SUMMARY AND IMPLICATIONS

The specific objectives of this expedition were to test gas hydrate formation models and constrain model parameters, especially models that account for the formation of concentrated gas hydrate occurrences through upward fluid and methane transport. During Expedition 311 a series of research sites across a transect through the Northern Cascadia margin was established by drilling and coring four sites extending from the westernmost accreted ridge (Site U1326, representing the youngest gas hydrate formation) and the first slope basin (Site U1325) to the second accreted ridge midslope (Site U1327) and a site at the eastern boundary of the inferred gas hydrate occurrence along this margin (Site U1329). We also cored and logged a cold vent (Site U1328), which represents a site of focused fluid and gas flow.

We collected the necessary data for characterizing gas hydrate proxies such as interstitial water chlorinities, core-derived gas chemistry, core physical properties, and downhole measured logging data. Geophysical characterization of each site was accomplished by the LWD/MWD program, conventional wireline deployments, and two VSP experiments. Extensive evidence for the presence of gas hydrate at all sites was collected from electrical resistivity and P-wave velocity logs, IR imaging, interstitial water and gas sampling, as well as from direct sampling of visible gas hydrate (Figs. F24, F25).

The main objectives of this cruise are considered fulfilled with only few elements having only partial success rates, such as temperature tool deployments and pressure coring, which are mainly a result of poor weather conditions.

The occurrence of gas hydrate has historically been inferred from the presence of a BSR in seismic images (Shipley et al., 1979). BSRs as ascribed to gas hydrates were first mentioned in the literature in 1977 from the Blake Ridge (e.g., Tucholke et al., 1977) and have since been a tool for gas hydrate detection. The BSR is the result of an impedance contrast generated by the transition from gas hydrate–bearing sediments above to gas hydrate–free and potentially free gas–bearing sediments underneath the interface. The presence of gas hydrate in sediments increases seismic P-wave velocity, whereas the presence of free gas (even small amounts) drastically reduces seismic P-wave velocity. The density is not expected to change much across this interface because the density changes introduced by the presence of gas hydrate and free gas are relatively small. The BSR reflection shows a phase-reversed polarity relative to that of the seafloor due to the negative impedance contrast at the interface.

The BSR has also been described as a frequency-dependent reflection (Chapman et al., 2002), which results from a finite-thickness gradient zone, in which velocity decreases from the elevated values above to low values below the interface. In conventional seismic surveys with low frequencies, the BSR appears as a strong reflection (see, e.g., Fig. F3), but with increasing frequency the magnitude of this reflection decreases.

BSRs have been widely observed in seismic data collected across the northern Cascadia margin (Hyndman and Spence, 1992; Hyndman, 1995) and were used to map the general distribution of gas hydrates along this margin (Fig. F1). At all sites visited during Expedition 311, BSRs were previously imaged with various seismic surveying techniques ranging over frequencies from 20 to 650 Hz (for a summary, see Hyndman et al., 2001). The depth to the BSR is a critical element in gas hydrate research because it is usually a measure of the base of the gas hydrate stability zone. However, it has also been shown that the BSR can occur away from the actual base of the stability zone at shallower or even deeper depths (Xu and Ruppell, 1999), where it depicts the first occurrence of free gas in the subsurface. Gas hydrate can also occur below the BSR, or even at a second BSR, in the form of Structure II gas hydrate.

The depth to the BSR in TWT has been determined from several crossing sections at each site visited during the expedition, and the average time was then converted to depth using velocities defined at Sites 889/890 from a VSP (Mackay et al., 1994). The BSR at Site 889 occurs at a depth of ~275 ms TWT, equivalent to a depth of 225 mbsf with the average velocity of 1636 m/s for the sediment column from seafloor to the BSR. This velocity was used across the entire margin for all sites visited and thus introduces a significant uncertainty in the depth assignment of the BSR. We therefore defined two extreme velocity profiles that describe sediments without gas hydrates (average velocity = 1619 m/s) and with concentrations well above what was observed (and inferred) at Site 889 (average velocity = 1653 m/s). The velocity increase appears very small; however, this is the result of only a thin sediment column over which the presence of gas hydrate was inferred (i.e., only 100 m above the BSR), leaving 125 m of sediment above this non-gas hydrate-bearing column. The range in velocity results then in a shift in the BSR depth by about

Throughout Expedition 311, special effort was made to better define the depth of the base of gas hydrate stability. Among the proxies used to define this boundary and to compare it to the seismically inferred BSR depth are

There are notable differences between the individual techniques, as they are highly dependent on how the measurements are conducted and the resolution or sensitivity of a particular measurement and as they are biased by core recovery (IR, pore water chemistry, and gas chemistry) and sampling density (e.g., frequency of temperature tool deployments and linear regression analysis).

