Also is known as the Super unsinkable ship as the background, "Poseidon" and the Earth all know the "Titanic" compared to the fate is inevitable. But compared to the Titanic's tragic love story took place, the story of Poseidon is more atmospheric, the film can explore the topic of "growth" to summarize, starring Amy ? Rosen said she was in this movie The task is "no strong-minded from a young girl mature rapidly growing autonomy as a woman" in the film, Amy, played by Jennifer has always regarded his father as the authority, "In the beginning was to find true love with no courage girls ", although in love with handsome young man, but did not dare tell the serious father, through a series of dangerous, she grown into an independent woman.
By Leave of Poseidon Marine Back to Table Of Contents I Large scale deepwater sediment I Imaging Through Gas Clouds: remobilisation: examples from North A Case History in the Gulf of Mexico Sea 3D seismic and outcrop I Neural Network Applications to I Off the shelf seismic data processing porosity Prediction; Hebron Field offshore Newfoundland, Canada I A seismic analysis of late albian age submarine fan I Depth Conversion Methodologies, Uncertainty, Quantification, and I Estimating TIV Anisotropy Application; Hebron Filed, Offshore Parameters in The Jeanne d'Arc Newfoundland Canada Basin at White Rose H-20 By Leave of Poseidon Poseidon was, like many of the Greek gods, difficult to please and dangerous when affronted. Those who chose to risk the sea and acquire marine seismic know well the special challenges attached to that endeavor. The unique problems in processing, acquisition, processing and interpretation require a different set of techniques and thinking, the consequence of which is that those who dare to choose this battle ground never forget that they work upon the seas of the world. Sponsor in part by Large scale deepwater sediment remobilisation : examples from North Sea 3D seismic and outcrop S.J.M. Molyneux 1; J.A. Cartwright2 and L Lonergan, T.H.Huxley School, Imperial College, London, UK. 1 PanCanadian Petroleum Ltd, Calgary. 2Department of Earth Sciences, Cardiff University, Cardiff, Wales The Palaeocene/Eocene to Mio-Pliocene sediments of the Central and Northern North Sea contain deepwater sediments with significant hydrocarbon reserves within submarine fan sands encased in mudrocks. Post depositional processes have significantly changed the original small and large scale geometry and reservoir characteristics of these deepwater sediments. Examples of large scale remobilisation from 3D seismic, well and core data from the above intervals vary from cross cutting 10’s of metres thick and 100’s of metres in extent intrusions of sands, kilometer scale seabed pockmarks, sand mounds with kilometre wavelengths deposited between a mud mounded sea floor and syn to post depositional controlled polygonally faulted deepwater sands. Outcrop analogues of large scale remobilisation are difficult to identify as the scale of these sandstone intrusion is often larger than the outcrop available. Within Upper Miocene Santa Cruz Mudstone, Santa Cruz, California exist one of only two kilometre scale sandstone intrusion complexes in the world. The intrusion of these sands is postulated to have been related to the expulsion of basinal fluids including hydrocarbons and related overpressure build ups, as proposed for several of the intrusive examples above in the Tertairy of the North Sea. The geoscientist should be aware of the small to very large scale nature of sediment remobilisation which can significantly change primary depositional geometries and physical properties of deepwater sediments. These above examples of remobilisation highlight the great variation in reservoir character, which may significantly affect the petroleum exploration and development of such reservoirs. Ignoring such features will lead to incorrect reservoir modelling and subsequent exploration and development “surprises”. Introduction This paper outlines examples of large scale remobilisation as examined from 3 D seismic datasets from Tertiary deepwater reservoirs of the North Sea. Section 1 reviews previous works on large scale remobilisation and the processes which may be involved. Section 2 documents a large scale intrusive complex in block 24/9 of the Norwegian Viking Graben (North Sea) in Palaeocene aged deepwater sediments of the Balder Formation. Section 3 outlines interpreted large scale sandstone intrusions within UK North Sea block 21/4 and 5 of Lower Eocene age deepwater sediments. Section 4 outlines Lower Eocene aged pockmarks of UK North Sea block 15/18 which occur within deepwater slope deposits. Section 5 describes kilometre scale interpreted deepwater sand and mud mounds of the Lower to Middle Miocene Utsira formation in UK North Sea blocks 16/27, 28 and 29. Section 6 presents the Upper Eocene Belton Member deepwater sands which are interpreted to have been controlled in their deposition by polygonally faulted lows and highs forming displacing the palaeo muddy seafloor. Section 7 reviews a coastal exposure of a large scale sandstone intrusion complex which extends over a kilometre within Upper Miocene Monterrey diatomaceous shales, at Yellow Bank Creek, Santa Cruz, California. Finally, section 8 summarises the potential effect large scale remobilisation may have on exploration and development of deepwater reservoirs. 1 - Review Of Large Scale Remobilisation Reservoir remobilisation can be defined as : “any comprehensive redistribution of reservoir units from their original depositional configuration due to pore pressure build up”. Small scale (mm to metre) remobilisation of deepwater sandstone reservoirs in the form of sandstone dykes and sills has been 1 recognized in outcrop and in core since the 1900’s (see review of sandstone intrusions in Allen, 1985 ). Large kilometre scale remobilisation of deepwater sediments has been rarely observed and only two documented outcrops exist worldwide ,(a) Ordovician 3 2 sandstone intrusion complex of the Southern Front Range, Colorado, USA (Harms, 1965 ) and (b) Upper Miocene sandstone intrusion 3 4 5 complex, Yellow Bank Creek, Santa Cruz, California, USA (Jenkins, 1930 ; Thompson, 1995 ; Thompson, 1999 ). Only recently has large 6 scale remobilisation been recognised subsurface with the advent of 3D seismic. Dixon et.al., 1995 highlighted large scale injection and remobilsation of sands from the Forth/Harding hydrocarbon field reservoired in North Sea Tertiary deepwater sediments. Thicknesses of sandstone intrusions reach upto 10’s of metres with lateral extent over 100’s of metres. Sandstone intrusions also occur within similar 7 8 9 aged Tertiary North Sea deepwater sediments of the Balder (Jenssen, et.al. 1993 ), Gryphon (Jaffri, 1993 ; Timbrell, 1993 ), and Alba 10 (Lonergan and Cartwright, 1999 ) hydrocarbon fields. It is important to note that these deepwater turbidite sands were deposited within extensive deepwater muds which were pervasively deformed by polygonal faults. Polygonal faults are a response of muds to expelling water in three dimensions almost immediatley upon 11 12 13 14 15 deposition and during burial (Cartwright 1994a , 1994b , Cartwright and Lonergan 1996 , 1997 , Cartwright and Dewhurst 1998 ; 16 Lonergan et.al. 1998 ). These faults on seismic data sets have a polygonal pattern in plan view, are spaced between 100 to 1000 metres apart and have displacements of between 10 to 100 metres. Below are a series of large scale remobilisation examples interpreted from 3D seismic, well and core data from the North Sea (section 2 to 7) and an outcrop example from California, USA (section 8). 2 - Norwegian North Sea Block 24/9 Palaeocene Balder Formation Large Scale Intrusive Complex Large and small scale sandstone injection features (discordant dykes, sills and clastic intrusions) are documented from 3D seismic, well log and core data in Late Palaeocene to Eocene sediments from Norwegian North Sea block 24/9. The study area is located in the axis of the South Viking Graben, with Tertiary deepwater sediments supplied mainly from the west. Two wells in the block (24/9-5 and 24/9-6) encountered oil in massive sandstones of late Palaeocene/early Eocene age (Balder Formation), corresponding to a high amplitude response on seismic. These sandstones are confined to an isolated east/west trending fault bounded low, and vary from a few metres to 40m thick. Deposition from detached sand-rich debris flows or high energy turbidity currents is suggested. Evidence for modification of depositional geometry is present at all scales. From seismic data, steeply dipping high amplitude continuous reflectors (10 to 50 msecs TWT/ 10-50 metres thick) can be followed from within the Balder, up along fault planes and into Eocene sediments above cross cutting up to 200m of stratigraphy at an angle of 10 to 30 degrees. The top of the Balder reservoir is complexly faulted, though many of these faults do not offset the base reservoir surface. Discordant sandstone dykes, shear fractures and compactional features in core suggest early sand remobilisation. These steeply dipping high amplitudes originating from the Balder below are interpreted to be 10 to 50m thick sandstone intrusions. Injection is postulated to have occurred by expelling basinal fluids and hydrocarbons during the Eocene to Miocene, which built up overpressure in the isolated Balder age sands below, with overpressure being released during during fault movement. The Balder discovery contains hydrocarbon reserves, but untested upside in the form of the injected sandstone high amplitude limbs could form a well-connected reservoir, cross-cutting stratigraphy, and increasing recoverable reserves. If successfully proven, field development would be optimised by locating producing wells in the higher injected portions, draining connected reservoirs in the Balder Formation. 