A Late Palaeocene–Early Eocene benthic foraminiferal record from Bovlstrup, Denmark, showing a remarkable agglutinated fauna

The Bovlstrup well, Denmark, provides a detailed record of benthic foraminifera from the Upper Palaeocene and Lower Eocene deposits. The investigated interval spans four litho-units: an informal Grey Clay unit, the Holmehus Formation, the Ølst Formation and the Røsnæs Clay Formation (Danian?–Ypresian). Five interval zones based on benthic foraminifera have been established. Three of these zones (Zones 2, 3, and 4) contain exclusively agglutinated faunas. No foraminifera have previously been found in the Ølst Formation (Late Thanetian–Early Ypresian), but at Bovlstrup the formation contains a remarkable low-diversity agglutinated fauna (Zone 4). A programme of relatively dense sampling yielded information that may be lost in commercial oil well analysis. The five foraminiferal zones at Bovlstrup are correlated to established North Sea zonations, and the recognition of the faunas of Zones 3 and 4 leads to the conclusion that the zonation of King (In: Jenkins, D. G. & Murray, J. W. (Eds), Stratigraphical Atlas of Fossil Foraminifera, Ellis Horwood, 1989) can be refined. The benthic faunas indicate changes in the bottom environment both at the sea floor and within the overlying water mass. A transition from a calcareous fauna to an agglutinated fauna is interpreted as the result of a change from a neutral to a slightly acidic environment at the sea floor. There is a fluctuation in water depth through the studied section with a minimum water depth during the Thanetian and Early Ypresian. Volcanic ash layers in the Ølst Formation presumably resulted in low pH values, thereby causing the extreme low diversity of the benthic foraminiferal faunas.


INTRODUCTION
This paper presents the results of an investigation of Upper Palaeocene to Lower Eocene foraminiferal faunas of the Bovlstrup well, East Jutland, Denmark (Fig. 1). Previous studics on the foraminiferal stratigraphy of the Upper Palaeocene-Lower Eocene sediments in Denmark (Franke, 1927: Berggren, 1960aHofker, 1966: Hansen & Anderwri, 1066: Hansen, 1968Dinesen, 1972;Petersen &k Buch, 1973;Dinesen et a/., 1077: Larsen & JGrgensen. 1977, have concentrated on the calcareous foraminifera from this time irittxval. The agglutinated foraminifera remain very poorly documented and are mentioned only by King (I994a,b). Th': Bovlstrup sequence, however, possesses relatively large nurnbers of agglutinated foraminifera, and the purpose of this paper is to document this rich arenaceous fauna. Three zones comprising exclusively agglutinated faunas have been established, including the first record of foraminifera from the Olst Formation close to the Palaeocene-Eocene boundary. 7'he faunas of the Bovlstrup well are described and co'rrelated below and any possible connection between the faunas and the lithological formations are assessed.
The well :site at Bovlstrup (Fig. 1) is approximately SO m above sea level and the investigated section from 133 to 212 m depth below surface (See Fig. 2) includes a sequence o f lithological units which represent different marine environments. The litho-units are, in ascending order: an inlorn-la1 Grey Clay unit, the Holmehus Formation, the 0lst Formation and the RGsnzs Clay Formation. The lithological units in the section are widely distributed in Denmark and are cc!'mprehensively described by Heilmann-Clausen et ul. (1985). They are equivalent to the VNe Formation (or diversity o f the benthic foraminiferal faunas. Maureen Formation equivalent), Lista, Sele/Balder Formations and the lower part of Horda Formation respectively (see Fig. 3b) (see also Isaksen & Tonstad, 1989;Knox & Cordey, 1992: O'Connor & Walker 1993: King, 1994a. The depth of the formation boundaries at Bovlstrup are indicated on Fig. 2. The precise location of the boundary between the Holmehus and 0lst Formation is uncertain in the Bovlstrup well (drawn as a diagonal) as the sediments are not characteristic of any of the formations.

