Nouveaux éléments sur la lithostratigraphie de la série éocène parisienne à partir de l’observation de centaines de sondages carottés (projet Éole et Grand Paris Express)

Extended abstract

The geology of the Parisian underground has been explored and recognized for a long time and the general succession and content or characteristics of the strata are generally well known (see explanatory notes accompanying the geological maps published by the BRGM, Soyer and Cailleux, 1960, or Filliat and Duvauchelle, 1981) and the geological knowledge may be very precise/detailed in certain areas. However, these historical works have also made it possible to highlight a great variability of the geology in its general framework and the difficulty of establishing a precisely reproducible lithostratigraphy for the whole of the Parisian subsoil.

The carrying out of many investigation surveys, in particular by drillings, for the new Eole and Grand Paris Express (GPE) metro projects has provided a considerable amount of new data about the Parisian geology allowing a better understanding, mainly in its periphery.

This article aims to present a synthesis of partly new observations, following the author's analysis of more than 800 core drillings carried out mainly along the Eole project and 15 and 16 lines of the GPE. The synthesis presented below focuses on the Eocene strata because they have been mainly recognized in boreholes, much more than the Upper Chalk of the Late Cretaceous, the Danian and Oligocene to Quaternary sedimentary units, which are therefore not presented here.

This article also addresses the weathering of Eocene strata, a phenomenon historically little described or even underestimated in the Parisian geological context.

Eocene lithostratigraphy

The Eocene series (Table 1) consists of an alternation of sand, clay, limestone and marl beds. Lateral variations are also observed, particularly in the Ypresian units. This variability results from the variation of the local depositional environments and more generally from the terrestrial to coastal contexts. Open marine facies are very rare, if not absent.

The Ypresian is classically divided into several lithostratigraphic units (Soyer, 1953 and 1966, Filliat and Duvauchelle, 1981), with in ascending stratigraphic order: 1) the terrestrial to lagoonal Sparnacian units, i) the ‘Argile plastique’ Fm at the base (floodplain to palustrine), overlain by ii) the ‘Sables d'Auteuil’ (shallow marine to lagoonal), then iii) the ‘Fausses Glaises’ (lagoonal to estuarine) and iv) the ‘Sables supérieurs’ (fluvial to estuarine), and 2), the shallow marine ‘Sables de Cuise’ Fm representing the Cuisian Stage (table 1).

However, the observation of numerous drill cores from the Ypresian enables to show that one frequently deviates from this lithostratigraphic scheme for the Sparnacian facies located above the ‘Argile plastique’ Fm.

The ‘Argile plastique formation is regularly observed and thus seems to constitute a continuous unit over the whole of the Parisian subsurface. It is mainly made up of characteristic variegated clayey facies (plurimetric to supradecametric in thickness) which often passes upwards to homogeneous (often dark) gray clayey facies a few meters thick. In certain areas, beds or sandy lenses plurimetric in thickness and plurimetric to pluridecametric in length are inserted within or at the base of the Formation (South or South-East of Paris), that may be interpreted as fluvial channels.

Above the ‘Argile Plastique’ Fm, the Ypresian strata are most often constituted by a set of dark siliceous and lignitic sands and gray-black organic clay associated to lignite beds and comprising silt laminae, the whole over a thickness of around twenty meters. It is mainly the vertical distribution and the proportion of these sandy and clayey facies that have historically led to the identification of the three distinct units in the reference lithostratigraphic scheme (‘Sables d'Auteuil’, ‘Fausses Glaises’ consisting of an alternation of sand and well-developed clay-lignite beds and the ‘Sables supérieurs’). But according to the observations on cores from the GPE and Eole, the distribution of the sandy and clay-lignite facies deviates commonly from the reference lithostratigraphic scheme and the distribution of Ypresian sandy and clayey terms is highly variable in the Parisian subsoil (Figure 3):

- It is therefore not possible to generalize the reality of the so-called ‘Sables d’Auteuil’ unit;

- The development and distribution of clay-lignite beds and lenses are highly variable, as illustrated by the sections in Figure 1;

- Most often, the upper part of the Sparnacian is composed of dominant siliceous and lignitic sands which constitute the ‘Sables supérieurs’ Fm, but clay-lignite lenses or beds are sometimes well developed over plurimetric to pluridecimetric thickness in this unit.

