The seasonal cycle in a coupled experiment with an atmospheric GCM and a two-layer equatorial ocean model

The mean state and the seasonal cycle in the tropical Pacific are studied, using a new coupled tropical ocean-global atmosphere model. The atmospheric component is a general circulation model and the oceanic component is a two and a half layer model of the tropical Pacific. The coupling is based on delocalized physics: the spatial resolution of the physics of the atmospheric component is the same as the spatial resolution of the oceanic model. No flux corrections are applied. A 31 year experiment has been made with the climatological observed sea surface temperature outside the area of coupling. We observe a quick drift of the model which, after three years, reaches a warm mean state. The temperature bias varies geographically between 1?^°C and 2?^°C, but, in spite of this default, the eastern part of the basin remains colder than the west. This contrast is shown to be dependent on the shoaling of the thermocline east of 160^°W. There is a significant seasonal cycle with an amplitude and phase of the seasonal variations which are well reproduced with respect to many other models. It is shown that interactions between the ocean and the atmosphere in the central and eastern Pacific are sufficient to explain the gross features of its evolution. In July, easterlies intensify in the Southern Hemisphere and lead to a strong upwelling and an enhanced evaporation in the eastern part of the basin. This induces a cooling throughout the area. The cooling reaches a first maximum in October in the easternmost part of the basin, then propagates westward along the equator with a decreasing amplitude. In January it is reinforced in the central part of the basin because of a divergence of the current, which is too strong. The mechanisms found here emphasize the role of the upwelling in maintaining the equatorial Pacific climate, and are in agreement with those found in other simplified coupled models.     

Potential decadal predictability and its sensitivity to sea ice albedo parameterization in a global coupled model

Decadal prediction is one focus of the upcoming 5th IPCC Assessment report. To be able to interpret the results and to further improve the decadal predictions it is important to investigate the potential predictability in the participating climate models. This study analyzes the upper limit of climate predictability on decadal time scales and its dependency on sea ice albedo parameterization by performing two perfect ensemble experiments with the global coupled climate model EC-Earth. In the first experiment, the standard albedo formulation of EC-Earth is used, in the second experiment sea ice albedo is reduced. The potential prognostic predictability is analyzed for a set of oceanic and atmospheric parameters. The decadal predictability of the atmospheric circulation is small. The highest potential predictability was found in air temperature at 2?m height over the northern North Atlantic and the southern South Atlantic. Over land, only a few areas are significantly predictable. The predictability for continental size averages of air temperature is relatively good in all northern hemisphere regions. Sea ice thickness is highly predictable along the ice edges in the North Atlantic Arctic Sector. The meridional overturning circulation is highly predictable in both experiments and governs most of the decadal climate predictability in the northern hemisphere. The experiments using reduced sea ice albedo show some important differences like a generally higher predictability of atmospheric variables in the Arctic or higher predictability of air temperature in Europe. Furthermore, decadal variations are substantially smaller in the simulations with reduced ice albedo, which can be explained by reduced sea ice thickness in these simulations.     

Climate model sensitivity to sea ice albedo parameterization

Summary Three one-year experimental simulations with the National Center for Atmospheric Research Community Climate Model (NCAR CCM) were performed with three sea ice albedo parameterizations and compared with control run results to examine their impact on polar surface temperature, planetary albedo and clouds. The first integration utilized sea ice albedos of the Arctic Basin for the spring and summer of 1977 derived from defence Meteorological Satellite Imagery (DMSP). The second simulation employed prescribed lead and melt pond fractions and an albedo weighting scheme. The third simulation involved the coupling of an interactive sea ice/snow albedo parameterization made a function of surface state.Results show that prescribed, and assumed true satellite sea ice albedos produced higher planetary albedos than those calculated with the standard CCM sea ice albedo scheme in the control run. As a result, lower temperatures (up to 0.5 K) and increased cloudiness are generated for the Arctic region. The standard CCM sea ice albedo scheme is used as an adjustment to maintain normal temperatures for the polar oceans. The radiative impact of leads and melt ponds warmed sea ice regions only for short time periods. The third scheme generated markedly lower planetary albedos (reductions of 0.07 to 0.17) and higher surface temperatures (up to 2.0 K) than control values.The CCM simulates a gradual decrease in spring and summer Arctic cloud cover whereas observations show a sharp spring increase. Examination of the CCM code, particularly the cloud parameterization, is required to address this problem.With 12 Figures     

