No description available.

In this thesis, we formulate a new 3-dimensional planetary geostrophic (PG) ocean general circulation model in spherical coordinates for use in ocean-climate studies. The model equations consist of full prognostic temperature and salinity equations and diagnostic momentum equations. The model is verified by comparison with results obtained by the well known Bryan-Cox model at coarse resolution. Extensive process studies are carried out to determine the roles played by various processes in determining the thermocline structure and the thermohaline circulation, especially the role of convective overturning. A secondary circulation theory which treats the thermohaline circulation as a geostrophic flow instead of frictional current is proposed, and the implications of the results to 2-dimensional latitude/depth models are discussed. / The model is also used to examine the stability and variability of the thermohaline circulation, especially the role played by air-sea heat flux at the surface. We show that the large reduction in the surface heat flux under mixed boundary conditions (restoring on temperature and flux on salinity) is essential for the occurrence of the "polar halocline catastrophe" (F. Bryan. 1986). Replacing the infinite heat capacity atmosphere implied by mixed boundary conditions with an atmosphere that can adjust thermally to the ocean, the amount of the heat flux reduction is less and thus the polar halocline catastrophe is less likely to occur; when it does occur, it is less severe. This is demonstrated by coupling our model to a simple zero heat capacity atmosphere (Schopf, 1983). Instead of a total collapse of the thermohaline circulation, the results show a decaying oscillation of 20 years period. / The ocean model is next coupled to a thermodynamic sea ice model to examine the effect of sea ice on the stability and variability of the thermohaline circulation. A robust 17-year period oscillation is obtained, and a new mechanism involving a feedback between ice cover and temperature is proposed. Conceptual models are formulated for further interpretation of the results, and to compare the thermal insulation and salinity rejection effects due to an anomalous ice cover, on the density of the surface water. The heat budget required for ice formation leads to a constraint on the convective overturning, and hence the transfer of heat and salt from the deep ocean to the bottom of the ice. As a result, the thermal insulation effect is dominant, at least for the case of annual mean surface forcing examined here.

An uncoupled, modified Hibler ice model has been applied to the Gulf of St. Lawrence for three different winters of varying severity, in order to examine interannual sea ice variability. The simulation was initialized with observed November sea surface temperatures, and forced by weekly geostrophic winds, monthly averaged meteorological data and model geostrophic surface currents. / Results showed a general correlation with observations, reproducing differences in the sea ice cover between the years chosen. Neglecting oceanic effects caused excessive ice formation in the northwest Gulf and produced discrepancies between the observed and modelled ice edge. / Sensitivity studies revealed a high sensitivity to variations in both the forcing fields and the model free parameters. Further modelling studies must include a coupled ocean component, and force the ice component with weekly meteorological data to improve the accuracy of the prediction.

<p>Western Long Island Sound (WLIS) bottom waters experience low dissolved oxygen (DO) levels in the summertime. The seasonal hypoxia in the WLIS motivated the development of a coupled biogeochemistry/hydrodynamic model named the System Wide Eutrophication Model (SWEM). A critical assessment of the SWEM model, however, reveals that vertical transport rates are grossly underestimated, indicating a need for data-based estimates of these rates. </p><p> I present a novel approach to estimate vertical mixing coefficients from time-series of measurements made at two or more depths by measuring the attenuation and phase lag of scalar signals generated periodically at the surface as they propagate downwards and estimate mean summertime depth-averaged downward fluxes of DO and heat of 14&plusmn;4 &mu;M day^&ndash;1 and 17&plusmn;10 Wm^&ndash;3, respectively. In order to assess the importance of horizontal transport in the bottom waters, I present an analysis of time-series of moored temperature, DO, and current observations in the hypoxic area of Long Island Sound and estimate mean near-bottom along-channel flux differences of DO and heat as 4&plusmn;6 &mu;M day^&ndash;1 and &ndash;5&plusmn;6 Wm^&ndash;3, respectively. I conclude that vertical transport forms the bulk of the physical supply of both DO and heat to the hypoxic zone. </p><p> When WLIS moored instrument records are examined, it is evident that near-bottom increases in DO and heat and a decrease in salt occur during the middle of the flood tide; an analysis of water mass signatures indicates that the transport involved is vertical and not horizontal. Temperature data from a thermistor string deployed in the WLIS for 16 days in August 2009 clearly shows internal waves and a pycnocline depression of approximately 25% of the water depth occurring at mid-flood. Near-bottom internal wave energy is correlated with near-bottom DO and temperature changes at both supertidal and subtidal scales, and I conclude that internal mechanisms are potentially important to vertical transport in the WLIS region. </p>

No description available.

An experimental investigation was performed to determine the drift velocities of oil slicks and the pile up of oil layers against barriers in the presence of water waves. Other investigations of the drift response were made by substituting thin plastic sheets in place of the oil slicks. Parameters were developed which govern these movements. Discrepancies in the literature between drift of oil lenses and the "Stokes' drift" were solved. It was discovered that the boundary layer between the drifting oil slick and wave motion has turbulent character and is not laminar as was derived from theory and assumed to be the case.

