Volcanic Stratigraphy, Petrology, Geochemistry and Precise U-Pb Zircon Geochronology of the Late Ediacaran Ouarzazate Group at the Oued Dar’a Caldera: Intracontinental Felsic Super-Eruptions in Association with Continental Flood Basalt Magmatism on the West African Craton (Saghro Massif, Anti-Atlas)

Abstract

The Ouarzazate Group in the Anti-Atlas Belt of southern Morocco, part of the West African Craton (WAC), is a significant Proterozoic lithostratigraphic unit formed during the late Ediacaran period. It includes extensive volcanic rocks associated with the early stages of Iapetus Ocean opening. Zircon U-Pb dating and geochemical analyses of the Oued Dar’a Caldera (ODC) volcanic succession in the Saghro Massif reveal two major eruptive cycles corresponding to the lower and upper Ouarzazate Group. The 1st cycle (588–563 Ma) includes pre- and syn-caldera volcanic succession characterized by basaltic andesite to rhyolitic rocks, formed in a volcanic arc setting through lithospheric mantle-derived mafic magmatism and crustal melting. A major caldera-forming eruption occurred approximately 571–562 Ma, with associated rhyolitic dyke swarms indicating a larger caldera extent than previously known. The 2nd cycle (561–543 Ma) features post-caldera bimodal volcanism, with tholeiitic basalts and intraplate felsic magmas, signaling a shift to continental flood basalts and silicic volcanic systems. The entire volcanic activity spans approximately 23–40 million years. This succession is linked to late Ediacaran intracontinental super-eruptions tied to orogenic collapse and continental extension, likely in association with the Central Iapetus Magmatic Province (CIMP), marking a significant transition in the geodynamic evolution of the WAC.

1. Introduction

Calderas are large collapsed depressions, typically circular in map view, with a diameter many times larger than that of an individual vent. The largest known calderas, with as much as 60 km in diameter, are associated with felsic pyroclastic eruptions that produce ash flow tuffs or ignimbrites having volumes exceeding 1000 km³. In addition to being a potential source of volcanic hazard, calderas and their associated volcanic activity play a key role in the formation of many ore deposits and are sources of geothermal energy. They also provide valuable insights into the generation and evolution of large volumes of felsic magma [1,2,3,4,5]. The evolution of calderas typically follows a progressive sequence marked by significant changes in the composition and physical properties of the erupted magmas. This evolution includes several stages: pre-caldera eruption, caldera-forming (syn-caldera) eruption, post-caldera eruption, caldera resurgence in some cases, and associated hydrothermal activity [1,3,4,6,7,8,9,10]. Caldera-forming eruptions constitute the most dynamic and catastrophic phase [3,6,8]. Before this phase, the magmatic system typically undergoes a prolonged development period, characterized by continuous magma recharge of the magma reservoir, magmatic differentiation resulting in increased silica content, and the formation of pre-caldera lava domes along with smaller-scale eruptions [3,4,6,7,8,9,11,12]. However, the deposits of pre-caldera eruptions are often poorly preserved due to subsequent collapse and infilling (erosion in ancient calderas), which complicates the study of sub-caldera magmatic system evolution and maturation [6,9]. Eroded caldera studies (e.g., [8]) and subsurface data for young examples (e.g., [13]) indicate that poorly sorted, pumice-rich pyroclastic flow deposits (ignimbrites) within many calderas compose significant portions of the total volumes erupted; the intracaldera deposits are typically significantly thicker than ignimbrites that outflowed from the calderas during the same eruptions. In ancient volcanic successions, intracaldera ignimbrite successions are very thick, with at least several hundred meters of thickness, sometimes even greater than 1 km in thickness, such as the Permian Ora caldera in northern Italy [14]. Most very thick intracaldera ignimbrite successions are welded and preserve fiamme welding textures and—if not too deformed—microscopic shard-welding textures. Interbedded with ignimbrites, there might be caldera lacustrine sediments, and close to the margin of the caldera, there could be coarse co-ignimbrite lag and caldera wall-collapse breccias, including some megabreccia blocks. The intracaldera succession might also be cut by dykes, and there could also be lava domes close to the original caldera margin fed by ring dykes, and sometimes inside the caldera. The transition to the outflow ignimbrite succession, if it is preserved, is most readily marked by a much thinner ignimbrite succession that is usually several tens of meters to ~200 m thick and becomes thinner away from the caldera margin, to as little as a few meters thick, or is entirely removed by post-eruption erosion. Outflow ignimbrites vary from welded near the caldera to non-welded further away from the caldera. If there are sediments interbedded with the outflow ignimbrite succession, they are likely to be fluvial in origin. Under the outflow ignimbrite there might be basal fallout deposits, but in ancient successions they could be highly altered and difficult to recognize, or they might be entirely absent if caldera collapse occurred at the same time as the eruption started [15,16].

In the Pan-African Anti-Atlas Belt of Southern Morocco along the northern margin of the West African Craton (WAC) (Figure 1A), Ediacaran magmatic activity is well represented by intrusive bodies, as well as volcanics associated with sedimentary sequences forming the Ouarzazate Group (“PIII” or XIII of [17]; Figure 1B). This magmatic activity has been considered a testimony of the Central Iapetus Magmatic Province (CIMP) large igneous province (LIP) (i.e., the Ouarzazate Event of [18,19,20,21,22,23,24]).

The Ouarzazate Group extends over a vast area of ~2 × 10⁶ km², dominated by pyroclastic ignimbrites and tuffs, with frequent andesite and basalt flows, with an estimated magma volume of ~1 × 10⁶ km³ and volcanic thicknesses of more than 2 km. The Ouarzazate Group (OG) did not experience Pan-African deformation and was deposited on a highly differentiated basement topography. The significant and abrupt variations in thickness over short distances suggest deposition in a post-collisional setting during extensional/transtensional movement [19,20,21,22,23,30,31,32,33,34,35,36,37,38,39,40,41,42]. The Ouarzazate Group magmatism has been linked to the gravitational collapse of the previously overthickened and weakened Pan-African orogenic belt [42]. The Pan-African Orogen collapsed along a low-angle normal fault during the waning stage of the Pan-African Orogeny (ca. 580 Ma to ca. 542 Ma), giving rise to a Basin-and-Range extensional province [43,44], in West Africa, involving unroofing metamorphic core complexes, syn-extensional plutonic bodies, dyke swarms, and volcanic successions [19,20,22,23]. Alternatively, a mantle plume might have been involved in the emplacement of this bimodal LIP with massive felsic magmatism followed by flood basalt volcanism. In that model, the mantle plume is responsible for the melting of the lower crust to produce the large felsic igneous province, followed by partial melting of the plume head that produced the flood basalts [18,20]. This Ediacaran magmatic event triggered intense hydrothermal activity, leading to the formation of several major mineral deposits, including Imiter and Zgounder (Ag-Hg), Bou Azzer (Co-Ni-As-Ag-Au), Iourirn (Au), and Bou Madine (Cu-Pb-Zn-Au-Ag) [22,23,45]. The most complete and well-preserved volcanic successions of the Ouarzazate Group can be found in the Ouarzazate, Siroua, and Bou Azzer regions, where the type of locality of the group was first defined [17].

We herein present volcanic stratigraphy, geochemistry, petrology, and precise U-Pb zircon geochronology for the Ouarzazate Group of the ODC volcanic succession located in the western part of the Saghro Massif (eastern Anti-Atlas, southern Morocco). The study aims to (i) describe the principal lithofacies of the volcanic succession and related intrusive rocks and suggest a facies model for their emplacement, (ii) elucidate their geochemical characteristics and magmatic signatures, and (iii) discuss their petrogenesis, mantle sources and the geodynamic implications of the Ediacaran Ouarzazate Group in western Saghro Massif. Furthermore, we provide a precise U-Pb zircon age for the Ouarzazate Group volcanic pile and place it in the context of the available U-Pb zircon data for the Saghro Massif in the eastern Anti-Atlas (Morocco).

2. Geological Background

The Anti-Atlas Belt in southern Morocco represents the most important segment of the Neoproterozoic, or Pan-African orogenic belt system within the northern part of the West African Craton (WAC) (Figure 1A). Although it experienced minor tectonic overprinting during the Variscan Orogeny [46,47], this mountain range is currently at an altitude exceeding 1 km over large areas due to Cenozoic uplift [48,49]. Therefore, the Anti-Atlas area presents vast exposures of WSW-ENE trending inliers, which consist of erosionally beveled Proterozoic basement rocks that are unconformably covered by little-deformed Ediacaran-Paleozoic sediments.

During the Proterozoic period, the Anti-Atlas belt was affected by two major orogenic events: the first during the Paleoproterozoic (Eburnean cycle) between approximately 2.20 and 2.03 Ga, documented only in the western part of the Anti-Atlas, and the second during the Neoproterozoic (Pan-African cycle) between ca. 885 and 541 Ma (e.g., [17,22,27,41,50,51,52,53]). The Paleoproterozoic basement outcrops exclusively southwest of the Anti-Atlas Major Fault (AAMF; [17]), but it might also underlie the Pan-African Orogen to the northeast [23,52,54,55]. It contains metasedimentary rocks, granites, paragneisses, and migmatites, with U–Pb zircon ages ranging from 2200 to 2030 Ma (e.g., [22,23,32,41,53,56,57,58]), intruded by dolerite dyke swarms with five age clusters: 2040, 1750, 1650, 1416–1380, and 885 Ma [52,59,60,61,62].

The Pan-African Orogen is described as a complex amalgamation of accreted oceanic crust terranes, some with island arcs, and is characterized by three main tectono-magmatic events [17,24,28,51,57,63,64,65]. The Neoproterozoic successions of the Anti-Atlas Belt associated with the Pan-African Orogeny are unconformably covered by the post-collisional, volcano-sedimentary Ouarzazate Group. The latter belongs to the Central Iapetus Magmatic Province (CIMP; [20,22,24,41,59,65,66,67]). The Neoproterozoic successions underlying the Ouarzazate Group are stratigraphically divided into lower units (Lkest-Taghdout, Bou Azzer, and Iriri groups), consisting mainly of basalt, quartz sandstone, and stromatolitic/oolitic carbonates, intruded by dolerite dyke swarms and sill complexes, which record deformation associated with Pan-African orogenic events, and upper units (Saghro and Bou Salda groups), consisting of metasedimentary rocks (turbidites, black shales, and sandstones) that were only affected by the latest stage of the Pan-African Orogeny (e.g., [22,23,27,28,32,38,57,59,68,69,70,71,72,73,74]).

In the eastern Anti-Atlas Belt (Saghro Massif), the Paleoproterozoic basement is not exposed on the surface. The oldest outcropping rocks consist of the Cryogenian turbiditic deposits (shales and greywackes) intercalated with mafic lava flows that were predominantly deposited at the initial stage of the Pan-African Orogeny [75,76]. These turbiditic deposits outcrop from west to east within the Sidi Flah, Qal’at Mgouna, Boumalne, and Imiter areas. They exhibit a low-grade deformation and metamorphose to greenschist facies [75,77]. The Cryogenian turbiditic deposits are unconformably overlain by a thick and widespread Ediacaran volcanic and volcaniclastic succession. Both Cryogenian and Ediacaran units are intruded by Pan-African plutonic bodies and dyke swarms (mafic and felsic). Calderas with large volumes of pyroclastic flow deposits (ignimbrites), and ash falls extensively developed over the large area of the Anti-Atlas are common in the Ouarzazate Group ([24,78] and references therein). In the Jebel Saghro massif, three large calderas have been identified, as described below [28,65,79,80].

The Qal’at Mgouna Caldera, approximately 6 km in diameter (Figure 2), is composed of two structural blocks divided by a southeast-trending (100°) ring fault. The northern block contains a >1500-meter-thick intracaldera succession, the main units of which are thick, welded ash-flow tuffs, lacustrine sedimentary rocks, and andesites. At the base of the intracaldera-filling sequence lies a microgranitic intrusion, interpreted as cogenetic with the volcanic units indicative of resurgent magmatism. The southern block preserves extracaldera sequences ranging from 0 to 500 m in thickness, consisting of fine-grained volcaniclastic sediments (interpreted as moat fill) and ash-flow tuffs, some of which are correlated with the intracaldera sequence [65,79]. The 6 km-wide Tizi n’Tasettift Caldera in the Imiter Subinlier shows similar geologic features to the Qal’at Mgouna caldera (Figure 2).

The ODC (Figure 2 and Figure 3) represents a large, rectangular-shaped volcanic depression infilled with crystal-rich to very crystal-rich, quartz-rich, biotite-bearing, moderately to densely welded ash-flow tuffs of rhyolitic composition. The caldera is approximately 11 × 18 km, elongated in a northeastward direction, and consistent with the dominant regional structure. It is bounded on the northwest and southeast by major strike-slip faults. The ODC formed in a pull-apart graben that developed along a northeast-trending, left-lateral strike-slip fault zone [28]. According to [28] the caldera fill preserved thickness exposed in the steep canyon of the Oued Dar’a River exceeds 500 m. Such an exceptionally thick accumulation of ignimbrites has been observed exclusively in caldera settings [15,16], and is typically associated with ring-fissure eruptions occurring contemporaneously with caldera collapse [1]. Using 500 m as an estimated average thickness of the intracaldera fill, the minimum eruptive volume of volcanic material was calculated to be approximately 100 km³, characteristic of a large caldera system, according to [1]. When the volume of the outflow facies (ash-flow tuff) is considered, the total eruptive volume almost doubles. According to [81], this large (200 km³) volume of ash-flow tuff corresponds to the Volcanic Eruption Index (VEI) ~7 for a single-eruption source of these rocks ([81]; see also Figure 1 and Table 1 in [82]). If we add additional volume inferred to be lost via uplift/erosion and indirectly recorded as polymictic conglomerates with volcanic pebbles at the top of the volcanic pile of the Ouarzazate Group, then the VEI could be as high as ~8.

3. Sampling and Methods

Research has been conducted through fieldwork focused on the geological characterization of the ODC volcanic succession. Lithofacies were identified based on petrographic observations for 150 thin sections, complemented by existing literature data (e.g., [28] and references therein). Additionally, whole-rock geochemical analyses by ICP-MS were performed on twenty samples, including major, trace, and rare earth elements (REE). To constrain the timing of volcanic events, three samples from the intra-caldera crystal-rich (syn-caldera-forming) and the rhyolite at the top (post-caldera), were dated using U–Pb zircon geochronology. The geochemical and geochronological data, together with insights from previous studies (e.g., [28] and references therein), provide a comprehensive framework for understanding the evolution and emplacement of the Oued Dar’a Caldera volcanic succession.

3.1. U–Pb Geochronology: CA-ID-TIMS U-Pb Zircon Dating

Zircon from the three volcanic units was dated employing chemical abrasion isotope dilution thermal ionization mass spectrometry (CA-ID-TIMS) methods at ETH Zurich (samples TIZ41 and TIZ46) and the University of Geneva (UNIGE, sample TIZ45). Zircon crystals were separated using standard techniques: crushing, milling, and sieving to less than 300 µm. This was followed by density separation using a Wilfley table, a Frantz magnetic separator, and methylene iodide heavy-liquid separation. Zircon U-Pb ages were determined according to the standard CA-ID-TIMS technique [84,85]. Individual zircons (euhedral, acicular, or prismatic) were hand-picked, annealed at 900 °C for 48 h, then chemically abraded at 210 °C for 12 h, following the technique of [84]. After removing leachates, the zircons were cleaned and placed overnight in 6N HCl on a hot plate at 80 °C. Afterwards, the zircons underwent four ultrasonic cleanings, each lasting approximately 15 min, alternating between water and weak nitric acid (HNO₃). Each single zircon was dissolved in pre-cleaned Savillex capsule with 4–6 mg of the EARTHTIME ²⁰²Pb-²⁰⁵Pb-²³³U-²³⁵U tracer solution [85,86], using 70 μL HF and traces of HNO₃ at 210 °C for 60h in Parr bombs. The residue was re-dissolved overnight in 25 μl 6N HCl. Samples were then dried down again and redissolved in 80 μL 3N HCl. U and Pb were separated using anion exchange chromatography [87] in 40 μL columns using ultra-pure HCl and H₂O, and finally dried down with 3 µL of 0.02N H₃PO₄. The isotopic analyses were performed on a TRITON Plus (ETH Zurich) and Triton (UNIGE) mass spectrometers equipped with a MasCom discrete dynode electron multiplier and Faraday cups connected to 10¹³ Ω amplifiers. The linearity of the multipliers was calibrated using U500, Sr SRM987, and Pb SRM982 and SRM983 solutions. The determination of the dead time as constant at 22.5 ns was made possible by the observation of up to 1.3 Mcps. The measurement of the Faraday/SEM yields was between 93 and 94%. Isobaric interferences from BaPO²⁺ and Tl⁺ were monitored on masses 201 and 203 and, since no statistically significant signal was observed on the controlled masses, no correction was applied. Lead isotopic fractionation was corrected based on the certified value of ²⁰²Pb/²⁰⁵Pb = 0.99924 ± 0.03%, 1σ of the EARTHTIME ²⁰²Pb-²⁰⁵Pb-²³³U-²³⁵U tracer. The U mass fractionation for the same analyses was calculated using the ²³³U-²³⁵U ratio of the double-spike solution (0.99506 ± 0.01%, 1σ). Both lead and uranium were loaded with 1 μl of silica gel–phosphoric acid mixture [88] on outgassed single Re-filaments. At ETH Zurich, Pb and U isotope ratios were analyzed using static multicollection on the Faraday cups, except for ²⁰⁴Pb, which was analyzed in the axial secondary electron multiplier [89,90]. At UNIGE, all Pb isotopes were analyzed in dynamic mode, using the secondary electron multiplier, while the U (as UO₂) isotopes were measured in static mode using a Faraday cup equipped with 10¹³ Ω amplifiers or, in case of insufficient U beam size, on the electron multiplier. Isobaric interference of ²³³U¹⁸O¹⁶O on ²³⁵U¹⁶O¹⁶O was corrected using an ¹⁸O/¹⁶O ratio of 0.00205±0.9%. The measured uranium isotopic ratios were corrected assuming a sample ²³⁸U/²³⁵U ratio of 137.818 ± 0.045 (2σ; [91]). All common Pb in the zircon analyses was attributed to the procedural blank with the following isotopic composition: ²⁰⁶Pb/²⁰⁴Pb = 17.10 ± 1.2, ²⁰⁷Pb/²⁰⁴Pb = 15.07 ± 0.7, ²⁰⁸Pb/²⁰⁴Pb = 36.17 ± 0.7 (1-sigma%) at UNIGE and ²⁰⁶Pb/²⁰⁴Pb = 18.10 ± 1.6, ²⁰⁷Pb/²⁰⁴Pb = 15.29 ± 1.7, ²⁰⁸Pb/²⁰⁴Pb = 36.95 ± 1.9 (1-sigma %) at ETH Zurich. Uranium blanks were <0.1 pg and a value of 0.05 533pg ± 50% was used for all data reduction. The initial statistics, data reduction, and age calculation were performed using the TRIPOLI (v.4.10.0.3) and Redux (v.3.7.1) software [92]. All ²⁰⁶Pb/²³⁸U and ²⁰⁷Pb/²⁰⁶Pb ratios were corrected for initial disequilibrium in ²³⁰Th/²³⁸U using Th/U [magma] = 3.5 ± 1 (1σ). The accuracy of the measured data was assessed by repeated analysis of the 100 Ma synthetic solution [93], yielding an internal reproducibility in ²⁰⁶Pb/²³⁸U dates better than 0.05%. The 100Ma synthetic solution measured during the period of zircon analyses with EARTHTIME ²⁰²Pb-²⁰⁵Pb-²³³U-²³⁵U tracer yielded a mean ²⁰⁶Pb/U = 100.175 ± 0.012/0.029/0.11 Ma (MSWD = 1.4, n = 8). All uncertainties reported are at the 2-sigma level, following x/y/z systematics of [94], where x = internal uncertainty, y = added systematic uncertainty, and z = added uncertainty of the decay constants. All data are documented in Table S1 in supplementary data, with internal errors only, including counting statistics, uncertainties in correcting for mass discrimination, and the uncertainty in the common (blank) Pb composition. The MSWD value of the weighted mean falls within the acceptable range of values at a 95% confidence level and for n-1 degrees of freedom, as defined by [95].

