Paul Boughton
With the oil and gas industry’s recent focus on 3D and 4D exploration technologies and powerful integrated workstations, the wealth of legacy 2D seismic data that exists has tended to be overlooked.
Today, existing conventional 2D data, of varying vintages, can be integrated with single and multiple 3D surveys, if available, and visualised and interpreted within the same, unified 3D space.
This capability brings a number of real benefits to the interpretation process.
For example, for the first time, the seismic interpreter can now view subsurface features present on 2D lines in a 3D space. Using value thresholding as a control allows rapid recognition of such features as channels and fault systems as they are sampled in 3D space by 2D slices through the data sets.
Furthermore, some of the more productive tools used for 3D interpretation are now available for 2D data as well, including automatic smart horizon tracking through the different surveys and automatic fault tracking on sections, along with smart display technology.
Next generation systems
Traditional 2D interpretation typically uses a series of section displays with interpretation performed on a section-by-section basis. However, with the introduction of the next generation of interpretation systems, full production interpretation functionality is now at the interpreter's fingertips for regional- and prospect-scale interpretation for single survey and/or multiple survey projects, including both 2D and 3D data.
Functionality includes data visualisation in 2D and 3D, survey balancing, and seismic-to-well ties. Classical manual interpretation and highly productive automatic horizon and fault interpretation tools are available, including mapping and plotting functionality, attribute generation, and horizon-based attribute extraction. The seamless combination of 2D and 3D view windows enables faster and more accurate interpretation in any geological environment.
Smart waveform-based horizon auto-tracking consists of propagating horizons through 2D surveys, offering hereditary path and threshold quality control tools similar to those used in 3D surveys. Direct interpretation of multi-z horizons, such as occur with over-thrusting and salt overhangs, along with auto-tracking tools and automatic triangulation of surfaces while picking, can significantly aid interpretation.
One issue that is inherent both within single 2D surveys as well as between multiple 2D surveys is data misties at intersection points, either as time shifts or as response changes. Within the same survey, the data misties are typically the result of recording in both the dip and strike directions. The strike direction records data updip (the shortest distance between shot/receiver and the reflector) and on the processed section. As a result the reflectors come in at an earlier time.
Between multiple 2D surveys, misties typically are compounded by multiple recording systems and different processing workflows, which result in both time and phase shifts. In the past, many long hours of the seismic processor’s time have been spent analysing and attempting to fix misties within and between surveys, in addition to many hours in meetings presenting test results prior to final application.
These issues can now be fixed effectively on-the-fly and saved permanently for future work. Mistie analysis between intersecting lines can be performed with a click of the mouse, with a table and map of intersections being displayed that show the time shift and phase characteristic measured. All that is required is to select which one to normalise to and the task is completed.
Horizon propagation
Although the actual horizon picking or reflector tracking component of the interpretation has traditionally been a very time-consuming task, horizon propagation is an efficient, field-proven method that searches for optimal waveform correlation along a seismic reflector. Enveloping reflectors can be used to guide horizon tracking from initial seed points or grids, ensuring the construction of reliable horizon interpretations.
Waveform-based horizon auto-tracking marks a significant step forward in computer-assisted interpretation, even in poor signal-to-noise areas. Auto-tracking can work at the regional scale by propagating horizons directly from data on disk at 8-, 16- or 32-bit data formats, or at the prospect scale utilising threshold rendering.
Threshold-based interpretation techniques can be used to reveal and place seeds on fragmented horizons with characteristic signatures, such as faulted channel systems viewed in 3D space. Propagation workflows enable enhanced user control, especially in geologically challenging areas where other auto trackers fail.
The interpreter is able to maintain total supervision over the handling of event character changes, for stopping at discontinuous events (faults, channels, etc.) and for methodical progress through areas affected by noise or other perturbing factors. Fast, true 3D visualisation, coupled with 2D scanning, enables rapid quality control of interpretations. The quality control process is further enhanced by the concurrent visualisation of well markers, faults and other horizons, ensuring that new picking is consistent with the structural framework.
This propagation in 3D space provides the highest quality event tracking through sub-sample precision. With 3D data, seed points are extended into picked regions by correlating trace shape and then propagating along the path of highest correlation, rather than enforcing a systematic, grid-based progression. This selective approach enables auto-tracking to be initially restricted to the highest quality seeds and correlations. With the trace shape analysis, doublets can be picked consistently, making the algorithm less prone to cycle skipping compared with amplitude-based auto-tracking.
