From Outcrop to Reservoir Simulation Model: Workﬂ Ow and Procedures

From outcrop to reservoir simulation model: Workﬂ ow and procedures

Håvard D. Enge* Department of Earth Science, University of Bergen, Post box 7800, N-5020 Bergen, Norway, and Centre for Integrated Petroleum Research, University of Bergen, Post box 7800, N-5020 Bergen, Norway Simon J. Buckley Centre for Integrated Petroleum Research, University of Bergen, Post box 7800, N-5020 Bergen, Norway Atle Rotevatn Department of Earth Science, University of Bergen, Post box 7800, N-5020 Bergen, Norway, and Centre for Integrated Petroleum Research, University of Bergen, Post box 7800, N-5020 Bergen, Norway John A. Howell Centre for Integrated Petroleum Research, University of Bergen, Post box 7800, N-5020 Bergen, Norway

ABSTRACT titative data on body geometry, and dynamic software; and (3) other applications of the col- investigation involves the simulation of fl uid lected data and the virtual outcrop. Since the pio- Advances in data capture and computer fl ow through the analog model. neering work of Bellian et al. (2005), there has technology have made possible the collection The work presented in this study dem- been a rapid increase in the application of lidar of three-dimensional, high-resolution, digital onstrates the utility of lidar as a data col- to the study and characterization of geological geological data from outcrop analogs. This lection technique for the building of more outcrops. Numerous groups are now working paper presents new methodologies for the accurate outcrop-based geocellular models. with such data, although, to date, no systematic acquisition and utilization of three-dimen- The aim of this publication is to present the methodologies for the collection, processing, sional information generated by ground- fi rst documentation of a complete workfl ow and utilization of these data have been published based laser scanning (lidar) of outcrops. A that extends from outcrop selection to model (e.g., Adams et al., 2007; Aiken, 2006; Deveu- complete workfl ow is documented—from investigation through the presentation of two gle et al., 2007; Enge et al., 2007; Enge et al., outcrop selection through data collection, worked data sets. 2006; Howell et al., 2007; Howell et al., 2006; processing and building of virtual outcrops— Jones et al., 2007; Lee et al., 2007; Martinsen et to geological interpretation and the building Keywords: outcrop analogs, laser methods, Pyr- al., 2007; Monsen, 2006; Oftedal et al., 2007; of geocellular models using an industry-stan- enees, sandstone, deltas, clinoforms, Canyonlands Olariu et al., 2005; Pedersen et al., 2007; Thur- dard, reservoir-modeling software. Data sets National Park, grabens, ramps, fault blocks, reser- mond, 2006). This paper documents a complete from the Roda Sandstone in the Spanish voir, modeling, analog simulation, fl uid. workfl ow, from outcrop selection through data Pyrenees and the Grabens region of Canyon- collection, processing, and interpretation, to the lands National Park, Utah, USA, are used to INTRODUCTION building of the geocellular model. The workfl ow illustrate the application of the workfl ow to is illustrated with two case studies that illustrate sedimentary and structural problems at a The intention of this study is to present new the application to sedimentary and structural reservoir scale. methodologies for the acquisition and utilization reservoir, geology-related problems. Subsurface reservoir models are limited of three-dimensional (3D) information gener- Lidar, which stands for light detection and by available geological data. Outcrop analogs ated by the ground-based laser scanning (lidar) ranging, includes both aerial and ground-based from comparable systems, such as the Roda of geological outcrops. In particular, the focus is techniques (Ackermann, 1999; Buckley et al., Sandstone and the Grabens, are commonly on (1) the accurate representation of geological 2006; Wehr and Lohr, 1999). Originally devel- used to provide additional input to models of entities from outcrops on a computer (referred to oped for aerial surveying, especially topographic the subsurface. Outcrop geocellular models in this paper as a “virtual outcrop”); (2) utilizing mapping, the technique allows the rapid collec- can be analyzed both statically and dynami- the virtual outcrop to extract data for the build- tion of spatially constrained point data that can cally, wherein static examination involves ing and testing of 3D geocellular models using capture the shape of a scanned feature (Baltsav- visual inspection and the extraction of quan- conventional, hydrocarbon reservoir-modeling ias, 1999; Baltsavias et al., 2001; Nagihara et

*[email protected]

Geosphere; December 2007; v. 3; no. 6; p. 469–490; doi: 10.1130/GES00099.1; 17 ﬁ gures; 1 table.

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

al., 2004). A geocellular model is a computer- fi eld; (4) safe access to vertical and sub-vertical rapidly in popularity (McCaffrey et al., 2005). based representation of a geological volume, portions of the outcrop; (5) the ability to iter- These methods were reviewed by Pringle et typically a subsurface reservoir. The model ate between the outcrop and the model during al. (2006) and include a variety of techniques comprises mapped surfaces that defi ne zones. the model-building phase; and (6) the ability for producing data of different resolutions Zones are populated with cells, which, in turn, to illustrate the model and outcrop side by side and accuracies. The application of laser scan- are assigned parameters such as porosity, per- for training and teaching purposes. The collec- ning as a method for ground-based geological meability, facies, etc. Such models are routinely tion of ground-based lidar data and the building fi eldwork is now proven (Bellian et al., 2005; used to visualize and simulate the subsurface of virtual outcrops provide a means to address Buckley et al., 2006; Leren, 2007; Pringle et in the oil industry. Given the poor resolution these issues. al., 2004a; Pringle et al., 2006; Redfern et al., of seismic data (e.g., Pickup and Hern, 2002) 2007). The employment of lidar and the cre- and sparse frequency of wells in most oil fi elds Previous Work Review ation of virtual outcrops from the point clouds (typical spacing ~1 km), outcrop data are com- provide a means for the rapid collection and monly used to provide information on interwell The application of digital data collection tech- interpretation of large volumes of accurate facies and structural architectures (e.g., Alexan- niques for outcrop studies is not new. Stafl eu et geometric outcrop data. A particular advantage der, 1993; Dreyer et al., 1993; Pickup and Hern, al. (1996) acquired photogrammetric stereopairs of terrestrial lidar scanning is that resultant sur- 2002; Reynolds, 1999) (Fig. 1). Reservoir mod- of carbonate rock outcrops to form digital eleva- faces are more effi cient to produce and have eling software has long been used to represent tion models (DEMs). These were then linked a higher accuracy potential than photogram- geological outcrops (Bryant et al., 2000; Bryant with petrophysical data to identify a relation- metric surfaces, especially in areas that exhibit and Flint, 1993; Dreyer et al., 1993; Joseph et ship between erosion and rock impedance. Xu high relief, such as good quality geological al., 1993), both for direct reservoir analogs and et al. (2000; 2001) used the Global Positioning outcrops (Baltsavias et al., 2001; Buckley et also as a tool for capturing structural and strati- System (GPS) and a refl ectorless laser to col- al., 2006). graphic architecture (e.g., Bellian et al., 2005; lect outcrop data and construct surfaces. Adams The techniques for collecting, preparing, Weber, 1986; White et al., 2004; Willis and et al. (2005) used real-time kinematic GPS and and presenting scan data in a geologically White, 2000). Key issues with the utilization a total station for recording 3D datapoints from meaningful context have been reviewed by of outcrop data have been: (1) the collection of the outcrop. These were combined with a DEM several authors (McCaffrey et al., 2005; Prin- suffi cient volumes of spatially accurate data; created from photogrammetry to form the basis gle et al., 2004a; Pringle et al., 2006). Other (2) correlation of surfaces over long distances for a geocellular outcrop model. examples (e.g., Bryant et al., 2000; Pringle et and between individual outcrops; (3) the rec- Recently, the use of modern data collec- al., 2004b) show that the use of digital spatial ognition of subtle dip and strike changes in the tion techniques in fi eld geology has increased information in outcrop modeling is increasing. The utilization of the collected data, especially for the building of geocellular models is only beginning to be addressed (Dreyer et al., 1993; Løseth et al., 2003). While recent studies by authors such as Bryant et al. (2000) and Bel- 100m Virtual Outcrop lian et al. (2005) have discussed the possibili- ties for broader geological application, as yet Seismic data very little has been published other than “state- 10m of-the-art” papers describing the potential of ~ parasequences the technology.

Simulation model Overview of the Paper 1m Grid-cell geological model This paper documents for the fi rst time a Grid-cell simulation “mini-model” systematic workfl ow from the collection of Logs 10cm raw scan data to utilization of the fi nal vir-

Vertical Thickness Core/ tual outcrop and the building and testing of plug ~ bed models. The workfl ow is illustrated by the construction of two, detailed geocellular mod- 1cm els. The resulting models range in size from 100 × 100 × 2 m to several kilometers wide Probe ~ lamina and tens or even hundreds of meters thick, respectively, and illustrate the utility of vir- 1mm 1cm 10cm 1m 10m 100m 1km 10km tual outcrop data. A key aspect of this paper is to document the workfl ow for the use of Horizontal length virtual data to solve specifi c geological problems. The workfl ows will be illustrated with Figure 1. Typical length scales of sedimentary structures in the horizontal and vertical reference to the two outcrop data sets, the directions, with scales of commonly used measurements and reservoir model sizes. Note background geology of which is summarized scale of the virtual outcrop, which covers a wide range of levels. Modifi ed from Pickup and briefl y in the next section. The workfl ow is Hern (2002). then illustrated, including the stages from

