Ts;A suite of physical property measurements was made to complement the other data sets taken on board and to support the main scientific objectives of Leg 204. Physical characteristics of the subsurface environment play an important role in determining the nature of fluid and gas migration, which, in turn, affect the nature of microbial communities and gas hydrate formation.
Sediment physical properties were routinely measured using the MST on whole-round cores and using discrete samples to measure bulk density, porosity, grain density, compressional wave (P-wave) velocity (VP), and magnetic susceptibility (MS). In addition, thermal conductivity was measured on whole cores.
Core handling was done routinely following ODP standards (i.e., after sectioning the core on the catwalk, the core sections were stored inside the laboratory to equilibrate to ambient room temperatures). This took, on average, 4 hr. After temperature equilibration, MST measurements were conducted on whole-round core, followed by measurements of thermal conductivity. The core sections were then split for discrete sampling for moisture and density (MAD) as well as for measurements of shear strength and VP with the Hamilton Frame velocimeter.
Brief summaries of each of the physical property measurement procedures used during Leg 204 are presented below, but the reader is referred to Blum (1997) for more detail. Apart from these routine measurements, which are made during most ODP legs, some special procedures and specific measurements related to hydrates were used during Leg 204. In particular, we collected IR images of the cores on the catwalk to identify and quantify the presence of gas hydrate. These measurements were supplemented by the use of handheld IR cameras and by self-logging temperature-probe measurements made both on the catwalk and in the core laboratory. We also used a Vertical Multi Sensor Core Logger (V-MSCL) on pressurized and repressurized cores in an attempt to help characterize the in situ properties of gas hydrate-bearing lithologies (see "Downhole Tools and Pressure Coring").
IR thermal imaging of the surface of the core liner was fully implemented during Leg 204. The initial development of the technique was accomplished during Leg 201, where IR imaging was shown to successfully identify thermal anomalies associated with gas hydrate and voids.
Thermal anomalies in marine sediment cores on short length scales (less than a few meters) could result from (1) adiabatic gas expansion, (2) gas exsolution from pore water, or (3) gas hydrate dissociation. All of these processes cool cores (Ussler et al., unpubl. data). However, discrete, strong cold anomalies in Leg 204 cores were shown to be directly associated spatially with gas hydrate. These negative temperature anomalies developed, for the most part, in response to gas hydrate dissociation after the core arrived on deck. Variations in heat capacity also impact core temperatures, but generally differences in thermal conductivity or density are relatively small and do not result in large enough heat capacity differences to cause discrete, negative thermal anomalies.
IR imaging on Leg 204 also confirms earlier thermistor measurements that show variations in thermal structure along entire cores, which are typically warmer at the bottom and cooler at the top (Ussler et al., unpubl. data). Gas expansion and gas exsolution may account for the observed gradient along cores. If so, the thermal structure developed principally during ascent of the core through the upper part of the water column. Alternatively, the thermal structure along entire cores may reflect differences in frictional heating during coring, creating warmer temperatures near the core barrel shoe (APC) or bit (XCB) and cooler temperatures near the core top. Analysis of data from the advanced piston corer methane (APCM) tool from Leg 204 is expected to help determine the origin of temperature gradients typically observed along each core. Regardless of the origin of the overall thermal structure of cores, the discrete, negative thermal anomalies associated with gas hydrate are superimposed on the broader gradient, providing a robust proxy for the location and abundance of gas hydrate in cores. Gas shows up as warm anomalies because of the low heat capacity of gas voids compared to sediment and in spite of the cooling effect of gas expansion.
