Gubbins D., Herrero-Bervera E. Encyclopedia of Geomagnetism and Paleomagnetism

Подождите немного. Документ загружается.

When examining a time sequence such as the GPTS it is crucial that

the observed events be ordered correctly. This, together with other dif-

ficulties, has led to severe problems with the land-based polarity

sequence. As one goes further back in time, the absolute error in dating

a rock soon becomes larger than the length of the shorter polarity inter-

vals. Because the land-based information comes from combined obser-

vations of magnetic polarity from rocks in widely spaced localities

worldwide and not from a single continuous sequence, the ordering

relies on the accuracy of the K-Ar dates. Hence, the ordering of events

is effectively indeterminate for events closer together than the dating

error. Consequently the length of the reliable land-based polarity time-

scale has been too short for reliable estimation of the statistical struc-

ture of the GPTS. Recently however, Kent and coworkers (e.g., Kent

and Olson, 1999) have shown the potential for GPTS extension in

thick, complete continental sedimentary sections.

The most useful data source for development of the GPTS has been

marine magnetic anomalies, and the reversal chronology is now quite

well determined for the past 165 Ma. The absolute error in the age

assigned to individual reversals will still often be greater than the

length of the shorter intervals, but the continuous nature of the data

source means that the ordering of events will be correct. Furthermore,

because of the way ages are assigned in the marine magnetic anomaly

timescales, the lengths of the intervals between reversals will typically

be about right.

Naturally though, there are difficulties and problems with the mar-

ine magnetic anomaly timescales. As we go further back in time the

record is more degraded, so the timescale from 120 Ma back to about

165 Ma is less reliable than that from the present back to 80 Ma.

Furthermore, because of subduction of the seafloor, any record prior

to about 170 Ma has been destroyed.

A major problem with marine magnetic anomalies is that it is fairly

difficult to identify all of the very short intervals, particularly if a short

interval appears between two relatively long intervals of the opposite

polarity. This had led to significant problems in interpreting the GPTS.

Despite the problems in the marine magnetic anomaly timescales,

these provide us with our best GPTS and so analyses of the statistical

structure of the geomagnetic reversal sequence have typically been

performed on these timescales.

Releva nt probabil ity distrib utions

It has been established that individual reversals take only a few thou-

sand years to complete, which is a short time relative to the average

interval between reversals. Thus it is a reasonable approximation to

assume that the reversals themselves are instantaneous relative to the

time constants of interest. It has already been noted that there appears

to be a random component in the reversal process, and so it is sensible

to look for a probabilistic description. Consider an interval of time D x

short enough that the probability of having two reversals in D x is neg-

ligible, but with probability of l D x of having a single reversal in D x.

This is then a Poisson process, and it is a simple matter to show that

the probability density p( x) of interval lengths x between reversals is

given by

p ð xÞ dx ¼ le

l x

d x; (Eq. 1)

where l is the rate of the process and m ¼ 1/l is the mean interval

length.

It is possible to test if the observed distribution of interval lengths is

compatible with the Poisson distribution. Initially it was concluded

that there was indeed compatibility but after further testing it was sug-

gested that there were too few short intervals in the observed record.

Naidu (1971) showed that a gamma distribution provided a good fit

to the then observed intervals of the Cenozoic timescale and Phillips

(1977) confirmed this in an extensive study of geomagnetic reversal

sequences. For a gamma process the probability density p (x) of inter-

val lengths x is given by

pð xÞ d x ¼

G ðk Þ

ðk LÞ

ðk  1Þ

k L x

d x

G ðk Þ

ðk  1Þ

l x

d x; l ¼ k L (Eq: 2)

where the mean interval length is now given by m ¼ 1/ L ¼ k/ l. G (k)is

the gamma function of k, given by

Gð k Þ¼

k  1

 z

d z (Eq. 3)

If k is an integer, then this is just the factorial function of k– 1, i.e.,

G (k ) ¼ (k– 1)!.

From Eqs. (1) and (2) it is apparent that the gamma process leads

to a family of distributions depending on the value of k, and that the

Poisson process is simply the special case of a gamma process with

k ¼ 1. Figure G38 shows appropriately scaled probability densities

for this family of distributions. It is immediately apparent that a

gamma process with k > 1 has far fewer very short intervals than a

Poisson process.

Statistical tools for analyzing the reversal sequence in terms of a

renewal process are given by McFadden (1984a) and McFadden and

Merrill (1986, 1993).

Impl ications of a gamma or Poiss on process

Consider the general probability of a reversal occurring at an interval

D x . Let x be the interval of time since the most recent reversal (i.e.,

no reversals have occurred in the interval 0 to x), and let f (x)D x be

the probability that a reversal will then occur in the interval x to

(x þ D x). Hence the function f(x) describes how the instantaneous

probability for the occurrence of another reversal varies with the time

since the last reversal and is given by

f ðxÞ¼

pðxÞ

1 

pðtÞdt

(Eq. 4)

Figure G39 shows f (x)/l plotted as a function of (xl). For a Poisson

process the occurrence of a reversal has no impact on the probability

of a future reversal. That is, the system has no memory. For k >1,

the probability for a future reversal drops to 0 as soon as a reversal

Figure G38 Probability densities of gamma distributions

plotted against the interval length scaled to the mean length.

The k ¼ 1 curve represents a Poisson distribution, which has a

relatively large number of shorter intervals compared with the

other distributions.

336 GEOMAGNETIC REVERSAL SEQUENCE, STATISTICAL STRUCTURE

occurs and then gradually rebuilds to its undisturbed value. That is,

the system has a memory of the previous reversal and this memory

causes inhibition of future events by depressing their probability of

occurrence.

Is the process gamma or Poisson?

As noted above, there may be too few very short intervals in the

observed timescale for a Poisson process, but a gamma distribution

provides a good fit. Thus it may initially appear as a clear case in favor

of a general gamma process, and indeed this was felt to be so for some

time. Analysis of early timescales showed that k was quite different for

the normal and reverse polarity sequences: at times it was about 4 for

the normal polarity sequence while it was close to unity for the reverse

polarity sequence (from about 40 to about 25 Ma); and in recent times

(the past 15 Ma) it was about 2½ for the reverse sequence while close

to unity for the normal sequence. This suggested a substantial asym-

metry between normal and reverse polarity, which was surprising

when theoretical considerations strongly implied that there should be

no asymmetry (see above and Geomagnetic field, asymmetries).

However, McFadden (1984a) showed that these estimates of k were

not robust and that the actual reversal process is much more likely to

be nearly Poisson. Obtaining a reliable GPTS from the marine mag-

netic anomaly record is not a trivial matter: the major problems relate

to accurate dating of the individual events and reliable recognition of

the shorter intervals. A small error in the dating of an individual rever-

sal has small consequences. By contrast, there are significant conse-

quences when a short polarity interval is actually missed. When a

short interval is not identified it is, in effect, combined with the preced-

ing and succeeding intervals of the opposite polarity. This means that

the short interval is missed from its own polarity sequence thereby

tending to increase the apparent value of k for that polarity sequence.

At the same time, it incorrectly produces a long interval of the opposite

polarity that is the sum of the interval preceding the missed short inter-

val, the short interval itself, and the succeeding interval.

McFadden (1984a) has shown that if n intervals from a gamma pro-

cess with index k are joined together then the resulting interval appears

to have been drawn from a gamma process with index nk. Therefore, if

the process is Poisson and a short interval is missed, the resulting long

interval of the opposite polarity appears to be an observation drawn

from a gamma process with k ¼ 3. Hence the parameter k can be a

fairly sensitive indicator of polarity intervals that have been missed.

