Identifying meteorological drivers of PM_2.5 levels via a Bayesian spatial quantile regression

Abstract

Recently, due to accelerations in urban and industrial development, the health impact of air pollution has become a topic of key concern. Of the various forms of air pollution, fine atmospheric particulate matter (PM 2.5 ; particles less than 2.5 micrometers in diameter) appears to pose the greatest risk to human health.While even moderate levels of PM 2.5 can be detrimental to health, spikes in PM 2.5 to atypically high levels are even more dangerous. These spikes are believed to be associated with regionally specific meteorological factors. To quantify these associations, we develop a Bayesian spatiotemporal quantile regression model to estimate the spatially varying effects of meteorological variables purported to be related to PM 2.5 levels. By adopting a quantile regression model, we are able to examine the entire distribution of PM 2.5 levels; for example, we are able to identify which meteorological drivers are related to abnormally high PM 2.5 levels.Our approach uses penalized splines to model the spatially varying meteorological effects and to account for spatiotemporal dependence. The performance of the methodology is evaluated through extensive numerical studies. We apply our modeling techniques to 5 years of daily PM 2.5 data collected throughout the eastern United States to reveal the effects of various meteorological drivers.

Decadal Trends in the Temperature Dependence of Summertime Urban PM2.5 in the Northeast United States

“…Early in the record, the conversion of these emissions to sulfate aerosols explains most of the observed temperature dependence. This explanation for the temperature dependence of PM 2.5 and sulfate differs from most prior analyses that focused on stagnation and the accelerated conversion of SO 2 to sulfate as temperature warms. − …”

Quantile based modeling of diurnal temperature range with the five‐parameter lambda distribution

“…The regression model for diurnal temperature range only includes spatial fixed effects. The inclusion of spatial random effects might improve model performance (Lum & Gelfand, 2012; Self et al, 2021). Additionally, attempts to estimate parameter uncertainty will be affected by the temporal autocorrelation of diurnal temperature range, which is statistically significant for lags up to approximately one week.…”

Quantile based modelling of diurnal temperature range with the five-parameter lambda distribution

“…The regression model for diurnal temperature range only includes spatial fixed effects. The inclusion of spatial random effects might improve model performance (Lum & Gelfand, 2012;Self et al, 2021). Additionally, attempts to estimate parameter uncertainty will be affected by the temporal autocorrelation of diurnal temperature range, which is statistically significant for lags up to approximately one week.…”