Accuracy of the LiDAR-derived DEM in dense shrub areas in mountainous NW US

Digital elevation model (DEM) is one of the most important information that can be derived from the airborne LiDAR data. However in areas covered by dense vegetation, the chance of laser passing through the canopy, hitting the ground and back to the receiver is limited, thus it is difficult to achieve accurate DEM in regions with dense vegetative covers (trees, shrubs and grasses). We considered the challenges in deriving accurate terrain elevation information from LiDAR data: identifying areas covered with dense vegetation, separating LiDAR point clouds into ground and non-ground returns, interpolating to create raster surface models. A case study in the semi-arid mountainous area of the Northwest US shows that dense shrub caused about 20 cm DEM error in vertical accuracy, in line with similar study results of LiDAR derived DEM from dense grass areas. Terrain, cover type and LiDAR acquisition parameters (scan or incidence angle, point density) can also affect accuracy.