Exhuming a cold case: The early granodiorites of the northwest Iberian Variscan belt—A Visean magmatic flare‑up?

In this study we report laser ablation–inductively coupled plasma–mass spectrometer U-Pb ages of granitoids from the so-called early granodiorites of the northwest Iberian Variscan belt. The U-Pb results attest to significant magmatic activity in Visean time (ca. 347–337 Ma) that generated a hitherto poorly constrained granitoid suite in the northwest Iberian tract of the western European Variscan belt realm. This early Carboniferous suite (ECS) is mainly composed of peraluminous cold and hot crustal granodiorites and monzogranites with minor associated mafic rocks that attest to minor involvement of mantle melting. Based on the geochronological and geochemical data, we compare the Visean granitoids with younger Variscan granitoids in northwest Iberia and, in view of the tectonothermal scenarios of the Variscan collision in northwest Iberia, propose a model for the genesis of the ECS in northwest Iberia that involves rapid melting upon fast exhumation of the thickened Gondwanan crust in the course of the protracted Variscan collision.


INTRODUCTION
Variscan granitoid magmatism in northwest Iberia is an example of the intimate link between granitoid magma production and plate convergencecollisional-postcollisional geodynamic scenarios. In the western European Variscan belt realm (WEVB, Fig. 1), the major stages of convergent tectonics and their aftermath are recorded by granitoid suites generated in a time span of ~50 m.y. . There is, however, some controversy regarding (1) whether there was a significant magmatic event in the early Carboniferous, for which there was previously scarce evidence (see following), and (2) whether the Variscan magmatic activity was continuous or occurred through discrete and relatively short-lived pulses. Unravelling these issues is one of the key elements to better constrain and interpret the different processes involved in the collisional scenarios that led to the Variscan orogeny in Iberia during late Paleozoic time (e.g., Matte, 2001;Martínez Catalán et al., 2007;Ballèvre et al., 2009;Arenas et al., 2016; and the subsequent development of the Ibero-Armorican arc (e.g., Weil et al., 2013, and references therein; Fig. 1).
The hypotheses that support rather continuous magmatic activity are mainly derived from the combined interpretation of radiometric ages of granitoids obtained by different methods 40 Ar-39 Ar, K-Ar; e.g., Serrano Pinto et al., 1987;Díaz Alvarado et al., 2013;Martínez Catalán et al., 2014) and the numerical modeling of the thermal evolution of the orogen, including radioactive heat production (e.g., Bea et al., 1999Bea et al., , 2003Bea, 2012;Alcock et al., 2015). In essence, these hypotheses view melt production in the collisional scenario mainly as the result of severe crustal thickening of a fertile (graywacke and pelite rich) crust followed by an interval of thermal relaxation (e.g., Martínez-Catalán et al., 2014) accompanied by melting of the thickened crust.
The advent of a large amount of more precise U-Pb age data on Variscan granitoids (and some volcanic rocks) from northwest Iberia in recent years (e.g., compilation in Gutiérrez-Alonso et al., 2011;Martínez Catalán et al., 2014) has provided a more focused picture of the magmatic history of the WEVB, providing solid ground for interpretations that link periods of more intense magmatic activity with large-scale crust-mantle processes involved in the collisional orogeny and the subsequent development of the West European Variscan Belt (e.g., Castro et al., 1999Castro et al., , 2000Fernández-Suárez et al., 2000;Villaseca et al., 2009Villaseca et al., , 2011Gutiérrez-Alonso et al., 2011;Orejana et al., 2012;López-Moro et al., 2018).
Based on those more precise U-Pb ages, three main pulses of magmatic activity seem to be well established.
3. A third suite of Variscan granitoids (the object of this study), putatively older than the syntectonic leucogranitoids, was postulated in the early works of Den Tex (1966), Capdevila (1969), and Capdevila et al. (1973), based mainly on the presence of a tectonic foliation and a more mafic composition than that of the synextensional granitoids. They are commonly labeled Granodioritas precoces (early granodiorites), following De Pablo Maciá (1981) and González Lodeiro et al. (1984), and were interpreted as being deeper seated intrusions than the POS. In the following we refer to this suite as the ECS (early Carboniferous suite). Gallastegui (2005) dated the intrusion of Bayo-Vigo applying U-Pb zircon dating, and Gloaguen (2006) and Gloaguen et al. (2006) used U-Th-Pb chemical dating of monazite to date the Chantada intrusion (both belonging to this group; Fig. 2), obtaining moderately precise ages between ca. 350 and 330 Ma. Moreover, there are K-Ar and Rb-Sr isochron ages ca. 340 Ma in granitoids putatively belonging to this suite in northwest Iberia (see compilation in Serrano Pinto et al., 1987). In addition, ca. 340 Ma zircon xenocrysts have been found in the ca. 320 Ma syntectonic leucogranitoids of the Tormes dome and surrounding areas (López-Moro et al., 2018), in the ca. 305 Ma Toledo Anatectic Complex (Martín Garro, 2015), and as detrital zircons in Variscan synorogenic sediments of the Cantabrian zone (Pastor- Galán et al., 2013) and the Central Iberian and Galicia Tras-os-Montes zones (Da Martínez Catalán et al., 2016). Furthermore, volcanic rocks (ash-fall beds) in the Cantabrian zone have been recently dated as ca. 340 Ma (Merino-Tomé et al., 2017). Monazite and rutile of this age have been found in in migmatitic rocks of the Basal Units of the allochthonous complexes in northwest Iberia (Abati and Dunning, 2002). However, despite the data cited above, the existence of a suite of Variscan granitoids older than ca. 330 Ma has been called into question (e.g., Rodríguez et al., 2007;Martínez Catalán et al., 2014;Alcock et al., 2015).
