Analytical approaches for characterizing drought rely predominantly on observations of precipitation, temperature and to some extent snowpack and streamflow, often aggregated to spatial units of climate divisions or states. Yet drought is a multi-faceted phenomenon, and the relationship of hydrologic aspects of drought to meteorological aspects of drought varies in space and with the annual cycle. Physically-based hydrology models such as those associated with the NASA/NOAA Land Data Assimilation System (LDAS) project can help to make this connection across space and time scales using a spatially consistent framework. One such model, the Variable Infiltration Capacity (VIC) model, has been applied at ½ degree spatial resolution to the continental US for the purpose of real-time hydrologic assessment and prediction, and for retrospective hydrologic analysis. The result is the University of Washington Experimental Surface Water (SW) Monitor, which since 2005 has provided daily-updating current analyses of nation-wide soil moisture, snow water equivalent and runoff, as well as a retrospective archive of similar products extending back to 1915. Recently, a forecast component has been added to the SW Monitor, enabling probabilistic hydrologic outlook products that are highly relevant to drought characterization and prediction. This presentation describes the SW Monitor and the development of forecast products related to drought.