The climate impact of ship NOx emissions: uncertainties due to plume chemistry

Ships are an important source of reactive trace gases in the marine atmosphere, comprising about 17% of total anthropogenic NOx emissions. In the marine environment, ship NOx emissions generate ozone (O₃) and hydroxyl radicals (OH) more efficiently than in continental air. Global atmospheric chemistry and transport models (CTMs) have found that ship NOx cools Earth's climate because reductions in methane radiative forcing (RF) due to the OH enhancements more than compensate for warming caused by ship-induced ozone. These past model studies, however, all assumed that the concentrated plumes of ship exhaust are instantly diluted into a grid cell spanning hundreds of kilometers. This expedient but inaccurate model assumption overestimates ozone and OH production, because the affected models bypass the early stages of plume evolution when high NOx concentrations intensify NOx chemical losses. We provide here the first estimate of RF from ship NOx that accounts for sub-grid-scale ship plume chemistry. First, we improve the plume-in-grid representation of exhaust gas chemistry, which is derived from a plume dispersion model, in the GEOS-Chem global CTM. The CTM now calculates methane oxidation within exhaust plumes for the first time, where OH concentrations are 2-3 times greater than background air. We also account for the effect of wind speed on ozone production and losses of NOx and methane in young plumes. We evaluate the CTM against airborne measurements of NOx and ozone over the ocean. The global ship-induced perturbations to ozone and methane concentrations in the improved model are smaller than suggested by the ensemble of past global modeling studies. If we assume instant dilution of ship NOx emissions in our CTM, we can reproduce the past model results, but ozone production is overestimated by 20% and the resulting ozone column enhancements and RF by 40%. Thus, the ozone and methane RF components from ship NOx are likely much smaller than suggested by past literature. Our best estimate of the total RF from ship NOx is -19 mW m^-2 for year 2000 emissions. We provide an error analysis of our results and compare the NOx effects against other climate effects of ships.