利用遙測估算農作物蒸發散量之研究

Abstract

本研究嘗試以Roerink等人（2000）建立的衛星遙測估算旱作蒸發散量SEBI(Simplified Surface Energy Balance Index)模式，應用於估計美國SGP’97試區的熱通量和蒸發散量，應用時加上大氣校正模式，以及由衛星影像作土地利用分類。並對台中霧峰農試所2000 年一期作水稻田利用SPOT衛星多光譜資料測試估計效果。在SGP’97試區方面，利用地面實測資料和衛星觀測建立回歸關係推估的結果較良好；各項能量通量估計值與觀測值間的誤差，大部分源自地表參數的反演誤差，而後者又與大氣校正模式或回歸關係的誤差有關。在農試所水稻田試區方面，由於SPOT衛星影像缺乏熱紅外光波段，因此利用SPOT 觀測資料反演的NDVI，透過Landsat衛星影像資料建立的回歸關係推估地表溫度，進而推估能量通量。但因此反演的地表參數間具有某種相關性，且水田的所有反射率一溫度資料均落於蒸發散控制區，不能建立模式所需熱通量分配的圖形。本研究也利用DMSV-1空載多光譜資料分析模式對水稻田的可行性，熱通量分配圖沒有某種程度相關，初估結果合理，大氣影響較小，待未來有測量資料時作進一步的比對分析。 This work intends to implement the Simplified Surface Energy Balance Index (5-SEBI) model by Roerink et al.(2000) to estimate the heat fluxes during The Southern Great Plain (SGP’97 in short) Hydrological Experiment in 1997. Two additional features are added to the 5-SEBI model in the work, performing atmospheric correction by radiative transfer model and land-use classification. Results show that partitioning of heat fluxes is in accordance with the measurement. Main cause of discrepancy is due to inaccurate surface parameter estimates. To estimate evapo-transpiration in rice paddy, surface temperature retrieval is compulsory. Since SPOT image is lack of thermal channel, this work uses the Landsat data to construct the regression function between NDVI and surface temperature. However, it is found that the derived surface temperature is collinear with the SPOT retrieved reflectivity. Also, it is realized that all reflectivity-temperature data of rice paddy should fall in the evaporation-controlled region. Application of satellite data for evaporation estimation in rice paddy requires further research.