spacer link to MAST page spacer logo image spacer
link to STScI page

Data Products

The file format description below pertains only to IMAPS-1 data. The status/format of IMAPS-2 files is not known.

IMAPS-1 Data

IMAPS-1 data sets primarily contain images of echelle spectra. The extraction of flux versus wavelength is left for the user although software written in IDL will be made available for this purpose. A complete description of how the IMAPS-1 data was obtained can be found in the paper by Jenkins, Reale, and Zucchino.

The IMAPS-1 data archive is mainly comprised of 2 types of data sets:

  • raw data sets containing the original uncorrected echelle spectral images for one observation,
  • corrected, coadded files containing the corrected, coadded image, the noise filter image, and wavelength information.
Both are stored using the Flexible Image Transport System (FITS) file format.

Raw Data Sets:

Each of the 643 raw data files contain a 320x256 integer array stored as a primary array fits file. The format for raw data set imaps_0507.fits for example, has the following structure:

     Total number of 2880-byte logical records =      58

        header:       1 record with 29 keywords
        data:        57 records

The IMAPS-1 header keywords contain general information on the observed target as well as comments (if any) recorded when the observation was made.

File names for the IMAPS-1 raw data sets are in the form "" where nnn is the original IMAPS exposure sequence number with values ranging from 0079 to 0721.

Corrected, Coadded Files:

Like the raw data, the original IMAPS data was stored in IDL unformatted data sets and was later converted to FITS files. The files contain all the information needed by the project-written IDL extraction software to calculate flux versus wavelength. In general, users will find the corrected, coadded files more useful than the raw files for data analysis.

IMAPS images were obtained in groups and coadded to improve S/N. Generally 2 to 5 images comprised a group. After coadding, the images were processed to remove numerous artifacts and geometric distortions (see Section 7.9 in the paper by Jenkins, Reale, and Zucchino).

There are currently 114 coadded, corrected IMAPS-1 files, although only 75 currently have wavelength assignments. (It is expected that wavelengths will be generated for all the coadded images, however if this is not possible, an alternate file format will need to be determined.)

The basic FITS structure (using file imaps1_337-339.fits as an example), is as follows:


     Total number of 2880-byte logical records =     917
           header:       1 record with 29 keywords
           data:       456 records
      BINTABLE extension:  IMAPS_DATA
           header:       2 records with 43 keywords
           data:         1 record
      IMAGE extension:  FPN1    
           header:       1 record with 13 keywords
           data:       456 records

The corrected, coadded images are stored as 640x512 pixel floating-point primary arrays. A 1-row binary table extension follows containing the following 10 fields:

  • WZERO: fiducial wavelength (in Angstroms),
  • MZERO: fiducial order number,
  • XZERO: x coordinate of the fiducial wavelength (in pixels),
  • YZERO: y coordinate of the fiducial wavelength (in pixels),
  • G: raw image scale factor,
  • NIMAGES: number of coadded images,
  • COEFFS: horizontal slope correction coefficients,
  • ORDER: 50-element vector of order numbers (i.e., 191 - 240),
  • WAVELENGTH: 50-element vector of starting wavelengths (in Angstroms),
  • DELTAW: 50-element vector of wavelength increments (in Angstroms),
  • YEST: 50-element vector of estimated order locations (in pixels).
The IMAPS wavelength assignments are based on the calculated order number, wavelength, and x and y position of one spectral feature within each corrected, coadded image. This calculated value was termed the fiducial wavelength and is stored in the binary table as shown above. (Note that any error in the fiducial wavelength would affect all the wavelengths assignments for that image.) Because observers could apply x and y offsets to view different portions of the echelle spectra however, not all the listed orders may appear in every image.

Also included in the file is the 640x512 pixel floating-point noise filter image (i.e., fpn1) which is stored as an image extension. Multiplying the corrected, coadded image by this template produces the original uncorrected, coadded image. (The fpn1 array is basically included for the IDL extraction software.)

The primary header contains the keywords from the raw data files plus additional information including the exposure sequence numbers of the files used in the coaddition, the midpoint of the observation start times, and the total exposure time.

The corrected, coadded file names are in the form "IMAPS1-nnn-mmm.fits" where nnn is the first, and mmm is the last exposure sequence number used in the coaddition.