Plate tectonics and mantle controls on plume dynamics
Summary (4 min read)
- Collectively, the above-mentioned studies highlight the need to investigate the coupled behaviour of mantle plumes, plate tectonics, large-scale mantle flow and basal thermochemical structures.
- Here the authors use time-dependent 3Dspherical numerical models of whole-mantle convection at Earth-like con-vective vigour and self-generating plate tectonics to jointly investigate the mechanisms linking plate tectonics, mantle convection and plume dynamics.
- In these models, drifting plumes arise self-consistently and dynamically interact with surface tectonics, large-scale and small-scale convection.
- Focusing on model plume conduits, the authors show that their number, lifetime, shape, temperature excess, rising speed, buoyancy and heat fluxes are comparable to observations.
- This serves as a basis to investigate the role of plate tectonics, plate layout, convective vigour of the lowermost mantle and basal thermochemical structures on plume spatio-temporal dynamics.
2.1. Numerical models of mantle convection with plate-like tectonics
- The authors track the evolution of compositional fields using the tracer-ratio method (Tackley and King, 2003) .
- Model continents are low-density lithospheric rafts, thicker and stiffer than the oceanic lithosphere to prevent their entrainment by the convective flow (Table 1 ).
- Their margins are recycled through time (Supplementary Movies 1-3), consistent with estimates of continental recycling at subduction zones (Coltice et al., 2019).
- The use of pseudo-plasticity favours Earth-like surface velocities and tectonics (e.g. Coltice et al., 2017) .
- As in Arnould et al. (2018) , the authors verify that Models 2-5 favour plate-like behaviour with Earth-like plateness (proxy for the degree of surface deformation localisation, Tackley, 2000) , surface mobility, plate velocities, heat flow and topography (Table 2 , Fig. 1 , Fig. S3 (a), Movie S1 and S2).
2.2. Automated detection of mantle plumes
- Therefore, in their models, "net rotation" is the net rotation of the mantle with respect to a fixed surface.
3. Phenomenology of model mantle plumes
- The authors first describe and quantify the spatio-temporal characteristics of model mantle plumes.
- The number of detected plumes is in broad agreement with other numerical models of whole-mantle convection although model parameters (including physical assumptions and surface boundary conditions) differ (e.g. Davies and Davies, 2009; Hassan et al., 2015; Li and Zhong, 2019) .
- On Earth, the number of hotspots of lower mantle origin remains debated.
- These results are compatible with the longest lived oceanic hotspot tracks for St-Helena (about 120 Myr), Tristan (about 120 Myr) and New England (about 130 Ma) plumes (Duncan, 1984; Williams et al., 2015) .
- Shorterlived hotspot tracks (e.g. Hawaii and Louisville active at least from 87 Ma) terminate at subduction zones, but may have been active longer, the potential oldest parts of their tracks having been subducted (e.g. Portnyagin et al., 2008) .
- Finally, geochemical analyses suggest that South Pacific Cook-Austral plumes have been active for at least 120 Ma (e.g. Konter et al., 2008) .
3.3. Shape and radius
- Plumes originate from topographic crests (Fig. 1 ), either along the edges or on the top of the basal thermochemical heterogeneities, as in previous models that consider dense and hot basal material (Garnero and McNamara, 2008; Hassan et al., 2015; Li and Zhong, 2017, 2019) .
3.5. Buoyant rising speed
- In Model 5 (thermochemical plumes), buoyant rising upper mantle speeds are only approximately 17 cm yr −1 because entrained denser basal material decreases the positive thermal buoyancy of plume material.
3.6. Buoyancy flux
- With ρ m the reference mantle density, α the reference thermal expansivity, ∆T the mantle plume temperature excess, A p the cross-sectional area of mantle plume conduits and v p the buoyant rising speed.
- This range of model plume buoyancy fluxes is comparable to estimates for present-day hotspots (Crosby and McKenzie, 2009) .
3.7. Heat flow
- Davies (1988) and Sleep (1990) estimated the contribution of plumes to the total surface heat flow as 2.5 TW from hotspot swells.
- The ratio of upper-mantle plume heat flow to core heat flow (about 50-60% in Models 1-5) is consistent with numerical models of incompressible and isoviscous (Labrosse, 2002) or temperature-dependent viscosity convection with different internal heating rates (Mittelstaedt and Tackley, 2006) . (2006) showed that plumes contribute to heat up sinking slabs, therefore losing some heat on their way up (Fig. S6 ).
- Using extended-Boussinesq (Leng and Zhong, 2008; Zhong, 2006) and compressible models (Bunge, 2005) , it was also suggested that the proportion of core heat flux advected by plumes decreases towards the surface following a steep plume adiabatic gradient.
3.8. Pulses of activity
- This temperature decrease is consistent with geological estimates of the temperature evolution of the Galapagos and Iceland plumes (Herzberg and Gazel, 2009) .
- Changes in plume buoyancy flux tend to occur after plume merging events or after interaction with a ridge (Fig. 2d ).
3.9. Absolute motions of plumes
- This reflects that some plumes are tilted by vigorous upper mantle convection.
