********* NOTE ON THESE FILES ******** 1 NOVEMBER 1994 ********************* The size distributions in these files were not divided by the sample volume, which is located in word 19. Consequently, the values in the size distribution arrays are in counts/size channel. Please make note of this. Corrected files, plus the 2D-C data, are being remade at this time. These should be ready in a week or so, providing that our troublesome EXABYTE drive can limp along without repair for a while. I'll keep you posted via this file. Robert A. Black, NOAA/AOML/HRD Ph. (305) 361-4314 ****************************************************************************** UPDATE: 1 DECEMBER 1994 The original files refered to in the previous note have been compressed and backed up on a TAR tape. The contents of this tape can be read on file tar_1_DEC_94.txt. Any of those files can be supplied on request. The new files contain both the 2D-P and 2D-C data. All known bugs have been fixed; these files are in the format described below. If you wish to see the original (unmerged) 2D-C and/or 2D-P files, please send me an e-mail message describing what you need. The next update will be for the first-look data from the NOAA WP-3D aircraft. Robert A. Black, NOAA/AOML/HRD (rblack@aoml.noaa.gov) PS: The size spectra files were compressed with the UNIX compress command on 11 May 1995 to save disk space. If this is a problem, please contact me via e-mail at the above address. I will uncompress the files for you, and leave them uncompressed for not less than 5 working days. R. A. Black ****************************************************************************** UPDATE: 18 November 1994 The PMS probe data files produced at NOAA/AOML/HRD in support of the TOGA-COARE project were reduced using methods essentially similar to those reported by Black and Hallett (JAS, 1986), henceforth, BH. In the case of the 2-D Mono-probe data obtained in rain by the NCAR Electra, the rejection criteria of BH were extensively altered so that most partial images were accepted as well. This was done because the NCAR data were exceptionally "clean", and free of splashes and other defects. All files are written in HP/Motorola (HILO) format as opposed to the DEC (LOHI) format. USERS OF DEC OR INTEL COMPUTERS MUST SWAP BYTES to read these files. ****************************************************************************** UPDATE: 18 July 1995 The exponential fit parameters given in the files do not accurately represent the spectra because the e-fit was performed on the raw counts before dividing by sample volume. The proper values can be obtained by dividing the N-zero parameter by the appropriate sample volume. R. A. Black ****************************************************************************** UPDATE: 20 January 1996 There is no temperature data in the PMS data record, so all data were processed as though the images were water. Frank Marks and I have recently completed a stratification of the samples by both altitude and temperature. Out of 14,557 6-s 2D-P spectra in the entire data set: a) 1,634 spectra were obtained at T < +3 C. These samples may contain ice. b) 2,284 spectra were obtained at altitudes < 1 km. c) 9,729 spectra were obtained at 1 km < alt. < 3.5 km d) 261 spectra were obtained at alt. > 3.5 km e) 605 spectra had sample volumes < 500 liters. These spectra are also suspect. Most (if not all) of the cold samples were obtained during the ferry to and from the region of interest. The ice data can be processed as ice, but since doing so requires hand inspection of the images to find the mixed and/or liquid phase data, this is extremely labor intensive. I will not do this for all the data any time soon, unless it is crucial to a specific analysis. If such is the case, or further information is necessary, feel free to contact me. E-mail is best; it is the only way to contact me during furlough periods. R. A. Black ****************************************************************************** a). NCAR rain data The NCAR 2-D mono-probes consist of a 0.2 mm resolution precipitation (2D-P) and a 0.025 mm resolution cloud (2D-C) probe. The size dependence of the 2D-C depth of field was NOT accounted for. These data were averaged