Questioning carbonate facies model definition with reference to the Lower Cretaceous Urgonian platform (SE France Basin)

“In the general field of stratigraphy and sedimentology, a facies model could be defined as a general summary of a particular depositional system involving many studies in both ancient rocks and recent sediments” (Walker, 1967, 1992) that is set up to compare, frame and interpret sedimentary systems. Facies models are used as a priori knowledge and conceptual tools to correlate outcrops and wells with the ultimate goal of predicting the complete sedimentary system. Facies models and correlations are based on stacking patterns, which are founded on the Walther’s law and sequence stratigraphy. In this context, it becomes logical to use depositional profiles as a chronostratigraphic means of spatial correlation of facies successions. Geologists working on inter-well stratigraphic correlations are constantly dealing with these questions (e.g.van Buchem et al., 2002; Borgomano et al., 2008; Pomar et al., 2015). Carbonate facies models are, by definition, conceptual and rarely based upon laterally continuous observations of ancient sedimentary records. Facies model definition from the coastline to the basin is based on palaeoenvironmental interpretations of incomplete dataset that are not physically correlated. Only a few examples of seismic-scale continuous outcrops provide in-depth understanding of facies spatial organization, belts, transitions and stratigraphic architecture over large distances exceeding the tens-of-kilometers (e.g. Cretaceous: Istria in Croatia, Maiella in Italy, Vercors in France and Maestrat in Spain: Tišljar et al., 1998; Stössel, 1999; Richet, 2011; Richet et al., 2011; Gili et al., 2016). Such exceptional outcrops could be used as a direct analogue of comparable size subsurface systems, but detailed, continuous facies observations over such distances are not frequent practice. Apart from these few exceptional case studies, outcrops are non-continuous and/or of a limited extent in comparison to tens-of-kilometer-wide Cretaceous platform systems (e.g.Everts et al., 1995; Bover-Arnal et al., 2009; Léonide et al., 2012). As a result, data integration of these punctual, spatially limited outcrops into larger-scale analogues must be undertaken with care. Every facies model results from interpretations, interpolations and extrapolations based on a priori knowledge and hypotheses that require genuine scientific validation loops (e.g.Borgomano et al., 2020). These facts imply that, in a strict sense, a data-based definition of a facies model is not directly feasible.

In the case of Early Cretaceous carbonate platforms, determination and correlation of synchronous strata relies on interpretations from oil field observations (e.g.Yose et al., 2006), relative and absolute dating (e.g.Clavel et al., 2014; Godet et al., 2016; Frau et al., 2018; Frau et al., 2020; Brigaud et al., 2020; Frau et al., 2021), and/or stacking patterns (e.g.Godet et al., 2010; Ferry and Grosheny, 2019). All these elements are integrated within a conceptual sedimentological, palaeoecological and diagenetic model, which is supported by fundamental and deterministic sedimentological and paleontological concepts (Wilson, 1975; Tucker and Wright, 1990; Schlager, 2005). Facies models are therefore based on a large array of data and techniques that all imply uncertainties and a large interpretative part starting with facies definition (cf.Lokier et al., 2016). Consequently, different facies models can be defined depending on interpreters from a given set of data. In turn, different facies models can lead to contrasting sequence stratigraphic interpretations and stratigraphic correlations (e.g.Borgomano et al., 2008; Lallier et al., 2016; Edwards et al., 2018).

There is hardly any doubt that workflows of facies model definition must explicitly show supporting background data, concepts, hypotheses and uncertainties in order to be shared, used and be able to evolve. While these tasks might appear relatively straightforward in theory, providing explicit geological background information and uncertainties remain challenging. The present study aims at analyzing the evolution of a facies model, documenting supporting data and concepts, and exploring a methodological workflow to establish a well-supported facies model. The well-studied Urgonian carbonate platform of SE France and Switzerland are used to tackle these objectives. This analytical method could have implications for sedimentological and stratigraphic studies of Middle East Cretaceous carbonates, as the Urgonian platform is retained as a valid analog (Borgomano et al., 2008, 2013; Massonnat et al., 2017).

For more than a century, and still ongoing today, numerous works have collected facies and palaeontological data leading to diverse conceptual models of the peri-Vocontian Urgonian platform (from d’Orbigny, 1850, to Frau et al., 2021). This great deal of effort has led to improving the overall knowledge and understanding of Lower Cretaceous carbonate systems. From an exhaustive literature review, we synthetise and classify the available data that support the interpretation of the published facies models (cf.Sect. 4). The data integration aims at unifying the diverse published facies tables and theoretical depositional profiles; and ultimately at defining a harmonised facies model, which can be used by everyone independently from research focus and approach. The facies model definition follows the approach of “infrastructure of carbonate platform” defined by Handford and Loucks (1993), which has the advantage of relating data and concepts to the controlling parameters of carbonate sedimentation at specific spatial scales. The synthetic facies table and model thus show a consistent continuum across all space and time scales, and propose quantifications that allow straightforward evaluation, criticism, validation and improvements (cf.Sect. 5). The facies model design will enable a direct comparison with other peri-Tethyan carbonate platforms and stratigraphic correlations. The model also provides data input for numerical models to challenge stratigraphic correlation hypotheses (both process- and geostatic-based modelling).

During the Barremian–Aptian period, carbonate platforms record a large development with an estimated extension of 4.63 or 10–22 · 10⁶ km² along the Tethyan margins (Philip, 2003; Pohl et al., 2020, respectively). Most carbonate platforms are characterised by the presence of rudists and benthic foraminifera associated with tropical palaeoenvironments (Douvillé, 1900; Rat, 1983; Gili and Götz, 2018). Rudists are considered to thrive in the neritic domain, under a tropical palaeoclimate and within shallow waters (Gili et al., 1995). The relatively symmetrical palaeolatitudinal distribution of biota between 35° N and 35° S is controlled by palaeoclimatic and palaeoceanographic parameters (Kiessling et al., 2003; Steuber et al., 2005; Pohl et al., 2019). In the Lower Cretaceous, rudist assemblages of the Tethyan domain exhibit contrasting palaeobiogeographies with two main provinces, namely (1) the northern Tethyan margin (European Province) and (2) the southern Tethyan margin (Arab-African Province: Masse, 1992; Masse and Fenerci-Masse, 2008; Skelton and Gili, 2012). Each province hosts specific genera and species that differ in time and space (Hughes, 1997; Skelton, 2003; Skelton and Gili, 2012; Masse and Fenerci-Masse, 2008, 2015; Masse et al., 2020, 2022). Rudist families, however, are the same along both margins and include Caprinidae, Requieniidae, Monopleuridae and Radiolitidae. In addition to paleontological attributes, both provinces exhibit carbonate facies analogies in terms of sedimentology and rock fabrics (Borgomano et al., 2013).

