Interpolate soundings made available

[akrherz]

A SHARPy developer provided me an interesting insight into interpolating sounding data

I’ll add that interpolation of the temperature and dew point profiles don’t seem to me to be a large issue. I believe the criteria for the standard and significant level selection is that the significant thermodynamic levels are selected in a way that the thermodynamic profile can be reconstructed via linear interpolation to a 1˚C (or 15% RH) accuracy, suggesting that interpolation is an expected component of the reconstruction process.

See a previous discussion within MetPy:: Unidata/MetPy#334

So U of Wyoming's archive may be doing this interpolation because it is the proper thing to do with the data. The thought never occurred to me that this interpolation could be a data-space saving thing done many moons ago within the reporting.

Anyway, an example interpolator was also provided, so perhaps I should implement this or at least make it an available option within the JSON API.

[wblumberg]

Hi @akrherz ...no problem quoting me verbatim...the example interpolator provided is something to strip the standard and significant levels from the high resolution data. It does not put it into the TTBB or TTAA formats, and we're not 100% sure this script does everything it is supposed to be doing. I would only pay attention to the README overall. The rules behind the level selection and interpolations are in the README, but you probably want to look into to GEMPAK's MR library, which correctly reconstructs a sounding from the TTAA/TTBB files. We've reconstructed soundings from the NSSL .gem archive through GEMPAK's SNLIST program. SNLIST uses the MR library to merge/interpolate the data together to reconstruct the sounding.

SNLIST: http://www.unidata.ucar.edu/software/gempak/tutorial_v2/snlist.html

Information/documentation about the library: https://github.com/Unidata/gempak/blob/4062d7bbc9c1c3a43a5ab51cc9073a4b28100400/gempak/txt/gemlib/mrlib.txt

The merge library: https://github.com/Unidata/gempak/blob/4062d7bbc9c1c3a43a5ab51cc9073a4b28100400/gempak/source/gemlib/mr/mruadt.f

Also, I stumbled upon this while searching for the above links. Looks like someone modified the mruadt.f program to interpolate winds also. I don't seem to remember GEMPAK interpolating the winds, but only the temperature and RH profiles. Not sure if it's helpful overall: http://madweather.blogspot.com/2007/08/avoiding-gempak-interpolation-when.html

[wblumberg]

Also, I think the upper air .gem sounding files should have international soundings in them. Some of the ones from GEMPAK files that we've decoded have international soundings in Europe, Asia, etc. I'm not sure what sites you could scrape to backfill that data though...