Rain-on-snow (ROS) is a hydrometeorological phenomenon in which liquid precipitation falls onto an existing snowpack, augmenting runoff through the combined effects of rainfall and accelerated snowmelt. Anthropogenic climate change is progressively shifting the rain-to-snow partitioning of precipitation and altering land-surface conditions across mid- to high-latitude mountainous regions, thereby heightening flood potential. Most previous work, however, has addressed ROS at regional scales and over historical periods; hemispheric-scale assessments of future ROS dynamics and their implications for flood hazard and societal exposure remain scarce. Here we apply 10 bias-corrected CMIP6 models together with ERA5-Land reanalysis data to project changes in ROS days across the Northern Hemisphere under four Shared Socioeconomic Pathway (SSP) scenarios. ROS days are coupled with flood frequency analysis to quantify changes in ROS flood occurrence, and gridded population and Gross Domestic Product (GDP) data are integrated to evaluate future population-socioeconomic exposure. Under low-to-medium emission scenarios, ROS days increase substantially over historical hotspots, whereas under high-emission scenarios they decline at mid- to high latitudes yet expand into previously unaffected high-latitude and inland cold regions. ROS flood days respond nonlinearly to ROS frequency because progressive snow water equivalent loss limits runoff generation, causing ROS floods to decrease in some mountainous areas even as ROS events become more frequent. Population-socioeconomic exposure exhibits a corresponding polarization: it declines in mid-latitude regions where snow cover is disappearing but rises sharply at high latitudes, with high-emission pathways accelerating the northward migration of disaster risk. These findings bridge critical gaps in large-scale ROS climatology and shed light on future changes in ROS-induced hydrological extremes. Besides, the findings facilitate the creation of regionally focused adaptation strategies and provide useful references for integrating climate model projections with remote sensing observations to improve future monitoring and risk assessment of ROS-related floods.
Feng et al. (Mon,) studied this question.