The accurate portrayal of the hydrological cycle is extremely important in land surface modeling. Central to this effort is the treatment of snowcover, as errors in this area can affect the entire simulation. Snow cover in the Eta Data Assimilation System (EDAS) is updated once per day with observations, and analysis shows that repeated melting and replenishing of snow pack occurs as a product of the updating process over significantly large areas of the United States. The root of this problem is a warm bias in the model which leads to excessive snowmelt at the edge of snowpacks. This repeated melting infuses the soil column with a large quantity of water that upsets the hydrological cycle. In an effort to quantify the impacts of this problem, an examination of output from the EDAS model was conducted, and a series of Mosaic Land Surface Model (LSM) simulations forced with output from the EDAS model were performed. These LSM runs used different updating and bias correction schemes, and made it possible to examine the importance of matching observed quantities used for updating model states, with model simulated quantities such as surface temperature.