Glacial termination: sensitivity to orbital and CO2 forcing in a coupled climate system model

 To study glacial termination and related feedback mechanisms, a continental ice dynamics model is globally and asynchronously coupled to a physical climate (atmosphere-ocean-sea ice) model. The model performs well under present-day, 11 kaBP (thousand years before present) and 21 kaBP perpetual forcing. To address the ice-sheet response under the effects of both perpetual orbital and CO₂ forcing, sensitivity experiments are conducted with two different orbital configurations (11 kaBP and 21 kaBP) and two different atmospheric CO₂ concentrations (200 ppmv and 280 ppmv). This study reveals that, although both orbital and CO₂ forcing have an impact on ice-sheet maintenance and deglacial processes, and although neither acting alone is sufficient to lead to complete deglaciation, orbital forcing seems to be more important. The CO₂ forcing has a large impact on climate, not uniformly or zonally over the globe, but concentrated over the continents adjacent to the North Atlantic. The effect of increased CO₂ (from 200 ppmv to 280 ppmv) on surface air temperature has its peak there in winter associated with a reduction in sea-ice extent in the northern North Atlantic. These changes are accompanied by an enhancement in the intensity of the meridional overturning and poleward ocean heat transport in the North Atlantic. On the other hand, the effect of orbital forcing (from 21 kaBP to 11 kaBP) has its peak in summer. Since the summer temperature, rather than winter temperature, is found to be dominant for the ice-sheet mass balance, orbital forcing has a larger effect than CO₂ forcing in deglaciation. Warm winter sea surface temperature arising from increased CO₂ during the deglaciation contributes to ice-sheet nourishment (negative feedback for ice-sheet retreat) through slightly enhanced precipitation. However, the precipitation effect is totally overwhelmed by the temperature effect. Our results suggest that the last deglaciation was initiated through increasing summer insolation with CO₂ providing a powerful feedback. Received: 22 February 2000 / Accepted: 17 September 2000     

Climate sensitivity due to increased CO2: experiments with a coupled atmosphere and ocean general circulation model

A version of the National Center for Atmospheric Research community climate model — a global, spectral (R15) general circulation model — is coupled to a coarse-grid (5° latitude-] longitude, four-layer) ocean general circulation model to study the response of the climate system to increases of atmospheric carbon dioxide (CO₂). Three simulations are run: one with an instantaneous doubling of atmospheric CO₂ (from 330 to 660 ppm), another with the CO₂ concentration starting at 330 ppm and increasing linearly at a rate of 1% per year, and a third with CO₂ held constant at 330 pm. Results at the end of 30 years of simulation indicate a globally averaged surface air temperature increase of 1.6° C for the instantaneous doubling case and 0.7°C for the transient forcing case. Inherent characteristics of the coarse-grid ocean model flow sea-surface temperatures (SSTs) in the tropics and higher-than-observed SSTs and reduced sea-ice extent at higher latitudes] produce lower sensitivity in this model after 30 years than in earlier simulations with the same atmosphere coupled to a 50-m, slab-ocean mixed layer. Within the limitations of the simulated meridional overturning, the thermohaline circulation weakens in the coupled model with doubled CO₂ as the high-latitude ocean-surface layer warms and freshens and westerly wind stress is decreased. In the transient forcing case with slowly increasing CO₂ (30% increase after 30 years), the zonal mean warming of the ocean is most evident in the surface layer near 30°–50° S. Geographical plots of surface air temperature change in the transient case show patterns of regional climate anomalies that differ from those in the instantaneous CO₂ doubling case, particularly in the North Atlantic and northern European regions. This suggests that differences in CO₂ forcing in the climate system are important in CO₂ response in regard to time-dependent climate anomaly regimes. This confirms earlier studies with simple climate models that instantaneous CO₂ doubling simulations may not be analogous in all respects to simulations with slowly increasing CO₂.A portion of this study is supported by the US Department of Energy as part of its Carbon Dioxide Research Program     

