Main content area

Estimation of the soil hydraulic properties from the transient infiltration curve measured on soils affected by water repellency

Moret-Fernández, D., Latorre, B., Giner, M.L., Ramos, J., Alados, C.L., Castellano, C., López, M.V., Jimenez, J.J., Pueyo, Y.
Catena 2019 v.178 pp. 298-306
coatings, equations, hydraulic conductivity, hydrophobicity, infiltration rate, models, soil water, topsoil, volumetric water content, water repellent soils, wettability
Estimation of soil sorptivity (S) and hydraulic conductivity (K) is fundamental to model the water infiltration into the soil. This process can be affected by soil water repellency, which is defined as a reduction in soil wettability due to coating of soil particles by hydrophobic substances. Unlike to wettable soils, this phenomenon can generate infiltration curves with double-slope shape: a transient infiltration curve followed by a steady-state section. Because the topsoil final volumetric water content (θ1) of the transient phase of the double-slope curve is not a measurable data, in principle, the standard model based on the Haverkamp et al. (1994) model cannot be used to estimate S and K. This work presents two different approaches based on the Haverkamp et al. (1994) equation, which allow estimating S and K from the first phase of a double-slope infiltration curve, when θ1 data are not available. The methods, which are based on the analysis of both short-medium time transient infiltration curve (Tr) and the combination of both short-medium transient and steady-state infiltration steps (Mx), were applied on 20 soils affected by different degrees of water repellency. The Haverkamp et al. (1994) model was also valid for infiltration curves measured on hydrophobic soils, and the final volumetric water content was not an essential data to estimate K and S. Although the steady-state infiltration rate (q1) calculated with Mx was about 26% larger than that estimated with Tr, comparable K and S values were obtained with both methods. Overall, a large dispersion on the estimate of θ1 was observed with both methods. The gravimetric time, tgrav, estimated in the studied soils was low, <500 s. While the Mx method required simpler numerical calculus, Tr looked like to be more robust and less subjective.