[icoads_usa] Improving the humidity calculations used for ICOADS as part of R3.0?

Scott Woodruff scott.d.woodruff at noaa.gov
Tue Feb 11 14:44:18 MST 2014


Hi Shawn and all,

I suspect (as USA staffing--and/or international partnership 
contributions--permit) we should plan to develop (and validate) new 
Fortran code for the WMO No. 8-based calculations. However (as more 
minor technical consideration), I'm not sure the resulting routine(s) 
would necessarily fit best in lmrlib, ref. an earlier comment I made 
quite a ways further upstream in this e-mail discussion:
>>>>>>>
>>>>>>> Developing further/better DPT/humidity calculation routines, e.g. to
>>>>>>> share with partners (plus ETMC, etc.), also sounds to me like a good
>>>>>>> idea. We'd have to think more about whether {lmrlib} would be 
>>>>>>> the best
>>>>>>> place for those though, or whether this might be another good 
>>>>>>> area for
>>>>>>> initiating GitHub-type collaboration (e.g. IMMA tools as recently
>>>>>>> provided by Philip Brohan: https://github.com/oldweather/IMMA).
Just to mention, in case useful at some point (e.g. for validation): In 
conjunction with the earlier JCOMM/SOT/TT-IS work on WMO TR 63 etc., 
Henry Kleta put together some code/testcases (but possibly not following 
Dave's earlier caveat "that the equations from the body of the text 
should be used rather than the single page in the attachment"):

On 3/6/12 5:49 AM, Kleta Henry wrote:
> Dear colleagues,
>
> a while ago a discussion concerning standards dewpoint algorithms started. With a lot of input from ETMC we agreed on providing programs, testbenches, and testdata to those wanting to implement WMO's formluae, given in WMO-No.8, 7th edition, 2008, Part I, Annex 4.B.
>
> Up to date I couldn't find a single implementation based on those formulas. Everyone (NMS and E-Logbook programmer) is basically using their own traditional formulas.
>
> That led to the fact, that I had to implement the WMO formulas on my own, as I didn't have a single implementation available.
> To assure the "quality" of what I did, I asked two colleagues to implement the formulas using different programming languages.  So now we have three versions available. Furthermore I have implemented a little IDL testbench to create data based on agreed testcases to allow manual checks of own implementations.
>
> Currently I can provide all programs "only" via my private website, as I do not have access to any other server.
>
> The programs can be found here:
> http://www.indigo2.de/download/dewpoint-humidity.exe
> http://www.indigo2.de/download/dewpoint-humidity.html
> http://www.indigo2.de/download/dewpoint-humidity.jar
>
> The IDL Testbench and the output of the agreed testcases (see attachement):
> http://www.indigo2.de/download/dewpoint.pro
> http://www.indigo2.de/download/dewpoint-idl.dat
...

I've added the red emphasis above. The current situation in terms of 
what appears to be a general lack of any WMO-certified software seems 
not very helpful all around (noting however this repository related to 
E-logbooks): 
http://www.wmo.int/pages/prog/amp/mmop/JCOMM/OPA/SOT/dewpoint-algo.html

In addition regarding already-available software resources, we have the 
old NOAA/PROFS (1981) full Fortran library of /Algorithms for 
thermodynamic calculations/:
     http://icoads.noaa.gov/software/other/profs
only a few of which are currently used in ICOADS processing, i.e. this 
small subset:
     http://icoads.noaa.gov/software/other/profs_short
probably exclusively for calculations of derived variables for the 
monthly summaries.

Agreed: very good suggestion to have a brief presentation on 
lmrlib/related ICOADS shared libraries at the IVAD1 Workshop. Since the 
emphasis thus far in developing lmrlib has been on historical 
calculations, I doubt you'll find much code of use there for translating 
modern SAMOS data. But see the list of functions at the top here: 
http://icoads.noaa.gov/software/lmrlib (e.g. Reaumur to Celsius; inches 
(French) Hg to millibars; etc.). -Scott

PS: Another somewhat cosmetic issue is that the "lmr" in lmrlib refers 
to the obsolete ICOADS Long Marine Report format (replaced now by IMMA 
format). If we made an new and ideally more collaborative start at 
renewed and expanded work on lmrlib (+ with or without new humidity 
calculation routines?), a better name would be nice.


On 2/7/14 7:26 AM, Shawn R. Smith wrote:
> Thanks Eric. I will add the WMO 8 code development to the discussion 
> for IVAD1 in June. Seems an appropriate topic when we discuss rdimma, 
> etc. We have no plans at present to calculate dewpoint for the SAMOS 
> data when we convert to IMMA1, but this procedure could be changed in 
> the future once a clear set of routines is available.
>
> Scott, Eric, or Steve - I think it would be beneficial to the IVAD 
> community to have a brief presentation on the lmrlib at IVAD1. I, for 
> one, am not all that familar with what this library offers. There may 
> be useful items in there for the SAMOS to ICOADS conversions, but this 
> is not something we have looked at closely. We work soley with the 
> rdimma code to do our format conversions (with no extra values being 
> calculated).
>
> Cheers,
>
> Shawn
>
> On Feb 7, 2014, at 9:20 AM, Eric Freeman wrote:
>
>> Hi Shawn,
>> I'm using these in Java, so not much use for the Fortran community.
>> It will be useful to have in lmrlib as I transition to Fortan on 
>> coming projects.
>> Thanks,
>> Eric
>> On 2/7/2014 9:18 AM, Shawn R. Smith wrote:
>>> I also agree with using WMO 8. Has anyone converted these equations 
>>> into program source code to allow dewpoint calculation from various 
>>> input humidity types? I know Mark Bourassa has a code like this, but 
>>> may not be based on WMO 8 equations.
>>>
>>> Shawn
>>>
>>> On Feb 6, 2014, at 7:36 PM, Steve Worley wrote:
>>>
>>>> I agree with this approach.  Can we put this in the library and 
>>>> validate the
>>>> code easily?
>>>>
...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/icoads_usa/attachments/20140211/4a859bf3/attachment.html 


More information about the icoads_usa mailing list