Specific Conductivity Synoptic Surveys to Map Groundwater-Surface Water Discharges in a Lowland River

Abstract:

Quantifying distributed groundwater-surface water (GW-SW) discharges at the watershed scale is a challenging but important aspect of tracking nonpoint-source pollutants and their potential impacts on riverine water quality. This work analyzes high spatiotemporal longitudinal river sampling along a 38 km reach of the lower Merced River (LMR) in California’s San Joaquin River (SJR) watershed and estimates GW-SW discharges using a simple discretized mixing model. The LMR flow is dam-controlled upstream and has several distributed surface discharges from SW diversions, SW canal inputs, and potential GW discharges influenced by the surrounding agricultural and municipal landscapes. We collected longitudinal datasets covering a wide range of flows (1.3 to 150 m³/s), sampling georeferenced water specific conductance (SC) at intervals of less than 1 minute. Whole-reach gradients of SW specific conductance (SC), representing GW-SW salinity loading (mS/cm/river km), decreased in proportion to the flow and stage, indicating either (1) a simple dilution effect or (2) reduced GW-SW discharges due to a reduced GW-SW hydraulic gradient at higher stages. With respect to distributed discharges, local gradients were significant (p < 0.05), with estimated local salinity loads varying from -0.4 to 8.9 mS/cm/river km for the LMR conditions. Using local groundwater salinity estimates, we inverted a discretized mixing model to estimate distributed groundwater fluxes. Summation of the fluxes agreed well with stream gage estimates. However, local flux estimates were nonuniform, with mid-reach values consistently increasing to a maximum value at intermediate stream flows, then decreasing for greater flows. Causes for this behavior remain uncertain but may be due to one or more of the following: (1) uncertainty in the groundwater salinity data, (2) salinity reduction in the hyporheic zone (e.g., denitrification), and (3) access to preferential GW-SW flowpaths at intermediate flows and stages. Overall, the methods described provide an efficient and economical means for estimating distributed GW-SW nonpoint-source discharges at the watershed scale.