Magmatic rifting of Pangaea linked to onset of South American plate motion

Erin K. Beutel

⁎

Department of Geology and Environmental Geosciences, College of Charleston, 66 George St., Charleston, SC 29424, United States

ABSTRACTARTICLE INFO

Article history:

Received 30 August 2007

Accepted 5 June 2008

Available online 21 June 2008

Keywords:

Pangaea

Stress

Finite element

Dikes

Magmatic rifting

Amagmatic rifting

The causes of the transition from amagmatic to magmatic rifting during continental break-up are not always

clear and have been often linked to the break-through of a plume. However, stress ﬁelds recorded in the crust

may offer new insights into the relationship between changes in the stress ﬁeld and the onset of magmatism.

In this paper stress ﬁelds recorded in the crust of North America are used to test possible causes of the break-

up of Pangaea and the transition from amagmatic to magmatic rifting. Finite element models reveal that the

most likely scenario for the break-up involves the initial northwest motion of North America at 230 Ma,

followed by a south-southeast motion of South America at 200 Ma with the initiation of magmatism, and

ﬁnally a weakened area between North and South America sometime soon after 200 Ma which resulted in

North America being dominated by northwest motion once again. It was also determined that plate

boundary structure and orientation play a large part in the recorded stress ﬁelds and must be taken into

consideration when modeling continental break-ups and rifting.

1. Background

Continental rifting and the transition from pure continental rifting

to seaﬂoor spreading may be investigated from both present day

activity, such as rifting in East Africa—the Red Sea, and from past

rifting events, such as the break-up of Pangaea. Present day studies,

such as those on the East African rift and Red Sea rift have the

advantage of real-time measurements and observations, while studies

of the past rifting events have the advantage of a completed timeline

rather than a snapshot waiting to develop. However, because of the

amount of time that has passed, investigations of past rifting events

sometimes smooth over the details such as rift orientation and

propagation because of a lack of data or because of scale issues. In this

paper I use these details to understand the stress ﬁeld in North

America at the time of the break-up of Pangaea, speciﬁcally the time

when the rifting went from amagmatic to magmatic. The possible

causes of these stress ﬁelds, and therefore the break-up of Pangaea,

are then determined using ﬁnite element models.

The break-up of Pangaea began around 230 Ma with the initial

rifting of the southeastern portion of North America (Schlische, 2002).

These early rifts can be found in both North and South Carolina

(Schlische, 20 02) a nd generally trend Northeast with regional

variations, however, it is unclear if this initial rifting also occurred

along the southern margin of North America as the few Mesozoic rift

basins from Georgia, Northern Florida, and Alabama are not precisely

dated. Around 200 Ma+/− 2Ma(Nomade et al., 2007) magmatic

injection in the form of a giant dike swarm began approximately

synchronously with the progression of rifting to the North-Northeast

(Hames et al., 2000; McHone, 2000; Salters et al., 2003; Schlische

et al., 2003; Beutel et al., 2005). Because most of the dikes and the

exposed ﬂows are dated to around 200 Ma, it is assumed that the

majority of the voluminous maﬁc magmas along the North American

margin from Florida to New England were emplaced within a 2–

3 million time period (Nomade et al., 2007). Initial maps combined

with new ﬁeld work indicate that the southeastern margin of North

America is dominated by NW trending dikes cross-cut by infrequent to

frequent N and NE trending dikes, however, as the swarm progresses

to the northeast the overall dike trend slowly rotates to become

predominantly NE (May, 1971; Ragland et al., 1983; McHone, 1988;

Marzoli et al., 1999; Salters et al., 2003; Beutel et al., 2005). Cross-

cutting relationships and geochemistry suggest that the NW trending

dikes were emplaced ﬁrst, followed by the N and ﬁnally the NE trending

dikes (Ragland et al.,1983; Beutel et al., 2005). Because of sedimentation

and logistical issues, it is not clear what the relationship between these

200 Ma dikes and the formation of the ﬁrstoceanic crust around 180 Ma.

However, it is clear that they represent the transition between the

amagmatic phase of rifting and magmaticrifting phase. This information

can be used in conjunction with a ﬁnite element model to determine the

stress ﬁeld during the break-up of Pangaea and the possible causes of

this stress ﬁeld as the rifting progresses from magmatic to amagmatic.

Because the regional stresses recorded in the rock represent snapshots

of what was occurring at the time, ﬁnding the sources of these stresses

should clarify possible causes for the break-up of the Pangaea by

indicating what was occurring prior to and during the break-up and

transition from amagmatic to magmatic.

Tectonophysics 468 (2009) 149-157

⁎ Tel.: +1 843 953 5591; fax: +1 843 953 5446.

