The 13C Chemostratigraphy of the Upper

NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 65

The 13C chemostratigraphy of the Upper Ordovician Mjøsa Formation at Furuberget near Hamar, southeastern Norway: Baltic, Trans-Atlantic, and Chinese relations

Stig M. Bergström, Birger Schmitz, Seth A. Young & David L. Bruton

Bergström, S.M., Schmitz, B., Young, S.A. & Bruton, D.L.: The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation at Furuberget near Hamar, southeastern Norway: Baltic, Trans-Atlantic, and Chinese relations. Norwegian Journal of Geology, Vol 90, pp. 65-78. Trondheim 2010, ISSN 029-196X.

Whereas no studies have previously been carried out on the δ13C chemostratigraphy of the Sandbian-Katian (Upper Ordovician) succession any- where in Norway, such investigations in Sweden and the East Baltic region have made the δ13C chemostratigraphy of that interval well known. In an attempt to document for the first time the presence of the globally distributed and stratigraphically important Guttenberg Carbon Isotope Excursion (GICE), 45 samples were collected at 2 m intervals through the Furnesfjorden and Gålås members of the Mjøsa Formation from two sections at Furuberget in the Nes-Hamar District 70 km north of the City of Oslo. Relatively high δ13C values, which are interpreted to represent the GICE, were obtained throughout the Gålås Member. The presence of this δ13C excursion in the middle part of the Mjøsa Formation, combined with biostratigraphic and other evidence, are used for detailed correlations of the Mjøsa Formation with other successions in Baltoscandia, North Ame- rica, and on the Yangtze Platform in southern China.

Stig M. Bergström, School of Earth Sciences, Division of Geological Sciences, The Ohio State University, Columbus, Ohio 43210, USA. E-mail: stig@ geology.ohio-state.edu. Birger Schmitz, GeoBiosphere Science Centre, Department of Geology, Lund University, Sölvegatan 12, Lund, SE-223 62 Lund, Sweden. E-mail: [email protected]. Seth A. Young, Department of Geological Sciences, Indiana University, Bloomington, Indiana 47405, USA. E-mail: [email protected]. David L. Bruton, Naturhistorisk Museum, University of Oslo, Sars gate 1, Oslo, Norway. E-mail: [email protected]

Introduction unique, position among the many Ordovician major stratigraphic units in the Oslo region. This prominent Although a large amount of work has been carried out formation, which reaches a thickness of approximately in recent years on the Sandbian and Katian δ13C che- 100 m or slightly more in many sections, is widely dis- mostratigraphy in the East Baltic region and Sweden (e.g. tributed round the northern part of Lake Mjøsa (Opal- Saltzman et al. 2003; Tobin et al. 2005; Kaljo et al. 2007; inski & Harland 1981) about 70 km north of the City Schmitz & Bergström 2007; Calner et al. 2010; Bergström of Oslo (Fig. 1). The Mjøsa Formation consists mainly et al. in press), no chemostratigraphic data have previously of relatively pure shallow-water limestones of baham- been published from the coeval successions in the Oslo itic type (Jaanusson 1973) containing locally Solenopora Region, or elsewhere in Norway. Indeed, the only δ13C bioherms. Its rather diverse, but unevenly distributed, chemostratigraphic information currently available from fossil fauna, part of which is still undescribed, includes the Ordovician of Norway is some data from the Hirnan- brachiopods, trilobites, corals and other shelly fossils tian succession in the Oslo area (e.g. Brenchley et al. 1997; along with stromatoporoids, algae, and microfossils. The Brenchley & Marshall 1999; Kaljo et al. 2004b; Bergström microfossils include relatively common ostracodes and et al. 2006). The present study is the first attempt to intro- conodonts, two groups that have not yet been subjected duce δ13C chemostratigraphy to a pre-Hirnantian forma- to detailed published studies. For faunal references, see tion in the classic Oslo Region Ordovician succession. Owen et al. (1990). As noted in several recent papers, since the 1990s, δ13C chemostratigraphy has become a very important tool for A significant number of the fossils present in the Mjøsa clarifying stratigraphic relationships at both the local and Formation differ markedly from those in most con- global scale. However, Ordovician chemostratigraphic temporaneous Baltoscandian faunas but have affinities studies are still in the pioneer stage with no such work yet to those in warm-water deposits in Laurentia (Jaanus- having been carried out in many, if not most, key succes- son 1979; Bergström 1997; Bergström et al. 1998). This sions round the world as is the case also in Norway. applies particularly to the conodont fauna that differs strikingly from that present in most coeval Baltoscan- In terms of lithology, fauna, and also economic impor- dic deposits but shows a remarkable similarity to that in tance, the Mjøsa Formation occupies a special, if not some formations in the North American Midcontinent, 66 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY

Fig. 1. Sketch-map of parts of the Nes-Hamar-Toten districts showing the location of the Furuberget Quarry northwest of Hamar (modified after Opalinski & Har- land 1981, fig. 2). Inset map (after Owen et al. 1990) illus- trates the distribution of Ord- ovician rocks (black) in the several districts in the Oslo region distinguished by Stør- mer (1953). Arrow indicates the location of the study area in the Lake Mjøsa region.

