Crustal transect across the North Atlantic

Mar Geophys Res (2008) 29:73–87 DOI 10.1007/s11001-008-9046-9 ORIGINAL RESEARCH PAPER Crustal transect across the North Atlantic R. Mjelde Æ T. Raum Æ A. J. Breivik Æ J. I. Faleide Received: 27 November 2007 / Accepted: 21 April 2008 / Published online: 14 May 2008 Ó Springer Science+Business Media B.V. 2008 Abstract Two dimensional crustal models derived from four different ocean bottom seismographic (OBS) surveys have been compiled into a 1,580 km long transect across the North Atlantic, from the Norwegian Møre coast, across the extinct Aegir Ridge, the continental Jan Mayen Ridge, the presently active Kolbeinsey Ridge north of Iceland, into Scoresby Sund in East Greenland. Backstripping of the transect suggests that the continental break-up at ca. 55 Ma occurred along a west-dipping detachment localized near the western end of a ca. 300 km wide basin thinned to less than 20 km crustal thickness. It is likely that an east-dipping detachment near the present day Liverpool Land Escarpment was active during the late stages of continental rifting. A lower crustal high-velocity layer (7.2–7.4 km/s) interpreted as mafic intrusions/underplating, was present beneath the entire basin. The observations are consistent with the plume hypothesis, involving the Early Tertiary arrival of a mantle plume beneath central Greenland and focused decompression melting beneath the thinnest portions of the lithosphere. The mid-Eocene to Oligocene continental extension in East Greenland is interpreted as fairly symmetric and strongly concentrated in the lower crustal layer. Continental break-up which rifted off the Jan Mayen Ridge, occurred at ca. 25 Ma, when the Aegir Ridge became extinct. The first ca. 2 m.y. of oceanic R. Mjelde (&)
T. Raum Department of Earth Science, University of Bergen, Allegt. 41, 5007 Bergen, Norway e-mail: [email protected] A. J. Breivik
J. I. Faleide Department of Geosciences, University of Oslo, P.O. Box 1047, Blindern, 0316 Oslo, Norway accretion along the Kolbeinsey Ridge was characterized by thin magmatic crust (ca. 5.5 km), whereas the oceanic crustal formation since ca. 23 Ma documents ca. 8 km thick crust and high magma budget. Keywords Crustal transect
North Atlantic
OBS data
Crustal evolution Introduction The area between Iceland and the Jan Mayen Fracture Zones represents one of the most complex tectonic regions of the Atlantic (Fig. 1). After the formation of the Caledonides, the area experienced several rifting episodes over a total of almost 350 m.y. culminating with continental break-up around magnetic anomaly 24B time (ca. 55 Ma, e.g. Brekke 2000). The break-up was associated with extensive magmatism, which most investigators relate to the arrival of a mantle plume (e.g. Eldholm et al. 1989; White and McKenzie 1989). Other proposed models for North Atlantic enhanced break-up magmatism include small-scale convection (e.g. Mutter et al. 1988), craton boundary effects (King and Anderson 1998) and fertile mantle melting (e.g. Foulger et al. 2005). One intriguing aspect concerning the evolution of the area is simultaneous continental extension along the East Greenland Margin and seafloor spreading at the Aegir Ridge (Kuvaas and Kodaira 1997). The Aegir Ridge became extinct just prior to magnetic anomaly seven (ca. 25 Ma), when seafloor spreading commenced along the Kolbeinsey Ridge separating the continental Jan Mayen Ridge from the East Greenland Margin (Vogt et al. 1980; Nunns 1983). Another significant observation is the asymmetry between break-up related magmatism and subsequent seafloor 123 74 spreading along the Aegir and Kolbeinsey Ridges: Earliest Eocene continental break-up produced large volumes of magma during and immediately after margin formation, but in the Norway Basin this magma pulse died down to ultra-slow, magma-starved seafloor spreading ca. 4 m.y. after break-up (Breivik et al. 2006). Rifting of the Jan Mayen micro-continent from Greenland during the Late Oligocene (30–25 Ma) was associated with minor magmatism, but transited to magma-robust seafloor spreading along the Kolbeinsey Ridge (Mjelde et al. 2007). Simultaneous continental extension and seafloor spreading has also been documented in the opening of the South China Sea (Clift et al. 2002) and in the Woodlark Basin (e.g. Mutter et al. 1996). Our understanding of the tectonic evolution of the North Atlantic has increased significantly following four different ocean bottom seismometer (OBS) surveys in 1988, 1995, 1996 and 2000 (Weigel et al. 1995; Kodaira et al. 1997, 1998; Raum 2000; Breivik et al. 2006). Here, we present new S-wave models of two OBS profiles, across the continental Møre Basin to the continent-ocean transition (COT) on the Møre Marginal High, and from the Møre Marginal High across the Aegir Ridge (Profiles 1-00 and Fig. 1 Bathymetry of the North Atlantic, with some of the main structural elements indicated. OBS profiles discussed in the text are indicated as white lines, and OBS profiles included in the transect are indicated as black lines. JLM: Jameson, Liverpool Land Margin 123 Mar Geophys Res (2008) 29:73–87 8A-96, Fig. 1). Finally, we link these crustal profiles to the earlier OBS investigations. The resulting compilation is a complete crustal transect from the Norwegian coastline into Scoresby Sund on East Greenland. We discuss the model in terms of the distribution of magmatism and modes of extension. The results are consistent with available multi-channel seismic (MCS), gravity and magnetic data, and the reader is referred to Weigel et al. (1995), Kodaira et al. (1997, 1998), Raum (2000) and Breivik et al. (2006) for details concerning the modeling. Data The present paper is based on the OBS profiles shown