Burst discontinuis due to glacier displacement in Sentinel-1 data

[fdelgadodelapuente]

Dear ISCE developers and users,

I have processed one-day Sentinel-1C/D interferograms in glaciers on the Patagonian ice fields. The data quality is spectacular but obviously the NS displacement of the ice results in displacement in the along track direction that is recorded in the ESD files. You can see that in certain regions of slow displacement in the glaciers the bursts show no discontinuities, but in other they do, like in this example.

Interferogram

ESD combine_IW3.int file

Is it possible to force the ESD correction to filter out somehow these strong azimuth displacements?

Thanks
Francisco

[piyushrpt]

You will have to write your own custom ESD filtering step. Essentially, all that step does is a histogram / median of overlap phase thresholded by coherence. If you use output of rdr2geo to create a mask of high velocity from existing maps like ITS LIVE - you can use that mask to filter out more points before performing that histogram / median.

[fdelgadodelapuente]

Thanks Piyush. I noticed that reading the runESD.py code that the calculation is pretty straightforward and does not take care of anything other than pixels of low coherence.

[fdelgadodelapuente]

Dear ISCE developers and users,

I found that those strong burst discontinuities are due to the ionosphere. They are gone outside of the deforming glaciers by including the burst overlap phase on the split spectrum computation in the topsApp.py input file

<property name="consider burst properties in ionosphere computation">True</property> 

This implies that the ionosphere is still quite strong even on latitudes as low as -49 (Patagonian Icefields), and that the effect is noticeable even on winter descending data acquired during the dawn.

Francisco,