Cross-correlation tools have been used to determine whether systematic trends in wavelength zero-points exist with respect to camera aperture, time, focus, and THDA for high-dispersion IUE/NEWSIPS data for all three cameras. Not surprisingly, a major residual error source is in an implied inconsistency in target-centering/guiding among different observations, some of which could arise from different thermal histories of the telescope. Even so, the wavelength calibration of the SWP high-dispersion data is surprisingly good. In fact, it can rival the wavelength precision of the Large Aperture GHRS spectra for stars observed many times over several epochs and in consistent instrumental conditions. Even over 10-15 years, we found evidence of significant velocity trends for only one or two stars ( 3.3.). At most these trends could amount to ± 3-5 km s-1 over this duration, and they may actually be smaller. We have also found that dedicated time-series observations of bright stars can substantially reduce apparent velocity errors. Yet, even for such data strings, changes in the spacecraft environment can produce spurious variations of ± 3 km s-1 over a timescale of a few hours. It is not clear that series of long-wavelength spectra exist in IUE database to detect such effects, but we expect that they are present nonetheless.
Using the GHRS atlas of 10 Lac as an secondary standard, we have found that the apparent radial velocity determined from high-dispersion images of the IUE SWP camera is zero, within the errors of measurement, with respect to the ultraviolet wavelength system defined by contemporary laboratory catalogs (Brandt et al. 1997). This result has been checked by lower-accuracy comparisons such as the difference between apparent radial velocities from SWP echellograms of two other presumed constant early-type stars with similar spectral type, HD 93521 and HD 60753, and also by comparing wavelengths of several interstellar lines with those measured by the GHRS. Additionally, and for all three IUE cameras, we have confirmed the accuracy of the zero-points with more limited sets of lines with the GHRS and STIS atlases of Procyon and Arcturus, respectively. We have compared the zero-points of the two IUE long-wavelength cameras by means of the Arcturus and Procyon atlases and also by compared wavelengths within a common narrow wavelength order among the three cameras. Neither of these comparisons led to a siginificant radial velocity offset between the IUE and HST instruments. The long-wavelength camera results are also in agreement with the ground-based velocity of Procyon, but less precisely so. Our estimate of the accuracy of in the zero-point error for the LWP and LWR cameras is itself imprecise, but judging from its consistency with the Arcturus and Procyon atlas wavelengths, it is probably no worse than ± 5 km s-1. Our findings are in apparent disagreement with those of González-Riestra et al. (2000). Their results led to shifts in the wavelength scales of the SWP high-dispersion and LWP/LWR low-dispersion data for the INES (IUE Newly Extracted Spectra) data product produced by the European Space Agency.
We have found no noticeable wavelength-dependent differences in zero-point between IUE SWP camera data for 10 Lac and the GHRS atlas of this star. However, note that this finding does not address possible errors within a particular order or measurements of a common feature appearing in adjacent orders, such as the Mg II h line (see Gonzalez-Riestra et al. 1998). Although we have not examined the data for this purpose, we have noticed occasional mismatches of 5-10 km s-1 from one end of an echelle order to another compared to theoretical spectral line syntheses. Such errors probably originate from an inadequacy of the calibration algorithms to determine dispersion nonlinearities accurately because only a few lines were available in individual orders of WAVECAL echellograms.
Finally, because the GHRS atlas of 10 Lac has a broad wavelength coverage and shows good correspondence in its wavelengths compared to laboratory data, we recommend the use of this atlas as a reference standard for comparison of astronomical mid-UV wavelengths with data from other missions.
The author gained valuable technical information presented herein while serving on the staff of the IUE Project at the Goddard Space Flight Center (NASA). It is our pleasure to acknowledge informative technical discussions with Dr. Catherine Imhoff, Mr. Randall Thompson, and Mrs. concerning questions of IUE operations practices and detailed contents of the IUE database. We also thank Dr. Rosario González-Riestra for providing us with a list of spectra used in her team's investigation of NEWSIPS data properties. We acknowledge helpful conversations with Drs. Richard Robinson, Charles Proffitt, and Mr. Don Lindler on the calibration of GHRS wavelengths, and Dr. Gill Nave for comments on the utility of various laboratory wavelength sources. We are also indebted to to Drs. Martin Snow and Brian Wood for providing their atlas data to MAST, and to Dr. Tom Ayres for providing it before publication. We are grateful to Drs. Rosario Gonzalez-Riestra and Jay Holberg for providing unpublished data from their work. We wish to thank Dr. Marc Postman for his patience.