such as the Lexington Limestone of Kentucky and coeval Spjeldnæs 1982) but the quarry succession, which is now strata in adjacent states (Bergström & Sweet 1966; Rich- exposed more extensively stratigraphically than in the ardson & Bergström 2003). past due to recent quarrying, has never been described in detail. The lower-middle part of the Lexington Limestone dis- plays the chemostratigraphically important Guttenberg δ13C excursion (GICE), one of the seven named Ordo- Distribution and geology of the Mjøsa vician δ13C excursions (Bergström et al. 2009a). The Formation GICE has been recorded at many localities in Lauren- tia (e.g. Ludvigson et al. 2004; Young et al. 2005; Barta In the region round the City of Oslo Ordovician rocks et al. 2007), the East Baltic region (Kaljo et al. 2007) and occur in a number of geographically more or less sepa- Sweden (e.g. Saltzman et al. 2003; Bergström et al. 2004) rated areas, some of which have their own lihologically and recently also in China (Bergström et al. 2009b). The and faunally specific Ordovician succession. The latter close similarity between the conodont faunas from the applies to our study area in the Lake Mjøsa region. For Lexington Limestone and the Mjøsa Formation, and the convenience, when discussing various parts of the Oslo fact that the latter is a shallow-water carbonate unit of Region below, we follow the standard district designa- the type that tends to preserve particularly well perturba- tions introduced by Størmer (1953) and subsequently tions in the marine carbon cycle, suggested to us that this used by many authors (Fig. 1). unit would be ideal for an attempt to establish the presence of the GICE for the first time in Norway. The pur- Occurrence and lithostratigraphy pose of this report is to document the δ13C chemostratigraphy through most of the Mjøsa Formation, and to The Mjøsa Formation is widely distributed in the Toten- describe the discovery of the GICE, in the well-known Nes-Hamar districts and can be studied in numerous, in outcrops at the Furuberget Quarry about 4 km northwest most cases now inactive, quarries as well as in road cuts of the centre of Hamar (Fig. 1). These outcrops have been and natural outcrops. Because of its resistant nature, it discussed repeatedly in the literature (see, e.g. Holte- tends to be better exposed than underlying shaly strata dahl 1909; Skjeseth 1963; Opalinski & Harland 1981; of the Furuberget Formation. The abundance of out- NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 67 crops is illustrated by the fact that Opalinski & Harland The middle portion of the formation was classified into (1981) listed 49 selected localities of the unit and several three, areally more restricted but at least broadly coeval, more are recorded in the literature. Although a thickness members, namely the Eina, Gålås, and Bergevika mem- figure of approximately 100 m has generally been given bers. For a summary of the characteristics of these mem- for the Mjøsa Formation, it should be noted that there is bers, see Owen et al. (1990). Although the Lower Paleo- no section described in which the entire unit and its top zoic sediments in the study area have been subjected to and basal contacts are well exposed. This also applies to considerable folding, faulting, and heating (approxi- its type section at the Eina Quarry in the Toten district, mately 300°C based on Conodont Alteration Index; cf. where most of the unit (85 m) and its basal contact are Bergström 1980), the prevailing opinion is that they are well exposed but where its uppermost portion and the autochthonous, at least in the Nes-Hamar area (Skjeseth top contact are covered (Opalinski & Harland 1981). 1963). For some different interpretations, see Spjeldnæs (1982). Hence, it is very unlikely that the studied succes- As just noted, the Mjøsa Formation rests, probably sion is located on a detached piece of Laurentia. conformably, on the Furuberget Formation (Fig. 2), the lithology of which is dominated by shales and siltstones Biostratigraphy and with subordinate limestones. The top of the Mjøsa Formation is a prominent karst surface with locally up No graptolites have been recorded from the Mjøsa For- to 3 m deep fissures filled by quartzitic material belong- mation and its shelly macrofossils have not provided ing to the overlying middle Llandovery Helgøya Member any very decisive evidence of the age of the unit. Early of the Sælabonn Formation (Strand & Henningsmoen students (Kiær 1897; Holtedahl 1909) referred the unit 1960; Skjeseth 1963; Worsley et al. 1983). The significant to the uppermost Ordovician (stage 5b; now Hirnan- stratigraphic gap separating these units includes the mid- tian), but this interpretation was challenged by Ray- dle and upper Katian Stage as well as the Hirnantian and mond (1916), who correlated the interval now known Rhuddanian stages (e.g. Dahlqvist & Bergström 2005). as the Mjøsa and Furuberget formations with a much For a detailed description of the lithologies of the Mjøsa older interval in the Oslo-Asker succession, namely that Formation, see Opalinski & Harland (1981) and Spjeld- between the Solvang (stage 4bδ) and Venstøp (stage næs (1982). Its bahamitic lithology and the common 4cα) formations. This correlation has subsequently been occurrence of cross-bedding, ripple marks, and bioherms followed by many authors (e.g. Størmer 1953; Skjes- formed by tabulate corals, Solenopora, and stromato eth 1963; Spjeldnæs 1982; Owen et al. 1990). Nõlvak & poroids, along with the local presence of desiccation Grahn (1993) suggested that the Mjøsa Formation ranges cracks and karst surfaces within the formation, indicate as high as the base of the uppermost Katian Pirgu Stage that most, if not all, of the Mjøsa Formation was depos- but they did not present any faunal evidence in support ited in shallow water. of this novel interpretation.

Opalinski & Harland (1981) subdivided the formation In a major, but unfortunately never completed, project into five named members (Fig. 2), two of which, the during the 1960s, György Hamar assembled a collection basal Furnesfjorden Member and the topmost Holtjern of more than 15 000 mostly fragmentary as well as ther- Member, are distributed across the Lake Mjøsa region. mally altered conodont elements from the Mjøsa and Furuberget formations at 16 localities in the Lake Mjøsa region. These collections, which have been kept at the Geologisk Museum (formerly Paleontologisk Museum) in Oslo, are now under restudy by one of us (SMB) and have been briefly referred to in several papers (e.g. Berg- ström 1997; Bergström et al. 1998), in which their strong Laurentian, rather than Baltic, faunal affinities have been noted. In Europe, this type of conodont fauna is previously known only from the Svartsætra limestone in the Trondheim region, central Norway (Bergström 1997). As noted in the latter paper, these conodont faunas lack representatives of Baltoniodus, Dapsilodus, Hamarodus, Peri- odon, Protopanderodus, and Scabbardella, which are the common taxa in the coeval Baltoscandic faunas. Instead, these Norwegian faunas are dominated by representatives of Aphelognathus, Oulodus, Phragmodus, and Plectodina. Fig. 2. Stratigraphic classification of the Upper Ordovician In the case of the Mjøsa and Furuberget formations, these succession in the Nes-Hamar-Toten districts. Note that faunas also contain less common elements of, for instance, rocks of middle-late Katian, Hirnantian, and Rhuddanian hyaline species, Polyplacognathus ramosus, Staufferella fal- age are missing at the unconformity below the base of the cata, and Yaoxianognathus abruptus, which are taxa that Sælabonn Formation. are otherwise unknown in Europe, or in the case of hya- 68 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY line forms only known from a few specimens collected in of important conodont species in the Mjøsa and Furu- Wales (Bergström 1964, 1971). The conodont faunas of berget formations and the Lexington Limestone and the the Mjøsa and Furuberget formations exhibit a particu- Tyrone Formation. The Norwegian data are based on 65 larly remarkable resemblance to the latest Sandbian-early samples containing a total of approximately 5800 cono Katian conodont faunas of the Ohio Valley Province of the dont elements. These were collected by György Hamar Midcontinent Realm (Sweet & Bergström 1984), such as from two sections at Furuberget. One of these includes those from the Lexington Limestone and Tyrone Forma- the upper 40 m of the Furuberget Formation and the tion of Kentucky (Bergström & Sweet 1966; Sweet 1979; lower 52 m of the Mjøsa Formation exposed in the rail- Richardson & Bergström 2003). way section south of the tunnel. The other section is a 20 m thick interval in the Furuberget quarry, the top Figure 3 shows a comparison between the vertical ranges