470 Geosphere, December 2007

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 From outcrop to reservoir simulation model: Workﬂ ow and procedures

the collection of the data and assembling the Nio, 1985). Virtual outcrops and 3D geocellu- mudstone, or are cemented, they are poten- virtual outcrop, to its utilization in reservoir lar models illustrate the lateral geometry within tially important barriers to horizontal and ver- modeling within computer-based tools. The the lobes and their constituent clinoforms. By tical fl uid fl ow within subsurface hydrocarbon former stage includes outcrop selection, data comparison to a conventional photograph, the reservoirs. Understanding their geometry is collection, a brief summary of data processing, inspection of these geometries is substantially critical to the modeling of intrazone reservoir and the generation of virtual outcrops ready eased by the three-dimensionality of the virtual heterogeneities within such systems as the for geo-interpretation. The latter stage incor- outcrop (Fig. 3). Also of interest is the lateral lower Brent Group from the North Sea and the porates export of data to geocellular modeling and down-dip transition from steeply dipping Halten Terrace, as well as the Tampen regions software, model building, grid creation and foresets to large sub-tidal bars. (e.g., Brekke et al., 2001; Corfi eld et al., 2001; population, and fi nally, model investigation Helland-Hansen et al., 1992). Clinoforms and fl ow simulation to test the sensitivity of the The Grabens, Canyonlands National Park, can also be used to map the evolving shore- model to reservoir fl uid fl ow. Utah line trajectory. Recent studies (Hampson and Storms, 2003) have highlighted the importance BRIEF SUMMARY OF THE GEOLOGY The Devil’s Lane area in the Canyonlands of documenting clinoform evolution through OF THE FEATURED DATA SETS Grabens was studied to test the feasibility of time as a means of predicting medium to short- lidar technology for collecting data that could term beach evolution on modern coasts. The Laser scanning, combined with traditional be used to address issues in structural geology accurate measurement of clinoform geometry fi eld techniques, has been used to collect two (Fig. 4A). As the name suggests, the Grabens is very diffi cult in the fi eld. The collection of high-resolution data sets from Spain and the region of Canyonlands National Park is a heav- lidar data and the building of virtual outcrops United States. The geological backgrounds to ily faulted area that has undergone deforma- address that issue and allow the study of indi- these study areas are described below. tion throughout the last 15 Ma, due to regional vidual clinoform bodies. Both outcrops have been subjected to exten- uplift and the collapse of a subsurface layer of Arrays of normal faults are connected sive previous studies (Leren, 2005; Leren, salt (Moore and Schultz, 1999; Trudgill and through relay zones (Cartwright et al., 1996; 2007; Lopez-Blanco, 1996; Lopez-Blanco Cartwright, 1994). The area features a series of Childs et al., 1995; Peacock and Sanderson, et al., 2003; Molenaar and Martinius, 1990, interconnected systems of horsts and grabens, 1994). In addition to providing important infor- 1996; Moore and Schultz, 1999; Peacock and a confi guration of faults and fault blocks mation on the evolution of fault systems, relay and Sanderson, 1991; Rotevatn et al., 2007; that is geometrically analogous to many sub- structures may also provide conduits for fl uid Trudgill and Cartwright, 1994; Yang and Nio, surface hydrocarbon reservoirs, e.g., in the fl ow across fault zones within hydrocarbon res- 1985), but include issues that can potentially North Sea (Færseth, 1996). The host rock to the ervoirs. Previous studies of fl ow through fault be resolved with more accurate geospatial data. faulting is the predominantly aeolian, Permian- zones have used synthetic and theoretical rela- Ground-based laser scanning data provide an aged Cedar Mesa Sandstone. The main feature tionships when accounting for relay zones in opportunity for the collection of very high of interest is a graben system featuring a right- the determination of fl ow across faults (Childs resolution data that can be targeted to address lateral step or shift of the bounding faults, et al., 1995). The aim of this study was to col- these issues. Within the Roda Sandstone, these resulting in a right-lateral step of the entire gra- lect a spatially accurate data set from an out- problems include the detailed delta-clinoform ben (Fig. 4B). This type of stepping or shifting cropping example and to dynamically test the and bedset geometries, and the correlation is common in graben systems and is related effects of this one case on simulated fl ow. across unexposed areas. In the Grabens area to the evolution of the faults through segment the detailed architecture of the fault overlap growth and linkage (Peacock and Sanderson, OUTCROP TO MODELING zone and distribution of antithetic structures 1991; Rotevatn et al., 2007). In the step-over WORKFLOW are issues of interest. Digitally collecting the area, the bounding faults constrain two oppo- spatial data also eases the export of the data to sitely dipping relay ramps, both of which are The modeling workfl ow includes a step- geocellular modeling software. swarmed by an array of smaller faults and frac- wise procedure from the selection of outcrops tures. It is the complexity of this structural fi eld through the collection of data to the creation Roda Sandstone, Spanish Pyrenees case that is sought to be captured using lidar of a virtual outcrop, geological interpretation, technology. and fi nally, building and testing of geocellular The Eocene Roda Sandstone crops out in the models based on these data. Spanish Pyrenees and is interpreted as a pre- Geological Issues to be Addressed dominantly siliciclastic, wave- and tide-infl u- Outcrop Selection enced, Gilbert-type delta system (Leren, 2005; The Roda Sandstone shows an excellent Leren, 2007; Lopez-Blanco, 1996; Lopez- example of seaward-dipping delta front clino- Outcrops are selected based on four criteria: Blanco et al., 2003; Molenaar and Martinius, forms. Within shoreface and shallow-water suitability to the problem, level of three dimen- 1996) (Fig. 2). The Roda comprises a series of delta systems, they typically dip between 1° and sionality, outcrop quality, and accessibility. It Gilbert-type lobes with steeply dipping clino- 3°, while in deeper water and bedload-domi- is important that the problem drives the selec- forms. The entire unit comprises six seaward nated deltas, they may dip at up to 30° (Ander- tion of outcrop; i.e., a geological or reservoir (south-westward) stepping, delta-front bodies. son et al., 2004; Bhattacharya, 2006; Gani and issue is identifi ed and the optimal outcrop(s) This study focuses on two of these packages. Bhattacharya, 2005; Gilbert, 1885; Nemec and are selected to address that problem, rather than Individual lobes are separated by cemented Steel, 1988). The clinoforms record the basin- just collecting data because the outcrop quality hardgrounds, and the distal toeset deposits are ward migration of the shoreline through time is good. The term “3D outcrop” has become reworked by strong west-northwest-directed, (Hampson, 2000; Howell et al., 2006). Because common in some areas of the geological com- ebb-tidal currents (Molenaar, 1990; Yang and clinoform surfaces are frequently draped with munity to describe outcrops in which there are

Geosphere, December 2007 471

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al. A Map below N PY RENEES Jaca Basin Axial zone

Trem Boixols p-Graus Basi Ainsa Basin n Montsec Palaeozoic

41 Serres marginals Mesozoic

TUGAL SPAIN Palaeogene

POR Neogene & Quatertnary 0 Ebro Basin Fault Thrust 0 100 km Barce- lona Fold

Las B Aras CAMPO thr N us t

urbon T

Ser

r ols

e iv R T a h en rust ab Is Upper Eocene-Oligo- cene conglomerates L a

Upper Cretaceous- F T

o a r m

Palaeocene c a u

s r

s c y i C Eocene n a

c a

l n t in h e SERRADUY e r r Preorogenic- o u

l s Mesozoic a t n

t i c

S R l Early Eocene folds i n o e e d RODA r s r a a

d Late Cret. folds a u n y t - ic l i Late Cret. thrusts n e Study Area S i s CAJIGAR 5km s y n c l

in e N 1km Study Area C Serraduy

7 28 2 Isabenaer Riv 8 1 La Puebla 12 27 35m de Roda 26 19

24 San Esteban VO del Mall Roda de 25 Isabena 22 20 21 23

Figure 2. (A) Geological map of the Spanish Pyrenees, with location of the study area at the northeastern ﬂ ank of the Tremp-Graus Basin. (B) Structural map of the Roda Folds in the western part of the Tremp-Graus Basin. See Figure 2A for location. Modiﬁ ed from Lopez-Blanco (1996), Lopez-Blanco et al. (2003), and Leren (2007). (C) (Left) Virtual outcrop (VO) detail example of the Roda Sandstone from road cut on the west side of the Isabena River, south of La Puebla de Roda. See map (right) for location. (Right) Location of study area. See Figure 2A for location.

472 Geosphere, December 2007

Limit ofBandC 10m Map view (Bird’s eye) irtual outcrop of same area as (A) of same area irtual outcrop ), viewed from varying perspective in D, ), viewed from 20m F 5m 20m 20m E 20m 10m

20m

C ii fBand B of Limit Vegetation: ~1million points, Rock face: ~ 33,000 polygons (17,500 points) 25m 20m 20m D A B C with map view insert. (C) Same as (B) with interpretations (colored lines) of clinoform outlines. (D, E, and F) Details from (C lines) of clinoform outlines. (D, E, and F) Details from (colored with map view insert. (C) Same as (B) interpretations and zoom the virtual outcrop. showing the advantage of being able to move, rotate, respectively, E, and F, Figure 3. Comparison between photograph and virtual outcrop. (A) A photograph from the distal part of the Roda Sandstone. (B) V the distal part of Roda Sandstone. (B) photograph from A (A) Figure 3. Comparison between photograph and virtual outcrop.

Geosphere, December 2007 473

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

a number of different orientations to the outcrop faces, and in which geological surfaces and bodies can be easily extrapolated. While such outcrops are clearly not 3D volumes, the UTAH, A key aspect is that they have a far greater utility Moab USA 10 than simple, single, straight cliff sections that 40 Green provide a two-dimensional section through the geology. To quantify the level of three- er N River dimensionality that an outcrop expresses, a Area Riv 191 shown new parameter, termed the Outcrop Area Ratio Devil’s Lane, (OAR), is proposed. This is the ratio between Colorado The Grabens, the plan view length of exposure and the plan Canyonlands view area it occupies. The Roda Sandstone has National Park an OAR of 2.13 and the Grabens is 2.4, i.e., the Roda area is slightly better, while both are 0km20 good (Figs. 3–5). Whereas the OAR provides information about the level of three-dimensionality that an outcrop expresses, it does not quantify the quality of the outcrop. An outcrop B can have a good OAR, but still can be poor in terms of outcrop (e.g., the Roda Sandstone has a good OAR, but a variable outcrop quality, especially in the more proximal parts). Out- crop quality is a function of vegetation and/or scree cover and is also considered. Preferably, the exposure should also be close to both strike

and dip (structural or depositional) and should 50m

5 intersect these directions as well. If these cri-

800m teria are met, the geology can be better repre- N sented in a 3D model. In a subsurface study, it is typical to either 300m 0km4 model the entire ﬁ eld or, in the case of very large ﬁ elds, to model a portion or sector. The size of the model is commonly limited by the

