4D Arctic: A Glimpse into the Structure and Evolution of the Arctic in the Light of New Geophysical Maps, Plate Tectonics and Tomographic Models
- First Online:
- Cite this article as:
- Gaina, C., Medvedev, S., Torsvik, T.H. et al. Surv Geophys (2014) 35: 1095. doi:10.1007/s10712-013-9254-y
- 1.7k Downloads
Knowledge about the Arctic tectonic structure has changed in the last decade as a large number of new datasets have been collected and systematized. Here, we review the most updated, publicly available Circum-Arctic digital compilations of magnetic and gravity data together with new models of the Arctic’s crust. Available tomographic models have also been scrutinized and evaluated for their potential to reveal the deeper structure of the Arctic region. Although the age and opening mechanisms of the Amerasia Basin are still difficult to establish in detail, interpreted subducted slabs that reside in the High Arctic’s lower mantle point to one or two episodes of subduction that consumed crust of possibly Late Cretaceous–Jurassic age. The origin of major igneous activity during the Cretaceous in the central Arctic (the Alpha–Mendeleev Ridge) and in the proximity of rifted margins (the so-called High Arctic Large Igneous Province—HALIP) is still debated. Models of global plate circuits and the connection with the deep mantle are used here to re-evaluate a possible link between Arctic volcanism and mantle plumes.
KeywordsArctic region Magnetics Gravity Tomography Mantle plume Volcanism
Prompted by recent climate change and economic motives, the interest for studying the Arctic has exponentially increased in the last decade. A wealth of new geophysical and geological data has been collected, older classified data have been made publically available, and several international efforts have contributed to large, regional data set compilations. This data collection indicates a much more complicated structure of the Arctic crust than previously thought (e.g., Mosher et al. 2012). As our knowledge about detailed structures of this remote region is increasing, new concepts and integrated data sets have to be employed for modelling its tectonic evolution. In particular, mantle-lithosphere connections may hold some of the keys to deciphering the timing and mechanisms of opening and closing oceanic basins for the last 200 million years.
Despite several decades of research, a consensus towards a model to explain all the geological and geophysical observations has not yet been reached. There are at least half a dozen kinematic models proposed for the evolution of the Amerasia Basin (Fig. 1)—an enigmatic submerged area that occupies more than half of the High Arctic region. Even the more recent evolution of the Arctic is not satisfactorily unravelled, and although a Mid–Late Paleocene age for the oceanic Eurasia Basin has been postulated, the exact timing of break-up is not known, nor are the plate boundary configurations mainly because the Eocene collision in SW Ellesmere/North Greenland has modified these plate boundaries.
In this contribution, we aim to demonstrate how recent data sets and models may change our view on the tectonic evolution of the Arctic. In Sect. 1, we will succinctly present the new geophysical maps that are used for the interpretation of main tectonic features in the Arctic region. A model for the evolution of the oceanic crust in the Eurasia and Amerasia Basins is developed in Sect. 2, together with a review of Iceland plume path models and possible connections with volcanism in the High Arctic. Section 3 reviews several tomographic models and speculates on the origin of observed mantle heterogeneities.
