The basic SCUBA jiggle map reduction consists of 4 steps, each of which is done with a separate task: reduce_switch, flatfield, extinction and rebin. For a complete reduction, you will also need some or all of the following tasks: change_flat, change_pointing, change_quality, despike, remsky, scuclip, scuover, and scunoise. For interactive despiking you will also need sclean or dspbol. For scan map data one additionally needs the tasks despike2, scan_rlb, and restore or remdbm.
We now try our first map (scan 86), which is a single observation (3 integrations) of
IRC10216,
one of our secondary calibrators, obtained on December 8 1997. Normally we would do the first three
steps with the script scuquick, i.e. scuquick -tau=0.185 -sub=lon IN 86 out=i86, but here we do
it step by step, so that you can see what is involved. First we execute reduce_switch, and
flatfield:
Next we want to correct for the atmospheric attenuation of the signal. At present there are two measures of the atmosphere’s opacity - , and skydips. A large amount of effort has gone into understanding the relationship between these two over the past few semesters and the results are presented in Archibald et al. [11]. The relationships found for the wideband filters now used are:
| (1) |
| (2) |
A careful observer will examine the data for the night, note whether it appears stable or not (it commonly ‘spikes‘ on nights with poor atmospheric conditions) and compare the results with the skydips obtained (one can use the Surf routine skydip to reduce the data if you want or use Orac-dr). The fits made to 450-dips are not always particularly good, there seems to be a problem with the minimization routine, and so the recommended method of calculating opacity at 450 is always to convert from the 850 or . Which you use is certainly open to debate; the skydips may well have the advantage of being made at the same azimuth as your source, but the is measured more regularly (every ten minutes or so). The preferred method used at the JAC is to fit a polynomial to the data, and convert the value from the fit to opacity at 450 and 850, these fits are regularly made and available at the JCMT’s calibration web page Orac-dr will make use of these fits if one configures it to do so.
Before getting too involved in correcting for sky-opacity it is worth keeping in mind that at 850, and particularly for faint sources, the exact value is not necessary – other uncertainties are likely to dominate. We therefore adopt a value for this observation. The extinction correction is applied by running extinction on the flatfielded data.
Note that extinction allows you to supply two values of opacity measured at different times if you want – in this case we have bypassed this option. This is also where the long wavelength array gets separated from the short wavelength one. When we want the short wavelength data we have to run extinction again and choose short as the . Our data are now extinction corrected, but still in instrumental units (Volts). In order to have a feeling for the true signal and noise level in our data, we therefore need to apply a scaling factor, FCF (Flux Calibration Factor), that converts the instrumental units to Jy/beam or Jy/. Calibration is discussed in detail in Section 7. For just a quick look we ignore the intricacies of calibration and use nominal FCSs, which for the current filters (850 W & 450 W) is 220 and 310 Jy/beam/V for 850 and 450 m, respectively. Since this map was taken with the old 850 m filter, 850 N, we use a different calibration factor, FCF = 280 Jy/beam/V, which is more appropriate. To scale our extinction corrected data we use the Kappa command cmult.
Here we gave the calibrated data set the extension _cal. Now we are ready to convert our extinction
corrected and calibrated data onto a spatial grid using the Surf task rebin.
The resulting map can be viewed with Kappa’s display
or by using Gaia. The resulting map does not look particularly nice, because we have not yet blanked
out any noisy bolometers, done sky noise reduction or despiking.