in Fig. 1. The OBS profiles used in the transect are indicated as black lines. The westernmost survey was performed in 1988 by use of the German icebreaker Polarstern (profile 1-88, Fig. 1). Two sources were used; (1) 32 dynamite shots, each 25 kg and (2) a single, 32 l Bolt airgun fired every 150–300 m. A total of seven land-stations and seven ocean bottom hydrophones (OBH) were used as recorders along the profile used for our transect. The land-stations were set up as geophone arrays. The OBH spacing was as large as 60 km for the offshore part of the profile, implying that the crustal resolution here is somewhat smaller than for the Kolbeinsey Ridge-Norway section. The processing was limited to band-pass filtering and gain adjustments, and the data were modeled by ray-tracing (forward modeling). The OBS models were subsequently constrained by gravity modeling. The reader is referred to Weigel et al. (1995) for details. The survey covering the area from the Kolbeinsey Ridge, across the Jan Mayen Basin to the Jan Mayen Ridge, was carried out in 1995 by use of the Norwegian Research Vessel, Håkon Mosby (profiles 3-95, 4-95, Fig. 1). A tuned air-gun array with a total volume of 42 l was shot every 100 m. Three-component, analogue OBSs, developed at Hokkaido University, were deployed at ca. 25 km intervals along the profiles making this section the best resolved part of the North Atlantic transect. As for the East Greenland margin data, the processing was limited to band-pass filtering and gain compensations, but travel-time inversion was performed in addition to forward ray-tracing, by use of the software developed by Zelt and Smith (1992). The P-wave models were then constrained by gravity- and S-wave modeling (Kodaira et al. 1997, 1998; Mjelde et al. 2002a, 2007). The data covering the eastern part of the Jan Mayen Ridge, the oceanic crust formed at the extinct Aegir Ridge, and eastwards across the COT to the Møre Marginal High, was acquired in 2000 by use of Håkon Mosby (profiles 8-00, Mar Geophys Res (2008) 29:73–87 75 1-00, Fig. 1). The source consisted of four air-guns with a total volume of 78 l, fired every 200 m. The seismic data were recorded by Hokkaido University OBSs deployed every ca. 25 km. The processing and modeling was identical to that of the Kolbeinsey Ridge—Jan Mayen Ridge data. Examples of data, travel-time curves and ray-paths are shown in Figs. 2–5. The reader is referred to Breivik et al. (2006) for further details on the P-wave and gravity modeling. The 250 km long profile extending from the Møre Marginal High, across the Møre Basin to the Norwegian coastline was acquired in 1996 by use of Håkon Mosby, the 78 l air-gun source fired every 200 m, and 10 three-component, analogue, Hokkaido University OBSs (profile 8A96, Fig. 1). The processing consisted of 5–12 Hz band-pass filtering, and a 4 s automatic gain control (AGC) window was used for scaling. The P-wave and gravity modeling is discussed by Raum (2000). Fig. 2 (a) Model for profile 1-00, from the Aegir Ridge to the Møre Basin, with calculated ray-paths for OBS 2. (b) Comparison between calculated (solid lines) and interpreted (hatched lines) for OBS 2. Sg2: S-wave in oceanic layer 2, Sg3a: S-wave in oceanic layer 3A, Sg3b: Swave in oceanic layer 3B, SMS: S-wave reflection at Moho. All these arrivals were P-to-S converted on the way down on top of oceanic layer 2. PnS: Critical P-wave reflection along Moho, P-to-S converted at the base of the crust. (c) Horizontal component data for OBS 2 (5– 12 Hz band-pass filtered) 123 76 Mar Geophys Res (2008) 29:73–87 Fig. 3 (a) Model for profile 1-00 with calculated ray-paths for OBS 4. (b) Comparison between calculated (solid lines) and interpreted (hatched lines) for OBS 4. Sg2: S-wave in oceanic layer 2, Sg3a: Swave in oceanic layer 3A, Sg3b: S-wave in oceanic layer 3B, SMS: Swave reflection at Moho, Sn: upper mantle S-wave refraction. All these arrivals were P-to-S converted on the way down at the seafloor. Pg3aS: P-to-S conversion at the top of layer 3A on the way up, Pg3bS: P-to-S conversion at the top of layer 3B on the way up. (c) Horizontal component data for OBS 4 (5–12 Hz band-pass filtered) Results S-wave modeling of arrivals interpreted from the OBS horizontal components is considered as the final step in state-of-the-art modeling of regional OBS data. The interfaces derived from the P-wave (and gravity) modeling are kept constant, and the modeling thus consists of obtaining the S-wave velocity, expressed as the Vp/Vs-ratio, for each layer, as well as the interface of P–S conversion for all interpreted arrivals. The ray-tracing module within RayInvr S-wave modeling of the Møre Basin-Aegir Ridge profiles The S-wave modeling of the Møre Basin-Aegir Ridge profiles (1-00 and 8A-96, Fig. 1) is based on the published Pwave and gravity models presented in the previous section. 123 Mar Geophys Res (2008) 29:73–87 77 Fig. 4 (a) Model for profile 1-00 with calculated ray-paths for OBS 6. (b) Comparison between calculated (solid lines) and interpreted (hatched lines) for OBS 6. Sg2: S-wave in oceanic layer 2, Sg3b: Swave in oceanic layer 3B, SMS: S-wave reflection at Moho. All these arrivals were P-to-S converted on the way down on top of oceanic layer 2. PnS: Critical P-wave reflection along Moho, P-to-S converted at the base of the crust. (c) Horizontal component data for OBS 6 (5– 12 Hz band-pass filtered) (Zelt and Smith 1992) has been used for this purpose. Lateral variations within layers are derived from the modeling of different OBSs along the profiles. Waves that have been converted on the way down are labeled PSSarrivals, and conversions on the way up are labeled PPSarrivals. The former are most useful, as they provide a direct estimate