Timing and geometry of Early Gondwana breakup

Abstract

The Mesozoic opening history of the Southern Ocean between South America, Africa, and Antarctica is one of the largest gaps in knowledge on the evolution of this region. Competing geodynamic models were published during the last two decades to explain the geophysical and geological observations. Here we report on aeromagnetic data collected along the East Antarctic coast during five seasons. These data provide new constraints on the timing and geometry of the early Gondwana breakup. In the Riiser-Larsen Sea/Mozambique Basin, the first oceanic crust between Africa and Antarctica formed around 155 Ma. In the west the Weddell Rift propagated from west to east with a velocity of 63 km/Myr between chrons M19N and M17N. At chron M14N, South America and Africa finally were split off the Antarctic continent. Stretching in the area of the South Atlantic started at the latest from 155 Myr onward. The different spreading velocities and directions of South America and Africa created at chron M9N the first oceanic crust in the South Atlantic. A new model indicates that the Karoo and Dronning Maud Land magmatism occurred well before any new ocean floor was created and therefore the first formation of new oceanic crust cannot directly be related to a plume interaction.

Full text

Timing and geometry of early Gondwana breakup
Wilfried Jokat, Tobias Boebel, Matthias Ko¨nig, and Uwe Meyer
1
Alfred Wegener Institute for Polar Research, Bremerhaven, Germany
Received 31 January 2002; revised 22 January 2003; accepted 15 May 2003; published 16 September 2003.
[1]The Mesozoic opening history of the Southern Ocean between South America, Africa,
and Antarctica is one of the largest gaps in knowledge on the evolution of this region.
Competing geodynamic models were published during the last two decades to explain the
geophysical and geological observations. Here we report on aeromagnetic data collected
along the East Antarctic coast during five seasons. These data provide new constraints on
the timing and geometry of the early Gondwana breakup. In the Riiser-Larsen Sea/
Mozambique Basin, the first oceanic crust between Africa and Antarctica formed around
155 Ma. In the west the Weddell Rift propagated from west to east with a velocity of
63 km/Myr between chrons M19N and M17N. At chron M14N, South America and
Africa finally were split off the Antarctic continent. Stretching in the area of the South
Atlantic started at the latest from 155 Myr onward. The different spreading velocities and
directions of South America and Africa created at chron M9N the first oceanic crust in the
South Atlantic. A new model indicates that the Karoo and Dronning Maud Land
magmatism occurred well before any new ocean floor was created and therefore the first
formation of new oceanic crust cannot directly be related to a plume interaction. INDEX
TERMS:1517 Geomagnetism and Paleomagnetism: Magnetic anomaly modeling; 1744 History of Geophysics:
Tectonophysics; 3040 Marine Geology and Geophysics: Plate tectonics (8150, 8155, 8157, 8158); 3005
Marine Geology and Geophysics: Geomagnetism (1550); 7220 Seismology: Oceanic crust; KEYWORDS:
Antarctica, Aeromagnetik, Gondwana, Mesozoic reconstruction
Citation: Jokat, W., T. Boebel, M. Ko¨ nig, and U. Meyer, Timing and geometry of early Gondwana breakup, J. Geophys. Res.,108(B9),
2428, doi:10.1029/2002JB001802, 2003.
1. Introduction
[2] How and when did Gondwanaland begin to separate?
What was the shape of the Southern Ocean in the Early
Cretaceous? For Late Cretaceous and Cenozoic times, most
of these questions were answered in the last two decades
through international research programs. Satellite altimeter
data in combination with marine magnetic data considerably
sped up this process as fracture zones were easily traced over
large distances allowing scientists to reconstruct the drift
paths of the continents in greater detail. For the Southern
Hemisphere this worked well to chron C34 based on an
extensive magnetic database. Further back in Mesozoic times
the model is less well constrained since a critical amount of
magnetic data is missing. The initial extent of the Southern
Ocean and the paleopositions of South America (SAM) and
Southern Africa (AFR) relative to Antarctica (ANT) forming
parts of Gondwanaland are still debatable [Segoufin, 1978;
Segoufin and Patriat, 1980; LaBrecque and Barker, 1981;
Martin and Hartnady, 1986; Kristoffersen and Haugland,
1986; Lawver et al., 1991; Cox, 1992; Elliot, 1992; Grunow,
1993b; LaBrecque and Ghidella, 1997; Livermore and
Hunter, 1996; Lawver et al., 1998].
[3] The center of this vast landmass was Antarctica. It is
the only part of Gondwanaland that had common bound-
aries to all fragments of the former supercontinent [Lawver
et al., 1991] (South America, Africa, Madagascar, India,
Australia, and New Zealand). However, details on the
timing and geometry of the Mesozoic breakup-related
basins are still largely unknown. Geological investigations
in SAM, AFR, and ANT revealed large igneous provinces
on the continents, which, when reconstructed represent one
of the most extensive volcanic provinces in the world of all
[Cox, 1992; Elliot, 1992; Brewer et al., 1992; Rapella and
Pankhurst, 1992]. Onshore, the volcanic activities started
around 200 Ma and terminated around 170 Ma. This
volcanism is believed to be closely related to breakup
processes [Cox, 1992; Elliot, 1992; Brewer et al., 1992;
Rapella and Pankhurst, 1992; Storey, 1991, 1995].
[4] Marine investigations in the 1970s and 1980s
[Kristoffersen and Haugland, 1986; Hinz and Kristoffersen,
1987; Kristoffersen and Hinz, 1991] discovered most of the
large scale geological features and allowed models to be
established for the initial breakup of Gondwana [LaBrecque
and Barker, 1981; Martin and Hartnady, 1986; Kristoffersen
and Haugland, 1986; Hinz and Kristoffersen, 1987;
Livermore and Hunter, 1996]. Magnetic anomalies of
Jurassic age were reported from the Weddell Sea [LaBrecque
and Barker, 1981]. The available information supported a
scenario of Gondwana breakup starting in the southern/
southwestern Weddell Sea and propagating clockwise
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. B9, 2428, doi:10.1029/2002JB001802, 2003
1
Now at Geoforschungszentrum, Potsdam, Germany.
Copyright 2003 by the American Geophysical Union.
0148-0227/03/2002JB001802$09.00
EPM 4 -1

around Antarctica from Jurassic times to present [Lawver et
al., 1991]. As no consistent magnetic data set was available
for the Weddell, Lazarew, and Riiser-Larsen Seas (Figure 1),
the time of initial breakup in the Weddell Sea was tentatively
dated as 160–180 Ma [LaBrecque and Barker, 1981].
Because of the slow growth of available data, a wide range
of frequently controversial geodynamic models were pub-
lished during the last two decades for the Mesozoic breakup
between SAM, AFR, and ANT.
[5] Most of the age models used geological information
from the onshore magmatism as anchor points for their
reconstructions. Often geological scenarios are in direct
conflict with geophysical models for the same time period
since unacceptable overlaps between the different continen-
tal fragments are introduced [Grunow, 1993b; Storey, 1991,
1996; Dalziel and Elliot, 1982; Jokat et al., 1996]. Paleo-
magnetic investigations provided no unequivocal solution to
that problem, since only a few Mesozoic rocks are exposed
for paleomagnetic sampling [Grunow et al., 1987, 1991;
Grunow, 1993a]. Despite all these efforts, high-quality
magnetic data were not available to reliably identify sea-
floor spreading anomalies along East Antarctica between
45°W and 40°E, which provide a widely accepted geo-
dynamic model for the southern ocean.
[6] To close this gap, the Alfred Wegener Institute con-
ducted new aeromagnetic surveys along the coast of Dron-
ning Maud Land (DML) annually since the austral summer
of 1996/1997. Within the East Antarctic Margin Aeromag-
netic and Gravity Experiment (EMAGE) project a total of
90,000 km of new aeromagnetic data were acquired along a
1200 km long segment of the East Antarctic coast (Figure 1).
The aircraft flight pattern between 18°Wand8°Ewas
extended by two helicopter magnetic surveys (total of
20,000 km) based on the R/V Polarstern during the 1999/
2000 season. This approach intended to provide unequivocal
age models for the more southerly spreading anomalies off
the Explora Escarpment (Figure 1, EE) and for the Riiser-
Larsen Sea (Figure 1, RLS).
2. Experimental Setup and Processing Steps
[7] The platform used for most of the airborne measure-
ments presented in this paper is the research aircraft Polar-2.
It is a Dornier-228 twin-turboprop aircraft equipped with
Figure 1. Recent continent configuration of the ‘‘Atlantic’’ Southern Hemisphere. The major tectonic
units as well as the plate boundaries and fracture zones are marked. The flight lines of the EMAGE
surveys in the Weddell, Lazarew, and Riiser-Larsen Seas are plotted in black and white. The water depths
are contours in 500-m intervals. Abbreviations are AFR, Africa; AFZ, Agulhas Fracture Zone; AR,
Astrid Ridge; EE, Explora Escarpment; FP, Falkland Plateau; GR, Gunnerus Ridge; MEB, Maurice
Ewing Bank; MR, Maud Rise; MOR, Mozambique Ridge; RLS, Riiser-Larsen Sea; SAM, South
America; SWIR, South West Indian Ridge.
EPM 4 -2 JOKAT ET AL.: EARLY GONDWANA BREAKUP

