Reorganization of Northern Peri-Gondwanan Terranes at Cambrian–Ordovician Times: Insights from the Detrital Zircon Record of the Ossa-Morena Zone (SW Iberian Massif) Open Access

Orogens often result from the amalgamation of a number of far-traveled continental pieces with different origin and age (e.g., [1]). Identifying and unraveling the evolution of each one of these pieces are usually difficult because a combination of diverse geological, geochronological, and paleomagnetic data is required. Nevertheless, a good characterization of the various terranes accreted in mountain belts is the key to elaborate reliable paleotectonic and paleogeographic reconstructions. In this regard, the number and width of oceanic domains opened at preorogenic times are on many occasions a matter of debate, which leads to contradictory plate reconstructions, especially for pre-Mesozoic times (e.g., [2–5]).

The multidisciplinary approach required to address paleogeographic/paleotectonic reconstructions has incorporated the systematic study of detrital zircon populations as a valuable tool to establish sediment provenance from different primary and/or recycled sources (e.g., [6, 7]). Despite a good number of limitations (e.g., [8–14]), detrital zircon studies have contributed to a better understanding of the preorogenic architecture of Paleozoic orogens (e.g., [15–18]). This is the case of the Late Paleozoic Variscan orogen in Western and Central Europe, in which the paleogeographic attribution of the different zones is mostly based on the comparison of detrital zircon spectra with the putative cratonic source areas (e.g., [19–23]). All the studies on the Variscan orogen support the presence of Gondwana-derived and Avalonia-derived terranes, which would have been separated by a large ocean, namely, the Rheic Ocean, opened at late Cambrian-early Ordovician time (e.g., [24–27]). The intervention of other lesser-scale oceans separating any Gondwanan-derived ribbon continental pieces is still under discussion (e.g., [2, 5]). With or without the intervention of narrow oceanic realms, there are marked differences in the detrital zircon record between Gondwanan-derived zones; some of them systematically contain a 1 Ga population interpreted to be sourced from the Arabian-Nubian Shield and/or the Saharan Metacraton, while others lack this 1 Ga population and have been interpreted as attached to the West African Craton (WAC) during the preorogenic evolution (e.g., [19, 21, 26, 28–31]).

Among the Variscan zones in the Iberian Massif (Figure 1(a)), most of them (Cantabrian Zone, West Asturian-Leonese Zone, and Central Iberian Zone) show detrital zircon spectra akin to Paleozoic rocks directly overlying the Saharan Metacraton [15, 16, 32]. In the southern sector of the Iberian Massif, some of the formations of the South Portuguese Zone (SPZ) show Avalonian affinity [17, 33], thus favoring the location of the Rheic suture along the contact between this zone and the Ossa-Morena Zone (OMZ) [34]. The latter is generally considered as a Gondwana-related continental piece, with or without a narrow Paleozoic oceanic realm in between it and the Central Iberian Zone (CIZ) (e.g., [35, 36]). Furthermore, the OMZ would have been located close to the WAC all along the pre-Variscan evolution, based on the lack of a 1 Ga detrital zircon population [26, 28, 30, 37]. Nevertheless, we recently reported the presence of a noticeable (c. 20%) 1 Ga detrital zircon population in Ordovician and Devonian rocks of the OMZ, thus proposing an eastward displacement of the OMZ (and counterparts through the Cantabrian Arc) related to the vanishing stages of the Cadomian orogeny [23]. In this paper, we present a wealth of U-Pb detrital zircon ages from the complete and very well-dated Cambrian sequence of the OMZ. The results serve to elaborate on the paleogeographic reorganization of northern peri-Gondwanan terranes during the Cambrian-Ordovician transition.

The Variscan belt constitutes a curved orogen that extends from Central and Western Europe to Northwestern Africa. This orogen resulted from the Devonian-Permian collision between Gondwana and Laurussia, with a number of peri-Gondwanan terranes also amalgamated in-between the two main plates (e.g., [5, 38–40]). Among the different transects of the Variscan belt, the Iberian Massif (central and western part of the Iberian Peninsula) is the most complete one and has been traditionally divided into two sectors: (i) a northern one with east vergence towards the Gondwanan foreland and (ii) a southern one with dominant SW vergence towards the Laurussian foreland. The northern sector is made up of three zones: Cantabrian, West Asturian-Leonese, and part of the CIZ [36], which appear overthrusted by the allochthonous ophiolite-bearing Galicia-Tras-os-Montes Zone (GTMZ), interpreted as an unrooted orogenic suture (Figure 1(a)) (e.g., [40]). The southern sector includes part of the CIZ, the OMZ, and the SPZ, with the contacts between the three zones representing orogenic sutures (Figure 1(a)) (e.g., [36, 41]).

The SPZ is considered to be part of the deformed Avalonian foreland, together with other terranes exposed in Southern England, Belgium, and Germany, in the so-called Rheno-Hercynian Zone (e.g., [39, 42–44]). As for the OMZ, it is usually correlated with the Saxo-Thuringian Zone (STZ) cropping out in northernmost Brittany, Belgium, Germany, and Czech Republic (e.g., [40] and references therein).

The OMZ-SPZ contact is usually considered to represent the Rheic Ocean suture [36, 45–49], although it rather appears as a cryptic suture [34] since the MORB-featured amphibolitic unit exposed along this boundary yielded Carboniferous ages [50]. To the north, the OMZ is separated from the CIZ by the Badajoz-Córdoba Shear Zone (BCSZ; Figure 1(b)) [36, 41, 51–55]. The interpretation of this tectonic boundary is controversial, with some authors considering it as a Variscan intracontinental shear zone (e.g., [35, 56–58]), others interpreting it as the root zone of the allochthonous units exposed in NW Iberia (e.g., [42, 59]), and others still interpreting it as being unrooted and part of the ophiolitic units of the GTMZ [60, 61].

