Title: The Arctic Ocean and its Role in Climate Change/Variability
Principal Investigator: Andrew Weaver
School of Earth and Ocean Sciences, University of Victoria
PO Box 3055, Victoria, BC, V8W 3P6
Tel: (250) 472-4001; Fax: (250) 472-4004
Institute of Ocean Sciences, PO Box 6000, Sidney BC, V8L 4B2
Tel: (250) 363-6585; Fax (250) 363-6746
Canadian Centre for Climate Modelling and Analysis,
University of Victoria, PO Box 1700, Victoria, BC, V8W 2Y2
Tel: (250) 363-8233; Fax: (250) 363-8247
Department of Atmospheric & Oceanic Sciences, McGill University
805 Sherbrooke St. W., Montreal, QC, H3A 2K6
Tel: (514) 398 3768; Fax: 514-398-6115
The purpose of this proposal is to develop a new component of the Canadian Climate Research Network aimed at understanding the Arctic Ocean and its role in climate change and climate variability. The ultimate goal of this proposal is to develop an improved representation of the Arctic Ocean, its sea ice cover and their coupling for use in the CCCma coupled model. In working towards this goal, two phenomenological questions will be addressed: 1) What processes drive interannual variability in the Arctic freshwater export, and how does this variability affect the global ocean and climate; 2) Is the recent observation of warm, subsurface North Atlantic water intrusion into the Arctic consistent with the response expected from enhanced anthropogenic greenhouse gas radiative forcing, or is it simply the Arctic response to low frequency variations in atmospheric circulation? Our approach will involve the analysis of output from the CCCma coupled model and the systematic testing of improved parameterizations and representations of sea ice and its interaction with the Arctic Ocean using the UVic coupled model. Further analysis of the most important aspects of these experiments will be examined through a reduced set of experiments using the more sophisticated CCCma coupled model.
The purpose of this proposal is to initiate new research into the role of the Arctic Ocean in climate change and climate variability. A team of researchers has been put together with expertise in the areas of ocean/ice/climate modelling, ocean data collection and data analysis to tackle the scientific questions outlined below. The theoretical understanding and modelling technology developed through the course of this project will benefit CCCma global climate modelling activities, and will compliment work being done in other CRN nodes and CRYSYS (section 3.2) The present proposal provides a complimentary oceanographic perspective to these initiatives, especially with regard to coupled sea-ice and ocean processes. In addition, it will provide an additional Canadian contribution to the WCRP ACSYS, CLIC and CLIVAR international programmes.
1.1 Overall Goals
This proposal has both a phenomenological and a technological component. The phenomenological component is designed to address the following two scientific questions:
In this subsection we review the observational, theoretical and modelling studies which motivate us to propose the research outlined in section 2. This subsection is broken down according to the two phenomenological and one technical question outlined above.
1.2.1 What processes drive interannual to interdecadal variability in Arctic freshwater export, and how does this variability affect the global ocean and climate?
Approximately 10% of the world's river runoff, accounting for ~3300 km3/yr (with large seasonal and interannual variability - Cattle, 1985), enters the Arctic Ocean, which occupies only ~5% of the total ocean surface area and ~1.5% of its volume. Bering Strait inflow represents the second largest freshwater source for the Arctic Ocean (~1670 km3/yr), with precipitation minus evaporation (~900 km3/yr ) and the import of freshwater in the Norwegian coastal current (~330 km3/yr ) accounting for the remainder of the freshwater sources. The sources of freshwater for the Arctic are balanced primarily by the export of sea ice in the East Greenland Current, which accounts for a freshwater loss to the Arctic and a gain by the Greenland Sea of ~2800 km3/yr (Aagaard and Carmack 1989). The exchange of water through the Canadian Archipelago and Fram Strait results in a loss of approximately 900 km3/yr and 820 km3/yr of freshwater, respectively. At low temperatures the density of sea water is largely controlled by salinity. As such, variations in the freshwater exchange (via both ocean and sea ice transports) between the Arctic and Atlantic Oceans are likely to affect the formation of deep and intermediate water masses there. Indeed, modelling results from Mauritzen and Hkkinen (1997) show that the thermohaline circulation increases by 10-20% in response to a decrease in sea ice export of 800 km3. The relative strength of the freshwater sources to the Nordic and Labrador Seas from the Arctic will also likely influence the preferred location and relative strengths of deep water formation. From a relatively short time series (1979-1985), Steele et al. (1996) show that simulated interannual variability in the outflow through the Canadian Archipelago is anticorrelated with the outflow through Fram Strait, with the Fram Strait anomalies leading the Canadian Archipelago anomalies by one year. This may explain why deep water formation in the Nordic and Labrador Seas have been observed to be out of phase in the past few decades. Alternatively, Mysak et al. (1990) and Mysak and Venegas (1998) suggest that this out of phase relationship may be associated with Greenland Sea ice export anomalies which freshen the surface waters and result in subsequent Labrador Sea ice anomalies once the freshwater has reached that region.
The large changes that occur in Arctic/North Atlantic freshwater exchange are epitomized by the Great Salinity Anomaly (GSA) of the late 1960s. This event freshened the upper 500 m of the northern North Atlantic with a freshwater excess of approximately 2000 km3 (or 0.032 Sv over a two year period). Dickson et al. (1988) traced this fresh anomaly as it was advected around the subpolar gyre for over 14 years. It originated north of Iceland in the late 1960s, moving southwestward into the Labrador Sea (1971-1973) and then proceeding across the north Atlantic, returning to the Greenland Sea in 1981-1982.
Several studies have examined the cause of the GSA and have generally determined that it was a result of Arctic/North Atlantic interactions. Both modelling (Hkkinen 1993) and observational (Walsh and Chapman 1990; Wohlleben and Weaver 1995; Mysak and Venegas 1998) studies concluded that strong northerly winds (in the Fram Strait region) caused an increased sea ice export into the Greenland Sea. The large freshwater flux anomaly that was associated with this transport was likely enhanced by the relatively large advection of thick ice from north of Greenland. Additionally, as simulated by Hkkinen (1993), increased oceanic transport of freshwater from the Arctic occurred. This was caused by fresh anomalies within the Siberian Sea that were advected across the Arctic, entering the Greenland Sea approximately 4 years later. During the GSA, the anomalous sea ice and oceanic freshwater transports were coincident, resulting in a significant and persistent freshening of the north Atlantic. This appears to have resulted in a reduction of deep water formation with winter convection in the Labrador sea limited to the upper 200 m (compared to 1000-1500 m for 1971-1973) (Lazier 1980).
North Atlantic SST records for the past century reveal slowly varying basin-scale changes including cold anomalies prior to 1920, warming from 1930-1940, and cooling again in the 1960s. Kushnir (1994) described the SST pattern associated with these long-term changes as uni-polar with a strong maximum around Iceland and in the Labrador Sea and a weaker maximum in a band near 35°N across the central Atlantic. The atmospheric pattern associated with the cooling in the 1960s has a negative pressure anomaly to the east of positive SST anomalies (also see Deser and Blackmon 1993 who suggest the pattern resembles the North Atlantic Oscillation [NAO]). Because the SLP anomalies appear downstream of the SST anomalies, these authors suggest that the atmosphere is responding to the ocean on these timescales.
Evidence for changes in the subpolar North Atlantic Ocean over similar timescales, compiled by Dickson et al. (1996), indicates that synchronous with the cooling in the late 1960s, convective activity reached a maximum in the Greenland Sea and a minimum in the Labrador Sea. These convective extremes occurred at the approximate time of the GSA. Since the early 1970s, the Greenland Sea has become progressively more saline and warmer through horizontal exchange with the deep waters of the Arctic Ocean. At the same time, the Labrador Sea has become colder and fresher as a result of local deep convection. Reverdin et al. (1997) explored patterns associated with salinity anomalies and found a single pattern explains 70% of the variance of lagged salinity anomalies. The pattern represents a signal originating in the Labrador Sea that propagates from the west to the northeast in the subpolar gyre. The strong correlation between salinity and sea ice in the Labrador Sea lead Reverdin et al. (1997) to link the salinity anomalies to the export of Arctic freshwater.
Delworth et al. (1993) described the first coupled ocean-atmosphere GCM study of long-term thermohaline variability. They associated the variability primarily with oceanic processes. Later Delworth et al. (1997) found salinity anomalies in the surface layer of the Arctic Ocean precede anomalies of the thermohaline intensity by 10-15 years. In agreement with the proposed climate cycle of Wohlleben and Weaver (1995), these Arctic freshwater anomalies are connected to the North Atlantic through SLP anomalies in the Greenland Sea resembling the pattern that Walsh and Chapman (1990) report preceded the GSA. Weaver and Valcke (1998) gave further evidence to suggest that the GFDL model thermohaline variability is a mode of the fully coupled atmosphere-ocean-ice system.
Thompson and Wallace (1998) recently analysed the northern hemisphere (north of 20°N) sea-level pressure field and noted that the dominant EOF revealed a spatial pattern whose centre was over the Arctic. Over the Arctic this EOF appeared zonally-symmetric, although signals of opposite sign were located over the Atlantic and Pacific Ocean, breaking down the zonal-symmetry. They argued that land sea contrasts were the source of this asymmetry. They further termed this pattern, which revealed strong seasonal-interdecadal variability, the "Arctic Oscillation" (AO) and noted that it resembled the NAO in the North Atlantic, although the NAO had no signature in the Pacific. Correlations between the AO index and Eurasian surface air temperatures were substantially higher than similar correlations using the NAO index. Finally, they noted that the AO consisted of two components: the first being of equivalent barotropic nature and extending well into the stratosphere, giving rise to a zonally-symmetric pattern; the second being confined to the troposphere and of a baroclinic nature, leading to an asymmetric zonal circulation. This oscillation, which contains the NAO as an apparent subset, appears to have important links to the variability of the Arctic Ocean, its sea ice and hence freshwater export. Indeed the AO may be an important natural mode of the climate system whose low-frequency modulation may be inherently linked to the observed northern hemisphere decadal-interdecadal climate variability.
1.2.2 Is the recent observation of warm, subsurface North Atlantic water intrusion into the Arctic consistent with the response expected from enhanced anthropogenic greenhouse gas forcing, or is it simply the Arctic response to low frequency variations in atmospheric circulation?
In the Arctic Ocean, water of Atlantic origin is evident as a relatively warm and salty water-mass at intermediate depths (200-600 m). Recent observations (e.g. Carmack et al. 1995; McLaughlin et al. 1996) indicate that the Atlantic layer within the Arctic Ocean has undergone large changes since 1990. These include a shift in the frontal structure which separates different Atlantic layer water masses (from the Lomonosov to the Mendeleyev ridge), and a significant warming of the Atlantic layer. By 1994, this warming extended across the Nansen, Amundsen and Makarov Basins.
Observations (Carmack et al., 1995; McLaughlin et al., 1996; Carmack et al., 1998a) have also shown that the recent changes in the Arctic, outlined above, are manifested by three processes:
In some recent experiments using the UVic coupled model (Wiebe and Weaver 1998) which examined the transient response of the climate system to increasing anthropogenic CO2, we found that subsurface warm waters intruded into the Arctic. Once in the Arctic, these waters were slowly advected and diffused throughout the basin, filling the middle layers with an anomalously warm water mass of Atlantic origin (see http://wikyonos.seos.uvic.ca/climate-lab/movies.html). Such warm water intrusion only existed in the experiments which replaced the usual horizontal/vertical mixing scheme with either a mixing scheme with reduced near-surface diffusivities, or the Gent and McWilliams (1990) parameterization for mesoscale mixing (which include a rotation of the diffusion tensor to be aligned along isopycnals). In the usual horizontal/vertical mixing case, there was enhanced intrusion of the surface waters which dramatically melted back the ice edge, exposing the ocean surface to the cold polar air. Strong convection then eliminated the signature of the warm Atlantic waters entering the Arctic. During the transient CO2 increase, the ice edge in the other cases was not as greatly affected so that the subsurface signature of the intrusion of warm Atlantic waters survived.
While this warming of the mid-depth Arctic Ocean is qualitatively similar to the aforementioned recent observed patterns of subsurface warming, there are several discrepancies, including a spreading timescale which is too slow and a maximum warming which is slightly too deep. The lack of a proper representation of the Arctic Ocean in the Wiebe and Weaver (1998) version of the coupled model severely limited our ability to quantitatively capture the dynamics of the region. Nevertheless, it is interesting that the observed warming trend is consistent with our model response to increasing atmospheric CO2, and we will address this issue in more detail in section 2.2.
An alternative explanation is afforded by the compilation of observational and modelling results of Proshutinsky and Johnson (1997) and Grotefendt et al. (1998). They find that the wind-driven circulation of sea-ice and the upper-ocean tends to alternate between two distinct modes: one with a large Beaufort gyre and transpolar drift directed towards Fram Strait; one with a small Beaufort gyre and transpolar drift directed towards the Canadian Archipelago. Fluctuations between these two modes occur on roughly decadal timescales. Their circulation patterns are broadly consistent with the observed shift in freshwater export from Fram Strait to the Canadian Archipelago, a corresponding shift in the import of North Atlantic Water, and a resulting displacement of water mass boundaries.
