Probabilistic soil-water characteristic curve of residual soils: A case history from Korea

Residual soils evolved from the weathering and decomposition of granites cover extensive area of the Korean Peninsula and generally exist in an unsaturated state. Soil-water characteristic curve (SWCC) is key information for estimating the response of unsaturated soils. This study presents the results of a copula-based probabilistic analysis of SWCC from incomplete data of site-specific residual soils. The uncertainty in the SWCC is characterized by two empirical parameters underlying the van Genuchten model of SWCC. A database of 52 SWCCs available in the literature for Korean residual soils is used to identify and simulate the probability distribution of bivariate SWCC parameters. The stochastic behavior of SWCC is optimized by testing a variety of marginal and copula functions. The descriptive statistics of the measured and simulated parameters are compared. The result indicates that the Gaussian copula with Weibull-truncnormal marginals is suitable for modeling the negative nonlinear dependency of the SWCC parameters of Korean residual soils. It is also shown that ignoring the negative correlation between SWCC parameters will not adequately represent that of the measured data. The probabilistic SWCC provides a rational basis for estimating key unsaturated soil properties such as water permeability, and shear strength, enabling more reliable geotechnical design under uncertainty, including slope stability and unsaturated seepage. The importance of a probabilistic SWCC is demonstrated through example calculations highlighting the variability in unsaturated soil properties induced by uncertainty in SWCC parameters.