The OMZ has specific stratigraphic and magmatic features that have been the base with which to decipher its paleogeographic and paleotectonic meaning. In this regard, the Ediacaran and Lower Paleozoic sequences in the OMZ are distinctive with respect to other Variscan zones in the Iberian Massif (e.g., [62, 63]). Furthermore, the OMZ is also characterized by prominent late Ediacaran and early-middle Cambrian to Ordovician magmatism (Figure 1(b)) (e.g., [64–66]).

The oldest rocks exposed in the OMZ are the slates, schists, and greywackes with minor amphibolite and black quartzite intercalations of the so-called Serie Negra Group (Figures 2 and 3) (e.g., [63]). These rocks were dated at late Ediacaran, based on its stratigraphic position underlying the Cambrian sequence, detrital zircon geochronological data [26, 28, 30, 67, 68], amphibolitic protolith radiometric dating [69], and cross-cutting relationships with Cambrian magmatic rocks [68, 70]. The Serie Negra Group is unconformably overlaid by the uppermost Ediacaran-lowermost Cambrian volcano-sedimentary sequence of the Malcocinado Formation, which is roughly coeval with the emplacement of a number of plutonic bodies [71, 72].

The Lower-middle Cambrian sequence rests unconformably on the Serie Negra Group and the Malcocinado Formation (Figures 2(a) and 3). A more detailed description of the Cambrian sequence is given below.

The Ordovician-Devonian succession is interpreted as a passive margin sequence deposited on the Gondwanan margin (e.g., [73]), coevally with the Rheic and other minor ocean expansion. The Ordovician sequence is made up of slates, siltstones, and sandstones, at times with Katian limestone intercalations (e.g., [62, 74, 75]). The Silurian corresponds to dominant graptolite-bearing black slates, with some chert, sandstone, and limestone intercalations. The Devonian succession consists of slates with siltstone intercalations (e.g., [74, 75]).

The synorogenic deposits are Upper Devonian-Lower Carboniferous greywackes, conglomerates, shales, and limestones, with volcanic intercalations [76–78], although locally Lower Devonian greywackes constitute the very first syncollisional sediments [79–81].

In this study, we have sampled detrital Cambrian rocks from two classical sections with abundant fossiliferous content, namely, the Córdoba and the Zafra sections, whose stratigraphic sequences are described in detail below.

The sedimentary successions in the Córdoba and Zafra areas are equivalent regarding the lower Cambrian but differ in the abundance of volcano-sedimentary rocks of middle Cambrian age.

In the Córdoba area, the Cambrian sedimentary sequence is deformed by an open upright syncline and unconformably overlies metamorphic rocks assigned to the Serie Negra Group (to the north) and vulcanites of the Malcocinado Formation (to the south; Figure 2(a)). The lower Cambrian is composed of conglomerates, sandstones, and shales of the Torreárboles Formation [82], followed by carbonate and siliciclastic alternations of the Pedroche Formation [83, 84] and purple slates (metatuffs?) with minor sandstone and limestone intercalations of the Santo Domingo Formation [85]. The Pedroche formation was dated at series 2 based on its fossiliferous content [86, 87]. Over the Santo Domingo Formation, decametric-thick quarzitic sandstone levels (Castellar Quartzite) grade upwards to an alternation of shales and sandstones (Los Villares Formation) [88], which constitutes the topmost outcropping Cambrian rocks in the Córdoba area (Figures 2(a) and 3(b)).

In the Zafra area, the Cambrian sequence also begins with the Torreárboles Formation, overlaid by the carbonate deposits of the Pedroche Formation (locally known as Alconera Formation) [89], the siliciclastic sediments of the Vallehondo Formation and the Castellar Quartzite. The sequence ends with the slates, sandstones, and metabasalts of the Playón Formation (Figures 2(b) and 3(a)). The Vallehondo Formation [90] includes a lower member mainly made up of slates (equivalent to the Santo Domingo Formation in the Córdoba section; Figure 3) and an upper member characterized by sandy slates with several intercalations of sandstones and volcanic rocks (equivalent to the lowermost Los Villares Formation in the Córdoba section). The upper Miaolingian (late Drumian-Guzhangian) Playón Formation overlies the Vallehondo Formation and consists in a thick sequence of slates and fine-grained sandstones with abundant felsic/mafic volcanic intercalations [91], which attest to the rifting stage that gave way to the opening of the Rheic Ocean [65, 90, 92–94].

This work reports detrital zircon U-Pb geochronological data obtained from 15 upper Ediacaran-middle Cambrian samples collected along the Córdoba and Zafra sections. The samples are listed in Table 1.

In the Córdoba area, samples COR13 and COR14 come from the lower levels of the Torreárboles Formation (Figure 3) outcropping in the Torreárboles hill, which is located in the northern limb of an open syncline affecting the Cambrian sequence (Figure 2(a)). Sandstone levels intercalated in the upper part of the Pedroche Formation (COR1 and COR2) were sampled along the CO-3404 road (Figures 2(a) and 3). Samples COR5 and COR6 correspond to sandstones from the Castellar Quartzite (Figure 3) collected close to the Santo Domingo hermitage. The Los Villares Formation was sampled in the northern limb of the syncline (COR8 and COR9; Figure 3). COR10 was sampled along the N-432a road from the Serie Negra Group (Figure 3).

In the Zafra area, samples ZAF13 and ZAF8 were collected from the Vallehondo Formation (upper member; [90]). Sample ZAF13 is located stratigraphically below the Castellar Quartzite (Figure 3) and was collected at the shore of the Albuera del Castellar Lake (Figure 2(b)); sample ZAF8 was taken in the upper part of the Vallehondo Formation (Figure 3), 1 km east of La Lapa along the EX-320 road (Figure 2(b)). Sample ZAF14 was collected 2 km SE of Zafra in the upper part of the Castellar Quartzite (Figures 2(b) and 3). The rest of the samples from the Zafra area were collected along the EX-320 road in the Serie Negra Group (ZAF1) and Torreárboles Formation (ZAF2 and ZAF3; Figures 2(b) and 3).