Quantum leaps
The interpreter can continue horizon picking by adding new seeds, or can switch modes to edge propagation to grow the region using all horizon picks as seeds. Such workflows eliminate the typical requirement for extensive 2D manual interpretation prior to running auto-tracking. This unique capability offers unprecedented benefits to both seismic data interpreters and processors alike, bringing quantum leaps in productivity and flexibility.
Another important capability is horizon auto-tracking on volumes that are larger than RAM by utilising efficient data caching from disk. Unprecedented speed and accuracy bring a major acceleration of the interpretation process. For very large regional projects, productivity and system efficiency can be increased dramatically by restricting project limits to user-defined areas or by culling 2D lines dynamically. This frees memory. For rapid image visualisation and efficiency, rendering is only performed on the visible part of the image. Display fidelity is preserved during data zooming through smart seismic trace decimation.
Viewing seismic data using 3D visualisation is an extremely powerful tool that can have an immediate impact on the interpreter’s understanding of the subsurface. However, this has not been exploited in the world of 2D seismic until now. Not only can 2D data be viewed in 3D space, but they can be interpreted in the same space. This is extremely useful for viewing and interpreting complex stratigraphic features such as channels, as well as complex fault systems.
Channel features that are sampled on multiple 2D lines can be viewed in 3D space using numerical thresholding to determine the relative spatial changes, and can then be interpreted in 3D space with surface triangulation, all within the same window to produce a continuous map of the channel feature. This also holds true for fault systems. A fault system can be viewed through the data, interpreted with projected intersections viewed in 3D space on adjacent lines, and the interpretation finalised. The fault plane is then available in 3D space for further analysis.
2D attributes in 3D space
Seismic attributes certainly are not new in the world of 2D – but looking at them in 3D space certainly is. The ability to rapidly change the viewed data from the seismic or from one attribute to another again allows the interpreter to visibly scan for anomalous features present within the 2D sampled subsurface, especially when using amplitude thresholding to see through the volume from line to line.
Certain attributes have tended to lend themselves to the 3D seismic world. The Coherence Cube technology, originally made famous by Amoco and brought to the market by The Coherence Cube Company, is one such attribute and has long been looked on as an exclusive 3D tool. However, run in a 2D mode and co-rendered back with the seismic or attribute data, it can bring new insights into subtle fault and fracture detection. And when used in conjunction with opacity, it can be an extremely useful tool for 2D interpretation. Fig. 4 shows the coherence attribute co-rendered with seismic on two intersecting lines in 3D space.
The vertical signatures of features that can easily be recognised on 3D seismic data from their spatial response can be observed on the 2D data, which is very effective when simultaneously working both 2D and 3D data in the same workspace.
There is little doubt that the exponential increase in the number of 3D seismic surveys in the late 1970s and early 1980s was the catalyst for developing early versions of 3D interactive interpretation workstations, of which some had token 2D survey capabilities. However, it was 3D that dominated the advancement into the world of 3D graphics, automatic tracking and visualisation in spite of the overwhelming amount of 2D versus 3D data being worked for exploration.
With ongoing advancements in seismic interpretation technology, 2D data can now be treated in a similar fashion to 3D data in three-dimensional space. Rapid scanning through parallel sets of lines can be achieved using line lists, whereby the interpreter can examine structural variations between lines or between 2D surveys, just like one would do for 3D surveys.
By looking at 2D data in the 3D environment, basin-scale features can be visualised, while regional aeromagnetic or gravimetric maps can be draped over the mapped structural time surfaces from the seismic. For long regional 2D lines, parameters can be selected to allow the interpreter to roam along the regional line and identify the detail within the big picture.
Capabilities like these are putting 2D seismic in a new light and providing 2D data sets with a renewed status in the interpretation world, especially with the ability to view and interpret 2D surveys along with 3D surveys. When both types of data are put in the same 3D space, a symbiotic relationship comes into play, creating a new dimension by utilising legacy 2D seismic data for analysing the subsurface.
Note: The data set used to create most of the figures was supplied by the Victorian State Government, Australia. The cultural data are provided to Paradigm under license by IHS Energy, copyright 2006.
