Page 256 - Geochemical Anomaly and Mineral Prospectivity Mapping in GIS

Data-Driven Modeling of Mineral Prospectivity 259

area calculations (i.e., as multiples of number of unit cells or pixels) in the application of
a raster-based or pixel-based GIS.
The application of point pattern analysis (Boots and Getis, 1988; Rowlingson and
Diggle, 1993), as shown in Fig. 8-1, can therefore be useful in deriving a preliminary set
of choices for a suitable N(•). However, the final choice of a suitable unit cell size must
also consider (a) the scale of the field geological observations used in constructing a
mineral deposit occurrence database, (b) the scales of the input maps or images of spatial
data of explanatory/predictor variables and (c) the scale of the desired output mineral
prospectivity map(s). These considerations were taken by Carranza et al. (2008b) in
data-driven modeling of prospectivity for alkalic porphyry Cu-Au deposits in British
Columbia. The current British Columbia MINFILE mineral inventory database (BCGS,
2007) contains records of prospect- to mine-camp-scale (usually larger than 1:10,000)
data for 356 locations of alkalic porphyry Cu-Au deposits. In contrast, the scale of the
geologic map is 1:250,000 (Massey et al., 2005), whereas the airborne magnetic and
gravity data were captured in 1-km and 2-km grids, respectively (Geoscience Data
Repository, 2006a, 2006b), which translate to map scales of about 1:400,000 and
1:800,000, respectively (see Hengl, 2006). Therefore, in view the different scales or
spatial resolutions of input spatial data of the explanatory/predictor variables, Carranza
et al. (2008b) used an ‘average’ unit cell size of 1 km. The use of a larger unit cell size
than indicated by the distance-probability relation can mean that more than one deposit-
type location is covered by a unit cell, and this was the case for some unit cells in the
British Columbia study. Here, however, some of the alkalic porphyry Cu-Au deposits
(e.g., Axe (Adit Zone), Axe (South Zone) and Axe (West Zone)) described in the
MINFILE database probably represent one large alkalic porphyry Cu-Au deposit so, if
that is the case, using a larger unit cell size than indicated by the distance-probability
relation is justifiable.
Whether a preliminary set of choices for a suitable N(•) indicated by the distance-
probability relation is adopted or adapted, the analysis of the rate of increase in the ratio
[N(D)] : [N(T)–N(D)] as function of equal-interval change in N(•), as illustrated in Fig.
8-3, is robust regardless of the number of deposit-type locations and the size of a study
area. Nevertheless, it is also imperative to verify if the most suitable N(•) suggested by
results of analyses depicted in Figs. 8-2 and 8-3 is reasonably consistent with the average
lateral extents (at prospect- to mine-camp-scales) of known occurrences of mineral
deposits of the type sought. Data of lateral extents of known occurrences of mineral
deposits of the type sought are, unfortunately, not available in many cases. If such is the
case, then the sorts of analyses demonstrated here, although mainly graphical, provide an
objective way of selecting the most suitable N(•) for GIS-based data-driven modeling of
mineral prospectivity. After having made a final objective choice of a suitable N(•), one
must next determine which of the known locations of mineral deposits of the type sought
are suitable in data-driven modeling of mineral prospectivity.