of the S-wave velocity in the layer of critical refraction. In order to assure that secondary P-waves are not mis-interpreted as S-waves, the vertical components are also used during the interpretation. S-wave arrivals generally appear with highest amplitudes on the horizontal components. Final discrimination between P- and S-waves is achieved through the use of particle diagrams. In the uppermost 1–2 km below the seafloor the Vp/Vsratio is generally found to be high (2.5–10), which is interpreted to indicate high porosities and the presence of relatively unconsolidated sediments (e.g. Mjelde et al. 2003). For deeper sedimentary layers the Vp/Vs-ratio is 123 78 Mar Geophys Res (2008) 29:73–87 Fig. 5 (a) Model for profile 1-00 with calculated ray-paths for OBS 10. (b) Comparison between calculated (solid lines) and interpreted (hatched lines) for OBS 10. Sg3b: S-wave in oceanic layer 3B, SMS: Swave reflection at Moho. All these arrivals were P-to-S converted on the way down on top of oceanic layer 2. PnS: Critical P-wave reflection along Moho, P-to-S converted at the base of the crust, Pg3aS: P-to-S conversion at the top of layer 3A on the way up. (c) Horizontal component data for OBS 10 p (5–12 Hz band-pass filtered) considered as a lithological indicator, with sandstone and shale as end-members. The Vp/Vs-ratios are generally found to be ca. 1.6 for sand and ca. 2.0 for shale, and carbonates have intermediate values (e.g. Domenico 1984). In crystalline rocks, a distinction can be made between felsic and mafic composition, with Vp/Vs-ratios of ca. 1.75 and 1.85, respectively (Holbrook et al. 1992). The uncertainty in the Vp/Vs-ratio is estimated to +/–0.05. The reader is referred to Mjelde et al. (2003) for a more detailed discussion on S-wave modeling, uncertainties and references to laboratory measurements on various lithologies. The final Vp/Vs-ratios are indicated as bold numbers in Figs. 6–7. The high Vp/Vs values for the uppermost layers (2.65–3.30) can be attributed to poorly consolidated sediments. The Vp/Vs-ratio decreases steadily with depth in the sedimentary section, which is indicative of increased 123 Mar Geophys Res (2008) 29:73–87 79 The Vp/Vs-ratios for the oceanic crust are relatively high (1.80–1.90), in agreement with a mafic composition (Fig. 7). There appears to be a slight reduction in the Vp/Vs-ratio with depth in the oceanic crust, which, if real, may be attributed to reduced fracture density with depth (Mjelde et al. 2002b). The Vp/Vs-ratios in the upper crust are somewhat higher near the Aegir Ridge, ca. 2.0. We interpret this anomaly in terms of increased crustal fracturing near the ridge (Flovenz 1980). Other factors affecting the Vp/Vsratio within oceanic crust are; the presence of partly serpentinized peridotities, which would increase the Vp/Vsratio, and increased Mg content at the expense of Fe, which tends to decrease the Vp/Vs-ratio (Mjelde et al. 2002a). We do not consider the resolution in our data to be high enough to discuss the model in these terms. The same applies to the upper mantle Vp/Vs-ratio estimates. The anomalously high upper mantle P-wave velocities between 270 and 290 km along the model (ca. 8.35 km/s, Fig. 6; unresolved for Swaves) are interpreted as partly eclogitized rocks, based on similarities with the southern Vøring Basin and eastern North Sea (Raum et al. 2006; Christiansson et al. 2000). The North Atlantic transect—tie between different surveys Fig. 6 Crustal velocity model for profile 8A-96, across the Møre Margin. P-wave velocities (from Raum 2000) are indicated as small numbers, and Vp/Vs-ratios as larger numbers. EFB/SDR: Eocene Flood Basalt/Seaward Dipping Reflectors. MMH: Møre Marginal High. OBS positions (1–9) are indicated on the seafloor. The Early Tertiary underplating is interpreted as a mixture of mafic, lower crustal intrusions and older continental blocks compaction and reduced porosity with depth. Borehole data indicate that carbonate rocks are sparse in the Møre Basin area (Brekke 2000), and the intermediate Vp/Vs-ratio in the sedimentary section below ca. 2 km depth (ca. 1.75) suggests that it is dominated by a mixture of sandstone and shale. The Vp/Vs-ratios in the upper- and lower-continental, crystalline crust are modeled to be ca. 1.70 and ca. 1.80, respectively. These values are consistent with felsic upper crustal and more mafic lower crustal compositions (Holbrook et al. 1992). We interpret the lower crustal layer as a mixture of mafic intrusions related to the last phase of rifting and break-up, and older, felsic continental lower crustal blocks. This interpretation is in agreement with models for the Vøring Margin located to the north (Fig. 1). An alternative hypothesis relates the high-velocity lower crustal body to high-grade Caledonian terranes (e.g. Gernigon et al. 2003). This hypothesis may be valid, provided that these terranes are of mafic composition (Mjelde et al. 2008b). The crustal scale models discussed above have been compiled into a North Atlantic transect in Fig. 8a, where the interpretation of the layers have been based on combined use of P-wave, S-wave and gravity modeling. The Pwave models are most important with this respect, as they provide the highest resolution. The different surveys, from which data were compiled into the transect, were designed to resolve relatively local features. Consequently, there are generally differences in azimuth between individual profiles in the transect, and the profiles are offset laterally at the tie-points (Fig. 1). The only exception is the tie between the easternmost Møre Basin and the Møre COT profiles. The latter profile, 1-2000, was acquired along a similar azimuth and with ca. 30 km overlap with the eastern Møre Basin