The Harz Mountains (Germany) – Cadomia meets Avalonia and Baltica: U–Pb ages of detrital and magmatic zircon as a key for the decoding of Pangaea's central suture Open Access

Supplementary material: Zircon U–Pb LA ICP-MS age data and method description are available at https://doi.org/10.6084/m9.figshare.c.6873591

Since the supercontinent Pangaea was defined for the first time (Wegener 1915), its configuration and outer shape were improved and perfected over the last decades, mainly by making use of a combination of palaeomagnetic, geochronologic, palaeoclimatic and palaeontological data (e.g. McKerrow and Scotese 1990; Torsvik and Cocks 2017). However, Pangaea's interior belts and sutures, which are related to its Devono-Carboniferous amalgamation, are still much debated. One of the most important mobile belts in the centre of Pangaea is the Alleghenian–Variscan Orogenic Belt. The orogen was first defined by Eduard Suess (1888) as ‘Variscisches Gebirge’, named after Curia variscorum, an area around the present city of Hof in Germany, where the German tribe of the Varisker lived. Even after more than 130 years of geological research and modelling, the belt is still enigmatic and many details are not understood. Many contrasting models exist for the Central European Variscides and its Avalonian–Cadomian precursors (e.g. Franke 2000; Murphy et al. 2006; Kroner et al. 2007; Murphy and Nance 2008; Nance and Linnemann 2008; Schulmann et al. 2009; Linnemann et al. 2010, 2014; Zeh and Gerdes 2010; van Staal et al. 2012; Kroner and Romer 2013; Franke et al. 2017). Hypotheses and models range from a stepwise terrane accretion onto the Laurussian margin and a partial re-opening of a narrow oceanic basin and the formation of an intermediate terrane (e.g. Franke 2000, Franke et al. 2017; Zeh and Gerdes 2010) to a two-plate model combined with an indenter (Kroner et al. 2007) named the ‘Armorican Spur’ (Kroner and Romer 2013). In this paper, we analyse the nature of the main Variscan suture in Central Europe using a dataset of detrital and magmatic zircon. The suture was formed during the amalgamation of the supercontinent Pangaea. We focus on the Harz Mountains in Germany, situated on the southeastern margin of the Rheno-Hercynian Zone (Fig. 1) and close to one of Pangaea's most important central sutures, the northern margin of the Mid-German Crystalline Rise (inset in Fig. 2). The main basis for our interpretations is a robust dataset of U–Pb LA-ICP-MS ages of detrital and magmatic zircon from the Harz Mountains. In addition, we have produced a dataset of U–Pb ages of sedimentary rocks from selected areas of the main continental blocks (Baltica, East Avalonia, West Africa), which were involved directly or indirectly in Variscan orogenic processes. Such data allow us to establish zircon provenance, and hence reconstruct sediment transport directions, provenance and the plate tectonic interaction of continents.

The Harz Mountains (German: Harz) form a mountain range and an important geological element of Central Europe (Figs 1 & 2). The Brocken is the highest mountain in the range (1141.2 m above sea level). It consists of a Variscan granite of the Brocken Plutonic Complex. During the Cold War, the border dividing East from West Germany ran through the mountain. The area has a great importance in the history of mining development, mainly for copper, iron, lead and silver since the Bronze Age. The Harz Mountains are ‘holy ground’ for the development of mining geology, melting technology, geological mapping and geological models. For these and many other reasons, the Harz Mountains are part of the UNESCO-Global Geopark Harz · Braunschweiger Land · Ostfalen.

The Harz Mountains are one of the type regions for the Hercynian Orogeny, a denomination often used in the Anglo-American geological literature instead of Variscan or Alleghenian–Variscan orogeny. Important contributions came from Lossen (e.g. 1869), who was one of the most famous mapping geologists of his time working in the Harz Mountains. Important steps in the frame of geological research were the establishment of a robust biostratigraphic dataset and lithostratigraphic subdivisions during the last 150 years, as summarized by Wachendorf (1986), Mohr (1993), Schwab (2008), Schwab and Ehling (2008a, b) and Schwab and Hüneke (2008).

The entire Paleozoic volcano-sedimentary rock sequence of the Harz Mountains was deformed into a strongly segmented fold and thrust belt during the final Variscan collision. Deformation and metamorphic overprint increased from diagenetic grade in the northwestern to greenschist grade metamorphism at the southeastern margin of the Harz (Ahrendt et al. 1983). Maximum temperatures of 300–350°C and pressures of 3–6 kb indicate subduction-related metamorphism (Siedel and Theye 1993; Massonne 1995). Franke et al. (2019) suggested the existence of a paired metamorphic belt at c. 380 Ma. Franke (2000) described the Harz Mountains as part of the Gießen-Werra-Harz-Gommern Nappes, which show strong parallels to the Lizard Ophiolitic Complex in Cornwall (UK).

The tectonic framework is characterized mainly by NW-facing folds and thrusts in the northwestern part and increasing strain towards the SE, where the Harz Mountains expose a complex association of thrust-bounded slices, imbrication structures, listric overthrusts and sheared lenses. The sequence often contains rock fragments of different lithology, size and/or age in a sheared shaly matrix. The northwestern margin of the thrust-bounded slices is marked by the Acker-Bruchberg Unit.

First nappe conceptions (Schriel 1925; Dahlgrün 1928) were picked up and modified by Schriel and Stoppel (1958), Walliser and Alberti (1983) and Franke (2000). Reichstein (1965) interpreted parts of the sheared block-in-matrix fabrics as tectonically overprinted submarine debris flows with pebbly mudstones, olistostromes and sliding nappes. Other authors followed this interpretation of sedimentary mixing (e.g. Lutzens 1972, 1991; Schwab 1976; Günther and Hein 1999).

The dimensions and ages of the suggested olistostromes are controversial. Lutzens and Schwab (1972) proposed olistostromes to be deposited in Lower Carboniferous time after greywacke deposition. The authors suggested a widespread distribution of olistostromes in the central and southeastern part of the Harz Mountains, with an overall thickness of more than 2000 m (Schwab 1980, 2008; Borsdorf et al. 1992; Herbig et al. 2018). Other authors (e.g. Reichstein 1970; Lütke 1978; Buchholz et al. 1990) suggested only smaller parts of the sequence to be locally re-deposited at various times prior to flysch sedimentation.

In contrast to the olistostrome model, other authors maintained the former pure tectonic conceptions and interpreted the block-in-matrix fabric as tectonic slices in extensive shear zones (Koll 1984; Wachendorf 1986; Friedel et al. 2019). Thin- and thick-skinned tectonics for large parts of the Harz Mountains were suggested by Wachendorf et al. (1995), Franke (2000) and Franzke (2001). Huckriede et al. (2004) suggested that the allochthonous units in the southeastern Harz Mountains form a duplex-like structure sandwiched between the Gießen-Harz Nappe and parautochthonous units at its base.

Radiometric ages in the Harz Mountains are scarce. For the granitoids of the Brocken Plutonic Complex, the gabbro-noritic Harzburg Plutonic Complexes and the Oker granite, thermal ionization mass spectrometry zircon ages were published pointing to intrusion ages in a time span of c. 293 to 295 Ma (Baumann et al. 1991). Zech et al. (2010) reported LA-ICP-MS zircon ages, which gave intrusion ages of 283 ± 2.1 Ma for the Brocken Granite and 283 ± 2.8 Ma for the Ramberg Granite. K–Ar ages (bulk) on detrital white mica were performed by Huckriede et al. (2004). The authors distinguished detrital input from various sources such as Cadomia, the Caledonian Orogen and early Variscan sources. In addition, very few LA-ICP-MS U–Pb analyses on detrital zircon were published. Meinhold et al. (2016) report results from one quartzite sample from the Wippra Unit, suggesting a late Devonian to Lower Carboniferous age for the uppermost lithostratigraphic section (Fütterungsberg-Metagrauwacke Formation). Three detrital samples from the Visean aged tectono-sedimentary Hüttenrode Mélange overlying the Blankenburg Unit point to a mixed provenance from Baltica and Cadomia (Zweig et al. 2016).

