**(GHMT) PROCESSING**

The goal of GHMT data processing is to obtain a magnetic reversal sequence that allows one to determine the magnetostratigraphy of a logged interval.

**GHMT measurement**

The GHMT consists of two sondes: SUMS (Susceptibility Magnetic Tool) and NMRS (Nuclear Resonance Magnetometer Tool). The former makes an induction-type measurement which records a signal related to formation susceptibility, while the latter is a high-precision nuclear magnetic resonance device, which measures the total magnetic field.

The magnetic induction (B) measured by the NMRS in a borehole is a function of location (p) and time (t) and can be expressed as a sum of different components (Pozzi et al., 1993):

B(p,t) = Br(p) + Ba(p) + Bt(p,t) + Bf(p)

where:

- Br(p): dipolar Earth's magnetic field

- Ba(p): anomaly field related to large scale heterogeneities

- Bt(p,t): transient variations of the magnetic field, of external origin

- Bf(p): field due to the magnetization (induced and remanent) of the sediment surrounding the borehole.

B(p,t) generally presents a shift towards high values at the top of the logged section; this effect is caused by the highly magnetized bottom hole assembly (pipe effect) and can be removed by modeling the pipe as a magnetic dipole.

**DATA PROCESSING**

After removal of the first three components and correction for pipe effect, the field due to the local magnetization (Bf(p)) is obtained. The susceptibility effect on the scalar total field magnetometer is calculated from the susceptibility tool (SUMS) and a transfer coefficient which depends on the geomagnetic location of the hole, the hole diameter, and a calibration ratio. To obtain the magnetostratigraphy from Bf(p), the susceptibility and total field measurements are combined to discriminate between induced and remanent magnetization. To obtain comparable data, the susceptibility effect and the remanent component are smoothed using Hanning filtering.

As the zero reference is not well defined and the volumes of investigation of the magnetometer and the susceptibility sonde are different, the polarity cannot be determined directly from the remanent component. The Koenigsberger coefficient, however, is supposed to be approximately constant for a given lithology. Then the slope of the mean square fit in a crossplot of the remanent versus the induced component may indicate the polarity of the paleomagnetic field (Barthès, 1991; Thibal, 1995). The correlation (positive slope) or anti-correlation (negative slope) between these two curves indicates respectively a normal or reverse polarity. The study of the correlation is performed in a sliding window versus depth. The length of a reversal and the time between two reversals is close to being random and the distance between them varies greatly. Therefore, the computation is performed several times by varying window heights and the results can be combined afterwards.

The number of window necessary for the interpretation is different for each hole and is chosen by the log analyst. A synthesis of the polarity sequence could be established considering that, for a given depth:

- if the slope is positive in every window, the final polarity is normal;

- if the slope is negative in every window, the final polarity is inverse

- if the slope is either positive or negative, the final polarity is undetermined.

The higher the amplitude of the slope, the stronger the correlation or anti-correlation between the curves.

This synthesis can be compared to a standard geomagnetic time scale to determine the magnetostratigraphy of the logged sediment.

**PROCESSED DATA**

For each hole logged with the GHMT, raw and processed GHMT data are displayed in an ASCII file. This file contains the following data from left to right:

- DEPTH (mbsf): sub-bottom depth

- MAGS (ppm^{ }SI): magnetic susceptibility, corrected for hole diameter

- BFI (nT): susceptibility effect

- BFIF (nT): filtered BFI

- MAGB (nT): raw total magnetic field

- BTCOR (nT): total magnetic field, corrected for pipe effect and present Earth’s magnetic field

- BTCORF (nT): filtered BTCOR

- REMA (nT): remanent component

- SLOPE (1-10): slope determined in the first to tenth window

Size of each window:

1 -> 11 samples = 1.52 m

2 -> 13 samples = 1.82 m

3 -> 17 samples = 2.43 m

4 -> 23 samples = 3.35 m

5 -> 31 samples = 4.57 m

6 -> 41 samples = 6.09 m

7 -> 53 samples = 7.92 m

8 -> 67 samples = 10.05 m

9 -> 88 samples = 13.25 m

10 -> 101 samples = 15.24 m

**BIBLIOGRAPHY**

- Barthès V., 1991. Traitement et interprétation des données GHMT. Rapport interne CEA/LETI/DSYS/SESA.

- Pozzi J.-P., Barthès V., Thibal J., Pocachard J., Lim M., Thomas T., and Pagès G., 1993. Downhole magnetostratigraphy in sediments: comparison with the paleomagnetism of a core. J. Geophys. Res., 98, 7939-7957.

- Thibal J., 1995. Analyse de l'aimantation des sédiments par diagraphies magnétiques. PhD thesis, Univ. of Paris XI Orsay.