葉面積指數(Leaf Area Index, LAI)是森林生態系中能量轉換的重要影響因子,在早期傳統推估LAI調查方法上既費時費力且花費成本高,且無法推估大範圍面積的LAI空間分佈資料,近年來藉由遙感探測影像之光譜特性,進行大面積LAI空間性資料推估為LAI量測發展趨勢。目前MODIS(Moderate-resolution Imaging Spectroradiometer)影像具有高時間解析力及提供免費檔案下載優勢,可供即時森林生態系監測之用。本研究利用野外調查之LAI資料與福衛二號影像之多種植生指標,進行LAI推估模式之建立,研究結果以NDVI冪次模式LAI=8.8997(NDVI)1.4937,相關係數為0.74,可作為推估LAI空間為依據。使用多期MODIS影像繪製全台灣LAI空間分布圖,並分別探討不同林型之LAI長期性及季節性變化。研究結果顯示LAI會因不同林型、天然災害、人為因素與環境因子改變而有所不同,從長期趨勢變化探討溫度及雨量對LAI的影響,顯示2001-2009年間台灣山區溫度上升約0.3~0.6℃,雨量則顯示乾旱年如2002年,2002年為高溫少雨之極端年份,其對於全台陸域植生LAI呈現下降趨勢。本研究以迴歸分析,進行各林型LAI長期變化趨勢推估,結果顯示10年間各林型之LAI有逐年下降之趨勢。季節性之LAI變化探討,結果顯示MODIS所推估之LAI可以反應林分之季節性變化。
Leaf Area Index (LAI) is one of the most important factors in forest ecosystem for energy fluxes. Traditional direct measurements of LAI are usually costly and time-consuming, and are unable to cover large areas. In recent years, there has been a tendency of estimating large-scale LAI based on the multiple spectral characteristics of satellite images. The NASA website also constantly provides free MODIS (Moderate-resolution Imaging Spectr- oradiometer) time series data useful for monitoring forest ecosystem. In this study, the relationships between the field measured LAI data and the vegetation indices derived from FS2 satellite images were examined. The result showed that Normalized Difference Vegetation Index (NDVI) was a good indicator of LAI. The regression equation, LAI=8.8997(NDVI)1.4937 (R2=0.74), was used to estimate the spatial distribution of LAI, and was subsequently applied to the MODIS imageries to analyze the spatial patterns and temporal trends. The results showed that the LAI varied with different forest types, natural disasters, human factors and environmental factors. Observed from the annual variability of LAI changes, the LAI of all forest types decreased as the temperature raised 0.3~0.6 ° C from 2001 to 2009. In terms of precipitation, the drought in 2002 significantly decreased the LAI of land vegetation. The results of time series analysis showed that there was a long-term decreasing trend in LAI for all forest types annually. The LAI derived from MODIS also clearly depicted the seasonal fluctuations of the forests.