Swift et al. (1998) show that the recent changes in the Arctic are likely caused by an increase in the temperature of the Atlantic waters which enter the Arctic Basin through Fram Strait. The anomalous warmth of these waters appears to be correlated with the NAO, and hence the AO, which corresponds to relatively warm air temperatures in the Greenland Sea region and thus a reduction in oceanic heat loss. An open question remains as to where the Arctic waters displaced through the intrusion of the Atlantic layer went, although enhanced transport through the Canadian Archipelago is plausible. This would be consistent with recent observations of anomalous cold and fresh waters in the Labrador Sea since the late 1980s (Dickson et al. 1996). Nevertheless, it is clear that channel-flow approaches to the movement of water masses through the Arctic Archipelago are woefully inadequate. Model parameterizations must account for: buoyancy-boundary currents; sub-basin circulation; tidal mixing and frontogenesis; wind-forced and ice-forced surface flows; thermohaline circulation within channels.
1.2.3 Develop improved parameterisations of processes important for the representation of the Arctic Ocean, its sea ice cover and their coupling for potential use in the CCCma coupled model.
The representation of the Arctic in the CCCma coupled model (Flato et al. 1998), and indeed all other coupled models, is currently very poor. The CCCma ocean model has a resolution of 1.875° in longitude by 1.856° in latitude on a spherical grid. Due to the convergence of meridians at high latitudes, Fourier filtering is necessarily employed, introducing unrealistic small-scale features into resulting the Arctic Ocean tracer, velocity and surface flux fields. In addition, substantial smoothing of topography is required as well as the inclusion of an artificial island at the North Pole. The sea ice component of the coupled model is the same as that used in earlier uncoupled CCCma climate change experiments (McFarlane et al., 1992) and retains only a zero-layer thermodynamic component (Semtner, 1976) with an improved representation of the energy balance at the ice-atmosphere interface. The salt or freshwater flux associated with sea ice growth or melt is calculated by assuming that salt is immediately expelled upon freezing, leaving pure ice. While the latest version of the CCCma coupled model does include a simple representation of sea ice dynamics (the cavitating fluid approach of Flato and Hibler 1992), the thermodynamics is still very crude and the Arctic Ocean is still poorly resolved. While the revised version of the ocean model no longer requires an island at the pole, Fourier filtering is still necessarily employed.
The crude representation of the Arctic Ocean and its sea ice cover in all coupled models is particularly important since these same coupled models all suggest that climate change, associated with increasing anthropogenic greenhouse gases, is amplified there. This northern amplification is associated with a northward retreat of sea ice cover and the subsequent reduction in the surface albedo (a positive feedback). Since Canada has a vast Arctic expanse and as climate change from coupled models is enhanced in the Arctic, we believe the improvement of the Arctic Ocean and its sea ice cover in the CCCma coupled model is both of particular Canadian interest, and would provide a unique contribution to international modelling efforts.
In addition, the CCCma coupled model (and indeed all coupled models) assumes that during the process of sea ice formation, brine rejection occurs on the scale of the ocean grid. This assumption is known to cause excessive convection and subsequent vertical mixing and oceanic heat loss (e.g. Duffy and Caldeira, 1997; Caldeira and Duffy, 1998). In reality, brine rejection occurs on very small spatial scales (hundreds of metres - Denbo and Skyllingstad, 1996). To account for this effect we (Duffy et al., 1998) have recently included a more realistic parameterization for brine rejection into the UVic coupled model. In this parameterization rejected salt is mixed to a depth which is calculated based on a prescribed density contrast relative to the surface, prior to the initiation of convection. This approach has the realistic property that rejected salt is mixed more deeply in regions where the vertical density stratification is weak, and less deeply in regions where the stratification is strong. The results from the inclusion of this parameterization, together with the Gent and McWilliams (1990) parameterization for mixing associated with mesoscale eddies, were dramatic and extremely encouraging. Spurious southern ocean convection was eliminated; the formation and representation of Antarctic intermediate Water was substantially enhanced; salinities were more realistic (both globally and locally); the overcooling problem of deep ocean temperatures when the Gent and McWilliams (1990) parameterisation is used was eliminated; sea ice extents were also improved.
2 Outline of proposed research
As in the previous section, this section will be focused around the same three themes. Section 2.4 outlines a list of milestones with which our progress can be compared. We have chosen to undertake the parallel approach of attempting to understand observed climate phenomena, while developing improved representations of the Arctic Ocean, its sea ice cover, and their coupling for two reasons. First, attempting to capture the essential dynamics and response of the coupled system involved in these phenomena provides a necessary validation test for both the UVic and CCCma coupled models. Second, once an initial understanding of the phenomena is obtained, we will undertake an exhaustive parameter sensitivity study to determine how their realization in the UVic coupled model responds to changes in the constituent components/parameterisations of the model. This will allow us to focus our attention on a narrow subset of parameterizations which need improvement/better representation in the CCCma coupled model. These improvements will be tested thoroughly using the UVic coupled model before being examined within the context of the more computationally expensive CCCma coupled model.
2.1 What processes drive interannual to interdecadal variability in Arctic freshwater export, and how does this variability affect the global ocean and climate?
Our conjecture is that Arctic sea ice export (and its relationship to the NAO and the Arctic Oscillation) plays an integral role in decadal-interdecadal North Atlantic ocean/climate variability.
The impact of ice export on climate variability will initially be addressed by applying an anomalous wind stress forcing to the UVic coupled model. The first 20 EOFs from NCEP reanalysis SLP data have been used to generate a synthetic anomalous wind stress field which has been applied over the North Atlantic Ocean. Initial results are extremely encouraging as the model reveals decadal-interdecadal variability around the North Atlantic which is intimately linked to the export of sea ice from the Arctic. Nevertheless, much analysis remains to isolate the dominant mechanism and timescale for the variability and to unequivocally prove that sea ice dynamics are crucial to the oscillation. These sensitivity analyses will involve 1) applying the anomalous wind forcing only over ice; 2) shutting off the oceanic ice advection; 3) adding the anomalous wind forcing effects to the model calculation of latent and sensible heat fluxes; 4) changing the number of categories in the thermodynamic component of the sea ice model; 5) removing the oceanic effects of brine rejection and sea ice melting; 6) using climatological (from the spin up of the coupled model) freshwater fluxes or heat fluxes to determine whether heat or freshwater flux changes amplify or diminish the variability. In addition, it will be important for us to determine, through sensitivity analyses to internal sea ice parameters (e.g., number of categories, shear strength, vertical temperature resolution, ice-ocean coupling parameter, albedo), whether or not the UVic coupled model allows self-sustained oscillations within a particular parameter range.
The modelling work at McGill will centre around the use of the new dynamic sea ice model of Tremblay and Mysak (1997) which has a rheology based on that of a granular material. Currently, the model domain consists of the Arctic Basin and Greenland-Norwegian Sea (extending to 65° N), and the ice model (which also has a thermodynamic part that is a modification of Hibler, 1979) is coupled to a mixed layer ocean model (with an annual mean ocean circulation) and a one layer thermodynamic atmospheric model, which allows for an ice albedo feedback. Land is represented by a 6 m thick layer with a constant base temperature. Under prescribed wind stress climatological forcing, the model yields a well-positioned ice edge in both winter and summer, and sea ice circulation and thickness distributions that are in good agreement with observations. Also, the commonly observed lead complexes, along which sliding and opening of adjacent ice floes occur in the Arctic sea ice cover, are well produced in the simulation. In particular, the shear lines extending from the western Canadian Archipelago toward the central Arctic, often observed in winter satellite images, are present. For further details of the climatological run, see Tremblay and Mysak (1997).
As a first step toward studying the internal variability of the Arctic sea ice cover using the Tremblay and Mysak (1997) model, Tremblay and Mysak (1998) examined the origin and space-time evolution of sea ice anomalies in the Beaufort and Chukchi seas. They found that an ice thickness anomaly placed in the western Arctic could survive a few seasonal cycles in the model as it was transported by the Beaufort Gyre toward the Transpolar Drift Stream and then exported out of the Arctic Basin into the Greenland Sea via Fram Strait. Building upon these results, Arfeuille (1999) used the Tremblay-Mysak sea ice model to study the interannual variability of the Arctic sea ice cover and sea ice export into the Greenland Sea during the 41-year period 1958-98. Monthly wind stress forcing derived from NCEP reanalysis data for this period were used to determine the year-to-year variations in the sea ice circulation and thickness in the Arctic Basin. Special attention was given to analysing the interannual variability of the sea ice volume in the Basin and the subsequent changes in the export of sea ice from this region into the Greenland Sea via Fram Strait. The comparative role of the sea ice thickness and velocity in the sea ice export anomalies was investigated, and the former was shown to be particularly important during certain large export years (e.g., 1967 and 1989). Mysak and his team at McGill, in collaboration with G. Arfeuille (who will be starting his PhD at the University of Victoria, under the co-supervision of Carmack and Weaver, in January 1999) will compare the interannual variability of Arctic sea ice export, documented in Arfeuille (1999), with results obtained from both the UVic coupled model and the CCCma model (discussed below).
In Mysak and Venegas (1998), a combined complex EOF analysis of 40 years of annual sea ice concentration data and winter sea level pressure (SLP) data revealed the existence of an approximately 10-year climate cycle in the Arctic and subarctic. The cycle is characterized in part by a clockwise propagating signal of sea ice concentration anomalies which start in the western Arctic and end up in the Labrador Sea. Mysak and his team at McGill will carry out an EOF analysis of the model results of Arfeuille (1999) to see whether the sea ice concentration anomalies, in the 41-year period studied, resemble those of Mysak and Venegas (1998). Finally, to delve further into the influence of sea ice export anomalies (on both interannual and interdecadal time scales) on the thermohaline circulation in the North Atlantic, they will replace the sea ice model in the UVic coupled model with the Tremblay-Mysak sea ice model. This will allow one to test whether ice export anomalies of the type simulated by Arfueille (1999) can produce perturbations in the transport of the thermohaline circulation, similar to those found by Mauritzen and Hakkinen (1997), and will allow for an comparison of the effects of using three distinct sea ice models (the two currently available in the UVic coupled model together with the Tremblay and Mysak, 1997, sea ice model).
The CCCma has conducted several multi-century climate simulations whose results will be analysed with regard to a variety of processes, such as: Arctic freshwater export, sea-ice export, and variations in ocean temperature/salinity structure and circulation. Flato will also be intimately involved in the efforts conducted using the UVic coupled model, and is currently conducting a range of sensitivity studies related to the role of sea ice dynamics in climate variability and change. Since it is impossible to conduct the large number of experiments needed to understand the sensitivity of sea ice models and their parameterisations in the CCCma coupled general circulation model, the UVic coupled model (with its simpler representation of the atmosphere) is viewed as an essential tool to further this understanding.
2.2 Is the recent observation of warm, subsurface North Atlantic water intrusion into the Arctic consistent with the response expected from enhanced anthropogenic greenhouse gas forcing? Alternatively, is it simply the Arctic response to low frequency variations in atmospheric circulation?
This question poses two hypotheses to explain the observed subsurface intrusion of warm North Atlantic waters, together with the shift of the Atlantic layer from the Lomonosov to the Mendeleyev Ridges. Our analysis and model experiments will be aimed at testing these hypotheses.
As noted earlier, initial results (Wiebe and Weaver, 1998), which incorporate more realistic representations of ocean mixing into the UVic coupled model revealed the warm, subsurface intrusion of Atlantic waters into the Arctic Ocean under transient CO2-increase experiments. The transient CO2-increase experiments (Boer et al., 1998a,b) conducted using the present version of the CCCma coupled model (Flato et al., 1998) does not exhibit such a result. This is, however, entirely consistent with the results of Wiebe and Weaver (1998) in which we showed that when the usual horizontal/vertical subgrid scale ocean mixing is used, the increased retreat of sea ice (under transient CO2-increase), leads to the exposure of ocean surface waters to the cold polar air which in turn induces unrealistically strong oceanic convection, wiping out any subsurface signature of warm water intrusion.
The CCCma are presently undertaking new transient CO2-increase experiments with a better representation of the ocean which includes the Gent and McWilliams (1990) parameterization (although they have included a small explicit background lateral diffusivity). The results from this experiment will be examined for evidence of subsurface intrusion of warm Atlantic water into the Arctic. In addition, the UVic coupled model which now uses a rotated coordinate grid (allowing for better resolution of the Arctic), and includes sea ice dynamics will be integrated under increasing CO2 radiative forcing to re-examine the subsurface Arctic intrusion issue.
As also noted earlier, an alternative explanation for the observed changes in Arctic water mass structure is provided by low frequency shifts in atmospheric-forced ice and ocean circulation. Simulations using the UVic coupled model driven by long-term (multi-decadal) timeseries of wind fields (based on observations) will be analysed to asses the extent to which observed changes in the Arctic water masses are consistent with atmospheric circulation variations.
