以雷達干涉技術偵測地表變形之研究

Abstract

合成孔徑雷達干涉技術(Interferometric Synthetic Aperture Radar, InSAR)為利用不同時間或不同位置的雷達天線所獲得的兩幅或多幅複數影像，再藉由相位值的差異量來量取地表的三維資訊。目前主要的應用包含數值地形模型(Digital Elevation Model, DEM)的產生、地震變形量(Deformation)的研究、地層下陷(Subsidence)的量測及火山監控、冰川漂移等。 台灣地區因位於地震帶上，地震發生的頻率相當頻繁，特別是一些較大規模的地震，常造成地表的嚴重變形現象，而威脅到人民生命財產的安全，如何有效的探討及監控其變形狀況，遂為一重要的研究課題。在變形量的監控上，傳統所使用方法的是以精密測量設備來進行，但因受限於儀器的精度及人力、物力的需求，所以不太容易進行大範圍、高精度、高密度的測量。雷達差分干涉技術(Differential InSAR, D-InSAR)利用不同時期的衛星影像進行偵測的處理，影像內容涵蓋的範圍大，且可以獲得高精度地表變動量的量測結果，是非常適合的偵測工具。 因為合成孔徑雷達干涉處理的步驟多，過程複雜不易了解，因此許多研究人員均以昂貴的套裝軟體進行處理。為有效的推廣干涉處理技術在各個領域的應用，本研究以公用程式為主，期有興趣及需求的使用者能廣泛應用。由於處理的過程繁複，為有效掌握處理結果，本研究探討處理的流程及產生結果的差異，並提出處理步驟中不同的處理方法及討論不同方法間的差異，藉以選擇較佳的處理模式。

為瞭解雷達干涉技術在台灣地區進行變形偵測的可行性，本研究以台灣中部地區集集地震及屏東平原地區的地層下陷兩不同區域、不同變形量特性案例進行研究分析。在集集地震所產生的地表偵測部分，因為台灣中部地區繁密的植物覆蓋及地震引起的變動產生不相關性，使得斷層東邊地區因低相關性而無法偵測變動情形；但斷層西部因地勢較平坦且變動量亦較小，因此以干涉技術可以清楚的偵測出其變化情形。D-InSAR與GPS資料都具有高精度的偵測結果，在垂直方向的比較結果均方根誤差為3.3 cm，兩者具有相當的一致性，因此在地震的變形偵測中，兩者的量測資料可以有良好的互補作用。 屏東平原位於台灣西南部，由於地體構造背景特殊，屬於快速沉降與堆積的區域，又由於近年來在沿海地區超抽地下水造成嚴重的地層下陷情形，因此，本研究利用ERS衛星雷達影像的差分干涉技術來偵測屏東地區的地表變形量，並與GPS資料及水準測量資料做比較，探討D-InSAR技術在台灣地區地層下陷偵測的精確度，並分析屏東地區的下陷狀況。實驗結果說明由D-InSAR技術計算的下陷量平均值與GPS及水準的下陷量平均值差異量均小於1 cm，且變化的趨勢一致；從干涉圖中可以明顯的看出林邊溪河口附近海岸為下陷最嚴重的地區，且該屏東平原地區在乾、濕季的下陷量明顯的不同，乾季的的下陷量大於濕季的下陷量。

Interferometric Synthetic aperture radar (InSAR) is an imaging technique for measuring the topography of surface, its changes over time, and other changes in the detailed characteristics of the surface. A radar sensor above the Earth detects tiny changes on the ground by very accurately measuring changes in the time delay, or phase, of a radar echo. Interferometric synthetic aperture radar and its spatially dense, accurate deformation measurements have advanced studies of the Earth’s crust. This technique has been widely used to produce digital elevation model (DEM) and to measure terrain deformation. This technique has also been applied to measure ocean currents, map hazards, and detect glacier motion. Taiwan lies at the boundaries of the Philippine Plate to the east and the Eurasian Plate to the west. The earthquakes occur often in Taiwan area and may cause heavy casualties and building damages. It is an important issue to determine the magnitude of deformation and monitor the fault movement. Researchers can avoid the limitations of the survey approaches and can begin to think about measuring the entire area’s deformation. Differential InSAR (D-InSAR) and its spatially dense, accurate deformation measurements have advanced studies of the Earth’s crust. A significant advantage of this technique is that it provides a comprehensive view of the motion detected for the entire area affected. It is expected that this type of result will supplement ground-based measurements, which are made at a limited number of locations. The processes of D-InSAR are complicated and difficult to be understood. To effectively promote the applications of D-InSAR, this study mainly applies the public programs to modeling the processes. This study emphasizes on the processes of D-InSAR and the differences of the results in order to effectively realize the experiment results. Moreover, we propose many alternative methods in the steps of the processes, and discusses the differences among them. Based on the comprehensive discussions among the alternative methods, a best method can be selected to apply on the actual practices. To realize the possibility of applying D-InSAR on the deformation detection in Taiwan, this study analyzes two particular cases located on different regions and characterized with different deformations. One case is the earthquake on 21 September 1999 near Chi-Chi in central Taiwan. The other case is the land subsidence in the Pingtung plain. We chose ERS radar images to detect the Chi-Chi earthquake’s deformation by D-InSAR technique. The data were used to generate a high-resolution, wide-area map of displacements in flat or semi-flat areas. The interferograms show radar line contours indicating line-of-sight changes corresponding to surface displacements caused by earthquake ruptures. These results were compared to synthetic interferograms generated by GPS data. Using a combination of both GPS and D-InSAR techniques, this study showed that the rms value calculated from the differences in vertical displacements between the methods was 3.3 cm. Thus, both techniques can be used to detect that magnitude of displacement. Both D-InSAR and GPS techniques have high precision. In our study, the results from using both techniques showed good correspondence, suggesting that these two approaches can be used as complementary tools. The Pingtung Plain of south-western Taiwan is located in a fast subsidence and deposition area due to its special tectonic setting. Serious land subsidence has occurred recently along coastal regions of Taiwan as a consequence of over-pumping of underground water. We used D-InSAR to detect land subsidence in the Pingtung Plain, southern Taiwan, between 1995 and 2000. The mean error was within 1 cm with comparison with GPS and levelling data. The results revealed that the critical subsidence region was located on the coast near the mouth of Linbien River. Experimental results indicated that subsidence was significantly higher during the dry season than during the wet season.