Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
11 INTRODUCTION Sea Level and Climate Change ERIC J. BARRON Pennsylvania State University STARLEY L. THOMPSON National Center for Atmospheric Research It has long been suggested that a change in global sea level produces a change in climate. This hypothesis re- ceives substantial support from the geologic record of sea- level change. Over a broad range of time scales (104 to 107 yr), eustatic sea level and continental flooding are well correlated with paleoclimatic data, particularly during the past 100 million yr (m.y.~. The relationship between sea level and climate has three components. First, climate directly influences sea-level variations largely through the processes that control the growth and decay of continental ice. At peak glaciation 18,000 yr ago, approximately one-sixth of the planet was covered by ice and continental glaciation, and sea level was lower by 85 to 130 m than it is at present (CLIMAP, 1976). Second, a variety of physical mechanisms exist whereby sea-level changes can directly affect global climate. By changing the nature of the atmosphere-surface interface, sea-level changes can alter the transfer of heat, moisture and momentum between the surface and the atmosphere. In addition, sea level can alter ocean currents by introduc- ing or removing geographic barriers. Ocean currents are a principal means of transferring heat from the tropics to polar regions and play a crucial role in controlling some regional climates of the present. Other potential direct influences of sea level on climate include effects on ice ~5 sheet size, ocean chemistry, and atmospheric content of carbon dioxide and other trace gases. Third, the correlation between sea level and climate during Earth history may be the product of an indirect association. One plausible indirect link is a relationship between increased volcanism, higher atmospheric carbon dioxide levels, and high global sea level caused by rapid sea-floor spreading (Berner et al., 19831. In this case, climate change would be directly related to changes in carbon dioxide levels, and the correlation of climate change with sea level would not be indicative of a direct cause- and-effect relationship. Only the direct influence of sea level on climate and the indirect association between sea level and climate are considered here. The direct influences of climate on sea level, through land ice volume or ocean temperature change, are considered in Chapters 10 and 13 (this volume), re- spectively. Simulations with global dynamical climate models and examples from the paleoclimatic record pro- vide insight into the importance of various physical mecha- nisms and the explanations of correlation between climate and sea-level variations during earth history. DIRECT EFFECTS OF SEA LEVEL ON CLIMATE The potential direct effects of sea level on climate can arise from physical mechanisms that fall into at least five categories: (1) regional changes in atmosphere-surface
186 coupling, (2) changes in ocean circulation, (3) changes in surface heat capacity, (4) changes at ice sheet-ocean mar- gins, and (5) changes in ocean-atmosphere chemical composition. The mechanisms are listed roughly in order of increasing uncertainty, although insufficient work has been completed in the investigation of the climatic effect of sea-level changes in every category. Hence the discus- sion that follows will necessarily rely heavily on work not originally conducted with an eye to sea-level change. In some cases, educated speculation will have to suffice until more research specific to the problem of sea level and climatic change is conducted. More is known about atmosphere-surface coupling than the other four categories, perhaps because atmospheric and oceanographic scientists have studied and modeled the general problem for many years. Ocean circulation modeling has also been conducted for many years, but the level of detail required for comprehensive study of the global effects of sea-level change has exceeded the capa- bility of even the largest computers. ECurrent global ocean models use about a 500-km grid. Typical sea-level changes (over geologic time) produce coast line changes on the order of 100 to 1000 km. Thus sea-level changes are either subgrid scale or, at best, barely resolvable (hence poorly treated) by current global ocean models.] Changes in ice sheet volume are usually thought of as influencing sea level; interest in the influence of sea level on ice sheets is relatively recent. Likewise, it is uncertain how sea-level changes can have a direct impact on ocean-atmosphere chemical composition, but such composition changes, if they exist, could cause substantial changes in global cli- mate by altering the "greenhouse" effect. Atmosphere-Surface Coupling Virtually every facet of Earth's climate is controlled or affected by the transfer