07

Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165 – 185 www.elsevier.com/locate/palaeo

Late Ordovician carbon isotope trend in Estonia, its significance in stratigraphy and environmental analysis Dimitri Kaljo *, Linda Hints, To˜nu Martma, Jaak No˜lvak, Asta Oraspo˜ld Institute of Geology at Tallinn Technical University, 7 Estonia Blvd., 10143 Tallinn, Estonia Received 24 November 2002; accepted 23 February 2004

Abstract Carbon isotope changes during most of Late Ordovician time (from the mid-Caradoc Kinnekulle K-bentonite until the beginning of the Silurian) were investigated. As the corresponding sequence of rocks is stratigraphically nearly complete in Estonia, an attempt was made to use it to elaborate the general pattern of carbon isotope changes in the Late Ordovician. Complications were caused by several local or regional hiatuses in the middle and late Caradoc and Hirnantian. A total of 385 whole rock samples were studied from eight drill cores in northern and central Estonia. The following positive carbon isotope events were observed: (1) the mid-Caradoc excursion (peak d13C value 2.2x) in the uppermost part of the Keila Stage, also known in Sweden; (2) the first late Caradoc excursion (1.9x) in the lower part of the Rakvere Stage; (3) the second late Caradoc excursion (2.4x) in the upper part of the Nabala Stage; (4) the early Ashgill excursion (2.5x) in the lowermost part of the Pirgu Stage; (5) the widely known large Hirnantian excursion (in Estonia the peak value reaches 6.7x) in the Porkuni Stage. The study interval comprises a long ( f 10 Ma) period characterized by low-magnitude carbon isotope changes and a following brief ( f 2 Ma) interval with large changes. No obvious lithological preference for hosting the positive shifts was recorded. In principle, the d13C values exceeding the background values may occur in all types of rocks present in a sedimentary basin. Several d13C positive excursions (values 1.5x to 3x) in the Mohawkian of North America are evidence that the minor Caradoc and early Ashgill d13C positive shifts in Baltoscandia may have counterparts in Laurentia. If correctly correlated, these shifts may have global significance. The Hirnantian excursion is usually linked to a major glacial event, even if some carbon cycling mechanisms are not completely understood. The environmental causes suggested for the earlier minor shifts range from global climatic and glacial events to very local changes in basin regime and sea level. Our study supports the primary role of climatic or climatically triggered oceanic processes. D 2004 Elsevier B.V. All rights reserved. Keywords: Carbon isotopes; Environmental reasons; Estonia; Late Ordovician; Stratigraphy

1. Introduction

* Corresponding author. Institute of Geology, Estonian Academy of Sciences, 7 Estonia Blvd., 10143 Tallinn, Estonia. Tel.: +372-2-454-653; fax: +372-6-312074. E-mail address: [email protected] (D. Kaljo). 0031-0182/$ - see front matter D 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.palaeo.2004.02.044

Stable carbon isotopes are widely used in geology for application such as chemostratigraphic correlation and interpretation of palaeoenvironmental parameters. The latest Ordovician isotope geology has received much attention, especially the Hirnantian positive

166

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

carbon isotope excursion (Brenchley et al., 1994, 2003; Marshall et al., 1997; Finney et al., 1999; Kump et al., 1999; Kaljo et al., 2001) that has been among the main clues for understanding glacial processes and mass extinctions at that time. Recently, the first data on the Caradoc and lower Ashgill stable isotope record have been published (Ludvigson et al., 1996, 2001; Patzkowsky et al., 1997; Ainsaar et al., 1999; Kaljo et al., 1999; Pancost et al., 1999; Fanton and Holmden, 2001), providing a basis for elucidation of the pre-Hirnantian environmental history. Whether there existed a long-lasting glacial epoch with several advances of glaciers (Patzkowsky et al., 1997; Hamoumi, 1999) remains debatable. The aim of this study is to elaborate the general pattern of the Late Ordovician d13C changes to serve as a point of comparison to other data sets and for the discussion of carbon cycling. In addition to previously published information, new data obtained from eight sections in Estonia are used. Naturally, the Estonian data alone are insufficient for a global survey, but because of a relatively complete stratigraphical succession and detailed biostratigraphy available, these may serve as a basis for further advancements. We tried to overcome difficulties caused by stratigraphical gaps in the sequence by studying a series of overlapping core sections to compile a complete carbon isotope curve that could serve as a tool for interregional correlation. The Hirnantian experience provides a clear model about the correlation possibilities (Brenchley et al., 2003). We shall discuss the possible correlation of several earlier carbon isotope shifts identified on Baltica and Laurentia that appear to be potentially global events. The general carbon isotope trend is also used to discuss reasons for climate changes before the end-Ordovician glacial event.

2. Geological setting The study area was a part of the Baltoscandian Basin located on the southwest margin of the Baltica palaeocontinent, which moved from the middle to low latitudes of the Southern Hemisphere during the late Ordovician (Torsvik et al., 1996). The basin embraces the Oslo Region, several outcrops in central and southern Sweden, and a continuous outcrop belt in

North Estonia and the environs of St. Petersburg in Russia. In subsurface, the basin rocks have been traced into the Moscow region in the east and into Latvia, Lithuania and northeast Poland in the south (Fig. 1A). All these areas except the Oslo Caledonian foreland basin (Bruton and Harper, 1988) were stable cratonic areas. The western part of the basin close to the Tornquist – Teisseyre Zone (craton margin) is a pericratonic sea with a ramp facies belt pattern, but its eastward gulf-like extension is an epicontinental sea with a characteristic pattern of the shelf facies. The Baltoscandian Basin has been divided (Ma¨nnil, 1966) into three broad facies belts (Fig. 1A) with specific rock and faunal associations corresponding roughly to the consecutive depth zones of the sea. Jaanusson (1995) used the term ‘‘confacies belts’’ for these units, and stressed that differences between East Baltic and Scandinavian faunas are in some cases (e.g. Livonian Tongue) more important than the lithological composition of the adjoining areas. The North Estonian – Lithuanian facies belt is represented by shallow water carbonate rocks (mainly limestones). The equivalent deep-shelf or ramp facies (the Central Baltoscandian facies belt) consists of more argillaceous wackestones, marlstones and mudstones, partly with limestone nodules in south Estonia, Latvia and eastern Sweden. The deepest part of the basin, represented by graptolitic black shales in South Scandinavia, belongs to the Scanian facies belt (Fig. 1A). The Oslo Region and Moscow Basin fit into the above facies model, but are separated because of the effects of the Caledonian tectonic processes (Oslo) or the location in the interior of Baltica (Moscow). The facies distribution in the Estonian part of the basin (our study area sensu stricto) generally follows the above pattern, but is rather variable through time due to changing environmental conditions (climate, sea level, influx of terrigenous material, etc.). Fig. 1B shows the facies distribution of the Vormsi Stage, which corresponds to a drowning episode, with high clay content of the carbonate rocks and graptolitic shales extending into south Estonia far from the basin depression. According to the facies model currently in use (Nestor and Einasto, 1997), four depth zones can be distinguished (Fig. 1B). The shallow shelf consists of skeletal limestones (grainstones and rudstones) with bioherms and carbonate mounds. The middle shelf is

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

167

Fig. 1. General facies zonation of the Baltoscandian basin and location of the studied core sections (black dots). (A) General facies belts (Ma¨nnil, 1966; Jaanusson, 1995 , modified): E—North Estonian, L—Lithuanian, CB—Central Baltoscandian with the Livonian Tongue (LT) included, S—Scanian; O—Oslo. Tornquist – Teisseyre (TTZ) and Sorgenfrei – Tornquist (STZ) zones by Tuuling (1998). Dotted line—outer limit of the area with continuous distribution of the Ordovician rocks. (B) Palaeogeography and facies of Vormsi age (early Ashgill). The belt without signature next to the land denotes uncertain transition between land and sea. Facies distribution north of the dotted line extrapolated from rocks occurring south of the line.

marked by the occurrence of skeletal pack- and wackestones and nodular micritic limestones with marly intercalations. The deep shelf consists of interbedded micritic limestones and graptolitic mudstones or marlstones with limestone nodules. Mudstones and graptolite shales occur in the basinal areas at the craton margin. The regional lithostratigraphy of the stratigraphical classification (Fig. 2) reflects changes in the rock content of formations along an onshore – offshore transect. The lithological logs of the cores analyzed (Figs. 3 and 4) illustrate the temporal variations in rock composition, and clearly show a cyclic pattern. The cyclicity may be linked to the sea level dynamics (high- and low-stands), changes in climate (alternation of arid and humid episodes), and/or oceanic processes because all are factors of primary importance that should be considered. Generalized Late Ordovician sea-level fluctuations are presented in Fig. 5, together with carbon isotope and climatic changes. Our study interval commences with rocks of the Keila and Oandu stages marking the end of the early-middle Caradoc high-stand and the following sea-level fall. The latter is clearly expressed by subregional stratigraphic gaps and shallow-water lithologies (Figs. 2 and 3), including carbonate mounds

surrounded by cystoid grainstones (i.e., the Vasalemma Formation) distributed in the Saku-Vasalemma area of northwest Estonia. The succeeding sea-level rise began in later Oandu time as evident by the argillaceous rocks of the Hirmuse Formation that overlie reef structures. The relatively high but variable sea level was highest at the Caradoc – Ashgill boundary (Vormsi Stage, Fig. 1B) and lasted until the middle of Pirgu time. It was followed by the end-Ordovician regression comprising a series of sea-level rises and falls. The regression reached its maximum in the middle of the Hirnantian (Porkuni) Stage and caused subaerial erosion and stratigraphic gaps, especially in North Estonia (Fig. 2). The late Oandu flooding event also marks a change in the sedimentary regime of the basin, highlighted by the dominance of algae among the skeletal components of the micritic rocks. The algae are especially abundant in the upper parts of the Rakvere and Nabala stages and in the middle of the Pirgu Stage (Po˜lma, 1972). The increased abundance of algae correlates with a temperature rise and an alternation of humid and arid climatic episodes (Kaljo et al., 1999) that are linked to the movement of Baltica closer to the equator (Nestor and Einasto, 1997). Dronov and

168

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Fig. 2. Stratigraphical classifications used in Estonia correlated with British and North American series (principally according to No˜lvak, 1997). Lithostratigraphy on the right is simplified, only formation names used in Table 1 and main gaps in sedimentation are shown. Ages from Tucker and McKerrow (1995), with additions noted in text. CBLT as in Fig. 1. Approximate positions of core sections against the lithostratigraphy are shown at the bottom.

