large phase variations at the coast - difficulty unwrapping / bridging?

[Jesse-kearse]

I have a question regarding phase anomalies at an urban coastline, where islands of reclaimed land are connected to the mainland by narrow land bridges. My input dataset is an Sentinel-1 SLC stack, processed using ISCE2 at full resolution, phase_linking with Miaplpy, and a single reference ifg network between dec 2016 - dec 2021 (~150 dates, reference in 20190630). The results look nice (please see "velocity") and the velocity values seem reasonable, apart from a large apparent "subsidence" within the coastal islands of reclaimed land. unwrapping in full resolution with snaphu worked really well for 95% of the study area, except for these coastal parts (please see "ifgram_20170228"). Connected components and coherence are mostly OK for parts of some of these islands, although coherence is very low in other places that are 'unconnected' in connComp. The time-series timeseries.h5 for 20170216, and 20170228 shows that these areas are OK in some of the images, but not in others, and show sharp displacement boundaries in (please see "ts_20170216", and "ts_20170228"). Correcting using bridging did not seem to produce any results using the default parameters. I have two questions: (1) How can I correct (either manually, or otherwise) these issues, given the large area and full resolution of the dataset, and (2) how can I assess from the InSAR data alone (without the use of additional ground deformation datasets) whether or not the displacement timeseries in such areas is reflecting actual ground deformation vs artificial errors related to unwrapping or other technical challenges. Any insight would be greatly appreciated.