Representing Urban Infiltration in the National Water Model

Since its inception, the National Water Model (NWM) has had a focus on forecasting the evolution of urban flooding and urban hydrology. Achieving this goal requires accurate hydrograph generation in basins containing urban and/or impervious surfaces. In this project, we explore how urban landcover is represented in the NWM land-surface model- Noah-MP - and the implications on hydrograph and water budget accuracies. Currently, the NWM relies on two simplifications in urban zones. The first is that all urban cells, irrespective of development intensity, are aggregated to a single class. The second is that all urban zones are parameterized with a limited water holding capacity that effectively limits infiltration. Using a collection of basins across CONUS, we run a series of experiments challenging these two assumptions. Experiment 1 removes the water holding capacity constraint in the current NWM configuration, while Experiment 2 introduces a new infiltration scheme that accounts for impervious surface using a mean condition. Experiment 3 uses this new scheme and introduces a graduated class of urban intensity (from the USGS National Land Cover Dataset) with an impervious percentage equal to the mean of each category's classification criteria. Experiment 4 introduces an impervious layer to the NWM parameter files to uniquely characterize each cell. The purpose of these experiments is to understand if, and to what level, added complexity in urban zones results in more accurate streamflow forecasts, and what characteristics are most critical.