A novel laboratory method to estimate the soil hydraulic property from the inverse analysis of an infiltration curve measured in finite columns

• Soil hydraulic properties are estimated from an infiltration plus drainage curve. • Two complementary methods, QEI and HYDRUS-VG, are used and compared. • The methods are evaluated on synthetic, sieved and undisturbed soil cores. • Robust relationships are found between the theoretical and optimized parameters. • The methods are a reliable, efficient and cost-effective approach. The van Genuchten-Mualem (VG) model, commonly used to describe soil hydraulic properties, requires five parameters: saturated hydraulic conductivity ( K s ), water retention curve parameters ( α and n ), and residual ( θ r ) and saturated ( θ s ) water contents. This study presents two complementary methods to estimate these parameters from the inverse analysis of a single infiltration curve followed by gravity drainage, measured at saturated conditions in finite, short soil columns. The first method, QEI, employs the Haverkamp formulation for 1D geometry. The second, HYDRUS-VG, combines the Haverkamp formulation with the HYDRUS-1D and the VG models. Assuming the initial water content of the soil as a measurable data equal to θ r , both methods take advantage of four singular characteristics of the infiltration curve followed by gravity drainage, such as the slope of the drainage section (related to K s ), the soil water increase ( Δθ T ), the infiltration time at the end of the transient phase, and the initial infiltration steps (related to sorptivity). The methods were evaluated on infiltration curves simulated on synthetic soils generated with HYDRUS-1D, and on experimental infiltration data measured in both sieved loam soil and undisturbed cores collected from the same loam soil under different tillage management practices: reduced tillage (RT), where chisel ploughing is used as primary tillage practice, and no-tillage (NT). A significant relationship (R 2 = 0.99; y = 1.00x + 5.2 10 −5 ; p < 0.001) was found between the theoretical K s , α , n , and θ s and the corresponding parameters optimized in the synthetic soils with both QEI and HYDRUS-VG. The methods were also able to detect differences in soil hydraulic properties between sieved soil and RT and NT managements, with RT soils exhibiting the highest values of K s , α , and n . The proposed methods offer a reliable, efficient, and cost-effective approach to estimate soil hydraulic properties, making them valuable tools for various hydrological, engineering or agricultural applications, among others.