Search

Absorption effects in the Earth attenuate high frequency seismic signal progressively with depth, reducing resolution and limiting the interpretability of the data. Conventional post-migration Q compensation methods often rely on artificial mechanisms to limit amplification of noise, at the expense of unintentionally restricting recovery of signal. We propose an amplitude Q compensation technique using 3D sparse inversion. Comparisons with a conventional approach on synthetic and field datasets highlight improved signal recovery and noise suppression with the proposed method.

Time-lapse seismic is now being used more frequently to assist reservoir development, prevent infrastructure damage or monitor geological storage. To better reveal true 4D signals while suppressing acquisition-related noise as a result of, for example, water velocity changes, source positioning errors etc., a new processing flow which focusses on correcting each noise-contributing factor based on its physical characteristics, has been developed to replace the conventional non-deterministic correction approach based on cross-survey matching. Our proposed flow is based on using common water bottom and the water-bottom travel time to invert each factor and correct for it, which allows for processing of each monitor survey independently and the possible acceleration of standard 4D processing timelines. We applied this workflow on two fields offshore Angola, one with strong subsidence and one without, and showed the superiority of this new approach to reveal the true 4D information. The subsidence effect, observable from the reservoir up to the water bottom, now better matches with the model of pressure changes in the new 4D results compared to legacy results. Even for field experiencing no subsidence effect, the time shift and NRMS maps obtained at the reservoir level are cleaner and easier to interpret from new flow.

Successful applications of full waveform inversion (FWI) to land datasets are far less numerous than marine applications, yet the development of dense, long-offset broadband acquisitions has presented promising opportunities. While challenges exist due to elastic effects, acoustic land FWI has been shown to provide accurate velocity models with a level of resolution traditionally seen only with marine data. The first successful land applications in Oman have been obtained on surveys with only minor variations in surface elevation, and have encouraged the development of FWI capabilities to handle more significant topography. We present a boundary-conforming free-surface topography method for FWI, cast in the curvilinear domain. In a synthetic example, we benchmark this approach against the use of an absorbing surface boundary and a replacement velocity in the air layer (the model extension method), and the method of applying statics shifts to compensate for elevation variations. Finally, we show two real data applications from North and South Oman where our free-surface topography tool illustrates imaging uplifts over FWI results obtained with an absorbing surface and legacy tomography.

Spanning decades of exploration and production in the United Kingdom Continental Shelf, many programs of towed streamer data have shaped our knowledge of the Central North Sea. However, the fundamental lack of illumination and azimuth/offset coverage provided by towed streamer geometries, remains a blocker to resolving the imaging challenges associated with many higher-risk Jurassic and Triassic plays. This means that existing streamer data is rapidly approaching the limit in the value it can add to our understanding of this mature basin. The Cornerstone ocean bottom node (OBN) program looks at using the well-known benefits of OBN data; full azimuths, long offsets and rich low frequencies, to provide a step change in imaging of this important region of the North Sea. This is achieved through improved model building, in particular the detail unlocked by full waveform inversion using the latest Time Lag cost function ( Zhang et al.,2018 ).Utilizing TL-FWI on this OBN data aimed at improving the entire section of the velocity model: the complex overburden, intra chalk and sub-chalk layers. In addition to the added illumination achieved from OBN data, the use of the multiples, further illuminates areas of the subsurface not captured in the primary wavefield.

Full waveform inversion (FWI) using diving waves has in recent years become a standard model-building tool in the Central North Sea (CNS). Below the maximum depth of diving wave penetration however, we have remained reliant on ray based tomographic methods, which although powerful, have many limitations.In this paper, we will describe the use of time-lag FWI (TLFWI). TLFWI uses a modified cost function which aims to minimize travel time differences between recorded and modelled shots. This mitigates many of the issues encountered in other FWI solutions when reflection information is included. The ability to reliably use reflection information make it possible to reduce our dependence on ray-based approaches at depth. This is of particular benefit in areas where these ray-based tomography methods are inherently limited such as fast, layered chalk layers. The application of this technology to a large survey in the Central North Sea will be described.

There have been many advances in natural language processing in recent years but most of the work have been focused on texts from a general domain or medicine and so datasets in the geology domain are sadly lacking. We demonstrate how existing taxonomy and geological texts can be used to address this issue and also show how named entity recognition and object detection can be used to retrieve information from a large number of documents.

The Greater Castberg survey was acquired in 2019 using a source-over-spread acquisition design with an additional source at the front of the streamers, towed by the receiver boat, to permit recording of longer offsets. Starting from an initial anisotropic model, time-lag full waveform inversion (TLFWI) was used to compute a 13 Hz velocity model over the 5000 km2 area, which led to a much improved migration image (e.g., better imaged reservoir flat spots). When pushing the inversion frequency up to 90 Hz, the resultant TLFWI velocity model enabled more detailed delineation of reservoir boundaries compared to the migration image. Furthermore, the FWI Image as an alternative view of the FWI velocity model, provided access to new reflectivity information, overcoming some limitations of current migration tools. The use of the full wavefield, including refraction, reflection of primary, multiple and ghost, greatly enhanced image without needing the complex data pre-processing required by conventional imaging. In the context of thick gas clouds, this new imaging technique provided accurate sub-gas reflectivity, which effectively enhanced event continuity compared to reverse time migration (RTM) results.

Offshore Gabon has several challenges for seismic exploration. The major difficulty is obtaining a detailed and accurate velocity model. Complicated salt structures with overhangs, variability within the salt and carbonate rafts with Karst features pose difficult challenges to conventional velocity model building. Moreover, these introduce illumination issues which are not addressed by traditional migration. This is particularly important in the pre-salt areas where the target generally lies. With this paper we show how the recent technological advances in velocity model building through Full Waveform Inversion (FWI) and particularly Time-Lag FWI (TL-FWI), followed by imaging through the Least-Squares (LS) migration have provided a step change improvement in the pre-salt image of narrow azimuth data acquired in the West African Atlantic margin of Gabon. To overcome cycle-skipping and amplitude discrepancy between synthetic and recorded data in the presence of sharp velocity contrasts and large scale geo-bodies, typical for Gabon context, we propose the use of a robust FWI cost function like the one employed in TL-FWI. The alignment of the resultant estimated velocity model with the geology of the margin allows the generation of considerable imaging uplifts. Additionally, the LS migration mitigates the noise generated by the migration operator and illumination deficiencies.

We discuss the processing and imaging challenges relating to a marine seismic survey acquired northwest of the Shetland Islands. The proximity of the survey to the islands forced the acquisition direction to be strike to the subsurface geology. A shooting vessel provided wide azimuth data to illuminate in the dip direction, and triple sources increased crossline sampling. We show how 3D source designature and source/receiver deghosting were required for this wide azimuth data. For demultiple we illustrate how utilizing narrow azimuth data in water layer multiple modelling was crucial for the wide azimuth multiple prediction. In addition we highlight the benefits of a wave equation deconvolution based approach in this shallow water setting. For velocity model building we show how using time-lag full waveform inversion helped resolve the specific challenges created by the geological setting. Finally we demonstrate how that imaging was aided by the use of least squares migration and the inclusion of the wide azimuth aspect.