Data integration using knowledge driven GIS modeling for Cu and Pb deposits Geology The study area is situated the in northwest part of Yunnan province between Latitude (26 07 to 28 66N) and Longitude (97 92 to 101 89 E),Fig (1)shows location map of study area. Study area, which is commonly called as “ three river” area. It is the combined part of Eurasia plate and Gondwana plate. The main rock associations in the region include variable composites of metavolcanics, principally Meta andesites, metatuff, schist, rhyolites, and sediment rocks such as limestone, sandstone, shale and conglomerate, and igneous rocks. The different kinds of rocks and formations in the study area are concisely described as follow:  Location map of Lanping basin study area, northwest, Yunnan GIS dataset GIS data include the following factors: Landsat ETM+ imagery path 132 row 42 was georeferenced for ground control points identified both on the image and topographic map. Host rocks were digitized from the 1:500 000 scale geological map of Yunnan province, the lithological units were added into the attribute table using Arc/info software. Major fault within the selected area are digitized from 1:500 5000 scale geological map of Yunnan province, other structures are derived from the interpretation of gravity and magnetic field data carried out by geophysical bureau of Yunnan province. These sets were merged into one layer as the first structural evidence. Major lineaments are extracted from landsat ETM+ imagery. Weights of Evidence A weight of evidence is quantitative method that was originally developed for spatial application of medical diagnosis. This method was developed by geological Survey of Canada, and adapted for mineral potential mapping with GIS (geographic information system) has been in use since the late 1980s . the GIS based mineral potential mapping process can be divided into four steps. 1. Building a special digital database. 2. Extracting predictive evidence for a particular deposit type based on an exploration mode. 3. Calculating weights for each predictive map or evidential theme. 4. Combining the evidential theme to predict mineral potential. Modeling typically processed in three phases, specification, prediction and testing. The specification of the model begins with the definition of a set of sites where some phenomena, such as mineral deposits, earthquake epicenters. Other specification required for model include a defined study area, preparation of data for use in evidential themes, exploration of spatial association between potential evidential themes and training points, and the generalization of evidential themes. The tools provided by the Arc_ Wofe extension are particularly valuable for spatial data exploration and generalization. The spatial association of each evidence map is assessed with respect to the location of known Cu and Pb occurrences. A pair of weights ( ) are determined from the degree of overlap between the known Cu and Pb deposits and the binary evidence map. The studentized C was useful for choosing the cutoff distance because it serves the measure of the certainty and uncertainty of the contras and the variation in the contrast for cumulative distances from the outline of the host rocks with respect of the known Cu and Pb occurrences. Calculating the weights of evidence of favorable rocks Favorable host rocks are buffered in order to define the potentially favorable lithologies that may have been covered by unfavorable rocks. The mineral occurrence point map is rasterized and crossed with a multi-class raster distance map of favorable host lithology to calculate the weights of evidence ( and ) for cumulative distance away from the favorable host lithology. The optimum cutoff distance is selected at 1400 meter away from the outline of a host lithology on the basis that the Studentized value of was estimated at this distance.  Host Rock map with Mineral Deposit. The contrast at this distance indicates a positive and strong spatial correlation between the mineral occurrences and host rocks domain with a 1400 m buffer zone and 10 (out of 19) occurrences. Figure.2 shows the map of host rocks domain with mineral occurrences. Table 1. shows weights of this favorable domain used for the final model are: W+ = 0.2533 if in a favorable host rock domain, W-= 1.3547 if not in a favorable host rock domain.  Calculating the weights of evidence of favorable structural evidence The faults which have spatial association with known mineralization in the area are successful representation of lineament features in the shaded relief depending on the choice of illumination azimuth, preferably perpendicular to the linear features.The resulting structural domain was rasterized and buffered at distances of 100 m and crossed with the raster mineral occurrence point map to estimate weights of evidence of the domain. The optimum buffer that resulted in the maximum Studentized value of was defined at 1100m. The resulting buffered structural domain covers the area and 8 out of 19 occurrences are present in this domain (See figure 4.2). The weights of evidence analysis reveals a strong correlation between the structural domain and mineral occurrences. Table 2 shows the weights which were used in the final model are: = 0.346 if in a favorable structural domain, =-0751 if not in a favorable structural domain. Table 2 Weights of Evidence Analysis of favorable Structural Domain   structural Map with mineral deposit point. Predictive map of studied area The predictive exploration model was generated by summing up the weights of evidence of three binary maps representing the epithermal Cu and Pb, recognition criteria in the Lanping basin. The various overlap combinations of the binary maps resulted in the highest cumulative weights in the area where all of the recognition criteria exist. The lowest weights are located in the areas with a scarcity of favorable conditions. The predictive weights range from -1.87(Min) to 0.432(Max) .This model assumes conditional independence of the input recognition criteria binary maps . Fig. 3 shows the predictive map of the area, the resulting predictive map was classified into three categories high, moderate and unfavorable.  potential Predictive Map of Cu and Pb Conclusions Conclusions The processing technique of the Arc-Wofe extension within GIS Arc view software are particularly suited for building the database, modeling the spatial correlations between geological features and known Cu and Pb occurrence, map, calculation and displaying the result. The predictive map highlight some unrecognized favorable area. Host rocks and deep faults have the strongest spatial associations with the known Cu and Pb, deposits. The predicted high-favorability zones do not show strong affinity with lineaments it has relatively weak spatial association with the Cu, Pb, deposits, which indicate these structures played less important roles in localizing mineralization on a regional-scale as compared to crustal scale faults in the study area. All the spatial associations are significantly positive. The predictive models generated five new target areas without known Cu, Pb, Zn deposits, especially in the northwest, southwest, southeast and northeast parts of the study area. Large body of these new target areas has high and moderate potential zones. Reference
Page 2 of 2
| Previous | |