1 Department of Geological Sciences, The Ohio State University, Columbus, OH 43210. E–mail: saltzman.11@osu.edu

2 Department of Geology, University of Cincinnati, Cincinnati, OH 45220.

3 Illinois State Geological Survey, 615 Peabody Drive, Champaign, IL 61820.

Key words: Carbon isotope. K–bentonite. Chatfieldian. North America. Sweden.

Two significant Carbon isotope excursions are currently recognized in the Upper Ordovician Global Series, one in the Hirnantian (Gamachian) Stage, and one in the early Chatfieldian North American Stage. Whereas the former excursion, known as the Hirnantian excursion, has received a substantial amount of recent study (see, e.g., Kaljo et al., 2001; Brenchley et al., 2003), relatively little has been published on the latter excursion (see, e.g., Ainsaar et al., 1999), which is now commonly referred to as the Guttenberg Isotope Carbon Excursion or GICE. During the past few years, we have studied the GICE in eastern and central North America and in Sweden. One aspect of our studies has been to explore the geographic and stratigraphic ranges of the GICE and its potential as a chronostratigraphic tool for assessment of local and regional stratigraphic relations. In the present paper, we will first briefly review the currently known distribution of the GICE and its relations to conodont and graptolite biostratigraphy. In a later part of the paper, we will discuss trans–Atlantic biostratigraphic relations and their bearing on relations between the North American Millbrig and the Baltoscandic Kinnekulle K–bentonites.

The GICE was first recognized in the Guttenberg Member of the Decorah Formation in Iowa (Hatch et al., 1987). Later studies (Ludvigson et al., 1996, 2000) have added much information about its occurrence and development. In this region, it is expressed as a relatively abrupt change in the d13C curve from –2‰ to –1‰ to about +2 to +3‰ (Figure 1). These high values are followed upwards by a relatively sudden drop down to previous values. The excursion occurs in an approximately 2–4 m thick interval in Minnesota that thins southeastward into Illinois and Iowa.

Figure 1. Comparison of the d13C excursion (GICE) in sections in the North American Midcontinent and Baltoscandia. Note similarity in shape of curve. The Wisconsin and Iowa sections are from Ludvigson et al. (2000), and that from Estonia from Ainsaar et al. (1999).

Our work has centered on sections in Kentucky, Tennessee, Virginia, Alabama, Oklahoma, Arkansas, and New York State (Saltzman et al., 2001; Bergström et al., 2001a, 2001b). The pre–excursion values in Kentucky are around +1‰ and the maximum excursion values near +3‰ (Figure 1), hence larger than in the Upper Mississippi Valley. As is the case in the latter region, the excursion occurs a short distance (in Kentucky less than 10 m) above the Millbrig K–bentonite. In Kentucky it begins in the Logana Member of the Lexington Limestone, in central Tennessee in the basal Hermitage Formation, in southwestern Virginia in the Trenton Limestone, in Oklahoma in the lower Viola Springs Formation, and in New York State in the Napanee Formation. It is also present in the Salona Formation of central Pennsylvania (Patzkowsky et al., 1997).

In North America (Figure 2), most of the GICE occurs in Sweet’s (1984) P. tenuis Conodont Zone, but it may possibly be present also in the uppermost part of the P. undatus Zone, and in the coeval portion of the A. tvaerensis Conodont Zone of Bergström (1971)(Figure 1). It is not yet clear if the GICE starts in the uppermost part of the C. americanus Graptolite Zone, or in the overlying O. ruedemanni Zone, but it does not appear to extend into the C. spiniferus Graptolite Zone.

Figure 2. Comparison if the range of GICE in North America and Baltoscandia and its relation to conodont and graptolite zones. Note the closely similar stratigraphic position of GICE to the Millbrig and Kinnekulle K–bentonites indicating no great age difference between these K–bentonites.

