Five pedotransfer functions for the van Genuchten model through Evolutionary Polynomial Regression.

Numerical modelling of unsaturated water flow requires knowledge of the soil-water characteristic curve and hydraulic conductivity function. These constitutive relationships are described through empirical models, whose parameters must be calibrated against collected data. Even though there are many techniques to obtain the hydraulic parameters, most of them are time-consuming, labor-intensive and expensive. Pedotransfer functions (PTFs), on the other hand, are simple analytical equations were the hydraulic parameters are estimated through physical information of the soil. PTFs decreases the cost and time needed to derive such parameters, however their predictions must be accurate enough in order to compensate their use. In this sense the present work applies Evolutionary Polynomial Regression procedure to obtain five analytical expressions for each van Genuchten parameter. The database was generated through the Rosetta black-boxed program.