Climate in Earth History: Studies in Geophysics (1982)

GARRETT W.BRASS, E.SALTZMAN, J.L.SLOAN II, and J.R.SOUTHAM

Rosenstiel School of Marine and Atmospheric Science

WILLIAM W.HAY

Joint Oceanographic Institutions, Inc.

W.T.HOLSER

University of Oregon

W.H.PETERSON

Cooperative Institute for Marine and Atmospheric Studies

Studies of climate frequently involve identifying a plausible forcing mechanism (e.g., solar fluctuations), hypothesizing the response, and attempting to verify the hypothesis with data. This is difficult because the mechanisms are many and frequently small in amplitude, the response complex, and the data meager. For changes in oceanic mixing on the geologic time scale, a significant forcing function has been the change in area, location, and configuration of marginal seas due to plate-tectonic motions and eustatic sea level fluctuations. The hypothesized response is that these changes influence the formation of oceanic deep water and cause the thermohaline circulation of the ocean to differ substantially from one age to another. We suggest that during much of the geologic past (especially in the late Cretaceous) bottom water was produced by the sinking of warm, salty water formed by evaporation in low-latitude marginal seas rather than by the sinking of cold water formed in polar and subpolar marginal seas. The possibility of the formation of warm saline bottom water (WSBW) and its role in climate was first suggested, to our knowledge, by Chamberlin in 1906.

We begin by discussing thermohaline circulation and the formation of oceanic deep water and suggest that marginal seas play a dominant role in deep water formation as well as in determining the characteristics of deep water. In the next two sections we present a simple model that provides a conceptual framework for evaluating the potency of marginal seas as deep-water sources and for understanding the formation of WSBW. The substantial geologic evidence for the occurrence of WSBW is the topic of the fourth section. We then show that paleogeography indicates that the configuration of marginal seas was favorable for the formation of WSBW at those times in the past when there is evidence in the geologic record (presented in fourth section) for the occurrence of WSBW. Finally, we present discussions of the consequences of WSBW on oceanic and atmospheric chemistry and on the climatic consequences of WSBW.

Thermohaline circulation is generally taken to be that circulation driven by density differences imposed at the ocean surface by interaction with the atmosphere. The bulk of the present ocean is filled with water from only a few source regions of restricted surface areas (e.g., the Norwegian Sea, the Weddell Sea, the Mediterranean Sea, and the Red Sea-Persian Gulf). This is verified by the fact that these deep waters essentially retain characteristics of their source regions. Simple box models using both radiocarbon and stable tracers (see Broecker et al., 1960; and Bolin and Stommel, 1961) yield estimates of the residence time of the deep water of about 1000 yr. If the ocean is assumed to be in a quasi-steady state (i.e., changes, if any, occur on time scales much greater than 1000 yr), the sinking of deep water in a few small regions must be compensated for by rather slow upwelling in the bulk of the ocean. If this upwelling were uniformly distributed (which it almost certainly is not) the upwelling velocity, based on the 1000-yr residence time, would be on the order of a few meters per year. In the present-day ocean the upwelling of dense, cold, deep water must be compensated for by a downward diffusive flux of negative buoyancy (heat in the present-day ocean) if the ocean is in a steady state.

Because of the small spatial scales and temporal variability of convective processes such as deep-water formation, there are few direct observations of deep-water formation. Furthermore, the small magnitude of the upwelling velocity precludes its direct observation. Finally, the details of oceanic diffusive processes are poorly known; therefore, most knowledge of thermohaline circulation and deep-water formation is based on indirect observation (e.g., the radiocarbon models mentioned above) and theoretical considerations. Because of these problems, knowledge of present-day thermohaline circulation is rudimentary compared with knowledge of the wind-driven circulation.

Rossby (1965) suggested that because the convective buoyancy transfer processes are so much more efficient than the diffusive processes, convective deep-water formation must be confined to small regions in order to maintain steady-state conditions. Observations indicate (see Gordon, 1975) that most deep water is formed in marginal seas or over continental shelves. Apparently, the trapping of water allows it to undergo intensive interaction with the atmosphere and thereby create a substantial buoyancy deficit. This suggests that accidents of geography may determine the sites of deep-water formation.

