Exploring the Relationship Between Chlorophyll-A and Other Water Quality Parameters by Using Artificial Neural Network Models: A Case Study of Lake Erie
Xue (Ashley) Hu, MLWS 2020
This study used a machine learning approach to model the water qualities of Lake Erie, Ontario, Canada. The data used in the modelling was obtained from the Environment and Climate Change Canada Agency for Lake Erie between 2000 and 2018 and included chlorophyll-a (CHLA), the dissolved oxygen (DO), total phosphorus (TP), total nitrogen (TN), temperature (T), pH, and depth. Several neural network (NN) models were selected for the data analysis, including the standard Neural Network (NN) model, the Simple Recurrent Neural Network (SRN) model, the Back Propagation Neural Network (BPNN) model and the Jump Connection Neural Network (JCNN) model. CHLA was selected as the key water quality indicators for eutrophication in Lake Erie. The above artificial neural network models were assembled. This study showed that the ANN ensemble model predicted the water quality of Lake Erie in a more timely and accurate manner, which aids in facilitating conventional water quality monitoring and reducing the risk of eutrophication in Lake Erie.