Below I summarize the progress on the research funded through the NSERC/WOCE Collaborative Special Project Grant. I have also enclosed a list of all publications (since 1992) which were supported through these funds. This progress report and the original WOCE proposal may be viewed on the world wide web at:
4.1 Decadal variability in a coupled OGCM/EMBM
A. Fanning, a PhD student, has developed and utilized an atmospheric energy-moisture balance model (EMBM, see Fanning and Weaver, 1995) coupled to an ocean general circulation model (the GFDL-MOM model, Pacanowski et al., 1993) in a series of experiments conducted in a single hemisphere (60deg. x 60deg.) basin, driven by zonally uniform wind stress and solar insolation forcing. The study examines the coupled system's sensitivity to resolution and oceanic parameters. We have already completed four experiments ranging from 4deg. x 4deg. resolution to 0.5deg. x 0.5deg. resolution, with appropriate horizontal viscosities, and diffusivities in each case. Poleward heat transport is shown to significantly increase from coarse to finer resolution, although the coupled atmosphere-ocean model results confirm that the time-variant (eddy) component of poleward heat transport counteracts increases in the time mean flow as suggested by Bryan (1986), yielding indistinguishable changes in heat transport from the moderately coarse (1deg.) to higher (0.5deg.) resolution. The net planetary heat transport, and atmospheric heat transport, also appear to converge as resolution is increased. To interpret these results, the heat transport has been decomposed into its baroclinic overturning (related to the meridional overturning and Ekman transports), barotropic gyre (that in the horizontal plane) and baroclinic gyre (the remainder) components. To further assess the results, we have repeated these same experiments under restoring boundary conditions (to apparent temperatures and salinities diagnosed from the 4deg. x 4deg. equilibrium state following Haney, 1971) to elucidate the differences between heat transport in the coupled versus uncoupled model. Currently we are extending the resolution studies (coupled and uncoupled models) to 1/4deg. x 1/4deg. as well as 1/5deg. x 1/5deg. resolution.
Of particular importance is that spontaneous decadal variability (period ~13 years) is found to exist in the 0.5 x 0.5 resolution case (in both the coupled and uncoupled model), with poleward heat transport changing by up to one third of the total oceanic heat transport over one oscillation in the thermohaline circulation. The oscillation is best described as an advective-convective mechanism, linked to the turning on and shutting off of convection in the northwest corner of the model domain. We find the variability is strongly linked to the value of the horizontal diffusivity utilized in the model. Increasing the diffusivity from 200 m2/s to 500 m2/s is enough to destroy the variability, while decreasing the diffusivity from 500 m2/s to 200 m2/s (in the 1deg. x 1deg. case) is capable of inducing the variability. The results of this research are currently being written up for publication.
In another project, A. Fanning has coupled the atmospheric energy moisture balance model to the realistic geometry global ocean general circulation model described by (Weaver and Hughes, 1995). To allow for the effects of ice in the polar regions, the coupled model contains a simple thermodynamic ice model (Semtner, 1976) which includes heat insulation, brine rejection, as well as the ice-albedo feedback effect. Precipitation over land is returned to the oceans through a series of river drainage basins as in Weaver and Hughes (1995), and we allow for the water vapor-planetary longwave feedback by employing the parameterization of Thomson and Warren (1982). The ocean model begins from the near equilibrium state of Weaver and Hughes (1995), while the atmospheric model begins from zero heat and moisture content. We have now integrated the coupled model to equilibrium. Our equilibrium climate, without flux adjustments, is very reasonable (see Fanning and Weaver, 1995).
Finally, Trevor Murdock has recently began his MSc. He will be investigating the effects on ocean circulation and global climate of the closing of ocean gateways (the isthmus of Panama and Drake passage) with this same coupled energy-moisture balance model. Mr. Murdock is currently taking courses but has become familiar with the coupled model.
4.2 Decadal variability in OGCMs with various subgrid-scale boundary layer dissipation parameterizations
T. Huck, a visiting PhD student from France, has developed a hierarchy of simplified thermohaline circulation models in order to study the effect of the momentum dissipation parameterizations on the large-scale ocean features. The main model is based on the Planetary Geostrophic equations, in a coarse resolution Cartesian beta-plane ocean; the choice of momentum dissipation includes the traditional Laplacian viscosity, biharmonic dissipation, and linear Rayleigh friction with different options to solve for the non-hydrostatic boundary layers (Salmon, 1986). In addition, the GFDL-MOM code has been utilized with the same geometry to provide a reference.
