A statistical analysis of particulate data sets in Brisbane, Australia

Abstract

This paper examines the relationship between various types of particulate data collected in Brisbane, Queensland, Australia—PM10, TSP, lead and nephelometer data. Statistical tests on the data show: the PM10 data collected at all sites throughout the city are closely correlated, the same for lead but not TSP data; PM10 and TSP data collected at the same site are closely correlated, and both are only weakly correlated with nephelometer data; the lead data sets at the same site are significantly correlated with the TSP, PM10 and nephelometer data; and windspeed data generally show significant negative correlation only with the lead data. The results indicate that PM10 data would appear to be approximately a constant proportion of the TSP data throughout the year, and there is no evidence to suggest that anthropogenic emissions only contribute to the PM10 fraction or that ‘natural’ sources contribute only to the coarser fraction of the TSP data. However the lead data and the nephelometer data arise from a distinctly different pattern of sources to the TSP and PM10 data, and events, with motor vehicle emissions being clearly important in the area studied (comprising more than 95% of all lead emissions).