In this section, we summarize some of the general conclusions that emerge from the above discussions of preliminary results at each site. We also point out additional questions that are the focus of ongoing postcruise data integration and modeling efforts.

Multiple proxies for hydrate presence and concentration are consistent and complementary. They indicate that gas hydrate is present, concentrated in relatively coarse-grained layers, over a broad depth range between the seafloor and the BSR. One of the main conclusions to emerge from the shipboard analysis of the data collected during Leg 204 is that gas hydrates are distributed through a broad depth range within the GHSZ. Electrical resistivity anomalies measured downhole with LWD and wireline logging, low-temperature anomalies measured with IR camera scans immediately after cores came on deck, low chloride concentration measured in interstitial waters, anomalously low C₁/C₂ ratios measured in vacutainer samples, and gas volumes measured from pressure core samples are all proxies for the presence of gas hydrate in the subsurface. All are consistently observed to start at a similar depth at a given site (~30–50 mbsf at sites away from the summit and at the seafloor near the summit). With the exception of local gas concentrations measured in PCS samples that are greater than the predicted solubility, none of these proxies extend below the BSR, which is the seismically predicted base of the gas hydrate stability field. The different proxies thus lead to consistent conclusions about the distribution and concentration of gas hydrate beneath Hydrate Ridge.

The various proxies for the presence of gas hydrate measure different length scales and have different sensitivity to hydrate concentration. Electrical resistivity, seismic velocity and attenuation, and temperature anomalies in recovered cores can be continuously measured on scales of centimeters and indicate that hydrate concentration probably varies considerably on scales of tens of centimeters. These data provide continuous profiles of relative hydrate concentration and indicate that hydrate is distributed over a broad depth range, with concentration controlled primarily by lithology. Exceptions are Site 1251 in the eastern slope basin, where hydrate is concentrated in a 10-m-thick zone just above the BSR, and sites near the summit, where it is concentrated near the seafloor. These geophysical parameters, however, do not give a direct measurement of concentration, although concentration can be estimated through physical models (e.g., through Archie's Law relation in the case of resistivity) (see Collett and Ladd, 2001).

The PCS data give a direct measurement of gas concentration in situ. The concentration of gas hydrate or free gas may be estimated by comparing the in situ gas concentration to the predicted in situ solubility. These measurements, however, are restricted to only a handful at each site because of logistical constraints. The ODP PCS provides measurements of concentration averaged >1 m, whereas the newer HYACE tools demonstrated the potential to provide additional information about the scale length of hydrate distribution and the presence of free gas within the GHSZ.

Chloride concentration measurements provide robust estimates of hydrate concentration if the background chloride concentration profile can be constrained. However, they are still spatially aliased because it is not practical to routinely obtain measurements less than several meters apart. Leg 204 has demonstrated that, except for near the summit where the background chloride concentration cannot be determined because of very rapid hydrate formation (see discussion of Site 1249), background chloride concentration can be well defined. Maximum concentrations are ~20% of pore space. It also demonstrates that it is essential to extend measurements well below the base of the GHSZ because there is a clear freshening of the interstitial waters as a result of diffusion of a low chloride concentration source at depth (probably water derived from dehydration of subducted minerals). Estimates that do not take this into account will overestimate the in situ hydrate concentration.

C₁/C₂ ratios have the potential to become a new proxy for hydrate concentration. Leg 204 data suggest preferential incorporation of ethane in the hydrate, which results in a signal that can be detected in routine headspace and vacutainer measurements. Additional analysis of the data from Leg 204 is underway to better calibrate quasicontinuous geophysical data using robust geochemical estimates of in situ hydrate concentration. This will result in greatly improved estimates of the total amount of methane and other hydrocarbon gases sequestered in the gas hydrate system on the Oregon continental margin. The calibrated downhole geophysical results will then be used to calibrate surface and seafloor seismic reflection and refraction data, which will provide tools to more robustly estimate hydrate distribution and concentration elsewhere.

Horizon A is an ash-rich layer that serves as a fluid pathway transporting methane and other hydrocarbons from the accretionary complex to the summit of Hydrate Ridge. Determining the origin and significance of a strong seismic reflection (Horizon A), which underlies the BSR and shallows toward the southern summit of Hydrate Ridge, was an important objective of Leg 204. This reflection was crossed at Sites 1245, 1247, 1248, and 1250. In the LWD data, it is characterized by a strong 2- to 4-m-wide double-peaked low-density and low-resistivity anomaly.

Cores at Sites 1245, 1248, and 1250 reveal several coarse-grained layers at this depth. Microscopic analysis of these sediments reveals that the sediment is composed primarily of relatively fresh glass shards indicating volcanic ash. The number and thickness of the ash layers could not be determined precisely because of generally poor recovery of this interval, probably because of high fluid content and possible overpressure. Grain density is low, probably because of vesicles within the glass shards. The age and provenance of the ash will be determined postcruise.

Although the physical properties of Horizon A are remarkably similar from site to site, hydraulic properties determined from chemical analysis of gases and interstitial waters are more complicated. Positive lithium anomalies are clearly associated with this horizon, supporting the interpretation that it is a conduit for fluids coming from greater depth. However, no thermal anomaly is detected, placing an upper limit on flow rate. Relatively high levels of higher hydrocarbons are observed in a 30- to 50-m-thick zone below the BSR at all sites, but the maximum level of higher hydrocarbons does not generally coincide with Horizon A (Fig. F15). A qualitative explanation for these observations is that diffusion away from Horizon A is superimposed on accumulation of hydrocarbons beneath the BSR. Postcruise modeling is needed to quantitatively test this hypothesis.

