Mantle convections theories
There are two widely accepted models for mantle convection:
- ‘Layered Mantle Convection’
- ‘Whole Mantle Convection’.
The layered mantle convection model (Figure.1) proposes that the boundary between the upper and lower mantle creates two separate convection cells (one in each layer). Little to zero subducted material will cross 660km discontinuity.
Evidence For:
-Different seismic waves deflection at 660km suggest change in upper and lower mantle composition and earthquakes ceases at 700km suggested the ceasing of slab sinking (van der Hilst, 1998).
-Compression-orientated nature of earthquakes just above the mantle transition zone, which is indicative of the slabs meeting resistance to further
subduction below the boundary, thus increasing stress on the slab. It is proposed that a small portion of slabs are able to sink into the lower mantle
due to accretion at the 660km boundary and subsequent slab breakthrough events.
Whole mantle convection mode( Figure.2) proposes an unimpeded single-layered convection system, whereby there is large continuous
flow of material between the upper and lower mantle, and this is strongly supported by modern seismic tomography (van der Hilst, 1998; van Kekan and
Ballentine, 1998: p.20). Whole mantle convection model would result in the deep mantle containing large amounts of young lithospheric plates, with the old lithospheric plates pushed up into the shallow mantle, with an eventual complete mixing of mantle material and a subsequent homogeneous composition of the mantle.
Evidence for: There is significant evidence from tomographic studies to suggest that the known slab exchange from upper to lower mantle far exceeds the level proposed by most layered convection models (Albarède and van der Hilst, 2002: p.2577).
Evidence against: Evidence to suggest that the mantle is heterogeneous in nature, with distinct domains within the deep lower mantle enriched in incompatible lithophyllic elements (Albarède and van der Hilst, 2002: p.2582).
- ‘Layered Mantle Convection’
- ‘Whole Mantle Convection’.
The layered mantle convection model (Figure.1) proposes that the boundary between the upper and lower mantle creates two separate convection cells (one in each layer). Little to zero subducted material will cross 660km discontinuity.
Evidence For:
-Different seismic waves deflection at 660km suggest change in upper and lower mantle composition and earthquakes ceases at 700km suggested the ceasing of slab sinking (van der Hilst, 1998).
-Compression-orientated nature of earthquakes just above the mantle transition zone, which is indicative of the slabs meeting resistance to further
subduction below the boundary, thus increasing stress on the slab. It is proposed that a small portion of slabs are able to sink into the lower mantle
due to accretion at the 660km boundary and subsequent slab breakthrough events.
Whole mantle convection mode( Figure.2) proposes an unimpeded single-layered convection system, whereby there is large continuous
flow of material between the upper and lower mantle, and this is strongly supported by modern seismic tomography (van der Hilst, 1998; van Kekan and
Ballentine, 1998: p.20). Whole mantle convection model would result in the deep mantle containing large amounts of young lithospheric plates, with the old lithospheric plates pushed up into the shallow mantle, with an eventual complete mixing of mantle material and a subsequent homogeneous composition of the mantle.
Evidence for: There is significant evidence from tomographic studies to suggest that the known slab exchange from upper to lower mantle far exceeds the level proposed by most layered convection models (Albarède and van der Hilst, 2002: p.2577).
Evidence against: Evidence to suggest that the mantle is heterogeneous in nature, with distinct domains within the deep lower mantle enriched in incompatible lithophyllic elements (Albarède and van der Hilst, 2002: p.2582).
ZONED MANTLE CONVECTION AND SELECTIVE
SUBDUCTION MODEL
‘Zoned Mantle Convection’, which maintains a concentration gradient within the mantle, where the composition of the mantle varies with depth. This model incorporates a process known as ‘selective subduction’. Components include ( see FIG.3):
- Deep mantle containing recycled plume heads from subducted oceanic plateaus, small amounts of primitive mantle
- Shallow mantle consists of subducted barren –no plume material- lighter oceanic plates.
- Selective subduction. That is how deep the slab sinks is dependent on thermal structure and composition of the subducted slabs. Thermal structure is a measure which includes direction and speed of convergence, age of lithosphere and speed of subduction. (Albarède and van der Hilst, 2002: p.2569).
- Compositional variation changes the buoyancies, causing them to sink to different depths. Mantle plumes are negatively buoyant and ocean crust in more buoyant.