McFadden and Merrill (1984) concluded that k for the observed

sequence was about 1.25 for the period from about 80 Ma to the

present. This suggested that several short intervals had not been iden-

tified in the GPTS. McFadden and Merrill (1993) discussed the possi-

ble alternative that the reversal process is actually gamma and showed

that this would imply that following a reversal the probability for

another reversal would be depressed for about 50 ka.

The observed sequence is statistically similar to a sequence that

would be obtained by taking a Poisson process with the appropriate

rate l and filtering the sequence so that any interval less than 30 ka

in duration is incorporated into the surrounding intervals of opposite

polarity. This is consistent with the conclusion of Parker (1997) that

the maximum resolution of the GPTS from marine magnetic anomalies

is about 36 ka.

Overall it is probably safe to conclude that the reversal process is

either Poisson or nearly Poisson, that there is no asymmetry between

the normal and reverse sequences, and that the polarity sequence can

be considered as a single sequence without regard to the actual polarity

of the individual intervals.

Nonstationarity in the reversal process

Recently the reversal rate has been about 4.5 Ma–1, around 40 Ma it

was about 2 Ma–1, and from about 83 Ma back to about 119 Ma there

were no reversals, the field having normal polarity. Clearly the rate at

which the reversal process occurs has not been constant through time.

This is of central geophysical interest because the existence of nonsta-

tionarity implies a change in the properties of the origin of the process,

typically in the boundary conditions.

Estimates of the reversal rate l, using the reversal chronology of

Cande and Kent (1995) for the interval 0 to 118Ma and that of Kent and

Gradstein (1986) for the interval 118–160 Ma, are shown in Figure G40.

These estimates have been obtained using a sliding window containing

50 polarity intervals. The Cretaceous Superchron, an interval of about

36 Ma from about 118 to 82 Ma when the polarity was normal and there

were no reversals, is very obvious in the sequence and requires explanation.

The characteristic time of these changes is the same as that asso-

ciated with changes in boundary conditions imposed by the mantle

on the outer core (e.g., Jones, 1977; McFadden and Merrill, 1984,

1993, 1995, 1997; Courtillot and Besse, 1987). Suggestions that spatial

variations in the core-mantle boundary conditions can affect the reversal

rate are now supported by some dynamo theory (e.g., Glatzmaier et al.,

1999). The variation in l shown in Figure G40 led to the interpretation

first articulated by McFadden and Merrill (1984), that changes in the

core-mantle boundary conditions gradually slowed the reversal process

Figure G40 Estimated reversal rate l for the past 160 Ma.

Constructed from the timescales of Kent and Gradstein (1986) and

Cande and Kent (1995), following the methods of McFadden

(1984a). From Merrill et al. (1996).

Figure G39 Illustration of the memory (inhibition) in a gamma

process and the absence of a memory in the special case of a

Poisson process (k ¼ 1). After Merrill et al. (1996).

GEOMAGNETIC REVERSAL SEQUENCE, STATISTICAL STRUCTURE 337

until eventually the process ceased, creating the superchron without any

reversals. Gradually the boundary conditions became more favorable for

reversals until eventually the reversal process restarted and gradually sped

up to its current rate.

Gallet and Hulot (1997), using the same timescales, proposed an

alternative nonstationarity. They suggested that the reversal rate had

been essentially constant from about 158 to 130 Ma and from about

25 Ma to the present, with an intermediate nonstationary segment

including the Cretateous Superchron. However, McFadden and Merrill

(2000) developed a statistical test that showed that the data demanded

the trends identified in Figure G40. Subsequently, using the methodol-

ogy developed by McFadden and Merrill (2000), Hulot and Gallet

(2003) analyzed the Gradstein et al. (1994) and Channell et al.

(1995) timescales. These timescales update the pre-superchron infor-

mation. Hulot and Gallet (2003) concluded that these newer timescales

show no long-term trend in reversal rate leading into the superchron.

Indeed, they suggest that the only precursor to the superchron was

perhaps a single unusually long interval (CM1n) just before the

superchron. If this is confirmed by future timescales then the interpre-

tation of the causes of the superchron will naturally change.

There is some fine structure observable in Figure G40, and the ques-

tion arises as to whether this structure is meaningful. Several authors

(Mazaud et al., 1983; Negi and Tiwari, 1983; Raup, 1985; Stothers,

1986; Marzocchi and Mulargia, 1990; Mazaud and Laj, 1991; Rampino

and Caldeira, 1993) have suggested either a 15- or a 30 Ma periodicity in

the reversal chronology record. Hulot and Gallet (2003) have recently

suggested a 20 Ma periodicity. Clearly the perceived periodicity is not

robust. McFadden (1984b) showed that similar apparent periodicities

are produced when using fixed-length sliding windows to analyze a

Poisson process with a linear trend in the reversal rate, but that such per-

iodicities are not observed when using sliding windows with a fixed

number of intervals (McFadden and Merrill, 1984). Lutz (1985), Stigler

(1987), McFadden (1987), and Lutz and Watson (1988) all showed that

the perceived periodicities are more likely an artifact of the methods of

analysis than real geophysical phenomena.

Acknowledgment

This paper is published with the permission of the Chief Executive

Officer, Geoscience Australia.

Phillip L. McFadden

Bibliography

Cande, S., and Kent, D.V., 1995. Revised calibration of the geomag-

netic polarity timescale for the Late Cretaceous and Cenozoic.

Journal of Geophysical Research, 100: 6093–6095.

Channell, J., Erba, E., Nakanishi, M., and Tamaki, K., 1995. Late Jurassic-

Early Cretaceous time scales and oceanic magnetic anomaly block

models. In Berggren, W., Kent, D., Aubry, M., and Hardenbol, J.

(eds.), Geochronology, Timescales and Global Stratigraphic Correla-

tion. Society of Economic Paleontologists and Mineralogists Special

Publications, 54:51–63.

Courtillot, V., and Besse, J., 1987. Magnetic field reversals, polar

wander, and core-mantle coupling. Science , 237: 1140–1147.

Cox, A., 1969. Geomagnetic reversals. Science, 163: 237–245.

Cox, A., Doell, R.R., and Dalrymple, G.B., 1963. Geomagnetic

polarity epochs and Pleistocene geochronometry. Nature, 198:

1049–1051.

Gallet, Y., and Hulot, G., 1997. Stationary and nonstationary behavior

within the geomagnetic polarity timescale. Geophysical Research

Letters, 24: 1875–1878.

Glatzmaier, G.A., Coe, R.S., Hongre, L., and Roberts, P.H., 1999. The

role of the Earth’s mantle in controlling the frequency of geomag-

netic reversals. Nature, 401: 885–890.

Glen, W. 1982. The Road to Jaramillo. Critical Years of the Revolu-

tion in Earth Science. Stanford: Stanford University Press.

Gradstein, F.M., Agterberg, F.P., Ogg, J.G., Hardenbol, J., van Veen, P.,

Thierry, J., and Huang, Z., 1994. A Mesozoic timescale. Journal of

Geophysical Research, 99: 24051–24074.

Hulot, G., and Gallet, Y., 2003. Do superchrons occur without any

palaeomagnetic warning? Earth and the Planetary Science Letters,

210: 191–201.

Jones, G.M., 1977. Thermal interaction of the core and the mantle and

long term behaviour of the geomagnetic field. Journal of Geophy-

sical Research, 82: 1703–1709.

Kent, D.V., and Gradstein, F.M., 1986. A Jurassic to recent chronol-

ogy. In Vogt, P.R., and Tucholke, B.E. (eds.), The Geology of

North America, Vol. M, The Western North Atlantic Region.

Boulder: Geological Society of America.

Kent, D.V., and Olsen, P.E., 1999. Astronomically tuned geomagnetic

polarity timescale for the Late Triassic. Journal of Geophysical

Research

, 104: 12831–12842.

Lutz, T.M., 1985. The magnetic reversal record is not periodic. Nature,

317: 404–407.