Holmer (2002) reached the same conclusion that a transition from cool-water temperate to warm-water carbonate sedimentation occurred during the late Oandu. Due to climate change, a sea-level rise, and a reduced influx of terrigenous material into the basin, upper Caradoc (Rakvere and Nabala stages) in north Estonia is dominated by micritic limestones intercalated with more argillaceous intervals. The limestones contain sparse skeletal particles of macrofossils, commonly concentrated along bedding planes. The limestone beds thin basinwards and grade into wackestones and marlstones (Mossen and Mo˜ntu formations). Lower-middle Ashgill (Pirgu Stage) carbonate rocks generally consist of more argillaceous micritic limestones (mainly wackestones) beds with the calcareous algae Palaeoporella and Vermiporella at some horizons. A few carbonate mounds similar to the Boda mounds in Dalarna, Sweden (Hints and Meidla, 1997) occur at some levels in the Moe Formation (Fig. 2). A diagnostic pre-Hirnantian brachiopod Holorhynchus has been identified in the uppermost beds of the Pirgu Stage (Hints, 1993; Brenchley et al., 1997), and,

together with the carbon isotope signature of the corresponding rocks, dates the Pirgu–Porkuni boundary and associated eustatic and isotope events. The Hirnantian Porkuni Stage comprises the top of the Ordovician sequence in Estonia, and contains two regressive cycles, that began with a brief flooding event (Kaljo et al., 2001; Brenchley et al., 2003). The ¨ rina Formation, and the upper is lower cycle is the A the Saldus Formation. Stratigraphic gaps occur at both the bottom and top of these units, and the most ¨ rina Formation and undersignificant gap caps the A lies the Saldus Formation. Siliciclastic rocks occur at ¨ rina Formation in north Estonia in the top of the A association with the boundary between these two cycles. Within the upper cycle, sandy limestones occur in the Saldus Formation in the Viljandi core (Fig. 4) and elsewhere in the offshore Central Baltoscandian facies belt. These features indicate significant sea-level changes and tectonic activity during the latest Ordovician. The hiatuses mentioned above (shown by discontinuity surfaces in Figs. 3 and 4) occur widely in the sections and are the main obstacle in establishing a

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

169

Fig. 3. Correlation and carbon isotope data of the core sections from northern Estonia (see Fig. 1B). Distribution intervals of some characteristic brachiopods and other fossils are shown to the right of the logs. Indexes of stratigraphical units: P—Pa¨a¨sku¨la, S—Saue, L—Lehtmetsa members of the Keila Stage, Sa—Saku Member and H—Hirmuse Formation of the Oandu Stage, V—Vasalemma Formation, S – V—transition from Saue to Vasalemma rocks.

170 D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Fig. 4. Correlation and carbon isotope data of the core sections from middle Estonia. For legend, refer to Fig. 3.

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

171

Fig. 5. A generalized carbon isotope trend correlated with several environmental events and processes (see text for explanation). Five main isotope excursions are marked with the corresponding names, doubtful shifts (?) established only in one section are shown by a dashed line. The sea level curve is compiled based on core sections included in Table 1. Occurrence levels of the K-bentonite layers (arrows) in Estonia (B1 – B7) according to authors’ data, those in Scandinavia (Kinnekulle) and North America (Deicke, Millbrig) according to radiometric ages by Min et al. (2001). Positions of the three Laurentian d13C shifts (black dots) are based on the following sources: the Carimona, upper Turinian—Ludvigson et al. (2001), the Guttenberg, lower Chatfieldian—Patzkowsky et al. (1997) and Bergsto¨m et al. (2001), and the Dunleith shift, higher in the Chatfieldian—Fanton and Holmden (2001). Content of the algal debris according to Po˜lma (1972). For climate episodes, see text (Jeppsson, 1990; Kaljo et al., 1999).

complete carbon isotope trend. To overcome this difficulty, we used a series of cores from different parts of the basin correlated by means of biostratigraphy and other chronostatigraphic criteria, including the isotope events identified in the drill cores. The isotope data are discussed in more detail in Section 4. The stratigraphical terminology used in the East Baltic area is based on stage-level chronostratigraphical units. The Estonian practice is to use regional stages that are not equivalent to the standard or international stages (of which only the Tremadocian and Darrivilian have formal approval). Most, but not all, of the ‘‘East Baltic’’ regional stages (Fig. 2) are

shorter than the standard ones. For example, should the Hirnantian become a standard stage, the Porkuni Regional Stage would be its full equivalent. We use terms like ‘‘middle’’ or ‘‘upper Caradoc’’ because there exist no internationally accepted subdivisions of the Caradoc and Ashgill stages or series.

3. Material and methods Carbon isotopes were determined by the wholerock analyses of samples from eight drill cores penetrating the late Ordovician of Estonia. The lithostratigraphy and biostratigraphy of the core sections (Fig.

172

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

2) was previously established based on macrolithological and microlithological descriptions of the rock sequences (Po˜lma, 1972; Oraspo˜ld, 1982a,b; Po˜lma et al., 1988) and detailed study of the distribution of chitinozoans (No˜lvak and Grahn, 1993; No˜lvak, 1999), ostracodes (Meidla, 1996) and brachiopods (Hints and Meidla, 1997; Hints, 1998). The location of the sections studied is shown in Fig. 1B and the corresponding isotope data are presented in Table 1. A few previously published isotope curves (not reproduced here) were also used when discussing the general trend (Fig. 5), for example Tartu and Ristiku¨la of Estonia (Ainsaar et al., 1999), and Fja¨cka of Sweden (Ainsaar et al., 2000). Some data from the Estonian Rapla and Kaugatuma sections were published previously (Kaljo et al., 1999), but additional analyses are included here. Data from the Ko˜rgessaare, Orjaku, Saku 1098A, Vasalemma 772 and Viljandi drill cores are published for the first time. The sampling strategy pursued a continuous and detailed isotopic characterization of the whole section. Sampling did not depend on fossil occurrences but did consider the stratigraphic context (lithology, unit thickness, position of unit boundaries). We used the same sampling strategy for whole-rock carbon isotope analyses that was adopted by Kaljo et al. (2001) with the aim of compiling an uninterrupted isotope curve. All together 385 samples were analyzed, with a mean sampling interval of 1 m (varying between 0.4 m in Rooku¨la and 1.5 m in Kaugatuma). Table 1 presents data of all carbon and oxygen isotope analyses, but this paper will focus on the carbon isotope record, since we consider oxygen data unreliable due to difficulties connected with the whole-rock method of isotope analysis used. The methodology was explained in detail earlier (Kaljo et al., 1997, 1998); here only the most important aspects are noted. The whole-rock samples were powdered to a < 10-ı`m grain size and reacted with 100% phosphoric acid at 100 jC for 15 min. The reproducibility of the results is better than 0.1x. Previous studies (Brenchley et al., 1994; Kaljo et al., 1997) show little diagenetic alteration of Baltic early Palaeozoic rocks, so we expect reliable carbon isotope analysis in whole-rock samples. Brenchley et al. (2003) discussed the reliability of isotope signals in the late Ordovician rocks of Estonia in detail newly and noted that the major changes in isotope values

Table 1 Whole rock isotopic data from the drill cores investigated d13C

d18O

Depth (m)

Stratigraphy

Kaugatuma 329.90 332.20 334.80 337.10 339.20 340.65 340.70 340.80 341.35 341.45 341.80 342.30 342.70 343.00 343.55 344.10 344.95 345.90 347.20 348.40 349.25 350.85 352.25 354.35 356.40 358.40 362.00 366.00 370.40 374.50 376.20 377.50 378.70 379.30 380.50 383.50 385.40

Juuru St. Juuru St. Juuru St. Juuru St. Juuru St. Juuru St. ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A Adila Adila Adila Adila Adila Adila Adila Adila Adila Adila Halliku Halliku Halliku Halliku Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Fja¨cka

1.28 0.69 0.48 0.31 0.64 0.15 4.39 4.35 3.69 3.94 2.49 1.80 1.76 1.17 0.81 0.54 0.28 0.45 0.19 0.33 0.31 0.35 0.71 0.05 0.29 0.32 0.65 0.31 0.76 0.57 0.63 2.03 2.46 2.18 2.00 1.94 1.49

3.96 3.77 3.71 3.78 3.77 5.05 3.20 4.21 4.93 5.15 5.01 3.32 3.99 3.93 3.69 3.23 3.02 3.85 3.14 4.22 4.04 3.43 3.44 3.81 3.33 3.30 2.51 3.90 4.12 4.10 3.98 4.38 4.29 4.50 3.85 3.52 3.25

Ko˜rgessaare 9.70 13.80 15.30 16.35 17.25 18.10 18.90 19.85 21.05 22.15 29.70 30.65

Saunja Saunja Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Rakvere St. Rakvere St.

