
1984), that identified multiple phases of river erosion and
sedimentation and demonstrated what they believed to be a
better temporal correspondence between anthropogenically
induced land-cover change and longer term river behaviour
in the Mediterranean. Nevertheless, the number of catchments
in the region where Holocene fluvial sedimentary
sequences are well constrained by multiple radiometric
dates remains surprisingly few. Since the middle of the
1990s, however, there has been an increase in the number of
published fluvial 14
C dates in non-archaeological contexts,
especially through the work of Benito and co-workers’
investigations of slackwater sediments in bedrock gorges
resulting from very large flood events (Benito et al., 2003;
Thorndycraft et al., 2005).
Research on Holocene river development in Great Britain
was both by European and North American standards
initiated relatively recently. In the late 1970s and early
1980s, the first studies tended to be focused either in small,
formerly glaciated upland catchments of northern (Harvey et
al., 1981) and western (Macklin and Lewin, 1986) Britain or
in larger lowland basins in the southern and eastern part of
the country, at the margin (Brown and Barber, 1985) or
beyond the last ice sheet (Robinson and Lambrick, 1984).
Throughout the 1980s and up to the early 1990s, opinions
on the primary controls of Holocene river dynamics in Great
Britain tended to be somewhat caricatured either into
Fcultural_ or Fclimatic_ schools of thought (e.g. Ballantyne,
1991). The consensus today favours the views of Macklin
and Lewin (as first set out by Macklin et al., 1992; Macklin
and Lewin, 1993) who consider river response to environmental
change over the Holocene in terms of a continuum
between the two forcing factors, each of which can vary in
importance over time and space both within a single
drainage basin and between catchments.
All published (up to the end of 2003) and unpublished
(available to the authors) 14
C dated Holocene fluvial units in
Great Britain (506), Poland (589) and Spain (99) have been
assembled into a database that for each 14
C date includes
information on drainage basin area, depositional environment
and type of material used for dating. Following the
methodology recently outlined by Macklin and Lewin
(2003), 14
C dates which coincided with an abrupt modification
in sedimentation style or rate (termed Fchange_ dates),
and attributed to major floods, were also identified. Longterm
records of flooding in the three areas were reconstructed
using the frequency distributions of all Fchange_
dates (263, 335 and 51 in Great Britain, Poland and Spain,
respectively). Dates were calibrated and then for each region
plotted as cumulative probability density functions (CPDFs)
in OxCal (version 3.9; Bronk Ramsey, 1995, 2001). The
calibration curve (INTCAL98; Stuiver et al., 1998),
however, exerts an influence on the CPDFs so that a date
coinciding with a plateau in the calibration curve may
produce a smaller, less well-defined peak than a date which
occurs at a steep section of the curve. Macklin et al. (2005)
have developed a correction to account for this by
generating a second CPDF plot using a simulated data set
of evenly distributed 14
C dates and subtracting this from the
CPDF plot of observed 14
C dates. The resulting probability
difference curves (PDCs; Fig. 2) show the difference
between the two CPDFs where peaks and troughs relate to
apportioned and dispersed probabilities of several dated
units. Changes in the height of the PDCs over the Holocene
are interpreted to reflect variations in the occurrence of
major floods. Multi-centennial deviations above, or below,
the zero point on the PDCs represent extended periods
during which flood units are more common (indicative of
major flooding episodes), or when they are rare or absent
(phases when large floods occurred very infrequently or
when conditions did not favour the preservation of flood
sediments).
Finally, we anticipated that by comparing the results of
identical PDC-based analyses we could also evaluate the
degree to which the palaeohydrological records of the three
regions may be contingent upon the choice of investigative
theme, and/or study catchment. Major issues such as the
representativeness of the fluvial archive as a record of
environmental change, and the effects of sampling or
preservation bias on long-term flood histories, have very
rarely been addressed empirically (but see recent papers by
Lewin and Macklin, 2003; Lewin et al., 2005 on this topic).
M.G. Macklin et al. / Catena 66 (2006) 145–154 147
Fig. 2. PDCs of 14
C dates associated with major flooding episodes in Great Britain, Spain and Poland plotted alongside the North Atlantic IRD record (Bond et al., 2001) and hydrological records from lakes in
southern Spain (Carrio´n, 2002) and mid-Europe (Magny et al., 2003). Episodes of European flooding are listed on the right-hand axis and episodes occurring in all three study regions are denoted with an asterisk.
M.G.Macklinetal./Catena66(2006)145–154148
Understanding these factors is, however, critically important
if the alluvial record is going to be used, for instance, to
better inform and improve present and future flood risk
estimates.
