|Title||New snow scheme in HTESSEL: description and offline validation|
|Publication Type||Technical memorandum|
|Secondary Title||ECMWF Technical Memorandum|
|Authors||Dutra, E, Balsamo, G, Viterbo, P, Miranda, PMA, Beljaars, A, Schär, C, Elder, K|
A new snow scheme for the European Centre for Medium-Range Weather Forecasts (ECMWF) land surface model has been tested and validated. The scheme includes a new parameterization of snow density, incorporating a liquid water reservoir, and revised formulations for the sub-grid snow cover fraction and snow albedo. Offline validation (covering a wide range of spatial and temporal scales) includes simulations for several observation sites from the Snow Models Intercomparison Project-2 (SnowMIP2), global simulations driven by the meteorological forcing from the Global Soil Wetness Project-2 (GSWP2), and by ECMWF ERA-Interim re-analysis. The new scheme reduces the end of season ablation biases from 10 to 2 days in open areas, and from 21 to 13 day in forest areas. Global GSWP2 results are compared against basin scale runoff and terrestrial water storage. The new snow density parameterization increases the snow thermal insulation, reducing soil freezing and leading to an improved hydrological cycle. Simulated snow cover fraction is compared against NOAA/NESDIS with a reduction of the negative bias of snow-covered area of the original snow scheme. The original snow scheme had a systematic negative bias in surface albedo, when compared against MODIS remote sensing data. The new scheme reduces the albedo bias, consequently reducing the spatial and time averaged surface net shortwave radiation bias by 5.2 W m-2 in 14% of the northern hemisphere land. The new snow scheme described in this report was introduced in the ECMWF operational forecast system in September 2009 (CY35R3).