In the present study, we focus on the Urgonian type platform of SE France and Switzerland. The term Urgonian herein refers to the depositional facies terminology defined by Masse (1966, 1976; also cf.Rat, 1983; Clavel et al., 2014) with “Urgonian sensustricto” designating the rudist facies that are embedded vertically and laterally within the “Urgonian sensulato” corresponding to the bioclastic limestones. The studied chronostratigraphic interval corresponds to the Barremian and Lower Aptian, according to the recent age reappraisal of Frau et al. (2018, 2021). The facies model results from the bibliographic synthesis of outcrop data stemming from the present-day Jura Chain and Helvetic Mountains to the North, the Mont d’Ardèche to the West, the Monts de Vaucluse and Calanques to the South, and the Verdon and the Alps to the East (Fig. 1). This area hosts the key localities of the Urgonian Jura–Subalpine platform to the North, the Bas-Vivarais platform to the west and the Provence platform to the South. Together, these platforms define the peri-Vocontian domain. Sediments deposited in the Vocontian Basin sensu stricto are not considered here. The basinal facies, which are disconnected from the platform, correspond to “pre-Vocontian” settings (sensuFrau et al., 2018; Tendil et al., 2018).

In general, a facies model is built by geologists to organize and extrapolate sedimentary data in space and time (Wilson, 1975; Walker, 1992). The model helps the interpretation of limited and incomplete sedimentary records and estimates the controlling parameters of sedimentation, allowing for depositional environment analyses and prediction. Based on the works of Walker (1992) and Argenio et al. (1997), a facies is herein defined as a volume of rock distinguished upon physical characteristics such as color, texture, components and sedimentary structures. A facies is a product of a specific depositional environment and may have an early diagenetic overprint, which is not the object of the present study. However, the palaeoenvironmental interpretation of facies is not necessarily unequivocal. In order to establish its depositional framework, a facies is generally framed into a facies association (sensuWalker, 1992; Argenio et al., 1997). A facies association corresponds to a genetically-related group of facies deposited together within a bedding interval or a depositional sequence (Collinson, 1969; Reading, 1996).

Here, the construction of the conceptual facies model is based on an exhaustive compilation of bibliographic data about Lower Cretaceous, Barremian–Aptian outcrops from SE France and Switzerland. This compilation first extracts supporting data of facies models only from studies that published a facies model representation and then labels each of them according to geographic area, age, model classification and scientific school. The bibliographic database allows synthesising the evolution of scientific approaches used by different academic teams (e.g. Grenoble, Lausanne, Lyon, Marseille and Neuchâtel) and resulting facies models of the Urgonian platform from Jura, Vercors and Provence (Figs. 2 and 4).

To build a constrained and consistent conceptual facies model, we propose a formal reproducible workflow that is based on the following four steps: (1) an exhaustive bibliographic review to extract the published data (e.g. sedimentary, chronostratigraphic and palaeoecological data) and concepts (e.g. facies belt organization, morphology of platform and carbonate factory) that support the facies model; (2) the standardisation of the various depositional facies data in a systematic and consistent way; (3) the definition of facies associations and controlling parameters for each facies association (palaeobathymetry and hydraulic energy); (4) the creation of a facies association table and a conceptual facies model. Iteration between the different steps should guarantee a great consistency between data and model. The table and model of depositional facies are based on the approach described by Handford and Loucks (1993) that realizes an infrastructure of carbonate platform elements from overall geometry or grain association of deposits to unitary facies. This approach aims at bridging the different scales of study of the carbonate system into a consistent and reproducible way. Supporting data of the facies model are thoroughly reported for each spatial scale and include outcrop locations within the interpreted palaeogeographic map, facies transition observations, temporal and spatial resolution of interpreted synchronous strata and concepts.

The historical evolution of Urgonian facies models is intrinsically linked to major outcrop surveys in different regions (i.e. Jura, Vercors and Provence) executed by different academic teams (Grenoble, Lausanne, Marseille and Neuchâtel: Fig. 2; see Rat, 1983; Clavel et al., 2014). These works match the evolution of concepts that influenced the construction of the different facies models (Fig. 4).

The peri-Vocontian Urgonian limestones went through several phases in the definition of sedimentary profiles (cf.Masse, 1992; Fig. 4): (a) pioneer works dealt with palaeogeographic occurrences of rudist and orbitolinid fossils before considering any geological model (Figs. 2 and 4); (b) these biota occurrences were then related to bio-sedimentary assemblages using actualism and following the definition of Pérès and Picard (1964, and references therein; Fig. 4); (c) in continuation, these bio-sedimentary associations were considered within a broader palaeoenvironmental context (e.g. palaeobiogeographic approach); (d) at this step only, sedimentary facies were described and associated to depositional palaeoenvironments in a spatial representation (Arnaud-Vanneau and Arnaud, 1976; Arnaud-Vanneau, 1979, Figs. 2a–2c).

The original Urgonian facies models were based on the Vercors platform. The interpreted sedimentary and palaeotopographic profiles were produced directly from the palaeontological and sedimentological correlations of putative contemporaneous depositional facies maps (Tab. 1).

• a Lower Barremian facies model describes an isolated platform high that is surrounded by two asymmetric slopes on both sides of the sedimentary profile (“shoal” or “haut-fond” model; Fig. 2a and Tab. 1). Ooid shoals with a mixture of bioclastic facies represent the shallowest facies association on the platform top. Coarser-grained facies including few coral facies occur on the interpreted windward side of the isolated platform high (i.e. Vocontian Basin, oceanic side), while finer and muddier facies are located on the interpreted leeward side (i.e. Jura, continental side). Bioclastic and very fine-grained bioclastic facies are shown on both interpreted steep and gentle slopes that reach basinal clay-rich deposits;

• an Upper Barremian facies model defines an attached platform (“platform” model; Fig. 2b and Tab. 1). Contemporaneous depositional facies include from the shallower to the deeper uncommon supratidal charophyte-rich marls, rudist and oncoidal facies, lagoonal muddy and sandy facies, coral facies at the margin, bioclastic facies on the slope and deeper micritic facies;

• Orbitolinid-rich platform (Lower Orbitolina beds, Ai1 sequence) and depressions (Upper Orbitolina beds; top Ai2 sequence) represent specific, non-contemporaneous depositional environments that relate with periods of terrigenous influx. Neither the orbitolinid facies nor the marly and orbitolinid-rich platform are shown in the facies models.

The integration of the two above mentioned palaeogeographic map-based models and the terrigenous-related “Upper Orbitolina bed” leads to the definition of a new conceptual facies model for the Urgonian limestones (Fig. 2c and Tab. 1). The conceptual facies model is based on the definition of specific paleoenvironments along the platform profile and is supported by biotic and foraminiferal association analyses. The model does not correspond to any contemporaneous sedimentary profile as it merges models defined for different time intervals (Lower Barremian, Upper Barremian and Lower Aptian; cf. above).

Orbitolinid marly facies are recorded over large areas of the platform and within channels (cf.Arnaud-Vanneau, 1979, 1980) during specific terrigenous-influenced periods (Lower Orbitolina Bed within the Ai1 sequence and Upper Orbitolina Bed at the top of Ai2 sequence, respectively). The new conceptual model only integrates the channelised facies of the Upper Orbitolinia Bed.

A new Urgonian facies model of a large spatial and temporal extent is defined integrating observations of different sedimentological schools from the Valanginian to Lower Aptian from different peri-Vocontian platforms in SE France (Fig. 2d and Tab. 1). The proposed conceptual facies model is based upon regional palaeogeographic maps. Little detail is provided about the facies model in the publication. The proposed facies model is very similar to the previously published models and shows the organization of the main theoretical facies belts of an attached platform.