The Central European Variscides have a general NE–SW strike and are subdivided from the NW to the SE into the Rheno-Hercynian, Saxo-Thuringian and Moldanubian zones (Kossmat 1927; Fig. 1). In addition, the Mid-German Crystalline Rise was introduced by Brinkmann (1948), and placed between the Rheno-Hercynian and Saxo-Thuringian zones. The Saxo-Thuringian and Moldanubian zones form part of Cadomia (Fig. 1) (Linnemann et al. 2014). The latter is characterized by the Ediacaran aged Cadomian basement and forms a large peri-Gondwanan crustal unit that stretches from the Bohemian Massif to Iberia (Fig. 1) (Linnemann et al. 2018). Above a Cadomian unconformity, Cambro-Ordovician to Lower Carboniferous sedimentary deposits cover major parts of the Saxo-Thuringian Zone (Linnemann et al. 2004). In contrast, the Rheno-Hercynian Zone is underlain by an East Avalonian basement of Ediacaran age (e.g. Wartenstein Gneiss; Dörr and Stein 2019) and a Cambro-Silurian cover sequence (Linnemann et al. 2012; Herbosch et al. 2020). A ‘Caledonian’ unconformity reflects a deformation event at c. 430–425 Ma (Herbosch et al. 2020). The unconformity is overlain by Upper Silurian to Lower Carboniferous strata of the Rhenish Massif (Walter 2006). Avalonia docked with Laurussia at c. 430–420 Ma (van Staal and Zagorevski 2012; van Staal et al. 2012; Torsvik and Cocks 2017). That collision caused the ‘Caledonian’ unconformity and placed East Avalonia at the southern margin of Baltica (Fig. 1).

The Harz Mountains form a WNW–ESE striking range at the southeastern margin of the Rheno-Hercynian Zone (Fig. 1). To the SE, the Harz Mountains are fault-bounded against the Mid-German Crystalline Rise (Franzke and Schwab 2011) (Fig. 2, inset). Internal structures show a pronounced NE–SW strike typical for the Variscan Orogen (Fig. 2). As summarized in Wachendorf (1986), Mohr (1993) and Schwab (2008), the Harz Mountains are subdivided traditionally into several tectonostratigraphic units with an individual lithostratigraphic subdivision (Figs 2 & 3). All tectonostratigraphic units except the Clausthal Unit (Oberharz) are rootless, which means having no observable contact with an underlying basement. Instead, it is suggested that all units except the Oberharz (NW of the Acker-Bruchberg Unit) are allochthonous and rest on thrusts (Franzke 2001; Franzke and Schwab 2011).

The Wippra Unit is the southeasternmost segment and fault-bounded at its southeastern margin against the Mid-German Crystalline Rise. The entire unit forms a steeply NW-dipping package of siliciclastic rocks and chert metamorphosed under greenschist facies conditions (Reichstein 1964; Franzke 1969; Schwab 1976). According to Ahrendt et al. (1996), K–Ar ages on metamorphic white mica scatter between c. 360 and 320 Ma and indicate a Variscan age for metamorphism. The lower strata form part of the Wippra Group (Fig. 3), which is composed, from the base upwards, of the Klippmühle Quartzite Formation (mainly greenish quartzite), the Biesenrode Formation (red phyllite and schist) and the Rammelburg Formation (phyllite) (Schwab and Ehling 2008b). For the Klippmühle Quartzite Formation, acritarchs establish its Arenigian (now Floian to Dapingian) age (Sehnert 1991). Acritarchs of Llanvirnian (Dariwillian) age were described from the Rammelsburg Formation (Burmann 1973). These Ordovician rocks have Gondwanan biogeographic affinities (Reichstein 1964; Sehnert 1991). A younger and fault-bounded segment of the Wippra Unit is represented by the Piskaborn Group (Fig. 3). The lowermost lithostratigraphic unit is the Pferdeköpfe Formation (green schist and minor quartzite), followed by an unnamed package of chert and coloured schist and shale (‘Buntschiefer’). The Fütterungsberg Metagrauwacke Formation (mainly greywacke) forms the youngest part of the group (Fischer 1929; Schwab and Hüneke 2008). A Devonian age is suggested for the Piskaborn Group based on lithostratigraphic correlation (Schwab and Hüneke 2008). An Upper Devonian to Lower Carboniferous age is possible for the uppermost part of the succession (Fütterungsberg Metagrauwacke Formation) (Meinhold et al. 2016). The Wippra Unit is generally steeply dipping and is limited by a major thrust fault at its northwestern margin (Fig. 2).

The next tectonostratigraphic segment NW of the Wippra Unit is represented by tectono-sedimentary mélanges, in which fragments and giant blocks of Ordovician to Lower Carboniferous age (often surrounded by a greywacke matrix) occur (Huckriede et al. 2004). In this paper, these mélanges are summarized as the Harzgerode Mélange Complex (Figs 2 & 3). The complex covers the largest parts of the eastern Harz Mountains (Fig. 3). Friedel et al. (2019) favour a pure tectonic origin for the mélanges. In our view, the mélanges contain both sedimentary and tectonic features and are classified here in terms of tectono-sedimentary mélanges. The Harzgerode Mélange Complex consists of an up to 1200 m thick pile of chaotic sediment with a sheared shaly to arenaceous matrix containing fragments of Silurian cherts and graptolite shales (Maronde 1968) associated with sandstones, mafic volcanic fragments, Lower Emsian calcareous sandstones, Devonian and Silurian neritic and pelagic limestones (Hercynian limestones; e.g. Hüneke 1998; Huckriede et al. 2004), bedded and in part silicified turbiditic limestones (‘Flinzkalke’) of Middle Devonian to Early Carboniferous age (Lütke 1978; Buchholz et al. 1991) and locally earliest Carboniferous calcareous conglomerates as the youngest component (Buchholz and Luppold 1990). From the Silurian to the Early Emsian, the faunas of the Hercynian limestone exhibit Armorican (Gondwanan) affinities (Franke 2000). Clasts of crystalline rocks such as granitoids or gneisses have not yet been reported (Schwab and Hüneke 2008). The Harzgerode Mélange Complex has often been interpreted as an olistostrome with various exotic olistoliths (Reichstein 1965; Schwab 1976; Lutzens 1979; Schwab and Ehling 2008a) (Fig. 3). Other authors interpret the Harzgerode Mélange Complex as the southeastern part of a large duplex-structure sandwiched between the Gießen-Werra-Harz-Gommern Nappe and other parautochthonous units (Franke 2000; Huckriede et al. 2004). The Harzgerode Mélange Complex forms a nappe, which is thrusted top-NW onto the Tanne and the Blankenburg units (Fig. 3). The core of the drilling ‘Götzenteiche’ (Fig. 3) shows a 1200 m thick Harzgerode Mélange Complex, which overlies greywackes of the Tanne Unit with a tectonic contact (Lutzens 1978; Lippert 1999). Further, the Harzgerode Mélange Complex is extended onto the Blankenburg Unit, where sub-units of the mélange complex (Hüttenrode, Bodetal and Harzrand mélanges; Lutzens and Schwab 1972) tectonically override volcano-sedimentary successions of the Blankenburg Unit, such as the Elbingerode Complex (Figs 2 & 3). The Harzgerode Mélange Complex is thrusted as a nappe for at least 50 km towards northwestern directions (Fig. 2).

The Südharz-Selke Nappe (Reichstein 1965) has a northern (Selke Nappe) and a southern (Südharz Nappe) segment (Fig. 2). The largest part of the Südharz-Selke Nappe is composed of the Middle to Upper Devonian sedimentary deposits of the Südharz-Selke Formation (quartzite, chert, shale, greywacke; Tschapek 1990). The upper part of the succession, the Südharz-Selke Greywacke Member, consists mainly of thick flysch sequences of Famennian age (Tschapek 1989, 1990; Ganssloser 1996). The Südharz-Selke Nappe overlies, separated by thrust faults, the Harzgerode Mélange Complex and the Tanne Unit (Fig. 2). At the bottom of the nappe, the tectono-sedimentary Stiege Mélange occurs, which consists of highly strained Emsian–Frasnian cataclasites covering condensed grey and coloured shales, radiolarian cherts, locally a Visean ‘wildflysch’ package and strongly sheared metabasalts. The latter exhibit both mid-ocean ridge basalt (MORB)-type (Platen 1991) and within-plate basalt compositions (Ganssloser et al. 1995). The Südharz-Selke Nappe is generally accepted as a thrust sheet (Schwab 1976, 1979) and an equivalent of the Giessen Nappe (Weber 1978; Walliser and Alberti 1983; Franke 2000). The Südharz-Selke Nappe is thrusted towards the northwestern for about 40 km (Fig. 2). Lithologies and stratigraphic ages of the Südharz-Selke Formation are very similar to the succession of the Piskaborn Group in the Wippra Unit (Fig. 3). Spatial arrangement and ages suggest a correlation of both units and a derivation of the Südharz-Selke Nappe out of the Wippra Unit (see later discussion).

