Surface Reflection Seismic Method: Coal Mines Field
1 like510 views
This document summarizes a study that used seismic modeling and analysis of seismic data to interpret the stratigraphy of a shallow coal deposit in central Alberta, Canada. 1D and 2D modeling helped relate seismic reflection character to stratigraphic variations in the coal zone. Seismic lines from the study area were reprocessed and interpreted, finding that reflection character could be quantitatively related to coal zone stratigraphy. The study demonstrated that detailed seismic interpretation is possible by relating reflection seismic data to variations in density logs from wells.
Surface Reflection Seismic Method: Coal Mines Field
- 1. CANAoIANJO”RN*L OF EYPLORAT,ON GEOPHYSlCS “OLD 24, NO. 2 (DECEMBER 1988), P 124~140 APPLICATION OF THE SURFACE REFLECTION SEISMIC METHOD TO SHALLOW COAL EXPLORATION IN THE PLAINS OF ALBERTA H.V. LYATSKY’ANDD.C. LAWTON~ ABSTRKT A seismicstudywar undertakenat theHighvale-Whitewood coatfield in ce”,ralAlbertaI” deterninewhetherreflectionseis- micsurveyscanbeusedtomap~tratigraphicandstr~cmraldetails of rtmttawPlains-typematdeposits. The study waspartly basedan one-dimensionalandtwo- dimensionalnumericalseismicmodellingusingsonicanddensity wet, togsto formulatea layeredearthmodel.I-D modelling showedthatsmallv*riationsin coal-zonestratigraphymayleadto significantchangesin theseismicresponseof thecoat-bearing interval. 2-D modellingshowedthat at source-receiveroffsets whichexceedthetargetdepth,reflectioncharacterisdegradeddue todifferential~OY~OU~.Thisresultsinalossofverticalresolution. However,examinationof reprocessedfield datain thestudyarea showedthat“ear~offsettracesarecontaminatedby groundroll, forcingatrade-offbetweentheneedforhighdataqualityandthe desireformaximumverticalresolution. Two reflecrionseismiclines fromthe Whitewoodminesite werereprocessedandinterpreted.tt wasfoundthatvariationsin reflectioncharactercouldberelatedquantitativelytothestratigra- phyof thecoalzone.Possiblelong-andshort-wavelengthstrut- furaldefomlatianwasak suggested.FracturingOftilecuatzme. probablyassociatedwith ghciotectonicthrusthutting, wasfound tocausealossofcoatretlectionalongpartsof theseismiclines. The use of reflection seismic methods in coal explo- ration has developed rapidly over the past I5 years. However, interpretations have been based primarily on the presence or absence of reflections from a coal zone, or on observation of fault-related vertical offsets of these reflec- tions (e.g., Clarke, 1976; Lepper and Ruskey, 1976; Peace, 1978: Ziolkowski and Lerwill, 1979; Schlicker and BGning, 1981). Sartorelli et al. (1985, 1986) discussed high-resolution shallow reflection seismic surveying for coal mining purposes, but their interpretation, too, was largely limited to observations of simply the presence or absenceof coal-related reflections. Fry and Orange (1982). who concentrated mostly on the study of coal-seam continuity, showed that I-D seismic modelling can be used to estimate thicknesses of coal beds. Lawton (1985) related seismic reflection character to coal depositional environments and proposed that a more detailed interpretation of seismic data can be made possi- ble by modelling. A pilot study of this nature was carried out by Lawton and Bertram (1983). but their interpretation was limited due to a lack of geophysical well logs, which are required for calibration of seismic data. The purpose of this study was to make a quantitative interpretation of reflection seismic data collected over the Highvale-Whitewood shallow coal deposit in the central Alberta Plains. The location of the study area is given in Figures 1 and 2; the deposit is presently being mined using open-pit techniques. I- and 2-D seismic modelling was undertaken to relate reflection character to stratigraphic variations in the coal zone. Additional objectives were to interpret the reflection seismic data in terms of geological features in the study area and to investigate the relationship between vertical resolution and field-acquisition geometry. The data base used in this study consisted of about 100 drillhole log suites and seven reflection seismic profiles. Most of the wells are located at the Highvale site. which is better explored than the Whitewood deposit and has a more continuous stratigraphy and less structural deformation. Density as well as sonic logs were used in this study, although the quality of sonic logs is generally poor, espe- cially at Whitewood, where all seven seismic profiles are located. ManuscriptreceivedAugust3I. 1988;revisedmanuscriptreceivedOctober26,1988. ~Formerly,Depanmentof Fedogy andGeophysics.Universityof Calgary,Calgary.AlbenaT2N IN4 presently.Cepartmenrof PhysicsandAstronomy, Universityof Victoria.Victoria,BritishColumbiaVW 2Y2 ‘DepartmentofGeologyandckophyrics,Lhiverrity ofCalgary.catgary,AlbertaT2N IN4 me a”thorSwouldliketn acknowledgethef”ll”wi”g organirationrandindividualswhoprwided&?,aandfundstar,hiEEtUdy.Drillholelagsan*loCation mapsweresuppliedby TrmsAltaUtilities Corp.2ndMonrncoCansuttantsLtd.Reflectionseismicdatawereprovidedby FewPhysi-ConCo.Ltd.and TransAttaUtilitiesCorp.Discussionswith Dr.L.V Hilts improvedourunderstandingof thegeologyof thestudyareaandhelpedusinterprettheseibmi~ data.Helpful criticismwasalsoprovidedby anonymousreviewers.Financialsupportfor this projectthroughtheAlberta/CanadaEnergyResources ResearchFund,ajoint programoftheFederalandAlbertaGovernmentsandadminisreredbyAlbmaEnergyandNaturalResources.isgratefullyacknowl- edged. 124
