To the Editor —

Whether subduction initiation is a forced or spontaneous process is a fundamental geological problem. Using drill core data from the Amami Sankaku basin, Arculus et al.1 suggest that subduction in the Izu–Bonin–Mariana arc began spontaneously. Here we argue that the evidence presented favours neither spontaneous nor forced initiation.

Some numerical models predict that the early stages of forced subduction initiation cause uplift, then subsidence, followed by extension and seafloor spreading in the upper plate2,3. The seafloor spreading generates juvenile oceanic crust in the upper plate as the subducting plate sinks and subduction becomes self-sustaining2,3,4. A similar process of juvenile crust generation is proposed for spontaneous subduction initiation4. Early uplift, however, is not expected in spontaneous initiation because there is no contraction that precedes subduction. Thus, determining whether uplift occurred prior to subduction can test for spontaneous versus forced initiation. To test for uplift, one must sample crust that was present when uplift occurred.

Arculus et al.1 find that the age and geochemistry of the basement rocks at site U1438 resemble juvenile fore-arc basalts of the Izu–Bonin–Mariana subduction system formed from seafloor spreading at the early stages of subduction initiation4. Their new findings suggest that the extent of across-strike juvenile crust formation at the onset of subduction in the nascent Izu–Bonin–Mariana system is larger than previously thought. However, the presence, nature and/or spatial extent of the juvenile oceanic crust are currently not known to be different between a forced and spontaneous subduction initiation event and thus cannot be used to argue for one mode over the other. In addition, since the younger juvenile crust will not record the uplift expected to accompany forced initiation, Arculus et al.1 were not able to use sedimentological evidence to test for early uplift in pre-subduction initiation basement rocks. We conclude that the data presented by Arculus et al.1 do not favour spontaneous or forced subduction initiation at the Izu–Bonin–Mariana arc.