We use seismic reflection data acquired by an over-winter expedition on drifting sea ice in the central Arctic Ocean to explore a possible spatial and temporal magmatic relation between the sub-bottom geology of part of the deep Arctic Ocean and the Mesozoic volcanic rocks found on the islands and the bordering continental shelf of Franz Josef Land and Svalbard. The new dataset from the North American segment (85–90°N) of the Lomonosov Ridge, central Arctic Ocean documents several Mesozoic volcanic pulses over a distance of c. 600 km along the ridge. This volcanism borders a domain of high magnetic field intensity over the adjacent Alpha Ridge in the deep basin where the magnetic source rocks and recent seismic reflection data indicate extensive Mesozoic magmatism. We suggest that the Mesozoic volcanism on the Lomonosov Ridge in its palaeo-position at the former continental margin north of Franz Josef Land and Svalbard spatially links the Mesozoic magmatic pulses of the continental High Arctic Large Igneous Province (HALIP) of polar Europe to volcanism on the adjacent Alpha Ridge in the deep Arctic Ocean. Increased input of heat to the upper crust on the Lomonosov Ridge enhanced maturation of hydrocarbon source rocks as manifested by the presence of gas or fluid escape pipes restricted to the area of volcanism.

Volcanic rocks, sills and dykes of Mesozoic age exposed by erosion on islands on the polar continental margin from the De Long islands on the Siberian continental shelf via Franz Josef Land and Svalbard to the Canadian Arctic (Fig. 1) are collectively known as the High Arctic Large Igneous Province (HALIP) (Tarduno et al. 1998; Maher 2001; Ernst and Bleeker 2010). The exposed HALIP rocks frame a distant area of the deep Mesozoic part of the Arctic Ocean (north of 78°N) characterized by high-amplitude magnetic field intensity (Fig. 1), a geophysical expression of massive basaltic magmatism (Piskarev 2004; Lebedeva-Ivanova et al. 2006; Vogt et al. 2006; Grantz et al. 2011; Saltus et al. 2011; Døssing et al. 2013; Estrada et al. 2016; Oakey and Saltus 2016). The main arguments for a spatial and temporal link between these two areas of Mesozoic magmatism are so far based on (1) an interpreted outlier of the high-amplitude magnetic field intensity zone from the Alpha Ridge (Fig. 1) onto an area of northern Ellesmere Island where HALIP-related magmatism is represented by intrusive and extrusive magnetic source bodies within the metamorphic Pearya Terrane (Funck et al. 2011; Estrada et al. 2016; Oakey and Saltus 2016) and (2) observations of the trends of dyke swarms and their interpreted offshore continuation into the deep basin (Embry and Osadetz 1988; Buchan and Ernst 2006; Jackson et al. 2010; Døssing et al. 2013; Estrada et al. 2016; Oakey and Saltus 2016). Dykes represent magnetic source bodies and their offshore continuation may be interpreted from linear short-wavelength magnetic field intensity trends projecting NW from Franz Josef Land and NE from the Canadian Arctic Islands converging towards a proposed plume centre at the southern end of the Alpha Ridge. These dykes suggest contemporary magmatism over a large area.

The temporal variation in intensity of HALIP magmatism on the islands in the High Arctic is manifested as one or more pulses within each subregion (Smith et al. 1976; Embry and Osadetz 1988; Dibner 1998; Tarduno et al. 1998; Maher 2001; Nejbert et al. 2011; Shipilov and Karyakin 2011; Corfu et al. 2013; Senger et al. 2014; Evenchick et al. 2015; Polteau et al. 2016; Dockman et al. 2018; Naber et al. 2020). Although only a limited number of basalt samples are available from the Alpha and Mendeleev ridges of the deep polar basin, all show ages in the range 127–90 Ma that fall within the HALIP time frame (Mudie et al. 1986; Jokat 2003; Morozov et al. 2013; Brumley 2014; Williamson et al. 2019; Mukasa et al. 2020).

To further explore the extent of Mesozoic magmatism in the High Arctic, we recognize the Lomonosov Ridge as bordering the area of the high magnetic field intensity variations in the Amerasia Basin over a distance of c. 1500 km (Fig. 1). The North American segment of the ridge was positioned within the realm of HALIP magmatism as an integral part of the pre-Cenozoic continental margin north of Franz Josef Land (Dibner 1998; Brozena et al. 2003; Glebovsky et al. 2006; Shipilov and Karyakin 2011; Minakov et al. 2017). Isolated seismic reflection evidence of volcanic rocks on the Lomonosov Ridge near the North Pole has been reported by Kristoffersen (2000) and on the southern end towards Canada by Kristoffersen and Mikkelsen (2006). Evidence of volcanism on the Central and the Siberian segments of the Lomonosov Ridge is not known (Jokat et al. 1992; Jokat 2003; Rekant and Gusev 2012; Sauermilch et al. 2018; Weigelt et al. 2020). However, in the adjacent deep Makarov and Podvodnikov basins (Fig. 1), deep low-frequency reflection segments within acoustic basement along with several basement features are interpreted as basalt flows and seamounts or volcanic centres (Nikishin et al. 2021b, 2023). More recently, Funck et al. (2022) and Funck and Shimeld (2023) presented seismic reflection and refraction information along a transect from the eastern flank of the Alpha Ridge across the Makarov Basin and the Lomonosov Ridge north of 88°N (Fig. 1). The data suggest an extensive package of volcanic material (v = 4.8–5.1 km s−1) below the slope of the Lomonosov Ridge facing the Makarov Basin and a 1–2 km thick layer of similar velocity extending from the eastern slope of the Alpha Ridge across the Makarov Basin onto the Marvin Spur. This material was interpreted as HALIP-related magmatism.

We present new seismic stratigraphic evidence for distinct magmatic pulses within a protracted period of late Mesozoic volcanism along the North American segment of the Lomonosov Ridge. From this, we hypothesize a direct link between the terrestrial exposures of HALIP pulses along the polar continental margin and magmatism on the adjacent Alpha Ridge in the deep polar basin.

Outline of the areal extent and intensity of HALIP magmatism

The terrestrial exposures of Mesozoic magmatism in the High Arctic are relatively small except for islands in the Franz Josef Land archipelago and in the Canadian Arctic (Fig. 1). The onshore area affected by magmatism has been extended by aeromagnetic and seismic data to include part of the continental shelf west of De Long Island (Nikishin et al. 2021b) and a major part of the northern Barents shelf (Grogan et al. 2000; Shipilov and Karyakin 2011; Polteau et al. 2016; Minakov et al. 2017; Nikishin et al. 2021a, b), and extends the coverage in the Sverdrup Basin (Buchan and Ernst 2006; Evenchick et al. 2015). The estimated volume of magma in separate sub-areas (Franz Josef Land–northern Barents Sea, Svalbard and the Canadian Arctic islands) exceeds the adapted definition of a large igneous province (Maher 2001; Ernst 2014; Senger et al. 2014; Polteau et al. 2016; Saumur et al. 2016).

