CJEES

Home
Peer Review
Editorial Board
Instructions
Online First
Latest Issue
Past Issues
Contact
Impact Factor
Reject Rate

 
You are here: Home » Past Issues » Volume 7, 2012 - Number 4 » IDENTIFICATION OF INLAND EXCESS WATER FLOODINGS USING AN ARTIFICIAL NEURAL NETWORK


« Back

Boudewijn VAN LEEUWEN, Gábor MEZŐSI, Zalán TOBAK, József SZATMÁRI & Károly BARTA
Department of Physical Geography and Geoinformatics, University of Szeged, POB 653 Szeged, H-6701,Hungary leeuwen@geo.u-szeged.hu, mezosi@geo.u-szeged.hu, tobak@geo.u-szeged.hu, szatmari@geo.u-szeged.hu, barta@geo.u-szeged.hu

IDENTIFICATION OF INLAND EXCESS WATER FLOODINGS USING AN ARTIFICIAL NEURAL NETWORK

Full text

Abstract:

Inland excess water is a partly natural and partly human induced phenomenon where areas are flooded with water that cannot find its way gravitationally to rivers or channels. These inundations cause large financial and environmental damage in the flat regions of the Carpathian basin. To understand where and why the inundations occur can help to take preventive measures and to reduce loses. Inland excess water is caused by a complex and interrelated set of factors. To study these factors, a new approach using a combination of an artificial neural network (ANN) and a geographic information system (GIS) has been developed. This article presents and evaluates the results of this approach. The network is integrated in a workflow that starts and ends using multiple spatial data sets in a GIS. The intermediate steps – the training and simulation of the ANN – are performed using a mathematically modeling environment which is controlled from within the GIS. This framework allows for the flexible use of different spatial data sets and experimentation with the settings of the neural network. The training validation shows that the relief is the most important factor in the study area, while other factors like distances to anthropogenic objects are of less importance. The simulation results show that the ANN – GIS framework is capable of accurately identifying inland excess water floodings.


Keyword: Inland excess water, artificial neural networks, GIS


(c) 2006 - 2017, Earth and Environmental Team
Design by Adrian Dorin