Structural Interpretation of Seismic Geologic Reality, Perspective, and 3-D Thinking
Published:January 01, 2016
- PDF LinkChapter PDF
Peter Boult, Brett Freeman, Graham Yielding, 2016. "Structural Interpretation of Seismic Geologic Reality, Perspective, and 3-D Thinking", 3-D Structural Interpretation: Earth, Mind, and Machine, Bob Krantz, Carol Ormand, Brett Freeman
Download citation file:
Our notion of reality in seismic interpretation and structural geology usually follows a series of careful observations and ideas that eventually crystallize into a best-case model. In most other branches of science the strength or reality of such models, or hypotheses, is increased by the number of robust tests that either refine or fail to disprove the original idea. However, geological models in the hydrocarbon exploration and production sector differ because the starting point for testing a hypothesis is usually an interpretation of seismic data or other remote measurements, rather than the direct observation of an effect.
The scientific method of prediction tested by observation is a key part of mapping three-dimensional (3-D) structures in the field and geological training. An analogous, rule-based approach also applies to the accurate creation of 3-D subsurface structural models. A defensible structural model must embody more than fault and horizon surfaces. It must also honor the rules of structural geology. Some simple rules are outlined in this chapter. These can be applied iteratively throughout the life of the seismic interpretation. Failure to honor structural rules leads to poor interpretations that may be compounded by a lack of appreciation of the importance of 3-D perspective. In this chapter, we also briefly explore the historical use and understanding of perspective.
Those in the exploration and production industry need to think carefully about how to leverage the 3-D interpretation and modeling process. Most importantly, since it is managers who control the exploration and production workflow, they above all need to be informed about the advantages of using a structurally qualified 3-D model in future projects.
The quest to present geological reality to stakeholders in the hydrocarbon exploration and production industry has its roots in the late renaissance of the 17th century and philosophers such as Rene Descartes (1596–1650), who grappled with the notion of reality itself and Francis Bacon (1561–1626), who was arguably the creator of the modern scientific method and empiricism. The list of notable scientists for geology should also include Nicolas Steno (1638–1686), who introduced the fundamental geological laws of superposition, the principles of original horizontality and crosscutting relationships. These people and their rules are considered the originators of modern geology and 3-D geological thinking.
Descartes’ notion of the self being separate from God and part of the natural world allowed those who followed him to think about how the natural phenomena they observed came about other than by the hand of God. Bacon embraced such thinking and applied deductive reasoning to science. The Baconian method of science is closely associated with empiricism, which is the scientific method most applicable to geology, in particular structural geology and seismic interpretation. This is because, unlike other sciences, whereby repeatability of experiments under laboratory conditions is a fundamental test of hypotheses, the large volume and time scale of geologic phenomena generally precludes the conventional scientific approach. In the discipline of structural geology, hypotheses by necessity need to be based on deductive reasoning, which in turn is based on prior knowledge and empirical correlations. However, when using empiricism as a scientific method, the fact that our knowledge is never certain ought to be taken into account. Hypotheses should be continually questioned to improve them, thus resultant theories are best expressed probabilistically.
The Royal Society of London, founded in 1660, was a place where natural philosophers such as Robert Hooke (1635–1703) and John Ray (1627–1705), both significant early contributors to geologic knowledge, came together to discuss science. Here the problems of geology were discussed alongside other sciences, but diagrams were seldom used to communicate ideas, especially those representing 3-D phenomena. To some extent this was perhaps due to printing limitations; however, maps and cross-sections were introduced very early, with one of the more famous being the Geological Map of Britain, including cross-sections, produced by William Smith in 1815 (Winchester, 2001). Nevertheless, it was another 100 years or more before geologists tried to represent their understanding of geologic reality in publications using 3-D diagrams.
In the exploration and production industry the availability of effective 3-D interpretation software has evolved steadily, building initially on legacy concepts of digital two-dimensional (2-D) interpretation. It is now possible to perform interpretation and modeling entirely within a 3-D visualization environment. It is also possible to interrogate, analyze, and improve interpretation incrementally. Later in this chapter we outline a set of structural geological rules that help to realize a defensible structural framework model (Krantz and Neely, 2016); but first we give a brief history of the use of 3-D thinking in the geological literature. We then discuss how structural geology differs from mainstream science, and where it fits into the exploration and production process.
Representing 3-D Ideas in 2-D and The Perspective Problem
The representation of 3-D phenomena on or within 2-D media such as paper or a computer screen requires an understanding of perspective. Long ago, Plato identified the problem and its solution through mathematical computation (Plato 360 BC). While in conversation with Socrates, he stated:
“The body which is large when seen near, appears small when seen at a distance? [-] Thus every sort of confusion is revealed within us; and this is that weakness of the human mind on which the art of conjuring and of deceiving by light and shadow and other ingenious devices imposes, having an effect upon us like magic [-] give way before calculation.”
Nevertheless, there are still seismic interpretation software systems in common usage, which have less than optimal implementations of this concept.
There are essentially two methods of displaying a scene in 3-D graphics. The first may be referred to as the isometric projection or, alternatively, the orthographic projection. All real-world coordinates are scaled equally into the coordinate frame of the screen. There is no relative distortion of the distances between coordinates. It has the advantage that angles, lengths, and areas are conserved with increasing distance from the observer. Its natural use in geoscience is principally for maps and cross- sections. However, isometric displays also have their place in seismic interpretation. For example, they can be useful when needing to compare data at different stages of processing, when panning through 3-D seismic cubes looking for fault trends, or when comparing positional relationships between near and distant objects. Although, paradoxically, there is no distortion, such 3-D views can give a sense of discomfort and disorientation to the observer. For this reason, an isometric projection is a poor choice when the goal of the exercise is to map the geological relationships that lead to a 3-D framework model. It is an acutely poor choice when the structures are geologically complex. If isometric is the only projection available, in our experience, it leads users to revert to 2-D representations of their data.
By contrast in the second method, perspective projection, scaling depends linearly on the distance to a particular coordinate. The apparent distance between far points is less than the same real distance between near points. It does not conserve angles, lengths, or areas, but it is a far more effective and natural visual environment for structural interpretation. It is a conundrum that as a subject perspective is more likely to be encountered in an art class or computer science class than on a geoscience course. However, its implementation underpins effective use of the digital environment for interpretation of complex geologic structures.
Here we define a concept of true perspective that states that spatial, angular, and length relationships, as viewed in a diagram or on a computer screen, should be indistinguishable from how a stationary observer would perceive the real object. The evolution of the concept in AAPG publications is outlined below.
A History of Perspective Use in the AAPG Bulletin
An investigation of the AAPG Bulletin archives allows one to appreciate how the use of perspective in published diagrams is only a recent phenomenon, but of course this does not reflect the history of 3-D thinking in geology itself. A search of papers between 1917 and 2013 (this survey was current in 2014), which uses the logical search criteria “block diagram” or “isometric” in the full text and “structure” in the abstract, or “3-D seismic” in the full text, returns 113 papers. Of these, 62 papers contain 3-D cube diagrams with sufficient information to allow for historical analysis of the type of perspective used to display them (Figures 1 and 2). The criteria used to categorize are as follows: the drawing has (1) no perspective and is purely isometric; (2) only one vanishing point; (3) two vanishing points; or (4) three vanishing points (Figure 3). Visual realism increases with the number of vanishing points but even at three we have only an approximation of true perspective, as will be explained later.
Using the above search criteria, the first 3-D block diagram in the AAPG Bulletin appears to have been published by Link in 1929. This was an isometric representation. However, it was not until 1942 (Butler, 1942) that the first block diagram was published in perspective. This had only one vanishing point (one-point perspective). It was another 24 years before Halbouty (1966) published a block diagram using a two-point perspective (two vanishing points). A further 40 years ensued before a diagram in three-point, or close to true perspective was published (Stewart and Davies, 2006), and in the last six years only 11, or 42% (n = 26), have been published this way. Furthermore, of the 11 diagrams that have three-point perspective or better, the vast majority are images taken directly from software packages and are not hand-drafted diagrams. To a large degree, this reflects the fact that perspective drawings are labour-intensive. They are graphical constructions that are artificially conditioned by the locations and orientations of sets of parallel lines. However, perhaps it also reflects a community view that isometric and perspective are interchangeable. While this may be acceptable for stationary images, that is certainly not the case in an animated 3-D environment. Here, the most effective representation for structural interpretation is true perspective and not isometric.
Perspective and Animation
The representation of geologic scenes using three-point perspective on paper or true perspective on a computer screen is one of the most important aspects of a display that captures visual realism. Equally important is the animation of the display, which needs to be carefully centered so that rotation, pan, and zoom do not produce sudden and unrealistic distortions. The rotation center is usually chosen to be either the center of gravity of the visible objects or the location of the cursor if it touches an object. Ideally the scene should behave as if it were held in the user’s hand. Mathematically the projection is simple, depending on nothing more sophisticated than similar triangles. The perspective transformation is calculated in terms of the relative distance from (a) the observer to the screen and (b) the observer to the object or coordinate (Treibergs, 2014). It is usually set to provide the most natural experience, and if the observation point is too close to the viewing plane, the projection mimics a fish-eye lens. On the other hand, by moving the observation point away from the viewing plane, the projection appears progressively more isometric.
Structural Models, The Exploration and Production Process, and Comparison with Laboratory-Based Sciences
The compaction of sediments to form rock and its subsequent deformation on the scale required to form commercial hydrocarbon reservoirs is impossible to replicate under laboratory conditions. This challenge, combined with the vagaries of poor seismic data gathered in two-way-time and the absence of robust velocity information, leads to a considerable degree of uncertainty in the information held in this primary data source. Velocity and imaging issues increase with structural complexity. The standard practice requires modeling for a best-case representation of structure in the subsurface. Such modeling will, in addition to the seismic data, incorporate parameters derived indirectly from wireline and perhaps geomechanical testing. The cost of gathering such data prohibits the normal refinement of models or hypotheses, which are commonly implemented in other sciences via repeated testing of the original idea using different datasets. Thus the seismic interpreter usually has to make the best of what they have and needs to resort to empirical rules to test their interpretation for plausibility.
Geological models in the exploration and production sector usually start with geophysics. Observations of geophysical effects may be either passive, such as magnetotelluric or gravity measurements, or induced, such as acoustic seismic or electromagnetic methods. This raw data is then subjected to a myriad of mathematical derivatives or cross-correlations, ideally validated against well or outcrop data, to produce perceived direct detection of lithology, reservoir properties, and in some cases, hydrocarbons. The geophysics workflow continually gathers momentum with the availability of new processing techniques such as spectral decomposition, acoustic/elastic impedance and other inversions, wave-form facies, and amplitude vs. offset analysis, while the workflow based on the interpretation of horizons and faults has languished. It is the latter workflow, with the building of structural framework models at its heart (Krantz et al., 2016) that can so often be grossly wrong. The geophysics workflow is more closely related to traditional science. It is heavily dependent on physics, where experimentation can be used to test theories while the structural framework workflow begins with an interpretation of processed geophysical data rather than the direct observation of an effect that can be modeled and checked against well data. This means that whatever tests we are able to apply have significant margins of error. They are further restricted by the intellectual issue of constructing a 3-D view of the perceived reality, initially in the interpreter’s mind and then in 2-D, either on paper or on the computer screen. This model building is an early stage of the exploration and production process and is of huge importance for this industry, where the building of a water-tight framework model is often the foundation for further geological, geophysical, or engineering purposes. The introduction of interpretation errors at this point, together with location uncertainty arising from the geophysics, will propagate throughout the subsequent workflow. Possibly these are the most significant factors that influence exploration risk, affecting perceived value and in particular drilling risk. Again, unlike other branches of science where a negative test can be viewed, to some degree, philosophically, a negative result from a drilling plan carries with it enormous cost implications.