- 2. SHALLOWREFLECTIONSEISMK COALEXPLOR*TtON 125 .?I P& RlYER / CT-*“” LGAR”*%, SCALE JY+d O-9, 5oclkm J-C-, Fig. 1. Mapof Albertashowingthe locationofthe studyarea. Fig. 2. Mapof the HighvaleandWhitewoodminesites. The stratified nature of the Highvale-Whitewood deposit and the small (generally under 3 m) thickness of individual coal seams mean that the reflection character of the coal zone results from the superposition of numerous, closely spaced and overlapping reflections. This makes resolution of individual interfaces difficult, and the interpretation of stratigraphic variations must be based on examination of the composite reflection character. Lateral character changes may be indicative of structural deformation or flu- vial channelling. GEOLOGK SETTING The Highvale-Whitewood mine sites are pan of the Late Cretaceous-Early Tertiary coal-bearing deposits in central Alberta. The Cretaceous-Tertiary stratigraphy of central Alberta has long been a subject of discussion, and a num- ber of different interpretations and nomenclatures have been proposed. The nomenclature adopted for this study is that from Gibson (19771, and the stratigraphic column is presented in Table 1. The coal deposit under study belongs to the Scollard Formation. The position of the Cretaceous- Tertiary boundary is at the base of the coal zone (Demchuk, 1987) and its equivalents elsewhere in Alberta (Sweet and Hills, 1984). This makes the considered deposit Early Paleocene in age. The coal zone consists of six major seams, as shown in PASKAPOO FM. SCOLLARD FM. BATTLE FM. & $ ii WHITEMUD FM. B w HORSESHOECANYONFM. BEARPAW FM. Table 1. Cretaceous-Tertiary stratigraphy in the study area (accordingto Gibson,1977).
- 3. 126 H.“. LYATSKYan*DC LAWTON Figure 3, with the thickest coals occurring near the top of the deposit. The stratigraphy varies slightly from Highvale to Whitewood, but reliable correlation exists for the entire coal field, and most Highvale seams have well-defined equivalents at Whitewood (Figure 3). Highvale seam I splits into two distinct seams at Whitewood (seam 1 upper and lower); Whitewood seam 2 has no equivalent at Highvale, and Highvale seam 2 corresponds to Whitewood seam 3. The coal-bearing interval is thicker at Whitewood as seams I and 3 in that area are spaced farther apart than their Highvale equivalents. The coal zone is underlain by inorganic sediments of the Scollard, Battle, Whitemud and Horseshoe Canyon Formations (Table I). Within the Scollard coal zone, the seams are separated by bentonitic sediments which have rather variable thicknesses (up to 3 m), especially at Whitewood. Overlying the coal-bearing interval are inorganic elastics of the upper Scollard Formation, containing fluvial sands, and the Paskapoo Formation. Glacial till several metres thick occurs at the surface. The total thickness of the coal-bearing interval is about 20 m, and its depth rarely exceeds 70 m in the study area. The seams nearest to the surface arc frequently disturbed by glaciotectonic deformation (Fenton, 1987) and typical structures are portrayed in Figure 4. While short-wave- length (several metres) deformation is probably beyond the limits of lateral resolution of the reflection seismic method, long-wavelength (tens of metres) features should be resolv- able. We have made a field observation that, when the coal zone is deformed, contortions are usually confined to the uppermost seams (Figure 4a). Anelastically deformable bentonitic interburden serves as a partial detachment zone, and deeper seamsare less deformed. Lawton (1985) summarised typical environments of coal deposition. He stressed particularly the differences between delta-plain coals, which are generally extensive, and fluvial coals, which are formed in restricted swamps and river flood plains and are often discontinuous. Meandering channels may periodically cut through coal- accumulating swamps, causing additional local discontinu- ides. The coals of the study area are probably of a mixed, fluvio-deltaic origin. Because of the continuity of the strata and the good data base of drillholes, the Highvale- Whitewood area is an excellent test site for seismic studies of shallow, layered deposits. ONE-DIMENSIONALNUMERKAL MODELLINGAT Htwtv,ux Examination of drillhole logs revealed little variation in coal-seam thickness and separation over much of the Highvale site. Therefore, a typical well (C-HV-X3-03), whose logs were considered reliable and whose location is shown in Figure 2, was