Content uploaded by Andreas Gobiet
Author content
All content in this area was uploaded by Andreas Gobiet on Feb 06, 2014
Content may be subject to copyright.
21st century climate change in the European Alps—Areview
☆
Andreas Gobiet
a,
⁎
, Sven Kotlarski
b
, Martin Beniston
c
, Georg Heinrich
a
, Jan Rajczak
b
,MarkusStoffel
c
a
Wegener Center for Climate and Global Change, University of Graz, Brandhofgasse 5, 8010 Graz, Austria
b
Institute for Atmospheric and Climate Science, ETH Zurich, Universitätstrasse 16, 8092 Zurich, Switzerland
c
Institute for Environmental Sciences, University of Geneva, Site de Battelle− Bâtiment D, 7, route de Drize− 1227 Carouge, Geneva, Switzerland
HIGHLIGHTS
• Warming is expected to accelerate throughout the 21st century in the Alpine region.
• Seasonal shifts in precipitation, global radiation, and relative humidity are expected.
• Precipitation and temperature extremes are expected to intensify.
• Snow cover is expected to drastically decrease below 1500–2000 m elevation.
• Further changes related to droughts and natural hazards are expected.
abstractarticle info
Article history:
Received 17 May 2013
Received in revised form 14 July 2013
Accepted 14 July 2013
Available online xxxx
Editor: D. Barcelo
Keywords:
Climate change
Alpine region
Extremes
Snow
Drought
Natural hazards
Reliable estimates of future climate change in the Alps are relevant for large parts of the European society. At the
same time, the complex Alpine region poses considerable challenges to climate models, which translate to uncer-
tainties in the climate projections. Against this background, the present study reviews the state-of-knowledge
about 21st century climate change in the Alps based on existing literature and additional analyses. In particular,
it explicitly considers the reliability and uncertainty of climate projections.
Results show that besides Alpine temperatures, also precipitation, global radiation, relative humidity, and closely
related impacts like floods, droughts, snow cover, and natural hazards will be affected by global warming.
Under the A1B emission scenario, about 0.25 °C warming per decade until the mid of the 21st century and accel-
erated 0.36 °C warming per decade in the second half of the century is expected. Warming will probably be
associated with changes in the seasonality of precipitation, global radiation, and relative humidity, and more
intense precipitation extremes and flooding potential in the colder part of the year. The conditions of currently
record breaking warm or hot winter or summer seasons, respectively, may become normal at the end of the
21st century, and there is indication for droughts to become more severe in the future. Snow cover is expected
to drastically decrease below 1500–2000 m and natural hazards related to glacier and permafrost retreat are
expected to become more frequent.
Such changes in climatic parameters and related quantities will have considerable impact on ecosystems and
society and will challenge their adaptive capabilities.
© 2013 The Authors. Published by Elsevier B.V. All rights reserved.
1. Introduction
Centrally located in the European continent and densely popu-
lated for most of its parts, the European Alps (Fig. 1)constitute
a dominant feature of the European landscape. Characterized by ex-
tensive lowlands, deeply incised valleys and peaking at an elevation of
more than 4800 m, the Alps are subject to a strong topographic variabil-
ity. The Alpine climate, its spatio-temporal variability and long-term
changes as well as its influence on various natural and socio-economic
sectors have been of high scientific interest for a long time (e.g., Auer
et al., 2007; Beniston and Jungo, 2002; Brunetti et al., 2009; Haeberli
and Beniston, 1998; OECD, 2007; Raible et al., 2006; Scherrer et al.,
2004; Stefanicki et al., 1998). This resulted in some of the world's lon-
gest observational time series of climatic parameters and a compara-
tively high observational network density (e.g., Barry, 1994).
A comprehensive overview on the Alpine climate, including its
major drivers and feedbacks to larger-scale flow conditions, has been
provided by Schär et al. (1998). Prominent features include distinct cli-
matic gradients in all three dimensions of space, the frequent occur-
rence of extreme precipitation events with associated hazards, the
Science of the Total Environment xxx (2013) xxx–xxx
☆
This is an open-access article distributed under the terms of the Creative Commons
Attribution-NonCommercial-ShareAlike License, which permits non-commercial use,
distribution, and reproduction in any medium, provided the original author and source
are credited.
⁎ Corresponding author. Tel.: +43 3163808448.
E-mail addresses: andreas.gobiet@uni-graz.at (A. Gobiet), sven.kotlarski@env.ethz.ch
(S. Kotlarski), martin.beniston@unige.ch (M. Beniston), g.heinrich@uni-graz.at
(G. Heinrich), jan.rajczak@env.ethz.ch (J. Rajczak), markus.stoffel@unige.ch (M. Stoffel).
STOTEN-15022; No of Pages 14
0048-9697/$ – see front matter © 2013 The Authors. Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.scitotenv.2013.07.050
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
importance of perennial snow and ice cover in the upper reaches as well
as the existence of various orographically triggered flow phenomena.
Due to various factors including the high population density, the
pronounced susceptibility to climate-related hazards, the importance
of Alpine water resources for energy production, the perception of the
Alps as “Water towers for Europe” (EEA, 2009), and the high economic
importance of both winter and summer tourism, robust and reliable
estimates of future climate change in the Alps are ultimately relevant
for large parts of the European society. Accordingly, numerous studies
have been concerned with the consequences of future climate change
on different sectors of the Alpine environment and its economy
(e.g., Beniston, 2007a; Beniston et al., 2011b; Elsasser and Bürki, 2002;
Haeberli and Beniston, 1998; OcCC, 2007; Steiger, 2010; Wolf et al.,
2012). Regional climate service initiatives including the Swiss CH2011
scenarios in the Western Alps (CH2011, 2011) or the Styrian STMK12
scenarios in the Eastern Alps (Gobiet et al., 2012) try to provide the cor-
responding climate scenario products in an end-user friendly and read-
ily accessible manner.
At the same time and due to a large number of small-scale atmo-
spheric phenomena, the Alpine region poses particular challenges to
climate models, which are routinely used to translate an expected
future increase in atmospheric greenhouse gas (GHG) and aerosol con-
centrations into climatic changes at regional to local scales. Previous
studies have for instance shown that today's state-of-the-art regional
climate models (RCMs) are able to represent the broad characteris-
tics of the Alpine climate, but that important mod el biases remain
for specific aspects (e.g. Frei et al., 2003; Haslinger et al., 2013;
Kotlarski et al., 2010; Prömmel et al., 2010; Rajczak et al., 2013;
Suklitsch et al., 2008, 2011). These biases partly translate into uncer-
tainties of future climate change projections, thereby adding to o ther
uncertainty sources lik e assumptions on future anthropog enic GHG
emissions or internal climate variability.