Lutz, T.M., and Watson, G.S., 1988. Effects of long-term variation on

the frequency spectrum of the geomagnetic reversal record. Nature,

334: 240–242.

Marzocchi, W., and Mulargia, F., 1990. Statistical analysis of the geo-

magnetic reversal sequences. Physics of the Earth and Planetary

Interiors, 61: 149–164.

Mazaud A., and Laj, C., 1991. The 15 Ma geomagnetic reversal peri-

odicity: a quantitative test. Earth and the Planetary Science

Letters, 107: 689–696.

Mazaud A., Laj, C., de Seze, L., and Verosub, K.L., 1983. 15 Ma per-

iodicity in the reversal frequency of geomagnetic reversals since

100 Ma. Nature, 304: 328–330.

McFadden, P.L., 1984a. Statistical tools for the analysis of geomag-

netic reversal sequences. Journal of Geophysical Research, 89:

3363–3372.

McFadden, P.L., 1984b. 15 Ma periodicity in the frequency of geo-

magnetic reversals since 100 Ma. Nature, 311: 396.

McFadden, P.L., 1987. Comment on “A periodicity of magnetic rever-

sals?” Nature, 330: 27.

McFadden, P.L., and Merrill, R.T., 1984. Lower mantle convection and

geomagnetism. Journal of Geophysical Research, 89: 3354–3362.

McFadden, P.L., and Merrill, R.T., 1986. Geodynamo energy source

constraints from paleomagnetic data. Physics of the Earth and Pla-

netary Interiors, 43:22–33.

McFadden, P.L., and Merrill, R.T., 1993. Inhibition and geomag-

netic field reversals. Journal of Geophysical Research, 98:

6189–6199.

McFadden, P.L., and Merrill, R.T., 1995. History of Earth’s magnetic

field and possible connections to core-mantle boundary processes.

Journal of Geophysical Research, 100: 317–316.

McFadden, P.L., and Merrill, R.T., 1997. Asymmetry in the reversal

rate before and after the Cretaceous normal polarity superchron.

Earth and the Planetary Science Letters, 149:43–47.

McFadden, P.L., and Merrill, R.T., 2000. Evolution of the geomagnetic

reversal rate since 160 Ma: Is the process continuous? Journal of

Geophysical Research, 105: 28455–28460.

Merrill, R.T., McElhinny, M.W., and McFadden, P.L., 1996. The Mag-

netic Field of the Earth: Paleomagnetism, the Core, and the Deep

Mantle. San Diego, CA: Academic Press.

Naidu, P.S., 1971. Statistical structure of geomagnetic field reversals.

Journal of Geophysical Research, 76: 2649–2662.

Negi, J.G., and Tiwari, R.K., 1983. Matching long-term periodicities

of geomagnetic reversals and galactic motions of the solar system.

Geophysical Research Letters, 10: 713–716.

Parker, R.L., 1997. Coherence of signals from magnetometers on par-

allel paths. Journal of Geophysical Research, 102:5111–5117.

338 GEOMAGNETIC REVERSAL SEQUENCE, STATISTICAL STRUCTURE

Phillips, J.D., 1977. Time-variation and asymmetry in the statistics of

geomagnetic reversal sequences. Journal of Geophysical Research,

82: 835–843.

Rampino, M.R., and Caldeira, K., 1993. Major episodes of geologic

change: correlations, time structure and possible causes. Earth

and the Planetary Science Letters, 114: 215–227.

Raup, D.M., 1985. Magnetic reversals and mass extinctions. Nature,

314: 341–343.

Stigler, S., 1987. A periodicity of magnetic reversals? Nature, 330:

26–27.

Stothers, R.B., 1986. Periodicity of the Earth’s magnetic reversals.

Nature, 332: 444–446.

Cross-references

Core-Mantle Boundary

Core-Mantle Boundary Topography, Implications for Dynamics

Geodynamo, Numerical Simulations

Geomagnetic Field, Asymmetries

Geomagnetic Polarity Timescales

Magnetic Field of Sun

Reversals, Theory

Superchrons, Changes in Reversal Frequency

GEOMAGNETIC REVERSALS, ARCHIVES

It is now widely accepted that the Earth’s magnetic field is generated

by electric currents in the iron-rich liquid outer core. A dynamo pro-

cess converts the energy associated with fluid convection within the

Earth’s core into magnetic energy. At the surface of the Earth, the field

varies on timescales that range over more than 18 orders of magnitude,

from less than a millisecond to more than 100 Ma. The most dramatic

field variations are reversals. Almost exactly one century ago, Bernard

Brunhes (1903, 1905) and his colleague David measured the magneti-

zation of a lava flow, which was magnetized in the opposite direction

to the present field. After investigating several possibilities, they con-

vinced themselves that this resulted from a magnetic field with its

magnetic north pole close to the south geographic pole. In other words

the field would have been flipped in the opposite configuration to the

present field with its south magnetic pole (in contrast to the current

belief the pole referred as the north magnetic pole is actually a south

pole which attracts the northern edge of the magnet) close to the north

geographic pole. It took about 50 years before the existence of the geo-

magnetic reversals was established. Between 1925 and 1935 the dis-

covery of new reversely magnetized rocks from different continents

convinced the Japanese Matuyama (1926, 1929) and the French scien-

tist Mercanton (1931) of their existence. However Louis Néel (1955)

and his colleagues reported that reversely magnetized rocks could

result from specific arrangements of the atoms lattices and the Japa-

nese Uyeda (1958) observed this mechanism in a dacite from Mount

Haruna. Further work demonstrated that this process was actually

associated with specific chemical configurations and thus very rarely

met in natural rocks. In the meantime, dating using new radiometric

techniques indicated that rocks of the same age had the same polarity.

Hence, it was obvious that the presence of reversed magnetization was

a worldwide phenomenon, not caused by self-reversing processes. The

discovery of the geomagnetic reversals allowed also understanding

why the basaltic seafloor magnetization was organized in alternances

of positive and negative giant anomalies parallel to the ridge axis. This

pattern actually reflects the succession of the two polarities of the geo-

magnetic field, which were initially recorded at the ridge axis and then

pushed away by the spreading seafloor.

Geomagnetic polarity timescale and the duration

of reversals

The first reversals were dated using the ages derived from land

sequences. This yielded the construction of the first geomagnetic

polarity timescale (GPTS), which covered the past 5 Ma (Cox et al.,

1963). Using the ages of these individual reversals, it was then possible

to interpolate or to extrapolate on the basis of the width of the marine

magnetic anomalies, which led to extend the timescale over the past

160 Ma (Figure G41). Beyond this period, we are faced to the absence

of magnetic anomalies. During the past 20 years much activity has been

devoted at studying long sequences of exposed sediments or lava flows in

order to build up the polarity timescale for the older periods.

The succession of polarity intervals during the past 160 Ma shows

the existence of periods of high reversal frequency, which alternate

with periods of low frequency. The mean reversal frequency is of the

order of 1 reversal per million years and the maximum value does

not seem to exceed 6 reversals per million years. During the period

extending from 118 to 83 Ma, the field remained in a virtually uninter-

rupted normal state. These 35 Ma in the upper Cretaceous are known

as the “Cretaceous normal superchron (CNS)” (between 120 and 83 Ma).

Before the Kiaman reversed superchron stretches undisturbed for 50 Ma

from the late Carboniferous to the middle Permian (between 310 and

260 Ma). The period between these two superchrons looks very much

like the magnetic record that has followed the CNS. A third superchron,

although apparently shorter in duration (20 ka), has been proposed from

magnetostratigraphic studies (Pavlov and Gallet, 1998). Thus, in the

present state of knowledge superchrons can be seen as a constant

characteristic of the polarity timescale. It was frequently claimed

that the slow decrease of reversal frequency before the Cretaceous

superchron (McFadden and Merrill, 1995) would led to the super-

chron, thus resulting from a long-term influence of the mantle

dynamics on core convection. Recent analysis (Hulot and Gallet,

2003) has shown that no long-term behavior over the 40 Ma pre-

ceding the Cretaceous superchron can be seen in the reversal rate

that could be invoked as announcing its occurrence 120 Ma ago.