Parts of the Blankenburg units were covered by a tectonic contact from the Tanne Unit, although outcrop is very poor. Here, the Tanne Unit is interpreted to form a nappe because of its sigmoidal and frayed outcrop shape, which is consistent with very low angle sole thrusts (Fig. 2). The drill core of the drilling ‘Götzenteiche’ shows that rocks of the Tanne Unit are overlain by the Harzgerode Mélange Complex. Both units are separated from each other by a thrust fault (Schwab 2008). From bottom to top, the Tanne Unit, the Harzgerode Mélange Complex and the Südharz-Selke Nappe form a nappe stack. The Tanne Unit is lithologically described in detail by Lippert (1999), and here interpreted as a mélange, which includes a little occurrence of Upper Silurian to lowermost Devonian graptolite shale and large units of Upper Devonian chert and greywacke (including minor red shale), followed by Tournaisian radiolarites and cherts (Culm Kieselschiefer Formation), which are overlain by Visean aged, very distal shaly turbidites (Culm Tonschiefer Formation, ‘Plattenschiefer’) and more proximal flysch deposits (Culm Grauwacke Formation) (Buchholz et al. 1990; Ganssloser 1996; Lippert 1999; Schwab and Ehling 2008a; Schwab and Hüneke 2008) (Fig. 3). The entire package is summarized in this paper as the Tanne Mélange (Fig. 2). The Tanne Unit is fault-bounded by sole thrusts and tectonically imbricated and overridden, at least in part, by the Südharz-Selke Nappe (Ganssloser 1996) (Fig. 2).

In the Blankenburg Unit, the sedimentary record ranges from the Lochkovian to the Visean (Fig. 3) (Wachendorf et al. 1995; Schwab and Hüneke 2008). The oldest part comprises shale with chert nodules and limestone turbidites of Lochkovian to Lower Emsian age, followed by the Hauptquarzit Formation (quartzite, Upper Emsian) and the Wissenbach shale, which is Upper Emsian to Eifelian in age. A Givetian aged diabase is followed by a package of Frasnian to Famennian shales and cherts, which interfinger with limestone turbidites (‘Flinzkalke’) and deepwater limestones (‘Hercynkalke’) (Fig. 3). The Lower Carboniferous part is composed of radiolarites and cherts (Tournaisian to lowermost Visean) with intercalated diabase, mid-Visean clay shale (Culm-Tonschiefer Formation) and greywacke (Culm-Grauwacke Formation) (Schwab and Ehling 2008b; Schwab and Hüneke 2008). The Blankenburg Unit is bounded by thrust faults towards the NW and became tectonically overridden top-NW by the nappe of the Harzgerode Mélange Complex (Figs 2 & 3). Mélanges of the latter form lithostratigraphic sub-units in the area of the Blankenburg Unit, such as the Hüttenrode Mélange (Zweig et al. 2016) and the Bodetal and Harzrand mélanges (Lutzens and Schwab 1972).

Within the Blankenburg Unit, the Elbingerode Complex forms a tectonic window that allows insight into pre-flysch deposits (Figs 2 & 3). Its section starts with the Wissenbach shale (Goslar Formation, Emsian–Eifelian) followed by the Elbingerode Schalstein Formation, which is composed of diabase, diabase tuffs and exhalative sedimentary iron ore, mostly hematite of the Lahn–Dill type. On top of submarine volcanoes, extensive Givetian to Frasnian aged reef limestones (Elbingerode Riffkalk Formation) occur. The Famennian brachiopod- and cephalopod-bearing limestone interfingers with coloured shales (Elbingerode Buntschiefer Formation) (Fig. 3). In Visean time, the clastic flysch sedimentation (turbiditic greywackes) terminated the Elbingerode reef build-up (Fig. 3) (Fuchs 1987; Friedel and Janssen 1988).

The Sieber Unit comprises, from the base upwards, quartzite (Emsian), shale and diabase (Emsian to Eifelian), chert (Frasnian to Tournaisian), clay schist (Tournaisian), alum schist (Tournaisian) and finally the Culm Grauwacke Formation (greywacke, Visean; Fig. 3) (Wachendorf 1966, 1986; Stoppel and Gundlach 1972; Solanwar 1978; Schwab and Ehling 2008a; Schwab and Hüneke 2008).

The section in the Acker-Bruchberg Unit starts with a Tournaisian alum schist followed by radiolarites and chert (Culm-Kieselschiefer Formation). Up-section, the Acker-Bruchberg Formation (Visean; Jäger and Gurski 2000) follows, which is subdivided into a lower Schiffelborn Member (quartz schist and platy quartzite) and an Upper Kammquarzit Member (thick-bedded high mature quartzite). In the Schiffelborn Member, a diabase is locally intercalated. The highly mature quartzites of the Acker-Bruchberg Formation represent a special facies and development among the Visean siliciclastics, which in all other tectonostratigraphic units of the Harz Mountains are composed of flysch deposits (greywacke, mudstone, olistostromes, shale). The Kammquarzit Member was derived from sands that were formed in shallow-marine environments. Sands were recycled several times and became finally re-deposited into deeper basins as turbidites and other submarine fan deposits (Jäger and Gurski 2000). The thickness of single beds and the degree of maturity increase up-section. The Visean age for the Acker-Bruchberg Formation was derived from conodonts (Büchner and Stoppel 1997) and spores (Jäger 1999). The Early Carboniferous Acker-Bruchberg Formation is part of the Hörre-Gommern Zone, and marks the NW margin of the allochthonous thrust-bounded slices in the Harz Mountains (Engel et al. 1983; Franke 2000).

In the Söse Unit (Fig. 2), Devonian sediments occur in thrust slices as Frasnian clay schist with limestone layers and Famennian red clay schist with carbonate nodules. Tournaisian clay and alum schist are followed by cherts (Culm-Kieselschiefer Formation, Lower Visean) and finally by greywackes of the Visean flysch deposits (Culm-Grauwacke Formation; Fig. 3; Meyer 1965; Wachendorf 1986; Schwab and Ehling 2008b; Schwab and Hüneke 2008). At the northwestern margin of the Söse Unit, the Oberharz Diabase Ridge occurs, which forms a long and narrow NE–SW oriented strip, mainly consisting of mafic volcanic rocks and showing a stratigraphic range from the Eifelian up to the Famennian (Wachendorf 1986; Ganssloser et al. 1995; Fig. 2).

The westernmost tectonostratigraphic segment is the Clausthal Unit (Figs 2 & 3). Its central element is the Oberharz Devon Antiform, which consists of the Kahleberg Group (Emsian, mainly quartzite) and the Oberharz Group (Eifelian to Famennian). The latter starts with the Wissenbach Shale (Upper Emsian to Eifelian). This part of the section also hosts the famous sedimentary-exhalative Rammelsberg ore deposit (mainly Pb, Zn). Shales of the Oberharz Group extend up to the Famennian and interfinger with carbonates and related lithologies of the Romkerhalle Formation. Other elements of the Devonian in the Clausthal Unit are the reef limestones and related sedimentary rocks of the Iberg Reef (Givetian to Frasnian; Franke 1973) (Figs 2 & 3). The Lower Carboniferous of the Clausthal Unit starts with Tournaisian alum schists followed by cherts and radiolarites (Culm-Kieselschiefer Formation, Tournaisian–Lower Visean). A mid-Visean alum schist package forms the transition to the flysch deposits of the Culm-Grauwacke Formation (Upper Visean). The greywackes of the latter form the largest part of the fill of the so-called Culm Basin of the Clausthal Unit (Figs 2 & 3) (Ribbert 1975; Buchholz et al. 2006, 2008).

The Brocken Plutonic Complex was intruded into the Blankenburg Unit, the Sieber Unit, the Acker-Bruchberg Unit and the Söse Unit after their deformation and final emplacement (Figs 2 & 3). The complex consists of several intrusion boosts including a syenogranite in the centre of the pluton (‘Kerngranit’). Its marginal varieties can be porphyric, aplitic, micro-pegmatitic, very fine and very coarse grained. Other granitoids are represented by granite senso stricto and a hornblende–augite bearing granite. At the easternmost margin of the complex, a north–south striking body of diorite and quartz diorite occurs (Schwab 2008). Reported ages of intrusion show a spread from 283 ± 2.1 (Zech et al. 2010) to 293–295 Ma (Baumann et al. 1991).

The Oker Granite is a small pluton that was intruded into deformed Devonian and Lower Carboniferous strata of the Clausthal Unit (Figs 2 & 3). The reported age of intrusion is at about 293–295 Ma (Baumann et al. 1991).