00m computing and software capacity and is also 3 dependent on the planned application of the model. Typical subsurface models are between 4 and 100 km2 and between 20 and 200 m 900m thick. When modeling outcrops, the size of the model is also dependent on the purpose. The

550m N smallest outcrop models (dm to m) are com- N monly used to address fl uid-fl ow behavior in individual bedforms (Jackson et al., 2005). 0m200 Models in the hundreds of square meters scale have been used to study bedform and individual architectural elements (Falivene et al., 2006; Pedersen, 2005; Vipond, 2005). Larger models at the interwell (1–2 km2) to the entire oil fi eld (up to ~100 km2) have also been built. Long-range lidar as a means of data collection Figure 4. (A) Generalized map with location of study area in Devil’s Lane, The Gra- and capture lends itself to all but the smallest bens, Canyonlands National Park, Utah, USA, and the main geographical features of these scales. outlined. (B) (Top right) Digital elevation model of The Grabens (Data available The effective range of a typical, refl ector- from U.S. Geological Survey, EROS Data Center, Sioux Falls, South Dakota, USA). less laser scanner on rock is currently ~600 m, (Left) Satellite photo of Devil’s Lane, showing the right-lateral step of the graben. although this varies depending upon the The numbers are used for the Outcrop Area Ratio (OAR). The OAR quantifi es the instrumentation, the refl ectivity of the rock, level of three dimensionality that an outcrop expresses. For the Devil’s Lane area, the angle of the scan, and the atmospheric con- the ratio between the plan view length of exposure; the lines outlined have a total ditions. Buckley et al. (2006) have described length of 3400 m, and the plan view area it occupies (1.04 km2) is 2.4. (Bottom right) how the outcrop must be within range of the Colored elevation model of Devil’s Lane showing one of the two ramps in the area. laser scanner, both horizontally and vertically.

474 Geosphere, December 2007

dGPS- antenna Nikon D100 Scanner LMS-Z420i

dGPS

Lap-top

Battery

Figure 5. The laser scanner and associated equipment setup in the ﬁ eld for digital outcrop collection. From the collection of the Roda Sandstone, Spain. (B) Laser scanning in Devil’s Lane, the Grabens, Canyonlands National Park, Utah, USA.

Geosphere, December 2007 475

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

If the distance is too great, then the returns will lenses, a dGPS setup, a laptop computer, tri- optimal lighting conditions for each scan posi- be scattered and not representative. The same pod, mounting, batteries, and cables. The cam- tion will typically dictate fi eldwork planning. is true, if the angle between scanner and target era and one of the dGPS antennas are mounted The time taken to collect data from a single is too wide. This is especially an issue when on the scanner head, while the second dGPS scan position is dependent on the resolution looking up at steep cliff faces from below. is located at a semi-permanent base station and the width of the scan. Under typical fi eld Preferably, the scanner can be positioned at a (Fig. 5). Software on the laptop controls both operating conditions, scanning and associated level at or close to half the total height of the the scanner and the camera, and records the photography takes around one hour to collect, outcrop. This position secures the lowest angle scans and the images taken for later texturing making it possible to collect up to eight to ten between the laser beam and the rock, giving of the virtual outcrop. The GPS readings are scans in a working day. the strongest laser return and, hence, a better stored in the GPS receivers and downloaded to This project used a Riegl LMS-Z420i scan- representation. The nature of the topography the project at a later stage. Use of dGPS allows ner (Riegl, 2006), together with a Nikon D100 may require that the scanner is positioned all of the scans to have a common coordinate camera and a set of dGPS receivers. Ashtech above or below this position, but the outcrops system, which is extremely useful during later Solutions 2.70 was used for processing the should be selected to minimize such effects. processing when results from numerous scan GPS data. Riegl’s own commercial software, Corners and bends in the outcrop will also locations are merged. RiSCAN PRO version 1.4.1, was used both result in less than optimal angles and shadows During data collection, the scanner continu- to control the scanner and camera and for post where not all of the outcrop can be seen from a ously emits a low-energy laser beam at the processing of point clouds and generation of single position. This will typically require the outcrop as it slowly rotates around its own axis the virtual outcrop. Registration of the dif- outcrop to be captured from various positions up to 360°, normally less than 180° for each ferent scan positions was carried out in Poly- with a high degree of overlap, both vertically scan position. The travel time of the refl ected works version 9.0.2, using a surface-matching and horizontally. light is used to calculate the distance to the approach to adjust the scan positions based on The laser return can be obstructed by any point of refl ection on the outcrop. Together, the overlap of each scan. object in the line of sight between the scanner the points produce a 3D point cloud. Using the and the outcrop. Consequently, care should be azimuth, inclination, and distance of the laser Data Processing and Generation of the taken to select scan positions where the outcrop return, the software calculates the XYZ coor- Virtual Outcrop is not obscured by obstacles, such as shrubs, dinate for each scan point and ensures a con- trees or other vegetation, large boulders, and sistent registration. The geometric relation- The data processing that leads to the creation masts or other man-made objects. Shadow- ship between the scanner and mounted SLR of a fi nished virtual outcrop is, at present, very ing from minor obstacles can be avoided by camera is calibrated at the scan site by using labor intensive. The procedure follows the fol- scanning from different positions. However, refl ectors (typically, six to eight) that are posi- lowing stages (Figs. 6 and 7): laser returns from the obstacles may need to tioned in different locations within the scan. be manually removed during post processing. Recording the refl ectors using the scanner and Stage 1. Post Processing of the GPS Data to This can be time consuming and should be camera allows the mounting calibration to be Include the Differential Correction taken into consideration during outcrop selec- updated, accounting for a very small change The GPS data for each scan position are pro- tion. Precipitation will return or diffract the in camera position when it is removed while a cessed relative to the static base station, so that laser beam, making arid areas most suitable for lens is changed. errors are minimized. application. Typically, a data set will have millions of The fi nal consideration for the selection of points with an accuracy of around ±0.02 m for Stage 2. Combination of Data from Single study areas is the portability of the system and each point when the laser is shot at a range of Scans into One Project access to the outcrops. The total weight for up to 600 m (Riegl, 2006). One or more series A single project will typically contain data the system used in this study is around 70 kg, of automatically registered digital photos are from between three and twenty scan locations. with batteries generally needing to be charged taken together with each scan by the mounted These data are combined into one data set that daily. While lighter systems exist and helicop- camera. The scanner has a vertical spread of will typically contain millions of data points. ters can go anywhere, our studies suggest that 80º, while the fi eld of view of the camera lens is The use of a single project coordinate system it is practical to work within ~2 km of vehicle commonly less, requiring additional photos to renders the possible combination of an unlim- access (which may include off-road jeeps and be taken with the mount for the camera tilted. ited number of scan positions and also allows quad bikes). This can signifi cantly affect out- In case of poor lighting conditions during scan- the integration of other data, such as sedimen- crop selection. ning, the camera can be dismounted from the tary logs, within one reference system. In summary, suitable outcrops: (1) address scanner, and photos can be taken separately, the geological problem; (2) are accessible even at a different time. These photos can be Stage 3. Coloring of the Point Cloud by vehicle; (3) have a high outcrop-area ratio manually referenced, but must be taken with Data from the photographs can be used to (OAR); (4) can be scanned at a close to hori- a camera that has been calibrated. The photos add RGB (red, green, blue) values to each of the zontal orientation; (5) have limited vegetation are used to color the point cloud (i.e., assign an points. This produces an image that looks similar cover; and (6) are in arid areas. RGB property to each point) and also to tex- to a “somewhat grainy” photograph. The colored ture the processed outcrop model (i.e., they are point cloud can be used for the mapping and cor- Data Collection draped onto the surface). The quality of the pho- relation of key surfaces and the identifi cation of tographs is a key aspect that controls the quality larger geobodies. Many groups working with Data collection in the fi eld requires a laser of the fi nal virtual outcrop. It is important that lidar data focus almost exclusively on the colored scanner, a digital single-lens refl ex (SLR) photos are taken in optimal lighting conditions point cloud. In the present study, a higher degree camera and photogrammetrically calibrated without strong shadows or haze. Obtaining of detail was required than is obtainable using the

476 Geosphere, December 2007

are removed from areas of little interest (e.g., scree slopes and foreground), while maintain- Data collection (points, photos, GPS) ing detail in areas where it is required. Nor- mally, processing using an octree ﬁ lter will produce a satisfactory result. The raw data are stored so that higher resolution, triangulated Time to virtual outcrops utilizing all available data can (1) Post processing of GPS data complete: be built from smaller areas of special interest, ~ hour if required.

Stage 5. Triangulation and the Creation of the DEM The points of the decimated cloud are con- (2) Combination of single scans Time to a) Register scans using matching complete: nected by triangles in a triangulation operation b) Merge/combine individual scans ~ hours to form a mesh surface, or digital elevation model (DEM) that can be textured (see Bellian Virtual Outcrop et al., 2005; Buckley et al., 2006 for details). While this process is largely automated, it Time to involves a series of user-deﬁ ned parameters

eneration of complete: that are required to produce a reasonable sur-

g (3) Coloring of point cloud << hour face, such as manually setting the maximum edge length and angle between two adjacent triangles. In this project, RiSCAN PRO version Time to 1.4.1 has been used for this purpose, although

processing and (4) Point cloud cleaning and decimation complete: experimental triangulation has been performed a ~ days using different software packages.