Superchrons might correspond to times when the process responsible

for geomagnetic reversals passed below a certain critical threshold, and

reversals could start again when that threshold was exceeded again.

Other analyses suggest that the superchrons indicate the existence of a

nonreversing state of the geodynamo in contrast to its reversing

state. Alternatively, changes between the two processes may depend

on factors directly connected to the boundary conditions, such as

the distribution of heterogeneities within the “D” layer (just above

the core-mantle boundary) that affect the flow pattern within the

core and therefore may not be related to the intrinsic time constants

of the geodynamo.

Figure G41 Geomagnetic polarity timescale for the past 160 Ma. Black (white) bars indicate periods of normal (reverse) polarity. Note

the existence of the long Cretaceous period without any reversal.

GEOMAGNETIC REVERSALS, ARCHIVES 339

The time span between two successive reversals appears to be ran-

dom. Many statistical analyses were aimed at detecting periodicities

or recurrent features, which would be hidden behind the succession

of reversals. The succession of polarity intervals is closely approxi-

mated by a Poissonian distribution or at least by a gamma process

if one assumes that short events have been missed in the reversal

sequence. One can always consider that subsequent refinement of the

reversal sequence can change this view. A critical aspect is the pre-

sence of field excursions which large deviations from the geocentric

axial dipole during periods of very-low-field intensity. Excursions

and reversals share common characteristics, which suggest that they

could be treated as manifestations of similar processes within the core.

In fact many authors proposed that excursions must be seen as aborted

reversals whereas others regard them as intrinsic secular variation in

presence of a weak dipolar field. In the first case, they should be trea-

ted at the same level as reversals in the statistics, which would provide

a very different picture of the field. A critical point to answer these

problems would be to determine whether excursions existed during

the superchrons.

Dating revers als and estimating their duration

The age of the successive reversals is critical for analyzing their suc-

cession. As described earlier most reversals were dated by extrapolat-

ing the ages of the recent reversals after combining the spreading rates.

Very significant progress was accomplished during the past 20 years

with the discovery that carbonated sediments revealed the succession

of the orbital cyclicities of the Earth (23 ka). The ages of the Pliocene

and Pleistocene reversals were refined by using this independent

method (Shackleton et al., 1990; Tauxe et al., 1992, 1996; Channel

and Kleiven, 2000) which relies on correlating climate proxies (such

as oxygen isotope ratios, susceptibility or density variations ...or sim-

ply counting the astronomical cycles recorded in continuous sedimen-

tary sequences) with calculated variations of the Earth’s orbit. Using

seafloor spreading rates for five plate pairs, Wilson (1993) has shown

that the errors in the astronomical calibration are not greater than 0.02

Ma (which corresponds to a precessional cycle of the Earth’s orbit) and

also that spreading rate can remain constant for several million years.

A critical question that comes to mind is how long it takes for the

field to reverse from one polarity to the other. Before dealing with this

aspect, it is important to determine when a reversal starts and when it

ends. There are many observations showing successions of large oscil-

lations (Hartl and Tauxe, 1996; Dormy et al., 2000) preceding or fol-

lowing polarity changes, and it is not clear whether or not they should

be incorporated in the reversal process. Such oscillations can be linked

to enhanced secular variation in presence of low dipole field. Alterna-

tively, they can be considered as successive attempts by the field to

reverse. Important also is the definition of a transitional direction,

which must exceed the normal range of secular variation (i.e., the

range of the field variations during periods of polarity). Limits on

the virtual geomagnetic pole (VGP) latitudes have been mostly used

since it is rare that VGPs reach latitudes lower than 60



, although a

strict definition should be restrained to positions lower than 45



(epi-

sodes of large secular variation can occasionally reach these latitudes).

The sharp transition between magnetic anomalies of opposite pola-

rities, the very narrow thickness of the intervals recording sedimentary

columns as well as the small number of lava flows with directions

being unambiguously identified as “intermediate” between the two

polarities were rapidly convincing and strong evidences that reversals

are short phenomena on a geological timescale. The data recorded

from the best continuous and documented sedimentary sequences indi-

cate that the jump between the two polarities is shorter than one pre-

cessional cycle (23 ka). If we refer to radiometric dating (e.g., K-Ar

or Ar-Ar techniques) of lava flows, the problem is delicate because

of the sporadic succession of the flows characterized by pulses of erup-

tions occurring over a short period and long intervals without any

magmatic event. A direct consequence is that sequences of lavas

provide only very partial records of one or several phases of the rever-

sal but not the entire process. Finally and even more critical is the fact

that uncertainties on the ages cannot be ruled out as this is clearly illu-

strated by the large number of studies performed on the last reversal

(Brunhes-Matuyama). The compilation of K-Ar and/or Ar-Ar dating

(Quidelleur et al., 2003) for 23 volcanic records indicates an age of

789  8 ka (total error). The tuning of the d

O records from sedimen-

tary sequences to orbital forcing models gives an age of 779  2ka

(Tauxe et al., 1996). Recently, Singer et al. (2005) mentioned that the

astronomical determination is close to the age of the lavas from Maui

(Hawaii) (776  2 ka), and therefore that the other volcanic records

are related to the onset of the transitional process. Consequently they

deduce that a field reversal would require a significant period. One

can oppose that there is no reason to consider that the onset of the rever-

sal was initiated at the same time everywhere but above all if we con-

sider the total uncertainties on the ages, there is no significant

difference. Actually this uncertainty leaves doubts as to the possibility

of constraining the duration of the transition with precision. There is

also no reason to consider that all reversals have the same duration.

Most studies converge to estimates between 5 and 10 ka but durations

as short as 1 ka or as long as 20–30 ka have been proposed also. Clem-

ent (2004) recently analyzed the four most recent reversals recorded in

sediments with various deposition rates and found a mean average dura-

tion of 7 ka for the directional changes (Figure G42). An interesting

characteristic is that the mean duration seems to vary with latitude, as

expected from simple geometrical models in which the nondipole fields

are allowed to persist while the axial dipole decays through zero and

then builds in the opposite direction.

The revers ing field

Recorded by sediments

There has been a great deal of speculations concerning the processes

governing field reversals. In order to decipher the mechanisms it was

important to focus on the morphology of the field during the transition

from one polarity state to the other. The first major step was to deter-

mine whether this “transitional” field would keep its dipolar dominant

character. To achieve this goal, the paleomagnetists took advantage of

Figure G42 (a) Geographical distribution of Matuyama-Brunhes

transitional VGPs derived from a selection of sedimentary records

of the last reversal (from Clement, 1991). (b) Longitudinal

distribution of transitional VGPs plotted as the number of

transitional VGPs in a sliding 30



wide longitude window. Note

the presence of two peaks over the American and Asian

continents.

340 GEOMAGNETIC REVERSALS, ARCHIVES

the concept of the VGP, which defines the position of the pole assum-

ing that the field is dominantly dipolar. The idea was very simple. If

the field remained dipolar, then any vector at the surface of the Earth

would point toward the same pole following the rotation of the dipole

with its north (south) pole passing to the (south), while in the opposite

situation the poles would be different at any site. Using the first

records from sediments, it was rapidly shown that the field was not

dipolar during reversals (Dagley and Lawley, 1974; Hillhouse and

Cox, 1976). In the meantime the first records of paleointensity estab-

lished that the field intensity was systematically very low during the

transitions. This decrease in dipole intensity is necessary before directions

depart significantly from dipolar ones (Mary and Courtillot, 1993) and

therefore consistent with the concept of a nondipolar field. The nature of

this nondipolar field is one of the major questions to elucidate.

