Effective Water Quality Monitoring Using Machine Learning and Satellite Imagery

Strategic Insights -

Researchers at the Graduate University of Advanced Technology have developed an innovative approach to monitor water quality and detect dead zones using satellite imagery and machine learning algorithms. The study employed seven AI models to estimate dissolved oxygen levels in Long Island Sound, New York, and found that the Extreme Gradient Boosting (XGBoost) model showed the highest prediction accuracy and lowest uncertainty. This breakthrough has the potential to enable efficient, scalable, and cost-effective monitoring of water quality and dead zones, informing decision-making for environmental management and ecosystem conservation.

Key Takeaways:

  • The research used Landsat 9 satellite imagery and seven AI models (AdaBoost, XGBoost, CatBoost, SVM, KNN, RF, and M5 Model Tree) to estimate dissolved oxygen levels in Long Island Sound, New York.
  • XGBoost showed the best performance among the AI models, offering the highest prediction accuracy and lowest uncertainty.
  • The study demonstrated the potential of integrating satellite-based remote sensing with AI models for water quality monitoring and dead zone detection.
  • The results of the study can inform decision-making for environmental management and ecosystem conservation.
  • The study has been peer-reviewed and published in Marine Pollution Bulletin.

Statistics:

  • 7 AI models were employed in the study (AdaBoost, XGBoost, CatBoost, SVM, KNN, RF, and M5 Model Tree).
  • 11 Landsat 9 bands were used to derive spectral properties for water quality assessment.
  • The study used observational DO data and spectral properties from Landsat 9 satellite imagery.
  • The XGBoost model showed the highest prediction accuracy (97.5%) and lowest uncertainty (5.2%) among the AI models.

Sources:

  • Water quality monitoring for coastal hypoxia: Integration of satellite imagery and machine learning models. Marine Pollution Bulletin, 2025;222:118735.
  • Graduate University of Advanced Technology, Department of Water Engineering, Faculty of Civil and Surveying Engineering.
  • Pergamon-elsevier Science Ltd, The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, England (Publisher contact information).