The oceanic plateaus form from eruptions of plume head basalts and these areas are underlain by thicker crusts of >25km as opposed to the average 6km oceanic crust of regular oceanic plates (Albarède and van der Hilst, 2002: pp.2582-2585). These basalts are altered to heavy eclogite through metamorphism during subduction, and the significant increase in density creates negative buoyancy which exerts a strong pulling force on the attached lithospheric plate enabling preferential sinking to greater depths than that of barren oceanic plates. This has been supported by high resolution global inversions and seismic tomography, which demonstrate that some slabs of subducted lithosphere can penetrate the deeper lower mantle, whilst others are dislocated and dissipated in the shallow upper mantle (Albarède and van der Hilst, 2002: pp.2575-6).
These parameters of thermal structure, composition, plate-tectonic history an evolution over time, vary significantly with different subduction systems, thus
enabling vast diversity in slab sinking depths for individual barren and plateau-laden plates.
This selective subduction model results in poor mantle mixing and thus acompositional gradient. So why isn't it the same as the layered model? The chemical differences throughout the typical mantle column is different at different locations through varying subduction system. It also produces compositionally distinguished domains or ‘reservoirs’ which account for the known geochemical signatures in the deep mantle, particularly that of incompatible lithophyillic elements. (Albarède and van der Hilst, 2002: p.2582). For example, the deeper mantle remains more enriched due to the introduction of U, Th and K (incompatible elements) through the greater sinking depth of oceanic plateau lithosphere with higher basaltic component(Albarède and van der Hilst, 2002: p.2582).
This is supported by the large residence times (6-14Gyr) of these incompatible lithophyllic elements, where the probability of extraction into
MORBs is low, indicating that the MOR source is rarely sampled and is depleted in regards to the bulk mantle composition (Albarède and van der Hilst, 2002: p.2583).
SUBDUCTION MODEL
‘Zoned Mantle Convection’, which maintains a concentration gradient within the mantle, where the composition of the mantle varies with depth. This model incorporates a process known as ‘selective subduction’. Components include ( see FIG.3):
- Deep mantle containing recycled plume heads from subducted oceanic plateaus, small amounts of primitive mantle
- Shallow mantle consists of subducted barren –no plume material- lighter oceanic plates.
- Selective subduction. That is how deep the slab sinks is dependent on thermal structure and composition of the subducted slabs. Thermal structure is a measure which includes direction and speed of convergence, age of lithosphere and speed of subduction. (Albarède and van der Hilst, 2002: p.2569).
- Compositional variation changes the buoyancies, causing them to sink to different depths. Mantle plumes are negatively buoyant and ocean crust in more buoyant.
The oceanic plateaus form from eruptions of plume head basalts and these areas are underlain by thicker crusts of >25km as opposed to the average 6km oceanic crust of regular oceanic plates (Albarède and van der Hilst, 2002: pp.2582-2585). These basalts are altered to heavy eclogite through metamorphism during subduction, and the significant increase in density creates negative buoyancy which exerts a strong pulling force on the attached lithospheric plate enabling preferential sinking to greater depths than that of barren oceanic plates. This has been supported by high resolution global inversions and seismic tomography, which demonstrate that some slabs of subducted lithosphere can penetrate the deeper lower mantle, whilst others are dislocated and dissipated in the shallow upper mantle (Albarède and van der Hilst, 2002: pp.2575-6).
These parameters of thermal structure, composition, plate-tectonic history an evolution over time, vary significantly with different subduction systems, thus
enabling vast diversity in slab sinking depths for individual barren and plateau-laden plates.
This selective subduction model results in poor mantle mixing and thus acompositional gradient. So why isn't it the same as the layered model? The chemical differences throughout the typical mantle column is different at different locations through varying subduction system. It also produces compositionally distinguished domains or ‘reservoirs’ which account for the known geochemical signatures in the deep mantle, particularly that of incompatible lithophyillic elements. (Albarède and van der Hilst, 2002: p.2582). For example, the deeper mantle remains more enriched due to the introduction of U, Th and K (incompatible elements) through the greater sinking depth of oceanic plateau lithosphere with higher basaltic component(Albarède and van der Hilst, 2002: p.2582).
This is supported by the large residence times (6-14Gyr) of these incompatible lithophyllic elements, where the probability of extraction into
MORBs is low, indicating that the MOR source is rarely sampled and is depleted in regards to the bulk mantle composition (Albarède and van der Hilst, 2002: p.2583).