GIS and Artificial Neural Network-Based Approach for Integrated Management in Costa Caparica, Portugal

Angeliki Peponi

Master's thesis

Abstract:

Monitoring and managing populated coastal areas is a very important task which continues to be a difficult venture for interest groups. In this study, the use of Artificial Neural Networks and GIS are applied as a robust tool in integrated coastal zone management in the community of Costa Caparica, Lisbon, Portugal, by modeling and forecasting the erosion changes for the year 2021. Artificial Neural Networks present noteworthy advantages in comparison with other methods used for prediction and decision making in urban coastal areas. Multilayer perceptron (MLP) type of ANNs are used to conduct sensitivity analyses on natural and social forces, and dynamic relations in the dune-beach system of Costa Caparica, in order to give a prognosis of the erosion changes. Numerous tests were performed in order to train the network by changing the number of hidden layers and hidden layers nodes. The most suitable topology was one hidden layer with seven hidden nodes. The sensitivity analysis indicated that the variables amongst all the input variables to the model, which trigger the erosion changes in the study area, are the number of residents and households. The developed methodology appears fitted to reality; however some further steps would make it better suited to the purpose of the thesis.

Abstract:

Monitoring and managing populated coastal areas is a very important task which continues to be a difficult venture for interest groups. In this study, the use of Artificial Neural Networks and GIS are applied as a robust tool in integrated coastal zone management in the community of Costa Caparica, Lisbon, Portugal, by modeling and forecasting the erosion changes for the year 2021. Artificial Neural Networks present noteworthy advantages in comparison with other methods used for prediction and decision making in urban coastal areas. Multilayer perceptron (MLP) type of ANNs are used to conduct sensitivity analyses on natural and social forces, and dynamic relations in the dune-beach system of Costa Caparica, in order to give a prognosis of the erosion changes. Numerous tests were performed in order to train the network by changing the number of hidden layers and hidden layers nodes. The most suitable topology was one hidden layer with seven hidden nodes. The sensitivity analysis indicated that the variables amongst all the input variables to the model, which trigger the erosion changes in the study area, are the number of residents and households. The developed methodology appears fitted to reality; however some further steps would make it better suited to the purpose of the thesis.

Keywords

Spatial analysis, Land use conflict, Artificial Neural Networks, Integrated Coastal Zone Management, Decision making

Keywords

spatial analysis, Decision making, Land use conflict, Artificial Neural Networks, Integrated Coastal Zone Management

Language used: English

Date on which the thesis was submitted / produced: 19. 4. 2016

Thesis defence

• Supervisor: Ing. Vojtěch Barták, Ph.D.

Czech University of Life Sciences Prague

Faculty of Environmental Sciences

Master programme / field:
Landscape Engineering / Landscape Planning