The aim of the present study, which was initiated in the framework
of the EU FP7 project ACQWA (http://www.acqwa.ch/; Beniston et al.,
2011a), is to review the current state-of-knowledge concerning 21st
century climate change in the Alps based on existing literature and
additional analyses of RCM simulations. In particular, it considers the
reliability and uncertainty of climate projections and not only changes
of the meteorological variables themselves, but also their impacts on
closely related natural systems. The paper is organized in the following
way: Section 2 briefly introduces and discusses the climate projections
used in this review. Section 3 summarizes the broad characteristics of
expected 21st century Alpine climate change and its temporal and spa-
tial variability. Projected changes of extreme precipitation and temper-
ature are presented in Section 4 and further aspects including floods,
droughts, snow, and natural hazards are addressed in Section 5.Finally,
a summary and discussion is given in Section 6.
2. Climate simulations and their uncertainty in the Alpine region
Future projections of regional climate are subject to different sources
of uncertainty stemming from the natural variability of the climate sys-
tem, unknown future GHG emissions, and errors and simplifications in
global climate models (GCMs) and RCMs or statistical downscaling
methods. The resulting uncertainties are often assessed by analyzing
ensembles of climate simulations, which sample the various sources of
uncertainty. In most parts of this review, future climate change in the
Alpine region and its reliability are assessed using the currently most
comprehensive ensemble of RCM projections for Europe from the
EU FP6 Integrated Project ENSEMBLES (van der Linden and Mitchell,
2009), which is based on the A1B emission scenario (Nakicenovic
et al., 2000) and covers the entire 21st century. Further sources of in-
formation include the PRUDENCE RCM projections (Christensen and
Christensen, 2007), which refer to the period 2071–2100 and are
basedontheA2andB2emissionscenarios(Nakicenovic et al.,
2000), and statistically downscaled regional scenarios from the global
CMIP3 multi-model dataset (Meehl et al., 2007).
Due to its dominant role in this review, som e aspects of the
ENSEMBLES multi-model d ataset are discussed in the following.
ENSEMBLES consists of 22 high resol ution R CM simulations until
mid of the 21st century (2050), 15 of them ranging until the end of
the century (2100). The simulations are available on a 25 km grid.
Fig. 1. Topography and location of the European Alps (based on the GTOPO30 digital elevation model, United States Geological Survey).
2 A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
The ensemble has been constructed by applying 17 different RCMs
driven by la teral boundary conditions from 8 different G CMs, and
theref ore mainly add resses model uncertainty. Uncertainty due to
natural variability is implicitly considered by using different GCMs.
Concerning future GHG emissions, only the A1B emission scenario
is used. Hence, three of the four major un certainty components are
at least roughly covered. Although only a small fraction of the possi-
ble GCM–RCM combinations could be realized and by far not all
available GCM simulations c ould be inclu ded due to computational
constraints, Heinrich et al. (2013a) could demons tra te that the
ENSEMBLES multi-model dataset does not underestimate GCM un-
certainty, compared to the much larger CMIP3 multi-model ensem-
ble. A rough est imate to which extent the overall uncertainty is
underestimated by using only one emission scenario can be obtained
from Prein et al. (2011), who showed that over Europe emission sce-
nario uncertainty is small in the first half of the 21st centu ry (b 10% of
the total uncertainty), but becomes considerable towards the end of
the century (30–40% of the total uncertainty of temperature). This
can be exemplified by the temperature evolution over Europe,
which only differs slightly between the A1B and the other two
major scenarios from the IPCC AR4 report (Solomon et al., 2007),
B1 and A2 before 2050. For the late 21st century, however, the tem-
per ature induced by the A1B s cenario differs remarkably from the
other emission scenarios, being about 1 °C wa rme r than B1 and
about 1 °C cooler than A2 (Fig. 2). This can be ex pected to be valid
also for the Alpine region and to affect also other meteorological
elements than temperature. For the interpreta tion of the results of
this review, it has to be k ept in mind that in many parts only the
A1B scenario has been analyzed.
Based on ensembles of climate simulations, the best estimate of
various aspects of expected climate change in this review is mostly
expressed by the multi-model mean or median. In addition, the uncer-
tainty or reliability is expressed by either comparing the results of sev-
eral single simulations or by displaying percentiles of the multi-model
ensemble spread. In some cases also more qualitative measures of
uncertainty are used.
3. The major patterns of climate change in the Alpine region
During the past decades the Alpine climate has been subject to
pronounced decadal-scale variability, but also to distinctive long-term
trends consistent with the global climate response to increasing GHG
concentrations. From the late 19th century until the end of the 20th
century Alpine temperatures have risen at a rate about twice as large
as the northern-hemispheric average, amounting to a total annual
mean temperature increase of about 2 °C (Auer et al., 2007). This
observed warming was comparatively homogeneous over the Alpine
region and was particularl y pronounced from 1980 onwards with
annual mean warming rates of about 0.5 °C per decade (EEA, 2009),
mainly caused by water va por-enhanced greenhouse warming
(Philipona, 2013). Past changes of total precipitation, in contrast,
considerab ly depend on the region, period, a nd season considered.
For Switzerla nd, Sc hmidli and F rei (2 005) and Widmann and Schär
(1997) show that mean precipitation has particularly increased in
the 20th century in fal l and winter. However, the analysis of
more recent data (1961–2012) confirms this trend only in northern
Switzerlan d in autumn (www.meteoswiss.ch ). In terms of spatial
patterns of annua l mean precipitation, Brunetti et al. (2006) show
that the north-western parts experienced slight precipitation increases
during the 20th century (mainly due to positive trends in winter and
spring), while the south-eastern Alps have been subject to a significant
drying (mainly caused by pronounced negative trends in autumn).
A more comprehensive overview of observed climate change in differ-
ent meteorological variables is given by Brunetti et al. (2009).Inthefol-
lowing Sections 3.1 and 3.2 the major patterns of expected monthly and
seasonal mean climate change in the Alpine region for various meteoro-
logical variables are described, mainly based on the ENSEMBLES multi
model dataset (Section 2) and the analysis of Heinrich et al. (2013b).
3.1. Spatial patterns of change
Fig. 3 depicts the spatial patterns of seasonal mean temperature (T),
precipitation (P), global radiation (G), relative humidity (RH), and wind
speed (WS) change in the Alpine region until the mid (2021–2050) and
the end of the 21st century (2069–2098) relative to the reference period
1961–1990 in summer and winter, expressed as multi model mean
change.
The projected changes of 2 m air temperature are positive for the
entire Alpine region in both time horizons and all seasons (spring and
autumn not shown). In summer, stronger warming in the southern
Alpine region and along the Western Alpine ridge is indicated. In winter,
regions south of the Alps show more moderate warming than the rest.
