Neodymium isotopic composition of Ordovician conodonts as a ...

Geochemistry Geophysics Geosystems

3

G

Article Volume 3, Number 2 23 February 2002 10.1029/2001GC000195

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES Published by AGU and the Geochemical Society

ISSN: 1525-2027

Neodymium isotopic composition of Ordovician conodonts as a seawater proxy: Testing paleogeography Cynthia A. Wright and Christopher R. Barnes Centre for Earth and Ocean Research, University of Victoria, Victoria, B. C., V8W 3P6, Canada ([email protected])

Stein B. Jacobsen Department of Earth and Planetary Sciences, Harvard University, 20 Oxford Street, Cambridge, Massachusetts, 02138 USA ( [email protected])

[1] To evaluate recent Ordovician paleogeographic reconstructions, eNd values were determined in conodonts as a proxy for the isotopic variations in ancient seawater using samples from the major cratons and microplates. Isotopic variations reveal the existence of distinct oceanic masses and epeiric seas that constrain the position of plates and terranes relative to these water masses and hence constrain models of regional and global paleogeography. The isotopic patterns show a consistent picture of the changes in the Ordovician oceans. The eNd values for water masses associated with Laurentia are strongly negative ( 28 to 18) in the Early Ordovician, evolving over time to higher values in the range 13 to 5. The Early Ordovician signature of Laurentia is in marked contrast to other cratons and microplates, which have a range of values from 10 to 5. Samples from South China show interesting signals that reflect a greater similarity to Laurentia than other peri-Gondwana terranes. The isotopic variations are a function of both regional geology and global tectonic processes, the most obvious being the Taconic Orogeny and onset of the closure of the Iapetus Ocean. Regional and global models of paleogeography are considered in light of these proxy signals. This study also reveals that conodonts are powerful geochemical tools for obtaining information on ancient water masses. The use of the Nd isotopic signatures from conodonts provides a method independent of paleomagnetism and biogeography to test paleogeographic models. By integrating such information, a more multidisciplinary approach is possible. Components: 9581 words, 5 figures, 2 tables, 1 data set. Keywords: Ordovician; neodymium; paleogeography; paleoceanography; conodonts; Paleozoic. Index Terms: 3030 Marine Geology and Geophysics: Micropaleontology; 9614 Information Related to Geologic Time: Paleozoic; 1010 Geochemistry: Chemical evolution; 1040 Geochemistry: Isotopic composition/chemistry. Received 18 June 2001; Revised 26 October 2001; Accepted 15 November 2001; Published 23 February 2002. Wright, C. A., C. R. Barnes, and S. B. Jacobsen, Neodymium isotopic composition of Ordovician conodonts as a seawater proxy: Testing paleogeography, Geochem. Geophys. Geosyst., 3(2), 10.1029/2001GC000195, 2002.

——————————— Theme: Geochemical Earth Reference Model (GERM) Guest Editor: Hubert Staudigel

Copyright 2002 by the American Geophysical Union.

1 of 17


3

G

wright et al.: isotopic composition

1. Introduction [2] Variations in the 143Nd/144Nd ratio due to the a-decay of 147Sm provides a powerful tracer to constrain the continental sources of dissolved and suspended materials transported to the oceans by rivers [Goldstein and Jacobsen, 1987]. On the basis of their short oceanic residence times, long halflife, and tight coupling behavior, Sm and Nd isotopes are useful for studying the geochemical evolution of the oceans and for tracking water masses by imparting a distinct 143Nd/144Nd and eNd value on the water column [Piepgras and Wasserburg, 1980, and references therein]. [3] Piepgras and Wasserburg [1980] assessed eNd measurement as a tracer in modern oceans and suggested that it may also be applicable for ancient oceans. Subsequent studies have shown that the Nd isotopic signatures from authigenic, detrital, and biogenic minerals from marine sediments can be used to track overlying water masses and also for tracking ancient seawater as the Nd isotopic signature of the water mass equilibrates with and is incorporated into the mineral matrix [Shaw and Wasserburg, 1985; Staudigel et al., 1985; Keto and Jacobsen, 1987, 1988; Grandjean et al., 1987; Holmden et al., 1996]. This has important implications for studying paleogeography, paleobiogeography, and paleoceanography. Repositories of these ancient isotopic signals include nodules, fossils, and phosphatic deposits [Grandjean et al., 1987]. [4] For recent biogenic material, the Nd signal is attained after death of the parent animal, becoming incorporated into the mineral lattice [Shaw and Wasserburg, 1985; Wright et al., 1984]. This signature is an isotopic fingerprint of the bottom waters at the time of death. Studies have shown that biogenic minerals give reliable and consistent proxy signals for contemporaneous water masses at the time of deposition. The fingerprint is picked up from the overlying water mass during burial and early diagenesis and is not a product of some diagenetic overprint [Wright et al., 1984; Shaw and Wasserburg, 1985; Grandjean et al., 1987; Holmden et al., 1996; Martin and Haley, 2000].

10.1029/2001GC000195

[5] Conodonts are the hard tissue dental apparatus of an extinct group of marine chordates that lived from the Late Cambrian to the Late Triassic. They are composed of hydroxy-apatite and, as such, represent sites for Sm and Nd incorporation. This enrichment and long-term mineral stability makes them a valuable proxy tool in establishing geochemical signals for ancient seawater. Conodonts have proven to be one of the most useful biostratigraphic groups for the Paleozoic-early Mesozoic, particularly the Ordovician where they experienced a dramatic evolutionary radiation. Ordovician stratigraphy is finely resolved on the basis of conodont biozonation. [6] For the early Paleozoic, lithological, geochemical, and paleontological data are all utilized to interpret ancient environments. Sedimentary analysis, stratigraphic correlation, paleomagnetic and isotopic signatures, and tectonic events assist in constraining paleogeographic patterns. However, tectonic deformation can make stratigraphic correlation difficult, and paleomagnetism, a dominant method for determining paleogeography, is also limited by overprinting from orogenic events and its inability to establish paleolongitude. To counter these limitations, the fossil record and faunal biogeography are often integrated with paleomagnetic results. However, the fossil record can also suffer from taphonomic and sampling biases. [7] The principle goal of this study is to use the Nd isotopic composition of conodonts as a proxy for recognizing ancient oceans and to test various proposed models of early Paleozoic paleogeography. The pattern of Nd isotopes from conodonts cannot provide direct evidence for paleogeographic positioning, but it can identify separate major ocean masses and indicate the relative juxtaposition of cratons and microplates through time. There are several paleogeographic models that reconstruct the position of the continents and ocean masses for the Ordovician, either on a global or regional basis. To date, the reconstructions of Scotese and McKerrow [1991] and Scotese [1997, 1999] are the most widely accepted, comprising global reconstructions based on paleomagnetic, paleoclimatic, faunal biogeographic, and tectonic data. 2 of 17


3

G


10.1029/2001GC000195

Table 1. Conodont Sample Information Sample (Number of Elements)

Location

Stratigraphic Unit

Stage or Biozone

Agec (Ma)

Chronostratigraphy

Sweden Australia Kazakhstan Newfoundland Alberta Newfoundland

Tremadoc Ceratopyge Lmst. C. angulatusb Ninmaroo Fm. C. proavus/H. simplexb Shabakty Fm. Euloma-Leiostegium Green Point Fm. C. lindstromib Survey Peak Fm. C. angulatusb Boat Harbour Fm. C. angulatusb

Tremadoc Tremadoc Tremadoc Tremadoc Tremadoc Tremadoc

494 510 508 505 505 500

57 [8] 56 [11] 47 [10] 33 [10] 34 [10] 2 [10] 48 [10] 53 [10] 44 [10] 50 [10] 45 [9] 7 [10] 26 [10] 46 [10] 51 [10] 14 [10] 15 [6] 60 [10]

Turkey Turkey Estonia Sweden Sweden Australia British Columbia Newfoundland Newfoundland Alberta Newfoundland Oklahoma Quebec Newfoundland Alberta Newfoundland Newfoundland Central China

