Identifying driving processes of drought recovery in the southern Andes natural catchments

Abstract

Study region: The natural river basins of Chile. Study focus: Drought effects on terrestrial ecosystems produce hydroclimatic stress with variable extensions. Particularly, hydrological drought duration can provide a better understanding of recovery together with catchment characteristics and climatology. This study focuses on the impacts of the multi-year drought experienced in Chile for more than a decade.The recovery of relevant catchment variables to quantify the drought termination (DT) and drought termination duration (DTD) after the hydrological drought is presented. A composite analysis of natural catchments using the CAMELS-CL data set discharge (1988–2020), k-NDVI (2000–2020), and soil moisture (1991–2020) provides the average response of the recovery after severe droughts. New hydrological insights for the region: This study demonstrates that local catchment properties can explain the recovery of studied variables after a hydrological drought.Explanatory variables from CAMELS-CL to derive the DT using random forest regression (RFR) were used with a strong correlation of 0.92, 0.84, and 0.89 for discharge, vegetation productivity, and soil moisture, respectively.The discharge patterns show longer recovery over environments dominated by shrublands with less precipitation and higher temperatures, in central Chile, while higher latitudes with higher vegetation cover, increasing precipitation, and lower temperatures present shorter recovery times. The vegetation productivity shows longer recovery over highly vegetated mountains in central Chile. The soil moisture recovery spatial distribution presented patterns that connect them with the discharge recovery. This work enables the identification of drought vulnerability, which is valuable for managing water resources and ecosystems and is helping to predict drought recovery periods in regions with a lack of observations.

Keywords

• CAMELS-CL data set
• Composite analysis
• Random forest regression
• Discharge
• Vegetation
• Soil moisture