Tsunami exposure estimation with land-cover data: Oregon and the Cascadia subduction zone Nathan Wood * Western Geographic Science Center, US Geological Survey,1300 SE Cardinal Court, Building 10, Vancouver, WA 98683, USA Keywords: Vulnerability Tsunami Oregon Cascadia C-CAP LandSat Thematic Mapper Land cover abstract A Cascadia subduction-zone earthquake has the potential to generate tsunami waves which would impact more than 1000 km of coastline on the west coast of the United States and Canada. Although the predictable extent of tsunami inundation is similar for low-lying land throughout the region, human use of tsunami-prone land varies, creating variations in community exposure and potential impacts. To better understand such variations, land-cover information derived from midresolution remotely-sensed imagery (e.g., 30-m-resolution Landsat Thematic Mapper imagery) was coupled with tsunami- hazard information to describe tsunami-prone land along the Oregon coast. Land-cover data suggest that 95% of the tsunami-prone land in Oregon is undeveloped and is primarily wetlands and unconsolidated shores. Based on Spearman rank correlation coefficients (r s ), correlative relationships are strong and statistically significant (p < 0.05) between city-level estimates of the amount of land-cover pixels classified as developed (impervious cover greater than 20%) and the amount of various societal assets, including residential and employee populations, homes, businesses, and tax-parcel values. Community exposure to tsunami hazards, described here by the amount and relative percentage of developed land in tsunami-prone areas, varies considerably among the 26 communities of the study area, and these variations relate to city size. Correlative rela- tionships are strong and significant (p < 0.05) for community exposure rankings based on land-cover data and those based on aggregated socioeconomic data. In the absence of socioeconomic data or community-based knowledge, the integration of hazards infor- mation and land-cover information derived from midresolution remotely-sensed imagery to estimate community exposure may be a useful first step in understanding variations in community vulnerability to regional hazards. Published by Elsevier Ltd. Introduction Disaster risk is a function of the hazards (natural or technological) and the societal vulnerability of a place and is influenced by human decisions and policies made before, during, and after an extreme event. Vulnerability, defined as the characteristics of a system that increase the potential for hazard-related losses, is determined by how a community occupies and functions in predicted hazard zones. Extreme natural events may cause damage to individual assets and communities, but the choices a society makes with regards to land use and its political, economic, and social systems prior to an event set the stage for these losses by creating unsafe conditions (Mileti, 1999; Wisner, Blaikie, Cannon, & Davis, 2004). * Tel.: þ1 360 993 8951; fax: þ1 360 993 8980. E-mail address: nwood@usgs.gov. Contents lists available at ScienceDirect Applied Geography journal homepage: www.elsevier.com/locate/apgeog 0143-6228/$ – see front matter Published by Elsevier Ltd. doi:10.1016/j.apgeog.2008.08.009 Applied Geography 29 (2009) 158–170