Long term stability in deep mantle structure: Evidence from the ~300 Ma Skagerrak-Centered Large Igneous Province (the SCLIP)

Abstract

On the basis of large areal extent (~ 0.5 × 106 km2), volume, brevity of eruption interval (± 4 My) and convergent dyke swarms, the flare-up of igneous activity at 297 Ma in NW Europe marks a typical Large Igneous Province (LIP): The Skagerrak-Centered LIP (SCLIP). LIPs are widely but not universally considered products of deep-seated mantle plumes: We test the idea that a Skagerrak plume rose from the core–mantle-boundary (CMB) by restoring the center of SCLIP eruption, using a new reference frame, to its ~ 300 Ma position in a Pangea A type reconstruction. That position (~ 11°N, 16°E, south of Lake Chad in Central Africa) lies vertically above the edge of the African Large Low Shear Velocity Province (LLSVP). It has previously been shown that eruption locations vertically above the edge of one or other of the Earth's two LLSVPs at the CMB characterize nearly all the LIPs erupted since 200 Ma. A deep-sourced SCLIP plume source implies that the edge of the African LLSVP at the CMB has not moved significantly with respect to the spin axis of the Earth during the past 300 My. This is a 30% longer duration for the stability of a deep mantle structure than has been previously demonstrated and suggests that the African LLSVP was at least established by early Permian (Pangea) times.

Full text

Page 1

Long term stability in deep mantle structure: Evidence from the ~300 Ma
Skagerrak-Centered Large Igneous Province (the SCLIP)
Trond H. Torsvika,⁎, Mark A. Smethursta, Kevin Burkeb,c, Bernhard Steinbergera
aCentre for Geodynamics, NGU, Leiv Eirikssons vei 39, N-7491 Trondheim, Norway
bDepartment of Geosciences, University of Houston, 312 S&R1 Houston Tx 77204-5007, USA
cSchool of Geosciences, University of Witwatersrand, WITS 2050, South Africa
Received 14 August 2007; received in revised form 2 December 2007; accepted 2 December 2007
Available online 15 December 2007
Editor: R.D. van der Hilst
Abstract
On the basis of large areal extent (~0.5×106km2), volume, brevity of eruption interval (±4 My) and convergent dyke swarms, the flare-up of
igneous activity at 297 Ma in NW Europe marks a typical Large Igneous Province (LIP): The Skagerrak-Centered LIP (SCLIP). LIPs are widely
but not universally considered products of deep-seated mantle plumes: We test the idea that a Skagerrak plume rose from the core–mantle-
boundary (CMB) by restoring the center of SCLIP eruption, using a new reference frame, to its ~300 Ma position in a Pangea A type
reconstruction. That position (~11°N, 16°E, south of Lake Chad in Central Africa) lies vertically above the edge of the African Large Low Shear
Velocity Province (LLSVP). It has previously been shown that eruption locations vertically above the edge of one or other of the Earth's two
LLSVPs at the CMB characterize nearly all the LIPs erupted since 200 Ma. A deep-sourced SCLIP plume source implies that the edge of the
African LLSVP at the CMB has not moved significantly with respect to the spin axis of the Earth during the past 300 My. This is a 30% longer
duration for the stability of a deep mantle structure than has been previously demonstrated and suggests that the African LLSVP was at least
established by early Permian (Pangea) times.
© 2007 Elsevier B.V. All rights reserved.
Keywords: Large Igneous Provinces; plumes; plate reconstructions; Pangea; large low shear velocity provinces; core mantle boundary; deep mantle stability
1. The Skagerrak-Centered Large Igneous Province
Assembly of Pangea, mainly during late Carboniferous
times (Torsvik and Cocks, 2004), was accompanied by intra-
continental rift initiation sporadically distributed over about
60% of the supercontinent's area (Burke, 1978). Rifts of this
population are well known in NW Europe (Wilson et al.,
2004), where at a time close to the Permo–Carboniferous
boundary (~300 Ma), as igneous activity in the rifts waned, a
flare-up of magmatism occurred. Dyke trends in Scania
(Sweden), the Oslo region (SE Norway) and Scotland indicate
it to be centred on the Skagerrak (Ernst and Buchan, 1997)
(Fig. 1). We estimate this event to have left a record over an
area of at least 0.5×106km2, twice the presently mapped area
of volcanics (228.000 km2), sills (14,000 km2) and dykes
(3353 km in total length) in Fig. 1. That record includes
rocks in the Oslo graben, the offshore Skagerrak graben, SW
Sweden (Västergötland and Scania), Northern Germany, Scot-
land, Northern England and the Central North Sea (Heeremans
etal.,2004a). Collectively we estimate that these rocks erupted
at 297±4 Ma (Figs. 1–2; Section 2) and represent a Large
Igneous Province (LIP) for which we suggest the name
“Skagerrak-Centered LIP” (SCLIP). Although a globally
agreed-upon definition of a LIP does not exist, Coffin and
Eldholm's(1994)originaldefinitionstressedtheimportanceof
areal extent (N0.1×106km2) of pre-dominantly mafic igneous
rocks. In the most recent definition, a LIP is defined as “mainly
Available online at www.sciencedirect.com
Earth and Planetary Science Letters 267 (2008) 444–452
www.elsevier.com/locate/epsl
⁎Corresponding author. University of Oslo, 0316Oslo Norway, and Schoolof
Geosciences, University of Witwatersrand, WITS 2050 South Africa.
E-mail address: trond.torsvik@ngu.no (T.H. Torsvik).
0012-821X/$ - see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsl.2007.12.004

