Geoscience Reference
In-Depth Information
pronounced when the body is at a depth of about ve or
more times its horizontal width or thickness, and is then
referred to as a thin body. When the body
geological plausibility is one of the most useful constraints
that can be applied to geophysical modelling, provided
plausibility is not confused with personal bias.
A common misconception about ambiguity in geophys-
ical modelling concerns the fact that a given set of obser-
vations may be matched by an in nite number of
subsurface physical property distributions. An in nite
number of solutions does not mean that the modelling
places no constraints on the subsurface. The skilful inter-
preter, through the experience of changing model param-
eters and observing the effects on the computed response,
can quantify the variation in the parameters that produce
an acceptable match to the data. In so doing, the modeller
should attempt to devise
is depth is
shallower, it appears wider to the survey and is then
referred to as a thick body. In this case its thickness
and physical property contrast more independently
in uence the response, so there is greater prospect of
resolving each parameter.
'
When the physical property of a body is a vector quan-
tity such as magnetism, or directionally dependent
(anisotropic), changes in the shape of the geophysical
response can occur because of changes either in the
orientation of the body or the direction of the physical
property vector/variations. This is particularly a problem
when analysing magnetic data ( Fig. 2.49c ).
￿
models that effect-
ively restrict the range of possible models, e.g. deepest and
shallowest likely source, extremes of dip etc. ( Fig. 2.49 ).
Simple-shaped models having few parameters are useful
in this respect, despite the fact that they may be hugely
simpli ed representations of the subsurface. Recent devel-
opments in inverse modelling pursue this strategy. For
example, Bosch and McGaughey ( 2001 ) describe the use
of inverse modelling to produce probability models de n-
ing the likelihood of encountering a particular lithotype at
a given point in the subsurface. In addition, the large
number of possible solutions highlights those features
having small variation between the solutions. These
'
'
end-member
'
Another source of non-uniqueness is associated with
resolution. Obtaining a single coherent property distribu-
tion as the solution model does not obviate the possibility
of the anomaly source actually being a collection of smaller
close-spaced bodies with different physical property
contrasts and depths, their properties being
to
obtain a single-source ( Fig. 2.49d ). In this case a source
comprised of en échelon veins is modelled as a vertical
sheet. Drillhole 1, designed to intercept the inferred source,
misses the actual source, but drillhole 2 would result in an
intersection.
'
averaged
'
elements are likely to be reliable elements of the
final solution model. Non-uniqueness also provides
insight into the geological possibilities of the subsurface,
particularly where little or no geological information is
available.
The final model(s) produced depends on many param-
eters associated with both the modelling and the survey
data. The choice of constraint determines which of the
many possible non-unique solutions will result from the
modelling process. Other parameters include the type of
model (1D, 2D etc.), the discretisation and cell size when
using cell-based models, the modelling algorithm, the type
of inversion algorithm, numerical accuracy, a priori geo-
logical knowledge and how it is used to control the model,
and subjective user decisions applied during the modelling.
The task for the interpreter then is to find the model(s) that
offers the most geologically plausible explanation of the
data. Given all the limitations imposed on the interpreter
and the assumptions adopted in building a model, it is a
reality of geophysical modelling that all models are a sim-
plified and imprecise depiction of the subsurface, but some
can be useful.
stable
'
2.11.4.1 Dealing with non-uniqueness
When modelling geophysical data it is important to
include all available complementary petrophysical, geo-
physical and geological information into the interpret-
ation to reduce the ambiguity of the result. In practice
this can be difficult to achieve as most geological infor-
mation is depth-limited, being mostly from the near-
surface, and drillhole information is often sparse and
shallow. Also, physical properties can vary by large
amounts over small distances.
A drillhole intersection providing a
'
tie-point
'
on a
source
s position will hugely reduce the range of possible
models. A set of petrophysical data can help limit the range
of the physical property and is preferable to just a single-
point measurement. A third constraint, which should be
used with caution, is geological expectation. Of course, the
fact that an anomaly in another part of the area was found
to be caused by, say, a flat-lying body of massive sulphide
does not mean that only bodies of this type should be used
to explain other apparently similar anomalies. Despite this,
'
 
Search WWH ::




Custom Search