Summertime evaluation of REFAME over the Unites States for near real-time high resolution precipitation estimation

Precipitation is the key input for hydrometeorological modeling and applications. In many regions of the world, including populated areas, ground-based measurement of precipitation (whether from radar or rain gauge) is either sparse in time and space or nonexistent. Therefore, high-resolution satellite-based precipitation products are recognized as critical data sources, especially for rapidly-evolving hydrometeorological events such as flash floods which primarily occur during summer/warm seasons. As " proof of concept" , a recently proposed algorithm called Rain Estimation using Forward Adjusted-advection of Microwave Estimates (REFAME) and its variation REFAMEgeo are evaluated over the contiguous United States during summers of 2009 and 2011. Both methods are originally designed for near real-time high resolution precipitation estimation from remotely sensed data. High-resolution Q2 (ground radar) precipitation data, in conjunction with two operational near real-time satellite-based precipitation products (PERSIANN, PERSIANN-CCS) are used as evaluation reference and for comparison. The study is performed at half-hour temporal resolution and at a range of spatial resolutions (0.08-, 0.25-, 0.5-, and 1-degree latitude/longitude). The statistical analyses suggest that REFAMEgeo performs favorably among the studied products in terms of capturing both spatial coverage and intensity of precipitation at near real-time with the temporal resolution offered by geostationary satellites. With respect to volume precipitation, REFAMEgeo together with REFAME demonstrates slight overestimation of intense precipitation and underestimation of light precipitation events. Compared to REFAME, It is observed that REFAMEgeo maintains stable performance, even when the amount of accessible microwave (MW) overpasses is limited. Based on the encouraging outcome of this study which was intended as " proof of concept" , further testing for other seasons and data-rich regions is the next logical step. Upon confirmation of the relative reliability of the algorithm, it is reasonable to recommend the use of its precipitation estimates for data-sparse regions of the world.