Originally submitted 8 December 1988 , accepted for publication 7 July 1989.
Advances in Ordovician Geology, C.R. Barnes and S. H. Williams (eds. ) Geological Survey of Canada. Paper 90-9. p. 283-298. 1991
The geographic position of major continental blocks in the Ordovician, determined primarily from lithofacies analysis and paleomagnetic studies, was the cause of a vastly different oceanic circulation pattern than that of today. The Ordovician was characterized by 1) a northern hemispheric ocean covering about one half of the Earth's surface; 2) a band of land and shallow shelf seas that occupied the tropics; and 3) a major landmass and associated shallow shelf seas in the southern hemisphere.
The trends during the Ordovician were a) the reduction of land and shallow shelves in the equatorial region and b) the movement of the southern landmass toward the south pole, culminating in a south polar ice cap during the Late Ordovician. The lack of meridional barriers in the northern hemisphere produced only zonal (east-west) surface currents: an easterly flowing current in the north polar zone; a westerly flowing current in the north temperate zone; and an easterly flowing current in the northern tropical regions.
The separation of the northern, or Boreal Ocean from the southern oceans (Iapetus and Proto-Tethys) by an equatorial barrier implies that the deep circulation of each ocean was independent. Only in the southern oceans with meridional barriers would the surface flow be similar in pattern to that of modern oceans. Trends toward an austral polar landmass, break-up of equatorial lands, and shallow shelf seas and generation of a polar ice cap during the Late Ordovician permitted establishment of connections between the oceans in the two hemispheres as well as significant deep ventilation from the edges of the polar ice cap during the Late Ordovician glaciation.
Monsoonal reversals of the equatorial currents would have occurred in the southern hemisphere due to the presence of the Gondwanan landmass. Such reversals would shift water mass boundaries on a seasonal basis and might explain apparent mixtures of marine faunal provinces ("Pacific" = tropical; "Atlantic" = temperate-subpolar), especially on the Gondwanan shore or the northeastern side of the Proto-Tethys ocean.
INTRODUCTIONThis paper uses fundamental oceanographic circulation principles to speculate on the major current patterns and water masses and their implications for the biogeography of the Ordovician Period as a function of paleogeogMphic reconstruction of the major continental blocks. This approach elaborates on the procedures used by Stehli (1965), Ross (1975), Berggren and Hollister, 1977, and Parrish (1982) and is an extension of our work on the development of paleooceanographic interpretative methods (Wilde and Berry, 1986).
The primary astrophysical assumptions are 1) the Earth was the same size and shape as it is today, 2) the Earth was rotating in the same direction (west to east), 3) the Earth was rotating at about the same angular velocity (one rotation = 24 hours), 4) the path of rotation was in the plane of the ecliptic (the sun on the equator), and 5) there was some seasonality (the pole of rotation at some angle to the plane of the ecliptic). In the atmosphere of a rotating, spherical Earth, incoming solar heating would be transferred polarward in three major atmospheric cells symmetrical about the incoming solar maximum near the geographic equator. This would produce climatic zones with rising air at the meteorological equator and at 60O, and sinking air at the poles and at 30 . The major climate zones would have been bounded by these vertical air flows and would have been similar to those of today with a tropical zone extending from about 30 N to 30 S centered on the geographic equator; two temperate zones from 30 to 60 ; and two polar zones from 60 to the pole. The degree designations for the boundaries are long-term mean values and would vary seasonally. The meteorological equator or thermal equator is the line that divides the polarward heat transfer in half. It will differ from the geographic equator (90 from each pole of rotation) 1) seasonally, moving with the sun north of the geographic equator in the boreal sumn1er and south of the geographic equator in the austral summer and 2) as a function of the land-sea ratio in each hemisphere shifting into the hemisphere with the most land.
The modern open ocean is density stratified and has two layers: 1) an upper wind-mixed isothermal layer (0 to 100 m), the temperature and salinity characteristics of which are determined by local surface conditions with a general latitudinal temperature decrease from the equator to the pole and 2) a deep isothermal layer (1000 m to the bottom) of relatively uniformly cold water with gradually increasing temperature a function of distance from the area of formation at high latitudes with little or no relationship to surface latitudinal temperature gradients. These two layers are separated by a boundary layer: the main pycnocline (100 to 1000 m), where the temperature and density are transitional with depth between the upper and deep layers. For this discussion of the Ordovician ocean, only circulation in the upper mixed layer is treated in detail. Although this would cover sites of deposition of most shelf deposits, deep sea fan or continental margin deposits would intersect the transitional pycnocline, as would outer shelf sediments during high stands of sea level. In any case, most of the pelagic and benthic faunas that were likely to be fossilized would live in the upper mixed layer, although they may have become fossilized in sediments in thepycnocline.
In the oceans, the movement of water is primarily the result of the interaction of frictional influence of the atmospheric winds on the ocean, and thermo-haline (density) stability. In most of the ocean, in the pycnocline and the deep layer, the effects of friction by surface winds or the bottom are negligible, so the circulation is dominated by density differences. Only in the surface mixed layer is wind friction the predominant driving force. However, the wind-driven circulation can be modified or even dominated by thermo-haline circulation in areas where density differences are sufficient to cause sinking or rising water before wind mixing and "homogenation'' can occur. At the sea surface, air flow connecting the rising or sinking air at the cell boundaries would produce easterly surface winds in the polar and tropical zones and westerly surface winds h1 the temperate zones. The frictional couple between the air l1ow and the sea surface generates water flow in the surface ocean to the depth of the influence of friction essentially defining the surface mixed layer. Ekman (1905) mathematically described this flow which is called the Ekman transport, and noted the net flow was 90 cum sol (to the right in the northern hemisphere and to the left in the southern hemisphere) to the wind. Low oceanic pressure is formed when two planetary Ekman transport vectors diverge (0 and 60 ) and high oceanic pressure is formed when planetary Ekman transport vectors converge (30 ). For oceanic pressure systems, like atmospheric pressure systems, the surface circulation is assumed to be geostrophic with cyclonic flow about low pressure systems and anticyclonic flow about high pressure systems. If there were no barriers in the oceans (i.e., no continents), the water circulation about such systems would be purely zonal or east-west. With meridional barriers, the north-south flow is permitted and the circulation can transfer heat to different latitudes. The Sverdrup-Ekman circulation model (Stommel, 1957) describes how Ekman transport due to planetary winds with appropriate meridional (north-south) boundaries creates quasi-circular pressure systems in the ocean.
