Numerous seismic investigations have been performed to evaluate the validity of the velocity behavior determined in laboratory measurements in in-situ hydrate structures and determine the cause of the BSRs.
Miller et al. 1991 used synthetic waveform modeling of vertical-incidence traces from offshore Peru to determine a thin free gas zone beneath the BSR. Performing waveform inversion with a constant Poisson's ratio of 0.4 on data offshore Vancouver Island, Singh et al. 1993 concluded that the BSR in this region is overlain by hydrate-bearing sediment having a P-wave velocity of approximately 1.7 km/s and is underlain by gas-saturated sediment having a P-wave velocity of about 1.4 km/s. Analysis on data from two drill sites offshore from Vancouver Island showed similar results MacKay et al. (1994). Hyndman and Spence 1992, on the other hand, did not find evidence of a free gas zone in this region when they conducted a comprehensive seismic analysis of the same data used by Singh et al. 1993, including vertical-incidence waveform modeling, reflectivity modeling of AVO, and calculation of interval velocities from RMS stacking velocities. They concluded that the strong BSR amplitudes are caused by high concentrations of hydrate in a thin layer just above the BSR and normal velocity brine-saturated sediment beneath. Seismic studies in the Beaufort Sea related the BSR occurence in that area to the presence of a free gas layer beneath the BSR Andreassen et al. (1995). This study was conducted using forward AVO modeling with a constant Poisson's ratio of 0.38 in the hydrate-bearing sediments.
Recent studies of the BSR occurring at the Blake Outer Ridge, offshore
Florida and Georgia, strongly support the hydrate-over-gas model for this
region. Rowe and Gettrust 1993 analyzed deep-towed
multichannel seismic data, estimating a significantly high P-wave velocity
of over 2.5 km/s
in a thick hydrate layer above the BSR and a low P-wave velocity of 1.4 km/s
in the gas-saturated region beneath. Similar results were obtained by
Korenaga et al. 1997 using full waveform and travel
time inversion of wide-angle data. Slightly lower P-wave
velocities of 1.9 km/s - 2.0 km/s in the hydrate-bearing sediments were
obtained
by Wood et al. 1994 through travel time velocity analysis
and acoustic waveform inversion; by Katzman et al.
1994 using traveltime inversion, AVO analysis and
synthetic modeling of the same wide-angle data as Korenaga et al.
1997; and by Lee et al.
1994 based on 1-D seismic AVO inversion and
assumed hydrate properties. Recent drilling results of ODP leg 164
Matsumoto et al. (1996) and vertical seismic profiling
(VSP) Holbrook et al. (1996) suggest P-wave velocities
of no more than 1.9 km/s in the hydrated sediment and the presence
of a thick gas layer (
100 m) with velocities between 1.4 km/s and
1.6 km/s.
Using walk-away VSP data from one of those wells which, however, did not
penetrate a visible BSR, Pecher et al. 1996 furthermore
suggested that the S-wave velocity of hydrated sediment was higher than
the one in the sediments underneath.
All of the seismic investigations inferred the P-wave velocities at the BSR directly from the data, but made assumptions about the possible S-wave velocity behavior at this transition. These assumptions were mainly based on laboratory measurements which were directly translated into a possible behavior of in-situ hydrate structures. Only Pecher et al. 1996, who used converted waves visible in walk-away VSP data investigated the behavior of the shear moduli directly. Their study, however, is limited to an area of no BSR occurence and is only a 1-D representation of an heterogeneous medium.