2. Geology of the Campine area

The geographic distribution of the Kasterlee Formation is limited to the Campine area (Fig. 1b). The formation is a micaceous, fine-grained sandy unit without macrofossils, slightly glauconitic and with intercalations of micaceous clay (De Meuter & Laga, Reference De Meuter and Laga1976). The lower boundary with the Diest Formation is often inconspicuous, but a discrete basal gravel has been observed in a few localities. Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007) proposed deposition at some time between 7.5 and 5.32 Ma (latest Tortonian to Messinian). Their palynological and geophysical study of the formation in the Dessel-2 borehole and the Olen outcrop (Fig. 1a, b) further revealed deposition in a shallow marine environment, characterized by a distinct shoaling phase in the upper part of the sequence (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007). It must be noted that the Olen outcrop is located near the southern boundary of Upper Miocene deposits in northern Belgium (Fig. 1a, b), and that deeper, neritic environments must have prevailed to the north. The present study area, Oud-Turnhout, is located approximately 20 km to the north of the Olen outcrop.

In parts of the Campine area, the shallow marine depositional setting of the Kasterlee Formation succeeded the distinctly transgressive, and fully marine environment of the underlying Upper Miocene Diest Formation (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007) (Fig. 2). The Diest Formation is a diachronous sequence genetically related to the eustatic sea-level rise following the sea-level low around the Middle–Upper Miocene boundary. Deposition started during early Tortonian times in the Campine area, and the centre of maximum deposition later shifted to the region north of Antwerp (Fig. 1) during late Tortonian to Messinian times (Louwye, De Coninck & Verniers, Reference Louwye, De Coninck and Verniers1999). This shift is expressed by distinctive prograding of the sequence to the northwest. The Deurne Sands are a local and marginal marine facies of the Diest Formation in the vicinity of the city of Antwerp, and are covered by the Lower Pliocene Kattendijk Formation. The age of the Diest Formation was already established from several locations in northern Belgium by Louwye (Reference Louwye2002), Louwye, De Coninck & Verniers (Reference Louwye, De Coninck and Verniers1999) and Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007): Tortonian in the greater part of the Campine area, early to middle Tortonian in the vicinity of the city of Antwerp (Deurne Sands), and Tortonian–Messinian north of Antwerp near the border with The Netherlands.

Figure 2.

Stratigraphic framework of the Neogene deposits in northern Belgium (after Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007).

The geographic distribution of the Poederlee Formation is restricted to the centre of the Campine area (Fig. 1b). De Meuter & Laga (Reference De Meuter and Laga1976) formally defined the Poederlee Formation as ‘fine, slightly glauconitic sand with small lenses of clay in the lower part; a base gravel of rounded quartz and flint, silicified limestones (the Hukkelberg gravel); a much oxidized upper part in the type region, sometimes containing limonitic sandstones with moulds of shells’. These authors designated the iron sandstone layers on top of the hills north of the village of Poederlee as the type section, and the village of Poederlee in the southern part of the Campine area as the type locality (Fig. 1a, b). The formation has a maximum thickness of about 10 m (Schiltz, Vandenberghe & Gullentops, Reference Schiltz, Vandenberghe and Gullentops1993). De Meuter & Laga (Reference De Meuter and Laga1976) broadly correlated the formation with the entire Lillo Formation from the Antwerp area (Figs 2, 3). The Poederlee Formation rests unconformably on the Kasterlee Formation throughout the study area (Buffel et al. Reference Buffel, Vandenberghe, Goolaerts and Laga2001; Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007) and is covered to the north by the Brasschaat Formation (Figs 1, 2), which was deposited in an estuarine environment. The absence of calcareous macrofossils, microfossils and organic-walled palynomorphs in the latter unit hampers its stratigraphic assessment.

Figure 3.

Previous correlations of the Poederlee Formation with members of the Lillo Formation in the Antwerp area.

