Predicting anaerobic membrane bioreactor performance using flow cytometry-derived numbers of high and low nucleic acid content. undefined

The performance of an anaerobic membrane bioreactor (AnMBR) is intricately related to its microbial community. Having a tool to monitor the microbial abundances in a rapid manner and to utilize the data to predict COD removal efficiency would facilitate the operation of AnMBR. This study aims to achieve this by operating a 35L pilot-scale AnMBR at varying hydraulic retention times (e.g., 24h, 20h, 16h and 12h). The AnMBR exhibited lower variability in COD removal and methane production when HRT decreased from 24h to 12h. Improved performance was positively correlated with the high and low nucleic acid (HNA and LNA) cell numbers but negatively correlated with the HNA/LNA ratio. This suggests that LNA cells, which were sorted and identified to include hydrolytic genera like Alcaligenes and Anaerovibrio, play important roles in COD removal. Stepwise robust linear regression model indicates that HNA/LNA ratio and LNA bacteria number could predict the COD removal efficiency with a coefficient of determination of 0.65. Further testing of the model by using data obtained from another independently operated reactor showed no significant difference between predicted values and the actual COD removal efficiencies. Overall, our findings suggest that flow cytometry can be used to monitor for HNA and LNA cell numbers, which in turn can be fitted into the developed regression model for constant monitoring of the AnMBR.