Introduction

Intraplate volcanism is defined as magmatic activity away from plate boundaries and without a clearly observable link to tectonic processes1. In contrast to passive mantle upwelling observed at mid-ocean ridges1,2, mechanisms behind the source of intraplate volcanism have included those related to regional mantle plume activity originating from the mantle transition zone (~410–660 km depth) or the mantle-core boundary region (~2900 km depth). Other models have argued for local processes such as lithospheric cracking and melting of chemically heterogeneous mantle3, small-scale mantle convection4, and/or melting of the asthenospheric mantle induced by shear deformation at the base of the lithosphere5. These models, or variations thereof, could explain the late Cenozoic intraplate volcanic activity within the Northeast (NE) Atlantic passive margins6,7,8,9,10 (Fig. 1). Mid-Cenozoic (<35–5 Million years ago; Ma) intraplate volcanism has also affected the NE Atlantic oceanic basins during various time intervals and tectonic changes, pointing towards complex geodynamic interactions between magmatic and tectonic processes2,11.

Fig. 1: NE Atlantic – Arctic regional setting with areas of postulated Late Cenozoic intraplate volcanism.
figure 1

Late Cenozoic (<35-25 Ma) intraplate volcanism adjacent to and within Svalbard from refs. 6,8,10,33; Interpreted seamounts in NE Atlantic from ref. 11; High Arctic Large Igneous Province (HALIP) from ref. 17; North Atlantic Large Igneous Province (NALIP) from ref. 43; Depth to lithosphere-asthenosphere boundary (LAB) and Moho temperature contours from ref. 54; Extent of Arctic Azolla from ref. 27. ACEX and ODP drilling sites, and location of used seismic reflection profiles (1-4) are indicated (see also S1). EVZ/WVZ: Eastern/Western volcanic zone, SMZ: Sparsely-magmatic zone. Box in bottom-right indicates a simplified Sr-Nd isotopic trend of Late Cenozoic basalts within the studied region (see also S6). AB Amundsen Basin, Bjø Bjørnøya, GR Gakkel Ridge, JMMC Jan Mayen micro-continent, LT Lena Trough, MJR Morris Jesup Rise, MS Mosby Seamount, NB Nansen Basin, SB Sophia Basin, S-N/Q Neogene/Quaternary basalts in Svalbard, YP Yermak Plateau. Inset map: Arctic Bathymetry (IBCAO) with areas affected by Late Cenozoic intraplate volcanism. Red box indicates area displayed. Other abbreviations/explanation in text and “Methods” section.

Full size image

Here, we use and analyze a compilation of 2D multi-channel seismic reflection profiles combined with available magnetic anomaly data in the southwestern (SW) Eurasia Basin region, located to the north of Svalbard and the Barents Shelf (Fig. 1). We focus on the tectono-magmatic processes occurring in this region since the early Cenozoic. Based on these data, we document late Cenozoic magmatic events that have affected the oceanic crust and the sediments in the SW Eurasia Basin. The resulting seismic database is a compilation of available reflection profiles collected over the last decades during various marine seismic campaigns (S1), including recently released profiles (i.e. NPD-1204-402, S2-5; and AWI-20130155; S6). We perform the interpretations using existing seismic-stratigraphy information tied to scientific boreholes. In doing so, we constrain the chronology of magmatic events postdating the emplacement of seafloor in the SW Eurasia Basin (Fig. 1 and Fig. 2 profiles i-ii; see “Methods” section for details). Magnetic anomaly data of the Arctic region12,13 together with compilations of geochemical data (S7) and mantle seismic tomography models (S8) were used to support the findings from the seismic reflection interpretation (see “Methods” section and Supplementary Information). Hydrothermal vents, chimneys, and evacuation structures imaged by seismic reflection data, and their morphology and location suggest relatively young magmatic origins, most probably no older than ~10 Ma. These findings suggest that intraplate volcanism, which occurred within continental Svalbard and offshore areas to the north, affected a greater area than previously anticipated. The geodynamic context and a discussion of possible triggers of these discovered volcanic features as well as their role during hydrothermal venting formation are presented below.

