SCUBA calibration: a quick guide

Introduction

This document is intended to serve as a simple guide to the main techniques to use in calibrating your SCUBA data. Two corrections must be made to convert your raw voltage measurements into the genuine flux-density of the object:

1. Flux attenuation by the atmosphere - Support scientist: Ming Zhu
2. Sensitivity fluctuations of the instrument and telescope - Support scientist: Remo Tilanus

Before observing: planning observations

1) Flux Attenuation by the Atmosphere

The flux density of the astronomical source gets attenuated by the atmosphere. The magnitude of the attenuation is a function of the atmospheric opacity along the line of sight. The atmospheric opacity is often referred to as "tau" and the process of correcting the flux is called the tau-correction.

The value of tau is obtainable through a number of sources, all of which give zenith tau (tau along the line of sight to the zenith), from which the tau value along the line of sight at a given elevation can easily be computed assuming a plane-parallel atmosphere. This is automatically done by the data reduction software packages SURF and ORAC-DR. The task of the observer is to provide the software with a value for the zenith opacity.

The major sources for this tau value are:

1. JCMT skydips
2. The JCMT water vapour monitor (WVM)
3. The 225 GHz CSO dipper
4. The 350 micron CSO dipper
5. Polynomial fits to the CSO dipper

By convention all measurements are scaled to tau at 225 GHz. Tau values can be easily transferred between different wavelengths and ORAC-DR automatically applies the necessary relations. The tau values from different sources should not deviate too far from one another, but they sometimes still do.

2) Sensitivity calibration of the telescope and instrument

Calibrating the sensitivity of the telescope and SCUBA is done by observing known calibrator sources at similar times to and with similar configurations as the science observations. Unfortunately, calibration in the submillimetre is always complicated by a lack of options. Our primary calibrators are the planets Mars, Uranus and Neptune. However, Mars often comes sufficiently close to the Earth that it becomes very large and our modelling of its brightness becomes less accurate. Hence at these times it can be subject to greater errors, especially at 450 µm.

If the planets are not usefully positioned for your observations, you will have to use a secondary calibrator. We currently have 6 such objects, and information about them is summarised below. As the table shows, most of these objects have peculiarities or subtleties which may make them unsuitable for your calibration. We have an ongoing program to find more secondary calibrators. More detailed information about the secondary calibrators, for use during data reduction, can be found here.

The availability of an planet or secondary calibrator on a given night can be determined using the sourceplot tool; simply type this at any JAC machine's prompt.

Finally, if for some reason you are unable to get any useful calibration observations, we calculate representative Flux Conversion Factors (FCFs) on approximately monthly timescales which can be used. These values are taken from observations of one or two calibrators and can be used to estimate SCUBA's general performance in a given time period. These values will never be as accurate as calibrations taken on the same night and in the same mode as your observation, but they are better than nothing!

Secondary Calibrators for SCUBA - useful observing information

Source name	R.A. (J2000)	Dec. (J2000)	Approx. 850 µm peak flux (Jy)	Approx. 450 µm peak flux (Jy)	Important usage notes
HL Tau	04 31	+18 13	2.35	9.4	May be variable at a very low level. Fairly compact for calibration purposes.
CRL 618	04 42	+36 06	4.6	10.9	Great source - compact, constant. Use whenever possible!
OH 231.8	07 42	-14 42	3 (variable)	N/A	CANNOT BE USED FOR 450 µm CALIBRATION. 850 µm flux varies by ± 15 %. Fairly compact for calibration purposes.
IRC 10216	09 47	13 16	7 (variable)	N/A	CANNOT BE USED FOR 450 µm CALIBRATION. 850 µm flux varies by ± 10 %. Extended source which should use a chop throw of 60" or larger.
16293-2422	16 32	-24 28	15.1	62.7	Very extended source, with a second source just 80" away. Best used only with a chop throw > 60" and a PA > 330° or < 150°. 450 µm numbers are subject to large errors.
CRL 2688	21 02	36 41	5.9	22.0	Extended source at 450 µm. For calibration at 450 µm, use a chop throw of 60" or larger.

During observing

1) Opacity correction during observing

On the summit one often wishes to have a good and fast result even if it is not perfect. To obtain this we strongly advise the use of ORAC-DR. By default ORAC-DR uses the value of the nearest preceding JCMT skydip result. It automatically corrects for elevation and wavelength (for both 850 and 450 microns). We recommend that you carry out skydips before and after every science observation and at least every two hours. If ORAC-DR finds no useable skydips it uses the value of the 225 GHZ CSO dipper, again applying necessary corrections. Quite often the results form ORAC-DR will only differ marginally from the final results in the offline data reduction.