2 New Regional Maps of the Arctic
List of gravity and magnetic digital gridded data for the Circum-Arctic region
ArcGP FA (Arctic Gravity Project)
Airborne, marine, ground and submarine data
5′ × 5′
CAMPGMP-GM FA/BA (Circum-Arctic Mapping Project-Gravity and Magnetic Anomaly maps)
Airborne, marine and surface data
10 × 10 km
EIGEN GL04C (Forste et al. 2008) for quality control of the long wavelengths
Gaina et al. (2011)
Satellite altimetry (ERS-2 and Envisat)
On land, the field have been augmented with the best available terrestrial gravity field complete global coverage
ARCS-2 FA (ARCtic Satellite-only)
Satellite altimetry (Envisat and ICEsat)
McAdoo et al. (2013)
World Digital Magnetic Anomaly Map (WDMAM)
Airborne, marine, surface and satellite data
5 km upward continued 5 km
Airborne, marine, surface and satellite data
2 km upward continued 1 km
Downward continued lithospheric magnetic field model MF6 derived from satellite data (Maus et al. 2009) was used as a regional reference surface
Gaina et al. (2011)
Airborne, marine, surface and satellite data
2 min upward continued 4 km
Interpolation between sparse track lines in the oceans was improved by directional gridding and extrapolation, based on an oceanic crustal age model. The longest wavelengths (>330 km) were replaced with the latest CHAMP satellite magnetic field model MF6
Maus et al. (2009)
Magnetic anomaly chrons dated by several age models and spreading rates used for synthetic model (Fig. 6)
Full spreading rate (km/Myr)
Horizontal resolution (km)
Maximum depth (km)
Ritsema et al. (1999)
Shapiro and Ritzwoller (2002)
Shapiro and Ritzwoller (2002)
Simmons et al. (2006)
Montelli et al. (2006)
Montelli et al. (2006)
S and surface waves
Lebedev and van der Hilst (2008)
Li et al. (2008)
Ritsema et al. (2011)
Jakovlev et al. (2012)
Schaeffer and Lebedev (2013)
To illustrate the distribution of short wavelength magnetic anomalies (which usually are produced by near-surface magnetized bodies), the CAMPGM magnetic anomaly map is also presented in a 3D view, draped on bathymetry and topography of the Arctic (IBCAO, Jakobsson et al. 2008) (Fig. 4d). The lithospheric magnetic field MF6—a model derived from satellite data and downward continued at the geoid height (Maus et al. 2009)—is draped on the Bouguer anomaly (BA) of the Circum-Arctic region. This combination of data sets aims to illustrate whether negative BA typical for areas of thick crust (mainly continental, cratonic area or large volcanic provinces) correlates with long wavelength structures of the magnetic field (which are mainly generated by deeper structures).
Some of the global and regional maps are regularly updated, especially with improved satellite data and any new available surface data (e.g., the DTU gravity anomaly map). In the case of magnetic anomaly compilations, international efforts for sharing data and to contribute to regional maps have proved extremely useful [like the WDMAM and CAMP-GM initiatives; (Korhonen et al. 2007; Maus et al. 2007; Gaina et al. 2011)]. Hopefully, the wealth of data that has been acquired under the Law of the Sea in the last couple of years will contribute to improve the existing geological and geophysical data sets of the Arctic region and will set the scene for more detailed “views” of the Arctic’s structure.
3 Oceanic Versus Continental Crust and Magmatic Activity in the Arctic Region
3.1 Amerasian Basin
3.1.1 Canada Basin
The most commonly accepted model for explaining the opening of the Canada Basin includes counterclockwise rotation (CCR) of Arctic Alaska away from the Canadian Arctic islands (Carey 1955). An alternative model was proposed by Koulakov et al. (2013) and presumes the clockwise rotation of central Arctic (Arctida Block) due to subduction of the oceanic lithosphere underneath the Anyui Suture. A more unconventional model postulated that this basin formed in North Pacific and was subsequently trapped in the Arctic region (Churkin and Trexler 1980).
The rotational model is based on paleomagnetic data (Halgedahl and Jarrard 1987), stratigraphic studies of the North Slope and Sverdrup Basin margins, and a fan-shaped magnetic pattern observed in the southern Canada Basin. Both stratigraphic studies (e.g., Embry 1990 and references therein) and magnetic anomaly patterns have been the subject of numerous discussions and reinterpretations. Lane (1997) and Grantz et al. (1998, 2011) proposed more complex models that include orthogonal or strike-slip motion, combined with a rotation in the later stages of opening.
3.1.2 Northwind Ridge and Chukchi Plateau
These submerged rifted plateaus are located north of the Chukchi Sea (Fig. 1) and are identified to be of continental nature. Data from cores collected from the southern Northwind Ridge show that Triassic and older strata were attached to both Arctic Canada and Alaska prior to the rifting that created the Canada Basin, but controversies exist in establishing the conjugate margins (for a review see Grantz et al. 2011). Younger sediments show that this continental sliver was isolated in the Early Jurassic (Grantz et al. 1998). Northwind Ridge was uplifted later (in the Paleocene), perhaps due to relative convergence with other adjacent tectonic blocks, while extension relative to the Chukchi Borderland created the Northwind Basin.
