群集分析法應用於大安溪事業區林地分類之研究

An Application of Clustering to Forestland Classification － An Illustration of Da-An-Shi Working Circle

林地分類為森林經營之基礎，為能有效發揮森林的各項功能，有賴良好林地分類體系的建立。本研究之試驗區域位於東勢林區管理處所轄之大安溪事業區為範圍，採用網格（200 m×200 m）作為基礎單元，萃取相關屬性資料，並利用群集分析法中階層式最小變異數法做同質性的歸群，最後根據各群所佔之比例來定位各林班或小班之分類適宜性，並由套疊921 後崩塌地圖層來探討分類的適宜性，所得之結果將比傳統的分林地分類方式更具合理性。 Forestland classification is a basic task for forest management. In order to bring the functions of forest into full play, it may rely on a well forestland classification system. The study area located in the Da-An-Shi working circle of Tung-Shi forest district. We used the grid (200 m×200 m) as basic units to extracts the related attribute data, and then applied to cluster analysis techniques of Minimum-Variance method. According to the cluster ratio it can evaluate the suitability of compartment or subcompartment. By overlaying the map of collapse in 921, we could find it will more reasonable than traditional forestland classification.

Forestland classification is a basic task for forest management. In order to bring the functions of forest into full play, it may rely on a well forestland classification system. The study area located in the Da-An-Shi working circle of Tung-Shi forest district. We used the grid (200 m×200 m) as basic units to extracts the related attribute data, and then applied to cluster analysis techniques of Minimum-Variance method. According to the cluster ratio it can evaluate the suitability of compartment or subcompartment. By overlaying the map of collapse in 921, we could find it will more reasonable than traditional forestland classification.