Models of thermohaline circulation have tended to treat the formation of deep water as a tacit assumption or as a response to global-scale forcing. In light of the apparent importance of marginal seas to deep-water formation, one of us (Peterson, 1979) has formulated a simple steady-state thermohaline convection model driven by imposed buoyancy sources. Because so little is known about thermohaline circulation, a simple horizontally averaged model exploring how buoyancy sources determine the stratification and vertical circulation seemed appropriate. The essential assumption of the model is that dense water sources have a sufficient buoyancy deficit to drive turbulent plumes that entrain water from the ocean’s interior, thus increasing the plume’s volume transport and decreasing their density as they penetrate. The present-day Mediterranean outflows lend credence to the entrainment assumption. The density of the outflow at the Straits of Gibraltar is greater than any density in the Atlantic Ocean (Wüst, 1961), but the Mediterranean outflow terminates at about 1200 m and spreads throughout virtually the entire North Atlantic as an intermediate salinity maximum. If the Mediterranean water were not diluted by entrainment it would sink to the bottom of the Atlantic Ocean.

The model is an extension of the pioneering work of Baines and Turner (1969) and can be envisioned as the filling of a large box from one or more steady buoyancy sources. The buoyancy sources drive turbulent plumes that interact with a laterally well-mixed interior by turbulent entrainment of nonturbulent interior fluid into the plume. The interior equation is a simple vertical balance of advection and diffusion of buoyancy. This interior balance is the same one much used by some geochemists (e.g., Craig, 1969), with the important exception that the vertical velocity is determined by the source strength and the plume dynamics.

The buoyancy flux is the product of the volume flux and the density difference between the plume and the interior. A buoyancy source is characterized by its initial buoyancy flux (F₀), defined as follows:

where Q is the volume flux, g is the acceleration due to gravity, ρ is the plume density, ρ_e is the environmental density, and ρ₀ is a reference density (a constant).

In the case of multiple plumes it is the one with the greater initial buoyancy flux not the one with the greater initial density that penetrates to the bottom and controls the stratification. The buoyancy fluxes of plumes with lesser F₀ vanish at intermediate levels, causing them to cease sinking and spread out horizontally. The model was extended to include sloping side walls and wall drag; these additions modify the details of plume termination but do not affect the qualitative results of the model.

Experiments with a two-plume model revealed some interesting behavior when the same total buoyancy flux was distributed differently between the two plumes. When the initial source strengths were between 10 and 90 percent apart, the pycnocline depth and depth of termination of the weaker plume were nearly the same. If the initial source strengths were nearly the same, a small perturbation could result in a dramatic change in the stratification and characteristics of the bottom water. We suggest below that this type of behavior may be reflected in the isotope record. Furthermore, it may

well be that this type of flip-flop behavior has ramifications for climate stability (see last section).

Evidence from the measurement of oxygen isotopes in the tests of benthic foraminifera indicates that in the past deep water has been as warm as 15°C (see next section). Peterson’s (1979; preceding section) model provides a conceptual framework for evaluating the relative potential of various marginal seas for the production of deep water. We suggest that at some time in the past, the strongest initial buoyancy flux, and therefore the bottom water, originated from a marginal basin subject to high evaporation rates and where the density deficit was caused by high salinity rather than low temperature. It is as if the present-day Mediterranean Sea became much larger or the evaporation rate increased such that its buoyancy flux became greater than that from any of the polar sources. According to our preliminary estimates, the present-day polar deep-water sources have about four times the initial buoyancy flux of the present-day Mediterranean. It is interesting to note that within the Mediterranean itself, the saltiest summer water is not sufficiently dense to sink; but winter cooling increases the density still more, and the resulting deep water has a temperature near 13°C, 10° cooler than typical summer surface temperatures (Lacombe and Tchernia, 1972).