skies for operation on snow. For safety and backup in
Antarctica, Polar-2 always operates in conjunction with its
sister aircraft Polar-4. Aeromagnetic sensing is achieved by
two Scintrex Cesium cell sensors installed in left and right
front wingtip housings. Before each survey, a dynamic
compensation flight pattern including rolls, pitches, and
yaws in the main four geomagnetic headings are flown
twice. The first set of maneuvers is used for in-flight
computing of the coefficients to eliminate the aircrafts
remanent and induced magnetic field components present
in the observations, the second set is for control and quality
checks. All instruments on board of the aircraft are previ-
ously checked for their magnetic noise level. Regarding
the presented measurements, the largest internal magnetic
noise is induced by the heating and platform control
systems of the aerogravimetric system with peak-to-peak
amplitudes of about 2 nT in the worst case. On most
flights a laser distance meter was operated to check the
flight altitude above ground. All flights were performed by
autopilot steering. The main inputs for the auto pilot
are barometric pressure for height control and GPS/INS
for heading control. The inertial navigation system on board
is a Honeywell LaserNav II. For scientific navigation
control two airborne Trimble 4000 SSI antennas and
receivers are operated. To enable offline kinematic DGPS
positioning, two to three ground reference stations were
employed to cover the survey area. In this study, more than
90% of the survey lines are distributed over the Weddell
Sea; thus GPS stations were active on the base of operation
and close to the shoreline. The remote GPS stations are
powered by solar energy buffered in batteries; the data are
stored on laptops. The same applies for the ground reference
magnetometer system which was regularly setup close the
individual base camp.
[8] The magnetic anomaly grid between 12°E and 20°W
along the continental margin off DML was calculated from
five different field campaigns between 1996 and 2002. After
individual processing of each field season the data were
upward continued to common flight levels and were
adjusted after merging. In detail, the following processing
steps were applied:
[9] 1. The data were generally edited, and the spikes were
removed.
[10] 2. Reducing the data for the ambient local field was
done by subtracting the calculated IGRF (International
Geomagnetic Reference Field) from the data.
[11] 3. Correcting for the daily variations of the Earth’s
magnetic field is a bit more difficult because flying over the
ocean means that base stations can be deployed only on one
side of the flight path. Thus the maximum difference
between the nearest base station and the measuring system
on board the plane was 500 km. For the helicopter-borne
survey between 64°S, 20°W and 68°S, 18°W it was even
twice that much. To overcome this problem, flight times
were adjusted to the predominantly magnetically quiet
hours to avoid measurements during magnetically disturbed
times. This allows the use of magnetometer data from the
Neumayer Station as base station data for the helicopter-
borne campaign and all other surveys for the times where no
other data are available.
[12] 4. Since the surveys were flown at different flight
levels all data were continued to a common datum of
1500 m. This means an upward continuation of 1300 m
for the helicopter borne survey and 100 – 500 m for all other
surveys. The equivalent source algorithm of Hansen and
Miyazaki [1984] has been used to calculate the continued
field.
[13] Finally, all campaigns were merged in one database
and a systematic adjustment/leveling was applied. The mean
mistie of all 3608 cross points before the adjustment was
31.7 nT and could be reduced to 9.4 nT after systematic
adjustment (Figure 2). Gridding was done using GMT
(Generic Mapping Tools by Wessel and Smith [1998])
gridding tools. The resulting grid has a cell size of 2 2
km and is plotted on a polar stereographic projection with
the Greenwich meridian as central meridian. The data are
illuminated from an azimuth of 340°(Figure 3b).
3. The New Magnetic Database
3.1. Antarctic-South American Sector
[14] Between 10°and 15°W longitude the most pro-
nounced magnetic anomalies are found north of 70°S
latitude (Figure 3). A sector has been chosen to extend the
fixed-wing data northward and to establish an unequivocal
age model for the opening of the Weddell Sea. Clear coast
parallel magnetic spreading anomalies are visible between
0°and 20°W. Starting in the north a pronounced negative
anomaly of 300 nT at 64°50
0
S, 18°W is observed. In
conjunction with the anomaly pattern farther north, this
Figure 2. Mistie values of the five magnetic surveys
before (grey; mean 31.7 nT) and after (black; mean 9.4 nT)
detailed leveling.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -3

anomaly is identified as chron C34 (Figure 3). C34N extends
southward to 68°S, 14°W. As expected, within the 390-km-
long Cretaceous Quiet Zone, no continuous magnetic
anomalies are present (Figure 3). The Mesozoic M1N
anomaly is located at 68°S, 14°W, striking SW-NE, and
continues to a pattern of short-wavelength, low-amplitude
anomalies. These anomalies are interpreted to represent M3
to M14N. A strong positive magnetic anomaly off the
Explora Wedge at 70°S, 12°W marks the boundary between
transitional and oceanic crust. From seismic reflection data it
is known that the Explora Wedge is build up by volcanic
seaward dipping units formed shortly before breakup [Hinz
and Kristoffersen, 1987; Jacobs et al., 1996].
[15] Farther southwest along the EE, progressively older
spreading anomalies occur. They represent chrons M16N to
M19N (Figure 3). This anomaly pattern suggests that the EE
was not formed during one magnetic chron but rather by a
propagating rift system. According to our anomaly identi-
fication the rift propagated with a speed of 6.3 cm/yr along
the EE toward the NE between M19N and M17N.
[16] The spreading velocity from M14N to M11N is
modeled with a 1.5 cm/yr half rate (Figure 4), decreasing
to 0.9–1.0 cm/yr half rate until C32N. The drop in spread-
ing velocity is associated with the occurrence of NW-SE
trending gravity ridges at 68°S first discovered on satellite
altimeter data [McAdoo and Laxon, 1996]. Combining the
new magnetic data with fracture zones derived from the
ERS-1 satellite altimeter data shows that the Falkland
Plateau forms the conjugate margin to the DML between
40°Wand8°W. In contrast to other published models
[Livermore and Hunter, 1996], we place the initial position
of SAM as close as possible to ANT. It is only this position
Figure 3a. Aeromagnetic data acquired with fixed wing and helicopter mounted sensors (black,
positive). The fixed-wing flight patterns were designed to cover the proposed continent-ocean boundary
[Jokat et al., 1996] in this area and to map magnetic anomalies as far north as 68°S with Dornier aircraft.
Because of the expected complexity of the early spreading history a line spacing of 9 km was chosen. The
flight level was 300 m during most surveys. To extend the survey area north of 68°S, a 130-km-wide
corridor (dashed box) was investigated with a helicopter mounted magnetic sensor at approximately
15°W. Some of the identified C and M series magnetic anomalies are annotated. The positive amplitudes
range between 40 and 800 nT.
EPM 4 -4 JOKAT ET AL.: EARLY GONDWANA BREAKUP