The Ramberg Granite was intruded into the deformed Blankenburg Unit (Figs 2 & 3). It comprises two lithologies, a granite and a syenogranite. An intrusion age of 283 ± 2.8 Ma is published by Zech et al. (2010).

The Harzburg Plutonic Complex, often referred simply as ‘Harzburger Gabbro’, is situated near the city of Bad Harzburg and was intruded after Variscan deformation into the Clausthal Unit and the Söse Unit (Figs 2 & 3). At its eastern margin, it intrudes into the Ecker Gneiss and interfingers with the Brocken Plutonic Complex (Fig. 2). Field relations do not tell exactly which of the plutonic complexes is older or younger than the other. A range for the age of intrusion is 293–295 Ma (Baumann et al. 1991), which is the same as that for the Brocken Plutonic Complex. The Harzburg Plutonic Complex forms a layered intrusion and consists of different mafic to ultramafic rocks. The main lithology is a gabbro–norite with subordinate gabbro, norite, diorite, quartz diorite and orthopyroxenite (Vinx 1982). The chemical composition of the Harzburg Plutonic Complex points to an island arc tholeiite (Sano et al. 2002). The most famous lenses that occur in the complex are those of harzburgite. The location in the valley of the Radautal near Bad Harzburg is the type locality for the peridotitic rock harzburgite, from which it was described and defined by Rosenbusch (1887).

The Ecker Gneiss is a 7 km² wide paragneiss block, which is pinched between the Harzburg and Brocken plutonic complexes (Figs 2 & 3). It is a garnet–biotite bearing, amphibolite-facies paragneiss (Martin-Gombojav 2003). According to SHRIMP U–Pb ages, the siliciclastic protolith must be younger than c. 410 Ma (Geisler et al. 2005). SHRIMP U–Pb inherited detrital zircon ages point to a Baltic provenance (Martin-Gombojav 2003). The Ecker Gneiss is intruded by the Brocken and the Harzburg plutonic complexes.

For U–Pb LA-ICP-MS detrital zircon analysis, we collected 17 samples from sedimentary rocks. In addition, nine samples of sedimentary rocks from other continents and terranes were involved in our detrital zircon study to characterize zircon provenances from areas that may have acted as sources for the sedimentary deposits in the Harz Mountains. A further seven samples of igneous rocks were taken for U–Pb LA-ICP-MS zircon dating. Exact sample locations including GPS coordinates are given in the supplementary material. Locations are indicated on the map (Fig. 2) and in the stratigraphic columns of the tectonostratigraphic units (Fig. 3).

From the Wippra Unit, two samples were collected. The first was taken from a quartzite of the Klippmühle-Quartzite Formation of the Wippra Group (sample 22Hz08), which is the oldest rock package in the Harz Mountains yielding Lower Ordovician acritarchs (Sehnert 1991). The other sample (Hz16) is from a quartzite bed, and was taken from the Pferdeköpfe Formation, which is the lowermost section of the Devonian Piskaborn Group. Sample H20 is a greywacke from the Südharz-Selke Greywacke Member of the Südharz-Selke Nappe. Another two samples (Hz09, 22Hz04) come from the Ecker Gneiss, which is a paragneiss. Sample Hz25 is a quartzite from the Kahleberg Group. From the Visean Kammquarzit Member, two samples of massive quartzite were taken (Hz01, Hz23).

Another group of samples for detrital zircon analysis was taken from tectono-sedimentary mélanges and flysch deposits, which are all Visean to Serpukhovian in age. The first sample is a greywacke (23Hz01) from the Harzgerode Mélange Complex. Another sample from the same unit is a sandstone fragment (Hz26), which was embedded in the mélange. Two more samples were taken from the Stiege Mélange, which underlies the Südharz-Selke Nappe. One sample is a quartzite clast (HzM18K), the other was taken from the greywacke matrix (HzM18M) of the mélange deposit. From the Culm-Grauwacke Formation of the Tanne Unit, we took a greywacke sample (Hz22). Out of the Hüttenrode Mélange (part of the Harzgerode Mélange Complex, which overlies the Blankenburg Unit), one quartzite clast (HzM21K) and two mudstones from the matrix were sampled (HzM19M, HzM21M). Finally, we collected a greywacke sample (Hz24) from the flysch deposit of the Culm-Grauwacke Formation in the Clausthal Unit.

From the Brocken Plutonic Complex, three samples were collected. One syenogranite sample comes from the top of the Brocken Mountain (Hz05). The other one is a syenogranite (HzBg1n) from the area around the village of Torfhaus. A third sample was taken from a diorite (Hz13) that forms the eastern margin of the complex. The outcrop is situated within the Thumkuhlental, 2 km SW of the Lossen Monument. Sample Hz11 is a granite from the Oker Granite. The sampling location is situated near Goslar and Oker. Finally, we took one granite sample (HzRAM1n) from the Ramberg Granite, which was taken east of the Hexentanzplatz area close to the city of Thale.

From the Harzburg Plutonic Complex, two samples were included in this study. At the harzburgite type locality at Kohlebornskehre, in the valley of the Radautal near Bad Harzburg (Rosenbusch 1887), a small lens of plagiogranite (trondhjemite) occurs within the harzburgite. From that lens, sample 22Hz1 was collected. About 20 m south of that locality, a sample of a diorite (Hz08) was taken.

To base tectonic interpretations and models upon the provenance of detrital zircon, it is necessary to know the composition of potential source areas. Thus, we included in our study nine representative samples from such geological units, which were directly or indirectly involved in Cadomian–Avalonian and Variscan orogenic processes. A quartzite sample (MOR A1) was taken from the Taghdout Group, which is Tonian in age and located in the Anti-Atlas of Morocco (West Africa) (Errami et al. 2021). Three further samples are greywackes from the Cadomian basement of the Saxo-Thuringian Zone, which are Upper Ediacaran in age (Katz1, AltTb1, Zir1). Data from these samples have been published by Linnemann et al. (2014). Samples Moelv3b and Moelv5c were taken from the Moelv Tillite (Hedmark Group, Oslo region, South Baltica), which is Cryogenian to Ediacaran in age (Bingen et al. 2005). From the Wartenstein Gneiss, located in the southern Rhenish Massif, sample S-Hun2 was taken. That paragneiss is Upper Ediacaran in age (Dörr and Stein 2019). In addition, we included two samples of Devonian sandstones from both sides of the Mid-German Crystalline Rise. Sample FELD2 was taken from the Taunus Quartzite Formation (Pragian–Lower Emsian) in the Feldberg area of the Rhenish Massif (Rheinisches Schiefergebirge). The other Devonian sample (LA5) was taken from the Hangender Quarzit Member (Gleitsch Formation, Famenne). The sample was obtained from the Schwarzburg Antiform of the Saxo-Thurigian Zone close to Saalfeld.

Heavy mineral separation and isotopic analyses of detrital and magmatic zircon were carried out at the GeoPlasmaLab of the Senckenberg Naturhistorische Sammlungen Dresden. For scanning electron microscopy (SEM), including cathodoluminescence imaging, an EVO 50 instrument (Zeiss) was used. Digital imaging was done using a Keyence Digital Microscope. The analyses of U–Th–Pb isotopes were performed with a RESOlution SE 193 nm excimer laser (Applied Spectra) connected to a single-collector ICP-MS ELEMENT 2XR (Thermo Fisher Scientific). For more details of the analytical procedure, see the Supplementary material. Concordia age plots were generated with Isoplot 4.15 (Ludwig 2008). Kernel density estimation plots were produced using the ‘detzrcr’ package for the statistical program R (version 3.6.1; Andersen et al. 2018).

For provenance studies, a total number of 3012 zircon grains were analysed for U–Th–Pb isotopes and U–Pb age dating. From this analysis, a total of 2391 were concordant in the range of 90 to 110%. Data are presented in Figures 4 to 8. From samples of the igneous rocks, 220 U–Th–Pb analyses were carried out, of which, 170 were concordant (90–110%). Data are shown in Figures 4 to 8.

Data for the establishment of zircon provinces are shown in Figure 4. Such data are necessary to discuss provenances, transport directions and continent interactions for the reconstruction of orogenic processes. Here, data from geological units are presented, which are typical of the main continental blocks, and which are as close as possible to our study area (Fig. 4).

