The challenge of relating the Kasimovian to west European chronostratigraphy: a critical review of the Cantabrian and Barruelian substages of the Stephanian Stage Open Access

The present work addresses analysis of the regional chronostratigraphic framework that has been developed, and is still the fundamental reference, for interpretation of Pennsylvanian (late Carboniferous) successions in western Europe. Integral to this chronostratigraphic construct is the relationship to biostratigraphy, and primarily to that of macrofloras, upon which the conceptual division of west European series and stages (later respectively re-assigned to stages and substages) was established. This work also presents radiometric dating results of pyroclastic ashfall tonsteins from the Cantabrian Mountains (NW Spain) that support a preliminary absolute time framework for current west European chronostratigraphy.

The west European area is a central element for the interpretation of the wider palaeogeographical and climatic history of the Pennsylvanian Period across Euramerica, at a time when the earliest extensive terrestrial wetlands, the coal swamp biome, developed. This period experienced significant global climate change (the Late Paleozoic Ice Age) and widespread orogenies resulting from the collision of the Gondwana and Laurussia plates. Interest has focused upon how this tropical biome interacted with the environmental factors, with potential relevance to modelling environmental changes taking place today (e.g. Gastaldo et al. 1996; Cleal and Thomas 1999, 2005). Understanding these global changes requires robust regional chronostratigraphic frameworks to provide for consistent global correlations for the Pennsylvanian.

The earliest collaborative attempts to develop a regionally applicable framework occurred during international conferences on Carboniferous stratigraphy, held in 1927 and 1935 at Heerlen in the Netherlands (Jongmans and Pruvost 1950; Wagner 1974, 1989; Wagner and Winkler Prins 2016). For convenience the resulting chronostratigraphic framework (Fig. 1) is often informally referred to as the Heerlen Scheme (Gothan 1952). After 1952, formal refinement of the chronostratigraphic framework became the responsibility of the IUGS Subcommission on Carboniferous Stratigraphy (SCCS), whose work has been discussed extensively by Wagner (1974, 1989), Wagner and Winkler Prins (2016) and Lucas et al. (2022a, b).

The Heerlen Scheme ultimately proved unusable at a global scale. Global stages are preferentially defined in open-marine successions (Remane et al. 1996), whereas the Heerlen Scheme had been mainly developed around the west European coal-bearing sequences where correlation depended heavily on the biostratigraphy of the floras, which show significant global provincialism (e.g. Vakhrameev et al. 1978; Meyen 1987; Cleal 1991; Opluštil et al. 2021a). Nevertheless, because of the economic importance of these coal-bearing deposits in widespread terrigenous successions across western Europe (Fig. 2), the stratigraphic divisions established within the Heerlen Scheme have continued to be used as a regional chronostratigraphy, recognized as the west European chronostratigraphic framework by the SCCS (Heckel 2004).

To make the west European regional framework compatible with modern stratigraphic practice (e.g. Hedberg 1965, 1968, 1976; Salvador 1994), boundary stratotypes for the then recognized stages were sought (van Leckwijck 1964). Initially this proved problematic for the upper Westphalian and lower Stephanian stages (now substages following Heckel 2004), as the classic areas for these intervals were in the intramontane basins of France and Germany, where the sequences tend to be incomplete and/or poorly exposed. Detailed investigation from the 1960s onwards in the Cantabrian Mountains of northern Spain, demonstrated a record of more or less continuous sedimentation through this interval with well-exposed successions of these strata. Boundary stratotypes were subsequently proposed, and some later ratified by the SCCS, in the Cantabrian Mountains for four substages spanning the upper Westphalian and lower Stephanian: Asturian (proposed; formerly Westphalian D), Cantabrian, Barruelian (formerly Stephanian A) and Saberian (proposed; formerly lower Stephanian B). The resulting chronostratigraphy (Fig. 3) has been widely accepted by investigators working in Europe (e.g. Waters in Waters et al. 2011; Opluštil et al. 2016a, 2017), parts of north Africa (Tawadros 2011) and Asia Minor (Opluštil et al. 2018), and in easternmost North America (Oleksyshyn 1976; Zodrow et al. 2016) and has proved to be a robust scheme for classifying and documenting strata of this age (e.g. Coquel and Rodríguez 1995; Schneider and Werneburg 2006; Cleal et al. 2009; Wagner and Álvarez-Vázquez 2010a; Opluštil et al. 2016a, 2017, 2021a; Besly and Cleal 2021).

Recently, some workers (e.g. Bashforth et al. 2021; Nelson and Lucas 2021) have argued that this chronostratigraphic framework is invalid, unnecessary or at least open to question, particularly for most of North America. To respond to these claims, the present report will review: (1) the historical and legal foundations of the two ratified substages and the proposed substages that embrace them; (2) the chronostratigraphic issues that led to their establishment with supporting stratotypes; (3) the biostratigraphic criteria used to correlate them; and (4) their distribution in the context of the palaeogeography and depositional environments of the time.

To provide a full contextual understanding of these issues has required an outline of the palaeogeography and tectonic development of the relevant European basins, and in particular the extensive foreland that developed south of the evolving Variscides (Fig. 4). This southern Variscan foreland, which hosted the reference successions in the Cantabrian Mountains, was separated from the northern foreland of western Euramerica by major orographic features. This resulted in the development of a quite distinct but widespread palaeogeographical realm in the southern foreland, with patterns of palaeofloristic development different from those seen in the northern foreland. It is also important to recognize the complex crustal shortening, buckling and rotation that affected the Iberian Peninsula in the latest Carboniferous–early Permian; any palaeogeographical reconstruction for the Late Pennsylvanian, which represents the Iberian Peninsula in its present-day area and geometry is grossly misleading as to the wide extent of the southern foreland area.

New high precision U–Pb radiometric dates (CA-ID-TIMS) are reported for four tonstein horizons that bracket the designated substage boundary stratotypes in the Asturian to Saberian succession. These include tonsteins close to the top boundaries of the Cantabrian and Barruelian substages, providing well-constrained ages for these boundaries. For the first time, it has been possible to provide a preliminary relative chronological framework for the upper Westphalian–mid-Stephanian interval, demonstrating the relative duration of the substages and allowing their correlation with other successions independent of biostratigraphy.

We conclude that these substages form part of a legitimate and scientifically effective regional chronostratigraphic scheme that continues to play a vital role in the geological study of the important Carboniferous coal-bearing deposits of central and eastern Euramerica.

The Heerlen Scheme originally recognized three stages in the Pennsylvanian Subsystem (Jongmans 1928; Fig. 1): the Namurian, Westphalian and Stephanian. The international decision on the mid-Carboniferous boundary (Engel 1989) later formalized that the base of the Namurian falls substantially below the base of the Pennsylvanian. From 1958, these stages were upgraded by the SCCS to series but were subsequently re-designated by the SCCS (Heckel 2004) as stages, and the component sub-units are now substages. The stratigraphy of the Namurian and lower Westphalian has been relatively uncontroversial, with boundary stratotypes now designated at eustatically controlled marine bands exposed in conserved sites in the British part of the northern Variscan foreland (Cleal and Thomas 1996).

Towards the top of the Westphalian, however, a combination of climate change linked with the onset of a major interglacial, and disruptive changes to basin structures due to the Variscan Orogeny, has made correlations more difficult, and this has hindered the development of a consistent regional chronostratigraphic scale. In the northern Variscan foreland area, above the Aegiranum Marine Band (marking the base of Westphalian C, now the Bolsovian), marine bands become sporadic and eventually disappear in these paralic basins, and the sequences become increasingly dominated by sparsely fossiliferous red-bed facies (e.g. Besly and Cleal 2021). As a result, a macrofloral biostratigraphy for the upper Westphalian became difficult to establish for the northern Variscan foreland except in a few parts of Britain and the Canadian Maritimes (e.g. Cleal 1978, 1997, 2007; Zodrow and Cleal 1985).

The stratigraphic units for the topmost Carboniferous were defined, only with very broad reference to certain macrofloral species of assumed biostratigraphic significance, in the various coalfields of the Massif Central of France (Munier Chalmas and de Lapparent 1893; Jongmans and Pruvost 1950), and in particular in the St Étienne Coalfield, which gave its name to the eponymous Stephanian Series (now stage). Nevertheless, this model was somewhat complicated at the first Heerlen Congress (1927), where it was asserted that the only area in western Europe where Stephanian strata overlie Westphalian strata is in the Lorraine–Saar–Nahe Basin. This resulted in the identification of the base of the Stephanian with the level of the Holz Conglomerate (Jongmans 1928), an interval that is widespread across the Lorraine–Saar–Nahe Basin at the base of the Stephanian Ottweiler Group (see Fig. 3). Further investigation of the succession underlying the Stephanian in this coalfield also identified that a significant succession of strata, the Assise de la Houve (see Fig. 3), occurred that was younger than the Westphalian C (now Bolsovian) as had been recognized in the paralic coalfields; this younger succession was designated Westphalian D (Darrah and Bertrand 1933; Bertrand 1937).

The designation of the base of the Holz Conglomerate as a reference for the base of the Stephanian marked a departure from the concept of defining the Stephanian in the French Massif Central, as developed by Munier Chalmas and de Lapparent (1893). However, there were no steps for identification of chronostratigraphic units for the Stephanian until a composite model for the Heerlen Scheme (Fig. 1) was developed by Jongmans and Pruvost (1950), which designated the Stephanian A, B and C as substages based on the succession in the St Étienne Coalfield.

The Jongmans and Pruvost model (1950) incorporated a number of inconsistencies that have later given rise to controversy. The conceptual stratotypes were designated in different palaeogeographical contexts: Westphalian A–C in the paralic coalfields of the northern Variscan foreland, Westphalian D in the intramontane Lorraine–Saar–Nahe Basin and Stephanian A–C in the intramontane coal basins of the French Massif Central. With the exception of the marine bands that defined the stages (now substages) of the lower Westphalian, the boundaries between successive stages remained undefined. The base of Westphalian D as a chronostratigraphic unit has remained undefined to the present; in its conceptual type area there is no longer access to the underground mine successions on which it was conceived. Unconformable contacts marked the base of the Stephanian in the Lorraine–Saar–Nahe Basin and the base of the stages Stephanian A, B and C. It remained an open question as to whether the base of the Stephanian should be defined at the base of the Holz Conglomerate or at the base of the Stephanian A stratotype in the St Étienne Coalfield. Also importantly, the unconformable contacts at the base of the Stephanian stage boundaries, particularly at the base of the Holz Conglomerate, could be seen to represent considerable periods of time that were not represented in the rock record (Bouroz 1968; Germer et al. 1968; Bouroz et al. 1972; Burger et al. 1997; Cleal 2008a). The relative scale of the various discontinuities in the Heerlen Scheme are illustrated in Figure 3 with reference to the preliminary absolute time framework, developed below in the present work. This illustrates those successions in western Europe that have played a part in the discussion on late Westphalian and Stephanian chronostratigraphic units.

It is apparent that in the northern part of western Europe there is no demonstrable continuity of sedimentation and corresponding rock succession across the interval represented in the Heerlen Scheme as Westphalian D through to Stephanian A. However, on the basis of floral biostratigraphy, a substantial thickness of strata of some thousands of metres was reported in the Cantabrian Mountains of northern Spain, which embraced the interval from the upper part of Westphalian D through to Stephanian A (Wagner 1966a, b). This formed the basis of the later detailed discussions within the SCCS leading to the recognition of the Cantabrian as a formal chronostratigraphic unit in the west European regional framework, as discussed below.

At the core of discussions around the Heerlen Scheme was whether this was truly a chronostratigraphic scheme, in the sense of Hedberg (1965, 1968, 1976), or was basically a biostratigraphic model. The Heerlen Scheme was initially founded on biostratigraphic correlations, but, during the second half of the twentieth century, the SCCS followed a programme, in line with IUGS guidelines, to establish a chronostratigraphic framework with defined stratotype sections (van Leckwijck 1964). While biostratigraphic criteria have continued to be used to aid correlation, successive meetings of the SCCS have explicitly recognized the Heerlen Scheme as a chronostratigrahic framework (George and Wagner 1972; Bouroz et al. 1978; Wagner et al. 1985) with a focus to identify a stratotype section to define formally each stage and substage boundary. Boundaries were now defined at a specific moment in time represented by a level in a rock sequence, and not by perceived changes in biotas and therefore biostratigraphy.

In this context, the Cantabrian Stage was introduced to cover the perceived time interval between the conceptual sequences for the upper Westphalian D (Faisceau de Steinbesch at the top of Assise de la Houve in Lorraine – Fig. 3) and the Stephanian A (Assise de Rive-de-Gier in the Saint-Étienne Coalfield – Fig. 3) (Wagner 1966b). This was precisely the same logic that had been applied to establishing the Westphalian D (Darrah and Bertrand 1933; Bertrand 1937) to represent a younger stratigraphic unit above the Westphalian C reference successions in the Nord-Pas-de-Calais and Ruhr coalfields.

Specific issues related to the substages of the current west European chronostratigraphic framework are discussed below; related megafloral biostratigraphy and summary description of the proposed and designated stratotypes are discussed subsequently.

The Westphalian D was originally conceived by Darrah and Bertrand (1933) and Bertrand (1937) as the uppermost part of the Westphalian Stage. It was identified by the presence of the medullosalean foliage Neuropteris ovata Hoffmann in the Assise de la Houve of the Lorraine–Saar–Nahe Basin, which indicated it was younger than most of the Westphalian successions recognized in the northern Variscan foreland such as in the Nord Coalfield of France and the Ruhr Coalfield of Germany.