3.2. Whole-Rock Geochemistry

Whole rock samples were first cut with a diamond-edged rock saw to remove weathered surfaces and then crushed to a fine powder in an agate mill. Powders from twenty samples (typically 0.1g) were weighed out into pre-acid-cleaned Teflon beakers. The powders were then dissolved in two steps, using the following reagents: (1) 1 ml of HF and 4 ml of HNO₃; and (2) 1 ml of HNO₃ and 3 ml of HCl (all acids used for dissolution were concentrated trace metal grade purity). After each step the samples were left closed on a hot plate at 100° C for 24 h and subsequently opened and dried. Following the complete dissolution of the powders, the dried samples were redissolved in 5 ml of 6N HCl. Splits were taken for elemental concentration measurements. For the major, minor, trace, and rare earth element concentration analysis, a split of 10 μl from each sample was evaporated to dryness, diluted 400 times with 5% HNO₃ (v/v), and spiked with 1 ppb Indium. The sample preparation was performed in a PicoTrace clean laboratory, and elemental concentrations were measured at the Yale Geochemistry Center (Yale University) on a Thermo Scientific Element XR (Waltham, MA, USA) inductively coupled plasma mass spectrometer (ICP-MS), using a quartz spray chamber as an introductory system. Values were normalized using routine measurements of USGS geostandards BHVO-2, SBC-1 and SGR-1b. Measurement precision for these standards was generally better than 5% (2σ) for the major, minor, trace, and rare earth elements.

4. Results

4.1. Volcanic Stratigraphy and Petrography of the ODC

The ODC, situated in the western part of the Saghro Massif of the Anti-Atlas (Figure 2 and Figure 3A,B), is one of the eroded and still well-preserved Ediacaran calderas in Morocco. It contains excellent outcrops of volcanic and volcaniclastic rocks associated with the Ouarzazate magmatic event (i.e., Ouarzazate Group or CIMP). The ODC preserves a comprehensive stratigraphic record of different evolutionary stages of the caldera, including the pre-caldera basaltic andesite-andesite, the syn-caldera crystal-rich tuff with variable crystal content (syn-caldera), and the post-caldera basaltic andesite-andesite, rhyolitic lava flows, and ignimbrites (Figure 2, Figure 3A,B and Figure 4). It provides a unique potential to study caldera-fill sedimentation (e.g., caldera-forming eruption deposits) and its implications for the dynamic post-collapse history of the region.

4.1.1. Pre-Caldera Stage

The pre-caldera stage is represented by basal conglomerates and a succession of basaltic andesite and andesitic lava flows associated with volcaniclastic rocks (Figure 4).

• Basal conglomerates

Conglomerates have a variable thickness of up to 300 m and are derived mainly from the reworking of the Saghro Group sediments (Figure 4). The basal conglomerate is preserved in grabens aligned with major fault zones. The well-rounded clasts, ranging from 1 to 30 cm in diameter, are composed of metasediments derived from the Saghro Group (quartzites, metapelites, and graphitic metapelites) as well as granodiorites and volcanic rocks, such as rhyolitic and dacitic tuffs. Ref. [77] placed this unit in the Lower Precambrian PIII, while [96] put it in the Middle PIII. Notably, the Ouarzazate Group corresponds to PIII in the older stratigraphic nomenclature of [17]. The conglomerate unit might mark the regional uplift and extend well beyond the area circumscribed by the caldera outer-ring faults [1].

• Lower basaltic andesite and andesitic lava flows

The lower basaltic andesite and andesitic lava flows are the oldest volcanic units of the ODC. They are associated with volcaniclastic and sedimentary deposits: microconglomerates and volcanic breccias, forming a sequence approximately 320 m thick. This unit outcrops on the northeastern outer slope of the caldera. The lower basaltic andesite and andesite flows exhibit both aphanitic and porphyritic textures. The flows are typically brecciated at the bottom and the top (flow breccias and/or blast breccias), while massive in their middle parts. In thin sections, andesites are porphyritic or microporphyritic, and mainly composed of plagioclase crystals with variable sizes (35%–55%) and pseudomorphs of ferromagnesian minerals (15%–25%) as well as minor amounts of clinopyroxene (5%–10%) and opaque oxides. The goundmass is microcrystalline with highly altered ferromagnesian minerals.

4.1.2. Syn-Caldera Stage

The syn-caldera stage is represented by crystal-rich to very crystal-rich, ash-flow tuffs, including both the intracaldera facies (caldera-forming eruption deposits) and the extracaldera facies (outflow sheets) (Figure 5A,B).

The caldera is filled with a relatively uniform deposit of ash-flow tuff (500m). The tuff is crystal-rich to very crystal-rich, contains quartz (25%–35%) ranging from 0.1 to 1 mm in length, biotite (2%–5%), a small percentage of aphanitic volcanic lithic fragments (10%–15%) and ranges in composition from dacite to rhyolite. In some areas, a well-developed eutaxitic texture is present, with quartz that is generally subhedral to anhedral, showing rounded embayments. Moderately welded, ash-flow tuff has groundmass with quartz, alkali feldspar, and oxides.

North of the caldera, the youngest volcanic unit consists of a moderately welded ash-outflow tuff that is rich in quartz and biotite crystals. This ash-flow tuff resembles the caldera fill deposits and was previously described as the outflow facies of the volcano [28]. The unit is massively bedded and exceeds 350 m in thickness. Similar outflow facies are also observed to the northwest and southeast of the caldera. While mineralogically and compositionally indistinguishable from the caldera-fill, these outflow deposits are notably thinner.

Two samples corresponding to the syn-caldera stage were dated by using the SHRIMP U-Pb zircon method [28]. Sample T4218 from the top of the 400 m-thick unit of crystal-rich and quartz-poor, moderately lithics-rich, moderately to densely welded dacitic ash-flow (outflow) tuff yielded a weighted average age of 571 ± 5 Ma. This unit is characterized by the presence of more than 20% feldspar phenocrysts, with only rare occurrences of quartz and no mafic minerals. It also contains a small percentage of aphanitic, volcanic lithic fragments. A well-developed eutaxitic texture is locally evident. A second geochronology sample (T4216) from the ODC fill represents a welded, rhyodacitic ash-flow tuff. U-Pb data for this sample form two clusters at ca. 575 Ma and ca. 545 Ma. Based on the geological argument that this unit is contemporaneous with the dacitic ash-flow tuff represented by sample T4218, ref. [28] concluded that the seven oldest zircons, yielding the age of 574 ± 7 Ma, most likely reflect the time at which the caldera tuff crystallized.

4.1.3. Post-Caldera Stage

• Upper basaltic andesite and andesitic lava flows

The upper basaltic andesite and andesitic lava flows (Figure 5C), along with associated volcaniclastic and pyroclastic deposits (including flow breccias, blast breccias, and pyroclastic fall deposits), represent, together with the latest rhyolitic lava flows and ignimbrites, the youngest volcanic units in the ODC succession. The andesitic flows are more extensive than the older, lower basaltic andesite and andesitic lava flows and form a sequence approximately 200 m thick. They crop out prominently along the southern outer slopes of the caldera (Figure 5C).

Locally, these units are intruded by andesitic dykes, typically meters thick, and trending north-south to northeast-southwest, which may represent their feeder systems. Notably, the stromatolites of Amane-n’Tourhart, located between the villages of Amane Issougrid and Ait Saoun on Ouarzazate–Agdz Road (Ouarzazate–Alougoum 1:100,000 map), are developed between the post-caldera upper lava flows (Figure 6A,B) ([40,96,97,98,99,100]).

In thin section, the andesitic rocks are typically porphyritic or microporphyritic and are composed primarily of plagioclase, the most abundant phase (60%–70%), ranging from 0.2 to 0.6 mm in length, with a weak normal zoning. Pyroxene phenocrysts (10–15 vol%), with high interference color and occasional minor opaque minerals, ranged in grain size from 0.3 to 0.8 mm, with occasional minor opaque minerals (Figure 7C). The groundmass is composed of a dense mass of plagioclase microlites measuring less than 0.1 mm and microcrystalline opaque oxides.

We interpret the upper basaltic andesite and andesite flows as products of post-caldera volcanic activity, based on their stratigraphic position above the caldera fill; this pattern is consistent with the post-collapse volcanism documented in other caldera systems [1,2,3]. The Amane-n’Tourhart alkaline lake is interpreted here as a post-caldera ephemeral lake rather than the moat fill of the Oued Draa Caldera.

• The caldera moat-filling sequences

The caldera moat-filling sequences crop out in two locations. At the southwestern part of the ODC, an ~6 km-long section of a classic caldera collapse margin is well exposed (Figure 3B). This section comprises approximately 200 m of coarse-grained, quartz-rich, granular volcaniclastic deposits that rest atop the intracaldera-fill tuffs. Interbedded within the volcaniclastic sequence is lacustrine, purple-green carbonate sequence. The entire section was interpreted as a result of moat sedimentation [28]. In the south-central part of the Tizgui 1:50,000 scale map, near the southeastern margin of the caldera at Jbel Tiglagal and Jbel Ayyous, more than 35 m of graben infill consist of mudflows and fine-grained, subaqueous volcaniclastic beds and a thin aphanitic volcanic unit (Figure 5D,E). The sedimentary sequence rests on a thin (1–2 m thick) flow of highly welded, siliceous, cineritic tuffs, and is overlain by 20–50 m thick, brown, crystal-poor rhyolitic lava flows that form the top of the volcanic pile (Figure 5D).

• Post-caldera rhyolites

The post-caldera rhyolites of the ODC magmatic system are exposed as 20–50 meter-thick, brown, crystal-poor lava flows at Jbel Tiglagal and Jbel Ayyous (Figure 5D), and as a 5–250 meter-thick succession of ignimbrites at Jbel Tasgdalt and the four-corner area encompassing the Tiwit, Qal’at Mgouna, Timdghas, and Nqob 1:50,000 scale geological maps [101,102]. Locally, these post-caldera rhyolites are intruded by andesitic dykes.

At Oued Alqantrat, an ~ 8-meter-thick, crystal-rich rhyolite ignimbrite is exposed near the base of the Upper Ouarzazate Group, just above the angular unconformity that separates the upper and lower parts of the Ouarzazate Group. In contrast, a regionally more extensive rhyolitic ignimbrite is found at Jebel Amgroud (Qal’at Mgouna 1:50,000 scale geological map), marking the western edge of a large rhyolitic mass that extends northward and eastward into adjacent map sheets. In the Tiwit 1:50,000 scale geological map, ref. [102] mapped this unit as the Assaka rhyolite ignimbrite. At Jebel Amgroud, the ignimbrite is approximately 5–7 m thick. Although it extends only for approximately 2 kilometers southwest of the dated sample locality [30], it continues for 10 kilometers to the northwest, where it notably thickens, reaching a maximum thickness of up to 250 m.

