Australia: The Land Where Time Began

A biography of the Australian continent 

Western Gondwana Craton – Modification by Plume-Lithosphere Interaction

The persistence of neutrally-to-positively buoyant and mechanically strong lithosphere which shields the cratonic crust from underlying mantle dynamics is generally believed to be responsible for the longevity of cratons. In this paper it is shown by Hu et al. that in South America and Africa large portions of the cratonic lithosphere, however, underwent significant modification during and following the Mesozoic, as is demonstrated by uplift and volcanism in the Cretaceous, high topography of the present, thin crust and the presence of seismically rapid though neutrally-buoyant anomalies in the upper mantle. It is suggested by Hu et al. that these observations reflect a permanent increase in lithosphere buoyancy that resulted from plume-triggered delamination of deep lithospheric roots that occurred during the Late Cretaceous and early Cainozoic. In these regions lithosphere has been re-established since then, as has been confirmed by its low heat flow of the present, high seismic velocities and realigned seismic anisotropy. Hu et al. concluded that the original lowermost cratonic lithosphere is compositionally denser than the asthenospheric mantle and when perturbed by underlying mantle upwelling can be removed. It is the buoyancy of the upper lithosphere, therefore, that perpetuates the stabilisation of cratons.

The stability and persistence of cratons, which are the long-lived portions of continents that are relatively stable, has been the focus of sustained research over the past 4 decades. According to the traditional hypothesis of isopycnicity lithospheres of cratonic mantle are highly depleted as a result of extensive melt extraction (1-4), which results in a highly-viscous thermal boundary layer that persists for billions of years that is neutrally to positively buoyant. Though this model is elegant, it does not easily explain apparent temporal variations in cratons, which include significant changes in elevation over time and episodic destruction of the deep lithosphere (3,4). It is suggested by recent measurements of seismic anisotropy that the upper portion, at <100-150 km depth of the cratonic lithosphere has a seismic fabric that differs from the underlying lithosphere (5-7). According to Hu et al. this new lithospheric structure is consistent with models that assume a compositionally stratified thermal boundary layer in which shallower depths are more depleted (8,9). Questions about the buoyancy and stability lithosphere are sparked by such models (2-4), especially with regards to density resulting from composition and temperature at different depths (8-10). Consequently, questions that concern how cratons remain stable for billions of years and their response to dynamic processes in the underlying asthenosphere are yet to be answered.

In this study Hu et al. investigated the density structure and temporal evolution of the cratonic lithosphere by the use of diverse observational constraints, which included topographic evolution, gravity anomalies, seismic tomography and anisotropy, tectonic reconstructions and mantle flow. In order to avoid the potential effects of subduction in the Mesozoic-Cainozoic on the evolution of the craton (2,11,12), they focused on cratons that border the passive margins of the South Atlantic. Precambrian regions of South America and Africa which host cratons and orogens from the Neoproterozoic (13) were included in the study area, regions that had in the past been part of western Gondwana. A prominent feature of these is that they have widespread surface topography of 1 km elevation or more either in the inside or on the edges of the São Francisco (SF), Congo (CG) and Kalahari (KH) cratons. In these regions the elevation is in sharp contrast to that of the other cratons such as the West African (WA), Amazonian (AZ) and North American (NA) cratons that are mostly of low topography (14,15).

Hu et al. say several key aspects related to the topographic evolution of the study area:

1)    The highest topography in the region is supported isostatically by sub-crustal lithospheric mantle.

2)    This high topography developed during and following the Late Mesozoic, as is indicated by the presence of earlier marine-lacustrine depositional environments (16,17) and of volcanic eruptions and associated surface uplift in the Cretaceous (20-22) in these regions.

3)    Beneath the study area the upper mantle contains anomalous zones which have high seismic velocity, though they are close to being neutrally buoyant, which is a feature that is characteristic of delaminated lithospheric mantle in cratons.

4)    Below these regions seismic anisotropy of lower cratonic lithosphere showing realignment to directions to the the mantle flow in the Cainozoic.

