2014 GSA Annual Meeting in Vancouver, British Columbia (19–22 October 2014)

Paper No. 74-10
Presentation Time: 3:35 PM

ENVIRONMENTAL LEAD CONTAMINATION: AN EXAMINATION OF GEOCHEMISTRY IN NW NIGERIAN GOLD MINING VILLAGES THROUGH MULTIVARIATE STATISTICAL CORRELATIONS, GIS GEOSTATISTICAL ANALYSIS, AND XRD CLUSTER ANALYSIS


TAYLOR, Lucy C.1, PICKERING, Rebecca A.2, BALLOW, William B.3, ELRICK, K.A.1 and DEOCAMPO, Daniel M.1, (1)Geosciences, Georgia State University, 24 Peachtree Center Avenue Northeast, Atlanta, GA 30303, (2)Geosciences, Georgia State University, PO Box 4105, Atlanta, GA 30302, (3)Department of Geosciences, Georgia State University, Atlanta, GA 30303

Artisanal gold mining has contaminated many villages in northwest Nigeria killing >400 children, and affected thousands. Ore is processed within the villages by mortar/pestle or gas powered grinding mills, washed, and mercury amalgamated to recover gold using home cooking pots. Processed ore and dust samples from children’s sleeping areas and play areas were collected from 54 villages by the USCDC and Nigerian colleagues in the summer of 2012. Our team at Georgia State University analyzed the samples using XRD, ICP-AES, and GIS.

Previous geochemical work has determined that 23 out of 772 samples contained lead (Pb) concentrations that exceed the US EPA Residential Soil Screening Level of 400 ppm. For instance Pb levels as high as 1,400 ppm in children’s sleeping areas, as high as ~61,000 ppm in ground ore, and as high as ~34,000 ppm in washed ore from ICP-AES analysis. In this study, we examine the multivariate correlations and GIS geostatistical analysis of the ICP-AES data as well as an XRD cluster analysis of 61 samples containing >100 ppm of Pb.

The multivariate matrix identified that Pb is strongly correlated with Cu and S where Pb is only slightly correlated with P and As. To illustrate and interpret the geochemistry spatially, Empirical Bayesian Kriging (GIS Geostatistical Analyst tool) was implemented on a region scale also revealing a strong spatial correlation between Pb and S. This indicates sulfide lead-bearing minerals, such as galena, are more likely a cause of Pb exposure than of phosphate lead-bearing minerals. In addition, XRD cluster analysis discriminated 3 main groups and one outlier (Rb-Li-Sb) within the 61 samples of >100 ppm Pb. Pattern analysis is underway to determine the statistical controls; preliminary results suggest the presence/absence of galena vs. cerussite has no controls over the cluster grouping, however, the two highest concentrations exceeding 3% Pb were grouped together.

These findings offer fundamental tools for future studies in this region by providing statistical correlations of geochemistry as well as visual illustrations of high risk areas in Northwestern Nigeria. Further work to be considered is C/N/S analysis to obtain a carbon number to determine the presence/absence of galena vs. cerussite in the area.