Probabilistic Estimation of Soil-Water Characteristic Curves of Fine-grained Soils

Abstract

The soil-water characteristic curve (SWCC), which describes the relationship between water content in soil and its matric suction, is important for studying the mechanical properties of unsaturated soil. Because measuring SWCC is time-consuming and labor-intensive, many empirical methods have been proposed for assessing SWCC. However, the uncertainty associated with empirical methods is rarely quantified. The purpose of this paper is to propose a probabilistic method for assessing the SWCC of fine-grained soils based on the particle size distribution and index properties. First, the van Genuchten model, which is a widely used SWCC model, is briefly presented. A comprehensive soil database was then compiled and the parameters of the van Genuchten model were calibrated for each soil sample. After that, a regression analysis was performed to estimate the mean value of the parameters of the van Genuchten model, and a correlation analysis was performed to estimate the covariance matrix of the parameters of the van Genuchten model. Finally, a joint probability density function of the parameters of the van Genuchten model is constructed. In this paper, Johnson distributions are used to construct the joint probability distribution of the parameters in the van Genuchten model that do not obey the multivariate normal distribution. The proposed method has been rigorously tested to assess its accuracy and applicability. The method proposed in this paper can provide a practical tool for estimating SWCC with limited data and accounting for the associated uncertainty.