IRAS Explanatory Supplement
X. The Formats of the IRAS Catalogs and Atlases
E. Low-Resolution Spectra

1. Catalog Header File
2. Spectra Records

E.1. Catalog Header File

As described in detail in Chapter IX, each sample in the spectrum corresponds to a certain in-scan distance of the source from the centerline of the spectrometer entrance aperture. ANGLE lists these angular distances.

Wavelength Calibration.: LAMBDA1, LAMBDA2

There is a non-linear relation between the displacement of the source from the centerline of the spectrometer and the sampled wavelength. The wavelengths corresponding to each sample (or ANGLE) are given in LAMBDA1 and LAMBDA2 for the two wavelength bands. The beginning and end of each spectrum contain measurements that lie outside of the wavelength coverage of the instrument but which can be used for electronic baseline determination. The wavelength values corresponding to these values of sample number (or ANGLE) are set to 0.

The items starting at bytes 0 through 25 are identical to the items with the same names in the point source catalog (see Section X.B.1). This is also the case for the items starting at bytes 71, 73, 75 and 960.

Number of Spectra: NSPECTRA, NACCEPT

NSPECTRA is the number of spectra of this source observed; NACCEPT are the numbers of 8-12 µm and the 11-25 µm spectrum halves ultimately averaged to make the entry in the catalog.

Characterization of the Spectrum.: LRSCHAR

A description of the method of characterization is given in Section IX.D. Table X.E.2 lists the spectral classes used to characterize the spectrum.

Quality of the Spectrum Halves:. SPQUAL

Depending on the signal-to-noise ratio of the 8-12 µm and 11-25 µm halves of the spectra, the number of accepted spectra halves and the difference in level of the baselines on either side of the spectrum halves, a quality digit is assigned to each half of the spectrum; 1 indicates good quality, 2 moderate quality, and 3 barely acceptable.

Scale factor For All Spectrum Flux Densities: SCALE

Multiplying the integers BASELINE, NOISE, and SPECTRUM by the factor SCALE converts the values into units of W m^-2µm^-1.

Baseline of Spectrum Halves: BASELINE

These four values give the average value of samples 1 through 20 (short wavelength end) and 81 through 100 (long wavelength end) of both the 8-12 µm and the 11-25 µm halves of the spectrum.

RMS Noise: NOISE

Using the twenty samples on the long wavelength end used for BASELINE, the rms noise per spectrum half is determined.

Signal-to-Noise Ratio: SNR

The average value of the samples in the wavelength ranges 8-12 µm and 11-25 µm, respectively, are divided by the NOISE values determined.

Baseline Asymmetry: ASYMM

This value indicates by what fraction of the average signal the baselines on the short and long wavelength sides of the spectrum halves differ. A large baseline asymmetry indicates that a confusing source may have contaminated the spectrum. The baseline asymmetry is usually large near the Galactic plane.

LRS/Survey Flux Ratio:. SRATIO

The ratio of the integrated flux (after convolution of the spectrum flux densities with the 12 µm band pass of the survey instrument) in the spectrum and the 12 µm survey flux is given in this item. Normally this value should be close to unity (Section IX.C). As the 11-25 µm part of the LRS spectrum hardly overlaps with the 25 µm survey band, a ratio of LRS/survey for this band is not significant.

The Spectrum: SPECTRUM

The integer values of the spectrum must be multiplied by SCALE for conversion to m^-2µm^-1. The wavelengths corresponding to the 100 samples given for the 8-12 µm and the 11-25 µm halves of the spectrum are given in the catalog header file. Values for non-significant wavelengths are set to zero. For baseline interpolation, either the sample numbers or the ANGLE (distance in arcmin from the spectrometer center line) can be used.