The post-caldera rhyolitic flows of the Upper Ouarzazate Group, above the regional unconformity, yielded dates of 558 ± 4 Ma at Oued Alqantrat and 556 ± 4 Ma to 555 ± 5Ma at Jebel Amgroud [28]. These ages are close to the 558 ± 5 Ma to 555 ± 5Ma dates for the Isk-n-Alla pink granite [28,103]. They may thus represent subaerial expressions of this intrusion found outside the caldera. The Jebel Amgroud rhyolite is a part of the same mapped unit as the dated Assaka rhyolitic ignimbrite on the Tiwit 1:50,000 map. Although part of the same stratigraphic unit, the ages obtained from the two locations differ slightly, and do not overlap within uncertainty. A slightly younger TIMS U-Pb age of 548 ± 3 Ma was reported [102], based on a sample collected approximately 60 m above a 250-meter-thick section. The age of 556 ± 4 Ma was obtained from the top of the Assaka rhyolite section [28].

In thin section, the rhyolitic lava flows are aphyric to sparsely porphyritic, with a dominant glassy to cryptocrystalline groundmass (~75%–85%), containing only rare crystals of quartz (5%–10%), feldspar (~10%), biotite (<5%), and oxide minerals. No fiamme were observed within the ignimbrites. Instead, micrometric, undeformed glass shards (with X- and Y-shaped morphologies) are dispersed throughout the lava flows (Figure 7D). Additionally, centimeter- to millimeter-sized globular pink fragments are locally scattered within the rock. The 556 ± 4 Ma to 555 ± 5 Ma dated rhyolite at Jebel Amgroud. This crystal-rich rhyolitic ignimbrite, is a maroon to purple color characterized by well-preserved eutaxitic texture. The unit is rich in angular white perthitic feldspar and albite phenocrysts (approximately 20%–40%), as well as angular to resorbed quartz grains. The groundmass includes feldspar and quartz shards, chloritized biotite, flattened pumice fragments, and fiamme. Additionally, miarolitic cavities are present, suggesting a late-stage, volatile activity during crystallization [28].

4.1.4. Intrusive Rocks

The Saghro Massif hosts a variety of Ediacaran intrusive rocks, ranging in composition from granite to gabbro [22,23,28,77,104,105,106]. The main plutonic bodies are located in the central part of the study area, particularly within the Bouskour and Qal’at Mgouna 1:50,000 scale geological maps (Figure 2 and Figure 3). Another suite of Ediacaran hypabyssal intrusions is also present in the Bouskour 1:50,000 scale geological map (Figure 3). Most of these intrusions either crosscut or crystallized contemporaneously with the volcanic units of the Lower Ouarzazate Group (588–571 to 563 Ma), with limited development in the Upper Ouarzazate Group (562 to 548–542 Ma). Two prominent rhyolite dyke swarms extend northward from Bouskour toward the Sidi Flah 1:50,000 scale geological map (Figure 3A). This spatial relationship suggests that the majority of shallow-level intrusive activity occurred prior to the regional unconformity separating the upper and lower parts of the group. Intrusive rocks in the area are generally found as both stock-like plutonic bodies and dykes or dyke swarms.

• Granitoid plutons

The most significant plutonic bodies in the study area were described and dated using the SHRIMP U–Pb zircon method [28]. From oldest to youngest, they include the Zouzmitane microgranite (588 ± 7 Ma; alternative age 588 ± 6 Ma), the Wizergane granodiorite (576 ± 5 Ma; 577 ± 6 Ma), the Assif Tagmout pink microgranite (574 ± 9 Ma; 577 ± 8 Ma), the Bouskour granite (570 ± 5 Ma), the Isk-n-Alla pink granite (559 ± 5 Ma; 558 ± 5 Ma), and the Tagmout gabbro (556 ± 5 Ma; 557 ± 5 Ma) [28].

The Bouskour granite exhibits a range of color (pink, white, and green), and contains quartz (30–40 vol%) with anhedral to subhedral grains, medium to coarse-grained biotite (5–10 vol%), and plagioclase (25–35 vol%) varying in size from 0.3 to 0.8 mm (Figure 5F and Figure 7E). The granite forms a northeast-southwest trending pluton that extends through the northeastern part of the Bouskour 1:50,000 scale geological map and into the south and central parts of the Sidi Flah 1:50,000 scale geological map (Figure 2 and Figure 3A,B; [104,107,108]).

The Isk-n-Alla granite is pink in color and fine- to medium-grained, it outcrops between the Sidi Flah Fault (SFF) and the Tagmout Graben Fault System (TGFS) (Figure 3A) [101]. The Isk-n-Alla pluton is approximately 10–46 km in width (from the Qal’at Mgouna 1:50,000 scale geological map) [101] northeastward into the adjacent Tiwit 1:50,000 map [102]. Ref. [77] initially mapped it as “granite alcalin d’Isk-n-Allah”. The granite crops out on a prominent horst structure, supporting the high peaks of Jebel Saghro, which rise above 2000 m. The main body of the pluton consists of a widespread pink granite, with a distinct red granite facies occurring along its margins. The granite typically forms spheroidally weathered blocks and towering vertical cliffs. It intrudes volcanic rocks of the Lower Ouarzazate Group and is itself cut by small plutons of gabbro and monzogabbro. Rhyolitic and microgranite dykes common elsewhere in the region are rare within the granite, suggesting that the Isk-n-Alla intrusion postdates much of the local felsic dyke emplacement. The pluton has a fine-grained to granophyric texture, with porphyritic quartz (~30%) and feldspar crystals. Porphyritic pyroxene (~15%) and scattered biotite crystals (~5%) are also present, as well as fine-grained, red hornblende along the pluton margin. The pluton contains numerous enclaves and xenoliths of dark, weathered, metamorphosed rhyolitic volcanic rocks. Contact-metamorphic assemblages in these rocks include apatite, rutile, sericite, and epidote.

Dark green to black gabbro occurs as small plutonic bodies and dykes, collectively covering only approximately 2.6 km² across the Bouskour, Sidi Flah, and Qal’at Mgouna 1:50,000 scale geological map sheets (Figure 3A). These gabbro form dark, hilly, massive outcrops with spheroidally weathered, fractured rocks. The largest of these bodies (2.1 km²) is known as the “Tagmout gabbro”. This gabbro pluton is located in the northern part of the TGFS, Qal’at Mgouna 1:50,000 scale geological map (Figure 3A; [101,109]). The gabbro intrudes volcanic rocks of the Lower Ouarzazate Group, Ediacaran granites, and the metasedimentary and metavolcanic units of the Sidi Flah Inlier. Specifically, the Tagmout gabbro cuts across the Isk-n-Alla granite and is itself crosscut by dykes. It is unconformably overlain by rhyolitic tuffs of the Upper Ouarzazate Group. Texturally, the Tagmout gabbro is fine-grained at its margins, coarsening up toward the center, with locally developed quartz-bearing monzogabbro. The gabbro exhibits a diverse and well-preserved petrographic assemblage. It is primarily characterized by euhedral plagioclase phenocrysts (40%–60%), averaging 1.5 mm in size, which are extensively altered to sericite, calcite, and epidote. Biotite, although relatively scarce, occurs as automorphic crystals and is notably rich in mineral inclusions. Plagioclase is extensively saussuritized and commonly exhibits a poikilitic texture, while olivine (<5%) is predominantly serpentinized. The remaining mineral assemblage includes amphibole-clinopyroxene (20%–30%), biotite, apatite, hematite, and rare occurrences of feldspar.

• Dyke Swarms

Numerous cryptocrystalline rhyolitic, microcrystalline microgranitic, cryptocrystalline andesitic, and microcrystalline doleritic dykes that intruded the lower units of the Ouarzazate Group are interpreted as feeder dykes for the overlying volcanic rocks of the upper part of the Ouarzazate Group. Key TIMS and SHRIMP U–Pb zircon dating for the upper part of the Ouarzazate Group include the Tachkakacht rhyolitic dyke in the Sagho Massif (543 ± 9 Ma; [22]), the Assaka rhyolitic ignimbrite (548 ± 3 Ma; [102]), the Takhakert rhyolite (550 ± 3 Ma; [45]), the Bou Madine rhyolitic dome in the Ougnat Massif (552 ± 5 Ma; [22]), and the rhyolite ignimbrite from the Tagragra of Tata (565 ± 7 Ma; [57]).

• Rhyolitic dykes

Rhyolitic dykes are found throughout the region, with the highest concentration occurring in the Jebel Saghro Inlier, Tizgui, and Qal’at Mgouna 1:50,000 scale maps (Figure 3A and Figure 5B). These dykes are broadly grouped into two geographically distinct swarms. The red dykes of the Bouskour Sidi Flah swarm and the white dykes of the Timijt swarm. Both swarms display a left-lateral offset of approximately 3 km along the Sidi Flah Fault (SFF) (Figure 3A).

The Bouskour–Sidi Flah dyke swarm displays a north–south trend and spans a width of approximately 5 km (Figure 3A). Dykes within the swarm are easily identifiable in the Google Earth imagery. The swarm’s widest section extends roughly 25 km, from south of Bouskour to north of the Sidi Flah Fault, where it becomes concealed beneath Neogene deposits. Assuming a conservative depth of 3 km, a width of 0.9 km, and a length of 25 km, the estimated volume of the dyke swarm is approximately 67 km³ [28].

The Bouskour Sidi Flah dykes are mostly aphanitic rhyolites that occasionally contain phenocrysts of quartz showing zoning and characteristic embayments, and fractured and altered plagioclase. The groundmass is composed of cryptocrystalline quartz and alkali feldspars, forming a fine-grained matrix, altered in secondary sericite. These dykes display laminated flow banding that runs parallel to their margins. They outcrop with thicknesses varying from less than 0.5 m to several meters, forming prominent ridges in the landscape and are clearly noticeable in satellite imagery. These dykes intrude on the lower units of the Ouarzazate Group and older formations, but are rare in the upper part of the Ouarzazate Group. They are interpreted as feeder dykes for volcanic units. A sample B1158 from the 15-meter-thick dyke within the Bouskour Sidi Flah swarm, which intrudes the 570 ± 5 Ma Bouskour granite, previously yielded a SHRIMP age of 563 ± 7 Ma [28].

In the western regions of the Bouskour and Tizgui 1:50,000 scale geological maps (Figure 2 and Figure 3), massive maroon to red rhyolite dykes (Figure 5B), roughly 400 m thick, are exposed in the north of the Azwou N Wallows Fault (AWF) and the western border of the ODC. These dykes are interpreted as components of a ring fissure system associated with the caldera’s collapse [28,83,108].

