Content uploaded by Geoffrey Klein
Author content
All content in this area was uploaded by Geoffrey Klein on Aug 30, 2018
Content may be subject to copyright.
ORIGINAL PAPER
Unchanged risk of frost exposure for subalpine and alpine plants
after snowmelt in Switzerland despite climate warming
Geoffrey Klein
1,2
&Martine Rebetez
1,2
&Christian Rixen
3
&Yann Vitasse
1,4
Received: 14 February 2018 / Revised: 17 May 2018 / Accepted: 23 June 2018 /Published online: 12 July 2018
#ISB 2018
Abstract
The length of the snow-free season is a key factor regulating plant phenology and shaping plant community composition in cold
regions. While global warming has significantly advanced the time of snowmelt and the growth period at all elevations in the
Swiss Alps, itremains unclear if it has altered the likelihood of frost risk for alpine plants. Here, we analyzed the influence of the
snowmelt timing on the risk of frost exposure for subalpine and alpine plants shortly after snowmelt, i.e., during their most
vulnerable period to frost at the beginning of their growth period. Furthermore, we tested whether recent climate warming has
changed the risk of exposure of plants to frost after snowmelt. We analyzed snow and air temperature data in the Swiss Alps using
six weather stations covering the period 1970–2016 and 77 weather stations covering the period 1998–2016, spanning elevations
from 1418 to 2950 m asl. When analyzed across all years within each station, our results showed strong negative relationships
between the time of snowmelt and the frequency and intensity of frost duringthe most vulnerable period to frost for subalpine and
alpine plants, indicating a higher frost risk damage for plantsduring years with earlier snowmelt. However, over the last 46 years,
the time of snowmelt and the last spring frost date have advanced at similar rates, so that the frequency and intensity of frost
during the vulnerable period for plants remained unchanged.
Keywords Air temperature .Alpine plants .Frost risk .Global warming .Snowmelt .Snow cover
Introduction
In mountainous regions, a significant decline of the snow cov-
er has been reported worldwide over the last decades (Park et
al. 2012; Pederson et al. 2013;Xuetal.2016), including the
European Alps (Klein et al. 2016;Marty2008;Valtand
Cianfarra 2010). Since the beginning of the 1970s in the
Swiss Alps, the shortening of the snow cover duration found
at elevations greater than 1100 m asl was mainly caused by
earlier snowmelt in spring, and, to a lower extent, by later
snow onset in autumn (Klein et al. 2016), in connection with
a stronger temperature warming in spring than in autumn
(Rebetez and Reinhard 2008; Serquet et al. 2013).
The timing of snowmelt, which greatly fluctuates from year
to year irrespective of elevation (Klein et al. 2016; Wheeler et al.
2014), is the main driver triggering the onset of growth of most
alpine plant species in spring (Gerdol et al. 2013; Inouye 2008;
Jonas et al. 2008; Sherwood et al. 2017; Vitasse et al. 2017).
The snow depth accumulated during winter was also found to
play a role in the abundance of flowers, as well as in the prob-
ability of frost damage in spring when the snow becomes too
thin to sufficiently protect overwintering plant tissues against
extreme low temperatures (Inouye et al. 2002). Conversely, a
deeper snow cover tends to delay the time of snowmelt, and
therefore shifts alpine plant growth to a warmer period of the
year with possibly fewer frost events (Jonas et al. 2008).
The beginning of the growing season for alpine plants is
primarily controlled by the timing of snowmelt and the
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s00484-018-1578-3) contains supplementary
material, which is available to authorized users.
*Geoffrey Klein
geoffrey.klein@unine.ch
1
Institute of Geography, University of Neuchatel,
Neuchatel, Switzerland
2
WSL Swiss Federal Institute for Forest, Snow and Landscape
Research, Neuchatel, Switzerland
3
WSL Institute for Snow and Avalanche Research SLF,
Davos, Switzerland
4
WSL Swiss Federal Institute for Forest, Snow and Landscape
Research, Birmensdorf, Switzerland
International Journal of Biometeorology (2018) 62:1755–1762
https://doi.org/10.1007/s00484-018-1578-3
subsequently air temperatures (Vitasse et al. 2017). On aver-
age, a duration of 2 to 3 weeks after the time of snowmelt was
observed before the beginning of plant height growth in the
Swiss Alps in snowbed conditions, corresponding roughly to
an accumulation of 100 growing degree days (abbreviated
GDD 100 thereafter) (Vitasse et al. 2017). Hence, the few days
following the time of snowmelt are critical for alpine plants, as
it is the period when plants lose progressively their freezing
resistanceacquired during winter and thus, when they become
most vulnerable to freezing damages (Rixen et al. 2012;
Sherwood et al. 2017). Nevertheless, at elevations above
2000 m asl, temperature below frost resistance of fully devel-
oped tissues of plants can still occur throughout the whole
growing season (Körner 2003).
