NEWDATA(SoilGrids-2-0): cSoil (CV, PI90 uncertainties)

[msteckle]

• Adds SoilGrids-2-0 and two different measures of uncertainty
• Cleans up Hoffman which used bnds to encode uncertainty
• Cleans up CARDAMOM uncertainty naming

In general, I figure we'll want to use the ODS uncertainty approach but be willing to expand outside of it where needed

• uncertainty variables are called {var}{uncert} or coefficient_of_variation
  • I think we should probably just call it coeffvar or something to be more like ODS
  • We should first check if {uncert} fits into one of the ODS uncertainty categories, but if it doesn't, we can make up our own
  • standard_name is "{var standard name} {ILAMB uncertainty standard name}" --- standard names aren't required, but we might want to come up with our own standards? Or maybe not since uncertainty is so complex. Maybe just leave it to the long_name and comment.
    • Right now, I believe we have lower_bound, upper_bound, standard_deviation, standard_error, relative_uncertainty (I did this for soilgrids since it wasn't technically coefficient_of_variation; maybe it should honestly be ratio_uncertainty? idk.)

There are other issues to address that I didn't in this PR:

• CLASS: *_sd (should be *sd)
• LORA: mrro_sd --> (should be mrrosd)
• WangMao: *_sd --> (should be *sd)
• Fluxnet: *_stderr --> (should be *ustd)
• WOA: ohc_stderr --> (should be *ustd)
• WOA climatologies: *_stderr --> (should be *ustd)