Determination of a suitable model for prediction of soil cation exchange capacity

Analysis and design of land-use management scenarios requires detailed soil data. The cation exchange capacity (CEC) of soil is a basic chemical property, as it has been approved that the spatial distribution of CEC is important for decisions concerning pollution prevention, crop and farming management. Since laboratory procedures for measuring CEC are cumbersome and time-consuming, it is essential to develop an indirect approach such as pedotransfer functions to predict this parameter from more readily available soil data. The aim of this study was to compare multiple linear regression, multiple non-linear regression, adaptive neuro-fuzzy inference system and artificial neural network including feed-forward back propagation (FFBP) model to develop PTFs for predicting paddy soils CEC in Guilan province, northern Iran. Two soil parameters including organic carbon and clay were considered as input variables for proposed models. 171 soil samples were used. The data set was divided into two subsets for calibration and testing of the models. The models prediction capability was evaluated by comparison with observed data through various descriptive statistical indicators include root mean square error, determination coefficient, mean bias error and relative improvement values. Results showed that the FFBP model had the most reliable prediction when compared with other models and that provide a new methodology with acceptable accuracy to estimate the CEC of soil that diminished the engineering effort, time and funds and can provide the scientific basis for the study of soil CEC and be helpful for the estimation of soil CEC in other places with similar conditions, too.