Please use this identifier to cite or link to this item:
標題: 協合遙測與數值地形模型於台灣冷杉生育地之模擬
Synergism of Remote Sensing and Digitall Terrain Model for Abies kawakamii Habitat Simulation
作者: 李旻旻
Lee, Min-Min
關鍵字: Abies kawakamii
Remote Sensing
Digital Terrain Model
Shelter Terrain
出版社: 森林學系
摘要: 台灣冷杉(Abies kawakamii)主要分佈在本省 3,000m 之高山區,是台灣主要河川之 發源地o 因此冷杉之分佈對高山地區的水土保持極具重要性o GIS 在近二十年來已成為協 助森林資源重要工具,而遙測綜觀覽要與多持攝像優點提供 GIS 資料庫更新最經濟有效 來源o 本研究目的係結合遙測與 DTM 探討台灣高山區冷杉之生態與分佈,並且建立 GIS 模式推測冷杉之潛在生育地o DTM 產生全區與冷杉之高程,坡度,坡向及水分梯度統計值 ,並與 Landsat TM 衛星影像套疊分析累計頻率圖及直方圖,以產生最佳之冷杉辨識波段 o本研究建立兩種 GIS 預測模式,分別為等權模式與非等權之卡方模式o 等權模式中每一 生育地因子之權重相等,非等權模式各因子之權重係以卡方檢定計算而得o 模式準確度評 估以模式建立用樣本及非模式建立用樣本完成o 卡方檢定結果顯示海拔為生育地之重要因 子;坡度,坡向及水分梯度為次要因子,模式準確度評估結果亦是如此o 屏障區假說可用 以解釋為何坡及水分梯度在冷杉生育地模式中微不足道o 重要的是模式加入 TM 影像導出 波段,其預測準確度較 DTM 模式穩定,顯示遙測與數值地形模型之結合,於冷杉生育地 之模擬,確實能產生協力增效作用o
Abies kawakamii are mainly distributed over Taiwan mountain areas with altit- ude around 3,000m where are the place of origin for Taiwan major streamsof fir- st order. Due to rugged terrain of Taiwan high mountain areas, landslide occu- rs frequently. Therefore, the distribution of Abies is veryimportant for soil and water conservation in Taiwan mountain areas. GIS hasbeen becoming an imp- ortant tool to and facilitate forest resource management.Remote sensing, provi- ding a synoptic view and multitemporal imaging for a large area is a cost-effe- ctive source for timely, accurate update of a GIS data base. The objectives o- f this study were to investigate the ecology anddistribution of Abies kawakami- i in Taiwan mountain areas by combining remotesensing and DTM data, and to pre- dict potential habitat for Abies. This studyuse DTM data to generate the stat- istics of elevation, slope, aspect, and moisture gradient for the entire area and Abies areas, as well as Landsat TMimage to identify the best discriminatin- g band between Abies and non-Abies viaspectral pattern analysis and cumulative- frequency graphs. Intervals within each habitat parameter were defined and a ssigned a score according to theirfrequency of occurrence at extant sites. Tw- o predictive GIS models were developed: an equally weight model and a chi-squa- re model. Each habitat parameter was weighted equally for an equal weight mode- l. For an unequal-weight chi-square model, the importance of each habitat par- ameter was evaluatewith a chi-square test. Two types of accuracy assessment w- ere performed usingthe Abies samples in model development and the Abies sample s not in model development. Chi-square test results indicate that elevation i- s the majorhabitat parameter, whereas slope, aspect, and moisture gradient wer- e the minorones. Model accuracy assessments provide the same results as those of the chi-square test. The topographic sheltering hypothesis can effectively explainwhy both aspect and mousture gradient factors become trivial parameter s for Abies habitat modeling. Of more signigicanceis that DTM models of eleva- tion and TM derivative band stability than DTM model. The results indicate the tsynergism of remote sensing and DTM can be achieved to improve the accuracyof- predicting Abies potential habitat.
Appears in Collections:森林學系



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.