The Lomonosov Ridge, central Arctic Ocean – the world's longest submarine ridge of continental origin: outline of the history of exploration, morphology, sediment deposition and exhumation of the North American and Central ridge segments Open Access

The Lomonosov Ridge divides the polar ocean into two major sub-basins: the Eurasia Basin and the Amerasia Basin (Fig. 1). The width of the ridge, as defined by the 2500 m isobath, ranges from 28 km in the central area near the North Pole to 70–200 km at both ends (Fig. 2). Its length is c. 1800 km, which is c. 50% longer than the central European Alps, and the height above the adjacent abyssal plains is comparable with the average height of the alpine peaks (2.5 km). The ridge has a similar average elevation difference as the deep part of the Gulf of California, but is three times its length.

The Lomonosov Ridge can be divided into three segments based on morphology and slight changes in orientation (Rekant et al. 2019): a Siberian Segment formed by a linear ensemble of narrow blocks; a Central Segment with two staggered larger blocks; and a North American Segment (NAS) dominated by smaller blocks in the narrow part near the North Pole and a large block south of 89° N (Fig. 2). The ridge structure near the North Pole is split by a 110 km long, 20 km wide closed basin with its floor c. 1600 m below the level of the ridge crest. A distinct bathymetric trend cutting obliquely across the ridge at 10° is present on all three ridge segments (Fig. 2).

The morphology of the Amerasia-facing side of the Lomonosov Ridge varies from the gentle slopes (<3°) of the Siberian Segment into the Podvodnikov Basin, to the relatively steep slopes (up to 6°) of the Central Segment into the Makarov Basin, grading into more gentle slopes along the NAS (Figs 2, 3). In the Eurasia Basin, the drop in basement elevation from the ridge to the deep basement below the abyssal plain takes place over 40–100 km and has a magnitude of 5–6 km, with a step-like morphology formed by rotated blocks (down towards the ridge) (Langinen et al. 2008; Poselov et al. 2011, 2012; Knudsen et al. 2017; Rekant et al. 2019; Savin et al. 2019; Weigelt et al. 2020; Nikishin et al. 2021; Funck et al. 2022). The distance from the foot of the slope of the Lomonosov Ridge to the position of Chron 24 varies from 60 to 80 km on the Siberian Segment to >100 km on the NAS (Savin et al. 2019; Funck et al. 2022).

The Lomonosov Ridge forms an oceanographic barrier across the Arctic Ocean (Schlosser et al. 1997). The crestal water depths range between 500 and 1500 m (Fig. 4). The barrier is separated from the continental margins at each end by passages, which reach depths of c. 1200 m north of Ellesmere Island and c. 1600 m north of the New Siberian Islands (Jakobsson et al. 2012). In part, these saddle points constrain the depths of the cyclonic circum-Arctic boundary current driven by thermohaline density differences and the potential vorticity (Aagard 1981; Nøst and Isachsen 2003; Yang 2005; Aksenov et al. 2011).

The most dynamic circulation involves the Atlantic layer (temperature >0°C), which extends from the halocline (c. 200 m) to 600–800 m depth, and the upper polar deep water, characterized by decreasing temperatures and increasing salinities down to c. 1500 m depth (Rudels et al. 1994, 1996; Jones et al. 1995; Jones 2001). The circum-Arctic boundary current is sourced by the inflow through the Fram Strait and from the Barents Shelf on the northern continental margin (Rudels et al. 1999, 2000; Aksenov et al. 2011). The two inflows mix laterally, along with contributions from brine rejection on the way towards the Lomonosov Ridge (Schauer et al. 1997; Karcher et al. 2007; Aksenov et al. 2011; Rudels et al. 2013). At the junction between the Lomonosov Ridge and the East Siberian continental margin, about half of the flow passes through the saddle point and the other half is deflected west along the Lomonosov Ridge in the Eurasia Basin (Woodgate et al. 2001). Flow on the Makarov Basin side of the ridge is from west to east (Jones et al. 1995; Jones 2001; Nøst and Isachsen 2003).

Flow across the Lomonosov Ridge has been both recorded directly by current meters (Aagard 1981) and inferred from water mass characteristics (Timmermanns et al. 2005; Björk et al. 2010, 2018). Intermittent flow may occur in both directions across the ridge. The two saddle points at the junction between ridge segments (Fig. 4) lower the elevation of the Lomonosov Ridge by c. 500 m. One connects with an intra-ridge basin near the North Pole, where the threshold on the Makarov Basin side is c. 1600 m and the other saddle point is at c. 84.5° N (the Oden Trough) with a sill depth of 1470 m (Björk et al. 2018).

Measured peak current velocities along and across the Lomonosov Ridge on the Eurasia Basin side are <12 cm s⁻¹ at 20–25 m above the seabed (Aagard 1981; Woodgate et al. 2001). The calculated velocities on the Makarov Basin side, based on the potential vorticity, are half or less of these values (Nøst and Isachsen 2003). The highest velocities are capable of the erosion of the medium silt-sized fraction of sediment deposited on the seabed (Ledbetter and Ellwood 1980; McCave 1984; McCave et al. 2017).

The water mass in the Arctic Ocean is stabilized by the distribution of salinity and temperature. As a consequence, the deeper waters (below 2500 m) are isolated from surface forcing. They are sourced by advection, with the circulation pattern steered by the bottom topography (Aagard 1981; Rudels and Carmack 2022).

Demenitskaya and Kiselev (1968) and Kiselev (1970) used shallow seismic refraction studies to document the presence of layered sediments thinning from c. 8 km near the shelf to 3 km at c. 85° N on the Siberian Segment of the Lomonosov Ridge. The underlying rocks had a velocity of c. 6.4 km s⁻¹. The deeper structure of the Lomonosov Ridge was first explored by seismic refraction and gravity measurements during the LOREX expedition (Fig. 1) near the North Pole (Forsyth and Mair 1984; Sweeney et al 1982). The results suggested that the ridge was a crustal block of Mesozoic and older sediments or low-grade metasediments above a lower crust of intermediate to basic crystalline rocks (v_p = 6.6 km s⁻¹). The Transarktika-1992 transect across the Siberian Segment at 83° 30′ N (Fig. 1) yielded similar results (Poselov et al. 2012). Recent transects across the NAS of the Lomonosov Ridge (Fig. 1, line LSL1601) show evidence of shallow volcanic sequences and a 5 km thick slab of high-velocity lower crust, interpreted as magmatic underplating (Funck et al. 2022).

Direct geological evidence from the ridge is very limited. The recovery of 23 cm of a breccia with siltstone clasts in a piston core (Fig. 1, location A) on the upper slope near the North Pole hinted at continental rocks (Grantz et al. 2001). However, the most definitive contribution to the pre-Mesozoic geology of the Lomonosov Ridge has been the recovery of c. 300 kg of dredged rock samples (Fig. 1, location B) on the NAS (Knudsen et al. 2017). The dominant rock type was metamorphic sandstone to silty mudstone of the greenschist facies, which yielded Ordovician to early Silurian ages. The non-metamorphic rocks were sandstones and siltstones (Knudsen et al. 2017). The Siberian Segment has been sampled at 82.54° N (Fig. 1, locations C and D) on slopes of stratified sediments outcropping on basement highs (Rekant et al. 2019). This hydrostatic coring effort recovered fragments of non-metamorphosed arkosic sandstone, which yielded Neoproterozoic detrital zircon ages.

The seismic reflection data show a planar angular unconformity across the Central Segment, suggesting that the ridge segment was once elevated above sea-level (Jokat et al. 1992). The acoustic image at the drill site of the Arctic Coring Expedition (ACEX) and the recovered stratigraphy is shown in Figure 5. The drilling was terminated at 427.63 m below the seafloor with the recovery of 1.4 m of silty clay of Campanian? age (Expedition 302 Scientists 2006), indicative of a setting proximal to a palaeocoastline (e.g. März et al. 2010).

The site subsided during the late Paleocene–Early Eocene, followed by an extended period within a neritic environment (O'Regan et al. 2008). The late Paleocene–early Eocene basal Unit 3 (91 m thick) of hard silty clay to mudstone is overlain by 90 m of mud-bearing biosiliceous ooze of mid-Eocene age (Unit 2). The palaeoenvironmental proxies suggest a well-stratified ocean with periods of ‘Black Sea’-type conditions during the late Paleocene–early Eocene (e.g. März et al. 2010). The estimated sea surface temperature of c. 25°C decreased gradually after c. 47 Ma (e.g. Stein 2019), with indicators for the presence of sea ice appearing at c. 46 Ma (St. John 2008; Stickley et al. 2008).

Deposits representing the latest Eocene to early Miocene appear to be missing (Sangiorgi et al. 2008), but with no indication of a major change in geochemical proxies across the stratigraphic gap (e.g. März et al. 2010). Alternatively, the sediment deposition was continuous and Units 1–5 accumulated at an extremely slow rate (Poirier and Hillaire-Marcel 2011; Chernykh and Krylov 2017). Above the apparent stratigraphic break is a 200 m thick section of late early Miocene–Recent soft clays, which document a rapid change from euxinic to fully marine oxic conditions. This shift has been related to the emerging Fram Strait gateway and connection to North Atlantic water masses (Jakobsson et al. 2007).

