Groundwater salinity prediction using adaptive neuro-fuzzy inference system methods: a case study in Azarshahr, Ajabshir and Maragheh plains, Iran

Autor: Hosnie Nazari, Behnam Taghavi, Farnusch Hajizadeh
Rok vydání: 2021
Předmět:
Zdroj: Environmental Earth Sciences. 80
ISSN: 1866-6299
1866-6280
DOI: 10.1007/s12665-021-09455-3
Popis: This study aims to investigate the groundwater salinity due to physical and chemical parameters using ANFIS-FCM and ANFIS-SCM methods in Azarshahr, Ajabshir and Maragheh plains situated in the Catchment Area of Urmia Lake, Iran. To this aim, 82 water samples were taken from wells and spring across the plains and chemically were analyzed in the laboratory. Descriptive statistics and correlation matrix of the studied parameters were obtained by SPSS software. Correlation matrix showed that four parameters including electrical conductivity (EC), dissolved oxygen (DO), total soluble solids (TDS) and pH had the highest correlations with salinity compared to the other parameters. Therefore, the mentioned parameters selected as inputs and salinity were the output according to the purpose of the study. After standardization, data were entered into the MATLAB environment and groundwater salinity was predicted using ANFIS-FCM and ANFIS-SCM methods. The models’ results showed that the estimated groundwater salinity for ANFIS-SCM model had very good accuracy and more correlation than the measured values. As a result, ANFIS-SCM intelligent method has been an effective, efficient and accurate method to estimate the parameters in the study area.
Databáze: OpenAIRE
Externí odkaz: