Verification of Satellite Derived Precipitation Estimates Over Complex Terrain: A Ground Truth Analysis for Nepal

Abstract

Precipitation estimates from the satellite-based Tropical Rainfall Measuring Mission (TRMM) instrumentation play a key role in flood analysis and water resource management across many regions of the world where rain gauge data are sparsely available. Previous studies have produced conflicting results regarding the accuracy of satellite-derived precipitation products, and several authors have called for further examination of their utility, specifically across the Himalaya Mountains region of southern Asia. In this study, daily precipitation estimates generated by TRMM were compared to daily precipitation measurements from a rain gauge network across the country of Nepal. TRMM data were statistically analyzed to quantify their representation of the gauge data during the four precipitation-defined seasons of Nepal. A detailed case study was assembled for the TRMM grid cell characterized by the greatest precipitation gauge density to develop a deeper understanding of local precipitation variability that the coarse resolution TRMM product cannot capture. The results illustrate that TRMM performs relatively well across all seasons, though the performance of TRMM during frozen precipitation events is not clear. In general, TRMM underestimates daily precipitation during the monsoon and pre-monsoon seasons, and overestimates during the winter and post-monsoon season. The case study analysis revealed a threshold for TRMM bias of 10-20mm of daily precipitation, overestimating lighter precipitation events while underestimating heavier precipitation events. Still, TRMM data compare favorably to gauge data, which contributes to the confidence with which they and other satellite-derived data products are used.