Reconstructing Digital Terrain Models from ArcticDEM and WorldView-2 Imagery in Livengood, Alaska

ArcticDEM provides the public with an unprecedented opportunity to access very high-spatial resolution digital elevation models (DEMs) covering pan-Arctic surfaces. As it is generated from stereo-pairs of optical satellite imagery, represents a mixture surface model (DSM) over non-ground areas and terrain (DTM) at bare grounds. Reconstructing DTM thus needed in studies requiring ground elevation, such as modeling hydrological processes, tracking change dynamics, estimating vegetation canopy height associated forest attributes. Here we proposed automated approach for two steps: (1) identifying pixels WorldView-2 imagery using Gaussian (GMM) local refinement by morphological operation, (2) generating continuous ArcticDEMs locations spatial interpolation methods (ordinary kriging (OK) natural neighbor (NN)). We evaluated our method three forested study sites characterized different cover topographic conditions Livengood, Alaska, where airborne lidar data available validation. Our results demonstrate that identification can effectively identify much lower root mean square errors (RMSEs) (<0.35 m) reference than comparative state-of-the-art approaches; NN performs more robustly OK; (3) DTMs GMM-based masks decrease RMSEs 0.648 m, 1.677 0.521 m Site-1, Site-2, Site-3, respectively. This viable means deriving high-resolution will be great value focusing on Arctic ecosystems, earth processes.