3 - Uk North Sea Blocks 21/4 And 5 Lower Eocene Aged Large Scale Sandstone Intrusions Kilometre scale, Lower Eocene “V” shaped amplitude anomalies, have been recognised on 3-D seismic surveys from the UK Central North Sea. These features are interpreted as concave upward conical sandstone intrusions. They are typically circular to elliptical in map-view with diameters of single cones ranging from 0.5 to 2km. Their sides are 100-300m tall, and dip between 5 and 15 degrees (compacted). Intrusion width is 10 to 50 msecs which calibrates to 10-50 metres of sandstone exhibiting a blocky, low gamma ray character on wireline logs. The conical sandstone geometries can be isolated or developed as compound structures amalgamated in clusters covering hundreds of square kilometres. These amplitudes match the geometries of polygonal faults which pervasively deform the encasing Lower Eocene muds. We considered several possible genetic models for these structures, including palaeo-pockmark craters, collapsed gas hydrate diapirs, and even isolated fan channel ribbon deposits. However, the preferred explanation is that upward migration of thermogenic methane generated from Jurassic source rocks released during earthquake or by auto-triggered failure of pressure seals led to inflated fluid pressures in isolated Palaeocene or Eocene deep water sandbodies. Catastrophic pressure inflation (internal blow-out) then led to large volume liquefaction of the sands, and upward conical injection which exploited pre-existing fault and fracture networks (polygonal faults), in the low permeability claystones of the Lower Eocene. We believe that these Lower Eocene intrusions are the largest reported subsurface examples of clastic remobilisation, which by their very scale may have been overlooked or misinterpreted in outcrop in other basins. 4 17 4 - Uk North Sea Block 15/18 Lower Eocene Aged Large Scale Pockmarks (Cole, D, 1998 ) A detailed 3D seismic interpretation has identified crater-like structures in the Paleogene sediments of block 15/18, Outer Moray Firth, UK North Sea. The structures range from 500 metres to 4 km in diameter and are between 50 to 200 metres deep. The structures are confined to one stratigraphic level, at the top of the early Eocene Balder Formation, and are restricted to an area of 15 x 15 km square within the available 3D seismic coverage. The craters appear to cut down into the top Balder Formation on 3D seismic data. They have seismically chaotic infill and mounded overlying reflectors. Planar high amplitude “wings” are located on the flanks of several of the structures. Well data suggest that the structures are themselves sand-filled, but occur within an argillaceous and tuffaceous unit. A 100m thick mudstone unit overlies the structures. The planar “wings” appear to be intruded sandstone dykes. Shallow gas eruptions on the early Eocene seafloor is the most likely mechanism for the formation of the craters as explosively formed seabed pockmark craters. The gas may have been sourced from deep Jurassic or Carboniferous sediments within the Witch Ground Graben, or it could have originated in the Palaeogene delta sediments within the study area. 6 3 Similar structures may prove to be effective hydrocarbon plays. The structures are large (100 x 10 m gross volume) and appear to be infilled with clean, massive sand. The suggested mechanism requires large-scale early Eocene hydrocarbon migration. If correct this is evidence of both a mature source and effective migration pathway. Migration of hydrocarbons would need to continue after deposition of the top seal to allow any entrapment. The overlying muds provide a competent top-seal to have forced sandstone diapirism. These muds may still act as an effective top-seal in restricted areas. 5 - Lower To Middle Miocene Utsira Formation Kilometre Scale Sand Mounds, Uk North Sea Block 16/27, 28 and 29 The Lower to Middle Miocene (Utsira Formation) of the UK North Sea blocks 16/27, 28 and 29 contain elliptical to circular mounded features, 1 to 2km wide, 0.5 to 10km long with a vertical relief of 50-100 metres. They onlap a muddy strata below of Oligocene age, have flank dips greater than 7 degrees (compacted) and are formed of massive sands as noted from well control. These mounds occur over 10’s of thousands of square kilometres of deepwater basinal shales of the Central North Sea. The Utsira sand mounds are interpreted to be lowstand turbidite sands sourced from marine sands to the North in the Northern North Sea 18 19 (Gregerson 1997 , 1998 ), some 300 km away. These sands deposited onto a mud mounded sea floor with kilometre wavelengths. 20 21 Anketell et.al. 1970 predicted such mounding within muddy substrates. Similar geometries have been noted by Davies et.al. 1999 22 within the Oligo – Miocene succession of the Faeoroe-Shetland Trough and Vogt (1997 ) in Pliocene sediments of the Norway Basin. 6 - Upper Eocene Belton Member Deepwater Sands Syn and Post Depositionally Polygonal Fault Controlled, Norwegian North Sea Block 24/9 Upper Eocene deepwater turbidite sands (Belton Member) of Norwegian offshore block 24/9 are expressed as a high amplitude reflection on 3D seismic data on a low amplitude palaeo sea floor of mud. Well data calibrates these amplitudes as blocky sands between 0 to 100 metres thick. These turbidite sands have been compartmentalised by polygonal faults which pervasively deform the muds. Amplitude maps and well data indicate that the sand is preferentially concentrated in the polygonally faulted lows on the Upper Eocene surface. These polygonal faults have fault displacements upto 300 metres in extreme cases, but more commonly between 10-50 metres and spaced 100 to 1000 metres. The preferred interpretation is that the Belton Member sands deposited as deepwater sands onto a polygonally faulted palaeoseabed, which concentrated deposition of sands within the downthrown lows. 7 - Upper Miocene Large Scale Sandstone Intrusion Complex, Santa Cruz, California, USA Giant scale sandstone intrusions are exposed on the coastline North West of Santa Cruz, California. They are the largest scale clastic intrusions in the world. Sandstone dykes and sills of between 1cm and 300metres width, intruded into Upper Miocene diatomaceous mudstones, probably as a function of minor Pacific Plate reorientation at 4Ma (Casey Moore pers.comm. 1999). The intrusion is related to movement along a series of wrench faults at this time, part of the San Gregorio/San Andreas fault zone (Casey Moore pers.comm. 1999). The dykes and sills in outcrop are fine to coarse grained poorly dolomite cemented clean massive sandstones. The sandstone is often partially stained with degraded oil which at Yellow Bank Creek, the site of a 300m wide sill/dyke, display as complex banding and balls/pillows of migrating hydrocarbons. These giant scale clastic dykes and sills may have been associated with migrating hydrocarbon fluids, expelling from offshore basins into this regionally higher, basin margin position. 