During winter, snow cover insulates alpine plants from
freezing temperature (Körner 2003), so that the timing of
snowmelt in spring determines when plant tissues are exposed
to atmospheric air temperature and potentially with freezing
temperatures. Freezing resistance of common plant species
from the European Central Alps inducing 100% of mortality
in the plant tissues was reported to range from −4to−16 °C,
with an average freezing resistance of −9 °C for most species
(Ladinig et al. 2013; Taschler and Neuner 2004).
Species growing at sites with little snowprotection, such as
ridges, are typically more freezing resistant than species grow-
ing under snowbed conditions (Nagy and Grabherr 2009). The
high interannual variability of the time of snowmelt shown in
Klein et al. (2016) might also alter the freezing resistance of
alpine and subalpine plants, which has been linked to the
fluctuations of the snow depth during the onset of spring
(Palacio et al. 2015).
Warming air temperatures since the 1970s have consider-
ably advanced the spring onset of growth below the treeline in
Western and Central Europe (Ahas et al. 2002;Menzeletal.
2006; Vitasse et al. 2018b), but little is known concerning
alpine plants in the Alps, as, to the best of our knowledge,
no long-term series of phenological observations are available
for plants above the treeline in these regions. Besides, only a
few studies describe the climatic conditions shortly after the
snowmelt in alpine and subalpine regions (Inouye 2008;
Inouye et al. 2002; Jonas et al. 2008; Wheeler et al. 2014),
likely due to the difficulty to obtain accurate meteorological
data at such elevations.
In a warmer climate, three different scenarios for the risk of
frost exposure for alpine plants could be expected (Vitasse et
al. 2018a): (i) an increase of the risk of frost exposure due to a
faster advance of the phenology compared to the frost-free
period, (ii) a decrease of the risk of frost exposure because
of a faster advance of the frost-free period compared to phe-
nology, (iii) no change in the risk of frost exposure due to a
similar advance of both phenology and frost-free period. With
a slower warming of minimum air temperatures compared to
maximum air temperatures during spring above 800 m in the
Swiss Alps over the last five decades (Vitasse et al. 2018a)and
an earlier time of snowmelt (Klein et al. 2016), the risk of frost
exposure for subalpine and alpine plants may not necessarily
decrease during their growth period in spring in a warmer
climate. Below the treeline in the Swiss Alps, the risk of frost
exposure and potential damage for some tree species has al-
ready increased during their leaf-out and flowering period at
elevations higher than 800 m over the period 1975–2016,
despite climate warming (Vitasse et al. 2018a). Hence, in-
creasing frequency of potentially damaging freezing events
might increase frost injuries (Wheeler et al. 2014), reduce
growth (Wipf et al. 2009), or increase the mortality of sensi-
tive frost plant species, such as shown for several species in
the Rocky Mountains (Inouye 2008).
Here, we examined long-term air temperature and snow
depth measurements in the Swiss Alps at six weather stations
during the period 1970–2016, together with data from another
meteorological network of 77 weather stations covering the
period 1998–2016, both located at elevations ranging from
1418 to 2950 m asl. We focused our analysis on events with
daily freezing temperatures below −4 °C during a period of
3 weeks enclosing the GDD 100, i.e., the most vulnerable
period for subalpine and alpine plants when growth begins.
Specifically, we aimed at (i) examining how the risk of frost
exposure for plants is related to the time of snowmelt, both in
its frequency and intensity, and (ii) testing whether the inten-
sity and the occurrence of frost events have changed over the
last five decades, when temperatures and the time of snowmelt
have significantly increased and advanced.
Materials and methods
Study sites
We analyzed air temperature and snow data from two inde-
pendent weather networks both located in Switzerland. One of
them was the IMIS network (Intercantonal Measurement and
Information System), which consists of high-elevation auto-
matic weather stations set up by the Swiss Federal Institute for
Snow and Avalanche Research (SLF). This network was
started in the beginning of the 1990s with a few stations and
has been steadily developed until reaching 103 stations in
2016. We selected 77 stations, ranging from 1630 to 2950 m
asl (Fig. 1), that covered a temporal period from 1998 to 2016
and provided accurate daily snow depth and temperature data
for at least 80% of the years during this period.
