Instrumental
To calibrate the instrumental polarizations, observe one
unresolved calibrator, whether polarized or not, at least three
times, more often if possible. Distribute these observations to cover
a range in parallactic angle of
: this
will separate any polarization of the calibrator from the required
instrumental terms (see Lecture 7). Programs involving long (
4 hr) syntheses of single sources will normally be able to derive the
instrumental polarization calibration from the observations of the
external gain calibrator. When determining the integration
time for the instrumental polarization
calibration, bear in mind that the cross-polarization (leakage) terms
whose relative amplitudes and phases are to be determined will likely
produce false polarized intensities that are only a few percent of the
flux density of the calibrator. The instrumental polarization
calibration must be done at each frequency for which polarimetry is
required. The most efficient way to do this is to cycle through the
frequencies used for the target observations each time you point the
array at the calibrator. If you omit an instrumental polarization
calibration, you will be unable to determine small degrees of
polarization reliably, or to deconvolve polarized extended structures
properly. (Antenna-to-antenna polarization differences distort the
polarization images in ways that do not satisfy the convolution
theorem.)
Position Angle
To calibrate the polarization position angle scale, observe 3C286 or 3C138 at least once during your observing run at each relevant frequency. You will determine the apparent position angles of the linear polarization of these sources after you have finished observing and after calibrating the total intensity data. The difference between the apparent and the nominal values of these position angles is corrected by adjusting the phase difference between the left and right circular polarizations, using a procedure that is described in detail in the AIPS Cookbook. It is advisable to alert the array operator to the presence of this calibration in your program, so that the observations of 3C286 or 3C138 can be extended or rescheduled if necessary to avoid losing them to equipment failure. Note that this calibration is essential if you wish to use your polarization position angle data, e.g. to determine magnetic field directions or to measure the Faraday rotation of a source.
Ionospheric Faraday Rotation
At wavelengths of 18 cm and longer, the position angle calibration
may appear to be time variable because of fluctuations in the
ionospheric Faraday rotation (Lecture 7). If you will use the
polarization position angle information at long wavelengths, it is
worth monitoring one polarized calibrator in the same part of the
sky as your target(s) throughout your observing, to see if its
apparent position angle changes significantly. If this calibration
shows that the ionospheric changes are less than about ,
it may be satisfactory to interpolate the observed position angle
changes as a function of time when you set the position angle scale by
adjusting the relative phase of the left and right circularly
polarized channels. If you see larger changes, especially in a
calibrator some tens of degrees from your target, it may be impossible
to correct for them accurately without using an ionospheric model.
Unfortunately, a source of critical-frequency data that was used to
estimate ionospheric models for many years at the VLA (via the FARAD
programs) is no longer available. Until it is replaced (possibly by
data from GPS systems), there is no way to repair VLA polarimetric
data sets in which the ionospheric Faraday rotations are large.
Observing a polarized calibrator can therefore be a ``warning light''
for such ionospheric Faraday rotation problems, but it does not
guarantee a way to correct them. Fortunately, ionospheric effects are
normally negligible above 4 GHz.