AbstractsGeography &GIS

Modeling mass care resource provision post hurricane

by Tammy Marie Poitras Muhs




Institution: University of Central Florida
Department:
Degree: PhD
Year: 2011
Keywords: Dissertations, Academic  – Sciences;Sciences  – Dissertations, Academic;Decision trees  – Florida;Food relief  – Florida;Hurricanes  – Florida
Record ID: 1893863
Full text PDF: http://digital.library.ucf.edu/cdm/ref/collection/ETD/id/4810


Abstract

Determining the amount of resources needed, specifically food and water, following a hurricane is not a straightforward task. Through this research effort, an estimating tool was developed that takes into account key demographic and evacuation behavioral effects, as well as hurricane storm specifics to estimate the number of meals required for the first fourteen days following a hurricane making landfall in the State of Florida. The Excel based estimating tool was created using data collected from four hurricanes making landfall in Florida during 2004-2005. The underlying model used in the tool is a Regression Decision Tree with predictor variables including direct impact, poverty level, and hurricane impact score. The hurricane impact score is a hurricane classification system resulting from this research that includes hurricane category, intensity, wind field size, and landfall location. The direct path of a hurricane, a higher than average proportion of residents below the poverty level, and the hurricane impact score were all found to have an effect on the number of meals required during the first fourteen days following a hurricane making landfall in the State of Florida.