3. The influence of drainage basin characteristics,
depositional environment and sedimentological factors
on the Holocene fluvial record
A total of nearly 1200 Holocene fluvial 14
C dates are
included within the British, Polish and Spanish databases
with Britain and Poland together contributing more than
90% of the radiometrically dated fluvial units (Table 1). The
proportion of 14
C dates that mark major changes in river
behaviour (coinciding with a modification in sedimentation
style or rate) is similar in all three areas, and generally just
exceeds 50%. Considering this from the reverse viewpoint,
nearly half of the 14
C dates obtained from fluvial
sedimentary sequences in Britain, Poland and Spain would
appear to relate to autogenic activity and, therefore, are of
relatively little value in interpreting external environmental
influence, or for reconstructing past hydrological events.
In Table 2, 14
C dates (and the proportion of these that
coincide with geomorphic change) are classified on the
basis of the size of drainage basin upstream from where 14
C
dating was undertaken, and significant differences are
evident between the three areas. In Great Britain, 44% of
the 14
C dates have come from relatively small river basins
with drainage areas of less than 100 km2
, compared to 22%
and 18% in similarly sized basins in Poland and Spain,
respectively. The majority of fluvial 14
C dates in Great
Britain come from catchments with drainage areas between
100 and 1000 km2
, whereas in Poland and particularly Spain
most reported 14
C dates come from the largest river basins
in both countries. These regional variations arise partly
because of physiographic differences between the three
areas, with river basins in continental Europe generally
being larger than those in Great Britain, but also because of
different research strategies adopted by local investigators.
This is also shown by the relationship between the
proportion of 14
C dates that mark geomorphic change and
catchment size. The proportions of so-called Fchange_ dates
in Great Britain systematically decrease with increasing
drainage basin area, while in both Poland and Spain the
reverse relationship is evident. Thus, in Great Britain,
catchments with drainage areas less than 10 km2
would
appear to be particularly sensitive to environmental changes
and/or favour the preservation of extreme flood events. By
contrast in Poland and Spain larger river basins with
drainage areas exceeding 10,000 km2
that have been
investigated produce a better record of allogenic forcing
over the Holocene. The relatively small number of 14
C dates
presently available from Spain makes it difficult to attribute
this pattern to any specific cause, and the relationship
between 14
C dates that mark geomorphic change and
drainage area may well be an artefact of the fact that the
Holocene fluvial histories of small catchments in Spain
(particularly those with drainage areas less than 10 km2
)
have been under-researched. However, the much larger
British and Polish databases appear to have been influenced
less by the deliberate or inadvertent selection of river basins
of a certain size for study. Therefore, variability in the
manner in which environmental change and flooding
episodes are recorded in the basins of varying size is likely
to reflect long-term differences in catchment hydrology,
river sediment transport processes and supply between the
two areas.
In Great Britain, for instance, examination of historical
channel change over the last 200 years in western and
northern catchments (Lewin, 1983, 1987) has shown a
tendency for average rates of geomorphic change to be
greatest in the middle reaches of rivers, or in the Fpiedmont
zone_ (Newson, 1981) where valleys open out from upland
areas but river gradients and stream power are still relatively
high. This may be one factor that accounts for the
progressive decrease in the percentage of change dates as
drainage area increases, particularly the abrupt fall off when
catchments exceed 1000 km2
that in the major easterly
draining rivers in northern Britain marks the downstream
limit of the piedmont zone. In larger British catchments,
although flood discharges are relatively high, river channel
gradients decrease to a point where stream powers are
generally too low to transport gravel-size sediment overbank
into long-term storage, thereby incorporating material of this
type into the alluvial Farchive_. Similarly, in small British
basins (10 km2
), the apparently high proportion of 14
C
dates that appear to reflect external forcing by environmental
change is likely to result partly from better slope-channel
coupling and higher rates of sediment supply but also
because of the greater geomorphological impact that
Table 1
Number of Holocene fluvial 14
C dates in Great Britain, Poland and Spain,
and those which are associated with geomorphic change
All 14
C dates FChange_ 14
C dates
Great Britain 506 263 (52%)
Poland 589 335 (57%)
Spain 99 51 (52%)
Total 1194 649 (54%)
Table 2
Number of fluvial 14
C dates classified by drainage area
Drainage area (km2
) Great Britain Poland Spain
<1 79 (60, 76%) 12 (0, 0%) 0 (0, 0%)
1 to 10 77 (59, 77%) 24 (0, 0%) 3 (1, 33%)
10 to 100 65 (32, 49%) 96 (54, 56%) 15 (5, 33%)
100 to 1000 214 (98, 46%) 95 (60, 63%) 22 (12, 55%)
1000 to 10,000 71 (14, 20%) 249 (150, 60%) 22 (12, 55%)
>10,000 0 (0, 0%) 113 (71, 63%) 28 (21, 75%)
Unclassified 0 (0, 0%) 0 (0, 0%) 9 (0, 0%)
The number and proportion (expressed as a percentage) of Fchange_ dates
are shown in brackets.
M.G. Macklin et al. / Catena 66 (2006) 145–154 149
localized extreme rainfall events have on small catchments.
