Prediction of Occurrence of Discrete Events - Geomathematics: Theoretical Foundations, Applications and Future Developments - page 146

Geochemical Signature for Au
mapid : hmn4

Legend

less favourable basins
more favourable basins
no data

20 km

Weight for feature present w+ = 1.0047

Weight for feature absent w− = 1.0047

No data for feature

w = 0

Fig. 5.4 Ternary pattern for geochemical signature (binary pattern for presence or absence of
favorable geochemistry plus binary pattern unknown) (Source: Bonham-Carter et al. 1988 ,Fig.2a)

If there is no spatial correlation between deposit points and map pattern:

p d p b p d

p b ðÞ ðÞ

p b p d p b p d

P ¼

where p ( b ) is proportion of study area occupied by map layer B , p ( d ) is total number
of deposits divided by total area, and with similar explanations for the other pro-
portions. A chi-square test for goodness of fit can be applied to test the hypothesis
that there is no spatial correlation between deposit points and map pattern. It is
equivalent to using the z -test to be described in the next paragraph. The variance of
the contrast C W + W is: ¼

1
nbd

1
nbd þ

1
n bd þ

1
n b d :

2 C

˃ ðÞ ¼

ðÞ þ

The subject of asymptotic estimates of variances will be discussed in more detail
in Sect. 5.1.4 .
Table 5.1 shows weights and contacts with their standard deviations for nine
binary or ternary map layers used in a subsequent application of WofE to Meguma
Terrain gold deposits (Bonham-Carter et al. 1990 ). Halifax Formation and Devo-
nian granite were not used in this WofE application because almost no occurrences