Plain Language Summary
The process of intra-oceanic subduction brings an oceanic slab under an overriding oceanic slab resulting in the formation of a convergent plate margin. Consequently, an oceanic island arc is formed in the upper plate, as is the case of the magmatically active arcs of southwest Pacific. Unlike continental magmatic arcs, intra-oceanic arcs are less studied because a large part of them is located below sea level, emerging as chains of small islands that constitute just the tops of large submarine volcanoes. In the northern Dominican Republic, recent geochemical studies of the Caribbean volcanic and plutonic rocks indicate that older tholeiitic and boninitic melts were successively replaced by younger island arc tholeiitic melts. This change in the compositional magmas, as well as related mantle sources, places important constraints on the magmatic and tectonic processes associated with the initiation and evolution of the Caribbean island arc. In this sense, the results presented in this work allow to be compared with the chemical stratigraphy observed in actual oceanic arcs and with the predictions of models for the initiation of intra-oceanic subduction, which constitutes one of the main questions not completely resolved of the global plate tectonics.
1. Introduction
The process of intra-oceanic subduction brings an oceanic slab under an overriding oceanic slab resulting in the formation of a convergent plate margin. Consequently, an oceanic island arc is formed in the upper plate, as is the case of the magmatically active arcs of Izu-Bonin-Mariana, South Sandwich and Lesser Antilles (Leat & Larter 2003; Stern, 2010; Arculus et al., 2015). Unlike continental magmatic arcs, intra-oceanic arcs are less studied because a large part of them is located below sea level, emerging as chains of small islands that constitute just the tops of large submarine volcanoes. Despite these difficulties, the magmatic processes in intra-oceanic arcs have been directly and indirectly studied from: (1) lower crust and upper mantle xenoliths erupted in active volcanoes (McInnes et al, 2001; DeBari & Green, 2011); (2) diving, dredging and drilling partial crustal exposures on the deep sea floor (Pearce et al., 1992; Taylor et al., 1994; Ishizuka et al., 2006; Reagan et al., 2010, 2019); and (3) from geophysical surveys of the island arc crust (Takahashi et al., 2008; Calvert, 2011).
Direct evidence of the processes controlling the evolution and formation of volcanic arcs also comes from the obducted sections of intra-oceanic arc lithosphere that form ophiolitic sequences in orogenic belts (Pearce, 2003; Stern et al., 2012). However, examples of well-preserved exhumed arc sections, complete from their mantle roots to upper volcano-sedimentary levels are very scarce. The best studied arc sections probably are: the Jurassic Talkeetna arc in south-central Alaska (Green et al., 2006; DeBari & Green, 2011; Kelemen et al., 2014); and the Cretaceous Kohistan arc in northern Pakistan (Garrido et al., 2006, 2007; Jagoutz et al., 2007, 2011, 2018; Dhuime et al., 2007; Burg, 2011; Bouilhol et al., 2015). Both Talkeetna and Kohistan paleo-arcs are compositionally stratified and contain a lower section made up of a basal ultramafic sequence of peridotite and pyroxenite, overlain by a mafic sequence of gabbroic rocks. To explain the genetic link between the ultramafic and mafic sequences two main hypotheses have been proposed.
The first hypothesis suggest that the ultramafic-mafic sequence, composed of dunites, wehrlites, pyroxenites, hornblendites and gabbronorites, may have crystallized in the upper mantle and lower crust from a single type of primitive arc magma [Mg#>60; where Mg# = molar 100×Mg/(Mg+Fetotal)] (Greene et al., 2006; DeBari & Green, 2011; Kelemen et al., 2014). The existence of primitive gabbronorites and the complementary compositions of the more evolved plutonic and volcanic rocks, together with the rather homogenous Nd-isotopic compositions of diverse igneous units of the arc, are put forward to argue for a common origin (magmatic or cumulative) for the ultramafic and mafic rocks in the crustal section through (simple) fractional crystallization (Greene et al. 2006; Kelemen et al. 2003; Rioux et al. 2007; DeBari & Green, 2011). Therefore, the gabbronorites would represent the crystallized cumulate pile and the erupted volcanic rocks the residual liquid following differentiation. This hypothesis is supported by experimental studies (e.g. Müntener et al., 2001; Villiger et al., 2004, 2007; Müntener & Ulmer, 2018), which successfully reproduced the formation of high-Mg# pyroxenites and complementary low-Mg# melts during the crystallization of anhydrous primitive magmas at lowermost arc crust conditions.
