Content uploaded by Gérald Giraud
Author content
All content in this area was uploaded by Gérald Giraud on Oct 30, 2015
Content may be subject to copyright.
Reanalysis of 44 Yr of Climate in the French Alps (1958–2002): Methodology, Model
Validation, Climatology, and Trends for Air Temperature and Precipitation
YVES DURAND,MARTIN LATERNSER,GE
´RALD GIRAUD,PIERRE ETCHEVERS,
BERNARD LESAFFRE,AND LAURENT ME
´RINDOL
GAME/CNRM-CEN (CNRS/Me
´te
´o-France), Saint-Martin d’He
´res, France
(Manuscript received 19 June 2007, in final form 18 August 2008)
ABSTRACT
Since the early 1990s, Me
´te
´o-France has used an automatic system combining three numerical models to
simulate meteorological parameters, snow cover stratification, and avalanche risk at various altitudes, as-
pects, and slopes for a number of mountainous regions in France. Given the lack of sufficient directly
observed long-term snow data, this ‘‘SAFRAN’’–Crocus–‘‘MEPRA’’ (SCM) model chain, usually applied to
operational avalanche forecasting, has been used to carry out and validate retrospective snow and weather
climate analyses for the 1958–2002 period. The SAFRAN 2-m air temperature and precipitation climatology
shows that the climate of the French Alps is temperate and is mainly determined by atmospheric westerly
flow conditions. Vertical profiles of temperature and precipitation averaged over the whole period for alti-
tudes up to 3000 m MSL show a relatively linear variation with altitude for different mountain areas with no
constraint of that kind imposed by the analysis scheme itself. Over the observation period 1958–2002, the
overall trend corresponds to an increase in the annual near-surface air temperature of about 18C. However,
variations are large at different altitudes and for different seasons and regions. This significantly positive
trend is most obvious in the 1500–2000-m MSL altitude range, especially in the northwest regions, and
exhibits a significant relationship with the North Atlantic Oscillation index over long periods. Precipitation
data are diverse, making it hard to identify clear trends within the high year-to-year variability.
1. Introduction
Since the early 1990s, Me
´te
´o-France has used an au-
tomatic system based on three numerical models to
simulate meteorological parameters, snow cover stratig-
raphy, and avalanche risk at various altitudes, aspects,
and slopes for a number of mountainous regions (mas-
sifs) in France (Durand et al. 1999). This SAFRAN–
Crocus–MEPRA
1
(SCM) model chain, usually applied
to operational avalanche forecasting, is used here for
retrospective snow and weather climate analyses.
A 10-yr snow climatology (1981–91) of the French
Alps has been established on the basis of modeled snow
data alone [i.e., not taking into account any snow mea-
surements (Martin 1995)]. These data have been used to
test snow sensitivity to input meteorological parameters
(Martin et al. 1994). Similar studies have been carried out
for Switzerland with a particular emphasis on long-term
trends (Laternser and Schneebeli 2003).
As far as we know, no practical climatological studies
on combined snow and meteorological parameters have
been carried out for the French Alps. Classical clima-
tological studies in France concentrate more on the
predominant low-elevation regions of the country and
focus mainly on air temperature and precipitation.
Moisselin et al. (2002) and Schmidli et al. (2002) discuss
precipitation trends in detail for the entire European
Alps. Frei and Scha
¨r (1998) have determined a high-
resolution precipitation climatology for the Alps based
on daily analyses using a methodology similar to ours
for this parameter. They also present a very complete
description of several prior climatological studies over
the Alps. Martin Beniston has also widely investigated
1
Here, SAFRAN stands for Syste
`me d’Analyse Fournissant des
Renseignements Atmosphe
´riques a
`la Neige (Analysis System Pro-
viding Atmospheric Information to Snow) and MEPRA stands for
Mode
`le Expert de Pre
´vision du Risque d’Avalanche (Expert System
for Avalanche Hazard Estimation).
Corresponding author address: Yves Durand, Me
´te
´o-France
CNRM-CEN, 1441 rue de la Piscine, 38400 Saint-Martin d’He
´res,
France.
E-mail: yves.durand@meteo.fr
VOLUME 48 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY MARCH 2009
DOI: 10.1175/2008JAMC1808.1
Ó2009 American Meteorological Society 429
mountain climates with particular emphasis on the en-
tire Alps in France and Switzerland (Beniston 2005;
Rebetez and Beniston 1998). Some regional studies
on precipitation have also been carried out (Berthelot
2004) in the southern French Alps.
The present study analyzes long-term climate series over
the entire French Alps. Using 44 yr of newly reanalyzed
atmospheric model data from the 40-yr European Centre
for Medium-Range Weather Forecasts (ECMWF) re-
analysis (ERA-40) project (ECMWF 2004), the SCM
model chain has been run on an hourly basis for a period
starting in winter 1958/59. Results include air tempera-
ture and precipitation trends, as well as average condi-
tions (spatial variability) and long-term trends (temporal
variability) for various snow-cover parameters.
The SCM chain has already been validated on nu-
merous occasions but only over the 1981–95 period in
mountainous areas. As few validation stations are avail-
able prior to 1980, the SAFRAN meteorological model
was validated using specific procedures. The results of
these comparisons show that the air temperature error
(RMS) is between 1.58and 28C and that the precipita-
tion error (only for nonzero precipitation) is unbiased.
These values are satisfactory even if few validation
stations are used by the model. Furthermore, these
magnitudes of error have been corroborated by other
studies over other geographical areas. Because of their
pertinence for snowpack evolution, these validation
tests only involve two of the nine parameters analyzed.
The obtained results show the ability of the SAFRAN
model to reproduce the main climatological features for
this mountainous region and to provide valid input data
for the Crocus snow model.
Because of the amount of information involved, this
study is dividedinto two papers. The first paper (this one)
focuses on data description, methodology, and validation
in relation to the SAFRAN meteorological model and
presents meteorological trends for air temperature and
precipitation (total and snow). A forthcoming paper will
deal with the results of the Crocus snowpack model,
providing a comprehensive snow climatology and long-
term snow trends. Results will be discussed, snow trends
considered in the light of the air temperature and pre-
cipitation trends revealed by this paper, and compari-
sons made with international snow studies (Rebetez and
Beniston 1998). The MEPRA expert system for avalanche
risk forecasting (Giraud 1993) is not used in this part of
the study and is therefore not discussed in this paper.
2. Models used
SAFRAN (Durand et al. 1993) is a meteorological
application that performs an objective analysis of weather
data available from various observation networks (in-
cluding radar and satellite data) over the considered el-
evations and aspects of the different massifs. SAFRAN
combines the observed information with a preliminary
estimation generally provided by numerical weather
forecasting models. The analysis method combines an
optimal interpolation every 6 h and a variational inter-
polation over 6-h windows, providing hourly data for the
main relevant atmospheric parameters affecting snow
surface changes (i.e., air temperature, wind speed, air
humidity, cloudiness, snow and rain precipitation, long-
wave radiation, and direct and scattered solar radiation).
Crocus (Brun et al. 1989, 1992) is a numerical snow
model used to calculate changes in energy and mass in
the snow cover. It uses only the meteorological data
provided by SAFRAN and simulates the evolution of
temperature, density, liquid water content profiles, and
layering of the snowpack at different elevations, slopes,
and aspects, including the internal metamorphism pro-
cesses. It is assumed that each simulated slope is free of
snow on 1 August of each year. The simulated snowpack
then evolves every hour from the first snowfall until
complete melting without reinitialization. The com-
puted snow state for a given hour is thus based only on
the snow state of the previous hour and on the atmo-
spheric forcing of the current hour.
In the present paper, the following output data will be
presented: air temperature and precipitation (total or
snow), all provided by SAFRAN; snow depth at ground
level as computed by Crocus will be in a forthcoming
paper. All these parameters are modeled for all 23 massifs
of the French Alps in 300-m-altitude steps over eleva-
tions ranging at the most from 300 to 3600 m MSL (Fig. 1;
Table 1). These different massifs have been defined for
their climatological homogeneity, especially with regard
to precipitation fields (Pahaut et al. 1991). They are those
used for operational avalanche hazard estimation in
France and their characteristics have been well known
for many years by local forecasters. Their boundaries co-
incide well with the main topographic features as shown
by Fig. 1. However, for each massif, only existing eleva-
tions are considered and no ‘‘fictitious’’ extrapolations are
made to higher or lower elevations, which can make
comparisons difficult between massifs for certain elevation
ranges. The output has an hourly resolution from 1 August
1958 to 31 July 2002 and covers 44 winter periods. By
convention, winters are referred to by the year of the main
part of the winter (e.g., 1959 means winter 1958/59).