of SAM in the very southern Weddell Sea that explains the
existing magnetic and gravity data.
3.2. Antarctic-African Sector
[17] Two additional new data sets describe the early plate
movements of the African continent. The first survey is
located in the Lazarew Sea south of the Maud Rise
(Figure 3); the second one is located in the RLS (Figure 5).
In the Lazarew Sea the magnetic anomalies are generally
oriented parallel to the coast, but show a more complex
pattern in direction and are less strong in amplitude than in
the west. Starting in the north a strong positive anomaly is
visible. In combination with the western survey, it is inter-
preted to represent M1N (Figure 3). The next anchor chron is
M12, which can be followed from the SAM sector (15°Wto
8°W) toward the east. However, at 69°S, 3°W there are
isolated anomalies, which are difficult to interpret. Thus no
reliable spreading velocities can be calculated from these
data between M1N and M12. The divergent movements of
SAM and AFR in this region are most likely the reason for
this complex magnetic pattern. A deep seismic refraction
profile acquired in the LS at 6°E helps to define the onset of
the oceanic crust in this region. The analysis of these wide-
angle data [Ritzmann, 2000] indicates that true oceanic crust
is found at 68°07
0
S, 6°E, which can be correlated with the
pair of strong magnetic anomalies in that region (Figure 3).
The slightly more northern anomaly might already represent
a true seafloor spreading anomaly. The area between the
continent-ocean transition (COT) at approximately 68°Sin
the Lazarew Sea and anomaly M0 at the southern part of
Maud Rise has the same width than in the region off 15°W.
Thus we propose that the Lazarew Sea opened in N-S
direction with the same spreading velocities like in the
eastern Weddell Sea.
[18] A better view of the early movements of the AFR
plate concerning its spreading velocities and direction is
available from the RLS (Figure 5). This magnetic data
survey covers the whole range of Mesozoic anomalies in
the RLS from its COT at 70°S to almost 62°S. Here,
anomaly M0 is identified at 63°S, 26°30
0
E. Southward
long-wavelength anomalies M1N to M4N occur, followed
by a number of high-frequency anomalies. In our interpre-
tation M21N is located at 67°S, 21°E. The spreading
velocity from M0 to M11N is 1.8 cm/yr half rate (Figure 6).
The azimuth of the magnetic anomalies is approximately
035°. The spreading velocity is twice as fast as in the Weddell
Sea just north of the EE for the same time period. Then,
between M16N and M12 the spreading velocities dropped
to 1.0 cm/yr half rate. This is the same time period as when
East and West Gondwana finally separated. From breakup
Figure 3b. EMAGE aeromagnetic data grid for the Weddell and Lazarew Seas acquired from 1996/
1997 to 2000/2001 (red, positive). All data were upward continued to a flight level of 1500 m. At 15°W
the magnetic anomalies are marked by dashed lines that were used later on for the reconstruction. They
are the same lines as displayed in Figure 7 for this area.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -5

until chron M16N the half rates had steadily decreased from
2.5 to 1.6 cm/yr half rate. The strike of the southernmost
magnetic anomalies is parallel to the continental margin.
Here, the oldest identified chron is M24N. Strong positive
anomalies up to 600 nT mark the onset of transitional crust
off RLS continental margin. Between these anomalies and
M24N the magnetic field is extremely smooth with low
amplitudes. They have most likely been formed during
the fast reversals between M24N and M29. However, it
is not clear if the slightly negative anomaly of 40 nT at
69°S represents M29 (160 Ma). It has not been included in
the modeling, as the identification is quite hypothetical due
to the probably thick sediments attenuating the magnetic
signals.
[19] The azimuth of the anomalies younger than M24N
changes steadily to 035°, while the oldest magnetic anoma-
lies indicate a breakup parallel to the present-day coast line.
Off Mozambique, the conjugate to the RLS, the oldest
magnetic anomaly was dated to represent M22 [Segoufin
and Patriat, 1980].
4. Early Gondwana Breakup: The New Model
4.1. Antarctica-South America
[20] All published models suffered from the lack of
identified spreading anomalies off the EE and in the
Lazarew Sea as well as the controversial interpretation of
the magnetic anomalies in the RLS and its conjugate margin
off Mozambique [Segoufin, 1978; Segoufin and Patriat,
1980; Bergh, 1987; Rao et al., 1992; Roeser et al., 1996]. In
Figure 7 a new model for the early Gondwana breakup
between SAM, AFR, and ANT is proposed. For the rotation
poles applied see Table 1. Starting in the west, the magnetic
data indicate that the age of the EE is not 180 Myr.
According to our model the EE at 15°W formed around
138 Ma. This is in excellent agreement with ODP drilling
results in the vicinity of the Wegener Canyon (ODP Sites
692 and 693), where shallow water black shales of Valan-
ginian-Hauterivian (Site 692, 138 – 124 Ma) and Albian/
Aptian age (Site 693, 110 Ma) were drilled [Mutterlose and
Wise, 1990]. The ODP results indicate that anoxic condi-
tions were present at that time. In the past, the presence of
the shallow water deposits was difficult to explain with an
overall age of 180 Ma for the EE. The new age model for
the breakup in the area of the drill holes avoids these
difficulties. During the deposition of the black shales the
continental margin was much younger than previously
assumed.
[21] Another consequence is that SAM is much closer
positioned to ANT than in most of the published models.
Although we describe here only coastal seafloor anomalies
between 20°W and 20°E, U.S.-Argentine-Chile aeromag-
netic data in the northern Weddell Sea and along the eastern
margin of the Antarctic Peninsula strongly support our
interpretation. A pronounced anomaly now dated to M1N
is continuous from our area of investigation to approxi-
mately 45°W[LaBrecque and Ghidella, 1997]. This is also
supported by the continuity of the basement ridges in the
satellite gravity data [McAdoo and Laxon, 1996]. The
basement ridges in combination with the new magnetic
data justify the position of SAM just north of the Filch-
ner-Ronne Shelf (Figure 7). A tentative correlation of
anomalies of the two different data sets allows the extrap-
olation of M10 and M11 far into the western Weddell Sea
Figure 4. Synthetic model for the Weddell Sea magnetic data in the marked corridor in Figure 3a. The
magnetic data are taken from the center of the corridor shown in Figure 3. KG85 for the Mesozoic
reversal timescale [Kent and Gradstein, 1985] has been used to establish the age model. The magnetic
layer is chosen to be at a constant 6-km depth.
EPM 4 -6 JOKAT ET AL.: EARLY GONDWANA BREAKUP

Figure 5. Riiser-Larsen survey made with the helicopter system of R/V Polarstern only. The swath is
65 km wide and more than 800 km long. The flight level was almost constant at 150 m during the survey.
The line spacing is 9 km. It samples all Mesozoic magnetic anomalies in the RLS. The strong positive
anomaly in the south marks the onset of transitional or rifted continental crust. The M anomalies were
named accordingly to the Mesozoic reversal timescale KG85 [Kent and Gradstein, 1985]. For the
location of the corridor see Figure 1.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -7

