Predicting water hyacinth expansion and identifying environmental drivers in Lake Tana using machine learning

This paper investigated the intensity and causes of water hyacinth expansion in Lake Tana, Ethiopia, using machine learning techniques. Water hyacinth poses a significant threat to freshwater ecosystems by disrupting ecological balance, degrading biodiversity, and affecting the livelihoods of communities that depend on lake resources. This research uses 20 years of weekly laboratory-based experimental data collected at 27 sampling sites around Lake Tana to model the relationship between water hyacinth expansion and key environmental drivers. Four conventional machine learning and five deep learning models were evaluated. Their performance was assessed using the coefficient of determination (R 2 ), mean absolute error (MAE), and root mean squared error (RMSE). The random forest machine learning model achieved R 2 , MAE, and RMSE values of 0.99, 4.70, and 14.95, respectively, which are the lowest among the trained models. Thus, the random forest model has achieved the best predictive accuracy. The most influential environmental predictors, total nitrogen, total phosphorus, chlorophyll-a, and pH, were identified. Consequently, the study concluded that integrating machine learning techniques into freshwater ecosystem monitoring and management can improve early detection, strengthen conservation strategies, and support efforts to mitigate further environmental degradation, including eutrophication, water pollution, invasive weed species, oxygen depletion, and climate change.