Page 2

mafic magmatic province with areal extent N0.1×106km2,
igneous volumes N0.1×106km3, and maximum life-spans of
~50 Myrs that are emplaced in an intra-plate setting and
characterized by igneous pulses of short duration (1–5 Myrs),
duringwhichalargeproportionofthetotaligneousvolumehas
been emplaced” (after Bryan and Ernst, 2008); the SCLIP
fulfils these criteria.
2. Was the SCLIP generated by a mantle plume?
It has been shown that at eruption all the LIPs of the past
200 My (except for the Columbia River Basalt LIP) lay over the
margins of one or other of the Earth's Large Low Shear Velocity
Provinces (LLSVPs, Garnero et al., 2007) at the core–mantle-
boundary (CMB) and are for that reason deep mantle plume-
derived (Burke and Torsvik, 2004; Davaille et al., 2005; Torsvik
et al., 2006), but other criteria have been widely used to
distinguish deep plume-derived provinces (see Courtillot et al.,
1999; Ernst et al., 2005). These criteria include a link to an age-
progressive volcanic track, link to uplift, geochemical signa-
tures (e.g. high3He/4He values) and the existence of giant dyke
swarms. The origin for igneous rocks of the Oslo Graben and
related areas of NW Europe is debated (Wilson et al., 2004).
The SCLIP clearly lacks an age-progressive volcanic track but
other provinces of presumed deep mantle origin (including
Karroo) do not have tracks, and in general tracks are rare or
poorly developed within continents. The reason for expecting a
volcanic track in a plume model are age data from the Oslo
Graben and related dykes and sills (mostly Rb/Sr) that suggest
long-lived and almost continuous magmatic activity from about
300 to 240 Ma (Fig. 2a). Because palaeomagnetic data imply a
change from tropical latitudes (ca. 12°N) in the Early Permian to
subtropical latitudes (29°N) by the Early Triassic (Fig. 2b, based
on a global compilation in Torsvik et al., in press), a deep plume
should have led to an age-progressive track of more than
2000 km in length, i.e. from South France to Oslo. However,
‘modern’ age data from NW Europe mostly yield SCLIP ages
between 290 and 300 Ma (Figs. 1 and 2b; Table 1), averaging to
297.5±3.8 Ma (simple mean age; 1σ error) or 297.4±0.2 Ma
(weighted mean age). From the Oslo graben, U/Pb ages (298±
3 Ma simple mean age) dating the initial and main rift phases
(Stages 2 and 3; Ramberg and Larsen, 1978; Olaussen et al.
1994) cluster between 302 and 298 Ma (Figs. 1 and 2b),
and not over an interval of 25–30 My as indicated by Rb/Sr
ages (Fig. 2a) that clearly underestimate the true age. Younger
magmatism occurred in the Oslo rift, such as Late Permian (?)
Fig. 1. MapofNWEuropeshowingthedistributionofc.290–300Mavolcanics(228.000km2),sills(14,000km2)anddykes(3353kmintotallength).Thesemagmatic
products with an estimated volume around 0.5×106km3form the SCLIP. Compiled ages (see also age histogram in Fig. 2) are all ‘modern’40Ar–39Ar ages (mostly
hornblende, biotite and feldspar ages), thermal ionization mass spectrometry U/Pb zircon and perovskite (OG) and baddelyeite (WS) ages, and high resolution ion
microprobe206Pb/238U zircon ages (NEGB). Ages (Table 1) are shown with 2σ errors. A simple mean age of SCLIP is 297.5±3.8 Ma (1σ, N=33 ages) or alternatively
297.4±0.2 Ma if ages are weighted by their individual errors. Ages in green are from sills/minor intrusions (Monaghan and Pringle, 2004; Van der Voo and Torsvik,
2004;Hamilton,personalcommunication2006),redfromdykes(VanderVooandTorsvik,2004)andblackfromvolcanics(BreitkreuzandKennedy,1999;Heeremans
et al., 2004a; Corfu and Dahlgren, 2006) and associated larvikites and lardalites⁎ (Dahlgren et al., 1996). Regions/features discussed in the text include: OG, Oslo
graben, SG, Skagerrak graben, VG, Västergötland; SC, Scania; NEGB, North-East German graben; CG, Central graben (North Sea); WMVI, West Midland Valley
Intrusions; WS, Whin Sill; IS, Iapetus Suture (mid Silurian); TS, Thor Suture (late Ordovician). Volcanic, sill and dyke outlines follow (Heeremans et al., 2004b).
445T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 3

