[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: metar netCDF to .lso for LEAD



On Fri, 18 Mar 2005, Tom Baltzer wrote:

> Hi Keith,
>
> I'll embed.
>
> > -----Original Message-----
> > From: Keith Brewster [mailto:address@hidden]
> > Sent: Wednesday, March 16, 2005 5:21 PM
> > To: Tom Baltzer
> > Cc: Kelvin Droegemeier; Dan Weber; Kevin Thomas
> > Subject: metar netCDF to .lso for LEAD
> >
> > Hi Tom,
> >
> > I have completed a first draft of a program to write the ADAS .lso
> > file from the UNIDATA netCDF file of metars that you supplied.  We
> > can use this as a stop-gap for now and later consider converting
> > this program to a subroutine embedded in ADAS...though the data
> > storage is significantly different than adas so we'd have to declare
> > a bunch of new variables, allocate and deallocate. This combined
> > with the netcdf/HDF4 linking incompatibility makes it seem better to
> > let it stand as a separate program. The program could use a few more
> > bells-n-whistles, but this should do the trick for now.
> >
>
> That's great!  I won't worry about trying to get the GEMPAK decoder as part
> of the LEAD LDM/Decoder setup.  Thanks!


hi,

the metar2nc decoder was written using "The manual on codes WMO 306"
section FM 15-IX Ext METAR as a guide line so values are interpreted
according to the manual and not as a direct conversion of a value.
Also, some of the errors are my misinterpretation of the wording in the manual.

>
> > One error I discovered (or maybe its a feature) is that the variable
> > which is supposed to be station pressure (labelled "air_pressure")
> > is really just the altimeter setting in the report converted to mb
> > from inches.

that is correct if the group is AIIII but not correct if it is QPPPP then
value is air pressure in hectopascals at the station.  Most european
stations report altimeter not air pressure. my mistake was the coversion
of altimeter to hectopascals and calling it air pressure. this will be
corrected.  there are routines to handles the conversion but they are
site dependant and usually off about 10% and the error gets worst with
higher elevations.

So its actually closer to sea level pressure than
> > station pressure.  I had a subroutine already to convert altimeter
> > setting to station pressure that I installed for my application, but
> > I wonder if other users/programs are using the wrong values for
> > pressure.
> >
> > Also I noticed some more-than-roundoff error in your conversions of
> > feet to meters (in cloud height)

cloud amount and cloud height. section 15.9  height is given by hhh each
subscripted with an s.  height shall be interpreted according to Code
table 1690.

table

000 < 30
001 30
002 60
...
...
999 30,000 or more

decoder routine:

# convert cloud height to  meters
sub cloud_hgt2_meters {

my( $value, $meters ) ;
( $value ) = @_ ;

if( $value == 999 ) {
        $meters = 30000 ;
} else {
        $meters = 30 * $value ;
}
return $meters
} # end cloud_hgt2_meters

in my opinion the calculation to meters is correct.


and knots to m/s in wind & gusts.

decoder code: it using 1knot/s = 0.5175m/s

                        if( $UNITS eq "KT" ) {
                                $UNITS = "m/s" ;
                                $convert = 0.5175 ;

seems correct to me.

 I
> > didn't dig into the decoder source code, but are you using something
> > simple like 3 ft per meter and 2 kts per m/s for that?
> >
> > Finally cloud report of OVC comes out with a "cloud fraction" of
> > 0.8.  Is that supposed to be eighths/10??  That is what the numbers
> > appear to be to me.  I can work with the numbers as given, but for
> > our cloud analysis purposes I'm multipling your value by
> > (10./8.=1.25) to get numbers in the range 0-1.0.


noop, it oktas of the sky ranging in eights of the sky.

From the manual:
Group NNN Category of the cloud amount, scattered, broken, or overcast,
given by three letter abbreviations SCT (1-4oktas), BKN (5-7oktas) or
OVC(8oktas)


As i said before the manual does calculation not directly and in some
instanteous so it gives large estimates.  i would suggest you get a copy
of the manual for your own work. also, you review of the manual code
would check if there are errors in the decoder. i'm always trying to make
the decoder more accurate according to the manual.

thanks,
robb...

> >
> > I was a little disappointed to see this netCDF file continue the
> > GEMPAK legacy of storing cloud information as low-middle-high
> > clouds.  This means if there are more than 3 layers, some
> > information is lost from the original report.
> >
> > -Keith
>
> I'm cc'ing Robb Kambic, the author of the metar2nc decoder to get him to
> weigh in on this.  Perhaps the decoder needs revisiting on these issues.
>
> Thanks,
> Tom.
>
>

===============================================================================
Robb Kambic                                Unidata Program Center
Software Engineer III                      Univ. Corp for Atmospheric Research
address@hidden             WWW: http://www.unidata.ucar.edu/
===============================================================================