The new seismic data from the NAS of the Lomonosov Ridge presented here (Fig. 2, white lines) extend the scientific drilling results from the Central Segment to explore the evolving early Cenozoic environment at the polar margins of Europe, Greenland and Canada.

The NAS of the Lomonosov Ridge is the least explored part of the ridge because it lies below the flow of heavily ridged sea ice exiting the Beaufort Gyre and moving out of the Arctic Ocean north of Greenland (Rigor et al. 2002). The frequency of pressure ridges where the ice thickness may exceed 10 m make it extremely difficult for a single ice-breaker to carry out seismic reflection surveys.

The scant seismic reflection data available for the NAS from stations on drifting sea ice include line drawings of returns from a sparker source made on Arlis-2 (Ostenso and Wold 1977), the LOREX transect using a small airgun source and electrostatic paper recording (Blasco et al. 1979), followed by the Soviet ice drift NP-28 using an explosive source of bundled detonators at 500 m shot distance. The NP-28 analogue recordings from an L-shaped sensor array with 545 m long arms were later digitized and processed (Langinen et al. 2008). Direct digital data were first recorded in short profiles from a small airgun source (Kristoffersen and Mikkelsen 2006; Jackson et al. 2010). The first extensive modern dataset was presented by Funck et al. (2022), who used two ice-breakers in tandem to successfully acquire three seismic reflection and refraction transects across the ridge (Fig. 1). The focus on the full geological ridge structure implied a trade-off against the resolution of details in the sedimentary section.

We pursued the traditional ice drift approach (Fram-2014/15) to obtain high-resolution seismic reflection data from the NAS of the Lomonosov Ridge (Fig. 2, profiles 2–5). A mobile hovercraft platform served as the base for an over-winter expedition on drifting sea ice (Kristoffersen et al. 2016, 2023). The craft, with a payload capacity of 2.2 tons and practical hover height of 0.5 m, was fully equipped as a scaled-down research vessel (Kristoffersen and Hall 2014). The seismic source was a (0.3 L) airgun powered by a dive compressor and fired under GPS control at 25 m intervals. The receiver was a single hydrophone suspended just below the ice–water interface to avoid cable strumming. This simple set-up exploits the low ambient noise level below an ice-covered ocean (Buck and Greene 1964; Buck 1968; Dyer 1984; Makris and Dyer 1986) and may achieve up to 2 km of sub-bottom penetration in 3 km water depth (Kristoffersen et al. 2021).

We revisited legacy seismic reflection data from the Central Segment acquired from ice-breaker Oden in 1996 (Fig. 2, profiles 6, 9 and 10). The operation from Oden used 2 × 1.3 L airguns as the sound source at 8 m depth and records were made with a 16-channel, 200 m long analogue streamer towed at c. 10 m depth with no depth control. A shot interval of 20 s resulted in a shot spacing of c. 20 m at an average speed of 2 knots (3.7 km h⁻¹).

We used the International Bathymetric Chart of the Arctic Ocean Version 3.0 to study the ridge morphology (Jakobsson et al. 2012). The coverage is made up of single-beam legacy data and single multi-beam tracks. Profiles 2–4 (Fig. 3) on the Lomonosov Ridge are within the area swath-mapped during SCICEX (Edwards and Coakley 2003). The slope profiles were derived using Global Mapper and GeoMapApp software (Ryan et al. 2009).

The raw seismic reflections generated by the 0.3 L airgun suspended at 5 m depth displayed a strong bubble pulse, as shown in Kristoffersen et al. (2021, their fig. 5). A complicating factor was the variation in the arrival time and the frequency of the bubble pulse because the air pressure could drop below the nominal 140 bar when the 40 L air reservoir was not recharged in time by the manually operated compressor. We used two processing approaches to enhance the data. One approach accurately tracks the bubble pulse time variation and, by trial and error, estimates the effective frequency content that best enhances the signal obscured by the bubble pulse. It then applies a minimum phase predictive deconvolution and removes random noise by stacking on common reflection surfaces (de Bazelaire 1988; Gelchinsky 1988). The other method for bubble pulse reduction uses the source signal recorded at 700 m depth to train a 2D convolutional neural network with a U-net architecture (Ronneberger et al 2015), as reported by Kristoffersen et al. (2023). The zero-phase data are displayed following the SEG polarity convention.

The NAS and Central segments of the Lomonosov Ridge are characterized by relatively flat tops flanked by increasing slopes (Fig. 2). The Makarov Basin side of the Central Segment has narrow slopes (c. 40 km wide, average 4.3° slope) with an elevation difference of c. 3 km (Fig. 3). A bathymetric swath survey was made by the nuclear submarine UNS Hawkbill in 1998 (Edwards and Coakley 2003) along c. 280 km of the ridge segment (Fig. 3, light shading in upper right panel). The UNS Hawkbill dataset shows gullies and canyons forming incisions in the slope; numerous gullies on the upper slope feed into 10–15 km wide canyons, which may have up to 500 m relief (Kristoffersen et al. 2007). The slope profile is roughly linear to the level of the abyssal plain (Fig. 3).

The seismic reflection data show that the slopes facing the Makarov Basin on the NAS and Central Segment of the Lomonosov Ridge are formed by old strata truncated at the seabed and covered by only a thin veneer of younger sediments (Fig. 2). The depth range of the truncations is from <400 m below the top of the ridge down to the level of the adjoining abyssal plains. Published seismic reflection data allowed us to follow the slope to c. 2 km below the abyssal plain (4 km depth) in the Makarov Basin and no truncation is observed (e.g. Evangelatos and Mosher 2016). The ridge elevation becomes subdued relative to the adjoining plains on the Siberian Segment south of 83° N (Fig. 2).

The Amundsen Basin side of the ridge is dissected by ridge-parallel faults, which give a complex morphology, and assessment of the slope is difficult (Cochran et al. 2006; Rekant and Gusev 2012; Knudsen et al. 2017; Nikishin et al. 2021).

The new seismic data explore the first-order regional acoustic stratigraphy of the NAS of the Lomonosov Ridge to c. 2 s travel time (c. 2 km) below the seabed (Figs 2 and 6–8). The top of both the NAS and the Central Segment is characterized by a flat-lying, <0.5 s thick, well-stratified sequence of uniform thickness above an unconformity or correlative conformity and is attenuated towards both ridge flanks (Figs 2, 6 and 7). The section on the NAS is c. 70% of the corresponding thickness at the ACEX drill site on the Central Segment. The unconformity at the base of the sequence is uneven below the northern part of the NAS and is absent between 87 and 88° N (Fig. 2, profiles 2 and 3). On the Central Segment, this unconformity forms a flat surface and a hiatus has been verified by scientific drilling (Figs 2, 5).

Several exceptions to laterally uniform horizons are specific to the NAS of the ridge (Figs 6, 7). In particular, on the Makarov Basin side, we observe an attenuated accumulation near the ridge flank at 150° W and a local ridge along the same flank further south between 87 and 88° N (Fig. 7a). These accumulations are internally stratified and initiated at a stratigraphic base characterized by high-amplitude reflections. In one case, an initial single crest shows slight migration away from the flank of the Lomonosov Ridge (Fig. 7a, lower left panel).

At the corresponding stratigraphic interval on the Amundsen Basin side of the Lomonosov Ridge is a stratified local accumulation (0.1 s high, >3 km wide) at 89° N, 90° W (Fig. 7b, lower right panel). The accumulation is tapered towards the centre of the ridge, but truncated by erosion beneath the ridge flank (Fig. 7b, line Fram 2014/14-3B). About 80 km to the north, the horizon above the high-amplitude reflection unit shows a slight differential increase towards the ridge flank (Fig. 6a, line 3A, light blue horizon). Further along the ridge, on the Central Segment, <5 km from drill site M004 of the ACEX, is a smaller corresponding accumulation indicated by reflections diverging towards the ridge flank (Fig. 7b, line AWI 9190). Higher in the stratigraphy, near the drill site location, is a dramatic erosional truncation (Fig. 7b, upper middle panel, red line), which is also present at the same stratigraphic level on the NAS (Fig. 7b; Fram-2014/15 line 3B, lower middle panel).

At 87° N, adjacent to the shallow southern end of the NAS, are c. 30 km wide sub-units of lens-shaped cross-section present at two distinct stratigraphic levels (Fig. 8, upper panel, red colour). The upper lens (B) is expanded towards deeper water by down-lapping strata (yellow lines). The thickness of lens A is 0.2 s (c. 200 m) and lens B about half of that. The profile crosses the northern end of a plateau roughly outlined by the present-day 1000 m depth contour (Fig. 8, middle right panel).

The seismic line Fram-2014/15 line 7 also crosses an enigmatic pyramid-shaped structure (c. 5 km wide, 1 km high) on the Amundsen Basin flank of the Lomonosov Ridge near 87° N, 60° W (Fig. 9, left panels). The true geometry (cone or ridge) remains unknown. The acoustic image of the upper 0.2 s of the structure appears as returns from random scatterers. In the deeper part, weak reflections project towards the centre and indicate a downward increase in slope for the core of the structure. This is in the opposite sense of a suspected velocity pull-up, suggesting growth by the addition of material in a central area. The reflection amplitudes are modest, except perhaps for the lower flanks (Fig. 9, left panel). The structure has no associated magnetic signature and is located at the foot of a small circular free air gravity anomaly of a few tens of milligals to the south (Gaina et al. 2009). Possible origins of this feature could be a horst, a carbonate mound or a volcano.

