Event-based evaluation of gridded precipitation datasets in the Zayu River Basin, Southeastern Tibetan Plateau

Study region The Zayu River Basin in the southeastern Tibetan Plateau. Study focus Utilizing a dense network of 18 newly deployed rain gauges, this study characterizes spatiotemporal rainfall variability and evaluates six gridded precipitation products. We integrated multi-scale assessments (daily to hourly) with an event-based error decomposition framework to diagnose systematic biases in satellite and reanalysis datasets. New hydrological insights for the region: Observations reveal early-morning rainfall peaks and substantial spatial heterogeneity. At the daily scale, gauge-based CHM (China Hydro-Meteorology Precipitation) outperforms others. For all the hourly products, evaluations expose a distinct spatial disparity: western tributaries exhibit higher detection skill, while eastern tributaries suffer severe overestimation. Specifically, GMCP (Global Multisource Merging-and-Calibration Precipitation) proves most robust. Error decomposition reveals that ERA5-Land (the land component of the fifth-generation ECMWF atmospheric reanalysis) biases stem from "drizzle effects" (false-end errors), whereas IMERG (Integrated Multi-satellite Retrievals for Global Precipitation Measurement) biases arise from stochastic detection failures (miss-events). These diagnostics offer critical guidance for correcting precipitation data in data-scarce alpine regions.