Integration of air quality data for improved estimates of PM2.5 source impacts

Air quality modeling;Particulate matter

Ivey, Cesunica Elizabeth ; Russell, Armistead G. Mulholland, James A. Chen, Yongsheng Odman, Talat M. Weber, Rodney J. Chang, Howard Civil and Environmental Engineering ; Russell, Armistead G.

University:Georgia Institute of Technology

Department:Civil and Environmental Engineering

关键词: Air quality modeling;    Particulate matter;   

Others  :  https://smartech.gatech.edu/bitstream/1853/55643/1/IVEY-DISSERTATION-2016.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

Human exposure to air pollution is a known risk factor for the development or worsening of diseases and increased mortality. Traditionally, air quality data availability for health analyses are spatially and temporally sparse, which limits the capabilities of the health studies and introduces bias into methods and results. This dissertation is a presentation of novel data assimilation methods for air quality modeling of PM2.5 source impacts for use in epidemiologic analyses. The presented methods improve spatially and temporally resolved source impact estimates and highlight uncertainties in presently used modeling techniques. The methods developed include a novel hybrid source apportionment method that uses observed and modeled data to generate spatially and temporally resolved source impacts, where some sources are commonly unresolved by traditional methods. The hybrid approach employs a nonlinear optimization method to generate adjustment factors that, when applied to CMAQ-DDM data, revise the original source impacts to better reflect observed concentrations. Additionally, a novel method for optimizing PM2.5 source profiles is presented which uses observed concentrations to generate a revised profile that reflects local conditions. Revised PM2.5 source profiles are generated for monitored locations across the United States, and the spatio-temporal characteristics of the revised profiles are analyzed. Further, a method is developed to address the modeling uncertainties associated with estimating concentrations and source impacts on secondary PM2.5. The method improves the estimates of sulfate, nitrate, ammonium, and secondary organic carbon concentrations, while providing an estimate of numerical bias in source impacts on the species. Overall, the methods and results presented in this dissertation provide insight on the greatest impacting PM2.5 sources in the U.S., such as vehicle emissions, coal combustion, biomass burning, and agricultural/livestock activities. The works presented here also provide insight on the biases present in source apportionment and PM2.5 modeling techniques. The methods and results presented offer benefits for scientists interested in policy and health applications, such as establishing NAAQS attainment for municipalities and further exploring links between human exposure to PM2.5 and adverse health effects.