Polar amplification in a coupled climate model with locked albedo

In recent years, a substantial reduction of the sea ice in the Arctic has been observed. At the same time, the near-surface air in this region is warming at a rate almost twice as large as the global average—this phenomenon is known as the Arctic amplification. The role of the ice-albedo feedback for the Arctic amplification is still a matter of debate. Here the effect of the surface-albedo feedback (SAF) was studied using a coupled climate model CCSM3 from the National Center for Atmospheric Research. Experiments, where the SAF was suppressed by locking the surface albedo in the entire coupled model system, were conducted. The results reveal polar temperature amplification when this model, with suppressed albedo, is forced by a doubling of the atmospheric CO₂ content. Comparisons with variable albedo experiments show that SAF amplifies the surface-temperature response in the Arctic area by about 33%, whereas the corresponding value for the global-mean surface temperature is about 15%. Even though SAF is an important process underlying excessive warming at high latitudes, the Arctic amplification is only 15% larger in the variable than in the locked-albedo experiments. It is found that an increase of water vapour and total cloud cover lead to a greenhouse effect, which is larger in the Arctic than at lower latitudes. This is expected to explain a part of the Arctic surface–air-temperature amplification.     

Global and regional coupled climate sensitivity to the parameterization of rainfall interception

A coupled land?Catmosphere model is used to explore the impact of seven commonly used canopy rainfall interception schemes on the simulated climate. Multiple 30-year simulations are conducted for each of the seven methods and results are analyzed in terms of the mean climatology and the probability density functions (PDFs) of key variables based on daily data. Results show that the method used for canopy interception strongly affects how rainfall is partitioned between canopy evaporation and throughfall. However, the impact on total evaporation is much smaller, and the impact on rainfall and air temperature is negligible. Similarly, the PDFs of canopy evaporation and transpiration for six selected regions are strongly affected by the method used for canopy interception, but the impact on total evaporation, temperature and precipitation is negligible. Our results show that the parameterization of rainfall interception is important to the surface hydrometeorology, but the seven interception parameterizations examined here do not cause a statistically significant impact on the climate of the coupled model. We suggest that broad scale climatological differences between coupled climate models are not likely the result of how interception is parameterized. This conclusion is inconsistent with inferences derived from earlier uncoupled simulations, or simulations using very simplified climate models.     

Objective comparison of patterns of CO2 induced climate change in coupled GCM experiments

 Four transient GCM experiments simulating the climatic response to gradually increasing CO₂, and two equilibrium doubled CO₂ experiments are compared. The zonally symmetric and asymmetric features of climate are both examined. Surface air temperature, sea level pressure, the 500 mb height and the relative topography between 500 and 1000 mb are analyzed. In the control simulations, the broad aspects of the present climate are in most cases well reproduced, although the stationary eddies tend to be less reliably simulated than the zonal means. However, the agreement between the four transient experiments on the geographical patterns of climate change is less impressive. While some zonally symmetric features, in particular the meridional distribution of surface air warming in the boreal winter, are rather similar in all models, the intermodel cross correlations for the zonally asymmetric changes are low. The agreement is largely restricted to some very general features such as more warming over the continents than over the oceans. The largest discrepancies between the two equilibrium-doubled CO₂ experiments and the transient experiments are found at the high southern latitudes, in particular in the austral winter. To identify the most robust geographical patterns of change in the transient experiments, the standard t test is used to determine if the four-model mean change is significantly above or below the global mean. Received: 18 January 1996 / Accepted: 5 July 1996     

