An area in the Magondi Belt, Zimbabwe, has been chosen for mineral potential mapping using the favourability functions approach.The available datasets comprising of an old geological map, a detailed airborne total magnetic field survey, and geochemical samples at the nodes of an exploration grid, have been integrated using seven different inference techniques through the joint probability function under the conditional independence hypothesis.A geological conceptual model has been adopted for the representation of the mineralization occurrence, in order to select appropriate geospatial indicators of mineralization, while existing mines in the area have been used as a training set for the model.Among the different integration techniques which have been tested, some have proven to be robust in correctly predicting all the known mine sites.It has been thus possible to draw favourability maps using existing data, which indicate the most promising areas for exploration and detailed mapping efforts for mineral exploitation.Even more important, using sensitivity analysis of the favourability functions allowed to evaluate the most important factors controlling mineralization occurrence, and thus worth additional future investigation.So far, a completely data driven approach has been supported because limitations due to lack of geological knowledge limited the possibility of using expert knowledge-based modifications of the conditional probabilities.Improvements in prediction can be achieved through a more detailed geological description of the area based e.g. on the interpretation of remotely sensed data.

Tsehaie Woldai

Validation and sensitivity analysis of a mineral potential model using favourability functions

ivySCI AI Smartly Parses PDF, Answers Researchers' Questions, and Helps You Understand Papers in Seconds

Journal Info