Assessing the tropospheric correction results

[Rino Salman]

Hi Yunjun,

I want to know the impact of tropospheric correction to my interferograms. To do so, I reconstruct interferograms using "Ifgram_reconstruction.py" with input files: timeseries.h5 (based on original interferograms without tropospheric correction I guess), timeseries_ERA5.h5, and timeseries_GACOS.h5 (produced using tropo_gacos.py). Then, I calculate the standard deviation for each reconstructed interferogram using the output file from "Ifgram_reconstruction.py" (ifgramStackRecon.h5/unwrapPhase). The results (attached, 1 circle=1 interferogram) show that standard deviations of the interferograms after the tropospheric corrections do not result in a significant impact compared to the standard deviations of the interferograms before the corrections.

Before I "blame" the pre-existing global atmospheric models (inaccurate, etc) in my study region (Sumatra, Indonesia), I want to clarify with you: 1) whether this is a correct way (compute the standard deviation before and after the correction) to assess the quality of tropospheric correction results, 2) "ifgram_reconstruction.py" uses timeseries.h5 as an input which is generated after step 5 (invert_network), so I think the reconstructed interferograms are not the same as the original interferograms after the ISCE processing, because there is weighting involved in invert_network, I am curious whether assessing the impact of tropospheric correction using timeseries.h5 is correct. What do you think?

Best,

Rino

[Zhang Yunjun]

Hi Rino,

1) whether this is a correct way (compute the standard deviation before and after the correction) to assess the quality of tropospheric correction results

This is up to the personal taste. I “believe” STD could be a good parameter for it ONLY IF the other phase contributions (deformation, iono, topo residual from DEM error, etc.) are properly removed. For Sentinel, topo residual is small, iono is small for descending (after phase deramping). So an area known for no deformation sounds good for this purpose.

You may also want to try the scatter plot of phase vs elevation, as several troposphere papers used.

I am curious whether assessing the impact of tropospheric correction using timeseries.h5 is correct.

Time-series file is the optimal estimate from the redundant interferograms, with decorrelation noise mitigated. I usually like to compare the time-series files directly (plotted in the same color range), and look/judge at the phase pattern manually since I know my study area very well.

Best,

Yunjun

--
You received this message because you are subscribed to the Google Groups "MintPy" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mintpy+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mintpy/1b551212-ff5d-48de-acbf-1bb84090324an%40googlegroups.com.
<trop_correction.pdf>

[Rino Salman]

Thanks Yunjun.