The spatially averaged warming is seasonally varying between + 1.2 °C
in spring and +1.6 °C in summer and winter until the mid of the 21st
century and +2.7 °C in spring and +3.8 °C in summer until the end
of the 21st century. As annual average, 1.5 °C warming (0.25 °C per
decade)isexpectedinthefirst half of the 21st century. Until the end
of the cen tury, warming is expected to accelerate and to amount
3.3 °C (0.36 °C per decade if only the second half of the 21st century
is considered). The warming sign al is very robust, which is indicated
by the fac t t hat all models agree on the sign of chan ge (Heinrich et al.,
2013b), but the amount of warming varies by about 3 °C betwee n the
lower and the higher estimates at the en d of the century .
Precipitation change patterns indicate less precipitation in summer,
particularly south of the Alps, and more precipitation in winter at the
end of the 21st century. A distinct impact of the Alpine ridge on the spa-
tial pattern is present in spring and autumn, although the area average
change is low due to a compensation of increases in the north and de-
creases in the south (not shown, Heinrich et al., 2013b). In numbers,
the spatial average change is seasonally varying between − 4.1% in sum-
mer and +3.6% in winter until the mid of the 21st century and − 20.4%
in summer and +10.4% in winter until the end of the 21st century.
However, the sign of change is highly diverse among the models. The
largest accordance is obtained for the decrease in summer precipitation
at the end of the 21st century, where 89% of the models agree in sign. In
a larger European context, the described pattern of change in the Alpine
region is part of a roughly dipolar north–south pattern with reduced
precipitation in Southern Europe in summer and increased precipitation
Fig. 2. Temperature evolution over Europe based on the CMIP3 simulations driven by the
emission scenarios A2 (red), A1B (green) und B1 (blue). The bold colored lines depict the
multi-model mean for each scenario; the shadings indicated the standard deviation.
Adoptedfrom Prein et al. (2011). (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)
3A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
in Northern Europe in winter and a transition zone which is shifting
northwards in summer and southwards in winter. This pattern is re-
ferred to as the European Climate change Oscillation (ECO; Giorgi and
Coppola, 2007) and the Alpine region is located at the transition zone
between increasing and decreasing precipitation, which is character-
ized by large unce rtainties of the proje cted precipitation changes
(Heinrich et al., 2013b).
Areas with increased and decreased global radiation largely cor-
respond to the opposite areas of precipitation cha nge. This is plaus i-
ble as the precipitation producing lower clouds reflec t the incoming
solar radiation. The spatial average change of global radiation is sea-
sonally varying between +0.4 W/m
2
in summer and − 1.2 W/m
2
in
winter until the mid of the 21st century and +3. 4 W/m
2
in summer
and − 4.0 W/m
2
in winter until the end of the 21st century. The spa-
tial patt ern also reveals a pronounced decrease until the end of the
21st century along the Alpine ridge in spring. Although uncertainty
is high in most regions and s easons, most RCMs agree on the de-
crease of global radiation along the Alpine ridge, particularly in
winter, and on increasing global radiation in summer at the end of
the century (Heinrich et al., 2013b).
The chan ge patterns of relative humidity are als o related to pre-
cipitation change, with incr eased humidity in regions with precipita-
tion increases and vice versa. This is again plausible, since it can be
related to the soil mo isture–atmosphere feedback. More precipita-
tion leads to wetter soils, which in turn increases mois ture flux due
to evapotranspiration into the atmosphere. This results in increased
humidity and potentially to cloud formation and precipitat ion, creat-
ing a positive feedback loop. In contrast, dry soils increase the sensi-
ble heat flux, resulting in a warmer, drier, and deeper boundary layer
which potentially inhibits convection and cloud formation (e.g.,
Alexander, 2011). In numbers, the spatial mean change in relative hu-
midity is seasonally varying between − 0.5% in winter and − 1.4% in
summer until the mid of the 21st century and 0.5% in winter and −
3.9% in summer until the end of the 21st century.
The projected changes of mean wind speed are close to zero in the
area average. Some decreases are found along the Alpine ridge and for
Fig. 3. Spatial pattern of expected seasonal mean change in the Alpine region for temperature (T), precipitation (P), global radiation (G), relative humidity (RH), and wind speed (WS)
relative to the reference period 1961–1990 in summer and winter. Left columns: 2021–2050, right columns: 2069–2098.
4 A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
the northern parts of the Mediterranean and the Adriatic Sea, particu-
larly in summer and autumn at the end of the 21st century. High accor-
dance between the RCM projections is only found in autumn with
decreasing wind speed in the southern parts until the end of the 21st
century (not shown, Heinrich et al., 2013b).
3.2. Annual cycle of change
Fig. 4 depicts the annual cycle of the spatially averaged monthly
mean change of T, P, G, RH, and WS in the Alpine region until the mid
(2021–2050) and the end of the 21st century (2069–2098) relative to
0
1
2
3
4
5
6
7
T [K]
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
−60
−50
−40
−30
−20
−10
0
10
20
30
40
P [%]
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
−25
−20
−15
−10
−5
0
5
10
15
20
25
30
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
−12
−10
−8
−6
−4
−2
0
2
4
6
RH [%]
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
−0.5
−0.4
−0.3
−0.2
−0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
2021-2050
2069-2098
0
1
2
3
4
5
6
7
−60
−50
−40
−30
−20
−10
0
10
20
30
40
−25
−20
−15
−10
−5
0
5
10
15
20
25
30
−12
−10
−8
−6
−4
−2
0
2
4
6
−0.5
−0.4
−0.3
−0.2
−0.1
0.0
0.1
0.2
0.3
0.4
0.5
0.6
G [W m^−2]
WS [m s^−1]
Fig. 4. Annual cycle o f expect ed monthl y mean chan ge in the Alpine region of te mperatur e (T), pr ecipitat ion (P), glo bal radiat ion (G), relat ive humi dity (RH ),andwindspeed
(WS) relative to the reference period 1961–1990. Left column: 2021–2050, right column: 2069–2098. The blu e line indicates the median, the grey area the 10–90 percentile
range. (For interpretation of the references to color in this figure legen d, the read er is referred to the we b version of this art icle.)
5A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
the reference period 1961–1990 in summer and winter, expressed as
multi model median (bold line) and the 10 to 90 pe rcentile-r ange
(grey areas). In these results, the missing RCM projections in th e
ENSEMBLES GCM–RCM simulation matrix have been statistically
reconstructed as described by Heinrich et al. (2013b).
Fig. 4 indicates stronger warming in late summer and winter than in
the rest of the year, which is more pronounced at the end of the 21st
century. The median warming of about 1.5 °C until the mid-century
and 3.3 °C until the end-century is associated with a bandwidth of
about 2 °C and 3 °C between the 10 and 90 percentiles, respectively.