3. Data and methods
The meteorological analyses are based on both con-
ventional observations and numerical atmospheric weather
430 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
model outputs. Conventional observations include var-
ious kinds of datasets extracted from the operational
databases of Me
´te
´o-France and ECMWF. They cover
the French Alps and adjacent areas of neighboring Italy
and Switzerland within a grid of 43.158–47.08N and
4.458–8.08E. Data are concatenated into several differ-
ent file types according to their contents and source. The
initial data had not been checked properly in terms of
quality (apart from quality flags at ECMWF and some
routine consistency checks at Me
´te
´o-France); however,
this was done automatically during subsequent SAFRAN
modeling.
All available conventional observations have been
used and are recorded in several files and databases
(Table 2). Air pressure, air temperature, wind (meridian
and zonal components), humidity, snow depth, new
snow, and various parameters for weather type and
cloudiness are available at their own observation fre-
quency (hourly or by steps of 3 or 6 h). Precipitation,
snow depths, and minimum–maximum temperatures are
available only on a daily basis. Radiosonde and pilot
balloon data from Lyon, Montelimar, Nı
ˆmes, Payerne
(Switzerland), and Torino (Italy) are also used. The
number of stations providing available data varies
greatly with the hour, day, and year considered and is
thus given only as a general indication. Individual files
are incomplete in roughly two-thirds of all cases, in
particular snow, weather type, and cloudiness along
with minimum–maximum temperature and new snow
amount. All these missing data and short observation
series are the main reason for using meteorological
analysis software such as SAFRAN that uses informa-
tion available on a daily basis, even if sparse, without
the constraint of full and homogeneous observation
series. However, since it takes observation errors into
account statistically, SAFRAN is not a ‘‘perfect’’ in-
terpolator scheme [i.e., it will never give the observation
TABLE 1. Details of the mountainous massifs of the French Alps
used in the SCM chain with their elevation range and geographic
region (cf. Fig. 1).
Massif Elev (m) Region
Chablais 600–2700 m NW foothills
Aravis 900–2700 m NW foothills
Bauges 600–2100 m NW foothills
Chartreuse 600–2100 m NW foothills
Vercors 600–2400 m NW foothills
Mont-Blanc 1200–3600 m North
Beaufortin 900–3000 m North
Haute-Tarentaise 900–3600 m North (interior)
Haute-Maurienne 1200–3600 m North (central, interior)
Vanoise 900–3600 m North
Maurienne 600–3000 m North (central)
Belledonne 600–3000 m North
Grandes-Rousses 900–3300 m Central
Thabor 1500–3000 m Central (interior)
Oisans 900–1200 m South (central)
Pelvoux 1200–3600 m South (central)
Champsaur 1200–3300 m South
De
´voluy 600–3000 m South
Queyras 1200–3000 m South (interior SE)
Parpaillon 900–3300 m South (interior SE)
Ubaye 1200–3000 m South (interior SE)
Alpes-Azure
´ennes 600–2700 m Far south
Mercantour 1200–3000 m Far south
FIG. 1. The 23 massifs of the French Alps between Lake Geneva to the north and the
Mediterranean Sea to the south and their underlying orographic features.
MARCH 2009 D U R A N D E T A L . 431
value (except by chance)], even if the observation point
corresponds exactly to an analysis point.
Prior to this study, the SCM chain had been run since
1981 using different numerical guess fields provided
by the available numerical weather prediction models
in use at Me
´te
´o-France. The most widely used is the
Action de Recherche Petite Echelle Grande Echelle
(ARPEGE) model (Courtier et al. 1991), with a present
dynamic resolution of about 20 km. This guess field has
more particularly been used by Quintana-Seguı
´et al.
(2008) for their work with SAFRAN. However, as
output from this model is not available prior to 1991,
we chose to use retrospective analyses from ERA-40
(ECMWF 2004) that provide a uniform coverage of
our entire study period even if their spatial resolution
is coarser. This version of the ECMWF assimilation
scheme uses both satellite and conventional observa-
tions to provide a full set of validated meteorological
analysis parameters from the surface to the 0.1-hPa
level (’65 km MSL) dating back to 1958 (the Interna-
tional Geophysical Year). For our purposes, six parame-
ters (P,Z,T,U,V,H; see Table 3) over a maximum of
16 elevation levels (from the surface up to the 300-hPa
level at ’8.5 km MSL) within a regular grid of 1.58
latitude–longitude were extracted over the entire pe-
riod. Horizontal and vertical downscaling operators
have also been developed to adapt these data from the
ECMWF MARS archive system to the concerned pa-
rameters and area of the Alps.
Even if the analyzed results cover the entire annual
period on an hourly time step and are therefore avail-
able at that scale, the results presented here in the dif-
ferent figures are mainly yearly or seasonal averages or
amounts at different elevations of two selected varia-
bles: near-surface air temperature (sometimes referred
to only as air temperature or temperature) and 24-h
rainfall (sometimes referred to only as precipitation).
Some finer results will be discussed in the text, in partic-
ular for half-seasons such as early summer or midwinter,
but will not be illustrated so as to avoid complicating
the figures. Elevation ranges are often referred to as
low (,1000 m), mid- (1000–2000 m), and high alti-
tudes (.2000 m); however, these terms should not be
taken too literally since they only represent a rough
graduation.
4. Model validation
Before running the two first models in coupled mode,
each was carefully validated in different contexts. Two
well-instrumented automatic sites, Col de Porte (1340 m,
Chartreuse massif) and Col du Lac Blanc (2800 m,
Grandes Rousses massif), are not included in the anal-
ysis system and are used for a daily local validation of
SAFRAN (Durand et al. 1993). Crocus has been vali-
dated at Col de Porte over several winter seasons (Brun
TABLE 2. Characteristics of the different observation sources
(see text).
Network,
observation
Temporal
resolution (h)
Period of
observation
No. of stations
(approx)
Synoptic, surface 6 1958–2002 30–200
Automatic, surface 1 1958–2002 30–200
Climatologic, surface 24 1958–2002 150–300
Vertical atmospheric 12 (6) 1958–2002 0–5
TABLE 3. ERA-40 output data parameters and corresponding elevation levels used in this study.
Level P(air pressure) Z(geopotential) T(air temperature) U(meridional wind) V(zonal wind) H(humidity)
Surface x X X X X
1000 hPa X X X X X X
925 hPa X X X X X X
850 hPa X X X X X X
775 hPa X X X X X X
700 hPa X X X X X X
600 hPa X X X X X X
500 hPa X X X X X X
400 hPa X X X X X X
300 hPa X X X X X X
Model level X
No. 51 900 hPa* X
No. 50 880 hPa* X
No. 48 820 hPa* X
No. 47 790 hPa* X
No. 45 720 hPa* X
No. 43 650 hPa* X
* The model level Pdepends on the surface pressure; therefore the values given are only approximate (ECMWF 2004).
432 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
et al. 1989, 1992) using measured meteorological data
from automatic weather stations. The SAFRAN and
Crocus models were assessed in coupled mode by
comparing simulated and measured snow depths at 37
sites over the 1981–91 period (Martin et al. 1994). The
quality of the simulations is satisfactory except at loca-
tions where snowdrifting is very frequent or where the
local meteorology significantly differs from regional
(i.e., here massif) meteorology. Results are better in the
northern Alps than in the southern Alps because of a
higher density of the snow weather observation network
in the northern Alps. As mentioned by Martin et al.
(1994), the maximum snow depth errors of the 37 sites
are usually less than 20 cm for test sites below 1500 m,
and 30 cm for other sites. The corresponding mean error
values are, respectively, 10 and 13 cm, corresponding to
18% and 12% of the observed mean snow depths. This
encouraging snow depth evaluation represents an indi-
rect validation of SAFRAN meteorological parameters
presented here, especially precipitation. The other results
concerning snow parameters will be presented in a
forthcoming paper. A global validation of SAFRAN ca-
pacities over all of France has also been performed
by Quintana-Seguı
´et al. (2008) and has confirmed the
unbiased characteristic of the results in a hydrological
context. Their results show an average value of the RMS
difference between SAFRAN output and observations at
about 1.58C for temperature, but at the same time point
out some problems concerning precipitation over moun-
tainous areas related to the sub-massif-scale variability.