for 6-seconds. However, the inter-record times on the files are not always 6-s apart because of the asynchronous data collection. Images whose largest dimension was within the diode array were sized by the equivalent circle area. Partial images whose longest dimension was on an edge were sized by the longest dimension, but these particles were counted as fractional images defined as the image area divided by the area of an equivalent circle of diameter equal to the length of the longest dimension. Images that occulted both end diodes were rejected. To account for the fact that the old PMS 2D Mono-probes stripped the first slice off each image, zero-area images were added into the computations as particles with area of 0.5 diodes. All other images had areas increased according to the following: Area (A) < 17, area=area +0.53707 +0.31552A -0.02030841A**2 +0.00050043A**3 Area > 16, area=area +2.0385 +0.016346A. This correction procedure forces zero-area images into the smallest size category, and increased the size of some images in the first few size channels up to the next larger size. Computed parameters include rain rate (mm/hr), radar reflectivity factor (Z mm**6/m**3), liquid water content in both g/m**3 and n/l, liquid mean volume diameter RADIUS (mm), and the y-intercept and slope of the liquid particle size distribution when plotted as Log(n/l) vs. diameter in millimeters. Measured parameters included in the file are date, time, sample volume (liters) for both probes, and elapsed over-load times in seconds for the averaging period. The physical records are organized as follows: Each averaging period of 2-D Mono-probe data from the NCAR ELECTRA contains 7 - 128-byte direct-access records written as 32 4-byte IEEE REAL format values associated with it. The first record of each group of 7 contains computed data, plus the date (YYMMDD) and time (seconds). The contents of the header record are: WORD CONTENTS 1 DATE (YYMMDD.) 2 TIME (SS.MSEC) (end time of averaging period) 3 5.5 cm radar attenuation from 2D-P (dBZ/km) 4 averaging time, seconds 5 True Airspeed (TAS), m/s 6 # records/time this file (7) 7 2DP liquid water content (LWC) (g/m**3) 8 2DP LWC (n/l) 9 2DP rain rate (mm/hr) 10 2DP ice water content (IWC) (g/m**3) (if any) 11 2DP IWC (n/l) 12 2DP radar reflectivity factor, Z (mm**6/m**3) 13 2DP liquid water (LW) Y-intercept, exponential fit) 14 2DP LW slope (exp. fit) 15 2DP ice water (IW) Y-intercept (exp. fit) 16 2DP IW slope (exp. fit) 17 2DP LW mean volume diameter (mvd) (mm) (ACTALLY, RADIUS) 18 2DP Elapsed o/l time (s), this period 19 2DP sample volume (l) 20 2DC LWC (g/m**3) 21 2DC LWC (n/l) 22 2DC rain rate (mm/hr) 23 2DC IWC (g/m**3) 24 2DC IWC (n/l) 25 2DC Z (mm**6/m**3) 26 2DC LW A0 (exp. fit) 27 2DC LW slope (exp. fit) 28 2DC IW A0 (exp. fit) 29 2DC IW slope (exp. fit) 30 2DC LW mvd (mm) (RADUIS, not diameter) 31 2DC Elapsed o/l time (s) 32 2DC sample volume (l) The second and third records are the 2D-P water size distribution as n/l per 0.2 mm size channel, and the fractional image data in the same units, respectively. If it is desired to exclude the partial images, simply subtract, element by element, the data in record 3 from that in record 2. The result will be the image data size spectra using the "center in" technique described by BH. Records 4 and 5 are the same as records 2 and 3, except for the 2D-C. These are divided into number per liter per 0.025 mm size channel. Records 6 and 7 are 2D-C graupel and 2D-C column data, unused for TOGA-COARE. b) NOAA/P3 2D-GREY PROBE DATA (I'm still working on it - not a high priority right now) The NOAA-P3 data are handled in a similar manner to the NCAR data, except that images with their longest dimension on an edge are not utilized. Another difference is that the NOAA aircraft used the new 2D-Greyscale probes along with an SEA Inc. model 200 data system. These differences require a much longer averaging period (at least 60 s) to obtain a stable size distribution. The 2D-Greyscale files are composed of 7 records each 256 bytes long consisting of 64 IEEE real-format words. Preliminary 2D-Grey probe file format As in the case of the 2-D files mentioned above, each N-second average is represented by a group of 7 