However, in Poland, given the extensive mountain headwaters
of the upper Vistula, the apparently low number of
14
C dates (especially those marking geomorphic change) in
catchments with drainage areas of less than 10 km2
must, to
a large degree, reflect the lack of research in river basins of
this size. The greater proportion of Fchange_ 14
C dates in
Poland’s larger river basins arises because of higher stream
powers particularly in the steeper gradient river systems
(e.g. upper Vistula, Raba and Wisloka) draining the foreland
of the Carpathian Mountains, which are also affected by
major ice-jam floods.
Depositional environment, as well as river activity rates
(both in the lateral and vertical sense) and style, has been
shown to have a major effect on the length and completeness
of the Holocene fluvial archive (Lewin and Macklin,
2003; Lewin et al., 2005), particularly the record of extreme
flood events (Macklin and Lewin, 2003). In order to explore
these relationships in more detail, each 14
C date has been
classified according to the depositional environment from
which it was collected. The majority of 14
C dates in all three
areas come from either palaeochannel fills or overbank
sediments, with the former providing the largest number of
14
C dates in Poland and the latter depositional environment
accounting for most 14
C dates in both Great Britain and
Spain (Table 3). However, with respect to 14
C dates that
pick out geomorphologically significant change in river
activity over the Holocene, flood basin sediments in Great
Britain, palaeochannel fills in Poland and overbank (slackwater)
sediments in Spain are the most important as
recorders of environmental change. Factors likely to favour
the recording and preservation of an environmental signal in
these depositional contexts are firstly, that major flood
events are commonly registered by abrupt changes in
sediment size and secondly, flood sediments are preferentially
incorporated into depositional niches located some
distance away from the main river channel or channel belt.
In Spain the absence of 14
C dates from palaeochannel fills is
an anomaly, especially as dateable material has been
recovered from channel bed and bar sediments (Table 3),
and must be attributed (assuming that 14
C dates have been
correctly classified) to researcher bias in the selection of
study sites for investigation.
Finally, to evaluate whether sampling strategies adopted
for dating varied between the three areas, and whether this
may have affected the fluvial record, 14
C dates were
classified by type of organic material used for dating (Table
4). A considerable range of material has been used and some
quite major differences are evident in the three regions. In
Spain, with its Mediterranean climate and a higher incidence
of natural fires in the landscape, charcoal is the most
common material that has been recovered and used for 14
C
dating. By contrast, within cooler and generally wetter
British and Polish catchments, samples from peat and wood
have together provided the largest number of 14
C dates. In
Poland, organic muds formed either in standing water within
cutoffs or in flood basins have also been used frequently in
dating. Similar dateable material, however, is rare within
British and Spanish alluvial sequences, probably as a
consequence of longer histories of land drainage resulting
in the loss of river wetlands (containing seasonal or more
permanent open water bodies) suitable for the deposition of
organic-rich muds.
4. The relationship between large-scale Holocene hydroclimatic
teleconnections in Europe and episodes of major
riverine flooding
One of the principal aims of this multi-national project
was to improve our understanding of the impact of climate
change on the spatial and temporal occurrence of extreme
hydrological events during the Holocene. The construction
of PDCs of 14
C dated flood units in Great Britain, Poland
and Spain provide, for the first time in Europe, probabilitybased
records of riverine flooding that extend over the entire
Holocene. Periods of more frequent flooding are recognised
by peaks (above zero) in the PDCs and are listed in Table 5.
From these, where PDC peaks in two or more areas fall
within a 100-year range, major flood episodes (possibly of
pan-European extent) have been identified. In Fig. 2, the
Holocene flood series for Great Britain, Poland and Spain
Table 4
Number of fluvial 14
C dates classified by type of organic material used for
14
C dating
Organic material used
for 14
C dating
Great Britain Poland Spain
Bone 17 (7, 41%) 0 (0, 0%) 2 (0, 0%)
Calcified roots 0 (0, 0%) 0 (0, 0%) 2 (0, 0%)
Charcoal 35 (10, 29%) 0 (0, 0%) 50 (36, 72%)
Organic mud 15 (10, 67%) 139 (90, 65%) 3 (1, 33%)
Peat 173 (120, 69%) 269 (144, 53%) 0 (0, 0%)
Plant macrofossils 19 (11, 58%) 48 (18, 38%) 0 (0, 0%)
Pollen 0 (0, 0%) 0 (0, 0%) 3 (1, 33%)
Shell 0 (0, 0%) 0 (0, 0%) 7 (6, 86%)
Soil 40 (32, 80%) 0 (0, 0%) 0 (0, 0%)
Wood 181 (65, 36%) 110 (75, 68%) 2 (0, 0%)
Unclassified 26 (8, 31%) 23 (8, 35%) 30 (7, 23%)
The number and proportion (expressed as a percentage) of Fchange_ dates
are shown in brackets.