With regard to precipitation changes, the annual cycles indicate in-
creases in winter and decreases in summer. This is hardly detectable
until the mid-21st century, but becomes a quite clear signal at the end
of the century. However, uncertainty is much larger than for tempera-
ture and only at the end of the century in a few months 85%–90% of
the models agree in the sign of change (January and February with in-
creases, June and August with decreases). The annual cycles of global ra-
diation and relative humidity change follow the precipitation change as
already discussed in Section 3.1. One remarkable feature of the
projected decrease of relative humidity in summer is its larger robust-
ness compared to precipitation change, indicated by the entire 10–90
percentile range being clearly negative. With regard to wind speed,
only a minor median decrease between 0.0 ms
− 1
and 0.2 ms
− 1
is indi-
cated, particularly in winter in the late 21st century.
3.3. Altitude gradients of change
Climatic changes can generally be expected to vary in all three
dimensions of space, i.e. vertical dependencies may arise in addition to
horizontal patterns of change. In this respect, one has to distinguish
between gradients in the free troposphere and the dependency of
near-surface climate change on the elevation of a specific site or
model grid cell. Assessing and understanding the latter type is of partic-
ular importance for impact assessments in Alpine terrain as estimates of
near-surface changes of meteorological variables are often used to drive
impact models of different kinds. Given the considerable temporal stor-
age of water in form of snow and ice in the higher regions of the Alps,
changes in high-elevation climate are of particular interest for research
on climate-related hydrological impacts in Alpine catchments.
Observational evidence suggests that in particular near-surface tem-
perature trends can considerably depend on elevation (e.g., Beniston
et al., 1997; Beniston and Rebetez, 1996; Diaz and Bradley, 1997;
Seidel and Free, 2003), with higher rates of warming often found at
high elevations. This rule of thumb, however, is not always true and de-
pends on the region and the period under consideration (see Rangwala
and Miller (2012) for a comprehensive review or, e.g., Böhm et al.
(2001)). The picture is even less coherent for precipitation changes.
The reasons for elevation-dependent temperature trends are manifold
and include changes in large scale atmospheric circulation (e.g., Ceppi
et al., 2012) as well as elevation-dependent changes in the surface ener-
gy balance induced by, e.g., snow cover changes (Kotlarski et al., 2012;
Scherrer et al., 2012) or changes in downward radia tion fluxes follow-
ing changes in atmospheric transmissivity (Marty et al., 2002;
Philipona, 2013). High-elevation sites can b e assumed to be par tly
decoupled from boundary layer processes, i.e. more strongly affected
by conditions in the free troposphere and less by local factors such as
air po llution or nea r-surface temperature inversions. For the Swiss
Alps and the period 1959–2008, Ceppi et al. (2012) identified in
their recen t study anomalously-strong warming at lo w elevations
in autumn and early winter and above-average spring temperature
trends at elevations close to the snow line. The latter can partly be at-
tributed to declining snow cover and an amplification of the general
warming by the snow albedo feedback (Scherrer et al., 2012). Several
pre vious studies applying global and regional climate models have
also confirmed the impo rtance of the snow albedo feedback for fu-
ture temperature changes in the Alps and other mountain regions
(Fyfe and Flato, 1999; Giorgi et al., 1997; Im et al., 2010). Recently,
Kotlarski et al. (2012) investigated the elevation dependency of
21st century near-surface cl imate change over Europe based on a
high-resolution climate change scenario carried out with the RCM
COSMO-CLM (Rockel et al., 2008). For the Alps, they found strong ev-
idence of an amplification of projected 21st centu ry warming at high
elevations, again presumably connected to the snow albedo feed-
back. Also summ er precipita tion changes were found to considerably
depend on ele vation, with strongest relative drying signals in the
lowlands.
We here present an extension of the study of Kotlarski et al. (2012),
taking into account ten regional climate simulations. They were carried
out by eight different RCMs which, in turn, were driven by six different
GCMs at their lateral boundaries. All experiments are provided by the
ENSEMBLES database at a horizontal resolution of about 25 km and
are based on the SRES A1B emission scenario (see Section 2). Fig. 5
shows the vertical profile of changes in 2 m temperature (a), precipita-
tion (b) and number of snow days (c) in the Alps until the end of the
21st century. Note that only elevations below about 2700 m are covered
by the model topographies and no information can be deduced for
higher elevations. This analysis confirms the results of Kotlarski et al.
(2012) and shows for most parts of the year an anomalous warming
at higher elevations, but also an amplified low-elevation warming in
summer (left row). In spring and partly also in summer and autumn,
the former effect can be re lated to a reduction of snow cover
(right row) and the snow-albedo effect. For precipitation (middle row)
no clear systematic altitude dependence of the climate change signal
can be found, except for a tendency towards a reduced summer drying
and a reduced winter moistening (in relative terms) at high altitudes
compared to low-lying regions. The 10-model ensemble reveals an aston-
ishing inter-model agreement on the shape of the vertical profiles, despite
consider able differences in the overall rate of change . For temperature
and precipitation, the latter is obviously controlled by the driving GCM
and also depends on the season.
4. Changes in precipitation and temperature extremes
4.1. Precipitation extremes
Heavy precipitation events possess the potential to cause natural
disasters and serious damage to infrastructure facilities. Subse-
quently suc h events can imply vast societal, economic and environ-
mental impact . In this respect and with anticip ated climate cha nge,
there is particular interest in the future behavio r of precipitation
extremes.
TheEuropeanAlpsarearegionofmajorconcernbyreasonoffre-
quent affliction by heavy precipitation events (e.g. the events in August
2005 in Switzerland; MeteoSchweiz, 2006). This is primarily due to
orographic mechanisms that extract ambient atmospheric moisture.
Also, the Alps are influenced by both Atlantic and Mediterranean cli-
matic regimes including events of stratiform and short-lived convective
nature (Frei and Schär, 1998). For this reason they feature a large spatial
variability of climatic regimes.
Assessing changes in heavy precipitation events and the hydrologi-
cal cycle in general contains the interplay of several complex processes
(Allen and Ingram, 2002; Emori and Brown, 2005; Held and Soden,
2006; O'Gorman and Schneider, 2009). These include thermodynamic
processes that can lead to an intensification of precipitation. The most
important is an increased moisture uptake capacity of air under warmer
conditions, but also possible changes in atmospheric stratification
(Christensen and Christensen, 2003; Frei et al., 1998; Pall et al., 2007).
Moreover, dynamic effects like changes in atmospheric circulation can
importantly contribute to an altered frequency of heavy precipitation
events.
