應用MODIS影像反演土壤含水率－以泰國MaeSa集水區為例

近年來熱帶地區國家，正逢快速的經濟發展，面臨大量的土地利用改變以增加農作面積或轉換成工業用地，為因應大量農業與工業的用水，水資源的分配與管理，逐漸成為重要的議題。在水的循環過程中，土壤含水率扮演著重要的角色之一，獲得土壤含水率的變化資訊，有助於強化水文模式，及了解不同土地利用覆蓋下的水資源的分布情形，對於分析未來的土地使用變更計劃所造成的水資源變遷，都可提供有用的資訊。本研究利用遙測技術結合熱慣量法，推估泰國北部MaeSa集水區內之土壤含水率。在遙測技術方面採用2002至2009年的美國中尺度影像光譜儀(Moderate Resolution Imaging Spectroradiometer, MODIS)及可與影像實際結合的表觀熱慣量法(Apparent Thermal Inertia, ATI)分別推估土壤深度10公分、100公分及200公分的土壤含水率。比對MODIS的推估值與地面觀測站的實測值的結果：10公分、100公分及200公分相關係數分別為0.80，0.84，及0.84。土深10公分、100公分及200公分的納許-史托克利夫效率係數值(Nash-Sutcliffe efficiency coefficient, E)，分別為0.57，0.537及0.492；而均方根誤差在各深度分別為0.055，0.025及0.029。綜合統計結果，土壤深度為100公分的含水率推估結果最佳。本研究成功地使用MODIS影像推估MaeSa集水區的土壤含水率，未來將可有效使用此方法於其他測站缺乏的集水區，以提供更豐富的土壤含水率資訊於相關領域管理應用。

關鍵字

MODIS ；遙測；土壤含水率；熱慣量

並列摘要

Tropical countries nowadays are at the economic developing stage. They are facing the rapid land use/ land cover change and increasing water demand for agriculture and industry. Therefore, hydrological issue becomes more and more important in these regions than ever before. In this study, we used thermal inertia approach with remote sensing technique to estimate the root zone soil moisture content, which plays a key role of hydrological cycle, in the tropical watershed of northern Thailand. The main purposes are focused on monitoring soil moisture content change which is part of an alert system for water supply and understanding the water distribution pattern corresponding to above land cover type and that is useful for land-use planning in the future. We analyzed different depths including 10cm, 100cm and 200cm root zone soil moisture contents (SMC) derived from the modified apparent thermal inertia approach (ATI) with Moderate Resolution Imaging Spectroradiometer (MODIS) imagery from 2002 to 2009. The results showed that the Pearson correlation coefficient and Nash-Sutcliffe efficiency coefficient between ground truth and retrieved SMC under all different land cover sites for three depths were respective above 0.80, 0.84, 084 and 0.57, 0.537, 0.492. The retrieved precision referred to Root Mean Square Error (RMSE) value were 0.055, 0.025 and 0.029 (m^3. M^(-3)) for respective three depths. The 100cm depth root zone SMC showed the optimal estimation in response to the upper precipitation characteristics and various land cover types. The successful root zone SMC estimation by using optical remote sensing data will still have a strong contribution for the tropical or subtropical mountain studies in the future.

並列關鍵字

MODIS ； Remote Sensing ； Soil Moisture content ； thermal inertia

國際替代計量

應用MODIS影像反演土壤含水率－以泰國MaeSa集水區為例

全文下載

主題瀏覽