In support of these modelling efforts, Carmack will focus on ocean field programmes in the Arctic to aid in ocean/ice model development and validation. These will be no-cost additions to this project as they will be exclusively funded by DFO and will take advantage of his access to Canadian Coast Guard icebreakers and arctic logistic bases. His work will entail: a)- observing change within the Arctic system; b)- measuring the transport of freshwater components through the Canadian Arctic Archipelago; and c)- analysing data related to T/S intrusions as an agent of thermohaline transition. The latter work will be carried out together with a postdoctoral research associate (B. May has expressed an interest in this project and his NSERC Postdoctoral Fellowship application is attached), in order to determine what these imply in terms of model transports and diffusivities
The investigation of the dynamics of thermohaline intrusions in the Arctic Ocean will involve the analysis and interpretation of hydrographic data. The observational results will be used for the development and improvement of theoretical models of thermohaline intrusions. A number of specific questions will be addressed. For example, how do the intrusions change as they advance from the boundaries into the interior of the ocean basins? How is the shape of the intrusions (i.e., saw tooth structure in temperature-salinity space) related to the vertical fluxes of heat and salt by double-diffusion? How can theories of thermohaline intrusions be improved to better match the observations?
This project will be extended to investigate the effects of mixing processes (i.e., thermohaline intrusions and diffusive-convection) on the larger-scale dynamics of the Arctic Ocean. This will involve the development and improvement of mixing parameterizations for inclusion into ocean general circulation models. Model simulations will then be performed with the UVic coupled model to determine how the large-scale dynamics are affected by the smaller-scale mixing processes. A number of specific questions will be addressed: Given a change in the Atlantic water inflow, how long does it take for this signal to propagate through the Arctic system? Given the observed change at the boundaries, are the lateral fluxes associated with thermohaline intrusions large enough to produce the observed warming in the ocean interior? At what rate is the warming transmitted vertically into the upper ocean by diffusive-convection?
2.3 Develop improved parameterisations of processes important for the representation of the Arctic Ocean, its sea ice cover and their coupling for potential use in the CCCma coupled model
The Arctic Ocean, its sea-ice cover, and their coupling is poorly represented in the CCCma coupled model, and our conjecture is that significant improvements can be made with little additional computational cost.
As noted in section 1.2.3, the present representation of the Arctic Ocean, its sea ice cover and their interaction in the CCCma coupled model, as in most current coupled models, is rather crude. We propose to undertake a number of sensitivity experiments with both the UVic coupled model and the CCCma coupled model to determine whether or not potential improvements are climatically important. If so, we envision their incorporation into the next generation CCCma coupled model.
The UVic coupled mode currently allows for two types of sea ice thickness-distribution (ITD). The first is the traditional `fixed-category' representation of the ITD used in Hibler (1980) and Flato and Hibler (1995). The second has a `Lagrangian' delta-function representation similar to those used in one-dimensional model studies of Bjork (1992) and Schramm et al. (1997). The `Lagrangian' delta-function ITD method has the additional option allowing for interior temperature resolution which gives the sea ice thermal inertia (unlike zero-heat capacity models) and more realistic heat conduction. The thermodynamic model also explicitly parameterises the effects of brine pockets as in Maykut and Untersteinter (1971). We use the elastic viscous plastic rheology for ice dynamics from Hunke and Dukowicz (1997), which approximates the Hibler (1979) viscous plastic model on synoptic and longer timescales. We propose to continue development of the ice component of the UVic model so that improvements can be assessed via sensitivity studies and, if appropriate, transferred to the CCCma model. Flato, who is responsible for the sea ice component of the CCCma coupled model will be intimately involved in the sensitively studies conducted with the UVic coupled model.
The sensitivity studies, while directed primarily at assessing the role of various processes which could potentially be included in the CCCma coupled model, also contribute directly to the objectives of the modelling component of ACSYS. Indeed the ACSYS Numerical Experimentation Group (of which Flato is a member and designated chair) is presently focusing on the coordination of thermodynamic sensitivity studies and comparisons of coupled ice-ocean models. The next ACSYS NEG meeting is planned for Victoria in 1999 and provides an opportunity for work done as part of this proposal to be communicated directly to the international modelling community.
The parameterization development activities outlined above are designed to compliment those already underway in the CRYSYS project. The latter is focused primarily on the sea-ice surface energy budget and in particular the role of snow on sea-ice (e.g., Flato and Brown, 1996; Hanesiak et al., 1998). By contrast, the parameterizations to be pursued under the current proposal are aimed at the representation of the thickness distribution and internal thermodynamic processes. As Flato is a co-PI in both projects, these activities will be closely coordinated and mutually beneficial.
The sensitivity to the different ice model components will be examined in the context of the global coupled ice/ocean/atmosphere model. This will include studies on the importance of the ice thickness distribution, the presence of vertical temperature resolution in the ice, and the presence of ice dynamics. The influence of these processes on determining the ice mass balance and ice/ocean/atmosphere exchange will be studied in simulations of present day climate and under transient CO2-increase conditions. This will allow us to examine how improvements in sea ice physics modify the simulated climate and its response to perturbations in the climate system.
Eby and Holloway (1995) proposed a complicated method for ocean grid rotation so that the Arctic could be better resolved in global ocean models. In their approach, a separate North Atlantic model, in which the North Atlantic and the Arctic were rotated so that the Equator of the ocean grid ran through the pole, was connected to a second global (less the North Atlantic) model along the original Atlantic equator. This approach required the continuous switching of boundary conditions between the two models. We have recently implemented a simpler grid rotation which rotates the grid globally and hence locates the singularity of the North Pole over central Greenland (for numerical purposes). Internationally, a grid rotation scheme has already been included in the French LMD global coupled model, while alternative schemes are being assessed for inclusion into the NCAR model. We will conduct sensitivity studies using both unrotated and rotated grid versions of the UVic coupled model to determine whether or not the rotation of the grid allows for a more realistic representation of the Arctic Ocean.
The coupling of a rotated grid ocean model to the CCCma atmospheric model could present some problems concerning the interpolation and conservation of air-sea-ice fluxes of heat and freshwater. We will further examine various techniques of interpolation which conserve both heat and freshwater using the rotated version of the ocean component coupled to an unrotated version of the atmospheric component of the UVic coupled model. In addition, through the careful comparison of experiments which use both rotated atmospheric and oceanic components, we will determine whether the interpolation of air-sea fluxes produces any significant differences. It is only after a careful analysis of these issues that the CCCma would be prepared to consider an ocean grid rotation in future versions of the CCCma coupled model.
