SAR Coherence over Himalayan glaciers - which co registration method to choose?

I am trying to create a SAR coherence image over some glaciers in the Himalayas in order to map debris-covered ice. I have downloaded two Sentinel 1 images of my area in IW format. When I am co-registering the data before carrying out all the steps to make a coherence image, what is the appropriate method to choose? I tried doing the S1 TOPs co-registration (with and without ESD) but I was concerned that the latter uses cross-correlation which I guess is not appropirate with an area containing flowing ice.

Does anyone have any suggestions on which method is the most applicable to use for a glaciarised area?


Here comes the step you could follow:


  1. Coregistration of data (Corregister TOPS Corr. And ESD Both are possible)

  2. Interferogram Formation

  3. TOPS debursting

  4. Topographic phase removal

  5. Goldstein phase filtering

  6. Multilook [10 (range); 2 (Azimuth)] for filtere5d phase;

  7. Reconstruct the phase after Multilooking by using the Band Maths Op;

  8. Export the result of step 4 to SNAPHU

  9. Export by SNAP to Unwrapping

  10. Import the result of SNAPHU
    11 - phase to displacement

  11. Range-Doppler Terrain Correction

  12. Export the result of step 10 as a KMZ

NB/ Either you could end up your processing until the step 7 by getting the coherence map, Or you could continue until to get the Differential Interferogram

1 Like

Hi falahfakhri,
Thanks for your response.

I have two little questions:

  1. It says in the help file that the ESD uses cross correlation to co-register the images - that is why I was thinking for setting where you may have significant changes between the two images, ESD was not appropriate?

  2. What is the Band Maths Op and why is it necessary?

Thank you for the help,


1 Like

In case your study is within one Burst, then you could use TOPS Corr., However, in case your study area is located within two bursts, then in this case use ESD to avoid jump in phase, but the ESD is only used in low deformation rate, However, there is a solution in case your study area is located within two bursts or more, besides the high deformation rate, then you could apply ESD to one image and follow this step by Corr. using TOPS corr.,

Use to enhance the visualization of your image only (In this case).

1 Like

Great, thank you very much. I will try it now and see if it works.


Also take a look at this topic,

Great. It all seems to be working. I am not interested in the DINSAR displacements, just coherence, so I was wondering if the band op influences the coherence image or just the phase? Also do you have more details on the band op?


As I explained to you previously, the band op (In this case) is only to enhance the visualization of your image, and your orginal data should be used, to get the real results of coherence.

Great, thank you

Hey, I am also working on glacier mapping using Sentinel 1 and working on it for the first time. Can someone help me regarding this ? Thanks