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

20010727: special metars revisited




Julian,

There are 2 parts to a surface file. One part holds the 
"decoded" information. The second part holds the "raw" information.
The raw information is stored unless the decoder is invoked with the -n flag
as Stonie mentioned.

The "decoded" part contains SA.* (regular metar) reports only. That is 
the part that you normlly see when you set SFPARM=dset in sflist. The typical
configuration of the pqact.conf action for dcmetr with -m 72 will
bin the data in to 20 minute groupings (3 per hour). Stations
like KGXY (Greeley, CO) report 3 regular metars per hour. Eg,

The second part of the file holds the "undecoded" SP and SA reports.
Each time bin in the file can hold 1 regular SA metar, and up to
6 SP special reports (160 characters per). You will see these when you set
SFPARM=text;spcl, or, to see everything that is in the file,
SFPARM=dset;text;spcl. The SPECIs will be indented in the output listing.

What Stonie said the other day is that the SPECIs are not "decoded". That is
true. They are only stored as raw ascii. The decoding effort is reserved for the
regular observations. 

Some of the logic behind this lies in the considerable effor NMAP places in
"time matching" different data sets. By placing the METAR observations into 20
minute bins, there is a regular interval for matching up loop frames. SPECIs 
would truely have to be stored not in bins, but at the observation time to
create a meaningful time matching of the data. That would be detrimental to
matching most data sets.

When you plot the 24 and 40 minute after the hour bins, you see there really are
a large number of regular stations reporting 20 minute observations. In
particular, mesonets in MN and IA provide very dense coverage.

I provided a sample SFLIST output for GXY and DEN below that follows what I just
said.

Steve Chiswell
Unidata User Support

 SFFILE   = metar
 AREA     = @gxy;den
 DATTIM   = 0200-0300
 SFPARM   = dset;text;spcl
 OUTPUT   = t
 IDNTYP   = stid
 GEMPAK-SFLIST>r

    STN    YYMMDD/HHMM      PMSL     ALTI     TMPC     DWPC     SKNT     DRCT
                            GUST     WNUM     CHC1     CHC2     CHC3     VSBY
                            P03D     P03I     MSUN     SNOW     WEQS     P24I
                            TDXC     TDNC     P03C     CTYL     CTYM     CTYH
                            P06I     T6XC     T6NC     CEIL     P01I
    DEN    010727/0200   1011.00    30.09    23.30    13.30    13.00    80.00
                        -9999.00 -9999.00   706.00  1302.00  2503.00    10.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00     9.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00   250.00 -9999.00
    GXY    010727/0200  -9999.00    30.06    24.00    15.00     7.00   120.00
                        -9999.00 -9999.00     1.00 -9999.00 -9999.00    10.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
    GXY    010727/0220  -9999.00    30.05    22.00    16.00     5.00   150.00
                        -9999.00 -9999.00     1.00 -9999.00 -9999.00    10.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
    GXY    010727/0240  -9999.00    30.06    22.00    15.00     5.00   150.00
                        -9999.00 -9999.00     1.00 -9999.00 -9999.00    10.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
    DEN    010727/0300   1012.00    30.10    20.60    13.90     8.00   180.00
                        -9999.00 -9999.00   656.00  1202.00  2203.00    10.00
                         1016.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00     1.60 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00   220.00 -9999.00
    GXY    010727/0300  -9999.00    30.07    21.00    15.00     6.00   210.00
                        -9999.00 -9999.00     1.00 -9999.00 -9999.00    10.00
                         2017.00 -9999.00 -9999.00 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00     1.70 -9999.00 -9999.00 -9999.00
                        -9999.00 -9999.00 -9999.00 -9999.00 -9999.00



KDEN 270153Z 08013KT 10SM FEW070CB SCT130 BKN250 23/13 A3009 RMK AO2 SLP110
VIRGA DSNT SE AND SW CB DSNT E MOV E SHRA DSNT E AND SW T02330133   
  KDEN 270203Z 09010KT 10SM FEW070CB SCT130 BKN250 23/14 A3009 RMK AO2 OCNL
LTGCG
  DSNT SW VIRGA DSNT SW CB DSNT E AND SW MOV E
KGXY 270155Z AUTO 12007KT 10SM CLR 24/15 A3006 RMK AO1  
  
