Flat earth compensation leaves unknown phase component remaining

We are working with TerraSAR-X repeat pass InSAR data. Flat earth removal was done with the following parameters:

  • 5th degree polynomial
  • 501 estimation points

When doing so, we observe two things:

  1. The flat earth phase seems to contain merely a linear phase component. Given the swath width, we would, however, expect a bit more nonlinearity in the signal.
  2. In any case, after flat earth removal, there seems to be a significant (flat earth-related?) phase component remaining.

Both phenomena are displayed in the attached PDF. Is that a known problem or are we doing something wrong?
TDX_example.pdf (686.9 KB)

I was processing S1B for a specific location at one point, and the flat earth phase removal seemed to be failing, but Abraun was able to process the same data using ASF verfex InSAR and that problem didn’t exist… never figured out what the problem was.