A study on rasterization method of sediment‐related disaster risk based on topographic and geological theme maps

Masayuki MATSUDA and Hiroaki NAKAYA

Abstract

Sediment‐related disasters (hereby SRD in this study) are triggered by inciting causes as heavy rain and earthquakes where lands are affected by primary factors as steep topography and geological conditions. Much study has been conducted on primary factors so far. The results, however, are not fully applied to disaster mitigation measures such as alerts and warnings due, in part, to overwhelming amounts of studies and their apparent inconsistency. Related suitable theme maps are handful if coverage of the entire national land is taken into account. Principal candidates among these are: landslide distribution maps based on air photo analysis by the National Institute of Disaster Prevention Science and Technology, SRD warning zone maps, and deep‐seated failure frequency maps by Sabo administration of the Ministry of Land, Infrastructure, Transport and Tourism. These maps are digitized, but not rasterized to be used readily via geographical information systems. We have rasterized these maps in a common way. 24‐year SRD records administered by Sabo administration are used to examine the quality of rasterization and data matching. The data matching shows that SRD warning zone maps are most fitted to general SRDs with hit rates of 7.36%, 3times higher than when primary causes are ignored completely, and missed rates of 10.91%. On the other hand, deep‐seated failure frequency maps combined with landslide distribution maps are most fitted to large‐scale sediment‐related disasters have 0.36% and 68.93% respectively. The results indicate that primary factors reflected to theme maps, even with coarse 1‐km spatial resolution, can provide useful information for disaster prevention. The rasterization method proposed in this study, when applied cautiously, is regarded conductive to bring parts of past studies into practices.

Key words

GIS, hazard mapping, rasterization