Using small watershed real-time streamflow gaging stations coupled with watershed modeling systems as real-time spatial precipitation networks

We have demonstrated that using short term forecasts can outperform nearby weather stations when used to force watershed modeling systems, as shown in Figure 1, a plot of NSE for the Climate Forecast System Reanalysis, which is basically a 32 year series of 6 hourly forecasts generated from the NCEP Global Forecasting System. The begs the question, how accurate are widely spaced meteorological stations at representing the spatial precipitation used in LDAS, and could these precipitation measurements be enhanced using real-time streamflow gaging stations? While weather radar and satellite remote sensing platforms have been valuable to enhance the spatial precipitable water, a stream flow gage could in fact be considered a direct measurement of precipitation events if basins are well parameterized and process based watershed models are used. Benefits of using these gaging stations as actual precipitation measurements is that there are fairly long continuous records for many streamflow gaging stations around the world as well as oftentimes extreme events that cause floodings are not captured by the closest weather station or well extrapolated by the satellite or radar based signals. In this poster presentation, we demonstrate the potential of using a parameterized watershed model on several small basins with streamflow gaging stations, and treating the model as an lm() like function in which given a streamflow response, a spatial precipitation even can be solved for.