Geochemistry of natural and anthropogenic fall-out (aerosol and precipitation) collected from the NW Mediterranean: two different multivariate statistical approaches |
| |
| Institution: | 2. Tsinghua University, Beijing, China;3. National Supercomputing Center in Wuxi, Wuxi, China;4. Maxeler Technologies, London, United Kingdom;5. Imperial College London, London, United Kingdom;1. University of Michigan College of Pharmacy, 4549 CC Little, 428 Church Street, Ann Arbor, MI 48109-1065, United States;2. Department of Clinical Pharmacy, University of Michigan College of Pharmacy, 3563 CCL, 428 Church Street, Ann Arbor, MI 48109-1065, United States;3. University of Michigan College of Pharmacy, 428 Church Street, Ann Arbor, MI 48109-1065, United States;4. University of Michigan College of Pharmacy, Department of Clinical Pharmacy, Room 2568B CC Little, 428 Church Street, Ann Arbor, MI 48109-1065, United States |
| |
| Abstract: | The chemical characteristics of the mineral fractions of aerosol and precipitation collected in Sardinia (NW Mediterranean) are highlighted by means of two multivariate statistical approaches. Two different combinations of classification and statistical methods for geochemical data are presented. It is shown that the application of cluster analysis subsequent to Q-Factor analysis better distinguishes among Saharan dust, background pollution (Europe-Mediterranean) and local aerosol from various source regions (Sardinia). Conversely, the application of simple cluster analysis was able to distinguish only between aerosols and precipitation particles, without assigning the sources (local or distant) to the aerosol. This method also highlighted the fact that crust-enriched precipitation is similar to desert-derived aerosol. Major elements (Al, Na) and trace metal (Pb) turn out to be the most discriminating elements of the analysed data set. Independent use of mineralogical, granulometric and meteorological data confirmed the results derived from the statistical methods employed. |
| |
| Keywords: | |
| 本文献已被 ScienceDirect 等数据库收录！ |
|