Arenig-Llanvirn Seydisehir Fm. Note a Sobova Lmst. D. hirundob Leetse Fm. Latorp Stg. B1b Lanna Lmst. B. navisb Segerstad Lmst. E. suecicusb Larapinta Gp. Upper Whiterockianb McKay Gp. Conodont Fauna C/Db Catoche Fm. P. carlae/S. ovatusb Green Point Fm. D. bifidus (bifidus)b Outram Fm. O. communisa Green Point Fm. I. v. maximusb Oil Creek Fm. Conodont Fauna 4b Le´vis Fm. D. dentatusb Green Point Fm. U. austrodentatusb Skoki Fm. Note a Table Head Gp. P. tentaculatusb Table Head Gp. P. tentaculatusb Kunitan Fm. P. serrusb

Arenig Arenig Arenig Arenig L. Llanvirn L. Llanvirn Arenig Arenig Arenig Arenig Arenig Arenig Arenig L. Llanvirn L. Llanvirn L. Llanvirn L. Llanvirn L. Llanvirn

482 477 480 478 472 472 491 485 485 485 478 477 476 476 473 472 472 472

30 [5] 36 [10] 6 [10] 13 [10] 55 [10] 39 [10] 61 [10]

Wales Manitoulin Island Colorado Ontario Argentina Australia Central China

Llandeilo-Caradoc Ffairfach Gp. A. inaequalisb Gull River Fm. Blackeriverana Harding Sst. Rocklandiana Whitby Fm. Maysvilliana Empozada Fm. A. superbusb Clearview Lmst. Brachiopod Fauna C/Db Baota Fm. Note a

M. Llandeilo Caradoc (Costonian) Caradoc (Soudleyan) Caradoc (Onnian) Caradoc (Marshbrook) Caradoc (Onnian) Caradoc (Marshbrook)

467 463 456 444 446 444 447

22 [12] 23 [10] 43 [10] 16 [10] 21 [7] 29 [10] 31 [9] 9 [10] 10 [10] 5 [10] 11 [9] 27 [10] 8 [10]

Spain Italy Germany Wales Sardinia Wales Norway Manitoba Manitoba Devon Island Manitoulin Island Southampton Anticosti Island

Ashgill Urbana Lmst. A. ordovicicusb Tonflaserkalk Lmst. A. ordovicicusb Kalkbalk Lmst. A. ordovicicusb Shoalshook Fm. A. ordovicicusb Dosmusnovas Fm. A. ordovicicusb Cruˆg Fm. A. ordovicicusb Gamme Fm. Note a Stony Mt. Fm. upper Richmondiana Stony Mt. Fm. upper Richmondiana Irene Bay Fm. mid-Maysvillea Georgian Bay Fm. upper Richmondiana Churchill River Gp. Richmondiana Ellis Bay Fm. Conodont Fauna 13b

Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill Ashgill

443 443 443 443 443 442 441 441 441 442 441 441 440

12 17 41 18

Manitoulin Island Anticosti Anticosti Island Siberia

Silurian Manitoulin Fm. Llandoverya Ellis Bay Fm. B. plicatispinaeb Bescie Fm. E. birminghamensisb Unit R/Member 74 A. acuminatusb

Llandovery (Rhuddanian) Ashgill (Hirnantian) Llandovery (Rhuddanian) Llandovery (Rhuddanian)

35 38 62 32 49 52

[10] [10] [10] [10] [10] [10]

[10] [7] [9] [10]

(Pusgillian) (Pusgillian) (Pusgillian) (Pusgillian) (Pusgillian) (Pusgillian) (Cautleyan) (Pusgillian) (Pusgillian) (Pusgillian) (Pusgillian) (Pusgillian) (Hirnantian)

438 440 437 438

a Unknown b c

or no biozone present, correlated on other stratigraphic information. Biozone present. Harland et al. [1989].

3 of 17


3

G


10.1029/2001GC000195

[8] According to this model, the major oceans surrounding the landmasses during the Ordovician were the Panthalassa, the Iapetus, and the Paleotethys. The major cratons at the time were Laurentia (North American craton, Greenland, Arctic Canada, northern Scotland), Gondwana (South America, Africa, Australia, New Zealand, east Antarctica, south central Europe, north and south China, plus other suspect terranes of the Middle East), Baltica (Balto-Scandia, East Russian Platform), Siberia, and Kazakhstan. In addition to the major cratonic landmasses, there were also a series of displaced peri-Gondwana terranes that included Avalonia, Meguma, Armorica, and Bohemia (sensu Van der Voo [1988]). Other than Avalonia, these terranes are not well constrained.

chosen for this study were carefully selected for their lack of adhering carbonate and/or clay debris. They were also chosen for having a conodont color alteration index of 5; most were CAI 2 1/2. Each sample consists of 10–15 conodont elements (10–100 mg total weight for each sample) of the same genus and, for nearly all samples, the same species. Sixtythree samples were chosen from four time slices through the Ordovician Period and earliest Silurian. They were also selected for their location in an effort to sample the major cratons and microplates and also for particular environmental gradients. One sample, 15, was an inarticulate brachiopod included for comparative purposes.

[9] The primary use herein of the Scotese [1999] model, however, does not preclude other models from being tested. Most paleogeographic models to date have been broadly similar pattern, but the model of Dalziel et al. [1994] differs considerably in the positioning of Gondwana in relation to Laurentia. A third paleogeographic model that differs slightly from the others is Paris and Robardet’s [1990] reconstruction of the Mediterranean region, which affects the location of the Avalonia microplate and the location of adjacent oceans.

2.2. Conodont Age Determinations

[10] Over 60 conodont samples were chosen to represent the Ordovician through to the earliest Silurian. These samples represent most of the major paleoplates for this interval. The data contained in this study are a substantial addition to the small existing Nd database for the early Paleozoic [Keto and Jacobsen, 1987, 1988; Holmden et al., 1996]. It is the first approach using conodonts on a global scale.

2. Samples and Methods 2.1. Conodont Sampling [11] A description of each sample is presented in Table 1. Supplemental data and also Wright [1995] provide detailed sample location information, taxonomomy, CAI values, and references. Conodont extraction procedures were as per Barnes et al. [1987]. The conodont elements

[12] The amount and accuracy of information varied between samples and not all could be dated with equal precision. On the basis of the literature and the personal knowledge of the collections, samples were placed into appropriate biozones (graptolite and/or conodont) and crosscorrelated into the chronostratigraphic timescale of Harland et al. [1990]. However, this timescale favors the British series and stages, and it was commonly necessary to correlate the biozone information with other regional stratigraphic timescales (e.g., International Union of Geological Sciences (IUGS) and Decade of North American Geology (DNAG)). The resulting age estimates are given in Table 1. This method allows the determination of absolute ages but does not permit the easy assignment of age errors. If errors need to be considered, then those associated with the biozone on the Harland et al. [1990] scale can be invoked. However, the sizes of the biozone errors are negligible since they would not have a significant effect on the final Nd isotopic values.

