4 Summary

  Within the framework of the development of the ESA INES Data Distribution System for the IUE Final Archive, IUE Low Dispersion spectra have been re-extracted from the bi-dimensional SILO files with a new extraction software. INES implements a number of major modifications with respect to the SWET extraction applied to the early version of the IUE Final Archive. The improvements in INES deal with the noise model, the optimal extraction method, the homogenization of the wavelength scale and the flagging of the absolute calibrated extracted spectra.

• The noise models for the different cameras have been re-derived to correct anomalies at high and low exposure levels in those used in SWET. The noise model results in a considerably more realistic estimate of the actual extraction errors in the IUE spectra.
• The algorithms to compute the camera background and the extraction profile are more consistent with the nature of the IUE detectors and result in a significantly improved data quality.
• Weak extended or miscentered spectra are more adequately handled. The fluxes of strong emission lines in weak continuum spectra are more reliable and consistent in the INES extraction.
• The handling and propagation of quality flags to the final extracted spectra has been improved. This implies a larger number of pixels flagged, but also a more

  correct information for the user of potential problems in the data.

• A major improvement has been reached in the removal of the solar contamination in LWP images after 1992.
• In order to facilitate an easier use of the data, all the spectra of a given range (short and long) have been resampled to a common wavelength scale.

As a general rule, INES data are similar or superior to SWET. Although the INES spectra may at times give somewhat lower signal-to-noise ratio than those obtained through SWET (e.g. when boxcar extraction is required to maintain data validity), the INES extraction results in a higher reliability of the IUEFA data, allowing direct intercomparison of all low resolution spectra, through an adequate treatment of errors, flags and warning messages in the image header.

Acknowledgements

We would like to acknowledge the support of all VILSPA staff, which collaborated actively to the development of the INES system.