Browsing by Author "Ru, Muye"

Air pollution, human health, and the economy are a connected system. Substantial efforts have been made in understanding and modeling these connections, so that we can predict the health and economic impacts of changes in air pollution to inform policymaking. However, simplifications and knowledge gaps are still present, and understanding of how these influence the entire system remain limited. This dissertation studies some of these simplifications and knowledge gaps, develops new tools to quantify them, and discusses their impacts to the system. These tools and impacts are used to facilitate more comprehensive evaluation of the impacts of air pollution, especially those that could have been underestimated, and to examine their sensitivity to certain assumptions. In general, air pollution has greater impacts than previously estimated, and future projections include assumptions that seem to be optimistic in the light of historical trajectories. The concluding chapter discusses implications for the sensitivities and uncertainties of the whole system, and how the system of economy-pollution-health should be integrated with the economy-climate system.

The relationships between economy and emissions of major air pollutants and greenhouse gases are basic inputs used to generate historical estimates and future emission scenarios. I show that these relationships vary over time across different widely-used global inventories, indicating the presence of large uncertainties within historical emission trajectories. More specifically, my examination of four major sectors (power, industry, residential, and transportation) and three pollutants (sulfur dioxide, carbon dioxide, and black carbon) demonstrates that long-term income-emission trajectories are both sector and pollutant specific. When assessing future projections of income-emission trajectories in reference scenarios, however, I show the persistence of faster rates of emission declines and estimates of earlier turnover incomes than estimated from historical data. This indicates some underlying uncertainties in such trajectories and that future projections of income-emission trajectories for integrated assessment should be used with appropriate caution.

In the next part of my analysis, I focus on the relationship between air pollution and health outcomes, which is based on epidemiological evidence. To date, epidemiologically-based quantitative relationships have been developed for many cardiovascular and respiratory diseases, as well as diabetes. A knowledge gap persists with respect to the exposure-response relationship between air pollution and incidence of dementia, due to relatively limited evidence. I assessed this exposure-response relationship by using a meta-analysis approach to collect data from existing epidemiological studies. I have developed an exploratory model of this relationship, and estimated that, globally, 1.1M [0.6M, 1.6M; 5-95% confidence] global incident cases and 0.34M [0.17M, 0.48M] premature deaths from dementia were attributable to ambient fine particulate matter (PM2.5) pollution in 2015. In addition, using model reconstructions of surface PM2.5 levels, I have shown that this burden of disease has grown 60~70% since 2000 as a consequence of globally increased exposures to ambient PM2.5. For the first time, our meta-analysis approach enables us to estimate that ambient PM2.5 pollution may be responsible for 15% of the premature deaths and 16% of the morbidity burdens associated with dementia across all risk factors.

The third linkage closes the loop of the air pollution, health, and the economy system, by estimating the direct economic costs associated with morbidity burdens (hospital admissions, emergency room visits, restricted activity days, etc.). Unlike mortality burdens, costs associated with morbidity burdens directly affect market activities. These costs include medical expenditures shared by households and the public health sector, productivity loss due to lost work hours, and the costs of private and public care for the ill. Identifying and estimating these costs is important for decision-making. For this analysis, I developed statistical models of seven exposure-response relationships for five acute morbidity endpoints, using data collected from meta-analysis. I quantified uncertainties associated with these exposure-response functions by randomization and simulation. I showed that these morbidity effects per unit increase in exposure generally reduce as exposures increase, unlike previous approaches that assume they are fixed over the entire range of exposures. Therefore, these functions are particularly useful to analyze impacts of air pollution in high-exposure regions or on a global scale.

I concluded that this dissertation demonstrated the importance of further understanding of uncertainties in this economy-pollution-health system. In particular, we should study interactions and propagations of uncertainty throughout this non-linear system. Moreover, research on the economy-pollution-health and the economy-climate systems should be more integrated, because the two systems overlap and the research methodology to analyze each also has many similarities. Insights from one system can inform challenges from the other system.