Volcanic Response to Post-Pan-African Orogeny Delamination: Insights from Volcanology, Precise U-Pb Geochronology, Geochemistry, and Petrology of the Ediacaran Ouarzazate Group of the Anti-Atlas, Morocco

Abstract

Post-collisional volcanism provides valuable insights into mantle dynamics, crustal processes, and mechanisms driving orogen uplift and collapse. This study presents geological, geochemical, and geochronological data for Ediacaran effusive and pyroclastic units from the Taghdout Volcanic Field (TVF) in the Siroua Window, Anti-Atlas Belt. Two eruptive cycles are identified based on volcanological and geochemical signatures. The first cycle comprises a diverse volcanic succession of basalts, basaltic andesites, andesites, dacites, and rhyolitic crystal-rich tuffs and ignimbrites, exhibiting arc calc-alkaline affinities. These mafic magmas were derived from a lithospheric mantle metasomatized by subduction-related fluids and are associated with the gravitational collapse of the Pan-African Orogen. The second cycle is marked by bimodal volcanism, featuring tholeiitic basalts sourced from the asthenospheric mantle and felsic intraplate magmas. These units display volcanological characteristics typical of facies models for continental basaltsuccessions and continental felsic volcanoes. Precise CA-ID-TIMS U-Pb zircon dating constrains the volcanic activity to 575–557 Ma, reflecting an 18-million-year period of lithospheric thinning, delamination, and asthenospheric upwelling. This progression marks the transition from orogen collapse to continental rifting, culminating in the breakup of the Rodinia supercontinent and the opening of the Iapetus Ocean. The TVF exemplifies the dynamic interplay between lithospheric and asthenospheric processes during post-collisional tectonic evolution.

1. Introduction

Magmas associated with continental collisional orogens can be subdivided into pre-collisional, syn-collisional, and post-collisional [1,2,3,4]. Pre-collisional magmatism takes place during oceanic subduction, forming continental arcs similar to the modern Andean system. Syn-collisional magmatism occurs during continental collision, deep subduction, and exhumation, often coinciding with ultrahigh-pressure metamorphism 5 to 10 million years later, potentially resulting from decompressional melting of the earlier subducted continental crust [2,4]. Post-collisional magmatism marks the end of the Wilson cycle and signals the beginning of anorogenic within-plate tectonic setting, corresponding to the unroofing, collapse, and extension of the orogenic belt.

The Neoproterozoic to early Cambrian breakup of Rodinia led to separation of Laurentia, Baltica, and Amazonia/West Africa. It was associated with large igneous provinces (LIPs), protracted multiphase rifting, and variable subsidence along continental margin segments (e.g., [5] and references therein). The 620–520 Ma Central Iapetus Magmatic Province (CIMP) was emplaced in multiple pulses (likely representing separate LIPs), linked to the breakup of Rodinia and the consequent opening of the Iapetus Ocean. It is well represented in Laurentia [5,6,7] and Baltica [8,9,10], but also in other previously attached blocks, such as the Amazonian and West African cratons [11,12].

In the Pan-African Anti-Atlas Belt along the northern margin of the West African craton (WAC), the CIMP LIP is well represented by intrusives, volcanics, and sedimentary sequences forming the Ouarzazate Group (“PIII” or XIII of [13]; Figure 1). The latter was not deformed during the Pan-African Orogeny, suggesting its deposition in post-collisional setting during the extensional/transtensional stage in the Pan-African Belt [12,14,15,16,17,18,19,20,21,22,23,24]. The Ouarzazate Group covers a large area of ~2 × 10⁶ km², with an estimated magma volume of ~1 × 10⁶ km³ and variable volcanic thickness between 300 and 2000 m, and resulted in intense hydrothermal activity, which produced diverse mineral deposits such as Imiter and Zgounder (Ag-Hg), Bou Azzer (Co-Ni-As-Ag-Au), Iourirn (Au), and Bou Madine (Cu-Pb-Zn-Au-Ag) [21,22,25,26].

The most-complete and best-preserved volcanic sections of the Ouarzazate Group are in the Ouarzazate–Siroua–Bou Azzer regions, where the “locus typicus” of the Ouarzazate volcanic series was defined for the first time [13]. While the geochemistry, petrology, and U-Pb geochronology of the Ediacaran volcanic formations of the Anti-Atlas CIMP (i.e., Ouarzazate Group) have been extensively studied (e.g., [11,23,24,28,29,30] and references therein) some questions remain to be addressed. Furthermore, the geodynamic significance of the CIMP remains a topic of contention and a limited number of studies have been conducted on the physical characteristics of these volcanoes, facies architecture, emplacement mechanisms, eruption style, and facies model.

Among the post-collisional, Pan-African volcanic complexes of the Anti-Atlas (Morocco), the Taghdout Volcanic Field (TVF) of the Siroua Window (Central Anti-Atlas) (Figure 1), with its well-established Ediacaran age, is particularly interesting for its structural context, due to the vicinity of the Anti-Atlas Major Fault (AAMF), multiple volcanic facies, and eruptive dynamics and yet easy access. Volcanic eruptions predominantly took place in a continental environment during the Ediacaran period, between 580 and 545 Ma ([11,19,24,31,32,33], this work). These eruptions occurred following the last tectonic event of the Pan-African orogenic cycle (885–545 Ma), which was dated at ca. 608 ± 7 Ma in the Bou Azzer Inlier [24]. In addition, the main tectonic event (collision) of this orogenic cycle is dated at ca. 647 Ma in the Siroua Window [34].

In this paper, we will present new volcanological, petrographic, mineralogical, and geochemical data for the Ediacaran volcanic succession of the TVF in the Siroua Window of the Central Anti-Atlas, aiming to (i) describe the main lithofacies of the volcanic succession and propose a facies model for their emplacement, (ii) clarify their geochemical characteristics and magmatic affinities, (iii) discuss their petrogenesis and mantle sources, (iv) propose a geodynamic model for the Ediacaran Ouarzazate Group of Morocco. In addition, we will present a precise zircon U-Pb age for the top of the Ouarzazate volcanic pile and place it in the context of the available zircon U-Pb data for the Siroua Window and the rest of the Central Anti-Atlas.

2. Geological Background

The West-African Craton (WAC) is one of the five Precambrian cratons that form the African Plate, alongside the Kalahari, Congo, and Tanzania cratons and Sahara Metacraton. It contains two Archean basement shields; the Reguibat Shield in the north and the Leo-Man Shield in the south. Their Archean (3.5 to 2.7 Ga) cores, including greenstone belts and tonalite–trondhjemite–granodiorite (TTG) plutons, are surrounded by the Paleoproterozoic domains (2.35–2.00 Ga) and are covered by the Meso- to Neoproterozoic and younger sedimentary basins [35,36].

The Anti-Atlas Belt in southern Morocco, located along the northern margin of the West African Craton (WAC), forms a broad (800 km long and 200 km wide) anticlinorium with an ENE–WSW trend parallel to the Alpine High-Atlas Mountains. The belt is separated from the High-Atlas and Meseta domains by the South Atlas Fault in the north (SAF; Figure 1). The Precambrian basement is well exposed in several erosional inliers (known as “boutonnières”), within the late Ediacaran and younger rock units (Bas Drâa, Ifni, Kerdous, Tagragra of Akka, Tagragra of Tata, Igherm, Siroua, Zenaga, Bou Azzer, Saghro, and Ougnat; Figure 1), and aligned along two major fault systems, the South Atlas Fault (SAF) and the Anti-Atlas Major Fault (AAMF; Figure 1).

The Anti-Atlas Belt contains the Paleoproterozoic basement (affected by the Paleoproterozoic Eburnean Orogeny) unconformably overlain by the Anti-Atlas Supergroup (including the Lkest-Taghdout, Bou Azzer, Iriri, and Saghro groups sensu [28]), which is in turn unconformably overlain by the Ouarzazate Supergroup (related to the Pan-African Orogeny). The two major crust-forming periods of tectonothermal activity have been recognized during the Proterozoic in the Anti-Atlas Belt and are described below.

(i) The Paleoproterozoic Eburnean Orogeny (2.20–2.07 Ga). The Paleoproterozoic basement rocks outcrop exclusively SW of the AAMF [13], but they may also constitute the basement of the Pan-African Orogen to the NE [22,37,38]. They contain metasedimentary rocks, granites, paragneisses, and migmatites, with zircon U–Pb ages ranging from 2200 to 2030 Ma (e.g., [19,22,39,40,41,42]), intruded by dolerite dyke swarms with five clusters of ages: 2040, 1750, 1650, 1416–1380, and 885 Ma [43,44,45,46]. The ca. 2040 Ma felsic intrusive magmatism in the Anti-Atlas inliers is interpreted to be derived from melting of the lower crust via mafic underplating associated with the Tagragra of the Tata LIP [43]. The 885 Ma mafic magmatism is linked to Rodinia, and the 1750, 1650, and 1380–1416 Ma magmatic events are associated with the Nuna/Columbia extension (cf. [47,48]).