Dat After the triangulated DEM is produced, it is prudent to carry out a visual quality check. It is commonly necessary to manually adjust the tri- (5) Triangulation and creation of DEM (mesh) Time to angulated surface due to erroneous points and a) Combination of points by triangulation complete: errors in the triangulation procedure. In-house b) Texture w. photos to create Virtual Outcrop days / weeks software has been developed to create a difference surface that records the spatial difference between the DEM and the original point cloud. This surface illustrates the degree of spatial error that has been introduced by the triangulation Interpretation of lines etc. on Virtual Outcrop process. This surface can be used to determine whether quality of the virtual outcrop within the areas of interest matches the original higher resolution point cloud (Fig. 8). Acceptable errors depend upon the proposed application of the Figure 6. Data collection and preparation of virtual outcrop workfl ow, and approximate virtual outcrop. time needed to complete each stage of the data processing part of the workfl ow. See text for explanation of each stage of the procedure. Stage 6. Texturing the DEM The high-resolution digital imagery captured with the scans or added to the project later is used to render the triangulated mesh. The greater resolution of the image data allows con- point cloud alone, and, therefore, virtual outcrops topography and other features that can produce tinuous coverage of the required geological fea- based on textured surfaces were generated. unwanted triangulation effects. tures. The rendering of the images is carried out To be able to generate a useable triangulated automatically in RiSCAN PRO, which selects Stage 4. Point-Cloud Cleaning and model, the point cloud must be decimated. This the optimum photography for each triangle. Decimation involves removal of a certain proportion of the This is typically effi cient, but results are vari- Given the limitations of currently available points to enable the surface to be triangulated able because the lighting conditions and qual- software, the raw, combined point cloud has to and visualized on a typical computer. It is not ity of the image portions selected for adjacent be modifi ed before it can be triangulated. This unusual to remove 50% of the points, although triangles are often different. This problem can involves a combination of both automated and this is not carried out in a uniform way. Built- be partially mitigated by removing very poor manual processing. The procedure includes in fi lter modes can perform different decima- photos from the project. It is also possible to cleaning, through the manual removal of veg- tion operations, e.g., by octree-fi ltering, and manually adjust the colors and lighting of pho- etation, and the fi ne tuning of sharp changes in manual editing can ensure that the most points tos so that they are more similar. This is done

Geosphere, December 2007 477

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al. A

25m

Figure 7. The three steps of virtual outcrop preparation, each with enlarged detail-box insert. (A) Point cloud colored from photos (but not textured). (B) Triangulated point cloud (green/blue colors). Also visible is colored point cloud of vegetation (top and bottom). (C) Triangu- lated model textured with digital photos to form a virtual outcrop. Also visible is colored point cloud of vegetation (top and bottom).

478 Geosphere, December 2007

Edge effects/ interpolation

Vegetation removal

Figure 8. Data are typically reduced before meshes are produced to save resources. A difference surface records the spatial difference between the digital elevation model and the original point cloud. This surface spatially illustrates the degree of error that has been introduced by the triangulation process. Vegetation removal is by far the biggest error source.

using image-editing software, such as Adobe Visual inspection of the data allows improved of the beds within the log are interpreted from Photoshop. understanding of bedforms and bedform the photographs and, ideally, calibrated to true The resultant textured DEM (the virtual out- geometries, the correlation of key surfaces, fi eld logs. Such logs can also be digitized and crop) captures the outcrop morphology and and, depending upon resolution, improved loaded into the reservoir modeling system as detail. The virtual outcrop can be loaded into a understanding of facies geometries and transi- wells (Falivene et al., 2006). Faults can be viewer and examined from any angle, and used tions (Fig. 3). In addition to qualitative visual mapped as planes, and accurate measurements for correlation and training. As each pixel in the inspection, a key utility of the virtual outcrop is of fault displacement along strike can also be virtual outcrop has an XYZ position, measure- the ability to extract quantitative spatial data— made directly from the virtual outcrop, pro- ments can be made and surfaces and features either manually or in an automated fashion. vided at least one continuous reference bed can be traced and digitized. Manual data extraction involves the user exists. viewing the virtual outcrop and manually There is currently no commercially avail- Working with the Virtual Outcrop digitizing points along a surface such as a bed able software for the automated extraction of boundary or fault plane. The points can then geologically meaningful spatial data from the A variety of both commercial and freely be stored and exported as individual points or virtual outcrop. Several research groups are available software can be used to visualize the polylines. Other forms of manual data extrac- working to create software using a number of virtual outcrop, although nothing designed spe- tion involve the measuring of surface strikes novel approaches including algorithms similar cifi cally for geological study is yet available, and dips using three user-selected points and to those used in the automated interpretation of and consequently, all have their limitations. In creating sedimentary logs. In the latter, a poly- seismic data (e.g., Monsen, 2006). This will be addition, in-house software has been developed line representing the log trace is highlighted on a signifi cant growth area in the near future. that permits rapid viewing and manipulation of the virtual outcrop. Points along the line that Once mapped and interpreted in three dimen- the large volumes of data on a typical personal represent bed boundaries are picked and used sions, the points, polyline, and log data can be computer. to generate a sedimentary log. The properties exported to a reservoir modeling software to

Geosphere, December 2007 479

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

allow the construction of surfaced-based geo- Model Building models. Outcrops can be used to provide direct cellular models. Data are typically exported in inputs for property modeling (e.g., shale bed ASCII (American Standard Code for Informa- A reservoir model is a 3D quantitative repre- lengths, fl uvial channel, width versus thick- tion Interchange) formats that are suitable for sentation of a volume of rock within a computer. nesses; see Reynolds, 1999) and can be mod- the chosen software. Reservoir modeling has become a necessary eled to understand the behavior of a particular and integrated part of predicting, planning, and type of system. Export to the Reservoir Modeling Package updating information concerning subsurface res- The challenges associated with modeling ervoirs, and, as a database, it comprises a large outcrops are somewhat different than those Irap Reservoir Modeling System (Irap RMS) amount of geological, petrophysical, and gen- faced when dealing with the subsurface. Most is a commercial reservoir-modeling package eral production data. The models have a wide reservoir modeling packages are designed for from Roxar that is widely used in the oil indus- area of application and are used for calculating modeling subsurface reservoirs on a scale of try for the visualization and simulation of sub- volumes, well planning, and predicting the paths several to tens of kilometers. Although most surface oil-fi eld data. This package can also be of fl uids during production. Models are limited of the tools and algorithms are scale indepen- used, with some modifi cation for handling and by the available geological data that are used to dent, some adjustments are necessary when visualizing data extracted from the virtual out- build them. Outcrop analogs from comparable working with outcrop data. Secondly, outcrop. The defi nition and mapping of the inter- systems can be used to provide additional input crops provide very high resolution informa- section of the geological surface and the out- to models, especially in an early stage of fi eld tion that is somewhat spatially limited. While crop in the virtual outcrop produces a series of development when subsurface data are limited this is superior for the data available from well points and lines that are exported to Irap RMS and there are no production data. During later logs in the subsurface, considerable extrapola- as DXF or text fi les, using Irap RMS internal fi eld life, analogs are more commonly used to tion is required away from, and between, cliff data format (Roxar, 2006). Other data types, improve understanding of the geological system sections. Additionally, models built to study such as sedimentary logs, can also be digitized that has controlled production and are used as stratigraphic issues require the removal of later and loaded into the modeling software. a quality check on history-matched dynamic tectonic deformation (tilting, folding, or even

2. Importing lines for surfaces 3. Generate surfaces 1. Interpretations on Virtual Outcrop

5. Grid design, property modeling & flow simulation 4. Zoning

Figure 9. Generalized workﬂ ow for reservoir model building. The method includes steps from interpreting on virtual outcrop and importing interpreted features into reservoir modeling software (steps 1 and 2), via surface generation (3) to building of reservoir models based on imported data (4 and 5). See text for further explanation.

480 Geosphere, December 2007

faulting) that are not relevant to the use of the ated, the scalar operations (e.g., depth = depth − achieved in a number of stochastic ways (using stratigraphic architecture as an analog. This is 2000 m) can be used to move the surfaces (and either Gaussian or Boolean type approaches— commonly done with specifi c structural restora- thus the model) into a typical depth regime for e.g., Falivene et al., 2006; Holden et al., 1998; tion packages, such as 3DMove (Fernandez et a reservoir. MacDonald and Halland, 1993) or by simple al., 2004). Finally, many of the algorithms used extrapolation of the facies body margins. In all for simulating fl uid fl ow are dependent on cer- Building a Fault Framework cases, the data are conditioned to the outcrop tain pressure, depth, and temperature relation- In traditional reservoir modeling of subsurface observations. Grids are normally designed to ships. Therefore, it can be necessary to “move” reservoirs, a key step after the initial data import specifi cally follow the key geological heteroge- the outcrop model to a typical reservoir depth and stratigraphic surface generation is building a neities because they control fl uid fl ow from for- (e.g., around 2000 m) for these calculations to structural framework based on the imported fault mation/reservoir level to lamina and pore level be relevant to reservoir-related issues. surfaces. In this paper, faults are a key feature of (Weber, 1986). Finally, both surfaces and grids Building the models involves a series of the Canyonlands case study (Figs. 4 and 11B). are adjusted to mapped faults, and the grid is stages, which are broadly similar to the proce- There is also a minor fault that has been modeled displaced accordingly. These faults potentially dure for modeling a subsurface data set (Fig. 9). in the Roda study area. The point data imported have a major infl uence on fl uid fl ow. These are discussed below. from the Canyonlands virtual outcrop represent Current limitations of computer hardware fault polygons extracted from the exposed fault and software restrict the number of cells that Surfaces scarps and fault surfaces in the virtual outcrop. can be represented within a cellular-reservoir When working with subsurface data, the Having extracted the fault data from the virtual simulation model, and, consequently, the reso- fi rst stage of the RMS modeling workfl ow is outcrop, and having measured the displacement lution of input data. To streamline the models to import seismically mapped stratigraphic and changes along strike in the virtual outcrop, the and save memory costs, the 3D grids are often structural (fault) surfaces. These surfaces are fault model is produced using the preexisting designed with a very large X and Y spacing, visually checked and tied in to the well data. algorithms in the RMS software package and with Z much smaller (e.g., 50 × 50 × 0.5 m). The imported data are then used to build a editing the resulting structural model as neces- This design refl ects the fact that, in most sedi- structural, surface-based framework for the sary (Fig. 11). The fault model is then used to mentary systems, the properties are more homo- modeling. Surfaces form the framework and re-grid the stratigraphic surfaces accounting for geneous in the X and Y direction, and it is the Z zone boundaries of the reservoir model and the displacement. The fault model is also impor- direction that needs to be captured with higher represent limits where changes in lithology and tant for the modeling grid. resolution. petrophysical properties occur. Faults are also Clinoforms of the Roda Sandstone exhibit a represented by surfaces. Grids and Grid Population systematic facies transition from delta front to Virtual outcrop data are somewhat different. After the surfaces are generated and adjusted, toesets, with no sharp facies changes vertically Polylines that represent the outcrop expression they are then used to create modeling zones. (Figs. 5 and 12). Their thickness varies from up of a surface are not in themselves continuous The 3D grid is created within each of the zones to a few meters in the proximal, up-dip portion surfaces; therefore, the surfaces need to be (Figs. 10B and 10C). The 3D grid is the cel- to zero or close to zero in the distal or down-dip generated from them. This is done statistically, lular framework in which all of the facies and portion. This allows the individual clinoforms to and RMS contains a variety of algorithms for property modeling within RMS take place. be presented in the model as a zone that is one the extrapolation of surfaces, each one produc- Grid scale and design is based upon the scale cell thick. For example, a model covering 200 × ing somewhat different results from the same and nature of the geology that is being modeled, 200 × 30 m and containing ten clinoforms, may input data. Visual quality check and compari- and there is a degree of fl exibility in the way in contain as little as 50–100 cells, although sev- son to the conceptual geological model is used which a grid can be built. To create a model- eral thousand would be more typical. This is in to determine which algorithms are producing ing grid, it is necessary to defi ne the grid type, contrast to a model built of an entire delta lobe the best results. As a general rule, the global the horizontal and vertical layout, and the cell of 2000 × 2000 × 50, in which several hundred b-spline produces the most geologically real- truncation. The resolution selected is usually a thousand cells may be used. istic results. In many cases, manual editing of compromise between necessary resolution and the surfaces away from the control points is computer memory limitation. Model Investigation required to further satisfy the conceptual model. The grids need to be populated with prop- The fi nal models can be analyzed both RMS requires that all of the surfaces cover the erties; in most models, these are facies based statically and dynamically. Static examination entire model area and do not cross each other (Fig. 12). In virtual outcrop models, properties involves the visual inspection and extraction of (Fig. 10). Editing is done by introducing guide at the outcrop are interpreted and placed directly quantitative data on body geometry, including points, guide contours, and by using trends to into appropriate grid cells or added from the the extraction of size variograms such as those guide the surfaces in the correct direction. sedimentary logs. Sedimentary logs imported as presented in Reynolds (1999) that can be used for It is useful to generate a surface that repre- deviated wells help to constrain the model, and the population of subsurface models where such sents the present-day topography to assist in facies modeling can be conditioned on wells. data are not available. Dynamic investigation the quality control of the stratigraphic surfaces Logs have to be “blocked,” or averaged, so that involves simulating the fl ow of fl uids through (Figs. 10E and 10F). If required, the removal each cell in the grid only contains one property. the model to understand how it would behave of tectonic dip can be performed in a separate There are a number of different ways that this as a reservoir. Dynamic simulation requires the software package as discussed earlier in this can be achieved. assignment of petrophysical properties to the paper. Stochastic algorithms can be used to The population of grid cells away from the grid cells. Petrophysical data from analogous reintroduce small-scale irregularities that are outcrop involves a degree of interpretation. subsurface systems can be used to populate lost between outcrops, if suitable (Falivene et Depending upon the conceptual geological the models on a facies-based approach. Using al., 2004). When the surfaces have been cre- model and the prior knowledge, this can be petrophysical numbers from outcrops does not

