中文摘要：

利用GPS数据改正InSAR大气影响中，GPS站网的低空间密度是限制改正精度的一个主要因素．文中引入大气传输模型（ATM）来考虑大气状态在时间上的演化，并兼顾了风向的估计，把GPS-ZWD（GPS湿延迟）时间序列观测值转换为空间上分布比较稠密的GPSZWD网络，然后利用该稠密网络在空间上内插产生水汽延迟图，来更好地模拟InSAR影像获取时刻的水汽场，提高大气改正精度．实验结果表明，在研究地区有限的数据条件下（即只有6个GPS点的情况下），GPS＋ATM算法在重现大气信号能力方面比单纯的使用影像获取时刻的GPS-ZWD数据要强，并且其对长波的大气误差去除更明显；用来做实验的三幅差分干涉图中，GPS＋ATM算法对两幅以长波信号为主的干涉图中的大气影响分别降低了21.7％和22．6％，比仅使用SAR过境时刻GPS—ZWD数据时的结果分别改进了4．5％和8．7％．而对以短波信号为主的干涉图没有明显提高．

英文摘要：

Abstract It is usually believed that the sparse spatial resolution of GPS stations is a main limitation factor for improving correction accuracy when we use GPS-ZWD data to mitigate atmospheric effect in repeat-pass InSAR measurement. In this paper, a simple atmospheric transport model （ATM） is induced to consider a temporal evolution of atmospheric state, GPS- ZWD time series around SAR overpass time can be transformed into a spatially denser network of GPS-ZWD control points by using the ATM after wind velocity is estimated. The denser network of GPS-ZWD enable estimation of ZWD difference maps to match better atmospheric phase distortions observed in the InSAR data. The experimental data analysis from the three differential interferograms of Shanghai demonstrates GPS ＋ATM method can reproduce the atmospheric signal with stronger ability than the method only using the GPS-ZWD data recorded at the radar observation times, and the reduction for long-wavelength signal is much remarkable under the condition with limited data resources of research area （only six GPS station）. After applying GPS＋ATM algorithm, the global root mean square （RMS） atmospheric phase fluctuations of the two interferograms dominating by the long wavelength signal, decrease by 21. 7% and 22. 6% respectively, with a improvement of 4.5% and 8.7% compared with the former method,but no improvement in another one dominating by the short-wavelength signal.