chosen for I-D modelling. The sonic and density logs from this well are presented in Figure 5, with individual coal seams identified. The top of the coal zone was encountered at a depth of 63 m, and seams I to 6 were penetrated. Transit-time variations can- not be easily correlated with lithologic changes, but density variations clearly show the location of coal seams. Therefore, as Lyatsky and Lawton (1989) have shown, reflection character of the seismic data is determined almost entirely by variations of density with depth, and velocity can be assumed constant for the entire coal zone. Figure 6 shows a synthetic seismogram generated from the logs in Figure 5. Normal polarity defines a reflection peak corresponding to an increase in acoustic impedance, and this convention will be used in all subsequent models. WHITEWOOD SEM.41 SEAM 2 HIGHVALE .15 Ill .lO Ill .sm Fig. 3. Stratigraphiccorrelationof coalseamsbetweenthe HighvaleandWhitewoodminesites(courtesyMnnencoLtd)
- 4. Fig. 4. Plaks illustrating giaciotectanic deformation of *earn 3 at Whitewood~ The thickness of the *earn shown is about 3 m. Plate 4a ShOWS shan-wavelengthdeformationwith faulting and folding; Plate4b ~Iiusfratesa Iong~wavelength,faultWwnded structure,with the fault location indicatedbythe arrow. The LWOlogs Sromwell C-HV-X3-03 and Ihc rc~ulumt time reflrctivit) sequewx are idso ~~OWIIin Figure 6. Sincc the highest t’requency of the seismic data recorded in the study area was ahout 130 Hr. a Kicker wa~elet with npproxi- nmtrly thi:; crnlral frequency was chosen for constructing the synthetic seismogram. The USCof ii zcm-phase Rickct wwelet is .justilicd hy the absence of a Ggnificanr depen- dence of retlectioll characw on wa1vcIc1phase in rhc study arca (Lyntsky, 19Xx). l’hc scismqrram was produced with a s~mplc imcrval OFl/X ms. If is ;rppxrnt l’rw~~the rcflec- tivity sequence in Figure 6 that high acoustic impedance conUxl within the coal-hearing infewal occur at Ihe COII- tilcts hctwccn coilI and interhurdcn. Neucrthcles. as Lyatsky (198X) has shown. intrahcd multiples generated within Ihe coal /one are ncgligihle. and they do not appeal to distort rhc CO&~IIIC rctlcction character.
- 5. I?.X (a) 2 z a w n II.-.,ShTSKY i,/ldD.C.I.AM;‘I‘ON TRANSIT TIME @s/H) DENSITY (Mg/m? 2 (b) 60 : __ 1.0 65 7c I! 7: - - 65 1 TO- SEAM2 A.$EAM 3 )P SEAM4 SEAM! Sk SEAM6 75 I / Fig. 5. LogsfromwellGHV~83-03:(a)sonic log:(b) densitylog. a -0.1 El.1 NOR --T- TYq--E , 0 _ 00 1-- ‘OLkR II i- 2L .E ? d E . 0 i L iii z 3 >:- i .... .... $ ............... 20 _..... SONIC LOG DENSITY LOG REFLECTIVITY SYNTHETIC (m/s) (Mg/m3) SEQUENCE SEISMOGRAM Fig. 6. Integratedlogs.time-reflectivitysequenceandsyntheticseismogramatwell GHV-83~03.
- 6. SHALLOWREFLECTlONSE,SM,CCOALEXPLORATION 129 Vertical resolution Further 1-D modelling was undertaken to assess the effect of variations in coal-zone stratigraphy on seismic data with different frequency bandwidths. Synthetic seis- mograms were generated by convolving Ricker wavelets, with dominant frequencies of SOHz, 100 Hz and 150 Hz, with the time-reflectivity sequence from well C-HV-83-03 (Figure 6). Seismograms were also generated after the logs had been manually edited, specifically with seams I and 2 replaced by interburden. These are the thickest seams and their seismic response dominates the overall reflection character of the coal zone. Both sonic and density logs were used for the modelling. Seismo,grams based on unaltered logs are presented in Figures 7~1,7b and 7c, for the 50-Hz, IOO-Hz and 150-Hz Ricker wavelets, respectively. Those in the middle row (Figures 7d, 7e, 7f) were generated with seam I replaced by interburden, and the bottom row (Figures 7g, 7h, 7i) shows the seismograms generated with seam 2 replaced by interburden. The 50-H.? seismograms (left column) show little chan,gein reflection character for all three models and indicate that conventional seismic data would be unable to resolve fine stratigraphic details of the coal zone. The IOO- Hz seismograms (central column) have a frequency content similar to that observed in the processed Whitewood field data and provide better resolution of geologic variations in the coal zone. For example, there is a significant reflection character difference between the seismogram generated from the model with seam I removed (Figure 7e) and that with seam 2 removed (Figure 7h). Thus, the results of this modelling indicate that the absence of seams I or 2 should be detectable in the seismic data. The synthetic seismo- grams generated with 150-H.? wavelets are shown in the right column and show major reflection character differ- cnces for all three models. Although field data with this bandwidth, are currently unavailable, continued advances in seismic xquisition and processing may realize this in the future. To demonstrate the importance of obtaining the maxi- mum possible bandwidth in reflection seismic data over shallow coal fields, Ricker wavelets of variable central fre- quency were used to generate