An advanced snow parameterization for the models of atmospheric circulation Ekaterina E. Machulskaya¹³, Vasily N. Lykosov²³ ¹Hydrometeorological Centre of Russian Federation, Moscow, Russia ²Moscow State University, Russia ³Institute for Numerical Mathematics, Russian Academy of Sciences, Moscow, Russia

Introduction Numerous observational studies and model simulations have shown that snow cover affects atmospheric circulation, air temperature, and the hydrologic cycle, due to its especial properties (high albedo, reduced roughness etc.) Snow is related to a number of feedbacks, the most obvious being the snow albedo feedback: a positive temperature bias larger snow melt, faster snow cover depletion decrease of surface albedo more absorption of solar radiation

Snow models description (1) Heat conduction Melting when snow temperature > 0°C or when soil surface temperature > 0°C Heat conduction Liquid water transport Gravitational compaction + metamorphosis Solar radiation penetration 1 layer Arbitrary number of layers, in this study 5 Numerical schemes Implemented processes COSMO INM

Implemented processes (2) - snow temperature,- snow liquid water content, - snow density,- snow specific heat content, - snow heat conductivity, - latent heat for freezing/melting, - melting rate,- refreezing rate,- infiltration rate due to gravity Water percolation: - snow hydravlic conductivity,- snow water holding capacity, - snow porosity Heat and water transport

Gravitational compaction and metamorphosis describes the gravity effect Where member describes the snow metamorphosis, = 75 Pa - the snow compaction viscosity Solar radiation penetration Implemented processes (3)

Data (1) Yakutsk Russia, East Siberia 62 N, 130 E boreal coniferous forest zone grassland site Valdai Russia, European part 58 N, 33 E boreal mixed forest zone grassland site

Data (2) Atmospheric forcing snow water-equivalent depth Valdai: 1966 – 1983 Yakutsk: 1971 – 1973 In winter: every 10 days In spring: more often Every 3 hours: air temperature air pressure air humidity wind speed at 10 m precipitation rate Estimated: shortwave radiation longwave radiation at 2 m Valdai: 1966 – 1983 Yakutsk: 1937 – 1984 Evaluation data

Correlation coefficient between time series of observed and simulated SWE (N = 221, p

Results discussion (2): SWE in Yakutsk observations COSMO INM

observations COSMO INM Results discussion (3): SWE in Valdai Days from Jan. 1 st, / / / /791979/ /69

COSMO TS INM TS COSMO SWE INM SWE Results discussion (3): Impact on the surface temperature (TS)

Summary A new advanced snow parameterization is suggested, implemented and tested by means of long-term data. This multilayered scheme takes into account the latent heat of the phase transfer of water and the interaction with radiative fluxes in the snowpack. In comparison with the more simple model incorporated in COSMO at present, the new more physical scheme represents the snow evolution more realistically, particularly during melting period. The implementation of the new scheme in COSMO is recommended since it can improve the quality of the surface air temperature prediction, particularly in spring. Results of the long-term continues integration with a real forcing data can be used as initial approximation fields for reanalysis of the surface temperature, snow mask and albedo for the adjustment of initial conditions of weather forecast model.

Futher possible directions of the study The Valdai observational data set includes data related to the snow density and albedo, as well as to the snow cover fraction. It is known that fractional snow cover, snow albedo, and their interplay have a considerable effect on the energy available for ablation (Slater et al., 2001; Luce et al., 1998). In alpine environment, elevation, aspect, and slope exert a major control on snow distribution affecting snow accumulation and snowmelt energetics (Pomeroy et al. (2003)). Different data sets that are obtained at present from different field experiments and regular observations (in mountain regions as well), allow to further evaluate the COSMO snow model and to understand to what extent the adequate simulation of different variables is important, in order to improve the prediction of snow evolution and surface air temperature.