Year of Award

2009

Document Type

Thesis

Degree Type

Master of Science (MS)

Degree Name

Geography (Cartography and GIS Option)

Department or School/College

Department of Geography

Committee Chair

David Shively

Commitee Members

Anna Klene, Tyron Venn

Keywords

Fire Management and Planning, Wildland Fire, Wildland-Urban Interface (WUI)

Abstract

Research indicates firefighting costs in the wildland-urban interface (WUI) are highly correlated with the number of homes threatened by wildfire. Therefore, knowing the location of structures is paramount for planners and fire managers attempting to reduce the threats posed to structures by wildfire, and for the attainment of land management goals and objectives for reducing hazardous fuels surrounding them. Yet, no national-level structure location dataset exists. Previous attempts, such as the SILVIS Lab’s product, to predict structure location and the extent of the WUI have relied on Census block-level data. While urban Census blocks are generally small in area, those corresponding to sparsely settled areas may contain many square miles of territory. Rural Census blocks can contain small clusters of homes in one area, but any large uninhabited regions in the remaining area can result in an average structure density that is lower than the federal WUI criteria. Additionally, the designation of an entire large Census block as WUI, when only a small portion of the block contains houses, simultaneously causes both an underestimation in the number of Census blocks that contain areas meeting the density criterion and overestimates the extent of the WUI. LandScan USA, created by researchers at the Oak Ridge National Laboratory, estimates the population distribution for the United States using Census blocklevel housing data and additional inputs including transportation infrastructure, land cover, elevation, and cultural criterion, such as recreational features, retail establishments, employment, and educational locations. In order to test the accuracy of the LandScan USA dataset for predicting structure locations in the WUI, this study measures the spatial coincidence between this dataset and county-level cadastral data in northwest Montana and compares those results to the SILVIS data. Additionally, each dataset was buffered 1½-miles and compared for spatial coincidence to measure the potential of the LandScan USA data to predict the location of the WUI. The findings reveal that the LandScan USA data do not adequately predict the location of structures for use in wildfire management and planning. However, this research does indicate that further research into LandScan USA’s ability to demarcate the WUI is justified.

Download

Share

COinS
 

© Copyright 2009 Jeffrey Daniel Kaiden