Network Analysis of Hurricanes Affecting the United States
Hurricanes affecting the United States cause severe damage and kill people. The risk of future hurricane activity along the coast is the subject of much scientific and public interest. While considerable work had been done to understand the occurrence of hurricanes along the coast, much less has bee...
Other Authors: | , , , , , , |
---|---|
Format: | Text |
Language: | English |
Published: |
Florida State University
|
Subjects: | |
Online Access: | http://purl.flvc.org/fsu/fd/FSU_migr_etd-4430 http://fsu.digital.flvc.org/islandora/object/fsu%3A182558/datastream/TN/view/Network%20Analysis%20of%20Hurricanes%20Affecting%20the%20United%20States.jpg |
Summary: | Hurricanes affecting the United States cause severe damage and kill people. The risk of future hurricane activity along the coast is the subject of much scientific and public interest. While considerable work had been done to understand the occurrence of hurricanes along the coast, much less has been done to examine the inter-relationships among the hurricanes. This dissertation concerns the relationships of hurricanes affecting the United States using methods of network analysis. Network analysis has been used in a variety of fields to study relational data, but has yet to be used in the study of hurricane climatology. The present work is largely expository introducing network analysis and showing how it can be applied to possibly better understanding regional hurricane activity as well as hurricane activity overtime. The research is divided into two cases. The first case consists of networks developed based on the relationships of spatial locations of landfalls and the second part consists of networks developed based on the relationships of the temporal occurrence of landfalls. In the first case, the network links coastal locations (termed nodes) with particular hurricanes (termed links). The topology of the network is examined using local and global metrics. Results show that certain regions of the coast (like Louisiana) have high hurricane occurrence rates, but not necessarily high values of network connectivity. Low values of connectivity indicate that hurricanes affecting Louisiana tend not to affect other regions. Regions with the highest values of connectivity include southwest Florida, northwest Florida, and North Carolina. Virginia which has a relatively low occurrence rate is well-positioned in the network having a relatively high value of betweenness. In the second case, the year-to-year variation in U.S. hurricane activity is examined by extending the ideas and concepts of network analysis for time series data. The "visibility" network link years experiencing a hurricane landfall with other hurricane landfall years ``visible" to each other through time. The topology of the visibility network is examined using local and global metrics. Results show that overall the visibility network has few years with many lines of visibility, therefore, many linkages to other years. Years with high hurricane count have more visibility in the network than those years that have less storms. Among years with high counts the years that are surrounded (before and after) with years of low counts will have greater visibility. The years 1886, 1893, 1955 and 2004 are highly visible in the network of U.S. hurricanes. A year is more central if it is a link in more visibility chains between other years in the network. Six conditional networks are constructed for the spatial and temporal networks based on years of below and above average values of important climate variables. Significant differences in the connectivity of the network are noted for different phases of the El Nino-Southern Oscillation. During El Nino years, when the equatorial waters of the eastern Pacific are warm, there tends to be shearing winds and subsidence over large portions of the North Atlantic where hurricanes form. These conditions lead to fewer hurricanes affecting the United States. More work is needed to better understand the details of how climate influences the network of landfalls. The scientific merit of the research is a better understanding of the relationships in the regional risk of hurricane activity. The broader impacts are an introduction of network analysis to hurricane climatology. A Dissertation submitted to the Department of Geography in partial fulfillment of the requirements for the degree of Doctor of Philosophy. Spring Semester, 2009. March 16, 2009. Robert Har, Hurricane Climatology, Hurricanes Includes bibliographical references. James B. Elsner, Professor Directing Dissertation; Robert Hart, Outside Committee Member; J. A. Stallins, Committee Member; Thomas Jagger, Committee Member. |
---|