Simple Questions and Rules to Test 3-D Interpretation of 2-D and 3-D Seismic Data
Without recourse to the standard scientific method of hypothesis testing by experiment, the modeling workflow must use other approaches to mitigate against structural uncertainty as well as absolute errors. Fortunately, the legacy of structural geology provides us with an intellectual schema that helps to distinguish between plausible and implausible models. In particular, the visual representation of fault surfaces and their relationships to the horizons they intersect is a useful basis on which to summarize very significant and diagnostic structural information.
Here we set out some simple questions and simple rules, which we have found useful for testing the geological credibility of 3-D seismic interpretations. Essentially, they are the same for 3-D and 2-D seismic data.
- Does the surface mesh derived from fault segments look orderly? A common mistake among both inexperienced and experienced interpreters is to pick fault segments only in the predominant dip-direction and ignore those in the strike direction. In both 2-D and 3-D seismic data, there is often valuable information in the strike line data that help to guide and define the shape of the fault surface (Figure 4). Sometimes the picking of fault segments is done on highly squeezed or vertically exaggerated sections. While these views can be helpful in identifying faults with small offsets, they can instill an impression of steep faults and divert from the true geometry. Furthermore, interpretations progressed in this way are often restricted to the reservoir interval and lead to extremely truncated models with incorrect fault dips. Our recommendation is that initially the fault segments should be meshed by a constrained triangulation, preserving the fault segment edges as triangle edges in the mesh (Figure 4). This immediately highlights gross inconsistencies between fault segments picked on in-lines and intersecting cross-lines. These are known as fault-pick misties. A simple question to ask at this stage is, for the given fault surface topography, “can the hanging wall slide on the footwall unimpeded?” If it can’t, then the interpretation needs attention. Only after these checks and corrections should the fault surface be meshed with a smoothed grid or triangulation (Figure 5).
Are the modeled fault surfaces curvy-planar in form? Fault surfaces that appear irregular (having high-frequency curvature or roughness) are unlikely, at a seismic modeling scale (see above), to be real. When roughness is admissible, given the movement caveat above, it should be preserved in the final model. If the fault is to be represented as a grid, the grid resolution should not be less than 1/4 the distance between fault segments or 2x the 3-D seismic bin size (Figure 5).
- Is the fault surface in the correct position? To check that a fault surface is correctly located in 3-D seismic data, the seismic volume should be sliced at a set distance from the fault’s hanging wall and footwall sides. The slices should then be projected onto the fault surface (Figure 6). If the fault is located correctly and has the correct dip, these projections will be fundamentally different with offset of reflectors on either side, revealing the displacement on the fault. This workflow is also ideal for locating intersecting or splay faults in 3-D seismic data.
For 2-D seismic data, the question of fault location is tackled by addressing fault aliasing problems. In other words, if the fault spacing is closer than the line spacing, it can be unclear which segments on which lines tie together in 3-D space (Freeman et al., 1990). Mapping and analyzing fault displacement, longitudinal and shear strain (e.g., Freeman et al., 2010) can largely alleviate this problem.
- Are the longitudinal and shear strains associated with the fault within acceptable limits? Individual faults should be checked for instantaneous longitudinal and shear strain (Figures 7 and 8) at an appropriate scale where possible. This is a measure of how much deformation is implied by the interpretation. It combines both horizon and fault information and helps to assess structural consistency. Acceptable upper bounds for longitudinal and shear strain are approximately +/−0.1 and +/−0.05, respectively (Freeman et al., 2010).
Both measures of strain are useful when checking for plausible fault length at a particular horizon or an upper/lower tip and inform about how far the mapped fault may extend beyond seismic resolution. They also highlight improbable horizon-offset/separation polygon geometry. Ipso facto, these should be visible in the horizon isochores.
