SPATIAL AND TEMPORAL VARIABILITY OF RIVERBED HYDRAULIC CONDUCTIVITY AT AN INDUCED INFILTRATION SITE

Under USEPA's Long Term 1 Enhanced Surface Water Rule, groundwater with characteristics closely correlated to nearby surface water can be designated as groundwater under the direct influence of surface water (GWUDI), resulting in required water treatment akin to surface-water treatment. Some recent studies, however, demonstrate contaminant attenuation due to riverbank filtration at suspect sites equal to or better than that achieved in surface-water treatment plants. As the USEPA prepares to more precisely define GWUDI in the Long Term 2 Enhanced Surface Water Rule, there is concern that while adequate attenuation may be achieved most of the time, the natural filtration capacity can be diminished due to riverbed scour during peak flow events. To better understand the dynamic nature of the riverbed and its role in riverbank filtration, this study investigates the spatial and temporal variability of riverbed hydraulic conductivity at a site of induced infiltration along the Great Miami River in southwest Ohio. At this site, 10 wells pump water from the outwash aquifer to supply 10% of Cincinnati's drinking water. A variety of methods are being employed to quantify riverbed conductivity including seepage metering, piezometer slug tests, and laboratory permeameter tests of sediment samples. Scour chains, load-cell sensors, and cross sectional profiles are being employed to quantify scour. Measures of scour in conjunction with conductivity measures should allow estimation of the impact of storm events. The hydraulic conductivity of the riverbed obtained through seepage metering before and after peak events indicates a highly variable, heterogeneous, less conductive (0.45 m/day) sediment layer overlying a more permeable cobble layer. The fine-sediment layer ranges in thickness from 0.15 to 0.22 meters. Field measurements of scour and deposition obtained through the use of scour chains and cross-sectional profiles indicate event-driven scour as great as 0.075 meters, with final composition ranging from net deposition of up to 0.10 meters to a net scour of 0.050 meters. These results will be compared to load-cell censor data from point bar depositional features. Temperature modeling of the riverbed over various periods will augment these methods and provide independent measures of hydraulic conductivity.