of heat, moisture, and momentum between the atmosphere and the underlying surface. For example, most of the heat that drives the atmospheric circulation is first absorbed as solar energy at the surface and then made available as sensible heat (direct tempera- ture change) and latent heat of condensation (heat released when evaporated moisture condenses to form rain or snow). Furthermore, the long-term circulation of the atmosphere is constrained to conserve absolute angular momentum. Momentum transferred to the Earth when the atmosphere "rubs" against the surface must be balanced by a reverse momentum transfer elsewhere arising from changes in wind speed and direction. Surface easterly winds repre- sent absolute momentum transfer from the Earth to the atmosphere, and surface westerlies vice versa. Changes in the regional distribution of surface "roughness" can di ERIC J. BARRON AND STARLEY L. THOMPSON rectly affect local surface winds and, owing to the momen- tum conservation constraint, global wind patterns as well. Surface Heat and Moisture Balance Heat balances typical of three surface types are shown in Figure 1 1.1. In general, the surface of the Earth receives a net positive flux of radiant energy (a gain of solar energy minus a loss of infrared energy) that is balanced by sensible heat and evaporative losses. Land in the subtropics is relatively dry and has a relatively high albedo, that is, reflectivity for solar energy. Most of the net radiative gain is balanced by sensible heat loss owing to the lack of moisture to be evaporated. In contrast, the adjacent subtropical oceans have low albedo and lose heat almost exclusively by evapo- ration. Changing subtropical land to ocean, e.g., by a sea- level increase, would produce a major change in the amount of heat and moisture received by the overlying atmosphere, perhaps producing a change in regional climate. As Fig- ure 1 1.1 shows, near equatorial land is qualitatively simi- lar to ocean because abundant vegetation makes it rela- tively dark and moist. Thus changing equatorial-type land to ocean would produce a smaller climatic effect than changing subtropical land to ocean. A local change in albedo of the surface can produce a large effect on the local surface energy balance, but what effect can it have on global-scale climate? A simple analy R Net Solar and I R Radiation ( Wm~2 ) E - Evaporation / Transpiration ( Wm~2) S Sensible Heat ( Wm~2 ) a _ Surface Al bedo 50 -15 -35 Subtropical Land R: Eji S 85 - 80 -5 Subtropical Ocean a=007 it L i__ ~ ~ _ Near 60 -45 - 1 5 Equatoria I Land Rll ED So| FIGURE 11.1 Heat balances for some representative surface types. On average, net radiative heating of the surface is bal- anced by evaporative and sensible heat loss. Evaporation is the preferred mode of heat loss at warm temperature; thus subtropi- cal oceans lose heat mainly through evaporation.
SEA LEVEL AND CLIMATE CHANGE sis based on our current understanding of global climate theory suggests that a global change in surface albedo of 0.01 will produce about a 1°C change in surface tempera- ture. The effect of less than global-scale changes in sur- face albedo will be smaller in general, but requires de- tailed modeling calculations to determine the actual sensi- tivity. We can now make a rough estimate of the effect of sea- level change on surface albedo and temperature. Suppose that in a latitude zone with an initial land fraction of 0.4, a sea-level rise reduces the land area fraction to 0.3 (a rather large change, probably requiring a sea-level increase well in excess of 100 m). Assume also that the difference in surface albedos between the covered land and the ocean is 0.1. The average change in surface albedo in the zone is thus (0.4 - 0.3) x 0.1 = 0.01. The predicted rise in surface temperature of the zone (assuming no interaction with adjacent zones or other effects) is about 1°C. Thus a major change in sea level is not likely to cause a large change in global-scale surface temperature from an albedo effect only. This approximate result is confirmed by more detailed calculations by Thompson and Barron (1981) that examined the effect on albedo of shifting the continents to the positions they had 100 m.y. ago (Ma). Factor of 2 changes in land area in the subtropics produced a change of only a few watts per square meter (W/m2) in the surface energy budget of the zone. Referring again to Figure 11.1, we see that plausible changes in surface wetness can produce much larger changes in the surface energy balance than can plausible changes in albedo. Sea-level change provides an ideal way to produce an extreme change in surface wetness. No calculations have been done with global climate models in which sea level has been changed and only surface wet- ness has been allowed to vary (i.e., no albedo changes or other effects). Thus it is difficult to say just how important