The majority of hydrologic modeling today still relies on point-based gauge measurements to describe the spatial distribution of precipitation. Although radar-derived rainfall from the National Weather Service (NWS) Weather Surveillance Radar-1988 Doppler (WSR-88D) network provides the complex spatial resolution that gauge networks lack, there has been a hesitation to use it instead of gauge data in hydrologic models. Previous research has shown that gauge adjusted radar-estimated precipitation is more accurate than that of non-adjusted radar data. One such radar adjustment algorithm is the NWS Multi-sensor Precipitation Estimator (MPE) which combines the spatial resolution of the radar with the relative accuracy of gauge measurements. We have employed a version of MPE at Florida State University (FSU) to create a historical precipitation dataset over Florida.

This paper will compare the use of quality controlled rain gauge data from the South Florida Water Management District (SFWMD) with the gridded FSU MPE dataset in the MIKE SHE hydrologic model. Both hourly and daily time steps will be examined. MIKE SHE is a fully integrated and distributed, physically based mathematical model that describes the flow within the entire land-based phase of the hydrological cycle. Our paper will consider the 1661 km2 Big Cypress Basin (BCB) in southern Florida. The basin is very flat, with an elevation difference of ~ 12 m over the entire domain. Stage heights from the model will be evaluated against observed daily stage heights provided by SFWMD. Differences between gauge versus multi-sensor precipitation at daily and hourly time steps will be quantified over a three year interval (2003-2005), over seasonal periods, and over multiple case studies. In summary, this study will evaluate the potential of using spatially distributed, multi-sensor rainfall for hydrological modeling within the relatively flat terrain of SFWMD.