The model includes orbitolinid facies within channels cutting through the main facies belts, while most orbitolinid facies are found during specific terrigenous-influenced periods (“Lower and Upper Orbitolina Beds”; cf. above). Corals and ooids are shown as contemporaneous sediments that are deposited next to each other both along dip and along strike. Such a spatial organization was not characteristically observed in the Vercors facies maps of Arnaud-Vanneau and Arnaud (1976; cf. above).

This conceptual facies model is defined for the Lower Cretaceous of SE France and strictly corresponds to the model published previously by Arnaud-Vanneau (1979; cf. above; Fig. 2e and Tab. 1). No data is provided about the facies model in the publication, but details about typical characteristics of Urgonian limestones. A novel approach is introduced using sequence stratigraphic concepts; the lateral facies succession along dip is related to vertical log succession, which allows for direct interpretation in terms of sea-level changes. Such a facies model definition “allows interpreting temporal (vertical) succession of interpreted palaeoenvironments (microfacies) using the Walther’s law in order to recognize main discontinuities in a sequence stratigraphic framework” (Wermeille, 1996).

In terms of facies distribution, corals and ooids are not laterally contemporaneous as in the 1982 model but along dip (cf.Sect. 4.1.3). Coral facies association corresponds to the infralittoral palaeoenvironment (i.e. euphotic to mesophotic). While the ooid facies association belongs to the same shelf margin zone as the coral facies association, the ooid facies association, by contrast to Arnaud-Vanneau (1979), is placed together with the bioclastic facies association into the circalittoral zone (i.e. oligophotic to aphotic; cf.Stein et al., 2012; Bastide, 2014). Orbitolinid facies are not mentioned in the sedimentary profile, which follows observed depositional facies maps (cf.Arnaud-Vanneau and Arnaud, 1976), but not the conceptual facies model of Arnaud-Vanneau (1979; cf. above).

The definition of the facies models is based on palaeoecological concepts (Tab. 1). Two different, exclusive facies models were defined for the Lower Cretaceous interval of the Helvetic realm based on the recognition of different platform-scale grain associations or carbonate factories, which are controlled by specific oceanographic parameters (e.g.Föllmi et al., 1994; James, 1997; Schlager, 2003; Fig. 2f). Model occurrence is related with the trophic conditions of the waters at a given period (e.g.Föllmi and Godet, 2013). A photozoan, distally-steepened ramp is promoted by oligotrophic conditions, while a heterozoan, homoclinal ramp developed during mesotrophic periods. The photozoan facies model refers to the model published by Arnaud-Vanneau (1980) and Arnaud-Vanneau et al. (1982; cf. above); a supplementary algal mat facies occur at the side of the emersive facies. The new Urgonian heterozoan facies model is only composed of heterotrophic biota (e.g. crinoids, bryozoans and sponges) plus oolitic deposits. The palaeotopographic profile of these models is conceptual and results from interpretation of facies belt organization and open versus restricted oceanic biota.

The facies model is related with palaeotopographic characteristics, which are nevertheless speculative (Tab. 1). The rimmed-shelf model, which implies reefal slopes both seawards and shorewards, is not supported by any observation, as it is purely conceptual by referring to the classification of Pomar (2001) and based on coral-reef-barrier actualism (cf.Arnaud-Vanneau and Arnaud, 2005; Godet et al., 2016). In earlier publications of the Urgonian facies model, such rim interpretation is never mentioned in the text, only small coral patch reefs and ooid shoals are referred to as constituents of the platform margin, and the term “lagoon” is used in a general sense for inner platform facies of carbonate platforms (e.g.Arnaud-Vanneau et al., 1987). It is noteworthy that the topographic rim has no impact on the proximal-distal and associated sea-level-based sequence stratigraphic interpretation, and that this rimmed-shelf model will be abandoned in later studies from the same academic teams (e.g.Bonvallet, 2015; Bonvallet et al., 2019).

In terms of sequence stratigraphy, microfacies types of the conceptual facies model are used to define the main transgressive and regressive trends. A “facies of transgression” is hence described to characterise the main transgressive phases (also cf.Blanc Aletru, 1995). This “facies of transgression” is not included into the general facies model, but in the sedimentary profile used as a basis of sequence stratigraphic analyses. Thus, such depositional facies and facies model definitions do not refer to specific palaeoecological interpretation at a given time but include processes through time and is directly tied to sequence stratigraphic interpretation.

The facies model is defined using palaeogeographic Cretaceous maps of depositional facies, palaeontological studies mostly from Provence (Masse, 1976, 1993; Masse and Philip, 1981), and integrating the Vercors model by Arnaud-Vanneau (1980; cf. above; Fig. 2g and Tab. 1). Facies belt transitions are not directly observed. Beyond palaeogeographic arguments (Masse and Philip, 1981; Masse, 1993), palaeontological and sedimentary facies compositions are interpreted in terms of both hydrodynamics and light-related zoning along a proximal-distal transect (i.e. infralittoral vs. circalittoral; cf.Pérès and Picard, 1964; Masse and Philip, 1981). For instance, occurrence of abundant miliolids and medium grain size (i.e. muddy sand) relates the rudist facies association with the inner platform, in which small requieniid rudist facies are interpreted as more proximal than large requieniid and caprinid rudist facies. No palaeobathymetry is assigned to specific depositional facies. Instead, a shallow platform top (including an inner and outer domain; < 30 m water depth) and a deep circalittoral (i.e. subtidal oligo- to aphotic) palaeoenvironment are defined. The “platform model” (i.e. non-rimmed, flat-topped platform) implies a significant palaeotopographic and/or environmental change at the platform break that is not directly observed on the field. Orbitolinid facies are here considered as either deep-water facies (not always occurring over the platform profile; Provence) or a specific, transient, ramp model (Vercors).

This facies model was applied on continuous outcrops (Richet et al., 2011; Léonide et al., 2012; Frau et al., 2021) and its chronostratigraphic framework was recently modified and refined (Frau et al., 2017, 2018; Tendil et al., 2018).

A different model was defined showing two non-contemporaneous facies models based on sea-level contexts (Quesne, 1998; Quesne and Bénard, 2006; Ferry and Grosheny, 2019). A rudist and marly facies model occur during periods of high sea level, while a bioclastic facies model is deposited during times of low sea level. This model is not further investigated here because no facies model definition and representation are provided beyond stratigraphic trends. To be noted that this non-contemporaneous model is in contradiction with the harmonised facies model of the present study.

Palaeogeographic mapping of synchronous stratigraphic units was the foundations of the facies models of Urgonian platforms from SE France and Switzerland (cf.Arnaud-Vanneau and Arnaud, 1976; Masse and Philip, 1981; Arnaud-Vanneau et al., 1982). The continuous lateral facies transition exhibited by models is not observed, but instead are the results of interpolations of distant stratigraphic log sections along pluri-kilometer-scale transects and showing temporal resolutions of ca. 0.5–1 Ma. These deterministic approaches strongly rely on sedimentological concepts and hypothetical principles, which are generally inspired from modern carbonate environments such as the Bahamas archetype (e.g.Ginsburg, 1956; Tucker and Wright, 1990). Additional applied basics include fundamental palaeontological and sedimentological interpretation of rudists and foraminifera, grain-size and hydrodynamic relationships (e.g.Wilson, 1975; Immenhauser, 2009). In this context, the ecological classification of bathymetric intervals of Pérès and Picard (1964), which comes from the modern, warm-temperate Mediterranean Sea is a seminal contribution (also cf.Betzler et al., 1997). Later facies models directly used the above-cited, earlier palaeogeographic map-based models to interpret local depositional facies data in terms of either palaeobathymetric and/or proximal-distal trends. Consequently, the Urgonian facies models represent conceptual interpretations of stratal successions and correlations. In several cases, most key data supporting facies models come from earlier studies of the same authors that are not shown herein (Tab. 1).