2.35 1.64 0.27 0.52 0.38 0.58 0.27 0.12 0.28 0.53 1.55 1.72

5.82 6.27 4.13 4.62 4.58 3.72 5.02 4.65 4.65 4.59 4.54 4.30

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185 Table 1 (continued)

173

Table 1 (continued) 13

d C

18

d O

Depth (m)

Stratigraphy

Ko˜rgessaare 31.80 32.65 33.90 37.65 38.55 41.40 43.05 44.35 44.85 45.20 46.05 47.05 48.00 49.00 49.95 51.10 52.05 53.10 54.15 55.05 56.10 57.05 58.20 59.05 60.15 61.15 62.20 63.20 64.10 65.15 66.20 67.30 68.25 69.10 69.95 70.60

Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Oandu St. Oandu St. Oandu St. Oandu St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St.

1.79 1.81 1.95 1.82 1.84 1.79 1.84 1.48 1.26 0.19 0.92 0.52 1.24 0.83 1.19 0.86 1.34 1.51 1.25 1.40 1.51 0.86 0.87 1.13 1.30 1.28 1.08 1.03 0.50 0.37 0.31 0.68 0.48 0.69 0.29 0.11

4.75 5.43 5.08 5.41 5.20 5.59 4.54 5.16 4.55 4.82 4.91 4.57 4.34 5.30 6.25 4.84 4.28 3.64 4.42 4.20 3.41 3.39 4.55 4.59 3.69 4.06 3.75 3.89 4.45 4.72 4.35 3.95 4.05 3.74 4.18 4.40

Orjaku 36.30 37.25 38.20 38.60 39.45 40.60 41.20 42.80 43.65 45.20 46.00 46.35 46.95 48.50

Juuru St. Juuru St. Juuru St. ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A Adila Adila Adila Adila Adila

0.36 1.62 2.29 2.98 4.88 4.55 4.79 2.82 1.90 0.09 0.34 0.57 0.57 0.13

5.90 6.94 6.43 5.26 5.42 5.17 6.66 7.39 5.36 5.43 4.32 4.65 4.51 5.31

Depth (m)

Stratigraphy

Orjaku 50.45 52.80 54.90 55.65 56.80 58.50 59.50 61.05 62.05 63.05 65.65 67.15 69.05 72.65 75.75 76.25 76.65 78.30 79.75 80.20 80.85 82.65 84.70 87.30 89.10 90.20 91.70 92.65 94.95 95.75 96.70 98.00 99.05 99.95 101.95 102.85 103.80 104.90 105.80 106.05 107.10 108.20 109.25 110.20 111.10 112.15 113.20 114.20 115.20 116.15 117.20 118.20

Adila Adila Adila Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Saunja Saunja Saunja Paekna Paekna Paekna Paekna Paekna Paekna Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St.

d13C 0.29 0.22 0.28 0.73 1.05 0.45 1.29 0.94 0.85 0.64 0.43 0.00 0.20 0.53 0.68 1.27 1.25 1.04 0.74 0.81 0.55 0.99 0.87 1.34 0.74 0.17 1.74 0.21 0.52 0.72 1.51 1.76 0.84 0.64 0.53 0.27 0.09 0.32 0.08 0.22 0.03 0.44 0.10 0.08 0.11 0.26 0.13 0.50 0.05 0.25 0.24 1.52

d18O 5.91 5.86 5.48 4.27 6.37 5.06 6.86 6.65 7.07 6.70 6.79 6.16 6.93 6.44 6.60 4.47 5.16 5.97 5.37 5.57 5.01 3.75 5.38 4.02 4.20 5.55 5.01 6.21 5.78 3.98 5.29 4.81 5.37 3.98 3.51 4.84 5.89 4.90 4.76 6.35 4.76 5.09 3.83 4.80 5.63 5.45 5.14 4.91 4.26 4.28 4.75 3.77

(continued on next page)

174

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Table 1 (continued)

Table 1 (continued) 13

d C

18

d O

Depth (m)

Stratigraphy

Orjaku 119.10 120.10 121.15 122.15 123.15 124.20 125.05 126.20 127.50 129.00 129.60 130.80 131.70 132.80 134.00 134.70 135.80 136.85 137.70 138.80 139.65

Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Rakvere St. Oandu St. Oandu St. Oandu St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St. Keila St.

1.40 1.90 1.53 1.72 1.62 1.77 1.61 1.67 1.50 1.27 0.89 1.03 0.83 0.98 1.09 0.93 1.05 0.72 1.02 1.10 0.98

4.64 4.43 5.40 4.64 5.55 4.35 4.62 4.28 5.48 4.55 4.07 4.05 3.92 4.74 3.80 4.79 4.74 5.17 4.65 4.02 4.34

Rapla 31.90 32.35 32.65 33.25 34.15 34.55 35.25 36.15 36.55 37.10 37.95 39.40 41.00 44.05 47.05 51.70 52.15 56.55 58.40 60.10 62.00 62.95 64.85 67.30 74.00 74.95 76.85 77.70 78.50

¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A ¨ rina A Adila Adila Adila Adila Adila Adila Moe Moe Moe Moe Moe Moe Moe Moe Moe Moe Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare

1.43 5.09 5.17 5.59 3.19 5.18 2.52 0.76 1.53 1.85 0.10 0.41 0.87 1.04 0.63 0.26 0.89 0.61 0.03 0.08 0.53 0.19 1.54 0.39 1.86 1.37 1.14 0.81 0.11

5.34 4.99 4.64 4.81 5.22 5.16 5.47 3.70 3.72 3.73 4.80 4.91 5.05 4.90 4.04 3.59 4.96 5.93 3.95 5.87 5.98 5.62 3.43 6.00 3.79 4.57 4.34 5.25 5.11

Depth (m)

Stratigraphy

Rapla 80.70 81.85 83.90 85.65 86.55 88.65 89.95 92.65 93.40 96.75 97.35 99.10 99.10 100.25 101.29 101.45 101.75 102.25 102.70 102.90 103.08 104.00 104.85 104.95 105.30 106.80 107.45 108.30 111.30 112.30 114.20 114.80 115.05 116.70 117.10 118.30 118.95 120.00 121.75 123.50 123.95 124.50 125.40 126.20 127.50 128.55 131.00 133.15 135.20 135.80 138.20 139.00

Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Ko˜rgessaare Saunja Saunja Saunja Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Paekna Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Ra¨gavere Hirmuse Kahula Kahula Kahula Kahula Kahula Kahula Kahula Kahula

d13C 0.78 1.05 0.05 0.18 0.31 0.89 0.92 0.63 1.12 0.83 0.45 0.34 0.34 0.56 0.61 1.03 1.61 0.89 1.67 1.62 1.60 1.76 1.05 1.01 0.44 0.14 0.17 0.05 0.10 0.66 0.86 0.45 0.19 1.39 1.05 1.61 1.50 1.91 1.82 1.82 1.74 1.64 1.16 0.78 0.69 0.80 0.69 0.30 0.33 0.09 0.82 0.73

d18O 3.95 3.47 5.51 5.96 4.81 4.17 4.17 5.36 6.41 5.55 3.80 3.75 3.75 3.76 5.19 4.41 3.64 6.82 3.83 3.55 3.49 3.72 4.01 4.08 3.99 5.22 5.83 4.04 5.15 5.16 3.82 5.83 4.53 4.41 4.96 4.51 5.01 4.25 4.98 5.04 4.66 5.22 6.66 3.81 4.09 4.09 4.44 4.42 5.53 5.27 3.42 3.52

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185 Table 1 (continued)

175

Table 1 (continued) 13

d C

18

d O

Depth (m)

Stratigraphy

Rapla 141.90 144.00

Kahula Kahula

0.72 0.57

4.77 3.64

Saku 1098A 1.20 1.20 1.50 2.00 2.30 3.20 3.50 4.05 4.50 4.85 5.60 7.10 7.60 8.30 8.90 9.95 10.05 11.30 12.80 13.60 15.10 15.90 17.60 19.60 22.15 25.20

Vas. Vas. Vas. Vas. Vas. Vas. Vas. Vas. Vas. Vas. Vas. KF., KF., KF., KF., KF., KF., KF., KF., KF., KF., KF., KF., KF., KF., KF.,

1.69 1.60 1.66 1.61 1.55 1.32 0.76 0.40 0.51 0.96 0.64 0.69 0.87 0.93 1.27 1.52 1.71 1.21 0.80 0.92 0.53 0.37 0.47 0.31 0.57 0.20

4.86 4.96 4.99 5.20 5.42 5.11 4.57 5.95 6.20 5.79 7.22 5.39 4.29 4.43 4.65 4.46 4.03 4.37 5.04 4.43 4.48 5.06 5.99 5.24 4.61 5.79

Vasalemma 772 3.60 Vas. Fm., mid. beds 4.00 Vas. Fm., mid. beds 6.90 Vas. Fm., mid. beds 7.60 Vas. Fm., mid. beds 8.00 Vas. Fm., mid. beds 10.45 Vas. Fm., lower beds 11.15 Vas. Fm., lower beds 12.20 Vas. Fm., lower beds 12.80 Vas. Fm., lower beds 14.40 Vas. Fm., lower beds 15.00 Vas. Fm., lower beds

2.12 2.03 1.16 1.78 1.48 1.48 1.25 1.06 1.07 0.79 0.71

6.19 8.43 4.29 3.99 5.54 4.65 5.53 5.47 5.42 5.52 5.71

Viljandi 275.20 276.20 276.40 276.70 277.00 278.10 278.90

0.57 1.53 1.68 3.21 3.20 2.24 2.57

5.39 4.89 3.63 4.82 5.16 5.30 4.81

Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Fm., Saku Mb. Lehtmetsa Mb. Lehtmetsa Mb. Lehtmetsa Mb. Lehtmetsa Mb. Lehtmetsa Mb. Lehtmetsa Mb. Lehtmetsa Mb. Saue Mb. Saue Mb. Saue Mb. Saue Mb. Pa¨a¨sku¨la Mb. Pa¨a¨sku¨la Mb. Pa¨a¨sku¨la Mb. Pa¨a¨sku¨la Mb.