Fig. 2: Acoustic and magnetic signatures of the oceanic basement and sedimentary structures in west Eurasia Basin.
figure 2

Expanded area in block plot is the magnetic anomaly map (CAMP compilation; ref. 13) with main oceanic crust structures as in Fig. 1. Geoseismic composite sections with seismic stratigraphic units and ties (2.6 Ma, 6–7/10 Ma, and 48 Ma Azolla markers) are shown as profiles i and ii, while selected portions of seismic reflections profiles are labeled as 1 to 3 and 4 (in Fig. 4). The magnetic anomaly and interpreted chrons (C2A to C24) based on CAMP grid (solid line) and NRL tracklines (stippled line) (refs. 12,13) are shown on top of Profiles i and 1(see also S2). Late Cenozoic magmatic additions are indicated with pink ellipses, including the region of observed VEHAI reflections. Seismic profile 2 from ref. 16 and profile 3 from ref. 17. See “Methods” section for details on the workflow to construct the geoseismic composite profiles. TAB Top Acoustic Basement. Other abbreviations as in Fig. 1. Overview of the region and profile locations are shown in the inset map.

Full size image

Results and Discussion

Late Cenozoic magmatism in the Eurasia Basin

The Eurasia Basin developed as the result of continental breakup north of the Barents and Kara continental shelves and subsequent seafloor spreading since the Paleocene, ~56 Ma14. The Nansen and Amundsen basins constitute the two conjugate sub-basins of the Eurasia Basin, and are situated to the south and north of the active ultraslow spreading mid-ocean Gakkel Ridge, respectively15,16 (Fig. 1 and Fig. 2 profile i). Magnetic chron 24 (C24) represents the oldest seafloor spreading anomaly unambiguously identified along both conjugate oceanic basins12,15,17 (Figs. 1 and 2). Our study area is in the proximity of this oldest oceanic crust of the SW Eurasia Basin, with a focus on the Nansen Basin flank adjacent to the Svalbard continental margin. Within the seismically observed passive margin sedimentary infill, we observe distinct and very-high acoustic impedance (VEHAI) seismic reflections in western Nansen Basin located below 6.8 s two-way traveltime (twt) at the base of sedimentary units (Fig. 2 profile 1). In contrast, the contemporaneous top acoustic basement further east within the Nansen Basin, and on the conjugate side within the Amundsen Basin, consists of smooth and rounded shapes without high-amplitude reflections (Fig. 2 profiles 2 and 3). The seismic features within and above the top acoustic basement in the SW Nansen Basin (i.e. VEHAI) exhibit transgressive or branched traits, and saucer-like shapes, typical of sill intrusion complexes (e.g. ref. 18) (Fig. 3a). The estimated dimensions of these complex reflections range from ~0.8-2.5 km in length and ~0.13-0.33 km in thickness, assuming a P-wave velocity of ~6.5 km/s. The VEHAI complex truncates otherwise coherent seismic reflectivity, and together with its undulating shape, hints at masked seismic units located below (Fig. 2 profile 1 and Fig. 3a; see also S5).

Fig. 3: Fluid evacuation features related to Late Cenozoic intraplate volcanism in SW Eurasia Basin.
figure 3

Enlarged selected portions of seismic data along NPD-1204-402 (Fig. 2 profile 1). a Deformation within basement and sedimentary structures, including the identified Mound-A and the Tayrona Vent, as well as Body-A due to sill intrusions and hydrothermal vent complexes associated to the VEHAI. b, c Mound-A and Tayrona Vent seismic morphologies. d Seismic chimney complex accompanied by velocity push-down effects from basement levels reaching the seafloor and aligned to seabed like-hill features. Shift in seismic facies is indicated by arrow position (See text). See also S3-5 for more details.

