Introduction

Plate tectonics is characterized by periods of gradual, broadly continuous, plate movement that are punctuated by relatively short times of plate reorganization1,2. These are due to the consumption of an oceanic plate, the collision of two continents, the cessation of subduction, or the breakup of a (super)continent3,4. These events disturb the plate kinematic status quo and force adjustments over the planet surface that affect ocean and atmospheric circulation and have been linked to perturbations in the carbon cycle5, amongst other things. Identifying and understanding plate reorganization events in deep time is only possible with full-plate topological reconstructions. These are well developed for the Mesozoic and younger6, but have only recently been proposed for the Paleozoic7,8 and now Neoproterozoic eras9,10. These allow regional plate tectonic induced phenomena to be understood in a global context.

Here we present an updated GPlates (www.gplates.org) model that uses recently published geological data from the terranes of the Arabian-Nubian Shield (ANS). By adding in geological data from relic volcanic arcs into these full-plate topological reconstructions of the ancient earth, we provide a new interpretation of the oceanic plate kinematic and dynamic evolution of the Neoproterozoic Mozambique Ocean. This led to the southward journey of Neoproterozoic India to collide against African Gondwana and the Australia-Mawson continent11 to form the kernel of Gondwana. This plate reorganization is coeval with the opening of the Pacific Basin12,13 and directly precedes the cataclysmic climatic perturbations of the Cryogenian.

Background

Plate tectonics has been causally linked to the Cryogenian climate instability14,15,16, to the coeval Neoproterozoic Oxygenation Event17,18, to the biosphere tumult that included the ecological takeover of eukaryote cells19,20 and, ultimately, to the evolution of metazoans21,22. The veracity of these hypotheses requires knowledge of the plate tectonic configuration and kinematics through the Neoproterozoic, which is missing, as most attempted reconstructions are “continental-drift” models with no full-plate circuit attempted11,23. A first attempt at a full-plate topological model for 1000–520 Ma was recently published9. However, this model focused predominantly on cratonic crust that had available palaeomagnetic data, and consequently simplified many complex areas around Neoproterozoic active margins. Mallard et al.24 demonstrated that active margins control plate size and number, therefore also control many of the parameters needed to understand the role of plate tectonics on broader earth systems. The ANS was an active margin and one of these areas simplified in Merdith et al.9, yet, it is one of the most critical for Neoproterozoic plate reconstructions as it is one of the most extensive areas of new Neoproterozoic crust on the planet25 (Fig. 1) and preserves evidence of subduction from pre-1 Ga until the Ediacaran26.

Fig. 1: Map of present northern Indian Ocean region.
figure 1

Showing the distribution of juvenile Stenian–Ediacaran crust, pre-Stenian exposed crust, pre-Stenian exposed crust reworked thermally and structurally during the Neoproterozoic, and Proterozoic sedimentary basins in NE Africa, Arabia, and the Indian subcontinent. The late Tonian plate reconfiguration is represented by the notable change from pre-720 Ma, approximately NE–SW sutures to post-720 Ma, approximately NNW–SSE striking sutures. MB Mozambique Belt, NED Northern Eastern Desert, CED Central Eastern Desert, SED South Eastern Desert.

The end of the Proterozoic eon is marked by some of the most dramatic events in Earth’s history, with this period of time being characterized by extensive changes in seawater chemistry demonstrated through the strontium, sulfur and carbon isotope records, large climatic extremes, and preservation of the Ediacaran faunal assemblage and the explosion of Cambrian fauna27. These global variations are concurrent with the amalgamation of Gondwana, and the closure of the Mozambique Ocean; representing one of the major and final Gondwana forming collisional zones28. As no in situ oceanic crust exists before ca. 200 Ma, the remnants of this major ocean gateway are only preserved in relic arc–arc, arc–continent, and continent–continent collisional zones, within the East African Orogen (EAO). The EAO is one of the largest orogens of the last billion years, which, in a reconstructed Gondwana, extends from Turkey and the Levant, in the north to Mozambique, Madagascar, Sri Lanka, and East Antarctica in the south29. Along strike, the orogen is divided into two. The Mozambique Belt lies in the south and is a tract of largely older continental crust, extensively deformed and metamorphosed in the Neoproterozoic/Cambrian30. The ANS makes up the north of the orogen26. The ANS, and adjacent Gondwanan rocks in North Africa and from east Arabia to NW India preserve the evidence we use to reconstruct the plate tectonic circuit as Neoproterozoic India converged and finally collided with the African parts of Gondwana26.

