Contrasting characteristics and drivers of dry and warm snow droughts in China's largest inland river basin

Abstract

Study region: The largest inland river basin in China (the Tarim River Basin) Study focus: This study introduces a novel classification scheme for dry and warm snow droughts with focus on their duration, severity, and intensity. The Nonparametric Standardized SWE Index (NSWEI) and the Standardized Precipitation Index (SPI) were employed to assess these drought types. The optimal parameters-based geographical detector (OPGD) model and machine learning (including random forest model and Shapley-XGBoost algorithm) were applied to reveal the spatial and time dynamic driving forces of droughts. New hydrological insights for the regions: Results indicated that warm-type droughts occurred more frequently and with greater intensity, exhibiting a larger spatial coverage compared to dry-type droughts, which can be attributed to warm-type droughts dominated by temperature and vapor pressure deficit, whereas dry-type are predominantly influenced by relative humidity and solar radiation. Moreover, the dynamics of snow droughts exhibited a contrasting pattern between the south (experiencing an upward trend) and the north (experiencing a descending trend). In the south, dry-type droughts were exacerbated by the strength of North Pacific Oscillation (PDO)-precipitation coupling, while warm-type droughts are influenced by the strength of North Atlantic Oscillation (NAO)-precipitation coupling and El Niño Southern Oscillation (ENSO)-precipitation.

Keywords

• Dry-type and warm-type snow droughts
• Snow drought characteristics and dynamics
• Optimal parameters-based geographical detector model
• Random forest
• Ocean-atmosphere couplings strength