The Westphalian D Working Group of the SCCS (Laveine 1977) analysed the biostratigraphic criteria used to identify the base of this substage (see also Cleal 1984a), but it was evident that the lack of suitable natural exposures in the Lorraine–Saar–Nahe area would render the area unsuitable for providing a boundary stratotype. Laveine (1977) suggested that the Central Asturian Coalfield (Fig. 5) of northern Spain was the most favourable area to provide an alternative standard section, and this was the basis of a preliminary proposal submitted by Wagner et al. (2002), including that the substage should be renamed Asturian. Although not formally ratified as of this report (2022), the Asturian is identified in the list of regional substages for Western Europe published by the SCCS (Heckel and Clayton 2006) and has gained wide acceptance amongst workers on European successions (e.g. Cleal et al. 2009; Waters et al. 2011; Opluštil et al. 2016a; Rodríguez et al. 2022).

Identification of the top of this substage proved more challenging, particularly as this would, by definition, constitute the base of the Stephanian. There are few places in Europe where the Westphalian–Stephanian boundary can be clearly demonstrated; usually, the grey Westphalian coal-bearing sequences are either truncated by a major unconformity or they transition into younger, poorly fossiliferous red-beds of uncertain age. The most notable exception is in the Lorraine–Saar–Nahe Basin where the upper Westphalian Assise de la Houve (or the Heiligenwalder Schichten in the German part of the basin) is overlain by the Stephanian Ottweiler Group, with the contact marked by the Holz Conglomerate. Consequently, early discussions on the definition of the Westphalian–Stephanian boundary focused on the Lorraine–Saar–Nahe Basin (e.g. Jongmans and Gothan 1937) until it became evident that the Holz Conglomerate represented a major non-sequence (e.g. Wagner 1964a; Germer et al. 1968; Kneuper 1971).

A proposed solution was to re-define the base of the Stephanian at a lower stratigraphical level in the Lorraine–Saar–Nahe Basin, at Tonstein 60 (e.g. Alpern and Liabeuf 1969; Fig. 3), but this was rejected because it would fundamentally alter the original concept of both the Westphalian and Stephanian (Bouroz et al. 1970). Although Tonstein 60 marks a significant change in the macrofloras corresponding to what is now known as the base of the Crenulopteris acadica Zone (discussed below), the upper Assise de la Houve macrofloras are still essentially Westphalian in character (Laveine 1977; Cleal 1984b).

Coincident with the debate on the Westphalian–Stephanian transition as seen in the Lorraine–Saar–Nahe Basin, research on the Pennsylvanian successions in the Cantabrian Mountains of northern Spain (Wagner 1965, 1966a, b) demonstrated that there is a thick sedimentary succession here equivalent to the interval missing from below the Holz Conglomerate. This has proved to be the best exposed such succession anywhere in Europe (Wagner 1965, 1966a, b; Bouroz et al. 1970; Boersma 1977; Laveine 1977) and the considerable thickness of these strata and the estimated time that appeared to be represented, led to the proposal (Wagner 1969; George and Wagner 1972) that a new chronostratigraphic time interval should be recognized for this succession. The Cantabrian Stage was conceived as occupying the interval between the Westphalian D Stage as defined in the Lorraine Coalfield and the Stephanian A as defined in the St. Étienne Coalfield.

There followed a lengthy and sometimes impassioned debate within the SCCS during the 1960s and 1970s (recorded in the published proceedings of the various Subcommission meetings, 1965, 1969, 1971, 1973) that focused on a small number of issues, which may be summarized:

• Should the succession in northern Spain be considered a separate stratigraphic unit (stage); or could some horizon within this succession be identified and accepted as a marker for the base of the Stephanian, and strata below and above this allocated respectively to Westphalian D and Stephanian A (see Boersma 1977);

• If a new Cantabrian Stage (now substage) were recognized as an independent stratigraphic unit, what biostratigraphic criteria (species) were indicative of the unit (Bode 1967; Boersma 1977)?;

• If the Cantabrian Stage were recognized, should this be considered the uppermost stage of the Westphalian or the basal stage of the Stephanian; in other words, is the base of the Stephanian to be defined at the base of the Cantabrian or at the top of this unit?

Nelson and Lucas (2021) have since criticized the 1972 decision as being premature, pointing out that there have been disagreements about choosing the basal boundary stratotype. Although there has been one formal change in the designated boundary stratotype (Knight and Álvarez-Vázquez 2021), the conceptual basis of the boundary using megafloral biostratigraphy (later defined as the base of the Odontopteris cantabrica Zone) has remained consistent since the late 1960s. This criterion is now used to indicate (but not define) the boundary in a naturally exposed stratotype section in the North Palencia Basin, now ratified as the boundary stratotype (Engel 1989). The base of the substage can be clearly recognized biostratigraphically in its type area and, similarly, its upper boundary, the base of the Barruelian, is also clearly identifiable on biostratigraphical criteria (Wagner et al. 1983; Wagner and Winkler Prins 1985a, b; Cleal et al. 2003; Wagner and Álvarez-Vázquez 2010a). The Cantabrian has also been recognized in other Euramerican sequences outside of northern Spain (e.g. Bouroz et al. 1970, 1972; Cleal 1978, 1997, 2007; Zodrow and Cleal 1985; Cleal et al. 2003, 2009; Opluštil et al. 2016a, 2017, 2018, 2021a). The present work reports the basis of recognizing a time interval of some 2.6 Ma for this substage (Figs 3 & 6 and later discussion), which corresponds to a similar length of time recognized for other substages in the Pennsylvanian Subsystem (Davydov et al. 2010, 2012).The debate about the definition and description of the Cantabrian substage has called into clear focus the distinction between chronostratigraphic units defined at boundary stratotypes in rock sequences (Hedberg 1965, 1968; Salvador 1994), and biostratigraphic zones defined purely on the distribution of the fossils. The macrofloral taxa found in the lower Stephanian tend to be long-ranging and reflect a gradually changing vegetation (e.g. Wagner et al. 1983; Wagner and Álvarez-Vázquez 2010a); there are no ‘index’ species that exclusively define the Cantabrian as a unit (see comments by Boersma 1977; Havlena 1977). The lack of specific index fossils has been used to question the validity of the Cantabrian (Nelson and Lucas 2021). These concerns fail to recognize that the Cantabrian is explicitly a chronostratigraphic unit. The linkage to biostratigraphy is discussed further below regarding the use of megafloral biozones, which are established as assemblage zones (in the sense of Salvador 1994), and therefore not defined by individual ‘index’ taxa. In summary, the Cantabrian is unequivocally defined by boundary stratotypes, is clearly identifiable biostratigraphically and represents a similar length of time to the other substages in the Pennsylvanian Subsystem.

The classic Stephanian A was based on the Faisceau de la Peronnière in the Assise de Rive-de-Gier (St Étienne Coalfield) of the Massif Central of France (Pruvost 1934; Jongmans and Pruvost 1950), and was later somewhat extended by correlation with the successions in the Cévennes Coalfield (Zone 2) and in the Carmaux Coalfield (Zone de Lentin) also in the Massif Central (Bouroz et al. 1969; Gras 1970; Bouroz and Doubinger 1978). The interval is characterized by a distinctive macroflora, today referred to as the Crenulopteris lamuriana Zone (see later).

However, the SCCS Working Group on the Stephanian (Bouroz et al. 1970) concluded that the Massif Central was unsuitable to locate a stratotype due to interrupted continuity of sedimentation, and limited exposure and accessibility (Liabeuf in discussion, Wagner 1969; Bouroz et al. 1970). It was therefore proposed to define the base of the Stephanian A in a stratotype in the Barruelo Coalfield (Fig. 5), in the North Palencia Basin (Wagner and Winkler Prins 1985a, b); this has been ratified by the SCCS and the substage renamed Barruelian (Engel 1989). Based on the range of the Crenulopteris lamuriana Zone here, the base of the Barruelian is located at the base of the coal-bearing Carboneros Member (Fig. 6). A pyroclastic tonstein from the Peñacorba Member, some 300 m stratigraphically below the Carboneros Member, has given a robust radiometric age (close to 305 Ma, further discussed below).

The upper Barruelian is absent from the Barruelo Coalfield but occurs in the lower part of the Sabero Coalfield (Fig. 5), some 65 km to the west, which also yields Crenulopteris lamuriana Zone macrofloras similar to those of the upper part of the Barruelo succession (Calero Member, Fig. 6); the correlation of the Barruelo and Sabero sequences is discussed later. In the upper part of the Sabero Coalfield succession, however, rather different, Alethopteris zeilleri Zone macrofloras occur, more similar to those of the Zone de Tronquié in the Carmaux Coalfield in the Massif Central of France. This is taken to mark the base of the Saberian, and therefore the top of the Barruelian (Knight and Wagner 2014); radiometric dating of tonstein bands that bracket the proposed boundary stratotype indicate an age of 303.5 Ma for this horizon (discussed further below).

The United States of America developed a different chronostratigraphic scheme for Pennsylvanian-age strata (as summarized by Menard 1979), and there has been some controversy about how these relate to the west European chronostratigraphic framework. In recent years, correlations of the grey, coal-bearing sequences (lower and middle Westphalian in Europe, Morrowan–Atokan in the North American mid-continent region) has improved significantly (Blake et al. 2002; Opluštil et al. 2021a) but correlations of the younger Pennsylvanian sequences between western Europe and North America remain controversial.

Various reports (e.g. Peppers 1996, DiMichele et al. 2009; Falcon-Lang et al. 2011b) have correlated the ‘traditional’ Westphalian–Stephanian boundary in Europe with the Desmoinesian–Missourian boundary in the mid-continent region of North America, but this is not based on detailed correlative evidence. Although the Westphalian–Stephanian and Desmoinesian–Missourian boundaries both mark significant changes in facies and palaeontological content, it is unlikely that these changes were being driven synchronously by the same forcing factors across Euramerica. There was undoubtedly significant global climate change occurring at about this time with the onset of a major inter-glacial of the Late Paleozoic Ice Age (e.g. Fielding et al. 2008; Montañez and Poulsen 2013), but the impact of that climate change on sedimentary patterns and biotas (in particular vegetation) varied considerably as it interacted with the dynamically changing landscapes in different parts of Euramerica. This was particularly evident in the areas adjacent to the Variscan–Appalachian Mountains that were undergoing significant tectonic disturbance at this time. For instance, the change from lycopsid-dominated to marattialean-dominated swamps that is taken as an important biostratigraphic indicator of the Desmoinesian–Missourian boundary in North America, has been shown to be highly diachronous across Euramerica (Cleal et al. 2009). It is apparent that for any attempt to correlate the chronostratigraphic frameworks of western Europe and the Appalachian and Mid-Continent basins of North America, it is essential to decouple the simplistic link between the Westphalian–Stephanian and the Desmoinesian–Missourian boundaries.

Any exercise of correlating the biostratigraphy of the various later Pennsylvanian successions in the European region must necessarily reflect the fact that this was a period of intense tectonic activity, related to the late phases of the Variscan Orogeny, which impacted on virtually all the coeval sedimentary basins and correspondingly on their fossil content. A palinspastic reconstruction of the principal European areas of sedimentation during the Late Pennsylvanian is presented as Figure 4. Correlations are hampered by the global climate change occurring at this time, reflected in the demise of the first (major) glacial phase of the Late Paleozoic Ice Age and the onset of the Late Pennsylvanian Interglacial (Fielding et al. 2008; Montañez and Poulsen 2013) which had a widespread effect on biotas, especially the floras (e.g. Gastaldo et al. 1996; Cleal and Thomas 1999, 2005; Cleal et al. 2009). This summary regional overview endeavours to underline the interpretation of an extensive southern foreland to the Variscides, and also the very different palaeogeographical and tectonic contexts in which important reference successions of western Europe were developed.

The broad overall tectonic picture invokes a once continuous orogenic belt involving the Ouachita–Appalachian–Variscan sectors that formed during the convergence of the Gondwana and Laurussia supercontinents and culminated with the completed assembly of Pangaea by the beginning of the Permian (Kroner and Romer 2013). The current orientation of the fragments of the Variscan orogenic belt and overprint of later orogenic episodes have given rise to complex and often disparate interpretative models for specific sectors of the European Variscides. A number of discrete tectonic zones (Figs 2 & 4) have traditionally been recognized within the widespread European fold-and-thrust structure and show general continuity along the orogenic belt (Franke 1989, 2000; McCann et al. 2008). However, the relationship of the Variscan lithotectonic units of the Iberian Peninsula to those of west-central continental Europe remains enigmatic.

The interpretation of an integrated picture of the Variscan Orogeny invokes multiple microplates, multiple subduction zones with opposed subduction polarities, complex thrust structures and nappe piles, regionally complex magmatic histories and major transcurrent faults (Kroner and Romer 2013). Episodic plutonism continued into the late Carboniferous and early Permian through the west European trace of the Variscides, including the Moldanubian Zone represented by the Bohemian Massif, Vosges Mountains and the Massif Central of France (Lardeaux et al. 2014; Tabaud et al. 2014; Žak et al. 2014), also in the Ibero-Armorican Arc (Ballèvre et al. 2014; Martínez Catalán et al. 2014) and the Pyrenees (Carreras and Druguet 2014; Denèle et al. 2014). Large scale dextral shear faulting is recognized as operating during the late Carboniferous–Permian in west and central Europe (Edel et al. 2014 and references therein).

