Increasing Assimilation and Forecast Skill Through Improved Land Surface States

Significant bias in land surface temperature forecasts in the NCEP regional and global models, both with and without snow, results in the rejection of many surface-sensitive atmospheric channels and limits data assimilation over land. The presence of snow within the the current operational forecast models requires the constraint of surface temperature to be at most freezing over the snow-covered portion of the model grid. This talk focuses on two approaches to removing this constraint in the presence of a vegetation canopy: (1) by maintaining the current Noah model structure, but adding a simple canopy representation and a diagnosed canopy temperature which is allowed to be above freezing and (2) by using the Noah-MP model, which contains an explicit canopy. The unmodified model and two new approaches are compared to satellite-observed surface temperature. Results are shown demonstrating the necessity of including a vegetation canopy for near-surface temperature prediction and reduction in the rejection rate of surface-sensitive channels.