Topographic phase removal does not seem to work

Hi,
I am working with a Sentinel-1 dataset, and I applied Radar->Interferometric->Products->Interferogram Formation.
Then Radar->Interferometric->Products->Topographic Phase Removal.
I used SRTM 1SecHGT(Auto Download).

In the image:
Left: Before topo remove
center: Topo phase
Right: After topo remove

It seems like topographic effect was not fully substracted, or am I wrong??

Thank you!
Marco

It looks like it worked - sometimes the atmospheric component is DEM-dependent due to atmospheric conditions at the time of acquisition. For example thick clouds in a valley.

I can see a similar situation in my case. Where the top image is the ifg before topo-phase removal, and the bottom one is after the removal.

Does that mean, the atmospheric fringes may resemble the topographic phase fringes? So they become more apparent after removing the topographic contribution, giving the illusion that there’s a topographic phase added instead of being substracted ?

Was it mentioned in literature that removal the erroneously calculated topographic phase contribution from the interferometric phase may leave remaining topographic relics or may add even more error by removing a portion of the deformation-related phase itself? If so, can you please mention the paper where this issue was encountered?