Assimilation of near-real time data products into models of an urban basin

The goal of this study was to determine if assimilating a combination of various derived data products can help circumvent some of the difficulties associated with urban watershed modeling. Combinations of the SNODAS (Snow Data Assimilation System) snow water equivalent data, the SMOS (Soil Moisture and Ocean Salinity) L2 soil moisture, and streamflow observations were used for the data assimilation schemes. Combinations of these observation data sets were assimilated into lumped conceptual rainfall-runoff models of the highly-urbanized Don River basin (in southern Ontario) to determine if assimilation of geophysical variables will have a significant impact on simulations and forecasting in an urbanized watershed. The Ensemble Kalman Filter (EnKF) data assimilation method was used for these analyses, with various rainfall-runoff models that include GR4J, HyMod, MAC-HBV, and SAC-SMA models. The best data assimilation scheme for hydrologic modeling involved using a combination of streamflow, soil moisture, and snow water equivalent while performing both state and parameter updating. These results suggest that using a combination of soil moisture and snow water equivalent from the SMOS and SNODAS data products can improve simulations and ensemble forecasts in an urban basin.