PNNL

Abstract

BACKGROUND: The spatial pattern of the uncertainty in climate air pollution health impact has rarely been studied due to the lack of high-resolution model simulations, especially under the latest Representative Concentration Pathways (RCPs). OBJECTIVES: We estimated county-level ozone (O3) and PM2.5 related excess mortality (EM) and evaluated the associated uncertainties in the continental United States in the 2050s under RCP4.5 and RCP8.5. METHODS: Using dynamically downscaled climate model simulations, we calculated changes in O3 and PM2.5 levels at 12 km resolution between the future (2057-2059) and present (2001-2004) under two RCP scenarios. Using concentration-response relationships in the literature and projected future populations, we estimated EM attributable to the changes in O3 and PM2.5. We finally analyzed the contribution of input variables to the uncertainty in the county-level EM estimation using Monte Carlo simulation. RESULTS: O3-related premature deaths in the continental U.S. were estimated to be 1,082 deaths/year under RCP8.5 (95% confidence interval (CI): -288 to 2,453), and -5,229 deaths/year under RCP4.5 (-7,212 to -3,246). Simulated PM2.5 changes resulted in a significant decrease in EM under the two RCPs. The uncertainty of O3-related EM estimates was mainly caused by RCP scenarios, whereas that of PM2.5-related EMs was mainly from concentration-response functions. CONCLUSION: EM estimates attributable to climate change-induced air pollution change as well as the associated uncertainties vary substantially in space, and so are the most influential input variables. Spatially resolved data is crucial to develop effective mitigation and adaptation policy.