granitic batholiths and Early Triassic dykes in the northernmost
areas (Torsvik et al., 1998), but we do not consider these related
to a plume and the SCLIP.
Evidence for normal to only slightly elevated mantle
temperature (Neumann 1994) has from time to time been used
as argument against a plume origin. LIP generating plumes have
risen from the CMB but near surface temperatures need not be
elevated because partial melting of mantle rock to generate
basalt can dissipate the heat associated with a newly risen
plume. Evidence of an elevated plume temperature in rock
composition is not to be expected and it is the large volumes
(~0.5 mill. km3; estimated from Breitkreuz and Kennedy, 1999;
Neumann et al., 2004) of erupted rock that indicate the short-
term proximity of a hot source. Because of (i) huge volume and
areal extent of erupted rock, (ii) short interval of eruption and
(iii) convergent dyke swarms we here test the idea (Section 4)
that the SCLIP was generated by a mantle plume, that we call
the Skagerrak plume.
3. Lithospheric controls on the Skagerrak plume
eruption site
The causal relation between plumes and rifting has been
much discussed (e.g. Courtillot et al., 1999). Some plumes of
presumed deep-seated origin, including the Central Atlantic
Magmatic Province (CAMP), Karroo, Deccan and Tristan
(Fig. 3), erupted into existing rifts (Burke et al., 2003). This can
be attributed to “upside-down drainage” by which a plume
impinging on the base of the lithosphere generates basaltic
magma that travels up-slope to a thin part of the lithosphere
(Sleep, 1997, 2006), such as an existing rift, before eruption.
The Skagerrak plume eruption site provides another example of
the eruption of a plume into a rift, although in this case the rift
has not been shown to be significantly older than the time of
eruption of igneous rocks of the SCLIP. From the evidence of
convergent dykes (Fig. 1) the estimated site where the centre of
the Skagerrak plume impinged on the lithosphere lies within a
~150 km diameter circular area around 57.5°N and 9.0°E in the
Skagerrak graben (Ernst and Buchan, 1997). The graben is
located, as many intra-continental rifts are, close to an older
suture zone, in this case the Late Ordovician Thor suture (Cocks
and Torsvik, 2005) (Fig. 1). The oldest rocks in the mapped
individual rifts of the Skagerrak graben and their extension into
the Kattegat are occurrences of Lower Rotliegende (~297 Ma)
volcanic rocks and a sill in well penetrations (Heeremans et al.,
2004a). We suggest that those are rocks related to the Skagerrak
plume and that the establishment of the rift structure may have
been as late as the time of the eruption.
Fig. 2. (a) Frequency distributions of published (1975–1998) isotopic age data from the Oslo rift and presumed related dykes and sills (Torsvik and Eide, 1998). The
majority of these isotope ages (82 out of 87) are Rb/Sr ages that often underestimate the true age (see text) (b) ‘Modern’ age data from NW Europe (Fig. 1; Table 1),
including U/Pb ages from the Oslo Graben. Collectively these SCLIP ages average to 297.4±3.9 Ma. 5 My time windows in histograms. The predicted latitude curve
for a location near Oslo (60°N, 10°E) is calculated from a global palaeomagnetic apparent polar wander path (Torsvik et al., in press) and shows that Oslo (Baltica/
Europe) drifted from low northerly tropical latitudes (Late Carboniferous) and into subtropical latitudes (Late Permian) during Late Paleozoic–Early Mesozoic times.
446T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 4