Juuru St., Silurian Juuru St., Silurian Saldus Saldus Saldus Saldus Saldus

Depth (m)

Stratigraphy

Viljandi 279.60 280.40 280.90 281.30 281.70 282.00 282.30 282.50 283.50 284.60 285.70 286.80 287.80 289.00 290.00 291.40 292.40 293.40 294.60 295.80 296.60 297.80 298.60 299.60 300.90 301.70 302.70 304.00 305.00 306.00 307.40 308.70 309.20 309.80 310.40 310.60 310.90 311.20 312.40 313.00 314.00 315.10 316.20 317.30 317.90 318.30 318.50 319.00 319.50 320.00 320.80 321.40

Saldus Saldus Saldus Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Halliku Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Jonstorp Tudulinna Tudulinna Tudulinna Tudulinna Tudulinna Tudulinna Tudulinna Tudulinna Tudulinna Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu

d13C 2.08 3.63 4.10 1.66 1.52 1.41 1.19 0.90 0.90 0.70 0.98 1.13 1.23 1.52 1.67 1.52 1.45 1.44 1.16 1.12 1.25 1.18 1.06 1.01 1.21 1.21 1.16 1.06 1.15 1.70 2.21 2.04 2.07 2.25 1.93 1.58 1.69 1.69 1.04 1.04 1.13 1.29 1.25 1.11 1.26 1.17 0.93 0.98 0.88 0.58 0.73 0.32

d18O 4.31 5.52 3.97 4.39 4.73 4.41 4.43 4.93 4.77 4.53 4.26 4.16 4.72 4.00 3.72 4.17 4.27 3.73 3.27 3.71 3.50 3.79 3.80 3.92 3.81 4.03 5.18 3.60 3.29 3.51 2.48 3.75 3.83 3.73 3.27 4.80 4.43 4.64 3.75 3.34 3.36 3.58 3.67 3.47 4.89 4.49 5.17 4.84 4.95 6.07 5.85 6.31

(continued on next page)

176

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Table 1 (continued) Depth (m)

Stratigraphy

321.80 322.20 322.50 322.80 323.00 323.30 323.60 323.80 324.10 324.60 324.80 325.10 325.40 325.80 326.80 328.00 328.90 330.00 330.90 331.50 332.10 333.30 334.50 335.70 337.00 337.90 338.90 340.00 341.00 342.00 343.00 344.00 345.00 346.00 346.90 349.80

Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Mo˜ntu Rakvere St. Rakvere St. Rakvere St. Variku (conditionally) Variku (conditionally) Variku (conditionally) Variku Variku Variku Variku Variku Variku Variku Variku Variku Variku (conditionally) Variku (conditionally) Kahula Kahula Kahula Kahula Kahula Kahula Kahula Kahula Kahula Kahula Kahula Haljala

d13C 0.63 0.49 0.41 0.21 0.52 0.77 0.66 0.66 0.54 0.62 0.78 0.93 1.05 0.93 1.06 1.12 1.11 0.66 1.51 1.00 1.23 1.19 1.44 1.71 2.12 1.74 1.20 0.63 0.62 0.24 0.37 0.46 0.43 0.59 0.52 0.67

d18O 4.69 5.31 5.97 6.37 6.17 4.43 4.75 4.78 4.03 4.19 4.06 5.20 3.30 3.95 4.14 3.96 4.59 5.21 3.71 5.63 4.92 4.35 3.99 4.31 4.13 4.36 4.01 5.02 5.27 5.41 4.27 4.51 4.21 3.81 4.23 3.63

For the stratigraphical nomenclature, see Fig. 2. Names alone are formations. St.—stage, Vas. Fm.—Vasalemma Formation, Mb— member, KF.—Kahula Formation, mid.—middle.

reflect primary composition. Comparison of our whole-rock isotope data (Kaljo et al., 1998, 2001) with those obtained for brachiopod shells from the Baltic Ordovician (Marshall et al., 1997; Brenchley et al., 2003) and from the Silurian of Gotland (Samtleben et al., 1996) shows only slight difference in d13C values but great similarity of the corresponding curves. The main advantage of the whole-rock method is that sampling could be performed at regular intervals, which is important in developing a continuous isotope curve. The oxygen isotope ratios are more

sensitive to diagenesis (Marshall, 1992) and therefore oxygen isotope data from whole-rock analysis is not trustworthy. Another difficulty arises from the fact that Baltic carbonate rocks are mostly highly variable mixtures of calcite and dolomite that have different oxygen isotope fractionation factors. Marshall et al. (1997) and Brenchley et al. (2003) provide additional information about the Hirnantian oxygen trend based on calcite from brachiopod shells.

4. Results Arranging the available Estonian d13C data (Table 1 and publications referred to below) by regional chronostratigraphic units (Fig. 2), the data indicate five positive excursions that alternate with intervals characterized by relatively low isotope values (f 1x or less). To put the isotope trend in a broader framework, the stratigraphical positions of the five positive d13C excursions are dated in terms of international stratigraphy based on current correlation schemes (Ma¨nnil and Meidla, 1994; No˜lvak, 1997, Fig. 2 here). 4.1. The lower mid-Caradoc (lower Keila stage) interval of low-isotope values At the Kinnekulle K-bentonite and in the lower part of the Keila Stage, the d13C values are low (0.5x to 0.7x), rising slightly in the middle of the stage (1.0x to 1.5x, Fig. 3). The bentonite marks the lower boundary of the Keila Stage practically over the whole of Estonia (Vingisaar, 1972; Bergstro¨m et al., 1995). This part of the stage is correlated with the uppermost Diplograptus multidens graptolite Biozone at the junction of the Amorphognathus tvaerensis and Amorphognathus superbus conodont biozones (Fig. 2; Viira and Ma¨nnik, 1997). 4.2. The mid-Caradoc (uppermost Keila and lower Oandu stage) positive excursion A positive excursion occurs in the uppermost Keila Stage and continues upward into the Oandu Stage. The peak level is well above the Kinnekulle K-bentonite (Fig. 3). The carbon isotope values reach 1.9x to 2.2x in the Tartu, Ristiku¨la (Ainsaar et al., 1999) and

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Viljandi sections (Fig. 4). The shift is not recorded in the northwestern and northern shallow-shelf areas of Estonia (Ko˜rgessaare, Orjaku, Rapla and Rooku¨la cores) due to a stratigraphic gap at this level, but was observed in the Vasalemma – Saku reef facies area (SW of Tallinn, Fig. 1B), in the Lehtmetsa Member of the Saku 1098A core and in the Middle Vasalemma Beds of the Vasalemma 772 core, where d13C values reach 1.7x and 2.1x, respectively (Fig. 3). According to Ainsaar et al. (1999), in the Ristiku¨la and Tartu cores the main excursion is followed by a smaller positive shift slightly higher, in the Oandu Stage. A second shift also occurs in the Saku Member and Upper Vasalemma Beds in the cores shown in Fig. 3. The correlation of the d13C peak interval in the reef facies and in the Viljandi core is well supported by occurrences of the alga Leiosphaeridia and the chitinozoan association (Fig. 3). These observations support the ideas of Ma¨nnil (1960) about the commencement (late Keila time) of the reef/carbonate mound formation in the North Estonian Facies Belt. The mid-Caradoc isotope excursion was first identified by Ainsaar et al. (1999). In addition to the five Estonian sites mentioned above, the shift has been identified also in the Kandava drill core in Latvia (Brenchley et al., 1996) and Fja¨cka outcrop in Dalarna, Sweden (Ainsaar et al., 2000). In the latter locality, the maximum values reach 2.1x. It should be noted that the excursion in the Fja¨cka section shows exactly the same doubled peak pattern of the curve as in Estonia. On the basis of the data available on the time – rock correlation, we date the mid-Caradoc excursion as an event corresponding in Baltica to the lower part of the Dicranograptus clingani graptolite Biozone, and not to the very beginning of the zone. In the global chitinozoan correlation scheme (Paris et al., 1999), this level is just above the Angochitina multiplex Subzone, which occurs in the lower part of the Keila Stage. 4.3. The mid-Caradoc (upper Oandu stage) interval of low isotope values In the upper part of the Oandu Stage d13C values remain low (mostly < 1x, Figs. 3 and 4), but rise in the top of the interval ( f 1.5x, Viljandi, Ristiku¨la). The last rise of the values may be the beginning of the next positive shift.