MATERIAL AND METHODS
A total of 53 ditch cuttings samples were prepared for foraminiferal analysis at an average 1 m spacing between 180 and 212 m, although spacing was variable (1-4 m ) between 133 and 178 m. The samples were washed through two sieves with mesh diameters of 0.1 and 1.0 mm, and the fraction between 0.1 and 1.0mm was examined. To avoid the formation of acidic solutions and the disintegration of the agglutinated foraminifera, a peptisizing agent (Na,P,O,, 10HZO) was used to disintegrate the sediments instead of hydrogen peroxide (H202). Where possible, at least 300 specimens were counted. In samples with few foraminifera the entire content was counted.
In order to solve some of the taxonomic problems, the internal structure of some agglutinated species excluding those that were pyritized were studied after staining with methyl violet: the specimens were soaked in methyl violet dissolved in anisoil. When the colour was absorbed by the test, they were dried and placed in uncoloured anisoil. The colour slowly disappeared until only the chambers remained stained. After drying, the internal structures of the chambers could be studied in transmitted light using a fresh drop o f anisoil. Sedimentological descriptions (see Fig. 3b) and the division into litho-units were performed by Heilmann-Clausen in accordance with his definitions ol the units (Heilmann-Clausen c~ d . , 19x5).
Six samples were analyzed for their content of dinoflagellates by C. Hcilmann-Clausen and referred to his xonation of the Viborg 1 borehole (Heilmann-Clauscn. 1985).

BENTHIC FORAMINIFERAL ZONATION AND PALAEOECOLOGY
The examined part of the Bovlstrup well has been subdivided into 5 interval zones and 3 subzones ( Fig. 321) according t o the definition of Hedherg (1976). 'l'he faund characteristics are described in ascending stratigraphic order in spite of the fact that the samples are ditch cuttings. Ecological information given by Murray (1971Murray ( , 1991 and Charnock & Jones (1990) about genera and extant species is uced as the basis for the present palaeoecological interpretations, with the proviso that the ecological requirements of the taxa remained unchanged in the intervening time.
Zone 1: The Eponides sp. 2-Bathysiphnn ex. gr. discreta Zone The zone consists of a single sample (212in below surface (mbs); Fig. 3a) of grey cdcnreous clay, and belongs to the informal Grey Clay litho-unit (Heilmann-Clausen cr id.. 1985). The lithological character, the foraminifera1 content and the age suggest that this is an equivalent to thc VZle Formation (or 'Maureen Formation equivalent') identifiable in the Danish and Norwegian sectors of the North Sea (Isaksen & 'l'onstad, 1989: Knox & Cordey. 1992: O'C'onnor Xr Walker, 1993. The bcnthic fauna is characterized by the presence o f low  (pers. comm.. 1993  Benthic loraminitera, Denmark lower part of Subzone .ZA consists o f a grey clayey sediment and is part of the informal Grey Clay litho-unit (VSle Formation equivalent). The remaining part o f Zone 2 comprise,< grelin-grey clay belonging to the Holmehus Formation ( Fig. 3b), which is equivalent to the Lista Formation in the North Sea (King, 199421).
Zone 2 is easily distinguished from Zone I by its toraminileral fauna o f exclusively agglutinated species (Fig.  3a) Murray, I 90 1 :I. Similar agglutinated foraminiferal ass e m b l a p havc been considered as indicators of reduced water circulation with low oxygen and high levels at the sediment,'water interface (Gradstein B( Berggren, 1981: Gradstein cr al., 1992. However, a very low content o f organic material (0.3-0.5%: H. Skovbjerg, pers. comm.. 1 988) and Zoophycos burrows i n the large cuttings trom the Holmehu,j Formation (also seen in outcrops; Heilmann-Clausen, 1989) are observed at Bovlstrup. These observations sugcest that the bottom waters were moderately to well oxygenated as indicated by Heilmann-Clausen el ul. (1985). A possible explanation to this paradox, could be that pH was \lightly below 7.8 as a result o f an increased CO, content through decomposition of organic matter in sinking particles (Moorkens, 1975: Boyle, 1988. This situation would lavour the removal of tlhe CaCO, and most of the organic matter although the bottom waters would still be moderately oxygenated. sufficient for the burrowing organisms to live. The removal of CaCO, could also explain the sudden change from a calcareous fauna t o an exclusively agglutinated fauna at the lower zonal boundary.
The Sulbzont 2B assemblage is apparently the result o f secondary dissolution where only strongly cemented species are prescrved. However, sonie aberrant specimcns o f .Sl,iroiilrc.i.unini irirr spt~crrihilis arid Huploplzragriioirlcs sp. I are observed at the boundary between Subzone 2A and Subzone 2B. These species are smaller in size than in Sub7ones 2A and 2C, while U.shekistaniri c~hriroir1e.s is larger. This is interpreted as a result of deteriorated living conditions for the foraminifera in Subzone 2B compared to the resl of Zone 2.
Zone 3: The Verneuilinoides subeocaenus-Haplophragmoides sp. 1 Zone The zone was identified in two samples (161-160mbs) of dark grey-green clay with a bluish tint and occurs at the transition between typical sediments of the Holmehus Formation and the 0lst Formation (Fig. 3b). As the sediments are not typical for any of thc formations the location of the formation boundaries are uncertain (Figs 2 and 3b).
Coarse-grained volcanic ash was identified in the sample at 161 ni (J. Eiriksson, pers. conim., 1993). At 160 in a peak in the K-curve or the gamma-ray log (Korsbech & Nielsen, 1089) indicates the presence of a glauconite layer corresponding to the glauconite horizon, which is normally observed at the base of the @lst Formation (Heilmann-Clausen er i d . , 1985).
The lower boundary o f the zone is characterized by thc change t o dominance of Ve~rrierrilinoirle~.s .srrhtwcaeni~s (restricted to this zone) and Ilaplophrrignioirlrs sp. I (Fig.  31). Another typical species in the zone is

