This brief message is to inform you of a problem that has just been identified [July 94]. The problem concerns only GHRS observations of moving targets; fixed target observations are not affected. The problem concerns how the onboard doppler compensation is applied to observations of moving targets.
As you probably know, the GHRS automatically corrects observations for the effects of the motion of the HST spacecraft. This correction is done in real time as an exposure is in progress. Every minute a new shift is calculated and if necessary a change in the deflection of the spectrum is made. These onboard corrections cannot be removed after the fact except as an overall shift of the spectrum, but the net doppler compensation applied is stored in the data header. As in the case of a heliocentric correction to radial velocity, the doppler compensation is calculated from the position of the object relative to the vector of motion of the spacecraft.
The problem identified is that the calculation does not use the true coordinates of the moving target but instead performs the calculation as though the object were at zero hours right ascension and zero degrees declination. The result is that individual spectra may be not merely shifted in wavelength erroneously, but may also be smeared because the compensation could be applied in the wrong sense. A series of exposures that were added together could have wrong relative shifts, leading to a degradation of the result.
Observations made for the comet Shoemaker-Levy 9 campaign (mid-July) and thereafter should not have a problem, but earlier observations might. We believe we have identified those observations for which this problem may be pertinent and we have notified the appropriate PIs. We have now tested an algorithm that computes the Doppler compensation that should have been applied as well as that which was, so that observers and archival researchers can correct their data. If you have questions on this matter, please contact the Help Desk.
This message is intended only to notify you that this problem exists so you can be aware that wavelength scales (and possibly line breadths) may be corrupted for moving targets. We have computed the magnitude of the effect for the modest number of exposures that are affected, as well as the compensation that should have been applied, and we find that the original spectrum can be properly reconstituted.