The sediments below the unconformity on the NAS persistently onlap the Lomonosov Ridge from the Makarov Basin and are abruptly truncated at the seabed along the slope of the ridge (Figs 2, 6). We observe an exception on the NAS, where a basement high reaches the level of the base-Cenozoic unconformity and forms a slope towards the Amundsen Basin with onlapping sediments (Fig. 2; transects 5 and 9). A coring attempt at the projected outcrop of the base of the Cenozoic section on the Makarov Basin side obtained a 20 cm sample of very cohesive, light cream-coloured, almost pure siliceous ooze (Fig. 6a, upper left panel).

The observable acoustic stratification below the regional unconformity is depth-limited by the strength of our seismic source, but suggests a maximum thickness of at least c. 0.7 s (>700 m) on the NAS and 1.3 s (c. 2.6 km) on the Central Segment (Jokat et al. 1995; Nikishin et al. 2021). The corresponding stratigraphy on the Siberian Segment appears mainly as infill between high-standing crustal blocks or as cover on rotated fault blocks (Fig. 2). On the Eurasia Basin side, the apparent dips are mainly related to the rotation of fault blocks (Fig. 2); see also Knudsen et al. (2017), Rekant et al. (2019) and Weigelt et al. (2020).

Unique to the NAS of the Lomonosov Ridge are high-amplitude reflection bands within the sediment section below the regional unconformity (Figs 2, 6b). Their acoustic signature and interpretation as volcanic flows and volcanic centres have been dealt with in detail by Kristoffersen et al. (2023) and are mentioned here for stratigraphic completeness only (Fig. 2).

Scientific drilling on the Central Segment of the Lomonosov Ridge has documented a c. 400 m thick Cenozoic section of hemipelagic sediments on top of the ridge (Expedition 302 Scientists 2006). The increasing slope of the uppermost ridge flanks appears to be directly related to the attenuation of the upper flat-lying Cenozoic sediments to a near-zero thickness above a horizontal unconformity (Figs 2, 6a). We interpret the geometry of the Cenozoic deposits to represent hemipelagic sediments settling out of a nepheloid layer advected in along the ridge by a combination of contour-following bottom currents and the relentless action of internal tides breaking over the slope (Cacchione and Drake 1986; Cacchione et al. 2002; Puig et al. 2004). High current velocities on the uppermost flank would reduce deposition and taper the accumulation towards the ridge perimeter (Bowden 1960; McCave 1984; Ribbe and Holloway 2001; McCave et al. 2017).

We assume the Makarov Basin slope received a dominantly siliciclastic sediment input from the south in its pre-Cenozoic position as the outer part of the Franz Josef Land continental margin (Glebovsky et al. 2006; Sømme et al. 2018; Blakey 2021). In the present position of the Lomonosov Ridge, the Mesozoic continental slope deposits appear as truncated layers towards the Makarov Basin (Figs 2, 6b) as a result of being cut off from the clastic continental shelf sediment input and subjected to submarine erosion at least since submergence c. 56 Ma (Expedition 302 Scientists 2006). We interpret the high slope angle on the lower slope as due to steepening by bottom current erosion. The slope appears roughly linear over an elevation difference of c. 5 km, a characteristic feature of low sediment input and a mature state (O'Grady et al. 2000). Sediments arriving from failure events on the higher slope have apparently not come to rest at the foot of slope, but have been redistributed by bottom currents over a larger area. Alternatively, accumulations around the foot of the slope may have become masked by turbidite deposition in the adjoining abyssal plain, where the sedimentation rate appears to be more than twice the rate of hemipelagic deposition on top of the ridge (Evangelatos and Mosher 2016; Nikishin et al. 2021).

The remarkable similarity between acoustic images from the NAS and the seismic image calibrated by scientific drilling on the Central Segment for the deposits above the regional unconformity enable an acoustic signature correlation and tentative lithostratigraphic extrapolation from the Central Segment to the NAS (Figs 5–8). The main acoustic attributes are as follows.

1. An upper, relatively transparent, c. 0.25 s thick acoustic interval that correlates with soft to hard clays (Unit 1) extending back to the late early Miocene (18.2 Ma).

2. Two closely spaced reverse polarity reflections. The first is associated with Unit 1/5, a mid-Eocene to late early Miocene hiatus or a condensed interval (Backman et al. 2008; O'Regan et al. 2008; Sangiorgi et al. 2008; Poirier and Hillaire-Marcel 2011; Chernykh and Krylov 2017). The second reflection arises from a c. 25% decrease in the bulk density at the base of the 25 m thick silty clay of Unit 1/6 of mid-Eocene age at the top (Expedition 302 Scientists 2006).

3. A 0.1 s thick interval characterized by low amplitudes associated with the biosiliceous silty clay to mudstone of Unit 2 (90 m thick). The age at the base is early Eocene.

4. A lower band of high-amplitude reflections related to Unit 3 (92 m thick), representing hard silty clay to mudstone with a basal age of late Paleocene.

A particular feature on the NAS is the presence of ridge-shaped accumulations inboard of the break in slope at the ridge perimeter (Fig. 7). This is within a stratigraphic interval corresponding to the biosiliceous ooze of Unit 2 of early to mid-Eocene age (Fig. 7b). The accumulations are commonly asymmetrical in cross-section, with a steep outboard slope and tapered towards the centre of the ridge. The internal stratification leads us to interpret the accumulations as sediment drifts, in spite of their position inboard of the ridge flank and the fact that, in some cases, their height relative to their width appears unusual. The features become wider and their heights are more reduced northwards along the both sides of the ridge. No associated moat is evident. The sediment drifts are often positioned on the upper continental slope (e.g. Rebesco et al. 2014), but not landward of the shelf edge. No analogy of a similar sediment drift/bottom topography relation has been found in the scientific literature. As this is a submarine ridge, we envision a scenario with deposition from sediment-laden bottom current flow across the Lomonosov Ridge in addition to flow along the bathymetric contours. A possible sediment source area would be the elevated plateau to the south (Fig. 7a).

The sediment lenses A and B at the southern end of Lomonosov Ridge have a cross-section tapered up-slope as well as downslope (Fig. 8). Their local extent and position next to a plateau suggest that the sediment source area was the elevated plateau. The seismic refraction data from the LORITA experiment (Jackson et al. 2010) show that the eroded plateau is mantled in the centre by exposed upper crustal rocks (v_p = 5.95 km s⁻¹), possibly framed by metasediments (v_p > 5.2 km s⁻¹). We interpret sediment lenses A and B to be a result of enhanced local clastic sediment input during periods of plateau uplift with subaerial exposure, high topographic gradients and erosion.

We tentatively interpret the pyramid-shaped structure on the flank of the Lomonosov Ridge at 87° N, 60° W to be a volcanic centre, but the lack of high-amplitude reflection segments appears problematic (Fig. 9). A small basement horst could initially have been present and the upward increase in slope does not exclude later tectonic movements. Another alternative could be a carbonate mound because a camera station on the flank of the structure shows abundant algal mats, which indicate the presence of methane at the seabed (e.g. Kravchishina et al. 2021). Methane supports the formation of authigenic carbonate and the construction of mounds (Judd and Hovland 2007). However, carbonate mounds, such as the Challenger Mound (Ferdelman et al. 2006; Henriet et al. 2014), appear not to reach dimensions comparable with this structure on the Lomonosov Ridge (Fig. 9). Submarine volcanoes can be outlined by high-amplitude seismic reflections, but are often not outlined in this way (Reynolds et al. 2014; Infante-Paez and Marfurt 2018; Bischoff et al. 2021). The reflectivity of the interior of a volcano sometimes shows the seismic character of a chaotic ‘salt and pepper’ configuration when tuff and breccias are the main volcanic rock components (Infante-Paez and Marfurt 2018). An interpretation of the pyramid-shaped submarine feature as a volcanic centre on the Lomonosov Ridge is not unreasonable in view of other volcanic centres present on the ridge c. 40 km to the west and in the Amundsen Basin 120 km to the east (Kristoffersen et al. 2021, 2023).

We interpret the smooth and even peneplain that forms the regional unconformity on the Central Segment to represent the erosion of relatively soft sedimentary rocks, whereas the more uneven interface at the northern end of the NAS (Fig. 6a) indicates a more resistant substratum of metamorphic and/or igneous rocks: This geology was also suggested by Funck et al. (2022). The sediments below the base-Cenozoic section on part of the Central Segment (Fig. 2, profiles 7–9) show prograding sequences towards the Makarov Basin, whereas onlapping geometries dominate on the NAS. The only sample recovered is an almost pure siliceous ooze (Fig. 6a, upper left), although the Mesozoic sediment section is expected to be predominantly siliciclastic with contributions of volcanoclastic(?) sediments based on its palaeoposition on the continental margin north of Franz Josef Land. Estimates of the age of the sediments below the base-Cenozoic unconformity on the Lomonosov Ridge can be made for the NAS based on an assumed age of the deepest and most extensive volcanic event (Kristoffersen et al. 2023) and, for the Central Segment, based on sedimentation rates.

