Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9

Geological Society, London, Special Publications Online First

The importance of natural fractures in a tight reservoir for potential CO 2 storage: a case study of the upper −middle Kapp Toscana Group (, Arctic )

K. Ogata, K. Senger, A. Braathen, J. Tveranger and S. Olaussen

Geological Society, London, Special Publications v.374, first published September 10, 2012; doi 10.1144/SP374.9

Email alerting click here to receive free e-mail alerts when service new articles cite this article Permission click here to seek permission to re-use all or request part of this article Subscribe click here to subscribe to Geological Society, London, Special Publications or the Lyell Collection How to cite click here for further information about Online First and how to cite articles

Notes

© The Geological Society of London 2012 Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9

The importance of natural fractures in a tight reservoir for potential CO2 storage: a case study of the upper Triassic–middle Jurassic Kapp Toscana Group (Spitsbergen, Arctic Norway)

K. OGATA1*, K. SENGER1,2,3, A. BRAATHEN1, J. TVERANGER2 & S. OLAUSSEN1 1University Centre in (UNIS), PO Box 56, 9171 Longyearbyen, Norway 2Centre for Integrated Petroleum Research (CIPR), Uni Research, Alle´gaten 41, 5020 Bergen, Norway 3Department of Earth Science, University of Bergen, Alle´gaten 41, 5020 Bergen, Norway *Corresponding author (e-mail: [email protected])

Abstract: In the Longyearbyen CO2 laboratory project, it is planned to inject carbon dioxide into a Triassic–Jurassic fractured sandstone–shale succession (Kapp Toscana Group) at a depth of 700– 1000 m below the local settlement. The targeted storage sandstones offer moderate secondary por- osity and low permeability (unconventional reservoir), whereas water injection tests evidence good lateral fluid flow facilitated by extensive fracturing. Therefore, a detailed investigation of fracture sets/discontinuities and their characteristics have been undertaken, concentrating on the upper reservoir interval (670–706 m). Datasets include drill cores and well logs, and observations of out- crops, that mainly show fracturing but also some disaggregation deformation bands in the sand- stones. The fracture distribution has a lithostratigraphical relationship, and can be subdivided into: (A) massive to laminated shaly intervals, offering abundant lower-angle shear fractures; (B) massive to thin-bedded, heterogeneous, mixed silty–shaly intervals, with a predominance of non-systematic, pervasive bed-confined fractures; and (C) massive to laminated, medium- to thick-bedded, fine- to coarse-grained sandstones with a lower frequency of mostly steep fractures. These domains represent pseudo-geomechanical units characterized by specific fracture sets and fracture intensity, with indicated relationships between the bed thickness and fracture intensity, and with domains separated along bedding interfaces. We discuss the impact of these lithostruc- tural domains on the fluid flow pathways in the heterolithic storage unit.

In the Longyearbyen CO2 laboratory project (LYB demonstrating the complete carbon value chain CO2 lab) the possible depth interval for an injection (Braathen et al. 2010). of carbon dioxide is a siliciclastic unit located at a In order to investigate the suitability of the depth of 700–1000 m in the subsurface beneath planned sequestration site, the Longyearbyen CO2 the High Arctic community of Longyearbyen, Sval- lab has drilled and cored four wells during the bard (Fig. 1). The project, initiated in 2007 by the period of 2007–2010 (the drill site locations are University Centre in Svalbard (UNIS), involves shown in Fig. 2), two of which penetrated the reser- partners from both academia and industry (Braathen voir (Dh2 and Dh4: Fig. 3). The target reservoir has et al. 2010; Baelum et al. 2012). The principal aim no mapped closure and it crops out approximately is to utilize the proximity of suitable reservoir rocks 15 km NE of the planned injection site. However, to store CO2 from the local coal-fuelled power borehole analyses on the Dh4 well recorded more plant. Research focuses on understanding the behav- than 50 bar underpressure at 870 m (Braathen et al. iour of CO2 in the subsurface while at the same time 2010) showing the presence of an efficient trap, contributing to the reduction of local emissions of probably either a stratigraphical or a combined CO2. The coal-powered co-generation 10 MW heat stratigraphical–structural trap, and supporting the and power plant, located only 5 km from the plan- sealing efficiency of the cap rock. Although there is ned injection site, emits approximately 64 000 no clear evidence of large faults or major structural tonnes of CO2 annually, from the combustion of features in the area close to the planned injection about 26 000 tonnes of locally mined coal. The pro- site, the reservoir interval is vertically sandwiched ximity of the energy source (i.e. coal mines), power between two main de´collement zones. These zones plant and injection site, in combination with the relate to the Palaeogene contractional event respon- fact that Longyearbyen is a nearly closed energy sible for the development of the transpressional fold- system, makes it a highly suitable candidate for and-thrust belt that characterizes the western part of

From:Spence, G. H., Redfern, J., Aguilera, R., Bevan, T. G., Cosgrove, J. W., Couples,G.D.&Daniel, J.-M. (eds) 2012. Advances in the Study of Fractured Reservoirs. Geological Society, London, Special Publications, 374, http://dx.doi.org/10.1144/SP374.9 # The Geological Society of London 2012. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

BFZ North Pole CANADA Sevemaya Zemlya RUSSIA

LFZ Arctic Ocean B Franz Josef Land RUSSIA Greenland DENMARK A Greenland Novaya Sea Zemlya RUSSIA Svalbard ISFJORDEN LYB CO 2 NORWAY Lab BFZ

LFZ Barents Sea

Arctic Circle NORWAY RUSSIA ICELAND SWEDEN a BELLSUND

b N

50 km Hornsund Fault Zone HORNSUND

LFZ Western hinterland

BFZ Basement-involved SØRKAPP N fold-thrust complex Thin-skinned fold-thrust belt 150 km Eastern foreland province c

Basement-involved Western hinterland fold-thrust complex Central Zone Eastern foreland province Hornsund Fault Zone West BFZ East LYB CO 2 LFZ Lab Cp Cp Tr Ca A Cp JC JC JC B Tr T 0 Cp 0 Cp Ca Cp Ca Ca D D D

50 km d

Fig. 1. Location and tectonic setting of the Svalbard archipelago and position of the Longyearbyen CO2 lab site (LYB CO2 Lab). The planned injection target, the Kapp Toscana Group, is sandwiched between the upper and middle de´collement zones associated with the West Spitsbergen fold-thrust belt, a direct result of the Palaeogene transpressional tectonics. (a) The location of the Svalbard archipelago in the context of the North Atlantic. The archipelago lies within the High Arctic, approximately half way between northern Norway and the North Pole. (b) Main structural lineaments on Svalbard, aligned predominantly NNW–SSE. (c) Tectonic overview map of Spitsbergen, illustrating the West Spitsbergen fold-and-thrust belt (modified from Bergh et al. 1997). (d) Schematic cross-section across Spitsbergen, emphasizing the LYB CO2 lab targeted reservoir/storage unit. The cross-section’s location is marked in (c). Figure modified after Bergh et al. (1997). BFZ, Billefjorden Fault Zone; LFZ, Lomfjorden Fault Zone; D, ; Ca, ; Cp, ; Tr, Triassic; JC, Jurassic–. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE

Svalbard (see Fig. 1c, d) (Bergh et al. 1997; Braathen drifting northwards during the deposition of the et al. 1999). Based on vitrinite data (Throndsen Mesozoic succession. The opening of the North 1982), the reservoir is thought to have experienced Atlantic led to dextral movement along the De maximum burial of approximately 4.5 km in depth Geer Zone (i.e. the Hornsund Fault Zone) between during the Eocene, prior to an estimated uplift of Svalbard and East Greenland during the Palaeogene about 3.5 km in the Oligocene–latest Neogene. (Eldholm et al. 1987; Braathen et al. 1995; Leever However, a number of workers using additional et al. 2011). This caused oblique compression and methods consider this to be an overestimation (see the development of the West Spitsbergen fold-and- the work of Blythe & Kleinspehn 1998 for an exten- thrust belt (WSFB), and the consequent Terti- sive review). Rapid unroofing during Quaternary ary CSB, a foreland to wedge-top basin comprising times is attributed to the widespread glaciations in an infill of mixed continental and marine clastics the area (Ingo´lfsson 2011). The deep burial led to (Steel et al. 1985; Braathen et al. 1997; Helland- mechanical compaction with dense grain packing as Hansen 2010). The development of the WSFB dur- well as extensive quartz cementation (Braathen et al. ing the Palaeocene–Eocene generated a series of 2010). Matrix permeability in Dh4, as measured eastwards-extending thrust sheets soled by de´col- from drill core plugs, is typically less than 2 mD lement zones along evaporite and shale intervals (where1 mD ¼ 10215 m2),withporosityvaluesvary- (Bergh et al. 1997). This folding and thrusting can ing from 8 to 18% (Farokhpoor et al. 2010, 2011). be linked to a significant fracturing. However, Late Despite these low measured permeabilities, three Cretaceous magmatic dyke–sill intrusion (Mina- water injection tests confirmed the good injection kov et al. 2012) and Quaternary unroofing and potential of the target reservoir. A 5 day injection decompaction probably also added fractures to the test conducted in August 2010 suggested an aver- succession. age flow capacity of 45 mD . m in the lowermost part of the reservoir (870–970 m: Braathen et al. Stratigraphical development 2010). The injectivity is thought to be primarily a function of the natural fractures identified both The targeted storage unit is an aquifer made up of a in the drill cores and in the outcrops. Accordingly, siliciclastic succession belonging to the upper Trias- a detailed understanding of the fracture systems sic–middle Jurassic Kapp Toscana Group. It com- observed in the reservoir is essential for predicting prises the De Geerdalen Formation and elements the behaviour of CO2 and other fluids in the reservoir. of the overlying Wilhelmøya Subgroup (Figs 2 & In this contribution we investigate the upper part 3) (Worsley 1973, 2008; Knarud 1980; Mørk et al. of the reservoir succession, presenting a case study 1982; Harland & Geddes 1997; Mørk & Worsley of selected examples from an extensive dataset of 2006). The 270 m-thick drilled section of the De natural fractures obtained from both borehole and Geerdalen Formation includes shales, siltstones outcrop data. When fully processed, the data will and sandstones deposited in a near-shore, paralic be utilized in fluid flow modelling. At this stage, (lagoonal–delta plain) environment with net sand- our aim is to highlight the importance of natural stone of approximate 30% (net/gross (N/G) ratio fractures when viewed as potential CO2 storage on of c. 0.3). Most palaeogeographical reconstructions Spitsbergen, as well as to illustrate the need for well- suggest a NW-prograding deltaic system (Mørk constrained and reliable outcrop data as input to the et al. 1982; Steel & Worsley 1984; Riis et al. 2008; reservoir modelling and simulation. Glørstad-Clark et al. 2010). The overlying 20 m- thick Norian–Bathonian Knorringfjellet Formation of the Wilhelmøya Subgroup is interpreted as a con- Geological setting densed siliciclastic unit (i.e. remanie´), probably a lag resulting from coastal reworking of deltaic sedi- Tectonic development ments (Mørk et al. 1982; Harland & Geddes 1997; Nagy et al. 2011). These units are overlain by a The Svalbard archipelago is an uplifted part of the 250 m-thick shale unit of the middle–upper Juras- NW margin of the Barents Shelf (Harland 1997). sic Agardhfjellet Formation and the 200 m-thick, Our study area is situated on the NW margin of the predominantly shaly, Lower Cretaceous Rurikfjellet Central Spitsbergen Basin (CSB), a major synclinal Formation: together these two units form the cap feature (see Fig. 1c, d). The CSB’s boundaries align rock of the reservoir/storage unit (i.e. ‘lower aqui- to the predominant NNW–SSE structural grain fer’) (Fig. 3). The Upper Jurassic Agardhfjellet For- prevalent on Svalbard (Nøttvedt et al. 1993). Follo- mation is widespread regionally and its time- wing the Caledonian Orogeny, assumed Devonian equivalent stratigraphical interval represents the crustal-scale extension and localized Carboniferous major source, and commonly the main cap rock, rifting, Svalbard evolved into a stable platform of many of the oil and gas fields on the Norwegian during the late Carboniferous–Mesozoic time, continental shelf (Spencer et al. 2008). Above the Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE cap rock, an approximately 60 m-thick deltaic unit, † asperity – wall-rock morphology along the dis- the Barremian Helvetiafjellet Formation, comprises continuity surface; an ‘upper aquifer’. Above this, the 120 m-thick over- † wall coatings and infillings – solid material burden consists of a 60 m-thick, Aptian–Albian, occurring as coatings or filling along the discon- mixed sandstone–shale unit of shallow-marine tinuity surface. to inner-shelf affinity belonging to the Carolinefjel- The differentiation between natural and drilling- let Formation, and a 60 m-thick interval of Quatern- induced fractures (excluded from the final count) ary and Holocene fluvial sand/gravel and marine was performed taking into account: (1) the overall clay. The thickness of the permafrost in the area appearance (e.g. fresh v. weathered); (2) their conti- is estimated to be approximately 100 m (Humlum nuity across the core (e.g. through-going v. core- et al. 2003). confined, point of origin/termination within v. outside the core); (3) the occurrence of mineraliz- Methods ation/coatings; (4) anomalous trends (e.g. abrupt hooking toward pre-existing fractures and core Borehole analyses boundaries; and (5) the matching/mismatching of the fracture sides (Kulander et al. 1977). As the An optical televiewer was used to characterize retrieved cores were not oriented, azimuth infor- the fractures between 442 and 706 m, whereas the mation is only available from televiewer data in narrow drill hole diameter (46 mm) in the deeper the uppermost 39 m of the reservoir interval (667– parts of the borehole prevented data acquisition 706 m total depth (TD)). from deeper levels (Elvebakk 2010). Detailed (1:10 scale) structural logging of the reservoir section (668–970 m) in the drill cores of Outcrop analyses Dh4 was carried out to describe the physical char- To complement the borehole information, fieldwork acteristics and frequency distribution of natural dis- was conducted at Deltaneset in Central Spitsber- continuities (as defined by Schultz & Fossen 2008), gen, where the reservoir interval reaches the surface primarily distinguishing between closed and open (for the location see Fig. 2). Natural fractures were ones, with a focus on fractures/cracks. The recog- counted along scanlines, allowing the construction nition of true pre-coring features has not been easy of frequency plots along individual intervals. The because of the high degree of decompaction and orientation of each fracture was measured, using drilling-induced stresses deforming the cores. The both a geological compass and a GeoClino digital characterization of discontinuities includes the fol- clinometer. As for the cores, natural fractures were lowing standard parameters (see Singhal & Gupta classified based on their mesoscale characteristics 2010 and references therein): (see earlier), including in this case observations of † orientation – spatial attitude and geometrical their outcrop-scale vertical continuity (bed-confined features (e.g. strike, dip-angle); v. through-going). Furthermore, stratigraphical log- † mid-point depth – vertical position of the dis- ging (1:50 scale) was conducted throughout the continuity half-way point within the well; Knorringfjellet Formation and the upper third of † spacing – perpendicular distance between adja- the De Geerdalen Formation in order to correlate cent discontinuities of the same set; the outcrop data to the borehole data. Scanlines were † persistence – minimum relative length of the surveyed along various stratigraphical intervals and discontinuity trace; within a range of different lithologies in order to † linear density – number of discontinuities per capture potential variations in the fracture character unit length; related to . Strike data are corrected for a † connectivity – intersection and termination magnetic declination of 68 east of true north. characteristics of discontinuities; A total of 59 scanlines were collected during the † relative aperture – relative perpendicular distance field campaign. In the present paper, 13 representa- between the two margins of the discontinuity; tive scanlines for the three major lithostructural

