B-Stars

The atmospheric features that mess up the red end of the spectrum may be calibrated out by multiplying by a calibration spectrum obtained from an observation of an object that is featureless in this region. A B star is usually used, hence the term `B-star calibration'.

The essential Figaro command for this calibration is `bsmult'. This multiplies the spectrum to be corrected by a B-star calibration spectrum-which is essentially obtained by fitting the continuum of a B star and then dividing that fitted continuum by the B-star observation. `bsmult' allows for the difference in air masses of the two observations, which is what makes it necessary, rather than, say, the simpler `imult'.

`bsmult' is a fairly straightforward operation. The problems come in generating the calibration spectrum. At present, the simplest way is to display the B-star spectrum using `splot' and then generate the continuum by hand using `cfit'. An alternative is to generate a mask for the lines in the spectrum using `mask' and then fit the masked continuum using `mcfit'. The main problem with the latter approach is that the chances are that very little of the spectrum is actually uncontaminated either by stellar lines or by the atmospheric bands.

(The best solution is use an automatic program that fits splines between points on the spectrum that are known to be uncontaminated, but such a program is not available yet-nor is a really good list of such points that will apply for all wavelength ranges and dispersions.)

The calibration spectrum should really be exactly 1.0 at all points not affected by the atmospheric bands, and it is probably worth displaying the calibration spectrum using `splot' and then using `cset' to set such regions to 1.0 interactively. (This is something else that could be made automatic eventually. Fortunately, it isn't necessary to do this very often.)

Note that `bsmult' requires that both the calibration and the object being corrected have valid air masses. Air masses are stored in NDF data files as `file.MORE.FIGARO.SECZ', and in DST data files as `file.OBS.SECZ'. If necessary they may be set by hand using, e.g.

   ICL> creobj type=_REAL dims=0 object=file.MORE.FIGARO.SECZ
   ICL> setobj object=file.MORE.FIGARO.SECZ value=1.4

Also note that the correction applied by `bsmult' is multiplicative-this means that it is not suitable for data that is in logarithmic flux units, such as magnitudes.