Datapages, Inc.Print this page

  Click to view article in PDF format.

PSFrom collection to utilization: Outcrop analog data in a 3D world*

By

John B. Thurmond1 

 

Search and Discovery Article #40126 (2004)

 

*Adapted from poster presentation at AAPG Annual Meeting, Dallas, Texas, April 18-21, 2004.

 1University of Texas at Dallas ([email protected])

 

Abstract 

The collection of three-dimensional data from outcrops is playing an increasingly important role in reservoir characterization studies. There are a variety of techniques that can be used to acquire three-dimensional data from outcrops, and each should be applied individually or in concert to collect data in specific circumstances. The current suite of emerging methods typically used in outcrop-scale measurement includes traditional surveying, direct GPS measurement, laser scanning (LIDAR), photogrammetry, and photorealistic mapping (texture draped geometry). Depending on the morphology, setting, and particular data needs of a specific outcrop, different methods can be used to acquire data. Case studies of individual outcrops will be shown to illustrate the problems and benefits of several of these methods. 

Once three-dimensional data is collected, utilizing the data can present its own set of challenges. Each collection method produces a different type of data, each of which requires a variety of processing and interpretation methods to utilize effectively. In most cases, there is also the need to integrate data from a variety of sources into a single interpretable data set. Again, case studies provide specific illustrations of effective methods that have been used in various projects to produce reservoir models from a variety of environments, including deep-water channel systems, heavily faulted fluvial environments, and carbonate build-ups.

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

uAbstract

uFigure captions

uMethods

  uGPS mapping

  uLaser Rangefinder mapping

  uPhotorealistic outcrop capture

uCarbonates--

  uGuadalupe Mountains

uSiliciclastics—

  uSouth-Pyrenean basin

uFault mapping—

  uCentral Utah

uAcknowledgments

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure Captions

Figure 1. Geologic data can be directly acquired in 3D through the use of RTK GPS receivers, providing a spatial accuracy of < 2cm. This is one of the most accurate methods, but also potentially one of the slowest. Rough terrain can provide a significant obstacle in this type of data collection

Figure 2. Laser scanners (one model shown here, being used by the author), can be used in two ways. First, a “single shot” can be taken for each point of 3D data to be collected. This method can be relatively slow and is slightly less accurate than GPS mapping, but it is not limited by difficult terrain. Typically, a combination of single-shot laser mapping and GPS mapping is used.

Figure 3. Laser scanners can also be used to collect detailed topographic data from the outcrop. This data, when combined with digital photography, can be used to produce photorealistic outcrop models, like the one illustrated in this figure. This method allows you to “take the outcrop home” and apply a variety of different interpretations.

Figure 4. Individual stratigraphic surfaces are walked with an RTK GPS system, providing a nominal mapping accuracy of approximately 50cm. These mapped surfaces, when combined with data from sample transects, can be used to produce geologic models of the 3D distribution of facies.

Figure 5. Geologic model of Last Chance Canyon mud-mounds, as mapped and interpreted in 3D. Blues are peloidal mudstones, greens are sparse fossil wackestones, and reds are lenses of fossil grainstone. Surface building and geologic modeling was performed using Roxar RMS.

Figure 6. Geologic model with overlay of USGS 10m DEM (transparent green surface) to show intersection of mud-mounds with topography.

Figure 7. Photorealistic model of main exposure of Ainsa turbidites shown with a DEM and high-resolution air photo. Because they are positioned with GPS, individual photorealistic segments can be integrated into a single framework.

Figure 8. Tore Løseth digitizing surfaces on the Ainsa model in the CAVE at Norsk Hydro. Photorealistic outcrops allow interactive interpretation sessions with a group of geologists. Interpretations can easily be changed at any time.

Figure 9. Digitized points can be combined with other data, including measured sections and wells to create three-dimensional surfaces that honor all available data. These surfaces can then be used to build a reservoir model, based on the properties observed between each bounding surface. In this figure, the photorealistic outcrop model can be seen in detail; digitized points are shown as yellow squares, and a channel bottom surface is shown in green. The channel surface curves off to the right in the distance.

Figure 10. The surfaces can be combined to build zone models of the outcrop, grouping elements with similar facies. These zone models can then be populated with physical properties, derived either from the outcrops themselves, or more realistic production values from subsurface analogs. In this figure, the zone model can be seen in relationship to the aerial photograph and DEM, first with a transparent ground surface (A), and then with an opaque surface (B).

Click to view sequence of Figure 10A and 10B.

Figure 11. The final result is a property model that can be used directly in a flow simulation. One great advantage of constructing a model using this method is that the outcrops can be seen juxtaposed with the data. This illustrates very well the problems involved with upscaling and the typical loss of geologic detail. Here, the lower channel surface can be seen on the photorealistic outcrop model to the right with a slice through the 3D porosity model.

Figure 12. Individual stratigraphic surfaces (left) and faults (right) both with and without USGS DEM data draped with high resolution oblique air photos and a geologic map overlay. Mapped GPS data is shown in the lower images as individual points. The stratigraphic surface is offset by a fault that tips out towards the left. The interpreted fault surface shows small-scale corrugations and relay zones that are captured during the mapping process. GPS mapping accurately captures these features and often forces mapped interpretations to be altered.

Figure 13. Several fault surfaces shown in a single visualization. High resolution aerial photographs overlain with a geologic map provide a link between the GPS mapping and previous data collection efforts. A previously collected database of fault properties was also located using GPS, providing a 3D browsable database of measurements. This can be used to propagate fault properties throughout the model and help to establish a quantitative link between fault geometry and fault properties.