Fig. 2. Location of the study area within the context of the regional geology of Central Spitsbergen. The inset map shows the study area at Deltaneset and the location (circled) of the stratigraphical log presented in Figure 7; the approximate grid co-ordinates of the stratigraphical log tying the various scanlines are: 0521174 East, 8697326 North (UTM Zone 33X, projection ED50). The cross-section illustrates the open nature of the reservoir succession, which rises to the surface in the NE, as well as the relationship of the drill site (Dh4) with the Deltaneset outcrop locality. Redrawn and modified after Major et al. (2001). Shapefiles of the geological units from the NPI-Geonet project are used. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE associations (LSAs) are presented (Table 1). These Successions of mudstone and shale, ranging in lithostructural domains or units are primarily thickness from metres to tens of metres, host the defined on the basis of the considered lithological majority of fractures. Fractures preferentially occur interval: (A) massive to laminated shaly intervals: within fine-grained lithologies (massive to lami- (B) massive to thin-bedded, heterogeneous, mixed nated and interbedded shales and siltstones) and, silty–shaly intervals; and (C) massive to laminated, among these, low-angle (08–458), usually slicken- medium- to thick-bedded, fine- to coarse-grained sided and non-mineralized fractures predominate. sandstones and conglomerates. Each domain is fur- High-angle (.458) fractures and veins preferably ther characterized by a distinct occurrence of preva- occur within coarser-grained units (fine to coarse lent systematic and non-systematic fracture sets, sandstones). The other closed (with infillings) preva- sedimentary facies, bed thicknesses and degree of lent discontinuity types recognizable within coarser- cementation. These units are regularly distributed grained lithologies are sharp hairline microfaults throughout the sedimentary succession and reflect (mostly normal) with very-fine-grained material and contrasting rheological behaviour, forming distinct calcite fillings, and disaggregation deformation intervals characterized by a specific type of fractur- bands (Fossen et al. 2007; Fossen 2010), usually ing and, thus, inferred to represent proxies of geo- bearing a phyllosilicate core. The fine-grained, infill- mechanical units (Shackleton et al. 2005). ing material can be interpreted as a shale gouge or Apart from the generalized scree/soil coverage membrane, possibly also related to the localized of the outcrops, we also took the frost–defrost dissolution of the cataclastic material into clay min- cycles and glacier dynamics-related slope unloading erals. Some of these features are millimetre- to into account as possible factors influencing fracture centimetre-thick, tabular deformation zones charac- patterns and apertures. terized by granular flow (no grain breakage), taken as evidence of soft-sediment deformation (e.g. folded laminae, fluidal structures). Deformation bands Fracture characterization comprise up to 30% of counted discontinuities in Borehole data medium-coarse sandstones. Mineralized fractures (mainly high-angle, calcite- and pyrite-filled veins) The reservoir interval in Dh4, from 672 to 970 m, seem to concentrate in coarse-grained lithologies, exhibits relatively high frequencies of both vertical especially within and around crystalline rocks (e.g. and horizontal sharp and tabular discontinuities doleritic intrusions) and carbonate-rich beds (e.g. (Schultz & Fossen 2008), particularly fractures and bioclastites). Open fractures predominantly display deformation bands. Manual counting identified 870 dip angles of less than 458 and are mostly hosted in individual, predominantly low-angled, and subordi- fine-grained lithologies. Less than 30% of open frac- nately high-angled, open fractures in this interval tures exhibit steep (.458) dips. These are concen- (Fig. 4). In the upper part of the reservoir (between trated in coarse-grained lithologies. 672 and 706 m TD) the acoustic televiewer iden- In general terms, high-angle fractures are predo- tified 35 fractures, whereas for the same interval minantly represented by Mode I fractures (i.e. joints), the manual logging revealed at least 108 fractures whereas low-angle ones belong to Mode II and III (Fig. 5). This incongruity is probably related to the fractures (i.e. shears/tears), with some evidence of predominance of discontinuities below the instru- mixed modes. Zones with high fracture frequencies ment resolution that are, instead, highlighted on the observed within the Janusfjellet Subgroup (i.e. part direct observation by the high decompaction of retri- of the lower Adventdalen Group) cap rock are asso- eved cores (e.g. strong preferential splitting of cores ciated with the upper main de´collement zone, along bedding surfaces, laminae and structural dis- appearing as centimetre- to metre-thick zones of continuities), which is probably supported by the unconsolidated, ‘crushed’ shales (shale gouge), sur- underpressure recorded in the reservoir. For the rounded by a dense network of generally conjugate, reservoir succession the N/G ratio is only around low-angle shear surfaces. Other minor peaks in frac- 25–30%, defined on the total sand thickness. ture frequencies are found within the reservoir

Fig. 3. Overview of the stratigraphy penetrated by the LYB CO2 lab project drilling campaign from 2007 to 2010 in Adventdalen, Spitsbergen. Boreholes Dh1 and Dh3 cored the cap-rock succession, whereas the Dh2 borehole cored the upper part of the target storage unit. The Dh4 borehole penetrated nearly the whole reservoir section. The black rectangle indicates the proposed injection target, which is further analysed in this paper, and the stratigraphical portion covered by analyses of cores. The red rectangle highlights the portion of the stratigraphical section investigated in the field (see Fig. 7 for correlation). Several igneous intrusions, up to 2 m thick, are penetrated by Dh4, and are illustrated in green. Please note that Dh1 and Dh2 are drilled at drill site 1 close to the airport, while Dh3 and Dh4 are drilled at drill site 2 in Adventdalen (see Fig. 2 for the exact locations). Table 1. Summary of the selected scanlines used in this contribution

Total No. of No. Fractures/m Fractures/m TG BC Fractures/m TG BC Lithology Considered Lithostructural No. of TG of BC (average) (maximum) fractures/m fractures/m (average) fractures/m fractures/m thickness association fractures fractures fractures (average) (average) length- (average) (average) (cm) (LSA) normalized length- length- Geological Society,London,SpecialPublicationspublishedonlineSeptember10,2012asdoi: normalized normalized

221 106 115 7.4 15 3.5 3.8 6.3 3 3.2 Fine- to 20 C medium-grained bioclastic bedset 317 147 170 12.7 18 5.8 6.8 9 4.1 4.8 Finely laminated 200 A silty–shale interval 74 38 36 8.0 10 4.1 3.9 2.1 1.1 1 Medium-grained 30 C sandstone bedset 192 107 85 6.4 13 3.5 2.8 5.4 3 2.4 Medium- to 150 C coarse-grained sandstone bed 144 88 56 4.8 12 2.9 1.8 4.1 2.5 1.5 Medium-grained 200 C OGATA K. sand-supported 10.1144/SP374.9 conglomerate 144 60 84 4.1 7 1.7 2.4 4.1 1.7 2.4 Thin-bedded 300 B shale–siltstone interval AL. ET 124 31 93 5.2 9 1.3 3.8 3.5 0.9 2.6 Highly bioturbated 50 B fine-grained sandstone 54 12 42 6.4 8 1.4 4.9 1.5 0.3 1.2 Highly bioturbated 50 B fine-grained sandstone 147 65 82 5.7 9 2.5 3.1 4.2 1.8 2.3 Highly bioturbated 50 B fine-grained sandstone 76 29 47 2.9 7 1.1 1.8 2 0.8 1.3 Fine- to 200 C medium-grained sandstone bed 85 23 62 8.5 10 2.3 6.2 2.4 0.6 1.7 Massive shale 150 A interval 76 31 45 4.6 10 1.8 2.7 2.1 0.8 1.2 Medium-grained 90 C sand-supported conglomerate 39 8 31 6.5 11 1.3 5.1 1.1 0.2 0.8 Medium-grained 90 C sand-supported conglomerate

The scanlines are collected across the stratigraphy of the upper reservoir section (for the location, see Fig. 7). See the text for an explanation of the shown parameters. OPEN DISCONTINUITIES CLOSED DISCONTINUITIES

TOP TOP TOP TOP Geological Society,London,SpecialPublicationspublishedonlineSeptember10,2012asdoi: a