A general comment following on from points 1, 2, 3, and 4 is that interpretation is often restricted to the reservoir interval of interest, which is almost another kind of 2-D view. Consistency comes from seeing the bigger 3-D picture, which means it is nearly always beneficial to pick the full vertical extent of faults and to map horizons above and below the intervals of interest.
- Are there any busts in the horizon pick and horizon structure maps? Finding horizon pick or map inconsistencies in complexly faulted datasets can be particularly difficult. For 3-D seismic data, a very common workflow is to attempt to use an autotracking algorithm (Figure 9A). There are primarily two methods for creating an autotrack. One is by using a primary linear seed-pick on a set of in-lines and cross-lines to control the autotrack. The other is repeated, limited autotrack propagation using seed-points. In both cases the resulting autotrack should be tested for correlation value, distance from seed and their product, and correlation confidence (Figures 9B and 9C). Whichever method is used, the original seed and autotrack parameters should always be preserved for each propagation event, to allow repeat of the autotrack experiment and controlled adjustment of input variables. Horizon busts are normally associated with large jumps in autotrack confidence.Figure 9.
For 2-D data, reflectors should ideally be correlated around loops using fence-like seismic displays within a 3-D display. This is in addition to checking ribbon maps, whereby seismic lines are colored according to their interpreted two-way time, for elevation busts.
For both 2-D and 3-D data, fault displacement analysis is another powerful QC tool for verifying horizon interpretation in the vicinity of faults. Displacement anomalies on otherwise sensible looking fault surfaces often indicate horizon mispicks/misties similar to the size of the wavelet cycle.
- Are horizon isochore maps consistent with the geological model? Syn-sedimentary growth faults, if normal, should show thickening on their hanging walls (Figure 9D and 10), or if reverse, on their footwalls. For post-depositional faults the thickness variation is constrained to +/−10% by the limits of longitudinal strain (Freeman et al., 2010). Usually it should be less than 10%. Horizon isochore data are very sensitive to subtle horizon busts that might not be discernible given a single horizon pick or map. Ideally isochore data should be displayed as an attribute of the raw horizon seed pick or autotrack data as well as an attribute of any finished, gridded, and mapped surfaces. The former method allows for immediate feedback to the interpreter as they interpret, rather than waiting until the horizon is complete or semi-complete, by which time making changes will be considerably more difficult.
Has a 3-D framework model been built as part of the interpretation process rather than as an afterthought? This is one of the most important aspects of good practice in seismic interpretation, especially in areas of complex geology. While the capability of building framework models has been possible since the early 1980s, it is still an all-too-common occurrence that the interpretation of horizons and faults is completed by one person before being handed on the next to build the framework model. The two people involved are commonly deemed the geophysicist and geologist and the handover of responsibility for a project often occurs at this point. It is our opinion that the framework model should be built and tested for structural validity by the same person who does the interpretation, that is, the geophysicist in a traditional work place, in an iterative manner.
Has the interpretation been done primarily in a 3-D environment with true perspective? It is almost too obvious to state that cognitive unloading of the intellectual task of imagining a 3-D model to a computer should be the norm. Nevertheless, due to the inertia, budget, or other reasons, many exploration and production departments and companies still operate with old workflow practices on modern computers that restrict or encourage their staff to interpret in section and map mode.
- Have fault statistical tools been used to test for the completeness of fault-set interpretation? Being able to break faults up into sets, add up fault throw or heave across a study area, investigate the power-law relationship of fault size in terms of maximum throw or length, plot rose diagrams of strike or poles to planes on a stereo-net, and so on. (Figure 11), not only increases structural understanding, but also helps identify areas of possible missing faults within the interpretation.Figure 11.
Simple structural geological rules such as those presented here can assist the interpreter to create a robust 3-D model (Figure 12A). Managers and joint venture partners can use the same rules to judge whether the model presented to them is viable or equally importantly, not viable. Displaying simple attributes such as dip and azimuth on a horizon surface (Figure 12B and C) can increase understanding of structural evolution, which may aid in trap integrity studies.