E-mail address: [email protected].

doi:10.1016/j.tecto.2008.06.019

Contents lists available at ScienceDirect

Tectonophysics

journal homepage: www.elsevier.com/locate/tecto

1.1. Dikes and stress

Dikes record the stress ﬁeld during their emplacement and thus the

stress ﬁeld during the change from amagmatic to magmatic rifting

during the break-up o f Pangaea. In general dikes a re injected

perpendicular to the least least-compressive regional stress ﬁeld,

however, pre-existing fractures, local stress ﬁelds, and basement

structures can cause deviation from the regional (state-wide) stress

ﬁeld (de Boer and Snider, 1979; Mchone et al., 1987; Fialko and Rubin,

1999; Ziv et al., 2000; Jourdan et al., 2006). According to Ziv et al. (2000),

one of three criteria must be met for dikes to follow pre-existing

fractures; one, that the fractures are close to perpendicular to the least

least-compressive stress ﬁeld (which means that the dikes will still give

us a general view of the stress ﬁeld at the time of their emplacement);

two, that the magma pressure is so high that it creates a scenario where

the shear to the opening of the dike walls is small; and three, the

effective dike-normal stress is small compared to the rock tensile

strength. Because these conditions are difﬁcult to maintain over long

distances and/or depths, most dikes affected by fractures rotate within a

short distance to become parallel to the ambient compressive stress ﬁeld

(Ziv et al., 2000). Local stress ﬁelds that affect dikes can be caused by the

injection of the dike itself and/or the magma body, the classic example of

magma bodies affecting dike orientation is seen at volcanoes where

dikes are injected radially about the domal uplift (Johnson, 1 96 1). The

best evidence for basement control of dike orientation comes from the

Karoo swarm in South Africa where Jourdan et al. (2006) recognized that

the Karoo swarm consists of both Proterozoic and Mesozoic dikes, both

of which follow the cratonic margin. This suggests that cratonic margins

and basement structures may strongly affect dike orientation.

1.1.1. Can CAMP dikes be used to reveal a regional (eastern North America)

stress history or are they affected by local (10's of km) structures and stress

ﬁelds?

The dikes and other Mesozoic igneous features of the east coast of

North America have been co llectively grouped into the Central

Atlantic Magmatic Province (CAMP) and have been used as evidence

both for a plume and against a plume (May, 1971; de Boer and Snider,

1979; McHone et al., 1987; Ernst and Buchan, 1997; King and

Anderson, 1998; Marzoli et al., 1999; Courtillot et al., 1999; Dalziel

et al., 2000; McHone, 2000; Janney and Castillo, 2001; Beutel et al.,

2005). In this study I will use the dike patterns to understand the

least-compressive stress regime during the transition from amagmatic

to magmatic in the break-up of Pangaea, however, this will only be

possible if none of the previously discussed outside inﬂuences have

has affected their orientation. Because the conditions under which

dikes follow pre-existing fractures cannot be maintained over long

distances unless the fractures are extensional stress perpendicular to

begin with, it is unlikely that the giant dike swarms of CAMP are overly

inﬂuenced by any non-stress responsive fractures. While the inﬂuence

of a large magma chamber due to a plume has been proposed to

explain the CAMP dikes, the dikes are not radial and in fact cross-cut.

Further, unless the plume was a moving target, the regional stress ﬁeld

would have exerted some inﬂuence on the dikes at a distance and no

evidence of a change in orientation along the strike of the dikes is

seen, suggesting that any magma chamber inﬂuence was extremely

localized (10s of km) and not visible. The other possibility, that pre-

existing basement structures could have inﬂuenced dike orientation,

is also unlikely given the multiple orientation of dikes within the

Carolinas, and because many of the dikes cross-cut or stop against the

Fig. 1. Sketch map of 200 my Pangaea with pre-South Carolina regional studies dike orientations (e.g. Marzoli et al., 1999), continental boundaries, and normal faults. Previously

mapped dikes are shown as short thick black lines (e.g. Marzoli et al., 1999), rough sketches of major normal faults in North America are shown as thin lines. Rose diagram in corner

shows true orientation of dikes in South Carolina and parts of North Carolina based on new ﬁeldwork and map compilations, while the adjacent grey box illustrates how the all the

mapped dikes and faults would overlie each other in the region indicated by the connecting lines. Light gray shading of 3 states along the southeastern portion of North America

indicates from North to South, the locations of North Carolina, South Carolina, and Georgia. The stippled area between the major continental outlines indicates both the continental

shelves, which we do not have good data for and missing continental pieces from the time of the suture—it is assumed that all three continents were acting as one at this time.

150 E.K. Beutel / Tectonophysics 468 (2009) 149-157

pre-existing structures (Ragland et al., 1983; Beutel et al., 2005). Thus,

it appears that the CAMP dikes were most likely inﬂuenced by regional

least-compressive stress ﬁelds rather than either pre-existing stress

ﬁelds or a magma body.

1.1.2. What do the dikes tell us?

Previous interpretations of the dike orientations have focused on

an explanation for the NW trending dikes that dominate South

Carolina and parts of Georgia and North Carolina because based on

sketch maps of the region they are both at a high angle to the expected

regional least-compressive stress ﬁeld during rifting and to the pre-

existing crustal structures (e.g. Schlische et al., 2003). Further, when

presented as a group on a sketch map, the CAMP dikes are often

mistaken as being regionally separate swarms of NW, N, and NE

trending dikes (Fig. 1). However, detailed mapping in South and North

Carolina suggests that the swarms overlap and cross-cut indicating a

non-radial source and a change in local conditions during emplace-

ment (Fig. 1). While only general maps of dike locations are available

for Georgia and Alabama, detailed ﬁeldwork and consolidation of

USGS quadrangle information indicates that a series of 200 Ma NW,

N, and NE trending igneous dikes are found throughout the Carolinas

(Beutel et al., 2005). Cross-cutting relationships indicate that the

oldest dikes are NW trending and the NE trending dikes are the

youngest with the N trending dikes coming in between. Though

all dike trends can be found in the Carolinas, the NW trending

dikes dominate in the south, while the NE trending dikes are more

dominant from Virginia north and N trending dikes are more domi-

nant in the middle. This indicates that the least-compressive stress

ﬁeld during emplacement of the dikes was rotating from NE ex-

tensional, to E extensional, to NW extensional in the southeastern

United States as rifting progressed northward (Fig. 2).