The Timijt dyke swarms are composed of aphanitic rhyolites, light-gray to light-pink in fresh exposures, but light tan to white when weathered. Ranging from 3 to 20 m in thickness, the dykes appear between the northern and central area of the Bouskour 1:50,000 scale geological map and the southwestern area of the Sidi Flah 1:50,000 scale geological map. They intrude both the Cryogenian and the Ediacaran plutonic rocks hosted within the Saghro and Ouarzazate groups (Figure 2 and Figure 3). These dykes display a porphyritic texture, characterized by well-developed quartz phenocrysts (20%–30%) with growth embayments, feldspar phenocrysts (35%) are also present, including both alkaline feldspar and plagioclase (10%). Automorphic brown biotite is scattered throughout the matrix, alongside accessory phases, such as hematite and magnetite. The groundmass is recrystallized and composed of fine quartz, albite, sericite, and epidote crystals.

The dykes were depicted on unpublished maps [110,111], with the latter naming them after an exposure near the Timijit village. [112] had previously mapped the rocks at Timijt as “rhyolites porphyriques.” [113] interpreted white rhyolites as volcanic layers interbedded within the Saghro Group, Cryogenian metasedimentary successions in the Sidi Flah Inlier. [28] correlated the Timijt dykes with more widespread light and colored rhyolitic dykes that trend north to northwest, extending from the Timijt village to Jebel Azariyf, and collectively referred to them as the Timijt dyke swarm (Figure 3A). Sample B1363 from the 10-meter-thick dyke of the Timijt dyke swarm previously yielded a SHRIMP age of 562 ± 5 Ma [28].

• Andesite dykes

The andesite dykes are well exposed in the eastern outcrop of the rhyolitic dykes of Bouskour and Timjjit, as well as in the eastern part of the Saghro Massif in the Imiter and Ougnat Inliers. The andesite dykes crop out over a length of several kilometers and with a thickness ranging from 1 to 2 m, cut through the volcanic formations of the Lower Ouarzazate Group, and trend northeast similar to the trend of the Tagmout graben. These andesitic dykes, dark brown-red to purple in color, are aphanitic and poor in quartz (<5%), with locally abundant plagioclase phenocrysts (50%–60%).

• Dolerite dykes

Dolerite dykes, identifiable in the field by their dark green color and characteristic alteration (Figure 5C,G), vary in thickness from a few centimeters to over two meters and extend up to 5 km in length, cutting through all the previously mentioned formations. They exhibit a porphyritic dolerite texture. The plagioclase is the dominant phase (approximately 50%–60%), occurring as euhedral to subhedral laths (0.5–2 mm), showing distinct albite twinning, zoning, and minor alteration. Clinopyroxene (10%–25%), forms as subhedral to anhedral grains with cleavage and locally developed alteration along grain boundaries. Orthopyroxene (5%–10%) forms elongated grains. Olivine (2%–5%) is present as subhedral grains that exhibit high interference colors and partial alteration. (Figure 7F).

4.2. Zircon U–Pb Geochronology

We dated three samples using the chemical abrasion isotope dilution thermal ionization mass spectrometry (CA-ID-TIMS) method (TIZ41, TIZ46, and TIZ45, all located on the 1:50,000 scale geological map of the Tizgui area) from the ODC volcanic succession, on the NW edge of the Anti-Atlas, Saghro Inlier.

The intracaldera TIZ41 sample corresponds to a dacitic to rhyolitic ash-flow tuff unit of the syn-caldera stage. It belongs to the 1st eruptive cycle or the Lower Ouarzazate Group. Six zircons were analyzed from sample TIZ41. Four out of six are statistically equivalent with a weighted mean ²⁰⁶Pb-²³⁸U age of 563.690±0.051 Ma (MSWD = 2.6) (Figure 8A).

The extracaldera TIZ46 sample is representative of dacitic to rhyolitic outflow of the same syn-caldera stage. It also belongs to the 1st eruptive cycle, or Lower Ouarzazate Group. Five zircons analyzed from sample TIZ46 yielded indistinguishable dates with a weighted mean ²⁰⁶Pb-²³⁸U age of 564.311 ± 0.045 Ma (MSWD = 1.5) (Figure 8B).

The sample TIZ45 corresponds to a rhyolitic ignimbrite located at the top of the volcanic pile above the unconformity (Figure 4 and Figure 5D). This rhyolite is similar in composition to the rhyolite lava flows at the top of the volcanic sequence of the ODC that covers the moat fill sequences of the caldera (carbonate and siliciclastic rocks). It belongs to the 2nd eruptive (post-caldera) cycle or the Upper Ouarzazate Group. Eleven zircons were analyzed from sample TIZ45, ten of which yielded indistinguishable dates with a weighted mean ²⁰⁶Pb-²³⁸U age of 561.61 ± 0.10/0.17/0.63 MSWD = 1.9, n = 10 (Figure 8C, Table S1).

4.3. Whole Rock Geochemistry

The whole-rock major and trace element data, including rare earth elements (REE), for the Ediacaran volcanic succession of the ODC, the intra-caldera crystal-rich (Syn-caldera), rhyolite, andesite and dolerite dykes, upper andesitic lava flows and rhyolite flows (Post-caldera) (western Saghro Massif) are provided in Table S2 of the Supplementary Material. In this study, new data are combined with the geochemical database compiled from literature sources ([83,101,108,109,110,114,115,116] see also [28] for more details). The literature database (Table S3 in the Supplementary Material) is primarily compiled from the geological mapping project at the 1:50,000 scale covering Tizgui, Bouskour, Qal’at Mgouna, Sidi Flah, and Timdghas map sheets. This work was conducted as part of the National Geological Mapping Program of the Geological Survey of Morocco, in collaboration with the United States Geological Survey (USGS) and the National School of Mineral Industry (ENSMR) (G. J. Walsh, R.W. Harrison, F. Benziane, M. Yazidi and collaborators, 2008) [108]. These data were used here to fill gaps in sample coverage across the western Saghro Massif and to support comparative analysis.

4.3.1. Classification of Magmatic Units of the Ediacaran Volcanic Succession of the ODC

The volcanic and subvolcanic rocks studied from both the 1st and 2nd eruptive cycles of the ODC volcanic succession exhibit variable contents of SiO₂ (46.70 to 79.70 wt.%), Al₂O₃ (8.95 to 18.20 wt.%), CaO (0.06 to 9.73 wt.%), and alkalis (Na₂O + K₂O = 3.54 to 12.50 wt.%). Loss on ignition (LOI) contents are variable (LOI= 0.14 to 6.95 wt.%) reaching values suggesting an impact of post-magmatic processes. Considering the mobility of alkalis, we did not use the TAS diagram [117] to determine the geochemical affinities of the studied rocks. For these purposes, we used the SiO₂ versus Nb/Y classification diagram (Figure 9; [118]). Samples of ODC are clearly plotted in the subalkaline field, ranging from mafic to felsic compositions. The only non-subalkaline samples, TIZ90, TIZ102, and TIZ103, which represent a dolerite dyke and rhyolitic lava flow, respectively, are plotted within the alkaline field, due to post-magmatic processes such as hydrothermal alteration or advanced fractional crystallization.

4.3.2. Geochemistry of the Igneous Rocks of the 1st Eruptive Cycle

The volcanic rocks of the 1st eruptive cycle of the ODC (Lower Ouarzazate Group) exhibit a wide range of silica contents (50.90–79.70 wt.%), spanning compositions from mafic to felsic. The mafic to intermediate rocks are represented by pre-caldera basaltic andesite to andesite lava flows, while the felsic end-members consist of dacitic to rhyolitic ash-flow tuffs (ignimbrites) that constitute both the caldera fill and associated outflow deposits.

• Mafic and intermediate rocks

The pre-caldera basaltic andesite and andesite lava flows from the 1st eruptive cycle of the ODC exhibit a wide range in SiO₂ (50.90–62.70 wt.%), CaO (1.39–6.19 wt.%), TiO₂ (0.53–1.76 wt.%), and MgO (1.56 to 6.50 wt.%) concentrations. The Mg# values (Mg# = Mg²⁺/(Mg²⁺ + Fe²⁺) vary between 0.34 and 0.59, with fairly variable transition element contents (Ni = 5–52 ppm, Cr = 5–190 ppm, Co 3.10–33.20 ppm, and Sc = 9–25ppm; Tables S2 and S3).

The chondrite-normalized REE diagrams of the pre-caldera basaltic andesite and andesite lava flows from the 1st eruptive cycle of the ODC are characterized by a slight enrichment in light rare earth elements (LREE) compared to heavy rare earth elements (HREE), with (La/Yb)_N = 0.58 to 7.85 and (La/Sm)_N =2.20 to 3.38. The HREE trends are generally flat to slightly declining, with (Gd/Yb)_N = 0.14–1.76. Sample T11.1 from published data shows a more enriched chondrite-normalized REE pattern, with a more pronounced negative Eu anomaly related to the plagioclase fractionation (Figure 10A). When normalized to the primitive mantle [119], the mafic and intermediate patterns display high LILE compared to HFSE and HREE, with negative Nb, P, and Ti and positive anomalies of K, Pb, and U (Figure 10B).

• Felsic rocks

The felsic rocks from the 1st eruptive cycle of the ODC display SiO₂ contents ranging from 51.80 to 79.70 wt.%, CaO from 0.06 to 7.44 wt.%, Al₂O₃ between 8.95 and 18.00 wt.%, and TiO₂ contents from 0.08 to 1.96 wt.%. Their compositions are predominantly peraluminous to metaluminous and range from high-K calc-alkaline to calc-alkaline (Figure 11A,B). Most of these felsic rocks fall into the metaluminous and peraluminous fields. The 1st cycle granitoids are metaluminous to peraluminous, possibly I-type to weakly S-type granites. The 2nd cycle granitoids are mostly peraluminous, corresponding to S-type granites (Figure 11A).

The Chondrite-normalized REE patterns of these felsic rocks display a consistent trend, characterized by an enrichment of LREE relative to HREE, with (La/Yb)_N values ranging from 0.11 to 17.30 and (La/Sm)_N from 1.25 to 8.51. The HREE patterns are generally flat to gently falling (Gd/Yb)_N = 0.06–2.57, accompanied by a negative Eu anomaly that is likely due to plagioclase fractionation (Figure 12A). Similar variations in incompatible element concentrations are evident on the primitive mantle-normalized trace element diagram, whereas all felsic rock groups exhibit negative anomalies in Nb, P, and Ti and positive anomalies of U, K, and Pb (Figure 12B; [119]).