  KDEN 270223Z 17008KT 10SM FEW070CB SCT130 BKN250 21/14 A3010 RMK AO2 OCNL
LTGCG
  DSNT W VIRGA DSNT SW-W CB DSNT W MOV NE TCU DSNT NE
KGXY 270215Z AUTO 15005KT 10SM CLR 22/16 A3005 RMK AO1  
  
KGXY 270235Z AUTO 15005KT 10SM CLR 22/15 A3006 RMK AO1  
  
KDEN 270253Z 18008KT 10SM FEW065 SCT120 BKN220 21/14 A3010 RMK AO2 SLP120 VIRGA
S-NW CB DSIPT SHRA DSNT SW T02060139 51016 $
KGXY 270255Z AUTO 21006KT 10SM CLR 21/15 A3007 RMK AO1 LTG DSNT W 52017 
  
 Parameters requested: SFFILE,AREA,DATTIM,SFPARM,OUTPUT,IDNTYP.
 GEMPAK-SFLIST>





On Thu, 26 Jul 2001, Julien Chastang wrote:

> Steve and Stonie,
> 
> I am reviving the thread concerning special metars from earlier this week.
> Laurie Carson and myself have re-read your e-mails carefully and we have
> furthermore done some of our own tests. (We had to do this because we were
> receiving conflicting assumptions from different people.) We basically
> wish to confirm that it is currently impossible to decode special metars
> into ASCII format with GEMPAK. We came to our conclusion by experimenting
> the pqact.conf file. As shown below, we attempted to decode the "regular"
> and "special" metars into different files, and run SFLIST on them. Running
> SFLIST on the "regular" metar decoded file did indeed yield a lot of data.
> Running SFLIST on the "special" metar decoded file gave us no data. I do
> not currently know whether the problem lies with dcmetr or SFLIST.
> 
> Curiously, the decoded ****_sao.gem file for special metars is not trivial
> in size (5M) compared with the regular metar file which is 15M. This *may*
> suggest that special metars have been decoded into the GEMPAK internal
> format (the ****_sao.gem file) by dcmetr, but SFLIST is not able to
> further decode that data into ASCII format. Therefore, the problem may lie
> with SFLIST and not dcmetr. What do you think?
> 
> ***************Snippet from pqact.conf************************
> 
> #Regular and special metars
> 
> DDS|IDS ^S[AP].* .... ([0-3][0-9])([0-2][0-9])
>         PIPE    decoders/dcmetr -b 9 -m 72 -s sfmetar_sa.tbl
>         -d data/gempak/logs/dcmetr.log
>         -e GEMTBL=/home/ldm/gempak-5.6/gempak/tables
>         data/gempak/surface/YYYYMMDD_sao.gem
> 
> # Regular metars
> 
> DDS|IDS ^SA.* .... ([0-3][0-9])([0-2][0-9])
>         PIPE    decoders/dcmetr -b 9 -m 72 -s sfmetar_sa.tbl
>         -d data/gempak/logs/dcmetr_r.log
>         -e GEMTBL=/home/ldm/gempak-5.6/gempak/tables
>         data/gempak/regular.metar/YYYYMMDD_sao.gem
> 
> # Special Metars
> 
> DDS|IDS ^SP.* .... ([0-3][0-9])([0-2][0-9])
>         PIPE    decoders/dcmetr -b 9 -m 72 -s sfmetar_sa.tbl
>         -d data/gempak/logs/dcmetr_sp.log
>         -e GEMTBL=/home/ldm/gempak-5.6/gempak/tables
>         data/gempak/special.metar/YYYYMMDD_sao.gem
> 
> *******************end************************************
> 
> Being able to decode special metars into an ASCII decoded form is
> important for our project. We were wondering how difficult it would be to
> modify the source in order to have special metars available in ASCII
> decoded form.
> 
> Thanks for your help and time.
> 
> -Julien.
> 
>