Document Type : Research Paper
Authors
1 Department of Civil Engineering, Faculty of Engineering, University of Birjand, Birjand, Iran.
2 Department of Civil Engineering, Faculty of Engineering, University of Birjand, Birjand, Iran
Abstract
Hydro-climatic research and water resource management in arid, data-scarce regions depend fundamentally on precise precipitation data. This study presents the first comprehensive, multi-scale evaluation of four prominent Gridded Precipitation Products (GPPs)—TRMM, CHIRPS, GSMaP, and ERA5—in the climatically challenging and sparsely gauged South Khorasan province of Iran (2010–2019). Using ground-based observations as a reference, GPP performance was evaluated across multiple timescales with a suite of statistical metrics. The evaluation framework leverages diagnostic visualizations, such as Taylor and performance diagrams, to provide deeper insights into error structures than can be achieved through traditional map-based assessments. The analysis revealed a clear performance ranking: the satellite-based TRMM and GSMaP consistently performed best, showing higher accuracy (median RMSE ≈ 2.91–3.05 mm/day), stronger correlation (median CC ≈ 0.63–0.65), and a more balanced detection skill (median CSI ≈ 0.43–0.45). In contrast, the ERA5 reanalysis product, despite achieving the highest Probability of Detection (POD ≈ 0.78), suffered from notable systematic biases and the largest random errors. Performance for all products degraded during the arid summer, and estimation errors systematically increased in wetter regions. We conclude that the gauge-adjusted satellite products, GSMaP and TRMM, provide the most dependable precipitation estimates for the study area. These findings offer a critical, evidence-based guide for selecting appropriate GPPs in this vulnerable environment and provide insights for future algorithm development.
Keywords
)