Hydrologic sensitivity to high spatiotemporal resolution precipitation measurements from phased-array radar

Weather radar is an attractive tool for use in hydrology due to its capability to measure the variability of rainfall at much higher resolution than is typically available by operational rain gauge networks. The use of next generation weather radar (NEXRAD) precipitation estimates has been demonstrated to improve hydrologic forecasting especially in ungauged regions. This study uses precipitation datasets from the National Mosaic QPE system (NMQ/Q2), which provides real-time precipitation estimates at 5-min/1-km2 over the conterminous US. In addition to NMQ/Q2, the National Weather Radar Testbed Phased Array Radar (PAR) is capable to adaptively scan storms at a temporal resolution of 1 min.

The aim of this study is to assess the sensitivity of the spatiotemporal variability of rainfall estimates on hydrological response at various catchment scales. We use the Coupled Routing and Excess Storage (CREST) distributed hydrological model to run the streamflow simulations. The CREST model, developed by the University of Oklahoma (hydro.ou.edu) and NASA SERVIR Project Team (www.servir.net), simulates the spatiotemporal variation of water and energy fluxes and storages on a regular grid with the grid cell resolution being user-defined, thereby enabling the different spatial scale applications. In this study, we evaluate the space-time scale sensitivity of the CREST model by using NMQ/Q2 and PAR high-resolution QPE observations at temporal resolutions from 1, 2, 3, 4, 5, 10, 30 and 60 min. The spatial resolutions range from 0.01*0.01 degree up to 0.25*0.25 degree. Simulations are evaluated in comparison to observations on the heavily instrumented Ft. Cobb basin in Oklahoma as well as on synthetic basins.