Chronostratigraphic framework and provenance of the Ossa-Morena Zone Carboniferous basins (southwest Iberia)

Carboniferous siliciclastic and silicic magmatic rocks from the Santa Susana–São Cristovão and Cabrela regions contain valuable information regarding the timing of synorogenic processes in SW Iberia. In this region of the Ossa-Morena Zone (OMZ), late Carboniferous terrigenous strata (i.e., the Santa Susana Formation) unconformably overlie early Carboniferous marine siliciclastic deposits alternating with volcanic rocks (i.e., the Toca da Moura volcano-sedimentary complex). Lying below this intra-Carboniferous unconformity, the Toca da Moura volcano-sedimentary complex is intruded and overlain by the Baleizão porphyry. Original sensitive high-resolution ion microprobe (SHRIMP) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) U–Pb zircon are presented in this paper, providing chronostratigraphic and provenance constraints since available geochronological information is scarce and only biostratigraphic ages are currently available for the Santa Susana–São Cristovão region. Our findings and the currently available detrital zircon ages from Paleozoic terranes of SW Iberia (Pulo do Lobo Zone – PLZ – South Portuguese Zone – SPZ – and OMZ) were jointly analyzed using the K–S test and multidimensional scaling (MDS) diagrams to investigate provenance. The marine deposition is constrained to the age range of ca. 335– 331 Ma (Visean) by new U–Pb data for silicic tuffs from the Toca da Moura and Cabrela volcano-sedimentary complexes. The Baleizão porphyry, intrusive in the Toca da Moura volcano-sedimentary complex, yielded a crystallization age of ca. 318 Ma (Bashkirian), providing the minimum age for the overlying intra-Carboniferous unconformity. A comparison of detrital zircon populations from siliciclastic rocks of the Cabrela and Toca de Moura volcano-sedimentary complexes of the OMZ suggests that they are derived from distinct sources more closely associated with the SPZ and PLZ than the OMZ. Above the intra-Carboniferous unconformity, the Santa Susana Formation is the result of the recycling of distinct sources located either on the Laurussian side (SPZ and PLZ) or Gondwanan side (OMZ) of the Rheic suture zone. The best estimate of the crystallization age of a granite cobble which was found in a conglomerate from the Santa Susana Formation yielded ca. 303 Ma (Kasimovian– Gzhelian), representing the maximum depositional age for the terrestrial strata. The intra-Carboniferous unconformity seems to represent a stratigraphic gap of approximately 12– 14 Myr, providing evidence of the rapid post-accretion and collision uplift of the Variscan orogenic belt in SW Iberia (i.e., the OMZ, PLZ, and SPZ). Published by Copernicus Publications on behalf of the European Geosciences Union. 1292 M. F. Pereira et al.: Carboniferous chronostratigraphy of the Ossa-Morena Zone


Introduction
The Variscan orogen that extends from central Europe to Iberia was reworked through discrete Carboniferous sedimentary cycles during the Laurussia-Gondwana convergence, giving rise to the formation of marine and terrestrial basins. In SW Iberia, stratigraphic correlation has been proposed for the Carboniferous synorogenic strata found in the three main tectonostratigraphic divisions of the Variscan orogen: the Ossa-Morena (OMZ), Pulo do Lobo (PLZ), and South Portuguese (SPZ) zones (Quesada and Oliveira, 2019, and references therein).
In this paper, sensitive high-resolution ion microprobe (SHRIMP) and laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) U-Pb analyses were performed on zircon grains from silicic volcanic, subvolcanic, and siliciclastic rocks sampled in the Santa Susana-São Cristovão and Cabrela regions (OMZ; SW Iberia). The aim of this geochronological study is to establish the chronostratigraphic framework of these Carboniferous strata and to discuss their provenance using a statistical approach (Kolmogorov-Smirnov test and multiscaling diagrams). Thus we pay tribute to José Ramón Martínez-Catalán, who devoted part of his career to investigating the Carboniferous synorogenic basins of NW Iberia.
The Toca da Moura volcano-sedimentary complex is mainly composed of pelites (i.e., Xistinhos; Teixeira, 1944;Fig. 3a) and greywackes associated with andesite to rhyolite volcanic rocks (lava flow and tuffs; Fig. 3c-e), andesitic basalt, chert layers (Gonçalves and Carvalhosa, 1984), and a few olistoliths of basalt and limestone. Siliciclastic rocks contain well-preserved in situ palynomorph assemblages of Tournaisian to Visean age and reworked palynomorphs ranging in age from the middle Cambrian to the early Tournaisian (Pereira et al., 2006;Lopes et al., 2014). Based on geochemical information, this volcanism was interpreted by Santos et al. (1987) as deriving from calc-alkaline magma produced in a continental magmatic arc. A stratigraphic correlation was established between the Toca da Moura volcanosedimentary complex and the Cabrela volcano-sedimentary   Gonçalves and Carvalhosa, 1984;Machado et al., 2012). Sampling locations of the Carboniferous sedimentary and igneous rocks used for geochronology are indicated with yellow stars. complex (Pereira et al., 2006), which is located 15 km to the NW in the Évora Massif (Pereira et al., , 2012a (Fig. 1b). The presence of variable-scale soft-sediment structures (i.e., slumps, intraclast conglomerates, and olistoliths) in both complexes indicates gravity-induced instability during marine sedimentation. Detrital zircon ages of a siliciclastic rock from the Cabrela volcano-sedimentary complex interbedded with silicic tuffs (Fig. 3f) are mainly middle to late Devonian (82 %) and early Carboniferous (14 %) and also include a few older grains (sample OM-200 from Pereira et al., 2012a).
The Santa Susana Formation (i.e., Santa Susana Basin; Domingos et al., 1983;Quesada et al., 1990;Oliveira et al., 1991) siliciclastic rocks that outcrop along a narrow, NNW-SSE-trending discontinuous band, which is 0.1-5 km wide and 12 km long, overlie unconformably the Baleizão porphyry and the Toca da Moura volcano-sedimentary complexes ( Fig. 2), forming the geological contact between these stratigraphic units often defined by faults (Gonçalves and Carvalhosa, 1984). The Santa Susana Formation is divided into two members (Machado et al., 2012, and references therein): (i) the lower member is mainly composed of coarsegrained sandstone and conglomerate beds , and these conglomerates include pebbles and cobbles of silicic porphyry, rhyolite, andesite, basalt, granite, felsic tuff, pelite, sandstone, greywacke, quartzite, phyllite, chert, and quartz ( Fig. 4e, f); (ii) the upper member represents a repetitive sequence of alternating beds of pelite and sandstone interbedded with coal seams and a few beds of conglomerate (Fig. 2). These terrestrial deposits were most probably deposited in an alluvial, fluvial-to-fluvial, and/or lacustrine (floodplain lakes and/or abandoned channels with abundant vegetation) system (Machado et al., 2012, and references therein). The plant fossils identified in the siliciclastic rocks of the Santa Susana Formation indicate a Moscovian-Kasimovian biostratigraphic age . Pelitic beds from the upper member include palynomorph assemblages assigned a Kasimovian age (Machado et al., 2012). Palynomorphs ranging in age from the middle Cambrian to the early Moscovian were also found in siliciclastic rocks of the Santa Susana Formation sampled from a borehole at a depth of around 400 m (Lopes et al., 2014). Detrital zircon ages from upper member sandstones (Dinis et al., 2018) are mainly distributed over the Devonian-Carboniferous (41 %-51 %), Paleoproterozoic (23 %-30 %), and Ediacaran-Cryogenian (16 %-23 %) groups, as well as a few Stenian-Tonian and Archean grains.
3 Rational and analytical methods U-Pb geochronology of detrital zircon from siliciclastic rocks has been extensively used in stratigraphic correlation studies for estimating the maximum depositional age and in-vestigating the provenance of sedimentary sequences (Fedo et al., 2003;Dickinson and Gehrels, 2009). The youngest detrital zircon grains found in siliciclastic rock commonly provide useful information about depositional age, especially in areas that experienced active volcanism during sediment accumulation (Gehrels, 2014). The maximum depositional age obtained for siliciclastic rock is often not necessarily coincident with the biostratigraphic age as defined by key fossil assemblages (Pereira et al., 2019). Therefore, in order to overcome any doubt about the true age of deposition, it is desirable that volcanic rocks interstratified with fossiliferous siliciclastic rocks should be dated (Fedo et al., 2003;Bowring et al., 2006). Furthermore, the application of zircon U-Pb geochronology to volcano-sedimentary and sedimentary sequences, which are separated by unconformities by means of the comparative analysis of their age populations, may be useful for estimating time intervals and revealing changes in provenance. Volcanic rocks that lie beneath or overlie sedimentary sequences and unconformities can provide maximum and minimum ages, respectively. When detrital zircon geochronology is linked to the geochronology of crosscutting younger igneous rocks, then both a maximum and minimum age bracket for deposition can be determined (Fedo et al., 2003).
In this study, SHRIMP U-Pb analyses were performed for the first time on magmatic zircon from two tuffs from the Toca da Moura volcano-sedimentary complex (TM-1 and SCV-2; Fig. 3c, d), one tuff from the Cabrela volcanosedimentary complex (CBR-12), one from the Baleizão silicic porphyry (SCV-30; Fig. 3b), and a cobble of granite (SCV-7; Fig. 4e) found in a conglomerate from the lower member of the Santa Susana Formation. Estimations of the crystallization age of samples SCV-2, TM-1, and CBR-12 (syndepositional volcanism) and sample SCV-30 (postdepositional) were used to validate the Tournaisian-Visean biostratigraphic age previously attributed to the Toca da Moura and Cabrela volcano-sedimentary complexes based on palynological assemblages (Pereira et al., 2006;Lopes et al., 2014). The presence of granite cobbles and pebbles in conglomerate layers from the lower Santa Susana Formation indicates denudation and recycling of a crystalline basement involving granite whose age is unknown. The dating of the granite cobble (sample SCV-7) is useful for discussing provenance and estimating the maximum depositional age of the Santa Susana conglomerate. In addition, LA-ICP-MS U-Pb analyses were performed on detrital zircon grains from two samples of sandstone from the upper and lower members of the Santa Susana Formation (samples SS-1 and SS-2, respectively; Fig. 5g, h) and a sample of pelite from the Toca da Moura volcano-sedimentary complex (sample TM-3; Fig. 5e). These new U-Pb data are useful for discussing provenance and determining the maximum depositional ages of the two sedimentary sequences separated by an intra-Carboniferous unconformity. Sample locations in the Santa Susana-São Cristovão region are indi-  Pereira et al., 2012a) were analyzed to test for the existence of pre-Devonian ages. The new U-Pb results obtained in the present study are compared with previously reported age spectra for pre-Kasimovian siliciclastic rocks from the OMZ, PLZ, and SPZ siliciclastic sequences of SW Iberia using statistical tools.
Zircon grains for U-Pb geochronology were selected using traditional techniques: density separation with a Wilfley table (Universidad Complutense de Madrid, Spain) and also a granulometric separation using sieves with a mesh size of less than 500 µm, density (panning) separation procedures, and mineral identification using a binocular lens and prepara-tion of epoxy resin mounts with zircon grains (Universidade de Évora, Portugal). U-Pb measurements were obtained at IBERSIMS (Universidad de Granada, Spain) using SHRIMP and also at the Senckenberg Naturhistorische Sammlungen Dresden (Museum für Mineralogie und Geologie, Germany) using LA-ICP-MS. U-Pb measurements using SHRIMP and LA-ICP-MS followed the procedures previously described by Dias da Silva et al. (2018) and Pereira et al. (2012a), respectively. U-Pb results are listed in Tables S1 and S2 in the Supplement. Concordia curves and weighted average means were obtained using Isoplot 4 (Ludwig, 2003) (Figs. 6 and 7). Kernel density estimation (KDE) diagrams were produced with 90 %-110 % concordant 206 Pb/ 238 U ages for grains younger than 1.0 Ga and 207 Pb/ 206 Pb ages for older grains (for further details, see Frei and Gerdes, 2009) using IsoplotR (e, f) conglomerate with pebbles of rhyolite (R), phyllite (P), felsic tuff (T), and quartzite (Q). (Vermeesch, 2018) (Fig. 8a, b). Cathodoluminescence imaging was performed at TU Bergakademie Freiberg (Germany) and at IBERSIMS.
The K-S test and the multidimensional scaling (MDS) technique were used in conjunction to compare populations of detrital zircon U-Pb ages obtained from the Carboniferous siliciclastic rocks of the Santa Susana-São Cristovão region using a method designed for a recent study of the provenance of Triassic sandstones (Gama et al., 2020, and references therein). The K-S test is a nonparametric statistical tool that has been successfully used for the comparison of two populations of detrital zircon U-Pb ages by evaluating whether they are significantly different, i.e., whether zircon age populations correlate with a similar source or not, regardless of whether they are of different sizes, while including at least 20 measurements (DeGraaff-Surpless et al., 2003). The proba-bility of the observed maximum vertical difference between the cumulative probability curves (D value) being unrelated to age differences between the two detrital zircon populations is given by a P value corresponding to a confidence interval of 95 % (Barbeau et al., 2009;Guynn and Gehrels, 2010) ( Fig. S1 in the Supplement). High P values and low D values indicate that the observed difference between the two detrital zircon populations may be explained by the existence of common sources (Gama et al., 2020, and references therein). K-S analyses were carried out using an Excel spreadsheet as described in Guynn and Gehrels (2010). The MDS technique provides a means for the comparison of samples based on quantified pairwise comparisons of their detrital zircon ages and is extremely useful for visualizing the degree of similarity between samples in two dimensions; i.e., greater distances between samples represent a greater degree of dissim-  ilarity between points on MDS diagrams (Vermeesch, 2013;Spencer et al., 2015;Wissink et al., 2018) (Fig. 9). MDS diagrams were produced using IsoplotR (Vermeesch, 2018).

Volcanic silicic rocks of the Toca da Moura and Cabrela volcano-sedimentary complexes
Sample SCV-2 is a fine-grained banded rhyolitic tuff consisting of quartz and K-feldspar phenocrysts and lithoclasts of varying size and shape (less than 1 mm in diameter) dispersed in an ash matrix (Fig. 5a). Zircon grains appear as stubby to elongated euhedral prisms (50-150 µm in diameter), mostly showing oscillatory concentric zoning growing on distinct cores or as simple crystals. There are some dark inclusions, unzoned patches, and transgressive variable luminescence and embayments. A total of 44 U-Th-Pb SHRIMP analyses of 44 grains yielded U contents ranging from 262 to 628 ppm. A group of 23 grains with 206 Pb/ 238 U ages (discordance ≤ 5 %) yielded a weighted mean 208 Pb/ 238 U age of 331 ± 4 Ma (mean square of weighted deviates; MSWD = 1.2; Fig. 6a), which probably represents the crystallization age of tuff. Sample TM-1 is a fine-grained banded rhyolitic tuff consisting of quartz, K-feldspar, and biotite phenocrysts, flattened dark-brown pumice (i.e., fiamme), and lithoclasts (less than 1 mm in diameter) enclosed in an ash matrix (Fig. 5b). The zircon population is characterized by stubby euhedralto subeuhedral-sized (30-100 µm in diameter) grains. Magmatic grains are either simple with concentric zoning or com-posite showing variable luminescence cores with concentric zoning, no zoning, or banded zoning. These cores are surrounded by overgrowths with concentric zoning and are occasionally diffuse or unzoned. A total of 120 U-Th-Pb LA-ICP-MS analyses yielded U contents ranging from 87 to 4136 ppm. There are 28 206 Pb/ 238 U ages (90 %-110 % of concordance) that yielded a weighted mean 208 Pb/ 238 U age of 341 ± 10 Ma with a very poor fit (MSWD = 6.9; Fig. 6b) as indicated by the scattering of ages along the Concordia curve. A coherent group of 21 grains with 206 Pb/ 238 U ages yielded a weighted mean 208 Pb/ 238 U age of 335 ± 6 Ma (MSWD = 1.5; Fig. 6b), providing the best age estimate for the volcanic rock (Fig. 6b). The youngest zircon grain (ca. 302 Ma) probably experienced Pb loss. The six oldest zircon grains present Paleoproterozoic (ca. 2 Ga), Neoproterozoic (ca. 715 Ma), and Devonian (ca. 395-378 Ma) ages, suggesting inheritance.
