Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4375487 | Ecological Informatics | 2006 | 9 Pages |
Abstract
This study demonstrated prediction of breeding success of Black-tailed Gulls in relation to the selected environmental factors through evaluation of relative importance in determining breeding success. The data were obtained from the 258 selected and 120 non-selected sites for breeding of the gulls during the breeding periods in 2002-2003. Breeding success at the selected sites, and environmental factors such as vegetation cover, vegetation height, rock cover, nest-wall, nearest distance between neighbors and slope, were measured at each sampling site. For predicting breeding success of Black-tailed Gulls, we used two different artificial neural networks in this study: self-organizing map (SOM) and multilayer perceptron (MLP). SOM was used to classify the sampling sites based on the environmental factors, whereas MLP was implemented to prediction of breeding success of the gulls at the non-selected sites based on environmental conditions. In our results, SOM discriminated clearly the sampling sites and presented differences in environmental factors between the selected and non-selected sites. Subsequently, the breeding success was accordingly predicted by MLP. Nest-wall was considered the most important environmental factor in determining survival status of the gulls. An increase in nest-wall and vegetation cover was required to support breeding of the specimens for managing the habitats for Black-tailed Gulls.
Keywords
Related Topics
Life Sciences
Agricultural and Biological Sciences
Ecology, Evolution, Behavior and Systematics
Authors
Who-Seung Lee, Young-Soo Kwon, Young-Seuk Park, Tae-Soo Chon, Jeong-Chil Yoo,