Almost all the samples are composed of more than 70% of quartz grains with a grain-size varying from 0.05-0.2 mm (COR1, COR2, COR6, and ZAF8) to 0.5-1 mm (COR5, COR8, COR9, COR13, COR14, ZAF13, and ZAF14). These rocks also include oxides, mica, and some grains of K feldspar and plagioclase. COR10 is strongly foliated; about 20% of the grains (mainly of quartz and oxides) are rounded and surrounded by the foliation. ZAF1 is a greywacke composed of quartz, K feldspar, plagioclase, and lithic clasts, included in a very fine-grained matrix; the clasts have grain-sizes between 0.2 and 1 mm. ZAF2 and ZAF3 contain about 30% of quartz grains with 0.1-1 mm size, surrounded by a fine-grained matrix made up of quartz, ores, and mica.

For each sample, 4-5 kg of rock was collected and processed in the laboratories of the University of Granada (Spain) in order to separate detrital zircon grains. This process included mechanical smashing in a jaw-crusher, sieving, density separation by panning of the 0.3-0.05 mm fraction, magnetic separation, and, finally, handpicking. The zircon grains were then mounted in epoxy discs, polished, cleaned, gold-coated, and analyzed using laser ablation inductively coupled plasma mass spectrometry (LA-ICPMS) at the John de Laeter Centre (JdLC) of the Curtin University (Perth, Australia). More details about the analytical procedure, including standard materials and references to the methodology, are detailed in Supplementary Material 1.

To grant a robust statistical representativeness of the dates, the zircon grains were imaged by cathodoluminescence (CL), once the U-Th-Pb analyses were performed. The CL images were performed at the JdLC using a Mira3 Field Emission SEM instrument. A selection of CL images, with the analyzed spot and the correspondent age, is included in Figures 4 and 5.

Raw data were statistically analyzed using IsoplotR [95] for calculating the mean square weighted deviation (MSWD) of the youngest detrital zircon population (minimum 3 data) and obtaining Multidimensional Scaling (MDS) plots. DensityPlotter 8.4 [96] was used to obtain the Kernel Density Estimators (KDE) and histograms (bandwidth: 30 Ma; bin: 30 Ma). The error associated with the ²⁰⁶Pb/²³⁸U ratio in zircon grains older than 1.5 Ga increases rapidly; therefore, this ratio was used only for ages younger than 1.5 Ga, while the ²⁰⁷Pb/²⁰⁶Pb ratio was used to calculate ages older than 1.5 Ga. Only data with a discordance value lower than 10% and f206 (common Pb) values lower than 1% were considered for the characterization and analysis of the samples. Errors are expressed at 1σ level. The complete tables of results are included in Supplementary Material 2.

(1) Serie Negra Group. COR10 is a quartzitic sandstone from the Serie Negra Group (Table 1). From this sample, 127 analyses were carried out, which yielded 76 concordant data (Figure 6(a)). Zircon grains have a length varying from c. 100 to 250 μm; they are rounded to elongated and present continuous oscillatory and sector zoning (Figure 4). The youngest detrital zircon population is late Ediacaran (⁠$564.8±2.6Ma$⁠, $n=5$⁠, and $MSDW=1.51$⁠) and, hence, coherent with the stratigraphic age of the sample.

The main detrital zircon population found in sample COR10 is late Tonian-Ediacaran. It includes c. 89% of the data (⁠$n=68$⁠; c. 848-555 Ma), with an Ediacaran mean age of $628.9±0.8Ma$ which roughly coincides with the main peak of the population. Two very minor detrital zircon populations have Orosirian-Statherian (3 grains; c. 1941-1889 Ma) and Rhyacian-Orosirian (4 grains; c. 2171-1951 Ma) ages. A metamorphic rim yielded early Ordovician age.

(1) Torreárboles Formation. Two hundred and fifty-nine analyses were carried out on two samples (141 in COR13 and 118 in COR14) yielding 207 concordant results (Figure 6(b)). Zircon grains from these samples are rounded to elongated, with lengths from c. 70 to 200 μm. The internal structures show frequent continuous oscillatory zoning and cores overgrown by rims, as well as a few sector zoning and homogeneous grains (Figure 4) The youngest detrital zircon populations are late Ediacaran in age (COR13: $567.3±3.6Ma$⁠, $n=3$⁠, and $MSWD=0.91$⁠; COR14: $575.7±2.6Ma$⁠, $n=5$⁠, and $MSWD=0.63$⁠).

The predominant detrital zircon population in these samples is late Tonian-Ediacaran (c. 810-557 Ma) and includes 41.5% of the data (⁠$n=86$⁠). The mean age of the population is $629.2±0.7Ma$⁠. Paleoproterozoic zircon grains are also common, and they group into two populations: a late Orosirian-Statherian one (c. 1941-1738 Ma, $n=28$⁠, 13.5% of the data with a mean age of $1905.8±2.8Ma$⁠) and a Rhyacian-early Orosirian one (c. 2216-1974 Ma, $n=55$⁠, 26.6% of the data with a mean age of $2108.2±1.9Ma$⁠). The latter is characterized by two second-order peaks at c. 2055 Ma and 2170 Ma. Nineteen data (9.2%) have Meso- to Neoarchean ages (c. 2925-2604 Ma, clustered around c. 2765 Ma). Scattered grains yielded Paleoarchean (⁠$n=6$⁠), Siderian-Rhyacian (⁠$n=7$⁠), and Ectasian-early Tonian (⁠$n=6$⁠) ages.

(1) Pedroche Formation. A total of 271 analyses were carried out on samples COR1 (138) and COR2 (133) (Table 1), yielding 209 concordant results (Figure 6(c)). Zircon grains have a length varying from c. 50 to 150 μm; they are mainly rounded, and their internal structures can be homogeneous or showing sector and continuous oscillatory zoning and cores overgrown by rims (Figure 4). The youngest detrital zircon populations are $579.0±1.5Ma$ (COR1, $n=16$⁠, and $MSWD=1.35$⁠) and $579.6±1.5Ma$ (COR2, $n=16$⁠, and $MSWD=1.25$⁠).

