Full probability design of soil slopes considering both stratigraphic uncertainty and spatial variability of soil properties

Stratigraphic uncertainty and soil spatial variability are the main components of slope uncertainty. However, these two uncertainties are not taken into account at the same time in the current slope design, which will inevitably lead to an inaccurate design scheme. This paper aims to propose a full probabilistic design method for soil slopes with modeling of these two types of uncertainties by generalized coupled Markov chain (GCMC) and random field simulation. The implementation procedure for full probability design of soil slopes considering the aforementioned two types of uncertainties is summarized. A soil slope is taken as an example for design using the borehole data in Perth, Australia. To illustrate the importance of considering these two types of uncertainties in slope design, the design results associated with two cases (i.e., only considering spatial variability of soil parameters and considering these two types of uncertainties) are compared. The influences of borehole layout scheme and spatial variability on soil slope reliability-based design (RBD) results are investigated. The results indicate that if only spatial variability of soil parameters is considered in slope RBD, the design results largely depend on the used stratigraphic profile. Moreover, both the location and number of boreholes have an impact on soil slope RBD results. To obtain the RBD results accurately, a sufficient number of boreholes should be provided, and the boreholes around the critical influence zone should be drilled with priority. The spatial variability, including both autocorrelation function (ACF) and scale of fluctuations (SOFs), has considerable influences on soil slope RBD results.