I l t r~~l o p h r a~~i i i o i r l r s
walreri. The diversity is lower than in Zone 2. The diatom C'osc~inorli.\cirs cl. sp. 1 is restricted to this zone.
The combination of Gloniospiru and I-lrip1ophrrrgmoide.s species (Fig. 3a) suggest an outer shelf environment (cf. Jones & C'harnock. 1985) while the low species diversity points to extreme bottom water conditions with unfnvourable living conditions. The zone is characterized by a very low number of species, and rather fluctuating concentrations (Fig. 3b) The Bovlstrup well is the first site where a foraminiferal fauna has been identified in sediments belonging to the @lst Formation. I t is possible that the lack of documentation of loraminifera from t h e @Is1 Formation is due to low sample resolution, a s it is in the case of the North Sea Balder Formation (M. Charnock, pers. comm., 1994). The very low species diversity points to an extreme palaeoenvironment. The dominance of Evoliitinella sp. 2 may be due to a high acid content in the water column caused by frequent volcanism (ash-layers, Fig. 3b). Such a situation has been reported by Finger & Lipps (1981) from an active Antarctic volcanic caldera: faunas at a water depth of 100-150 m were dominated herc by Trochammina species for several years after an eruption.
Pyrite and gypsum are present in the sediment and indicate depleted oxygen conditions at the sea floor. 'The sediment contains more particles of the silt fraction than in the underlying Lones. This may be interpreted as an indication of a slight decrease in water depth, but the silt fraction also includes fine-grained ash particles (J. Eiriksson, pers. comm., 1993).

