PP22A-01 INVITED 10:20h
The Dispersal of the 8.2 ka Lake Agassiz Drainage Waters Into the NW Atlantic: Working Hypothesis
The so-called "8.2 ka event", often associated with the drainage of glacial Lake Agassiz through Hudson Strait into the Labrador Sea that occurred some 8.4 ka ago, has been "seen" in various types of records. Absent in the Antarctica ice records, it appears in some Arctic ice records but not others. It is clearly recorded in the GRIP and North GRIP records, although with distinct features (notably a reverse deuterium-excess trend, suggesting that a proximal source of moisture could be involved; V. Masson pers. com.). Several paleoclimate sequences from Western Europe depict highly variable climate oscillations that are often attributed to this event. Such oscillations generally differ in their duration (from less than 40 years up to several hundreds of years) and, in many cases, do not differ significantly from the subsequent climate oscillations of the Holocene. Elsewhere, continental evidence for a major climate event is often unclear or questionable. Most scientists evoke a collapse or a strong reduction in the rate of the Atlantic Meridional Overturning (AMO) as an explanation for their observations, thus linking the pulses of freshwater release, the thermohaline circulation (THC), and the atmospheric response that would have resulted in largely spread "climatic anomalies". Some model experiments seem to support this scenario, but they have not necessarily been run with realistic freshwater release rates or routes. Most however would not allow a short "THC-excursion" thus raising concern about a less than 40 yr-event as reported in some records. Therefore, some strong marine "evidence" for at least a reduction of the AMO is needed. Unfortunately, most marine records have a time resolution unsuitable to ascertain a short duration reduction of the AMO, and/or are smoothed due to benthic mixing, thus dampening high frequency signals into the stochastic noise. When reported in marine records from the North Atlantic, the presence of this 8.2 ka event seems generally inconclusive. However, off Hudson Strait, near the outlet of the final drainage of Lake Agassiz, a few marine records show a very high time resolution (up to 15 cm/year), and minimum benthic mixing. In these records, some evidence for a very short "sedimentological event" has been found, but none show any indication for a significant concomitant change in sea-surface salinity or temperature, based on micropaleontological or isotopic data. The question thus arises of the fate of the near 160 * 103 km3 of drainage waters released into the Labrador Sea, possibly in a couple of years, based on glaciological models. We hypothesize here that this water merged into the Labrador Sea, after a long subglacial tunnel trajectory, as a dense turbiditic flow, with possibly two spreading routes: one surface branch, nearshore, on the Labrador Shelf, and a deep underflow branch flowing practically along the NAMOC route. In both cases, the influence on open ocean sea surface conditions would have been negligible. In this scenario, the opening of Hudson Bay and the final collapse of the Laurentide ice sheet into two small residual ice caps would have resulted in a major atmospheric reorganization that could explain both long distance correlative climate responses and the influence of more proximal moisture sources indicated by the GRIP records.
PP22A-02 INVITED 10:35h
Reconstructing Paleocirculation Using d13C and 231Pa/230Th
Past changes in ocean circulation may be estimated from reorganization of subsurface water masses as indicated by nutrient proxies such as d13C. These reconstructions are invaluable, yet the changes they reveal are not necessarily related to the climatically significant transport of heat. Better estimates of this transport may be made using dynamical proxies of the rate of overturning. The burial ratio of unsupported 231Pa: 230Th in bulk North Atlantic sediments, age-corrected to the time of deposition, provides such a proxy. Uranium is well mixed in the ocean, and its decay produces 231Pa and 230Th at a constant activity ratio (0.093). Because the residence time in seawater with respect to removal by particles of Th is on the order of decades and that of Pa on the order of centuries, subsurface advection leads to a deficit in buried 231Pa/230Th in the North Atlantic. Changes in this burial ratio thus vary with the residence time of water in the basin, and are inversely proportional to the rate of meridional overturning and export of subsurface waters. The combination of d13C and 231Pa/230Th can therefore provide more complete insight into past circulation changes. Glacial, deglacial, and interglacial paleo-circulation configurations and rates have now been investigated with this paired approach. Large changes in water mass structure characterize the Atlantic Ocean at the last glacial maximum (LGM), yet the overall rate of meridional overturning was reduced by no more than approximately one third. In contrast, briefer millennial-scale reductions were larger in scale and were accompanied by dramatic climate changes in the North Atlantic region. Evidence from d13C suggests similarly dramatic changes during the glacial period and even during the Holocene. Only the glacial and deglacial events appear to have comparable impact on the rate of meridional overturning. Taken together, the results point to the transient nature of the response to forcing by surface buoyancy fluxes and the importance of location in such forcing. The paired results help constrain the effects of additional physical and bio-logical effects on d13C and 231Pa/230Th and confirm the value of these proxies for reconstructing ocean circulation.