The northern margin of significant Variscan deformation is relatively well defined and identified as the Variscan Front (Figs 2 & 4), which is the northern boundary of the Rhenohercynian Zone. This zone lies north of and external to the line of the major suture, the Rheic Suture, which marks the margin of the strongly deformed internal zone of the orogenic belt (Kroner et al. 2008). The northern foreland of the Variscides comprises (1) the Rhenohercynian Zone, (2) a contiguous foredeep basin area lying immediately north of the Variscan Front characterized by rapid deposition, but which was subject to significant crustal shortening and deformation (e.g. Jones 1991; Gayer et al. 1993), and (3) a cratonic area with less deformed sequences, separated from the foredeep area by a forebulge known as the Wales–Brabant ‘Massif’ (Deckers and Rombaut 2021). This extensive foreland area north of the Variscan Front contained interconnected flexural basins with broadly similar basin fills, including the important European coalfields (Fig. 2), from South Wales and the Pennine Basin of Britain, extending across Nord-Pas-de-Calais, Limburg–Campine, the Ruhr and through to the Upper Silesian Basin of Poland, the Dobrudzha in Bulgaria and Zonguldak in northern Turkey (Opluštil and Cleal 2007; Cleal et al. 2009).

In the internal, deformed core of the central and west European Variscides (Moldanubian Zone), compressional tectonics had effectively ceased by the late Carboniferous, and a common late-stage history can in general terms be recognized. Extensive dextral wrenching occurred through the internal zones of the Variscides, leading to a number of fault-bounded graben, half-graben and narrow troughs that hosted contemporaneous coal-bearing successions. These basins tended to be relatively small, forming in isolation from each other and with their own internal drainage systems (Timmerman 2004); they include the Lorraine–Saar–Nahe Basin, the Saale Basin, the Late Pennsylvanian (Stephanian) basins of Bohemia, the Stangnock Basin of the eastern Alps and the basins of the Massif Central of France. In the Massif Central the location of individual basins is controlled by the interaction of large-scale wrench faults and subsidiary splay faults (Arthaud and Matte 1977), which remained active and controlled the sedimentary basin-fill.

The west European end of the Variscides is characterized by the presence of the large, curved oroclinal feature, involving the South Armorican Massif and NW Iberian Massif, referred to as the Ibero-Armorican Arc (Ballèvre et al. 2014), of which the core, in current orientation concave to the east, comprises the tightly curved Cantabrian Orocline (Gutiérrez-Alonso et al. 2012). This has been interpreted to be only a segment of a more continuous ‘S’-shaped fold belt, extending into the Central Iberian Orocline (Martínez Catalán 2011; Shaw et al. 2012). The mechanism of buckling this double oroclinal feature remains an open discussion for which a number of interpretations have been developed (see discussions in Martínez Catalán 2011; Shaw et al. 2012; Weil et al. 2012; Fernández-Lozano et al. 2016). However, there is a general consensus that the oroclinal trace was originally a linear feature reflecting the transition from a hinterland, comprising the internal zone of an Iberian Orogenic Belt, which under current orientation lay to the west, and a foreland sedimentary belt to the east, marginal to the Paleotethys Ocean (Shaw et al. 2012; Weil et al. 2012); this forms a southern foreland area (Fig. 4) to the overall Variscan belt. This model envisages that a linear fold and thrust belt developed in the foreland area during Moscovian times and was subsequently reconfigured by longitudinal arc-parallel shortening, giving rise to buckling through the Kasimovian and Gzhelian, culminating in closure to the present oroclinal configuration before earliest Permian time (Weil et al. 2012).

Casas and Murphy (2018) have modelled the configuration of an essentially linear orogenic margin or foreland belt for the Iberian Peninsula through successive stages of compression and buckling to the current post-Variscan configuration of the Cantabrian Arc at the beginning of the Permian. This construct provides an interpretation of the relative geographical relationship for the Cantabrian area, and the Massif Central of France, the two key reference areas for Late Pennsylvanian chronostratigraphy. It is also a guide to the relative position of other structural units such as the Pyrenees. Figure 4 incorporates the model of this foreland area through the late stages of the Pennsylvanian.

The presence of a broad area identifiable as a southern foreland to the Variscan orogenic belt, extending from the Iberian Peninsula into the current area of southeastern Europe, has been intimated as part of a number of palaeogeographical interpretations (Opluštil and Cleal 2007; Cleal et al. 2009; Merino-Tomé et al. 2009). The sedimentary basin-fill at the core of the Ibero-Armorican Arc has for many years been recognized as comprising the east and southwards facing foreland basin to the orogenic core of the Iberian Massif (Wagner et al. 2002, Merino-Tomé et al. 2009). However, the palaeogeographical significance of a wider area of Late Pennsylvanian sedimentation on the southern side of the Variscan Mountains, and its influence on a regionally distinct floral biota was first explicitly recognized by Cleal et al. (2015) with respect to the fossil flora of the Velebit Mountains of Croatia (‘I’ in Fig. 4). Within the area subsequently affected by intense deformation in the Alpine orogenic episode, similar allochthonous, tectonically isolated basins occur in the Jesenice area of Slovenia (Pšenička et al. 2015) and the Italian Carnic Alps. In the latter area the Nassfeld Basin (‘H’ in Fig. 4) includes a marine and paralic succession of Kasimovian through to at least Sakmarian age, formed on the southern margin of the Variscan Mountains (Novak et al. 2019; Opluštil et al. 2021b). Specific reference to a southern foreland of the Variscan Mountains has been first mentioned by Besly and Cleal (2021) and further developed by Opluštil et al. (2021a).

A characteristic feature of the Late Pennsylvanian (Stephanian) in all areas of the Variscan orogenic belt is the frequent and widespread occurrence of volcanic rocks, both as extrusives and as air-fall tuffs; the location and character of late Carboniferous volcanism in central Europe has been comprehensively summarized by Timmerman (2004, 2008). However, the source areas and parent volcanic edifices are generally not easily recognized. Sites of explosive volcanism were a feature of a number of sedimentary basins, including the Intra-Sudetic Basin of SW Poland (Awdankiewicz 2004) and the Kladno-Rakovník Basin of west-central Czech Republic (Třtěno Volcanic Complex; Opluštil et al. 2016b). In the Massif Central of France, a considerable number of air-fall tuff horizons have been recognized through the Stephanian succession (Bouroz 1968: some 45 cinerite–tonstein bands in the correlated successions of the St-Étienne, Carmaux, Cevennes and Decazeville coalfields), and local extrusive volcanism is represented by rhyolite lava flows, such as the Gore Vert de Grand Croix in the Assise de Rive de Gier (Doubinger et al. 1995). In the Pyrenees locally thick volcano-sedimentary units occur within discrete intramontane basins; ignimbrites, pyroclastic flow deposits and lahars are a feature of the Unidad Gris attributed a mid-Stephanian (Stephanian B) age (Martí and Gisbert 1985). In the south-central Iberian Peninsula, recent radiometric dating (current investigation by the authors) indicates that the local volcanic edifice in the Puertollano Basin (Ciudad Real) is of mid-Stephanian age.

The dynamic tectonic evolution of the west European area of the Variscides during the Late Pennsylvanian resulted in four distinct palaeogeographical areas and depositional environments (Fig. 4) in west and central Europe: (a) the northern foreland area, (b) intramontane basins within the Variscan internal zones, (c) the southern foreland and (d) the marine Picos de Europa platform.

The Variscan northern foreland is characterized by numerous large coal basins where extensive mining has resulted in a wealth of sedimentary and palaeontological data (Cleal et al. 2009). The coals are the result of peat deposits formed by the coal swamp biome that during the early half of the Westphalian extended across much of the foreland, including both the foredeep adjacent to the orogen and the cratonic area to the north. However, the coal swamp biome started to contract across the northern foreland during the middle – late Westphalian. Ever-wet, coal-swamp conditions had all but disappeared from the northern, cratonic parts of the foreland by the mid-Bolsovian (Dreesen et al. 1995), with later deposits being predominantly red-beds with only a few coals (Besly 1988; Besly and Cleal 1997, 2021; see Fig. 3). The swamps persisted in the more southern foredeep areas such as South Wales, Somerset and Germany (e.g. Josten et al. 1984; Thomas and Cleal 1994; Cleal 1997, 2007; see Fig. 3) albeit with some better-drained ‘dry-spots’ supporting non-swamp vegetation (e.g. Falcon-Lang et al. 2011a; Wood et al. 2022), but even here the swamps had disappeared by the early Cantabrian.

The uppermost Carboniferous deposits of the northern Variscan foreland are alluvial red-beds with local disconformities, and have been recognized offshore of western Ireland, through the English Midlands, the southern North Sea and in isolated European successions through to eastern Germany (e.g. Besly 1988; Besly and Cleal 1997, 2021; Fig. 4). Correlation of these red-beds is hindered by the absence of marine bands or any other marker horizons, and a very poor fossil record mainly limited to occasional vertebrate body and trace fossils (Werneburg and Schneider 2006; Schneider et al. 2020).

Intramontane basins developed in the internal sectors of the Variscides during the late Westphalian and Stephanian: in the Massif Central, Lorraine–Saar–Nahe Basin, Bohemia, and the Pyrenees (Fig. 2). These were fault-bounded basins, each with its own internal fluvial and/or lacustrine systems and no marine influence. Basin development was subject to strong tectonic control resulting in relatively short sequences bounded by depositional hiatuses. Doubinger (1960) demonstrated that there was some endemicity in the vegetation between basins, but the macrofloras have nevertheless been the main basis of stratigraphical correlation, supplemented more recently by palynology, insect wings and conchostracans (Schneider and Scholze 2018; Schneider et al. 2020), and tonstein petrography and radiometric dating (e.g. Bouroz 1967; Pellenard et al. 2017).

Based on hydrogen isotope data, Dusséaux et al. (2019) estimated that the Massif Central basins were 3000–4000 m above sea level during the late Westphalian, and Becq-Giraudon and van den Driessche (1994) and Becq-Giraudon et al. (1996) interpreted deposits in the Autunian basins of Central France as periglacial, indicating elevations of 4000–5000 m. Based on the distinctive nature of their fossil floras, Holub et al. (1977) and Tenchov (1976, 1977) estimated elevations of 1000–2000 m for the Bohemia and Svoge coal basins, and Opluštil (2005) estimated that the Bohemia basin was formed at about 1000 m above sea level based on modelling of drainage patterns.

Regional strike-slip faults (Courel 1987) and extensional faulting controlled the development of a number of isolated basins, some with high sedimentation rates and thick successions in excess of 5000 m (e.g. the Cévennes and St Étienne coalfields). Synsedimentary differential uplift within basins resulted in local half-graben structures, and gravity-controlled folding and décollement with substantial thrust displacements, which has significantly disrupted sedimentary continuity (Bouroz 1978; Djarar et al. 1996). Sedimentation at basin margins was commonly on large alluvial fans, which more distally become interbedded with coal-bearing successions, often with numerous seams, some of which attain great thickness (up to 50 m at Montceau-les-Mines); these are associated with well-developed seat-earths that may become locally bauxitic (Besly 1987; Courel 1987) and occasional localized lacustrine intervals also occur. The lower part of the St Étienne Coalfield succession, comprising the Assise de Rive-de-Gier (Fig. 3), was the original type of the Stephanian A, now Barruelian (Jongmans and Pruvost 1950; Doubinger and Vetter 1985a, b); only the relatively thin (50 m) Faisceau de la Péronnière was coal-bearing and yielded fossils (Bouroz et al. 1972). Overlying the Assise de Rive-de-Gier, the Assise de St Étienne (Fig. 3) was regarded as the type of the Stephanian B (Jongmans and Pruvost 1950). However, the two intervals are now known to be separated by a sizeable stratigraphical gap (Mattauer and Matte 1998) representing the interval now represented by the Saberian (Knight and Wagner 2014).

This basin formed in a half-graben located between the Rhenohercynian and Saxothuringian zones of the Variscides (see Fig. 2). The Saarbrücken Group (terminology of the German part of the basin, equivalent to Assise de la Houve, see Fig. 3) consists of a continuous succession of c. 2500 m of upper Duckmantian to Asturian coal-bearing grey strata (Kneuper 1966, 1971; Cleal 1984a; Schäfer 1989, 2011). The Saarbrücken Group is overlain by 2500 m of Stephanian alluvial and lacustrine red-beds, known as the Ottweiler Group (Fig. 3), with the contact marked by the Holz Conglomerate (Kneuper 1971); the base of the Holz Conglomerate has been shown to represent a significant hiatus of c. 3.6 Ma (Cleal 2008a; see Fig. 3). A literature review by Cleal (2008a) has shown that the rather limited lower Ottweiler Group floras belong to the Alethopteris zeilleri Zone, indicating a Saberian age.

The late Carboniferous continental deposits of the northern Czech Republic are remnants of a large intra-montane basin-complex including the Central and Western Bohemia, Intra-Sudetic, Krkonoše-Piedmont, Mnichovo Hradiště and Česká Kamenice basins (Pešek 2004). They formed on a high-altitude plateau (Opluštil 2005), on uplifted granites and high-grade metamorphic rocks, as fault-related extensional or strike-slip basins, whose formation was accompanied by explosive acid volcanism that produced extensive ash deposits. The overall succession comprises basin-wide alternations of grey, coal-bearing sequences and fluviatile red beds with impersistent high ash coals, and also a number of grey lacustrine horizons (Opluštil et al. 2013, 2016a, b, 2017). The lower–middle Westphalian (Radnice Member in central and western Bohemia, Žacléř Formation in the Intra-Sudetic Basin) consists mainly of grey, coal-bearing sequences containing remains of coal swamp vegetation that has a broadly similar composition to that found in the northern foreland sequences (Cleal 2008b, c; Šimůnek and Cleal 2020). There is a major hiatus before the next-youngest strata: the Nýřany Member in central and western Bohemia, and the Svatonoviče Member in the Intra-Sudetic Basin. Macrofloras from both members are of the upper Crenulopteris acadica and Odontopteris cantabrica zones indicating late Asturian and Cantabrian ages (Opluštil et al. 2016a, 2017); radiometric dating supports interpretation of the Nýřany Member as ranging from latest Moscovian to early Kasimovian (Opluštil et al. 2016a). Another significant hiatus intervenes before the latest successions in these basins, which are dominated by red-bed facies. Ashfall tuffs have yielded radiometric dates that indicate these successions as ranging from near the Kasimovian–Gzhelian boundary through the Gzhelian (Opluštil et al. 2016a); macrofloras represent a Mesophytic vegetation with conifers and peltasperms that thus far do not provide a basis for biostratigraphical correlation.

wThese comprise a number of elongate, structurally confined basins (Surroca–Ogassa in Fig. 2) in a narrow strip along the southern margin of the axial zone of the Pyrenees (Laumonier et al. 2014; Casas and Murphy 2018). They classify as ‘molassic’ and reflect the initiation of Late Pennsylvanian continental deposition on a metamorphic and plutonic basement. The stratigraphy of these basins is recognized as a series of units separated by disconformities (Broutin and Gisbert 1985). The oldest unit, the Aguiró Formation, is preserved as alluvial fan deposits and has yielded a late Asturian to earliest Cantabrian macroflora (Talens and Wagner 1995). The overlying Unidad Gris is mostly andesitic volcanics with some ignimbrites (Martí and Gisbert 1985), and in the upper part includes lacustrine and alluvial sediments with seat-earths and a number of coal seams with associated fossil flora (Besly and Collinson 1991). The flora has been ascribed a Stephanian B–C age (Talens and Wagner 1995), suggesting a considerable time gap at the unconformable contact with the Aguiró Formation. Palynology yielded a characteristically Stephanian hygrophilous microflora, with many fewer xerophytic elements compared with other intramontane basins (Juncal et al. 2019).