If ORAC-DR cannot find a suitable skydip or a CSO tau value it will not carry out a tau correction and so you will need to use SURF to get an immediate estimate. In SURF the zenith tau for a given wavelength has to be entered by hand. We recommend you use the values of the WVM or, if that device is also not working, the values of the 350 micron CSO dipper. As the values from these instruments represent tau at 225 GHz they have to first be corrected to the respective wavelength using the following relations:

• tau_{850 micron} = 4.02 * tau_{225 GHz} - 0.004


• tau_{450 micron} = 26.2 * tau_{225 GHz} - 0.367

The conversion from zenith tau to the value appropraite for the actual elevation is applied automatically by SURF.

Note that you can find out which tau value is displayed by the JCMT query tool if you hover your mouse pointer over the left hand side value in the display.

2) Sensitivity calibrations during observing

1. What sort of observations should I make?

   The most important rule is that you should preserve the observing mode of your source observations in your calibration observations. The FCF depends on the observing mode and the chop throw used so your calibration observations should mirror your source observations in this respect. For this reason it is best to include a calibration within your MSB, and then provide clear notes as to how often you actually want calibrations to be done, for example. If you anticipate using IRC 10216, 16293-2422 or CRL 2688 as a calibrator therefore, you should consider the notes above regarding the required chop mode for these sources and where possible, change your science chop mode to match.

2. How often should I calibrate?

   This is a decision which will vary greatly from program to program, and will also be affected by the sheer availability of calibrators during your observations. You should make an absolute minimum of one calibration observation during a shift, just to ensure SCUBA is working as expected (diagnosing any problems is always easier with brighter, known sources). Furthermore, you should be aware that the dish shape can change significantly with temperature and so you should aim to calibrate at least four times a night, sampling the time periods well before and after sunset and sunrise. It is also possible that elevation has some effect on the FCF, although we believe it is less important than the temperature. However, your need and ability to remove all these factors from affecting your calibration depend on the accuracy that the program requires, and on source availability. Generally you should try to calibrate as close as possible in time to your science observations, but this is often not possible and you will sometimes find yourselves doing calibrations for several projects at once.

3. How do I assess SCUBA's sensitivity?

   Running ORAC-DR at the summit will provide you with near-instantaneous FCFs. These values depend upon temperature and (therefore) time of day as well as SCUBA's sensitivity so it is difficult to provide precise expected values. However, approximately monthly values are posted here, which can be used as a guideline to typical values. Your support scientist and/or operator should also be able to advise you as to current typical values.

After observations are completed

1) Offline opacity correction

With offline data reduction one wants to obtain the best precision. For that we recommend using the polynomial fits to the 225 GHz CSO dipper which are provided on a quasi-daily basis by the JAC. If no polynomial fit is available for the time of a given observation, observers can contact the JAC to request one. ORAC-DR uses the numerical representation of the fits and automatically applies corrections for the wavelength and to the line of sight elevation. For use with SURF one has to enter a suitable value by hand. SURF automatically transfers zenith tau to the tau along the line of sight but wavelength corrections have to be applied manually (see above.)

As an alternative to the polynomial fit one can use the data from the WVM, which can be queried from the JCMT calibration pages. If neither of the above are possible we recommend using alternatives in the following order of preference: JCMT skydips, the 225 GHz CSO dipper, the 350 micron CSO dipper. The dipper values can be queried from the CSO web pages - do not use the tau value from the log.

If you are unsure how and where to find the values of the different devices, please consult the list below or contact your support scientist.

Locating opacity values and further information

You can find the different tau values and further information in the following locations:

• On the summit the WVM data is displayed in a graphical tool which should be started by the "observer_up" command.
• The OMP Query Tool displays the WVM data if available. If not, the second choice is currently the 225 GHz CSO dipper, and the third choice is the 350 micron dipper.
• The engineering screen at the summit displays the value of the 225 GHz CSO dipper as does the automatic logbook and the data headers. The data headers also record WVM tau at the start and end of every observation.
• The values of the skydips at different wavelengths are displayed bt ORAC-DR. They are also (wrongly) displayed by the online VAX data reduction.

You can find the interactive WVM archive at: http://www.jach.hawaii.edu/JACpublic/JCMT/Observing_info/Weather/wvm.html

and the historical archive at: http://www.jach.hawaii.edu/JACpublic/JCMT/Observing_info/Weather/WVM/historical_wvm_archive/

You can find the data from the two CSO dippers at: http://ulu.submm.caltech.edu/csotau/2tau.pl

You can find the polynomial fits to the 225 GHz CSO dipper at: http://wwww.jach.hawaii.edu/JACpublic/JCMT/Continuum_observing/SCUBA/astronomy/calibration/taufits

Finally, you can find more information on opacity correction on the JCMT calibration pages at: http://www.jach.hawaii.edu/JCMT/continnuum/calibration/calibration.html.