The ‘Sables de Cuise’ Fm, which belongs to the Cuisian Stage, contains glaucony which gives it a greenish tint (P. Diffre, 1980; Filliat and Duvauchelle, 1981). The ‘Sables de Cuise’ shall be distinguished from the underlying and lateral ‘Sables supérieurs’ by their petrography, mineralogy, paleontological content and facies, their depositional environment and former substage to which they were ascribed (glauconiferous sands containing marine molluscan shells, foraminifera and vertebrate remains, fully marine and Cuisian vs more or less lignitic, silty and clayey coarse to fine sands, very often pyritic, containing wood, leaves and other vegetal remains, fluviatile to estuarine and Sparnacian.

The Parisian Lutetian series is composed, in ascending order by the more or less fossiliferous ‘Calcaire Grossier’ Fm and ‘Marnes et Caillasses’ Fm. At the time of the intense exploitation of the ‘Calcaire grossier’ in the South of Paris, the quarrymen established a fine lithostratigraphy of the Formation (Tableau 2) by identifying and naming each bed. This detailed lithostratigraphic nomenclature is not generally found on the logs of the GPE-Eole boreholes, with the frequent exception of a few emblematic hard beds (the ‘Banc de Roche’ in particular).

More schematically, the ‘Calcaire grossier’ Fm is frequently divided into upper, middle and lower parts (see the Paris geological map explanatory notes at 1/50,000 scale, Filliat and Duvauchelle, 1981 or Merle Coord, 2008):

- The upper part is made up of several rather hard beds or of the single ‘Banc de Roche’ bed which marks the top of the Formation. One to three flint beds are frequently observed in this upper part;

- The middle part is made of medium to coarse limestone, plurimetric in thickness, generally not truly rocky;

- The lower part is plurimetric in thickness, glauconiferous and therefore characterized by a gray-green tint. Quartz grains associated with limestone and glaucony give a more sandstone character to this lower part.

At the base of the Lower glauconiferous ‘Calcaire Grossier’, a coarse quartzo-calcaro-glauconitic sandy to microconglomeratic and shelly facies corresponds to the ‘Glauconie Grossière’ unit as defined by R. Soyer (cf. Soyer and Cailleux, 1960) and is rather often observed here and there. This thin (generally several decimeters) and very discontinuous loose bed should not be confused with the glauconiferous, thick and essentially rocky Lower ‘Calcaire Grossier’ Fm.

The ‘Marnes et Caillasses’ Fm is mainly made up of an alternation of carbonate beds of moderate “hardness” (“semi-rocky limestone”) and hard beds or fragments mainly present at the base of the Formation (cf. ‘Banc de Rochette’). A few thin clay beds are regularly observed. To the North and East of Paris, beds of albastroïd gypsum (“ante-ludian” gypsum) are intercalated, with a thickness of several decimeters or meters. To the East of Paris, their cumulative thickness reaches 10 to 20 m (Figure 6 and Toulemont, 1987).

Many analyzes of the carbonate content were carried out on samples during the GPE and EOLE reconnaissance studies and in particular a series of systematic analyzes was performed every 20 cm on two core drillings of the EOLE project. The results (summary Table 2) indicate that the samples analyzed mainly correspond to pure limestone (or

dolomite) (>90% carbonates) or marly limestone (70-90% carbonates) but rarely to marls (<70% carbonates), whatever the texture of the beds analyzed. Apart from areas with developed gypsum, the ‘Marnes et Caillasses’ Fm is therefore essentially made up of limestone or dolomite, of variable texture/hardness, but little or no of marl.

Concerning the composition of the carbonate beds of the ‘Marnes et Caillasses’ Fm, 4 semi-quantitative XRD analyzes carried out on EOLE samples (Table 3) indicate very variable proportions of calcite and dolomite.

The gypsum present in the ‘Marnes et Caillasses’ Fm is subject to dissolution by the underground circulating water, which leads to the formation of cavities or, more often, of (very) destructured or decompressed zones (diffuse dissolution). These dissolution zones are unevenly distributed depending on the initial composition of the ‘Marnes et Caillasses’ Fm (presence or absence of gypsum) and the very irregularly developed effective dissolution.

The Bartonian is composed mainly of sands (‘Sables de Beauchamp’) and limestone (‘Calcaire de Saint-Ouen’).

The ‘Sables de Beauchamp’ Fm is made up of fine to very fine calcareous sands, often also silty to clayey. Since the work of J. Simonnet (2010), a clay level is frequently described in the middle position in the ‘Sables de Beauchamp’ Fm under the term “median clay screen”. On the geological sections (logs) of the GPE drillings, such a median clayey interval is quite often described from the visual observation of the cores (Figure 8). However, the analysis of logs of natural radioactivity (gamma-ray logging) carried up in the ‘Sables de Beauchamp’ Fm does not enable to confirm this observation as they highlight a random variability of the radioactivity and therefore probably of the clay content in these strata (Figure 9). Hence, most often, these logs do not identify a more median clayey interval in median position. It is therefore prudent not to generalize the existence of a median clayey interval within the ‘Sables de Beauchamp’ Fm.