Main differences between facies models are justified by the different conceptual approaches followed by authors (Fig. 4). The evolution of concepts about carbonate systems over the last century has influenced the definition of facies models. Improvement in depositional palaeoenvironment characterisation has forced diverse interpretative hypotheses about the palaeotopographic profile, the microfacies distribution along the palaeobathymetric profile and stacking pattern, the palaeoenvironmental influence on the carbonate factory and palaeontological successions.

The resulting facies belts of the models are overall closely similar to one another, not only showing a common origin but the consistency of sedimentary facies data across the entire peri-Vocontian region. Later studies, which observed lateral continuity of facies belt transitions over several kilometers, confirmed the general trends of Urgonian facies models (Bièvre and Quesne, 2004; Richet, 2011; Richet et al., 2011; Léonide et al., 2012; Frau et al., 2021). Two major differences between facies models, however, include the interpretative palaeoenvironmental setting of the orbitolinid facies and the interpretative palaeotopographic profile of ramp vs. flat-topped platform. The former point deals with different marly orbitolinid sediments, which are considered as either (i) proximal channelised structures in the Vercors (Arnaud-Vanneau, 1979; Arnaud-Vanneau et al., 1982), (ii) distal deposits in Provence (cf.Masse and Fenerci-Masse, 2011, Sect. 5.1.5; Tendil et al., 2018), or (iii) a distinct, terrigenous-influenced facies model such as the Lower Orbitolina beds in the Vercors and Helvetic Alps (Arnaud-Vanneau and Arnaud, 1976; Föllmi et al., 2006). The latter point of discrepancy about the palaeotopographic profile is critical for the interpretation of sea-level variations because the hypothesis of a flat-topped platform creates a dichotomy between proximal-distal vs. palaeobathymetric interpretation of facies associations (cf.Masse and Fenerci-Masse, 2011, Sect. 4). Platform-top facies showing the same or overlapping palaeobathymetric ranges do not allow for depositional facies change interpretation in terms of water-depth variation, and thus for the Walther’s law application.

The construction of the above-cited Urgonian facies models can be classified into three main approaches that are (1) topographic reconstruction, (2) microfacies succession and (3) depositional environment. Such a classification is deterministic and simplistic as these classes are neither independent nor exclusive. These approaches appear, however, to represent scientific evolution in facies model definition beyond the Urgonian realm (e.g. Middle East region).

This approach has been used in the Barremian–Aptian formations from the Middle East (e.g.Alsharhan, 1993; Hughes, 1997; Cantrell et al., 2014) and the Jura (Godet, 2006; Godet et al., 2010; see Sect. 4.1.6). It follows the work of Wilson (1975, Chapter XI) and is based on the analogy with modern reef-building corals at the origin of morphological reliefs on the seafloor (e.g.Arnaud-Vanneau and Arnaud, 2005). Following this “standard” platform profile, the rudists are considered as reef builders forming a barrier separating an internal (inner platform, lagoon, “shelf lagoon”) from an external domain (outer shelf). These shelf models imply a lagoonal depression with water depths in the order of 10 m (e.g.Alsharhan 1995; Godet et al., 2010). However, rudists were not shown to build up such morphological anomalies as it is the case for coral reefs (e.g.Gili et al., 1995; Quesne and Bénard, 2006). Outcrop and subsurface studies show instead that rudist limestones are deposited as extensive, tabular topographies without adjoining slope (Gili et al., 1995, 2016; Skelton and Gili, 2012; Skelton et al., 2019). Consequently, the relatively deep palaeobathymetric estimates of the sedimentary profile appear not to be supported by facies and palaeontological data.

This model, which was established from the end of the 80s by the Grenoble University and is based on the Vercors outcrops (e.g.Arnaud-Vanneau et al., 1987; Adatte et al., 2005; see Sect. 4.1.4), coincides with the onset of sequence stratigraphic concepts (Sarg, 1988; Jacquin et al., 1991; Handford and Loucks, 1993; Fig. 4). It has been mainly used for Urgonian limestones of the Jura–Subalpine domain (e.g.Adatte et al., 2005; Godet et al., 2010). In these models, interpretation of facies distribution is based upon the Modern Mediterranean ramp model of Pérès and Picard (1964). The facies and microfacies zonations are defined on a theoretical sedimentary profile by specific, exclusive palaeobathymetric ranges with a single polarity organizational trend from the coast to the basin. These models allow interpreting vertical successions of microfacies using the Walther’s law to define sedimentary discontinuities and stacking patterns within a sequence stratigraphic framework (e.g.Catuneanu et al., 2011; Pomar et al., 2015). Several studies have shown, however, some limitations of this approach on carbonate sedimentary systems; the distribution of depositional facies is a function of biocenosis and palaeoenvironment, and shows lateral changes within the same palaeobathymetric ranges (cf.Rankey, 2002; Wright and Burgess, 2005; Purkis et al., 2015, 2019; Pomar, 2020; Bover-Arnal et al., 2022).

The facies distribution is based on palaeontological and palaeoenvironmental interpretations. These interpretations infer relative, non-exclusive palaeobathymetric ranges and positions along ecological successions from proximal/shallow to distal/deep waters of specific biota associations (cf.Masse, 1976, 1991). No quantification is provided. Palaeontological data offer a chronostratigraphic framework so that ecological successions can be related to contemporaneous sedimentary profile. This model distinguishes an infralittoral (euphotic/mesophotic) and circalittoral (oligophotic/aphotic) zone (cf.Pérès and Picard, 1964; Betzler et al., 1997). By contrast with the Modern warm-temperate Mediterranean model defined by Pérès and Picard (1964), the infralittoral zone contains the whole tropical platform top (e.g.Masse and Philip, 1981; Masse, 1991; Masse and Fenerci-Masse, 2011). Drowning discontinuities (sensuSchlager, 1981; Masse and Fenerci-Masse, 2011), which mark drastic carbonate production changes, define sequence boundaries and thus sedimentary sequences. The model ignores published sea-level curves. Chronostratigraphic resolution is then limited to correlation based on carbonate platform demise, biostratigraphy and isotopic chronostratigraphy.

The first, “topography-based approach” implies the use of a conceptual model of ramp vs. shelf that is not supported by available data. Moreover, sedimentary profile morphologies are still debated in many cases (e.g.Rosen and Taylor, 1990; McCall et al., 1994; Pomar et al., 2017). The second, “microfacies-based model” uses exclusive palaeobathymetric ranges that are not well constrained, and might lead to some interpretation pitfalls that are related with the concepts of facies mosaic and carbonate sequence stratigraphy (e.g.Purkis et al., 2015; Borgomano et al., 2020, respectively). In the present study, the third, “palaeontology-based approach” is followed because it represents the most descriptive method using well-defined, consistent palaeoecological and sedimentological concepts. In addition, this latter approach allows for evolution because lateral and vertical facies distributional scheme does not show a forced, unique facies succession.

