Do CSIA data from aquifers inform on natural degradation of chlorinated ethenes in aquitards?

Back-diffusion of chlorinated ethenes (CEs) from low-permeability layers (LPLs) causes contaminant persistence long after the primary spill zones have disappeared. Naturally occurring degradation in LPLs lowers remediation time frames, but its assessment through sediment sampling is prohibitive in conventional remediation projects. Scenario simulations were performed with a reactive transport model (PHT3D in FloPy) accounting for isotope effects associated with degradation, sorption, and diffusion, to evaluate the potential of CSIA data from aquifers in assessing degradation in aquitards. The model simulated a trichloroethylene (TCE) DNAPL and its pollution plume within an aquifer-aquitard-aquifer system. Sequential reductive dechlorination to ethene and sorption were uniform in the aquitard and did not occur in the aquifer. After 10 years of loading the aquitard through diffusion from the plume, subsequent source removal triggered release of TCE by back-diffusion. In the upper aquifer, during the loading phase, δC-TCE was slightly enriched (up to 2‰) due to diffusion effects stimulated by degradation in the aquitard. In the upper aquifer, during the release phase, (i) source removal triggered a huge δC increase especially for higher CEs, (ii) moreover, downstream decreasing isotope ratios (caused by downgradient later onset of the release phase) with temporal increasing isotope ratios reflect aquitard degradation (as opposed to downstream increasing and temporally constant isotope ratios in reactive aquifers), and (iii) the carbon isotope mass balance (CIMB) enriched up to 4‰ as lower CEs (more depleted, less sorbing) have been transported deeper into the aquitard. Thus, enriched CIMB does not indicate oxidative transformation in this system. The CIMB enrichment enhanced with more sorption and lower aquitard thickness. Thin aquitards are quicker flushed from lower CEs leading to faster CIMB enrichment over time. CIMB enrichment is smaller or nearly absent when daughter products accumulate. Aquifer CSIA patterns indicative of aquitard degradation were similar in case of linear decreasing rate constants but contrasted with previous simulations assuming a thin bioactive zone. The Rayleigh equation systematically underestimates the extent of TCE degradation in aquifer samples especially during the loading phase and for conditions leading to long remediation time frames (low groundwater flow velocity, thicker aquitards, strong sorption in the aquitard). The Rayleigh equation provides a good and useful picture on aquitard degradation during the release phase throughout the sensitivity analysis. This modelling study provides a framework on how aquifer CSIA data can inform on the occurrence of aquitard degradation and its pitfalls.