The primary benefits of using IR cameras (rather than running a hand down the length of the core) include the following: (1) more precise identification of thermal anomalies, (2) the estimation of hydrate volume in processed images, and (3) determinations of shapes of gas hydrate. The IR camera is also quicker and simpler to use and has a much higher spatial resolution than an array of thermistors. Hydrate veins or lenses, hydrate nodules, and disseminated gas hydrate were all identified. The resolution of thermal anomalies observed indicates that the camera can detect small volumes of gas hydrate if they are adjacent to the core liner. Determining precise, quantitative volumetric estimates of gas hydrate in cores was an objective during Leg 204, but realizing this objective will require further postcruise analysis of collected data. IR images were used to do the following:
Ts;
Two ThermaCam SC 2000 cameras (FLIR Systems) and an AVIO Neo Thermo model TVS-610 (Nippon Avionics Co., Ltd.) were used to map temperature variations along core. The FLIR Systems cameras provide temperature-calibrated images over a temperature range from -40° to 
1500°C. For shipboard measurements, the cameras were set to record a more limited range of temperatures from -40° to 120°C (Range 1). To perform the critical task of rapid identification of gas hydrate within the core on the catwalk, one of the FLIR Systems IR cameras was mounted on a track above the catwalk and driven automatically by a stepper motor controlled by a LabView software program. The camera was mounted in such a way that a 20-cm field of view along the core was obtained with the camera lens located 55 cm from the top of the surface of the core liner. To minimize the effect of external IR radiation reflecting off the core-liner surface, the camera was enclosed within a cardboard sheath covered on the outside with aluminized Mylar and lined on the inside with black felt (Fig. F10). After operations at Site 1244 were complete, the catwalk was also shaded with cardboard and the camera enclosure was extended to minimize ambient IR reflections from the core liner.
Images and data for each core were acquired immediately after the core liner was wiped dry. The mounted camera was moved down the core in 20-cm increments, starting 10 cm from the core top. Each image was saved with a unique identifier (header file) that included information about the position of the center of the image. Initially, all images were automatically saved over the shipboard computer network. Network slowdowns early in the cruise necessitated modification of the system to save images for a single core directly on the computer controlling the camera. Images were then manually transferred over the network to a laptop computer dedicated to processing of the IR images. Camera span and level parameters were automatically adjusted to optimize visual contrast on the computer screen for the expected downcore temperature variation. A physical properties scientist or technician observed the scan results visually, either by following the camera-mounted monitor or by looking at the monitor connected to the computer controlling the scan. As soon as each scan was completed, scientists typically examined the scan results on the laptop computer running ThermaCam Researcher software in the core laboratory. The locations of thermal anomalies were marked on the core and whole-round samples (e.g., hydrate and microbiology samples) were collected as defined by the core-sampling plan for the hole. After catwalk sampling, the cores were cut into 1.5-m sections and many cores were reimaged to provide an image sequence matching the length of the curated core sections.
The track-mounted IR imaging camera was supplemented on most cores by discrete imaging using a second FLIR Systems camera or an AVIO Neo Thermo camera in a handheld mode (Fig. F10). In this mode, a physical properties scientist walked along the catwalk with the IR camera pointed perpendicular to the core liner, usually just after the track scan was completed. This approach provided immediate confirmation of temperature anomalies from a viewpoint rotated 90° from the orientation of the track scan. If anomalies were detected, they were marked on the core, stored on the camera's memory card, and logged on a log sheet; later, the files were transferred to the ship's computer network for storage, analysis, and archiving. Temperature anomalies detected with the handheld camera were compared with the track scans, and decisions were made to guide the sampling of any gas hydrate present.
In some instances, subtle thermal anomalies or observations of gas escaping from holes drilled into the core liner triggered the decision to perform an additional track scan. Results from successive scans could then be directly compared; typical results show the spatial expansion of cold zones over time periods of only a few minutes.
To develop downcore profiles of temperature anomalies, the following procedures were established, building on the experience from Leg 201:
Following this process, the data files were combined to provide profiles of downhole temperature anomalies. Where core recovery was >100%, the overlapping depth interval was manually edited and the data were deleted from the resulting profile.
Downcore thermal profiles and IR images were used to extract individual negative thermal anomalies (cold zones on the core liner). Parameters obtained were (1) core and section (if available); (2) depth interval on uncut liner (top and bottom); (3) T (peak anomaly temperature-background temperature); and (4) gas hydrate texture or shape description and other comments on anomaly shape. A unique identifier was assigned to each IR anomaly for reference in text and figures. Results were tabulated for each site. Thermal anomalies were identified from the downcore temperature profiles derived from the images as described in the previous section. An analysis of thermal data on board showed that T values indicative of hydrate were insensitive to ambient catwalk temperature and illumination conditions. The T values provide an approximate measure of hydrate abundance, albeit influenced by the proximity of hydrate to the core liner. Gas hydrate undergoing dissociation and directly in contact with the core liner produces a larger T than hydrate insulated from the liner by sediment.