The role of the mechanical energy budget is analyzed in relation to the small viscosity behaviour of an idealized model of the wind-driven circulation. This is addressed theoretically and numerically in part one. In part two the connection to the real ocean is made through an analysis of the energy source with real data. / The theoretical arguments based on the QG energy equation show that, with the assumption that the maximum velocities occur at inertial length scales or smaller, a Sverdrup interior is consistent with the small Rossby number assumption only when the frictional parameters exceed critical values. For frictional parameters smaller than these values, valid solutions must decrease the energy source. This is possible for non-Sverdrup solutions since the energy source is dependent on the solution. / The numerical study is focused on insensitivity to frictional parameters in the nonlinear Stommel-Munk model. Dependence of the multiple solutions of the steady state model on the boundary layer Reynolds number, Re, are investigated by varying the eddy viscosity for fixed wind forcing. An important finding is the tendency to decrease the energy source for solutions that are nonsymmetric about the centre latitude. Antisymmetric solutions display the opposite behaviour, and diverge more quickly with increasing Re. Also interesting is the tendency for the total energy and transport to become less sensitive to eddy viscosity with increasing Re. The robustness of the results to dynamic boundary condition, symmetry and strength of wind stress, time dependence and bottom friction are tested. A significant, though not surprising, result is that the no slip condition leads to high Rossby number in the boundary layer at much lower Re than for the free slip condition. / It is demonstrated that recent advances in altimetry measurements allow for a reasonable estimate of the rate of mechanical energy transfer from the atmospheric winds to the surface geostrophic velocity integrated over large regions. (Local values are highly uncertain due to the very large uncertainty in the marine geoid undulation field.) The feasibility and methodology of a quantitative uncertainty estimate is also demonstrated. The estimate of the energy source term allows comparison of the theoretical and numerical results with the real ocean.

Circulation and mixing processes in the St. Lawrence Estuary are examined using both newly acquired and historical observations, as well as numerical modeling. In the Summer of 1997 a field experiment was undertaken to verify the prediction of a three-dimensional numerical model on the occurrences of shear instabilities within the upper estuary. The instabilities were found as predicted and documented from acoustic imaging, current profiler and density measurements. The instabilities, as in the case of Kelvin-Helmholtz instabilities, first develop in the form of wavelike disturbances before they break. The unstable waves have wavelength of around 140--150 m and a vertical extent of between 10 and 25 m. It is shown that the modeled gradient Richardson numbers near the pycnocline are reasonably well reproduced and appear to be useful for the prediction of flow instabilities in such a complex environment. / The seasonal variability of the general circulation and mixing is examined with a newly developed laterally averaged numerical model. The model takes into account tidal forcing and a realistic topography while forced with a large seasonal change in the river discharge. The model is able to reproduce important processes associated with the tidal forcing and mean conditions (density field, residual circulation). The numerical results show that mixing in the St. Lawrence Estuary is not evenly distributed. Mixing is found to be large near the sill at the head of the Laurentian Channel, as previously thought, as well our new results show that equivalent levels of mixing are also found near English Bank and near Ile-aux-Coudres. During the spring freshet, when the river discharge is 40% higher than normal, the limit of salt intrusion is moved seaward by approximately 12 km, the residual circulation is intensified by approximately 10%, and the turbulent vertical mass flux increases throughout owing to greater availability of freshwater. The results also suggest that during high runoff periods, the vertical stability of the water column increases in the upper estuary. / The model is also used to assess the impacts of climate change on the estuarine dynamics. (Abstract shortened by UMI.)

A simple box model of a a high-latitude two-layer ocean, first developed by Martinson et al., 1981, is applied to four Arctic regions connected together: the Greenland Sea, the Norwegian Sea, the Arctic Ocean, and the Greenland Gyre. The latter, in fact, is a small convective region embedded in the northwest corner of the Norwegian Sea region. The model for each region consists of a thermodynamic ice layer that covers two layers of salty water which can, under specific conditions, become statically unstable and hence create a state of active overturning. The system is forced by monthly mean atmospheric temperatures in the four regions, as well as by continental runoffs and by inflows from adjacent oceans. / The model predicts the ice thickness, and the temperature and salinity of the water in the upper layer for the four regions. Also determined are the water temperature and salinity for the lower layer in the Arctic Ocean box. The convective state of any given region, i.e., whether it is in an active overturning mode or not, is also obtained continuously. / The different output variables of the model, which are the response to climatological forcing conditions, compare favourably with observed data. In this control run, the Arctic Ocean region is characterized by continuous ice cover only during winter, and the Norwegian Sea region never forms an ice cover. Another feature of the model is the winter time occurrence of convective overturning in the upper 200 m in the Greenland Gyre region. (Abstract shortened by UMI.)

The winter watermass properties and circulation of the Gulf of St. Lawrence and St. Lawrence Estuary are examined using a combination of new wintertime observations and numerical modelling. Ten moorings were deployed over the winter 2002/3, providing the first continuous observations of the upper 200m of the water column in winter. A harmonic analysis of current data in the estuary confirms the existence of a semi-diurnal internal Poincare wave. The analysis also provides strong evidence of a neap-tide intensification of the estuarine circulation. The cold intermediate layer (CIL) in the estuary is seen to be renewed abruptly in late March, with the advection of a wedge of near-freezing water from the gulf. / An eddy-permitting sea ice-ocean model is used with the observations to quantify the formation and circulation of the CIL. As a first step, a two-equation second moment turbulence closure parameterization is implemented to better account for vertical mixing processes. The model sensitivity to this parameterization is tested and differences of several salinity units are found in the surface layer throughout the southern half of the gulf. The model reproduces the strong spring renewal event of the CIL into the estuary, and helps to explain the role of local dynamics, including the coupled circulation in the northwest gulf, in controlling the exchange processes at depth. The model results suggest the existence of a separated Gaspe Current that persists for more than a month. A significant correlation is found between the inflow through the Strait of Belle Isle and outflow through Cabot Strait. In order to isolate the effects of this inflow on the CIL heat content, we examine a sensitivity experiment in which the Strait of Belle Isle is closed. This simulation shows that the inflow has a relatively small effect on the total CIL heat budget. With the Strait of Belle Isle closed, the outflow of CIL through Cabot Strait is reduced by 61%. It is estimated that only 64% of CIL inflow through the Strait of Belle Isle in fall and winter contributes directly to the heat content of CIL present in the gulf.