profile, 8A-96. In the transect, we have used the entire length of profile 1-00, because this profile covers the COT on the Møre Marginal High. The lateral offset between Møre profile 1-00 and Jan Mayen profile 8-00 at the tie-point near the Aegir Ridge is almost 300 km. However, the locations where the two profiles cross the COT at the Møre Margin and eastern Jan Mayen Ridge, are very close to conjugate (Kodaira et al. 1998; Breivik et al. 2006). The westward continuation of profile 1-00 across the Aegir Ridge, i.e. the eastermost ca. 70 km of profile 7-00 (Fig. 1), has a very similar crustal structure to profile 8-00, whose eastern end maps the older Jan Mayen Fracture Zone. This northernmost profile gives the most representative image of the oceanic crust formed 123 80 Mar Geophys Res (2008) 29:73–87 Fig. 7 Crustal velocity model for profile 1-00, from the Aegir Ridge to the Møre Margin. P-wave velocities (from Breivik et al. 2006) are indicated as small numbers, and Vp/Vs-ratios as larger numbers. Uncertain estimates are within brackets. EFB/SDR: Eocene Flood Basalt/Seaward Dipping Reflectors. OBS positions (1–13) are indicated on the seafloor. The lower crustal layer is interpreted as Early Tertiary underplating east of the COB, and oceanic layer 3B further westwards on the western side of the Aegir Ridge, because profile 7-00 is strongly influenced by the Iceland Plume at less than 100 km distance from the Aegir Ridge (Mjelde et al. 2008a). The Jan Mayen Ridge is mapped by profiles 8-00 and profile 4-95, which are offset by ca. 30 km on the ridge. The overall crustal models are very similar, in terms of layering and crustal thickness (ca. 15 km; Kodaira et al. 1998; Mjelde et al. 2008a). In the transect we have used the model of profile 4-95 since this profile crosses stretched crust at the eastern margin of the Jan Mayen Basin. The azimuths of profile 4-95 and profile 3-95 are exactly the same. The latter is offset ca. 30 km to the south and extends from the Jan Mayen Basin to the Kolbeinsey Ridge. The westernmost profile was compiled by Weigel et al. (1995). It includes two relatively short profiles acquired within Scoresby Sund, and a longer profile extending across the Kolbeinsey Ridge (Fig. 1). The trend of the westernmost profile is EW, and its tie-point with profile 3-95 is located ca. 30 km further north. The Moho is poorly constrained in the horizontal ranges 130–170 and 270–380 km (Figs. 8a, e and 9a). 123 Fig. 8 Combined (unsmoothed) transect. (a) Present-day model, (b) c reconstruction at 25 Ma, (c) 52 Ma, (d) 55 Ma (North Atlantic breakup). (e) The original version of the westernmost part of the model, redrawn from Weigel et al. (1995; the vertical scale is the same as in A, but the horizontal scale has been doubled). SDR: Eocene Flood Basalt/Seaward Dipping Reflectors, LLE: Liverpool Land Escarpment (detachment), JMR: proto Jan Mayen Ridge, AR: Aegir Ridge, COB: Continent-Ocean-Boundary detachment, MM: Møre Margin. The brown lower crustal layer is interpreted as non-intruded lower crust, whereas the dark grey lower crustal layer (underplating) is interpreted as a mixture of mafic, lower crustal intrusions and older continental blocks landward of the COB, and oceanic layer 3B elsewhere. Dots on the seafloor in a) are OBS positions Mar Geophys Res (2008) 29:73–87 81 123 82 123 Mar Geophys Res (2008) 29:73–87 Mar Geophys Res (2008) 29:73–87 b Fig. 9 Smoothed and interpreted transect. (a) Present-day model, (b) reconstruction at 25 Ma, (c) 52 Ma, (d) 55 Ma (North Atlantic breakup), (e) same as D with Early Tertiary underplating removed. See text for details. SDR: Eocene Flood Basalt/Seaward Dipping Reflectors, LLE: Liverpool Land Escarpment (detachment), JMR: proto Jan Mayen Ridge, AR: Aegir Ridge, COB: Continent-Ocean-Boundary detachment, MM: Møre Margin. The brown lower crustal layer is interpreted as non-intruded lower crust, whereas the dark grey lower crustal layer (underplating) is interpreted as a mixture of mafic, lower crustal intrusions and older continental blocks landward of the COB, and oceanic layer 3B elsewhere. Dots on the seafloor in a) are OBS positions Discussion 83 can partly be related to the Jan Mayen Basin being significantly narrower along profile 3-95. Furthermore, the thickness of the Cenozoic sedimentary section decreases southwards (Kodaira et al. 1998). The adjustments performed in Fig. 9a represent roughly the average depths of interfaces for profile 3-95 and 4-95. The two profiles mapping the Kolbeinsey Ridge match perfectly across the oceanic layer 2/3A interface. However, the layer 3A/3B and Moho interfaces are significantly deeper on the westernmost profile. This difference is partly real, since an OBS profile acquired along the Kolbeinsey Ridge revealed these interfaces to be 1 and 1.5 km deeper ca. 30 km north of profile 3-95 (Kodaira et al. 1997). The transect—adjustment and interpretation of interfaces The transect—main provinces Figure 9a shows a modified version of the transect. Here we have adjusted the position of most interfaces in the tiepoints between the profiles. Some of these adjustments are straightforward, because the uncertainty in depth to interfaces generally increases towards the ends of the profiles. However, the larger differences observed in Fig. 8a are real, due to offsets between the profile’s tie-points. The adjustments made in Fig. 9a represent averaging of the interfaces at the tie-points, performed in order to make the general aspects of the transect stand out more clearly. It is important to emphasize that these modifications contain elements of interpretation. No significant adjustments were needed for the profiles mapping the Møre Marginal High. On the western side of the Aegir Ridge, the