From the second and long-term weather network used in
this study, we selected 11 weather stations from the Swiss
Federal Office of Meteorology and Climatology
(MeteoSwiss) providing more than 45 years of daily snow
data and described in Klein et al. (2016). From these 11 weath-
er stations, we selected six stations covering the period 1970–
1756 Int J Biometeorol (2018) 62:1755–1762
2016 that provided daily temperature data in at least 90% of
the years and that were situated at approximately the same
altitude range than the IMIS weather stations, i.e., ranging
from 1418 to 2540 m asl (Fig. 1).
Temperature and snow data
All climatic parameters analyzed during the vulnerable period
for subalpine and alpine plants were calculated for all stations
and years that provide at least 80% of available data during
that period. For the MeteoSwiss stations, daily snow depth
was manually recorded every morning, whereas daily mini-
mum and maximum air temperatures were automatically mea-
sured. For the IMIS stations, snow depth was automatically
monitored every half hour through an ultrasonic sensor situ-
ated 6 m above the ground (SR50, Campbell Scientific, USA).
Daily morning snow depth values were then extracted for a
better comparison with the MeteoSwiss stations. Temperature
and snow data for both networks were manually checked and
tested for outliers and missing data.
Each year was considered as the period ranging from 1
September until 31 August of the following year. For each
year, the time of snowmelt was defined as the first snow-free
day after an at least 40-day snow-covered period (~ 6 weeks)
from 1 September until 31 August, following the methodolo-
gy used by Klein et al. (2016).
Data analysis and statistics
In this study, we analyzed the risk of frost exposure for plant
species growing through the altitudinal range of both IMIS
and MeteoSwiss selected stations, i.e., from 1418 to 2950 m
asl, which includes both subalpine and alpine plant species.
We defined the vulnerable period for alpine plants as the
weeks enclosing the time when an accumulation of 100 degree
days was reached since the time of snowmelt (GDD 100),
corresponding roughly to the onset of growth at plant commu-
nity scale (Vitasse et al. 2017). Specifically, we considered the
duration of this vulnerable period as the time ranging from
7 days before the GDD 100 to 14 days following this GDD
100, corresponding to a total duration of 21 days. Choosing a
shorter or a longer period (7 or 14 days before and after the
GDD 100) for defining the duration of this period did not
change the final results.
Among the two networks, non-relevant vulnerable periods
for plants were found for 4 stations-years, because the GDD
WFJ 254 0
GRH 1970
ARO 0
DAV 1560
GRC 1550
SMM 1418
ZER22750
YBR 2 17 01
VLS 2 207 0
VIN2 2730
VDS 2 23 90
TUM2 2195
TUJ3 22 20
TUJ22270
TIT2 2140
TAM 3 2 1 7 0
TAM22460
STN 2 29
STH 2 17 80
SPN3 2420
SMN22520
SHE2 1850
SCB 2 17 70
SCA 3 23 30
SCA2 2030
SAA 2 24 80
PUZ 2 21 95 PMA 2 2 430
OFE2 2360
OBW32200
OBW2 24
OBM2 2110
NEN2 23 25
NAS2 2350
NAR22070
MUT2 2 474
MUO2 2 083
MUN2 2 210
MTR2 189 0 MES2 238 0
MEI2 221 0
MAE2 2165 LUK2 2550
KLO 2 21 40
JAU21716
ILI 2 2 02 0
HTR3 2200
HTR2 215 0
GUT32200
GUT22110
GOR2 2950
GOM3 2430
GOM2 2450
GLA21630
GAN2 2717
GAD2 2060
FUS22390
FUL22610
FRA2 2100
FNH2 2240
FLU2 2390
FIS22160
FIR2 2110
FAE 2 1 9 7 0
ELM2 2050
ELA22725
DAV2 2560
CON2 2 230
CMA2 2 330
CHA22220
BOV22700
BER2 2450
BEL22556
BED3 2100
ARO 3 26 10
ARO2 2850
ANV3 2590
ALI 2 171 6
Bern
ROA2 187 0
ORT21830
CAM22220
BED 2 24 50
ANV2 2630
0 30609012015
Kilometers
IMIS stations
MeteoSwiss stations
WSL/LWF, Flurin Sutter, 2018
Source: B FS GEO STAT / Bundesamt für L andest opographi e
175
30
10
Fig. 1 Map of Switzerland showing the location and elevation of the 83 weather stations of both IMIS and MeteoSwiss networks used in the analyses
Int J Biometeorol (2018) 62:1755–1762 1757
100 was not reached before 1 September that defines the be-
ginning of the following year and were thus discarded from
the analysis (for the IMIS network: station GAN2 at 2717 m
asl in 1999 and 2004 and for the MeteoSwiss network: station
WFJ at 2540 m asl in 1978 and 1980).
