Solid earth tides in GUNW product

[dbekaert]

Below is the outline of the steps to add solid earth tides to the GUNW product:

What

Earth tides: The Solid Earth tide is the displacement of the solid earth's surface caused by the gravity of the Moon and Sun.

Dependencies

• (pysolid)[https://github.com/insarlab/PySolid] available on conda as:
  conda install -c conda-forge pysolid

Implementation

• Can either be a stand-alone wrapper script or directly be integrated in the topsApp workflow.
  Benefit of a stand-alone wrapper script allows also for backward compatibility in GUNW-products and push in the layer in past products and helps with integration into ARIA-tools.
• Solid earth tides are smooth and of long-wavelength (hence no need to generate them at the IFG resolution). As they are corrections of the whole solid earth, the topography impact to the ENU displacements is small and neglected (single layer is sufficient), however the looking geometry of the satellite will add a topography dependent component when projecting into the LOS. We will generate solid earth tides calculation at the same posting as the geometry layers (e.g. 0.1 degree) and the same vertical levels.

pseudo cost steps

• Extract the dates (reference, secondary), satellite UTC time, wavelength from the GUNW meta-data (can be done the same way as done for RaiDER). Extract the incidence and heading angle 3D cubes, and corresponding lon lat grid (0.1 lateral posting).
• Use the lon lat grid (0.1 lateral posting) to compute the earth tides in ENU for reference and secondary dates at time of satellite. This leads to three 2D rasters (ENU) for the reference and secondary dates each.
• For Reference and secondary project the solid earth tides from ENU into the LOS. Use the wavelength, and the incidence and azimuth 3D cubes. Easiest is to replicate the 2D E, N, and U component into a 3D cube (assume that there is no variation in Z) and then do the element wise operations for projection. An example how you to compute tides over a grid and projection is documented in the pysolid notebook.
• Difference the reference and secondary LOS solid earth tides to compute the differential This would now be a coarse 2D grid at 0.1 degree posting of solid earth tides.
• Push the layer into the product alike what is done for tropo. hard code version in product to be 1c for development purposes on ARIA-tools side.
• Work on ARIA-tools to add in support for solid earth tides. This would be much alike the tropo and geometry meta-data. See ticket here on how this would feed into mintpy
• Inspect and verify the signal is correct. You could compare with already existing implementation in Mintpy to ensure all signs are and magnitudes are correct etc.