Introduction

Large Igneous Provinces (LIPs) are defined as large volumes of predominantly mafic rocks characterized by a high rate of magma accumulation and unrelated to plate-tectonic processes, i.e., formed far away from plate boundaries within intraplate tectonic environments1,2,3. Within continents, such a sudden occurrence of continental flood volcanism is usually preceded by a rapid uplift of the surface topography of 0.5–2 km within a few Myr4,5. Most commonly, both the transient dome-shaped surface uplift6 and the subsequent intraplate magmatic activity7 are attributed to mantle plumes8, seismically detected thermal9,10 or thermal–chemical11,12 anomalies in the Earth’s mantle13,14. Importantly, these upwelling structures are not limited to the classic (“primary”) Morgan-type plumes15 that rise from the mantle–core boundary (~2900 km) throughout the entire mantle but also include so-called “secondary” plumes16 rooted in the upper-lower mantle transition zone (MTZ: ~410–660 km)17. Such small-scale anomalies in the upper mantle (also called “baby” plumes)18 could originate from “primary” (super)plumes ponding at the 660 km phase change boundary19 or be the result of deep dehydration of oceanic slabs stagnating in the lower part of the MTZ20,21,22.

Although the formation of LIPs is by definition not causally linked to plate-tectonic processes, Precambrian records in southern Africa show that LIPs may occur during supercontinental assembly23 through thermal blanketing beneath the growing continent24 and without support from mantle plumes25. In contrast, most Phanerozoic plume-related LIPs are known to be associated with the break-up of continents and the subsequent opening of large oceanic basins. This is evidenced by the spatial and temporal correlation between the major continental flood basalts formed in the last 300 Myr and the different phases of fragmentation of Pangea, the youngest supercontinent in Earth’s history26 (see also Table 1). In addition, recent compilations of continental and oceanic LIPs (including those found not only on the present-day seafloor but also in ophiolites) have shown a statistical correlation between the number of mantle plume events and supercontinental cycles since >1000 Myr27.

Table 1 Break-up-based classification of Large Igneous Provinces (LIPs).
Full size table

However, the concept of plume-induced continental rupture can be challenged by the following observations: (1) long-lasting (up to ∼200 Myr) phases of near-amagmatic rifting preceding emplacement of LIPs28 and (2) the coincidence of the location and orientation of break-up axes with pre-existing suture zones29. In conjunction with arguments in favor of low (50–100 K) potential temperature anomalies beneath commonly accepted mantle plumes, such as the Iceland plume30, not only has the active role of LIPs in Pangea break-up31 been questioned, but even the existence of mantle plumes themselves32,33. It should be noted that all of these “anti-plume” views are generally at odds with deep mantle geochemical signatures of the LIPs34 and other volcanic hotspot lavas35.

To reconcile these apparent contradictions between active (driven by mantle plumes) and passive (driven by far-field tectonic forces) mechanisms of continental rifting and subsequent break-up, attempts have been made to develop transitional36 or combined37,38,39 passive-active scenarios. More recently, intermediate types of rifting-to-break-up systems (such as the so-called “semi-active” and “semi-passive”) have also been identified18. With this in mind, we examine major continental LIPs emplaced since the Late Paleozoic and propose a new classification consistent with their relationship to the disintegration of Pangea.

LIPs without break-up (LIP-Shirker)

We begin our overview of Phanerozoic LIPs with two examples that, contrary to common expectations, are completely ineffective in terms of continental break-up and subsequent onset of oceanic spreading. Therefore, we refer to this type of LIP as LIP-Shirker.

The Permo-Triassic Siberian Traps (~250 Ma40) is an archetypal example of a continental LIP (Fig. 1a), best known for its proven impact on the largest known mass extinction event11,41. Despite the concurrent development of a rift system beneath the West Siberian Basin42,43, one of the largest hydrocarbon provinces in the world44,45, the Siberian Traps did not result in plate rupture and subsequent formation of an ocean basin.

