INVESTIGATING THE TEMPORAL VARIABILITY OF RIVERBED HYDRAULIC CONDUCTIVITY USING TEMPERATURES

In this study we investigated the temporal variability of the riverbed hydraulic conductivity in a stream-aquifer system using temperature data. The site of the study is the Bolton wellfield, which is a site of induced infiltration along the Great Miami River in Fairfield, Ohio. The working hypothesis for the study is that the Great Miami riverbed generally has a lower hydraulic conductivity than the underlying glacial outwash aquifer. During storm events, we hypothesized that scour takes place, cleansing the riverbed of some of its clogging, thereby increasing its conductance and the amount of river water that infiltrates into the aquifer. This effect is important because scour mostly occurs during times of maximum hydraulic-head gradient and therefore during times with the greatest driving force of river-water infiltration. In this study, we estimate riverbed hydraulic conductivity and its small-scale temporal variability using temperature profiling and modeling. The results from temperature modeling are then compared to riverbed hydraulic conductivity values obtained from seepage meters, slug tests and constant head permeameter tests. Temperature and water level data from the river and groundwater were continuously monitored using thermistors, mini-trolls and a load cell for about two years. Data from two riverbed piezometers, two monitoring wells and the Great Miami River were used in a two-dimensional groundwater flow and heat transport model in VS2DH, a graphical software package that simulates heat transport and groundwater through variably saturated porous media. Riverbed hydraulic conductivity estimates were determined by adjusting the model's riverbed hydraulic conductivity value in order to match simulated groundwater temperatures to observed temperatures in the piezometers and monitoring wells. Preliminary results from temperature modeling show the average riverbed hydraulic conductivity to vary about two orders of magnitude (from 10-1 to 100 meters/day). These results are comparable to results from seepage metering, permeameter and slug tests, which yielded averages hydraulic conductivities of 0.12m/day, 6m/day and 9.6m/day respectively. If the higher conductivity values correspond to periods of high river stage then there is sufficient evidence to support our hypothesis above.