The sand of the ‘Sables de Beauchamp’ Fm is generally loose but rocky and more or less hard intercalations decimetric to supra metric in thickness are observed irregularly in boreholes. The composition of these hard intervals is not the same everywhere: they are either siliceous (quartzitic), possibly very cohesive, very hard and very abrasive, or mainly calcareous (sometimes fossiliferous), concentrated in certain areas (cf. GPE St-Denis-Pleyel station).

Sandstone-siliceous (quartzitic) intercalations generally do not correlate from one borehole to another, their geometry is of the lenticular type and their dimensions are sometimes very small (decimetric thickness for a pluridecimetric length). Lightly or moderately indurated sand beds or lenses are observed as well, reflecting the more or less developed phenomenon of silicification.

The calcareous sandstones identified are sometimes correlated between two or three boreholes and for their part reflect a greater lateral extension than the siliceous (quartzitic) sandstone lenses, which enables to suggest a synchronous sedimentary or diagenetic origin of the sand cementation by carbonate as indicated by the presence of fossils (or their imprint) in these intervals.

The ‘Calcaire de Saint-Ouen’ Fm is made of a limestone facies which has strong similarities with the one of the ‘Marnes et Caillasses’ Fm. It is distinguished from these by generally more developed weathering due to its frequent superficial position (along lines 15 East and 16) and by a proportion of gypsum never as high as in the ‘Marnes and Caillasses’ Fm. The complete thickness (not eroded) of this Formation ranges most often from less than 10 m up to 15-20 m (10-12 m in average).

When located at depth and not or only slightly weathered, the ‘Calcaire de Saint-Ouen’ Fm is composed mainly of “semi-rocky” limestone, comprising intercalations or subdecimetric to suprametric fragments of harder limestone (more "rocky") and some dark and thin clay beds. According to the analyzes carried out for the GPE, the dominant limestone facies is generally marly limestone (<30% clay), and contains little marl in the proper sense, as in the ‘Marnes et Caillasses’ Fm. hard (“rocky”) limestone intercalations are rather localized in the lower part of the Formation.

Gypsum is irregularly associated with the ‘Calcaire de Saint-Ouen’ Fm in diffuse form or locally in well-defined massive beds, but of limited multi-decimetric to metric thickness. Note that the lower part of the Masse et marnes du Gypse (see below) represent the upper Bartonian and the end of the sequence of the Calcaire de Saint-Ouen and the Sables verts (Gely, 2016).

The base of the ‘Calcaire de Saint-Ouen’ Fm is represented over an approximately metric thickness by the Mortefontaine and Ducy Beds. However, these are not always easy to identify. According to our observations, the Mortefontaine Bed includes at the top a hard and fossiliferous limestone bed (with Avicula francei) 15-20 cm thick, overlying a pluridecimetric sandy unit. The Ducy Bed is represented by a limestone bed also 15-20 cm, but generally soft to crumbly and sometimes very little individualized.

The “Sables verts” or ‘Sables de Montceau’ unit constitutes a sandy-marly bed, metric to rarely plurimetric in thickness, sometimes difficult to identify or even not identified on cores due to characteristics that are often not enough distinctive compared to the units that encompass it.

Our observations lead us to consider that the ‘Sables verts’ unit is not continuous and regular but rather constitute a discontinuous lenticular bed, or even several ones potentially shifted in altitude in the series and possibly recurrent in the transition interval between the ‘Calcaire de Saint-Ouen’ Fm and the ‘Masses et Marnes du Gypse’ Fm.

Gypsum is rarely present within the “Sables verts”. However, beneath the Clichy hillock and to the North of it (L16), a gypsum bed is frequently described at the base of the ‘Sables verts’ over a thickness of several decimetres (up to 1.15 m thick in the survey).

The Priabonian is made up of the main part of the ‘Masses et marnes du Gypse’, ‘Calcaire de Champigny’ and ‘Marnes supragypseuses’ Formations.

The ‘Masses et Marnes du Gypse’ Fm represents one of the emblematic lithostratigraphic units of the Parisian Paleogene series due to the presence of thick layers of gypsum within it, which were widely quarried under Paris in the past. We include in this Formation some intercalated marl beds and the ‘Marnes infragypseuses’ and/or ‘Marnes à Pholadomyes’ unit, classically distinguished. Laterally, towards the East – North-East (North of the Marne valley along line 16) and towards the South, the ‘Masses et Marnes du Gypse’ pass to the ‘Calcaire de Champigny’ Fm, which presents its own and very distinctive lithological characteristics.

