Wilfred Whiteside, Wenlong Xu, Zhiming Li, Ashley Lundy, Itze Chang - TGS-NOPEC Geophysical Co.
Significant improvements in defining salt boundaries and in positioning reflectors and salt overhangs by properly accounting for the effect of anisotropy were achieved by TGS-NOPEC Geophysical Co. (TGS) in its Mississippi Canyon Revival survey covering some 660 Gulf of Mexico OCS blocks. Success hinged on the strategic combination of a well tying anisotropic sediment model, anisotropic pre-stack Kirchhoff migration, modeling of salt bodies with overhangs, and iterations of both supra- and subsalt tomography on a regional scale.
The Mississippi Canyon area of the GoM has many deepwater discoveries and is an area of intense interest. It can be challenging due to the large volumes of subsurface salt in the area. Seismic imaging with accurately positioned events both laterally and vertically is essential for exploration, but these large salt bodies make this difficult.
660 OCS blocks of the Mississippi Canyon Revival survey (highlighted) were imaged.
The key to TGS’ strategy was the proper calculation of the effects of anisotropy, or the directional dependence of seismic velocity through a rock layer. The recently completed case study of anisotropic model building, and the results proved the value of these calculations: Subsalt events were placed more accurately and image quality was dramatically improved. Properly accounting for the effects of anisotropy also affected the measured depth of a reservoir, as well as the lateral positioning and the estimated volume of the pay zone.
The goals of this project were accurate event placement and improved imaging of steep dips, salt boundaries, and subsalt events. To accomplish these goals, all available well information was used to calibrate the seismic velocity model. A well tying anisotropic sediment model was used in combination with anisotropic pre-stack Kirchhoff migration, modeling of salt bodies with overhangs, and iterations of both supra- and subsalt tomography, all on a regional scale.
The finalized seismic image volume placed events with a much higher degree of accuracy than previous work and improved image quality. Salt boundaries and steep or overturned events were brought out that previously were either inferior or failed to image. The deep and subsalt events were more focused and continuous with improvements in lateral positioning and structure. The results yielded a model suitable for a higher end imaging algorithm, such as reverse time migration.
Project preparation
The MC Revival survey is in an area of the GoM that has large volumes of salt which make imaging a challenge because salt interferes with the accurate focusing and positioning of events essential for seismic success.
TGS prepared by collecting all necessary data. Prior model building and imaging of this survey had been performed using an isotropic wave equation migration, and an isotropic velocity model. Extensive well log data also were available and allowed for calibration of the isotropic seismic velocities. This provided the starting point for the anisotropic model building.
As a supplement, three additional 3D surveys were prepared to enhance signal-to-noise via swell noise removal and Radon demultiple. Careful matching tied the surveys to each other. Additionally, water depth and starting salt surfaces from these surveys were available along with horizons from surrounding 2D data. This information provided the ability to form an integrated set of horizons extending well beyond the imaging area.
Initial anisotropic model building
The previous isotropic wave equation migration image with two well logs posted on the image (gamma ray and sonic logs) had several discrepancies to address. For instance, the depth of the top of the carbonate layer was 610 m (2,001 ft) deeper than the top of the layer found in the well logs. In addition, the steep salt boundaries were not well focused in the image due to the dip limitation of wave equation migration.
The lateral positioning and curvature of the top of the Oligocene layer under edge of salt is more accurate.
Once the anisotropic Kirchhoff migration using the anisotropic velocity model was run, the accuracy of event placement improved. As a result of these calculations, the depth of the top of the carbonate layer now matches the well depth. In addition, the steep salt flank and overhang of the left side of the salt body are clearly focused in the anisotropic prestack depth migration image.
Also of interest in the anisotropic migration is that the structure of the carbonate layer changes under the edge of the salt. This can be seen comparing contours of this key layer on the isotropic and anisotropic migrations, respectively. The apex of the event shifts laterally 400 m (1,312 ft) and the curvature of the event is decreased on the anisotropic migration.
This illustrates that anisotropy can have important consequences in reservoir analysis. Failure to account for anisotropy can affect not only the measured depth of a reservoir, but also the lateral positioning and the estimated volume of the pay zone.
Velocity model update
The next step was to update the anisotropic sediment model. This required two passes of grid based tomography, which required 3D anisotropic pre-stack Kirchhoff migration and automatic dip estimation run for each iteration. In addition, the salt model was built in four stages to include salt overhangs in the model.
The Kirchhoff, with its ability to image steep and overturned events, made proper modeling of the salt overhangs possible. With the overhangs properly accounted for, the salt bases are focused more clearly.
Salt overhangs are now properly structured and salt bases are more focused.
Sediment structure in the vicinity of salt was also greatly improved. The isotropic wave equation migration and anisotropic Kirchhoff images in an area where the Kirchhoff algorithm and more accurate salt modeling were critical in imaging the steep sediments near the salt are shown.
The next step was to migrate the data with the completed salt model and to compute residual curvatures for use in the final tomographic inversion update. The updated regions of the model included both sediment under salt and sediment away from salt. Care was taken not to include deep areas with a higher proportion of remnant multiple energy in the analysis. The model was updated from the inversion results, yielding the final anisotropic velocity model.
The final imaging step was performed with the anisotropic pre-stack Kirchhoff migration. In its initial version with the isotropic wave equation migration image, the salt model lacked needed overhangs, and the wave equation migration did not capture the steep dips needed to image them. Also, the base of salt is highly curved and unfocused.
In the corresponding final anisotropic image using the new anisotropic model, the steep salt overhangs are imaged better on both sides of the salt body, and the base of salt is quite flat and well focused. Additionally, the sediment truncation against the salt is much better defined than in the wave equation migration. The shadow zone beneath the overhang on the right side of the salt body shows sediments being imaged that are not apparent in the wave equation migration.
As an added benefit, the improved anisotropic model was available to use in imaging with anisotropic prestack reverse time migration. Reverse time migration has the best of both worlds: the advantage of the wave equation migration for imaging under salt along with the ability of Kirchhoff to image steep and overturned events. Application of the reverse time migration algorithm further enhanced the salt boundaries and sediment layers under the salt.
The processing group at TGS continues to apply these methods to similar, ongoing prestack depth migration projects. One project they are currently working on involves TGS’ Sophie’s Link and Eastern Mississippi Canyon surveys, which will cover 553 and 227 blocks in the GoM, respectively.
The eventual goal is to have the entire Gulf Coast and GoM area covered with anisotropy processing.
