Gravity and magnetic methods - Geophysics for the Mineral Exploration Geoscientist - page 105

complex correction to implement as it requires the topog-
raphy and its density distribution be accurately known; but
this is usually dif cult to achieve, so it is prone to error.
Digital terrain data provide the essential heights and
shape of the topography needed for calculating the terrain
correction. The ability to mathematically describe the ter-
rain in terms of the digital data depends on the nature of
the data and its resolution. Some airborne gravity systems
are equipped with a laser scanner which maps the terrain
in an across-line swathe below the aircraft, providing ter-
rain information of sufficient resolution and accuracy for
the terrain correction. For ground surveys, data may be
true point (spot) heights taken from aerial photography or
contour maps, or may be average heights of compartments
subdividing the photography or contour map. It is import-
ant that the actual elevations of the gravity stations match
the equivalent points on the DEM as discrepancies in
heights and locations are sources of errors. The DEM
may lack suf cient resolution to adequately de ne small
topographic features and large abrupt surface irregularities,
such as the edges of steep cliffs and the bases of steep hills, in
the immediate vicinity of the gravity station, an additional
source of error for stations affected in this way. During the
gravity survey it is necessary to record details of small local
features manually. Leaman ( 1998 ) provides practical advice
regarding terrain effects close to the gravity station.
Various procedures for geometrically describing the
digital terrain, so that its volume above the datum level
can be calculated, have been implemented. The most
common approach divides the terrain into a large number
of volume elements (voxels), usually flat-top juxtaposed
prisms, extending down to the datum level. Their gravita-
tional attractions are computed and summed at each grav-
ity station ( Fig. 3.16c ) and the whole process repeated for
each gravity station. Given that topography closer to the
gravity station exerts greater in uence than more distant
terrain features, ef cient algorithms implement smaller
prisms with more realistic, i.e. topography resembling,
upper surfaces close to the station in order to minimise
errors. Distant features, having a smaller gravity effect, are
described more crudely with larger and fewer voxels, which
also signi cantly reduces computing resources. Speci c
densities can be assigned to each prism but, as with the
Bouguer density, using variable density implies significant
understanding of the density variation in the survey area
and its (distant) surrounds.
For gravity stations located near or over a mass of water,
such as a lake, the terrain correction needs to correct down

0 2

Metres

Figure 3.17 Estimated magnitude of the terrain corrections, in gu,
due to small-scale topographic features close to the gravity station.
Redrawn, with permission, from Leaman ( 1998 ) .

Bouguer density. The process is repeated for each gravity
station as they will all (or mostly) have different heights so
the Bouguer slab at each station has different thickness
( Eq. (3.17) ), and they all have a different relationship with
the topography. Alternatively, the Bouguer correction can
be ignored, and the gravitational attraction of the undulat-
ing terrain surrounding the gravity station can be computed
as a full terrain correction with respect to the datum level.
The gravitational attraction of topography depends on
the size of the topographic features, and decreases with
their increasing distance from the gravity station.
Depending on the desired accuracy of the survey and the
ruggedness of the terrain, this means that relatively small
features close to the survey station, such as culverts, storage
tanks, reservoirs, mine dumps, open-pits and rock tors, can
have a signi cant effect on the measured gravity and can be
a major source of error in high-resolution gravity work
( Fig. 3.17 ; Leaman, 1998 ) . In addition, topographic fea-
tures tens of kilometres away from the station, and even
very large mountain ranges more than a hundred kilo-
metres away, may need to be accounted for. The effects
of more distant features appear as a regional gradient in the
data so their effect could be removed with the regional field
(see Section 2.9.2 ). The terrain correction is by far the most