changes in wetness are compared to albedo changes, but we can get some indication from simulation studies that have changed land surface properties to investigate other problems. Shukla and Mintz (1982) performed two simu- lations with a global climate model: one in which all land was assumed to be completely dry (no evaporation al- lowed) and the other assuming land to be completely wet (like the ocean). Thus the simulations can give an indica- tion of what might happen if very dry desert were changed to ocean without any albedo change. The calculations showed the dry land case producing summer land surface temperatures 10° to 20°C higher than the wet case. Clearly, evaporation from the surface can have a powerful cooling effect. Moisture flux from the surface not only transports heat but causes changes in atmospheric water vapor content that can affect cloudiness and precipitation. Yeh et al. ~7 (19841 have done simulations with a global climate model in which soil moisture was perturbed to see the effect on surface temperature and rainfall. They found, in general, that increased surface evaporation led to increased precipi- tation, but not necessarily in the same region as the in- creased evaporation. Increasing evaporation in a surface divergence region (a region of downward air motion) led to increased precipitation in adjacent convergence regions, but not locally. The explanation of the effect is that it is difficult to form precipitation in regions of downward motion regardless of the amount of water vapor in the air. Excess humidity in subsidence regions is moved by winds to regions of upward motion and then rained out. More recently, Gordon and Hunt (1987) have further empha- sized the role of simulated surface hydrology and precipi- tation patterns. From the above discussion, it can be concluded, for example, that replacing land with ocean in the normally dry subtropics might not lead to any local rainfall increase, but might instead produce increased equatorial rainfall. Aside from noting these potential local and nonlocal mechanisms, more specific statements cannot be made about the effects of sea-level change on rainfall without more-detailed global climate model calculations. What happens if surface albedo and moisture availabil- ity are changed simultaneously? Henderson-Sellers and Gornitz (1984) performed climate model experiments in which the effects of Amazon deforestation were exam- ined. The deforestation was simulated by changing the surface albedo from 0.1 1 to 0.19 and by greatly decreasing the potential for surface evaporation and transpiration. These changes are similar to what one would expect from a sea-level drop, i.e., replacing sea with land. The re- searchers found that there was no net surface temperature change in the deforested area. The surface albedo and wetness changes for this particular case produced nearly complete compensating effects. There was, however, a local decrease in rainfall of about 10 percent over the deforested area, as expected from the discussion above. There were no global-scale climate effects that could be reliably detected above the normal model variability. Even though these results are only suggestive, they do indicate that the direct climate effect of large sea-level changes may not be large even on regional scales. \ .~ ~ ~ ~ J .~ .^ ~ ~ ~ ~ _ Surface Roughness The transfer of heat, moisture, and momentum between the atmosphere and surface is propor- tional to the surface drag coefficient (CD), a nondimen- sional quantity that expresses the proportionality between the friction force per unit area at the surface and the square of the surface wind speed. The value of CD is strongly dependent on the "roughness" of the surface as measured by the characteristic height of surface protrusions such as
188 TABLE 1 1.1 Typical Values of Surface Drag Coefficient, CD (X10-3) (from Garratt 1977) D Ocean Wind speed 5 m/s 1 Wind speed 20 m/s 2 Land Desert 3 Tropical forest 25 Typical 1 0 rocks or plants. Typical values of CD for land and ocean are given in Table 11.1. In general, the drag coefficient over ocean is an order of magnitude smaller than it is over land. This has the effect of (a) tending to produce de- creased surface fluxes over ocean and (b) producing a higher surface wind speed over ocean by way of compen- sation. [Substantial momentum transfer between the Earth and the atmosphere is also created by large-scale orogra- phy, the so-called "mountain pressure torque" (see Holton, 1979, pp. 264-265~. It is unlikely that reasonable sea- level changes alone would produce climatologically im- portant changes in topographic heights relative to sea level.] It is not clear what effect a change in sea level could have on large-scale climate through a change in CD. Some sensitivity studies with climate models (e.g., Hansen et al., 1983) have shown that the simulated climates