2. SCUBA details
SCUBA is the new sub-millimeter continuum instrument for the James Clerk Maxwell Telescope on Mauna Kea, Hawaii (for a review see Holland et al. 1999). It uses two arrays of bolometers to simultaneously observe the same region of sky, in diameter. The arrays are optimized for operations at 850 and 450 µm. Fully sampled maps of the region can be made by "jiggling" the array. Thus this mode is appropriate for mapping the better localized GRB error boxes as well as those of the X-ray transients. This mode can also be used to look for extended quiescent counterparts.
Deeper photometry can be performed using any pixel of these arrays. There are also dedicated photometry pixels for 1100, 1350, and 2000 µm observations (that cannot be used at the same time as the arrays). This photometry mode is appropriate for well localized radio or optical transients.
Scan mapping of larger GRB error boxes using the 450:850 filters is also possible. While this mode is ideal for mapping the long thin GRB error boxes that are obtained using triangulation between satellites, e.g. (Hurley et al. 1997), the sensitivity is greatly reduced.
Given the rapid dissemination of candidate optical and radio transients, the photometry mode is the one that we use most often. Except for the 1997 Dec 16 map of GRB 971214, all the results presented in this paper use the photometry mode. In principle, the most sensitive measurements can be made at 1350 µm, though the 850 µm array has an advantage because the multiple bolometers permit a good sky noise subtraction. In photometry mode, for an integration time of 2 hours, we would expect to achieve an rms mJy at 1350 µm, mJy at 850 µm, and mJy at 450 µm. The sensitivities depend significantly on the weather, particularly at the shorter wavelengths. The jiggle maps at full resolution give an rms that is a factor times higher.
To ensure a homogeneous sample we used a similar calibration and reduction procedure for all the observations. The main steps during the data reduction were the application of the flat-field, removal of systematically noisy bolometers, despiking, and sky-noise removal. The flat-field corrects each bolometer both for a systematic relative gain factor and for its exact position on the sky. In photometry observations of a point source with an array, one pixel, typically the central one, is always used to measure the source while all the other bolometers monitor the surrounding sky. Excluding the noisy bolometers and after despiking, the median value of these sky bolometers for each 18-sec integration is adopted as the instantaneous local sky level and this is subtracted from all the data in that integration. Sky-noise removal thus attempts to best account for the absolute sky level as seen by the central pixel during each integration. No systematic effects were observed associated with the exact choice of sky bolometers i.e. whether the whole array was used or only the inner ring around the central bolometer.
During an observation the secondary is chopped between the source and sky at 7 Hz. This is done mainly to take out small relative DC drifts between the bolometers, and also to remove any large-scale sky variations. Thus sky subtraction also happens for observations using the single-bolometer photometric pixels, but it is not as accurate as for the arrays. The term "integration" time in this paper always refers to the "on+off" time, including the amount of time spent while chopped off-source. An 18 sec integration thus amounts to a 9 sec on-source observation time.
A typical measurement consisted of 50 integrations of 18 seconds. Each observation of a source in general consists of several such measurements with pointing and calibration observations in between. Calibration involves establishing both the opacity and the gain for each observation. The opacities at 850 and 450 µm were measured from skydips while using the continuously monitored 1.3 mm opacity as a guideline for any trends between the skydips. For the absolute flux calibration the gain was measured using planets or standard SCUBA secondary calibrators. The dominant uncertainties in the flux calibrations are transient effects such as thermal relaxation of the dish during the night, the weather, and pointing errors resulting from an inaccurate track model. Based on observed variations of the gain factor and signal levels we estimate typical systematic uncertainties in the absolute flux calibrations of 10% at 850 µm and 20% at 450 µm. In general the rms errors of the observations presented here are larger than this uncertainty.
The reduction of map observations is similar to the photometry ones, except that care is taken that no low-level extended emission is subtracted by the sky-noise removal algorithm. As a final step the data is gridded onto a rectangular grid.
© European Southern Observatory (ESO) 1999
Online publication: June 18, 1999