PERSPECTIVES ON THE ANDEAN ARC FROM THE IZU-BONIN-MARIANA ARC SYSTEM IN THE WESTERN PACIFIC

The Izu-Bonin-Mariana (IBM) arc system extends 2800km (40% as long as the Andes) from near Tokyo, Japan to Guam, USA (and beyond), in the Western Pacific. IBM lies at the opposite end of the spectrum of convergent-margins from the Andes, with which it differs greatly in crustal type and thickness, age of subducted lithosphere, topographic expression, dip of subducted slab, earthquake hazard/plate coupling, and behavior of back-arc regions. Most of the variability can be ascribed to the fact that IBM subducts the oldest seafloor on the planet (Jurassic and Cretaceous) and lies on the Philippine Sea plate which itself is subducted westward, and so is under strong extension, whereas the Andes subducts younger lithosphere and is under compression. Unlike the Andes, IBM has no flat slab segments. The behavior of the subducting lithosphere differs further, in that beneath NE Japan and IBM there is a double seismic zone defined by two parallel planes separated by 30-40km, whereas beneath the Andes there is generally no double seismic zone or parallel planes are separated by ~15km. The sedimentary stratigraphy varies simply from north to south on the incoming Pacific plate and this subduction input is reflected in sympathetic changes in trace elements (e.g., Th/La) in arc lavas. The history of IBM since subduction began about 43 Ma is better understood than any other convergent margin. We infer a time-integrated rate of crust formation of ~100 km3/km-Ma, much larger than the 20-40 km3/km-Ma for arc growth inferred by Reymer and Schubert (1984). The lack of an accretionary wedge or contamination by continental inputs makes it a particularly valuable natural laboratory. The IBM mostly erupts basalts but dacites are increasingly recognized. The northern IBM arc is about 22km thick, with a felsic middle crust; this middle crust is exposed in the collision zone at the northern end of IBM. Development of thick felsic middle crust indicates that crustal fractionation is an efficient process that begins early in the evolution of an arc system. Evolution of basalt to dacite on some islands can be described by simple fractional crystallization models, but felsic melts may also be generated by anatexis of amphibolite-facies lower crust.