Converted-wave processing of 4C OBC data

Abstract

Abstract A four component ocean bottom cable (OBC) survey was acquired to investigate the use of using P and P-Sv converted wave AVO anomalies to predict lithology and fluid properties of the subsurface. Extracting lithology and fluid informationfrom P and P-Sv waves requires preservation of seismic amplitudes. We discuss a processing flow that attempts topreserve the amplitude information necessary for lithology prediction. Introduction The OBC data used in this abstract were collected by Geco Prakla during December of 1998 and shared by Burlington Resources Inc., Marathon Oil Company, Mobil Oil Company, Oryx Energy Company and Vastar Resources Inc.. The survey area is in the Gulf of Mexico where both fizz water and commercial deposits of natural gas exist. The objective of the survey is to use P-SV mode converted wave in conjunction with the P wave data to distinguish fizz water from the commercial gas reservoir. Processing of these data is an important step. Since the converted wave follows an asymmetric raypath, processing of the converted wave data is more difficult. This article describes the effort we made in processing the 4C data, especially the converted wave data. The key issue in the converted wave processing is to bin the traces to a certain gather so that (a) the data can be stacked to an image with good signal to noise ratio; (b) the amplitude versus offset information can be analyzed. We tested the two available converted wave binning techniques: asymptotic converted point (ACP) binning and common scatter point (CSP) binning. We found the CSP technique yielded a more satisfactory result. We will discuss the use of this technique. The combination of Vp and Vs information is considered an important factor for lithology prediction. We show a simple way to estimate P and S velocities from the P and P-Sv rms velocity analysis. The converted wave has a very different AVO response from the P wave and thus angle stacks and fluidline may not be directly applied to converted wave. AVO modeling based on the estimated velocity information should be used to guide the converted wave processing. We show an example of using P-Sv AVO modeling. There are a few articles regarding the OBC converted wave processing in the recent years (for examples, Thomsen, 1998; Yuan, et al., 1998). But each OBC geometry is designed uniquely for each region, therefore, the corresponding processing flow is unique for the goal defined in the above. We hope our current processing experience will contribute to the pursuit of the converted wave technology. Geometry and Data Figure 1 shows the geometry of one the four 2-D lines from this survey. This line consists of six adjacent cable positions with the first stations of the new cable position overlaying the last station of the previous cable position. The data are recorded at 12 second length and with 2 ms sampling rate. The sources are generated with 5400 cu in clustered air-gun array. Figure 2 shows a field data record with (a) time square energy composition and (b) automatic gain control (AGC). Analysis based on the criteria used by the Yuan et al. (1998) how that the X component energy is mainly due to P-Sv onverted-waves at the reflectors. This is confirmed by our AVO modeling which will be discussed in the later section of this article. Since P and P-Sv are well separated in the Z and X components for this particular data set, the P and P-Sv waves separation is not done. If there is mixture between the P and the converted waves, we hope the stacking will attenuate them in the final images. Processing Figure 3 shows the processing flowchart. Before the processing start, the common mid poi