Introduction to this special section—AVO

Document Sample
Introduction to this special section—AVO Powered By Docstoc
					   Introduction to this special section—AVO
   AVO (amplitude variation with offset) analysis was for- applied to identifying subcrop reservoirs associated with other
   merly introduced by Bill Ostrander (in a presentation at the gorges in Sacramento Basin as well as in the search for hydro-
   1982 SEG Annual Meeting), for confirmation of bright spots carbons trapped beneath submarine canyons in deepwaters
   and other anomalous reflections seen on RAP (relative-ampli- basin worldwide.
   tude preserved) sections. Since then, we have come a long way             AVO anomalies are usually located by crossplotting a size-
   in extracting more information from prestack seismic ampli- able window of data in the far-angle versus near-angle or AVO
   tudes. Afew crucial points that have emerged in all these analy- gradient versus intercept data planes and then looking for clus-
   ses are: (1) an understanding of the assumptions underlying ters of points that stand out from the background majority of
   the theory being put to use and how well these assumptions points. Such a process is repeated for the zones of interest in
   are met in practice, (2) the realization that AVO reliability the survey under investigation. Swan (“Automatic compen-
   strongly depends on the quality of the processing, and (3) sation of AVO background drift”) introduces a novel idea of
   awareness of the importance that correct interpretation tech- processing the near and far or gradient and intercept data, so
   niques be employed to extract meaningful information from that the background majority always lies within a predefined
   the AVO attributes.                                                   region in the data plane. AVO anomalies can then be detected
        If the processing and analysis effort is not sincere enough, by examining data at fixed locations that are separated from
   AVO can very well stand for all very obscure, and then one can the fixed background region. In this way, large quantities of
   forget about benefitting from it. On the other hand, if we try data can be quickly examined for AVO anomalies and so the
   and understand the phenomena that distort amplitudes in the method is suitable for reconnaissance AVO.
   prestack domain and make all possible efforts to estimate and             Prestack seismic gathers that are NMO-corrected (using
   remove or compensate for energy losses that seismic waves analytic curves) may not result in flat gathers as required for
   suffer, or reduce noise contamination and other related effects, AVO analysis. There may be different reasons for this, but if
   we can make an important difference.                                  gathers are not flat, AVO analysis could lead to misleading
        Over the last few years, newer technology and/or tech- results. Gulunay et al. (“Gather flattening”) illustrate a method
   niques have evolved which allow bet-                                                          of flattening such gathers, which they
   ter extraction of useful information            ... if we try and understand the phenomena refer to as a brute-force type method.
   from prestack data. These include                     that distort amplitudes in the prestack Their method utilizes event tracking
   improved imaging and noise sup-               domain and make all possible efforts to esti- with a two-trace cross-correlation
   pression, relying on less simplistic            mate and remove or compensate for energy algorithm and moveout functions that
   assumptions, and using more sophis-              losses that seismic waves suffer, or reduce can be smoothed spatially as well as
   ticated analysis methods. A continu- noise contamination and other related effects, over time before applying corrections
   ing objective of research and                          we can make an important difference. to the gathers. Spatial smoothing of
   technology development is for AVO to                                                          moveout functions is done in the inline
   mean all very obvious. Given the vari-                                                        and crossline directions. The authors
   ability of data quality, processing streams, and analysis meth- demonstrate consistent flat gathers after application of this
   ods, a continuing challenge will be to properly ascertain the approach.
   applicability of the method in specific localities and in quan-           For 2D surface seismic data (gathers), the fold and offset
   titatively incorporating AVO results into risk assessment and are usually found to have a one-to-one correlation. But in 3D
   probabilistic reserves assessment.                                    CDP gathers, the fold is generally low for near and far offsets
        This special section provides a snapshot of the current state- and traces, while intermediate offsets dominate the overall fold.
   of-the-art in the application in AVO analysis. The 11 papers In 3D AVO analysis, evenly offset-spaced gathers generated
   cover a variety of topics from case studies to application of from supergathers and partial offset/angle stacks are com-
   newer ideas, and demonstrate that the AVO technology enve- monly used for AVO inversion. Xu and Chopra (“Benefiting
   lope is indeed still being vigorously expanded.                       from 3D AVO by using adaptive supergathers”) show that
        Young and Tatham (“Fluid discrimination of poststack doing this will lower the reliability of AVO inversion for 3D
   bright spots using lambda-mu-rho (LMR) inversion in the AVO. They present an approach wherein adaptive super-
   Columbus Basin, offshore Trinidad”) discuss an interesting gathers are generated to overcome the effect of 3D foldage.
   case study wherein the evaluation of bright spot targets is car- The real data examples that they show demonstrates the ben-
   ried out using the LMR attribute application and lowering risk efit of an adaptive supergather approach for improving the
   in drilling prospects. The gas sands exhibiting bright spots in reliability of AVO inversion.
   the Columbus Basin have lower acoustic impedance than the                 Offset-dependent tuning for thin beds and wavelet stretch-
   sealing shales, exhibiting class 3 AVO anomalies. Based on the ing due to NMO correction of seismic gathers causes prob-
   LMR analysis, only nine of the 15 prospective intervals eval- lems in AVO analysis. Many studies have been carried out for
   uated in this study proved gas-saturated and the other six were an analytical understanding of NMO stretching and offset-
   “false bright spots.”                                                 dependent tuning and their correction to improve AVO fidelity.
        May et al. (“Amplitude anomalies in a sequence strati- Xu and Chopra (“Improving AVO fidelity by NMO stretch-
   graphic framework: exploration successes and pitfalls in a sub- ing and offset dependent tuning corrections”) describe the
   gorge play, Sacramento Basin, California”) present a implementation of NMO stretching and thin-bed tuning cor-
   convincing AVO case study from Sacramento Basin, California, rections for production AVO analysis. Both synthetic and real
   where the authors claim 100% success in drilling subgorge data examples from Alberta show that these corrections are
   traps. The approach followed was to identify and understand necessary for reliable AVO analysis.
   the amplitude anomalies associated with the subgorge traps                Xu and Tsvankin (“A case study of azimuthal AVO analy-
   within their sequence stratigraphic framework, followed by sis with anisotropic spreading correction”) discuss the appli-
   recognizing the AVO signatures of gas-charged sandstones and cation of azimuthal moveout and AVO analysis on P-wave
   finally avoiding lithologic pitfalls. Such an approach can be reflection seismic data for fracture characterization of reser-

