File format for InSAR data

[brendanjmeade]

celeri.py currently uses pandas to read in a dataframe from a .csv file with the columns

"lon"
"lat"
"depth"
"line_of_sight_change_val"
"line_of_sight_change_sig"
"look_vector_x"
"look_vector_y"
"look_vector_z"
"reference_point_x"
"reference_point_y"

Anything else that should be added/modified? Thoughts welcome.

[jolivetr]

Hi Celeri devs,

It would be great to have the possibility to include a full covariance with the InSAR data. In general, we (at least I) estimate on the original data a gaussian or Laplace covariance decay as a function of distance between pixels on insar data and we use this decay to build this covariance. Considering that no one will invert 1e6 pixels, it could remain tractable. For each InSAR data, it would have to be a different covariance function.

Let me know if you need further details but this is described in Lohmann & Simons 2005 or in Jolivet et al 2012 (JGR).

Cheers, R

[brendanjmeade]

@jolivetr Thanks for this idea! Some followup questions: Is this analytic or data? If data, how is it stored (.hdf, .csf)?

[jolivetr]

We could start with a simple analytic expression (an amplitude and a length scale) so that céleri populates the Cd Matrix on his own. Otherwise, hdf5 is the modern way of doing things but I have to say I like plain binary files :-)