Two objectives were pursued in order to provide some answers

to this question—the first one relied on the acquisition of multiple

records of the same transition. It was proposed that because the Earth’s

rotation keeps a major role in field regeneration the transitions would

be dominated by axisymmetrical components (quadropolar or octupo-

lar). This suggestion could be tested at least at the first order by refer-

ring again to the useful concept of the VGPs. Indeed to satisfy the

axial symmetry, the VGPs always follow the great circle passing

through the observation site (or its antipode), which implies to study

the same reversals at different sites. Records from sediments were

appropriate because the transitions can be identified without ambiguity

and thus correlated from widely separated sites. In the meantime the

development of cryogenic magnetometers was very helpful as they

provided the possibility of measuring weakly magnetized sediments.

The results established that the VGP trajectories were effectively con-

strained in longitude but in many cases 90



away from the site meri-

dian rather than centered above it.

The second objective was to investigate whether successive rever-

sals were characterized by some recurrent or persistent features. This

required studying sequences of reversals at the same site. Again long

sequences of marine sediments were appropriate for this kind of study

as well as the use of VGP paths, despite the dominance of nondipolar

components. The most appealing observation emerged from a selec-

tion of sedimentary records (Clement, 1991; Tric et al., 1991), which

were all showing VGP paths within two preferred longitudinal bands,

over the Americas and eastern Asia. Going one step further Laj et al.

(1992) noted that these areas coincide with the cold circum Pacific

regions in the lower mantle (outlined by seismic tomography), thereby

suggesting that density or temperature conditions in the lower mantle

could control the geometry of the reversing field. However, this obser-

vation was controversial because it relied on a selection of records.

Another intriguing characteristic (Valet et al., 1992; McFadden et al.,

1993) was the fact that these VGP paths were also found 90



away

from the longitude of their sites despite their relatively wide geo-

graphic distribution (Figure G43), which could suggest some artifacts

in the recording processes. Several studies effectively questioned the

fidelity of sediments as recorders of the field variations, particularly

during periods of low field intensity. Many factors (compaction, align-

ment of the elongated particles) reduce the inclination of the magneti-

zation, a process that moves the VGPs paths far away from the

longitude of the observation sites (Rochette, 1990; Langereis et al.,

1992; Quidelleur and Valet, 1994, 1995; Barton and McFadden,

1996,). There is also some indication that for some sediments the

magnetic torque generated by a weak field is too low to provide

accurate orientation of the magnetic grains, leaving in this case a promi-

nent role to the hydrodynamic forces. These problems suggest that sedi-

ments could not be as appropriate as it was originally thought to study

reversals.

Recorded by volcanic lava flows

One must thus turn toward volcanic records keeping in mind that in

this case we are faced to the very discontinuous character of the

eruption rates. Following the presentation of preferred longitudinal

bands, Prévot and Camps (1993) compiled all volcanic records with

VGPs latitudes lower than 60



, considering only one position for poles

that were identical or very close to each other. They observed that no

preferred longitudinal band emerged from this database. Love (1998)

questioned this interpretation, arguing that similar directions from suc-

cessive flows should not be averaged, because they do not necessarily

result from a very rapid succession of eruptions. Instead he treated

each individual flow as a single time event (implying no correlation

between successive flows (Figure G44), or identically that the duration

between flows was larger than the typical correlation times of secular

variation). Because VGPs obtained from volcanics can reach latitudes

as low as 45–50



during episodes of large secular changes (and of

course excursions), it is also important to restrain the analysis to the most

transitional directions, i.e., those with VGP latitudes less than 45



Unfortunately in this case the number of points becomes too small to per-

form any robust analysis. This illustrates the difficulties of finding

detailed records of reversals but also implies and confirms that indeed

transitions occur very rapidly.

Using another selection of records, Hoffman (1991, 1992, 1996)

pointed out the existence of clusters of VGPs in the vicinity of South

America and above western Australia. Because of the apparent longevity

of these directions they were interpreted as indicating the existence of a

persistent inclined dipolar field configuration during the reversal pro-

cess. This is an attractive suggestion, which would establish some link

between the sedimentary and the volcanic records but limited to a selec-

tion of data.

It is striking that these interpretations depend on the chronology of

the lava flows. Volcanism is mostly governed by short periods of

intense eruptions alternating with quiet intervals. It is usually admitted

that the active periods can be very short with respect to the intervals of

quiescence. An indirect “magnetostratigraphic” indica tion has been

given by three parallel sections of Hawaii (Herrero-Bervera and Valet,

1999, 2005), which are not distant, by more than a few kilometers.

They all recorded the same reversal but do not show the same succes-

sions of transitional directions. Clusters of similar directions can be

present in one section without being recorded 2 km away. Similarly

Figure G43 Distance between the mean longitudes of the virtual

geomagnetic poles (VGPs) and their site meridian. Note that most lie



away from their site longitude.

GEOMAGNETIC REVERSALS, ARCHIVES 341

apparent rapid changes can be observed in detail in one case and be

absent 2 km away. In the first case, the field remained in the same

position during a period short enough for not being recorded in the

nearby section. In the second case, the field variations occurred either

over a short time interval or the field moved very rapidly and was thus

recorded only at a single location. Thus we cannot rule out that the

apparent concentrations of VGPs close to South America and western

Australia can be purely coincidental so that it cannot be considered yet

as certain that long-lived transitional states represent an actual charac-

teristic of the reversing field.

Another approach is to rely exclusively on volcanic records with

well-defined pre- and posttransitional directions and a sufficient num-

ber of intermediate directions. These records are more significant in

terms of transitional field characteristics because they are not asso-

ciated with uncertainties about the origin of the directions and their

stratigraphic relation. The distribution of poles extracted from this lim-

ited database does not display any preferred location, nor does it show

any evidence for systematics in the reversal process (Figure G45).

Note also the presence of clusters at various longitudes and latitudes.

Despite the difficulties inherent to the interpretation of sedimentary

and volcanic transitional directions, the two kinds of data may share

some common characteristics. We already noted a possible link

between the volcanic clusters and the preferred VGP paths in sedi-

ments but outlined some difficulties in reconciling these two aspects.

Another important issue is that the two kinds of records are associated

with very different timing in the acquisition of their magnetization, the

almost instantaneous cooling of the lavas being almost opposite to the

slow processes governing the lock-in of magnetization in sediments.

It is thus more justified to attempt a comparison by considering sedi-

ments with very high deposition rates in order to reach a better resolu-

tion (for magnetization acquisition) thus closer to the volcanic

characteristics. A few detailed sedimentary records with deposition

rates exceeding 5 cm per 1 ka have been published (Valet and Laj,

1984; Clement and Kent, 1991; Channell and Lehman, 1997). A domi-

nant and common characteristic is that they display a complex structure

with large directional variations preceding and/or following the transi-

tion which reminds the features seen in the volcanics (Mankinen et al.,

1985; Chauvin et al., 1990; Herrero-Bervera and Valet, 1999). These

large loops share similarities with the secular variation of the present

field. This observation reinforces the simplest model initially sug-

gested by Dagley and Lawley (1974) of a rather complex transitional

field which would be dominated by nondipole components following

the large drop of the dipole field (Valet et al., 1989; Courtillot et al.,

1992).

Figure G45 Positions of the VGPs derived from the most detailed volcanic records of reversals published so far. In contrast with

Figure G44 no Icelandic sequence of superimposed flows met the selection criteria. Note also the large number of clusters

(surrounded by circles) at various locations of the globe.