Cloud albedo feedback and the super greenhouse effect in a global coupled GCM

Two competing cloud-radiative feedbacks identified in previous studies i.e., cloud albedo feedback and the super greenhouse effect, are examined in a sensitivity study with a global coupled ocean-atmosphere general circulation model. Cloud albedo feedback is strengthened in a sensitivity experiment by lowering the sea-surface temperature (SST) threshold in the specified cloud albedo feedback scheme. This simple parameterization requires coincident warm SSTs and deep convection for upper-level cloud albedos to increase. The enhanced cloud albedo feedback in the sensitivity experiment results in decreased maximum values of SST and cooler surface temperatures over most areas of the planet. There is also a cooling of the tropical troposphere with attendant global changes of atmospheric circulation reminiscent of those observed during La Niña or cold events in the Southern Oscillation. The strengthening of the cloud albedo feedback only occurs over warm tropical oceans (e.g., the western Pacific warm pool), where there is increased albedo, decreased absorbed solar radiation at the surface, stronger surface westerlies, enhanced westward currents, lower temperatures, and decreased precipitation and evaporation. However, the weakened convection over the tropical western Pacific Ocean alters the large-scale circulation in the tropics such that there is increased upper-level divergence over tropical land areas and the tropical Indian Ocean. This results in increased precipitation in those regions and intensified monsoonal regimes. The enhanced precipitation over tropical land areas produces increased clouds and albedo and wetter and cooler land surfaces. These additional contributions to decreased absorbed solar input at the surface combine with similar changes over the tropical oceans to produce the global cooling associated with the stronger cloud albedo feedback. Increased low-level moisture convergence and precipitation over the tropical Indian Ocean enhance slightly the super greenhouse effect there. But the stronger cloud albedo feedback is still the dominant effect, although cooling of SSTs in that region is less than in the tropical western Pacific Ocean. The sensitivity experiment demonstrates how a regional change of radiative forcing is quickly transmitted globally through a combination of radiative and dynamical processes in the coupled model. This study points to the uncertainties involved with the parameterization of cloud albedo and the major implications of such parameterizations concerning the maximum values of SST, global climate sensitivity, and climate change.Support is provided by the Office of Health and Environmental Research of the U.S. Department of Energy, as part of its Carbon Dioxide Research Program.The National Center for Atmospheric Research is sponsored by the National Science Foundation.     

Capturing the Atlantic cold tongue and coastal upwelling in an intermediate-level ocean model coupled to a regional climate model

Coupled atmosphere–ocean general circulation models are known to have difficulties simulating the cold tongue in the equatorial Atlantic Ocean. Here a regional climate model coupled to an intermediate-level mixed layer ocean model with Ekman dynamics is developed and used to better understand the seasonal evolution of the equatorial Atlantic cold tongue and upwelling off western Africa. Parameterization improvements are made to an earlier version of the ocean model to account for the variations in temperature and shearing stress at the base of the mixed layer. 90-km resolution sensitivity tests demonstrate that the development of the equatorial Atlantic cold tongue in the boreal spring/summer is captured only if seasonal variations in the temperature at the base of the ocean mixed layer are included. The development of cold temperatures off the northwest African coast in the late boreal winter/spring is found to be primarily associated with the net radiation balance as shortwave warming of the mixed layer is relatively low while latent cooling is relatively high yielding a net cooling of mixed layer temperatures, consistent with other studies. The westward extension of the Atlantic cold tongue is primarily due to the horizontal advection of cool water from the South Atlantic African coast. This coastal cooling is associated with vertical diffusion and vertical entrainment, while the vertical entrainment has a secondary and more localized role over the equatorial Atlantic.     

