Digital Elevation Modelling for natural hazard risk assessment

Introduction The application of geographic information systems (GISs) to natural hazard risk management is a relatively new and emerging science. Coppock (1995) notes that GIS has made a contribution to various facets of natural hazard risk management since risk is a multi-dimensional and multidisciplinary phenomenon, which have a spatial component, whatever their initial focus. Hence, the success of GIS implementation for natural hazard risk reduction can be contingent upon the availability of spatial data. The digital elevation model (DEM) is a key form of spatial data . A DEM is a spatially referenced continuous surface representing the topography of an area. A DEM surface may be represented by a grid, where each cell in the grid indicates a ground elevation. DEMs are usually stored as computer files and are key to most GIS spatial databases. Ground elevations modelled by DEMs are important for a number of natural hazard risk management applications including flood inundation modelling (storm tide and riverine), landslide susceptibility modelling and bushfire risk mapping. Risk management personnel may not be directly responsible for DEM creation but may need to be familiar with the general concepts and key terms. Concepts include DEM accuracy, resolution, spatial extent, currency and fitness-of-use. Presenting the basic theory of DEM creation, and DEM characteristics may help risk managers understand and better utilise this important, and increasingly common, model of spatial data. Spatial data issues associated with DEMs including error, accuracy, resolution and scale are also critical to other spatial data used in GIS-based natural hazard risk management. Data themes can include physical hazard zonations, building and lifeline databases, and census data. Examining these issues should also be a first step when organisations integrate GIS with risk management processes such as AS/NZS 4360 (1995). For example, the first stage of AS/NZS 4360 is titled ‘Establishing the Context’ and focuses on institutional issues surrounding the implementation of the standard. This stage is important since it identifies the aims, objectives and GIS modelling limitations of the project. Similarly, issues of error, accuracy, resolution, scale and spatial extent should be considered. DEMs are examined as a casestudy of these issues. Cairns in Far North Queensland is the study site, and reference is made to storm tide inundation modelling. Issues associated with riverine flooding are similar, and the general concepts are applicable to other natural hazard risk assessments. In Queensland, DEMs have recently been used for developing inundation evacuation plans, and therefore, issues of accuracy and error are critical for decision making. Issues to be discussed also include, DEM availability in Australia, surface interpolation procedures and DEM input data sources. Since risk results derived from DEMs are as accurate as the input data used to derive them, attention is given to DEM error assessment. End-users of DEMs perceive levels of accuracy in the elevation data that are optimistic. Exploratory spatial data error analysis is shown to be a critical aspect of DEM creation, and for determining the fitness-of-use of spatial data.