Sensitivity of Simulations of Global O3 and its Radiative Forcing to Selected Representations of Emissions, Meteorology, and Radiation Transfer

Abstract

Recently, Acquah et al. (ACP, 2025) reported on the sensitivity of simulated global O3 to employed emission inventories in the atmospheric chemistry general circulation model EMAC. Somewhat complementarily, we quantify here how much the simulated O3 and the associated effect on the Earth's radiation balance can be affected by specific (other) choices during design of the model setup. Regional changes of up to several percent in annual mean near-surface O3 mixing ratio occurred, for instance, when we exchanged the soil NOx emissions parameterization. Another, maybe less obvious reason for differences can be the choice of nudged vs. free-running meteorology, due to the effect on thunderstorms and, consequently, on lightning NOx generation. Leveraging the TAGGING submodel for process-resolved source attribution, we found changes induced by the different meteorology setups to be of the same order of magnitude. We demonstrate that such setup changes can reinforce each other, so that regional differences of several 10% in annual mean surface O3 mixing ratio can occur. Furthermore, we found similarly large differences in the radiative forcing when we used two different schemes for radiation transfer. We conclude that it is important to keep such effects in mind when interpreting model output, and that simulation setup details must be carefully documented for meaningful comparisons - even between simulations with the same model.