(ii) The Neoproterozoic Pan-African Orogeny (885–545 Ma). The Neoproterozoic sequences of the Anti-Atlas Belt associated with the Pan-African Orogeny are unconformably overlain by the post-collisional, volcano-sedimentary Ouarzazate Supergroup. The latter belongs to the Central Iapetus Magmatic Province (CIMP; [11,18,21,26,43,49,50]). The Neoproterozoic successions underlying the Ouarzazate Supergroup are subdivided into the lower units (Lkest-Taghdout, Bou Azzer, and Iriri groups), affected by all Pan-African orogenic events, and upper units (Saghro and Bou Salda groups), only affected by the latest stage of the Pan-African Orogeny (e.g., [21,22,28,40,43,51]).

The Lkest-Taghdout Group consists mainly of basalt, quartz sandstone, and stromatolitic/oolitic carbonate, intruded by dolerite dyke swarms and sill complexes [28,52]. Its lower and middle parts are Paleoproterozoic ([27,53] and references therein), while the upper part corresponds to the rifting at the northern portion of the WAC during the late Neoproterozoic at 885 Ma [27,54]. The rifting culminated with the opening of an oceanic basin on the northern margin of the WAC. The relics of the oceanic crust (Bou Azzer Group) are preserved in the Bou Azzer and Siroua inliers as strongly sheared, allochthonous ophiolite complexes ([23,55] and references therein). The ophiolitic assemblage mainly consists of ultramafic rocks (serpentinites with few chromite pods), but also includes mafic metacumulates, metabasaltic sheeted dykes, and pillow lavas [56,57,58,59]. Their geochemical signatures point to emplacement in a supra-subduction setting [59,60,61,62,63,64,65,66]. Zircon U–Pb dating of plagiogranite intrusions in the Siroua Inlier constrained oceanic accretion at ca. 760 Ma [59,67]. Subsequent to the opening of the ocean basin, subduction under an island arc was initiated (recorded by the Iriri Group in the Siroua Inlier; [19,68,69]). Migmatite from the Iriri Group yielded SHRIMP U–Pb ages at 743 ± 14 and 663 ± 13 Ma [19]. The older age refers to the emplacement of the migmatite protolith in an island–arc setting, whereas the younger age corresponds to ophiolite obduction and arc accretion followed by the regional metamorphism at the northern edge of the WAC in the Anti-Atlas Belt [19]. The development of the Anti-Atlas Belt oceanic arc complexes occurred during three magmatic flare-ups at ca. 750, 700, and 650 Ma that were punctuated by an early, major tectonic episode between 730 and 700 Ma [65,70]. A garnet Sm–Nd date of 647.2 ± 1.7 Ma was recently obtained for the regional metamorphism within the Tasriwine ophiolite complex of the Siroua Inlier [34]. This age is more than 15 million years younger than the previous age estimate [19]. North of the AAMF, in the eastern Anti-Atlas Belt (Figure 1), the oldest rocks are thick deposits of basalts and turbidites of the Saghro Group, with detrital zircon U-Pb ages forming peaks at ca. 880 to 615 Ma [38]. The depositional age of and geodynamic constraints for the Saghro Group, unconformably overlying late Ediacaran Ouarzazate Group, are still debated and controversial [22,38,71,72,73,74,75,76,77,78]. Several authors suggested the Saghro Group was deposited in the early Neoproterozoic and age equivalent to the Tachdamt–Bleida passive continental margin sequence in the Bou Azzer Inlier [19,51,72,79,80,81]. Based on this interpretation, the age of the Saghro Group was considered to be Tonian to Cryogenian and its deformation has been linked to the main stage of the Pan-African Orogeny (650 Ma; [22,73,79]). Alternatively, it has been suggested that the Saghro Group was deposited in a back-arc setting [71,72,74,75,82] between the WAC margin and the northern Cadomian arc [29,76,77,78,83].

The post-collisional period (630–545 Ma) evolved from a transpressional to a transtensional tectonic setting, with the emplacement of large, high-potassium calc-alkaline to shoshonitic volcanic/plutonic bodies, accompanied by deposition of sedimentary successions in pull-apart basins and grabens. Deposition of the Bou Salda/Tidilline and Ouarzazate groups corresponds to the final stage of the Pan-African Orogeny. The Bou Salda/Tidilline groups (600–580 Ma) are characterized by poorly sorted sediments and diamictites associated with volcanic and intrusive rocks. The Bou Salda Group could be older than 600 Ma [19]. The Ouarzazate Group (580–545 Ma) is mainly composed of high-potassium, calc-alkaline volcanic, volcanoclastic, and plutonic rocks that are widely distributed in the High-Atlas and Meseta (e.g., [16,18,21,22,26,49,83,84,85,86,87,88]. The Ouarzazate Group was followed by deposition in the foreland basin represented by the Taroudant and Tata groups [49,84].

The Taghdout Volcanic Field (TVF, Siroua Window, Central Anti-Atlas) comprises volcano-sedimentary sequences represented by basalts, andesites, dacites, rhyolites, ignimbrites, conglomerates, sandstones, siltstones, and carbonates. They form volcanic successions that unconformably overlie the Neoproterozoic to Paleoproterozoic basement in the Aguerzoula, Taghdout, and Tinzaline subinliers, located to the north of the MAAF. This defines the northern border of the Zenaga Inlier and the margin of the WAC (Figure 2). The TVF, 26 km long and 8 to 10 km wide, forms a NW–SE arc, with dips around 20° towards the north or northeast; it is displaced by multiple strike–slip faults, running NE–SW to E–W.

3. Materials and Methods

3.1. Stratigraphy and Petrography of Volcanic Sequences

This study is the result of extensive fieldwork on the Taghdout Volcanic Field (TVF), including the study of outcrops and the stratigraphic logging of eight sections. We have produced a comprehensive volcanic stratigraphy of the area, identifying the products of each individual eruption, their relative stratigraphy, and surface area. Volcanic stratigraphy is an essential tool for understanding the evolution and volcanic history of this volcanic field and for establishing a correct facies model for its emplacement and architecture. Petrographic analysis was carried out on a total of 200 thin sections, allowing the study of textures and mineralogical composition.

3.2. High-Precision U-Pb Geochronology Employing CA-ID-TIMS Technique

Selected zircons from sample SI88 were analyzed for high-precision geochronology using chemical abrasion, isotope dilution, thermal ionization mass spectrometry (CA-ID-TIMS) at ETH Zurich. The crystals were annealed at 900 °C for over 60 h and subsequently loaded in Teflon microcapsules for chemical abrasion with concentrated HF and placed in a high-pressure Parr bomb at 190 °C for 12 h. Afterwards, the zircons were transferred to Teflon beakers and washed in 6N HCl on a hot plate (at 80 °C) and in 4N HNO₃ in an ultrasonic bath. The crystals were then loaded back into the pre-cleaned microcapsules for dissolution with concentrated HF and trace HNO₃. The EARTHTIME ²⁰²Pb-²⁰⁵Pb-²³³U-²³⁵U isotope tracer solution (ET2535; [89,90]) was added to ensure complete sample–spike equilibration. The zircons were dissolved in Parr bombs over 60 h at 220 °C.

The solutions were subsequently dried down and redissolved in 6 N HCl at 190 °C for 4–6 h in the Parr bombs. Then, the solutions were dried down and redissolved in 3N HCl for column chemistry, modified after [91], which efficiently separates Pb and U from other elements. Uranium and Pb were collected and dried down with one drop of 0.02 M H₃PO₄. Finally, the samples were loaded onto outgassed, zone-refined Re filaments with a silica gel emitter (modified after [92]).

