See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4020525

Applications of GPR for surface mining

Conference Paper · June 2003

DOI: 10.1109/AGPR.2003.1207302 · Source: IEEE Xplore

CITATIONS READS

17 996

2 authors:

Jan Francke Richard Yelf

International Groundradar Consulting Inc University of Liverpool University College London
31 PUBLICATIONS   206 CITATIONS    14 PUBLICATIONS   249 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mining Applications View project

Aeolian GPR Oman View project

All content following this page was uploaded by Jan Francke on 10 April 2020.

The user has requested enhancement of the downloaded file.

 APPLICATIONS OF GPR FOR SURFACE MINING

Jan C. Francké
Associated Mining Consultants Ltd.
4389 Osler Street, Vancouver, B.C., Canada
[email protected]

ABSTRACT PLACER GOLD AND DIAMOND MINING

Over the past 25 years, a myriad of applications for GPR Placer deposits occur in fluvial environments where
technology have evolved ranging from archaeology to precious metals or gemstones have been eroded or
subsurface mapping on Mars. However, it is only recently weathered from primary veins or pipes, transported, and re-
that GPR has been applied on a production basis for surface deposited amongst sands and gravels within active water
mining. courses or paleochannels. Although the scale of placer
mining ranges from artisanal gold panners to massive
Due to the generally resistive nature of the media, GPR is dredging operations, the importance of a priori knowledge
an obvious candidate for alluvial gold and diamond resource of the complex subsurface geology of the paleofluvial
exploration in aggregate-filled paleochannels. By environment is critical to resource exploration, project
correlating the radar reflectors to known geological features economics, as well as mine engineering. Of primary
detected by boreholes or trenches, GPR has been used on a importance is information related to bedrock topography
large scale for preliminary exploration, as well as on a local and gain-size distribution within the fluvial facies. Mine
scale for the three-dimensional reconstruction of complex geologists require information on the bedrock topography,
braided paleochannel systems. Although GPR has been as the highest concentrations of precious metals or
attempted historically at a variety of placer sites, newly- gemstones are generally found in coarse-grained gravels in
developed visualization techniques have enabled greater bedrock troughs such as paleothalwegs. The financial
exploitation of the richness GPR data affords. Data are viability of a placer mine often depends on the stripping
presented which illustrate the utility of GPR in the three- ratio, which is the volume of overburden material required
dimensional mapping of various fluvial paleofeatures, to to be removed to access the high-grade gravels. Data
depths frequently exceeding 40 m. related to the thickness of overlying sands as compared to
the thickness of coarse basal gravels may determine that a
A demanding application for GPR is that presented by project is not economic. In addition, open pit placer
tropical weathering environments. With the dramatic operations generally need to be de-watered during mining,
growth of interest in lateritic and bauxitic resource making data related to bedrock topography and grain-size
exploration, fuelled by increasing demand and new distribution important for mining engineers.
processing technologies, the need for accurate resource
delineation and careful mine planning becomes paramount. Traditionally, boreholes or trenches are placed at set
The traditional use of borehole grids to calculate ore intervals across a known or suspected paleochannel in order
reserves has proven to be neither sufficiently accurate nor to map bedrock topography and fluvial lithology. However,
cost-effective at many sites due to the complexity of these techniques are costly and often produce results which
tropical weathering profiles. Although conventional are unrepresentative of the true complexity of a deposit. In
wisdom dictates that radar surveys are usually unsuccessful order to mitigate these drawbacks of traditional exploration,
in regions with a high clay fraction, GPR has emerged as geophysical techniques are increasingly employed to
the most suitable geophysical tool to complement borehole provide data continuity between boreholes or test pits.
grids in addressing project geology, resource delineation, Techniques such as refraction seismic, electrical imaging or
and mine planning issues in the high clay fraction soil found electromagnetic surveying suffer from either inadequate
in most tropical weathering environments. resolution or high costs. GPR offers an ideal solution to the
problem of paleochannel mapping using sparse boreholes or suppression, before being tessellated into squares of
test pits [1]. approximate dimensions x = 5 m, y = 2 m. Experimentation
with finer tessellations has resulted in an escalation of
Figure 1 shows GPR data acquired at a placer gold deposit computation time with minimal increase in the final
in South America which have been interpreted based on two interpretability of results. Each square is analysed based on
diverse processing regimes. The data were acquired using its texture and linearity of radar reflections, and assigned a
50 MHz unshielded bi-static antennas. The data shown on colour code. The resultant section may be clipped to
the upper profile were processed using standard steps, highlight only the radar “texture” variations in the
including noise suppression and gaining. Clearly evident on paleochannel using the bedrock profile interpreted on the
this section are various radar reflectors as detected by the 50 original radar data. Applications of this method of
MHz antennas, whose wavelength of radar energy suggests processing have resulted in variable results with its success
a minimum thickness of radar facies of approximately 2 m. apparently dependant on the sharpness of lithological
Along with such fluvial features as channel fills, cross-beds variations, as well as the degree of water saturation. Radar
and longitudinal bar foresets, the bedrock topography is wave velocities are diminished in highly saturated
well imaged, showing the paleothalweg to be situated at aggregates, thereby shortening wavelengths and increasing
approximately position 130E. profile resolution. The profile presented illustrates an ideal
case where the differentiation of sands from basal gravels
The lower profile depicts data which have been processed and cobbles was possible. This type of processing is most
using a newly-developed semi-automated segmentation and suitable for diamond or gold exploration, where the highest
classification routine designed for GPR data acquired in grade coarse gravels may be differentiated by surrounding
placer environments. Using algorithms commonly data based on their radar signature.
employed in GIS applications for texture and linearity
analysis of remote sensing imagery, the purpose of these The relatively low cost of large-scale GPR acquisition,
routines is to provide a consistent and autonomous method coupled with its ability to provide horizon continuity
of classifying grain-size distributions within paleochannels. between sparse boreholes allows the geo-referenced three-
Fine-grained sands and silts tend to be characterised by dimensional reconstruction of individual paleofluvial facies
coherent linear radar reflections, whereas the coarse-grained for deposit modelling. At most placer deposits, the high-
basal gravels and cobbles generally exhibit neither yielding coarse gravel bodies must be mapped in three
coherence nor linearity. Radar data initially undergo dimensions, by tracing buried paleochannels with little or
standard processing steps such as gaining and noise no surficial expression. The use of extensive GPR profiling

