Comparison of liquid phase and remote sensing measurements of nitrous oxide emission from biological nitrogen removal at a wastewater treatment facility

Abstract

Direct emission of the greenhouse gases methane and nitrous oxide (N2O) constitute a significant fraction of the overall carbon footprint of wastewater treatment. Measurement methods to identify emission sources and to quantify emissions are key in mitigating these direct emissions. Nitrous oxide is formed in biological nitrogen removal process units, which are the main source of N2O emission from wastewater treatment.Liquid phase sensors (LPS) have recently been developed and installed at various Danish wastewater treatment plants to measure N2O concentrations in the liquid phase of biological nitrogen removal tanks. These sensors can be used to implement adjustments on the operation of the plant (for example duration of aeration), which affects N2O emission. In addition, LPS can be utilized to calculate N2O emission through mass transfer modelling. However, there is a need for validation of liquid-based modelled emission rates against measurement methods, which measure direct N2O emission rates. In this study, emission rates determined by two remote sensing methods, the tracer gas dispersion method (TDM) and Eddy covariance method (EC) were compared to LPS derived N2O emission rates.TDM relies on continuous, controlled release of a gaseous tracer at the source combined with downwind measurements of concentration of target gas (N2O here) and tracer gas (often acetylene - C2H2).  This method is well-established, validated, and has been used to quantify fugitive emissions from various sources such as landfills, composting plants, biogas plants, etc. EC is a stationary method, which relies on high-frequency measurements of N2O concentration and wind vector on a tower near the source. EC can be set up for continuous monitoring, while TDM as applied here is a discrete measurement method.In the study, N2O emission rates were measured over a period of 1.5 years at a relatively large wastewater treatment plant in the greater Copenhagen area. TDM measurements were conducted on 15 measurement days covering both periods of relatively high and low N2O emission rates. TDM measurements were compared to LPS derived emission rates, where N2O emission was measured using sensors in four of eight process units for biological nitrogen removal. Overall, daily average emission rates between approximately 0.38 and 13.4 kg N2O h-1 were measured. High emission rates of 120 kg N2O h-1 were observed on a day, where plant maintenance is believed to be the cause of unusual high emission. Emission rates from simultaneous TDM measurements and LPS derived values (n=43) showed good correlation (R2=0.70). On average, emission rates from TDM were 35% higher than LPS rates. The model implementation to derive LPS determined emission rates was further developed during the study, and the listed results were the final values after some correction. Several factors can explain the difference – including liquid sensor drift, which for the specific sensors tends towards lower N2O concentration readings than actual concentrations. Continuous EC measurements showed the same emission dynamics as measured by the liquid sensors located inside the footprint of the station.