synthetic seismograms for the same geologic situations as those modelled in Figure 7. The results are presented in Figure 8. In each row, the cen- tral frequency of the Ricker wavelet used was 97 Hz in the left column, 111 Hz in the central column, and 129 Hz in the right-hand column. This range in bandwidth is equiva- lent to th;lt observed in the Whitewood field data. The three geologic situations were modelled in the same order as in Figure 7 and were found to produce seismic respons- es with different reflection character, particularly in data which contain the highest frequencies. For example, the absence of seam I may not be detectable at 97 Hz (Figures 8a, d) but is immediately apparent at 129 Hz (Figures 8c, f), Furthermore, in the 97-Hz data, it is more difficult to tell the difference between the absence of seam I (Figure 8d) and seam 2 (Figure 8g) than at 129 Hz (Figures 8f, i). COAL ZONEI 50 Hz 100 Hz 150 Hz Fig. 7. Synthetic seismogramsat well C-HV-83~03showing the dependence01coal-zonereflectioncharacteron the bandwidthof seismicdata. Seismogramsrepresentingall Highvalecoal seams are shown in the top row.with seam 1 replacedwith interburden (middlerow) and with seam 2 replacedwith interburden(bottom row). Seismogramswere generatedwith 50~HzRickerwavelets (lettcolumn).100.HzRickerwavelets(middlecolumn),and 150~Hz Rickerwavelets(rightcolumn). TWO-DIMEYSIONALNUMERICALMODELLIY(; The second component of the study was two-dimension- al seismic modelling of the Highvale-Whitewood coal deposit. The main purpose of this work was to investigate the effect of source-receiver offset on reflection character, and to test the receiver-array geometry used for reflection seismic data acquisition over shallow targets. This was achieved by examination of 2-D synthetic seismograms and subsequent processing of these synthetic data to pro- duce stacked sections. Snell’s-law ray-tracing techniques were used in this work. Modelling parameters Sonic and density logs from well C-HV-83-03 were again used to provide the model of the coal zone stratigra- phy. Although field seismic data are available at Whitewood only, the use of the same logs as in the previ- ous section allows comparison between 2-D and I-D mod- els, and the continuity of Highvale coal seamsmakes a lay-
- 7. I30 II.“. LYATSKYandD.C.LAWTON -j@-q--r* C”AL1 -----.ZONE ’ 1’-----r;‘:::I:I:il~~~;1 .-----. ,-----. f 1 07 Hz 111 HZ 120 HZ Fig. 8. Syntheticseismogramsat well C~HV~83~03showing co& zone reflection~charactervariations for the range in bandwidths observedin Whitewoodfield seismicdata. Theseismagramswere derivedfromthe samemodelsas thosein Figure7 andwere gent erated with 97~HzRicker wavelets (left column), Ill-Hz Ricker wavelets(centralcolumn),and 129.Hz Rickerwavelets (right COI- “nq wed-earth assumplion realistic. In order to cover a wide range of source-receiver offsets in the modelling. the far offset was chosen to be 150 m. This exceeds the depth to the coal done by a fxcor of about 2.5. A group interval and a near offset of 2 m were chosen to allow detailed exami- nation of variations in retlection character with source- receiver offset. An averegc velocity of 2000 m/s was assumed helween fhe grwnd surlhce and the top 01 the logged interval, this value being based on refraction veloci- ties obtained from the Whitcwood reflection survey. Vertical resolution Two-dimensional synthetic seismograms were generated for the same geological models used in Figure 8. Figure 9 shows the results for Y7-Hz and 12Y-Hz Ricker wavelets, with the lower-frequency data in the left column. Cornpain& these scismogrems with their I-D equivalents (Figure 8) shows that uwclct inlcrlrencc occurs at offws exceeding the depth to the co% zone. This is caused by dif- ferential NM0 rcsultinp in offset-dependent tuning. The loss in vertical resolution is especially severe in the seis- mogram shown in Figure Yb, where seismic character is more complex than in the other seismograms. In this cxam- pie, the second peak (arrowed) is resolvable only at offset distances less than about 50 m Incorporation of variations of amplitude and phase with incidcncc angle into the seis- mograms does nof reduce the loss of ~.cwIuti~n with source-receiver offset (Lyarsky, IYXX). Offset-dependent tuning, as shown in Figure Y, also result in reduced bandwidth of data in stacked sections. ‘This aspect was studied by generating a series of shot gath- ers and processing them into u slacked section. The seome- try used was a Y&trace split-spread configuration with a f;lr offset 01 144 m, and nominal I?-fold coverage. Several NMO-corrcctcd CDP gathers from lhe synthetic data arc (a, 97 Hz bl 129 Hz Fig. 9. Syntheticshot gathersat well C-HV-83-03.generatedwith thesamemodelsusedin Figures7 and8.The seismogramsin the leftcolumnweregeneratedusing97~HzRickerwaveletsandthose in the right columnwith 12g~HzRickerwavelets.In Figuregb, the second peak (arrowed)is attenuatedwithin an offset distance01 50 m.This is causedbyo&et-dependentinterference.