2.3. Analytical Methods and Results [13] Conodont elements were removed from their mounts and thoroughly washed 3 times in ultra pure water and allowed to air dry. The samples were then dissolved in 1 mL of 4N HCl. 4 of 17


3

G

10.1029/2001GC000195


Table 2. Neodymium Isotopic Results for Ordovician - Silurian Conodont Apatite 147

eNd(0)c

eNd(T)

±2s

18 99 43 12 22 14

10.14 15.14 14.36 23.13 25.91 24.93

7.14 9.47 10.27 17.69 20.82 18.86

0.35 1.94 0.84 0.24 0.43 0.27

Arenig-Llanvirn 0.1219 0.511324 0.1660 0.511256 0.1398 0.511314 0.1420 0.511279 0.1304 0.511192 0.2791 0.511520 0.0876 0.510651 0.0943 0.510359 0.1066 0.510806 0.0962 0.510660 0.1089 0.510867 0.1409 0.510765 0.1212 0.510931 0.1178 0.510904 0.1056 0.510968 0.1213 0.510189 0.1250 0.510899 0.1589 0.511024

15 39 15 49 23 19 17 12 19 31 44 26 52 116 36 90 19 14

10.22 11.55 10.41 11.10 12.80 6.39 23.37 29.07 20.34 23.19 19.15 21.14 17.90 18.43 17.17 32.40 18.52 16.08

5.62 9.68 6.93 7.76 8.80 11.36 16.53 22.73 14.76 16.97 13.78 17.74 13.31 13.63 11.67 27.85 14.20 13.8

0.29 0.76 0.29 0.96 0.45 0.37 0.33 0.24 0.37 0.61 0.86 0.51 1.02 2.27 0.70 1.76 0.37 0.27

9.37 84.9 187.4 10.2 35.6 5.16 6.68

Llandeilo-Caradoc 0.1274 0.511312 0.1793 0.510758 0.1946 0.510894 0.1366 0.511158 0.1293 0.511349 0.1234 0.511354 0.1316 0.511074

30 13 13 29 13 30 26

10.45 21.28 18.62 13.46 9.73 9.63 15.10

6.32 20.25 18.50 10.05 5.89 5.49 11.39

0.59 0.25 0.25 0.57 0.25 0.59 0.51

Armorica Armorica Armorica Avalonia Avalonia Avalonia Baltica Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia

50.4 11.7 20.3 9.21 72.1 13.7 16.1 45.9 293.0 124.0 13.0 46.4 24.5

Ashgill 0.1583 0.1490 0.2090 0.2837 0.1827 0.1779 0.1339 0.1119 0.1055 0.0985 0.1120 0.1312 0.1300

0.511352 0.511297 0.511433 0.511569 0.511565 0.511405 0.511162 0.511301 0.511259 0.511070 0.511170 0.510832 0.511090

17 44 13 67 13 15 13 90 41 16 21 28 23

9.67 10.75 8.09 5.43 5.51 8.64 13.38 10.67 11.49 15.18 13.23 19.83 14.79

7.50 8.05 8.78 10.35 4.72 7.57 9.85 5.89 6.35 9.64 8.46 16.14 11.04

0.33 0.86 0.25 1.31 0.25 0.29 0.25 1.76 0.80 0.31 0.41 0.55 0.45

Laurentia Laurentia Laurentia Siberia

22.3 24.8 21.5 20.6

Silurian 0.1065 0.1178 0.1079 0.1291

0.511283 0.511046 0.510992 0.511593

46 18 14 38

11.02 15.65 16.71 4.96

5.97 11.24 11.75 1.18

0.90 0.35 0.27 0.74

Sm/ Nda

143

Sample Number

CratonPaleoplate

Nd, ng

35 38 62 32 49 52

Baltica Gondwana Kazakhstan Laurentia Laurentia Laurentia

89.4 5.55 10.4 35.7 158 14.9

Tremadoc 0.1492 0.1097 0.1337 0.1123 0.1178 0.1016

0.511328 0.511072 0.511112 0.510663 0.510521 0.510571

57 56 47 33 34 2 48 53 44 50 45 7 26 46 51 14 15 60

Armorica Armorica Baltica Baltica Baltica Gondwana Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia Laurentia South China

6.90 17.1 406.0 20.2 17.0 37.7 14.3 35.0 9.79 4.91 4.06 50.3 25.6 12.6 11.4 47.6 151 22.0

30 36 6 13 55 39 61

Avalonia Laurentia Laurentia Laurentia Gondwana Gondwana South China

22 23 43 2 21 29 31 9 10 5 11 27 8 12 17 41 18

144

144

Nd/ Ndb

±2s 10

a Uncertainty b c

is 0.5%. Data normalized to 146Nd/144Nd = 0.724134. Deviations in parts 104 from present day CHUR value of

143

Nd/144Nd = 0.511847 [Jacobsen and Wasserburg, 1984].

5 of 17


3

G


10.1029/2001GC000195

Panthalassa Ocean

-9.5

-10.7

Panthalassa Ocean

Kazakhstan -10.3

-20.8

Siberia Laurentia

northern Paleotethys Ocean

-10.7

Iapetus Ocean -17.7, -18.9

-7.1

Baltica Gondwana

southern Paleotethys Ocean

Gondwana

Figure 1. Paleogeography and eNd variations for samples from the Tremadoc Interval (510 – 493 Ma). Map taken from Scotese [1999]. Kazakhstan island arcs added. The eNd values shown in ovals are comparative values from the literature from other crustal sources or conodonts. See text for details.

[14] The chemical separation methods for Sm and Nd are the same as those used previously by Keto and Jacobsen [1987, 1988] for brachiopods and conodonts [see, also, Jacobsen and Dymek, 1988]. The total procedural blank for the conodont measurements was in the range of 10–30 pg for Nd and for Sm it was in the range of 2–6 pg. [15] The Nd isotopic compositions were measured using a Finnigan Mat 262 multicollector mass spectrometer in static mode. The Sm isotopic compositions were measured by ion counting on a Finnigan MAT THQ Thermal Quadrupole Mass Spectrometer operated in peak jumping mode. Sm-Nd ratios are accurate to better than 0.5%. The Nd isotopic ratios were corrected for mass fractionation using the exponential law and using 146Nd/144Nd = 0.724134 [cf. Wasserburg et al., 1981]. The value obtained for the Caltech Nd-beta standard with these procedures was 143Nd/144Nd = 0.511095 with an uncertainty of 0.000005. [16] The Nd and Sm isotopic compositions are also reported and represented as parts per 104 deviations from the present chondritic uniform reservoir (CHUR) value: 143Nd/144Nd CHUR(0) = 0.511847

[Jacobsen and Wasserburg, 1980]. This parameter together with the estimated stratigraphic age of the samples is used to calculate the initial Nd isotopic compositions, eNd(T ), relative to the CHUR evolution curve. [17] Table 2 lists the results of the Nd isotopic analysis for 48 samples. The 143Nd/144Nd values range from 0.510359 to 0.511593 and the eNd values, overall and through time, range from 29.1 to 5.0. The uncertainties for most of the samples are within 1 e-unit, while only four have errors between 1 and 2 e-units.

3. Discussion 3.1. Nd Isotopic Values and Their Relation to Sedimentary Sources [18] For a comprehensive analysis, it is important to discuss the Nd isotopic values in the context of regional geological associations, as well as to consider the importance of erosional source material to the fossil signature. Only a few studies have tried to assess the isotopic composition of seawater for this interval [Keto and Jacobsen, 1987, 1988]. Most studies have focused on using the Nd isotopic 6 of 17


3

G


10.1029/2001GC000195

-11.4 -15.7, -18.4

-11.7, -17.0 -16.5 -17.4 -13.8

-11, -17

-27.9 -13.3 -13.6 to -22.7

-6.9 to -8.8 -9.7

-5.6

Amorica Avalonia

Figure 2. Paleogeography and eNd variations for samples from the Arenig-Llanvirn Interval (492 – 469 Ma). Map taken from Scotese [1999]. Kazakhstan island arcs added. The eNd values shown in ovals are comparative values from the literature from other crustal sources or conodonts. See text for details.

composition of the sediment to quantify provenance sources.

3.2. Tremadoc Interval [19] The data for this interval (510–493 Ma) are shown in Figure 1. Globally, eNd values range from 20.8 to 7.1 with the most negative values being associated with the marginal oceans of the Laurentian craton. These nonradiogenic signals from Laurentia are likely a result of the averaging of various Precambrian provinces, which at certain intervals of the Early Paleozoic would have been exposed to erosion. [20] The sample from Baltica ( 7.1) shows a substantially different signature than Laurentia. Baltica is generally composed of Archean continental crust with Svecofennian and Grenville aged material and also various volcanic terranes with mid-oceanic ridge basalt (MORB) characteristics. Goldstein and Jacobsen [1987] noted that orogenic areas might erode faster than stable older cratonic regions leading to isotopically younger rocks being eroded faster yielding higher eNd values. [21] Kazakhstan shows a slightly more negative signature but with a younger crustal source than

Laurentia. A single previous eNd value from conodonts exists for Kazakhstan, also from the Batyrby section. Keto and Jacobsen [1988] report a eNd of 10.7, identical to the value obtained in this study. [22] The sample from Australia, representing equatorial Gondwana, shows a value similar to that of Kazakhstan. Sediments from the neighboring Amadeus Basin show a similar with an eNd value of 9.7 [Zhao et al., 1992]. Their study showed that the major source of sediments to the Amadeus Basin was linked to the surrounding igneous rocks, the Arunta and the Musgrave Blocks, reflecting a close regional source for sedimentary material rather than being predominately cratonic material [Zhao et al., 1992].

