March 8, 1996 MEMORANDUM TO: Record FROM: Dennis A. Keyser -- W/NMC22 Subject: Changes to OI/SSI-Data Preprocessor (July 12, 1995 Version -- UPDATE # 4) The July 12, 1995 version of the data preprocessor program (PREPDATA) was updated for the fourth time. UPDATE #4 was implemented on 1200 UTC 27 March 1996 on the CRAY (GLOBAL, ETA versions). All changes now affect only OBUFR = TRUE switch. Except where noted, the changes below affect both the GLOBAL and ETA models. ********************* U P D A T E # 4 *********************** I. SPECIFIC OPERATIONAL CHANGES A. CHANGES TO ALL DATA TYPE PROCESSING 1) ETA: The level pressure is now stored in the NOBS array as ten times the nearest integer of the pressure. This always results in a zero for the tenths value. Prior to this, the nearest integer of the pressure multiplied by ten was stored in the NOBS array. The value in the NOBS array is subsequently encoded in BUFR. This temporary fix will allow the REGIONAL BUFR/PREPDA level pressure data to be identical to its value in the old REGIONAL PREPDA format when it is converted back to this format in the first step of the FERR program. The old PREPDA format stores only whole pressure; the BUFR/PREPDA encodes pressure times ten. B. CHANGES TO UPPER-AIR DATA PROCESSING 1) ETA: Several temporary changes are made in the REGIONAL BUFR/PREPDA processing of upper-air data to allow the REGIONAL BUFR/PREPDA data to appear more like that in the old REGIONAL PREPDA output. The latter format is what is expected by the ETA FERR and OI-analysis codes. Specific changes are: a) The deepened temperature is set to missing in the profile if the temperature has a bad quality marker. b) The surface level height (elevation) quality marker is set to good (1) rather than 6 (neutral). c) Interpolated or calculated data on levels where the data are missing or bad can now get a good (1) quality marker if the data on both spanning levels have a good quality marker. Before, such data would receive a neutral (6) quality marker. d) The check for duplicate pressure or height levels is now performed in the same way as for the old REGIONAL PREPDA processing. e) The quality markers for wind and deepened data generated on levels where the data are missing or bad are now determined in the same way as for the old REGIONAL PREPDA processing. f) The test for determining a PIAL report is now performed in the same way as for the old REGIONAL PREPDA processing. g) The test for the last (highest) "good" data level (mass and wind) is now performed in the same way as for the old REGIONAL PREPDA processing. 2) The report sequence number is now stored in NOBS(12) of the header for all reports in the "ADPUPA" file. It is subsequently encoded into BUFR. C. CHANGES TO SURFACE DATA PROCESSING 1) Marine reports only: If the reported station pressure is non-missing and originally has a neutral (6 or 7) quality marker, then the sea-level pressure quality marker is transferred to the station pressure if the sea-level pressure quality marker is keep (0) or purge (14). Prior to this, the station pressure quality marker remained neutral. 2) All surface mass reports (land and marine): The height (elevation) quality marker is now set to that of the station pressure prior to storing the report in the NOBS array and subsequently encoding the report into BUFR. Prior to this, the surface height always received a neutral (6) quality marker since there is no reported quality marker for the elevation. ** - Change 1) above is needed because O.P.C. places a keep flag 'H' or a purge flag 'P' only on the sea-level pressure quality marker in the ON29 SFCSHP file when it quality controls marine pressure data. The logic in PREPDATA is designed to look first for a station pressure for all surface reports. If the station pressure is found, then it's value along with its quality marker are passed into the NOBS array and subsequently encoded into BUFR. In this case, the sea-level pressure and its quality marker are never used by PREPDATA, and the O.P.C. `H' and 'P' quality markers are never honored. In cases where the station pressure is missing, the sea-level pressure and its quality marker are transferred to the station pressure slot in the NOBS array, provided the elevation is less than 7.5 meters. In this case the O.P.C. 'H' and 'P' quality marks are honored. Change 1) is not in effect for the O.P.C. corrected 'C' quality marker on sea-level pressure. The correction is applied only to the sea- level pressure observation and not to the station pressure observation. In all such cases that I've examined, the station pressure has been missing, so the good (1) quality marker associated with the `C' is automatically transferred to the station pressure, along with the corrected sea-level pressure observation. ** - Change 2) above is needed because even with change 1), the O.P.C. keep or purge flag on pressure is only written into the pressure quality marker location. The logic within subsequent GLOBAL data processing programs (OICQBUFR, SSIANL), is designed to "key" on the height quality marker in the processing of surface mass reports (both land and marine). The logic within the subsequent ETA data processing program FERR likewise was keying on the height quality marker to re-generate the mass-level quality marker in the old PREPDA format. (This was actually an error in the FERR data transformation logic, which is being corrected soon to always use the pressure quality marker as the level quality marker for all surface mass reports.) The exception in both cases is when the pressure has purge or bad quality marker. In this case the report is not used by either analysis regardless of the other quality markers. As a result, only the O.P.C. purge quality marker would be honored by either analysis without change 2). The O.P.C. keep flag would never be honored by the either analysis. Actually, for all surface mass data (land and marine), a keep, good, or suspect quality marker on the pressure would not honored by either analysis without this change. D. CHANGES TO SATELLITE WIND DATA PROCESSING 1) GLOBAL (AVN/FNL) only: NESDIS GOES-8 and GOES-9 cloud drift winds became operational at 18Z 21 March 1996 (with AVN run). 2) ETA (EDAS/MESO) only: NESDIS GOES-8 and GOES-9 cloud drift winds became operational at 12Z 27 March 1996 (with MESO run).