Comprehensive evaluations of gridded precipitation datasets across diverse climate zones of Brazos River Basin, Texas, USA

Accurate precipitation datasets are essential for informed water, land, and agricultural management. This study evaluates four gridded precipitation datasets, Daymet (version 4 R1), PRISM, IMERG (version 07), and CHIRPS (version 2.0), against GHCN (version 3) observations across climate zones of the Brazos River Basin, Texas, USA (1998–2020). The evaluation examined precipitation occurrence, intensities, totals, and extremes. Event detection and intensity were assessed using categorical metrics and probability density functions (PDFs), while totals from daily to annual scales were evaluated with four continuous statistics. Ten extreme precipitation indices evaluated each dataset’s ability to capture precipitation extremes. A Comprehensive Rating Index (CRI) was applied to synthesize overall performance. The results show that PRISM consistently outperformed the other datasets in detecting precipitation occurrence, simulating intensities, estimating totals, and capturing most precipitation extremes across most climate zones. Daymet and CHIRPS performed poorly in precipitation detection, while CHIRPS and IMERG were less effective in estimating intensity. Daymet performed poorly for daily estimates, and IMERG had lower accuracy at the monthly, winter, and annual scales. IMERG and CHIRPS underperformed for summer precipitation, with CHIRPS also weakest in representing most precipitation extremes. These findings also underscore the variability in dataset performance across climate zones. Therefore, PRISM is recommended for gridded applications in the Brazos River Basin due to its accuracy across precipitation characteristics and temporal scales. For station-level or region-specific studies, selecting the top-performing dataset for each climate zone of the study area based on relevant temporal scales is advised.