128Z.Zhangetal./MarineandPetroleumGeology34(2012)119e133

a

b

c

Figure9.SyntheticseismicdiagramsfromthewedgemodelsinFigure8forhydrate-andgas-bearingsedimentsandhydrate-over-gasbearingsediments.ARickerwaveletof50Hzdominantfrequencywaschosentogeneratezero-offsetsyntheticseismicdata.TheP-wavevelocityoftheclayis1810m/sandthedensityis2.06g/cc.Theelasticpropertiesofwater-bearingsand,hydrate-bearingsandandfree-gas-bearingsandwerederivedfromthewell-logdataandtherockphysicsmodel.(a)alsoincludesplotsofamplitudeverses

thickness.

hydratelayer.Forexample,ifhydratesaturationis75%,thethick-nessofahydratelayerat12Hzis19m(14msTWT).

Figure11bshowsthenormalizedamplitudewithrespectto12Hzfrequencywith30%freegassaturationbelowagashydrate-bearing

x10

-3

layerwithvariablehydratesaturations(model“c”inFigure8c).AstraightlinehasaFourieramplitudeof900(Fig.11b).Abovethestraightline,theamplitudesstarttoincreasewithincreasingthicknesses.Thus,weusedthevalueof900asthe

reference

x10

-3

a

10

4

b

10

9

12Hz

Relativeenergystrength

3.532.521.510.5

12Hz

Relativeenergystrength

10

20

30

Thickness(ms)

40

50

87654321

Frequency(Hz)

30

Frequency(Hz)

2020

30

4040

50

10

2030Thickness(ms)

4050

050

Figure10.(a).TheFourierspectraofthemodel“a”with75%hydratesaturationinFigure8,(b)theFourierspectraofmodel“c”with75%hydratesaturatedsandovera30%gassaturatedsandinFigure8.Thecolorbarshowsdifferentenergylevelin(a)and(b).(Forinterpretationofthereferencestocolourinthis?gurelegend,thereaderisreferredtothewebversionofthisarticle.)

Z.Zhangetal./MarineandPetroleumGeology34(2012)119e133129

a

1900180017001600150014001300

b

Fourieramplitude

Fourieramplitude

1200110010009008007006005004003002001000

1

3

5

7

9

111315171921232527293133353739414345

11001

3

5

7

9

111315171921232527293133353739414345

Thickness(ms)Thickness(ms)

Figure11.(a)The12HzFourieramplitudeversusthicknessplotsfromthemodel“a”and(b)model“c”inFigure8.TheamplitudesarenormalizedwithacalibrationpointderivedfromtheLWDdataintheGC955-Hwell(shownasastarinFigure11a).

amplitudeforthecaseofhydrateoverfreegas,whichisahydratelayer19mthick(16.2msTWT)ifhydratesaturationis50%.

Figure12containsaseriesofcrosssectionsforthesixzonesofinterestasdelineatedinFigure2withthenormalizedamplitudesofthewaveformpeakre?ectorfortheinterpretedtopofthegashydrate-bearinglayerplottedacrossthe6zones.Themaximumpeakamplitudeswerepickedalongthesixzonesonthetracesofa12Hzdominantfrequencysub-band,whichwerederivedfromRockSolidAttributesintheKINGDOMsoftwaresuite(Stanley,2008).ThepeakamplitudeswereproportionaltothenormalizedamplitudebycalibratingwiththeLWDderivedgashydratesatu-rationsintheGC955-H(Fig.11a).

Figure12alsodepictstheseismicinferredoccurrenceofgashydrateandfreegasasmodeledinthisstudy.Amplitudeandwaveformanalysesdescribedinabovesectionswereusedinourinterpretationfortheoccurrenceofgashydrateandfreegas.Normalizedamplitudeslowerthanthereferenceamplitudeof900formodel“c”hydrate-bearingsedimentsareinterpretedtobegas-bearingzoneswith/withoutthinhydrate-bearinglayerabove.Normalizedamplitudesgreaterthanthereferenceamplitudeof

Zone1WellGC955Q

Amplitude

Amplitude

Zone2

200

WellGC955H

200

400Distance(m)Zone3

600

800

400

Distance(m)Zone4

600800

Amplitude

Amplitude

Distance(m)Zone5

Distance(m)Zone6

Distance(m)

Amplitude

Distance(m)

Gassandswith/withoutthinhighconcentrationhydrateorthick

lowconcentrationhydrateaboveit

Gassandswiththickhighlyconcentratedhydrateaboveit

Amplitude

Thickhighlyconcentratedhydrate-bearingsands

Figure12.Theseismicinterpretedhydratelayerandhydratewith/withoutfreegasbeneathitinthesixidenti?edzonesinFigure2inferredfromthe12Hzdominatedfrequencysub-band(Fig.11).