The morphological spectrum of siliclastic passive continental margins shows recognizable patterns (Emery 1980; O'Grady et al. 2000), with the sediment supply, margin maturity and action of internal tides (Cacchione et al. 2002) considered to be the most important factors shaping the continental slope.

The speed of the thermohaline and vorticity-driven bottom currents flowing along the upper slope of the Lomonosov Ridge (Nøst and Isachsen 2003; Yang 2005; Rudels and Carmack 2022) is, in part, related to the steepness of the slope (Bowden 1960). Friction and the Coriolis force induce flow across the ridge, as measured directly by Aagard (1981) and inferred from the water mass properties by Timmermanns et al. (2005) and Björk et al. (2010, 2018). Flow across the ridge is most likely to be considerably more complex than hitherto recognized from only considering the saddle points on the ridge (Fig. 4). This complexity may be illustrated by a textbook example presented by Nikishin et al. (2021) of current deposits perched on top of, and in between, a number of local bathymetric highs at 84° N on the Siberian Segment (Figs 4, 7b, upper right).

High velocities along the uppermost flank reduce deposition and taper the accumulation towards the ridge perimeter (Bowden 1960; McCave 1984; McCave et al. 2017), which is evident for the Cenozoic sediment section in all profiles across the Central Segment and the NAS of the Lomonosov Ridge (Fig. 2). A sediment drape is dominant on the Siberian Segment (Fig. 2). Van der Krogt (2018) compiled and interpreted Parasound sub-bottom profiling data (maximum penetration 200 m) from the central Arctic Ocean. The steepest slopes along the entire length of the Lomonosov Ridge are associated with hyperbolic returns, which arise from the bottom roughness and characterize ridge-parallel areas of bottom current erosion.

Steep continental slopes have been considered to reflect the initial shear or oblique extension before break-up (Minakov et al. 2012; Evangelatos and Mosher 2016). However, relatively steep slopes, such as the Makarov Basin side of the Lomonosov Ridge (Fig. 3), are also found at rifted margins, such as portions of the US east coast (Pratson and Haxby 1996) or the Lofoten margin (Meza-Cala et al. 2021). Truncated layering, as observed on the Makarov side of the Lomonosov Ridge, is also a characteristic feature of bottom current action on slopes of sediment-starved bathymetric structures, such as the Naturaliste Plateau (Borissova 2002) and the Shatsky Rise (Clark et al. 2018).

A striking feature of the Makarov Basin side of the Lomonosov Ridge is a steep slope directly adjoining the abyssal plain with no transitional rise (Fig. 3). This is very different from the exponential shape considered to result from a balance of sediment input, transport, deposition and erosion (Adams and Schlanger 2000; Mosher and Yanez-Carrizi 2021). The slope angle and linear geometry on the Makarov Basin slope suggest that sediment input has been missing and any material generated by mass wasting on the slope has been effectively moved along by contour-following currents.

The geometry of the Cenozoic sediment deposits on the Lomonosov Ridge has the form of a drape on the broad Siberian Segment, although a flat-lying cap tapered at the edges is present on top of both the Central Segment and the NAS (Fig. 2). The latter two segments also have steeper bounding slopes. Mid- to Late Quaternary to Recent sedimentation rates along the length of the ridge (Fig. 10) vary by a factor of five, with the highest rate at the Siberian end (Nørgaard-Pedersen et al. 2007; Polyak and Jakobsson 2011). The Quaternary sediment deposition on the Lomonosov Ridge reflects the Pleistocene bottom current circulation in the Eurasia Basin, combined with upstream glacial sediment input to the Laptev Sea and Barents Sea margins. The cyclonic bottom flow is partially deflected poleward where the Lomonosov Ridge meets the East Siberian shelf (Woodgate et al. 2001) and the decreasing along-ridge sedimentation rate reflects the progressive downstream depletion of the nepheloid layer towards Canada/Greenland.

The details of the stratigraphy at the ACEX drill site on the Central Segment of the Lomonosov Ridge have been extended by acoustic signature correlation to the NAS, as shown in Figure 5, and to the Siberian Segment by Weigelt et al. (2014, 2020). The pattern of sub-bottom acoustic reflections on the NAS shows a broad similarity with the calibrated reflectivity sequence at the ACEX drill site, although distinct reflections in the acoustic stratigraphy above Unit 3 on the Siberian Segment are less clear (Fig. 5). The total thickness of the Cenozoic section on the NAS (88° N) is about half the thickness on the Siberian Segment (81° N), mostly due to a large Neogene along-ridge depositional gradient (Fig. 10).

In detail, the equivalent of Unit 3 (92 m of late Paleocene–early Eocene silty clay) appears to have a similar thickness on the Central and Siberian segments, but this is only half of that on the NAS. The overlying equivalent of Unit 2 (90 m of biosiliceous silty clay) is c. 50% thicker on the NAS in the vicinity of the North Pole compared with the ACEX drill site on the Central Segment, whereas the corresponding thickness on the Siberian Segment is not obvious from the acoustic reflection pattern (Fig. 5). There are also probable along-ridge differences in the dominant lithology because the Siberian Segment appears to lack features related to mobilized sediments within the stratigraphic interval corresponding to the biosiliceous clay of Unit 2. Evidence for the mobilization of the biosiliceous ooze is observed at six slide scar locations near the ACEX drill site on the Central Segment (Kristoffersen et al. 2007) and four gas pipe locations on the NAS (Kristoffersen et al. 2022). This suggests that terrestrial input dominates over the biosiliceous component in the sediments corresponding to Unit 2 towards the Siberian margin (early to mid-Eocene). Local uplift of the Lomonosov Ridge created a plateau north of Ellesmere Island and became an important early Eocene sediment source for local flanking deposits as well as sediment drifts to the north along the NAS (Fig. 5, left panel, star symbol).

The high-amplitude reflections in the seismic image of the lower part of the Cenozoic section (Unit 3) could be related to the transformation of opal A to opal C/T (O'Regan et al. 2010). In particular, the content of opal A is high (30–70%) throughout Unit 2 (220–314 m.b.s.f.) and in the upper part of Unit 3 (O'Regan et al. 2010). Because the down-core reduction in porosity beginning at c. 285 m.b.s.f. (the lower two-thirds of Unit 2 and the upper half of Unit 3) is not represented by any changes in sediment composition, an opal A to opal C/T phase transition has been predicted (O'Regan et al. 2010, their fig. 10). This appears problematic because the opal C/T phase only makes up <10 wt% in the interval 300–340 m.b.s.f. In addition, the phase transition is strongly dependent on temperature (Kastner et al. 1977; Hesse and Schacht 2011). Given a temperature gradient of 30.5°C km⁻¹, the Cenozoic section at the ACEX drill site would have been subject to temperatures <12°C, which is below the known level of active silica transformation (Kastner et al. 1977; Hein et al. 1978; Davies and Cartwright 2002).

With this in mind, we scrutinized the available seismic datasets from the Central Segment and the NAS of the Lomonosov Ridge for possible acoustic expressions of a phase transformation front, as documented from other settings with silica-rich sediments (Davies and Cartwright 2002; Meadows and Davies 2007). The result was negative. The lower 0.15 s thick band of high-amplitude reflections maintains its character in detail on seismic line AWI 91-090, which crosses the top of the Central Segment, including the drill site. The corresponding acoustic response on the NAS shows similar qualities and suggests any contribution from the conversion of opal A to opal C/T is not significant.

In summary, we suggest that the Siberian Segment of the Lomonosov Ridge was more accessible to clastic input during the early Eocene from the time of submergence of the ACEX drill site on the Central Segment. Biosiliceous input to the NAS became relatively more plentiful during the same time period. The large along-ridge sediment thickness gradient from the late early Miocene onwards suggests a major hemipelagic input from a Barents Sea/Laptev Sea shelf source region.

Sediments representing deposition between the mid-Eocene and the late early Miocene were not obtained at the ACEX drill site on the Central Segment of the Lomonosov Ridge (Expedition 302 Scientists 2006). The recovered section demonstrates no major change in grain size nor the total organic carbon content across the stratigraphic break, nor any indication of the reworking of older sediments above the break (Expedition 302 Scientists 2006; Stein et al. 2006; Sangiorgi et al. 2008). Two questions arise: (1) is this a hiatus or deposition at a very slow rate; and (2) how representative is this apparent stratigraphic gap at the drill site for the depositional history of other parts of the Lomonosov Ridge?

The close similarity between the acoustic reflection images at the ACEX drill site on the Central Segment and areas on the NAS strongly suggest that the drilling results are not strictly site-specific. The apparent identical sequences of acoustic reflections observed as far away as the north flank of the Alpha Ridge also support a broader environmental significance (Bruvoll et al. 2010, their fig. 16). There is, however, an important ACEX site-specific caveat that has been consistently overlooked: the drill site location (M004) is <5 km away from an area with seismic evidence of heavy erosion on the flank of the Central Segment at about the level of the apparent hiatus (Fig. 7b). Evidence of erosion at the same stratigraphic level is also present on the NAS (Fig. 7b, lower right panel). In spite of this, the geochemical data from the recovered ACEX section show no major change in the detrital source area across the hiatus (März et al. 2010).

Along-ridge variations in the sedimentation rate on the Lomonosov Ridge have also been significant. The acoustic thickness of Unit 2 and Unit 1/6 at the ACEX site is about half the thickness of the corresponding stratigraphic interval on the NAS near the North Pole c. 260 km away (Figs 5, 9). The sedimentation rate at the ACEX drill site decreases to about one-third above the hiatus – that is, from the late early Miocene onwards (Backman et al. 2008). This decrease could well have occurred closer to the mid-Eocene.

We argue that the seismic reflection data suggest an environmental change. Low sedimentation rates combined with along-ridge variations are more realistic than non-deposition in explaining an apparently incomplete stratigraphic representation at the ACEX drill site (Poirier and Hillaire-Marcel 2011; Chernykh and Krylov 2017). A complete switch-off in hemipelagic sediment input or contributions from biogenic fallout is unlikely.

The accumulations interpreted as sediment drifts on the NAS of the Lomonosov Ridge (87–89° N) represent an energy level of the Eocene bottom current circulation not reported anywhere else in the Mesozoic and younger sediment stratigraphy of the Arctic Ocean Figs 6, 7). The build-up of sediment drifts in a setting like the sediment-starved Lomonosov Ridge would involve local winnowing of fine silt to clay-sized particles and material advected in and settled out of the nepheloid layer (e.g. Rebesco et al. 2014; Gardner et al. 2017). The suspended particle load is most often concentrated in the first hundred metres above the bottom (Gardner et al. 2017), although measurements over the top of the Alpha Ridge show an increase in particles towards the bottom in the lower 500 m of the water column (Hunkins et al. 1969). Sediment drifts are mainly formed during extreme current events rather than from the average-intensity background flow (Gardner et al. 2017; Salim et al. 2018; Thran et al. 2018; Miramontes et al. 2019).

The present-day maximum flow across the Lomonosov Ridge (at an angle of 20° to the contours) was observed as three-day pulses with a maximum velocity of 12 cm s⁻¹ measured 25 m above the seabed. The pulses were hypothesized to be buoyancy oscillations trapped between the pycnocline and the bottom and align with the ridge slope (Aagard 1981; Holloway and Merrifield 1999). Bottom current speeds of this magnitude are sufficient to transport, and even erode, non-cohesive silt-sized particles (McCave 1984; McCave and Hall 2006; Gardner et al. 2017). Sediments accumulate in areas where the bottom current velocity is reduced off to the side of the main flow (e.g. Rebesco et al. 2014). The position of the sediment ridges along and inboard of the edge on top of the Lomonosov Ridge indicate significant flow obliquely across the ridge in both directions, interacting with the flow along the slopes. An important element would be Eocene sediment input to a west-flowing contour current along the Makarov Basin slope of the Lomonosov Ridge, concurrent with the deposition of sediment lenses A and B (Figs 7, 8). The Eocene sediment source was the erosion of the exhumed southern plateau of the Lomonosov Ridge. Possible sediment sources for drifts on the Amundsen Basin side of the Lomonosov Ridge are more elusive, but were most likely from upstream along-slope erosion.

We interpret the two events leading to the deposition of sediment lenses A and B adjacent to the shallow southern plateau of the Lomonosov Ridge to be a result of uplift from the tectonic far-field effects of the Greenland plate indenting the High Arctic (Døssing et al. 2014; Piepjohn et al. 2016). These lenses led us to reconsider the youngest part of the stratigraphy of a seismic profile acquired during the GreenIce 2004 project across the western flank of the same plateau (Kristoffersen and Mikkelsen 2006). Below a 0.125 s (c. 100 m) thick cover is a lens-shaped body that extends downslope c. 30 km from the top of the acoustic basement on the plateau (Fig. 8, red deposit, lower left panel). We suggest that this accumulation is equivalent to lenses A and B seen at the northern end of the same plateau.

The largest lens-shaped deposit A attained an apparent maximum thickness of c. 200 m in a short time, probably of the order of <2 Myr, corresponding to the base of Unit 3 (latest Paleocene). Sediment lens B represents a new surge in sediment supply corresponding to the basal part of Unit 2 (late Ypresian; just after Chron 22, c. 48 Ma). We suggest that the uplift and increased erosion of the southern end of the Lomonosov Ridge directly reflect peaks in the north–south component of compression during the Eurekan Orogeny. These events relate to the two distinct tectonic phases discussed by Ricketts and Stephenson (1994) and Piepjohn et al. (2016). Plateau uplift and the earliest Eocene deposition of sediment lens A (c. 55 Ma) correlate with: (1) the uplift and formation of syntectonic intermontane basins on Ellesmere Island (Fig. 7), dated as Early Eocene by the palynology of lignite-bearing sediments (Ricketts and Stephenson 1994); (2) the deposition and deformation of the clastic Margaret Formation in Stenkul Fjord in southeastern Ellesmere Island (Von Gosen et al. 2019); and (3) the episodic exhumation of northern Ellesmere Island (Vamvaka et al. 2019). In Spitsbergen, it correlates with a shift in sediment provenance from east to west represented by the Hollenderdalen Formation (Steel et al. 1985; Helland-Hansen 1990; Bruhn and Steel 2003; Sætre 2011; Petersen et al. 2016; Helland-Hansen and Grundvåg 2021).

The early to mid-Eocene (48 Ma) deposition of lens B is correlated with a pulse of NW–SE contraction, expressed by thrusting and left-lateral shear (Ricketts and Stephenson 1994; Saalmann et al. 2005; Harrison 2008). Thrusting and thermal resetting of the Kap Washington volcanics in North Greenland (49–47 Ma; Tegnér et al. 2011) appears to be contemporaneous with the uplift of the source region for sediment lens B. This is also the time of initiation of extensive volcanism at the northeastern end of the ‘Yermak block’ and the formation of a volcanic plateau at least 60 km across, making up the northern tip of Morris Jesup Spur and the Yermak Plateau (Chron 22–Chron 18, 49.5–39 Ma). This event preceded the separation of the two crustal structures at c. Chron 15 (Kristoffersen et al. 2020).

We note that the full circuit of spreading centres (the North Atlantic north of the Charlie Gibbs Fracture Zone, the Reykjanes Ridge and the Labrador Sea) defines a change in the direction of motion of Greenland relative to North America in the Eurekan tectonic domain at about Chron 21, which is compatible with the sequence of deformation events during the Eurekan Orogeny (Piepjohn et al. 2016). This change appears not to be resolved if the main emphasis is on the magnetic lineations and fracture zones in the Labrador Sea and Baffin Bay (Oakey and Chalmers 2012).

The disconformity on top of the Central Segment of the Lomonosov Ridge was first detected by refraction measurements (Demenitskaya and Kiselev 1968) and the angular discordance revealed in the first seismic reflection data across the ridge (Jokat et al. 1992). The stratigraphic break has been documented by scientific drilling as an c. 20 Myr erosional hiatus (Campanian to late Paleocene) at the drill site (Backman et al. 2008). This planar unconformity below the Cenozoic sediments on the Central Segment of the Lomonosov Ridge has been eroded in sediments with a compressional velocity of 4.0–4.3 km s⁻¹ overlying 0.7–1.5 km thick layers of 4.7–5.2 km s⁻¹ velocity. The underlying metamorphic? or igneous? upper crust has a compressional velocity of 5.8–6.3 km s⁻¹ (Jokat et al. 1995). The high-amplitude band of reflections (Unit 3 equivalent) at the northern end of the NAS is underlain by a stratified interval c. 0.1 s thick and an angular unconformity is only apparent directly below the Makarov Basin flank (Fig. 6a, line 3A). However, 70 km to the east is a base-Cenozoic unconformity present across the entire ridge (Fig. 6a, line 3B). The disconformity has an extent that includes the southern plateau and part of the northern end of the NAS, the entire Central Segment and, as documented by Sauermilch et al. (2018) and Weigelt et al. (2020), the most elevated parts on the Siberian Segment (Fig. 2, yellow line). The piece-wise reflective substratum on the Siberian Segment suggests the erosion of local sedimentary basins separated by metamorphic or igneous basement (Sauermilch et al. 2018; Weigelt et al. 2020).

No information exists of the undisturbed older stratigraphy below the NAS, except for a single sediment sample recovered during the Fram-2014/14 ice drift (Fig. 6a, line 3A, upper left side). A tentative age is Campanian–Maastrichtian (Stein 2019). The pre-Cenozoic sediment section on the NAS has intermittent high-amplitude reflection bands within the Mesozoic stratigraphy present over a distance of 600 km along the ridge segment (Kristoffersen et al. 2023). The amplitudes and geometry of the reflections (roughness, discontinuities and overlaps) suggest volcanic flows and episodes of volcanism. The attributes and implications have been detailed by Kristoffersen et al. (2023). A particular feature of the magmatic activity is the staircase pattern of lava flows displayed at 87° N (Fig. 6b). These volcanic events have been correlated with four major magmatic pulses in the adjoining High Arctic land geology because the NAS of the Lomonosov Ridge was part of the pre-Cenozoic continental margin north of Franz Josef Land (Nejbert et al. 2011; Corfu et al. 2013; Evenchick et al. 2015; Polteau et al. 2016; Dockman et al. 2018; Abashev et al. 2020).

