CGRG Bibliography of Canadian Geomorphology
Search Results
Author : Han, G.; and Shi, Y.
Date : 2008.
Title : An Atlantic Canadian Coastal Water Level Neural Network Model (ACCSLENNT).
Publication : Water, Weather, and Climate: Science Informing Decisions. Canadian Meteorological and Oceanographic Society 2008 Congress. May 25-29, 2008, Kelowna, British Columbia.
Issue :
Page(s) :
Abstract
Coastal sea level information is essential for coastal zone management, navigation and oceanographic research. However, long-term sea level observations are usually available at a limited number of locations only. This study discusses a method based on Neural Networks (NN) to predict sea levels at a specified coastal site from the data gathered at other nearby or remote permanent stations. A simple 3-layered feed-forward back-propagation network and a neural network ensemble, named ACCSLENNT, were developed to establish the nonlinear relationship of sea level data among stations by learning historical characteristics between them. Hourly sea level observations at five stations along the coast of Atlantic Canada, St. John’s, Argentia, North Sydney, Belledune and Halifax, are used to formulate and validate the ACCSLENNT model. Qualitative and quantitative comparisons of the network output with target observations showed that appropriately trained NN models are able to provide robust long-term predictions of both tidal and non-tidal sea levels when only short-term sea level data are available. The results indicate that the NN models in conjunction with limited permanent stations are able to supplement long-term historical sea level data along the Atlantic Canadian coast. The approach can be readily applied to other coastal regions.
Bibliography of Canadian Geomorphology