I would say this makes sense if
- your displacement over the various images is constant
- you have enough image pairs to cancel out atmospheric patterns (if present in one image it might superimposing the fringes in other images)
- if the displacement patterns of all image pairs make sense and are more or less comparable. “Displacement everywhere” is always suspicious
Comparing ascending and desecending data is always a good idea but you are right - combining them to vertical displacement is not so trivial. Phase to Displacement Units? - #29 by mengdahl