Use of 4 km, 1 hr, precipitation forecasts to drive a distributed hydrologic model for flash flood prediction

Use of a distributed hydrologic model to improve upon the capabilities of the National Weather Service (NWS) Flash Flood Guidance (FFG) system (typically applied to ungauged locations) is an active area of research. An ongoing research project at the NOAA NWS Hydrology Laboratory (HL) involves the following steps: (1) parameterize a distributed hydrologic model at a scale commensurate with the available multi-sensor precipitation estimates, (2) derive estimates of flood flow using a regional empirical relationship, (3) run retrospective simulations, (4) evaluate statistical procedures to produce probability of flooding estimates on a grid, and (5) validate the results using observed streamflow data from small watersheds (<260 km2) that were not used to estimate model parameters. A practical limitation of the research to date is that we have only examined the use of radar-based precipitation observations, not high resolution precipitation forecasts. Thus, the only lead time is a function of basin response time, which may be of limited practical value on very small basins. This paper will expand upon the current work by using 1 hr forecast precipitation grids from the HL Multisensor Precipitation Nowcaster (MPN) to drive the model. Candidate study areas include basins covered by the Tulsa, Oklahoma radar and the Sterling, Virginia radar.