To perform this test we undertook laser ablation-inductively coupled plasma-mass spectrometer (LA-ICP-MS) U-Pb dating of zircons from 13 samples taken from intrusions ascribed to the early granodiorites (Fig. 2) in northwest Iberia. Thorough reviews of the main geological and petrological features of these granitoids can be found in González Lodeiro et al. (1984), Bellido Mulas et al. (1987, and Gallastegui (2005). As shown here, our work proves beyond reasonable doubt that such a group of older granitoids exists in northwest Iberia. Their ages are constrained between 347 and 337 Ma (Visean), an age bracket similar to that of their west European counterparts (see references in paragraph above).

REGIONAL SETTING
The collision between Laurussia and Gondwana and several putative microplates resulted in the late Paleozoic orogen in central and western Europe known as the Variscan orogen (e.g., Nance et al., 2010), that led to the amalgamation of Pangea. The WEVB is classically divided into a number of zones based on the differences in the early Paleozoic stratigraphy as well as structural style, metamorphism, and magmatism (Lotze, 1945;Julivert, 1971;Franke, 1989;Martínez Catalán et al., 1997Ballèvre et al., 2014). These zones broadly correspond to increasing distance from the Gondwanan margin toward the Rheic Ocean (Fig. 1). www.gsapubs.org | Volume 10 | Number 2 | LITHOSPHERE The northwest Iberian massif ( Fig. 1) includes Neoproterozoic rocks with Gondwanan (North Africa) affinity, which form the basement for late Ediacaran-early Cambrian subduction-related and early Paleozoic passive margin sequences, respectively (e.g., Rodríguez Alonso et al., 2004;Murphy et al., 2008;Fernández-Suárez et al., 2014;Rubio-Ordóñez et al., 2015). In Late Devonian and Carboniferous time, the collision of the passive margin of Gondwana (acting as the lower plate) with Laurussia (upper plate) resulted in the Variscan orogen and ultimately in the amalgamation of Pangea (e.g., Matte, 1986Matte, , 2001Murphy et al., 2009;Kroner and Romer, 2013).
The earliest record of Variscan deformation in Iberia is preserved at ca. 400 Ma in the allochthonous units (Figs. 1 and 2) (e.g., Dallmeyer and Gil Ibarguchi, 1990;Gómez Barreiro et al., 2006); the continentcontinent collision occurred later, ca. 365 Ma (e.g., Dallmeyer et al., 1997;López-Carmona et al., 2014), when the Gondwanan, most external, part of the margin (the Basal Units, Fig. 2) attained its maximum depth of burial under the continental units now preserved in the upper units (Fig.  2). Deformation migrated toward the east (in present-day coordinates), culminating with the development of the foreland fold and thrust belt (the Cantabrian zone) during Late Mississippian-Early Pennsylvanian time (ca. 325-320 Ma). After these events, the orogenic edifice was buckled around a vertical axis, giving rise to the orocline known as the Iberian Armorican arc, or the Cantabrian orocline (e.g., Weil et al., 2013;Pastor-Galán et al., 2015, 2016. A sequence of magmatic events is recorded in the rocks affected by the Variscan orogeny of Iberia. Some of them developed prior to the Variscan orogeny, while others took place coevally, or immediately after, and can be summarized as follows. (1) A subduction-related Cadomian (ca. 600 Ma) magmatic event dominated by I-type granitoids and volcanic rocks (Fernández-Suárez et al., 1998;Rubio-Ordóñez et al., 2015), which is scarcely represented in northwest Iberia. (2) A voluminous extensionrelated late Cambrian to Early Ordovician magmatic event (ca. 490-470 Ma) is generally interpreted to be linked to the opening of the Rheic Ocean (Valverde-Vaquero and Dunning, 2000;Murphy et al., 2006;Bea et al., 2006b;Montero et al., 2007;Díez Montes et al., 2010;Talavera et al., 2013). (3) A scarcely represented volcanic event ca. 400-390 Ma is interpreted to represent the extension associated with the far-field effect of the Rheic ridge subduction under its northern margin . (4) Carboniferous synorogenic (Variscan) magmatism was concentrated between ca. 325 and 315 Ma (see Introduction). (5) Postorogenic magmatism peaked ca. 305 and 290 Ma, when voluminous granitoids and some mafic rocks with their extrusive equivalents were emplaced and erupted in both the internal and external zones of the orogen.

MAIN FEATURES OF THE ECS GRANITOIDS
The dated intrusions belonging to the ECS show elongated shapes that follow the structural grain of the area where each of them was emplaced Visean magmatic flare-up in Variscan northwest Iberia | RESEARCH (Fig. 2). This feature has been classically interpreted as the result of their emplacement having occurred at deeper crustal levels than that of the younger granitoid groups (Capdevila et al., 1973;Bellido Mulas et al., 1987). Figure 2 is a simplified geological map of northwest Iberia showing the plutons dated in this study and the location of each sample.
Here we present a brief description of the samples from this study together with a summary of the main features of the intrusions taken from the literature (González Lodeiro et al., 1984;Bellido Mulas et al., 1987;Barrera et al., 1989;Gallastegui, 2005).