130Z.Zhangetal./MarineandPetroleumGeology34(2012)119e133

300formodel“a”hydrate-bearingsedimentsareinterpretedtobe

thickhydrate-bearinglayerswithoutfreegasbelow.Someofthe

highamplitudere?ectorsinFigure2donotnecessarilycorrespond

tothethick,highsaturatedgashydratelayers.Forexample,there

arehighamplitudesinzone6,butthehydratethicknessinthezone

isrelativelythinasshowninFigure12.3.5.2DmodelingForward2Dmodelingwasusedtostudytheseismicresponsesinasequenceofthinandthicksandbedscontainingeithergashydrateat50or70%saturation,orfreegasat30%saturation.Thegeologicmodelwasgeneratedbasedontheinterpretation

fromFigure13.Comparisonbetween(a)?eldstackedseismicsectionand(b)compressional-wavevelocitymodelusedtogeneratethesyntheticstack,and(c)syntheticstackedseismicsection.

Z.Zhangetal./MarineandPetroleumGeology34(2012)119e133131

seismicandloggingdatafromGC955.A35HzRickerwaveletwasusedinthisgeologicmodel.Compressional-wavevelocitiesofclay-richsectionswerederivedfromJIPLegIIloggingdataasshowninFigure13b.Densitiesoftheclay-richsectionswerecalculatedfromamodi?edGardner’sequationbycalibratingwithLWDdata.Withinthesand-richzones,theelasticparametersforvariablegashydrateandfree-gassaturationswerecomputedfromtherockphysicsmodelpreviouslydiscussedinthisreport.

Theestimatedwell-log-andmodel-derivedvelocityanddensitydistributionswereinputintoHampson-Russell’sAVOmodelingpackage(Russelletal.,2001).ThesyntheticamplitudeswerecomputedfromtheZoeppritzequationsintheray-tracingmodelingalgorithm.NMOcorrectionwasappliedinthe?nalmodel.Theeffectsofconvertedwaves,inelasticity,andanisotropywerenotconsideredinthemodel.Overall,thereisagoodagreementforamplitudeandwaveformatthetopofseismicinferredsandlayersbetweensyntheticdatageneratedfrommodeldataand?eldseismicdata(Fig.13aandc).ThebaseofthesandlayerswerenotrecoveredinsyntheticdatabecausenotallsandlayerswerepenetratedatGC955wells.Forexample,onlythehydratesaturationsectionatthetopofthesandlayerwasloggedatGC955-Q.Wesuspectthatamorecomplex2-Dmodelcouldproducethecomplexamplitudepaintingsatthebaseofthesand,iftherewerelogcontrols.

Theamplitudefeaturesseeninthe?elddata(Figs.2and13a)havegashydratesaturationandthicknesscontributions,whicharesimulatedbythe2Dmodeling(Fig.13bec).Theamplitudeanom-alieshavepatternsconsistingofhighwaveformpeakoverhighwaveformtrough(probablyrelativelythickbedofgashydrate)andlowpeakwaveformoverhighwaveformtrough(probablyrela-tivelythinbedofgashydrate),asanticipatedfromtheanalysispresentedearlierinthisreport.VelocitymodelingdataforZones1,2,and3(Fig.2)arealsopresentedinFigure13b.HighpeakamplitudeanomaliesinZone1showhighgashydratesaturationof50%underlainbyfreegasbetweensp30andsp40.WithinZone2,however,theamplitudeanalysisrevealsan18mthickgashydratelayerwithoutfreegasbelowassociatedwithdominanthighpeakamplitudesbetweensp65andsp80.Zone3showpeakanomalies(aquatoblueinFigure13c)changingwithlayerthicknessvariation(from2mto19m)with50%gashydratesaturation.4.Discussion

4.1.Thedistributionofgashydrateandfreegas

Althoughseismicsignaturesofgashydratevarysigni?cantlyacrossGC955,themostprominentseismiccharacteristicinthestudyareaareacousticbrightspots,whicharehighamplitudeanomaliesinthestackedseismicsection.Theamplitudeanomaliesaredistributedinarangefrom3.15sto3.4sTWTthatareequiv-alenttoameasureddepthofapproximately400mbsfto680mbsf(Fig.2).Speci?cdepthsofsomeoftheanomaliesarepresentedinTable1.

Sixzones(Zone1toZone6inFigure2andTable1)havebeenanalyzedinthisstudyfortheoccurrenceofgashydrateandfreegasbasedonamplitudestrengthandpatternsinfullbandand12Hzsub-bandseismicdatafromGC955.Zone1ischaracterizedbychaotichighamplitudere?ectorswithhighpeakamplitudeevents(aquainFigure2andTable1)overhightroughamplitudeanom-alies(yellowinFigure2andTable1)atthecrestofastructuralhigh.Theseamplitudesareinterpretedtobefreegasoverlainbythickhighlyconcentratedhydrate-bearingsandlayers(yellowinZone1inFigure12).Logdatashowanincreaseinvelocityandresistivityintheseismic-inferredgashydrateoccurrencesinZone1.Gashydratesaturationsestimatedfromtheresistivitylogsareapproximatelybetween40%and75%(Guerinetal.,2009).Ablankedzoneabove