July 10, 2001 MEMORANDUM TO: Record FROM: Dennis A. Keyser -- NP22 Subject: Changes to IBM-SP Data Preprocessor (PREPDATA) (February 17, 2000 Version -- UPDATE # 5) The February 17, 2000 data preprocessor program (PREPDATA) was updated for the fifth time. UPDATE #5 was implemented on the IBM-SP machine at 12Z on July 10, 2001. Except where noted, the changes below affect all five versions: ETA, AVN, FNL, RUC and CDAS. **************** I B M - S P U P D A T E # 5 *************** I. GENERAL CHANGES 1) Added new subroutine TIMCHK which performs time window checking for all reports. This step had been included in subroutine RPTLBL, but it was separated to allow separate time window checking to be performed on sounding/radiance, precipitable waterm and cloud-top reports all generated from a single decoded GOES sounding/radiance/retrieval report which otherwise only needs to call subroutine RPTLBL a single time for each decoded report. [MAIN, PREP, RPTLBL, TIMCHK, GETC06, SATEDS, SATBFR, GOESCHK, SFCDTA, GETSMI, GETSCATT] 2) Existing namelist switches GOESND, GOESPW and GOESRD expanded from logical scalars to 1-dimensional arrays of size 2. The first word refers to GOES 5x5 fov sounding/radiance/retrieval data (low-resolution, as before) and the second word refers to GOES 1x1 fov sounding/radiance/retrieval data (high-resolution, new). All default to 2*FALSE. [MAIN, PREP, GOESDG, DATACARDS] 3) Added new namelist switch GOESCT, a logical 1-dimensional array of length 2 which controls the processing of GOES cloud top reports from either 5x5 fov sounding/radiance/retrieval data (word 1) or 1x1 fov data (word 2). Defaults to 2*FALSE. [MAIN, PREP, GOESDG, DATACARDS] 4) Modified namelist switches IWINDO, SATMST, ISATLS, IRTRV and KTOP to be 3-dimensional arrays of length (5,4,4) rather than 2-dimensional arrays of length (5,6). The second dimension now refers specifically to the satellite (even-1, odd-1, even-2, odd- 2 for GOES; even, odd, reserved, reserved for RTOVS/ATOVS). The third dimension refers to the type of report being processed (RTOVS/ATOVS soundings/radiances, GOES soundings/radiances, GOES precipitable water retrievals, GOES cloud top retrievals). See code for defaults. (Note: There are two new possible values for IRTRV - 118 which means exclude only unknown path types and 4 which means process pnly unknown path types. Both apply only to GOES data since cloud top reports have unknown path type.) [MAIN, SATEDS, SNDTBL, SATBFR, GOESDG, DATACARDS] 5) As a result of I.4, namelist switches JSATLS and FLGOES are no longer needed and are removed. [MAIN, SNDTBL, GOESDG, DATACARDS] 6) Modified subroutine SNDTBL to add new input argument IDSAT (length 4, where 1 is GOES even-1 sat. or RTOVS/ATOVS even sat, 2 is GOES odd-1 sat. or RTOVS.ATOVS odd sat., 3 is GOES even-2 sat., and 3 is GOES odd-2 sat.). These replace input arguments IEVSAT and IODSAT, and allow the pairs of even and odd GOES satellite numbers to to printed out in the table information. [MAIN, SATEDS, SNDTBL, SATBFR, GOESDG] 7) Changed the length of the array holding counts of total satellite (RTOVS/ATOVS or GOES) sounding/radiance/retrieval data (NUMOB)from (5,4,2,2) to (5,4,2,4,4); and changed the length of the number kept (KEPOB), and number of moisture data kept (MSTOB) from (5,4,2) to (5,4,2,4). These changes allow counts to be kept based on pairs of even or odd GOES satellites rather than just a single even and a single odd satellite (the expansion of the fourth element in NUMOB from 2 to 4) and based on the 4 types of reports that can be processed (the new fifth element in NUMOB and the new fourth element in MSTOB and KEPOB. [MAIN, SATEDS, SNDTBL, SATBFR, GOESDG] 8) Subroutine SNDTBL has been extensively modified to present an improved satellite sounding/radiance/retrieval table, especially for the GOES data. [MAIN, SNDTBL] 9) Added a new Table D sequence PLEVLG which will be used to define the Table A entry GOESND pressure level sequence. It is different from PLEVL (which still defines pressure level sequence for all