Thesis (Ph. D.)--UCLA, 2009. / Vita. Includes bibliographical references (leaves 153-161).

The components of the northern hemisphere cryosphere and their relationship to Atlantic tropical cyclone activity are investigated in this study. Multiple ordinary least-squares regression with a stepwise selection procedure is used to develop a new statistical forecasting scheme for 13 seasonal tropical cyclone parameters at four lead times for the period 1980-2010. Sea ice area and sea ice extent in 10 geographic regions, snow cover extent in three geographic regions and five indices reflecting major modes of climate variability were analyzed as possible predictors. Three model groups, based on predictors, were constructed and evaluated: 1) only climate mode predictors, 2) only cryosphere predictors, and 3) both cryosphere and climate mode predictors. Models using only climate mode predictors showed poor predictability of the tropical cyclone parameters across all four lead times while the models using only cryosphere predictors and those using both sets of predictors showed improved predictability. Baffin Bay and Hudson Bay sea ice area were found to be the most significant predictors, exhibiting an inverse relationship with overall tropical cyclone activity. The developed models were also compared to current operational statistical models of tropical cyclone activity. While the operational models were generally more skillful, June hindcasts of major hurricanes outperformed the operational models by as much as 20%.

The seasonal pattern of ice motion in response to wind forcing and potential consequences to upwelling on the Mackenzie Shelf are considered using satellite-derived ice motion data from the National Snow and Ice Data Center and the NCEP 10 m wind data. The frequency of strong upwelling-favorable alongshore ice motion is high in early winter (November and December) compared to middle and late winter (January to May).For periods when the alongshore component of the wind is upwelling-favorable, the ratio of ice drift divided by wind speed on the Mackenzie Shelf is 0.024 in November and0.008 in March; we conjecture that this ratio decreases as winter progresses because the internal ice stress becomes stronger as both ice thickness and ice concentration increase. This constitutes a possible 10-fold decrease in the seasonal transmission of wind stress to the underlying water from November to March. This ratio in May (0.015) is higher than that in March. We suggest that it is because the internal ice stress becomes weaker as ice concentration decreases on the Mackenzie Shelf in May. Hence, under the same wind forcing, the potential for winter upwelling on Mackenzie Shelf may be enhanced if climate warming results in reduced ice thickness and/or ice concentration. Numerical model results show that the stress on the shelf could be reduced because of the internal ice stress from the pack ice over the deep ocean when the ice moves like a rigid body. We found that the model results are not realistic when the ice strength is 5,000 Nm-2. When the ice strength is 27,500 Nm-2, the model results are more realistic. / Science, Faculty of / Earth, Ocean and Atmospheric Sciences, Department of / Graduate

No description available.

No description available.

No description available.

Climate modelling and high-latitude marine navigation require improved information on sea-ice floe extents and dynamics. New satellite sensors provide raw data of this nature but the volume of information makes human analysis impractical. To address this problem, a software system for automatic tracking of sea-ice floes in satellite imagery has been designed and evaluated. Using a recurrent neural network model, experiments were conducted to discover suitable design parameters. Simulated imagery time-sequences of increasing complexity were produced to train successive models. The networks produced were evaluated based on performance and reliability. A small-scale working system, able to map multiple input features in image sequences to Cartesian coordinates, was produced. Results show that a recurrent neural network is suitable for the tracking task and has advantages in robustness and speed over other approaches. Recurrency (feedback) was found to be crucial in achieving good performance.

Thesis (Ph.D.). / Written for the Dept. of Atmospheric and Oceanic Sciences. Title from title page of PDF (viewed 2008/02/12). Includes bibliographical references.

Thesis (Ph.D.)--University of Delaware, 2008. / Principal faculty advisors: Chandra Kambhamettu, Dept. of Computer & Information Sciences; and Cathleen A. Geiger, Dept. of Geography. Includes bibliographical references.

Thesis (Ph. D.)--University of Washington, 1996. / Vita. Includes bibliographical references (leaves [133]-142).