When taken together, the analyses of Hu et al. indicate that the western Gondwana cratonic lithosphere was modified by delamination that began in the Late Mesozoic, as a result of its interaction with mantle plumes.

Origins of high cratonic topography

This crust, that is relatively thin, cannot compensate for high topography, and therefore must reflect positive buoyancy from the underlying mantle. Density differences between cratonic and non-cratonic crust also cannot explain the topographic variation that is observed, as the high topography occurs within only part of the CG craton.

In order to gain further insight into the variation of the topography, Hu et al. estimated the lithospheric residual topography by removing contributions of the crust and the sublithospheric mantle, neither of which can explain the high surface topography.

As well as the lack of correlation of the crustal thickness with topography, as has been discussed above, neutral-to-negative dynamic topography has been revealed in the study region by recent geodynamic studies (28-30), a result that was confirmed by the calculations of Hu et al. that used recent tomographic images. As a consequence, the lithospheric residual topography displays a peak-to-trough variation of up to 2 km, with the positive areas correlating with the topography of the prominent land surface. Most of the residual topography is present inside the cratons, with only a minor portion over the Neoproterozoic (Braziliano or Pan African) orogens. It is suggested by this result that high cratonic topography, in regions where the crust is thin, reflects heterogeneities in lithospheric buoyancy. This conclusion is supported further by a similar analysis of the gravity field, with negative lithospheric gravity anomalies occurring in regions of high topography.

Hu et al. examined seismic, heat-flow, and xenolith data, in order to evaluate whether the heterogeneity of the lithospheric buoyancy that is inferred has a thermal or compositional origin. It is indicated that cold lithosphere extends down to depths of more than 200 km (34,37)  by the presence of high shear velocities  (31-36), low attenuation (37) and cold geotherms that are inferred from heat flow (38) beneath both regions of high topography of KH, CG and SF and the 2 cratons of low topography of WA and AZ. Thermal thickness of lithosphere is inferred from shear velocities (34) vary by less than 25 km among the cratons KH, CG, SF,WA AND AZ.

Also, residual topography does not consistently correlate with regional variations in thermal buoyancy of the mantle that is inferred by attenuation (37), either within or between the cartons. The thermal thickness of the lithosphere that is inferred from heat flow (38) varies by less than 100 km, with cratons of high topography being slightly thinner. Similar to slightly higher (by less than 150 k) temperature beneath the HK craton compared to intact cratons such as the Slave, are revealed by estimates of lithospheric temperature from mantle xenoliths. The associated thermal buoyancy could account for less than 500 metres of topography, which contrasts sharply with the variation that is much larger, up to 2 km, of lithosphere residual topography, when an upper limit of 150 K of these temperature variations that are inferred within the lower 100 km of the lithosphere is taken. Therefore it is concluded by Hu et al. that a significant fraction of the apparent lithospheric buoyancy reflects composition.

Uplift in the Cretaceous due to lithospheric delamination

Hu et al. also evaluated the temporal evolution of cratons in order to understand better the origin of high topography cratons. Since the Cretaceous the SF craton has gained more than 1 km of surface elevation, relative to the adjacent AZ craton (16). As a result, shallow marine to lacustrine sedimentary strata dating from the Late Jurassic to Early Cretaceous in eastern Brazil have been elevated to well above sea level (16). According to Hu et al. most of the uplift probably occurred during the Late Cretaceous associated with denudation of more than 3 km in eastern Brazil; with the greatest denudation occurring along the northeastern edge of AZ (16) and over the entire SF (20,21), regions that have positive residual topography at the present. Similarly, southern Africa, which is a region that had undergone subsidence during the Jurassic, as is indicated by the presence of sedimentary deposits that date to this age (17) – experiencing rapid uplift and denudation during the Cretaceous, as has been revealed by kimberlite diatremes (19), by thermochronology studies of the KH (22), and by the evaluation of the history of sedimentation along the southern coast of Africa (39).