The West African Zircon Province is characterized by a large peak at c. 1950 to 2200 Ma (Fig. 4), which corresponds to the Eburnean Orogeny in the West African Craton. There is another pronounced Paleoproterozoic zircon peak of c. 2200 to 2500 Ma. Archean zircon ages scatter in the range of c. 2500 to 3300 Ma, and were formed during the Liberian (c. 2600–2700 Ma) and Leonian orogenies (c. 3100–3300 Ma) in West Africa (Fig. 4). Very old zircon grains (c. 3400 to 3600 Ma) were derived from the oldest nuclei of the West African Craton (Potrel et al. 2005). There was a long period of tectono-magmatic quiescence in the West African Craton, during which no zircon was formed. This time span ranges from c. 900 to 1600 Ma and is referred to as the West African magmatic gap (Fig. 4; e.g. Linnemann et al. 2014).

The Cadomian Zircon Province is typical for the sedimentary rocks of the Cadomian basement, which was consolidated in latest Ediacaran time (Linnemann et al. 2014). Such rocks occur between Iberia and the Bohemian Massif, forming the core of Cadomia (Fig. 1), a large peri-Gondwanan tectonostratigraphic unit at the northern periphery of West Africa (Linnemann et al. 2018, 2022). That unit probably never left the Gondwanan margin (Linnemann et al. 2004). The Cadomian Zircon Province has a close relation to the West African Zircon Province because portions of West African rocks became recycled and involved in Cadomian magmatic and sedimentary processes (Linnemann et al. 2014). Thus, the Cadomian Zircon Province is characterized by an inherited Eburnean peak and the same pattern of age distribution among the Archean zircons (see Fig. 4). In addition, the Cadomian zircon pattern reflects also the West African magmatic gap (Fig. 4). The main peak is composed of Ediacaran zircon ages ranging from c. 540 to 700 Ma. Such zircons seem to correspond to Cadomian arc magmatism and the igneous activity of Pan-African events during the amalgamation of Gondwana (Fig. 4).

To characterize the zircon populations of Southern Baltica, detrital zircon spectra from the Moelv tillite (Hedmark Group) in southern Sweden, north of Oslo, are presented. The tillite has in common with the Cadomian basement, an Ediacaran age (Bingen et al. 2005). In contrast to West Africa and Cadomia, the South Baltica Zircon Province is characterized by abundant Tonian, Mesoproterozoic and late Paleoproterozoic zircons in the age range of c. 900 to 1800 Ma (Fig. 4). Most zircon ages fall in the Mesoproterozoic (1000–1600 Ma), with distinct peaks at c. 1000–1200 and 1500 Ma (Fig. 4). A few zircon grains are of Lower Ediacaran and Paleoproterozoic age.

Outcropping basement units of East Avalonia are very scarce on the European mainland. One of the few occurrences is represented by the Wartenstein Gneiss Complex at the southern margin of the Rhenish Massif (Dörr and Stein 2019). Analysis of that paragneiss shows a distinct peak of Ediacaran zircon ages in the range of c. 570 to 720 Ma. A few ages scatter in the Mesoproterozoic interval (1000–1600 Ma). This pattern is consistent with data from the Avalonian Lower Paleozoic rocks of the Brabant and Stavelot-Venn massifs (Linnemann et al. 2012; Herbosch et al. 2020), and thus typical of an East Avalonian Zircon Province. Mesoproterozoic ages point to a possible linkage between East Avalonia and Southern Baltica in Neoproterozoic time.

Important sedimentary rocks of the Harz Mountains are Devonian in age. Therefore, typical zircon patterns of Devonian rocks from the Rheno-Hercynian Zone (Rhenish Massif) and from the Saxo-Thuringian Zone are included. Kołtonik et al. (2018) showed that Famennian sandstones from the Rheno-Hercynian Zone yield zircon populations typical for Baltica. In this paper, we add analyses of detrital zircon from the Taunus Quartzite Formation (Pragian–Lower Emsian = Siegenian). The Taunus Quartzite occurs on the southern margin of the Rhenish Massif, and most zircon ages show the pattern of the South Baltic Zircon Province (Fig. 4). An additional major peak of c. 440–420 Ma points to the input of detritus from Silurian arc magmatism. In complete contrast, Devonian siliciclastic sedimentary rocks from the Saxo-Thuringian Zone are characterized by inherited and recycled detrital material from the Cadomian basement only (Fig. 4). This observation represents an important tool to distinguish transport directions and provenances for sedimentary detritus in the Harz Mountains. During Variscan collision and formation of flysch sediments during Visean time, mixing and sediment homogenization of both sources occurred (uppermost histogram shown in Fig. 4).

Described zircon provinces enable to distinguish between nappes and tectonostratigraphic units derived from the Rheno-Hercynian margin from such nappes and units, which have an origin from the Saxo-Thuringian margin. In the Harz Mountains, Rheno-Hercynian nappes are characterized by a majority of the South Baltica Zircon Province, whereas in Saxo-Thuringian nappes, the Cadomian Zircon Province dominates. These Precambrian fingerprints are dominant and visible in any of the zircon populations of each sample investigated. Strong dominance of the South Baltica Zircon Province points to a sedimentary transport from the Rheno-Hercynian shelf into the basin, because the latter was supplied with detritus from South Baltica. Most of the Avalonian basement of the Rheno-Hercynian Zone was buried in Devono-Carboniferous times, and thus probably contributed little detritus to marine shelves. A strong contribution from the Cadomian Zircon Province indicates sediment supply from the Saxo-Thuringian Zone, which was part of Cadomia and situated at the periphery of West Africa (peri-Gondwana) (Linnemann et al. 2004).

Younger parts of the zircon populations of each sample in our study were divided into time frames that are characterized by distinct tectono-magmatic episodes (Figs 5–8). Zircons in the age range of c. 900 to 680 Ma are thought to be related to the dispersal of the supercontinent Rodinia and the pre-Gondwanan re-organization of plates and terranes. Gondwana assembly and related orogenic events such as the Avalonian and Cadomian magmatic arcs and magmatic impulses fall in a time window of c. 540 to 680 Ma. The opening and closure of the Rheic Ocean happened from c. 540 to 420 Ma. Variscan orogenic processes were active in a time frame of c. 420 to 330 Ma. In addition, the diagrams show also the maximum depositional ages (MDA), which deliver important age constraints because the sediment must be younger than its youngest zircon population. For fossil-free deposits or metamorphic rocks, this can be a very useful and powerful tool.

A first group of Devonian sedimentary rocks from the Harz Mountains shows a strong sedimentary supply from the Rheno-Hercynian shelf, which is characterized by the South Baltic Zircon Province. These are the sandstone of the Kahleberg Group (Hz25; MDA 471 ± 7 Ma, Clausthal Unit), the paragneiss samples from the Ecker Gneiss (Hz09, 22Hz04; MDA 410 ± 3 Ma) and the quartzite samples of the Kammquarzit Member (Hz01, MDA 462 ± 7 Ma; Hz23, MDA 428 ± 7 Ma; Acker-Bruchberg Unit) (Fig. 5). All five samples show a strong dominance of zircons reflecting the South Baltica Zircon Province, with portions in the range of 71 to 89% (Fig. 5). These findings confirm previous investigations (e.g. Franke et al. 1978; Mecklenburg et al. 1984; Jäger 1999; Franzke 2001; Huckriede et al. 2004). In addition, the Baltica Zircon Province is reflected by two samples from the Hüttenrode Mélange (Fig. 8). Zircon populations of one matrix sample (Hz21M) and one quartzite fragment (HzM21K) are dominated by Baltica-derived zircon grains. The Hüttenrode Mélange as part of the nappe of the Harzgerode Mélange Complex was thrusted for a distance of at least 10–15 km over the Blankenburg Unit. According to our interpretation, the Hüttenrode Mélange assimilated, during its tectonic transport, sedimentary material from the quartzite of the Emsian Hauptquarzit Formation (Fig. 3) of the underlying Blankenburg Unit. The Hüttenrode Mélange incorporated Baltica-derived zircon by abrading rocks of the underlying rocks of the Blankenburg Unit.

The zircon population patterns from the samples mentioned above qualify the Blankenburg Unit and the Acker-Bruchberg Unit to the group of the Rheno-Hercynian nappes (Fig. 9). The Ecker Gneiss is placed as a 7 km² wide unit between the Brocken Plutonic Complex and the Harzburg Plutonic Complex. Its zircon record relates the paragneiss block to the Rheno-Hercynian Zone. The Oberharz Devon Antiform represents the parautochthonous part of the Rheno-Hercynian Zone (Fig. 9).

