@EJFielding I selected a reference point after obtaining the horizontal displacement map, subtracted the value of its horizontal displacement, and obtained agreement with the measurements of GPS points. Also, after obtaining the vertical displacement map, I selected another reference point and subtracted the value of its vertical displacement, and there was a convergence with the measurements of the vertical displacements of GPS points . Is the work I did correct? Thank you very much for always helping me
When comparing to GPS measurements, you should always project the 3D GPS measurements into the original InSAR line-of-sight directions, not compare them to the derived horizontal and vertical displacements.
2 Likes
hi Mr Mirzade. did you learn how to use GBIS for modeling source of earthquake? and how can we get the initial parameters and file for GBIS input?
@EJFielding Could you elaborate more on this statement (Aug 2023), please. It is interesting to have your arguments.
Thank you for your time.
When you have two InSAR measurements from ascending and descending orbits of a satellite system, you can’t estimate the full 3D surface displacements accurately. In particular, the north component of displacement can’t be separated from the vertical. Unless you can assume the north component is zero, then the vertical estimate will include some north displacement. By projecting the full 3D GPS displacements into the two InSAR LOS directions, you will accurately include correct amount of the north component in the comparison to the InSAR.
1 Like
@EJFielding Dear Dr. Fielding, thank you for your valuable experience here. Could you please suggest software for reprojection of the full 3D GPS displacements into the two InSAR LOS directions or provide the equations to do this.
Thank you.
First calculate the look vectors [ux, uy, uz] at your GPS station location, for ascending and descending tracks. Next calculate the LOS displacement (asc and desc) for GPS by adding the products of lookvector by GPSdisplacement: (ux * GPSeast) + (uy * GPSnorth) + (uz * GPSvert) . You then need to adjust the reference date to compare the GPS measurements with the InSAR time series.
Maybe the look vectors have to be calculated before, but now many InSAR packages include either the look vectors computed for the specific frame or the information necessary to calculate them (incidence angle, heading angle).
You can do it all in a bash script for simplicity.
@Simonst7 Do you suggest to use SNAP for this task?
I am trying to perform DinSAR on an earthquake event. The master image is a scene but the slave image falls on two overlapping images. coregistering them has been difficult since each of the images have 3 sub swaths. subsetting and splitting these subswaths before coregistering has not worked. Any advice on how I can subset and coregister these images. or merge both slave images before co registering with the master image. I have tried everything and it just won’t work.