TOP

e g h TOP TOP AUA RCUE N CO AND FRACTURES NATURAL b c TOP 10.1144/SP374.9

TOP d f i

I. II. Fig. 4. Synthesis of fracture information gathered from the drill

162 107 98 118 74 73 70 60 109 380 290 98 54 7 2 2 36 3 cores of Dh4. (a) Weathered (oxidized) fracture. Knorringfjellet Formation, depth 676.56–676.83 m. (b) Hairline weathered (oxidized) fracture. De Geerdalen Formation, depth 698.17– 2 STORAGE 698.34 m. (c) Hairline fresh fractures. De Geerdalen Formation, depth 716.77–716.91 m. (d) Highly fractured interval. De Geerdalen Formation, depth 753.00–753.36 m. (e) Disaggregation deformation band, with phyllosilicate membrane and approximately 1 cm displacement. Knorringfjellet Formation, depth 679.71–679.86 m. (f) Disaggregation deformation band. Knorringfjellet Formation, 0.8 cm displacement, depth 678.54– 678.64 m. (g) Calcite veins in sandstone. De Geerdalen Formation, depth 920.75–920.96 m. (h) Calcite veins in shale. De Geerdalen Formation, depth 966.42–966.48 m. (i) Mixed calcite–pyrite, thick vein in shale. De Geerdalen Formation, depth 955.61– 955.70 m. The graphs indicate the relative fractions of the open and closed types for different dip angles (to the left; diagram I) and in different host lithologies (to the right; diagram II). The absolute Lithologies number of discontinuities within each category is given at the top of each bar. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

Fig. 5. Wireline logs and fracture measurements for the target reservoir interval in the upper part of the De Geerdalen Formation and the Knorringfjellet Formation, as cored by Dh4. The wireline tools record the gamma ray, resistivity and temperature logs (Elvebakk 2010). Major changes in the resistivity and calculated geothermal gradient logs generally line up with heavily fractured intervals, possibly suggesting enhanced fluid flow. The fracture intensity plots are subdivided into three dip angle categories, as well as into open and closed fracture intensity curves. GR, gamma ray; Res, resistivity (ohm . m); Temp, temperature (8C); TV int, televiewer intensity; f/m, number of fractures per metre. section related to specific stratigraphical intervals discussed earlier. Furthermore, closed discontinu- (see Fig. 5). ities generally occur only within coarser-grained The projected reservoir interval (storage unit) in units, especially in the upper third of the Knorringf- the Knorringfjellet Formation was fully logged jellet Formation, peaking at approximately 678 m. using wireline tools recording gamma ray, resistiv- ity and temperature readings, as well as an acoustic Outcrop data televiewer (Elvebakk 2010). This allows correla- tion of the fractures registered by the televiewer The presented scanlines dataset was primarily col- and the manual count (see Fig. 5). As the drill cores lected in a valley located about 1 km east of Konus- were not oriented during recovery, the azimuthal dalen (for the location see Fig. 2). The valley has information cannot be compared. The stratigraphi- previously been studied for biostratigraphical pur- cal fracture intensity (i.e. on the vertical axis), how- poses (Wierzbowski et al. 1981; Chlebowski & ever, indicates some agreement, particularly within Wierzbowski 1983; Wierzbowski 1989), and is cur- the uppermost part of the De Geerdalen Forma- rently being investigated for its Jurassic marine tion. The fracture frequency is clearly highest for reptiles and hydrocarbon seeps (Hammer et al. the 08–308 dip angle category; this is generally con- 2011). Outcrops are often scree-covered and scan- centrated within the finer-grained lithologies, as lines are, therefore, correspondingly short due to Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE limited exposure (Fig. 6). Nonetheless, the valley contacts), sometimes showing deflections where provides full exposure of the Knorringfjellet For- they propagate across different lithologies. mation, as well as the uppermost 57 m of the De It should be noted that the fracture sets character- Geerdalen Formation, and thus serves well for com- izing the conglomerate beds of the Knorringfjellet piling a representative stratigraphical log for corre- Formation display evidence of generalized shear- lation to the Dh4 borehole, approximately 20 km ing, with a relatively high frequency of striated away (Fig. 7). (i.e. slickensided) walls and, sometimes, other kin- Also in this case, the high-angle fractures are pre- ematic indicators (e.g. calcite steps, en echelon dominant and are mainly represented by Mode I frac- veins). Such structures are distributed in both high- tures (i.e. joints), while the subordinate low-angle angle and low-angle fractures, exhibiting preferen- ones are expressed by Mode II and III fractures tial strike-slip and dip-slip movements, respectively (i.e. shears/tears), preferentially occurring within (Fig. 9). coarse-grained and fine-grained intervals, respect- Throughout the investigated section, some evi- ively. These structures have been distinguished in dence of calcite and quartz veining and cementation two main populations, through-going (TG) or bed- was observed within carbonate layers (e.g. lime- confined (BC), onthe basis of theirpersistence within stones, bioclastites) and medium- to thick-bedded the considered lithostratigraphical interval (e.g. sandstones and conglomerates, respectively. Enhan- bed, bedset, lithological unit), their cross-cutting ced, localized calcite cementation occurs within and relationships with the host lithologies and, where around fossiliferous carbonate beds and limestones, possible, their continuity at the outcrop scale. In especially in the upper part of the De Geerdalen this framework, compared to BC fractures, TG frac- Formation; whereas the Knorringfjellet Forma- tures are defined by sets that are more homogeneous. tion is generally characterized by quartz-cemented, Localized high dispersions of structural orien- coarse-grained beds. Enhanced cementation and tations are observed within the laminated to thin- veining was also observed within and around deci- bedded, heterogeneous intervals (i.e. lithostructural metre- to metre-sized dolerite dykes and sills in the domain B), while the massive, generally homo- lower–middle part of the De Geerdalen Formation. geneous rocks are characterized by relatively more Localized oxide-cemented parts are also observed even and clear distributions (i.e. lithostructural throughout the section, especially in the upper- domains A and C). most De Geerdalen Formation and Rurikfjellet In general, the high-angle fractures seem to be Formation. organized into two main sets, trending approxi- Iron and manganese oxides forming dark, mately east–west and NNE–SSW, and two subor- reddish- to purplish-coloured coatings on fracture dinate and less defined NW–SE and NE–SW sets. walls are widely distributed, especially on most con- The low-angle fractures, prevailing within the fine- tinuous open cracks and on weathered surfaces. grained, massive to laminated intervals (i.e. litho- structural domain A), are, instead, represented by two principal sets striking approximately WNW– Discussion and conclusions ESE and NNE–SSW (Fig. 8). The above-described fracture sets are expressed The structural characterization of the reservoir sec- differently within the investigated section, with evi- tion was performed on the upper 100 m of the Kapp dent strike rotations of the main trends. In the lower Toscana Group, from which borehole and outcrop part of the exposed De Gerdaleen Formation, the data were available and thus comparable in the data- main fracture set trends approximately east–west, base. Utilizing both data sources, we discuss the with a subordinate set trending NNE–SSW; while genesis and the evolution of the investigated struc- the upper part of this formation is, instead, charac- tures, their lithostratigraphical distribution and terized by WNW–ESE and ENE–WSW sets, their possible implications for subsurface fluid flow with minor occurrence of the east–west and the in the reservoir and the cap rock. NNE–SSW sets. The fracture sets of the Knorring- fjellet Formation are represented by three main Tectonic events and related structures ENE–WSW, NE–SW and NW–SE sets, with the secondary occurrence of the NNE–SSW set, defining Based on cross-cutting relationships and the phys- in places a single broad fan of fractures (see Fig. 7). ical/geometrical characteristics of the investigated Systematic high-angle fractures do not show structural discontinuities, we propose the following clear terminations within specific lithologies, ran- chronological subdivision from oldest to youngest domly tipping out into each defined lithostructural (Fig. 10): domain, while low-angle ones seems to preferen- † High-angle disaggregation deformation bands tially die out at well-defined lithological boundaries (clay-filled microfaults) affecting coarse-grained (e.g. master bedding surfaces and abrupt lithological beds. These are attributed to the early phases Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

Knorringfjellet Fm. CD_KD_4 CD_KD_5 ~ S CD_KD_3 ~ N

De Geerdalen Fm.

