Estimation of water retention curve of granular soils from particle-size distribution - a Bayesian probabilistic approach

This study proposes two empirical relationships to estimate the parameters of van Genuchten's formula for modeling the water retention curve from the particle-size distribution. The relationships are determined by the Bayesian probabilistic method for selecting the most plausible class of models based on a database of 90 soil samples. The highest plausibility model among the selected relationships shows that the parameter alpha can be expressed as a first-order function of the particle size at 50% passing (d(50)) and parameter n is expressed as a third-order polynomial of the reciprocal of the standard deviation of geometric mean particle size (sigma(g)). The predictability of proposed relationships for other soils outside the calibrated database is also presented. It is found that the model prediction is highly consistent with the measurements for sands. However it only matches well with the measurements in the low suction regime for soils with at least 20% of fines content.