177

4.4. The first late Caradoc (lower Rakvere stage) positive excursion A positive excursion (1.9x) in the lower part of the Rakvere Stage is well developed in the northwest sections noted above, but it is poorly developed (1.1 – 1.2x) in the southern, deeper shelf sections (Viljandi, Ristiku¨la). The lower boundary of the Rakvere Stage is marked by a sharp lithological change (appearance of micritic limestones), which reflects a change in the sedimentary regime in the Baltoscandian Basin (Po˜lma, 1982; Nestor and Einasto, 1997). The boundary coincides with the beginning of the Cyathochitina angusta chitinozoan Subzone (No˜lvak and Grahn, 1993), and the appearance of new ostracode associations in different parts of the basin (e.g. Olbianella cf. braderupensis, Disulcina perita perita, Meidla, 1996). However, the precise position of the boundary in the Viljandi section remains unclear, and lower Rakvere beds are probably missing. This interpretation is based on the occurrence of the chitinozoans Sphaerochitina? sp. nov. and Spinachitina cervicornis, characteristic of the Oandu Stage, together with C. angusta in the uppermost Variku Formation and in overlying beds up to 324.7 m. The lack of the lower Rakvere beds is also confirmed by the low d13C values that are not typical of lower Rakvere strata in other sections studied. The positive excursion has been established in several Estonian sections containing ‘‘Rakvere-type’’ micritic limestones. Therefore, it appears to be a local phenomenon related to the presence of hiatuses, and needs additional study, especially in sections where early Rakvere time is represented by different lithofacies. The most reliable dating of this event in terms of graptolite zonation seems to be the upper part of the Dicranograptus clingani Biozone based on microfossil and macrofossil correlations (No˜lvak, 1997), including graptolites of the Mossen Formation (Jaanusson, 1982). 4.5. The late Caradoc (upper Rakvere and lower Nabala stages) interval of low isotope values The next interval characterized by relatively low carbon isotope values corresponds to the upper part of the Rakvere and the lower part of the Nabala stages (late Caradoc). The d13C values are close to 0 ( 0.5x

178

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

to + 0.6x). An exception is the Rapla core where values reach 1.7x in the middle of the Paekna Formation (Fig. 2). The reason for this rise is not yet fully understood, but it may refer to another gap in the succession of the Nabala rocks in the neighboring cores, unusual diagenetic effects, or a drilling defect. This doubtful excursion is not considered below. 4.6. The second late Caradoc (upper Nabala stage) positive excursion A positive d13C shift occurs in the upper part of the Nabala Stage (Saunja Formation). The peak values decrease in an offshore direction: Ko˜rgessaare 2.4x, Orjaku and Rapla 1.8x to 2.0x, Viljandi 1.3x (Figs. 3 and 4). The dating of the shift is complicated by the lack of clear biostratigraphical criteria, but the formation is lithologically distinct in the North Estonian Belt because it consists of micritic limestones underlying the argillaceous rocks of the Vormsi Stage. In the Viljandi core, the Saunja Formation is very thin and several discontinuity surfaces occur at its upper boundary. The upper Nabala event is very similar to the preceding positive excursion with respect to the host lithology and section location where the shift has been identified; only d13C maximum values are slightly higher (2.4x). Considering the position of the shift in the sequence and the occurrences of Rectograptus gracilis and Archaeoretiolites regimontanus in the Saunja Formation (Ma¨nnil, 1990), the event level is probably within the Pleurograptus linearis Biozone. 4.7. The latest Caradoc – earliest Ashgill (Vormsi stage) interval of low isotope values The second late Caradoc positive excursion is followed by an interval of low but variable carbon isotope values ( 0.8x to + 1.2x, in the topmost part 1.7x) corresponding to the Vormsi Stage (at the Caradoc –Ashgill junction). The interval is observed in the Orjaku, Rapla and Viljandi cores (Fig. 3). 4.8. The early Ashgill (lowermost Pirgu Stage) positive excursion A distinct positive excursion occurs in the lowermost part of the Pirgu Stage. At Kaugatuma (maxi-

mum d13C values reach 2.5x), the shift interval occurs in shallow-shelf skeletal packstones and biohermal (stromatactis) rocks. The isotope values decrease seaward (at Viljandi 2.2x) in accordance with a general pattern, but the Rapla (1.9x) and Orjaku (1.3x) sections show very low values compared to the Kaugatuma section. This could mean that part of the excursion (especially at Orjaku) is missing. Biostratigraphically, the shift occurs in the Tanuchitina bergstroemi chitinozoan Biozone, just above the Acanthochitina barbata Subzone that marks the uppermost Vormsi Stage. According to the current stratigraphical scheme (No˜ lvak, 1997, Fig. 2), the excursion probably occurs in the lower part of the Dicellograptus complanatus biozone. 4.9. The Ashgill (Pirgu Stage) interval of low isotope values The carbon isotope values remain low (close to or below 0x) throughout most of the Pirgu Stage (middle Ashgill) until the very top where a rise begins that leads into the Hirnantian positive excursion. The Viljandi core (Fig. 4) is an exception, showing a positive excursion (d13C values reaching 1.5 –1.7x, Table 1) in the middle of the Halliku Formation between the last occurrence of Conochitina rugata and the first appearance of Tanuchitina anticostiensis. This shift needs confirmation in other sections, but it may indicate that the northwestern sections like Rapla and Orjaku are incomplete in this interval. If this is the case, the carbon isotope trend and the current chitinozoan biozonation (Fig. 2) will need to be revised based on data from the Viljandi core. 4.10. The Hirnantian (Porkuni Stage) positive excursion The Ordovician carbon isotope record in Estonia is capped by a major positive excursion in the Porkuni Stage (Brenchley et al., 1994, 2003; Kaljo et al., 2001; ¨ rina new data herein). The d13C values in the A Formation (lower Porkuni Stage) reach 5.6x in Rapla, 4.9x in Orjaku and 4.4x in Kaugatuma. The Saldus Formation (upper Porkuni Stage) occurs Viljandi core and has a d13C value of 4.1x. All the Porkuni shelf sections in the North Estonian Belt (Fig.

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

¨ rina subaerial erosion, 2) are truncated due to post-A so the peak isotope values reflect both the Hirnantian trend and the depth of erosion. Some erosion is also expected in Viljandi, however, the isotope value is uncommonly high for the upper Hirnantian. The d13C shifts in the sections of the Porkuni Stage must be considered against the background of the Hirnantian isotope record. The Hirnantian positive excursion has been recognized in many regions of the world, and the reported peak d13C values reach 7x, although the values are usually 5 –6x or less. Brenchley et al. (2003) proposed a composite model for the Hirnantian carbon isotope event based on data from the Baltic (Marshall et al., 1997; Kaljo et al., 2001), Dob’s Linn, Scotland (Underwood et al., 1997), and the Monitor Range, Nevada (Finney et al., 1999). According to this model d13C values began to rise slowly in the uppermost Rawtheyan (Belonechitina gamachiana chitinozoan Biozone), more rapidly in the lower Hirnantian (Spinachitina taugourdeaui Biozone) and reached a peak ( f 7x) in the Conochitina scabra Biozone ( = upper part of the Normalograptus extraordinarius graptolite Biozone). The isotope values fell slightly in the Normalograptus persculptus Biozone, followed by a rapid fall to the beginning of the Silurian, when the values return to the pre-Hirnantian level. Simultaneously with the rising limb of the carbon isotope excursion the Hirnantia community appeared in the Baltic area. Rong et al. (1999) show that this characteristic relatively cool-water assemblage persisted into N. persculptus time and even into the very beginning of the Llandovery. In view of the Brenchley et al. ¨ rina For(2003) model, the shift observed in the A mation represents the lower (extraordinarius) part and that from the Saldus Formation the upper (persculptus) part of the Hirnantian trend.

5. Discussion 5.1. General pattern The previous section describes one major and four minor positive carbon isotope excursions from the upper Ordovician of Estonia that occur in at least two drill cores. These data reveal the general pattern of carbon isotope changes during late Ordovician time

179

(Fig. 5), and the relation of the isotope curve to the environmental events and the K-bentonites used to obtain radiometric ages of the sampled intervals. The isotope curve was compiled using the following procedure: (1) the whole sequence was divided into 26 working units of more or less uniform duration; (2) the absolute age chronology suggested by B. Webby for biodiversity analysis in the IGCP Project 410 was used to calculate mean ages of the working units; (3) mean d13C values were calculated for every unit (using the analyses in Table 1) and plotted against the scale of the radiometric ages. Mean values were used to construct the curve plotted in Fig. 5, so some of the peak values presented in Figs. 3 and 4 are higher, however the general trend is similar. The interval covered by the five positive carbon isotope excursions ranges from 455 Ma (age of the Kinnekulle K-bentonite, Min et al., 2001) to 443 Ma (age of the Ordovician – Silurian boundary, Tucker and McKerrow, 1995). The magnitude of the carbon isotope shifts divides this interval into two clearly different periods. The first 10 Ma is marked by relatively low magnitude isotope shifts, with several minor positive excursions (close to 2x) occurring every 1 – 1.5 Ma. The variation is even smaller (around 1x) during the last 3 Ma (the Pirgu interval of low isotope values). In contrast, the latest Ordovician (less than 2 Ma) includes a sizable isotope excursion of up to 7x, about three times the magnitude of the earlier shifts. The precise time durations are estimates, but the difference between pre-Hirnantian and Hirnantian isotope records is evident. This difference is undoubtedly connected to the contrast between the environmental changes associated with the Hirnantian glaciation and the preceding less intense arid – humid climatic fluctuations. Against this background, it is interesting to note the stepwise increase in peak values of the pre-Hirnantian excursions through time. This slight tendency seems to point to possible intensification of some environmental processes during the pre-glacial Caradoc and Ashgill. 5.2. Relationship of carbon isotope values with facies and bathymetry The main data about relationships between facies and isotope excursions are summarized in Table 2.

