Estimation of Final Soil Infiltration Rate Using Fuzzy Clustering Algorithm, Nero Fuzzy (ANFIS) and Fuzzy Inference System (FIS) (A Case Study: Behshahr Plain, Galougah, Mazandaran, Iran)

Infiltration plays an important role in surface and subsurface hydrology and it is a key factor in the rainfall and runoff equations. The use of new approaches that have no limitations of common theoretical and empirical methods to determine infiltration relationships, will minimize the necessity of time consuming and costly experiments to determine permeability values and will make it possible to estimate the functional values. In this regard, in the present study the amount of soil permeability was estimated in Behshahr plain of Galougah located in Mazandaran province, using Fuzzy Inference System (FIS), Fuzzy Clustering Algorithm (FCA) and Nero-Fuzzy (ANFIS); so that, initial soil moisture content, soil organic matter content and soil lime content were considered as input parameters, and final soil infiltration rate was considered as output parameters of the models. Finally, the results obtained by the three mentioned modes were compared to the observed values by single-ring approach. According to the achieved results, Nero-Fuzzy approach with a mean deviation of 0.0042 cm/min, BIAS value of 0.6754 cm/min, Root-Mean-Square Error of 1.2096 cm/min and correlation coefficient of 0.9233 showed the most appropriate performance to estimate soil infiltration rate among the studied models; while, Fuzzy Clustering Algorithm with a mean deviation of 0.0075 cm/min, BIAS value of 2.1165 cm/min, Root-Mean-Square Error of 2.0244 cm/min and correlation coefficient of 0.8776, and Fuzzy Inference System with a mean deviation of 0.0161 cm/min, BIAS value of 2.5042 cm/min, Root-Mean-Square Error of 2.4533 cm/min and correlation coefficient of 0.8167 were placed in the next ranks respectively. Also, the highest correlation between observed and estimated values was seen in Nero-Fuzzy model (R2=0.85), and the two other studied models including Fuzzy Clustering Algorithm (R2=0.77) and Fuzzy Inference System (R2=0.66) are at the next ranks respectively. At the end of this research providing more data of soil physical and chemical characteristics as well as permeability amounts has been recommended in order to more accurate estimation of the studied models.