Geomagnetic Field Behavior at High Latitudes from a Paleomagnetic Record
from Eltanin Core 27-21 in the Ross Sea Sector, Antarctica
Luigi Jovane1-2, Gary Acton1, Fabio Florindo2, and Kenneth L. Verosub1
1
2
Department of Geology, University of California, Davis, CA, 95616, USA.
Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143, Roma, Italy.
Abstract
We present a high-resolution paleomagnetic record from 682 discrete samples from
Eltanin 27-21 (69.03°S 179.83°E), a 16-meter long piston core recovered in 1968 by the
USNS Eltanin as part of Operation Deep Freeze. After removal of a low-coercivity
overprint, most samples yield stable characteristic remanent magnetization directions.
The down-hole variation in the magnetic inclination provides a well-resolved
magnetostratigraphy from the Brunhes Chron (0-0.78 Ma), through the Reunion
Subchron (2.128-2.148 Ma), and into Chron C2r.2r. The sedimentation rates are
sufficiently high that even short-term geomagnetic features, like the Cobb Mountain
excursion, are resolved. The record from Eltanin 27-21 provides new insights into the
behavior of the geomagnetic field at high latitudes, about which very little is currently
known. Using the variability in the inclinations during stable polarity intervals, we
estimate that the dispersion in the paleomagnetic pole position over the past ~2 Ma is
31.3°  4.3°, which is significantly higher than observed at low to mid latitude sites. We
believe that the high dispersion observed at Eltanin 27-21 reflects higher geomagnetic
field variability that is consistent with numerical modeling of the geodynamo. This
modeling suggests that patterns of fluid flow in the outer core are strongly influenced by
the inner core and predicts that geomagnetic field behavior maybe different inside and
1
outside of the tangent cylinder, which is defined as the cylinder coaxial with Earth’s
rotation axis and tangent to the inner core/outer core boundary. The tangent cylinder
intersects the Earth’s surface in the polar regions at 69.6° latitude.
Keywords: paleomagnetism, geomagnetism, magnetostratigraphy, Eltanin, Ross Sea,
Antarctica, tangent cylinder, geodynamo, Cobb Mountain Subchron, Reunion Subchron.
Introduction
Numerical modeling of the geodynamo (Glatzmaier and Roberts, 1995; Kuang and
Bloxham, 1997) and studies of present and recent geomagnetic field configurations that
estimate the geomagnetic field at the core/mantle boundary (e.g., Gubbins and Bloxam,
1985; Bloxham and Jackson, 1992; Jackson et al., 2000; Wardinski and Holme, 2006)
indicate that the behavior of the field may differ inside and outside of the tangent cylinder
owing to the physical barrier that the inner core imposes on fluid motions that occur in
the outer core. The tangent cylinder is the region encompassed by a cylinder, coaxial with
Earth’s axis of rotation and tangent to the inner core/outer core boundary. The tangent
cylinder intersects the surface of the Earth in the polar regions at latitudes of 69.6° North
and South. How (or even whether) the differences in geomagnetic field behavior manifest
themselves at the Earth’s surface can only determined by high-quality, high resolution
paleomagnetic studies from high latitudes, of which there are surprisingly few.
One potential source of material for such studies would be marine cores from the
continental margins of Antarctica. Antarctic climate is a main driver of global climatic
2
variation, and long, continuous sedimentary records from Antarctica have historically
been considered an important and challenging archive of global paleoclimatic history
(Kennett, 1977). However, due to low core recovery, stratigraphic discontinuities, and
poor microfossil abundances, it has proven difficult to obtain high-resolution sedimentary
records
from
Antarctica
with
unambiguous
age-models
and
uncomplicated
magnetostratigraphies.
Here we report new paleomagnetic results from samples that were collected from a
piston core from the Ross Sea, Antarctica, almost forty years ago. Despite the age of the
samples, we find that they contain an excellent paleomagnetic record that provides a
well-resolved magnetostratigraphy and new information about geomagnetic field
behavior at high latitudes.