Expedition 311 featured 31 temperature tool deployments in an attempt to characterize the thermal regime of sites drilled along the Expedition 311 transect (Table T2). Three standard IODP temperature tools were deployed during the expedition, including the APCT (nine times), DVTP (nine times), and DVTPP (five times). We also deployed the new APC3 tool a total of eight times. Figure F26 shows a compilation of all in situ temperature estimates from Expedition 311 compared to results from Site 889. The implied heat flow, assuming a constant thermal conductivity of 1.1 W/m·K, is also shown and compared to the regional heat flow determined by Hyndman and Wang (1993). Heat flow across the lower slope (Sites U1325, U1327, and U1328) appears to be depressed compared to the regional heat flow pattern, consistent with perturbation by a high sedimentation rate and upward fluid advection (Hyndman and Davis, 1992). Postcruise analysis will focus on detailed examination of data uncertainties and on processes to explain inter- and intrasite variation.

The primary goal of the temperature tool deployment program during Expedition 311 was to obtain the data needed to calculate the depth to the base of the gas hydrate stability zone at each of the sites visited during the expedition. Gas hydrate exists under a limited range of temperature and pressure conditions such that the depth and thickness of the zone of potential gas hydrate stability can be calculated. Most gas hydrate stability studies assume that the pore pressure gradient is hydrostatic (9.795 kPa/m). However, the seafloor temperature and geothermal gradient for any given site can be highly variable. The temperature data acquired during this expedition have been used to make a preliminary estimate of the depth to the base of the gas hydrate stability zone at each site. In these calculations a pure methane gas chemistry was assumed for the gas hydrate, whereas the interstitial water salinities from the onboard analysis of core samples were used to estimate the depth to the base of the methane hydrate stability zone at each site. The results of these preliminary calculations have been reported as range of depths for each site:

For the most part, the calculated depth to the base of the methane hydrate stability zone for each site falls near the estimated depth of the BSR as inferred from seismic data. In the case of Sites U1325 and U1326, however, it appears that the estimated depth of the BSR is less than what would be expected from the gas hydrate stability calculations. It should also be noted that the linear regression used to estimate the subsurface temperatures in the above calculations may not be appropriate for environments that are affected by fluid advection (e.g., the cold vent Site U1328). A more detailed temperature measuring program is required to better resolve the base of the gas hydrate stability zone and will be addressed in the Phase II drilling under this project by deployments of distributed fiber-optic temperature cables that allow continuous recording of the temperature field on a vertical sampling density of <1 m.

The occurrence of gas hydrate can be documented by a variety of methods. The most obvious and fundamental method is by direct visual observation of cores. During Expedition 311, visual observation of gas hydrates was primarily associated with the cold vent location, Site U1328, where a thick section of highly concentrated gas hydrate–bearing strata was inferred from LWD/MWD and previous precruise studies. From cores in this interval, numerous gas hydrate pieces were recovered from the upper ~35 mbsf, some measuring up to 5–8 cm in diameter (Fig. F25). Gas hydrate was also observed at other sites, more typically seen as small nodules or filling pores within sandy layers (Fig. F24).

Because gas hydrates, especially disseminated gas hydrates, begin to dissociate during core recovery, visual observation may not be possible because they will have decomposed before the cores can be processed on deck. Thus, indirect or proxy methods related to the physical and chemical consequences of dissociation must be used to document gas hydrate occurrences.

