本研究以林試所蓮華池試驗林之蛟龍溪集水區為試區，旨在結合Markov模式與SRES氣候預設情境，探討未來土地利用變遷與氣候變遷對GWLF模式流量模擬的影響，並評估其對集水區流量之衝擊。研究方法係先數化集水區1971及1998年土地利用型圖，並根據該資料建立Markov模式和預測其未來土地利用變遷情形，結果顯示民墾地面積由1971年的0.36%和1998年的2.40%，增加為2025年的3.29%、2052年的6.13%和2079年的8.87%，明顯呈現上升趨勢。至於整合Markov模式模擬結果與SRES情境修正之氣候變遷資料，供分析土地利用變遷與氣候變遷對GWLF模式流量模擬之影響、以及評估其對集水區流量的衝擊結果指出，蛟龍溪集水區在利用Markov模式修正未來土地利用參數後，其月平均流量、年總流量及年平均流量無論在未來短期、中期或長期之模擬結果均有所變化，同時根據土地利用與氣候變遷資料，該集水區在豐水期發生暴雨與洪水的頻率將大幅增加，反之在枯水期時，河川流量將變為乾旱。
The Markov model and the special report on emissions scenarios (SRES) climate scenario were applied to investigate the influence of future land-use changes and climate changes on the streamflow simulation of the generalized watershed loading function (GWLF) model, and also to assess the impacts on the Jiao-Long watershed, of central Taiwan. Land-use maps in 1971 and 1998 were first digitized. The Markov model was then constructed according to those land-use data and applied to predict future land-use changes. Results of the Markov prediction indicated that the area of unlawful cultivation would increase from 0.36% in 1971 and 2.40% in 1998 to 3.29% in 2025, 6.13% in 2052, and 8.87% in 2079. In addition, the climate change data retrieved from the SRES were finally integrated with the Markov predictions in order to analyze the influence of future land-use changes and climate changes on the streamflow simulation of the GWLF model, and also to assess the impacts on the watershed. Results showed that, regardless of whether considering the short-, middle- or long-term simulation, values of monthly average streamflow, annual total streamflow, and annual average streamflow all changed. Furthermore, land-use changes and climate changes would lead to increases in the occurrences of storms and floods in the wet season, and drought in the dry season.