The Mesozoic section of volcanic rocks on Bennett Island in the De Long archipelago on the East Siberian continental shelf (Fig. 1) is 250–350 m thick and includes up to 20 basalt flows resting on a thin (c. 20 m thick) Early Cretaceous sandstone and carbonaceous mudstone (Kos'ko and Trufanov 2002; Drachev and Saunders 2003; Kos'ko and Korago 2009; Drachev et al. 2010; Tegner and Pease 2014). Basalt flows are also present on the nearby Zhokhov and Vel'kitsii islands (Silantyev et al. 2004) and geophysical data suggest that the volcanic rocks exposed on the islands are part of a larger (about 350 km long and 150 km wide) offshore area of Mesozoic volcanic rocks on the surrounding De Long High (Drachev et al. 2018; Nikishin et al. 2021b). About 200 km to the WSW of Bennett Island is Kotel'nyi Island, with an Aptian–Albian sedimentary section that includes lenses of acidic tuff capped by a rhyolite sheet (Kos'ko and Trufanov 2002). Mesozoic volcanic rocks are also inferred from seismic data to floor the Anisin Rift Basin about 150 km north of Kotel'nyi Island (Nikishin et al. 2021b).

Mesozoic magmatism in the Franz Josef Land archipelago (Fig. 1) and the adjacent northern Barents Sea is dominated by sills and dykes (Dibner 1998; Polteau et al. 2016; Shipilov 2016; Minakov et al. 2017; Karyakin et al. 2021). The sills exposed in the archipelago vary in thickness from 20–30 m to 100 m. Dykes with a NW trend are present across the islands and the surrounding shelf area, but are particularly abundant on the large islands in the southeastern part of the archipelago (Shipilov 2016). The extrusive magmatic activity in Franz Josef Land is represented by basalt flows with intercalated terrestrial sediments and an apparent thickness of more than 300–380 m (Embry 1992; Dibner 1998). Individual flows are 10–71 m thick and one volcanic edifice <200 m high is present on Salisbury Island in the central part of the archipelago.

Sills and dykes also dominate the Mesozoic magmatism (Fig. 1) in the Svalbard area (Grogan et al. 2000; Maher 2001; Nejbert et al. 2011; Senger et al. 2014). Individual sills may be up to 50 m thick whereas dykes (10–15 m thick) are less common (Burov et al. 1976; Maher 2001). Lava flows are known only from the islands of Kong Karls Land in the east, where at least six flow episodes are recognized (Smith et al. 1976; Olaussen et al. 2019). Basalt flow thicknesses range from 5 to 43 m.

The c. 8 km wide outcrop of the Kap Washington volcanic group extends for 45 km along the north coast of Greenland (Fig. 1). The sequences of lavas and tuffs with an estimated thickness >5 km are overridden from the south by early Paleozoic metamorphic rocks (Dawes 1971; Soper et al. 1982; Brown et al. 1987; Tegnér et al. 2011). Swarms of vertical alkaline dykes up to 25 m in thickness are abundant out to about 50 km to the west of the Kap Washington volcanic rocks and within 150 km to the south and east (Higgins et al. 1981; Soper et al. 1982; Thórarinsson et al. 2015, and references therein). The swarms can be grouped in north–south, NW–SE and east–west trends.

In the Canadian Arctic archipelago, Mesozoic magmatic events are dominantly sills and dykes (Blackadar 1964; Osadetz and Moore 1988; Williamson 1988; Buchan and Ernst 2006, 2018) and estimated to be 3–5 times the volume of extrusive rocks (Saumur et al. 2016). The sills and dykes, commonly diabasic rocks, occur over an area about twice as large to the south and west of the extrusive rocks (Evenchick et al. 2015). The main depocentres of subaerial flood basalts and pyroclastic rocks are on Axel Heiberg Island and the northwestern Ellesmere Island (Embry and Osadetz 1988; Evenchick et al. 2015; Estrada et al. 2016). The first volcanic rocks appear as a few c. 10 m thick lava flows in the lower part of the early Late Valangian to late Aptian, 600–1600 m thick siliclastic Isachsen Formation (Embry and Osadetz 1988). The age of the volcanic rocks is late Hauterivian–early Barremian (c. 130 Ma). Higher in the Isachsen Formation is the Aptian (130–113 Ma), c. 300 m thick Walker Island Member, where basalt flows interbedded with quartz sandstone and pyroclastic sediments make up 220 m of the section. Individual flows are 5–30 m thick (Embry and Osadetz 1988). Subsequent extrusive volcanism is represented by the upper Albian Strand Fiord Formation, which includes a more than 800 m thick succession of tholeiitic subaerial basalt flows, dykes, sills and pyroclastic deposits with minor interbedded lacustrine sediments (Ricketts et al. 1985; Dostal and MacRae 2018). The main depocentre is in the western part of Axel Heiberg Island and individual flows range in thickness from 6 to 60 m. The youngest massive volcanism is represented by the Hansen Point Volcanics, a c. 1000 m thick section of lavas and pyroclastic deposits with intercalated sediments preserved in fault-bounded outcrops on the coast of northwestern Ellesmere Island (Trettin and Parrish 1987; Embry and Osadetz 1988; Estrada and Henjes-Kunst 2013). This volcanic succession is made up of chemically evolved rhyodacite, trachy-andesite and rhyolitic rocks.

The products of Mesozoic magmatism in the High Arctic have compositions that group into a tholeiite suite and a suite of more alkaline volcanic rocks (Tegnér et al. 2011). Predominantly tholeiitic volcanic lavas, dykes and sills were emplaced in the De Long Islands, the Franz Josef Land–Svalbard area and the Canadian Arctic during the time frame 130–80 Ma, whereas rocks of the mildly alkaline suite are younger (85–60 Ma) and restricted to northern Ellesmere Island and northern Greenland (Bailey and Brooks 1988; Amundsen et al. 1998; Dibner 1998; Ntaflos and Richter 2003; Drachev and Saunders 2006; Tegnér et al. 2011; Jowitt et al. 2014; Tegner and Pease 2014; Senger et al. 2014; Estrada 2015; Evenchick et al. 2015; Dostal and MacRae 2018). Both rock suites include varying degrees of crustal contamination (Amundsen et al. 1998; Ntaflos and Richter 2003; Bédard et al. 2021a, b). The tholeiitic suite indicates a higher degree of partial melt at shallow depth whereas the composition of the mildly alkaline rocks implies a lower degree of partial melt at deeper levels (Jowitt et al. 2014). The tholeiitic rocks are flood basalts suggested to be associated with rifting and opening of the Canada Basin (Weigand and Testa 1982; Worsely 1986; Embry and Osadetz 1988; Amundsen et al. 1998; Grogan et al. 2000; Maher 2001; Drachev and Saunders 2006; Ernst and Bleeker 2010; Corfu et al. 2013; Estrada 2015; Estrada et al. 2016). The younger alkaline magmatism present in northern Ellesmere Island and northern Greenland appears to be spatially related to magmatism at the junction of three lithospheric plates during the Late Cretaceous–Early Cenozoic opening of the Eurasia Basin and therefore unrelated to the main phase of HALIP (Tegnér et al. 2011).