The area now preserved in the Cantabrian region (Fig. 5) exposes a wide area of marine shelf sediments in an apparently passive margin setting from the Early Devonian, and evolving as a foreland basin through the course of the Carboniferous. A broad pattern of near-shore sediments to the west and deeper offshore sediments to the east (García-Alcalde et al. 2002) persisted into the Carboniferous, when the Ibero-Armorican orogenic belt formed a rising and encroaching hinterland to the west of the foreland depositional area. Intensive deformational events, reflecting the complex geometry of the originally linear orogenic belt, and the north–south compression leading to buckling of the Ibero-Armorican Arc, started in the foreland area in the early Westphalian (late Langsettian). This phase of deformation (the Palentian folding phase of Wagner 1965) was sharply delimited in time, and generated folds, large scale thrusting and nappes. Thereafter, the Late Pennsylvanian succession of the Cantabrian area can be interpreted as three basin-fill successions (Fig. 6) punctuated by brief, diachronous and locally intense deformational episodes.

Succession 1: from late Langsettian through to mid-Asturian time, predominantly marine but dominated by terrigenous sediments commonly associated with delta lobes, fan deltas and the important coal-bearing sequence of the Central Asturian Coalfield; limestones are generally subordinate. Now preserved in the tightly deformed structure of the Cantabrian Arc, it is clear that this depositional basin encompassed a much wider palaeogeographical extent (4 in Fig. 4). A mid-Asturian deformational episode, the Leonian folding phase of Wagner and Martínez-García (1974), mainly affected northern León and central Asturias, giving rise to a strongly angular unconformity. Further east in Palencia it is evident only as a widespread disconformity associated with uplift and some extensional faulting.

Succession 2 (post-Leonian in Fig. 5): the oldest deposits, of mid-late Asturian age, occupy the strongly subsiding North Palencia Basin, limited to the west by onlap on to the previously folded succession (Tejerina area, Wagner et al. 1969) and to the east by synsedimentary faulting that separated a mainly siliciclastic basin from a more stable carbonate platform in the Barruelo area (Wagner et al. 1977). The basin-fill (the lower part of Area B in Fig. 4) repeats the pattern of predominant terrigenous sediments against the western border, while marine influence increases eastwards. Alternating marine and non-marine intervals have been interpreted to reflect a number of marine transgressions driven primarily by local tectonic control (Wagner and Fernández García 1983). The later part of the post-Leonian succession is marked by widespread development of coals and seat-earths in the Barruelo Coalfield. The tectonic movements of the Asturian Folding Phase occurred in mid-Barruelian time (Wagner 1965; Wagner and Martínez-García 1974) with locally intense folding and thrusting, particularly over the area of the post-Leonian sedimentation and specifically in the North Palencia Basin, although in the core of the present tight Cantabrian Orocline, the Picos de Europa carbonate platform records only uplift and erosion.

Succession 3 (post-Asturian in Fig. 5): the Asturian deformation (Wagner 1965) was of short duration, and the oldest sediments of the succeeding succession, recognized in the Sabero Coalfield, are of mid-late Barruelian age. This succession is now represented in a number of tectonically isolated coalfields preserved around the perimeter of the Cantabrian Orocline. The overall model is of a single basin-fill (Area ‘B’ of Fig. 4) on-lapping westwards (Wagner and Castro 2011) on to hinterland marginal to the main orogenic belt. The hinterland was locally highly dissected and probably affected by some synsedimentary faulting, so that some segmentation and differential subsidence, possibly as wedge-top basins, cannot be ruled out. The western marginal area of the basin-fill presents a paralic environment interpreted to connect eastwards to a wider area of more pervasive marine influence. Marine sediments are proven in the lower part of the Sabero succession (Eagar and Weir 1971; Iwaniw 1984). The coal-bearing successions throughout these coalfields are characterized by extensive development of stigmarian seat-earths, and large arborescent lycopsid tree compressions are a characteristic feature of the coal seam roof measures; these are common features through to the youngest successions recognized in the post-Asturian sequence, at Puerto Ventana, where a pyroclastic ashfall tonstein has provided a high precision U–Pb CA-TIMS date of 300.5 ± 0.3 Ma (horizon 11 on Table 1).

The broader interpretation of the orogen of the Iberian Massif and southern foreland of the Variscides (Casas and Murphy 2018), supports an interpretation of the geological context of the Puertollano Coalfield, Ciudad Real (Area ‘A’ of Fig. 4), which has hitherto been enigmatic. Descriptions of the succession, as representing an alluvial plain with lacustrine intervals, have emphasized that this basin was probably coastal (Wagner 1999; Wagner and Álvarez-Vázquez 2010a, 2015), although no direct evidence has been identified. The presence of the fossil remains of euselachian sharks and acritarchs (Soler-Gijón 1999) has been considered strongly suggestive of, at least sporadic, connection to the sea. There is strong evidence of wide lateral extent for this basin, far greater than its present area (Wagner 1985), with a rising hinterland to the west and potential marine connection to the east; a position analogous to the Stephanian basins of the Cantabrian region. The fossil flora comprises a record of predominant hygrophilous elements (lycopsid trees, marattialean ferns, sphenopsids), to a large extent characteristic of the post-Asturian succession of the Cantabrian area, plus conifers, assumed to be drifted ‘extra-basinal’ elements (Wagner and Álvarez-Vázquez 2010a; Álvarez-Vázquez et al. 2018, 2022). A number of samples of the volcanic horizons associated with exploitable coals have been submitted for radiometric dating and formal publication of results is still pending; provisional results (the present authors and Opluštil pers. comm.) indicate dates somewhat older than those already published for the Saberian succession (Knight and Wagner 2014) and consistent with the transition Barruelian–Saberian.

Further east in the heavily tectonized Alpine-Carpathian domain, a number of allochthonous Upper Pennsylvanian sequences provide further evidence of the extensive southern Variscan foreland. In the Carnic Alps of Italy and Slovenia, the Auernig–Nassfeld Basin (‘H’ in Fig. 4) has transgressive–regressive cycles with marine limestones containing mid-Kasimovian to mid-Gzhelian faunas (Novak et al. 2019) and terrestrial intervals with upper Crenulopteris lamuriana to lower Sphenophyllum angustifolium zone macrofloras (Pšenička et al. 2015; Opluštil et al. 2021b). Similar sequences are seen in the Balkans (Vozárová et al. 2009). For instance, the Lika region and the Velebit Mountains of Croatia (‘I’ in Fig. 4) have predominantly marine carbonates with fusulinid faunas (Sremac 2005) interbedded with terrestrial clastics containing Sphenophyllum angustifolium Zone macrofloras (Cleal et al. 2015). The fusulinoidean foraminifera faunas in these carbonate successions broadly correspond to the succession of fusulinid zones recognized in the Picos de Europa Province (Merino-Tomé et al. 2009) and the global framework (Davydov et al. 2012).

In the context of the southern foreland of the Variscides, the Picos de Europa platform (Figs 4 & 5) represents a unique example of a large, long duration carbonate platform. Its development has been interpreted as being controlled by its location in the distal part of a large marine foreland basin; asymmetric subsidence generated a rapidly subsiding foredeep adjacent to the advancing orogenic front while the more distal and slowly subsiding sector allowed development of the carbonate platform (Merino-Tomé et al. 2009). Its current position in the core of the Cantabrian Orocline and its structural history are directly linked to the closure of the Ibero-Armorican arc and oroclinal bending; it comprises an imbricate structure of thrust sheets with translation towards the south and SW. Emplacement commenced during the latest Moscovian (mid- to late Myachkovian) and continued through into the Gzhelian (Merino-Tomé et al. 2009). A rich fauna, in particular of fusulinids and brachiopods, supports correlation of successive depositional sequences across the interval from late Moscovian (upper Myachkovian), through the Kasimovian and into the Gzhelian (Sánchez de Posada et al. 1999; Merino-Tomé et al. 2006). The most complete stratigraphic record from upper Moscovian through to mid-Kasimovian is found in the Las Llacerias section in the NW part of the unit; fusulinoidean biozones (Merino-Tomé et al. 2006) were correlated with those of the Russian Platform and with the somewhat sparse fusulinid record of the North Palencia Basin (van Ginkel 1965). Conodonts from the Las Llacerias section (Méndez 2006) show affinities with those of the Moscow and Donets basins and broadly support the stratigraphical interpretation established on the fusulinoids.

Unravelling the Pennsylvanian history of western Europe has necessarily been heavily dependent on floral biostratigraphy for regional correlations as no other group of fossils is as ubiquitous in these coal-bearing successions. Grand'Eury (1877) and Kidston (1894) recognized that different stratigraphical intervals contained distinct floral assemblages, and these assemblages were given zonal names by Zeiller (1894) and Bertrand (1914, 1918, 1920); see also Bertrand (1937) for a summary. A similar approach was later used by Read (1947) and Read and Mamay (1964), who advanced a macrofloral zonation for the sequences in North America. However, none of these were biozones in the sense that we would recognize today, rather they were unbounded assemblages characteristic of the lithostratigraphical intervals in which they occur, rather than intervals defined by the intrinsic distribution of the fossils.

A change occurred with the work of Dix (1934, 1937), who, for the first time, used stratigraphical range charts to define biozones in the Namurian–Westphalian coal-bearing sequences of South Wales. The biozones were defined purely by the detailed ranges of the fossil-taxa, and not by the age or lithology of the strata they were found in, and so avoided potential circular arguments when making stratigraphical correlations (Burek and Cleal 2005). A similar approach was also used by Bell (1938) for the Sydney Coalfield in the Canadian Maritimes. Dix's ideas were further developed by Wagner (1984), who used range charts to try and recognize floral biozones through the entire Carboniferous and across the palaeoequatorial wetlands of Euramerica. However, Wagner (1984) was partly reverting to the older ideas of Zeiller and Bertrand in that he tried to make the biozones at least partly reflect the chronostratigraphical intervals. Although Wagner (1984) emphasized the important distinction between biostratigraphy and chronostratigraphy, there was a certain logic to this as some of the stages/substages had originally been defined in part using palaeobotanical indicators (Jongmans and Pruvost 1950; Bode 1970). However, it meant that Wagner's (1984) biostratigraphy does not fully reflect the dynamics of vegetation changes occurring in this biome. Consequently, Cleal (1991) refined Wagner's (1984) zones, identifying subzones corresponding to the change in the macrofloras recognized in the Dix (1934) scheme. Although a compromise, the scheme now seems to provide a better reflection of the vegetation changes that occurred in the wetland biome during the Pennsylvanian, and has proved to be of practical merit (as reviewed by Opluštil et al. 2016a).

Wagner (1984) described his zones as concurrent range zones but they are in fact assemblage zones: the zonal boundaries are defined at biohorizons (sensuSalvador 1994; Cleal 1999) where particular taxa appear (First Appearance Data – FADs) or sometimes disappear (Last Appearance Data – LADs) from the succession. However, as many of the macrofloral species are long ranging and the changes in fossil plant assemblages gradational, the choice of biozone boundaries may sometimes seem somewhat arbitrary (Cleal 1991). Moreover, there is often no particular taxon whose range defines a zone, which has caused problems with the naming of zones. Wagner (1984) often named a zone after two of the species that typically occurred, but this led to a rather cumbersome nomenclature, so Cleal (1991) argued that just one taxonomic name should be used for each zone. Following Salvador (1994), these names are merely labels to identify the zone and there is no implication that the eponymous taxon is either restricted to or occurs throughout the zone. For instance, Alethopteris zeilleri occurs both above and below the zone of that name, but is nevertheless a characteristic species of those floras (Wagner 1968). Although the zones may be indicated by such taxa, they are defined by biohorizons.

Accurately placing the biohorizons can be difficult, as plants respond to various drivers such as climate and landscape change, as well as phyletic change. It has been shown in middle Westphalian sequences, for instance, where there is a tight chronological control of the correlations based on widespread isochronous marine bands, there are variations in the detailed distribution of some of the macrofloral taxa (e.g. Cleal 2005, 2007; Cleal et al. 2009). To test for some of these problems, Laveine (1977) developed a model for identifying the base of what is now known as the Linopteris obliqua Zone. Rather than just using the FAD of Neuropteris ovata, which was traditionally used to define the base of the zone, he identified a series of other biohorizons (‘events’) just above and below that level. The base of the zone should only be placed at the lowest occurrence of N. ovata if the other biohorizons followed the normal pattern seen in most sequences. Using this principle, Zodrow and Cleal (1985) were able to show that the base of the Linopteris obliqua Zone was significantly lower stratigraphically in the Sydney Coalfield of the Canadian Maritimes (‘1’ on Fig. 4) than the FAD of N. ovata. A similar, approach was subsequently used by Cleal et al. (2003) to develop a sequential biohorizons model to help in the identification of the base of the Odontopteris cantabrica Zone.

