Abstract / Description of output
Two coupled climate–chemistry model experiments for the period 1990–2030 were conducted: one with a fixed climate and the other with a varying climate forced by the is92a scenario. By comparing results from these experiments we have attempted to identify changes and variations in physical climate that may have important influences upon tropospheric chemical composition. Climate variables considered include: temperature, humidity, convective mass fluxes, precipitation, and the large-scale circulation. Increases in humidity, directly related to increases in temperature, exert a major influence on the budgets of ozone and the hydroxyl radical: decreasing O3 and increasing OH. Warming enhances decomposition of PAN, releasing NOx, and increases the rate of methane oxidation. Surface warming enhances vegetation emissions of isoprene, an important ozone precursor. In the changed climate, tropical convection generally reduces, but penetrates to higher levels. Over northern continents, convection tends to increase. These changes in convection affect both vertical mixing and lightning NOx emissions. We find no global trend in lightning emissions, but significant changes in its distribution. Changes in precipitation and the large-scale circulation are less important for composition, at least in these experiments. Higher levels of the oxidants OH and H2O2 lead to increases in aerosol formation and concentrations. These results indicate that climate–chemistry feedbacks are dominantly negative (less O3, a shorter CH4 lifetime, and more aerosol). The major mode of inter-annual variability in the is92a climate experiment is ENSO. This strongly modulates isoprene emissions from vegetation via tropical land surface temperatures. ENSO is also clearly the dominant source of variability in tropical column ozone, mainly through changes in the distribution of convection. The magnitude of inter-annual variability in ozone is comparable to the changes brought about by emissions and climate changes between the 1990s and 2020s, suggesting that it will be difficult to disentangle the different components of near-future changes.