Development of pedo transfer functions (PTFs) to predict soil physico-chemical and hydrological characteristics in southern coastal zones of the Caspian Sea

The research presented in this paper attempts to develop a more realistic model using multi-layer perceptron (MLP), a feed forward artificial neural network (ANN), instead of traditional models like multiple linear regression (MLR) for predicting some soil physico-chemical and hydrological properties. The study area (Guilan Province) which is located in northern Iran bordering to south side of Caspian Sea in a coastal zone has udic and thermic soil moisture and temperature regimes respectively. The estimated soil parameters were CEC, EC, ESP, MWD and final steady-state infiltration rate (IR). Although these parameters can be measured directly, their measurement is difficult and expensive, so pedotransfer functions (PTFs) provide an alternative by estimating these parameters from more readily available soil data. In order to predict the mentioned parameters, soil sampling was conducted at 500 points in the region. Measured soil variables included texture, O.C, porosity, EC, CEC, SAR, ESP, MWD, soluble cations and anions and IR. Then, ANN and MLR models were tested. The data set was divided into two subsets for calibration (80%) and testing (20%) of the models and their normality were tested by Minitab software and Kolmogrov-Smirnov method. In order to evaluate the models, root mean square error (RSME) was used. The comparison of RSME for two mentioned models showed that the ANN model gives better estimates rather than the MLR model. So that the levels of RMSE and R2 derived by ANNs models for EC, CEC, ESP, MWD and IR were 0.24, 0.96; 1.25, 0.90;  0.18, 0.94;  0.04, 0.84 and;  1.55, 0.92 respectively while these parameters for MLR models were 1.98, 0.73;, 7.92, 0.60;  1.13, 0.66;  0.187, 0.51 and;  9.45, 0.57 respectively. The superiority of ANN models compared with MLR models was probably due to a nonlinear relationship between the dependent and independent variables. Furthermore, results indicated that training is very important in increasing model accuracy for one region.   Key words: Artificial neural networks, multiple linear regressions, soil properties

[1]  M. Meirvenne,et al.  Predictive Quality of Pedotransfer Functions for Estimating Bulk Density of Forest Soils , 2005 .

[2]  L. Baker,et al.  Optimisation of pedotransfer functions using an artificial neural network ensemble method , 2008 .

[3]  A. Page.  Methods of soil analysis. Part 2. Chemical and microbiological properties. , 1982 .

[4]  M. Schaap,et al.  Neural network analysis for hierarchical prediction of soil hydraulic properties , 1998 .

[5]  Fereydoon Sarmadian,et al.  Modeling of Some Soil Properties Using Artificial Neural Network and Multivariate Regression in Gorgan Province, North of Iran , 2008 .

[6]  P. Garofalo,et al.  Application and Evaluation of the SWAP Model for Simulating Water and Solute Transport in a Cracking Clay Soil , 2005 .

[7]  B. Gérard,et al.  Modeling hydraulic properties of sandy soils of Niger using pedotransfer functions , 2007 .

[8]  G. Wessolek,et al.  Reply to comments on ``Evaluation of pedo-transfer functions for unsaturated soil hydraulic conductivity using an independent data set'' , 2002 .

[9]  J. A. Kittrick,et al.  A PROCEDURE FOR THE PARTICLE‐SIZE SEPARATION OF SOILS FOR X‐RAY DIFFRACTION ANALYSIS , 1963 .

[10]  Cathy A. Seybold,et al.  Predicting Cation Exchange Capacity for Soil Survey Using Linear Models , 2005 .

[11]  J.H.M. Wösten,et al.  Testing an Artificial Neural Network for Predicting Soil Hydraulic Conductivity , 1996 .

[12]  M. Rashidi,et al.  Prediction of soil exchangeable sodium percentage based on soil sodium adsorption ratio , 2009 .

[13]  M. Schaap,et al.  Using neural networks to predict soil water retention and soil hydraulic conductivity , 1998 .

[14]  Amit Kumar,et al.  An evaluation of artificial neural network technique for the determination of infiltration model parameters , 2006, Appl. Soft Comput..

[15]  H.W.G. Booltink,et al.  Neural network models to predict soil water retention , 1999 .

[16]  D. W. Nelson,et al.  Total Carbon, Organic Carbon, and Organic Matter 1 , 1982 .

[17]  Sanjeev Kumar,et al.  A pedo-transfer function (PTF) for estimating soil bulk density from basic soil data and its comparison with existing PTFs , 2002 .

[18]  Budiman Minasny,et al.  Comparison of different approaches to the development of pedotransfer functions for water-retention curves , 1999 .

[19]  Nader Fathianpour,et al.  Neural network models to predict cation exchange capacity in arid regions of Iran , 2005 .

[20]  Alex B. McBratney,et al.  A comparison of prediction methods for the creation of field-extent soil property maps , 2001 .

[21]  Peter Keith Woodward,et al.  Artificial neural network for stress-strain behavior of sandy soils: Knowledge based verification , 2005 .

[22]  Ferhat Karaca,et al.  NN-LEAP: A neural network-based model for controlling leachate flow-rate in a municipal solid waste landfill site , 2006, Environ. Model. Softw..

[23]  P. Jardine,et al.  Using soil physical and chemical properties to estimate bulk density , 2005 .

[24]  Budiman Minasny,et al.  The neuro-m method for fitting neural network parametric pedotransfer functions , 2002 .

[25]  D. W. Nelson,et al.  Total Carbon, Organic Carbon, and Organic Matter , 1983, SSSA Book Series.

[26]  Özer Çinar,et al.  Comparison of artificial neural network and regression pedotransfer functions for prediction of soil water retention and saturated hydraulic conductivity , 2006 .

[27]  A. Mermoud,et al.  Comparative analysis of three methods to generate soil hydraulic functions , 2006 .

[28]  Y. Pachepsky,et al.  Artificial Neural Networks to Estimate Soil Water Retention from Easily Measurable Data , 1996 .

[29]  A. Klute.  Methods of soil analysis. Part 1. Physical and mineralogical methods. , 1988 .

[30]  A. Klute,et al.  Methods of soil analysis , 2015, American Potato Journal.