Analyses were conducted on a Thermo TRITON Plus thermal ionization mass spectrometer equipped with an axial secondary electron multiplier and Faraday cups connected to 1013 Ω amplifiers in a setup described in [93,94]. Lead and U isotopes were analyzed using static multicollection in the Faraday cups, except for ²⁰⁴Pb, which was analyzed in the axial secondary electron multiplier. Pb and U mass fractionations were corrected using the double spike (ET2535) and assuming a sample ²³⁸U/²³⁵U ratio of 137.818  ±  0.045 (2σ; [95]). Data reduction and visualization were performed using Tripoli and UPb Redux software (v. 3.7.1) [96]. ²⁰⁶Pb/²³⁸U dates were corrected for initial ²³⁰Th/²³8U disequilibrium using a constant Th-U partition coefficient ratio of 0.2. All dates are reported with 2-sigma uncertainty which are presented as μ ± x|y|z, where μ is the date, x is the uncertainty including only analytical uncertainties, y includes the uncertainty of the tracer calibration, and z also includes the uncertainty on the decay constant.

3.3. Electron Microprobe Analyses

Electron microprobe (EMP) analyses of clinopyroxene were carried out with a CAMEBAX microprobe (Camparis, University Pierre and Marie-Curie of Paris VI, Paris, France) using oxide standards and natural minerals for calibration. Operating conditions included an acceleration voltage of 15 kV, a beam current of 40 nA, a minimum beam diameter of 1 µm, and a counting time varying between 10 and 40 s, depending on the element. The analytical accuracy and precision estimated from repeated analyses on natural standards are within ±2% for major elements and ±5% for minor elements. The units of the clinopyroxene formula have been calculated on the basis of 6 oxygen atoms.

3.4. Whole Rock Geochemistry

To investigate the tectonic setting of the Taghdout Volcanic Field (TVF), we combined our own geochemical data [14,97] and previously published literature data [98,99].

Element concentrations for whole rock analyses were determined at the “Service d’Analyse des Roches et des Minéraux” of the Geochemical and Petrographic Research Centre in Nancy (SARM Laboratory-CRPG-CNRS, Nancy, France). To avoid contamination, crushing and pulverization were carried out using agate mills. Powders were treated with LiBO₂, dissolved with HNO₃, and analyzed and calibrated with international geostandards (basalt BE-N, basalt BR, dolerite WS-E, and microgabbro PMS). Major and trace elements were analyzed by inductively coupled plasma–atomic emission spectroscopy (ICP-AES) and inductively coupled plasma–mass spectrometry (ICP-MS), respectively. The analytical procedures are described in detail at https://sarm.cnrs.fr/index.html/ (URL accessed on 28 January 2025). The asterisk-marked rocks were also analyzed for REE and additional trace elements by the neutron activation method. The method used was pure instrumental activation analysis (without chemical separation) using epithermal neutron irradiation (OSIRIS reactor at Saclay, France, “Commissariat de l’Enérgie Atomique, Groupe des Sciences de la Terre, Laboratoire Pierre Süe”). Irradiation was carried out in Cd vials, followed by several measurements taken at different intervals from 8 days to 1 month after irradiation [100]. The reference standard used was GNS [101], while the sample standard used was BEN [102].

4. Results

4.1. Fieldwork Observations, Volcanic Stratigraphy, and Petrography

4.1.1. The Volcanic Succession of the Jbel Adrar Tougmast Type Section

The detailed distribution of facies both laterally and vertically was examined and is illustrated in Figure 3. Depending on the area, a thickness between 420 and almost 1500 m of effusive, pyroclastic, and siliciclastic rocks was deposited, representing between 480 and 600 km³ of volcanic and sedimentary deposits. Accumulation of these volcano-sedimentary sequences is the result of a volcanic evolution that differed in time and space. The most complete volcanic successions are exposed at Adrar Tougmast, Bou Hanou, South Aguerzoula, and North Aguerzoula (Figure 3, columns A, B, F, and H).

The Jbel Adrar Tougmast lithostratigraphic column (Figure 3, column A) was measured in the section at the base of Jbel Adrar Tougmast. It represents a succession of lithofacies that are separated by an angular unconformity into two main formations: the Adrar Tougmast Formation at the base, overlain by the Tizi-n-Bachkoun Formation (Figure 4A).

The Adrar Tougmast Formation

The formation consists of the following sequences described in ascending order:

(i) A 140 m thick sequence of basalt, andesite, and hybrid lava flows and associated volcaniclastic facies that unconformably overlies the Paleoproterozoic basement with quartz sandstones of the Taghdout-Lkest Group. It begins with about 40 m of porphyritic andesites with plagioclase megacrysts. In the thin sections, porphyritic andesites have large euhedral phenocrysts of mostly plagioclase and less commonly hornblende and pyroxene. Less common are phenocrysts of biotite, apatite, ilmenite, magnetite, and titanomagnetite (Figure 5A).

(ii) A 110 m thick sequence of porphyritic rhyolite and dacite lava flows separated by a lenticular conglomerate (middle conglomerate). The porphyritic rhyolite consists of quartz, oligoclase, sanidine, and, less commonly, amphibole, biotite, and pyroxene phenocrysts in a fine-grained matrix of mainly quartz and feldspar. The porphyritic dacite shows quartz vein stockwork.

(iii) A 140 m thick succession of crystal-rich, quartz- and biotite-bearing, moderately welded ash flow tuff. The phenocryst content is >30%, typically of feldspar, quartz, and biotite, with a few percent of porphyritic, juvenile fragments and aphanitic, volcanic lithic fragments. Locally, the eutaxitic texture is well developed in this unit (Figure 5B). An interval of red siltstones occurs near the top of the sequence, indicating a volcanic quiescence.

(iv) A 200 m thick sequence of welded, rhyolitic ignimbrite with fiammes and scattered lithic fragments (Figure 4B and Figure 5C). At least three cooling units can be distinguished (Figure 4C,D).

Tizi-n-Bachkoun Formation

Above the angular unconformity, the Tizi-n-Bachkoun Formation is composed of:

(v) A 200 m thick, siliciclastic fluvial sequence of conglomerates, sandstones, and siltstones (upper conglomerates).

(vi) A 100 m thick sequence of basalt and basaltic andesite lava flows (Figure 4A).

(vii) Fifty-meter-thick conglomerates that comprise poorly sorted boulder beds, horizontally stratified pebbly to blocky to clast-supported beds, and lenticular sandstones. Boulders are subrounded to subangular and are occasionally imbricated. Oversized boulders (up to 160 cm) of lava flow and ignimbrite are occasionally present.

(viii) One-hundred-meter-thick eutaxitic ignimbrites, typically with fiammes and glass shards (Figure 5D), and matrix-supported lapilli tuffs with vitroclastic matrix. Some lapilli tuffs lack matrix and show mostly unfractured crystals but contain oriented, devitrified fiamme. These textural features suggest rheomorphic ignimbrites. Locally, ignimbrites are associated with rhyolitic domes and lava flows.

(ix) Ninety-meter-thick fluvial conglomerates (upper conglomerates).

4.1.2. Stratigraphic Correlation with Other Regions

The lithostratigraphic columns established for other areas show significant differences relative to the Jbel Adrar Tougmast section, indicating lateral variations within the volcanic succession.

(i) In the Bou Hanou area (Figure 3, column B), there is an increase in the thickness of the volcanic pile compared to the Jbel Adrar Tougmast section due to the multiple basaltic and andesitic lava flows with sandstone intercalations and a thicker sequence of rhyolitic ignimbrites with up to six cooling units. Furthermore, the siliciclastic sedimentary sequence that precedes deposition of the basalt and basaltic andesite flows of the Tizi-n-Bachkoun Formation is absent there. Instead, lenticular conglomerate layers and lacustrine, stromatolitic limestones characterize the latter formation in this area.

(ii) In the Lizate area, volcanic pile is up to 435 m (Figure 3, column C) with a decametric interval (maximum thickness of 50 m) of polymictic conglomerates (lower conglomerates) having quartzite, schist, and granite pebbles at the base of the volcanic pile. The porphyritic rhyolite and dacite lava flows and the welded rhyolitic ignimbrite with fiammes of the Adrar Tougmast Formation are also absent. Further, the upper sequence of the volcanic succession (the Tizi-n-Bachkoun Formation) is missing there, suggesting a topographic high (uplift) in this area during the emplacement of the units of the second eruptive cycle.

(iii) In the Tinzaline area (Figure 3, column D), the volcanic pile thickness is reduced compared to the Jbel Adrar Tougmast section and most of the upper units of the Tizi-n-Bachkoun Formation are absent.

(iv) In the Halouqt area (Figure 3, column E), the rhyolitic, crystal-rich tuff sequence of the Adrar Tougmast Formation and basaltic and basaltic andesite lava flows of the Tizi-n-Bachkoun Formation are missing.

