The effect of future outdoor air pollution on human health and the contribution of climate change

At present, exposure to outdoor air pollution from ozone and fine particulate matter (PM2.5) causes over 2 million deaths per year, due to respiratory and cardiovascular diseases and lung cancer. Future ambient concentrations of ozone and PM2.5 will be affected by both air pollutant emissions and climate change. Here we estimate the potential impact of future outdoor air pollution on premature human mortality, and isolate the contribution of future climate change due to its effect on air quality. We use modeled present-day (2000) and future global ozone and PM2.5 concentrations from simulations with an ensemble of chemistry-climate models from the Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP). Future air pollution was modeled for global greenhouse gas and air pollutant emissions in the four IPCC AR5 Representative Concentration Pathway (RCP) scenarios, for 2030, 2050 and 2100. All model outputs are regridded to a common 0.5°x0.5° horizontal resolution. Future premature mortality is estimated for each RCP scenario and year based on changes in concentrations of ozone and PM2.5 relative to 2000. Using a health impact function, changes in concentrations for each RCP scenario are combined with future population and cause-specific baseline mortality rates as projected by a single independent scenario in which the global incidence of cardiopulmonary diseases is expected to increase. The effect of climate change is isolated by considering the difference between air pollutant concentrations from simulations with 2000 emissions and a future year climate and simulations with 2000 emissions and climate. Uncertainties in the results reflect the uncertainty in the concentration-response function and that associated with variability among models. Few previous studies have quantified the effects of future climate change on global human health via changes in air quality, and this is the first such study to use an ensemble of global models.