WRCC models and estimation
           It is noted that most of the instruments reported above have limited range of suction measurement. Therefore, several empirical or semi-empirical mathematical models have been developed for representing WRCC. The parameters of these models can be obtained by using the limited range of measured suction data. Once the parameters are known, the same can be used for extrapolating or interpolating the results by using the WRCC model. These WRCC parameters are important input for many of the mathematical models dealing with unsaturated soils. Two such models which are used widely in the literature are van Genuchten (1980) model and Fredlund and Xing (1994) model represented by Eqs. 2.35 and 2.36, respectively.

2.35

2.35

where θ(ψ) is the volumetric water content at any suction, ψ; θ_r is the residual volumetric water content; θ_s is the volumetric water content at saturation; a_vg and a_f are fitting parameters primarily dependent on the air entry value (AEV); n_vg and n_f are fitting parameters that are dependent on the rate of extraction of water from the soil; m_vg and m_f are fitting  parameters which depend on θ_r; h_r is the suction (in kPa) corresponding to residual state. There are several such simplified and complex models reported in the literature for defining WRCC.
           The experimental procedures adopted for determining SWCC are time consuming and cost-intensive. Therefore, attempts have been made by researchers to develop functions (such as pedo transfer functions) for the quick estimation of WRCC without performing extensive suction measurements. In such cases, WRCC is based on soil physical parameters that can be quickly determined in the lab. This indirect approach is less time-consuming, simple, and more economical. However, such estimations can be soil specific and case-specific and would depend mostly on the data used for developing the procedure.

           Therefore, estimated WRCC should be used with caution and only in those cases where suction measurements cannot be performed. For important projects it is always preferable to obtain measured WRCC for the soil. For more details on WRCC estimation, readers are requested to go through the wide range of literature available (Fredlund et al. 1998; Mbagwu and Mbah 1998; Fredlund et al. 2002; Matula et al. 2007; Nimmo et al. 2007; Soil vision 4.10)
 
Complexity in modelling the behaviour of unsaturated soil
            As discussed above, all the properties of unsaturated soil such as seepage, strength, and volume change behaviour are dependent upon the suction existing in the soil. These properties changes when the state of unsaturation changes and suction changes. The state of unsaturation is defined by θ, w or S_r. As against the steady state behaviour in saturated soil, an unsaturated soil therefore exhibit transient behaviour. The complex behavioural modelling of unsaturated soil is explained with respect to unsaturated hydraulic conductivity (k_u) and flow as an example.
            In the case of saturated soil, hydraulic conductivity (k_s) remains constant with time. This is mainly due to the fact that all the pores are filled with water. In the case of unsaturated soil there is retention forces (suction) acting on water that would restrict easy movement. Due to this, hydraulic conductivity of unsaturated soil drastically reduces and is essentially a function of water content or suction present in the soil. For a particular soil, k_u increases as suction decreases till it approaches ks as shown in Fig. 2.13 (Malaya 2011). This is mainly due to the fact that suction decreases due to the increase in water content which results in the increase in k_u. Therefore, it is clear that k_u is a function and changes with water content or suction. These functions are highly non-linear and laborious to determine experimentally. Mostly k_u functions are estimated indirectly from water-retention characteristic curve (or SWCC), which is relatively easy to determine experimentally.