(v) In the Aguerzoula Inlier (Figure 3, columns F, G, and H), the rhyolitic, crystal-rich tuff sequence (Figure 3, columns F and H) and dacite lava flows (Figure 3, column G) are absent. Hybrid lavas are generally abundant. In Tizi-n-Bachkoun, the sequence of crystal-rich tuffs and ignimbrites is well exposed with several cooling units. The basaltic and basaltic andesite lava flows of the Tizi-n-Bachkoun Formation often show intercalations of red siltstones. The rhyolitic bodies often form domes in association with ignimbrites, having small thickness and short lateral extent.

4.2. CA-TIMS Zircon U-Pb Geochronology

We analyzed a total of six single zircon crystals from sample SI88. Five out of six zircons yielded indistinguishable 230Th-corrected ²⁰⁶Pb/²³⁸U dates with a weighted mean of 557.613 ± 0.043 Ma (MSWD = 0.33; see Supplementary Materials Table S1). One zircon yielded a younger date of ca. 556 Ma (Figure 6) that we consider reflecting residual Pb loss. The weighted mean ²⁰⁶Pb/²³⁸U date of SI88 tightly constrains the age of the upper part of the volcanic pile and together with previous ages constrains the duration of volcanic activity in the Taghdout Volcanic Field and the Ouarzazat Group as a whole (see Section 5.3).

4.3. Whole-Rock Geochemistry

The representative whole-rock major and trace element data (REE included) of the Ediacaran TVF volcanics are presented in the Supplementary Materials (Table S2). The used geochemical database is compiled from both the literature data [14,97,98,99] and new data presented herein.

The lithostratigraphic data for the Ediacaran TVF allow for distinguishing the Adrar Tougmast and the Tizi-n-Bachkoun formations, separated by an angular unconformity and characterized by very different volcanic facies. This was taken into account to group the studied volcanic rocks (including pyroclastics) into two main cycles: (i) the first cycle (1st cycle) corresponds to the Adrar Tougmast Formation that forms the lower part of the TVF volcanic pile and that includes lavas of basaltic to rhyolitic compositions, but also hybrid lava flows and ignimbrites, and (ii) an upper sequence (2nd cycle) corresponding to the Tizi-n-Bachkoun Formation, also composed of basaltic to rhyolitic lavas rocks and including breccia and ignimbrites.

4.3.1. Classification of Volcanic Rocks of the Ediacaran Taghdout Volcanic Field

Since the studied rocks show evidence of greenschist-facies metamorphism (association of albite, actinolite, and chlorite) and because high-field strength elements like Nb and Y are less susceptible to alteration during metamorphism than Na₂O and K₂O, we used the SiO₂-Nb/Y diagram [103], considered to be more reliable to distinguish the geochemical affinities of altered rocks than alkali-based schemes. On such a diagram (Figure 7), the volcanic rocks of both eruptive Ediacaran TVF cycles are plotted in the subalkaline field ranging in composition from basalts to rhyolites (SiO₂ = 44.6 to 80.8 wt%). According to the systematics shown in Figure 7, the compositional range of the 1st cycle comprises basalts, andesites, rhyodacites/dacites, and rhyolites, while the 2nd cycle ranges in composition from basalts to rhyolites through basaltic andesites. In order to facilitate rock description, they were arbitrarily subdivided into “mafic” and “felsic” rocks, these terms being used in a sensu lato. The mafic group includes basalts and andesites whereas the felsic group includes rhyodacites, dacites, and rhyolites.

4.3.2. Geochemistry of Volcanic Rocks of the 1st Cycle of the Taghdout Volcanic Field

The volcanic rocks of the first eruptive cycle of the TVF (i.e., the Adrar Tougmast Formation) range in SiO₂ content from 47.7 to 80.8 wt%, from mafic to felsic compositions.

Mafic Rocks

The mafic rocks of the 1st eruptive cycle of the TVF range in SiO₂ content from 47.7 to 57.5 wt% and have Mg# values (Mg# = Mg²⁺/(Mg²⁺ + Fe²⁺) from 0.10 to 0.61, with correspondingly variable transition element contents (Ni = 8–68 ppm, Cr = 5–188 ppm, Co 16–54.4 ppm, Sc = 20.7–30.4 ppm). In agreement with their subalkaline characteristics, TiO₂ contents are low, from 0.98 to 2.30 wt%.

Primitive mantle-normalized trace element “spidergrams” for the mafic rocks of the 1st eruptive cycle of the TVF (Figure 8) are characterized by variable fractionation of REE, with enrichments of more incompatible LREE over HREE, corresponding to (La/Yb)_N = 1.53–8.37 (Figure 8A). They are characterized by significant negative anomalies of Nb ((Th/Nb)_N = up to 8.39), Ta, and Sr and less pronounced Ti anomaly (Figure 8B).

Felsic Rocks

The herein considered felsic rocks of the 1st eruptive cycle of the TVF are characterized by SiO₂ contents varying from 57.58 to 80.80 wt%, with Mg# values ranging from 2.24 to 43.10 and TiO₂ contents from 0.71 to 1.64 wt%. They have largely peraluminous to slightly metaluminous, high-K calc-alkaline to calc-alkaline compositions (Figure 9A,B). Their REE patterns are characterized by significantly more pronounced Eu negative anomalies (Eu/Eu* = down to 0.12) compared with the mafic rocks from the same cycle (Eu/Eu* = 0.75 to 0.95).

The felsic rocks, plotted in the normalized multielement diagrams (“spider diagrams”), correspond to different stages of magma evolution, with almost all of them showing similar primitive-mantle-normalized patterns (Figure 10A,B), potentially indicating comagmatism. The patterns are all characterized by a significant enrichment in LILE (Rb, Ba, K, and Th), significant and systematic Sr and Ti negative anomalies, with most of them also having marked Nb and Ta negative anomaly ((Th/Ta)_N = up to 4.14; (Th/Nb)_N = up to 9.44).

4.3.3. Geochemistry of Volcanic Rocks of the 2nd Eruptive Cycle of the TVF

The volcanic rocks of the 2nd eruptive cycle of TVF (the Tizi-n-Bachkoun Formation) range in SiO₂ content from 44.6 to 79.2 wt%, corresponding to ultramafic to felsic composition. This cycle, in contrast to the first one, is clearly bimodal with a significant gap in silica content between 56.2 and 68.9 wt%.

Mafic Rocks

Mafic rocks of the 2nd eruptive cycle of the TVF are characterized by SiO₂ content ranging from 44.6 to 55.0 wt% and have Mg# varying from 0.22 to 0.50 and variable transition element content (Ni = 13–181 ppm, Cr = 17–311 ppm, Co 12–66 ppm, Sc = 25.9–37.5 ppm). TiO₂ contents are low, 0.96 to 2.41 wt%.

When compared with the mafic rocks of the 1st cycle, the mafic lithologies of the Tizi-n-Bachkoun Formation are significantly less enriched in incompatible elements, as shown, for example, by La (6.98 to 31.6 ppm vs. 21 to 56.15 ppm for the 1st cycle), and have less fractionated REE patterns ((La/Yb)_N = up to 8.37) compared to the 1st cycle (up to 13.15). In the normalized multielement diagrams (“spider diagrams”), basalt patterns for the second eruptive cycle of the TVF (Figure 8A) are more variable in LILE (Ba, U, Th) concentrations, having strong Nb, Sr, and Ta negative anomalies (e.g., (Th/Nb)_N = 3.50–5.25).

Felsic Rocks

Felsic rocks of the 2nd eruptive cycle of the TVF have peraluminous, calc-alkaline to high-K calc-alkaline compositions (Figure 9A,B). They show significantly fractionated LREE compared to HREE ((La/Yb)_N = 3.13–12.16; 7.03 on average) and systematic negative anomalies in Eu (Figure 10A). In the normalized multielement diagrams (“spider diagrams”), rhyolites of the 2nd eruptive cycle of the TVF display uniform patterns (normalized to the primitive mantle, [105]; Figure 10B). The patterns are characterized by a significant enrichment in LILE (Rb, Ba, K, and Th) and pronounced negative anomalies in Nb and Ta (normalized Th/Ta ratio varies from 5.07 to 8.58 with an average of 6.83 and normalized Th/Nb ranges from 3.23 to 15.98 with an average of 10.69). Systematic negative anomalies are also observed for Ti.

4.4. Mineral Chemistry of Clinopyroxene

Clinopyroxene (cpx) in the studied rocks occurs as either a matrix component or as phenocrysts but is rarely well preserved. In most cases, clinopyroxene is entirely replaced by calcite, epidote, iron oxides, and chlorite, forming ghosts only recognizable by their habit.