(Figure 7c). A Pacific position of the northern tip of the
AP relative to Patagonia is the consequence of our model,
which is in contrast to some models based on the geology
in SAM. The new model does not explain the occurrence
of mid-Paleozoic to lower Mesozoic accretionary prism
along the west coast of Patagonia [Ling and Forsythe,
1987; Grunow et al., 1992; Mukasa and Dalziel, 1996].
However, the continuous anomalies of the U.S.-Argentine-
Chile aeromagnetic data (USAC) in the northwestern
Weddell Sea leave little choice for a different interpreta-
tion. No new constraints on the movements of the West
Antarctic terranes can be deduced from the new data
beside the position of the Antarctic Peninsula relative to
the Filchner-Ronne Shelf.
[22] In strong contrast to other models, our data show no
evidence for large-scale strike-slip movements along the EE
during the drift period of the continents. This has been
postulated in almost all published models [Martin and
Hartnady, 1986; Kristoffersen and Haugland, 1986; Lawver
et al.,1991;Cox,1992;Elliot,1992;Grunow,1993b;
Livermore and Hunter, 1996]. We used the Beattie Anom-
aly (Figure 7a) and a strong magnetic anomaly along the
Heimefrontfjella (SKA-Sverdrupfjella-Kirvanveggen-
Anomaly; Golynsky and Aleshkova, 2000) as starting point
for our model.
[23] The evidence given here for propagation of the
Weddell Sea rift system indicates that the first oceanic crust
should have formed in the southwestern corner of the
Weddell Sea. Details on its age and opening history are
unknown, since no high-quality data yet exist. Here, a more
complex spreading system including the Antarctic Peninsula
and East Antarctica might have existed. If the rift had
propagated from this region with the same speed as we have
observed north of EE (63 km/Myr), it would have needed
approximately 16 Myr to have propagated the 1000 km from
the margin of the Antarctic Peninsula. The oldest identified
magnetic anomaly along the EE is M19N (144 Ma). If, as a
rough estimate, we add the 16 Myr, an age of 160 Myr for the
oldest crust in the southern Weddell Sea can be suggested.
Toward the Lazarew Sea the marine magnetic anomalies
continuously change their spreading direction shortly after
the breakup. However, the direction at the time of breakup
for AFR still is parallel to the present ANT coastline. The
NNE strike of the seafloor anomalies just south of Maud
Rise (Figure 3) indicates that there was already extension
between SAM and AFR while both continents moved
northward as one plate. The survey in the Lazarew Sea
shows a pronounced positive anomaly close to 66°S. It is
dated as chron M1N.
4.2. Antarctica-Africa
[24] The ‘‘African margin’’ of ANT continues into the
eastern RLS. Here, conflicting studies for the age of the
RLS exist [Bergh, 1987; Rao et al., 1992; Roeser et al.,
1996]. The new helicopter magnetic data acquired between
19°E and 25°E provide an age close to 160 Ma for the
oldest parts of the RLS. Thus our data confirm the Roeser
model [Roeser et al., 1996] although differing in detail.
The new data indicate that the earliest spreading direction in
the area investigated was not as oblique as previously
Figure 6. Synthetic model for the Riiser-Larsen Sea. The center trace of the corridor has been chosen as
reference. The magnetic layer is chosen at a constant depth of 6 km. Closer to the margin, there is a strong
misfit in amplitude between theoretical and field data. This is most likely due to a thicker sediment cover.
However, no published information is available to better constrain our model in this part.
EPM 4 -8 JOKAT ET AL.: EARLY GONDWANA BREAKUP

suggested till chron M23 [Bergh, 1987; Rao et al., 1992;
Roeser et al., 1996]. Applying the different spreading
directions and velocities of AFR and SAM, this breakup
model implies no strike slip along the EE during the drift
period. For the initial position of AFR relative to ANT we
considered two pronounced magnetic anomalies along
the Cape Fold Belt (Figure 7a, Beattie-A) and in DML
(Figure 7a, western part of SKA) to be a continuous
prebreakup feature [De Beer and Meyer, 1983; Corner
and Groenewald, 1991; Golynsky and Aleshkova, 2000].
The Beattie Anomaly in southern Africa has a strike length
of approximately 900 km. Its counter piece in Antarctica
has a similar signature in terms of amplitude (100 –900 nT),
wavelength, and strike length (>600 km). The age of the
Beattie Anomaly is unknown, and speculations range
from 500 Ma (Cape orogeny [Corner et al., 1991]) to a
Grenville age [De Beer and Meyer,1983;Corner and
Groenewald, 1991]. Both features may belong to the
Namaqua-Natal-Maud Belt, which can be partly mapped
in southern Africa and Dronning Maud Land. Thus this is
an anchor point for AFR relative to ANT for our start
model. To explain the magnetic anomalies in the RLS and
their conjugate in the Mozambique Basin, AFR has to be
shifted to the east relative to Antarctica. In our model the
Figure 7a. Plate reconstructions for 155 Ma based on the new magnetic data. At this time only seafloor
spreading in the RLS constrain our model. For the initial fit between AFR and ANT, pronounced
magnetic anomalies on both continents are used [Corner and Groenewald, 1991; Golynsky and
Aleshkova, 2000]. The possible connection (hatch area) between the Beattie Anomaly and the
Sverdrupfjella-Kirvanveggen Anomaly (SKA, dashed line) is marked. Please note that the spreading
history in the westernmost Weddell Sea is unknown, and therefore the extension of this early rift basin is
not plotted. The bold line indicates the position of the spreading ridges. Note that the spreading history
between the Antarctic Peninsula and SAM is not well constrained. This area is marked by question
marks. Furthermore, no efforts have been undertaken to remove any stretching of the Falkland Plateau,
since it is not well known. So, the fit can be enhanced if the extensional history is better constrained.
Furthermore, some of the geology, which is relevant for the position of SAM relative to the AP, is
included. The locations of the mid-Paleozoic to lower Mesozoic accretionary prism are taken from Ling
and Forsythe [1987], and the description of the Roccas Verdes ophiolites is from Mukasa and Dalziel
[1996] and Grunow et al. [1992]. Abbreviations are AP, Antarctic Peninsula; AR, Astrid Ridge; BDS,
Botswana Dyke Swarm; Beattie-A, Beattie Anomaly; EE, Explora Escarpment; FRS, Filchner Ronne
Shelf; GR, Gunnerus Ridge; IND, India; MAD, Madagascar; MOZB, Mozambique Basin; RLS, Riiser-
Larsen Sea; RVO, Roccas Verdes ophiolites; SKA, Sverdrupfjella-Kirvanveggen Anomaly; SRI, Sri
Lanka.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -9

early rotation poles explain these movements as a conse-
quence of the divergent drift of AFR and SAM relative to
ANT. In this case the strike-slip movements occurred in the
area of the LS between 160 and 140 Ma. It might be that
during this period the Agulhas Fracture Zone (Figure 1) was
already active as an intracontinental fault zone. However,
the exact timing of these movements is not known. For an
enhanced model the amount of stretching beneath Mozam-
bique is critical to close the gap in our reconstruction for the
eastern RLS (Figure 7a).
[25] Independent evidence for such early strike-slip
movements comes from the analysis of dolerite dikes in
the western Dronning Maud Land [Grantham, 1996]. Two
different strike directions in western Dronning Maud Land
were found, a NE and slightly west of north directions.
Combining this information with dike orientations in south-
ern Africa [Grantham, 1996], a strike-slip movement of
ANT in easterly directions between 190 and 200 Ma has
been postulated. In a second stage the slightly west of north
dikes intruded at circa 170 – 180 Ma.
[26] The pronounced differences in spreading velocity
and direction of the SAM and AFR plates caused a constant
stretching of the continental crust along their southern
boundaries. The South Atlantic started to open around
M9N [Rabinowitz et al., 1983] as a consequence of these
divergent plate movements.
5. Implications for Understanding of the Onshore
Geology and the Opening of the South Atlantic
[27] In this section the relationship of the new model with
the onshore geology is discussed. It will focus on some
striking similarities and is by far not complete. The new data
have some far-reaching consequences for the interpretation
of onshore geology in SAM, AFR, and DML. Onshore
magmatism occurred as early as 200 Ma, 193 Ma, 178 Ma,
165 Ma, 150 Ma, and 137 Ma in the Karoo Province
[Brewer et al., 1992], the Cape Province, Lesotho, Swazi-
land, and the Lebombo Monocline [Cox, 1992; Elliot, 1992;
Brewer et al., 1992]. In the conjugate DML province
magmatism appeared at circa 182 Ma and at circa 172 Ma
[Cox, 1992; Elliot, 1992; Brewer et al., 1992]. The magma-
tism clearly predates the opening of the RLS by some 12 –
22 Myr.
Figure 7b. Plate reconstructions for 145 Ma. Here the model is constrained by spreading anomalies in
the Weddell Sea and RLS. In the Lazarew Sea, no spreading was active. The hatched area west of Astrid
Ridge (AR) indicates the location of a shallow sea or a still subaerial region. The thin lines in the Weddell
and Riiser-Larsen Seas indicate dated magnetic anomalies. The dotted box in the RLS indicates the
magnetic anomalies from the Mozambique Basin. Abbreviations are AR, Astrid Ridge; BDS, Botswana
Dyke Swarm; Beattie-A, Beattie Anomaly; EE, Explora Escarpment; FP, Falkland Plateau; FRS, Filchner
Ronne Shelf; GR, Gunnerus Ridge; IND, India; MAD, Madagascar; MOZB, Mozambique Basin; RLS,
Riiser-Larsen Sea; RVO, Roccas Verdes ophiolites; SKA, Sverdrupfjella-Kirvanveggen Anomaly; SRI,
Sri Lanka; WS, Weddell Sea.
EPM 4 -10 JOKAT ET AL.: EARLY GONDWANA BREAKUP