Weyl (1968) noted that cold deep-water formation would be inhibited by a slight lowering of surface salinity. Freshening of the surface layer in the regions of cold deep-water formation would cause bottom-water temperatures to increase by drastically reducing the buoyancy flux from polar sources. Weyl (1968) suggested perturbations in the water-vapor transport between the Atlantic and Pacific Oceans as a mechanism for lowering the surface salinity. It has been shown by Lazier (1973) that in the middle 1960’s a decrease in the salinity of the outflow from the Arctic Ocean caused a freshening of the surface layer of the Labrador Sea, which in turn caused about a five year hiatus in the usual wintertime deep convection.

Another mechanism for freshening the surface layers may be removal of salt in basins where evaporites are being precipitated.

Measurement of oxygen isotopes on the tests of benthic foraminifera from deep sea cores (see Chapter 18) indicates that the temperature of bottom water has decreased from approximately 15°C at the end of the Cretaceous to 3°C at present (see Figure 7.1). Superimposed on the generally decreasing isotopic temperature trend is a series of step changes. The abruptness of these changes may be due to hiatuses in sedimentation, or they may reflect a series of shorter time-scale transitions in bottom-water temperature. If these stepwise changes are eventually confirmed, Peterson’s competing plume model suggests an explanation. Deep water is produced by multiple sources, both warm and cold. The temperature and other characteristics of the bottom water are determined by the deep-water source area that produces the largest buoyancy flux. These source strengths have varied through time as a consequence of changes in areas, location, and configurations of marginal seas owing to plate motions and eustatic sea-level changes. In Cretaceous times, warm water sources were dominant over cold-water sources, whereas today cold bottom water is produced at high latitudes. This change required at least one transition from a regime dominated by WSBW to one dominated by cold polar bottom water. Because the plume termination depths are sensitive to small differences in source strengths, several transitions from one regime to the other might be expected as a result of variations in buoyancy fluxes from competing deep water sources during the time when the source strengths were nearly the same.

A simple box model of an evaporative basin shows that the buoyancy flux depends only on the evaporative flux (i.e., area of basin times evaporation rate) (Peterson et al., 1981; Brass et al., 1982) until the onset of halite precipitation, when the buoyancy flux begins to decrease because of salt removal. An evaporative basin insufficiently concentrated to deposit salt is the most effective source of WSBW.

Epicontinental seas as sources of WSBW provide the link by which the solid Earth forces oceanic and atmospheric circulation. Eustatic sea-level fluctuations can be caused by variations in the global seafloor spreading rate (Hays and Pitman, 1973). The area of continent flooded by a given increase in sea level is a function of the global hypsography at that time. The size, and hence drainage efficiency, of the continents directly controls the shape of the hypsographic curve (Harrison et al., 1981; Hay et al., 1981). Thus, times of increased seafloor spreading and continental breakup generate large areas of epicontinental seas both by raising sea level and by lowering the elevation of the continents.

Epicontinental seas producing WSBW must be located within the zone of net evaporation (10–40° N and S) associated with the descending branches of the atmospheric Hadley cell circulation. The distribution of Mesozoic and Cenozoic evaporite deposits strongly suggests that this zone has remained stationary during the last 120 million years (m.y.). Figure 7.2 shows the areas of flooded continents and marginal seas in 10° latitude intervals over the last 100 m.y., as measured from the paleogeographic maps of Barron et al. (1981). The most dramatic change during this time interval has been the decrease in area of shallow seas in the high evaporation belt (10–40°).

Because area is one of the important factors determining the buoyancy flux from evaporative basins, the decrease in the area of evaporative marginal seas over the last 100 m.y. strongly suggests that the rate of production of WSBW has declined over the same time. The shape of the curve of marginal

FIGURE 7.1 Oxygen isotope paleotemperatures from Savin (1977), copyright Annual Reviews, Inc.

sea area versus time is similar in shape to the oxygen isotopic temperature record curve from benthic foraminifera. The decrease in area of evaporative marginal seas and the decrease in the temperature of bottom water during the Cenozoic suggests a transition in the mode of deep-water formation in which cooling at high latitudes has played an increasingly important role,

Changes in the latitudinal distribution of epicontinental seas due to the motions of the lithospheric plates also occur. This movement may transport flooded regions into or out of the net evaporation belt. These movements appear to have had little effect on the areas producing WSBW in the Mesozoic and Cenozoic but may have been more important in earlier times.