The major currents generated by this system can also be identified with climatic zones, although the flows are connected at zonal boundaries.
In the tropics, there are two westward flowing equatorial currents astride the meteorological equator, which are the northerly and southerly limbs of the equatorial divergence (oceanic low pressure). The equatorial currents flow in the same direction due to the shift of the Coriolis deflection and the sense of geostrophic flow across the geographic equator. The equatorial currents transfer water to the west creating an accumulation of water at the western barrier. As the centre of the atmospheric low pressure at the meteorological equator is characterized by rising air, surface wind flow is diminished, as signified by the calms of the doldrums. Thus, part of the water accumulation is relieved by oceanic easterly flow embedded in the centre of the equatorial low pressure. One such current is the easterly flowing equatorial counter-current, seen best developed on the eastern side of the ocean. As this type of counter-current develops as a response to the accumulation of water on the west, the amount of transport and the extent of the counter-current would-vary directly with the width of the particular equatorial ocean and the fetch of the tropical trade winds. The transport of surface counter-current is small and restricted to the narrow zone of the doldrums, so that the westward zonal flow in the tropics is dominant.
In the temperate zone, the open ocean circulation is dominated by the oceanic high pressure centred in middle latitudes at about 30 . Such systems, with meridional barriers, will produce polarward warm currents derived from the. warm water piled up on the western side of the system in the tropics and equatorward cool currents on the eastern side. In general, the polarward flow is strongest as a result of western intensification (Stommel, 1948; Munk and Carrier, 1950).
In the polar regions, the circulation is cyclonic about oceanic low pressure centred at about 60 . With meridional barriers, the circulation will be cyclonic with equatorward cold currents flowing along the western barrier and polarward warm to cool currents along the eastern barrier. The temperature of polarward flow is a function of the strength of the western intensification polarward warm current about the neighboring oceanic high pressure at midlatitudes. As the area of the polar regions is much smaller than either the temperate zone or the tropical zone, the volume of surface water formed in the polar regions is relatively small. The coherence and transport of the north-south components of circulation depends on the configulation and the meridional continuity of the barriers. Thus, a definite continental barrier such as the Americas will produce a strong western intensification as seen in the Gulf Stream. Seasonal conditions
The planetary oceanic circulation can be modified on a seasonal basis, il the welds produced by atmospheric pressure systems oppose and override the planetary winds or are able to generate water currents counter to those produced by first-order geostrophic processes.
Monsoons are produced when there is sufficient land at middle latitudes to modify the circulation derived from global oceanic conditions. The landmass or continental effect (don Arx, 1962, p. 164; Barron, 1981) is due to the difference in heat capacity and albedo between land and water, because water has a greater heat capacity and lower albedo than land. Thus, the air temperature over land is highly variable as a function of insolation compared to water, with seasonal atmospheric high pressure created over land in the winter (sinking cold air) and atmospheric low pressure created over land in the summer (rising hot air). These continental atmospherie pressure systems reverse with season, air flows from the ocean to the land in the hemispheric summer and from the land to the ocean in the hemispheric winter. Where the con- figuration of the landmass results in wields flowing counter to the direction of the usual planetary winds, current reversals will occur in the ocean. In the opposite season, the monsoonal winds and the planetary wields will be in the same direction, with an increase in current speed and transport in the ocean. In the tropics, reversals of the trade winds in the hemispheric summer would occur if the continental pressure system were centered to the east of the oceanic area. Such a situation is present today in the tropical northern Indian Ocean.
Midlatitude west coast gradient currents
During the winter, off western coasts in the temperate zone, the presence of atmospheric low pressure systems offshore creates localized polarward surface winds along the coasts, and Ekman transport against them. This pile-up of water creates polarward current flow along the coast, counter to the equatorward flow on the eastem side-of the oceanic high pressure system. The Davidson Current (Sverdrup et al., 1942), off the west coast of the United States, is an example of such winter seasonal currents.
WATER MASS FORMATIONWater masses are packets of water with relatively uniform or systematically varying temperature and salinity. Their distribution and coherence, therefore, would be density related. However, the distribution of surface water masses, due to the primary influence of the wind-driven circulation in the upper mixed layer, is the result of a combination of both thermohaline (density) and wind-driven factors. Surface water masses are formed and identified by the temperature-salinity conditions generally as the result of latitudinal changes in temperature (polarward cooling) and salinity (evaporation versus precipitation). Surface circulation will transport water from areas of primary formation, with gradual changes due to mixing or dilution with smaller water masses, or latitudinal temperature change, until the individual water mass encounters another water mass with differing properties. If the density difference is sufficient, the denser water mass will sink under the lighter mass at a density convergence, forming a boundary zone between the two water masses. This density convergence is different from the wind-driven convergence produced by the opposition of two Ekman transport vectors. Thus the boundaries of watermasses are only subparallel with the climatic boundaries where the flow is zonal, and diverge significantly where the flow is meridional at continental ban iers. Water mass nomenclature used here follows that in Sverdrup et al. (1942) and Tchernia (1980).
Equatorial water masses are of relatively low density, because they are wammer and generally less saline than other oceanic water, due to high precipitation in the tropics. More dense but basically tropical masses, or central water, show a wider range of warm temperatures, as a result of meridional flow and mixing and generally higher salinity at its polarward edge, as the evaporation maximum occurs at the latitude of the center of the oceanic high pressure system. Surface waters in temperate to subpolar latitudes are cool with relatively low salinity due to higher precipitation at the boundary between the temperate and polar zones. The paired equatorial convergences define the region of the doldrums and highest rainfall. This least dense water floats on top of more dense tropical waters and is carried in the equatorial countercurrent. As the salinity difference caused by precipitation is slight, due to the enormous volume of the oceans, the paired equatorial convergences are weak. The equatorial convergences lie embedded within the major wind-driven planetary divergence associated with the oceanic low pressure near the equator (Sverdrup, 1947). Accordingly, at the global scale of the maps figured here, the equatorial water masses are merged with the central water masses. At the subtropical convergence, the warmer but more saline central water sinks with the colder but less saline subpolar waters. This boundary generally is polarward of the center-line between the oceanic middle latitude high with wanm central water flowing westward on the equatorward part of the anticyclonic oceanic high, and cooler subpolar water flowing east on the other side of the convergence. At the subtropical convergence, central water sinks, forming the upper part of the pycnocline and subpolar water sinks, forming intemediate water at the base of the pycnocline. During present glacial climates, there is a major convergence in the polar regions, where dense, cold polar water sinks. Foaming deep and bottom water below the pycnocline. The high latitude wind-driven planetary divergence occurs between the subtropical convergence and the polar convergence at the center of the oceanic low-pressure system.
