蘭陽平原於1991 年國道五號開工，2000 年農業發展條例修正通過後，因預期心理大量興建農舍，此一農地轉用與農地破碎現象已被重視，故本研究以1977 年至2006 年之蘭陽平原為例，並以系統生態學（system ecology）中的系統觀點（system approaches），來解構農地地主對於農地轉用建地（農舍）過程，影響農地轉用地因子，並考量農地破碎化程度、土地價格、可轉用量等條件，加以建立及驗證蘭陽平原農地轉用假說。
本研究經文獻回顧提出蘭陽平原農地轉用初步假說及其系統圖（system diagram）後，經訪談加以修正蘭陽平原農地轉用假說，共八項，並按農地轉用因子的影響期間，提出1977 年、1995 年的農地轉用系統圖，供圖資分類及遴選地景指數之依據。經統計軟體─主成分分析，加以檢視農地破碎與轉用的因果關係，分別修正農地轉用系統圖，再藉由逐步迴歸分析，加以驗證農地轉用假說及農地轉用條件假說，並按驗證結果，分別提出1977 年、1995 年的全區、都市地區、非都市地區農地轉用系統圖。藉由此一滾動式反覆修正的過程，有助於討論過去文獻、專家學者、農地地主觀點及統計分析結果之差異性，且以系統圖方式表現，亦有助於檢視轉用因子於農地轉用過程的重要性程度差異。
農地轉用假說驗證結論如下：農地破碎化會影響農地轉用率，耕地（含旱作地、稻作地）具有解釋力。生活機能影響對農地轉用具重要影響力，但隨時間變遷影響因子不同。緊鄰住宅、商業、寬四米含以上的道路緊鄰、大專院校的農地，及離河流較近的農地，因景觀或投資效益佳，對於農地轉用具有解釋力。其中並以旱作地較容易受鄰居影響。土地價格會影響農地轉用，並以後期較為顯著。整體而言，農地轉用具有次要火車
站的衛星效應現象。農地轉用受所住宅數量影響，具群聚特徵，並以後期的都市地區影響較為明顯。可轉用廢耕地、稻作地面積比率會影響農地轉用率，並以全區與非都市地區受影響較明顯。整體而言，30 年間的蘭陽平原農地轉用，受後期（1995-2006 年）影響最多。
Since the Fourth Amendment of the gricultural Development Act (ADA) was passed in 2000, rapid increase in farmhouse development has led to agricultural landscape change and fragmentation in Lanyang Plain. The purpose of this thesis was to use a systems approach to study the important factors driving the decision for agricultural landowners to develop their
farm lands in Lanyang Plain during 1977‐2006.
Eight specific hypotheses were submitted for describing agricultural landscape change, and system diagrams were used to display the interrelationships between factors and agriculture landscape change in 1977, 1995 and 2006. We divided Lanyang Plain into 245
cells (cell size, 1500X1500 m2) and calculated the variables for each cell in ArcGIS. We also applied stepwise regression analysis to examined specific hypotheses regarding agricultural landscape change.
The empirical results showed that fragmentation of agricultural landscape will affect the rate of agricultural landscape change of arable land (including rainfed land and rice fields). Access to services (including road, residences and business) will influence agricultural landscape change rate, but the important factors change with time. Adjacency to residences, businesses, university and rivers, will improve agricultural landscape change rate, especially rainfed land, probably because these landscape types provide better scenery or higher investment expectations. Land prices will affect agricultural landscape change rate, and this effect is more significant in 1995‐2006. The distance to railway station will affect agricultural
landscape change rate. Agricultural landscape change is spatially clustered, especially in 1995‐2006. The percentage of land still available for development as regulated under the Agricultural Development Act in abandoned fields and rice fields will affect the rate of agricultural landscape change, especially in non‐urban areas. Overall, the major changes to agricultural lands in Lanyang Plain occurred during 1995‐2006.
