unwrapping error of large earthquakes based on ALOS2 ScanSAR data

[Viecgg]

I used alos2App.py to generate the displacement field of the 2025 Myanmar earthquake using ALOS-2 ScanSAR data. The workflow went smoothly, and I successfully obtained the interferogram, unwrapped displacement map, connected component map, and the *.ino.geo file, as shown below. However, there appears to be some unwrapping error in the filt_250216-250330_5rlks_28alks.unw.geo file. I would like to know how to reduce or correct this unwrapping error.

Here is my configuration file:

<property name="reference directory">./ALOS2_0216/downloads/</property>
<property name="secondary directory">./ALOS2_0330/downloads/</property>
<property name="reference frames">[3150, 3200, 3250]</property>
<property name="secondary frames">[3150, 3200, 3250]</property>
<property name="starting swath">1</property>
<property name="ending swath">5</property>

<property name="dem for coregistration">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property>
<property name="dem for geocoding">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property>
<property name="water body">./wbd/swbdLat_N12_N27_Lon_E092_E101.wbd</property>

<property name="use GPU">True</property>

[Viecgg]

@EJFielding Dear Dr. Fielding, I would really appreciate your suggestions or any advice on this issue.

[EJFielding]

Hello @Viecgg, we got similar results processing this coseismic interferogram. I decided it was going to be too much work to get a better phase unwrapping, so I did not attempt it. There is very little high-coherence phase in the area near the main rupture, so there is not much surface displacement that could be unwrapped correctly. We don't have tools built into ISCE2 to do the required manual intervention to fix phase unwrapping. The basic method is to mask out the phase along the main rupture to separate the phase on the two sides and unwrap them without the unwrapping program trying to connect across the rupture. In this earthquake on this track the fringes are only very close to the main rupture, so it will be difficult to separate the good phase from the bad phase.

[Viecgg]

Hi Dr. @EJFielding , thank you very much for your reply. Since it’s quite difficult to obtain a reliable unwrapped phase from this ALOS-2 ScanSAR track, I’m wondering if there might be an alternative approach to derive azimuth displacement.

I came across another discussion (#674 ), which mentioned that alos2App.py does not support dense offset. Dr. @CunrenLiang suggested that alos2burstApp.py could be used to extract azimuth displacement from ALOS-2 ScanSAR data if I understand correctly. I’d appreciate it if you could kindly provide some guidance or configuration *.xml files on how to proceed with this method.

Thank you again for your time and suggestions!

[fdelgadodelapuente]

Hi @Viecgg,

If you have access to the now very old and deprecated ROI_PAC, you can use its implementation of the branch and cut algorithm. That way you set a seed point to start the phase unwrapping and then it will only unwrap good coherent areas stopping at regions separated by lots of residues. Afterwards you can set bridges to manually connect individual sections that you now the number of integers of 2pi with respect to the previous unwrapped patches. Branch and cut is much better than SNAPHU_MCF for unwrapping barriers separated by lots of residues due to a surface rupture. Keep in mind that ROI_PAC is useless nowadays unless you need to reprocess some very specific ENVISAT/ERS/RADARSAT-1/JERS interferograms, or use its branch and cut unwrapper.

Francisco

[Viecgg]

Hi Dr. @fdelgadodelapuente,

Thank you for your detailed reply. Since I don’t have access to the old version of ROI_PAC, my plan is to unwrap the interferogram by splitting it into upper and lower parts along the fault trace. I will then bridge the two unwrapped phases by adding multiples of 2π. I’ll share my results if this method works.

[CunrenLiang]

An example xml file is here https://github.com/isce-framework/isce2/blob/main/examples/input_files/alos2/alos2burstApp.xml From: ***@***.***> on behalf of Viecgg ***@***.***> Reply-To: isce-framework/isce2 ***@***.***> Date: Saturday, July 26, 2025 at 10:10 AM To: isce-framework/isce2 ***@***.***> Cc: Cunren Liang ***@***.***>, Mention ***@***.***> Subject: Re: [isce-framework/isce2] unwrapping error of large earthquakes based on ALOS2 ScanSAR data (Discussion #974) Hi Dr. @EJFielding , thank you very much for your reply. Since it’s quite difficult to obtain a reliable unwrapped phase from this ALOS-2 ScanSAR track, I’m wondering if there might be an alternative approach to derive azimuth displacement. I came across another discussion (#674 ), which mentioned that alos2App.py does not support dense offset. Dr. @CunrenLiang suggested that alos2burstApp.py could be used to extract azimuth displacement from ALOS-2 ScanSAR data if I understand correctly. I’d appreciate it if you could kindly provide some guidance or configuration *.xml files on how to proceed with this method. Thank you again for your time and suggestions! — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[Viecgg]

Hi Dr. @CunrenLiang, thank you very much for your prompt and kind reply. I’ll give it a try.

[EJFielding]

Let us know how the along-track interferometry with alos2burstApp works. I thought about running that but did not have time to try it.

