In the previous section, we discussed the impact of Leg 204 on our understanding of gas hydrate processes in an accretionary complex based on information on the gas hydrate distribution. In this section, we discuss the techniques whereby this gas hydrate distribution was determined. Because gas hydrate is not stable at the sea surface, these estimates are based on various proxies. Some of these (e.g., pressure cores and pore water Cl– anomalies) provide well-constrained but incomplete estimates of the gas hydrate content of the subsurface because they sample only a small fraction of sediments within the GHSZ, whereas others (e.g., resistivity and sonic logs and infrared thermal anomalies) provide good spatial coverage, but algorithms for quantifying gas hydrate from the observations must be calibrated and verified by other techniques. Using data from Site 1244, Figure F8 illustrates the variety of length scales to which these different data sets are sensitive. It also provides additional evidence for lateral heterogeneity between holes spaced tens of meters apart at a given site. Table T2 shows that estimates of the average gas hydrate content of the sediments derived by different techniques are generally similar. It is also apparent that regions characterized by many faults, as imaged by the seismic data (e.g., Site 1244 and 1246), show more variability between holes at a particular site than regions characterized by simple structure (e.g., Site 1245).
Pressure core samplers allow us to recover short (~1 m) sediment cores at in situ pressure conditions (Dickens et al., 1997). This is the only way to recover all in situ methane and thus directly measure the total gas concentration at discrete depths within sediment sequences. The gas hydrate content of the core can then be calculated from the total methane concentration assuming thermodynamic equilibrium and that porosity and pore water chemistry are known (Dickens et al., 2000).
A record-breaking number of pressure core samples for a single cruise was obtained during Leg 204. As the most direct measurement of gas hydrate content, results from these cores provide a good estimate of the average amount of gas hydrate present within the study region (Milkov et al., 2003); they also provide critical information to test estimates of gas hydrate content obtained by other means (Tréhu et al., 2004b). In general, however, pressure core data provide limited constraint on the details of gas hydrate distribution because the number of pressure core samples at a given site is too few to define the vertical variation in gas hydrate content.
Leg 204 was a testbed for a new generation of pressure core samplers (the HYACE system), which was designed so that cores recovered under pressure can be logged at in situ conditions and as pressure is released (Fig. F8A). The HYACE cores provided critical information about how gas hydrate is distributed within cores at in situ conditions and how it responds as pressure is released (Tréhu, Bohrmann, Rack, Torres, et al., 2003; Tréhu et al., 2004b).
Gas hydrate dissociation is strongly endothermic, resulting in cold spots in cores (Figs. F6C, F6D, F7C, F8B), which were originally detected by hand (e.g., Paull, Matsumoto, Wallace, et al., 1996). Infrared cameras were first introduced to ODP during Leg 201 (Ford et al., 2003) and were used systematically to scan all core recovered from within or near the GHSZ during Leg 204. These scans were invaluable for rapidly identifying gas hydrate samples for special shipboard experiments (e.g., Riedel et al., 2006) or for preservation in liquid nitrogen for future studies (e.g., Abegg et al., this volume).
An important advantage of this technique, which distinguishes this approach from the continuous geophysical records obtained by downhole logging, is the ability to directly sample lithologies and pore waters associated with the IR anomalies. Initial attempts to calibrate the anomalies recorded in catwalk core scans using coincident measurements of chloride concentration were presented by Tréhu et al. (2004b). These data provide important information on the spatial distribution of gas hydrate, including the thickness, spacing, and shape of gas hydrate lenses and nodules (Weinberger et al., 2005). Although the amplitude of any particular anomaly depends on how the hydrate is distributed in the core, the average gas hydrate content of the sediments can be estimated using simple, empirical calibrations based on pore water Cl– data (Tréhu et al., 2004b).
The most widely used geochemical proxy for gas hydrate presence is based on the accurate measurement of dissolved chloride (Cl–) in the pore fluids. During core recovery, gas hydrate dissociates, resulting in dilution of the Cl– concentration by addition of water sequestered in the gas hydrate lattice prior to core recovery. The use of the Cl– proxy is predicated on the assumption that the background Cl– concentration is known and that the rate of hydrate formation is slow enough that high Cl– anomalies resulting from salt exclusion during hydrate formation have been removed by diffusion and advection (Hesse and Harrison, 1981; Ussler and Paull, 2001). For example, in Figure F8C, different gas hydrate estimates are derived if the background Cl– content of the pore water is assumed to be equal to that of seawater or if only the discrete low Cl– spike is attributed to gas hydrate dissociation (i.e., using an empirical baseline). With the latter assumption, the maximum gas hydrate content indicated by Cl– data at Site 1244 is ~8% of the pore space and the average gas hydrate content of the gas hydrate–bearing sediment that extends from 30 to 130 mbsf is ~3% (Tréhu et al., 2004b). Increased freshening of pore waters with subseafloor depth and with distance from the deformation front at Sites 1244–1246, 1251, and 1252 (Fig. F2A) has been used to separate effects of clay dehydration reactions and gas hydrate dissociation on the dissolved Cl– distribution (Torres et al., 2004a). The observations are consistent with enhanced conversion of smectite to illite, driven by an increase in temperature and age of the accreted sediments. The gas hydrate pore space saturations derived from Cl– data in Figures F6 and F7 were also determined using an empirical baseline.