Fig. 1: End-member LIP types.
figure 1

a Siberian Traps as an example of LIP-Shirker. Hot upwelling material spreads laterally below the lithosphere-asthenosphere boundary (LAB) to form a mushroom-shaped plume head feeding widespread magmatic activity over an area of up to ~5.0 × 106 km2. Plume-induced continental rifting (the West Siberian rift system) terminated in ~10 Myr during the Middle Triassic42 without rupturing the lithosphere. Geological map is modified from ref. 40. b Deccan LIP as an example of LIP-Producer. Rapid penetration of the mantle plume through the lithospheric mantle, which is progressively thinned in the active-passive mode37,79, leads to continental break-up shortly after (in a few Myr) emplacement of LIP-related volcanism, which in this case is much more spatially restricted, covering an area of ~1.5 × 106 km2. Réunion hotspot track is after refs. 70,71. Paleo-position of the Seychelles microcontinent at ~63 Ma is after ref. 67. Note the oblique orientation of the pre-magmatic rifts/basins (CB—Cambay Basin; KB—Kutch Basin; NR—Narmada-Zone Rift)75,76,77 with respect to the break-up axis (future Carlsberg Ridge) between the Indian Plate and the Seychelles78.

Full size image

Coincidence or not, another LIP-Shirker developed around the same time: the Late Permian (~260 Ma) Emeishan LIP was emplaced on the western margin of the Yangtze Cratonic block of South-West China46,47. Similar to the Siberian Traps, the Emeishan LIP was accompanied by plume-induced continental rifting6,48, which took place in the Panxi region above the plume head49, possibly also in the presence of far- and near-field tectonic stresses50.

These two examples document that without appropriate geodynamic conditions (e.g., the presence of a pre-existing weak zone at the site of LIP emplacement and/or high-level and long-lasting far-field extensional stresses), even the most voluminous extrusions of flood basalts are insufficient to cause continental break-up with their impact limited to aborted rift systems.

Although the criteria for defining LIPs are not quite met, the Tertiary volcanic provinces of western and central Europe are probably another example of a “Shirker”. In this case, the European Cenozoic Rift System (ECRIS) was likely activated during the Paleogene51 by a mantle plume or a system of small (“secondary”) plumes that have been magmatically active since the Paleocene52 and can still be detected through seismic tomography53,54. The change in geodynamic regime at ~35 Ma (transition from progressive closure of the Neo-Tethys Ocean to development of back-arc basins over retreating slabs)55 has aborted the ECRIS and made the European volcanic province an equivalent counterpart to the more voluminous LIP-Shirkers discussed above.

LIPs with break-up

In contrast to the LIP-Shirker end-member, most cases of LIP emplacement result not only in rifting but also in rupture of the continental lithosphere26. Considering the relative role of mantle plumes associated with LIPs in the break-up process, we introduce the following types: LIP-Producer, LIP-Trigger, and LIP-Attractor. Their characteristic features are described below along with the natural examples and the criteria for distinguishing them.

LIP-Producer

The LIP-Producer corresponds to a scenario in which plume emplacement determines the location and timing of continental rupture. Prominent examples are the Madagascar and Deccan LIPs (Fig. 1b), where the rapid eruption of intraplate flood basalts occurred at ~93–9056 and ~66–64 Ma57, respectively. The continental lithosphere overlying the corresponding mantle plumes was therefore effectively weakened by basal thermo-mechanical erosion58,59,60 and the reduction in the long-term brittle strength of rocks exposed to melt percolation61,62,63. In addition, this weakened lithosphere was subjected to slab pull forces by the continuous subduction of the Neo-Tethys Ocean floor64, sometimes enhanced by a double subduction with two nearly parallel, north-dipping subduction zones between the Indian and Eurasian plates65. Under such favorable conditions, corresponding to the active-passive scenario when the mantle plume is combined with far-field tectonic extension37,38,39, the continents were broken-up in only a few Myr after the formation of the Madagascar and Deccan LIPs, resulting in the successful separation of Madagascar and India at ~84 Ma66 and India and the Seychelles at ~63 Ma67. Consistent with the classic concept that flood basalts represent the “head” of the plume and that continued magmatism along hotspot tracks is associated with the remaining plume “tail”68, the Madagascar and Deccan LIPs mark the spatial and temporal beginning of well-known oceanic hotspot tracks that terminate at the current position of the Marion69 and Réunion70,71,72 plumes, respectively. It is also important that the Mesozoic extensional basins that preceded the emplacement of both the Madagascar73,74 and Deccan LIP75,76,77 are characterized by a strong obliquity with respect to the future break-up axis78.