Not only are intra-continental rifts commonly developed, as
the Skagerrak graben is, in proximity to older suture zones but
intra-continental rifts are also sometimes reactivated in later
tectonic episodes. The Oslo rift has been suggested to overlie
andtorecordreactivationofaLatePrecambrian (Vendian)intra-
continental rift. The presence of such a rift has been inferred
from the proximity to the Oslo graben of the 583±15 Ma Fen
alkaline igneous rock and carbonatite complex (Burke, 1978;
Meertetal.,1998)andthecontemporaryAlnöcarbonatite(584±
7 Ma) in Sweden (Meert et al., 2007). Both are considered
related to a Vendian rifting event that immediately preceded the
formation of a rifted continental margin to Baltica in latest
Precambrian times (Cocks and Torsvik, 2005).
Plume-related basalt may erupt into a rift at a site that oc-
cupies a relatively small area, possibly no more than 200 km in
diameter, but magma from such small sources has been shown
capable of propagating rapidly, presumably within the litho-
sphere, for more than 1000 km from original plume eruption
sites (Marzoli et al., 1999). Propagating magma forms dykes,
sills and lavas that erupt into existing and commonly actively
extending rifts. CAMP plume magma erupted at ~200 Ma into
an active Triassic rift of the east coast of North America and
propagated within rifts as far as Canada and Brazil (Marzoli
et al., 1999). SCLIP magma appears to have propagated to
Scotland(~1000km)andNorthernEnglandformingdykes,sills
and lavas that we assign to the SCLIP on the basis of age, dyke
trends (Fig. 1) and volume. Those rocks were erupted into the
Midland Valley rift that had been active through much of
Carboniferous time (Monaghan and Pringle, 2004). SCLIP
magma also propagated to the north and south of the Midland
Valley in a 500 km wide dyke swarm and into Northern England
in the Great Whin sill of County Durham (Fig. 1). To the east of
the plume eruption site rocks of the SCLIP have been penetrated
in wells in rifts beneath the Kattegat but SCLIP volcanic rock is
muchmoreabundantinthesub-surfaceoftheNorthGermanand
North Sea foreland basins (Heeremans et al., 2004b).
4. Relationship of the Skagerrak LIP eruption site to the
deep mantle
Today two large low shear wave velocity (δVs) regions of the
deep mantle, the LLSVPs, are recognized as the most prominent
features of all global shear-wave tomographic models. Those
well-defined African (Fig. 3) and Pacific LLSVPs are isolated
within the faster parts of the deep mantle. Using existing
tomographic results Torsvik et al. (2006) found steep horizontal
δVsgradients to be concentrated along the margins of the two
LLSVPs at the CMB close to the 1% slow δVscontour, which
forms faster–slower-boundaries (FSB) within the deep mantle.
LIPs at the Earth's surface when reconstructed over the past
200 My to their original eruption sites become concentrated
radially above the margins of the African (Fig. 3) and Pacific
LLSVPsclosetotheCMB.Thiscorrelationhasbeenestablished
by using paleomagnetic as well as fixed and moving hotspot
reference frames. The long-term persistence of the LLSVPs is
consistent with independent evidence that they are composi-
tionally distinct and are not just simply hotter than the material
making up the rest of the deep mantle (Torsvik et al., 2006 and
references therein). The finding that the majority of LIP sources
have been generated from sites on the present day LLSVP
margins at the CMB requires that within resolution those mar-
gins have remained in their present positions with respect to the
spin axis for at least 200 My. That led us to consider whether
older LIPs and large magmatic events not classified as LIPs in
the literature would yield similar results. An obvious LIP to
consider was the ~251 Ma Siberian Trap LIP but when that LIP
is reconstructed paleomagnetically to its eruption site, it is clear
that it was not associated with either the African or the Pacific
LLSVP,butatitseruptiontimerosefromaFSBlinkedtoasmall
isolated low-velocity region at the CMB that borders a prom-
inent slab graveyard volume (Fig. 3).
The establishment of an ‘absolute’ reference frame (Torsvik
etal.,inpress)forthemotionofobjectsattheEarth'ssurfaceisa
required critical input to mantle convection simulations oriented
Table 1
‘Modern’ age data from NW Europe (Figs. 1 and 2b)
Locality, countryRock type Age±2σ MethodMineralRef
Larvik, Norway
Skien, Norway
Larvik, Norway
Skien, Norway
Brunlanes, Norway
Brunlanes, Norway
Arendal, Norway
Central Graben,
North Sea
Mt. Billingen, Sweden
Mt. Billingen, Sweden
Scania, Sweden
Scania, Sweden
Peterhead, Scotland
Cathcart, Scotland
Lennoxtown, Scotland
Carskeoch, Scotland
Androssan, Scotland
Craigen-Avisyard,
Scotland
Whin Sill, England
Whin Sill, England
Petersberg, Germany
Petersberg, Germany
Wetting, Germany
Friedland core,
Germany
Mirow core, Germany
Kotzen core, Germany
Penkun core, Germany Lava
Lobenjun, Germany
Friedland core,
Germany
Mirow core, Germany
Landsberg, Germany
Holzmuhlenthal,
Germany
Schwerz, Germany
Lardalite
Ignimbrite 298±2
Larvikite
Lava
Lava
Lava
Dyke
Lava
292±1 U/Pb
U/Pb
U/Pb
U/Pb
U/Pb
U/Pb
40Ar/39Ar
40Ar/39Ar
Zircon
Titanite
Zircon
Perovskite
Perovskite
Perovskite
Whole rock 3
Whole rock 7
1
2
1
2
2
2
299±1
299±1
300±1
302±1
296±10
299±3
Sill
Sill
Dyke
Dyke
Dyke
Sill
Sill
Intrusion
Sill
Intrusion
298±2
300±2
287±2
300±10
297±12
292±1
292±3
295±1
298±1
302±1
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
40Ar/39Ar
Plagioclase
Whole rock 4,3
Plagioclase
Plagioclase
Plagioclase
Biotite
Biotite
Hornblende 5
Hornblende 5
Biotite
4,3
4,3
4,3
3
5
5
5
Sill
Sill
Lava
Lava
Lava
Lava
294±2
297
292±6
294±6
297±6
297±8
40Ar/39Ar
U/Pb
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
Plagioclase
Baddelyeite 6
4,3
8
8
8
8
Lava
Lava
297±6
297±6
298±6
298±6
299±12
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
8
8
8
8
8
Lava
Lava
Lava
Lava
Lava
300±6
301±6
302±6
206Pb/238U Zircon
206Pb/238U Zircon
206Pb/238U Zircon
8
8
8
Lava 307±6
206Pb/238U Zircon8
Ref = Reference; (1) Dahlgren et al. (1996), (2) Corfu and Dahlgren (2006),
(3)VanderVooandTorsvik(2004),(4)Timmerman,pers.comm.,(5)Monaghan
and Pringle (2004), (6) Hamilton, pers. comm., (7) Heeremans et al. (2004a),
(8) Breitkreuz and Kennedy (1999).
447 T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 5

