摘要：Critical gaps in the amount, quality, consistency,availability, and spatial distribution of rainfall data limit extremeprecipitation analysis, and the application of gridded precipitation datais challenging because of their considerable biases. This study correctedAsian Precipitation Highly Resolved Observational Data Integration TowardsEvaluation of Water Resources (APHRODITE) estimates in the YarlungTsangpo–Brahmaputra River basin (YBRB) using two linear and two nonlinearmethods, and their influence on extreme precipitation indices was assessedby cross-validation. Bias correction greatly improved the performance ofextreme precipitation analysis. The ability of four methods to correctwet-day frequency and coefficient of variation were substantially different,leading to considerable differences in extreme precipitation indices. Localintensity scaling (LOCI) and quantile–quantile mapping (QM) performedbetter than linear scaling (LS) and power transformation (PT). This studywould provide a reference for using gridded precipitation data in extremeprecipitation analysis and selecting a bias-corrected method for rainfallproducts in data-sparse regions.
