Modeling Versatile and Dynamic Anaerobic Metabolism for PAOs/GAOs Competition Using Agent-based Model and Verification via Single Cell Raman Micro-spectroscopy

Abstract

Side-stream enhanced biological phosphorus removal process (S2EBPR) has been demonstrated to improve performance stability and offer a suite of advantages compared to conventional EBPR design. Design and optimization of S2EBPR require modification of the current EBPR models that were not able to fully reflect the metabolic functions of and competition between the polyphosphate-accumulating organisms (PAOs) and glycogen-accumulating organisms (GAOs) under extended anaerobic conditions as in the S2EBPR conditions. In this study, we proposed and validated an improved model (iEBPR) for simulating PAO and GAO competition that incorporated heterogeneity and versatility in PAO sequential polymer usage, staged maintenance-decay, and glycolysis-TCA pathway shifts. The iEBPR model was first calibrated against bulk batch testing experiment data, being proved to perform better than the previous EBPR model for predicting the soluble orthoP, ammonia, biomass glycogen, and PHA temporal profiles in a batch starvation testing under prolonged anaerobic conditions. We further validated the model with another independent set of batch anaerobic batch testing data that included high-resolution single-cell and specific population level intracellular polymer measurements enabled by the single-cell Raman micro-spectroscopy technique. The model accurately predicted the temporal changes in the intracellular polymers at cellular and population levels within PAOs and GAOs, and further confirmed the proposed mechanism of sequential polymer utilization, and polymer availability-dependent and staged maintenance-decay in PAOs. These results indicate that under extended anaerobic phases as in S2EBPR, the PAOs may gain competitive advantages over GAOs due to the possession of multiple intracellular polymers and the adaptive switching of the anaerobic metabolic pathways that consequently lead to the later and slower decay in PAOs than GAOs. The iEBPR model can be applied to facilitate and optimize the design and operations of S2EBPR for more reliable nutrient removal and recovery from wastewater.