4.3.3. Geochemistry of the Igneous Rocks of the 2nd Eruptive Cycle

The volcanic rocks of the ODC 2nd eruptive cycle (Upper Ouarzazate Group) display a wide range in silica content (SiO₂: 46.70–76.10 wt.%), encompassing compositions from mafic to felsic. The mafic to intermediate rocks include post-caldera products such as basaltic and andesitic lava flows (pre-561 Ma), andesitic feeder dykes, dolerite dykes, and the ca. 556–557 Ma Tagmout gabbro. The felsic rocks comprise rhyolitic lava flows and ignimbrites and the ca. 559–555 Ma Isk-n-Alla pink granite.

• Mafic and intermediate rocks

The post-caldera mafic rocks from the 2nd eruptive cycle of the ODC exhibit a large range in SiO₂ (46.70–69.20 wt.%), CaO (0.33–9.73 wt.%), TiO₂ (0.32–3.26 wt.%), and MgO (0.18–7.47 wt.%) contents. The Mg# values range from 0.04 to 0.65, and concentrations of transition elements are highly variable (Ni = 1.03–169 ppm, Cr = 1.73–469 ppm, Co = 1.33–116 ppm, Sc = 5.28–43 ppm; Tables S2 and S3).

The mafic rocks from the 2nd eruptive cycle of the ODC exhibit chondrite-normalized REE patterns characterized by a general enrichment in LREEs relative to HREEs, with (La/Yb)_N values ranging from 0.54 to 7.26 and (La/Sm)_N from 0.98 to 3.52. The LREE segment shows a steady decrease in concentrations from La to Sm, while the HREE patterns are relatively flat to gently declining (Gd/Yb)_N = 0.15-2.58). A slight negative Eu anomaly is observed in most samples, likely reflecting plagioclase fractionation (Figure 10A). The overall patterns are coherent among the samples, with minor variations in the amplitude of enrichment or depletion.

The primitive mantle-normalized multi-element patterns of the mafic rocks from the 2nd eruptive cycle of the ODC display high concentrations of LILEs such as Rb, Ba, Th, U, and Pb, and display negative anomalies in Nb and Ta [119]. The patterns also show strong depletions in Nb, P, and Ti (Figure 10A,B). HFSEs such as Zr display slight depletions or weak positive anomalies, depending on the sample. The overall patterns are consistent across samples, with only minor variations in the degree of enrichment or depletion.

• Felsic Rocks

The felsic rocks from the 2nd eruptive cycle of the ODC have SiO₂ contents ranging from 66.10 to 76.10 wt.%, CaO from 0.06 to 7.44 wt.%, Al₂O₃ between 8.95 and 18.00 wt.%, and TiO₂ from 0.08 to 1.96 wt.%. Their composition is generally peraluminous, with one sample plotting in the metaluminous field, ranging from high-K calc-alkaline to calc-alkaline (Figure 11A,B).

The chondrite-normalized REE patterns of the felsic rocks exhibit enrichment in LREEs relative to HREEs, with (La/Yb)_N ranging from 0.20 to 15.42 and (La/Sm)_N from 0.95 to 5.20. The LREEs decline steadily from La to Sm, while the HREE patterns are flat to gently declining, with (Gd/Yb)_N values between 0.05 and 1.92. Most samples exhibit a slightly negative Eu anomaly (Figure 12A). Overall, the patterns are similar across samples, with minor variations in enrichment level.

5. Discussion

5.1. Age of the ODC Igneous Succession

The compilation of geochronologic, U-Pb zircon data for the igneous rocks of the Ouarzazate Group (i.e., CIMP event: [22,28,45,58,65,80,102,103,106,124,125,126,127,128,129,130,131] and this study) in the Saghro Massif of the Anti-Atlas is summarized on Figure 2 and Figure 13 and references to the literature data and to our own data are compiled in Table S4.

We dated three felsic samples from the ODC volcanic succession in the northwestern part of the Saghro Inlier, Anti-Atlas Morocco. Two of the samples (TIZ41, TIZ46) belong to the 1st eruptive cycle of the Ouarzazate Group, while sample TIZ45 is assigned to the 2nd eruptive cycle of the same group (see Section 4.2 for further details on the samples location, nature, and dating).

Both TIZ41 and TIZ46 samples are syn-caldera deposits, the first being intracaldera and the second extracaldera rocks. The ages of the rocks studied are similar: TIZ41 = ²⁰⁶Pb-²³⁸U age of 563.690 ± 0.051 Ma (MSWD=2.6); TIZ46 = 564.311 ± 0.045 Ma (MSWD=1.5); TIZ45 = ²⁰⁶Pb-²³⁸U 561.61 ± 0.10/0.17/0.63Ma (MSWD = 1.9). Samples TIZ41 and TIZ45 are stratigraphically located close to the top of the volcanic sequence, while the sample T4216 of [28], (rhyolitic tuff), close to the bottom of the volcanic sequence, yielded a SHRIMP zircon ages of 574 ± 7 Ma (573 ± 5Ma), indicating that the ODC volcanism lasted for at least 12 Ma. During this period extracaldera flows were deposited, represented by samples TIZ46 (this study) and T4218 [28], a rhyodacitic tuff that yielded a SHRIMP U-Pb zircon age of 571 ± 5 Ma (571 ± 6 Ma).

Based on the U-Pb zircon dates, the age of magmatism during deposition of the Ouarzazate Group in the Caldera of the Oued Dar’a region can be bracketed between 588 ± 6/7 Ma (age of Zouzmitane microgranite [28]), which belongs to the 1st eruptive cycle (Lower Ouarzazate Group), and 548 ± 3 Ma, the age of the Assaka rhyolitic ignimbrite [102] from the uppermost part of the volcanic pile northeast of Timdghas (2nd eruptive cycle or Upper Ouarzazate Group). The magmatic rocks of the Tagmout graben yielded U-Pb zircon ages between 557-556 ± 5 Ma for the gabbro [28] and 542 ± 5 Ma for the granodiorite [131] Overall, the duration of volcanic activity of the Ouarzazate Group is estimated to span approximately 29–46 million years. The angular unconformity between the 1st eruptive cycle (Lower Ouarzazate Group) and the 2nd eruptive cycle (Upper Ouarzazate Group) of the ODC region must be older than 561.61 ± 0.10/0.17/0.63 Ma (sample TIZ45, this study). The age of the unconformity is thus estimated to be close to ca. 562 Ma (Figure 14). This is close to the age of the angular unconformity between the Lower and Upper Ouarzazate Groups in the Siroua Window [42] and Agadir Melloul Inlier [38,132,133], estimated at 560 Ma. The duration of the hiatus represented by this unconformity remains uncertain, however, it is a critical factor in assessing whether the two volcanic cycles are genetically related.

5.2. Architecture of the ODC: Unique Late Ediacaran Intra-Continental Silicic Super-Eruptions Within a Complex Caldera

The Bouskour Sidi Flah and Timijt rhyolitic dyke swarms have been dated using the SHRIMP zircon method, yielding ages of 563 ± 7 Ma (564 ± 7 Ma) and 562 ± 5 Ma, respectively [28]. These ages are very close to those obtained for the studied samples: sample TIZ41 from the intracaldera facies is dated at 563.690 ± 0.051 Ma, and sample TIZ46 from the extracaldera facies (outflow sheets) at 564.311 ± 0.045 Ma. In contrast, the age of the rhyolitic ignimbrite (sample TIZ45 from this study), dated at 561.61 ± 0.10/0.17/0.63 Ma, is slightly younger, indicating that the Bouskour–Sidi Flah rhyolitic dyke swarms predate this sample and thus the unconformity between the Lower and Upper Ouarzazate Groups. This observation suggests that the Bouskour Sidi Flah and Timijt rhyolitic dyke swarms are part of the 1st eruptive cycle (Lower Ouarzazate Group) and are ultimately connected to the syn-caldera stage. Consequently, the caldera may be larger than previously thought, and these dyke swarms could represent part of its footwall (Figure 15A,B). In addition, the caldera, originally estimated to be approximately 11 km wide and 18 km long [28], is now, with these new constraints, expanded to approximately 35 km in width and 50 km in length. The caldera fill exposed in the steep canyon of the Oued Dar’a has a preserved thickness of over 500 m. This substantial deposit of ignimbrite is a rare occurrence, typically found in caldera formations [1,15,16], as well as in eruptions from ring fissures coinciding with caldera collapse [1]. Assuming a uniform thickness of 500 m for the caldera fill, the minimum eruptive volume of volcanic material is estimated to be approximately 875 km³, which is consistent with the scale expected of large calderas. However, when the volume of outflow-facies ash-flow tuffs is considered, the total estimated volume nearly doubles [28]. This huge volume of ~1750 km³ of ash-flow tuff corresponds to a VEI of 8 for the eruption [81] (see also Figure 1 and Table 1 in [82]). If we add in the extra volume inferred from uplift/erosion predicted to have removed an expected upper part, marked by polygenic conglomerates with volcanic pebbles at the top of the Ouarzazate Group volcanic pile, then the VEI would likely exceed 8. The ODC is thus an Ediacaran very large eruption or super-eruption (supervolcano) like the 640 ka Yellowstone Caldera [28,82].

The western boundary of the caldera (Figure 15) is marked by a ring-fracture vent zone, ~0.75 km wide, filled with flow-banded, aphanitic rhyolites exhibiting a vertical flow fabric. Ring dyke rhyolites are found along the Jbel Azouguiyghn–Tazouli Fault and Awzou N Wallows Fault (AWF). The arcuate shape of rhyolitic dyke swarms and their high density in certain areas suggest that the ODC was formed by the coalescence of multiple collapse structures (Figure 15). The latest major collapse, associated with the eruption of voluminous, crystal-rich rhyolitic to dacitic ash-flow tuffs, likely occurred around the time when the dated c.a. 563 Ma cooling unit, stratigraphically overlying the collapse breccias, was deposited. The ODC thus appears to be a composite caldera rather than a simple, single-collapse structure. The caldera collapse geometry is of the piecemeal or multicyclic caldera type (see Figure 6, p. 654 of [8]). Table S5 compares the ODC with some well-documented calderas or cauldrons. With an estimated diameter of 35 × 50 km and an eruption volume exceeding 1750 km³, the Oued Dar’a Caldera is one of the largest calderas documented globally. Its size is comparable to that of other known calderas, such as La Garita (35 × 75 km), Toba (30 × 80 km), and Yellowstone (45 × 85 km), and its eruption volume places it in the same class as these high-volume volcanic events. Structurally, Oued Dar’a is a caldera cluster, similar to the Yellowstone, Chegem, and Crater Lake systems known for their complex, large-scale collapses and prolonged volcanic activity. Notably, Oued Dar’a Caldera is one of the largest known Ediacaran piecemeal calderas (Table S4 in the supplementary data).