5 8 - Exploration and Development Issues Related to Large Scale Remobilisation of Deepwater Reservoirs Interpretation of deepwater sandstone reservoirs is normally based on the interpretation of reservoir geometry that has a primary depositional origin. However, seismic and core evidence from isolated deepwater sandbodies in the Palaeocene and Eocene of the North Sea and the Californian field analogue reviewed above shows that original depositional geometries can be significantly modified by sub- surface remobilisation of sands by processes related to elevated pore fluid pressures which may involve hydrocarbons including gas. Important points to note regards large scale sandstone remobilisation are: (a) Post depositional remobilisation of deepwater sediments is an important element in reservoir interpretation and modelling. (b) Remobilisation involves in situ, higher pore pressure related processes such as fluidization and liquifaction of porous, poorly cemented sandstones at some stage during burial and reconfiguration of the original depositional geometry. (c) Remobilisation changes reservoir properties and geometry: reservoir characterisation must be approached with this in mind. (d) Failure to identify remobilised reservoirs prior to completion of the reservoir model will lead to unexpected results during development drilling. (e) Large scale intrusion complexes may be related to tectonic activity and related flow which may also be associated with the migration of hydrocarbons from basinal kitchen areas. The presence of gas may be a large factor in heightened pore pressures associated with intrusions. (f) Relationships to active faulting and folding such as large scale polygonal faulting of clay dominant systems may affect primary deposition of deepwater sand reservoirs and significantly change final reservoir geometry. (g) Geometry of typical dyke-sill complexes is highly variable and complicated. Wide varieties of geometries exist, from simple vertical dykes and horizontal sills to more complex sill-dyke geometries with conical, rectilinear or radial patterns, sea floor mounds, and pockmarks. Large scale deepwater sediment remobilisation is a poorly documented and understood phenomena. The field geologist, seismic interpreter and petroleum geologist should bear in mind the scale and likely reconfiguration remobilisation may have on deepwater sediments in outcrop, well and seismic data. Incorrect recognition of remobilised sediments may for the petroleum geoscientist lead to subsequent exploration and development “surprises”. Acknowledgements The authors wish to thank Fina North Sea Ltd for funding the authors PhD research at Imperial College, London UK with special thanks to Mick Cope and Rod Laver. Fieldwork in California was sponsored by an American Association of Petroleum Geologist and Petroleum Exploration Society of Great Britain student fund, and a London University Research Grant. Many thanks for discussion on the process of sediment remobilisation go to Edwin Tervoort, Dustin Lister, Richard Evans, Nick Lee, Dave Dewhurst and Richard Jolly while at Imperial College, London, UK. Finally, thanks to PanCanadian Petroleum Ltd for encouraging me to present this work, especially Ian McIlreath of the Calgary office and Alan Booth of the UK office. Special thanks too Chris Paton and Pentagraphix for computing advice during construction of this paper. Views in this paper are those of the authors. References Allen, J.R.L., Principles of Physical Sedimentology, George, Allen and Unwin, 1985. 272p. Harms, J.C., Sandstone Dikes in relation to Laramide Faults and Stress Distribution in the Southern Front Range, Colorado. Geological Society of America Bulletin, 1965, 76, 981-1002. Jenkins, O.P., Sandstone dikes as conduits for oil migration through shales. Bulletin of the American Association of Petroleum Geologists, 1930, 14, 411-421. Thompson, B.J., Geometry and fluid flow mechanisms of the bituminous sandstone intrusion at Yellow Bank Creek, western Santa Cruz County, California : Thesis for Masters Degree, University of California, Santa Cruz, unpublished. 1995. Thompson, B.J., Garrison, R.E. and Moore, J.C., A late Cenozoic intrusion west of Santa Cruz, California : fluidized flow of water and hydrocarbon-saturated sediments. In : Late Cenozoic fluid seeps and tectonics along the San Gregorio fault zone in the Monterrey Bay region, California (eds) Garrison, R.E., Aiello, I.W. and Moore, J.C., Annual Meeting of the Pacific Section AAPG, Monterrey, California, April 1999, 1999. Dixon, R.J., Schofield, K., Anderton, R., Reynolds, A.D., Alexander, R.W.S., Williams, M.C. and Davies, K.G. , Sandstone diapirism and clastic intrusion in the Tertiary Submarine fans of the Bruce-Beryl Embayment, Quadrant 9, UKCS. In: Hartley, A.J. and Prosser, D.J.(eds), Characterization of Deep Marine Clastic Systems, Geological Society Special Publication, 1995. 94, 77-94. Jenssen, A.I., Bergslien, M., Rye-Larsen, M. and Lindholm, R.M., Origin of complex mound geometry of Palaeocene submarine-fan th sandstone reservoirs, Balder field, Norway. From: Petroleum Geology of Northwest Europe Proceedings of the 4 Conference (edited by J.R.Parker), 1993. 135-143. Jaffri, F., Cross-cutting sand bodies of the Tertiary, Beryl Embayment, North Sea, 1993. Unpublished PhD thesis, University of London, UK. Timbrell, G., Sandstone architecture of the Balder Formation depositional system, UK Quadrant 9 and adjacent areas. From: Petroleum th Geology of Northwest Europe: Proceedings of the 4 Conference (edited by J.R.Parker), 1993. 107-121. Lonergan, L. and Cartwright, J.A., Polygonal faults and their influence on deepwater sandstone reservoir geometries, Alba field, UK Central North Sea, 1999. AAPG Bulletin, 83, 410-432. 6 Cartwright, J.A., Episodic basin-wide fluid expulsion from geopressured shale sequence in the North Sea basin, Geology, 1994a. 22, 447-450. Cartwright, J.A., Episodic basin-wide hydrofracturing of overpressured Early Cenozoic mudrock sequences in the North Sea basin, Marine Petroleum Geology, 1994b. 11, 587-607. Cartwright, J.A and Lonergan, L., Volumetric contraction during the compaction of mudrocks: a mechanism for the development of regional-scale polygonal fault systems, Basin Research 1996. 8, 183-193. Cartwright, J.A. and Lonergan, L., Seismic expression of layer-bound fault systems of the Eromanga and North Sea Basins, Exploration Geophysics, 1997. 28, 323-331. Cartwright, J.A. and Dewhurst, D.N. Layer-bound compaction faults in fine grained sediments, GSA Bulletin, 1998. 110; no 10; 1242-1257. Lonergan, L., Cartwright, J.A. and Jolley, R., The geometry of polygonal fault systems in Tertiary mudrocks of the North Sea, Journal of the Structural Geology , 1998. 20, no.5, 529-548. Cole, D., Large crater-like structures in the Palaeogene sediments of Block 15/18, Outer Moray Firth, UK North Sea, 1998. Unpublished MSc thesis, Imperial College, London. Gregerson, U., Michelson, O. and Sorenson, J.C., Stratigraphy and facies distribution of the Utsira Formation and the Pliocene sequences in the northern North Sea, Marine and Petroleum Geology , 1997. 14, no 7/8, 893-914. Gregerson, U., Upper Cenozoic channels and fans on 3D seismic data in the northern Norwegian North Sea, Petroleum Geoscience, 1998. 4, 67-80. Anketell, J.M., Cegla, J. and Dzulynski, S., On the deformational structures in systems with reversed density gradients. Rocznik Polsk. Towarz. Geol., 1970. 40, 3-30. Davies, R., Cartwright, J.A. and Rana, J., Giant (km) scale differential compaction/density inversion pillows in fine grained sediments, Geology, 2000., in press. Vogt, P.R., Hummock fields in the Norway Basin and Eastern Iceland Plateau: Rayleigh – Taylor instabilities? Geology, 1997. 25, no.6, 531-534. 7 E x am p les o f rem o b ilisatio n a ffects o n d eep w a te r re serv o irs D E P O S IT IO N A F T E R R E M O B IL IS AT IO N a) C H A N G E IN R E S E RV O IR G E O M E T R Y b) M A S S IV E B L O C K Y SA N D STO NE IN T R A -R E S E RV O IR S H A L E S C H A N N E L IS E D T U RBITIQ UE SA N D ST O N ES N O IN T E R N A L C L AY B R E A K S C H A N G E IN R E S E RV O IR P R O P E RT IE S c) DYKE OWC S IL L IS O L AT E D C H A N N E L S S A N D S C H A N G E IN C O N N E C T IV IT Y O F O R IG IN A L LY IS O L AT E D R E S E RV O IR S PRO DUCER YES NO d) OWC OWC C H A N G E IN T O P R E S E RV O IR S U R FA C E A N D IN R E S E RV O IR V O L U M E T R IE S 8 M o d el o f larg e scale rem o b ilisatio n featu res in th e E o cen e to M io cen e o f th e N o rth S ea SH EL F P o ck m a rk s (C o le 1 9 9 8 ) SLO PE U tsir a m o u n ds L arg e an d sm a ll a nd c o llap se sc a le c lastic in tru sio n s C o n e sh ap e d fe a tur es o v e r B A S IN (N o rw ay 2 4/9 ) in tru sio n s a lo n g m ud m o u n d s p oly g o n al fau lts (U K B lo c ks 2 1 /4 + 5 ) G as m o und s P oly g o n a l g ro w th (B r o o k e et a l ., 1 99 5 ) fa u lts (B e lto n M e m b e r) S an d p ro n e S an d v o lca n o S h ale p ro n e T u rb id ite Ju rassic h y d ro c arb o n ric h so u rc e ro ck s M ig ratin g S an d filled ch an n e l an d h y d ro carb o n s p o ck m a rk p o n d ed san d s P re J u ra ss ic (Tim b rell, 1 9 9 3 ) (N .B . S alt w ith in th e P erm ian m ay b e d iap iric ac tin g as a co n d u it fo r escap in g b asin al flu id s ) F au lt P e rm ian salt C retace o u s C h alk 9 Im p act o f p o ly g o n al fau ltin g o n reserv o ir g eo m e try ( C artw rig h t et. A l., 1 9 9 9 ) D E PO S IT IO N A L : L A K E S U P E R IO R 2 4 /9 B E