Understanding Long-Term Streamflow Response to Snowfall Change: Insights From a Multivariate Analysis

Abstract

Ongoing climate change is modifying snow dynamics, further altering streamflow (Q) in cold regions. Despite extensive research over the past decades, the impact of changes in snow on mean annual Q remains debated, and the underlying mechanisms driving these responses are still unclear. In this study, we employ the Budyko framework to establish multivariate relationships between a Budyko model parameter and several variables, including snowfall fraction (f_s), across 931 carefully selected snow-influenced catchments in the Northern Hemisphere. Our results show that the average elasticity of Q to f_s across all catchments is 0.13. However, mean annual Q responses vary among catchments, with 78% of catchments experiencing decreased Q and others exhibiting increased Q with declining f_s. These variations are manifested in decreased warm-season Q and increased cold-season Q, respectively. Additionally, attribution analysis of Q change highlights a critical role of f_s changes as a second-order control on Q after precipitation. Examining possible mechanisms behind the varying Q responses reveal that the influence of f_s on snowmelt intensity and snow season available energy is more pronounced in catchments exhibiting positive Q feedback to f_s. In contrast, its impact on rainfall intensity and the asynchrony between available energy and liquid water input is greater in catchments displaying negative Q feedback to f_s. Our findings enhance the understanding of Q responses to changing f_s and provide valuable insights for water resource management in snow-influenced regions in a warming climate.