Figure 1: Arc position versus subduction parameters for data and models.
a, Calculated depth Dmelt of the shallowest portion of the P, Tmelt-based melting field (compare figures 1 and 2 of ref. 7). Calculations were carried out on a 1-km finite-volume mesh9, for dip of 30° to 70° in steps of 10°, and for convergence rate V from 30 to 100 mm yr−1, in steps of 10 mm yr−1; these ranges include the parameters of the calculations of ref. 7. The points correspond to the minimum depths of melting calculated according to the hypothesis and methods of Grove et al.7 for a 2.3 × 2.3-km resampled grid (open triangles) and for a 0.25 × 0.25-km resampled grid (filled triangles). b, Diamonds show the depth of the slab Dslab, determined seismologically2 (error bars as described by ref. 2); filled triangles show the calculated Dslab below the locus of shallowest melting, for the 0.25 × 0.25-km resampled grid from panel a. The red triangles correspond to the corrected values of Dslab for the combinations of dip and convergence rate used by ref. 7 (T. Grove et al., personal communication). The grey line corresponds to a constant Dslab = 62 km. c, This panel corresponds to the lower 200 km of figure 3 in ref. 7. Points as for panel b, plotted for the horizontal distance between the trench and the arc, which is equal to Dslab/tan(Dip), the quantity on the y axis of figure 3 of ref. 7. The grey line corresponds to Dslab = 62 km and demonstrates the spurious correlation referred to in the main text.
