3. Methodology
Both the geometrical characteristics of rock mass structures and rock mass strength could be controlled by a fault within a certain area (Osmundsen et al., 2009). The results of geometrical characteristics of rock mass structures and rock mass strength within the same fault zone should be consistent approximately if the approaches are used suitably. Hence, we firstly explored the spatial variation in the geometrical characteristics of the rock mass structures. Rock mass structures at the slope scale were identified and measured using a UAV at five selected sites at varied distances from the YLTP Fault core (Fig. 1), with the consideration that exhumation doesn’t influence fracture measurements at the surface (Savage & Brodsky, 2011). The selection of the sites was based on the outcrop rock mass conditions and the rock mass structures present. The horizontal distances of the five sites from the YLZP Fault core are 0.5 km, 3.0 km, 3.4 km, 8.5 km and 13.5 km (Fig. 1). To get precise geometrical data of rock mass structures, we set at least six ground control points (GCP) at each site when flying UAV. The UAV used in our study is Phantom 4 RTK that provides real-time, centimeter-level positioning data for improved absolute accuracy on image metadata (https://www.dji.com/ca/ phantom-4-rtk). To satisfy the requirement of data resolution, we ensured lateral overlap ratio of aerial photography by UAV more than 65% and heading overlap ratio more than 75%. We sub-sampled point clouds to a minimum point spacing of 0.1 m by Agisoft Photoscan (AgiSoft LLC, 2010).
At each site, the same window (100 β…Ή 100 β…Ή 100 m) was selected for measuring the dip/dip direction and spacing of all visible rock mass joints structures by PhotoScan, Coltop (Jaboyedoff et al.,2007) (Figs. 4a and b) and ESRI ArcMap 10 software. We generated the stereographic projections by inputting the data into Rocscience DIPS 7.0 software. We selected different appropriate viewpoints in point cloud model of PhotoScan to generate orthographic projection images according to the occurrence of each joint, and then the image data with scale were imported into ArcMap. By ArcMap, we vectorized each joint and measured discontinuity spacing in detail. The joint size measured is based on the quantity of data obtained by UAV, with a minimum joint spacing of 0.3m (Fig. 4b).
Fracture density is an important parameter in quantifying the geometrical character of the rock mass (Faulkner et al., 2010). To estimate fracture density, we used three-dimensional geomechanical data to provide a joint volume count (Jv), which we then took as a measure of block size and of the total number of joints encountered in a cubic meter of the fractured rock mass (Palmstrom, 2005). After measuring the spacing of the joints, we calculated mean value of each group of joints. Using the mean spacing values of the joint sets, we calculated Jv as follows (Palmstrom, 2005):
Jv=\(\frac{1}{S1}+\frac{1}{S2}+\frac{1}{S3}+\ldots+\frac{1}{\text{Sn}}\)(1)
where 𝑆𝑖 is the mean joint spacing for each joint set, for 𝑖 = 1, 2, . . ., 𝑛.
To verify the results of joint spacing and fracture density Jv at the five sites (1-5), we independently measured fallen block sizes using the UAV and Photoscan imagery (Fig. 2).