Identification of trigger factors selecting for polyphosphate- and glycogen-accumulating organisms in aerobic granular sludge sequencing batch reactors

Nutrient removal performances of sequencing batch reactors using granular sludge for intensified biological wastewater treatment rely on optimal underlying microbial selection. Trigger factors of bacterial selection and nutrient removal were investigated in these novel biofilm systems with specific emphasis on polyphosphate- (PAO) and glycogen-accumulating organisms (GAO) mainly affiliated with Accumulibacter and Competibacter, respectively. In a first dynamic reactor operated with stepwise changes in concentration and ratio of acetate and propionate (Ac/Pr) under anaerobic feeding and aerobic starvation conditions and without wasting sludge periodically, propionate favorably selected for Accumulibacter (35% relative abundance) and stable production of granular biomass. A Plackett-Burman multifactorial experimental design was then used to screen in eight runs of 50 days at stable sludge retention time of 15 days for the main effects of COD concentration, Ac/Pr ratio, COD/P ratio, pH, temperature, and redox conditions during starvation. At 95% confidence level, pH was mainly triggering direct Accumulibacter selection and nutrient removal. The overall PAO/GAO competition in granular sludge was statistically equally impacted by pH, temperature, and redox factors. High Accumulibacter abundances (30–47%), PAO/GAO ratios (2.8–8.4), and phosphorus removal (80–100%) were selected by slightly alkaline (pH > 7.3) and lower mesophilic (<20 °C) conditions, and under full aeration during fixed 2-h starvation. Nitrogen removal by nitrification and denitrification (84–97%) was positively correlated to pH and temperature. In addition to alkalinity, non-limited organic conditions, 3-carbon propionate substrate, sludge age control, and phase length adaptation under alternating aerobic-anoxic conditions during starvation can lead to efficient nutrient-removing granular sludge biofilm systems.