Abstract
The tracking of hurricanes, largely controlled by the organization of the presiding pressure systems, determines whether or not any given hurricane will strike a coastline. Some of the climatic influences, such as the North Atlantic Oscillation, show annual- or decadalvariability. This means that particular locations will have typical hurricane tracks that may vary with the climate. Therefore, it makes physical sense to summarize large sets of hurricane tracks by creating an average track. A hurricane climatology describes the typical hurricane to affect a location. This dissertation proposes expanding the hurricane climatology by adding a spatial dimension in the form of an average track. This is referred to as a spatial hurricane climatology. Since a hurricane track is a polyline, the construction of a spatial hurricane climatology requires averaging spatial polyline data. The technique introduced in this dissertation uses distance maps to average a set of polylines. Three applications of spatial hurricane climatologies are also detailed in this work. First they are used to construct historical hurricane chronologies. This has the possibility of providing an additional 150 years of hurricane data, providing a glimpse into hurricanes prior to the American industrial revolution. The second application is a risk analysis of local-scale hurricane winds. The technique uses statistics of past hurricanes and places them in a deterministic model. This can be performed for any coastal area, and provides wind gusts and economic loss estimations for a once-in-100-year event. Because the statistics are easy to manipulate, this allows for simple analysis of the affects of climate change. This is done as the final application of the technique. These are only a few examples of the uses of spatial hurricane climatologies, and the ideas presented in this research provide a basis for future studies on spatial hurricane patterns, as well as the analysis of spatial polyline data in general.
|