Figure 1 : Processed and interpreted GPR data shown in upper profile. Lower profile shows classified data after texture
and linearity analysis.
to reconstruct the paystreak facies offers dramatic cost standard Internet browser, project geologists can be
saving over drilling alone. The example shown in Figure 2 provided with valuable a priori insight into a deposit.
illustrates the capability of sufficiently dense GPR data
acquired along transects aligned perpendicular with the Although it is evident that the use of test pits or boreholes to
paleochannel. The target in this example was a clearly- correlate GPR data acquired in placer environments is
definable coarse gravel and cobble basal facies generally critical, of similar importance is an understanding of the
situated within the paleothalweg. A series of 100 m-spaced inherent differences in material properties identified by each
GPR profiles were processed using the aforementioned exploration method. Typically, the definitions of lithology
texture analysis routines. To further automate and ensure identified by boreholes are the results of a combination of
consistency of the interpretations, an image segmentation mining economic criteria and geological concepts.
routine was employed whereby data exhibiting similar Borehole lithologies are defined using methods varying
values were grouped together by pairwise data clustering from simple visual inspection to detailed geochemical cut-
with topological consideration. This method of offs. Converse to this scheme of horizon definition, GPR
segmentation is commonly used in medical imaging relies on significantly different rock properties to produce
systems to automatically classify various image attributes. coherent reflections or changes in the frequency content of
Upon segmenting the sharp contrast between the target the data, such as grain-size variations, changes in
facies from the overlying sands and silts, the shape of the conductivity, or even the presence of magnetite commonly
gravel body was extruded in three dimensions using non- associated with placer gold. With the understanding that
uniform rational b-spline (NURBS) modelling based on the strict correlations may not always be possible between
GPR slices. By geo-referencing such gravel bodies and geological horizons and radar facies, the geophysicist must
incorporating boreholes, topography and air photos in an reconcile these disparate data types to arrive at reasonable
interactive three dimensional model constructed in virtual and representative interpretation. The increase in geological
reality modelling language (VRML) and viewable on any insight garnered by the ability of GPR to provide horizon

Figure 2 : VRML model of basal gravels in a braided paleochannel system, showing segmented GPR data slices, extruded
interpretation of gravel body, boreholes and surface features.
continuity between boreholes is significant, gained at a making them effectively transparent to radar energy.
fraction of the costs associated with a dense drilling grid.
Laterite resource modelling is a process of systematically
compiling all known information regarding a deposit in
LATERITE AND BAUXITE MINING order to define the volume of ore contained and its intrinsic
economic value. Traditionally, this process has been highly
The suitability of GPR to mine exploration and speculative in nature, as the variability of layer thicknesses
development in tropical weathering environments has been between even closely-spaced boreholes can be extreme.
discussed in previous publications [2] – [5]. In both lateritic The issue of horizon continuity is of great importance in
and bauxitic deposits, the need for accurate resource mine modelling and resource definition. In laterite deposits,
definition prior to mining is paramount. A major economic continuity of layer thickness is difficult to prove with
factor controlling the development of these projects is the drilling alone. Variography studies using decreasing drill
significant capital cost due to their size. Developing lateritic spacing rarely converge on a single reliable estimate, as
or bauxitic deposits often requires large investment in even the finest drill spacing is usually greater than the
infrastructure due to remote tropical locations and the spatial frequency of the variations in profile thickness. The
processes’ heavy reliance on large-scale materials handling. limitations associated with this inferential approach to
Exploration costs such as the saturation drilling needed to resource estimation can be mitigated with the use of GPR.
define a laterite nickel resource entail significant expense Limonite is commonly homogenous with relatively constant
and may prevent a project from progressing. Mapping the grade and density, allowing the use of a relatively fewer of
extent of the mineralised layers using high-resolution strategically-placed boreholes combined with detailed radar
geophysics changes often significantly reduces to need for imaging along drill lines.
saturation drilling, thereby lowering exploration costs.
Figure 3 illustrates typical post-processed GPR data
Although a number of geophysical methods such as acquired in a lateritic environment. The upper profile was
refraction and reflection seismic, SASW and electrical processed using a typical regime of gains and band-pass
imaging have been applied with varying success to tropical
weathering environments [4], GPR has emerged as the
method of choice for both exploration and development
stages of mining projects. Profile depths in excess of 60 m
have been recorded with low-frequency antennas and
confirmed by boreholes. The excellent penetrability of
radar energy in these environments can be explained by the
additive effects of a strong target contrast and lack of
attenuation.