Modern interpretation and modeling software is very sophisticated. It is now easy to generate beautifully persuasive realities that owe more to software design and programmatic flair than to geological realism. To put it another way, with 3-D seismic workstations, it is now possible to make bad interpretations in much less time than before. Furthermore, the availability of the 3-D interpretation/modeling workspace has had little impact on the requirement to understand structural geology in 3-D as opposed to the apparent structure of a single, autotracked horizon. Unfortunately, a pretty map or 3-D model does not necessarily equate with geologic reality. A good structural model is not simply a collection of water-tight fault and horizon surfaces. It is a set of surfaces that have to honour the rules of structural geology, which have been formulated by researchers in the field and in laboratories over the last 150 years. Hypotheses have been discussed, disproven, reformulated, and approved by consensus through peer reviewed scientific literature, and interpreted structural models should be tested by these same rules. In our opinion the best way to do this is during the interpretation process in an iterative manner rather than afterwards, as is still the case in many exploration and production companies today.
Consider the concept of “Model-dependent realism”, which Hawking and Mlodinov (Ch. 3, 2010) coined, whereby reality is merely a constructed best-case model of the facts we are aware of. The key word to focus on in his statement is aware. Thus reality, which in our case includes a good structural model interpreted from seismic data, is dependent on the awareness or education of both the interpreter and the software engineer or computer scientist who built the interpretation system in the first place. In other words, an interpreter cannot be expected to create an accurate representation of reality if they are not aware of what that reality might be.
Over the last 30 years or so the quality of seismic data has improved immensely. In the early 1980s, unmigrated time sections of faulted structure barely resembled geology. The interpreter really had to interrogate and interpret the geophysical image to finish with something looking geological. With modern day data acquisition, pre- or post-stack time or depth migration, seismic data looks like geology from the outset. Modern data has democratized the interpretation process so that anyone can think they can do it. Likewise, with the plethora of 3-D mapping and modeling platforms available, incorrect representations of perspective notwithstanding, anyone can be a modeler, if the seismic looks like geology, the interpretation and model looks real. It is our contention that if left untested for structural validity, it remains a sketch.
The Baconian scientific method of prediction, tested by observation and in our case using 3-D structural geologic rules, is the fundamental key to the process of mapping structures in the field and from subsurface data, representing them in three-dimensions and understanding their formation. Because of this, a majority of older structural geologists will have undergone undergraduate and postgraduate training in 3-D thinking by way of producing geological maps and cross-sections from the two-plus dimensional data of surface exposures and local structural measurements. Spending months or even years mapping in the field trying to unravel outcrop information in the Rockies, Alps, Australian outback, and others is a fundamental part of learning how to take data gathered on a two-plus dimensional plane and construct a 3-D model. Indeed, since three dimensions is the universal currency of structural geology, they most likely chose their vocational path because they were good at thinking in three dimensions.
However, with the recent move of many western universities toward user-pays funding models, expensive field-based mapping courses have progressively been dropped. Thus many geoscientists now graduate with minimal experience in reconstructing data gathered in the field to create map, cross-section, and 3-D models. Increasingly, fieldwork is confined to postgraduate and post doctorate level. Even postgraduate projects are skewed toward the use of industry data, workstation interpretation, and digital 3-D modeling. The result is that geologists with a background in structural fieldwork are becoming increasingly rare. However, possession of 3-D spatial skills and the building of 3-D water-tight geologic models of faults and horizons from seismic interpretation are fundamental to the creation of a reality for exploration and production managers, on which they make multi-million dollar decisions.
The use of quantitative structural geology is greatly undervalued in industry practices. For example, it is common for seismic imaging to deteriorate with depth. Here, the associated uncertainty can be reduced considerably with the application of structural rules. However, the industry response to this situation is much more likely to be focused on the considerably more expensive route of reprocessing existing seismic or shooting new seismic.