256-byte long records. Each record consists of 64 IEEE real format 32-bit words. The first of these records contains the date, time, and parameters derived from the 2-D data. Similarly, the second and third records contain the 2DG-P water and ice distributions, and the fourth, fifth, sixth, and seventh records contain the 2DG-C water, ice, graupel, and column distributions, respectively. All concentratons are given in number/liter. The derived parameters (as of 12/5/94) are as follows: WORD PARAMETER NAME UNITS 1 DATE (YYMMDD) 2 TIME (HHMMSS) 3 RAIN RATE mm/hr (MAX of 2D-P and 2D-C) 4 LWC g m-3 (MAX of 2D-P and 2D-C) 5 IWC g m-3 (MAX of 2D-P and 2D-C) 6 2D-P Radar reflectivity dBZ 7 5.5 cm 2-way attenuation dBZ/km 8 Averaging time seconds 9 True air speed m/s (related to the clock rate) 10 2D-P LWC g m-3 11 2D-P IWC g m-3 12 2D-C LWC g m-3 13 2D-C IWC g m-3 14 2D-C Sample vol. L 15 2D-P Sample vol. L 16 2D-C Rain rate mm/hr 17 2D-P Rain rate mm/hr 18 number of rejected records this interval 19 2D-C Radar reflectivity dBZ 20 2D-P "K" factor from Heymsfield et al., (1986) 21 unused 22 unused 23 2D-C "K" 24 unused 25 sum 2D-C pixel area mm**2 26 average 2D-C pixel area mm**2 27 sum 2D-P pixel area mm**2 28 average 2D-P pixel area 29 2D-P overload count (not valid for Grey probe data) 30 2D-C overload count 31 2D-P Liquid counts 32 2D-C Liquid counts 33 2D-P Ice counts 34 2D-C Ice counts 35 unused 36 unused 37 unused 38 unused 39 unused 40 unused 41 unused 42 unused 43 unused 44 unused 45 unused 46 unused 47 unused 48 unused 49 unused 50 unused 51 unused 52 unused 53 unused 54 unused 55 unused 56 unused 57 unused 58 unused 59 unused 60 2DG-C total count (?) 61 2DG-P total count (?) 62 2DG-C counts/record (?) 63 2DG-P counts/record (?) 64 Bulk ice density g m-3 c) NASA DC-8 2D-GREY PROBE DATA Ask Dean Churchill (dean@coconut.rsmas.miami.edu) when these data will be ready. I can display images locally, but that's it. Don't expect too much: time searches on EXABYTE (8 mm) tapes are excruciatingly slow. Please be advised: the NASA DC-8 flies too fast for the 2-D probes to accomodate, thus 2-D pixels in these data are not square. Probe resolutions (parallel to the diode array) are 20 and 40 microns. d) FORTRAN PROGRAMS FOR A QUICK LOOK. Any or all of these are available on request. All are written in FORTRAN, and most are de-bugged on the HP 700 series workstations. Of course, I make no warranty for the performance of these routines on other computers. FDUMP FDUMP prints a specified number of direct-access records to a printer or the screen. This is useful for determining if the file was created properly. The user is responsable for output formatting. USAGE: fdump2 1 999 CHECK CHECK makes sure the times on the file wre created properly, and it allows the user to change the date and times as necessary to account for times extending past midnight, and other time problems. USAGE: check3 2-d_ filename EXTRACT2D EXTRACT2D creates an ASCII file out of the computed parameters from the 2-D data file. (This one doesn't work properly as of 15 Nov. 1995) USAGE: extract2d EXTR_HIST EXTR_HIST creates an ASCII file out of the selected 2-D size distribution(s) The user selects the parameter, start, and stop times, plus provides the name of the output file. CONCAT2D CONCAT2D concatenates 2-D files if necessary. This routine is also useful for removing unwanted data from the direct acess files. USAGE: concat2d PMS_EDIT PMS_EDIT allows the user to modify and change any element in any one of the 7 arrays (records) that make up a 2-D n-second average. Only one element may be changed at once, but this 1 parameter of the n-second average may be altered on all of the 7 record n-second averages in the entire file at once. USAGE: pms_edit PMS_MERGE PMS_MERGE assembles the (separate) 2D-P and 2D-C n-second average files into 1 file, and resolves the conflicts in the record times such that if one time is more than 1 averaging time smaller, the data read from that file makes a separate entry in the merged file. If the time difference is less that half the averaging time, the 2D-P and 2D-C data are merged assuming the 2D-P time is applicable. USAGE: pms_merge Robert A. Black NOAA/AOML/HRD e-mail rblack@aoml.noaa.gov