Simulating the Spatiotemporal Variations of δ15N in Atmospheric NOx Using CMAQ to Assess the Role of Atmospheric Physical Processes over the Pearl River Delta Region, China

Reactive nitrogen oxides (NOx = NO + NO2) are critical drivers of atmospheric chemistry, with far-reaching consequences for air quality, climate, and ecosystem health. Stable nitrogen isotopes provide a useful constraint for NOx source attribution, but source-derived isotopic signals can be substantially modified by atmospheric processes after emission. In this study, we used a previously developed 15N-enabled NOx emission dataset based on the 2008 Multi-resolution Emission Inventory for China (MEIC) as input to the Community Multiscale Air Quality modeling system (CMAQ) to simulate the spatiotemporal variation in δ15N(NOx) over South China, with a focus on the Pearl River Delta region. In the simulations, 14NOx and 15NOx were implemented as nonreactive tracers under three scenarios to isolate the effects of atmospheric physical processes, without explicitly simulating NOx oxidation chemistry, isotope fractionation, or nitrate formation. Under the “emission + transport + default deposition”, simulated δ15N(NOx) values generally ranged from approximately −2‰ to +6‰, with most areas showing values between +2‰ and +4‰. Compared with the corresponding “emission-only” baseline, atmospheric transport, mixing, and deposition tended to increase δ15N(NOx), especially in rural and low-emission regions, by redistributing anthropogenic NOx signals. Default deposition had only a minor effect on δ15N(NOx), whereas enhanced deposition, used to approximate effective near-source removal, produced spatially heterogeneous but mostly modest changes. Simulated δ15N(NOx) values were both systematically lower than measured δ15N(NO3−) at the rural Dinghushan and urban Guangzhou sites under all scenarios. This discrepancy indicates that atmospheric physical processes alone cannot fully explain observed nitrate isotope variability, and that isotope fractionation during NOx oxidation, source-endmember uncertainty, and emission inventory biases likely need to be considered. This study serves as a proof of concept and a necessary step toward fully 15N-enabled CMAQ simulations for evaluating atmospheric reactive nitrogen sources and improving NOx emission inventories.