- 8. SHALLOWREFLECTtONSEISMlCCOALEXPLORATtoN 131 presented in Figure IO. Near-offset traces are shown on the right, and far-offset traces on the left of each gather. It is apparent ,that the frequency content of the far-offset data has been reduced due to offset-dependent tuning and NM0 stretch (Buchholtz, 1972; Dunkin and Levin, 1973; Taylor, 1984). The impact of this frequency loss when the gathers are stacked is illustrated in Figure I I. The data in Figure I I a were obtained by stacking only near-offset (0 to 48 m) traces from each CDP gather, whereas the section in Figure I Ib war based on all traces. Loss of resolution in the full stack is demonstrated by the observation that the second peak (arrowed) in Figure Ila is severely attenuated in Figure I Ib. Thus, the full-fold stack would not yield the best possible interpretation of the data. W~rtwvoo~ FWLU DATA Data acquisition The locations of the two seismic profiles used in this study are shown in Figure 12. The data were acquirrd by Gee-Phys-Con. Ltd. in December 1986. A twenty-four channel, split-spread receiver geometry was used in the field, with single 14.Hz geophones. Single geophones were preferred over arrays to avoid smearing of the wide-angle reflections from the shallow coal zone. Both the geophone interval and the near offset were 5 m, with a far offset of 60 m. A shot was fired at every station, resulting in 12.fold subsurface coverage. The data were recorded at a sampling interval of II4 ms. An explosive source was used, com- prised of I5 to 30 cm of primacord placed in shotholes I m deep. Data processing The initial step in the processing sequence was the appli- cation of a refraction statics correction to field data using a method described by Lawton (1989). The purpose was to eliminate the effects of topography and low-velocity 50 75 Fig. 10. NMO-correctedCDP gathers at well C-HV-83-03,based on 2-C synthetic seismogramsusing a 129-Hz Ricker wavelet. Loss of resolutionon the far traces 01each gather is caused by destructiveinterferenceand NM0 stretch. COAL ZONE I COAL ZONE I (a) ;i .5 ii F Fig. 11. Stackedsectionsbasedon CDPgathersshownin Figure 10: (a) incorporatingall tracesfromeach gather:(b) incorporating onlythefour nearest-offsettracesfromeachgather. glacial till. The datum was chosen at an elevation of 780 m, and the till (velocity of 400 to 500 m/s) was “replaced” with bedrock material having a velocity of 1850 m/s, as determined by refraction profiling. All subse- quent processing was performed using the Teknica seismic processing package at the University of Calgary. A fre- quency bandwidth from 30 Hr to 240 Hz was maintained during data processing, retaining the recorded sample interval of l/4 ms. Several field records from Line 4 are shown in Figure 13; thcsc records have been corrected for spherical diver- gence and have been trace-equalized. At .shon traveltimes on near (< 30 m) traces, these data are dominated by ground roll. This is significant as 2-D modelling has shown that near-offset (less than target depth) recording is required to achieve the best vertical resolution. Consequently, a loss of near-offset data could have a detri- mental effect on data quality in the final stack. The brute stack shown in Figure l4a was produced with only the near-offset (5 to 30 m) data from Line 4. The extremely poor quality of this section renders it of limited use for interpretation. In comparison, the far-offset (35 to 60 m) stack in Figure l4b has a much higher S/N. However, the use of a partial range of offsets reduced the subsurface coverage to 6M) percent from the 1200 percent recorded, and the use of source-receiver offsets exceeding the target
- 9. 132 H.“. LY.4TSKYandD.C.LAWl~“N IQSP 100 iIIII iIi I / I I FP 30 1111111dlII I I WE 021 SP 100 1 WW&-022 . . l-- ---------- “88 020. *“L+L!8- ----- --- ww88moo5 l WW88-024 1! - WwRR~nI0 / 1 p8L-o,3 WW88-025; pjE3iy . 1 . . . I !.WWELOO3 ! wwm-018 ww**-017 ww86-oo?I/- Q!_"gT A / iwm-001 . .;"ww88-008~A 1 km Fig. 12. Mapofthe Whitewoodminesite.showingthe lxations of seismicLines4 and6.the deepwelt,andthe 1988drilling program(COUP tesyTransAltaUtilitiesCorp.). depth may result in reduced frequency content in the data after correction for normal moveout (NMO). Nevertheless, all of the stacked sections from Whitewood displayed here- after were based on far-offset data only, in order to provide maximum SIN, cvcn at the expense of some loss of resolu- tion. Alternative processing techniques. such as surgical muting of the ground roll, were found to be less effective than the approach discussed abovc. The degradation of near-offset data due to shot-generat- ed noise suggests that all geophones should bc placed with- m a source~recciver offset window where contamination of rctlection data is minimired yet optimum resolution is retained. This technique, called the “optimum window technique”, was proposed by Hunter et al. (1984). A simi- lar approach was favoured by Varsek and Lawton (I 985). The importance of weathering-statics corrections in shallowrellection data is illustrated in Figure IS, which shows two stxkcd sections of Line 6. Figure 1% was obtained by processing the data corrected for topography only. whereas Figure ISb was generated with the data cot- rccted for the effects of topography and the low-velocity till. The poor quality of the section in Figure ISa demon-
- 10. I33 REFLECTION GROUND ROL Fig. 13. Examplesof field recordsfromLine4. Fig. 14. Stackedsectionsof Line4: (a) usingtraceswith source-receiveroffsetsof lessthan30 m;(b) usingoffsetsgreaterthan35 m strates the need to correct data for weathering statics to improve coherency of seismic events and to help estimate the location of reflectors in the subsurface. The quality of stacked data was further improved by the subsequent appli- cations of an automatic trim-statics correction and a front- end mute. As it is not alweys possible to distinguish direct arrivals from reflections produced by glacial till. only the first I5 ms were muted to avoid possible destruction of useful data. A post-stack frequency-wavenumber filter was applied to enhance subhorizontal events. Processed stacked sections from Lines 4 and 6 are shown in Figures I6 and 17, respectively. Two stacks were generated for each seismic line: a structural section in the upper half of each panel and, in the lower half, a section which has been flattened on a strong event which occurs between 75 and 80 ms. This reflection represents the top of
