Navigating Carbon Limits: Lessons Learned from Small and Full-scale S2EBPR Operation in a C-limited Facility

Introduction Water reclamation facilities use enhanced biological phosphorus removal (EBPR) to prevent the release of phosphorus into aquatic environments. Phosphorus-accumulating organisms (PAO) drive EBPR by using rbCOD to perform phosphorus cycling. However, PAO may compete with other desirable heterotrophs and glycogen-accumulating organisms (GAO). Sidestream EBPR (S2EBPR) is designed to reduce upsets from low or unreliable rbCOD in influent wastewater. In this study, Calumet WRP conducted a small-scale and full-scale study of S2EBPR to achieve effluent P goals below 1 mgP/L. This study evaluated the performance of S2EBPR with and without external carbon dosing, compared small-scale findings to full-scale performance, and assessed the PAO and GAO community over time. Methods Small- and full-scale reactors (SSR and FSR) were operated at Calumet WRP for 1 year, in 2019 and 2021, respectively (Figure 1a and 1b; also see Sabba et al. 2023 and Farmer et al. 2023 for additional details about the FSR). The SSR was supplied with external C addition in 3 phases: 50, 25 and 0% C. The FSR was also operated with changing C dosing, 2 periods with C and 2 periods without. Composite samples of reactor effluent were collected on a daily and/or weekly basis for TP, OP NH3−N, NOx, TKN, BOD5, total and volatile suspended solids and VFA and measured with EPA and APHA standard methods. 16S rRNA gene sequencing was performed on samples collected from SSR and FSR, and metagenomic sequencing was performed on select samples from the FSR. Results SSR S2EBPR findings Batch testing was to measure P uptake on nitrate and oxygen in the SSR (Figure 2a). Oxygen-driven P uptake appeared to be the main contributor to P removal. However, the weekly average P removal observed in the reactor was significantly correlated with nitrate-driven rate, not the oxygen-driven rate, suggesting that nitrate-driven P uptake was important for overall performance (Figure 2b). Overall, these results suggest a higher proportion of PAO can denitrify than previously reported (Nielsen et al., 2019). Figure 3a and 3c shows the relative abundance of PAO and GAO population in the SSR during the three C phases. Despite the C reduction, overall, the S2EBPR shows stability of PAO population, including the canonical PAO Accumulibacter, with only a slight decrease in relative abundance of Tetrasphaera. Interestingly the decrease of C to 0% did not lead to exclusion of GAO but rather, a stable population of Ca. Competibacter and Micropruina was established. FSR S2EBPR findings Figure 3b displays the relative abundance of PAO and GAO in the FSR during the four C phases. Both PAO and GAO populations increased and established during the two phases where C was dosed. Notably, Ca. Competibacter increased during phase III (C on) and persisted during phase IV (C off). To further investigate the trends observed in the FSR with regards to PAO and GAO and connect them with denitrification-based P uptake observed in Figure 2, we next analyzed the denitrification pathway using metagenomic sequencing data with a focus on denitrifying PAO and GAO (known as DPAO and DGAO). We mapped our metagenomic reads to curated denitrifying genes from Ca. Accumulibacter, Dechloromonas, Tetraphaera, and Ca. Competibacter. We found that the complete denitrification pathway was encoded by DPAO and DGAO, but there were considerable differences in the gene abundances amongst the organisms. For example, most nitrate reductases mapped to narG from Ca. Accumulibacter, but the nitric oxide reductases mapped to nor genes from Ca. Competibacter, suggesting that the DPAO and DGAO may have been partial denitrifiers. Our metagenomic findings also point to the possibility of N2O accumulation due to incomplete denitrification, though further analysis of the denitrification pathway completeness considering other denitrifying bacteria is needed to increase our confidence. Comparison of SSR and FSR performance External C dosing had a substantial impact on P removal performance in both SSR and FSR. In the SSR, the median OP removal was 90% with the 50% C dose, compared to 83% with the 25% C dose and 50% without dosing. In the FSR, both the 50% and 25% C dosing conditions achieved less than 1mgP/L with a median effluent 0.49 and 0.40 mgP/L, respectively. The FSR exhibited similar trends; the median OP removal was 90% with C dosing (average of 3.3 L/min) compared to a median of 15% OP removal, as well as occasional net OP release, without C dosing. The median effluent OP from the FSR was 0.29 mgP/L under the high C dosing condition, also achieving the less than 1mgP/L goal. Overall, the performance was comparable between the SSR and the FSR, despite the differences in C dosing regimes and reactor configurations. Benefits and significance This presentation will highlight the successful scale-up of an S2EBPR process from a small lab-scale system to a full-scale demonstration, as well as the potential role of DPAO in combined P and N removal. For practitioners, this work will provide insights into optimizing EBPR systems at low carbon facilities, understanding how DPAO can contribute to P and N removal, and operating relevant pilot studies to inform full-scale work. For researchers, our findings on denitrification gene abundances are relevant for further study of truncated denitrification pathways in DPAO and DGAO.