Page 24 - Geochemical Anomaly and Mineral Prospectivity Mapping in GIS

Predictive Modeling of Mineral Exploration Targets 19

Fig. 1-5. Deposit occurrences overlaid on (or crossed with) classified Cu data. Cross-table output
showing number of deposit occurrences in every class of Cu data. The analysis indicates that >75
percentile Cu values are significant Cu anomalies.

Specific tools for certain data analysis may not be available in some GIS software
packages. In such cases, data must be exported or converted to formats supported by
other computer software packages that provide the specific tools of interest. The
examples given here and several other forms of data analysis demonstrated in the
succeeding chapters support the creation of maps of evidential features (e.g., significant
geochemical anomalies), which are eventually integrated to model mineral prospectivity.
Note that the example of data analysis via two-map overlay of Cu data and mineral
deposit occurrence data (Fig. 1-5) is already a form of data integration.

Data integration
The behaviour of indirectly observable and complex real-world system of interest,
such as a geochemical anomaly or mineralisation, is controlled by several interacting
processes. In order to predict the behaviour of such systems, it is instructive to combine
or integrate sets of data, pieces of geo-information or models representing the individual
processes involved. In a GIS, a predictive model or map is usually derived by combining
predictor maps (Fig. 1-2) via a computational function that aptly characterises the
interactions or relationships among the processes that control the behaviour of a system
of interest:

predictive mod el = ( f predictor maps ).

There are different forms of the computational function f. The choice of a computational
function depends on whether the predictive model is stochastic, empirical or hybrid
stochastic-empirical. Unlike predictive modeling of geochemical anomalies, which can
be stochastic, predictive modeling of mineral prospectivity usually makes use of