Igneous differentiation in arcs

This chapter assesses crystallization-differentiation of convergent-margin magmas both from the standpoints of natural rock suites and from laboratory phase-equilibrium experiments. Crystallization-differentiation of dry subalkaline and alkaline basalts are summarized at the beginning for reference and context. Experiments crystallizing arc basalts through andesites at middle to upper crustal pressures (100–400 MPa), saturated either with H2O or mixed H2O-CO2, produce liquid lines of descent similar to arc igneous suites, with higher fO2s giving results closer to calc-alkaline magma series. Early enrichment of melts in SiO2 with little or no enrichment in FeO* (total Fe as FeO) comes about by the combined effects of H2O suppressing crystallization of plagioclase relative to olivine, shifting plagioclase to more calcic – therefore lower SiO2 – compositions, and suppressing silicate minerals to greater extents than Fe-rich spinel. Also contributing to SiO2 enrichment is crystallization near the liquidus of higher proportions of olivine relative to calcic clinopyroxene, compared with MORB, but it is unclear if this is brought about by H2O or elevated pressure. Exhumed arc terranes, however, generally fail to expose evidence for primitive liquids having reached and differentiated at middle to upper crustal depths, at least in any abundance. Instead, most differentiation in arcs appears to take place in the lower crust or near the crust-mantle transition, and although hydrous experiments at such pressures produce liquids broadly similar to arc suites in their SiO2 enrichment, the experimental liquids differ in becoming peraluminous at early stages of differentiation while melts have basaltic, basaltic andesitic, or andesitic SiO2 concentrations. Crystallization of abundant calcic clinopyroxene close to the liquidus causes the high-pressure experimental liquids to become peraluminous with only small increases in SiO2 concentration. In contrast, natural arc volcanic and plutonic suites become gradually less metaluminous with increasing SiO2, crossing to peraluminous at dacitic to rhyolitic SiO2 concentrations. Hybridism, encompassing magma-mixing, assimilation of old crust, and incorporation of near-solidus melts from only shortly older deep, antecedent arc intrusions may accompany natural crystallization-differentiation, preventing melts from attaining peraluminous compositions at low-SiO2 concentrations. If so, then nearly all arc suites are to substantial extents composites, rather than preserving simple, progressive liquid lines of descent. Some comments and assessments are also made on the participation of garnet in arc differentiation and on the production of arc batholiths.