Discriminating soil-contamination sources using combination of magnetic parameters

Abstract

This paper demonstrates the efficiency of magnetic techniques to provide identification and differentiation of contaminating emission sources. Three areas from different emission sources were selected for this study: roadside soils (27 samples), soils near a power plant (37 samples), and soils near a cement plant (32 samples). Self-organizing map (SOM) has become a largely used methodology to classify large amounts of data. From the U-matrix it is easy to see that the bottom three rows of the SOM formed a very clear cluster. The soils near the power plant exhibited high values of χfd and χARM/SIRM ratio. The separating factor between roadside soils and soils near the cement plant is that the former had higher value of χ. Discriminant analysis of the four magnetic variables indicated that the magnetic parameter χARM/SIRM was the most powerful in discriminating among soil samples from the three areas. Biplots combining magnetic parameters allow the identification and differentiation of different pollutant emission sources. These results lead to the conclusion that the χARM/SIRM ratio is a good indicator in discriminating different contamination sources, which is consistent with the previous result. Soil samples from roadside and near the power plant were dominated by ferrimagnetic materials while soil samples near the cement plant had low χARM/SIRM and relatively high SIRM/χ values, indicative of being contaminated by hematite.