Finally, we have already demonstrated the importance of the parameterization of local effects of brine rejection within the context of the UVic coupled model (Duffy et al. 1998). The improvements realized with this parameterisation were so great that we will conduct an experiment with the CCCma coupled model which includes this effect. The substantial reduction of spurious ocean convection may further reduce, if not eliminate, the need for large flux adjustments in the Southern Ocean domain of the CCCma coupled model and will also reduce them substantially near the ice edge in the north Atlantic.
2.4 Annual Milestones
b) Initial analysis of T/S intrusions and their implications for oceanic advection and diffusion.
c) Two manuscripts to be written discussing the role of Arctic Freshwater export (via the ocean and sea ice) on the decadal-interdecadal variability of the North Atlantic Climate.
d) Initial analysis of statistical relationships between Arctic climate indices.
b) Completion of sea ice sensitivity analysis in UVic coupled model.
c) Assessment of the hypotheses regarding the observed subsurface Arctic intrusion of warm North Atlantic waters.
d) Analysis of freshwater export and role of sea ice dynamics from CCCma model.
e) Completion of analysis of statistical relationships between Arctic climate indices.
Transfer of technology to CCCma in terms of:
(ii) Completion of analysis of role of Arctic freshwater forcing on
North Atlantic decadal-interdecadal climate variability.
The team of researchers are well linked with related international research efforts which are summarized below:
Weaver is currently a member of the Steering Committee for the Arctic System Science/Ocean-Atmosphere Interactions component of the NSF Arctic Systems Science Program. He was also a member of the US National Academy of Sciences Committee on Major Ocean Programs and served on the local organizing committee for the WCRP ACSYS 2nd Scientific Conference held in Orcas Island in November 1997. He is currently a Lead Author of Chapter 8 of the IPCC 3rd Scientific Assessment.
Carmack is and has been heavily involved in national and international committees dealing with Arctic climate issues. He was a charter member of the ACSYS Scientific Steering Group, and Chair of the Canadian ACSYS committee which produced the Canadian ACSYS Science Plan. He is a past member of the U.S. Polar research board and RSC Canadian Arctic Panel and has served as Chief Canadian Scientist for the 1994 Arctic Ocean Section and the 1997/98 JOIS/SHEBA Program.
Flato is currently a member and designated Chair of the ACSYS Numerical Experimentation Group and its Sea-Ice Model Intercomparison Project, the PI of the WCRP/CMIP cryosphere subproject, a Co-I in CRYSYS and a member of the NASA/NSIDC scientific advisorary group. In addition, Flato is a co-developer of the CCCma global coupled model.
Mysak is a Fellow of the Royal Society of Canada and Member of the Order of Canada. He is currently involved in the Canadian CRYSYS programme and serves on the Scientific Advisory Committee for the International Arctic Research Centre in Fairbanks, Alaska.
Weaver has already secured matching NSERC Strategic support for the three year initial duration of this project. In Year 3 we anticipate approaching NSERC for a new strategic initiative to support Canadian university participation in the international WCRP ACSYS programme, which will be in place until 2003 as WCRP CLIC (Cryosphere and Climate) programme is developed. In addition, a collaboration with Phil Duffy (Lawrence Livermore National Laboratories) will be continued.
3.2 Relationship to CRYSYS, ACSYS and CLIVAR
The Canadian CRYSYS project encompasses a wide range of cryospheric research, with a particular focus on remote sensing. Although sea-ice is one of the topics being addressed by CRYSYS, the focus is on improved understanding of surface processes, particularly as they affect remote sensing. The underlying Arctic Ocean is not included in CRYSYS. The present proposal compliments CRYSYS by providing an oceanographic perspective and a related focus on internal ice thermodynamic processes, ice dynamics and ice-ocean coupling. As Flato is a co-I involved in several specific collaborative CRYSYS projects, the activities in the present proposal will be closely coordinated and mutually beneficial. Where possible, developments of surface process parameterizations (e.g., snow melt, albedo) obtained as part of CRYSYS activities, will be incorporated into the UVic and CCCma sea-ice component models.
This proposal is directly related to the ocean and sea ice components of the ACSYS modelling programme (WCRP 1994). Our sea ice modelling sensitivity analysis directly addresses most of the questions on page 49 of WCRP (1994) in the Arctic Numerical Experimentation Programme. Our attempts to develop better ice-ocean coupling techniques is also central to the Ocean Process Models (page 51) of the ACSYS Science plan. Finally, the contribution by Carmack will directly address (page 56) the data requirements for Arctic model initialization, forcing and validation.
As noted by WCRP (1994) on page 57: "Two sets of problems concerning the role of the Arctic in the climate system can be addressed with coupled atmosphere-ocean circulation models: the influence of the Arctic on natural climate variations and the impact on the response of the climate system to global warming. These model investigations will be promoted and co-ordinated by the WCRP research programme on climate variability (CLIVAR). The contribution of ACSYS to these activities should be to ensure that polar processes are properly represented in the coupled model. This includes the delivery of an optimised dynamic-thermodynamic sea-ice model and accurate ocean physics, tailored for the Arctic area.". Thus our two central scientific questions are key to one of the central goals of ACSYS and CLIVAR. Indeed, as a group, we will be positioned to make a major contribution to WCRP activities in this area. Given Canada's abundance of Arctic territory and its attempts to reinforce sovereignty of this area, we believe it is in the nation's interest to undertake research in this region.
Finally, Flato has recently been tasked with organizing the next ACSYS NEG meeting in Victoria in 1999. As part of international ACSYS NEG meetings, it is normal for local sea-ice modelling efforts to be presented before the committee. As such, in 1999 we will have an opportunity for direct contribution to international ACSYS initiatives.
4 Training, management and technology transfer with the CCCma
4.1 Training of Highly Qualified Personnel
The training of highly qualified personnel is an integral part of this project and the majority of the requested funds would be used to support salary costs for 3 post-docs and 4 graduate students, whose supervision will be distributed amongst the investigators. These young researchers would be exposed to a dynamic research environment which transcends traditional disciplines. Carmack is committed to providing opportunities for young investigators to participate in Arctic field experiments aboard Canadian Coast Guard icebreakers in order to provide them with an understanding of the nature of the Arctic and its environment, as well as the meaning and limits of data collected in remote regions.
We have recently contacted and set up an advisory panel for this Arctic Initiative. Francis Zwiers (Chief, CCCma, AES) would act as chair of the panel and provide a link with the CCCma modelling activities; Barry Goodison (Chief, Climate Processes Research Division, AES) would provide linkages with CRYSYS; Ken Denman (DFO) would provide important links to DFO; Jacques Derome (McGill) would provide links with the CRN Climate Variability work as well as represent the interests of the University community as a whole. This advisory panel would oversee our research and ensure that it contributes directly and successfully to AES/CCCma coupled modelling activities, while also providing important links to CRYSYS, DFO and the Climate Variability Node.