Two representative samples from the basaltic andesite of the Adrar Tougmast Formation (Sample YS 60) and basalt of the Tizi-n-Bachkoun Formation (Sample YS 22) were selected for pyroxene composition analyses (Supplementary Materials Table S3).

The clinopyroxenes of the basaltic andesites of the Adrar Tougmast Formation (1st cycle) are augite with an average value of En43-Wo42-Fs15 (Figure 11A). Their chemical characteristics indicate that they are pyroxenes typical of non-alkaline basalts: having high silica content (51.48 wt% ≤ SiO₂ ≤ 53.53 wt%) with relatively low alumina (1.62 wt% ≤ Al₂O₃ ≤ 3.29 wt%, with an average of 2.23 wt%) and titanium (0.40 wt% ≤ TiO₂ ≤ 0.78 wt% with an average of 0.54 wt%) contents and low alkaline and alkaline earth contents ([Ca + Na] < 1) (Figure 11B). The clinopyroxenes of the Tizi-n-Bachkoun Formation (2nd cycle) basalts are also augites with an average value of En42-Wo40-Fs18 (Figure 11A), showing identical non-alkaline chemical characteristics: high silica content (48.87 wt% ≤ SiO₂ ≤ 53.62 wt%), low titanium (0.45 wt% ≤ TiO₂ ≤ 1.03 wt%, with an average of 0.78 wt%), aluminum (1.38 wt% ≤ Al₂O₃ ≤ 3.43 wt% with an average of 2.58 wt%), and alkaline and alkaline earths ([Ca + Na] < 1) concentrations. The calc-alkaline nature of the magmas from which the clinopyroxenes crystallized is indicated in Figure 11C.

5. Discussion

5.1. Depositional Environment of the Ediacaran Volcanism of the TVF

Geological and paleogeographic frameworks of the region during the Ediacaran period (intracontinental basin developed during the gravitational collapse of the Pan-African Orogen), lithofacies (a′a lava flows and lacustrine and fluvial deposits), the large volume of outcropping materials (480 km³ to 600 km³), and short duration of volcanic eruptions (9 Ma minimum to 23 Ma maximum) clearly point to a large igneous province (LIP) emplaced in a continental setting. Characteristics of the lava flows indicate predominantly subaerial eruption (e.g., a′a lava). Pyroclastic rocks contain juvenile, pyroclastic fragments with significant vesiculation (50 to 80% vesicles by volume which may suggest strombolian eruptions). Conglomerates and domal and crinkly laminated stromatolites are interbedded in the volcanic succession. These lithologies carry sedimentological features, indicating continental, fluvial, and lacustrine sedimentary environments.

5.2. Facies Analysis of the TVF

The goal of facies analysis is to reconstruct spatial and temporal variations in the geometry of volcanic, volcaniclastic, and sedimentary facies associations. It results in stratigraphic reconstructions, characterizing the volcanic succession over the entire volcanic region. This approach, developed by [111], aims to define on a regional scale a framework of volcanic processes at the surface, over a given eruptive cycle, in a given continental and/or marine environment, depending on the nature and type of the erupted products (e.g., mafic vs. felsic; lava flows, domes, pyroclastites), the relative volume and their spatial distribution, and the stratigraphic and lateral relative abundance of sedimentation style (e.g., fluvial, lacustrine, marine, etc.). Establishing a volcanic facies model based on this analysis allows comparison with other examples and integration into a geodynamic context of the studied volcanism.

During the 1st eruptive cycle, the Ediacaran volcanism of the TVF had the following volcanological characteristics: (1) eruptions occurred in a subaerial continental environment; (2) stratigraphic and lateral facies variations were abrupt; (3) non-bimodal volcanism was characterized by volumetric predominance of rhyolitic and andesitic magmas over basaltic, dacitic, and hybrid magmas; (4) deposition of mafic (basalts and andesitic basalts) and intermediate (andesites, dacites, and hybrid lavas) magmas occurred mainly in the form of lava flows, while sedimentation of felsic magma (rhyolites) was in the form of pyroclastic deposits (aerial fallout deposits, ignimbrites), domes, and rare lava flows of limited extension. Epiclastic deposits (conglomerates, sandstones, and siltstones) are mostly lenticular and limited in their extent. They indicate fluvial and/or lacustrine terrestrial sedimentary environments; (5) a very dispersed vent system was predominantly oriented parallel to the ESE–WNW fracture zone. Combined, these characteristics indicate a central volcano setting.

During the 2nd eruptive cycle, the Ediacaran volcanism in the TVF had the following volcanological characteristics: (1) eruptions also occurred in subaerial continental environments; (2) volcanism was characterized by the volumetric predominance of basaltic and andesitic magmas (basaltic andesites); andesites sensu stricto, dacites, and hybrid lavas were absent; (3) mafic eruptions occurred in the form of lava flows; (4) sedimentary deposits associated with this period are carbonates and siliciclastic deposits, indicating lacustrine and fluvial terrestrial environments. The fluvial sequences could have been very thick and of great lateral extent. These siliciclastic intervals thus reflect long periods of erosion between short-lived volcanic episodes; (5) contact of the basaltic succession with the underlying units is always marked by an angular unconformity; (6) volcanism occurred in an area located next to the margin of a continent, where a regional uplift was locally compensated by a subsidence, which occurred after volcanism cessation; (7) basaltic magmas have an anorogenic character and a tholeiitic affinity; (8) the closely spaced vent system was oriented parallel to the NE–SW structure, indicating a fissure-type volcanism.

During the late stage of the 2nd eruptive cycle, the Ediacaran volcanism of the TVF had the following volcanological characteristics: (1) eruptions also occurred in subaerial continental environments; (2) volcanism was characterized by the volumetric predominance of felsic magmas (rhyolites) expressed in the form of effusive rhyolitic bodies and a succession of deposits of pyroclastic flows (ignimbrites), which rest on various units of the underlying volcanic pile. The effusive rhyolitic bodies correspond to a short-distance lava flows (coulées) and domes associated with proximal deposits of pyroclastic flows and rhyolitic fall deposits.

The facies models that are the best match for the TVF are those associated with continental environments, namely: continental stratovolcanoes, continental basalt successions, and continental felsic volcanoes [111]. However, none of these models involves bimodal volcanism. Thus, the facies model for the TVF necessarily needs to combine these three facies models. The volcanological facies model of the TVF is that of a “continental stratovolcano” during the 1st volcanic cycle. The TVF then evolved to become compatible with the facies model of “continental basaltic successions” during the early stage of the 2nd eruptive cycle and finally with that of “continental felsic volcanoes” towards the end of the 2nd eruptive cycle.

5.3. Depositional Age of the TVF Volcanic Succession

The compilation of U-Pb zircon geochronological data for the igneous rocks of the Ouarzatate Group (i.e., CIMP) in the Central Anti-Atlas is summarized in Figure 12 and references to the publications where these ages were reported are compiled in Table S4.

The rhyolitic sample SI88 from the top of the 2nd eruptive cycle of the TVF (i.e., the top of the Ouazazate Group) yielded a weighted mean ²⁰⁶Pb/²³⁸U date of 557.613 ± 0.043 Ma (MSWD = 0.33, n = 5/6) on zircon analyzed by the CA-ID-TIMS method.

In the Siroua Window, on the northern margin of the Agadir Melloul Inlier, an ignimbrite sample yielded a weighted mean ²⁰⁶Pb/²³⁸U date of 575.0 ± 0.8 Ma (MSWD = 0.16, n = 3) on zircon analyzed by the CA-ID-TIMS method [33] and 572 ± 5 Ma on zircon analyzed by the SHRIMP method [24]: the sample OU72B of [33], described in [24] as a rhyolitic crystal tuff, TBOB647, from the base of the Adrar-n-Takoucht Formation (i.e., the basal part of the Ouazazate Group; Figure 13) that unconformably overlies the Wawkida Group (equivalent of the Tidilline Group of the Bou Azzer Inlier). Other samples from the same formation, OU71 and OU75 of [33], corresponding to well-stratified ash tuffs and massive ignimbrites, gave a weighted mean ²⁰⁶Pb/²³⁸U date 571.1 ± 1.1 Ma (MSWD = 1.4, n = 2) and 569.7 ± 0.7 Ma (MSWD = 0.82, n = 3; oldest dates) for zircons analyzed by the CA-ID-TIMS method, respectively.