The Ricobayo pluton (GP-2) is the southernmost of the studied intrusions. It intruded early Paleozoic sedimentary rocks of the autochthonous sequences of the Central Iberian zone (CIZ). The intrusion is made up mainly of heterogranular biotite and two-mica monzogranites, partially episyenitized (López-Moro et al., 2013), displaying a cataclastic to protomylonitic texture imparted by a late to post-Variscan shear zone (Villalcampo shear zone; González-Clavijo et al., 1993) dated as 306 ± 6 Ma .
The Espenuca intrusion (samples GP-5, GP-6, and GP-7) is an elongated pluton, more than 65 km in length, intruding the GTM parautochthonous unit and the allochthonous complexes (upper unit of the Órdenes complex). It is mainly composed of biotite ± muscovite monzogranites and was strongly affected by the Valdoviño fault in its western part (Ortega et al., 1994). Ortega (1998) obtained an Rb-Sr isochron age of 319 ± 20 Ma for the main facies of this granitoid.
The Santa Comba granodiorite, also known as Negreira granodiorite in the regional literature (Parga Pondal, 1956), is a north-south-elongated body (sample GP-8) intruding the GTM parautochthonous unit. The studied intrusion is affected in its northern boundary by the Pico Sacro detachment, which is the lower limit of the allochthonous units in this area, and interpreted to be of late Carboniferous age (Gómez- Barreiro et al., 2010). This intrusion is locally affected by the Padrón dome migmatites that are interpreted to have formed ca. 320 Ma (Díez Fernández et al., 2017, and references therein).
Sample GP-10 was collected in the Meabia pluton, which is the only one of the studied plutons to have clearly intruded in the allochthonous units (Lalín Forcarey Unit; Klein, 1982;Marquínez, 1984).
The Bayo-Vigo granodiorite is a north-south-elongated intrusion located to the west of the westernmost allochthonous unit (Malpica-Tui Complex) and separated from it through the ca. 307 Ma Malpica-Lamego shear zone ( 40 Ar-39 Ar; Rodríguez et al., 2003;Gutiérrez-Alonso et al., 2015). There are minor Mg-K biotite-rich quartz-diorite rocks associated with this intrusion. These mafic rocks are known in the Variscan regional literature as Vaugnerites; because of the widespread use of this term (cf. von Raumer et al., 2014, and references therein) we retain its use herein.
The Bayo-Vigo intrusion was studied in detail by Gallastegui (2005), who obtained a U-Pb age (isotope dilution-thermal ionization mass spectrometry on multigrain zircon fractions) of 349 ± 15 Ma. Two samples from this intrusion were collected for this study: sample GP-11 represents the main granodiorite facies and sample GP-12 represents the spatially associated vaugnerites. Both samples were taken ~1 km apart.
Sample GP-13 was collected in the Avión intrusion, a deformed northsouth-elongated granodiorite pluton (Monteserín, 1981;Barrera et al., 1989) that intrudes the parautochthonous units and the southernmost part of the Basal Units cropping out in the Lalín-Forcarey domain (Ordenes Complex).

GENERAL GEOCHEMICAL FEATURES AND CLASSIFICATION OF THE ECS GRANITOIDS
The rocks putatively belonging to the ECS show a considerable variety of mineralogical and geochemical features, and the granitoids are in some cases associated with volumetrically minor mantle-derived mafic rocks, including vaugnerites (see compilation in Gallastegui, 2005).
Although the main aim of this study is to prove the hitherto questioned existence of an early Carboniferous granitoid suite in northwest Iberia, as well as to propose a geodynamic scenario for its genesis, we use whole-rock major and trace element analyses (analysis performed by Actlabs, www.actlabs.com) of the dated rocks (Table 1) along with some geochemical data from Gallastegui (2005) for the Bayo-Vigo granodiorites and vaugnerites, López-Moro et al. (2013) for the Ricobayo granite, and from the Spanish geological survey (IGME, Instituto Geológico y Minero de España) reports for the Chantada-Taboada, Meabia, and Avión intrusions (Barrera et al., 1989) to provide an overview of the main geochemical features of this granitoid suite.
Based on their major element composition and following the nomenclature of Frost et al. (2001), the granitoid rocks of the ECS of Iberia (Table 1; Fig. 3) can be classified as magnesian (with a transition to ferroan in the most silicic terms), alkali-calcic to calc-alkalic and peraluminous (Shand, 1943), with the exception of the more mafic terms, which are metaluminous. Note that in the SiO 2 versus Na 2 O + K 2 O-CaO diagram the more silicic terms shift to more calc-alkalic compositions, as noted by Frost et al. (2001) for the postorogenic Caledonian granites of Britain.
The analyzed samples have moderate rare earth element contents with LaN/LuN values between 7 and 40, and all samples display a negative Europium anomaly with Eu/Eu* values ranging between ~0.3 and 0.8 (Table 1; Fig. 4A).
The normal mid-oceanic ridge basalt (N-MORB) normalized plots (Fig. 4B) show a marked enrichment in the incompatible elements with positive anomalies in K and Pb and negative anomalies in Nb and Ti.
In the trace element discrimination diagrams Rb versus Y + Nb and Y versus Nb (Pearce et al., 1984;Figs. 4C, 4D), the ECS rocks straddle the fields of syncollisional granitoids and that of volcanic arc granites, a feature characteristic of many granitoids that are neither true volcanic arc nor Himalayan type (cf. Hopkinson et al, 2017). The same pattern is displayed by the granitoids of the POS of northwest Iberia on these diagrams (cf. Fernández-Suárez et al., 2000. Although the composition of the mafic rocks (vaugnerites) in the Bayo-Vigo intrusion has been plotted in Figures 3 and 4, our focus is on the granitoids. In the context of the objective of this work, the presence of these mafic rocks, that also occur in other intrusions of the ECS (Bellido Mulas et al., 1987;Gallastegui, 2005) but have not been dated, is used as an argument in the discussion to strengthen the notion that the genesis of the ECS was accompanied by some degree of partial mantle melting  Watson and Harrison (1983). ASI and Tzirc values are not given for samples GP5, GP6, and GP9 because they are altered.