CD_KD_2 CD_KD_1 ca. 10 m CD Strat. Log

through-going fracture bed-confined fracture scanline path

CD_KD_5 considered interval ca. 2 m

A a

CD_KD_4 ca. 2 m

B b

CD_KD_3 ca. 1 m ca. 50 cm c

C

CD_KD_2 ca. 1 m D

CD_KD_1 ca. 1 m d

Fig. 6. Overview of the investigated section (location informally defined as Criocerasdalen, approximately 1 km east of Konusdalen), illustrating the location of the some selected scanlines (traces and scanline identities are labelled in red) as well as the stratigraphical log (yellow dashed line). For scanline descriptions see Table 1. The location of the Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE

of deformation, within basically unconsolidated unloading joints (Engelder 1987), newly created rocks.Calcite-filledveinsmaydevelopinthecon- or, most probably, developed through the reacti- tact zones of igneous intrusions due to expansion vation and interconnection of earlier discontinu- and shrinkage during the heating–cooling cycle, ities resulting from the generalized extension and possibly related aureole metasomatism and caused by Pliocene and Quaternary uplift and hydrothermal circulation. subsequent unroofing. Oxide coatings on the † Predominant low-angle fractures affecting fine- walls often characterize such open cracks, which grained horizons. These relate to layer-parallel indicates subsequent circulation of fluids. slip caused by a near-horizontal compression, probably associated with the crustal shortening Lithostructural domains and fracture density during the Tertiary episode of transpression. The competence contrast between layers could faci- The fracture sets characterizing the studied strati- litate bedding-parallel slip along weak (fine- graphical section have been subdivided into three grained) layers and subsequent strain accommo- main lithostructural domains (A, B and C). The dation-induced fracturing in intercalated, stiffer classification is based on lithology, compositions beds (e.g. coarse grained), probably reactivat- (e.g. sand/mud ratio, coarse- v. fine-grained popu- ing the older high-angle discontinuities. Layer- lations), texture (homogenerous v. heterogeneous) parallel shearing is evidenced by the predomi- and sedimentary structures (e.g. massive v. lami- nance of low-angle slickensided fractures and nated). By combining these lithostructural associ- their apparent arrangement in conjugate systems. ations with the considered thicknesses and the Moreover, high- and low-angle, strike-slip and fracture densities (i.e. fractures per metre (frac- dip-slip shear fractures in the thick conglomerate tures/m)), based in this preliminary analysis, the beds of the Knorringfjellet Formation reflect an following pattern emerges (Fig. 11): approximately NE–SW-oriented compression, as is also recorded in the western fold-and-thrust † a positive distribution for domain A – fracture belt (Bergh et al. 1997; Braathen et al. 1997). The densities increase with increasing interval general lack of sealed fractures in fine-grained thicknesses; lithologies may be attributed to the compartmen- † a slightly negative distribution for domain B – a talization of mineralizing fluids provided by moderate decrease of fracture densities with layer-parallel discontinuities in shaly layers (i.e. increasing interval thickness; permeability anisotropy) with restricted com- † a negative distribution for domain C – frac- munication between fine-grained beds and sand- ture densities decrease with increasing interval stone layers. Considering the general inferred thickness. patterns of intersection between these differ- Whereas an increase in fracturing is expected ent fracture sets, it is possible to argue for an within stiffer and thinner beds (Lorenz et al. 2002; enhanced and generalized vertical connectivity, Shackleton et al. 2005 and references therein), an with three main trends of lateral connectivity opposite trend can be discerned for the shaly inter- directed approximately NNE–SSW, NW–SE vals of domain A. Although only a part of the full and east–west. In this framework, highly minera- dataset has been included here, on-going analyses lizing permeating fluids (such as those resulting suggest this pattern to be consistent for the whole from the water–rock interactions with carbon- database. ate-rich and igneous lithologies) could cause This observation may be explained by an enhanced cementation of fractures within coarse- enhanced strain concentration within the metre- grained beds and in the nearby host rocks (see thick, homogeneous, massive to laminated shaly below). intervals caused by predominantlygeneralized layer- † The most continuous, non-mineralized and wea- parallel shearing. The high content of organic thered open fractures could be interpreted as material, the general lamination and the compaction

Fig. 6. (Continued) stratigraphical log is illustrated in Figure. 2. In addition, five photo-mosaics with relative interpretations of the represented scanlines are shown. Through-going (red) and bed-confined (black) fractures are labelled in the interpretations. Stereoplots and fracture frequency histograms for these scanlines are shown in Figure 7. The red circle in CD_KD_1 and CD_KD_2 indicates the 20 cm-long measuring handle for scale. Locations of close-up photographs (lens cap for scale; 6.7 cm in diameter) are marked in the photo-mosaics: (a) high-angle fracture network in an approximately 2 m-thick, coarse-grained layer (i.e. Slottet Bed of the lowermost Knorringfjellet Formation); (b) jigsaw erosional relief of a medium- to coarse-grained sandstone bedset highlighted by tens of centimetre-spaced, high-angle, through-going fractures; (c) high- and low-angle fracture network within a metre-thick shaly interval; (d) detail of the closely spaced high-angle fractures within a centimetre-thick, well-cemented carbonate bed. Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

Fig. 7. Correlation of the Dh4 with the measured section at the Deltaneset field locality. The CD stratigraphical log (for location see Fig. 6) correlates fairly well the Knorringfjellet Formation, and the upper third of the De Geerdalen Formation with the upper part of the reservoir in the Dh4 (sedimentary log from Dh4 is compiled by Atle Mørk: see Braathen et al. 2010). Stereoplots and histograms showing the fracture distribution of each scanline are represented beside the stratigraphical log in correlation with the labelled scanline location to highlight their relationships with the considered lithostratigraphical intervals. In the histograms, through-going (red) and bed-confined (grey) fractures are distinguished, while the sedimentary succession is subdivided into three lithostructural associations (LSA), A, B and C, on the basis of their sedimentary facies, lithology and fracture patterns (see the text for an explanation). foliation (i.e. pseudo-bedding) of these lithologies In this framework, a strain decoupling due to fric- may have contributed to the development of the tional slip and interface opening at the main litho- low-angle shear fractures. logical boundaries (e.g. master bedding planes, Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE

Fig. 8. Summary stereoplots showing the entire dataset from the selected scanlines (for a description of each scanline see Table 1, and for locations see Figs 6 & 7): (a) planes to poles; and (b) 1% area contours density plot.

WILHELMØYA SUBGROUP Knorringfjellet Fm. CD_KD_4 CD_KD_5 Strike slip Equal Angle slickenlines

Slickenlines trend/plunge a

Dip slip slickenlines

Great Circle: N = 17 ; first plane = 1 ; last plane = 17 b Pattern = solid

Fig. 9. Stereoplot of the shear fractures and the kinematic indicators recognized within the coarse-grained, thick-bedded and well-cemented beds of the Knorringfjellet Formation in the investigated section: for a description of the considered scanlines see Table 1 (locations are shown in Figs 6 & 7). These shear systems suggest an overall NE–SW-oriented axis of horizontal contraction (red arrows). (a) An example of strike-slip slickenlines in a high-angle shear fracture (compass used for scale); and (b) an example of dip-slip slickenlines in a low-angle shear fracture (lens cap used for scale). Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

CAUSES AND STRUCTURAL MECHANISMS PRODUCTS

Unroofing Activation, reactivation and uplift of early structures

Reworking (opening) of decompaction and preceding fractures generalized extension

Deformation bands (HA) Tectonic/sedimentary + hairline fractures (HA) burial and horizontal Slickensided, polished, compression conjugate fractures (LA)

layer-parallel shearing Brittle micro-faults (HA) caused by crustal shortening Pyrite mineralization Calcite mineralization

Disaggregation Gravity and deformation bands (HA) differential A Ductile micro-faults (HA) compaction B C Ductile shear zones (HA and LA)

MESOZOICigneous PALEOGENE intrusions NEOGENE Diffused soft sediment deformation structures A Shaly/massive-laminated HA: High Angle B Silty/thin-bedded LA: Low Angle C Sandy/medium-thick bedded NOT TO SCALE

