Snow model comparison to simulate snow depth evolution and sublimation at point scale in the semi-arid Andes of Chile

Abstract. Physically based snow models provide valuable information on cover evolution and are therefore key to water availability projections. Yet, uncertainties related modelling remain large as a result of differences in the representation physics meteorological forcing. While many studies focus evaluating these uncertainties, no model comparison has been done environments where sublimation is main ablation process. This study evaluates case semi-arid Andes Chile aims compare two with different complexities, SNOWPACK SnowModel, at local point over one season evaluate their sensitivity relative parameterisation For that purpose, forced (i) most ideal set input parameters, (ii) an ensemble physical parameterisations, (iii) biased Results indicate depending forcing, roughness length z0, albedo parameterisation, fresh density parameterisation. The uncertainty caused by forcing directly bias chosen. Even though show significant complexity, choice least importance, both data was same order magnitude highly influenced precipitation uncertainties. ratio ranges agreement for models: 36.4 % 80.7 SnowModel 36.3 86.0 SNOWPACK, models.