ERS-ENVISAT Permanent Scatterers Interferometry

The Permanent Scatterers (PS) technique is a powerful and fully operational tool for monitoring ground deformations on a high spatial density grid of point-wise targets, exploiting long series of SAR data [1], [2], [3]. The most attractive aspect of this approach is the capability of providing measurements relative to individual radar targets with unprecedented precision. Up to now, PS analyses have been carried out on ERS, RADARSAT, and JERS data sets. The purpose of this presentation is to discuss the feasibility of updating results obtained by means of a PS analysis on ERS interferometric data using ENVISAT ASAR images. In particular, the main goal is to stitch coherently the new ENVISAT ASAR measurements to already available ERS displacement time series relative to individual PS. To this end, we will model the interferometric phase of point-wise targets taking account the different ERS and ENVISAT carrier frequencies. Then, we identify the main constraints to be met at individual Permanent Scatterers in order to guarantee the feasibility of coherent stitching. I. ERS AND ENVISAT: AUTOAND CROSSINTERFEROGRAMS Besides several important technological improvements, the key difference between ERS and ENVISAT is the 30 MHz central frequency change from 5.3 to 5.33 GHz. This change entails advantages as well as disadvantages: classical interferometry on distributed scatterers is made impossible, unless the normal baseline ranges from –1000 m to – 3000 m [4], [5], [6] so that the wavenumber shift ensures sufficient range common band. Thus, high baseline cross-interferograms can be generated and very high precision DEM estimation becomes possible. A very interesting application could be the topographic characterization of flood plains. However, volumetric decorrelation effects are likely to be very strong and, therefore, extremely short revisiting times would be important to guarantee at least a negligible temporal decorrelation. To this end it should be kept in mind that ERS2 operates now in Zero Gyro Mode and, therefore, looking for short temporal baseline pairs, constraints for ERS1 With cross-interferograms we refer to interferograms between ERS and ENVISAT images. ENVISAT cross-interferometry on distributed scatterers are set on the Doppler Centroid value as well. Since, of course, these rather strict requirements are not systematically met, two different strategies can be considered for continuing the phase histories of individual scatterers. In the case of point scatterers (or at least scatterers with a reduced slant range extension) the baseline range allowing for a prediction of the phase signature from the ERS to the ENVISAT operating frequency is extended and crossinterferograms can be created without strict constraints on the normal baseline. In fact point scatterers are imaged coherently with both systems by definition. Their phase histories can be stitched moving from one to the other frequency. Of course this requires the correction of the deterministic phase terms depending on the scatterer position coupled with the normal baseline and the frequency shift (the expression is provided in the next paragraph). Alternatively, we can consider classical auto-interferograms only and bridge the frequency change in a similar way to what is carried out in Small Baseline Interferometry [7]. This can be performed by combining the two distinct interferogram classes (namely ERS-ERS and ENVISATENVISAT) basing on a model adopted for motion (the interferogram classes can be temporally intertwined, as long as ERS-2 operations are continued). Two main advantages of using cross-interferograms are: The possibility of long term interferometry from 1991 to the end of the life of ENVISAT, allowing one to measure very slow earth motions on a pixel by pixel basis in a unique coherent time series (all data are referred to a single master acquisition); The possibility of determining with high precision the location of the scatterer, exploiting the slant range dependent phase shift in cross-interferograms. In fact, the phase shift due to the change of frequency from f0 to f0+∆f, for a given PS with slant range position and elevation respectively ∆r and ∆q (both relative to the center of the sampling cell taken as origin of the coordinates) is: