Fix names/units for variables on METAR netcdf files

[dopplershift]

There are some curious issues in the netcdf files for METAR data on http:https://thredds.ucar.edu:

• Names are not all CF-Compliant: e.g. inches_ALTIM, hectoPascal_ALTIM
• The units on altimeter setting are 'inches'. This should be changed to udunits for 'in Hg'

Would be nice to make the new units compatible with MetPy/Siphon.

[lesserwhirls]

If you look in the idd/pqact/cdl directory, you will see where these are
set (I think). Should be an easy fix without having to rework the decoders.
I'll take a look and see.

Sean

On Sunday, June 21, 2015, Ryan May [email protected] wrote:

There are some curious issues in the netcdf files for METAR data on
http:https://thredds.ucar.edu:

• Names are not all CF-Compliant: e.g. inches_ALTIM, hectoPascal_ALTIM
• The units on altimeter setting are 'inches'. This should be changed
  to udunits for 'in Hg'

Would be nice to make the new units compatible with MetPy/Siphon.

—
Reply to this email directly or view it on GitHub
#140.

[dopplershift]

It's also odd that it calls these "altim" which I think means altimeter setting, but then it also calls them "pressure at station"; these are two completely different things.

[dopplershift]

More things I have noticed:

• The data returned (at least in CSV) have the field time rather than date, as in NCSS grid as point.
• String format for times is %Y-%m-%d %H:%M:%SZ--note the lack of a T in the middle
• Times are returned with increasing date, but decreasing time

[lesserwhirls]

I wonder what they are called in the WMO or NWS decoding docs.

On Sunday, June 21, 2015, Ryan May [email protected] wrote:

It's also odd that it calls these "altim" which I think means altimeter
setting, but then it also calls them "pressure at station"; these are two
completely different things.

—
Reply to this email directly or view it on GitHub
#140 (comment).

[dopplershift]

Also, wondering why BJC is the nearest station to 40, -105. Why no Boulder or Longmont obs?

[lesserwhirls]

The station list may not be up to date. It's hand maintained, I think.

On Sunday, June 21, 2015, Ryan May [email protected] wrote:

Also, wondering why BJC is the nearest station to 40, -105. Why no
Boulder or Longmont obs?

—
Reply to this email directly or view it on GitHub
#140 (comment).

[JohnLCaron]

reading metars from metar2nc bastard files. whats the long range fix here?

[dopplershift]

Long range fix captured in more appropriate venue on Unidata/TdsConfig#2.