Please send any comments on this document to Jim Lewis.
There is a good deal of concern regarding the removal of 2d instrumental and sky signatures from infrared imaging data. The ability to model these effects out properly will determine to a large degree the observing strategies with WFCAM. Members at CASU have spent a lot of time working with data from three infrared imagers (CIRSI, UFTI, ISAAC) to get a feel for the stability of the various instrumental signatures that can occur in such detectors.
Fringing is a severe problem with UFTI in the H-band, as it occurs with an amplitude of +/- 5% of sky. This can be removed during the sky subtraction phase using a coaverage of the frames taken over the course of an hour or so. Although the mean fringe intensity varies over this timescale, it is still possible to remove the fringing to a level of less than 0.1% with suitable scaling. With the current service dataset there doesn't appear to be any variation between individual fringes. A sky frame from one part of the night can be fringe corrected with a similar sky frame from another part of the night. Based on our experience with the three imagers, the fringing on UFTI seems to be particularly bad. Similar H-band sky frames with ISAAC show no trace of fringing.
The stability of dark frames, flat fields and reset anomaly are all crucial. Dark frames on all three imagers show mainly reset anomaly. The latter appears to be a function of exposure time and incident flux, so it is important that dark frames be taken for each exposure time used during the course of a night as they can be used to remove the reset anomaly for the corresponding target images. With ISAAC the dark frames appear to be stable from night to night. UFTI seems to suffer from a peculiar form of persistence which is visible in the dark frames after exposure to twilight sky. This is what makes using twilight flats on UFTI impossible. It has been agreed that the best way to flat field data is to use twilight flats. As there is only a small window of opportunity for taking twilight flats during the course of a night, it is essential that the instrumental flat field be stable from night to night or even over the course of weeks, so that library flats can be used when none are available during a night. Tests with ISAAC appear to suggest that flats are stable for that instrument at least over the course of a week and even on timescales of a month.