In the following, published data of Urgonian facies models are harmonised into a synthetic conceptual facies model (Fig. 3). Data supporting each definition step of the facies model are documented from the broadest to the finest spatial scale following the infrastructure framework of Handford and Loucks (1993) to explicitly describe the interpreted controlling parameters of the sedimentary model (Tabs. 2 and 3). Only carbonates are analysed; siliciclastics such as turbiditic sandstones (e.g. Vocontian Basin; Parize et al., 2007) are not included.

The facies model is defined by a palaeoecological facies belt succession constrained at a given time as far as geological record and dating allow. This choice is driven by the fact that at any given time, only contemporaneous palaeoenvironmental conditions control the biota and grain distribution, not the evolution of palaeoenvironmental conditions through time. Thus, neither climate change nor sea-level context is considered in theory. In practice however, the temporal integration of “contemporaneous sedimentary records” potentially biases observed palaeoecological lateral successions (e.g. occurrence of emersive facies within a subtidal facies belt resulting from accommodation infill up to sea level).

Urgonian biotic association is consistently set into the broader context of tropical carbonates (Tab. 3). The determination of grain association is interpretative (cf.Kindler and Wilson, 2010) and a debate about the photozoan character of rudists remains (e.g.Gili et al., 1995; Skelton and Gili, 2012; Skelton, 2018; Pohl et al., 2019). At a platform scale however, rudist platforms are consistently placed into the photozoan grain association (sensuJames, 1997) and the tropical carbonate factory (T-factory sensuSchlager, 2005; Michel et al., 2019). These platforms are thus interpreted to have developed in an oligotrophic, tropical palaeoceanographic context as witnessed by the coral patch belt that occupies the external parts of the platform (cf.Föllmi et al., 2006; Pohl et al., 2019; Michel et al., 2019). Accumulations of heterozoan grain associations, which are characterised by either (i) crinoids, bryozoans, brachiopods and siliceous sponges, or (ii) orbitolinids and detrital grains, are also described (Arnaud-Vanneau and Arnaud, 1976; Arnaud-Vanneau et al., 1982; Föllmi et al., 1994, 2006; Bodin, 2006; Michel et al., 2019). The differentiation of these grain associations led to the recognition of three exclusive Urgonian carbonate factories at the platform scale, namely the “photozoan rudist platform”, “heterozoan bryozoan-crinoid ramp” and “marly orbitolinid ramp”, which are diachronic and succeeded from one another through time as a result of palaeoceanographic changes (palaeotemperatures, nutrient concentrations and terrigenous influences; cf.Bodin, 2006; Föllmi et al., 2006; Masse and Fenerci-Masse, 2011).

The main uncertainty of facies models lies in the palaeotopographic profile (cf.Pomar, 2001). Geological data poorly allow for the reconstruction of original angle of repose. Consequently, major discussions about the palaeobathymetry of carbonate platform margins are still ongoing (i.e. surface waves vs. fair-weather and storm wave bases vs. internal waves; Immenhauser, 2009; Peters and Loss, 2012; Pomar et al., 2012; Purkis et al., 2015; including the Urgonian platform; Masse and Fenerci-Masse, 2011; Fig. 5). The concept of T-factory for rudist platforms would imply a flat-topped stratigraphic architecture and thus a flat-topped sedimentary profile as a general predictive rule (cf.Schlager, 2005).

On the field, continuous outcrop beddings and palaeogeographic reconstructions based on section correlations show that the tens-of-kilometer-wide, flat inner rudist platform is likely to be separated from the slope environment by a coral or ooid facies belt, the width of which can reach a few kilometers (cf.Tab. 3; e.g.Masse and Philip, 1981; Arnaud-Vanneau et al., 1982; Richet et al., 2011; Tendil et al., 2018). This latter interpreted platform margin shows high-energy, grainstone sediments along the coral and large-sized rudist facies (e.g. caprinid; Masse, 1976; Fenerci-Masse et al., 2005; Masse and Fenerci-Masse, 2008; Bastide, 2014; Fig. 5). This belt can locally form slight reliefs but does not create a real barrier associated with a depression, as per the definition of a lagoon (Gili et al., 1995; Quesne and Bénard, 2006; Skelton and Gili, 2012). The slope consists of calcarenite to calcisiltite facies passing to marly basinal facies. This lateral facies succession supports increasing palaeodepths and sloping geometries of a few degrees maximum from the platform margin to the basin (Tab. 3; e.g.Everts et al., 1995; Léonide et al., 2012; Richet, 2011).

This general sedimentary profile defines a tropical, unrimmed flat-topped platform (sensuHandford and Loucks, 1993; Pomar, 2001) with shallow-water platform-top and margin facies developed within the infralittoral, euphotic zone (Figs. 5 and 6). This flat-topped interpretation is supported by stratigraphic observations of rudist infilling of accommodation space and persistent shallowing-upward sequences reaching sea level (cf.Gili et al., 1995; Masse and Montaggioni, 2001; e.g. Gorges de La Nesque: Léonide et al., 2012). Such a levelling pattern is consistent with shallow-water, inner-platform rudist palaeoecology and massive rudist carbonate production (Steuber, 2000; Gili and Götz, 2018; Pohl et al., 2020).

Both heterozoan and orbitolinid carbonate systems are consistently referred to as ramps (e.g.Föllmi et al., 1994, 2006). Even though no palaeotopographic profiles were directly observed, lateral successions of grainy facies are considered to represent diagnostic features (e.g.James, 1997; Ahr, 1998).

The lateral succession of facies on tropical-type platforms was described in a standard facies model defined by Wilson (1975). This model has become a widely accepted framework for the representation of carbonate facies models (e.g.Tucker and Wright, 1990; Flügel, 2004; Schlager, 2005). In this model, the facies associations are represented as successive belts along dip. This model shows limitations such as a sharp transition between facies belts and the lack of asymmetry of the platform (e.g. leeward to windward; Schlager, 2005). In the present study, the facies are defined according to Wilson (1975)’s model and placed on a three-dimensional depositional model to image the possible lateral facies changes along strike (e.g. coral and ooid facies association; Fig. 3).

Four genetically related zones are distinguished from proximal to distal, namely the inner platform, outer platform, slope and basin; the inner platform being subdivided into supratidal, intertidal and euphotic subtidal zones (Fig. 3 and Tab. 2; cf.Sect. 5.5 for details). Facies textures and biota of each zone are related to common interpretation in terms of light penetration, hydrodynamics and water depth (Figs. 5 and 6 and Tab. 3; cf.Pérès and Picard, 1964; Immenhauser, 2009; Purkis et al., 2019; e.g.Masse et al., 2003). Large interpretative overlaps exist regarding facies distribution in terms of palaeobathymetry. These overlaps imply significant uncertainties as to their interpretation in terms of vertical stacking patterns and sea-level evolution versus facies mosaic due to contemporaneous hydrodynamic processes (cf.Wilkinson and Drummond, 2004; Wright and Burgess, 2005; Purkis et al., 2015). In turn, these uncertainties limit interpretation of exclusive trends in sedimentary sequences.