Fig. 3. Comparison of ranges of virtually all conodont species in the Mjøsa Formation and uppermost Furuberget Forma- tion (black bars) and the lower part of the Lexington Limestone and the upper Tyrone Formation of Kentucky (open bars). The Norwegian data are based on collections made by György Hamar from the railway and quarry exposures at Furu- berget. Note that there is a more than 25 m thick unsampled interval between these outcrops. The data of the Kentucky composite section are from the Camp Nelson exposures (Tyrone Formation; cf. Bergström 2002), the Frankfort East section (Curdsville and Logana members; cf. Bergström & Sweet 1966; Richardson & Bergström 2003), and the Minerva drill-core (lower Grier Member; Sweet 1979). Based on conodont data, the Mjøsa Formation is referable to the North American Chatfieldian Stage and its base is equivalent to the base of the Lexington Limestone. NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 69 of which is approximately 20 m below the top of the Mjøsa Formation. We interpret this collected interval as corresponding to the upper 20 m of the Gålås Mem- ber. Most of the conodont taxa in the study interval were readily identified, especially as the Norwegian specimens could be compared directly with specimens from Ken- tucky. However, as noted by Sweet (1981), the taxonomy of some of the morphotypes of Aphelognathus and Oulo- dus, which show considerable morphological variation, is not yet well understood. In the present account, we have adopted a rather broad species concept when deal- ing with the few taxa of these genera in the collections at hand. Of special biostratigraphic significance are the ranges of Plectodina tenuis, Phragmodus flexuosus, and Phragmodus undatus that indicate that most, if not all, of the Mjøsa Formation is referable to the Phragmodus undatus and Plectodina tenuis North American Mid- continent conodont zones of Sweet (1984). The superja- cent Belodina confluens Zone might possibly be present in the uppermost part of the Mjøsa Formation but the diagnostic zone index has not yet been found in the unit. Also, the data at hand indicate that the upper 40 m of the Furuberget Formation at its type locality (Owen et al. 1990) in the Furuberget railway cut represents the Belo- dina compressa Zone of Sweet (1984). Interestingly, the conodont data suggest that the boundaries between the Tyrone/Lexington and Furuberget/Mjøsa formations are equivalent. We conclude that based on the conodont biostratigraphy, the Mjøsa Formation represents an interval Fig. 4. A. Sketch-map of the Furuberget Quarry and adja- corresponding to at least the lower and middle parts of cent area showing the location of the Entrance Road sec- the North American Chatfieldian Stage. Paradoxically, tion and the position of the collected Quarry section (map the biostratigraphic range of the Mjøsa Formation in reference 091441; 1:50 000, Series M711, Sheet 1961, IV terms of the Baltoscandic regional stage succession is edition 1975). Extensive outcrops of the Furuberget and much more difficult to establish based on the conodont Mjøsa formations occur also along the railway near the collections at hand. shore of Lake Mjøsa (Holtedahl 1909). B. Schematic North-South cross section of Furuberget (modified after 13 Spjeldnæs 1982) illustrating the major synclinal structure  C samples and laboratory methods and the locations of the study sections along the northern and southern limbs of the syncline. Study sections

Among the stratigraphically nearly complete sections of the Mjøsa Formation recorded from the region round Lake Mjøsa, we selected those exposed at Furuberget tion along the east side of the quarry entrance road just approximately 4 km northwest of Hamar for our collect- north of a red house about 150 m north of the quarry ing of chemostratigraphic samples. At this locality there entrance (Fig. 4). Starting at the level of the lithologi- are excellent exposures of the Mjøsa Formation and the cally not conspicuous formation contact, we collected 15 underlying Furuberget Formation along the railroad near samples at 2 m stratigraphic intervals from the succes the shore of Lake Mjøsa, along the entrance road to the sion of south-dipping well-bedded limestones that ends Furuberget quarry, and inside this large quarry (Fig. 4). at a forest track that runs up the hillside just north of As has been illustrated by several authors (e.g. Spjeldnæs the red house. We observed no evidence of significant 1982), the exposures along the entrance road as well as in faulting in this exposure of the Mjøsa Formation but the northernmost part of the quarry are in the northern the underlying shaly strata of the Furuberget Formation limb of a synclinal structure with an east-west oriented are in part strongly tectonized. In terms of structural axis, whereas those in most of the quarry, especially in its complexity, they compare well with the coeval beds in recently excavated eastern part, are located in the south- the railroad section to the west that was strikingly illu ern limb of that structure (Fig. 4). The upper Furuberget strated by Holtedahl (1909, fig. 8). Using the lithostratig- Formation as well as the lower Mjøsa Formation, and raphy proposed by Opalinski & Harland (1981), the col- the formation contact, are exposed in a continuous sec- lected 28 m thick sampled succession is referable to the 70 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY

Furnesfjorden Member, the basal member of the Mjøsa 13 18 Formation. Table 1. C and O sample values from the two study sections at Furuberget. Standard The other, and significantly thicker, section was sampled deviation is 0.09 and 0.23, respectively. along the eastern and northeastern walls of the eastern- Entrance road section most, recently quarried, part of the quarry. We collected (measured from the base of the Mjøsa Formation) this section, which is just east and north of a water-filled δ13 δ18 pit, starting from the lowest exposed north-dipping bed M C C at the south-east end of the pit and continuing up to the 0 2.4 -6.75 base of a lithologically conspicuous succession of silt- 2 2.22 -9.06 stones, which is interpreted to represent the Holtjern 4 1.28 -9.63 Member of Opalinski & Harland (1981), high up on the 6 1.60 -7.26 north wall of the quarry. A single sample, which proved 8 1.64 -4.87 13 to contain too little carbonate to yield a δ C value, was 10 1.17 -8.61 collected from the base of the latter member, which has 12 1.01 -9.14 an estimated thickness of about 20 m, but because of the difficulty and physical risk involved in climbing higher 14 1.01 -9.62 parts of the steep slope without body harness and rope, 16 0.99 -8.59 no further sampling was done from that member. A total 18 0.93 -8.90 of 30 samples were collected at 2 m intervals from this 20 1.77 -6.12 section, which is a lithologically relatively uniform and 22 1.09 -8.75 apparently largely, if not entirely, unfaulted succession 24 1.44 -8.60 of poorly fossiliferous limestones. This entire succes- 26 1.35 -8.97 sion appears to represent the Gålås Member of Opal- 28 1.64 -8.86 inski & Harland (1981). Because the thickness of this member at Furuberget given by the latter authors is Quarry section (measured from the oldest strata exposed in the quarry wall to the base of the Holtjern Member) 36.5 m, we expected that our sampled succession would overlap at least 10 m with the entrance road section. 0 2.64 -3.14 However, as shown below, the δ13C chemostratigraphy 2 1.56 -7.05 strongly suggests that this is not the case; rather there 4 2.46 -6.00 is a stratigraphic gap, presumably of no more than a 6 1.85 -5.97 few m, between the collected successions. Based on our 8 2.30 -6.62 measurements, and the thickness figure of the Holtjern 10 1.94 -6.42 Member (25.9 m) given by Opalkinski & Harland (1981), 12 1.93 -7.83 the thickness of the Mjøsa Formation at Furuberget would amount to more than 111.9 m, which is signifi- 14 1.85 -7.68 cantly larger than the thickness commonly cited for the 16 1.89 -7.62 formation (100 m). 18 2.33 -7.55 20 1.60 -7.51 Laboratory treatment of samples 22 1.63 -7.78 24 1.83 -7.77 Powdered, homogenized bulk-rock samples were ana- 26 2.56 -7.27 lyzed for 13C/12C and 18O/16O ratios at the Stable 28 2.18 -7.35 Isotope Laboratory, Department of Geology, Copen 30 2.33 -7.57 hagen University. The samples were dissolved in vacuum in 100% phosphoric acid at 25 °C. The carbon dioxide 32 2.43 -6.72 evolved was analyzed in a Finnigan-MAT 250 mass- 34 2.07 -7.93 spectrometer. The results are reported in per mil (‰) 36 1.84 -8.19 deviations from the V-PDB (Vienna-Pee Dee Belemnite) 38 2.13 -8.23 13 standard. Reproducibility is better than ±0.03 ‰ for δ C 40 2.34 -8.11 18 and ±0.05‰ for δ O expressed as ± σ (standard devia- 42 1.72 -8.68 13 tion) for 10 identical samples. Only the δ C results of 44 1.68 -7.30 the 45 samples studied are dealt with herein and listed in 46 1.81 -8.83 Table 1. 48 1.76 -8.31 50 1.75 -6.89 52 1.23 -8.70 54 1.40 -9.06 56 0.94 -7.16 NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 71

13C chemostratigraphy

The δ13C curve of the Mjøsa Formation

A composite δ13C curve through most of the Mjøsa Formation based on the samples from the two sampled sections is illustrated in Figure 5. The samples from the lowermost two meters of the Mjøsa Formation show unusually elevated δ13C values in comparison with those from presumably coeval strata elsewhere (e.g. Young et al. 2005) but it is appropriate to note that these values have a significantly lower laboratory precision than is the case with most other samples. Regardless of the possible significance of this, the δ13C baseline values of the lower part of the Furnesfjorden Member appears to be ~+1‰. There is an indication of a small increase in δ13C values in the uppermost portion of the sampled entrance road succession that could represent a beginning of a δ13C excursion but its possible significance requires further study.

