Toward effective emissions of ships in global models

Abstract

The dispersion and chemical conversion of emissions in the near-field of a single ship are studied with two different modelling approaches to explore the differences between gradual dispersion and instantaneous dilution into a box with a size comparable to the large grid boxes of global scale models or satellite data. While both techniques use the same photochemical box model to solve the chemical equations, the dilution of the exhaust into the background air is different. One approach uses a Gaussian plume model and accounts for the expansion phase of a plume. The other one instantaneously disperses the emissions over a large gridbox, a technique commonly used by large scale models. In a first step we show that differences in the time evolution of ozone between the two model approaches are large for the case studied here, where emissions from a large container ship are released into the marine boundary layer with neutral stability. For emissions at noon, the differences in the ship induced ozone change at the reference time when both boxes are of equal size chosen to be 60 km are largest. The ozone change is then overestimated by the global-model approach by a factor of three. This results from the neglect of sub-grid scale OH loss due to NO2 oxidation in the global-model approach, which inhibits hydrocarbon oxidation and thus ozone production. Smallest differences are encountered for emissions released around sunset. One possibility to account for these sub-grid processes in global-models is the use of effective emissions, i.e. actual emissions are changed and emissions of additional compounds like ozone are introduced in a way that they take sub-grid processes into account. In a second step we present effective emissions for the particular case discussed above. It is shown for this case that the method is able to account for the neglect of sub-grid processes in global models for different emission times and emission strengths.