PRELIMINARY RESULTS AND DISCUSSION

Gas hydrates in the samples were classified from the CT images based upon their shape. The results are listed in the Table T1 as "CT shape." The classification is generally sourced on terms that have been defined for the description of hydrates visualized by thermal imaging using IR cameras (Tréhu, Bohrmann, Rack, Torres, et al., 2003; J.L. Weinberger, pers. comm., 2005). The IR method uses the endothermic dissociation of hydrate, which forms cold thermal anomalies on the core liner. Such defined IR shapes are listed in Table T1 for comparison.

The following fabric types, listed from small to large hydrate features, were identified on the CT images (see "Supplementary Material" for CT images):

1. Disseminated hydrate (Fig. F1): smallest pieces of gas hydrate with grain sizes up to 3 mm, which are disseminated in the sediment; such hydrate grains are not connected. Sometimes grains are hardly visible, which may partly be due to dissociation during recovery. Based on their small mass, such grains dissociate very rapidly after leaving the gas hydrate stability field in comparison to hydrate pieces of larger size.
2. Veinlet (Fig. F1): thin tabular hydrate, 1 mm thick in different orientations. Veinlets can appear as disks in three dimensions and thus are the small or thin form of veins.
3. Vein (Fig. F2): similar to veinlets but differ in thickness from 1 mm up to 10 cm. Veins, as well as veinlets, are nonparallel to bedding planes and transect the sedimentary bedding. The dip angle varies from nearly horizontal to vertical. Analyzing the dip angles of veins, it is obvious that almost all angles are >30° with a tendency to angles >60°.
4. Layer (Fig. F3): differs from veins because of the orientation in the sediment: a layer transects the core conformably to bedding. Hydrate layers are often intercalated between fine-grained sediment layers.
5. Nodule (Fig. F4): spherical to oblate features, typically 1–5 cm in diameter, sometimes with edges.
6. Massive hydrate (Fig. F5): thicker than ~10 cm and less than ~25% intercalated sediment.

In many cases the difference between such defined shapes is based only on the orientation of the hydrate within the sedimentary bedding. An important example: veins and layers are comparable in size, but veins cross the bedding planes and layers are intercalated among horizontal sediment layers. In this study we did not use the term "lens" because the sample size was too small to observe a tapering margin. Nodules sometimes look like small lenses. They are described as spherical to oblate, and if their size exceeds the core diameter they appear as layers. Furthermore, massive hydrate may also represent a layer parallel to bedding, which on a larger scale most probably looks like a lens.

Other features often observed are bubble fabrics in which the bubble size and the density of bubbles vary. A bubble fabric often occurs in the outer part of the core slices (Fig. F1) and is caused by the formation of gas bubbles in a soupy sediment that results from the dissociation of hydrate during core recovery and the sampling procedure on the catwalk. Because the samples were immediately frozen in liquid nitrogen, the bubbles were conserved.

The density of the frozen mud containing the bubbles is higher than the density of hydrate but lower than the ambient sedimentary matrix. Disseminated hydrate seems to be particularly susceptible to rapid hydrate dissociation. Because of the effective water release from such hydrate, dissociation of disseminated hydrates in fine-grained sediments leads to the formation of soupy sediments where gas bubbles are constantly being released. Very often this results in soft-sediment deformation of various dimensions (Fig. F1). Small amounts of gas are released when pressure reduction reduces solubility.

Based on our CT analyses, hydrate from deeper in the GHSZ predominantly appears as veins or veinlets with dipping angles of >30°, which are fracture or channel fillings that occurred in the sediment sequence because of tectonic movements of Hydrate Ridge. Sometimes before these fractures were filled completely by hydrate, they were filled in situ by fluids or free gas and may represent structures through which free gas could move rapidly through the sediment column to the surface. The mechanism for supply of free gas from a horizon characterized as Reflector A from beneath the bottom-simulating reflector (BSR) and the transport mechanism through the GHSZ has been described by Tréhu et al. (2004). Free gas discharge from the seafloor into the water column of southern Hydrate Ridge has also been observed during various cruises (e.g., Suess et al, 1999; Heeschen et al., 2003). The rim of these channels or fractures within the GHSZ may be stabilized by hydrate formation at the sediment/gas interface and may explain why free gas can move upward away from any water contact. After the decrease of gas flow, the hydrate filled up the entire volume of the inner part of the channel or fracture. An example of such a steep-dipping vein is given in Figure F2B. The visible length of the x-axis of this hydrate vein (Section 204-1248C-10H-1) reached a length of 13.5 cm. Adding another 4.8 cm, measured on a subsample, it sums to a length of >18 cm, which is still not the total length because top and bottom have not been preserved.