Comparison of Spatial Patterns of Aridity in Argentina Based on Different Climatic Indices and Datasets

Abstract

Aridity significantly affects ecosystems, agriculture and water resources. Characterising its spatial patterns is key for designing adaptation and mitigation strategies against advancing desertification, but comparative assessments of how different indices and datasets represent these patterns are still lacking. This study analyses the spatial patterns of aridity in Argentina during 1961–2020 by comparing different climatic indices and datasets. Monthly precipitation and temperature data from observations and gridded datasets—CRU, WORLDCLIM and ERA5—were used. Potential evapotranspiration (PET) was estimated using Thornthwaite, Holdridge and Hargreaves–Samani methods. Aridity indices were calculated at annual and seasonal scales, selected for their frequent use in recent literature and ability to provide complementary perspectives on aridity. The performance of the gridded data was evaluated using various statistical comparison metrics against observations. The results show that aridity indices based on mean temperature or PET, particularly those using the Thornthwaite method, are more consistent in representing climate patterns in Argentina compared to indices based on temperature ranges. On one hand, CRU and WORLDCLIM exhibited similar performance, with relatively low biases, high spatial coherence with observations, and good representation of statistical distributions, while on the other hand, ERA5 showed significant errors in representing climate categories and estimating changes in aridity, especially during summer and in areas with complex topography. These findings highlight the importance of carefully selecting the aridity index and dataset, as this decision affects the representation of aridity patterns and their changes, which are crucial for climate monitoring and decision-making in regions vulnerable to desertification.