Performance Evaluation of Global Precipitation Datasets in Northern Mexico Drylands

Abstract Precipitation is a fundamental process in the hydrological cycle; hence, reliable and accurate information on the spatiotemporal rainfall distribution is critical for climate change modeling, local decision-making, and the development of climate adaptation policies. In Mexico’s drylands, the rainfall gauge network is highly heterogeneously distributed and frequently with incomplete datasets. Here, we examine how this deficient rainfall information can be best compensated by information derived from global precipitation datasets. We applied a performance evaluation of Climate Hazards Group Infrared Precipitation with Station data (CHIRPS), CFS-2, AgERA5, PERSIANN-CDR, and TerraClimate and compared these data with observed monthly precipitation records (1983–2018) for dryland regions associated with a network of Participatory Social–Ecological Observatories including Mediterranean climate in northwest (NW) Mexico, coastal arid climate in the Sonora Desert, and semiarid climate in the Chihuahua Desert. We compared monthly and annual rainfall of the global datasets with observed data with Pearson’s correlation coefficient, means, standard deviations, root-mean-square errors, and Taylor diagrams. The results indicate that CHIRPS and AgERA5 can reproduce the precipitation cycle at monthly and annual scales; also, the interannual variability is well captured. Our results suggest that for Mexican drylands, global precipitation datasets can be used to understand drought patterns, for hydrological modeling, for local decision-making, and for the development of urgently needed climate adaptation policies.