3.3. Arenig-Llanvirn Interval [23] For this time interval (492–469 Ma; Figure 2), the global range of eNd for seawater is from 27.8 to 5.6. Marginal ocean masses for the Laurentian craton begin to show a shift toward more radiogenic eNd values. The western margin of Laurentia bordering the Panthalassa Ocean has slightly more negative signals, with input remaining predominately from older Precambrian continental crust. 7 of 17


3

G

10.1029/2001GC000195


-5.5

-18.5

-5.9

-20.3 -14.7, -16.8 -6, -10.1 -9 -19.6 -6, -10

-4, -10

-11.4

-6.3

Figure 3. Paleogeography and eNd variations for samples from the Llandeilo-Caradoc Interval (468 – 444 Ma). Map taken from Scotese [1999]. Kazakhstan island arcs added. The eNd values shown in ovals are comparative values from the literature from other crustal sources or conodonts. See text for details.

The least negative value of 11.7 is the shallow water Skoki Formation from Alberta. The neighboring values ( 17 (Outram Formation, Alberta); 16.5 (McKay Group, British Columbia)) are slightly older, deeper water deposits, and the more radiogenic signal displayed by the Skoki could be the response to different input sources. Recent evidence suggests that during the early Paleozoic, the present-day western Cordillera was subjected to three phases of extension and rifting [Cecile et al., 1997]. [24] The Iapetus margin of Laurentia shows a narrow range of eNd values from 14.8 to 13.3. The strongly negative value of 22.7, from the Catoche Formation of Western Newfoundland, represents an outer platform facies at the onset of a regional transgressive phase. The coeval Cow Head Group at St. Paul’s Inlet represents a deeper slope facies and shows a western Iapetus Ocean signature of approximately 14. Additionally, as reefal deposits built up at the shelf-break, the degree of circulation between the platform and the outer slope would have decreased [James et al., 1989] resulting in two isotopically distinct water masses. The younger Table Head Group represents a collapsing continental margin and progressive drowning. It has a similar eNd value as those of the slope

facies of the Cow Head Group and the Le´vis Formation. [25] Other Nd isotopic data for Laurentia are sparse. Boghossian et al. [1996] sampled the Survey Peak, Glenogle, and Mount Wilson sedimentary rocks (Western Laurentia) and found eNd isotopic values of 18.4, 15.7, and 15.8, respectively. They suggest these values are the result of the source materials being derived from proximal Precambrian basements. For eastern Laurentia, Miller and O’Nions [1984] reported Nd isotopic values from early Arenig aged sediments from Quebec with values of 17 and 11. [26] The eNd values for Baltica ( 6.9 and 8.8) are consistent with those from the Tremadoc. The eNd values for the Armorican microplate are similar to eNd values of 9.7 and 5.6. Sediment sources for the Armorican microplate have been linked to the Cadomian-Pan-African basement of Gondwana [Noblet and Lefort, 1990]. [27] The sample from South China shows an eNd signal of 13.8 and suggests input from an older continental source than that contributing to the southern Paleotethys (e.g., Armorica) or the eastern Iapetus waters off Baltica. The rocks of the Yangtze 8 of 17


3

G


10.1029/2001GC000195

-1.2 -5.9 -16.1 -6.4

-9.6

-6.0 -11.0 to -11.8

-6

-8.5

-9.9 -7.6 -7, -8

-4.7 -7.5

-8.1

-8.8

Figure 4. Paleogeography and eNd variations for samples from the Ashgill-early Silurian Interval (443 –436 Ma). Map taken from Scotese [1999]. Kazakhstan island arcs added. Kazakhstan island arcs added. The eNd values shown in ovals are comparative values from the literature from other crustal sources or conodonts. See text for details.

Platform are predominately Late Archean and Paleoproterozoic with surrounding younger orogenic belts.

3.4. Llandeilo-Caradoc [28] Sampling distribution for this time interval (468 – 444 Ma; Figure 3) is limited. Laurentia shows a rather large range of eNd values, from 10.1 to 20.3. This can be explained by regional variation in depositional environments. The Gull River Formation (eNd = 20.3), in Ontario, represents a lagoonal sequence and interaction with the outer platform and open ocean was restricted. The Harding Sandstone (eNd = 18.5), in Colorado, represents a shoreline transgressive sequence with the major sediment source being dominated by comparatively older continental crust. These values are distinctly different from the Whitby Formation (eNd = 10.1) in Ontario, which represents the initial phase of collapsed foreland basin that allowed an influx of open Iapetus waters. Sediment sources for Whitby included the uplifts associated with the Taconic Orogeny to the east. [29] Comparative values in the literature from this time interval include a sample from Tennessee with an eNd value of approximately 6 [Shaw and Wasserburg, 1985]. This is slightly more radiogenic

than values from Tennessee conodonts from east of the Saltzville Fault (Blockhouse Shale), which yielded a value of 8, and a younger sample taken west of the fault (Holston Fm.) that yielded a value of 9 [Keto and Jacobsen, 1987]. Other interpreted values for Laurentia include 6.8 (Oklahoma) and 19.6 (Pennsylvania) [Keto and Jacobsen, 1987]. Holmden et al. [1996] using a conodont sample from below and above the Millbrig and Deicke Kbentonites in Iowa found a Nd signal of 14.7. [30] Clastic sediments from the Taconic and Sevier Basins were analyzed for eNd by Andersen and Samson [1995]. The Taconic Basin showed values from 12.6 to 8.5 and those from the Sevier Basin had a range of 6.2 to 10.0. A major phase of volcanism provides several potential sources of isotopically younger material including the Bronson Hill Anticlinorium (New England Appalachians) [Andersen and Samson, 1995], Barnard volcanics (Vermont), Ascot volcanics (Que´bec) and Carolina Terrane (Virginia, North & South Carolina) (S. Samson personal communication, 1995). [31] The eNd value for Avalonia ( 6.3) can be attributed to a mixture of continental crust and juvenile arc material. The Avalonian basement consists of Late Precambrian igneous rocks and Ordovician volcanics. The composition of the var9 of 17


3

G

10.1029/2001GC000195


0 Baltica Laurentia

-5

Kazakstan

ε Nd(T)

Armorica

-10

Avalonia South China

-15

Peri-Gondwana (northern) Siberia

-20 -25 -30 SIL.

430

ASHGILL

ARENIG LLANVIRN

LLANDEILO CARADOC

450

470

TREMADOC

490

510

Age Figure 5. Nd isotopic evolutionary trends for the Ordovician and Early Silurian Oceans based on conodonts (and one brachiopod) from this study.

ious basement rocks, 4 to 6 [Noble et al., 1993] could easily account for the seawater signal attained from conodonts in our study. The overlapping values for Baltica, Avalonia, and Armorica suggest that these waters may have been mixed. [ 32 ] Data from Welsh sediments [Miller and O’Nions, 1984; Thorogood, 1990] and the Caban Conglomerate [Evans, 1992] for this time interval show for an epsilon range from approximately 10 to 4 and are consistent with the sediment values from earlier time slices for the region. Thorogood [1990] reported Tremadoc aged sediments from Wales having a eNd = 7.7 and 8.2, and Evans [1992] reported Arenig-Llanvirn aged samples from the Caban Conglomerate with values of 6 to 9.5. [33] The northeastern margin of Gondwana, represented by a sample from the Bowan Park Group of east-central Australia, shows a eNd signature of 5.5. This suggests input from juvenile volcanic sources and continental crust. The Bowen Park Group represents deposition in the western flank of a volcanic setting known as the Molong High [Webby and Packham, 1982]. These volcanics could represent the source for the juvenile Nd isotopic source.