(for discussions on the occurrence, age, and genesis of durbachite and/ or vaugnerite rocks in the Variscan belt of Europe, see Buda et al., 2004;Scarrow et al., 2009;von Raumer et al., 2014;Kubínová et al., 2017). Figure 5 shows some aspects of a broad comparison between the ECS and the POS (Fernández-Suárez et al., 2000. This comparison is pertinent because both suites of granitoids are relatively similar in their main petrological and geochemical features (e.g., Capdevila et al., 1973;Fernández-Suárez et al., 2000) and both show associated mantle-derived mafic rocks, at variance with the syntectonic leucogranitoids that are associated with migmatites, and are the product synextensional melting of the middle crust (e.g., López Moro et al., 2018, and references therein). Figure 5A shows a KDE (kernel density estimation) diagram of ASI values in samples from the ECS and the POS. Note that the peak of ASI values is ~1.05 in the POS and ~1.2 in the ECS. This is consistent with the involvement of a higher proportion of more aluminous rocks in the genesis of the ECS. This observation is also consistent with the trend defined by the ECS granitoids in the Ba-Zr diagram ( Fig. 5B; Clemens, 2014) that is rather characteristic of S-type granitoids. For comparison, representative Ba-Zr analyses of the POS have been plotted in this diagram. Although both suites (ECS, POS) are not strikingly different in this diagram, the ECS seems to show a better positive correlation of Zr and Ba. As both suites are mainly crustal and likely derived from different admixtures of metaigneous and metasedimentary protolith rocks (see following), it makes sense that the plot does not show a clear-cut difference between them. This observation is consistent with the postulated more I-type nature of most granitoids belonging to the POS see Villaseca et al., 1998, for a discussion on the significance of the I/S classification scheme in Iberian Variscan granitoids). The Th versus U diagram ( Fig. 5C) shows that for similar whole-rock Th contents, the POS granitoids (with only two exceptions) have significantly lower whole-rock U contents; at face value, this is consistent with the involvement of a more pelitic (and hence more aluminous) source in the genesis of the ECS.
This apparent geochemical source-composition difference between the ECS and the POS is also evident in the different apparent zircon inheritance patterns of both suites (see following).

GEOCHRONOLOGY
We collected 13 samples corresponding to intrusions ascribed to the early granodiorite group in the regional literature (López-Plaza and Martínez Catalán 1987;Bellido Mulas et al., 1987;Gallastegui, 2005) for this study (location in Fig. 2). Among all samples, sample GP-1 (not reported herein) did not contain any early Carboniferous zircon and sample GP-9 was devoid of zircon. All but one zircon analyzed in sample GP-1 were older than ca. 480 Ma and therefore can be considered as inherited zircon crystals. In this sample only one Variscan concordant zircon was found, with an age of ca. 317 Ma. For this reason, sample GP-1 has been excluded from this study, as there is no evidence that it belongs to the ECS.

Analytical Procedures
Zircons were separated at the University of Salamanca (Spain); ~8 kg of sample were crushed in a jaw crusher and sieved for the 63-400 µm fraction. Concentrates were obtained using a Wilfley table, Frantz isodynamic magnetic separator, and heavy liquids (diiodomethane).
Zircon crystals of all grain sizes and morphological types were selected, mounted in resin blocks, and polished to half their thickness at the Museum für Mineralogie und Geologie (Senckenberg Naturhistorische Sammlungen Dresden) for subsequent analysis. All grains were documented by backscattered electron and cathodoluminescence (CL) images using a www.gsapubs.org | Volume 10 | Number 2 | LITHOSPHERE scanning electron microscope (JEOL JSM-820 at the Research Assistance Centre of Geological Techniques, Complutense University, Madrid, Spain) to study their internal structure to select the best areas for laser ablation. Zircon was analyzed for U, Th, and Pb isotopes by LA-ICP-MS using a Thermo-Scientific Element 2 XR sector field ICP-MS coupled to a New Wave UP-193 excimer laser system. A teardrop-shaped, low-volume laser cell by Ben Jähne (Dresden, Germany) was used to enable sequential sampling of heterogeneous grains (e.g., growth zones) during time-resolved data acquisition. Each analysis consisted of ~15 s background acquisition followed by 35 s data acquisition, using a laser spot size of 25 µm. We analyzed 60 zircon crystals from each sample. If necessary, a common-Pb correction based on the interference-and background-corrected 204 Pb signal and a model Pb composition (Stacey and Kramers, 1975) was carried out. The necessity of the correction is judged on whether the corrected 207 Pb/ 206 Pb is outside of the internal errors of the measured ratios. Discordant analyses were generally discarded. Raw data were corrected for background signal, common Pb, laser-induced elemental fractionation, instrumental mass discrimination, and time-dependent elemental fractionation of Pb/Th and Pb/U using an Excel spreadsheet program developed by Axel Gerdes (Institute of Geosciences, Johann Wolfgang Goethe-University Frankfurt, Frankfurt am Main, Germany). Reported uncertainties were propagated by quadratic addition of the external reproducibility obtained from the standard zircon GJ-1 (~0.6% and 0.5%-1% for the 207 Pb/ 206 Pb and 206 Pb/ 238 U, respectively) during individual analytical sessions and the within-run precision of each analysis. Only concordant results (95%-105% concordance) were considered for single analysis. Concordia diagrams (2σ error ellipses) and concordia ages (95% confidence level) were produced using Isoplot/Ex 3.75 (Ludwig, 2012). (For further details on analytical protocol and data processing, see Frei and Gerdes, 2009.)