Return to top.

 

Methods 

GPS Mapping (Figures 1 and 4) 

The most straightforward method for mapping geologic surfaces in 3D is to simply walk them out with a high-precision GPS system. This type of mapping is most appropriate on accessible outcrops and in places where contacts or facies boundaries are subtle. 

Normally, Real-Time Kinematic GPS systems are used, which provide an accuracy of approximately 2cm. Due to the practicalities of efficiently walking on a geologic surface, the normal accuracy is about 50 cm.

 

Advantages:                  High accuracy (50 cm or less).

                                    Accurate data distribution (no surface, no data) eliminates “interpolation” problems.

 

Disadvantages:  Slow (5-10 km per day).

                                    Re-interpretation can require re-mapping.

 

Laser Rangefinder Mapping (Figures 2 and 3

Often, it is not physically possible to “walk” on a stratigraphic surface, so other techniques are required. Reflectorless Laser Rangefinders, coupled with high-precision GPS receivers, provide the opportunity to capture data from such locations. These systems integrate an EDM for distance measurement with a digital compass and inclinometer, so 3D position can be measured remotely. However, this technique requires surfaces that are visible from a remote location.

 

Advantages:                  Fast.

 

Disadvantages:  Lower accuracy (varies with distance).

                                    Encourages interpolation/extrapolation.

                                    Re-interpretation can require re-mapping.

 

Photorealistic Outcrop Capture (Figures 7 and 8

Using a scanning laser system, coupled with a high-accuracy GPS system, it is possible to scan the topography of an outcrop with a high degree of accuracy. Digital photographs can then be accurately mapped to the topography, which provides an accurate, three-dimensional model of the outcrop. This model can then be interpreted in 2D (on the photographs) or in 3D (on the model), and re-interpreted as often as necessary. Some laser systems provide the capability of scanning color simultaneously with distance, but the data sets are incredibly difficult to work with, even with the most advanced 3D workstations available. However, 3D geometry combined with texture data from photographs renders extremely quickly, even on laptops (ask for a demo!).

 

Advantages:                  Fast.

                                    High accuracy (5-10cm pixel error (!)).

                                    Provides images with data, so interpretations are believable.

                                    Data sets can be re-interpreted as paradigms change.

 

Disadvantages:  Expensive equipment required.

                                    Processing can be intensive.

 

Carbonates 

Location:    Last Chance Canyon, Guadalupe Mountains, New Mexico (Figures 5 and 6

Problem:    Excellent 3D exposures of a mixed carbonate and siliciclastic system. Antecedent topography is an important control on subsequent facies deposition. Mapping of paleogeomorphological surfaces and subsequent facies in 3D provides the opportunity to build models which predict facies deposition as a function of topography for specific systems. 

Techniques:    Currently, only GPS mapping has been applied. Photorealistic mapping scheduled for Q2 of 2004. Sample transects integrated into a 3D VRML data base containing outcrop photos, photomicrographs and sample descriptions (the live visualization has been demonstrated at oral and poster presentations). 

Results:    Current:  3D geologic model of (hydrodynamic?) mud-mounds. Model provides evidence for re-interpretation of processes controlling mud-mound deposition.

Future:  Interpreted photorealistic model of Last Chance Canyon. This will provide a framework for building a 3D geologic model of the canyon, which will be an excellent research and teaching data set.

 

Siliciclastics 

Location:    Eocene Ainsa II deepwater channel and lobe complex of the South-Pyrenean Foreland Basin, Spain (Figures 7, 8, 9, 10, and 11

With:    Ole J. Martinsen and Tore M. Løseth, Jan Rivenaes, and Kristian Soegaard, Norsk Hydro Research Centre 

Problem:    These deep-water siliciclastics are an excellent analog for active production fields in offshore Angola. Accurate 3D data captured from outcrop facies relationships is used to build a reservoir model, which can help predict heterogeneity in the subsurface and will also make an excellent teaching data set.  

Techniques:    Numerous integrated photorealistic models were used to collect accurate 3D representations of the outcrop. Surfaces were interpreted in an immersive visualization environment (a CAVE) and, combined with spatially accurate 3D structural models, produced by the University of Barcelona. 

Results:    A reservoir model was produced from the 3D data acquired from the outcrops.

 

Fault Mapping 

Location:    Various faults in central Utah (Figures 12 and 13

With:    Rod Myers, Peter Vrolijk, and Tom Hauge, ExxonMobil Upstream Research 

Problem:    Fault systems have complicated 3D geometries, and fault properties vary as a function of geometry. Mapping of fault geometries and properties in 3D allows these relationships to be determined quantitatively, allowing them to be applied algorithmically to fault geometries observed in the subsurface. 

Techniques:    Many of these faults are difficult to see from a distance, and outcrops are generally accessible, so direct GPS mapping was used to collect data. Individual faults and stratigraphic surfaces were walked out, and the intersection of topography and the surfaces was sufficient to produce a 3D model. 

Results:    A faulted framework model was produced in gOcad using the 3D data collected in the field. This model will be used to visualize fault properties as a function of geometry.        

 

Acknowledgments 

The author would like to thank the following individuals and companies for financial, conceptual, and/or fieldwork support for these projects: 

Norsk Hydro Research Centre

ExxonMobil Upstream Research

NSF Graduate Fellowship

Carlos Aiken

Xueming Xu

Janok Bhattacharya

 

The author thanks Roxar for the generous donation of their software to UTD.

 Return to top.