Snow depth and streamflow relationships in large North American watersheds

Abstract

Snowmelt runoff in the spring is an important component in regional hydrologic systems in the northern United States and Canada, having a vital influence on water resources. In northern latitude rivers, snowmelt runoff provides a considerable volume of freshwater to drive circulation in the Arctic Ocean. This project defines and analyzes patterns of snow volume and discharge in major North American watersheds and determines the strength and form of the associated relationships. The results are used to develop statistical models applicable to individual watersheds, including the Yukon and Mackenzie basins in northern Canada and Alaska, the Saskatchewan basin in southern Canada, and the Missouri and upper Mississippi watersheds in the northern United States. It is shown that snow volume can predict winter and early spring discharge in all watersheds in the study region, with the best model performance in the higher-latitude Yukon and Mackenzie basins during late fall and winter accumulation. In the lower-latitude Missouri and upper Mississippi basins, despite additional influences of rain on discharge patterns, the statistical models based on snow volume were still able to estimate streamflow with percent relative error around 50%. To improve modeled discharge estimates during peak spring runoff, additional snow cover variables, including the value and timing of peak snow volume and the duration of snowmelt, were compared with peak annual discharge during periods of intense snowmelt. Significant results were found to occur in the Yukon and Saskatchewan basins owing to the extreme ensitivity to snowmelt runoff and fast river response times. Copyright 2008 by the American Geophysical Union.