Geophysics is a branch of physics that is mainly concerned about under- standing and describing the physical behaviour and activities of the earth’s geological system. Usually, seismic data is acquired at the surface and the corresponding signals go through a sequence of preprocessing steps to filter out the noise and enhance the quality of the measurements. These mea- surements are then transformed into a so-called reflectivity image (snapshot in time) of the subsurface via the deployment of the so called migration al- gorithms. Extensive studies and great effort is usually made to determine a suitable acquisition geometry design for optimal illumination of every imaging grid point in the subsurface in the studied area. However, within the geo- energy industry it has always been an undisputed believe that improved and better results can be obtained only if more data is acquired with denser sam- pling at both the source’s and receiver’s side. This ”linear” way of thinking is also consistent with the conceptual assumptions of most current migration algorithms. The challenge is of course to achieve the same, or maybe even better, results with less data. This is in accordance with the currently ongoing en- ergy transition, which is forcing the geophysical scientific community to shift their focus from the acquisition and processing of large and expensive surface seismic surveys toward optimized management, in terms of data acquisition, processing and production, of the existing hydrocarbon-based reservoirs. Es- pecially datasets with sparse acquisition geometry like 3D Ocean Bottom Node (OBN) and 3D Borehole Seismic Data (BSD) surveys, where we have measurements at a limited number of sensors along the ocean bottom or in the borehole but usually with dense sources sampling at the surface, can greatly benefit from such a development. 3D borehole geophysics, which is the main subject of this thesis, has for a long time been an underdeveloped and, therefore, an unappreciated compo- nent within most geophysical organisations. This is mainly because accurate results are usually obtained only in the immediate vicinity of the borehole and their quality decays rapidly in the lateral extent. However, and especially in the marine case, 3D BSD surveys are rich in higher-order scatterings that can have significant added value when combined with unconventional and non- linear inversion-imaging algorithms like Full Wavefield Migration (FWM) and Joint Migration Inversion (JMI). Furthermore, in combination with modern measurements techniques (like Distributed Acoustic Sensing (DAS) technol- ogy), continuous and permanent monitoring of existing and new reservoirs – whether hydrocarbon-based or geothermal – can easily be realised. In this thesis the recently developed inversion-imaging algorithms FWM and JMI are extended to the 3D case and further engineered to properly han- dle the special acquisition geometry of 3D BSD surveys and exploit the full potential of the total wavefield available in 3D borehole seismic data. First, a more complete and comprehensive derivation of the involved gradients, for the reflectivity image and velocity model update, is presented. This makes the combination of one-way tomography of the direct wavefield with reflec- tion tomography of the other energy modes (primary reflections, higher-order scatterings of the up- and down-going wavefield) a straightforward process. Then, an effective strategy of the application of the 3D JMI algorithm to 3D BSD is developed and, with the presented examples, it will be demonstrated that, for instance, the standard and conventional separation of the up- and down-going wavefield of 3D BSD becomes an obsolete process. Along the same lines, we will show that integration of surface seismic data and 3D BSD, or even multi 3D BSD surveys, in one inversion process produces more accurate and geologically consistent solutions. Next, the capability of the 3D FWM algorithm together with 3D BSD surveys for solid reservoir monitoring will be demonstrated. After that the challenges of the current acoustic imple- mentation of the FWM JMI algorithms will be discussed, especially the effect of the mode converted waves on the velocity model gradient. Finally, some suggestions are made for further enhancement of the JMI algorithm, partic- ularly at the side of the migration velocities update. This can be achieved by the combination of complementary and effective objective-functions, which makes the JMI algorithm more robust especially in the case of geological environment with high velocity contrast.@en

