Spatial regression modelling of particulate pollution in Calgary, Canada.

摘要: The study presents a spatial analysis of particulate pollution, which includes not only matter, but also black carbon, pollutant growing concern for human health. We developed land use regression (LUR) models two matter size fractions, PM2.5 and PM10, δC, an index calculated from carbon (BC)—a component PM2.5—which indicates the portion organic versus elemental BC. LUR were estimated over Calgary (Canada) summer 2015 winter 2016. As all samples exhibited significant autocorrelation, autoregressive lag (SARlag) error (SARerr) computed. SARlag preferred pollutants in both seasons, yielded goodness fit aligned with or higher than values reported literature. consistent sets predictors, representing industrial activities, traffic, elevation. obtained model coefficients then combined local variables to compute fine-scale concentration predictions entire city. predicted concentrations slightly lower less dispersed observed ones. Consistent pollution records, prediction maps road network, areas, eastern quadrants Lastly, results corresponding PM 2010 2011 considered. While small 2010–2011 sample hampered multi-temporal analysis, we cautiously note comparable seasonal patterns association fine fractions 5-year interval.