On the use of soil hydraulic conductivity functions in the field

The estimation of soil water fluxes using the Darcy–Buckingham flux-gradient approach is, after a century of use, still problematic under field conditions. Two features of the soil hydraulic conductivity (K) function are the main causes of failure, first the exponential character of the K versus soil water content (θ) relations, which lead to large variations in K for minimal variations of θ. Due to this, the level of precision of field measurements of θ and the spatial variability of θ itself, make the deterministic estimation of soil water fluxes unfeasible using this approach. Secondly, the spatial variability of the parameters of the K(θ) relations also contribute heavily to errors in soil water flux estimation from site to site. In a coffee crop water balance experiment, soil water fluxes below the root zone were estimated over one year, comparing the use of a soil hydraulic conductivity function obtained in the field, with an indirect climatologic approach in which the deep drainage is estimated from a water balance excess. Five replicates gave the possibility of calculating variances of both forms of calculation and their respective coefficients of variation (CV). Results show that CVs of the estimates made through the Darcy–Buckingham approach varied from 78 to 122%, in comparison to 8–23 for the indirect climatologic approach. It is therefore concluded that Darcy–Buckingham approach used deterministically under field conditions does not yield consistent results.