Modeling of upwelling early warning system using water quality sensor device and automatic weather system integrated with hybrid FIS GA

It is a common fact that one of the adverse effects of upwelling in lakes is fish die-offs in floating net cages or other forms of aquaculture. A solution to prevent or reduce this unfortunate effect of upwelling is to create an upwelling early warning system that can predict the probability of upwelling events within a certain period. This research proposes such a system that can measure the parameters for upwelling prediction. The proposed system uses a sensor device that combines a water quality sensor device and an Automatic Weather System (AWS), integrated with the hybrid Tsukamoto fuzzy inference system using genetic algorithm (hybrid FIS GA) method. The sensor device used in this system has been successfully tested in laboratory and field experiments. From the test results, the water quality sensor successfully measures the water quality parameters, namely water pH, Dissolved Oxygen (DO), Oxidation-Reduction Potential (ORP), Electrical Conductivity (EC), and Resistance Temperature Detectors (RTD) or water temperature with 80% accuracy. Moreover, the AWS succeeds in measuring weather parameters, namely wind speed wind direction, rainfall, and air temperature with 100% accuracy. These measurements produce data for the hybrid FIS GA method to predict upwelling events.