Bayesian estimation of soil-water characteristic curves
Jie Zhang, Shimin Yang, L.L. Zhang, Mei Zhou
Abstract
The soil-water characteristic curve (SWCC) is a significant prerequisite for studying the mechanical properties of unsaturated soil. As experimental measurement of the SWCC is time-consuming, empirical methods have been suggested to estimate the SWCC. However, the uncertainty associated with SWCC can be substantial. In this paper, a hybrid method based on Bayes’ theorem is suggested to estimate the SWCC, where an empirical method can be used to provide prior knowledge about the SWCC, and a limited quantity of measured data are used to update the SWCC. The Bayesian model is then solved with a Markov Chain Monte Carlo simulation. Through the suggested method, the valuable information provided by the empirical method can be combined with the measurement data. The suggested method can not only provide the best estimate about the SWCC, but also account for the associated uncertainty. Also, the effect of more measured points on the estimation of SWCC can be quantified. The suggested method provides a practical means to estimate the SWCC using a limited amount of data.