Application of Artificial Neural Network Models to Estimate Water Levels in Large Area of Everglades in South Florida

study and analysis of water level behavior helps to understand the behavior of any hydrological process in a watershed. Water resources management studies include several methods and techniques to understand different hydrological phenomena among those data driven models (DDM) is one. Solomatine et.al. (2009) mentioned that, DDM is based on data analysis of a system, in order to find connection between different system variables without explicit knowledge of the physical behavior of the system. In fact, Loucks et. al. mentioned that, the applications where computational efficiency is crucial or if there is lack of understanding in underlying relationship DDM serves better. In this paper, application of DDM has been done to forecast water levels. Study area considered for the present research was Everglades which is a complex natural ecosystem in Florida, USA spread across a wide area. This area is inhabited by the fauna which is extremely sensitive to minor changes in the water level. Such situations makes the water level prediction crucial to protect the ecosystem. In this study, artificial neural networks (ANN) technique was used with backpropagation algorithm was used to build model. The area was then divided within different cluster models which were built analyzing their influence on group of water level measuring gages located in the area.