Dissecting the variable source area concept - Flow paths and water mixing processes

Variable source areas (VSAs) are hot spots of hydrological (saturation excess runoff) and biogeochemical processes (e.g. nitrogen, phosphorus, organic carbon cycling) in the landscapes of the northeastern U.S. The prevalence of shallow, highly transmissive soils, steep topography, and impeding layers in the soil (i.e. fragipan) have long been recognized as first-order controls on VSA formation. Nevertheless, there is still understanding to be gained by studying subsurface flow processes in VSAs. Thus, we instrumented (trenched) a 0.5 ha hillslope in the southern tier of New York State, U.S.A. and measured water fluxes in the trench, upslope water table dynamics, surface and bedrock topography in conjunction with isotopic and geochemical tracers in order to four-dimensionally characterize (XYZ and Time) subsurface storm flow response within the VSA for five storm events. We used tracer-based hydrograph separation models and physically measured flow components to separate temporally (i.e. event and pre-event) and spatially shallow water from above the fragipan layer (including both surface runoff and shallow interflow) and deeper water from below the fragipan layer. Shallow water (event/pre-event) contributions were greatest during storms with wet antecedent conditions and large rainfall amounts (> 15 mm), when soils above the fragipan were saturated, prohibiting deep percolation through cracks in the fragipan. Shallow water contributions were well correlated to the saturated contributing area. During these events, the pre-event shallow water peaked on the rising and falling limb, which can be explained by flushing of pre-event water from macropores on the rising limb and subsequent drainage of pre-event water from micropores into macropores on the falling limb. During events with dry antecedent conditions, greater amounts of event water (24 - 28 %) are proportionally contributed by surface runoff in the top 10 cm of the soil through macropores than by shallow interflow from the soil-fragipan interface. Pre-event deeper water contributions to total trench discharge varied between 15 and 65% but were independent of total rainfall amounts, rainfall intensities, and water table dynamics. Our results have important implication for the protection of streams from dissolved pollutant transport and recommend that preference be given to variable-width buffers over fixed-width stream buffers.