林木健康指標之建立與效度評估

Validity Analysis and Development for the Index of Tree Health

近年來，對於森林健康監測（Forost Health Monitoring, FHM）之研究越來越受到重視，許多國家陸續進行大面積的森林健康狀況調查。本研究主要以阿里山事業區內的不同林型為研究對象，透過地面調查健康指標之建立進行地面樣區之每木調查，調查的項目包括葉面積指數、樹冠密度、樹冠透視度、樹冠梢枯等，研究中將上述變數量化成為森林健康指標，藉此評估分析樣區之健康狀態。本研究針對森林健康指標項目進行效度評估，利用誤差矩陣表之Kappa值進行效度評估分析，結果顯示，健康指標調查項目之總體精度介於72.91%-93.81%之間，其中以冠層狀態之調查準確度最高Kappa值為0.929、活冠層比例之調查準確度最低Kappa值為0.637；而在整體林分健康度之評估中，總體分類準確度及Kappa值分別為92.86%及0.826。本研究結果顯示，研究中所建立之健康指標可有效的評估林分之健康狀態。 In recent years, the research on forest health monitoring (Forest Health Monitoring, FHM) has been receiving increasing attention; many of countries have been conducting large areas of forest health survey. The purpose of this research is according to different forest type as the research object in the Alishan Working Circle. We have used a repeatable system to assess several parameters (Leaf Area Index, Crown density, Crown Transparency, Crown Dieback, and unusual phenomenon of other growth), and the above-mentioned parameters quantified to health indicator. This study is aimed at the validity assessment of forest health indicators in the Alishan plantation, utilizing en or matrix Kappa value to proceed the assessment of validity analysis. The results showed that the overall accuracy of health indicator value ranged from 72.91% to 93.81%. It also showed that the crown defoliation had the highest degree of accuracy (Kappa value=0.929) and the live crown ratio had the lowest degree of accuracy (Kappa value=0.637). This result of study shows that set up healthy indicator can effective assessment states of forest health.

In recent years, the research on forest health monitoring (Forest Health Monitoring, FHM) has been receiving increasing attention; many of countries have been conducting large areas of forest health survey. The purpose of this research is according to different forest type as the research object in the Alishan Working Circle. We have used a repeatable system to assess several parameters (Leaf Area Index, Crown density, Crown Transparency, Crown Dieback, and unusual phenomenon of other growth), and the above-mentioned parameters quantified to health indicator. This study is aimed at the validity assessment of forest health indicators in the Alishan plantation, utilizing en or matrix Kappa value to proceed the assessment of validity analysis. The results showed that the overall accuracy of health indicator value ranged from 72.91% to 93.81%. It also showed that the crown defoliation had the highest degree of accuracy (Kappa value=0.929) and the live crown ratio had the lowest degree of accuracy (Kappa value=0.637). This result of study shows that set up healthy indicator can effective assessment states of forest health.