THE APPLICATION OF EVIDENTIAL BELIEF FUNCTION IN THE INTEGRATION OF REGIONAL GEOCHEMICAL AND GEOLOGICAL DATA OVER THE IFE-ILESHA GOLDFIELD,NIGERIA

There are several methods available for the integration of geological, geophysical and even remotely sensed data sets. Several published rep orts have discussed the successful application of these integration te chniques, including geologically orientated geographical information s ystems (GIS), to non-renewable resource exploration. Many geoscience d ata sets often have only partial coverage, and in most cases have diff erent spatial resolution. These shortcomings plague statistical integr ation schemes employed in resource evaluation. In this paper, the evid ential belief function approach is examined and applied to the integra tion of geological and geochemical data sets over the Ife-Ilesha goldf ield, southwest Nigeria. In using this approach, nine maps of trace el ement concentrations (whole rock) and that of the bedrock geology of t he test area were digitised. The characteristic grid values of the dig itised maps were converted to degrees of belief, with a Au bearing min eral deposit as the target proposition. These resulting spatial data s ets were pooled using the Dempster-Shafer combination rule. The result ing maps define portions of the study area with known auriferous miner al occurrences and even outline areas of past mining operations. There are additional areas with high interest that are not linked with know n mineralisation and could therefore represent viable exploration targ ets. The Dempster-Shafer method of evidential reasoning appears to be suitable in identifying areas of known Au mineralisation in the test a rea. One of the major difficulties of this method is that the evidenti al belief function depends on the exploration target, which can only b e defined, at present, by the interpreter's intuitive and qualitative knowledge of the environmental proposition being explored. (C) 1997 El sevier Science Limited.