台湾浊水溪冲积扇地表变形监测雷达干涉技术与资料融合方法

The Rate of Subsidence of the Choushui River Alluvial Fan (Taiwan) Based on Persistent Scatterer Interferometry and Leveling Data

浊水溪冲积扇为台湾目前最严重的地层下陷区,已设置包括DInSAR、GPS、水准测量、磁环分层式监测井与地下水位井等多元监测系统。大范围监测地层下陷方法中的水准测量有取样性不足之缺点,而DInSAR也易受相位不相关与大气误差影响而致精度降低。本研究利用2006~2008年期间共20幅ENVISAT卫星影像,搭配PSI技术,有效降低DInSAR的误差影响,同时获得107.6像素/km2的取样密度,弥补了水准测量取样密度仅0.2个/km2的取样性不足,透过295个监测点交叉验证,PSI与水准测量的成果显示两者的均方根误差为0.6cm。本研究亦发展了资料融合方法,有效结合水准测量与PSI成果,实验结果显示融合后的成果更能精确展现下陷的范围与下陷中心的变化,同时该成果与水准测量之均方根误差降至0.4cm。

With abundant paddy fields and fishing ponds,the Choushui River Alluvial Fan（CRAF） is one of the most important agricultural areas in Taiwan;however,it is subject to severe subsidence caused by groundwater abstraction.We use 20 ENVISAT images,taken between 2006 and 2008,to determine the rate of vertical deformation over CRAF using persistent scatterer interferometry（PSI）,which effectively reduces the errors associated with conventional Differential Interferometric Synthetic Aperture Radar（DInSAR） techniques.The pixel density over CRAF is 107.6 pixels/km2,compared with 0.2 points/km2 for the leveling benchmarks.Rates of vertical displacement derived from PSI and leveling are in good agreement,root mean square（RMS） error is 0.6 cm/year,but PSI provides a higher spatial resolution.We also develop a method of combining the two data sources that treats the high-precision,low point-density leveling data as a smoothed correction of the PSI data.The merged field represents the overall subsidence pattern and finer details better than the original sources alone,and has an associated RMS error of 0.4 cm/year when compared with the leveling data.