Background
The first systematic piston coring of the Southern Ocean was conducted during the
1960s by the USNS Eltanin as part of Operation Deep Freeze. Of particular importance
for Antarctic research were Cruise 27 in 1968 and Cruise 32 in 1969, which collected
several dozen cores from the Ross Sea sector (Fig. 1). Subsequent piston coring in the
Ross Sea was done by the USCGC Glacier (1970s), R/V Polar Duke (1980s), and the
R/V Nathaniel B. Palmer (1990s to present).
In the first comprehensive study of piston cores from the Antarctic region, Watkins
and Kennett (1972) described the magnetostratigraphy and foraminiferal assemblages of
Eltanin cores from the Southern Ocean (Cruises 16, 26, 27, 34, 35, 36 and 39). In that
work, they identified a normal polarity zone at ~12 to ~15 mbsf as the Gauss epoch.
3
Subsequently Fillon (1975) studied radiolarian and foraminiferal assemblages of Eltanin
cores from Cruises 27 and 32 and re-interpreted this normal polarity zone as the Olduvai
Chron, based on his identification of reworked Gauss age fossils in Brunhes deposits.
In the study of Watkins and Kennett (1972), the sampling was done at a 50-100 cm
sampling interval and the paleomagnetic measurements were done using a Foster spinner
magnetometer. Step-wise alternating field demagnetizations were carried out until the
strength of the magnetization fell below the sensitivity of the magnetometer, which
usually occurred at about 15 mT. Besides the magnetostratigraphic interpretations given
above, few details concerning the paleomagnetic results were presented for the sites and
no paleomagnetic or magnetostratigraphic data were explicitly presented for the Eltanin
27-21 core. Similarly, there are no published reports of paleomagnetic studies of piston
cores from the Ross Sea sector collected by the USCGC Glacier, R/V Polar Duke or the
R/V Nathaniel B. Palmer.
Last year, we visited the Antarctic Marine Geology Research Facility in Tallahassee,
Florida to assess the suitability of using old cores from the Ross Sea for new
paleomagnetic studies. During the visit, we discovered a drawer in the cold room
containing 682 oriented samples from Eltanin 27-21. The samples were in plastic boxes,
2 cm on a side and represent an almost continuous sampling for the core. Eltanin 27-21 is
a 16-meter-long core collected at a water depth of 3456 meters in drift deposits off the
northern tip of the West Antarctic Rift System. The coring site is located at 69.03°S
179.83°E which is ~400 km northwest of Cape Adare (Figure 1). This core was relatively
uniform in texture and color and was composed of silty clay with 60-70% clay (Goodell,
1968).
4
Magnetic Measurements
After re-attaching broken fragments and stabilizing the specimens within their plastic
sample cubes, we measured their magnetization using a 2G Enterprises automated
cryogenic magnetometer (Model 755), operated in a discrete-sample mode. In-line
alternating field (AF) demagnetization was done at steps of 5, 10, 15, 20, 25, 30, 35, 40,
50, and 60 mT.
Virtually all samples displayed linear decay of their NRM on orthogonal vector plots
as they were progressively demagnetized (Fig. 2), which allowed us to determine the
direction of the characteristic remanent magnetization (ChRM) using principal
component analysis (Kirschvink, 1980). Maximum angular deviation (MAD) angles,
were < 10° for 74% of the samples. Larger MAD angles were generally associated with
more weakly magnetized samples, which mainly occur within polarity transitions. Many
samples possessed a small, often nearly vertical upward, secondary overprint that, in
almost all cases, could be removed by alternating field demagnetization at a level
between 10 and 20 mT. This low-coercivity overprint is probably a viscous
magnetization acquired during initial coring operations and subsequent storage and
handling, and/or the effect of a possible previous demagnetization to 15 mT. The
remanence measurements (NRM before and after demagnetization), PCA results,
colatitudes, polarity interpretation, and other results determined for each sample are given
in supplementary Table A1.
The intensity of the natural remanent magnetization (NRM) ranged from 1.86 x 10-7
to 1.88 x10-5 A/m with a mean of 4.23 x10-6 A/m (Fig. 3). Samples at 10.02, 11.76 and
5
15.65 meters have very high intensities, which is probably related to the presence of
manganese-stained pebbles or rock fragments (Goodell, 1968) or possibly ice-rafted
debris. Many of the samples from the uppermost part of the core (< 50 cm depth) have
anomalously shallow directions, which may result from soft sediment deformation during
coring or sampling.
