Integrating CFD modelling, neural networks and remote sensing: controlled prediction of chlorophyll-a concentration in the Mejillones of South Bay
The geomorphologic characteristics of the Mejillones of South Bay (The Bay) in northern Chile influence the oceanic dynamics inside the Bay. The presence of the Günther tropical undercurrent brings tropical water to the region, whereas the Humboldt polar current brings cool water to the same area. The combination of geomorphology and currents generates conditions responsible for producing a powerful upwelling phenomenon in the Bay. A novel approach that integrates both remotely sensed data and in situ information, including oceanic variables and bathymetry, is proposed, which when combined with a computational fluid dynamic (CFD) model predicts patterns of surface distributions of variables such as chlorophyll-a concentration and temperature. The proposed system uniquely generates a controlled map via the application of digital image processing and an artificial neural network.