REGIONAL HEAVY MINERAL SURVEY IN THE EXPLORATION FOR GOLD USING REGRESSION - GRENVILLE PROVINCE, SOUTHWESTERN QUEBEC

A geochemical reconnaissance heavy mineral survey has been carried out in the area of the Gatineau River Valley, over a part of the Grenville Province of Quebec. The majority of samples show gold concentrations below the detection limit of 5 ppb, and anomalies are erratic. To circumvent these problems and to present an accurate image of the gold-bearing potential of the area, a regression analysis has been applied to the samples whose gold concentrations were above the detection limit. The regression of gold on chalcophile pathfinder elements (As, Se and Pb) explains 11% of the gold variance. Bedrock lithologies, overburden cover and physico-chemical characteristics are important factors influencing the chalcophile element dispersion pattern. Principal component analysis was used to characterize the principal inter-element relationships and to identify the factors influencing the dispersion of chalcophile elements. A filtering procedure was used to remove the background effects influencing the pathfinder elements, and to enhance the influence of mineralization. This procedure uses standard formulas for partial reconstruction of the data matrix from results of principal component analysis. The regression model constructed with these corrected data explains 20% of gold variance. It constitutes an improvement over the model which uses non-filtered data. The equation inferred from regression was applied to the whole data set and provided an estimation of gold at each sampling point. Exploration targets are defined from highest gold concentrations estimated by regression. The regression model aided in the definition of some well defined anomalies associated with geological structures favourable to gold mineralization.