TEMPORAL VARIABILITY OF RIVERBED HYDRAULIC CONDUCTIVITY AT A WELLFIELD IN SOUTHWEST OHIO

The dynamic nature of a riverbed was investigated at the Bolton wellfield, a site of induced infiltration along the Great Miami River in southwest Ohio. The wellfield lies in a glacial-outwash, buried-valley aquifer and comprises 10 municipal wells supplying 12% of the water for the City of Cincinnati. Our study focused on a single site associated with one of the supply wells. The principal questions were to what degree high-stage events scour the riverbed and to what extent scour changes the overall hydraulic conductivity of the system.

Horizontal and vertical hydraulic conductivity (Kv and Kh) of the riverbed and underlying sediment were measured using seepage meters, slug tests, and laboratory permeameter tests of split-spoon core samples. Modeling of temperature fluxes between the river and aquifer allowed estimation of the overall conductivity of the system. Riverbed scour and deposition were measured using cross-sectional profiles, scour chains, and a load-cell pressure sensor buried in the riverbed.

29 seepage-meter Kv values had a geometric mean of only 0.12 m/day and showed no temporal trends. Slug tests performed at depths below the riverbed of between 0.7 and 2.2 m yielded Kh values averaging 19.2 m/d (Kv of perhaps 9.6 m/d). Three split-spoon samples were taken to depths not exceeding 0.12 m before hitting very coarse material. Permeameter-based Kv of these samples averaged 6.0 m/d. Scour-chains indicated that high-stage events (0.34 to 7.05 m above base-flow stage) resulted in a maximum of 0.061 m of scour. Load-cell pressure sensor data likewise indicated that fluctuation of sediment height during the study period did not exceed about 0.05 m.

Based on these measurements we hypothesized a three-layer conceptual model. The top 0.06 m of the riverbed is transient with a Kv of about 6 m/d. Underneath is a stable, very low Kv (~0.1 m/d) colmation layer comprising large gravel and cobbles but clogged with fine sediment. The large particles protect this layer from even very high-stage events. Underneath this layer, the Kv increases dramatically. Temperature modeling indicates the overall effective Kv of this system is about 3 m/d. The loss of the transient layer does not have much impact of the overall conductivity of the system. Loss of the colmation layer would have a much greater effect, but has not been observed.