Accurate prediction of available water supply from snowmelt is needed if the myriad of human, environmental, agricultural, and industrial demands for water are to be satisfied, especially given legislatively imposed conditions on its allocation. Robust retrievals of hydrologic basin model variables (e.g., insolation or areal extent of snow cover) provide several advantages over the current operational use of either point measurements or parameterizations to help to meet this requirement. Insolation can be provided at hourly time scales (or better if needed during rapid melt events associated with flooding) and at 1-km spatial resolution. These satellite-based retrievals incorporate the effects of highly variable (both in space and time) and unpredictable cloud cover on estimates of insolation. The insolation estimates are further adjusted for the effects of basin topography using a high- resolution digital elevation model prior to model input. Simulations of two Sierra Nevada rivers in the snowmelt seasons of 1998 and 1999 indicate that even the simplest improvements in modeled insolation can improve snowmelt simulations, with 10%–20% reductions in root-mean-square errors. Direct retrieval of the areal extent of snow cover may mitigate the need to rely entirely on internal calculations of this variable, a reliance that can yield large errors that are difficult to correct until long after the season is complete and that often leads to persistent underestimates or overestimates of the volumes of the water to operational reservoirs. Agencies responsible for accurately predicting available water resources from the melt of snowpack [e.g., both federal (the National Weather Service River Forecast Centers) and state (the California Department of Water Resources)] can benefit by incorporating concepts developed herein into their operational forecasting procedures.