HALIP magmatism through time

Mesozoic magmatism in the High Arctic appears long lived (130–60 Ma) with activity clustered within periods of a few million years (Fig. 1). In the case of Franz Josef Land, combined stratigraphic and radiometric evidence documents older magmatism during a brief period at c. 180 Ma and another (c. 157 Ma) determined by 40Ar/39Ar geochronology (Karyakin et al. 2021; Karyakin and Aleksandrova 2022). We briefly summarize the available radiometric age information as follows.

De Long Islands

The age of the volcanism on Bennett Island as determined by radiometric K–Ar dates (130–100 Ma) falls within the Hauterivian to early Albian (Fedorov et al. 2005; Drachev and Saunders 2006). The maximum age is supported by the presence of spores in the underlying coal-bearing unit (Kos'ko and Trufanov 2002) and a possible later peak between 115 and 110 Ma was suggested by Drachev and Saunders (2006).

Franz Josef Land

Attempts to decipher a time-line for the Mesozoic magmatism on Franz Josef Land based on 40Ar/39Ar geochronology showed ages ranging from Early Jurassic to Early Cretaceous (Karyakin and Shipilov 2009) and an Early Jurassic event has recently been corroborated by stratigraphic evidence (Karyakin and Aleksandrova 2022). The relative probability age distribution is centred around 130 Ma (Abashev et al. 2020). However, basalts from 50 sites appear to have normal remanent magnetization (Abashev et al. 2020), which suggests cooling through the Curie temperature within the geomagnetic normal polarity Superchron 121.4–83.65 Ma (Ogg 2020; Zhang et al. 2021). New thermal ionization mass spectrometry (TIMS) U–Pb measurements on sills from two sites on Franz Josef Land also give ages of 122–123 Ma suggesting short-lived volcanic activity (Corfu et al. 2013). However, the stratigraphic evidence compiled by Embry (1992) indicates four volcanic episodes between the Valangian–Hauterivian (c. 135 Ma) and early Albian (c. 110 Ma).


The ages of intrusive rocks in Svalbard show peaks at 115, 100 and 78 Ma (Nejbert et al. 2011), but the six new TIMS U–Pb ages available fall between 120 and 125 Ma (Polteau et al. 2016). Lava flows on Kong Karls Land are present within Barremian (131–126 Ma) sediments (Smith et al. 1976; Grogan et al. 2000).

Sverdrup Basin

Four Mesozoic magmatic pulses and associated transgressive–regressive cycles have been constrained by biostratigraphy (Embry and Osadetz 1988) and further refined by radiometric dating. Initial deposition of volcanoclastic rocks in the Isachsen Formation during the Hauterivian (c. 130 Ma) was followed by three intervals of sill emplacement between 130 and 120 Ma (Evenchick et al. 2015; Dockman et al. 2018). The major pulse of volcanism represented by continental flood basalts of the Strand Fiord volcanic series on Axel Heiberg Island took place between c. 105 and 95 Ma (Tarduno et al. 1998; Villeneuve and Williamson 2003; Davis et al. 2017) followed by the Hansen Point Tholeiite Suite emplaced at 97–93 Ma and the Audhild Bay Alkaline Suite at 83–73 Ma on the north coast of Ellesmere Island (Estrada et al. 2001; Estrada and Henjes-Kunst 2013; Estrada et al. 2016; Dockman et al. 2018; Dostal and MacRae 2018; Naber et al. 2020).

North Greenland

The alkali dyke swarms in northern Greenland appear to be nearly coeval. U–Pb ages for the north–south and east–west dykes are c. 81 Ma (Thórarinsson et al. 2015) and 40Ar/39Ar ages for the NW–SE swarm are c. 85 Ma (Kontak et al. 2001). The c. 5 km thick section of the Kap Washington Volcanic Group at the northern coastline was emplaced during a subsequent alkaline magmatic phase between 71 Ma (Tegnér et al. 2011; Thorarinsson et al. 2011) and 64 Ma (Larsen 1982; Estrada et al. 2001).

In summary, Mesozoic magmatism in the High Arctic is within the definition of a large igneous province (<50 myr, Bryan and Ernst 2008) if the life span is associated with emplacement of the tholeiite suite of rocks (130–80 Ma) as suggested by Tegnér et al. (2011). However, the documented Jurassic periods of volcanism on Franz Josef Land suggest a more complex history (Karyakin et al. 2021; Karyakin and Aleksandrova 2022). More robust radiometric dates suggest shorter periods of high magmatic intensity as well as overlap in timing between the different areas (Corfu et al. 2013; Estrada and Henjes-Kunst 2013; Evenchick et al. 2015; Polteau et al. 2016; Dockman et al. 2018; Abashev et al. 2020; Naber et al. 2020).

Seismic data acquisition

Seismic data acquisition across the North American segment of the Lomonosov Ridge between the North Pole and Canada–Greenland was first achieved from camps on drifting sea ice: the LOREX experiment in 1979 (Weber 1980), the Soviet ice drift station North Pole-28 (Gramberg et al. 1991; Langinen et al. 2009) and the Fram-2014/15 ice drift (Kristoffersen et al. 2016). In particular, North Pole-28 crossed the Lomonosov Ridge three times in 1985 (Fig. 2, fine white line). Blasting caps were fired at 500 m interval and the seismic signals recorded on analogue tape and later digitized (Langinen et al. 2009). The area between the North Pole and Canada–Greenland is within the realm of old ice flowing out of the Arctic Ocean and the sea ice is thickened by closely spaced (3–4 km−1) ice ridges (Wadhams 1981; Rigor et al. 2002; Rothrock et al. 2008). Penetration by ship is critically dependent on the temporal ice pressure even for two icebreakers in tandem (Jokat 1999, 2003; Marcussen and the LOMROG III Scientific Party 2012; Mosher et al. 2013). An unprecedented incentive for seismic surveys including the northern part of the Lomonosov Ridge was created by the geological documentation required by the circum-arctic nations for extension of the respective legal continental shelves. The preferred strategy was two icebreakers in tandem with the trailing vessel towing a 100–200 m long streamer recording the signal from seismic source arrays with a volume of 8–18 l (Marcussen and the LOMROG III Scientific Party 2012; Mosher et al. 2013; Nikishin et al. 2021a; Funck et al. 2022). The exception was the Arktika-2014 survey, which used a 600 m long cable and seismic source volumes in the range 17–34 l (Nikishin et al. 2021a). The Canadian effort included several multichannel seismic transects across the Lomonosov Ridge (Fig. 2, LSS 1601 and 1603) between 90 and 130°W (Evangelatos and Mosher 2016; Funck et al. 2022) and the Danish–Greenland effort a 10 km long multichannel seismic section across the ridge perimeter at 89°N (Marcussen and the LOMROG III Scientific Party 2012; Knudsen et al. 2017). Several seismic transects (not shown) have been made across the narrow part of the ridge in the sector 170°E–170°W near the North Pole (Kristoffersen 2000; Lebedeva-Ivanova 2010; Nikishin et al. 2021b; Funck et al. 2022).