Our hypothesis, that plate motions and sea-level changes provide the mechanisms responsible for variations in the production of deep water, has many consequences for the chemistry of the ocean and atmosphere. The chemical state of the ocean at times when WSBW was dominant would have been very different from that which exists at present, and evidence of these differences should be present in the chemistry and isotopic composition of marine sediments. The solubilities of many gases are strongly dependent on temperature and salinity, and

the increased temperature of WSBW at its source (about 15°C versus −2°C for present-day cold bottom water), where it is in contact with the atmosphere, reduces the concentration of dissolved gases such as molecular oxygen and carbon dioxide in water subsequently transported to the deep sea. As a consequence, WSBW variations should leave a signature in the accumulation rates of organic carbon and carbonate in the deep sea,

The distribution of both oxygen and carbon dioxide within the ocean is the result of the formation and destruction of organic matter, exchange with the atmosphere, and physical transport processes. As a result of respiration by animals and bacteria, oxygen is depleted in most subsurface ocean waters. The longer a water mass is isolated from the atmosphere the lower its oxygen content becomes. This fact has enabled oceanographers to use O₂ as a circulation tracer.

Anoxic conditions develop in deep water when the consumption of oxygen by organisms exceeds the rate of oxygen supply. When anoxic conditions exist, organic carbon accumulates at a higher rate. Deep-sea anoxic events have been observed in the geologic record by many authors (e.g., Degens and Stoffers, 1974; Thierstein and Berger, 1978). Many mechanisms have been proposed to explain these events, including stagnation resulting from a layer of fresh or brackish surface water or expansion of the oxygen-minimum layer by inhibited oceanic circulation. These events may also be explained within the context of our model as resulting from a decrease in ventilation of

FIGURE 7.2 Area of epicontinental seas versus age, by latitude band (from the maps of Barron, 1980).

the deep ocean because of reduced oxygen solubility in the source regions without requiring any change in circulation rate. The oxygen content of WSBW may have been similar to the oxygen content of the Mediterranean deep water, which south of France and in the Adriatic contains about 220 µmoles per kilogram of O₂ (Miller et al., 1970). In contrast, North Atlantic deep water near its source contains 310 µmoles of O₂ per kilogram (Bainbridge, 1980). The Mediterranean escapes anoxic conditions owing to the short residence time of its deep waters (100 yr, Lacombe and Tchernia, 1972) and its low nutrient content (Miller et al., 1970), which restricts biological productivity.

The CO₂ system is of interest because changes in atmospheric CO₂ may affect climate by altering the Earth's thermal balance. Dissolved inorganic carbon is present in seawater in the form of bicarbonate and carbonate ions in addition to the dissolved gas. The ocean is a much larger reservoir of CO₂ than is the atmosphere and ultimately determines the atmospheric CO₂ content.

Significant variations in rates of accumulation of carbonates and organic carbon in the ocean basins suggest corresponding variations in atmospheric CO₂ in the past. The decrease in solubility of CO₂ with increasing temperature suggests that WSBW production leads to an increase in atmospheric CO₂. Thus, at times in the geologic past when WSBW production was large, atmospheric CO₂ levels may have been larger than at present and may have played a major role in forcing climate.

The terrestrial record shows that the late Cretaceous climate was more equable (i.e., had a lesser meridional gradient) with much milder conditions at high latitudes than at present, as evidenced by the occurrence of tropical and temperate faunas and floras in higher latitudes than at present (Barron, 1980). We believe that the change to the present climate may be a consequence of the decrease in WSBW production, which is, in turn, a consequence of tectonic and eustatic activity. Two mechanisms may explain the effects of WSBW production on climate: (1) modification of heat transport via both the atmosphere and ocean and (2) changes in the transparency of the atmosphere to incoming solar and outgoing infrared radiation due to increases in the atmospheric water vapor and CO₂.