Given the small number of previous studies and the
lack of widely distributed mountain meteorological
observations, especially over the 1958–80 period, the
SAFRAN results have been validated mainly in terms
of air temperature and precipitation. The choice of
these two parameters is also related to their impact on
snow evolution. As no observed data are directly rep-
resentative of the massif scale, results were validated us-
ing the ability of the SAFRAN model to simulate precise
geographical locations by including their main surround-
ing topographical features (Durand et al. 1999) through
appropriate downscaling procedures. In a first step, 43
such sites (Table 4) were selected using two criteria:
available meteorological data during most of the period,
and sites well distributed over the whole Alps.
For the entire 44-yr period, a first run was carried out
without the observations from these selected sites. This
first experiment and the corresponding results will be
hereinafter referred to as ‘‘WITHOUT.’’ In a second
step, a new run referred to as ‘‘WITH’’ was carried
out using all the observations. Objective comparisons
between the raw observed data and corresponding
SAFRAN WITH and WITHOUT data were then per-
formed and the results illustrate both the quality of the
analysis and the pertinence of the observations.
a. Air temperature
Figure 2 shows the RMS values of the difference be-
tween measurements (43 sites, some with sometimes
sporadic data) and SAFRAN analyzed fields for the
annual mean air temperature over the 44 yr for the two
datasets (WITH represented by the solid line and
WITHOUT by the broken line). No constraint is ap-
plied to the analysis scheme and the presented RMS
values therefore include the observation errors of the
concerned sites, which increase the results. It is difficult
to separate the relative share of these two errors. As a
general indication, the values of RMS temperature ob-
servation error used in our area of the Alps range from
18to 1.58C depending on the site and result from our
own monitoring and experience. These values are close
to those suggested, for instance by Fuentes and Heimann
(1996). The WITH set exhibits higher quality because of
the additional information of the test sites, but the RMS
difference between the two sets is low, from 0.18Cin
2000–09 to 0.38C in the 1960s. The SAFRAN analyses
are globally better at the end of the millennium because
of the improvement of the snow and weather network
with an increasing number of meteorological observa-
tions. For extreme values (details not shown here), the
RMS values vary from 0.18C in 1996 for WITH simu-
lations for a site in the southern Alps to 4.78C in 1986 for
WITHOUT simulations for a site in the central Alps.
The minimum (TN) and maximum (TX) daily air tem-
peratures have also been compared. The bias is gener-
ally positive for TN (mean value of 1.28C for WITH and
1.98C for WITHOUT set) and negative for TX (mean
value of 21.08C for WITH and 21.48C for WITHOUT
set).
The relative decrease in performance over the 1980s
has been identified to a lack of information in the da-
tabases, mainly due to the demise of the French manual
observation network that had not yet been compensated
for by the deployment of the new snow weather ob-
servation network in mountainous areas—which clearly
produced a positive impact during the 1990s. The au-
tomatic observation network was also, at that time, only
in its infancy, with data difficult to integrate in the
analysis scheme. As described farther on (and also vis-
ible in Fig. 9), the 1980s are also representative of a net
change in the warming temperature trend (Trenberth
et al. 2007), especially concerning the daily minima that
are not used explicitly by SAFRAN. This phenomenon
could also partially explain the bad RMS values of that
critical period if we could establish the sensitivity of this
MARCH 2009 D U R A N D E T A L . 433
particular parameter with respect to the final result
in the framework of a reduced observation network;
however, this point is still under investigation.
The presented results are also an indirect evaluation
of SAFRAN accuracy according to the information
used and of the sensitivity of the analysis scheme to its
input data. When the observation network is sparse, as
during the 1980s, the analysis error is close to the guess-
field error. On the other hand, a denser observation
network implies less analysis error in the vicinity of the
observation errors. As a whole, the magnitude of these
differences is very close to those obtained by Quintana
(RMS value of about 1.58C), who used SAFRAN over
all of France with a majority of low-elevation areas as
previously indicated (Quintana-Seguı
´et al. 2008).
Some detailed results concerning seven selected sites
(one for each French Alps department) are shown
in Table 5 and concern TN–TX and mean daily tem-
perature (TM) verifications. The values correspond to
RMS differences and bias differences. Results of the
TABLE 4. List of the 43 selected validation sites used in Fig. 2 (comparisons WITH–WITHOUT insertion in the analysis scheme) with
their main characteristics. The 21 sites with the longest observed series and used in Fig. 4 (comparison with observations) are in italic. The
seven ‘‘department representative’’ sites used in the SAFRAN validation (Tables 5, 6) are in boldface.
Name Department Massif Alt (m MSL) Lat (8N) Lon (8E) Aspect
Les Gets Haute Savoie Chablais 1172 46.17 6.67 E
Les Contamines Haute Savoie Chablais 1180 45.77 6.73 N
La Clusaz Haute Savoie Aravis 1180 45.9 6.42 NW
Vallorcine Haute Savoie Mt Blanc 1300 46.03 6.93 NE
Chamonix Haute Savoie Mt Blanc 1042 45.91 6.86 F
Megeve Haute Savoie Mt Blanc 1104 45.85 6.62 SW
Les Houches Haute Savoie Mt Blanc 1008 45.89 6.81 F
Beaufort Savoie Beaufortin 1030 45.69 6.57 F
Bourg St Maurice Savoie Beaufortin 868 45.61 6.76 F
Val d’lse
`re Savoie Hte Tarentaise 1850 45.44 6.98 F
Pralognan Savoie Vanoise 1420 45.38 6.72 F
Peisey-Nancroix Savoie Vanoise 1350 45.32 6.46 F
St Martin de B. Savoie Vanoise 1500 45.27 6.51 F
Bonneval Savoie Hte Tarentaise 1830 45.37 7.04 NW
Bessans Savoie Hte Tarentaise 1715 45.32 6.99 F
Valloire Savoie Maurienne 1296 45.08 6.43 F
St Sorlin Savoie Maurienne 1650 45.22 6.23 S
Chambery Savoie Bauges 239 45.65 5.88 F
St Pierre de C Ise
`re Chartreuse 945 45.33 5.82 F
St Pierre d’E Ise
`re Chartreuse 644 45.38 5.85 F
Revel Ise
`re Belledonne 630 45.19 5.87 F
La Ferrie
`re Ise
`re Belledonne 1082 45.28 6.07 F
SMH Ise
`re Chartreuse 200 45.17 5.77 F
Allemond Ise
`re Belledonne 1270 45.16 6.08 F
Autrans Ise
`re Vercors 1090 45.17 5.54 F
Villard de L. Ise` re Vercors 1050 45.07 5.55 F
Besse Ise
`re Grdes Rousses 1525 45.07 6.17 F
Vaujany Ise
`re Grdes Rousses 772 45.13 6.08 N
Bourg d’Ois Ise
`re Grdes Rousses 720 45.03 6.18 F
St Christ. En O. Ise
`re Oisans 1570 44.95 6.17 SW
La Grave Hautes Alpes Oisans 1780 45.03 6.3 SW
St Etienne en D. Hautes Alpes Devoluy 1300 44.7 5.93 S
Lus La Cr. Hautes Alpes Devoluy 1059 44.68 5.71 SW
Briancxon Hautes Alpes Le Pelvoux 1324 44.91 6.61 F
Le Monetier Hautes Alpes Le Pelvoux 1490 44.97 6.5 F
Orcie
`res Hautes Alpes Champsaur 1435 44.68 6.32 F
St Veran Hautes Alpes Queyras 2010 44.7 6.87 SW
Arvieux Hautes Alpes Queyras 1675 44.75 6.75 F
Ceillac Hautes Alpes Queyras 1665 44.67 6.77 SE
Embrun Alpes Hte Provence Parpaillon 871 44.57 6.5 S
Barcelonnette Alpes Hte Provence Ubaye 1155 44.39 6.67 F
Beuil Alpes Maritimers Alpes-Asure
´ennes 1465 44.1 7 F
Luceram Alpes Maritimers Mercantour 1420 43.93 7.37 S
434 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
comparisons for the WITH and WITHOUT experi-
ments are indicated in a similar way. The results of the
comparisons are obviously better for WITH experi-
ments (because of the use of the additional observations
in the analysis scheme), except for Luceram, a site re-
cently set up in the southern French Alps (detailed re-
sults not shown) with few available data and for which
insertion in the SAFRAN data decreases analysis per-
formance. Except for this particular site, the bias is al-
ways positive for TN and negative for TX, as already
mentioned. This supports our previous idea concerning
the possible weaknesses of the diurnal cycle analyzed by
SAFRAN (underestimation of the amplitude, but less
error on the mean temperature value) and the possible
improvements that could be achieved by using explicitly
the information on these extremes. This also shows that
the analysis process is not trivial and that the charac-
teristics of each site have to be carefully taken into ac-
count, which is not yet the case for Luceram. The same
analyses were performed using only winter data and
both bias and RMS show similar values (not shown
here).
b. 24-h precipitation
As 24-h precipitation is not a continuous daily pa-
rameter, direct comparisons are not easy. We therefore
used two types of comparisons.