The data reported here were acquired from ice drift station Fram-2014/15 (Fig. 3) deployed in early September 2014 in the Makarov Basin from the icebreaker Polarstern (Kristoffersen et al. 2016). A favourable sequence of large-scale atmospheric circulation cells from the end of October 2014 to early March 2015 drove the sea ice cover in a see-saw pattern across the Lomonosov Ridge at an average rate of 4.5 km day−1 (Figs 1 and 2). The resulting trajectory represented three-and-one-half crossings of the microcontinent between the North Pole and Canada–Greenland. The platform was a Griffon TD-2000 hovercraft (Fig. 3) with a payload capacity of 2.2 tons, equipped as a scaled-down research vessel (Kristoffersen and Hall 2014). Its mobility was a major advantage relative to traditional temporary ice camps, as the ice dynamics during the polar night forced four site relocations including two events with complete destruction of the camp environment (Kristoffersen et al. 2016). A 0.3 l airgun source was suspended at 5 m depth below the ice and powered by a diver compressor. Whereas the seawater temperature is a constant −1.7°C, the ambient air temperatures during the winter ranged between −30 and −43°C. The seismic wave field generated at 25 m shot intervals was received by a single hydrophone just below the ice–water interface to avoid cable strumming, sampled at 2 ms sample rate and logged in SEGY format. An ocean covered by a sea ice lid offers exceptionally low ambient noise levels (Buck and Greene 1964; Buck 1968; Makris and Dyer 1986), which allowed up to 2 km of sub-bottom penetration in 3 km water depth. A global positioning system controlled acquisition procedure made it possible for a single person to manage a 24/7 seismic acquisition campaign for over 7 months. In a later experiment, we have recorded the downgoing source signal from the Bolt PAR 0.3 l airgun by a Magseis™ seismic node suspended at 700 m water depth.

Seismic processing

Seismic processing of the single-channel data was initiated with rejection of random noise by a time–frequency-domain procedure (Shafi et al. 2009). The Magseis records of the source signal were used to train a 2D convolutional neural network with a U-net architecture (Ronneberger et al. 2015) for reduction of the relatively strong bubble pulse. The training was facilitated by use of 250 pairs of images with and without bubble energy. Varying coherent dipping reflection events were then enhanced by a 2D structural tensor filter (Morelatto and Biloti 2013). The source signal arriving at the seabed has a dominant frequency of c. 60 Hz decreasing to c. 40 Hz at 1 s depth (c. 1 km), which implies a vertical resolution of 8–10 m and a threshold for lateral separation of c. 250 m at 1 km sub-bottom. The zero-phase data are displayed following the SEG-polarity convention. For velocity–depth conversion, we assume a compressional velocity of 1.62 km s−1 for Cenozoic sediments, 2.0 km s−1 for Mesozoic sediments and 4.0 km s−1 for volcanic rocks (Jokat et al. 1992; O'Regan et al. 2010; Funck et al. 2022).

To enhance the deeper reflection events in the transects across the Lomonosov Ridge (Fig. 4), we use a greyscale display with illumination from above. The absolute value of the signal amplitudes has been rotated by 90° followed by bandpass filtering.

The Lomonosov Ridge between the North Pole and Ellesmere Island (the North American segment) is capped by a flat-lying section of <0.4 s (c. <320 m) thick sediments. This section is identified as Cenozoic (Fig. 4) by acoustic signature correlation (Fig. 5) with the drilled section at the ACEX drill site (Langinen et al. 2009; Kristoffersen et al. 2022). South of 88°N, the underlying Mesozoic sediments form a wedge onlapping gently sloping acoustic basement from the Makarov Basin side (Figs 4 and 5, lines 4 and 6/7) whereas parallel basement ridge topography north of 89°N frames small intra-ridge basins (Funck et al. 2022). The Mesozoic sediment wedge is floored by or includes bands of high-amplitude reflections extending downslope beyond the limits of our data (Figs 4c, d and 5). The high-amplitude reflection bands may have a thickness of more than 0.2 s (c. 400 m) and are characterized by minor discontinuities at length scales of hundreds of metres, small internal overlaps and abrupt reflector terminations (Fig. 6). In some places, the bands are associated with 4–10 km wide symmetrical mound-like structures (Fig. 7a–c) with heights up to 0.6 s (c. 600 m). The far end of the high-amplitude reflections extends downdip for more than 20 km and beyond the limits of our data (Figs 4 and 5). The deepest reflection band on line 6/7 forms a staircase pattern shallowing updip towards the SE (Fig. 5). This staircase involves a difference of c. 0.6 s (c. 600 m) in stratigraphic levels between the base of the staircase pattern to the NW and the top of the basal high-amplitude reflection band to the SE. The high-amplitude reflections at the shallowest level grade downslope over a distance of about 4 km into a high-frequency pattern of imbricated interfaces with steeper distal dips truncated at the top by more horizontal events (Fig. 7d).

Lava flows and volcanic centre

We interpret the high-amplitude reflections on top of the Lomonosov Ridge (Figs 47) as basalt flows based on the large number of comparable acoustic responses calibrated by drilling on the Norwegian continental margin, in the Faeroes–Shetland area, in the South China Sea, on the continental margin south of Australia and off New Zealand (Kjørboe 1999; Planke et al. 2000; Expedition 324 Scientists 2010; Schofield et al. 2012; Wright et al. 2012; Reynolds et al. 2017; Infante-Paez and Marfurt 2018; Sun et al. 2019; Walker et al. 2019). Minor reflector discontinuities and small overlaps (Fig. 6) suggest compound braided lava lobes (Walker 1971; Sun et al. 2019; Bischoff et al. 2021; Soule et al. 2021). In some cases, the reflections appear laterally uniform over distances of more than several kilometres (Figs 46). Lava flow morphology depends on viscosity, slope and magma supply, with extrusion rate being the dominant parameter (Bonatti and Harrison 1988; Griffiths and Fink 1992; Gregg and Fornari 1998; Deschamps et al. 2014). Lava flows at 88°N cover the Makarov Basin side of the Lomonosov Ridge from the middle of the Lomonosov Ridge (Fig. 5, upper panel). The upper mid-ridge part forms a staircase pattern towards successively shallower stratigraphic levels (Fig. 5, lower panel). We interpret this staircase pattern to reflect distinct pulses of extrusive magmatism tentatively numbered 1–4. The top lava flow at each level forms an outlier that extends several kilometres over the underlying sediments downslope. At the shallowest level, pulse 4 is associated with a low-relief mound-like structure (Figs 5 and 7d) interpreted to represent a volcanic centre. The volcanic centre is flanked to the west by a prograded geometry interpreted as a possible hyaloclastite–lava delta, suggesting volcanism at a shoreline (Garcia and Davis 2001; Garcia et al. 2007; Jerram et al. 2009; Calvès et al. 2011; Abdelmalak et al. 2016).The two largest mound structures (8–10 km wide, 0.6 km high) are flanked by symmetric high-amplitude reflections with frequent minor discontinuities, suggesting dominantly compound lava flows extruded from a volcanic centre (Fig. 7a, c and d) (Bischoff et al. 2021). The volcanic centre VC1 (line Fram-2014/15-3A) is considered to represent a complex volcano associated with lava flows downslope at two stratigraphic levels (Figs 4a and 7a). Centre VC2 is also a complex volcano with a central cone (2.5 km wide and 0.4 km high) flanked by a 1.5 km wide berm (Fig. 6c). The volcano is located on the southern slope of the Lomonosov Ridge at 88°N and associated with the oldest volcanic episode (Fig. 8). The deepest lava sequence is present in all seismic profiles and appears to have the largest areal extent of all (Figs 4 and 8). In the absence of velocity information, the thickness of the volcanic sequences is unknown. The limitations of our seismic source strength, anelastic damping and internal scattering of the signal in the basalt layers limit our ability to define any basalt thickness greater than 0.5 km assuming an interval velocity of 4 km s−1 (Shaw et al. 2008; Petersen et al. 2015). We note a general absence of high-amplitude reflection segments characterized by transgressive terminations cross-cutting younger stratigraphic horizons, a characteristic of intrusions (Malthe-Sørenssen et al. 2004; Thomson and Hutton 2004; Planke et al. 2005). This absence is remarkable in light of the abundance of intrusions present in the northern Barents Sea (Polteau et al. 2016) and the Franz Josef Land region (Shipilov 2015; Nikishin et al. 2021b).

A time-line for volcanism on the Lomonosov Ridge

The flat-lying upper part of the sediment section on the North American segment of the Lomonosov Ridge has acoustic characteristics that closely correspond to the reflectivity pattern of Cenozoic sediments at the ACEX drill site (Figs 1 and 5). The exception is an attenuated (50%) Miocene interval (Fig. 5) (Kristoffersen et al. 2022). The underlying Mesozoic sediment package shows no apparent regional unconformities above the lowest stratigraphic level of the volcanic rocks (Fig. 5). As a first approximation, we calculate an average Mesozoic sedimentation rate between a tentative age of 122 Ma for the end of the oldest volcanic pulse (VP1) and the base of the ACEX section at 56 Ma. The estimated rate of deposition (c. 1.5 cm ka−1) is comparable with the average rate for the Cenozoic section at the ACEX site (Backman et al. 2008). Given this tentative depth–age frame, we superimpose the published relative probability curves of radiometric age determinations for Franz Josef Land from Abashev et al. (2020), for Svalbard from Nejbert et al. (2011) and for the Sverdrup Basin from Dockman et al. (2018). U/Pb TIMS ages from Franz Josef Land and Svalbard from Corfu et al. (2013) and Polteau et al. (2016) are indicated in Figure 5 by a white star. We suggest that there is a first-order correspondence between volcanic pulses documented in Franz Josef Land, Svalbard and the Sverdrup Basin and the seismic stratigraphic documentation of volcanic pulses on Lomonosov Ridge presented in Figure 5. The end of pulse VP1 is assumed to be 122 Ma. The volcanic productivity during pulse VP2 decreased by about 100 Ma, followed by a minor pulse at about 95 Ma. Pulse VP3 ended by about 80 Ma. The last pulse, VP4, which produced the hyaloclastite delta, may have ended at or after about 70 Ma (Fig. 5). We note the apparent lack of hyaloclastite deltas at the flow terminations of the preceeding volcanic pulses 2 and 3, indicating general syndepositional lava extrusion throughout the period of volcanism on the Lomonosov Ridge.

Physical properties and seismic imaging of volcanic rocks

High amplitude of the reflected signal implies high acoustic impedance (the product of velocity and density) contrast across a layered rock interface. For well-lithified sediments, the difference in compressional velocity is the most significant factor. Unfortunately, our simple set-up with a single hydrophone as receiver gave no velocity information. Candidates for high-velocity sediments are carbonates, particularly well-cemented shallow-water carbonates (Amselmetti and Eberli 2012; Burgess et al. 2013) or chert (Wilkens et al. 1993), and the igneous alternative is volcanic rocks such as lava flows and intrusions. The presence of carbonates to account for the high impedance contrasts observed on Lomonosov Ridge is less likely, as Mesozoic stratigraphic equivalents are not observed on Franz Josef Land (Ershova et al. 2017) or Svalbard (Stemmerik et al. 1999; Worsley 2008) or in the Sverdrup Basin (Reid et al. 2007; Embry and Beauchamp 2008). Diagenetic fronts within biosiliceous sediments may also generate strong seismic reflections (Meadows and Davies 2009; Stratford et al. 2018) and siliceous sediments have been recovered in short cores from the Alpha Ridge (Kitchell and Clark 1982; Davies and Kemp 2016). However, no unequivocal seismic signatures related to silica diagenesis have been observed so far (Bruvoll et al. 2010; Nikishin et al. 2021b). For volcanic rocks, the compressional velocities of massive tabular basalts are c. 6 km s−1, for compound basalts–hyaloclastite rocks 4–5 km s−1 and for intrusive rocks >6 km s−1 (Sigurðsson et al. 2000; Christie et al. 2006; Nelson et al. 2009). The density of basalts is >2.7 g cm−3 whereas undeformed continental margin sediments commonly have velocities <3 km s−1 and densities <2.5 g cm−3 (Storvoll et al. 2005; Japsen 2018). Therefore, impedance contrasts across sediment–basalt interfaces are several times the contrasts across sediment–sediment interfaces and are associated with corresponding strong seismic reflections (Thomson 2005; Haug Eide et al. 2017; Toonen 2017). High signal amplitudes can also occur from constructive interference between the top and base of a basalt layer or sill, suggested by modelling to be when the thickness is 30–40 m (Smallwood and Maresh 2002). When a stack of basalts includes intercalated sediments, the rock sequence will act as a low-pass filter (Maresh and White 2005; Rohrman 2007). Although high-amplitude seismic reflections are a non-unique indicator of a sediment–volcanic interface, additional attributes such as greater interface roughness, lateral amplitude variations represented by frequent lateral discontinuities and split reflections are features commonly associated with deposition of volcanic material such as lobes of lava flows (Schofield et al. 2012; Reynolds et al. 2017; Infante-Paez and Marfurt 2018). In particular, associated mound-like high-amplitude reflection geometries suggest a non-sedimentary origin (Fig. 7). An alternative interpretation for high reflection amplitudes and lateral variability may be sand bodies within deltaic deposits, as documented in detail by Ramsden et al. (2005) and Ni et al. (2022). However, this alternative interpretation requires an independent recognition of a relevant deltaic setting. Although we recognize the uncertainties in assessing a depositional environment from single seismic lines, the consistent amplitude characteristics associated with all interfaces interpreted as volcanic rocks (Figs 47) and comparisons with calibrated seismic images from a number of continental margins suggest that the high-amplitude reflection sequences on the Lomonosov Ridge are related to volcanism. To place the evidence of magmatism on the Lomonosov Ridge in a regional perspective, we first discuss its relation to HALIP magmatism in the Franz Josef Land and the Svalbard area (Fig. 8) and subsequently consider a link to the High Arctic Magnetic High domain (Fig. 9).

