ESEL Paper Review_20131122
By Hong Guo
Phone: (+82) (0)10 82276568
1, Title and Author
Title: An ANN application for water quality forecasting
Journal: Marine Pollution Bulletin
Sundarambal Palani*, Shie-Yui Liong, Pavel Tkalich
Tropical Marine Science Institute, National University of Singapore, 14 Kent Ridge Road, Singapore 119223, Singapore
2. Summary of Paper
? In this study, artificial neural networks (ANNs) were used to predict and forecast quantitative characteristics of water bodies. The true power and advantage of this method lie in its ability to (1) represent both linear and non-linear relationships and (2) learn these relationships directly from the data being modeled. ? The study focuses on Singapore coastal waters. The ANN model is built for quick assessment and forecasting of selected water quality variables at any location in the domain of interest ? The results show the ANN’s great potential to simulate water quality variables. Simulation accuracy, measured in the Nash?Sutcliffe coefficient of efficiency (R2), ranged from 0.8 to 0.9 for the training and overfitting test data. Thus, a trained ANN model may potentially provide simulated values for desired locations at which measured data are unavailable yet required for water quality models.
? Temperature model results
? The results show that adequate temperature prediction/ forecasting was obtained with temperature data as the only input. The neural network is able to simulate the water temperature with an accuracy of a degree or less
? DO model results
? Typical DO forecasting results are shown in Fig. 6b for the training, overfitting test, and validation data sets. There was a small amount of uncertainty in the DO prediction and
forecasting for some field measurements and ANN-predicted values, but this was likely due to unknown local factors. Even though there was uncertainty in the training and overfitting test sets during model construction, the performance accuracy of the developed ANN DO prediction and forecasting model
? This model could be used in parallel with physics- based models as a new prediction tool. It can identify important parameters for enabling both selective physical/chemical monitoring and quick water quality assessment of Singapore seawater. The limitations of this study include its limited data set.
? The lack of fit between the observed and estimated data indicates that new patterns must be incorporated into the model, and thus the model should be recalibrated and revalidated as more data are collected.
This research would be fully studied and applied to the research of the modeling for the food supernatant assessment
5. Contact (Mail address): firstname.lastname@example.org (S. Palani).