An optimal merging technique for high-resolution precipitation products

Abstract

Precipitation products are currently available from various sourcesat higher spatial and temporal resolution than any time in the past.Each of the precipitation products has its strengths and weaknessesin availability, accuracy, resolution, retrieval techniques and qualitycontrol. By merging the precipitation data obtained from multiplesources, one can improve its information content by minimizing theseissues. However, precipitation data merging poses challenges of scale-mismatch,and accurate error and bias assessment. In this paper we presentOptimal Merging of Precipitation (OMP), a new method to merge precipitationdata from multiple sources that are of different spatial and temporalresolutions and accuracies. This method is a combination of scaleconversion and merging weight optimization, involving performance-tracingbased on Bayesian statistics and trend-analysis, which yields mergingweights for each precipitation data source. The weights are optimizedat multiple scales to facilitate multiscale merging and better precipitationdownscaling. Precipitation data used in the experiment include productsfrom the 12-km resolution North American Land Data Assimilation (NLDAS)system, the 8-km resolution CMORPH and the 4-km resolution NationalStage-IV QPE. The test cases demonstrate that the OMP method is capableof identifying a better data source and allocating a higher priorityfor them in the merging procedure, dynamically over the region andtime period. This method is also effective in filtering out poorquality data introduced into the merging process.