نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشگاه آزاد اسلامی واحد اهواز
2 استاد گروه خاکشناسی، واحد علوم و تحقیقات، دانشگاه آزاد اسلامی، تهران، ایران
3 استاد گروه مهندسی آب، دانشگاه شهرکرد، شهرکرد، ایران
4 دانشیار گروه آب، واحد اصفهان، دانشگاه آزاد اسلامی، اصفهان، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Monitoring and ensuring the quality of river water is critical for environmental and agricultural sustainability. In this study, we aimed to reduce the parameters for assessing water quality by employing Principal Component Analysis (PCA) and simulating the impact of land use and rainfall on various water quality attributes of the Zayandeh-Rood River using Artificial Neural Networks (ANN). Water quality data from 12 stations along the river were collected and PCA was applied to reduce the dataset influencing water quality. After extracting land use and rainfall data within the watershed, we modeled the impact of these factors on certain water quality attributes (4608 data points) using the backpropagation algorithm and calculated the Mean Squared Error (MSE). PCA analysis revealed that the first six components accounted for 94.78% of the total variance. ANN training results showed a determination coefficient (R²) of 0.79 and RMSE of 0.20. The R² for SAR, TDS, EC, and Na were 0.88, 0.84, 0.83, and 0.83, respectively, demonstrating high prediction accuracy. The Nash-Sutcliffe Efficiency (NSE) values were above 0.6 for all measured factors except pH, CO3, and flow rate (NSE < 0.5), indicating the model’s effectiveness. The results highlighted that reduced factors in PCA matched the highest R² in ANN, establishing PCA as a useful tool for better managing river water quality with fewer measurements.
کلیدواژهها [English]
)