3.1.3 Alpha and Mendeleev (AMR)
The nature of both the Alpha and Mendeleev Ridges remains speculative because extensive volcanism has overprinted and complicated the original geophysical signatures. Maastrichtian fossils have been recovered (Clark et al. 1980), and sediments from the cores (Late Campanian to Early-Maastrichtian, 76–69 Ma, Davies et al. 2009) provide a minimum age constraint. Volcanic material, 82–72 Ma in age, has been described from the Alpha Ridge (Jokat 2003), and the similarity with the Iceland–Faeroes plateau led Vogt et al. (1979), Lawver et al. (1983) and Forsyth et al. (1986) to suggest an “hotspot-related” origin. Aiming to clarify the origin and structure of basement and sediment cover of Alpha and Mendeleev Ridges, a wealth of geophysical data and piston core samples has been acquired in the past few years. A continental-like intruded crust has been suggested by Kaminsky (2005), an oceanic origin by Jokat (2003), and undecided (rifted volcanic continental margin or oceanic plateaus formed at spreading centres) by Dove et al. (2010). Some authors (e.g., Filatova and Khain 2010) propose that these ridges might be the remnant parts of a larger continental plate which existed in the Arctic and was destroyed due to different tectonic episodes. Recent seismic studies have provided evidence for continental crust (about 32 km thick) underlying at least part of the Mendeleev Ridge (Lebedeva-Ivanova et al. 2006). Bruvoll et al. (2012) concluded that the upper volcanic carapace of Mendeleev and north-western Alpha Ridge were most likely emplaced during a brief igneous episode—no later than Campanian (80 Ma)—as part of the latest events of the Late Cretaceous Circum-Arctic volcanism.
New high-resolution magnetic anomaly data collected along the Alpha Ridge, situated near the southern Lomonosov Ridge and Greenland, revealed dyke swarm-like linear magnetic anomalies that resemble anomalies identified in Ellesmere Island and Franz Josef Land (Døssing et al. 2013). These authors concluded that part of the Alpha Ridge is highly attenuated continental crust formed by poly-phase break-up with LIP volcanic addition, but they also interpreted Barremian (or alternatively lower Valanginian–Barremian) seafloor spreading anomalies in the Makarov Basin region.
3.1.4 Lomonosov Ridge
The western boundary of the Eurasian margin, the Lomonosov Ridge, indisputably rifted away from the northern Barents Sea during the Paleocene as a narrow microcontinent (ca. 55 Ma, e.g., Srivastava 1985). Its continental nature was recognized from seismic data that showed horsts and grabens (e.g., Jokat et al. 1992), and it is the only Arctic location drilled under the IODP programme (IODP Expedition 302, 2006). Recent investigations have shown a more complex nature with detailed pull-apart basins in its central part, a volcanic plateau close to Greenland, and extended crust at its opposite end, close to the East Siberian Shelf (e.g., Lebedeva-Ivanova et al. 2006; Jackson et al. 2010).
3.1.5 Podvodnikov and Makarov Basins
The Podvodnikov Basin occupies the space between Mendeleev Ridge, the East Siberian Shelf and the Lomonosov Ridge. The Makarov Basin is much deeper, with its northward continuation bordered only by the Lomonosov and Alpha Ridges. Seismic reflection and refraction studies reveal a thicker oceanic crust for the Makarov Basin (~22 km) and part of the Podvodnikov Basin (~20 km) (Sorokin et al. 1999; Lebedeva-Ivanova et al. 2006). Regarding the formation of these basins, one end-member hypothesis considers the Makarov Basin as the continuation of the Canada Basin, thus implying a similar age and structure. Recently, Døssing et al. (2013) interpreted N–S magnetic lineations as chrons M16n–11An.1n (Early Valanginian–Late Hauteverian, 138–132 Ma) and chrons M10n/9n–4n (Barremian, ca. 129–126 Ma). Alternatively, a Late Cretaceous–Early Tertiary age has been postulated. Weber (1990) suggested an age between 118 and 56 Ma, and Mickey (1998) an age of 95 to 67 Ma.
These uncertainties led Alvey et al. (2008) to examine three new plate tectonic scenarios for the age and opening of the Makarov and Podvodnikov Basins, in which they assigned an Early- to Mid-Cretaceous age to the Canada Basin and varied the age and the crustal nature between (and within) the Alpha Ridge and Mendeleev and Lomonosov Ridges. This range of oceanic age distributions has been used to determine the crustal thickness of these basins, and the results were compared with the estimates from refraction data modelling. They concluded that the Podvodnikov Basin is probably Late Cretaceous in age. Our interpretation is that these basins, whether they are floored by oceanic or extended continental crust, are the result of the Late Cretaceous–Cenozoic extension between the North America and Eurasia, as predicted by the regional model of Gaina et al. (2002). New seismic profiles acquired across the Lomonosov Ridge and adjacent basins interpreted by Langinen et al. (2009) have confirmed that the Marvin Spur is a sliver of continental crust (as also suggested by Cochran et al. 2006) and that part of the Makarov Basin probably formed in the Early Tertiary on thinned continental crust.