Fig. 10. Proposed model for the evolution of the investigated structural discontinuities within the different lithostructural domains and their relationships with the main regional tectonic events. erosional surfaces, diagenetic horizons) may have thought to be prone to opening when subjected to played a major role in the structural segmentation even relatively small pressure increments, especially of the stratigraphical section (Cooke & Underwood if the reservoir underpressure is considered. Coarser- 2001 and references therein). grained sandstone intervals, however, are thought to enhance the vertical communication. Silty intervals, Implications for fluid flow defined by the lithostructural domain B, are charac- terized by the predominance of non-systematic, Given the abundance of both vertical and horizontal pervasive bed-confined fractures, and represent a open fractures observed in the drill cores and in out- combination of both diffused and baffled lateral crops, and very low matrix permeability, fractur- and vertical flow between the two proposed end ing appears to be relevant for fluid circulation. members. This is supported by borehole water injection tests The interconnectivity of fracture clusters at and results obtained from the wireline logs, particu- mechanical interfaces is critical to allow the verti- larly the resistivity and temperature logs (see Fig. 5). cal migration of fluids through a layered medium The block diagram in Figure 12 shows a sche- (Gudmundsson et al. 2010; Larsen & Gudmundsson matic representation of the three-dimensional frac- 2010). Fracture length and lateral connectivity ture network within the individual lithostructural between vertical, open fractures are also key factors domains A, B and C. Considering the abundance of for horizontal fluid migration (Odling et al. 1999), lower-angle fractures within the finer-grained units, since they are expected to cause a pronounced per- and reiterating that fluid flow in the tight reservoir is meability anisotropy that steers the fluid flow in a essentially a function of the fracture network, these specific direction. In the discussed case, besides a are thought to represent zones of enhanced lateral generalized vertical connectivity, the main lateral connectivity. Even if such fractures could be sealed spreading of the flow is assumed to be directed by elastic strain due to vertical stress, they are roughly WSW–ENE, according to fracture set Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE

Fig. 11. (a) Cross (X–Y ) plot showing the relationships between the different lithostructural domains, average fracture densities and considered bed thicknesses. Note the opposite trends shown by lithostructural association A with respect to B and C. (b) Whisker-plot showing a summary of the fracture frequency distributions and fluctuations for each scanline (see a for a colour-code explanation). In each box, min–max lower and upper quartile, and median are displayed.

intersections and the mean strike of the measured hydrocarbons. The storage of CO2 in depleted subvertical fractures (see Figs 8 & 12). hydrocarbon fields is tempting, since the lowered pressure allows more CO2 to be injected before Significance of fractures in the overburden the capillary entry pressure of the cap rock is exceeded. However, owing to the lower interfacial Fractured cap rocks, such as in the Snorre Field tension in the CO2–brine system compared to the (Caillet 1993), may cause leakage of CO2 and/or hydrocarbon–water system, the cap-rock sealing Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL.

Equal Area C.I. = 1.0 % / 1 % area Equal Area C.I. = 1.0 % / 1 % area Lower 1 % Area Contours Lower 1 % Area Contours Diffused/baffled Hemisphere Hemisphere lateral and vertical connectivity

N = 469 C N = 822 Enhanced vertical connectivity B

A Equal Area C.I. = 1.0 % / 1 % area Lower 1 % Area Contours Hemisphere

Potential fluid migration pathways

A Shaly/massive-laminated Enhanced lateral B Silty/thin-bedded connectivity C Sandy/medium-thick bedded not to scale N = 402