It is important to note that depth measurements were recorded on uncut core liners before sectioning and removal of gas voids. Hence, depth assignments do not precisely match the curated depths of core sections; typically depth differences are <1 m. Hydrate samples are included on IR anomaly tables and provide specific data on curated and uncut core-liner depth discrepancies.
Extracted thermal anomalies were plotted as a function of depth and, for many sites, compared with pore water saturation (Sw) derived from resistivity logs using Archie's Relation (see "Downhole Logging"). The visual comparison of the two data sets provided a useful means of assessing the overall similarity of the two methods in detecting differences in gas hydrate abundance as a function of depth. Calculation of Sw is ordinarily used to estimate volume fraction of pore space occupied by pore water in a homogeneous media as opposed to a more resistive fluid such as gas or liquid hydrocarbons. In the GHSZ, it is assumed that the resistive material is gas hydrate, which may be present in veins that are very heterogeneous. This heterogeneity may cause large errors when calculating Sw in a hydrate-bearing sediment and hence should be used with caution. Sw estimates from Hydrate Ridge typically range from 1.0 to 0.5 (usually 1.0 to 0.85), but in some cases (Site 1249) values as low as ~0.1 were obtained. These values are interpreted as
including all hydrate regardless of size of hydrate features (not just pore scale hydrate). IR anomalies and Sw were plotted back-to-back with depth on the ordinate, T increasing to the left, and Sw increasing to the right on the abscissa. The position of the Sw plot extending to the right from the abscissa provides a visual depiction of the estimated hydrate concentration for direct comparison with the IR thermal anomalies. Values >1 (physically impossible) are truncated in this and other plots of Sw in this volume.
A variety of gas hydrate shapes were observed during the leg, providing significant new insight into how those shapes and textures are distributed in marine sediments. Usage of terms for describing hydrate shapes evolved during the leg as sites were drilled. Defining terms used is particularly important for hydrate studies because of the ephemeral nature of gas hydrates. Gas hydrates must be either described within minutes after sampling and prior to dissociation or they must be preserved in pressure vessels or liquid nitrogen where, for the most part, they are not available for further description shipboard. However, IR imaging provided permanent information on hydrate shapes to a resolution of ~0.5 cm. These IR images together with visual observations and photographs of gas hydrate form the basis for the terms described below to define macroscopic (scales approximately 0.5 cm) geometries of gas hydrates:
- Layer: tabular gas hydrate feature that transects the core conformable to bedding. Its apparent thickness is, typically, on the order of a few centimeters. Layers thicker than ~10 cm are generally considered massive hydrate (interval 204-1250C-11H-3, 94-95 cm). See Figure F11F.
- Lens: a hydrate layer or other feature with tapering margins. Many hydrate layers may be lenses on a scale larger than the core diameter. See Figure F11.
- Vein: tabular gas hydrate feature that transects the core at an angle to bedding. Its apparent thickness is, typically, on the order of a few centimeters. Veins thicker than ~10 cm are generally considered massive hydrate (interval 204-1244C-8H-1, 47-52 cm). See Figure F11D.
- Veinlet: thin, tabular hydrates ~1 mm thick or less, commonly present adjacent to veins or layers and oriented in mutually orthogonal directions. Veinlets are visible by eye or with the aid of a hand lens but were commonly the smallest macroscopic hydrate features observed during Leg 204. See Figure F8C in the "Site 1249" chapter.
- Nodular: spherical to oblate features typically 1-5 cm in diameter. Two-dimensional circular shapes in IR images on core liners are usually described as nodular, recognizing that 3-D shapes may actually be blades or rods. Shipboard examination of liquid nitrogen-preserved cores clearly shows that some nodular IR features are, in fact, nearly spherical in shape, whereas others are rod or blade shaped with long dimensions significantly greater than the core diameter (interval 204-1244C-10H-2, 70-103 cm). See Figure F11B.
- Disseminated: hydrate grains less than ~3 mm distributed throughout the sediment matrix. This includes grain size ranging from ~3 mm to pore scale and produces IR anomalies that have diffuse boundaries. IR data typically cannot distinguish between disseminated hydrate and a single hydrate nodule in the center of the core. Any distributed form of hydrate with a small dimension <3 mm may be described as disseminated (interval 204-1249C-2H-1, 108-140 cm). See Figure F11A.