For calculating the day of the year (abbreviated DOY here-
after) of the GDD 100 from the time of snowmelt, we first
computed the daily mean temperature for both networks,
based on the mean of the daily minimum and maximum air
temperature. We then accumulated all daily mean temperature
values >0 °C from the time of snowmelt, until reaching
100 °C.
Yearly snowmelt date anomalies of each IMIS and
MeteoSwiss weather station were determined by calculating
the difference between the yearly time of snowmelt and the
mean time of snowmelt of each station over the period 1998–
2016, if at least 50% of the snowmelt dates were available.
In our analysis, we considered frost events below −4°C
during the vulnerable period for plants, as this threshold cor-
responds to the lowest freezing resistance of numerous com-
mon plant species from the European Central Alps, typically
growing in snowbed conditions where the IMIS stations are
located (Ladinig et al. 2013; Taschler and Neuner 2004). The
intensity of frost was calculated by extracting the absolute
minimum air temperature occurring during this vulnerable
period. The last frost day of the season was defined as the last
occurrence of frost below −4 °C for each year (1 September–
31 August).
General spatial and temporal patterns analyzed in this
study were tested across all stations within each IMIS and/
or MeteoSwiss network, by using mixed effect models
with stations or elevation as a random effect. Different
model types were tested for each analysis (linear and
non-linear models, such as polynomial or exponential
models). The best model for each relationship was then
selected based on the lowest Akaike information criterion
(AIC). Comparisons between each model (mixed or fixed
effect models) were conducted using ANOVA to test
whether they significantly differ. Detailed statistics of each
selected mixed effect model are presented in Supplementary
Material (Online Resource 1).
All individual temporal analyses reported in this study were
performed for each MeteoSwiss station, by applying the non-
parametric Theil-Sen estimator slope, combined with a Mann-
Kendall significance test over the common temporal period
for all six stations (1970–2016), as most of the analyzed pa-
rameters were not following a normal distribution (verified
using Shapiro tests). No consistent breakpoints were detected
(tested by step-wise regressions) in the temporal trends for all
parameters and stations over the study period.
All analyses, tables and figures were performed using R 3.3
(R Core Team 2016) and the following R-packages: EnvStats,
Kendall, Hmisc, nlme, and plotrix.
Results
Temporal trends of the monthly minimum
and maximum air temperatures
A global increase of the monthly mean minimum air tempera-
tures was detected over the study period at the six MeteoSwiss
stations in the Swiss Alps, but with strong disparities across
seasons (Online Resource 2). While only slight warming was
detected during winter, minimum air temperatures increased
considerably during spring and summer, especially during the
snowmelt period from April to June, with rates ranging on
average from + 0.42 ± 0.05 °C decade
−1
to + 0.72 ± 0.06 °C
decade
−1
in May and April, respectively (Fig. 2). Similar re-
sults were found for the warming rate of mean maximum air
temperatures, with slightly higher values than for minimum air
temperatures in spring, ranging from + 0.52 ± 0.13 °C de-
cade
−1
inMayto+0.78±0.11°Cdecade
−1
in April, and with
lower values from September to December (Fig. 2).
Relationships between snowmelt and frost frequency
and intensity within stations
The mixed effect models, both with stations or elevation as a
random effect, showed significant exponential and linear relation-
ships across all stations for both IMIS and MeteoSwiss networks
(P< 0.001), between snowmelt anomalies and frost frequency or
Fig. 2 Estimated trends (slope per decade) with associated standard
errors for the monthly mean minimum (blue bars) and maximum air
temperatures (red bars), averaged from the six MeteoSwiss stations over
the period 1970–2016 and calculated from the Theil-Sen tests. The
significance level of the Theil-Sen slopes was calculated with Mann-
Kendall tests and is indicated with stars (*P< 0.05, **P< 0.01, and
***P<0.001)
1758 Int J Biometeorol (2018) 62:1755–1762
frost intensity during the vulnerable period for subalpine and al-
pine plants across the study years (Fig. 3).Theearlierthetimeof
snowmelt, the more frequent and intense the frost events during
the vulnerable period for plants in the Swiss Alps, irrespective of
elevation, the temporal period analyzed (1998–2016 or 1970–
2016), or the network used (IMIS or MeteoSwiss).