In Sweden (Figure 2), it starts in strata traditionally correlated with the lower part of the D. clingani Graptolite Zone, and the entire excursion apparently lies within that graptolite zone. It occurs in the uppermost part of the A. tvaerensis Conodont Zone, but the level of the base of the overlying A. superbus Conodont Zone remains undetermined, although it is somewhere in the middle of the Moldå Formation. Accordingly, in terms of conodont biostratigraphy, the GICE occupies a closely similar stratigraphic interval in Baltoscandia as in North America.

Two graptolite species are of special interest for the trans–Atlantic biostratigraphic correlation of the GICE interval between Baltoscandia and the American Midcontinent, namely Climacograptus bicornis and C. spiniferus, which are used as zone fossils in North America but also occur in the Baltoscandic succession. In southern Baltoscandia, C. bicornis ranges from the upper N. gracilis Zone to well into the D. clingani Zone and it appears that the GICE overlaps the uppermost range of C. bicornis. In the East Baltic, C. spiniferus ranges through the Rakveran and lower Nabalan Stages (Figure 2), and it seems as if the excursion ends near the level of appearance of C. spiniferus (Figure 2). In North America, C. bicornis ranges into the lowermost Chatfieldian in Oklahoma (Finney, 1986) but the precise position of the top of its range, and the base of the overlying C. americanus Zone, are poorly constrained in our study sections. The characteristic species C. spiniferus appears somewhat higher in the succession (Figure 2). We conclude that the relations between the GICE and the ranges of these graptolite species are closely similar in Baltoscandia and North America.

As shown above, the GICE in North America and Baltoscandia occupies a very similar stratigraphic position in relation to the North American Millbrig and the Baltoscandic Kinnekulle K–bentonites. Based on biostratigraphy and then available isotopic age evidence, Huff et al., (1992) correlated these giant ash falls and speculated that they might have originated from the same eruption(s). This correlation has subsequently become somewhat controversial based on geochemical evidence. In a recent study, Min et al. (2001), using 40Ar/39Ar dating on biotite from the Millbrig and Kinnekulle K–bentonites, concluded that there is an age difference of 6.8±2.8 Ma between these ash layers, the Kinnekulle being much older than the Millbrig. This is in contrast to previously published 206U/238Pb ages for these ash beds (Tucker and McKerrow, 1995) that were given as 456±1.8 for the Kinnekulle and 453±1.3 for the Millbrig. Based on the biostratigrahic data presented above, and the fact that the GICE occupies a strikingly similar stratigraphic position to these ash beds in North America and Baltoscandia, we conclude that an age difference of 6 Ma between these K–bentonites as improbable. Indeed, this presumed 6 Ma age difference is larger than the difference between the isotopic date of the Millbrig (448±2.0 Ma) presented by Min et al. (2001) and that commonly used for the top of the Ordovician (443 Ma), a stratigraphic interval that in North America is subdivided into 10 graptolite, 10 conodont, and 17 chitinozoan zones. Based on the evidence we have presented above, we conclude that if there is an age difference between the Millbrig and the Kinnekulle, it is much smaller than that suggested by Min et al. (2001).

The early Chatfieldian Guttenberg d13C isotope excursion (GICE) is widely recognizable in carbonate facies successions in eastern and central North America and in Baltoscandia, and it is shown to be an excellent tool for precise local and trans–Atlantic chemostratigraphic correlations. The shape of the GICE d13C curve, with two peaks at some localities, is similar in North America and Baltoscandia, and in both these regions, the GICE occupies a very similar stratigraphic position in the conodont and graptolite zone schemes. Furthermore, it has a virtually identical stratigraphic position in relation to the North American Millbrig and the Baltoscandic Kinnekulle K–bentonites. Based on this and the biostratigrahic evidence at hand, we believe the recently published opinion based on 40Ar/39Ar isotopic dating that there is an age difference of 6 Ma between these K–bentonite beds is incorrect, perhaps due to thermal resetting, and their coeval nature remains credible.

Accepted: June 15, 2003