Talc, the softest mineral in the world, is a hydrated magnesium sheet silicate, formed by transformation of existing rocks under hydrothermal activity. Properties of talc—including its resis- tance to heat, electricity, and acids, its platyness and its insolubility—make it a very important industrial mineral. The open-cut mining of talc can therefore be an attractive and profitable industry, pro- vided exploration and production costs are not excessive. WMC Resources first became involved in talc exploration at Three Springs, Western Australia, in 1960. There, talc mineralization is hosted in the Proterozoic Moora Group dolomite, which unconformably overlies Archaean basement and is intruded by numerous dolerite dykes and sills. The location of the talc mineralization is closely related to the structural setting. In the late 1980s, WMC investigated the physical proper- ties of the talc mineralization and its dolerite/dolomite contacts, and con- cluded that the acoustic contrasts between talc and the host rock should be sufficient to produce a reflection. However, it was recognized that there may be difficulties associated with the seismic imaging of the talc mineraliza- tion because the talc bodies are shallow, have a complex geometry, and are inter- sected by many faults. Such conditions could significantly scatter the seismic energy, resulting in a deteriorated image of the subsurface. In 1993, following the results of the wireline logging, a 700-m seismic line was recorded over a known talc deposit, in order to test the applicability of seis- mic methods for talc exploration at Three Springs. A Betsy gun system was the energy source. Shots were fired in 1-m holes, 2.4 m apart; receiver stations were 1.2 m apart. A 24-channel Bison record- ing system was used to stack repeated shots. Average fold was six. The line was over a shallow talc mineralization in the form of domes. The near-surface geology consists mainly of crusty, highly hetero- geneous lateritic material. Initial processing of the data in 1993 produced a poor result. Minor reflec- tions were observed in the field records, but the resulting seismic section was uninterpretable and the talc mineraliza- tion could not be imaged. In 2000, the data was reprocessed at Curtin University. The major prestack process- ing effort was spent on obtaining a pre- cise statics solution required for data with high-frequency content. After this was achieved, good quality final time and depth images were produced. The results of this study suggested that seis- mic methods could be a valuable explo- ration tool for talc exploration in this area. Data processing. Inspection of the shot records revealed that 1-2 traces nearest to the shot have to be edited due to amplitude clipping. This resulted in fur- ther lowering nominal fold from 6 to around 5. However, weak reflections in several field shot records (Figure 1) encouraged further reprocessing. A two- layer model was used to compute refrac- tion statics. Subsequent processing steps involved signal to noise ratio (S/N) improvement through amplitude com- pensation (including Q-compensation), f-k filtering, and deconvolution (Figure 2). Figure 3a shows a brute stack, after application of refraction statics and S/N ratio improvement. Figure 3b is a final time stack with surface consistent resid- SEPTEMBER 2002 THE LEADING EDGE 923 Shallow high-resolution seismic imaging of the Three Springs talc mine, Western Australia MILOVAN UROSEVIC and BRIAN EVANS, Curtin University, Perth, Western Australia LISA VELLA, WMC Resources, Perth, Western Australia Figure 1. Selected raw field shot records. A poor signal-to-noise ratio is apparent, but weak reflections are present in the data. b) a) c) Figure 2. Signal enhancement: (a) Band-pass filtered (80-450 Hz), (b) f-k filter and band-pass filter applied, and (c) Q-compensation, f-k filter, predictive deconvolution and 80-450 Hz band- pass filter applied to the data. A time-variable Q-compensation was applied using at time t=0, Q=80 and for t=100 ms, Q=120, and thereafter.
Transcript
Talc, the softest mineral in the world,is a hydrated magnesium sheet silicate,formed by transformation of existingrocks under hydrothermal activity.Properties of talc—including its resis-tance to heat, electricity, and acids, itsplatyness and its insolubility—make ita very important industrial mineral. Theopen-cut mining of talc can therefore bean attractive and profitable industry, pro-vided exploration and production costsare not excessive.
