[S-IT04_28AM1] Fluid flow, deformation and physical properties of the subduction boundary and forearc mantle

[SIT04-02] Effects of a local deepening of slab-mantle decoupling depth on slab surface temperature

*Manabu MORISHIGE¹, Peter E. VAN KEKEN² (1.JAMSTEC, 2.University of Michigan)

Keywords:subduction zone, slab-mantle decoupling depth, slab surface temperature

In the subduction zone, we generally observe low seismic attenuation in the forearc mantle. In addition, surface heat flow shows low value in the forearc and a sudden transition to high value in the arc. These observations suggest that the forearc mantle is cold and is not involved in the corner flow in the mantle wedge. We can understand it in terms of slab-mantle decoupling depth (D_dec). Above D_dec, the mantle does not move with the slab just beneath it. Therefore, it becomes cold quickly due to the cooling from the overriding plate and the slab. Below D_dec, on the other hand, the mantle moves with the slab. It keeps this part of mantle warm by advection of hot material due to the corner flow. Thus, D_dec is a key parameter which strongly affects thermal structure in the subduction zone. Comparison of the observed surface heat flow and the one predicted with 2D numerical model suggests that D_dec does not vary much for each subduction zone and is 70-80km, but in each subduction zone D_dec may show some degree of along-arc variation. One such example is the junction between Japan and Kurile arcs, where the down-dip limit of thrust type earthquake is locally deepened by around 15km. In this presentation, we investigate the effects of a local deepening of D_dec on slab surface temperature.Toward the goal, we use time-dependent 3D finite element models to compute mantle flow and temperature. Only mantle wedge is treated as a dynamic entity. We use a simple slab geometry and assume a local deepening of D_dec to see its effects. We find that the increase in slab surface temperature at D_dec is larger where we assume a deepening of D_dec, which produces a warmer region there. It is caused by 3D flow in the mantle wedge due to along-arc variation of D_dec. We also calculate surface heat flow from obtained thermal structure, but it does not show significant along-arc variation. These results do not change even when we use a realistic slab geometry which is similar to that of the junction between Japan and Kurile arcs. While the surface heat flow anomaly and deepening of the seismic belt in S. Hokkaido cannot be easily explained by these models, the temperature excursions at the slab surface are significant. These models predict potentially strong variations in the conditions that the fluids leave the slab, which may be visible by various new geothermometers, such as those based on the H₂O/Ce ratio.