Article Title: EEstimating cation exchange capacity using fractal parameters of particle size distribution by artificial neural networks and regression models

The cation exchange capacity (CEC) is one of the most important soil chemical properties that influences other soil properties and is a suitable index of soil quality and use. Direct measurement of the CEC is difficult، costly and time-consuming. Therefore، the objective of this study was to predict the CEC by pedotransfer functions (PTFs) using readily available soil properties as predictors. The data were supplied from two provinces of Iran: 129 soil samples were collected from dominant soil series of Hamadan and Guilan provinces. The soil particle size distribution (PSD)، organic matter، CEC and pH have been measured in the laboratory. In this research in addition to conventional predictors of soil CEC such as soil textural fractions، pH and organic matter، fractal parameters of four PSD models including Bird، Perrier-Bird، Kravchenko-Zhang and Yang et al. have been used to predict CEC. In order to evaluate the predictability of CEC from fractal parameters of PSD، regression and artificial neural networks (ANNs) techniques have been used. The results showed that entering fractal parameters of PSD especially the parameters of Bird and Perrier-Bird to the models improved the statistical criteria of the models. In addition، ANN models performed better than regression models. Using new predictors in the PTFs improved the prediction of CEC.