So far in Part 3 I have developed an approach to mantle convection based on thermal boundary layers and used this to look at the behaviour of the two boundary layers in the mantle for which there is good evidence. Here I consider how these parts assemble into a coherent picture and look at some immediate implications about how the system does and does not work.

It is also an appropriate place to discuss alternative views. Some of these are in direct opposition to the picture developed here, such as that the mantle is divided into two layers that convect separately. Others are different ways of looking at the system that have been used in the long debate. Some of these carve the total system up in a different way. Some are complementary, and useful for bringing out particular aspects, while others are unprofitable or potentially misleading.

The mantle as a dynamical system

The picture that has emerged here is of a mantle system in which two thermal boundary layers have been identified on the basis of observational evidence, one comprising the plates and the other giving rise to plumes. The boundary layers appear to transport heat at substantially different rates, and they manifest quite different geometries and flow patterns. The differences in geometrical patterns are inferred to be due to different mechanical properties of the boundary layers, and such differences can be plausibly justified on the basis of our understanding of material properties of rocks. The boundary layers seem to operate with a lot of independence, since plume locations correlate only weakly with spreading centres, the sites of (passive) upwelling in the plate-scale flow.