Sample CBR-12 is a fine-grained rhyolitic rock in which feldspar and quartz phenocrysts and lithic fragments occur embedded in an altered, very fine-grained matrix of quartz, sericite, and chlorite, including devitrified shards. Zircon (40-150 µm in diameter) appears as stubby to elongated euhedral prisms, mostly showing oscillatory concentric zoning and sometimes being disturbed by inclusions. A few crystals show banded zoning or are diffuse or unzoned. A total of 32 analyses were performed on this silicic volcanic rock yielding U contents ranging from 288 to 2587 ppm. A total of 22 analyses with discordance ≤ 5 % yielded a weighted mean 208 Pb/ 238 Th age of 335±2 Ma (MSWD = 1.2; Fig. 6c). The oldest two grains yielding 206 Pb/ 238 U ages of ca. 389 and 371 Ma are interpreted as representing xenocrysts.

Baleizão porphyry
Sample SCV-30 is a porphyritic rhyodacite-rhyolite consisting of quartz, plagioclase, K-feldspar, biotite, and amphibole phenocryst (less than 3 mm in diameter) embedded in a finegrained silicic matrix (Fig. 5c). The zircon population contains grains (30-140 µm in diameter) from subrounded subhedral to prismatic euhedral. Prisms are equant to moderately elongate showing a simple internal structure characterized by concentric and sector zoning to unzoned. A concentric zoned or unzoned rim surrounds unzoned cores of a few composite grains. A total of 37 U-Th-Pb SHRIMP analyses for sample SCV-30 yielded U contents ranging from 74 to 4290 ppm. For zircon, 25 analyses were obtained with discordance ≤ 5 %, distributed along the concordia curve from ca. 355 to 288 Ma, and yielded a weighted mean 208 Pb/ 238 Th age of 315 ± 6 Ma (MSWD = 12; Fig. 7a). Some of the spread observed could be due to the presence of inheritance. The oldest 10 grains yielding 206 Pb/ 238 U ages of ca. 355-337 Ma probably represent xenocrysts derived from the Toca da Moura volcano-sedimentary complex. The 11 grains in the age range ca. 334-311 Ma yielded a weighted mean 208 Pb/ 238 U age of 318 ± 2 Ma (MSWD = 0.95; Fig. 7a), which is regarded as the best estimate for the crystallization age of subvolcanic silicic rock.

Cobble of granite found in a conglomerate from the Santa Susana Formation
Sample SCV-7 is a cobble (20 cm in diameter) of pinkish medium-grained granite consisting of quartz, alkali feldspar, and biotite (Fig. 5d) Sample TM-3 is a laminated, poorly sorted siltstone with quartz-rich silt layers containing feldspar and tourmaline grains and lithoclasts (Fig. 5e), which are intercalated with darker layers of clay. The zircon population is mostly characterized by stubby to elongated prismatic small grains (less than 100 µm in diameter). It includes simple and composite zircons showing concentric, sector, and banded zoning. Of a total of 82 U-Th-Pb LA-ICP-MS analyses with U contents ranging from 19 to 4630 ppm, 36 zircon grains yield 90 %-110 % concordance. The number of grains of sample TM-1 does not conform to the minimum of 60-100 grains often used in provenance studies (Vermeesch, 2004), and therefore percentages based on the proportions of ages need to be interpreted with caution.  Pereira et al., 2006;Lopes et al., 2014). Sample CBR-11 is a fine-grained poorly to moderately sorted siltstone consisting predominantly of quartz and a few feldspar grains and lithoclasts enclosed in silt-clay-sized particles (Fig. 5g). Most zircon grains are small (less than 100 µm in diameter) and euhedral to subeuhedral. They are simple grains (short, stubby to equant prisms) with oscillatory concentric and banded zoning, and only a few are composite grains with rounded cores. Of a total of 20 U-Th-Pb LA-ICP-MS analyses with U contents ranging from 54 to 1379 ppm, 10 grains yielded 90 %-110 % of concordance. Five grains are Paleozoic (Carboniferous: ca. 359, 351, and 346 Ma; Cambrian: ca. 514 and 511 Ma) and five are Precambrian (Paleoproterozoic: ca. 2.4, 2.1, and 1.8 Ga; Mesoproterozoic: 1 Ga; Neoproterozoic: ca. 603 Ma). By combining our new data with those from sample OM-200 (Pereira et al., 2012a) collected from the same quarry as sample CBR-11, it was found that the detrital zircon population (CB, N = 54; Fig. 8a) is largely dominated by Paleozoic grains (90 %): late to middle Devonian (68 %), early Carboniferous (15 %), Cambrian (4 %), and early Devonian (2 %) grains, being distinct from sample TM-3 described above (Fig. 8a). The number of grains of sample CB (N = 54), despite being larger than that of sample TM-1, does not conform to the minimum of 60-100 grains, and therefore the proportions of ages obtained also need to be interpreted with caution. The youngest zircon population (N = 5), ranging from ca. 353 to 346 Ma, suggests a Tournaisian maximum depositional age which is slightly older than the sedimentary age inferred from biostratigraphic constraints (late Tournaisian to middle to late Visean; Pereira et al., 2006).