As evident from these examples, on the one hand, the role of the LIP-Producer is dominant because plate-tectonic forces alone (e.g., slab pull by long-lived subduction since the Paleozoic)64 were not sufficient to localize rifting at the site and with the same orientation as the axis of eventual break-up. On the other hand, as shown by contrary LIP-Shirker end-member cases (Siberian Traps and Emeishan LIP), complete rupture of the lithosphere would not be possible without appropriate (extensional) far-field stresses determining the orientation of the break-up axis79,80. Paradoxically, the East African Rift System (EARS) could be an example of both end-members. The ongoing purely active rifting in its Eastern Branch, established in the Miocene81 after the emplacement of the Kenya plume82,83,84,85 at ~40–45 Ma86, did not yet evolve into the rupture of the African Plate. Obviously, this is due to an unfavorable tectonic setting with regional far-field compression by mid-ocean ridges surrounding the African continent87,88,89. In the absence of a plate-tectonic reorganization, the EARS will be aborted, similar to the fate of the ECRIS and other rift systems associated with LIP-Shirkers. In an opposite scenario where a switch in the tectonic regime can make break-up possible, the East African volcanic province will be an equivalent of LIP-Producer.

Importantly, mantle plume impingement can produce not only (super)continent rupture, but also the separation of relatively small continental ribbons or microcontinents90,91. In particular, several continental fragments are known to have drifted away from northern Gondwana during the Paleozoic (the Avalonia terranes and the Cimmerian blocks)92 and Mesozoic (the Apulian microcontinent)93, possibly due to the combined effect of slab pull by continuous subduction of paleo-oceans (from Proto-Tethys to Neo-Tethys)94 beneath the active margin of opposing continents (from Laurentia to Laurasia)95 and mantle plumes periodically arriving at the base of the African lithosphere96.

LIP-Trigger

In contrast to the LIP-Producer scenario, where thermal and magmatic weakening caused by a mantle plume leads to initial localization of pre-break-up deformation in the lithosphere, the magmatic activity associated with an LIP-Trigger is preceded by continental rifting, situated in the area of future lithospheric rupture and, crucially, of the same orientation as the final break-up axis78. Prominent examples are the Central Atlantic Magmatic Province (CAMP) and the Karoo LIP, emplaced at ~20197,98 and ~184–182 Ma99,100,101,102, respectively. In these cases, the role of the corresponding plumes was limited, only enhancing the ongoing localized extension related to pre-magmatic rifting which in both regions was already operating for several tens103,104 to several105,106 Myr. Therefore, the CAMP and the Karoo LIP activated (but did not cause!) lithospheric rupture that occurred relatively soon (~10–15 Myr) after their emplacement, opening the Central Atlantic at ~190 Ma107 and separating South Africa from Antarctica and Madagascar at ~168–164 Ma108.