It is not considered to be likely that dynamic topography due to sublithospheric convection was the cause of uplift during the Cretaceous and high residual topography of the present, for 2 reasons.

1)    Subsidence, not uplift, in Eastern South America (28,29), has been revealed by recent estimates of changes in dynamic topography that is due to mantle convection since 100 Ma. Similarly, little uplift took place in southern Africa during the Cainozoic (28,29).

2)    Residual topography that has been observed in the study area has a wavelength of less than 500 km, though it is usual for dynamic topography to occur a wavelength of more than 1,000 km.

Hu et al. also examined the possible role of lithospheric delamination in increasing buoyancy of cratonic lithosphere, because delamination can cause isostatic uplift, as dynamic topography in unable to explain uplift that has been observed. During the Late Cretaceous a rapid increase in heat flow in South America and southern Africa (19,21), and contemporaneously, the disappearance of high-pressure garnet facies in South America (21), has provided independent evidence of the removal of the lowermost lithosphere. Also, the temporal correlation among the 2-pulse history of kimberlite eruption with a composition that is more fertile from the 2^nd phase (40), and the sedimentation rate circum-Africa (39), and the unroofing history of KH, further support the regional modification of southern African lithosphere (40).

According to Hu et al. mantle upwelling in plumes in regions that are far from subduction zones represent the most likely candidate for the triggering of delamination of lithosphere. It was found by reconstructing the trajectory of Atlantic hotspots (41) back to the Early Cretaceous by the use of recent plate reconstruction (43), that hotspot tracks from the Cretaceous coincide with regions where the topography is high in South America and southern Africa. The matching of hotspots with volcanism that dates to the Late Cretaceous in Brazil and South Africa is an independent validation of this reconstruction. It is implied by this correlation that there is a potential causal relationship among these phenomena. The most intense kimberlitic volcanism from the Late Cretaceous in the SF and KH cratons primarily occurred along their edges, which possibly implies the location of initial lithospheric delamination. In southern Africa thermochronology studies reveal there was a migration of exhumation from the edge to the interior of KH during the timespan 90-60 Ma (22), which is consistent with a progressive peeling off of cratonic lithosphere towards the centre.

A local increase in buoyancy and surface uplift can result from warming of the lithosphere and delamination of high density materials. It is suggested by the coherently fast seismic velocities of the thick lithospheres (31-36) and the low surface heat flow of the present (38) that thermal perturbation that was caused by the plume-lithosphere interaction in the Cretaceous has now largely dissipated. This disappearance is consistent with the time that would be required for the delaminated lower lithosphere to be replaced by a new thermal boundary layer. The high topography of the present must, therefore, reflect a permanent loss of a compositionally dense layer from the original lithosphere.

It has been revealed by 3-D representation of mantle seismic structure that there are multiple voluminous anomalies of high velocity within the sublithospheric mantle below the margins of the south Atlantic. It seems these mantle structures are well resolved, as they are similar among different topographic models. These anomalies could not represent slabs that have been subducted as subduction has not occurred in these regions since the early Mesozoic. They were previously interpreted and cold downwelling along lithospheric margins, as a result of edge-driven convection (43). It is suggested by Hu et al. that instead they represent the foundered segments of lithospheric mantle.

A separate study in which Hu et al. evaluated quantitatively the mantle flow associated with the subduction of the Nazca Plate relative to other anomalies of high velocity by matching simultaneously predictions of mantle flow to observations of surface wave anisotropy and shear wave splitting (44), supported this concept. Fits to regional and local anisotropy measurements require the Nazca slab to be the dominant cause of downwelling and that other high-velocity anomalies are mostly drifting passively and generating negligible vertical mantle flow. The upper-mantle anomalies of the south Atlantic are required by this result to be close to neutral buoyancy.

