Comparison of regression pedotransfer functions and artificial neural networks for soil aggregate stability simulation

Abstract

Abstract: Simulation of soil aggregate stability is a suitable method for saving time and cost spent for direct measurement. This research comprises regression pedotransfer functions (RegPTFs) and artificial neural networks (ANNs) for estimation of soil aggregate stability. 100 soil samples from forest and pasture's soils of Guilan Province in Iran were collected and geometric mean diameter (GMD), %silt (Si), %clay (Cl), %sand (Sa), bulk density (BD), equivalent carbonate calcium (CaCO 3), particle density (PD), soil mechanical resistance