Massive hydrate lenses extend to a depth of ~30 mbsf near the summit of South Hydrate Ridge, and hydrate formation here is very rapid. Prior to Leg 204, it was known that methane bubbles were venting from the southern summit of Hydrate Ridge and that massive hydrate was present at the seafloor; however, the rate of hydrate formation and the depth to which massive hydrate is present were unknown. Seismic reflection data suggested that this zone extends to ~30 mbsf (Trehu et al., 2002) (Fig. F7). Electrical resistivity anomalies recorded in LWD data, however, indicate extremely high resistivities (approximately two orders of magnitude greater than observed at other sites) from 0 to 40 mbsf. The resistivity is higher than at other sites logged during Leg 204 to a depth of ~70 mbsf.

In the upper 10–20 mbsf near the summit at Sites 1249 and 1250, interstitial fluids are high chlorite brines, cores contain pervasive gas hydrate veins and nodules, and pressure core samples indicate the presence of massive hydrate. Moreover, core recovery from this interval was poor, probably because of massive hydrate presence. Below 15–30 mbsf, high-chloride brines give way to the low-chloride anomalies characteristic of dissociation of hydrate relative to background interstitial waters that do not retain the signature of hydrate-driven brine, presumably because of much slower rates of hydrate formation. At these depths, core recovery improved significantly. Both of these observations indicate a change from a shallow zone in which methane is provided through vigorous advection and hydrate forms rapidly to a zone in which methane flux is slower and hydrate formation is slower, perhaps because of limited availability of water. Below ~20–30 mbsf, lenses of massive hydrate containing significant pockets of free gas are not likely, although both hydrate and free gas in disseminated form may be present in relatively large concentration. The cause of the very high resistivity from 30 to 70 mbsf remains poorly understood and is a focus of postcruise research as is estimation of hydrate formation rate from the positive chloride concentration measurements.

Free gas is trapped in gas hydrates tens of meters beneath the seafloor at the southern summit of Hydrate Ridge. Gamma density logging of a HYACE RPC core from 14 mbsf at Site 1249 indicated the presence of several layers of massive hydrate, one of which contained material with very low density indicating the presence of free gas. This is the first direct evidence for free gas in the GHSZ, although the presence of free gas had previously been hypothesized based on observations of gas bubbles emanating from the seafloor (Suess et al., 2001; Torres et al., 2002). The new data provide strong support for the hypothesis that hydrate formation locally uses up all available water, isolating pockets of free gas. Modeling of elastic wave velocities and attenuation should further constrain the presence and distribution of free gas. The presence of free gas shallows within the stability zone has important implications for seafloor stability.

Lithology is a major factor influencing hydrate concentration. Integrated analysis of Horizon B provides an excellent example of the impact of lithology on hydrate distribution (see discussion of Sites 1244 and 1246). Detailed analysis of physical properties and IR thermal anomalies combined with lithologic description demonstrates that gas hydrates are concentrated in the coarse-grained layers in this ash-rich turbidite section.

Ethane is enriched beneath the BSR. All sites show an abrupt decrease in C₁/C₂ at the BSR except for Site 1252 (where there is no BSR). This effect is most evident at Sites 1244 and 1251, where this signal is not obscured by additional effects of Horizon A and results from an increase in C₂ rather than a decrease in C₁. There are at least two possible mechanisms to explain this observation. In one model, the BSR serves as a barrier to upward flow of ethane. In the other model, ethane is preferentially incorporated into hydrate; dissociation of the hydrate at a later time in response to tectonic uplift recycles ethane into the free-gas zone beneath the hydrate. Relative enrichment of ethane observed in gases from several dissociated hydrate samples support the second mechanism. Discontinuities at the BSR in several other chemical species support the first mechanism. It is possible that both mechanisms operate simultaneously.

There is less free gas beneath the BSR here than beneath north Hydrate Ridge or the Blake Plateau. One of the more surprising results of Leg 146 on north Hydrate Ridge and Leg 164 on the Blake Ridge was the depth to which free gas is present beneath the BSR. Seismic experiments at both of these sites indicated that free gas is present in the sediments for several hundred meters below the BSR (MacKay et al., 1994; Holbrook et al., 1996; Trehu and Flueh, 2001). In contrast, PCS measurements of gas concentration and seismic measurements during Leg 204 indicate that free gas is present beneath the BSR but only in thin layers.

The accretionary complex is permeable and is a source of fresh water, which must be accounted for when estimating hydrate concentration from chloride concentration anomalies. Data acquired during Leg 204 confirm the interpretation that the boundary in the 3-D seismic reflection data between stratified sediments and seismically incoherent material represents an unconformity between slope basin material and older indurated and fractured sediments. In addition, interstitial waters from sites on the eastern flank of Hydrate Ridge (Sites 1244, 1251, and 1252) show a decrease in chloride concentration and an increase in lithium with depth, which indicates that fresh water migrates from deeper in the accretionary complex where it probably originates from dehydration of subducted oceanic crust and sediment. The slope of the mixing curve between seawater and water from this deep source changes at the top of the accretionary complex, implying that the accretionary complex is more permeable than the overlying deformed slope basin and that diffuse vertical migration of fluids through this material is significant. Although this freshening effect has been documented previously from Cascadia (Kastner et al., 1995), implications for estimation of hydrate concentration from salinity has not been fully appreciated. Leg 204 provides a systematic transect across the margin that can be integrated with structural information to constrain depth and volume of dewatering and the mechanism of fluid expulsion.