The ChRM inclinations were used to determine magnetic polarity. The lack of
azimuthal orientation of the core did not pose a problem for magnetostratigraphic studies
because the inclination of the geomagnetic field at the site latitude is very steep (~±80°).
The directional record from Eltanin 27-21 defines five normal and five reversed
polarity zones (Fig. 2). Our correlation of this polarity zonation to the Geomagnetic
Polarity Time Scale of Gradstein et al. (2004) is straightforward (Table 1). From 0.16 to
5.08 mbsf, the polarity is normal, which we interpret as the Brunhes Chron (C1n).
Because we do not know how much sediment might be missing from the top of the core,
we cannot determine the exact age of the top of the core. From 5.08 to 11.84 mbsf, there
is a reversed polarity interval that we interpret as the youngest part of the Matuyama
Chron (C1r), including the normal Jaramillo Subchron (C1r.1n) from 5.90 to 6.59 mbsf
and the Cobb Mountain excursion (C1r.2r-1n) from 7.16 to 7.35 mbsf. We recognize the
Olduvai Subchron (C2n) as the normal polarity interval below C1r and extending from
11.84 to 14.29 mbsf. Below this is a reversed polarity interval that persists until the end
of the core. We interpret this as Chron C2r. At 64 cm from the base of the core, there is a
single sample with normal polarity that we attribute to the Reunion Subchron (C2r.1n;
2.128-2.148 Ma). This correlation is consistent with the low-resolution study of radiolaria
and foraminifera assemblages done by Fillon (1975). The base of the core is therefore
6
older than 2.148 million years, and rates of sedimentation for individual polarity zones
range from 0.41 cm/k.y. to 1.46 cm/k.y., with a mean rate of 0.76 cm/k.y. If the same
rate of sedimentation prevailed above the Brunhes/Matuyama boundary, the Brunhes
portion of the core would encompass 670 k.y. This suggests that perhaps about 85 cm or
111 k.y. of the Brunhes Chron is missing from the top of the core or that sedimentation
rates were about 15% slower during the Brunhes Chron.
Discussion
Geomagnetic field behavior at high latitudes
The paleomagnetic record from Eltanin 27-21 is of unusually high quality for an
Antarctic core. The sedimentation also appears to be relatively continuous; no apparent
hiatuses occur; and the sediment is relatively fine grained throughout, all of which are
features favorable for paleomagnetic studies and somewhat uncommon in glacial
environments.
One way of assessing the quality of the data from Eltanin 27-17 would be to
determine the mean paleomagnetic directions for normal and reversed polarity intervals
and to compare their directions to each other and to the direction of a geocentric axial
dipole.
Because our samples are not azimuthally oriented, we cannot do this with
directions, and because we are at high latitudes, we have to be careful about using the
arithmetic mean of the inclinations. Nevertheless, we note that the arithmetic means of
the inclinations of the normal samples (-73.4°) and of the reversed samples (72.2°) have
7
almost identical absolute values and compare favorably with the inclination of geocentric
axial dipole at the sampling site, which is 79.1°.
Although the sedimentation rates are moderate to low for marine sections, the overall
resolution of the paleomagnetic signal is relatively high as is apparent from the shortlived features that are recorded, such as the Cobb Mountain excursion and the Reunion
Subchron. These features have only been recognized previously near Antarctica in cores
from ODP Leg 178; namely Sites 1095D, 1101A and 1095D for the Cobb Mountain
excursion and Sites 1096C and 1101A for the Reunion Subchron (Acton et al., 2002).
The boundaries of most polarity zones in Eltanin 27-21 are also quite sharp and welldefined, and in some cases, polarity transitions are preceded by well-resolved precursors.
In particular, the small inclination fluctuation at 5.19-5.25 mbsf, about 15 k.y. before the
Brunhes/Matuyama boundary, appears to correspond to transition-field instability DIP-1
of Hartl and Tauxe (1996). The same fluctuation is present at ODP Site 1095A (Acton et
al., 2002) and in lavas on La Palma, Canary Island (Singer et al., 2002).
