Global systematics of arc volcano position.

Global systematics in the location of volcanic arcs above subduction zones are widely considered to be a clue to the melting processes that occur at depth, and the locations of the arcs have often been explained in terms of the release of hydrous fluids near the top of the subducting slab (see, for example, refs 3-6). Grove et al. conclude that arc volcano location is controlled by melting in the mantle at temperatures above the water-saturated upper-mantle solidus and below the upper limit of stability of the mineral chlorite and in particular, that the arc fronts lie directly above the shallowest point of such melt regions in the mantle. Here we show that this conclusion is incorrect because the calculated arc locations of Grove et al. are in error owing to the inadequate spatial resolution of their numerical models, and because the agreement that they find between predicted and observed systematics arises from a spurious correlation between calculated arc location and slab dip. A more informative conclusion to draw from their experiments is that the limits of chlorite stability (figure 1b of ref. 7) cannot explain the global systematics in the depth to the slab beneath the sharply localized arc fronts.