The presence of snowcover strongly influences energy exchanges between
the land surface and the atmosphere, and so influences the
modification of air masses. Accurate snow analyses are thus required
for numerical weather prediction, but some studies have found only
weak positive impacts on forecast skill from the assimilation of snow
observations from existing sources. This presentation will investigate
the information content of different types of direct and remote snow
observations for constraining the evolution of snow model state
variables and resulting impacts on simulated surface energy fluxes.