Additional details about facies, facies associations and supporting data of facies association definition can be found in Tables 2 and 4.

Facies textures and biota within an inner-platform palaeogeographic setting indicate an overall restricted, sheltered environment that is located within the supralittoral to infralittoral zones (i.e. supratidal, intertidal and euphotic subtidal).

The supratidal environment is rarely preserved in Urgonian records as a result of initial thin deposits and preferential erosion of these most proximal portions of the inner platform (Fig. 7 and Tab. 4). The most proximal, emersive facies show well-known sedimentary and palaeontological signals of continental, beach and restricted palaeoenvironments (Upper Barremian, G. sartousiana ammonite zone, southern Provence; cf.Masse et al., 2003; Tabs. 2 and 4). In addition, in relatively distal palaeogeographic zones of the platform in southern Vercors and northern Provence, hundreds-of-meters-scale sedimentary records of subaerial exposure are interpreted from hardground surfaces overprinted by meteoric diagenesis and from shoal tops that show keystone vug-bearing facies within rudist-dominated successions (Fouke et al., 1996; Richet, 2011; Léonide et al., 2012).

Interestingly, these latter emersive and beach facies surround rudist facies at the top of aggrading, shallowing-upward sequences. In a strict sense, these features highlight the ability of rudist facies to fill up accommodation space, locally allowing islands to form, rather than a strict palaeoecological lateral succession (cf.Gili et al., 1995; Masse et al., 2003). The latter interpretation would imply a facies belt shift over time in the sense of the Walther’s law. In terms of facies model, this sedimentary pattern implies that (i) palaeobathymetric range of rudist facies nearly reaches the sea surface (Masse et al., 2003), and (ii) rare records support a strict, consistent lateral transition between beach and rudist facies belts.

The extent of the peritidal areas remains difficult to estimate due to the limited preservation of proximal settings (Fig. 7 and Tab. 4). Characteristic facies and grains include stromatolites, charophytes, gastropods, miliolids, and peloids (Arnaud-Vanneau and Arnaud, 1976; Masse et al., 2003; Tendil et al., 2018; Fig. 6 and Tabs. 2 and 4).

This proximal subtidal part of the inner, flat-topped platform is characterised by a protected palaeoenvironment, the peloidal-foraminiferal facies association and miliolid, mollusc, microbial, algal and non-skeletal facies (e.g.Tendil et al., 2018; Tab. 2). Only few, small, scattered records occur locally throughout the peri-Vocontian region. This apparent scatter of peloidal and miliolid facies as well as intertidal and emersive facies within rudist facies association records of the peri-Vocontian platform does not allow for the direct observation of the facies belts and belt transitions. In practice, a continuous gradual trend exists in facies description between the peloidal-foraminiferal and rudist facies association end members that is not well imaged using log correlations (Tendil, 2018; Fig. 7). At a regional scale, palaeogeographic maps show that the proximal inner-platform facies belt would have been located in areas where no Urgonian record can be observed at surface due to a complex structural setting (proximal Jura-Subalpine and Bas-Vivarais Platform in the surroundings of the Massif Central palaeohigh; Masse and Philip, 1981; Arnaud-Vanneau et al., 1982; Tendil et al., 2018; Barbarand et al., 2020; Fig. 7). Thus, such a peloidal-foraminiferal facies end member is locally recorded in the Vercors, the Subalpine Chains and in North and South Provence (e.g.Masse et al., 2003; Frau et al., 2018; Tendil et al., 2018; Tab. 4).

The euphotic subtidal environment of the inner platform is dominated by rudist-bearing to rudist-rich facies with reference biostratigraphic materials being found at Cassis and Orgon (Urgonian limestone sensu stricto as defined by Masse, 1976). Such rudist facies cover most of the peri-Vocontian platforms of the Upper Barremian (Frau et al., 2018). They are roughly estimated to represent 50–60% of Vercors facies on palaeogeographic maps (cf.Arnaud-Vanneau, 1980) and 40% of Provence platform sediments both on palaeogeographic maps and in log sections (cf.Tendil et al., 2018; Tab. 4).

Rudist facies display a wide range of textures from wackestone to grainstone and floatstone and are made up of a diverse rudist fauna (Masse, 1992; Masse et al., 2003, 2020; Fenerci-Masse et al., 2005; Tab. 2). A well-defined palaeontological trend exists; while the proximal setting is dominated by small-sized rudists (e.g. requieniids), the distal part of the inner platform and the platform margin are characterised by the occurrence of large-sized rudists (e.g. caprinids; Masse et al., 1998; Masse and Fenerci-Masse, 2008).

The outer platform corresponds to the platform margin, high-energy and open-marine infralittoral zone (euphotic to mesophotic subtidal; Masse, 1976; Arnaud-Vanneau, 1980; Everts et al., 1995; Masse and Fenerci-Masse, 2011; Figs. 4–6 and Tab. 3), and is also referred to as the “shelf-break”. This depositional environment located at the platform-to-slope transition is dominated by bioclastic, echinoderm-rich facies, with local contribution of coral reefs and ooid-dominated deposits (Tab. 2). The bioclastic facies represents a large part of the Urgonian limestones sensu lato (as defined by Masse, 1976), and are found from the margin to the proximal slope. By contrast, both coral and oolitic sedimentary bodies show a limited, hundreds-of-meters to kilometers-scale lateral extent, and predominantly form at the platform margin (Tab. 4). Locally, the transition between the inner and outer platform environments is marked by the deposition of coarse-grained, coral and rudist fragment rudstones. These rudstone deposits possibly originate from short-lasting, storm-related events creating channel-like systems, as suggested by Léonide et al. (2012; Provence platform).

Coral and coral-rudist gravel facies associations are recorded as patches in the surroundings of the high-energy platform margin (Masse and Philip, 1981; Masse and Fenerci-Masse, 2008; Léonide et al., 2012; Gili and Götz, 2018); they are found at the transition between the rudist and bioclastic facies association (e.g.Richet et al., 2011; Léonide et al., 2012). Coral facies records are not rare but limited in spatial distribution overall; they are roughly estimated to represent 4% of Provence records (cf.Tendil, 2018; Tab. 4).

As coral facies, ooid facies occur in limited, localised zones that are interpreted as outer platform setting (Richet et al., 2011; Léonide et al., 2012; Tab. 4). Ooidal facies association includes both ooidal and oobioclastic packstone and grainstone facies (Tab. 2). Typical ooidal and coral deposits appear to exclude one another at any given time period (Arnaud-Vanneau and Arnaud, 1976; Richet et al., 2011). Such a direct contact, however, is interpreted in Provence in the Gorges de La Nesque area, where a fault separates both facies association (Léonide et al., 2012). The ooid facies association passes laterally into bioclastic facies (Arnaud-Vanneau and Arnaud, 1976; Richet et al., 2011; Léonide et al., 2012; Tendil et al., 2018).

