An assessment of the performance of the Monitor for AeRosols and GAses in ambient air (MARGA): a semi-continuous method for soluble compounds

Abstract

Ambient air monitoring as part of the US Environmental Protection Agency's (US EPA's) Clean Air Status and Trends Network (CASTNet) currently uses filter packs to measure weekly integrated concentrations. The US EPA is interested in supplementing CASTNet with semi-continuous monitoring systems at select sites to characterize atmospheric chemistry and deposition of nitrogen and sulfur compounds at higher time resolution than the filter pack. The Monitor for AeRosols and GAses in ambient air (MARGA) measures water-soluble gases and aerosols at an hourly temporal resolution. The performance of the MARGA was assessed under the US EPA Environmental Technology Verification (ETV) program. The assessment was conducted in Research Triangle Park, North Carolina, from 8 September to 8 October 2010 and focused on gaseous SO2, HNO3, and NH3 and aerosol SO42-, NO3-, and NH4+. Precision of the MARGA was evaluated by calculating the median absolute relative percent difference (MARPD) between paired hourly results from duplicate MARGA units (MUs), with a performance goal of ≤ 25%. The accuracy of the MARGA was evaluated by calculating the MARPD for each MU relative to the average of the duplicate denuder/filter pack concentrations, with a performance goal of ≤ 40%. Accuracy was also evaluated by using linear regression, where MU concentrations were plotted against the average of the duplicate denuder/filter pack concentrations. From this, a linear least squares line of best fit was applied. The goal was for the slope of the line of best fit to be between 0.8 and 1.2. The MARGA performed well in comparison to the denuder/filter pack for SO2, SO42−, and NH4+, with all three compounds passing the accuracy and precision goals by a significant margin. The performance of the MARGA in measuring NO3- could not be evaluated due to the different sampling efficiency of coarse NO3- by the MUs and the filter pack. Estimates of "fine" NO3- were calculated for the MUs and the filter pack. Using this and results from a previous study, it is concluded that if the MUs and the filter pack were sampling the same particle size, the MUs would have good agreement in terms of precision and accuracy. The MARGA performed moderately well in measuring HNO3 and NH3, though neither met the linear regression slope goals. However, recommendations for improving the measurement of HNO3 and NH3 are discussed. It is concluded that SO42-, SO2, NO3-, HNO3, NH4+, and NH3 concentrations can be measured with acceptable accuracy and precision when the MARGA is operated in conjunction with the recommendations outlined in the manuscript.