Buffel et al. (Reference Buffel, Vandenberghe, Goolaerts and Laga2001) regarded the Poederlee Formation as a transitional facies between the fully marine Lillo Formation in the west (the Antwerp area) and the continental/estuarine Mol Formation to the east (Fig. 1b). Based on geometrical arguments and palaeontological evidence, including molluscs, these authors correlated the Poederlee Formation with parts of the shallow marine Lillo Formation (Oorderen Sands, Kruisschans Sands and lower part of Merksem Sands) from the Antwerp area (Fig. 3). They also correlated the Brasschaat Formation with the upper part of the Merksem Sands and Zandvliet Sands of the Lillo Formation in the Antwerp area. Buffel et al. (Reference Buffel, Vandenberghe, Goolaerts and Laga2001) proposed a Piacenzian age for the Poederlee Formation and a late Piacenzian age for the Brasschaat Formation, a viewpoint followed by Vandenberghe et al. (Reference Vandenberghe, Van Simaeys, Steurbaut, Jagt and Felder2004).

3. Historical background

According to Mourlon (Reference Mourlon1882), A. Dumont was the first to describe the sandy deposits in the hills around Poederlee around 1850, and the latter assumed a correlation with the Upper Miocene Diest Formation based on the ferruginous character of the deposits. Cogels & Van Ertborn (Reference Cogels and Van Ertborn1881) were the first to coin the term ‘La roche de Poederlé’ after a mapping campaign. A detailed overview of the literature regarding the stratigraphy of the Poederlee Formation until 1935 was presented by Gullentops & Huyghebaert (Reference Gullentops and Huyghebaert1999).

It was only after 1950 that new insights into Neogene stratigraphy of Belgium arose, based on observations made during large excavation works for the Antwerp harbour (de Heinzelin, Reference de Heinzelin1955). De Heinzelin grouped both the Diest Formation and the superjacent Kasterlee Formation in the Upper Miocene ‘Diestien’ stage, whereas the Kattendijk Formation, the Luchtbal and Oorderen Sands (as the Kallo Sands) were placed in the Pliocene ‘Scaldisien’ stage. Furthermore, a new Pleistocene stage, the ‘Merksemien’, was then established and consisted of the Kruisschans Sands and Merksem Sands. Based on the faunal similarity between the Merksem Sands and the Poederlee Formation (Fig. 3), de Heinzelin (Reference de Heinzelin1955) considered the Poederlee Formation as Lower Pleistocene, hence excluding the presence of Pliocene deposits in the Campine area.

A heterogeneous gravel layer at the base of the Poederlee Formation was observed in the Campine area (Gulinck, Reference Gulinck1960), and was named the Hukkelberg gravel after its locality. Geets (Reference Geets1962) studied mollusc moulds from the limonitic sandstone layer in the upper part of the Poederlee Formation, and suggested a correlation with the Kruisschans Sands of the ‘Merksemien’ stage (the present-day Lillo Formation) as defined by de Heinzelin (Reference de Heinzelin1955). Although the fossil content (mainly bivalves) is strongly weathered and limonitized, Tavernier & de Heinzelin (Reference Tavernier and de Heinzelin1962) identified Cardium parkinsoni, Corbulomya complanata and Mya arenaria, which allowed correlation with the Kruisschans Sands and Merksem Sands (‘Merksemian’ stage) (Fig. 3). These authors considered the Poederlee Formation as a transgressive unit of Pleistocene age in the entire the Campine area, with the Hukkelberg gravel at its base. Gulinck (Reference Gulinck1962) followed these ideas and also considered the Poederlee Formation as a lateral equivalent of the ‘Merksemian’ stage. Vandenberghe et al. (Reference Vandenberghe, Laga, Steurbaut, Hardenbol, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998) agreed with Tavernier & de Heinzelin (Reference Tavernier and de Heinzelin1962) that the Poederlee Formation is the lateral equivalent of the Merksem Sands, although they suggested that sedimentological similarities rather point at an equivalence with the lower part of the Lillo Formation. They attributed a Piacenzian age to the Poederlee Formation.

