Heterodyne Calibration

Observations with RxA and HARP are calibrated on a T_A^* scale (in Kelvin), in which the spectra are corrected for atmospheric attenuation and scattering and spillover seeing ambient temperature.
This is done by combining measurements of hot and cold loads and knowledge of the sky and ambient temperatures. Note there are special considerations when observing targets with a strong continuum such as the moon, the Sun, or planetary atmospheres.

The temperature scale is checked by observing a number of standard sources at certain standard frequencies (depending on the observing program for the night). Also the bandwidth chosen depends on the observing program. Some of these source/frequency observations have been observed very frequently (e.g. CO in IRC+10216, CRL2688 and CRL618) and can be used to monitor the calibration of the telescope. Other combinations have been observed only once.

Results are given separately for `Ū`ū, RxA3 (retired 230GHz receiver) and HARP

In addition regular observations are made of Mars, Uranus, and Jupiter in order to monitor the main beam efficiency η_mb (and the aperture efficiency η_a) of the JCMT. Also some observations of the full Moon have been made. Most observations are made for RxA3 at 230.538 GHz (CO 2-1) and for HARP at 345.796 GHz (CO 3-2), both for a bandwidth of 1000 MHz. (On Mars, the spectral regions from 345.7-345.93 GHz and 230.4-230.7 GHz are excluded from efficiency calculations to avoid the  CO 3-2 and 2-1 absorption lines.)

Results are given separately for RxA3 and HARP.

For point sources the conversion from T_A^* in Kelvin to flux density in Jansky for the JCMT is
S(Jy) = 15.6 T_A^*(K) / η_a.

In general, our rule of thumb is that night time spectral standard observations should be within 10% of the ‘canonical’ value (see below for a more thorough discussion of the true uncertainties). For convenience, we have analysed observations (considered as ‘good’, and not including very anomalous results or poor baselines) towards some of our standard sources at CO 3-2 (HARP) and CO 2-1 (RxA3). The mean peak and integrated value, along with the standard deviation and percentage error are shown in the table below for these sources. Please note that in general the distribution may not be Gaussian (due to pointing errors, which will tend to skew the distribution below the ‘true’ value).  This does not include any RxA3m observations, and excludes day time observations and observations when RxA3 is believed to have been misaligned.

Considerations for Calibration Uncertainty

Categorising and understanding each component of the inherent scatter in the Standard observations is a difficult task that is not well-defined. For small sources compared to the beam, the typical uncertainties in peak flux measurements tend to be in the range of 10-20%, with brighter sources (such as our calibrators) nearer to 10% and faint peaked sources/marginal detections closer to 20%. For observations of extended (especially diffuse) structures, the observing mode, data reduction, and method of analysis can all affect the uncertainties with no simple characterisation of the uncertainties. For extended, diffuse structures, varying uncertainty can be introduced depending on the size of the region and the chop-angle used for background subtraction (as extended emission components can simply be “chopped out”), so this is difficult to analyse systematically. Indeed it is difficult to constraint faint, extended flux measurements obtained by any millimetre or submillimetre telescope to better than 30%.

A few issues to consider when using the HARP instrument are:

1. The SSB noise factor when observing far from the center of the band (5 GHz IF). While most observations are centred optimally on this IF, data can be noticeably affected across the 2 GHz band.

2. The nominal value assumed for the separation between HARP receptors is 30″, but that is only correct to the 2” level.

3. Calibration differences between receptors will also play a role when comparing observing modes because rasters will move the array more across the source than jiggles, where each receptor stays its own small area of the map. Standard source observations are performed by “staring” at the source with just the pointing receptor (H05) so there is no way to collect enough information on a night-to-night basis to compare receptors without the calibration time becoming prohibitive. For more information, see the work by Curtis et al. (2009) in assessing individual factors for different receptors to mitigate striping artefacts: (Section 3.1, “The HARP flat-field”)

https://academic.oup.com/mnras/article/401/1/455/1007610#92356426

This is also described by Jenness et al (2009) for the ORAC-DR pipeline here: (Section 4.8, “Flat-fielding”)

https://academic.oup.com/mnras/article/453/1/73/1747692#25829784