Abstract: This study aims to provide a method for developing artificial neural networks in estuaries as emulators ofprocess-based models to analyse bathing water quality and its variability over time and space. The methodology forecasts the concentration of faecal indicator organisms, integrating the accuracy andreliability offield measurements, the spatial and temporal resolution of process-based modelling, and the decrease in computational costs by artificial neural networks whilst preserving the accuracy of re-sults. Thus, the overall approach integrates a coupled hydrodynamic-bacteriological model previouslycalibrated withfield data at the bathing sites into a low-order emulator by using artificial neural net-works, which are trained by the process-based model outputs. The application of the method to the EoEstuary, located on the northwestern coast of Spain, demonstrated that artificial neural networks areviable surrogates of highly nonlinear process-based models and highly variable forcings. The resultsshowed that the process-based model and the neural networks conveniently reproduced the measure-ments ofEscherichia coli(E. coli) concentrations, indicating a slightly betterfit for the process-basedmodel (R2¼0.87) than for the neural networks (R2¼0.83). This application also highlighted that dur-ing the model setup of both predictive tools, the computational time of the process-based approach was0.78 times lower than that of the artificial neural networks (ANNs) approach due to the additional timespent on ANN development. Conversely, the computational costs of forecasting are considerably reducedby the neural networks compared with the process-based model, with a decrease in hours of 25, 600,3900, and 31633 times for forecasting 1 h, 1 day, 1 month, and 1 bathing season, respectively. Therefore,the longer the forecasting period, the greater the reduction in computational time by artificial neuralnetworks.
