Artificial neural network based nowcasting model for beach water quality

Recreational water users may be exposed to elevated pathogen levels that originate from various point and non-point sources. Current daily notifications practice depends on microbial analysis of indicator organisms such as Escherichia coli (E. coli) that require 18-24 hours to provide sufficient response. This research evaluated the use of Artificial Neural Networks (ANNs) for real time prediction of E. coli concentration in water at Toronto beaches (Ontario, Canada). The nowcasting models were developed in combination with readily available real-time environmental and hydro-meteorological data during the bathing season (June-August) of 2008 to 2012. The results of the developed ANN models were compared with historic data and found that the predictions of E. coli concentrations generated by ANN models slightly outperforms than currently used persistence model with better accuracy. The best performing ANN models for each beach are able to predict approximately 74% to 82% of the E. coli concentrations.