[28] During the Karoo magmatism in Africa, the exten-
sive Lebombo volcanics were emplaced. These sequences
have an age of 190 Myr. In various models the Lebombo
sequences have been interpreted to be the conjugate of the
Explora Wedge volcanic sequences [Cox, 1992; Elliot,
1992]. A 190 Myr age for the EE cannot be confirmed by
the new magnetic data. If the model is applied that volcanic
seaward dipping reflector sequences erupted in space and
time close before the first oceanic crust formed [Hinz
and Kristoffersen, 1987; White and Mackenzie, 1989], the
Explora Wedge has an age of approximately 138 Myr at
15°W. Hence the wedge is significant younger than 190 Myr,
and the conjugate shoulder of the Lebombo volcanic
sequences may be located between the Heimefrontfjella-
Kirvanveggen (Figure 7) and the EE. As a consequence, a
long-lived rift system stretching the continental crust be-
tween AFR and ANT was active for more than 30 Myr in this
area. In DML only a few volcanic rocks with an age of 190
Myr were found in the Vestfjella [Peters, 1989]. That a
thermal event influenced the tectonic evolution of DML is
Figure 7c. Plate reconstructions for 131 Ma. East (ANT) and West Gondwana (AFR/SAM) finally
separated. A continuous spreading system between the Weddell and Riiser-Larsen Seas was established.
The new COT is labeled Explora Wedge-East. The dot-dash-dotted line in the western Weddell Sea
represents a pronounced magnetic anomaly digitized from LaBrecque and Ghidella [1997]. The dotted
box in the RLS indicates the magnetic anomalies from the Mozambique Basin. Abbreviations are AP,
Antarctic Peninsula; AR, Astrid Ridge; BDS, Botswana Dyke Swarm; Beattie-A, Beattie Anomaly;
EA, East Antarctica; EE, Explora Escarpment; EW-E, Explora Wedge-East; FRS, Filchner Ronne
Shelf; GR, Gunnerus Ridge; LS, Lazarew Sea; MEB, Maurice Ewing Bank; MOZB, Mozambique
Basin; MOZR, Mozambique Ridge; RLS, Riiser-Larsen Sea; RVO, Roccas Verdes ophiolites; SKA,
Sverdrupfjella-Kirvanveggen Anomaly; WS, Weddell Sea.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -11

suggested by apatite fission track analysis [Jacobs et al.,
1996]. The apatite fission track data show that the onset of
tectonically induced denudation causing basement cooling
started no earlier than circa 140 Ma. This correlates remark-
ably well with the first formation of oceanic crust north of
this area. With the onset of seafloor spreading, most of the
heat beneath the continent was channeled through the new
mid-ocean ridge system.
[29] On the conjugate margin, a large volcanic feature, the
Astrid Ridge, has been postulated to be closely related to
the breakup [Bergh, 1987; Rao et al., 1992; Roeser et al.,
1996]. From the new magnetic data it is obvious that
oceanic crust formed in the RLS from approximately 160
Ma onward, while west of the Astrid Ridge, no seafloor
spreading anomalies are visible. Here, stretched continental
crust was present. In our interpretation Astrid Ridge marked
a continental margin from M24N to approximately M12.
This area was highly stretched before breakup, and this
might have created local volcanism during the initial and
final breakup of Africa. It is therefore very likely that the
African plate did not act as a single plate for some time.
Since no basement rocks were recovered so far from the
Figure 7d. Plate reconstructions for 122 Ma. AFR and SAM are by now separated, as they move
northward with respect to ANT at different spreading velocities. The connection of the spreading systems
in the Natal Basin and Weddell Sea/LS is not well constrained and therefore not plotted. The dot-dash-
dotted lines in the western Weddell Sea represent pronounced magnetic anomalies digitized from
LaBrecque and Ghidella [1997]. Please note that the position of SAM relative to the AP is not correct,
since no constraints exist on its position from the existing magnetic data. The dotted box in the RLS
indicates the magnetic anomalies from the Mozambique Basin. Abbreviations are AP, Antarctic
Peninsula; AR, Astrid Ridge; BDS, Botswana Dyke Swarm; Beattie-A, Beattie Anomaly; EE, Explora
Escarpment; EW-E, Explora Wedge East; FP, Falkland Plateau; FRS, Filchner Ronne Shelf; GR,
Gunnerus Ridge; LS, Lazarew Sea; MAD, Madagascar; MEB, Maurice Ewing Bank; MOZB,
Mozambique Basin; MOZR, Mozambique Ridge; RLS, Riiser-Larsen Sea; RVO, Roccas Verdes
ophiolites; SKA, Sverdrupfjella-Kirvanveggen Anomaly; SRI, Sri Lanka; WS, Weddell Sea.
EPM 4 -12 JOKAT ET AL.: EARLY GONDWANA BREAKUP