The Secretariat for this group will be located in room 296b of the Gordon Head Complex. Wanda Lewis is in place to handle all correspondence associated with this project as well as distributing funds and maintaining the budget. The group will meet twice a year. One of these meetings will be at the annual CMOS congress with the additional meeting in Victoria. We have already requested a special Arctic session at the 2000 CMOS conference in Victoria, B.C.
4.3 Technology Transfer
During the course of the above collaborative research, various model enhancements and process parameterisations will necessarily be developed, particularly with regard to representation of Arctic Ocean circulation, water mass formation and mixing, ice-ocean coupling, and sea-ice processes. These will be available for use in the CCCma global model. The method for technology transfer to the CCCma will come directly via Greg Flato who is a co-PI of our project.
In order to facilitate close collaboration and smooth technology transfer, one of the post-doctoral fellows supported under this proposal would work directly with the CCCma and their coupled model. This will involve analysing model results, implementing improved parameterizations of ice-ocean processes and conducting process studies. By providing explicitly for the development and study of such parameterizations in the CCCma model, in collaboration with the broader range of process studies to be undertaken within the proposed Node, technological developments will be directly available within the CCCma modelling environment.
5 Proposed Budget
The proposed budget shown below is for each of the 1999-00, 2000-01
and 2001-02 fiscal years.
|Postdoctoral Research Associates||
5.1 Budget Details:
1) A total of 3 postdoctoral research associates would be supported at a rate of 40,000 per annum each (including benefits) throughout the initial three years of funding. The postdoctoral research associates would be expected to interact with all investigators in this project and their supervision will be distributed amongst the PIs. Cecilia Bitz (UVic) will remain on for the first phase of this project and Brian May (Dalhousie) has expressed an interest to join the group (see attached NSERC Postdoctoral Fellowship application). While we are not able to offer postdoctoral positions until funding is confirmed, there have been numerous inquiries from potential applicants from the best national and international institutions, so that we do not expect there to be any problem filling these positions. For example, Alexandre Perchat from Toulouse in France has applied to work on this project as part of his French Military Service requirement.
2) In the first year four graduate students would be supported. Three graduate students at UVic have already been accepted although funding for them (with the exception of Roth who holds an NSERC) has not been secured: 1) Michael Roth (PhD); 2) Linda Waterman (MSc); Gilles Arfeuille (PhD). A fifth graduate student would be accepted in Year 2 to stagger the student involvement at different stages of the research program since new scientific questions that arise in Year 1 will need to be addressed in Year 2. The funding is at the NSERC rate of $16,500 pa. Based on past experience, the PIs feel confident that excellent scholarship students will also be attracted to join the project. All UVic PIs would be on the supervisory committees for each student.
3) Travel expense for students, research associates and university investigators to the annual CMOS meeting each year. Travel is also to be used for Mysak and some of his team to participate in the annual meeting in Victoria.
4) Publication charges and other operating costs for university participants. The costs have been increased in Year 2 in anticipation of substantially increased publication charges.
5) Partial salary for secretary (W. Lewis) to work at Secretariat as well as communication charges (courier, phone fax etc.)
Aagaard K, Fahrbach E, Meinke J, Swift JH. 1991. Saline outflow from the Arctic Ocean: Its contribution to the deep waters of the Greenland, Norwegian and Iceland Seas. J. Geophys. Res. 96:20433-20441.
Arfeuille G. 1999. Modelling the interannual variability of the Arctic sea ice cover. MSc thesis, McGill Univ., Montréal, Québec.
Bjork G. 1992. On the response of the equilibrium thickness distribution of sea ice to export, mechanical deformation, and thermal forcing with application to the Arctic Ocean. J. Geophys. Res. 97:11,287-11,298.
Boer GJ, Flato GM, Reader MC, Ramsden D. 1998. A transient climate change simulation with historical and projected greenhouse gas and aerosol forcing: experimental design and comparison with the instrumental record for the 20th century. Clim. Dyn., submitted.
Boer GJ, Flato GM, Ramsden D. 1998. A transient climate change simulation with historical and projected greenhouse gas and aerosol forcing: projected climate for the 21st century. Clim. Dyn., submitted.
Caldeira K, Duffy PB. 1998. Sensitivity of simulated CFC-11 distributions in a global ocean model to the treatment of salt rejected during sea-ice formation. Geophys. Res. Let. 25:1003-1006,.
Carmack EC, Killworth P. 1978. Formation and inter-leaving of abyssal water masses off Wilkes Land, Antarctica. Deep-Sea Research, 25:357-369.
Carmack EC, Macdonald RW, Perkin RG, McLaughlin FA, Pearson RJ. 1995. Evidence for warming of Atlantic water in the southern Canadian Basin of the Arctic Ocean: Results from the Larsen-93 expedition. Geophys. Res. Let. 22:1061-1064.
Carmack EC, Aagaard K, Swift JH, MacDonald RW, McLaughlin FA, Jones EP, Perkin RG, Smith JN, Ellis KM, Killius LR. 1998a. Changes in temperature and tracer distributions within the Arctic Ocean: results from the 1994 Arctic Ocean section. Deep-Sea Res. 44:1487-1502
Carmack EC, Matear R, Perkin RG, McLaughlin FA, MacDonald RW, Aagaard K. 1998b. Combined influence of geothermal heating and biological regeneration on the relict deep waters of the southern Canadian Basin. J. Geophys. Res. submitted.
Cattle H. 1985. Diverting Soviet rivers: Some possible repercussions for the Arctic Ocean. Polar Rec. 22:485- 498.
Delworth T, Manabe S, Stouffer RJ. 1993. Interdecadal variations of the thermohaline circulation in a coupled ocean-atmosphere model. J. Clim. 6:1993-2011.
Delworth T, Manabe S, Stouffer RJ. 1997. Multidecadal climate variability in the Greenland sea and surrounding regions: a coupled model simulation. Geophys. Res. Let. 24:257-260.
Denbo D, Skyllingstad ED. 1996. An ocean large-eddy simulation model with application to deep convection in the Greenland Sea. J. Geophys. Res. 101:1095-1110.
Deser C, Blackmon ML. 1993. Surface climate variations over the North Atlantic Ocean during winter: 1900- 1989. J. Clim. 6:1743-1753.
Dickson RR, Meincke J, Malmberg SA, Lee AJ. 1988. The "great salinity anomaly" in the Northern North Atlantic 1968-1982. Prog. Oceanogr. 20:103-151.
