Gravity and magnetic methods - Geophysics for the Mineral Exploration Geoscientist - page 175

(B in Fig. 3.71c ) . It was not considered justi able to model
this part of the pro le using the model representing the
Wallaby alteration pipe, but a simple and geologically
plausible interpretation is that the northern margin is
parallel to the southern margin.
The top of the prism model, shown in Fig. 3.71b , agrees
quite well with the Euler solutions except on the northern
margin, which is to be expected since this part of the data
were not modelled. The modelling was undertaken without
reference to the known susceptibility distribution, but
these data from beneath the modelled profile are in excel-
lent agreement with the depth to the top of the source and
its dip. The absolute and relative susceptibilities are also
correctly represented, notably the lower-susceptibility core
of the alteration zone. The depth to the base of the anom-
aly source was poorly predicted; but the geometry of the
deeper parts of an anomaly source is usually not well
constrained because the observed response is dominated
by the response of the shallower parts of the source (see
Sections 3.10.1.1 ).

performed better, but this is partly due to being con-
strained by a forward modelling result which accurately
predicted the susceptibility distribution of the subsurface.

3.11.2.3 Discussion
The modelling of the Wallaby anomaly demonstrates that
it is good practice to begin by forward modelling the data,
incorporating all available information in order to under-
stand the range of possible source geometries, and then use
3D inversion to refine the model. Proceeding directly with
inverse modelling is a very high-risk strategy, especially
when it is unconstrained. In common with all magnetic
modelling, the results are more likely to be correct in terms
of source depth and extent for shallow regions of the
subsurface, but information about deeper regions is likely
to be less accurate. This also applies to source dip, which is
a critical parameter in designing a drilling programme to
test the source of an anomaly (see Section 2.11.4 ) . This
example also shows that although the Wallaby anomaly is a
relatively isolated feature in the aeromagnetic data, even
here modelling was hindered by interference from other
anomalies. 3.11.2.2 Inverse modelling

Inverse modelling was also applied to the Wallaby mag-
netic anomaly using a 3D voxel-based algorithm (see
Section 2.11.1.2 ) to explore the range of possible source
geometries. Firstly, a half-space model was used and the
inversion applied unconstrained. It was then reapplied but
constrained by the results of the 2.5D modelling, with the
inversion adjusting the body parameters only as necessary
to t the data. The computed responses are shown as
contours in Fig. 3.71d ; note the excellent match between
these and the actual data. The subsurface susceptibility
distributions produced by the modelling are shown in
Figs. 3.71e and f for the principal profile and two depth
slices, along with equivalent displays of the downhole
(observed) susceptibility data.
The inversion results from the shallower parts of the
subsurface, i.e. the 200 m depth slice and the upper part of
the cross-section, agree quite well and are a good repre-
sentation of the observed susceptibility distribution. The
spatial extent of the anomalous susceptibility is correctly
de ned, including the lower-susceptibility core, although
the absolute value of susceptibility was less well predicted
by the unconstrained inversion. At greater depths, the
correspondence between the modelled and observed distri-
butions deteriorates, especially for the unconstrained
inversion which has incorrectly predicted a steep northerly
dip for

3.11.3 Magnetic responses from an
Archaean granitoid-greenstone terrain:
Kirkland Lake area.

-

Archaean granitoid greenstone terrains contain a wide

variety of lithotypes ranging from very weakly magnetised
sedimentary and felsic igneous rocks through to ultra-
magnetic iron formations. Moreover, metamorphic grade
is generally quite low and, although brittle and ductile
deformation is ubiquitous, the overall structure is normally
not too complicated at the scale of hundreds of metres to
kilometres. Aeromagnetics is a useful aid to geological
mapping in these environments at both the regional and
prospect scale.
The most magnetic lithotypes found in greenstone belts,
in addition to iron formations, are ultrama c rocks such as
komatiites and some types of granitoids. Metamorphic
grade is normally greenschist to amphibolite facies, so
ma c rocks tend to be weakly magnetic (see Fig. 3.48 ) .
Other usually weakly magnetised lithotypes include felsic
and intermediate igneous rocks, and clastic sedimentary
rocks. Structures such as faults and shear zones are import-
ant targets because of their association with mineral
deposits. They are often recognised as offsets and trunca-
tions to stratigraphic anomalies, and may appear as zones the source. The constrained inversion has