August 25, 1995 MEMORANDUM TO: Record FROM: Dennis A. Keyser -- W/NMC22 Subject: Changes to OI/SSI-Data Preprocessor (July 12, 1995 Version) The July 12, 1995 version of the data preprocessor program (PREPDATA) was updated for the second time. UPDATE #2 was implemented as a CRISIS JIF on 1200 UTC 28 August 1995 on both the HDS and CRAY machines. Except where noted, the changes below affect both versions of the data preprocessor. ******************** U P D A T E # 2 *********************** I. GENERAL CHANGES 1) The default value for namelist switch FWINDO(5) has been changed from 300. to 600. (hours). II. SPECIFIC OPERATIONAL CHANGES A. CHANGES TO RECONNAISSANCE AND DROPWINSONDE DATA PROCESSING (See end of this document for more information on the corrections made to the processing of these data.) 1) In the previous version, the reported elevation was assigned as the height of the surface level for type 182 reports. This was and still is correct for type 182 splash-level drops which report an elevation of zero over water. However, the type 182 low-level reccos, which report an estimate of sea-level pressure, report the flight-level altitude as elevation. Since these flights are several hundred meters above the ground this lead to large residuals in the heights of these estimated sea- level pressures. The program GLOBAL OIQC program had thrown out most of the data, but some had gotten through. The change will hardwire the height of these low-level recco report type 182 surface data to zero. 2) GLOBAL: The previous version contained a lower-limit of 938 mb for accepting type 182 pressures. Pressures below this limit were flagged as bad, and the report was not used by the analysis. During strong hurricanes, a valid sea-level pressure less than this limit could be reported by a drop. This lower limit is being changed to 800 mb to allow such data to pass through. If a type 182 pressure truly is bad, the GLOBAL OIQC should flag it. The REGIONAL version is not being changed, since the REGIONAL OIQC program's response to bad pressures here is uncertain. 3) GLOBAL: All surface mass data in the type 132 drops, as well as the estimated sea-level pressure (and height) in the type 132 low-level reccos, are now flagged for non-use by the analysis regardless of its quality or value. This is being done for two reasons: a) This surface information is already provided to the analysis in the type 182 reports. This will remove the redundancy of information. b) The type 182 pressures are quality controled by the GLOBAL OIQC program, but the type 132 surface (sea- level) pressures are not. In addition, the CQCHT program does not quality control type 132 reports. If the reported sea-level pressure is bad, it would get through to the analysis as a type 132. 4) In a rerun case for 06Z 07 February 1995, a dropwinsonde was incorrectly coded into ON29 with all cat. 2 level pressure values of zero. In the previous version's GLOBAL processing of the type 182 sea- (splash) level pressure report, the 0 mb pressure was flagged as bad (since it was below the acceptable limit of 938 mb), but it was retained since there was no lower limit for processing surface (sea-level) pressure. (In the REGIONAL, levels with a pressure below the limit were/are skipped, so the report was not processed.) The subsequent GLOBAL PREVENTS program failed when it encountered this 0 mb pressure value. PREPDATA is changed to place a lower-limit on the sea- level pressure in a type 182 report. The report will not be processed if the pressure is less than 1 mb in the GLOBAL and 50 mb in the REGIONAL. 5) GLOBAL: The time window for processing type 182 reports is changed from 3-hours to 6-hours on each side of the cycle time. This will be consistent with the time window for preocessing type 132/232 recco/drops (see I.1 above). III. SPECIFIC NON-OPERATIONAL CHANGES A. CHANGES TO GOES-I SATELLITE RETRIEVAL PROCESSING 1) The temperature, height, and specific humidity on the "surface" level of a sounding are now given at best a suspect quality mark. (This is not done for output to the 36-day archive.). In the REGIONAL, the surface level of the mass report gets a suspect quality mark of '3'. =================================================================== Corrections to the Processing of Reconnaissance Data Dennis Keyser 12Z 28 August 1995 Prior to 12Z 28 August 28 1995 the oi-data preprocessor (PREPDATA) processed entire flight-level reconnaissance and dropwinsonde reports with report types 132(mass) and 232(wind). In addition, the surface level of the drops (splash-level) and the estimated sea-level pressure reported from a low-level recco flight were processed as marine surface reports with type 182(mass). Several problems in PREPDATA's processing of the type 182 reports were discovered in August 1995. I am submitting a crisis JIF of PREPDATA to correct these problems, which I have listed below. This will be implemented at 12Z on 28 August 1995. 1) Prior to the fix: PREPDATA was assigning the reported elevation to the height of the surface level in type 182. This was correct for splash-level drops which report an elevation of zero over water. However, the low-level reccos report the flight-level altitude as elevation. Since these flights are several hundred meters above the ground this led to large residuals in the heights of these estimated sea-level pressures. OIQC in the GLOBAL system had thrown out most of the data, but some had gotten through. Both the REGIONAL and GLOBAL versions of PREPDATA are being changed to hardwire the height of these surface data to zero. 2) Last week (prior to the fix) during hurricane Felix, several drops reported a sea-level pressure less than the lower limit of 938 mb for acceptance by PREPDATA. ALL surface types with pressures below this limit were flagged as bad. The GLOBAL version of PREPDATA is being changed to accept pressures as low as 800 mb for type 182 (the limit is still 938 mb for all other surface types). If a type 182 pressure truly is bad, the GLOBAL OIQC should flag it. The REGIONAL version of PREPDATA is not being changed, since the REGIONAL OIQC's response to bad 182 pressures is uncertain. 3) The GLOBAL version of PREPDATA is being changed to flag all surface mass data in the type 132 drops and the estimated sea-level pressure (and height) in the type 132 low-level reccos, regardless of its quality or value. This is being done for two reasons: a) This surface information is already provided to the analysis in the type 182 reports. This change will remove the redundancy of information. b) The type 182 pressures are QC'd by the GLOBAL OIQC, but the type 132 surface (sea-level) pressure is not In addition, CQCHT does not QC type 132 reports. Prior to the fix, if the reported sea-level pressure was bad, it would have gotten through to the analysis as a type 132. The REGIONAL version of PREPDATA is not being changed to flag such data. 4) In a rerun case for 06Z 07 February 1995 using the version of PREPDATA before this fix (by Bert Katz), a dropwinsonde was incorrectly coded into ON29 with all significant level pressure values of zero. PREPDATA has a lower limit for processing upper-air pressure levels; 1 mb in the GLOBAL system and 50 mb in the REGIONAL system. All levels with pressure below this limit are skipped. Thus, these levels with 0 mb pressure were not included in the type 132/232 drops report. In the processing of the type 182 sea- (splash) level pressure report, the 0 mb pressure was flagged as bad (since it was below the acceptable limit of 938 mb), but it was retained since there was no lower limit for processing surface (sea-level) pressure in the GLOBAL system. (In the REGIONAL system, levels with a pressure below the limit were skipped, so the report was not processed.) The subsequent GLOBAL PREVENTS program failed when it encountered this 0 mb pressure value. The REGIONAL and GLOBAL versions of PREPDATA are being changed to place a limit on the sea-level pressure in a type 182 report. The report will not be processed if the pressure is less than 1 mb in the GLOBAL and 50 mb in the REGIONAL (same as pressure check for all levels in the upper-air data). 5) The time window for processing reccos and drops in the GLOBAL version of PREPDATA is 6-hours on each side of the cycle time for types 132/232. However, for the type 182 it had been just 3-hours on each side of the cycle time. The GLOBAL version of PREPDATA is being changed to give type 182 reports a 6-hour window about the cycle time.