臺灣北中部海岸保安林健康指標之研究

Health Indexes of Coastal Protection Forests in Northern and Central Taiwan

劉玲華(Ling-Hua Liu) ； 葉慶龍(Ching-Long Yeh) ； 王兆桓(Chao-Huan Wang)

《國立臺灣大學生物資源暨農學院實驗林研究報告》 21卷2期 (2007/06) Pp. 147-160

https://doi.org/10.6542/EFNTU.200706_21(2).0005

摘要

本研究以臺灣北中部海岸保安林為對象，測計與林木健康有關的變數，應用因素分析與鑑別分析，建立臺灣北中部海岸保安林的健康指標，並建立鑑別函數協助健康狀況的預測。結果顯示，因素分析歸類出5個林木健康因素：1.枝條活力指標；2.樹皮損傷指標；3.枝條蟲食指標；4.根部損傷指標；5.樹冠活力指標。上述5個因素的共通性總和佔總變異的92%，具相當高的解釋力，並用於林木分級。在原始變數及林分解釋變數逐步鑑別分析之鑑別函數擊中率達92%及88%，表示此鑑別函數的預測能力很高，此種分析程序不僅可以建立臺灣北中部海岸保安林的健康指標，同時可以評估個別林木的健康狀況，使經營者能以客觀有效的方式進行林木的保育及管理。

關鍵字

健康指標 ； 因素分析 ； 逐步鑑別分析

並列摘要

Variables related to tree health of coastal protection forests in Northern and Central Taiwan are measured and analyzed in this study. By conducting factor analysis and discriminate analysis, a discriminate function was constructed to predict health indicators of coastal protection forests in the Northern and Central Taiwan. Results indicate that there are five factors related to tree health: (1) branch vigor (2) bark damage (3) branch insect damage (4) root damage and (5) crown vigor. The above five factors explain as high as 92% of the cumulative total variance and then used to classify tree health. The result of stepwise discriminate analysis using original observed variables and stand variables shows that the hit rates are 92% and 88%, respectively, it represents that the constructed discriminate functions are highly predictable. The above procedure can construct health indexes of coastal protection forest in Northern and Central Taiwan. In addition, it can also evaluate health conditions for individual tree and assist managers to conserve and manage the forests by objective and effective ways.

並列關鍵字

Health Index ； Factor Analysis ； Stepwise Discriminate Analysis