Walsh et al., 2012 [28] interpreted the ODC as a “resurgent caldera” following the model of [1]. They based this interpretation on the restricted, ring-like distribution of moat-fill deposits and the close association of these sediments with an exceptionally crystal-rich (>40%–50%) tuff. Although the crystal-rich volcanic rocks resemble holocrystalline textures, detailed inspection reveals a rapidly cooled, devitrified matrix comprising at least 20%–30% of the rock’s volume. Notably, such rocks were only noticed in the western part of the ODC.

5.3. Depositional Environment and Facies Model for the Volcanic Succession of the ODC

To date, facies models have been the primary approach for specifically representing volcanic stratigraphy and the relationships among lithostratigraphic units defined by their lithologic characteristics. Facies analysis of volcanic deposits has been widely used to develop these facies models [15,16]. The generalized facies models developed from studies of modern volcanoes provide essential reference frameworks that represent important facies associations and, where possible, their spatial and genetic relationships. These facies models also provide a schematic summary of the paleoenvironmental and paleogeographic context of the volcanic field. If successful, these facies models can be useful for understanding ancient volcanic successions and for conducting basin analysis involving volcanic successions.

The geological and palaeogeographical context of the ODC during the Ediacaran period (intracontinental basin developed during the gravitational collapse of the Pan-African Orogen), the lithofacies (a’a lava flows without pillow lavas, lacustrine and fluviatile deposits), the large volume of outpoured magma (>1750 km³), and the duration of volcanic eruptions (23 million years minimum, up 40 million years maximum) suggest emplacement of a large igneous province (LIP) in a continental environment. The facies models that might be relevant are those where volcanism took place in a continental setting, namely continental strato-volcanoes, continental basalt successions, and continental felsic volcanoes [15,16]. However, none of them involves bimodal volcanism. Thus, the model of the ODC volcanic succession necessarily combines all of these three facies models. The volcanological facies model for the Caldera of Oued Dar’a includes “continental strato-volcanoes” with the development of large calderas during the 1st eruptive cycle. It then switches to a “continental basaltic successions” type during the 2nd eruptive cycle, and, finally, concludes with the “continental acid volcanoes” type at a later stage of the 2nd eruptive cycle.

5.4. Petrogenesis and Tectonic Setting

5.4.1. Petrogenesis

Most of the magmatic rocks from both the 1st and 2nd eruptive cycles of the ODC display Nb/Y ratios <0.8, a geochemical feature indicative of sub-alkaline affinity. However, three samples (TIZ90, TIZ102, and TIZ103) of the post-caldera rocks from the 2nd eruptive cycle exhibit an alkaline affinity (Figure 9; [118]).

Mafic rocks from both the 1st and 2nd eruptive cycles of the ODC display Mg# values below 0.59 and 0.65, respectively. These values suggest that the mafic rocks originated from relatively evolved magmas rather than from primary mantle-derived melts (Figure 16). In addition, the moderately low MgO contents (1.56 to 6.50 wt.% and 0.18 to 7.47 wt.%) and low Ni concentrations (<52 ppm, and <169 ppm) in these mafic rocks are substantially lower than those typically associated with primary basaltic magmas (MgO = 10–15 wt.% and Ni = 400–500 ppm) [134,135,136]. These characteristics further support the interpretation that the parental magmas underwent fractional crystallization, likely involving early removal of olivine and/or clinopyroxene. Almost all of the mafic rock samples from both 1st and the 2nd eruptive cycles show weak positive Eu anomalies (Eu/Eu* = 1.1–1.2), implying plagioclase crystallization (Figure 10A).

Diagrams plotting Th against compatible elements such as Sc, Ni, Cr, and Co, where Th serves as an indicator of partial melting extent and magma differentiation, are valuable for assessing the roles of fractional crystallization vs. partial melting in petrogenesis of igneous rocks [137,138,139,140,141,142]. Significant variations in the Th content, particularly when accompanied by an inverse correlation with the compatible elements associated with mantle residual phases or early crystallization minerals (e.g., Sc, Ni, Cr, and Co), reflect differences in the degree of partial melting. Conversely, strong co-variation in Sc, Ni, Cr, and Co with Th suggests the influence of fractional crystallization. As shown in Figure 17, the inverse relation between Th and these compatible elements suggests fractional crystallization as the primary cause for the chemical variability in the studied volcanic rocks.

The La/Nb vs. Nb/U, Nb/Th vs. Nb/La, and La/Sm vs. Nb/La diagrams (Figure 18) indicate that the mafic rocks from the 1st and 2nd eruptive cycles of the ODC are compositionally closer to the Upper Continental Crust (UCC), with a relatively minor contribution from the Lower Continental Crust (LCC) (Figure 18A). These trace element ratios suggest that the mafic rocks of the 1st eruptive cycle likely originated from a mantle source that interacted with continental crust materials or subduction-related components. In contrast, dolerites of the 2nd eruptive cycle exhibit a lower degree of crustal influence and point to a more enriched or primitive mantle source, possibly indicating a shift in the tectonic environment or in the nature of the magmatic processes involved (Figure 18).

One of the notable features of the mafic and intermediate rocks from the 1st and 2nd eruptive cycles of the ODC is the presence of marked negative anomalies in Nb, P, and Ti (Figure 10B). In addition, this Nb depletion is further confirmed by the Nb/U ratios shown in Figure 18A, which are significantly lower than the range of OIB and MORB values (47 ± 10; [143,144]), representing oceanic basalts formed away from subduction influence. However, as continental crust is largely derived from arc magmas [145,146], Nb negative anomalies are also typical of the continental crust [147], complicating the distinction between a supra-subduction origin and crustal contamination in the studied mafic rocks. However, the mafic and intermediate rocks from the 1st and 2nd eruptive cycles of the ODC display marked negative P and Ti anomalies that are characteristic of the continental crust rather than arc magmas [148].

Figure 18. Trace element ratio diagrams for crustal contamination assessment for the analyzed ODC volcanic successions of the Saghro Massif. (A) La/Nb vs. Nb/U; (B) Nb/Th vs. NB/La; (C) La/Sm vs. Nb/La. Data for upper continental crust (UCC) and lower continental crust (LCC) are from [147]. Data for oceanic basalts (OIB) and MORBs are from [119,143]. CC—continental crust; NMORB—normal mid-ocean ridge basalts; EMORB—enriched mid-ocean ridge basalts. Dataset source: [28,83,101,108,109,110,114,115,116] and our own data.

Figure 18. Trace element ratio diagrams for crustal contamination assessment for the analyzed ODC volcanic successions of the Saghro Massif. (A) La/Nb vs. Nb/U; (B) Nb/Th vs. NB/La; (C) La/Sm vs. Nb/La. Data for upper continental crust (UCC) and lower continental crust (LCC) are from [147]. Data for oceanic basalts (OIB) and MORBs are from [119,143]. CC—continental crust; NMORB—normal mid-ocean ridge basalts; EMORB—enriched mid-ocean ridge basalts. Dataset source: [28,83,101,108,109,110,114,115,116] and our own data.

The nature of the mantle source and melting regime of the Oued Dar’a mafic magmas are revealed by the trace element ratios shown in the Nb/La vs. La/Yb and (Gd/Yb)_N vs. (La/Sm)_N diagrams (Figure 19). In the Nb/La vs. La/Yb diagram (Figure 19A), the samples from both eruptive cycles mainly plot within the lithospheric mantle and the mantle influenced by the asthenosphere-lithosphere interaction fields. This distribution suggests derivation from a heterogeneous mantle source variably enriched by asthenospheric input or metasomatic processes. A modified mantle reservoir’s partial melting or progressive enrichment trend is supported by the positive correlation between Nb/La and La/Yb. The majority of the samples in the (Gd/Yb)_N vs. (La/Sm)_N diagram (Figure 19B) fall inside or close to the spinel peridotite field, suggesting melting under spinel stability conditions in the upper mantle. However, a single dolerite dyke plots in the garnet field, indicating deeper melting.

5.4.2. Tectonic Setting

On the Zr/Y vs. Zr and V vs. Ti diagrams (Figure 20A,B), the mafic rocks from the 1st and 2nd cycles plot between the MORB and within-plate basalt fields, suggesting a transitional mantle source with both depleted and enriched (E-MORB-like) components, consistent with magmatism linked to intraplate rifting or early-stage lithospheric stretching [150]. In the La-Y-Nb ternary diagram [151] (Figure 20C), the mafic rocks from the 1st eruptive cycle display calc-alkaline basalt composition and define a trend towards the alkali basalt domain, which is typically associated with intracontinental rift setting and plume-related magmatism [152]. Some samples plot near the Upper Continental Crust (UCC), suggesting crustal contamination (Figure 18A). In contrast, the mafic rocks from the 2nd eruptive cycle show a compositional trend evolving from the calc-alkaline basalt towards the MORB, reflecting increased contribution from a depleted asthenospheric mantle source, likely due to decompressional melting in an extensional regime [135,153]. Together, these trends suggest a tectonic evolution from an initial rift-related magmatic phase (1st cycle), characterized by partial melting of an enriched lithospheric mantle, to a more advanced extensional or post-orogenic setting (2nd cycle), marked by deeper asthenospheric input. This geochemical evolution supports the interpretation that the ODC formed in a post-orogenic extensional regime, likely linked to the late evolution of the Pan-African Orogeny and associated with regional lithospheric thinning and mantle upwelling during the late Ediacaran.

Further insights into the mantle source characteristics and potential crustal contributions are provided by the Th/Yb vs Nb/Yb diagram (Figure 20D; [154]). In this diagram, samples from both volcanic cycles of the ODC fall along or near the compositional array of enriched mantle, trending from E-MORB toward within-plate basalt (Figure 20D). The rocks are notably enriched in Th relative to Nb, suggesting a contribution from subcontinental lithospheric mantle (SCLM), metasomatized by subduction-related fluids or melts [155,156] or, alternatively, assimilation of crustal components during magma ascent [157]. The dolerites from the 2nd cycle plot closer to the MORB–OIB array, implying a predominantly asthenospheric source with limited crustal interaction [135,154].