Fig. 12. Schematic diagram of the interconnectivity and intensity of fractures within the three lithostructural associations: A, B and C (see the text for details). Corresponding stereoplots show the total considered fractures (plane to poles) for each lithostructural association. The finer-grained lithostructural domain A is typically characterized by conjugate, lower-angle shear fractures, leading to enhanced lateral connectivity. The coarser-grained units, however, are typically dominated by steep fractures, and may thus enhance vertical connectivity. Note also the mixture of steep and lower-angle fractures within the silty interval of lithostructural domain B. pressure must be determined prior to injection Local enhanced sealing processes and (Li et al. 2006). The increase in pressure related to reservoir segmentation the potential CO2 injection furthermore may lead to the reactivation of faults within the overburden Localized veining and enhanced cementation occurs (Chiaramonte et al. 2008). Considering the poten- within and around beds or intervals characterized by tial storage of CO2 within the coarser-grained inter- specific compositions and sedimentary features: vals of the Knorringfjellet Formation, we must also increased calcite and quartz precipitation relate to consider possible leakage pathways from these the diagenesis of carbonate, crystalline and thick- units. bedded, coarse-grained layers, respectively. Such On Spitsbergen, even though the Janusfjellet diagenetic horizons may represent intervals with a Subgroup cap rock is highly fractured, the differen- baffled connectivity due to fracture sealing and, tial pressure in the target reservoir and the overbur- thus, possibly contribute to vertical compartmentali- den affirms the efficient sealing potential of the unit. zation of the reservoir/storage unit. Reactivation and opening of the predominantly Moreover, the presence of dolerite intrusions low-angle fractures may lead to enhanced lateral within the reservoir unit is manifested both in Dh4, migration at the base and/or within the cap-rock where a 2 m-thick sill occurs at 950 m, and by succession but, owing to the low vertical connec- seismic data that suggest thicker sills beneath the tivity, the cap rock is deemed to be sealing. This well’s total depth of 970 m. Such bodies are also view has recently been corroborated by water injec- found in outcrop: intrusive rocks of the Diabasod- tion tests. den Suite occur over large parts of the Svalbard Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE archipelago and are generally thought to be of Late suggests that the succession is likely to be suitable Jurassic or Early Cretaceous age (Burov et al. for storing the modest, test-site-scale CO2 emissions 1977; Maher 2001, Nejbert et al. 2011). In outcrop, from the local power plant. sills exhibit a thickness ranging from 1 to more than 30 m, with off-cutting dykes generally minor and localized. An exception is the probable feeder Concluding remarks dyke at Diabasodden, adjacent to the Deltaneset field locality. The drilled dolerite in Dh4 shows, on a Carbon dioxide may potentially be injected into a small scale, that increased fracturing occurs around siliciclastic succession on Spitsbergen, Svalbard, the intrusion, probably associated with focused given the good injectivity of this naturally fractured strain during Tertiary contractional tectonics. Fur- tight sandstone–shale reservoir/storage unit. Frac- thermore, millimetre- to centimetre-thick, calcite- ture sets have been identified in both drill core and filled fractures in the vicinity of this and other, even outcrop data, and three major lithostructural thinner, dolerite bodies in Dh4 suggest increased domains (A, B and C) can be identified throughout mineralized fluid circulation, possibly due to syn- the investigated section. In the massive to laminated, emplacement metasomatism and hydrothermalism, shale-dominated unit A, low-angled slickensided and related calcite precipitation, within the contact open fractures dominate and are probably related aureoles, or post-emplacement water–rock interac- to the regional Tertiary transpression. The other end tions. The mechanical deformation of the sedimen- member, the massive to laminated, sand-dominated tary succession could therefore either be related to unit C, is characterized by steeply dipping fractures, the intrusion phase (e.g. magma flow, bulging and probably reworked/reactivated and interconnected thermal shrinking) and/or to the subsequent tec- during the later Quaternary unroofing. The inter- tonic- related mechanical decoupling due to rheolo- mediate heterogeneous, laminated to thin-bedded, gical differences between igneous units and the host silty unit B displays a predominantly non-systematic, rock at the post-intrusive stage. dispersed fracture network, representing a mixed In this framework, dolerite bodies may play a member in-between. Such fracture associations are double role in controlling the fluid flow: (1) acting considered to represent the key factor in control- as conduits due to along-dyke wall and internal frac- ling the internal connectivity of the reservoir and turing; and (2) acting as baffles or barriers for the the direction of the fluid flow, by influencing lateral wall-rock fracture flow due to cemented fractures and vertical fluid migration, respectively. in the contact aureole. Depending on their three- The LYB CO2 Lab project represents a unique dimensional architecture and fracturing-sealing test site for potential CO2 storage in a naturally frac- processes, dolerite intrusions could serve to com- tured tight siliciclastic reservoir. The large amount partmentalize the reservoir into high-conductivity of available data from different methods, coupled fracture corridors and low-conductivity domains of with the possibility of detailed studies on the reser- cemented fractures. Further work addressing these voir and cap rock directly in outcrop, allow a com- issues is currently in progress. prehensive characterization of the potential storage framework. The future exploitation of unconven- Implications for CO storage tional reservoirs for subsurface CO2 disposal needs 2 to have the same degree of knowledge developed The fractured nature of the reservoir, together with for the currently emerging hydrocarbon exploration its low matrix porosity, complicates the volumetric for shale gas/oil and enhanced geothermal systems calculation of the potential CO2 storage reservoir. (EGS). This should highlight the importance of Numerical modelling has shown that, depending detailed structural studies, especially for the stimu- on the chemical composition of the fluid, the reser- lation (i.e. hydrofracturing) of tight reservoirs, the voir and cap rocks, CO2 may both enhance porosity construction of datasets gathering key parameters through dissolution yet also decrease porosity and for the characterization of the vertical–lateral frac- permeability through the precipitation of calcite ture connectivity within the reservoir–cap rock (Gherardi et al. 2007). In the LYB CO2 lab case, system and, therefore, the prediction/mitigation of the storage potential of the target aquifer is funda- the risk of leakage towards the surface. mentally dependent on the interconnectivity, length and aperture of the fracture system. Further, the This work is part of the ‘Geological input to Carbon dynamic interaction between the fractures and the Storage (GeC)’ project funded by the CLIMIT programme of the Research Council of Norway, with K. Senger’s field- host-rock matrix represents a substantial source of work supported by an Arctic Field Grant from the Svalbard uncertainty (Farokhpoor et al. 2010, 2011). At this Science Forum. A. Rittersbacher, D. Richey and L. Farrell stage it is still unclear whether the Kapp Toscana are thanked for their assistance in the field. The GeC Group sediments are, indeed, optimal for storing project team works in close co-operation with the Long- large amounts of CO2, but acquired evidence yearbyen CO2 lab project (http://co2-ccs.unis.no), which Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 K. OGATA ET AL. is currently funded by Gassnova, ConocoPhillips, Statoil, frictional slip and interface opening. Journal of Struc- Statkraft, Store Norske, LNS, Lundin Norway and Baker tural Geology, 23, 223–228. Hughes. Shapefiles from NPI-Geonet have been used to Engelder, T. 1987. Joints and shear fractures in rock. In: construct the geological maps. The North Africa Research Atkinson, B. K. (ed.) Fracture Mechanics of Rock. Group at the University of Manchester is thanked for orga- Academic Press, Orlando, FL, 27–69. nizing the ‘Naturally Fractured Hydrocarbon Reservoirs’ Eldholm, O., Faleide,J.I.&Myhre, A. M. 1987. workshop where this work was initially presented. Finally, Continent-ocean transition at the western Barents we would like to thank G. Spence and T. Bevan for having Sea/Svalbard continental margin. Geology, 15, supported the publication of this work, and R.H. Gabriel- 1118–1122. sen and C. Bond for their constructive reviews. Elvebakk, H. 2010. Results of Borehole Logging in Well LYB CO2, Dh4 of 2009, Longyearbyen, Svalbard. NGU Report, 2010.018. Geological Survey of References Norway (NGU), Trondheim. Farokhpoor, R., Torsæter, O., Baghbanbashi, T., Baelum, K., Johansen, T. A., Johnsen, H., Røed, K., Mørk,A.&Lindeberg, E. G. B. 2010. Experimental Ruud, B. O., Mikkelsen,E.&Braathen, A. 2012. and numerical simulation of CO2 injection into Subsurface geometries of the Longyearbyen CO2 upper-Triassic sandstones in Svalbard, Norway. In: lab in Central Spitsbergen, as mapped by reflec- SPE International Conference on CO2 Capture, tion seismic data. Norwegian Journal of Geology,in Storage, and Utilization held in New Orleans, Louisi- press. ana, USA, 10–12 November 2010. Society of Pet- Bergh, S. G., Braathen,A.&Andresen, A. 1997. Inter- roleum Engineers, Richardson, TX, Paper SPE- action of basement-involved and thin-skinned tecton- 139524. ism in the Tertiary fold-thrust belt of central Farokhpoor, R., Baghbanbashi, T., Torsæter, O., Spitsbergen, Svalbard. AAPG Bulletin, 81, 637–661. Lindeberg,E.&Mørk, A. 2011. Experimental and Blythe,A.E.&Kleinspehn, K. L. 1998. Tectonically simulation analysis of CO2 storage in tight and frac- versus climatically driven Cenozoic exhumation of tured sandstone under different stress conditions. In: the Eurasian plate margin, Svalbard: fission track ana- SPE Europec/EAGE Conference and Exhibition, lyses. Tectonics, 174, 621–639, http://dx.doi.org/10. Vienna, Austria, 23–26 May 2011. Society of Pet- 1029/98TC01890. roleum Engineers, Richardson, TX, Paper SPE- Braathen, A., Bergh,S.G.&Maher,H.D.Jr. 1995. 143589. Structural outline of a Tertiary Basement-cored uplift/ Fossen, H. 2010. Deformation bands formed during soft- inversion structure in western Spitsbergen, Svalbard: sediment deformation: observations from SE Utah. kinematics and controlling factors. Tectonics, 14, Marine and Petroleum Geology, 27, 215–222. 95–119. Fossen, H., Schulz, R. A., Shipton,Z.K.&Mair,K. Braathen, A., Bergh,S.&Maher, H. 1997. Thrust kin- 2007. Deformation bands in sandstone – a review. ematics in the central part of the Tertiary transpres- Journal of the Geological Society, 164, 755–769. sional fold-thrust belt in Spitsbergen. NGU Bulletin, Gherardi, F., Xu,T.&Pruess, K. 2007. Numerical mod- 433, 32–33. eling of self-limiting and self-enhancing caprock Braathen, A., Bergh,S.G.&Maher,H.D.Jr. 1999. alteration induced by CO2 storage in a depleted gas Application of a critical wedge taper model to the Ter- reservoir. Chemical Geology, 244, 103–129. tiary transpressional fold-thrust belt on Spitsbergen, Glørstad-Clark, E., Faleide, J. I., Lundschien,B.A.& Svalbard. Geological Society of America Bulletin, Nystuen, J. P. 2010. Triassic seismic sequence strati- 111, 1468–1485. graphy and paleogeography of the western Barents Braathen, A., Baelum,K.et al. 2010. Longyearbyen Sea area. Marine and Petroleum Geology, 27, CO2 lab 2007–2009 – Phase 1. Final Report. UNIS 1448–1475. Report, 2010-02, 69. University Centre in Svalbard Gudmundsson, A., Simmenes, T. H., Larsen,B.& (UNIS), Longyearbyen. Philipp, S. L. 2010. Effects of internal structure and Burov, P., Krasil’scikov, A. A., Firsov,L.V.& local stresses on fracture propagation, deflection, and Klubov, B. A. 1977. The age of Spitsbergen Dolerites arrest in fault zones. Journal of Structural Geology, (from isotopic dating). Norsk polarinstitutt A˚ rbok, 32, 1643–1655. 1975, 101–108. Hammer, Ø., Nakrem,H.A.et al. 2011. Hydrocarbon Caillet, G. 1993. The caprock of the Snorre Field, seeps from close to the Jurassic-Cretaceous boundary, Norway: a possible leakage by hydraulic fracturing. Svalbard. Palaeogeography, Palaeoclimatology, Pala- Marine and Petroleum Geology, 10, 42–50. eoecology, 306, 15–26. Chiaramonte, L., Zoback, M., Friedmann,J.&Stamp, Harland, W. B. 1997. Chapter 3: Svalbard’s geological V . 2008. Seal integrity and feasibility of CO2 seques- frame. In: Harland, B. (ed.) The . tration in the Teapot Dome EOR pilot: geomechanical Geological Society, London, Memoirs, 17, 23–46. site characterization. Environmental Geology, 54, Harland,W.B.&Geddes, I. 1997. Chapter 18: Triassic 1667–1675. history. In: Harland, B. (ed.) The Geology of Sval- Chlebowski,R.&Wierzbowski, A. 1983. Pyroclastic bard. Geological Society, London, Memoirs, 17, material from the Upper Triassic deposits of Sassen- 340–362. fjorden, Spitsbergen. Polar Research, 1, 75–81. Helland-Hansen, W. 2010. Facies and stacking patterns Cooke,M.L.&Underwood, C. A. 2001. Fracture ter- of shelf-deltas within the Palaeogene Battfjellet For- mination and step-over at bedding interfaces due to mation, Nordenskio¨ld Land, Svalbard: implications Geological Society, London, Special Publications published online September 10, 2012 as doi: 10.1144/SP374.9 NATURAL FRACTURES AND CO2 STORAGE

