Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 91 of the 91 references it cites were found in a registry or an index; 25 of the 61 sources its claims cite were checked for support.

Initiation of plate tectonics from post‐magma ocean thermochemical convection

Open access
Foley, Bradford J., Bercovici, David, Elkins‐Tanton, Linda T. · 2014 · Journal of Geophysical Research: Solid Earth

Abstract

Abstract Leading theories for the presence of plate tectonics on Earth typically appeal to the role of present day conditions in promoting rheological weakening of the lithosphere. However, it is unknown whether the conditions of the early Earth were favorable for plate tectonics, or any form of subduction, and thus, how subduction begins is unclear. Using physical models based on grain‐damage, a grainsize‐feedback mechanism capable of producing plate‐like mantle convection, we demonstrate that subduction was possible on the Hadean Earth (hereafter referred to as proto‐subduction or proto‐plate tectonics), that proto‐subduction differed from modern day plate tectonics, and that it could initiate rapidly. Scaling laws for convection with grain‐damage show that though either higher mantle temperatures or higher surface temperatures lead to slower plates, proto‐subduction, with plate speeds of ≈1.75 cm/yr, can still be maintained in the Hadean, even with a CO 2 rich primordial atmosphere. Furthermore, when the mantle potential temperature is high (e.g., above ≈2000 K), the mode of subduction switches to a “sluggish subduction” style, where downwellings are drip like and plate boundaries are diffuse. Finally, numerical models of post‐magma ocean mantle convection demonstrate that proto‐plate tectonics likely initiates within ∼100 Myr of magma ocean solidification, consistent with evidence from Hadean zircons. After the initiation of proto‐subduction, non‐plate‐tectonic “sluggish subduction” prevails, giving way to modern style plate tectonics as both the mantle interior and climate cool. Hadean proto‐subduction may hasten the onset of modern plate tectonics by drawing excess CO 2 out of the atmosphere and cooling the climate.

Overview

CoChat Summary

TL;DR

Using grain-damage scaling laws and numerical post-magma-ocean convection models, the paper argues that Hadean proto-subduction was geodynamically possible and could initiate rapidly, even with high mantle and surface temperatures. This early subduction was likely sluggish and drip-like rather than modern plate tectonics, and may have helped cool the climate and mantle.

Key findings

  • Scaling laws show proto-subduction can be maintained in the Hadean: at \(T_i = 2100\) K and \(T_s = 500\) K, plate speed is \(\approx 1.75\) cm/yr; at present-day surface temperature, plate speed decreases from \(\approx 5.25\) cm/yr at \(T_i = 1600\) K to \(\approx 3.25\) cm/yr at \(T_i = 2100\) K, and at a Venusian \(T_s = 750\) K it is \(\approx 0.4\) cm/yr.
  • At high mantle potential temperatures (at or above \(\approx 2000\) K for \(T_s = 273\) K), subduction switches to “sluggish subduction,” with drip-like downwellings and diffuse plate boundaries, though appreciable lithospheric recycling remains.
  • Numerical models indicate proto-plate tectonics likely initiates within \(\sim 100\) Myrs of magma ocean solidification; the lag time is \(\approx 100\) Myrs at \(T_i \approx 2000\) K and decreases to \(\approx 100\) yrs at \(T_i \approx 2100\) K for \(f \approx 10^{-9}\), with overturn times of \(\sim 10^4\) yrs for \(R_s = 300\) km and \(\sim 10^6\) yrs for \(R_s = 30\) km.
  • Estimated convective heat loss is \(\approx 115\) mW m\(^{-2}\) (\(\approx 60\) TW) at \(T_s = 273\) K and \(T_i = 2100\) K, and \(\approx 75\) mW m\(^{-2}\) (\(\approx 38\) TW) at \(T_s = 500\) K; Hadean radiogenic heating is estimated at \(\approx 70\)–100 TW, but melting/volcanism could make up the difference.

Methods

  • The study combines scaling laws for convection with grain-damage from Foley and Bercovici [2014] with numerical infinite-Prandtl-number, Boussinesq thermo-chemical convection models that include grain-damage and a composition equation.
  • Numerical experiments use a 4 × 1 finite-volume domain, typically 512 × 128 resolution (1024 × 256 for high \(Ra_0\) or \(D\)), with post-magma-ocean initial conditions: solidus-like temperature profile, static mantle, uniform initial fineness, and a small random perturbation; parameters varied include \(D\), \(Ra_0\), \(\mu_l/\mu_m\), \(h_l/h_m\), \(Q'\), \(Le\), and \(B\).
  • The authors measure overturn, onset, and lag times, derive an empirical lag-time scaling law, and apply it to Hadean Earth conditions using assumed damage, healing, and lithosphere parameters.