Buffel et al. (Reference Buffel, Vandenberghe, Goolaerts and Laga2001) studied the Kasterlee and Poederlee formations north of the type locality in four boreholes, and one borehole, located at Oud-Turnhout, is studied herein (Fig. 1a, b). The Kasterlee Formation consists lithologically of glauconitic sands that are fining upward in the unit. The uppermost part of the formation is a zone of poorly cemented, ferruginous sandstone fragments with internal moulds of dissolved molluscs. A basal gravel of small quartz pebbles (the Hukkelberg gravel) separates the overlying Poederlee Formation from the Kasterlee Formation. The lithology of the Poederlee Formation consists of fine-grained glauconitic sands with dispersed shell debris layers. The well-sorted, fine-grained sands in the upper two metres of the Poederlee Formation are called the Heieinde facies or unit (Fig. 4). Molluscs indicate that the Poederlee Formation north of the type locality is a lateral equivalent of the Oorderen Sands Member of the Lillo Formation (Fig. 3). Also, reworked shells from the Luchtbal Sands Member occur in the Poederlee Formation, suggesting that the Luchtbal Sands were deposited in the greater part of northern Belgium, but eroded before the deposition of the Poederlee Formation in the Campine area. Based on a re-evaluation of the mollusc fauna, Buffel et al. (Reference Buffel, Vandenberghe, Goolaerts and Laga2001) refuted the biostratigraphic correlation of the iron sandstone layer at the top of the Poederlee Formation with the Kruisschans Sands proposed by Geets (Reference Geets1962). Buffel et al. (Reference Buffel, Vandenberghe, Goolaerts and Laga2001) rather proposed a correlation of this layer with the base of the Merksem Sands of the Lillo Formation. In summary, they correlated the Poederlee Formation with the Oorderen Sands to Merksem Sands interval in the Antwerp area. Vandenberghe et al. (Reference Vandenberghe, Van Simaeys, Steurbaut, Jagt and Felder2004) considered the Poederlee Formation equivalent to large parts of the Oorderen Sands, Kruisschans Sands and Merksem Sands, and correlated the lower boundary with sequence boundary Za2 at 4.04 Ma (Hardenbol et al. Reference Hardenbol, Thierry, Farley, Jacquin, de Graciansky, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998). An exhaustive overview of the stratigraphy of the Kasterlee Formation is given in Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007).

Figure 4.

Lithological column and stratigraphic assignment of the units in the studied interval of the Oud-Turnhout borehole. The depths of the samples are referred to the Belgian Ordnance Datum (Tweede Algemene Waterpassing).

4. The study site

The studied borehole Oud-Turnhout (Lambert coordinates X: 192210, Y: 218431; Fig. 1a, b) was drilled in 1998 in the framework of a geological mapping project in Flanders. The borehole was entirely cored to a depth of 50 m (Fig. 4), and is now stored at the Belgian Geological Survey (archive number 017E/0399). It is located in the Campine area, north of the outcrop area of the Poederlee Formation. The lowest lithological unit (50.00 m to 38.71 m) corresponds with the Kasterlee Formation (De Meuter & Laga, Reference De Meuter and Laga1976). It is characterized by a dominant grain size of 125–250 μm, a mica and glauconite fraction, dispersed sand concretions, and varying clay content. Loosely compacted, glauconitic sandstone fragments with internal moulds of shells occur in the top of the sequence, together with fossil wood fragments. The base of this unit was not reached in this borehole.

At the base of the overlying Poederlee Formation (38.71–30.50 m), a layer of small, discoid-shaped quartz pebbles separates this unit from the Kasterlee Formation (Fig. 4). Between 38.71 and 33.30 m, the lithology consists of fine-grained sand with an admixture of coarse glauconite grains and mica. Clayey intercalations and bioturbations are present. Between 33.30 and 30.50 m, well-sorted, fine-grained sand with dispersed pebbles, some wood fragments, and traces of bioturbation are the Heieinde facies of this formation (Buffel et al. Reference Buffel, Vandenberghe, Goolaerts and Laga2001). An abrupt change to the glauconitic, medium-grained sand of the Brasschaat Formation occurs above the Heieinde facies at 30.50 m. A total of 19.50 m of cored sediments from the Kasterlee and Poederlee formations are thus recovered between 50.00 and 30.50 m depth.

6. Palynological analysis: results

In most samples, pollen grains are much more abundant than dinoflagellate cysts (Fig. 5). Only samples 37.38 m and 35.53 m yield more dinoflagellate cysts than pollen grains. Unfortunately, terrestrial palynomorphs are of limited biostratigraphic value in this study because of the selective loss of small specimens after sieving on a 20 μm mesh. The preservation of the marine palynomorphs is good to excellent. The presence of well-preserved protoperidinioid dinoflagellate cysts indicates that little or no post-depositional oxidation took place.