We can use the volcanic information to arrive at a tentative age for the sediment section below the base-Cenozoic unconformity on the NAS of the Lomonosov Ridge. We assume that the basalt age information (122 Ma) from Franz Josef Land and Svalbard (Corfu et al. 2013) correlate with the oldest and most extensive volcanism on the NAS. This implies an average sedimentation rate for the overlying post-Barremian–late Paleocene sediment section to be 1.5 cm ka⁻¹ and comparable with the deposition rate for the Cenozoic part of the section (1.7 cm ka⁻¹) at the ACEX drill site (Backman et al. 2008).

For the Central Segment, we estimated the age of the sediment section below the base-Cenozoic unconformity by extrapolating the sedimentation rates. The ACEX drill site encountered the regional angular unconformity at 404.8 m.b.s.f. with a transition from the silty clay of late Paleocene age to the recovery of 1.4 m of clayey mud to silty sand of Campanian age somewhere between 404.8 and 424.5 m.b.s.f. (Expedition 302 Scientists 2006). The section below the regional unconformity dips towards the Makarov Basin with an apparent maximum thickness of 1 s TWTT (c. 2 km) below the ridge flank and a compressional velocity of 4.0–4.6 km s⁻¹ (Jokat et al. 1995). We therefore assume that the youngest prograded sediments, below the unconformity at the ridge flank on the Makarov Basin side, must have been derived from the erosion of the ridge and can be no younger than late Paleocene. Furthermore, we assume that the c. 2 km thick section below the unconformity was deposited at a rate roughly equivalent to the average rate for the Eocene part of the Cenozoic section at the ACEX drill site (1.74 cm ka⁻¹). This would have required an estimated 115 Myr to represent mid-Jurassic to Paleocene sedimentation.

Upper crustal velocities (4.7–5.1 km s⁻¹) at the northern end of the NAS (90–150° W; Fig. 5a) of the Lomonosov Ridge have been interpreted as volcanic rocks (Forsyth and Mair 1984; Funck et al. 2022). The geology of the subvolcanic part of this segment has only been sampled at 89° N (Fig. 1, site B), which yielded Caledonian metasediments deformed to the greenschist facies (Knudsen et al. 2017). The compressional velocities of the upper crust at the southern end of the NAS of the Lomonosov Ridge (Fig. 7) are 5.9–6.1 km s⁻¹ (Jackson et al. 2010). Although there is no direct relation between the compressional wave velocity and the lithology of igneous and metamorphic rocks, the seismic velocity information still yields supplementary constraints on the geology. In general, the compressional velocities in the upper crust, as documented in published seismic refraction experiments, are higher than the velocity of undisturbed sediments and fall in the range of igneous and metamorphic rocks (Fig. 11, the Lomonosov Ridge, NAS south).

Relative motion between Greenland and North America was initiated with the formation of a >200 km wide transition zone of unknown crustal composition in the Labrador Sea prior to seafloor spreading starting at Chron 27 (Roest and Srivastava 1989; Chalmers and Laursen 1995; Chian et al 1995; Srivastava and Roest 1999; Oakey and Chalmers 2012). The tectonic effects of a rift phase and pre-Chron 24 seafloor spreading are manifested in the High Arctic by Paleocene exhumation and compression in Ellesmere Island (Arne et al. 1997; Grist and Zentilli 2006; Vamvaka et al. 2019), in northeastern Greenland (Japsen et al. 2021), along the east coast of Greenland (Guarnieri 2015), along major faults in northeastern Spitsbergen (Dörr et al. 2012, 2019) and on the west coast of Svalbard (Lepvrier 1992; Teyssier et al. 1995; Blythe and Kleinsphen 1998; Jones et al. 2017). Note that Chron 24 is characterized by a positive magnetic amplitude doublet and the generation of oceanic crust had to be well underway during the preceding reverse sub-Chron 24r in order for the first positive amplitude to stand out. This suggests that seafloor spreading in the Eurasia Basin had to start around the Palocene–Eocene boundary at c. 57 Ma (Ogg 2020).

About 6 Myr after Chron 27, a new branch of the mid-ocean rift system had opened the Norwegian–Greenland Sea and propagated into the Arctic Ocean (Heezen and Ewing 1961; Karasik 1968; Pitman and Talwani 1972; Kristoffersen and Talwani 1977; Srivastava 1978; Vogt et al. 1979; Chalmers and Laursen 1995; Roest and Srivastava 1989). A deep basin c. 250 km wide (equal to the width of the present-day Red Sea) was now present beyond Chron 24 in the western Eurasia Basin (Figs 12, 13). This basin was partly floored by extended continental crust on both sides (Lutz et al. 2018; Kristoffersen et al. 2021; Funck et al. 2022). Any attempt to reconstruct the plate motion during a continental rift phase is challenged by the lack of quality fiducial markers in the continental plate boundary geology. Calculated stage poles for the pre-Chron 24 motion between Greenland and Europe will have large uncertainties, rendering the qualitative pre-break-up palaeostress approach of Guarnieri (2015) as a useful alternative.

The Eurasia Basin formed as a c. 1700 km long sliver and the Lomonosov Ridge rifted off the Barents–Kara Sea continental margin (Brozena et al. 2003; Glebovsky et al. 2006; Lebedeva-Ivanova 2010; Gaina and Jakob 2019). Seafloor spreading in the Labrador Sea gave Greenland a northward component of motion relative to North America and the tectonic effect north of Greenland was the deformation in the Canadian Arctic islands known as the Eurekan Orogeny (Thorsteinsson and Tozer 1970; Kerr 1977) and in Svalbard as the West Spitsbergen fold–thrust belt (Dallmann 2015).

The Lomonosov Ridge subsided below sea-level and sediment deposition resumed at the ACEX drill site above a ridge-wide regional unconformity in the late Paleocene (Expedition 302 Scientists 2006). The western Ellesmere and Axel Heiberg islands were the site of slowly subsiding deltas during the Late Cretaceous (Ricketts and Stephenson 1994), while tectonic activity occurred in northern Ellesmere Island in the form of minor brittle strike-slip faulting (Piepjohn et al. 2013). Most of these two islands were alluvial or river-dominated delta plains during the early Eocene sea-level highstand, with sediment transport towards the west and south. Subsequently, a number of smaller intermontane basins formed outboard of the major thrusts or reverse faults. The major deformation appears to have been over by the mid-Eocene (Ricketts and Stephenson 1994). Detailed analysis of faults and shear planes suggest that the Eurekan deformation had two distinct phases: one with predominantly ENE sinistral strike-slip tectonics (late Paleocene to Early Eocene), which was followed by NW–SE contraction during the early to mid-Eocene (Lepvrier et al. 1996; Saalmann et al. 2008; Piepjohn et al. 2013, 2016; Von Gosen et al. 2019; Brandes and Piepjohn 2021). Both phases generated local uplift of the southern plateau on the Lomonosov Ridge north of Ellesmere Island (Fig. 8).

Sediments below the base-Cenozoic unconformity on the NAS and Central Segment of the Lomonosov Ridge dip towards the Makarov Basin, but with an important difference: progradation is dominant on the Central Segment and onlap on the NAS (Fig. 2). The presence of volcanics over c. 600 km along the length of the NAS, and evidence of four magmatic pulses interpreted as High Arctic large igneous province magmatism (Kristoffersen et al. 2023), allow us to suggest the stratigraphic age of the overlying and onlapping sediment section to be post-Barremian (Fig. 6b). Sediments on the Central Segment prograde towards the Makarov Basin and extrapolations from the ACEX drill hole presented earlier suggest a post-mid-Jurassic to late Paleocene age. These ridge segments must therefore have been part of high-standing topography north and east of Franz Josef Land (Figs 12–14). The high ground, probably also including crust, later extended and down-faulted inboard of Chron 24 during the rift phase in the Eurasia Basin. The southern plateau of the Lomonosov Ridge north of Ellesmere Island had a palaeoposition north of Nordaustlandet and Spitsbergen (Fig. 13). The Cretaceous high ground, which included the plateau, a Yermak/Morris Jesup block and the adjoining margin between Svalbard and Kvitøya (exhumation 3 km) was the likely source area for Cretaceous and early Cenozoic sediment deposition on Spitsbergen and the northern Barents Sea (Fig. 13).

The high topography along the northern margin collapsed as a result of Late Cretaceous rifting, but the timing of the pre-late Paleocene collapse is uncertain. The term collapse is justified as progradation on the Central Segment below the base-Cenozoic unconformity is dominantly towards the Makarov Basin side and, to a minor degree, towards the Eurasia Basin (Fig. 2). An exception is the area near the North Pole on the NAS.

The average width of the Eurasia Basin at the time of Chron 24 was comparable with the present-day Red Sea (Fig. 13). The basin was terminated in the east by the Laptev Shelf and in the west by a transform fault (the De Geer Fault) NW of Svalbard (Kristoffersen et al. 2020). The entire Central Segment of the Lomonosov Ridge was at sea-level and was part of the Siberian Segment (Sauermilch et al 2018; Weigelt et al. 2020), as suggested by the extent of the angular unconformity at the base of the high-amplitude reflection band (the top of Unit 4). The NAS of the Lomonosov Ridge between 87 and 89° N was submerged below the wave base (Fig. 13).