In the Wagner (1984) and Cleal (1991) biostratigraphic scheme, there are two biozones occurring in the lower Stephanian (broadly equivalent to the Kasimovian Stage). The following discussion will summarize these two zones, plus the zones immediately above and below them. The main emphasis will be given to the criteria used to define the biohorizons that indicate the boundaries between the zones. It should be emphasized that these biozones are based on the hygrophilous swamp vegetation associated with the coal-bearing sequences. During the Middle and Late Pennsylvanian, this vegetation was progressively replaced in the northern Variscan foreland by more mesophilous vegetation dominated by conifers, cycadophytes and peltasperms as substrates became drier due to a combination of Variscan orogenesis and climate change. This change from hygrophilous to mesophilous vegetation was historically regarded as of biostratigraphical significance but is now recognized to be strongly diachronous depending on basin development (Galtier et al. 1992; Boyarina 2021). It is desirable that a separate biostratigraphy for the remains of the mesophilous vegetation be developed, to run in parallel with the Wagner (1984) scheme for the hygrophilous vegetation, but no such scheme has yet been proposed, and to develop one is beyond the scope of the present work. Although progressively contracted in aerial extent through the Late Pennsylvanian, the hygrophilous vegetation remained sufficiently widespread for the Wagner (1984) biostratigraphy to be recognizable over large parts of Europe, and even in parts of eastern North America (Opluštil et al. 2021a), where the evidence suggests the biozones are relatively isochronous (Opluštil et al. 2016a).

Originally named the Lobatopteris vestita Zone by Wagner (1984), it had to be renamed due to confusion over the taxonomic nomenclature of the eponymous species (see Wittry et al. 2015). The base of the zone is placed at the biohorizon in the middle Asturian where there is a marked change in the macrofloras, with a significant increase in abundance and diversity of marattialean ferns (also seen in the palynological record, see Laveine 1977). It, in fact, marks one of the most significant changes in the vegetation of the coal swamp biome and has been linked with climate change and the start of the Late Pennsylvanian Interglacial (e.g. Phillips and Cecil 1985; Gastaldo et al. 1996), orogenic landscape change linked with the Variscan Orogeny as especially evidenced in Variscan Euramerica (Cleal and Thomas 1999; Cleal et al. 2007), or a combination of the two (Cleal et al. 2011). In addition to a marked increase in abundance of marratialeans (notably Cyathocarpus, Crenulopteris and Diplazites), the base of the zone is marked by the appearance of medullosaleans Alethopteris grandinii (sensuCleal and Cascales-Miñana 2019), Alethopteris pseudograndinioides, Alethopteris serlii and Callipteridium armasii. This change was identified by Laveine (1977) at the level of Tonstein 60 in the Lorraine–Saar–Nahe Basin (see Fig. 3) and was later found in South Wales (Cleal 1978, 2007), the Sydney Coalfield in the Canadian Maritimes (Zodrow and Cleal 1985), northern Spain (Wagner and Álvarez-Vázquez 1991) and the Donets Basin (Boyarina 2016). It also probably corresponds to the base of the Neuropteris flexuosa Zone in the Read and Mamay (1964) scheme (as reviewed by Opluštil et al. 2016a).

Cleal (1991) found a second change in the middle of the zone with the appearance of Dicksonites plukenetii, Acitheca polymorpha and Crenulopteris acadica. This corresponded to the boundary between zones H and I in the Dix (1934) scheme, and was also identified at the level of Tonstein 40 in Saar–Lorraine by Laveine (1977). This biohorizon was used to divide the zone into lower (Crenulopteris micromiltonii) and upper (Dicksonites plukenetii) subzones.

Wagner (1984) established this zone so that its base coincided with the base of the Cantabrian (sub)stage in its stratotype, although he emphasized that the zonal base does not define the substage – it is merely an indicator of it. Compared with the base of the Crenulopteris acadica Zone, this represents a less sharply demarcated change in the macrofloras; although there are some well-defined FADs at or about this level (e.g. of Odontopteris cantabrica, O. brardii, Sphenophyllum oblongifolium and Callipteridium striatum). Wagner (1984) also emphasized important transitional changes, notably of Crenulopteris acadica to C. lamuriana, Alethopteris pseudograndinioides to Alethopteris zeilleri and Callipteridium armasii to C. striatum.

In order to provide a clearer definition, Cleal et al. (2003) identified four macrofloral biohorizons in the type section for the Asturian–Cantabrian boundary in the Guardo–Cevera Coalfield (Fig. 2) in Palencia (Spain) based mainly on data from Wagner et al. (1983) and Wagner and Winkler Prins (1985a, b).

• Biohorizon 1: FAD (rare) of Alethopteris pseudograndinioides var. subzeilleri.

• Biohorizon 2: FAD of Alethopteris bohemica and Pseudomariopteris corsinii; there is also a significant increase in abundance of A. pseudograndinioides var. subzeilleri (and a concomitant decrease in abundance of the type variety).

• Biohorizon 3: FAD of Odontopteris cantabrica, O. brardii, Sphenophyllum oblongifolium and Alethopteris barruelensis; also LAD of Mariopteris nervosa. There are also rare occurrences of Alethopteris zeilleri and Crenulopteris lamuriana.

• Biohorizon 4: FAD of Alethopteris leonensis (sensuWagner and Álvarez-Vázquez 2010b) and Callipteridium striatum.

While questioning the use of fossil plant biostratigraphy in general, Nelson and Lucas (2021) specifically questioned the taxonomic validity of Odontopteris cantabrica and considered it a problematic species of limited value ‘for delimiting a chronostratigraphic unit’. When first describing O. cantabrica, Wagner (in Wagner et al. 1969) mainly distinguished it from Odontopteris schlotheimii (≡ Odontopteris osmundaeformis auct.) because it occurred in much older floras, with a significant age gap between the two sets of specimens. Wagner suggested that O. cantabrica has smaller, slenderer pinnules but, since he figured only a few fragments with the protologue, the case was not strong. However, larger specimens of O. cantabrica have since been illustrated by Zodrow (1985) and Zodrow and D'Angelo (2019), which confirm Wagner's observation, especially when compared with more recently documented examples of O. schlotheimii (e.g. Barthel 2006; Boyarina 2007). Cuticles might help resolve this taxonomic problem but, although they are known from O. cantabrica (Cleal et al. 2007), they have so far not been obtained from O. schlotheimii. Nevertheless, the concerns expressed by Nelson and Lucas (2021) do not recognize that the megafloral zones are assemblage zones, they are not defined by the range of the specific named taxon applied to the zone and, as biozones, cannot be definitive of a chronostratigraphic unit.

Wagner (1984) regarded the base of this zone as being more or less coincident with the base of the Barruelian Substage as it is now defined (Wagner and Winkler Prins 1970, 1985a, b). The macrofloras again tend to show a gradational change through this interval, with no clear-cut change allowing a defining biohorizon to be identified. Nevertheless, the macrofloras show a clear change in character, which merits the zonal distinction.

The base of the Crenulopteris lamuriana Zone is most clearly marked by the FADs of Nemejcopteris feminaeformis, Cyathocarpus arborea (‘Pecopteris arborescens’ sensuCorsin 1951) and Eusphenopteris rotundiloba. There has been considerable confusion about the taxonomy of C. arborea and it has been reported in the literature in much earlier macrofloras. As pointed out by Corsin (1951), however, this was mainly based on small fragments from the terminal parts of pinnae (see also discussion by Zodrow 1990), and more typical examples with long, parallel-sided pinnae bearing small dentate pinnules do not occur below strata now assigned to Barruelian.

The transition between Crenulopteris acadica and C. lamuriana is completed at this level, with the complete disappearance of the former with its parallel-sided pinnatifid pinnules, and a significant increase in abundance of pinnae with smaller, more tapered and less lobed pinnules typical of the latter (Wittry et al. 2015). The base of the Crenulopteris lamuriana Zone has been best documented in NW Spain, notably in the Barruelo Coalfield (Wagner and Winkler Prins 1970, 1985a, b), and floras from the upper part of the zone reported from the Sabero Coalfield in Spain (e.g. Knight 1974). The classic Stephanian A floras of the Assise de Rive de Gier in the St. Étienne Coalfield of central France also belong to this zone (e.g. Doubinger and Vetter 1985a; Doubinger et al. 1995), as do those of the Zone de Lentin of the Carmaux Coalfield. In the Donets Basin of Ukraine, the base of the zone has been identified in the middle Torestskian ‘Regional Stage’, at about the level of limestone O₁ (Boyarina 2016, 2019, 2021). Elsewhere in Euramerica, however, strata of this age are either missing (e.g. in the Lorraine–Saar–Nahe Basin) or are in facies representing better drained substrates in which the macrofloras are dominated by mesophilous vegetation (e.g. Conemaugh Group of the Appalachians, Týnec Formation of Bohemia).

Wagner (1984) defined this zone so that its base approximated to the Barruelian–Saberian boundary, marked by the FADs of Cyathocarpus jongmansii, Pecopteris ameromii, Lobatopteris corsinii, Sphenopteris castelii, Sph. mathetii, Oligocarpia letophylla, Mixoneura subcrenulata and M. wagneri, and the LAD of Alethopteris barruelensis; Crenulopteris lamuriana also becomes very rare and only extends in to the basal part of the zone. The zone represents some of the best-documented, middle Stephanian floras of Europe, notably from the Sabero, Ciñera-Matallana and La Magdalena coalfields in Spain (e.g. Wagner 1963, 1964b; Knight 1974; Castro Martínez 2005a, b; see Wagner and Álvarez-Vázquez 2010a for a review), the Zone de Tronquié of the Carmaux Coalfield in southern France (Doubinger and Vetter 1969), most of the Ottweiler Group of Saarland in Germany (Cleal 2008a), the Slaný Formation in Bohemia (Šimůnek in Pešek 2004), the Carnic Alps (Opluštil et al. 2021b), and the upper Torestskian ‘Regional Stage’ in the Donets (Boyarina 2016, 2019, 2021). As with the Crenulopteris lamuriana Zone, the Alethopteris zeilleri Zone appears to be missing from the Appalachians in North America due to drier substrates supporting a more mesophilous vegetation.

The profound changes in terrestrial habitats that took place during the late Westphalian–early Stephanian (Moscovian–Kasimovian) affected the different palaeogeographic areas of the Variscides in different ways. The coal-swamp biome that covered large areas of the northern Variscan foreland during the early Westphalian, started to contract during the late Westphalian (Gastaldo et al. 1996; Cleal and Thomas 2005; Cleal et al. 2009). This coincided with changes in vegetation composition in the coal-swamp biome, most notably a decline in dominance of the arborescent lycopsids as the ecological framework taxa. This is most clearly seen in the record of anatomically preserved floras in coal-balls (e.g. Phillips et al. 1985) and a significant increase in abundance of monolete spores produced by marattialean ferns that largely replaced the arborescent lycopsids (e.g. Peppers 1996). Demonstrably during the earliest Stephanian, the coal-swamp biome continued to contract and then disappear over the northern foreland area of Euramerica.

However, the southern foreland displays a rather different dynamic. This may have been a coastal area of very considerable extent, as suggested by the extended interpretation of the Iberian orogenic belt and the now isolated successions in the Carnic Alps and Croatia, although much of the evidence of this extent has since been consumed by latest Variscan tectonics and the subsequent Alpine orogeny. In this southern foreland and many of the intramontane basins, the coal-swamp biome persisted, evidenced by the record of arborescent lycopsids and stigmarian palaeosols, throughout the Late Pennsylvanian; evidence of a Mesophytic flora and aridification with red-bed facies only becomes apparent in the early Permian (Wagner and Martínez García 1982). Similar longer-term persistence of the coal-swamp biome is evident in the the coastal deposition of the Donets Basin and Illinois Basin. Nevertheless, shortly after the early Permian the swamps had disappeared from the whole of Euramerica, including the Variscides, although the biome had migrated eastwards and now covered large areas of Cathaysia (Hilton and Cleal 2007).

The complex interaction of vegetation with climate, palaeogeography and orogenesis during the Late Pennsylvanian has inevitably produced a heterogeneous macrofloral and palynological biostratigraphic record. The plant biostratigraphy, as summarized by Opluštil et al. (2021a), that has underpinned the development of the West European regional chronostratigraphy for this time interval is clearly determinable where the wetland biome persisted. Although there have been some disagreements about the detailed correlation of some of the zones and therefore substages (e.g. Bouroz et al. 1970; Knight and Wagner 2014; Boyarina 2019), the broad pattern in the biostratigraphy is relatively consistent across palaeotropical Euramerica (Opluštil et al. 2016a, 2021a).

Where the coal-swamp biome was replaced by Mesophytic vegetation dominated by conifers, peltasperms, cycadophytes and other gymnospermous groups (sensuKerp 1996; Cleal and Cascales-Miñana 2014, 2021), these biozones obviously cannot be recognized. However, this does not invalidate the value of the biozones in the Upper Pennsylvanian. The geological record of this swamp biome is extensive in the Upper Pennsylvanian and has been intensively investigated, in part due to its economically important coal resources. Moreover, the regional stratigraphic resolution that can be achieved with these biozones is far superior than that obtained with the Mesophytic fossil record of this age. It is also important to remember that these substages are not defined by the biozones but by the stratigraphical levels in their relevant boundary stratotypes, which as will be shown later represent accurately determined time-planes. It is possible to correlate these time-planes and, therefore, the substage boundaries with reasonable accuracy with the North American sequences (e.g. Heckel 2008) independent of biostratigraphic indices.