Several observational studies have investigated changes in heavy
precipitation at global to regional scale (Solomon et al., 2007). For the
6 A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
Alpine region, Schmidli and Frei (2005), show that precipitation in
Switzerland has intensified in the 20th century. Furthermore, Schmidli
and Frei (2005) and Widmann and Schär (1997) show that mean pre-
cipitation has particularly increased in fall and winter. However, long-
term reconstructions of precipitation (Pauling et al., 2006; Pauling and
Paeth, 2007; Casty et al., 2005)andflood-records (Schmocker-Fackel
and Naef, 2010a,b) for the European Alpine region show large decadal
variations in the frequency of heavy events and floods. Therefore, recent
changes might be explained by natural climate variability. In addition,
the small scales of heavy precipitation events tend to suffer from
greater uncertainty than is the case for regional atmospheric pat-
terns, especially when projecting their future behavior.
In a recent study Rajczak et al. (2013) have assessed projected
changes in Alpine precipita tion and its extremes in lar ge detail.
The study used a set of 10 regional climate simulations from
the ENSEMBLES multi model data set at a resolution of 25 km
(see Section 2). Based on this study, Fig. 6 presents and overvi ew
on projected 21st century changes (2070–99 compared to 1970–99)
for impact-relevant precipitation indices at the seasonal scale for the
Alpine region. The Figure illustrates, that the frequency of wet days
(left column) is projected to substantially decrease in summer across
the entire Alpine region, whereas in fall and spring substantial reduc-
tions are only projected for southern Alpine regions. In winter, no
clear changes in precipitation frequency are obvious. However, for
some southern Alpine areas projections suggest an increased number
of wet days, which could be due to changes in atmospheric circulation.
Projections of mean precipitation (middle column) show overall
increases in winter and dec reasing signals in summer, as already
pre sented in Section 3.1. Model agreement on changes is more
obv ious in Alpine foreland regions than in central Alpine areas.
Especially in winter and also for extreme diagnostics (right column),
signals are smaller and afflicted with model uncertainty for inner-
alpine regions. In spring and fall, the Alpine ridge presents a part of a
distinct transition zone separating increasing signals in the north from
decreasing signals in the south of Europe.
Considering extreme precipitation events, Fig. 6 (right column) pre-
sents projected changes in the 5-year return value of daily precipitation
events as estimated by generalized extreme value theory. The Alpine re-
gion is expected to experience an increase in the intensity of extreme
pre cipitation events in all seasons and for most regions. This is equiv-
alent to a reduction (increase) of return periods (values) under fu-
ture climatic conditions with respect to present-day conditions.
Exceptional from th is gen eral behavior are only summer-time events
in the southern Alpine region, where return peri ods associated with
small return periods tend to decrease. The most substantial and
widespread intensifications are projected to occur in fall and in the
northern Alpine region, where changes in intensity amount up to
+30%, representing more than a halving of return periods (not shown,
Rajczak et al., 2013).
It is obvious that changes in m ean (frequency) and heavy precip-
itation do not scale proportionate. Depending on season and loca-
tion, projected chan ges are even oppo sitiona l, highlighting not only
a
0 2 4 6 8
0
500
1000
1500
2000
2500
Temperature change [°C]
Elevation [m]
Elevation [m]
Elevation [m]
Elevation [m]
FJD
MAM
JJA
NOS
Ensemble mean
ARPEGE-driven BCM-driven HadCM3Q0-driven
HadCM3Q3-driven HadCM3Q16-driven ECHAM-driven
Precipitation change [%]
b
-80 -60 -40 -20 0
0
500
1000
1500
2000
2500
Snow day change [days season
-1
]
c
02468
0
500
1000
1500
2000
2500
02468
0
500
1000
1500
2000
2500
02468
0
500
1000
1500
2000
2500
-40 20 0 20 40
0
500
1000
1500
2000
2500
-40 20 0 20 40
0
500
1000
1500
2000
2500
-40 20 0 20 40
0
500
1000
1500
2000
2500
-40 20 0 20 40
0
500
1000
1500
2000
2500
-80 -60 -40 -20 0
0
500
1000
1500
2000
2500
-80 -60 -40 -20 0
0
500
1000
1500
2000
2500
-80 -60 -40 -20 0
0
500
1000
1500
2000
2500
Fig. 5. Elevation dependency of mean seasonal changes in (a) 2 m temperature (b) precipitation and (c) number of snow days in the Alps (based on 100 m elevation bins). For the latter,
a snow day threshold of 3 mm we has been applied. Results are based on ten GCM–RCM chains of the ENSEMBLES project, all of them assuming the SRES A1B emission scenario. Changes
refer to the period 2070–2099 with respect to the reference period 1961–1990. The color indicates the driving GCM. The Alpine domain is defined according the Steger et al. (2013).
7A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
the complexity of the hydrological response to climatic change, but
also the increasing probability of both more intense precipitation
and drought conditions.
4.2. Temperature extremes
Anomalously warm months or seasons in the last 1–2 decades of the
observational record have prompted many publications to explore and
explain these events (e.g., Beniston, 2004, 2005; Schär et al., 2004;
Luterbacher et al., 2007). For example, the links between soil–moisture
deficits and heat waves have recently attracted much interest as an
explanatory causal mechanism (e.g., Seneviratne et al., 2006). Vautard
et al. (2007) have shown that the northward spread of droughts
that originate in the Mediterranean during winter yield preconditions
capable of triggering intense and persistent heat waves in Europe.
Such conditions are likely to be amplified in the future, since higher
mean temperatures facilitate the exceedance of thresholds considered
to be extreme (e.g., taking the 90% quantile as the threshold of extreme
temperature).
While it is difficult to use RCM results to directly quantify changes in
extremes, their statistics can nonetheless help identifying the possible
change in frequency of anomalously hot seasons in the future. For ex-
ample, Beniston (2007b) showed that RCM outputs from the EU FP5
project PRUDENCE (see Section 2) can be used to define the “envelope”
of quantiles around the mean monthly temperatures of a future climate.
An example is provided in Fig. 7, where the mean monthly course of
daily maximum temperature in a scenario climate is plotted for Basel
(north–west Switzerland) in a high-emissions scenario (A2). The RCM
ensemble statistics also serve to define the quantile boundaries around
the multi-model mean. For the hottest months of July and August, the
mean Tmax is close to 31 °C (i.e., about 6 °C more than today), while
the 90% quantile is around 39 °C. In order to compare the future climate
statistics with those that have been observed in a recent past, the 2003
monthly statistics and summer Tmax have been added. The observed
statistics show that Tmax in June and August was not only well beyond
Fig. 6. Projected changes in wet day frequency [days N 1 mm], mean precipitation, and the 5-year return value of 1-day precipitation events (left to right) for the four climatological
seasons (top to bottom) for the European Alpine region. Colored contours show the median change signals from a 10-member multi-model ensemble and are expressed as the percentage
change for period 2070–2099 with respect to period 1970–1999. Hatching denotes agreement in the sign of change in 90% of the considered models. The results are based on a set of 10
ENSEMBLES RCMs at a resolution of 25 km. Bold lines indicate the 700 m-isoline as presented by the E-OBS topography (Haylock et al., 2008). Figure and results are adapted from Rajczak
et al. (2013).
