I wouldn't trust that graphic at the top of the article to be very accurate. It has an obvious acquisition footprint that was not resolved in processing. Those WNW-ESE stripes should've been resolved before publishing by ground-truthing the stripes using benchmarks established inside the mapped area so that the end result wouldn't suggest higher/lower subsidence along tracks than seen on parallel offset from tracks. That's just sloppy.

The striping can have multiple sources so they need to study why there is an obvious footprint and then make the appropriate corrections.