چكيده به لاتين
Water quality monitoring is one of the most important activities in managing water resources. This process is conducted to assess the current state of water, identify changes over time, and predict future trends. Given the importance of water monitoring for providing drinking water, monitoring reservoirs is of great significance. To evaluate the water quality of the Wadi Al-Dhiyagh Dam in Oman, key parameters such as chlorophyll-a, dissolved oxygen, temperature, and turbidity were estimated over the period from January to December 2023. The use of modern technologies has enabled the collection of unprecedented environmental data. Utilizing technologies such as remote sensing and machine learning allows access to accurate information regarding various environmental parameters such as temperature, humidity, air pollution, and water quality. Satellites with high-resolution imaging capabilities and extensive coverage facilitate the monitoring of environmental changes on a large scale. Ultimately, combining these technologies with machine learning, a branch of artificial intelligence, enables systems to learn from data and make accurate predictions.
Using machine learning algorithms (Ada Boost, Bagging Regressor, Extra Tree, Random Forest, Gradient Boost, KNN, LGBM, and Decision Tree), the data collected from satellites can be analyzed to uncover hidden patterns. These patterns can be utilized for predictions. Satellite data from the Sentinel-2 sensor was used as the primary source of information in this study. By leveraging machine learning algorithms, the values of quality parameters at the water surface and the depth of the reservoir were estimated. Subsequently, using multicriteria decision-making methods, the best time and place for water extraction from the dam were determined.
The results of this research showed that the use of satellite data and machine learning can provide an efficient and accurate method for continuous monitoring of water quality in water resources. The results indicated that the temperature quality parameter achieved good results with individual models at the surface, while the other quality parameters performed well with combined models. The highest correlation coefficients for the estimates of the parameters were 0.84, 0.77, 0.99, and 0.98, respectively. Additionally, at the depth of the reservoir, the quality parameters of temperature and dissolved oxygen were estimated with individual models, while chlorophyll-a and turbidity were estimated with combined models, achieving high correlation coefficients of 0.97, 0.93, 0.89, and 0.97, respectively. In decision-making for water extraction from the intake tower located in the reservoir, multicriteria decision-making methods such as COPRAS, CODAS, and MABAC were used to determine the appropriate level