Forty-nine percent of the data (⁠$n=103$⁠) is included in a late Tonian-Ediacaran (c. 835-543 Ma, mean age of $628.7±0.6Ma$⁠) detrital zircon population, characterized by a main peak at c. 610 Ma and a very minor one at c. 805 Ma. Paleoproterozoic grains define two detrital zircon populations, which include Rhyacian-early Orosirian (c. 2207-1958 Ma, $n=48$⁠, 23.0% of the data) and late Orosirian-Statherian (c. 1944-1799 Ma, $n=24$⁠, 11.5% of the data) dates. The mean ages of these populations are $2071.9±2.1$ and $1894.3±3.2Ma$⁠, respectively, and correspond to the peaks in the histogram and KDE plots (Figure 6(c)). A few data (⁠$n=7$⁠, 3.3%) form a very minor detrital zircon population of Stenian-early Tonian (c. 1103-963 Ma) age. Seventeen grains (8.1% of the data) have Meso to Neoarchean ages (c. 2878-2607 Ma) clustered at c. 2670 and 2830 Ma. Finally, very scarce grains yielded Paleoarchean (⁠$n=1$⁠), late Neoarchean-Siderian (⁠$n=3$⁠), Ectasian (⁠$n=3$⁠), and Tonian (⁠$n=3$⁠) ages.

These samples correspond to the Castellar Quartzite (COR5 and COR6) and Los Villares Formation (COR8 and COR9).

(1) Castellar Quartzite. In samples COR5 and COR6, 295 analyses were carried out (121 and 174, respectively), 205 of which yielded concordant results (Figure 6(d)). Zircon grains from these samples are rounded to elongated with lengths of c. 50-150 μm. Internal structures show frequent continuous oscillatory zoning, rims overgrowing older cores, and some sector zoning (Figure 4). Both samples have late Ediacaran youngest detrital zircon populations (COR5: $559.0±2.3Ma$⁠, $n=7$⁠, and $MSWD=1.18$⁠; COR6: $557.1±3.43Ma$⁠, $n=3$⁠, and $MSWD=0.85$⁠).

The main detrital zircon population includes 143 late Tonian-Ediacaran grains (c. 776-542 Ma, 69.8% of the data) with a peak broadly coinciding with the Ediacaran mean age (⁠$628.9±0.5Ma$⁠). Seventeen percent of the data (⁠$n=34$⁠) have Rhyacian-early Orosirian ages (c. 2213-1962 Ma) clustered in a peak with Orosirian mean age (⁠$2089.7±2.4Ma$⁠). Minor detrital zircon populations yielded late Orosirian-Statherian (⁠$1806.0±6.3Ma$⁠, $n=9$⁠, 4.4% of the data between c. 1915 and 1709 Ma) and Neoarchean-early Siderian (scattered data between c. 2682 and 2439 Ma, $n=9$⁠, 4.4% of the data) ages. Other scattered zircon grains have Mesoarchean (⁠$n=2$⁠), late Siderian-Rhyacian (⁠$n=4$⁠), and late Calymmian-Tonian (⁠$n=4$⁠).

(2) Los Villares Formation. Samples COR8 and COR9 were collected from the upper and lower Los Villares Formation, respectively. Two hundred and sixty-eight analyses were carried out (149 in COR8 and 119 in COR9) and yielded 187 concordant results (Figure 6(e)). The grains are mainly elongated, with lengths of c. 70-150 μm, sector and continuous oscillatory zoning, and cores overgrown by rims (Figure 4). The youngest detrital zircon populations in both samples gave very similar late Ediacaran ages (COR8: $583.0±3.3Ma$⁠, $n=3$⁠, and $MSWD=1.11$⁠; COR9: $601.9±1.65Ma$⁠, $n=15$⁠, and $MSWD=1.3$⁠).

In these samples, the late Tonian-Ediacaran detrital zircon population is very dominant and includes 76.5% of the data (c. 884-574 Ma, $n=143$⁠, mean age of $645.4±0.6Ma$⁠). Minor detrital zircon populations cluster in three peaks at c. 1875, 2055, and 2180 Ma. The youngest of these peaks is defined by 15 data with late Orosirian-Statherian ages (c. 1948-1839 Ma, 8.0% of the data) and a Statherian mean age (⁠$1875.8±4.1Ma$⁠). The other two minor peaks include Rhyacian-early Orosirian data (⁠$n=21$⁠, 11.2% of the data, c. 2229-1969 Ma, mean age of $2127.1±3.0Ma$⁠). Scattered grains yielded Neoarchean-Siderian (⁠$n=7$⁠) and late Stenian (⁠$n=1$⁠) ages.

(1) Serie Negra Goup. Sample ZAF1 was collected from the late Ediacaran Serie Negra Group (Table 1). One hundred and forty analyses were carried out on the same number of detrital zircon grains, yielding 111 concordant ages (Figure 7(a)). Zircon grains are elongated, with lengths of c. 100-150 μm. The internal structures (Figure 5) show some cores overgrown by rims and both sector and continuous oscillatory zoning. The youngest detrital zircon population, composed of 4 data between c. 597 and 589 Ma, gave an Ediacaran mean age (⁠$593.1±1.6Ma$ and $MSDW=1.26$⁠).

The predominant detrital zircon population includes 55 data (49.5% of the data) of late Tonian-Ediacaran (c. 847-557 Ma) age and yields a latest Cryogenian mean age (⁠$640.7±0.5Ma$⁠) with a main Ediacaran peak at c. 630 Ma. The Rhyacian-Orosirian population represents 22.5% of the data (⁠$n=25$⁠; c. 2260-1985 Ma; mean age of $2129.4±2.5Ma$⁠), which appear clustered around two peaks at c. 2210 and 2045 Ma. Minor detrital zircon populations have Orosirian-Statherian (c. 1920-1782 Ma, $n=5$⁠, 4.5%, mean age of $1866.4±6.7Ma$⁠) and Mesoarchean-Siderian (c. 3135-2440 Ma, $n=21$⁠, 18.9%, with two minor peaks at c. 3025 and c. 2580 Ma) ages. A few scattered zircon grains yielded Paleoarchean (⁠$n=1$⁠), Siderian (⁠$n=1$⁠), Statherian (⁠$n=1$⁠), and Tonian (⁠$n=2$⁠) ages.