Water masses, because of both the uniformity and the continuity of variation in temperature and salinity within each water mass, commonly define major biogeographic regions (e.g., La Fond and La Fond, 1971; McGowan, 1972). In this paper, the figures depict the equatorial and central water masses as warm water, subpolar water masses as cool water, and polar water masses as cold water.
PALEOGEOGRAPHIC RECONSTRUCTIONSThe representations of significant continental blocks given in the figures presented here are based on the reconstructions of Scotese (1986). As most of the fossil record is in marine rocks, the outline of the continental blocks presented by Scotese (1986) must include the continental shelf and even parts of the slope and rise, with sea level somewhere inside the given outline. For the Sverdrup-Ekman model, the meridional barriers are not coh1cident with sea level, but with the edge of the continental block on the slope and rise. Thus, lack of detailed knowledge of the location of sea level on a given block at any given time is not a critical problem in reconstruction of the major global currents. Such imprecision is a problem for coastal currents, particularly for seasonal monsoonal and gradient currents. We have chosen to illustrate currents flowing over continental blocks seaward of the position of the major shield and platform areas, assuming these areas would have been shallow enough to be considered a barrier in the Sverdrup-Ekman model. However, the depicted currents would not flow necessarily over shallow platforms and regions of continental shelf shallower than 100 m. Areas within a continental block boundary known to be deep water continental margin, such as the Upper Ordovician section at Dobs' Linn, Scotland, are indicated as oceanic.
Kirshvink (1988, pers. comm.) has developed paleogeographic reconstructions that show differences in block orientation, particularly for Kazakhstan and North China, compared with Scotese (1986). Such differences would not affect the major currents, so are not included here.
RELEVANCE TO THE ORDOVICIANImparting climatic zones and seasonality to the Ordovician is a risky business, particularly using analogues of the modern interglacial climate with a relatively high temperature contrast between the polar and the tropical latitudes. Although the Upper Ordovician is considered to have had a glacial climate, most of the Ordovician is considered to have had a milder climate than today (Frakes, 1979; Brenchley, 1984; Barnes, 1986). Also, the position of the continents and other landmasses, whicl1 might aLlect physical oceanography, is different from today (Scotese, 1986; Cocks and Fortey, 1982). Thus, the position of the climatic belts is conjectural. The approach used here is to assume that the oceans, at least since the late Precambrian, covered a greater surface area than the land. Accordingly, the primary driving force for surface circulation is that of wind blowing over water. A first-order circulation model would be that of a limitless ocean, similar to the Hough-Goldsborough models described by Stommel (1957). Thus, the influence of the varying positions of (he continental blocks would be to impose modifications to the primary pattern, chiefly in pemmitting meridional flow, as described by the Sverdrup-Ekman equations.
For times when the climate was nonglacial, it is difficult to postulate the extent of the variability of the seasons. However, with a rotating spherical Earth, there always will be latitudinal differences in insolation and heat flux away from the zone of maximum insolation at the meteorological equator. Seasons occur as a result of the angle of the pole of rotation to the plane of the ecliptic, causing a variation in the length of the diurnal light-dark cycles both annually and latitudinally. Only if the pole of rotation was perpendicular to the plane of the ecliptic, so that light-dark cycles would be the same latitudinally, would there not be any seasons. However, there still would be climatic belts and a latitudinal difference in insolation with the heat flux from the equator to the poles. Vanyo and Awrarnik (1985), using data derived from stromatolites, indicated that the obliquity of the Earth's rotation to the ecliptic at 850 Ma was similar to the modem. Thus, seasonality in the Ordovician similar to modern conditions is not unreasonable.
The Ordovician was prior to the development of vascular land plants that reduce the land albedo. The terrestrial albedo in the Ordovician was probably high, similar to that of deserts. Therefore, even with a relatively mild global climate, if the landmasses were in the appropriate locations, regional modifications of the wind patterns, such as monsoons, could occur. The extent of modification would be a function of the seasonality and the areal extent of the land and its latitudinal position. Because large contiguous landmasses occurred on the Gondwanan continent in the Ordovician, it may be assumed that monsoons also would have occurred then. In this paper, due to lack of land in the northern hemisphere, figures designated northern hemispheric summer typify mean annual nonseasonal conditions. Due to the presence of land in the southern hemisphere, figures designated northern hemispheric winter or austral summer show the maximum reversal effects of the austral monsoons.
Both the Ekman transport and the velocity of geostrophic currents vary with the angular rotation of the Earth or planetary vorticity (Defant, 1961, p. 383). If the number of days in a year was significantly more in the Paleozoic (410 days in a year in the Silurian [Mazzullo, 1971]), then the geostrophic current velocities would be about 15 per cent more for the same pressure gradient seen today and net transport of water arid circulation would be more vigorous. No current velocities or transport volumes are expressly quantified here, so any change in the artgular rotation of the Earth would have a relative effect.
The International Union of Geological Sciences (IUGS) has published correlations charts for Australia, New Zealand, and Antarctica (Webby et al., 1981); Canada (Barnes et al., 1981); China (Shen-Fu, 1980); Kazakhstan and Middle Asia (Nikitin et al., 1986); the Near and Middle East (Dean, 1980); South America (Acenolaza and Baldis, 1987); southwestern Europe (Hamman et al., 1982); and the United States (Ross et al., 1982). The major lithofacies patterns for intervals wilhhl the Ordovician may be constructed on the paleogeographic maps from data from these charts. Thick carbonate sections are found in blocks within the tropical regions as plotted by Scotese (1986). Throughout the Ordovician, North America, Australia, North China, Siberia (Taimyr) and parts of Kazakhstan were sites of carbonate sequences presumably deposited in shallow shelf seas. Starting in the Caradoc, the Balto-Scandia-Russian Platfomm also showed shallow shelf carbonate deposition. Thick shallow shelf sea accumulations of calcium carbonate only form in tropical and subtropical latitudes as a function of maximum light availability, because year-around light refraction drops markedly at about 35 degrees from the equator, which is the present poleward limit of Bahamian environments.