Therefore based on ﬁeld evidence from dikes and rif ts I

reconstructed the following stress changes along the east coast of

North America from ~230 Ma to 190 Ma and conclude based on these

changes that a plume as the cause of them is unlikely. Between 230 Ma

and ~200 Ma the southeastern edge of North America was slowly

rifting and undergoing a NW extensional stress while the rest of the

continental edge appears to be quiescent (though a lack of data from

the southern margin does not preclude early rifting here) (Fig. 2).

Around 200 Ma east–west rifts started to develop along the southern

margin of North America and northeast trending rifts began appearing

along the present day northeastern United States, this indicates that

the extensional stress ﬁeld was N–S along the southern margin of

North America, NW–SE along the southeastern margin of North

America, and NNW–SSE along the eastern margin of North America

(Fig. 2). However, based on the dike orientations, there existed at the

same time a NE–SW extensional stress ﬁeld in the southeast that

rapidly (~2 Ma) rotated to NW–SE. Dikes along the eastern margin of

North America north of Virginia are parallel to the rifts and therefore

apparently emplaced by the same least-compressive stress ﬁeld that

generated the rifts, indicating that unlike in the south the extensional

stress that created the rifts was likely responsible for the dike

orientation as well. This stress story, as summarized below, is used to

determine the viability of the ﬁnite element models via stress

ﬁeld

orientation comparison.

2. Model

A quadrilateral plane strain ﬁnite element grid using Gobalt and

Atkinson's (1996) FElt model was constructed to determine the

possible causes of the rotating stress ﬁelds during the transition from

amagmatic to magmatic rifting of Pangaea. These static elastic models

are essentially solutions to Hooke's law, which describes the

deformation of elastic solids. If we view the models as simple grids

of elastic springs and nodes, we provide the construction of the grids,

the properties of the springs, and the initial forces. The program solves

for the propagation of these forces across the grid, taking into account

the neighboring springs. Overall, the grid is isotropic and composed of

plane stress quadrilateral elements, except at the plate boundary zones

between the continents and the South Georgia Rift, which feature

elements specially aligned and shaped to represent these boundaries.

The ﬁnite element code requires thickness for the z-dire ction,

Poisson's ratio, Young's Modulus, and densities for all materials in

the program. Strength of the continents is given based on the elastic

moduli and an input thickness. The following parameters were input;

continental and rift mechanical properties and locations, stress/force

properties and locations, and motion constraints. Essentially the

model started with the outline of the continental shelves input into

the ﬁnite strain grid with a one order of magnitude weaker elastic

moduli and 3 times thinner than the surrounding continents (original

continental material properties: Young's Modulus 40 GPa, Poisson's

ratio .2, density of 2650 kg/m

). As an elastic model the results show

only the instantaneous stress and do not take into account viscous

ﬂow.

Fig. 2. From top to bottom the left column shows the estimated stress directions per

time period based on the faults and dikes of that age. The right column is a series of

cartoons illustrating the general orientation of the resultant extensional stress features

(dikes and faults) created at that time, these are underlain by the dikes and faults

created at the previous time and shown in grey. Double headed arrows indicate

tensional stress directions, shorter arrows indicate estimated smaller magnitude

stresses based on the prevalence of faulting or diking in that area, in the right-hand

column the thick bars indicate dikes while the lines with arrows indicate normal faults.

Stress Field Summary Time 1: NW trending extensional stress concentrated along the

southeast coast of North America. Time 2: N trending extensional stress along the south

margin of North America, NE trending stress along the southeast coast of North America

decreasing and/or rotating to NW extensional as you move northeast along the present

day coast of North America. Time 3: NW extensional stress along the eastern margin of

North America, concentrated along the present day northeast coast of the United States

portion of North America.

151E.K. Beutel / Tectonophysics 468 (2009) 149-157

Applied forces were calculated based on a general rate of motion

and mass for continents, because I am examining relative stress

orientation and strength rather than absolute numbers within an

order of magnitude will give the same results. Constraints on motion

were simpliﬁed into able to move in all directions or ﬁxed in space and

unable to move at all. Several models were constructed to test

numerous possibilities including a radial slab suction/rollback, a

vertical plume with radial compressional stress, and different applied

stresses to the continents. The radial slab pull model is constructed

based on the reconstructions of Pangaea that show it completely

surrounded by inward dipping slabs (Golonka, 2007). Approximate

slab-suction forces were based on the average rates of continental

motion, which was then translated into an outward force applied to

the margin. Vertical loading and topographic doming with radial

compression were combined to determine the approximate magni-

tude of the radial compressive force that might have been caused by

the intrusion of a plume. The other possible force tested was the

motion of the continents as driven by a mantle basal drag force. This

was also approximated based on the approximate rate and direction of

motion of the continents. The assumption being that the continents

are more likely to be driven by mantle ﬂow against a continental keel

and not asthenospheric ﬂow along the bottom of an oceanic plate, the

relative importance of this force is highlighted in Conrad and Lithgow-

Bertelloni (2006).