The lowermost sample of the quarry section has a markedly higher δ13C value of +2.64‰, which in magnitude is higher than that of the GICE recorded in Swedish (Saltzman et al. 2003; Bergström et al. 2009b; Barta et al. 2007), Estonian (Kaljo et al. 2007), and most North American (Young et al. 2005) successions. Because this high δ13C value, and similar high values obtained from the overlying interval of the Gålås Member up to at least the 40 m level, are of GICE magnitude (most are between +2‰ and +2.5‰), combined with that fact that accord- Fig. 5. The δ13C curves through the Mjøsa Formation at ing to conodont biostratigraphy they occur in the same the Entrance Road and Quarry sections. Note that the δ13C stratigraphic interval as the GICE in North America, chemostratigraphy suggests that these sections do not over- we identify this excursion as the GICE. Above the 40 m lap stratigraphically; rather, they appear to be separated by level, there is a somewhat gradual fall in the δ13C curve an unstudied, presumably quite thin, stratigraphic inter- that reaches a baseline value of <+1‰ at the top of the val. The figure also shows that the GICE is restricted to the Gålås Member. We interpret this decline in the δ13C val- sampled quarry succession, which is interpreted to repre- ues as marking the end of the GICE, which appears to sent the Gålås Member of Opalinski & Harland (1981). have a stratigraphical thickness of >50 m. The fact that in most other Baltoscandic sections, the thickness of the GICE interval is <10 m shows that the rate of net deposition of the Mjøsa Formation was several times higher than that in the more central portions of the Baltoscan- dic craton. Also in North America and China, the thickness of the GICE interval is at most localities <20 m. We the recorded thickness of the Mjøsa Formation (approxi- know only one succession, namely in the Tarim Basin, mately 100 m) elsewhere in the Lake Mjøsa region. China (Bergström et al. 2009b), where the GICE interval appears to be significantly thicker than in the Mjøsa Formation but the biostratigraphic dating of the Tarim Regional comparison sequence is still tentative and more than one δ13C excursion may be involved. As noted above, based on conodont biostratigraphy, it is far easier to establish equivalents to the Mjøsa Formation As noted above, the rather conspicuous difference in the North American Midcontinent than in the Swed- between the δ13C values at the top of the entrance road ish and East Baltic successions. We will therefore start section and the bottom of the quarry section suggests our review of the regional stratigraphical relationships that there is no stratigraphic overlap, but rather a strati- of the Mjøsa Formation with a comparison with the Lex- graphic gap, between these sections. At the present ington Limestone of Kentucky. This is followed by sepa- time, the size of this gap remains undetermined but it is rate discussions of the relationships to key successions in unlikely to amount to more than a few meters in view of Sweden, Estonia, China, and the Oslo Region, Norway. 72 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY

Fig. 6. Comparison of the δ13C curves of the Mjøsa Formation and the coeval part of the Lexington Limestone in Kentucky. The latter curve is a composite based on δ13C values from the Frankfort East section (Curdsville and Logana members) and the Minerva drill-core (Grier Member) (after Young et al. 2005, 2008). Note the similarity between these δ13C curves despite the fact that the GICE is approximately five times thicker in the Mjøsa succession than in the Lexington Limestone. Also note the close correlation between the lower members of the two formations. The geographically widespread M4/M5 sequence boundary of Holland & Patzkowsky (1996) is located at the base of the Lexington Limestone. In stratigraphically nearly complete successions, the prominent Millbrig K-bentonite is present at, or very near, the top of the Tyrone Limes- tone.

North America be taken as a third- order sequence boundary in such a classification but whether or not it is associated with a As shown in Figures 3 and 6, both in terms of biostratig- notable stratigraphic gap is currently unknown. How- raphy and δ13C chemostratigraphy, and also lithostratig- ever, there is a distinct change in the conodont fauna at raphy, there are surprising similarities between the Lex- the base of the Mjøsa Formation at Furuberget (Fig. 3). ington Limestone and the Mjøsa Formation despite the fact that these units are likely to have been widely sepa- The interval between the base of the Lexington Lime- rated geographically during their deposition in Late stone and the base of the Mjøsa Formation and the level Ordovician time. of the beginning of the GICE is characterized by baseline δ13C values. This interval is referred to as the Curdsville The base of the Lexington Limestone is an unconformity, Member in Kentucky and the Furnesfjorden Member in which is marked by a distinct facies change and a minor the Lake Mjøsa region. The vertical scale in the strati- stratigraphic gap. This level represents the geographically graphic column in Figure 6 is based on the reasonable widespread third-order M4/M5 sequence boundary of assumption that the GICE had the same geochronologi- Holland & Patzkowsky (1996). The basal contact of the cal duration globally. Interestingly, as illustrated in this Mjøsa Formation is not as clearly marked lithologically figure, this results in that also the Curdsville and Furnes- as that of the Lexington Limestone but it certainly repre- fjorden members were deposited during approximately sents a general change in deposition from dominant silic- the same time span although they differ substantially in iclastics to dominant carbonates. This level could well thickness. Hence, it appears as if these members are close NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 73 equivalents. Further, the range of the GICE coincides Sweden, Estonia, and Latvia with that of the Logana Member in Kentucky and that During the last decade, a substantial amount of che- of the Gålås Member in the Lake Mjøsa region, which mostratigraphic work involving the GICE has been car- indicates that also these members are to be regarded as ried out in Sweden (e.g. Bergström et al. 2001; Saltzman coeval. The Logana Member is generally considered to et al. 2001, 2003; Bergström et al. 2004; Tobin et al. 2005; have been deposited in deeper water than the Curdsville Calner et al. 2010; Bergström et al. 2010) and Estonia and Member (Brett et al. 2004) but if the same applies to the Latvia (e.g. Ainsaar et al. 1999, 2004a, 2004b; Kaljo et Gålås Member compared to the Furnesfjorden Member al. 2004a, b, 2007), and the stratigraphic position of the requires further studies. The Logana Member has an GICE is now well established in several key sections in abrupt upper contact used as the M5A/M5B sequence these countries. In Figure 7, we compare the δ13C curve boundary by Brett et al. (2004). Because the upper con- of the Mjøsa Formation with those from Fjäcka, Sweden, tact of the Gålås Member, which is also lithologically well and the Mehikoorma (421) drill-core from southeastern marked in the Furuberget Quarry, has been described Estonia. as a regionally distributed erosion surface by Opalinski & Harland (1981), it may well correlate with the M5A/ In the well-known Fjäcka section in central Sweden (for M5B sequence boundary. It appears that equivalents to stratigraphy, see Bergström 2007), the GICE starts in the the Holtjern Member should occur in the Grier Member middle part of the Skagen Formation and ranges into of Kentucky but chemostratigraphic and/or biostrati- the middle portion of the overlying Moldå Formation graphic evidence for this interpretation is not yet avail- (Saltzman et al. 2003; Barta et al. 2007; Bergström et al. able. At any rate, based on the data now at hand, it would 2009a). Based on δ13C chemostratigraphy, we correlate seem that the top of the Mjøsa Formation at Furuberget the Swedish GICE interval with that in the Gålås Member is unlikely to be younger than the base of the Shermanian of the Mjøsa Formation. The lower portion of the Skagen Substage of the Chatfieldian Stage in terms of the North Formation has baseline δ13C values similar to those of the American standard regional stage classification. For fur- Furnesfjorden Member, which suggests that these two ther discussion of the GICE and its occurrences across intervals may be coeval although they are of very differ- North America, see, for instance, Saltzman et al. (2003), ent thickness and have very different faunas. The contact Young et al. (2005), and Barta et al. (2007).