180

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

Table 2 Relationships between carbon isotope excursions and corresponding facies Excursion

Main lithologies of host rocks

Onshore/offshore changes in isotopic values (d13C)

Mid-Caradoc

Argillaceous limestones, marls, upper part silty to siltstones, reef area—micritic carbonate mounds, grainstones, packstones Micritic limestones, abundant algal skeletal particles Wackestones, partly argillaceous, partly with biohermal and/or micritic patches Large variety of shallow (grainstones to micritic limestones) and deeper shelf rocks (marls and siltstones, nodular limestones) Different rocks: micritic and more or less argillaceous limestones, partly alternating, marl- and mudstones, limestones with some content of skeletal sand, wackestones

No clear trend observed

Late Caradoc (1st and 2nd) Early Ashgill Late Ashgill (Hirnantian)

Intervals between excursions (cf. Section 4)

There is no obvious lithological preference for hosting the positive shifts as high d13C values may occur in all lithologies. The absence of preference was an expected, but not always obvious, observation given the general nature of carbon isotope fractionation. However, some specific details should be discussed in more detail. In the late Caradoc micritic and skeletal limestones (wackestones), small differences in lithology correlate with carbon isotope shifts. For example, in the Rapla core positive shifts correlate with minima of the calcareous algal debris content in the rock sequences, and d13C values are low within intervals where algae constitute is over 50% of the skeletal material (Kaljo et al., 1999). Considering the co-occurrence of high algal abundance and low clay contents, we could interpret the algal maxima as arid episodes as suggested by the climatic – oceanic model by Jeppsson (1990). This suggests that climatic changes and processes related to carbon fractionation are of more importance than lithology in determining the carbon isotope values. Changes in carbon isotope values along the bathymetric profile in the Ordovician sections of Estonia are not as clear as in Silurian strata within the Baltic region. A clear decrease of d13C values (from 5.2x to 3.1x) was reported in the early Wenlock rocks along an onshore-to-offshore profile (Kaljo et al., 1998).

Values decrease seaward by about 1x Decreasing trend evidently as above, but overshadowed by erosion No clear trend observed in North Estonia, partly due to erosion; seaward decrease recorded in South Estonia Values vary at low levels; no clear bathymetric trend observed

The isotope values presented above for Ordovician strata along a shelf profile illustrate the same pattern of decreasing values in an offshore direction. The main uncertainty is if the depth difference (shallow to middle shelf) is sufficient to really reveal a bathymetric influence. An additional uncertainty is the position of some of the studied boreholes within the complicated bottom topography of the late Ordovician Baltoscandian Basin. 5.3. Correlation with the Laurentian events Several studies that documented the d13C positive excursions (values 1.5 – 3x) in the Mohawkian sections of North America have been published recently (Ludvigson et al., 1996; 2001; Patzkowsky et al., 1997; Pancost et al., 1999; Byers et al., 2001; Fanton and Holmden, 2001). The exact positions of these excursions in relation to graptolite biozonation are not precisely known, but these shifts should be compared to our results. A lower positive carbon isotope excursion (1.5x) has been identified (Ludvigson et al., 2001) in the Carimona and Castlewood members of Iowa just above the Deicke K-bentonite (age 449.8 F 2.3 Ma according to Min et al., 2001), and an upper excursion (2.5x) occurring near the Elkport K-bentonite in the Guttenberg Member of Iowa, Illinois and Minnesota. In Pennsylvania, a larger

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

positive isotope shift (3x) begins just below the Millbrig K-bentonite (age 448.0 F 2.0 Ma according to Min et al., 2001) and reaches maximum values above it in the uppermost Phragmodus undatus and lowermost Phragmodus tenuis conodont Zone (Patzkowsky et al., 1997). Bergsto¨m et al. (2001) considered the excursion at the Millbrig K-bentonite (also named the Guttenberg) coeval with several others in Laurentia and with the Baltoscandian mid-Caradoc shift above the Kinnekulle K-bentonite (age 454.8 F 2.0 Ma according to Min et al., 2001). Byers et al. (2001) established a 3x excursion in the Northern Mississippi Valley in the base of the upper carbonate part of the Decorah Formation. From the Late Ordovician of Iowa, Fanton and Holmden (2001) reported at least four positive carbon isotope excursions above the Decorah Formation, three in the Dunleith Formation, and one in the Wise Lake Formation (although their exact placement in the graptolite biozonation is uncertain). Based on the above data and the recent correlations of Leslie and Bergstro¨m (1995), a mid-Mohawkian (late Turinian –Chatfieldian) set of d13C shifts that are stratigraphically positioned using lithostratigraphy, K-bentonites (Deicke, Millbrig and Elkport), and biostratigraphy. These shifts would correspond to the mid-Caradoc and first late Caradoc excursions identified in this paper. However, the excursions appear to be younger if the shifts are dated on the basis of the radiometric ages of the Deicke and Millbrig K-bentonites (Min et al., 2001). The radiometric dates indicate an early Cincinnatian age correspondingly to a position just below and above the Caradoc –Ashgill junction (449 Ma, Figs. 2 and 5), equivalent to the second late Caradoc and the early Ashgill excursion discussed above. We favor the first correlation for several reasons. It is unclear if the Kentucky K-bentonites studied by Min et al. (2001) occurred within the Diplograptus multidens biozone since no biostratigraphic data was presented. We also question their conclusion that the duration of this biozone is at least 6.8 F 2.8 Ma. We suggest that the correlation of the dated Kentucky K-bentonites with others established in Iowa and Pennsylvania remains undetermined and requires additional study. It seems more likely that the Kentucky K-bentonites, dated by Min et al. (2001), are lower Cincinnatian, and that they may

181

not be the same as K-bentonites associated with the isotope excursions in Iowa and Pennsylvania given the biostratigraphic information currently available. The Estonian record (Fig. 5) suggests that isotope shifts as well as K-bentonites may occur in the lower Cincinnatian strata. Ainsaar et al. (1999) discussed the correlation of the mid-Caradoc shift in Baltica with those in Iowa and Pennsylvania using biostratigraphical data. They noted some differences in the environmental background, but accepted the synchrony of the excursions. We agree that the stratigraphical positions of the midCaradoc excursions in Estonia, Sweden, and Iowa (Ludvigson et al., 1996; Pancost et al., 1999) coincide, but the excursion in Pennsylvania commencing below the Millbrig K-bentonite (Patzkowsky et al., 1997) seems to be slightly earlier than the supposedly synchronous shift well above the Kinnekulle K-bentonite in Baltica. Unfortunately, due to differences in biogeography, the biostratigraphical criteria are inconclusive for the correlation of this interval, but it should be borne in mind that in Estonia there occur several Kbentonites in mid-Caradoc (Fig. 5) as is the case in Laurentia (Emerson et al., 2001; Ludvigson et al., 2001). In spite of these correlation problems, the occurrence of the Laurentian excursions noted above indicates that the minor d13C positive shifts (values close to 2x) identified in the Baltoscandian Caradoc and early Ashgill may have counterparts in North America, but their precise correlation requires additional biostratigraphical and/or geochronological study. If such correlations were successful, these pre-Hirnantian excursions would have global or at least inter-regional significance for understanding the late Ordovician carbon cycle and resolving some of the complicated problems of correlation, sedimentary history and palaeogeography involving Laurentia and Baltica. The Hirnantian excursion is a good example of this, and its inter-regional correlation has demonstrated the value of the simultaneous application of biostratigraphical and carbon isotope data. 5.4. Environmental interpretation The Hirnantian carbon isotope excursion is a well-established global event (Brenchley et al.,

182

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

2003), and the pre-Hirnantian excursions may be significant on an inter-regional scale. The Hirnantian excursion is usually interpreted as an event caused by glaciation (Marshall et al., 1997), but a weathering hypothesis has also been suggested (Kump et al., 1999). Despite the different causal mechanisms proposed to explain the excursion, both models rely on atmospheric CO2 levels as the main driver of the climate, and conclude that positive d13C and d18O excursions reflect glaciations. The minor pre-Hirnantian d13C shifts may be related to glaciations, but other global or local causes have been suggested: (1) Increased productivity and rates of organic carbon burial may have drawn down atmospheric pCO2 and induced global cooling, although not necessarily ice-sheet formation (Patzkowsky et al., 1997). (2) The mid-Caradoc excursion is related to a sedimentary hiatus in onshore areas, suggesting that better ventilation of the ocean and the influx of nutrients enhanced primary productivity and depleted of seawater in 12C during the eustatic lowstand period (Ainsaar et al., 1999). (3) The negative correlation of the algal content and the late Ordovician d13C curve may indicate that the positive excursions occur mostly during cooler climatic episodes (Kaljo et al., 1999), as suggested by the Jeppsson’s (1990) oceanic model for Silurian rocks of Scandinavia. (4) The Turinian – Chatfieldian boundary interval (mid-Mohawkian) with several d13C excursions coincided in part with major palaeoceanographic changes that resulted in cessation of carbonate accumulation and the development of a starved submarine surface across large areas (Ludvigson et al., 2001). (5) The Decorah positive carbon isotope excursion that occurs at the base of carbonate facies may have been induced by an interval of high phytoplankton productivity in the epeiric sea (Byers et al., 2001). (6) Sea-level change may have forced changes in carbon isotope values as a result of increased seawater exchange between neighboring water masses with different 13C enrichment (Fanton and Holmden, 2001).

