Predicting dissolved oxygen levels in aquatic ecosystems using machine learning models

Öz

This research investigates the performance of several machine learning algorithms in forecasting dissolved oxygen (DO) levels in the Brisbane River, utilizing physicochemical parameters alongside water flow data. We examined algorithms such as Linear Regression, Support Vector Regression, Random Forest, Gradient Boosting, XGBoost, and K-Nearest Neighbors, employing evaluation metrics including Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), R-squared (R²), and Mean Absolute Percentage Error (MAPE). Among the models, ensemble techniques, particularly Random Forest and XGBoost, exhibited enhanced predictive accuracy and robustness in identifying intricate, non-linear relationships. Analysis revealed that key variables, including pH, salinity, and specific conductance, were significant predictors, a finding corroborated by the correlation matrix. This study underscores the promise of machine learning, particularly ensemble approaches, in improving water quality monitoring and management, providing valuable insights for ecological sustainability and informed policymaking.

Anahtar Kelimeler

• Machine Learning Algorithms
• Aquatic Ecosystems
• Brisbane River
• Water Quality

Makine öğrenimi modelleri kullanarak su ekosistemlerindeki çözünmüş oksijen seviyelerinin tahmini

Öz

Bu araştırma, Brisbane Nehri'ndeki çözünmüş oksijen (DO) seviyelerini tahmin etmek için çeşitli makine öğrenimi algoritmalarının performansını, fizikokimyasal parametreler ve su akış verileri ile birlikte incelemektedir. Lineer Regresyon, Destek Vektör Regresyonu, Random Forest, Gradient Boosting, XGBoost ve K-En Yakın Komşu algoritmaları değerlendirilmiş; Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), R-squared (R²), and Mean Absolute Percentage Error (MAPE) gibi değerlendirme metrikleri kullanılmıştır. Modeller arasında, özellikle Random Forest ve XGBoost gibi topluluk (ensemble) teknikleri, karmaşık ve doğrusal olmayan ilişkileri belirlemede üstün tahmin doğruluğu ve dayanıklılık sergilemiştir. Analizler, pH, tuzluluk ve spesifik iletkenlik gibi anahtar değişkenlerin önemli öngörücüler olduğunu ve bu bulguların korelasyon matrisi ile desteklendiğini ortaya koymuştur. Bu çalışma, makine öğreniminin, özellikle topluluk yaklaşımlarının, su kalitesinin izlenmesi ve yönetiminin iyileştirilmesindeki potansiyelini vurgulamakta ve ekolojik sürdürülebilirlik ile bilinçli politika oluşturma için değerli içgörüler sunmaktadır.

Anahtar Kelimeler

• Makine Öğrenimi Algoritmaları
• Su Ekosistemleri
• Brisbane Nehri
• Su Kalitesi

Kaynakça

1. F. Granata, S. Zhu, and F. di Nunno, “Dissolved oxygen forecasting in the Mississippi River: advanced ensemble machine learning models,” Environmental Science: Advances, 2024, doi: 10.1039/d4va00119b.
2. Z. Wang, Q. Wang, Z. Liu, and T. Wu, “A deep learning interpretable model for river dissolved oxygen multi-step and interval prediction based on multi-source data fusion,” Journal of Hydrology, vol. 629, p. 130637, Feb. 2024, doi: 10.1016/J.JHYDROL.2024.130637.
3. H. Lim, H. Shin, J. Lee, J. Do, I. Song, and Y. Jin, “Prediction of Dissolved Oxygen Factor at Oncheon Stream Watershed Using Long Short-Term Memory Algorithm,” Water (Switzerland), vol. 16, no. 17, Sep. 2024, doi: 10.3390/w16172363.
4. F. H. Garabaghi, S. Benzer, and R. Benzer, “Modeling dissolved oxygen concentration using machine learning techniques with dimensionality reduction approach,” Environmental Monitoring and Assessment, vol. 195, no. 7, pp. 1–23, Jul. 2023, doi: 10.1007/S10661-023-11492-3/FIGURES/21.
5. M. Rajesh and S. Rehana, “Impact of climate change on river water temperature and dissolved oxygen: Indian riverine thermal regimes,” 123AD, doi: 10.1038/s41598-022-12996-7.
6. https://waterdata.usgs.gov/monitoring-location/07374000. [Accessed: 13-September-2024].
7. Liu, W., Lin, S., Li, X., Li, W., Deng, H., Fang, H., & Li, W. (2024). Analysis of dissolved oxygen influencing factors and concentration prediction using input variable selection technique: A hybrid machine learning approach. Journal of Environmental Management, 357, 120777, doi: 10.1016/J.JENVMAN.2024.120777.
8. S. Shadkani, Y. Hemmatzadeh, A. Saber, and M. Mohammadi Sergini, “Enhanced predictive modeling of dissolved oxygen concentrations in riverine systems using novel hybrid temporal pattern attention deep neural networks,” Environmental Research, vol. 263, Dec. 2024, doi: 10.1016/j.envres.2024.120015.