The industry noticeably still harbors the cultural divide of geology and geophysics, and this division often guides the workflow that an interpreter will use to generate their representation of subsurface reality. An out and out geophysicist is more likely to be interested in squeezing the last drop of physics from the seismic cube rather than testing that their horizon and fault picks make sense in the context of a structural model. There are also the issues of the time made available to do a good interpretation and the effect of corporate inertia. Interpretation, and especially the creation of water-tight geologically plausible models, is a lengthy process, and even in the face of expert opinion and advice a poor model may stand because (1) the interpreter will be reluctant to agree that the model is wrong and/or (2) it is too late to make significant changes because the resulting static and dynamic cellular models are already in use by reservoir geologists and engineers respectively. Furthermore, the quest for good 3-D interpretation is sometimes hindered by management, which more often than not underestimates the time to create a good product and often still requires 2-D, paper-based or PowerPoint reports to make their decisions. The most pragmatic way to combat these issues is to allot more time to interpretation so that when the 3-D structure is challenging, simple rules can be applied iteratively as the seismic interpretation evolves and allow the time for management to see the data and interpretation in true 3-D perspective.
Working in 3-D, true perspective is also an obvious step forward. If the interpreter favors traditional flat screen map and section-based views, then the final product will tend to be inferior. On the other hand, if their work is undertaken in a 3-D true perspective environment and simple, structurally-derived attributes for fault and horizon surfaces make sense when plotted in appropriate graphs, then in our opinion the geological model is likely to be more robust and closer to reality.
Corporate inertia and the cultural divide between geology and geophysics are an unnecessary expense for the exploration and production industry, yet they are alive and thriving in the modern industry, especially where seismic interpretation occurs. The human resources sector needs to be educated to encourage more crossovers between geophysicists and structural geologists trained in field interpretation. We propose that the title structural geophysicist be used.
For the industry to apply a truly multidisciplinary approach to its workflows, it should reconsider the 3-D process in some detail. It should also try to educate and inspire software vendors to develop innovative, state-of-the-art, and integrated products that enable the testing of 3-D interpretations during the interpretation process and provide exposure to these kinds of ideas through training across the board. Chiefly, since it is managers who are responsible for the exploration and production workflow, they above all must be educated about the many advantages of using a structurally qualified 3-D model.
Geologists have explored 3-D graphical representations over a long period of time, progressing from isometric to perspective drawings.
True perspective representations form the most natural visualization/interpretation environment.
Visualization and interpreting seismic in a 3-D environment assists the evolution of a robust 3-D model.
The geological framework is the main product of the interpretation process. Its plausibility can be tested using the simple rules stated within this chapter.
Geophysical imaging is usually imperfect, but structural geology holds some of the missing parts of the interpretation puzzle and can reduce structural uncertainty. Interpretation should honor the rules of structural geology, and with good training and appropriate software it is easily possible to get better quality interpretation from poor 3-D data and 3-D quality interpretation from 2-D data.
The relationship between industry and geoscience is complex. Individual company cultures may actually work against the efforts of geological 3-D analysis and thinking in its efforts to solve 3-D problems.
We acknowledge both Clare Bond and Doug Goff for their comments to help improve the content of this chapter. We thank Cue Petroleum Pty. Ltd, Melbourne, Australia, for permission to display examples of 3-D models in Figure 12.
Figures & Tables
3-D Structural Interpretation: Earth, Mind, and Machine
Three-dimensional geologic interpretation of surface and subsurface data requires integration and application of both geologic knowledge and spatial cognitive skills. Much surface geologic mapping still employs pen and paper techniques, but subsurface interpretation is usually accomplished using sophisticated visualization software. In both cases, successful interpreters use mental models that bridge internal and external forms of 3-D visualization to construct 3-D geologic interpretations. This AAPG Memoir 111 sets out to understand more about the convergence of geology, 3-D thinking, and software, which collectively provide the basis for truly effective interpretation strategies. It should appeal to all geologic interpreters, and especially those who investigate and teach interpretation skills.