Fig. 7. Comparison of the GICE curves from Furuberget, Fjäcka (after Saltzman et al. 2003), Mehikoorma (421) drill-core (after Kaljo et al. 2007), and Puxihe Quarry (after Bergström et al. 2009b). Note the similarity in baseline and peak excursion values. The M4/M5 sequence boundary of Holland & Patzkowsky (1996) is inferred to be at the Kinnekulle K-bentonite in the Swedish and Estonian successions, at the base of the Pagoda Formation on the Chinese Yangtze Platform, and conceivably at the base of the Mjøsa Formation at Furuberget. 74 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY between the Skagen Formation and the underlying Dalby the central part of the Yangtze Platform. Although the Formation, which is marked by the distinctive and wide- GICE interval in these sections is vastly different from spread Kinnekulle K-bentonite (Bergström et al. 1995), that of the Mjøsa Formation in terms of fauna, lithofa- may well be equivalent to the base of the Mjøsa Forma- cies, and stratigraphic thickness, the GICE provides an tion and represent a lithologic change that is coeval important correlation link to the Norwegian formation. with that at the base of the Mjøsa Formation. However, The beginning of the GICE in most of the central Yang- in Sweden there is a change from dominating carbonate tze Platform sections is several m above the base of the to dominating calcareous mudstone whereas there is the Pagoda Formation, the basal contact of which represents opposite lithologic change in the Norwegian study area. a conspicuous lithologic change from the siliciclastics of The base of the Skagen Formation may also be equivalent the Miaopo Formation to the rather pure carbonates of to the North American M4/M5 sequence boundary. The the Pagoda Formation. It is tempting to speculate that latter is located at, or very close to, the level of the Millbrig the base of the Pagoda Formation may be a correlative K-bentonite, which has been regarded as equivalent to the with the base of the Mjøsa Formation (Fig. 7), especially Kinnekulle K-bentonite (Huff et al. 1992; Saltzman et al. as there is conodont evidence that the topmost Miaopo 2003). No trace of the Kinnekulle K-bentonite has been formation is coeval with the topmost Dalby Formation in found in the Lake Mjøsa region although this ash bed is Sweden (Bergström et al. 2009b). well represented in the Oslo-Asker District (Bergström et al. 1995). However, it is conceivable that this soft clay Oslo-Asker District, Norway bed, which ought to be located at, or close to, the contact between the Mjøsa and the Furuberget formations, was As noted previously, our record of the GICE in the Mjøsa squeezed out during the severe folding and faulting that Formation is the first of this important δ13C excursion in affected particularly the shaly beds of the Furuberget For- the Upper Ordovician successions in Norway. Accord- mation below the base of the Mjøsa Formation in the Fur- ingly, the GICE cannot yet be used for establishing the uberget outcrops and elsewhere in the study area. presently somewhat enigmatic position of the Mjøsa Formation in terms of the standard succession in the In Estonia, the base of the Keila Stage is taken to be at Oslo-Asker District. Recent authors have correlated this the base of the Kinnekulle K-bentonite. The lower part formation with the Solvang and Venstøp formations of the Keila Stage is characterized by baseline δ13C values in the latter district (e.g. Owen et al. 1990), perhaps at (Fig. 7) but near the middle of the Keila interval, there is least partly based on the record of the trilobite Toxochas- a moderate increase in δ13C values that marks the begin- mops extensa in the Furnesfjorden and Eina members ning of the GICE. The GICE ranges upwards into the (Opalinski & Harland, 1981) and in the Solvang Forma- Oandu Stage (Kaljo et al. 2007) but establishing its upper- tion (Størmer 1953). As noted by Jaanusson (1976) and most range is complicated by the presence of significant McNamara (1980), this species appears in Sweden at stratigraphic gap(s) within the Oandu Stage, particu- the base of the Moldå Formation and ranges in Estonia larly in the shelf area (Ainsaar et al. 1999, 2004a), which from the Johvi to the Oandu stages, hence within part of cuts out part of the GICE. However, in the Mehikoorma the GICE interval. Thus, there is agreement between the (421) drill-core (Martma 2005) from southeastern Esto- occurrence of this species and the δ13C chemostratigra- nia, which represents a basin or middle-outer ramp phy of the Mjøsa Formation. Indirect evidence provides depositional environment with a stratigraphically more a clue to the position of the GICE in the Oslo-Asker suc- complete succession than that on the shelf in northern cession. That is, if the Kinnekulle K-bentonite is situated Estonia, the end of the GICE is in the middle part of the at approximately the same level as the base of the Mjøsa Oandu Stage. Based on δ13C chemostratigraphy, we con- Formation as suggested above, then the latter level should clude that the Mjøsa Formation is likely to correlate with correspond to the level of this ash bed in the Oslo-Asker an interval from the base of the Keila Stage to approxi- succession, which is somewhat above the middle of the mately the top of the Oandu Stage, although it might Arnestad Formation (Bergström et al. 1995). This is sup- possibly range a little higher. This is consistent with the ported by the occurrence of many fossils in the upper fact that a few conodonts with Laurentian affinities have part of this formation that suggest a correlation of the been recorded from strata of Oandu age in Estonia, espe- latter interval with the lower part of the Skagen Forma- cially from the Saku Member of the Vasalemma Forma- tion in Sweden (Owen et al. 1990; Hansen & Harper tion (e.g. Viira 1974; Sweet & Bergström 1984; Bergström 2008). Such a correlation of the base of the Mjøsa For- 1997). It is of interest to note that the latter formation is mation is significantly lower stratigraphically than that similar to the Mjøsa Formation also in containing bio- shown by Owen et al. (1990). The position of the top of herms (Hints 1990; Hints & Meidla 1997). the Mjøsa Formation in terms of the Oslo-Asker succession remains uncertain although it is probably below the China Venstøp Formation. Hamar (1966) recorded from the top part of the Solvang Formation Amorphognathus com- In a recent study (Bergström et al. 2009b), the GICE was plicatus and Hamarodus europaeus (=n. genus & n. sp. in recorded for the first time in Asia, where it was recog- Hamar 1966), two distinctive species that appear a little nized in the Pagoda Formation at several sections in above the GICE in Sweden and Estonia (Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 75

only in the relatively shallow water successions elsewhere (2009b). This suggests that the top of the Gålås Mem- in Baltoscandia but also in comparable platform sequences ber is older than the top part of the Solvang Formation. in North America and China. This evidently implies that The stratigraphic relations of the GICE interval between this event was global and that it may be related to eustasy. the Lake Mjøsa region and the Oslo-Asker District may Also, the surprising fact that some facies changes in the be more precisely clarified by a δ13C study of the Frog- Mjøsa Formation that are used for its classification into nerkilen and Solvang formations that we plan to carry members can be directly correlated with apparently coeval out as part of a second phase of the present project. lithological changes in the Kentucky succession was not expected and it may also reflect eustatic control.