adjustments to the layer 2/3A and 3A/3B interfaces are within modeling uncertainties. The slightly increased total crustal thickness on the western side of the ridge (compared to the eastern side) is tied to increased magma budget in the northern Norway Basin (Breivik et al. 2006; Mjelde et al. 2008a). On the eastern side of the Jan Mayen Ridge, we have increased the depth to the top of the Eocene basalt ca. 1 km in agreement with the observed eastward continuity across the COT on profile 8-00. The deeper interfaces have been merged in agreement with the COT interpretation discussed by Mjelde et al. (2008a). The same applies to the lower crustal high-velocity layer (‘underplating’). We emphasize that we interpret the continental, lower crustal high-velocity layer as being a mixture of break-up related mafic intrusions and older continental, lower crustal blocks. In Figs. 8 and 9 the continental, intrusive lower crust is marked by the same colour as oceanic layer 3B (dark grey), because the layers are genetically related across the COTs. Non-intruded continental lower crust (East Greenland and most of the Jan Mayen Ridge) is marked as dark brown. On the western side of the Jan Mayen Ridge, differences between profiles 3-95 and 4-95 The westernmost 230 km of the profile extends from underplated continental crust within the inner parts of Scoresby Sund over to oceanic crust formed at the Kolbeinsey Ridge (Fig. 9a). Exposed basement windows in East Greenland reveal that the crystalline basement dominantly consists of Mid-Proterozoic, Caledonian migmatites and Caledonian intrusions. The Jameson Land Basin is made up of sedimentary rocks ranging in age from Devonian to Cretaceous. Just east of the present day coastline (at 170 km on the profile), the Liverpool Land Escarpment marks the western boundary of a deep, Cenozoic basin underlain by thin crust of uncertain origin. Based on our reconstruction, we interpret the western part of this basin to be thinned continental crust in agreement with Weigel et al. (1995). The lower crustal high-velocity bodies (7.4– 7.6 km/s) are interpreted by Weigel et al. (1995) as Early Tertiary intrusions/underplating. This magmatism was tied to the continental rifting of Greenland/Jan Mayen from the Møre margin, while the rifting of the Jan Mayen block from Greenland in the Late Oligocene was apparently amagmatic prior to the onset of seafloor spreading (Weigel et al. 1995; Kodaira et al. 1998). Oceanic crust formed at the Kolbeinsey Ridge from ca. 25 Ma to present, is located between 230 and 630 km on the profile (Fig. 9a). The transition from thinned continental crust in the Jan Mayen Basin to thick oceanic crust (ca. 8 km) appears very abrupt, but a profile extending obliquely across the COT revealed that the oldest oceanic crust is thin (ca. 5.5 km; Kodaira et al. 1998) in agreement with slow spreading and a low magma budget (White et al. 1992). The thin crust is also observed close to Greenland between distances 230 and 300 km. The larger uncertainty along this profile implies that the zone of thin crust may be significantly narrower than indicated in the model. Our interpretation of the COT at 230 km implies that spreading along the Kolbeinsey Ridge has been symmetric and continuous since 25 Ma (Kodaira et al. 1997), with increased 123 84 crustal thickness from ca. 23 Ma due to increased influence from the Icelandic Plume to the south (pulsing plume; Mjelde et al. 2008a). Alternatively, the increase in melting at ca. 23 Ma could be related to interaction between the spreading ridge and a heterogeneous mantle source (Foulger et al. 2005). Our interpretation locates the COT in the centre of the thinned Jan Mayen Basin and provides a straightforward correlation of layers across the COT. The apparent absence of oceanic layer 2 between 230 and 370 km is most likely due to increased alteration and closure of pore space caused by the increased overburden. A similar strong increase in P-wave velocity in the uppermost part of the oceanic crust has been documented beneath thick Cenozoic wedges west of the Barents Sea (Breivik et al. 2003). The Jan Mayen Ridge is located between 630 and 780 km. The western part represents continental crust, thinned prior to continental break-up at 25 Ma. Dynamical modeling based on strength estimates suggests that the continental thinning commenced at ca. 42.5 Ma (Mjelde et al. 2008a). A layer of pre-Cretaceous rocks, up to 3 km thick, is located along the western flank of the ridge. Most of the thinning occurred in the lower crust, and this western margin is classified as amagmatic (Kodaira et al. 1998). Any underplating in this area would have been revealed by the relatively dense data coverage. The eastern side of the ridge represents the volcanic margin rifted off the Møre Margin at ca. 55 Ma (Mjelde et al. 2008a). The COT east of the Jan Mayen Ridge between 780 and 810 km consists of underplated continental crust and anomalously thick oceanic crust, formed during the first ca. 3 m.y. of spreading along the Aegir Ridge. The ca. 12 km thick oceanic crust formed during North Atlantic break-up is generally interpreted to be related to the arrival of a mantle plume, currently located beneath Iceland (e.g. Eldholm et al. 1989; White and McKenzie 1989). Thinner oceanic crust generated at the Aegir Ridge from 52 Ma to ca. 25 Ma (Breivik et al. 2006), is located between 810 and 1,180 km in the model. The spreading was asymmetric, with higher spreading rates to the east, and ca. 1 km thicker crust on the western side (Breivik et al. 2006). However, if we take into account synchronous continental rifting west of the Jan Mayen Ridge and seafloor spreading at the Aegir Ridge, then total spreading appears to be