Introduction to volcanism in Antarctica: 200 million years of subduction, rifting and continental break-up

Gondwana break-up volcanism commencing at c. 190 Ma may have been driven by the effects of a large mantle plume (Storey 1995) and it is represented in Antarctica by two major voluminous volcanic provinces (Fig. 1a). One is a mafic large igneous province (LIP) that crops out throughout the Transantarctic Mountains and in Dronning Maud Land, with correlatives in South Africa, Tasmania, Australia and New Zealand (Elliot et al. 2021; Elliot and Fleming 2021; Luttinen 2021). The other consists of a series of felsic flare-ups that affected the entire Antarctic Peninsula and extended into southern South America (Chon Aike province: Riley and Leat 2021a, b). The mafic volcanism in Antarctica is known as the Ferrar Supergroup but representatives in Dronning Maud Land show greater compositional affinities to coeval mafic lavas and sills in the Karoo of South Africa. The mafic LIP comprises voluminous flood lavas and thick dolerite sills. The event probably had a total volume of more than 0.5 × 10⁶ km³ but it may have been emplaced in a very short time, perhaps as little as 400 kyr (Marsh 2007; Burgess et al. 2015). Magma in the sills is estimated to have travelled c. 4000 km laterally, making it the longest interpreted magma flow on Earth (Elliot et al. 1999; Leat 2008). The earliest eruptive phase was characterized by phreatomagmatic pyroclastic eruptions from nested maar–diatreme vent complexes collectively called phreatocauldrons (White and McClintock 2001).

By contrast, the felsic Chon Aike province was erupted in three major pulses, individually c. 6–10 myr in duration, between 189 and 153 Ma (Pankhurst et al. 2000; Riley et al. 2001; Riley and Leat 2021a, b). Each corresponds to a predominantly explosive volcanic flare-up (sensu Paterson and Ducea 2015) resulting mainly in the eruption of rhyolite ignimbrites. A prominent Pacific-ward progression observed in the ages of the volcanism has been linked to the impact and sublithospheric melting effects of the large spreading head of the same plume responsible for the Ferrar–Karoo LIP. An arc-like (subduction-influenced) mantle source, rather than a plume, has also been postulated for the mafic LIP (Choi et al. 2019; Elliot and Fleming 2021; Panter 2021).

Coincident with the break-up magmatism, the Pacific margin of Gondwana was already the locus of a major long-lived continental magmatic arc, the products of which are today widely preserved throughout the Antarctic Peninsula (Fig. 1b–e) (Leat and Riley 2021a, b). The volcanism in the arc, and its plutonic equivalents, have been extensively dated, revealing that the activity probably took place as a series of magmatic flare-ups and lulls (Riley et al. 2018). This is particularly evident in the South Shetland Islands, which is the most intensively dated arc-related region (Haase et al. 2012; Leat and Riley 2021a). The locus of active arc volcanism in Palmer Land and much of Graham Land moved into the forearc region during the Late Cretaceous, probably due to orthogonal subduction coupled with enhanced time-integrated growth of an extensive accretionary complex. By contrast, the volcanism migrated obliquely away from the trench in the South Shetland Islands possibly due to a highly oblique subduction vector causing much more limited accretion and possibly forearc erosion (Smellie 2021a). The magmatic activity shut down progressively in a clockwise direction, commencing in Marie Byrd Land in the mid-Cretaceous (Larter et al. 2002). Subduction is active today (at a very slow rate) only at the northern tip of the Antarctic Peninsula, where a small ensialic marginal basin populated with numerous submarine volcanic centres opened up in response to plate boundary forces, including slab rollback (Haase and Beier 2021; Smellie 2021b). The marginal basin includes Deception Island, one of Antarctica's most active volcanoes (Geyer et al. 2021), which underwent a major caldera collapse eruption c. 4 kyr ago that dispersed ash more than 4000 km in an arcuate swathe across the Scotia Sea and much of East Antarctica (Antoniades et al. 2018). Coincidentally, from c. 12 Ma an extensive back-arc mafic alkaline volcanic field developed to the rear of Graham Land in the James Ross Island region (Fig. 1f) (Haase and Beier 2021; Smellie 2021b). It is dominated by the very large shield volcano of Mount Haddington, which is predominantly glaciovolcanic and contains an unrivalled record of the Antarctic Peninsula Ice Sheet (Smellie et al. 2008).

Following the progressive shut down of subduction, ‘windows’ opened up in the downgoing oceanic slab, allowing ingress and decompression melting of mantle unaffected by subduction metasomatism. As a consequence, several small monogenetic alkaline volcanic fields were constructed from c. 7.5 Ma, mainly along the flanks of the Antarctic Peninsula (Fig. 1f) (Hole et al. 1995). The outcrops are overwhelmingly glaciovolcanic and, like Mount Haddington, they also preserve a uniquely valuable record of the Antarctic Peninsula Ice Sheet, as well as influencing our understanding of glaciovolcanism (Smellie and Edwards 2016; Smellie and Hole 2021). New research suggests that melting of the downgoing slab may be involved in the generation of the slab-window basalts (Hole 2021).

