November 6, 1997

MEMORANDUM TO:  Record
FROM:           Dennis A. Keyser -- W/NP22
Subject:        Changes to OI/SSI-Data Preprocessor
                (May 1, 1997 Version -- UPDATE # 3)

The May 1, 1997 version of the data preprocessor program
(PREPDATA) was updated for a third time.  UPDATE #3 was
implemented on 1200 UTC 16 July 1997.  All changes now affect
only OBUFR = TRUE switch.  Except where noted, the changes below
affect both the GLOBAL and RGL/ETA models. PREPDATA now
exclusively uses JBUFR file input.

*********************  U P D A T E   # 3  ***********************

I.  GENERAL CHANGES

     1) Removed more logic testing for SEQNTL=FALSE, OBUFR=FALSE,
NLWRD=4, and now also KMAX=12 since these are never satisfied
with output now only to PREPBUFR file.

     2) Streamlined do loop processing in a few places to speed
up the program a bit.  This change also allowed subroutine TAB to
be removed.

     3) Added new Namelist switch GOESRD which, when =TRUE, will
turn on the processing of GOES-I radiances.  The default (=FALSE)
is still used operationally (see III.A below).

     4) Removed the Namelist switch IAFTFI since it is obsolete
and no longer needed.


II. OPERATIONAL CHANGES

A. CHANGES IN THE PROCESSING OF AIRCRAFT DATA

     1) A change in W3LIB routine IW3GAD (picked up by this new
version of PREPDATA) now removes the filter on Air France AMDAR
reports.  These are now being processed operationally.






B. CHANGES IN THE PROCESSING OF SATWND DATA

     1) For U.S. satellite-derived winds, the 7'th and 8'th
characters of the station id (generated now by a corresponding
change in W3LIB routine IW3GAD, picked up by this new version of
PREPDATA) are now encoded in the PREPBUFR file.  These 2-
characters are read in through unpacked ON29 category 8 code
figure 920.  The actual ID is only 7-characters long, so the 8'th
character is a blank here. [The change in IW3GAD actually
reprocesses the JBUFR stn id to make it appear more like it did
in the old ON29 processing.  Characters 3-5 are a serial index,
which allows the report ids to now be unique through 9999 reports
of a particular satellite/product combination (the new 7'th
character allows for 9999 rather than just 999 combinations)].


III. NON-OPERATIONAL (BUT-SOON-TO-BE) CHANGES

A. CHANGES IN THE PROCESSING OF GOES-I SOUNDING DATA

     1) Modified processing to read from a single JBUFR file
using W3LIB routine W3UNPK77 rather than the old read from two
WMO BUFR files which used W3FI77.  Also reads in and processes
18-channels of brightness temperature (based on new namelist
switch GOESRD,see I.2 above).


B. CHANGES IN THE PROCESSING OF NEXRAD (VAD) WIND DATA

     1) The quality marker for VAD wind levels with an input RMS
vector error tag of "B" (approx. 3.9 knots) has been downgraded
from neutral to suspect (at the request of NOAA/FSL).

     2) GLOBAL ONLY: Data are now processed only at the reported
levels.  Before, data were also interpolated to the special
25/50mb auxiliary levels used by the REGIONAL versions.  These
levels were also processed.


C. CHANGES IN THE PROCESSING OF WIND PROFILER DATA

     1) GLOBAL ONLY: Data are now processed only at the reported
levels.  Before, data were also interpolated to the special
25/50mb auxiliary levels used by the REGIONAL versions.  These
levels were also processed.