We dub this type of LIP as LIP-Trigger because in this case pre-LIP localized deformation (rifting) could ultimately be terminated by a break-up of the continent, even without the involvement of a plume that merely accelerates (or triggers) this process without playing a dominant role. Coincidence or not, both LIP-Triggers (the CAMP and the Karoo LIP) lack hotspot tracks, unlike the LIP-Producers (the Deccan and Madagascar LIPs), which, as described above, have a clear expression in the form of tracks imprinted by deep-seated Réunion and Morion plumes. Therefore, LIP-Triggers seem to be preferably associated with bundles of “secondary” plumes rather than with individual “primary” plumes (Fig. 2a). This assumption is indirectly supported by (1) the presence of numerous small plumes in the central Atlantic Ocean (Azores, Canary, and Cape Verde)109, which are obviously much younger (Cenozoic) in age110,111, but could indicate a “secondary” plume fabric, that established in the region during the Late Triassic–Early Jurassic and is still in operation; (2) a larger area of scattered magmatism associated with LIP-Triggers (e.g., ~3 × 106 km2 for the Karoo LIP) than in the case of a spatially more concentrated magmatic area typical of LIP-Producers (e.g., ~1.5 × 106 km2 for the Deccan LIP); and (3) strong multivariance in proposed positions for the center of the Karoo plume112. We should note, however, that despite our hypothesis of numerous “secondary” plumes below the CAMP and the Karoo LIP this is not a general requirement for the LIP-Trigger, which in principle can also develop with a single “primary” plume, as exemplified by the Kerguelen oceanic plateau formed in the Early Cretaceous as a LIP over the Kerguelen hotspot113. The first magmatism associated with the Kerguelen plume has been dated to ~130 Ma in southwestern Australia114 and ~132 Ma in southeastern Tibet115, roughly contemporaneous with the initial break-up of India–Antarctica116 and India–Australia117, which in turn was preceded in both cases by continental rifting since ~160 Ma118, consistent with the LIP-Trigger scenario.

Fig. 2: Other types of LIPs associated with break-up.
figure 2

a Central Atlantic Magmatic Province (CAMP) as an example of LIP-Trigger. Simultaneous emplacement of numerous “secondary” plumes16 at the base of the pre-thinned lithosphere results in scattered magmatism that is relatively quickly followed (in ~10 Myr) by plate rupture. Concentrated volcanism coincident with plume-related lithospheric break-up forms large and thick wedges of so-called seaward dipping reflector sequences (SDRs) that mark volcanic passive margins (VPMs), in contrast to adjacent non-volcanic passive margins (NVPMs) without SDRs that presumably develop as part of a purely passive rifting-to-break-up system. Note that the VPMs of the Central Atlantic are separated from the NVPMs between Newfoundland and Iberia by a series of major transform faults. CFZ Canary Fracture Zone, NAFZ Newfoundland-Azores Fracture Zone, NFD Newfoundland. b Afar LIP as an example of LIP-Attractor. Soft point in the lithosphere created by plume-induced thermal anomaly at ~30 Ma121,122 controls the direction of horizontal propagation from the Carlsberg Ridge, which causes progressive break-up in the Gulf of Aden from the east (~19–17 Ma)119 to the west (~1 Ma)120. AFFZ Alula-Fartak Fracture Zone, SSFZ Shukra El Sheik Fracture Zone. Two maps in a and b are modified from ref. 78.

Full size image

LIP-Attractor

For both LIP-Producer and LIP-Trigger, continental break-up always occurs first above the plume emplacement area, forming volcanic passive margins (VPMs), and then propagates laterally, developing non-volcanic passive margins (NVPMs), which are usually abruptly separated from volcanic counterparts by transform faults78. On the contrary, in the case of the Afar LIP (Fig. 2b), the NVPMs at the eastern tip of the Gulf of Aden began to develop much earlier (~19–17 Ma119) than the VPM formation within the Afar triple junction (~1 Ma120). This spatial and temporal evolution of the break-up of the Gulf of Aden could be explained as follows: a domain of the continental lithosphere, that was rheologically weakened by the massive melting event of the Afar plume at ~30 Ma121,122, directed the westward lateral propagation of seafloor spreading from the mid-oceanic Carlsberg Ridge, which had already been active since ~63 Ma67 and then approached this thermal sublithospheric anomaly acting as a soft point123. We, therefore, dub this type of LIP as LIP-Attractor.

