The recently developed stochastic approach to discontinuity shear strength (StADSS) avoids scale effects by predicting peak and residual shear strength using full scale discontinuity data. StADSS has been applied to mortar replicas of perfectly matched specimens in the laboratory, but application to large in situ discontinuities is yet to be undertaken. Paramount to the in-situ application of StADSS is the adequate collection of discontinuity roughness information. This paper presents an approach to digitize the seed trace of in-situ discontinuities. Several identical images of a seed trace with varying pixel densities were analyzed to determine what pixel to millimeter ratio was sufficient to enable a Gaussian distribution of gradients to develop. It was found that a pixel density of 6-7 pixels per millimeter of seed trace length was sufficient to develop the parameters of a suitable Gaussian distribution of gradients, which could then be used for further statistical analysis.
The stochastic approach to discontinuity shear strength (StADSS) assumes that the profile of a discontinuity exposed on a rock face (referred to as the "seed trace") is representative of the roughness contained with the rock mass. This new stochastic approach bypasses the scale effect that plagues shear strength predictions (both peak and residual) by using full scale discontinuity roughness data and avoiding scale dependent roughness descriptors (Buzzi & Casagrande, 2018; Buzzi et al., 2017; Casagrande, 2018; Casagrande et al., 2018; Jeffery, 2021; Jeffery et al., 2021; Jeffery et al., 2018).
The approach was validated in the laboratory by series of direct shear tests at small scale (100mm per 100 mm, Casagrande (2018)) and large scale (2m per 2m, Jeffery (2021)). After large scale validation under controlled conditions, Jeffery (2021) recommended large scale, in-situ application of StADSS for proper field validation. To enable large scale application of StADSS in the field, a seed trace digitization methodology must be developed. Ideally, seed trace data must be digitized at high resolution (≤ 1 mm intervals) to gain enough data to characterize roughness at full scale (Buzzi & Casagrande, 2018; Buzzi et al., 2017; Casagrande, 2018; Casagrande et al., 2018; Jeffery, 2021; Jeffery et al., 2021).