Table 3
Number of fluvial 14
C dates classified by depositional environment
Depositional environment Great Britain Poland Spain
Channel bed and
bar sediments
61 (22, 36%) 82 (48, 58%) 11 (1, 9%)
Palaeochannel fills 147 (59, 40%) 241 (152, 63%) 0 (0, 0%)
Overbank sediments 177 (99, 55%) 183 (114, 62%) 64 (44, 69%)
Flood basin sediments 112 (78, 70%) 46 (21, 46%) 10 (2, 20%)
Debris flow/colluvial
sediments
9 (5, 55%) 15 (0, 0%) 8 (4, 50%)
Unclassified 0 (0, 0%) 22 (0, 0%) 6 (0, 0%)
The number and proportion (expressed as a percentage) of Fchange_ dates
are shown in brackets.
M.G. Macklin et al. / Catena 66 (2006) 145–154150
are compared with hydrological records from lakes in
southern Spain and mid-Europe, and the North Atlantic
IRD record. Major flooding episodes in Europe (PDC peak
in two or more regions) are shown by solid horizontal grey
lines, while PDC peaks that are evident only in one of the
three flood records are denoted by broken horizontal grey
lines. Fifteen major periods of flooding are discernible, six
of which (at c. 570, 660, 860–880, 1930–1950, 2370–2810
and 4840–4860 cal. BP) are recorded in river basins across
Europe. Comparison with the similarly well dated, lakelevel
record of mid-Europe shows that 11 out of 15 peaks of
flooding coincide with phases of higher lake level. It is also
noteworthy that the majority (7 out of 11) of PDC peaks
occur either at the very beginning or during the early part of
the higher lake-level phases. This would indicate not only a
common underlying control but also that river basins in
Europe (through changes in the frequency and magnitude of
geomorphologically Feffective_ floods) respond more rapidly,
and perhaps more sensitively, to short-term, external
climate forcing than do lake systems in the region. This is a
significant finding particularly in the context of the current
research on detection and attribution of climate change
signals associated with global warming, based on hydrological
indicators.
Out of the four major riverine flooding episodes (at c.
860–880, 1040, 1830–1950 and 6790–6820 cal. BP) that
do not coincide with phases of higher lake level in midEurope,
three are recorded within the last 2000 years. The
most prominent of these are those at c. 860–880 and 1830–
1950 cal. BP, which occur at the same time as significant
agricultural expansion, deforestation and land-use change in
the Roman period (Great Britain and Poland) and during the
Middle Ages (all three regions). As has been demonstrated
by many earlier studies (Butzer, 1980; Macklin et al., 1992;
Macklin and Lewin, 1993; Kalicki, 1996; Coulthard and
Macklin, 2001; Benito, 2003; Macklin and Lewin, 2003),
although large-scale forest removal and agricultural development
augmented both runoff and sediment supply, the
geographically widespread nature of these sedimentation
events suggests that climate-related changes in flood
frequency and magnitude played a role as well. Indeed,
documentary records of flooding in both Great Britain
(Brown, 1998) and Spain (Benito et al., 1996) show that the
early part of the Middle Ages was characterized by an
increased frequency of large floods, which corresponded to
changes of prevailing atmospheric circulation patterns
affecting western Europe at that time. This period of marked
climatic variability is not, however, manifested in the lake
record of middle Europe. This is possibly due to its short
duration but also may reflect the relative insensitivity in this
part of Europe of lake systems to hydroclimatic changes
except those of an extreme nature lasting several centuries.
Before 5000 cal. BP most of the major flood periods
match Holocene cooling phases marked by IRD events in
the North Atlantic Ocean (Fig. 2). However, the widespread
so-called 8.2 ka cold event (Alley et al., 1997), coincides
with prominent troughs in all three flood records, indicating
a very low incidence (or preservation) of major floods.
Indeed, in both Great Britain and Spain the period c. 7600–
8400 cal. BP appears from the fluvial record to have been
particularly dry, an interpretation also supported by very low
lake levels documented in southern Spain (Carrio´n, 2002) at
this time. After 5000 cal. BP, the association between
marine ice-drift indices and riverine flooding becomes much
weaker with only four (at c. 570, 660, 1040, 2730–2810 cal.
BP) out of the nine post 5000 cal. BP major flood episodes
corresponding with IRD peaks in the North Atlantic sector.
This weaker coupling between cooling in the North Atlantic
Ocean and hydrological change since c. 5000 cal. BP is also
apparent in the mid-European lake record (Magny et al.,
2003). A similar mid-Holocene hydroclimate Fsystem
switch_ (Steig, 1999) has been shown recently in many
other European terrestrial (Leuschner et al., 2002; Magny
and Haas, 2004) and marine (Hall et al., 2004) proxyclimate
records, and would suggest external forcing of
hydroclimate during this period not only by changes in
ocean circulation but also increasingly by variations in solar
activity as well (Macklin et al., 2005).
