Assessment and quantification of NOx sources at a regional background site in North China: Comparative results from a Bayesian isotopic mixing model and a positive matrix factorization model.

Regional sources of nitrogen oxides (NOx) in North China during summer were explored using both a Bayesian isotopic mixing model and a positive matrix factorization (PMF) model. Results showed that the nitrogen isotope (δ15N) composition of particulate nitrate (NO3-) varied between -8.9‰ and +14.1‰, while the oxygen isotope (δ18O) composition ranged from +57.4‰ to +93.8‰. Based on results from the Bayesian isotopic mixing model, the contribution of the hydroxyl radical (•OH) NOx conversion pathway showed clear diurnal fluctuation; values were higher during the day (0.53 ± 0.16) and lower overnight (0.42 ± 0.17). Values peaked at 06:00-12:00 and then decreased gradually until 00:00-06:00 the next day. Coal combustion (31.34 ± 9.04%) was the most significant source of NOx followed by biomass burning (25.74 ± 2.58%), mobile sources (23.83 ± 3.66%), and microbial processes (19.09 ± 5.21%). PMF results indicated that the contribution from mobile sources was 19.83%, slightly lower as compared to the Bayesian model (23.83%). The PMF model also reported a lower contribution from coal combustion (28.65%) as compared to the Bayesian model (31.34%); however, the sum of biomass burning and microbial processes in the Bayesian model (44.83%) was lower than the aggregate of secondary inorganic aerosol, sea salt, and soil dust in PMF model (51.52%). Overall, differences between the two models were minor, suggesting that this study provided a reasonable source quantification for NOx in North China during summer.