Prediction of the Soil-Water Retention Curve of Loess Using the Pore Data from the Mercury Intrusion Technique

The soil-water characteristic curve (SWCC) is a crucial input parameter for describing the distribution of soil moisture and water movement in various environmental and geotechnical challenges. It is widely recognized that the SWCC is controlled by the soil's pore structure. Attempts to estimate the SWCC using pore data obtained through the mercury intrusion porosity (MIP) technique have been conducted. However, the performance of MIP estimation remains uncertain and requires further validation with experimental data. In this study, the accuracy of MIP estimation is validated using intact and compacted loess samples prepared at different water contents, specifically dry of optimum (8%), optimum (17%), and wet of optimum (19%). The results reveal that intact and dry of optimum specimens exhibit relatively good pore connectivity, with more point-to-point contacts between particles. Conversely, specimens compacted under optimum and wet of optimum conditions exhibit poor pore connectivity, with more isolated pores, particularly in the wet-of-optimum specimen. The SWCC predictions based on MIP data are accurate for intact and dry-of-optimum compacted specimens, but significant errors occur for the optimum and wet-of-optimum specimens. Prediction accuracy using MIP data is closely linked to the soil's pore connectivity. Despite a tenuous theoretical basis in the high suction range where adsorption forces dominate, a strong consistency between predictions and measured data across a wide suction range (e.g., 10-104 kPa) instills a high level of confidence in using MIP data to predict the wetting SWCC. The contact angle required for the prediction is suggested as a fitting parameter.