Well, part one has not been greeted with murmurs of agreement or howls of protest. Perhaps, as a gentle introduction it is lacking in controversial material. Let’s see if part 2 stirs any feelings.
Part 2
Is mantle convection possible? The first thing to recognise is that convection does not necessarily require that the rock at the bottom of the column be less dense than the rock at the top. All that is needed to start convection is for the rock at any point in the column to become sufficiently less dense than the rock directly above it; or to become sufficiently more dense than the rock directly below. In fact, even horizontal differences in density can initiate convective movement. Once started, the rate at which convection will proceed depends to a great extent on the plasticity of the rock. In general, the hotter the rock in any given environment, the more readily it will flow, and the more easily convection should progress.
Consider first rock towards the top of the mantle where pressure is a less important factor than it would be at the base. Local variations in temperature may occur, caused, perhaps by inhomogeniety in radioactive elements. Theory maintains that the hotter rock will be lighter than the colder rock, so, given that the material is sufficiently plastic, the hotter rock will rise, relative to the colder. Obviously the colder rock must sink in order to maintain isostatic balance.
At greater depth in the mantle gravity plays a more important role. The downward pressure exerted by the overlying column of rock at any given point increases with depth, as does the temperature. To decide whether convection will occur, it is necessary to balance increased density due to compression against decreased density resulting from heating.
Let us assume that if the mantle were homogeneous, evenly heated from below, and uniformly cooled from above, a state of balance would eventually be reached in which hot heavy rock would remain at the bottom and light cool rock would remain at the top. Heat transfer would be by conduction, and everything would be stable. This is what the Earth’s mantle would be like if these conditions prevailed. However, such is not the case. The study of seismic waves has shown that the mantle is not homogeneous. In particular, seismic tomography has revealed marked inhomogeniety in the upper and lower regions of the mantle.
One factor that may add to doubts about mantle convection is that many simplified explanations of the process, such as that cited above, tend to suggest that it is a straightforward case of Rayleigh-Benard convection. Such explanations often draw a parallel between mantle convection and a “pot on the stove” situation. It must be remembered that in Rayleigh-Benard convection heating is from below, and cooling from above. While it is undoubtedly true that the mantle is cooled from above, heating is by no means solely from below. A significant portion of the mantle’s heating is from within, and the source of this heat, radioactive material, is not evenly distributed.
Although it is true to say that the mantle is cooled from above, this cooling is most certainly not uniform. Again seismic tomography provides valuable information. The deep ocean trenches and island arcs that plate tectonics theory identifies as areas in which oceanic plates are being subducted, have been studied extensively. Traces of relatively cold material that in terms of proportions and angle of apparent descent fit the theory of subduction have been identified. These reach down to considerable depths in the mantle.
It is widely accepted in scientific circles that although there is now ample evidence that tectonic plates move about the Earth’s surface, the actual mechanism is not fully understood. To some extent this leaves scope for a range of other theories such as the expanding Earth and Kevin Mansfield’s “collision” theory. Possibly professional geologists in general lack the time and inclination to give more than a cursory glance towards non-standard ideas, in addition to any potential damage to their careers that seeming to entertain crackpot theories might bring. We amateurs suffer no such constraints, so perhaps we have a “duty” to question everything, including established wisdom.
It has been argued that if there were convection cells in the mantle discontinuities would be significantly displaced by the ascending and descending limbs. This seems like a logical line of reasoning, but it may be based on an over simplistic view of nature of the mantle.
One possibility is that there is more than one level of convection. For example, there is a major discontinuity at about 660 km depth. It has been suggested that convection may occur below that, with a separate set of convection cells above that level. The chemical composition of basalts from mid-ocean ridge systems indicates that the magma from which they are formed may have originated at no great depth. Although this has been used as an argument in favour of layered convection, such may not actually be the case. Mantle convection would not necessarily require melting to the point of magma formation. There is good evidence that mineralogical changes are brought about by changing pressure at different levels in the mantle. Thus the composition of the magma may reflect the level at which melting took place, rather than the depth at which an ascending limb started its journey.
Convection is much more likely to be possible in the mantle if the dominant mineralogical changes are structural, rather than compositional. Compositional changes would require the addition or removal of chemical elements; whereas structural changes require only readjustment of existing elements. Let us look at the dominant mineral type of the mantle; olivine. Common olivine is stable down to a depth of about 410 km, at which point pressure causes it to convert to wadsleyite. (Temperature plays a part in the transformation, but the prime mover is pressure). Wadsleyite is stable down to about 520 km where it transforms into ringwoodite. All these minerals share the same chemical composition; only the structure is different. Transformation in either direction can take place easily given the necessary changes in pressure; however, these transformations produce discontinuities, and it has been argued that if convection occurred in the mantle these various discontinuities would be displaced either up or down by the rising or falling limbs of each convection cell. One has to ask if any evidence of this has been found.
Seismic tomography has revealed that beneath the Japanese trench the 660 km discontinuity is displaced downward by about thirty kilometres. The descending material is cooler than the surrounding mantle, suggesting that it is a subducting plate that is causing the depression of the discontinuity. If, as geologists seem increasingly to be thinking, subduction is the major driving force of plate movement, it would be reasonable to expect the downward displacement of discontinuities to be more marked, and easier to detect than upward displacement.
One might wonder why displacements of this nature are not detected more commonly. In fact it is much more likely that seismic tomography will trace a descending plate down as far as the 660 km discontinuity. A probable reason for this might be that the spinel to perovskite transformation that takes place at about 660 km depth is endothermic; thus it would be likely to decrease the temperature of the mantle locally, therefore increasing the relative buoyancy of the descending material.
In part 3 I propose to look at the question of vanishing oceanic crust.