As Berry and Boucot (1973) discussed, based on Latest Ordovician-Early Silurian worldwide correlations, glacial and glaciomarine strata developed across North Africa, eastem South America, the modem Middle East, and some sites in Spain during the Late Ordovician. Simultaneously, glacioeustatic sea level lowering worldwide led to exposure of wide areas of fommer continental shelf. The effects of major continental glaciation about a pole in central Africa are seen in the stratigraphic record both in the forth of glacial and glaciation-related deposits, and in the form of a major depositional hiatus. The geographic location of the glacial beds is consistent with the plate reconstructions from remanent magnetism indicated by Scotese (1986).
Ordovician graptolite faunal provincialism has been related to oceanic surface water temperatures by Skevington (1974,1976). He pointed out that the most species-rich faunas are those that occur about the craton-platform areas that were sites of thick accumulations of shallow shelf carbonates. The species-rich faunas, those of the "Pacific" Faunal Region. were most likely tropical. The other graptolite faunas of the Ordovician (e.g., Erdtmann, 1984) have been grouped in the "Atlantic" Faunal Region. Such appellations using modem geographic names can be misleading, particularly if the the faunal region was based on climatic conditions. According to the present reconstructions, the "Pacific" region actually encompassed the tropical and subtropical areas, whereas the ''Atlantic" region was in middle to high southern latitudes.
Based on considerations of prominent lithofacies patterns and on a consistent biogeography for graptolites, the rock and fossil data regarding the latitudinal (climatic zonal) positions for major cratons and shelf seas are consistent with paleomagnetic data. Accordingly, the maps showing major marine surface currents and water masses for the various depicted time intervals from the Late Cambrian to Early Silurian reflect consideration of three lines of evidencebearing on paleogeography: paleomagnetism, lithofacies pattems, and biogeography.
Figures I to 6 depict the principal features of the surface oceanography from the Late Cambrian through the Early Silurian. The following text concerning each of the maps is, in effect, an expanded figure caption showing how the basic oceanographic principles are used to reconstruct the oceanoaphic features. Specific references to sections orcolumns from the JUGS compilation charts or other compilations are given in the text to demonstrate the interaction among paleogeography and lithofacies to reconstruct the paleo-oceanography.
LATE CAMBRIANFigures I A, B represent the surface oceanic current patterns in the Late Cambrian, and set the scene for the Ordovician. The major feature in the north polar region is open ocean, with continental blocks all south of 35 N.
A zonal Boreal Ocean covered the north polar and north temperate zones. This ocean had two embayments extending to the equator between northern Gondwana (New Guinea and Australia) and Laurentia, each with a longitudinal extent of almost 180 . Depending on the level of these blocks in the northern hemisphere, the actual Boreal Ocean may have extended to the equator and covered the whole northern hemisphere. Temperate to cool water dominated in the Boreal Ocean due to the lack of significant meridional barriers to advect wamm water from the tropical areas. Clockwise circulation in the two weak high pressure systems in the two boreal embayments advected cool water equatorward along the west coast of North America and New Guinea, and wamm water polarward in the Arctic Islands and south Australia.
Emergent land on the equator included the Laurentian, Siberian, and Australian-Antarctic shields. Thus, the global oceanic low pressure systems probably were fragmented, with the major systems south of the high pressure centres of the Boreal Ocean embayments. The water current flow was east to west, parallel to the equator.
The African, South American, and Antarctic shield areas were in the southern hemisphere as part of Gondwana. A relatively small Austral Ocean was divided by separate blocks consisting of Siberia, Kazakhstan, and Balto-Scandia and the Russian Platform. These barriers produced two counterclockwise oceanic high pressure systems, which moved warm water south along eastern North America and Siberia and Kazakhstan, and cool water north along western Siberia and through Turkey, Iran, and South China. The Austral Ocean was connected to the Boreal Ocean on the west through a seaway between Laurentia and Gondwana (South America). This would produce a third counterclockwise oceanic high pressure circulation gyre in the southern hemisphere. The Gondwanan shields form the only continuous meridional barrier on the eastern side of the Austral Ocean. Thus, the easternmost oceanic high pressure gyre would be the most coherent. The two western high pressure gyres have limited eastern barriers, so would have relatively weak northern meridional flow on the eastern side. The two westem-most gyres are commonly combined into the Iapetus Ocean, whereas the easternmost gyre is included in the Proto-Tethyian Ocean.
The Austral Ocean was open to the south pole with the Balto-Scandian barrier at about 60 S. Circulation was clockwise, governed by the oceanic low pressure.
Figure I A shows the effects of monsoons produced by the concentration of continental blocks in the southern hemisphere. During the southern hemisphere summer, anti-trade monsoon winds would be generated along the coastal and platform areas of Gondwana. In contrast to the equatorial currents shown in Figure I B, these monsoonal currents would drive warm water eastward and polarward through North China, Bomeo, Sumatra, New Guinea, and South China, countering the cool flow from the midlatitude oceanic high pressure systems. In the northern hemisphere, cool water would be advected southward along New Zealand and the east coast of Australia.
Tropical divergences would occur along the equator with the most persistent in the Arctic Islands and southern Siberia. Due to monsoonal current reversals, upwelling occurred along the equator during the northern hemisphere summer in North China, New Zealand-Australia, and southern Laurentia. Near 60 S, planetary divergences occurred in North Africa, Iberia-Amorica, Balto-Scandia, and Turkey and the Near East. Seasonal divergences near 30 S in the austral summer produced upwelling off western coasts in South China. Upwelling would persist in the austral winter as a result of offshore advection due to the meeting of monsoonal warm and equatorward cool currents.
EARLIEST ORDOVICIAN: TREMADOCThe major paleogeographic changes from the Upper Cambrian were I ) the shift of both Laurentia and Gondwana to the south, increasing the latitudinal extent of the zonal Boreal Ocean; 2) the northward drift of a) Balto-Scandia, closing the Iapetus gap and increasing the width of the Tomquist Sea between Balto-Scandia and the parts of Western Europe attached to North Africa and the main Gondwanan continent and of b) Siberia, maintaining the open ocean connection between Iapetus and Proto-Tethys in the Austral Ocean (Figure 2 A, B).