The primary objective of GPR surveying in lateritic or

bauxitic environments is to image the base of the economic
weathering zone, be it either parent bedrock or a ferricrete
base. In the case of laterites, the reflections from the base
of the weathered zone are caused by a dramatic variation in
bound water saturation. Both limonite and saprolite have a
very high bound water content, whereas crystalline bedrock
has a low water content due to its lack of porosity. Thus, the
interface between the weathered material and rock produces
a strong radar reflection. In the case of bauxitic
environments, the pisolitic bauxite is highly resistive and
contrasts strongly with the basal ferricrete.

Conventional wisdom as developed in temperate regions

dictates that radar surveys are usually unsuccessful in area
with a high clay fraction, and laterites are, by definition,
highly weathered clays. Surveys at dozens of these sites
around the world have demonstrated that the weathering
processes which formed the laterite deposit has leached the
saprolitic clays of their conductive mineralogy, thereby Figure 3 : Amplitude (top) and frequency (bottom)
processing of GPR data acquired at a laterite deposit.
filtering. Interpretable is the limonitic to rocky saprolitic REFERENCES
transition zone, demarked by the solid line. Due to the
homogenous and highly-weathered nature of the overlying [1] Yelf, R., “Application of GPR to large scale production
limonite, the rocky saprolite layer provides a strong contrast mining of alluvial gravels, bauxite and laterite
for radar reflections. With boreholes spaced every 100 m, deposits”, Proceedings of the Tenth Australian Society
the large rocky saprolite pinnacle would not have been of Exploration Geophysicists Conference, Sydney,
imaged without GPR. Previously hidden features such as Australia, February 1997.
this are of critical importance to geologists in calculating
resource estimates, as well as to mine engineers when [2] Francké, J., and Nobes, D., “A preliminary evaluation
planning site trafficability. of ground penetrating radar for nickel laterite
exploration,” Proceedings of the Eighth International
The lower profile shows the same profile segment processed Conference on Ground Penetrating Radar, Gold Coast,
using the Hilbert transform in the frequency domain. Australia, pp. 7-12, May. 2000.
Although the rocky saprolite structure shown in the
amplitude data is indiscernible, the ultimate base of [3] Francké, J., “The Application of Geophysics to Nickel
weathering is clearly shown as a transition from high- Laterite Resource Evaluation.” Master’s thesis,
amplitude low-frequency data characteristic of weathered University of Canterbury, 2001.
material to low-amplitude high-frequency data
characteristic of parent bedrock. When combined, the two [4] Francké, J., and Parkinson, G., “The new role of
interpretations provide a more complete understanding of geophysics in nickel laterite exploitation and
the subsurface geology. development”, Proceedings of Mining Millennium /
PDAC Convention, Toronto, Canada, pp. 204-210,
March 2000.
CONCLUSIONS
[5] Queen, L., Parkinson, G., Francké, J., and Brown, E.,
GPR is gaining acceptance as standard practice for “The integration of ground penetrating radar in
reconnaissance alluvial exploration. The resistive soils of resource delineation for nickel laterites in Papua New
most alluvial environments yield impressive penetration Guinea,” Proceedings of the Pacific Exploration
depths and profile resolution. Faster and more Technologies Symposium, Nadi, Fiji, pp. 87-90,
representative of geology than test pits and boreholes alone, September 1998.
GPR may be utilised to model the three-dimensional extent
of gold and diamond-bearing fluvial sequences. A variety of
sediment structures including channel fills, cross-beds and
longitudinal bar foresets are commonly identified. In these
environments, penetration with the radar is generally
excellent and frequently exceeds 40 m. These data may then
be used in a mining package for the 3D reconstruction of
the paleochannel system. The data can be used to delineate
the extent of the paystreak, to estimate resource volume and
to select strategically placed drilling and exploration targets.

One of the most successful applications for GPR methods in

mineral exploration has been in the laterite and bauxite
sectors. GPR has the ability to continuously image the
details of the laterite profile. Radar images have shown that
the boundaries of the layering are surprisingly variable
between boreholes. These abrupt changes in horizon
elevations are generally undetectable by drilling at any
economic spacing. By using GPR to accurately map the
volume of the deposit, coupled with sparser drilling to
confirm the layer identification and grade, a
geoscientifically sound resource estimate may be made
quickly and economically.
View publication stats