Incorporating Watershed-Scale Groundwater/Surface Water Interactions to Better Understand How ENSO/PDO Teleconnections Affect Streamflow Variability in Geologically Complex, Semiarid, Snow-Dominated Mountainous Watersheds

Abstract:

In the Southwestern U.S., warm anomalies in the El Nino-Southern Oscillation (ENSO) are associated with increased probability of wetter than normal winter precipitation. For semiarid, snow-dominated mountainous watersheds, teleconnections, such as ENSO, may strongly affect the magnitude and timing of snowmelt pulses in streamflow. In examining stream-gage data, an implicit assumption is made that all the streamflow generation processes operative within the watershed are captured by the stream gage. However, zones of strong groundwater discharge to the stream alternating with zones of strong recharge from the stream may emerge in geologically complex watersheds. The spatial complexity of these groundwater/surface water interactions may not be captured in the stream-gage discharge data. This may not be a problem in watersheds where streamflow is generated primarily by shallow, fast runoff processes. In that case, changes associated with ENSO can be quickly apparent in streamflow (i.e., an increase in snowpack associated with warm ENSO anomalies will quickly translate to increases in daily and peak streamflow). However, the spatial complexity of groundwater/surface water interactions creates a problem in geologically complex watersheds where interactions with deep, regional groundwater are present. In this case, we test the hypothesis that the combined effect of complex geology and deep groundwater interactions creates phase shifts between peak snowpack, onset and peak of snowmelt pulses, and teleconnection indices. Using time-series analysis, the relationships between teleconnections, and metrics for snowpack and streamflow are evaluated for selected watersheds in New Mexico, Arizona, and Colorado. A phase shift (lag) is observed between the Oceanic Nino Index (ONI) and onset and peak of snowmelt pulses in streamflow in snow-dominated watersheds with complex geology across scales of 50 to 1600 km². Additionally, strong relationships between teleconnections and streamflow are not observed at larger snow-dominated watersheds with known significant deep groundwater contributions. We conclude that the effects of complex geology and deep groundwater interactions must be considered when quantifying the effects of ENSO and other teleconnections on the timing of the onset of the snowmelt pulse.