Temporal trends of the frost day frequencies
and intensities
The date corresponding to the GDD 100 after the time of
snowmelt has advanced significantly across all six available
MeteoSwiss stations during the period 1970–2016, when
using both stations or elevation as a random effect (P<
0.001) (Fig. 4a). Individual rates per station ranged from −
4.3 ± 0.1 to −7.0 ± 0.1 days decade
−1
(Online Resource 3).
However, no significant general patterns were found for the
temporal variations of the frequency or intensity of frost dur-
ing the vulnerable period for plants over the period 1970–
2016 (respective pvalues of 0.45 and 0.13) (Fig. 4b, c).
Occurrence of the last frost day of the season
The last frost day of the season (1 September-31 August) advanced
significantly during the period 1970–2016 across all six
MeteoSwiss stations, when using both stations or elevation as a
random effect (P< 0.001) (Fig. 5a). These trends ranged from −
1.0 ± 0.2 to −10.0 ± 0.4 days decade
−1
depending on stations
(Online Resource 3). As both the time of snowmelt (see Klein et
al. 2016) and the occurrence of the last frost day of the season
advanced over the last five decades at similar rates, we did not find
any general pattern across all six MeteoSwiss stations for the du-
ration of the period between the time of snowmelt and the occur-
rence of the last frost day of the season (Pvalue 0.32) (Fig. 5b).
Discussion
Relationship between snowmelt and frost exposure
for plants
Through the analysis of two independent high-elevation
weather networks in the Swiss Alps over the period 1970–
2016, our study demonstrates the strong connection between
the time of snowmelt and the spatial and temporal distribution
of the risk of frost events during the early growing season for
subalpine and alpine plants. Specifically, our analysis focused
on the most vulnerable period for plants to freezing events, i.e.,
the time shortly after snowmelt occurring generally between
spring and early summer in subalpine and alpine regions.
Fig. 3 Relationships within
stations between snowmelt
anomalies and frost frequency (a,
b) or frost intensity (c, d) during
the vulnerable period for plants,
separated for each IMIS and
MeteoSwiss networks over the
periods 1998–2016 and 1970–
2016, respectively. The black line
corresponds to the predicted
values from the best mixed effect
model (with stations as a random
effect), plotted when significant at
P<0.05
Int J Biometeorol (2018) 62:1755–1762 1759
On average, we found that in years with early snowmelt,
the frequency and intensity of frost were higher, regardless
of elevation. This finding is in agreement with results of
previous studies conducted in the Rocky Mountains
(Inouye 2008; Inouye et al. 2002)orintheSwissAlps
(Wipf et al. 2009). Our results were consistent across both
IMIS and MeteoSwiss networks and showed a consistent
relationship between snowmelt and frost risk, whether we
looked at numerous stations over a short period of time or at
only a few stations over 46 years. Our results were also
consistent for stations located in dryer or more humid re-
gions (data not shown), according to the yearly mean pre-
cipitations map in Switzerland over the 1981–2010 period
(MeteoSwiss website, unpublished work).
Despite a very high interannual variability of both plant
phenology and snowmelt timing, our methodology for ana-
lyzing the risk of frost exposure for subalpine and alpine
plants provided robust results, as findings were very similar
when testing different durations for the vulnerable period
around the GDD 100. This approach may thus be mainly
valid for alpine species inhabiting snowbed conditions,
where plants are generally more sensitive to frost, but also
for species which are able to start their growth before the time
of snowmelt, such as Crocus albiflorus, one of the first spe-
cies to start growing and flowering when the snow cover
becomesverythin(Rixenetal.2008). It may, however, be
less valid for species inhabiting ridge habitats (e.g.,
Loiseleuria procumbens), which are generally more freezing
resistant than snowbed species. Our findings suggest that
early snowmelt and a long growing season are not necessar-
ily beneficial for plants, as during such years, plants are more
exposed to freezing temperatures during the vulnerable peri-
od of initial growth. Furthermore, earlier phenology was
found to be a costly strategy for certain alpine plants when
their habitat faces temperature warming (Scheepens and
Stöcklin 2013).
