Search

Geological sequestration of CO2 is considered to be an important greenhouse gas mitigation technology (Washington et al. 2009). Current public resistance to onshore sequestration sites makes offshore sites an attractive alternative. The Sleipner gas field in the North Sea is such a site and has been the worlds first industrial scale CO2 storage project. An effective CO2 storage requires monitoring and verification over large areas. Time lapse seismic monitoring is an integral tool to do so. Here we use time lapse elastic 2D full wave form inversion (FWI) to monitor CO2 in the Sleipner aquifer reservoir and retrieve the amount of free CO2. Inverted P-wave velocities (Vp) are related to gas saturations using a fluid substitution model employing Gassmann theory assuming patchy saturation. We find that a considerable amount of the injected CO2 could be in dissolved form by 2006 and hence saline aquifers could be a natural reservoir for CO2 storage.

Full waveform inversion is increasing being used in oil industry to quantify P and S-wave velocities of the sub-surface. So far, waveform inversion methods have been either 3D acoustic or 2D elastic. Furthermore, only low frequencies have been used because of the high computation cost. The application of elastic full waveform for 3D reservoir monitoring is still beyond the reach of present day computing technology. Here, we demonstrate the application an injected grid method, where the forward modelling for iterative elastic waveform inversion is performed in a small volume surrounding the reservoir, reducing the cost significant for full waveform. The application of algorithm is demonstrated on Marmousi II data set in a time-lapse mode.

Since the 1980’s a significant amount of research in the field of full waveform has been conducted, and we are now at the stage where more accurate descriptions of the subsurface are being sought in addition to the elastic parameters. Elasticity accounts for materials that have a capacity to store mechanical energy with no dissipation of energy. A more accurate description of real earth media takes into account the time-dependent nature of previous stress states. In other words, the material possess a characteristic referred to as a memory effect (Robertsson et al. 1994).

E014 CO2 Thin Beds Imaging Using Full Waveform Inversion–Applications to Synthetic and Real Time–Lapse Data A. Gosselet* (Institut de Physique du Globe de Paris (IPGP)) & S.C. Singh (Institut de Physique du Globe de Paris) SUMMARY Global warming attributed to the greenhouse gas effect has generated growing interest in geological sequestration of CO2 over the past decade. Seismic monitoring is a key process in the strategy to reduce atmospheric CO2 emissions from a security standpoint and for sequestration optimization itself. Application of time domain full wave form inversion is proposed to estimate subsurface elastic properties using time-lapse data. The method

C031 Elastic Full Waveform Inversion of Multi- Component OBC Seismic Data – Application to Alba Field UK Central North Sea T.J. Sears* (University of Cambridge) S.C. Singh (Institut de Physique du Globe de Paris) & P.J. Barton (University of Cambridge) SUMMARY Elastic full waveform inversion of multi-component ocean-bottom cable seismic data has previously been applied to a range of synthetic examples in both one- and two-dimensions demonstrating the ability to quantify both P- and S-wave velocities in depth. Here results from application to a real dataset from the Alba Field UK Central North Sea are presented. Data were inverted in

Elastic full waveform inversion of seismic reflection data is an alternative to conventional stack based processing. Both travel-time and amplitude information within shot domain data are used to invert for the sub-surface velocity, from a starting model containing the long wavelength features. Although full waveform inversion is computationally intense, the ability to invert for both P- and S-wave velocity, in a data driven manner, makes the computational burden worthwhile. Building on previous developments in waveform inversion, such as the benefit of wide-aperture data, we consider multi-component ocean-bottom cable seismic data. Synthetic examples show that it is possible to invert for and to quantify both P- and S-wave velocity. Example synthetic work is presented, followed by a summary of work in progress on real data. Inversion for the shallow sub-seabed region is of particular interest. Successful inversion for the shallow region may be used to help solve shear-wave static problems observed in OBC data. This idea is explored in a synthetic example, and acts as an objective for real data application, with results to be presented.

