The Arctic region is warming considerably faster than the rest of the globe1, with important consequences for the ecosystems2 and human exploration of the region3. However, the reasons behind this Arctic amplification are not entirely clear4. As a result of measures to enhance air quality, anthropogenic emissions of particulate matter and its precursors have drastically decreased in parts of the Northern Hemisphere over the past three decades5. Here we present simulations with an Earth system model with comprehensive aerosol physics and chemistry that show that the sulfate aerosol reductions in Europe since 1980 can potentially explain a significant fraction of Arctic warming over that period. Specifically, the Arctic region receives an additional 0.3 W m−2 of energy, and warms by 0.5 °C on annual average in simulations with declining European sulfur emissions in line with historical observations, compared with a model simulation with fixed European emissions at 1980 levels. Arctic warming is amplified mainly in fall and winter, but the warming is initiated in summer by an increase in incoming solar radiation as well as an enhanced poleward oceanic and atmospheric heat transport. The simulated summertime energy surplus reduces sea-ice cover, which leads to a transfer of heat from the Arctic Ocean to the atmosphere. We conclude that air quality regulations in the Northern Hemisphere, the ocean and atmospheric circulation, and Arctic climate are inherently linked.
A. Asmi is acknowledged for help with the observational data. This work benefited from discussions with R. G. Graversen, A. Lewinschal, G. Messori, M. Salter, J. Nilsson and F. Pausata. The research leading to these results has received funding from the Nordic Centres of Excellence CRAICC and eSTICC, Swedish Environmental Protection Agency projects SCAC and CLEO, Norwegian Research Council projects EVA (grant no. 229771) and NOTUR (nn2345k), European FP7 Integrated projects PEGASOS (no. 265148) and ACCESS, and European Research Council Grant ATMOGAIN (no. 278277). The Swedish National Supercomputing Centre and NordStore (project ns2345k) are acknowledged for computational resources for running the simulations.