| Biochemical, sulfate (SO42-) reducing permeable reactive barriers (PRBs) containing decomposable solid organic materials can be used for in-situ bioremediation of ground water impacted by acid mine drainage (AMD). A new biochemical model for bioremediation of AMD based on SO42- reduction and precipitation of metals coupled to, and limited by, the decomposition of solid organic materials (e.g., leaf mulch compost, wood chips, sawdust, alfalfa) was formulated and integrated either as a stand-alone algorithm or coupled to flow and multi-species contaminant transport (i.e., MODFLOW with RT3D). The model combines the primary processes of (i) decomposition of polysaccharides in solid organic materials, (ii) SO42- reduction based on lactate, (iii) precipitation of insoluble metal sulfides, (iv) partial volatilization of hydrogen sulfide (H 2S) to the gas phase, and (v) reversible dissolution-precipitation of calcite (CaCO3(s)) and metal carbonates (e.g., siderite, FeCO 3(s)). The model was evaluated and calibrated using experimental data from the literature pertaining to four batch-equilibrium (no-flow) tests, two closed-system column tests, three open-system bioreactor tests, and a full-scale PRB. The comparisons with experimental data are performed in terms of both effluent and resident concentrations and cumulative mass, and provide insight into the effects of different model kinetics, slow solid decomposition, multiple-phase precipitation, and temperature and seasonality, as well as the appropriate ranges of values for a variety of model parameters.; Additionally, as common to all PRBs, heterogeneity in aquifer hydraulic conductivity (K) also influences the performance of biochemical PRBs, with preferential flow and contaminant transport exposing the PRB to spatially variable groundwater seepage velocities at the location of the PRB, with localized values that may be sufficiently high such that the required PRB thickness is greater than that evaluated for homogenous aquifer. Also, heterogeneity induces divergent flow patterns such that the PRB length required to capture or encompass the contaminant plume at the location of the PRB is greater than that for homogenous aquifer. New probabilistic factors of safety related to scaling the PRB thickness and length amidst heterogeneity were quantified using flow and particle tracking (i.e., MODFLOW with MODPATH) in conjunction with 650 synthetic stochastic heterogeneous aquifers (Turning Bands method) for different levels of aquifer heterogeneity, correlation structure anisotropy ratios, and distances from source zone to PRB. The significant magnitudes of such scaling factors emphasize the importance of site characterization as a prerequisite to PRB design. |