The oceanic circulation would have been driven by a combination of wind forcing and the seasonal buoyancy flux related to heating or cooling (Sofianos and Johns 2003).

Today, the circulation in the largest inland seas on Earth is mainly wind-driven (Naithani et al. 2003; Troitskaya et al. 2015; Markova and Bagaev 2016). In particular, the Central Segment of the Lomonosov Ridge (the ACEX drill site) was facing the area between Franz Josef Land and Severnaya Zemlja, where the West Siberian Seaway entered the Arctic Ocean between the early late Paleocene and the end of the early Eocene (Akhmetiev et al. 2012). This is the time of formation of sediment drifts on the Lomonosov Ridge (Fig. 12). The effect of a seaway is powerful because it allows the meridional transport of water and the flux of potential vorticity essentially determines the direction of circulation in the polar basin (Yang 2005). We speculate that the water advected into the proto-Eurasia Basin through this shallow waterway was relatively fresh and had water mass properties that may not have been entirely representative of the contemporary Arctic Ocean, as is often inferred (Expedition 302 Scientists 2006; Moran et al. 2006; Sangiorgi et al. 2008; Stickley et al. 2008; Waddell and Moore 2008). The strongest boundary currents were associated with the Central Segment and the NAS of the ridge (Fig. 12). The passage formed by the Klenova Valley north of Ellesmere Island allows restricted deep onward connection with the Amerasia Basin (Figs 12, 13).

We group the observed exhumation along the polar continental margin into pre-Eurekan events (pre-seafloor spreading in the Eurasia Basin) and Eurekan events related to Greenland acting as an independent plate (Chrons 24–15), but some published exhumation age ranges overlap.

The latest Jurassic–Early Cretaceous deposits on Svalbard and the northernmost Barents Sea document exhumed areas to the north, feeding into a vast low-gradient regional landscape where global sea-level changes had a significant impact on the position of the shoreline (Steel and Worsley 1984; Maher 2001; Dypvik et al. 2002; Smelror et al. 2009; Midtkandal et al. 2019). At the end of the Hauterivian stage (132.6–126.5 Ma), the northern Barents Sea was a shallow shelf with the shoreline positioned along the western and northern perimeters of Svalbard/Nordaustlandet (Grundvåg et al. 2017) and at the southern end of the Franz Josef Land archipelago (Smelror et al. 2009). Sediment input was from the NW, with depositional lobes reaching the central Barents Sea (Grundvåg et al. 2017). Uplift in the north during the Barremian (126.5–121.4 Ma) created fluvial to paralic conditions over eastern and southern Spitsbergen and a shallow shelf embayment in the central Barents Sea. A regional hiatus, which includes Spitsbergen south of Isfjorden, is, in part, a result of forced regression (Grundvåg et al. 2017; Midtkandal et al. 2019).

A shift to sediment input from the NE indicates the increasing elevation of Franz Josef Land and its environs. The high ground to the north prevailed at least into Albian times (Dypvik et al. 2002; Smelror et al. 2009; Grundvåg et al. 2019). The erosional unconformity truncated Albian rocks at the southern tip of Spitsbergen (76° 30′ N) and Permian rocks at Ny-Ålesund (76° N) (Jochmann et al. 2019). Deposition in the Central Basin in Spitsbergen (Van Mijenfjorden Group, 2100 m thick) started in the Paleocene at 61.5 Ma (Jones et al. 2017) and the youngest Aspelintoppen Formation (Eocene) of coastal plain deposits may extend into the Oligocene (Bruhn and Steel 2003; Dallmann 2015; Helland-Hansen and Grundvåg 2021). The sediment input for the Paleocene part of the succession (700 m thick) was from the east and NE, with a shift to a western input from small catchments along the rising West Spitsbergen fold–thrust belt at the Paleocene–Eocene boundary (Bruhn and Steel 2003; Elling et al. 2016; Petersen et al. 2016; Helland-Hansen and Grundvåg 2021). This shift appears not to be associated with any erosion surface (Bruhn and Steel 2003).

A northern source area implies a positive topography and uplift along the northern continental margin from Svalbard to Franz Josef Land, including the Lomonosov Ridge in its pre-rift palaeoposition, as well as Ellesmere Island and the intervening areas, such as the Yermak/Morris Jesup Spur block (Fig. 13). Although post-Devonian sediments are absent in northern Spitsbergen, a total thickness of c. 3 km of Upper Carboniferous to Upper Jurassic strata present in central and southeastern Spitsbergen (Harland 1997; Dallmann 1999) could possibly have extended to the north (Dørr et al. 2013). The Morris Jesup Spur, a submarine plateau with a palaeoposition north of Svalbard (Fig. 14), shows truncated, west-dipping sequences of sediments and volcanics with an unknown amount of pre-Eocene material removed (Kristoffersen et al. 2021).

The magnitude and timing of erosion may be addressed by multiple methods, as outlined in several papers (e.g. Corcoran and Doré 2005; Anell et al. 2009; Henriksen et al. 2011; Lasabuda et al. 2021). The exhumation in Franz Josef Land is estimated to be 1.5–2 km based on vitrinite reflectance data from the Hayes-1 borehole in the central part of the archipelago (Sobolev 2012) and the velocity–depth function in the two other boreholes gives a minimum of 1 km (Fig. 14). Apatite fission track data from northern Spitsbergen and Nordaustlandet suggest two periods of enhanced cooling (Dörr et al. 2012). The oldest period (mid-Jurassic to Early Cretaceous) mostly affected the eastern part Nordaustlandet and northwestern Spitsbergen. The youngest period (Late Cretaceous to Paleocene) involved the exhumation of local fault-bounded blocks in the eastern part of northern Spitsbergen, where up to 4 km of sediments were removed (Dörr et al. 2019).

Blythe and Kleinsphen (1998) recognize exhumation between 70 and 50 Ma in the Central and Kongsfjorden basins in Spitsbergen and deposition in the Central Basin started c. 61.8 Ma (Jones et al. 2017). The Central Basin extended further to the north because vitrinite reflectance data from a newly discovered locality of Paleocene rocks north of Isfjorden (600 m above sea-level) indicates +2 km of burial before folding and thrusting (Jochmann et al. 2019). The subsequent Cenozoic history of the basin was influenced by the rising West Spitsbergen fold–thrust belt along the coast (Throndsen 1982; Muller and Spielhagen 1990; Blythe and Kleinsphen 1998; Bruhn and Steel 2003; Dørr et al. 2018), where the estimated net erosion is up to 3 km north of 74° N (Henriksen et al. 2011).

Exhumation on Ellesmere Island, inferred from cooling documented by apatite fission track combined with U–Th–Sm/He(AHe) thermochronology, is mainly associated with the major faults and cores of anticlines (Arne et al. 1998, 2002; Grist and Zentilli 2006; Vamvaka et al. 2019). The Eurekan exhumation was most pronounced in the Pearya terrane along the northern margin of Ellesmere Island (Fig. 12). Three separate periods are recognized (55–48, 44–38 and 34–26 Ma), with an estimated erosion within each interval of c. 2–3, 2–3 and 3–4 km, respectively (Vamvaka et al. 2019). A pre-Eurekan event with an apatite fission track age of 63 ± 3 Ma is also recognized in the Eureka Sound Group sediments, suggesting a rapid exhumation pulse of ≥0.4 km Myr⁻¹.

In northern Greenland, apatite fission track data record a mid-Paleocene (c. 60 Ma) compressional event with exhumation along the large fault systems of the Harder Fjord and the Trolle Land faults, as well as in the Wandel Sea Basin (Japsen et al 2021). This is supported by the results of strain analysis (Manby and Lyberis 2000; Guarnieri 2015; Svennevig et al. 2016).

Another approach to estimate the maximum past sediment overburden is based on the increase in the compressional wave velocity from mechanical compression and porosity reduction with increasing burial depth (e.g. Japsen 1998). The temperature becomes high enough at 2–3 km depth to initiate quartz cementation and chemical compression becomes increasingly important (Bjørkum 1996; Oelkers et al. 1996; Storvoll et al. 2005). These processes reduce the porosity and increase the velocity. The smectite content in shales also strongly influences the seismic velocity, with lower velocities for higher clay contents (Thyberg et al. 2000). Pore pressure above the hydrostatic value (overpressure) is also associated with lower velocities (Japsen 1998; Dvorkin et al. 1999). The seismic velocities on the Lomonosov Ridge and the continental margin between Franz Josef Land and Svalbard are generally higher than the velocities for marine shale, sandstone or carbonate lithologies (Fig. 12). Because the effects of compaction are largely irreversible, a likely explanation for the high velocities at shallow levels may be exhumation after undergoing maximum pre-Cenozoic burial.

