Predicting Biological Phosphorous Removal Performance through Measuring Intracellular Carbon Storage Levels

INTRODUCTION Over the last decade, Clean Water Services (CWS) has been studying biological phosphorus removal (BPR) at full-scale to gain insight into potential process control parameters to inform process stability and predict upsets. In the search for stability indicators CWS focused on residual phosphorus uptake (RPU) testing [2], online stability analyzers [3], functional gene expression [4], and measuring carbon storage compounds – specifically PHA (polyhydroxyalkanoates). PHA is known to drive phosphorus uptake [5]. CWS therefore hypothesized that RPU and intracellular PHA content are directly correlated. While RPU is a clear indicator of BPR health, the procedure is labor intensive and only performed weekly. If PHA is a meaningful indicator of BPR health, its ease of measure could provide daily information on BPR health. Initial investigations indicated a strong Monod-like relationship between phosphorus uptake rate and biomass PHA content [2]. However, subsequent full-scale testing pairing only RPU with PHA present in the last aerobic zone (same source of biomass for the RPU test) revealed a poor relationship when evaluated over a long period of time [6]. CWS examined multiple PHA profiles collected on several basins throughout the 2024 dry weather phosphorus permit season. The results demonstrated that, while the expected Monod-type relationship is consistently observed on a given day, the relationship is highly variable both in time and across basins [Figure 1] [7]. OBJECTIVES Given previous results, the following objectives were laid out for testing during the summer of 2025. •Objective 1: Determine if the PHA concentration at the end of the anaerobic zone acts as an improved indicator of BPR performance when correlated with RPU. •Objective 2: Determine if anaerobic PHA, when examined in the context of a wider range of BPR relevant parameters, can expand our understanding of BPR stability and health. •Objective 3: Measure the PAO and GAO relative abundance in the basins to examine if biomass abundance can explain PHA variability. METHODOLOGY PHA grab samples were taken weekly at the end anaerobic zone for eight basins across two CWS facilities, Durham and Rock Creek. PHA samples were measured in-house [8]. RPU was measured weekly on biomass sampled from the last aerobic zone. DNA samples were shipped to the University of Idaho for quantitative polymerase chain reaction (qPCR) testing and subsequent GAO and PAO quantification (16S rDNA and ppk for PAOs). Several relevant predictor variables were evaluated for correlation with anaerobic zone PHA (see Figures 5 and 6 for a complete list). Correlation statistics with PHA were interpreted with Pearson or Spearman tests with alpha values of 0.05. FINDINGS / STATUS Objective 1: Using anaerobic PHA as a replacement indicator for RPU had mixed results [Figure 2,3]. While Durham basins 2, 4, and 5 showed a significant linear correlation, other basins – particularly those with periods of poor BPR stability – did not. Longer anaerobic detention time is known to improve BPR performance [9]. Further investigation found that the average anaerobic PHA concentration was at least 150% greater in Durham basin 5 than any other basin [Figure 4]. Related, the anaerobic detention time for Durham basin 5 is roughly twice that of any other basin [Figure 4]. Long term operation of Durham basin 5 has clearly demonstrated that an aeration basin designed with current best industry knowledge for BPR is more stable and experiences fewer process upsets [data not shown]. Objective 2: The correlation analysis [Figure 5] found that at Durham, basin 5 had a significant relationship between the end anaerobic zone PHA and over half of the potential predictor variables tested. The other Durham basins had a few significant correlations amongst them, but no universal variable. Looking at the Rock Creek analysis [Figure 6], there were only seven instances of potential correlation – none of which were significant across all basins. Objective 3: Microbial measurements revealed higher GAO levels at Rock Creek than Durham [Figure 7,8], but the differences did not translate into predictable impacts on PHA or basin performance on initial review. PRELIMINARY CONCLUSIONS The relationship between end anaerobic PHA and RPU is better than that found previously for end aerobic PHA. However, the relationship is not universal, even across identical basins at Durham, and is non-existent for any Rock Creek basin. Moreover, end anaerobic PHA is not predictive of secondary effluent phosphorus removal performance. PHA also did not show any consistent trends across basins when searching for correlations with any other predictive parameters. A greater number of potential relationships were identified for Durham Basin 5, with the differentiating factor appearing to be linked to the higher anaerobic detention time. This finding of increased anaerobic detention time reaffirms the industry's understanding of appropriate basin design for improved BPR performance [9]. Based on these results, we do not believe that anaerobic PHA concentration is a reliable replacement for RPU as a stability indicator. It is also not a strong indicator of BPR health. This presentation will summarize CWS findings around the utility of PHA as a BPR health indicator. Further analysis of the relationship between PHA and PAO/GAO population abundance will also be presented.