Sensitivity of projected PM2.5- and O3-related health impacts to model inputs: A case study in mainland China

Abstract

In China, fine particulate matter (PM_2.5) and ground-level ozone (O₃) are anticipated to continuously affect large populations in the coming decades. Simulations of the levels of these pollutants largely depend on emissions inputs, which are highly uncertain both in magnitude and spatial distribution. Our goal was to explore sensitivities of projected changes in PM_2.5- and O₃-related short-term health impacts in mainland China to emissions and other model inputs. We simulated winter PM_2.5 and summer O₃ concentrations using the Weather Research and Forecasting model coupled with Chemistry (WRF-Chem) for both 2008 and 2050. We used three emission inventories in 2008 and four emissions scenarios in 2050. The resulting air pollutant concentrations were combined with eight population projections and three concentration-response functions (CRFs) to estimate future PM_2.5- and O₃-related health impacts including total, cardiovascular, and respiratory mortalities in mainland China. Multivariate analysis of variance was used to apportion the uncertainty due to different model parameters. Combinations of different parameters produced a wide range of national PM_2.5- and O₃-related mortalities. CRFs and present emissions each contribute 38%-56% and 20%-28% of the total sum of squares for PM_2.5-related mortalities. Future emissions are the largest source of uncertainty in O₃-related mortality estimates, contributing 24%-48% of total sum of squares. Our results suggest that conducting more epidemiological studies and constraining the present day emissions are essential for projecting future air pollutant-related health impacts in mainland China.

Keywords: China; Emissions; Mortality; Ozone; PM(2.5); Sensitivity analysis.