On the evolution of marginal basins

Abstract

This paper presents a critique of the models proposed to explain the formation of back arc basins: those marginal basins located behind active or inactive trench systems and whose origin is inferred to be subduction related. Several marginal basins are not generically back arc, but rather are related to the initiation of subduction, continental rifting, or plate boundary readjustments. Models of back arc basin formation must be consistent with the observed characteristics of back arc basins, which include the following: (1) the crustal structure and magnetic lineation patterns of back arc and major ocean basins are similar, implying that the processes of crustal accretion in both tectonic environments are also similar; (2) arc‐trench systems without currently spreading back arc basins are in the majority; (3) although petrographically similar to and within the compositional range of midocean ridge basalts (MORB), back arc basin basalts (BABB) show consistent geochemical differences from N‐type MORB and are in many respects transitional toward island arc tholeiites (IAT); (4) distinct BABB and IAT mantle sources may be in extremely close proximity during the initial rifting of an island arc; (5) rifting of the volcanic arc may occur on either side of the line of active volcanoes (±50 km) and may vary from one side to the other along strike; (6) a remnant arc is not always developed; (7) back arc spreading may be initiated above subducted lithosphere, but with time the center of spreading migrates away from the arc and no longer overlies a Benioff zone; and (8) there is no simple correlation between the timing of global plate reorganizations and the formation of back arc basins. The three main classes of models proposed to explain back arc basin formation are mantle diapirism, induced aesthenospheric convection, and global plate kinematics. Because the driving mechanism for the first two classes of models is provided by the local subducting slab, these models fail to explain the temporal and spatial distribution of back arc basins. The third class of models proposes that back arc basins should form whenever global plate interactions require divergence between the overriding plate and the trench line, the latter being observed to move only in a seaward direction with respect to a hot spot reference frame. However, the tectonic setting of several back arc basins suggests that they represent more than just a passive response to kinematic boundary conditions. For example, some dynamic back arc force seems required to move the New Hebrides, Tonga, and South Sandwich forearc plates seaward. Neither the local nor the global models adequately specify the necessary and/or sufficient conditions for back arc basin formation. A better understanding of the processes associated with the initiation of rifting and of subduction is required. Arc rifting and back arc spreading are both secondary phenomena to subduction, but they are also primary extensional phenomena, fundamentally analogous to continental rifting and the initial spreading of major ocean basins.