Contributors: Samuel Morin (METEO-FRANCE), Raphaëlle Samacoïts (METEO-FRANCE), Hugues François (INRAE), Bruno Abegg (HSG)
1. Short description
The C3S European Tourism "Mountain Tourism Meteorological and Snow Indicators" dataset provides pan-European information relevant to snow reliability, hence operating conditions of winter ski resorts under past and future climate conditions. It consists of 39 indicators characterizing atmospheric and snow conditions, with and without snow management (grooming, snowmaking), computed in a similar manner for all mountain regions in Europe at the scale of NUTS-3 regions and by steps of 100 m elevation on flat terrain. The 39 indicators are grouped in the following 7 groups: PR=precipitation indicators, Tas=temperature indicators, SD=snow depth indicators, SWE=snow water equivalent indicators, MM-PROD=amount of snow produced, WBT: wet bulb temperature-based indicators, BS-ES: beginning and end of season indicators. Past climate conditions based on reanalysis data span the period from 1961 to 2015, and future climate
conditions are based on an ensemble of adjusted climate projections spanning the period from 1950 to 2100, with 2006-2100 covered by RCP2.6 (2 GCM/RCM pairs) and RCP4.5 and RCP8.5 (9 GCM/RCM pairs each). This makes it possible to compute not only mean values of indicators for future time periods (2021-2040, 2061-2080 and 2081-2100) but also the multi-model standard deviation and quantiles of annual values (Q10, Q20, Q50, Q80 and Q90), which are critical for the assessment of snow reliability of mountain regions in a multi-annual perspective. The files are provided in NetCDF format.
Figure 1: NUTS-3 regions in Europe where mountain regions are present at 800 m elevation above sea level. Data represented correspond the change between two aggregated indicators, see below for details.
2. Description of the climate modeling chain
2.1. Input data, pre-processing and climate models
MTMSI data are based on the following input data:
- Surface atmospheric fields of the UERRA 5.5 km reanalysis (Soci et al., 2016) were used, for 5652 points selected depending on their location (presence within a mountain NUTS-3 region, see Figure 1) and elevation (by steps of 100 m within NUTS-3 regions). In addition, 932 points were selected to cover all European NUTS-3 regions, at their mean elevation. Note that while UERRA 5.5 km points all have different coordinates, within a given NUTS-3 region they were all assigned the same latitude and longitude (barycenter of the NUTS-3), and the elevation was rounded to the nearest 100 m elevation band. The selection was performed by H. François, INRAE Grenoble, based on the 2018 NUTS-3 inventory and a 30 m resolution DEM for the entire European domain.
- Atmospheric fields of GCM/RCM pairs from the EUROCORDEX dataset were used, as follows: SMHI-RCA4/MOHC-Hadgem2-ES, HIST, RCP4.5, RCP8.5
CNRM-ALADIN53/CNRM-CERFACS-CNRM-CM5, HIST, RCP4.5, RCP8.5
IPSL-INERIS-WRF331F/IPSL-CM5A-MR, HIST, RCP4.5, RCP8.5
MPI-CSC-REMO2009/MPI-M-MPI-ESM-LR, HIST, RCP2.6, RCP4.5, RCP8.5 SMHI-RCA4/ICHEC-EC-EARTH HIST, RCP2.6, RCP4.5, RCP8.5
SMHI-RCA4/CNRM-CERFACS-CNRM-CM5, HIST, RCP4.5, RCP8.5 SMHI-RCA4/IPSL-CM5A-MR, HIST, RCP4.5, RCP8.5
SMHI-RCA4/MPI-M-MPI-ESM-LR, HIST, RCP4.5, RCP8.5 CCLM4-8-17/MPI-M-MPI-ESM-LR, HIST, RCP4.5, RCP8.5
This comprises 20 future climate change scenarios for the 21stcentury. In total, this corresponds to 2305 model years (taking into account the reanalysis).