3.2 Eurasian Basin
Well-preserved linear magnetic anomalies (isochrons) that are relatively easy to identify have allowed a straightforward interpretation of the Eurasian Basin (Figs. 4, 6b). Most authors have identified chron 24 (ca. 54 Ma) as the oldest magnetic isochron, spawned by seafloor spreading between the Lomonosov Ridge and the Eurasian margin (Srivastava 1985; Gaina et al. 2002; Glebovsky et al. 2006). Other studies have identified an abandoned extinct ridge (ca. 55 Ma) in the proximity of Lomonosov Ridge (Brozena et al. 2003). If correct, this structure implies that the opening of the Eurasian Basin was linked to the evolution of Baffin Bay and the Labrador Sea, but the restoration of this plate boundary is made difficult by the subsequent Eurekan compressional phase. New studies suggest that the previously interpreted chron 25 (ca. 56 Ma) may represent serpentinized exhumed mantle formed before break-up together with the highly thinned continental crust that can be observed only on the Lomonosov Ridge conjugate margin (Minakov et al. 2012). The age of the oldest oceanic crust in our Eurasia Basin model (Fig. 6) is taken as 56 Ma (C25).
3.3 Cretaceous to Present Magmatic Activity in the Arctic
Scattered magmatic areas have been identified in the Arctic along the continental margins (North Greenland, East Siberian Islands, Svalbard, Franz Josef Land, e.g., Maher 2001; Buchan and Ernst 2006; Kuzmichev et al. 2009; Tegner et al. 2011; Corfu et al. 2013; Nejbert et al. 2011) on submerged microcontinents (Chukchi Plateau) and submarine plateaus of unknown or controversial nature (Alpha and Mendeleev Ridges, Yermak and Morris Jesup Plateaus, e.g., Jokat et al. 1995 and 2008) and within the extended continental crust (e.g., Barents Sea, Minakov et al. 2012). Volcanism spanning from about 130 to 60 Ma probably occurred during at least two phases—an initial tholeiitic phase (130–80 Ma) and a second alkaline phase (85–60 Ma) (for a review see Tegner et al. 2011) and is labelled as the High Arctic Large Igneous Province (HALIP) in the literature (see Buchan and Ernst (2006) for a review). The tholeiitic phase was probably the result of mantle plume activity.
Assuming that the Iceland plume was active since the Cretaceous, its restored Early Cretaceous position is located in the Arctic region, as suggested in the early 1980s by Morgan (1983). The Lawver and Müller (1994) classical paper showed that the computed Iceland plume position at 130 Ma assuming a fixed hotspot reference frame (Müller et al. 1993) was very close to the presumably massive Early Cretaceous volcanism on the Alpha Ridge, as also suggested by Forsyth et al. (1986).
Compared to the fixed hotspot model (Lawver and Müller 1994), the combined moving hotspot- and paleomagnetic-based models (Torsvik et al. 2008 and Doubrovine et al. 2012) predict the Early–Mid-Cretaceous position for the Iceland hotspot in the western and central Ellesmere Islands, respectively, more than 1,000 km away from the previously inferred location of a fixed mantle plume (Fig. 8). The new predicted hotspot tracks are farther away from the Alpha and Mendeleev Ridges, but closer to the Axel-Heiberg Island dykes dated by the 40Ar–39Ar method as 128.2 ± 2.1 Ma (Villeneuve and Williamson 2003).
Possible links to other HALIP volcanic regions can be better visualized on maps showing the position of tectonic blocks in an absolute reference frame, and we have computed plate reconstruction models for 130, 110, 90 and 61 Ma that may represent the more intense magmatic periods reflected by the ages of volcanism in Greenland, Ellesmere Islands, Svalbard, Franz Josef Land and possibly New Siberian Island, Chukchi, and Alpha–Mendeleev Ridges (Fig. 7). In addition, the so-called Plume Generation Zones (PGZs)—regions at the core–mantle boundary where plumes initiate and eventually generate large igneous provinces and surface hotspot volcanism through time (e.g., Torsvik et al. 2006)—is also shown on these reconstructions. Arctic magmatic provinces may be related and generated from a PGZ from ca. 130 to 100 Ma, as North Greenland, Ellesemere Islands, Svalbard, Franz Josef Land and possibly Alpha Ridge were in the proximity of its northernmost prolongation, but ~115–80 Ma magmatism on the Chukchi Plateau, New Siberian Island and Mendeleev Ridges is probably not related to it, unless it can be shown that the PGZ-generated volcanism extends on very large areas (with a radius of more than ca. 1,000 km). In Paleocene time, the northern PGZ coincides well with the NAIP volcanism, including the Disko Island (~62 Ma, Storey et al. 1998) magmatism that cannot be well explained by the new Iceland track locations based on the moving hotspot absolute reference frames of Torsvik et al. 2008 and Doubrovine et al. 2012 (Figs. 7, 8).