Over both the instrumental and documentary record, the
most severe and widespread floods in Europe have generally
occurred during cooler periods of climate, when the lateral
temperature gradient is steeper and circumpolar air masses
Table 5
Periods of major flooding in Great Britain, Spain and Poland
Great Britain Spain Poland
570 570 570
660 660 660
860 865 880
1040 1040
1290 1310
1650
1950 1930 1830
2280
2550 2560
2730 2750 2810
3540 3510
4520
4840 4860 4840
5540 5640
5730
5920
6250
6820 6790
7260
7590 7590
8400
8820
8870
9010
9330
9440 9530
10490
11160 11200
Flooding episodes recorded in two or more areas are shown in bold (ages
cal. BP).
M.G. Macklin et al. / Catena 66 (2006) 145–154 151
are shifted southwards favouring more frequent and
enhanced occurrence of meridional wind patterns in mid
latitudes (Benito et al., 1996; Rumsby and Macklin, 1996).
Comparing the timing of widespread flooding episodes
identified in this paper with climate reconstructions from
lake and marine records also suggests that a similar
relationship between climate and flooding in Europe existed
during much of the Holocene (i.e. flood frequency and
magnitude were higher during relatively cool, wetter
periods). Furthermore, our data generally supports Magny
et al.’s (2003) hydrological reconstruction for the 8.2 ka
cold event and other Holocene climate cooling phases,
although the flood record indicates that cyclonic activity
during these later periods extended further southwards in
Europe, to around 36-N in southern Spain. This would
indicate that the timing of hydrological change in the midlatitudes
of Europe from between 59- and 36-N, and at least
as far east as 24-E (eastern Poland), was broadly synchronous
during the Holocene, especially with respect to the
incidence of extreme hydrological events.
Identification of periods during the Holocene characterized
by significantly fewer 14
C dated fluvial units in Great
Britain, Poland and Spain is fairly straightforward (Table 6),
although attributing these to either a lower frequency of
large floods or local preservation factors is more problematic.
Since c. 6000 cal. BP, four episodes of apparently
reduced river activity at c. 1420–1690, 2180–2350, 2950–
3400 and 4870–5400 cal. BP are recorded in all three
regions, which suggest that these were actually times when
large floods were relatively uncommon in Europe. There
are, however, major differences in the overall temporal
distribution of Holocene fluvial units between Great Britain,
Poland and Spain, shown clearly by the contrasting shapes
of the three PDCs (Fig. 2). In Great Britain and Spain, the
probability of flood unit preservation progressively
decreases with age until c. 8000 cal. BP, indicating the
generally eroding nature of British and Spanish river
systems until this time (at least those which have been
studied) and the resulting loss of older units. Alluvial units
deposited before c. 8400 cal. BP, however, are better
preserved in both areas. In the case of Great Britain, this
has been partly attributed to downcutting in many river
valleys during the early Holocene (Lewin and Macklin,
2003), but it may also reflect the size of the forcing events
between c. 11,200 and 9000 cal. BP, which in Spain was a
period characterized by a series of very large floods (Benito
et al., 2003). The form of the age–frequency plot for dated
flood sediments in Poland indicates a rather different set of
preservation factors have operated to produce the temporal
bias observed in the fluvial record. The broadly negative
exponential relationship between age and the number of
flood units found in the larger alluvial river systems of
Poland back to c. 5500 cal. BP is a pattern that would be
expected to be associated with long-term lateral migration
of river channels, and limited channel incision or aggradation
(Lewin and Macklin, 2003). Immediately prior to
c. 5500 cal. BP the higher frequency of flood units suggests
a different river activity style in the early Holocene of
episodic valley floor aggradation followed by incision that
favoured alluvial preservation. Thus, while the timing of
many of the high-frequency Fspikes_ (related to climatically
controlled change in flood occurrence and magnitude) in
the PDCs are very similar in Great Britain, Poland and
Spain, the primary determinant on preservation potential, as
reflected by the overall form of the PDC in each area, is the
regional context and local river system behaviour.
5. Conclusions and future research priorities
This paper provides a probability-based record of
riverine flooding in Europe during the Holocene and has
been constructed using a new, large database of nearly
1200 14
C dated fluvial deposits from Great Britain, Poland
and Spain. Employing a novel, sedimentologically based
methodology recently developed by Macklin and Lewin
(2003) and Macklin et al. (2005), phases of significantly
higher, and lower, flood occurrence have been identified
during the Holocene and dated. Fifteen major periods of
flooding are evident (Table 5; Fig. 2), eleven of which
coincide with phases of lake level rise in mid-Europe. This
demonstrates a common underlying climatic control but also
somewhat surprisingly indicates that river basins can
respond, through variations in the frequency and magnitude
of major floods, more rapidly to climate change than lake
systems. Indeed, punctuated instability appears to be an
inherent, and natural, characteristic of many rivers in Europe
over the last 11,500 years. We have also addressed, using
one of the largest Holocene river databases so far constructed
in Europe, major issues such as the Frepresentativeness_ of
the fluvial archive as a record of environmental change (cf.
