Jj Zhu

3D coseismic Displacement of 2010 Darfield, New Zealand earthquake estimated from multi-aperture InSAR and D-InSAR measurements

Differential interferometric synthetic aperture radar (D-InSAR) measures ground deformation only along the line-of-sight (LOS) of the radar, which limits the capability of D-InSAR in investigating the surface damages and the focus mechanisms of earthquakes. We do a three-dimensional (3D) decomposition of the coseismic displacement of the Darfield, New Zealand earthquake that occurred on 3 September 2010 by exploiting the Multi-Aperture InSAR (MAI) and D-InSAR measurements from both ascending and descending L-band PALSAR data. Due to the dispersive nature of the ionosphere and the slight Doppler shift between the forward- and backward-looking interferograms, the ionospheric effects can be more serious in MAI measurements than in D-InSAR. We propose mitigating the ionospheric effects in the MAI processing with the directional filtering and interpolation procedure that has been applied in Offset-tracking. The rupture revealed by the 3D surface displacement fits closely to the Greendale fault, which is believed to be responsible for the earthquake. The horizontal ground motions, mostly eastwards in the hanging wall and westwards in the footwall, reached up to 2.5 m and are anti-symmetric with respect to the Greendale fault. Up to 2.5 m subsidence occurred in the hanging wall, while uplift is found in the footwall with an extreme case of 1.6 m in the far left of the fault. This makes us conclude that the Greendale fault is a normal and dextral strike-slip. It is seen that the MAI measurements are very helpful in the derivation of 3D coseismic displacement fields as it provides more accurate displacement estimation in the north–south direction.

Characterizing sudden geo-hazards in mountainous areas by D-InSAR with an enhancement of topographic error correction

Differential interferometric synthetic aperture radar (D-InSAR) has been viewed as a promising technique in monitoring sudden geo-hazards (e.g., earthquake and landslide) in mountainous areas. However, the tough natural settings (e.g., steep slopes and vegetation) pose the D-InSAR technique to face many challenges. Among them, phase residuals induced by inaccurate topographic heights that can result in intolerable error have not been paid adequate attention. We present, in this paper, a new strategy of using D-InSAR measurements to characterize sudden geo-hazards with an emphasis on the correction of topographic errors. In the proposed strategy, a least squares model with an outlier detector is constructed to estimate the topographic errors from multi-baseline wrapped differential interferograms, and the error-prone phase unwrapping procedure is not needed. The new strategy is applied to the ALOS PALSAR images acquired for monitoring a giant mudslide occurred in Zhouqu County, China. After refining the topographic height originally from the inaccurate ASTER GDEM, notable improvements to the D-InSAR measurements can be clearly seen, which is helpful to better interpret the deformation signals associated with the mudslide event. It is observed that the Zhouqu mudslide had caused large ground movements in the Luojiayu and Sanyanyu groove valleys. In addition, we find that the Suoertou landslide has been experiencing moderate ground movements during the geologic event.

Helmert Variance Component Estimation-based Vondrak Filter and its Application in GPS Multipath Error Mitigation

Vondrak filter is a unique smoothing method that aims to find a balance between the smoothness of the filtered data series and the closeness of the filtered series to the original one. The key element of the Vondrak filter is the determination of the smoothing factor, which controls the degree of smoothing. We propose in this paper a new smoothing factor determination method that is based on the Helmert variance component estimation for the Vondrak filter. Experiments with simulated and real datasets indicate that the proposed method can select the optimal smoothing factor, and separate the signals and random noise at different noise levels as long as the noise level is lower than the magnitude of the signals successfully. By exploiting the day-to-day repeating property of GPS multipath errors, we can use the proposed method to correct GPS measurements for the multipath errors. We first use the proposed method to separate the multipath signals from noise, and then use the separated multipath signals to correct the subsequent sidereal day’s multipath errors of GPS survey. The results show that the accuracy of the GPS positioning is improved significantly after applying the proposed methods. Comparisons with some well-known filters are also made.