Limitations

  • Stated: the scaling-law analysis assumes statistically steady-state convection and does not predict initiation timing; the numerical models are not intended to replicate the post-impact Earth because its high Rayleigh number and viscosity ratio are not numerically tractable, and they exclude thermal evolution, melting effects, and detailed carbon cycling.
  • Stated: chemical buoyancy variations from mid-ocean ridge melting are neglected, the damage partitioning fraction \(f\) and healing activation energy \(E_h\) are uncertain, and Hadean zircon interpretations remain controversial.
  • Evident from the text: the models use a simplified single-phase diffusion-creep rheology, assume a pinned state throughout the mantle, neglect dislocation creep in the mantle interior, and are 2-D, which may overstate time-dependence relative to 3-D.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1002/2014jb011121

Cites

91 of 91 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Plots of plate velocity (a) and heat flow (b) versus mantle potential temperature, at three different surface temperatures: T s = 273 K (red line), T s = 500 K (blue line), T s = 750 K (green line). The dashed black line shows v m / 10 (Equation (11)), giving the boundary between the nominally plate-tectonic regime and a non-plate-tectonic, 'sluggish subduction' regime.
Figure 1 Plots of plate velocity (a) and heat flow (b) versus mantle potential temperature, at three different surface temperatures: T s = 273 K (red line), T s = 500 K (blue line), T s = 750 K (green line). The dashed black line shows v m / 10 (Equation (11)), giving the boundary between the nominally plate-tectonic regime and a non-plate-tectonic, 'sluggish subduction' regime.From the PDF
Figure 2 Images from a thermo-chemical model showing showing A) the initial condition, B) the initial overturn, C) post-overturn stagnant lid convection, D) the initiation of subduction, and E) the final statistical steady-state convection. Each image shows (from top to bottom), the horizontal surface velocity ( u ′ surf ), the fineness field ( A ′ ), the viscosity field ( µ ′ ), the temperature field ( T ′ ), and the composition field ( C ′ ). The model shown here uses Ra 0 = 10 6 , D = 10 -2 , H = 10 3 , E ′ v = E ′ h = 23 . 03, B = 1, and Le = 10 4 .
Figure 2 Images from a thermo-chemical model showing showing A) the initial condition, B) the initial overturn, C) post-overturn stagnant lid convection, D) the initiation of subduction, and E) the final statistical steady-state convection. Each image shows (from top to bottom), the horizontal surface velocity ( u ′ surf ), the fineness field ( A ′ ), the viscosity field ( µ ′ ), the temperature field ( T ′ ), and the composition field ( C ′ ). The model shown here uses Ra 0 = 10 6 , D = 10 -2 , H = 10 3 , E ′ v = E ′ h = 23 . 03, B = 1, and Le = 10 4 .From the PDF
Figure 3 Velocity (A) and Nusselt number (B) time-series data for the numerical result shown in Figure 2. Black dots labeled A-E mark the location of the images shown in panels A-E of Figure 2
Figure 3 Velocity (A) and Nusselt number (B) time-series data for the numerical result shown in Figure 2. Black dots labeled A-E mark the location of the images shown in panels A-E of Figure 2From the PDF
Figure 4 Plot of whole mantle rms velocity ( v ′ rms ), surface velocity ( u ′ surf ), timeaveraged rms and surface velocity ( 〈 v ′ rms 〉 and 〈 u ′ surf 〉 ), and the overturn time t ′ ot and onset time t ′ onset . The numerical result shown here uses Ra 0 = 10 6 , D = 2 × 10 -3 , H = 10 3 , E ′ v = E ′ h = 23 . 03, and Le = 10 4 .
Figure 4 Plot of whole mantle rms velocity ( v ′ rms ), surface velocity ( u ′ surf ), timeaveraged rms and surface velocity ( 〈 v ′ rms 〉 and 〈 u ′ surf 〉 ), and the overturn time t ′ ot and onset time t ′ onset . The numerical result shown here uses Ra 0 = 10 6 , D = 2 × 10 -3 , H = 10 3 , E ′ v = E ′ h = 23 . 03, and Le = 10 4 .From the PDF
Figure 5 Overturn time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). All the models use m = 2, p = 4, and thermo-chemical models use B = 1 and Le = 10 4 , except in (E) and (F) where those parameters are varied. Models in (G) are purely internally heated, while the rest of the models are bottom heated with Q ′ = 0. All models have an initial perturbation amplitude of 5 × 10 -3 , except for those in (H) where the amplitude is varied. The rest of the parameters are (A) H = 10 3 , Ra 0 = 10 6 , E ′ v = E ′ h = 23 . 03 (resulting in µ l /µ m = 10 5 and h l /h m = 10 -5 ); (B) H = 10 3 , D = 2 × 10 -3 , E ′ v = E ′ h = 23 . 03; (C) H = 10 3 , D = 2 × 10 -3 , Ra 0 = 10 6 , E ′ h = 23 . 03; (D) H = 10 3 , D = 2 × 10 -3 , Ra 0 = 10 6 , E ′ v = 23 . 03; (E), (F), (G), and (H) D = 2 × 10 -3 , H = 10 3 , Ra 0 = 10 6 , E ′ v = E ′ h = 23 . 03.
Figure 5 Overturn time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). All the models use m = 2, p = 4, and thermo-chemical models use B = 1 and Le = 10 4 , except in (E) and (F) where those parameters are varied. Models in (G) are purely internally heated, while the rest of the models are bottom heated with Q ′ = 0. All models have an initial perturbation amplitude of 5 × 10 -3 , except for those in (H) where the amplitude is varied. The rest of the parameters are (A) H = 10 3 , Ra 0 = 10 6 , E ′ v = E ′ h = 23 . 03 (resulting in µ l /µ m = 10 5 and h l /h m = 10 -5 ); (B) H = 10 3 , D = 2 × 10 -3 , E ′ v = E ′ h = 23 . 03; (C) H = 10 3 , D = 2 × 10 -3 , Ra 0 = 10 6 , E ′ h = 23 . 03; (D) H = 10 3 , D = 2 × 10 -3 , Ra 0 = 10 6 , E ′ v = 23 . 03; (E), (F), (G), and (H) D = 2 × 10 -3 , H = 10 3 , Ra 0 = 10 6 , E ′ v = E ′ h = 23 . 03.From the PDF
Figure 6 Onset time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). Parameters are listed in the caption to Figure 5.
Figure 6 Onset time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). Parameters are listed in the caption to Figure 5.From the PDF
Figure 7 Lag time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). Parameters are listed in the caption to Figure 5.
Figure 7 Lag time versus D (A), Ra 0 (B), µ l /µ m (C), h l /h m (D), Le (E), B (F), Q ′ (G), and initial perturbation amplitude (H). Parameters are listed in the caption to Figure 5.From the PDF
Figure 8 Plot of the overturn time versus Ra 0 with an Ra -1 0 power law.
Figure 8 Plot of the overturn time versus Ra 0 with an Ra -1 0 power law.From the PDF
Figure 9 Empirical fits to the numerical results for t ′ lag as a function of D (A), Ra tot (B), and µ l /µ m (C). The empirical fits to the corrected lag time (corrected lag time is defined in the text below equation (15)) in (A) and (C) are to simple power law functions, with the results listed in the figure legend, while the fit in (B) is to equation (14). Parameters for the numerical model results displayed here are listed in the caption to Figure 5.
Figure 9 Empirical fits to the numerical results for t ′ lag as a function of D (A), Ra tot (B), and µ l /µ m (C). The empirical fits to the corrected lag time (corrected lag time is defined in the text below equation (15)) in (A) and (C) are to simple power law functions, with the results listed in the figure legend, while the fit in (B) is to equation (14). Parameters for the numerical model results displayed here are listed in the caption to Figure 5.From the PDF
Figure 10 Plot of the lag time versus interior mantle potential temperature, at two different values of f : f = 10 -9 (red line) and f = 10 -7 (blue line). For each value of f , curves are plotted with B = 0 (solid line), and B = 3 (dashed line).
Figure 10 Plot of the lag time versus interior mantle potential temperature, at two different values of f : f = 10 -9 (red line) and f = 10 -7 (blue line). For each value of f , curves are plotted with B = 0 (solid line), and B = 3 (dashed line).From the PDF

