Preferential Flow Effects on Subsurface Contaminant Transport in Alluvial Floodplains

Abstract

For sorbing contaminants, transport from upland areas to surface water systems is typically considered to be due to surface runoff, with negligible input from subsurface transport assumed. However, certain conditions can lead to an environment where subsurface transport to streams may be significant. The Ozark region, including parts of Oklahoma, Arkansas, and Missouri, is one such environment, characterized by cherty, gravelly soils and gravel bed streams. Previous research identified a preferential flow path (PFP) at an Ozark floodplain along the Barren Fork Creek in northeastern Oklahoma and demonstrated that even a sorbing contaminant, i.e., phosphorus, can be transported in significant quantities through the subsurface. The objective of this research was to investigate the connectivity and floodplain-scale impact of subsurface physical heterogeneity (i.e., PFPs) on contaminant transport in alluvial floodplains in the Ozarks. This research also evaluated a hypothesis that alluvial groundwater acts as a transient storage zone, providing a contaminant sink during high stream flow and a contaminant source during stream baseflow. The floodplain and PFP were mapped with two electrical resistivity imaging techniques. Low-resistivity features (i.e., less than 200 O-m) corresponded to topographical depressions on the floodplain surface, which were hypothesized to be relict stream channels with fine sediment (i.e., sand, silt, and clay) and gravel deposits. The mapped PFP, approximately 2 m in depth and 5 to 10 m wide, was a buried gravel bar with electrical resistivity in the range of 1000 to 5000 O-m. To investigate the PFP, stream, and groundwater dynamics, a constant-head trench test was installed with a conservative tracer (Rhodamine WT) injected into the PFP at approximately 85 mg/L for 1.5 h. Observation wells were installed along the PFP and throughout the floodplain. Water table elevations were recorded real-time using water level loggers, and water samples were collected throughout the experiment. Results of the experiment demonstrated that stream/aquifer interaction was spatially non-uniform due to floodplain-scale heterogeneity. Transport mechanisms included preferential movement of Rhodamine WT along the PFP, infiltration of Rhodamine WT into the alluvial groundwater system, and then transport in the alluvial system as influenced by the floodplain-scale stream/aquifer dynamics. The electrical resistivity data assisted in predicting the movement of the tracer in the direction of the mapped preferential flow pathway. Spatially variable PFPs, even in the coarse gravel subsoils, affected water level gradients and the distribution of tracer into the shallow groundwater system.