According to the model of Hu et al. (44), the mantle underlying an overriding plate, such as South America, undergoes Poiseuille flow (pressure driven) that is driven by a pressure gradient. Beneath plates that are not close to a subduction zone, in contrast, mantle undergoes plate flow that is driven by Couette flow (drag driven). As a consequence, mantle structure of the present further supports the lithospheric delamination hypothesis in that the geometry and the location of these anomalies of high velocity are best explained as fragments of foundered lithosphere left behind by South America and Africa as they moved away from each other since the Cretaceous. The stronger westward upper-mantle Poiseuille flow drawn by the down-going Nazca slab (44) may be reflected in the tendency for the delaminated materials to remain on the South American side of the Atlantic.

Cainozoic realignment of lithospheric seismic anisotropy

Seismic anisotropy that is similar to that of recent asthenospheric flow should be displayed by lithosphere that has undergone shear during the Cainozoic. Another important test of the model of Hu et al. (44) can be provided by comparison of azimuthal anisotropy (6) with that of what is predicted by Cainozoic mantle flow, as a function of depth. Robust predictions of anisotropy within the asthenosphere underlying continental South America and the surrounding oceans allows the application of modelling results to the interior of the cratonic lithosphere.

It is suggested by the lack of correlation between the observed anisotropy in the uppermost <100 km of the mantle with that of the asthenosphere that fabrics at this depth are mostly ‘fossil’ signals, meaning relicts of shear prior to the Cainozoic. The correlation clearly improves in most non-cratonic regions at depths of ≥100 km, the correlation clearly improves in most non-cratonic regions, which is consistent with their thin lithospheres and an increasing influence of shear in the Cainozoic. Anisotropy that is compatible with shear from the Cainozoic at greater than 150 km depth is not displayed by the region of the East African Rift, which Hu et al. suggest is probably because of perturbations by upwelling of the upper mantle. Contrasting with this, anisotropy correlation is poor below the low topography cratons, which are inferred to have intact mantle lithospheres (AZ and WA). It is emphasised by this lack of correlation that deformation of the mantle did not occur within the intact cratonic lithosphere during the Cainozoic (5,6).

In most cratonic regions of high topography, however, anisotropy that is observed aligns with mantle shear that has been predicted at depths as shallow as 100 km, such as in SF. In cratonic  regions with the most high topography, however, anisotropy that is observed aligns with mantle shear that has been predicted at depths as shallow as 100 km, such as in SF. When a different anisotropy model (45) was used there was a similar correlation of anisotropy that is observed and predicted results. It is suggested that the lithospheric fabrics of these cratons were reset by this obvious regional realignment of lower lithosphere anisotropy with mantle flow in regions where there was modification during the Mesozoic. This conclusion was confirmed by another anisotropy model (46) for depths of less than 150 km, though at greater depths significant discrepancy is displayed. As this model fails to match their asthenospheric deformation, that is well-predicted, and probably also fails to represent the shear wave-splitting observations over South America (44), Hu et al. suggest that the other 2 anisotropy models (6,45) represent better lowermost lithosphere fabrics in the study area.

Mechanically, resetting of fabric could be the result of either local deformation of the original lithosphere (11) or growth of a new thermal boundary layer following delamination. Hu et al. suggest a combination of both mechanisms. Isostatic uplift of topography, on the one hand, requires dense lithospheric materials to be removed. On the other hand, the high velocity (more than 2 %) lithosphere at 200 km and below requires compositional (e.g. garnet or MORB) as well as their thermal effects. Hu et al. therefore propose that delamination occurred mostly within the lowermost cratonic lithosphere where there are high density materials, and that lithosphere was not necessarily removed entirely below about 100 km, instead undergoing shear that is sufficient to realign it crystallographic anisotropy with that of the ambient mantle. It is suggested by Hu et al. that this concept can be verified further by an examination of the spatial extent of anisotropy realignment, where the area in Africa that is realigned is much larger than the zones of high residual topography, with the former also correlating closely with the province of kimberlites from the Cretaceous (47).