- 11. Fig. 15. Stackedsectionsof Line6: (a)correctedfor elevationstaticsonly;lb) comxted for elevationandweatheringstatics the Horseshoe Canyon Formation and will hereafter he referred to as the Horseshoe Canyon Marker. or tICM. Structural sections reveal the gross deformational and structural patterns along the two profiles. However, they still contain residual static anomalies that were not totally removed either hy refraction or trim-statics corrections. The imperfect performance of the refraction-statics proce- dure can he explained by the location of some of the causative velocity anomalies below the depth imaged by refmction profiling. The flattened sections were used to interpret reflection character variations and minor stw- tural features. Thus. the two displays complement cxh other and were used together Sorinterpretation. Gross geologic structure From drillhole data at the Whitewood mint site (Figure 12), it is known that the dip of the coil1 is not uniform, and a slight depression in the coal zone exists in the middle part of Line 4. This structural style is seen in Figure Iha. However, the coal-related seismic went contains signifi- cant character variatiotn and rchidual staic croon. which introduce uncertainties in the interpretation of the depth to the coal in the section. Consequently. the interpretation of the gross structure was assisted by the observation of the HCM, which represents a conformable and locally isochronous geologic surface and is the most consistent rcflcction in the seismic section. In Figure 16x1,tt depres- hion is found in the tlCM between shotpoints (SP) 230 and 140. Since the coal-related event follows the shape of the HCM, the deformation must postdate the deposition of the Cd. On Line 6, the depression in the coal LOX appears to he masked by residual static anomalies in the central and northrm portions of the seismic line (Figure 17a).although rcllections in the southern part of the line (SP 235 to SP 27% do chow a northward dip. The ~pparcnt structure between SP 120 and SP 140 is considcrcd to be a long- us:,vzlength static anomaly. As in Line 4, seismic structural analysis was helped by examining the HCM. Although the coal reflection in Line 6 (Figure l7a) generally mimics the structure 01 the HCM, locally there are time differences bctwecn the two events. For cnample, north of SP I75 and SP 140. the coal-HCM isochron incrcaxs abruptly and faulting of the co& zone is suspected. Reflection character variations Figures IXa and l8b respectively show geologic inter- pretations of Lines 4 and 6. These are the most favoured intcrprctations based on modellin& and specific examples are discussed below. Stratigraphic control for the interpre- tation was provided by a deep well (WW-X7-015) drilled near Line 4, 40 m cast of SP 210 (Figure 12). The I-D syn- thetic seismogram based on this well is included in Figure
- 12. COAL HCM 100 m Fig. 16. Final stacked sections of Line 4: (a) structural section; (b) flattened on the HCM. 135 COAL 2 2 HCM 100 m Fig. 17. Final stacked sections of Line 6: (a) structural section: (b) flattened on the HCM.
- 13. 136 H.“. LYAI-SKYandD.C.LAWTON 19, where the central portion of Line 4 is reproduced. This seismogram was generated assuming an invariant velocity of 2000 m/s for the entire modelled interval, based on a checkshot survey, and using the density log to compute the reflectivity sequence. Since the average central frequency of the coal reflection in the stacked section (Figure 19) is about 100 Hz, a Ricker wavelet with a similar central fre- quency was used for the modelling. The coal-r&ted cvcnt and the HCM can easily be correlated between real and synthetic data. The strong peak observed in the synthetic seismogram just below the coal rcflcction corresponds to a sandstone body which appears in Line 4, just north of SP 210 (Figure IXa). The high-amplitude trough-peak pair found in the synthetic seismogram after the HCM (Figure 19) corresponds to a deep coal deposit in the Horseshoe Canyon Formation. This horizon is poorly resolved in the field data due either to penetration problems resulting from too small a charge size or, more likely, to velocity varia- tions at this depth unaccounted for in the modelling. For interpreting the remainder of Line 4, and also for Line 6, the density data from well WW-X7-015 and other drillholes in the vicinity of the seismic lines (Figure 12) were combined to form a density log considered to be typi- cal for the Whitewood coal zone. This log, shown in Figure 20, was used f.or I-D modelling to assist in the interpretation of the seismic data. Cart was taken to (a) include even the relatively thin wal and bentonite beds in the model. Seam 1 is absent in this area and the coals rep- resent seams 2 to 6. Figure 21 shows a part of the seismic section from the northern end of Line 4. At SP 121 a synthetic seismogram is shown, based on the density log in Figure 20, assuming a constant velocity of 2000 m/s. The synthetic seismogram is dominated by a tight (half-period of 4 ms) trough-peak pair preceded by another. smaller peak. This character is contributed mostly by seam 3. Seams4 to 6 are interpreted to be thin and contribute little to the coal-related event at this Iocatiun. A similar character is observed on the south end of Line 4 (Figure l6bj, confirming that seams 2 to 6 are also present there. On this sane line, the small early peak disappears near SP 200, although the main peak- trough pair persists, with interruptions and minor varia- tions, as far north as SP I60 (Figure Ihb). The loss of the side lobe before the coal done may be due to interference arising from the thinning of the interval between coal and the base of glacial till. At SP 202. a strong event is evident immediately after the coal reflection (Figure 16b). It is interpreted to be a reflection from a sandstone in the lower Scollard Formation, as noted above. In Line h. a similar went exists only at the north end of the profile (Figure 17b). On Line 4, between SP 200 and SP IX0 (Figure IO)_the LINE 4 Fig. 18. Geologic interpretations of Whitewood seismic sections: (a) Line 4; (b) Line 6.