Model tests within four orders of magnitude were conducted for all

applied forces and the relative strength and orientation of the

resulting stresses did not change. Therefore, the forces chosen for

the models were based on the approximate relative strength of the

forces in general and on the approximate magnitude of those stresses

as listed in Fowler (1997).

Constraints on the motion of continents was limited to the end-

members of either free to move in three-dimensions or ﬁxed in space.

When constraints on motion were applied they were applied to all of

the nodes in the center of a continent assuming that the mechanism

holding the continent in place would be a lack of forces moving it and

a resistance to motion along the base of the cratonic portions of the

continent as a whole.

Given the scale of the elements in the model the only pre-existing

crustal elements portrayed are the plate boundary zones (or zones

where the continents meet) and the South Georgia Rift, which was

Fig. 3. A–J: Each ﬁgure A–J is divided into three boxes. The ﬁrst box to the far left is the result of the ﬁnite element model and shows the maximum stress intensity as background

colors, blue is compressional and red is extensional, and the maximum and minimum stress vectors, black is extensional and white is compressional. The north arrow indicates

present day north while the orientation of the continents reﬂects their approximate position at 230 Ma. The second box shows possible rift and/or dike orientations given the stress

orientations shown in the ﬁrst box. Dikes and faults are only indicated for the margins of the North American continent despite stresses being transmitted throughout the continent

because ﬁeld work suggests that the dikes stop at the Brevard fault, a major Appalachian plate boundary zone that runs parallel to the east coast. The fault is not in the model, but any

stresses behind its approximate location on the model should be viewed with suspicion. And the third box shows a cartoon of model parameters, plate boundary lines are shown

(note that a plate boundary line is shown between Florida and the red of North America, this actually represents the South Georgia rift and a possible plate boundary attachment of

Florida, models run without it are mildly different), as are the applied stresses (shown as arrows), and the constraints on movement (ﬁxed areas are shown with black dots). Finally,

areas of extreme weakness (areas of crust 3-orders of magnitude weaker than the surrounding crust) are shown as shaded grey.

152 E.K. Beutel / Tectonophysics 468 (2009) 149-157

believed to be active during this time. For this reason the only inter-

pretations of stress results that will be analyzed are those on the

margins as other, non-modeled crustal zones would affect stress

distribution in the interior.

The locations and shapes of the plate boundaries or margins of the

continents are based on reconstructions of Pangaea and current

continental crust outlines (e.g. Torsvik and Van der Voo, 2002). The

addition of a boundary between Florida and the rest of North America

is based on multiple lines of evidence that there was a zone of

thinning there. Several researchers have pointed out that Florida does

not appear to be North American in origin and it is of ten not even

depicted as attached until during/after the break-up of Pangaea and it

is separated from North America by the Brunswick Magnetic anomaly

(e.g. Heatherington and Mueller, 1999; Heatherington et al., 1999;

Ford and Golonka, 2003). The Brunswick Magnetic anomaly also

marks the South Georgia Rift system, which appears to cut parallel to

the probable suture between the Suwanee Terrane (Florida) and the

rest of North America (McBride, 1991).

3. Results (Table 1)

Fig. 3 shows several example model results, the postulated dike

ﬁelds, and the applied forces and Table 1 is a verbal summary of Fig. 3.

Postulated dike and fault orientations are only modeled for the South

and Eastern margins of North America despite the transmissions of

stresses into the continent because the model does not take into

account faults and plate boundaries within the continent known to

have halted the progression fractures. For example, the CAMP dikes

along the southeastern margin of North America all cease (except one)

at the Brevard fault zone in the Appalachians. Because these barriers

and crustal features were not taken into account the stresses from the

interior of the continent are less likely to be accurate.

In the ﬁrst set of three models there are no ﬁxed continents, the

plate boundary between the continents is modeled as one order of

magnitude lower strength than the surrounding crust, and the only

force is a force applied to the North American craton. In the ﬁrst

column the results of the models are displayed as a background

maximum stress intensity with the maximum and minimum stresses

plotted on top. Fig. 3A shows the results of a NW stress on the North

American continent, which results in a relatively uniform magnitude

stress ﬁeld with the correct distribution of stresses to have produced

NW trending dikes in the far southern United States and northeast

trending stresses along the east coast. By changing the applied stress

to more northerly the assumed dike orientation (perpendicular to the

least least-compressive stress) becomes predominantly NW trending

along the east coast of North America, once again the stress magnitude

is relatively uniform. Finally, in Fig. 3C the actual areas of weakness are

changed to become much sharper with deﬁned strong transform

Fig. 3 (continued ).

153E.K. Beutel / Tectonophysics 468 (2009) 149-157

Fig. 3 (continued ).

154 E.K. Beutel / Tectonophysics 468 (2009) 149-157

zones between them and the stress applied to North America is once

again northeast and an increase in NW trending dikes along the

southeastern North American continent is observed.

In Fig. 3D–F a northeastern trending force is applied to North

America, Africa is ﬁxed and the stress on South America and the

strength and the width of the plate boundary area is changed. In Fig. 3D

Africa is ﬁxed, North America is pushed to the northeast, and the plate

boundary zones remain linear and one order of magnitude lower

strength than the surrounding crust. This results in a concentration of

stress along the east coast of North America, including northwest

trending tensional stress along the southeastern coast, which would

create northeast trending dikes or rifts, and decreasing and more varied

tensional stress along the northeastern coast of North America. In

Fig. 3E the applied force to North America remains northeast and an

applied south-southeasterly force is applied to South America, the plate

boundary remains the same. This results in an almost radial stress

pattern around the three continents with northeastern trending

tensional stresses along the southeast coast of North America changing

to northwestern trending tensional stresses northward along the coast.