The sample OU13 [33], a crystal-rich pyroclastic tuff with a vitroclastic texture (TBTB329; [24]) from the Tadoughast Formation (Figure 13), yielded a weighted mean ²⁰⁶Pb/²³⁸U date of 567.0 ± 1.5 Ma (MSWD = 1.6, n = 4) on zircon analyzed by the CA-ID-TIMS method and 565 ± 6 Ma on zircon analyzed by the SHRIMP method [24].

In the Agadir Melloul Inlier, close to the Siroua Window, the sample OU49 (red hyaline tuff with a greater proportion of shards than crystal clasts) and sample OU87 (massive grey ignimbrite; [33]) from the same Tadoughast Formation (Ait Hamd Section) yielded a weighted mean ²⁰⁶Pb^/238U date of 567.3 ± 1.2 Ma and a weighted mean 206Pb/238U age of 567.0 ± 1.5 Ma (MSWD = 0.64, n = 7) on zircon analyzed by the CA-ID-TIMS method, respectively. In the same section, an andesitic flow located below the massive ignimbrite of the Tadoughast Formation (Figure 13) yielded an age of 566 ± 6 Ma on zircon analyzed by the SHRIMP method (sample AMTB061) [24]. In the Jbel Iguiguil Section, the sample OU63 [33] from the same Tadoughast Formation yielded an age of 569.2 ± 1.7 Ma on zircon analyzed by the CA-ID-TIMS method.

A grey, welded massive lapilli tuff (sample AMTB065) gave a weighted mean age of 556 ± 5 Ma on zircon analyzed by the SHRIMP method [24]. This sample was collected in the Agadir Melloul Inlier near Jbel Iguiguil and belongs to the Fajjoud Formation. The lapilli tuff with a vitroclastic texture lacks lithic clasts and has only a few crystal clasts. The sample (located at the top of the unit) is stratigraphically equivalent to our rhyolitic sample SI88 at the top of the Ouarzazate Group and it yielded an age of 557.613 ± 0.043 Ma on zircon analyzed by the CA-ID-TIMS method.

Rhyolitic ignimbrites of High Irhiri (Siroua), corresponding to the base of the Ouarzazate Group, gave a Concordia zircon age of 580 ± 12 Ma by the ID-TIMS method [31]. This unit unconformably overlies the Amassine granite, which yielded a U-Pb age on zircon 661 ± 23 Ma by the SIMS method [31] and 601 ± 2 Ma and 599 ± 1 Ma by the SHRIMP method [112].

Rhyolites at the base of the Tikhfist Formation (Tiouin Subgroup) and the Aguins Member (Tawzzart Formation, Tafrant Subgroup), which are from the base and the top of the Ouarzazate Group volcanic succession, respectively, gave identical SHRIMP ages on zircon: 571 ± 8 Ma and 577 ± 6 Ma [19]. Both rhyolites belong to our 1st eruptive cycle of the TVF.

Zircon dates for the Zgounder rhyolitic domes, aligned E–W, cluster into four groups: (1) ca. 815 Ma date, corresponding to zircon’s inherited cores reworked from the Neoproterozoic crust formed during the Pan-African, 885–555 Ma orogenic cycle [27]; (2) 610 ± 7 Ma; (3) 578 ± 4 Ma, corresponding to the emplacement of multiple rhyolitic intrusions and the large magma input to the upper crust of the Pan-African Orogen during the late Neoproterozoic; and (4) 564 ± 15 Ma, a less-precise date obtained on hydrothermally altered zircon domains in U-rich magmatic zircons. The latter date is the best estimate for the timing of hydrothermal albitization/mineralization in the Zgounder epithermal Ag–Hg deposit [32].

In the Tifnoute Valley, south of the Jbel Toubkal peak, rhyolite lava flows of the volcano-sedimentary succession of the Tachakoucht Formation, which corresponds to the base of the Ouarzazate Group according to [13], yielded a Concordia zircon age of 574 ± 2 Ma ([113]; LA-ICP-MS zircon method). A rhyodacitic ignimbrite and andesitic tuff of the andesitic volcanic succession of the Toubkal Massif, corresponding to the top of the volcanic pile of the Ouarzazate Group, have been dated by the LA-ICP-MS method on zircon at 577 ± 3 Ma and 562 ± 1 Ma, respectively [113].

The giant batholith of the Askaoun granodiorite has been dated at 575 ± 8 Ma by the SHRIMP method on zircon [19] and at 574 ± 3 Ma by the LA-ICP-MS method on zircon [113]. The Askaoun granodiorite also yielded a younger date on zircon at 558 ± 8 Ma ([114]; LA-ICP-MS method). The Imourkhssen granitic pluton, intruding the Askaoun granodiorite, has been dated at 562 ± 5 Ma ([19]; SHRIMP method on zircon), 561 ± 3 Ma ([114]; LA-ICP-MS method on zircon), and 558 ± 1 Ma ([16]; LA-ICP-MS method on zircon). The Ougougane granite, which corresponds to the felsic dyke of the Douar Eç-çour dyke swarm, yielded a date of 550 ± 1.5 Ma [16,113,115].

U-Pb zircon dates in Figure 12 and Table S4 were obtained by laser ablation–inductively coupled plasma–mass spectrometry (LA-ICP-MS), secondary ion mass spectrometry (SIMS)/sensitive high-resolution ion microprobe (SHRIMP), and chemical abrasion–isotope dilution–thermal ionization mass spectrometry (CA-ID-TIMS) techniques. The accuracy and precision of in situ U-Pb zircon dates obtained by the LA-ICP-MS, SIMS, and SHRIMP methods usually do not satisfy the requirements for high-resolution correlations among widely distributed rocks (e.g., [116]).

Based on the CA-ID-TIMS U-Pb zircon dates, the age of magmatism during deposition of the Ouarzazate Group in the Siroua Window can be bracketed between 575 ± 0.7 Ma, the age of the Adrar-n-Takoucht Formation [33], which could be the lateral equivalent of the Adrar Tougmast Formation of the TVF, and 557.613 ± 0.043 Ma, the age of the Tizi-n-Bachkoun Formation, obtained in this study, which could correspond to the Fajjoud Formation in the Agadir Melloul Inlier near Jbel Iguiguil ([24], Figure 13). The duration of the volcanic activity of the Ouarzazate Group could be estimated to be around 18 Myrs. The angular unconformity between the Adrar Tougmast and Tizi-n-Bachkoun formations of the TVF must be older than the 561 ± 6 Ma SHRIMP zircon U-Pb age for the Fajjoud Formation [24]. The age of the unconformity is thus estimated to be 560 Ma. This is the same as the age of the angular unconformity between the Lower and Upper Ouarzazate groups of Saghro [29]. It is uncertain how much time was taken by the hiatus along this unconformity, but it clearly has bearing on whether the two volcanic cycles of the TVF are genetically related.

5.4. Petrogenesis of the TVF

5.4.1. Fractional Crystallization vs. Crustal Contamination

The studied volcanics range in composition from ultramafic (2nd cycle) and mafic (1st cycle) to felsic, clearly indicating magma evolution through time.

Incompatible trace element ratios are largely unaffected by partial melting and fractional crystallization processes ([117] and references therein). The Zr vs. Nb diagram (Figure 14) shows that, except for rhyolites and ignimbrites of the 1st cycle, all the other lithologies of both cycles are characterized by Zr/Nb ratios of ≈20, suggesting that they are comagmatic, i.e., that they are derived from magmas generated from a common mantle source and later affected by closed-system magmatic evolution processes. The rhyolites and ignimbrites of the 1st cycle, having significantly lower Zr/Nb ratios (as low as 6.4), are clearly not comagmatic with other studied lithologies from the TVF in the Siroua Window.

Even considering mafic rocks from both cycles (see Section 4.3.2 and Section 4.3.3), the effects of magma evolution processes are evident. Indeed, taking into account Mg# up to 0.61 (1st cycle) and 0.50 (2nd cycle), even the most primitive rocks of the TVF are clearly not compatible with magmas generated in equilibrium with residual mantle mineralogy, i.e., primary magmas (Mg# = 0.68 to 0.75). This interpretation is also supported by low concentrations of transition trace elements (e.g., Ni ≤ 68 ppm; Cr ≤ 311 ppm) as compared with those typical for primary magmas (Cr ≥ 1000 ppm, Ni ≥ 400–500 ppm, Co ≥ 50–70 ppm; [118,119,120]; see Figure 15).

