Warming-induced cryosphere changes predict drier Andean eco-regions

Abstract

Abstract Climate change impacts on humans and ecosystems depend on the intensity, timing, and spatial variability of these changes. While considerable attention has been paid to current and future changes in temperature patterns, comparatively less attention has been devoted to water availability for humans and ecosystems. The aridity index (AI), the ratio of precipitation to potential evapotranspiration, is a common metric used to assess water availability within ecosystems. However, the role of snow in AI calculations within snowy eco-regions is often neglected, resulting in an incomplete understanding of water balance dynamics in these environments. In this study, we estimate aridification under ongoing climate change in Andean eco-regions (AEs), focusing on two horizons: 2050–2060 and 2090–2100. Using monthly TerraClimate data from 2013–2018, we calculated a mean AI for each AE, taking into account the absence of snow (pixels with a snow water equivalent (SWE) < 10 mm/month) and its presence (AI-snow; pixels with a SWE > 10 mm/month). We show that AI allows to differentiate the eco-regions, but that the incorporation of snow in the AI calculation highlights the heterogeneity of aridity conditions within some eco-regions with energy-limited regimes (AI > 1) in the snowy zones and water-limited regimes (AI < 1) elsewhere. Analysis of the CORDEX-SAM regional projections for the periods 2050–2060 and 2090–2100 indicates a general shift towards drier conditions prevailing over wetter conditions in most eco-regions, notably: the Southern Andean Steppe, the Central Andean Wet Puna, the Santa Marta Páramo, and the Peruvian Yungas. The projected reduction in snowfall in CORDEX-SAM, coupled with glacier volume loss, appears to be contributing to the prevalence of aridification across many AEs. These findings highlight potential transitions towards aridification in diverse eco-regions, with repercussions on water availability for humans and ecosystems.