More Accurate Subsurface Imaging - CWI

Seismic images need to provide accurate information from the seabed down to the reservoir. For without a solid description of the shallow geology there can be no certainty that our calculations regarding the deeper reservoir layers are correct.

Key Benefits

  • More accurate imaging and modeling of the entire substrata from the seabed to the reservoir
  • Improved accuracy of reservoir characterization
  • Better detection of potential shallow hazard threats

Producing a detailed and accurate image of the reservoir rock and the geological structures that contain hydrocarbons deep inside the subsurface is crucial for placing wells and planning effective recovery. For decades 3D seismic has been the method of choice for creating such an image of the subsurface.

Imaging a hydrocarbon reservoir with seismic waves is complicated, as many kilometers of rocks and sediment lie between the surface and the target, ’obscuring’ the view. The speed with which a seismic signal travels through these various rocks, and the amplitudes of the sound waves that are reflected back to the surface, need to be calculated very precisely to produce an accurate subsurface image.

Very shallow rock formations can be the most challenging to measure. However, it is precisely those very near surface layers that oil companies are increasingly concerned about, as these may pose hazards for drilling and infrastructure. They also need to ensure that no production-related contamination occurs through near surface fractures and pathways. Shallow gas accumulations or paleochannels filled with soft and unconsolidated sediments can also have a serious effect on the propagation of seismic waves and our ability to create a reliable reservoir image.

More than the sum of its parts

So, how can we compute accurate velocity information in areas of complex shallow overburden geology? The answer may lie in a new processing workflow that makes use of various wavefield recordings from the co-located hydrophone and vertical velocity sensors on GeoStreamer® which record different properties of the same seismic wave. The sensors produce complementary recordings from which we can calculate local wavefield components. It is those components, in addition to the original hydrophone and velocity sensor recordings, that are utilized in a new velocity model building and imaging workflow called Complete Wavefield Imaging (CWI).

CWI Waveforms

The CWI workflow is comprised of three main elements: Reflection Tomography, Full Waveform Inversion (FWI) and Separated Wavefield Imaging (SWIM®).

  1. Reflection tomography – ensures a globally consistent initial velocity model as a starting point for full waveform inversion, improving its stability.
  2. Dual-sensor streamer data that is very rich in low frequency amplitudes has been shown to further improve the resolution and accuracy of the shallow velocity model derived from FWI, resulting in an overall more accurate model for imaging.
  3. The final step consists of computing SWIM gathers using a breakthrough technology that permits a much more extensive subsurface illumination of the earth, by including sea-surface reflections. This provides more reliable information about the quality and accuracy of the FWI velocity updates in the very shallow subsurface.

Each of the three components of the CWI workflow employs a different wavefield. Reflection tomography uses primary reflections only, whereas FWI utilizes the very low frequencies in the refracted waves, and SWIM uses reflections from the sea surface. SWIM gathers have been crucial in ensuring that the final velocity model is globally consistent and suitable for producing accurate depth images, from the shallow layers beneath the seafloor all the way down to deeper reservoir targets.

Putting it all together – A case study example from the North Sea

In 2009, the first 3D GeoStreamer survey in the North Sea was acquired over the Edvard Grieg, Johan Sverdrup and Luno fields in the southern part of the Utsira High in the Norwegian sector of the North Sea. The giant Johan Sverdrup discovery, made by Lundin and partners Statoil and Maersk Oil in 2010, is one of the five largest oil discoveries ever on the Norwegian continental shelf. The discovery in well 16/2-6 at Avaldsnes, on the south-eastern flank of the Utsira High, was followed by 30 appraisal wells, including eight sidetracks in PL265, PL501 and PL502 (Blocks 16/2, 16/3 and 16/5).

The main reservoir at a depth of 1900 m is composed of the Upper Jurassic Draupne sandstone dominated by coarse sandstones with average permeability in excess of 30 Darcies. The Avaldsnes High appears to play a key role in the distribution of the high-quality reservoir sandstones up to a distance of more than 10 km from the paleo-shoreline.  The reservoir is complex and the field represents a very challenging area for high-resolution imaging. Accurate seismic-to-well depth ties are particularly critical, given that the reservoir is estimated to be less than 50 m thick for much the Johan Sverdrup discovery. Historically a variety of near-surface velocity anomalies would have made this impossible. Improving reservoir delineation and solving depth conversion problems were essential steps during the field appraisal.

Velocity depth model (depth (m)A) Wavelet shift tomography with the Kirchhoff depth migrated image overlaid. B) FWI with Kirchhoff overlaid. Data: Lundin

 

Kirchhoff PSDM v. SWIMA) Traditional Kirchhoff PSDM of shallow overburden. B) SWIM depth image of the same shallow section. Notice the improvement in the top section of the image and clearer deeper sections due to improved illumination and increased fold. Data: Lundin

The asset team launched a pre-stack depth imaging project, using the existing dual-sensor towed streamer data, to create a velocity model that would accurately account for small-scale heterogeneous near-surface velocity variations. An older model built on conventional 3D data was used as the starting point to speed up the process. However, due to the shallow water depth (85-115m) in the survey area, conventional reflection tomography on the legacy project had failed to produce a sufficiently accurate shallow overburden model. To avoid possible instabilities in the later application of FWI, reflection tomography was applied to the GeoStreamer data, to improve the initial velocity model with a particular focus on a more precise estimation of both velocities and the laterally-consistent anisotropy parameters in the shallow overburden. These velocity updates provided a much-improved match between modeled and observed refraction data, permitting the subsequent FWI updates to resolve additional high-resolution velocity variations associated with channels, pockmarks and gas chimneys in the shallow overburden.

SWIM angle gathers were used to validate the longer wavelength features of the updated velocity model. These are not easy to observe or cross-check in the data domain using FWI. The additional illumination achieved by imaging all wavefields provides a significantly enhanced higher resolution shallow overburden image. This was of particular importance as a shallow wedge covers large parts of the field. The long wavelength velocity variations associated with this wedge structure have a significant impact on the vertical position of the target sands in respect to the oil-water contact. For the deeper part of the overburden, particularly the chalk layer and the target zone, high-resolution reflection tomography was applied. Significant improvements were made in the quality of the final high-resolution subsurface image compared to the legacy PSDM data, and the combined project objectives of improved reservoir delineation and seismic-to-well depth ties were achieved.

Johan Sverdrup reservoir horizonsSubstantial depth shifts apparent on the CWI model are highlighted at a virtual well location, suggesting a much-improved match with the reservoir horizons at various well locations. Data: Lundin

Knowing the shallow brings the deep into focus

Oil companies require a very accurate and reliable structural image of the reservoir. Producing such an image using seismic data requires detailed knowledge of the propagation speed of sound in all the rocks overlying the reservoir. Estimating velocity information in the very shallow section of the overburden, closest to the ocean floor, is especially challenging – particularly in areas of relatively shallow water. Any errors in estimating the propagation speed of seismic waves in the near surface will lead to significant uncertainties in the position and shape of the reservoir structure at depth.

There is another good reason for requiring an accurate picture of the sub-surface from shallow to deep. A more detailed image of the rocks at or close to the seabed can help oil companies to place seafloor equipment, identifying potential hazards and avoiding structural changes which may result in the opening of faults and fractures during production.

So no matter how deep the targets, to improve the reliability of reservoir estimates and reduce the risk of finding and producing them, CWI provides an important addition to the reservoir imaging toolbox.