Astrid Ridge, no further constraints on its age and compo-
sition are available.
[30] The new magnetic data show that basin evolution in
the area investigated was different from that previously
suggested. No simple rift system propagated from west to
east. Moreover, three large separate ocean basins formed
during Jurassic times. One was located in the western
Weddell Sea, propagating from west to east; the second in
the Mozambique Basin/RLS, and the third north of Mada-
gascar [Goodlad et al., 1982] in the Somali Basin. Here,
Mesozoic anomalies back to chron M24N are identified.
While these ocean basins grew, AFR and parts of the
Falkland Plateau still were connected to Antarctica. Here,
the lithosphere of the old craton was strong enough to resist
the final separation.
[31] The oldest magnetic spreading anomalies for the
opening of the South Atlantic are chron M10N in the Natal
Basin [Goodlad et al., 1982] and its conjugate, the Falkland
Plateau [Martin et al., 1982]. Chron M9N (‘‘Cape Sequen-
ces’’) is reported from the southernmost South Atlantic for
the first oceanic crust there [Rabinowitz and LaBrecque,
1979]. Although the magnetic anomalies in the South
Atlantic and the Natal Basin are not well constrained, it
seems likely that the spreading centers evolved more or less
at the same time. Together with the earliest magnetic
anomalies of M14N along the East Antarctic margin in
the Lazarew Sea, a complex ridge/rift system must have
existed around 140 Ma in that area. The final separation of
Antarctica and Africa, the opening of the Natal Basin, and
thus the initial opening of the South Atlantic occurred
within approximately 10 Myr (M14N-M9N). The large
strike-slip movement of the Falkland Plateau/South
America along the Agulhas Fracture zone is a direct
consequence of the different spreading azimuths of the
SAM and AFR plates. No major plate reorganization was
necessary between AFR and SAM to open the South
Atlantic. As a consequence, the continental crust between
SAM and AFR began to stretch at least with the occurrence
of the first M anomalies in the RLS at about 155 Ma.
Evidence for a rift phase in the South Atlantic is reported
from the Orange Basin. At 155 Ma, rift-related volcanism
was terminated by a regional uplift and block rotation [Light
et al., 1992]. At least 27 Myr of divergent stress finally
resulted in the formation of new oceanic crust in the Natal
Basin and the South Atlantic.
6. Remaining Problems
[32] In the following, the remaining problems on the
origin of some structures will be discussed. The new data
provide no constraints for the initial movement of West
Antarctica and here especially the AP. There is an ongoing
debate how far the AP did overlap with Patagonia. In our
model the AP has been left in its present-day position
relative to East Antarctica. The magnetic data along the
eastern margin of the AP show no unique correlation
[Ghidella et al., 2002; Kovacs et al., 2002]. In these two
publications the same magnetic data set has been interpreted
in a completely different way.
[33] From onshore geological mapping no unique con-
straints exist, but a northerly position of SAM in the
Weddell Sea is preferred. Constraints might be derived from
onshore Permo-Carbonifereous and Jurassic accretionary
prism rocks along SAM [Ling and Forsythe, 1987; Grunow
et al., 1992; Mukasa and Dalziel, 1996]. It cannot be ruled
out the AP and SAM underwent a more complex geological
history than the current magnetic data can explain, e.g., that
the AP had a more southern position during the creation of
the Permo-Carbonifereous units. Since the conjugate
anomalies were destroyed by the overriding Scotia plate,
the oceanic basement in southern Weddell Sea is covered
by a thick pile of sediments [Rogenhagen and Jokat, 2000],
and the spreading velocities might have been very slow
(<1 cm/yr [Ghidella et al., 2002]), it will be difficult to
derive a conclusive geophysical model for this area from the
existing data. It is therefore necessary to acquire more
detailed magnetic data in the western Weddell Sea.
[34] Paleomagnetic data [Grunow et al., 1987] indicate a
post-Mid-Jurassic 30°clockwise rotation of the AP relative
to East Antarctica. However, the data do not provide good
constraints for the longitudinal position of the blocks
relative to each other. Between both blocks a huge sedi-
mentary basin, the Filchner-Ronne Shelf, exists. Following
results from seismic refraction and reflection work along the
Filchner Ronne Shelf [Hu¨ bscher et al., 1996; Jokat et al.,
1997], we suggest that the AP was attached closer to East
Antarctica during the early rift period in the Mesozoic.
According to these results the basement beneath the
Filchner-Ronne region was stretched by factor of 1.5 to
3.0. Evidences for strike-slip movements have not been
reported. However, the stretching led not to a formation of a
vast amount of oceanic crust between AP and East Antarc-
tica [Hu¨ bscher et al., 1996] making a movement of the
EWM terrane from the southern Weddell Sea to its present-
day position extremely unlikely [Grunow, 1993b]. Our data
do not support any Mesozoic subduction of Jurassic Wed-
dell Sea oceanic crust beneath the eastern margin of the AP.
For the interpretation of the paleomagnetic data of this
region we prefer an early model for the West Antarctic
terranes of Storey [1991]. Here, the paleomagnetic data
Tab l e 1 . Euler Poles Applied to Describe West and East
Gondwana Movements
a
Time,
Ma
Latitude,
deg
Longitude,
deg
Rotation,
deg Reference
79.08 4.7 320.3 16.04 C33o [Royer et al., 1988]
83.00 5.7 320.8 17.85 C34y [Royer et al., 1988]
127.50 14.46 339.11 47.05 this study
130.00 13.10 337.75 47.67 this study
131.00 11.59 336.45 47.40 this study
135.00 8.25 333.09 47.22 this study
141.00 7.66 331.16 49.24 this study
144.50 4.59 327.93 49.34 this study
146.50 3.63 327.19 49.71 this study
148.00 2.92 325.60 49.81 this study
153.00 3.23 325.60 51.86 this study
159.00 4.64 324.80 55.29 this study
200.00 4.65 324.80 55.29 this study
a
See Figure 7. ATLASWIN 1.11 and TIMETREK 3.1 software were used
to calculate the reconstruction. Poles describe the reconstruction of ANT
with respect to AFR only. The direction and velocities of the two spreading
systems in the Weddell Sea and the Riiser-Larsen Sea are described by a
clockwise rotation of ANT relative to AFR. For the early movements
between SAM and AFR we have no constraints. From previous studies
[Lawver et al., 1998] it is suggested that they did not behave as rigid blocks.
Therefore we introduced no new Euler poles for SAM.
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -13