It is inferred that the results would be dominantly northwest trending

dikes or rifts in the southeastern United States and northeast trending

dikes or rifts in the northeastern United States. Finally, in Fig. 3F the

applied forces to North and South America remain the same, but an

area of signiﬁcantly weakened crust is placed along the northern edge

of South America. This results in the east coast of North America being

subject to a northwesterly tensional stress that would result in north-

east trending dikes or rifts mostly concentrated along the northeast

margin of the present day United States.

In Fig. 3G and H the effect of weak plate boundary zones is

examined for the scenario of a ﬁxed Africa, a northeast trending North

America, and a south-southeast trending South America. While there

is some small-scale variation along the rifts, strengthening (but not

removing the changes in the grid shape) the South Georgia Rift or the

other plate boundaries, has little effect on the overall stress ﬁeld.

In Fig. 3I the force applied to South America is changed from south-

southeast to southeast, this dramatically changes the stress ﬁeld in

Table 1

Refers to Fig. 3 Results: Listed as probable dike

and normal fault trends along

the south and eastern coasts of

North America

Model parameters: Unless

otherwise speciﬁed suture

zones between continents are

modeled as one order of

magnitude weaker and thinner

than surrounding continental

crust. The South Georgia Rift is

modeled in the same manner.

A N to NW trending along the

Southern margin of North

America, abrupt change in trend

around Georgia/South Carolina

border to NE trending.

NW trending motion for North

America—no motion or

resistance to motion for Africa

or South America.

B NW trending along South and

Southeastern margin of North

America.

NNW trending motion of North

America—no motion or

resistance to motion for Africa

or South America

C N to NW trending along the

Southern and Southeastern

margin of North America,

change in trend around Georgia/

South Carolina border to NE

trending, trend change is

marked by an area of highly

variable stress direction

coinciding with the area of most

deﬁned rift basins.

NW trending motion for North

America—no motion or

resistance to motion for Africa or

South America. Suture zones

along east coast made to

resemble rift basins with north–

south trending rift basins of one

order of magnitude weaker and

thinner crust separated by strong

(same as the rest of the

continent) areas of continent

with structures at right angles to

the rift zones.

D Generally NE trending along the

Southeastern margin of North

America with possible

N trending along the Southern

margin.

NW trending motion for North

America—no motion or

resistance to motion for South

America. Africa is held ﬁxed'

‘ﬁxed’ in place so it cannot

move in response to North

America's motion.

E NE trending along the Southern

margin of North America

changing to NW trending along

the Southeastern margin and

ﬁnally changing to NE trending

along the Northeastern margin

of North America.

NW trending motion for North

America and a SSW trending

motion for South America.

Africa is held ‘ﬁxed’ in place so

it cannot move in response to

North America's motion.

F Generally NE trending along the

entire Eastern margin of North

America.

NW trending motion for North

America and a SSW trending

motion for South America.

Africa is held ‘ﬁxed’ﬁxed’ in

place so it cannot move in

response to North America's

motion. The area between

South and North America is

modeled as 3 orders of

magnitude lower strength than

the continental crust.

G NE trending along the Southern

margin of North America

abruptly switching to NW

trending along the Eastern

margin of North America.

Applied continental motions are

thesameAsinFig. 3F, North

America is moving to the NW,

South America is moving to the

SSW , and Africa is held ﬁxed. The

suture between South America

and North America is removed as

a zone of weakness to see the

effect this suture had on the

distribution of stress along the

coast.

H NE trending along the Southern

margin of North America,

switching gradual to NW

trending along the Eastern

margin of North America.

As in

Fig. 3F, North America is

moving to the NW, South

America is moving to the SSW,

and Africa is held ﬁxed. The

suture between South America

and North America is only

modeled as weak around Florida

and through the South Georgia

Rift.

(continued on next page)

Table 1 (continued)

Refers to Fig. 3 Results: Listed as probable dike

and normal fault trends along

the south and eastern coasts of

North America

Model parameters: Unless

otherwise speciﬁed suture

zones between continents are

modeled as one order of

magnitude weaker and thinner

than surrounding continental

crust. The South Georgia Rift is

modeled in the same manner.

I NE trending along the Southern

and Southeastern margins of

North America, switching to NW

trending in the mid-section of

the Eastern margin.

As in Fig. 3F, North America is

moving to the NW, South

America is moving to the SSW,

and Africa is held ﬁxed. In this

model the suture between

North and South America is

simpliﬁed and the South

Georgia Rift is removed from

the zones of weakness.

J N trending dikes throughout the

North American Eastern margin.

All of the continents are subject

to an outward motion (radial

about the triple junction)

applied to their outer

boundaries due to a likely slab-

suction force.

K Small stresses in the continents

result in small N trending dikes

along the Southeastern margin

and NW trending dikes along the

Southern Margin.

A negative vertical loading is

applied to an area around the

triple junction and a radial

outward force modeled on

expected topographic

expression is emplaced around

the vertical load.

155E.K. Beutel / Tectonophysics 468 (2009) 149-157

North America including northwest trending stresses in the south-

eastern United States and northeast trending tensional stresses in the

northeastern United States.