3D Vertical Seismic Profile (VSP) surveys play an important role in seismic exploration and reservoir development. This is mainly due to the higher data-quality and broader frequency bandwidth relative to surface seismic data, which as a result provides higher-resolution image around the borehole. Conventional VSP imaging techniques mainly utilize the separated and deconvolved Pwave reflections which provide an excellent image around the borehole but the quality and the coverage degrade both laterally away from the borehole and spatially with increasing depth. Moreover, the final migration velocity model is commonly built from checkshots as the initial flatlayered model and updated using first-break travel-time tomography and slowness inversion; however, the update only covers the cone-shaped area around the borehole which the direct-wave passed through. The full 3D borehole signal, which includes all multiple scattering from both up- and down-going wavefields data, should be used to improve the subsurface illumination and to accurately update both the image and the velocity model away from the borehole. The Joint Migration Inversion (JMI) technique offers a comprehensive solution by comparing the modeled data to the measured data iteratively. The JMI forward modeling algorithm is based on the reflectivity and the propagation velocity in which it may provide an alternative to resolve our non-linear inversion problem faced in traditional Full Waveform Inversion (FWI). Consequently, it yields a true-amplitude reflectivity image with a relatively higher resolution and larger lateral extent as well as a smooth updated background velocity model. In this paper, we apply the 3D JMI to 3D borehole field data from the Arabian Gulf of Saudi Arabia.


@en

3D borehole-related seismic data can provide us with accurate velocity models and high-resolution images. However, conventional imaging algorithms that assume primary reflection energy will cover only a limited areaaround the borehole. This illumination problem can be overcome by including surface-related and internal multiples in the imaging algorithm to enhance the illumination of the subsurface and image areas that are beyond the coverage of primary reflections. The industry standard for 3D borehole imaging is to use the first breaks to update the velocity model and consider only the primary reflections for imaging. It has been demonstrated that multiple scattering and the down-going wavefield can tremendously extend the imaging area around the borehole. Moreover, primary and multiple scattering contain valuable and crucial information that
can enable us to update the velocity model beyond the coverage area of the first arrivals. In this paper we have employed the full wavefield of the 3D borehole data - containing all order scattering in both up/ down-going wavefields - in one integrated inversion-imaging process as proposed by the joint migration inversion methodology. The final result is a smooth accurate background velocity model along with a true amplitude reflectivity high-resolution image with and maximum lateral extent.@en

3D borehole-related seismic data has superior quality and higher-frequency content compared to surface seismic data. These unique properties make it possible to produce high-resolution images and accurate velocity models especially around the borehole. However, using conventional imaging algorithms, that assume primary reflection energy, will retrieve only a limited area around the borehole. This problem can be overcome by including surface-related and internal multiples in the imaging algorithm to enhance the illumination of the. In addition, on-the-fly the velocity model can be updated using the so-called Joint Migration Inversion process, which explains the full wavefield seismic data in terms of reflectivity and a propagation velocity model. To augment the results, datasets from different wells in the area can reinforce each other by simultaneous inversion to assure the consistency and improve the quality of the results. To steer and constrain the velocity estimation, the estimated reflectivity in the JMI process can be used as additional constraint for the velocity updating process. In this paper we have deployed the full wavefield of the 3D borehole data, from two different wells, containing all orders scattering, both up- and down-going wavefields, in one integrated inversion-imaging process as proposed by the JMI methodology. The final result is a smooth accurate background velocity model along with a true amplitude reflectivity image with high resolution and maximum lateral extent. @en

Surface-related and internal multiples will enhance the illumination of the subsurface and enable us to image areas that are beyond the coverage of primary reflections, especially in the case of datasets with a sparse acquisition geometry like 3D Borehole-related seismic data. In addition, shadow zones and acquisition geometry footprints, like what we encounter in 3D surface seismic, can be overcome by including the multiple scattering in the imaging process. In conventional imaging, the input data is first linearized, by suppressing the multiples, to make it suitable for linear migration algorithms and the associated velocity models are updated accordingly. However, it has been demonstrated that the nonlinearity in the data adds extra sensitivity to erroneous propagation operators. This unique property puts additional constraints on the possible solutions for the velocity model and will help expediting and steering inversion algorithms toward more reliable estimates. Furthermore, nonlinear migration algorithms must have a closed-loop architecture to ensure consistency between the input data and the estimated model parameters. In this paper we propose to deploy both surface seismic and 3D VSP data simultaneously in one joint migration inversion process, such that-at least within an area around the well location-an optimum high-resolution image and a better background velocity model is obtained and the lateral continuity and consistency is guaranteed. Furthermore, by including all multiple scattering in the inversion process - as proposed by joint migration inversion- a smooth velocity model will be obtained along with a true amplitude reflectivity image with high resolution.

@en