news & views

GEODYNAMICS

When plumes tickle continents

Continental stability may be linked to a shallow, buoyant mantle layer, and the deepest craton roots can be
destabilized and removed by mantle plumes.

Eric Debayle

C
ratons form the core of continents. long-term stability of cratons is due to the and why cratons remain unchanged for
Seismic images1 show that these old shallower buoyant lithosphere only. billions of years.
and stable Precambrian cores are If continental roots are cold, they should Hu et al.5 propose that the answer to
underlain by high-velocity roots or keels that in principle be denser than the surrounding this question lies in the buoyancy of the
can extend to depths greater than 200 km. mantle at the same depth and sink. However, shallow cratonic lithosphere. They use a
These keels are probably composed of cold, analysis of xenolith samples reveals that combination of topography, gravity and
dry and highly viscous upper mantle rocks2,3. cratonic roots are depleted in dense basaltic seismic data to analyse the structure of
Analysis of peridotite xenoliths, mantle components compared to the surrounding cratons in South America and Africa,
rocks erupted at the surface, indicate that mantle. So, the excess density due to cooler which once formed the western part of
these keels formed early in the history of temperatures could be cancelled out by the supercontinent Gondwana. Some of
cratons4. The great antiquity of keels implies the light composition6, providing a way these cratons display widespread high
that the deep roots of the continents are to steady the highly viscous and cold but surface topography of 1 km or more.
stable and have largely been preserved depleted keels beneath cratons over billions The researchers demonstrate that the
from the erosive influence of mantle of years. This classical view of keel stability high topography is supported by buoyant
convection since their formation. Writing has, however, been refined as a number of lithospheric heterogeneities: the crust is
in Nature Geoscience, however, Hu et al.5 studies revealed that the lithosphere beneath too thin to support the high topography,
report that the deepest cratonic lithosphere cratons may be stratified7 and the mantle and the dynamic topography created by
can be removed when perturbed by mantle root of many cratons has been removed or the convective sub-lithospheric mantle
upwellings, challenging the classical view of modified8. If the roots beneath cratons extends beneath both elevated and lower
stable cratonic roots. They propose that the can be destabilized, it is then unclear how areas, and thus cannot be the cause either.

Ocean Craton with high topography Sedimentary basin

Crust
Buoyant upper lithosphere
Oceanic lithosphere (highly depleted)

Mantle Restored roots Lower lithosphere

plume (mostly thermal) (less depleted)

Anisotropy aligned with Cenozoic mantle flow

410 km

660 km

Delaminated roots
(neutrally buoyant)

Fig. 1 | Schematic illustration of the present-day lithosphere structure for South American and African cratons. Hu et al.5 use a variety of data sets to
argue that the stability of these cratons relies on the shallow and buoyant upper lithosphere. They suggest that the dense roots (blue) of some cratons and
eventually some parts of the lower lithosphere were removed by Cenozoic mantle plumes and replaced with more buoyant mantle. The delaminated roots
sank and became trapped in the mantle transition zone. The lost cratonic roots are then restored to their original positions (green layer). Patterns of Cenozoic
mantle flow (seismic anisotropy, dark half-barbed arrows) are frozen in the restored lithosphere and in some parts of the lower lithosphere. Figure adapted
from ref. 5, Macmillan Publishers Ltd.