A total of 74 dinoflagellate cyst, eleven acritarch and five chlorophycean algae taxa are recorded in the Kasterlee and Poederlee formations together (Fig. 5). The assemblages in both formations are dominated by gonyaulacacean species. Highest numbers of protoperidiniacean cysts are recorded in the Kasterlee Formation (sample 41.75 m). Reworked pre-Neogene dinoflagellate cysts, pollen, spores and acritarchs are present in every sample (Fig. 5), but their numbers are relatively unimportant except in samples 34.70 m and 34.24 m, where respectively 1082 and 1443 reworked dinoflagellate cysts per gram are recorded.

A total of 47 dinoflagellate cyst taxa are recorded in the Kasterlee Formation. Undetermined specimens of the genera Spiniferites and Achomosphaera are grouped as Spiniferites/Achomosphaera spp. This group constitutes at least ~ 50 % of the assemblage in every sample, with the highest abundance of almost 65 % in sample 40.66 m. Lingulodinium machaerophorum and Operculodinium centrocarpum s.s. are other species with high relative abundances. Operculodinium tegillatum is well represented in the lowest sample (42.74 m).

Fifty-three dinoflagellate cyst taxa are recorded in the Poederlee Formation. The Spiniferites/Achomosphaera spp. group is abundant to dominant in every sample. Other common species are Achomosphaera andalousiensis suttonensis, Operculodinium centrocarpum sensu Wall & Dale (Reference Wall and Dale1966) and Spiniferites coniconcavus. The taxa Barssidinium pliocenicum, Bitectatodinium raedwaldii, Operculodinium centrocarpum s.s., Operculodinium? eirikianum eirikianum and Operculodinium israelianum have high relative abundances at the base of the unit.

Heteraulacacysta sp. A of Costa & Downie (Reference Costa, Downie, Montadert, Roberts, De Charpal and Guennoc1979) is recorded in almost every sample, and is very abundant at the top of the Kasterlee Formation (sample 38.96 m) and in sample 34.24 m of the Poederlee Formation. This species is a common element of Miocene (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007) and Pliocene assemblages (Head, Reference Head1993; Louwye, Head & De Schepper, Reference Louwye, Head and De Schepper2004; De Schepper, Head & Louwye, Reference De Schepper, Head and Louwye2009).

Cyclopsiella? trematophora (Fig. 7l) is the most common acritarch. Impagidinium solidum (sample 35.53 m; Fig. 7e, f) and the acritarch Lavradosphaera lucifer (sample 36.71 m; Fig. 7m–o) were recorded for the first time in the North Sea Basin.

Operculodinium tegillatum was defined by Head (Reference Head1997) from the mid-Pliocene of eastern England. According to De Schepper & Head (Reference De Schepper and Head2008b), the lowest occurrence of this species was recorded from DSDP Hole 603C in the western North Atlantic and is located at 5.02 Ma (based on unpub. data of M. J. Head). The morphologically similar Operculodinium? pontis was described for the first time by Zevenboom & Santarelli in Zevenboom (D. Zevenboom, unpub. Ph.D. thesis, Univ. Utrecht, 1995) from the Upper Miocene and Pliocene of the Boxmeer well in the Netherlands. Louwye & Laga (Reference Louwye and Laga1998) identified Operculodinum? pontis in the Upper Miocene Diest Formation and the Pliocene Kattendijk and Lillo formations in the Kalmthout well of northern Belgium. Louwye (Reference Louwye1999) formally defined Operculodinium antwerpensis from the Upper Miocene Diest Formation, and considered Operculodinium? pontis Zevenboom & Santarelli in Zevenboom (D. Zevenboom, unpub. Ph.D. thesis, Univ. Utrecht, 1995) as a manuscript name and synonymous. Louwye (Reference Louwye1999) regarded the continuous periphragm and the proximally to medially joined processes a sufficient argument to differentiate Operculodinium antwerpensis from Operculodinium tegillatum. However, new observations from the Kasterlee Formation showed both morphological criteria to be insufficient for discriminating between both species. The degree of perforation of the tegillum, and the degree of connection between processes, medially or proximally, is a variable feature within this taxon (Fig. 6i–l). Based on these observations, Operculodinium antwerpensis Louwye (Reference Louwye1999) is here considered as a junior synomym of Operculodinium tegillatum Head (Reference Head1997).

