Imaging of gas hydrate bearing marine sediments by full waveform inversion of multichannel seismic data from Krishna-Godavari basin, India

Tuesday, 16 December 2014
Maheswar Ojha1, Mrinal K Sen2 and Kalachand Sain1, (1)National Geophysical Research Institute, Hyderabad, India, (2)University of Texas at Austin, Austin, TX, United States
Limitation of offset, lack of low frequencies, and low velocity contrast at shallow marine sediment make it difficult to apply full waveform inversion (FWI) to conventional marine multichannel seismic data. To mitigate these problems, we make use of moderate offset conventional end-on multichannel seismic data, which are transformed to split spread configuration using reciprocity. Data redundancy in split-spread configuration provides better constraints in inversion resulting in a geologically meaningful velocity model. Here, we employ an acoustic FWI [1] in frequency domain with a constant density to resolve fine scale velocity structure of gas hydrate bearing sediments in Krishna Godavari offshore basin from eastern Indian offshore. Conventional multi channel surface seismic data were acquired in 2010 [2] using 360 channels with 12.5m receiver interval, 25m shot spacing and 100m near offset for the investigation gas hydrate in this region. In end-on configuration, we have 400 shot gathers with 360 channels each, which were transformed to split-spread gathers. We make use of a smoothed interval velocity derived from root mean square (RMS) velocity as an initial model with a grid spacing of 12.5m and ran inversion for fourteen frequencies (8 to 21 Hz). The split-spread dataset is able to resolve subsurface features much better than the conventional end-on data. For example, we notice marked improvement in identifying the free gas layer below the hydrate bearing sediments and many structural features like faults are conspicuous in the final image. However, shallow structures with low impedance contrast could not be resolved further using higher frequencies since a lot of high frequency noises/artifacts are created in the model. We use the resulting high-resolution velocity model together with rock-physics model to estimate hydrate saturation within the sediments.
  1. Virieux, J.; Operto S. Geophysics, 2009, 74, WCC127–WCC152.
  2. Sain, K.; Ojha, M.; et al. JGSI, 2012, 79, 553-556.