For the sites previously selected in each French Alps
department (seven sites, already used in Table 5),
objective comparisons are computed between the
observed and the WITH and WITHOUT corre-
sponding SAFRAN analyzed quantities only when
the daily quantity is higher than 0.2 mm. Table 6
shows this comparison using different statistical
parameters such as the average of observed data and
SAFRAN analyzed data, standard deviation of the
difference, and the correlation coefficient. As for
temperature, the Luceram site shows the worst re-
sults for all statistical parameters, which reinforces
our previous doubts concerning the use of these data
in the analysis scheme. However, the weak differ-
ences between WITH and WITHOUT results are
representative of the good stability of the SAFRAN
scheme, especially in relation to precipitation, for
TABLE 5. Illustration of the spatial variability of the SAFRAN analyses; the table shows for seven sites the rms and bias difference for
minimum (TN), maximum (TX), and average (TM) air temperature (8C) over the whole period and the two experiment sets (WITH and
WITHOUT).
Site
No. of
cases
WITH WITHOUT
TN (8C) bias
(rms)
TX (8C) bias
(rms)
TM (8C) bias
(rms)
TN (8C) bias
(rms)
TX (8C) bias
(rms)
TM (8C)
bias (rms)
Chamonix 13 695 12.2 (3.1) 21.0 (2.3) 10.7 (1.7) 12.9 (3.8) 21.2 (2.6) 10.8 (1.9)
Bessans 6921 13.4 (4.6) 20.5 (2.8) 11.5 (2.8) 13.6 (4.9) 20.5 (2.9) 11.6 (2.9)
Villard de Lans 13 138 11.0 (2.8) 21.0 (2.1) 20.0 (1.7) 12.2 (3.7) 21.3 (2.5) 10.5 (1.9)
Lus La Croix Haute 15 930 10.6 (2.0) 20.3 (1.5) 10.2 (1.2) 11.9 (3.4) 20.8 (2.2) 10.5 (1.7)
Orcie
`res 12 249 10.6 (2.3) 22.2 (3.1) 20.8 (1.7) 11.4 (2.7) 22.6 (3.4) 20.6 (1.7)
Barcelonnette 14 246 13.9 (4.9) 21.7 (2.7) 11.1 (5.8) 14.9 (5.8) 22.6 (3.6) 11.2 (2.3)
Luceram 2791 22.5 (3.6) 10.6 (1.8) 21.0 (1.7) 22.4 (3.4) 10.1 (1.8) 21.1 (1.9)
FIG. 2. Global comparison of the RMS evolution of annual mean air temperature (8C) for the
two SAFRAN runs with and without the 43 selected observations.
MARCH 2009 D U R A N D E T A L . 435
which characteristic spatial lengths are smaller than
for temperatures.
For all the 43 sites, contingency tables of daily pre-
cipitation have been compiled to compare obser-
vations (rows in the table) and SAFRAN analyzed
data (columns). Seven classes from near 0 (#0.2
mm) to high precipitation (.40 mm) were defined.
Table 7 shows the result in terms of percentages in
the different classes for the WITH and WITHOUT
simulations. For the WITH results, the value of the
Hansen and Kuiper Skill Score (Wilks 1995) of
0.607 as well as the percentage of well-classified
cases of 68.2% (based on the diagonal elements of
Table 7) are globally quite satisfactory. Concerning
the WITHOUT experiment, the comparisons be-
tween the two sets show that the SAFRAN analy-
ses without using the data of the additional sites are
slightly worse with a Hansen and Kuiper skill score
of 0.575.
In both experiments, SAFRAN analyses give values
lower than measurements, especially for high values of
precipitation. Despite this, these validations globally show
the ability of SAFRAN to reproduce the mountain me-
teorological climatology of precipitation since 1958 but
with a slight bias generally due to local effects. However,
the real-time operational runs show that this does not
affect the results for climatological purposes (Martin et al.
1994; Martin 1995; Quintana-Seguı
´et al. 2008).
c. Analyzed temperature and precipitation trends
Before discussing the final SAFRAN output analyzed
on the massif scale in terms of temporal trends for
precipitation and temperature, a quick overview of the
observation series will be provided to evaluate the
modeled results. The purpose is to assess the possibility
of drawing conclusions with only the modeled fields at
locations or areas where no observations are available.
For this, 21 observation sites with more than 10 000 data
(two observations per day) were selected from the ini-
tial list of 43 sites previously used (Table 4). For ex-
ample, Fig. 3 shows the annual observed temperature
for representative locations of the three main geo-
graphical areas: Nice for the southern Alps, Annecy for
the northern Alps, and Villard-de-Lans for the central
Alps. All locations exhibit a clear temperature increase
over the past 40 yr of about 1.58C for Annecy and Nice
and 1.18C for Villard-de-Lans (with higher variability).
These results, required for the same period as the
modeled results, are relevant for the last 45 yr, but
cannot be extrapolated to longer periods such as the
whole century. They exclude particularly the important
TABLE 6. Illustration of the spatial variability of the SAFRAN analyses and observations for the precipitation parameter (mm day
21
);
the table shows for seven sites the observed and modeled mean values, the averaged differences (bias), the standard deviations (std), and
correlations over the whole period and the two experiment sets (WITH and WITHOUT).
Site
No. of
cases
Mean
observed
values
(mm day
21
)
WITH WITHOUT
Mean
analyzed
values
(mm day
21
)
Difference bias
(std) Correlation
Mean
analyzed
values
(mm day
21
)
Difference bias
(std) Correlation
Chamonix 7881 6.98 8.56 21.58 (4.22) 0.91 8.52 21.54 (4.58) 0.90
Bessans 7591 5.18 5.91 20.73 (3.94) 0.90 5.47 20.29 (4.80) 0.84
Villard de Lans 7874 7.47 7.8 20.33 (3.95) 0.93 7.92 20.45 (4.76) 0.89
Lus La Croix
Haute
7268 6.66 7.02 20.36 (4.45) 0.90 6.7 20.40 (4.71) 0.89
Orcie
`res 6156 8.87 8.38 10.49 (4.30) 0.94 8.27 10.60 (4.80) 0.92
Barcelonnette 5814 4.72 6.35 21.63 (4.76) 0.92 6.39 21.67 (4.82) 0.92
Luceram 1280 5.38 5.38 22.38 (9.62) 0.77 7.73 22.35 (10.21) 0.73
TABLE 7. Daily precipitation contingency table between obser-
vations (lines) and SAFRAN (columns) for the two experiments
sets (WITH and WITHOUT) according to seven classes (mm
day
21
).
WITH
#0.2
mm
0.2–2
mm
2–5
mm
5–10
mm
10–20
mm
20–40
mm
.40
mm
#0.2 mm 81.5 16.4 1.7 0.4 0.1 0.0 0.0
0.2–2 mm 9.5 60.8 23.4 5.5 0.8 0.1 0.0
2–5 mm 0.6 18.8 47.5 27.8 5.0 0.3 0.0
5–10 mm 0.2 2.9 19.9 50.6 24.8 1.6 0.0
10–20 mm 0.1 0.4 2.8 19.1 61.0 16.4 0.2
20–40 mm 0.1 0.1 0.3 2.3 25.2 65.5 6.5
.40 mm 0.3 0.3 0.3 0.4 2.6 38.1 58.1
WITHOUT
#0.2 mm 81.5 16.0 1.9 0.5 0.1 0.0 0.0
0.2–2 mm 12.7 56.6 22.9 6.5 1.2 0.1 0.0
2–5 mm 1.5 21.4 43.3 27.1 6.2 0.4 0.0
5–10 mm 0.4 4.8 21.1 46.2 25.3 2.1 0.1
10–20 mm 0.3 1.0 4.0 20.1 57.3 17.2 0.3
20–40 mm 0.1 0.3 0.7 3.2 26.5 62.4 6.9
.40 mm 0.1 0.2 0.3 0.8 3.7 39.3 56.4
436 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
1940s and 1950s decades, as can be seen when com-
paring Figs. 3c and 3d for the Annecy series.