Verification

Claims the paper makes

17 Valid · 6 Not valid · 21 Not checked · 117 Own results and interpretations

91 more claims · 4 with their cited sources checked

Create a free account to see all 161 claims and to verify the 3 references still open.

See all claims — free account
  • It may state two separate claims as one. 53
  • Probably not a substantial claim. 21

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Research works printed but not linked: 1

  1. Batchelor G., 1967, An Introduction to Fluid Dynamics

Cited by

See all 98

Metadata

Authors
Foley, B. J., Bercovici, D., & Elkins‐Tanton, L. T.
Affiliations
  • Carnegie Institution for Science (US)
  • Yale University (US)
  • School of Earth and Space Exploration Arizona State University Tempe Arizona USA
Corresponding author
Foley, Bradford J.
Venue
Journal of Geophysical Research: Solid Earth
Published
2014-11
Volume
119
Issue
11
Pages
8538-8561
ISSN
  • 2169-9313 (ISSN-L)
  • 2169-9356
OpenAlex
W2019434814
Cited by
98
References
87
Citations by year
2026: 8 · 2025: 7 · 2024: 8 · 2023: 7 · 2022: 9 · 2021: 8 · 2020: 13 · 2019: 6 · 2018: 12 · 2017: 6 · 2016: 11 · 2015: 3
FWCI
5.93
Type
journal article
Language
en
Open access
Locations
  • Journal of Geophysical Research Solid Earth (publishedVersion)
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
  • AGU Fall Meeting Abstracts
Funders
NSF (EAR-1135382)
Topics
  • Physical Sciences › Earth and Planetary Sciences › Geochemistry and Petrology › Geological and Geochemical Analysis
  • Physical Sciences › Earth and Planetary Sciences › Geophysics › High-pressure geophysics and materials
  • Physical Sciences › Earth and Planetary Sciences › Space and Planetary Science › Astro and Planetary Science
Keywords
plate tectonics initiation, Hadean Earth, magma ocean solidification, mantle convection, grain damage, mantle potential temperature, Hadean zircon, mantle cooling, atmospheric CO2 removal, scaling laws

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.