Incorporate Satellite Precipitation Product into Radar-Gage Multisensor Precipitation Estimation Algorithm

Abstract: This work is to develop and test a new and enhanced fusion module in the Multisensor Precipitation Estimator (MPE) that would more effectively integrate real-time satellite quantitative precipitation estimates (SQPE) to complement the strengths of ground radar and gauge observations. This module contains a preprocessor that mitigates the spatial location error and systematic bias in SQPE, and a two-way blending routine that statistically fuses adjusted SQPE with radar estimates. To reduce spatial location mismatch, a centroid-moving location matching algorithm is developed to translate the SQPE fields so that its mass centroid coincide with that of radar estimates. Systematic bias and false alarms are reduced through a simple quantile matching algorithm with bias-corrected radar estimates as the reference. The products of this module are validated using independent gauge data. Preliminary results of the validation confirm that the method not only corrected the satellite systematic bias and enhanced the quality of satellite precipitation quality, but also improved the similarity of geographical distribution patterns between satellite and radar products. The effectiveness of the blending module is also tested through data denial experiment, and the results of the experiment show that the blending helps reduce the discontinuities along the boundary of effective radar coverage and improve the root mean square error.