Nature Geoscience | www.nature.com/naturegeoscience

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
news & views

Additionally, xenolith data show that lower density than the original delaminated suggest that cratons affected by subduction
temperature variations are only moderate lithosphere, so that cratons that were once have experienced a different story.
at less than 150 K, so thermal buoyancy delaminated have increased their buoyancy Hu et al.5 propose that plume-induced
cannot be responsible for the raised and maintained high topography. Patterns of delamination can destabilize a high-
topography. Instead, the uplift must be seismic anisotropy in the restored cratonic density layer located at the base of cratons
sustained by compositional heterogeneities roots align fairly well with the Cenozoic and eventually some parts of the lower
in the shallow lithosphere. patterns of mantle flow in the lower lithosphere. The buoyancy of the shallower
Seismic images reveal anomalous lithosphere beneath the high-topography lithosphere ensures the long-term stability
regions in the mantle transition zone, where cratons, adding support to the authors’ of cratons, whereas the delaminated root
seismic waves travel unusually fast. Hu and model of root delamination and would be thermally restored in a relatively
colleagues propose that these anomalies Cenozoic restoration. short period of time. This view of cratonic
represent remnants of cratonic roots that The multiple high-velocity anomalies evolution may explain how cratonic cores
have been removed by mantle plume activity observed in the mantle transition zone were remain stable in the light of an emerging
and then sunk into the mantle. Sedimentary previously interpreted as small-scale mantle picture of remarkably layered and complex
deposits constrain the timing of craton convection at the edge of cratons9, but it cratonic roots. ❐
uplift to the Cretaceous, so the researchers now seems the anomalies might instead be
reconstruct the trajectory of southern delaminated parts of cratonic roots. The Eric Debayle
Atlantic hotspots, thought to be caused by anomalies do not perturb the patterns of Laboratoire de Géologie de Lyon: Terre, Planètes,
mantle plumes, back to the Early Cretaceous. seismic anisotropy, so they are probably Environnement, Université Lyon 1, Ecole Normale
They find that Cretaceous-aged hotspot neutrally buoyant and stay trapped in the Supérieure de Lyon and CNRS, Villeurbanne, France.
tracks indeed coincide with regions of high transition zone. e-mail: Eric.Debayle@ens-lyon.fr
topography in both South America and The variety of observations used in the
Africa. This correlation supports the idea research each has their own limitations Published: xx xx xxxx
that Cretaceous mantle plumes could have (one may, for example, argue that the https://doi.org/10.1038/s41561-018-0074-z
triggered delamination of a high-density interpretation of anisotropy is at the limit
References
layer, enriched in garnet-peridotite at the of current models resolution). However, 1. Gung, Y., Panning, M. & Romanowicz, B. Nature 422,
base of the cratonic lithosphere, at depths taken together the observations provide a 707–711 (2003).
greater than 200 km. If correct, tomographic set of constraints for a novel scenario 2. Jaupart, C., Mareschal, J. C. & Guillou-Frottier, L. J. Geophys. Res.
103, 15269–15286 (1998).
and heat flow data imply that the thermal of cratonic evolution. It is important to 3. Hirth, G., Evans, R. L. & Chave, A. D. Geochem. Geophys. Geosyst.
perturbation caused by plume–lithosphere now test the proposed scenario. Mapping 1, C000048 (2000).
interaction during the Cretaceous has now simultaneously seismic attenuation and 4. Pearson, D. G., Carlson, R. W., Shirey, S. B., Boyd, F. R. & Nixon, P. H.
Earth Planet. Sci. Lett. 134, 341–357 (1995).
largely disappeared. velocity may help disentangle the thermal 5. Hu, J. et al. Nat. Geosci. https://doi.org/10.1038/s41561-018-
The authors demonstrate that or compositional origin of lithospheric 0064-1 (2018).
delaminated cratonic roots can be thermally heterogeneities, and therefore distinguish 6. Jordan, T. H. Nature 274, 544–548 (1978).
7. Yuan, H. & Romanowicz, B. Nature 466, 1063–1069 (2010).
restored back to their original positions intact from restored lithosphere. We must
8. Carlson, R. W., Pearson, D. G. & James, D. E. Rev. Geophys. 43,
over a period of 90 million years (Fig. 1). also check whether this model applies RG1001 (2005).
Importantly, the restored lithosphere has elsewhere. A number of observations 9. King, S. & Ritsema, J. Science 290, 1137–1140 (2000).

Nature Geoscience | www.nature.com/naturegeoscience

© 2018 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.