Fig. 4 Temporal variations of the yearly aGDD 100 calculated from the
time of snowmelt and frost bfrequencies, and cintensities calculated
during the vulnerable period for plants across the six MeteoSwiss
stations over the period 1970–2016. The black line corresponds to the
predicted values from the best mixed effect model (with stations as a
random effect), plotted when significant at P<0.05
Fig. 5 Temporal trends of the
yearly alast frost day of the
season and bduration between
this last frost day and the time of
snowmelt across all six
MeteoSwiss stations over the
period 1970–2016. The black line
corresponds to the predicted
values from the best mixed effect
model (with stations as a random
effect), plotted when significant at
P<0.05
1760 Int J Biometeorol (2018) 62:1755–1762
Trends in temperatures, timing of snowmelt,
and frost exposure
We found that both minimum and maximum air temperatures
have increased over the period 1970–2016 for all six
MeteoSwiss stations used in this study, and particularly during
the period where snowmelt generally occurs at these eleva-
tions (April–June), with average rates exceeding 0.5 °C de-
cade
−1
for the maximum air temperatures. This result is con-
sistent with previous studies, showing similar temperature
trends during spring and summer (Rebetez and Reinhard
2008). The high consistency observed among the six
MeteoSwiss stations indicates a general warming that may
not be related to local climate conditions only. The stronger
temperature increase observed in spring, corresponding to the
mean snowmelt period, could be partly explained by the
snow-albedo positive feedback loop described by Scherrer et
al. (2012). This aforementioned study showed that around the
snow line in the Swiss Alps, a spring day without snow cover
is on average 0.4 °C warmer than a spring day with snow
cover at the same location (Scherrer et al. 2012).
Despite the strong relationships found between snowmelt
anomalies and the frequency or intensity of frost events during
the vulnerable period for subalpine and alpine plants, no con-
sistent temporal trends were found for the frequency or inten-
sity of frost over the period 1970–2016. The stable frost ex-
posure risk for plants found in this study may be explained by
the compensatory effect of a similar increase in minimum and
maximum air temperatures observed over the period 1970–
2016. Increasing maximum air temperatures have contributed
to the advance of both snowmelt and spring phenology, while
increasing minimum temperatures have delayed the last po-
tentially damaging frost, resulting in an overall unchanged
risk of frost damage.
A faster worldwide increase of maximum air temperatures
than minimum air temperatures in spring in high-elevation re-
gions (Rangwala et al. 2013), as well as a reduction of the snow
cover thickness and duration at all elevations in the Swiss Alps,
strongly connected to temperature warming (Schmucki et al.
2015;Stegeretal.2013) are expected over the next decades.
This prediction, following the general trends of temperature
warming and snowmelt observed since 1970 in the Swiss
Alps, suggests that the risk of frost exposure for subalpine
and alpine plants might not be reduced and may even increase
in the near future. Longer growing seasons with unchanged
risk of frost damage may help plants adapted to such harsh
environment to persist longer when more competitive lowland
species migrate upslope (Matteodo et al. 2013; Steinbauer et al.
2018), as plants have generally a stronger freezing resistance at
higher elevation (Sierra-Almeida et al. 2009). An increase in
plant height and biomass production is also expected by the
end of the century in the Swiss Alps, in connection with an
earlier time of snowmelt and onset of growth for plants, but
without taking into account the risk of frost exposure for plants
(Rammig et al. 2010; Carlson et al. 2017).
However, with the predicted reduction in snow cover du-
ration over the next decades, we may expect that the snow
cover will become too thin, removing the snow insulation
effect against late frost events, eventually resulting in an in-
crease of the frost exposure for plants. Warming air tempera-
tures was also shown to increase the freezing sensitivity of
plants during the beginning of their growing season (Martin
et al. 2010). With future climate warming and a weaker
protecting effect of the snow cover against frost, the risk of
exposure to frost damage for subalpine and alpine plants may
increase over the next decades.
Conclusions
The time of snowmelt is a major factor determining the expo-
sure of subalpine and alpine plants to late frost events, as
plants become coupled with surrounding air temperature. By
using long-term series of snow and temperature parameters at
high-elevation in the Swiss Alps, we showed that an early time
of snowmelt generally leads to an increasing frequency and
intensity of frost during the vulnerable period for plants, irre-
spective of elevation or the temporal period analyzed (1998–
2016 or 1970–2016). However, despite climate warming and
the general decline of both snowmelt timing and last frost day
of the season in the Swiss Alps, our study suggests that the
frequency and intensity of frost during the vulnerable period
for plants have remained unchanged over the period 1970–
2016. This absence of trends may be explained by the similar
increase of minimum and maximum air temperatures found
over the same period, which has shifted spring phenology and
the last occurrence of potentially damaging frost to a same
extent. Longer growing season with unchanged risk of frost
damage may help plants adapted to such harsh environment to
persist longer, whereas new thermophile species colonizing
from lowlands areas could experience severe frost slowing
down their upward shift. It remains a future research challenge
if our results also hold in a global context of different alpine
climates. In oceanic regions with unpredictable climate, frost
can occur at any time of the year and hence, the freezing
resistance of plants can be higher than in regions with predict-
able snow cover (Bannister 2007; Bannister et al. 2005; Venn
et al. 2013). Understanding the role of frost events in different
climates will considerably improve our predictions of vegeta-
tion changes under ongoing climate.