9. Li, Q., He, J., Mu, D., Liu, H., & Li, S. (2025). Dissolved Oxygen Modeling by a Bayesian-Optimized Explainable Artificial Intelligence Approach. Applied Sciences, 15(3), 1471, doi: 10.3390/app15031471. [10] https://www.busan.go.kr/ihe/index. [Accessed: 13-September-2024].
10. Y. Hu, C. Liu, and W. M. Wollheim, “Prediction of riverine daily minimum dissolved oxygen concentrations using hybrid deep learning and routine hydrometeorological data,” Science of The Total Environment, vol. 918, p. 170383, Mar. 2024, doi: 10.1016/J.SCITOTENV.2024.170383.
11. N. Shaghaghi et al., “DOxy: A Dissolved Oxygen Monitoring System,” Sensors, vol. 24, no. 10, May 2024, doi: 10.3390/s24103253.
12. S. Lin, D. C. Pierson, R. Ladwig, B. M. Kraemer, and F. R. S. Hu, “Multi-Model Machine Learning Approach Accurately Predicts Lake Dissolved Oxygen With Multiple Environmental Inputs,” Earth and Space Science, vol. 11, no. 7, Jul. 2024, doi: 10.1029/2023EA003473.
13. S. Rajagopal, S. S. Ganesh, A. Karthick, and T. Sampradeepraj, “Environmental water quality prediction based on COOT-CSO-LSTM deep learning,” Environmental Science and Pollution Research, Sep. 2024, doi: 10.1007/s11356-024-34750-4.
14. X. Nong, C. Lai, L. Chen, D. Shao, C. Zhang, and J. Liang, “Prediction modelling framework comparative analysis of dissolved oxygen concentration variations using support vector regression coupled with multiple feature engineering and optimization methods: A case study in China,” Ecological Indicators, vol. 146, p. 109845, Feb. 2023, doi: 10.1016/J.ECOLIND.2022.109845.
15. K. Roushangar, S. Davoudi, and S. Shahnazi, “The potential of novel hybrid SBO-based long short-term memory network for prediction of dissolved oxygen concentration in successive points of the Savannah River, USA,” Environmental Science and Pollution Research, vol. 30, no. 16, pp. 46960–46978, Apr. 2023, doi: 10.1007/S11356-023-25539-Y/FIGURES/13.
16. M. M. Bolick, C. J. Post, M. Z. Naser, and E. A. Mikhailova, “Comparison of machine learning algorithms to predict dissolved oxygen in an urban stream,” Environmental Science and Pollution Research, vol. 30, no. 32, pp. 78075–78096, Jul. 2023, doi: 10.1007/S11356-023-27481-5/TABLES/7.
17. X. Wang, Y. Li, Q. Qiao, A. Tavares, and Y. Liang, “Water Quality Prediction Based on Machine Learning and Comprehensive Weighting Methods,” Entropy 2023, Vol. 25, Page 1186, vol. 25, no. 8, p. 1186, Aug. 2023, doi: 10.3390/E25081186.
18. https://www.kaggle.com/datasets/downshift/water-quality-monitoring-dataset/data [Accessed: 3-September-2024].
19. Jordan, M. I., & Mitchell, T. M. “Machine learning: Trends, perspectives, and prospects.” Science, vol. 349, no. 6245, pp. 255-260, 2015. doi: 10.1126/science.aaa8415.
20. Gauss, C. F. Theoria motus corporum coelestium in sectionibus conicis solem ambientium. Hamburg: Friedrich Perthes und I.H. Besser, 1809.
21. Cortes, C., & Vapnik, V. “Support-vector networks.” Machine Learning, vol. 20, no. 3, pp. 273-297, 1995. doi: 10.1007/BF00994018.
22. Chen, T., & Guestrin, C. “XGBoost: A scalable tree boosting system.” Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, Aug. 2016, pp. 785-794. doi: 10.1145/2939672.2939785.
23. Friedman, J. H. “Greedy function approximation: A gradient boosting machine.” Annals of Statistics, vol. 29, no. 5, pp. 1189-1232, 2001. doi: 10.1214/aos/1013203451.
24. Cover, T., & Hart, P. “Nearest neighbor pattern classification.” IEEE Transactions on Information Theory, vol. 13, no. 1, pp. 21-27, 1967. doi: 10.1109/TIT.1967.1053964.
25. Breiman, L. “Random forests.” Machine Learning, vol. 45, no. 1, pp. 5-32, 2001. doi: 10.1023/A:1010933404324.
26. Gauss, C. F. Theoria motus corporum coelestium in sectionibus conicis solem ambientium. Hamburg: Friedrich Perthes und I.H. Besser, 1809.
27. Willmott, C. J., & Matsuura, K. “Advantages of the mean absolute error (MAE) over the root mean square error (RMSE) in assessing average model performance.” Climate Research, vol. 30, no. 1, pp. 79-82, 2005. doi: 10.3354/cr030079.
28. Willmott, C. J., & Matsuura, K. “On the use of dimensioned measures of error to evaluate the performance of spatial interpolators.” International Journal of Geographical Information Science, vol. 20, no. 7, pp. 801-820, 2006. doi: 10.1080/13658810600661574.
29. Pearson, K. “Mathematical contributions to the theory of evolution—III.” Philosophical Transactions of the Royal Society of London, Series A, vol. 187, pp. 253-318, 1896.
30. Makridakis, S., Wheelwright, S. C., & Hyndman, R. J. Forecasting: Methods and Applications. John Wiley & Sons, 1998.