METHODS

The U.S. Department of Energy provided funding support to deploy the proVision tool during Leg 204. The basic technology behind this tool is similar to modern wireline nuclear magnetic resonance (NMR) technology (Kleinberg et al., 2003; Horkowitz et al., 2002), based on measurement of the relaxation time of the magnetically induced precession of polarized protons. A combination of magnets and directional antennas are used to focus a pulsed, polarizing field into the formation. Figure F3 shows a schematic of the proVision tool. The antennas on the proVision tool measure the relaxation times of polarized molecules in the formation, which can be used to assess sediment porosities.

During Leg 204, the proVision tool acquired formation and engineering information in memory and transmitted some data to the surface via MWD. The relaxation time spectra were recorded downhole and porosity estimates were transmitted to the surface in real time. These spectra were stacked in postprocessing to improve the measurement precision. Data were also acquired while pulling the tool upward (sliding, not rotating) over short open-hole intervals to compare measurements with and without the effect of lateral vibrations induced while drilling and rotating.

In recent years there have been significant developments in the field of NMR well logging (reviewed by Horkowitz et al., 2002). Similar to neutron porosity devices, NMR tools primarily respond to the presence of hydrogen molecules in the pore fluids in rock formation. Unlike neutron porosity tools, however, NMR tools use the electromagnetic properties of hydrogen molecules to analyze the nature of atomic interactions within pore fluids. Relative to other pore-filling constituents, gas hydrates exhibit unique chemical structures and hydrogen concentrations. In theory, therefore, it should be possible to develop NMR well logging evaluation techniques that would yield accurate reservoir porosities and water saturations in gas hydrate–bearing sediments.

Under the effect of a strong magnetic field, hydrogen nuclei tend to align with the induced magnetic field. A certain amount of time, called the longitudinal magnetization decay time (T₁), is required for this alignment. When the magnetic field is pulsed, the hydrogen nuclei returns to a disordered state with a characteristic relaxation time, called the transverse magnetization relaxation time (T₂). T₂ depends on the relaxation characteristics of the hydrogen-bearing substances in the rock formation. For example (Fig. F4), T₂ for hydrogen nuclei in solids is very short, however, T₂ for hydrogen nuclei in fluids can vary from tens to hundreds of milliseconds depending on fluid viscosities and interactions with nearby surfaces (reviewed by Kleinberg et al., 2003).

When deployed in the LWD tool configuration, the NMR measurement represents the in situ NMR properties of hydrogen in the formation. Initially, the hydrogen atoms are aligned in the direction of a static magnetic field (B₀). The hydrogen atoms are then tipped by a short burst from an oscillating magnetic field that is designed so that they precess in resonance in a plane perpendicular to B₀. The precession of the hydrogen atoms induces a signal in the tool's antenna, and the decay of this signal is measured from the echo amplitudes in the pulse sequence from which T₂ is calculated. Because the formation contains hydrogen in different forms (in water in large pores and small pores, bound in clay minerals, and in gas hydrate), there is a distribution of T₂ times; for Leg 204 sites, T₂ times were recorded in a range set between 3 ms and 3 s.

The proVision signal investigates an ~15-cm cylindrical volume of the borehole, and for a 9 -in bit size, the depth of investigation of the measurement is ~5 cm into the formation. In most cases, drilling with the LWD tools proceeded at ~25 m/hr. Using this relatively slow average penetration rate, enhanced NMR spectral resolution and a data sampling rate of approximately one sample per 15-cm depth interval was achieved. This high data density improved NMR spectral resolution after postprocess stacking and enhanced the overall logging data quality.

Lateral tool motion may reduce proVision data quality in some circumstances. Thus, accelerometers and magnetometers are contained in the tool to measure downhole motion and to evaluate data quality. The maximum resolvable relaxation time can be computed from this information. In addition, after reaching the total depth in three of the LWD holes on Hydrate Ridge (Holes 1244D, 1246A, and 1250B), the LWD tools were pulled upward while sliding (without rotating) for ~30 m to compare the measurements while drilling downward over the same interval. Comparing the proVision spectra with and without the effect of the lateral vibration due to drilling allowed further refinement of the measurement resolution. Table T1 shows the holes in which sliding tests were conducted. The sliding data were not processed; they were used by the logging engineer to verify the quality of the proVision spectra being collected during drilling.

The proVision data quality is high throughout most of the logged interval in all nine holes drilled during Leg 204. ProVision data quality is degraded, however, when the distance between the tool sensor and the wall of the borehole is greater than 1 in. To evaluate this, the differential caliper log (DCAL) is recorded by the LWD density tool and provides a measure of this distance. DCALs from the LWD drill holes measure values <1 in over 90%–95% of the total interval drilled in all nine LWD holes. In each hole, however, the uppermost 10–30 meters below seafloor (mbsf) was typically washed out to >1 in because of drilling disturbance of the softer subseafloor sediments. Deeper intervals in only one hole (Hole 1244D) show deflections of the DCAL measurement of up to 1 in where borehole breakouts occurred below 250-m depth. NMR measurements may be degraded in these particular intervals.

During ODP, some shipboard data processing is conducted in order to verify data content and quality. In addition, both real-time and memory data from LWD and conventional wireline logging data are transmitted via Inmarsat B satellite from the JOIDES Resolution to LDEO and, in some cases, then transferred to Schlumberger for reprocessing and data quality assessment. The general data processing for the LWD logging data from Leg 204 followed the steps as described below (for more detailed description of the Leg 204 logging data processing effort, see the processing documentation notes under each site entry on the ODP Logging Services Web site: http://www.ldeo.columbia.edu/BRG/ODP/DATABASE/DATA/search.html):

1. Depth shift: Original logs are first depth-shifted to the seafloor. The seafloor depth was determined by the step in gamma ray and resistivity values at the sediment/water interface.
2. Neutron porosity data processing: The neutron porosity measurements are corrected for bit size, temperature, mud salinity, and mud hydrogen index (mud pressure, temperature, and weight).
3. Density data processing: Density data are processed to correct for the irregular borehole using a technique called "rotational processing," which is particularly useful in deviated or enlarged boreholes with irregular or elliptical shapes. This statistical method measures the density variation while the tool rotates in the borehole, estimates the standoff (distance between the tool and the borehole wall), and corrects the density reading.
4. Resistivity data: The resistivity curves are sampled at a 0.0304-m (1.2 in) sampling rate.
5. NMR: The T₂ distribution is the basic output of NMR measurement. It is further processed to give the total pore volume (the total porosity) and pore volumes within different ranges of T₂, such as the bound- and free-fluid volumes. However, it is important to note that in a gas hydrate–bearing section, the NMR tool only measures the liquid-filled portion of the porosity. The NMR tool used during Leg 204 was an experimental tool; the processing was performed on shore by Schlumberger.