Gas hydrate dissociation releases gas and water during core retrieval, which should physically disturb the original sedimentary fabric. Previous observations have indicated two textures that are a product of dissociation (Westbrook, Carson, Musgrave, et al., 1994; Kastner et al., 1995; Lorensen et al., 2000). Soupy texture describes intervals that are water saturated and have lost all primary sedimentary structures, which are thought to represent areas that once contained larger pieces or veins of gas hydrate. Intervals with many small, round voids are termed mousselike and are interpreted to represent dissociation of disseminated gas hydrates. During Expedition 311, soupy and mousselike textures were common and were associated with intervals where direct and other indirect evidence of gas hydrate were noted. Site U1329 was the only site where these textures were not observed. At the cold vent Site U1328 in the upper shallow gas hydrate zone (lithostratigraphic Unit I), soupy texture was common, but not below this unit, whereas mousselike textures were observed in the lower units. Sites U1327, U1326, and U1325 also noted the presence of soupy and mousselike textures. At Site U1326 the lower part of a section with a notable IR anomaly (Section 311-U1326-18X-4) was sectioned and opened immediately. The sediment in this core had a very unusual texture, remaining cohesive and "foamy" rather than "soupy" or "moussy." Sediment in this section lost its foamy texture in less than an hour, collapsing to a stiff, dry sediment that occupied an estimated 25% of the original volume in the core liner.

The presence of gas hydrate in situ has often been inferred by freshening of pore water chlorinity/salinity relative to a defined baseline (e.g., Kastner et al., 1995). When gas hydrate forms in the subsurface the process initially increases pore water chlorinity/salinity because gas hydrate formation excludes the salt from the crystal structure. Over geologic time, the excess salt is removed by advective and diffusive processes and the background pore water chemistry is reestablished. In case of rapid and relatively recent gas hydrate formation, this process is incomplete and high–pore water chlorinity/salinity brines are observed (e.g., Site 1249 of Southern Hydrate Ridge during ODP Leg 204; Trehu, Bohrmann, Torres, Rack, et al., 2003). Such a system was also cored at the cold vent Site U1328 during this expedition.

If a core is retrieved on the catwalk that had gas hydrate present in situ, then gas hydrate has most likely completely dissociated, leaving freshwater behind, which can be detected by interstitial water analyses. A summary of all interstitial water chlorinity profiles linked to LWD/MWD RAB-derived borehole images is shown in Figure F27. The definition of an interstitial water chlorinity baseline has proven difficult because the effect of pore water freshening from gas hydrate dissociation and advection of fresher pore fluids from greater depth cannot easily be distinguished with chlorinity and salinity measurements alone. Additional shore-based pore water analyses, such as strontium, boron, or iodine isotopes, are required to distinguish these processes (Teichert et al., 2005; Fehn et al., in press).

The catwalk procedure for collecting interstitial water samples during this expedition had been adapted from previous "gas hydrate" expeditions and was strongly linked to IR imaging. The presence of gas hydrate in the core can be detected (if not visually) by a decrease in core temperature due to the cooling upon gas hydrate dissociation. One of the early systematic approaches to core temperature measurements was during Leg 164 using an array of digitally recorded thermocouple probes (Paull, Matsumoto, Wallace, et al., 1996). Initial development of the IR imaging technique was accomplished during Leg 201, where thermal anomalies in IR images were associated with gas hydrate and voids (Ford et al., 2003). Systematic IR thermal imaging of the surface of the core liner was first fully implemented during Leg 204 (Trehu, Bohrmann, Torres, Rack, et al., 2003). During Expedition 311 a slightly different version of the IR system was used to systematically scan the core for IR cold spot anomalies. A summary of all IR profiles from continuous coring are compared to Archie-derived pore water saturations (S_w) and LWD/MWD RAB-derived borehole images, shown in Figure F28.

Generally, several interstitial water samples were collected from most cores that did not show IR cold spot anomalies to define the in situ background pore water chemistry. In addition, we directly sampled IR-imaged cold spots in cores to measure the amount of pore water freshening associated with gas hydrate–related IR temperature anomalies. This will help to further calibrate the IR measurements and delineate anomalous trends from the background pore water chemistry. Experiments during Leg 204 showed that cold spot–related interstitial water freshening is focused spatially to a range of a few centimeters around the IR anomaly (Trehu et al., 2004). Each IR cold spot sampled was therefore carefully analyzed during this expedition. Once in the chemistry laboratory, the handheld IR camera was used to isolate the cold anomaly so a subsample of the sediment that had contained gas hydrate could be separated from background sediment (Fig. F25). The cold material was squeezed separately from the background material. In many cases, it was found that the sediment hosting the cold spot was coarser-grained sands and silts and the surrounding clay did not exhibit freshening.

