When DS Does Not Equal 0: A Multi-year Climate Signal, Mediated through Groundwater, Controls Hydrologic Response to Climate Change in Seasonally Snow-covered Mountain Headwaters.

A common, often implicit, assumption in assessing climate change impacts on water resources is that changes in groundwater storage are sufficiently small relative to streamflow and evapotranspiration (ET) that they do not play a significant role in interannual variability in water supply. However, a rapidly growing body of research from catchment hydrology suggests that dynamic, multi-year groundwater storage is responsible for the common observation that catchments rapidly release groundwater in response to rainfall or snowmelt. Similarly, recent ecohydrological research invokes the presence of variable groundwater stores in explaining patterns in vegetation growth and multiyear resistance to drought. Because inferences from these disparate fields both suggest that the residence time of this water is longer than a year and that it is closely coupled to both ET and streamflow, our objectives were to: 1) quantify the interannual variability in size of this stored water, and 2) evaluate the influence of this storage on hydrologic partitioning and annual water yield.

We focused on twelve, seasonally snow-covered, headwater catchments in Utah, USA that are the primary water sources for over 1 million urban residents and a diverse range of agriculture. Using over a century of climate and streamflow data, supplemented by hydrological tracers, we demonstrate that January baseflow, measured under seasonal snow cover, is a reliable metric of dynamic groundwater storage. Interannual variability in this storage followed a consistent temporal pattern across all catchments with the timing related to multi-year climate variability. Variability in this storage was the strongest predictor of water yield in 10 out of 12 catchments and a multiple linear regression model combining antecedent storage with melt dynamics was able to explain greater than 90% of the variability in annual water yield across all catchments. Together, these results suggest the presence of large, variable, and dynamic groundwater stores that provide a multi-year climate memory to hydrologic partitioning and annual water supply. Assuming that changes in storage on an annual time step are insignificant misses these long term climate signals which may exacerbate or attenuate hydrological response to a changing climate.