were explained by small relative movements of the West
Antarctic crustal blocks, but leaving them more or less at
the same position relative to East Antarctica they have
today. The conflict with the interpretation of the
paleomagnetic data [Grunow, 1993b] might be explained
with the usage of different reference poles than derived in
this study.
[35] In the ‘‘African sector’’ of Antarctica the open
questions focus mainly in the LS and the Lebombo/Mozam-
bique region. In Figure 7a the direct neighborhood of the EE
with the Lebombo volcanics is striking, but both features
were most likely created during a different time period. The
question is, where the conjugate features for the Lebombo
volcanics, if any, are located. We speculate that the area
north of the magnetic SKA somehow marks the region,
which was affected by the Jurassic rift process in this
particular sector. According to our model the LS has been
an area of extensive stretching for at least 20 Myr. The
amount of stretching in Mozambique and DML is still
unknown. Thus we chose a loose fit for the oldest recon-
struction. The obvious gap between Mozambique and the
Astrid Ridge (Figure 7a) might be closed if we have better
constraints on stretching parameters.
[36] In the same area the Mozambique Ridge (MOZR)
formed some time between 145 and 131 Ma according to
our model. There is no conclusive evidence on its origin
(oceanic versus continental). No magnetic anomalies are
visible in the LS that shed new light on this problem. From
our reconstruction we prefer the interpretation of the MOZR
as a continental sliver once attached to Antarctica. Hence
the southern margin of the Mozambique ridge was the
conjugate part of AFR to the LS continental margin before
the Natal Basin formed. During the rift processes in the
Natal Basin and the LS the crust might be heavily intruded
and overprinted by simultaneous volcanism along its north-
ern and southern boundaries. For a closer fit of the con-
tinents and a better understanding of the breakup processes
it is furthermore essential to investigate in greater detail, if
the spreading in the MOZB and RLS was symmetrical or
not. From the existing databases it is also not clear if a
propagating rift system existed here or if both basins opened
in one step.
7. Implications for Breakup Processes
[37] The timing and geometry of our new breakup model
shows in time no direct influence of a hot spot. The initial
volcanism in SAM and AFR occurred between 200 and
180 Ma, approximately 40 – 20 Myr before the first oceanic
crust formed in the area investigated. The large time delay
between the Karoo magmatism and the initial separation of
SAM/AFR and ANT indicates that the mantle plume did not
provide the essential trigger for Gondwana breakup. How-
ever, it might have controlled the ultimate position of the
major fault systems. The scenario described here shows that
divergent mantle flows existed in the central part of Gond-
wanaland and caused the separation of AFR/SAM from
ANT. For unknown reasons the fluid dynamics in the
mantle caused spreading velocities on the surface which
differ as much as 100% (SAM, 0.9 cm/yr; AFR, 1.8 cm/yr)
and in spreading direction by 60°(SAM, 335°; AFR, 035°)
for the overlapping time period. The transition between the
two spreading regimes is in the Lazarew Sea. Here, a diffuse
N-S spreading is found. May be here no regular spreading
could be established due to the emplacement of the large
Maud Rise complex. The differences in spreading velocities
and directions were stable more or less through the Meso-
zoic and Cenozoic [Royer et al., 1988].
[38] For the breakup between West and East Gondwana,
only a single mantle source is required for our geometrical
reconstruction. The different spreading velocities and direc-
tions are modeled by a clockwise southward rotation of
ANT relative to AFR/SAM. However, the model does not
provide any constraints for mantle processes during Gond-
wana breakup.
[39]Acknowledgments. We would like to thank the flight crews
(aircraft/helicopter) for their essential and excellent support during the field
seasons. Special thanks to D. Steinhage for his careful planning in the field of
the last campaign. Without the constant support of Heinz Miller this project
would have been significantly delayed. H. Roeser, B. Schreckenberger, and
G. Eagles provided essential support for the modeling of the synthetics. For
the modeling, Magbath software has been used, which is originally based on
work of D. Naar, J. Morgan, and D. Wilson.
References
Barker, P. F., and J. P. Kennett, Proceedings of the Ocean Drilling Program,
Scientific Results, vol. 113, Ocean Drill. Program, College Station, Tex.,
1990.
Bell, R. E., J. M. Brozena, W. F. Haxby, and J. L. LaBrecque, Continental
margins of the Western Weddell Sea: Insights from airborne gravity and
Geosat-derived gravity, in Contributions to Antarctic Research I,Antarct.
Res. Ser., vol. 50, edited by D. H. Elliot, pp. 91 – 102, AGU, Washington,
D. C., 1990.
Bergh, H. W., Underlying fracture zone nature of Astrid Ridge off Antarc-
tica’s Dronning Maud Land, J. Geophys. Res.,92, 475 – 484, 1987.
Brewer, T. S., J. M. Hergt, C. J. Hawkesworth, D. Rexmand, and B. C.
Storey, Coats Land dolerites and the generation of Antarctic flood basalts,
in Magmatism and the Causes of Continental Break-up, edited by B. C.
Storey, T. Alabaster, and R. J. Pankhurst, Geol. Soc. Spec. Publ.,68,
185 – 208, 1992.
Corner, B., and P. B. Groenewald, Gondwana reunited, S. Afr. Trans. Nav.
Antarkt.,21(2), 172, 1991.
Corner, B., J. C. Maccelari, and S. Niccol, Major magnetic anomalies in
Western Dronning Maud Land; their possible origin and correlates to
southern Africa, paper presented at Sixth International Symposium on
Antarctic Earth Sciences, Natl. Inst. of Pol. Res., Tokyo, 1991.
Cox, K. G., Karoo igneous activity and the early stages of the break-up of
Gondwanaland, in Magmatism and the Causes of Continental Break-up,
edited by B. C. Storey, T. Alabaster, and R. J. Pankhurst, Geol. Soc. Spec.
Publ.,68, 137 – 148, 1992.
Dalziel, I. W. D., and D. H. Elliot, West Antarctica: Problem child of
Gondwanaland, Tectonics,1(1), 3 – 19, 1982.
De Beer, J. H., and R. Meyer, Geoelectrical and gravitational characteristics
of the Namaqua-Natal mobile belt and its boundaries, Spec. Publ. Geol.
Soc. S. Afr.,10, 91 – 100, 1983.
Elliot, D. H., Jurassic magmatism and tectonism associated with Gondwa-
naland break-up: An Antarctic perspective, in Magmatism and the
Causes of Continental Break-up, edited by B. C. Storey, T. Alabaster,
and R. J. Pankhurst, Geol. Soc. Spec. Publ.,68, 165 – 184, 1992.
Ghidella, M. E., G. Yanez, and J. L. LaBrecque, Revised tectonic implica-
tions for the magnetic anomalies of the western Weddell Sea, Tectono-
physics,347, 65 – 86, 2002.
Golynsky, A. V., and N. D. Aleshkova, Regional magnetic anomalies of the
Weddell Sea region and their geological significance, Polarforschung,
67(3), 101 – 117, 2000.
Goodlad, S. W., A. K. Martin, and C. J. H. Hartnady, Mesozoic magnetic
anomalies in the southern Natal Valley, Nature,295, 686 – 688, 1982.
Grantham, G. H., Aspects of Jurassic magmatism and faulting in western
Dronning Maud Land, Antarctic: Implications for Gondwana break-up, in
Weddell Sea Tectonics and Gondwana Break-up, edited by B. C. Storey,
E. C. King, and R. A. Livermore, Geol. Soc. Spec. Publ.,108, 63 – 71,
1996.
Grunow, A. M., New paleomagnetic data from the Antarctic Peninsula and
their tectonic implications, J. Geophys. Res.,98, 13,815 – 13,883, 1993a.
Grunow, A. M., Creation and destruction of Weddell Sea floor in the
Jurassic, Geology,21, 647 – 650, 1993b.
EPM 4 -14 JOKAT ET AL.: EARLY GONDWANA BREAKUP