Over a period of a few million years in the Cretaceous, Antarctica's magmatism shifted from arc-related to rift-related during the latest stage of Gondwana break-up that separated Zealandia from Antarctica (c. 90–83 Ma). Rare Late Cretaceous alkaline intraplate magmatism is found in southern Zealandia (Weaver and Smith 1989), including HIMU-like ocean island basalt (OIB) (Panter et al. 2006; Hoernle et al. 2020), and mafic dykes are present along the Ruppert and Hobbs coasts of Marie Byrd Land (Storey et al. 1999). However, the episode was largely amagmatic. The extension also probably caused widespread topographical lowering until much of the region in Marie Byrd Land (and coeval terrain in New Zealand) subsided down to, or close to, sea level (LeMasurier and Landis 1996).

Renewed extension during the Cenozoic resulted in the creation of the WARS, a very large continental rift characterized by widespread alkaline volcanism (Fig. 1d–f) (Siddoway 2008; Smellie and Martin 2021; Smellie and Rocchi 2021; Wilch et al. 2021). The rift contains numerous large and small volcanoes with basalt–trachyte, phonolite and rhyolite compositions (Martin et al. 2021; Panter et al. 2021b; Rocchi and Smellie 2021). Remote sensing studies have also suggested that numerous additional volcanic centres may be widespread beneath the West Antarctic Ice Sheet, including several that may be active (Behrendt et al. 1994; van Wyk de Vries et al. 2018; Quartini et al. 2021). The origins of the volcanism are disputed, and variable roles have been inferred for deep mantle plumes and shallow thermal anomalies with associated edge flow (LeMasurier and Landis 1996; Rocchi et al. 2005; Martin et al. 2021; Panter et al. 2021b; Rocchi and Smellie 2021). They are all linked within the broad concept of a diffuse alkaline magmatic province (DAMP), which also includes Late Cretaceous and Cenozoic volcanism in eastern Australia, Tasmania and Zealandia (Finn et al. 2005). The striking compositional commonality throughout the DAMP region has been explained by melting of a mantle source component inferred to underlie the entire region, with the characteristics of HIMU-like OIB. The mantle reservoir may have been emplaced as a large plume head, of late Cretaceous age or older, or lithosphere that has been metasomatized, or both sources exist (Panter 2021). Like alkaline volcanism in the Antarctic Peninsula, the volcanism in the WARS is also predominantly glaciovolcanic, and it has also provided uniquely important information on the development of the West and East Antarctic ice sheets (Smellie and Rocchi 2021; Wilch et al. 2021).

The tectonic setting of some of Antarctica's volcanoes is enigmatic, however. They include at least three small monogenetic phreatomagmatic edifices in the southern Transantarctic Mountains (upper Scott Glacier) and beneath the East Antarctic Ice Sheet, less than 300 km from the South Pole (Fig. 1e) (Smellie et al. 2021), and Gaussberg, which is an isolated pillow volcano on the East Antarctic coast far from any other expression of volcanism on the continent (Fig. 1g) (Smellie and Collerson 2021). The upper Scott Glacier and nearby subglacial centres are outside of the WARS and they are believed to have formed in response to the detachment and sinking of lithosphere into the convecting mantle beneath the East Antarctic Craton (Panter et al. 2021a). The composition of Gaussberg is unique in Antarctica and very rare worldwide, being formed of ultrapotassic lamproite. Its origin is probably linked to a small deep-sourced mantle plume, distinct from the large Kerguelen plume, that incorporated a component derived from ancient subducted sediment (Murphy et al. 2002).

Antarctica also contains several large active volcanoes (Geyer 2021) (Fig. 1g). Only two have been observed in eruption (Mount Erebus and Deception Island); both have been intensively investigated (Oppenheimer and Kyle 2008; Geyer et al. 2021; Sims et al. 2021). Mount Erebus also hosts the world's only semi-permanent phonolite lava lake. The presence of relict heat (Mount Berlin, Mount Melbourne and Mount Rittmann) and abundant englacial and marine tephras sourced in Mount Takahe, Mount Berlin, Mount Waesche, Mount Rittmann and, possibly, The Pleiades indicate that many others were active in recent geological time (<10 ka: Lee et al. 2019; Dunbar et al. 2021; Gambino et al. 2021; Narcisi and Petit 2021; Di Roberto et al. 2021). Three of the volcanoes are, or have been, monitored (Deception Island, Mount Erebus and Mount Melbourne) but only one has published hazard and risk assessments (Deception Island: Bartolini et al. 2014; Pedrazzi et al. 2018; Geyer et al. 2021).

The authors gratefully acknowledge the collaborations and conversations with colleagues, too many to mention, over the many years that we have been actively engaged in Antarctic research. We are also grateful for the support provided by the various national funding agencies, without which none of our studies could have been implemented.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data sharing is not applicable to this article as no datasets were generated or analysed during the current study.