Reactive transport modeling of early diagenesis in a reservoir lake affected by acid mine drainage: Trace metals, lake overturn, benthic fluxes and remediation

The Sancho Reservoir in SW Spain has been impacted by acid mine drainage (AMD) since the Tharsis mine stopped activity in 1998. As a result, the reservoir exhibits low pH (~3.5) and high aqueous concentrations of sulfate, aluminum, iron and trace metals. Thus far, removal of contaminants by sediment burial has not been as effective as expected in improving water quality within the reservoir. To inform potential remediation strategies, a 1-D, non-steady-state reactive transport model with a comprehensive set of equilibrium and kinetic biogeochemical reactions is used to simulate the fate of trace metals and acidity in sediments affected by AMD. Two realizations of the model account for the spatial heterogeneity of bottom water oxygenation. A “permanently oxic” model represents shallow sediments above the thermocline, while a “holomictic” model represents the deeper sediments where bottom water oxygen levels oscillate between completely anoxic and oxic as a result of water-column overturn. The model is calibrated against an extensive dataset on the depth distributions of pore water and solid phase species. Model results imply that, under permanently oxic conditions, the sediments act as a sink for acidity (H+) and aqueous Al, Zn, Cu, Co and Ni, but act as a source of aqueous Mn, Fe and As. The latter are released to the overlying water as a result of Mn and Fe (oxy)hydroxide reductive dissolution in the sediments. Below the thermocline, when bottom waters become anoxic, metal sulfides precipitate in the sediment. When the bottom waters subsequently become oxic, the metal sulfides are oxidized along the downward-penetrating oxygen front and the associated metals are released to the overlying water. On the order of 35% of the sediment pools of sulfide-bound Zn, Cu, Co and Ni, and ~25% of FeS are thus reoxidized. However, overall the sediments act as a net sink for the pollutants considered in the model. On an annual basis, about 10% of the total elemental masses of S, Al, Zn and Cu present in the water column of the reservoir are removed by burial in the sediments, but only ~2% for Co and Ni. For Fe, Mn and As, the corresponding values are 80, 70 and 98% respectively. The model predicts that, if AMD input to the reservoir were to completely cease, the sediments would reach a new steady state with negligible release of aqueous contaminants to the overlying water column within a few years.