Richards, 2002), as well as the effect of sampling and
Table 6
Periods characterized by significantly fewer 14
C dated flood units in Great
Britain, Spain and Poland
Great Britain Spain Poland
260–560
690–760 690–760
950–1040 1060–1260
1420–1550 1300–1870 1540–1690
2180–2340 2000–2350 2150–2750
2950–3400 2830–3440 2870–4830
3560–4170
4350–4830
4870–5320 5960–5590 4860–5600
6270–6690 5660–7220 6000–6150
6400–6670
6860–10440 7300–7530 7420–7570
7610–8780 7680–7980
9020–9270 8630–9000
9530–10280 9710–10180
Episodes in the late and middle Holocene recorded in all three areas are
shown in bold (ages cal. BP).
M.G. Macklin et al. / Catena 66 (2006) 145–154152
preservation bias on reconstructing longer-term flood
histories.
Although recent developments in the numerical modelling
of Holocene drainage basin and alluvial valley floor
evolution are showing considerable promise (e.g. Coulthard
and Macklin, 2001; Coulthard et al., 2005), scale and
process representation (for instance, lateral channel movement),
as well as the availability of suitable data to drive
models (climate and flood series), remain as significant
constraints in their wider use. The empirical database
approach outlined in this paper offers a complementary
data analysis tool to numerical modelling. In interpretation
terms, it is similarly process-based but it presently has the
significant advantage of being able to readily utilise the
many thousands of 14
C dates that have already been
obtained from fluvial contexts in Europe, and elsewhere
in the world, over the last 30 or so years. However,
answering some research questions may need a more
purposive sampling and study strategy than has hitherto
been adopted. The Holocene alluvial archive currently
consists of a somewhat disparate set of site studies
conducted for a variety of purposes. Although the metadata
analysis procedure we have adopted has facilitated the
reconstruction of major hydrological events from the
European Holocene fluvial record, it has also identified
certain periods (e.g. the early Holocene in Great Britain),
regions (e.g. small river basins in Spain and Poland) and
depositional settings (e.g. palaeochannels in Spain) in each
of the three study areas for which there is currently very
little information. These need to be targeted in future
investigations. Key research questions that have emerged
from this present study are: 1. How did Holocene river
dynamics and flooding impact on human settlement and
exploitation of river valleys in the prehistoric and early
historic periods? 2. To what degree does land cover in a
river basin have to change in order to either amplify or
reduce the climate signal recorded in fluvial sediments? 3. Is
the location of land-use change within a catchment
important in this respect? Further interrogation and development
of the Holocene fluvial database is required to
address these topics, as well as to try and evaluate
differential response of contiguous catchments to identical
and/or simultaneous climate/land-use forcing. Such information
could be invaluable for improving forecasts of
possible river system response to present and future
environmental change. Finally, the data analysis methodology
outlined here is generic and could readily form the basis
for compiling and extending records of extreme flood
events worldwide, particularly in ungauged catchments and
regions where documentary records of flooding do not exist.
Acknowledgements
Funding for this collaborative research project was
principally provided by ICSU (International Council for
Science), NERC (NER/A/S/2001/00454 to Macklin and
Lewin) and Historic Scotland (Macklin and Johnstone) and
this support are gratefully acknowledged. We also wish to
thank Dr Paul Brewer, Dr Simon Gittins, Dr Marco van de
Wiel, Professor Andreas Lang and Dr Barbara Rumsby for
their help in developing the database and for their very
useful comments on this paper.
References
Alley, R.B., Mayewski, P.A., Sowers, T., Stuiver, M., Taylor, K.C., Clark,
P.U., 1997. Holocene climatic instability: a prominent, widespread
event 8200 yr ago. Geology 25, 483–486.
Ballantyne, C.K., 1991. Late Holocene erosion in upland Britain: climatic
deterioration or human influence? Holocene 1, 81–85.
Benito, G., 2003. Palaeohydrological changes in the Mediterranean region
during the Late Quaternary. In: Gregory, K.J., Benito, G. (Eds.),
Palaeohydrology: Understanding Global Change. Wiley and Sons,
Chichester, pp. 123–142.
Benito, G., Machado, M.J., Pe´rez-Gonza´lez, A., 1996. Climate change
and flood sensitivity in Spain. In: Branson, J., Brown, A.G., Gregory,
K.J. (Eds.), Global Continental Changes: The Context of Palaeohydrology,
Special Publication-Geological Society of London,
vol. 115, pp. 85–98.
Benito, G., Sopen˜a, A., Sa´nchez-Moya, Y., Machado, M.J., Pe´rezGonza´lez,
A., 2003. Palaeoflood record of the Tagus River (central
Spain) during the Late Pleistocene and Holocene. Quaternary Science
Reviews 22, 1737–1756.
Berendsen, H.J.A., Stouthamer, E., 2001. Palaeogeographic Development
of the Rhine–Meuse Delta. Van Gorcum, Assen.
