LAUSR

Multicomponent seismic data analysis enhances confidence in interpretation by providing mode-converted PS data for imaging the subsurface. Integrated interpretation of PP and PS data begins with the identification of reflections corresponding to similar geologic events on both data sets. This identification is accomplished by carrying out well-log correlation through the generation of PP and PS synthetic seismograms. There are a few issues associated with the approach. One of the issues is that PS data have lower resolution than PP data. This presents difficulties in the correlation of equivalent reflection events on both data sets. Even if few consistent horizons are tracked, the horizon-matching process introduces artifacts on the PS data mapped in PP time. In this paper, we elaborate on such challenges with a data set from the Anadarko Basin in the United States. We then propose a novel workflow to address the challenges. Introduction Multicomponent seismic data analysis enhances confidence in interpretation by providing mode-converted PS data for imaging the subsurface. For compressional energy that is incident on a rock interface at an angle different from normal incidence, partitioning takes place in transmitted and reflected compressional (P) and shear (S) components. Mode-converted shear (PS) energy is primarily recorded on the radial component of three-component receivers. Because these waves follow different travel paths with different wavelengths, they “see” the subsurface differently than P-waves. Consequently, PS seismic data exhibit significant changes in amplitude and character of seismic events that may not be seen on PP seismic data. When PP and PS data are analyzed together, more confident interpretation takes place, yielding important rock property estimates such as VP /VS. Because VP values exhibit overlap in different rock types, the added value that VS brings makes VP /VS an important parameter. It was established early on that this ratio is a good lithology indicator (Tatham and Stoffa, 1976; Tatham, 1982; Pardus et al., 1990), is good at identifying limestone/shale boundaries (Goldberg and Gant, 1988) and sand/ shale ratios in channels (Garotta et al., 1985), and is sensitive to fluids (Dillon et al., 2003) and porosity (Pigott et al., 1990). Such applications have continued to be demonstrated in the literature, but it is important to note that they can only be successful if the determined values of VP /VS are accurate. Conventional vertical-component seismic data allow the estimation of mode-converted shear at oblique angles of incidence, which after simultaneous inversion, yields VP /VS. Multicomponent seismic data provide the actual recorded horizontal radial component, which is what we refer to as “PS data.” PS data are expected Satinder Chopra and Ritesh Kumar Sharma to provide significant information to PP seismic data interpretation, but the differences in timescales prevent an easy visual comparison between them. To accomplish this, PS data are converted to PP two-way traveltime. Once this is done, not only can the coupled interpretation of PP and PS data be carried out, but both types of data can be put through an integrated or joint impedance inversion, yielding VP /VS data. Integrated interpretation of PP and PS data begins with the identification of reflections corresponding to similar geologic events on both data sets. This identification is accomplished by carrying out well-log correlation through the generation of PP and PS synthetic seismograms. If check shots or vertical seismic profile data are not available, slight stretching/squeezing may be necessary. One way to generate a PS synthetic seismogram is to use VS and density curves to generate a PS elastic gather with a wavelet extracted from PS stacked data and then stack it. The stacked gather trace can be correlated with the PS stacked data. The other method is to generate angle-dependent PS reflectivity at 10° or 12° and use it to generate the PS synthetic seismogram. It is assumed here that a shear sonic curve is available, and both synthetic seismograms are generated over the same range of frequency bandwidth as the input seismic reflection data. Such a correlation helps with the visual identification of events on PS sections at the location of the well, considering their character, relative amplitudes, and approximate traveltimes. In a similar way, reflection events are identified on PP data. When the events of interest are identified and correlated on the PP and PS sections (through respective synthetic seismograms) at the location of the wells, horizons are picked on the data volumes on an interpretation workstation. The peak amplitude on the PP and PS data represents an increase in elastic impedance across the interface that it is representing. Mis-ties may be seen on such synthetic-seismic correlations, and one should keep an open mind while analyzing the reasons. While performing event correlations between PP and PS data, a peak on PP data is expected to correlate with an equivalent peak on PS data. However, this is not always the case (e.g., the oil sands area in northern Alberta, Canada, where the Paleozoic marker is a difficult pick). The Paleozoic marker is a weathered unconformity between the Paleozoic carbonates and Cretaceous clastics. The relative compressibility and rigidity of the weathered carbonates as well as the tuning artifacts make the seismic response exhibit a peak at places, which becomes a zero crossing or a trough at others (Anderson and Larson, 2006). In such cases, a prominent horizon above or below can be used for horizon picking. Next, the equivalent correlative events on the PP and PS data volumes are used to map or shrink the PS timescale to the PP timescale, a process referred to as “registration.” This step has TGS, Calgary, Alberta, Canada. E-mail: [email protected]; [email protected]. https://doi.org/10.1190/tle39010047.1 D o w n lo a d e d 0 9 /0 4 /2 0 t o 7 0 .7 7 .2 1 4 .1 3 8 . R e d is tr ib u ti o n s u b je c t to S E G l ic e n s e o r c o p y ri g h t; s e e T e rm s o f U s e a t h tt p s :/ /l ib ra ry .s e g .o rg /p a g e /p o lic ie s /t e rm s D O I: 1 0 .1 1 9 0 /t le 3 9 0 1 0 0 4 7 .1