Full size image

We identify a triangular mound feature, informally named Mound-A, which rests on top of the base of seismic unit NB-3 (~6.3 s twt; ~10 Ma) (Fig. 2 profile 1 and Fig. 3a). Mound-A has stronger seismic amplitudes compared to adjacent reflections, stands ~0.35 km tall at its longest segment, and has a base width of ~3 km (Fig. 3b). Below the mound, we observe distorted reflectivity patterns, reminiscent of deformation, and vertical propagation of climbing sill intrusions (ref. 18) (Fig. 3a). Mound-A is internally built by stacked, double-verging, downlapping reflections, typical of volcano-sedimentary edifices with multiple growth phases (e.g. Bight Basin, offshore southern Australia ref. 19 and Gulf of California ref. 20) (Fig. 3a, b). The seismic data (Fig. 2 profiles 1 and 2; see also S2-5) further highlight disturbances in the acoustic response as potential fluid escape conduits, vents, or chimneys, indicated by reduced similarity between traces and total trace energy within the interpreted conduits (Mound-A summit; Fig. 3a, b). One of them originates from the tip of a saucer-shaped reflection complex cross-cutting and deforming the area within Mound-A (Fig. 3a, b). The association with conduits, and possibly contemporaneous faults, corresponds to the location of a fluid escape feature, here informally referred to as the Tayrona Vent (Fig. 2 profile 1). Its main feeder pipe reaches the base of a surface that has an inferred age of sediments younger than ~2.6 Ma, and has dimensions of ~0.4 km vertically and ~2.5 km horizontally. From 2D seismic data, the Tayrona Vent has a dome shape crater-hosted mound structure with several intra-vent reflections truncated against its walls (Fig. 3c). These features are typical of hydrothermal vent complexes identified in other volcano-sedimentary regions in the NE Atlantic18,21, Mediterranean (e.g. ref. 22), southern Australia (e.g. refs. 19,23), and South Africa (e.g. ref. 24). Body-A, a submarine paleo-slide25 (Fig. 2 profile i), displays a more subdued and chaotic reflection amplitude but retains layered units (Fig. 3a). These packages appear thrusted and imbricated into the southern side of the Tayrona Vent (Fig. 3a–c).

Nature and origin of fluid escape features

The interpreted seismic features indicate strong fluid evacuation restricted to the SW Eurasia Basin (Fig. 2 profile 1 and S3). Some of these fluid-escape features reach the present-day seafloor (Figs. 3d and S3, 4). We favor a late Cenozoic sill intrusion scenario since it can explain their visiblity on seismic reflection data (see S5): hydrothermal vent systems emplaced in organic-rich sedimentary basins must develop within a few hundred years of the related sill intrusions i.e., before the magma cools down18,26. We postulate that the widespread and thermally immature Arctic Azolla interval was deposited27 only after this late Cenozoic intraplate volcanic activity occurred (Figs. 1 and 2; see “Methods” section for details). The Eocene (~49 Ma) in age Azolla organic-rich layer serves as a candidate source rock in the Arctic Ocean, including the SW Eurasia Basin28, for generating the identified vents and fluid features observed on seismic reflection data (Fig. 4). Late Cenozoic intraplate volcanism has also been observed in the Sophia Basin (SB in Fig. 1), an engulfed depocenter located offshore northern Svalbard and adjacent to the SW Eurasia Basin. Its basement consists of magmatically-reworked extended continental or transitional crust (i.e. continent-ocean transition, COT; refs. 14,29) overprinted by Miocene volcanic activity (~13 Ma Mosby Seamount; ref. 30) (Figs.1 and 2 profile ii). The acoustic basement and sedimentary structures in the Sophia Basin show clear signs of volcanism via high-amplitude reflections indicating sills or igneous bodies intruding before and during ~10 to 7-6 Ma and until ~2.6 Ma below undisturbed and syn-volcanic deposits, respectively (Fig. 4a1-a2; see also S6). This multi-stage volcanism seems to correspond to the development of Mound-A and the Tayrona Vent at the maximum ages of ~10 Ma and ~2.6 Ma, respectively (Fig. 4b1-b4). Other regional seismically-imaged hydrothermal vents with similar proposed ages and formation mechanisms have been previously suggested, namely on the Lomonosov Ridge31 (“Pockmarks” in Fig. 1) and within the Vestbakken Volcanic Province (VVP) SW of Bjørnøya21 (Fig. 1).

Fig. 4: Schematic evolution of basement and sedimentary structures in northern Svalbard and SW Eurasia Basin due to Late Cenozoic intraplate volcanism.
figure 4

a Seismic profile 4 (AWI-20130155; S6) segments (a1 and a2) are located within the Sophia Basin (SB in Fig. 1). Sills and igneous additions are observed modifying the acoustic basement and deforming sedimentary reflections below (post-deposition; a1) and above (syn-volcanic doming; a2) the 6-7/10 Ma marker. For location of Profile 4 see Figs. 1, 2, and S1. YP-2/3 represents stratigraphic seismic units in the Yermak Plateau. See more information about seismic survey and seismc stratigraphy in Methods. b Cartoon showing the proposed evolution in the SW Eurasia Basin with respect to recent igneous activity observed in adjacent offshore areas and onshore northern Svalbard: (b1) Stage 1 corresponds to undeformed area and thermally immature Azolla unit. b2 Stage 2 corresponds to deformation of basement due to first igneous intrusion event accompanied with early Mound-A formation. b3 Stage 3 consists of a second event of magmatism with climbing sills and associated deformation within the sedimenatary structure, as well as early Tayrona Crater formation. b4 Stage 4 corresponds to continuous hydrothermal circulation across the entire sedimentary successions, including, locally, a thermally mature Azolla unit. VEHAI (very-high acoustic impedance) seismic reflections.

Full size image

An alternative interpretation of the acoustic basement in the SW Eurasia Basin is that of exhumed and serpentinized oceanic mantle driven by the tectonic mode of slow seafloor spreading at the Gakkel Ridge, with a minor addition of basalts at the top of rotated blocks32. Magnetic anomaly data compilations within the Arctic region have been used to address the nature of the oceanic crust in Eurasia Basin12,13,15,33. This data presents a ~2800 km2 semi-circular region in the SW Eurasia Basin where the western terminations of seafloor spreading anomalies of chrons C24 to C21 (i.e. Eocene; early Cenozoic) are disrupted and coincide with the mapped VEHAI (Fig. 2 and S2). Existing seismic reflection profiles in the same area reveal an elevated basement block, previously interpreted as exhumed mantle34, yet the top acoustic basement presents VEHAI-like reflections with associated seismic fluid escape features like the ones presented here (e.g. Fig. 3) disturbing younger mid-Cenozoic (i.e. <35-23 Ma) sedimentary units. In the SW Eurasia Basin, oceanic upper mantle serpentinization products have recently been considered as a mechanism in hydrothermal circulation and the generation of seismic fluid indicators in regions located between spreading anomalies C23 to C11 (ref. 35). In contrast to an active sill intrusion scenario, our analysis instead suggests that these features correspond to a passive scenario whereby fluid migration occurs above the crest of smooth and rounded basement highs, which are in turn driven by sedimentary faults and normal compaction processes (see seismic flat spot in Fig. 2 profile 2). The nature and seismic responses of the Tayrona Vent and other fluid escape features imaged by our data support an active scenario of magmatic sills intruding immature early Cenozoic Azolla-rich sediments during the late Cenozoic (Figs. 3 and 4). In addition, the presence of late Cenozoic magmatic sills beneath the Tayrona Vent area may explain observed deformation within the oceanic basement and the surrounding areas (Figs. 3a and S5). This deformation, in the form of faulting, is a fundamental mechanism that would lead to sustained fluid circulation after sill intrusion24. Similar fault-sill-fluid interactions have been documented in other regions with hydrothermal vent complexes (e.g. Mediterranean22, offshore southern Australia23,36).

Hydrothermal systems due to recent intraplate volcanism

Neogene (23–2.6 Ma) and Quaternary (<2.5 Ma) intraplate volcanism with associated hydrothermal systems have been observed elsewhere surrounding the SW Eurasia Basin. These include, Quaternary volcanoes and active thermal springs onshore North Svalbard8, and offshore as volcanic centers in the Sophia Basin (i.e. ~13 Ma Mosby Seamount; ref. 30), as well as areas further south in the Bjørnøya region within the Vestbakken Volcanic Province (VVP)10,21 (Figs. 1 and 4). Geochemical data derived from the basalts extruded at various locations show similar patterns in trace elements compositions. The light incompatible elements present signatures typical of E-MORB and OIB basalts (S7), suggesting an enriched mantle source and/or sub-continental lithospheric components10,37,38,39. Isotopic data (S7), on the other hand, have opposing Sr-Nd trends between basalts in VVP and North Svalbard10,40, indicating magmatic mixing and early-stage enrichment in trace elements concentrations with respect to the source of basalts along the central Lena Trough and Gakkel Ridge38,39,41 (Fig. 1). Therefore, the on-axis Gakkel Ridge basalts possibly originate from a more depleted magmatic source than the intraplate Neogene/Quaternary basalts on Svalbard and at the Mosby Seamount (Fig. 1). Intermittent volcanic activity (~35 Ma and ~10 Ma) at the Yermak Plateau and Morris Jesup Rise31,33,42, partially-volcanic ridges located north of Svalbard and Greenland, respectivelly – could be related to an elevated mantle temperature anomaly (i.e. “Yermak hotspot”; ref. 7). Such a hotspot could be accounted for by a present-day low seismic velocity anomaly observed in the upper mantle beneath NE Atlantic-Arctic region, extending to at least ~300 km depth43,44. This mantle-derived anomaly may have existed at earlier times and therefore could have facilitated the Miocene magmatism described here (i.e. at ~10 Ma; Fig. 5 and S8) and may explain the observed first-order bathymetric/subsurface structure and magnetic features (e.g. refs. 27,29,37,45) of the Yermak Plateau-Morris Jesup Rise system (Fig. 2). In this sense, our studied area might represent the effects of a mantle plume feature analogous to those modeled for the East African (e.g. ref. 2) and Afar-Red Sea rifted margins (e.g. refs. 46,47,48). Another example of apparently similar geodynamic setting is the SE Australia-Tasmania passive margin, where the volcanism is probably related to the thinning of the lithosphere due to interactions with a convective oceanic mantle asthenosphere (e.g. ref. 49). Furthermore, melting of a heterogeneous upper mantle – could also explain the enriched trace element patterns in basalts recovered adjacent to and within Svalbard41,50 (Figs. 1 and S7).

Fig. 5: Conceptual geodynamic representation of Late Cenozoic and recent volcanically driven hydrothermal activity within northern Svalbard and SW Eurasia Basin.
figure 5

a, b 10 Ma and present-day (0 Ma) absolute plate reconstructions11 that overly the present-day 250 km and 56 km horizontal slices of the seismic tomography model NAT2021 (ref. 43), respectively (see also S8). The slow/negative S-wave velocity anomalies (contoured with dotted lines) at this depth may indicate the higher-than-normal mantle temperatures linked to the mantle upwelling that affected the study region in the Miocene. Possible areas of lateral mixing of the oceanic asthenosphere with sub-continental lithospheric components are schematically indicated and should correspond to a region at 30-40 km depth. c Hot asthenosphere along the western Gakkel Ridge and Yermak Plateau regions ascends and underplates the sub-lithospheric mantle domains of northern Svalbard/Barents Sea and SW Eurasia Basin. A possible interplay between the lateral flow of the hot mantle asthenosphere, the thinned lithosphere, and detachment fault systems leads to intraplate volcanic activity with focused hydrothermal activity. The crustal transects depict a recent (<3 Ma) setting where simultaneous uplift of the shelf due to volcanism favors erosion and progradation of sedimentary units. Transects and structural features in (i) and (ii) are adapted from refs. 14,32,53. The inset map in bottom left shows the location of the main features relative to the present setting of the study area. Other abbreviations as in Fig. 1 and details in text and “Methods” section.

Full size image

According to other geodynamic scenarios, the observed trace elements concentrations in the Gakkel Ridge samples41,51 could have been attained due to the melting of mantle affected by earlier subduction of a Cretaceous (~120 Ma) subducted slab, now postulated to be located under Greenland52. In this way, the enriched basalt compositions with anomalous trace elements concentrations, positive lead-Pb signatures and arc composition with an addition of recycled crustal components can be explained51. However, the involvement of subduction-related melts has been deemed not likely on Svalbard37,40 despite the Eurekan orogeny phase during the Eocene when Greenland’s northward motion may have resulted in lithospheric impingement and shortening in the SW Eurasia Basin12. Alternatively, mantle xenoliths in Quaternary basaltic lavas on Svalbard indicate young (<2.6 Ma) metasomatic reactions40. Our interpretations, although limited by the narrow collections of sampling and data, support that the enrichment in trace elements concentrations, including Pb, in the late Cenozoic basalts in North Svalbard region must have been obtained via percolation of magmatic fluids during basaltic eruptions37,38. The intense hydrothermal activity suggested by seismic reflection data (Figs. 3 and 4) is interpreted here as direct evidence of such mechanisms operating at the northern Svalbard passive margin since at least ~10 Ma (Fig. 5). Interpretation of seismic reflection data indicates detachment fault systems and exhumed mantle during Early Cenozoic continental breakup within the SW Eurasia Basin and surrounds32,34,53. We propose that these faults were later used as pathways for magma transport at a shallower crustal levels, and may have assisted the formation of VEHAI structures atop a synchronously deformed oceanic basement (Fig. 5 and S5). These faults can also help seawater to percolate to depths of ~30-40 km38,39 and lower the temperature of partial melting by ~200-300 °C9 within the region of thinned lithosphere6,54 (Figs. 1 and 4). These processes could have facilitated the formation of favorable conditions for hydrothermal circulation across the sedimentary successions of Svalbard and the SW Eurasia Basin, as the ones identified in the offshore area, the Mound-A and Tayrona Vent (Fig. 5).

Implications for Arctic tectono-stratigraphy

The late Cenozoic volcanism described in this study is located close to the Fram Strait, the oceanic gateway between the Arctic and NE Atlantic which formed during mid-Miocene times (~20–15 Ma; e.g. refs. 45,55) west of Svalbard. Signs of late Cenozoic intraplate volcanism in sedimentary units of the SW Eurasia Basin might have been accompanied by a shift in seismic-stratigraphic facies from the base of seismic unit NB-3 (with a maximum age of ~10 Ma) and upwards (Figs. 3 and 5). The shift in seismic facies consists of stacked, northerly prograding seismic sequences that transition to gentler-to-flatter dips (<3°) towards anomaly C24 and further oceanward with sets of piled, imbricated, chaotic, and acoustically brighter units (e.g. Body-A in Fig. 2 profile 1 and S2–4). Similar facies changes in acoustic data have been observed in the western Amundsen Basin at ~18-10.6 Ma (units AB-4-5; ref. 56), whereas in the western Nansen Basin the changes are visible in sedimentary units of ~20-16 Ma age31 (Fig. 2 profiles i and 1-3). Despite the lack of clear constraints on early icehouse conditions in the Eurasia Basin, these shifts in seismic facies are interpreted to indicate the onset of glaciations across the High Arctic region31,56. The latest Miocene-Quaternary conditions on the continental shelves of Svalbard and the northern Barents Sea that resulted in glacial sedimentation-erosion cycles may have started at ~6 Ma55,57,58. This may have been preceded by a thermal uplift of the Svalbard region lithosphere as postulated by several authors (e.g. refs. 6,9,44,54,59,60) (Fig. 5). This late Cenozoic uplift episode, supported by geological data, may have been triggered by the same thermal anomaly observed in the present-day upper mantle, which we ascribe to the magmatic activity and changes in sediment deposition in the SW Eurasia Basin described by seismic reflection data.

