Modelling Heavy Metal Concentration in Kainji Hydropower Reservoir, Nigeria

ABSTRACT The primary function of impounding water at the Kainji Dam is to generate hydropower, while it also offers opportunities for development of fishery, irrigation, recreation and navigation. The operation of the Kainji dam reservoir in Nigeria over the years has concentrated solely on power generation with little or no cognizance given to environmental aspect. The objectives of the study were to model concentration of selected heavy metals in sediment of the reservoir using Multilayer Perceptron Neural Network (MLPNN) and Radial Basis Function Neural Network (RBFNN) were used to model sediment quality parameters at selected location in the reservoir. The results showed that the application of the NN approaches to model the sediment quality parameters gives good outcomes for all the selected locations with strong correlation coefficient of at least 0.72. Also the Root Mean Square Error (RMSE) and Mean relative Error (MRE) are found within acceptable limits. Concentrations of Cu2+, Pb2+ and Cr3+ in sediment vary between 27.9 to 68.6 mg/kg, 21.9 to 35.0 mg/kg and 39.7 to 95 mg/kg respectively. These fell within the tolerance limits of 80, 35 and 95 mg/kg stipulated by Washington Department of Ecology Sediment Quality Guidelines (WDOESQG) for concentration of heavy metals in sediment. The modelling approaches can be adopted in similar studies.