In the Kohistan paleo-arc, however, the scarcity of rocks with intermediate Mg# values between high-Mg# dunites-wehrlites-pyroxenites and overlying gabbros, as well as the existence of significant variations in the Sr-Nd-Pb isotope data between these groups of rocks, rule out a simple fractional crystallization relationship between the ultramafic and mafic sequences. These petrological characteristics and REE numerical modeling suggest a second hypothesis for the origin of the ultramafic sequence by melt-rock reaction at the expense of the sub-arc oceanic mantle. (Garrido et al., 2006, 2007; Dhuime et al., 2007; Burg, 2011). Although predicted by crystal fractionation models, a thick ultramafic layer of cumulates is nevertheless absent in the crustal section of both arcs. This absence has been interpreted as a consequence of delamination of dense, unstable lower crust and/or convective thermomechanical erosion of the sub-arc lithosphere (Jull & Kelemen, 2001; Garrido et al., 2006, 2007; Dhuime et al., 2007; Kelemen et al., 2014). Later studies establish a more complex magmatic evolution for the Kohistan arc that includes different mantle sources for the ultramafic and mafic rocks throughout an extended period of ca. 30 Ma. This evolution includes a first stage of extensive boninitic magmatism connected with initiation of subduction, followed by a tholeiitic magmatism second stage associated with the building of a mature arc. This last stage culminates with granitic magmatism that produces intra-crustal differentiation (by fractionation process), associated with delamination and/or erosion of the lower arc crust (Dhuime et al., 2007; Jagoutz et al., 2011, 2018; Jagoutz & Schmidt, 2012; Stern, 2010; DeBari & Green, 2011).
A multi-stage tectono-magmatic evolution has also been proposed to explain the characteristics of the mantle and crustal sections of the Puerto Plata ophiolitic complex (PPC), which constitutes a segment of the Caribbean, intra-oceanic island arc (Escuder-Viruete et al., 2006, 2014). Currently preserved at several places in the Greater Antilles, the Caribbean island arc contains volcanic rocks as old as Late Aptian to Lower Albian in northern and central-eastern Dominican Republic (Kesler et al., 2005; Lewis et al., 2002; Escuder-Viruete et al., 2006, 2014; Jolly et al., 2006; Proenza et al., 2006; Marchesi et al., 2006; Rojas-Agramonte et al., 2011, 2016; Hastie et al., 2013; Torró et al., 2017). Following Draper et al. (1994), the arc is generally interpreted to have formed in a supra subduction zone (SSZ) setting at the leading edge of the Caribbean plate by SW-directed subduction (present-day coordinates) of the proto-Caribbean lithosphere.
In the northern Dominican Republic, geochemical studies of the Caribbean volcanic rocks indicate that older LREE-depleted tholeiitic and boninitic melts were successively replaced by younger island arc tholeiitic (IAT) melts (Escuder-Viruete et al., 2006, 2014). This change in the compositional magmas, as well as related mantle sources, places important constraints on the magmatic and tectonic processes associated with the initiation and evolution of the Caribbean island arc. These changes coincide with the chemical stratigraphy observed in actual oceanic arcs (Ishikawa et al., 2002; Ishizuka et al., 2006, 2011; Reagan et al., 2010, 2019) and with models for the initiation of intra-oceanic subduction (see review in Stern & Gerya, 2018). Recent advances in regional geological knowledge have made it possible to identify the plutonic rocks that constitute the lower crust of the Caribbean arc and their complementary volcanic rocks in the upper crust, which have been very little studied.
Here, we combine field mapping, petrological, mineralogical and geochemical data in order to characterize the lower crust of the Caribbean island arc exposed in the Rio Boba mafic-ultramafic plutonic sequence in the northern Dominican Republic. The main objective is to establish the petrogenetic relationships among the cumulate pyroxenites and gabbronorites of the plutonic complex, and the structurally adjacent mafic metavolcanic rocks of the Puerca Gorda Schists. These relationships allow us to (1) constraint the main differentiation processes in the magmatic system, (2) reconstruct the crustal section of the intra-oceanic Caribbean island arc, (3) place constraints on the nature of parental magmas during subduction zone infancy, and (4) propose regional correlations based on a spatial/temporal evolution in stages for the arc magmatism.