LT O N ALBA ( W attru s e t. a l., 2 00 0 ) P O ST-D E P O S IT IO N A L : T IE R S IN T R A T IE R B E D L E N G T H S H O R T E N IN G P O ST-D E P O S IT IO N A L : -IN JE C T IO N 2 4 /9 FA U LT S N U C L E AT E T R IG G E R E D B Y P R E -E X IS T IN G ON SANDBODY FA U LT S FA U LT M A R G IN S 10 L o catio n o f 2 D an d 3 D se ism ic d a tasets u sed in th is N o rth S ea stu d y. (p artly fro m P E S G B , 1 9 9 7 ; Jo n es et.al. 1 9 9 9 ; P erg ru m an d S p en cer, 1 9 9 0 ) G ra b e n in g ik N o r th V AY M EN T E A S T SH E T L A N D B E RY L E M B P L AT F O R M 24 /9 D GE RI RD FO AW CR U T S IR A H IG H G 15 /18 FL UN RO A DS DE P N UR H O R D A P L AT F O R M 1 6/27 16 /28 O ute r M ora y , 29 F irth 2 1/3 , 4 an d 5 JAE REN E HIG H I DG DR EA F O RT IE S-M O N T R O S E TE RH H IG H N O R WAY PE UK Ce n tr al Gr ab en FORTH APPR OA C HES JO H I B A SIN SE GH PH IN E 0 m ile s 50 S e ism ic lin es E oc en e S he lf e dg e 3 D se ism ic stu d y a re as U pp er Ju ra ssic th at h as S c ale o il m atu re (P erg ru m an d S p en c er, 19 9 0 ) 11 “Off the Shelf” Seismic Data Processing David Campden, Andy Dyke, Josef Heim & Elaine Jeong, WesternGeco, Calgary Summary Seismic reflection data offshore Eastern Canada are heavily contaminated with complex water-bottom multiples, caused by a hard and irregularly shaped sea floor. In addition, some target reservoir intervals require imaging of steeply dipping structures in an environment influenced by salt tectonics. This paper discusses the use of free surface multiple suppression techniques and Kirchhoff pre-stack time migration methods and compares results with more conventional multiple suppression and imaging techniques. Introduction Briefly, the structural styles offshore Nova Scotia have been controlled by the high sedimentation rates associated with the progradation of the late Jurassic to early Cretaceous Sable Delta complex. This sedimentary loading propagated a series of large, down-to-basin, listric faults, as well as initiating the movement of deeply buried Triassic salt. The latter accentuated the growth of syndepositional faults and the formation of numerous localised sedimentary basins (CNSOPB, 1997). Much of the current hydrocarbon exploration interest in this region is in the transition zone from shallow (~30 m) to deep (~3 km) water, or off the continental shelf. Associated with this large change in water depth are deep channels in the sea floor, created by turbidite flows, carrying sediments from the shallower regions down the continental slope to the deep ocean floor. This combination of structurally complex geology and a highly irregular water bottom is not friendly towards standard "off-the-shelf" processing methods. The main problems are complex water bottom multiples and seismic velocity fields that have a high degree of lateral variability. While standard post-stack time migrations and Radon de-multiple techniques provide a first-pass solution, more sophisticated techniques are required to unravel the structural complexity and better suppress the water-bottom multiples. Free Surface Multiple Elimination Numerous publications on the subject of surface related multiple elimination (SRME) exist (e.g. Verschuur, 1992). The method itself requires no knowledge of velocities, or where the multiples are generated. It does, however, require very regular geometry, with a shot at every receiver location and all offsets from zero to the maximum offset to be present. SRME will, in theory, model any multiples with a reflection point at the free surface (i.e. it will not attack inter-bed multiples). Since the multiple model that SRME produces is based on a spatial convolution of the data with itself, the source signature must be estimated and removed from the model before its subtraction from the input data. In practice, a simple subtraction of the source-signature-deconvolved multiple model from the input data does not work! An adaptive subtraction process must be used, consisting of residual matching filters to apply minor phase and time shifts to the multiple model prior to subtraction (e.g. Nekut and Verschuur, 1998). These residual differences between the multiple model and input data result from errors in estimating the source signature and from 3D effects. Offshore Eastern Canada, where the sea floor topography can be very complex, it is the 3D effects that really cause the problems. Extension of the 2D version of SRME to 3D (van Dedem and Verschuur, 1998; van Dedem et.al., 1999; Nekut, 1999) requires a significant increase in computational effort and is still in the experimental stage (i.e. large scale 3D applications of the technique have yet to be published). An inverse scattering formulation of free surface multiple elimination has been used for the results in this paper. This differs from SRME in that it models the source as a monopole, rather than a vertical dipole in the water (Weglein, 1999). This more correct modelling of the source can give better results for steeply dipping multiples compared with SRME. The data example presented here is a 2D line from offshore Nova Scotia. It Source spacing: 25 m was acquired parallel to the continental shelf, thus cutting across a number Receiver spacing: 12.5 m of channels created by turbidite flows. The acquisition parameters are Source to near trace distance: 196 m given in Table 1. Number of receivers: 576 CMP spacing: 6.25 m Figure 1 shows a stack of the line, with a Radon de-multiple applied, CMP fold: 144 specifically directed towards removing the water-bottom multiple. While Table 1: 2D line acquisition parameters this technique works well where the water bottom is regular, it is less successful below where the water bottom contains changes in dip. Inverse Scattering Multiple Attenuation (ISMA) was applied to the line by first summing adjacent receivers, so the receiver and shot spacing were the same. Next, the data were extrapolated to zero offset using a Hyperbolic Radon Transform applied in the CMP domain. The data were then sorted back to shot domain for ISMA to be applied. After estimation and removal of the source signature, a residual matching process was carried out. This is the most problematic and subjective part of the process. Allowing too much freedom to match the multiple model with the input data results in primaries being removed. Too little freedom, and artefacts can be introduced by the multiple model. 12 Figure 2: shows the ISMA de-multipled stack (no additional multiple suppression has been used). This shows that in the areas below water- bottom channels and where the water bottom is dipping, ISMA does a better job of suppressing the more complex multiples than the Radon transform. 2.0 3.0 Two-way time, s 4.0 5.0 Figure 1: Stack with Radon de-multiple. The multiples generated from the relatively flat water bottom have been suppressed. The circled areas show residual multiples from below the water bottom with steeper dips. 2.0 3.0 Two-way time, s 4.0 5.0 Figure 2: Stack with ISMA. Multiples from below the water bottom with steeper dips have been suppressed relative to the Radon de-multiple stack. Imaging methods Generally speaking, post-stack time migration is not a good imaging method for East Coast Canada data. Figure 3 shows a post-stack time migration from a portion of the line shown in Figures 1 and 2. This is supposed to show a salt upheaval below the Base Cretaceous. However, steep dips are poorly, if at all, imaged and there are migration "smiles" between 4 and 5 seconds. 13 Pre-stack time migration (PreSTM) was performed on the line using two different methods. First, Common Offset Time Migration (COTiM) was performed using all 144 offsets. This form of PreSTM has been widely used in the industry (i.e. MOVES from Veritas DGC, as well as COTiM from WesternGeco). The final COTiM migration is shown in Figure 4 and shows a generally "cleaner" image compared with the post- stack time migration, with less evidence of migration "smiles". Kirchhoff migration was used for the second PreSTM method, the results of which are shown in Figure 5. Comparing Figures 4 and 5, it is clear that the Kirchhoff migration has imaged the steep dips significantly better than COTiM. There are still some conflicting dips in the image (on the left flank of the salt upheaval), but these are likely to be out-of-plane (3D) effects. Kirchhoff migration uses a double square root (DSR), which is a description of the offset term, in the design of the migration operator. Although the migration part of COTiM is zero offset (single square root), the combination of NMO, DMO and zero-offset migration is attempting to mimic the DSR. In general, Kirchhoff is a more expensive method of PreSTM compared with the COTiM approach, but with increasing computer power, more efficient algorithms and a better looking final image, it is becoming an increasingly attractive option. 