All of these attributes attest to the ability of Eltanin 27-21 to provide new insights
into the behavior of the geomagnetic field at high latitudes. One immediate observation is
the absence of long intervals of low inclination in the Brunhes Chron. Previous
paleomagnetic studies of two cores from the Arctic Ocean (Nowaczyk et al., 1994;
Nowaczyk and Knies, 2000; Nowaczyk et al., 2001) had suggested that at high latitudes,
geomagnetic excursions might be more frequent, of longer duration and/or of different
character than corresponding excursions observed at lower latitudes. In some of these
Arctic cores, reversed polarity or excursional intervals comprised over 30% of the record
over the past few hundred k.y. The inclinations from Eltanin 27-21 and from other
8
Antarctic sites (Acton et al., 2002) are not consistent with these observations. Either the
Arctic sediments are much older than Brunhes age, with the thick polarity zones being
polarity subchrons and chrons rather than excursions; or deviations from dipole fields are
highly accentuated for high northern latitudes but not for high southern latitudes during
this period; or sedimentation rates were fortuitously high during excursions in the many
parts of the Arctic. Arguments against older sediment ages for the Arctic sites have been
provided in a recent review by Spielhagen et al. (2004). If these arguments are correct,
then the first alternative is not viable.
A new approach for analyzing inclination-only data
The paleomagnetic record from Eltanin 27-21 also allows us to estimate the
dispersion of the geomagnetic field at the latitude of the site. However, this analysis must
be done carefully because even methods designed to avoid statistical biases that occur
when averaging inclination-only data, such as that proposed by McFadden and Reid
(1982), have been shown to have problems at high latitudes (Arason and Levi, 2006). To
avoid these problems, we combined a new maximum likelihood method developed by
Arason and Levi (2006) with a Monte Carlo simulation method that maps the region of
possible solutions for various combinations inclinations vs. directional angular dispersion
or paleolatitude vs. virtual geomagnetic pole (VGP) angular dispersion.
With this
approach we are able to obtain unbiased estimates of the inclination, paleolatitude,
dispersions, and their uncertainties.
For the simulation, we generated 10,000 Fisherian deviates for paleomagnetic poles
with known mean inclinations and dispersions that span a large grid of possible values.
9
We then mapped the region in this grid that had statistical properties indistinguishable
from the observed data (Fig. 4). This approach yielded results virtually identical to the
Arason-Levi method but also allowed us to determine the uncertainty in the observed
inclination and paleolatitude, with the uncertainty in the dispersions being determined
using the method of Cox (1969). In contrast, the Arason-Levi method determines the
uncertainty for a Fisherian distribution with a true inclination and dispersion equivalent to
the maximum likelihood estimate. This underestimates the true uncertainty in the
observations.
As shown in Table 1, the absolute values of the unbiased mean inclinations for the
normal and reversed polarity samples are statistically indistinguishable from each other,
with both being nearly vertical. As noted above, the biased arithmetic means of normal
and reversed inclinations are also indistinguishable. Thus, regardless of the method used
to calculate mean inclinations, the inclination data for the site pass the paleomagnetic
reversal test. The 95% confidence interval for the unbiased mean of the reversed polarity
samples (76.9° to 90°) differs insignificantly from that of the normal polarity samples
(81.5° to 90°).
The unbiased reverse mean also differs insignificantly from the expected inclination
of a geocentric axial dipole (GAD), which is 79.2° for the site, whereas the mean of the
normal polarity samples and the mean of all the samples are marginally steeper than the
expected inclination of a GAD field. We note however that when estimating unbiased
inclinations for high paleolatitude sites, the dispersion and inclination co-vary strongly,
so that combinations of solutions exist ranging from those that have steeper inclination
with higher dispersion or shallower inclination with lower dispersion (Fig. 4). Thus, if
10
data have slightly higher than expected dispersion, due to sample orientation errors,
measurement errors for weakly magnetized samples, small amounts of coring
deformation, or non-dipole components of the geomagnetic field, then the unbiased mean
inclination will tend to be steeper. One could assume a geomagnetic dispersion model to
address this problem (e.g., as was done by Cox and Gordon, 1984), but no single
dispersion model is widely accepted and, in fact, a goal of the study is to put bounds on
the VGP dispersion and inclination.
