Variation in estimated ozone-related health impacts of climate change due to modeling choices and assumptions

Abstract

Background: Future climate change may cause air quality degradation via climate-induced changes in meteorology, atmospheric chemistry, and emissions into the air. Few studies have explicitly modeled the potential relationships between climate change, air quality, and human health, and fewer still have investigated the sensitivity of estimates to the underlying modeling choices.

Objectives: Our goal was to assess the sensitivity of estimated ozone-related human health impacts of climate change to key modeling choices.

Methods: Our analysis included seven modeling systems in which a climate change model is linked to an air quality model, five population projections, and multiple concentration-response functions. Using the U.S. Environmental Protection Agency's (EPA's) Environmental Benefits Mapping and Analysis Program (BenMAP), we estimated future ozone (O(3))-related health effects in the United States attributable to simulated climate change between the years 2000 and approximately 2050, given each combination of modeling choices. Health effects and concentration-response functions were chosen to match those used in the U.S. EPA's 2008 Regulatory Impact Analysis of the National Ambient Air Quality Standards for O(3).

Results: Different combinations of methodological choices produced a range of estimates of national O(3)-related mortality from roughly 600 deaths avoided as a result of climate change to 2,500 deaths attributable to climate change (although the large majority produced increases in mortality). The choice of the climate change and the air quality model reflected the greatest source of uncertainty, with the other modeling choices having lesser but still substantial effects.

Conclusions: Our results highlight the need to use an ensemble approach, instead of relying on any one set of modeling choices, to assess the potential risks associated with O(3)-related human health effects resulting from climate change.

Figure 1

The structure of the analysis of O₃-related impacts on human health attributable to climate change. GHG, greenhouse gas.

Figure 2

Estimated national summertime (June–August) O₃-related nonaccidental mortality due to simulated climate change between 2000 and ca. 2050 (C–R function from Bell et al. 2004). We estimated that –0.6 deaths were based on the WSU climate change model–air quality model and Census 2000 population data.

Figure 3

Estimated national and regional summertime (June–August) O₃-related nonaccidental mortality due to simulated climate change between 2000 and ca. 2050 (C–R function from Bell et al. 2004; ICLUS-A1 population projection).

Figure 4

Cumulative probability density functions of national population-weighted summertime O₃ concentration changes between 2000 and ca. 2050 from the seven sets of climate change–air quality modeling results (ICLUS-A2 population projection; other population projections yielded similar results).