Tony Rebec and Tony Marsh are with Paradigm Geophysical Corp, Houston, Texas, USA. www.pdgm.com
Today, existing conventional 2D data, of varying vintages, can be integrated with single and multiple 3D surveys, if available, and visualised and interpreted within the same, unified 3D space.
This capability brings a number of real benefits to the interpretation process.
For example, for the first time, the seismic interpreter can now view subsurface features present on 2D lines in a 3D space. Using value thresholding as a control allows rapid recognition of such features as channels and fault systems as they are sampled in 3D space by 2D slices through the data sets.
Furthermore, some of the more productive tools used for 3D interpretation are now available for 2D data as well, including automatic smart horizon tracking through the different surveys and automatic fault tracking on sections, along with smart display technology.
Next generation systems
Traditional 2D interpretation typically uses a series of section displays with interpretation performed on a section-by-section basis. However, with the introduction of the next generation of interpretation systems, full production interpretation functionality is now at the interpreter's fingertips for regional- and prospect-scale interpretation for single survey and/or multiple survey projects, including both 2D and 3D data.
Functionality includes data visualisation in 2D and 3D, survey balancing, and seismic-to-well ties. Classical manual interpretation and highly productive automatic horizon and fault interpretation tools are available, including mapping and plotting functionality, attribute generation, and horizon-based attribute extraction. The seamless combination of 2D and 3D view windows enables faster and more accurate interpretation in any geological environment.
Smart waveform-based horizon auto-tracking consists of propagating horizons through 2D surveys, offering hereditary path and threshold quality control tools similar to those used in 3D surveys. Direct interpretation of multi-z horizons, such as occur with over-thrusting and salt overhangs, along with auto-tracking tools and automatic triangulation of surfaces while picking, can significantly aid interpretation.
One issue that is inherent both within single 2D surveys as well as between multiple 2D surveys is data misties at intersection points, either as time shifts or as response changes. Within the same survey, the data misties are typically the result of recording in both the dip and strike directions. The strike direction records data updip (the shortest distance between shot/receiver and the reflector) and on the processed section. As a result the reflectors come in at an earlier time.
Between multiple 2D surveys, misties typically are compounded by multiple recording systems and different processing workflows, which result in both time and phase shifts. In the past, many long hours of the seismic processor’s time have been spent analysing and attempting to fix misties within and between surveys, in addition to many hours in meetings presenting test results prior to final application.
These issues can now be fixed effectively on-the-fly and saved permanently for future work. Mistie analysis between intersecting lines can be performed with a click of the mouse, with a table and map of intersections being displayed that show the time shift and phase characteristic measured. All that is required is to select which one to normalise to and the task is completed.
Horizon propagation
Although the actual horizon picking or reflector tracking component of the interpretation has traditionally been a very time-consuming task, horizon propagation is an efficient, field-proven method that searches for optimal waveform correlation along a seismic reflector. Enveloping reflectors can be used to guide horizon tracking from initial seed points or grids, ensuring the construction of reliable horizon interpretations.
Waveform-based horizon auto-tracking marks a significant step forward in computer-assisted interpretation, even in poor signal-to-noise areas. Auto-tracking can work at the regional scale by propagating horizons directly from data on disk at 8-, 16- or 32-bit data formats, or at the prospect scale utilising threshold rendering.
Threshold-based interpretation techniques can be used to reveal and place seeds on fragmented horizons with characteristic signatures, such as faulted channel systems viewed in 3D space. Propagation workflows enable enhanced user control, especially in geologically challenging areas where other auto trackers fail.
The interpreter is able to maintain total supervision over the handling of event character changes, for stopping at discontinuous events (faults, channels, etc.) and for methodical progress through areas affected by noise or other perturbing factors. Fast, true 3D visualisation, coupled with 2D scanning, enables rapid quality control of interpretations. The quality control process is further enhanced by the concurrent visualisation of well markers, faults and other horizons, ensuring that new picking is consistent with the structural framework.
This propagation in 3D space provides the highest quality event tracking through sub-sample precision. With 3D data, seed points are extended into picked regions by correlating trace shape and then propagating along the path of highest correlation, rather than enforcing a systematic, grid-based progression. This selective approach enables auto-tracking to be initially restricted to the highest quality seeds and correlations. With the trace shape analysis, doublets can be picked consistently, making the algorithm less prone to cycle skipping compared with amplitude-based auto-tracking.