U-Pb Results: The Visean Event
Results of LA-ICP-MS U-Pb zircon analyses are listed in Table 2 and representative CL images of the analyzed zircons are shown in Figure 6. For the sake of clarity, we address first the results of U-Pb analyses interpreted to date the age of magmatic crystallization in each intrusion and then discuss the inherited zircon ages and inheritance patterns. Figure 7 shows the Wetherill concordia diagrams obtained from the U-Pb data interpreted to represent the magmatic crystallization age of each intrusion. A summary of the obtained concordia ages and the number of analyses is shown in Table 3; the ages range from 347 to 337 Ma. Because several of the used samples were collected in the same pluton, there are some considerations.
Note that hitherto, in the regional literature the intrusions from which samples GP-3 and GP-4 were collected are considered to be a composite intrusion, but our results indicate that sample GP-3 is ~5 m.y. older than GP-4, so they are not, strictly speaking, coeval, although they form a continuous body of granitoid exposure (Fig. 2).
Samples GP-6 and GP-6A (Irixoa-Espenuca granite) were taken ~25 m apart. The results of U-Pb analyses of both samples are shown separately in Table 2, but the data have been combined in one concordia plot, as the analyses are equivalent and the pooled 17 analyses yield a concordia Visean magmatic flare-up in Variscan northwest Iberia | RESEARCH age of 338.8 ± 2.0 Ma, considered to represent the crystallization age of this granitoid. Although samples GP-5, GP-6, and GP-7 belong to what was previously considered a single intrusion (the Espenuca pluton; Ortega et al., 1994;Ortega, 1998), our data show that samples GP-5 and GP-6 yield the same age within analytical uncertainty (ca. 338 Ma), and sample GP-7 ( Fig. 2) with a concordia age of 343.1 ± 2.6, is slightly older.
It is noteworthy that Figure 2 shows that samples GP-3, GP-4, GP-5, GP-6, and GP-7 represent the largest present-day outcrop of ECS granitoids in northwest Iberia, and taken as a whole seem to represent a large intrusion built over an apparent time span of ~5 m.y. (343-338 Ma).

Zircon Inheritance and Zircon Saturation Temperatures
The results of LA-ICP-MS U-Pb analyses older than ca. 347 Ma are considered to be inherited crystals in the studied samples. They are provided in Table 2, keyed to the corresponding sample reference number and summarized in Table 3. In addition, Figure 8 shows the concordia plots (Figs. 8A, 8B) of inherited zircons and the KDE plots ( Fig. 8C).
Because the CL-guided analyses ( Fig. 6) were aimed primarily at establishing crystallization ages for the studied intrusions, a few preliminary considerations are in order.
1. In zircons with a clear core-rim structure, some cores were analyzed (Fig. 6), but most cores were often avoided to maximize the number of ages corresponding to the magmatic event that generated the granitoid suite. In some cases, analyses performed on zircon centers with no observable core-rim structure yielded ages older than 347 Ma, and are thus interpreted to represent zircon xenocrysts incorporated into the magma.
2. The same approach was followed in every sample when selecting spot analyses.
3. Given the above, the inherited zircons found in the studied samples may be at best a broad proxy for the proportion of zircon xenocrysts, but do not necessarily provide an accurate datum for the proportion of the total inherited zircon component in a given sample.
4. The following discussion is solely based on concordant (or near concordant) analyses. Therefore, the weight of those zircons older than ca. 347 Ma that yielded discordant ages remains unknown.
Assuming that the real spectrum of inheritance may be different from that apparent from the concordant analyses obtained for each sample, as the methodological approach has been the same for all samples, we consider that comparative values of apparent inheritance between samples are still worth reporting and considering. A brief summary of the main inherited zircon results in each sample is shown in Table 3.
As regards U-Pb ages of inherited zircon in the studied samples, data are presented in Table 2, summarized in Table 3, and illustrated in Figure  8. Our data show that the inheritance pattern in the ECS as revealed by concordant analyses is characterized by the following age populations, from youngest to oldest: (1) ca. 400-390 Ma (4%), (2) ca. 490-470 (55%), (3) ca. 760-540 Ma (31%), (4) ca. 1500-900 Ma (6%), and (5) pre-Neoproterozoic (3%). Therefore ~85% of the apparent inherited population in the ECS is dominated by Ediacaran-Cryogenian zircon (31%) and early Paleozoic zircon (55%). In the northwest Iberian Variscan realm these ages correspond to well-known magmatic events (see Introduction) related to the Cadomian arc orogeny (e.g., Fernández-Suárez et al., 2011) and the Ollo de Sapo event related to the lithospheric thinning caused by the opening of the Rheic ocean (e.g., Murphy et al., 2006;Linnemann et al., 2008). Ollo de Sapo is the regional term for a thick succession of augen gneisses and porphyroid acid volcanics of Early Ordovician age that crop out in many sections of the Variscan belt in western Europe (e.g., Díez Montes et al., 2010;Montero et al., 2007Montero et al., , 2017, and references therein). www.gsapubs.org | Volume 10 | Number 2 | LITHOSPHERE     Visean magmatic flare-up in Variscan northwest Iberia | RESEARCH If we compare the apparent age spectrum of inherited zircons in the ECS and the POS (Fig. 8C), it is noteworthy the POS peak age is late Ediacaran (ca. 550 Ma), interpreted as corresponding to zircons derived mainly from metaigneous rocks of the Cadomian arc and/or their erosion products (Fernández-Suárez et al., 1998. In contrast, the age peak of the inherited population in the ECS is ca. 485 Ma (Fig. 8C), an age corresponding to the voluminous magmatic event that took place in Iberia on the extended Gondwanan margin in Early Ordovician time, recorded in the Ollo de Sapo rocks (see Fig. 2).