7. Biostratigraphy

7.a. Kasterlee Formation

The dinoflagellate cyst assemblage recorded from the Kasterlee Formation in the Oud-Turnhout borehole compares well to other records from the same formation in shallower environments identified in the Dessel-2 borehole and Olen outcrop (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007), although the relative abundance of certain species (e.g. Gramocysta verricula) is lower. Only four dinoflagellate cyst species encountered in the Kasterlee Formation have a biostratigraphic value and allow correlations to the timescale of Lourens et al. (Reference Lourens, Hilgens, Shackleton, Laskar, Wilson, Gradstein, Ogg and Smith2005).

Trinovantedinium ferugnomatum is recorded in low abundance in the Kasterlee Formation (Figs 5, 6h). Trinovantedinium ferugnomatum was recorded for the first time by de Verteuil & Norris (Reference de Verteuil, Norris, Head and Wrenn1992) in the Middle and Upper Miocene from the Maryland and Virginia Coastal Plain. The highest occurrence (HO) is placed in the dinoflagellate cyst zone DN10 of late Tortonian to Messinian age (de Verteuil & Norris, Reference de Verteuil and Norris1996) (Fig. 8). This species has also been recorded from Middle and Upper Miocene deposits of the southern North Sea Basin (Louwye, Reference Louwye2002; Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007; Louwye & Laga, Reference Louwye and Laga2008) and the Porcupine basin (Louwye et al. Reference Louwye, Foubert, Mertens and Van Rooij2008). Records from the Mogarts Beach Member of the Yorktown Formation (Upper Pliocene) by de Verteuil & Norris (Reference de Verteuil, Norris, Head and Wrenn1992) and the Lower Pliocene Kattendijk Formation of northern Belgium (Louwye, Head & De Schepper, Reference Louwye, Head and De Schepper2004; De Schepper, Head & Louwye, Reference De Schepper, Head and Louwye2009) probably represent reworking.

Figure 8.

Left panel: absolute and relative age, geomagnetic polarity, calcareous microfossil biozones, and Miocene dinoflagellate cyst zonation of the USA eastern coast. Time scale used is the ATNTS 2004 (Lourens et al. Reference Lourens, Hilgens, Shackleton, Laskar, Wilson, Gradstein, Ogg and Smith2005), but we follow Gibbard et al. (Reference Gibbard, Head and Walker2009) for the position of the Pliocene–Pleistocene boundary. Abbreviations in Polarity column are OLD – Olduvai, KAE – Kaena, MAM – Mammoth, COC – Cochiti, SID – Sidufjall, THV – Thvera. The dinoflagellate cyst zonation for the Miocene (DN zones) is from de Verteuil & Norris (Reference de Verteuil and Norris1996) and for the Pliocene (RT zones) from De Schepper & Head (Reference De Schepper and Head2009). Middle panel: stratigraphic range of selected dinoflagellate cysts used for the relative age assessment of the Kasterlee and Poederlee formations, for which a maximum and possible (minimum) age assessment is given. Right panel: combined dinoflagellate cyst and sequence stratigraphic age assessment for the Kasterlee and Poederlee formations. Their position in relation to other Neogene stratigraphic units in northern Belgium and to sequence stratigraphy of Hardenbol et al. (Reference Hardenbol, Thierry, Farley, Jacquin, de Graciansky, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998) is also given. Regressive (R) and transgressive (T) cycles are after Haq, Hardenbol & Vail (Reference Haq, Hardenbol and Vail1987). Thin dashed horizontal lines represent correlation and age indication. Arrows indicate time ranges. HML – Northern Hemisphere Mid-latitude; EQ – Equatorial. Ranges of dinoflagellate cysts and position of the Neogene units of northern Belgium are based on this study and de Verteuil & Norris (Reference de Verteuil, Norris, Head and Wrenn1992, Reference de Verteuil and Norris1996), Louwye (Reference Louwye1999, Reference Louwye2002), Warny & Wrenn (Reference Warny and Wrenn2002), De Schepper, Head & Louwye (Reference De Schepper, Head and Louwye2004), Louwye, Head & De Schepper (Reference Louwye, Head and De Schepper2004), Munsterman & Brinkhuis (Reference Munsterman and Brinkhuis2004), Williams et al. (Reference Williams, Brinkhuis, Pearce, Fensome, Weegink, Exon, Kennett and Malone2004), Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007), Louwye & Laga (Reference Louwye and Laga2008), De Schepper & Head (Reference De Schepper and Head2008a,Reference De Schepper and Headb), De Schepper, Head & Louwye (Reference De Schepper, Head and Louwye2009) and references therein.

