Quality Control and
Data Processing

The data are meant to provide, per Giraffe and UVES setup, a reference for achievable resolution, and a template for subtraction of sky/solar contamination.

The data were acquired on six nights between 2004-09-26 and 2004-12-02:

DATE	GIRAFFE settings	resolving power	UVES settings
2004-09-26	Medusa2 HR379.0 - HR572.8	19,000	580 nm
2004-09-28	Medusa1 HR599.3 - HR920.5	19,000	520nm, 580nm, 860nm
2004-10-11	Medusa1 LR385.7 - LR881.7	7,000	580nm
2004-11-26	Argus LR379.0 - LR881.7	11,000
2004-11-30	Argus HR379.0 - HR548.8	30,000
2004-12-02	Argus HR504.8 - HR920.5	30,000

Detailed information about the instrument is found here.

The spectra were acquired on the evening twilight sky ("twilight flats"). All fibres were illuminated with scattered solar light. The integration time is indicated in the plots.

Data reduction

Giraffe

The data were processed with the standard GIRAFFE pipeline, using calibration raw and product files taken at most a few days after the twilight spectra. The data processing included:

• bias subtraction,
• fibre localization and extraction (using average extraction) of the flat field,
• fibre extraction (using average extraction) of the sky exposures,
• division by the extracted flat (providing correction of fibre transmission, fibre-to-fibre transmission variation, fringing)
• wavelength calibration,
• rebinning.

All spectra are named by the tag "GI_SRBS" (which stands for "rebinned GIRAFFE spectrum"), with timestap and setup information added. Each GI_SRBS file has a GI_SRBE counterpart where "E" stands for error and contains the extraction error (standard deviation). Dividing both files by each other yields the per-fibre signal-to-noise ratio.

UVES

The spectra were reduced using the standard FLAMES-UVES pipeline and calibration data usually taken a few days before or after the observations. The reduction steps include: signal extraction per fibre, flat-fielding, fibre efficiency correction, wavelength calibration, and merging of orders. More information about FLAMES-UVES data reduction can be found here.

How to use

Giraffe

Per setup, plots are either organized as "1. Overview", or as "2. Details". Use the top horizontal navigation bar to switch between both.

The overview part has:

• the mean spectrum, averaged across all fibres with signal (Medusa: about 128; Argus: about 310), excluding the 5 SIMCAL fibres
• the signal from a selected fibre, about the middle one

The 'details' part shows the central fibre spectrum only, distributed across four panels in order to show the full spectrum.

In both plot sets, telluric absorption lines are marked redward of about 600 nm. These data are taken from a yet unpublished atlas of telluric absorption lines, derived from UVES telluric standard star data. The UVES data were taken at higher resolution (R = 50,000). The list has been selected to show only the lines deeper than 95% of the UVES continuum. This line list is used to mark potential telluric lines in the GIRAFFE data (by blue vertical bars).

The fringing redwards of about 750 nm has been completely removed by the flat fielding process. It remains visible in the S2N plot, because of the extraction error file precisely recording the error variations due to fringing. The fringing is most prominent in the single-fibre S2N plot, and is smeared out in the fibre-averaged plot.

All spectra are plotted on an approximate flux scale which still contains the ratio of solar spectral slope and lamp efficiency.

Some spectra show in their overview plot a curvature towards their red end (averaged plot only, no. 1). This is due to the slit curvature resulting in a shorter spectral range for the first and the last fibres. Hence the longest wavelengths in these spectra have only a few fibres contributing to the signal. This effect is of course not visible in the single-fibre plot (no. 2).

The horizontal navigation bar links to:

• the corresponding 'details' or 'overview' plot,
• the corresponding postscript file,
• the fits file with the data,
• the error fits file,
• and the tar file which has all spectra and error files.

UVES

The plots are organized per central wavelength and per CCD (lower/upper chip of UVES). For each combination of wavelength and CCD, an overview plot and five detailed plots are available. They can be reached using the vertical navigation bar. Telluric absorption lines are marked redward of about 600 nm. See above for more details.

The overview plots show the extracted spectrum of one fibre from one well exposed observation. The spectra have been rebinned to 0.1 Å resolution and are splitted over five panels for clarity.

The detailed plots are indicated by their first wavelength in the vertical navigation bar. They contain an average over all observations of the same setting from one night (26 September 2004 for 580 nm, 28 September for 520 and 860 nm) and across the fibres. Due to the inclination of the fibre traces on the chip, the available wavelength region is slightly different for each fibre. Therefore, only a selected number of fibres has been used for averaging in order to minimize artefacts at the edges.

Some gaps in the extracted spectra are present redward of 940 nm due to non-overlapping spectral orders and reduction artefacts.

The horizontal navigation bar links to:

• the corresponding postscript file,
• the fits file with the data,
• the error fits file (overview plots only),
• and the tar file which has all spectra and error files.

This atlas is the result of a joint effort between Paranal Science Operations (principal investigator: Jonathan Smoker and the Kueyen team; data acquisition) and QC Garching (Reinhard Hanuschik and Burkhard Wolff, data reduction and presentation).