Results and Discussion

Model contraints

In this paper, we use previously published geological data to constrain our full-plate topological model for the evolution and closure of the Mozambique Ocean and the amalgamation of central Gondwana. The full-plate model is based on geological and paleomagnetic data and is part of the first published self-consistent model of global plate tectonics over the last billion years10. We emphasize that this is a model, and although we argue that it best represents the geological and paleomagnetic data available in 2021, it is not a unique solution and is subject to improvement with more data and better interpretations. The model, however, does present interesting implications for the progression of plate tectonics over this time, the distribution of plates, of continents and oceans and leads to hypotheses for plate tectonic influence of earth-surface systems that we begin to explore in this paper.

The ANS is laced with suture zones that represent collisions between different terranes as subduction zones consumed the intervening oceanic crust (Fig. 1). A dramatic feature of the region is that pre-715 Ma sutures are aligned approximately ninety degrees from post-715 Ma sutures26,31. This observation reflects a major change in plate convergence direction and we use this as the start of a higher-order reconstruction of this region in a full-plate context.

The Mozambique Ocean, Azania, and Afif-Abas

The Mozambique Ocean closed as Neoproterozoic India converged on the African parts of Gondwana (Kalahari, Congo, Sahara) to form central Gondwana29. The East African Orogen resulted from the collision between these major continents and amalgams of smaller terranes, during the Neoproterozoic to early Cambrian. Sandwiched within the EAO lies a broad band of Archean to Paleoproterozoic crust that was identified by Collins and Windley32 as a microcontinent (subsequently named “Azania”), whose remains are found in southern India, central Madagascar, Somalia, eastern Ethiopia, and Arabia (Fig. 1). In Yemen, the Al-Mafid Terrane is correlated with Azania32 and this is separated from a second pre-Neoproterozoic terrane called the Abas Terrane by a Neoproterozoic arc terrane (the Al Bayda terrane). Because of this, Collins and Windley32 suggested that a second microcontinent existed that they called Afif-Abas due to the continuation of the Abas terrane into Saudi Arabia as the Afif Terrane.

Azania, and Afif-Abas, are interpreted to have collided with the eastern margins of the Congo craton and Saharan Metacraton by ~630 Ma to form the East African Orogeny sensu stricto. A younger orogeny (ca. 570–520 Ma), was interpreted to represent the final collision between India and the amalgamated Africa/Arabia and called the Malagasy orogeny11.

The Eastern Margin of the EAO (NW India to Oman)

The easternmost margin of the northern East African Orogen is the boundary between the Mesoproterozoic terranes of India and the Stenian–Tonian crust that extents west from the Delhi-Aravalli Orogen. This has been interpreted to be the eastern margin of the northern East African Orogen. During the Stenian and Tonian, progressive arc accretion of volcanic arc rocks onto the NW margin of Neoproterozoic India occurred; extending into the basement rocks of Pakistan and the inliers of Oman33,34. This was later covered by an extensive Cryogenian–Ediacaran passive margin succession, with comparable sequences continuing into the Cambrian35.

The Arabian-Nubian Shield

The ANS is dominated by low grade volcano-sedimentary sequences and associated plutonic and ophiolitic remnants. The tectonic history of the ANS is complicated and preserves a complex mix of terranes, accreted arcs that record subduction polarity reversals that are reviewed and summarized in a number of papers26,31. There are no reliable paleomagnetic data available to constrain these blocks, so we have constrained their positions by their relation to each other and through plate kinematic constraints.

The oldest terrane in the ANS is the late Mesoproterozoic Sa’al Metamorphic complex (1.03–1.02 Ga) in Sinai, marking the initiation of magmatism in the northern-most ANS36,37 (Fig. 1). The location of this Stenian terrane in the reconstruction is uncertain, but coeval subduction-magmatism occurred within the Saharan Metacraton (see below).

The Tonian to Cryogenian history of the ANS is marked by formation of oceanic volcanic arcs and continental volcanic arcs built on Azanian (or Afif-Abas) crust that amalgamated to form a larger intra-Mozambique ocean terrane separate from both Neoproterozoic India and African Gondwanan continents. A number of terranes in the ANS are correlated as equivalents, separated by the opening of the Red Sea, from south to north, these are the Asir and Tokar/Barka terranes, the Haya and Jiddah terranes, the Hijaz and Gabgaba/Gebeit terranes, and the Eastern Desert and Midyan terranes26. It is unclear whether the combined Asir-Tokar/Barka terrane and Haya-Jiddah terranes were ever on separate plates as, in Saudi Arabia, no clear suture is seen between them. In SE Sudan and Eritrea, the Barka suture does appear as the site of ocean closure, so these may form a complex middle Tonian amalgam.