2.2. Downscaling and adjustment method
The ADAMONT method (Verfaillie et al., 2017) was used to adjust the EUROCORDEX GCM/RCM pairs using the UERRA 5.5 km reanalysis as an observation reference. This method carries out quantile mapping on daily data. Quantile mapping functions are computed as a function of weather type (4 weather types considered, based on synoptic fields from 500 hPa geopotential height for the driving GCM) and season (4 seasons : DJF, MAM, JJA, SON). The adjustment was performed using UERRA 5.5km data from 01/01/1980 to 01/01/2012 and GCM/RCM pairs from 01/01/1974 to 01/01/2006. In the special case of the GCM/RCM pair involving the MOHC-HadGEM2-ES GCM only, here are the time periods used for the adjustment : UERRA 5.5km data from 01/01/1987 to 01/01/2012 and MOHC-HadGEM2-ES/RCM pairs from 01/01/1981 to 01/01/2006).
As indicated above, all UERRA 5.5 km points within a given NUTS-3 are affected the same latitude and longitude, hence they all have the same corresponding GCM/RCM point (for a given GCM/RCM grid geometry). This ensures consistency, within a given NUTS-3 region, of the climate change signal, although this may inhibit potential elevation dependent signals inherited from the GCM/RCM. This approach is similar to previous studies carried out in French mountain regions (Verfaillie et al., 2017, 2018, Spandre et al., 2019).
Following the application of the quantile mapping on daily data, 6-hourly data sets are generated following a disaggregation method using reanalysis data as guess data for the shape of the diurnal cycle (see Verfaillie et al., 2017) for more information.
The code of the ADAMONT v1.0 method is available as an open git repository after free registration at https://opensource.cnrm-game-meteo.fr/projects/adamont.
2.3. Impact model
Reanalysis and adjusted climate projections were used as such for atmospheric indicators (e.g., temperature, wet bulb temperature, precipitation). For snow indicators, the Crocus snowpack model was used (Vionnet et al., 2012). Crocus is a multi-layer snowpack model embedded in the ISBA land surface model within the SURFEX model (Masson et al., 2013). Crocus makes it possible to account for grooming and snowmaking (Spandre et al., 2016), based on physical representation of these snow management practices and operational rules (Spandre et al., 2016, 2019). For the simulations with snowmaking, the wet-bulb temperature threshold for snowmaking was set to -5°C, the maximum wind speed threshold was set to 4.2 m s-1, the density of machine-made snow was set to 600 kg m-3, and the production rate of machine made snow was set to 1.2 10-3 kg m-2 s-1. In the simulations, between November 1 and December 15, a 30 cm deep "base layer" (150 kg m−2, corresponding to 30 cm of snow at 500 kg m−3 typical density on ski slopes) is produced, weather conditions permitting, regardless of natural snowfalls during the period.
Between December 15 and February 28, snow is produced if meteorologically possible so as to maintain a total snow depth of 60 cm. After March 1, no more snow is produced.
The UERRA 5.5 km atmospheric fields and adjusted EUROCORDEX data are available at 6-hourly time resolution, and were used as such to drive SURFEX/ISBA-Crocus. Note that SURFEX was run at 15 min internal time resolution, with appropriate time interpolation of the 6-hourly driving variables (Soci et al., 2016). Crocus output data were stored at daily time resolution, using the value at 6 UTC. Natural snow refers to natural processes only (no grooming nor snowmaking). Groomed snow refers to natural processes and the effects of grooming. Managed snow refers to natural processes and the effects of grooming and snowmaking.
SURFEX is available online through a Cecill-C licence : http://www.umrcnrm.fr/surfex/spip.php?article437
2.4. Sectoral Impact Indicators
A total of 39 annual-scale indicators were generated. Taking into account the multiple datasets used to generate them, this corresponds to 91065 annual scale indicators (with a value for each of the 6584 points). For each indicator, multi-annual/multi-model aggregated values were computed for various 30 and 20 years time periods, described below.