2) Offline sensitivity calibration

Further details can be found in the photometry cookbook, the mapping cookbook and these additional notes which update the mapping cookbook after the 1999 SCUBA upgrade. The calibration procedure varies according to whether you have Photometry or Map data.

Photometry data calibration

Photometry data calibration is simple, since all you need to do is generate a single peak flux conversion factor (FCF) to multiply your science result by. The one key stage is in the choice of ANALYSIS mode in SCUPHOT. In a single photometry integration, the secondary mirror of the JCMT actually moves to make a small 9-point square map, with 2" spacing. The three options for dealing with these 9 data points are SAMPLES, AVERAGES and PARABOLA.

SAMPLES simply views each integration as supplying 9 independent data points and retains all 9N data points in its final statistics, where N is the number of integrations. AVERAGES instead averages the 9 points of an integration together first, and then has N average data points in its final analysis. Whilst these two methods will produce the same mean flux value, they can generate different errors, even accounting for the sqrt(9) factor. For small data sets, SAMPLES is statistically more sensible. Finally, PARABOLA fits a 2-dimensional function to the 9 points to find a peak value, and then uses N peak data points in its final analysis.

For bright sources (such as calibrators), the PARABOLA method is the most useful since it can correct for small pointing errors, and is directly comparable with making a jiggle map of the source. However, for fainter sources, the SAMPLES or AVERAGES methods are likely to be the more reliable. In fact, if the source is perfectly centred in the beam, there is a direct mathematical relation between the average and peak analysis modes for a bright source, depending on the beamsize. Hence at 850 μm, the PARABOLA:SAMPLES flux ratio is 1.075:1, whereas at 450 μm, the PARABOLA:SAMPLES flux ratio is 1.24:1. Both of these values have been verified empirically.

Hence, we recommend using PARABOLA for your calibrator observation and SAMPLES for your science target (unless it is very bright), and using the above ratios to convert the result from the PARABOLA value to a SAMPLES value.

To summarise, the FCF is given by:

FCF = S(expected) / V(measured),

where S(expected) and V(measured) are the expected and measured flux of the calibrator respectively.

Expected fluxes for the planets are generated by the FLUXES program - use the value marked "Flux in Beam". If your calibrator is a secondary calibrator, you can find the expected fluxes on our secondary calibrator webpage.

Map data calibration

We have two major ways of calibrating map data, Jy per beam or Jy per unit area. Choosing between these two methods depends mainly upon the nature of your source and your required results. For a much more in-depth discussion, you should consult the mapping cookbook.

"Per beam" calibration

This method, which produces an FCF in Jy per beam, is most easily used to calibrate peak flux measurements. For unresolved calibrators and science targets, this is of course equivalent to the total or integrated flux. For planets, the program FLUXES calculates the expected flux in the beam. This can then be compared directly to the peak of the planetary map to calculate the FCF. We also provide peak values for the secondary calibrators.

To summarise, the peak FCF is given by:

FCF(peak) = S(expected) / V(measured),

where S(expected) and V(measured) are the expected and measured peak fluxes respectively of the calibrator source.

"Per unit area" calibration

The alternative is to use an aperture on your calibrator source and add up all the flux inside, dividing by the area of one square to get an FCF in Jy per square arcsecond. This is often useful for those sources which are obviously resolved by the JCMT's beam. It also appears to be a stabler measure of the telescope's sensitivity. It is important to use the same size aperture on your calibrator as you use on your science object.

Again, the program FLUXES can be used to calculate the FCF for planets. For Uranus and Neptune, which are always relatively small, the total flux at the appropriate wavelength should be used, assuming you are using an aperture larger than the beam. Mars, however can be larger and so you will need to ensure you use a large enough aperture. We also provide values for the secondary calibrators using a 40" aperture. For the three secondary calibrators which appear to be either unresolved or very marginally resolved by the JCMT at 850 μm (CRL 618, HL Tau and OH 231.8), this 40" aperture value can also be used for larger apertures to reasonable accuracy. However, for the other calibrators at 850 μm and for all the secondary calibrators at 450 μm this is not a valid assumption and these numbers should not be used for larger apertures. Planets are your only useful calibration source in this case.

Probably the simplest way to actually measure the flux within an aperture is to use the "Aperture photometry" function within the "Image analysis" menu in the Starlink GAIA tool. This allows you to centre a circular aperture on an object, and to increase or decrease its size until the suitable aperture is reached. It will then tell you the total flux within that aperture.

To summarise, the integrated FCF is given by:

FCF(integrated) = S(expected) / [V(measured) x A],

where S(expected) and V(measured) are the expected and measured total fluxes respectively of the calibrator in a given aperture and A is the area of a single pixel in square arcseconds.