Temporal variations of the sulfur and carbon isotope signatures during ASR of aerobic groundwater in a pyritic aquifer: Development of a reactive transport modelling framework and application to a field data set

Henning Prommer, CSIRO Land and Water, Floreat, WA, and The University of Western Australia.
Co-authors: Carlos Descourvieres, Simone Seibert, Grzegorz Skrzypek

ABSTRACT

Oxidation of sedimentary pyrite and organic carbon are the main drivers for water quality changes during managed aquifer recharge operations in anaerobic aquifers, when aerobic water is injected (e.g., Hartog et al., 2002, Prommer and Stuyfzand, 2005, Descourvieres et al., 2010). Particularly the oxidation of pyrite may trigger secondary reactions such as mineral buffering (e.g., by carbonates and Fe(II)-aluminosilicates) and trace metal release. These processes can be recognised indirectly by monitoring changes in the sulphur and carbon isotope compositions (d13C, d34S) in the groundwater. Subsequently geochemical and reactive transport modelling can be used for a detailed, spatially and temporally resolved quantification of these processes.

In the present study, we extend the existing reactive multi-component transport model PHT3D (Prommer et al., 2003) to allow the quantification of isotope mixing and fractionation processes that affect  temporal and spatial changes in d13C and d34S values during the injection of aerobic water into deep, anaerobic aquifers. Under these conditions d34S is predominantly affected by pyrite oxidation, which causes a 34S-depletion. The d13C value is mainly controlled by organic matter oxidation as well as the carbonate buffering reactions that are triggered during pyrite oxidation.

The model is applied to the interpretation of data collected during an ongoing aquifer storage and recovery (ASR) experiment in the anoxic, pyritic Leederville aquifer beneath Perth, Western Australia, where aerobic groundwater from a shallow aquifer is used as injectant. The monitored stable isotope composition of two elements (C, S) in three chemical phases (DIC, DOC and SO4) is used to constrain the calibration of the numerical model and to underpin our conceptual understanding of the predominant reactive processes.