We investigated three types of transient flow experiments (Onestep, Multistep, Continuous) with regard to their
suitability for the determination of hydraulic properties of large undisturbed soil samples. The methods differ in
the way a pressure gradient is applied which induces the flow of water out of or back into the soil core. The
instantaneous application of one large pressure step is called Onestep, the application of several smaller
increments is called Multistep, and the smooth continuous change of the pressure gradient is called Continuous.
The three methods were compared with regard to the uniqueness, identifiability and stability of the estimated
parameters of the hydraulic functions. The functions used were able to represent unimodal or bimodal pore
systems, with or without hysteresis.
The results for a sandy forest soil show that the Onestep method yields nonunique solutions. Further, since the
method does not represent conditions that occur in nature, it appears not suitable for estimating the soil
hydraulic parameters. In contrast, the Multistep and Continuous methods lead to a unique parameter
identification. Both methods yield the same soil water retention curves q(y) for the respective soil core.
However, the parameter for the saturated hydraulic conductivity, Ksat, is insensitive to the optimization of the
outflow data of the Continuous flow experiments. The optimized Ksat values are in some cases one order of
magnitude smaller than those obtained by the Multistep method. This in turn yields different K(q) relationships
for the two methods. The hysteretic properties of
the analyzed soils could not be adequately described by the Kool and Parker (1987) hysteresis model.
With the present level of information it can be concluded that the Multistep and Continuous method are equally suitable in the estimation of the soil hydraulic functions of soil cores.
