Abstract. We investigate the evolution of the ridge produced by the convergent motion of two substrates, on which a layer of a non-Newtonian power-law liquid rests. We focus on the self-similar regimes that occur in this process. For short times, within the linear regime, the height and the width increase as t 1/2 , independently of the rheology of the liquid. In the selfsimilar regime for large time, the height and the width of the ridge follow power laws whose exponents depend on the rheological index.
IntroductionMountains belts arise due to the shortening of the crust that occurs when two lithospheric plates collide, or when a plate is subducted beneath an other. However there is also abundant geological evidence showing that extension occurs in the central part of many mountain belts, notwithstanding shortening is taking place. This apparently paradoxical fact suggested one of us [1] the idea that mountain building occurs as a consequence of a dynamic balance between the shortening of the plate and the spreading flow, that occurs because an isostatically compensated range is not in hydrostatic equilibrium, and tends to spread and collapse unless restrained by appropriate stresses. Indeed, it can be easily estimated that the characteristic time for the collapse of a mountain range resulting from root spreading is of the same order of magnitude as the times involved in orogeny, which means that both processes (shortening and lateral spreading) occur simultaneously. Starting from these considerations the scaling laws for the evolution of orogenic belts were derived, based on simple physical hypotheses about the viscous flow caused by the shortening of the Earth's crust [1], which is assumed to have a velocity that depends on the depth.Various physical models have been used to simulate the build-up and the evolution of mountain belts. All are extensions and variations of the thin viscous sheet model of England and McKenzie [2;3]. A discussion and a classification of these models has been made by Medvedev and Podladchikov [4;5], to which the reader is referred for more details. All these numerical calculations attempt to describe specific orogenies as realistically as possible. No effort has been made to gain a deeper physical insight of the process. In particular no scaling laws have been derived from these models. Unlike in [1] most thin viscous sheet models assume that the velocity field does not depend on the vertical coordinate, so that the main gradient is horizontal. It was