Prediction of water quality under the impacts of fine dust and sand storm events using an experimental model and multivariate regression analysis

Abstract

Many regions of the world, especially arid and semiarid areas, occasionally experience fine dust and sandstorms, known environmental problems that make normal life difficult. Since the intrusion of large amounts of dust into treatment plants may significantly change the water quality indices, the main goal of this study was to estimate these indices during the events, which can help decision-makers to improve water quality. To achieve relationships using nonlinear multivariate regression analysis, a long-term (three years: April 2017–February 2020) experimental study of water quality parameters including total dissolved solids (TDS), hydrogen content (pH), electrical conductivity (EC), chlorine (Cl), total hardness, sodium (Na), and magnesium (Mg) for water samples from wastewater treatment plants in Sistan region (Iran) was conducted where is one of the most popular regions in the world with high amount of annual fine dust level. Analysis of ANOVA showed that of all the independent parameters considered in this study, water quality parameters strongly correlated with monthly mean sand and dust storm index (SDSI), wind speed, temperature, and the number of monthly windy days. For the regression analysis, 25 months of data were used for the simulation process and 10 months for validation. The final results showed that the relationships obtained from the nonlinear multivariate regression analysis could predict the water quality indices very well (with R2 more than 0.75) except for Mg with R2 equal to 0.55. In addition, the maximum mean relative error belongs to Mg (10.8%) and then Na (9.9%) whereas the minimum mean relative error belongs to pH (2.6%) and then EC (2.9%).