The Influence of Anthropogenic Aerosol on Multi-Decadal Variations of Historical Global and Regional Climate

Aerosol presents a large source of uncertainty in climate simulations. The representation of aerosol, and their interaction with clouds, in climate models is changing rapidly. It is important to gauge model performance in simulating these processes, and which aspects of aerosol-climate interaction contribute to uncertainty, to enable effort to be prioritized. Future changes in aerosol entwine air-quality and climate change mitigation options. With rapid reductions in aerosol emissions anticipated in the near-term, it is important to constrain uncertainty to facilitate sound decisions on future policy. We show that climate models that include a representation of the indirect effect of aerosol better reproduce inter-decadal variability in historical global-mean near-surface temperatures, particularly the cooling in the 1950s and 1960s, compared to models with representation of the aerosol direct effect only. Using an adaptive decomposition technique to identify nonlinear trends, analysis of single forcing runs from CMIP5 simulations shows that the mid-twentieth century temperature hiatus is likely to have been influenced strongly by anthropogenic aerosol forcing. In addition to global impacts, aerosol can have a pronounced influence on local climate. Using case studies from regions with large responses to aerosol forcing, we investigate inter-model differences in aerosol burden, and in the sensitivity of atmospheric metrics to aerosol changes. We find a large range of sensitivities to aerosol perturbations, as well as considerable differences in the mass loading of some species on regional scales. This inter-model diversity in aerosol burden and representation of aerosol-cloud interaction can produce substantial variation in simulations of climate variability on multi-decadal timescales.