Search

To better understand deepwater imaging challenges in the Gulf of Mexico, we constructed a large 3D model based on the complex salt geology of the Garden Banks protraction area. We simulated a regional wide-azimuth (WAZ) streamer survey with offsets of ±8 km inline and ±4 km crossline over this model. Using the true velocity model, reverse time migration of this data set produced a usable subsalt image nearly everywhere. The same synthetic seismic data set was passed to interpreters and geophysicists to process and image as if it were real field data. They did not get to see the true velocity model. Instead they performed conventional “migrate, pick, and flood” top-down velocity-model building, followed by final imaging through their interpreted velocity model. Where the salt was relatively simple, we found that the interpreted velocity model was reasonably accurate. Reverse time migration of the seismic data through these parts of the velocity model produced an image that was imperfect, but still usable for exploration-scale, structural interpretation. Where the salt structure was complex, however, it was sometimes grossly misinterpreted. The ensuing large-scale errors in the interpreted velocity model resulted in an unusable, shattered subsalt image. We next simulated a low-frequency, ocean-bottom-node (OBN) acquisition, with offsets up to 30 km in all azimuths. We started with the imperfect model produced by the interpreters and investigated what it would take for full-waveform inversion (FWI) to successfully find and fix the gross interpretation errors. Initial results suggest this “interpretation followed by FWI” methodology could produce a velocity model capable of generating an adequate subsalt image, but it may require ultrawide offsets (30 km), ultralow frequencies (below 2 Hz), and improved FWI algorithms to do so.

To realize the vision of automatic velocity model building with full-waveform inversion, a paradigm shift in the way seismic data are acquired is necessary. In December 2017 – January 2018 BP acquired 1000km of source lines over the deepwater sub-salt Mad Dog field in the US Gulf of Mexico using our prototype low-frequency source (LFS), Wolfspar. This was the first industry test of a new acquisition paradigm: a 3-D marine survey designed primarily for velocity-model building in areas of complex geology, using a source specifically designed for that task. It was a simultaneous-source survey, with two airguns and the LFS being recorded by the same ocean-bottom nodes. The Field Trial demonstrated that the source could operate for the required length of time for a 3-D survey, operate simultaneously with airguns without interfering with the airgun data or impacting operations, record low-frequency data at offsets wider than has been possible with airguns, and the resulting data could successfully be used for Full-Waveform Inversion to create an improved velocity model. We believe that the future of seismic acquisition in areas with complex geology will require bespoke surveys for velocity model building, and this experiment was a necessary first step.

The permanent ocean-bottom array at the Valhall Field in Norway provides an excellent source of passive seismic data to test what might be accomplished with seismic interferometry. The array was installed in 2003 (Kommedal et al., 2004) and data can be recorded for long periods in all weather conditions. The subsurface structure is well known, both from numerous wells and from seismic imaging. During periods without active seismic acquisition at Valhall, there is abundant passive energy in the data over a wide range of fre-quencies.

Subsalt imaging has been a long-term challenge for the oil and gas industry. The substantial progress made in data acquisition and imaging since the late 1990s has made some subsalt imaging problems tractable, but building earth models that enable imaging under complex salt remains a challenge. Labor-intensive workflows remain industry standard practice. Not only are these costly and time consuming, they have also performed poorly in many areas of economic interest. Various automatic model-building tools have been proposed to overcome these disadvantages. One such tool, full-waveform inversion (FWI), has already revolutionized velocity-model building in areas with shallow gas. Prior to 2006, imaging in these areas had been considered challenging and labor intensive, just as imaging under complex salt remains today. Modeling indicates that low frequencies and wide offsets may be the key to success when building velocity models using FWI. Just how low and how wide that may be required for FWI success depends on the particular problem. At the Atlantis Field in the deepwater Gulf of Mexico we recently acquired wide-offset ocean-bottom-node data with conventional airguns. By taking care during the acquisition, we recorded usable signal down to a lower frequency than previously achieved. We then applied FWI to the resulting data set and used the resulting velocity model, unmodified, to reverse time migrate the seismic data. It produced some of the best subsalt images of the Atlantis reservoir structure ever seen. Furthermore, the FWI velocity model revealed several major interpretation errors in the legacy salt model; thus the FWI result also offered an excellent basis for updating the salt model with the conventional workflow. These results demonstrate that with appropriate seismic data to support it, and with due care taken during processing and inversion, FWI truly offers a paradigm shift in model building and imaging in areas of complex salt.

