Not logical result of interferometry

Hello everyone,
I calculated the interferometry for two Sentinel 1A SAR data for the centeral of Iraq I followed the below steps to get the final result

  1. Coregistration of data
  2. Interferogram Formation
  3. TOPS debursting
  4. Topographic phase removal
  5. Goldstein phase filtering
  6. Multilooking (optional)
  7. Export the result of step 5 to SNAPHU
  8. Unwrapping 9. Import the result of SNAPHU
  9. Phase to displacement
  10. Range-Doppler Terrain Correction
  11. Export the result of step 10 as a TIFF
    I am not happy with the final result because the displacement (LOS ) is not logical

could anyone have an explain to that?

I don’t see you mention Apply Orbits. For InSAR you should apply the precise ones.

1 Like

It is already involved in the S-1 TOPS Coregistration step (there is Apply-Orbit-File page). Do you agree with me or not?

Provided that the processing is without error (it looks ok to me), you could be seeing atmospheric/ionospheric ramps.

Yes, it was without error. Could you tell me how we can check the atmospheric/ionosperic ramps?

They are an error-source in SAR interferometry. Perhaps you should try with some other datasets to see whether your chain produces reasonable results somewhere else?

Thanks, I will do that, but what I understand from you that the procedure is fine and every step is Okay, is that correct?

Hi Arsalan,

Are your results obtained over an urban area?
It looks that there is a strong linear trend present in your results. This is usually due to orbital errors as mengdahl mentioned.
If you have applied the precise orbit files in your processing, then what you see is a strong signal delay due to ionosphere and troposphere (mainly troposphere).

What you can do is not to select image pairs that have been acquired under sever weather conditions (clouds, high pressure, high humidity). If all these parameters are very high when the images was acquired then your interferogram formation will be heavily affected, hence, your displacement results.

For better results, obtain a few images (discard the ones acquired under turbulent weather) create a stack and then calculate the map displacement. The more images you include in your process, the more the atmospheric error is minimized.

1 Like

Hi Johnhan,
Thank you for your response, the area is in the Mesopotamia. This area is suffering from sand storm and sand dunes distribution, In the area of interest, there are a few towns and cities. the two data was processed are acquired in November 2014 and 2015.
the main aim of this work is to monitor the sand dunes movement (3D).


Applying differential interferometry for detecting changes on the ground using only two images is not the ideal scenario. The simple answer is due to atmospheric errors. Let’s say that the first image of your area of interest was acquired under good weather conditions and the second image was acquired under severe weather conditions (stormy weather) this can add a significant error in your measurements (adding extra fringes on the interferogram) which will be interpreted as displacement.

In order to avoid such situations, you need to make sure that the images were not acquired under stormy weather. You can combine multiple images to produce a stack and then estimate the displacement map.

To move one step further, you can you PyAPS ( which is a python library for removing atmospheric noise or PSI technique

1 Like

Another thing is to check for suitable baseline conditions:
Chapters 1.3 and 1.4.1

I have check it, it is less than 100 m