[Viecgg]

Hi Dr. @EJFielding and Dr. @CunrenLiang, I derived the along-track interferometry by alos2burstApp.py successfully as shown below. However, the obtained azimuth displacement is kind of weird — it shows some fringes along the east-west direction instead of a clearly north-south displacement. Is there any setting that might be wrong? Could you give me some suggestions to get a reliable azimuth displacement of this ALOS2-ScanSAR track?

Here is my configuration file alos2burstApp.xml:

<property name="reference directory">../ALOS2_ISCE2/ALOS2_0216/downloads/</property>
<property name="secondary directory">../ALOS2_ISCE2/ALOS2_0330/downloads/</property>
<property name="reference frames">[3150, 3200, 3250]</property>
<property name="secondary frames">[3150, 3200, 3250]</property>
<property name="starting swath">1</property>
<property name="ending swath">5</property>

<property name="dem for coregistration">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property>
<property name="dem for geocoding">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property>
<property name="water body">./wbd/swbdLat_N12_N27_Lon_E092_E101.wbd</property>

Here is the derived along-track interferometry result by alos2burstApp.py:

Here is the azimuth displacement of the same ALOS2 track from JAXA(https://www.eorc.jaxa.jp/ALOS/jp/library/disaster/dis_myanmar_earthquake_20250409_j.htm), How can I get a azimuth displacement like this?

[CunrenLiang]

That’s caused by ionosphere. You can remove it by masking the along fault areas and fitting a 2d polynomial to the rest of the result. From: ***@***.***> on behalf of Viecgg ***@***.***> Reply-To: isce-framework/isce2 ***@***.***> Date: Sunday, July 27, 2025 at 1:45 PM To: isce-framework/isce2 ***@***.***> Cc: Cunren Liang ***@***.***>, Mention ***@***.***> Subject: Re: [isce-framework/isce2] unwrapping error of large earthquakes based on ALOS2 ScanSAR data (Discussion #974) Hi Dr. @EJFielding and Dr. @CunrenLiang, I derived the along-track interferometry by alos2burstApp.py successfully as shown below. However, the obtained azimuth displacement is kind of weird — it shows some fringes along the east-west direction instead of a clearly north-south displacement. Is there any setting that might be wrong? Could you give me some suggestions to get a reliable azimuth displacement of this ALOS2-ScanSAR track? Here is my configuration file: <property name="reference directory">../ALOS2_ISCE2/ALOS2_0216/downloads/</property> <property name="secondary directory">../ALOS2_ISCE2/ALOS2_0330/downloads/</property> <property name="reference frames">[3150, 3200, 3250]</property> <property name="secondary frames">[3150, 3200, 3250]</property> <property name="starting swath">1</property> <property name="ending swath">5</property> <property name="dem for coregistration">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property> <property name="dem for geocoding">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property> <property name="water body">./wbd/swbdLat_N12_N27_Lon_E092_E101.wbd</property> Here is the derived along-track interferometry result by alos2burstApp.py: image.png (view on web) Here is the azimuth displacement of the same ALOS2 track from JAXA(https://www.eorc.jaxa.jp/ALOS/jp/library/disaster/dis_myanmar_earthquake_20250409_j.htm), How can I get a azimuth displacement like this? image.png (view on web) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[CunrenLiang]

In addition, ionospheric effect varies with acquistion time. If you want to do a fair comparison, you need to use the same acquisitions. From: ***@***.***> on behalf of Viecgg ***@***.***> Reply-To: isce-framework/isce2 ***@***.***> Date: Sunday, July 27, 2025 at 1:45 PM To: isce-framework/isce2 ***@***.***> Cc: Cunren Liang ***@***.***>, Mention ***@***.***> Subject: Re: [isce-framework/isce2] unwrapping error of large earthquakes based on ALOS2 ScanSAR data (Discussion #974) Hi Dr. @EJFielding and Dr. @CunrenLiang, I derived the along-track interferometry by alos2burstApp.py successfully as shown below. However, the obtained azimuth displacement is kind of weird — it shows some fringes along the east-west direction instead of a clearly north-south displacement. Is there any setting that might be wrong? Could you give me some suggestions to get a reliable azimuth displacement of this ALOS2-ScanSAR track? Here is my configuration file: <property name="reference directory">../ALOS2_ISCE2/ALOS2_0216/downloads/</property> <property name="secondary directory">../ALOS2_ISCE2/ALOS2_0330/downloads/</property> <property name="reference frames">[3150, 3200, 3250]</property> <property name="secondary frames">[3150, 3200, 3250]</property> <property name="starting swath">1</property> <property name="ending swath">5</property> <property name="dem for coregistration">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property> <property name="dem for geocoding">./dem/demLat_N12_N27_Lon_E092_E101.dem.wgs84</property> <property name="water body">./wbd/swbdLat_N12_N27_Lon_E092_E101.wbd</property> Here is the derived along-track interferometry result by alos2burstApp.py: image.png (view on web) Here is the azimuth displacement of the same ALOS2 track from JAXA(https://www.eorc.jaxa.jp/ALOS/jp/library/disaster/dis_myanmar_earthquake_20250409_j.htm), How can I get a azimuth displacement like this? image.png (view on web) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>