Grunow, A. M., D. V. Kent, and I. W. D. Dalziel, Mesozoic evolution of the
West Antarctica and the Weddell Sea Basin: New paleomagnetic con-
straints, Earth Planet. Sci. Lett.,86, 16 – 26, 1987.
Grunow, A. M., D. V. Kent, and I. W. D. Dalziel, New paleomagnetic data
from Thurston Island: Implications for the tectonics of West Antarctica
and the Weddell Sea opening, J. Geophys. Res.,96, 17,935– 17 ,954,
1991.
Grunow,A.M.,I.W.D.Dalziel,T.M.Harrison,andM.T.Heizler,
Structural geology and geochronology of subduction complexes along
the margin of Gondwanaland; new data from the Antarctic Peninsula
and southernmost Andes, with Suppl. Data 92 – 32, Geol. Soc. Am. Bull.,
104(11), 1497 – 1514, 1992.
Hansen, R. Q., and Y. Miyazaki, Continuation of potential fields between
arbitrary surfaces, Geophysics,49(6), 787 – 795, 1984.
Hinz, K., and Y. Kristoffersen, Antarctica: Recent advances in the under-
standing of the continental shelf, Geol. Jahrb., Reihe E,37, 54, 1987.
Hu¨bscher, C., W. Jokat, and H. Miller, Structure and origin of southern
Weddell Sea crust: Results and implications, in Weddell Sea Tectonics
and Gondwana Break-up, edited by B. C. Storey, E. C. King, and R. A.
Livermore, Geol. Soc. Spec. Publ.,108, 201 – 211, 1996.
Jacobs, J., N. Kaul, and K. Weber, The history of denudation and resedi-
mentation at the continental margin of western Dronning Maud Land,
Antarctica, during break-up of Gondwana, in Weddell Sea Tectonics and
Gondwana Break-up, edited by B. C. Storey, E. C. King, and R. A.
Livermore, Geol. Soc. Spec. Publ.,108, 191 – 199, 1996.
Jokat, W., C. Hu¨ bscher, U. Meyer, L. Oszko, T. Scho¨ne, W. Versteeg, and
H. Miller, The continental margin off East Antarctica between 10°W and
30°W, in Weddell Sea Tectonics and Gondwana Break-up, edited by B. C.
Storey, E. C. King, and R. A. Livermore, Geol. Soc. Spec. Publ.,108,
129 – 141, 1996.
Jokat, W., N. Fechner, and M. Studinger, Geodynamic models of the Wed-
dell Sea embayment in view of new geophysical data, in The Antarctic
Region: Geological Evolution and Processes, edited by C. A. Ricci, pp.
453 – 459, Terra Antarct., Siena, Italy, 1997.
Kent, D. V., and F. M. Gradstein, A Cretaceous and Jurassic geochronology,
Geol. Soc. Am. Bull.,96, 1419 – 1427, 1985.
Kovacs, L. C., P. Morris, J. Brozena, and A. Tikku, Seafloor spreading in
the Weddell Sea from magnetic and gravity data, Tectonophysics,347,
43 – 64, 2002.
Kristoffersen, Y., and K. Haugland, Geophysical evidence for the East
Antarctic plate boundary in the Weddell Sea, Nature,322, 538–541,
1986.
Kristoffersen, Y., and K. Hinz, Evolution of the Gondwana plate boundary
in the Weddell Sea area, in Geological Evolution of Antarctica, edited by
M. R. A. Thomson, J. A. Crame, and J. W. Thomson, pp. 225– 230,
Cambridge Univ. Press, New York, 1991.
LaBrecque, J. L., and P. F. Barker, The age of the Weddell basin, Nature,
290, 489 – 492, 1981.
LaBrecque, J. L., and M. E. Ghidella, Bathymetry, depth to magnetic base-
ment, and sediment thickness estimates from aerogeophysical data over
the western Weddell Basin, J. Geophys. Res.,102, 7929 – 7945, 1997.
Lawver, L. A., J. Y. Royer, D. T. Sandwell, and C. R. Scotese, Evolution of
the Antarctic continental margins, in Geological Evolution of Antarctica,
edited by M. R. A. Thomson, J. A. Crame, and J. W. Thomson, pp. 533 –
540, Cambridge Univ. Press, New York, 1991.
Lawver, L. A., L. M. Gahagan, and I. W. D. Dalziel, A tight fit-Early
Mesozoic Gondwana, a plate reconstruction perspective, Mem. Natl. Inst.
Polar Res., Spec. Issue,53, 214 – 229, 1998.
Light,M.P.R.,M.P.Maslanyj,andN.L.Banks,Newgeophysical
evidence for extensional tectonics on the divergent margin offshore
Namibia, in Magmatism and the Causes of Continental Break-up, edited
by B. C. Storey, T. Alabaster, and R. J. Pankhurst, Geol. Soc. Spec. Publ.,
68, 257 – 270, 1992.
Ling, H. Y., and R. D. Forsythe, Late Paleozoic pseudo-albaillellid radi-
olarians from southernmost Chile and their geological significance, in
Gondwana Six; Structure, Tectonics, and Geophysics,Geophys.
Monogr. Ser., vol. 40, edited by G. D. McKenzie, pp. 253–260,
AGU, Washington, D. C., 1987.
Livermore, R. A., and R. J. Hunter, Mesozoic seafloor spreading in the
southern Weddell Sea, in Weddell Sea Tectonics and Gondwana Break-
up, edited by B. C. Storey, E. C. King, and R. A. Livermore, Geol. Soc.
Spec. Publ.,108, 227 – 241, 1996.
Martin, A. K., and C. J. H. Hartnady, Plate tectonic development of the
southwest Indian Ocean: A revised reconstruction of East Antarctica and
Africa, J. Geophys. Res.,91, 4767 – 4786, 1986.
Martin, A. K., S. W. Goodlad, and D. A. Salmon, Sedimentary basin in-fill
in the northernmost Natal Valley hiatus development and Agulhas current
palaeo-oceanography, J. Geol. Soc.,139, 183 – 201, 1982.
McAdoo, D. C., and W. Laxon, Marine gravity from Geosat and ERS-1
altimetry in the Weddell Sea, in Weddell Sea Tectonics and Gondwana
Break-up, edited by B. C. Storey, E. C. King, and R. A. Livermore, Geol.
Soc. Spec. Publ.,108, 155 – 164, 1996.
Mukasa, S. B., and I. W. D. Dalziel, Southernmost Andes and South Geor-
gia Island, North Scotia Ridge; zircon U-Pb and muscovite
40
Ar/
39
Ar age
constraints on tectonic evolution of southwestern Gondwanaland, J. S.
Am. Earth Sci.,9(5 – 6), 349 –365, 1996.
Mutterlose, J., and S. W. Wise, Lower Cretaceous nannofossil biostratigra-
phy of ODP Leg 113 Holes 692B and 693A, continental slope off East
Antarctica, Weddell Sea, Proc. Ocean Drill. Program Initial Rep.,113 ,
325 – 352, 1990.
Peters, M., Igneous rocks in western and central Neuschwabenland,
Vestfjella and Ahlmannryggen, Antarctica. Petrography, geochemistry,
geochronology, paleomagnetism, geotectonic implications, Ber. Polar-
forschung,61, 1989.
Rabinowitz, P. D., and J. L. LaBrecque, The Mesozoic South Atlantic
Ocean and Evolution of its continental margins, J. Geophys. Res.,84,
5973 – 6002, 1979.
Rabinowitz, P. D., M. F. Coffin, and D. A. Falvey, The separation of
Madagascar and Africa, Science,220, 67 –69, 1983.
Rao, D. G., A. K. Chaubey, and T. Ramprasad, The Cretaceous-Tertiary sea
floor off Dronning Maud Land, Antarctica, Tectonophysics,205, 447–
452, 1992.
Rapella, C. W., and R. J. Pankhurst, The granites of northern Patagonia and
the Gastre Fault System in relation to the break-up of Gondwana, in
Magmatism and the Causes of Continental Break-up, edited by B. C.
Storey, T. Alabaster, and R. J. Pankhurst, Geol. Soc. Spec. Publ.,68,
209 – 220, 1992.
Ritzmann, O., Refraktionsseismische Untersuchungen am Kontinentalrand
der Ostantarktis, Diploma thesis, 89 pp., Rheinische-Friedrich-Wilhelms-
Univ. Bonn, Bonn, Germany, 2000.
Roeser, H. A., J. Fritsch, and K. Hinz, The development of the crust off
Dronning Maud Land, East Antarctica, in Weddell Sea Tectonics and
Gondwana Break-up, edited by B. C. Storey, E. C. King, and R. A.
Livermore, Geol. Soc. Spec. Publ.,108, 243 – 264, 1996.
Rogenhagen, J., and W. Jokat, The sedimentary structure in the western
Weddell Sea, Mar. Geol.,168, 45 – 60, 2000.
Royer, J. Y., P. Patriat, H. W. Bergh, and C. Scotese, Evolution of the
Southwest Indian Ridge from the Late Cretaceous (anomaly 34) to the
Middle Eocene (anomaly 20), Tectonophysics,155, 235 – 260, 1988.
Segoufin, J., Anomalies magnetique mesozoique dans le bassin de Mozam-
bique, C. R. Sceances Acad. Sci., Ser. 2,287D, 109 – 112, 1978.
Segoufin, J., and P. Patriat, Existence d’anomalies mesoziques dans le
bassin de Somali; Implications pour les relations Afrique-Antarctique-
Madagascar, C. R. Sceances Acad. Sci., Ser. 2,291B, 85 – 88, 1980.
Storey, B. C., The crustal blocks of West Antarctica within Gondwana:
Reconstruction and break-up model, in Geological Evolution of Antarc-
tica, edited by M. R. A. Thomson, J. A. Crame, and J. W. Thomson, pp.
587 – 592, Cambridge Univ. Press, New York, 1991.
Storey, B. C., The role of mantle plumes in continental break-up case
histories from Gondwanaland, Nature,377, 301 – 308, 1995.
Storey, B. C., Microplates and mantle plumes in Antarctica, Terra Antarct.,
3(2), 91 – 102, 1996.
Wessel, P., and W. H. F. Smith, New improved version of Generic Mapping
Tools released, Eos Trans. AGU,79, 579, 1998.
White, R. S., and D. P. Mackenzie, Magmatism at rift zones: The generation
of volcanic continental margins and flood basalts, J. Geophys. Res.,94,
7685 – 7729, 1989.

T. Boebel, W. Jokat, and M. Ko¨ nig, Alfred Wegener Institute for
Polar Research, Postfach 12061, Columbusstrasse, D-27568 Bremerhaven,
Germany. (jokat@awi-bremerhaven.de)
U. Meyer, Geoforschungszentrum, Telegrafenberg, D-14473 Potsdam,
Germany. (umeyer@gfz-potsdam.de)
JOKAT ET AL.: EARLY GONDWANA BREAKUP EPM 4 -15