for subsurface reservoir prediction. Sedimentology, 57, Nejbert, K., Krajewski, K. P., Dubin´ ska,E.&Pe´cskay, 190–208. Z. 2011. Dolerites of Svalbard, north-west Barents Sea Humlum, O., Instanes,A.&Sollid, J. L. 2003. Perma- Shelf: age, tectonic setting and significance for geotec- frost in Svalbard: a review of research history, climatic tonic interpretation of the High-Arctic Large Igneous background and engineering challenges. Polar Province. Polar Research, 30, 7306. Research, 22, 191–215. Nøttvedt, A., Livbjerg, F., Midbøe,P.S.&Rasmus- Ingo´lfsson,O´ . 2011. Fingerprints of Quaternary glacia- sen, E. 1993. Hydrocarbon potential of the Central tions on Svalbard. In: Martini, I. P., French,H.M. Spitsbergen Basin. In: Vorren, T. O., Bergsager, & Pe´rez Alberti, A. (eds) Ice-Marginal and Perigla- E., Dahl-Stamnes, Ø. A., Holter, E., Johansen, cial Processes and Sediments. Geological Society, B., Lie,E.&Lund, T. B. (eds) Arctic Geology and London, Special Publications, 354, 15–31. Petroleum Potential. Norwegian Petroleum Society, Knarud, R. 1980. En sedimentologisk og diagenetisk Special Publications, 2, 333–361. undersøkelse av kapp toscana formasjonens sedimeter Odling, N. E., Gillespie,P.et al. 1999. Variations in pa˚ Svalbard. MSc thesis, University of Oslo. fracture system geometry and their implications for Kulander, B. R., Dean,S.L.&Barton, C. C. 1977. fluid flow in fractures hydrocarbon reservoirs. Pet- Fractographic Logging for Determination of Pre- roleum Geoscience, 5, 373–384. core and Core-induced Fractures – Nicholas Combs Riis, F., Lundschien, B. A., Høy, T., Mørk,A.&Mørk, No. 7239 Well, Hazard, Kentucky. Energy Reasearch M. B. E. 2008. Evolution of the Triassic shelf in the and Development Administration, Morgantown northern Barents Sea region. Polar Research, 27, Energy Research Center, Morgantown, WV. 318–338. Larsen,B.&Gudmundsson, A. 2010. Linking of frac- Schultz,R.A.&Fossen, H. 2008. Terminology for struc- tures in layered rocks: implications for permeability. tural discontinuities. AAPG Bulletin, 92, 853–867. Tectonophysics, 492, 108–120. Singhal,B.B.S.&Gupta, R. P. 2010. Applied Hydro- Leever, K. A., Gabrielsen, R. H., Faleide,J.I.& geology of Fractured Rocks. Kluwer, Dordrecht. Braathen, A. 2011. A transpressional origin for the Shackleton, J. R., Cooke,M.L.&Sussman, A. J. 2005. West Spitsbergen fold and thrust belt – Insight from Evidence for temporally changing mechanical strati- analog modeling. Tectonics, 30, TC2014, http://dx. graphy and effects on joint-network architecture. doi.org/10.1029/2010TC002753. Geology, 33, 101–104. Li, Z., Dong, M., Li,S.&Huang, S. 2006. CO2 sequestra- Spencer, A. M., Briskeby,P.I.et al. 2008. Petroleum tion in depleted oil and gas reservoirs – caprock geoscience in Norden – exploration, production and characterization and storage capacity. Energy Conver- organization. Episodes, 31, 10. sion and Management, 47, 1372–1382. Steel,R.J.&Worsley, D. 1984. Svalbard’s post- Lorenz, J. C., Sterling, J. L., Schechter, D. S., Caledonian strata – an atlas of sedimentational pat- Whigham,C.L.&Jensen, J. J. 2002. Natural frac- terns and paleogeographic evolution. In: Spencer, tures in the Sprayberry Formation, Midland Basin, A. M. (ed.) Petroleum Geology of the North European Texas: the effects of mechanical stratigraphy on frac- Margin. Graham & Trotman, London, 109–135. ture variability and reservoir behavior. AAPG Bulletin, Steel, R. J., Gjelberg, J., Helland-Hansen, W., 86, 505–524. Kleinspehn, K., Nøttvedt,A.&Rye-Larsen,M. Maher,H.D.Jr. 2001. Manifestations of the Cretaceous 1985. The Tertiary strike-slip basins and orogenic High Arctic Large Igneous Province in Svalbard. belt of Spitsbergen. In: Biddle,K.T.&Christie- Journal of Geology, 109, 91–104. Blick, N. (eds) Strike-Slip Deformation, Basin For- Major, H., Haremo, P., Dallmann,W.K.&Andersen, mation, and Sedimentation. SEPM, Special Publi- A. 2001. Geological Map of Svalbard, 1:100,000, Sheet cations, 37, 339–359. C9G Adventdalen. Norwegian Polar Institute, Tromsø. Throndsen, T. 1982. Vitrinite reflectance studies of coals Minakov, A., Mjelde, R., Faleide, J. I., Flueh, E. R., and dispersed organic matter in Tertiary deposits in the Dannowski,A.&Keers, H. 2012. Mafic intrusions Adventdalen area, Svalbard. Polar Research, 1982, east of Svalbard imaged by active-source seismic tom- 77–91. ography. Tectonophysics, 518, 106–118. Wierzbowski, A. 1989. Ammonites and stratigraphy of the Mørk,A.&Worsley, D. 2006. Triassic of Svalbard and Kirnmeridgian at Wimanfjellet, Sassenfjorden, Spits- the Barents shelf. NGF Abstracts and Proceedings, 3, bergen. Acta Palaeontologica Polonica, 34, 355–378. 23–29. Wierzbowski, A., Kulicki,C.&Pugaczewska,H. Mørk, A., Knarud,R.&Worsley, D. 1982. Deposi- 1981. Fauna and stratigraphy of the uppermost Triassic tional and diagenetic environments of the Triassic and the Toarcian and Aalenian deposits in the Sassen- and Lower Jurassic succession of Svalbard. In: fjorden, Spitsbergen. Acta Palaeontologica Polonica, Embry,A.F.&Balkwill, H. R. (eds) Arctic 26, 195–253. Geology and Geophysics. Canadian Society of Pet- Worsley, D. 1973. The Wilhelmøya Formation – a new roleum Geologists, Memoirs, 8, 371–398. lithostratigraphical unit from the Mesozoic of eastern Nagy, J., Hess, S., Dypvik,H.&Bjærke, T. 2011. Marine Svalbard. Norsk polarinstitutt A˚ rbok, 1971, 7–16. shelf to paralic biofacies of Upper Triassic to Lower Worsley, D. 2008. The post-Caledonian development of Jurassic deposits in Spitsbergen. Palaeogeography, Svalbard and the western Barents Sea. Polar Research, Palaeoclimatology, Palaeoecology, 300, 138–151. 27, 298–317.