A Comparative Evaluation of Gauge-Satellite-Based Merging Products Over Multiregional Complex Terrain Basin

Abstract

Satellite precipitation products (SPPs) have become a significant data source in hydrometeorology, especially for ungauged or sparsely gauged basins, and satellite estimates can supplement the lack of precipitation information. However, satellite remote sensing precipitation products have various degrees of uncertainty which may be related to studied area, precipitation type, and other factors. In addition, the applicability of the same satellite product may not be consistent in different regions. This study focuses on statistical evaluations of five SPPs (TRMM_3B42v7, TRMM_3B42RT, PERSIANN_CDR, CMORPH_CRT, and IMERG_Final) and further improves their accuracy by using the principal component regression-based merging model over two different topography regions. The utilities of all the precipitation sets were then investigated using the Xinanjiang hydrological model to access the influence of the combined product in hydrological response and relative improvements. Results show that the original TRMM_3B42v7, CMORPH_CRT, and IMERG_Final products exhibit much better performance than other two products for precipitation estimation, while the accuracy of TRMM 3B42RT varies greatly between two regions. The merged IMERG_Final and CMORPH_CRT significantly improved the results in two basins. In summary, the merging method combines satellite- and ground-based precipitation to produce a finer spatial distribution, while attention should be paid to the effect of merging method with respect to different satellite information within different basins. This study provides quantitative precipitation estimation analysis of SPPs in different regions and the applicability of merging methods for different SPPs, and thus, will provide useful guidelines for future utilizations of SPPs and merging schemes in regional hydrological applications

Keywords

• Climate precipitation center morphing method (CMORPH)
• hydrological model
• integrated multisatellite retrievals for global precipitation measurement (IMERG)
• precipitation estimation from remote sensing information using artificial neural network (PERSIANN)
• precipitation data merge
• satellite precipitation