In the South Atlantic124, break-up at 133–125 Ma125 also propagated (in this case from south to north) toward the Paraná-Etedeka LIP, which erupted almost simultaneously at ~134–132 Ma126. This sequence has been hypothesized as evidence for a non-plume mechanism of South Atlantic opening127 and Pangea fragmentation in general31. However, it is more likely that we are dealing here with another LIP-Attractor, where the influence of the plume is restricted to directing the propagation of oceanic spreading that is already operating.

It should be mentioned that from the perspective of the opening of the Red Sea, the Afar LIP can also be considered as LIP-Producer. In this case, initial continental rifting occurred almost simultaneously with plume emplacement119, while the break-up was considerably delayed until ~4 Ma at the southern tip, without yet reaching the northern segments of the Red Sea128. This delay between the LIP formation and the resulting plate rupture (not typical of the classic LIP-Producers, where break-up usually develops much more rapidly) may be due to a much reduced Neo-Tethys slab pull since the onset of the progressive collision between the Arabian and Eurasian Plates in the Bitlis suture zone at ~40–30 Ma129. Consequently, it appears that in the classic active–passive scenarios corresponding to the LIP-Producers, the time interval between LIP emplacement and final break-up is controlled primarily by a combination of the efficiency of lithosphere weakening by magmatism (volume, temperature, and water content of plume material) and the level of external extensional forces.

“Re-awakened” and “dormant” LIPs (LIP-Dornröschen and LIP-Sleeper)

Despite decades of extensive study and the accumulation of a vast amount of geologic and geophysical data130, the North Atlantic Igneous Province (NAIP) and the associated evolution of lithospheric rupture and oceanic spreading are still the subject of vigorous debate about the mechanisms that favor either plume-induced131,132 or purely passive rifting and break-up133.

To reconcile ongoing controversies about the actual causes of the emplacement of voluminous magmatism in the NAIP, the following key elements of the evolution and structure of the North Atlantic realm should be taken into account as constraints: (1) the spatial variation in the thickness of the continental lithosphere134,135,136; (2) the relative motion trajectory of the Iceland plume since the Cretaceous137,138,139; (3) the timing and spatial distribution of the magmatism related to the NAIP140,141,142; and (4) the timing of the opening of the Labrador Sea/Baffin Bay143,144 and the North Atlantic145,146. According to these data, the lithosphere is thinned beneath the central part of Greenland, as evidenced by low seismic velocities in the mantle134,135,136. This seismically detected zone of relatively thin lithosphere crossing Greenland from west to east is also reflected in high measured147 and predicted148 heat flow, intense basal ice melt149, and increased crustal thickness due to magmatic underplating150 and intrusions in the middle and lower crust151. Given the spatial correspondence with the reconstructed paths of the Iceland hotspot137,138,139, most recent studies interpret these features as relict signatures of the passage of the Iceland mantle plume beneath Greenland from at least 90 to ~60 Ma148,149,150,151. In contrast, according to more traditional views, the Iceland plume was emplaced on the eastern edge of Greenland, as supported by observations of radiating and circumferential dyke swarms152, whereas part of its large, flattened head rapidly extended into western Greenland153, causing the quasi-simultaneous magmatic activity (the NAIP) on both sides of Greenland around 60 Ma140,141,142. However, the relatively close location of the reconstructed position (~130–120 Ma) of the so-called High Arctic Large Igneous Province, which includes the exposed components on Ellesmere Island, Spitsbergen, and perhaps northern Greenland154, and the corresponding segment of the hotspot track132 is indicative of the Iceland plume, which is more than twice as old as the NAIP137. Furthermore, the Iceland hotspot can even be traced back to West Siberia in the Late Permian–Early Triassic155, so that the Siberian Traps (~250 Ma40) could hypothetically be the very first magmatic manifestation of the Iceland plume.