Magmatic activity on the Lomonosov Ridge at the Mesozoic polar continental margin

The volcanism on the Lomonosov Ridge (Figs 47) is of Mesozoic age and refers to a palaeoenvironment prior to the Cenozoic opening of the Eurasia Basin (Vogt et al. 1979; Brozena et al. 2003; Glebovsky et al. 2006; Blakey 2021). We follow the criteria for a pre-drift reconstruction given by Glebovsky et al. (2006) based on matching conjugate continent–ocean crustal boundaries as defined by the Bouguer gravity gradient (Fig. 8). The maximum Bouguer gradient reflects the change in thickness between continental and oceanic crust. The principal difference between the pre-drift continent–ocean boundary solution of Glebovsky et al. (2006) and a Chron 25 reconstruction of Gaina et al. (2002) is that the former places the Lomonosov Ridge about 85 km farther west relative to Franz Josef Land. The initial position of North America relative to Europe based on Glebovsky et al. (2006) aligns the Pearya Terrain Boundary as defined by Poselov et al. (2011) with an initial rift direction of the Eurasia Basin (PTB in Fig. 8). Major faults in Ellesmere Island may converge southeastwards and have originated as splays from the De Geer Fault or a proto Hornsund Fault Zone (Fig. 8).

The magnetic field intensity in the Franz Josef Land archipelago is characterized by distinct lineated contours (Fig. 8) and the dominant magnetic source rocks are considered to be dolerite dykes (Dibner 1998; Shipilov 2016; Minakov et al. 2017; Shipilov et al. 2018). Dykes varying in thickness from a few metres to 20–25 m are present over the entire archipelago (Dibner 1998). Basalt flows with interbedded sediments and coal seams have a total thickness of >380 m and the total thickness of sills is greater (Embry 1992; Dibner 1998). The broad high magnetic field intensity variations over the Franz Josef Land archipelago project NNW across the continental shelf towards the North American segment of the Lomonosov Ridge (Fig. 8). However, on the Lomonosov Ridge, the main magnetic anomaly is ridge-parallel and continuous along the Makarov Basin side from north of 87°N to the North Pole (Fig. 9). Any cross-ridge magnetic anomaly trends representing dyke swarms in this area are not evident, even at the highest data resolution published. The absence may or may not be due to overprint by later volcanism as suggested by Døssing et al. (2013). The ridge-parallel positive magnetic anomaly appears to be spatially related to the presence of basalt in the seismic stratigraphy with one exception (Figs 4 and 9). At 87°N, the eastern two-thirds of line 6, which includes part of the volcanic section, borders a broad magnetic low (Fig. 9). This feature is labelled ‘low B’ in figure 3b of Døssing et al. (2013). It is important to recognize that the magnetic signature of volcanic rocks may be highly variable and a non-unique characteristic for the presence or absence of volcanic rocks. Whereas massive basalt has a relatively high intensity of magnetization, principally from a remanent component, siliceous lavas and tuffs tend to be relatively weakly magnetized (Flinn and Morgan 2002; Ganerød et al. 2010). A case in point is the contrast in magnetic expressions of outcrops of the Kap Washington Group volcanic series on the north coast of Greenland (Fig. 8), where the bulk of the material is pyroclastic rocks and siliceous lavas (Batten et al. 1981). Whereas a magnetic anomaly of a few tens of nanoTeslas (nT) is associated with a volcanic section with an estimated thickness of c. 3 km in the western part (Lockwood Island), the amplitude is more than 200 nT over the outcrop of a c. 5 km thick section at Kap Washington (Brown et al. 1987; Tegnér et al. 2011; Jokat et al. 2016). Also, a magnetic low is present between the Morris Jesup Spur and Morris Jesup Rise (Fig. 9, lower middle part) where a continuous basalt section between the two features is clearly documented in the seismic stratigraphy (figs 6 and 10 of Kristoffersen et al. 2021).

Volcanic flows extend downslope from the Lomonosov Ridge towards the Makarov Basin (Figs 8 and 9). At 89°N, basalt is present below the middle of the ridge (Figs 4b and 6b), whereas farther to the west, acoustic basement reaches the base Cenozoic unconformity. Here, the type of basement rocks is uncertain (Figs 4b and 9). Rocks below the basement interface in this area appear to lack the crude acoustic stratification characteristic of rocks interpreted as basalts (Fig. 6b). The adjacent area is crossed by line LSL1601 (130–170 km; fig. 3c of Funck et al. 2022) and the seismic refraction results suggest rocks with a velocity of 4.6–4.8 km s−1. Metasediments are an alternative explanation as suggested by Funck et al. (2022), but would require steep dips to inhibit seismic reflection images of horizontal stratification. We concur with the suggested interpretation of Funck et al. (2022) of volcanics or metasediments for the basement rocks on this part of Lomonosov Ridge.

The palaeo-position of the Lomonosov Ridge at the Franz Josef Land continental margin suggests that Mesozoic volcanism on the Lomonosov Ridge was an integral part of HALIP magmatism in the greater northern Barents Sea and Franz Josef Land domain (Figs 1 and 8). The eastern extent of volcanism on the Lomonosov Ridge corresponds closely to the known extent of magmatism in the eastern part of the Franz Josef Land archipelago, whereas its western extent is more elusive (Fig. 8). Although eruptive fissures and dykes on the ridge would be expected to represent a continuation of trends observed in the archipelago, the dominant ridge-parallel positive magnetic anomaly suggests that the most voluminous volcanism on the ridge occurred during the initial pulse below the slope facing the Makarov Basin (Fig. 9). Material with a velocity of 4.8–5.1 km s−1 has a thickness of up to 2 km (fig. 9 of Funck et al. 2022), but it is not possible to ascertain if the entire section is volcanic.

