IRAS Explanatory Supplement
VIII. Sky Coverage, Confusion, Completeness and Reliability
D. Point Source Catalog Reliability and Completeness
D.1 Definitions, Assumptions and Limitations
Chapter Contents | Authors | References
Table of Contents | Index | Previous Section | Next Section
The completeness of the catalog above a given flux density at a specified wavelength is defined as the function of true sources above that flux density which are present in the catalog. The completeness can be quantified by statistical analysis of the detection history in an area with multiple HCON coverage under the assumption that two different detections of a real source are two statistically actually independent events, each with the same probability.
The reliability of the catalog is defined as the probability that a catalog entry corresponds to a true, celestially fixed source of radiation. An unreliable catalog entry would be created whenever a spurious HCON is confirmed by another spurious HCON on any subsequent survey scan. The survey strategy was developed so that observations of one area were repeated several times, thus lowering the probability that a chance detection would be confirmed.
Ideally, the completeness and reliability of sources with flux densities in a given range would be calculated separately for each wavelength. However, because band-merging occurred before hours-confirmation, completeness and reliability unavoidably depend on the spectrum of a source. In particular, two very different definitions of reliability must be understood. The reliability of a source, which is primarily discussed in this chapter, refers to the probability that a source in the IRAS catalog is a true celestial object measured in at least one wavelength band. The reliability of a flux measurement, on the other hand, refers to the probability that the flux density quoted for a source in a given band is a true measurement and not one due to spurious effects such as noise, radiation hits or asteroids. The concept of high and medium quality flux densities is critical in this discussion (Section V.H.5). A high quality flux density measurement is one which has met all the confirmation criteria on timescales of seconds, hours and weeks. The existence of a source is guaranteed and the quality of its flux density measurement in a band is assured according to the completeness and reliability values discussed in this chapter if, and only if, the source has a high quality flux density in that band. If a source has only a moderate quality flux density in a given band, then lower standards of reliability apply (Section VIII.D.4.c). This dual concept of reliability was necessitated by a desire to provide flux information in as many bands as possible for sources whose existence was guaranteed in at least one band.
The estimates of completeness and reliability given here are valid only for sources that are truly point-like. Sources which are slightly extended may still achieve an acceptable correlation with the point-source template in the in-scan direction, but extend across more than three detectors in the cross-scan direction. The point-source processing software resolved such sources into a string of two or more point sources with a spacing of a few arc minutes in the cross-scan direction. Successive passes may have resolved the source in different ways, causing the completeness of such sources to be poor. This problem is a common occurrence with "cirrus" at 100 µm and in all bands in the Galactic plane near the Galactic center. In regions of high source-density such strings of sources were discriminated against by the "confused-neighbor" and "weaker-neighbor" rejection criteria (see Section V.H.6).
In addition, the completeness estimates are valid only for point sources without close neighbors. Close neighbors can (i) cause the detections of both sources to be lost when the confused detections of both sources no longer match a point source template; (ii) cause detections from one source to be lost because of source shadowing (Section V.C); (iii) create cross-scan confusion and alter significantly the cross-scan position of the source.
Table of Contents | Index | Previous Section | Next Section