Modelling Soil Water Contents at Field Capacity and Permanent Wilting Point Using Artificial Neural Network for Indian Soils

Abstract

Soil field capacity (FC) and permanent wilting point (PWP) are important input parameters in many biophysical models. Although these parameters can be measured directly, their measurement is quite difficult and expensive. Pedotransfer functions (PTFs) provide an alternative by estimating soil parameters from more readily available soil data. A study has been conducted to evaluate PTFs of FC and PWP created using artificial neural networks (ANNs). A total of 721 different sampling locations spread all over India are selected to develop PTFs using ANN. Results indicate that six neurons in hidden layers are best suited for prediction of FC and PWP. The statistical criteria (value of R2, RMSE, MBE, ME, and d) is used to evaluate ANN, indicated an unbiased and higher predictability of developed models.