are not very sensitive to CD. On the other hand, it is well known that changes in momentum exchange at the surface can have a substantial effect on atmospheric circulation systems (Horton, 1979, p. 260~. It is not possible at this time to say how important sea-level-induced changes in surface rough- ness are in relation to changes in albedo or wetness. Combined Surface Effects Barron and Washington (1984' performed two climate model simulations for con- tinental positions of 100 Ma. One case was for present day sea levels and the other included a sea level estimated to be 330 m higher than the present that is thought to have occurred during the Cretaceous period. This is a massive sea-level increase that would have caused a 20 percent decrease in global land area. Barron and Washington found that the sea-level change produced systematic local climate changes only in the subtropics and tropics. In these regions, changing land to ocean produced local cooling (the evaporation increase outweighed the albedo decrease) and slight warming elsewhere. Zonally aver- aged temperatures changed by less than 2°C at any latitude (see Figure 11.2~. The global net effect on surface tem ERIC J. BARRON AND STARLEY L. THOMPSON perature was small. The postulated change in Cretaceous sea level produced a smaller global climate effect than did correspondingly large changes in topography and conti- nental positions. Ocean Circulation In a number of instances, sea-level changes have been invoked as a mechanism of modifying the ocean surface and thermohaline circulations. The sea-level ocean circu- lation mechanism involves bathymetric control of current directions, importance of barriers in basin-basin commu- nication, and the role of shallow epicontinental seas and marginal basins in the formation of deep water. Each example cited below illustrates the potential for sea-level changes to influence climate, but none of the studies to date provides quantitative links between sea-level-induced circulation changes and climate change. The high-velocity core of the Gulf Stream scours the Blake Plateau, producing a 50- to 75-km-wide band of erosional topography. Pinet and Popenoe (1982) used seismic stratigraphy and a series of drill holes to show that the axis of the Gulf Stream has shifted hundreds of kilo- meters repeatedly during the past 20 m.y. The position of the Gulf Stream axis is well correlated with sea-level 310 305 A 300 o 295 290 llJ rL 285 L3J 280 275 270 _ 26s 315 I ~ 1 l- I 1 1 1 1 1 1 1 1 1 1 l l 1 : _ _ ~ - ,/ Low Sea Level ~/ --- High Sea Level ~< it' . ,, 1,, I,, 1 1 1 1, 1 1 1 1 1 an 60 30 0 ~90 LAT ITUDE FIGURE 11.2 A comparison of the zonally averaged surface temperature for a general circulation model experiment with Cretaceous continental positions with low sea level (present-day total area) and no topography and an experiment with Cretaceous continental positions with high sea level (about 20 percent of continental area flooded) and no topography.
SEA LEVEL AND CLIMATE CHANGE 1. it= ~ 3so~ ~ ~ ~ ~ 1 If/ ~ CON N ECTE D BASINS GROWI N G /ICE SHEET 2 ~WATER TRANSFER / ~: ~ ~/ :'~i./'7/< LOWERED / ¢/////////////// SEA LEVEL 3. ~ SEA ICE EVAPORI TES 4;~ LESS SALINE OCEAN FIGURE 11.3 A hypothetical example of the potential for sea- level change to affect climate through a mechanism that relies on isolation of ocean basins. (1) The world ocean and a subtropical sea communicate across a shallow strait. (2) A global climate change induces ice sheets to grow. Lowered sea level isolates the subtropical sea, which begins to evaporate. Fresh water is transferred to the world ocean by precipitation. (3) Lower salin- ity of the world ocean promotes a sea ice increase, which acts as a positive feedback to the global cooling. variations. During lowstands of sea level, a broad bathymet- ric bulge, called the Charleston bump, deflects the Gulf Stream offshore to a position across the Blake Plateau and along the steep continental slope just north of the Blake Spur. During sea-level highstands, the Gulf Stream paral- lels the Florida-Hatteras Slope. The Gulf Stream has a notable impact on regional climates, thus the potential for climate change. However, the extent of possible climatic change due to such movements in current direction in response to sea-level variation is unknown. The elevation of a barrier above sea level or subsidence of a ridge below sea level has been cited as a mechanism of climatic change in numerous instances. The subsidence of the Greenland-Iceland-Faeroe Ridge (about 38 Ma) probably resulted in a significant interaction between the Arctic and Atlantic oceans; subsidence of the Walvis Ridge off South Africa (about 20 Ma) greatly influenced the Bengula current system; and formation of the Isthmus of Panama (about 2.3 Ma) eliminated equatorial Atlantic-Pacific flow (Berggren, 19821. Berggren stressed the role of ocean