ULTRA-FAST SOURCE-TO-SURFACE MOVEMENT OF MELT AT ISLAND ARCS FROM RA-TH SYSTEMATICS

The velocity of melt ascent from its source through Earth’s mantle and crust to the surface is extremely hard to determine, even though such measurements would place important constraints on the mechanisms of melt transport and the physical behavior of the mantle during partial melting. Island arc lavas have Ra–excesses that extend to higher values than those observed in mid-ocean ridge or ocean island basalts. The Ra–excesses are largest in the most primitive lavas and positively correlated with the fluid-sensitive index Ba/Th and therefore appear to have been introduced into the base of the mantle melting column by fluids released from the subducting plate. Preservation of the Ra signal requires transport to the surface arguably in only a few hundreds of years and directly constrains the average melt velocity to the order of 1000 meters per year. Thus, melt segregation and channel formation can occur rapidly in the mantle and island arc melting models need to incorporate high melt ascent velocities. Such conditions probably reflect the vigor of fluid-induced melting compared with mid-ocean ridges and ocean islands where melting rates are dictated by upwelling rates on the order of a few cm per year. However, it may be that these velocities are generally applicable in which case the widely held notion of basalts migrating slowly through the mantle in interconnected but vanishingly small pore spaces will have to be abandoned. The island arc Ra–excesses also decrease with increasing magmatic differentiation. If these decreases reflect the time taken for crystal-liquid differentiation, then this must take < 8,000 years and might be almost instantaneous if such decreases are due to mixing processes. Thus, if melts do stall in the crust it is generally for short periods of time (< 8000 years) and it appears that the majority of island arc basalts and andesites spend less than a few thousand years traversing the crust. This leaves very little time for crustal interaction (e.g. MASH) which must either occur very quickly or else may be restricted to continental areas with very thick lithosphere.