Selenopemphix armageddonensis is present in low abundances in three samples of the Kasterlee Formation (Figs 5, 6f). It was previously recorded in the latest Tortonian to Messinian Kasterlee Formation by Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007). The latter authors give an extensive overview of published Miocene records of this species, and assume no in situ Pliocene records. According to Williams et al. (Reference Williams, Brinkhuis, Pearce, Fensome, Weegink, Exon, Kennett and Malone2004), Selenopemphix armageddonensis has a lowest occurrence at 9.0 Ma in equatorial areas and at 7.45 Ma in the northern hemisphere mid-latitudes. Based on data from Zevenboom (D. Zevenboom, unpub. Ph. D. thesis, Univ. Utrecht, 1995) from the Mediterranean realm, Williams et al. (Reference Williams, Brinkhuis, Pearce, Fensome, Weegink, Exon, Kennett and Malone2004) place the HO of this species at the Miocene–Pliocene boundary (5.32 Ma) in equatorial areas. Warny & Wrenn (Reference Warny and Wrenn2002) recorded Selenopemphix armageddonensis in the Messinian of Morocco and southern Spain, and place the HO also at the Miocene–Pliocene boundary (Fig. 8). Single records of Selenopemphix armageddonensis are reported in Pliocene deposits from the southern North Sea Basin (Louwye, Head & De Schepper, Reference Louwye, Head and De Schepper2004), from the Utsira Formation in the northern Viking Graben (Piasecki, Gregersen & Johannessen, Reference Piasecki, Gregersen and Johannessen2002) and from the Yorktown Formation in Virginia, USA (Edwards et al. Reference Edwards, Barron, Bukry, Bybell, Cronin, Poag, Weems, Wingard, Horton, Powars and Gohn2005). The latter records thus most probably represent reworking, since Selenopemphix armageddonensis is not reported from other Lower Pliocene assemblages.

Barssidinium taxandrianum is a rare species, recorded only from the southern North Sea Basin in Upper Miocene deposits. It was recorded for the first time in the Tortonian Diest Formation from northern Belgium by Louwye (Reference Louwye1999), and later also by Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007) and Louwye & Laga (Reference Louwye and Laga2008). Other records are from the early to middle Tortonian Deurne Sands Member (lower part of the Diest Formation) (Louwye, Reference Louwye2002), and the late Tortonian to Messinian Kasterlee Formation (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007). No Pliocene records are known.

A detailed overview of the stratigraphic occurrences and ranges of Gramocysta verricula is given by Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007). Geographically, the species is apparently restricted to the North Sea Basin, the western North Atlantic realm and the eastern Mediterranean, and has a published HO at the base of Chron C3r (middle Messinian) in the Miocene of Morocco (Warny & Wrenn, Reference Warny and Wrenn2002) (Fig. 8). No in situ Pliocene records are known. The morphology of Gramocysta verricula from the Kasterlee Formation is illustrated in Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007, fig. 8h–j).

Labyrinthodinium truncatum truncatum does not occur in the Kasterlee Formation of this study, but was recorded from the very basal part of this formation in the Olen outcrop (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007). It is a common species in the greater part of the Tortonian Diest Formation of northern Belgium. Its HO is situated in the upper part of Chron C4n at c. 7.5 Ma, that is, the latest Tortonian (de Verteuil & Norris, Reference de Verteuil and Norris1996; Munsterman & Brinkhuis, Reference Munsterman and Brinkhuis2004; see Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007 for an overview of the stratigraphic ranges) (Fig. 8). Species with a known post-Miocene lowest occurrence (LO) are not encountered in the Kasterlee Formation.