[34] Similarly, the sample from the Precordilleran terrane of Argentina shows a signal of 5.9. Source regions could have been from the stable Gondwana craton, while the isotopically juvenile input could be from rift related volcanics or from an arc-continent collision as the Famatina terrane accreted to the margin [Astini et al., 1995]. This Precordilleran terrane is believed to have docked onto the Panthalassa margin of Gondwana by the Caradoc thus the eNd value is not inconsistent. [35] The South China sample, with an eNd value of 11.4, is slightly more radiogenic than the sample from the Arenig-Llanvirn in that area. Comparatively, its signature is most similar to the Laurentian Iapetus margin and is in contrast to other Gondwana values.

3.5. Ashgill and early Silurian [36] The data for this interval (443–436 Ma) are shown in Figure 4. Laurentian Iapetus margin eNd values for the Ashgill are higher than those in past intervals and fall close in range to values from Baltica and Avalonia. Samples from the interior of the craton show a trend toward more 10 of 17


3

G


radiogenic values likely from erosion and transport of island arc complexes associated with the Taconic Orogeny. One sample from the Churchill River Group, northern Hudson Bay, still retains a more non-radiogenic signal of 16.1 and indicates that input is still dominantly from older Precambrian basement. [37] Even though Avalonia had rifted from Gondwana and was drifting northward with Baltica, there is still narrow range between these values and those of the Armorican margin. This indicates that the waters of the Paleotethys Ocean may have been well mixed. Limited data for this time interval includes a sample from Tennessee produced by Keto and Jacobsen [1987] and sediments from Wales [Thorogood, 1990] with values of 8 to 7. [38] Across the Ashgill-Llandovery boundary, Laurentian eNd values of 12 to 6 illustrates a consistent source of material, while the Siberian sample of 1.2 is considerably more radiogenic than any other sample analyzed. This sample represents a displaced terrane (Kolyma Terrane) that was not part of the stable Siberian craton. Ophiolites juxtaposed to the Paleozoic carbonates are thought to be of early Paleozoic age [Van der Voo, 1990] and could be a cause for the more radiogenic signature.

3.6. Nd Isotopic Evolutionary Trends for the Ordovician and Early Silurian Oceans [39] It is apparent from Figure 5 that for most of the Early to Middle Ordovician, Laurentian seawater exhibits eNd values that are distinct ( 29 to 13) from Baltica, Avalonia, and the various peri-Gondwana terranes ( 5 to 10). Eastern Laurentia shifts over time to more radiogenic values and becomes increasingly similar to Baltica, Avalonia, and Amorica. The eNd values derived from Early-Middle Ordovician conodonts and brachiopods from the Baltic platform range between 8.5 and 4.5 [Felitsyn et al., 1998] and are consistent with our envelope that represents the eastern Iapetus signal [40] For Laurentia, there is a general trend toward increasing values until 470 Ma, followed by shifts toward lower values around 460–450 Ma.

10.1029/2001GC000195

This is followed by the final change toward increasing values. It is also apparent (Figure 5) that South China falls within the envelope of Laurentia rather than that of the upper curve, which contains all the peri-Gondwana microplates. [41] The shifts in eNd values are most dramatic for Laurentia, and the more intense sample distribution allows for a more detailed analysis of the marginal oceans of this craton. The general increase toward more radiogenic values can be interpreted as the reduction of erosion of older Precambrian terranes due to the widespread deposition of carbonates during this time. This elimination of older continental crustal sources results in the dominant input being of younger crustal material mixed with isotopically juvenile volcanics produced by arc volcanism and continent-arc collisions. This widespread carbonate deposition is the result of the Caradoc transgressions, the largest in the Phanerozoic, that would have reduced barriers between ocean masses. The Caradocian excursion in the Nd isotopic curve from 470–460 Ma still needs further investigation. Indeed, this may be an artefact of sampling and may serve to illustrate the effects of localised signals. [42] Other secular variation curves have been published by Keto and Jacobsen [1987] for North America (Laurentia) and Europe (Baltic). When the Keto and Jacobsen [1987] curve is recalibrated to the Harland et al. [1990] timescale, it is complementary to this study in the overall trends. The Keto and Jacobsen [1987] Early Ordovician curve is not adequately sampled to confirm the initial trend toward increasingly radiogenic values in the early Arenig, but it does show an abrupt shift toward nonradiogenic input followed by resumption toward more radiogenic values by the end Ordovician. Their curve also illustrates the presence of two distinct water masses in the broad Iapetus region, one associated with North America and the other with Europe. They suggested that the values that represent the waters off North America are typically Panthalassic, while those off Baltica represent true Iapetus waters. Our alternative interpretation consistent with their data is that the values associated with North America, except for western samples, do not represent values for the Panthalassa Ocean but rather represent western Iapetus Ocean 11 of 17


3

G


values and those of Baltica represent the eastern Iapetus Ocean. [43] Holmden et al. [1996] found that the eNd values from Laurentian conodonts at 454 Ma showed systematic changes associated with input from continental sources with different ages. They argued that caution should be used when considering oceanic secular variation curves since epeiric seawater would have been dominated by regional riverine material. They concluded that the Nd isotopic values for epeiric, near shore, and oceanic environments can be grouped into aquafacies. [44] Although there is a clear trend in our data that can serve as oceanic secular variation, there is also considerable regional control in some of the values. This illustrates the forcing that rivers and eperic waterbodies can have on the Nd isotopic signal. Our samples deliberately included regional environmental gradients and the isotopic values from these samples (e.g., Skoki versus Outram/McKay; Catoche versus Cow Head; Harding Sandstone/ Gull River versus Whitby; Churchill River) supports the concept of regional controlling factors.

3.7. Paleogeographic Models [45] The ability to test paleogeographic reconstruction models for any period in geologic time is a function of the size and quality of the database. With the exception of Laurentia and Avalonia, the published Nd isotope data set for the Ordovician is modest but defines the Nd isotopic ranges for the various water masses associated with the major Ordovician oceans and epeiric seas. [46] The equatorial position for the Laurentian craton has been widely accepted for many years [Scotese and McKerrow, 1991]. The Nd isotopic data generated from the eastern and western margins of this craton show how the epeiric seas and adjacent oceans changed through time with respect to tectonics, erosional fluxes, and eustasy. [47] Paleomagnetic data for Baltica are also reasonably well constrained, and the faunal and paleoclimate data are in agreement with initial southerly high latitude during the early Ordovician with a drift toward more equatorial positions in the

10.1029/2001GC000195

Middle and Late Ordovician (for review see Torsvik et al. [1992]). Data are also in agreement with a convergence of the paleopoles for Laurentia and Baltica in the Late Silurian [Torsvik et al., 1990]. The Nd isotopic values for Baltica determine the range of values for the eastern Iapetus Ocean. There are no significant shifts in eNd epsilon units associated with its northward drift, although it does show a slightly more nonradiogenic component during the Ashgill possibly due to increased influenced from materials shed from Laurentia and increased water mass mixing with western Iapetus waters as the Iapetus Ocean closed. [48] The reconstruction by Paris and Robardet [1990] shows the Armorican plate is rotated into a more easterly position and Avalonia is placed as a southwest tongue of Baltica. In review, Fortey and Cocks [1992] criticized this reconstruction for using an inappropriate fossil group (acritarchs) to constrain the biogeographic patterns and that the authors failed to include critical data that did not support their reconstruction. Unfortunately, the Nd isotopic data from this study do not resolve this conflict. There is a general homogenization of isotopic signals from Avalonia, Armorica, and Baltica. The data do not show any evidence of a discrete Rheic Ocean or the South Armorican Ocean as reported by Paris and Robardet [1990] to have existed. The data defines isotopic values for the eastern Iapetus and the southern Paleotethys and do not indicate any other regionally significant water body during the Ordovician Period. This may be due to strong east-west ocean circulation in this region, which would have a homogenizing influence, and/or similarities in sedimentary provenance due to the peneplaned nature of Baltica at that time. [49] The positioning of Baltica proposed by Dalziel et al. [1994] is difficult to reconcile. This model does not violate the paleomagnetic constraints; however, at 450 Ma, Baltica is a considerable distance from Laurentia. Additionally, this model does not account for the positioning of Avalonia or Armorica. Even if thse microplates are assumed to be in the same position as the Scotese and McKerrow [1991] and Scotese [1999] models, the Nd isotopic results from this study cannot resolve this 12 of 17


