Dendrograms, heat maps and principal component analysis – the practical use of statistical methods for source apportionment of trace elements in PM10

Abstract

Dendrogram (DE), heat map (HM) and principal component analysis (PCA) methods were used in order to identify possible emission sources of As, Cd, Co, Cr, Hg, Mn, Ni, Pb, Sb and Se in ambient PM10 collected in the surroundings of working power plants. Each statistical tool resulted in slightly different clusters. The best approximation of possible emission sources was received by the use of statistical analysis of trace-element concentrations combined with characterization of the sampling sites. In the study, PCA was indicated as the most useful statistical tool for source apportionment of trace elements in PM10. Major sources identified by PCA included: (1) coal combustion, (2) soil and road-dust resuspension, (3) the use of pesticides and (4) waste incineration.

Keywords:

• Dendrogram
• heat map
• PCA
• PM10
• trace elements
• outdoor air

Additional information

Funding

This work was supported by the Faculty of Energy and Environmental Engineering, Silesian University of Technology (statutory research). All the results used in this paper have benefitted from funding from the National Science Centre Poland under grant 2012/05/B/ST10/00524