Unraveling operational drivers of nitrous oxide emissions in biological wastewater treatment systems through machine learning analysis of multi-decadal datasets

This study focused on the development of machine-learning- (ML-) based strategies for mitigating nitrous oxide (N 2 O) emissions from various wastewater treatment systems in the United States measured using a benchmark USEPA-endorsed protocol. Results revealed that in general, poor process performance correlated with higher N 2 O emissions. Specifically, local variables including zone-specific dissolved oxygen, ammonia, and nitrite concentrations and global variables including effluent nitrite and nitrate concentrations contributed positively towards N 2 O emissions from both aerobic and anoxic zones of the process bioreactors. The optimal operational conditions identified for minimizing N2O emissions included, operation of aerobic and anoxic zones at DO > 4 mg O 2 /L and < 1 mg O 2 /L, respectively, coupled with appropriate solids retention times (SRTs) that maximize process performance. Accordingly, our results strongly underscore the utility of ML models in combination with bioprocess fundamentals for predicting and mitigating N 2 O emissions, while concomitantly optimizing wastewater treatment operations.