Organic geochemical studies focused on volatile hydrocarbon gases from headspace samples, void gases, and selected gas hydrate samples. At all sites, analyses showed that methane is the dominant hydrocarbon gas within the cored intervals. However, the presence and distribution of higher hydrocarbons (C₂₊) provided useful information about the occurrence and possibly the type of gas hydrate. In particular, studies have shown that gas hydrate forms with the preferential selection of ethane into the gas hydrate structure, which is released during dissociation. At most sites, where other visual or proxy data indicated the presence of gas hydrates, void gases and selected gas hydrate and PCS samples contained elevated ethane concentrations and decreasing C₁/C₂ ratios (Fig. F29). In some intervals, elevated propane and isobutane were also observed, which suggest there may also be Structure II gas hydrate present in the cored interval as well.

Because gas hydrate is characterized by unique chemical compositions and distinct electrical resistivity and acoustic physical properties, it is possible to identify and further characterize gas hydrate–bearing sediment properties, including porosities and gas hydrate saturations, with downhole logging tools commonly deployed from the ship during IODP. In the case of Expedition 311, the standard downhole wireline logging program was augmented with a precoring LWD/MWD dedicated drilling-logging program. The downhole LWD/MWD logging data acquired during the start of this expedition was used to direct the special pressure coring program and make decisions on other critical special tool deployments. The LWD/MWD data, along with downhole wireline logging data, were also used to identify and quantify the occurrence of gas hydrate at each site drilled along the Expedition 311 transect.

As shown in Figure F30, the LWD/MWD-derived RAB images of borehole resistivities reveal numerous apparent high-resistivity intervals (shown as light color bands) within the stratigraphic section above the projected depth of the BSR. Also shown in Figure F30 are the Archie-derived water saturations (which is the mathematical complement of gas hydrate saturation) for each of the LWD/MWD-logged holes. Except for the cold vent Site U1328, the downhole logging–inferred gas hydrate occurrences appear to be first observed at a minimum depth of ~70–100 mbsf; and in most cases we see evidence for gas hydrates to about the base of the predicted methane hydrate stability zone. The apparent low water saturations below the BSR likely depict the occurrence of free gas. The downhole logging data from Site U1329 revealed relatively little evidence of gas hydrate. The downhole logging data from Sites U1327 and U1326, however, are dominated by the appearance of thick high-resistivity intervals near a depth of 100 mbsf at each site. In the case of Site U1327, we see an 18 m thick (120–138 mbsf) high-resistivity interval with peak LWD/MWD resistivity values exceeding 12 m, whereas in Hole U1326A we see a 35 m thick interbedded section (72–138 mbsf) of gas hydrate–bearing sands and non-gas hydrate–bearing clay sections. The logging-measured resistivities in gas hydrate–bearing sands at Site U1326 exceed 40 m for part of the section. Archie analyses of these high-resistivity intervals at both sites indicate gas hydrate saturations possibly exceeding 50%–75% of pore volume. The gas hydrate occurrence at Site U1325, however, appears to be distributed throughout the cored section of the site with the gas hydrate concentrated in thinly bedded sand units measuring between 5 and 20 cm thick. The downhole LWD/MWD logging data from Site U1328 reveal a very distinct zone of high resistivity extending from the seafloor to a depth of ~50 mbsf. Archie analysis of the resistivity logging data from Site U1328 indicates that this near-surface high-resistivity interval may contain a significant amount of gas hydrate. The closer examination of the RAB images from the cold vent Site U1328 also reveals numerous high-angle fractures cutting through the shallow gas hydrate–bearing section, which may have served as migration conduits for gas from potential deeper sources.