Implications for the density and evolution of cratonic lithosphere

The density of the lowermost cratonic lithosphere must be greater than that of the surrounding asthenosphere for it to delaminate. Also, as the craton topography of the presence in the study area rises significantly higher than the pre-delamination topography that is reflected in uplift that occurred in the Cretaceous and present positive residual topography, the lithosphere must have also been denser than the thermal boundary layer that formed subsequently. According to Hu et al. this proposal explains why ‘intact cratons’, such as WA and AZ, have topography that is lower than that of ‘delaminated cratons’, e.g. SF, CG and KH. The proposal of Hu et al. contrasts with the traditional view that cratonic roots have approximately buoyancy at all depths (1,3,4). Instead, Hu et al. proposed that in cratons the upper lithospheric mantle is highly depleted and chemically buoyant, possibly even more so than has yet been inferred from xenoliths in the mantle (4,40). The lower lithosphere in cratons is more fertile, and grades downwards into a boundary layer that is purely thermal, which contains zones or layers at more than 200 km depth enriched in high density minerals.

The model of Hu et al. is consistent with the occurrence of garnet-peridotite in lower portions of cratons that are intact, such as Slave and KH cratons that are from pre-Cainozoic times (3,4,21,40), and with high seismic velocities in the average lowermost 150-200 km cratonic lithosphere (35). Assuming a thickness of 50 km of this high-density layer, 20% excess garnet-peridotite relative to a lowermost lithosphere that is neutrally buoyant are required by the residual topography and gravity, if the density of garnet-peridotite is about 10% higher than that of peridotite (4). Hu et al. proposed that this high-density layer may have been emplaced during the formation of the craton (48) and/or represents secular accumulation of basaltic melts from the deep mantle (49).

The lithosphere-density model of Hu et al. also explains several enigmatic properties of the prominent high velocity seismic anomalies of the upper mantle below the south Atlantic. Included among these are anomalies in regions that are far from subduction zones. Notably, delaminated lithosphere that consists of dense garnet-peridotite and buoyant harzburgite should have small initial negative buoyancy with a magnitude that would also decrease with increases of depth and temperature. Within the warming delaminated root at a depth of 660 km a delayed post-garnet transformation would generate additional positive buoyancy and prohibit further penetration into the lower mantle. The stagnation of the delaminated materials at the base of the upper mantle is explained:

1)    First, by this concept, and

2)    second, their minimal net buoyancy, therefore avoiding the generation of local convection, as is required by seismic anisotropy.

3)    Finally, the high seismic velocities of these anomalies are consistent with the enrichment of harzburgite and garnet at temperatures that are relatively low.

It was noted by Hu et al. that the minimum depth of about 100 km of the anisotropy realignment correlates broadly with the mid-lithospheric discontinuities (MLDs) that are commonly present beneath continents (26,51-54). It is implied by this correlation that the MLD may represent a layer that is mechanically weak inside the lithosphere, below which shear deformation of more than 100 km and ultimately delamination at greater than 150-200 km depth could occur more easily than in the layer of lithosphere above. Hu et al. suggest this conclusion is consistent with recent suggestions that the MLD represents the top of a weak layer of solidified volatile-rich melt accumulation (49,55). A decoupling surface between the upper and lower lithosphere could be served by such a weak layer during extension (56), thereby facilitating mobilisation and delamination of dense deep lithosphere.

The results presented in this paper, when taken together, indicate that lowermost cratonic mantle lithosphere may be removed episodically when mantle processes perturb them sufficiently, while the buoyant lithosphere that is depleted tends to remain stable with the result that its crust maintains cratonic characteristics over geological time. The key features of the destabilisation of cratonic lithosphere, during which deformation and removal of the dense root facilitates kimberlite volcanism (18-22), isostatic uplift of the surface, and shallow Moho and thin crust. It is further proposed by Hu et al. that zones in which lower lithosphere was removed takes 10s of millions of years to recover thermally, though the density of the new thermal root would be less than that of the intact root. Cratons without hotspot activity in the Phanerozoic and subduction that is nearby, such as those in the Northern Hemisphere, have not undergone delamination since the Mesozoic and therefore display negative residual topography and mantle gravity anomalies that are positive, which contrasts with cratons that are affected by subduction, such as those in East Asia (11) and western North America (12), and cratons that are affected by mantle plumes, such as those in Southern Hemisphere.