- 14. 137 25 75 100 m Fig. 19. Centralpoltion of Line 4 (flattened),including a syntheticseismogrambas?don the density log fromwell WW-87.015 assumingan intervalvelocity012000m/s. DENSITY 0”4g/m3) generated Fig. 20. Synthetic density log used for modelling at Whitewood, basedon logsfromwellWW-87.015andotherwells in the area,as shownin FiglJre12. coal-zone reflection loses amplitude and, in places, disap- pears completely. Similar anomalies are found elsewhere on Line 4, between SP 160 and SP 140 (Figure 16b), and on Line 6, between SP 140 and SP 120 (Figure 17b). It was initially interpreted that these anomalies indicated removal of the coal due to fluvial channelling. However, recent drilling revealed that the coal zone is continuous in these anomalous areas, although it is fractured and seam 3 is sometimes thinned. One-dimensional modelling showed the thinning of seam 3 to be insufficient to produce signifi- cant variations in reflection character, suggesting that these seismic anomalies may be a response to changing petro- physical properties of the coal. Problems with the data, such as unresolved, short-wavelength statics, are not sus- pected since there is no loss of coherency in the HCM reflection. However. it is significant to note that seam 3 occurs at a depth of less than 10 m, and only about I m of inorganic bedrock separates it from glacial till. It is possi- ble that the small acoustic impedance contrast between till and coal may account for the loss of the coal-related event. for the inorganic bedrock layer is too thin and weathered to produce a significant reflection. A possible geologic expla- nation of the fracturing of the coal in the anomalous areas is long-wavelength structural deformation, perhaps of the type illustrated in Figure 4. Glaciotectonic thrusting in the area is documented by Fenton (1987) and minor faulting
- 15. 100 m Fig. 21. Seismic section from the north end of Line 4. with a synthetic seismogram based on the log in Figure 20. generated assuming an int&al velocity of 2000 m/s. has been seen in drillcore by Monenco Ltd. Faulting is also interpreted to occur on Line 4. herween SP 245 and SP 230, where the isochron between the coal zone and the HCM increases (Figure l&t). On Line 6, the coal-zone rrllection is present as a dw- hlet at SOms at the south end of the line (Figure 17b). This is interpreted to indicate the presence of scam I (Figure IXh), based on 1-D modelling of the coal zone, as shown by the synthetic seismogram near SP 260 in Figure 22. The till-coal interval is thicker here than at the south end of Line 4, and the event at 35 ms at SP 260 may represent a reflection from the baseof the glacial till. North of SP 245. the character of the coal-zone reflection changes to a trough-peak pair, as shown hy the synthetic seismogram north of SP 240; this character is similar to that observed on Line 4 (Figure 21). The co&related event on Line 6 becomes discontinuous between SP 230 and SP 190 (Figure l7b). This area coincides with a zone of large weathering-static anomalies (Figure 15~1).suggesting glaciotectonic deformation or scouring of bedrock overly- ing the coal zone. Within this region; between SP 2 IS and SP 200, two high-amplitude reflections occur between 40 and 60 ms. These events are interpreted to rcprcscnt local restoration of a relatively undisturbed coal zone, or erratic material within the glacial till. The central part of Linr 6 is shown in Figure 23. The zone of extensive deformation of coal ends at SP 1x0, and an undisturbed coal retlection is observed south of this point. In the disturbed zone, the coal-related event is repre- cented not by a trough-peak pair. hut hy a more complex reflection pettcm. The top peak (at 35 ms) may represent the base-of-till reflection. or reflections from remnants of scam I. Combined with the event at 50 ms, this peak forms a rcflcction configuration similar to that on the south end of the profile (Figure 22). However, the change in reflec- tion character at SP IX0 and SP I45 (Figure 23) coincides approximately with the structural deformation discussed prrviously. The interpretation favoured is that the coal zone has been deformed in these areas by glaciotectonic ICI0 m Fig. 22. Seismic section from the south end of Line 6, with syn- thetic seismograms representing seams 1 to 6 (SP 260) and seams 2 to 6 (SP 240).
