Estimation of sodium absorption ratio (SAR) and electrical conductivity (EC) using hybrid models of support vector regression and random forest-case study of Qatur river, west of Iran

Document Type : Research Paper

Authors

1 Department of water engineering, faculty of agriculture, Urmia university, Urmia, Iran

2 Urmia university

3 Urmia University

4 University of Tabriz

Abstract

In the field of irrigation, water quality plays an effective role in determining the compatibility of water with soil and different crops. As a valuable tool for water quality assessment, the Wilcox diagram enables stakeholders to promote sustainable agricultural practices by reducing potential risks related to water salinity and sodium content. However, measuring sodium absorption ratio (SAR) and electrical conductivity (EC) parameters is associated with various challenges such as cost and time, calibration error, and maintenance problems. Based on this, this research was carried out to predict SAR and EC parameters using easily accessible parameters including sulfate and chloride in Qatur rivers located in the northwest of Iran from 1993 to 2019. The models used included random forest (RF), support vector regression (SVR), random forest combined with genetic algorithm (RF-GA), and support vector regression hybridized with fruit fly algorithm (SVR-FOA). The analysis of the results of this research showed that the RF-GA model had the least error in the Qatur river station with root mean square error of 0.809 meq/lit and 0.146 dS/m for SAR and EC parameters, respectively. Finally, the evaluation of the utilized parameters indicated that chloride plays an important role in predicting SAR and EC parameters, and if sulfate is not measured, it can be used as a promising alternative.

Keywords

Main Subjects



Articles in Press, Accepted Manuscript
Available Online from 20 November 2024
  • Receive Date: 29 July 2024
  • Revise Date: 13 November 2024
  • Accept Date: 20 November 2024
  • Publish Date: 20 November 2024