Sources & Further reading

Hu, J., et al. (2018). "Modification of the Western Gondwana craton by plume–lithosphere interaction." Nature Geoscience 11(3): 203-210.

1 Jordan, T. H. Composition and development of the continental tectosphere.

Nature 274, 544–548 (1978).

2. Durrheim, R. J. & Mooney, W. D. Evolution of the precambrian lithosphere:

seismological and geochemical constraints. J. Geophys. Res. 99,

15359–15374 (1994).

3. Carlson, R. W., Pearson, D. G. & James, D. E. Physical, chemical, and

chronological characteristics of continental mantle. Rev. Geophys. 43,

RG1001 (2005).

4. Lee, C.-T. A., Luffi, P. & Chin, E. J. Building and destroying continental

mantle. Annu. Rev. Earth Planet. Sci. 39, 59–90 (2011).

5. Yuan, H. & Romanowicz, B. Lithospheric layering in the North American

craton. Nature 466, 1063–1069 (2010).

6. Yuan, K. & Beghein, C. Seismic anisotropy changes across upper mantle

phase transitions. Earth. Planet. Sci. Lett. 374, 132–144 (2013).

7. Debayle, E. & Kennett, B. L. N. The Australian continental upper mantle:

structure and deformation inferred from surface waves. J. Geophys. Res. Solid

Earth 105, 25423–25450 (2000).

8. King, S. D. Archean cratons and mantle dynamics. Earth Planet. Sci. Lett.

234, 1–14 (2005).

9. Eaton, D. W. & Perry, H. K. C. Ephemeral isopycnicity of cratonic mantle

keels. Nat. Geosci. 6, 967–970 (2013).

10. Kaban, M. K., Mooney, W. D. & Petrunin, A. G. Cratonic root beneath North

America shifted by basal drag from the convecting mantle. Nat. Geosci. 8,

797–800 (2015).

11. Griffin, W. L., Andi, Z., O’Reilly, S. Y. & Ryan, C. G. Phanerozoic evolution of

the lithosphere beneath the Sino-Korean Craton. Mantle Dyn. Plate Interact.

East Asia 27, 107–126 (1998).

12. Levander, A et al. Continuing Colorado plateau uplift by delamination-style

convective lithospheric downwelling. Nature 472, 461–465 (2011).

13. Alkmim, F. F. et al. Kinematic evolution of the Araçuaí–West Congo orogen

in Brazil and Africa: nutcracker tectonics during the Neoproterozoic assembly

of Gondwana. Precam. Res. 149, 43–64 (2006).

14. Kaban, M. K., Schwintzer, P., Artemieva, I. M. & Mooney, W. D. Density of

the continental roots: compositional and thermal contributions. Earth Planet.

Sci. Lett. 209, 53–69 (2003).

15. Mooney, W. D. & Kaban, M. K. The North American upper mantle: density,

composition, and evolution. J. Geophys. Res. 115, B12424 (2010).

16. Arai, M. Chapadas: relict of mid-Cretaceous interior seas in Brazil. Rev. Bras.

Geoci. 30, 436–438 (2000).

17. Catuneanu, O. et al. The Karoo basins of south-central Africa. J. Afr. Earth

Sci. 3, 211–253 (2005).

18. Harman, R., Gallagher, K., Brown, R., Raza, A. & Bizzi, L. Accelerated

denudation and tectonic/geomorphic reactivation of the cratons of

northeastern Brazil during the Late Cretaceous. J. Geophys. Res. 103,

27091–27105 (1998).

19. Hanson, E. K. et al. Cretaceous erosion in central South Africa: Evidence

from upper-crustal xenoliths in kimberlite diatremes. South Afr. J. Geol. 112,

125–140 (2009).