Expedition 311 had the most ambitious pressure coring and onboard pressure core analysis program ever attempted in the history of ocean drilling. Pressure cores retrieved at in situ pressures were used to determine methane hydrate quantity using degassing techniques and mass balance calculations and methane hydrate distribution using nondestructive measurement of the physical properties of the cores at in situ pressures. Large improvements in temperature control over previous expeditions (e.g., Leg 204; Tréhu, Bohrmann, Torres, Rack, et al., 2003) made the recovery and analysis of pressure cores more practical. Pressure coring is crucial for understanding the concentration of gas hydrate and free methane gas in marine sediments, their nature and distribution, and their effect on the intrinsic properties of the sediment.

Pressure cores were collected using the IODP PCS and HYACINTH FPC and HRC pressure corers. The PCS is a downhole tool designed to recover a 1 m long sediment core with a diameter of 4.32 cm at in situ pressure up to a maximum of 69 MPa (Pettigrew, 1992; Graber et al., 2002). Two types of wireline pressure coring tools were developed in the European Union–funded HYACE/HYACINTH programs: a percussion corer (FPC) and a rotary corer (HRC), which were designed to cut and recover core in a wide range of lithologies where gas hydrate–bearing formations might exist.

During Expedition 311, the PCS retrieved pressurized sediments for onboard degassing experiments. Controlled release of pressure from the PCS through a manifold permits

A total of 24 PCS runs were carried out during Expedition 311, of which 16 were successful and returned sediment under pressure. Each PCS was X-rayed before and after the degassing experiment for core characterization and to guide subsampling for interstitial water and gas headspace. Gas hydrate and free gas concentration estimates were determined from the mass balance calculations and are superimposed on the Archie-derived water saturation logs in Figure F30. The PCS-derived pore water gas hydrate and free gas concentrations agree well with estimates from LWD/MWD data. The results of all 16 successful PCS measurements are displayed in Figure F31, showing the calculated methane concentrations as function of depth and their position relative to the methane hydrate phase and methane solubility boundaries defined from the Xu model (Xu, 2002, 2004). The largest amount of gas hydrate within a PCS core was 40% of the pore space as inferred from degassing of Core 311-U1326C-12P, recovered from a depth of 83.7 mbsf within a high–electrical resistivity zone (Fig. F30). Other significantly large amounts of gas hydrates were inferred from Cores 311-U1328B-4P (15%) and 311-U1328E-10P (22%) at the cold vent (Table T3). A very large amount of free gas was inferred from the degassing of Core 311-U1328E-13P (58%) recovered from ~233 mbsf near the base of the gas hydrate stability zone. However, the core did not show any typical degassing features after the experiment, such as gas expansion cracks or large voids combined with mousselike or soupy texture;, it was concluded that the free gas was accidentally trapped within the outer chamber of the PCS and as it flowed into the borehole from a free gas–bearing formation.

In total, three HRC cores and one FPC core were stored under in situ pressures for future shore-based analyses. Onboard Expedition 311 these cores were X-rayed and analyzed with the multisensor core logger to measure density and P-wave velocity.

The presence of gas hydrate in marine sediments is mainly inferred from the presence of BSRs on seismic sections. Quantification of gas hydrate concentrations from remote sensing techniques such as seismic or CSEM methods was attempted earlier at this margin (e.g., Yuan et al., 1996, 1999; Yuan and Edwards, 2001; Hyndman and Spence, 1992), but these efforts need further calibration from borehole data, especially P-wave velocity, density, porosity, and electrical resistivity. Although borehole data may provide the needed calibration data, strong lateral heterogeneity in the derived gas hydrate concentrations and physical properties between adjacent holes only several tens of meters apart was observed at multiple sites visited during this expedition. Structural control of the gas hydrate occurrence, local dip of the formation, and other tectonic complications make estimates of gas hydrate concentration from remote-sensing techniques a challenging task and will be addressed postcruise in various studies.

In recent years significant progress has been made to address some of the key issues on the formation, occurrence, and stability of gas hydrates in nature. The concept of a "gas hydrate geologic system" approach is gaining some acceptance. In this approach the individual factors that contribute to the formation of a gas hydrate occurrences, such as gas source, gas and water migration, emplacement, and growth of the gas hydrate in a suitable host sediment, can be identified and quantified.