The rocks related to the Ollo de Sapo event (dominated by rhyolites, ignimbrites, tuffs, and graywackes with high proportion of volcanic component; e.g., Díez- Montes, 2007) are strongly peraluminous. In contrast, Cadomian magmatism is metaluminous to weakly peraluminous (Fernández-Suárez et al., 1998;Rubio-Ordóñez et al., 2015). Therefore, the apparently more peraluminous nature of the ECS (cf. Fig. 5A) is consistent with the observation that the ECS granitoids might have recycled a higher proportion of the Ollo de Sapo protolith rocks than the POS granitoids, and likely also include a higher proportion of pelite component.
Another difference in the inheritance pattern between the ECS and the POS is the presence in the ECS of ca. 400-390 Ma zircon ( Fig. 8C; see Introduction). This inherited component is also present in the syntectonic leucogranites (ca. 320 Ma; López-Moro et al., 2018) but absent in the POS. The origin of the zircons in this age group is not clear but zircons of this age are present in some mafic volcanic rocks in central Iberia  and are also abundant as detrital components in Devonian-Carboniferous sequences of southwest Iberia, where they have been interpreted as derived from the dismantling of (intra?) Rheic oceanic arc rocks (Pereira et al., 2018).
Sample GP-12 (vaugnerite, Bayo-Vigo intrusion) contains the ca. 400-390 Ma inherited component (and apparently none of the others). This observation, pointing to some degree of crustal involvement, is consistent with the hypothesis of Scarrow et al. (2009), who argued that vaugnerites could represent magmas derived from a mantle metasome that intruded a crustal anatectic zone where some degree of mixing and mingling may take place.
The zircon saturation temperatures of the ECS granitoids, based on our analyses and those available in the literature (Fig. 9A), have a bimodal distribution with maxima ~803 °C and 704 °C, essentially corresponding to the apparent occurrence in the suite of hot and cold granitoids (sensu Miller et al., 2003) with an apparent predominance of the former. This is consistent with the Ba and Pb contents and Pb/Ba ratios of these granitoids (Table 1; Fig. 9B). On the diagram of Finger and Schiller (2012) they plot on both sides of the empirical line that separates primary lower temperature S-type (muscovite ± incipient biotite breakdown) and higher temperature S-types (muscovite + biotite breakdown), which represent the predominant type in our data set in agreement with the apparent predominance of hot granites.
In our case study there seems to be little or no correlation between the inheritance features and other geochemical parameters (e.g., ASI, SiO 2 content, maficity) or crystallization age within the suite, in agreement with the observations made by Miller et al. (2003).

DISCUSSION
The primary conclusion of this study is that it proves the existence of a volumetrically significant Visean magmatic event in northwest Iberia. Except for sample GP-1, the age of which remains unconstrained, the age data of all other samples in this study confirm the existence of an early granodiorite group between 347 and 337 Ma, as assumed by previous work. Therefore, the ECS constitutes the oldest known manifestation of Variscan magmatism in northwest Iberia.
The first objective of this study has been fulfilled and our results challenge any model, hypothesis, or speculation that does not take this evidence into account.  www.gsapubs.org | Volume 10 | Number 2 | LITHOSPHERE Moreover, our study has not included all intrusions that are possible candidates to belong to the ECS in northwest Iberia. Accordingly, it is to be expected that further geochronological work will reveal the existence of more intrusions in this age bracket in the northwest Iberian realm of the WEVB (e.g., northern Portugal, southeast CIZ). In addition, the syntectonic leucogranitoids generated ca. 320 Ma and the POS contain abundant ca. 340 Ma inherited zircon (Martín Garro, 2015;López-Moro et al., 2018), even in areas without any known exposures of the ECS granitoids. This is another argument in favor of the idea that the ECS represents a larger than hitherto considered magmatic event in northwest Iberia.

CONCORDANT ANALYSES DISCONCORDANT ANALYSES
As reported here, the ages of the dated intrusions range between ca. 347 and 337 Ma. These ages suggest that the magmatic event that produced this suite of Visean granitoids in northwest Iberia was relatively short lived (ca. 10 Ma) when compared with the time span of the POS in the same geological realm (ca. 310-287 Ma, a duration of >20 m.y.; Gutiérrez-Alonso et al., 2011, and references therein).
What we consider geologically relevant based on our data is that, in devising a geodynamic scenario for the genesis of the ECS in northwest Iberia, it may be worth pursuing the notion that this event was short lived and voluminous (see Fig. 2) and thus could be considered as a sort of magmatic flare-up.