(1) Torreárboles Formation. Two hundred and eighty analyses (140 in sample ZAF2 and 140 in sample ZAF3) were carried out on detrital zircon grains from these samples, 190 of which yielded concordant results (Figure 7(b)). The zircon grains from these samples have a length from c. 80 to 120 μm, and they are mainly elongated. Their internal structures include continuous oscillatory zoning and cores overgrown by rims (Figure 5). The youngest detrital zircon population of these two samples are Ediacaran (ZAF2; $596.6±1.9Ma$⁠, $n=4$⁠, and $MSWD=1.46$⁠) and Early Cambrian (ZAF3; $526.7±1.6Ma$⁠, $n=3$⁠, and $MSWD=0.61$⁠) in age.

The main peak in the distribution plot (c. 645 Ma) is defined by 40.5% of the data (⁠$n=77$⁠), with late Tonian-Ediacaran ages (c. 814-555 Ma) and a late Cryogenian mean age (⁠$646.7±0.4Ma$⁠). Another important detrital zircon population is the Rhyacian-early Orosirian one, including 25.9% of the data (⁠$n=49$⁠) between c. 2189 and c. 1985 Ma (mean age of $2093.7±1.7Ma$⁠). Sixteen percent of the data (⁠$n=31$⁠) have Mesoarchean-early Siderian ages ranging from c. 3194 to c. 2367 Ma but show a rather dispersed distribution. Two minor detrital zircon populations yielded Orosirian-Statherian (c. 1926-1836 Ma, $n=8$⁠, 4.2% of the data, mean age of $1889.8±6.5Ma$⁠) and Stenian-early Tonian (c. 1113-866 Ma, $n=8$⁠, 4.2% of the data, mean age of $947.2±1.8Ma$⁠) ages. Finally, several scattered grains have Paleoarchean (⁠$n=7$⁠), Siderian-early Rhyacian (⁠$n=3$⁠), Statherian (⁠$n=2$⁠), Ectasian (⁠$n=1$⁠), and early Cambrian (⁠$n=4$⁠) ages.

Among the Miaolingian samples from the Zafra area (Table 1), ZAF8 and ZAF13 were collected from the Vallehondo Formation, while ZAF14 belongs to the Castellar Quartzite, which, in this area, has been described as an intercalation within the Vallehondo Formation (Figures 2(b) and 3; Palacios et al., [90]). ZAF13 was collected below the Castellar Quartzite, while ZAF8 is from above. Nevertheless, the two Vallehondo Formation samples yielded very similar results and are described together.

(1) Castellar Quartzite. One hundred and forty analyses were carried out on detrital zircon grains from sample ZAF14 and yielded 94 concordant results (Figure 7(c)). Zircon grains are rounded to elongated, with lengths of c. 100-150 μm, cores overgrown by rims, and sector and continuous oscillatory zoning (Figure 5). The youngest detrital zircon population includes 6 dates and is late Ediacaran in age (⁠$571.5±2.4Ma$ and $MSWD=0.91$⁠).

The main peak observable in the KDE plot is marked by the Cryogenian-Ediacaran data (c. 712-564 Ma, $n=48$⁠, 51.1% of the data, mean age of $615.4±0.9Ma$⁠). Similarly, to the other samples so far described, ZAF14 is characterized by a late Orosirian-Statherian detrital zircon population (c.1954-1837 Ma, mean age of $1901.9±2.9Ma$⁠), but in this case, it includes up to 26.6% of the data (⁠$n=25$⁠). Minor populations consist of Rhyacian-early Orosirian (c. 2178-2013 Ma, $n=9$⁠, 9.6% of the data, mean age of $2092.5±4.2Ma$⁠) and late Neoarchean-Siderian (c. 2515.0-2358.8 Ma, $n=5$⁠, 5.3% of the data, peak at c. 2410 Ma) dates. Scattered grains have Mesoarchean (⁠$n=1$⁠), Statherian (⁠$n=2$⁠), late Stenian-early Tonian (⁠$n=3$⁠), and early Cambrian (⁠$n=1$⁠) ages.

(2) Vallehondo Formation. A total of 280 analyses were carried out on samples ZAF8 (140) and ZAF13 (140), yielding 210 concordant results (Figure 7(d)). The zircon grains from these samples have lengths of 100-150 μm and mainly elongated shapes. The internal structures show frequent sector zoning and cores overgrown by rims, as well as some continuous oscillatory zoning (Figure 5). The youngest detrital zircon populations in the two samples are late Ediacaran in age (ZAF8: $574.6±1.2Ma$⁠, $n=21$⁠, and $MSWD=0.96$⁠; ZAF13: $575.7±2.2Ma$⁠, $n=7$⁠, and $MSWD=0.96$⁠).

Sixty-four percent of the data (⁠$n=135$⁠) have late Tonian-Ediacaran ages (c. 845-545 Ma) organized in a main peak at c. 595 Ma and with an early Ediacaran mean age (⁠$621.5±0.5Ma$⁠). Most of the remaining data are distributed in 4 minor peaks with Stenian-early Tonian (mean age $1006.8±2.6Ma$⁠, c. 1118-908 Ma, $n=16$⁠, 7.6% of the data), late Orosirian-Statherian (mean age $1862.3±3.3Ma$⁠, c. 1942-1754 Ma, $n=16$⁠, 7.6% of the data), Rhyacian-early Orosirian (mean age $2118.6±1.9Ma$⁠, c. 2190-1965 Ma, $n=26$⁠, 12.4% of the data), and Mesoarchean-Siderian (c. 2819-2447 Ma, $n=8$⁠, 3.8% of the data, peaked at c. 2705 Ma) ages. Scattered zircon grains have Paleoarchean (⁠$n=2$⁠), late Siderian (⁠$n=1$⁠), and Statherian-Stenian (⁠$n=6$⁠) ages.