The tradeoffs between these parameters should be kept in mind. In any case, the
slightly steeper normal polarity inclination may be indicative of the geomagnetic field or
may be caused by additional dispersion unrelated to the geomagnetic field. It is also
worth noting that this steeper than expected direction suggests that inclination flattening,
which is often a concern when dealing with paleomagnetic directions from sediments
(e.g., Kodama, 1997; Tauxe, 2005), is negligible or non-existent in these sediments.
Testing the tangent cylinder hypothesis
Using the Monte Carlo simulation, we determined that the best estimate for the VGP
dispersion for the Eltanin 27-21 core is 33.5° with lower and upper 95% confidence
limits of 31.7° and 35.6°. For the same data, the Arason-Levi method gives a value for
the dispersion of 39.0°. As noted above, tradeoffs between inclination (or geomagnetic
paleolatitude) and the inclination (or VGP) dispersion occur. Because our best estimate of
the inclination from all the data is slightly steeper than the GAD direction, the tradeoff
could yield a dispersion estimate that is slightly larger than the true value, if the true field
direction is that of a GAD. Obviously, the site has not moved significantly over the past 2
m.y., and so the true paleolatitude is approximately equal to the current site latitude. The
11
paleolatitude calculated from the inclinations, however, is based on the assumption that
the field is a GAD. The steeper than expected direction and higher than expected
paleolatitude therefore could indicate that non-dipole components have been large
enough to steeper the inclination significantly at the site, that dispersion has been high, or
that both non-dipole field components and dispersion are high at the site. Even if we
assume the true inclination is the GAD value of 79.2°, the best estimate of the dispersion
is 28.8°1.8°. This provides a lower bound for the VGP dispersion. In the following
discussion, we will use a conservative dispersion estimate of 31.3°4.3°, which spans the
range of possible values between the case where the inclination is estimated from the data
and the case where the inclination is assumed to be that of the present-day GAD. This
value is consistent with that found for the dispersion during any individual polarity
interval (Table A2). Furthermore, no notable difference in dispersion exists between
normal and reversed polarity intervals (Table A2).
The dispersion for Eltanin 27-17 is shown in Figure 5 along with expected values
computed by Quidelleur and Courtillot (1996).
Also shown are estimates of the
dispersion determined by Tauxe et al. (2004) from 37 Brunhes and Matuyama age lava
flows in McMurdo Sound and by Baraldo et al. (2003) from 27 lava from Deception
Island spanning a 100 k.y. period about 150,000 years ago.
Our estimate of the dispersion is considerably higher than the value predicted by
paleosecular variation models and also higher than those reported in the other two
studies, which involved fewer samplings of the geomagnetic field over shorter intervals
of time. Given all of the indicators for the fidelity of the Eltanin 27-17, we do not believe
that the dispersion is an artifact of a noisy paleomagnetic record. Furthermore, our results
12
appear to be consistent with the results of a number of dynamo models that were
successful in explaining the origin and overall behavior of the geomagnetic field.
These models showed that the poloidal and toroidal flux distributions of the field had
very different patterns inside and outside the tangent cylinder (Glatzmaier and Roberts,
1995; Kuang and Bloxham, 1997). Subsequent modeling showed that these differences
could be explained by the presence of polar vortices in the core (Olson and Aurnou,
1999). Other evidence for the importance of polar vortices comes from thermal
convection experiments in a rotating fluid (Aurnou et al., 2003). Detailed analyses of
present and recent geomagnetic field configurations have indicated that fluid motions in
the core associated with polar vortices in the tangent cylinder result in high latitude flux
patches on the core/mantle boundary with high magnetic density (Olson et al., 2002;
Hulot et al, 2002; Gubbins et al., 2006). These flux patches strengthen the non-dipole
components of the geomagnetic field at high latitudes relative to the dipole component.
The net effect would be greater dispersion at the Earth’s surface within or near the
tangent cylinder, as we observe at the site of Eltanin 27-21, which lies virtually on the
tangent cylinder.