Historically, seismic migration has been the practice (science, technology, and craft) of collapsing diffraction events on unmigrated records to points, thereby moving (“migrating”) reflection events to their proper locations, creating a true image of structures within the earth. Over the years, the scope of migration has broadened. What began as a structural imaging tool is evolving into a tool for velocity estimation and attribute analysis, making detailed use of the amplitude and phase information in the migrated image. With its expanded scope, migration has moved from the final step of the seismic acquisition and processing flow to a more central one, with links to both the processes preceding and following it. In this paper, we describe the mechanics of migration (the algorithms) as well as some of the problems related to it, such as algorithmic accuracy and efficiency, and velocity estimation. We also describe its relationship with other processes, such as seismic modeling. Our approach is tutorial; we avoid presenting the finest details of either the migration algorithms themselves or the problems to which migration is applied. Rather, we focus on presenting the problems themselves, in the hope that most geophysicists will be able to gain an appreciation of where this imaging method fits in the larger problem of searching for hydrocarbons.

The ninth International Workshop on Seismic Anisotropy (91WSA) was held on March 26–31, 2000, at Camp Allen Conference Center, about 20 miles from Houston, Texas. The 91WSA organization committee is listed above and was chaired by Leon Thomsen. Below are the abstracts of the 58 papers presented at 91WSA. Please see http://www.seg.org/9iwsa for expanded abstracts and proceedings of 91WSA.

Abstract Because of its location, the magnitude 5.2 Green Canyon earthquake of 10 February, 2006 has been of considerable interest to the oil industry. Unfortunately, because the nearby stations were all located to the North (onshore), its location and depth were only poorly constrained by the traditional worldwide seismic network. Fortunately, the Atlantis 3D survey happened to be ongoing at the time of the earthquake and recorded it on over 500 4C ocean-bottom nodes. The Atlantis array was closer to the event than any station on shore, and recorded the event from the South. These data thus provide an opportunity to dramatically refine our knowledge of the earthquake's location. We preserved 2.5 hours of data spanning the time of the earthquake and analyzed it for signals, both man-made and natural. Although there are a large number of overlapping signals in the data, they can be clearly separated by their distinct arrival times, temporal frequencies, and phase velocities across the 10km × 6km array. At frequencies of 5Hz and above repetitive air-gun signals from 4 ongoing seismic surveys dominate the data. These account for all the surveys shooting in the Northern Gulf of Mexico at the time. At frequencies of 1Hz and below there is a nearly continuous background noise coming primarily from the South and East that rumbles along for the entire 2.5 hours, with a phase velocity across the array of about 2000 m/s. This is most likely ocean wave-generated noise. At frequencies below 2Hz the earthquake signal dominates the data from 04:14:34 UT to about 04:20. It begins with several discrete arrivals from the NNW over a span of about 20 seconds, which are then followed by a drawn-out coda from the NW lasting over 8 minutes. At 04:22:13 a second event arrives from the SE, which is in turn followed by its own coda from the SSE, which lasts about 3 minutes. (It's not yet clear what this second event might be.) At about 04:27 a 3rd arrival, a weak, diffuse event from the ESE, fades in and lasts about 3 minutes. These data have been provided to earthquake experts and we expect to have a revised location for the event by the time of the OTC meeting in May. Preliminary indications are that the earthquake was well to the North of the originally published position. Making use of the Atlantis OBS-node array The geometry of the active nodes at the time of the earthquake is shown in Figure 1. We have a large, dense array (at least at low frequencies!) in this dataset, which allowed us to recycle a program called "Radar" for analyzing the signals in the data. "Radar" was originally written for examining surface-wave noise in land seismic data. It performs planewave stacks across the array and produces a graphical result showing the time, phase velocity, and azimuth of arriving energy. Figure 2 shows that "Radar" was able to correctly determine the azimuths of all the active air-gun surveys in the Northern Gulf of Mexico shooting at the time of the dataset.

We extracted 2.5 hours of continuous data from the Atlantis OBSnode array around the time of the 10 February, 2006 magnitude-5.2 Green-Canyon earthquake, for the purpose of characterizing the location, mechanism, and significance of this unusual event. The array, consisting of about 500 active 4C nodes, was several tens of kilometers to the South of the earthquake. Typical earthquakes radiate the bulk of their energy at frequencies much lower than the 10Hz geophones used in the nodes were designed to record, so the dataset also served as a testbed for understanding how standard exploration-seismic geophones might be used at frequencies below 5Hz. At traditional frequencies of 10Hz and above the Atlantis airgun signal completely dominates the data. However, filtering away frequencies above 2Hz removes the Atlantis airgun signal and reveals a strong series of arrivals from the earthquake. At “earthquake” frequencies of 2Hz and below, the Atlantis array becomes densely sampled in space (i.e., with more than 2 receivers per wavelength), allowing it to be formed into a powerful directional antenna. Beam steering the array reveals many interesting signals in the data. Examining our own airguns, we find that ”inconsequential” differences in our airgun arrays had unexpected and significant effects. Beam steering resolves a complex sequence of distinct arrivals from the Green-Canyon earthquake spanning 8 minutes of time, and two other mysterious events subsequent to the Green-Canyon earthquake that appear to be temporally related to it, but cannot be traditional aftershocks because they arrive from different azimuths.