PP22A-03 INVITED 10:50h
Bipolar Seesaw Concepts: Do we Understand North-South Climate Connections?
Recent ice core results from Greenland bring the number of Dansgaard/Oeschger (D/O) events to 25 [NorthGRIP Members, 2004]. It is surprising that the earliest D/O event occurs around the beginning of glaciation. Model simulations have shown that collapses and rapid onsets of the Atlantic thermohaline circulation explain many features of abrupt change found in different paleoclimatic archives. In this lecture we give an overview of the hierarchy of models describing the north-south climate connection during abrupt climate change. Changes in the Atlantic thermohaline circulation have a far-field effect which was first described by a simple bipolar seesaw [Crowley, 1992; Broecker, 1998] and later by the thermal bipolar seesaw [Stocker and Johnsen, 2003]. The latter concept has resolved the controversy of lead and lag between climate signals in the northern and southern hemispheres. However, its major shortcoming was that the southern damping time scale, derived from high-resolution, synchronised ice core records from Greenland and Antarctica, was about 1000 years and hence much longer than the typical adjustment times of the Southern Ocean of about 100 years. Using a coupled model of reduced complexity, but including a comprehensive ocean circulation component, we could show that the major north-south response to a shut down of the thermohaline circulation is well described by an extended bipolar seesaw concept with physically plausible time scales. Changes in the density structure, caused by the freshwater discharges, induce anomalous ocean circulations in the South Atlantic and modify the heat and water exchange with the Southern Ocean [Knutti et al., 2004]. This {\it thermal freshwater seesaw}, in which both temperature and freshwater anomalies contribute to the changes in southern temperature is able to explain most of the millennial variability in different paleoclimatic records during the last ice age. Open questions regard the trigger and location of thermohaline shut-downs and what the influence of southern freshwater releases on the climatic signals in Greenland and Antarctica might be. \scriptsize Broecker, W.S., {\it Paleoceanogr., 13}, 119-121, 1998. Crowley, T.J., {\it Paleoceanogr., 7}, 489-497, 1992. Knutti, R., et al., {\it Nature, 430}, 851-856, 2004. NorthGRIP Members, {\it Nature, 431}, 147-151, 2004. Stocker, T.F., and S.J. Johnsen, {\it Paleoceanogr., 18}, 1087, 2003.
PP22A-04 11:05h
Is there a bi-Polar Ocean Seesaw?
In ice records from Antarctica and Greenland, it has been noted that when one hemisphere is warm, the other hemisphere is cold and vice versa. Some paleodata and ocean-only modeling imply that this may be because the ocean thermohaline circulation (THC) behaves as a seesaw swinging from a regime with weaker North Atlantic Deep Water (NADW) formation to a regime with stronger NADW formation. These northern hemisphere regimes are associated with stronger and weaker Antarctic Deep Water formation, respectively. The seesaw supposedly can be driven by change of freshwater flux either in the northern North Atlantic or in the Southern Ocean. Here we investigate this issue using a fully coupled advanced ocean-atmosphere-land surface-sea ice model . The Manabe Climate Model (MCM) is used in idealized scenarios that we believe are favorable to alleged bi-polar seesaw regimes. The GFDL R30 atmosphere-ocean model is forced with a freshwater flux anomaly of 1 SV persisting for 100 model years beginning from a control integration. After 100 years, this additional freshwater flux is turned off and the model climate is allowed to recover. In one integration, the additional freshwater is added in the high latitudes of the North Atlantic Ocean. In the second integration, the excess of freshwater is added in the high latitudes of the Southern Ocean. In both cases, oceanic convection and the associated THC local to the freshwater input is greatly hindered and slowed down. This results in a large local cooling and freshening of the surface waters. The hemisphere with the freshwater input tends to cool at the surface, while the opposite hemisphere tends to warm. This supports the seesaw hypothesis. However, the response of most the climate system is much more complicated, and the simplistic seesaw concept needs to be reconsidered. These features will be discussed.
PP22A-05 11:20h
The Transient Atmosphere-Ocean Response to Fresh Water Perturbations Applied to the North Atlantic Basin
The issue of climate system instability involving rapid changes in the thermohaline circulation (THC) of the Atlantic Ocean is a continuing focus of climate dynamics research. As a contribution to the WCRP/CLIVAR coupled model and paleoclimate model inter-comparison projects (CMIP/PMIP), a series of experiments have been performed in which a range of different fresh water forcing anomalies have been applied to the North Atlantic Basin between 50 and 70 degrees north latitude. In the first experiment an equilibrated modern pre-industrial simulation, constructed using the NCAR CCSM, version 1.4, is forced with a fresh water forcing (FWF) anomaly of 0.1 Sv for 100 years and subsequently allowed to recover. A second experiment is also performed in which the 100 year FWF anomaly is increased to 1 Sv and the climate once more allowed to fully recover after this forcing is removed. This paper will describe the transient evolution of the global climate system before, during, and after the FWF anomaly is applied. Of particular interest is the large difference in response between the 0.1 Sv and 1 Sv experiments. The 0.1 Sv FWF anomaly induces only a small change in the North Atlantic THC, whereas the 1 Sv FWF anomaly induces a near complete shutdown of the overturning circulation. This version of the CCSM is therefore considerably less sensitive to "hosing" than some earlier models were found to be, with significant implications concerning the question as to whether THC collapse should be viewed as a probable consequence of ongoing global warming.