Conclusion/Summary
The forty-year-old Eltanin 27-21 core has yielded a high-resolution paleomagnetic
record that provides detailed ages for sediments that have previously proven difficult to
date and a rare high-fidelity record of geomagnetic field behavior at high latitude. The
magnetostratigraphy extends from the Brunhes Chron (C1n; 0-0.78 Ma), through the
Reunion Subchron (C2r.1n; 2.128-2.148 Ma), and into but not through Chron C2r.2r
13
(2.148-2.581 Ma). The sedimentation rates are sufficiently high, ranging from 4.1 to 14.6
m/m.y. with an average of 7.6 m/m.y., that even short-term geomagnetic fluctuations, like
the Cobb Mountain and DIP-1 events and details of geomagnetic secular variation are
well recorded. To date, this is the highest latitude magnetostratigraphic record for
Pliocene-Pleistocene sediments in the Ross Sea sector, providing important constraints
for future coring and for further studies on old cores around Antarctica.
The angular dispersion in the paleomagnetic pole position caused by geomagnetic
secular variation over the past ~2 Ma is conservatively estimated to be 31.3°  4.3°. This
is significantly higher than observed from low to mid latitude sites where dispersion is
generally < 20° and is somewhat higher than observed at the few high-latitude sites
studied (dispersions of ~16°-26°) and than predicted by secular variation models
(dispersions of ~20°-25°). We suggest that the high dispersion observed at Eltanin 27-21
reflects higher geomagnetic field variability that occurs nearly directly over the tangent
cylinder.
REFERENCES : [need to be prepared in base of the journal]
Acton, G.D., Guyodo, Y., and Brachfeld S.A., 2002, Magnetostratigraphy of sediment
drifts on the continental rise of West Antarctica (ODP Leg 178, Sites 1095, 1096,
and 1101), in Barker, P.F., Camerlenghi, A., Acton, G.D., and Ramsay, A.T.S.,
Proceedings of the Ocean Drilling Program, Scientific Results, Volume 178: College
Station, Texas, Ocean Drilling Program, p. 1000.
Arason, P., and Levi, S., 2006, The Maximum likelihood solution to inclination-only
data: Eos trans. AGU, v.86 (62), Fall Meet., Abstract GP21B-1312.
14
Aurnou, J.M., S. Andreadis, L. Zhu, and P.L. Olson, 2003, Experiments on convection in
Earth's core tangent cylinder: Earth and Planetary Science Letters, v. 212, p. 119134.
Baraldo, A., Rapalini, A.E., Böhnel, H, and Mena1, M., 2003, Paleomagnetic study of
Deception Island, South Shetland Islands, Antarctica: Geophysical Journal
International, v. 153, p. 333-343.
Bloxham, J., and Jackson, A., 1992, Time-dependent mapping of the magnetic field at the
core-mantle boundary, J. Geophys. Res., 97, 19537–19563.
Cox, A., 1969. Confidence limits for the precision parameter , Geophys. J. R. Astron.
Soc., 18, 545-549.
Cox, A., and Gordon, R.G., 1984, Paleolatitudes determined from paleomagnetic data
from vertical cores: Reviews of Geophysics and Space Physics, v. 22, n. 1, p.47-72.
Fillon, R.H., 1975, Late Cenozoic Paleo-Oceanography Of Ross Sea, Antarctica:
Geological Society of America Bulletin, v. 86, n.6, p. 839-845.
Glatzmaier, G.A., and Roberts, P.H., 1995, A three-dimensional convective dynamo
solution with rotating and finitely conducting inner core and mantle: Physics of the
Earth and Planetary Interiors, v. 91, p. 63-75.
Goodell, H.G., 1968, USNS Eltanin Cruises 16-27, core descriptions: Sedimentology
Research Laboratory Contribution, n. 25, Department of Geology, Florida State
University, Tallahassee, 247 p., http://www.arf.fsu.edu/publications/ELT_16_27.pdf.
Gradstein, F.M., Ogg, J.G., Smith, and A.G., 2004, A Geological Time Scale 2004:
Cambridge Univ. Press, Cambridge, 589 p.
Gubbins, D., and Bloxham, J., 1985, Geomagnetic field analysis – III. Magnetic fields on
15
the core-mantle boundary, Geophys. J. R. Astron. Soc., 80, 695– 713.