Another group of sedimentary rocks is strongly related to a sedimentary supply from the Saxo-Thuringian shelf, because the Cadomian Zircon Province is dominant (Figs 6–8). The sample from the Klippmühle Quartzite Formation of the Wippra Unit (22Hz08; MDA 530 ± 4 Ma; Fig. 6) is the stratigraphically oldest sample in our study, and Lower Ordovician in age. Ordovician ages of metasediments in the Wippra Unit are biostratigraphically documented by Burmann (1973), Ackermann (1985) and Sehnert (1991). The sample 22Hz08 shows 51% of Cadomian detrital zircon and a West African input of 14% (Eburnean and Liberian input). Both together form the Cadomian Zircon Province (65%) and indicate a dominant sedimentary supply from the Saxo-Thuringan Zone. Surprisingly, that sample also has a portion of 14% of the South Baltica Province, which points to a pre-Ordovician interplay between Baltica and Cadomia. Discussing this is beyond the scope of this paper.

The sample from the Pferdeköpfe Formation (Hz16; Piskaborn Group, Wippra Unit; Fig. 6) has a MDA of 393 ± 5 Ma and is thus Eifelian or younger in age, as biostratigraphically documented by Burmann (1973), Ackermann (1985) and Sehnert (1991). The Cadomian Zircon Province dominates because 42% of the zircons are of Cadomian arc material, 7% Eburnean and 1% Archean input. The West African (1000–1600 Ma) gap is clearly visible. Only two zircon grains fall in that interval, which may be recycled from older strata. Pronounced peaks reflect igneous activity in the source area during opening and closure of the Rheic Ocean (28%) and Variscan-related magmatism (16%).

Sample Hz20 was taken from the Südharz-Selke Grauwacke Member (Südharz-Selke Formation, Südharz-Selke Nappe; Fig. 6). The MDA of 370 ± 4 Ma is consistent with the Famennian age suggested by Tschapek (1990). Hz20 shows a very similar zircon age distribution pattern as sample Hz16 from the Wippra Unit. The Cadomian Zircon Province dominates because of 39% of Cadomian arc material, 6% Eburnean and 1% Archean input. The West African (1000–1600 Ma) gap is clearly visible. Only one inherited zircon age falls in that interval. Again, pronounced peaks reflect igneous activity in the source area during opening and closure of the Rheic Ocean (10%) and Variscan-related magmatism (33%). The latter points to an exhumed source area characterized by felsic arc-related rocks. Similarities of both samples allow a correlation between the Piskaborn Group and the Devonian strata incorporated into the Südharz-Selke Nappe, as suggested by Wachendorf (1968), Schwab (1976), Franzke (1969) and Ganssloser (1996). Thus, according to spatial arrangement, stratigraphic ages and similarities in lithology and detrital zircon populations, the Südharz-Selke Nappe seems to be derived from the Piskaborn Group of the Wippra Unit (Fig. 9). The entire steeply dipping Wippra Unit is assumed to represent a marginal part of the Saxo-Thuringian margin, which is tectonically detached from the main part of Saxo-Thuringia by the final emplacement of the Mid-German Crystalline Rise (Fig. 9).

For the Stiege Mélange, which is the basal part the Südharz-Selke Nappe, a MDA of 328.8 ± 2.9 Ma (Serpukhovian) was calculated from the youngest zircon population of the greywacke sample HzM18M (Fig. 6). Therefore, the emplacement of the entire Südharz-Selke Nappe took place after c. 329 Ma in Serpukhovian (Namurian) time or later. The chaotic tectono-sedimentary Stiege Mélange is, as the rest of the Südharz-Selke Nappe, derived from Saxo-Thuringian sources (Fig. 9), as demonstrated by the zircon population pattern of greywacke sample HzM18M and sample HzM18K, which is a fragment of a sandstone incorporated in the mélange. Both samples show a pronounced pattern typical for the Cadomian Zircon Province (Fig. 6). In addition, there is a major part of detrital material derived from Variscan arcs in the samples of the Südharz-Selke Nappe, which ranges from 26 to 64%. As mentioned above for the greywacke of the Südharz-Selke Formation, large parts of the source area are characterized by felsic arc-related rocks. A good candidate for this is the arc-complex of the Mid-German Crystalline Rise, which would have been active from the Famennian onwards.

Sample Hz22 is a greywacke and was collected from the Culm-Grauwacke Formation of the Tanne Unit, which forms a nappe thrusted onto the Blankenburg Unit and is itself tectonically overlain by the nappe of the Harzgerode Mélange Unit (Fig. 9). Zircon populations of sample Hz22 (Fig. 7) show a MDA of 357 ± 6 Ma and are dominated by the pattern of the Cadomian Zircon Province (50%) and 32% of detrital material derived from a Variscan arc source (Fig. 7). Both features point to a source area in the Saxo-Thuringian Zone and the Mid-German Crystalline Rise, and qualify the Tanne Unit as a Saxo-Thuringian nappe (Fig. 9). One greywacke sample from the Harzburg Mélange Complex (23Hz01; Fig. 7) shows strong similarities to sample Hz22 from the Tanne Unit. With a MDA of 354 ± 2 Ma and a dominant zircon population showing the pattern of the Cadomian Zircon Province (52%) and a portion of 34% of zircon grains derived from a Variscan arc source, a classification of the Harzburg Mélange Complex as a Saxo-Thuringian nappe is very likely (Fig. 9). One sample from a fragment of a sandstone clast in the mélange (Hz26; Fig. 7) shows a MDA of 520 ± 8 Ma, a strong portion typical for the Cadomian Zircon population (68%) and no Variscan detrital ages. Surprisingly, there is a portion of 14% of detrital zircon ages typical for Southern Baltica (Fig. 7). Therefore, the sandstone fragment could be derived from Devonian sandstones showing a Baltic provenance, which occur at some localities in the Mid-German Crystalline Rise (Zeh and Gerdes 2010).

The Harzgerode Mélange Complex is thrusted onto the Blankenburg Unit. In that unit, the complex is represented by several mélanges, such as the Hüttenrode Mélange at Königshütte. From the sandstone matrix of that mélange, a mudstone sample (HzM19M) was collected. The MDA is 328.4 ± 2.8 Ma, thus the final emplacement of the Hüttenrode Mélange and the entire Harzburg Mélange Complex took place not earlier than Serpukhovian time. As sample Hz22 (Tanne Unit) and sample 23Hz01 (Harzgerode Mélange Complex), the mudstone sample HzM19M shows a very strong and dominating portion of 72% of zircons from a source characterized by felsic Variscan arc-related rocks (Fig. 8). The potential source area of that zircon population is the Mid-German Crystalline Rise, which was obviously available for erosion in Serpukhovian time. Zircons representative of the Cadomian Zircon Province are visible but show only a portion of 5%. Instead, a portion of 14% of Baltica-derived zircons occurs (Fig. 8). Presence of the latter may be explained by the presence of Baltica-derived Devonian sandstones in the potential source area of the Mid-German Crystalline Rise (Zeh and Gerdes 2010). Another possibility could be the accumulation of Baltica-derived detrital material from the underlying Blankenburg Unit during the tectonic transport of the mélange. A major portion of Variscan arc material and the presence of zircons representing the Cadomian Zircon Province in sample HzM19M relate the Hüttenrode Mélange as part of the Harzgerode Mélange Complex to the group of the Saxo-Thuringian nappes (Fig. 9).

Greywacke sample Hz24 (Fig. 8) was collected from the flysch deposits of the Culm Basin in the Clausthal Unit (Figs 2 & 3). A MDA of 339 ± 6 Ma points to an onset of flysch deposition during the Visean or younger. Available palaeontological data constrain a Serpukhovian (Namurian) age of sedimentation (Wachendorf 1986). A dominant portion of 64% of zircons derived from a Variscan arc source and a portion of 27% of zircons representative of the Cadomian Zircon Province show a clear provenance from the Mid-German Crystalline Rise and rocks of the Saxo-Thuringian Zone for the flysch deposits in the Culm Basin of the Clausthal Unit (Fig. 9). Remarkably, there is no visible input from the Rheno-Hercynian margin. The present dataset demonstrates that all flysch deposits and related tectono-sedimentary mélanges derived from the Mid-German Crystalline Rise and the Saxo-Thuringian Zone (Fig. 9). The arc magmatism in the Mid-German Crystalline Rise started in Famennian time (see detrital input of 33% in the Südharz-Selke Grauwacke Formation; Fig. 6). Detrital input from the Mid-German Crystalline Rise is permanently high and increased from 32% in the Tanne Unit and from 33% in the Harzgerode Mélange Complex up to 72% in the Hüttenrode Mélange. In the flysch deposits of the Clausthal Unit, the portion of the detrital contribution from the Mid-German Crystalline Rise is likewise very high (64%). All in all, a stepwise exhumation of the plutonic arc rocks of the Mid-German Crystalline Rise during Visean and Serpukhovian times is detectable from the detrital zircon record (Fig. 9). According to Walliser and Alberti (1983), Rheno-Hercynian and Saxo-Thuringian nappes may summarize as a Harz Nappe System (Fig. 9).

