Quality Control and
Data Processing

* central wavelength, dichroic filter, binning
** as measured from arc lamp exposures of that period

Reduction

The spectra fill the slit continuously, very much like a flat field lamp. Each spectrum has been reduced with the UVES pipeline. The data have been de-biased, flat-fielded, re-aligned row-by-row ('2D extraction', see below) so that the tilted echelle orders become aligned with the X axis, and then wavelength calibrated.

Since there is no real spatial information, this pipeline product has then been collapsed in Y direction. Also all exposures per setting have been co-added. The resulting 6 spectra (2 blue ones and 2x2 red ones) have then been merged into the final solar spectrum.

2D Extraction

Neither optimal nor average extraction give useful results here. The 2D extraction mode of the UVES pipeline has been used instead.

Both optimal and average extraction depend on an unresolved continuum which is used to find the signal and then extract it (either using a Gaussian fit or by averaging the signal in the object window). Optimum extraction simply fails since no signal is found (the solar spectrum is taken by the pipeline as background signal). Average extraction also fails since the signal and the sky are identical and the net spectrum is zero.

The 2D extraction uses the order table to define the extraction slit which is then used to extract the flux binwise (i.e. follow the inclined echelle orders while preserving the spatial information along the slit). The final product of the pipeline then is a 2D reduced spectrum where one dimension is wavelength and the other one the Y coordinate in cross-dispersion direction (along the slit).

Some remarks

The echelle orders are merged in a rather crude way by the pipeline (simply by averaging overlapping regions). In some cases sudden steps in the spectral slope appear due to this procedure. The nominal start wavelength of each order is marked in the plots by the vertical broken red line.

Generally the continuum slope in the data has been modified only by the flattening process which corrects for the blaze function. No additional rectification has been applied.

No attempt has been made to remove telluric absorption lines from the spectra.

Spectral gaps occur at 5710 - 5820 A and 8405 - 8650 A, due to incomplete overlap of the spectral range of the gratings. Above 9760 A the orders do not overlap anymore.

There are some problems with the extracted spectra due to imperfect flat fielding. These show in some cases outliers (caused by bad columns) which divide into the spectra and appear, in the non-averaged (2D) format, as inclined step function. In the collapsed version this artefact is correspondingly spread in wavelength space. In most cases the effect is barely noticable (faint artificial 'emission'), but in a few cases the spectrum is strongly affected (especally in the 6700/6800 A region). In any case, the ranges are marked in the plots by a horizontal red bar. Find the table of wavelength ranges affected here.

Plots

Spectrum

The complete UVES solar spectrum can be downloaded as FITS file here (4.0 MB FITS file; click on right mouse button, "Save Link as ..."). Its format is one-dimensional, with X coordinate in Angstroems. Uniform step size is 0.0124862 A.