Global volcanic arc magma composition correlates with thickness of both arc crust (Moho depth) and arc lithosphere (LAB depth): A revealing message on arc basement histories and arc magmatism

Abstract

Past fifty years have seen mounting publications on the genesis of volcanic arc magmas. While details remain debated, it is generally agreed that arc magmas result from slab-dehydration induced mantle wedge melting followed by crustal level differentiation of varying extent and sophistication. Two recent arc magma studies deserve particular attention because they attempt to discuss globally unifying controls on arc magma composition. Both Harvard study (Turner and Langmuir, 2015a,b) and Rice study (Farner and Lee, 2017) show correlations of arc magma composition with crustal thickness and both ascribe the crustal thickness as the principal control on their observed magma compositional variations, yet the physical role of the crustal thickness in their interpretations is markedly different because of (1) the ambiguous use of “crust” and (2) their different magma compositional ranges chosen in discussion. The Harvard study only uses basaltic samples corrected to MgO = 6.0 wt.% to discuss mantle processes and interprets the arc crustal thickness as restricting the mantle wedge melting, i.e., the extent of melting decreases with increasing crustal thickness. The Rice study uses samples of all compositions (basaltic to rhyolitic), whose extent of differentiation increases with increasing crustal thickness, interpreted as Moho-crossing mantle wedge melts travelling greater vertical distance with greater degree of cooling and erupting more evolved compositions above thicker crust than melts erupted above thinner crust without need of invoking mantle wedge processes. We commend these efforts and approve their different approaches but emphasize that the unifying understanding of global arc magmatism requires clearly defined Moho (the base of the crust) and LAB (the lithosphere-asthenosphere boundary) and their intrinsic controls on mantle wedge melting (Harvard Study model) and crustal level magma differentiation (Rice Study model) beneath global arcs. In this study, we use chemical compositions of 36,945 global arc volcanic samples provided by the Rice study together with the literature data on seismic Moho and LAB depths of these sample locations to establish (1) the correlation of crustal thickness interval averaged magma composition with crustal thickness and (2) the correlation of lithosphere thickness interval averaged magma composition with lithosphere thickness. These correlations reaffirm our understanding that the lithospheric mantle must exist beneath volcanic arc crust with a globally averaged LAB/Moho depth ratio of 3.26 ± 0.63, i.e., the arc crust is on average about 31.8% ± 6.1% of the lithosphere thickness. This knowledge forms a solid constraint essential for models of global arc magmatism.