symmetric. Anomalously thick oceanic crust (8–12 km) formed from ca. 55 to ca. 52 Ma off the Møre Marginal High, is located between 1,180 and 1,250 km. With the exception of the Cenozoic section, the area further eastwards corresponds stratigraphically to the Jameson Land sequence on East Greenland, with sedimentary rocks ranging in age from Devonian to Cretaceous (Weigel et al. 1995; Brekke 2000). With reference to the COT, the lower crustal underplating extends further eastwards beneath the Møre Basin than 123 Mar Geophys Res (2008) 29:73–87 westwards beneath the Jan Mayen Ridge, but if underplating under eastern Greenland is included it becomes more symmetric. An explanation for this observation is discussed below. Conjugate margin reconstruction The data presented here is not sufficient to perform a complete tectonic reconstruction of all layers in the transect. However, simple closure of oceanic crust back to the time of break-up provides valuable images for increasing our understanding of the area’s geological evolution (Fig. 9). In Fig. 9b (25 Ma) we have removed all oceanic crust formed at the Kolbeinsey Ridge along with all sediments deposited after ca. 25 Ma. In order to compensate for the observed basin subsidence associated with the thick Cenozoic sedimentary sequence on the East Greenland margin, we have lifted the part of the model located east of 170 km ca. 3 km. We infer that part of the subsidence most likely can be attributed to an east-dipping normal fault near the Liverpool Land Escarpment. We recognize that this simple restoration need to be refined with tectonic backstripping taking compaction and flexural rigidity into account. Figure 9c shows the model after removal of the ca. 42.5–25 Ma extension between the Jan Mayen Ridge and East Greenland, as well as the ca. 52–25 Ma (thin) oceanic crust formed at the Aegir Ridge. The entire sequence of Cenozoic sediments has been removed as well. The Jan Mayen models indicate that the pre-Cretaceous basin is presently located beneath the western flank of the Jan Mayen Ridge, indicating that the east-dipping normal fault at the Liverpool Land Escarpment may have been important during the extension (Fig. 9d). However, faults along the western flank of the Jan Mayen Ridge are dominantly dipping westwards (Kuvaas and Kodaira 1997). The resolution in our models is not large enough to allow distinction between pure- or simple shear deformation for this part of the area. Figure 9d represents crustal structure of the North Atlantic continental break-up, ca. 55 Ma. Breivik et al. (2006) and Mjelde et al. (2008a) argued that the COT on the Møre Margin and eastern Jan Mayen Ridge Margin could be confined to a crustal scale west-dipping detachment that nucleated the continent-ocean-boundary (COB). The main movement along the detachment is assumed to have occurred during the continental thinning prior to the onset of the thermal anomaly. Based on the present model it cannot be excluded that the final break-up was more symmetric, but we note that the proposed detachment matches the dipping Moho beneath East Greenland. From Fig. 9d it is clear that the Early Eocene break-up occurred Mar Geophys Res (2008) 29:73–87 near the western end of the highly stretched area between Norway and Greenland. This observation is in agreement with the hot-spot hypothesis locating the Iceland mantle plume beneath central Greenland at the time of break-up (e.g. Müller et al. 1993). Decompression melting would be concentrated in areas with thinned lithosphere producing the underplated lower crust (e.g. White and McKenzie 1989). Note that the reconstruction indicates that the lower crustal 7.5 km/s velocity body, observed east of Liverpool Land (150–230 km in Fig. 9d), belongs to the Palaeocene/ Early Eocene underplating beneath the proto-Jan Mayen Ridge. Similar lower crustal high-velocity bodies have been modeled in the central Vøring Basin (Mjelde et al. 1997). Figure 9e shows the model before emplacement of this body east of Liverpool Land. We have here assumed that the entire body was emplaced immediately prior to and during break-up, but emphasize that the intrusive period may have started at least 10 m.y. earlier. Note that the continental crust near the COB may be strongly intruded on both sides. On the western side of the COB, the intruded zone may be at least 50 km wide. The mixture of continental and oceanic crust can be exemplified by the piece of pre-Cretaceous rock included on the western side of the COB in Fig. 9d, which in Fig. 9c has been completely assimilated within the underplating. Figure 9d shows that at the time of break-up at ca. 55 Ma, the postulated east-dipping fault near the Liverpool Land Escarpment was assumed to be of significance during the extension west of the Jan Mayen Ridge, and it matches the west-dipping fault that nucleated the COB. It is thus possible that the extension west of the Jan Mayen Ridge was inititated some million years prior to the continental break-up. Further evidence in favor of East Greenland extension may come from paleomagnetic studies of plate movements relative to the hot-spot reference frame (Torsvik et al. 2001). These studies show that Eurasia and the Greenland plates followed the same tracks during Cretaceous and Early Tertiary until continental break-up at ca. 55 Ma. However, after break-up Greenland continued its northwestward movement bringing East Greenland/protoJan Mayen Ridge towards the centre of the Iceland plume, whereas Eurasia switched to a more northerly propagation. Relative to the plume, the Eurasian plate has moved significantly less than the Greenland plate after break-up. Summary and conclusions A Vp/Vs-model from the Møre Margin COT across the extinct Aegir Ridge