Siliciclastic rocks from the Santa Susana Formation
Sample SS-2 represents medium-to coarse-grained, poorly sorted sandstone. It is mainly composed of lithoclasts (siltstone, mudstone, quartzite, phyllite, rhyolite, basalt) and quartz grains but also includes muscovite and feldspar grains (Fig. 5g). The zircon population is mostly characterized by stubby to prismatic and subrounded to subangular grains (120-300 µm in diameter). Morphologically simple and composite grains were found. Cathodoluminescence imaging shows that most zircon grains have concentric oscillatory zoning or irregular zoning, or they are banded or unzoned. In total, 153 U-Th-Pb LA-ICP-MS analyses were performed on detrital zircon grains. They show U contents ranging from 15 to 6158 ppm. A population with 51 grains yielding U-Pb ages with 90 %-110 % concordance (Fig. 8b) Machado et al., 2012;Lopes et al., 2014).

K-S test and MDS analysis: results
We recognize that the representativeness of the detrital zircon grains of samples TM-1 and CB is not the most recommended for provenance analyses. However, we consider it important to present a preliminary comparison of detrital zircon populations of Visean marine siliciclastic rocks from the Toca da Moura and Cabrela volcano-sedimentary complexes because it seems to indicate variability in the sources. A table showing the K-S results (referred to as Fig. S1 throughout the text) can be found in the Supplement.
The K-S test performed on the Santa Susana sandstones shows that the detrital zircon populations of sample SS-2 (lower member) and the SS upper member (i.e., it includes samples StSz2 and StSz4 from Dinis et al., 2018) are not significantly different (all ages: P value = 0.169; pre-Carboniferous ages: P value = 0.879) at the 5 % confidence level (Fig. S1). A comparison of samples SS-1 and SS-2 reveals that they are significantly different (P value ≤ 0.01). Unlike sample SS-2, the sample SS-1 detrital zircon population is significantly different (P value < 0.01) from the SS upper population (Fig. S1), indicating that they are derived from distinct sources. Besides this, sample SS-1 is much closer to that of SS upper (D value = 0.323) and more distant from sample SS-2 (D value = 0.465) with regards to the distance between cumulative probability curves (Fig. 8c).
In Fig. 9a, the MDS diagram produced with all ages shows sample SS-1 adjacent to Cabrela and Mértola siliciclastic rocks, while sample SS-2 is near the Mira, Santa Iria, and Represa detrital zircon populations. In the MDS diagram for pre-Carboniferous ages, sample SS-2 is juxtaposed with sample TM-3 and closest to the Mira, Phyllite-Quartzite, and Tercenas formations (Fig. 9b), suggesting likely sources. Nevertheless, the probable contribution to SS-2 samples of sediment derived from the oldest siliciclastic rocks from the PLZ, SPZ (i.e., Pulo do Lobo, Gafo, Ribeira de Limas, Atalaia, and Ronquillo formations; P-G-R-A-R), and OMZ sources cannot be excluded. Their detrital zircon populations are not significantly different enough (all ages: P value = 0.003) and not significantly different (pre-Carboniferous ages: P value = 0.113-0.165) at the 5 % confidence level (Fig. S1). This similarity is also illustrated in the approximation between the SS-2, P-G-R-A-R, and OMZ populations in the MDS diagrams (Fig. 9).
K-S test results for the comparison between samples SS-2 and TM-3 indicate that they present not significantly different detrital zircon populations (all ages: P value = 0.399; pre-Carboniferous ages: P value = 0.0.411) at the 5 % confidence level (Fig. S1). Furthermore, their cumulative probability curves are much closer (Fig. 8d): D values are 0.195 (all ages) and 0.203 (pre-Carboniferous ages) (Fig. S1). The close relationship of the two detrital zircon populations suggests that the Toca da Moura volcano-sedimentary complex directly supplied sediment to the Santa Susana Basin. However, the relationship described above does not extend to the entire Santa Susana Basin since sample SS-1 presents a greater degree of similarity with the Cabrela detrital zircon population as regards the proximity between cumulative probability curves (Fig. 8d) and MDS diagrams (Figs. 9a, b).
In addition, Cabrela siliciclastic rocks are significantly different at the 5 % confidence level from sample TM-3 (P values < 0.01) as regards the significant distance between them on the MDS diagram (Fig. 9a, b) and the significant distance between cumulative curves (Fig. 8d) with a D-value range of 0.712-0.731 (Fig. S1). The difference found in the detrital zircon populations suggests that Cabrela and Toca da Moura siliciclastic rocks probably derived from different sources.
As a result of the K-S test and MDS analyses, the Horta da Torre Formation is significantly different (Fig. S1) and is clearly separate (Fig. 9a, b) from all the other detrital zircon populations, ruling out the possibility of it being a source for the Toca da Moura and Cabrela volcano-sedimentary complexes or Santa Susana Formation siliciclastic rocks.

