1998AJ....115..745V

We have conducted a survey of IRAS images of the high-latitude molecular cirrus clouds cataloged by Magnani, Blitz, & Mundy. This paper reports the data reduction methods used in our survey and also studies the photometric accuracy that can be achieved for faint extended sources on IRAS image products. The principal topics covered are the modeling and removal of sky backgrounds at high Galactic latitude, and as a function of ecliptic latitude; the relative accuracy of different background models and methods of background removal; the integration of source flux on background-subtracted images; the proper definition of flux error bars, their calculation, and the minimum uncertainties set by the performance of the IRAS detectors; the creation of multiple-wavelength composite images; and the creation of temperature and opacity maps. We find that most error bars quoted for IRAS fluxes in the literature are incorrect. We conclude that the limiting surface brightness to which Sky Flux Plates may be used to study faint extended sources is approximately 0.05 MJy.sr^–1. The photometric issues discussed herein are relevant to all IRAS image products, including Sky Flux Plates, co-added images, and the IRAS Sky Survey Atlas (ISSA). We compare the accuracy of data reduction on different products and discuss their relative advantages and disadvantages. We are particularly concerned with the accurate removal of zodiacal backgrounds near the ecliptic plane, a region known as the ISSA Reject Set.