Elastic full waveform inversion of seismic reflection data represents an alternative to conventional stack based processing. Both travel-time and amplitude information within shot domain data are used to invert for the subsurface velocity, from a starting model containing the long wavelength features.

OBC technology provides information on both pressure and 3 components particle velocity recorded at the seafloor level, allowing complementary PP and PS imaging. In the following paper, we describe a new elastic decomposition scheme in the tau-p domain that separates upgoing-downgoing P and S wavefields and eliminates all water layer related multiples. We introduce frequency dependant operators to calibrate the geophones and to characterize the P wave reflection coefficient at the seafloor. The process does not require any a priori knowledge of seafloor parameters. We start by describing theoretical aspects for complete decomposition. Then, we detail a multi-step procedure for estimating the shear wave velocity at the seafloor, for extracting the upgoing S wavefield and finally for removing peg-leg arrivals. We show successful tests on a real dataset. The results suggest promising avenues for further processing.

P068 Full waveform inversion of multi-component OBC seismic data Abstract 1 Advances in computer technology have made full waveform inversion a more realistic and achievable process. In addition to overcoming the computational challenges involved the inverse problem itself has been helped through different techniques. The use of long-offset data and preliminary inversion of lower frequencies are seen to mitigate problems of nonlinearity. Medium wavelength features thought missing from near offset reflection data may then be resolved and convergence to a unique solution is made possible. However while the inversion of compressional (P) -wave velocity from marine streamer data is more

A16 DATA-DRIVEN P-S SEPARATION IN OBC DATA EDME P. 1 SOUDANI M.T.A. 2 3 BOELLE J.L. 2 SINGH S.C 1 . Abstract 1 OBC technology provides information on both pressure and particle velocity recorded at the seafloor level allowing complementary PP and PS imaging. In the following paper we describe a new elastic decomposition scheme that separates upgoing-downgoing P and S wavefields in the��-p domain. We introduce frequency dependant operators to calibrate the geophones so that the process does not require any a priori knowledge of seafloor parameters. We start by describing theoretical aspects for complete decomposition. Then we detail

According to ray theory, the traveltime T(R,S) between any source S and receiver R is related to the slowness distribution ŋ at any point X by the integral: T(R,S) = ∫ г(ŋ) ŋ(X)dl, where г(ŋ) is the ray path which yields the minimum traveltime and dl is the arc length. Because the ray path г(ŋ) depends upon the actual slowness ŋ, the inverse problem consisting of recovering the slowness distribution from the traveltime T(R,S) is non linear. A common approach is to linearize the traveltime with respect to a known slowness and iteratively update the current estimate. This requires ray tracing to compute the traveltime misfit.

Seismic reflection profiles across many continent al margins have imaged "bottom simulating reflectors" (BSRs), which are interpreted as marking the base of a methane hydrate stability field. BSRs often have very high amplitude and can be continuous for many kilometres, so they provide an ideal target for a waveform inversion scheme. Knowledge of the detailed velocity structure in the vicinity of the BSR allows us to determine whether it is underlain by a free gas zone, and estimate the thickness of such a zone where present.

Deep seismic reflection profiles show that the lower part of the continental crust often contains many strong sub-horizontal reflections. These reflections are interpreted as a sequence of finely layered structuren. To confirm or invalidate this interpretation, we can use a seismic waveform inversion to estimate the reflectivity in the lower crust. However, it is very difficult to invert for the whole data set, partly because of large computation time in modelling and partly because the reflected energy in the upper crost dominates the misfit function.

Seismic waveform inversion consists of minimizing misfit between observed and calculated wavefields. Since this misfit function could be highly non-linear, the only way to be sure of arriving at the global solution is to use a Monte-Carlo search. This is feasible if the number of unknown parameters is not too large. If it is large, as is the case with the seismic inverse problems, it is extremely expensive to use the Monte-Carlo search to optimize for a single misfit function over the whole model space. However, if one uses a number of misfit functions depending up on the sensitivity of the model parameters on the data, one can implement the Monte-Carlo method efficiently. Fortunately, this is possible as the different wavelengths of the model parameters influence the different parts of the data.