The bioclastic facies association appears as the volumetrically dominant Urgonian sediments in SE France (Masse and Fenerci-Masse, 2011; Richet et al., 2011; Tendil et al., 2018). According to palaeogeographic maps, the bioclastic facies characterise the platform margin, namely the transition between the platform top and the slope environment (Arnaud-Vanneau and Arnaud, 1976; Arnaud-Vanneau et al., 1982; Tendil et al., 2018). These calcarenite facies consist of undetermined grains and fragments of rudists, crinoids, echinoderms, bryozoans, brachiopods and orbitolinids. Sand wave and wave ripple sedimentary structures are common. This facies association is interpreted as reworked sediments of the outer platform and proximal slope (e.g.Masse, 1993; Léonide et al., 2012). More specifically, these bioclastic facies are interpreted as genetically-related to the rudist facies association and includes an increasing proportion of in situ echinoderm and bryozoan grains towards more distal and aphotic settings (e.g.Léonide et al., 2012). In Provence, the bioclastic facies association forms a facies belt varying from 5 up to 20 km in width (Léonide et al., 2012; Tendil et al., 2018; Tab. 4).

The slope corresponds to the circalittoral (oligophotic to aphotic subtidal) zone that is characterised by lower light and hydrodynamic levels compared to the outer platform. Other terms that are used in literature include “external slope” and “outer shelf”. The slope is composed of very fine bioclastic facies made up of fragments of sponges, echinoderms, bryozoans and a significant part of transported sediments (e.g. mass flow; South Vercors and North Provence; Arnaud-Vanneau and Arnaud, 1976; Tendil et al., 2018; Tabs. 2 and 4). Orbitolinid facies were reported from this depositional environment in Provence (Gorges de La Nesque; Léonide et al., 2012; Tendil et al., 2018).

In many cases, the so-called basin strictly corresponds to the termination of the continental shelf (i.e. pre-Vocontian settings sensuFrau et al., 2018) and is mostly characterised by fine-grained calcisiltite facies (Tabs. 2 and 4). The Vocontian Basin sensu stricto is located further seaward and delineated by major faulting systems (e.g. Ventoux–Lure axis in northern Provence). These basinal environments correspond to the bathyal zone and are characterised by fine-grained transported, hemipelagic and pelagic marl mudstone facies (cf.Tendil et al., 2018).

A facies model of ancient sedimentary system is rarely built upon absolute and continuous real data, but rather on limited, recorded parameters generally extrapolated with deterministic interpretations (Fig. 7). In terms of data, spatial and temporal resolutions strongly limit the realism and accuracy of the model. Urgonian facies model data show an outcrop spacing varying from 3 to 55 km, and a temporal resolution between ca. 1 and 3 Ma based on biostratigraphic interpretation, increasing to 0.5 Ma based on lithostratigraphic and sequence stratigraphic interpretation (Tab. 1). In terms of interpretation, the use of concepts plays a leading role in defining a facies succession and thus a facies model (Tab. 3). Actualism views have a strong influence on the interpretation of facies distribution and their controlling parameters, and even more in a context of incomplete and uncertain data records (Ginsburg, 1956; Pérès and Picard, 1964).

Studying historical evolution can help understanding such a facies model definition of a specific zone (Figs. 2 and 3). For instance, a modern-type platform profile including a reef barrier built up to sea level is now well accepted not to be used as a direct analogue for ancient systems (e.g.Rosen and Taylor, 1990; McCall et al., 1994; Gili et al., 1995; Pomar, 2020). The birth and early developments of sequence stratigraphy had a strong imprint in the way a facies model and palaeobathymetric profile were defined (e.g.Arnaud-Vanneau et al., 1987; Wilkinson et al., 1996; Purkis et al., 2015).

The starting point of facies model definition is the depositional facies classification (Tab. 2). This basic point should not be considered as straightforward, because such a basic sedimentological observation does not provide absolute datasets but includes geological interpretation (cf.Lokier et al., 2016). Palaeontological and palaeoecological analyses represent further basics of facies model definition, even though high taxonomic level determination can be of limited value (Fig. 6). Urgonian facies definition benefits from the comparison of palaeontological and sedimentological studies for more than a hundred years that provide a great robustness in interpretation (Fig. 2).

A key constraint of facies model definition is the location of depositional facies on palaeogeographic maps. Such maps were utilised to show a convincing consistency of platform-scale, general Urgonian facies model (Masse and Philip, 1981; Arnaud-Vanneau et al., 1982; Tendil et al., 2018; Fig. 7 and Tab. 4). At this large scale, however, smaller-scale lithofacies heterogeneity is generally difficult to capture. Lithofacies heterogeneity is more easily recognized studying stacking patterns on log sections from which facies variability is directly observed (e.g.Borgomano et al., 2002; Godet et al., 2010). In a log section context in turn, the spatial interpretation of temporal stacking patterns in terms of either Walther’s law or facies mosaic is not straightforward (e.g.Purkis et al., 2015, 2019). A higher spatial resolution is required to be able to determine a better constrained facies model.

Continuous outcrop studies provide invaluable, high-resolution data about facies succession and transition. Lateral palaeontological and sedimentological transitions are directly observed on the kilometer-to-ten-of-kilometer-long outcrops and tens-to-hundreds-of-meters log section correlations of Gorges de La Nesque in Provence, Gresse-en-Vercors and the Cirque d’Archiane in Vercors (Léonide et al., 2012; Richet et al., 2011; Everts et al., 1995, respectively); these outcrops being in the surroundings of the palaeo-platform margin (Fig. 7). In a more proximal setting of the Barremian platform, down to 100-m-distant log sections were correlated in the Cassis area in Provence (Fenerci-Masse et al., 2005). As an example, the Gresse-en-Vercors outcrop directly shows lateral facies transitions from rudist to coral-rudist, coral-bioclastic and bioclastic facies along a continuously-walked-on-the-field time line (Richet, 2011; Richet et al., 2011). This lateral facies succession is consistent with platform-scale log correlations and palaeogeographic maps (Fig. 7). Such a consistency between rock observations at outcrop and platform scales and conceptual model strongly constrains the overall facies belt scheme (Figs. 4 and 7 and Tabs. 3 and 4).

Beyond the general, consistent facies distribution of these continuous outcrops, localized observations of limited amount of keystone-vug facies, coral–rudist rudstone facies, oolitic facies and orbitolinid–green algae grainstone facies highlight less constrained spatial and/or temporal sedimentological patterns such as transport or accommodation infilling events (Fig. 7 and Tab. 4). These patterns are considered as spatially random processes affecting a limited volume of sediments that show only minor imprints into the process-based facies model (Fig. 3). The possibility of the distinction between such minor vs. more general facies patterns remains questionable using a limited dataset.

Because the records of the Lower and Upper Orbitolinid Beds in the Vercors and the I. giraudi and M. sarasiniPalorbitolina-rich beds in Provence tend to exclude other platform-top sediments (Arnaud-Vanneau and Arnaud, 1976; Tendil et al., 2018, respectively), a specific facies model for orbitolinid-rich deposition is considered. A distinct facies model is justified by specific palaeoenvironmental conditions; terrigenous influences are repeatedly observed and invoked in association with Cretaceous orbitolinid-rich sedimentation (general discussion and Spanish Albacete-Prebetic case study – Vilas et al., 1995; Vercors, France – Arnaud-Vanneau and Arnaud, 1976; Helvetic Alps, Switzerland – Stein et al., 2012; Lusitanian Basin, Portugal – Burla et al., 2008; Kharaib and Shuaiba formations including the Hawar member of the Arabian Platform, Saudi Arabia, UAE and Oman – Borgomano et al., 2002; Davies et al., 2002; van Buchem et al., 2002; global pattern – Michel et al., 2019).