The role of low clouds in determining climate sensitivity in response to a doubling of CO2 as obtained from 16 mixed-layer models

The effects that low clouds in sub-tropical to tropical latitudes have in determining a given model’s climate sensitivity is investigated by analyzing the cloud data produced by 16 “slab” or mixed-layer models submitted to the PCMDI and CFMIP archives and their respective response to a doubling of CO₂. It is found that, within the context of the 16 models analyzed, changes of these low clouds appear to play a major role in determining model sensitivity but with changes of middle cloud also contributing especially from middle to higher latitudes. It is noted that the models with the smallest overall cloud change produce the smallest climate sensitivities and vice versa although the overall signs of the respective cloud feedbacks are positive. It is also found that the amounts of low cloud as simulated by the respective control runs have very little correlation with their respective climate sensitivities. In general, the overall latitude-height patterns of cloud change as derived from these more recent experiments agree quite well with those obtained from much earlier studies which include increases of the highest cloud, decreases of cloud lower down in the middle and lower tropospheric and small increases of low clouds. Finally, other mitigating factors are mentioned which could also affect the spread of the resulting climate sensitivities.     

ENSO variability and atmospheric response in a global coupled atmosphere-ocean GCM

The interannual variability associated with the El Ni?o/Southern Oscillation (ENSO) cycle is investigated using a relatively high-resolution (T42) coupled general circulation model (CGCM) of the atmosphere and ocean. Although the flux correction is restricted to annual means of heat and freshwater, the annual as well as the seasonal climate of the CGCM is in good agreement with that of the atmospheric model component forced with observed sea surface temperatures (SSTs). During a 100-year simulation of the present-day climate, the model is able to capture many features of the observed interannual SST variability in the tropical Pacific. This includes amplitude, lifetime and frequency of occurrence of El Ni?o events and also the phase locking of the SST anomalies to the annual cycle. Although the SST warming during the evolution of El Ni?os is too confined spatially, and the warming along the Peruvian coast is much too weak, the patterns and magnitudes of key atmospheric anomalies such as westerly wind stress and precipitation, and also their eastward migration from the western to the central equatorial Pacific is in accord with observations. There is also a qualitative agreement with the results obtained from the atmospheric model forced with observed SSTs from 1979 through 1994. The large-scale dynamic response during the mature phase of ENSO (December through February) is characterized by an eastward displacement and weakening of the Walker cell in the Pacific while the Hadley cell intensifies and moves equatorward. Similar to the observations, there is a positive correlation between tropical Pacific SST and the winter circulation in the North Pacific. The deepening of the Aleutian low during the ENSO winters is well captured by the model as well as the cooling in the central North Pacific and the warming over Canada and Alaska. However, there are indications that the anomalies of both SST and atmospheric circulation are overemphasized in the North Pacific. Finally, there is evidence of a coherent downstream effect over the North Atlantic as indicated by negative correlations between the PNA index and the NAO index, for example. The weakening of the westerlies across the North Atlantic in ENSO winters which is related to a weakening and southwestward displacement of the Icelandic low, is in broad agreement with the observations, as well as the weak tendency for colder than normal winters in Europe. Received: 31 October 1995 / Accepted: 29 May 1996     