Figure 20. Tectonic discriminant diagrams for the analyzed mafic rocks from the ODC volcanic field of the Saghro Massif. (A) Geochemical characterization of the ODC mafic and intermediate volcanic rocks in the Zr/Y vs. Zr diagram [158], E-MORB (enriched mid-ocean ridge basalt), IAB, Island Arc Basalt; MORB, Mid-Ocean Ridge Basalt; WPB, Within Plate Basalt; (B) Ti vs. V diagram [159], IAT—Island Arc Tholeiite, IBM—Izu-Bonin-Mariana arc, MORB—Mid-Ocean Ridge Basalt, WPB—Within Plate Basalt; (C) Geochemical characterization of the ODC mafic and intermediate volcanic rocks in the La/10–Y/15–Nb/8 diagram of [151], N-MORB—normal mid-ocean ridge basalt). 1, orogenic domain; 2, intermediate domain (post-orogenic series with crustal participation and “source effect”); 3, anorogenic domain; LCC, lower continental crust; UCC, upper continental crust. PIAT, Primitive Island Arc Tholeiite; MORB, Mid-Ocean Ridge Basalt. MT, mixing line between the MORB E or OIB subdomain and the calc-alkaline orogenic basalt domain; (D) Th/Yb vs Nb/Yb diagram after [154], OIB—Oceanic-Island Basalt, IAB—island-arc basalt. Dataset source: [28,83,101,108,109,110,114,115,116] and our own data.

Figure 20. Tectonic discriminant diagrams for the analyzed mafic rocks from the ODC volcanic field of the Saghro Massif. (A) Geochemical characterization of the ODC mafic and intermediate volcanic rocks in the Zr/Y vs. Zr diagram [158], E-MORB (enriched mid-ocean ridge basalt), IAB, Island Arc Basalt; MORB, Mid-Ocean Ridge Basalt; WPB, Within Plate Basalt; (B) Ti vs. V diagram [159], IAT—Island Arc Tholeiite, IBM—Izu-Bonin-Mariana arc, MORB—Mid-Ocean Ridge Basalt, WPB—Within Plate Basalt; (C) Geochemical characterization of the ODC mafic and intermediate volcanic rocks in the La/10–Y/15–Nb/8 diagram of [151], N-MORB—normal mid-ocean ridge basalt). 1, orogenic domain; 2, intermediate domain (post-orogenic series with crustal participation and “source effect”); 3, anorogenic domain; LCC, lower continental crust; UCC, upper continental crust. PIAT, Primitive Island Arc Tholeiite; MORB, Mid-Ocean Ridge Basalt. MT, mixing line between the MORB E or OIB subdomain and the calc-alkaline orogenic basalt domain; (D) Th/Yb vs Nb/Yb diagram after [154], OIB—Oceanic-Island Basalt, IAB—island-arc basalt. Dataset source: [28,83,101,108,109,110,114,115,116] and our own data.

5.5. Geodynamic Setting of the ODC Volcanic Succession: A Pre-Iapetus Association of a Silicic Large Igneous Province (SLIP) with a Continental Flood Basalt Large Igneous Province (LIP)

The volcanic stratigraphy, geochemistry, and U-Pb zircon geochronology data provided in this study, combined with the literature data (e.g., [28] and references therein) bring new insights into understanding the depositional age, petrogenesis, environment, and facies models of the ODC volcanic succession. Two consecutive eruptive cycles separated by a regional unconformity estimated at 562 Ma can be distinguished in the 588–542 Ma ODC volcanic succession (ODCVS) of the Ouarzazate Group in the Anti-Atlas (Saghro Massif, southern Morocco) based on structural, volcanological and geochemical data.

The 1st eruptive cycle (588–571 to 563 Ma) corresponds to the Lower Ouarzazate Group, and records the pre-caldera stage (lower basaltic andesite-andesite lava flows and associated volcanoclastic rocks) and the syn-caldera stage (dacitic to rhyolitic ash-flow tuff) that comprises intracaldera facies (caldera-forming eruption deposits) and extracaldera facies (outflow sheets). The Bouskour Sidi Flah and Timijt rhyolitic dyke swarms are manifestations of the syn-caldera stage based on their ages and geochemistry. Consequently, the caldera may have been larger than previously thought, and these dyke swarms could represent part of its structural wall. In addition, the caldera, originally estimated to be ~11 km wide and 18 km long [28], is now thought to span approximately 35 km in width and 50 km in length. The ODC is comparable in size to the 640 Ka Yellowstone caldera, indicating that it is a product of a very large Ediacaran eruption or super-eruption (supervolcano). The duration of the volcanic activity of the syn-caldera stage could be estimated to be between 7 and 10 Myrs based on U-Pb zircon ages. Both products of the pre-caldera and syn-caldera stages show the volcanological characteristics of “continental strato-volcanoes”. They also exhibit arc calc-alkaline affinities. Mafic magmas were originated from a lithospheric mantle and/or through asthenosphere-lithosphere interaction and are associated with the gravitational collapse of the Pan-African Orogen. In contrast, felsic rocks were primarily generated through partial melting of crustal lithologies, although contribution from an enriched mantle source contaminated with crustal components cannot be entirely neglected (Figure 21).

The 2nd eruptive cycle (562 to 548–542 Ma) corresponds to the Upper Ouarzazate Group and records the post-caldera stage (upper basaltic andesite-andesite lava flows and associated volcanoclastic rocks), rhyolitic lava flows, and ignimbrites. Mafic rocks, including intrusive rocks, display continental flood basalt and calc-alkaline basalt affinity. Felsic rocks were generated by partial melting of crustal lithologies or an enriched mantle contaminated with crustal components (Figure 21). These units exhibit volcanological features consistent with facies models for continental basalt successions and continental felsic volcanoes.

The duration of the volcanic activity of the Ouarzazate Group could be estimated to be approximately 23–46 million years. The ODC volcanic succession represents a unique association of late Ediacaran intra-continental super-eruptions associated with flood basalt magmatism on the West African Craton. This magmatism marks the transition from orogen collapse to continental rifting, leading to the opening of the Iapetus Ocean. It belongs to the so-called Central Iapetus Magmatic Province (CIMP), or the Ouarzazate Event. In detail, the Ouarzazate Group appears to be an amalgamation of two large igneous provinces (a compound LIP, in contrast to a simple LIP like the Central Atlantic Magmatic Province): a silicic large igneous province (SLIP) between 580 and 563 Ma, and a continental flood basalt large igneous province (CFB LIP) between 561 and 542 Ma.

During the late Ediacaran, ca. 590-540 Ma, the CIMP was emplaced in multiple pulses in the northwestern part of the WAC, preceding the opening of the Central Iapetus Ocean. According to [24,160], a mantle plume was responsible for melting the lower crust to generate the SLIP, followed by partial melting of the plume head [19,20,42,130]. These rocks, exhibiting several successive cycles with alkaline, calc-alkaline and shoshonitic signatures, are interpreted as due to a post-collisional transtensional regime related to the waning stages of the Pan-African Orogeny. Contrary to common assumptions, not all LIPs are dominated by basaltic rocks. Some LIPs, particularly those linked to continental breakup, are characterized by felsic volcanic rocks, with basalt being a minor component [19,161,162], similar to the ODC magmas, which exhibit continuous andesite-dacite-rhyolite composition. Ref. [163] reported that the Ouarzazate Group comprises both a SLIP and a LIP, co-occurring at approximately 580 Ma. The SLIP initiated glaciation, and the subsequent LIP, characterized by flood basalts, concluded a brief glacial period [160,163]. According to [162], the maximum duration of magmatic activity in large igneous provinces can reach 50 Ma, comparable to the record of magmatic activity in the Ouarzazate Group (46 Ma). Therefore, further investigation into the timing of putative siliceous large igneous province activity needs to be undertaken through a more comprehensive analysis of the zircon record preserved in the associated magmatic and sedimentary rocks of the Ouarzazate Group. Additional U-Pb zircon geochronology is needed to establish the duration of magmatism for this group across the entire Anti-Atlas Mountains region.

6. Concluding Remarks

• According to their volcanological, stratigraphical, and geochemical characteristics, two consecutive eruptive cycles can be distinguished in the ODC volcanic succession of the Ouarzazate Group exposed in the Saghro Massif (Anti-Atlas Belt). The two eruptive cycles are separated by an unconformity estimated at 562 Ma, and correspond to Lower and Upper Ouarzazate Group, respectively.
• The 1st eruptive cycle (588–571 to 563 Ma) corresponds to the Lower Ouarzazate Group, and records the pre-caldera stage (lower basaltic andesite-andesite lava flows and associated volcanoclastic rocks) and the syn-caldera stage (dacitic to rhyolitic ash-flow tuff) that comprises intracaldera facies (caldera-forming eruption deposits), extracaldera facies (outflow sheets), and the Bouskour Sidi Flah and Timijt rhyolitic dyke swarms, as well as the ring dykes of the ODC. Mafic magmas of the 1st eruptive cycle clearly show chemical characteristics of calc-alkaline basalts emplaced in the aftermath of the Pan-African Orogeny. Mafic magmas were derived from a lithospheric mantle and a mantle influenced by the asthenosphere-lithosphere interaction and are temporally associated with the gravitational collapse of the Pan-African Orogen. Felsic rocks were generated by partial melting of continental crust. The 1st eruptive cycle is characterized by the formation of a large caldera system. The ODC spans approximately 35 km from east to west and 50 km from north to south and has an intracaldera ignimbrite layer that is over 500 m thick in places. This corresponds to a minimum eruptive volume of approximately 875 km³. Including the volume of the outflow-facies ash-flow tuffs nearly doubles the total eruptive volume to approximately 1750 km³, an amount typical of large caldera-forming events. If this volume were to be emplaced in a single eruption, it would correspond to a Volcanic Explosivity Index (VEI) of 8. The ODC is a late Ediacaran very large eruption or super-eruption (supervolcano) comparable in size to the 640 ka Yellowstone Caldera. The facies model for the ODC succession deposited during the 1st eruptive cycle is that of continental stratovolcanoes.
• The 2nd eruptive cycle (561 Ma to 548–542 Ma) corresponds to the Upper Ouarzazate Group and records the post-caldera stage (upper basaltic andesite-andesite lava flows and associated volcanoclastic rocks) and the terminal rhyolitic lava flows and ignimbrites. The 2nd cycle is marked by bimodal volcanism, featuring tholeiitic basalts (Continental Flood Basalts, CFB) also sourced from a lithospheric mantle and a mantle influenced by the asthenosphere-lithosphere interaction and felsic intraplate magmatisms. The proposed facies model is a composite of two conventional facies models for continental basaltic successions (continental flood basalt, CFB), and a continental silicic volcano type.
• The Ouarzazate Group of the Saghro Massif appears to be an association of two large igneous provinces (a compound LIP versus a simple LIP like the Central Atlantic Magmatic Province): a silicic large igneous province (SLIP) between 588 and 563 Ma, and a continental flood basalt large igneous province (CFB LIP) between 561 and 542 Ma. The Ouarzazate Group belongs to the Central Iapetus Magmatic Province (CIMP), which preceded the opening of the Central Iapetus Ocean.