Mega-Scale Broadband PSDM in the North Sea

Super-dataset sets new quality standard to help identify and exploit remaining opportunities in this mature area

Contributed by CGG

This line from the Cornerstone PSDM dataset co-rendered with the final velocity model, traversing 250 km through the Central North Sea area, shows how TomoML provides geologically consistent regional trends as well as local detail and definition (image courtesy of CGG Data Library).

Recovering the remaining and bypassed hydrocarbons in mature areas requires new approaches. A great deal of seismic exploration has taken place within the North Sea, using a variety of acquisition configurations, including the latest broadband solutions such as BroadSeis™. Most surveys in this mature area have been processed and reprocessed multiple times as techniques have evolved; using time migration and now mainly pre-stack depth migration (PSDM).

Within the Central North Sea area alone, CGG’s non-proprietary library includes data from 37 acquisition phases (26 main and 11 infill) spanning a 15-year period from 1999 to 2014. Collectively referred to as Cornerstone, this library covers approximately 35,000 km2, including 9,000 km2 with dual azimuth. Multi-client surveys provide a cost-effective means of acquiring high-quality data. Even in relatively mature basins, such as the North Sea, where acreage is shared by many operators, multi-client surveys can be a valuable tool for development.

The majority of seismic projects within the North Sea independently focus on specific small areas, primarily restricted by the high work effort involved in integrating all available data (wells, horizons, vertical seismic profiles, etc.) into the velocity model, combined with the lengthy time scales required to update the model using traditional tomography methods. The development of TomoML, an algorithm for multi-layer tomography, now allows completion of complex PSDM projects within a significantly reduced time scale while simultaneously improving image quality.

Taking full advantage of TomoML, CGG has recently completed a major project to combine its entire Cornerstone Central North Sea data library into a single, seamless 35,000 km2 volume of high-quality broadband PSDM seismic, along with associated velocity and anisotropy models. This work required only 18 months to achieve. The final data can be used to perform rapid AVO screening, pore pressure prediction and prospect evaluation on a regional scale but with the level of detail usually associated with smaller, reservoir-focused projects.

A vital new addition was the application of a proprietary pre-imaging deghosting technique known as Ghost Wavefield Elimination (GWE). This improved resolution by extending the seismic bandwidth. It also facilitated a robust merge across the multiple vintages, although low-frequency matching between conventional and broadband acquisitions was challenging due to the lack of recorded lows in the legacy seismic. In addition, all datasets were re-datumed to mean sea level during GWE. The processing sequence was further refined with advanced demultiple techniques including shallow water demultiple and residual peg-leg multiple suppression, applied to selected parts of the dataset.

This vast project did not involve only seismic data. Information from 225 wells was used in the construction of the initial velocity and anisotropy models, with the TTI axis constrained by a total of eight structural horizons. The North Sea exhibits a complex layered geology with abrupt changes in acoustic impedance that are observed in the well logs. The majority of these key reflection boundaries have an excellent spatial interpretability. This facilitated the construction of a layer-based initial model which matched the geology from the outset.

Two passes of TomoML were used to optimize the velocities for final imaging. Each pass incorporated dip-constrained inversion, which performs a 3D joint inversion of residual moveout and offset-dependent dip picks, in order to estimate shallow velocity variations caused by Quaternary channels. The multi-layer tomography updated the entire depth range of the model from 0 to 10 km, with shallow channels handled by the dip constraint and anisotropy parameters adjusted as necessary to preserve the imaging velocity. Final multi-layer well calibration achieved an absolute average depth miss-tie of less than two per cent for all horizons, without introducing distortions into the model velocities or seismic image.

The final model reveals geologically consistent regional trends in the velocities with excellent stability and a good correlation with the seismic image. A velocity depth slice through the Quaternary glacial channels reveals high-frequency details which are well correlated with the complex channel structure and show consistent velocity trends extending over many kilometers. This geological consistency is also observed in the deeper section where conventional methods often fail to deliver spatially stable results. The stable velocities improved the imaging of the pre-Cretaceous section, which was further enhanced by the bandwidth extension from GWE.

As new and improved technologies such as BroadSeis become available, continued evolution and refinement of data and models enable operators within a mature basin, such as the North Sea, to maximize hydrocarbon recovery. The creation of this PSDM super-dataset provides consistent regional coverage with local detail and definition, setting a new quality standard in North Sea seismic. This will enable reliable identification and exploitation of any remaining opportunities within this mature area.

Visit CGG on booth #440 for a closer look at our latest Cornerstone data and state-of-the-art processing and imaging capabilities. We will also be presenting daily on this topic in our booth theatre - check with booth staff or visit cgg.com/eage2015 for the latest schedule.