We have combined all the in situ sonic measurements made on Hydrate Ridge during Leg 204 in order to draw a complete overview of the implications of these data for the Hydrate Ridge system. The synthetic seismograms, calibrated or not by the VSP, enable a clear correlation between specific features in the logs and distinct reflections in the 3-D seismic data set. In addition to the BSR, the most significant reflectors are the various horizons below the gas hydrate stability field that contribute to the gas supply of the gas hydrate reservoir. Horizon A is the main free gas conduit, but Horizons B and B´ also display significant amounts of free gas. The Gassmann model also suggests a pervasive presence of free gas under the southern Hydrate Ridge summit.
The sonic logs also provide new estimates of gas hydrate saturations that agree overall with predictions based on the resistivity logs but differ in the details of the distribution. The different methods indicate an overall gas hydrate saturation of 10%–20% near the crest of southern Hydrate Ridge, decreasing to ~5% on the flanks. When present below the BSR, free gas saturation does not exceed a few percent of the pore space.
The cementation model used in this study is by no means the only formulation to derive gas hydrate saturations from velocity logs, and other models could possibly provide more reliable estimates. In particular, the suggestion of gas hydrate below the bottom of the gas hydrate stability field in several holes is obviously inaccurate because it is thermodynamically impossible and is a clear indication of the limits of this formulation. Guerin et al. (1999) successfully used this model in a significantly different environment, Blake Ridge, which is a passive margin where the penetrated sediments had an almost uniform lithology. Such description does not apply to the Hydrate Ridge sediments, which have a more diverse composition and, more significantly, are being actively deformed. The reasons why we chose this model for this study was its ease of use, its apparent accuracy in the previous study, and the fact that, unlike other effective media models, it offers a physical description of the grain-scale interaction between grains and hydrate. Its failure at more reliably identifying gas hydrate on Hydrate Ridge shows that the assumption that the gas hydrate/cement forms uniformly on spherical grains is an oversimplification of hydrate distribution and of the pore space. Guerin and Goldberg (2005) show that cementation occurs in the presence of gas hydrate but that friction between gas hydrate and the grains is responsible for the energy dissipation measured. Such friction implies that pore scale distribution is heterogeneous, which is hardly compatible with a uniform coating of grains by gas hydrate.
However, our goals were to provide not only a general overview of the implications of the acoustic logs on Hydrate Ridge, including new independent gas hydrate estimations to complement existing estimates, but also the integration of the logs and of the VSP with the seismic data, rather than a study of different possible models to determine gas hydrate saturations from acoustic logs. The overall agreement between the estimates derived from the resistivity and the velocity logs within the gas hydrate stability zone shows that it provides a reasonable estimate of global gas hydrate saturation, which has been shown to vary significantly between authors (Milkov et al., 2003).
Finally, the sonic logging waveforms offer a qualitative indication of the presence of gas hydrate and free gas. The sonic amplitudes are strongly dependent on the nature and the frequency of the source, and the high frequency dipole appears to be the most sensitive in this low-hydrate saturation environment.
The agreement between the different indicators described is best in the crest sites, in Holes 1247B and 1250F, which are the sites with the highest gas hydrate saturations. This suggests that the effect of gas hydrate on sonic velocity through cementation, and more significantly on sonic attenuation, requires significant amounts of gas hydrate. Surprisingly, although Site 1247 is located near the top of the mound, and the BSR is clearly present at this site, very little gas hydrate was recovered in the cores during Leg 204. The hydrate saturations derived from the elastic logs indicate significant values that were not identified at the time of drilling. Similarly, at Site 1250, the resistivity logs indicate low hydrate saturations immediately above the BSR whereas the bulk modulus and the elastic properties suggest high saturations, which are necessary to generate such a strong BSR. These two sites have the most clearly defined BSR of the visited sites. Overall, the results confirm that accurate identification of gas hydrate can be made only by the combination of various measurements in order to properly estimate the effect of heterogeneous free gas and gas hydrate distribution on the elastic and electrical properties of the formation.