2.0 3.0 Two-way time, s 4.0 5.0 Figure 3: Post-stack time migration 2.0 3.0 Two-way time, s 4.0 5.0 Figure 4: COTiM pre-stack time migration (144 offset slots) 14 2.0 3.0 Two-way time, s 4.0 5.0 Figure 5: Kirchhoff pre-stack time migration Acknowledgements The authors wish to thank GSI for allowing the publication of this data and also to Pete Stewart and Björn Mueller from WesternGeco in Houston for their technical support. References CNSOPB, 1997; Technical Summaries of Scotian Shelf Significant and Commercial Discoveries. van Dedem, E., Verschuur, D. and Schonewille, M., 1999; 3-D surface-related multiple prediction and data reconstruction: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 1068-1071. van Dedem, E. J. and Verschuur, D. J., 1998; 3-D surface-related multiple elimination and interpolation: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 1321-1324. Nekut, A. G. and Verschuur, D. J., 1998; Minimum energy adaptive subtraction in surface-related multiple attenuation: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 1507-1510. Nekut, A. G., 1998; 3D surface-related multiple prediction, Annual Meeting Abstracts, Society Of Exploration Geophysicists, 1511-1514. Verschuur, D. J. and Berkhout, A. J., 1992; Surface-related multiple eliminations: Practical aspects: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 1100-1103. Weglein, A.B., 1999; Multiple attenuation: an overview of recent advances and the road ahead, The Leading Edge, 18, 40-44. 15 A Seismic Analysis of a late Albian age Submarine fan, Glenwood Prospect, Jeanne d'Arc Basin, Canada. Torbjorn Fristad*, Andrew Barnwell*, *Norsk Hydro, and Edward Stacey, Petro-Canada Oil and Gas. Summary Reservoir Prior to this work, amplitude anomalies (bright spots) inferred to In order to predict the likely reservoir parameters, potential be late Albian age submarine fan deposits were mapped on 2D analogue submarine fans from the North Sea were reviewed. seismic data in the northern Jeanne d'Arc Basin. One of these The Glenwood fan is small (3-8kms) in length, and its seismic prospects was acquired in 1999 by a group operated by Norsk expression indicates a single package consisting of low-angle Hydro (33.34%), and partners Petro-Canada (33.33%) and Mobil reflectors with rare clinoforms and poor internal character. The (33.33%) (fig.1). The prospect is a deep marine fan, which is morphology is lobate, which may be channelised. These likely to be sand-prone, with good-excellent reservoir potential. characteristics indicate a sand-rich fan (Reading & Richards, Now covered with 3D seismic the Glenwood anomaly has been 1994). Such fans have high N/G (> 0.7), but are usually fairly reinterpreted and mapped. Other methods, including AVO, AVO thin (<100m). These types of fans may be deposited during attributes, amplitude and structure maps and wavelet highstands, due to shoreface reworking. The amplitude classification were utilised in order to enhance our anomalies also suggest a sand model. If so, then Glenwood understanding of this anomaly. This talk will outline our appears to be a close analogue to the Miller Field of the UK approach into the characterisation of the anomaly. sector; data from Miller and other sources were used as reservoir input parameters for volumetric calculations. Introduction The Glenwood prospect is positioned on license EL 1047 and is Geophysical Interpretation positioned north of White Rose and is situated just off the Outer Initially the 1998 seismic amplitude anomaly was mapped on 2D Ridge in the northern Jeanne d'Arc Basin. (Fig. 1). The data. New 3D seismic data confirmed the presence of this Glenwood prospect has a potential STOOIP of 328 mmbbls of oil amplitude anomaly and enhanced our understanding of the or 3.1 TCF of gas. The seismic anomaly is interpreted to be a geology, as a feeder system was evident up-dip of the previously mid-Upper Cretaceous aged, submarine fan deposit. The trap is interpreted crestal area. entirely stratigraphic. The top seal will be a deep-marine shale and with an up dip seal within the main feeder channel. The The seismic anomaly was mapped at a centre of a trough source rocks are deeply buried and will be gas prone or over (negative amplitude). The amplitude anomaly should correspond mature under the license, but oil in nearby wells suggest the to an acoustic impedance drop, increasing the likelihood for possibility of an oil case. having a porous rather than a calcite cemented sand. The prospect is defined by a strong seismic amplitude anomaly, which maps out as a set of fan lobes, which occur at slightly different stratigraphic levels, with a maximum length of about 8kms. The top of the prospect, at the up dip seal location, lies at 3710m depth. It has a maximum thickness of about 70m and a 2 combined area of about 130 km Glenwood Figure 1: Location map of the Glenwood prospect. Methodology The first objective of this study was to review the regional petroleum geology of the Glenwood license using an integrated geological and geophysical approach. There are no wells drilled on the license, but several exist in the vicinity. These wells, plus all available internal and published data, were used in Figure 2: Outline of Glenwood fans. conjunction with the seismic database to formulate the evaluation. Most well data were obtained from the GSC 'Basins' database, available via the Internet. 16 The prospect is within a stratigraphic sequence not previously Later modeling performed by Veritas suggest that a high porosity penetrated by wells, but seismic correlation strongly suggests sandstone (>12%) and hydrocarbon filled interval could explain the main reservoir lies between the top of the Ben Nevis Fm and the low rigidity/incompressibility and also the low amplitudes the overlying Cenomanian unconformity, which means it is latest observed on the seismic. Albian to Early Cenomanian in age. The geological age of the seismic anomaly was detained using regional seismic control. Tie lines from both North Ben Nevis and White Rose clearly indicate that the Glenwood prospect is Tight Sands Carbonates located between the Ben Nevis Fm and the Cenomanian λ/µ Mu-Rho (rigidity) Unconformity. This suggests a Late Albian age for the submarine lowoil) ine ) ( (br fan and confinement within the Nautilus Fm of nd λ= 2µ top d sa Gas Sands fille Sands oil of nd se ba d sa Stratigraphically, the fan may be equivalent to the Eider oil fille /µ Formation as seen in the Whale Basin, which is the first oo nw ly d hig hλ Gle oma transgressive sand on the Aptian unconformity (McAlpine, 1990), Coal an Shale Inversion Lambda-Rho (incompressibility) A seismic inversion was done by CGG. The purpose of the seismic inversion was to substantiate the thickness modelling Figure 4: Lambda-Rho Vs Mu-Rho crossplots showing litholog. and to confirm an acoustic impedance drop for the Glenwood anomaly. An acoustic impedance drop was confirmed by the From the above observations, there is a positive case for inversion. The thickness of the low acoustic impedance interval assuming a hydrocarbon filled reservoir in the Glenwood was determined by measuring the time and depth converted prospect. However, the AVO attributes are closely associated using a velocity of 3600 m/s. An isochore map was then with the outline of the amplitude anomaly and no DHI's are contoured using the measured thickness and the amplitude map obvious on the seismic data. Leading us to two conclusions, first that the AVO is being influenced more by the porosity that AVO Modelling and Analysis the presence of Hydrocarbon or the reservoir is completely filled. The main objective of the AVO analysis was to determine the AVO response from a potential reservoir at Glenwood. Both the Depth Conversion modelling and AVO analysis where done by Veritas. Models The methodology that was chosen for depth conversion was the where prepared to determine the theoretical response of a wet, application of a linear velocity model down to Base Tertiary and oil and a gas filled reservoir. Based on the modelling results it then a second velocity function from Base Tertiary to the should be possible to distinguish between a water and character anomaly. hydrocarbon filled reservoir, but to distinguish between gas and oil an absolute calibration point is needed. The second velocity model down to the Glenwood anomaly proved to be difficult. Stacking velocities and linear velocity Wet function both gave unrealistic velocities for the interval. A velocity function based on the Base Tertiary to A Marker (or equivalent) interval for all wells in the Jeanne d'Arc Basin was Oil therefore chosen. The resulting interval velocity map shows a slight increase in velocity versus time to the anomaly. 