Figure G44 (a) Map of gray-scale histogram of the VGPs from

volcanic lava flows with latitudes that are low enough for being

considered as transitional (from Love et al., 1998). (b) Histogram

showing their longitudinal distribution. All lava flows were

considered and thus not necessarily incorporated within a

sequence of overlying flows. The similar directions that could

be linked to a phase of rapid volcanic eruptions were considered

as a result of a random process and thus were all taken in

consideration. The database is strongly dominated by Icelandic

lava flows.

342 GEOMAGNETIC REVERSALS, ARCHIVES

Fast impulses during rever sals?

The existence of fast impulses during the 16 Ma old reversal recorded

at the Steens Mountain in Oregon (Mankinen et al., 1985) was sug-

gested from a puzzling progressive evolution of the paleomagnetic

directions in the interiors of two transitional lava flows (Coe and Pré-

vot, 1989; Camps et al., 1999). Each lava unit recorded a complete

sequence of directions going all the way from that of the underlying

flow to the direction of the overlying flow. In the absence of any clear

evidence for anomalous rock magnetic properties, these features have

been interpreted in terms of very fast geomagnetic changes, which

would have reached 10



and 1000 nT per day. For comparison values

typical of the present-day secular variation of the field (of internal ori-

gin) are of the order of 0.1



and 50 nT per year, i.e., some 10

times

slower. In this specific case the timing of these fast changes can be

constrained by estimates of the cooling times of individual flows.

However, such rapid changes do not seem to be compatible with

accepted values of mantle conductivity (Ultré et al., 1995). As a con-

sequence this interpretation of the magnetization generated exciting

controversy. Additional detailed investigations have been conducted

in order to see whether this situation could not have arisen because

of remagnetization of the flows. Remagnetizations of lava flows, yield-

ing complex or unusual directions, have been detected at several loca-

tions where reversals have been recorded. A first interesting example

was given by Hoffman (1984) from Oligocene basaltic rocks. Valet

et al. (1998) observed the coexistence of both polarities (with similar

characteristics as for the directions recorded at Steens Mountain)

within flows marking the last reversal boundary (0.78 Ma) at the Can-

ary Islands, and also in a lava flow associated with the onset of the

upper Réunion subchron (2.13 Ma) in Ethiopia. In these cases, a

purely geomagnetic interpretation would imply that a full reversal took

place in only a few days. Similarly to the Steens Mountain, there is no

striking difference between the rock magnetic properties of these units

and the rest of the sequence, but a scenario involving thermochemical

remagnetization is not incompatible with the results. Thermochemical

magnetic overprinting can be particularly serious when it affects a flow

emplaced at a time of very low field intensity, sandwiched between

flows emplaced at a time of full (stable polarity) intensity. Recent

investigation of additional flows at Steens did not shed more light on

this problem but rather casts doubts on a geomagnetic interpretation

of the paleomagnetic directions (Camps et al., 1999). Therefore, the

existence of very large and rapid changes that have been documented

from a single site by a unique team remains controversial. In the mean-

time no alternative explanation has been completely accepted yet.

Field intensity variations across reversals

Variations in field intensity accompanying the reversals provide impor-

tant and unique information concerning the transition itself but also the

evolution of the dipole prior and after the reversals. We mentioned ear-

lier that a significant decrease of the dipole was reported since the ear-

liest studies. It is well established, based on all sedimentary as well as

volcanic records that field intensity drops significantly and in most

cases these changes last longer than directional changes (see e.g.,

Lin et al., 1994; Merrill and Mc Fadden, 1999). Note that similar

drops have been mentioned in all records of excursions with one

exception (Leonhardt et al., 2000). Initially, the records were restrained

to the transitional interval or to a few thousand years preceding and

following the reversal. Several long records of relative paleointensity

have now been obtained using sequences of deep-sea sediments, which

make possible to observe the evolution of the field over a long period.

These independent records from sediment cores in different areas of

the world can be stacked together to extract the evolution of the geo-

magnetic dipole moment (Guyodo and Valet, 1996, 1999; Laj et al.,

2000). There is a remarkable consistency between the first stacks pub-

lished for the past 200 and 800 ka and a similar approach performed

from field measurements immediately above bottom seafloor magnetic

anomalies (Gee et al., 2000). There is also a good agreement with the

cosmogenic isotopic records (Frank et al., 1997; Baumgartner et al.,

1998; Carcaillet et al., 2003, 2004; Thouveny et al., 2004). Apart

many other interesting observations a dominant feature is the existence

of a large drop of intensity about every 100 ka, which coincide with

excursions reported from various sequences in the world.

The most recent curve of relative paleointensity was extended to the

past 2 Ma (Valet et al., 2005) and is in good agreement with the absolute

dipole moments derived from volcanic lavas, which were used for cali-

bration. It shows that the time-averaged field was higher during periods

without reversals but the amplitude of the short-term oscillations

remained the same. As a consequence, few intervals of very low inten-

sity and thus less instability are expected during periods with a strong

average dipole moment, whereas more excursions and reversals are pro-

duced during periods of weak field intensity. Prior to reversals, the axial

dipole decays during 60 to 80 ka, but rebuilds itself in the opposite direc-

tion in a few thousand years at most (Figure G46). The most complete

volcanic records confirm that recovery following a transition is short

and culminates to very high values. The detailed volcanic records includ-

ing determinations of absolute paleointensity provide support for such

an asymmetry. Strong posttransitional field values have been reported

in a Pliocene reversal recorded at Kauai (Bogue and Paul, 1993) and in

the upper Jaramillo (0.99 Ma) subchron recorded from Tahiti (Chauvin

et al., 1990) as well as for the lower Mammoth reversal (3.33 Ma) from

Hawaii (Herrero-Bervera and Valet, 2005). The same characteristics

emerge also from the 60 Ma oldest record obtained so far in Greenland

(Riisager and Abrahamsen, 2000), from the 15 Ma old Steens Mountain

reversal (Prévot et al., 1985) and from the last reversal (0.78 Ma)

recorded from La Palma in the Canary Islands (Valet et al., 1999).

Conclusion and perspectives

Several hundreds reversals have been documented from geological

records and it is not unlikely, yet not demonstrated, that reversals

always accompanied the existence of the geomagnetic field. Their

internal origin neither makes any doubt. Many numerical models for

the Earth’s dynamo were produced during the last decade, including

three-dimensional self-consistent dynamos that exhibit magnetic rever-

sals. However, mostly for computation difficulties the parameters used

remain far away from the Earth. This demonstrates the importance of

Figure G46 Field intensity variations across the five reversals

occurring during the past 2 Ma. In this figure we superimposed

the changes in dipole moment during the 80 and 20 ka time

intervals, respectively preceding and following each reversal.

Note the 60–80 ka long decrease preceding the reversals, and the

rapid recovery following the transitions.

GEOMAGNETIC REVERSALS, ARCHIVES 343

accumulating data. For about 30 years the paleomagnetists attempted

to acquire as many detailed records as possible using the magnetic

memory of sediments and lava flows. One of the first objectives was

to determine whether the field keeps its dipolar character when rever-

sing. The complexity of the directional changes shown by the detailed

records and the large decrease of the field intensity indicates that the

dipolar component strongly decreases by at least 80%, if not vanishing

completely. One of the major constraints is the rapidity of the reversal

process. There is no clear estimate for reversal duration, which may

vary. Indeed it seems easier to isolate transitional directions for some

reversals than for some others. There is no estimate for a lower limit

of the duration of a transition which could be as short as a few hun-

dreds years, if not less. It is reasonable to consider that the upper limit

does not exceed 20 ka. After many years the suitability of sedimentary

records (which have the advantage of preserving continuous informa-

tion on field evolution) has been heavily questioned because their

direction of magnetization can be affected by other factors (climate,

alignment of the magnetic grains, postdepositional reorientations),

particularly in presence of low field intensity. It is thus wise to turn

also toward volcanic records despite their intrinsic limits in terms of

resolution and dating. If we refer to the existing volcanic database, dif-

ferent views are presently defended regarding the field configuration

during the short transitional period. Some claim that there is a domi-

nance of the pole positions within preferred longitudinal bands, parti-

cularly within the American and Australian sectors, while others

oppose rock magnetic artifacts and defend that the distribution of the

transitional directions is typical of a nondipole field that would be

similar to the present one. These two views have different implications

and impose different constrains. The first one assumes that the lower

mantle exerts some control on the reversal processes while the alterna-

tive interpretation defends that the transitional field would result from

intrinsic processes linked to the dynamic of the core fluid. Another

aspect is the existence of precursory events. The complexity of the field

evolution prior to reversals depicts some “excursions” of the directions

that are interpreted as precursory events. This observation can be

linked to the long-term decay of the dipole component prior to the

reversal, which is responsible for the complexity of the directional

changes observed at the surface. Under this scenario the “precursory”