Compared to PCS samples, more chloride samples are obtained in each hole and each sample averages over a smaller distance; nonethe-less, considerable statistical undersampling remains. Another important use of chloride data during Leg 204 was to calibrate data from IR temperature scans (see "Infrared Camera Scans of Cores"). This line of inquiry was further developed during IODP Expedition 311 (Expedition 311 Scientists, 2005). One limitation of Cl– as a gas hydrate proxy during Leg 204 was that it cannot be used when hydrate is forming more rapidly than the excluded salts can be removed by diffusion or fluid advection. This was the case in the shallow subsurface near the summit (Fig. F5H). However, in this case, the resultant brines provide important constraints on system dynamics, as discussed in "Transport-Dominated Regime." There may also be problems defining a baseline when gas hydrate is evenly disseminated in the pore space over tens of meters in depth.
A possible new geochemical proxy is discussed by Milkov et al. (2004a). Ethane-enriched Structure I gas hydrate solids are buried more rapidly than ethane-depleted dissolved gas in the pore water because of upward advection of pore water due to compaction. With subsidence beneath the GHSZ, the ethane (mainly of low-temperature thermogenic origin) is released back to the dissolved gas–free gas phases and produces a discontinuous decrease in the C1/C2 vs. depth trend. These ethane fractionation effects may also be useful to reconstruct upward migration of the base of the gas hydrate stability field in the past (Claypool et al., this volume). For example, release of ethane from gas hydrate may be responsible for the increase in ethane in pore waters from beneath the BSR at Site 1251.
Various downhole geophysical logs were used to estimate the gas hydrate content of the subsurface, including electrical resistivity (Tréhu, Bohrmann, Rack, Torres, et al., 2003; Lee and Collett, this volume; Janik et al., 2003), sonic velocity (Lee and Collett, this volume; Guerin et al., this volume), and NMR (Collett et al., this volume), and were critical for determining the gas saturation in Horizon A (see "Transport-Dominated Regime").
Geophysical logs provide a number of important advantages: (1) they are relatively quick and economical to acquire; (2) they provide nearly continuous coverage of the subsurface, except for in the upper few tens of meters where log quality is bad (e.g., Figs. F5D, F7B, F8D); (3) they image the subsurface with the highest resolution, with some techniques providing centimeter-scale azimuthal images of the borehole (e.g., Fig. F8D); and (4) they provide unique information on in situ physical properties (e.g., density, electrical resistivity, seismic velocity, and in situ stress), which cannot be measured on core samples because of degassing and deformation during core recovery (e.g., Fig. F5D, F5G).
However, a number of assumptions are needed to derive estimates of the gas hydrate content of the sediments from geophysical logs, and it is difficult to directly test the validity of these assumptions, particularly when no core data are available from a logged hole, as is the case with most LWD data and when core recovery is poor. For example, electrical resistivity data cannot distinguish between gas hydrate and free gas. Comparisons between different holes that are several tens of meters apart are difficult because it is clear that there is considerable lateral heterogeneity in gas hydrate distribution (see "Lateral Heterogeneity in Gas Hydrate Content").
Seismic data acquired during Leg 204, including vertical, offset, and walkaway VSPs and regional high-resolution seismic reflection and refraction profiles, provide the largest-scale estimates of gas hydrate distribution. Successful VSPs were acquired at Sites 1244, 1247, and 1250 (Tréhu et al., this volume; Guerin et al., this volume). These data indicate that, in contrast to results from Legs 146 and 164 where free gas is present in a zone tens to hundreds of meters thick beneath the BSR (e.g., MacKay et al., 1994; Holbrook et al., 1996), free gas beneath the GHSZ at SHR is restricted to narrow, stratigraphically controlled conduits (e.g., Fig. F5G). Walkaway VSPs and regional ocean bottom seismograph (OBS) refraction profiles, which indicate velocity anisotropy beneath the summit but not beneath the adjacent basin, help to characterize gas hydrate distribution away from the drill sites. These data suggest that gas hydrate near the base of the GHSZ in the summit region is concentrated in vertical fractures that are aligned along the ridge axis and does not significantly impact the stiffness of the sediment (Kumar et al., 2006, in press).