WMC Resources first becameinvolved in talc exploration at ThreeSprings, Western Australia, in 1960.There, talc mineralization is hosted inthe Proterozoic Moora Group dolomite,which unconformably overlies Archaeanbasement and is intruded by numerousdolerite dykes and sills. The location ofthe talc mineralization is closely relatedto the structural setting. In the late 1980s,WMC investigated the physical proper-ties of the talc mineralization and itsdolerite/dolomite contacts, and con-cluded that the acoustic contrastsbetween talc and the host rock shouldbe sufficient to produce a reflection.However, it was recognized that theremay be difficulties associated with theseismic imaging of the talc mineraliza-tion because the talc bodies are shallow,have a complex geometry, and are inter-sected by many faults. Such conditionscould significantly scatter the seismicenergy, resulting in a deteriorated imageof the subsurface.
In 1993, following the results of thewireline logging, a 700-m seismic linewas recorded over a known talc deposit,in order to test the applicability of seis-mic methods for talc exploration at ThreeSprings. A Betsy gun system was theenergy source. Shots were fired in 1-mholes, 2.4 m apart; receiver stations were1.2 m apart. A 24-channel Bison record-ing system was used to stack repeatedshots. Average fold was six. The line wasover a shallow talc mineralization in theform of domes. The near-surface geologyconsists mainly of crusty, highly hetero-geneous lateritic material.
Initial processing of the data in 1993produced a poor result. Minor reflec-tions were observed in the field records,but the resulting seismic section wasuninterpretable and the talc mineraliza-tion could not be imaged. In 2000, the
data was reprocessed at CurtinUniversity. The major prestack process-ing effort was spent on obtaining a pre-cise statics solution required for datawith high-frequency content. After thiswas achieved, good quality final timeand depth images were produced. Theresults of this study suggested that seis-mic methods could be a valuable explo-ration tool for talc exploration in thisarea.
Data processing. Inspection of the shotrecords revealed that 1-2 traces nearestto the shot have to be edited due to
amplitude clipping. This resulted in fur-ther lowering nominal fold from 6 toaround 5. However, weak reflections inseveral field shot records (Figure 1)encouraged further reprocessing. Atwo-layer model was used to compute refrac-tion statics. Subsequent processing stepsinvolved signal to noise ratio (S/N)improvement through amplitude com-pensation (including Q-compensation),f-k filtering, and deconvolution (Figure2). Figure 3a shows a brute stack, afterapplication of refraction statics and S/Nratio improvement. Figure 3b is a finaltime stack with surface consistent resid-
SEPTEMBER 2002 THE LEADING EDGE 923
Shallow high-resolution seismic imaging of the Three Springs talcmine, Western AustraliaMILOVAN UROSEVIC and BRIAN EVANS, Curtin University, Perth, Western AustraliaLISA VELLA, WMC Resources, Perth, Western Australia
Figure 1. Selected raw field shot records. A poor signal-to-noise ratio is apparent, but weakreflections are present in the data.
b)a) c)
Figure 2. Signal enhancement: (a) Band-pass filtered (80-450 Hz), (b) f-k filter and band-passfilter applied, and (c) Q-compensation, f-k filter, predictive deconvolution and 80-450 Hz band-pass filter applied to the data. A time-variable Q-compensation was applied using at time t=0,Q=80 and for t=100 ms, Q=120, and thereafter.
Poststack processing started with f-x deconvolution to further improve S/Nratio while preserving signal integrity(Figure 4a). To generate the velocity fieldfor migration, we carried out a prelimi-nary interpretation and then assignedvelocities measured in the laboratory toselected layers. Two different migrationtechniques were tested: phase shift andfinite difference. In this case, these twotechniques produced practically equiv-alent results. Figure 4b is the final time-migrated image. A second phase ofinterpretation was carried out on themigrated seismic data and included log
data and local geologic information.Velocities measured on core sampleswere subsequently assigned to thisrefined geologic model. The derivedvelocity field, after smoothing, was usedfor time-to-depth conversion. Figure 5shows an expanded portion of the finalseismic line covering the talc mineral-ization. The time-migrated section isFigure 5a and the equivalent depth sec-tion is Figure 5b.