It has been shown that the availability of large quantities of hydrocarbon gas from either microbial or thermogenic sources or both is an important factor controlling the formation and distribution of natural gas hydrates (Kvenvolden, 1988). Carbon isotope analyses indicate that the methane in most oceanic gas hydrates is derived from microbial sources. However, molecular and isotopic analyses indicate a thermal origin for portions of the gas in several of the gas hydrate occurrences on the Cascadia margin (Pohlman et al., 2005). Additional factors controlling the availability of gas are the geologic controls on fluid migration. If effective migration pathways are not available, it is unlikely that a significant volume of gas hydrate would accumulate. Therefore, geologic parameters such as water and gas chemistry, as well as sediment permeability and the nature of faulting, must be evaluated to determine if the required gas and water can be delivered to the sedimentary section potentially hosting the gas hydrate.

Analysis of interstitial waters during Expedition 311 gave us important insight to fluid and gas sources and their movement along the Cascadia margin. At Sites U1327 and U1329 we see general decreases in salinity and chlorinity with depth, suggesting dilution. It was concluded that the dilution profile observed is primarily dominated by diffusive communication with an advective low-chlorinity fluid system at greater depth and not only to gas hydrate dissociation. In addition to chlorinity, the concentrations of Na, K, Ca, and Mg also decrease with depth, supporting the suggestion of communication with a deeper-sourced fresher fluid. At the vent Site U1328, salinities and chloride concentrations remain almost constant below the highly concentrated surface section of gas hydrate, again suggesting communication with a fluid at greater depth. But in this case the waters have undergone less freshening than at Site U1327. The salinity and chlorinity profiles at Sites U1326 and U1325, however, trend toward concentrations higher than seawater, which was interpreted as indicative of low-temperature diagenetic reaction. A plausible candidate for such a reaction is the alteration of volcanic ash to clay minerals and/or zeolites. These reactions consume water and, hence, increase in situ the chlorinity values. The analysis of core-derived hydrocarbon gases along the Expedition 311 transect, when linked to interstitial water chemistries, also show communication with a deeper source.

Hydrocarbon headspace gas measurements show that methane is the dominant hydrocarbon gas within the cored interval at every site along the transect. Although ethane and more complex hydrocarbon gas concentrations were low, the occurrence of more complex hydrocarbon gases generally indicates the contribution of a deeper gas source, much like the interstitial water chemistry results. It was also shown that the methane needed to sustain the shallow gas hydrate formation at the Site U1328 vent site was likely supplied along faults or fracture zones from a deeper source. It is also notable that the concentration of ethane and other gas hydrate–forming gases, including propane and isobutane, increase within the gas samples collected from the cores crossing the depth of the seismically inferred BSR at several sites. This gas geochemistry also strongly suggests the presence of a mixture of Structure I and Structure II gas hydrate.

It can be shown that by integrating the results of the interstitial water and gas geochemistry studies that the occurrence of gas hydrate along the Expedition 311 transect is controlled in part by the upward advection and diffusion of methane and other gas hydrate–forming gases from a deeper source. In the case of vent Site U1328 we see the influence of a focused flow system that has contributed to the development of a localized concentrated gas hydrate accumulation.

For the most part, the interpretation of downhole logging data and linked IR imaging and interstitial water analyses from the sites drilled along the Expedition 311 transect indicate that the occurrence of concentrated gas hydrate is mostly controlled by the presence of sand-rich sediments, which provide the conduits for gas migration and the containers in which for gas hydrate to grow. The occurrence of the relatively thick, highly concentrated gas hydrate accumulations within 100 mbsf at Sites U1326 and U1327 are also partially controlled by the presence of suitable host sands; but in these two cases we may also be seeing the effect of regional gas solubility in water on the occurrence of gas hydrates. We see relatively little evidence of gas hydrate or gas above these intervals at any other site visited during Expedition 311, other than the Site U1328 vent site at which free gas has been observed venting from the seafloor (Fig. F32).

Thus, preexpedition models that had predicted the occurrence of a highly concentrated gas hydrate section (50–100 m thick) just above the base of the gas hydrate stability zone do not to take into account the geologic complexity revealed during Expedition 311, in that the occurrence of gas hydrate appears to require the contribution from mixed gas sources, including a potential deep thermogenic gas source. The occurrence of migration conduits and host sands also controls the spatial distribution of gas hydrates on the Cascadia margin.