towards an improved understanding of the deep mantle. In
making a palaeomagnetic reference frame for movement back to
297MasothatwecoulddeterminewhethertheSkagerrakplume
overlay the margin of an LLSVP at the CMB we used zero
longitudinal average motion of Africa as the best possible
approximation(BurkeandTorsvik,2004;Torsviketal.,inpress)
in a Pangea A type configuration (Van der Voo, 1993), i.e. NW
Africa located adjacent totheAtlanticmarginof NorthAmerica.
The SCLIP eruption site is reconstructed to ~11°N and 16°E
(south of Lake Chad, Central Africa). The restored SCLIP
eruption site indeed projects radially downward onto the margin
oftheAfricanLLSVPattheCMB,andnearthe1%slowcontour
in the SMEAN (Becker and Boschi, 2002) tomography model at
depth 2800 km (Fig. 4). This result is comparable to findings for
practically all LIPs since 200 Ma and for most major hotspots
(Steinberger, 2000; Courtillot et al. 2003; Montelli et al., 2004,
2006) argued to have a deep plume origin (Fig. 3). One may
argue that these findings are sensitive to the choice of tomog-
raphy model: Torsvik et al. (2006) examined and compared
several global shear-wave tomographic models at the CMB.
In Figs. 3–4, the 297 Ma SCLIP reconstruction is draped on
SMEANbutforcomparisonandasasensitivitytestwehavealso
reconstructed SCLIP on two pure Dq tomography models. Both
models also show that the SCLIP eruption site projects onto the
LLSVPmarginattheCMB,neartheca.1%slowcontourofKuo
et al. (2000) and the ca. 0.8% slow contour in the Castle et al.
(2000) model (Fig. 5). Globally, these contours correspond
approximately to the 1% slow contour of SMEAN, because the
area withb−1% in the SMEAN and Kuo et al. (2000) models is
approximately equal to the area with b−0.8% in the Castle et al.
(2000) model. The choice of tomography model is thus not
critical to conclusions linking SCLIP to long-term heterogene-
ities in the deep mantle.
Possible sources of error to which we knew our reconstruc-
tion was subject included (i) longitudinal uncertainty (assump-
tion of zero longitudinal average motion of Africa), (ii) relative
plate circuits (including our choice of a Pangea A reconstruc-
tion), (iii) plume advection, (iv) non-dipole field contributions,
and (v) true polar wander (TPW). The overall direction of
postulated TPW for the last 130 Ma (Besse and Courtillot, 2002;
Torsvik et al., in press) is roughly from the present day location
of NE Canada/Greenland (60°W) towards the North Pole. This
Fig. 3. Reconstructed LIP eruption sites in the Indo-Atlantic realm (annotated circles; extracted from Torsvik et al., 2006) and hotspots (crosses; Steinberger, 2000;
Montelli et al., 2006) draped on the SMEAN (Becker and Boschi, 2002) shear wave velocity anomaly (δVs) model for 2800 km depth. The −1% contour (1% slow) is
shownin grey. Horizontal gradients (red-to-blue contours)in shear wave velocity anomaly(%δVs/°) are indicatedwhere theyexceed 0.16%/°.Hotspotsargued to have
originated from deep plumes (IL, Iceland; AZ, Azores; CY, Canary; TT, Tristian; RU, Reunion; AF, Afar; AS, Ascension; CR, Crozet; CV, Cap Verde; KE, Kerguelen)
are shown as enlarged red crosses (Courtillot et al., 2003; Montelli et al., 2004, 2006). Our reconstruction of the Skagerrak-Centered LIP (SCLIP, 11°N, 16°E; open
star) shows that it was located near the 1% slow contour and a high gradient, in the vicinity of several active hotspots (Cameroon, Darfur and Tibesti) and some
1500 km from the present Afar hotspot. LIP abbreviations (mean eruption ages in Ma) are: AF, Afar Flood Basalt (31); GI, Greenland/Iceland (54); DT, Deccan Traps
(65); SL, Sierra Leone Rise (73); MM, Madagascar/Marion (87); BR, Broken Ridge (95); CK, Central Kerguelen (100); SK, South Kerguelen (114); RT, Rajhmahal
Traps (118); MR, Maud Rise (125); PE, Parana-Etendeka (132); BU, Bunbury Basalts (132); KR, Karroo Basalts (182); CP, CAMP (200); ST, Siberian Traps (251);
SCLIP, Skagerrak-Centered LIP (297).
448 T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 6

approximately corresponds, in a reference frame in which the
Earth's spin axis is fixed, to a rotation of the African LLSVP
around an equatorial location close to its center of mass, very
close to the reconstructed SCLIP location (11°N, 16°E). Pro-
vided that TPW was in approximately the same or opposite
direction before—and this is in fact expected if LLSVPs are
associated with highs of the degree-2 non-hydrostatic geoid—
the SCLIP will stay close to the LLSVP boundary regardless of
TPW, and our result is therefore not sensitive to whether or not
there has been TPW.
The reconstructed SCLIP eruption site is based on a global
palaeomagnetic apparent polar wander (APW) path. To test the
latitudesensitivitywealsocomputedthelatitudefortheestimated
SCLIP centre (57.5°N, 9°E) using only 285–305 Ma European
palaeomagnetic poles (Van der Voo and Torsvik, 2004). A Euro-
pean only mean pole (pole latitude/longitude=41.9°N/167.7°E;
95percentconfidencecircle,A95=1.9°;numberofpoles,N=35)
yieldalatitudeof11.0±4.8°andthusidenticaltoourglobalAPW
path connected through relative plate circuits. Repeating this
exercise with only palaeomagnetic data from the Oslo rift and
related rocks yield a latitude of 6.9±4.5° (N=6 poles; 285–
305 Ma age range; pole=37.7°N/167.3°E, A95=5.1°), a some-
what lower latitude than our global or purely European mean
models but within errors of both.
Fig. 4. Expansion/zoom of the box region in Fig. 3 in which we show the distribution of c. 290–300 Ma volcanics, sills and dykes (as Fig. 1). These magmaticproducts
(the SCLIP), along with present coastlines for Scandinavia (Baltica), the British Isles, NW Europe (Avalonia), Greenland (Laurentia) and pre-existing sutures between
Laurentia–Baltica (Iapetus) and Baltica–Avalonia (Tornquist/Thor) are all reconstructed to 297±10 Ma using a global running mean (20 Ma window) palaeomagnetic
reference frame (Torsvik et al., in press), constructed in South African co-ordinates and restoring all locations to South Africa using appropriate plate circuits in a
Pangea A-type configuration. The resulting ‘absolute’ Euler pole for European elements is latitude=22.3°N, longitude=78.4°E, angle=54.5°, and for Greenland
latitude=36°N, longitude=76.1°E, angle=68.8° relative to the spin-axis; at this time Europe was moving NNW with a speed of ca. 2.7 cm/yr. According to our
reconstruction SCLIP is estimated to have erupted radially above the 1% slow contour at the CMB.
449T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 7