The age of the Kasterlee Formation from the Oud-Turnhout borehole relies on the ranges of key dinoflagellate cyst species such as Selenopemphix armageddonensis, Trinovantedinium ferugnomatum, Barssidinium taxandrianum and Gramocysta verricula, and the absence of Labyrinthodinium truncatum truncatum and typical Pliocene species. Because only the upper part of the sequence was analysed (Fig. 4), it is obvious that an age estimate for the entire formation cannot be proposed. The formation at Oud-Turnhout is younger than 7.5 Ma and older than 5.32 Ma.

8. Palaeoenvironmental analysis

9. Discussion and conclusions

In the Oud-Turnhout area, the sediments of the Kasterlee Formation were deposited in a neritic environment as attested to by the dominant neritic dinoflagellate cyst assemblage. Terrestrial palynomorphs significantly outnumber the dinoflagellate cysts in almost every sample and are indicative of the nearby presence of the continent. The quasi-absence of freshwater algae points to a limited fluviatile input. The Kasterlee Formation in the area of Oud-Turnhout was thus deposited in deeper neritic environments compared to the Dessel-2 borehole and Olen outcrop (Fig. 1), where a shallow marine facies is recorded following deposition in a truly coastal environment, characterized by a distinct shoaling phase (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007). Therefore, the Kasterlee Formation shows in the Campine area a deepening of the depositional environment in a northerly direction.

The stratigraphic ranges of the dinoflagellate cysts confine the age of the Kasterlee Formation in the Oud-Turnhout borehole between 7.5 and 5.32 Ma (latest Tortonian to Messinian), comparable to the age proposed by Louwye et al. (Reference Louwye, De Schepper, Laga and Vandenberghe2007). The unconformity at the top of the Kasterlee Formation can be correlated with the major sequence boundary Me2 of Hardenbol et al. (Reference Hardenbol, Thierry, Farley, Jacquin, de Graciansky, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998), which is dated at 5.73 Ma. The combination of dinoflagellate cyst stratigraphy with sequence stratigraphy narrows the time frame for deposition between 7.5 and 5.73 Ma (Fig. 8).

Messinian depositional sequences are characterized globally by regressive trends, caused by lowering sea level due to global increasing ice volume (Adams et al. Reference Adams, Benson, Kidd, Ryan and Wright1977). A δ¹⁸O isotope study with benthic foraminifers of a Late Miocene to Early Pliocene sequence from ODP Site 982 (North Atlantic) revealed a late Messinian glacial period between 6.26 to 5.50 Ma (Hodell et al. Reference Hodell, Curtis, Sierro and Raymo2001). This glacial period consists of eighteen glacial–interglacial cycles, of which TG20 at 5.75 and TG12 at 5.51 Ma are the most intense. The sequence boundary Me2 at 5.73 Ma (Hardenbol et al. Reference Hardenbol, Thierry, Farley, Jacquin, de Graciansky, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998), correlated to the unconformity at the top of the Kasterlee Formation, is thus likely related with glacial event TG20. According to Hodell et al. (Reference Hodell, Curtis, Sierro and Raymo2001), the magnitude of the associated sea-level drop is difficult to determine in open marine settings, but estimates range from ten metres (Aharon et al. Reference Aharon, Goldstein, Wheeler and Jacobson1993) to tens of metres (Braga & Martin, Reference Braga and Martin1996). In summary, the regressive trends and shoaling observed in the Kasterlee Formation (Louwye et al. Reference Louwye, De Schepper, Laga and Vandenberghe2007) can be set against the backdrop of the onset of the late Miocene glaciation. Although the effect of a sea-level drop is first observed in coastal areas, the top of the Kasterlee Formation in the deeper (Oud-Turnhout area) and shallower facies (Dessel and Olen areas) shows no age difference. The Messinian glaciation ended with the transition from TG12 to TG11 just before 5.5 Ma, after which sea levels started to rise again.