The youngest detrital zircon populations (YDP) in the studied samples are Ediacaran, except for sample ZAF3 that yielded Terreneuvian age (c. 527 Ma). Leaving aside this sample and those from the late Ediacaran Serie Negra Group, no systematic variation in the YDP is observed along the Cambrian stratigraphic sequence, with all values ranging from c. 557 to c. 602 Ma (Figure 8). YDP are used to constrain maximum depositional ages (MDA) when no true depositional ages (TDA) derived from paleontological dating and/or direct numerical dating of volcanic intercalations are available. This is not the case in the studied samples, which have been accurately dated based on their fossiliferous content [83, 90, 97–99]. Thus, the TDA of the studied samples is generally between c. 35 and 60 Ma younger than MDA, with this difference decreasing towards the bottom of the Cambrian stratigraphic sequence (Figure 8). This trend is well in accordance with the results obtained from Ordovician-Devonian OMZ metasedimentary rocks, with YDP ranging from 582 to 592 Ma and differences between MDA and TDA of 100-180 Ma [23].

According to most authors, the Cambrian-Devonian succession of the OMZ represents a pre- (early Cambrian) and synrift (middle Cambrian) to passive margin (Ordovician-Devonian) transition, coeval to the opening of the Rheic, and other minor peri-Gondwanan oceans (e.g., [36, 65, 73]). The comparison between TDA and MDA supports this tectonic interpretation, with the maximum differences between TDA and MDA occurring during early Devonian time, i.e., at the end of the period of oceanic expansion that preceded the Variscan orogeny. Nevertheless, the prominent presence of lower Cambrian to Lower Ordovician plutonic and volcanic rocks outcropping in the OMZ (e.g., [65] and references therein) compared with the absence of these ages in the detrital zircon populations found in coeval and younger sedimentary rocks is intriguing. This is especially true in the case of the Cambrian rocks of this study, supposedly deposited during the rifting stage in the northern Gondwanan margin, when differential block uplift and rotation might have favored the incorporation of synrift igneous material into coeval or slightly younger sediments. The most plausible explanation for the absence of Cambrian detrital zircon populations (only very few scattered grains were found in the studied samples) is that this magma-rich rift environment was below sea level at that time and was not exhumed to the surface until the collisional and/or postcollisional stages of the Variscan orogeny during Carboniferous-Permian time. Sample ZAF3 from the Torreárboles Formation breaks this rule, since its MDA and TDA are virtually coincident, which probably means that synrift igneous-derived detrital zircons were occasionally incorporated during the deposition of these Terreneuvian detrital sequences of the OMZ.

To sum up, the comparison between MDA and TDA in the Lower Cambrian-Lower Devonian sequence of the OMZ is compatible with a rift to drift transition, in which the difference between TDA and MDA increases upwards from being almost coeval (in a lower Cambrian sample) to c. 180 Ma (Lower Devonian), while the input from intrabasinal-derived rift-related magmatic sources was generally very limited.

The Cambrian samples studied in this work were deposited during a time lapse of c. 40 Ma (c. 540-500 Ma), during the rifting process that finally led to the formation of the northern Gondwanan margin. The dominant detrital zircon population in all the samples is the late Tonian-Ediacaran one (850-550 Ma), which represents between 40 and 90% of the concordant analytical results (Figures 6 and 7). Zircon grains from these ages can be attributed to the prominent long-lived Cadomian magmatic arc and the partially coeval Pan-African orogen, whose materials are dominant in the vast majority of Cambrian to Carboniferous metasedimentary rocks of the Variscan orogen (Figure 9) (e.g., [23, 31, 100–104] and references therein, [105–110]).

The second order detrital zircon populations are Rhyacian-early Orosirian and late Orosirian-Statherian (c. 2200-1950 and 1950-1700 Ma, respectively) in age, together accounting for 8-36% of the concordant ages (Figures 6 and 7). Several samples also contain Mesoarchean-early Siderian grains representing up to 19% of the concordant data (Figures 6 and 7). The commonly invoked source for these detrital zircon populations is the WAC (Eburnean and Leonian-Liberian orogenies, respectively) (e.g., [111] and references therein); although, they are also compatible with a Tuareg Shield origin (see [104] and references therein).

Samples ZAF8 and ZAF13 slightly depart from the above-mentioned pattern, with the minor presence of a Stenian-early Tonian detrital zircon population peaked at c. 1 Ga, which represents c. 8% of the concordant results (Figure 7(d)). Similar results to the ones obtained in these two samples were reported from Cambrian-Lower Devonian rocks of the Moroccan Variscides, where the c. 1 Ga subordinate detrital zircon population has been attributed to intermittent and distant sources located at the Saharan Metacraton [100, 102, 103, 107, 108].

Previous geochronological data on detrital zircon grains from Cambrian OMZ rocks (e.g., [26, 30, 68, 110]) yielded results totally coincident with the ones reported here, i.e., a main population of late Tonian-Ediacaran (850-550 Ma; Figure 8) ages, with minor peaks centered at c. 2 Ga, and lack of Stenian-early Tonian detrital zircon grains (except for very minor and local occurrences). These data, combined with the results from samples of the Serie Negra Group, have been used to locate the OMZ close to the WAC during Ediacaran-Cambrian times, in a western position with respect to the other Variscan zones [26, 28, 30, 68]. Nevertheless, Solís-Alulima et al. [110] proposed that the Sierra Albarrana Domain, located in the northern part of the OMZ and made up of poorly dated Cambrian rocks, is akin to the CIZ and was placed close to the Saharan Metacraton at Cambrian times. This proposition is difficult to reconcile with the virtual absence of c. 1 Ga detrital zircon grains in the Sierra Albarrana rocks, since the presence—not the absence—of a noticeable peak of that age is taken as diagnostic of a Saharan Metacraton provenance (e.g., [15]).