Concluding remarks In this connection it is appropriate to refer to a recent study of Ordovician sea level changes in the Oslo Region The inferred relationships between the Mjøsa Forma- by Nielsen (2004). Within the stratigraphical interval tion and some North American, Baltoscandian, and Chi- dealt with in the present paper, he considered the Arnes- nese stratigraphical units that have been revealed from tad Formation to represent a deepening event (Arnestad the present, in some respects preliminary, investigation Drowning Event) and the overlying Frognerkilen For- are summarized in Figure 8. Our study demonstrates the mation a brief shallowing episode (Frognerkilen Low- usefulness of δ13C chemostratigraphy to unravel relations stand Event) of late Keila and Oandu age. The Nakkhol- that would be very difficult, if not impossible, to recognize men Formation was interpreted to reflect a new sea level using traditional lithostratigraphy and/or biostratigraphy. rise (Nakkholmen Drowning Event) and the superja- Our chemostratigraphically based conclusions regarding cent Solvang Formation to have been deposited during the stratigraphic relations of the Mjøsa Formation are con- a new regression (Solvang Lowstand Event). We agree sistent with the available, in most cases less decisive, bio- with these sea-level interpretations but believe that his stratigraphic evidence. Of perhaps more general interest is attempts to correlate these events to North America are the discovery that the base of the Mjøsa Formation may out of phase with current trans-Atlantic correlations. represent a depositional event that can be recognized not For instance, the Frognerkilen Lowstand Event, which

Fig. 8. Stratigraphic summary chart showing proposed correlation between the successions discussed in the text. The δ13C chemostratigraphy suggests that the Gålås Member corresponds to the upper Keila and lower Oandu Baltoscandian Regio- nal Stages. In the absence of graptolites, the precise level of the base of the Mjøsa Formation in terms of the Global Katian Stage remains unknown. However, in view of the fact that the base of this stage is slightly below the beginning of the GICE in the Katian GSSP section in Oklahoma (Goldman et al. 2008), it is likely that the base of the Katian Stage is coeval with a level in the upper part of the Furnesfjorden Member. The M4/M5 sequence boundary at the base of the Chatfiel- dian Stage is interpreted to be coeval with the level of the Kinnekulle K-bentonite in Sweden and the East Baltic region and with the probably unconformable base of the Pagoda Formation on the Yangtze Platform in China. The Kinnekulle K-bentonite is present in the middle part of the Arnestad Formation in the Oslo-Asker district, where it is not associated with a lithologic change, unconformity, or other indication of a depositional break. Nevertheless, this level may be approxi mately coeval with the base of the Mjøsa Formation. The age of the top of the Mjøsa Formation is currently uncertain but it is unlikely to be younger than the Baltoscandic Oandu Stage and the North American Balsamian Substage (Barta et al. 2007) of the Chatfieldian Stage. Note that the precise relations between the boundaries of the Baltoscandian standard stages in the study interval and the formations in the Oslo-Asker District are still not precisely established. 76 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY should be well within the GICE interval, is regarded ning, Rapporter och Meddelanden 128. as corresponding to the Turinian/Chatfieldian Stage Bergström, S. M., Chen Xu, Guttiérrez-Marco, J. C. & Dronov, A. 2009a: The new chronostratigraphic classification of the Ordovi- boundary rather than to an interval well up in the Chat- cian System and its relations to major regional series and stages and fieldian Stage. At the present time, we are unable to apply δ13C chemostratigraphy. Lethaia 41, 97-107. Nielsen’s (2004) event stratigraphy to the Mjøsa Forma- Bergström, S. M., Chen Xu, Schmitz, B., Young, S. A., Rong, J.-Y., Salt- tion or Lexington Limestone, and further studies are zman, M. R. 2009b: First documentation of the Guttenberg δ13C clearly needed to sort out which were true eustatic events excursion (GICE) in Asia: Chemostratigraphy of the Pagoda and and which apparent changes of sea level were caused by Yanwashan formations in southeastern China. Geological Magazine local epeirogeny. 146, 1-11. Bergström, S. M., Hamar, G. & Spjeldnæs, N. 1998: Late Middle Ord- ovician conodonts with Laurentian affinities from the Mjøsa and Acknowledgements. We are indebted to the Naturhistorisk Museum, Furuberget formations in southeastern Norway. Seventh Interna- Oslo for providing access to Hamar’s conodont collections and for tional Conodont Symposium held in Europe, Bologna-Modena 1998, making a vehicle available for the fieldwork. We thank the authorities of the Furuberget Quarry (now known as the Franzefoss Miljøkalk Abstracts, 13-14. Quarry) for permission to study the succession there and to collect rock Bergström, S. M., Huff, W. D., Kolata, D. R. & Bauert, H. 1995: samples, and Håkan Bergström for companionship in the field. We very Nomenclature, stratigraphy, chemical fingerprinting, and areal dis- much appeciate the comments on the manuscript by the journal revie- tribution of some Middle Ordovician K-bentonites in Baltoscandia. wers Per Ahlberg and Dimitri Kaljo. The laboratory work on the iso- GFF 117, 1-13. tope samples was performed by Bjørn Buchardt and Bo Petersen at the Bergström, S. M., Huff, W. D., Saltzman, M. R., Kolata, D. R. & Leslie, Stable Isotope Laboratory, Institute of Geography and Geology, Univer- S. A. 2004: The greatest Volcanic Ash falls in the Phanerozoic: sity of Copenhagen. Trans-Atlantic relations of the Ordovician Millbrig and Kinnekulle K-bentonites. The Sedimentary Record 2, 4-8. Bergström, S. M., Saltzman, M. R. & Schmitz, B. 2006: First record of References the Hirnantian (Upper Ordovician) δ13C excursion in the North American Midcontinent and its regional implications. Geological Ainsaar, L., Meidla, T. & Martma, T. 1999: Evidence for a widespread Magazine 143, 657-678. carbon isotope event associated with late Middle Ordovician sedi- Bergström, S. M., Saltzman, M. R., Kolata, D. R. & Huff, W. D. 2001: mentological and faunal changes in Estonia. Geological Magazine The Ordovician (Chatfieldian) Guttenberg Carbon Isotope Excur- 136, 49-62. sion. 2. Significance of clarifying Baltoscandic and trans-Atlantic Ainsaar, L., Meidla, T. & Martma, T. 2004a: The middle Caradoc facies biostratigraphic relations and the equivalence of the North Ameri- and faunal turnover in the Late Ordovician Baltoscandian paleo- can Millbrig and Baltoscandic Kinnekulle K-bentonites. Geological basin. Palaeogeography, Palaeoecology, Palaeoclimatology 210, 119- Society of America Abstracts with Programs 33(4), A-41. 133. Bergström, S. M., Schmitz, B., Saltzman, M. R. & Huff, W. D. 2010. Ainsaar, L., Meidla, T. & Tinn, O. 2004b. Middle and Upper Ordovi- The Upper Ordovician Guttenberg δ13C excursion (GICE) in North cian stable isotope stratigraphy across the facies belts in the East America and Baltoscandia: Occurrence, chronostratigraphic signi- Baltic. In: Hints, O. & Ainsaar, L. (Eds.). WOGOGOB-2004. 8th ficance, and paleoenvironmental relationships. Geological Society of Meeting of the Working Group on the Ordovician Geology of America Special Paper, 466, 37-67. Baltoscandia. Conference materials. Abstract and Field Guidebook, Bergström, S. M. & Sweet, W. C. 1966: Conodonts from the Lexington Tartu, 11-12. Limestone (Middle Ordovician) of Kentucky and its lateral equiva- Barta, N. C., Bergström, S. M. & Saltzman, M. R. 2007: First record lents in Ohio and Indiana. Bulletins of American Paleontology 50, of the Ordovician Guttenberg δ13C excursion (GICE) in New York 269-441. State and Ontario: Local and regional chronostratigraphic impli- Brenchley, P. J. & Marshall, J. D. 1999. Relative timing of critical events cations. Northeastern Geology and Environmental Sciences 29, 276- during the Late Ordovician mass extinction---new data from Oslo. 298. In: Kraft, P. & Fatka, O. (Eds.). Quo Vadis Ordovician?. Acta Uni- Bergström, S. M. 1964: Remarks on some Ordovician conodont faunas versitatis Carolinae, Geologica 43(1/2), 187-190. from Wales. Acta Universitatis Lundensis, II(3), 1-65. Brenchley, P. J., Marshall, J. D., Hints, L., & Nõlvak, J. 1997: New data Bergström, S. M. 1971: Conodont biostratigraphy of the Middle and solving an old biostratigraphic problem: The age of the upper Ord- Upper Ordovician of Europe and eastern North America. Geologi- ovician brachiopod Holorhynchus giganteus. Journal of the Geologi- cal Society of America Memoir 127, 83-159. cal Society London 154, 335-342. Bergström, S. M. 1980: Conodonts as paleotemperature tools in Ordo- Brett, C. E., McLaughlin, P. I., Cornell, S. R., & Baird, G. C. 2004: Com- vician rocks of the Caledonides and adjacent areas in Scandinavia parative sequence stratigraphy of two classic Upper Ordovician and the British isles. Geologiska Föreningens i Stockholm Förhand- successions, Trenton Shelf (New York-Ontario) and Lexington lingar 102, 377-392. Platform (Kentucky-Ohio): implications for eustasy and local tec- Bergström, S. M. 1997: Conodonts of Laurentian faunal affinities from tonism in eastern Laurentia. Palaeogeography, Palaeoclimatology, the Middle Ordovician Svartsætra limestone in the Trondheim Palaeoecology 210, 295-329. region, central Norwegian Caledonides. Norges Geologiske Under- Calner, M., Lehnert, O., & Joachimski, M. 2010: Carbonate mud søkelse Bulletin 432, 59-69. mounds, conglomerates, and sea-level history in the Katian (Upper Bergström, S. M. 2002: Guide to some Ordovician conodont locali- Ordovician) of central Sweden. Facies 56, 157-172. ties in Kentucky. Annual Meeting of the North-Central and South- Dahlqvist, P. & Bergström, S. M. 2005: The lowermost Silurian of Eastern Sections of the Geological Society of America, Lexington, Jämtland, central Sweden: conodont biostratigraphy, correlation Kentucky, 2002, Pander Society Field Trip Guide, 1-22. and biofacies. Transactions of the Royal Society of Edinburgh, Earth Bergström, S. M. 2007: The Ordovician conodont biostratigraphy in Sciences 96, 1-19. the Siljan region, south-central Sweden: a brief review of an inter- Goldman, D., Leslie, S. A., Nõlvak, J., Young, S., Bergström, S. M. & national reference standard. In Ebbestad, J.O. R., Wickström, L. M. Huff, W. D. 2007: The Global Stratotype Section and Point (GSSP) & Högström, A. E. S. (Eds.): WOGOGOB 2007, Ninth meeting of for the base of the Katian Stage of the Upper Ordovician Series at the Working Group on Ordovician Geology of Baltoscandia. Field Black Knob Ridge, southeastern Oklahoma, USA. Episodes 30, 258- Guide and Abstracts, 26-41, 63-78. Sveriges Geologiska Undersök- NORWEGIAN JOURNAL OF GEOLOGY The δ13C chemostratigraphy of the Upper Ordovician Mjøsa Formation 77

