ANALYSIS OF LIDAR POINT DATA AND DERIVED ELEVATION MODELS FOR MAPPING AND CHARACTERIZING BOULDERY AND BLOCKY LANDFORMS IN THE FORESTED ALLEGHENY MOUNTAINS OF WEST VIRGINIA

Bouldery and blocky landforms dominate Allegheny Mountain landscapes underlain by Pottsville Group sandstones in West Virginia. Most of these bouldery landforms are difficult to map using conventional methods because of a widespread forest cover, locally dominated by non-deciduous species, such as Rhododendron maximum, Picea rubens, and Tsuga canadensis. A recent study of the viability of LiDAR-derived elevation data in accurately mapping and characterizing bouldery landforms showed both disappointing output from ground-returns data and considerable potential in extracting landforms using last-returns data, supplemented by other data such as LiDAR return intensity and aerial imagery. With initial parameters set to avoid the misidentification of trees or buildings as ground surface, a TerraScan software ground-returns classification removed most bouldery and blocky features less than 5 to 10 m in width. Isolated blocks and boulders were dropped almost universally from the TerraScan ground-returns data, whereas rock cities with little separation of blocks were portrayed more accurately. Hillshade imagery derived from classified ground-returns data was inadequate for depicting bouldery landforms. Manipulation of last-returns data was labor intensive, but generally captured bouldery and blocky landforms better than ground-return algorithms. Last-returns elevation and LiDAR intensity data delineated bouldery landforms in open areas with good success. Identifying and describing boulders and blocks under a tree canopy required reliable ground classification of LiDAR points, which was obtained in this study through iteration of Prologic LiDAR Explorer classifications with varied input parameters. Individual blocks and boulders were not all discernable, but areas of irregular rugged topography were better represented using Prologic LiDAR Explorer than by the default TerraScan software ground data classification. Index overlay for likelihood of presence of bouldery landforms using supervised classified aerial imagery and LiDAR-derived parameters in a raster environment proved to be an alternative means of detecting bouldery landforms.