Influence of Terrain Model Resolution on Rockfall Dispersion and Tortuosity of Rockfall Paths

The geology of the Snake River Plain region of southwestern Idaho contains terraces formed from downcutting of the Boise River and columnar jointed basalt cliffs from lava flows. Within the city limits of Boise, Idaho, rockfall events originate from these columnar jointed basalt cliffs. Whitney Terrace, a rockfall-prone location in southwestern Boise, was chosen as the study site. Site-specific LiDAR data was used to create a bare earth terrain model, and rockfall boulders were geolocated and measured to be used as inputs into rockfall models. The slope below the cliff face was divided into four distinct slope zones, each with its own average slope, sequentially decreasing in gradient away from the cliff. RocFall3, a three-dimensional rockfall modeling program, was used to model the site and simulate rockfall boulders. Rigid body simulations were performed to investigate the effect of terrain resolution, boulder size, and boulder shape on tortuosity and dispersion area. Two boulder shapes (hexagons and spheres), four boulder sizes, and five terrain model resolutions were used, resulting in 20,000 single boulder rockfall simulations. Tortuosity was defined as the ratio of the total boulder path length to the straight line distance of the path. It was found that as the terrain model resolution increased, the tortuosity values for spherical and hexagonal boulders minimally increased. Spherical boulders experienced a decrease in tortuosity values as their mass increased. Hexagonal boulders did not show a consistent trend with increasing mass. Spherical boulders had larger observed tortuosity values than the more angular-shaped hexagonal boulders. Results showed that spherical boulders have longer runout distances and greater lateral movements than hexagonal boulders; however, the tortuosity values are generally similar for the two shapes. A water droplet analysis was performed to investigate if the runout paths generated could be compared to lumped mass low-energy rockfall simulations. The results were comparable at the lower terrain resolutions. The dispersion area was defined as the area bound by the seeder location and the outermost end locations of the simulated boulders. As the terrain model resolution increased, the dispersion areas increased. Hexagonal boulders produced a rounded cone dispersion area, while the spherical boulders produced an elongated end cone shape. In terms of mass, for the spherical boulders, dispersion areas generally decreased as the mass increased. For the hexagonal boulders, there was not a clear relationship between dispersion area and mass. It was found that a local variation in topography (a soil mound) influenced the shape of the spherical boulder dispersion but not the shape of the hexagonal boulder dispersion. The vast majority of the spherical boulders came to rest in the lowest gradient slope zone. The hexagonal boulders came to rest in the two lowest slope zones. There are three recommendations for future work with this research. Additional boulder shapes should be simulated to examine how shape can further influence the tortuosity of boulder paths and dispersion areas. The idea of applying tortuosity to boulder paths could be expanded upon to include a relationship between slope angles and boulder energies because it was noted that all boulder paths were generally straight in high-angle slope zones and paths became tortuous in lower-angle slope zones. Currently, the dispersion area includes significant portions of the slope with no rockfall boulders. The definition of dispersion area could be modified so it does not include the seeder location but only encapsulates the end locations of boulders, defining the outer boundary of deposited boulders.