The application of the previously validated algorithm to the magnetic records from the GPIT on the FMS-sonic tool string allows us to compare our filtering algorithm with the standard algorithm (here referred to as the Schlumberger algorithm) routinely used to compute tool orientation during ODP legs. Magnetic logs of FMS-sonic tool string passes 1 and 2 are of excellent quality, as attested by the repeatability of the intensity and inclination of the local total field (Figs. F4A, F4B, F6A, F6B). Figure F6C shows the tool declination (tdec) of FMS-sonic passes 1 and 2, integrating total field declination (Dec) and the rotational component (Rot). From the bridge located at ~560 mbsf, guided by hole shape (Fig. F6D) and cable properties, both passes use the same path, thus restricting the study to the first pass. Figure F6E–F6H compares the tool string rotational components isolated using our proposed algorithm (cosine filter of 4 m width and boxcar filter of 10 m width) and that of the Schlumberger algorithm. As attested by the residual (
rot) between these two components (Fig. F6F, F6H), results are in good accordance (mostly within 5°) and validate the use of the Schlumberger algorithm in such an environment. A comparison of the declination obtained at this site using both algorithms is given in Figure F4D.
The two-step demonstration above shows that independent of the geometric parameters of the tool string (length, diameter, and number of points of contact), the proposed algorithm can be validated by comparing the GBM direct (optical-gyro based) rotation record with a filtered version of its bulk magnetic parameters. This well-documented algorithm was next compared with the standard Schlumberger algorithm, initially developed for weakly magnetized (sedimentary) formations. The good agreement between results validates the use of the latter in magnetically contrasted and layered formations. As an additional test for the validity of these algorithms, we applied both to magnetic records from reference Hole 735B, which were acquired using other tool string configurations.
Hole 735B was logged during ODP Legs 118 and 176, providing a common logged section of more than 500 m. During Leg 118, downhole magnetic measurements were performed with the same U.S. Geological Survey (USGS) susceptibility probe used during ODP Leg 102 (Bosum and Scott, 1988) and the nonoriented magnetometer in the Schlumberger GPIC run in conjunction with the Schlumberger Lithodensity Tool (LDT). During ODP Leg 176, magnetic data were recorded by the Schlumberger GPIT, first run in conjunction with the Dual Resistivity Lateral Log (DLL) tool string and then with the FMS-sonic tool string. Because digital records of the USGS magnetometer are not available, direct validation as in the previous data set is impossible. However, the set of magnetic logs recorded by the GPI modules gives us the opportunity to compare the standard Schlumberger algorithm used at this time with the filtering algorithm previously validated.
The lithostratigraphy of Site 735 (50–600 mbsf) is presented in detail in Dick et al. (1991) for the Leg 118 cores and Dick, Natland, Miller, et al. (1999) for the Leg 176 cores. Gabbroic rocks constitute >99 vol% of the total section. This section is composed of five principal gabbroic units and more secondary intrusive units, a number of them late small ferro-gabbro intrusions (Fig. F7A). The igneous stratigraphy is primarily controlled by the interconnection between deformation, magma segregation, and crystallization (Dick, Natland, Miller, et al., 1999). Shipboard laboratory measurements on minicores from Hole 735B (core recovery = 86.6%) indicated that the gabbroic rocks have a mean MS of 23 x 10–3 SI (maximal value ~ 10–1 SI) and a mean NRM of 2.5 A/m (26 x 10–5 to 130 A/m). In accordance with core measurements (Fig. F7B, dots), the MS log (Fig. F7B, continuous line) made with the USGS susceptibility tool (30–490 mbsf) indicates that the MS of gabbro in the upper 215-m section of the hole is extremely variable, with one thin zone approaching 0.5 x 10–3 SI and many thin anomalies peaking out between 10 x 10–3 and 25 x 10–3 SI, but with many other values approaching zero. Midway downhole (centered at ~245 mbsf), a 60-m zone of high MS (~30 x 10–3 to 55 x 10–3 SI) occurs in a gabbroic section containing anomalously high concentrations of magnetite and ilmenite. Below the magnetite-ilmenite-rich gabbros, a 100-m-thick magnetically quiet zone occurs, with MS averaging 1 x 10–3 to 10 x 10–3 SI. Below this zone, another highly variable 100-m-thick interval occurs near the bottom of the hole, with MS in the range of 5 x 10–3 to 40 x 10–3 SI (Pariso et al., 1991; Kikawa and Pariso, 1991). This interval is also characterized by high magnetite and ilmenite concentration in the drill cores.
Figure F7C shows that the intensities of the total magnetic field for both the GPIC (Leg 118, in red) and the GPIT (Leg 176, FMS in blue and DLL in green) have the same shape and relative amplitudes but are offset by a constant. The shift between the two tool generations could have resulted from calibration problems and/or imperfect magnetic isolation of the GPIC. Although in sedimentary sections, total magnetic intensity and inclination are close to the expected values at this latitude (38,348 nT and –61.5°, respectively), two significant increases in the intensities as well as changes in the inclination logs can be observed in oxide-rich sections (Fig. F7C, F7D): the first between 220 and 285 mbsf and the second between 410 and 510 mbsf.
Figure F7E presents the tool declination (tdec) log, which contains both the rotational component and declination of the local magnetic field, as recorded by the three tool strings. For these tool strings, isolated tool rotation (rot) deduced by our filtering operation is compared with that deduced from the Schlumberger algorithm (Fig. F7F–F7H). For the two algorithms and independent of the lithology, the short and no-contact-point DLL tool string continuously rotated (mean rate of rotation = 1 rotation/~25 m) over the whole section (~60–580 mbsf). On the contrary, because of friction between the tool string and borehole walls, the long and one-contact-point (centralizer) LDT tool string as well as the long and multiple-contact-points (four pads and centralizer) FMS-sonic tool string did not rotate significantly during the logging operation. This small rate of rotation may also be indicative of hole ellipticity, deviation, or directional borehole damage caused by the extensive fishing operations following the break-off of drill pipe in the hole. Independent of the tool string geometry, residuals (rot) between the developed algorithm and the Schlumberger algorithm are <6° (not shown), thus confirming the use of the Schlumberger algorithm in such formations. On the other hand, observed tool rotations at this site confirm the initial finding that tool rotation is intrinsically of higher wavelength (approximately tool string length) than the local fluctuations of the magnetic field (meter scale). In such cases, filtering techniques using filters between 2 and 10 m length lead to correctly oriented borehole wall images.
