Publication Type:
ThesisSource:
Department of Geology, Laurentian University, Volume MSc, p.155 (2007)Abstract:
This study describes a new technique to identify and quantify areas of host rock alteration to aid in the search for Archean volcanogenic massive sulphide (VMS) deposits. Major element oxide data from a suite of unaltered Cenozoic volcanic rocks is used to develop a set of reference curves to calibrate a new correction procedure to detect and correct for metasomatism, refining the Beswick and Soucie (1978) correction procedure. The weight percent values for the major element oxides in the Cenozoic suite were converted to molecular proportions and were plotted in the form of log X/Z vs. log Y/Z to represent primary, unaltered compositional trends for each oxide tested. These compositional trends were modeled by regression analysis to produce a best-fit curve. A sequential correction process was developed that corrects for mobility of eight major element oxides in the Cenozoic data. Using the regression curves as a reference, samples which plot away from the trend are adjusted back to the trend using the regression equation where two of the components were fixed or assumed to be immobile or conserved.<br/>These regression curves and equations for the Cenozoic suite were then compared to a data set from an area of Archean volcanic rocks in the Confederation Lake area of northwest Ontario, which contains a past-producing VMS mine. The regression equations from the Cenozoic compositional trends were used to quantitatively estimate the relative individual oxide gains and losses for the Confederation Lake data set in the same manner as the Cenozoic samples.<br/>The relative oxide gains and losses from the Confederation Lake area were then used in a discriminant function analysis to develop a formula, which separated mineralized (massive sulphide bearing) host rocks from the general population of regionally metamorphosed Archean volcanic rocks. Individual samples, when processed through the discriminant analysis function, were ultimately expressed as a probability of being mineralized. <br/>The new technique was successfully tested by including data from other Archean greenstone belts to confirm its ability to identify VMS-mineralized host rocks.