Application of Artificial Neural Networks and MATLAB Models for Assessing Pollution Levels in River Benue, Benue State, Nigeria.

The impact of pollution on freshwater resources has significant implications for public health and environmental sustainability. This study investigates the pollution levels of River Benue in Benue State, Nigeria, with the objective of evaluating its suitability for potable water supply. A comprehensive analysis was conducted using standard analytical techniques including titrimetry, gravimetry, potentiometry, and spectrophotometry to determine key physico-chemical parameters during both dry and rainy seasons. Results revealed seasonal variations in water quality indicators: total hardness ranged from 22.28 mg/L (dry season) to 41.87 mg/L (rainy season), while electrical conductivity measured 91.35 µS/cm in the rainy season and 54.10 µS/cm in the dry season. Total dissolved solids (TDS) were higher in the rainy season (76.02 mg/L) compared to the dry season (32.31 mg/L), indicating greater ionic presence during rainfall. Dissolved oxygen concentrations were 3.18 mg/L (rainy) and 3.72 mg/L (dry). Phosphate levels were 0.73 mg/L (rainy) and 0.85 mg/L (dry), while sodium concentrations were 2.78 mg/L and 2.95 mg/L, respectively. Calcium values were 1.57 mg/L (rainy) and 2.07 mg/L (dry). Chemical oxygen demand (COD) recorded 114.37 mg/L in the rainy season and 84.97 mg/L in the dry season. Biochemical oxygen demand (BOD) exceeded the World Health Organization (WHO) limits for drinking water, with values of 24.87 mg/L in rainy season and 25.63 mg/L in dry season). Elevated concentrations of heavy metals, including lead 0.00865 mg/L in dry season; 0.00281 mg/L in rainy season and copper 1.24 mg/L in dry season; 0.9734 mg/L in rainy season, poses potential health risks to aquatic life and communities relying on the river for domestic use. A Pearson Product Moment Correlation analysis yielded a coefficient (r) of 0.91, suggesting a strong positive correlation and relatively consistent water quality status across seasons. Furthermore, Artificial Neural Network (ANN) and MATLAB were employed to support the data analysis and predict pollution patterns. The integration of ANN techniques demonstrated high predictive accuracy, highlighting their effectiveness in water quality assessment and environmental monitoring.