The strata on the Siberian Segment (83–84° 30′ N) below the equivalent of the early Eocene and younger section penetrated at the ACEX drill site are mildly deformed sediments with velocities of 2.3–4.0 km s⁻¹ or higher (Jokat 2003; Sauermilch et al. 2018; Weigelt et al. 2020). We may also extract a crude velocity–depth function from the Transarctica-1992 line at 83° 30′ N, as given by Poselov et al. (2012, their fig. 2). The vertical difference between the smoothed velocity gradient from the seismic reflection/refraction results and the gradient for marine shale or sandstone lithologies suggests a possible exhumation of up to 3 km for the northern part (82° 20′–85° N), but insignificant uplift in the southern part (80–81° 30′ N) (Fig. 4, two upper left panels).

The Cenozoic section on the Central Segment of the Lomonosov Ridge is unconformably underlain by rocks with seismic velocities >4 km s⁻¹, as recorded by three sonobuoys (Jokat et al. 1992). The velocity–depth trends below the unconformity are shallower than generally expected for shale or sandstone lithologies and may be explained by up to 3 km of pre-Cenozoic exhumation (Fig. 14).

The LOREX refraction line near the North Pole shows a refracting interface of 4.7 km s⁻¹ at c. 1 km depth (Fig. 14, lower left velocity panel). Recent seismic reflection and refraction data suggest that this represents basalt (Funck et al. 2022). If we exclude this interface, then the remaining velocity–depth relation still suggests exhumation of the order of kilometres. At the southern end of the NAS, the LORITA wide-angle refraction and reflection line (Jackson et al. 2010) suggests insignificant exhumation for a predominant sandstone lithology and of the order of <3 km for a shale lithology (Fig. 14).

The late Mesozoic tectonic history of Severnaya Zemlya in the east is unknown (Lorenz 2005). The bedrock consists of Neoproterozoic–Permian strata. Only a small remnant of Cretaceous sediment is present on October Revolution Island.

The Mesozoic section on Franz Josef Land is up to 2 km thick with three major breaks (Embry 1992; Dibner 1998). A mid-Jurassic break (Toarcian, 190–180 Ma) separates lower Jurassic and older sandstones from the overlying shale–siltstones. A base-Hauterivian (132–127 Ma) unconformity separates lower Cretaceous sandstone from the overlying c. 500 m thick section of subaerially erupted volcanics. The volcanism ended in the Albian and on top of the volcanics is a 50 m thick sandstone of Cenomanian age. Dibner (1998) gives velocity and lithological information for the Nagurskaya (Nag) and Severnaya (Sev) boreholes on the easternmost and westernmost part of Franz Josef Land, respectively (Fig. 14, upper right panel). The velocity–depth values are lower than on the Lomonosov Ridge, except for short high-velocity intervals related to basalt or carbonate rocks. The sediments are mostly mudstone or silty mudstone, which suggest that post-Early Cretaceous exhumation of >1 km may have taken place. Vitrinite reflectance measurements on samples from the Hayes-1 borehole about midway between these two sites are consistent with 1.5–2.0 km of overburden being removed (Sobolev 2012).

The outer continental shelf north of 80° N and west of 30° E between Franz Josef Land and Svalbard has a sediment thickness of up to 3.5 km above the acoustic basement (Geissler and Jokat 2004; Lasabuda et al. 2018). The entire section is considered to be of Cenozoic age. Velocity–depth information for the upper 4 km have been obtained from sonobuoy measurements (Geissler and Jokat 2004) and form separate envelopes for the outer and inner shelf environments (Fig. 14, right panels). The deposits above the acoustic basement (4.5–5.0 km s⁻¹) on the outer shelf are characterized by a velocity–depth function that closely follows the gradient of a marine shale lithology. The sediment thickness below the inner shelf >30 km from the present shelf edge is <1 km and the acoustic basement velocities lie in the range 5–6 km s⁻¹.

The area between Franz Josef Land and Spitsbergen was a shallow marine carbonate platform during several periods between the late Carboniferous (Muscovian) and the mid-Permian (Wordian) (Smelror et al. 2009; Dallmann 2015). High compressional velocities (4–6 km s⁻¹) at shallow depth may therefore, in part, relate to the presence of carbonate-dominated sediments (Anselmetti and Eberli 2001). In this case, any underlying shale or sandstone lithology would present a velocity inversion that would be difficult to detect in unreversed sonobuoy measurements. As a consequence, the velocity–depth data from the inner shelf are basically inconclusive, but not inconsistent with subaerial exhumation of several kilometres (Fig. 14).

In summary, the high seismic velocities at shallow depth over a 400–500 km wide swath along the Mesozoic polar continental margin between Svalbard and Franz Josef Land, including the Lomonosov Ridge, Ellesmere Island and Morris Jesup Spur, suggest the removal of several kilometres of pre-Cenozoic overburden during a pre-Eurekan period, which we relate to a rift stage preceding seafloor spreading in the Eurasia Basin (Fig. 14). The eroded topography to the north became a major sediment source area for Svalbard and the northern Barents Sea during the Early Cretaceous in a sector between the NE and NW, as seen from Svalbard. The uplift along different parts of the northern margin may have varied with time, as reflected in shifts in the direction of sediment input to Svalbard. The high topography also shed sediments to the north towards the Makarov Basin (Figs 2, 14).

The seismic data calibrated by scientific drilling on the Central Segment of the Lomonosov Ridge show that the Central Segment had been above sea-level, but was levelled by erosion by the end of the Paleocene (Expedition 302 Scientists 2006). Similarly, the NAS had been above sea-level near the North Pole, near sea-level between 88° 30′ N and 87° 30′ N, and subaerial south of 87° N (Fig. 14). The Siberian Segment of the Lomonosov Ridge was probably a string of narrow crustal blocks that were eroded and eventually formed a string of islands with shallow water in between (Fig. 14). Uplift and rifting along the northern continental margin and subsequent separation of the Lomonosov Ridge enhanced erosion, followed by a reduction in the source area as the Eurasia Basin formed. The exhumation during the subsequent Eurekan Orogeny was largest in northern Ellesmere Island and northern Svalbard.

The Lomonosov Ridge is a continental sliver that divides the Arctic Ocean into a Cenozoic part and a Mesozoic part. The ridge elevation above the adjacent abyssal plains is equivalent to the average height of the European Alps (2.5 km) or the deepest part of the Gulf of California. The presence of a submarine ridge was first inferred from its effect as an oceanographic barrier in the polar ocean.

The top of the Lomonosov Ridge has been sediment-starved since the earliest Cenozoic and restricted to hemipelagic input advected in by bottom currents and deposited out of the nepheloid layer. The total thickness of the Cenozoic section on the NAS (at 88° N) is about half the thickness on the Siberian Segment (at 81° N), mostly due to a large Neogene along-ridge depositional gradient. The thickness of sediments equivalent to Unit 3 (late Paleocene–early Eocene silty clay) on the NAS is about half the thickness of the corresponding units on the Central and Siberian segments, whereas the overlying Unit 2 (biosiliceous silty clay) is thicker (50%) on the NAS than at the ACEX drill site on the Central Segment. The thickness of the equivalent of Unit 2 cannot be reliably determined for the Siberian Segment and the lack of features related to mobilized sediments suggests a reduced biosiliceous component towards the Siberian continental margin.

Increased cross-ridge bottom current flow starting in the earliest Eocene is documented by the presence of sediment drifts on top of the ridge along parts of the ridge flanks. Drift accumulations are observed on the NAS and also on the Amundsen Basin side of the Central Segment of the Lomonosov Ridge next to the ACEX drill site. The drifts on the Makarov Basin side were sourced from the south by the erosion of a plateau exposed by two ridge uplift events north of Ellesmere Island during the early to mid-Eocene. We suggest that increased oceanic circulation coincided with a seaway connection to Tethys and later to the West Siberian Basin during the Thanetian–Ypresian stage. The alternative of a regional environmental change, with a very low rate of sediment deposition rather than non-deposition, during the mid-Eocene to late early Miocene interval is supported by the striking similarity between the acoustic patterns at the ACEX drill location and the pattern on the NAS, as well as on the north flank of the Alpha Ridge.

We used the seismic velocity gradients in the pre-Cenozoic rocks and assumed a dominant siltstone lithology to obtain estimates of exhumation in the range 1–3 km along parts of all three segments of the Lomonosov Ridge. We infer that the Lomonosov Ridge was an integral part of a Late Jurassic and Cretaceous elevated northern continental margin and northern sediment source region extending from Severnaya Zemlya to North Greenland and Ellesmere Island.

The logistic support provided by A. Tholfsen, the Alfred 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 this research possible. A.J. Bugge generously contributed to the neural network approach and J.E. Lie advised on the seismic processing details. I. Gimse of Magseis Fairfield kindly provided a seismic node for observations of the seismic source signature. The enthusiasm of G.B. Larsen, H. Jahre and H. Brekke was crucial 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 A. Embry, M. Rebesco and an unknown reviewer, who improved the presentation of our results.

YK: conceptualization (lead), data curation (lead), formal analysis (lead), investigation (lead), methodology (lead), project administration (lead), writing – original draft (lead); JKH: conceptualization (equal), funding acquisition (equal), resources (lead), writing – review and editing (supporting); EHN: formal analysis (equal), methodology (equal), software (lead), visualization (supporting).

This work was funded by Lundin Energy Norway and Oljedirektoratet. 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 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.