Spatiotemporal modeling of air pollutants and their health effects in the Pittsburgh region

Air pollutants have been associated with adverse health outcomes such as cardiovascular and respiratory diseases through epidemiological studies. Spatiotemporal and spatial statistics are widely used in both exposure assessment and health risk estimation of air pollutants. In the current paper, spatiotemporal and spatial models are developed for and applied to four specfic topics about air pollutants: (1) estimating spatiotemporal variations of particulate matter with diameter less than 2.5 um (PM2.5) using monitoring data and satellite aerosol\ud optical depth (AOD) measurements, (2) estimating long-term spatial variations of ozone (O3) using monitoring data and satellite O3 profile measurements, (3) spatiotemporal associating acute exposure of air pollutants to mortality, and (4) spatiotemporal associating chronic air pollution exposure to lung cancer incidence. Environmental, socioeconomic and health\ud data from Allegheny county and the State of Pennsylvania are collected to illustrate these techniques.\ud \ud The public health significance of these studies includes characterizing the exposure level of air pollutants and their health risks for mortality caused by cardiovascular and respiratory diseases and lung cancer incidence in the Pittsburgh region and developing novel spatiotemporal models such as spatiotemporal generalized estimating equations for the regression analysis of spatiotemporal counts data, especially for the massive spatiotemporal\ud data used in epidemiological studies