Abstract Archive Select a year below to view:
Hydrogeologic Thresholds Affect Groundwater-Surface Water Interactions of the Big Sunflower River at Sunflower, Mississippi
Proceedings of the 2019 Mississippi Water Resources Conference
Year: 2019 Authors: O'Reilly A.M., Holt R.M., Davidson G.R., Patton A., Rigby J.R., Barlow J.
At a site near Sunflower, Mississippi, observations of Big Sunflower River stage and adjacent groundwater level indicate a disproportionately larger response in groundwater level above a certain stage elevation. This suggests the river contributes water to the aquifer at a higher rate above a threshold that defines high-stage versus lower-stage periods. This is in contrast with the common assumption of a linear relation between river leakage and water-level difference between the river and aquifer. Heterogeneity in the riverbed or near-field geology can impart such a threshold effect on groundwater-surface water (GW-SW) exchange. Variations in the texture of riverbed sediments and lithologic variations in underlying geology are examples of common heterogeneities. Hydrologic interaction with these heterogeneities leads to distinct types of behavior that switch when river stage or groundwater level rises above or falls below the interface. A simple dynamic water-balance (linear reservoir) model was developed to investigate this phenomenon at the study site on the Big Sunflower River. Four conceptual models, each of which consists of a perched aquifer that receives recharge from the riverbank and loses water to the underlying Mississippi River Valley alluvial aquifer, were tested: homogenous riverbank and aquifer lithology, two-layer riverbank and homogenous aquifer lithology, two-layer riverbank and two-layer aquifer lithology, and homogenous riverbank and two-layer aquifer lithology. Models were run using hourly observed river stage, calibrated to a 382-day period of water-level measurements in a nearby well, and rerun for the entire 1,278-day period of record. All models matched observed groundwater levels reasonably well, with a maximum root-mean-square error (RMSE) of 0.46 m. However, the heterogeneous models matched high-stage events substantially better than the homogeneous model, with the best performance (RMSE of 0.27 m) by the model incorporating threshold effects controlled by both two-layer riverbank and two-layer aquifer lithology. Substantial flow occurs through both the upper riverbank, representing about 30% of the flow through the lower riverbank, and the upper aquifer, representing about 25% of flow through the lower aquifer. These results illustrate the importance of considering threshold effects on GW-SW interactions. Regional-scale, numerical, groundwater models typically do not incorporate threshold effects, because the causative heterogeneities exist at a scale smaller than a single grid cell. Models such as those developed here can be incorporated into numerical groundwater models to better simulate intra-cell processes that cannot be effectively modeled at the regional scale.