Climatic phenomena, such as the El Nino/Southern Oscillation, have been shown to be related to local patterns of land surface hydrologic variables, such as streamflow. Persistence of the climate signal can be exploited to make seasonal forecasts of precipitation and temperature, which in turn can be used to predict runoff and streamflow. For seasonal forecast lead times, the predictability of runoff, and ultimately streamflow, is derived both from predictability of the climatic forcing signal; and persistence in the initial moisture conditions of the land surface, specifically hydrologic storage as soil moisture and snow water equivalent. Identification of the importance of the space-time distribution of hydrologic storage has important implications for targeting locations for ground observation efforts and for validation of remotely sensed products of snow and soil moisture. We evaluate relationships among surface hydrologic variables, and their implications for hydrologic forecasting, using a derived data set of soil moisture, snow water equivalent, and runoff generated using a macroscale hydrologic model forced with gridded surface observations. Statistical analysis of the data set provides insights into the predictability of runoff over the Mississippi River basin, for forecast lead times up to one year. These analyses show that, in addition to information in the climatic signal, significant seasonal forecast skill can be obtained through knowledge of the land surface moisture storage (soil moisture and snow), particularly in the western portions of the basin along the Rocky Mountains. The land surface moisture storage has a significant influence on surface runoff at lead times up to a year for forecasting of winter runoff, and at two to three season lead times for predicting runoff during other seasons. We propose a simplified index, based on the climatological variability in the soil moisture storage, and evaluate its ability to capture the importance of soil moisture and snow on seasonal streamflow forecasts.