voir in the Rulison Field in Colorado. Their adopted process-     attributes and how in the study area the derivation of litho-
ing sequence includes advanced anisotropic travel time and        facies and clay volumes derived from PC2 attributes led to
amplitude inversion designed for wide-azimuth and long-off-       improved reservoir characterization.
set data. Prior to the azimuthal AVO analysis, to account for          Brown et al (“AVO monitoring of CO2 sequestration: A
amplitude distortions in the overburden, the authors applied      benchtop modeling study) demonstrate that AVO attributes
a moveout-based anisotropic spreading correction, which           could be used as discriminators for the presence of CO2. Using
helps in accurate estimation of reflection coefficients. The      laboratory and numerical experiments, the authors attempt
derived AVO attributes exhibit a pronounced character, so that    to upscale and predict the effects of small-scale heterogeneities
not only the gradient anomaly is well defined but the domi-       on AVO response at the seismic scale. Besides the presence of
nant fracture azimuth is in good agreement with EMI logs and      CO2, the authors suggest that it is possible to track changes
the direction of the fault systems, as also supported by geo-     in CO2 saturation over time.
logic evidence.                                                        Generally, AVO analysis utilizes Zoeppritz equations or
    Ren et al (“Spectra crossplot”) propose a spectra crossplot   their approximations for computing reflection coefficients,
technique for a quick and useful discrimination between gas       which in turn assumes that the incident wave impinging at
and wet reservoirs for AVO class 2 anomalies. The authors         an interface is a plane wave. Since incident waves originate
have done a detailed thin-bed signature analysis and notice       from point sources, it would be more appropriate to consider
that the maximum spectral amplitude difference between            a point source in a homogeneous material giving rise to a
water and gas-saturated reservoirs is near the tuning fre-        spherical wave. Ursenbach et al (“Efficient spherical-wave
quency of the thin-bed and that the maximum spectral ampli-       AVO modeling”) introduces the idea of spherical-wave AVO
tude difference increases with incident angle. They then          and present a way of calculating spherical-wave reflection coef-
demonstrate that on a crossplot between the near-offset and       ficients. Their analysis shows that spherical-wave effects are
far-offset spectral amplitudes (obtained by performing spec-      noticeable, not only near the surface, but even in deep reflec-
tral decomposition on near-offset and far-offset seismic sec-     tions if there is a critical angle in the data. Also, the spherical-
tions) around the peak frequency, data points with high slope     wave reflection coefficients are only weakly dependent on the
and high far amplitude values illuminate the gas reservoirs.      choice of the wavelet as long as the average frequency is spec-
    Singh (“Lithofacies detection through simultaneous inver-     ified correctly. We hope the readers find this section interest-
sion and principal component attributes”) demonstrates the        ing as well as informative. TLE
application of principal component analysis on attributes
derived from simultaneous inversion of seismic data from off-                                                 —SATINDER CHOPRA
shore peninsular Malaysia. Of the various attributes derived,                                   Arcis Corporation, Calgary, Canada
the author shows how flat spot features could be observed on
Poisson impedance and second principal component (PC2)                                                         —JOHN P. CASTAGNA
attributes, and not seen on acoustic and shear impedance                                                University of Houston, USA

                                                                                                 DECEMBER 2007    THE LEADING EDGE   1507

Shared By: