Watershed- to continental-scale influences on winter stormflow in the Southern Blue Ridge Mountains

Abstract

Spatial and temporal influences on the winter (December-March) stormflow characteristics of fifteen United States Geological Survey (USGS)-gaged watersheds in the Southern Blue Ridge Mountains are identified: (1) watershed-scale differences in geomorphology; (2) continental-scale teleconnections during periods of wetness/dryness (based on the relative amount of winter precipitation over a consistent 20 year dataset); and (3) land cover in the context of soil parent material (e.g., development on alluvium/colluvium). Multiple regression was used to determine how much variance could be explained in five hydrologic variables describing the flashiness of peak flow (three original metrics), total seasonal flashiness (Richards Baker flashiness index), and the ratio of total winter stormflow to total discharge (the stormflow index). Models were constrained to three uncorrelated (|0.65|) variables to prevent overfitting to the dataset. Average-, dry-, and wet-years were subset using the z-scores for winter precipitation derived from the 4 km monthly PRISM (Parameter-elevation Relationships on Independent Slopes Model) dataset, for the period of 1986–2006. Relief, slope, and landscape connectivity explain the majority of explained variance in all five of the hydrologic variables during all time periods. During dry-, average-, and wet-years, atmospheric circulation patterns (i.e., North Atlantic Oscillation and Pacific/North American Pattern) explain more variance than total seasonal precipitation (PRISM), which is not true in the majority of the all-years models. Land cover explains only a small portion of the variance in regional stormflow and only when sub-divided based on soil parent material. Results provide a framework for connecting watershed-scale characteristics to regional- and continental-scale processes.