Zone 5: The Eponides plummerae-Neoeponides karsteni Zone
This zone identified in 4 samples (136-133 mbs) is developed in the red clays of thc Rgsnas Clay Formation (Fig. 3b). The lowermost sample in the zone has a high content of glauconite but no foraminifera are present.
The glauconite layer observed at the basc of the zone presumably indicates a brief pause in sedimentation. The succeeding sediment is red, carbonate rich, plastic clay pointing to a well oxygenated water mass. The genus composition (Bulimina, Cihicidoide.s and Eponides) indicates an outer shelf to upper bathyal palaeoenvironment (Murray, 1991). This interpretation is supported by the diverse fauna and the increasing amount of planktic foraminifera, which indicate an increasing water depth.

CORRELATION AND AGE Zone 1 : Danian? to Selandian
The presence of Lenticuliria platypleura, Anomalinoirles ruhiginosus, and Bulimina midwayensis suggest correlation of Zone 1 with the North Sea Benthic foraminiferal zonc NSB l b of King (1989). The upper boundary of Zone 1 is sharply defined as calcareous taxa are absent in the overlying interval. This phenomenon was described by King (1989) at the upper boundary o f Zone NSB lb. An equivalent of King's NSB Ic Subzone was not identified. In addition, the co-occurrence of Eponides lunatus and Lenticulina platypleura ( L . niultLformis of Doppert & Neele, 1983) implies correlation of Zone 1 with Zone FJ from the Netherlands (Doppert & Neele, 1983).

Zone 2: Late Selandian to Thanetian
The presence of Spiroplectunimina spectahilis, which has its last appearance at the top of Zone 2, and Rzehakitirr epigona. Rec~irvoides ex. gr. turhiriutus. Buthysiphon ex. gr. discreta. Ammodiscus crrtaceoiis and Karrc~rulina c o n u e r s~~ indicate a correlation of Zone 2 with Zone NSA 1 (North Sea Agglutinated) of King (l98Y). The Holmehus Formation has previously been assigned to Zone NSP 3 and Subzone NSB Ic equivalents of NSA Ib by King (1989, compare figs 9.3 and 9.9). The fact that Trochammina ruthvmmitrrayi is only found in the lowermost part of the zone is in opposition to the observations of King (1989) but in agreement with those of Gradstein 6 " ul. (1994). Bovlstrup Zone 2 corresponds to the majority of the Trochammina riithvennir*rrayi-lir~ticiilol,hragmiiim l~aupr~ra Zone of Gradstein ?( al. (1992Gradstein ?( al. ( , 1994 (1994). An influx of agglutinated specics seen at the transition between Zones FJ and FI in the Dutch area (Doppert & Neele, 1983) also infers correlation to Zone 2.
The boundary between the informal Grey Clay litho-unit and the Holmehus Formation is placed at 202m well depth as cuttings from both litho-units occur in the sample at this level (Heilmann-Clausen, pers. comm.. 1988). This boundary occurs within Zone 2 in the Bovlstrup well. Trochammina riithuenmurrayi is found below the lithological boundary and Cenodiscus cf. sp. T3 above.
A sample 3 m above the lower boundary of Zone 2 contained a Viborg Zone 3 dinoflagellate flora, and two samples from the uppermost part of Zone 2 possessed Viborg Zone 4 floras (Heilmann-Clausen, 1985: pers. comm., 19x8). This implies that Zone 2 corresponds to both a part of dinoflagellate Zone 3 and probably most of Zone 4 of Viborg (Heilmann-Clausen, 1985) and to Zone NSA Ib (King, 1989). Therefore, Zone 2 of Bovlstrup can probably be referred to nannoplankton Zones NP 6-NP 8 (see also King. 1989), indicating a possible Late Selandian to Thanetian age which also is in agreement with the correlations of Gradstein et a/. (1994).
The boundary between the Selandian and Thanetian (Fig.  3b) is tentatively placed half way between the sample yielding the Viborg Zone 3 dinoflagellate flora and the lowermost sample referred to dinoflagellate Zone 4.