- Massive: the presence of hydrate in core greater than ~10 cm in thickness and with less than ~ 25% intercalated sediment (Section 204-1249C-1H-CC). See Figure F11C.
The MST has four physical property sensors mounted on an automated track that sequentially measure MS, gamma ray attenuation (GRA) density, VP , and natural gamma emissions on intact whole-round core sections. We concluded that at low count times the natural gamma data would not contribute to the leg objectives. Therefore, we did not run the natural gamma system during Leg 204. However, we did have the opportunity to test the new Geotek Non Contact Resistivity (NCR) system.
Whole-core MST measurements are nondestructive to sediment fabric and can be used as proxies for other data as well as for facilitating core-to-core correlation between adjacent holes at the same site or among different sites. These correlations between holes can be used to construct composite stratigraphic sections at a single site and to correlate core data with measurements made by downhole tools. Each device has an intrinsic downcore spatial resolution determined by its design specifications (see discussion of each measurement below). Data quality is a function of both core quality and sensor precision. Optimal MST measurements require a completely filled core liner with minimal drilling disturbance. Precision is a function of measurement time for MS and GRA density but not for VP. The spatial interval used for all sensors was generally set at 2.5 cm.
At the end of Hole 1251B operations, we discovered, while checking the new position of the NCR system, that the positions of the other sensors on the track as entered into the configuration set up in the controlling computer had a small error. The positions are measured from the end of the water check sample to the sensor. Table T3 shows the original positions (for Hole 1251B; before 29 July 2002 at 2235 hr) that are in error, the actual measured position, and the new positions as entered (for Hole 1251B; before 29 July at 2235 hr).
Prior to the adjustment and including data from previous legs, we concluded that offsets between MST data and core position are in error by 1.3 cm, whereas after the change, the error was reduced to 0.5 cm.
Whole-core volume MS was measured using a Bartington MS2 meter coupled to a MS2C sensor coil with a diameter of 8.8 cm operating at 565 Hz. The measurement resolution of the MS2C sensor is a function of many factors but is generally taken to be around 3-4 cm. The instrument has two fixed integration periods of ~1 and 10 s. During Leg 204, MS was routinely measured at a spacing of 2.5 cm, with the average of three 1-s data acquisitions being recorded for each sample location, unless otherwise stated. The instrument automatically zeroes and records a free-air value for MS at the start and end of each section run. Instrument drift during a section run is then accommodated by subtraction of a linear interpolation between the first and last free-air readings. Drift-corrected MS data were archived as raw instrument units (SI) and not corrected for changes in sediment volume. Therefore, the data is reported as "uncorrected volume susceptibility."
GRA density was determined using the GRA densitometer. This sensor system measures the attenuation (mainly by Compton scattering) of a gamma beam caused by the average electron density in the gamma path. A well-collimated gamma beam (primary photon energy of 662 keV) is produced from a small (370 MBq; ~1994) 137Cs source (half-life = 30.2 yr) and passes through an assumed known thickness of sediment (internal diameter of core liner). At the start of Leg 204, we discovered that the energy counting window was incorrectly set, as it was counting a significant number of low-energy photons (Compton scattered). For the most accurate results, it is important to only count the unattenuated photons (662 keV); hence, the window should be set evenly around the 137Cs peak but beginning at the base of the "Compton trough."
The measurements are most empirically related to the bulk density of the material, and hence, the data are often referred to as "wet bulk density." However, we believe this to be slightly confusing and prefer to refer to this data set as "gamma density," or GRA density, which we then compare to wet bulk density measured by MAD gravimetric techniques. Although the empirical calibration procedure for GRA is based on bulk density measurements (i.e., of a known graduated aluminum and water standard), the measurements will vary from true gravimetric bulk density because of variations in mineralogy. Gamma attenuation coefficients for different materials vary as a function of atomic number. Fortuitously, most earth-forming minerals have similar and low atomic numbers (similar to aluminum). Consequently, the correlation of GRA density and bulk density is usually very good. In summary, GRA density should be considered as the density of sediment and rocks as determined from GRA measurements using aluminum and water as reference materials.
