A41P-07
Anthropogenic Aerosol Dimming Over Oceans: A Regional Analysis
Abstract:
The role of anthropogenic aerosols in shaping 20th century SSTs through alteration of surface solar radiation (SSR) is still subject to debate.Identifying and quantifying the relationship between aerosol-induced changes in SSR and the corresponding SST response is difficult due to the masking effect of numerous feedback mechanisms and general variability of the atmosphere-ocean system. We therefore analysed potential anthropogenic aerosol effects on SST with a cascade of experiments of increasing complexity: From atmosphere-only over mixed-layer ocean (MLO) experiments, to fully coupled transient ocean-atmosphere simulations, with and without greenhouse gases and / or aerosols, using the general circulation model ECHAM with explicit aerosol representation.
We find anthropogenic aerosols to be crucial to obtain realistic SSR and SST patterns, although co-location of changes in individual variables (aerosol optical depth, SSR, SST) is weak. The effect of greenhouse gases and aerosols in the MLO simulations is essentially additive on global and regional scales, an assumption frequently made in the literature. With atmosphere-only simulations we identified regions most prone to anthropogenic aerosol dimming throughout the 20th century using a strict criterion. From MLO equilibria representative of different decades throughout the 20th century, we identified ocean regions, whose SSTs are most sensitive to changing anthropogenic aerosol emissions. The surface temperature response patterns in our MLO simulations are more sensitive towards the choice of prescribed deep-ocean heat flux if anthropogenic aerosols were included as compared to greenhouse gas only simulations. This implies that ocean dynamics might mask some of the response and cautions against the use of just one set of deep-ocean heat fluxes in MLO studies.
Our results corroborate not only the relevance of anthropogenic aerosols for SST responses, but also highlight the complexity and non-locality of the processes involved.