excursions simply reflect the dominance of the nondipole part of the

field, which will then prevail during the transition. Finally, a fast and

strong recovery takes place immediately after the transition. The

amplitude of this restoration phase is certainly critical as recent obser-

vations suggest that the dipole field strength could be a dominant fac-

tor controlling the frequency of reversals.

Jean-Pierre Valet and Emilio Herrero-Bervera

Bibliography

Barton, C.E., and McFadden, P.L., 1996. Inclination shallowing and

preferred transitional VGP paths. Earth and the Planetary Science

Letters, 140: 147–157.

Baumgartner, S., Beer, J., Masarik, J., Wagner, G., Meynadier, L., and

Synal, H.-A., 1998. Geomagnetic modulation of the

Cl flux in

the Grip ice core, Greenland. Science, 279: 1330–1332.

Bogue, S.W., and Paul, H.A., 1993. Distinctive field behaviour follow-

ing geomagnetic reversals. Geophysical Research Letters, 20:

2399–2402.

Brunhes, B., 1903. Sur la direction de l’aimantation permanente dans

diversees roches volcaniques. Comptes Rendus De l Academie

Des Sciences Paris, 137: 975–977.

Brunhes, B., 1905. Sur la direction de l’aimantation dans une argile de

Pontfarein. Comptes Rendus De l Academie Des Sciences Paris,

141: 567–568.

Camps, P., Coe, R.S., and Prévot, M., 1999. Transitional geomagnetic

impulses hypothesis: geomagnetic fact or rock magnetic artifact?

Journal of Geophysical Research, 104(B8): 17747–17758.

Carcaillet, J., Thouveny, N., and Bourlès, D., 2003. Geomagnetic

moment instability between 0.6 and 1.3 Ma from cosmonuclide

evidence. Geophysical Research Letters , 30(15): 1792.

Carcaillet, J., Bourlès, D.L., Thouveny, N., and Arnold, M., 2004. A high

resolution authigenic

Be/

Be record of geomagnetic moment varia-

tions over the last 300 ka from sedimentary cores of the Portuguese

margin. Earth and Planetary Science Letters, 219:397–412.

Channell, J.E., and Kleiven, H.F., 2000. Geomagnetic paleointensity

and astronological ages for the Matuyama-Brunhes boundary and

the boundaries of the Jaramillo subchron: paleomagnetic and oxy-

gen isotope records from ODP site 983. Philosophical Transac-

tions of the Royal Society of London A, 358: 1027–1047.

Channell, J.E.T., and Lehman, B., 1997. The last two geomagnetic

polarity reversals recorded in high deposition-rate sediments drifts.

Nature, 389: 712–715.

Chauvin, A., Roperch, P., and Duncan, R.A., 1990. Records of geo-

magnetic reversals from volcanic islands of French Polynesia.

Journal of Geophysical Research, 95: 2727–2752.

Clement, B.M., 1991. Geographical distribution of transitional

VGPs: evidence for non zonal equatorial symmetry during the

Matuyama-Brunhes geomagnetic reversal. Earth and Planetary

Science Letters, 104:48–58.

Clement, B.M., 2004. Dependence of the duration of geomagnetic

polarity reversal. Nature, 428: 637–640.

Coe, R.S., and Prévot, M., 1989. Evidence suggesting extremely rapid

field variation during a geomagnetic reversal. Earth and Planetary

Science Letters, 92: 292–298.

Courtillot, V., Valet, J.-P., Hulot, G., and Le Mouël, J.-L., 1992. The

Earth’s magnetic field: which geometry? EOS Transactions, Amer-

ican Geophysical Union, 73(32): 337.

Cox, A., Doell, D.R., and Dalrymple, G.B., 1963. Geomagnetic polar-

ity epochs and Pleistocene geochronometry. Nature, 198: 1049.

Dagley, P., and Lawley, E., 1974. Paleomagnetic evidence for the tran-

sitional behaviour of the geomagnetic field. Geophysical Journal

of the Royal Astronomical Society, 36: 577–598.

Dormy, E., Valet, J.-P., and Courtillot, V., 2000. Observational con-

straints and numerical dynamos. Geochemistry, Geophysics and

Geosystems, 1: Vol1, N



10, doi:10,1029/2000GCD00062.

Frank, M., 1997. A 200 ka record of cosmogenic radionuclide produc-

tion rate and geomagnetic field intensity from

Be in globally

stacked deep-sea sediments. Earth and Planetary Science Letters,

149: 121–129.

Gee, J., Cande, S.C., Hildebrand, J.A., Donnelly, J.A., and Parker,

R.L., 2000. Geomagnetic intensity variations over the past 780 ka

obtained from near-seafloor magnetic anomalies. Nature, 408:

827–832.

Guyodo, Y., and Valet, J.P., 1996. Relative variations in geomagnetic

intensity from sedimentary records: the past 200 thousand years.

Earth and the Planetary Science Letters, 143:23–26.

Guyodo, Y., and Valet, J.P., 1999. Global changes in intensity of

the Earth’s magnetic field during the past 800 ka. Nature, 399:

249–252.

Hartl, P., and Tauxe, L., 1996. A precursor to the Matuyama/Brunhes

transition-field instability as recorded in pelagic sediments. Earth

and the Planetary Science Letters, 138: 121–135.

Herrero-Bervera, E., and Valet, J.-P., 1999. Paleosecular variation dur-

ing sequential geomagnetic reversals from Hawaii. Earth and the

Planetary Science Letters, 171: 139–148.

Herrero-Berrvera, E., and Valet, J.P., 2005. Absolute paleointensity

from the Waianae volcanics (Oahu, Hawaii) between the Gilbert-

Gauss and the upper Mammoth reversals. Earth and the Planetary

Science Letters, 234: 279–296.

Hillhouse, J., and Cox, A., 1976. Brunhes-Matuyama polarity transi-

tion. Earth and the Planetary Science Letters, 29:51–64.

Hoffman, K.A., 1984. Late acquisition of “primary” remanence in

some fresh basalts: a case of spurious paleomagnetic results. Geo-

physical Research Letters, 11: 681–684.

344 GEOMAGNETIC REVERSALS, ARCHIVES

Hoffman, K.A., 1991. Long-lived transitional states of the geomag-

netic field and the two dynamo families. Nature, 354: 273–277.

Hoffman, K.A., 1992. Dipolar reversal states of the geomagnetic field

and core-mantle dynamics. Nature, 359: 789–794.

Hoffman, K.A., 1996. Transitional paleomagnetic field behavior: pre-

ferred paths or patches? Surveys in Geophysics, 17: 207–211.

Hulot, G., and Gallet, Y., 2003. Do superchrons occur without any

paleomagnetic warming? Earth and the Planetary Science Letters,

210: 191–201.

Laj, C., Mazaud, A., Weeks, R., Fuller, M., and Herrero-Bervera, E.,

1992. Geomagnetic reversal paths. Nature, 351: 447.

Laj, C., Kissel, C., Mazaud, M., Channell, J.E.T., and Beer, J., 2000.

