Search

Exploration on the African continent has traditionally been concentrated in prolific basins on the Sub-Saharan Western margin, in detriment to the East. Proven prolific basins, existing infrastructure, favourable and progressively transparent regulation; along with the analysis of prospective conjugate basins of Brazil are just some of the reasons for creating a positive environment for international investors. A report submitted in 2012 by the USGS, predicts undiscovered mean gas resources of more than 370 TCFG in offshore East Africa, including Tanzania, Mozambique, Madagascar and Seychelles. World-class gas discoveries have been made between 2010 and 2013 in the offshore regions of Tanzania and Mozambique, along with onshore discoveries in Uganda and Kenya thus reinvigorating exploration interest along the Eastern African margin. The Zambezi channel, offshore Mozambique, has traditionally been the locus of intense academic research, with a number of 2D seismic and gravimetric and magnetic acquisition campaigns deployed in recent years. However, the region still remains poorly understood. With this in mind, in 2017 CGG acquired a high-resolution 3D seismic survey located in the outer Zambezi Delta Basin, west of the Beira High. This seismic survey, in conjunction with newly acquired high-resolution shipborne grav/mag survey and access to data from multiple wells in the area, aims to bring a new and fresh dimension on the geological understanding of the basin.

As the key step in subsalt imaging, conventional salt model building has typically relied on manual interpretation of the salt geometry, which proves to be difficult for resolving complex salt models. As a result, subsalt imaging has been approaching a plateau in the last few years, largely due to our inability to improve the accuracy of salt models. Fast forward to today, where the recent success of using full-waveform inversion (FWI) to automatically update salt models and significantly improve subsalt images has opened the door to a new era of subsalt imaging. As FWI for salt model updates prefers data with good low frequencies, long offsets, and full azimuth, sparse node-for-velocity surveys were proposed to serve as an economic yet suitable acquisition solution for large-scale subsalt exploration. Due to FWI’s prior inability to update salt models for field data, the feasibility of sparse node-for-velocity surveys was previously only studied based on synthetic data, from which the conclusions might not be immediately applicable to field data. Using our recently-developed FWI algorithm that proves to work on salt in field data and by decimating the densely-acquired Atlantis ocean bottom node (OBN) data, we studied the impact of sparse node data for FWI salt model updates. Based on the understanding gained from this, we further proposed and validated methods to improve FWI results with sparse node-for-velocity data.

Areas with complex overburdens pose a major challenge for seismic imaging of deep targets. Reduced illumination can compromise velocity estimation and the image at the target level. A solution to this problem is given by ocean bottom node (OBN) acquisitions. They provide increased illumination through larger offsets and full azimuthal coverage. In these complex cases, another issue that might arise is that of internal multiples. Strong reflectors in the overburden can generate internal multiples, which will be imaged incorrectly as artifacts. These artifacts may have detrimental effects on interpretation and amplitude analysis. The pre-salt in the Santos Basin, offshore Brazil is a good example of this scenario, where both illumination and internal multiples contribute to affect imaging in the target areas. While internal multiple attenuation (IMA) methods have been widely studied in the case of streamer acquisitions, a full solution tailored to OBN acquisitions is still lacking. We propose a solution based on a method recently proposed for streamer acquisitions.

Santos Basin has become one of the most prospective oil provinces in the world. Due to the geological complexity, OBN acquisition has emerged as an imaging solution. It provides the full-azimuth and long-offset illumination FWI needs to realize its potential. Given the right data, a reliable FWI engine and a good starting model, we are able to derive accurate velocities to impact the imaging of the salt and pre-salt. Two major challenges remain. FWI is dependent on the initial model and anisotropy remains a challenge. Better acquisitions together with advanced multi-parameter FWI technology are still needed to overcome these challenges.

Stampede field is a faulted subsalt four-way reservoir in Green Canyon, Gulf of Mexico. Imaging for part of the field has remained challenging due to interference from the complex overburden, which carries large velocity errors and creates non-uniform illumination for the subsalt. Before correcting the velocity error, least-squares reverse time migration (LSRTM) does not produce desirable subsalt image, even when using newly acquired ocean bottom node (OBN) data. With an improved OBN full-waveform inversion (FWI) model, combined with the benefits from the full-azimuth and long-offset coverage of OBN data, LSRTM greatly improves the subsalt image. However, further improving LSRTM is still challenging due to the remaining velocity uncertainty and un-modeled physics, as well as un-attenuated multiples and converted waves.

See document

In this case study, we discuss the impact of using a not-so-appropriate Wide Azimuth Data (WAZ) data which lacks good usable low frequencies and long offset full azimuth coverage for FWI based model building. We use Time-Lag FWI (TLFWI) demonstrated to be an appropriate algorithm by Zhigang et al 2018. We observe significant uplift in the shallow velocity model and salt overhang definitions at most locations. In this abstract, we discuss the limitations of using WAZ data for FWI and the additional effort to obtain better starting models to overcome the lack of sufficient diving energy and lack of low-frequency signal (<3Hz). We also briefly discuss the uncertainties in the inverted models esp. beyond the penetration depth of transmission waves.

we present a FWI scheme based on the quadratic Wasserstein metric, with adaptive normalization and integral wavefield. We show that this scheme has better convexity than traditional metrics, and therefore can mitigate cycle-skipping issues, while being insensitive to amplitude effects. We demonstrate the effectiveness of our approach with a streamer data set in an area of complex salt geometry. In addition, we show that this approach can work naturally on reflection data, without the extra procedure of scale separation by either decomposition or demigration. With these learnings, we believe that the fundamentals of FWI research are starting to converge.

The Gini 3D survey, acquired in the summer of 2018 in the Delaware Basin, provided a test area for which this type of design and acquisition could be tested and compared with traditional operations used onshore US. The test was planned to be operationally efficient & cost effective (in terms of equipment), with the aim of proving that blended acquisition can be as effective as traditional designs in terms of imaging. Operationally the test exceeded expectations and with processing still ongoing, initial results show that the two designs are comparable.