February 6, 2007 MEMORANDUM TO: Record FROM: Dennis A. Keyser -- NP22 Subject: Changes to IBM-CCS Data Preprocessor (PREPDATA) (September 9, 2004 Version -- UPDATE # 5) The September 9, 2004 data preprocessor program (PREPDATA) was updated for the fifth time. UPDATE #5 was implemented on the IBM-CCS machines at 12Z on February 6, 2007. Except where noted, the changes below affect all eight versions: NAM, NDAS, GFS, GDAS, RUC, CDAS, RTMA and AFOSPLT. *************** I B M - C C S U P D A T E # 5 ************** I. GENERAL CHANGES 1) Increased size of arrays SWINDO_e, SWINDO_l, SWNLND, JSWIND and JSMASS from (6,7,10) to (6,8,10) to allow EUMETSAT satellites to be split into separate processing for odd (*,5,*) and even (*,6,*) satellites rather than all EUMETSAT satellites in (*,5,*) as before [(*,7,*) and (*,8,*) now refer to AQUA and TERRA satellites, respectively instead of (*,6,*) and (*,7,*) as before] (defaults are same as previous EUMETSAT-all for both odd and even satellites) [MAIN, DATACARDS, BLOCK] II. OPERATIONAL CHANGES SPECIFIC TO A DATA TYPE A. CHANGES TO SATELLITE-DERIVED WIND DATA PROCESSING 1) Modified EUMETSAT processing to consider "odd" and "even" satellites separately for process/no-process/flag all/flag every-other, time window, land/no-land, all by lat band and product type. [MAIN, SWNTBL, GETC06, DATACARDS] 2) GFS, GDAS, CDAS, CDC, NAM, NDAS only: With the expansion of the “satwnd” dump to now include EUMETSAT visible and water vapor tanks, the METEOSAT-5 visible data are now processed and used by the assimilation. The METEOSAT-5 WV data are now flagged for non-use by the assimilation. [UPSTREAM SATWND DUMP] 3) GFS, GDAS, CDAS, CDC, NAM, NDAS only: IR and visible winds from the odd EUMETSAT satellite (currently METEOSAT-8) are flagged so they will not be assimilated by analysis (both even and odd satellites continue to flag water vapor winds). The METEOSAT-8 will be introduced into the "satwnd" dump within the next two weeks. [DATACARDS] 4) AFOSPLT only: IR and visible winds from the odd EUMETSAT satellite (currently METEOSAT-8) are tossed so they will not be plotted by AFOSPLT (since they are not yet assimilated by the GBL-SSI) (both even and odd satellites continue to toss water vapor winds). The METEOSAT-8 will be introduced into the "satwnd" dump within the next two weeks. [DATACARDS] 5) GFS, GDAS, AFOSPLT, CDAS, CDC only: EUMETSAT IR, water vapor and visible satellite winds are now processed with a time window of -0.75 to +0.25 hours about cycle time (the time window had been +/- 0.75 hours about cycle time). This change was made because the future METEOSAT-8 wind sets are available hourly (unlike the METEOSAT-5 wind sets which are available every 90-minutes), and only one wind set time is ever assimilated (here the METEOSAT-8 wind set from 30-minutes prior to cycle time will be assimilated once these are ready to be used). This change does not affect the METEOSAT-5 data. [DATACARDS] 6) NAM, NDAS only: EUMETSAT IR, water vapor and visible satellite winds are now processed with a time window of -0.75 to +0.25 hours about cycle time (the time window had been -1.25 to +0.25 hours about cycle time). This change was made because the future METEOSAT-8 wind sets are available hourly (unlike the METEOSAT-5 wind sets which are available every 90-minutes), and only one wind set time is ever assimilated (here the METEOSAT-8 wind set from 30-minutes prior to cycle time will be assimilated once these are ready to be used). This change does not affect the METEOSAT-5 data. [DATACARDS] 7) GFS, GDAS only: A bug which caused Aqua and Terra MODIS deep-layer water vapor winds to be treated as though they were cloud-top water vapor winds (and thus assimilated instead of tossed) was corrected. As such they are no longer lumped together with the cloud-top water vapor winds. The datacards had been set to toss these, however, since they did not harm the analysis it was decided that the deep-layer water vapor winds should continue to be assimilated. The datacards were changed to process these instead of toss them. [IW3UNPBF, DATACARDS] 8) GFS, GDAS only: Aqua and Terra MODIS IR winds and water vapor (deep-layer and cloud top) winds are now quality controlled according to the value of the NESDIS Recursive Filter Function (RFF) read in (like for GOES satellite-derived winds). Here, winds with RFF < 50 are marked as bad (13), winds with RFF 50-55 are marked as suspect (3), winds with RFF 56-84 are marked as neutral (2), and winds with RFF > 84 are marked as good (1). Prior to this, all winds received a neutral quality mark (2). [IW3UNPBF, GETC06] 9) Modified to handle future GTS GOES winds and future hi-density (BUFR) JMA winds. [IW3UNPBF] 10) Stores percent confidence based on EUMETSAT Quality Index with forecast consistency test (BUFR code table 0-33-197) in category 8 code figure 356. [IW3UNPBF] 11) Stores percent confidence based on EUMETSAT Quality Index without forecast consistency test (BUFR code table 0-33-198) in category 8 code figure 357. [IW3UNPBF] 12) Accounts for EUMETSAT winds now possibly providing an additional quality indicator - NESDIS Recursive Filter Function (RFF). [IW3UNPBF] B. CHANGES TO GOES SATELLITE SOUNDING PROCESSING 1) RUC only: GOES-11 cloud-top data are now assimilated (like GOES-12), they are no longer flagged (switch IRTRV changed from 9999 to 0 for GOES ODD-2 cloud-top data). [DATACARDS] 2) RUC only: GOES-11 clear layer precipitable water data are now assimilated (like GOES-12), they are no longer flagged (switch IRTRV changed from 9999 to 544 for GOES ODD-2 precipitable water data). Cloudy layer precipitable water data from both GOES-11 and GOES-12 continue to be flagged. [DATACARDS] 3) NAM, NDAS only: GOES-11 clear layer precipitable water data are no longer flagged (switch IRTRV changed from 9999 to 0 for GOES ODD-2 layer precipitable water data). Like GOES-12 clear layer precipitable water data (which was already not flagged), they are monitored, but not used, by the NAM-GSI. [DATACARDS] 4) NAM, NDAS only: GOES-11 and GOES-12 cloudy layer precipitable water data are no longer flagged (switch IRTRV changed from 9999 to 0 for GOES ODD-2 layer precipitable water data and changed from 544 to 0 for GOES EVEN-1 layer precipitable water data). They are monitored, but not used, by the NAM-GSI. [DATACARDS]