Bond, G., Kromer, B., Beer, J., Muscheler, R., Evans, M.N., Showers, W.,
Hoffman, S., Lotti-Bond, R., Hajdas, I., Bonani, G., 2001. Persistent
solar influence on North Atlantic climate during the Holocene. Science
294, 2130–2136.
Bronk Ramsey, C., 1995. Radiocarbon calibration and analysis of
stratigraphy: the OxCal program. Radiocarbon 37, 425–430.
Bronk Ramsey, C., 2001. Development of the radiocarbon program OxCal.
Radiocarbon 43, 355–363.
Brown, A.G., 1998. Fluvial evidence of the medieval warm period and the
late medieval climatic deterioration in Europe. In: Benito, G., Baker,
V.R., Gregory, K.J. (Eds.), Palaeohydrology and Environmental
Change. Wiley and Sons, Chichester, pp. 43–52.
Brown, A.G., Barber, K.E., 1985. Late Holocene paleoecology and
sedimentary history of a small lowland catchment in central England.
Quaternary Research 24, 87–102.
Butzer, K.W., 1980. Holocene alluvial sequences: problems of dating
and correlation. In: Cullingford, R.A., Davidson, D.A., Lewin, J.
(Eds.), Timescales in Geomorphology. Wiley and Sons, Chichester,
pp. 131–142.
Carrio´n, J.S., 2002. Patterns and processes of Late Quaternary environmental
change in a montane region of southwestern Europe. Quaternary
Science Reviews 21, 2047–2066.
Coulthard, T.J., Macklin, M.G., 2001. How sensitive are river systems to
climate and land-use changes? A model based evaluation. Journal of
Quaternary Science 16, 346–351.
Coulthard, T.J., Lewin, J., Macklin, M.G., 2005. Modelling differential
catchment response to environmental change. Geomorphology 69,
222–241.
Hall, I.R., Bianchi, G.G., Evans, J.R., 2004. Centennial to millennial scale
Holocene climate-deep water linkage in the North Atlantic. Quaternary
Science Reviews 23, 1529–1536.
Harvey, A.M., Oldfield, F., Baron, A.F., Pearson, G.W., 1981. Dating of
post-glacial landforms in the central Howgills. Earth Surface Processes
and Landforms 6, 401–412.
M.G. Macklin et al. / Catena 66 (2006) 145–154 153
Johnstone, E., Macklin, M.G., Lewin, J., 2006. The development and
application of a database of radiocarbon dated Holocene fluvial deposits
in Great Britain. Catena 66, 14–23. doi:10.1016/j.catena.2005.07.006
(this volume).
Kalicki, T., 1996. Climatic or anthropogenic alluviation in central European
valleys during the Holocene? In: Branson, J., Brown, A.G., Gregory,
K.J. (Eds.), Global Continental Changes: The Context of Palaeohydrology,
Special Publication-Geological Society of London, vol. 115,
pp. 205–215.
Klimek, K., Starkel, L., 1974. History and actual tendency of flood-plain
development at the border of the Polish Carpathians. Abhandlungen der
Akademie der Wissenschaften in Goettingen. Mathematisch- Physikalische
(Klasse 3) 29, 185–196.
Kozarski, S., 1983. River channel changes in the middle reach of the
Warta valley, Great Poland Lowland. Quaternary Studies in Poland 4,
159–169.
Kozarski, S., Rotnicki, K., 1977. Valley floors and changes of river channel
pattern in the North Polish Plain during the Late-Wu¨rm and Holocene.
Questiones Geographicae 4, 51–93.
Leuschner, H.H., Sass-Klaassen, U., Jansma, E., Baillie, M.G.L., Spurk,
M., 2002. Subfossil European bog oaks: population dynamics and longterm
growth depressions as indicators of changes in the Holocene
hydro-regime and climate. Holocene 12, 695–706.
Lewin, J., 1983. Changes of channel patterns and floodplains. In: Gregory,
K.J. (Ed.), Background to Palaeohydrology. Wiley and Sons, Chichester,
pp. 303–319.
Lewin, J., 1987. Historical river channel changes. In: Gregory, K.J., Lewin,
J., Thornes, J.B. (Eds.), Palaeohydrology in Practice: A River Basin
Analysis. Wiley and Sons, Chichester, pp. 161–175.
Lewin, J., Macklin, M.G., 2003. Preservation potential for Late Quaternary
river alluvium. Journal of Quaternary Science 18, 107–120.
Lewin, J., Macklin, M.G., Johnstone, E., 2005. Interpreting alluvial
archives: sedimentological factors in the British Holocene fluvial
record. Quaternary Science Reviews 24, 1873–1889.
Macklin, M.G., Lewin, J., 1986. Terraced fills of Pleistocene and
Holocene age in the Rheidol Valley, Wales. Journal of Quaternary
Science 1, 21–34.
Macklin, M.G., Lewin, J., 1993. Holocene river alluviation in Britain.
Zeitschrift fu¨r Geomorphologie. Supplementband 88, 109–122.
