Phase Referencing and Sign Convention of Tropospheric and Ionospheric Phase Delay Estimations

[alnguyen01]

I'm currently using the 'correct_ionosphere' and 'correct_troposphere' steps within MintPy to calculate their corresponding two-way phase delay and remove these phase estimations manually from interferograms that are affected by these errors. It looks like the estimated ionospheric time-series (./ion.h5) are phase-referenced (iono_split_spectrum.py), but is the tropospheric estimated phase delay time-series (./inputs/ERA5.h5) also phase referenced? I was unable to find any information within the tropo_pyaps3.py script about this tropo_pyaps3.py.

Additionally, are the sign conventions of both these phase delay estimation time-series consistent with the original ifgramStack.h5/unwrapPhase being "+" away from the radar? I see that the sign is reversed in the tropo_pyaps3.py script tropo_pyaps3.py line 573, but is this the same for the ionospheric time-series?

Attached here is an example of the two-way phase delay estimation (date1 one-way delay - date2 one-way delay for the tropo time-series) and the correction for a given manually-phase-referenced scene-wide interferogram over northern Alaska.

Here is an example of the ionospheric two-way phase delay estimation.

And here is an example of adding the tropospheric and ionospheric phase delays to then subtract them from the unwrapped phase:

I manually phase-referenced in the ifgramStack.h5/unwrapPhase using a set of reference pixels, as shown below in the image of average coherence over the stack of same ints used in the MintPy analysis (black x = MintPy reference pixel, and red x's = additional pixels selected using a priori information of relatively stable pixels).