Water Quality Assessment and Forecasting Along the Liffey and Andarax Rivers by Artificial Neural Network Techniques Toward Sustainable Water Resources Management

This research evaluates water quality in two contrasting hydro-climatic regions: the River Liffey in Ireland and the Andarax River in Spain. It utilizes an Artificial Neural Network (ANN) to simulate potential changes in key water-quality parameters based on field measurements. The ANN models showed strong predictive efficiency and performance, achieving R2 values of 0.89 for dissolved oxygen (DO), 0.98 for electrical conductivity (EC), 0.87 for pH, 0.95 for total dissolved solids (TDS), and 0.96 for turbidity. The root mean-square-error (RMSE) values for important parameters were DO (1.25 mg/L), EC (48.06 µS/cm), and turbidity (8.9 FNU). The models were able to capture complex nonlinear relationships under different environmental conditions. The results showed that DO levels in the Liffey will decline by up to 20% over the next decade due to rising nutrient pollution, while TDS levels in the Andarax River are expected to rise by approximately 15% during the same period as a result of ongoing agricultural runoff. The study also simulated potential future hypothetical scenarios by applying the model to four different “what-if” situations. Overall, the research underscores the significance of machine learning in understanding intricate water-quality dynamics.