A first set of experiments has been done under an atmospheric forcing limited to restoring boundary conditions for the surface layer temperatures. This leads to an equilibrium state after some 3000 years. Planetary Geostrophic dynamics prove to yield a satisfying framework for the mid-latitude basin studied here, as the results with horizontal Laplacian viscosity compare very well with the GFDL-MOM model case under the same conditions. Results indicate also that the vertical momentum dissipation has a very limited influence on the equilibrium temperatures and velocities. The Laplacian viscosity at coarse-resolution produces unexpectedly strong vertical velocities, especially along the boundaries. Around the thermocline depth, these spurious boundary vertical transports are comparable to the total interior upwelling. A better agreement between downwelling vertical velocities and convection is found with linear friction, either using a vorticity closure for the tangential velocities along the lateral walls (Winton, 1993), or relaxing the hydrostatic approximation via a vertical friction (linear with the vertical velocity [Salmon, 1986]). In these cases, the vertical velocity fields are much smoother, not so strongly perturbed near the boundaries: the deep water is slightly colder (0.1deg. C), and the polar heat transport 8% larger, although the meridional overturning streamfunction is much weaker, dropping from 15 to 9 Sverdrups. This is not a consequence of the 'no normal flow' boundary conditions, as the use of free-slip boundary conditions with the Laplacian or biharmonic viscosity does not resolve these problems. This comparison will be reported in a paper to be submitted by the end of the year.
The second objective of these models concerns decadal oscillations and their driving mechanisms. Oscillations have occurred in these thermally-only-driven experiments, under restoring boundary conditions for temperature with long restoring time scales, or more readily with constant heat flux. A wide range of tests have shown that convection is not necessary to the oscillation's mechanism, but that a critical damping factor is the horizontal eddy-diffusivity. This variability also occurs on an f-plane, where their amplitude grows with the Coriolis parameter. The geographical distribution of the surface heat flux is of primary importance. The heat flux fields deduced from restoring experiments never lead to oscillations, as opposed to longitudinally uniform fluxes (Greatbatch and Zhang, 1995). Research is in progress to clarify the driving mechanism by simplifying the oscillation to its necessary elements and comparing its behaviour according to the momentum dissipation and boundary conditions.
4.3 Variability as a function of mean climatic state
Tertia Hughes recently spent one week at GFDL in Princeton acquiring the GFDL coupled climate model for use on my local work station cluster. We now have this model up and running and will use it to investigate questions concerning the existence of climate variability in the coupled climate system and how it varies as the mean climatic state changes. Sophie Valcke (an NSERC postdoctoral fellow) and Sheng Zhang (a research associate) will work on various aspects of this project.
4.4 Global ocean modelling
Development of the global ocean model continued during this period, with an investigation of the seasonal cycle, and its relevance to the annually-averaged water mass properties and overturning circulation. In a first phase, a seasonal cycle was added only to the temperature and salinity restoring boundary conditions, and the North Atlantic overturning was found to increase by a few Sverdrups. A seasonal cycle on the winds was then also added, changing the annual mean overturning again by a small amount. Preliminary investigations suggested that variability of the Mediterranean outflow temperature and salinity was an important influence, however parallel experiments in an alternate geometry with no Mediterranean Sea showed that other factors could also be contributing. In particular, large changes in convective patterns near the western boundaries of the oceans under seasonal winds gave rise to concern whether the on/off character of the standard convective parameterization was appropriate for representing seasonal excursions of the mixed layer. Annual and seasonal experiments with the isopycnal and Gent-McWilliams formulations (which can replace convection in regions of steeply-sloping isopycnals) were therefore commenced, but have not yet been fully analyzed due to an unusual numerical problem which is currently being investigated by both Daniel Robitaille and Michael Eby in idealized box models.
Another aspect of the global modelling project has been to assess the equilibrium circulation produced under observed heat and freshwater fluxes and surface wind stresses from the recent da Silva et al. (1994) climatology. Some preliminary treatment was needed to make the fluxes suitable for forcing an ocean model, since the open water bias of the observations caused heat losses in the Arctic to be overestimated. However, even after adjusting the fluxes to be consistent with the climatological ice cover, excessive cooling of the northern oceans led to the dominance of northern deep water formation while southern sinking was essentially suppressed. Resolution is thought to be a major issue here, as suggested by Tziperman and Bryan (1993) and Schiller (1995). It remains an open research question how to improve parameterizations in coarse resolution ocean models enough to resolve this incompatibility. One issue which we are exploring is the suggestion of enhanced vertical mixing in western boundary currents compared to the quiescent interior. For this purpose, we have constructed a stripped-down version of the Bryan-Cox primitive equations model which has no barotropic mode, a linear equation of state, fixed surface temperatures/ densities, and spatially-varying vertical diffusivity, and are currently running tests of the sensitivity of the overturning circulation to inhomogeneous kv in a simple flat-bottomed basin geometry.