These proposed mechanisms (carbonate accumulation and sea level changes included) could be interpreted as results of climatic or climatically triggered oceanic processes, similar to those proposed for environmental changes connected with the Hirnantian glacial event. Some mechanisms, such as high primary productivity, are important causes of positive carbon isotope shifts, but are difficult to prove independent of the isotope shifts. In addition, tectonic and cosmic influences should also be considered, but these are also linked to climatic changes related to glacial or, at least, cooling processes. It appears that the minor pre-Hirnantian isotope shifts record a long epoch of alternating cooler and warmer climatic episodes in the late Ordovician, ending with the brief (0.5 Ma, Brenchley et al., 1994) but severe Hirnantian glacial episode. This interpretation is supported by occurrences of glacial sediments on Gondwana (Hamoumi, 1999), although precise dating of these rocks is not yet possible. Ultimately, the isotope data may lead to a revision of concept of a stable late Ordovician greenhouse period (Morrow et al., 1995). In many cases, different interpretations of isotopic events are arguable or equally probable. We have demonstrated how the integration of carbon isotope, biostratigraphic, and other geological data may help evaluate alternative explanations. The continued study of these inter-relations will help solve the complicated problems of correlation, sedimentary history and palaeogeography involving Baltica, Gondwana and Laurentia.

6. Conclusions 1. One major global Hirnantian (Brenchley et al., 1994) and four minor (d13C values 1.9 –2.5x) pre-Hirnantian positive carbon isotope excursions are recorded in upper Ordovician strata of the Baltic region. The minor shifts may be global events but their correlation to other areas of Laurentian and Gondwanan events requires additional study. 2. The magnitude of the carbon isotope changes divides the study interval into a long ( f 10 Ma) low period with only minor variations, and a brief period of more intense changes. This difference

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

reflects environmental changes connected with the Hirnantian glacial event and the preceding cooling episodes of lesser intensity. 3. The Hirnantian excursion is usually linked to a glaciation. For the minor shifts, various global and regional environmental causes have been suggested related to global climate changes, glacial events, oceanic ventilation, and sea level. Our interpretation supports the primary role of climatic or climatically triggered oceanic processes linked to alternations of arid and humid episodes. 4. More attention should be paid to the trend of the d13C values, and less to their numerical values. The latter are dependent on various factors, including study methods and depth of the sedimentary environment. However, no obvious lithological preference for hosting the positive shifts was observed and, in principle, positive excursions of d13C values may occur in all types of rocks forming in a sedimentary basin.

Acknowledgements The authors thank A. Noor for linguistic help. We are grateful to D. Osleger and J.A. Simo for constructive reviews and to M.T. Harris for his many-sided advice. The study was partly supported by the Estonian Science Foundation (grants 4674 and 5042).

References Ainsaar, L., Meidla, T., Martma, T., 1999. Evidence for a widespread carbon isotopic event associated with late Middle Ordovician sedimentological and faunal changes in Estonia. Geol. Mag. 136, 49 – 62. Ainsaar, L., Martma, T., Meidla, T., 2000. The mid-Caradoc carbon isotopic event and its expression in Siljan District, central Sweden. J. Conf. Abstr. 5 (2), 123 – 124. Bergstro¨m, S.M., Huff, W.D., Kolata, D.R., Bauert, H., 1995. Nomenclature, stratigraphy, chemical fingerprinting and areal distribution of some Middle Ordovician K-bentonites in Baltoscandia. Geol. Fo¨ren. Stockh. Fo¨rh. 117, 1 – 13. Bergsto¨m, S.M., Saltzman, M.R., Huff, W.D., Kolata, D.R., 2001. The Guttenberg (Chatfieldian, Ordovician) 13 C excursion (GICE): significance for North American and trans-Atlantic chronostratigraphic correlations and for assessment of the age relations between the North American Millbrig and the Balto-

183

scandic Kinnekulle K-bentonites. Geol. Soc. Amer. Annual Meeting, Nov. 5 – 8. Final Program. Abstract No. 19370. Brenchley, P.J., Marshall, J.D., Carden, G.A.F., Robertson, D.B.R., Long, D.G.F., Meidla, T., Hints, L., Anderson, T.F., 1994. Bathymetric and isotopic evidence for a short-lived Late Ordovician glaciation in a greenhouse period. Geology 22, 295 – 298. Brenchley, P.J., Hints, L., Marshall, J.D., Martma, T., Meidla, T., No˜lvak, J., Oraspo˜ld, A., 1996. Isotopic data combined with biostratigraphy: an Ordovician case study. The Third Baltic Stratigraphical Conference. Abstracts. Field Guide. Tartu, 20 – 21. Brenchley, P.J., Marshall, J.D., Hints, L., No˜lvak, J., 1997. New isotopic data solving an old biostratigraphic problem: the age of the upper Ordovician brachiopod Holorhynchus giganteus. J. Geol. Soc. (Lond.) 154, 335 – 342. Brenchley, P.J., Carden, G.A., Hints, L., Kaljo, D., Marshall, J.D., Martma, T., Meidla, T., No˜lvak, J., 2003. High resolution isotope stratigraphy of Late Ordovician sequences: constraints on the timing of bio-events and environmental changes associated with mass extinction and glaciation. Geol. Soc. Amer. Bull. 115, 89 – 104. Bruton, D.L., Harper, D.A.T., 1988. Arenig-Llandovery stratigraphy across the Scandinavian Caledonides. In: Harris, A.L., Fettes, D.J. (Eds.), The Caledonide/Appalachian Orogen. Geol. Soc. London, Spec. Publ., vol. 38, pp. 247 – 268. Byers, C.W., Simo, J.A., Emerson, N.R., 2001. Paleoceanography of the Decorah Formation (Ordovician, Mohawkian), Northern Mississippi Valley. Geol. Soc. Amer. Annual Meeting, Nov. 5 – 8. Final Program. Abstract No. 27536. Dronov, A., Holmer, L., 2002. Ordovician sea-level curve: Baltoscandian view. In: Satkunas, J., Lazauskiene, J. (Eds.), The Fifth Baltic Stratigraphical Conference ‘‘Basin Stratigraphy—Modern Methods and Problems’’, Extended Abstracts. Geol. Surv. Vilnius, Lithuania, pp. 33 – 35. Emerson, N.R., Fournelle, J., Simo, J.A., Byers, C.W., 2001. Correlation of (Mohawkian) K-bentonites by using apatite chemistry: implication for Decorah Formation carbonate and shale depositional sequences. Geol. Soc. Amer. Annual Meeting, Nov. 5 – 8. Final Program. Abstract No. 22926. Fanton, K.C., Holmden, C., 2001. Sea level forcing of Late Ordovician carbon isotope excursions. Geol. Soc. Amer. Annual Meeting, Nov. 5 – 8. Final Program. Abstract No. 14469. Finney, S.C., Berry, W.B.N., Cooper, J.D., Ripperdan, R.L., Sweet, W.C., Jacobson, S.R., Soufiane, A., Achab, A., Noble, P., 1999. Late Ordovician mass extinction: a new perspective from stratigraphic sections in central Nevada. Geology 27, 215 – 218. Hamoumi, N., 1999. Upper Ordovician glaciation spreading and its sedimentary record in Moroccan North Gondwana margin. In: Kraft, P., Fatka, O. (Eds.), Quo Vadis Ordovician. Acta Univ. Carol., Geol., vol. 43 (1/2), pp. 111 – 114. Hints, L., 1993. Holorhynchus (Pentamerida, Brachiopoda) in the Upper Ordovician of Estonia. Eesti NSV Tead. Akad. Toim., Geol. 42, 118 – 123. Hints, L., 1998. Oandu Stage (Caradoc) in central North Estonia. Proc. Estonian Acad. Sci. Geol. 47, 158 – 172. Hints, L., Meidla, T., 1997. Pirgu Stage. In: Raukas, A., Teeduma¨e,