North Atlantic paleointensity stack since 75 ka (NAPIS 75) and

the duration of the Laschamp event. Philosophical Transactions

of the Royal Society of London A, 358: 1009–1025.

Langereis, C.G., van Hoof, A.A.M., and Rochette, P., 1992. Longitu-

dinal confinement of geomagnetic reversal paths. Sedimentary arte-

fact or true field behavior? Nature, 358: 226–230.

Love, J.J., 1998. Paleomagnetic volcanic data and geometric regularity

of reversals and excursions. Journal of Geophysical Research,

103(B6): 12435–12452.

Leonhardt, R., Hufenbecher, F., Heider, F., and Soffel, H.C., 2000.

High absolute paleointensity during a mid Miocene excursion of

the Earth's magnetic field. Earth and the Planetary Science Let-

ters, 184(1): 141–154.

Lin, J.L., Verosub, K.L., and Roberts, P.A., 1994. Decay of the virtual

dipole moment during polarity transitions and geomagnetic excur-

sions. Geophysical Research Letters, 21: 525–528.

Mankinen, E.A., Prévot, M., Grommé, C.S., and Coe, R.S., 1985. The

Steens Mountain (Oregon) geomagnetic polarity transition. 1.

Directional history, duration of episodes and rock magnetism.

Journal of Geophysical Research, 90: 10393–10416.

Mary, C., and Courtillot, V., 1993. A three-dimensional representation

of geomagnetic field reversal records. Journal of Geophysical

Research, 98: 22461–22475.

Matuyama, M., 1929. On the direction of magnetisation of basalt in

Japan, Tyosen and Manchuria. Proceedings of the Imperial Acad-

emy of Japan

, 5: 203.

McFadden, P., and Merrill, R., 1995. Fundamental transitions in the

geodynamo as suggested by paleomagnetic data. Physics of the

Earth and Planetary Interiors, 91: 253–260.

McFadden, P.L., Barton, C.E., and Merrill, R.T., 1993. Do virtual geo-

magnetic poles follow preferred paths during geomagnetic rever-

sals? Nature, 361: 342–344.

Mercanton, P.L., 1926. Inversion de l’inclinaison magnétique aux âges

géologiques. Nouvelles constattations. Comptes Rendus De l Aca-

demie Des Sciences. Paris, 192: 978–980.

Merrill, R.T., and McFadden, P.L., 1999. Geomagnetic polarity transi-

tions. Reviews of Geophysics, 37: 201–226.

Néel, L., 1955. Some theoretical aspects of rock magnetism. Advances

in Physics, 4: 191.

Pavlov, V., and Gallet, Y., 1998. Upper Cambrian to Middle Ordovi-

cian magnetostratigraphy from the Kulumbe river section (north-

western Siberia). Physics of the Earth and Planetary Interiors,

108(1): 49–59.

Prévot, M., and Camps, P., 1993. Absence of longitudinal confinement

of poles in volcanic records of geomagnetic reversals. Nature, 366:

53–57.

Prévot, M., Mankinen, E.A., Coe, R.S., and Grommé, C.S., 1985. The

Steens Mountain (Oregon) geomagnetic polarity transition. Field

intensity variations and discussion of reversal models. Journal of

Geophysical Research, 90: 10417–10448.

Quidelleur, X., and Valet, J.-P., 1994. Paleomagnetic records of

excursions and reversals: possible biases caused by magnetiza-

tion artefacts? Physics of the Earth and Planetary Interiors, 82:

27–48.

Quidelleur, X., Valet, J.-P., Le Goff, M., and Bouldoire, X., 1995.

Field dependence on magnetization of laboratory redeposited

deep-sea sediments: first results. Earth and the Planetary Science

Letters, 133:311–325.

Quidelleur, X., Carlut, J., Soler, V., Valet, J.-P., and Gillot, P.Y., 2003.

The age and duration of the Matuyama-Brunhes transition from

new K-Ar data from LaPalma (Canary Islands) and revisited

Ar/

Ar ages. Earth and the Planetary Science Letters, 208:

149–163.

Riisager, P., and Abrahamsen, N., 2000. Paleointensity of west Green-

land Paleocene basalts: asymmetric intensity around the C57n-

C26r transition. Physics of the Earth and Planetary Interiors,

118:53–64.

Rochette, P., 1990. Rationale of geomagnetic reversal versus rema-

nence recording process in rocks: a critical review. Earth and the

Planetary Science Letters, 98:33–39.

Shackleton, N.J., Berger, A., and Peltier, W.R., 1990. An alternative

astronomical calibration of the lower Pleistocene timescale based

on ODP Site 677. Transactions of the Royal Society of Edinburg

Earth Science, 81: 251–261.

Singer, B.S., Hoffman, K.A., Coe, R.S., Brown, L.L., Jicha, B.R.,

Pringle, M.S., and Chauvin, A., 2005. Structural and temporal

requirements for geomagnetic reversal deduced from lava flows.

Nature, 434: 633–636.

Tauxe, L., Deino, A.D., Behrensmeyer, A.K., and Potts, R., 1992. Pin-

ning down the Brunhes/Matuyama and Upper Jaramillo bound-

aries: a reconciliation of orbital and isotopic timescales. Earth

and the Planetary Science Letters, 109: 561–572.

Tauxe, L., Herbert, T., Shackleton, N.J., and Kok, Y.S., 1996. Astro-

nomical calibration of the Matuyama Brunhes Boundary: conse-

quences for magnetic remanence acquisition in marine carbonates

and the Asian loess sequences. Earth and the Planetary Science

Letters, 140: 133–146.

Thouveny, N., Carcaillet, J., Moreno, E., Leduc, G., and Nérini, D.,

2004. Geomagnetic moment variation and paleomagnetic excur-

sions since 400 ka BP: a stacked record of sedimentary sequences

of the Portuguese margin. Earth and Planetary Science Letters,

219: 377–396.

Tric, E., Laj, C., Jehanno, C., Valet, J.-P., Mazaud, A., Kissel, C., and

Iaccarino, S., 1991. A detailed record of the upper Olduvai polarity

transition from high sedimentation rate marine deposits of

the Po valley. Physics of the Earth and Planetary Interiors, 65:

319–336.

Ultré-Guérard, P., and Achache, J., 1995. Core flow instabilities and

geomagnetic storms during reversals: the Steens mountain impul-

sive field variation revisited. Earth and the Planetary Science

Letters, 139:91–99.

Uyeda, S., 1958. Thermoremanent magnetism as a medium of paleo-

magnetism with special references to reverse thermoremanent mag-

netism. Japanese Journal of Geophysics, 2:1–123.

Valet, J.-P., and Laj, C., 1984. Invariant and changing transitional field

configurations in a sequence of geomagnetic reversals.

Nature,

311(5986): 552–555.

Valet, J.-P., Tauxe, L., and Clement, B.M., 1989. Equatorial and mid-

latitudes records of the last geomagnetic reversal from the Atlantic

Ocean. Earth and the Planetary Science Letters, 94: 371–384.

Valet, J.P., Tucholka, P., Courtillot, V., and Meynadier, L., 1992.

Paleomagnetic constraints on the geometry of the geomagnetic

field during reversals. Nature, 356: 400–407.

Valet, J.-P., Kidane, T., Soler, V., Brassart, J., Courtillot, V., and

Meynadier, L., 1998. Remagnetization in lava flows recording

pretransitional directions. Journal of Geophysical Research,

103(B5): 9755–9775.

Valet, J.P., Meynadier, L., and Guyodo, Y., 2005. Geomagnetic field

strength and reversal rate over the past 2 Million years. Nature,

435: 802–805.

GEOMAGNETIC REVERSALS, ARCHIVES 345