gateways in modifying heat transport by the oceans. Sea-level variations have the potential to have breached or created surface circulation barriers during Earth history, but few instances related specifically to sea level are documented, and in every case the climatic implica- tions are qualitative. Sea-level change can isolate or reconnect small basins. Particularly in subtropical high-evaporation zones or re- gions where precipitation greatly exceeds evaporation, isolation of a basin from the main ocean can result in large density contrasts (i.e., hypersalinity or freshening of the basin). Thierstein and Berger (1978) hypothesized that reconnection of a temporarily isolated basin can result in an injection event that favors either abyssal stratification or surface stratification depending on the salinity charac- teristics. Brass et al. (1982) showed that deep water can form in marginal seas in the subtropics due to high evapo- ration (warm, salty bottom water). These authors suggest that changes in the size and configuration of marginal seas in the subtropics owing to sea-level change may have controlled deep-water formation during different periods of earth history. Deep water tends to form in semirestricted basins or marginal seas because the isolation allows the water mass to obtain different density characteristics, through atmospheric interaction, in comparison with the main ocean. These marginal seas are certainly susceptible to sea-level variations that may modify their size and configuration. Basin isolation and induced salinity vari- ations may also result in climate change because fresh water freezes more readily (see hypothetical example in Figure 11.3~. The potential of this mechanism to modify climate is largely unexplored. Annual Temperature Cycle The heat capacity or thermal inertia of the surface is a strong determinant of the annual cycle of temperature at a given latitude. The thermal inertia of the ocean is depend- ent on the depth of the mixed layer, ranging from tens of meters to in excess of 200 m. In contrast, the thermal inertia of land is roughly equivalent to a 1.0-m depth of water. Extensive continental flooding is likely to modify the local thermal inertia, and hence influence the annual cycle of temperature. In experiments using the seasonal zonal energy balance climate model of Thompson and Schneider (1979) to in- vestigate the contrast between the climate of the Creta- ceous and that of the present (Barron et al., 1981), the changes in land fraction associated with Cretaceous geog- raphy resulted in a 3° to 5°C reduction in the amplitude of the annual cycle of surface temperature in the Northern Hemisphere mid-latitudes (i.e., the summers were cooler and winters warmer than present). The mid-latitude ocean fraction increased by approximately 20 percent in the Cretaceous case. In this model the zonal thermal inertia
190 was computed from an area-weighted harmonic mean assuming no land-sea heat transfer within a zone. More recently, the general circulation model experi- ments with Cretaceous geography and mean annual isola- tion of Barron and Washington (1984) have been extended to a seasonal mixed-layer experiment. The zonally aver- aged amplitude of the annual cycle of surface temperature in the Northern Hemisphere mid-latitudes was reduced 4° to 8°C in comparison with a present-day control experi- ment. The largest component of this zonally averaged difference is regions that were land initially but become oceanic when the sea level was increased. If this compo- nent of the zonal average is removed, the change in ampli- tude is 1° to 3°C. The above two simulations illustrate the potential for sea-level fluctuations to produce changes in seasonality. However, each experiment described above includes a number of variables that could influence seasonality (e.g., ice-albedo feedback), and consequently the importance of thermal inertia has not been isolated. Ice Sheets The influence of sea level on ice sheets became a topic of interest when the potential instability of the present West Antarctic Ice Sheet became widely known. This ice sheet is the smaller of two major ice masses covering Antarctica. Unlike the East Antarctic Ice Sheet, the west- ern ice sheet is mostly grounded below where sea level would be if the ice were removed. Apparently, the ice sheet exists only because floating ice shelves at its edge act as buttresses to prevent the ice sheet from quickly flowing out into the ocean. These ice shelves are them- selves stabilized by friction against protruding islands, underwater rises, and the sides of the bays in which the shelves sit. Such marine ice sheet-shelf systems have been shown to be potentially unstable to perturbations such as ocean warming or sea-level rise (Thomas et al., 19791. Evidence and modeling studies (Thomas and Bentley, 1978) indicated that the West Antarctic Ice Sheet