documents a clear lateral increase in the Vp/Vs-ratio from ca. 1.75 in the felsic, continental crust to ca. 1.85 in the mafic, oceanic crust. A similar increase is 85 observed with depth in the Møre Basin (1.70–1.80), which is consistent with felsic upper crust and mafic lower crust (underplating). Increased upper crustal values (ca. 2.0) near the Aegir Ridge are interpreted as increased crack porosity. The crustal models are integrated with earlier models from the Aegir Ridge, across the continental Jan Mayen Ridge, the presently active Kolbeinsey Ridge, to Scoresby Sund on East Greenland. The compilation represents a complete, 1,580 km long crustal transect across the North Atlantic. Reconstruction and interpretation of the transect suggest that continental break-up at ca. 55 Ma occurred near the western end of a ca. 300 km wide basin thinned to less than 20 km crustal thickness. A lower crustal high-velocity layer, interpreted as comprised of mafic intrusions/underplating, underlies the entire basin. The COB is related to a west-dipping, crustal detachment. The first ca. 3 m.y. of seafloor spreading was characterized by the formation of anomalously thick oceanic crust, probably due to high melting potential provided by elevated mantle temperatures. Subsequent seafloor spreading along the Aegir Ridge occurred under a low magma budget creating anomalously thin oceanic crust. Simultaneously, continental stretching occurred in East Greenland/Jan Mayen. The model indicates that part of the extension may have been accommodated along an eastdipping detachment near the present day Liverpool Land escarpment. This fault may have been active prior to the commencement of oceanic spreading along the Aegir Ridge. Fairly symmetric extension led to continental breakup west of the Jan Mayen Ridge at ca. 25 Ma. At the same time, the Aegir Ridge became extinct. The first ca. 2 m.y. of oceanic accretion along the Kolbeinsey Ridge was characterized by a low magma budget and formation of thin crust, but increased activity from the North Atlantic plume from ca. 23 Ma has resulted in accretion of ca. 8-km-thick oceanic crust, which is ca. 2 km thicker than the global average. Early Tertiary break-up magmatism was caused by hot asthenospheric material emplaced within zones of lithospheric extension along the line of continental break-up. This mantle material had increased melt potential but was spent during a few million years of seafloor spreading (e.g. Holbrook et al. 2001). Symmetrical V-shaped gravimetric ridges, which can be traced back to ca. 48 Ma, document large-scale asthenospheric flow both north (Aegir) and south of Iceland. Such flow is predicted by geodynamical models of mantle plumes, but has yet to be predicted by other mechanisms (Mjelde et al. 2008a). The continental stretching leading to Late Oligocene break-up between East Greenland and Jan Mayen was amagmatic. Only after a couple of million years of seafloor spreading on the Kolbeinsey Ridge was the influence of the 123 86 Iceland hotspot established, resulting in thicker than normal oceanic crust. This asymmetry may be related to largescale pulsing of the North Atlantic plume, being relatively weak in the late Eocene-Oligocene (e.g. Smallwood and White 1998). Alternatively, it may suggest direct plumeridge interaction from 23 Ma. Acknowledgements The crew of R/V Håkon Mosby and engineers from the University of Bergen are greatly acknowledged for their skills and help in acquiring the OBS data from the Norwegian Coastline to the Kolbeinsey Ridge. We also thank H. Shimamura, H. Shiobara, S. Kodaira, Y. Murai and engineers from Hokkaido University for invaluable participation in planning and executing these surveys, as well as for initial processing of the OBS data, and we thank Beata Mjelde for drawing figures. The Norwegian Petroleum Directorate (NPD), Statoil and Norsk Hydro funded these projects, and we thank in particular to E. Bråstein and H. Brekke (NPD), S. Thorbjørnsen (Statoil), as well as G. Haatvedt and R. Karpuz (Hydro). The modeling was done with help of the inversion/forward modeling software developed by C. Zelt (Rice University, Houston). The present paper can to a large extent be attributed to the fruitful environment for scientific discussions the first author experienced during a research stay in 2006 at SOEST, University of Hawaii. The paper was finalized during a research stay in 2007 at Institute for Seismology and Volcanology, Hokkaido University. We finally thank Peter Clift and two anonymous reviewers for very constructive comments. References Breivik AJ, Mjelde R, Grogan P, Shimamura H, Murai Y, Nishimura Y (2003) Crustal structure and transform margin development south of Svalbard based on ocean bottom seismometer data. Tectonophysics 369:37–70 Breivik AJ, Mjelde R, Faleide JI, Murai Y (2006) Rates of continental break-up magmatism and seafloor spreading in the Norway Basin—constraints on Iceland plume activity. J Geophys Res 111:1–17 Brekke H (2000) The tectonic evolution of the Norwegian Sea continental margin, with emphasis on the Vøring and Møre Basins. In: Nøttvedt A (ed) Dynamics of the norwegian margin. Geological Society of London Special Publication 167, pp 327–378 Christiansson P, Faleide JI, Berge AM (2000) Crustal structure in the northern North Sea: an integrated geophysical study. In: Nøttvedt A (ed) Dynamics of the norwegian margin. Geological Society of London Special Publication 167, pp 15–40 Clift P, Lin J, Barckhausen U (2002) Evidence of low flexural rigidity and low viscosity lower continental crust during continental break-up in the South China Sea. Mar Pet Geol 19(8):951–970 Domenico SN (1984) Rock lithology and porosity determination from shear and compressional wave