Hillslope-Scale Spatial Variability of Snowmelt Runoff via Interflow in a Minnesota Forest

English Abstract:

In humid temperate catchments with permeable surface soils underlain by confining soil horizons, lateral interflow is an important runoff mechanism, particularly during the spring snowmelt period. Numerous physical and biological factors affect the spatial variability of interflow runoff. From a hydro-biogeochemical perspective, flow path (e.g. soil contact), residence time (affecting soil-water partitioning of chemicals), and magnitude of flow (controlling how much water or solute is delivered to a particular area) are important, but spatially-variable factors controlling downstream or down-gradient biogeochemical processing. To characterize the spatial variability of interflow across a north-facing forested hillslope, three separate 10 m wide interflow collecting trenches were dug to the confining layer across an approximate 75 m wide hillslope area. Continuous measurements of runoff, and spatially distributed perched water table elevations, soil moisture, and snowmelt rates were made through the spring snowmelt period. Stable isotope-based hydrograph separation was also applied to each of the hillslope plots to assess variability in event vs. pre-event water signatures. The runoff response to a rain-on-snow event (112 mm total input) varied from 40 to 72 mm. While hydrograph shapes and inter-site runoff amounts differed considerably during the rising limb, broad characteristics such as peak flow lag times were similar (45-47 hours). Hydrograph recession analysis revealed three distinct drainage periods; an early, relatively rapid draining of nearer-surface, high conductivity soils over a period of 15-30 hours, a middle recession ranging from 45 to 110 hours that was distinctively longest at the hillslope plot where toe slope soils were comparably deepest, and a longer draining component that is too ephemeral to be characterized as baseflow. Variability in the contribution of event water across the hillslope was high during the rising limb, but relatively low (~60 ± 20%) after peak flow. Overall, this work provides a needed baseline from which the uncertainty of future impacts can be quantitatively assessed and demonstrates the likelihood that small-scale differences in soil depth, particularly at the toe slope, are important to runoff response at this scale.