Eutrophication analysis of water reservoirs by remote sensing and neural networks

Algal blooms of the water are an important variable for the analysis of freshwater ecosystems, which are relevant not only for human populations but also for plant and animal diversity. Monitoring algal blooms from space allows for a continuous and automatic control without the necessity of water sampling and human intervention. However, it is a very challenging task, which becomes particularly difficult when dealing with cyanobacteria blooms. Water limnology, satellite imagery and neural networks can be used as an ensemble of remote sensing and machine learning technologies in order to estimate the concentration of algal blooms from space. This paper describes empirical algorithms to this end, which incorporate information from the multi-spectral instrument of Sentinel-2 satellite. This approach is applied to the Cefni Reservoir (Anglesey, U.K.), by using spatial and temporal scales. Algae estimation is accomplished using different types of neural and fuzzy neural networks and the experimental results are very accurate, therefore proving the reliability and accuracy of the proposed approach for monitoring water reservoirs by using remote sensing and neural networks tools