Ash-fall tuffs (pyroclastic tonsteins) are widely occurring in the Pennsylvanian successions of the Cantabrian area. Robust radiometric dates have now been obtained from a number of these horizons using high precision U–Pb CA-ID-TIMS (Chemical Abrasion-Isotope Dilution-Thermal Ionization Mass Spectrometry) analysis. These dates have allowed construction of a preliminary time framework constraining the boundary stratotypes recognized in the Cantabrian area, presented here as Figure 6. This figure also includes, for the lower part of the Westphalian only, reference ages from the northern foreland area of Europe, (numbered horizons 1–3) that provide an age constraint to the boundaries between the Langsettian and Duckmantian (Pointon et al. 2012) and the Duckmantian and Bolsovian (Pointon et al. 2012; Waters and Condon 2012). For the Cantabrian area, individual tonstein bands for which radiometric dating results are available are numbered successively 4–11 on Figure 6.

Merino-Tomé et al. (2017) provide a comprehensive overview of ashfall tuffs in the Cantabrian area of northern Spain available for radiometric dating; dating was undertaken using the laser ablation process (LA-ICP-MS U–Pb dating). Discussion in Merino-Tomé et al. (2017) addresses the precision of LA-ICP-MS dating, which is lower than that using the CA-ID-TIMS method. A considerable discrepancy can be appreciated between ages for the same sample determined by LA-ICP-MS and CA-ID-TIMS (for example for horizon 5, comparative results for sample TC3 in Table 1). Schaltegger et al. (2015) provide an overview of the appropriate applications and shortcomings of available U–Pb radiometric methods, concluding that for timescale research, the highest precision and accuracy is provided by the CA-ID-TIMS method using EARTHTIME calibrated isotopic tracers. For the time framework developed in the present work (Fig. 6), only radiometric dates by the CA-ID-TIMS method with EARTHTIME calibrated isotopic tracers have been used (see Table 1). For the dated samples from the Cantabrian area, with one exception of a duplicate analysis performed on one sample from horizon 8 (see Table 1), all reported ages using CA-ID-TIMS were obtained by the Pacific Centre for Isotopic and Geochemical Research, a unit of the Department of Earth, Ocean and Atmospheric Sciences of the University of British Columbia, Vancouver, Canada (EOAS-UBC). Interpreted ages for all samples are based on weighted ²⁰⁶Pb/²³⁵U dates reported at the 2-sigma confidence level in the three-error format of Schoene et al. (2006) – ± X includes internal errors only, largely composed of analytical (counting statistics), mass fractionation and common lead composition uncertainties; ± (Y) error includes X plus isotopic tracer calibration uncertainty; ± [Z] includes, additionally to X + Y, uranium decay constant errors.

The time framework for the Asturian through to Saberian in the Cantabrian region, has been constructed around four tonstein horizons (horizons 5, 7, 8, 9 on Fig. 6), considered here to be robustly defined on the basis one or more samples with recent radiometric dates. Table 1 presents the dating results from individual samples; horizon 5 is supported by one CA-ID-TIMS date only although this analysis displays strongly concordant analytical parameters (see Fig. 7); horizons 7 and 8 are each supported by three independent dates from separate samples; horizon 9 is supported by two independent dates from geographically separate sample points. That a number of separate samples give closely coincident dates for any one horizon is here considered to provide a robust confirmation of the reproducibility of the results and reliability of the age assigned to the horizon.

The position of each dated horizon in a documented stratigraphic succession, and the geographical relationship of samples that contribute to the date, is of key importance for the interpretation of the age relationship with boundary stratotypes. Commentary on the location of samples that contribute to the dating of each horizon is provided in the following ‘Summary of stratotypes and reference sections’; concordia diagrams are presented for a sample of each of the four reference tonstein horizons. Table 1 provides reference to the published sources that first recognized the presence of the individual pyroclastic tonstein bands as key markers and the geographical location of the dated samples.

It falls beyond the scope of the present work to report more widely on the radiometric dates currently available or in process in the successions of the Cantabrian region. However, it is pertinent to underline that not all pyroclastic tonsteins prove suitable for reliable radiometric dating, often through complex zircon systematics due to the presence of older material incorporated in the horizon during the explosive event or due to pervasive lead-loss. This has proved to be the case for at least one horizon in the higher part of the Sabero succession and also for the only horizon thus far located for sampling in the Ciñera-Matallana Coalfield (horizon 10 on Table 1, Fig. 6). However, explosive volcanism, reflected in pyroclastic tonsteins, certainly continued through to the latest Pennsylvanian. A horizon in the Puerto Ventana Coalfield (horizon 11, Sample P168; see Table 1) has provided a reliable date of 300.24 ± [0.40] Ma, which agrees closely in age with the result from a third party laboratory that analysed the same band (O. Merino-Tomé pers. comm.).

The currently reported time framework (Fig. 6) has benefitted from collaboration with an on-going research programme at the University of Oviedo to generate revised dates using high precision CA-ID-TIMS dating, through which Dr O. Merino-Tomé kindly made available the tonstein sample (TC3, horizon 5) from the Agapita coal seam (Pozo Maria Luisa, Asturias).

The locations of the proposed and ratified stratotypes are presented on Figure 5, with their relationship to the depositional basins and successions of the Cantabrian Mountains. All the proposed and ratified stratotypes are freely accessible surface exposures and are in the process of geoheritage protection; all are registered as LIGs (lugares de interés geológico; sites of geological interest), which constitutes the first step towards formal legal protection. Additionally, in the case of the Barruelian stratotype, this has also been adopted as a site of special protection by the local municipal administration. Similar levels of local protection are anticipated for all the stratotype sections.

The time relationship of reference successions in the Cantabrian area is presented in Figure 6, in relation to an absolute timescale and the accepted chronostratigraphy of the global framework and macrofloral biozonation. Columns in this illustration represent continuity of sedimentation and do not directly reflect the thickness of sediments. The general character of the sedimentary environment is illustrated on the basis of five categories only: limestone-dominated successions, marine clastic successions, terrigenous successions with coals and palaeosols, sandstone-dominated successions and successions of coarse clastics and conglomerates.

The tonstein horizons that generate key radiometric dating results are highlighted on Figure 6 and numbered as in Table 1. In addition, the age relationship framework for the substages also draws on zonal correlation established by fusulinid foraminifera, allowing close comparison with the platform carbonates and predominantly marine succession of the Picos de Europa Province (Merino-Tomé et al. 2006, 2009) and through this, with the marine succession of the Russian Platform and the corresponding global stages.

This substage has achieved a wide level of acceptance in the context that it replaces the previous usage of Westphalian D; it is identified in the list of regional substages for Western Europe published by the SCCS (Heckel and Clayton 2006). However, the Asturian has only been outlined in a preliminary proposal (Wagner et al. 2002), and a formal boundary stratotype has not yet been identified, which would be necessary before submission of a formal proposal for ratification to the SCCS.

The SCCS working group to address the Westphalian C–D boundary (Laveine 1977) developed a number of palaeobotanical and palynological criteria, which together could be expected to define the base of this chronostratigraphic unit in comparable successions of western Europe. These criteria included the top of the range of Paripteris spp., the base of the range of Linopteris obliqua var. bunburii and the base of the epibole of the complex around Neuropteris ovata. Laveine's (1977) report also suggested that the Central Asturian Coalfield was potentially the most favourable location for selection of a Westphalian D stratotype in Western Europe. It was suggested that the Caleras Beds (Fig. 6; otherwise referred to as ‘paquete Caleras’, an informal unit based on mining practice) were clearly lower Westphalian D and that the contact Westphalian C–D would lie in the underlying Tendeyón Beds. Wagner and Álvarez-Vázquez (1991) undertook a wide-ranging review of the floral characterization of the Westphalian D Stage in NW Spain. The three floral biozones recognized by Cleal (1984a, 1997, 2007) for the Westphalian D of the Lorraine–Saar–Nahe Basin and South Wales were equally recognized in NW Spain, the two lower zones (obliqua and micromiltoni zones respectively) being identified in the Central Asturian Coalfield.

Within the coalfield, two structurally defined areas contribute to the the analysis of the later Westphalian succession: the Riosa-Olloniego subarea to the west and the Aller-Nalón subarea, which covers the main productive part of the coalfield (see Fig. 5). These two subareas are separated by the La Justa–Aramil thrust belt. Correlation between these subareas has been based on lithostratigraphical features and biostratigraphy using fossil flora and the evidence of fusulinid and brachiopod faunas (García-Loygorri 1974). This demonstrated that the succession of the Riosa-Olloniego subarea covered the Westphalian C–D boundary, which was likely to be situated within the Canales Formation (Fig. 6) as defined by Pello and Corrales (1971). The floral record for the Riosa–Olloniego sector (Jongmans and Wagner 1957; Wagner and Álvarez-Vázquez 1991, 2010a) comprises collections across a section covering the productive coal seams in a succession extending from above the Mieres Conglomerate Formation and through the underlying Canales Formation. Flora from above the Mieres Conglomerate is recorded primarily from surface localities and included Neuropteris ovata Hoffmann and Linopteris obliqua Bunbury throughout. In contrast, the flora of the Canales Formation was collected exclusively from underground mine sections; N. ovata was notably absent, although Pecopteris unita Brongniart, considered by Laveine (1977) to appear at the same level as N. ovata, makes its appearance at the top of the Canales Formation. Wagner and Álvarez-Vázquez (1991) concluded that the upper part of the Canales Formation corresponds to the transition Westphalian C–D.

The proposed boundary stratotype (Wagner et al. 2002) is located along a disused mineral railway track in the Riosa Valley of the Riosa–Olloniego subarea, within a succession extending across the Canales Formation and its contact with the succeeding Mieres Conglomerate Formation (Pello and Corrales 1971). The sedimentology of this interval has been studied by Fernández (1990); within a stack of deltaic sequences, alluvial sediments host relatively thin but worked coals, and interbedded wave-dominated shore-line conditions can also be identified. The fossil content is rich and varied; in addition to fossil plants, marine faunas comprise brachiopods, ostracods and bivalves, although systematic documentation is still incomplete. Cleaning and further detailed measurement in conjunction with conservation measures are required to establish if a satisfactory stratigraphic boundary can be defined, either at the upper contact of the Canales Formation or at a point within this formation, which is attributed a stratigraphic thickness ranging 700 to 1000 m (Fernández 1990).

Recognition of a Bolsovian–Asturian boundary stratotype in the Central Asturian Coalfield offers the advantage that it lies within an area with a long and continuous, well-documented succession of Westphalian strata from Langsettian through to late Asturian. In contrast to other areas in Europe, there is generally good natural and permanent exposure and access. Extensive description of the productive coal-bearing succession of the Central Asturian Coalfield was provided in the publications of the 10th Carboniferous Congress held in Madrid in 1983 (Leyva et al. 1985; Luque et al. 1985; Sáenz de Santa María et al. 1985). Palynological studies have been undertaken in numerous sections of the Central Asturian Coalfield (Chateauneuf 1973; Luque et al. 1985), providing a record closely comparable to those of corresponding successions in the north European paralic belt. The correlative succession of the Riosa-Olloniego section in the Aller-Nalón subarea includes goniatites (Anthracoceras cambriense Bisat) in the Generalas Beds and a locally rich fauna of corals and brachiopods (Feys et al. 1974).

The productive coal-bearing succession in the Aller–Nalón subarea also hosts a number of pyroclastic horizons (Feys et al. 1974; Merino-Tomé et al. 2017). A tonstein sample (TC3, collected in the mine Pozo Maria Luisa, Langreo) from the Agapita coal seam (horizon No. 5 on Fig. 6; Table 1), has been submitted for high precision U–Pb CA-ID-TIMS dating by the EOAS-UBC laboratory, Vancouver. This has yielded a date 0f 309.67 ± [0.37] Ma; the corresponding concordia diagram is presented as Figure 7. This horizon, also recorded as the Lozanita tonstein (Feys et al. 1974), is widely recognized across the Central Asturian Coalfield, near the top of the paquete Sotón (in the succession of informal lithostratigraphic units recognized in the mining industry, Feys et al. 1974). This places it clearly above the correlative horizon of the Mieres Conglomerate and therefore clearly within, although within the lower part of, the proposed reference succession for the Asturian Substage. There is an approximate stratigraphic thickness of some 700 m between the horizon of the Agapita–Lozanita tonstein and the correlative position of the top of the Canales Formation, considered essentially coincident with the base of the Asturian (ex-Westphalian D) substage (Wagner et al. 2002). The recently obtained radiometric date by the CA-ID-TIMS method for this tonstein is, as yet, the best guide to estimating an absolute age for the base of the Asturian substage; using a very approximate scaling factor, an indicative age of c. 310.7 Ma has been estimated (Fig. 6).

An informal proposal for recognition of the Cantabrian Stage (Wagner 1969) was accompanied by a report on a proposed boundary stratotype (Wagner et al. 1969) near the village of Tejerina, province of León (Fig. 5). The lower part of the reference section, comprising nearly 600 m of strata, included an extensive flora that corresponded to the upper part of Westphalian D. This section then traverses a relatively thin (45 m) marine intercalation of coastal (neritic) aspect, the Barranquito Marine Formation (Fig. 6), which is succeeded by a coal-bearing succession of over 300 m with a characteristic flora of Stephanian aspect, including Odontopteris cantabrica, Alethopteris bohemica and Pseudomariopteris cordato-ovata (Wagner et al. 1969). However, this Tejerina section, located on the NW margin of the North Palencia Basin, initially proved problematic to correlate with more complete successions across the wider area of the basin to the east.

