Understanding how water behaves in unsaturated soil is essential for geotechnical engineering. Soil-water characteristic curves are used to estimate permeability, shear strength, tensile strength, and other hydraulic properties, and they are especially important in seepage modelling and slope stability analysis. Yet soil behaves differently during drying and wetting, and ignoring this hysteresis can lead to analyses that do not reflect real field conditions, particularly during rainfall and snowmelt infiltration.

This issue becomes even more important in dual-porosity soils, where water moves through both macropores and micropores and the soil-water characteristic curve becomes bimodal. Such soils are common in engineered fills, compacted pavement materials, and coarse colluvial deposits. Because many geotechnical failures occur during wetting rather than drying, using only drying curves can produce inaccurate or overly simplified predictions of infiltration and stability.

The study matters because it fills a missing piece in the field. According to the authors, equations had been developed for drying bimodal SWCC, but not for wetting bimodal SWCC. At the same time, many existing fitting models are either mathematically cumbersome or rely on parameters that have little physical meaning. A simpler model with physically interpretable parameters could therefore make both laboratory interpretation and engineering application more practical.

How was the study carried out?

The researchers developed a new mathematical equation to model the bimodal wetting SWCC across the full suction range. The model was designed so that its parameters correspond to physically meaningful features of the wetting curve, such as water saturation points, delimiting point, water entry point, and associated moisture contents. This makes the equation easier to interpret and potentially more useful in practical analysis than purely empirical best-fit models.

To test the model, the team used eleven bimodal wetting datasets. Five datasets came from previously published studies, while six additional datasets were generated experimentally using engineered soils prepared from different proportions of silty sand and inert kaolin. The soils covered a broad range of moisture contents and dry densities, helping the authors evaluate the model across different dual-porosity conditions.

The experimental part of the study also introduced a new way to measure wetting SWCC. The authors combined a high suction polymer sensor (HSPS) for the lower suction range with a dewpoint potentiometer (WP4C) for higher suction values. According to the paper, this method enabled faster wetting measurements and helped close an important measurement gap that can arise with more conventional testing combinations.

What did the study find?

The proposed equation performed very strongly across all eleven datasets. The authors report that the model achieved coefficients of determination (R²) close to 1.0 and root mean square error (RMSE) values approaching 0, indicating a very close fit between the predicted curves and laboratory or published data. Overall, the paper reports an average R² of about 0.998 and an average RMSE of about 0.005 across the validation datasets.

Beyond fit quality, the model’s main advantage is interpretability. The authors emphasize that, unlike many earlier fitting equations, all parameters in the proposed model have physical meaning in relation to the wetting SWCC. This means that important variables can be read directly from the curve, which may simplify calibration, reduce empirical assumptions, and make the results easier to use in engineering applications such as infiltration modelling, capillary barrier design, and geobarrier analysis.

The study also found that the new HSPS-based wetting measurement approach produced bimodal wetting curves comparable to those obtained using other methods, while offering the advantage of shorter testing time. Together, the new equation and the new measurement workflow provide a more realistic way to represent wetting behaviour in dual-porosity soils, particularly in problems driven by rainfall, where wetting conditions are the most relevant.

The authors note, however, that the model was validated on a limited range of soils, from silty sand to coarse clay, and that volume change effects were not included. They recommend further testing on a wider variety of soils and acknowledge that some finer details of hysteresis, such as ink-bottle effects, are not fully captured by the simplified scaling approach used in the model.