Chronostratigraphic framework
The geochronological data presented in the present study provide the basis for the first chronostratigraphic record for the Carboniferous basins of the Santa Susana-São Cristovão region (SW Iberia). Dating of silicic volcanic rocks interbedded in the Toca da Moura and Cabrela volcano-sedimentary complexes constrain the interval of felsic magmatism to having occurred from ca. 335 to 331 Ma (Visean; Fig. 6), complementing currently available biostratigraphic information for Toca da Moura and Cabrela siliciclastic rocks (Pereira et al., 2006;Lopes et al., 2014). U-Pb ages of the youngest detrital zircon grains from the siliciclastic rocks of the Toca da Moura and Cabrela volcano-sedimentary complexes (TM-3 and CB, respectively; Fig. 8a) provide maximum age constraints for these marine deposits. Their maximum depositional ages (ca. 351-348 Ma; Tournaisian) are slightly older than currently available biostratigraphic ages (Pereira et al., 2006;Lopes et al., 2014) but provide confirmation that both marine deposits are broadly contemporaneous. Furthermore, the best estimate of the crystallization age of the Baleizão silicic intrusion provides a minimum age of 318 ± 2 Ma (Bashkirian; Fig. 7a) for the intra-Carboniferous unconformity. Zircon extracted from a pebble of granite found in a Santa Susana conglomerate yielded a crystallization age of ca. 303 Ma for plutonic rock (Fig. 7b). This age estimate overlaps with the age range of ca. 305-303 Ma (i.e., the maximum depositional age range) obtained for the youngest population of detrital zircon grains from sandstone of the upper member (Dinis et al., 2018), complementing the currently available biostratigraphic information for the Santa Susana Formation (Machado et al., 2012;Lopes et al., 2014). Given the findings described above, a stratigraphic interval of approximately 13-17 Myr can be established for the intra-Carboniferous unconformity, marking a change in depositional environment from marine to terrestrial in the OMZ. Basin-drainage and infill patterns most probably changed due to the rapid uplift of the Variscan-Appalachian orogenic belt, active during the waning stages of the Laurussia-Gondwana collision (i.e., late Carboniferous).

Provenance and evolutionary model
An initial important finding provides evidence that they are derived from different sources. The TM-3 population presents 64 % Precambrian detrital zircon grains, while the CB population contains only 10 % (Fig. 8a). Toca da Moura siliciclastic rocks have a greater affinity with the Phyllite-Quartzite, Tercenas, Santa Iria, and Represa formations (Fig. 9), indicating that detrital zircon populations were reproduced faithfully in SPZ and PLZ (Laurussiantype) sources. A contribution from the oldest siliciclastic sequences of PLZ (Pulo do Lobo, Atalaia, Gafo, and Ribeira de Limas formations) and OMZ (Gondwanan-type) sources cannot be ruled out for sample TM-3 (Fig. 9). The number of late to middle Devonian zircon grains in sample TM-3 (6 %) is smaller than that of the CB population (68 %) (Fig. 8a), suggesting that Cabrela siliciclastic rocks were most likely derived largely from a Devonian source consistent with a limited contribution from recycled ancient rocks (Pereira et al., 2012a). This indicates that the origin of the Visean Toca da Moura and Cabrela basins is most likely more closely linked to sources located in the SPZ and PLZ (Laurussian type) than in the OMZ (Gondwanan type). The evidence in the Visean Toca da Moura basin for the dissection of the inactive Devonian magmatic arc and the erosion of its plutonic roots, together with the recycling of the PLZ and SPZ Frasnian-Tournaisian siliciclastic sequences and OMZ basement rocks, differs from the evidence in the Cabrela basin. The significance of the involvement of distinct sources is that part of the region located on the boundary between the OMZ, PLZ, and SPZ (SW Iberia) was subjected to uplift while the remaining part underwent flexural subsidence. A similar tectonic setting has been put forward as an explanation for differences in stratigraphy found in the Pedroches synorogenic Figure 9. Multidimensional scaling diagrams (Vermeesch, 2018) applied to the U-Pb ages (90 %-110 % concordance) of detrital zircon grains from the Toca da Moura (TM-3) and Cabrela (CB) volcano-sedimentary complexes and the Santa Susana Formation (SS1, SS2, SS upper member) and different potential sources: OMZ (Linnemann et al., 2008;Pereira et al., 2008Pereira et al., , 2012c, PLZ (Pereira et al., 2017a;Pérez Cácerez et al., 2017), and SPZ (Braid et al., 2011;Pereira et al., 2012aPereira et al., , 2014aRodrigues et al., 2014). basin located along the OMZ-Central Iberian Zone boundary (Armendáriz et al., 2008, and references therein) (Fig. 1).
