Coupling snowpack and groundwater dynamics to interpret historical streamflow trends in the western United States

Authors


Correspondence to: Gordon Grant, PNW Research Station, USDA Forest Service, Corvallis, OR 97331, USA.

E-mail: ggrant@fs.fed.us

Abstract

A key challenge for resource and land managers is predicting the consequences of climate warming on streamflow and water resources. During the last century in the western United States, significant reductions in snowpack and earlier snowmelt have led to an increase in the fraction of annual streamflow during winter and a decline in the summer. Previous work has identified elevation as it relates to snowpack dynamics as the primary control on streamflow sensitivity to warming. But along with changes in the timing of snowpack accumulation and melt, summer streamflows are also sensitive to intrinsic, geologically mediated differences in the efficiency of landscapes in transforming recharge (either as rain or snow) into discharge; we term this latter factor drainage efficiency. Here we explore the conjunction of drainage efficiency and snowpack dynamics in interpreting retrospective trends in summer streamflow during 1950–2010 using daily streamflow from 81 watersheds across the western United States. The recession constant (k) and fraction of precipitation falling as snow (Sf) were used as metrics of deep groundwater and overall precipitation regime (rain and/or snow), respectively. This conjunctive analysis indicates that summer streamflows in watersheds that drain slowly from deep groundwater and receive precipitation as snow are most sensitive to climate warming. During the spring, however, watersheds that drain rapidly and receive precipitation as snow are most sensitive to climate warming. Our results indicate that not all trends in western United States are associated with changes in snowpack dynamics; we observe declining streamflow in late fall and winter in rain-dominated watersheds as well. These empirical findings support both theory and hydrologic modelling and have implications for how streamflow sensitivity to warming is interpreted across broad regions. Copyright © 2012 John Wiley & Sons, Ltd.

Ancillary