Incoming solar radiation is more intense in low than in high latitudes. The outgoing reradiation is also more intense in low latitudes; however, the difference is much less, and there is a net heating in low latitudes. This heating imbalance requires a poleward transport of heat across latitude to maintain the entire Earth's mean temperature distribution in steady state. This transport is accomplished by sensible heat transport in the atmosphere and ocean and by latent heat transport in the form of water vapor in the atmosphere. The poleward transport of sensible heat in the atmosphere for the same circulation intensity would be curtailed by an equable climate because of the reduced meridional temperature gradient. Lorenz (1967) noted that in the present-day atmosphere most of the mid-latitude

poleward transport of both heat and moisture is accomplished by eddy transports, These eddies are the mid-latitude storm systems that are driven primarily by baroclinic instability, which represents the conversion of available potential energy into kinetic energy. There would be less available potential energy in an equable climate, and, therefore, this mechanism should be less effective. Manabe and Wetherald (1980) suggested that large-scale monsoonal circulations might accomplish the mid-latitude transport of moisture. Our model requires an increase in net evaporation from low latitude marginal seas to produce WSBW and implies a freshening of surface water at high latitudes to inhibit the sinking of cold water. The link between these two phenomena is higher poleward transport of water vapor in the atmosphere with its equivalent latent heat. Manabe and Wetherald studied the climatic structure in a model atmosphere forced by an increase in CO₂ content and found a striking increase in latent heat transport One of the causes for the reduced thermal gradient was the poleward retreat of highly reflective snow cover. “Another important reason for the general reduction of the meridional temperature gradient is the large increase in the poleward transport of latent heat due to the general warming of the model atmosphere” (Manabe and Wetherald, 1980, p. 102). They concluded that the increased poleward latent heat transport may account in part for the equable Mesozoic climate (Manabe and Wetherald, 1980, p. 117). We suggest that the forcing mechanism for an increase in CO₂ abundance may be changes in the area of marginal seas in the high net evaporation regions and the production of WSBW.

It is difficult to assess the importance of oceanic sensible heat transport when the ocean is filled with WSBW. The presentday oceanic heat transport is not well known and is usually measured indirectly as a residual required to complete a heat balance (e.g., Vonder Haar and Oort, 1973). Bryan (1962) suggested that the vertical thermohaline circulation is more important in the oceanic heat budget than is the wind-driven horizontal circulation. Thus, WSBW formed in low latitudes might well accomplish some of the required poleward heat transport.

We have emphasized steady-state situations based on the notion that climate tends to reside in one of many possible quasisteady states until it is perturbed into another quasi-steady state by some event. It is Important to understand the steady-state WSBW system before studying the transitions between WSBW and cold bottom water. However, it is clear that the transitions would be accompanied by substantial perturbations in the heat budgets and would have dramatic consequences for oceanic and atmospheric chemistry. An understanding of how these transitions occur will be a valuable contribution to understanding how climate evolves.

From these discussions it is evident that the ocean of the geologic past may have had substantially different thermohaline circulation and bottom water characteristics. It has been suggested (e.g., Weyl, 1968) that the modern ocean provides a stabilizing influence on climate because of the difference in characteristic time scales of the ocean and the atmosphere (about 1000 yr versus a few months). It is also reasonable to suggest that a substantial change in the characteristics of the ocean should have substantial effect on the climate.

This chapter is a contribution of the Rosenstiel School of Marine and Atmospheric Sciences. This research was supported by grants from the National Science Foundation, the Office of Naval Research, the National Oceanic and Atmospheric Administration, and the Petroleum Research Fund of the American Chemical Society.

Bainbridge, A.E. (1980). GEOSECS Atlantic Expedition, Vol. 2, Sections and Profiles, U.S. Government Printing Office, Washington, D.C., 198 pp.

Baines, W.D., and J.S.Turner (1969). Turbulent buoyant convection from a source in a confined region, J. Fluid Mech. 37, 51–80.

Barron, E.J. (1980). Paleogeography and climate, 180 million years to the present, Ph.D. Dissertation, U. of Miami, Miami, Fla., 270 pp.

