Variability and Uncertainty Analysis of N2O Emissions from WWTP to Improve the Accuracy of Emission Factors and the Design of Monitoring Strategies

Investigations on global nitrous oxide (N2O) emissions from wastewater treatment plants (WWTPs) often overlook data variability and uncertainty by utilizing averaged emission factors (EF and N2O emitted/total nitrogen load). The variability of the EF is system-specific: technology, operation, influent characteristics, and microbial community. The uncertainty of the EF is associated with the monitoring campaign: number of measuring locations, analytical methods, data frequency, and duration. This study quantified the spatial and temporal variability of N2O emissions and EF from liquid measurements of N2O in a surface-aerated WWTP over a 110-day campaign. The gas stripping model (kLaN2O) yielded a minimum measurement error of 10.4% and 13.0% for the TNload. Daily patterns of N2O concentrations and emissions were not normally distributed but positively skewed, suggesting that log-distributions are more accurate descriptors (11.6% of the days emitted 50% of emissions). The EF was 0.30 +/- 1.29% (n = 117602), and the higher standard deviation of the mean highlights the inaccuracy of normal distributions for N2O data sets, whereas the median and 2.5-97.5th quantiles and/or continuous log-distributions are more informative. An in silico analysis indicates that monitoring 5 out of 30 min instead of continuously provides reasonable accuracy and precision while allowing for six different monitoring locations.