Gubbins, D., and Bloxham, J., 1987, Morphology of the geomagnetic field and
implications for the geodynamo: Nature, v. 325, n. 6104, p. 509-511.
Gubbins, D., Jones, A.L., and Finlay, C.C., 2006, Fall in Earth's magnetic field is erratic:
Science, v. 312, n. 5775, p. 900-902.
Jackson, A. Jonkers, A. R. T., and Walker, M. R., 2000, Four centuries of geomagnetic
secular variation from historical records, Phil. Trans. R. Soc. London, Ser. A, 358,
957– 990.
Kennett, J.P., 1977, Cenozoic evolution of Antarctic glaciation, the Circum-Antarctic
Ocean, and their impact on global paleoceanography: Journal of Geophysical
Research, v. 82, p. 3843–3860.
Kuang, W., and Bloxham, J., 1997, An Earth-like numerical dynamo model: Nature, v.
389, n. 6649, p. 371-374.
Hartl, P, and Tauxe, L, 1996, A precursor to the Matuyama/Brunhes transition-field
instability as recorded in pelagic sediments: Earth and Planetary Science Letters, v.
138, n. 1-4, p. 121-135.
Hulot, G, Eymin, C., Langlais, B., Mandea, M., and Olsen, N., 2002, Small- scale
structure of the geodynamo inferred from Oersted and Magsat satellite data: Nature,
v. 416, p. 620-623.
Kennett, J.P. and Fillon, R.H., 1970, Micropaleontological and associated studies of
Southern Ocean deep-sea cores: Antartic Journal of the United States, v. 5, n. 5, p.
181-182.
Kennet. J.P., and Watkins, N.D., 1970, Geomagnetic Polarity Change, Volcanic maxima
16
and Faunal Extintion in the Soouth Pacific: Nature, v. 277, p. 930-934.
Kirschvink, J. L., 1980, The least-squares line and plane and the analysis of
paleomagnetic data: Geophysical Journal of the Royal Astronomical Society, v. 62, p.
699-718.
Kodama, K. P., 1997. A successful rock magnetic technique for correcting paleomagnetic
inclination shallowing: Case study of the Nacimiento Formation, New Mexico, J.
Geophys. Res., 102, 5193–5205.
Jackson, A., Jonkers, A.R.T., and Walker, M.R., 2000, Four centuries of geomagnetic
secular variation from historical records: Philosophical Transactions of the Royal
Society, v. 358, n. 1768, p. 957-990.
McFadden, P.L., and Reid, A.B., 1982, Analysis of palaeomagnetic inclination data:
Geophysical Journal of the Royal Astronomical Society, v. 69, n.2, p.307-319.
Nowaczyk, N.R., Frederichs, T.W., Kassens, H., Norgaard-Pedersen, N., Spielhagen,
R.F., Stein, R., and Weiel, D., 2001, Sedimentation rates in the Makarov Basin,
central
Artic
Ocean:
A
paleomagnetic
and
rock
magnetic
approach:
Paleoceanography, v. 16, n. 4, p. 368-389.
Nowaczyk, N.R., and Knies, J., 2000, Magnetostratigraphic results from eastern Arctic
Ocean-AMS 14C ages and relative paleointensity data of the Mono Lake and
Laschamp geomagnetic reversal excursions: Geophysical Journal International, v.
140, p. 185-197.
Nowaczyk,
N.R.,
Frederichs,
T.W.,
Eisenhauer,
A.,
and
Gard,
G.,
1994,
Magnetostratigraphic data from late Quaternary sediments from the Yermak Plateau,
Arctic Ocean: Evidence for four geomagnetic polarity events within the last 170 Ka
17
of the Brunhes Chron: Geophysical Journal International, v. 117, p. 453-471.
Olson, P., Sumita, I., and Aurnou, J., 2002, Diffusive magnetic images of upwelling
patterns in the core: Geophysical Journal International, v. 107, n. B12, p. 2348,
doi:10.1029/2001JB000384 .
Olson, P., and Aurnou, P., 1999, A polar vortex in the Earth's core: Nature, v. 402, p.
170-173.
Quidelleur, X., and Courtillot, V., 1996, On low-degree spherical harmonic models of
paleosecular variation: Physics of the Earth and Planetary Interiors, v. 95, n. 1-2, p.
