instrument capable of detecting natural remanent magnetizations of less than 10
−12 A/m, on cylindrical samples of
approximately 10 cm
3 taken from sedimentary or volcanic
formations and as perfectly oriented in space as possible. This
orientation in space of the sample collected in the field is
essential in order to determine the direction of the magnetization vector (represented by inclination and declination) and
then to calculate the position (longitude and latitude) of the
virtual magnetic pole associated with it. To obtain a reliable
virtual magnetic pole, several criteria must be met. First of
all, one needs to have a sufficient number of samples. The
number of samples collected is usually eight per site, and the
number of sampling sites within a geological formation is at
least six sites. As in the processing of the signal, increasing
the number of sites and samples per site improves the
signal-to-noise ratio (thus reducing the sources of errors). The
uncertainty on the virtual magnetic pole is of the order of a
few degrees (1° = 111 km) in the best case, which still represents a few hundred kilometers. The displacement of continent of less than several hundred kilometers is therefore
difficult to detect, which, taking into account the average
speed of plate drift (*4 cm/year), represents a time span of
about ten million years. The dating of the sampled geological
formations must also meet a standard of robustness in order
to produce a reliable virtual magnetic pole. The primary
character of the magnetization of the ferromagnetic carriers
must be verified. The magnetization signal must be synchronous or recorded in the millions of years (<5 Ma) after
diagenesis (in the case of sedimentary formations). This is to
ensure that the magnetization age is essentially the same as
the “stratigraphic” age of the formation. When these two ages
differ, a subsequent re-magnetization has occurred (in
response, for example, to the burial at several kilometers
deep, which has the effect of imprinting a new magnetic
signal in the sample masking partially or wholly the primary
magnetization).
Tests have been developed to evaluate the quality of the
measured primary natural remnant magnetization (NRM).
The first is the reversal test. On average, the magnetic field
experiences several reversals per million years (there are,
nevertheless, periods of several tens of millions of years
during which no reversal is observed, Chap. 7). Over a time
period of less than a few million years (<5 Ma), the calculated virtual magnetic pole must be statistically identical,
regardless of the polarity of the magnetic field. This test
increases the confidence in the primary character of the
NRM and ensures that the secular variation of the magnetic
field resulting from the movements of molten iron in the
outer core has indeed been averaged. The secular variation
of the magnetic field gives rise to a rapid drift of the magnetic pole which can move the virtual pole by more than 10°
from its mean position. Thus, the direction provided by a
single lava flow cooled in less than one year does not
correspond to the average direction of the dipole. A second
test consists of checking the reliability of the magnetic pole
determined from sampling sites for which the dip in the
geological layers differs. Indeed, the geological formations
can be subjected to deformations. To calculate the virtual
magnetic pole, the paleo-horizontality of the sampled sedimentary formation that prevailed at the moment of the
acquisition of the magnetic signal must be restored. Taking
into account the tectonic (tilt) correction for each sampling
site must have the effect of clustering the set of magnetic
directions obtained for each site. If this structural correction
does not have the desired effect on the data, this means that
the magnetization was acquired during or after deformation.
However, these tests are not sufficient to certify the
robustness of the data that can be affected by various biases.
Regional tectonics in areas of active collisions such as in the
Alps or by rifting in Afar could lead to horizontal block
rotation that is well recorded by paleomagnetism. These
regional tectonics can mask the large drifts of plates or larger
masses. A second bias is related to the geometry of the
magnetic field. The calculation of a virtual magnetic pole is
based on the assumption of a geocentric axial dipole. In the
case of samples collected from sedimentary formations, the
magnetic measurements are carried out on cylinders of
approximately 10 cm
3 . Given the low accumulation rate of
sediments and the effects of diagenesis (such as compaction),
a cylinder of 10 cm
3 may represent a time period of several
thousand years, and the measured magnetic direction is in
some sense an “average” direction of this time period. For a
sampling site, the magnetic direction is the average of the
magnetic directions of the samples taken from a unit of a
sedimentary formation. It is therefore possible that the time
interval associated with the mean magnetic direction represents several thousand years, or even more. The variations of
the magnetic field are thus smoothed, and the magnetic
direction measured is indeed that produced by a geocentric
axial dipole field in the study area under consideration.
In the case of volcanic series, the thermoremanent magnetization (TRM) is acquired during the cooling of the lava flow.