3.11. Topographic swell
- The swell extent and amplitude depend on 1/ plume buoyancy flux that decreases over time (Fig. 2 and Movies S1 and S2), 2/ the nature of the impacted lithosphere (thickness, type, plate boundary proximity) and 3/ the relative motion between the plume and the lithosphere, which can shear the conduit and result in an asymmetric shape of the plume trail (Fig. S2 , Movies S1 and S2; see also Arnould et al., 2019) .
3.12. Plume-ridge interactions
- Finally, some plumes can cross ridge axes and contribute to the propagation of a new spreading axis (Fig. 6c ).
- This case shares similarities with the Réunion hotspot crossing the Central Indian Ridge about 30 Myr ago and its putative link to the formation of the Rodriguez ridge (Morgan, 1978) .
4. Sources of plume drift
- The authors used the above analysis of the characteristics of model plumes to classify the lateral plume motions into categories.
4.1. "Fixed" plumes
- These plumes are rooted at saddle points of basal flow and therefore remain stable throughout their lifetime.
- This stability is favoured by the absence of plate tectonics at the surface, since the number of fixed plumes decreases to between 25% (Model 5) and 50% (Model 3) when platelike behaviour occurs (Fig. 5a ).
- In Li and Zhong (2019) , the proportion of fixed plumes is 10-20%, possibly due to the deep mantle rotation rotation induced by imposed plate velocities (Rudolph and Zhong, 2014) .
4.2. Basal mantle flow entrainment
- IDs 18 and 18) coincides in direction and magnitude with lowermost mantle flow in their vicinity (Fig. 5a and Table 2 ).
4.3. Slab-induced drift
- Have not yet been documented on Earth, which could be explained by the rare occurrence and short duration of such model events.
5.1. Limitations of models and analysis
- This threshold was defined by comparing the number of plumes detected by the algorithm to the number of mantle plumes detected visually.
5.2. What controls plume dynamics?
- Here, the authors focus on the causes of lateral motions of already developed plumes and do not investigate the controls on the position of their source, which is likely to also affect their dynamics (e.g.
5.2.1. Indirect control by plate tectonics
- The authors show that planetary surface dynamics exerts a first-order control on plume drift: if the surface is in stagnant lid, stable or slowly moving plumes predominate.
- The impossibility for mantle plumes to drift rapidly in the absence of plate tectonics and lithospheric thickness heterogeneities was noted by Zhong (2009) who studied the origin of Martian volcanism.
- Plate-like behaviour promotes faster plume motion (Fig. 5 ) due to the interaction of plume conduits with slabs in the upper or the lower mantle.
5.2.2. Limited stabilisation of mantle plumes by mid-ocean ridges
- It has been proposed that mantle plumes can be pinned to stable ridges (Tarduno et al., 2009) .
- In their models, three plume behaviours arise depending on the type of spreading centre they interact with.
- The possible slow motion of the Azores plume could fall in that category (Arnould et al., 2019) .
- Finally, mantle plumes interacting with fast-migrating ridges, usually neighbouring small plates affected by fast reorganisations, tend to move laterally faster and more eratically (Fig. 6d ).
5.2.3. The role of the mantle environment on plume dynamics
- Model mantle plumes rooted at the edges of basal thermochemical heterogeneities tend to be more mobile (Fig. 5a ) and are more likely to be deflected (Fig. 4 ) than purely thermal plumes (Davaille et al., 2002; Li and Zhong, 2019) because the entrainment of dense material by plume conduits slows their rise by a factor of two in their experiments.
5.2.4. Plume characteristics are not diagnostic of plume motions
- Moreover, plume tilt angles do not correlate with plume drifting rates (Fig. 8b ).
- Instead, fast-moving plumes are characterized by small tilt angles.
- This suggests that intrinsic plume characteristics cannot be used as a diagnostic for plume drift.
5.3. Implications for absolute plate reconstructions
- Moreover, their results show that plumes with different drifting rates can coexist within the same global convective system, depending on their location and potential interactions with slabs, plate tectonics and regional convective flow (Fig. S9 ).
- Indeed, model plume velocities can exceed 2 cm yr −1 , consistent with geochronological and paleomagnetic observations for Hawaii (Tarduno et al., 2003; Konrad et al., 2018) and Kerguelen (Antretter et al., 2002) .
- The authors study therefore reconciles contradictory observations of plume drift and suggests that defining a global reference frame based on hotspot tracks to reconstruct past absolute plate motions requires the careful selection of slow-moving plumes based on paleomagnetic and geochronological data.
- The intrinsic properties of plumes observed from the surface (age, excess temperature, buoyancy flux, rising speed, tilt angle) are not diagnostic of plume motion since they do not correlate with lateral plume velocity.
Did you find this useful? Give us your feedback
Cites background from "Plate tectonics and mantle controls..."
...For example, all the following studies model plate tectonics, but the geometry of the model can be complex (e.g., a 3D spherical domain like Arnould et al. (2020)) or simple (e.g., a 1D or 2D domain like Bercovici and Ricard (2014))....
Related Papers (5)
L. Cserepes, Dave A. Yuen
Robert A. Duncan, Mark A. Richards
Dinzia G. Farnetani, Mark A. Richards
Geoffrey F. Davies
Norman H. Sleep, Mark A. Richards +1 more