Methods

Seismic dataset and stratigraphic framework

Sedimentary age control in the Eurasia Basin remains uncertain due to lack of litho/biostratigraphic ties and calibration from boreholes. The Arctic Coring Expedition (ACEX) located on the Lomonosov Ridge near the North Pole (Integrated Ocean Drilling Program, IODP Expedition 302; Fig. 1), retrieved cores down to uppermost Cretaceous units (>65 Ma; ref. 61). Drilling along the flanks of the Yermak Plateau yielded sediment records back to the Late Miocene (~6 Ma; Ocean Drilling Program, ODP sites 910 C, 911 A, 912 A; ref. 57) (Fig. 1). We constructed two composite geoseismic sections (i.e. Figure 2 profiles i and ii; see also ref. 62) using seismic data from earlier expeditions in the western Eurasia Basin (Supplementary Fig. 1, S1), together with seismic stratigraphy previously established by maximum age relationships between onlaps observed in reflection profiles and magnetic anomaly-derived seismic top basement ages, and sonobuoy data for sediment velocities (i.e. ARCTIC’91: AWI-91-090/097/098/100/101/102, ref. 63; AMORE and ARKTIS-XV/2: AWI-99176/20010100, ref. 64, AWI-99165/99166: ref. 29; AWI-20020300: ref. 65; ARK2004: AWI-2004-0005/0100, ref. 42; MSM31: AWI-2013-0130/0140/0141/0155/0160, ref. 66; NPD-POLAR-16/17, ref. 25; PMGRE: NP-28, ref. 67). Additional data were used to complement the observations of this study, including the IBCAO grid (v.4.2, ref. 68), oceanic crust structures and features (from refs. 69,70,71), Cenozoic domes (modified from ref. 72), in combination with other reflection profiles (i.e. Arktika: ARC11-003/004/005/006, ref. 16; PANORAMA-1: BGR-13-206/207, ref. 34, BGR-13-201/202/204, ref. 32; BGR-18-112: ref. 73; LOMROG: ref. 17; NPD-BARMOD: ref. 74). We implemented the stratigraphic framework of ref. 25 for mapping seismic units as it presents the latest seismic stratigraphy with good ties to existing profiles in focus study area. Two distinct seismic horizons are constrained in the Yermak Plateau, one of 2.6 Ma and another of 6–7 Ma (refs. 29,57; and seismic horizon MB07 – 6.4 Ma – tied to ODP 909, ref. 65), that correlate towards the Eurasia Basin side as 2.6 Ma and 10 Ma respectively (ref. 25) (profiles i and ii; Fig. 2).

Seismic interpretation

The recently released seismic reflection profile NPD-1204-402 acquired by the Norwegian Petroleum Directorate (NPD, now Norwegian Offshore Directorate; Sokkeldirektoratet - Sodir) offers a unique opportunity to visualize the SW Eurasia Basin’s sedimentary architecture and basement structure (Fig. 2 profile 1 and S2-4). For a report on the complete processing workflow and metadata please see Sodir site75. This seismic profile provides fine vertical resolution (0.1–0.3 ms ~20–60 m) from the seabed reflection down to 6.8 s two-way traveltime (twt), although more reflectivity with various resolutions is visible down to 8 s twt in some areas. A tentative Eocene (~49 Ma) organic-rich Azolla interval61 is correlated on the utilized seismic sections as a semi-continuous and relatively high-amplitude seismic reflector occasionally truncated against the top acoustic basement and is buried by only ~2 km of sediments at water depths of ~4 km (Fig. 2 profiles i-ii and 1–3). The Azolla event corresponds to the base of unit NB-1B in the Nansen Basin25, whereas to the near top of unit AB-1a in the Amundsen Basin56. Seismic attributes analysis is implemented to support the interpretation of features discussed in this work (e.g., hydrothermal vents/fluid conduits, sill intrusions, and deformation associated to intraplate volcanic activity; see S5). The analysis of the original trace amplitudes along seismic profile NPD-1204-402 (S5a) was performed in PetrelTM using the seismic volume attributes module and included: RMS amplitudes, which is the root-mean-square amplitudes computed over a specified time window (in this case 10 milliseconds; S5b); the relative acoustic impedance, which is an estimation of the acoustic impedance calculated by integrating the trace over time, and then passing the result trough a high-pass Butterworth filter (with cut-off frequency of 10 Hz; S5c); and the edge enhancement, which is based on an algorithm that includes a contrast filter guided by the relative dip of reflections (S5d). Insights from magnetic anomaly data complement the seismic interpretation within regions of sparse sampling coverage (see S1 and S6). Magnetic data is publicly available through the US Naval Research Laboratory (NRL; Arctic surveys 1998/99; ref. 12), National Centers for Environmental Information/National Oceanic and Atmospheric Administration (NCEI/NOAA76), and Circum-Arctic mapping project (CAMP13) (Figs. 2 and S2).