Figure 4a exhibits the mean temperature trend for 21
observation sites and the corresponding SAFRAN an-
alyzed values downscaled at these points. As explained,
it is difficult to simulate precise geographical locations
with the modeled results, which do not take into account
small-scale orographic effects at these locations. In ad-
dition, some observation sites have also been greatly
influenced by surrounding urbanization, as is probably
the case for Megeve. However, the mean air tempera-
ture trends for the 21 sites are 0.0258Cyr
21
for the
observed data and 0.0288Cyr
21
for the SAFRAN
simulated data. The results are therefore of the same order
of magnitude even if the analysis overestimates the values
for many points such as those in the Vercors massif.
Similar remarks can be made for the precipitation
trends shown in Fig. 4b, especially in the northern Alps
where both SAFRAN analyses and observations show a
small temporal increase, often overestimated by the
model. Trends are rather weak in the southern Alps for
this parameter. As very few observation series cover the
full temporal period and as the observations are above
all representative of the winter season, these values are
difficult to interpret both in time and spatially. The
mean precipitation trends, presented in Fig. 4b, are
11.6 mm yr
21
for the observed data and 12.6 mm yr
21
for
the SAFRAN simulated data. Note the clear latitudinal
difference with a positive trend both for the observa-
tions of the northern Alps (those from about 1 to 30 on
the xaxis in Fig. 4b) and the SAFRAN results, and no
real trend in the south. However, even if the positive
FIG. 3. Annual mean air temperature with its increase and temporal trend for three representative observed series locations: (a) Nice,
(b) Villard-de-Lans, and (c),(d) Annecy.
MARCH 2009 D U R A N D E T A L . 437
trend values are consistent with those of Fig. 9, they are
not statistically significant. Moisselin et al. (2002) point
out the lack of consistency and of significance of most of
the observed precipitation series over the southeast of
France and these data are the main SAFRAN inputs. It
is therefore impossible to draw valid conclusions on the
scale of the observation site concerning these trends.
However, considering cross validations only, which is
our purpose here, we observe a consistent positive trend
for precipitation in the northern Alps and no trend
in the southern Alps, both for observed data and
SAFRAN results.
5. Massif-scale climatology
a. SAFRAN air temperature climatology
The climate of the French Alps is temperate (Fig. 5)
with annual mean air temperature at 1800 m MSL
varying from 3.48C in the north (Chablais massif) to
5.18C in the south (Mercantour massif) near the Medi-
terranean Sea. This latitudinal variability is globally
consistent with that observed for France as a whole
at lower elevations (annual mean air temperature of
12.98C for the city of Toulouse in the south and 10.08C
for the city of Lille, 850 km to the north). The variations
FIG. 4. Observed and SAFRAN-analyzed annual mean (a) temperature and (b) precipitation trends for 21 sites in the French Alps
since 1958.
438 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
are slightly higher in winter (from 21.48to 10.48C; Fig.
5, top right panel) than in summer (from 18.38to
19.98C; Fig. 5, bottom right panel). This low variability
with latitude over these two seasons is partially due to
the fact that results shown concern near surface condi-
tions at a constant midaltitude elevation, which implies
a partial influence of the more smoothed free atmos-
phere conditions. In addition, the strongest latitudinal
gradient occurs mainly during the intermediate seasons
according to the latitudinal variations of the polar front
over France.
b. SAFRAN precipitation climatology
The climate of the French Alps is mainly determined
by a northwesterly atmospheric flow as can seen in
Fig. 6, which shows annual mean precipitation at the
massif scale. This influence is visible both in summer
and winter (rhs of Fig. 6), with more convective pre-
cipitation in summer. Year-to-year variability of annual
precipitation can be very high (commonly 100% for
annual data and much more seasonally) and regional
trends exist (next Fig. 9b). Frei and Scha
¨r (1998), along
with Beniston (2005), also insist on the dynamic inter-
action between weather systems and mountains and on
the influence of sea moisture, especially during south-
erly conditions. At 1800 m MSL, the maximum annual
precipitation amounts to nearly 2000 mm in the north-
western foothills (particularly Chartreuse and Aravis),
and decreases to less than half that amount toward the
southeast (831 mm for Queyras). A secondary maxi-
mum is located in the extreme southeast associated with
the occurrence of northward Mediterranean flows. Note
the small difference between summer and winter in the
massif precipitation distribution despite the differences
in meteorological patterns and types of precipitation.
The snow fraction is about half in the northwest and
only one-third in the south. In this respect, the three
southernmost massifs (Ubaye, Alpes-Azure
´ennes, and
FIG. 5. SAFRAN (left) annual and (right) (top) winter and (bottom) summer mean values for air temperature at an elevation of 1800 m
with the same color code for each.
MARCH 2009 D U R A N D E T A L . 439
Mercantour) get less snow than the overall driest massif
(Queyras); Ubaye receives only an average of around
280 mm of snow water equivalent, as compared with
944 mm of total precipitation.
c. Mean vertical gradients
Four main areas of the Alps (Fig. 7a) regrouping the
different massifs of Fig. 1 have been determined by
expert meteorological clustering. The first split sepa-
rates areas of greatest dissimilarity and divides roughly
the northern from the southern areas. This line does not
really run W–E, but rather SW–NE. Note that Haute-
Maurienne in the central east has a pronounced south-
ern influence. The second split separates the north-
western foothills from the central ranges. Whereas
Belledonne, Beaufortin, and Mont-Blanc have a rather
foothill character, Vercors (the southernmost foothill
massif) resembles more the central massifs. The third
split divides the southern Alps, notably with Ubaye in-
cluded in the far south. These subdivisions can be con-
sidered logical, except perhaps for Belledonne, which
has close ties to Chablais-Mt Blanc (rather than to its
immediate neighbors), and De
´voluy that is closely re-
lated to Queyras-Parpaillon.
For each area (Fig. 7a) thus representative of the
snowpack conditions, low-atmosphere vertical gradients
have been computed for the near-surface temperature
(e.g., the massif averaged air temperature at 2 m at
different surface elevations) and for the annual rainfall.
As shown in Figs. 7b,c, these gradients are very linear
over the entire area. Note that this is not imposed by the
analysis scheme (Durand et al. 1993) and results directly
from the processing of the observations and the ERA-
40 fields. From north to south, the mean near-surface
vertical temperature gradient varies from 25.08to 25.58C
(1000 m)
21
. These rates are very close to those com-
puted by Rolland (2003) over the Italian Alps. The
annual vertical rainfall gradients exhibit a larger lat-
itudinal dependence with respective values from north
to south of 294, 195, 172, and 178 (1000 m)
21
.
FIG. 6. As in Fig. 5, but for precipitation.
440 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
6. Temperature and precipitation trends
SAFRAN analyzed temperature and precipitation
trends are shown in detail for the entire French Alps and
the different areas defined in Fig. 7a at midaltitude (1800
m MSL). When appropriate, the situations for particular
massifs and at other elevations will be also discussed.
All the SAFRAN analyzed values are compared with
the North Atlantic Oscillation (NAO) index
2
(freely
available online at http://www.cpc.noaa.gov/products/
precip/CWlink/pna/nao_index.html) variations over the
whole study period to better explain observed features.
As explained by Beniston (2005), NAO is well repre-
sentative of the decadal-scale variability of the climate
in the Alps, especially at high elevation. Even if its in-
fluence is more pronounced during the winter season
when the westerly meteorological flows are more in-
tense, all the results presented here cover the complete
year. In addition, the daily variability of all the involved
parameters is such that a temporal filter has to be used
to remove interannual noise. Figure 8 shows the varia-
tion of the correlation coefficient between the daily
NAO index and spatially averaged SAFRAN parame-
ters at 1800 m MSL according to different temporal
FIG. 7. (a) The four main areas of the Alps and the SAFRAN-averaged vertical gradient for the (b) near-surface air temperature and
(c) annual rainfall.
2
The NAO index used here is the one computed daily by the
National Oceanic and Atmospheric Administration/National
Weather Service/Climate Prediction Center and is constructed by
projecting the daily (0000 UTC) 500-hPa height anomalies over
the Northern Hemisphere onto the first leading modes of the ro-
tated empirical orthogonal function (REOF) analysis of monthly-
mean 500-hPa heights over the 1950–2000 period.