The older, Tonian to earliest Cryogenian, amalgamation history of the ANS is marked by approximately NE–SW oriented sutures (present orientation) between juvenile Neoproterozoic ocean-arc terranes. The oldest of these sutures is between the Jiddah-Haya and Gabgaba/Gebeit-Hijaz terranes (the Bi’r Umq–Nakasib suture), and is dated at ca. 780–750 Ma26. This suture created the kernel of a late Tonian microcontinent. The Midyan-Eastern Desert collided with this kernel ca. 715 Ma along the Yanbu–Sol Hamed suture31. Both of these sutures evolved from SE-dipping subduction zones31.

The older NE–SW sutures are bound by younger NNW–SSE Cryogenian to Ediacaran sutures and terranes that represent a fundamental kinematic change in Mozambique Ocean subduction. The oldest of these is the 680–640 Ma Nabitah suture, which forms the eastern margin of the intra-Mozambique Ocean island-arc terrane microcontinent (discussed above), against Tonian–Cryogenian continental arcs built on the Afif-Abas microcontinent. This now enlarged Cryogenian Afif-Abas microcontinent collided with the active margin of the Sahara Metacraton along the Sudanese Keraf Suture. This collision occurred in late Cryogenian to early Ediacaran times (ca. 650–580 Ma)38. Further to the east, in the most easterly exposed terrane, the Saudi Ar Rayn Terrane, juvenile calc-alkaline magmatism stretches from ca. 690 to 615 Ma39. Turbiditic sediment deposition in the Ad Dawadimi basin that separates the Ar Rayn Terrane from the Afif-Abas microcontinent continued until at least 620 Ma, but was locally intruded by ca. 630 Ma adakitic magmas40. This sequence was metamorphosed to greenschist-facies grades at ca. 620 Ma41. Further east still, broad N–S magnetic highs, beneath the Arabian Phanerozoic sedimentary sequence42, suggest younger arc terranes now buried beneath the Rub al-Khali Basin. The transition to post-tectonic magmatism within the eastern terranes of the ANS begins from 605 Ma39 and pull-apart basins developed along the large strike-slip faults that cut the region43. Post-tectonic magmatism begins in western Ethiopia at ca. 572 Ma44.

The final collision between Neoproterozoic India and the, by then amalgamated, Azania/Congo Craton occurred at ca. 570–540 Ma, closing the final strand of the Mozambique Ocean11,28. This suture lies beneath the Phanerozoic cover between the exposed Saudi and Yemen basement and Mirbat in SW Oman. It appears to be imaged by shear-wave anisotropy variations seen directly west of Mirbat45. In western Oman, latest Ediacaran–Cambrian deformation is seen in the subsurface, its limit is known as the Western Deformation Front and the deformation associated with this is known as the “Angudan event”. The sub-Rub al-Khali suture has been traced south within reconstructed Gondwana to Madagascar where it has been correlated with the Antsaba shear zone of NW Madagascar46, the Betsimisaraka suture32 and into the Palghat–Cauvery Suture of southern India47. This Palghat–Betsimisaraka–Antsaba–Western Deformation Front suture represents the final suture of the Mozambique Ocean9,11.

The Western Margin of the EAO (the eastern Saharan Metacraton)

The Saharan Metacraton is still very poorly known, but extensive late Mesoproterozoic subduction-related magmatism is found in Chad and west and north Sudan48. To the west of this, in eastern Sudan and western Ethiopia magmatism associated with early Neoproterozoic subduction characterizes terranes that are thought to have formed over westward dipping subduction zones44,49. This longevity of subduction, which also includes that seen in the Sinai36,37, demonstrates that the EAO extends back into the Stenian, or even earlier, when terrane accretion and subduction-zone magmatism initiated against the Paleoproterozoic kernel of the “metacraton”. The NE margin of the Congo Craton, in Uganda, preserves orogenesis that begins with Tonian subduction-zone magmatism in the Karamoja Belt that is coeval with terranes in Sudan50.