Geochemistry of Late Cenozoic igneous rocks

Available relevant geochemical dataset of Late Cenozoic (i.e., Neogene, Quaternary) basalts and xenoliths in basaltic lavas adjacent to and within Svalbard were compiled for the purpose of determining possible sources of the volcanic products. The final goal was to establish processes responsible for the formation of the reported volcanic-related hydrothermal vents reveled by seismic reflection data (e.g., Tayrona Vent) in this study. The compilation consists of published trace elements compositions and isotopic (Sr, Nd, Pb) data. Spider-diagrams were constructed with sample compositions normalized relative to primitive mantle77 (S7a). Standardized isotopic cross-plots were, in addition, constructed to further discriminate magmatic sources37,40 (S7b–d). In these cross-plots are also included illustrating the fields/trends of end-member magmatic sources (i.e. HIMU, EM-1/2), depleted mantle (DMM), and continental crust (CC) signatures, as well as other typical geochemical isotopic character of basalts worldwide37,40,78. Sample locations are indicated by each sample zone in a map of depth to lithosphere-asthenosphere boundary (LAB; ref. 79) (S7e). Sample references and in-depth geochemical investigations are listed as follows: Mosby Seamount, ~13 Ma-Sophia Basin (ref. 30); SW Bjørnøya, ~7-3 Ma (ref. 10); Quaternary xenoliths and lavas in Svalbard (Type 1/2, T-1/2; refs. 37,40); Neogene and Quaternary basalts in Svalbard (refs. 38,39); basalts in central and northern Lena Trough, CLT and NLT/GR-WVZ (ref. 41); and basalts along Gakkel Ridge’s western volcanic/sparsely-magmatic zone, WVZ-GR and SMZ-GR (ref. 51). Supplemnetary table file is included with this work and contains the geochemical data compiled from the basalts referred above.

Mantle tomography in NE Atlantic-Arctic

We have inspected the mantle tomographic model of ref. 43, NAT2021, which is an azimuthally anisotropic S-wave model of the upper mantle and transition zone under the North Atlantic and western part of the Arctic regions. This model is the most up-to-date and high-resolution model of the crust and upper mantle structure, and is specifically optimized for this region. The model resolution is explicitly variable and changes both, laterally and vertically, depending on the data sampling. We extracted several horizontal slices of the tomographic model at depths between 56 and 485 km aiming to map the lateral and depth extent of negative (seismically slow) mantle anomalies that may indicate a higher than ambient mantle temperature. Such a mantle anomaly is identified in the upper mantle SW of our study area; at depths exceeding 300 km, this anomaly is more regional, starting beneath NE Greenland and continuing to Svalbard and northern Barents Sea (S8). Other tomographic models (another recent S-wave model, 3D2018 by ref. 80, and the composite model SMEAN2 by ref. 81) are presented here for comparison, show negative mantle anomalies in our extended study area at 100 km and at 300 km and deeper. The Tayrona Vent (depicted by a red star, S8), identified in this study, is usually sitting at the eastern boundary of these mantle anomalies. Although the mantle structure is based on present-day seismological data, we suggest that the deep-seated negative mantle anomaly may illustrate the structure of the mantle in the past and was possibly influenced and connected to the complex development of the Iceland Plume in the NE Atlantic (ref. 43).