MARCH 2009 D U R A N D E T A L . 441
FIG. 8. Correlations (vertical axis) between the daily NAO index and spatially averaged SAFRAN parameters at 1800 m MSL with
different temporal filter lengths (horizontal axis, in days): (a) temperatures over the entire French Alps and the other areas defined in Fig.
7a and (b) precipitation of the same areas with the same color code.
442 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
filter lengths for temperature (Fig. 8a) and precipitation
(Fig. 8b) and for the different geographical areas pre-
viously defined in Fig. 7a. We can see that without the
filtering of several years (see the xaxis in days), the
correlations are not significant. On the other hand, long
filtering periods are impossible with our sample of about
45 yr. However, we can observe that for temperature
(Fig. 8a), a minimal total sampling interval of 4 yr is
necessary to reach a minimum correlation over the en-
tire Alps and their northern and central areas whereas
the southern Alps exhibit weaker values. Precipitation
(Fig. 8b) does not show any significant value, especially
in the southern areas. These features have already been
identified by several authors including Beniston (2005,
and references therein) using observed series and are
mainly due to the lower influence of the strong Atlantic
flows on the southern Alps climate. The discrepancy
between temperature and precipitation is mainly due to
their different horizontal characteristic scales.
For this part of the study, 3 supplementary years of
analyzed values have been added to the previous 44
available years, extending the study period to 2005. This
was done because we thought it was important to take
into account in our results the severe decrease of the NAO
index during these years and the corresponding tem-
perature variations. However, given that no ERA-40
guess field was available at the time for SAFRAN, we
used the operational daily ARPEGE fields (Courtier
et al. 1991) as described in Durand et al. (1999).
The total 4-yr sampling interval (triangular shape, 2 yr
before, 2 yr after) was chosen for all the following ana-
lyses and figures concerning the annual variability of
1800 m MSL near-surface temperature and precipita-
tion over different areas of the Alps and for the NAO
index. A larger value would not have been significant
given the 48 available years. In this section, total pre-
cipitation (rain and snow) is mainly discussed.
a. Temperature trends
SAFRAN filtered daily temperatures over several
areas are presented in Fig. 9a together with the filtered
NAO index. The mean values over the entire Alps
(black curve) exhibit the classical shape of the last years
characterized by a plateau until the 1970s, followed by a
more pronounced increase of about 118C. All the dif-
ferent regional areas show the same features modulated
by the latitudinal variability and temporal smoothing.
These characteristics have already been pointed out by
Trenberth et al. (2007) for a larger spatial scale but with
the same magnitude for the temperature trend, and are
mainly due to the increase of the daily minimum tem-
peratures as quoted by Moisselin et al. (2002) and
Beniston (2005). On a large scale, Pro
¨mmel et al. (2007)
have identified a relationship between strong westerly
flows across the North Atlantic and a positive NAO
index, which results in a correlation between NAO and
air temperature that is quite large and positive in the
north of the Alps but smaller in the south. Beniston and
Jungo (2002) widely studied the relationship between
NAO index and the temperature and pressure fields in
Switzerland with a particular emphasis on the increased
values since the 1980s. They successfully linked high-
value NAO periods over the entire Alps to high pressure
blocking events accompanied by vertical circulation in-
ducing compressional warming, subsiding velocities, and
decreasing cloudiness and thus positive temperature
anomalies. Scherrer et al. (2006) have also shown an
enhanced occurrence of blocking-type high pressure
systems over Europe and its link with NAO. According
to Fig. 8a, the correlation between filtered NAO index
and temperature over our working area is about 0.7, which
corroborates the previous results and indicates a mutual
influence on this finer scale. The latitudinal variability is
consistent with the discussion in Pro
¨mmel et al. (2007).
Detailed results (not shown here) show an overall rise
of about 11.58C for Chablais (the northernmost French
massif) over the last 30 yr. The winter half-year increase
was almost 128C and the summer increase about 11.58C
with a constant very limited variation but higher
variability during late summer. All foothill massifs
(Chablais–Vercors) including Mont-Blanc, Beaufortin,
and Belledonne show in general the same behavior.
Chartreuse is the most extreme massif with a net
winter rise of almost 12.58C. The Mercantour massif is
well representative of the central and southern massifs.
The most striking difference to the northern massifs is
a strong temperature decrease in early winter (228C)
since the mid-1980s followed by only a slight increase
in midwinter but an increasing trend in late winter (up
to 138C), which implies only a slight increase (10.58C)
on the scale of the overall winter season. All central
and southern massifs roughly follow this pattern with
Haute Tarentaise-Vanoise-Maurienne being the least
distinctive and Queyras-Parpaillon-Ubaye being the
most pronounced.
b. Precipitation trends (rain and snow)
Figure 9b shows the same results for precipitation. As
in several other studies, no clear temporal trend or clear
relationship with the NAO index can be found for any
of the concerned areas that exhibit only a clear lat-
itudinal variability between the northern and southern
Alps. A linear fitting procedure was performed for the
curves representative of the northern and southern
areas in relation to the validation process presented in
Fig. 4 and the very small trends observed in the northern
MARCH 2009 D U R A N D E T A L . 443
observations. However, these indications of a possible
small positive trend in the north and of a very flat shape
in the south are not statistically significant and allow no
conclusions to be drawn.
These results could appear to be contrary to other
studies such as that of Quadrelli et al. (2001), who
showed, over a much larger area of the Alps (about 15
times bigger than ours), a clear negative correlation
between NAO index and the first component of an EOF
decomposition of the winter precipitation field. In fact,
the numerous differences with our experiment, in par-
ticular our study area for which the main climatological
FIG. 9. Annual variability of the NAO index (right vertical axis) and spatially averaged SAFRAN values at 1800 m MSL with a
temporal filter of 4 yr: (a) temperatures for the entire French Alps and the other areas defined in Fig. 7a; and (b) precipitation for the
same areas with the same color code. In addition, a linear fit for the northern and southern areas is shown in (b).
444 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
precipitation features are more represented by their
second EOF component and the use of yearly precipi-
tation at midaltitudes (1800 m MSL) over 44 yr, make
comparisons and conclusions difficult because of these
scale and decomposition effects.
However, all these considerations and study com-
parisons lead to questions, especially for this precipita-
tion parameter presenting high local variability and
links with large-scale circulation patterns that are diffi-
cult to state. Pro
¨mmel et al. (2007) mention the north-
ern deviation of the westerly winds in the southern Alps
together with decreased precipitation amounts. In their
study of accumulated new snow totals (a quantity rela-
tively well related to precipitation) over the Swiss Alps,
Scherrer and Appenzeller (2006) observe a correlation
of their first orthogonal mode with surface pressure
anomalies over southeastern Europe. Quadrelli et al.
(2001), over their large area, show a good correlation
between their first precipitation mode and north–south
fluctuations of the Atlantic midlatitude westerlies,
whereas their second mode, much less correlated to
NAO, is more influenced by northwest flows. Indeed,
the latter meteorological situations are basically the
rainiest over our working area (Fig. 6) during winter,
whereas the summer season is more influenced by con-
vection. These two points can partially explain our very
weak correlation with NAO.
Detailed results (not shown here) show flat mean
shapes in Chablais both for winter and summer seasons
but with larger interannual and interseasonal variations.
The Grande Rousses massif presents a significant in-
crease during the summer period (;70 mm per decade)
and is one of the only massifs to show a small positive
trend, whereas Mercantour shows a small negative trend
especially during the winter season. However, Chablais
presents two extreme values for the last two winters
(2001 and 2002) and Mercantour includes three very
high values during recent winters (1997, 1998, 2001)
while its snowfall rises to a high point at the end of the
1970s before dropping.
c. Annual distribution
The annual distribution of monthly mean tempera-
ture and precipitation at 1800 m MSL is presented in
Fig. 10. The previously seen marked temperature in-
crease is clear in Fig. 10a (left: entire Alps; top right:
northern Alps; bottom right: southern Alps) both in
winter and summer seasons. The winter season exhibits
fewer cold events, begins a bit later in the north, and
ends earlier in the south. The summer season becomes
clearly warmer over a longer time. The transition period
between winter and summer temperatures appears to be
decreasing (as shown by the ‘‘green’’ area) in all regions,
which denotes shorter intermediate seasons, consistent
with the present personal feelings of many inhabitants.
