Title: Development of Predictive Rockfall Runout Maps Based on Calibrated Rockfall Models
Program: Master of Science in Civil Engineering
Advisor: Dr. Nick Hudyma, Civil Engineering
Committee Members: Dr. Bhaskar Chittoori, Civil Engineering, and Dr. Mary MacLaughlin, Civil Engineering
Cliffs formed in columnar jointed basalt have the proper geometry and discontinuities to initiate rockfall events. There are a number of locations in Boise, ID that may experience rockfall events from columnar jointed basalt cliffs. Publicly available LiDAR data coupled with site reconnaissance activities were used to develop calibrated two-dimensional rockfall models at a study site in southeast Boise. Photogrammetry was used to determine the volume of seven calibration boulders. The study site, located on Whitney Terrace, consists of a columnar jointed basalt cliff followed by a slope with a subdivision at the toe of the slope.
The two common approaches used to generate rockfall models are the Lumped Mass and Rigid Body methods. At this site, the Rigid Body method that incorporated vegetative properties was demonstrated to be the most appropriate method for the rockfall model. Rockfall model calibration was performed by comparing the model-simulated runout distances and geolocations of the seven calibration boulders. The rockfall models were considered calibrated when the location of the calibrated boulder was within one-half of the standard deviation of the mean runout distance of the furthest traveling grouping of simulated rocks. The calibration provided a range of restitution coefficients that were used to develop predictive runout maps for rocks of various masses for the site. The rock sizes used for the predictive runout maps are based on the range of sizes of runout boulders at the site.
The predictive runout maps showed site vegetation acts as a barrier or energy dissipater for 50 kg and 100 kg simulated rocks. This mass of rocks had shorter runout distances compared to the 400 kg, 800 kg, and 1500 kg simulated rocks. There was only one location within the study area where large, 800 kg and 1500 kg, simulated rocks would runout past the study area and into the subdivision.
Three recommendations are provided for future work. Firstly, since size and shape are important inputs to Rigid Body rockfall models, more effort should be used to accurately determine these properties. Second, publicly available LiDAR data had to be augmented with field data. Future studies should incorporate site-specific LiDAR data collected with either a terrestrial LiDAR scanner or UAV LiDAR. Finally, three-dimensional rockfall models should be used because they incorporate full-site topography and allow simulated rocks to spread downslope and laterally.