It thus represents a (quasi-) instantaneous photography of the
Earth’s magnetic field. To overcome the effect of the secular
variation, the average direction of a large number of sites must
be measured, so as to tend towards the direction which would
be obtained with an axial geocentric dipole. However, the
presence of a persistent quadrupole term in the paleomagnetic
data has been identified as a possible source of errors. The
presence of quadrupole terms of about a few percent (the level
generally observed) implies an error of a few degrees in latitude
at the equator. This does not therefore significantly affect the
virtual magnetic pole which is calculated assuming the axial
geocentric dipole hypothesis. There is also a possibility that the
magnetic field could include an octupole component affecting
Asia during the Tertiary and Pangea during the Permian. If we
28
F. Fluteau and P. Sepulchre
−12 A/m, on cylindrical samples of
approximately 10 cm
3 taken from sedimentary or volcanic
formations and as perfectly oriented in space as possible. This
orientation in space of the sample collected in the field is
essential in order to determine the direction of the magnetization vector (represented by inclination and declination) and
then to calculate the position (longitude and latitude) of the
virtual magnetic pole associated with it. To obtain a reliable
virtual magnetic pole, several criteria must be met. First of
all, one needs to have a sufficient number of samples. The
number of samples collected is usually eight per site, and the
number of sampling sites within a geological formation is at
least six sites. As in the processing of the signal, increasing
the number of sites and samples per site improves the
signal-to-noise ratio (thus reducing the sources of errors). The
uncertainty on the virtual magnetic pole is of the order of a
few degrees (1° = 111 km) in the best case, which still represents a few hundred kilometers. The displacement of continent of less than several hundred kilometers is therefore
difficult to detect, which, taking into account the average
speed of plate drift (*4 cm/year), represents a time span of
about ten million years. The dating of the sampled geological
formations must also meet a standard of robustness in order
to produce a reliable virtual magnetic pole. The primary
character of the magnetization of the ferromagnetic carriers
must be verified. The magnetization signal must be synchronous or recorded in the millions of years (<5 Ma) after
diagenesis (in the case of sedimentary formations). This is to
ensure that the magnetization age is essentially the same as
the “stratigraphic” age of the formation. When these two ages
differ, a subsequent re-magnetization has occurred (in
response, for example, to the burial at several kilometers
deep, which has the effect of imprinting a new magnetic
signal in the sample masking partially or wholly the primary
magnetization).
Tests have been developed to evaluate the quality of the
measured primary natural remnant magnetization (NRM).
The first is the reversal test. On average, the magnetic field
experiences several reversals per million years (there are,
nevertheless, periods of several tens of millions of years
during which no reversal is observed, Chap. 7). Over a time
period of less than a few million years (<5 Ma), the calculated virtual magnetic pole must be statistically identical,
regardless of the polarity of the magnetic field. This test
increases the confidence in the primary character of the
NRM and ensures that the secular variation of the magnetic
field resulting from the movements of molten iron in the
outer core has indeed been averaged. The secular variation
of the magnetic field gives rise to a rapid drift of the magnetic pole which can move the virtual pole by more than 10°
from its mean position. Thus, the direction provided by a
single lava flow cooled in less than one year does not
correspond to the average direction of the dipole. A second
test consists of checking the reliability of the magnetic pole
determined from sampling sites for which the dip in the
geological layers differs. Indeed, the geological formations
can be subjected to deformations. To calculate the virtual
magnetic pole, the paleo-horizontality of the sampled sedimentary formation that prevailed at the moment of the
acquisition of the magnetic signal must be restored. Taking
into account the tectonic (tilt) correction for each sampling
site must have the effect of clustering the set of magnetic
directions obtained for each site. If this structural correction
does not have the desired effect on the data, this means that
the magnetization was acquired during or after deformation.
However, these tests are not sufficient to certify the
robustness of the data that can be affected by various biases.
Regional tectonics in areas of active collisions such as in the
Alps or by rifting in Afar could lead to horizontal block
rotation that is well recorded by paleomagnetism. These
regional tectonics can mask the large drifts of plates or larger
masses. A second bias is related to the geometry of the
magnetic field. The calculation of a virtual magnetic pole is
based on the assumption of a geocentric axial dipole. In the
case of samples collected from sedimentary formations, the
magnetic measurements are carried out on cylinders of
approximately 10 cm
3 . Given the low accumulation rate of
sediments and the effects of diagenesis (such as compaction),
a cylinder of 10 cm
3 may represent a time period of several
thousand years, and the measured magnetic direction is in
some sense an “average” direction of this time period. For a
sampling site, the magnetic direction is the average of the
magnetic directions of the samples taken from a unit of a
sedimentary formation. It is therefore possible that the time
interval associated with the mean magnetic direction represents several thousand years, or even more. The variations of
the magnetic field are thus smoothed, and the magnetic
direction measured is indeed that produced by a geocentric
axial dipole field in the study area under consideration.
In the case of volcanic series, the thermoremanent magnetization (TRM) is acquired during the cooling of the lava flow.
It thus represents a (quasi-) instantaneous photography of the
Earth’s magnetic field. To overcome the effect of the secular
variation, the average direction of a large number of sites must
be measured, so as to tend towards the direction which would
be obtained with an axial geocentric dipole. However, the
presence of a persistent quadrupole term in the paleomagnetic
data has been identified as a possible source of errors. The
presence of quadrupole terms of about a few percent (the level
generally observed) implies an error of a few degrees in latitude
at the equator. This does not therefore significantly affect the
virtual magnetic pole which is calculated assuming the axial
geocentric dipole hypothesis. There is also a possibility that the
magnetic field could include an octupole component affecting
Asia during the Tertiary and Pangea during the Permian. If we
28
F. Fluteau and P. Sepulchre