Zone 3: Thanetian
The presence of Vernruilinoitlrs subrocaenus in Zone 3 and the sparse, generally poorly preserved fauna correspond to the faunal characteristics of lower part of Zone NSA 2 (King, 1989) although King did not recognize this assemblage as distinct. The NSA 2 Zone of King is correlated to his NSP 4 Zone indicated by the presence of Co.scin<xliscirs sp. 1 and Coscinodiscirs sp. 2. In Bovlstrup Zone ?#, however, none of the characteristic diatoms of NSP 4 were found, but they are present in the above lying Zone 4. For this reason Bovlstrup Zone 3 is correlated with the lower part of NSA 2. The present features were probably only clbserv'ed due to a close sampling interval. In a commercial oil well, where the sampling interval is commcinly 1 0 m , a Zone 3 equivalent might easily escape sampling, and Zones 3 and 4 would accordingly be registered as one.
Gradstein ef al. (1994, p. 37) state that the NSA 2 Zone of King (1989) is found only in the southern part of the North Sea. They find the Coscinodiscus assemblage but not the Vrrnrii/inoidi~.s .subeocuenu.s assemblage. As Zone 3 (characterized by Vrmuilinoirles suheocueniis) is represented only in a 2 m int'erval it is easy to understand why it was not found in exploration wells in the northern North Sea.
The top of Bovlstrup Zone 3 probably corresponds to bioevent M5 of Mudge & Copestake (1992) and O'Connor & Walker (1993). These authors state that the event is isochronous over large parts of the basin as it probably reflects a change in the bottom water conditions.
Unfortunately no dinoflagellate analysis were carried out in Zen'? 3 and the chronostratigraphic allocation of the zone is thus uncertain. Bovlstrup Zone 3 falls between samples dated )using dinoflagellates to be Early Thanetian (Viborg Zone 4, Heilmann-Clausen, pers. comm.) and samples dated to Latr. Thanetian to Early Ypresian. We therefore suggest Zone 3 to be of Thanetian age.

Zone 4: Late Thanetian to Early Ypresian
Thc faunal composition of Zone 4 (almost exclusively Evolutiric4la ip. 2 and Vrrneuilinoides sp. 1) combined with the presence of pyritized and siliceous diatoms (Coscinodiscus sp. I. Coscinodiscus sp. 2, Triceratium sp. 1, Coscinodiscics sp. 7 and Coscinodisciis sp. 11) are rather unique features and have not previously been registered in Denmark .
Evolutinella sp. 2 is found only in a very narrow time interval in the Late Palaeocene to Early Eocene (Charnock & Jones, 1990).
Coscinodix-us sp. I and other diatoms occur frequently in the lowermost part of the FI Zone of the Netherlands (Doppert & Neele, 1983), hut the foraminiferal fauna in this interval shows no resemblance to the Bovlstrup Zone 4 fauna. Zone 4 is, nevertheless, tentatively correlated to the lowermost part of the FI Zone, to the Coscinodiscus Zone of Gradstein et a/. (1992Gradstein et a/. ( , 1994, and to the NSP 4 Zone of King (1989) because of the presence of Coscinodiscus sp. 1.
Comparable foraminiferal faunas have been found in the Untereozan 1 of Germany (Wick, 1943;Bettenstaedt et al., 1962) and in the basal Ieper Formation of Belgium (Willems, 1983).
Dinoflagellates were studied in three samples from Zone 4 (156, 154, and 152mbs). The flora placed the samples in Zone 7 of Viborg and thus with the upper part of the 0lst Formation (Heilmann-Clausen, 1985, p. 30) indicating a Late Thanetian to Early Ypresian age. The dinoflagellates suggest a tentative correlation of Bovlstrup Zone 4 with nannoplankton Zones N P 9-10 of Martini (1971) (see also Gradstein et a/., 1994, p. 37). The formal boundary between the Palaeocene and the Eocene is not yet internationally agreed upon, but it is normally placed at, or close to the N P 9-10 boundary.
The lithological boundary between the 81st Formation and the R@snxs Clay Formation has been placed between 136 and 137 m below the surface (Heilmann-Clausen pers. comm., 1988).