The zonal portion of the Boreal Ocean expanded to about 30 N, limiting meridional flow in the northern hemisphere to the tropic zone with a weak development of oceanic high pressure in subtropical to southern north temperate regions. With the shift of the major landmasses to the south, the westward flowing north equatorial current expanded meridionally in the tropical regions. The preponderance of zonal flow in the northern hemisphere would produce a temperature distribution in the surface ocean, with isotherms parallel to lines of latitude with minor deflections of cold water to the south on the eastern side (Western Canada and New Guinea), and of warm water to the north on the western side (Arctic Islands and New Zealand) of the boreal embayments.
In the southern hemisphere, the westward flowing south equatorial current was segmented by the the Laurentian, Siberian, and Gondwanan shields. The northern movement of Balto-Scandia and the southerly shift of South America combined to merge the oceanic high pressure gyres of the Upper Cambrian into one gyre in the Iapetus Ocean with weak northerly flow on the eastern side. Circulation in the midlatitude Proto-Tethys oceanic high would be strong on the western side and weak on the southerly limb at the juncture with Iapetus.
At high southern latitudes, the Tomquist Sea and its Proto-Tethys counterpart were astride the planetary oceanic low pressure system producing strong clockwise circulation ccntred at about 60 S.
During the austral summer in the Tremadoc (Figure 2 A) current reversals would develop in the subtropical regions along the shores of the Gondwanan continent similar to those hi the Upper Cambrian. Although the direction of the equatorial current during the austral monsoon would be reversed (actually, by modern analogy the equatorial countercurrent) to flow east; there would be little thermal change in the northern hemisphere due to the already weak meridional flow. The major changes in temperature would be in the south. hemisphere in South China where cool water advecting from the south would be replaced by warm tropical water driven by the monsoon.
Major tropical divergences would occur along the equator in western America (Mayflower Schist, Ross et al. , no.14) and Arctic Islands in Laurentia, Siberia, North China (Hsienerhtai Em, Erdtmann , no. 19), Weddel Sca (Antarctica), and southernmost South America. In the southern hemisphere, the planetary divergence at 60 S would affect North Africa, Maguma (Barnes et al. , no. 68), Avalonia (Bames et al. , no. 87), South Ireland and Britain, the Baltic area (Barnes et al. , no. 66), the southern Ukraine, Asia Minor (Karadere Formation, Dean , no. 4), the Levant, and northwestern Arabia.
Less intensive seasonal upwellings would occur in the summer in southwestern China (Pingtou Fonnation, Shen-Fu , no. I) to Iran south of 30 S, augmented by offshore flow caused by the Gondwanan monsoonal currents merging with the cool eastern gyro of the oceanic Proto-Tethyan high. Depending on the latitude of Laurentia, potential upwelling would occur on the east side of the Iapetus high pressure gyre from eastern Canada (Levis Formation, Bames et al. , no. 66), Scotland, and northern Newfoundland (Cow Head, Barnes et al. , no. 74). In the northern hemisphere, upwelling might occur in New Zealand (Webb Formation, Webby et al. , no.59) as a result of offshore flow where the northern branch of the Gondwanan monsoonal current flowing south meets the easterly deflection of the north equatorial current.
EARLY ORDOVICIAN (ARENIG)The major change in the Arenig (Figure 3 A, B) from the Tremadoc was the continued northerly drift of Balto-Scandia and Siberia, and the southerly drift of Gondwana. Southern Ireland and Great Britain, and Iberia-Amorica began to drift northward, away from North Africa and Gondwana.
Zonal flow continued to dominate the circulation in the north polar and temperate belts. The movement of the Siberian Shield astride the equator deflected the west-flowing north equatorial current slightly. In the southern hemisphere, the twin barriers of Balto-Scandia and Siberia emphasized the independent oceanic midlatitude high pressure systems in the Iapetus and Proto-Tethyian oceans. However, some transport probably occurred, of cool water from Iapetus eastward across the north side of Balto-Scandia, and of warm water from tropical Proto-Tethys westward across Siberia and Kazakhstan. At high latitudes, clockwise circulation about the oceanic low pressure system centered at 60 S and the Tornquist Sea persisted, connecting Iapetus and Proto-Tethys on the south. The increase in land in the south polar zone probably reduced the transport volume in the southern part of the gyre, which would cause increased northerly advection along the west coast of Balto-Scandia, or more likely, against the Gondwanan barrier to the east of Balto-Scandia.
The effect of the Gondwanan monsoon probably vanished in the northern hemisphere due to the continued southerly drift of the main Gondwanan continent and concentration of landmasses in the southern hemisphere. however, as a consequence of this, monsoonal conditions intensified in the subtropical regions of the southern hemisphere. During the austral summer, monsoonal warm water would flow south through South China and into Central Asia, whereas during the austral winter the monsoon would intensify the planetary oceanic high pressure flow of cool water northward into the subtropics even as far as North China.
With landmasses on the equator, the planetary equatorial divergence was segmented into three parts. The longest extends from the west coast of Laurentia (Phi-Kappa Formation, Ross et al. , no. 9 and Kanosh Formation, Ross et al.  no. 25) to the east coast of Gondwana (Victoria, Australia, Webby et al.  no. 23; Lancefield, Erdtmann , no. 17). Smaller segments of the equatorial oceanic low pressure zone were between Laurentia and Siberia and between Siberia and Gondwana, over what is now Kazakhstan, North China, Sumatra, and Borneo. Upwelling in the Gondwanan sector would be reduced by the monsoon, especially south of the equator.
High latitude divergence at about 60 S in the southern hemisphere would be seen in Venezuela (Cerrojon Formation, Acenolaza and Baldis , no.59); southern Appalachia (Suwanee Basin, Ross et al. , no. 100), northwest to west Africa, Nova Scotia (Halifax, Barnes et al., no. 67); Avalonia (Wabana Formation, Bames et al.[l981], no. 83); Wales, Denmark to the Ukraine on Balto-Scandia (Lindstrom and Vortische ); northwestern Arabia-Asia Minor (Bosporus, Dean , no. 4; Sedisehir Formation, Dean , nos. 5,6,7,8-12).
Summer seasonal upwelling in the southern hemisphere would occur at midlatitudes off the western coasts of Central America-southern Mexico, and northern Norway-Spitzbergen (Valhalfonna, Archer and Fortey ; Fortey ), with the best developed system in western South China (Ningkou Shale, Shen-Fu , nos. XV,XVI; Chungyi Formation, Shen-Fu , no. XVII, and Indo-China).
