Subset is showing upliftment and subswath is showing subsidence

I was processing 3 bursts but I tried to make subset to get small area and I get upliftment in some area. Again I processed 3 bursts of that sub swath I am getting subsidence in those areas where there was upliftment in subset image.
What might be the reason and how to get reliable data?

This is subset extracted from 3 bursts.
hilly region is showing upliftment.

Again I processed the 3 bursts to cross verification the same area is subsiding.

Have a look at this.

What might be the reason and how can I improve the data realibility?

have you seen these?

the strongly changing values suggest that the phase quality is bad in those areas. The satellite image already indicates that this coincides with forest areas where no displacement can be measured.

Can you please show a screenshot of the interferogram after filtering?

1 Like

This is topo-phase

This is phase

I think high vegetation and high temporal phase decorrelation is main reason for the same. How can we find that what is the rate of subsidence For longer duration of time? I just want to know that how much land subsidence is there during last 4-5 years?

as you see, the strange patterns in your displacement are those with noisy phase. They produce random results during the unwrapping.
You can reduce the amount of phase noise by an image pair with shoter temporal baseline. A period of 3 years is way too long - pairwise DInSAR is not working over such large periods. So the contribution of displacement to the interferogram is superimposed by other effects. Please try an image pair with 12 or even 6 days baseline.

If you want to retrieve subsidence over a longer time, you have to combine displacements of multiple pairs.

1 Like

@ABraun thanks for your instant reply. your suggestions are very valuable. How to combine displacements of multiple pairs? Is it similar to PSI? PSI will provide horizontal velocity?

you can combine several displacement pairs as provided in this tutorial: (example given at the end).

PSI is another method, more robust and with a higher suitability for time-series, but also more compilcated to achieve.

Neither of them provide horizontal displacements, but you can generate them by combining displacements of ascending and descending results. We have many discussions on this topic in this forum.


the stacking technique is also presented here: Wright & Stow (1999) Detecting mining subsidence from space

I might add this point at the end of the Sentinel-1 TOPS interferometry tutorial at a later time as well.

1 Like