1999 Fall AGU Abstract (Marton)

Marton, F. C., C. R. Bina, S. Stein, and D. C. Rubie, Mineralogy and the regulation of subduction rates, Eos, Transactions of the American Geophysical Union, 80, Fall Supplement, T11G-08, 1999.

T11G-08

The negative thermal buoyancy of cold subducting lithosphere is believed to be a major control on subduction rates. Recently, attention has been given to the influence of mineralogy on the buoyancy of subducting slabs and its possible influence on subduction rates [Rubie, 1993; Kirby et al., 1996]. Changes in the density of the slab relative to the surrounding mantle via phase transformations of (Mg,Fe)₂SiO₄ should have a significant effect on subduction rates. In addition to deflecting the depths of equilibrium phase transitions, cold temperatures in the interiors of slabs may kinetically inhibit these transitions [Rubie and Ross, 1994]. The resulting metastable wedges of olivine might then act as ``parachutes,'' slowing the slabs' descents [Kirby et al., 1996]. Previous work [Marton et al., 1999; Schmeling et al., 1999] has shown that metastable wedges can slow slabs by up to 30% and that negative feedback between the wedge sizes and the subduction rates may help to control plate speeds.

We use thermo-kinetic modeling to predict mineralogy and calculate the driving buoyant forces acting on subducting slabs. Calculated terminal velocities (CTVs) are found via balances of the driving forces and the opposing viscous drag forces. These CTVs are then fed back into the model, adjusting the durations of the iterations. The resultant CTVs and cross-sectional areas of the metastable wedges achieve a steady state within a few My after the slabs' tips exit the transition zone. Over short time scales, however, the velocities and areas appear to be exhibit the postulated negative feedback. This is examined further by increasing and decreasing the CTVs for short durations and then allowing the velocities and areas to freely adjust themselves. Areas and CTVs return to their steady state values within a few My, showing that, in our model, the negative feedback does occur.