was much larger 18,000 yr ago during the height of the last glacial episode. This increased size was made possible by the almost 100-m lower sea level at that time. The hy- pothesis is that the rise in sea level associated with the melting of the great Northern Hemisphere ice sheets caused a substantial collapse of the large West Antarctic Ice Sheet until it stabilized at its present size. A further sea-level rise could presumably act to unpin the buttressing ice shelves that allow the remnant ice sheet to exist. Although the massive East Antarctic Ice Sheet is in no danger of total collapse, it too has undergone fluctuations in size associated with sea-level changes. During the last glacial maximum this ice sheet probably extended 75 to 90 ERIC J. BARRON AND STARLEY L. THOMPSON km farther onto the continental shelf than at present (Alley and Whillans, 1984~. The rise in sea level at the end of the Ice Age produced a rapid retreat response at the ice sheet edge that propagated as a wavelike reduction of ice sheet thickness to the interior of the ice sheet. Although the basic link between large sea-level changes and ice sheet growth and decay is fairly clear, it is not obvious that small sea-level changes in isolation have a substantial effect on global climate. The principal effect on climate would probably arise from the change in area of the ice sheet and associated sea ice. A further regional climatic effect would come from changes in ice sheet topography. If large glacio-climatic changes in sea level are needed to have significant effects on ice sheets, then any sea-level-induced ice sheet changes would act only as positive climatic feedbacks rather than actual driving forces for large climatic changes. The level of positive feedback could be quite small, e.g., a 50-cm rise in sea level due to ocean thermal expansion could cause such a small reduc- tion in ice sheet size that any positive feedback on ocean temperature would be negligible. Ocean Chemistry The most uncertain potential direct effect of sea level on climate involves the conjectured relation between sea level, ocean-atmosphere chemistry, and the greenhouse effect. Studies of cores drilled in ice sheets have shown that atmospheric CO2 concentration was about two thirds of its present value during the last glacial maximum. Since CO2 is a greenhouse gas, the change in atmospheric com- position apparently acted to substantially augment the postglacial warming (Thompson and Schneider, 1981). It is not clear what caused the atmospheric CO2 in- crease at the time of deglaciation, but it is known that the changes must have originated in the oceans (Broecker, 1984~. An initial hypothesis (Broecker, 1982) was that increased sea level due to melting Northern Hemisphere ice sheets covered previously exposed continental shelf. Increased sedimentation of organic matter onto the shelf would have removed phosphorous from the ocean. Since phosphorous is a limiting nutrient, such a loss would de- crease ocean phytoplankton productivity, reduce the up- take of CO2 in surface water, and hence produce an atmo- spheric CO2 increase. A second hypothesis (Berger, 1982) was that the postglacial sea-level rise encouraged deposi- tion of carbonate sediments on the shelf, an action that would have generated an increased CO2 content in the surface water and the atmosphere. The source of carbon- ate to be deposited, and hence CO2, would have been dissolution of deep-sea sediments. However, more recent hypotheses of the glacial to interglacial CO2 change have not directly involved sea-level changes (Broecker, 1984~.
SEA LEVEL AND CLIMATE CHANGE Whether sea-level changes can directly affect atmospheric CO2 concentration is an unresolved question. INDIRECT SEA-LEVEL AND CLIMATE ASSOCIATIONS The global cooling trend over the past 70 m.y. is well correlated with a gradual decline in global sea level or increase in global land area. The global cooling from the Cretaceous to the present day has been estimated to be in the range of 6° to 12°C in globally averaged surface tem- perature (Barron, 1983~. The decrease in global sea level is estimated to be as much as 300 to 400 m (Hardenbol et al., 1982) and an increase in total land area of 20 percent (Barron et al., 1980~. Early comparisons between sea- level variations associated with this long-term trend and paleotemperature (e.g., Damon, 1968) have been used to infer a strong causal relationship between sea level and climate over the past 100 m.y. Barron and Washington (1984) performed an extreme sea-level sensitivity experiment with a general circulation model of the atmosphere. This experiment compared mean annual simulations for Cretaceous geography with flooded continents and with present land area (nonflooded conti- nents), which were described earlier. The globally aver- aged surface temperature response to increased sea level and decreased