Snow is a critical component of the hydrologic cycle accounting for a significant fraction of the available freshwater resources in many parts of the northern hemisphere. This two-part talk will discuss some recent work related to the assimilation of space-based measurements into a land surface model for the purpose of improved snowpack characterization. The first talk will highlight a study assimilating Gravity and Recovery Climate Experiment (GRACE) terrestrial water storage (TWS) information into the NASA Catchment Land Surface Model (Catchment) within the Mackenzie River basin located in northwest Canada. When compared against an advanced, ground-based snow product, results show that assimilation (relative to the open-loop simulation) reduced the RMSE in modeled snow water equivalent (SWE) by as much as 25%. However, the utility of GRACE measurements in the Mackenzie is limited by a variety of factors, including the effects of post glacial rebound. The second talk discusses the development of an artificial neural network (ANN) using Catchment output to predict Advanced Microwave Scanning Radiometer (AMSR-E) brightness temperatures (Tb) across North America as a function of snowpack state. The ANN will ultimately be used as a predicted measurement model within a data assimilation framework for the assimilation of AMSR-E Tb. Over the course of the 9-year study period, results show the ANN predictions are unbiased, yield a reasonable amount (<10 K) of RMSE, and demonstrate significant skill at predicting interannual variability (anomaly correlation of 0.6-0.7). In addition, the ANN demonstrates considerable skill at predicting Tb when the snowpack is ripe (wet) with even greater skill at predicting Tb when the snowpack is dry. Ultimately, the results from these two studies could be merged to yield improved snowpack estimates at the continental scale via simultaneous assimilation of GRACE TWS and AMSR-E Tb.