Higher Resolution Seismic Imaging using Holography

Abstract

Holographic Imaging of the earth subsurface using seismic survey data produces substantially greater spatial and time resolution than may be obtained by traditional signal processing methods. For this reason such imaging is often called “Highest Possible Resolution” or “High Definition” Imaging. This new Paradigm for holistic seismic imaging is based on waveform interference patterns. Migration as we know it is wave-based and attempts to reconstruct wavefields at any point within the earth using approximations to the wave equation. The reconstructed wavefield is a migrated section. In our case, it is important to distinguish between wavefield reconstruction and the imaging process. Wavefield reconstruction is the determination of the wavefield over an area of interest. Imaging on the other hand, creates a picture of the reflecting surfaces within an area of interest. The input to the migration process is the seismic data, sometimes referred to as a hologram. This takes into account all the recorded signals received at the earth’s surface. The first step of migration process is wavefield reconstruction. The second step is imaging. Attainable resolution using Holographic Imaging techniques represents the highest possible values with limits imposed by the imaging approximations, estimated propagation velocities, noise, and the geologic character. Such character relates to the sediment deposition and its energy, having specific expression in terms of variations over the effective Fresnel Zone and vertical grading or transitioning of the lithologies. For such imaging, as typical of Holographic methods, illumination bandwidth is largely incidental, and frequencies in the image domain may range between 3 and 10 times the usual input bandwidths, and possibly greater, as geology ultimately permits. This paper presents advantages and merits of Holographic Imaging by examples and comparisons supported by the modeling. One of the other advantages in having broader bandwidths for both frequency and wavenumber is seen in the detailed velocity analyses. High density and high definition velocity analyses enables one after imaging, to see subtle stratigaphic details and fault planes and also in the deeper seismic data, making it easier and even practical to interpret the tectonic style and stratigraphy. 3811 SEG Denver 2010 Annual Meeting © 2010 SEG Availability of more computing power now places us in a strong position to expand innovative geophysical technologies that will be both cost effective and provide better seismic interpretation. The greater resolutions place heavy demands on traditional display capabilities which fall short. Considerations here are necessarily then directed to extending visual dynamic range via specially designed color inversion data presentations. Holographic Seismic Imaging: When we view a holographic image we do not ask about the characteristics of the illumination. The image is as it is; the illumination is incidental. This is not the case with a seismic image. Dr Eisner speaks to this issue in his contribution (see References). A forum in Houston in 1977, “The Stationary Convolutional Model”, firmly established as a model for each trace of a processed seismic image, a reflection coefficient series convolved with an unchanging propagating wavelet. A noise term was added to further represent unwanted signal returns and noise. According to this model the frequency content of the wavelet and the trace spacing limited the image frequency and wavenumber content, and hence the attainable resolution. This viewpoint prevails today even for circumstances where it need not apply. Mechanics of seismic acquisition as practiced, and much of the subsequent processing are in fact holographic. They employ focusing of energy via imaging curves related to wave propagation in a subsurface using some information about geometries and velocities. Diffraction imaging is employed by holographic methods rather than relying principally on reflections. We note Figure 1, a surface seismic survey as indicated would not normally image the vertical boundary between lithologies A and B. Detection of this boundary would involve diffraction events. Diffractions however, are not an innate property of a subsurface. In this same Figure 1, a source and receiver in the indicated borehole would see this same vertical boundary as a reflection. Diffraction imaging is the more general in nature. This viewpoint is discussed by Professor E. A Robinson (see References ) and illustrated by a model calculation shown in Figures 6 and 7 or as discussed in Part 3 of the Neidell series of papers “ Perception in Seismic Imaging”. We will discuss this shortly. Figure 2 shows a seismic image of a salt dome with 110ft x 110ft bins spatial sampling. The High Definition image in Figure 3 was processed at a 1ms sample rate, and with 27.5ft x 27.5ft bins. From the wiggle trace displays one can fully appreciate the differences both from an image point of view and a resolution point of view. A corollary to Huygens’ principal states that the information contained within a recorded seismic wavefield does not depend simply on the wavefield sampling and source properties, but on the information contained within the wavefield and that the geologic structure can be considered as composed of many diffraction points whilst the resulting seismic section is the superposition of all the diffraction curves (Eisner, 1998.) Figures 4 summarizes the main points that relate to the traditional seismic processing 3812 SEG Denver 2010 Annual Meeting © 2010 SEG view of how we handle seismic image formation as compared to the non-traditional holistic approach. Figure 5 shows a portion of a depth model that is used for a wave field simulation. We see 10, 20, 30, 40,and 50 ft square indentations. The model response (1 ft traces and 1msec samples) from the wave equation calculation along with the interpretation is shown in Figure 6. The model represents a noise free seismic survey traditionally imaged. Using Holistic imaging now we input only every 10 trace and every 10 time sample (one hundredth the data first used).. One can see in Figure 7 that the model response based on the wavefield imaging reconstruction and interpretation look almost identical to the one derived originally from the full data set using standard methods. Figure 8 and 9 compare a standard imaging of a marine profile from the Bonaparte Basin, Offshore Australia. The trace interval is 18.75m with a 4 ms time sampling for the “production” section, and 4.69m, 1 ms sampling for the High Definition section. The comparison even at this larger scale is quite remarkable. The HiDef Image shows steeply dipping events such as small faults and possible wrench or fluid flow features. The close-up comparison of just one interesting feature presented in Figures 10 and 11 makes the same points as previously emphasized. In general then the better imaging will lead to more detailed interpretations and improved understanding. As a corollary, all attribute calculations and data analyses achieve better performance with higher resolution data. Up to this point our main focus has related largely to the improvements on the imaging. Higher Resolution Velocity Analysis: Typically, moveout velocity analysis is often presented in the form of a velocity scan or as a contoured display. It provides velocity information as an imaging parameter, and semi-quantitative information about the wavefield character and general coherence of the events it identifies. Figure 12 shows comparisons of a velocity analysis for the production data after PSTM, and after the Holographic Imaging. The Holographic result shows better defined events having more contour closures, as well as more deep events. Hence, these higher frequency events must represent real subsurface returns because they have a necessary wavefield character – moveout. Extended Visual Dynamic Range Seismic Displays : In general, usual seismic data displays are rapidly overwhelmed by the information content of these holographic imaging methods. To attain maximum benefit from such processing, amplitudes are scaled as velocities (a simple inversion) using an approximate low frequency trend developed only from the moveout derived velocities. These velocity values are displayed using an Extended Visual Dynamic Color Display. Such displays present 5 times the information of typical color displays and 25 times more than black and white data presentations. Typical velocity increments used might 3813 SEG Denver 2010 Annual Meeting © 2010 SEG be 400 ft/sec. Such displays aid significantly in recognizing lithology, geopressure, porosity and possible hydrocarbons, particularly in high velocity or consolidated lithologic circumstances (including carbonates). One data set is a small 3D survey over an onshore reef platform at depths greater than 15,000 ft. The views offered by the Imaging from time slices and vertical sections offer remarkable direct insights into the development of the reefs on the platform, their hydrocarbon potential, the cyclic development of the carbonate lithology, shales and anhydrites, and the correlation with the well control (Figure 13 and 14) In order to focus on the desired information new display formats have been developed as noted which almost directly map objectives and facilitate volumetric estimates. One of these displays highlights those zones most likely to be gas reservoirs. Seismic Images (time slices) for the reef study look remarkably like modern aerial photographs over reef complexes. The study is significant in that it readily identifies and explains a high volume producing gas well and also non-commercial borehole.