HTML automatically generated with rman
Table of Contents


hi - ascii format for (ADC and other) HI surveys


#include <hi.h>


hi(5) describes the raw ascii HI survey dataformat from the standard NSSDC data archive. Five datasets are currently available from the Astronomical Data Center (ADC) in a more-or-less uniform format (albeit not FITS). Other datasets have been converted and added to this database to provide an all sky coverage. Users can define their own format, for which an extra slot is available. Subsequent processing (mapmaking) can be done with hispectrum(1NEMO) .

The ascii format is basically as follows: each record, with line length reclen (excluding the newline), consists of a small header followed by nchan values of the spectrum. The header must at least contain the pointing position, glon and glat, and sometimes variables that are only known to the decoding routines (see decode.f). All values in the header are integers, typically 10 times or 100 times, the actual real value, but this is certainly no requirement. It merely saves space.

Here are some properties of the available surveys from $HI_DIR:

id    reclen    nrec    nchan    Survey    Authors    Source
hh    643    134932    100    Hat Creek High-Latitude     (Heiles+ 1974-1976)     ADC 8007
pk    420    9891    64    Parkes High-Latitude        (Cleary+ 1979)           ADC 8008
ar    396    55470    56    Argentina High-Latitude     (Colomb 1980)           ADC 8009
bl    703    22261    124    Bell Laboratories           (Stark+ 1990)            ADC 8010
ww    1204    38961    238    Berkeley Low-Latitude         (Weaver, Williams 1973)    ADC 8011
jo    1314    38961    260    Berkeley Low-Latitude         (Weaver, Williams 1973)    JO fits
kj    1514    17901    300    Parkes Southern Low-Latitude     (Kerr+ 1986)       JO fits
lb    1130    1842    224    Galactic Center                 (Burton, Liszt, 1983)    Liszt fits

Some telescope properties, in particular the conversion from antenna temperature to brightness temperature, are:

id    beam    spacing    spacing    TA/TB
            (deg)    (deg)    (km/s)    conversion
hh    0.5    0.5    2 :    ?    Hat Creek 26m (85ft)
pk    0.5    0.5    2 :    ?    Parkes 64m (210ft)
ar    0.5    2    4    ?    PPI 30m
bl    2    2    5.2767    ?    7m Bell Labs
ww    0.5    0.25    1.055    ?    T_A Hat Creek 26m (85ft)
jo       0.5      0.25     1.0    ?    T_A Hat Creek 26m (85ft)
kj    0.5    0.25    1.0    0.975    T_A Parkes 18m
lb    0.5    0.5    2.75    0.66    NRAO 140ft


On a UNIX system one can read the ADC tapes as follows:
    # ADC tape 1, skip the software, and read the data
    mt -f /dev/rmt12 fsf 1
    dd if=/dev/rmt12 of=hh.hi ibs=8464  cbs=643 conv=unblock
    dd if=/dev/rmt12 of=pk.hi ibs=8464  cbs=420 conv=unblock
    dd if=/dev/rmt12 of=ar.hi ibs=8464  cbs=396 conv=unblock ?
    dd if=/dev/rmt12 of=bl.hi ibs=14060 cbs=703 conv=unblock
    # ADC tape 2, skip the software, and read the data
    mt -f /dev/rmt12 fsf 1
    dd if=/dev/rmt12 of=ww.hi ibs=12040 cbs=1204 conv=unblock
    # JO fits files
    fitstohi in=fits id=482:702:1 lon=240:350:0.5 out=kj.hi cdelt1=1000
    # BL fits cube
    fitstohi in=hicenter.vlb out=lb.hi
The conv=unblock causes a newline to be inserted after each record, such that (in principle) fgets(3S) and wc(1) can be used for processing (but see BUGS below).

On a VMS system the DUMP_NSDC program (generally found in the first file on the ADC tape) can be used to convert (ADC) tape data into a disk file.


The pk format contains a NULL (0x00) at position 36 in each record. It separates header from spectrum. decode_pk can handle this, but it is not always clear that fgets would, hence programs like hispectrum should use fread(3) with the correct record-length to read data.

The ww seems to contain bad data for L=11.5, 29 and 91.5. Use jo instead, which has been derived from the equivalent VB-FITS files.

Ra and Dec are not known for the WW, LB and KJ datasets.

If you modify/add a new dataformat, you have to modify hi.h, decode.f and the switch-table in hispectrum.c. This should have been clearly isolated to one routine.


LB: The original data from H.S.Liszt (hicenter.vlb) is a FITS cube of spectra in l and b, with the velocity axis specified in m/s. Reference: W. B. Burton and H. S. Liszt, 1983 A&A Suppl 52, p. 63.

KJ: The original data from P. Jackson consist of 221 separate VB fits files, each one at a different galactic longitude. Created by J. Ohlmacher and P. Jackson (11-dec-84) Reference: Kerr, F. J., Bauer P.F., Jackson, P.D. & Kerr, M., 1986 A&A Suppl 66, p. 373.

JO/WW: Gridded fits data are 481 separate VB fits file, each on a different galactic longitude. Created by J. Ohlmacher and P. Jackson (14-dec-81) Reference: Weaver, H. & Williams, D.R.W., 1974 A&A Suppl 17, p. 1.

PK Reference: Cleary, M. N., Haslam, C. G. T. & Heiles, C. 1979 A&A Suppl 36, p.95-127.

Adc Acknowledgements

By using these programs and the data that they access, the user also agrees to the following:
  1. If data from any survey are used in a publication, the user will reference the original authors of the survey, as if the original authors had themselves provided the data to the user, by quoting the appropriate publication in the usual fashion.
  2. .
  3. Whenever data obtained from the NSSDC are used in a publication acknowledgment shall be made to Dr. Wayne H. Warren Jr. and/or the Astronomical Data Center/National Space Science Data Center for storing and distributing the data.
  4. .
  5. If the survey data or software described herein are used, acknowledgment shall be made to William T. Reach for creating the tape and writing the software.

There is no implied guarantee that these routines will perform the tasks they claim. In other words, user beware--check the results before using them.

See Also

hispectrum(1NEMO) , fitstohi(1NEMO) , fitshicube(1NEMO) , decode.f


Peter Teuben, William T. Reach (decode.f)


$HI_DIR/??.hi                 all ascii versions [200 Mb]
$HI_DIR/??.hi.Z             compressed ascii [49 Mb]
$HI_DIR/??.hi.gz             compressed ascii
$HI_DIR/fits/hifits.*         WW (1-481) and KJ (482-702) VB maps [33 Mb]
$HI_DIR/fits/hicenter.vlb    Burton & Liszt VLB cube [1 Mb]
$NEMO/usr/pjt/hi                   NEMO programs source code

Update History

29-jun-92    V1.0: Created                       PJT
 7-jul-92    documented all associated data        PJT
8-apr-94        added JO dataset                      pjt
11-apr-94    renamed .raw extension to .hi          pjt

Table of Contents