A Multistep Methodology for Building a Stochastic Model of Gold Grades in the Disseminated and Complex Deposit of Casas Novas in Alentejo, Southern Portugal

This paper proposes a multistep approach for creating a 3D stochastic model of gold grades in a complex disseminated auriferous deposit located in the Alentejo region, southern Portugal. The approach involves the following steps: (i) the creation of a 3D low-resolution vector object model of two geological domains that better discriminate gold grades; (ii) the definition of three intervals for gold grades (low, intermediate, and high values); (iii) the estimation of the probability of each grid block belonging to each interval constrained by the geological domains; (iv) the calculation of local conditional cumulative distribution functions (CCDFs) of gold grades, for each block, taking into account the geological domains and the borehole data; and finally (v) the simulation of images of gold grades. The results demonstrate the usefulness and accuracy of the proposed procedure, as they are in compliance with the conceptual model, the distribution laws are conditioned by the geological domains, and the transition of grades between domains is continuous.