Large lakes, with horizontal dimensions of several hundred km and depths of several hundred meters, have their own interesting dynamics. In winter they are isothermal, but by late summer a strong thermocline has developed. The uppermost 5 m may be isothermal at 20 deg C, the depths below 15 m may be isothermal at 5 deg C, and there can be a strong thermocline with a change of 15 C degrees over 10 m. This is a stronger thermoclines than is usually found in the ocean, and it is a significant challenge for an ocean model to represent the dynamics.

Two numerical ocean models, the Princeton Ocean Model (POM) and the DIECAST ocean model are applied to the model the wind forced upwelling dynamics in large lakes with this thermocline. The models are first applied to a circular basin of diameter 100 km and of maximum depth 100 m. Both constant depth and parabolic bathymetries are used. After the wind stops, for light upwelling, and internal Kelvin wave progresses around the basin. For strong upwelling when the thermocline breakes the surface, the upwelling front progression resembles an internal bore. Finally, the models are applied to a realistic case of strong upwelling in Lake Michigan.

Existing two dimensional turbulent boundary layer modeling parameters fail to adequately model the physics of complex three dimensional turbulent boundary layer flow. This failure is illustrated through several lags in turbulence parameters as the flow develops. The turbulent shear stress angle lags the flow gradient angle and the behavior of the Reynolds shear stresses lag that of the mean flow. A minimum set of equations to account for the three dimensionality of turbulent flows of engineering interest is proposed. Eleven sets of experimental data for complex three dimensional turbulent boundary layer flows of engineering interest were examined. From these data, new algebraic turbulence modeling parameters were developed to support the six equations mentioned above. The resulting parameters display a good degree of universality for the variety of experimental geometries that were examined.

The simulations show that the tropical warming and the cooling are controlled dynamically by the ocean currents in the Pacific. The warming in the cold tongue is caused by the reduction of cold advection off the equator in the surface ocean, and by the reduction of upwelling cold advection in the thermocline. The warming in the surface warm pool is caused by the reduction of the westward cold advection from the cold tongue. The cooling in the thermocline warm pool is caused by the reduction of the equatorward warm advection. These oceanic advections are mainly controlled by the reduction of the meridional divergent (convergent) current off (into) the equator, the westward equatorial current, and upwelling current near the Peru coast. These currents, in turn, are controlled dynamically by the reduction of atmospheric trade winds. Therefore, the tropical temperature trend is controlled dynamically by the reduction of trade winds in the Pacific. The thermodynamic effect through the surface net heat flux between the ocean-atmosphere interface seems not to be a critical factor to affect the tropical ocean temperature trend. In the North Pacific mid-latitudes, however, the thermodynamic effect seems more important to the surface western and central ocean. But for the eastern and lower ocean, the dynamic effect on the oceanic temperature trend is still dominant.

1. The beginning of re-intensification occurred when the upper level potential vorticity (PV) anomaly intensified and moved southward approaching the lower level PV anomaly.

2. Ex-Earl moved into an area of favorable extratropical cyclogenesis downstream of an approaching mid-level trough(corresponding to the upper level PV anomaly).

3. The system approached a pre-existing region of low level PV anomaly

Ensemble forecasts from 17 members (9 spectral models including control run, 8 grid model ) were used to analyze the performance of the CMC ensemble forecast members. Approximately half of the 17 ensemble forecast members (initialized at 00 UTC 03 September and 00 UTC 04 September) successfully traced the decay of ex-Earl and the subsequent re-intensification after 00 UTC 05 September. Members, which failed to track the remains of Earl because they couldnUt sustain the lower level PV anomaly associated with ex-Earl, developed a new (extratropical) cyclone north of the analyzed track. Many members, among those which had "good" tracks, did not deepen the low significantly (below 980 hPa). This owes to underestimation of the warm core thermal anomaly ,the upper level PV anomaly, the cold-air intrusion after 00 UTC 05 September. Re-intensification was best traced by members which successfully captured the above features.

The physical variations associated with these alternating periods of dominance by the two slope water components have been accompanied by systematic variations of dissolved oxygen and nutrients: Labrador Slope Water leads to low nitrate and high oxygen concentrations relative to the times when WSW prevailed. The significant changes of the dissolved oxygen and nitrate concentrations are coherent with those in temperature and salinity, in many ways similar to an episode that occurred in the mid 1960s. The relationship between the hydrographic properties, the nutrients and the dissolved oxygen for Scotian Shelf waters will be discussed, and the role of the Labrador Current in these changes will be examined.