land area in the model was -0.2°C (Figure 11.2~. In the subtropics, where the majority of the flood- ing occurred, increased evaporative cooling (about 60 W/m2) and a small increase in cloud cover (2.5 percent) compensated for the surface albedo change. The results of the sea-level climate model sensitivity experiment bring into question the direct explanation of the global cooling trend as a function of sea-level and surface-albedo variations. The alternative possibilities are (1) inadequate model sensitivity, most likely because of a lack of a seasonal cycle or a fully resolved coupled ocean model, or (2) an indirect association of sea level and paleo- temperatures. Point (1) is discussed in detail by Barron and Washington (1984), and here we will introduce only one plausible indirect association between sea level and paleotemperatures. Berner et al. (1983) and Lasaga et al. (1985) performed calculations with a geochemical model based on the carbonate-silicate geochemical cycle to test the possibility of atmospheric CO2 concentration changes on time scales of 1 m.y. over the past 90 m.y. On this time scale, CO2 removal from the atmosphere is largely dependent on weathering of exposed silicate rocks (with higher sea level this area decreases). One explanation of higher global sea levels is increased rates of seafloor spreading (see Harri- son, Chapter 8, this volume). If seafloor spreading rates are high, then a logical implication is greater volcanism 191 and hence higher CO2 input into the atmosphere (see Arthur et al., 1985, for a discussion on substantially higher vol- canic input estimates for the Cretaceous). Berner et al. (1983) suggested that this substantially higher volcanic input of CO2 cannot be compensated by increased rate of weathering of exposed silicate rocks. In this scenario, high paleotemperatures would result from a CO2-induced warming that would be associated with high sea level, but the warming would not be a direct response to sea-level . . variations. The problems with the atmospheric CO2 model of Berner et al. (1983) include estimating actual rates of CO2 de- gassing through time, estimating actual seafloor spreading rates (Berner et al. used four different estimates of seafloor spreading variations), and determination of how directly weathering of silicate and carbonate rocks responds to the level of atmospheric CO2 concentrations (Berner and Barron, 1984~. In addition, Shackleton (1985) suggested that the carbon isotope record may limit the potential variability of atmospheric CO2. Despite some unresolved problems, the CO2 model of Berner et al. (1983) and Lasaga et al. (1985) would solve the problem of explain- ing the long-term correlation between sea level and paleo- temperatures in light of small climate model sensitivity to large-scale continental flooding, and is a good example of a potential indirect association between sea level and cli- mate. SUMMARY The direct effects of sea level on climate include changes in atmosphere-surface coupling, ocean circulation, ther- mal inertia, ice sheet-ocean interactions, and changes in ocean-atmosphere chemical composition. In only a few cases has the direct effect of a specific variable been iso- lated. In the majority of the climate model simulations, the model experiments provide insight into the importance of various physical mechanisms, but a series of sensitivity experiments should be conducted to isolate the importance of variables such as surface albedo, surface wetness, sur- face roughness, and thermal inertia. Many of the argu- ments presented for the importance of various mecha- nisms (e.g., ocean gateways that influence oceanic heat transport) are purely qualitative. The first steps toward placing these concepts on a more firm physical foundation must be taken if we are to demonstrate the potential role of sea level in climate change. Further, correlations between sea level and climate in the geologic record may not be a product of the direct mechanisms described here. The possibility exists, for example, that the causes of sea-level change may influ- ence climate (e.g., tectonic control on sea level and vol
192 canic CO2 emissions), and the paleoclimatic associations with sea level may be indirect. REFERENCES Alley, R. B., and I. M. Whillans (1984). Response of the East Antarctic Ice Sheet to sea-level rise, J. Geophys. Res. 89, 6487-6493. Arthur, M. A., W. E. Dean, and S. O. Schlanger (1985~. Changes in the global carbon cycle and climate during the mid-Creta- ceous: Their relationship to volcanism and possible changes in atmospheric CO2, in The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present, E. T. Sundquist and W. S. Broecker, eds., Geophys. Monogr. Ser. 32, American Geophysical Union, Washington, D.C. Barron, E. J. (19831. A warm equable Cretaceous: The nature of the problem, Earth Sci. Rev. 19, 305-338. Barron, E. J., and W. M. Washington (1984~. The role of geo- graphic variables in explaining paleoclimates: Results from Cretaceous climate model sensitivity studies, J. Geophys. Res. 89, 1267-1279. Barron, E. J., J. L. Sloan, and C. G. A. Harrison (19801. Potential significance of land-sea distribution and surface albedo vari- ations as a climatic forcing factor; 180 m.y. to the present, Palaeogeogr. Palaeoclim. Palaeoecol. 