velocity. Geophysics 49:1188– 1195 Eldholm O, Thiede J, Taylor E (1989) Evolution of the Vøring volcanic margin, Proc. ODP, Sci. Results 104, College Station, TX (Ocean Drilling Program), pp 1033–1065 Flovenz OG (1980) Seismic structure of the Icelandic crust above layer three and the relation between body wave velocity and the alteration of the basaltic crust. J Geophys Res 47:211–220 Foulger GR, Natland JH, Anderson DL (2005) A source for icelandic magmas in remelted Iapetus crust. J Volcanol Geotherm Res 141:23–44 Gernigon L, Ringenbach JS, Planke S, Le Gall B, Jonquet-Kolstø H (2003) Extension, crustal structure and magmatism at the outer 123 Mar Geophys Res (2008) 29:73–87 Vøring Basin, Norwegian margin. J Geol Soc London 160: 197–208 Holbrook WS, Mooney WD, Christensen J (1992) The seismic velocity structure of the deep continental crust. In: Fountain DM, Arculus R, Kay RW (eds) Continental lower crust, development in geotectonics, vol 23. Elsevier, Amsterdam, pp 1–43 Holbrook WS, Larsen HC, Korenaga J, Dahl-Jensen T, Reid ID, Kelemen PB, Hopper JR, Kent GM, Lizarralde D, Bernstein S, Detrick RS (2001). Mantle thermal structure and active upwelling during continental breakup in the North Atlantic. Earth Planet Sci Lett 190:251–266 King SC, Anderson DL (1998) Edge-driven convection. Earth Planet Sci Lett 160:289–296 Kodaira S, Mjelde R, Shimamura H, Gunnarsson K, Shiobara H (1997) Crustal structure of the Kolbeinsey ridge, N. Atlantic, obtained by use of Ocean bottom seismographs. J Geophys Res 102:3131–3151 Kodaira S, Mjelde R, Gunnarsson K, Shiobara H, Shimamura H (1998) Structure of the Jan Mayen micro-continent and implications for its evolution. Geophys J Int 132:383–400 Kuvaas B, Kodaira S (1997) The formation of the Jan Mayen microcontinent: the missing piece in the continental puzzle between the Møre-Vøring Basins and East Greenland. First Break 15(7):239–247 Mjelde R, Kodaira S, Shimamura H, Kanazawa T, Shiobara H, Berg EW, Riise O (1997) Crustal structure of the central part of the Vøring Basin, mid-Norway margin, from ocean bottom seismographs. Tectonophysics 277:235–257 Mjelde R, Aurvåg R, Kodaira S, Shimamura H, Gunnarsson K, Nakanishi A, Shiobara H (2002a) Vp/Vs-ratios from the central Kolbeinsey ridge to the Jan Mayen Basin, North Atlantic; implications on lithology, porosity and present-day stress field. Mar Geophys Res 23:125–145 Mjelde R, Kasahara J, Shimamura H, Kamimura A, Kanazawa T, Kodaira S, Raum T, Shiobara H (2002b) Lower crustal seismic velocity-anomalies; magmatic underplating or serpentinized peridotite? Evidence from the Vøring margin, NE Atlantic. Mar Geophys Res 23:169–183 Mjelde R, Raum T, Digranes P, Shimamura H, Shiobara H, Kodaira S (2003) Vp/Vs-ratio along the Vøring margin, NE Atlantic, derived from OBS-data; implications on lithology and stressfield. Tectonophysics 369:175–197 Mjelde R, Eckhoff I, Solbakken S, Kodaira S, Shimamura H, Gunnarsson K, Nakanishi A, Shiobara H (2007) Gravity and Swave modeling across the Jan Mayen ridge, North Atlantic; implications for crustal lithology and continental break-up processes. Mar Geophys Res 28:27–41 Mjelde R, Breivik AJ, Raum T, Mittelstaedt E, Ito G, Faleide JI (2008a) Magmatic and tectonic evolution of the North Atlantic. J Geol Soc London 164:1–12 Mjelde R, Faleide JI, Breivik AJ, Raum T, Wilson J (2008b) Lower crustal composition and crustal lineaments on the Vøring margin, NE Atlantic. Tectonophysics (in press) Müller RD, Royer JY, Lawver LA (1993) Revised plate motions relative to the hotspots from combined Atlantic and Indian Ocean hotspot tracks. Geology 16:275–278 Mutter JC, Buck WR, Zehnder CM (1988) Convective partial melting, a model for the formation of thick basaltic sequences during the initiation of spreading. J Geophys Res 93:1031–1048 Mutter JC, Mutter CZ, Fang J (1996). Analogies to oceanic behaviour in the continental break-up of the western Woodlark Basin. Nature 380:333–336 Nunns AG (1983) Plate tectonic evolution of the Greenland-Scotland ridge and surrounding regions. In: Bott MHP, Saxow S, Talwani M, Thiede J (eds) Structure and development of the Greenland- Mar Geophys Res (2008) 29:73–87 Scotland ridge; new methods and concepts. Nato Adv. Res. Inst., Plenum Press, New York, pp 11–30 Raum T (2000) Crustal structure and evolution of the Faroe, Møre and Vøring margins from wide-angle seismic and gravity data. Ph.D. thesis, University of Bergen, pp 114 Raum T, Mjelde R, Shimamura H, Murai Y, Bråstein E, Karpuz RM, Kravik K, Kolstø HJ (2006) Crustal structure and evolution of the southern Vøring Basin and Vøring transform margin, NE Atlantic. Tectonophysics 415:167–202 Smallwood JR, White RS (1998) Crustal accretion at the Reykjanes ridge. J Geophys Res 103:5185–5201 Torsvik TH, Van der Voo R, Meert JG, Mosar J, Walderhaug HJ (2001) Reconstructions of the continents around the North Atlantic at about the 60th parallel. Earth Planet Sci Lett 187:55–69 Vogt PG, Johnson GL, Kristjansson L (1980) Morphology and magnetic anomalies North of Iceland. J Geophys 47:67–80 87 Weigel W, Flüh ER, Miller H, Butzke A, Dehghani GA, Gebhardt V, Harder I, Hepper J, Jokat W, Kläschen D, Kreymann S, Schüßler S, Zhao Z (1995) Investigations of the East Greenland continental margin between 70° and 72°N by deep seismic sounding and gravity studies. Mar Geophys Res 17:167–199 White RS, McKenzie D (1989) Magmatism at rift zones: the generation of volcanic continental margins and flood basalts. J Geophys Res 94:7685–7729 White RS, McKenzie D, O’Nions J (1992) Oceanic crustal thickness from seismic measurements and rare earth element inversion. J Geophys Res 97:19683–19715 Zelt CA, Smith RB (1992) Seismic traveltime inversion for 2-D crustal velocity structure. Geophys J Int 108:16–34 123