The indicators are grouped in 7 groups
- PR=precipitation
- Tas=temperature
- SD=snow depth
- SWE=snow water equivalent
- MM-PROD=amount of snow produced
- WBT: wet bulb temperature
- BS-ES: beginning and end of season
PR:
snowfall-amount-winter: sum of snow precipitation, from November of year N to following April (included)
precipitation-amount-winter: sum of total precipitation, from November of year N to following April (included)
Tas:
tas-11: mean temperature for November of year N tas-12: mean temperature for December of year N tas-01: mean temperature for January of year N tas-02: mean temperature for February of year N tas-03: mean temperature for March of year N
tas-04: mean temperature for April of year N
tas-winter: mean temperature for November of year N to April of year N+1 (included)
SD:
sd-days-05-NS: Number of days with at least 5 cm of natural snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-05-GS: Number of days with at least 5 cm of groomed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-05-MS: Number of days with at least 5 cm of managed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-30-NS: Number of days with at least 30 cm of natural snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-30-GS: Number of days with at least 30 cm of groomed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-30-MS: Number of days with at least 30 cm of managed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-50-NS: Number of days with at least 50 cm of natural snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-50-GS: Number of days with at least 50 cm of groomed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-50-MS: Number of days with at least 50 cm of managed snow on the ground, starting on August 1st of year N to July 31st of year N+1
sd-days-Xmas-NS: Number of days with at least 30 cm of natural snow on the ground, from December 22 of year N to January 4 (included) of year N+1
sd-days-Xmas-GS: Number of days with at least 30 cm of groomed snow on the ground, from December 22 of year N to January 4 (included) of year N+1
sd-days-Xmas-MS: Number of days with at least 30 cm of managed snow on the ground, from December 22 of year N to January 4 (included) of year N+1
sd-days-PUR-NS: Number of days with at least 30 cm of natural snow on the ground, from December 4 of year N to December 10 of year N (included)
sd-days-PUR-GS: Number of days with at least 30 cm of groomed snow on the ground, from December 4 of year N to December 10 of year N (included)
sd-days-PUR-MS: Number of days with at least 30 cm of managed snow on the ground, from December 4 of year N to December 10 of year N (included)
SWE:
swe-days-100-NS: Number of days with an amount of at least 100 kg m-2 of natural snow on the ground, starting on August 1st of year N to July 31st of year N+1
swe-days-100-GS: Number of days with an amount of at least 100 kg m-2 of groomed snow on the ground, starting on August 1st of year N to July 31st of year N+1
swe-days-100-MS: Number of days with an amount of at least 100 kg m-2 of managed snow on the ground, starting on August 1st of year N to July 31st of year N+1
swe-days-120-NS: Number of days with an amount of at least 120 kg m-2 of natural snow on the ground, starting on August 1st of year N to July 31st of year N+1
swe-days-120-GS: Number of days with an amount of at least 120 kg m-2 of groomed snow on the ground, starting on August 1st of year N to July 31st of year N+1
swe-days-120-MS: Number of days with an amount of at least 120 kg m-2 of managed snow on the ground, starting on August 1st of year N to July 31st of year N+1
MM-PROD:
mm-prod: Annual amount of machine made snow produced (in kg m-2), from August 1st of year N to July 31st of year N+1
WBT:
wbt-2-hrs: Early season potential snowmaking hours (for wet bulb temperature lower than -2°C), from November 1st, year N to December 31st, year N.
wbt-5-hrs: Early season potential snowmaking hours (for wet bulb temperature lower than -5°C), from November 1st, year N to December 31st, year N.
BS-ES:
beginning-season-30-NS: Beginning of season, i.e. first date of the longest continuous period with at least 30 cm of natural snow on the ground (from August 1st of year N to July 31st of year N+1)
end-season-30-NS: End of season, i.e. last date of the longest continuous period with at least 30 cm of natural snow on the ground (from August 1st of year N to July 31st of year N+1)
beginning-season-30-GS: Beginning of season, i.e. first date of the longest continuous period with at least 30 cm of groomed snow on the ground (from August 1st of year N to July 31st of year N+1)
end-season-30-GS: End of season, i.e. last date of the longest continuous period with at least 30 cm of groomed snow on the ground (from August 1st of year N to July 31st of year N+1)
beginning-season-30-MS: Beginning of season, i.e. first date of the longest continuous period with at least 30 cm of managed snow on the ground (from August 1st of year N to July 31st of year N+1) end-season-30-MS: End of season, i.e. last date of the longest continuous period with at least 30 cm of managed snow on the ground (from August 1st of year N to July 31st of year N+1)
Aggregated data were computed as follows:
- Values for the period 1961-1990 and 1990-2015 based on reanalysis data.