30, 17~0. Barron, E. J., S. L. Thompson, and S. H. Schneider (1981~. An ice-free Cretaceous? Results from climate model simulations, Science 212, 501-508. Berger, W. H. (19821. Increase of carbon dioxide in the atmos- phere during deglaciations: The coral reef hypothesis, Naturwis- senchaften 69, 87-88. Berggren, W. (1982~. Role of ocean gateways in climatic change, in Climate in Earth History, Studies in Geophysics, National Academy Press, Washington, D.C., pp. 118-125. Berner, R. A., and E. J. Barron (19841. Comments on the BLAG model: Factors affecting atmospheric CO2 and temperature over the past 100 million years, Am. J. Sci. 284, 1183-1192. Berner, R. A., A. C. Lasaga, and R. M. Garrets (19831. The carbonate-silicate geochemical cycle and its effect on atmos- pheric carbon dioxide over the last 100 million years, Am. J. Sci. 283, 641-683. Brass, G., E. Saltzman, J. Sloan, J. Southam, W. Hay, W. Holzer, and W. Peterson (1982~. Ocean circulation, plate tectonics, and climate, in Climate in Earth History, Studies in Geophys- ics, National Academy Press, Washington, D.C., pp. 118-125. Broecker, W. S. (1982~. Glacial to interglacial changes in ocean chemistry, Prog. Oceanogr. 11, 151-197. Broecker, W. S. (1984). Carbon dioxide circulation through the ocean and atmosphere, Nature 308, 602. CLIMAP (19761. The surface of the ice-age Earth, Science 191, 1131-1137. Damon, P. (19681. The relationship between terrestrial factors and climate, Meteorol. Monogr. 8, 106-111. Garratt, J. R. (19771. Review of drag coefficients over oceans and continents, Mon. Weather Rev. 105, 915-929. ERIC J. BARRON AND STARLEY L. THOMPSON Gordon, H. B., and B. G. Hunt (1987). Interannual variability of simulated hydrology in a climate model: Implications for drought, Climate Dynamics l, 113-130. Hansen, J., G. Russell, D. Rind, P. Stone, A. Lacis, S. Lebedeff, R. Ruedy, and L. Travis (1983~. Efficient three-dimensional global models for climate studies: Models I and II, Mon. Weather Rev. 111, 609-662. Hardenbol, J., P. R. Vail, and J. Ferrer (1982). Interpreting paleoenvironments, subsidence history, and sea-level changes of passive margins from seismic and biostratigraphy, in Cl~- mate in Earth History, Studies in Geophysics, National Acad- emy Press, Washington, D.C., pp. 139-153. Henderson-Sellers, A., and V. Gornitz (1984~. Possible climatic impacts of land cover transformations, with particular empha- sis on tropical deforestation, Climatic Change 6, 231-257. Holton, J. R. (1979~. An Introduction to Dynamic Meteorology, 2nd ea., Academic Press, New York, 391 pp. Lasaga, A. C., R. A. Berner, and R. M. Garrels (1985~. An improved geochemical model of atmospheric CO2 fluctuations over the past 100 million years, in The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present, E. T. Sundquist and W. S. Broecker, eds., Geophys. Monogr. Ser. 32, American Geophysical Union, Washington, D.C., pp. 397~1 1. Pinet, P. R., and P. Popenoe (1982~. Blake Plateau: Control of Miocene sedimentation patterns by large-scale shifts of the Gulf Stream axis, Geology 10, 257-259. Shackleton, N. (1985~. The ocean carbon isotope record: Impli- cations for the history of global carbon cycling, in The Carbon Cycle and Atmospheric CO2: Natural Variations Archean to Present, E. T. Sundquist and W. S. Broecker, eds., Geophys. Monogr. Ser. 32, American Geophysical Union, Washington, D.C. Shukla, J., and Y. Mintz (19821. Influence of land-surface evapotranspiration on the Earth's climate, Science 215, 1498-1501. Thierstein, H. R., and W. H. Berger (1978~. Injection events in ocean history, Nature 276, 461-466. Thomas, R. H., and C. R. Bentley (1978~. A model for Holocene retreat of the West Antarctic Ice Sheet, Quat. Res. 10, 150-170. Thomas, R. H., T. J. O. Sanderson, and K. E. Rose (1979~. Effect of climatic warming on the West Antarctic Ice Sheet, Nature 277, 355-358. Thompson, S. L., and E. J. Barron (19811. Comparison of Cre- taceous and present Earth albedos: Implications for the causes of paleoclimates, J. Geol. 89, 143-167. Thompson, S. L., and S. H. Schneider (19791. A seasonal zonal energy balance climate model with an interactive lower layer, J. Geophys. Res. 84, 2401-2414. Thompson, S. L., and S. H. Schneider (1981~. Carbon dioxide and climate: Ice and ocean, Nature 290, 9-10. Yeh, T.-C., R. T. Wetherald, and S. Manabe (1984). The effect of soil moisture on the short-term climate and hydrology change A numerical experiment, Mon. Weather Rev. 112, 474-490.