Abstract
Cluster analysis is shown to be a useful technique for the identification and characterization of atmospheric aerosol particles. A general procedure for combining the cluster analysis results from a large number of samples is described and shown to be successful for comparing such samples. Two similarity measures, one not previously described, are evaluated for use in cluster analysis. The new analysis procedure was applied to the examination of 11 coarse-fraction and 19 fine-fraction aerosol samples collected during September and October, 1984 at Ny Alesund, Spitzbergen, Norway, ~ 1400 km north of the Arctic Circle. The aerosol at this site consists primarily of Si-, Cl- and S-rich particles, the relative abundance of each varying from sample to sample. Considerable differences in composition were observed between simultaneously collected coarse- and fine-fractions.
Original language | English (US) |
---|---|
Pages (from-to) | 1649-1657 |
Number of pages | 9 |
Journal | Atmospheric Environment (1967) |
Volume | 21 |
Issue number | 7 |
DOIs | |
State | Published - 1987 |
Keywords
- Aerosol
- Arctic haze
- atmospheric
- cluster analysis
- multi-variate analysis
- particles
- scanning electron microscope
ASJC Scopus subject areas
- General Environmental Science
- General Earth and Planetary Sciences