We have examined the distribution of microbial iron reduction potential in coastal plain sediments of the Delmarva Peninsula at a field site near Oyster, Virginia. The stratigraphy at the site consists of sandy shoreface sediments interfingered with back-barrier muds and peat. Several boreholes were sampled below the water table at approximately 15 cm intervals over a depth of about 6.5 m. We measured the microbial iron reduction potential (MIRP) by determining the amount of Fe(II) evolved during anaerobic incubation of the sediment in a buffered slurry containing artificial groundwater, substrate, and nutrients. We also analyzed the samples for hydraulic conductivity, grain size, organic matter, bulk density, porosity, and extractable iron (extractions included 0.5 M HCl and citrate-dithionite). We analyzed the data using multivariate statistics, neural networks, and geostatistics.

Results indicate that MIRP at the site occurs primarily in sandy sediment, and does not occur in peat or mud samples. The peat and mud sediment samples had relatively high organic content but very low concentrations of extractable ferric iron. Even within the sand and muddy sand samples the distribution of positive MIRP was very heterogeneous, with 70% of the muddy sand samples and 39% of the sand samples exhibiting no activity. Examination of results from the positive samples indicate that activity was associated with coarser grain size, less clay, and higher concentrations of extractable ferric iron.

Paired incubations were also performed in which the sediment was amended with artificial hydrous ferric oxide (HFO) as an electron acceptor. The incubations with added HFO indicate that although microbial iron reducers are not active in the peat and mud under current conditions, they are present and remain capable of reducing ferric iron.

Because geology and geochemistry exert a strong influence on the distribution of MIRP, the microbial response is spatially continuous with a correlation range of about 1 m, similar to the range of the hydraulic conductivity and about half that of extractable ferric iron content (~ 2 m). We have used models of this spatial continuity to predict the hydraulic conductivity and MIRP at unsampled locations, which we plan to use as the basis for modeling reactive transport at the site.