184

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185

A. (Eds.), Geology and Mineral Resources of Estonia. Estonian Academic, Tallinn, pp. 82 – 85. Jaanusson, V., 1982. Ordovician in Va¨stergo¨tland. In: Bruton, D.L., Williams, S.H. (Eds.), Field Excursion Guide. IV Intern. Symp. Ordovician System. Paleont. Contr. Univ. Oslo, Oslo, vol. 279, pp. 164 – 183. Jaanusson, V., 1995. Confacies differentiation and upper Middle Ordovician correlation in the Baltoscandian Basin. Proc. Estonian Acad. Sci., Geol. 44, 73 – 86. Jeppsson, L., 1990. An oceanic model for lithological and faunal changes tested on the Silurian record. J. Geol. Soc. (Lond.) 147, 663 – 674. Kaljo, D., Kiipli, T., Martma, T., 1997. Carbon isotope event markers through the Wenlock – Pridoli sequence in Ohesaare (Estonia) and Priekule (Latvia). Palaeogeogr. Palaeoclimatol. Palaeoecol. 132, 211 – 224. Kaljo, D., Kiipli, T., Martma, T., 1998. Correlation of carbon isotope events and environmental cyclicity in the East Baltic Silurian. In: Landing, E., Johnson, M.E. (Eds.), Silurian cycles— linkages of dynamic stratigraphy with atmospheric, oceanic and tectonic changes. New York State Mus. Bull., vol. 491, pp. 297 – 312. Kaljo, D., Hints, L., Hints, O., Martma, T., No˜lvak, J., 1999. Carbon isotope excursions and coeval biotic-environmental changes in the late Caradoc and Ashgill of Estonia. In: Kraft, P., Fatka, O. (Eds.), Quo vadis Ordovician? Acta Univ. Carolinae, Geol., vol. 43, pp. 507 – 510. Kaljo, D., Hints, L., Martma, T., No˜lvak, J., 2001. Carbon isotope stratigraphy in the latest Ordovician of Estonia. Chem. Geol. 175, 49 – 59. Kump, L.R., Arthur, M.A., Patzkowski, M.E., Gibbs, M.T., Pinkus, D.S., Sheehan, P.M., 1999. A weathering hypothesis for glaciation at high atmospheric pCO2 during the Late Ordovician glaciation. Palaeogeogr. Palaeoclimatol. Palaeoecol. 152, 173 – 187. Leslie, S.A., Bergstro¨m, S.M., 1995. Revision of the North American late Middle Ordovician standard stage classification and timing of the Trenton transgression based on K-bentonite bed correlation. In: Cooper, J.D., et al. (Ed.), Ordovician Odyssey: short papers for the Seventh Intern. Symp. on the Ordovician System. Pacific Section, Society for Sedimentary Geology, Fullerton, CA, pp. 49 – 54. Ludvigson, G.A., Jacobson, S.R., Witzke, B.J., Gonzalez, L.A., 1996. Carbonate component chemostratigraphy and depositional history of the Ordovician Decorah Formation, Upper Mississippi Valley. In: Witzke, B.J., Ludvigson, G.A., Day, J. (Eds.), Paleozoic Sequence Stratigraphy: Views from the North American Craton. Geol. Soc. Am. Special Paper, vol. 306, pp. 67 – 86. Ludvigson, G.A., Witzke, B.J., Gonzalez, L.A., Carpenter, S.J., Hasiuk, F.J., 2001. New complexities in the expression of Late Ordovician carbon isotope excursions straddling the Turinian – Chatfieldian boundary interval. Geol. Soc. Amer. Annual Meeting, Nov. 5 – 8, 2001. Final Program, Abstract No 27487. Ma¨nnil, R., 1960. The stratigraphy of the Oandu (‘‘Vasalemma’’) Stage. Eesti NSV Tead. Akad. Geol. Inst. Uurimused 5, 89 – 122 (In Russian, with English Abstr).

Ma¨nnil, R., 1966. Evolution of the Baltic Basin during the Ordovician Valgus, Tallinn. In Russian, with English Abstr. Ma¨nnil, R., 1990. The Ordovician of Estonia. In: Kaljo, D., Nestor, H. (Eds.), Field Meeting Estonia 1990: An Excursion Guidebook. Estonian Academy of Sciences, Tallinn, pp. 11 – 20. Ma¨nnil, R., Meidla, T., 1994. A. The Ordovician System of the East European Platform (Estonia, Latvia, Lithuania, Byelorussia, parts of Russia, the Ukraine and Moldova). In: Webby, B.D., Ross, R.J., Zhen, Y.Y. (Eds.), The Ordovician System of the East European Platform and Tuva (southeastern Russia), vol. 28, A IUGS Publ., Boulder, CO, pp. 1 – 52. Marshall, J.D., 1992. Climatic and oceanographic isotopic signals from the carbonate rock record and their preservation. Geol. Mag. 129, 143 – 160. Marshall, J.D., Brenchley, P.J., Mason, P., Wolff, G.A., Astini, R.A., Hints, L., Meidla, T., 1997. Global carbon isotopic events associated with mass extinction and glaciation in the late Ordovician. Palaeogeogr. Palaeoclimatol. Palaeoecol. 132, 195 – 210. Meidla, T., 1996. Late Ordovician ostracodes of Estonia. Fossilia Baltica, vol. 2. Tartu Univ. Press, Tartu. 222 pp. Min, K., Renne, P.R., Huff, W.D., 2001. 40Ar/39Ar dating of Ordovician K-bentonites in Laurentia and Baltoscandia. Earth Planet. Sci. Let. 185, 121 – 134. Morrow, J.R., Schindler, E., Walliser, O.H., 1995. Phanerozoic development of selected global environmental features. In: Walliser, O.H. (Ed.), Global Events and Event Stratigraphy in the Phanerozoic. Springer, Berlin-Heidelberg, pp. 53 – 61. Nestor, H., Einasto, R., 1997. Ordovician and Silurian carbonate sedimentation basin. In: Raukas, A., Teeduma¨e, A. (Eds.), Geology and Mineral Resources of Estonia. Estonian Academic, Tallinn, pp. 192 – 204. No˜lvak, J. 1997. Ordovician. Introduction. In: Raukas, A., Teeduma¨e, A. (Eds.), Geology and Mineral Resources of Estonia. Tallinn, Estonian Academic Publishers, pp. 52, 54, 55. No˜lvak, J., 1999. Ordovician chitinozoan biozonation of Baltoscandia. In: Kraft, P., Fatka, O. (Eds.), Quo vadis Ordovician? Acta Univ. Carol., Geol., vol. 43, pp. 287 – 290. No˜lvak, J., Grahn, Y., 1993. Ordovician chitinozoan zones from Baltoscandia. Rev. Palaeobot. Palynol. 79, 245 – 269. Oraspo˜ld, A., 1982a. Lithology of the Vormsi Stage in Estonia. Acta et Comment. Univ. Tartuensis, 527. To¨id Geoloogia Alalt, vol. VIII, pp. 51 – 74. In Russian, with English Abstr. Oraspo˜ld, A., 1982b. Lithological characterisation of the boundary beds of the Vormsi and Pirgu stages in Central Estonia. Acta et Comment. Univ. Tartuensis, 527. To¨id Geoloogia Alalt, vol. VIII, pp. 75 – 90. In Russian, with English Abstr. Pancost, R.D., Freeman, K.H., Patzkowsky, M.E., 1999. Organicmatter source variation and the expression of late Middle Ordovician carbon isotope excursion. Geology 27, 1015 – 1018. Paris, F., Vernier, J., Achab, A., Albani, R., Ancilletta, A., Asselin, E., Chen, X., Fatka, O., Grahn, Y., Molyneux, S., Nolvak, J., Samuelsson, J., Sennikov, N.V., Soufiane, A., Wang, X., Winchester-Seeto, T., 1999. Correlation of Ordovician regional chitinozoan biozonations. In: Kraft, P., Fatka, O. (Eds.), Quo vadis Ordovician? Acta Univ. Carol., Geol., vol. 43, pp. 291 – 294. Patzkowsky, M.E., Slupik, L.M., Arthur, M.A., Pancost, R.D., Freeman, K.H., 1997. Late Middle Ordovician environmental

D. Kaljo et al. / Palaeogeography, Palaeoclimatology, Palaeoecology 210 (2004) 165–185 change and extinction: Harbinger of the Late Ordovician or continuation of Cambrian patterns? Geology 25, 911 – 914. Po˜lma, L., 1972. Skeletal debris content and composition in the sediments of the northern facial belt of the East Baltic Ordovician (Rapla borning). Eesti NSV Tead. Akad. Toim., Keem. Geol. 21, 326 – 332 (In Russian, with English Abstr.). Po˜lma, L., 1982. Comparative Lithology of the Ordovician Carbonate Rocks in the Northern and Middle East Baltic Valgus, Tallinn. In Russian, with English Abstr. Po˜lma, L., Sarv, L., Hints, L., 1988. Lithology and Fauna of the Type Sections of the Caradoc Series in Northern Estonia Valgus, Tallinn. In Russian, with English Abstr. Rong, J.-Y., Chen, X., Harper, D.A.T., Mitchell, C.E., 1999. Proposal of a GSSP candidate section in the Yangtze Platform region, S. China, for a new Hirnantian boundary stratotype. In: Kraft, P., Fatka, O. (Eds.), Quo vadis Ordovician? Acta Univ. Carol., Geol., vol. 43, pp. 77 – 80. Samtleben, C., Munnecke, A., Bickert, T., Pa¨tzold, J., 1996. The Silurian of Gotland (Sweden): facies interpretation based on stable isotopes in brachiopod shells. Geol. Rundsch. 85, 278 – 292. Torsvik, T.H., Smethurst, M.A., Meert, J.G., Van der Voo, R.,

185

McKerrow, W.S., Brasier, M.D., Sturt, B.A., Walderh aug, H.J., 1996. Continental break-up and collision in the Neoproterozoic and Palaeozoic—a tale of Baltica and Laurentia. Earth Sci. Rev. 40, 229 – 258. Tucker, R.D., McKerrow, W.S., 1995. Early Paleozoic chronology: a review in light of new U – Pb zircon ages from Newfoundland and Britain. Can. J. Earth Sci. 32, 368 – 379. Tuuling, I., 1998. Shipborne geophysical study of an Ordovician – Silurian carbonate platform, Fa˚ro¨ – Hiiumaa area, northeastern Baltic Sea. Medd. Stockh. Univ. Inst. Geol. Geokem. 301, 9 – 11. Underwood, C.J., Crowley, S.F., Marshall, J.D., Brenchley, P.J., 1997. High-resolution carbon isotope stratigraphy of the basal Silurian Stratotype (Dob’s Linn, Scotland) and its global correlation. J. Geol. Soc. (Lond.) 154, 709 – 718. Viira, V., Ma¨nnik, P., 1997. Conodonts. In: Raukas, A., Teeduma¨e, A. (Eds.), Geology and Mineral Resources of Estonia. Estonian Academic, Tallinn, pp. 241 – 244. Vingisaar, P., 1972. On the distribution of the main metabentonite stratum (d; XXII) in the Middle Ordovician of Baltoscandia. Eesti NSV Tead. Akad. Toim., Keem. Geol. 21, 62 – 70 (In Russian, with English Abstr.).