Acknowledgements We are grateful to Christoph Marty for providing
IMIS temperature and snow data and to MeteoSwiss for providing
long-term series of temperature and snow data. We also thank Flurin
Sutter for drawing the map of the selected stations shown in Fig. 1and
Bradley Carlson for his editorial improvements of the manuscript.
Int J Biometeorol (2018) 62:1755–1762 1761
References
Ahas R, Aasa A, Menzel A, Fedotova V, Scheifinger H (2002) Changes
in European spring phenology. Int J Climatol 22:1727–1738
Bannister P (2007) Godley review: a touch of frost? Cold hardiness of
plants in the southern hemisphere. N Z J Bot 45:1–33
Bannister P, Maegli T, Dickinson KJ, Halloy SR, Knight A, Lord JM,
Mark AF, Spencer KL (2005) Will loss of snow cover during cli-
matic warming expose New Zealand alpine plants to increased frost
damage? Oecologia 144:245–256
Carlson BZ, Corona MC, Dentant C, Bonet R, Thuiller W, Choler P
(2017) Observed long-term greening of alpine vegetation—a case
study in the French Alps. Environ Res Lett 12:114006
Gerdol R, Siffi C, Iacumin P, Gualmini M, Tomaselli M (2013) Advanced
snowmelt affects vegetative growth and sexual reproduction of
Vaccinium myrtillus in a sub-alpine heath. J Veg Sci 24:569–579
Inouye DW (2008) Effects of climate change on phenology, frost damage,
and floral abundance of montane wildflowers. Ecology 89:353–362
Inouye DW, Morales MA, Dodge GJ (2002) Variation in timing and
abundance of flowering by Delphinium barbeyi Huth
(Ranunculaceae): the roles of snowpack, frost, and La Nina, in the
context of climate change. Oecologia 130:543–550
Jonas T, Rixen C, Sturm M, Stoeckli V (2008) How alpine plant growth is
linked to snow cover and climate variability. J Geophys Res
113(G03013):10. https://doi.org/10.1029/2007JG000680
Klein G, Vitasse Y, Rixen C, Marty C, Rebetez M (2016) Shorter snow
cover duration since 1970 in the Swiss Alps due to earlier snowmelt
more than to later snow onset. Clim Chang 139:637–649
Körner C (2003) Alpine plant life: functional plant ecology of high moun-
tain ecosystems. 2nd ed. Springer, Berlin - Heidelberg
Ladinig U, Hacker J, Neuner G, Wagner J (2013) How endangered is
sexual reproduction of high-mountain plants by summer frosts?
Frost resistance, frequency of frost events and risk assessment.
Oecologia 171:743–760
Martin M, Gavazov K, Körner C, Hättenschwiler S, Rixen C (2010) Reduced
early growing season freezing resistance in alpine treeline plants under
elevated atmospheric CO2. Glob Chang Biol 16:1057–1070
Marty C (2008) Regime shift of snow days in Switzerland. Geophys Res
Lett 35(L12501):5. https://doi.org/10.1029/2008GL033998
Matteodo M, Wipf S, Stöckli V, Rixen C, Vittoz P (2013) Elevation
gradient of successful plant traits for colonizing alpine summits
under climate change. Environ Res Lett 8:024043
Menzel A, Sparks TH, Estrella N, Koch E, Aasa A, Ahas R, ALM-
KÜBLER K, Bissolli P, Braslavská O, Briede A (2006) European
phenological response to climate change matches the warming pat-
tern. Glob Chang Biol 12:1969–1976
Nagy L, Grabherr G (2009) The biology of alpine habitats. Oxford
University Press, Oxford
Palacio S, Lenz A, Wipf S, Hoch G, Rixen C (2015) Bud freezing resis-
tance in alpine shrubs across snow depth gradients. Environ Exp Bot
118:95–101
Park S-H, Lee M-J, Jung H-S (2012) Analysis on the snow cover varia-
tions at Mt. Kilimanjaro using Landsat satellite images. Korean J
Remote Sens 28:409–420
Pederson GT, Betancourt JL, McCabe GJ (2013) Regional patterns and
proximal causes of the recent snowpack decline in the Rocky
Mountains, US. Geophys Res Lett 40:1811–1816
Rammig A, Jonas T, Zimmermann N, Rixen C (2010) Changes in alpine plant
growth under future climate conditions. Biogeosciences 7:2013–2024
Rangwala I, Sinsky E, Miller JR (2013) Amplified warming projections