Interpretation of results. Final inter-pretation of the data was based on phys-ical property results, geologic knowledge
at the mine site, and seismic data. Thevicinity of the open-cut pit was certainlyhelpful in this process (Figure 6). Thephotograph shows two talc domes inthe open-cut pit, after mining. The roadways (benches) around the pit and theexcavation truck give a sense of relativesize. The seismic line was recorded 100m away from the pit, at a slight angle tothe open cut, but a comparison shows aremarkable similarity. It was certainlyvery encouraging to see good agreementbetween the seismic data and the maingeologic features exposed in the opencut. Clearly, the seismic method worked
924 THE LEADING EDGE SEPTEMBER 2002
Figure 3. (a) Brute stack with refraction statics and NMO correction applied and (b) final stackwith residual statics applied.
Figure 4. (a) Final stack and (b) phase shift migrated using a single velocity function.
b)
a)
b)
a)
well.The 700-m seismic line crossed two
known talc mineralizations of domeshape. Two parts of the final depth sec-tion covering the eastern and westernextensions of the talc mineralization areenlarged in Figures 7a and 7b, respec-tively. The main faults and the bound-aries of the talc domes have beeninterpreted with the help of borehole
data and geologic cross-sections.Reflection amplitudes were used in theinterpretation process according to theimpedance contrast, measured on coresamples, between massive talc/dolomiteand talc/dolerite rocks.
Conclusions. We have shown that talcmineralization can potentially be imagedwith high-resolution seismic methods,
SEPTEMBER 2002 THE LEADING EDGE 925
Figure 5. An enlargement surrounding the domed feature is shown (a) as a migrated time sec-tion and (b) as a depth section after time-to-depth conversion was performed. The possible extentof mineralization is marked in (a).
b)
a)
Figure 6. The seismic profile recorded about 100 m from and parallel to the open-cut pit, closelyresembles observed local geological features.
providing that a reasonable effort, time,and care are put into data processing.One of the most important steps in high-resolution data processing is accuratecomputation of refraction and residualstatic corrections. This step was also cru-cial for successful reprocessing of the
Three Springs seismic data. Interactivevelocity analysis was carried out beforeand after residual static computations.The estimate of the best stacking veloc-ities was not reliable because of the veryshort spread used during data acquisi-tion (28.8 m). To overcome these prob-
lems, at least partially, we adjusted stack-ing and migration velocities accordingto the laboratory measurements of ultra-sonic velocities on core samples.
The quality of the final seismic sec-tion enabled mapping the extent of talcmineralization. A subsequent surveywas commissioned, using survey para-meters adjusted in accordance with theanalysis of 1993 data set. These resultswill be reported in the future. TLE
Acknowledgments: We thank Luzenac Australia(the current owner of Three Springs talc mine)for its kind permission to publish this paper.Thanks are also due to WMC Resources Ltd., par-ticularly Exploration and Group Technology, forits support of this work. Finally, we acknowledgethe work of current and past employees of ThreeSprings talc mine, which contributed to this study.Milovan Urosevic and Brian Evans are supportedby the Improved Seismic Imaging Program of theAustralian Petroleum Cooperative ResearchCentre and the Curtin Reservoir GeophysicsConsortium. We thank Landmark Graphics for itsUniversity software grant.
Milovan Urosevic, a senior research fellow atCurtin University, has a BS from BelgradeUniversity, an MS from the University ofHouston, and a PhD from Curtin University. Hisinterests include high-resolution seismic methodsand multicomponent seismology.
Brian Evans, an associate professor at CurtinUniversity, has an MS from WAIT and PhD fromCurtin University. He is the author of the SEGMonograph Seismic Data Acquisition inExploration. His research interests include frac-ture imaging and time lapse developments.
Lisa Vella, a senior geophysicist with WMCResources, has a BSc (honors) from the Universityof Sydney. Her research interests include improveddetection of ore bodies, using integrated geophys-ical methods.
926 THE LEADING EDGE SEPTEMBER 2002
Figure 7. Seismic interpretation of talc mineralization: (a) talc-east and (b) talc-west.