Geosphere, December 2007 481

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al. A B

C D

E F

G H

Figure 10. (A, B) Examples of surfaces in reservoir modeling software based on features (visible in red and blue) imported from the virtual outcrop. (C) Geocellular model of (B). (D) Transparent surface and cross sections. (E, F) Examples of surfaces and DEM, in (F) with cross sections. (G) Surface and crossing fault-surface. (H) Three reservoir modeling software-generated surfaces visualized with virtual outcrop.

482 Geosphere, December 2007

N N

Figure 11. Based on data imported from the virtual outcrop, a structural framework of faults is built. (A) This fault model is used to adjust the reservoir model. (B) Any surfaces and reservoir zones in a model can be adapted to the modeled faults, and the grid can be displaced accordingly. The block is 800 m × 1300 m in size.

Typical size facies model Typical mini-model

400m 400m 30m

2000m 3000m

Figure 12. Typical geocellular reservoir models. Color codes for lithofacies: yellow—sand; gray—sand/mud (heterolithics); red—sand (tidal bar). (Left) A geocellular model of the Roda Sandstone at typical oil-ﬁ eld scale, built on the basis of virtual outcrop data. (Right) At top, an outcrop-based mini-model with surfaces and zones generated from features imported from the virtual outcrop. At bottom, a cross section of the mini-model with one isolated facies visualized (red—sand (tidal bar)).

Geosphere, December 2007 483

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

necessarily give the desired analog values and the effect of the fault framework and the pres- while the production well produces fl uid (oil) sometimes can be diffi cult to collect correctly ence of relay ramps on fl ow was devised. Both until the water reaches it (water breakthrough). due to weathering or accessibility issues. On the two-phase fl uid fl ow and streamline simulation The result shown in Figures 15 and 17 demon- other hand, petrophysical data from outcrops models were run. Streamlines follow the path- strates that, despite the apparent connectivity can give a higher sample resolution and better way of a particle of fl uid through the volume problems shown in Figure 14, communication control, if collected in a systematic manner, e.g., at different time steps. To isolate the effects of is preserved, due to the overlap between the by a facies approach (e.g., Froster et al., 2004). the faults, the bulk rock properties were set as faults in the graben over-step area. Simulation All faults are also assigned values for trans- a homogenous volume. Porosity was set at 30% of a comparable but unfaulted volume illustrates missibility. This value determines the degree for the entire model, and, correspondingly, per- that the presence of the faults increases the tor- to which any fault in the model permits fl uids meability was set to 1000 mD (Kh) and 100 mD tuosity of the fl ow path and delays the time of to pass through the fault. This has a signifi cant (Kv). Net/gross ratio was set at one. Fault trans- breakthrough (Fig. 16). impact on reservoir performance, but often the missibility was set to zero to make the faults The infl uence of the faults can be quanti- lack of such data for subsurface faults makes the completely sealing. fi ed by running streamline simulation on an assignment of transmissibility values seem more Two wells were placed on opposite sides of unfaulted and faulted model. This simulation like guesswork. By using realistic values based the overstepping fault system—one injection was undertaken with three different petrophysi- on empirical data from actual faults in the fi eld, well and one production well (Fig. 13). A pro- cal setups in the host rock (Table 1) to determine a greater understanding of how faults affect fl uid fi le between the two wells illustrates the lack of whether the faults have more or less effect in fl ow in subsurface reservoirs can be achieved. two-dimensional lateral connectivity, due to the lower (or higher) permeability settings. For the Canyonlands reservoir model, fl uid- faults, while a visual inspection of the 3D model The simulation shows that the lower the per- fl ow simulation has been conducted, demon- would predict some fl ow between the wells; fl ow meability and porosity, the greater the dissimi- strating the fi nal part of the workfl ow. As an simulation allows this to be quantifi ed (Figs. 13 larity between the faulted and unfaulted res- initial approach, an experiment to investigate and 14). Water is injected into the injection well ervoir. In other words, in this particular case,

Figure 13. Faulted reservoir model screen capture showing well placements. Note that the injection well and the production well are placed on opposite sides of the graben system. The block is 800 m × 1300 m.

484 Geosphere, December 2007

a better reservoir quality lowers the infl uence TABLE 1. PERMEABILITY AND POROSITY VALUES OF SETUPS IN THE of the fault (Table 1). The streamline results HOST ROCK USED IN THREE DIFFERENT STREAMLINE were confi rmed by dual phase, fl uid-fl ow SIMULATIONS ON AN UNFAULTED AND A FAULTED MODEL simulations (Fig. 17). These detailed results Kx Ky Kz Porosity Unfaulted, time Faulted, time are beyond the scope of this paper and will be 1000 1000 100 0.3 4008 9448 published elsewhere. 100 100 10 0.25 4396 26,568 10 10 1 0.15 6220 325,033 Error Examination Note: See text for discussion of results.

There is high potential for errors to exist and propagate throughout the workfl ow. The aim of this work is a general improvement in reservoir pled at discrete intervals. This means that the In particular, the focus has been on (1) the modeling by using accurate spatial data, while geological surfaces are likely to be defi ned with accurate representation of geological entities minimizing the error sources at each processing higher accuracy, which, thus, will allow more from outcrops within a computer (referred to stage to minimize the overall uncertainty in the accurate reservoir models. in this paper as a virtual outcrop); (2) utilizing fi nal model. In the past, with only approxima- the virtual outcrop to extract data for build- tions of geometric information taken at discrete CONCLUSIONS ing and testing 3D geocellular models using intervals during sedimentary logging, undefi ned conventional hydrocarbon, reservoir-modeling error could be introduced to the extrapolation of Geometric data from outcrops and the model- software; and (3) applications of the collected geology surfaces (e.g., Jones et al., 2004; Prin- ing of outcrops using subsurface technology has data and the virtual outcrop. gle et al., 2004b). This, in turn, may result in started to bridge the gap between well bore and This paper documents a complete work- further error when a 3D grid was made, thus seismic methods and fi ll the gaps in our under- fl ow—from outcrop selection through data affecting the geometry of the resulting model, standing of the 3D geometries of geological sub- collection, processing and building of a virtual any volumetric calculations made, and fl uid surface entities. Qualitative and quantitative out- outcrop, and geological interpretation—to the fl ow simulations. Modeling of the outcrop crop analog studies can be used for this purpose. building of the 3D geocellular models. The geometry using terrestrial laser scanning gives In this study, we have presented new meth- workfl ow is illustrated with two case stud- far better constraints on the available outcrop odologies for the acquisition and utilization of ies that illustrate the application to sedimen- exposure. The stratigraphic layers can be fol- 3D information generated by the ground-based tary and structural-reservoir, geology-related lowed continuously, instead of only being sam- laser scanning (lidar) of geological outcrops. problems.

Injection Production Well Well

Permeable Rock Impermeable Rock

Figure 14. Vertical two-dimensional section (profi le) of the reservoir model between the two wells (see Fig. 13 for location). The two-dimensional section implies a lack of communication between the two wells due to vertical separation of reservoir unit. The streamlines simulation and fl ow simulation investigate whether this is also the case in the third dimension. The profi le is 800 m across.