4.5 CFCs and global ocean models
A global version of the GFDL model was run using three different sub-grid mixing schemes: lateral/vertical, an isopycnal mixing scheme (Redi, 1982; Cox, 1987), and the Gent and McWilliams (1990) parameterization. The results from these three runs were analyzed by using CFC-11 as a time-dependent passive tracer, and by comparing with observations. the Gent and McWilliams parameterization improves the CFC-11 distributions when compared to both of the other schemes, especially in the southern ocean, where the "bolus" transport canceled the mean advection of tracers and hence caused the Deacon Cell to disappear. These results suggest that the asymmetric response found in CO2 increase experiments, whereby the climate over the Southern Ocean does not warm as much as in the northern hemisphere, may be an artifact of the particular sub-grid scale mixing schemes used.
The sensitivity of the Gent and McWilliams parameterization was also evaluated by introducing an additional stability dependent diapycnal mixing, following Tandon and Garrett (1995). This additional mixing does not affect the surface buoyancy flux, but do affect the circulation of passive tracers in the deep ocean. These results will be analyzed by looking at the role of boundary layer versus interior mixing.
4.6 Finite element modelling
This four year project involves a systematic procedure for model development. The last progress report outlined the main accomplishments to date. Since that time the barotropic finite element model has been tested in its global version. Time-dependent and nonlinear terms have also been included. In addition a detailed diagnostic calculation of the annual mean, annual cycle and interpentadal variability of the North Pacific Ocean circulation has been completed (Myers and Weaver, 1995).
Bryan, K., 1986: Poleward buoyancy transport in the ocean and mesoscale eddies. J. Phys. Oceanogr., 16, 927-933.
Cox, M.D., 1987: Isopycnal Diffusion in a Z-Coordinate Model. Ocean Modelling, 74, 1-5.
da Silva, A.M., C.C. Young and S. Levitus, 1994: Atlas of surface marine data 1994. Volume 1: Algorithms and procedures. NOAA Atlas NESDIS 6. U.S. Government Printing Office, Washington, D.C.
Fanning, A.F. and A.J. Weaver, 1995: An atmospheric energy moisture-balance model for use in climate studies. J. Geophys. Res., submitted.
Gent, P. R. and J. C. McWilliams, 1990: Isopycnal Mixing in Ocean Circulation Models, J. Phys. Oceanogr., 20, 150-155.
Greatbatch, R.J., and S. Zhang, 1995: An interdecadal oscillation in an idealized ocean basin forced by constant heat flux. J. Climate, 8, 81-91.
Haney, R.L., 1971: Surface thermal boundary condition for ocean circulation models. J. Phys. Oceanogr., 1, 241-248.
Myers, P.G., and A.J. Weaver, 1995: On the circulation of the North Pacific Ocean: Climatology, seasonal cycle and interpentadal variability, Prog. Oceanogr., submitted.
Pacanowski, R., K. Dixon and A. Rosati, 1993: The GFDL Modular Ocean Model Users Guide. GFDL Ocean Group Technical Report #2, 46pp.
Redi, M.H., 1982: Oceanic Isopycnal Mixing by Coordinate Rotation J. Phys. Oceanogr, 12, 1154-1158. Salmon, R., 1986: A simplified linear ocean circulation theory. J. Mar. Res., 44, 695-711.
Schiller, A., 1995: The mean circulation of the Atlantic Ocean north of 30 S determined with the adjoint method applied to an ocean general circulation model. J. Mar. Res., 53, 453-497.
Semtner, A.J., 1976: A model for the thermodynamic growth of sea ice in numerical investigations of climate. J. Phys. Oceanogr., 6, 379-389.
Tandon, A., and C. Garrett, 1995: On as Recent Parameterization of Mesoscale Eddies. J. Phys. Oceanogr, submitted.
Thompson, S.L., and S.G. Warren, 1982: Parameterization of outgoing infrared radiation derived from detailed radiative calculations. J. Atmos. Sci., 39, 2667-2680.
Tziperman, E. and K. Bryan, 1993: Estimating global air-sea fluxes from surface properties and from climatological flux data using an oceanic general circulation model. J. Geophys. Res., 98, 22629-22644.
Weaver, A.J., and T.M.C. Hughes, 1995: On the incompatibility of ocean and atmosphere models and the need for flux adjustments. Clim. Dyn., in press.
Winton, M., 1993: Numerical Investigations of Steady and Oscillating Thermohaline Circulations. PhD thesis, University of Washington. 155 p.