Barron, E.J., C.G.A.Harrison, J.L.Sloan, and W.W.Hay (1981). Paleogeography, 180 million years ago to the present, Eclogae Geol. Helv. 74, 443.

Bolin, B., and H.Stommel (1961). On the abyssal circulation of the world ocean-IV, Deep Sea Res. 8, 95–110.

Brass, G.W., J.R.Southam, and W.H.Peterson (1982). Warm saline bottom water in the ancient ocean, Nature 296, 620–623.

Broecker, W.S., R.Gerard, M.Ewing, and B.C.Heezen (1960). Natural radiocarbon in the Atlantic Ocean, J. Geophys. Res. 65, 2903–2931.

Bryan, K. (1962). Measurements of meridional heat transport by ocean circulation, J. Geophys. Res. 67, 3404–3414.

Chamberlin, T.C. (1906). On a possible reversal of deep-sea circulation and its influence on geologic climates, J. Geol. 14, 363–373.

Craig, H. (1969). Abyssal carbon and radiocarbon in the Pacific, J. Geophys. Res. 74, 5491–5506.

Degens, E.T., and P.Stoffers (1974). Stratified waters as a key to the past, Nature 263, 22–26.

Gordon, A.R. (1975). General ocean circulation, in Numerical Models of Ocean Circulation, proceedings of a symposium held at Durham, N.H., October 17–20, 1972. National Academy of Sciences, Washington, D.C.

Harrison, C.G.A., G.W.Brass, E.Saltzman, J.Sloan III, J. Southam, and J.M.Whiteman (1981). Sea level variations, global sedimentation rates and the hypsographic curve, Earth Planet. Sci. Lett. 54, 1–16.

Hay, W.W., E.J.Barron, J.L.Sloan II, and J.R.Southam (1981). Continental drift and the global pattern of sedimentation, Geol. Rundsch. 70, 302–315.

Hays, J.D., and W.C.Pitman (1973). Lithospheric plate motions, sea level changes and climatic and ecological consequences, Nature 246, 18–22.

Lacombe, H., and P.Tchernia (1972). Caractères hydrologiques et circulation des eaux en Méditerranée, in The Mediterranean Sea, D.J.Stanley, ed., Dowden, Hutchinson and Ross, Stroudsburg, Pa., pp. 25–36.

Lazier, J.R.N. (1973). The renewal of Labrador Sea water, Deep Sea Res. 20, 341–353.

Lorenz, E.N. (1967). The Nature and Theory of the General Circulation of the Atmosphere, WMO No. 218, T.P. 115, World Meteorological Organization, Geneva.

Manabe, S., and R.T.Wetherald (1980). On the distribution of climate changes resulting from an increase in CO₂ content of the atmosphere, J. Atmos. Sci. 37, 99–118.

Miller, A.R., P.Tchernia, and H.Charnock (1970). Mediterranean Sea Atlas, Woods Hole Oceanographic Inst., Woods Hole, Mass., 190 pp.

Peterson, W.H. (1979). A Steady Thermohaline Convection Model, Ph.D. Dissertation, U. of Miami, Fla., 160 pp.

Peterson, W.H., G.W.Brass, and J.R.Southam (1981). The formation of warm saline bottom water in ancient oceans, Ocean Modeling 38, 1–7.

Rossby, H.T. (1965). On thermal convection driven by nonuniform heating from below: An experimental study, Deep Sea Res. 22, 853–873.

Savin, S.M. (1977). The history of the Earth’s surface temperature during the past 100 million years, Ann. Rev. Earth Planet. Sci. 5, 319–355.

Thierstein, H.R., and W.H.Berger (1978). Injection events in ocean history, Nature 276, 461–466.

Vonder Haar, T.H., and A.H.Oort (1973). New estimate of annual poleward energy transport by northern hemisphere oceans, J. Phys. Oceanogr. 3, 169–172.

Weyl, P.K. (1968). The role of the oceans in climate change: A theory of the ice ages, Meteorol. Monogr. 8, 37–62.

Wüst, G. (1961). On the vertical circulation of the Mediterranean Sea, J. Geophys. Res. 66, 3261–3271.