Styles of mantle convection and their influence on planetary evolution Les styles de convection du manteau et leur influence sur l'évolution des planètes

Mantle convection is the method of heat elimination for silicate mantles in terrestrial bodies, provided they are not too small or too hot. Bodies that are small (∼Moon or smaller, possibly even Mercury) may rely largely on conduction or melt migration, and bodies that are very hot (Io, very early Earth) may use massive melt migration (magma oceans) and heat pipes. In the standard, simple picture, we can use scaling laws to determine the secular cooling of a planet, likelihood and form of volcanism, and the possibility of a core dynamo. Contrary to popular belief, small planets do not cool faster than larger planets (provided they convect) but they do tend to have a slightly lower internal temperature at all times and thus may cease to be volcanically active at an earlier epoch. On the other hand, a larger volume fraction of a small planet may be involved in melt generation. However, our understanding of heat transfer by mantle convection is limited by three very important, largely unsolved problems: The complexities of rheology, the effects of compositional gradients, and the effects of phase transitions, especially melting. The most striking manifestation of the role of rheology lies in the difference between a mobile lid mode (plate tectonics for Earth) and a stagnant lid mode (other large terrestrial bodies). This difference may arise because of the role of water, but perhaps also because of melting, or size (gravity), or the vagaries of history. It has profound effects for the differences in history of Earth, Venus and Mars, including their surface geology, volatile reservoirs and magnetic fields. Since thermal convection is driven by small density differences, it can also be greatly altered or limited by compositional or phase effects. Melt migration introduces additional complications to the heat transport as well as being a source for the irreversible differentiation that might promote layering. Our limited understanding and ability to model these processes continues to limit the development of a predictive framework for the differences among the terrestrial planets.  2003 Academie des sciences/Editions scientifiques et medicales Elsevier SAS. All rights reserved.