Geostatistical and Multivariate Analysis of Heavy Metal Pollution of Coal-mine Affected Agricultural Soils of North-eastern India

Abstract

Total concentrations of heavy metals in the soils of mine drainage and surrounding agricultural fields in the
Ledo coal mining area of Tinsukia district, Assam, India, were investigated using statistics, geostatistics
and GIS techniques. The amounts of iron (Fe), manganese (Mn) and zinc (Zn) were determined from 83
soil samples collected within the contaminated area. The mean concentration of Fe, Mn and Zn were
28585, 627 and 227 mg kg-1, respectively. The greatest and the smallest standard deviation were observed
in the Fe (7506) and pH (0.44), respectively. All heavy metals exhibited a medium variation (15-50%).
Analysis of the isotropic variogram indicated that the Mn and Zn semivariograms were well-described with
the exponential model, with the distance of spatial dependence being 1083 and 994 m, respectively, while
the Fe semivariogram was well-described with the spherical model, with the distance of spatial dependence
being 1784 m. Thus, the length of the spatial autocorrelation was much longer than the sampling interval of
500 m. The spatial distribution maps of Fe, Mn and Zn showed that high concentration of heavy metals was
located in the low-lying rice field and near coal mining site. Multivariate statistical analyses and principal
component analysis suggested that Fe was derived from anthropogenic sources, particularly coal mining
activities, whereas Mn and Zn were derived from lithogenic and/or anthropogenic sources.