Vulnerability of critical slopes by using continuous Bayesian networks

ABSTRACT The paper deals with the vulnerability assessment of a critical slope in Mexico throughout the use of a Bayesian network with continuous variables that consider the slope geometry and random soil properties and random rainfall features. The soil properties change according to the random rainfall intensity, which is modeled as exponentially distributed, making the process highly dynamical. The influence of each parameter, especially the phreatic water level, on the safety level is explored through continuous functions. Practical mitigation recommendations may be devised by analyzing the BNet and modeling variations on these variables.