The depositional environment of the Poederlee Formation was neritic with an open ocean influence at first, changing into more shallow waters in the proximity of the coast and fresh water influence. Deposition likely took place during a transgressive to highstand phase, under a high and possibly maximal sea level, until the accommodation space was filled. At the edge of the North Sea Basin, accommodation space is limited; hence, from halfway up the Poederlee Sands, fresh water influences are recorded together with increased pollen influx showing the proximity of the continent.

The stratigraphic range of marker dinoflagellate cyst species confine the age of the Poederlee Formation between 3.71 and 2.74–2.72 Ma. Given the stratigraphic range of Impagidinium solidum (3.87–3.15 Ma), the Poederlee Sands could be older than 3.15 Ma. Based on the range of marker species and the overall dinoflagellate cyst assemblage, the Poederlee Formation is correlated with the Oorderen Sands. The Oorderen Sands are placed in the sequence between the Pia1 (3.21 Ma) and Pia2 (2.76 Ma) sequence boundaries (Hardenbol et al. Reference Hardenbol, Thierry, Farley, Jacquin, de Graciansky, Vail, de Graciansky, Hardenbol, Jacquin and Vail1998; De Schepper, Head & Louwye, Reference De Schepper, Head and Louwye2009). Therefore, also the lower boundary of the Poederlee Sands should be correlated to the Pia1 (3.21 Ma). This sequence boundary is likely related to the global glaciation event of Marine Isotope Stage M2 at around 3.30 Ma (e.g. Lisiecki & Raymo, Reference Lisiecki and Raymo2005; De Schepper, Head & Groeneveld, Reference De Schepper, Head and Groeneveld2009). Miller et al. (Reference Miller, Kominz, Browning, Wright, Mountain, Katz, Sugarman, Cramer, Christie-Blick and Pekar2005) and also Dwyer & Chandler (Reference Dwyer and Chandler2009) estimate the sea-level low stand at MIS M2 at approximately 65 m below present. The Pia2 sequence boundary (Fig. 8) corresponds with the sea-level drop of about 75 m around 2.74 Ma (Miller et al. Reference Miller, Kominz, Browning, Wright, Mountain, Katz, Sugarman, Cramer, Christie-Blick and Pekar2005) at the intensification of the Northern Hemisphere glaciation.

The combined evidence from dinoflagellate cysts and sequence stratigraphy suggests that the Poederlee Formation was deposited between 3.21 and 2.76 Ma, but may have even been deposited in the narrow time window between 3.21 and 3.15 Ma. Sediments from the eastern English sector of the southern North Sea Basin often have been deposited in a short amount of time, even decadal timescales (Head, Reference Head1998; Balson, Reference Balson, Daley and Balson1999). Thus, deposition of the Oorderen Sands and Poederlee Formation likely took place during the Mid-Pliocene warm period or PRISM time slab (3.27–2.97 Ma), when sea level was high enough and accommodation space was adequate for sediments to be deposited. Sea level was higher by around 35 ± 18 m (Dowsett & Cronin, Reference Dowsett and Cronin1990), but more recent reconstructions suggest a sea level of 20 to 25 m higher relative to today (Miller et al. Reference Miller, Kominz, Browning, Wright, Mountain, Katz, Sugarman, Cramer, Christie-Blick and Pekar2005; Dowsett, Reference Dowsett, Williams, Haywood, Gregory and Schmidt2007; Dwyer & Chandler, Reference Dwyer and Chandler2009).

The magnitude of the hiatus between the Miocene and Pliocene series in northern Belgium varies considerably from region to region, and no continuous section exits. In the vicinity of the city of Antwerp, the hiatus between the Diest Formation (more specifically the Deurne Sands Member) and the Zanclean Kattendijk Formation can be estimated at a minimum of 3.2 million years (Fig. 8). The hiatus is at least 2.52 million years in the Campine area, where the Kasterlee Formation is covered by the Poederlee Formation. No precise estimation can be given for the areas where Miocene deposits are covered by the continental/estuarine Mol Formation for which no precise age can be proposed (Fig. 8). The major cause of this variation is the shallow depositional environment at the southernmost border of the North Sea Basin, where glacio-eustatic or local tectonically induced sea-level fluctuations determine deposition or non-deposition in marginal marine areas.