270. Oslo Region, Norway. 30. Stratigraphy of the Mjøsa Limestone Hamar, G. 1966: The Middle Ordovician of the Oslo region, Norway. in the Toten and Nes-Hamar areas. Norsk Geologisk Tidsskrift 61, 22. Preliminary report on conodonts from the Oslo-Asker and Rin- 59-78. gerike districts. Norsk Geologisk Tidsskrift 46, 27-83. Owen, A., Bruton, D. L., Bockelie, J. F. & Bockelie, T. G. 1990: The Hansen, J. & Harper, D. A. T. 2008: The late Sandbian-earliest Katian Ordovician successions of the Oslo Region, Norway. Norges Geolo- (Ordovician) brachiopod immigration and its influence on the giske Undersøkelse Special Publication 4, 1-54. brachiopod fauna in the Oslo region, Norway. Lethaia 41, 25-35. Raymond, P. E. 1916: Expedition to the Baltic Provinces of Russia and Hints, L. 1990: Locality 2:2. Vasalemma Quarry. In: Kaljo, D. & Nestor, Scandinavia. Bulletin of the Museum of Comparative Zoology at H. (Eds.). Field Meeting Estonia 1990. An excursion guidebook, 131- Harvard College 56,179-286. 133. Institute of Geology, Estonian Academy of Sciences, Tallinn. Richardson, J. G. & Bergström, S. M. 2003: Regional stratigraphic rela- Hints, L. & Meidla, T. 1997: Ordovician. Oandu Stage. In: Raukas, A. tions of the Trenton Limestone (Chatfieldian, Ordovician) in the & Teedumäe, A. (Eds.). Geology and Mineral resources of Estonia, eastern North American Midcontinent. Northeastern Geology and 52-88. Estonian Academy Publishers, Tallinn. Environmental Sciences 18, 93-115. Holland, S. M. & Patzkowsky, M. E. 1996: Sequence stratigraphy and Saltzman, M. R., Bergström, S. M., Huff, W. D. & Kolata, D. R. 2001: long-term paleoceanographic changes in the Middle and Upper The Ordovician (Chatfieldian) Guttenberg Carbon Isotope Excur- Ordovician of eastern United States. Geological Society of America sion. 1. New data from the eastern North American Midcontinent Special Paper 306, 117-129. and Baltoscandia. Geological Society of America Abstracts with Pro- Holtedahl, O. 1909: Studien über die Etage 4 des norwegischen Silur- grams 33(4), A-41. systems beim Mjøsen. Skrifter Norske Videnskaps-Akademi i Oslo. Saltzman, M. R., Bergström, S. M., Huff, W. D. & Kolata, D. R. 2003: 1. Matematisk-Naturvidenskapelig Klasse 1909(7), 1-76. Conodont and graptolite biostratigraphy and the Ordovician (Early Huff, W., D., Bergström, S. M. & Kolata, D. R. 1992: Gigantic Ordo- Chatfieldian, Middle Caradocian) δ13C excursion in North Ame- vician volcanic ash fall in North America and Europe: Biological, rica and Baltoscandia: implications for the interpretation of the tectonomagmatic, and event-stratigraphic significance. Geology 20, relations between the Millbrig and Kinnekulle K-bentonites. In: 875-878. Albanesi, G. L., Beresi, M. S., & Peralta, S. H. (Eds.). Ordovician Jaanusson, V. 1973: Aspects of carbonate sedimentation in the Ordovi- from the Andes. INSUGEO, Serie Correlación Geológica 17, 137-142. cian of Baltoscandia. Lethaia 6, 11-34. Schmitz, B. & Bergström, S. M. 2007: Chemostratigraphy in the Jaanusson, V. 1976: Faunal dynamics in the Middle Ordovician Swedish Upper Ordovician: Regional significance of the Hir- (Viruan) of Balto-Scandia. In: Bassett, M. G. (Ed.). The Ordovician nantian δ13C excursion (HICE) in the Boda Limestone of the Siljan System: Proceedings of a Palaeontological Association Symposium, region. GFF 129, 133-140. 301-326. University of Wales Press, Cardiff. Skjeseth, S. 1963: Contribution to the geology of the Mjøsa districts Jaanusson, V. 1979: Ordovician. In: Robison, R. A. & Teichert, C. and the classical Sparagmite area in southern Norway. Norges Geo- (Eds.). Treatise on Invertebrate Paleontology. Part A. Introduction, logiske Undersökelse 220, 1-126. A136-A166. Geological Society of America and University of Kansas. Spjeldnæs, N. 1982: The Ordovician of the districts around Mjøsa. In: Kaljo, D., Hints, L., Martma, T., Nõlvak, T. & Oraspõld, A. 2004a: Late Bruton, D. L. & Williams, S. H. (Eds.). IV. International Sympo- Ordovician isotope trend in Estonia, its significance in stratigraphy sium on the Ordovician System. Field excursion guide. Paleontolo- and environmental analysis. Palaeogeography, Palaeoclimatology, gical Contributions of the University of Oslo 279, 148-163. Palaeoecology 210, 165-185. Størmer, L. 1953: The Middle Ordovician of the Oslo region, Nor- Kaljo, D., Martma, T., Neuman, B. E. E. & Rønning, K. 2004b. Carbon way. 1. Introduction to stratigraphy. Norsk Geologisk Tidsskrift 31, isotope dating of several uppermost Ordovician and lower Silu- 37-141. rian sections in the Oslo Region, Norway. In: Hints, O. & Ainsaar, Strand, T. & Henningsmoen, G. 1960: Cambro-Silurian stratigraphy. L. (Eds.): WOGOGOB-2004. 8th meeting of the Working Group on In: Geology of Norway, 128-169. Norges Geologiska Undersøkelse the Ordovician Geology of Baltoscandia. Field Guide and Abstracts, 208. 51-52. Tartu. Sweet, W. C. 1979: Conodonts and conodont biostratigraphy of post- Kaljo. D., Martma, T. & Saadre, T. 2007: Post-Hunnebergian Ordovi- Tyrone Ordovician rocks of the Cincinnati region. United States cian carbon isotope trend in Baltoscandia, its environmental impli- Geological Survey Professional Paper 1066-G, G1-G26. cations and some similarities with that of Nevada. Palaeogeography, Sweet, W. C. 1981. Genus Aphelognathus Branson, Mehl, & Bran- Palaeoclimatology, Palaeoecology 245, 138-155. son, 1951 and Genus Oulodus Branson & Mehl, 1933. In: Zie- Kiær, J. 1897: Faunistische Uebersicht der Etage 5 des norwegischen gler, W. (Ed.), Catalogue of Conodonts, IV, 27-56, 193-221. E. Silursystems. Skrifter Norske Videnskaps-Akademi i Oslo. 1. Mate- Schweizerbart’sche Verlagsbuchhandlung, Stuttgart. matisk-Naturvidenskapelig Klasse 1897 (3), 1-76. Sweet, W. C. 1984: Graphic correlation of upper Middle and Upper Ludvigson, G., Witzke, B. J., Gonzáles, L. A., Carpenter, S. J., Schnei- Ordovician rocks, North American Midcontinent Province, U.S.A. der, C. L. & Hasiuk, F. 2004: Late Ordovician (Turinian-Chatfiel- In: Bruton, D. L. (Ed.). Aspects of the Ordovician System. Palaeon- dian) carbon isotope excursions and their stratigraphic and palae tological Contributions from the University of Oslo 295, 23-35. oceanographic significance. Palaeogeography, Palaeoclimatology, Sweet, W. C. & Bergström, S. M. 1984: Conodont provinces and bio Palaeoecology 210, 187-214. facies of the Late Ordovician. Geological Society of America Special Martma, T. 2005: Ordovician carbon isotopes. In: Põldvere. A. (Ed.). Paper 196, 69-87. Mehikoorma (421) drill core. Estonian Geological Sections. Bulletin Tobin, K. J., Bergström, S. M. & de la Garza, P. 2005: A mid-Carado-

