Winter/Spring 2005

This presentation will begin with an introduction to the MSC's data assimilation and prediction infrastructure which is the foundation on which this activity is built. It will give an overview of the evolution of data assimilation within MSC and provide an update on recent developments in variational data assimilation. Examples of current and proposed research projects using these methods in collaborative projects at Dalhousie University will also be mentioned.

Vortex dipoles are formed in a viscous fluid when a force is applied locally to some volume of fluid. If the force acts impulsively, a translating vortex dipole is generated. If the force starts at t = 0 and then acts continuously a starting jet with a dipole at its front is generated. Solutions for unsteady viscous flows generated by the action of continuous or impulsive localized forces are obtained in Oseen approximation. The solutions are compared with direct numerical simulations of vortex dipoles as well as with laboratory experiments. The comparison shows good quantitative agreement in both cases.

A physical problem where the localized force acting continuously on fluid is placed in a uniform stream is equivalent to a problem of a fixed body in a uniform stream, while the couple of forces acting in opposite direction are equivalent to the problem of self propelled body moving at constant velocity through a fluid. The solutions for the two dimensional far-field wake are obtained in both cases. At a certain Reynolds number, wakes become unstable and form vortex streets. New series of high-resolution 2D numerical simulations is performed to study the characteristics of the wakes including the shedding frequency for a wide range of control parameters such as translational velocity, magnitude and spatial extent of a localized force. The results of numerical experiments of unsteady wake flow show existence of a great variety of flow regimes and are in good qualitative agreement with laboratory experiments.

Freshening imposed on the deep intermediate layers by extreme convection in the early 1990's penetrated to the deep layers. We discuss the nature of the time lag between the intermediate and deep freshening peaks, which can be partially attributed to a spreading of LSW across the subpolar gyre of the North Atlantic. Combining BIO observations in the Labrador Sea with measurements to the east of Greenland (collected by European teams), we obtain a fairly accurate estimate of the LSW spreading time from its source to different sub-polar basins.

The unprecedented warming of the Labrador Sea is partitioned into a change in the water masses and a structural shift in the water mass distribution. The major change is marked by the appearance of a large volume of warm and salty water from the Irminger Sea. This water spread westward from its usual location along the eastern boundary, reaching the midpoint of the Labrador Basin, and filling the whole eastern part of the basin between 100 m and 800 m. Due to this expansion of the Irminger water, the water at 700 m in the eastern part of the Labrador Sea became about 0.6 warmer and 0.05 saltier between 2003 and 2004. Recent observations reveal fresh and cold anomalies that are propagating from the Denmark Strait, entering the deep and bottom layers of the Labrador Sea, and spreading westward. The freshest event was recorded in the early spring of 2004 in the northern Irminger Sea and is expected to arrive to the abyss of the Labrador Sea in 2005. The continuation of Labrador Sea monitoring in 2005 and beyond, combined with observations at the various sources of its deep and bottom waters, will provide the most reliable indication of the operation and regimes of the grand ocean conveyor.

To assess the role of SSTs on the decay rate of Hurricane Juan during its approach to Nova Scotia, we use a mesoscale model of the atmosphere called the Mesoscale Compressible Community (MC2) model. The model consists of a fixed and nested 3 km grid driven by a coarser 12 km grid, and is initiated using a synthetic hurricane vortex constructed from observational information such as storm size and intensity, thus giving a decent representation of the real storm. Multiple experiments were conducted for two SST configurations: 1) observed SST and 2) climatology SST representative of late September. Results from the 3 km simulations initiated approximately 1 day prior to landfall (and others initiated 2 days prior to landfall) indicated that the intensity of Hurricane Juan.s maximum surface winds near landfall were 5-8 m s^-1 (~ ± 1.5 m s^-1) weaker in the climatology-SST simulations. The destructiveness of these winds was only 50-70% of the observed (warmer than normal) SST runs. This difference suggests that damage from Hurricane Juan could have been much less had SSTs been near normal! Model performance is measured using surface weather data, as well as data collected from a research aircraft that flew into the storm just prior to landfall.

I will also discuss how the model system is being applied to other tropical cyclone events that affected eastern Canada and discuss how the system can be used to improve forecasting of these storms in the future.