Thorium, as a highly incompatible element, can be used as an index for both partial melting and fractional crystallization processes [122,123,124,125,126]. Figure 16 shows inverse relation between Th and compatible elements, suggesting fractional crystallization rather than partial melting as the main cause of the chemical variability in the studied TVF volcanic rocks. Indeed, transition elements are known for their high compatibility with the ferromagnesian silicates (olivine and pyroxene) that first crystallize from mafic magmas [127], resulting in the lowering of their concentrations as fractional crystallization advances. For the chemical variation of mafic rocks, the role of plagioclase fractionation was also important as evidenced by Sr and Eu negative anomalies (see Figure 8), which are accentuated in the felsic rocks (Figure 10). The felsic rocks are, in addition, characterized by significant Ti negative anomalies, whose meaning will be discussed later.

The diagrams of Figure 16 also show the role of partial melting in the generation of mafic magmas in the 2nd eruptive cycle.

One of the most striking characteristics of the mafic rocks of the TVF is the presence of pronounced Nb and Ta negative anomalies. The Nb depletion is also supported by the Nb/U ratios (Figure 17), which are clearly below the value of 47 ± 10 of MORB and OIB [128,129], i.e., of the oceanic basalts generated away from the influence of subduction processes.

Since Nb and Ta are highly immobile in aqueous fluids generated by dehydration of subducting slabs [131,132], supra-subduction-related magmas are typically characterized by negative Nb and Ta anomalies [133], which translate, for example, to average Nb/U ratios of about 4.6 [134]. However, considering that the continental crust grows at the expense of arc magmas [135,136,137], an unavoidable consequence is that the Nb and Ta negative anomalies are also a characteristic of continental crust [138]. This makes it difficult to distinguish between the supra-subduction origin of the studied mafic rocks and the effects of basaltic magmas’ contamination with crustal materials. However, at odds with arc magmas, continental crust is characterized by a P depletion, resulting in large negative anomalies (so called the “crust composition paradox” [139]). The lack of significant P negative anomalies, in association with strong negative Nb and Ta anomalies, is an indication of a metasomatized mantle enriched by subduction-related fluids rather than contamination by continental crust.

The fact that, except for the rhyolites and ignimbrites of the 1st cycle, all the other lithologies have similar Zr/Nb ratios (≈20) (Figure 14) suggests that even for the most evolved felsic rocks of the 2nd cycle the effects of crustal contamination were negligible. In this perspective, significant negative Ti anomalies shown by felsic rocks of the 2nd cycle are here linked to the crystallization of Ti-bearing oxides during magma evolution. The same conclusion can be extracted from the Zr/Hf ratios. Indeed, some of the felsic rocks of the 1st cycle are characterized by Zr/Hf values clearly lower than the chondritic value (≈38) presented by the 2nd cycle mafic to felsic rocks and by the mafic lithotypes of the 1st cycle.

5.4.2. Magma Source Composition and Magmatic Processes

From the above discussion, the source-inherited negative Nb and Ta anomalies must have been retained, which strongly points to a supra-subduction mantle source, metasomatized by aqueous fluids derived from a dehydrating subduction slab. This is further supported by the high normalized LILE/HFSE ratios for the mafic magmas (see Figure 8) and by the calc-alkaline characteristics of the mafic magmas inferred from clinopyroxene mineral chemistry (see Section 3.4; Figure 11).

As previously shown, even considering only the mafic rocks, the effects of fractional crystallization are evident. Olivine, clinopyroxene, and plagioclase were probably the dominant phases during the initial stages of crystallization. These phases do not significantly alter ratios involving elements incompatible with them, like Ce/Yb and La/Ta [105]. This allows the use of such ratios even for rocks with Mg# down to 30 to characterize signatures inherited from magma source processes. For values of Mg# inferior to 30, such ratios increase significantly as a result of magma evolution processes.

The Ce/Yb ratio of mantle magmas reflects melting pressure/depth, where high values indicate garnet involvement, owing to a significantly higher garnet distribution coefficient for Yb relative to Ce, at odds with that observed for the other common mantle residual phases, not capable of such significant Ce-Yb fractionation [140,141,142]. Consequently, partial melting of garnet peridotite yields Ce/Yb ratios exceeding 20 [143], while partial melting of spinel peridotite results in lower magma Ce/Yb ratios. Additionally, the distinction between asthenospheric and lithospheric magma sources can be informed by the La/Ta ratio [144,145,146]. Igneous rocks with La/Ta values < 22 are partial melts and inferred to be from asthenospheric sources with minimal or no lithospheric contamination, while [146] suggested that La/Ta ratios ranging from 22 to 30 might indicate lithospheric mantle contamination. The diagram Ce/Yb vs. La/Ta (Figure 18) suggests magma genesis close to the garnet–spinel peridotite zones mainly into the lithosphere.

The experimental work of [140,147] on the CMAS system indicates that the spinel–garnet transition occurs at ~2.4–2.5 GPa and 1440 °C, despite being dependent on peridotite fertility [148]. The positioning of TVF mafic rocks in Figure 18 suggests that primary magmas for most of the samples were generated at pressures around 2.5 GPa.

Since the two eruptive cycles are considered comagmatic here, as inferred from similar Zr/Nb ratios, the lower Th concentrations of the second eruptive cycle mafic rocks, compared to rocks with a similar degree of evolution of the first eruptive cycle (Figure 16), likely indicate a higher degree of partial melting.

This view is endorsed by the lower LREE concentrations of the 2nd cycle mafic rocks. Moreover, the variation in the extent of melting is also observed for a single cycle. Indeed, considering the mafic rocks of the 2nd cycle, a large range of Th contents, in association with high concentrations of compatible elements (Figure 16), suggests multiple and variable degrees of partial melting magmatic events. In summary, if most of the chemical variability in the studied rocks is due to fractional crystallization (see Section 5.4.1), a variable extent of partial melting events also contributed to the heterogeneity observed.

5.5. Tectonic Setting

The geochemistry of mafic rocks and of clinopyroxenes can be used, with caution, to constrain the tectonic setting that prevailed at the time of their emplacement (e.g., [134] and references therein). The TVF exposed in the Siroua Window of the Central Anti-Atlas belongs to the Ouarzazate Group ([11,19,24,31,32,33], this work). Two eruptive cycles are distinguished in the TVF, according to their volcanologic characteristics and geochemical signatures. The mafic rocks of both eruptive cycles carry pronounced Nb and Ta negative anomalies (see Section 5.4.1) and are characterized by high LILE/HFSE ratios (Figure 8), both interpreted as a signature of a magma source metasomatized by aqueous fluids associated with dehydration of a subduction plate. These characteristics correspond to the positioning of the studied mafic rocks in the tectonic setting discriminant plots of Figure 19 (see also Figure 11), where an orogenic (s.l.) setting is suggested. The tectonic setting of the TVF will be further refined taking into account their age and location of emplacement (see Section 5.6.2).

5.6. Geodynamic Implications

The Pan-African Orogeny corresponds to a series of major Neoproterozoic mountain-building events that contributed to the formation of the Gondwana and Pannotia supercontinents about 600 million years ago (Figure 20).

This orogenic cycle spans some 300 million years (most of the late Neoproterozoic), which is comparable to those during the assembly of the Columbia/Nuna and Rodinia (Grenville) supercontinents. This extended timeframe suggests that the Pan-African orogenic cycle consists of multiple orogenic episodes.

For comparison, the Cordilleran/Andean Orogeny, along the western margin of the Americas from Alaska to Chile, has developed over the past 380 million years, a duration comparable to that of the Pan-African orogenic cycle. However, the Cordilleran/Andean Orogeny includes six distinct episodes, each lasting between 30 and 90 million years, such as the Antler (380–340 Ma), Sonoma (270–240 Ma), Nevada (180–140 Ma), Sevier (140–50 Ma), Laramide (70–35 Ma), and the ongoing Cascade and Andean orogenies. These episodes reflect different geodynamic settings, such as flat-slab subduction and collision with volcanic arcs and oceanic plateaus. The Cordilleran has been characterized as an accretionary orogeny [154], in contrast to the Pan-African Orogeny, which was primarily a collisional orogeny until the Ediacaran, when it was followed by the Cadomian accretionary orogenesis [155].

5.6.1. Previous Geodynamic Models

The Anti-Atlas Belt of Morocco is a key region for understanding Proterozoic accretion and break-up events along the margins of the West African Craton (WAC), given that a significant Pan-African Orogen segment is prominently exposed there. The Anti-Atlas Belt is considered to be a part of a 6000 km long Neoproterozoic “ring of fire” along the northern margin of the WAC [83]. This area contains extensive outcrops of the Neoproterozoic Pan-African Belt and the underlying Paleoproterozoic foreland, which have been intensely studied over many years (e.g., [13]). Traditionally, the Precambrian formations of the Anti-Atlas Belt have been divided into three major lithosomes (e.g., [13]): PI (Archean to Paleoproterozoic), PII (early and middle Neoproterozoic), and PIII (late Neoproterozoic) (Figure 1). However, this lithostratigraphic classification has recently been challenged [19,28,29,40,43,44,45,156].

