An azimuth dependant processing pilot study was carried out in a large Middle East Field to evaluate if this technology has the potential to successfully identify fracture permeability pathways. The field is heavily faulted and fractured with good well control and therefore is a good candidate to perform this study. The success criteria for the Azimuthal processing are:• Improved fault imaging relative to the available conventional processed seismic volume; • Obtain information about seismic anisotropy in the reservoir zones.This anisotropy will be linked in a full evaluation to fault & fracture density and orientation. The anisotropy can be measured via differences in seismic travel times or amplitudes / seismic attributes measured in the different azimuth seismic cubes. Azimuthal anisotropy from a 3D land seismic dataset acquired in the U.A.E. has been analyzed using wide azimuth processing. Two different processing methods and flows were tested to derive optimum processed volumes. In both methods raw CMP gathers, after convolution, residual statics, and inter-bed multiple elimination were used as input data for the azimuth stack processing sequence. The two methods are • Azimuth Sectoring • Common Cartesian Offset Bins (CCOB)Both processing methods have their benefits, one big advantage of CCOB is that you can stack very fast different individual azimuths together and get a sharper image, which results in better interpretation. Azimuth sectors both parallel and perpendicular to the three major fault system orientations, were imaged separately to produce the six final azimuth volumes. Comparisons between the different azimuth sectors were used to detect azimuthal differences in velocities and amplitudes that could be correlated with fault and fracture orientation and magnitude.The interpretation and validation of the results suggest that value is maximized by integrating multiple attributes that include horizon mapping for time differences, amplitude extractions for reflectivity differences and result validations with available well calibration. The azimuth sector results have aided in the quantification of fault presence, magnitude of throw and suggests that fractured zones can be identified which may indicate higher permeability pathways within the reservoir. Another important learning from this case study is to use an integrated approach during processing and interpretation and don't look only at one single part, e.g. velocity cube.Overall the results of this carbonate Azimuthal Pilot for fault and fracture characterization has produced encouraging results and valuable lessons learned to aid future studies.

Multiples are successfully removed from seismic data to increase the signal-to-noise ratio in a 3D land dataset acquired in Abu Dhabi. The process aims at removing in user specified spatial-temporal windows only that multiple energy (surface or inter-bed), interfering with particular primary seismic events. The anti-multiple is designed in the f-xy domain and applied prestack, on constant offset time migrated data volumes. Subtraction of the multiple events from the original seismic data is shown to produce an improved subsurface seismic image that is more suitable for interpretation and attribute analysis. The ability to remove multiple events without affecting the primary signal is a crucial issue in seismic processing, as spurious energy at a target level leads to suboptimal images and adds uncertainty to reservoir characterization. The state of the art in attenuating multiple arrivals involves a two steps process. First the multiples are predicted via a data driven technique. Then the predicted events are matched to the multiples actually present in the data and removed with some matching filter. The removal assumes that the kinematics of the multiple events are correctly derived and that the prediction locates them correctly in time and space. The technique applied on the Abu Dhabi dataset focuses on a particular family of reverberating events that invade the target area. This is a marked difference with respect to general techniques that try to predict the whole multiple wavefield. The technique requires the strong impedance contrast that generates the multiple (surface related or internal) to be reasonably flat. When this requirement is met the prediction of the multiple is simple and avoids all shortcomings of general techniques such as SRME with irregular acquisition geometries, poor signal to noise ratio or variable near surface effects. Moreover we believe that this requirement is often met in the Arabic Peninsula. Introduction The need for seismic processing to more effectively remove multiple energy noise while preserving at the same time the primary events is growing in important due to both structural interpretation and advance geophysical analyses that require higher quality seismic data. Multiples can corrupt the primary seismic events at the target and lead to incorrect seismic attributes and erroneous interpretations. Multiple attenuation as applied in this land seismic study was a two step process. In the first step the multiples, which include both surface and inter-bed related multiples are predicted using a data driven process that does not involve any prior information concerning the subsurface velocities. In the second step the predicted multiple events are matched to the multiple events present in the seismic data and removed using a derived matching filter. This approach assumes that the dominant multiple events in the zone of interest are correctly derived and that the prediction locates the multiples at their proper seismic time. The method is applied in a way that minimizes any possible interference between the primary seismic events and the multiples to preserve the primary seismic event amplitudes. Figure 1 compares a simple reflection, surface related multiple and interbed multiple to illustrate their differences. The seismic data used in this study has strong multiple energy that has similar moveout to the primary events and is difficult to remove without damaging the primaries. The multiple method used in this study did not require moveout differences between the primary and multiple events in order to be able to remove the multiples.

Multiple attenuation was successfully performed on land seismic data in the difficult Abu Dhabi sand dune and sabkha environment. The keys to success were careful analysis plus selection and application of carefully designed data-adaptive multiple attenuation workflows. The demultiple success on the South East Abu Dhabi seismic data described in this paper was the result of a unique workflow developed specifically for this survey. Development of the workflow depended on rigorous analysis and consistent feedback from interpreters and geophysics specialists. Multiples with similar prestack moveout with offset to the primary events were successfully modeled and removed from seismic data to produce higher quality final seismic images. Onshore seismic data from the high sand dune South-Eastern Abu Dhabi area suffers from strong ambient noise and multiples that make exploration efforts difficult. Although a variety of techniques and algorithms exist for modeling and removing multiples from seismic data, there is no single method that is appropriate for all cases. Successful projects for the most difficult onshore multiple problems are often the result of careful analysis, selection and application of the right technology and the use of an appropriate data-adaptive workflow. The demultiple success on the South East Abu Dhabi seismic data demonstrates that with the proper tools and feedback it is possible to developed a workflow that can attenuate even the most difficult multiple noise. Introduction Onshore seismic data from the high sand dune South-Eastern Abu Dhabi area suffers from strong ambient noise and multiples that make exploration efforts difficult. Figure 1 is an elevation map of the study area with the large sand dunes shown in blue. The height of the dunes ranges from ~50–70m above MSL. During 2006 a multiple attenuation pilot was conducted over this approximately 300 sq km area. The objectives included multiple attenuation and improving the pre-stack and post-stack signal-to-noise ratio such that the data would be appropriate for AVO and reservoir characterization. The results from the multiple attenuation pilot demonstrated that both pre and poststack seismic data quality can be improved sufficiently to satisfying the pilot study objectives. The modern toolkit for multiple removal from pre-stack seismic data contains a variety of techniques designed to use various attributes of the multiples, such as their differential moveout, predictability, etc. Despite decades of research, no single algorithm or technique emerged to solve all of our multiple problems. Successful projects, especially for the most difficult multiple problems are the result of careful data analysis and the appropriate choice of algorithms by experienced geophysicists. Often the best results arise from a carefully designed workflow of data preconditioning and several multiple attenuation steps. Orthogonal geometry data sets, particularly those from high ambient noise environments, present one of the more difficult demultiple problems, at least for pre-stack data. Orthogonal geometries used for more economically feasible survey designs often provides poor or irregular sampling in one or more data domains.

No abstract