Application of Artificial Neural Network to Define Water Quality in Zaribar Lake in Kurdistan of Iran

摘要

Utrification is one of the most important phenomenons for water quality in lakes and dams which causes the decrease of Dissolved Oxygen in water and endanger to aquatic life and reduction of water quality. Therefore controlling this phenomenon is vital. For this purpose, having enough data and analyses and interpretation of it can be helpful for water quality management. In this study, some water quality data in Zaribar Lake in Kurdistan North West of Irons is analyzed using Artificial Neural Network Multilayer Perception (MLP) type . Twelve years, monthly Total Phosphorus(TP), Total Nitrogen(TN) and BOD parameters were analyzed and effects of the TP and TN for prediction of amounts of pollution obtained. The results show that the MLP with the TP and TN parameters inputs presents an acceptable prediction for BOD which existing BOD data and predicted ones have acceptable conformity with correlation coefficient R=0.9.Also, the rate of error for predication of BOD when TP parameter was input of the MLP is less than TN as input to the network, it may be that TP has effective role to increase pollution in the lake than TN. Hence, applying the MLP may be helpful for water quality management in the lake.