In view of the above, magmatic activity in Baffin Island and Western Greenland during the Paleocene was probably triggered by spreading axis propagation from the Labrador Sea toward the Baffin Bay, where continental break-up was progressively initiated from ~64 Ma143 to ~56 Ma144, respectively. The overlap of the propagating spreading ridge with the ~100–80 Ma-dated segment of the Iceland hotspot track near the West Greenland passive margin139 (Fig. 3a) enlivened the hot material that had resided there already for ~20–40 Myr without signs of excessive volcanism. The simultaneous approach of the actual “tail” of the Iceland plume to the eastern edge of thick Greenland lithosphere139 allowed horizontal flow of the plume material along the adjacent thin-lithosphere corridors135 connecting the eastern edge of Greenland with the Southern Scandinavia and Scotland/Ireland156,157. A similar mechanism of propagation of hot plume material toward areas distant from its original emplacement along the elevated lithosphere–asthenosphere boundary158,159,160 has been proposed for Late Mesoproterozoic mafic magmatism in the southwestern USA (the Keweenawan LIP and the Southwest Laurentia LIP)161 and Cenozoic alkaline volcanism in Africa162 and Arabia163,164. The arrival of Iceland plume hot material beneath thin European lithospheric segments has led to intense plume-related magmatism in the British Isles area165, quickly followed by the break-up in the North Atlantic sensus stricto (i.e., between Greenland and Europe) along the Reykjanes–Aegir–Mohns Ridge at ~54–53 Ma145, preceded by the most long-lasting rifting since the Late Paleozoic166,167. The contemporaneous (~62–58 Ma) magmatism over a vast area from Baffin Island to the British Isles142 is thus driven by these two independent processes—spreading axis propagation and plume conduit motion—which happen to coincide in time (Fig. 3b).

Fig. 3: North Atlantic Igneous Province (NAIP) as an example of LIP-Dornröschen (or “re-awakened” LIP).
figure 3

a Thickness of continental lithosphere (LAB depth), based on seismic tomography model by ref. 135. A ~60 Ma plate reconstruction and the Iceland plume trajectory are based on the global moving hotspot reference frame of ref. 139. Blue circles indicate the approximate location of two NAIP regions. b Schematic profile showing: (1) “re-awakening” of plume material seeded since ~90 Ma on the western edge of Greenland by propagation of the spreading axis from the Labrador Sea, and (2) flow of hot material delivered from the plume source on the eastern edge of Greenland toward the thinner lithosphere of Europe. These independent but coincident events led to the simultaneous (at ~62–58 Ma)142 emplacement of the Baffin Island/West Greenland and British Tertiary Igneous Provinces, which together form the NAIP. Note that the vertical scale is exaggerated on this and other (Figs. 1 and 2) schematic profiles. Map in a is modified from ref. 132.

Full size image

The evolution of the North Atlantic region shows that a thermal anomaly hidden for a while beneath thick lithosphere can be “re-awakened” (or “re-initialized”) by the lateral propagation of spreading ridges or by tapping its source under thinner segments of overlying lithosphere due to horizontal plate movements. We dub this type of LIPs as LIP-Dornröschen (“re-awakened” LIP). We expect that the term LIP-Dornröschen (LIP-Sleeping Beauty) may be applicable to a broad family of LIPs, including those from the Precambrian. For example, several LIP pulses in the period 1800–1600 Ma, which formed on the supercontinent Columbia during ancient plate motion over a single stationary Xiong’er plume in the North China Craton168, can be regarded as the earliest known manifestation of a LIP-Dornröschen-like scenario in Earth history. On the contrary, the youngest case of delayed LIP outpouring might be the Columbia River Basalt LIP (~17–16 Ma169) associated with the Yellowstone mantle plume170, with a history appearing to extend to at least ~56 Ma, as indicated by offshore volcanism on the Siletzia oceanic plateau171.

