Geochemical Classification of Archean Volcanic Rocks in the Blake River Group, Abitibi Greenstone Belt, Québec

Sarane Sterckx1, Pierre-Simon Ross1 and Jean Goutier2, (1)Institut national de la recherche scientifique, Centre Eau Terre Environnement, Quebec City, QC, Canada, (2)Géologie Québec, Rouyn-Noranda, QC, Canada

Contact First Author: Sarane Sterckx; sarane.sterckx@ete.inrs.ca

Abstract ID#: 34952

 

English Abstract:
The Archean Abitibi Greenstone Belt is one of the best places in the world to explore for volcanogenic massive sulphide deposits (VMS), with 810 Mt of ore known so far. The Blake River Group (BRG) in the Rouyn-Noranda area includes two important mining camps, Noranda and Doyon-Bousquet-LaRonde, representing together about 374 Mt. VMS deposits are partially controlled by stratigraphy and, for the BRG, few regional marker horizons are known. Therefore, an improved knowledge of the volcanic stratigraphy would be valuable, and a chemo-stratigraphic approach could help, especially if combined with the large high-precision U-Pb geochronology dataset now available.

We have compiled more than 2000 complete geochemical analyses (major and trace elements) for the Québec side of the BRG. With such a large database, multivariate statistical analysis is needed to utilize the data to its full potential. The BRG has been divided into a number of volcanic formations. We have so far focussed on two of these, the Noranda formation (hosting many VMS deposits of the Noranda camp) and the Reneault-Dufresnoy. For each formation we first split the data into mafic-intermediate and felsic samples, and discard a few outliers. Then we use hierarchical clustering analysis to place samples into groups. Examination of spidergrams and extensive tests leads us to use the elements Th, Nb, La, Ti, and Yb for the hierarchical clustering analysis. These elements (except La) are known to be largely immobile during hydrothermal alteration and metamorphism. The dendrograms shows the hierarchy of clusters, and we decide on the optimal number of groups based on (1) geological/geographical consistency of the groups; (2) maximizing the differences between groups; (3) maximizing the similarity within groups. We interpret those groups as volcanic units having different geochemical signatures, an interpretation consistent with the stratigraphic polarity and the U-Pb data.