Displacement Time Series Question

[ballistic]

Hello,

I have been teaching myself to create displacement products with Sentinel 1, SNAP and Snaphu. I am at the point where I can reliably create a displacement product but I fear I lack some of the more subtle understanding of the parameters when I run my graphs, which I have attached here.

I have been reading many of the links users such as ABraun have shared and I appreciate all the effort that they take in trying to help us beginners.

As a learning exercise I have been trying to measure displacement over time at a dam. Currently, Persistent Scattering methods and Stamps is beyond my abilities, so I have been making several displacement products, adding a coherence band, and masking the displacement by a minimum coherence of .45 or greater. As I understand it, atmospheric effects cannot be controlled for so I have been measuring dam displacement as a value relative to a baseline area near the dam. Is this the proper way to proceed in this case?

Second question is I have been processing my images as follows for time series: A-B, B-C, C-D, D-E, etc.
Should I instead process them A-B, A-C, A-D, A-E, etc? The dam has seasonal subsidence and uplift from what I can tell, but I want to make sure I am doing this correctly so far.

Thanks for any advice!

[ABraun]

@ballistic

As I understand it, atmospheric effects cannot be controlled for so I have been measuring dam displacement as a value relative to a baseline area near the dam. Is this the proper way to proceed in this case?

yes, this is the most robust way with classic InSAR approches - I would have recommended the same. Good job on getting so far by the way

@ballistic

Second question is I have been processing my images as follows for time series: A-B, B-C, C-D, D-E, etc.
Should I instead process them A-B, A-C, A-D, A-E, etc? The dam has seasonal subsidence and uplift from what I can tell, but I want to make sure I am doing this correctly so far.

This entirely depends on what you do with the information. AB, AC, AD, AE gives you the displacement between the first data and many others. The results area easier to compare because you see how the dam dynamics change over time. The downside is the increasing tempral decorrelation which leaves you with less high coherence areas.
On the other side, AB, BC, CD… gives you the displacement for a defined time step and you would rather add them up.

The differences are elaborated at the end of this tutorial: https://eo-college.org/resources/insar_deformation/

But in the end, all depends on if the unwrapping produced usable results over your dam.

[swadhin]

@ballistic @ABraun how do I produce the time series of the displacements over the time period? Can the Pin tool be useful, if yes , how?

[ABraun]

Please have a look at this answer:

Multi-Temporal InSAR by SNAP

Hello everyone. I want to do a dinsar with multiple images. I know that we can only use two images at a time but how should I do insar for one year’s data. I have 26 images (don’t want to do SBAS and PS right now) and want to make a time series of deformation. I used the stacking tool to find my master image. should I make the interferograms based on this image? or instead, when I have two images (at a time), should I use the first date as the master and the second date as the slave?