The precipitation (Fig. 10b, same areas as in Fig. 10a)
does not exhibit any temporal structure and we see
mainly the latitudinal gradient as well as the main fea-
tures of the southern Alps: dry in summer and winter and
storms in autumn. Particularly in winter, year-to-year
variability can be very high and appears to have in-
creased even more in recent years. While the last de-
cade is generally marked by low precipitation, some
outstanding maximum years clearly stand out. The early
winters of 1997, 1998, and 2001 have beaten all records
in the south, but below 2000 m MSL precipitation fell
predominantly in the form of rain. The midwinters of
1995 and 1999 brought record precipitation in the north
falling as snow down to 1000 m MSL and the far south
received large snow amounts down to low levels in 1993
and 1995. The year 2001 was an outstanding year for
late-winter record snowfalls throughout the French
Alps except in the far south (Mercantour, Alpes-
Azure
´ennes) and Haute-Maurienne in the east. Even if
it is standard to split the Alps into a northern and
southern part, the central massifs, in particular, can
show major deviations. Particularly in early summer,
these central massifs can differ considerably from both
northern and southern massifs (results not illustrated
here), showing a strong increase (up to 100% over the
whole period for Grandes-Rousses and Pelvoux).
d. Vertical trends
At different elevations (from 600 to 3600 m MSL),
Table 8 presents, among other features, the Spearman’s
rank correlation coefficient ‘‘r’’ computed for the daily
near-surface SAFRAN analyzed temperatures over the
entire area of the Alps for the new 47-yr period. This
coefficient is simply a special case of the Pearson product-
moment coefficient in which the data are converted to
rankings before calculation, and has been widely used
by many authors such as Moisselin et al. (2002) for trend
detection. Here, its vertical variation shows a clear
positive increase with time especially at midelevations
(1500–2000 m MSL). The corresponding significance
has been evaluated through a Student’s ttest with a
95% confidence interval [the corresponding tvalues are
presented in Table 8 in the t(r) column]. As the t
threshold corresponding to our sample size is about 2,
all levels except the highest (3600 m MSL) present a
significant positive near-surface temperature increase
over the limited study period. Even though Spearman’s
method does not require the assumption that the re-
lationship between the variables is linear, many studies
(such as Trenberth et al. 2007 and references therein)
have computed linear fits, but generally over longer
MARCH 2009 D U R A N D E T A L . 445
FIG. 10. Annual distribution of SAFRAN monthly-mean values at 1800 m MSL over the entire French Alps (23 massifs) for (a)
temperature (left: entire Alps; top right: northern Alps; bottom right: lower southern Alps) and (b) daily mean values for precipitation
(same geographical display). Years on horizontal axis and months on vertical axis.
446 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
periods. We have also determined such fits at the dif-
ferent elevations despite the 47 available years, which
implies results only representative of this period. The
limited accuracy of the linear assumption is visible through
the values of the square of the correlation coefficient
(column R
2
in Table 8) between raw and fitted values
where only midelevation values are of little significance.
However, all vertical levels (except the highest) exhibit
a positive linear trend (the a column) corroborated by
their 95% confidence interval (6acolumn) with a clear
emphasis at midelevation and a weaker signal higher.
Discussion of these results is hampered by the char-
acteristics of the analyzed temperature, which here is
representative of the near-surface conditions but at dif-
ferent mountainous elevations. It is therefore the ‘‘sub-
tle’’ result of surface and free atmosphere conditions with
the interaction of the orographic features and effects
such as sun occultation or meteorological-induced cir-
culation. The highest elevations can therefore be as-
sumed to be more representative of the free atmosphere
conditions, which implies a reduced, or very weak, pos-
itive temperature trend. At low elevations, the temper-
ature trend is superimposed on other phenomena such as
valley effects, boundary layer processes, local observa-
tion site characteristics, and less sun radiance, which in-
troduce noise in the positive signal. A complementary
explanation of this vertical variability can be found in the
behavior of the NAO index for which the fluctuations are
linked to pressure field anomalies. Over a large part of
our study period, the observed positive NAO fluctuations
(Fig. 9) are thus representative of more frequent high
pressure situations and of the induced vertical tempera-
ture inversions for which the tops are generally located
within these midelevations. These phenomena could also
be increased by the winter snow cover decrease at these
elevations and by increased summer dryness (not pre-
sented here).
The mean values obtained at midelevations corre-
spond to those given by Trenberth et al. (2007) but with
a larger confidence interval, mainly due to our short
time series. Rebetez and Reinhard (2007) find a slightly
higher value (0.0578Cyr
21
) for 12 Swiss stations over the
1975–2004 period. Beniston and Jungo (2002) also deter-
mined an altitudinal variation of temperature anomalies
with minimum values at low elevations. These results
canalsobecomparedtotheobservedtrendvaluesinFig.4,
which well illustrate the variability in our mountainous
area.
The similar study for precipitation (not shown here)
does not show any significant results for our area over
the same considered time period.
e. Link between temperature and precipitation trends
Looking at snow precipitation trends in the light of
temperature trends reveals that in the north, falling tem-
peratures are associated with slightly rising snowfalls
(early winter) and rising temperatures cause diminish-
ing snowfalls (midwinter–early summer). Constant late-
summer temperatures show no impact on snow precipi-
tation trends, as would be expected. However, the ex-
ample of Mercantour in the far south shows that strongly
dropping early winter temperatures do not necessarily
result in increasing snowfalls, since total precipitation is
also decreasing. At Grandes-Rousses, in the central
part, we see that strongly rising late-winter temperatures
have hardly any effect either on snow or rain precipita-
tion, but a strong early summer temperature increase is
accompanied by a very strong rainfall increase and a
slight snowfall decline. Finally, near-constant late-sum-
mer temperatures are accompanied by a strong positive
rainfall trend but have no effect on snowfall.
Beniston (2003) studied the possible impacts of these
climatic trends in mountainous areas on hydrology, snow
conditions, glacier vegetation, and tourism; he mentions
more particularly several research works with SAFRAN-
Crocus. Some elements of our study are in common with
those of Beniston, especially the uncertainties concern-
ing precipitation and the NAO–temperature link.
7. Summary
The validations presented here and based on the
SAFRAN analysis process show the robustness of the
models used and their ability to reproduce the main
TABLE 8. Temperature trends for the entire Alps study area at
different elevations (from 600 to 3600 m MSL). The ‘‘r’’ column
indicates Spearman’s rank coefficient, and ‘‘t(r)’’ represents the
corresponding Student’s tfunction. The ‘‘a’’ and ‘‘b’’ columns in-
dicate, respectively, the linear trend (8Cyr
21
) and the residual (8C)
of the associated linear fit. The ‘‘6a’’ column (8Cyr
21
) represents
the confidence interval (95%) of the aparameter and the column
R
2
is the square of the correlation coefficient of the linear fit. The
computation is performed over 47 yr.
Alt
(m) rt(r)a(8Cyr
21
)6ab(8C) R
2
600 0.48 3.6 0.020 0.010 9.7 0.24
900 0.49 3.8 0.020 0.010 8.2 0.24
1200 0.63 5.5 0.026 0.011 6.8 0.35
1500 0.70 6.6 0.034 0.011 5.4 0.46
1800 0.70 6.6 0.033 0.010 3.9 0.47
2100 0.66 6.0 0.031 0.010 2.7 0.44
2400 0.61 5.1 0.029 0.010 1.0 0.39
2700 0.49 3.7 0.020 0.010 20.7 0.24
3000 0.41 3.1 0.016 0.011 22.2 0.17
3300 0.35 2.5 0.014 0.011 23.9 0.12
3600 0.22 1.5 0.009 0.010 25.7 0.06
MARCH 2009 D U R A N D E T A L . 447
meteorological features of several mountainous obser-
vation sites even when data are deliberately omitted
from the analyses. The analyzed results on the massif
scale can be considered to be representative of the cli-
matology of the French Alps study area at different
elevations during the considered period.
The annual mean air temperature at 1800 m MSL
varies from 3.48C in the north (Chablais massif) to 5.18C
in the south (Mercantour massif). The variations are
slightly higher in winter (from 21.48to 10.48C) than in
summer (from 18.38to 19.98C).
Year-to-year variability of annual precipitation can be
very high (commonly 100% for annual data and much
more seasonally) and regional trends exist. At 1800 m
MSL, the maximum annual precipitation amounts to
nearly 2000 mm in the northwestern foothills (particularly
Chartreuse and Aravis), and decreases to less than half
that amount toward the southeast (831 mm for Queyras).
A secondarymaximum is located inthe extreme southeast
associated with the occurrence of northward Mediterra-
nean flows.
