This study examines the changing roles of temperature and precipitation on snowpack variability in the Northern Hemisphere for Second Generation Canadian Earth System Model (CanESM2) historical and future (2006-2100) climate simulations. The strength of the linear relationship between monthly snow water equivalent (SWE) in January-April and precipitation P or temperature T predictors is found to be a sigmoidal function of the mean temperature over the snow season up to the indicated month. For

more »

... predictors, the strength of this relationship increases for colder snow seasons, whereas for T predictors it increases for warmer snow seasons. These behaviors are largely explained by the daily temperature percentiles below freezing during the snow accumulation period. It is found that there is a threshold temperature (25618C, depending on month in the snow season and largely independent of emission scenario), representing a crossover point below which snow seasons are sufficiently cold that P is the primary driver of snowpack amount and above which T is the primary driver. This isotherm allows one to delineate the snowclimate regions and elevation zones in which snow-cover amounts are more vulnerable to a warming climate. As climate projections indicate that seasonal isotherms shift northward and toward higher elevations, regions where snowpack amount is mainly driven by precipitation recede, whereas temperature-sensitive snowcovered areas extend to higher latitudes and/or elevations, with resulting impacts on ecosystems and society.