Tropical Pacific internal atmospheric dynamics and resolution in a coupled GCM

A noise reduction technique, namely the interactive ensemble (IE) approach is adopted to reduce noise at the air–sea interface due to internal atmospheric dynamics in a state-of-the-art coupled general circulation model (CGCM). The IE technique uses multiple realization of atmospheric general circulation models coupled to a single ocean general circulation model. The ensembles mean fluxes from the atmospheric simulations are communicated to the ocean component. Each atmospheric simulation receives the same SST coming from the ocean component. The only difference among the atmospheric simulations comes from perturbed initial conditions, thus the atmospheric states are, in principle synoptically independent. The IE technique can be used to better understand the importance of weather noise forcing of natural variability such as El Niño Southern Oscillation (ENSO). To study the impact of weather noise and resolution in the context of a CGCM, two IE experiments are performed at different resolutions. Atmospheric resolution is an important issue since the noise statistics will depend on the spatial scales resolved. A simple formulation to extract atmospheric internal variability is presented. The results are compared to their respective control cases where internal atmospheric variability is left unchanged. The noise reduction has a major impact on the coupled simulation and the magnitude of this effect strongly depends on the horizontal resolution of the atmospheric component model. Specifically, applying the noise reduction technique reduces the overall climate variability more effectively at higher resolution. This suggests that “weather noise” is more important in sustaining climate variability as resolution increases. ENSO statistics, dynamics, and phase asymmetry are all modified by the noise reduction, in particular ENSO becomes more regular with less phase asymmetry when noise is reduced. All these effects are more marked for the higher resolution case. In contrast, ENSO frequency is unchanged by the reduction in the weather noise, but its phase-locking to the annual cycle is strongly dependent on noise and resolution. At low resolution the noise structure is similar to the signal, whereas the spatial structure of the noise deviates from the spatial structure of the signal as resolution increases. It is also suggested that event-to-event differences are largely driven by atmospheric noise as opposed to chaotic dynamics within the context of the large-scale coupled system, suggesting that there is a well-defined “canonical” event.     

Empirical Estimates of Global Climate Sensitivity： An Assessment of Strategies Using a Coupled GCM

A control integration with the normal solar constant and one with it increased by 2.5% in the National Center for Atmospheric Research （NCAR） coupled atmosphere-ocean Climate System Model were conducted to see how well the actual realized global warming could be predicted just by analysis of the control results. This is a test, within a model context, of proposals that have been advanced to use knowledge of the present day climate to make ＂empirical＂ estimates of global climate sensitivity. The scaling of the top-of-the-atmosphere infrared flux and the planetary albedo as functions of surface temperature was inferred by examining four different temporal and geographical variations of the control simulations. Each of these inferences greatly overestimates the climate sensitivity of the model, largely because of the behavior of the cloud albedo. In each inference the control results suggest that cloudiness and albedo decrease with increasing surface temperature. However, the experiment with the increased solar constant actually has higher albedo and more cloudiness at most latitudes. The increased albedo is a strong negative feedback, and this helps account for the rather weak sensitivity of the climate in the NCAR model. To the extent that these model results apply to the real world, they suggest empirical evaluation of the scaling of global-mean radiative properties with surface temperature in the present day climate provides little useful guidance for estimates of the actual climate sensitivity to global changes.     

The albedo of temperate and boreal forest and the Northern Hemisphere climate: a sensitivity experiment using the LMD GCM

A deforestation experiment is performed using the Laboratoire de Meteorologie Dynamique Atmospheric General Circulation Model (LMD GCM) to determine the climatic role of the largest vegetation formation in the Northern Hemisphere, localized mostly north of latitude 45°N, which is called the temperate and boreal forest. For this purpose, an iterative albedo scheme based on vegetation type, snow age, snowfall rate and area of snow cover, is developed for snow-covered surfaces. The results show a cooling of Northern Hemisphere soil and an increase in the snow cover when the forest is removed, as found by previous similar experiments.In our study this cooling is related to different causes, depending on the season. It is linked to modifications in the soil radiative properties, like surface albedo, due to the disappearance of forest, and consequently, to a greater exposure of the snow-covered soil underneath. It is also related to alterations in the hydrological cycle, observed mainly in summer and autumn at middle latitudes. The model shows a strong sensitivity to the coupled surface albedo — soil temperature — fractional snow cover response in the spring. A later and longer snowmelt season is also detected.This study adds to our understanding of climatic variation on longer time scales, since it is widely accepted that the formation and disappearance of different vegetation formations is closely related to climatic evolution patterns, in particular on the time scale of the glacial oscillations.     