A Late Tonian Plate Reconfigurationn

The model presented here (Fig. 2) is developed from the reconstruction of Merdith et al.9. Details of both geological and paleomagnetic data used to constrain cratonic configurations, positions, and motions are provided therein. Here, data and observations discussed above for the terranes of northern Africa, Arabia, and NW India have been integrated10. These define a marked change in Mozambique Ocean subduction kinematics at ca. 720 Ma, from a predominately N–S to E–W striking subduction system (Fig. 2). The timing of these observed changes are broadly coeval with the start of Neoproterozoic India’s southern progression from polar regions to lower latitudes29,51,52 and we suggest that they represent a plate reorganization in this hemisphere. This kinematic shift is coeval with sedimentological and kinematic estimates for the breakup of the core of Rodinia. The ancestral Pacific basin is constrained to open before 725 Ma based on kinematic constraints12 and separation of the Kalahari and Congo continents is also consistent with voluminous magmatism along the southern Congo margin at ca. 750 Ma53. This late Tonian plate reorganization heralds the start of a shift of continental crust away from the northern hemisphere into the southern hemisphere (Figs. 2 and 3). This narrows, then eventually closes the equatorial Mozambique Ocean. The model implies that most continental blocks were concentrated in the southern hemisphere in the Late Neoproterozoic (Fig. 3). If this reflects the real distribution, then it would have interesting consequences for ocean/atmosphere circulation. One possible effect would be to compartmentalize ocean gyres in the southern hemisphere, while removing obstacles for hemispherical circulation in the northern hemisphere (Fig. 2). This shift from bi-hemisphere continent distribution towards a world with a pan-northern hemisphere ocean and continents concentrated in the southern hemisphere coincides with end of whole-earth glaciations that characterize the middle Neoproterozoic. In contrast to the Sturtian and Marinoan whole-earth glaciations, Ediacaran glaciations, such as the Gaskiers glaciation, appear more regional in scale15,54. Williams and Schmidt55 hypothesized that the mid-Ediacaran Shuram/Wonoka negative carbon isotope anomaly represents an unprecedented perturbation of the world ocean. We speculate that the plate tectonic driven bifurcation of the planet into continent and ocean latitudinal hemispheres may be a major control on this oceanic perturbation and climate switch—a consequence of the late Tonian plate reorganization.

Fig. 2: Global full-plate topological reconstructions.
figure 2

a 850 Ma, b 750 Ma, c 700 Ma, d 600 Ma modified from Merdith et al.9,10, focusing on detail within the ANS. A-A Afif-Abas, Am Amazonia, ANS Arabian-Nubian Shield terrane, Az Azania, Ba Baltic, Bo Borborema, Bu Butana, C Congo, Ca Cathaysia, Ch Chortis, DML Dronning Maud Land, G Greenland, H Hoggar, I Neoproterozoic India, K Kalahari, L Laurentia, Ma Mawson, NAC North Australian Craton, N-B Nigeria-Benin, NC North China, O Oman, Pp Paranapanema, R Rayner, RDLP Rio de la Plata, S Seychelles, SM Sahara Metacraton, SF Saõ Francisco, SMC Sa’al Metamorphic Complex, Sr Sri Lanka, Si Siberia, SAC South Australian/West Australian Craton, SES South Ethiopian Shield, Ta Tarim, WAC West African Craton, WES West Ethiopian Shield, Yg Yangtze. Colors refer to modern continents that the terranes now make up: Red = Asia, Blue = Africa, Gold = South America, Green = Europe, Yellow = North America, Purple = Australia, Teal = Antarctica, Lilac = India, Madagascar, and eastern Arabia. Blue dashed lines indicate schematic ocean currents.

Fig. 3: Latitudinal distribution of continental crust with respect to time.
figure 3

a Percentage of continental crust in either hemisphere as a function of time. Sturtian 717–663 Ma,57,58, Marinoan 645–635 Ma,15, and Gaskiers 579 Ma,54 glacial durations indicated. b Histograms of latitudinal distribution of continental area with respect to time (Ma). Both as implied in Merdith et al.10. If these models reflect reality, the late Tonian sees the beginning of a long-term trend in continental crust moving to southern latitudes, first into tropical regions through the Cryogenian, promoting a high albedo world. The Ediacaran sees a shift to lower latitudes, which would suppress albedo. C Cambrian.

Methods

This manuscript is based on a full-plate tectonic reconstruction of the last billion years that has been developed on the open access software GPlates (www.gplates.org). All GPlates files needed to reconstruct the model are available here—https://zenodo.org/record/3647901. The methodology for constructing full-plate tectonic reconstructions are detailed in Merdith et al.10 and also in Domeier and Torsvik56. Computation of latitudinal surface area was done using pyGplates (www.pygplates.org). To extract the latitudinal distribution of continent area, we first created a global equal-area mesh. We then used a grid-intersection between the nodes of the mesh and the polygons of the plate model that represent continental crust to estimate the area of crust at each latitude. These were summed for each timestep to create an array from 1000–520 Ma of the latitudinal distribution of crust on the earth.