Fig. 5. Similar to Fig. 4 but simplified reconstructions draped on Dq tomography models of Kuo et al. (2000) and Castle et al. (2000). In these models our
reconstruction of SCLIP is also estimated to have erupted radially above the African LLSVP margin, and near the ca. 1% (Kuo et al., 2000) and ca. 0.8% (Castle et al.,
2000) slow contour at the CMB.
Fig. 6. Early Permian (ca. 297 Ma) Pangea A type reconstruction draped on SMEAN assuming that present-day deep mantle structures, the antipodal Africa (centred
beneath Pangea) and Pacific LLSVPs, have existed at least since the Early Permian and have sourced most LIPs erupted at Earths surface. The Gondwana superterrane
formed at ca. 550 Ma and included most of South America, Africa, Madagascar, India, Arabia, East Antarctica and Australia. Laurussia formed by the Caledonian
(Silurian) collision of Laurentia, Baltica, Avalonia and intervening terranes. Subsequently collided with Gondwana to form Pangea at the end of the Carboniferous.
Siberia, North China and South China were not part of Pangea in the early Permian (Torsvik and Cocks, 2004).
450 T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 8

Most agree that before the onset of Pangea breakup, the
Jurassic Pangea A reconstruction is correct with NW Africa
adjacent to NE America (Fig. 6). For Permian times, however,
palaeomagnetic poles do not fully agree witha Pangea Afit (e.g.
Van der Voo and Torsvik, 2001, 2004) due to substantial con-
tinental overlap of Gondwana and Laurussia in the equatorial
realm. One possible option to avoid this continental overlap (not
favored geologically by the authors) is to modify the Pangea A
reconstruction by moving Gondwana eastward (alternatively
Laurussia westward) in order to avoid the overlap and later
juxtapose them by major lateral shear (as much as 8000 km in
some models; i.e. Pangea C) before Jurassic Pangea break-up. If
so, this would seriously affect our interpretations, violating the
‘zero-Africa’ approach, and thus the Pangea A reconstruction in
Fig. 6 would be invalid. For many reasons (Van der Voo and
Torsvik, 2001, 2004; Torsvik and Cocks, 2004) we consider
PangeaAastheonlyviablereconstructionfortheEarly Permian
(Fig. 6).
Other solutions to the Pangea enigma is to hypothesize
sedimentary inclination shallowing or alternatively octupole
(G3=g03/g01) field contributions (Van der Voo and Torsvik,
2001; Torsvik and Van der Voo, 2002), and in both cases the
Pangea misfit can be explained by a too-far-south position for
Laurussia (including the SCLIP) and a too-far northerly position
of Gondwana. At around SCLIP eruption time, Torsvik and Van
der Voo (2002) hypothesize 10% octupole contributions but
since G3 is latitude dependent and most palaeomagnetic
sampling sites for Laurussia were at low latitudes, the overall
effect is very small, and Laurussia is estimated to be only 1–2°
too far south at ca. 300 Ma (see Torsvik and Van der Voo, 2002;
their Fig.9).G3contributions are thus notcritical toconclusions
linking SCLIP to heterogeneities in the deep mantle.
5. Conclusions
(1) The SCLIP 297 Ma eruption event in NW Europe covered
vast areas, estimated to at least 0.5 million km2and thus
larger than some other magmatic events characterized as
LIPs (e.g. Columbia River and Maud Rise).
(2) Eruption occurred within a relatively short time-span,
estimated to ±4 My with existing isotope data but future
high-precision U/Pb data may show that the time-span
was even shorter.
(3) Based on these characteristics, along with the occurrence
of convergent dykes in the Skagerrak graben we conclude
that the SCLIP can be characterized as a LIP, and magma
propagated more than 1000 km (Scotland) away from its
centre (ca. 57.5°N, 9°E)
(4) The restored SCLIP eruption site projects radially down-
ward onto the CMB at the margin of the African LLSVP,
and we arguethe SCLIP eruption event was triggered by a
deep plume from the CMB.
(5) The SCLIP and practically all other LIPs for the last
300 Ma have erupted radially above the African or Pacific