Quantum leaps
The interpreter can continue horizon picking by adding new seeds, or can switch modes to edge propagation to grow the region using all horizon picks as seeds. Such workflows eliminate the typical requirement for extensive 2D manual interpretation prior to running auto-tracking. This unique capability offers unprecedented benefits to both seismic data interpreters and processors alike, bringing quantum leaps in productivity and flexibility.
Another important capability is horizon auto-tracking on volumes that are larger than RAM by utilising efficient data caching from disk. Unprecedented speed and accuracy bring a major acceleration of the interpretation process. For very large regional projects, productivity and system efficiency can be increased dramatically by restricting project limits to user-defined areas or by culling 2D lines dynamically. This frees memory. For rapid image visualisation and efficiency, rendering is only performed on the visible part of the image. Display fidelity is preserved during data zooming through smart seismic trace decimation.
Viewing seismic data using 3D visualisation is an extremely powerful tool that can have an immediate impact on the interpreter’s understanding of the subsurface. However, this has not been exploited in the world of 2D seismic until now. Not only can 2D data be viewed in 3D space, but they can be interpreted in the same space. This is extremely useful for viewing and interpreting complex stratigraphic features such as channels, as well as complex fault systems.
Channel features that are sampled on multiple 2D lines can be viewed in 3D space using numerical thresholding to determine the relative spatial changes, and can then be interpreted in 3D space with surface triangulation, all within the same window to produce a continuous map of the channel feature. This also holds true for fault systems. A fault system can be viewed through the data, interpreted with projected intersections viewed in 3D space on adjacent lines, and the interpretation finalised. The fault plane is then available in 3D space for further analysis.
2D attributes in 3D space
Seismic attributes certainly are not new in the world of 2D – but looking at them in 3D space certainly is. The ability to rapidly change the viewed data from the seismic or from one attribute to another again allows the interpreter to visibly scan for anomalous features present within the 2D sampled subsurface, especially when using amplitude thresholding to see through the volume from line to line.
Certain attributes have tended to lend themselves to the 3D seismic world. The Coherence Cube technology, originally made famous by Amoco and brought to the market by The Coherence Cube Company, is one such attribute and has long been looked on as an exclusive 3D tool. However, run in a 2D mode and co-rendered back with the seismic or attribute data, it can bring new insights into subtle fault and fracture detection. And when used in conjunction with opacity, it can be an extremely useful tool for 2D interpretation. Fig. 4 shows the coherence attribute co-rendered with seismic on two intersecting lines in 3D space.
The vertical signatures of features that can easily be recognised on 3D seismic data from their spatial response can be observed on the 2D data, which is very effective when simultaneously working both 2D and 3D data in the same workspace.
There is little doubt that the exponential increase in the number of 3D seismic surveys in the late 1970s and early 1980s was the catalyst for developing early versions of 3D interactive interpretation workstations, of which some had token 2D survey capabilities. However, it was 3D that dominated the advancement into the world of 3D graphics, automatic tracking and visualisation in spite of the overwhelming amount of 2D versus 3D data being worked for exploration.
With ongoing advancements in seismic interpretation technology, 2D data can now be treated in a similar fashion to 3D data in three-dimensional space. Rapid scanning through parallel sets of lines can be achieved using line lists, whereby the interpreter can examine structural variations between lines or between 2D surveys, just like one would do for 3D surveys.
By looking at 2D data in the 3D environment, basin-scale features can be visualised, while regional aeromagnetic or gravimetric maps can be draped over the mapped structural time surfaces from the seismic. For long regional 2D lines, parameters can be selected to allow the interpreter to roam along the regional line and identify the detail within the big picture.
Capabilities like these are putting 2D seismic in a new light and providing 2D data sets with a renewed status in the interpretation world, especially with the ability to view and interpret 2D surveys along with 3D surveys. When both types of data are put in the same 3D space, a symbiotic relationship comes into play, creating a new dimension by utilising legacy 2D seismic data for analysing the subsurface.
Note: The data set used to create most of the figures was supplied by the Victorian State Government, Australia. The cultural data are provided to Paradigm under license by IHS Energy, copyright 2006.
Tony Rebec and Tony Marsh are with Paradigm Geophysical Corp, Houston, Texas, USA. www.pdgm.com