3

G


debate. Nd analyses could resolve this debate, but it would require more extensive sampling. The lack of carbonates for northern Africa and northern South America will make sampling for conodonts difficult. [50] Paleomagnetic results [Trench and Torsvik, 1992, and references therein] are in general agreement that Avalonia was situated at high southerly latitude during the early Ordovician. It contains a Pan-African basement and was likely adjacent to the northern margin of Africa, and the Armorican plate Avalonia apparently rifted from the margin of North Africa during the Arenig and subsequently travelled north [Scotese and McKerrow, 1991]. One scenario suggests that it came into contact with Baltica, possibly during the late Ashgill, and both travelled northwest until their eventual collision with the northern and eastern margins of Laurentia during the Acadian Orogeny in the Late Silurian [Trench and Torsvik, 1992]. Armorica has a similar basement to Avalonia [Van der Voo, 1990] and northeastern Africa [Noblet and Lefort, 1990]. Paleomagnetic data suggest similar high southerly latitude during much of the Ordovician [Torsvik et al., 1990]. [51] None of the samples represent the northern margin of Africa so it is not possible in this study to compare the positioning of Avalonia against this region of Gondwana. However, the data for Armorica, Baltica, and Avalonia overlap in this interval, even when Avalonia had supposedly begun its northward drift. Alternatively, Armorica may have rifted and been in closer proximity to Avalonia and Baltica than previously considered. However, there are no paleomagnetic data to support the rifting of Armorica until the Late Devonian [Van der Voo, 1988]. [52] The positioning and orientation of south China has been a considerable problem, and paleomagnetic data have not solved this issue (for review see Li et al. [1993]). Faunal data suggest affinities between the south China and Australian blocks of Gondwana [Young, 1990] supporting the positioning of south China close to Australia in many reconstructions. There are only two samples from south China, one in the Arenig and the other in the

10.1029/2001GC000195

Llandeilo-Caradoc. Both of these sites have Nd isotopic values that are similar to those for Laurentia. Since there are no data from Indochina, north China, or Iran, it is difficult to substantiate the position of south China relative to these other peri-Gondwana blocks. The isotopic values are different from the values of the Armorican and Avalonian microplates in the southern Paleotethys. It is quite possible that south China was separated from Gondwana and in closer proximity to Laurentia than suggested by the Scotese and McKerrow and Scotese models. [53] Although there are some apparent discrepancies in the literature (for review see Li et al. [1993]), the southern paleopoles of the major blocks of Gondwana generally group around northwest Africa. With the general paucity of carbonates, relatively few samples were analyzed, and the majority of these represent peri-Gondwana terranes. For the Tremadoc, the sample from the Ninmaroo Formation, Georgina Basin, and northern Territory represents the Nd isotopic composition of seawater residing in an inner cratonic basin and interacting with the northern Paleotethys. This value is also similar to the value from Kazakstan ( 10.3) and suggesting that these two sites were close enough in proximity to influence the same body of water. [54] During the Caradoc, the Clearview Limestone Member, New South Wales, is situated on the stable craton in most reconstructions. However, VandenBerg and Stewart [1992] suggested that the Molong High, of which the Clearview Limestone represents a section of the western flank, was an offshore volcanic island-arc complex. The signature of the sample suggests significant isotopically juvenile input, but it also detects some mixing with stable continental crust. It is in general agreement with the regional geological reconstruction with the western flank of the Molong High being bordered by volcanics on the east and the stable platform craton on the southwest. The Nd isotopic data cannot delimit how far this island was offshore, but it is apparent that continental input was still a contributor to the arc basin. This value is also important because of its similarity to that of the Argentine Precordillera. 13 of 17


3

G


[55] The origin and migration of the Argentine Precordillera are only now becoming understood [Astini et al., 1995]. Unfortunately, the single Nd isotopic analysis from this study cannot assist in directly constraining any one model and or assist in tracking the evolution of the terrane. The sample from the Precordillera has an age of 446 Ma. This postdates its proximate accretion with Gondwana. A way to test the reconstruction would be to sample the Precordillera terrane through time to see if it initially shows a Laurentian signature with gradual shifts to more radiogenic values associated with the crossing of the Iapetus and eventual collision with Gondwana. The sampling for such a study has been completed by one of the authors (C.R.B.). [56] Torsvik et al. [1995] pointed out that in Dalziel et al.’s [1994] model, the distance of separation between eastern Laurentia and the Argentine Precordillera would be small; hence the Iapetus Ocean would have to be narrower than predicted by other models. However, the higher radiogenic signature from this region of Gondwana would not be expected if it were in such close proximity to Laurentia with its more nonradiogenic values. The lack of carbonates in South America is in contrast to Laurentia, and the differences in faunal provinces also fails to support the model of Dalziel et al. [1994]. The eNd values from Australia ( 5.5) and the Argentine Precordillera ( 5.9) are more supportive of the Scotese and McKerrow [1991] and Scotese [1999] models. [57] There are no systematic paleomagnetic measurements that delineate the position of the Kazakhstan terranes in the Ordovician [Li et al., 1993], and the Nd isotopic values from this study provide limited information except for providing the isotopic signal for the northern Paleotethys in the Tremadoc. It is in good agreement with those from northern Gondwana (Ninmaroo Formation, Australia) supporting the positioning in the Scotese and McKerrow [1991] model. Since Kazakhstan was apparently a series of arcs and exotic terranes future work will need intensive sampling to understand the evolution of the components of this microplate.

10.1029/2001GC000195

[58] Torsvik et al. [1995] published paleopoles for the Siberian craton, but paleopoles for the Kolyma Terrane are not known. Brachiopod evidence from the terrane suggests that it was in close proximity to the northern edge of Laurentia [Rong and Harper, 1988]. However, the eNd data suggest that Kolyma was not influenced by sediments shedding from Laurentia. For most of the Ordovician, the major land masses were situated in the Southern Hemisphere, and the northern Panthalassa Ocean had essentially unobstructed circulation above 30 north. This oceanic area would have been predominately influenced by oceanic crustal material rather than continental crust. Hence, the signature of the Kolyma terrane represents the signal of the northern Panthalassa Ocean. It is likely that this terrane, and perhaps the Siberian craton, was further north than the Scotese and McKerrow [1991] and Scotese [1999] models predict.

