應用3S地球空間技術於植群空間分布型態之探討

Application of 3 S Geo-spatial Technology to Examine the Spatial Distribution Pattern of Vegetation

植群生態學在探討植物社會分布之法則，並尋求植群分布與環境變數之關係。傳統植群生態調查在樣區數目、大小、位置等因素限制下，常造成研究結果的偏差。本研究應用3S技術－－地球空間資訊系統、遙感探測及全球定位系統，探討中興大學惠蓀林場台灣杜鵑及紅檜空間分布型態，分析兩者在海拔、坡度、坡向與坡面位置的分布型態，並以x2及K-S檢測評估四者之相對重要性。研究顯示二者於海拔之相對重要性最高，坡度和坡面位置相近層次，而坡向最小，且海拔遠高出另三者。它們較能適應中海拔日照強、士壤淺薄、乾旱貧瘠之中坡或上坡到山脊之環境，大致與傳統研究相符。惟取樣必須完全採行多樣區、多樣本分散分布的方法，使樣本更具代表性，以避免統計檢測偏差，尤其是因植群樣本於海拔垂直分布之涵蓋不全所造成。後續研究將發展日照及土壤溼度等環境變數，從而建立模式以推估植群潛在生育地。 Vegetation ecology is to examine the rules that govern the distribution of plant community and to seek the relationship between the vegetation distribution and environmental variables. Conventional forest ecology studies have been subjected to statistical biases owing to the constraints of the size, number, and location of sampling plots, thereby making the inferences from different studies for the same vegetation often contradictory to each other. The study was attempted to examine the spatial pattern of Formosan rhododendron and Taiwan red cypress in the Huisun Experimental Forest Station using 3S geo-spatial technology: geo-spatial information system (GIS), remote sensing (RS), and global positioning system (GPS). Specifically, it was to analyze the spatial distribution of the two tree species on four topographic factors: elevation, slope, aspect, and terrain position and to assess their relative importance for each species by using Chi-square test and K-S test. For the two species, the relative importance of elevation is the highest, slope and terrain position are the second and close to each other, aspect is the lowest, and elevation is much more important than the other three factors. In comparison with other tree species, they may be more adaptable to the poor environment from a steeper midslope or upslope to a ridge top, with elevations from 1,000 to 2,400 m, receiving higher light intensity, with thinner, drier, and less fertile soils. The analysis results are generally consistent with the conventional studies. However, sampling must fully take a multi-plot, multi-sample dispersion approach so that the samples can be more representative, thereby avoiding statistical biases due to a scarcity of some portions of data for a habitat factor, particularly the vertical distribution of a tree species on elevation. Follow-up studies will develop proxy indicators for environmental variables, such as solar radiation and soil moisture, and develop the model based upon these variables to predict the potential habitat of the two species.

Vegetation ecology is to examine the rules that govern the distribution of plant community and to seek the relationship between the vegetation distribution and environmental variables. Conventional forest ecology studies have been subjected to statistical biases owing to the constraints of the size, number, and location of sampling plots, thereby making the inferences from different studies for the same vegetation often contradictory to each other. The study was attempted to examine the spatial pattern of Formosan rhododendron and Taiwan red cypress in the Huisun Experimental Forest Station using 3S geo-spatial technology: geo-spatial information system (GIS), remote sensing (RS), and global positioning system (GPS). Specifically, it was to analyze the spatial distribution of the two tree species on four topographic factors: elevation, slope, aspect, and terrain position and to assess their relative importance for each species by using Chi-square test and K-S test. For the two species, the relative importance of elevation is the highest, slope and terrain position are the second and close to each other, aspect is the lowest, and elevation is much more important than the other three factors. In comparison with other tree species, they may be more adaptable to the poor environment from a steeper midslope or upslope to a ridge top, with elevations from 1,000 to 2,400 m, receiving higher light intensity, with thinner, drier, and less fertile soils. The analysis results are generally consistent with the conventional studies. However, sampling must fully take a multi-plot, multi-sample dispersion approach so that the samples can be more representative, thereby avoiding statistical biases due to a scarcity of some portions of data for a habitat factor, particularly the vertical distribution of a tree species on elevation. Follow-up studies will develop proxy indicators for environmental variables, such as solar radiation and soil moisture, and develop the model based upon these variables to predict the potential habitat of the two species.