Another relevant datum issuing from this work is the age of the vaugnerite rocks in the Bayo-Vigo intrusion, which is coeval with many vaugnerite and/or durbachite rocks in the Variscan realm of western Europe (e.g., von Raumer et al., 2014). This is important because there are also younger vaugnerite and/or durbachite rocks in the Variscan realm (e.g., Moyen et al., 2017) that would be linked to different petrogenetic-geodynamic scenarios. In this regard, note that in Fernández-Suárez et al. (2000) it was shown that hornblende-bearing peridotites (known as in the regional literature as cortlandites) spatially associated with the ca. 325 Ma Vivero granitoid intrusion (also in the northwest Iberian realm) represent a later event of mantle melting ca. 290 Ma.
Essentially, the ages of sample GP-12 proves that there are mantlederived rocks associated and coeval with the ECS of granitoids, and that they are linked to their petrogenetic-geodynamic scenarios.
The ages obtained for the vaugnerite (sample GP-12; 339 ± 1 Ma) and the main granodiorite facies (sample GP-11; 337 ± 1 Ma) of the Bayo-Vigo intrusion (Fig. 2) overlap within error, albeit the mafic rocks yield a concordia age ~2 m.y. older than the granodiorites. These ages indicate that although essentially coeval, the crystallization of mafic magmas took place slightly earlier than that of the main granodiorite unit. This observation points to a slight age difference in the spatially related mafic and granite rocks and is consistent with Gallastegui's (2005) interpretation (based on field, mineralogical, geochemical, and isotope data) that the mafic and granitoid magmas, although physically mingled, had essentially separate chemical evolutions. This is also in agreement with some of the current views on the origin of the chemical diversity in granitoids that place little emphasis on actual magma mixing as a source of chemical variation (e.g., Clemens and Stevens, 2012).
Based on the information reported, and on previous studies in the area, a hypothesis for the generation of the ECS of northwest Iberia needs to consider the following.
1. The ECS of northwest Iberia represents an apparently short lived event with generation of a significant volume of granitoid magma in a relatively short period of time (~5-10 m.y.). It is convenient to highlight here that the apparent short duration of the ECS event contrasts with the longer time span of the POS in the same northwest Iberian realm in which there seems to have been a protracted time span of granitoid production between ca. 310 and 287 Ma (although mostly between 305 and 290; Gutiérrez-Alonso et al., 2011).
2. The Visean magmatic event reported herein generated mainly peraluminous crustal granitoids that recycled mostly the early Paleozoic and late Ediacaran basement rocks of northwest Iberia. Melt temperatures were ~700 to >800 °C with an apparent predominance of hot granitoids generated through melting processes that likely involved muscovite and muscovite + biotite breakdown reactions. In the intrusion of Bayo-Vigo, where we have proven the coeval nature of vaugnerites and granitoids, the main granite facies is a hot granitoid (sample GP-11). It is remarkable that ash-fall beds in the Carboniferous strata of the neighboring Cantabrian zone of northwest Iberia dated as ca. 340 Ma (Merino Tomé et al., 2017) indicate that this event also produced volcanic activity, the magnitude of which it is not possible to constrain with available geological evidence, in particular owing to the poor preservation of Visean sedimentary rocks throughout the CIZ.
3. There is some involvement of mantle melting, as attested to by the presence of coeval vaugnerite rocks.
Based on these three points, we must invoke a geodynamic scenario compatible with relatively fast production of significant amounts of crustal granitoid melts and coeval production of minor melts extracted from the underlying subcontinental lithospheric mantle in the broader scenario of the Variscan collision in the northwest Iberian Variscan realm.
It is also important to note that these granitoids appear as elongate intrusions that follow the structural grain of the emplacement region and appear to be absent in the most external parts of the orogen (West Asturian Leonese and Cantabrian zones; Figs. 1 and 2), at variance with the POS that is present across the entire orogen, including the foreland fold and thrust belt (Cuesta and Gallastegui, 2007;Gutiérrez-Alonso et al., 2011).
Several models involving varied and contrasting geodynamic scenarios have been proposed to explain the origin of early Carboniferous Visean magmatic flare-up in Variscan northwest Iberia | RESEARCH magmatism in the European Variscan Belt. We argue that these models are difficult to extrapolate to the northwest Iberian Variscan realm, as they cannot be easily reconciled with available geological observations in the region. Models invoked to explain the origin of this magmatic event in the Variscan Belt include: (1) persistent subduction under the Gondwanan margin until ca. 340 Ma (Janoušek and Holub, 2007); (2) slab break-off of the southward-directed subduction of the Rheic oceanic crust (e.g., Kubínová et al., 2017); (3) mantle wedge metasomatism (Finger et al., 2007); (4) late collisional slab sinking and subduction inversion (von Raumer et al., 2014); (5) relamination, rheological weakening, and radiogenic heat (Schulmann et al., 2014); and (6) strike-slip wrenching (Edel, 2001;Rossi et al., 2009).
All of these models except 5 have in common that magmatism is thought to have occurred in the upper plate, which is a primary inconsistency with geological data available for the northwest Iberian realm, which indicate that all the granitoids dated in this work intruded in the lower plate ( Fig. 10; see following).