Geosphere, December 2007 485

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

Figure 15. Streamlines simulation results. Screen captures are shown from various time steps between t01 to t45. Each streamline represents the pathway of a droplet of water from the injector to the producer. In the ﬁ nal time step, the streamlines that have achieved communication between the injector and the producer are colored red. The block is 800 m × 1300 m.

486 Geosphere, December 2007

Injector

Producer

Figure 16. Unfaulted reservoir model screen capture showing well placements. Note that the injection well and the production well are the same as shown in Fig. 16, but the graben system and the associated faults are not present. The block is 800 m × 1300 m.

Outcrops are selected based on four criteria: Models are limited by the availability of the ous surfaces. Therefore, the surfaces need to be suitability to the problem, level of three dimen- geological data that are used to build them. statistically generated from them. After the sur- sionality, outcrop quality, and accessibility. Outcrop analogs from comparable systems can faces are generated and adjusted to faults, they The data processing that leads to the creation be used to provide additional input to models, are then used to create modeling zones. The 3D of a fi nished virtual outcrop is, at present, very especially in an early stage of fi eld development grid is created within each of the zones and is labor intensive. The procedure follows the fol- when subsurface data are limited. During later the cellular framework in which all of the facies lowing stages: (1) post processing of the GPS fi eld life, analogs are more commonly used to and property modeling within RMS take place. data to include the differential correction; (2) improve understanding of the geological sys- The fi nal models can be analyzed both combination of data from single scans into tem and for quality checking history-matched statically and dynamically. Static examination one project; (3) coloring of the point cloud; (4) dynamic models. Outcrops can be used to pro- involves the visual inspection and the extraction point-cloud cleaning and decimation; (5) trian- vide direct inputs for property modeling. They of quantitative data on body geometry. Dynamic gulation and the creation of the DEM; and (6) can also be modeled to understand the behavior investigation involves simulating the fl ow of texturing the DEM. of a particular type of system. Building the mod- fl uids through the model to understand how it A variety of commercial, freely available and els involves a series of stages, which are broadly would behave as a reservoir. Dynamic simula- in-house software is used to visualize and pro- similar to the procedure for modeling a subsur- tion requires the assignment of petrophysical cess the virtual outcrop. Once mapped and inter- face data set. properties to the grid cells. preted in three dimensions, the point, polyline, Contrary to seismically mapped stratigraphic There is high potential for errors to exist and and log data can be exported to reservoir-mod- and structural (fault) surfaces from the subsur- propagate throughout the workfl ow. The aim of eling software to allow the building of surface- face, polylines that represent the outcrop expres- this work is a general improvement in reservoir based geocellular models. sion of a surface are not in themselves continu- modeling by using accurate spatial data, while

Geosphere, December 2007 487

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

minimizing the error sources at each processing stage to minimize the overall uncertainty in the ﬁ nal model. Consequently, geometric data from outcrops, and, most recently, the modeling of Production well Water outcrops, has started to bridge the gaps in our injection understanding of the 3D geometries of geologi- well t1 cal subsurface entities.

ACKNOWLEDGMENTS

This work is sponsored by the Norwegian Research Council and Statoil via the RUM project. We are grateful to Tony Reynolds and Jamie Pringle for constructive reviews on the initial version of this manuscript. We acknowledge all of our collaborators within the Virtual Outcrop Geology program at the Centre for Integrated Petroleum Research for assis- tance with ﬁ eldwork and software issues. We thank Statoil for the permission to use the satellite-photo in Figure 4B. Thanks to Roxar for providing the RMS software package and to Midland Valley for supplying 3DMove. t2 REFERENCES CITED

Ackermann, F., 1999, Airborne laser scanning—Present status and future expectations: ISPRS Journal of Pho- togrammetry & Remote Sensing, v. 54, p. 64–67, doi: 10.1016/S0924-2716(99)00009-X. Adams, E.W., Csoma, A.E., Grélaud, C., Loïc, B., and Hol- lis, C., 2007, Improving Middle East carbonate reservoir models with digital outcrop modeling: Abstract Volume, Understanding earth systems pursuing the checkered fl ag: 2007 AAPG Annual Convention and Exhibition (April 1–4, 2007) Long Beach, California: Los Angeles, USA, American Association of Petro- leum Geologists. Adams, E.W., Grotzinger, J.P., Watters, W.A., Schroeder, S., McCormick, D.S., and Al, S.H.A., 2005, Digital characterization of thrombolite-stromatolite reef distribution in a carbonate ramp system (terminal Protero- zoic, Nama Group, Namibia): American Association of Petroleum Geologists Bulletin, v. 89, p. 1293–1318, doi: 10.1306/06160505005. t3 Aiken, C.L., 2006, Ten years of 3D digital mapping of outcrops geology by the University of Texas at Dallas: Real Earth Models, Abstract volume: Unlocking 3D earth systems—Harnessing new digital technologies to revolutionize multi-scale geologic models: Dur- ham, United Kingdom, Geological Society of America Penrose Conference, 17–21 September 2006, Durham University. Alexander, J., 1993, A discussion on the use of analogues for reservoir geology: Advances in reservoir geology, Vol- ume 69: London, United Kingdom, Geological Society of London, p. 175–194. Anderson, P.B., Chidsey, T.C., Jr., Ryer, T.A., Adams, R.D., and McClure, K., 2004, Geologic framework, facies, paleogeography, and reservoir analogs of the Ferron Sandstone in the Ivie Creek area, east-central Utah, in Chidsey, T.C., Jr., Adams, R.D., and Morris, T.H., eds., Regional to wellbore analog for fl uvial-deltaic reservoir modeling; the Ferron Sandstone of Utah: Ameri- can Association of Petroleum Geologists Studies in Figure 17. Two-phase, fl uid-fl ow simulation of the Devil’s Lane Model. Water is being Geology 50: Tulsa, Oklahoma, United States, Ameri- can Association of Petroleum Geologists, p. 331–356. injected into the right-hand well to displace oil that is produced from the right-hand well. Baltsavias, E.P., 1999, A comparison between photogram- Colors indicate fl uid saturation (red—water; pink—oil; the greens and blues are interme- metry and laser scanning: ISPRS Journal of Photo- diate). The three time steps, t1–t3, demonstrate how the water front (greens and blues) grammetry & Remote Sensing, v. 54, p. 83–94, doi: 10.1016/S0924-2716(99)00014-3. is moving (communicating) across the critical step-over area of the bounding faults. This Baltsavias, E.P., Favey, E., Bauder, A., Boesch, H., and Pat- particular simulation was run over a period of ten years with synthetic shallow-marine eraki, M., 2001, Digital surface modelling by airborne laser scanning and digital photogrammetry for glacier sedimentary facies, and water breakthrough in the producing well was reached within less monitoring: Photogrammetric Record, v. 17, p. 243– than fi ve years (not shown in any of these fi gures). 273, doi: 10.1111/0031-868X.00182. Bellian, J.A., Kerans, C., and Jennette, D.C., 2005, Digital outcrop models: Applications of terrestrial scanning

