Hydro-Mechanical Properties of an Unsaturated Frictional Material

2.6. IDENTIFICATION OF HYDRAULIC FUNCTIONS USING INVERSE
PROCEDURES
soil-water characteristic curve are necessary to estimate reliable unsaturated hydraulic
conductivity functions.
- Inverse procedure using experimental results derived from the Multistep method
Unsaturated hydraulic conductivity was directly estimated using tensiometer measure-
ments during multistep outflow experiments and in a additionally performed steady-
state downward flow experiment (Klute & Dirksen 1986b). Three different soils were
investigated (loamy sand, sandy loam, loam). The direct estimation of the unsaturated
hydraulic conductivity using tensiometer measurements only gave 2 results for each soil,
that is not very meaningful. The comparison of the results showed that the unsaturated
hydraulic conductivity calculated from steady-state downward flow experiment were
underestimated by the inversely determined unsaturated hydraulic conductivity using
multistep outflow data. The authors found the multistep method is superior to the
one-step method and is more feasible for simulating unsaturated flow.
A comparison of inverse solutions for determination of unsaturated hydraulic functions
using both one-step and multistep experimental data was conducted by Eching & Hop-
mans (1993) on silt loam, loam, sandy loam and fine sand for one drainage cycle. To
avoid the lack of uniqueness simulations were performed not only in combination with
initial and final water content and pressure date, but also including soil water pressure
head data. Therefore a pressure cell was additionally instrumented with a tensiometer
sensor and pressure transducer. The authors found that the identification of soil-water
characteristic curves by using one-step and multistep outflow data are improved when
using simultaneously measured water content and corresponding suction data.
Both one-step and multistep approaches for the estimation of unsaturated hydraulic
functions using inverse modeling were also compared by Dam et al. (1994) on a loam.
Using the outflow measurement data only the multistep outflow method resulted in
unique estimates while the one-step outflow method often yielded in non-unique solu-
tions. Major problem of the one-step pressure method is the fast change in pressure,
that does not occur in nature and that may leads to non-equilibrium condition in the
investigated soil (Dam et al. 1992, 1994).
The multistep method was reevaluated by Fujimaki & Inoue (2003), who checked the re-
liability of the multistep method through comparisons with independently measured un-
saturated hydraulic conductivity. Unsaturated hydraulic conductivity was determined
using outflow data and i) an optimization method and tensiometer readings, ii) a direct
method using tensiometer readings in one depth and iii) a direct method using ten-
siometer readings in two depth. Best agreement was found between the direct measured
results. The observed results using optimization method were underestimated by the
measured results.
59