40 API Oil Acknowledgements We wish to thank WesternGeco, CGG Canada Services Ltd, Gas Veritas Exploration Services, Norsk Hydro and Petro-Canada Oil and Gas for their permission to present this talk. References Fristad, T., Groth, A., Yielding, G., Freeman, B., 1997, Figure 3: Models showing differences in expected response. Quantitative Fault Seal Prediction: A Case Study from Oseberg Syd, Norwegian Petroleum Society (NPF), Special Publication No.7 McAlpine, K.D., 1990, Mesozoic Stratigraphy, Sedimentary The AVO analysis consisted of two steps. First AVO attributes Evolution, and Petroleum Potential of the Jeanne D'Arc were determined. Secondly, inversion of two AVO attributes, the Basin, Grand Banks of Newfoundland. Geol. Survey Canada expected P-Wave and S-wave reflectivity were extracted. The - Paper 89-17. results of the inversions were used to calculate rock properties. Poupon, M. & Palmer, G., 1999, Finding Channel Sands with Rock property (λρ incompressibility and µρ rigidity) and cross- Seismic Facies Analysis and Litho-seismic Modelling, plots were used to assess fluid content and lithology. Offshore Magazine March 1999 Reading, H.G. & Richards, M.R., 1994, Turbidite Systems in Cross-plots suggest low rigidity and incompressibility, which Deep-Water Basin Margins Classified by Grain Size and normally would be associated with shale or coal/tuff This is not Feeder System, Bull. AAPG, v 78, no 5, pp792-822 compatible with the morphology of the low impedance anomaly. 17 Estimating TIV Anisotropy Parameters in the Jeanne d’Arc Basin at White Rose H-20 Well Emery, David J., Brown, R. Jim, and Enachescu, Michael E. Summary In the summer of 1999 three delineation wells (L-08, A-17 and N-30) were drilled in the White Rose field of the Jeanne d’Arc Basin. Interpretation of the VSP and log information indicates that the seismic velocities increase reasonably linearly with depth within the Tertiary section followed by an abrupt velocity increase in the penetrated Cretaceous. Evaluation of anisotropy parameters using check shots and stacking velocities indicates a ~6% difference between v(0°) and v(90°) (Hedlin, 1999). During the summer of 2000, a single delineation well, H-20, was drilled into the northern part of the White Rose South Avalon Pool (figure 1). The H-20 well was targeted to penetrate the hydrocarbon bearing Avalon Formation through a small intra-reservoir fault with a throw of approximately 25 m. Due to the deviated nature of the H-20 well the VSP program was designed with a walkabove configuration. Rig time and equipment availability provided the opportunity to also acquire a single offset over a limited depth range. This paper deals with the evaluation of this limited VSP information for anisotropy parameters. The H-20 direct-arrival for the offset VSP indicates that the Tertiary age Banquereau Formation of Tertiary age is weakly anisotropic (δ = 0.03 ε = 0.07). Evaluation of the interval velocities derived from VSP shows a direct relationship between shale compaction and degree of anisotropy. The VSP interval velocity information also indicates that the Santonian age Dawson Canyon shales are highly anisotropic (δ = 0.40 ε = 0.80). Figure 1: White Rose H-20 Offset VSP located on a N-S 3-D seismic line VSP Source approx. 1000 m north of well VSP Receivers 2080-2500 m MD H-20 E-09 L-08 Ocean Bottom Velocity increase from 1850 to 2300 m/s Banquereau Fm. Unc. Eocene Unc. Unc. Base Tertiary Unc. Dawson Canyon Fm. oir Avalon Reserv 500 m Theory The evaluation of the White Rose H-20 VSP for rock anisotropy was performed using Thomsen’s (1986) equations for weak elastic anisotropy. The velocity of P-waves in such media is defined as: v(θ) = v0 (1 + δ sin θcos θ + ε sin θ) 2 2 4 (1) where v(θ), the P-wave velocity for a phase angle θ, is determined by the vertical velocity (v0) and the two anisotropy parameters δ and v ε (Thomsen, 1986, equation 16a). 18 A short-spread assumption is appropriate for both surface reflection seismic, where the effective offset is less than the depth of investigation, and for VSP data, where the offset is less than half the depth of investigation. Using Thomsen’s (1986) equation for a short- spread the δ parameter can be estimated using the NMO velocity: 2 δ = ½[(vNMO/v0) –1] (2) To estimate a range for ε, a least-squares solution was determined for a varying set of ε and δ values. The least-squares solution attempts to minimize the traveltime and NMO velocity differences using the straight-ray angle for the average phase angle and the isotropic ray length for raypath distance. Assuming that the majority of the anisotropy effects on the direct arrivals occur in the 2050 m interval above the shallowest geophone location in the offset VSP, the least-squares solution should represent the average anisotropy parameters for the Banquereau Formation. Two different assumptions were used to estimate the range of interval anisotropic parameters for the 420 m span in the offset VSP. The first approximation was for elliptical anisotropy (δ = ε) discussed by Daley & Hron (1977), because of its algebraic simplicity. Substituting ω for δ and ε in Thomsen’s equation results in: ω = [(v(θ) / v0) – 1] / sin θ 2 (3) The second estimation was for intrinsic anisotropy where ε > δ as proposed by Berryman (1979) and Thomsen (1986). Using a ratio of 2δ = ε, Thomsen equation can be modified by letting τ = δ = ε /2 resulting in: τ = ((v(θ) / v0) – 1) / (sin θ + sin θ) 2 4 (4) Results The White Rose H-20 VSP values for δ, vary from 0.029 to 0.048, with the Tertiary section (above 2230 m) having the lowest values (table 1). The stacking velocities from the 3-D surface seismic compares with the observed moveout velocity from the VSP. The corresponding log derived formation velocities are overestimated by 4-6%. Since vNMO is derived from the lithology above each sample point, and since there is no significant change in δ until a depth of 2230 is reached, it would be reasonable to assume that the Tertiary is only slightly anisotropic. The deviation in δ below 2230 m would indicate a sharp change in properties below this point and a significant increase in anisotropic parameters. The best fit for a least-squares error between the calculated and observed moveout velocities occurs at ε = -0.46 with a δ = +0.21. This solution falls far outside the expected anisotropy parameter range for the White Rose area. The solution for elliptical anisotropy (ω = δ = ε = 0.04) is closer to the expected range and is situated at a local minimum. Around this solution is a group of solutions from δ = 0.02 ε = 0.10 to δ = 0.04 ε = 0.03 with approximately the same degree of error. The answer of δ ≅ .03 ε ≅ .07 falls at the center point of this cluster of solutions and represent the average anisotropy parameters in the Banquereau Formation. 19 Table 1: H-20 VSP summary Average Average Average Minimum Maximum Group Interval Velocity v (NMO) v (NMO)^ δ δ Angle Tertiary Shale 2050 to 2220 2168 2179 2236 0.028 0.036 25.11 Shale 1 2050 to 2120 2158 2171 2228 0.029 0.036 25.57 Shale 2 2130 to 2220 2176 2187 2242 0.028 0.034 24.69 Eocene Peak 2230 to 2250 2188 2200 2259 0.029 0.036 23.54 Paleocene 2260 to 2370 2208 2219 2282 0.032 0.038 23.32 Dawson Canyon 2400 to 2460 2243 2257 2341 0.042 0.048 22.33 ^ NMO velocity estimated using raypath length from isotropic GXII model Interval Offset Phase Elliptical Intrinsic 2δ = ε Estimate Velocity Velocity Angle ε=δ δ ε Error* Tertiary Shale 2694 2800 31.18 0.147 0.116 0.232 0.06 Shale 1 2632 2684 31.60 0.072 0.056 0.113 0.14 Shale 2 2745 2845 30.80 0.139 0.110 0.220 0.12 Eocene Peak 2846 3686 29.13 1.246 1.007 2.015 0.69 Paleocene 2994 3163 29.32 0.235 0.190 0.380 0.11 Dawson Canyon 3726 4485 37.00 0.562 0.413 0.826 0.24 D Canyon (GXII) 3726 6376 37.28 1.938 1.418 2.836 0.24 * Estimated Error represents a 1ms time shift and the effects on the calculated Elliptical parameters Evaluating the interval velocity behavior provides insights into the discrete anisotropy parameter of individual geological layers. The velocity in the Tertiary sequences is relatively linear with crossover in the velocity display (figure 2). This may indicate that the Tertiary is predominately isotropic with the net anisotropic effect occurring because of the intrinsic layering. Assuming elliptical anisotropy the Tertiary would have a ω = 0.15 which is higher than the δ determined from vNMO values. Using the anisotropy relationship 2δ = ε the results for the Tertiary would be ε = 0.23 with a δ of 0.11 or higher than the values determined from the average velocities. The high velocity anomaly from 2230 to 2260 m feature a sharp increase in anisotropy parameters to ω = 1.2 or ε = 2.0 and δ = 1.0. The anisotropy parameters are unexpectcally high. The sonic and density logs indicate only a small incremental change in rock properties where the surface seismic indicates a relatively strong impedance response. While, geologically this interval corresponds with the Eocene unconformity, the high velocity obtained for the interval is most likely the effect of time sampling errors (figure 2). The Paleocene shales (2260 to 2370) are just slightly deformed and have undergone significant compaction with a v(0) = 2994 m/s and ω = 0.23 or ε = 0.38 and δ = 0.19. The Paleocene shale interval has a consistent velocity shift and an constant set of anisotropy parameters. This interval appears to be moderately anisotropy. At 2380 m there is a pronounced shift in the interval velocities and calculated anisotropy parameters. This shift correlates with the Base Tertiary unconformity and marks the change from parallel layering to moderately faulted and folded geometry. The underlying Dawson Canyon Formation is made up of compacted Cretaceous shale and limestones with dips varying from 3° to 10°. The GXII raypath-model solution indicates a substantial offset velocity (7450 m/s) and an unrealistic anisotropy effect ω = 1.9. While the ray-traced data may need checking, the straight-ray approximation still demonstrates a significant anisotropy effect ω = 0.56 or ε = 0.83 and δ = 0.41. While the investigated Dawson Canyon interval is limited on the offset VSP, it would be reasonable to conclude that these Senonian shales are strongly anisotropic (ε > 0.4). 20 Velocity (m/ms) 2000.0 3000.0 4000.0 5000.0 2000 Velocity Walkabove (Check Shot) Velocity 1000m Offset (Straight Ray) Velocity 1000m Offset (model) 2100 Tertiary Shale v0 = 2700m/s group angle =25º phase angle = 31º ω = δ = ε = 0.15 or ε = 0.23, δ = 0.12 2200 Depth (meter SS) Eocene Unconformity Paleocene Shale 2300 v0 = 3100 m/s group angle =23º phase angle = 29º ω = δ = ε = 0.23 ε = 0.38, δ = 0.19 Base Tertiary Unc 2400 Dawson Canyon Shale v0 = 3900 m/s group angle = 22º phase angle = 37º ω = δ = ε = 0.56 ε = 0.83, δ = 0.41 2500 Figure 2: White Rose H-20 Interval Velocity Profile - The blue line represents the velocities for the zero offset VSP, the red line for the 1000-m offset VSP using a straight line geometry and yellow using the raypath length from the GXII model. 21 Conclusions Previous work done by Hedlin (1999) indicated that the Banquereau Formation Tertiary shales are slightly anisotropic with an approximate 6% difference between v(0°) and v(90°). The offset VSP average velocities agree with this observation and gives a range of anisotropy parameters from δ = 0.02 and ε = 0.10 to δ = 0.04 and ε = 0.03. Since the White Rose VSP represents a short spread geometry, the dominant Thomsen (1986) parameter should be δ because of its control on wavefront shape from 0° to 30º. This is reflected in the limited anisotropy parameter range estimated from the VSP (δ = 0.029 to 0.048). The least square solution of ε = 0.07 with a δ = 0.03 would be consistent with the both surface 3-D seismic and VSP data. The interval velocities indicate that the shales in the White Rose area have varying degrees of anisotropy parameters. The accuracy of measuring the anisotropy parameters of these shales at White Rose, is unfortunately limited, since only a single offset was used. The interval anisotropy parameters indicate that the lower portion of Banquereau Formation is weakly anisotropic (ω = 0.15 or ε = 0.23 with δ = 0.11), the Paleocene shales are moderately anisotropic (ω = 0.23 or ε = 0.40 and δ = 0.19) and the Dawson Canyon Formation is strongly anisotropic (ε > 0.4). Acknowledgments Special thanks to Larry Mewhort and Ken Hedlin for their technical evaluation and application assistance, Larry Mayo for design and acquisition of the VSP. References Berryman, J.G., 1979, Long-wave elastic anisotropy in transversely isotropic media: Geophysics, 44, 896-917. Brown, R.J., Lamoureux, M.P., Slawinski, M.A. and Slawinski, R.A., 2000, Direct traveltime inversion of VSP data for elliptical anisotropy in layered media: CREWES Research Report, 12, 19-34. Daley, P.F. and Hron, F., 1977, Reflection and transmission coefficients for transversely isotropic media: Bull. Seis. Soc. Am. Vol 67, 661-675. Enachescu, M.E., 1987, Tectonic and structural framework of the northeast Newfoundland continental margin in Beaumont, C. and Tankard, A. J., Eds., Sedimentary basins and basin-forming mechanisms: Can. Soc. Petr. Geol., Mem. 12, 117-146. Enachescu, M.E., 1988, Extended basement beneath the intracratonic rifted basins of the Grand Banks of Newfoundland: Can. J. Expl. Geophys., 24, No. 1, 48-65. Grant, A.C. and McAlpine, K.D., 1990, The continental margin around Newfoundland in Keen, M.j. and Williams, G.L., Eds., Geology of the Continental Margin of East Canada: Geol. Surv. Can., Geology of Canada, No. 2, 239-292. Hedlin, K., 1999, White Rose time-depth analysis: Husky Oil Operation Ltd, internal report. Thomsen, L., 1986, Weak elastic anisotropy: Geophysics, 51,1954-1966. Schlumberger, 2000, White Rose H-20 VSP report: Husky Energy, internal report. 22 Imaging Through Gas Clouds: A Case History In The Gulf Of Mexico S. Knapp1, N. Payne1, and T. Johns2, Seitel Data1, Houston, Texas: WesternGeco2, Houston, Texas Summary Results from the worlds largest 3D four component OBC seismic survey will be presented. Located in the West Cameron area, offshore Gulf of Mexico, the survey operation totaled over 1000 square kilometers and covered more than 46 OCS blocks. The area contains numerous gas invaded zones and shallow gas anomalies that disturb the image on conventional 3D seismic, which only records compressional data. Converted shear wave data allows images to be obtained that are unobstructed by the gas and/or fluids. This reduces the risk for interpretation and subsequent appraisal and development drilling in complex areas which are clearly petroleum rich. In addition, rock properties can be uniquely determined from the compressional and shear data, allowing for improved reservoir characterization and lithologic prediction. Data acquisition and processing of multicomponent 3D datasets involves both similarities and differences when compared to conventional techniques. Processing for the compressional data is the same as for a conventional OBC survey, however, asymmetric raypaths for the converted waves and the resultant effects on fold-offset-azimuth distribution, binning and velocity determination require radically different processing methodologies. These issues together with methods for determining the optimal parameters of the Vp/Vs ratio (γ) will be discussed. Finally, images from both the compressional (P and Z components summed) and converted shear waves will be shown to illuminate details that were not present on previous 3D datasets. Introduction Multicomponent 3D surveys have become one of the leading areas where state of the art technology is being applied to areas where the exploitation potential has not been fulfilled due to difficulties in obtaining an interpretable seismic image (MacLeod et al 1999, Rognoe et al 1999). In areas where conventional datasets suffer from poor data quality due to the presence of gas and/or fluids in the pore spaces, it has been shown that ocean-bottom acquired multicomponent surveys have improved images (Thomsen, et al 1997, Granli et al 1999). For marine surveys where the energy source is towed just below the sea surface, the initial downgoing wave is always compressional and mode conversion to shear waves takes place at reflector boundaries (Thomsen 1999). For clarification, this paper will describe compression as P-P waves and converted shear as P-S waves, indicating one mode conversion. The primary difference between the two is that the shear waves are unaffected by the presence of gas or fluids and only detect lithologic variations. The resultant converted shear seismic image, therefore, contains interpretable reflectors that were not present on the compressional data. Acquisition Over a six-month period from December 1999 until May 2000 a large multicomponent seismic survey was acquired in the West Cameron area, Gulf of Mexico. This was the first commercial application of a new ocean bottom cable and sensor package designed specifically for multicomponent applications. Cables were deployed under tension and the no-slack design with 50-meter group intervals helped improve the correspondence between preplotted and actual location. Receiver positioning was confirmed using both a high frequency acoustic method and a first break based technique. The survey was acquired using swath style acquisition comprised of 102 swaths in 9 overlapping zones for a total of 18,700 CMP kilometers and 930 square kilometers of full fold coverage. The area included numerous platforms and subsurface obstructions requiring extensive planning, especially to ensure sufficient fold of coverage for both the compression and shear data. Commercial ship traffic caused both noise problems and an operational hazard for vessel maneuvers. Processing Both field processing and conventional shore-based processing were utilized in this survey. In order to ascertain data quality and permit preliminary evaluation of the imaging results, displays ranging from shot records to 3D inlines, crosslines and timeslices were created onboard using optimized processing parameters that were determined from a 2D test line that was acquired and processed previously. The shot records with the navigation merged were then brought onshore and proceeded through a complete isotropic processing sequence at Schlumberger’s Houston office. This included creation of both a P-P wave data volume by combining the hydrophone and geophone using a non-linear summation technique (Moldoveanu, 1996) and a radial and transverse volume from the P-S data recorded on the two horizonal geophones. Deliverables include pre stack migrated gathers and migrated volumes for both P-P and P-S wave modes. The large size of the survey resulted in several modifications to the basic P-S converted wave processing sequence (Brzostowski et al 1999, McHugo et al 1999). Resolution of the receiver statics and determination of the appropriate binning Vp/Vs ratio (γVTI) as discussed in Thomsen, 1999 proved to be challenging given the volume of data. We were frequently evaluating the length of the error bar, then iterating to achieve a result which would accomplish our primary objective: imaging through the gas areas. A demonstration of the spectacular 23 results can be seen in comparisons between the compressive and radial shear data volumes. Figure 1 is an inline from a portion of the survey where the P-P image, shown on the left, is distorted by the presence of a large gas cloud. The equivalent section from the P-S volume shows the dramatic increase in reflection events, and hence the interpretability of the data. Figure 2 is a crossline that transects the same gas cloud and shows a similar high quality image that can now be used to reduce the risk for further exploration. There are multiple areas throughout the survey where this level of improvement can be demonstrated by comparing timeslices through the two data volumes (Figure 3). Post stack versus pre stack migrated results for both P-P and P-S waves will be compared to identify improvements in data quality with the added benefit of providing gathers suitable for AVO analysis. Future Work As can be seen by these examples the results are very encouraging, however, further analysis is necessary to fully realize the added value of this dataset. Special processing to determine the effect of azimuthal anisotropy and depth domain imaging would be of interest. A definitive evaluation of the obvious benefits versus potential inadequacies from limited azimuth acquisition needs to be addressed. References Brzostowski, M., Altan, S., Zhu, X., Barkved, O., Rosland, B. and Thomsen, L., 1999, 3-D converted-wave processing over the Valhall Field: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 695-698. Granli, J. R., Arntsen, B., Sollid, A. and Hilde, E., 1999, Imaging through gas-filled sediments using marine shear-wave data: Geophysics, 64, no. 3, 668-677. MacLeod, M.K., Hanson, R.A., Bell, C.R., and McHugo, S., 1999, “The Alba Field ocean bottom cable seismic survey: Impact on development,” SPE Paper 56977, Offshore Europe Conference, Aberdeen. McHugo, S., Probert, A., Hodge, M., Kristiansen, P., Painter, D. and Hadley, M. E. A., 1999, Acquisition and Processing of a 3D Multicomponent Seabed Data from Alba Field - a Case Study from the North Sea : 61st Mtg. Eur. Assoc. Expl Geophys., Extended Abstracts, European Association Of Geophysical Exploration, Session:6026. Moldoveanu, N., 1996, US Patent and Trademark Office, “Method for attenuation of reverberations using a pressure-velocity bottom cable,” Patent No. 5621700. Rognoe, H., Amundsen, L. and Kristensen, A., 1999, Improved structural and stratigraphic definition from 3-D/4-C data - Statfjord Field: Annual Meeting Abstracts, Society Of Exploration Geophysicists, 736-739. Thomsen, L. A., Barkved, O., Haggard, B., Kommedal, J. and Rosland, B., 1997, Converted-wave imaging of Valhall reservoir: 61st Mtg. Eur. Assoc. Expl Geophys., Extended Abstracts, European Association Of Geophysical Exploration, Session:B048.B048. Thomsen, L., 1999, Converted-wave reflection seismology over inhomogeneous, anisotropic media: Geophysics, 64, no. 3, 678-690. 24 P-P Section P-S Section (Radial) Figure 1: Inline migration comparison P-P Section P-S Section (Radial) Figure 2: Crossline migration comparison 25 Figure 3: Timeslice comparison 26 Neural Network Applications to Porosity Prediction; Hebron Field Offshore Newfoundland, Canada Stan M Volk - ExxonMobil Development Company, John D. Logel - Petro-Canada Oil and Gas, Craig Coulombe - Chevron Resources Canada Introduction The Hebron Field is located approx. 350 kilometers SSE of St. John's Newfoundland. The filed has three pay zones and an estimated 2 billion barrels in-place. The advent of seismic attributes has allowed more information to be extracted from the seismic data. A method of determining which information to use is highly valuable. Neural networks can be employed to isolate the attributes, which in combination can best predict desired log responses. Neural networks were employed in an attempted to predict the porosity prediction in these three pay zones. Methodology Porosity was predicted from five porosity curves in the field using linear regression and neural networks. Attributes were calculated from seismic data, impedance, coherency, and velocity cubes. Numerous iterations were made using predictive and multi-layered feed forward networks. A best fit solution was generated for each pay zone. Conclusion A neural network application enhanced the porosity distribution prediction and increased the correlation by an average of 30 percent for all three pay zones. Comparisons to seismic inversions also showed a marked improvement using neural networks. 27 Depth Conversion methodologies, uncertainty, quantification and application; Hebron Field, Offshore Newfoundland Canada John D. Logel - Petro-Canada Oil and Gas Craig Coulombe - Chevron Resources Canada Stan Volk - ExxonMobil Canada Introduction The Hebron field is located 350 kilometers SSE of St. Johns, Newfoundland. The field has three pay zones and estimated 2 Billion Barrels in-place. The field has very subtle relief and some of the accumulations are very sensitive to velocity determination and depth conversion. The field was depth converted using two different methods to better quantify uncertainty, and because neither method could be proved to be superior to the other. Methodology Seismic time to depth conversion is a process to convert from recorded seismic travel times to depth. This process is carried out to determine drilling depth, reservoir understand and structural validity. The process is to build a detailed velocity model that is robust enough to allow for lateral and vertical variations, that ties the wells and does not induce anomalies or artifacts. The process can be done a number of different ways from very simply extrapolating time/depth tables from wells to pre-stack coherency inversion and tomography. The basic data can include strictly well data or a combination of well and seismic processing velocities. The difference being in the time and cost expenditure. The methods used for the Hebron field mapping was to use a layercake based method that employed both a tied and smoothed stacking velocity based model and a well based method that would account for the instantaneous change of velocity as a function of depth. This change is expressed as an acceleration value (k) and surface intercept of (V0). Quantifying errors The use of two equal and “valid” velocity models continued to be a source of uncertainty and question. The differences were small, as shown in the difference map in Figure 1. But in some accumulations, accounted for over 100 million barrels. To try and resolve the problem and converge on a single model, other source of data were incorporated. These included average and interval velocity mapping, cross- validation with well control, fault seal and juxtaposition analysis, hydrocarbon contact correlation and pre-stack depth tomography. Conclusion These methods were integrated together in an overall velocity model building and used independently and in combination to try and encompass a wider range of gross-rock-volume (GRV) uncertainty. 28 Zero contour Figure 1. Difference map between Stacking velocity based model and Instananeous Velocity Model 29
Pages to are hidden for
"By Leave of Poseidon"Please download to view full document