Over the past 4 decades, different models have emerged to explain the geodynamic evolution of SW Iberia with the subduction polarity being widely discussed (Quesada et al., 1994;Castro et al., 1996;Ribeiro et al., 2007;Simancas et al., 2009;Braid et al., 2011;Pérez-Cáceres et al., 2015;Díez Fernández et al., 2016;Pereira et al., 2017a). Although in many paleogeographic reconstructions of the Devonian Iberia is flanked by the Rheic and Paleo-Tethys oceans (Stampfli et al., 2002(Stampfli et al., , 2013Cocks and Torsvik, 2006;Stampfli and Kozur, 2006;Torsvik et al., 2012;Arenas et al., 2014;von Raumer et al., 2016), solely the subduction of the Rheic Ocean is considered in the present geodynamic models for Iberia. This geodynamic model of a single ocean has been the trigger of numerous discussions about whether the active magmatic arc was located in Laurussia or Gondwana margins. Our challenging recent proposal considers that SW Iberia's geodynamic evolution could have been linked to the closure of these two oceanic basins (Pereira et al., 2020). The main assumption is that SPZ and PLZ (Laurussian side) and OMZ (Gondwana side) have experienced different and independent geodynamic evolutions before they were juxtaposed with the Variscan sinistral orogen-parallel motion (Pérez-Cáceres et al., 2016). As illustrated in Fig. 10a, the subduction of the Rheic Ocean floor beneath the Laurussian margin during the late Devonian (Pérez-Cáceres et al., 2015;Pereira et al., 2017a, and references therein) caused the onset of the Rheic magmatic arc in the Meguma terrane and related synorogenic basins. A slab rollback mechanism similar to the one that caused the opening of the Okinawa Trough behind the Ryukyu-type subduction in the Pacific Ocean (Yamaji, 2003;Boutelier and Cruden, 2013) could explain the lithosphere extension on the Laurussian side and the late Devonian siliciclastic sedimentation in SPZ and PLZ (Pereira et al., 2017a). The Laurussian active margin was progressively accreted to the Gondwana passive margin facing the Rheic Ocean during the late Devonian (Fig. 10a). A nappe stack was gradually emplaced in the Gondwana margin as a consequence of ongoing continental collision (Pérez-Cáceres et al., 2015;Díez Fernández and Arenas, 2015, and references therein;Díez Fernández et al., 2016) and orogenic gravitational collapse (Dias da Silva et al., 2020). Following the closure of the Rheic, Tournaisian magmatic episodes were associated with lithospheric extension in the Laurussian (Pyrite Belt volcano-sedimentary complex and Gil Márquez pluton) and Gondwanan (Beja Igneous Complex) margins. A mechanism of steepening and break-off of the Rheic Ocean slab beneath the Laurussian margin, as in the modern analog of Eastern Anatolian Alpine-Himalayan collisional mountain belt (Sengör et al., 2003;Keskin, 2007), was possibly the main reason for the early Carboniferous magma generation in the SPZ ; simultaneously the Meguma terrane experienced rapid uplift and terrestrial sedimentation (Fig. 10b1). At the same time, on the Gondwanan side, the upwelling of the asthenosphere could have triggered the partial melting of crustal materials and lithosphere extension (Pereira et al., , 2012b, creating the right conditions for the onset of gneiss domes in the OMZ (Dias da Silva et al., 2018). The emplacement of voluminous subductionrelated magmas (Santos et al., 1990;Castro et al., 1996;Je-sus et al., 2007;Pin et al., 2008;Lima et al., 2012;Pereira et al., , 2015aMoita et al., 2009Moita et al., , 2015, including some with boninitic (Castro et al., 1999) and adakitic (Lima et al., 2013) affinities, was coeval with flexural subsidence, marine sedimentation, and volcanism (Toca da Moura, Cabrela, and Los Pedroches basins) (Fig. 10b2). The early Carboniferous thermal anomaly recorded in the OMZ has been interpreted as resulting from the emplacement in the middle crust of a large volume of mantle plume-related magmas (Simancas et al., 2006). Other studies have suggested that voluminous early Carboniferous magmatism could have resulted from the subduction of an oceanic ridge, creating a slab window beneath the OMZ (Castro et al., 1996(Castro et al., , 1999Díaz-Azpiroz et al., 2006). This model uses the Chile ridge that plunges beneath the South American Plate in Patagonia (Breitsprecher and Thorkelson, 2009) as a modern analog. Our model assumes the subduction of a ridge of the Paleo-Tethys Ocean's lithosphere beneath the Gondwana margin ( Fig. 10b2) instead of the Rheic Ocean's lithosphere as a hypothesis to be further explored since the magmatic activity has extended to the Serpukhovian and Bashkirian in the OMZ (Pavia pluton, Valencia del Ventoso plutonic complex, and Baleizão porphyry). Simultaneous with the putative subduction of the Paleo-Tethys Ocean ridge, other regions of the Variscan-Appalachian orogenic belt, mostly on the Laurussian side (Fig. 10b1), have experienced an oblique collision, rapid uplift, and terrestrial sedimentation (Pereira et al., 2020).
