Using Chemical and Physical Data to Delineate Storm-Event Flow Pathways in Lowland Watersheds
Griffin, Michael Patrick
Callahan, Timothy J
MetadataShow full item record
Land use and land cover changes have been impacting the hydrologic processes of coastal-plain watersheds of the southeastern United States. To better understand effects of land use change on these watershed processes, it is critical to understand the groundwater and surface-water interactions in the natural, undeveloped state. End-member mixing analysis (EMMA), a hydrogeochemical, multiple-tracer approach, was used for assessing contributing sources to stream flow. With EMMA, the stream is visualized as a constantly changing mixture of various source waters, or end-members. With the application of EMMA, we assume that stream chemistry changes rapidly, but end-member chemistry remains relatively constant. Though it has been used effectively in landscapes with a steeper topographic gradient, the use of EMMA in the Lower Coastal Plain (LCP) has produced mixed results. Delineating stormwater pathways in the LCP is challenging due to its unique characteristics, such as shallow water table and minimal hydraulic gradient. In this project, the application of EMMA to explain LCP hydrology was explored. The objectives of this study were: (1) to build EMMA models for comparing stream flow response to storm events in two first-order, forested, lowland watersheds through the use of chemical hydrograph separation, (2) to elucidate runoff processes in these watersheds, and (3) to use these findings to characterize stormwater pathways in coastal watersheds of different soil types.We hypothesized that hydrogeochemical modeling can permit estimation of end-member contribution to stream flow and that hydrological conditions in forested, lowland watersheds are impacted by watershed soil characteristics, antecedent soil moisture, and evapotranspiration conditions. To test these hypotheses, EMMA models were created for three sites. Upper Turkey Creek (UTC), in the Francis Marion National Forest, outside of Charleston, SC, drains 694 hectares. Adjacent to UTC, Watershed 80 (WS-80), in the US Forest Service's Santee Experimental Forest near Charleston, SC, drains 155 hectares characterized by clayey soils. Upper Debidue Creek (UDC), near Clemson University's Baruch Institute in Georgetown, SC, drains 162 hectares dominated by sandy soils. Samples from streams, water-table wells, piezometers, lysimeters, and rain gauges at these three sites were analyzed for major cation and anion concentrations; trends in tracer concentrations were assessed with principal components analysis (PCA); and results were used to construct EMMA models for each site.Hydrographs were created for fifteen storm events in a variety of conditions. Results showed that storm-event stream water at clayey WS-80 is 55%-67% rainwater, whereas rainwater only accounted for 28%-34% of storm-event stream water at sandy UDC. Thus, soil type played a critical role in hydrological processes; while groundwater processes dominated at the sandy UDC site, the clay layer at WS-80 limited groundwater flow, causing water to travel to the stream more rapidly and closer to the surface. Rainfall quantity was a poor predictor of runoff response; antecedent soil moisture (ASM), as indicated by riparian water-table depth, enhanced predictions of runoff in response to storm events. Evapotranspiration was shown to rapidly lead to dry ASM conditions in the growing season. Results underscored the importance of characterizing soil type, antecedent soil moisture, and evapotranspiration conditions when assessing the potential runoff response of storm events in minimally disturbed, lowland, forested watersheds. In a time of rapid land-use change, this study contributes to a need for a growing understanding of groundwater and surface-water interactions in these watersheds. This information is critical to land managers and policymakers who oversee the management of these areas.