- Values for the period 1986-2005, 2021-2040, 2061-2080 and 2081-2100 for GCM/RCM data.
For each of these 20 or 30 years periods, the following statistics were computed:
- Mean and standard deviation (across GCM/RCM pairs for a given RCP) for multi-annual averages
- Quantile of annual values (Q10, Q20, Q50, Q80 and Q90) across all available GCM/RCM pairs for a given time period and RCP.
Note that the mean and standard deviation for RCP2.6 values are only based on 2 available GCM/RCM pairs (9 pairs for RCP4.5 and RCP8.5).
2.5. File formats
File containing the indicators are provided in netCDF files. Aggregated and annual files share the same structure. There is one coordinate for the locations (referred to as "Number_of_points"), with 6584 points. Each location is characterized by a lat/lon (for the centroid of the NUTS-3), and elevation. The NUTS3_ID variable (which has "Number_of_points, nchar") as a dimension and uses 5 characters, makes it possible to assign each location to a NUTS-3 and elevation. There is either one file per year, or one file per aggregated time period.
Below is an example of a netCDF file structure (output of a ncdump command). All standard data processing software (python, R, FORTRAN, matlab etc.) have netCDF reading capabilities.
netcdf tas-03_CNRM-ALADIN53.CNRM-CERFACS-CNRM-CM5_RCP45_2033 {
dimensions:
Number_of_points = 6584 ; nchar = 5 ;
time = UNLIMITED ; // (1 currently)
variables:
float LAT(Number_of_points) ; LAT:units = "degrees_north" ; LAT:_FillValue = -9999999.f ; LAT:long_name = "latitude" ;
float LON(Number_of_points) ; LON:units = "degrees_east" ; LON:_FillValue = -9999999.f ; LON:long_name = "longitude" ;
char NUTS3_ID(Number_of_points, nchar) ; NUTS3_ID:long_name = "NUTS regions" ; NUTS3_ID:out_name = "nuts" ; NUTS3_ID:type = "character" ;
float ZS(Number_of_points) ; ZS:units = "m" ; ZS:_FillValue = -9999999.f ; ZS:long_name = "altitude" ;
float tas(time, Number_of_points) ; tas:_FillValue = -9999999.f ; tas:missing_value = -9999999.f ; tas:standard_name = "air_temperature" ; tas:units = "K" ;
tas:long_name = "Mean air temperature for a given month or time period" ; tas:comment = "Average of 6-hourly air temperature for all dates in a given month or
period of year N" ;
tas:out_name = "tas" ; tas:type = "real" ;
float time(time) ; time:standard_name = "time" ; time:long_name = "time" ;
time:units = "hours since 2032-08-01 06:00:00" ; time:calendar = "standard" ;
time:axis = "T" ;
// global attributes:
:CDI = "Climate Data Interface version 1.8.2 (http://mpimet.mpg.de/cdi)" ;
:Conventions = "CF-1.7" ;
:institution = "Centre National de Recherches Meteorologiques" ;
:contact = "samuel.morin@meteo.fr" ;
:experiment_id = "RCP45" ;
:experiment = "REF" ;
:driving_model_id = "CNRM-CERFACS-CNRM-CM5" ;
:driving_model_ensemble_member = "r1i1p1" ;
:frequency = "year" ;
:institute_id = "CNRM" ;
:model_id = "CNRM-ALADIN53" ;
:project_id = "EUROCORDEX" ;
:domain = "EUROCORDEX plaine" ;
:product = "output" ;
:references = "https://opensource.cnrm-game-meteo.fr/projects/adamont" ;
:NCO = "\"4.5.5\"" ;
:nco_openmp_thread_number = 1 ;
:CDO = "Climate Data Operators version 1.8.2 (http://mpimet.mpg.de/cdo)" ;
:creation_date = "2019-03-27T22:01:53Z" ;
:institute = "Centre National de Recherches Météorologiques" ;
}
2.6. Uncertainty or skill assessment
The chain of models used for this product was tested in previous applications, using similar climate projection input data (Verfaillie et al., 2018, Spandre et al., 2019) and are considered state-of-the- art. The UERRA 5.5 reanalysis has undergone evaluation as part of the UERRA European Project although a published evaluation in mountainous areas is still missing. It is considered that, due to the larger amount of input data, the use of a true surface reanalysis system (including precipitation analysis), and the higher resolution (5.5 km) than previous products (e.g, E-OBS), that this dataset is more fit-for-purpose than alternative products at the European scale. Nevertheless, the quantity and quality of input data to the UERRA 5.5 km varies across the domain, so that heterogeneities in the dataset are likely. This probably constitutes the best available data set at European scale, still subject to significant improvements for the future.