The ‘Masses et Marnes du Gypse’ Formation (MMG Fm) is present in its completeness only when it is covered by the younger Oligocene strata. This is particularly the case in the hillocks of Paris (‘Butte de Montmartre’) and around (‘Butte de Clichy’, etc.) where the Formation it is still protected from weathering and erosion by the overlying strata still in place and in particular the almost impermeable ‘Argile verte de Romainville’ unit (Table 1).

The composition of the MMG Fm and in particular the proportion of gypsum (thickness and number of gypsum beds) appears to be variable at the Parisian and surroundings scale: four thick beds or ‘Masses’ (whose the fourth an lower one is now included in the Bartonian series, Gely 2016) of variable thickness (up to 20 m for the higher ‘Masse’) are clearly identified in the North-East of Paris (Clichy Hillock), whereas in the South-West of Paris in particular, gypsum is encountered in borehole only on one or more layers much thinner.

In the low areas of the Paris region (such as the ‘Plaine de France’) between the hillocks and plateaus, the MMG Fm is generally represented above the ‘Sable verts’ unit only by an interval of marly more or less clayey soils generally quite compact and of rather homogeneous composition. This marly facies is interpreted as the residual lateral equivalent of the MMG Fm in which the gypsum initially present has been dissolved and the residual marls compacted following the dissolution. These residual marls are located in the plain where the impermeable strata which overlay the ‘Masses et Marnes du Gypse’ Fm and protect them from dissolution have been previously eroded.

The ‘Calcaire de Champigny’ Fm represents a lateral variation of the ‘Masses et Marnes du Gypse’ Fm and is characterized in particular by the disappearance of the gypsum beds. The cores taken in boreholes for the South-East part of the GPE show a great diversity of facies, from limestones (and marls?) more or less soft and limestones (dolomitic?) rocky (hard) and vacuolar. Characteristic dark brown cherts were regularly observed within a plurimetric interval of limestone.

The ‘Marnes supragypseuses’ Fm is made up of the ‘Marnes d'Argenteuil’ in the lower position and the ‘Marnes de Pantin’ above; they represent the lithostratigraphic units of the latest Eocene.

The analyzes on samples of the carbonate content carried out as part of the studies of different sections of the GPE indicate that these are indeed marls. However, these analyzes make it possible to determine an overall more calcareous lithology of the ‘Marnes de Pantin’ and a correlatively more clayey one in the ‘Marnes d'Argenteuil’, finally a few beds of hard limestone (CaCO3 >90%) have been highlighted in the ‘Marnes de Pantin’.

The superficial weathering of Eocene strata

Parisian Cenozoic geological formations are generally highly weathered when lying in superficial position and subject to rainwater. This weathering is notably responsible for the almost total dissolution of the gypsum in the ‘Masses et Marnes du Gypse’ Fm not protected on the topographic surface and their transformation into residual marls.

Similarly, the ‘Calcaire de Saint-Ouen’ Fm, frequently in a superficial position in the plains to the North (Plaine de France) and to the East of Paris, then appears in the form of a generally soft and/or very fragmented limestone facies (unstructured) reflecting a marked superficial weathering. In this weathered context, intervals or beds of harder limestone are often preserved as angular fragments. Some fragments are silicified and reflect additive weathering phenomena comparable to the superficial silicifications observed in Oligocene strata (‘meulières’ in particular).

The ‘Marnes et Caillasses’ and the ‘Calcaire grossier’ Fms are rarely in a superficial position, but when this is the case, these two Formations also undergo clear weathering. Contrary to the deep weathering of the ‘Marnes et Caillasses’ Fm by dissolution of gypsum, the superficial weathering of these and of the ‘Calcaire grossier’ Fm is little known and has sometimes been underestimated although in fact often observed (Egal and Piraud, 2018 and Figure 21, see also Quesnel’s observations in Grataloup et al., 2020). However, it leads to a potentially significant weakening of the geomechanical characteristics of the Formations.

In some areas, the ‘Calcaire grossier’ Fm in deep position shows dark and weathered intervals that could result from weathering phenomena by the water table movements, notably when it was in a much lower position than it is today (during the last glacial period? Or possibly during Lutetian times, during short emersion events, F. Quesnel, written communication).

Conclusion

The observations and descriptions here reported highlight the strong variability and heterogeneity of the Eocene series under Paris and its periphery within the large well-defined lithostratigraphic entities. The causes of these variabilities are largely of sedimentary origin but also in a notable way due to “deep” dissolution of gypsum and superficial weathering phenomena.

Keywords: Eocene, Paris, Geology, Boreholes, Grand Paris Express, Eole.