Modeling the geometry of a mineral deposit domain with a Potential Field
Résumé
The potential field method, successfully applied in geological modeling, has been developed in the very different context of mineral deposit modeling, where in comparison to geological mapping we have many drill holes but usually few structural data. After a reminder of the fundamentals of the method, new advances related to the incorporation of soft data and to the assessment of the uncertainty are presented. The principle of the method is to derive the geometry of the domain under study from a 3D interpolation of a scalar field, known as the potential field, by cokriging from information on contacts with drill holes and on structural data linked with the gradient of the potential field. Still the information brought by the drill holes is much more abundant than just the transitions from outside to inside the domain (or the reverse). Soft information is efficiently added to the hard contact data by means of additional control points processed with the Gibbs sampler algorithm. Finally the potential field approach provides side products such as the cokriging variance, and the gradient of the estimated potential field, which can be turned into an uncertainty on the location of the domain boundary or used to map the probability that a specific location lies within the domain. The proposed potential field method is put into practice in a case study of a gold rich porphyry deposit, La Colosa. On the exploration drillholes, hardness measurements related to lithology and alteration have been collected by Equotip equipment. After simple pre-processing of this data, one domain has been modeled and probability map has been calculated.