Intercomparison and validation of snow albedo parameterization schemes in climate models

Snow albedo is known to be crucial for heat exchange at high latitudes and high altitudes, and is also an important parameter in General Circulation Models (GCMs) because of its strong positive feedback properties. In this study, seven GCM snow albedo schemes and a multiple linear regression model were intercompared and validated against 59 years of in situ data from Svalbard, the French Alps and six stations in the former Soviet Union. For each site, the significant meteorological parameters for modeling the snow albedo were identified by constructing the 95% confidence intervals. The significant parameters were found to be: temperature, snow depth, positive degree day and a dummy of snow depth, and the multiple linear regression model was constructed to include these. Overall, the intercomparison showed that the modeled snow albedo varied more than the observed albedo for all models, and that the albedo was often underestimated. In addition, for several of the models, the snow albedo decreased at a faster rate or by a greater magnitude during the winter snow metamorphosis than the observed albedo. Both the temperature dependent schemes and the prognostic schemes showed shortcomings.     

Sensitivity of tropical Atlantic climate to mixing in a coupled ocean–atmosphere model

The sensitivity of tropical Atlantic climate to upper ocean mixing is investigated using an ocean-only model and a coupled ocean–atmosphere model. The upper ocean thermal structure and associated atmospheric circulation prove to be strongly related to the strength of upper ocean mixing. Using the heat balance in the mixed layer it is shown that an excessively cold equatorial cold tongue can be attributed to entrainment flux at the base of the oceanic mixed layer, that is too large. Enhanced entrainment efficiency acts to deepen the mixed layer and causes strong reduction in the upper ocean divergence in the central equatorial Atlantic. As a result, the simulated sea surface temperature, thermocline structure, and upwelling velocities are close to the observed estimates. In the coupled model, the seasonal migration of the Intertropical Convergence Zone (ITCZ) reduces when the entrainment efficiency in the oceanic mixed layer is enhanced. The precipitation rates decrease in the equatorial region and increase along 10°N, resulting in a more realistic Atlantic Marine ITCZ. The reduced meridional surface temperature gradient in the eastern tropical Atlantic prohibits the development of convective precipitation in the southeastern part of the tropical Atlantic. Also, the simulation of tropical Atlantic variability as expressed in the meridional gradient mode and the eastern cold tongue mode improves when the entrainment efficiency is enhanced.     

Sensitivity of the thermohaline circulation and climate to ocean exchanges in a simple coupled model

 A new simple, coupled climate model is presented and used to investigate the sensitivity of the thermohaline circulation and climate to ocean vertical and horizontal exchange. As formulated, the model highlights the role of thin, ocean surface layers in the communication between the atmosphere and the subsurface ocean. Model vertical exchange is considered to be an analogue to small-scale, diapycnal mixing and convection (when present) in the ocean. Model horizontal exchange is considered to be an analogue to the effects of the wind-driven circulation. For small vertical exchange in the ocean, the model exhibits only one steady-state solution: a relatively cold, mid-high-latitude climate associated with a weak, salinity-driven circulation (“off ” mode). For large vertical and horizontal exchange in the ocean, the model also exhibits only one steady-state solution: a relatively warm, mid-high-latitude climate associated with a strong, thermally-driven circulation (“on” mode). For sufficiently weak horizontal exchange but large enough vertical exchange, both modes are possible stable, steady-state solutions. When model parameters are calibrated to fit tracer distributions of the modern ocean-atmosphere system, only the “on” mode is possible in this standard case. This suggests that the wind-driven circulation in consort with diapycnal mixing suppresses the “off ” mode in the modern ocean-atmosphere system. Since both diapycnal mixing and the wind-driven circulation would be expected to increase in a cold climate with greater meridional temperature gradients and enhanced winds, vertical and horizontal exchange in the ocean are probably associated with strong negative feedbacks which tend to stabilize climate. These results point to the need to resolve ocean wind-driven circulation and to greatly improve the treatment of ocean diapycnal mixing in more complete models of the climate system. Received: 16 November 1999 / Accepted: 19 June 2000