3.8. Testing Displaced Terranes [59] The data from conodonts in this study can help constrain the paleogeographic position of ‘‘displaced’’ terranes. Nd isotopic data exist from the sediments of the Southern Uplands of Scotland, the Dalradian sediments of Scotland, and the eastern Central Mobile Belt of western Newfoundland [60] The Dalradian (Late Precambrian-early Paleozoic) sediments of Scotland are a thick sequence of unfossiliferous metasediments, sandwiched between the Highland Border Complex and the Durness Limestone, both of which have Ordovician strata that contain faunas with Laurentian or Appalachian affinities [Curry and Williams, 1984]. O’Nions et al. [1983] report eNd values of 19 to 18 for Tremadoc Dalradian slates, indicating that these sediments were influenced predominately by old Archean continental crust. These values fall within the eNd curve of Laurentia for that time and supports the idea that the Dalradian was deposited on the margin of Laurentia along the Iapetus Ocean. [61] The Southern Uplands (SU) of Scotland is thought to represent an accretionary prism deposited on the northwest margin of the Iapetus Ocean. Stone and Evans [1995] presented a thorough Nd 14 of 17


3

G


isotopic analysis of greywacke samples from the SU, including previously published values of O’Nions et al. [1983]. The data suggested that the SU in the Ordovician had two provenances, a volcanic arc and a basement provenance. The arc shows a characteristic radiogenic signal of 3 to 2 whereas the basement has a range of 12 to 8. By the Ashgill and Llandovery, these provenances were indistinguishable with a range of values from 8 to 5, supporting the SU being placed in the northwest region of the Iapetus Ocean off the northeast (present day) margin of Laurentia. [62] The eastern Central Mobile Belt (ECMB) of Newfoundland includes the Gander Zone, parts of the Exploits Subzone and the Avalon Zone. Nd isotopic values for metasediments with a minimum age of 470 Ma have a range of the eNd values from 8 to 6.5 and the granitoid suites show a range of 8 to 2 [Kerr et al., 1995]. The epsilon values indicate that the affinities for the ECMB in the Arenig are with the eastern Iapetus waters (Avalonia and Baltica). This is supported by faunal evidence from the Arenig of the Exploits Subzone where graptolites and trilobites show Avalonia/periGondwana affinities [Williams et al., 1992].

4. Conclusions [63] Ordovician conodont samples were chosen for Nd isotopic analysis to attain proxy signals for ancient oceans contemporaneous with the timing of deposition through four time slices of the Ordovician and early Silurian. Once delimited, the Nd isotopic signature of these ancient oceans can be linked to regional and global geological processes and be used to constrain models of proposed paleogeography. [64] As Laurentia was separated from other landmasses, its east and west margins show different isotopic values associated with the western Iapetus and the Panthalassa Ocean (south of 30N), respectively. These values are distinct from Baltica, Gondwana, and peri-Gondwana terranes that delineate the Nd isotopic composition of the eastern Iapetus Ocean, southern and northern Paleotethys Ocean. Changing sea level and paleogeographic configurations and the influence of the Taconic

10.1029/2001GC000195

Orogeny are reflected in the changing isotopic values as the signatures for Baltica/Avalonia/Amorica and eastern Laurentia become increasingly similar as the former drifts northward. By the Late Ordovician, there is complete overlap in the values from both sides of the Iapetus and the southern hemispheric oceans become isotopically homogenous, signaling enhanced circulation as a result of converging land masses and reduction of paleoceanographic barriers due to the extensive Caradoc transgression. [65] Changes in the isotopic signal can be correlated with eustasy, paleogeography, and orogenic events. Unfortunately, the Nd isotopic results cannot resolve the reconstructions of Paris and Robardet [1990] nor the one of Dalziel et al. [1994]. There is generalized broad scale support for the reconstruction as modelled by Scotese and McKerrow [1991] and Scotese [1999] with respect to the major craton configurations. On a finer scale, there are some microplate configurations that can be called into question. [66] This study has shown that conodonts can be used to determine the Nd isotopic signature of ancient oceans providing an independent technique for testing paleogeographic reconstructions. This is critical since it provides a means of dealing with plate configurations without dealing with the potential errors involved in paleomagnetism. Further sampling and analysis could tighten the Nd isotopic ranges and increase the resolution of the different oceanic masses and epeiric seas. This may well reveal finer scale patterns not detectable in broad reconstructions. Additionally, as the global ocean values are delineated, it will become easier to model the relative position of displaced terranes by placing them in the appropriate ocean mass or tracking their oceanic migrations. Finally, the Nd isotopic data can be used to integrate the various paleoenvironmental models and yield a more comprehensive picture of the early Paleozoic world.

Acknowledgments [67] We thank the reviewers (S.J. Goldstein and two anonymous reviewers) as well as the Guest Editor H. Staudigel for their helpful comments on the manuscript. C. R. Barnes grate15 of 17


3

G


fully acknowledges financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC), and Stein Jacobsen is grateful for the financial support from the U.S. National Science Foundation grant EAR-9616072.

References Andersen, C. B., and S. D. Samson, Temporal changes in Nd isotopic composition of sedimentary rocks in the Sevier and Taconic foreland basins: Increasing influence of juvenile sources, Geology, 23, 983 – 986, 1995. Astini, R. A., J. L. Benedetto, and N. E. Vaccari, The early Paleozoic evolution of the Argentine Precordillera as a Laurentian rifted, drifted, and collided terrane: A geodynamic model, Geol. Soc. Am. Bull., 107, 253 – 273, 1995. Barnes, C. R., Z. Youqiu, L. Jeppsson, D. Fredholm, W. J. Varker, A. Swift, G. K. Merrill, L. Jeppsson, and K. J. Dorning, Recent developments in rock disintegration techniques for the extraction of conodonts, in Conodonts: Investigative Techniques and Applications, edited by R. L. Austin, pp. 35 – 53, Ellis Horwood, Chichester, New York,1987. Boghossian, N. D., P. J. Patchett, G. M. Ross, and G. E. Gehrels, Nd isotopes and the source of sediments in the miogeocline of the Canadian Cordillera, J. Geol., 104, 259 – 277, 1996. Cecile, M. P., D. W. Morrow, and G. K. Williams, Early Paleozoic (Cambrian to Early Devonian) tectonic framework, Canadian Cordillera, Bull. Can. Pet. Geol., 45, 54 – 74, 1997. Curry, G. B., and A. Williams, Lower Ordovician brachiopods from the Ben Suardal Limestone Formation (Durness Group) of Skye, western Scotland, Trans. R. Soc. Edinburgh Earth Sci., 75, 301 – 310, 1984. Dalziel, I. W. D., L. H. Dalla Salda, and L. M. Gahagan, Paleozoic Laurentia-Gondwana interaction and the origin of the Appalachian-Andean mountain system, Geol. Soc. Am. Bull., 106, 243 – 252, 1994. Evans, J. A., Geochemical and isotope composition of pebbles from the Caban Conglomerate Formation and their bearing on the source of Welsh Paleozoic sedimentary rocks, Geol. Mag., 129, 581 – 587, 1992. Felitsyn, S. A., U. Sturesson, L. Popov, and L. Holmer, Nd isotopic composition and rare earth element distribution in early Paleozoic biogenic apatite from Baltoscandia: A signature of Iapetus ocean water, Geology, 26, 1086 – 1092, 1998. Fortey, R. A., and L. R. M. Cocks, The early Paleozoic of the North Atlantic region as a test case for the use of fossils in continental reconstructions, Tectonophys., 206, 147 – 158, 1992. Goldstein, S. J., and S. B. Jacobsen, The Nd and Sr isotope systematics of river-water dissolved material: implications for the sources of Nd and Sr in seawater, Chem. Geol. Isotope Geosci. Sect., 66, 245 – 272, 1987. Grandjean, P., H. Cappetta, A. Michard, and F. Albare´de, The assessment of REE patterns and 143Nd/144Nd ratios in fish remains, Earth Planet. Sci. Lett., 84, 181 – 196, 1987.