The gamma source collimator is 5 mm, which produces an effective downcore spatial resolution of ~1 cm. Following our detailed setup and calibration procedure, we logged the graduated aluminum and water standard as a check to confirm accurate calibration. These data indicate an excellent calibration and illustrate the downcore spatial resolution (Fig. F12). The minimum integration time for a statistically significant GRA density measurement is 1 s. During many ODP legs, a count time of 2 s is used. However, we considered this count period to be too short and used a 5-s count time throughout Leg 204 in order to improve precision. A freshwater control was run with each section to measure instrument drift. GRA data are of highest quality when measured on nongassy APC cores because the liner is generally completely filled with sediment. In XCB cores, GRA measurements can often be unreliable (unless the sample points are very carefully chosen) because of the disturbance caused by the mixing of drilling slurry and core biscuits.
Electrical resistivity of sediment cores has been measured during a number of ODP legs in an ad hoc manner on split cores. This has usually been done by inserting electrodes into the split core or by placing an array of electrode pads on the open surface of split cores. It has long been recognized that resistivity measurements provide important information that can be used to further understand sediment facies and geochemical processes. Combination logs of resistivity and density provide pertinent lithologic information (grain size/permeability/tortuosity) that cannot be achieved with other nondestructive measurements. Pore water salinity also influences the resistivity of sediments; hence, resistivity is valuable as a continuous log for geochemical purposes. However, it has proved a difficult parameter to measure in a fully automated, nonintrusive, and routine manner. The relatively new Geotek NCR system, which overcomes these difficulties, was installed for trial purposes during Leg 204 to test the possible value of this new sensor system.
Measurement of sediment resistivity using the NCR system can be made rapidly on a whole core in the plastic liner and, hence, is an ideal additional component to the MST. Consequently, during the first part of Leg 204, it was installed as an integral component of the MST for test purposes. It was not available for routine measurements until Hole 1251B but was then used at a number of sites during the remainder of the leg. The NCR technique operates by inducing a high-frequency magnetic field in the core from a transmitter coil, which in turn induces electrical currents in the core that are inversely proportional to the resistivity. A receiver coil measures very small magnetic fields that are regenerated by the electrical current. To measure these very small magnetic fields accurately, a difference technique has been developed that compares the readings generated from the measuring coils to the readings from an identical set of coils operating in air. This technique provides the accuracy and stability required. Resistivities between 0.1 and 10 
m can be measured at spatial resolutions along the core of ~2-4 cm. As with other parameters, the measurements are sensitive to core temperature and should be obtained in a stable temperature environment for best results.
Calibration was achieved by filling short lengths (~25 cm each) of core liner core with water containing known concentrations of NaCl. This provides a series of calibration samples with known resistivities that are then placed on the MST and logged. The logged results are illustrated in Figure F13A, which shows the raw output data in millivolts decreasing with decreasing salinity. The drop in output between the calibration sections illustrates the effect of having complete insulating gaps between the samples. Averaged values in each salinity are then plotted against the theoretical resistivity in Figure F13B, which provides a power-law calibration equation.
Transverse VP was measured on the MST track with the P-wave logger (PWL) for all cores at a routine sample interval of 2.5 cm. The PWL transmits a 500-kHz P-wave pulse through the core at a specified repetition rate (50 pulses per second). The transmitting and receiving ultrasonic transducers are aligned so that wave propagation is perpendicular to the core axis. Core diameter is measured using two displacement transducers that are mechanically linked to the ultrasonic transducers. The recorded velocity is the average of the user-defined number of acquisitions per location (10 during Leg 204). Calibrations of the displacement transducers and measurement of electronic delay within the PWL circuitry were conducted using a series of acrylic blocks of known thickness and P-wave traveltime. Repeated measurements of VP through a core liner filled with distilled water at a known temperature were used to check calibration validity. The Hamilton Frame discrete method of measuring velocity (PWS3) on the split track was used to occasionally check the MST values. In fact, we discovered a significant PWS3 measurement error (50 m/s), caused by a worn displacement transducer, which was corrected by adjusting the transducer to a new location.
