Potential contamination of stream waters by ultramafic mining sediments: Identification of geochemical makers (New Caledonia)

Aurélie Boula (1, 2) , Christine Laporte-Magoni (1, 2) , Peggy Gunkel-Grillon (1, 2) , Olivier Bour (3) , Nazha Selmaoui-Folcher (1, 2, 4)

1 UNC - Université de la Nouvelle-Calédonie
2 ISEA - Institut de sciences exactes et appliquées
3 GR - Géosciences Rennes
4 IPAL - Image & Pervasive Access Lab

The ultramafic massifs of New Caledonia are mainly composed of regolites with lateritic soils rich in metals such as Co, Fe, Mg and Ni. The nickeliferous ore is exploited by opencast mines. After rainfalls, erosion and particle runoff cause the mobilization of metals and metalloids in surface waters. The objective of this study is (i) to determine the geochemical background of metals and metalloids in surface waters draining a mining ultramafic massif, (ii) to identify the geochemical markers influenced by the mining of nickeliferous ore (iii) and to define discriminant values of the mining impact. Hydrochemical monitoring was carried out on five streams draining the Koniambo massif over a 15-years period. Twenty-four elements were analyzed for their total concentration and 18 for their dissolved concentration (<0.45 μm). The descriptive analysis shows that one river has higher total concentrations of As, Co, Cr, Ni, Mn, Al, Fe, Si and Mg and higher dissolved concentrations of Al, Cr, Ni than the other 4 rivers. This difference results from an old mine whose erosion exacerbates sediment transport. These 5 rivers drain the same ultramafic massif of small surface area (21 km2). The description by the boxplot method of the 4 non-impacted rivers allowed us to define a geochemical background. The graphical comparison between the geochemical background with the median values of the river impacted, validated by the ANOVA one way, allowed us to identify markers of the mining anthropization. The use of a supervised classification as a decision tree ranks the markers according to their discriminating power. In addition, this method determines the range of threshold values that separate impacted from non-impacted rivers. The elements Cr, As, Si, Mg and Al are identified as discriminating markers for total concentrations. Only Cr has been identified as a discriminating metal for the dissolved fraction with an interval around 58 μg/l. This concentration is close to the limit defined by the WHO for drinking water which is 50 μg/l.

Mots clés: hydrogeochemical background; ultramafic complexes; surface waters; total metal concentration; dissolved metal concentration; machine learning

Domaines: Sciences de l'environnement; Hydrologie; Géochimie

Aurélie Boula, Christine Laporte-Magoni, Peggy Gunkel-Grillon, Olivier Bour, Nazha Selmaoui-Folcher. Potential contamination of stream waters by ultramafic mining sediments: Identification of geochemical makers (New Caledonia). Journal of Geochemical Exploration, 2022, 232, pp.106879. ⟨10.1016/j.gexplo.2021.106879⟩. ⟨insu-03330547⟩