A second significant finding is that detrital zircon populations from the Santa Susana Formation (samples SS-1 and SS-2) also show significant differences (Figs. 8 and 9). Basal conglomerate (sample SS-2) presents a greater percentage of Precambrian grains (64 %) than uppermost sandstone (SS-1 sample; 28 %) and presents a great degree of affinity with the detrital zircon population of sample TM-3. Sample SS-2 presents a great degree of similarity with the detrital zircon populations of overlying SS upper-member sandstones (samples StSz-2 and StSz-4; Dinis et al., 2018) sampled as part of the same stratigraphic profile. SS-2 and SS upper age populations show a great degree of affinity ( Fig. 9), suggesting that detrital zircon grains were mainly derived from the erosion of the Toca da Moura volcanosedimentary complex, the Santa Iria and Represa formations (PLZ), and the Mira Formation (SPZ). However, regarding the detrital zircon grains with pre-Carboniferous ages, additional contributions from other PLZ (Pulo do Lobo, Atalaia, Gafo, and Ribeira de Limas formations), SPZ (Brejeira, Phyllite-Quartzite, Tercenas, and Ronquillo formations), and OMZ sources cannot be ruled out (Fig. 9). The zircon age population of sample SS-1, which is distinct from the SS-2 population, presents a great degree of affinity with the CB population, suggesting lateral changes in sources during the deposition of Santa Susana's uppermost sandstones. The great degree of affinity of the SS-1 with both the Cabrela volcano-sedimentary complex and Mértola Formation detri-tal zircon populations suggests a close association between the two and a common source. Cabrela and Mértola siliciclastic rocks may be regarded as the main source for sample SS-1 and an intermediate sediment repository as they are derived from the erosion of a Devonian source partially represented by the Cercal porphyries from the SPZ. As a result of rapid uplift, the progressive erosion of the Devonian magmatic arc (including its plutonic roots), and of PLZ, SPZ, and OMZ rocks, is evidenced in the Santa Susana Formation. The volumetrically significant contribution of Carboniferous sources to the Santa Susana Basin fill confirms derivation from the erosion of the following: (i) Pyrite Belt volcanic rocks and Phyllite-Quartzite, Tercenas, Mértola, Mira, and Brejeira siliciclastic rocks (SPZ); (ii) the Santa Iria and Represas formations (PLZ); (iii) Gil Marquez granitic rocks and other plutons of the Sierra del Norte Batholith (SPZ and PLZ); (iv) the Beja Igneous Complex, which includes the Baleizão porphyries (OMZ) and Évora and Pavia plutonic and high-grade metamorphic rocks (OMZ); and (v) the Cabrela and Toca da Moura volcano-sedimentary complexes (OMZ) and Mértola turbidites (SPZ). From the late Carboniferous to early Permian, large-scale strike-slip motions have juxtaposed OMZ with PLZ and SPZ (García-Navarro and Fernández, 2004;Pérez-Cáceres et al., 2016) and simultaneously with the rapid uplift of the Variscan orogenic belt (Fig. 10c). In Kasimovian-Gzhelian, sedimentation probably occurred through the opening of the pull-apart terrestrial basin (Santa Susana Basin) related to the movement of major strike-slip faults (i.e., Porto-Tomar fault zone; Pereira et al., 2010;Machado et al., 2012;Gutiérrez-Alonso et al., 2015) during the progressive uplift and buckling of the linear Variscan-Appalachian orogenic belt (i.e., OMZ, PLZ, and SPZ; Fig. 10c), related to a change in the regional stress field that produced the Greater Cantabrian Orocline . U-Pb dating of magmatic zircon extracted from a pebble of granite (ca. 303 Ma; Fig. 7b) found in a conglomerate of the Santa Susana Formation lower member suggests provenance from the direct erosion of Permo-Carboniferous plutons (i.e., original primary source), such as Santa Eulália-Monforte granitic and gabbro-dioritic rocks (OMZ). This ca. 303-297 Ma calc-alkaline plutonic suite is coeval with the Nisa-Albuquerque and Los Pedroches batholiths, located on the OMZ-Central Iberian Zone boundary (Fig. 1), probably representing magmatism related to the eastward migration of the Paleo-Tethyan arc (Permo-Carboniferous Pyrenees plutonic and volcanic rocks; Pereira et al., 2014Pereira et al., , 2015bPereira et al., , 2017b. The Permo-Carboniferous OMZ plutons were emplaced at shallow crustal levels consistent with the low assimilation of country rocks and the sharp contacts, and, therefore, they may have experienced denudation shortly after crystallization without unrealistic uplift rates being required.

Conclusions
The main conclusions of this study are the following.

Visean marine deposition in the Santa Susana-São
Cristovão and Cabrela regions is constrained to the age range of ca. 335-331 Ma by the new U-Pb data for volcanic rocks intercalated within siliciclastic rocks of the Toca da Moura and Cabrela volcano-sedimentary complexes.
2. U-Pb dating of the Baleizão porphyry provides a minimum age of 318 ± 2 Ma (Bashkirian) for the overlying intra-Carboniferous unconformity.
3. Visean siliciclastic rocks from the Cabrela and Toca de Moura volcano-sedimentary complexes are derived from distinct sources, which probably include a Devonian continental magmatic arc, and are likely to be more closely associated with the Laurussian-type sources (SPZ and PLZ) than the Gondwanan-type sources (OMZ).
4. Terrestrial siliciclastic rocks from the Santa Susana Formation are probably the result of the recycling of distinct sources associated with the SPZ, PLZ, and OMZ.
5. The best estimate of crystallization of a granite pebble found in Santa Susana Formation conglomerate suggests a maximum depositional age of ca. 303 Ma (Kasimovian-Gzhelian); together with the youngest U-Pb ages (< ca. 318 Ma) of detrital zircon grains, these findings provide evidence of the denudation of primary crystalline sources during the rapid post-accretion and collision uplift of the Variscan orogenic belt in SW Iberia (i.e., Gondwanan-and Laurussian-type sources).
6. The intra-Carboniferous unconformity that separates the Toca da Moura volcano complex and the Baleizão porphyry from the Santa Susana Formation indicates a notable time interval of approximately 13-17 Myr.
Data availability. The only data related to this publication regard the U-Pb results that are listed in Tables S1 and S2 of the Supplement.
Author contributions. MFP designed the study and, together with CG, ÍDdS, and JBS, performed the geological mapping, sampling, and petrography. The geochronological study was conducted by CG, ÍDdS, MH, and AG and was supervised by MFP and UL. MFP, CG, íDdS, and JBS wrote the paper. Funding acquisition was achieved by MFP, CG, and ÍDdS. All authors have read and agreed to the published version of the paper.
Competing interests. The authors declare that they have no conflict of interest.
Special issue statement. This article is part of the special issue "The Iberian Massif in the frame of the European Variscan Belt". It is not associated with a conference.