The steps to formalize a boundary stratotype have been summarized by Knight and Álvarez-Vázquez (2021). Central to the difficulty of establishing firm correlation when the Tejerina section was first described, was that detailed mapping and investigation of the North Palencia Basin (Fig. 5: Valderrueda–Guardo–Cervera Coalfield; La Pernía, Castillería and Redondo coalfields) was in progress and continued through the 1970s. With the intention to establish formally the base of the Cantabrian Stage in a more continuous succession, with terrigenous strata with the transitional fossil flora interbedded with well-documented limestones, the SCCS Stephanian Working Group (Bouroz et al. 1972) proposed that this should be defined at the base of the Lores Limestone in the La Pernía area. Up to this point the boundary and stratotype had not been subject to a vote; the boundary at the base of the Lores Limestone was ratified in the 1971 SCCS meeting in Krefeld (George and Wagner 1972). There was presumed equivalence of this limestone with the Barranquito Marine Formation. The Lores Limestone yielded an important fauna of fusulinid foraminifera to van Ginkel (1965) who assigned this to his Fusulinella Zone (subzone B; subdivision B₃), correlated with the Myachkovsky Horizon of the upper Moscovian Stage.

Continuing investigation of the North Palencia Basin recognized a thick succession of up to 5500 m of alternating marine and terrigenous facies from which a very extensive fossil flora was collected in the terrigenous sections (Wagner and Fernández García 1983). Correlation was largely based on mapping and identification of the marine intervals. In summary, it was found that the Barranquito Marine Formation of Tejerina in northern León in fact correlated to a horizon substantially higher stratigraphically than the Lores Limestone (see Fig. 6).

Wagner (1983) developed the argument that the original intent in selection of the base of the Lores Limestone as the Cantabrian boundary stratotype was that it should be a marine interval equivalent in position to the Westphalian D–Cantabrian interface recognized at Tejerina. It was argued that the boundary stratotype should be adjusted in the light of new information. In the Guardo Coalfield fossil flora evidence suggested that there was a significant floral change reflecting the passage into Cantabrian strata, in effect the base of the Odontopteris cantabrica Zone, at the level of the terrigenous Choriza Formation. A formal proposal was developed (Wagner and Winkler Prins 1985a) that a revised boundary stratotype should be defined in the Guardo Coalfield, at the base of the marine Villanueva Formation immediately underlying the Choriza Formation; the location of the stratotype is illustrated on Figure 8. This proposal was accepted by ballot of SCCS members (Engel 1989). A correlative horizon of the Villanueva Formation contains fusulinid fauna indicative of Myachkovian age (the Urbaneja Limestone, van Ginkel 1971), suggesting that the base of the Cantabrian can be considered somewhat older than the base of the Kasimovian.

An overview by Wagner and Winkler Prins (1985b) of the sedimentological and palaeontological characterization of the Cantabrian succession in the North Palencia Basin, included the ranges of flora, algae, foraminifera, brachiopods, molluscs, ostracodes, corals and sponges. An important correlative horizon is the Brañosera Formation (indicated on the Barruelo column of Fig. 6), reflecting a major marine transgression across the extent of the basin and providing unequivocal correlation from the Guardo–Cervera Coalfield (Taranilla Marine Formation) eastwards to the La Pernía area and the Rubagón Valley area of the Barruelo Coalfield (Fig. 5). The Brañosera Formation, with a stratigraphic thickness of over 1200 m in the Rubagón Valley, is succeeded by the Peñacorba coal-bearing unit and at some 300 m stratigraphically above this the Carboneros Member (Fig. 6). The base of the Carboneros Member is taken as the base of the succeeding Barruelian substage (Wagner and Winkler Prins 1985a) and is correspondingly the top boundary of the Cantabrian substage. The fossil floral assemblage of the Carboneros Member corresponds very closely with that of the classic reference section for the Stephanian A of the Assise de Rive-de-Gier, St. Étienne Coalfield of the Massif Central, France (Wagner and Winkler Prins 1970, 1985b).

A number of samples of a pyroclastic tonstein band (horizon 7 on Table 1; Fig. 6) have been collected from small tips associated with the Peñacorba coal seams (samples P178, P179; see Fig. 9) and also from the tip of the Peñacorba Colliery at Barruelo de Santullán (sample P174, Table 1; Fig. 9). These three geographically separate samples have been submitted for radiometric dating (U–Pb CA-TIMS), and have yielded very closely coincident ages (Table 1), which, in conjunction with similar petrographical descriptions, confirms they are the same horizon, suggesting a high degree of reproducibility and confidence in the relative accuracy of the dates. The date of sample P174 at 305.02 ± [0.45] Ma (concordia diagram on Fig. 10) can be considered definitive for the age of the Peñacorba Member; an indicative age of 304.9 Ma has been assumed for the contact Cantabrian–Barruelian (Fig. 6; Table 2).

The boundary stratotype of the Barruelo substage, defined at the base of the Carboneros Member, is exposed in the cutting of a disused mining railway line near the hamlet of Helechar west of Barruelo de Santullán (Fig. 9); a detailed stratigraphic section across the boundary is provided in Wagner and Winkler Prins (1985b). The Carboneros Member is succeeded by a marine unit with some brachiopod, bivalve and gastropod fauna, and thereafter, the succession comprises the Calero Member, which hosts the main productive sequence of coal seams of the coalfield. However, still within the Calero Member there are thin marine bands, including a Lingula band and a calcareous lens between Seams III and IV with brachiopods and fusulinid foraminifera. For the latter, identifications have been reported in Wagner et al. (1977); these, including Fusulina and Pseudotriticites species, were considered insufficient for exact age dating but suggestive that this would be no higher than Kasimovian.

The top of the Calero Member represents the youngest strata recorded in the depositional history of the North Palencia (post-Leonian) Basin. Sedimentation is interpreted to have ceased with the initiation of uplift and locally intense deformation representing the Asturian Folding Phase, a relatively short-lived event within the Barruelian substage. The oldest sediments deposited after the Asturian Folding Phase occur in the lower part of the Sabero Coalfield, some 65 km to the west of Barruelo (Fig. 5); Knight (1971, 1974, 1983a, b, 1985) detailed the Sabero succession and its flora. The flora of the lower part of the Sabero succession was shown to be closely similar and effectively contemporaneous with the flora of the Calero Member and also identifiable with late Stephanian A as characterized in the coalfields of the Massif Central of France (Knight 1971). It was concluded that the area of the Cantabrian Mountains conserved a record of virtually continuous sedimentation through the Stephanian A (Bouroz et al. 1972). This conclusion was sustained in the proposal for a new Barruelian Stage (Wagner and Winkler Prins 1985a); it was proposed that the boundary stratotype would be defined in the Barruelo Coalfield, and that the upper part of the stage would be characterized by the succession in the lower part of the Sabero Coalfield. It was assumed that the upper contact of the Barruelian Stage (now substage) would later be defined in the Sabero succession.

Commencing in Sabero, the overall pattern of the post-Asturian basin succession is of basin-fill with sediment on-lapping up a basin margin to the west. The Sabero succession is characterized by a number of major inundation events, possibly linked to global eustasy, each of which initiates a coarsening upwards cycle (Knight et al. 2000), marked by characteristically fine lacustrine shale units with Leaia bands at the base and culminating with alluvial sediments and coal formation. These cycles have allowed robust correlation with the Ciñera–Matallana Coalfield further west (Wagner and Castro 2011; Knight and Wagner 2014). The first cycle, near the base of the Sabero succession, is marked by a marginal marine interval that thickens eastwards to fully marine sediments with Lingula (Iwaniw 1984). Bivalve faunas from horizons throughout the lower part of the Sabero and Ciñera-Matallana coalfields and the upper part of the La Magdalena Coalfield, have been studied by Eagar (Eagar and Weir 1971; Eagar 1985), with the conclusion that in many cases they present evidence of near-marine and brackish environments. The near-coastal, paralic, sedimentary environment is closely analogous to that of the succession of early Barruelian age in the Barruelo Coalfield and, overall, the composition of the flora in terms of the representation of the principal taxonomic groups is closely similar (see record in Wagner and Álvarez-Vázquez 2010b).

Interpretation of the overall floral composition from the lower part of the Sabero succession concluded that it should be considered marginally younger than that of the Calero Beds, but that it can be compared closely to the flora of the Zone de Lentin, in the Carmaux Coalfield of the Massif Central (Knight 1974, 1983a). The corresponding conclusion was that the top of the Stephanian A (now Barruelian) should be recognized at the base of the Quemadas Formation (Fig. 6), which also marks the base of one of the widespread inundation cycles. As discussed above, the Crenulopteris lamuriana megafloral zone has been considered to coincide with the Barruelian Stage (now substage) in its type area (Wagner 1984), and the upper boundary conceptually represents passage to the succeeding Alethopteris zeilleri zone. The well-exposed surface section, near the village of Saelices, which provides the type section for the Sucesiva Formation and the base of the Quemadas Formation, has been designated the boundary stratotype for the proposed Saberian substage, which succeeds the Barruelian (Knight and Wagner 2014).

The Sabero succession hosts at least ten pyroclastic tonsteins (ash-fall tuffs), of which five have been identified in the lower part of the Sabero succession now attributed to the Barruelian substage (Knight et al. 2000). The highest recognized tonstein horizon in the Barruelian succession (in the Sucesiva Formation) has been submitted for high precision U–Pb (CA-ID-TIMS) radiometric dating. This horizon (horizon No. 8 on Fig. 6; Level 21 of Knight et al. 2000), collected on the Saberian boundary stratotype section, occurs at approximately 100 m stratigraphically below the base of the Quemadas Formation (Knight and Wagner 2014). Two separate samples (P08, P32) from different positions on adjacent outcrops have been dated by the EOAS-UBC laboratory and give very closely similar dates; a duplicate sample of P08 has been submitted to a third-party laboratory and gave a marginally older date (see Table 1; Merino-Tomé pers. comm.). The closely coincident dates are considered to indicate a high degree of reproducibility and confidence in the relative accuracy of the dates. The radiometric date obtained from Sample P32 is 303.51 ± [0.39] Ma (concordia diagram illustrated on Fig. 11); this has been used as a key reference to estimate an indicative date for the Barruelian–Saberian boundary.

The concept that the top of the then proposed Barruelian Stage would be defined in the Sabero Coalfield (in the post-Asturian succession) was expressed explicitly by Wagner and Winkler Prins (1985a); implicit was the recognition that the top of the Barruelian would need to be defined by the boundary stratotype of a new succeeding chronostratigraphic unit, still to be defined. Wagner and Álvarez-Vázquez (2010a) introduced the term ‘Saberian’ and informally proposed that this should be recognized as the chronostratigraphic unit above the Barruelian and should coincide with the Alethopteris zeilleri macrofloral zone. A formal proposal for the recognition of the Saberian substage in the west European chronostratigraphic scheme (Knight and Wagner 2014) has not yet been submitted for ratification through the SCCS.

In the proposed stratotype section the Quemadas Formation commences with a thick lacustrine shale succession marking a widespread inundation event. It precedes a coarsening-upwards basin fill that culminates in alluvial sediments with thick workable coal seams. The inundation event of the Quemadas Formation has been correlated to a similar coarsening-upwards lithological cycle commencing with the Tabliza Horizon at the base of the Pastora Formation in the Ciñera–Matallana Coalfield (Fig. 5; see relevant column on Fig. 6), some 20 km to the west (Knight and Wagner 2014). The upper part of the Sabero succession includes at least two further inundation–lacustrine–basin-fill cycles (Knight et al. 2000) that can be correlated to the Ciñera–Matallana succession. A single sedimentary basin appears proven, with consistent westwards expansion and onlap (Wagner and Castro 2011), although the relationships with other tectonically isolated areas of terrigenous coal-bearing successions (Fig. 5: La Magdalena, El Bierzo, Carrasconte, Puerto Ventana, Canseco-Rucayo, Viego-Salamón) are less clear without specifically correlatable lithologies. All these areas include successions that have yielded floras that can be identified with the Alethopteris zeilleri macrofloral zone. In these coalfields there is abundant evidence that the sedimentary environment, although dominated by lacustrine and alluvial plain sediments, was in a near-coastal environment with marine influence evident from bivalve assemblages (Eagar and Weir 1971; Eagar 1985), reinforced by the find in the Ciñera–Matallana Coalfield of a pygidium of the trilobite Cinerana matallensis Gandl (Gandl 2021).

Thus far, palynological records in the post-Asturian succession are almost entirely lacking, in part reflecting the relatively high rank of the coals. However, in the boundary stratotype for the Saberian, the Quemadas Formation has been the subject of detailed palynological sampling by research personnel of the University of Vigo. Although this remains a work-in-progress awaiting formal publication, Juncal et al. (2019) report that the palynoflora of the base of the Quemadas Formation corresponds to the Angulisporites splendidus-Latensina trileta (ST) Zone of Clayton et al. (1977), and correlates entirely to a position within the Gzhelian. Specifically noted is the abundance of Lycospora pusilla, Densosporites spp. Laevigatosporites spp. Thymospora sp., Florinites spp., and also the constant presence of Disaccites non striatiti. Additionally, the presence of acritarchs was recorded (M. Juncal, pers. comm.), strongly suggestive of marine influence.

