On the use of a water mask close to the sea

[svenborgstrom]

Hi Andreas,
my AOI is close to the sea, so I decided to use a water mask to filter out the contribution of water during processing (S1, descending, 85 images).
B.t.w. using the mask, I missed a consistent part of the info on the costline close to the sea, as you can see from the figure; conversely, I got a good result without masking, contrary to what I would expected. In any case, I can only see a few pixels in the sea without using a mask, so I'm wondering if masking is really necessary or not.
I used default values in the config.json file; keep in mind I have no experience on PS processing as I am a routine SBAS user.
Cheers,
Sven.

[Andreas-Piter]

Hi Sven,
Thanks for sharing this interesting case. In general, masking is not necessary as the quality pixels in the water is anyway low and will be filtered out. However, if you have a water mask, it doesn't hurt using this extra information.
Nevertheless, it is strange that the displacement could not be observed for such a significant part of the area. This problem should not be related to the masking.

I assume the problem could be related to the atmospheric filtering. So please check out the following questions:

1. Are the first-order points well distrbuted over the study area?
2. Does the time series of the first-order points look reasonable?
3. Were enough points found for atmospheric filtering?
4. Do you use spatial or spatiotemporal unwrapping?

As far as I know, your study area is characterized by non-linear displacement signals. However, a sufficient number of points with linear displacement is required for the atmospheric filtering approach.

Please try for the moment to run the processing without atmospheric filtering and see if it works.

Or if you don't mind, would you share your input dataset (inputs/slcStack and innputs/geometryRadar) with me so that I can investigate the error?
Best, Andreas