Solutions to seismic resolution and characterization challenges in offshore Thailand

Abstract

Listen

Translate article

Abstract A high-density 3D survey was acquired during 2004 in the Gulf of Thailand with shallow source and streamer depths, a customized source array and adherence to strict spatial sampling criteria (Ramsden et al., 2004). Spectacular resolution improvements were achieved in comparison to legacy 3D data acquired in 1984 with coarser acquisition parameters and much larger source and streamer depths. Carter and Pambayuning (2009) observed, however, that the absence of very low frequencies in the shallow-tow 2004 dataset combined with noisy side lobe effects on extracted wavelets presented challenges to reservoir characterization. Careful reprocessing and merging of the 1984 and 2004 datasets into one dataset resulted in broader frequency bandwidth on the low end, a cleaner wavelet, and a better platform for high-end interpretation. Nevertheless, it is noted that source and receiver ghost effects greatly reduced the available low and high frequencies in both the 1984 and 2004 datasets, and thus on the merged result too. The ultimate scenario available now is that a single 3D survey can be efficiently acquired using technology that avoids the effects of both source-side and receiver-side ghosting, thus recovering strong low frequency and high frequency content, improves signal penetration, and yields remarkably clean pre-stack data for reservoir characterization and high-end interpretation. I introduce here the combination of dual-sensor streamers and an entirely new source technology that together yield frequency bandwidth recovery that is limited at the high frequency end only by the Nyquist frequency of the acquisition system and the effects of anelastic attenuation in the earth. Significantly, very low frequency amplitudes are simultaneously boosted by up to an order of magnitude in comparison to deep towed conventional source and streamer arrays. No assumptions are made regards artificial data whitening in processing, and the technology is fully compliant regards 3D pre-stack amplitude preservation. Introduction Two overlapping 3D seismic surveys were acquired with different survey azimuths in 1984 and 2004 over the Jasmine oil field in block B5/27, Gulf of Thailand. The 1984 acquisition parameters were typical of those used in the early days of offshore 3D, with relatively deep 7 m source depth and 10 m streamer depth. The well-known "receiver ghost" phenomena (Carlson et al., 2007) results in the data being dominated by strong low frequencies, and lacking adequate high frequency content. Conversely, the 2004 acquisition parameters pursued the high frequency bias achieved by a 2D site survey in 2000, and used a 4 m source and streamer depth combined with a spatially compact source array and 2D symmetric sampling principles (Ramsden et al., 2005). Figure 1 shows a zoomed comparison of pre-stack migrated data for the three surveys discussed in Ramsden et al. (2005), and Figure 2 shows superimposed amplitude spectra. Note how the high resolution 3D acquisition and processing delivers better event character and coherency in comparison to 2D data with comparable frequency content. The compact source array used in 2004 was designed to minimize source directivity, delivering high frequency output for large source emission angles to shallow targets. Furthermore, 12.5 m dual-source shot interval complemented by shot interpolation in processing was designed to deliver unaliased diffraction hyperbolae, thus enabling the successful application of high resolution linear and parabolic noise and multiple removal algorithms in processing.