488 Geosphere, December 2007

lidar technology in stratigraphic modelling: Jour- logical Society of America Penrose Conference, 17–21 American Association of Petroleum Geologists Bulle- nal of Sedimentary Research, v. 75, p. 166–176, doi: September 2006, Durham University. tin, v. 89, p. 507–528, doi: 10.1306/11230404036. 10.2110/jsr.2005.013. Falivene, O., Arbues, P., Gardiner, A., Pickup, G., Munoz, Jones, R., McCaffrey, K.J.W., Imber, J., Wightman, R., Healy, Bhattacharya, J.P., 2006, Deltas, in Posamentier, H.W., and J.A., and Cabrera, L., 2006, Best practice stochastic D., Guo, J., and Smith, S., 2007, Quantitative structural Walker, R.G., eds., Facies models revisited: Tulsa, facies modeling from a channel-fi ll turbidite sand- analysis of reservoir outcrop analogues, Abstract vol- Oklahoma, SEPM (Society for Sedimentary Geology) stone analog (the Quarry outcrop, Eocene Ainsa Basin, ume: From outcrop to asset: Recent advances in digi- Special Publication 84, p. 237–292. northeast Spain): American Association of Petro- tal outcrop data collection and modelling techniques, Brekke, H., Sjulstad, H.I., Magnus, C., and Williams, R.W., leum Geologists Bulletin, v. 90, p. 1003–1029, doi: 10–11 January 2007, University of Manchester, Atmo- 2001, Sedimentary environments offshore Norway— 10.1306/02070605112. spheric & Environmental Science: Manchester, United An overview, in Martinsen, O.J., and Dreyer, T., eds., Falivene, O., Arbues, P., Howell, J.A., Fernandez, O., Munoz, Kingdom, The Geological Society. Sedimentary environments offshore Norway—Palaeo- J.A., and Marzo, M., 2004, Facies trends and modeling Jones, R.R., McCaffrey, K.J.W., Wilson, R.W., and Hold- zoic to Recent, Volume 10: Amsterdam, Norwegian strategies within slope turbidite systems: Insights from sworth, R.E., 2004, Digital fi eld data acquisition: Towards Petroleum Society Special Publication, Elsevier Sci- 3D modeling of outcrop and subsurface data from the increased quantifi cation of uncertainty during geological ence B.V., p. 7–37. middle Eocene Ainsa system, south Pyrenean foreland mapping: Geological prior information, informing science Bryant, I.D., Carr, D., Cirilli, P., Drinkwater, N., McCor- basin, Spain, American Association of Petroleum Geol- and engineering, Volume 239: London, United Kingdom, mick, D., Tilke, P., and Thurmond, J., 2000, Use of 3D ogists annual meeting, Volume 13: Tulsa, Oklahoma, Geological Society of London, p. 43–56. digital analogues as templates in reservoir modelling: USA, American Association of Petroleum Geologists Joseph, P., Hu, L.Y., Dubrule, O., Claude, D., Crumeyrolle, Petroleum Geoscience, v. 6, p. 195–201. and Society for Sedimentary Geology, p. 43. P., Lesueur, J.L., and Soudet, H.J., 1993, The Roda del- Bryant, I.D., and Flint, S.S., 1993, Quantitative clastic reser- Fernandez, O., Munoz, J.A., Arbues, P., Falivene, O., and taic complex (Spain); from sedimentology to reservoir voir geological modelling: Problems and perspectives, Marzo, M., 2004, Three-dimensional reconstruc- stochastic modelling, Subsurface reservoir character- in Bryant, I.D., ed., The geological modelling of hydro- tion of geological surfaces: An example of growth ization from outcrop observations, Volume 51: Paris, carbon reservoirs and outcrop analogues, Volume 15: strata and turbidite systems from the Ainsa Basin France, Technip, p. 97–109. Oxford, International, Blackwell, p. 3–20. (Pyrenees, Spain): American Association of Petro- Lee, K., Bouroullec, R., Tomosso, M., and Ambrose, W., Buckley, S.J., Howell, J.A., Enge, H.D., Leren, B.L.S., leum Geologists Bulletin, v. 88, p. 1049–1068, doi: 2007, Integration of ground penetrating radar with con- and Kurz, T.H., 2006, Integration of terrestrial laser 10.1306/02260403062. ventional stratigraphic and LIDAR data to investigate scanning, digital photogrammetry and geostatistical Froster, C.B., Snelgrove, S.H., and Koebbe, J.V., 2004, Mod- the three-dimensional geometry of a tidal-inlet reser- methods for high-resolution modelling of geological elling permeability structure and simulating fl uid fl ow voir analog, Upper Ferron Sandstone, Utah, Abstract outcrops: Dresden, Germany, ISPRS Commission V in a reservoir analog: Ferron Sandstone, Ivie Creek Volume, Understanding earth systems pursuing the Symposium September 25–27, Volume 36(B5), Inter- Area, East-Central Utah, in Chidsey, T.C., Jr., Adams, checkered fl ag: 2007 AAPG Annual Convention and national Archives of the Photogrammetry, Remote R.D., and Morris, T.H., eds., Regional to wellbore ana- Exhibition (April 1–4, 2007) Long Beach, California: Sensing and Spatial Information Sciences: Commission log for fl uvial-deltaic reservoir modeling: Ferron Sand- Los Angeles, USA, American Association of Petro- V, WG III, http://rcswww.urz.tu-dresden.de/~isprs/pro- stone of Utah, Volume 50: Tulsa, Oklahoma, American leum Geologists. ceedings/paper/1228_Dresden06.pdf. Association of Petroleum Geologists, Studies in Geol- Leren, B., 2005, Sandbody architecture of basin margin Cartwright, J.A., Mansfi eld, C., and Trudgill, B.D., 1996, ogy, p. 359–382. shallow-marine delta systems: The Eocene Roda Sand- The growth of normal faults by segment linkage: Mod- Færseth, R.B., 1996, Interaction of Permo-Triassic and stone, Tremp-Graus Basin, Spain and the Cemalettin ern developments in structural interpretation, validation Jurassic extensional fault-blocks during the develop- Formation, Boyabat Basin, Turkey, Abstracts volume: and modelling, Volume 99: London, United Kingdom, ment of the northern North Sea: Journal of the Geologi- Calgary, Canada, 2005 American Association of Petro- Geological Society of London, p. 163–177. cal Society of London, v. 153, p. 931–944. leum Geologists Annual Convention (June 19–22) Childs, C., Watterson, J., and Walsh, J.J., 1995, Fault overlap Gani, M.R., and Bhattacharya, J.P., 2005, Lithostratigraphy Technical Program. zones within developing normal fault systems: Journal versus chronostratigraphy in facies correlations of Leren, B.L.S., 2007, Complexity in low-accommodation of the Geological Society of London. Quaternary deltas: Application of bedding correlation, fan-deltaic systems deposited in tectonically active Corfi eld, S., Sharp, I., Haeger, K.-O., Dreyer, T., and Under- River deltas, concepts, models, and examples, Volume basins. Examples from the Eocene Roda Sandstone, hill, J., 2001, An integrated study of the Garn and Melke 83: Tulsa, Oklahoma, USA, SEPM (Society for Sedi- Spain, the Eocene Cemalettin Formation, Turkey and formations (Middle to Upper Jurassic) of the Smør- mentary Geology), p. 31–48. the Jurassic Oseberg Formation, North Sea [Ph.D. the- bukk area, Halten Terrace, mid Norway, in Martinsen, Gilbert, G.K., 1885, The topographic features of lake sis]: Bergen, University of Bergen, Norway. O.J., and Dreyer, T., eds., Sedimentary environments shores, Annual Report: U.S. Geological Survey, v. 5, Lopez-Blanco, M., 1996, Estratigrafi a secuencial de siste- Offshore Norway—Palaeozoic to Recent, Volume 10: p. 75–123. mas deltaicos en Cuencas de Antepais: ejemplos de Amsterdam, Norwegian Petroleum Society Special Hampson, G.J., 2000, Discontinuity surfaces, clinoforms, Sant Lorenc del Mont, Montserrat y Roda (Paleogeno, Publication, Elsevier Science B.V., p. 199–210. and facies architecture in a wave-dominated, shoreface- Cuenca de Antepais Surpirenaica) [Ph.D. thesis], Uni- Deveugle, P.E.K., Jackson, M.D., Hampson, G.J., Calvert, shelf parasequence: Journal of Sedimentary Research, versitat de Barcelona, Spain. C.S., Stewart, J., Sprague, A.R., and Farrell, M.E., 2007, v. 70, p. 325–340. Lopez-Blanco, M., Marzo, M., and Munoz, J.A., 2003, Low- High-resolution, three-dimensional outcrop modeling of Hampson, G.J., and Storms, J.E.A., 2003, Geomorpho- amplitude, synsedimentary folding of a deltaic com- a fl uvial-dominated deltaic reservoir analogue, Abstract logical and sequence stratigraphic variability in plex; Roda Sandstone (Lower Eocene), South-Pyre- volume: From outcrop to asset: Recent advances in digi- wave-dominated, shoreface-shelf parasequences: Sedi- nean Foreland Basin: Basin Research, v. 15, p. 73–95, tal outcrop data collection and modelling techniques, mentology, v. 50, p. 667–701, doi: 10.1046/j.1365- doi: 10.1046/j.1365-2117.2003.00193.x. 10-11 January 2007, University of Manchester, Atmo- 3091.2003.00570.x. Løseth, T.M., Thurmond, J., Søgaard, K., Rivenaes, J.C., and spheric & Environmental Science: Manchester, United Helland-Hansen, W., Ashton, M., Lomo, L., and Steel, R., Martinsen, O., 2003, Visualization and utilization of 3D Kingdom, The Geological Society. 1992, Advance and retreat of the Brent Delta: Recent outcrop data: Barcelona, Spain, American Association Dreyer, T., Falt, L., Høy, T., Knarud, R., Steel, R., and Cuevas, contributions to the depositional model, Geology of the of Petroleum Geologists International Convention. J.-L., 1993, Sedimentary architecture of fi eld analogues Brent Group, Volume 61: London, United Kingdom, MacDonald, A.C., and Halland, E.K., 1993, Sedimentology for reservoir information (SAFARI): A case study of the Geological Society of London, p. 109–127. and shale modeling of a sandstone-rich fl uvial reser- fl uvial Escanilla Formation, Spanish Pyrenees, in Flint, Holden, L., Hauge, R., Skare, O., and Skorstad, A., 1998, voir; upper Statfjord Formation, Statfjord Field, north- S.S., and Bryant, I.D., eds., The Geological modelling of Modeling of fl uvial reservoirs with object modern North Sea: American Association of Petroleum hydrocarbon reservoirs and outcrop analogues, Volume els: Mathematical Geology, v. 30, p. 473–496, doi: Geologists Bulletin, v. 77, p. 1016–1040. 15: London, International Association of Sedimentologists 10.1023/A:1021769526425. Martinsen, O., Thurmond, J., Hunt, D., Løseth, T.M., and Special Publication, p. 57–79. Howell, J.A., Buckley, S.J., Carlsson, C., Enge, H.D., Kurz, Gillespie, P., 2007, Outcrops—From fully qualitative Enge, H.D., Buckley, S.J., Howell, J.A., Vassel, Å., Leren, T.H., and Rotevatn, A., 2007, Virtual outcrop geology— to fully quantitative: A review of recent advances in B.L.S., and Martinius, A.W., 2007, Laser scanning as Data collection, processing and application, Abstract technology, methodology, and Application, Abstract a tool in sedimentology—Case studies from shallow volume: From outcrop to asset: Recent advances in digi- volume: From outcrop to asset: Recent advances in deltaic systems from the Ferron Sandstone, USA and tal outcrop data collection and modelling techniques, digital outcrop data collection and modelling tech- the Roda Sandstone, Spain, Abstract volume: From 10–11 January 2007, University of Manchester, Atmo- niques, 10-11 January 2007, University of Manchester, outcrop to asset: Recent advances in digital outcrop spheric & Environmental Science: Manchester, United Atmospheric & Environmental Science: Manchester, data collection and modelling techniques, 10–11 Janu- Kingdom, The Geological Society. United Kingdom, The Geological Society. ary 2007, University of Manchester, Atmospheric & Howell, J.A., Vassel, Å., Aune, T., Buckley, S.J., and Enge, McCaffrey, K.J.W., Jones, R.R., Holdsworth, R.E., Wilson, Environmental Science: Manchester, United Kingdom, H.H., 2006, From outcrop to fl ow simulation: The R.W., Clegg, P., Imber, J., Holliman, N., and Trinks, The Geological Society. effect of shallow marine clinoforms on simulated pro- I., 2005, Unlocking the spatial dimension: Digital Enge, H.D., Leren, B.L.S., Howell, J.A., Buckley, S.J., and duction from outcrop analogs in Utah: Houston, Texas, technologies and the future of geoscience fi eldwork: Martinius, A.W., 2006, Laser scanning as a tool in sedi- USA, Abstracts Volume, American Association of Journal of the Geological Society of London, v. 162, mentology—A case study of the 3D sandbody architec- Petroleum Geologists 2006 Annual Convention, April p. 927–938. ture of the shallow marine Roda Sandstone, N. Spain, 9–12, p. 49–50. Molenaar, N., 1990, Calcite cementation in shallow marine Abstract volume: Unlocking 3D earth systems— Har- Jackson, M.D., Yoshida, S., Muggeridge, A.H., and Johnson, Eocene sandstones and constraints of early diagenesis: nessing new digital technologies to revolutionize multi- H.D., 2005, Three-dimensional reservoir characteriza- Journal of the Geological Society of London, v. 147, scale geologic models: Durham, United Kingdom, Geo- tion and fl ow simulation of heterolithic tidal sandstones: p. 759–768.

Geosphere, December 2007 489

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021 Enge et al.

Molenaar, N., and Martinius, A.W., 1990, Origin of nodules Pedersen, C., 2005, Reservoir modelling of small-scale het- Thurmond, J., 2006, Five years of 3D outcrop analysis: The in mixed siliciclastic-carbonate sandstones, the lower erogeneities in a the sand-rich channelised turbidite past, present, and the future, Abstract volume: Unlock- Eocene Roda Sandstone Member, southern Pyrenees, system; Ross Formation, SW Ireland [M.S. thesis]: ing 3D earth systems—Harnessing new digital tech- Spain: Sedimentary Geology, v. 66, p. 277–293, doi: Bergen, Norway, University of Bergen. nologies to revolutionize multi-scale geologic models: 10.1016/0037-0738(90)90064-Z. Pedersen, C., Howell, J.A., Lien, T., Vipond, E., Thurmond, Durham, United Kingdom, Geological Society of Molenaar, N., and Martinius, A.W.,1996, Fossiliferous J., Løseth, T.M., and Martinsen, O., 2007, Reservoir America Penrose Conference, 17–21 September 2006, intervals and sequence boundaries in shallow marine, modelling of small-scale heterogeneities in a channel- Durham University. fan-deltaic deposits (early Eocene, southern Pyrenees, ized turbidite systems: Digital outcrop studies of the Trudgill, B., and Cartwright, J., 1994, Relay-ramp forms Spain): Palaeogeography, Palaeoclimatology, Pal- Ross Formation, Western Ireland, Abstract Volume, and normal-fault linkages, Canyonlands National aeoecology, v. 121, p. 147–168, doi: 10.1016/0031- Understanding earth systems pursuing the checkered Park, Utah: Geological Society of America Bul- 0182(95)00073-9. fl ag: 2007 American Association of Petroleum Geolo- letin, v. 106, p. 1143–1157, doi: 10.1130/0016- Monsen, E., 2006, Automated extraction of bed boundar- gists Annual Convention and Exhibition (April 1–4, 7606(1994)106<1143:RRFANF>2.3.CO;2. ies and fractures, Abstract volume: Abstract volume: 2007) Long Beach, California: Los Angeles, USA, Vipond, E., 2005, Modelling small-scale heterogeneities in Unlocking 3D earth systems— Harnessing new digi- American Association of Petroleum Geologists. a channelised turbidite system: Digital outcrop study tal technologies to revolutionize multi-scale geologic Pickup, G.E., and Hern, C.Y., 2002, The development of from the Ainsa I Turbidite System [M.S. thesis]: Ber- models: Durham, United Kingdom, Geological Society appropriate upscaling procedures: A symposium on gen, Norway, University of Bergen. of America Penrose meeting, 17–21 September 2006, upscaling reservoir properties, upscaling downunder, Weber, K.J., 1986, How heterogeneity affect oil recovery, in Durham University. Volume 46, 2–3: Dordrecht, Netherlands, Kluwer Aca- Lake, L.W., and Carroll, H.B., Jr., eds., Reservoir char- Moore, J.M., and Schultz, R.A., 1999, Processes of fault- demic Publishers, p. 119–138. acterization: Orlando, Academic Press, p. 487–544. ing in jointed rocks of Canyonlands National Park, Pringle, J., Gardiner, A., and Westerman, R., 2004a, Vir- Wehr, A., and Lohr, U., 1999, Airborne laser scanning—An Utah: Geological Society of America Bulletin, v. 111, tual geological outcrops; fi eldwork and analysis made introduction and overview: ISPRS Journal of Photo- p. 808–822, doi: 10.1130/0016-7606(1999)111<0808: less exhaustive?: Geology Today, v. 20, p. 67–72, doi: grammetry & Remote Sensing, v. 54, p. 68–82, doi: POFIJR>2.3.CO;2. 10.1111/j.1365-2451.2004.00450.x. 10.1016/S0924-2716(99)00011-8. Nagihara, S., Mulligan, K.R., and Xiong, W., 2004, Use of Pringle, J.K., Howell, J.A., Hodgetts, D., Westerman, A.R., White, C.D., Willis, B.J., Dutton, S.P., Bhattacharya, J.P., a three-dimensional laser scanner to digitally capture and Hodgson, D.M., 2006, Virtual outcrop models of and Narayanan, K., 2004, Sedimentology, statistics, the topography of sand dunes in high spatial resolu- petroleum reservoir analogues: A review of the current and fl ow behavior for a tide-infl uenced deltaic sand- tion: Earth Surface Processes and Landforms, v. 29, state-of-the-art: First Break, v. 24, p. 33–42. stone, Frontier Formation, Wyoming, United States, in p. 391–398, doi: 10.1002/esp.1026. Pringle, J.K., Westerman, A.R., Clark, J.D., Drinkwater, Grammer, G.M., Harris, P.M., and Eberli, G.P., eds., Nemec, W., and Steel, R.J., 1988, Fan deltas: Sedimentology N.J., and Gardiner, A.R., 2004b, 3D high-resolution Integration of outcrop and modern analogs in reservoir and tectonic settings: Glasgow, Blackie, 444 p. digital models of outcrop analogue study sites to con- modeling: American Association of Petroleum Geolo- Oftedal, B.T., Ahokas, J.M., Martinius, A.W., Suter, J., Næss, strain reservoir model uncertainty: An example from gists Memoir 80, p. 129–152. A., and Olsen, T., 2007, Building multiscale outctop Alport Castles, Derbyshire, UK: Petroleum Geosci- Willis, B.J., and White, C.D., 2000, Quantitative outcrop data libraries (The Lower Jurassic Neill Klinter Group, ence, v. 10, p. 343–352. data for fl ow simulation: Journal of Sedimentary East Greenland) for improved calibration of earth mod- Redfern, J., Hodgetts, D., and Fabuel-Perez, I., 2007, Digital Research, v. 70, p. 788–802. els of shallow marine heterolithic reservoirs offshore analysis brings renaissance for petroleum geology outcrop Xu, X., Aiken, C.L., Bhattacharya, J.P., Corbeanu, R.M., mid-Norway, Abstract volume: From outcrop to asset: studies in North Africa: First Break, v. 25, p. 81–87. Nielsen, K.C., McMechan, G.A., and Abdelsalam, Recent advances in digital outcrop data collection and Reynolds, A.D., 1999, Dimensions of paralic sandstone M.G., 2000, Creating virtual 3-D outcrop: The Lead- modelling techniques, 10-11 January 2007, University bodies: American Association of Petroleum Geologists ing Edge, Society of Exploration Geophysicists, v. 19, of Manchester, Atmospheric & Environmental Science: Bulletin, v. 83, p. 211–229, doi: 10.1306/00AA9A48- p. 197–202. Manchester, United Kingdom, The Geological Society. 1730-11D7-8645000102C1865D. Xu, X., Bhattacharya, J.P., Davies, R.K., and Aiken, C.L., Olariu, C., Bhattacharya, J.P., Xu, X., Aiken, C.L., Zeng, Riegl, 2006, Terrestrial Scanner Overview http://www.riegl. 2001, Digital geological mapping of the Ferron Sand- X., and McMechan, G.A., 2005, Integrated study of com/terrestrial_scanners/terrestrial_scanner_over- stone, Muddy Creek, Utah, with GPS and refl ectorless ancient delta-front deposits, using outcrop ground- view/terr_scanner_menu_all.htm. laser rangefi nder: GPS Solutions, v. 5, p. 15–23, doi: penetrating radar, and three-dimensional photorealis- Rotevatn, A., Fossen, H., Hesthammer, J., Aas, T.E., and 10.1007/PL00012872. tic data: Cretaceous Panther Tongue Sandstone, Utah, Howell, J.A., 2007, Are relay ramps conduits for fl uid Yang, C.S., and Nio, S.D., 1985, The estimation of palaeo- USA, in Giosan, L., and Bhattacharya, J.P., eds., River fl ow? Structural analysis of a relay ramp in Arches hydrodynamic processes from subtidal deposits using Deltas—Concepts, models, and examples, Volume 83: National Park, Utah, in Lonergan, L., Jolly, R.J.H., time series analysis methods: Sedimentology, v. 32, Tulsa, Oklahoma, SEPM (Society for Sedimentary Rawnsley, K., and Sanderson, D.J., eds., Fractured res- p. 41–57, doi: 10.1111/j.1365-3091.1985.tb00491.x. Geology) Special Publication, p. 155–177. ervoirs, Volume 270: London, The Geological Society Peacock, D.C.P., and Sanderson, D.J., 1991, Displacements, of London, Special Publications, p. 55–71. segment linkage and relay ramps in normal fault zones: Roxar, 2006, User guide interactive reservoir analysis package Journal of Structural Geology, v. 13, p. 721–733, doi: (IRAP) RMS 8.0: Roxar Software Solutions, 2320 p. 10.1016/0191-8141(91)90033-F. Stafl eu, J., Schlager, W., Everts, A.J.W., Kenter, J.A.M., Peacock, D.C.P., and Sanderson, D.J., 1994, Geometry and Blommers, G., and van Voorden, A., 1996, Outcrop development of relay ramps in normal fault systems: topography as a proxy of acoustic impedance in syn- MANUSCRIPT RECEIVED 26 FEBRUARY 2007 American Association of Petroleum Geologists Bulle- thetic seismograms: Geophysics, v. 61, p. 1779–1788, REVISED MANUSCRIPT RECEIVED 6 JUNE 2007 tin, v. 78, p. 147–165. doi: 10.1190/1.1444094. MANUSCRIPT ACCEPTED 18 JULY 2007

490 Geosphere, December 2007

Downloaded from http://pubs.geoscienceworld.org/gsa/geosphere/article-pdf/3/6/469/854380/i1553-040X-3-6-469.pdf by guest on 26 September 2021