A characteristic feature of the early Saberian interval is the presence of a number of pyroclastic tonsteins, reported respectively from the Ciñera–Matallana (Bieg and Burger 1992) and Sabero (Knight et al. 2000) coalfields. For definition of the Barruelian–Saberian boundary, the dating of the lowest tonstein horizon in the Saberian succession provides a constraining date above this boundary; this tonstein has been identified in the lower part of the Herrera Beds, in the Sabero Coalfield (horizon No. 9 on Fig. 6; Level 31 in Knight et al. 2000). This tonstein horizon is known at numerous points in the coalfield, occurring within the floor strata of one of the most widely worked coal seams – Capa 4N; sample P93 was collected from a surface exposure associated with a mining collapse (date reported first in Knight and Wagner 2014), and sample P76 was collected from workings in the 6th level of Pozo Sotillos (see Table 1). The two samples provided near identical dates (Table 1); the concordia diagram for Sample P76, dated 302.14 ± [0.37] Ma, is illustrated as Figure 12. This horizon is located at some 375 m above the base of the Quemadas Formation (Knight and Wagner 2014). Taking account of the highest dated tonstein in the Barruelian, in the same sedimentary sequence (horizon 8 on Fig. 6; samples P08, P32), a scaled age of 303.5 Ma is suggested herein as the indicative age for the Barruelian–Saberian boundary.

The age relationships of key reference successions in the Cantabrian region are presented on Figure 6. The time horizons for the chronostratigraphic units defined in this region are, insofar as possible, constrained by radiometric dating of pyroclastic tonsteins. The stratigraphic columns for each of the sections illustrated for this region are correlated also on lithological and sedimentological characteristics and where possible on the basis of fossil content, and specifically fossil flora and fusulinoidean foraminifera. This allows that a preliminary chronostratigraphic framework can be essayed for the upper Westphalian–mid-Stephanian interval, embracing the ratified boundary stratotypes of the Cantabrian and Barruelian substages and the proposed stratotypes of the substages immediately above and below these. It is recognized that this is a preliminary step in the process to construct a definitive composite chronostratigraphic framework, which would integrate a range of biostratigraphical successions and also reflect calibration of rates of sedimentation to support a scaled projection of age across the intervals from the dated horizons to designated substage boundaries. Nevertheless, the position of dated tonsteins close to the top boundary of the Cantabrian and Barruelian substages, respectively, provides well-constrained ages for these horizons.

Table 2 presents the indicative ages estimated from the present study in the Cantabrian region for the boundary of each of the ratified and proposed substages of the west European framework. This table also summarizes the degree of age constraint around each boundary and the scale of projection between dated horizons and adjacent substage boundaries.

Preliminary proposals by Bouroz et al. (1978) for a global Carboniferous chronostratigraphic scheme reflected the principles set out in the International Stratigraphic Guide (Salvador 1994; Remane et al. 1996). The object was to identify a marine sequence with a complete succession of faunas that would provide the primary signal for recognizing the successive chronostratigraphic boundaries. The Russian Platform and adjacent areas of the Ural Mountains (Wagner et al. 1979; Einor 1996) were identified as providing the best available model for the Pennsylvanian Subsystem, and the former area was formally adopted as the global reference (Heckel 2004). This is still a work-in-progress (Lucas et al. 2022b) with significant unresolved discrepancies remaining in the stage stratotypes (Heckel 2004, fig. 1). Nevertheless, by integrating numerous observed palaeontological events (mainly conodont-based) with radiometric ages, a composite standard succession of Pennsylvanian ages/stages has been constructed (Davydov et al. 2012; Aretz et al. 2020; Cohen et al. 2022): Bashkirian (Lower Pennsylvanian), Moscovian (Middle Pennsylvanian), Kasimovian and Gzhelian (together Upper Pennsylvanian).

As pointed out by Wagner (2017), however, the reliance on conodonts as the main indices for this global scheme hinders its global application, as these fossils are rare or absent from many areas. Although there are marine deposits throughout the lower Stephanian of the Cantabrian region of northern Spain, conodont evidence is equivocal for correlation (Méndez 2006). Fusulinoidean biostratigraphy and transgressive-regressive cycles in the Picos de Europa province (Merino-Tomé et al. 2006) have the potential for correlation with the cyclic sequences in the Donets and Russian sequences (e.g. Davydov et al. 2012). However, these correlations are hindered by the presence of numerous stratigraphical discontinuities in the Spanish sections. In the North Palencia Basin, brachiopods, foraminifera and conodonts have been used to correlate the Brañosera Formation in the middle Cantabrian with a level close to the Moscovian–Kasimovian boundary, but regional tectonism in this area makes it difficult to correlate the observed transgressive–regressive cycles with those observed in the Ukrainian and Russian sequences (Wagner et al. 1977).

It is therefore necessary to use ‘secondary’ criteria (sensuRemane 2003) to correlate the lower Stephanian substages with the global stages as defined in Russia. Palynofloras are of limited value due to the paucity of records in the Cantabrian region, but macrofloras have considerable potential for correlating with the mixed marine–non-marine sequences in the Donets Basin. The macrofloral biozones identified by Boyarina (2016, 2019; see also Shchegolev and Kozitskaya 1975) in the Kasimovian of the Donets seem to follow a broadly similar pattern to those seen in western and central Europe (Opluštil et al. 2021a).

Relevant correlations, primarily using megafloral biozones, with the Donets succession, and by extension to the global chronostratigraphic construct, are summarized for the Cantabrian through to Saberian substages.

Boyarina (2016, 2019) placed the base of the Odontopteris cantabrica Zone at about coal n₁, below limestone N₂ in the middle Isaevsky Donets Regional Stage (upper Moscovian). Davydov et al. (2010) published a date at this horizon of 307.26 ± 0.11 Ma. This is close to the c. 307.5 Ma suggested here for the base of the Cantabrian.

Boyarina (2016, 2019) placed the base of the ‘Lobatopteris lamuriana’ Zone at about limestone O₁ in the middle Toretsky Donets Regional Stage (middle Kasimovian). Davydov et al. (2010) dated this as c. 305.5 Ma, which is slightly older than the 304.9 Ma estimated herein for the base of the Barruelian.

Assuming the base of the Saberian, and therefore of the Alethopteris zeilleri Zone, is at c. 303.5 Ma, in the Donets this would correspond with about limestone O₅ at the base of the Kalinovsky Regional Stage, in the upper Kasimovian (Davydov et al. 2010). Shchegolev (1988) and Boyarina (2016, 2019) described a significant floristic change at this level, corresponding to a change to a more wetland vegetation with more lycopsids and sphenopsids (Boyarina 2021). Boyarina (2016, 2019) suggested that the base of the Alethopteris zeilleri Zone and therefore of the Saberian was rather lower, at the middle Kasimovian limestone O₄³, but her range charts are more compatible with this biohorizon occurring at a higher level in the Kasimovian as indicated by the radiometric dates.

In summary, the published and new radiometric dates reported here support a time framework (Fig. 6) that places the base of the Cantabrian substage at c. 307.5 Ma and the top boundary of the Barruelian substage at c. 303.5 Ma, which closely corresponds to the time interval of the Kasimovian Stage, for which the reference age range is 307.0–303.7 Ma (Davydov et al. 2012). The top boundary of the Barruelian corresponds closely to the time horizon of the base of the Gzhelian (303.7 Ma in Davydov et al. 2012).

The present analysis of the history and rationale of the Cantabrian, Barruelian and Saberian substages of the Stephanian has confirmed the following.

1. The Upper Pennsylvanian of the Cantabrian region of northern Spain occupied the southern foreland of the Variscan Mountains. It was deposited in a paralic environment with mixed marine and terrigenous sediments, offering a range of biostratigraphically significant fossil groups, including macrofloras, foraminifera, and some conodonts. Although subjected to moderate tectonic disturbance, the succession is essentially continuous. The vegetation dynamics of the area were rather different from that of the other parts of Euramerica, especially a much later persistence of the coal swamp biome.

2. Based on this Spanish succession, two regional substages have been formally recognized: the Cantabrian and Barruelian. The Barruelian was proposed to replace the concept of the Stephanian A as defined by the Assise de Rive-de-Gier of the St Étienne Coalfield, but which proved to be too incomplete to act as a satisfactory stratotype. The Cantabrian was established to fill the time-gap between the traditional concepts of Westphalian D (Asturian) and Stephanian A (Barruelian). The Saberian Substage has been proposed to fill, in part, the time-gap between the Barruelian and the classic Stephanian B of the St Étienne Coalfield.

3. The proposal to recognize the Cantabrian and Barruelian was considered by the SCCS over more than 20 years (1965–89). There were several field excursions by subcommission members, numerous publications documenting the stratigraphy and fossil-content of the sections, and an open postal ballot that produced overwhelming support for ratification. These are the only ratified substages of the Stephanian; they are also the most completely documented chronostratigraphic units for the later Pennsylvanian in the central and west European framework. The Cantabrian and Barruelian substages have been officially ratified by the IUGS Commission on Stratigraphy.

4. Following normally accepted international procedures, the Cantabrian and Barruelian are formally defined by boundary stratotypes: respectively at the base of the Villanueva Formation in the Velilla de Tarilonte Section, Guardo Coalfield (province of Palencia) for the Cantabrian, and at the base of the Carboneros Member in the Helechar Section, west of Barruelo de Santullán (province of Palencia) for the Barruelian. The boundary stratotype for the base of the proposed Saberian Substage, which would mark the top of the Barruelian, is at the base of the Quemadas Formation in the Arroyo de Saelices Section in the Sabero Coalfield (province of León). Detailed correlation between the sections in the Guardo–Cervera, Tejerina, La Pernía–Castillería, Barruelo and Sabero coalfields allows compound unit stratotypes to be constructed for the Cantabrian and Barruelian substages.

5. High precision radiometric dates (U–Pb CA-ID-TIMS) are reported from four pyroclastic tonstein horizons, which provide close constraints on the Cantabrian–Barruelian time interval. The base of the Cantabrian substage (c. 307.5 Ma) corresponds with uppermost Myachkovian (i.e. marginally older than the base of the Kasimovian – 307 Ma); the base of the Barruelian substage (c. 304.9 Ma) is late Kasimovian; and the base of the Saberian (c. 303.5 Ma) is essentially coincident with the base of the Gzhelian (303.4 Ma). The Cantabrian and Barruelian together, therefore, encompass the entire Kasimovian Stage.

6. The bases of the Cantabrian, Barruelian and Saberian in their respective stratotypes are indicated (but not defined) by the bases of the Odontopteris cantabrica, Crenulopteris lamuriana and Alethopteris zeilleri macrofloral zones. These are assemblage zones, rather than taxon range or interval zones, and the criteria used for placing the zonal boundaries are clarified. Comparison of these boundaries with radiometric data from other basins indicates that they represent relatively isochronous levels over palaeotropical Euramerica.

7. Using a combination of radiometric dates and biostratigraphy, the Cantabrian, Barruelian and Saberian offer the prospect of correlation with most of the terrestrial Upper Pennsylvanian sequences of Euramerica. Effective biostratigraphic correlation has been established with the Donets succession, the various Variscan intramontane basins, and other parts of the southern foreland in southeastern Europe (the Carnic Alps and Croatia). This has confirmed that there is no other comparable, essentially complete and fossiliferous succession in Europe, which could provide alternative stratotypes to cover the time interval of these substages.

8. In other parts of Euramerica, notably the northern Variscan foreland of Britain and the Appalachians, and the mid-continent area of North America, a combination of climatic, orographic and palaeogeographical changes meant that the coal swamp biome was replaced by Mesophytic (conifer, peltasperm, cycadophyte dominated) vegetation during the Late Pennsylvanian. Since the Cantabrian, Barruelian and Saberian are recognized biostratigraphically on criteria derived from the coal swamp biome, these substages are not readily recognizable in these other areas. However, this does not invalidate these Stephanian substages, which can still be recognized in non-coal-swamp areas by integrating radiometric and sedimentological evidence. They represent a specific interval of time and provide a much better stratigraphic resolution than any yet proposed framework using more mesic biotas.

All the three authors were at various times postgraduate students of Robert (Bob) H. Wagner (Sheffield, later Córdoba), and are pleased to acknowledge the shared experience of his vision of stratigraphic palaeobotany and his profound knowledge of the Pennsylvanian geology of the Iberian Peninsula. The co-ordination of information on the results of radiometric dating has been fundamental to the present work and the authors acknowledge very helpful meetings and discussion with Óscar Merino-Tomé of the Department of Geology, University of Oviedo, who, with other departmental colleagues also provided helpful guidance on fusulinid biostratigraphy. O. Merino-Tomé is thanked for making available the sample of the Agapita tonstein, from which radiometric dating is reported in this work. Luis Sardina is thanked for his assistance in fieldwork and in particular for help in locating the tonstein associated with the Peñacorba coals. Special thanks are due to Juan I. Peláez who has been a constant colleague in the fieldwork for reporting on the stratotypes in the Cantabrian region and also has undertaken preparation of some figures. Manuel Juncal is thanked for providing a preliminary overview of palynological investigation of the Saberian boundary stratotype undertaken at the University of Vigo.

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.

JK: conceptualization (lead), data curation (equal), formal analysis (lead), funding acquisition (equal), investigation (equal), methodology (equal), project administration (equal), resources (equal), software (equal), supervision (equal), validation (equal), visualization (equal), writing – original draft (equal), writing – review & editing (equal); CJC: investigation (equal), methodology (equal), writing – review & editing (equal); CÁ-V: conceptualization (supporting), data curation (equal), formal analysis (equal), investigation (equal), writing – original draft (supporting), writing – review & editing (equal).

The research reported in this work contributes to Project MCIU-19-PGC2018-099698-B-100 supported by the Spanish Government. However, the authors of this paper have received no direct financial support; fieldwork, radiometric dating and the preparation of illustrations have been self-funded.

The analytical procedures for radiometric dating have been documented by the Pacific Centre for Isotopic and Geochemical Research, EOAS Department, University of British Columbia, Vancouver, Canada and have been made available to the authors for separate publication, in preparation, specific to reporting of radiometric dates in the Pennsylvanian successions of Spain. In all other areas of the present work, data-sharing is not applicable.