Quality Flag Statistics

Quality flags pertaining to potentially unreliable photometry or astrometry are stored in the catalogue. Flags are constructed autonomously during pipeline processing and the details of the flagging algorithms are provided in the section Data Processing: Quality Flagging.

Filter dependence

The figure below displays a typical XMM-OM field through three filters, UVW2, UVW1 and B. As a consequence of both the spectral properties of typical field sources and the effective areas of the filters, the B image is the most crowded and source count rates through B, with a few exceptions, are larger. Five of the catalogue quality flags, readout streaks, smoke rings, diffraction spikes, mod-8 noise and bright neighbours, are the result of bright sources in the field. Many of these qualities can be seen by eye in the B image, as well as source saturation and an enhanced background at the center of the image caused by diffuse light reflected from outside the field of view. 

As we step through the filters towards the UV wavelengths, both the effective area, intrinsic brightness of typical celestial sources and background photon rates decrease. As a consequence so does the number of quality issues. Diffuse scattered light, a readout streak and smoke rings are still visible in the UVW1 image but at a reduced frequency and intensity relative to the B image. The UVW2 image contains a faint central enhancement but no other quality issues related to bright sources.

A typical OM field observed through the UVW2, UVW1, and B filters. The effective area of the OM, source brightnesses (in general), and source density increase towards the optical. The number of quality issues therefore increase towards optical wavelengths.

Quality Statistics

Inspection of the catalogued sources reveal the same trends across the filters. Those flags that are typically set when sources are bright or have bright neighbours dominate the optical images but become less frequent in the UV. In contrast, bad pixels, image edge and extended source flags reveal a relatively flat distribution across the filters. Central enhancement statistics also remain flat despite quantity of scattered light falling as the wavelength decreases. This is because the flagging algorithm makes no attempt to quantify scattered light which depends on pointing direction and exposure time. Sources will simply be flagged if they are located within the fixed-size, central cone whose characteristics have been iterated upon by screening multiple observations. So while, in real terms, the fraction of UV sources found over a scattering-enhanced background is smaller than the optical fraction, our conservative approach to source quality means that this property is not apparent in the quality flags.

At face value, the fraction of edge-of-image sources appears large. But each field is often comprised of multiple sub-images increasing the probability of an image edge impinging upon a source. Edges have an impact on both astrometry and photometry. With reference to both the many potential image windows and spacecraft pointing drift, particular notice of the edge flag should be taken when searching for source variability.

The table below gives the number of sources with each type of flag for each filter (see Data Processing: Quality Flagging for a description of the meanings of the flags). 

Table giving the statistics on the quality flagging of the sources. The statistics were compiled using XMM-SUSS2.1