Macklin, M.G., Lewin, J., 2003. River sediments, great floods and
centennial-scale Holocene climate change. Journal of Quaternary
Science 18, 101–105.
Macklin, M.G., Passmore, D.G., Rumsby, B.T., 1992. Climatic and cultural
signals in Holocene alluvial sequences: the Tyne basin, northern
England. In: Needham, S., Macklin, M.G. (Eds.), Alluvial Archaeology
in Britain. Oxbow Press, Oxford, pp. 123–139.
Macklin, M.G., Johnstone, E., Lewin, J., 2005. Pervasive and long-term
forcing of Holocene river instability and flooding in Great Britain by
centennial-scale climate change. Holocene 15 (7), 937–943.
Magny, M., Haas, J.N., 2004. A major widespread climatic change around
5300 cal. yr BP at the time of the Alpine Iceman. Journal of Quaternary
Science 19, 423–430.
Magny, M., Be´geot, C., Guiot, J., Peyron, O., 2003. Contrasting patterns of
hydrological changes in Europe in response to Holocene climate
cooling phases. Quaternary Science Reviews 22, 1589–1596.
Newson, M.D., 1981. Mountain streams. In: Lewin, J. (Ed.), British Rivers.
Allen and Unwin, London, pp. 59–89.
Pope, K.O., van Andel, T.H., 1984. Late Quaternary alluviation and soil
formation in southern Argolid: its history, causes and archaeological
implications. Journal of Archaeological Science 11, 281–306.
Richards, K., 2002. Drainage basin structure, sediment delivery and the
response to environmental change. In: Jones, S.J., Frostick, L.E. (Eds.),
Sediment Flux to Basins: Causes, Controls and Consequences, Special
Publication-Geological Society of London, vol. 191, pp. 141–160.
Robinson, M.A., Lambrick, G.H., 1984. Holocene alluviation and hydrology
in the upper Thames basin. Nature 308, 809–814.
Rotnicki, K., 1983. Modelling past discharges of meandering rivers. In:
Gregory, K.J. (Ed.), Background to Palaeohydrology. Wiley and Sons,
Chichester, pp. 321–354.
Rumsby, B.T., Macklin, M.G., 1996. River response to the last Neoglacial
(the FLittle Ice Age_) in northern, western and central Europe. In:
Branson, J., Brown, A.G., Gregory, K.J. (Eds.), Global Continental
Changes: The Context of Palaeohydrology, Special Publication-Geological
Society of London, vol. 115, pp. 217–233.
Starkel, L., 1960. Development of the relief of the Polish Carpathians in the
Holocene (in Polish with English summary). Prace Geograficzne,
Institute of Geography, Polish Academy of Sciences 22, 1–239.
Starkel, L., 1991. The Vistula River valley: a case study for central Europe.
In: Starkel, L., Gregory, K.J., Thornes, J.B. (Eds.), Temperate Palaeohydrology:
Fluvial Processes in the Temperate Zone During the Last
15000 Years. Wiley and Sons, Chichester, pp. 171–188.
Starkel, L., Kalicki, T., Kra˛piec, M., Soja, R., Ge˛bica, P., Czyz˙owska,
E., 1996. Hydrological changes of valley floor in the upper
Vistula basin during late Vistulian and Holocene. In: Starkel, L.
(Ed.), Evolution of the Vistula River Valley During the Last 15000
Years, Geographical Studies Special Issue, vol. 9. Polish Academy of
Sciences, Institute of Geography and Spatial Organization, Wrocaaw,
pp. 7–128.
Starkel, L., Soja, R., Michczyn´ska, D.J., 2006. Past hydrological events
reflected in the Holocene history of Polish rivers. Catena 66, 24–33.
doi:10.1016/j.catena.2005.07.008 (this volume).
Steig, E.J., 1999. Mid-Holocene climate change. Science 286, 1485–1487.
Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A.,
Kromer, B., McCormac, G., van der Plicht, J., Spurk, M., 1998.
INTCAL98 radiocarbon age calibration, 24000–0 cal BP. Radiocarbon
40, 1041–1083.
Thorndycraft, V.R., Benito, G., 2006. Late Holocene fluvial chronology in
Spain: the role of climatic variability and human impact. Catena 66,
34–41. doi:10.1016/j.catena.2005.07.007 (this volume).
Thorndycraft, V.R., Benito, G., Rico, M., Sa´nchez-Moya, Y., Sopen˜a, A.,
Casas, A., 2005. A long-term flood discharge record derived from
slackwater flood deposits of the Llobregat River, NE Spain. Journal of
Hydrology 313, 16–31.
Vita-Finzi, C., 1969. The Mediterranean Valleys: Geological Changes in
Historical Times. Cambridge University Press, Cambridge.
Vita-Finzi, C., 1976. Diachronism in Old World alluvial sequences. Nature
263, 218–219.
M.G. Macklin et al. / Catena 66 (2006) 145–154154