Snow cover has great impact on both hydrological cycle and atmospheric process. However, most global and regional model only have very simple parameterizations. It is a challenging scientific task to develop a physical based snow model with sufficient simplicity for global and regional climate studies.
A three layer snow model based on Anderson and Jorden's parameterizations has been developed for this purpose. A great deal of effort has been made in selecting and presenting the physical processes for heat energy transfer and mass transfer in snow layers. It includes the development of a scheme of layer snow cover, and finding an accurate and efficient numerical method for the model.
The physical processes in the model contains the self weight effect of snow mass on snow compaction, which affects snow density and depth, and partitioning the melting water outflow from a sublayer into runoff and downward infiltration flow, etc. Enthalpy is used as a prognostic variable to more efficiently deal with phase change and reduce computational errors.
The scheme of layering the snow cover is critical for prediction. The total number of layers is not more than three. The thickness of each layer is variable with time and increases with the snow depth. An implicit numerical scheme is used for the surface layer and an explicit scheme for other layers, which save computational time and increase the accuracy of the prediction.
A preliminary comparison shows this model is able to produce similar results as Jorden's fine layer model. Snow data from six Russian stations and Col de Porte of France have been used to validate the model's simulations of snow cover, snow depth, snow water equivalent, snow surface temperature, and runoff. SSiB has also been tested using these stations data for comparisons. The results from layer snow model show good agreement with the observations are better than those from SSiB simulations in thick snow cases. A coupled SSiB/layer snow model has also been tested.