MIDDLE ORDOVICIAN: (LLANDEILO-CARADOC)In the Middle Ordovician (Figure 4 A, B), Gondwana continued its southern drift. The major change from the Arenig was the formation of an eastern meridional barrier in the Iapetus Ocean, because of the merging of Laurentia and Siberia in the tropics and the near attachment of Balto-Scandia to Siberia. In high southern latitudes, the Rheic Ocean opened to separate the Meguma, Avalonia, South Ireland-Great Britain, Belgium sliver from Iberia-Amorica and North Africa. Accordingly, the Iapetus and Proto-Tethyan oceans were more sharply delineated, although a seaway connection probably existed between them.
In the northern hemisphere, zonal circulation continued in the Boreal Ocean north of 30 N, potentially extending to the equator with only weak meridional flow west of Laurentia and east of New Zealand-Australia. Due to the shift of Gondwana toward the south. the norm equatorial current flowed westward, relatively unimpeded from the west coast of Laurentia to Siberia with only the Laurentian and Siberian shields under shallow water or emergent. The south equatorial current was blocked on the west by the Australian-Antarctic shields, but re-formed in Southeast Asia through North China and Kazakhstan until the current was blocked by the Siberian Shield.
At midsouthern latitudes, three oceanic high pressure gyres were maintained. With South America drifting south of 30 S, a distinct high pressure system could develop in the Proto-Pacific. This system was linked with the Iapetus anticyclonic gyre in the gap between Mexico and South America south of 30 S. At this time, both Iapetus and Proto-Tethys had well developed meridional barriers, although eastward flow from Iapetus probably leaked into Proto-Tethys via the Tomquist Sea. The circulation in the Rhaeic Ocean and the southemmost part of Proto-Tethys were governed by cyclonic flow about the 60 S oceanic low pressure system.
Monsoonal current reversals during the austral summer would occur along the northwest shores of Gondwana (Figure 4 A) due the continuing presence of the large landmass to the south.
Divergences (up welling)
The planetary divergence at the equator was divided into two major oceanic low pressure systems with the most persistent open ocean system located between the west coast of Laurentia (Phi Kappa, Ross et al. , no.9; Vimini/Valmay, Ross et al. , nos. 12,13,14,17,18; in Canada: Skoki Formation, Bames et al. , nos. 4,5,7,8,9 and Road River Formation, Barnes et al. , nos. 10,12,13) and the north coast of Gondwana (Victoria, Webby et al. , nos. 23,26,27). The system between Gondwana and Siberia would be modified in the southern hemisphere by the austral winter monsoons with modified upwelling in eastern Kazakhstan -and central North China.
In the vicinity of 60 S, upwelling along the planetary divergence would occur in southern Appalachia (Suwanee US, no. 100); west Africa, Iberia (Hamman et al. , nos. 27,28); west Armoricia (Urville, Le Pissot, Hamman et al. , nos.3,4,5; Schistes d'Angers, Hamman et al. , nos. 7,8); Belgium, and northwest Arabia (Jordan, Dean , no. 20; Arabia Hanadir Shale, Dean , no. 21).
Seasonal upwellings in the summer along western coasts of middle latitude oceanic high pressure systems would occur in the southern hemisphere of Central America, western Norway, and western South China and Indo-China (Shen-Fu , nos. XIII,I). The seasonal upwellings in China at middle latitudes would be augmented by the divergence of the currents caused by the Gondwanan summer monsoon.
LATE ORDOVICIAN (ASHGILL)Gondwana continued to move south with the South American, African and parts of the Antarctic shields in high southem latitudes astride the south pole (Figure 5 A). The Tomquist Sea closed, linking the Meguma through Belgium blocks with Balto-Scandia. Iapetus narrowed with a seaway at the equator separating Laurentia-Siberia from BaltoScandia. The Rhaeic Ocean widened, which acted as the major water communication between the widening South Pacific and Proto-Tethys.
Zonal flow in the Boreal Ocean continued above 30 N. The alignment of the Laurentia, Siberia, and Balto-Scandia blocks, which reduced Iapetus essentially to an embayment, created a meridional barrier in the tropical regions with significant meridional flow on both sides of that block. The most persistent equatorial currents flowed to the west between Laurentia and northern Gondwana. The major oceanic high was in the South Pacific, with smaller highs in Proto-Tethys and Iapetus. With ice in the south polar regions and limited tropical circulation in the northern Iapetus and Proto-Tethys, the surface waters in the southern hemisphere, except in the western South Pacific, were noticeably cooler than in the Middle Ordovician. Circulation about the 60 S oceanic low was limited by the extensive landmasses and the possibility of ice shelves covering significant parts of the gyre. With glaciation in the southern hemisphere and the lowering of sea level, the areal extent of the shelf seas was limited so that the meridional barriers would be more unifomm. These conditions would lead to more intense circulation in the open ocean.
Glaciation in the Ashgill produced conditions favoring the formation of cold deep water during the austral winter, which could affect the depositional environment on the continental rise as seen in deposits on deep sea fans. Prior to the glaciation, deep sea fan deposits such as at Dob's Linn, Scotland were anoxic black shale. For pre-Ashgill times, in tropical areas, the waters in the lower part of the mixed layer and the pycnocline were derived from the sinking of relatively warm central water at the subtropical convergence. However, the upper Ordovician deposits are more oxic mudstone, indicating ventilation of middle depths. The narrowing of Iapetus and the position of the Gondwanan shore near the pole suggests that Iapetus was ventilated by cold deep water sinking during the formation of sea ice in the austral winter. Also, during glaciation, subpolar water must have been colder and more oxygen-rich. This water, also sinking at the subtropical convergence, may have chilled the overlying central water and ventilated it by vertical advection. Thus, although Dob's Linn was associated with tropical Laurentia, at this time, with wamm surface waters, the water in the pycnocline overlying the offshore fans was cold and relatively oxygen-rich. On the other hand, deep sea fan deposits on the continental slope and rise from the late Ordovician of China continue to be anoxic black shale. Accordingly, cold water sinking from the Gondwanan ice shelf at the south polar convergence might have been blocked from entering the Proto-Tethyan Ocean due to the bathymetric ridge-rise associated with the opening of the Rhaeic Ocean. The ridge apparently was high enough to, at least partially, block subpolar water sinking at the subtropical convergence, permitting dilution by the larger volume of central water formed in the larger Proto-Tethys. Altemately, this deep water may not have contained sufficient oxygen to vertically ventilate the pycnocline or to compensate for the organic productivity in Proto-Tethys.
