Nova Terra 52

subduction initiation, where the subduction erosion might have played an important role on the arc magmatism. The first igneous event recorded in the SW Iberian Massif is rep- resented by metabasites intercalated within the lower (and older) Montemolín Formation. According to the geochemical data, the mafic protoliths may be interpreted as primary calc-alkaline arc melts (Hastie et al., 2007). These melts would derive from a mantle source, enriched and modified by subducted slab components, as suggested by Straub et al. (2014, 2020) and Stern (2020), during subduction initiation processes. This enrichment is also visible through the high values of the Nd model ages ranging between 1316 and 2010 Ma which overlap the model ages range obtained for the Serie Negra Group, suggesting analogous isotopic sources. The comparison e Nd of the Serie Negra Group also suggests these metamafic rocks hold mantle juvenile influence shown through their more positive e Nd (t) values (Supplementary Table S4 and Fig. 12), ranging in a narrow range from slightly negative to posi- tive values. Accordingly, it is not possible to rule out the participation of a pre-existing (non-preserved/unexposed) crustal source, such as a crystalline fore-arc basement, or even another non-preserved metasedimentary sequence. In this regard, it is worth mentioning that Cadomian magmatic activity in African peri -Gondwana can be tracked as back in time as c. 750 Ma (via zircon inheritance), but the oldest metasedimentary series found in Iberia so far is some 150 m.y. younger. If those numbers are observed for the sec- tion of the arc preserved in the Mérida Massif, the oldest metased- imentary series ( 600 Ma) is 50 m.y. younger than the oldest magmatism deduced from ages of inherited zircons ( 645 Ma). The isotopic signature of the metabasites of the Montemolín Fm. largely coincides with the range of those of North African gran- itoids, which appear intruding older series between c. 609 and 580 Ma (Fig. 12a, Errami et al., 2009; El Haïbi et al., 2021; and ref- erences therein). These model ages are assumed as a product of recycling of old WAC sources, mixed with juvenile mantle sources. On the other hand, although the extensive presence of inherited zircon in all metaigneous complexes imply strong recycling of the arc itself, the model ages calculated for these rocks do not share the old isotopic sources of the enriched mantle wedge shown by the metamafic rocks, not being possible the complete derivation from the same source area. A more juvenile source must necessar- ily be involved in the generation of these metagranitic melts. Juve- nile isotopic sources derived from the oceanic plate incoming in the beginning of the subduction initiation (c. 600 Ma) are recorded through the Calzadilla Ophiolite in the studied arc section (Arenas et al., 2018). Only the participation of ancient cortical sources, such as those involved in the enrichment of the wedge mantle, mixed with more juvenile derived from the mafic materials of the oceanic plate incoming, is able to produce the model ages shown by meta- granitic complexes (Fig. 12a). This agreed with the involvement of juvenile material and/or limitation of the crustal contribution to these melts indicated by the more radiogenic values of e Nd (t) and a wide dispersion in the Sr values, which could be due to geodynamic reasons. Unlike the metamafic rocks, the greater depletion observed in the HFSE and HREE, together with high values of La/Yb or Sr/Y of the San Andrés orthogneisses suggest derivation from a primary source, classically associated with adakitic signature magmatism. This singular fea- ture makes the involvement (through extreme fractionation) of the metabasites into the generation of the San Andrés assemblage unlikely. Classical slab-derived adakites differ from the rocks with adakitic signature studied in this work in that the latter are essen- tially peraluminous. However, the San Andrés assemblage shares geochemical features with adakitic rocks derived from subduction of continental crust (Stern and Kilian, 1996; Wang et al., 2008; Eyuboglu et al., 2013, 2018; Azizi et al., 2019), such as low contents of MgO, Ni, Cr and V, which are inconsistent with the direct melt- ing of the mantle