Perpustakaan Digital ITB

Exploration activity is indispensable to minimize the risks inherent in the mining industry where exploration must provide complete information about the deposit such as the resource estimation model. Resource estimation using conventional or linear geostatistical methods sometimes produces less accurate results. Therefore the non-linear geostatistical method such as multivariate geostatistics is expected to provide more accurate estimation. The case study for this research is Pongkor gold vein located in the district of Bogor, West Java, Indonesia_ Pongkor gold vein deposit composed of gold and silver mineralization. This study aims to estimate the blocks and their variance estimation on Au-Ag grades using Ordinary Cokriging (COK) method as one of the multivariate geostatistical methods. However the linear geostatistical method such as Ordinary Kriging (OK) as frequently used methods was performed for comparison. Both methods were validated using composited grade data by considering the slope of regression and coefficient of correlation between estimated grades and composited actual grades. The Au-Ag grades data was divided into two parts, i.e. data with and without top cut. Variogram of Au-Ag grades and their cross-variogram were used as estimation parameters for COK and OK. The correlation coefficient of estimate grades and composite sample grade showed the strong correlation with coefficient more than 0.70. COK produced smaller estimation variances on Au-Ag estimated grades than OK method. The cumulative distribution frequency of estimate grades and kriging variance depicted the difference and accuracy in estimation results of both COK and OK methods. The data with and without top cut showed no improvement in the estimate grade and validation results.