The three major granitoid intrusions of the Brocken Plutonic Complex, the Oker Granite and the Ramberg Granite are part of our geochronological study. We also measured two samples from the mafic to ultramafic Harzburg Plutonic Complex (Fig. 9).

Three samples were dated from the Brocken Plutonic Complex. The first was a syenogranite (Hz5) taken from the top of the Brocken Mountain (‘Kerngranit’). Of 118 measured zircon grains, 38 analyses were concordant. The most concordant grains of the youngest zircon population (n = 8, 95–101% conc.) were used to calculate a concordia age at 295 ± 3 Ma (Fig. 10), which is interpreted to be the age of intrusion. The sample yielded two concordant grains from an older basement, with ages at 1376 ± 36 and 2346 ± 42 Ma. A zircon grain from sample Hz5 is pictured in Figure 10. It shows the typical shape of zircons grouped in the P4 field of the zircon crystal typological classification diagram of Pupin (1980), which allows an estimate for the magma temperature of about 800°C. Another sample from the Brocken Plutonic Complex is a syenogranite (HzBg1n) that was taken east of the village of Torfhaus. Of 30 analysed grains, 15 were concordant in the range of 90–100%. From such zircons, the most concordant ones (97–100%) were selected (n = 6) to calculate a concordia age at 295.7 ± 3.8 Ma (Fig. 10). This result is interpreted to reflect the age of intrusion, which is, within error, in line with the age of Hz5. No older basement zircons were detected in sample HzBg1n. The next sample (Hz13) comes from a diorite, which forms a narrow strip at the eastern margin of the Brocken Plutonic Complex (Fig. 2). Of 47 analyses, 26 were concordant in the range of 90 to 110%. From the most concordant analyses (n = 5, 95–108% conc.), the calculated concordia age is 300.4 ± 3.2 Ma (Fig. 10). We interpret this as the age of intrusion for the diorite, which is about 5 Ma older than the other granitoid samples of the Brocken Plutonic Complex. No older zircon ages of an underlying basement were detected in sample Hz13.

The sample from the Oker Granite (Hz11) was collected in the Okertal near Bad Harzburg. Of 30 analysed grains, 28 zircons were concordant. From the most concordant zircons (n = 5, 99–100% conc.), a calculated concordia age gave 303.3 ± 3.6 Ma, which is assumed to be the age of intrusion (Fig. 10). A few older zircons could be identified, which are aged at 611 ± 13, 1773 ± 34 and 1862 ± 21 Ma.

The sample from the Ramberg Granite (HzRAM1n) was taken east of the Hexentanzplatz area, near Thale. From 28 measured grains, 16 analyses were concordant (92–101%), and the three youngest zircon ages were used for a concordia age calculation resulting in an age of intrusion at 295.0 ± 4.5 Ma. No inheritance from an older basement was found.

In summary, our new data show that granitoid intrusions in the Harz Mountains were intruded in a time window of c. 5 Myr (c. 295 to 300 Ma) during uppermost Carboniferous to lowermost Permian times (Gzehlian to Asselian). A plot of the five inherited zircon grains is shown in Figure 12a. These new ages contrast with the published ages of Zech et al. (2010), which are younger. This is due to the use of other standards and the improvement of technology and methods.

In this study, two samples from the Harzburg Plutonic Complex were included. The first (22Hz01) comes from a plagiogranite (trondhjemite) sampled from a small lens discovered within the harzburgite at its type locality in the Kolebornskehre of the Radautal, close to Bad Harzburg (Fig. 10d). The sample yielded a few quite large (c. 300–400 µm) zircons (Fig. 11). In this case, zircons were not embedded in resin and polished. Instead, such zircons were analysed by LA-ICP-MS from the tape on which the grains were mounted, in order to get the youngest age possible from the very marginal zones of the few crystals (Fig. 11b). A concordia age at 329 ± 2 Ma was calculated from seven spot analyses (Fig. 11a). According to this age, the harzburgite from its type locality is about 329 ± 2 Ma old or even older, because the plagiogranite was originated from partial melting of the harzburgite. From the plagiogranite, a few c. 150 µm large crystals of bytownite (Ca)[(Si,Al)₄O₈] were discovered (Fig. 11c). An energy-dispersive X-ray (EDX) analysis showed that no Na occurs in the crystals (Fig. 11c), which underlines the ultramafic nature of the rock from which the plagiogranitic melt was derived.

Vinx (1982) interpreted the harzburgite as a cumulate within the Harzburg Plutonic Complex. However, this is now in conflict with the available ages. Baumann et al. (1991) published intrusion ages for the Harzburg Plutonic Complex in a range of 293–295 Ma. If the harzburgite would have formed cumulate lenses in the plutonic complex, it should have the same age. Instead, the new age for the plagiogranite at 329 ± 2 Ma questions this interpretation. According to this age, the harzburgite should be older than 329 ± 2 Ma. It is more likely that it forms a xenolith and a remnant of a formerly existing larger body of peridotite. The harzburgite may represent a piece of older pre-329 Ma oceanic crust. This would be more in line with Sano et al. (2002), who outlined that the chemical composition of the Harzburg Plutonic Complex points to an island arc tholeiite. The entire plutonic complex may have resulted from the recycling and re-melting of a larger piece of oceanic crust, and the lenses of the harzburgite form small remnants of peridotite, which was formerly part of an area of oceanic crust.

Another sample of a diorite (Hz08) was taken from the Harzburg Plutonic Complex to verify the age determination of Baumann et al. (1991). No zircon younger than c. 374 Ma (Upper Devonian) could be found in the sample. All obtained zircon ages point to a strong portion of old sedimentary material, which contaminated the diorite. Data on the inherited ages are shown in a plot in Figure 12b. The diorite assimilated a sedimentary rock, in which the zircon grains were typical of the South Baltica Zircon Province (Fig. 12b).

There is broad agreement that East Avalonia collided with southern Baltica at a time frame of c. 450 Ma (e.g. Franke et al. 2017; Torsvik and Cocks 2017). As a consequence, the Rheic Ocean started to subduct top-NW underneath the southeastern margin of East Avalonia (Fig. 13). A magmatic arc was formed at c. 435–420 Ma (Fig. 13), which produced large plutons and volcanics (e.g. Sommermann et al. 1992, 1994; Sommermann 1993; Brätz 2000; Linnemann et al. 2012; Herbosch et al. 2020). Contemporaneously, large volumes of sediment derived from South Baltica started to overstep and cover East Avalonia from earliest Devonian time onwards (Fig. 13; see also sample FELD2 in Fig. 4). Furthermore, the entire Rheno-Hercynian shelf, which covered East Avalonia in the Rhenish Massif, experienced massive sedimentary siliciclastic supply from South Baltica during Devonian time (Kołtonik et al. 2018) (Fig. 13). The Rheic Ocean was closed by c. 420 Ma, and Cadomia collided with the southeastern margin of East Avalonia (Fig. 13) (Kroner et al. 2007). Sands derived from Baltica started to overstep the suture of the Rheic Ocean and were deposited on the northeastern margin of Cadomia (Fig. 13), which is represented in Central Europe by the Saxo-Thuringian Zone (Linnemann et al. 2004). In East Avalonia, Silurian arc magmatism became extinct after ridge collision and slab break-off (Fig. 13). Its arc rocks were covered by late Silurian and Devonian siliciclastics (Struve 1973, 1982). During the beginning of the Middle Devonian, Cadomia rifted away from East Avalonia and a new narrow oceanic basin opened, which is called the Rhenish Ocean (e.g. Franke et al. 2017). Here, the term Rhenish Seaway is preferred, because the narrow oceanic basin existed only for about 30 Myr, which is approximately the lifetime of the recent Red Sea (Stern and Johnson 2019). During break-up from East Avalonia, Cadomia (Saxo-Thuringian Zone) carried sands that were originally derived from South Baltica at its northwestern margin (Fig. 13) (Zeh and Gerdes 2010). The onset of top-SE directed closure of the Rhenish Seaway started at c. 360 Ma, as indicated by the intrusion of the Frankenstein Gabbro at 362 ± 2 Ma (Kirsch et al. 1988), which may reflect ridge subduction (Fig. 12). Intense arc magmatism in a time span of c. 360–330 Myr formed the Mid-German Crystalline Rise on the northwestern rim of Cadomia (Fig. 13) (e.g. Zeh and Gerdes 2010).

