High-resolution DEM generated from LiDAR data for water resource management

Terrain patterns play an important role in determining the nature of water resources and related hydrological modelling. Digital Elevation Models (DEMs), offering an efficient way to represent ground surface, allow automated direct extraction of hydrological features (Garbrecht and Martz, 1999), thus bringing advantages in terms of processing efficiency, cost effectiveness, and accuracy assessment, compared with traditional methods based on topographic maps, field surveys, or photographic interpretations. However, researchers have found that DEM quality and resolution affect the accuracy of any extracted hydrological features (Kenward et al., 2000). Therefore, DEM quality and resolution must be specified according to the nature and application of the hydrological features. The most commonly used DEM in Victoria, Australia is Vicmap Elevation delivered by the Land Victoria, Department of Sustainability and Environment. It was produced by using elevation data mainly derived from existing contour map at a scale of 1:25,000 and digital stereo capture, providing a state-wide terrain surface representation with a horizontal resolution of 20 metres. The claimed standard deviations, vertical and horizontal, are 5 metres and 10 metres respectively (Land- Victoria, 2002). In worst case, horizontal errors could be up to ±30m. Although high resolution stereo aerial photos provide a potential way to generate high resolution DEMs, under the limitations of currently used technologies by prevalent commercial photogrammetry software, only DSMs (Digital Surface Models) other than DEMs can be directly generated. Manual removal of the nonground data so that the DSM is transformed into a DEM is time consuming. Therefore, using stereo aerial photos to produce DEM with currently available techniques is not an accurate and costeffective method. Light Detection and Ranging (LiDAR) data covering 6900 km² of the Corangamite Catchment area of Victoria were collected over the period 19 July 2003 to 10 August 2003. It will be used to support a series of salinity and water management projects for the Corangamite Catchment Management Authority (CCMA). The DEM derived from the LiDAR data has a vertical accuracy of 0.5m and a horizontal accuracy of 1.5m. The high quality DEM leads to derive much detailed terrain and hydrological attributes with high accuracy. Available data sources of DEMs in a catchment management area were evaluated in this study, including the Vicmap DEM, a DEM generated from stereo aerial photos, and LiDAR-derived DEM. LiDAR technology and LiDAR derived DEM were described. In order to assess the capability of LiDAR-derived DEM for improving the quality of extracted hydrological features, sub-catchment boundaries and drainage networks were generated from the Vicmap DEM and the LiDAR-derived DEM. Results were compared and analysed in terms of accuracy and resolution of DEMs. Elevation differences between Vicmap and LiDAR-derived DEMs are significant, up to 65m in some areas. Subcatchment boundaries derived from these two DEMs are also quite different. In spite of using same resolution for the Vicmap DEM and the LiDARderived DEM, high accuracy LiDAR-derived DEM gave a detailed delineation of sub-catchment. Compared with results derived from LiDAR DEM, the drainage networks derived from Vicmap DEM do not give a detailed description, and even lead to discrepancies in some areas. It is demonstrated that a LiDAR-derived DEM with high accuracy and high resolution offers the capability of improving the quality of hydrological features extracted from DEMs.