Linking Lomonosov Ridge volcanism to the High Arctic Magnetic High domain

The western slope of the North American segment of the Lomonosov Ridge borders an area of the Amerasia Basin north of 80°N (Fig. 1) unique in its association with high magnetic field intensity variations and complex trends (Fig. 9). The trends are disordered compared with the magnetic field variations in the neighbouring Eurasia Basin as well as other basins in the world's ocean (Taylor et al. 1981; Maus et al. 2009; Saltus et al. 2011). In particular, the magnitude of amplitude variations appears to be similar over the whole area irrespective of depth to acoustic basement. Oakey and Saltus (2016) have proposed the name High Arctic Magnetic High domain (HAMH) and given a succinct review of the evolution of ideas about the geological interpretation of the HAMH. The HAMH includes the Alpha and Mendeleev ridges, suggested by many to be spatially and temporally related to the High Arctic Igneous Province (e.g. Maher 2001; Vogt et al. 2006).

The contact between the crust of the North American segment of the continental Lomonosov Ridge and the Makarov Basin crust within the HAMH domain is considered a Mesozoic lithospheric plate boundary (Cochran et al. 2006; Langinen et al. 2009; Miller and Verzhbitsky 2009; Grantz et al. 2011; Shephard et al. 2013; Evangelatos and Mosher 2016; Evangelatos et al. 2017; Funck et al. 2022; Nikishin et al. 2023). The contact is associated with contrasts in both seismic velocity and density in the lower crust (Evangelatos et al. 2017; Funck et al. 2022). Our seismic stratigraphic evidence shows that flows from several magmatic events on the North American segment of the Lomonosov Ridge (70–120°W) extend downslope towards the plate boundary (Fig. 9). Whether there is a direct spatial link between HALIP magmatism on Franz Josef Land and the Alpha Ridge via the Lomonosov Ridge and the Makarov Basin is an open question (Evangelatos and Mosher 2016; Evangelatos et al. 2017; Nikishin et al. 2021b; Funck et al. 2022). The eastern slope of Alpha Ridge adjacent to the North American segment of the Lomonosov Ridge is covered by a single seismic line that combines reflection and refraction data (Fig. 1, line LSL-1604). A high-amplitude acoustic basement reflection is followed by strong reverberations with embedded weak events of an apparent dip towards the higher elevation of the Alpha Ridge (fig. 4 of Funck and Shimeld 2023). This upper 1–2 km thick layer with a velocity of 4.9 km s−1 is interpreted as a volcanic section and includes the Fedotov Seamount where recovered hyaloclastic rocks show an 40Ar/39Ar age of 90 Ma (Williamson et al. 2019; Funck and Shimeld 2023). The volcanic layer extends northwards below the Makarov Basin and includes the upper basement rocks of Marvin Spur. Any direct connection to the Lomonosov Ridge is not resolved in the regional seismic reflection line LSL1603/04 of Funck and Shimeld (2023). The seismic refraction results from line LSL1601, on the other hand, show basement velocities in the range 4.8–5.1 km s−1 from the HAMH domain of the Alpha Ridge and on to the Lomonosov Ridge suggesting a continuous layer of predominantly volcanic rocks at 88°N (fig. 18 of Funck and Shimeld 2023). Seismic resolution is the critical issue for robust documentation of the volcanic architecture of the upper crust. In the case of the Alpha and Lomonosov ridges the water depth and sediment overburden are roughly similar. Why is it possible to image lava flows on the Lomonosov Ridge (Figs 4 and 5) and on the Morris Jesup Spur (figs 7b and 10 of Kristoffersen et al. 2021) using a 0.3 l airgun source and a single hydrophone, but very difficult on the Alpha Ridge using a seismic source volume more than 50 times larger or about four times the source strength (Landrø and Amundsen 2010)? Are there fewer lava flows or is the seismic source spectrum too skewed towards low frequencies? Perhaps anelastic attenuation (Q-factor) caused by a thick package of predominantly tuff is a contributing factor?

The radiometric age determinations of basalt samples from the Alpha Ridge are at present limited to four locations (Fig. 1). A trachybasalt sample from the Trushkin Seamount at the western end of the ridge (Fig. 1, site 4) has been obtained by remote drilling and dated at 127 Ma by the U–Pb method (Morozov et al. 2013). Two sites (Fig. 1, sites 2 and 3) yielded 40Ar/39Ar ages of 90 Ma (Williamson et al. 2019) and 89 Ma (Jokat et al. 2013). Also, the sediments above acoustic basement recovered in short cores from the Alpha Ridge may be at least as old as the Maastrichtian or Campanian (Mudie et al. 1986; Dell'Agnese and Clark 1994; Firth and Clark 1998). On the Mendeleev Ridge, basalts from the Healy Spur have ages of 105–100 and 90–70 Ma whereas basalts from a lower scarp on the northern end of Northwind Ridge gave ages of 118–112 Ma (Mukasa et al. 2020).

In summary, we observe extensive magmatism partly expressed by discrete pulses over a distance of c. 600 km along the Lomonosov microcontinent bordering the High Arctic Magnetic High domain (Figs 4 and 5). The main volcanism on the Lomonosov Ridge is limited to the North American segment of the ridge, which had a palaeo-position proximal to Mesozoic magmatic activity on Franz Josef Land (Fig. 8). Volcanism documented on the adjacent Alpha Ridge around 127 and 90 Ma falls within known HALIP episodes on the adjacent continental margins. The first magmatic pulse on the Lomonosov Ridge appears to be the most voluminous (Fig. 9), which is in line with the main Hauterivian–early Aptian activity in Franz Josef Land (Shipilov 2016; Abashev et al. 2020; Karyakin et al. 2021) as well as the age from the Trushkin Seamount at the western end of the Alpha Ridge (Morozov et al. 2013). We suggest that the first and most voluminous volcanic pulse on the Lomonosov Ridge was part of magmatism that in space and time involved the entire HALIP domain (Fig. 5). Later extrusion events on the Lomonosov Ridge appear to be of more local significance. The youngest magmatic pulse VP4 on the Lomonosov ridge ended after about 70 Ma and may partly be related to volcanism associated with the opening of the Eurasia Basin including the initial emplacement of basalt in the Morris Jesup Spur and Morris Jesup Rise area (Kristoffersen et al. 2021) and the Kap Washington Group north of Greenland (Estrada et al. 2010; Tegnér et al. 2011; Thorarinsson et al. 2011, 2015).