No specific palaeobathymetry is assigned to these orbitolinid-rich facies associations (Fig. 6). If orbitolinid grains can be found in diverse facies and certain orbitolinid facies can be found in specific palaeoenvironments (e.g.Masse, 1991; Léonide et al., 2012), orbitolinid-rich deposits are considered as opportunist and thus interpreted to occur in palaeoenvironments in which other biota do not occur (e.g. terrigenous-influenced settings; e.g.Föllmi et al., 2006; Michel et al., 2019).

A facies model definition results from the integration of chronostratigraphic, palaeontological and sedimentological data and their interpretations into a consistent conceptual framework of correlations (Tabs. 1–4). The key principles of definition include (i) building consistent conceptual models based on lateral palaeoecological succession concepts, palaeogeographic maps, direct lateral facies transition observations and consistent stacking trends, (ii) showing existing models and their local and general differences, and (iii) providing locations of records (e.g.Tendil, 2018), an estimate of recorded volumes, a detailed chronostratigraphic framework, and the basic hypotheses and uncertainties associated with the model (e.g. absence of record of most proximal Urgonian records; Fig. 7).

In many cases, the location of recorded facies and their relative quantities within the palaeogeographic context are neither shown nor mentioned due to limited temporal and spatial resolutions and preservation, which effects can hardly be distinguished and constrained. These data, however, are key for further studies opening to re-evaluation and re-interpretation of previous works in the light of advancing science. Questions about (i) the importance of the poorly preserved, most proximal deposits of a sedimentary system and (ii) the ramp vs. shelf interpretation in the absence or poor preservation of platform margin records are first concerned by such data (e.g. Lower Cretaceous Urgonian limestones – Masse and Fenerci-Masse, 2011; Oligocene Lessini Shelf – Pomar et al., 2017,vs.Bosellini et al., 2020; Miocene Yadana Platform – Paumard et al., 2017,vs.Teillet et al., 2020).

The Urgonian facies model provides a robust, platform-scale belt scheme and large-scale facies patterns (kilometers-to-tens-of-kilometer-scale; Figs. 4–7). In this context however, finer, hundreds-of-meters-scale facies heterogeneity cannot be captured precisely. Specific geobody distribution and continuous outcrop studies are required to study these finer-scale heterogeneities (e.g.Borgomano et al., 2002; Fenerci-Masse et al., 2005; Richet et al., 2011; Léonide et al., 2012). Numerical modelling can be used to tackle these questions (e.g.Budd et al., 2016; Massonnat et al., 2017; Lanteaume et al., 2018; Madjid et al., 2018; Tomassetti et al., 2018).

The facies model forms the basis to predict empty spaces of stratigraphic correlations and reservoir models. Facies of the facies model are used as elementary bricks, the consistent distribution (i.e. facies belts) of which allows analyzing rock property distribution and scale change in static modelling. The facies model defined in this study and associated physical parameters (hydrodynamics, distance to the coast and spatial organization; Figs. 4, 5 and 7) can be used as a constraint and input (external drift) for facies modelling. The graphic model representation and facies extent provide rules to determining proportional cubes for this type of platform (e.g.Tendil, 2018).

In general, models with standard facies belts are statically constructed. The static assumption is important to consider during the construction of stratigraphic models because carbonate systems largely evolve through time. These temporal variations generate transient and unstable behaviors. Nevertheless, the morphology of the Urgonian sediments makes it possible to analyse facies stability (cf.Schlager, 2005) during a favorable period because the production area is of a large extent (e.g. rudist facies association). A facies model aims at characterizing such a facies stability pattern. In addition, transient events on the Urgonian platform such as orbitolinid mass occurrences appear to be the effect of palaeoclimate changes (Arnaud-Vanneau and Arnaud, 1976; Föllmi, 2012), which in turn could be triggered by volcanic events (Tendil et al., 2018). Prediction of such a facies change depends on the understanding of palaeoenvironmental controlling mechanisms.

The method of the facies model construction promotes the use of process-based modelling. This model provides elements to define sedimentary supply, bathymetric and hydrodynamic profiles. The use of these parameter values on synchronous sedimentary layers will allow refining and challenging facies transitions. The chronostratigraphical assumption of the synchronous interpretation, which stands as the “article of faith” of facies model definition, needs constant cross-testing with new available data and concepts, as well as proof of concept by numerical modelling. Process-based modelling can provide ranges of uncertainties on qualitative and quantitative facies distribution. For example, a point to explore would be the quantitative relationship between the production of the inner-platform rudists and the outer-platform bioclastic facies. Process-based models could allow balancing and quantifying the needed production and transport of sedimentary facies to fill in a defined accommodation space.

Based on a literature review and a contextualization of interpreted data and models, this study provides a unified facies model for the Urgonian carbonate platforms of SE France. The facies model is described within an infrastructure framework sensuHandford and Loucks (1993) ranging from the overall carbonate system to the unitary depositional facies. The model analysis makes it possible to discuss and argue each element of the carbonate platform at a regional scale. We therefore propose a three-dimensional sedimentary model with facies belts along dip and lateral facies variability. Sedimentary facies are related with physical characteristics and processes such as palaeobathymetry and hydrodynamics that make the model readily suitable for static and forward modelling.

The proposed facies model is defined as a contemporaneous sedimentological and palaeoecological system and placed into a consistent palaeoecological framework. Data supporting the model including location and volume estimates are explicitly mentioned, as well as fundamental assumptions and uncertainties related to the model definition. These elements are considered as prerequisite prior to the building up of a robust facies model and its subsequent comparison to others.

Specific remarks on facies model definition can be put forward:

• the actualistic palaeotopographic model of lagoonal depression-marginal reef relief-steep slope of carbonate platform is broadly used even though no observation supports such an interpretation, which has long been cautioned (e.g.Rosen and Taylor, 1990; Gili et al., 1995; Pomar, 2020).

• building a facies model results from a complex integration of conceptual models along with spatial and temporal data interpretations. Model definition is based on key geological hypotheses that are generally not referred to in studies, while these assumptions could evolve with scientific advances.

G.M., J.B., M.R., J-P.R., C.D. and P.L. conceived the research project and provided a continuous input throughout the study. C.L., A.T., F.B., C.F., J.M. and M.B. realized the bibliographic compilation and data processing. C.L. and J.M. led the writing of the manuscript. All authors have defined the proposed facies model.

Thank you to all the reviewers who contributed to the great improvement of the manuscript. This study was funded by Total Exploration and Production in the ALBION project (https://www.ep.total.com/en/expertise/reservoir/albion-continuum-dynamic-analogues-heart-carbonate-reservoir). This paper strongly benefited from discussions with Jean-Pierre Masse, François Fournier (Aix-Marseille Université), Pierre Masse, Jérémy Robinet and Emmanuel Dujoncquoy (TOTAL S.E., Pau).