The PWL was generally not used when cores were taken with the XCB, as the poor core quality precluded reliable velocity measurements. Normally, the undisturbed biscuits of primary sediments are surrounded by lower-velocity drilling slurry. Consequently, in XCB cores, we attempted to obtain velocity measurements by extracting undisturbed fragments of material and trimming them for use in the split-core PSW3 velocity system (see below).
Sediment VP was also measured on the split liner with the ODP standard Hamilton Frame PWS3 system, which is composed of two transducer pairs that take measurements transversely through the core liner or directly on sediment fragments ("chunks"). The system determines VP based on the traveltime of a 500-kHz wave between a pair of piezoelectric crystals separated by a variable distance (measured using displacement transducers). System accuracy was checked by measuring the velocity of acrylic samples. In addition, the PWS1 and PWS2 velocity sensors were used on selected cores to determine velocity anisotropy in the upper 10 mbsf, where gas expansion effects least affected the sediment quality.
Routine sampling frequency and location for P-wave measurements were generally coincident with the discrete MAD samples. Note that the velocity data stored in the Janus database are uncorrected for in situ temperature and pressure. These corrections can be made using the relationships outlined in Wyllie et al. (1956), Wilson (1960), and Mackenzie (1981). Poor core quality in deeper sections (generally in XCB cores) and abundant gas-expansion cracks and voids inhibited reliable measurements throughout Leg 204.
Thermal conductivity measurements on whole-core samples were made using the TK04 (Teka Bolin) system described by Blum (1997). Measurements were generally made once per core (generally within Section 3) but were increased to three per core where downhole temperature measurements were taken. The measurement system employs a single needle probe (Von Herzen and Maxwell, 1959) heated continuously in "full-space configuration." At the beginning of each measurement, temperatures in the samples were monitored automatically, without applying a heater current, until the background thermal drift was <0.04°C/min. Once the samples were equilibrated, the heater circuit was closed and the temperature rise in the probe was recorded. The needle probe contains a heater wire and calibrated thermistor. The probe is assumed to be a perfect conductor because of its high conductance relative to the core sediments. With this assumption, the temperature of the superconductive probe has a linear relationship with the natural logarithm of the time after the initiation of heating,
where
The thermal conductivity (k) was determined, using equation 6, by fitting the temperatures measured during the first 150 s of each heating experiment (for details see Kristiansen, 1982; Blum, 1997).
The reported thermal conductivity value for each sample is the average of three repeated measurements. Data are reported in W/(m·K), with measurement errors of 5%-10% in high-quality cores.
MAD parameters were determined from wet mass, dry mass, and dry volume measurements of split core sediments after method C of Blum (1997). Push-core samples of ~10 cm3 were placed in 10-mL glass beakers. In stiffer sediments drilled with the XCB where no push-core samples could be retrieved, samples were taken in small biscuits. Care was taken to sample undisturbed parts of the core and to avoid drilling slurry. Immediately after the samples were collected, wet sediment mass (Mwet) was measured. Alternatively, the sample was temporarily covered with parafilm to prevent any moisture loss prior to weighing. Dry mass and volume were measured after samples were heated in an oven at 105°C for 24 hr and allowed to cool in a desiccator. Sample mass was determined to a precision of 0.01 g using two Scientech 202 electronic balances and a computer averaging system to compensate for the ship's motion. The reference mass was always set to be within a 5-g margin of the actual sample weight. Sample volumes were determined using a helium-displacement Quantachrome penta-pycnometer with a precision of 0.02 cm3. Volume measurements were repeated five times, until the last two measurements exhibited <0.01% standard deviation. A reference volume was used to calibrate the pycnometer and to check for instrument drift and systematic error. After a calibration run for all five cells, two runs of measurements of five samples each were conducted before repeated calibration. The time required to run 10 samples, including a calibration, was ~2.5 hr. Sampling frequency was typically one per section, but it was decreased if uniform lithology was encountered throughout the core.
Moisture content, grain density, bulk density, and porosity were calculated from the measured wet mass, dry mass, and dry volume as described by Blum (1997). Corrections were made for the mass and volume of evaporated seawater using a seawater density of 1.024 g/cm3 and a salt density of 2.20 g/cm3.