for high altitude regions of the northern hemisphere mid-latitudes
from CMIP5 models. Environ Res Lett 8:024040
Rebetez M, Reinhard M (2008) Monthly air temperature trends in
Switzerland 1901–2000 and 1975–2004. Theor Appl Climatol 91:
27–34
Rixen C, Freppaz M, Stoeckli V, Huovinen C, Huovinen K, Wipf S
(2008) Altered snow density and chemistry change soil nitrogen
mineralization and plant growth. Arct Antarct Alp Res 40:568–575
Rixen C, Dawes MA, Wipf S, Hagedorn F (2012) Evidence of enhanced
freezing damage in treeline plants during six years of CO2 enrich-
ment and soil warming. Oikos 121:1532–1543
Scheepens J, Stöcklin J (2013) Flowering phenology and reproductive
fitness along a mountain slope: maladaptive responses to transplan-
tation to a warmer climate in Campanula thyrsoides. Oecologia 171:
679–691
Scherrer S, Ceppi P, Croci-Maspoli M, Appenzeller C (2012) Snow-
albedo feedback and Swiss spring temperature trends. Theor Appl
Climatol 110:509–516
Schmucki E, Marty C, Fierz C, Lehning M (2015) Simulations of 21st
century snow response to climate change in Switzerland from a set
of RCMs. Int J Climatol 35:3262–3273
Serquet G, Marty C, Rebetez M (2013) Monthly trends and the corre-
sponding altitudinal shift in the snowfall/precipitation day ratio.
Theor Appl Climatol 114:437–444
Sherwood J, Debinski D, Caragea P, Germino M (2017) Effects of exper-
imentally reduced snowpack and passive warming on montane
meadow plant phenology and floral resources. Ecosphere 8(3):
e01745. https://doi.org/10.1002/ecs2.1745
Sierra-Almeida A, Cavieres LA, Bravo LA (2009) Freezing resistance
varies within the growing season and with elevation in high-
Andean species of central Chile. New Phytol 182:461–469
Steger C, Kotlarski S, Jonas T, Schär C (2013) Alpine snow cover in a
changing climate: a regional climate model perspective. Clim Dyn
41:735–754
Steinbauer MJ, Grytnes J-A, Jurasinski G, Kulonen A, Lenoir J, Pauli H,
Rixen C, Winkler M, Bardy-Durchhalter M, Barni E (2018)
Accelerated increase in plant species richness on mountain summits
is linked to warming. Nature 556:231–234
Taschler D, Neuner G (2004) Summer frost resistance and freezing patterns
measured in situ in leaves of major alpine plant growth forms in relation
to their upper distribution boundary. Plant Cell Environ 27:737–746
R Core Team (2016) R: A Language and Environment for Statistical
Computing. Vienna, Austria
Valt M, Cianfarra P (2010) Recent snow cover variability in the Italian
Alps. Cold Reg Sci Technol 64:146–157
Venn SE, Morgan JW, Lord JM (2013) Foliar freezing resistance of
Australian alpine plants over the growing season. Austral Ecol 38:
152–161
Vitasse Y, Rebetez M, Filippa G, Cremonese E, Klein G, Rixen C (2017)
‘Hearing’alpine plants growing after snowmelt: ultrasonic snow
sensors provide long-term series of alpine plant phenology. Int J
Biometeorol 61:349–361
Vitasse Y, Schneider L, Rixen C, Christen D, Rebetez M (2018a) Increase
in the risk of exposure of forest and fruit trees to spring frosts at
higher elevations in Switzerland over the last four decades. Agric
For Meteorol 248:60–69
Vitasse Y, Signarbieux C, Fu YH (2018b) Global warming leads to more
uniform spring phenology across elevations. Proc Natl Acad Sci
115:1004–1008
Wheeler J, Hoch G, Cortés AJ, Sedlacek J, Wipf S, Rixen C (2014)
Increased spring freezing vulnerability for alpine shrubs under early
snowmelt. Oecologia 175:219–229
Wipf S, Stoeckli V, Bebi P (2009) Winter climate change in alpine tundra:
plant responses to changes in snow depth and snowmelt timing.
Clim Chang 94:105–121
Xu Y, Ramanathan V, Washington W (2016) Observed high-altitude
warming and snow cover retreat over Tibet and the Himalayas en-
hanced by black carbon aerosols. Atmos Chem Phys 16:1303–1315
1762 Int J Biometeorol (2018) 62:1755–1762
