The knowledge of soil water retention characteristic (SWRC) is crucial for solving many soil and water management problems related to agricultural, ecological, hydrological and environmental issues. However, the information on this soil property is usually missing in soil databases, since the methods for determination of SWRC are cumbersome and expensive. The dearth of this information in developing countries located in the tropics is even worse than elsewhere due to additional problems linked to personnel training and acquisition of needed equipment for these measurements. Pedotransfer functions (PTFs) which are used to estimate soil water retention characteristics (SWRC) from other available soil properties has been considered as an attractive alternative for direct SWRC measurement.

Since developing new PTFs for particular regions is very arduous and mostly requires a large soil database of good quality, utilizing existing PTFs where possible is highly recommended. However, many PTFs have limited portability and predictive capacity, since they have been derived for specific soils within limited geomorphic and geographic domain. Predominant clayey soils being mainly exploited for intensive rice-based agriculture in the tropical Vietnamese Mekong Delta (VMD) have distinct characteristics, which implies that PTFs derived elsewhere could not be directly utilized without evaluating.

Therefore, the objective of this study was to evaluate the applicability and reliability of published PTFs in predicting SWRC for tropical VMD soils. The predictive performance of a number of well-known regression based and pattern-recognition based PTFs developed from soils in both tropical and temperate regions were evaluated based on the dataset of 160 soil samples. The sampling locations were selected so that the dataset well represents the distribution of different soil types and textures in the delta. The ME, SDE and RMSE are validation indices used to assess PTFs? performance.

The results of this study show that most of the investigated PTFs which were derived from soils in both tropical and temperate climates have somewhat limited predictive potential for tropical Mekong Delta soils. The difference in soil texture (i.e. clay and sand content) as well as the mineralogy of the clay fractions in PTFs calibration datasets compared to evaluated dataset are mainly explain the inadequate performance of the investigated  PTFs for tropical Mekong delta soils. In order to acquire accurate and reliable SWRC estimations for further uses in the Delta, it is necessary to develop new PTFs based on an extensive database of soils of tropical deltas.