The closure of the Rhenish Seaway caused collision tectonics, which led to the origin of complicated nappe structures (Gießen-Harz-Werra-Gommern Nappes; Fig. 13) at the southern margin of East Avalonia (e.g. Eckelmann et al. 2014; Mende et al. 2019; Nesbor 2021). The most important part of that nappe system is represented by the Harz Mountains, showing a tectonic style of thin- to thick-skinned tectonics (Franzke 2001). Sedimentary deposits of the Harz Mountains represent shelf sediments from both the northwestern (Rheno-Hercynian) and the southeastern (Saxo-Thuringian) margins of the Rhenish Seaway. In addition, widespread distributed radiolarites in mid-Devonian to Lower Carboniferous time document true deep-sea environments far below the carbonate compensation depth. The harzburgite in the Harzburg Plutonic Complex is older than 329 ± 2 Ma (this study) and could act as another hint of pre-existing oceanic crust and true oceanic settings. The magmatic arc that developed during the closure of the Rhenish Seaway is represented by the Mid-German Crystalline Rise (Zeh and Gerdes 2010).

Collision tectonics and obduction of the orogenic system onto the Rheno-Hercynian (East Avalonian) margin started in the Harz Mountains by the onset of the deposition of flysch and tectono-sedimentary mélanges in latest Visean to Serpukhovian times (Fig. 14). Marginal shelf deposits on both sides of the ocean and the oceanic sediments were thrusted slice by slice towards the NW to form nappes and the tectonostratigraphic zones of the Harz Mountains (Figs 9, 14 & 15). Sedimentary deposits originated at the southeastern margin of the Rhenish Seaway were deposited at the margin of Saxo-Thuringia (Cadomia) and show a distinct relation to the Cadomian Zircon Province. Such sedimentary sequences occur in the Wippra Unit (Wippra and Piskaborn groups, samples 22Hz08 and Hz16, respectively; Fig. 6) and in the Südharz-Selke Nappe (sample Hz20). The concept of thin- to thick-skinned tectonics allows to easily establish a connection between the steeply dipping Wippra Unit and the Südharz-Selke Nappe (Fig. 9). Sedimentary deposits from the northwestern shelf of the Rhenish Seaway were formed on the southern margin of the Rheno-Hercynian Zone (East Avalonia). They are dominated by detrital input from South Baltica, and thus, the zircon populations are typical of the South Baltica Zircon Province. Such samples are the ones from the Kahleberg Group (Hz25), the Ecker Gneiss (Hz09, 22Hz04) and the Kammquarzit Member (Hz01, Hz23) (Figs 6 & 9). The Blankenburg Unit and the Acker-Bruchberg Unit form Rheno-Hercynian nappes and thrust sheets (Fig. 9). The zircon record of the Ecker Gneiss points to a derivation from Devonian sedimentary rocks of the Rheno-Hercynian Zone.

Sedimentary transport during flysch sedimentation was generally top-NW oriented (Plessmann and Wunderlich 1961; Ribbert 1975; Ganssloser 1996; Franzke 2001). First tectono-sedimentary mélanges and flysch sediments are represented by the Tanne Unit and the Harzgerode Mélange Complex (Hz26, 23Hz01) (Figs 9 & 14). Input from former shelf sediments from the southeastern (Saxo-Thuringian) margin dominates because of the fingerprint of the zircons typical of the Cadomian Zircon Province. Some input from South Baltica is explainable by the erosion of its sands, which were stored in the Mid-German Crystalline Rise (Zeh and Gerdes 2010). According to our data, the nappe of the Tanne Unit is thrusted onto the Blankenburg Unit (Figs 9 & 14). According to the MDA of 328.4 ± 2.8 Ma (sample HzM19M), the Harzgerode Mélange Complex (including the Hüttenrode Mélange) was thrusted onto the Tanne and Blankenburg units in Serpukhovian time (Fig. 9). The Südharz-Selke Nappe tectonically overstepped the Harzgerode Mélange Complex in Serpukhovian (Namurian) time (see also Lutzens and Schwab 1972; Schwab 1977; Ganssloser 1996). Emplacement of the Südharz-Selke Nappe took place during Serpukhovian time because of the MDA of 328.8 ± 2.9 Ma of the Stiege Mélange Formation, which forms the basal part of the Südharz-Selke Nappe (sample HzM18M; Fig. 6). The Tanne Unit, the Harzgerode Mélange Complex (including the Hüttenrode Mélange) and the Südharz-Selke Nappe (including the Stiege Mélange) are related to Saxo-Thuringian nappes (Fig. 9).

The dataset shows that all flysch deposits and related tectono-sedimentary mélanges became derived from the Mid-German Crystalline Rise and the Saxo-Thuringian Zone (Fig. 9). The arc magmatism in the Mid-German Crystalline Rise started in Famennian time. Detrital input from the Mid-German Crystalline Rise is high in Lower Carboniferous strata and increases from 32% in the Culm-Grauwacke Formation of the Tanne Unit up to 72% in the Hüttenrode Mélange. In the flysch deposits of the Culm Basin in the Clausthal Unit, the portion of the detrital contribution from the Mid-German Crystalline Rise is still 64%. A stepwise exhumation of the plutonic arc rocks of the Mid-German Crystalline Rise during Visean–Serpukhovian times is detectable (Figs 7–9). The final emplacement of all nappes and flysch deposition in the Harz Mountains took place in Serpukhovian time or later.

Emplacement of the large plutons in the Harz Mountains occurred very much later than the final emplacement of the tectonostratigraphic units of the Harz Nappe System. The latter were finally emplaced in Serpukhovian time or later (pre-300 Ma). Intrusion of the Brocken Plutonic Complex, the Oker Granite and the Ramberg Granite occurred post-deformative in a time window of c. 300 to 295 Ma and forms the upper limit for the emplacement of the tectonostratigraphic units in the Harz Mountains. The time frame 300 to 295 Ma is characterized by a completely different tectonic regime dominated by extension of crust in Central Europe, connected with the opening of the major sedimentary basins such as the Saale Basin, which accumulated large volumes of the Variscan molasse (Upper Carboniferous to Rotliegend, Lower Permian) (Fig. 15). Plate tectonic re-organization led to a change from a compressive to an extensional regime. The latter enabled the emplacement of the major granitoid plutons of the Harz Mountains (Fig. 15). The same regime is also responsible for the intrusion of the gabbroic–noritic Harzburg Plutonic Complex. Inherited zircon from both the granitoids and the Harzburg Plutonic Complex yield remarkable volumes of East Avalonian and South Baltica-derived zircon grains (Fig. 12a, b). Such data witness an underlying East Avalonian (Laurussian) basement underneath the Harz Mountains (Figs 9, 14 & 15). Our data show that a two-plate model of closure of the Rheic Ocean at c. 420 Ma (Kroner et al. 2007; Kroner and Romer 2013) could work, but the re-opening of a narrow ocean in Pangaea's interior in mid-Devonian time, such as the Rhenish Seaway, is required (Figs 13–15). Terrane transfer between the two plates as suggested by Franke et al. (2017) and Zeh and Gerdes (2010) is not confirmed by our dataset.

The authors thank the team of the ‘UNESCO-Global Geopark Harz · Braunschweiger Land · Ostfalen’ for their support during sample collection. We also thank Stephan Schnapperelle (Halle), Wolfgang Franke (Gießen) and an unknown reviewer for their fruitful reviews. Many thanks to Rob Strachan (Portsmouth) for improving the English language of the manuscript.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

UL: conceptualization (lead), data curation (lead), investigation (lead), methodology (lead), supervision (lead), visualization (lead), writing – original draft (lead), writing – review & editing (lead); MZ: data curation (equal), investigation (equal), methodology (equal), writing – original draft (equal); writing – review & editing (equal); MZ-H: data curation (equal), investigation (equal), methodology (equal), writing – original draft (equal); TV: data curation (equal), investigation (equal), methodology (equal), writing – original draft (equal); JZ: data curation (equal), investigation (equal), methodology (equal), writing – original draft (equal); JH: data curation (equal), investigation (equal), methodology (equal), writing – original draft (equal); AG: data curation (equal), formal analysis (equal), investigation (equal), methodology (equal), writing – original draft (equal); KM: data curation (equal), formal analysis (equal), investigation (equal), methodology (equal), writing – original draft (equal); RK: investigation (equal), methodology (equal), resources (equal); FK: data curation (equal), investigation (equal), writing – original draft (equal).

Our research was funded by the Senckenberg Collections of Natural History Dresden, the Senckenberg Society for Natural Research and the Leibniz Association (Germany).

All data generated or analysed during this study are included in this published article (and its supplementary information files).