Summary:  The 'G' or 'T' solutions obtained by gaincal for calibrators for which the flux density was unknown and assumed to be 1 Jansky are correct in a time- and antenna- relative sense, but are mis-scaled by a factor equal to the inverse of the square root of the true flux density. This scaling can be corrected by enforcing the constraint that mean gain amplitudes determined from calibrators of unknown flux density should be the same as determined from those with known flux densities. The fluxscale task exists for this purpose.

Before running fluxscale, one must have first run setjy for the reference sources and run a gaincal that includes reference and transfer fields. The reference field(s) should be standard flux density calibrators for which an accurate flux density (or better, a model image, especially if it is a resolved source) is known. The transfer fields are all other calibrators, typically point sources (to a good approximation), and for which actual flux density is unknown. After running fluxscale, the output fluxtable will have been scaled such that the correct scaling will be applied to the transfer sources. If incremental=True, a simple incremental calibration table will be generated that contains a single antenna-based solution per field embodying the required scale factors. If incremental=False, a copy of the input caltable is generated, with the required scale factors applied to the transfer fields.

fluxscale applies the constraint that net system gain was, in fact, independent of field, time, and direction, on average, and that field-dependent gains in the input caltable are solely a result of the unknown flux densities for the calibrators. Using time-averaged gain amplitudes, the ratio between each ordinary calibrator and the flux density calibrator(s) is formed for each antenna and polarization (which they have in common). The average of this ratio over antennas and polarizations yields a correction factor that is applied to the ordinary calibrators' gains.

The square of the gain correction factor for each calibrator and spw is the presumed flux density of that calibrator (if the assumed flux density when solving was 1 Jy), and is reported in the logger. The errors reported with this value reflect the scatter in gain ratio over antennas and polarizations, divided by the square root of the number of  antennas and polarizations available. If the flux densities for multiple spws exist, fitted spectral index and (for nspw>2) curvature are also reported. The fit is done for

$log(S_\nu) = a_o + a_1*(log(\nu/\nu_0)) + a_2*(log(\nu/\nu_0))**2$.

The reference frequency, $\nu_0$ (the mean of $log(\nu)$) is reported in the logger along with the flux density at that frequency. The fit results are also reported in the returned Python dictionary which takes the form:

{fieldIdstr: {spwIdstr: {'fluxd':array([I,Q,U,V]),
                         'fluxdErr': corresponding errors,
                         'numSol': corresponding no. of solutions}
              'fieldName': field name,
              'fitFluxd': fitted flux density at the reference frequency,
              'fitFluxdErr': fitted flux density error,
              'fitRefFreq': reference frequency,
              'spidx': a_0, a_1, a_2
              'spidxerr': errors in a_0,a_1, a_2}
 'freq': (center) spw frequencies
 'spwID': list of spw IDs,
 'spwName': list of spw names}

where fieldIdstr and spwIdstr are field Id and spw Id in string type, respectively. The 'spidx' coefficients, $a_0$, $a_1$, and $a_2$ are the $log(S_{\nu=\nu0})$, the spectral index, and the curvature, respectively.

The calibrator models are currently not revised within the MS to reflect the flux densities derived by fluxscale. Use setjy to set these, if necessary.

The constant gain constraint is usually a reasonable assumption for the electronic systems on typical antennas. It is important that external time- and/or elevation-dependent effects are separately accounted for when solving for the gain solution supplied to fluxscale, e.g., gain curves, opacity, etc. (see gencal). 

The fluxscale results can also be degraded by poor pointing during the observation. The parameters, gainthreshold and antenna (and timerange/scan) can be used to control the data to be used in the flux derivation in such cases. The gainthreshold parameter sets the range of the input gain to be used in terms of the percentage deviation from their median values (per field, per spectral window).

If the reference and transfer fields were observed in different spectral windows, the refspwmap parameter may be used to achieve the scaling calculation across spectral window boundaries. In general, this will yield less accurate flux density calibration.

The fluxscale task can be executed on either 'G' or 'T' solutions, but it should only be used on one of these types if solutions exist for both and one was solved relative to the other (use fluxscale only on the first of the two).

ALERT: 'GSPLINE' solutions from gaincal are not supported in fluxscale.


Using resolved calibrators

If the flux density calibrator is resolved, the assumption that it is a point source will cause solutions on outlying antennas to be biased in amplitude. In turn, the fluxscale step will be biased. In general, it is best to use a model for the calibrator, but if such a model is not available, it is important to limit the solution on the flux density calibrator to only the subset of antennas that have baselines short enough that the point-source assumption is valid. Such a subset of antennas can be selected for the fluxscale calculation using the antenna parameter, which uses standard antenna-selection syntax. Specifying something in antenna also reveals timerange and scan selection parameters which enable more specific selection on these axes.

Alternatively, limiting the fluxscale calculation to antennas on unresolved baselines can be effected by using antenna and uvrange selection when solving for the flux density calibrator in gaincal. Please see the Examples section.