The rift to drift transition in the OMZ deduced from the stratigraphic and magmatic records (e.g., [65, 73, 112]) is also reflected in the available detrital zircon spectra. In this respect, the OMZ middle Ordovician-Lower Devonian rocks contain an important Stenian-early Tonian detrital zircon population (up to 27% of the concordant data; Figures 8 and 9) [23] and, hence, strongly contrast with the Cambrian rocks reported here and in previous works (Figure 10), which have only sporadic Stenian-early Tonian detrital zircon grains, or more frequently show a total absence of these ages. We tentatively attributed the presence of an apparent Stenian-early Tonian detrital zircon population in Ordovician-Lower Devonian OMZ rocks to an Ediacaran-earliest Cambrian eastward translation of this continental piece during the vanishing stages of the Cadomian orogeny, i.e., from an Ediacaran location close to the WAC to a Cambrian position close to the Saharan Metacraton [23].

There is an almost consensus in primarily (Ediacaran time) locating the OMZ close to the WAC [26, 28, 30, 68]. Furthermore, the putative correlation of the OMZ in Iberia with the STZ in Central Europe, which is widely accepted (e.g., [38, 39, 40, 113, 114]), reinforces this paleogeographic attribution. Thus, the OMZ-STZ would constitute an elongated continental terrane (Figure 11(a)), with a c. 2 Ga Paleoproterozoic WAC basement which was dredged in the Galicia Bank [115, 116] and crops out locally in the Cherbourg-Trégor region of NW France [117–119] and the STZ (e.g., [40] and references therein). Finally, the Sm-Nd isotopic signature of Ediacaran rocks from the OMZ and STZ also supports the affinity with the Paleoproterozoic WAC [120, 121].

Given that the available detrital zircon data cover now the whole OMZ preorogenic stratigraphic sequence (Upper Ediacaran-Lower Devonian; Figures 8 and 9), we will try in the next paragraphs to refine our initial hypothesis for the incorporation of a noticeable 1 Ga detrital zircon population at Ordovician time.

In this regard, a first issue to consider is the possibility of intermediate sediment repositories (ISR) [122] for the late Stenian-early Tonian detrital zircon grains, which would prevent from stablishing a direct connection with the primary source. Nevertheless, the lack of a significant unconformity between the Cambrian and Ordovician-Devonian rocks in the OMZ undermines this possibility. Furthermore, the occurrence of a minor 1 Ga detrital zircon population is only occasional in Cambrian OMZ rocks, as well as in coeval rocks from other areas located close to the WAC (e.g., Anti-Atlas). Consequently, these rocks cannot be considered an ISR for the Stenian-early Tonian detrital zircon grains found in the Ordovician-Devonian OMZ sequence (up to 27%), since sediment recycling should maintain the detrital zircon content in the primary/intermediate source [122] but not selectively augment the content of one of the populations. Therefore, a direct connection with a primary 1 Ga source seems justified by the high percentage of this population in the Ordovician-Devonian OMZ rocks. Furthermore, an MDS plot of the available Ediacaran-Devonian OMZ samples also supports this fundamental difference between pre-Ordovician and Ordovician-Devonian OMZ rocks (Figure 10). Thus, Ediacaran-Cambrian samples cluster in two groups characterized by a slightly different percentage distribution of the major populations. In contrast, Ordovician-Devonian samples cluster in a clearly separated third group, reflecting the presence of an increasing number (≥20% of the data) of Stenian-early Tonian dates, absent or very scarce in older samples. We can conclude that the most plausible explanation for this contrast in the Stenian-early Tonian detrital zircon content is related to a change in sediment provenance from a WAC source (Cambrian) to a Saharan Metacraton one (Ordovician-Devonian), where these zircon grains were available from a primary source located in the present-day interior of Africa (e.g., [15]).

A second issue to be addressed is whether the Cambrian-Ordovician change in the detrital zircon record of the OMZ is due to a reorganization of the drainage systems and/or to an eastward displacement of the OMZ-STZ terrane. This Cambrian-Ordovician change is coeval to the rift-drift transition culminating in the formation of the northern Gondwanan margin, which, in turn, might have favored sediment input from the Saharan Metacraton. Thus, relief decay associated with passive margin development may have destroyed putative topographic barriers (remains of the Cadomian magmatic arc and Pan-African belt) that impeded sediment supply from the east (Saharan Metacraton). Furthermore, drainage system expansion during this stage is also plausible as attested by the eastern Gondwana super-fan system [15], which might have enlarged its depositional area since Ordovician time. If this was the case, the Cambrian-Ordovician change in the detrital zircon content of OMZ rocks would be simply due to a shift in the sedimentary systems with the addition of an eastern Gondwanan source to the sediment supply.

Alternatively, the change in the OMZ Cambrian-Ordovician detrital zircon record might be due to a tectonically driven eastward displacement that brought this terrane to a closer position with respect to other zones of the Iberian Massif (CIZ, WALZ, and CZ). In this regard, Azor et al. [23] proposed that the eastward translation of the OMZ occurred during the vanishing stages of the Cadomian orogeny (latest Ediacaran-earliest Cambrian). Nevertheless, the data provided here constrains the change in the detrital zircon record at late Cambrian to early Ordovician time, coeval with the rift-to-drift transition and not with the subduction-to-rift one as initially claimed [23].

To sum up, two alternative scenarios can explain the change in the detrital zircon record of the OMZ: (i) a drainage system reorganization favored by the rift to drift transition with no significant translation of the OMZ terrane, or (ii) an eastward displacement of this terrane facilitated by plate tectonic kinematics. In the following paragraphs, we explore an intermediate scenario, which is maybe more likely considering the whole available data.