10.1029/2001GC000195

Harland, W. B., R. L. Armstrong, A. V. Cox, L. E. Craig, A. G. Smith, and D. G. Smith, A Geological Time Scale, 263 pp., Cambridge Press, New York, 1990. Holmden, C., R. A. Creaser, K. Muehlenbachs, S. M. Bergstrom, and S. A. Leslie, Isotopic and elemental systematics of Sr and Nd in 454 Ma biogenic apatites: Implications for paleoseawater studies, Earth. Planet. Sci. Lett., 142, 425 – 437, 1996. Jacobsen, S. B., and R. F. Dymek, Nd and Sr isotope systematics in clastic metasediments from Isua, West Greenland: identification of pre-3.8 Ga differentiated crustal components, J. Geophys. Res., 93, 338 – 354, 1988. Jacobsen, S. B., and G. J. Wasserburg, Sm-Nd isotopic evolution of chondrites, Earth Planet. Sci. Lett., 50, 139 – 155, 1980. James, N. P., R. K. Stevens, C. R. Barnes, and I. Knight, Evolution of a Lower Paleozoic continental-margin carbonate platform, northern Canadian Appalachians, in Controls on Carbonate Platform and Basin Development, SEPM Spec. Publ. No. 44, edited by P. D. Crevello, J. L. Wilson, J. F. Sarg and J. F. Read, pp. 123 – 146, Tulsa, Oklahoma,1989. Kerr, A., G. A. Jenner, and B. J. Fryer, Sm-Nd isotopic geochemistry of Precambrian to Paleozoic granitoid suites and the deep-crustal structure of the southeast margin of the Newfoundland Appalachians, Can. J. Earth Sci., 32, 224 – 245, 1995. Keto, L. S., and S. B. Jacobsen, Nd and Sr isotopic variations of Early Paleozoic oceans, Earth Planet. Sci. Lett., 84, 7 – 41, 1987. Keto, L. S., and S. B. Jacobsen, Nd isotopic variations of Phanerozoic paleoceans, Earth Planet. Sci. Lett., 90, 395 – 410, 1988. Li, Z. X., C.McA. Powell, and A. Trench, Paleozoic global reconstructions, in Paleozoic Vertebrate Biostratigraphy and Biogeography, edited by J. A. Long, pp. 25 – 58, Belhaven Press, London, 1993. Martin, E. E., and B. A. Haley, Fossil fish teeth as proxies for seawater Sr and Nd isotopes, Geochim. Cosmochim. Acta, 64, 835 – 847, 2000. Miller, R. G., and R. K. O’Nions, The provenance and crustal residence ages of British sediments in relation to paleogeographic reconstructions, Earth Planet. Sci. Lett., 68, 459 – 470, 1984. Noble, S. R., R. D. Tucker, and T. C. Pharaoh, Lower Paleozoic and Precambrian igneous rocks from eastern England, and their bearing on late Ordovician closure of the Tornquist Sea: Constraints from U-Pb and Nd isotopes, Geol. Mag., 130, 835 – 846, 1993. Noblet, C., and J. P. Lefort, Sedimentological evidence for a limited separation between Armorica and Gondwana during the Early Ordovician, Geology, 18, 303 – 306, 1990. O’Nions, R. K., P. J. Hamilton, and P. J. Hooker, A Nd isotope investigation of sediments related to crustal development in the British Isles, Earth Planet. Sci. Lett., 63, 229 – 240, 1983. Paris, F., and M. Robardet, Early Paleozoic paleobiogeography of the Variscan regions, Tectonophys., 177, 193 – 213, 1990. 16 of 17


3

G


Piepgras, D. J., and G. J. Wasserburg, Neodymium isotopic variations in seawater, Earth Planet. Sci. Lett., 50, 128 – 138, 1980. Rong, J., and D. A. T. Harper, A global synthesis of the latest Ordovician Hirnantian brachiopod faunas, Trans. R. Soc. Edinburgh Earth Sci., 79, 383 – 402, 1988. Scotese, C. R., Paleogeographic Atlas. PALEOMAP Prog. Rep. 90-0497, 45 pp., Dep. of Geol., Univ. of Tex. at Arlington, Arlington, 1997. Scotese, C. R., Paleogeographic Map Archive. Still images used: 510.eps, 470.eps, 460.eps, 440.eps, PALEOMAP Project, Dep. of Geol., Univ. of Tex. at Arlington, Arlington, 1999. Scotese, C. R., and W. S. McKerrow, Ordovician plate tectonic reconstructions, in Advances in Ordovician Geology, edited by C. R. Barnes and S. H. Williams, Geol. Surv. Can. Pap. 90-9, 271 – 282, 1991. Shaw, H. G., and G. J. Wasserburg, Sm-Nd in marine carbonates and phosphates: Implications for Nd isotopes in seawater and crustal ages, Geochim. Cosmochim. Acta, 4, 503 – 518, 1985. Staudigel, H., P. Doyle, and A. Zindler, Sr and Nd isotope systematics in fish teeth, Earth Planet. Sci. Lett., 76, 45 – 56, 1985. Stone, P., and J. A. Evans, Nd-isotopic study of provenance patterns across the British sector of the Iapetus Suture, Geol. Mag., 132, 571 – 580, 1995. Thorogood, E. J., Provenance of the pre-Devonian sediments of England and Wales: Sm-Nd isotopic evidence, J. Geol. Soc., London, 147, 591 – 594, 1990. Torsvik, T. H., M. A. Smethurst, J. C. Briden, and B. A. Sturt, A review of Paleozoic paleomagnetic data from Europe and their paleogeographical implications, in Paleozoic Paleogeography and Biogeography, edited by W. S. McKerrow and C. R. Scotese, Geol. Soc. London Mem., 12, 25 – 41, 1990. Torsvik, T., M. A. Smethurst, R. Van der Voo, A. Trench, N. Abrahamsen, and E. Halvorsen, Baltica. A synopsis of Vendian - Permian paleomagnetic data and their paleotectonic implications, Earth Sci. Rev., 33, 133 – 152, 1992. Torsvik, T. H., J. Tait, V. M. Moralev, W. S. McKerrow, B. A. Sturt, and D. Roberts, Ordovician paleogeography of Siberia and adjacent continents, J. Geol. Soc. London, 152, 279 – 287, 1995. Trench, A., and T. H. Torsvik, The closure of the Iapetus

10.1029/2001GC000195

Ocean and Tornquist Sea: new paleomagnetic constraints, J. Geol. Soc. London, 149, 867 – 870, 1992. VandenBerg, A. H. M., and I. R. Stewart, Ordovician terranes of the southeastern Lachlan Fold Belt: Stratigraphy, structure and paleogeographic reconstruction, Tectonophysics, 241, 159 – 176, 1992. Van der Voo, R., Paleozoic paleogeography of North America, Gondwana, and intervening displaced terranes: Comparisons of paleomagnetism with paleoclimatology and biogeographical patters, Geol. Soc. Am. Bull., 100, 311 – 324, 1988. Van der Voo, R., Phanerozoic paleomagnetic poles from Europe and North America and comparisons with continental reconstructions, Rev. Geophys., 28, 167 – 206, 1990. Wasserburg, G. J., S. B. Jacobsen, D. J. DePaolo, M. T. McCulloch, and T. Wen, Precise determination of Sm/Nd ratios, Sm and Nd isotopic abundances in standard solutions, Geochim. Cosmochim. Acta, 45, 311 – 323, 1981. Webby, B. D., and G. H. Packham, Stratigraphy and regional setting of the Cliefden Caves Limestone Group (Late Ordovician), central-western New South Wales, J. Geol. Soc. Aust., 29, 297 – 317, 1982. Williams, S. H., W. D. Boyce, and S. P. Colman-Sadd, A new Lower Ordovician (Arenig) faunule from the Coy Pond Complex, central Newfoundland, and a refined understanding of the closure of the Iapetus Ocean, Can. J. Earth Sci., 29, 2046 – 2057, 1992. Wright, C. A., Neodymium isotopic composition of Ordovician-Silurian conodonts, M.Sc. Thesis, University of Victoria, Victoria, British Columbia, 1995. Wright, J., R. E. Seymourand H. F. Shaw, REE and Nd isotopes in conodont apatite: Variations with geological ages and depositional environments, in Conodont Biofacies and Provincialism, edited by D. L. Clark, Spec. Pap. Geol. Soc. Am., 196, 325 – 340, 1984. Young, G. C., Devonian vertebrate distribution patterns and cladistic analysis of a paleogeographic hypothesis, in Paleozoic Paleogeography and Biogeography, edited by W. S. McKerrow and C. R. Scotese, Geol. Soc. London Mem., 12, 243 – 255, 1990. Zhao, J. X., M. T. McCulloch, and V. C. Bennett, Sm-Nd and U-Pb zircon isotopic constraints on the provenance of sediments from the Amadeus Basin, central Australia: Evidence from REE fractionation, Geochim. Cosmochim. Acta, 56, 921 – 940, 1992.

17 of 17