Estimation of Windbreak Changes in Yi-Lang Coastal Zone Based on Land Use Distance Variables

臺灣位於東亞季風帶，冬季東北季風及夏季的颱風風速強勁，經常對臺灣的海岸地區造成嚴重的風害，因此海岸地帶以栽種防風林來抵禦強風的侵襲。然而，由於海岸土地的開發所導致的環境變遷，造成海岸防風林分布範圍的變化及組成的差異，而反過來影響海岸地帶土地利用的品質及可能的開發潛力。因此若能掌握防風林的變遷趨勢，有效實施防護措施，則可確保防風林的穩定成長，達到維護海岸環境品質的目的。本研究假設防風林邊緣的土地利用變遷與其鄰近土地利用的空間分佈情況有關，因此利用鄰近土地使用的變遷趨勢，可以進一步推測防風林可能發生的變動。首先利用宜蘭縣三個鄉鎮三版土地利用調查資料，分析防風林帶邊緣的變遷與鄰近土地利用分佈的關係。設計土地利用距離變數描述1982年與1988年林地邊緣各種土地利用在林地邊緣不同距離處的出現機率。再以計算所得的土地利用距離便數值，進行兩種區別分析取得對應的區別函數。一為僅以1988年林地是否變動為依據的林地變動區別函數，另一種則為以單種土地利用的變動為依據的個別土地利用變動區別函數。然後利用1988年的各種土地利用距離變數值，分別以兩種區別函數去進行林地變遷的推估。結果發現兩類區別函數對三鄉鎮防風林的變遷推測準確度均高於52％，而蘇澳鎮更高達84％，顯示此種方法有其發展和應用上的價值。Taiwan is located in the East Asia Monsoon Climatic Zone. Coastal zones in Taiwan are constantly subjected to strong northeasterly winds in the winter and destructive typhoons in the summer. As a result, Coastal windbreak plantations are very common in these areas. The existence of a coastal windbreak belt has proven beneficial to crop yields, and effective in preventing the inland migration of coastal sand dunes. However, coastal zone development has resulted in environmental changes, which in turn influence the growth and distribution of coastal windbreaks. Furthermore, the changes in the condition and distribution of coastal windbreak have impacts on the environmental quality and landuse potential of the coastal zone areas. It is hypothesized that the boundary shifting of windbreaks is closely related to the spatial land use pattern in their neighborhood. The study collected land use digital data of 3 Yi-Land Counties derived from surveys conducted in 1982, 1988 and 1994. Based on the data, the probability for each land use type to occur within a certain distance from a grid cell in the windbreaks boundary was calculated, and the data were used for discriminant analysis to derive two sets of discriminant functions. One is to differentiate whether or not a grid cell in the windbreak boundary changed its land use during the period between 1982 and 1988, and the other is to differentiate whether the cell changed to particular type of land use. The accuracy rates of both methods for the three districts of Yi-Lang County were all above 52%, while the best estimations were achieved at the Su-Ao District, which reached 84%. The proposed methods proved to be potentially useful for estimating local, small-scale land use changes