4 Circum-Arctic Mantle Imaged by Tomography Models
The structure of the mantle in the Arctic region has rarely been discussed in studies dedicated to global tomographic models, either because of the lack of proper seismic ray path coverage or because of a lack of interest in this important area. Several studies present adjacent areas like the North Pacific (e.g., Gorbatov et al. 2002) and North Atlantic (e.g., Rickers et al. 2013) where subduction processes and mantle plumes may have generated interesting structures prone to be imaged by seismology. Exception to this are studies of Levshin et al. (2001) and Jakovlev et al. (2012). Fortunately, in the last decade, due to the availability of much improved global earthquake catalogue data that can resolve the northern higher-latitude area, increased resolution global and regional models are now able to shed light on the upper and lower structures of the mantle.
In contrast, the two models seem to differ in imaging the crust of the SW Barents Sea, with the SL2013sv model showing large velocities (standing presumably for thick crust) and the IPGG12 model showing small velocities (Fig. 9—area indicated by white circles). Conversely, the region of northern Greenland has small velocities in the SL2013sv model and large velocities in the IPGG12 model (Fig. 9—area indicated by grey circles). Both these regions are at the boundary between thick cratonic crust and thinner crust of sedimentary basins and, at the same time, closer to active mid-ocean ridges; therefore, more information from improved tomographic and crustal studies will help understand the plate boundary evolution processes. It should be noted that the IPGG12 model provides the P-velocity anomalies, whereas SL2013sv images the S-anomalies. Also, at 50 km depth, the ray paths of body waves used in the IPGG12 model in most areas do not form a sufficient intersection system; thus, the result might represent averaged values of crustal and mantle anomalies down to 100 km depth. The differences observed in Fig. 9 can be partly explained by these two reasons.
Upper mantle imaging strongly depends on the crustal model used, and the interpretation of anomalies is hampered by the complexity of the mantle structure, temperature, chemistry and anisotropy (e.g., Bastow 2012). Therefore, besides presenting the available models for the Arctic region upper mantle, we will restrain from further interpretation of these models.
The lower mantle is probably more robustly imaged, and we may use these models to decipher the earlier evolution (before the development of preserved oceanic basins) of the Arctic region. We have selected a range of global models aiming to illustrate how additional data and techniques have improved the quality of the High Arctic lower mantle images. More publicly available global models were scrutinized, but some of them presented seriously “smeared” structures and were not selected here. Vertical cross-sections along five profiles that run from the North American–Greenland margins to the Eurasian margin were extracted from six tomographic models (Table 3).
All cross-sections remarkably show three groups of positive mantle anomalies: one composed of a sub-horizontal or slightly southward-dipping feature, around 1,000 km depth observed on profiles A and B (I in Fig. 11), another sub-horizontal anomaly at around 1,500 km is seen on profiles A, B and C (II in Fig. 11), and another steeper, dipping northward segmented feature starts below 1,000 km and is reaching the core–mantle boundary (CMB) and is seen on profiles B to E (III in Fig. 11). 3D images of isosurfaces showing these positive anomalies (possibly representing cold subducted material) are also shown in Fig. 11 to better illustrate the extent of the “slab graveyard” under the High Arctic region.
We attempt to interpret the nature and age of these anomalies by (1) applying an age–depth relationship and (2) analysing kinematic models of the Arctic. Several studies are now suggesting that the depth of positive anomalies observed in the mantle tomographic models could be roughly correlated with the age of assumed subducted oceanic slabs. This link was established by a series of observations and models, and several simple assumptions have to be imposed in order to achieve a very nascent linear relationship. We use two models that suggest depth–age relationship for interpreted subducted slabs: van der Meer et al. (2010) and Steinberger et al. (2012) (Fig. 11). According to these models, the three groups of sinking slabs may have ages between 85 and 95 (55) for group I, around 130 (80) for group II, and between 100 and 200 (75–200) for group III (ages according to Steinberger et al. (2012) are in italic) (Fig. 11).