55-77.
Singer, B. S., Relle, M.K., Hoffman, K.A., Battle, A., Laj, C., Guillou, H., and
Carracedo, J. C., 2002, Ar/Ar ages from transitionally magnetized lavas on La Palma,
Canary Islands, and the geomagnetic instability timescale: Journal Geophysical
Research, v. 107, n. B11, p. 2307, doi:10.1029/2001JB001613.
Spielhagen, R. F., Baumann, K.-H., Erlenkeuser, H., Nowaczyk, N. R.,
Norgaard-
Pedersen, N., Vogt, C., and Weiel, D., 2004, Arctic Ocean deep-sea record of
northern Eurasian ice sheet history, Quat. Sci. Rev., 23, 1455-1483.
Tauxe, L., 2005, Inclination flattening and the geocentric axial dipole hypothesis, Earth
Planet. Sci. Lett., 233, 247-261.
Tauxe, L., Gans, P., and Mankinen, E.A., 2004, Paleomagnetism and Ar-40/Ar-39 ages
from volcanics extruded during the Matuyama and Brunhes Chrons near McMurdo
Sound, Antarctica: Geochemistry, Geophysics, Geosystems, v. 5, Q06H12.
Wardinski, I., and Holme, R., 2006, A time-dependent model of the Earth's magnetic
field and its secular variation for the period 1980-2000, J. of Geophys. Res., 111,
18
B12101.
Watkins, N.D. and Kennett, J.P., 1972, Regional sedimentary disconformities and upper
Cenozoic changes in bottom water velocity between Australia and Antartica, in
Hayes, D.E., Antarctic Oceanology II, The Australian-New Zealand Sector: Antartic
Research Series, v. 19, p.273-294.
FIGURE AND TABLE CAPTIONS:
Figure 1: Locations for Eltanin 27-21, DSDP Leg 28, ANDRILL, CIROS, and CRP drill
sites are shown on the bathymetric map of the Ross Sea sector.
Figure 3: Orthogonal vector component diagrams and curves of intensity decay for AF
demagnetization. Closed (open) circles denote the projection on the horizontal
(vertical) plane.
Figure 3: Age model of the Eltanin 27-21 core based on the magnetostratigraphy. Normal
remanent magnetization (NRM) (A) and inclination of the ChRM (B) are
associated to the GPTS of Gradstein et al. (2004). The slopes of the lines give
the sedimentation rate.
Figure 4: A map of the misfit, as measured by a chi-squared statistic, between a grid of
19
possible model values with known paleolatitudes and dispersions and the
observed values of the paleolatitude and the dispersion at Eltanin 27-21. The
overall best-fit (lowest chi-squared) is shown by the star. The 95% confidence
region as defined by the chi-squared values is given by the contour (dark green
region). Had we assumed the field was a geocentric axial dipole (GAD), the best
estimate of the dispersion would be that shown by the diamond. The 95%
confidence limits for the best-fit and GAD dispersion estimates (black error
bars) are calculated using the method of Cox (1969). The white error bar gives
an estimate of the predicted dispersion from the C1 model of Quidelleur and
Courtillot (1996). Other paleolatitude and dispersion estimates are from the
arithmetic mean (square), the method of McFadden and Reid (1984) (polygon),
and the method of Arason and Levi (2006) (circle).
Figure 5: Angular dispersion value for the Eltanin 27-21 core, McMurdo Sound (Tauxe et
al., 2004) and Deception Island (Baraldo et al., 2003), on the C1 model of
Quidelleur and Courtillot (1996). The dashed vertical lines give the latitudes of
the intersection of the tangent cylinder with Earth’s surface.
Table 1. Magnetostratigraphy and Sedimentation Rates for Eltanin 27-21
Table A1: Paleomagnetic data for Eltanin 27-21. The volume of the samples has been
assumed of 4 cm3, which is not a bad estimate of the current dry volume.
20
Table A2. Estimates of unbiased mean inclination, paleolatitude, directional and VGP
dispersions (S), precision parameters (k), and related statistics for Eltanin 27-21
calculated using the methods of McFadden and Reid (1982), Arason and Levi
(2006) and a Monte Carlo Simulation method (see text).
21