In Fig. 3J and K two other possible scenarios are examined. In Fig. 3J

the results of a slab-rollback force applied to the edge of the model

such that a radial tensional stress was applied to the margins of the

supercontinent is examined. This results in a non-radial stress pattern

with east–west stresses dominating along the coast of North America.

In Fig. 3K the results of a plume model are examined. A large area of

radial stress around the triple junction between North America, South

America, and Africa, was applied similar to what might be expected

from a plume dome. Interestingly, though this creates extremely large

tensional stress at the triple junction, there is very little stress

transmitted through the continents, though some east–west tensional

stresses were located along the east coast of North America. This

would result in north–south trending rifts or dikes.

4. Analysis

Given the results of the models, it appears that the most likely

scenario for the creation of initial northeast trending rifts along the

southeastern United States followed by northwest trending dikes in

the southeast with northeast trending dikes and rifts in the north-

eastern United States is the following: Africa is ﬁxed (or resistant to

movement); North America begins to move to the northwest around

230 Ma creating northeast trending rifts in the southeastern United

States (Fig. 3D); Around 200 Ma South America begins to move to the

south-southwest, this creates enough of a stretch in the crust to cause

decompression melting which leads to the injection of northwest

trending dikes in the southeastern United States and northeastern

trending rifts and dikes along the northeastern coast of the United

States (Fig. 3 E); Finally, as South America pulls away from North

America an area of weakened crust decouples South America's

inﬂuence on North America's stress ﬁeld and results in a concentration

of northwestern trending tensional stress along the east coast of North

America, this results in the injection of northeast trending dikes along

the east coast, but predominantly in the northeast (Fig. 3F). The

generation of the melt volumes generally associated with volcanic

margins via stretching of the continental crust has been shown to be

viable and is inﬂuenced by lithospheric architecture, mantle tem-

perature, and the rate of spreading (Van Wijk et al., 2001). Given these

parameters it appears that continental thinning could produce the

melt volumes seen in the CAMP province. Further, because the CAMP

dike swarms consist of the three geochemically distinct sources, with

probable depth constraints on their origin (e.g. McHone, 2000), it is

more likely that decompression melting at various depths and of a

heterogeneous mantle would result in the observed geochemical

variabilities observed in the North American Central Atlantic

Magmatic Province volcanics associated with the break-up of Pangaea.

This model explains the apparently rapidly changing stress ﬁeld in

southeastern North America (from NW tensional to NE tensional and

back to NW tensional) and why the northeastern coast of North

America has a different stress history (dominantly NW tensional).

These ﬁnite element model results suggest that the initial north-

eastern motion of North America resulted in the opening of rifts along

the southeast coast and not the northeast coast because of the

orientation of the weakened plate boundary, which acted as a stress

guide. Further the models suggest that the initiation of the movement

of South America to the south-southeast not only acted as a catalyst

for magmatism, but also rotated only the southern stresses in North

America to northeast tensional. This suggests that the stress ﬁeld

generated by South America's motion was diminished with distance

from the source and by the trend of the weakened plate boundary

zones. Finally, it appears that continued magmatism and stretching in

the Gulf of Mexico would have essentially detached South America

from North America. This would have resulted in the stress ﬁeld in

North America changing back to what it was before South America

began to move (hence beginning with NE trending normal faulting

and ending with NE trending dikes with NW trending dikes in be-

tween while South America was exerting some inﬂuence). Because the

exerted stress ﬁeld along the southern portion of North America de-

creased dramatically with the complete separation of South America

from North America, it appears that tensional stress in the crust is

most ampliﬁed when there is resistance to the tension.

4.1. Other scenarios

It is also clear from the models that while plate boundary orien-

tation may be important in affecting where stress ﬁelds change, the

plate boundary zone does not have to be weak and areas of ﬁxed crust

or applied force that mimic the plate boundary zone shape have the

same effect (Fig. 3G and H). The effect of the shape of the plate

boundary zone is also seen in Fig. 3C where transforms were inserted

between the rifts, this causes localized (directly adjacent to the margin

—probably 10s10's of km) changes in the stress ﬁeld, but does not

signiﬁcantly deviate the regional (eastern North America) stress ﬁeld.

Two non-craton driven scenarios were also examined, that of a radial

slab-retreat tensional force applied to the supercontinent and a radial

stress pattern caused by a plume. While the radial slab retreat did not

create the stress ﬁelds preserved in the rocks, it is also clear that slab

retreat can affect the stress ﬁeld within the continent and it likely was

similar in orientation to the modeled craton driven forces in models

3D–F during the break-up of Pangaea. The plume scenario has been

given much credence as a cause of the break-up of Pangaea and the

radial' ‘radial’ dike swarm in the southeastern United States (May,

1971; de Boer and Snider, 1979; Wilson, 1997; Ernst and Buchan, 1997;

Courtillot et al., 1999; Marzoli et al., 1999; McHone, 2000; Dalziel et al.,

2000; Janney and Castillo, 2001; Storey et al., 2001), however, model

results suggest that the stress generated by a plume would not be

transmitted very far in a continent and that the orientation of that

stress is incompatible with the observed dikes and rifts. Certainly the

presence of zones of weakness between the continents may have

affected the ability of stresses to propagate into continent. However,

the orientation of the stresses around a domal plume feature would

not result in the multiple stress ﬁelds indicated by the overlapping

dike swarms in Southeastern North America and there must have

been some zone of thinning or weakness prior to rifting or doming for

the continents to break as they did.