PP22A-06 11:35h
Effects of Surface Freshwater Fluxes on the Ocean General Circulation: A Theoretical Perspective
The changes in ocean circulation that are inferred from paleoclimate records are frequently ascribed to changes in the density of surface waters at high latitudes arising from anomalies in surface freshwater fluxes. This hypothesis is in apparent contradiction with two "theorems" based on the equations of fluid dynamics: the theorem of Sandstrom (1908) and the theorem of Paparella and Young (2002). The theorem of Sandstrom states that the supply and removal of buoyancy at the surface cannot drive the ocean general circulation. Previous attempts to formalize the argument are based on the equation for the circulation of the vector velocity round a material line. The theorem of Paparella and Young implies that a hypothetical ocean forced only by a buoyancy difference at the surface would not exhibit the observed levels of kinetic energy dissipation (as inferred from microstructure observations); other energy sources must be responsible for the flux of kinetic energy in the ocean. Unlike the Sandstrom theorem the theorem of Paparella and Young is based on the full set of the equations of motion (in the Boussinesq approximation) and could be viewed as a more robust statement. Both theoretical arguments suggest that surface buoyancy fluxes constitute an ineffective mechanism for supplying energy to the ocean general circulation. In our presentation we will discuss the strengths and limitations of both arguments and attempt to clarify their paleoclimate implications.
PP22A-07 11:50h
Does the Oceanic Meridional Overturning Cell Really Have More Than one Stable State ??
Numerical climate and ocean models consistently show that the meridional overturning cell (MOC) has two stable states for the same fresh water flux (into the ocean). One of these two states usually corresponds to a high northward transport of surface water (and, hence, to a warm northern hemisphere climate) whereas the other corresponds to low northward transport (and, hence, to a cool northern hemisphere climate). This two-states scenario leads to the broadly quoted concern that our present day warm climate can perhaps spontaneously flip to a much cooler state. We present new analytical and numerical runs showing that these two states are an art-effect of the high eddy diffusivities most often used in the commonly employed numerical models. For the Atlantic, these diffusivities artificially introduce unrealistic upwelling into the thermocline within the limits of the Atlantic itself. These, in turn, introduce the familiar hysteresis corresponding to the two-states. In the real ocean, the convection in the north Atlantic is too strong and the vertical diffusivities in the thermocline are too small to allow for such upwelling within the limits of the Atlantic itself. As a result, the water which ultimately sinks in the north Atlantic is drawn into the Atlantic from regions far away--the Southern and Indian oceans. Both the analytical and the numerical results show that the two states converge into one in the limit of small diffusivity. This one-state scenario doesn't at all mean that the MOC cannot collapse due to a large fresh water flux--it certainly can do so if the fresh water flux is large enough. Rather, it means that there are no two-states for the same fresh water flux, implying that the transition from warm to a cold state cannot happen spontaneously.
PP22A-08 12:05h
Is a Shutdown of the Thermohaline Circulation Irreversible?
The thermohaline circulation (THC) in the North Atlantic plays a vital role in explaining past abrupt climate changes and in maintaining the current climate. Its remarkable nonlinear dynamics, first demonstrated by Stommel, has been supported by different types of models. This has led to concern that global warming may shut down the THC irreversibly, with consequent catastrophic climate changes, particularly for Europe. However, recent simulations by complex atmosphere/ocean general circulation models show a great suppression of the nonlinear response of the THC to external freshwater forcing. In this study a suite of models is used to investigate the nonlinear response of the THC to freshwater addition. It is found that the THC simulated by an ocean general circulation model responds very differently depending on whether it is uncoupled or coupled to an atmosphere general circulation model. The THC shuts down irreversibly in the uncoupled ocean general circulation model (OGCM) simulations, but reversibly in the coupled atmosphere/ocean general circulation model (AOGCM) simulation. This occurs because of a crucial negative feedback in the AOGCM simulation that cannot occur in the OGCM simulations. Analysis of Stommel's 2-box ocean model within different parameter regimes supports this finding. Thus, the irreversible shutdown of the THC caused by freshwater addition appears to be a model artifact rather than a likely outcome of global warming.