Detection of Seasonal Changes in Heavy Metal Concentrations in the Soil-Irrigation Water System using Machine Learning
Detection of Seasonal Changes in Heavy Metal Concentrations
in the Soil-Irrigation Water System using Machine Learning
Sukiasyan Astghik,
Yesayan Patrik,
Kirakosyan Armen,
Ghazaryan Armine,
Simonyan Gevorg
Summary
Key words: soil, water, heavy metals, distribution, pollution, geo-ecology, environmental chain, modeling, PYTON/Simulink
Applying a balanced approach to predicting the spatio-temporal changes in ecosystem pollution allows us to assess their possible consequences. In this context, the study of changes in the concentration and probable migration pathways of heavy metals in the environment plays an important role. The aim of this work was to study the seasonal distribution of some heavy metals in the “soil-irrigation water” system using modern machine learning approaches. According to the results obtained, it is shown that the concentrations of V, Fe, Zn and Cu in irrigation water and soil samples vary at different times of the year. At the same time, a positive correlation is observed for the concentrations of V, Cu and Zn in water and soil samples, and a negative correlation for the concentration of Fe. Basically, the lowest values are observed in spring, while the highest distribution coefficient is observed for iron, and the lowest for copper. The machine learning method showed that the influence of geographical location on the concentration of metals in water is greater than the influence of the season. This may affect secondary processes, such as the rate of metal dissolution and sedimentation. The applied significance and novelty of the work is due to the fact that the distribution coefficients of metals iron, vanadium, copper and zinc in the soil-irrigation system are seasonal.