Decades of study of arc volcanism have led to a plethora of models to account for global and local observations of arc magma chemistry. Perhaps the most obvious first-order observation is that stratovolcanoes built on thick continental crust have substantially elevated incompatible element abundances compared to island arc volcanics. Hypotheses for these differences include (1) a greater influence of differentiation and contamination, (2) slab diapirs and subduction erosion that provide different source materials to the melting regime, (3) a variable slab flux controlled by the slab thermal structure, (4) heterogeneity of the ambient mantle wedge, and (5) variations in extent of melting caused by wedge thermal structure. Here I evaluate these models using arc geochemistry, experimental results, physical models, and tectonic constraints, and show that the first three models fail various tests as explanations for these first order observations. There is compelling evidence for the latter two models, however, which in combination provide straightforward explanations for the first-order features of arc geochemistry that can be quantitatively verified. Perhaps the most interesting outcomes of this quantitative analysis is that melting of both sediments and mafic ocean crust are necessary to account for the compositions of all arc-front stratovolcano lavas, and that trace element ratios that have historically been associated with an “aqueous fluid” component are better accounted for by variations in ambient mantle composition alongside variations in the sediment compositions.
Get the latest
Subscribe to our newsletters.