1996), parameterised using the ePTF. Where possible, soil data from the experiments were used as inputs for the ePTF.
To evaluate the performance of the published PTFs as well as the ePTF, at each matric suction, the values measured and estimated for either hydraulic property were compared using a series of statistical goodness-of-fit measures.
The performance of the ePTF was evaluated using the values of [R.sup.2] and RMSE; again the hydraulic conductivity values were In-transformed for these analyses.
These values were better than those of any other PTF and thus indicated that these models should be selected for the final ePTF.
The final ePTF for estimating soil water retention was then composed from the equations from Wosten et al.
As a preliminary evaluation of the performance of the ePTF, its estimates were compared against the measured values in the calibration dataset.
The performance of the ePTF for estimating the soil water retention characteristics showed promising results against the testing dataset.
No significant deviation for the general performance of the ePTF was found even for soils that were not typical of the datasets used in the development of the original PTFs, such as soils containing high levels of allophane, pumice, or organic matter.
Due to the limited hydraulic conductivity data, it was not possible to properly test the ePTF on the testing dataset.
A functional evaluation of the ePTF was performed using the APSIM model parameterised with the outputs of the ePTF.
Overall, but especially for the soil water retention curve, the ePTF can be considered as a suitable alternative for obtaining the hydraulic characteristic curves for New Zealand soils.
The test against an independent dataset showed that the ePTF can be applied to the entire moisture range and for different soil types, including allophanic and pumice soils.