LLSVP margin and CMB heterogeneities must therefore
haveremainedthe same, atleast since shortlyafter Pangea
formation.
Acknowledgements
We thank Michel Heeremans for providing digital outlines of
volcanics, sills and dykes, Martin Timmerman for detailed
information on unpublished40Ar–39Ar ages (quoted in Van der
Voo and Torsvik, 2004), Mike Hamilton for information on the
297MaU/PbbaddelyeiteagefromtheWhinSill(UK),andBjørn
T. Larsen, Mike Gurnis, Rob Van der Voo, Vincent Courtillot and
Conall Mac Niocaill for stimulating comments and suggestions.
NFR, NGU and Statoil ASA are thanked for financial support
(PETROMAKS GPlates: A Novel Exploration Tool Frontier
Science and Exploration no. 176531).
References
Becker, T.W., Boschi, L., 2002. A comparison of tomographic and geodynamic
mantle models. Geochem. Geophys. Geosys. 3. doi:2001GC000168.
Besse, J., Courtillot, V., 2002. Apparent and true polar wander and the geometry
of the geomagnetic field over the last 200 Myr. Journal of Geophysical
Research-Solid Earth 107. doi:1029/2000JB000050.
Burke, K., 1978. Evolution of continental rift systems in the light of plate
tectonics. In: Ramberg, I.B., Neumann, E.R. (Eds.), Tectonics and
Geophysics of Continental Rifts in Tectonics and Geophysics of Continental
Rifts. Reidel Publishing Company Dordrecht, pp. 1–9.
Burke, K., Torsvik, T.H., 2004. Derivation of Large Igneous Provinces of the
past 200 million years from long-term heterogeneities in the deep mantle.
Earth Planet. Sci. Lett. 227, 531–538.
Burke, K., MacGregor, D.S., Cameron, N.R., 2003. Africa's petroleum
systems: four tectonic “aces” in the past 600 million years. In: Arthur, T.J.,
Macgregor, D.S., Cameron, N.R. (Eds.), Petroleum Geology of Africa;
New Themes and Developing Technologies. Geol. Soc. Lond. Spec.
Public., vol. 207, pp. 21–60.
Breitkreuz, C., Kennedy, A., 1999. Magmatic flare-up at the Carboniferous–
Permian boundary in the NE German basin revealed by SHRIMP zircon
ages. Tectonophysics 302, 307–326.
Bryan, S., Ernst, R., 2008. Revised definition of Large Igneous Province (LIP).
Earth Science Reviews 86, 175–202.
Castle, J.C., Creager, K.C., Winchester, J.P., van der Hilst, R.D., 2000. Shear
wavespeedsat the baseof the mantle.J. Geophys.Res.105,21.543–21.558.
Cocks, L.R.M., Torsvik, T.H., 2005. Baltica from the late Precambrian to mid
Palaeozoic: the gain and loss of a terranes's identity. Earth Sci. Rev. 72,
39–66.
Coffin, M.F., Eldholm, O., 1994. Large Igneous Provinces: crustal structure,
dimenstions, and external consequences. Reviews of Geophysics 32, 1–36.
Corfu, F., Dahlgren, S., 2006. Perovskite U–Pb age and Pb–Pb isotopic
characteristics of the earliest, Carboniferous melilitite–melilite nephelinite–
basanite volcanic series, Oslo rift. Association of Canada, Annual Meeting,
Montréal, May 14–17, Abstract 31, pp. 30.
Courtillot, V., Jaupart, C., Manighetti, I., Taponier, P., Besse, J., 1999. On causal
links between flood basalts and continental breakup. Earth Planet. Sci. Lett.
166, 177–195.
Courtillot, V., Davaille, A., Besse, J., Stock, J., 2003. Three distinct types of
hotspots in the Earth's mantle. Earth Planet. Sci. Lett. 205, 295–308.
Dahlgren, S., Corfu, F., Heaman, L.-M., 1996. U–Pb isotopic time constraints,
and Hf and Pb source characteristics of the Larvik plutonic complex, Oslo
Paleorift. Geodynamical and geochemical implications for the rift evolution.
V.M. Goldschmidt Conference, Ruprecht-Karls Universität, Heidelberg,
Germany, p. 120.
Davaille,A.,Stutzmann,E.,Silveira,G.,Besse,J.,Courtillot,V.,2005.Convective
patterns under the Indo-Atlantic. Earth Planet. Sci. Lett. 239, 233–252.
Ernst, R.E., Buchan, K.L., 1997. Giant radiating dyke swarms: their use in
identifying pre-Mesozoic large igneous provinces and mantle plumes. In:
Mahoney, J., Coffin, M. (Eds.), Large Igneous Provinces: Continental,
Oceanic, and Planetary Volcanism. Geophysical Monograph Series Amer-
ican Geophysical Union Washington DC, vol. 100, pp. 297–333.
451T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452