Monsoonal circulation in the Indo-Australian-Antarctic regions of Gondwana likely intensified during glaciation, particularly along the northeastern subtropical coast.
Equatorial upwelling along the planetary divergence would persist from the west coast of Laurentia (Antelope Valley, Ross et al. , no. 17; Road River Formation, Barnes et al. , no. 12) to the east coast of Gondwana (Australia, Melbourne, Webby et al. , nos. 25,26,27; Victoria [Webby et al. , nos. 29,30,31). The austral polar latitude divergence was limited by ice and land with upwellings seen in Iberia (Orea Shale, Harnman et al. , nos. 23,24,26); western France (glaciomarine pelites and fragments, Hamman et al. , nos. 1,2,3); and northwest Arabia (Syria, Dean , no. 19).
LOWER SILURIAN (LLANDOVERY)Deglaciation occurred with the shift of Gondwana away from the south pole (Figure 6A, B) Siberia moved above 30 N, reducing the zonal Boreal Ocean. Laurentia moved south as South America moved north, constricting the connection between the Proto-South Pacific with the austral Iapetus and Rhaeic oceans. This constriction would force more cool water along the coast of Central America and Mexico in a manner similar to the modern Peru-Chile (Humboldt) Current, which is enhanced by the constriction of the Drake Passage. The limited supply of cool water moving into the southern Iapetus and Rhaeic oceans permitted more warm water from the tropics to flow into these areas, particularly on the western sides of the oceans.
SUMMARYGiven the basic assumptions discussed above, oceanographic conditions in the Ordovician have been reconstructed. The characteristic features of oceanic conditions in the Ordovician include:
1. Zonal circulation in the polar and most of the temperate regions of the northern hemisphere due to a globe girdling Boreal Ocean.
2. Meridional barriers in the tropical northern hemisphere, which permitted weak north-south flow in the sub-tropics.
3. Westward-flowing equatorial currents were discontinuous until the Late Ordovician as the result of numerous equatorial landmasses and shallow seas. Although landmasses would impede equatorial zonal flow, a mixture of tropical and subtropical waters would produce vast areas of central water with relatively uniform temperature conducive to carbonate deposition in shallow areas. Planetary divergence and upwelling near the equator would be limited by the reduced wind fetch caused by the discontinuity of true oceanic depths there.
4. Three small to modest-sized oceans existed at middle latitudes in the southern hemisphere, with at least partial meridional barriers. The easternmost, the Proto-Tethys, had the most persistent barrier on its eastern shore. As a consequence of the north-south barriers, warm water flowed polarward on the western side of the oceans and cool water flowed equatorward along the eastern side Seasonal upwellings would enhance productivity at middle latitudes along the eastern coast of these oceans.
5. Monsoonal conditions produced by the concentration of landmasses in Gondwana produced seasonal current reversals on the subtropical Gondwanan shore of ProtoTethys. This permitted mingling of tropical (old Pacific) and subpolar/temperate (old Atlantic) faunas.
6. Austral polar and subpolar areas were land during most of the Ordovician. Increased productivity at the planetary divergence at the polartemperate boundary would have occurred in embayments of the three southern oceans.
7. Pycnoclinal and deep circulation was governed chiefly by sinking of warm central water until the Late Ordovician. During glaciation, colder and denser intermediate and deep waters ventilated the pycnocline, particularly in the narrow Iapetus Ocean. However, tectonic ridge-rise systems, such as those that produced the Rhaeic Ocean could block pycnoclinal ventilation in tropical regions of ProtoTethys.
Consequences of the evolving Ordovician oceanic circulation patterns would be: 1) enhanced upwelling (and increased organism diversity) on the southeastern side of a bounded southern ocean (Iapetus in the early Ordovician); 2) inhibited upwelling at the equatorial planetary divergence during the presence of the equatorial barrier; 3) significant deep ventilation in the global ocean during glaciation; and 4) low new nutrient availability in an isolated Boreal Ocean.
ACKNOWLEDGMENTSI thank C.R. Barnes, E.G. Kauffman, R.J. Ross, and O.H. Walliser for their encouragement of our interdiscipli- nary activities. W.B.N. Berry, M.S. Quinby-Hunt, and an anonymous reviewer contributed greatly to the clarity of the manuscript. M. Krup did her usual heroic job in compiling the various drafts into clear and concise illustrations. This is contribution number MSG 8Oo-005 of the Marine Sciences Group of the University of California-Berkeley.
REFERENCESAcenolaza, F.G. and Baldis, B. 1987: The Ordovician System of South Amcrica. International Union of Geological Sciences, Publicalion no. 22, 64 p.
Archer, J.B. and Fortey, R.A. 1974: Ordovician graptolites from the Valhallfonna Formation, northern Spitsbergen. In Graplolite Studies in Honour of O.M.B. Bulman, R.B. Rickards, D.E. Jackson and C.P. Hughes (eds.); The Palaeontological Associalion, London, p. 87-97.
Barnes, C.R. 1986: The faunal extinction event near the Ordovician-Silurian boundary: a climatically induced crisis. In Global Bio-Events, O. Walliser (ed.); Springer-Verlag, Berlin, p. 121-126. Barnes, C.R., Norford, B.S., and Skevington, D. 1981: The Ordovician System in Canada. International Union of Geological Sciences, Publication no. 8, 27 p.
Barron, E. J. 1981: Paleogeography as a climatic forcing factor. Geologishes Rundschau, v 70, p 737-747.
Berggren, W.A. and Hollister, C.D. 1977: Plate tectonics and paleocirculation-commotion in the ocean. Tectonophysics, v. 38, p. 1148.
Berry, W.B.N. and Boucot, AJ. 1973: Glacio-euslalic control of Late Ordovician-Early Silurian plalform sedimentation and faunal changes. Geological Society of America, Bulletin. v. 84. p. 275-284.
Brenchley, P.J. 1984: Late Ordovician extinctions and their relationship to the Gondwana glaciation. In Fossils and Climate, P.J. Brenchley (ed.); John Wiley, New York, p. 291-316.
Cocks, I..R.M. and Fortey, R. A. 1982: Faunal evidence for oceanic separations in the Palaeozoic of Britain. Journal of the Geological Society of London, v. 139, p. 465-478.