other Table A entries) in that in includes the cloud top information sequence (CTINFO which itself contains the new sequences CTPEVN, CTPBAK and CTPPST) and it doesn't include any wind information sequence. Also, because of this change, PLEVL no longer includes the precipitable water information sequence since no Table A entries need it. [BUFRTABLE] II. OPERATIONAL CHANGES SPECIFIC TO A DATA TYPE A. CHANGES TO WIND PROFILER DATA PROCESSING 1) Subroutine IW3UNPBF modified to remove a temporary fix which hardwired the vertical significance (VSIG) to 2 (if surface, never was) or 64 (winds-by-height) for wind profilers from PIBAL bulletins. This had been in place because the profilers from PIBAL bulletins do not contain VSIG. The decoder has since been modified to hardwire these values. [IW3UNPBF] B. CHANGES TO GOES DATA PROCESSING 1) Added new subroutine GOESCHK which performs the checks on GOES sounding/radiance/retrieval reports that had been done in subroutine GOESDG [e.g., report outside domain, outside time window, over land (when only want over ocean), over ocean (when only want oner land), or with unwanted path]. This makes the code more consise now that these checks are performed separately on sounding/radiance, precipitable water, and cloud-top reports all generated from a single decoded GOES sounding/radiance/retrieval report. [MAIN, GOESDG, GOESCHK] 2) Subroutine W3UNPKB7 modified to generate proper station id character 8 for reports with an unknown satellite path (either K or S for even-1 or even-2, or N or T for odd-1 or odd-2 satellite). It also writes an H in the first character of the station id for 1x1 fov reports to identify them as such. Note: All of this is invoked only if the station id wasn't already generated in the upstream program BUFR_DUPSAT. [W3UNPKB7] 3) Subroutine W3UNPKB7 modified to recognize BUFR type/subtype 003/002 (GOES 1x1 fov data) as part of the "GOESND" dump file. Reports in this subtype currently do not have any decoded profile levels or radiance channels. Reports in this subtype write cloud cover (BUFR c.f. 0-20-010, per cent) in output category 8, code figure 258 rather than cloud amount as the old 5x5 fov (003/001) reports do. [W3UNPKB7] 4) Eta and RUC versions only: Can now read in GOES sounding/radiance/retrievals reports from GOES 1x1 fov (high- resolution) data, included in the GOESND data dump file under BUFR type/subtype 003/002. Currently, only GOES cloud top retrievals are available in the input dump. The GOES cloud top reports are encoded into the PREPBUFR file under report type 151 (switch GOESCT=FALSE,TRUE). They have station id's with first character H, and the eighth character identifies the satellite under the unknown path type combinations (see II.A.2). Information encoded into report type 151 include the new mnemonics CDTP (cloud top pressure observation), TOCC (cloud cover observation), GCDTT (cloud top temperature observation), CDTP_QM (cloud top pressure quality marker), CDTP_PC (cloud top pressure program code), and CDTP_RC (cloud top pressure reason code). In addition, slots are left open for other programs to someday encode CDTP_AN (cloud top pressure analyzed value), CDTP_FC (cloud top pressure forecast value), and CDTP_OE (cloud top pressure observation error). Currently, a report will not be written into the PREPBUFR file if the cloud top presssure observation is missing. These data are not yet assimilated (report type 151 is not recognized), but will eventually be assimilated into the RUC analysis and used directly by the Eta model. [MAIN, FILLX, SNDTBL, GOESDG, W3FIZZ, FIZZ01, BUFRTABLE] 5) Subroutine GOESDG has been extensively modified to streamline the processing of GOES reports. [MAIN, GOESDG] 6) Set GOES precipitable water report reason codes (PW1R, PW2R, PW3R, PW4R) to 100 (were missing). [MAIN, FIZZ01] C. CHANGES TO SSM/I DATA PROCESSING 1) Set SSM/I precipitable water report reason code (PWR) to 100 (was missing). [MAIN, FIZZ01] 2) Set SSM/I rainfall rate reason code (REQ6_RC) to 100 (was missing). [MAIN, FIZZ01]