8 A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
the norm of the reference period 1961–1990, but that it was slightly
higher than the average Tmax of the future climate at the end of the
21st century. Similarly for the daily course of Tmax during the
2003 summer, the observed statistics show that some days of the
heat-wave were close to or even in excess of the 90% quantile of
thefutureclimate.
By analyzing these statistics more closely, it is possible to define
what would be the frequen cy of occurrence of observed record-
breaking seasons in the future. Schär et al. (2004) sugg ested that
one summer in two by 2100 would be at least as hot as the 2003 sum-
mer. The statistics exa mined by Beniston (2007b) not only confirm
this finding, but also show that 6 winters out of 10 would be as hot
as the record-breaking winter 2006–2007, 7 springs in 10 as hot
as the record spring of 2007, and 6 autumns i n 10 as hot as the
2006 autumn.
5. Further aspects
5.1. Floods
Major floods have been rather s carce in the Alps over much of the
20th century before several catastrophic floods occurred during the
last three decades, with associated substantial increa ses in flood-
related losses (Bezzola and Ruf, 2009).InthecaseofSwitzerland,
16 major fl oods have been recorded since the early 19th century,
with nine of these during the last 30 years. Over the course of the
past 500 years, catchments in northern Switzerland were affected
by marked fl uctuations in flood frequencies, with periods of frequent
flooding AD 1560–1590, 1740–1790, 1820–1940, and since the 1970s
(Schmocker-Fackel and Naef, 2010a). During several of these periods,
debris flows have been more frequent in the Valais Alps as well
(Stoffel and Beniston, 2006; Bollschweiler and Stoffel, 2010;see
Section 5.4). Although there is reason to believe that climate change
will likely lead to more and more severe floods in the Alps, the cur-
rent increase in flood frequency remains, for the time being, compa-
rable to past periods of increased flood frequency mentioned a bove
(Schmocker-Fackel and Naef, 2010a). Allamano et al. (2009) detected
an increase in flood peaks in Swiss rivers over the course of the last cen-
tury, possibly influenced by the large events over the past forty years,
and have ascribed this increase to increasing temperature and precipi-
tation (Allamano et al., 2009). In other regions of the Alps, however,
trend detection is not feasible due to short observational records and
because extreme events are rare per definition (IPCC, 2012
). At the
same time, apparent changes in flood frequency and magnitude have
also been aggravated by the concreting of river reaches and the water-
proofing of settled areas. At almost 2.5 billion Euros, the August 2005
flood (MeteoSchweiz, 2006) represents the most extensive financial
loss for Switzerland ever caused by a single natural disaster, despite
the fact that several floods of the 19th century actually matched or
even exceeded the extent of damage of the 2005 event (FOEN, 2007).
The occurrence of heavy or extended precipitation events will likely
increase in a future greenhouse climate (see Section 4.1)andpossibly
cause more frequent severe flooding events in Europe (e.g., Christensen
and Christensen, 2003), despite the general drying of future summers
(Section 3.1). It thus seems possible that the size and frequency of
winter and spring floods will increase, in particular north of the Alps
and at altitudes up to 1500 m above sea level (KOHS, 2007). According
to Allamano et al. (2009), and assuming a 2 °C increase and a 10%
increase in precipitation intensity, the return period of a current-day
100-year winter flood could be reduced to a 20-year event. Summer
floods are by contrast expected to occur less frequently in the future
(KOHS, 2007). South of the Alps, floods are predicted to become more
severe in all seasons except for summer (OcCC, 2007).
5.2. Droughts
Drought can be regarded as a natural recurrent phenomenon
which occurs on a variety of d ifferent temporal and spatial sca les
and si gn ificantly affe cts natural and socio-e conomic systems. The
vulnerability of the Alpine region to drought was clearly revealed
in 2003 with, e.g., large scale losses in agriculture and forestry, low-
ering of the ground water level, shortages in the generation of hydro-
power electricity, and pronounced snow and glacier-melt leading to
increased rock and ice falls in the mountains (see Section 5.4; Gr uber
et al., 2004; Jolly et al., 2005; Fischer et al., 2007; Garcia-Herrera
et al., 2010).
During the 20th century, evidence for increasing drought risk was
found especially in central, eastern, and southern Europe (e.g., Szinell
et al., 1998; Lloyd-Hughes and Saunders, 2002; Bonaccorso et al.,
2003; Dai et al., 2004; Trnka et al., 2009; Briffa et al., 2009). For the
Alpine region, van der Schrier et al. (2007) investigated monthly mois-
ture variability for the period of 1800–2003 based on a simple drought
index and found that the late 1850s into the 1870s and the 1940s to
the early 1950s stand out as persistent and exceptionally dry periods.
The driest summers on record, in terms of the amplitude of the index
averaged over the Alpine region, are 1865 and 2003. For Switzerland,
Rebetez (1999) investigated the change in frequency of drought epi-
sodes based on precipitation and detected increased drought frequency
and persistence during the 20th century in the southern parts.
Future changes in drought conditions are primarily driven by altered
precipitation regimes along with increased evapotranspiration related
to higher temp eratures and increased water demand (Briffa et al.,
2009; Sheffield and Wood, 2008
). Based on the a nalysis of GCMs,
Giorgi (2006) detected the Mediterranean and n orth-ea stern parts
of Europ e as most responsive regions to clim ate change by the end
of the 21st century worldwide and the high vulnerability of southern
European regions to future drought regimes was underpinned by vari-
ous other studies (e.g., Lehner et al., 2006; Burke and Brown, 2008;
Sheffield and Wood, 2008; Warren et al., 2009). For Switzerland,
changes in hydrological drought characteristics until the mid and
the end of the 21st centu ry are available from the EU FP7 project
DROUGHT R&SPI (Alderlies te and van Lanen, 2013). In this project,
a set of three statistically downscaled and bias corrected GCM pro-
jections forced by the SRES A2 and B1 emission scenarios are used
to drive a set of differen t Global Hydrological Models (GHMs) and
Land Surface Models (LHMs). As expected, the changes are generally
larger for the A2 emission scenario. However, the change in number
of drought events doesn't show a c lear pattern among the emission
scenarios and future time periods. Until the end of the 21st century
(2071–2100), the expected multi-model mean c hanges fo r the B1
(A2 ) scenario are: + 4.7% (+ 4.7%) for the number of drought
events, +89.9% (+143.1%) for the average duration, +375.2%
Fig. 7. Future (2071–2100) monthly-mean daily maximum temperatures (Tmax) for
Basel, Switzerland (mean and quantiles of the RCM ensemble). The monthly and
summer daily Tmax for 2003 are also in clude d (red lines) (Adapted from Beniston,
2007b). (For interpretation of the references to color in this figure legend, the reader is re-
ferred to the web version of this article.)