Dickson RR, Lazier J, Meincke J, Rhines P, Swift J. 1996. Long-term coordinated changes in the convective activity of the North Atlantic. Prog. Oceanogr. 38:205-239.
Duffy PB, Caldeira K. 1997. Sensitivity of simulated salinity in a three-dimensional ocean model to upper-ocean transport of salt from sea-ice formation. Geophys. Res. Let. 24:1323-1326.
Duffy PB, Eby M, Weaver AJ. 1998. Effects of sinking of salt rejected during formation of sea ice on results of a global ocean-atmosphere-sea ice climate model. Geophys. Res. Let. submitted.
Eby M, Holloway G. 1994. Grid transform for incorporating the Arctic in a global ocean model. Clim. Dyn. 10:241-247.
Flato GM, Brown RD. 1996. Variability and climate sensitivity of landfast Arctic sea ice. J. Geophys. Res., 101:25,767-25,777.
Flato GM, Hibler III WD. 1992. Modeling pack ice as a cavitating fluid. J. Phys. Oceanogr. 22:626-651.
Flato GM, Hibler III WD. 1995. Ridging and strength in modeling the thickness distribution of Arctic Sea ice. J. Geophys. Res., 100:18,611-18,626.
Flato GM, Boer GJ, Lee WG, McFarlane NA, Ramsden D, Reader MC, Weaver AJ. 1998. The Canadian Centre for Climate Modelling and Analysis Global Coupled Model and its Climate. Clim. Dyn., submitted.
Gent PR, McWilliams JC. 1990. Isopycnal mixing in ocean general circulation models. J. Phys. Oceanogr. 20:150-155.
Grotefendt K, Logemann K, Quadfasel D, Ronski S. 1998: Is the Arctic Ocean warming? J. Geophys. Res., 103:27,679-27,687.
Hkkinen S. 1993. An Arctic source for the Great Salinity Anomaly: A simulation of the Arctic ice-ocean system for 1955-1975. J. Geophys. Res. 98:16397-16410.
Hanesiak JM, Barber DG, Flato GM. 1998. The role of diurnal processes in the seasonal evolution of sea ice and its snow cover. J. Geophys. Res. in press.
Hibler III WD. 1979. A dynamic thermodynamic sea ice model. J. Phys. Oceanogr. 9:817-846.
Hibler III WD. 1980. Modeling a variable thickness sea ice cover. Mon. Weather Rev. 108:1943-1973.
Hunke EC, Dukowicz JK. 1997. An elastic-viscous-plastic model for sea ice dynamics. J. Phys. Oceanogr. 27:1849-1867.
Kushnir Y. 1994. Interdecadal variability in the Northern Hemisphere sea surface temperature and associated atmospheric conditions. J. Clim. 7:141-157.
Lazier JRN. 1980. Oceanographic conditions at Ocean Weather Ship Bravo, 1964-1974. Atmos-Ocean, 18:227-238.
Mauritzen C, Hkkinen S. 1997. Influence of sea ice on the thermohaline circulation in the Arctic-North Atlantic Ocean. Geophys. Res. Let. 24:3257-3260.
Maykut GA, Untersteinter N. 1971.Some results from a time dependent thermodynamic model of sea ice. J. Geophys. Res. 76:1550-1575.
McFarlane NA, Boer GJ, Blanchet J-P, Lazare, M. 1992. The Canadian Climate Centre second generation general circulation model and its equilibrium climate. J. Clim. 5:1013-1044.
McLaughlin FA, Carmack EC, Mcdonald RW, Bishop JKB. 1996. Physical and geochemical properties across the Atlantic/Pacific water mass front in the southern Canadian Basin. J. Geophys. Res. 101:1183-1197.
Mysak, LA and SA Venegas, 1998, Decadal climate oscillations in the Arctic: A new feedback loop for atmosphere-ice-ocean interactions. Geophys. Res. Lett. 25: 3607-3610
Mysak LA, Manak DK, Marsden RF. 1990. Sea-ice anomalies observed in the Greenland and Labrador Seas during 1901-1984 and their relation to an interdecadal Arctic climate cycle. Clim. Dyn. 5:111-133.
Proshutinsky AY, Johnson MA. 1997. Two circulation regimes of the wind-driven Arctic Ocean. J. Geophys. Res. 102:12493-12514.
Reverdin G, Cayan D, Kushnir Y. 1997. Decadal variability of hydrography in the upper northern North Atlantic in 1948-1990. J. Geophys. Res. 102:8505-8531.
Schramm JL, Holland MM, Curry JA, Ebert EE. 1997.Modeling the thermodynamics of a sea ice thickness distribution 1. Sensitivity to ice thickness resolution. J. Geophys. Res. 102:23,079-23,091.
Semtner AJ. 1976. A model for the thermodynamic growth of sea ice in numerical investigations of climate. Phys. Oceanogr. 6:379-389 .
Steele M, Thomas D, Rothrock D, Martin S. 1996. A simple model study of the Arctic Ocean freshwater balance, 1979-1985. J. Geophys. Res. 101:20833-20848.
Swift JH, Jones EP, Aagaard K, Carmack EC, Hingston M, MacDonald RW, McLaughlin FA, Perkin RG. 1998. Waters of the Makarov and Canada basins. Deep-Sea Res. 44:1503-1529.
Thompson DJ, Wallace JM. 1998. The Arctic Oscillation signature in the wintertime geopotential height and temperature fields. Geophys. Res. Let. 25:1297-1300.
Tremblay L-B, Mysak LA. 1997. Modeling sea ice as a granular material, including the dilatancy effect. J. Phys. Oceanogr. 27:2342-2360.
Tremblay L-B, Mysak LA. 1998. On the origin and evolution of sea-ice anomalies in the Beaufort-Chukchi Sea. Climate Dynamics 14:451-460.
WCRP 1994. Arctic Climate System Study (ACSYS) Initial Implementation Plan. WCRP report 85, WMO/TD- No. 627, Geneva, 66pp.
Walsh JE, Chapman WL. 1990. Arctic contribution to upper-ocean variability in the North Atlantic. J. Clim. 3:1462-1473.
Weaver AJ, Valcke S. 1998. On the variability of the thermohaline circulation in the coupled model. J. Clim. 11:759-767.
Wiebe EC, Weaver AJ. 1998. On the sensitivity of global warming experiments to the parameterisation of sub- grid scale ocean mixing. Clim. Dyn. submitted.
Wohlleben TMH, Weaver AJ. 1995. Interdecadal climate variability in the subpolar North Atlantic. Clim. Dyn. 11:459-467.