Improving the representativeness of unsaturated hydraulic parameter estimation using repeated drainage–imbibition tests

Abstract

Water movement in the vadose zone plays a critical role in groundwater recharge, contaminant transport, slope stability, and landfill performance. However, the determination of unsaturated hydraulic properties remains challenging because estimated parameters are highly sensitive to experimental conditions, monitoring locations, and wetting–drying history. This study proposes an experimental and analytical framework for improving the representativeness of unsaturated hydraulic parameter estimation using repeated drainage–imbibition (RDI) tests and inverse modeling. One-meter soil columns packed with homogeneous coarse and fine sands were instrumented with volumetric water content and matric potential sensors at multiple depths. Repeated drainage–imbibition cycles were conducted under controlled boundary conditions, and the collected datasets were analyzed using inverse modeling to estimate the soil–water characteristic curve (SWCC) and hydraulic conductivity function (HCF) parameters. The results showed that hydraulic responses varied significantly with measurement depth, drainage–imbibition phase, and repetition cycle. Consequently, hydraulic parameters derived from individual tests differed substantially, indicating that a single experiment or monitoring location may not adequately represent the overall hydraulic behavior of the soil. By integrating datasets from repeated cycles, more stable and representative parameter sets were obtained. Incorporating hysteresis into the inverse modeling further improved prediction accuracy, particularly for the fine sand where wetting–drying effects were more pronounced. In contrast, parameters estimated using pedotransfer functions showed relatively limited agreement with experimental observations under dynamic conditions. The proposed RDI-based framework provides a robust approach for reducing uncertainty and deriving representative unsaturated hydraulic properties under cyclic boundary conditions. The methodology has potential applications in vadose-zone characterization, groundwater recharge assessment, landfill cover design, and L-NAPL contamination analysis.