"Ten years of operational numerical simulations of snow and"
Ten years of operational numerical simulations of snow and mountain weather conditions and recent developments at Météo-France. Y. Durand, G. Guyomarc'h, L. Mérindol, J.-G. Corripio, G. Giraud, E. Brun, E. Martin. Météo-France, Centre d'Études de la Neige (CNRM), 1441 rue de la Piscine, F 38406 Saint Martin d'Hères. Introduction: The development of snow CROCUS is thus forced by the SAFRAN evolution numerical modelling in the objective analysis which provides hourly beginning of the nineties at the Snow values of temperature, wind, cloudiness, Study Centre (CEN) of Météo-France has humidity, precipitation conditions and induced the necessity of building a full incoming radiations. Except for the suite of software including an objective precipitation where only a daily analysis analysis program suited for mountainous by Optimal Interpolation (OI) is performed conditions and able to determine relevant and then split at hourly time steps, all the meteorological near-surface conditions. main quantities are hourly analysed The purpose was thus to implement an through a mixed method (OI and operational survey of the snow pack variational) due to the multi-purpose (including its internal characteristics and software design. Several algorithms, such stratigraphy) in mountainous areas where as a precise determination of the snow/rain no routine snow observation is available. limit altitude and the use of satellite Since the winter season 1991-92, the first information are also incorporated. version of this chain has been used The spatial working grid of the whole operationally by the French forecasters in chain is irregular and corresponds to the the Alps (1995-96 in the Pyrenees). massif scale (~400 km2). The models are Several improvements and developments run for seven slope aspects at different have been then carried out by automatic elevations with 300m vertical steps. It is monitoring and contacts with the users. worth noticing that the full system is never The chain is also being used for researches re-initialised along the year with snow in hydrology and in climate change impact observations; the snow evolution at every assessment. location being only forced by the meteorological inputs. An avalanche hazard diagnosis is then provided at each study point by the MEPRA model. Results are deducted from the estimated mechanical properties of each snow layers. The risk of spontaneous avalanche is thus based on the comparison between the shear strength and the shear stress applied to each layers. In case of wet snow, the risk can be increased by melting destabilisation or decreased by refreezing. Concerning the accidental risks, the Flow chart of the full chain. predicted shear stress is increased by the effect of overloading due to a skier and The present models: The suite is also typical structures of slabs and weak organised around the CROCUS snow layers are also searched into the snow model which calculates the energy and profile. The risk estimation is completed mass evolution of the snow cover with a with a classification of the avalanche type. vertical discretization up to 50 levels in The use of such indications is presented in including snow metamorphisms and a the following figure. representation of each snow crystals types. Figure illustrating the underlying CROCUS- MEPRA processes (vertical structure, used variables, type of crystals, diagnostic of an accidental risk on the Belledonne massif, 1 January 96 at 12 UTC) A forecast version of this chain is also available and runs at a range of 2 days. It is based on downscaling operators treating Figure showing 10 years of comparisons (81-91) between measured snow depth (dotted line) and the information from larger scales NWP simulated (solid line) at the Tignes ski resort models (Arpège or Aladin) and on the use (Vanoise massif) of observations of analogous weather situations of the past. The current developments: The present line of research aims at a better spatial resolution of the provided results, especially for the avalanche diagnoses, at a finer spatial scale of about 1 km, what we call "local scale". This requires to better take into account the fine scale forcing of the local topography, the effects of which are less important at the massif scale and different other related phenomena. An important point is thus to insert the complex effects of the snowdrift which transported amounts cannot be ignored so as its implications on the snow stability. Figure showing the Alpine massifs irregular grid The treatment of such small but important and four 24h precipitation fields at level 1800 m on phenomena is presently a challenge due to a flat aspect. From the left, you find: the SAFRAN the permanent interaction between snow analysed field on 11 January 99, the final and weather conditions and to their very forecasted field at this date and the Arpège downscaled field and the analogous past situation. fine scale resolution and high complexity which often exceeds the present state of the The operational validation is obtained by art in current modelling. The used simulating precise locations where regular solutions are fine mesh modelling and manual or automatic snow and weather several downscaling operators. Among observations are performed without these operators are the local simulation of incorporating these data in the analysis the snow profiles, the modifications of the scheme or by comparison with different velocity field, the snow drift occurrence independent fields as on the previous and the precipitation redistribution. We are figure presently working on these different subjects as well as on the validation of the results. The SAFRAN wind field is adapted to the Related applications: SAFRAN and local topography by a small numerical CROCUS have been widely used for more model of potential vorticity, SAMVER, specific research applications. We shall and through statistical relationships with principally quote climate change with the main features of the local topography. modifications of the SAFRAN outputs These last results have been obtained by according to several climate change the use of comparisons between SAMVER scenarii. and MESO-NH on common runs. The snow drift effects are simulated on a test area (2x3 km) on a grid of 45m mesh (model SYTRON2) and are presently under evaluation. Figure showing the impact of a scenario of a temperature increase of 1.8 °C on the snow on ground duration in the different alpine massifs at level 1500m asl. The duration is reduced from 5 to 4 months in the Northern Alps and from 3 to 2 months in Southern massifs. Several other studies, especially for hydrology have also successfully been done. An operational suite of automatic estimation of hydrological discharges based on the SAFRAN analysis is being implemented over the whole French territory. Figure showing the SYTRON2 6hrs modelled snowdepth difference on 22 March 2002 due to References: snow drift phenomena. The scene is located in the Brun, E., P. David, M. Sudul and G. Brugnot 1992. A Grandes Rousses massif (orography near 2700m numerical model to simulate snow cover stratigraphy asl plotted in bold lines) with eroded areas in for operational avalanche forecasting. J. Glaciol., brown colours and accumulation in green colours. 38(128), 13-22. The mesh size is about 45m and the plotted wind is Corripio, J. G., 2003. Snow surface albedo estimation using terrestrial photography, Int. Journal of Remote obtained from SAMVER. Sensing . Submitted. Durand, Y., G. Giraud, E. Brun, L. Mérindol and E. The validation of such experiment is very Martin, 1999. A computer-based system simulating swowpack structures as a tool for regional avalanche difficult. The present used methodology is forecast. J. Glaciol., 45(151), 466-484. the comparison of the observed surface Guyomarc'h G , L. Mérindol, 1998. Validation of an snow albedo (by the use of digital application for forecasting blowing snow. Ann. Glaciol., 26, 138-143. photographs) and the modelled resulting Lafore, J. P., J. Stein, N. Asencio, P. Bougeault, V. albedo when the snow crystal features have Ducrocq, J. Duron, C. Fischer, P. Hereil, P. Mascart, been modified by the transport effects and J. P. Pinty, J. L. Redelsperger, E. Richard, and J. Vila-Guerau de Arellano, 1998: The Meso-NH the deposition. Atmospheric Simulation System. Part I: Adiabatic Other developments are also in progress in formulation and control simulations. Annales the framework of the chain as new stability Geophysicae , 16, 90-10. Martin E., E. Brun, Y. Durand, 1994. Sensitivity of the diagnostics linked to micro mechanical French Alps snow cover to the variation of climatic studies or as new estimations of the liquid variables. Ann. Geophy. 12, 469-474. water fluxes. Mérindol L., Y. Durand and G. Guyomarc’h. 2000: “Simulation of Snowdrift over Complex Terrain”. ICAM 2000, Innsbrûck, 11-15/9/2000.