9A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
(+467.4%) for the average deficit volume, and +70.6% (+67.1%) for
the average intensity.
Calanca (2007) investigated the occurrence and severity of droughts
for the Alpine region until the end of the 21st century based on results
from a RCM. In this study, a simple soil water balance model was driven
by a single RCM simulation from the PRUDENCE project forced by the
SRES A2 emission scenario. The conclusions are that there are clear
indications that in the future the Alpine region will increasingly suffer
from droughts. The RCM projects a decrease in the frequency of wet
days of about 20% with respect to the growing season of summer
crops (April to September) which consequently results in an increase
of drought frequency from about 15% to more than 50%. Furthermore,
the results indicate an overall shift of the distribution towards higher
severity of drought events.
In order to account for the uncertainties in RCM projections,
Heinrich and Gobiet (2012) used a set of eight downscaled and error-
corrected RCM projections from ENSEMBLES forced by the SRES A1B
emission scenario and calculated a set of commonly used drought indi-
ces for nine European subregions until the mid of the 21st century
(2021–2050). The study revealed that there is large structural uncer-
tainty among the different drought indices and climate scenarios for
the Alpine region. Statistical significant changes in drought characteris-
tics are only obtained for indices which account for the effect of increas-
ing air temperature. The self-calibrated Palmer Drought Severity Index
(scPDSI), which is based on simplified soil water balance model,
shows an increase of +7.1% for the average length, +24.7% for the
average magnitude, and +4.1% for the average area of dry events with
respect to the baseline period of 1961–1990.
In summary, a rather systematic picture of increasing drought char-
acteristics is projected for the near- as well as long-term future for the
Alpine region. The exact numbers of the changes are subject to consid-
erable uncertainties which can be related to the different climate sce-
narios applied in the various studies and structural uncertainty of the
models which are used for the quantification and definition of drought.
Although the projected changes in the Alpine region are more dissonant
and less reliable than for, e.g., Southern Europe, they still indicate in-
creasing drought stress and call for flexible and adaptable water man-
agement strategies.
5.3. Snow
Snow in a populated and economically-diverse region such as the
Alps plays an important role in both natural environmental systems,
(e.g., hydrology and vegetation), and a range of socio-economic sectors
(e.g., tourism or hydropower). Shifts in snow amount and duration as a
result of a changing climate are likely to impact upon these systems in
variousways(Beniston, 2012). The behavior of the snowpack is ob-
viously related to geographic characteristics, in particular altitude,
orientation, expo sure to dominant atmospheric flows, and location
at the bottom of a valley floor (possibly subject to t emperature
inversions), slopes or mountain tops.
The 20th century has already seen significant changes in snow
amo unt and duration, which generally exhibit a large degree of
interannual and inter-decadal variability. The observational record
shows periods of snow-abundant winters (e.g., in the 1960s) and
snow-sparse seasons as experienced from the latter part of the
1980s to the mid-1990s for example. In some instances, particularly
snow-sparse winters seem to be related to the positive (or warm)
phase of the North Atlantic Oscillation (NAO; e.g., Beniston, 1997),
but the NAO is by no means the only explanat ory facto r that explains
the variability of snow in the Alps. For example, Scherrer and
Appenzeller (2006) suggest that half the variability of Alpine snow
cover is related to the establishment of blocking pat terns over
Eur ope, not always related to influence of the NAO. For the Swi ss
Alps, Marty (2008) has ident ified what appears to be a regime shift
in snow, i.e., a stepwise decline in snow amount and duration in
the 1980s, with no defi
nite trend occurring since then.
As a “rule of thumb”, the average level of the snowline rises by
roughly 150 m per degree Celsius. With regional climate model projec-
tions suggesting wintertime increases of temperature of 2 °C–4°Cin
the latter part of the 21st century (e.g., CH2011, 2011), this implies an
upward snowline shift by 300–600 m. However, it has to be noted
that this simple concept might overestimate the rise of the snow line,
as it doesn't regard the effect of temperature inversions and cooling by
melting precipitation (Unterstrasser and Zaengl, 2006). While changes
in precipitation patterns are also likely to influence the abundance and
geographic distribution of snow, a number of studies have emphasized
the fact that in a warmer climate, temperature is likely to be the domi-
nant control on snow cover, and the wintertime precipitation increases
that most models project for the Alps (Section 3.2) will not compensate
the large los ses in snow volume that more elevated temperatures
will induce.
Because of their relatively coarse grids, regional climate models have
in the past not proven sufficient to provide detailed information on
snow precipitation and snow amount, although Steger et al. (2013)
have shown that current generations of RCMs now capture the spatial
and seasonal snow variability, but with over- or under-estimations of
quantities according to altitude. Assessing the behavior of snow in com-
plex topography can also involve interfacing techniques that enables to
estimate of snow depth and duration at a very local scale. Statistical
downscaling techn iques can be applied, but also physicall y-based
snow models that use RCM-generated outputs as initial and bound-
ary conditions for snow a nd surface ener gy-balance models at the
very local scales (e.g., Martin et al., 1996; Uhlmann et al., 2009)
have proven to b e fairly powerful tools t o assess snow in the Alps.
Studies by Beniston et al. (2003) Uhlmann et al. (2009),orSteger
et al. (2013), among others, all agree on the likelihood of seeing large
decreases in Alpine snow amount and duration below about 1500 m el-
evation, and even above 2000 m the declines in snow amount are a fea-
ture common to the different methodologies applied to future climatic
conditions (Fig. 8). In addition, Steger et al. (2013) show that the reduc-
tion of snow cover is greatest in the spring; the snow cover thus exhibits
an asymmetric reduction within the winter season.
Beniston et al. (2011b) have att empted t o see whe the r snow-
abu ndant winters may still occur on occasion by the end of the 21st
century, based on a typology of winters using joint quantiles of tem-
per ature and precipitation. It was shown that when a combinati on of
warm and dry winter days is less frequent, more snow can fall and
remain on the grou nd. Transposing this particu lar temperature-
pre cipitation mode to the future, the authors have shown that the
Fig. 8. Snow volume under current climate and a possible future climate with winters 4 °C
warmer than today (slightly warmer than median estimate for the end of the 21st century,
see Section 3.2). The spread within the two curves indicates the variability of winters
(“snow-sparse” to “snow-abundant”). Total snow volume is computed as the average
snow depth multiplied by the surface area on which it lies, for elevation levels between
ranging from 200 m to 4500 m in Switzerland.