In the new interpretation, the oldest Paleoproterozoic (PI) rocks in the Anti-Atlas Belt are divided into several complexes, such as the Zenaga and Kerdous (in the Siroua and Kerdous basement inliers). The overlying PII lithosome deposited on this basement is collectively referred to as the Anti-Atlas Supergroup, which consists of four volcano-sedimentary units (Lkest-Taghdout, Bou Azzer, Iriri, and Saghro), considered to be Neoproterozoic in age. Several intrusive suites (Ifzwane and Toudma) are also recognized, although they are still poorly dated. These volcano-sedimentary units have been associated with the early passive margin, oceanic, and island-arc environments, with ages ranging from 885 to 640 Ma. The earliest Pan-African deformation (660–640 Ma) is attributed to ocean basin closure, southwest-directed thrusting, and accretion of island-arc remnants [19,28,29,40].

Recent studies have provided new structural, geochronological, and geochemical data (e.g., [22,157]) and references therein). However, many critical questions about the accretionary history of the Anti-Atlas Belt are still debated (e.g., [157] and references therein). Such questions include the origin and tectonic framework of the Upper Ediacaran volcano-sedimentary cover series (i.e., the Ouarzazate Group), which unconformably overlies the deformed Pan-African units and immediately precedes the Ediacaran–Cambrian deposits that record the Basal Cambrian Carbon Isotope Excursion (BACE) [49].

Several models have been proposed for the origin and tectonomagmatic evolution of the Pan-African Anti-Atlas Belt (e.g., [12,19,29,30,83,158,159,160,161,162,163,164,165,166,167,168,169]). Previous models diverge on several aspects, particularly the early stages in the geodynamic evolution of the Pan-African Orogen. It is still unclear whether the subduction zone associated with accretion of an island arc dipped towards the north or towards the south. On the other hand, all these models (with the exception of the scheme of [160]) agree in that they consider the volcanism at the end of the Neoproterozoic as post-collisional (late-to-post-orogenic) and they link its origin to subduction of oceanic lithospheric plates; obducted ophiolites of Bou Azzer and Siroua are preserved remnants of these plates. The extensional stage (ca. 606–545 Ma) in the aftermath of the collision was marked by molasse deposition in a successor basin and igneous events (represented by the Ouarzazate Group), followed by the marine transgression recorded by the Taroudant Group, which straddles the Ediacaran–Cambrian boundary.

5.6.2. Towards a New Geodynamic Model for the Ediacaran Magmatism in Morocco

The geological and geochronological data for the Ediacaran volcanic rocks from the TVF of the Siroua Window in the Anti-Atlas Belt indicate that the volcanism occurred at ca. 575–557 Ma (based on TIMS zircon U-Pb data; [33] and this study). Considering that the collisional processes in the Pan-African Orogen may have occurred at approximately 647 Ma [34], such ages confer a post-collisional setting to the studied TVF. Volcanic eruptions occurred in a continental setting [16,24,32,33,113] in two successive, eruptive cycles distinguished according to their volcanological characteristics and geochemical signatures (see above).

The post-collisional timing of the TVF volcanism impedes its genetic relation with coeval active subduction event(s). This post-collisional magmatism suggests that an orogenic collapse following collision led to gravitational and thermal instability of the lithosphere, inducing magmatism [3]. Such melting events would be facilitated by the existence of low-solidus metasomatized lithospheric domains imprinted by fluids generated during previous subduction processes (e.g., [170]). This scenario explains the geochemical characteristics of the TVF, namely the Nb-Ta negative anomalies and the high LILE/HFSE ratios (Figure 8).

In general, two models have been proposed to explain post-collisional magmatism associated with an orogen collapse: lithospheric delamination and convective erosion [3,4,171,172,173,174,175,176,177], both leading to a reduction in lithosphere thickness. This lithospheric thickness reduction/thinning is also imperative in the present case. Indeed, some melts were generated at the spinel–garnet transition, at pressure close to 2.5 GPa (see Section 5.4.2). Considering a crustal thickness of about 40 km and mean density of the crust and lithospheric mantle of 2700 kg/m³ and 3200 kg/m³, respectively, such pressure would correspond to ~84 km in depth. This depth is likely significantly shorter than the thickness of the lithosphere in the aftermath of collision, thus requiring significant thickness reduction in order to make it feasible for lithospheric melting at ~84 km in depth by interaction with the hotter asthenosphere. The oldest mafic TVF magmatism is dated at 575 Ma, i.e., some 72 Ma after the end of collision [59]. This suggests a long-lasting process of lithospheric thinning (s.l.) before melting started.

No single model can explain all the events in the Anti-Atlas collisional belt. Figure 21 illustrates our proposed model for the unroofing and collapse in the collisional orogen of the Pan-African Anti-Atlas Belt: (1) oceanic crust subduction between ca. 777 and 760 Ma ago released fluids from a dehydrating subduction plate [59] that metasomatized the overlying asthenosphere and lithosphere mantle. (2) Continental collision between the West African Craton and the northern block led to continental subduction and subsequent exhumation with syn-collisional magmatism at ca. 650–640 Ma (e.g., [178]). (3) As a consequence of orogenic collapse, the lithospheric mantle has been thinned. (4) A reduction in the lithospheric thickness and increase in heat transfer from the asthenosphere with magmatism started at ca. 575 Ma and continued for ca. 15 million years (until ca. 560 Ma). (5) Associated delamination of the lithospheric mantle and consequent isostatic adjustment ended with complete leveling of the orogen and deposition of the shallow-marine, mixed siliciclastic–carbonate sequence of the late Ediacaran–early Cambrian Taroudant Group, which marks the end of the Pan-African orogenic cycle.

The delamination model suggests that the dense base of an orogenically thickened continental lithosphere detached and sank into the asthenosphere. Numerical modeling [171] suggests that a tectonic event could have initiated an elongated conduit connecting the hot asthenosphere to the crustal base, creating a localized thermal anomaly that facilitates voluminous crustal melting. This process may explain formation of the felsic rocks from the first eruptive cycle. However, felsic rocks of the second cycle, with Zr/Nb ratios similar to the associated mafic rocks, are more likely to result from fractional crystallization rather than crustal melting.

In the Siroua Window (Central Anti-Atlas), lithospheric instabilities could have occurred near the end of early Pan-African subduction (ca. 762 Ma ago) due to slab break-off or rollback, potentially coupled with the gravitational collapse of the Central Anti-Atlas block. While delamination is a widely discussed mechanism for forming large igneous provinces [179], applying this model to the Pan-African Belt requires further careful consideration of its temporal and geochemical complexities [177].

6. Concluding Remarks

1. According to their volcanological, stratigraphical, and geochemical characteristics, two consecutive eruptive cycles can be distinguished in the Taghdout Volcanic Field (TVF) of the Ouarzazate Group exposed in the Siroua Window (Anti-Atlas Belt).

2. The precise geochronological dating of the FVF rocks (575 Ma to 557 Ma) allows for assigning them to a post-collisional stage in the Pan-African Orogen evolution.

3. Mafic subalkaline (calc-alkaline) magmas clearly show chemical characteristics indicative of sources metasomatized by fluids derived from dehydrating subduction slab(s), likely in the aftermath of the Pan-African Orogeny.

4. Mafic rocks that erupted over a period of ca. 18 Ma evolved from primary magmas originated by melting of a peridotitic lithospheric mantle during collapse of the Pan-African Orogen. Chemical composition of post-collisional mafic magmatism evolved systematically over time with the mafic rocks of the first eruptive cycle (ca. 575 to 560 Ma) being derived from magmas with a lower degree of partial melting than those of the second cycle (ca. 557 Ma), suggesting an increased heat transfer from asthenosphere associated with its upwelling.

5. Our model implies a significant reduction in the lithospheric thickness during the orogenic collapse, probably involving a long-lived, sublithospheric thermal erosion, followed by a rapid delamination causing a large-scale, isostatic adjustment. Afterwards, surface relief diminished and Ediacaran–Cambrian, mixed siliciclastic–carbonate sediments covered the craton, recording the end of the Pan-African orogenic cycle, the Ediacaran–Cambrian transition, and the Basal Cambrian Carbon Isotope Excursion (BACE).