Dean, W.T. 1980: The Ordovician System in the Near and Middle East. International Union of Geological Sciences, Publication no. 2. 22p.
Defant,A. 1961: Physical Oceanography. Pergamon, Oxford v. 1 729 p., v 2, 598 p.
Ekman, V.W. 1905: On the influence of the Earth's rotation on ocean currents. Arkiv for matematik, astronomi, och fysik, v. 2, no. 11, p. 1-52.
Erdtmann, B.-D. 1982: Palaeobiogeography and environments of planktic dictyonemid graptolites during the earliest Ordovician. In The Cambrian-Ordovician Boundary: Sections, Fossil Distributions and Correlations, M.G. Bassett and W.T. Dean (eds.); National Museum of Wales Geological Series no. 3, p. 9-27.
_____1984: Outline ecostratigraphic analysis of the Ordovician graptolite zones in Scandinavia in relation to the paleogeographic disposition of the lapetus Ocean. Geologica et Palaeontologica, v. 18 p. 9- 15.
Fortey, R.A. 1984: Global earlier Ordovician transgressions and regressions and their biological implications. In Aspects of the Ordovician System, D.L. Bruton (ed.); Palaeontological Contributions from the University of Oslo, p. 37-50.
Frakes, L.A. 1979: Climates Throughout Geologic Time. Elsevier, Amsterdam,310 p.
Hamman, W., Robardet, M., and Romano, M. 1982: The Ordovician System in Southwestern Europe. International Union of Geological Sciences, Publication no. 11, 47 p.
La Fond, E.C. and La Fond, K.G. 1971: Oceanography and its relation to marine production. In Fertility of the Sea, J.D. Costlow (em); Gordon and Beach, New York, p.241-265.
Lindstrom, M. and Vortisch, W. 1953: Indications of upwelling in the Lower Ordovician of Scandinavia, Part B. In Coastal Upwelling, J. Thiede and E. Suess (eds.); Plenum, New York. p. 535-551.
Mazzullo, S. J. 1971: Length of the year during the Silurian and Devonian periods Gcological Society of America, Bulletin, v. 82, p. 1085,1086.
McGowan, J.A. 1972: The nature of oceanic ecosystems. In The Biology of the Oceanic Pacific, C.B. Miller (cd.): Oregon University Press Corvallis, Orcgon, p. 9-28.
Munk, W.H. and Carrier, G.F. 1950: The wind-driven circulation in ocean basins of various shapes. Tellus, v. 2, p. 158-167.
Nikitin, I.F., Apollonov, M.K., Tzaj, D.T., Koroijov, V.G., Kim, A.l., Erina, N.V., Larin, N.M., and Golikov, A.N. 1986: The Ordovician System in Kazakhstan and Middle Asia. Iniernational Union of Geological Sciences, Publication no. 21, 34 p.
Parrish, J.T. 1982: Upwelling and petroleum source beds. with reference to Paleozoic. American Association of Petroleum Geologists, Bulletin v. 66. p. 750-774.
Ross, R. J. 1975: Early Paleozoic trilobites, sedimentary facies, lithospheric plates and ocean currents. Fossils and Strata, no. 4, p. 307-329.
Ross, R.J., Adler, F.J., Amsden, T.W., Bergstrom, D., Bergstrom, S.M., Carter, C., Churkin, M., Cressman, E.A., Derby, J.R., Dutro, J.T., Ethington, R.L., Finney, S.C., Fisher, D.W., Fisher, J.H., Harris, A.G., Hintze, L.F., Ketner, K.B., Kolata, D.L., Landing, E., Neuman, R.B., Sweet, W.C., Potjeta, J., Jr., ['otter, A.W., Rader, E.K., Repetski, J.E., Shaver, R.H., Thompson, T.L., and Webers, G.F. 1982: The Ordovician System in the United States. International Union of Geological Sciences, Publication no. 12, 73 p.
Scotese, C.R. 1986: Phancrozoic reconstructions: a new look at the assembly of Asia University of Texas Institute for Geophysics, Technical Repon no.66.,54p.
Shen-Fu, S. 1980: The Ordovician System in China. International Union of Geological Sciences, Publication no. 1, 7p.
Skevington, D. 1974: Controls influencing the composition and distribution of Ordovician graptolite faunal provinces In Graptolite Studies in Honour of O.M.B. Bulman, R.B. Rickards, D.E. Jackson and C.P. Hughes (eds.); The Palaeontological Association, London, p. 59-73.
1976: A discussion of the factors responsible for the provincialism displayed by graplolite faunas during the Early Ordovician. In Graploliles and Slrafigraphy, D. Kaljo and T. Koren (ads.); Academy of Sciences of Estonian SSR Institut of Geology, Tallinn, p. 180-201.
Stehli, F.G. 1965: Paleontologic technique for defining ancient ocean currents. Science, v. 148. p. 943-946.
Stommel, H. 1948: The westward intensification of wind-driven ocean currents. Transactions of the American Geophysical Union, v. 29, p. 202-206.
_____1957: A survey of ocean current theory. Deep-Sea Research, v.4,p.149-184.
Sverdrup, H.V. 1947: Wind-driven currents in a baroclinic ocean; with application to the equatorial currents of the eastern Pacific. Proceedings of the National Academy of Sciences, v. 33, p. 318-326.
Sverdrup, H.V., Johnson, M.W., and Flemming, R.H. 1942: The Oceans. Prcnticc-Hall, Engiewood Cliffs, NJ., 1042 p.
Tchernia, P. 1980: Descriptive Regional Oceanography. (translated by D. Densmore). Pergamon Press, Oxford, 253 p.
Vanyo, J.P. and Awramik, S.M. 1985: Stromatolitcs and Earth-Sun-Moon dynamics. Precambrian Research, v. 29. p. 121-142.
Von Arx, W.S. 1962: Introduction to Physical Oceanography. Addison-Wesley, Reading, Massachusetts, 422 p. Webby, B.D. et al. (eds.) 1981: The Ordovician System in Australia. New Zealand and Antarctica. International Union of Geological Sciences, Publication no.6.,64 p. .
Wilde, P. and Berry, W.B.N. 1986: Role of oceanographic factors in the generation of global bio-events. In Global Bio-Events, O. Walliser (ed.); Springer-Verlag, Berlin, p. 75-91.