A first option to be considered is whether the OMZ might have been located close to the WAC all along the Ediacaran-Devonian timespan, or a certain amount of tectonically driven eastward translation to close to the Saharan Metacraton is required. In this regard, the Ediacaran OMZ samples only contain Paleoproterozoic and Neoproterozoic detrital zircon grains, plausibly derived from the WAC and the surrounding Pan-African and Cadomian mountain belts, with no input from the Saharan Metacraton. The Cambrian samples reported here have very similar detrital zircon spectra, though in some of them, a minor 1 Ga population is also present. This can be related to a distant and intermittent Saharan Metacraton source, but the tectonic setting might have caused eastward translation that favored the incorporation of the 1 Ga detrital zircon grains. Despite the lack of reliable geometric and kinematic data on the Cadomian orogeny, the distribution and geochemistry of late Ediacaran rocks in the OMZ were unanimously attributed to a magmatic arc formed over a subduction zone affecting the northern Gondwanan margin at that time [26, 64, 121, 123]. Based on the proposal of Linnemann et al. [26] for the Cadomian subduction kinematics and the direct evidence of Pan-African collisional kinematics [124], Azor et al. [23] suggested a scenario with right-lateral displacement at latest Ediacaran-early Cambrian times (Figure 11(b)), shortly after the interruption of the Pan-African convergence. Thus, the OMZ terrane (and its European counterparts) would have undergone eastward displacement during the vanishing stages of the Cadomian orogeny. This displacement would have translated the OMZ close to the Tuareg Shield (Figure 11(b)), which, in turn, would have provided most of the detrital zircon grains found in the Cambrian rocks reported here. Interestingly, the Tuareg Shield is characterized by the absence of c. 1 Ga zircon grains ([104] and references therein), in accordance with the detrital zircon spectra found in the Cambrian rocks of the OMZ. During early-middle Cambrian time, after this dextral displacement of the OMZ, rift-related sedimentation and magmatism prevailed [65, 125].

The rift-to-drift evolution of the OMZ and other European counterparts occurred in the context of Rheic and other minor ocean opening at late Cambrian-early Ordovician times. In this situation, the reorganization of the drainage systems feeding the OMZ might have been the ultimate responsible for the incorporation of 1 Ga detrital zircon grains coming from the Saharan Metacraton (Figure 11(c)). By admitting an eastward translation of the OMZ during the vanishing stages of the Cadomian orogeny to a paleoposition close to the Tuareg Shield, we do not need to invoke additional tectonic displacement during the rift-to-drift transition, and, hence, we claim for the simplest explanation, i.e., a variation of the drainage system caused the change in the OMZ detrital zircon record with no significant tectonic displacement (Figure 11(c)). However, both boundaries of the OMZ are important tectonic contacts with pre-Variscan and Variscan displacement, and, hence, the putative influence of tectonic displacement cannot be excluded.

• (i)

  Two classical sections of the Ossa-Morena Zone (Córdoba and Zafra; OMZ) with well-exposed and palaeontologically dated Cambrian rocks, have been analyzed to obtain detrital zircon U-Pb geochronological data. All of the studied samples show a dominant late Tonian-Ediacaran population which represents 40-90% of the concordant ages and peaks at c. 600 Ma. Secondary populations include Rhyacian-early Orosirian and late Orosirian-Statherian ages which account for 8-34% of the concordant data and show peaks centered at c. 2100 and 1900 Ma. Mesoarchean-early Siderian ages usually define a minor population representing 3-19% of the concordant data and show peaks between c. 2450 and 3050 Ma. Most of the studied samples lack a Stenian-early Tonian population, except for two samples that show a minor peak (3-8% of the data) at c. 1000 Ma

• (ii)

  The youngest detrital zircon population in most of the Cambrian rocks of the OMZ is late Ediacaran (c. 557-602 Ma), with no systematic variation along the stratigraphic sequence. By taking these ages as maximum depositional ages (MDA) and comparing them with true depositional ages (TDA) of the Cambrian rocks of the OMZ, it can be inferred that the TDA of the studied samples is about 35-60 Ma younger than the MDA, with this gap increasing towards the top of the stratigraphic sequence. Only one sample yielded a youngest detrital zircon population with early Cambrian age (i.e., close to the stratigraphic age), which can be attributed to detritus derived from intrabasinal magmatic activity during the rifting that affected the northern Gondwana margin at Cambrian time

• (iii)

  The detrital zircon contents found in the Cambrian rocks of the OMZ suggest derivation from the West African Craton and/or the Tuareg Shield and hence can serve to locate this terrane in a western position, with respect to the other Iberian Variscan zones at the onset of the rifting stage that opened the Rheic Ocean and drifted Avalonia from Gondwana

• (iv)

  The detrital zircon spectra of the Cambrian OMZ rocks have a marked difference with middle Ordovician-Lower Devonian ones. The latter contain a noticeable Stenian-early Tonian detrital zircon population with a peak at c. 1000 Ma, while the former lack this population, or it only represents a very minor population. This difference is explained in terms of a reorganization of the drainage systems during the rift-to-drift transition occurred at late Cambrian-early Ordovician time, which would have favored the incorporation of detrital zircon grains derived from the Saharan Metacraton

• (v)

  Previously, the latest Ediacaran-earliest Cambrian evolution of the OMZ involved significant eastward displacement from an initial position close to the West African Craton to a final one close to the Tuareg Shield. This eastward translation occurred during the vanishing stages of the Cadomian orogeny, which might have changed the plate kinematic scenario from subduction to right-lateral shearing

All data supporting the results of our study can be found in the figures, tables, and supplementary material submitted to the journal.

The authors declare that there is no conflict of interest regarding the publication of this paper.

This study was funded by the Consejería de Transformación Económica, Industria, Conocimiento y Universidades de la Junta de Andalucía through grants (P20_00063 and PY20-01387) and the Ministerio de Ciencia e Innovación of the Spanish government through grants (PID2020-118822GB-I00 and PID2020-119651RB-I00). Comments and suggestions made by an anonymous reviewer and Dr. Tobias Stephan have greatly contributed to improve the discussion and conclusions of this paper. We also thank Dr. Lorenzo Valetti for revision of the English style. We want to thank specially Noreen Evans, Hugo Olierook, and Bradley McDonald for the geochronological work carried out at the John de Laeter Centre (Curtin University, Perth, Australia).