ISSA Explanatory Supplement
Gain and Offset Corrections
A file containing corrections, gains and DC offsets, accompanies the IRAS Zodiacal History File (ZOHF) Version 3.0. A description of how the corrections were derived is presented below along with statistics on the corrections.
Version 3.0 ZOHF was used to compute average profiles of the zodiacal emission versus inclination (azimuth angle about the Earth-Sun axis) for the entire IRAS dataset following the method described in Boulanger and Pérault 1988. A zodiacal light profile was computed for each scan by linear interpolation of the nearest average profiles. It was assumed that the zodiacal light dependence on elongation and time was linear between two consecutive average profiles. Gain and DC offset corrections were obtained for each scan by deriving a linear correction that forced the lower envelope of the scan to match the zodiacal light profile computed for that scan. This fitting process was iterated three times, discarding all points with residuals larger than 5 sigma from one iteration to the next. These corrections force the scan to match the average zodiacal light measured by all scans with the same SOP (Survey Observing Plan, roughly a half day of observations) and elongation.
This procedure is valid only if the correction fit is made over data points for which the Galactic emission is negligible compared to the magnitude of the corrections, which are typically a few percent of the zodiacal emission. This condition was satisfied by using only points at high Galactic latitude (|b| >= 25°), where a good correlation exists between the IR and H I emission (Boulanger and Pérault 1988). The Galactic emission in this region is negligible compared to the zodiacal light at 12 and 25 µm and the Galactic contribution at 60 µm was removed using H I data (Boulanger and Pérault 1988). No correction factors were derived at 100 µm due to the variations in the IR-H I correlation across the sky, which prevented subtraction of the Galactic emission with sufficient accuracy.
Correction factors were measured only for scans for which at least 60 data points (30° of scan length) satisfy the selection criteria for low Galactic emission described in the previous paragraph. Therefore, no gain and offset corrections were obtained for short scans and scans which have too few points in regions of low Galactic emission. The file gives correction factors for about 80% of the scans longer than 30° . Statistics on the gain and offset corrections are presented in Tables F.1 and F.2. Table F.1 gives the average value and the root mean square dispersion of the gain and offset corrections. At all wavelengths the average gain and offset corrections are close to one and zero, respectively. This shows that the corrections do not change the overall calibration of the data. The root mean square dispersion of the gain corrections is about 3% for each of the three wavelengths. Table F.2 gives a histogram of the gain corrections. The gain and offset corrections are plotted against SOP and elongation in Figures F.1(a) - (c). These figures show that there is no systematic effect in the corrections with respect to elongation and SOP.
|Number of Scans||Offsets
(2) number of scans without a fit; most of these are short scans
(3) number of scans with a poor fit (correlation coefficient smaller than 0.96)
(4) average offset correction
(5) standard deviation of offset corrections
(6) average gain correction
(7) standard deviation of gain corrections
(8) average amplitude of residuals (RMS dispersion) after subtraction of the fit
|Range||Number of Gains|
|12 µm||25 µm||60 µm|
Figure F.1(a) Gain and offset corrections versus
elongation and SOP, 12 µm
Figure F.1(b) Gain and offset corrections versus
elongation and SOP, 25 µm
Figure F.1(c) Gain and offset corrections versus
elongation and SOP, 60 µm