Evaluation of precipitation data from CHIRPS images in dry and tropical (San Luis Potosi) and temperate (Mexico State) basins in Mexico

Abstract

Hydrological modeling requires precipitation data as a basic input since it represents the primary water source entering the hydrological system. These data are mainly obtained from weather stations, but in many cases, the number of stations is low, and may have missing data, leading to errors in modeling. With technological advances in remote sensing, freely accessible satellite precipitation data are gaining importance as a potential input for hydrological modeling. Thus, the objective of the present research is to evaluate two sources of precipitation data: weather stations and CHIRPS (Climate Hazards Center InfraRed Precipitaion Station) imagery. Both sources were compared across three basins with contrasting climates (arid, temperate, and tropical). The analysis was based on statistical measures, such as Spearman’s correlation coef ficient ρ (rho), Root Mean Square Error (RMSE), and the Wilcoxon test to detect correlations and statistical dif ferences. The results revealed a positive and significant correlation (P < 0.05) between both data sources across the three basins, though with variations depending on the climate type. The RMSE (< 21 mm) were recorded in temperate and arid regions, while in the tropical region, CHIRPS imagery underestimated precipitation. The greatest similarity (P > 0.05) between the two sources was found in the temperate region, which suggests that CHIRPS imagery is a viable alternative to weather stations for hydrological modeling, especially in temperate climate basins.