Interestingly, the western (Baffin Island and Western Greenland) and eastern (British Isles) components of the North Atlantic “Sleeping Beauty” show some similarities to the LIP-Attractor and the LIP-Trigger, respectively (see Fig. 3 and Table 1), so a multilevel classification for these and/or other LIPs (and/or LIPs components) could probably be developed in the future.

The mantle plume underlying the Manus back-arc basin (Western Pacific Ocean), which was discovered by seismic tomography without showing up as an evident hotspot172, is likely a good example of a possible future LIP-Dornröschen that is currently still “dormant” (LIP-Sleeper). Future high-resolution seismic tomographic studies in continental and oceanic lithospheric environments will be of particular importance in discovering new “dormant” plumes or hidden hotspots173, which may provide additional constraints on plate motion history174 and will likely help uncover new aspects in the geodynamics of mantle-lithosphere interactions.

Conclusions and future outlooks

For more than 25 years, the prevailing view of the scientific community on the causal links between mantle plumes and the break-up of (super)continents has changed considerably. Traditionally, it was postulated that the emplacement of LIPs played a key role in Pangea fragmentation26. However, recognition of certain limitations of this concept28,29 has gradually led to a fundamental reassessment of the causes of continental break-up31, which ultimately rules out the necessity175 and even the existence of a mantle plume component32,33. As a consequence, purely passive continental rifting models are now even being applied to regions such as East Africa176,177,178,179, where both geophysical83,85 and geochemical82,84 observations unequivocally indicate the presence of mantle plumes, rooted in a common large-scale source corresponding to a first-order mantle structure such as the African superplume180,181.

To prevent further counterproductive discussions advocating end-member views of classic “passive versus active“ rifting debate182, a parallel of the more general “plates versus plumes“ controversy32,33, we propose here a new classification of LIPs that illustrates the variability in the possible role of mantle plumes in the process of continental break-up (in addition to the well-known diversity in geochemically based classifications)23,183. In particular, we demonstrate that, on the one hand, LIPs indeed cannot always be causally linked to lithospheric rupture. As the geologic records of the Siberian Traps and the Emeishan LIP (LIP-Shirkers) show, even very voluminous magmatism associated with active rifting does not always by itself lead to continental break-up in the absence of favorable tectonic conditions. On the other hand, regional far-field extension alone also does not act as an efficient break-up mechanism, as most non-volcanic passive margins (e.g., Newfoundland–Iberia, Equatorial Atlantic) are the result of horizontal propagation from adjacent areas of plume-activated spreading78. This highlights the key role of combined active–passive mechanisms38,184, where the site of localized deformation is determined by the plume, while the orientation of the rift and spreading axis is controlled by the direction of external extension. This active-passive scenario corresponds to the LIP-Producer exemplified by the Deccan and Madagascar LIPs.

In addition, plumes can play a limited, yet important and sometimes even definitive role in the dynamics of plate rupture. These include (1) initiating break-up when rifting is already underway, thereby determining the timing of lithospheric rupture (LIP-Trigger, e.g., the CAMP and the Karoo LIP), and (2) creating mechanically soft zones in the lithosphere that spatially control the direction of propagation of oceanic spreading (LIP-Attractor, e.g., the Afar and Paraná-Etedeka LIPs).

Two remaining types of LIPs are LIP-Dornröschen (Iceland plume) and LIP-Sleeper (Manus basin). This implies that interpretation of the timing of LIP emplacement made from a mantle dynamics perspective27 should be handled with caution because of possible delays between the timing of upwelling in the mantle and its detectable magmatic manifestation at or near the Earth’s surface.

Although our classification is currently based on the best-known examples of continental Phanerozoic LIPs (Table 1), it should also be relevant to the future study of other LIPs (including Precambrian and oceanic), providing a generic guide for further studies of intraplate magmatism in terms of its relationship to plate rupture.