Shear strength measurements are extremely difficult to obtain in the gas hydrate-bearing sediments encountered during this leg. Gas expansion and the presence of hydrate destroy the sediment fabric and do not allow for reliable shear strength measurements. Consequently, the automated shear vane equipment was not used except on selected cores, normally at shallow depth above 10 mbsf. However, we measured shear strengths using a handheld Torvane wherever the sediment conditions permitted. The device is quick to use and comes with three vane sizes: small, medium, and large (19, 25, and 48 mm diameter), providing maximum shear strength measurements up to 250, 100, and 20 kPa, respectively.
Most of the cores recovered during Leg 204 suffered significant disturbance caused by gas-expansion effects. Expansion of free gas, exsolution of dissolved gas, and dissociation of hydrate can cause gas expansion effects. It is worth noting here that a unit volume of free gas at 1000 m below the sea surface will expand by a factor of 100 by the time it reaches the surface (ignoring any temperature expansion or further exsolution). Another way of visualizing the expansion of gas during the core recovery process is to consider that a 0.1-mm-diameter free gas bubble in the sediment matrix at a depth of 1000 m will become a 4.5-mm-diameter gas bubble at the sea surface. When relatively low volumes of gas exsolve from pore fluids during core recovery, small bubbles will form in the sediment matrix with only minor amounts of core-volume expansion. As the gas volumes become greater, the sediment structure begins to fracture and a large amount of core expansion occurs, forming a series of large gas voids in the core. This process takes some time but is readily visible on the catwalk. When core expansion occurred, the core was generally pushed back together prior to sectioning, but inevitably these fractures remain in the core and affect the physical property measurements to some extent.
MS measurements are probably the least affected by core expansion, as these measurements depend only on the sediment/mineral volume. After the core has been pushed back together, the overall mineral volume changes over a 4-cm interval (the approximate spatial resolution of the sensor) is fairly small; therefore, significant effects are only seen where there are still clear gas gaps (greater than ~1 cm).
GRA measurements can be significantly affected by gas voids and sediment cracks. The downcore spatial resolution is ~1 cm, and hence, all cracks, no matter how minor, will show up in the GRA density profile. With significant degassing in many cores, this effect shows up as a very "noisy" data set. These data are, in fact, quite accurately representing the state of the core but of course do not represent the in situ condition. The closest representation of the in situ condition is achieved by taking the upper values of the GRA data envelope. However, even these values are often lower than the in situ densities because of gas in the matrix. The reader should see the data summaries and comparisons with downhole logs.
Gas expansion has the largest effect on the measurement of ultrasonic VP . Even very small amounts of free gas in the sediment matrix (<1%) will cause a significant decrease in velocity. More importantly, the very high attenuation of P-waves at ultrasonic frequencies in sediments with even very low volumes of free gas makes the measurement of velocity all but impossible. During Leg 204, we generally encountered gas-rich sediments in which VP measurements were possible only in the upper few meters. In practice, we observed that VP was generally only measurable above the sulfate/methane interface (SMI). Below the SMI, the gas exsolved generally prevented the acquisition of reliable data.
Gas expansion has little, if any, effect on the gravimetric determination of bulk density if the samples are carefully chosen. Small gas bubbles that form in the sediment matrix will not cause any effect on the wet-weight measurement or the dry-volume measurement and, hence, on the calculated bulk density and porosity values. The only time that gas expansion will have an effect on these measurements is when cracks in the sediment have caused premature drying. Even this effect is negated if the sampling is carried out promptly after core splitting.
Nearly all of the above effects of core disturbance caused by gas expansion do not occur when the physical property measurements are made at in situ pressures. This is one of the primary reasons for collecting pressure cores and analyzing them in the V-MSCL (see "Downhole Tools and Pressure Coring" for information about the use of the systems of pressure coring and to the appropriate site chapter where results are described in detail). Where comparisons between different data sets are appropriate, they are included in the site chapters.
To illustrate the pervasive nature of gas expansion throughout Leg 204, even in cores that did not apparently expand on the catwalk, we imaged an APC section through the liner before it was split. Figure F14 shows the nature of this cracking in a typical APC core (Section 204-1251G-1H-6). The image was obtained with the Geotek XY-DIS using a procedure known as "virtual slabbing." This is achieved by calibrating the camera on a white sheet of paper wrapped around the core rather than the flat white tile. A white calibration is achieved that compensates for the changes in illumination across the curved surface, providing a flat looking "slabbed" final image.
