FLUXSTAR computes a flux calibration curve from a standard star spectrum and a file containing the correct flux levels as a function of wavelength. The input spectrum must be wavelength calibrated, be corrected for atmospheric extinction, and have its intensities on an uncalibrated F-lambda scale. The routine locates the flux points given in the input file, and finds the star's average flux over the specified wavelength bin of the flux point. A set of correction points are thus defined, which consist of the correct fluxes, reduced to the Hayes-Latham Vega calibration (Hayes & Latham, 1975 ApJ, 197, 593), divided by the observed intensity of the standard star. A spline is drawn through these points and replaces the standard star spectrum to give the flux curve. If you wish to use a polynomial rather than a spline, use the POLY=n keyword, where n specifies the degree of the polynomial to use.
The FLUX command takes the correction points defined above, and uses a spline (or polynomial) to define a flux calibration buffer for the input spectrum. The spectrum is then calibrated by multiplication by this buffer. The separate FLUXSTAR and FLUX commands permit the calibration of spectra on different wavelength scales than the standard star spectrum.
The standard star's flux measurements are read in from a file, which is assumed to be in the spectrum directory (see PRINT DIRECTORIES), with extension .FLX unless specified otherwise. The file is headed by the stars apparent V magnitude, and then its magnitude at 5556 A. Generally these are identical, and will only be different if a (grey) change in the Vega calibration is made. Each line of the file will contain a flux point specified by its wavelength, magnitude per unit frequency, and bin width in angstroms. The points must be in order of increasing wavelength, but there is no strict format that must be observed.
Below is an example of a standard star flux file:
The keywords 'AVE' and 'WT=' allow the averaging with weights of multiple flux curves. The default weighting, used for the first flux curve as well as for those using 'AVE', is 1. Anytime neither keyword is specified a fresh flux curve is started (again, with a weight of 1.). Note that the averaging can handle two flux curves which overlap (but do not necessarily match perfectly) in wavelength, and can even create a flux curve from two curves with completely disjoint wavelength scales, but it cannot insert flux points in the midst of an existing flux curve.
The SYSA keyword produces a point-by-point flux curve instead of a smooth, spline fitted flux curve. This is done by removing the known stellar absorption lines from the observed standard star spectrum. The atmospheric bands, however, are not removed. The result is that you end up with a flux curve which can correct for the atmospheric bands. A drawback for SYSA is that any "anomalous" absorption lines not known to the program will appear in the response curve and can lead to extraneous features in the data.
The SYSC keyword produces a compromise response curve with the best features of both "system A" and (the default) "system B". In this option the system A curve is smoothly fitted by a spline (at the usual knot points) while the B and A atmospheric bands (+/- 65 Angstroms) are retained. The curve is piecewise continuous. It is free from glitches introduced by spurious absorption features in the stellar spectra, retains a higher accuracy outside the extreme knot points, and compensates for the atmospheric bands in the data.