Most models proposed for the Variscan evolution of northwest Iberia (e.g., Martínez Catalán et al., 2002Catalán et al., , 2007Ribeiro et al., 2007;Gómez Barreiro et al., 2010;Pastor-Galán et al., 2013; agree that the continental margin of Gondwana underwent westward (in present-day coordinates) directed subduction down to ~80 km (26 kbar in eclogites; Rodríguez et al., 2003;22 kbar in blueschists;López-Carmona et al., 2013   www.gsapubs.org | Volume 10 | Number 2 | LITHOSPHERE was followed by the exhumation of the subducted Gondwanan rocks that were subsequently overthrust onto what is currently known as the parautochthon and autochthon domains (Fig. 10B). Exhumation took place ca. 360-350 Ma (Santos Zalduegui, 1995;Rodríguez et al., 2003;López-Carmona et al., 2014), and final emplacement of the allochthonous units happened through an out-of-sequence thrusting event ca. 350-340 Ma (Dallmeyer et al., 1997;Martínez Catalán et al., 1996Gómez Barreiro et al., 2006López-Carmona et al., 2014). The emplacement of the allochthonous units was coeval with widespread recumbent folding and a subsequent thrusting event in the internal zones of the autochthonous units (Rubio Pascual, 2013;Martínez-Catalán et al., 2014). Given that the studied granitoids mostly intrude the parautochthonous unit around the allochthonous complexes (Fig. 2), it is important to keep in mind the above information in order to propose a tentative scenario for the generation of the ECS in northwest Iberia. Figure 10 illustrates our favored hypothesis for the genesis of the ESC of Iberia and endeavors to take into account the observations presented here, as well as the general framework of the geodynamic scenarios of the Variscan collision in northwest Iberia.
Our starting point is the widely agreed upon scenario in which the most seaward part of the Gondwanan continental platform was over ridden by another continental block (generally considered to have been Laurussia). The Gondwanan lower plate was subducted to a depth of ~70 km (ca. 370-360 Ma; Fig. 10A, see preceding references) before starting its exhumation path. The exhumation of this portion of the subducted Gondwanan passive margin caused its overthrusting onto the Paleozoic sedimentary sequence located in the non-subducted Gondwanan realms (i.e., the parautochthonous and autochthonous CIZ units; Figs. 1 and 2).
Exhumation of the subducted units was accompanied by shortening in the parautochthonous and autochthonous units, recorded as recumbent folding in the overthrusted sedimentary sequence ca. 360-350 Ma (Dallmeyer et al., 1997;Fig. 10B). Together with the superposition of the allochthonous complexes onto the parautochthonous and autochthonous units, the shortening also led to crustal duplexing at lower levels, causing the duplication of the crust and substantial increase in crustal and lithospheric mantle thickness and arguably some lithospheric mantle wedging (Fig. 10B).
Alternating episodes of shortening and extension happened subsequently as deformation progressed toward the foreland (Dallmeyer et al., 1997;Martínez Catalan et al., 2007;Gómez-Barreiro et al., 2006;Rubio-Pascual et al., 2013;. The earliest extensional events are likely to have occurred after the initial thickening in the lower plate and to have been localized in domains where maximum thickness was attained under the allochthonous and parautochthonous units (Fig. 10C).
In our interpretation, the presence of this thickened crust (~70 km) may have been ultimately responsible for the extension that in turn triggered crustal (and minor mantle) melting. Melts thus produced would have migrated upward until their final emplacement at mid-crustal levels, although some of these melts were erupted and preserved as ash-fall deposits (Merino-Tomé et al., 2017) (Fig. 10C). Crustal melts would have formed the granitoid intrusions that we studied and the minor vaugneritic bodies could have been generated by melting of the lithospheric mantle when some of the mantle wedges mentioned herein were decompressed upon exhumation. In this context we tentatively envisage a pattern whereby the higher temperature granitoids might have been generated in melting loci where the mantle was involved (providing extra heat upon melting) and the cold granitoids generated in melting loci where only crust was involved. Exhumation happened at a fast rate (2.0-2.5 cm/yr) consistent with the apparent short time span of granitoid production that our geochronological data suggest (López-Carmona et al., 2014). Therefore, we argue that fast exhumation of a thickened and fertile crust (and some wedged-in lithospheric mantle) led to rapid melt production (magmatic flare-up) that seems to have ended ca. 337 Ma.
Subsequent shortening ca. 335-325 Ma resulted in a rethickening of the crust, followed by extension and the generation of the metamorphic domes (Martínez Catalán et al., 2014, and references therein) where the studied rocks are currently exposed, surrounded in some cases by a younger generation of crustal leucogranitoids formed ca. 320 Ma (López-Moro et al., 2017;Díez-Fernández et al., 2017). The presence of abundant ca. 340 Ma zircon xenocrysts in the younger leucogranitoids (e.g., Martín Garro, 2015;López-Moro et al., 2018) indicate that the ECS granitoids were involved (i.e., acted as protolith rocks) in the generation of the synextensional leucogranitoids ca. 320 Ma.

CONCLUSIONS
The geochronological work carried out in this study proves that the so-called early granodiorites in northwest Iberia represent a distinct, Visean magmatic flare-up in Variscan northwest Iberia | RESEARCH short-lived, relatively voluminous magmatic event that took place in Visean time, between ca. 347 and 337 Ma.
This magmatic event produced mainly crustal peraluminous granitoids with minor associated mafic rocks and volcanics.
The hypothesis proposed to explain the genesis of this early Carboniferous magmatic event in the context of the Variscan collision in the WEVB involves subduction of the Gondwanan crust under Laurussia or another Gondwanan element followed by rapid exhumation of the thickened crust. Fast exhumation led to rapid melt production and a relatively short lived magmatic event (flare-up) recorded in the ECS granitoids that we studied.
The existence of this Visean magmatic event explains the abundance of ca. 340 Ma zircon xenocrysts in younger granitoids of northwest Iberia, leading to the interpretation that the ECS granitoids acted as a source rock in subsequent Variscan crustal melting episodes.