Page 9

Ernst, R., Buchan, K.L., Campbell, I.H., 2005. Frontiers in Large Igneous
Province research. Lithos 79, 271–297.
Garnero, E.J., Lay, T., McNamara, A.K., 2007. Implications of lower mantle
structural heterogeneity for existence and nature of whole mantle plumes. In:
Foulger, G.R., Jurdy, D.M. (Eds.), Plates, Plumes, and Planetary Processes.
GSA Special Paper, 430, pp. 79–102.
Heeremans, M., Timmerman, M.J., Kirstein, L., Faleide, J.I., 2004a. New
constraints on the timing of late Carboniferous–early Permian volcanism
in the central North Sea. In: Wilson, M., Neumann, E.R., Davies, G.R.,
Timmerman, M.J., Heeremans, M., Larsen, B.T. (Eds.), Permo–Carbonifer-
ous Magmatism and Rifting in Europe. Geological Society London Special
Publication, vol. 223, pp. 177–194.
Heeremans, M., Faleide, J.I., Larsen, B.T., 2004b. Late Carboniferous–Permian
of NW Europe: an introduction to a new regional map. In: Wilson, M.,
Neumann, E.R., Davies, G.R., Timmerman, M.J., Heeremans, M., Larsen,
B.T. (Eds.), Permo–Carboniferous Magmatism and Rifting in Europe.
Geological Society London Special Publication, vol. 223, pp. 75–88.
Kuo, B.-Y., Garnero, E.J., Lay, T., 2000. Tomographic inversion of S-SKS times
for shear velocity heterogeneity in D'': Degree 12 and hybrid models.
J. Geophys. Res. 105, 28.139–28.157.
Marzoli, A., Renne, P.R., Piccirilo, 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.
Meert, J.G., Torsvik, T.H., Eide, E.A., Dahlgren, S., 1998. Tectonic significance
of the Fen Province, S. Norway: constraints from geochronology and
paleomagnetism. J. Geology 106, 553–564.
Meert, J.G., Walderhaug, H.J., Torsvik, T.H., Hendriks, B.W.H., 2007. Age and
paleomagnetic signature of the Alnø Carbonatite Complex (NE Sweden):
additional controversy for the Neoproterozoic Paleoposition of Baltica.
Precambrian Res. 154, 159–174.
Monaghan, A.A., Pringle, M.S., 2004. 40Ar/39Ar geochronology of Carboni-
ferous–Permian volcanism Midland Valley, Scotland. In: Wilson, M.,
Neumann, E.R., Davies, G.R., Timmerman, M.J., Heeremans, M., Larsen,
B.T. (Eds.), Permo–Carboniferous Magmatism and Rifting in Europe.
Geological Society London Special Publication, vol. 223, pp. 219–241.
Montelli, R., Nolet, G., Dahlen, F., Masters, G., Engdahl, E., Hung, S., 2004.
Finite-frequency tomography reveals a variety of mantle plumes. Science 303,
338–343.
Montelli, R., Nolet, G., Dahlen, F., Masters, G., 2006. A catalogue of deep
mantle plumes: new results from finite-frequency tomography. Geochem.
Geophys. Geosyst. 7. doi:10.1029/2006GC001248.
Neumann, E.R., 1994. The Oslo rift P–T relations and lithospheric structure.
Tectonophysics 240, 159–172.
Neumann, E.-R., Wilson, M., Heeremans, M., Spencer, E.A., Obst, K.,
Timmermann, M.J., Kirstein, L., 2004. Carboniferous–Permian rifting and
magmatismin southernScandinavia,the North Sea andnorthernGermany:a
review. In: Wilson, M., Neumann, E.R., Davies, G.R., Timmerman, M.J.,
Heeremans, M., Larsen, B.T. (Eds.), Permo–Carboniferous Magmatism and
Rifting in Europe. Geological Society London Special Publication, vol. 223,
pp. 11–40.
Olaussen, S., Larsen, B.T., Steel, R., 1994. The Upper Carboniferous–Permian
Oslo Rift; basin fill in relation to tectonic development. In: Embry, A.F.,
Beauchamp, B., Glass, D.J. (Eds.), Pangea; global environments and
resources, Memoir-Canadian Society of Petroleum Geologists, 17. Canadian
Society of Petroleum Geologists, Calgary, AB, Canada, pp. 175–197.
Ramberg, I.B., Larsen, B.T., 1978. Tectonomagmatic evolution. In: Dons, J.A.,
Larsen, B.T. (Eds.), The Oslo Paleorift: A Review and Guide to Excursions.
Bulletin Norges Geologiske Undersøkelse, vol. 337, pp. 55–73.
Sleep, N.H., 1997. Lateral flow and ponding of starting plume material.
J. Geophys. Res. 102, 10001–10012.
Sleep,N.H.,2006.Mantleplumesfromtoptobottom.EarthSci.Rev.77,231–271.
Steinberger, B., 2000. Plumes in a convecting mantle: models and observations
for individual hotspots. J. Geophys. Res. 105, 11127–11152.
Torsvik, T.H., Cocks, L.R.M., 2004. Earth geography from 400 to 250 million
years: a palaeomagnetic, faunal and facies review. Journal Geol. Soc. Lond.
161, 555–572.
Torsvik, T.H., Eide, E.A., Meert, J.G., Smethurst, M.A., Walderhaug, H.J.,
1998. The Oslo Rift: New palaeomagnetic and40Ar/39Ar age constraints.
J. Geophys. Int. 135, 1045–1059.
Torsvik, T.H., Eide, E.A., 1998. NGU GEOCHRON: Database and analysis for
Norwegian isotope geochronology. NGU Open File Report 98.003, 54 pp.
Torsvik, T.H., Smethurst, M.A., Burke, K., Steinberger, B., 2006. Large igneous
provinces generated from the margins of the large low-velocity provinces in
the deep mantle. Geophys. J. Intern. 167, 1447–1460.
Torsvik, T.H., Van der Voo, R., 2002. Refining Gondwana and Pangea
Palaeogeography: estimates of Phanerozoic (octupole) non-dipole fields.
Geophys. J. Int. 151, 771–794.
Torsvik, T.H., Müller, R.D., Van der Voo, R., Steinberger, B., Gaina, C., in press.
GlobalPlateMotionFrames:Towardaunifiedmodel,ReviewsofGeophysics.
Van der Voo, R., 1993. Paleomagnetism of the Atlantic, Tethys and Iapetus
Oceans. Cambridge University Press, Cambridge. 411 pp.
Van der Voo, R., Torsvik, T.H., 2001. Evidence for Permian and Mesozoic non-
dipole fields provides an explanation for the Pangea reconstruction
problems. Earth Planetary Science Letters 187, 71–81.
Vander Voo, R., Torsvik,T.H., 2004. The quality oftheEuropean Permo–Triassic
paleopolesanditsimpactonPangeareconstructions.In:Channel,J.E.T.,Kent,
D.V., Lowrie, W., Meert, J.G. (Eds.), Timescales of the Paleomagnetic Field,
Geophysical Monograph Series 145. American Geophysical Union, Washing-
ton DC, pp. 29–42.
Wilson, M., Neumann, E.R., Davies, G.R., Timmerman, M.J., Heeremans, M.,
Larsen, B.T. (Eds.), 2004. Permo–Carboniferous Magmatism and Rifting in
Europe. Geological Society London Special Publication, vol. 223. 498 pp.
452 T.H. Torsvik et al. / Earth and Planetary Science Letters 267 (2008) 444–452