نوع مقاله : مقاله پژوهشی
موضوعات
عنوان مقاله English
نویسندگان English
Groundwater sources are essential for drinking water, agriculture, and industrial use. Assessing the quality of these sources, including water hardness, is a crucial aspect of hydrogeological studies. This research aims to estimate and model the level of groundwater hardness in the coastal areas of the Caspian Sea, which is vital for effective management planning and land preparation. We employed a hybrid support vector regression model enhanced by firefly and chicken swarm algorithms. For this purpose, we selected various quality parameters from Babolsar piezometer wells located in Mazandaran Province, which include bicarbonate, chloride, sulfate, magnesium, and calcium, as input parameters, with water hardness as the output parameter, analyzed on a monthly basis over a statistical period from 2012 to 2022. The performance of the models was evaluated and compared using the correlation coefficient, root mean square error (RMSE), and Nash-Sutcliffe efficiency coefficient. The results demonstrated that the hybrid support vector regression model combined with the firefly algorithm achieved the highest correlation coefficient of 0.980, the lowest RMSE of 0.115 ppm, an average absolute error of 0.087 ppm, and a Nash-Sutcliffe value of 0.950. Overall, our findings indicate that artificial intelligence hybrid models perform effectively in estimating groundwater quality parameters.
کلیدواژهها English