Uncertainties from climate projections are accounted for using a multi-model approach, which uses a larger number of GCM/RCM pairs than previous studies at the pan-European scale, and comparable number to local studies (e.g. Marke et al., 2015, Spandre et al., 2019).
The snowpack model Crocus features excellent performances in model intercomparison exercises (Krinner et al., 2018). Within the scope of this product, the main limitation to the use of the Crocus model is the similar snow management parameters (grooming, snowmaking) across Europe, although regional variations in management practices already occur and are likely to evolve in the 21st century, due to changes in technology and adaptation of management practices to climate change.
2.7. Known issues/caveats
The main issue/caveat identified on this data set is related to the geographical setting employed. Indeed, there is a trade-off between the representation of spatial variability within mountain ranges and the specification of a pan-European product with features as homogeneous as possible. In order to operate on a manageable number of points, while representing the elevation dependence of changes in the mountain environment, the choice was made to define the indicators for NUTS-3 regions. This is consistent with previous studies, and makes it possible to combine the indicators with other socio-economic indicators. However, given the size of NUTS-3 regions, it was necessary
to select points, in order to represent 100 m elevation bands, which are sometimes distant of more than 100 km. NUTS-3 are purely administrative borders, which implies that they sometimes do not align with the physical geography.
Figure 2 shows an example of the selection of UERRA 5.5 km points for the Vaud canton in Switzerland (NUTS-3 CH011). It shows that sometimes UERRA-5.5 km points used for different elevation bands can be located quite far away.
While for each point the climate projection information remains adequate, in such cases, the vertical lapse rate of the indicators can be nonlinear, because, while elevation is the primary control for mountain climate, smaller scale processes operate within NUTS-3 regions. Sometimes, points for a given elevation were selected outside the strict boundaries of NUTS-3 regions (sometimes the same point was used for several NUTS-3, but the climate projections are different between difference points from the EUROCORDEX GCM/RCM pairs were used).
Alternatives for this caveat were sought, but not found within the time frame of the production schedule. For example, it was not possible to combine several points within a given NUTS-3 for a given elevation and perform some spatial averaging: this would have smoothed out meteorological conditions, in particular precipitation events, and resulted in unrealistic snow model outputs both for reanalysis and climate projections. As a consequence, care should be exercised when analyzing small scale features such as the elevation dependence of the indicators within a given NUTS-3 or differences between neighbouring NUTS-3 regions for a given elevation.
Example of UERRA 5.5 km point selection for various elevation levels for the CH011 (Vaud canton) NUTS-3, located in Switzerland.
2.8. Use cases and potential further developments
The C3S European Tourism MTMSI is not meant to replace higher resolution products which are available in some European countries, and provide a more detailed view of the future of snow conditions in European ski resorts, accounting, for example, for slope, aspect, local phenomena and local snow management practices. However, given that the workflow for the generation of the product is homogeneous at the pan-European level, it is useful to compare the main features of past and future snow conditions at the pan-European level, or to compare distant destinations (e.g., compare Scandinavia and Eastern Europe for a given elevation and time horizon). Furthermore, where no other source of information is available, it provides an original outlook on future meteorological and snow conditions in mountain regions.
It is the first version of a pan-European product for mountain (ski) tourism under climate change, which holds significant potential for applications but can be improved in several areas, such as:
- Refined methodology for location/elevation issues
- Improvements of the regional reanalysis and availability of more GCM/RCM pairs
- Improvements of the adjustment method
- Improvements of the snowpack model.
Additional indicators could be developed for year-round mountain tourism, or for complementary winter tourism indicators.
Product versioning
Time aggregation | Variable | Experiment | Regional climate model | Global climate model | Period | Multi-model statistic | Year | Issue | Solved in version(s)(date published) | Version to use |
|---|---|---|---|---|---|---|---|---|---|---|
| Annual data | Annual amount of machine made snow produced | RCP4.5 | WRF331F (IPSL, France) | CM5A-MR (IPSL, France) | - | - | 2038 | Corrupted file | 20210318 | v1.1 |
| Climatology | Start and end of season, natural, groomed and managed snow | All experiments | - | - | All periods | All statistics | - | Wrong computation of statistics at low elevation or poor snow cover | 20210318 | v1.1 |
References
Krinner, G., Derksen, C., Essery, R., Flanner, M., Hagemann, S., Clark, M., Hall, A., Rott, H., Brutel- Vuilmet, C., Kim, H., Ménard, C. B., Mudryk, L., Thackeray, C., Wang, L., Arduini, G., Balsamo, G., Bartlett, P., Boike, J., Boone, A., Chéruy, F., Colin, J., Cuntz, M., Dai, Y., Decharme, B., Derry, J., Ducharne, A., Dutra, E., Fang, X., Fierz, C., Ghattas, J., Gusev, Y., Haverd, V., Kontu, A., Lafaysse, M., Law, R., Lawrence, D., Li, W., Marke, T., Marks, D., Ménégoz, M., Nasonova, O., Nitta, T., Niwano, M., Pomeroy, J., Raleigh, M. S., Schaedler, G., Semenov, V., Smirnova, T. G., Stacke, T., Strasser, U., Svenson, S., Turkov, D., Wang, T., Wever, N., Yuan, H., Zhou, W., and Zhu, D.: ESM-SnowMIP: assessingsnowmodels and quantifyingsnow-relatedclimate feedbacks, Geosci. Model Dev., 11, 5027- 5049, https://doi.org/10.5194/gmd-11-5027-2018, 2018.
Marke, T., U. Strasser, F. Hanzer, J. Stötter, R.A. Wilcke, and A. Gobiet, Scenarios of Future Snow Conditions in Styria (Austrian Alps).J. Hydrometeor.,16, 261–277, https://doi.org/10.1175/JHM-D-14-0035.1, 2015.
Soci, C., Bazile, E., Besson, F., and Landelius, T., 2016. High-resolution precipitation re-analysis system for climatological purposes, Tellus A, 68, 29879.
Spandre, P., S. Morin, M. Lafaysse, Y. Lejeune, H. François and E. George-Marcelpoil, Integration of snow management processes into a detailed snowpack model, Cold Reg. Sci. Technol., 125, 48- 64, doi:10.1016/j.coldregions.2016.01.002, 2016.
Spandre, P., François, H., Verfaillie, D., Pons, M., Vernay, M., Lafaysse, M., George, E., and Morin, S.: Winter tourism under climate change in the Pyrenees and the French Alps: relevance of snowmaking as a technical adaptation, The Cryosphere, 13, 1325-1347, https://doi.org/10.5194/tc-13-1325-2019, 2019.
Verfaillie, D., Déqué, M.,Morin, S., and Lafaysse, M.: The method ADAMONT v1.0 for statistical adjustment of climate projections applicable to energy balance land surface models, Geosci. Model Dev., 10, 4257-4283, https://doi.org/10.5194/gmd-10-4257-2017, 2017.
Verfaillie, D., Lafaysse, M., Déqué, M., Eckert, N., Lejeune, Y., and Morin, S.: Multi-component ensembles of future meteorological and natural snow conditions for 1500 m altitude in the Chartreuse mountain range, Northern French Alps, The Cryosphere, 12, 1249-1271, https://doi.org/10.5194/tc-12-1249-2018, 2018.
Vionnet, V., Brun, E., Morin, S., Boone, A., Faroux, S., Le Moigne, P., Martin, E., and Willemet, J.-M.: The detailed snowpack scheme Crocus and its implementation in SURFEX v7.2, Geosci. Model Dev., 5, 773-791, doi:10.5194/gmd-5-773-2012, 2012.