As Kuzmichev (2009) suggests, the South Anyui subduction zone was a continuation of the trench situated south of the Arctic Alaska and associated accreted terranes (that latter became the Angaucham Suture, see Nokleberg et al. 1998). His “bipolar opening” of the Arctic Ocean as a back-arc basin driven by a northerly directed subduction zone seems to agree with the shallower slab remnants imaged by seismic tomography under the Arctic region (Figs. 11, 12), which probably resulted from the episodic northward subduction events that have been partially terminated and relocated by obductions. However, the exact geometry of the Mesozoic trenches south of Chukotka and Arctic Alaska remains to be clarified. A southward-dipping subduction might have existed as well and would have been a more efficient mechanism to explain continental ribbon detachment (by slab pull) and subsequent collision with the Siberian Craton (in the same manner as the Paleotethys closure formed the Neotethys). This model is supported by volcanism south of the South Anyui Suture (on Anyui-Svyatoinos arc, see Zonenshain et al. 1990) and implies the existence of a set of trenches with reversed polarities, a configuration found today in the SW Pacific.
5 Summary and Concluding Remarks
The Eurasian Basin is undoubtedly floored by oceanic crust. Based on potential field data, the margins seem to have experienced a short-lived break-up, but transitional-type crust (with serpentinite ridges, e.g., Minakov et al. 2012) or an older-than-C24 opening (Brozena et al. 2003) are possible interpretations for the Eurasian flank of the Lomonosov Ridge. Although a long-lived mega strike-slip boundary along the entire Nares Strait length is now discounted, a possible connection with the Baffin Bay/Labrador Sea is not excluded as suggested by large-scale plate kinematics.
Shortly before the opening of the Eurasia Basin, extension is predicted by large-scale plate tectonic models (e.g., Gaina et al. 2002), and this could have triggered continental splinters from the Amerasian side of the Lomonosov Ridge to be detached from the ridge and form small basins, floored by either extended continental crust or even oceanic crust (for example, Makarov and possibly part of the Podvodnikov Basin).
Pseudo-linear magnetic anomalies, a sinuous feature that can be interpreted from potential field data as an extinct ridge, and a thin crust predicted by the gravity anomaly inversion are consistent with an oceanic type crust in the Canada Basin. Although a definitive consensus toward the mechanism of this basin formation is not reached, a “rotational” model is partially supported. A “bipolar” opening due to southward and northward-directed subduction along the South Anyui and Angaucham trenches could provide a viable mechanism for detaching several continental blocks and to form the Amerasian Basin. Global tomography models image the slab graveyard below the South Anyui Suture and its surface expression, but more sophisticated methods are required to extract all the information from these data and models.
The timing and the extent of High Arctic Large Igneous Province (HALIP) is not well matched by paleopositions of the Iceland plume. However, part of the HALIP was formed at times when the Plume Generation Zone (Torsvik et al. 2006) was located beneath the North Atlantic and High Arctic regions.
Numerous questions regarding the configuration of High Arctic remain unanswered. Among them are the nature of the Alpha Ridge and timing and extent of volcanism and a proper understanding of the terrane amalgamation around the High Arctic region (NE Asia, North Pacific and NW America). In this study, we have used only the freely available data and published results but, in the near future, the access to the wealth of data collected in the Arctic in the last decade as well as new data collected for “Law of the Sea” exploration will unquestionably shed more light upon the unresolved issues.
Two anonymous reviewers and the Guest Editor are thanked for their useful suggestions. We thank A. Schaeffer and S. Lebedev for a preprint of their SL2013sv paper and model. G. Laske and co-authors are also thanked for an early distribution of CRUST1.0 model. CG, THT and SCW acknowledge financial support from the European Research Council under the European Union’s Seventh Framework Programme (FP7/2007–2013)/ERC Advanced Grant Agreement Number 267631 (Beyond Plate Tectonics). CG also acknowledges financial support from the Norwegian Research Council (225027/H30 4DARCTIC). Financial support for SM was provided by Det norske oljeselskap. IK is supported by Russia-Norwegian Project RFBR No. 12-05-93085 Norv_a.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.