Uncertainty Analysis of Ozone Attributable to Aviation NOx Emissions

Reported tropospheric ozone burden changes due to aviation NOx range from 2.5 to 16 TgO3/TgN across different chemical transport models, but the parameters contributing most to this uncertainty have not been identified. We quantify changes in aviation-attributable ozone by varying a set of model parameters using the Community Earth System Model version 2 (CESM2), including with a newly developed configuration (CESM-GC), which implements the GEOS-Chem chemistry modules (chemical mechanism, photolysis, emission, and deposition). Our investigation focuses on how nudging strength, lightning NOx intensity, vertical resolution, chemistry mechanisms, and photolysis schemes affect the ozone burden sensitivity to aviation NOx. We test different nudging strengths (12 hr, 30 hr, 50 hr, and a vertically-variable nudging strength), lightning NOx intensities (2.7, 5.3, 8.0 TgN/year), with vertical resolutions (32L, 70L, and 110L). We exploit the modularity of CESM-GC which enables us to test sensitivities to using the FAST-JX versus TUV photolysis schemes and the GEOS-Chem versus MOZART-TS1 chemical mechanism. For each configuration, we conduct simulations with and without aviation NOx. The base case (MOZART-TS1, 32L, TUV, 12 hr, 5.3 TgN/year) shows an annual average ozone burden change due to aviation NOx of 9.4 TgO3/TgN, with seasonal variation from 8.2 to 10.7 TgO3/TgN. Across 10 configurations, the annual average ozone sensitivity ranges from 7.9 to 11.7 TgO3/TgN, and considering seasonal variations, from 1.5 to 24.9 TgO3/TgN. We also examine each parameter's impact. Changes in lightning NOx intensity alter background NOx concentration, with higher NOx concentration decreasing ozone sensitivity due to aviation NOx. We find seasonally dependent responses to changes in both the chemistry scheme and the nudging strength. Changes in the photolysis scheme affect J(NO2) and J(O3) rates, with J(O3) being 40 to 50% higher in TUV compared to FAST-JX, and ozone sensitivity to aviation NOx being 9% higher with FAST-JX than with TUV. We infer that the range of the sensitivity of the tropospheric ozone burden to aviation NOx reported in the literature likely reflects inter-model differences in the particular model configuration, pointing to the need for modular approaches such as now possible in CESM2.