Response of climate to regional emissions of ozone precursors: sensitivities and warming potentials

Open Access
Publisher Website
Google Scholar
Abstract
The response of climate to ozone perturbations caused by regional emissions of NOx or CO has been studied through a sequence of model simulations. Changes in O3 and OH concentrations due to emission perturbations in Europe and southeast Asia have been calculated with two global 3-D chemical tracer models(CTMs; LMDzINCA and Oslo-CTM2). The radiative transfer codes of three general circulation models (GCMs; ECHAM4, UREAD and LMD) have been used to calculate the radiative forcing of the O3perturbations, and for a subset of the cases full GCM simulations have been performed with ECHAM4 and UREAD. The results have been aggregated to a global number in two ways: first, through integrating the global-mean radiative forcing of a sustained step change in emissions, and second through a modified concept (SGWP*) which includes possible differences in the climate sensitivity of O3, CH4 and CO2 changes. In terms of change in global tropospheric O3 burden the two CTMs differ by less than 30%. Both CTMs show a higher north/south gradient in the sensitivity to changes in NOx emission than for CO. We are not able to conclude whether real O3 perturbations in general have a different climate sensitivity from CO2. However, in both GCMs high-latitude emission perturbations lead to climate perturbations with higher (10–30%) climate sensitivities. The calculated SGWP*, for a 100 yr time horizon, are negative for three of the four CTM/GCM combinations for European emissions (−9.6 to +6.9), while for the Asian emissions the SGWP* (H= 100) is always positive (+2.9 to +25) indicating a warming. For CO the SGWP* values (3.8 and 4.4 for European and Asian emissions respectively, with only the Oslo-CTM2/ECHAM4 model combination) are less regionally dependent. Our results support the view that for NOx, regionally different weighting factors for the emissions are necessary. For CO the results are more robust and one global number may be acceptable. DOI: 10.1111/j.1600-0889.2005.00152.x