Geophysical data acquisition, processing and interpretation - Geophysics for the Mineral Exploration Geoscientist - page 61

modelled so it cannot be properly removed but, neverthe-
less, the requirement for removal remains. There exists an
extensive literature on regional removal, often referred to
as

a)

Gravity
(gu)

50

, and various techniques
are used. Hearst and Morris ( 2001 ) show how subjective
regional removal can give rise to con icting interpretations
of gravity data from the well-known Sudbury structure in
Ontario, Canada.
The two most common approaches to removing the
regional response involve either using a smooth mathemat-
ical function to approximate the regional or filtering the
data to remove the longer-wavelength variations (see
Frequency/wavelength in Section 2.7.4.4 ) . Neither approach
is universally applicable, notwithstanding the fact that
determining whether a particular approach is

' regional - residual separation '

40

Observed

30

20

Regional

10

0

20

Residual

10

0

0 200

Gossan

Metres

is
a purely subjective decision. It is often the case that the
process needs to be repeated with different interpolation/

' correct '

Mineralisation

Quartoze sedimentary rocks
Graphitic, chloritic sedimentary rocks

residual response is
obtained. Ultimately, the task for the interpreter then is to
find a residual response, from an in nite number of possi-
bilities, which is the most geologically plausible. The target
anomaly will inevitably be distorted in some way.
We describe here some aspects of regional removal,
emphasising the practicalities of the process. At this stage
it is worth saying that avoiding the problem altogether
may, in some situations, be the best approach. Sometimes
it may be better to account for all scales of variation in the
observed data, even if this necessitates making gross
assumptions about the deep geology of the area, rather
than remove some arbitrary regional component. At least,
in this case, the assumptions about the local geology are
known to the interpreter. On the other hand, not to
remove some form of the regional variation can severely
limit the interpretation of important local responses.

filter parameters until an ' acceptable '

b)

TMI
(nT)

40

Observed

30

20

Regional

10

0

-10

20

Residual

10

0

Surface

0 200

Marcasite
mineralisation

Metres

c)

SP (mV) SP (mV)

0 200 400 600 800 0 200 400 600

Ore
intersection

2.9.2.1 Interpolation methods
Surface and curve fitting techniques involve interpolating

the broader regional response into areas dominated by the

Residual

Figure 2.37 Examples of regional removal. (a) Gravity data and
geology from the Murray Brook massive Cu - Pb - Zn sulphide
deposit, New Brunswick, Canada. Redrawn, with permission, from
Thomas ( 1999 ). (b) Magnetic data across a Mississippi Valley-type
Pb - Zn deposit, Pine Point area, Northwest Territories, Canada.
Redrawn, with permission, from Lajoie and Klein ( 1979 ) . (c)
Downhole self-potential data from a drillhole intersecting the Joma
pyrite deposit, Trøndelag, Norway. Redrawn, with permission, from
Skianis and Papadopoulos ( 1993 ).

Observed

25

Regional

0

Metres