Incorporation of geological constraints and semivariogram scaling law into geostatistical modeling of metal contents in hydrothermal deposits for improved accuracy

• DOI: 10.1016/j.gexplo.2021.106901
• Publication Date: 2021-09-29T07:30:10Z
• Authors: Katsuaki Koike; Takuya Kiriyama; Lei Lu; Taiki Kubo; Mohamad Nur Heriawan; Ryoichi Yamada
• Keywords: 01 natural sciences; 0105 earth and related environmental sciences

Abstract Detailed metal distributions in a deposit can be used to determine the generation mechanism and process of the deposit type and improve mining development efficiency. Two geostatistical methods were developed that incorporate geologic information and a semivariogram scaling law to improve the spatial estimation accuracy of metal contents at the fine scale. In the first method, a binary dataset is prepared by assigning 1 to the location of the targeted geologic category in each borehole column and 0 elsewhere. Principal component analysis is implemented to generate a geologic model using principal values and ordinary kriging (OK). By overlapping with a 3D geologic model, OK of the main metal contents is then performed by limiting the content data with the same geologic category as the calculation point. This method is termed kriging with a geological constraint (KGC) and focuses on the correlation between the geologic category and metal content. The next step downscales the KGC model using a semivariogram scaling law to obtain finer spatial resolution (KGC-DS). This law is experimentally derived by repeatedly thinning out the original sample data and approximating the relationship between the mean distance of the data pair and semivariogram parameters (range, nugget effect, and sill) using a power function. This approximation can produce semivariograms in a virtual scenario with more sample data than the original. These two methods were applied to two types of hydrothermal deposits: one of the largest kuroko (volcanogenic massive sulfide) deposits in northern Japan using 77 vertical or sub-vertical drilling data and one porphyry copper deposit in northern Sulawesi, Indonesia using 58 drilling data. In the first case study, the accuracy in the high-content zones increases and the metal content distribution becomes considerably more detailed than the OK result by reducing the smoothing effect. The superposition of the geologic model and high-content zones can be used to trace the ore solution flows and interpret the deposit formation process. For the second case study, oxidation, supergene, and hypogene mineralization zones with high copper contents are clearly detected and the relationship between high-content zones and hydrothermal alteration type is identified. The results of these two case studies therefore demonstrate the effectiveness of KGC and KGC-DS.