20. Cogné, N., Gallagher, K. & Cobbold, P. R. Post-rift reactivation of the

onshore margin of southeast Brazil: evidence from apatite (U–Th)/He and

fission-track data. Earth Planet. Sci. Lett. 309, 118–130 (2011).

21. Read, G. et al. Stratigraphic relations, kimberlite emplacement and

lithospheric thermal evolution, Quiricó Basin, Minas Gerais State, Brazil.

Lithos 77, 803–818 (2004).

22. Stanley, J. R., Flowers, R. M. & Bell, D. R. Kimberlite (U–Th)/He dating links

surface erosion with lithospheric heating, thinning, and metasomatism in the

southern African Plateau. Geology 41, 1243–1246 (2013).

23. Laske, G., Masters, G., Ma, Z. & Pasyanos, M. Update on CRUST1.0 - A

1-degree Global Model of Earth’s Crust. Geophys. Res. Abstr. 15,

EGU2013–2658 (2013).

24. Reid, A. B., Ebbing, J. & Webb, S. J. Comment on ‘A crustal thickness map of

Africa derived from a global gravity field model using Euler deconvolution’

by Getachew E. Tedla, M. van der Meijde, A. A. Nyblade and F. D. van der

Meer. Geophys. J. Int. 189, 1217–1222 (2012).

25. Assumpção, M., Feng, M., Tassara, A. & Julià, J. Models of crustal thickness

for South America from seismic refraction, receiver functions and surface

wave tomography. Tectonophysics 609, 82–96 (2013).

26. Liu, L., K. Liu and S. Gao. Lithospheric layering beneath southern Africa

constrained by S-to-P receiver functions. In AGU Fall General Assembly

2016. abstr. DI51A-2660 (American Geophysical Union, 2016).

27. Globig, J. et al New insights into the crust and lithospheric mantle structure

of Africa from elevation, geoid, and thermal analysis. J. Geophys. Res. Solid

Earth 121, 5389–5424 (2016).

28. Shephard, G. E., Müller, R. D., Liu, L. & Gurnis, M. Miocene drainage

reversal of the Amazon River driven by plate-mantle interaction. Nat. Geosci.

3, 870–875 (2010).

29. Flament, N., Gurnis, M. & Müller, R. D. A review of observations and models

of dynamic topography. Lithosphere 5, 189–210 (2012).

30. Moucha, R. & Forte, A. M. Changes in African topography driven by mantle

convection. Nat. Geosci. 4, 707–712 (2011).

31. French, S., Lekic, V. & Romanowicz, B. Waveform tomography reveals

channeled flow at the base of the oceanic asthenosphere. Science 342,

227–230 (2013).

32. Ritsema, J., Deuss, A., van Heijst, H. J. & Woodhouse, J. H. S40RTS: a

degree-40 shear-velocity model for the mantle from new Rayleigh wave

dispersion, teleseismic traveltime and normal-mode splitting function

measurements. Geophys. J. Int. 184, 1223–1236 (2011).

33. Pasyanos, M. E., Masters, T. G., Laske, G. & Ma, Z. LITHO1.0: an updated

crust and lithospheric model of the Earth. J. Geophys. Res. 119,

2153–2173 (2014).

34. Priestley, K. & McKenzie, D. The relationship between shear wave velocity,

temperature, attenuation and viscosity in the shallow part of the mantle.

Earth Planet. Sci. Lett. 381, 78–91 (2013).

35. Adams, A. & Nyblade, A. Shear wave velocity structure of the southern

African upper mantle with implications for the uplift of southern Africa.

Geophys. J. Int. 186, 808–824 (2011).

36. Feng, M., Assumpção, M. & Van der Lee, S. Group velocity tomography and

lithospheric S-velocity structure of the South American continent. Phys. Earth

Planet. Inter. 147, 315–331 (2007).

37. Dalton, C. A., Bao, X. & Ma, Z. The thermal structure of cratonic lithosphere

from global Rayleigh wave attenuation. Earth Planet. Sci. Lett. 457,

250–262 (2017).

38. Artemieva, I. Global 1 degrees x 1 degrees thermal model TC1 for the

continental lithosphere: Implications for lithosphere secular evolution.

Tectonophysics 416, 245–277 (2006).

39. Guillocheau, F. et al. Quantification and causes of the terrigeneous sediment

budget at the scale of a continental margin: a new method applied to the

Namibia–South Africa margin. Basin Res. 24, 3–30 (2012).

40. Griffin, W. L. et al. The origin and evolution of Archean lithospheric mantle.

Precambrian Res. 127, 19–41 (2003).

41. Courtillot, V., Davaille, A., Besse, J. & Stock, J. Three distinct types of

hotspots in the Earth’s mantle. Earth Planet. Sci. Lett. 205, 295–308 (2003).

42. Müller, R. D. et al. Ocean basin evolution and global-scale plate

reorganization events since Pangaea breakup. Annu. Rev. Earth Planet. Sci. 44,

107–138 (2016).

43. King, S. & Ritsema, J. African hot spot volcanism: small-scale convection in

the upper mantle beneath cratons. Science 290, 1137–1140 (2000).

44. Hu, J., Faccenda, M. & Liu, L. Subduction-controlled mantle flow and seismic

anisotropy in South America. Earth Planet. Sci. Lett. 470, 13–24 (2017).

45. Schaeffer, A. J., Lebedev, S. & Becker, T. W. Azimuthal seismic anisotropy in

the Earth’s upper mantle and the thickness of tectonic plates. Geophys. J. Int.

207, 901–933 (2016).

46. Debayle, E., F. Dubuffet, and S. Durand, An automatically updated S-wave

model of the upper mantle and the depth extent of azimuthal anisotropy.

Geophys. Res. Lett. 43, 674–682, (2016).

47. Yaxley, G. M. et al. The discovery of kimberlites in Antarctica extends the

vast Gondwanan Cretaceous province. Nat. Commun. 4, 2921 (2013).

48. Walter, M. J. Melting of garnet peridotite and the origin of komatiite and

depleted lithosphere. J. Petrol. 39, 29–60 (1998).

49. Rader, E. et al. Characterization and petrological constraints of the

midlithospheric discontinuity. Geochem. Geophys. Geosys. 16,

3484–3504 (2015).

50. Stixrude, L. & Lithgow-Bertelloni, C. Thermodynamics of mantle minerals-II.

Phase equilibria. Geophys. J. Int. 184, 1180–1213 (2011).

51. Wittlinger, G. & Farra, V. Converted waves reveal a thick and layered

tectosphere beneath the Kalahari super-craton. Earth Planet. Sci. Lett. 254,

404–415 (2007).

52. Sodoudi, F. et al. Seismic evidence for stratification in composition and

anisotropic fabric within the thick lithosphere of Kalahari Craton. Geochem.

Geophys. Geosys. 14, 5393–5412 (2013).

53. Selway, K., Ford, H. & Kelemen, P. The seismic mid-lithosphere discontinuity.

Earth Planet. Sci. Lett. 414, 45–57 (2015).

54. Fischer, Karen M., Ford, Heather A., Abt, David L., Rychert, Catherine A. The

Lithosphere–Asthenosphere boundary. Ann. Rev. Earth Planet. Sci. 38

551–575, (2010).

55. Chen, L. et al. Presence of an intralithospheric discontinuity in the central

and western North China Craton: Implications for destruction of the craton.

Geology 42, 223–226 (2014).

56. Liao, J., Gerya, T. & Wang, Q. Layered structure of the lithospheric mantle

changes dynamics of craton extension. Geophys. Res. Lett. 40, 5861–5866 (2013).

57. Jelsma, H., Barnett, W., Richards, S. & Lister, G. Tectonic setting of

kimberlites. Lithos 112S, 155–165 (2009).