S1a debursting results look strange


I try to make interferograms with S1a data in SNAP 3.0. I follow the TOPS Interferometry tutorial for S1a from May 2015. It is working, so that I get an interferogram. But I have still stripes after running the debursting step and at the end after the goldstein filtering I have these phase jumps between the subswaths and the bursts.
My preprocessing chain is: Import Sentinel 1 > S1 TOPS Coreg. (3x (for each sub swath)) > Interferogramm formation > (3x (for each sub swath)) > 3x debursting > merging > topo phase removal > goldstein filtering.

I have here a zoomed screenshot after the debursting step of the first subswath:

I try to understand, why I have still these effects. Can somebody help me? Do I have to run the radiometric calibration before the coreg. process? Or is it included in this process?
Thank you very much in advance.

It should be possible to remove these by using the Spectral Diversity step before interferogram formation (Open products -> TOPSAR split -> apply orbit files -> back geocode ->Enhanced Spectral Diversity etc…) . However when I have tried to use this it hasn’t actually made any difference to the stripes. Does it work for you?

Hey Amy,
thanks for replying. I just try this way. But I I thought if I choose the S1 TOPS Coreg. method, that all these steps like split, apply orbit file, back geocoding and esd would have been included.
I let you know, if I have results.

Hey Amy,

so I have run your suggested processing chain, but I still have stripes in the sub-swath between the bursts.
What are you doing, if this also did not help you ?

Does anybody else have a solution for that?

@HHVice I am in the same situation and have this problem for almost all of the pairs I am using. I have not been able to remove the stripes and have posted this problem a few times but not heard any other suggestions. Does anybody have any advice?

@amylauraparker I also have tried several ways. SNAP 2.0, 3.0, Coreg with or without ESD,… I can not produce images without these stripes like in the tutorial from the esa. I found many papers concerning producing ifgs from the IW mode from S1a and in every paper it is mentioned, that the ESD step is important. But for me it did not work.

If you have precise orbits, the TOPS backgeocoding with ESD should be able
to precisely coregister the data so that you may produce an interferogram
without phase jumps. However, phase jumps can still occur in scenes where
there is movement or an atmospheric component to the signal.

Thanks @lveci. Unfortunately it seems in many cases the ESD is failing to make a difference and that even with the precise orbits the phase jumps still occur due to TOPS mode acquisition rather than a deformation/atmospheric component. I’ve found that they often occur when the coverage of the swath has been adjusted slightly. Do you have any other suggestions of things to try? Speaking to other colleagues I know this is an issue with GMT5SAR GAMMA etc also.

Please update to version 3.0.3 of S1TBX (released today) to correct a bug in the orbit interpolations. This has been causing phase jumps in a number of products.

@HHVice I’m getting similar problems.
Could you tell us where is located the footprint of the SAR images you are using?


the footprint is located in the north of Chile in the Atacama.
Yours as well?
I have still problems to process the data with SNAP, so I switch to GMTSAR.

My is located in northern Finland.
My post about this issue: S1 TOPS (ESD) Coregistration

I’m trying to understand and if possible to overcome this problem. With these artifacts I can’t continue my research.

Could you tell me:

  1. What are the names of the two products that you have used to generate the interferogram above? I will repeat your processing.
  2. Have you gotten better results using GMTSAR?

the two products have the name:

I still work on GMTSAR/GMT5SAR, just started. I will let you know, if it is better. But I’ve heard only good things about it and it should be faster. SNAP is only able to take one core on my machine.

Did you get any progress with your data?
I still wanna mention, that I had the same abnormality, that the phase jumps at the top of the image are not so big like the ones at the bottom.

Thank you.
I think I solved your problem. My problem seems to be more complex.
See here:

thanks for the troubleshooting and all the efforts on that.
But I stay now with GMT5SAR, which is also 4free to use and until yet I have no phase jumps at all.
Is this this issue, which you have found, an error in general in the code?

Hi there,

I’m sorry, I know it’s been a while since you posted this, but I am experiencing the same problems of incoherence between bursts in my interferograms and I would like to try your solution, could you please help me? This is my processing chain:

“Back-geocoding —> ESD --> rgFlt --> azFlt --> ifg --> deb --> dinsar --> flt --> ML”

My question is: how do I compute a new deramp demod phase before the ESD operator performs the azimuth shift?

The range ans azimuth filters don’t support burst data. I don’t believe they are needed for S1.

Thanks. Could these be the source of the inter-burst correlation issues?