The convection cell mechanism
This concept initially attracted a great deal of support (Griggs, 1939; Holmes, 1965). It was argued that the heat generated in the interior of the Earth (largely as the result of radioactive decomposition of minerals) must escape through the mantle by convection and by conduction at the lithospheric boundary, and that this accounted for the high temperature gradient of about 25°C near the surface of the Earth. The density anomalies in the mantle, which are necessary to drive currents at a rate comparable with that of continents, may be easily met, in that the required temperature difference between the up-going and down-going currents, at a specific depth, need only be small (see Chapter 1).
In this early period, it was thought that plate motion was dominated by convectional movements in the asthenosphere beneath the plates, in much the same manner in which surface 'skin' in a cooking pot is driven into ridges by the mobile convectional movements of the more fluid content of the pot. This concept was supported by a persuasive paper (and film) of a model in which David Griggs (1939) 'built a mountain chain' by causing opposing currents to develop in a viscous fluid, dragging a lighter, plastic crust to form a geosyncline. In this phase, the crust became compressed and thickened. The crust was then uplifted to form a mountain chain, once the motivating currents were stopped (Figure 3.1).
However, it was soon realised that small plates would require to be driven by a convection cell in the asthenosphere of a comparable plan-area. For a very much larger body, such as the Pacific plate, a correspondingly large plan-area convection cell would be required, for it is difficult to envisage how a relatively large number of small convection cells could combine to provide a resultant force capable of driving such a large mass as the Pacific plate at a relatively high velocity.
Furthermore, hotspots, some of which are presumably rooted in the mantle beneath the plates, are almost stationary, as regards lateral movement, while the plates move at significantly higher velocities. Hence, it was rendered even more difficult to envisage how plates can be driven by mantle convection, while hotspots are relatively unaffected. This difficulty was further enhanced (Turcotte and Schubert, 1982; Bott, 1993), once it was realised that oceanic plates were underlain by the low-viscosity zone (LVZ) that might be 50200 km thick and in which the coefficient of viscosity is at least an order of magnitude less than that of the mantle in general. Consequently, the possible coupling between a mantle convection cell and the overlying plate was seen to be exceedingly improbable. In motoring terms, there was a 'slipping clutch' between the engine and the drive shaft.
Anderson et al. (1992) expressed what is, we believe, the generally held view that plates slide on the LVZ because it presents a relatively small, unit area, resistance to plate movement. The plume hypothesis, mentioned above and also in Chapter 1 (Morgan, 1971) which refers to upwelling columns in the mantle, has been used to explain the initiation of spreading-ridges, hotspots and a wide range of igneous provinces (all of which will receive comment in later chapters). The plume mechanism was originally considered to be
Figure 3.1 A mechanical scale-model designed by Griggs to simulate the action of convection currents in the mantle by using rotating drums. (a) This figure shows how down-folding of the crust has developed in response to the inward drag of the currents in the mantle. (b) Shows crustal development when only one drum is rotated. The 'continent' to the right is thickened and shortened as the result of lateral transfer of crust from the left. It will be noted that the mountain chain is transposed towards the non-rotating drum. It is held in that position because the right-hand wall of the tank prevents lateral motion of the crust. If the wall were distant from the drums, so that the crust could freely move to the right, it is probable that the mountain chain would not have developed.
responsible for the generation and continued development of a spreading-ridge. However, there are certain obvious problems in applying this concept, namely that all spreading-ridges migrate and, moreover, they frequently jump, i.e. they show abrupt displacement along transform faults (Figure 3.2). The reasons for such abrupt displacements is not well understood. However, less rapid differential movement of a spreading-ridge can be induced by the behaviour of strong crustal elements in the oceanic lithosphere. For example, the Nazca/Sala-y-Gomez Ridge abuts the S American plate (Figure 3.3). This ridge is too thick, wide and strong to permit its easy subduction beneath the S American plate. Consequently, it acts as a battering-ram which locally closes the trench, and also prevents current activity of previously active volcanic cones in the Andes, in the latitude of the ridge. However, S America is advancing westward and is, therefore, also pushing the 'battering-ram' westward, thereby, causing differential displacement of that ridge. It is not easy to see how a surface feature, such as the Nazca/Sala-y-Gomez Ridge can disrupt a deep-seated wall-like plume.
Was this article helpful?