Low-atmosphere vertical gradients have also been
computed and exhibit a very linear shape over the entire
area. From north to south, the mean near-surface ver-
tical temperature gradient varies from 25.08to 25.58C
(1000 m)
21
. The annual vertical rainfall gradients ex-
hibit a larger latitudinal dependence with values from
north to south of 294, 195, 172, and 178 mm (1000 m)
21
.
In terms of an overall temporal trend for the 1958–
2002 observation period, the annual air temperature rose
by about 18C, mainly during the 1980s and 1990s. However,
variations of this trend are large for different altitudes,
seasons, and regions. The trends are most pronounced
between 1500 and 2000 m MSL and exhibit some rela-
tionships with the NAO variations especially for north-
ern massifs. Temperatures have risen in spring and fallen
in autumn, reducing the intermediate seasons. This
temperature drop in autumn and early winter is also at
the root of the most striking regional differences. A
large year-to-year variability is another common char-
acteristic, often deviating far from smoothed trend lines.
Temperatures have remained relatively homogeneous
at high elevations, without significant trends.
Precipitation variability is very high, making it hard to
detect clear trends. No relationship with NAO has been
detected or any clear tendency or temporal trend. Re-
gional differences split the French Alps into a northern
and southern part. Whereas variations in the north are
greater in summer, the southern massifs show higher
variability in winter.
Acknowledgments. We are grateful to all those who
have developed and used SAFRAN software, helping to
make it a reliable tool. We are also indebted to the
ECMWF, who carried out the ERA-40 simulations that
are the basis for this study, to the NOAA/NWS/CPC for
the daily NAO index, and to several colleagues of
Me
´te
´o-France who helped us to collect and process dif-
ferent climatological series. We also thank numerous
people together with the three anonymous reviewers, the
editor, and the native English translator who all helped
us improve both the language and content of this paper.
REFERENCES
Beniston, M., 2003: Climatic change in mountainous regions: A
review of possible impacts. Climatic Change, 59, 5–31.
——, 2005: Mountain climates and climatic changes: An overview
of processes focusing on the European Alps. Pure Appl.
Geophys., 162, 1587–1606.
——, and P. Jungo, 2002: Shifts in the distribution of pressure,
temperature and moisture and changes in the typical weather
patternsin the Alpine region in response to the behaviour of the
North Atlantic Oscillation. Theor. Appl. Climatol., 71, 29–42.
Berthelot, M., 2004: Variabilite
´interannuelle des pre
´cipitations
dans les Alpes du Sud pour la pe
´riode 1951-2000 (in French).
[Available online at http://perso.wanadoo.fr/alexandre.letort/
berthelot/alpessud.htm.]
Brun, E., E. Martin, V. Simon, C. Gendre, and C. Cole
´ou, 1989: An
energy and mass model of snow cover suitable for operational
avalanche forecasting. J. Glaciol., 35, 333–342.
——, P. David, M. Sudul, and G. Brugnot, 1992: A numerical
model to simulate snow cover stratigraphy for operational
avalanche forecasting. J. Glaciol., 38, 13–22.
Courtier, P., C. Freydier, J. F. Geleyn, F. Rabier, and M. Rochas,
1991: The ARPEGE project at Me
´te
´o-France. Proc. Seminar
on Numerical Methods in Atmospheric Models, Vol. 2, Read-
ing, United Kingdom, ECMWF, 191–212.
Durand, Y., E. Brun, L. Me
´rindol, G. Guyomarch, B. Lesaffre, and
E. Martin, 1993: A meteorological estimation of relevant
parameters for snow models. Ann. Glaciol., 18, 65–71.
——, G. Giraud, E. Brun, L. Me
´rindol, and E. Martin, 1999: A
computer based system simulating snowpack structures as a
tool for regional avalanche forecast. J. Glaciol., 45, 469–485.
ECMWF, 2004: ‘‘ERA-40: ECMWF 45-years reanalysis of the global
atmosphere and surface conditions 1957-2002’’ by Uppala, S.,
and Coauthors. ECMWF Newsletter, No. 101, ECMWF, Read-
ing, United Kingdom, 2–21.
Frei, C., and C. Scha
¨r, 1998: A precipitation climatology of the
Alps from high-resolution rain-gauge observations. Int. J.
Climatol., 18, 873–900.
Fuentes, U., and D. Heimann, 1996: Verification of the statistical-
dynamical downscaling in the Alpine region. Climate Res., 7,
151–168.
Giraud, G., 1993: MEPRA: An expert system for avalanche risk
forecasting. Proc. Int. Snow Science Workshop: A Merging of
Theory and Practice, Breckenridge, CO, American Associa-
tion of Avalanche Professionals, 97–104.
Laternser, M., and M. Schneebeli, 2003: Long-term snow climate
trends of the Swiss Alps (1931-99). Int. J. Climatol., 23, 733–750.
Martin, E., 1995: Mode
´lisation de la climatologie nivale des alpes
francxaises, application des te
´chniques de re
´gionalisation a
`
l’e
´tude de l’impact d’un changement climatique sur l’en-
neigement (Modelling the snow climate of the French Alps,
448 JOURNAL OF APPLIED METEOROLOGY AND CLIMATOLOGY VOLUME 48
application of regionalisation techniques for studying the
impact of climate change on the snow cover). Ph.D. thesis,
Paul Sabatier University, 244 pp.
——, E. Brun, and Y. Durand, 1994: Sensitivity of the French Alps
snow cover to the variation of climatic variables. Ann. Geo-
phys., 12, 469–477.
Moisselin, J. M., M. Schneider, C. Canellas, and O. Mestre, 2002:
Les changements climatiques en France au XXe sie
`cle (Cli-
mate change in France during the 20th century). Meteor-
ologie, 38, 45–56.
Pahaut, E., E. Brun, and G. Brunot, 1991: L’organisation de la
pre
´vision du risque d’avalanches en France. Proceedings of
the Symposium ANENA -CISA-IKAR, Chamonix, 4-8 June
1991, G. Brugnot et al., Eds., ANENA, 50–56.
Pro
¨mmel, K., M. Widmann, and J.-M. Jones, 2007: Analysis of the
(N)AO influence on Alpine temperature using a dense dataset
and high-resolution simulations. Extended Abstracts, 29th Int.
Conf. on Alpine Meteorology, Vol. 1, Chamberg, France,
Me
´te
´o-France, 233–236. [Available from Me
´te
´o-France/CNRM,
42 av. Gaspard Coriolis, 31057 Toulouse, France.]
Quadrelli, R., M. Lazzeri, C. Cacciamani, and S. Tibaldi, 2001:
Observed winter Alpine precipitation variability and links
with large-scale circulation patterns. Climate Res., 17, 275–
284.
Quintana-Seguı
´, P., and Coauthors, 2008: Analysis of near-surface
atmospheric variables: Validation of the SAFRAN analysis
over France. J. Appl. Meteor. Climatol., 47, 92–107.
Rebetez, M., and M. Beniston, 1998: Changes in temperature
variability in relation to shifts in mean temperatures in the
Swiss Alpine region this century. The Impacts of Climate
Variability on Forests, M. Beniston and J. L. Innes, Eds.,
Springer, 49–58.
——, and M. Reinhard, 2007: Monthly air temperature trends in
Switzerland 1901–2000 and 1975–2004. Theor. Appl. Clima-
tol., 91, 27–34, doi:10.1007/s00704-007-0296-2.
Rolland, C., 2003: Spatial and seasonal variations of air tempera-
ture lapse rate in Alpine regions. J. Climate, 16, 1032–1046.
Scherrer, S.-C., and C. Appenzeller, 2006: Swiss Alpine snow pack
variability: Major patterns and links to local climate and large-
scale flow. Climate Res., 32, 187–199.
——, M. Croci-Napoli, C. Schwierz, and C. Appenzeller, 2006:
Two-dimensional indices of atmospheric blocking and their
statistical relationship with winter climate patterns in the
Euro-Atlantic region. Int. J. Climatol., 26, 233–249.
Schmidli, J., C. Schmutz, C. Frei, H. Wanner, and C. Scha
¨r, 2002:
Mesoscale precipitation variability in the region of the Eu-
ropean Alps during the 20th century. Int. J. Climatol., 22,
1049–1074.
Trenberth, K. E., and Coauthors, 2007: Observations: Surface and
atmospheric climate change. Climate Change 2007: The
Physical Science Basis, S. Solomon et al., Eds., Cambridge
University Press, 235–336.
Wilks, D. S., 1995: Statistical Methods in the Atmospheric Sciences:
An Introduction. Academic Press, 467 pp.
MARCH 2009 D U R A N D E T A L . 449