Thermal aspects of HALIP magmatism

The supply of magma in the HALIP domain would enhance heat transport to the upper continental crust. However, the measured conductive heat flow on the Alpha Ridge appears low, 50–55 mW m−2, with an uncertainty of about 20% (Ruppel et al. 2019). Values from the northern part (c. 120–165°W) of the North American segment of the Lomonosov Ridge are 60–70 mW m−2 (Sweeney et al. 1982; Shephard et al. 2018), and measurements in three deep boreholes in the Franz Josef Land archipelago give 77–80 mW m−2 (Khutorskoi et al. 2008). Evidence of increased heat input to the upper crust on the North American segment of the Lomonosov Ridge is documented by the recent discovery of gas or fluid escape pipes penetrating the Mesozoic sediment section above the basalt (Figs 8 and 10). On the conjugate Franz Josef Land continental margin, Sokolov et al. (2017) have mapped areas of gas flares emanating from the boundary between Triassic sediments and Jurassic clays below the seabed SW of Franz Josef Land (Fig. 10). Also, bitumen is found in Lower Jurassic and Upper Triassic strata in the vicinity of dykes on islands in the central (Hayes Island) and eastern (Graham Bell and Wilchek islands) part of the Franz Josef Land archipelago (Bezrukov 1997; Klubov et al. 1997, 1999). The combined absence of volcanism and evidence of thermogenic gas on the Central and Siberian segments of the Lomonosov Ridge (Jokat et al. 1992; Jokat 2003; Rekant and Gusev 2012; Sauermilch et al. 2018; Weigelt et al. 2020) strongly suggest that HALIP volcanism enhanced maturation of pre-HALIP hydrocarbon source rocks on the North American segment (Stein 2007; Mann et al. 2009; Moore et al. 2011). However, lavas extruded on the surface cool quickly, as convective heat transport is important in the upper 500–1000 m of highly fractured extrusive volcanic units. Conductive cooling becomes dominant at deeper levels (Flovenz and Saemundsson 1993). Horizontal or vertical intrusions at depth create a contact aureole where the width is defined as the zone of raised vitrinite reflectance <1%Ro (e.g. Aarnes et al. 2010; Maré et al. 2013). A compilation of studies (table 1 of Aarnes et al. 2010) of contact metamorphism of sheet intrusions (width D) into shales and mudstone suggests that intrusions of D <10 m create an aureole of width less than D and intrusions <100 m wide are associated with aureoles <2D wide. The aureole thickness increases with the background temperature and sequential intrusions at different stratigraphic levels will raise the background temperature (Zalan et al. 1990; Brown et al. 1994; Shutter 2003; Aarnes et al. 2011). Heat from massive Mesozoic intrusions into Permian sediments, Triassic clinoform sequences and Jurassic synclines (Polteau et al. 2016; Sømme et al. 2018; Johansen et al. 2019) most probably enhanced hydrocarbon maturation in the Franz Josef Land archipelago (Shipilov 2015) and in the Svalbard area (Mørk and Bjorøy 1984; Brekke et al. 2014; Lundschien et al. 2014). As the presence of gas or fluid escape pipes on the Lomonosov Ridge is restricted to the area of volcanism, we suggest that undetected Mesozoic intrusions may be present on the North American segment of the ridge. By the same reasoning, the relatively low heat flow values on the Alpha Ridge may indicate predominantly extrusive HALIP magmatism, which cooled relatively quickly (Funck and Shimeld 2023; Nikishin et al. 2023).

The traditional ice drift approach to exploration of the Arctic Ocean is still valid for areas difficult to access by icebreaker surveys. We have over-wintered on the drifting sea ice (c. 4.5 km day−1) using a hovercraft platform to explore the geology of the submarine Lomonosov Ridge from the North Pole towards Canada and Greenland. The new seismic reflection data document Mesozoic magmatism over a distance of c. 600 km along the Lomonosov Ridge at the boundary between the microcontinent and the High Arctic Magnetic High domain, which is considered a signature of extensive Mesozoic magmatism in the deep basin between the North Pole and Alaska. The concept of a large Mesozoic igneous province built from discrete volcanic pulses established from the land geology is captured in a seismic image at 87°N on the Lomonosov Ridge in the form of a staircase pattern formed by several discrete events of enhanced magma productivity during a period spanning a 0.6 s (c. 600 m) section of the Mesozoic stratigraphy. The first magmatic pulse on the ridge appears as the most extensive and may have been part of plume activity that involved the entire area covered by the High Arctic Large Igneous Province. Later volcanic episodes on the Lomonosov Ridge appear to be more local events. We speculate that the last magmatic pulse (c. 70 Ma) is possibly related to volcanism associated with opening of the Eurasia Basin and emplacement of basalts on the Morris Jesup Spur and Morris Jesup Rise as well as the volcanic rocks representing the Kap Washington Group on the north coast of Greenland. Volcanism on the Lomonosov Ridge as well as pipe structures representing gas-charged fluid migration in the overlying sediments are restricted to the North American segment of the ridge, which formed the pre-Cenozoic continental margin north of Franz Josef Land. We suggest that heat from undetected intrusions together with increased overburden from extrusive magmatism enhanced maturation of hydrocarbon source rocks on the North American segment of the Lomonosov Ridge.

This paper is dedicated to senior engineer O. Meyer in recognition of three decades of outstanding technical support for our polar research projects. The Fram-2014/15 ice drift was sponsored by Blodgett-Hall Polar Presence LLC, Lundin-Norway, the Norwegian Petroleum Directorate and the Nansen Environmental and Remote Sensing Center. The logistic support provided by A. Tholfsen, A. Wegener Institute of Polar and Marine Research, the Norwegian Air Force and the Danish Air Force, along with technical assistance by Senior Engineer O. Meyer and Griffon Hoverworks, made it all possible. A. J. Bugge generously contributed to the neural network approach and J. E. Lie advised on processing details. I. Gimse of Magseis Fairfield kindly provided a seismic node for observation of the seismic source signature. The enthusiasm of G. B. Larsen, H. Jahre and H. Brekke were critical to the realization of a Norwegian ice drift in the central Arctic Ocean only 118 years after Nansen's drift with the Fram. We much appreciate the constructive comments from reviewers O. Galland and D. Mosher which improved the presentation of the results.

YK: conceptualization (equal), formal analysis (lead), funding acquisition (supporting), investigation (lead), methodology (lead), project administration (lead), resources (equal), visualization (lead), writing – review & editing (lead); EHN: formal analysis (equal), software (lead), visualization (equal); JKH: conceptualization (equal), funding acquisition (lead), resources (equal), writing – review & editing (supporting)

Funding support from the Blodgett-Hall Polar Presence LLC, Lundin Energy Norway, Oljedirektoratet and Nansen Environmental and Remote Sensing Center.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The dataset acquired and analysed in the current study is available from the corresponding author on reasonable request.

This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License (http://creativecommons.org/licenses/by/4.0/)