Vegetation Classification and Mapping of the Upstream Watershed of Eliaonan River

• Main Authors: Chien-Chun Liao, 廖健均
• Other Authors: Ching-Long Yeh
• Format: Others
• Language: zh-TW
• Published: 2008
• Online Access: http://ndltd.ncl.edu.tw/handle/32482165165588568397

碩士 === 國立屏東科技大學 === 森林系所 === 96 === The studies were collecting the literature of vegetation ecology at Eliaonan River upstream watershed, as well as investigated the region that lack of vegetation data by quantitative methods. Furthermore, we combined the historical plots with field data for analysis. According to the vegetation classification system of Taiwan, the name and vegetation types in this area was re-evaluated. The vegetation key and the real vegetation map draw at Alliance level were provided for ecology management and conservation. The vegetation map was mapped (or predicted) using the Generalized Addition Model, to examine the accuracy of prediction. In the aspect of vegetation mapping, a total of 84 quantitative plots and 7 environmental factors were collected. There are 5 formations, 5 alliances and 7 associations when classified by DCA, Cluster Analysis, fidelity of species and vegetation synoptic table. The result of CCA show that the significant factors which affect the distribution of vegetation and species are topography, altitude, stone and WLS. Vegetation mapping was making at alliance level which was obtained from the result of vegetation classification. There were 24 representative vegetation categories, Pasania kawakamii ALL., Rhododendron formosanum ALL., Carpinus kawakamii ALL., Chamaecyparis formosensis-Neolitsea acuminatissima ALL., Tsuga chinensis var. formosana ALL., secondary forest, buildings and roads, stream and bare area.Among them, the biggest alliance of natural forest and the artificial forest unit are Pasania kawakamii ALL and Acacia confusa ALL., which occupy 1,648.98 ha (41.17%) and 783.55 ha (19.56%), respectively. In the forecast vegetation map aspect, it demonstrated that the accuracy of major alliance can above 50%, excepting the vegetation the group Rhododendron formosanum ALL. and Carpinus kawakamii ALL. After reposition of forecast vegetation, the accuracy of vegetation map will increased obviously. The result implies that it was feasible to forecast vegetation map by using the Generalized Additive Models, but it must be proved by lots of studies. Although the model cannot forecast vegetation map correctly, it can apply on draft drawing. Furthermore, referring to the vegetation key for vegetation mapping will increase the speed of vegetation mapping.