2009-08-012024-05-18https://scholars.lib.ntu.edu.tw/handle/123456789/705199摘要：台灣雲霧森林（霧林）是一個生物多樣性非常高的生態系統，基於長時間高雲霧籠罩及潮濕的環境因子影響，雲霧的沈降量非常高且具有較低之蒸散量，所以此系統是下游生態系統重要的水分來源。根據跨政府氣候變遷小組（IPCC）的預測，全球氣候變遷將會對此台灣特有霧林及其優勢樹種台灣扁柏（Chamaecyparis obtusa var. formosana）構成極大的威脅，所以瞭解台灣霧林物質與能量的轉換，是一個非常重要的課題。本研究以宜蘭棲蘭山區霧林為樣區，用光學遙測的技術來：（一）估測台灣扁柏現今碳貯存（carbon stock）總量；（二）模擬棲蘭山區霧林年碳吸存淨量[又稱初級生產淨量(Net Primary Production, NPP)]。本研究將用影像分割及物件化的演算法把樹冠投射面積從高空間解析度QuickBird影像中計算出，在結合現地資料後，研究者將可計算出台灣扁柏的碳貯存總量。樹冠層孔隙度（gap fraction）將用光譜混合分析模型從QuickBird影像中粹取出，此參數可以利用在光合作用有效光的運算，將光合作用有效光和氣候及土壤資料套入CASA（Carnegie-Ames-Stanford Approach）NPP模型，再與現地資料及其他模型推導值做比較與校正後之後，本研究將可準確估算棲蘭山區霧林NPP。研究者將會分享研究成果並移轉這些技術至相關政府及學術單位，在與其他專家討論後，此研究可做為台灣其他生態系碳貯存量及NPP監測的一個基準。<br> Abstract: A cloud forest is an ecosystem covered by a high incidence of low-level cloud at the canopy level. It is generally found in a tropical or subtropical montane region. Because a cloud forest usually develops on the saddles of mountains, where moisture introduced by settling clouds is effectively retained, it becomes an important water source for its downstream ecosystems and a habitat for various floras and faunas. The Intergovernmental Panel on Climate Change (IPCC) stated that global climate change would have severe threats to endemic (local) species, which may include yellow cypress (Chamaecyparis obtusa var. formosana), the most dominant and endemic tree species in the Taiwan’s cloud forest. Therefore, it is important to understand the carbon dynamics of this unique ecosystem. The main objective of this research is using optical remote sensing techniques to estimate total carbon stock (aboveground + belowground) and net uptake [also known as the Net Primary Production (NPP)] of a yellow cypress forest in the Chi-Lan Mountain region of the northeastern Taiwan. An image segmentation algorithm will be applied on a high spatial resolution QuickBird satellite image to extract projected tree crown area of yellow cypress, and it will be converted to total carbon by referring to field measured allometry. The Carnegie-Ames-Stanford Approach (CASA) NPP model will be used to estimate annual net carbon uptake of the forest by integrating gap fraction derived from QuickBird multi-spectral bands (to predict absorbed photosynthetically active radiation); and climate and substrate data (to modify the maximum light use efficiency). CASA will be tuned and predicted NPP will be calibrated by referring to ground data and NPP derived by another model. Results will be shared and approaches will be discussed with related government agencies and academic institutions; the refined work procedure of proposed research can be a standard for estimating carbon budgets of other ecosystems in Taiwan.Carnegie-Ames-Stanford Approach (CASA)霧林樹冠層孔隙度影像分割初級生產淨量QuickBird光譜混合分析台灣扁柏Carnegie-Ames-Stanford Approach (CASA)cloud forestgap fractionimage segmentationNet Primary Production (NPP)QuickBirdspectral mixture analysisyellow cypress (Chamaecyparis obtusa var. formosana)