5. Conclusions

The break-up of Pangaea and the transition from amagmatic rifting

to magmatic rifting was likely driven by the craton driven motion of

North and South America. Finite element models indicate that the

complicated stress regime recorded along North Americas east coast is

best recreated by assuming a structural boundary between the con-

tinents and an initial northwestern motion of North America followed

by a south-southwest motion of South America and the initiation of

magmatic rifting. The ﬁnal changes in the stress ﬁeld appear to have

been caused by the separation of North America and South America by

an area of weakness. The onset of magmatic rifting is clearly asso-

ciated with a change in the stress ﬁeld in the southeastern United

States (going from Northeast rifting to Northwest trending dikes) and

therefore the initiation of the movement of South America. This sug-

gests that magmatism is a result of lithospheric thinning rather than

a particular upwelling, as an upwelling would have initiated motion

in North America and South America at the same time, whereas a

combination of slab retreat and mantle ﬂow might allow for the

30 million year delay in onset of South America's movement. It is also

clear that magmatism and weakened crust strongly affect the regional

continental stress ﬁeld as shown by the period of weakened crust

north of South America and its effective decoupling of the South

156 E.K. Beutel / Tectonophysics 468 (2009) 149-157

American continent. Thus it is apparent that the complicated stress

ﬁelds preserved in amagmatic and magmatic features can illustrate

the importance of continental motion, boundary shape, and coupling

when considering the break-up of Pangaea and the transition from

amagmatic to magmatic rifting. The results also suggest that a plume

is an unlikely source of the CAMP magmatic province and the break-

up of Pangaea.

Acknowledgements

Sincere thanks to 3 anonymousreviewers for their helpful comments

on content and presentation, model clarity and ﬁgure clarity were

greatly improved by their comments. Special thanks to my chair, Mitch

Colgan for supporting my research efforts via meeting support and

fostering a supportive environment within the department.

References

Beutel, E.K., Nomade, S., Fronabarger, A.K., Renne, P.R., 2005. Pangaea's complex

breakup: a new rapidly changing stress ﬁeld model. Earth and Planet. Sci. Lett. 236,

471–485.

de Boer, J.Z., Snider, F.G., 1979. Magnetic and chemical variations of Mesozoic diabase

dikes from eastern North America: evidence for a hotspot in the Carolinas? Geol.

Soc. Am. Bull. 90, 185–198.

Conrad, C.P., Lithgow-Bertelloni, C., 2006. Inﬂuence of continental roots and astheno-

sphere on plate–mantle coupling. Geophys. Res. Lett. 33 L05312L.

Courtillot, V., Jaupart, C., Manighetti, I., Tapponier, P., Besse, J., 1999. On causal links

between ﬂood basalts and continental break-up. Earth Planet. Sci. Lett. 166,

177–195.

Dalziel, I.W.D., Lawver, L.A., Murphy, J.B., 200 0. Plumes, orogenesis, and super-

continental fragmentation: Earth Planet. Sci. Lett. 178, 1–11.

Ernst, R.E., Buchan, K.L., 1997. Giant radiating dyke swarms; their use in identifying pre-

Mesozoic large igneous provinces and mantle plume. In: Mahoney, J.J., Cofﬁn, M.F.

(Eds.), Large Igneous Provinces; Continental, Oceanic, and Planetary Flood Volcanism.

Geophys. Monogr., 100, pp. 297–333.

Fialko, Y.A., Rubin, A., 1999. Thermal and mechanical aspects of magma emplacement in

giant dike swarms. J. Geophys. Res. 104, 23,033–23,049.

Ford, D., Golonka, J., 2003. Phanerozoic paleogeography, paleoenvi ronment and

lithofacies maps of the circum-Atlantic margins. Mar. Pet. Geol. 20, 249–285.

Fowler, C.M.R., 1997. The Solid Earth. Cambridge University Press, United States of

America, p. 472.

Gobalt, J.L., Atkinson, D.C., 1996. FElt Users Guide and Reference Manual. San Diego, San

Diego Press, University of San Diego. 234 p.

Golonka, J., 2007. Late Triassic and Early Jurassic palaeogeography of the world.

Palaeogeogr. Palaeoclimatol. Palaeoecol. 244 (1–4), 297–307.

Hames, W.E., Renne, P.R., Ruppel, C., 2000. New evidence for geologically instantaneous

emplacement of earliest Jurassic Central Atlantic magmatic province basalts on the

North American margin. Geology 28, 859–862.

Heatherington, A.L., Mueller, P.A., 1999. Lithospheric sources of North Florida, USA

tholeiites and implications for the origin of the Suwannee terrane. Lithos 46,

215–233.

Heatherington, A.L., Mueller, P.A., Nutman, A.P., 1999. A Jurassic granite from Southern

Georgia, U. S. A. Silicic, extension- related magmatism along the Southeastern

Coastal Plain. J. Geol. 107, 375–384.

Janney, P.E., Castillo, P.R., 2001. Geochemistry of the oldest Atlantic oceanic crust

suggests mantle plume involvement in the early history of the central Atlantic

Ocean. Earth Planet. Sci. Lett. 192, 291–302.