or the reaction of oceanic type adakitic melts with mantle. Moreover, these rocks have more evolved isotopic composition and higher K 2 O, Th and Th/La values than the tradi- tional slab-melts derived adakites (Defant and Drummond, 1990; Kelemen et al., 2003; Plank, 2005), indicating that their sources were not (only) basaltic juvenile crust. The path followed by the samples on the Sr/Y vs. Y and (La/Yb) N vs. Yb diagrams suggests slab melting with a garnet amphibolite restite (Defant and Drummond, 1990; Münker et al., 2004; Macpherson et al., 2006). Subduction to great depths (eclogite facies; also suggested by dm average calculated for these ratios) of crustal materials rich in equilibrium with garnet and or amphibole, would generate a melt with adakitic signature relatively rich in Sr and depleted in HREE and HFSE and, on the other hand, a residue more enriched in Y + HREE. However, the resulting melt (Wang et al., 2008; and ref- erences therein) would have a composition still far from the pera- luminous character shown by the San Andrés metaigneous complex. Large LREE and LILE enrichment shown by this metaig- neous complex is compatible with enrichment by crustal assimila- tion in the latest stages of its generation. This is in agreement with the enrichment due to melting of sediments and/or contamination at upper levels of the continental crust, recognizable through its high Ce/Yb and Th. The igneous activity recorded in this section allows the recognition of an adakitic signature in the metagranitic rocks, at least until c. 550 Ma. The rocks belonging to the Valverde metaigneous complex show a model age range equivalent to that obtained for the San Andrés rocks, while the more negative e Nd (t) values suggest higher crustal contribution and/or reworking of pre- vious igneous rocks for the generation of this assemblage. In arcs or arc segments where basalts are absent or less significant, and ada- kitic rocks comprise a significant proportion of the magmatism as in the case of Mérida Massif, melting of the subducted slab, includ- ing crustal eroded components, has been suggested as the source of these rocks (Kilian and Stern, 2002; Bindeman et al., 2005). Sub- duction only of metasediments equivalent to the preserved Serie Negra Group can be ruled out, because these studied sequences do not have the same isotopic leverage as Late Neoproterozoic granites. Recently, Sarrionandia et al. (2020) reported a U-Pb zircon age of 534 Ma for the andesitic rocks of the Malcocinado Forma- tion, which appears unconformably about the Serie Negra Forma- tion. Part of these andesites is also interpreted as adakitic rocks formed in a mature stage of the peri -Gondwanan volcanic arc. Subduction-related magmatic activity remained active at c. 541 Ma, when large volumes of igneous rocks of intermediate- mafic composition intruded into the studied section of the arc. These metaigneous complexes differ geochemically from previous adakite-like metaigneous complexes, reflecting differences in their petrogenetic mechanisms and/or sources. The Valle Real and Don Álvaro metaigneous complexes share similar geochemical, iso- topic, and geochronological features, which is compatible with derivation from a single magmatic source. The least fractionated REE patterns, with slight enrichment in LREE and values close to unity in the HREE, along with low Sr/Y and La/Yb ratios, and clearly negative Eu and Sr anomalies are typical of a ‘‘normal” calc- alkaline arc signature (Defant and Drummond, 1990; Martin, 1999), derived from hydrous metasomatism of an overlying mantle wedge at lower pressure in equilibrium with plagioclase. A High U/ Th ratio close to those indicated as typical of a volcanic arc (c. 40; Sun and Sterns, 2001) suggests a greater contribution from meta- somatized mantle for the generation of these magmas. The Valle Real and Don Álvaro metaigneous complexes do not show atypical values in the Ce/Yb ratios that suggest modification by fluids or sig- nificant crustal input. Ediacaran to Lower Cambrian xenoliths in the intermediate metagranitoids have been traditionally consid- ered to have been derived directly from melting of the Serie Negra E. Rojo-Pérez, U. Linnemann, M. Hofmann et al. Gondwana Research 109 (2022) 89–112 &KDSWHU