FILE FORMAT FOR THE NCAR ARCHIVE OF NMC UPPER AIR ADP DATA
Paul Mulder
August 1977
rev. Gregg Walters
February 1992, May 2008 and March 2009
This writeup describes the archive file format of NMC observational data for
the period 1973 to the present.
Archive files contain one or more files, each marked by a header record and
often a trailer record. Records are ordered in a file as follows with a maximum
of 6440 ASCII characters per record.
HEADER RECORD (beginning of a file)
1 thru N DATA RECORDS
TRAILER RECORD (not always here)
HEADER RECORD (beginning of next file)
1 thru N DATA RECORDS
TRAILER RECORD (not always here)
REPEATED TO END OF DATA
END OF FILE (end of archive file)
Header records have the following form (first 30 characters):
1 2 3 4 5 6 7 8 9 10
-------------------------------
1 H H H H Y Y M M D D
2 GROUP NAME ------ CODING
3 W A S H I N G T O N
HHHH - GREENWICH HOUR (e.g. 1800)
YY - YEAR OF CENTURY (e.g. 73)
MM - MONTH (e.g. 01)
DD - DAY (e.g. 01)
GROUP NAME
ADPUPA - UPPER AIR ADP DATA (RAOBS AND PIBALS)
AIRCFT - AIRCRAFT DATA
AIRCAR - AIRCRAFT DATA
SIRSOB - SIRS SOUNDINGS
SATWND - SATELLITE WINDS
UPABOG - UPPER AIR BOGUS
ADPSFC - LAND SURFACE ADP
SFCSHP - SHIP SURFACE ADP
SFCBOG - SURFACE BOGUS
CODING (NCEP)
B1 - BINARY DIRECT ACCESS
B2 - BINARY SEQUENTIAL ACCESS
C1 - CODED DIRECT ACCESS
C2 - CODED SEQUENTIAL ACCESS
The 10 characters "WASHINGTON' should appear in positions 21 through 30.
Sometimes only "WASH" appears.
The remainder of the header record is filled with either blanks or 9s. No
further meaningful information is contained.
Data records have the form described in NMC OFFICE NOTE 124. Data records
contain a variable number of variable length reports.
NOTES FROM DENNIS JOSEPH'S TELEPHONE CONVERSATION WITH JIM MCDONALD ON
MARCH 16, 1989
Current cutoffs for final analyses which get into the NCAR archive are 06Z for
the 00Z final analysis and 21:30Z for the 12Z analysis. These are nominal
times and the actual times float somewhat and probably cannot be recovered
exactly.
The ADP receipt times are included in the data record. For surface data, the
receipt time is included in the record identification section immediately
following the observation time (same form as observation time). For upper air
data the receipt time is put in special category 08 fields with an ID code of
105. The parts of the raob and wind reports are identified with letters (next
page) and the time of receipt is given in the data section of the category 08
fields.
Jim will send us information on these and other category 08 fields. His general
comment was that except for a few remote regions of the globe, all data in the
ADP files gets into the analysis; but the data cutoff is later than the analysis
cutoff time. For all data the the nomial data cutoff was 9 hours after after
the observation time until the summer of 1988. On Aug 2, 1988, the 00Z
observation time cutoff for upper air data was increased to 15 hours and on
August 4, the 12Z observation time cutoff for upper air data was increased to
15 hours. The actual cutoff (dump time) should be in the ADP headers following
the word "WASHINGTON" in columns 53 through 56 (hundredths of an hour). It
appears that only ADPUPA (raobs) and ADPSFC (surface synoptic) files contain
cutoff times in the header.
Letter Identifiers for upper air parts
A raob part A
B raob part B
C raob part C
D raob part D
I winds part A
J winds part B
K winds part C
L winds part D