Zone 5: Ypresian
The presence of Gaudryinu hiltermani, Pseudoclavulina anglicu and Turrilina breuispira in Bovlstrup Zone 5 justifies a correlation with Zone NSB 3a of King (1989). This is further supported by the planktic fauna of the sample at 133 m, which is totally dominated by Subhotina ex. gr. linaperta. Zone 5 is consequently placed in Zone NSP 5a (King, 1989). Furthermore, Zone 5 equates with part of the Subbotina patagonica Zone of Gradstein et al. (1992Gradstein et al. ( , 1994. The dominance of Subbotinii ex. gr. linaperta and the presence of Pseudohasterigrrina wilcoxensis indicate a correlation with P6-P8 of the standard planktic foraminiferal zonation of Blow (1979) (see also Gradstein eta/., 1994). The top of the acme of Suhborina ex. gr. linuperta is correlated to a position in P8 by King (1989) and Mudge & Bujak (1994). Bovlstrup Zone 5 may therefore correspond to NP 11-12. This is based on the correlation with NSP 5a (see also King, 1983) and on the fact that the Rasnas Clay Formation has previously been correlated to NP 11 -12 (Heilmann-Clausen, 1989). Zone 5 is consequently of Ypresian age.

SUMMARY AND DISCUSSION
The Bovlstrup well provides new information on foraminiferal faunas and palaeoenvironments of upper Palaeocene and lower Eocene deposits in Denmark. Especially significant is that whereas no foraminiferal faunas have previously been encountered in the 0lst Formation, this interval in the Bovlstrup well contains a remarkable low-diversity agglutinated fauna.

Stratigraphy
The five foraminiferal zones at Bovlstrup are readily referred to the established North Sea zonation of King (1983,1989). Relatively dense sampling (maximum 4 m between the samples), however, yielded information that may be lost in commercial oil wells.
Zone 1 is placed in Subzone NSB Ib of King (1989) and a Danian to Selandian age is suggested. Fluctuations in the faunal diversity in Zone 2 (Fig. 3b) formed the basis for a division into 3 subzones. These are correlated with NSA 1 Zone of King ( 1989) King (1989) and an Ypresian age for this m n c is suggested.

Litho-units
The foraminiferal zones of the Bovlstrup well occur in the following litho-units ( Fig. 3b) al. (1988) and the base of sequence 1.2 of Michelsen at ul. (in press).
The increasing water depth from the shelf environment in Bovlstrup Zones 3 and 4 to the outer shelf t o upper bathyal as indicated by thc increasing amount o f planktic foraminifera in Zone 5 , may correspond to TA 2.5-2.6 of Haq et al. (1988) and sequence 2 of Michelsen el al. (in press) The benthic Caunas also indicate radical changes in the bottom water conditions. We interpret the transition from the calcareous fauna of Zone 1 to an exclusively agglutinated fauna in Zone 2 as the result of a change to a slightly acidic environment at the sea floor. The low diversity of the fauna in Zone 3 indicates that bottom water conditions may have deteriorated. The volcanic ash layers o f Zone 4 presumably resulted in low pH values. leading to benthic faunas with extremely low diversities. This may correspond to the short-term negative excursion in both carbon and oxygcn isotopcs observed in DSDP wells worldwide at the Palaeocene-Eocene boundary (Pak dz Miller. 1992, p. 419). as the C ' 0 2 from the volcanism may have caused polar warming. This event led to the largest deep-water benthic foraminiferal turnover of the Cenozoic. After a pause or reduction in sedimentation, indicated by a fossil-free, glauconite-rich sample at the boundary between Bovlstrup Zones 4 and 5 , the oceanographic setting changed dramatically to one of improved water circulation with oxygenated bottom waters