Johnson, R.B., 1961. Patterns and origin of radial dike swarms associated with West

Spanish peak and Dike Mountain, South Central Colorado. Geol. Soc. Am. Bull. 72,

579–590.

Jourdan, F., Feraud, G., Bertrand, H., Watkeys, M.K., Kampunzu, A.B., Le Gall, B., 2006.

Basement control on dyke distribution in Large Igneous Provinces: case study of the

Karoo triple junction. Earth Planet. Sci. Lett. 241, 307–322.

King, S.D., Anderson, D.L., 1998. Edge-driven convection. Earth Planet. Sci. Lett. 160,

289–296.

Marzoli, A., Renne, P.R., Piccirillo, E.M., Ernesto, M., Bellieni, G., De Min, A., 1999.

Extensive 200 million-year-old continental ﬂood basalts of the Central Atlantic

Magmatic Province. Science 284, 616–618.

May, P.R., 1971. Pattern of Triassic–Jurassic diabase dikes around the North Atlantic in

the context of the predrift positions of the continents. Geol. Soc. Am. Bull. 82,

1285–1292.

McBride, J.H., 1991. Constraints on the structure and tectonic development of the early

Mesozoic South Georgia Rift, Southeastern United States; seismic reﬂection data

processing and interpretation. Tectonics 10 (5), 1065–1083.

McHone, J.G., 1988. In: Manspeizer, W. (Ed.), Tectonic and paleostress patterns of

Mesozoic intrusions in eastern North America: in, Triassic–Jurassic Rifting :

continental breakup and the origins of the Atlantic Ocean and passive margins.

Elsevier, New York, pp. 607–620.

McHone, J.G., 2000. Non-plume magmatism and tectonics during the opening of the

central Atlantic Ocean. Tectonophysics 316, 287–296.

McHone, J.G., Ross, M.E., Greenough, J.D., 1987. Mesozoic dyke swarms of Eastern North

America. In: Halls, H.C., Farhig, W.F. (Eds.), Maﬁc Dyke Swarms. Geol. Ass. Can. Spec.

Pap., 34, pp. 279–288.

Nomade, S., Knight, K.B., Beutel, E., Renne, P.R., Verati, C., Feraud, G., Marzoli, A., Youbi,

N., Bertrand, H., 2007. Chronology of the Central Atlantic Magmatic Province;

implications for the central Atlantic rifting processes and the Triassic–Jurassic biotic

crisis. Palaeogeogr. Palaeoclimatol. Palaeoecol. 244 (1–4), 326–344.

Ragland, P.C., 1991. Mesozoic igneous rocks. In: Horton, J.W., Zullo, V.A. (Eds.),

The Geology of the Carolinas, pp. 171–190. Carolina Geological Society Fiftieth

Anniversary Volume.

Ragland, P.C., Hatcher Jr., R.D., Whittington, D., 1983. Juxtaposed Mesozoic diabase dike

sets from the Carolinas. Geology 11, 394–399.

Salters, V.J.M., Ragland, P.C., Hames, W.E., Milla, K., Ruppel, C., 2003. Temporal chemical

variations within lowermost Jurassic tholeiitic magmas of the Central Atlantic

Magmatic Province. In: Hames, W.E., McHone, J.G., Renne, P.R., Ruppel, C. (Eds.), The

Central Atlantic Magmatic Province: Insights from Fragments of Pangaea. Geophys.

Monogr., 136, pp. 163–177.

Schlische, R.W., 2002. Progress in understanding the structural geology, basin evolution,

and tectonic history of the eastern North American rift syste. In: LeTourneau, P.M.,

Olsen, P.E. (Eds.), The Great Rift Valleys of Pangaea in Eastern North America, vol. 1.

Columbia University Press, New York.

Schlische, R.W., Withjack, M.O., Olson, P.E., 2003. Relative timing of CAMP, rifting,

continental breakup, and basin inversion; tectonic signiﬁcance. In: Hames, W.E.,

McHone, J.G., Renne, P.R., Ruppel, C. (Eds.), The Central Atlantic Magmatic Province;

Insights from Fragments of Pangaea. Geophys. Monogr., 136, pp. 33–59.

Storey, B.C., Leat, P.T., Ferris, J.K., 20 01. The location of mantle–plume centers during the

initial stages of Gondwana breakup. In: Ernst, R.E., Buchan, K.L. (Eds.), Mantle Plumes:

Their Identiﬁcation Through Time. Geol. Soc. Am., Special Paper, 352, pp. 71–80.

Torsvik, T.H., Van der Voo, R., 2002. Reﬁning Gondwana and Pangaea palaeogeography:

estimates of Phanerozoic non-dipole (octupole) ﬁelds. Geophys. J. Int. 151, 771–794.

Van Wijk, J.W., Huismans, R.S., ter Voorde, M., Cloetingh, S.A.P.L., 2001. Melt generation

at volcanic rifted margins: no need for a mantle plume. Geophys. Res. Lett. 28 (20),

3995–3998.

Wilson, M., 1997. Thermal evolution of the Central Atlantic passive margins: continental

break-up above a Mesozoic super-plume. J. Geol. Soc. (London) 154, 491–495.

Ziv, A., Rubin, A.M., Agnon, A., 2000. Stability of dike intrusion along preexisting

fractures. J. Geophys. Res. 105 (B3), 5947–5961.

157E.K. Beutel / Tectonophysics 468 (2009) 149-157