10 A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
number of snow-abundant winters by the end of the century would
rep resent a 1-in-30 yea r's event, compared to 8 in the reference
climate (1961–1990).
5.4. Natural hazards in the Alps
Changes in temperature and precipitation are considered to likely
have a range of secondary effects, including on the occurrence of natural
hazards in mountain environments. However, while there is theoretical
understanding for increased mass-movement activity as a result of pre-
dicted climate change, changes can hardly be detected in observational
records (Stoffel and Huggel, 2012).
5.4.1. Glacier retreat and related hazards
One of the most obvious consequences of climate change at high-
elevation sites is the widespread retreat and disintegration of glaciers
(Zemp et al., 2007; Diolaiuti et al., 2011). The consequences for natural
hazards following increasingly rapid changes in glacier geometry are
multiple and include the formation of ice-marginal lakes, ice avalanches,
and mass movements originating from the recent debuttressing of pre-
viously glacierized walls and hill slopes.
Rapid lake formation and growth that has been accelerated in recent
years is a global phenomenon but has been observed in much of the
Swiss Alps (Künzler et al., 2010; Werder et al., 2010). Several lakes
have formed within the past decade at the terminus of glaciers where
subglacial topography has been overdeepened by the glacier. Positive
feedback processes, mainly related to the thermal energy of water,
accelerated glacier melt and has been observed to result in the forma-
tion and extensive growth of proglacial lakes over periods of only a
few years (Kääb and Haeberli, 2001). Some of the lakes have become
major tourist attractions, but considerable concern exists about hazards
in case of a lake outburst that could be triggered by ice avalanches or
rockfalls following debuttressing of the steep lateral slopes (Dalban
Canassy et al., 2011).
By way of example, downwasting of the Lower Grindelwald glacier
in its terminal part resulted in a loss of between 60 and more than
80 m of ice thickness between 1985 and 2000 (Paul and Haeberli,
2008). In recent years a glacial lake started to form in the terminus
area of the glacier (Fig. 9). In 2004 and 2005, the lake had a limited vol-
ume but has subsequently grown continuously in the spring and early
summer seasons, resulting in lake volumes of 250,000 m
3
in 2006,
1.3 million m
3
in 2008 and 2.5 million m
3
in May 2009 (Werder et al.,
2010) and the occurrence of a glacier-lake outburst flood in 2008
(Worni et al., in press). The rock slope failures above the terminus of
the Lower Grindelwald glacier (Fig. 9b) are yet further textbook exam-
ples of glacier retreat, downwasting and associated debuttressing
effects on rock slope stability, and could in fact serve as a model case
for increasingly destabilized future high-mountain environments
(Stoffel and Huggel, 2012). Similar examples can be f ound in the
Mount Blanc mass if, where the Brenva and Triolet rock avalanches
(18th and 20th centuries; Deline, 2009) ha ve been considered char-
acteristic examples of rock slope instability related to glacial
oversteepening or debuttress ing.
The current rapid glacier downwasting is likely to promote many
rock slope failures at rather short future time scales, probably on t he
ord er of decades. For the future, Gl acier downwasting is expected to
result in the formation of further ice-marginal lakes and subsequent
pro blems of glacier-lake outburst floods (Frey et al., 2010; Worni
et al., 2012).
5.4.2. Permafrost thawing and mass movements
Important effects of climate change on mountain slope stability
are also related to the warming and thawing of permafrost. Permafrost
exists in many steep rock slopes in high-mountain environments
(Salzmann et al., 2007) and its degradation due to global warming can
affect slope stability. Although this link might be intuitively clear, the
mechanisms of permafrost degradation and related slope stability re-
main complex and only poorly understood (Gruber and Haeberli,
2007). A number of recent slope failures have been documented in per-
mafrost areas, and related to increasing temperatures in general or to
the heat wave of summer 2003 and the related excessive thawing of
the active layer of permafrost bodies in particular (Gruber et al., 2004;
Stoffel et al., 2005; Fischer et al., 2011; Ravanel and Deline, 2011).
Changes in sediment supply and land-use are further key deter-
minants for mass-movement frequency and magnitude. Recent ob-
servations in the Swiss Alps indicate that sediment supply can in
fact change significantly as a result of permafrost degradation of
rock and scree slope s or mass movements related to other processes
(Huggel et al., 2012). Average flow veloc ities of rock glaciers h ave in-
creased drastically in many parts of the Alps (Kääb et al., 2007; Roer
et al., 2008), probably as a result of increasing mean annual air tem-
peratures (Kääb et al., 2007). As such, warming has been reported to
exert indirect control on debris-flow magn itude and frequency
(Stoffel et al., 2011) through the delivery of larger quantities of sed-
iment into the debris-flow channels under current conditions than in
the past (Lugon and Stoffel, 2010). As a consequence , the volume of
the largest debris flows has risen by on e order of mag nitude since
the 1920s (Stoffel, 2010) and is likely to further increase with ongo-
ing p ermafrost degradation (Stoffel and Beniston, 2006).
The temporal frequency of debris flows, was, in contrast, not directly
affected by these changes, as their release depends primarily on meteo-
rological triggers such as intense rainfall in summer. Triggering
Fig. 9. (a) Glacial lake at Lower Grindelwald glacier with destabilized moraine that partly failed on 22 May 2009. The dashed line indicates the failed mass. The volume of the landslide was
about 300,000 m
3
, with 100,000 m
3
reaching the lake and generating an impact wave. A glacier lake outburst flood was recorded at the site in 2008. The volume of the lake reached
N 2.5 × 10
6
m
3
water in 2009 (Source: www.gletschersee.ch). (b) Rockslide from the Eiger resulting from the debuttressing after the retreat of the Lower Grindelwald glacier (adapted
from Stoffel and Huggel, 2012).
11A. Gobiet et al. / Science of the Total Environment xxx (2013) xxx–xxx
Please cite this article as: Gobiet A, et al, 21st century climate change in the European Alps—A review, Sci Total Environ (2013), http://dx.doi.org/
10.1016/j.scitotenv.2013.07.050
meteorological conditions have been shown to occur less frequently
under current climatic conditions as compared to those of the late
19th and early 20th centuries (Stoffel et al., 2011; Schneuwly-
Bollschweiler and Stoffel, 2012), and are not expected to increase in a
future greenhouse climate (Stoffel et al., submitted for publication). Re-
cent debris flows from other regions of the Swiss Alps tend to confirm
the recent magnitude increase and related changes in erosive power
which have proven to be sufficient to remobilize large amounts of sed-
iment on Holocene fans (Stoffel and Huggel, 2012).
Despite uncertainties, recent developments at high-elevation sites
have shown clearly that the sensitivity of mountain and hill slope
systems to climate change is likely to be acute, and that events beyo