6. Geological Survey of Estonia, Tallinn, 25-27. cian (453Ma) drawdown in atmospheric pCO2 without ice sheet McNamara, K. J., 1980: Taxonomy and distribution of chasmopine tri- development? Palaeogeography, Palaeoclimatology, Palaeoecology lobites. Geological Magazine 117, 65-80. 226, 187-204. Nielsen, A. T. 2004: Ordovician sea levels changes: A Baltoscandian Viira, V. 1974: Ordovician conodonts of the East Baltic. Institute of Geo- perspective. In: Webby, B., Paris, F., Droser, M. L. & Percival, I. logy, Estonian Academy of Sciences, Tallinn. 142 pp. G. (Eds.). The Great Ordovician Biodiversification Event, 84-93. Worsley, D., Aarhus, N., Bassett, M. G., Howe, M. P. A., Mørk, A. & Columbia University Press, New York. Olaussen S. 1983: The Silurian succession of the Oslo Region. Nor- Nõlvak, J. & Grahn, Y. 1993: Ordovician chitinozoan zones from ges Geologiske Undersøkelse 384, 1-57. Baltoscandia. Review of Palaeobotany and Palynology 79, 245-2269. Young, S. A., Saltzman, M. R. & Bergström, S. M. 2005: Upper Ordo- Opalinski, P. R. & Harland, T. L. 1981: The Middle Ordovician of the vician (Mohawkian) carbon isotope (δ13C) stratigraphy in eastern 78 S. M. Bergström et al. NORWEGIAN JOURNAL OF GEOLOGY

and central North America: Regional expression of a perturbation of the global carbon cycle. Palaeogeography, Palaeoclimatology, Palaeoecology 222, 53-76. Young, S. A., Saltzman, M. R., Bergström, S. M., Leslie, S. A. & Chen Xu 2008: Paired δ13Ccarb and δ13Corg records of Upper Ordovician (Sandbian-Katian) carbonates in North America and China: Impli- cations for paleoceanographic change. Palaeogeography, Palaeocli- matology, Palaeoecology 270, 166-178.