Spatio-temporal performance of monthly gridded precipitation products in Patagonia

Abstract

Reliable precipitation information is essential for climate monitoring, hydrological assessment, and water resource management, particularly in regions with strong climatic gradients and limited observational coverage. Patagonia presents major challenges for precipitation estimation due to its complex topography and sparse gauge network. This study evaluates the spatio-temporal performance of six monthly gridded precipitation products (GPCCv2022, CRU TS v4.09, MSWEPv2.8, CHIRPSv3.0, JRA-3Q, and ERA5-Land) over Patagonia during 1981–2020 using observations from 345 meteorological stations. The assessment combines climatological analyses, statistical performance metrics, spectral characterization, and trend evaluation to provide an integrated perspective on dataset reliability across this climatically heterogeneous region. Results show broad agreement among datasets in reproducing mean precipitation patterns over eastern Patagonia, whereas substantial discrepancies emerge along the Andes and southern sectors. In terms of precipitation magnitude, GPCC exhibits the lowest bias, with a median overestimation of only 0.2%, while ERA5-Land shows the largest systematic overestimation, with median precipitation amounts 59.7% higher than observed. Regarding temporal variability, MSWEP achieves the highest median correlation with observations (+0.856), whereas CHIRPS shows the lowest performance (+0.730). Overall performance, assessed through the Kling–Gupta Efficiency, identifies MSWEP as the best-performing product across most stations, while ERA5-Land ranks as the least efficient dataset. Spectral analyses indicate that all products reproduce the dominant annual cycle but generally overestimate its amplitude and underestimate variability at shorter and longer timescales. Trend analyses reveal agreement in the sign of large-scale precipitation changes, although important regional differences in magnitude persist. These findings provide guidance for precipitation dataset selection in Patagonia and other data-sparse regions.