Elucidating the relationship between gaseous O2 and redox potential in a soil aquifer treatment system using data driven approaches and an oxygen diffusion model

Knowledge concerning the redox potential (Eh) conditions in the vadose zone of a soil aquifer treatment (SAT) system constitutes valuable information for assessing water quality and operational efficiency. The complex nature of Eh conditions in SAT limits the ability to predict and quantify them using detailed models. Alter-natively, data-driven models can be used for predictions and relationship analysis. Hourly measurements of Eh, volumetric water content (theta), soil temperature (T), and gaseous oxygen (O2) were obtained at multiple depths of a SAT vadose zone. A correlation analysis showed that O2 correlated with Eh for most temporal components. Only the monthly component of T and the daily component of theta were correlated with Eh. A detailed multiple linear regression (MLR) analysis illustrated that the gaseous O2, at shallow depths, can explain the majority (above 80%) of the Eh variability. The MLR curve demonstrated breakpoints in the Eh response to O2 at shallow depths, which were identified using a piecewise regression. These breakpoints explain, in part, the different stages of microbial activity throughout the SAT wetting and drying cycles. Combining an oxygen transport analytical model with the piecewise regression enabled the Eh prediction through easy-to-acquire T and theta measurements. At deeper depths, the Eh-O2 relationship demonstrates a step-function characteristic, which indicates that the changes in Eh occur due to the arrival of a low-Eh solution. Thus, the Eh dynamic in the SAT vadose zone is mostly controlled by aerobic conditions.