- 16. SHALLOWREFt,ECTIONSE,SM,CCOALEXPLORATlON 139 thrust faulting. Since the ice advanced generally from the north (Fenton, 19X7), a northward dip of the fault plane is expected, as shown by the interpretation in Figure IRb. C0NCLUS10NS This study has allowed us to arrive at the following con- clusions: 1. 2. 3. 4. 5, 6, Reflection seismic data have been proven useful in delineating structural deformation and variations in coal thickness and stratigraphy in the Highvale- Whitewood coal field. Relatively minor variations in coal-zone stratigraphy can produce noticeable changes in the seismic response of the coal-bearing interval. Numerical modelling showed that, at source-receiver offsets exceeding target depth, degradation of reflec- tion character and of vertical resolution occurs due to offset-dependent tuning interference and NM0 stretch. The contamination of near-offset data by shot-gener- ated noise forcer a trade-off between the need for noise reduction and the desire to retain the highest resolution possible. Retlection character can be used to map areas where the coal is shallow gnd heavily fractured. Interpretation of two seismic sections from Whitewood showed that the character of the coal reflection is not always related solely to geologic variations in the coal zone. Reflection character is also affected by the proximity of the glacial till above the coal and sandstone below it. This indicates that the coal zone should not be regarded in isolation, whether seismically or geologically, and that study of overlying and underlying sediments is also required. REFERENCES B”chh”,tr. H.. ,972,A noteOnipli *isrortiandue10dynamic(NM”, corrrcrionr:GeqJhys.Prasp.m195402. Claikr.A.M.. Iwh, Seismicsurveyingandmineplanning:theirrekation- shipsandapplication.ih Muir. W.L.F..Ed..Cd cnploration:PrOC. FirstInternat.CoalEnpl.Symp..London,England,158~191. “emchuk.T.D..19x7.Palynumatigraphyof Paleocenestrataor rilecen- trd AihenaPlains:M.Sc.thesis,Univ.of Alberia. Dunkin.J.W.andkin. EK., 1973.Effectof normalmoveoutonit SeiS- micpuke:Gcnphysics38,h35-642. Fenron.MM., ,987,A modelfur glacialtect”niPmLakeW~hamunarea. Albena.GreatPlains.NarlhAmerica:2ndappmximalion:Presentedat fk xtttil lnlrmal.congress.hltema,.Unionthatem.Fe.. orrawa. Fry,K.C.andOrange,ADS..1482.High-rrrolulionseismic:A practical approachtu coalenphmtion:in Gurgei,K.D..Ed..Pruc.Fifth Symp. I%“,. RockyM”unrainclxil,:Uldl Ceol.Min.suw.Buli. 118.211L?18. Ciibron.D.W.,,977,UpperCielaceousandTcrtiilryc”&bearl”g strataI” he D~umhellrr~Adlcyregion,RedDeerRivervalky, Alberta:Geol. sun,.can..Pqm X-35. Hunter.,.A.. Pulla”, SK. Burns,K.A..Gagne.R.M. andGood.R.L.. IP84.Shallowre3nic refkction ‘““ppi”g “f tile ovcrburdenendrock inrerfarrwith theengineeringscimugrilph~ somesnnpletechnques: Geophysici‘w 13X1-1385. Lawron.D.C.. 1985.Seismicfacisi analysisof delta-plaincods from Camiose.Alherra.anti lku~tiine coalsfrom Pictoucoalfield.Nova Scotia:Bull.Am.ASS”.PefrGeal.69,2120-2129. 19xl).Computation“f refractionstaticC”rrecli”“< usingfirst hrrakmvrllime differences:Geophysics.sublniited. andBenram,MB.. ,983,Renrc,i”n seismicin”entigati”nsof wesamC;m8diii”co;!lfields:unpub.Rep. Lepper.C.M. andRu*kry.E. 1976.High-resolutionseismicretkctian ~echiquesfor mappingcoalsc*m hm he surfacr:U.S.Bureauof Mines.CoalMineHs;dthandSafetyProgram.Tech.PmgreasRrp.101. 100 m Fig 23. Seismicanwnaliesof Line6: (a)in thecentre;(b) atthe northend
- 17. 140 H.". LYATSKYandD.C.tAWTON Lyatsky. H.“., ,988, Refkction SeirmiC study of a shdlow coil, field in central Alberta M.SC. ,kSlS. Univ. of cd!$lry. ~ and Lawton, IX., 19X9, Densiry~based rekctivity in seismic exploration fur cval in Albena. Canads: Ceophyr~ci. submiffed. Peace, D.G.. 1978. Surface reflection seismic - looking undeifraund from rhc surface: in Argilll. 0.0.. Ed., Coal cnploration. 2: Proc. Second Internat. Cuvl Enpl. symp.. Denver. 230-266. Sanorelli. AN., Henderson, I.D. and Lawtun, DC., 1486, Development of high resulutiun reflection seismic techniques for shallow depths in Alhena: Fksen,c* at tile ,986 Na,. Conv.. Can. sue. Enpi. Geophyr.. wgary. __, Hmdrnon, J.D., Kozenberg, G.S. and Prsowrki, M.. 1985. An overview of he uses of surface geophysical methods in wesfern Canadian Prairies cwd mining areas: Presented at the 55th Am. tntemat. Mr!&. sot. Expl. Geophys., W*rhington. Schlicker. H. and BBning, C., 1981, Reflection seismic and wirclinc cur- ing: 7-W” “pfional systems for expl”rlng the Saar con, depusit: is Argnll, 0.0.. Ed.. ma, exploration. % Pmt. Third tnlemaf~ co?,, Expl. Symp.. ciilguy. 229.238. Swcci. A.R. and Hills, L.V.. 1984, A palynoiog~c;ii and scdimenlological analysis of tile Creraceoua~Tcrtiwy boundary. RCd Deer River vauq. Alhenn. Canada ~Absrract,: in Sixth lnfelnai. Pely”“l”gicsl Cod. Calgary. IhO. Taylor. 0.. 19X4.Seismic rei~l~ti~n and field design: Success and faiiure ai Tahcr. A,hc”a: J. Can. sot. EYp,. CCOphys.20. 7-22. Mrs& J.L. and Lawton, U.C., 1985. The seiigun ~ pan It: Dara acquisi- iion fwn a study of slacked sections: J. Can. sue. En*,. Gcuphys. 21. 77.88. Ziolhowski, A. and Lerwill. W.E., 1979. A simple approach I0 thigh rtx- hion seirmic prufilinp for cud Gcuphyr. Pmsp. 27, 360~393.