Heterogeneous snowpack response and snow drought occurrence across river basins of northwestern North America under 1.0°C to 4.0°C global warming

Anthropogenic climate change is affecting the snowpack freshwater storage, with socioeconomic and ecological impacts. We present an assessment of maximum snow water equivalent (SWE max ) change in large river basins of the northwestern North America region using the Canadian Regional Climate Model 50-member ensemble under 1.0 °C to 4.0 °C global warming thresholds above the pre-industrial period. The projections indicate steep SWE max decline in the warmer coastal/southern basins (i.e., Skeena, Fraser and Columbia), moderate decline in the milder interior basins (i.e., Peace, Athabasca and Saskatchewan), and either a small increase or decrease in the colder northern basins (i.e., Yukon, Peel, and Liard). A key factor for these spatial differences is the proximity of winter mean temperature to the freeze/melt threshold, with larger SWE max declines for the basins closer to the threshold. Using the random forests machine-learning model, we find that the SWE max change is primarily temperature controlled, especially for warmer basins. Further, under a categorical framework of below-normal SWE max defined as snow drought (SD), we find that above-normal temperature and precipitation are the dominant conditions for SD occurrences under higher global warming thresholds. This implies a limited capacity of precipitation increase to compensate the temperature-driven snowpack decline. Additionally, the frequency and severity of SD occurrences are projected to be most extreme in the southern basins where current water demands are highest. Overall, the results of this study, including insights on snowpack changes, their climatic controls, and the framework for SD classification, are applicable for basins spanning a range of hydro-climatological regimes.