Below we give more details on changes that are likely to be noticed by users in their data.

The SDC table (SDCTAB) has columns A2CENTER and CRPIX2, which give the location of the spectrum in the cross-dispersion direction. A2CENTER is the location in the input image, and CRPIX2 is the corresponding location in the output image. The rows of the SDC table were not selected by APERTURE. When data are taken using one of the CCD pseudo-apertures (i.e. apertures with names ending in "E1"), the location of the target will be quite different from the values given in the SDC table, resulting in an incorrect value for CRPIX2 in the 2-D rectified image. This is now rectified. (OPR 42410)

One dimensional interpolation of interorder light did not provide an adequate description of the scattered light beneath each echelle order. The error in saturated line cores can be 5% of the local continuum in the FUV. Don Lindler (Advanced Computer Concepts) and Chuck Bowers (STIS IDT at GSFC) have developed a two-dimensional iterative deconvolution algorithm that reduces the residual error by a factor of 10. This algorithm is now implemented in calstis6. (OPR 42441)

OPRs related to 42441 (which are also closed) are: 42442 KWDB Add 8 new STIS keywords needed by SC2D algorithm 42444 OPUS Populate new STIS keywords for SC2D algorithm 42445 CDBS Make new reference table types for SC2D algorithm

Recently delivered flat fields for the CCD had dust motes flagged with 1024 (small blemish), and these regions cover a significant fraction (about 6.5%) of the image. Of order 75000 pixels are flagged with 16 (hot pixel) in the dark files. These data quality flags in the reference files get propagated to the science data during calibration. They are also propagated during calibration of the wavecal files. Most of these pixels, however, are only marginally affected, because the dust motes and hot pixels are reasonably well corrected by applying the reference files. It was felt that it is appropriate to maintain these data quality flags in the science data to warn the user that there might be problems in certain areas of the image. It is our experience, however, that pixels flagged as small blemish or hot pixel are so well corrected that they can safely be used for wavecal processing, which is now implemented. (OPR 42727)

When basic2d (calstis1) is run, the PHOTMODE keyword is updated in the primary header, based on values of other keywords such as APERTURE, OPT_ELEM, and DETECTOR. It is intended that PHOTMODE could be used as a value for the obsmode parameter in synphot tasks. The STIS header keyword values are not always the same as the keywords in the synphot graph table, however. For example, OPT_ELEM for imaging data can be "MIRNUV", while the graph table only contains an entry for "MIRROR". This results in a warning from synphot that keyword mirnuv is not used. The fix in calstis 2.9 is such that the PHOTMODE output can now be correctly used in synphot. (OPR 42728)

CALSTIS uses cubic interpolation within 1-D arrays such as the PHOTTAB and APERTAB. If there is a discontinuity in one of these reference tables, the interpolated values can be wildly incorrect, which can affect the flux-calibrated values and the photometry keywords. This is fixed with linear interpolation. (OPR 42765)

Calstis no longer makes use of the MOFFSET1 and MOFFSET2 keywords. The wavecal processing includes these offsets when finding the MSM offsets. The MAMA offset coefficients table (MOFFTAB, _moc.fits) was still being opened and read, but the coefficients were effectively not used because the offsets are set to zero. This means that a user running calstis had to have a copy of a MOFFTAB even though it really isn't needed, and the code was misleading because it looks as if this table is actually used. This problem is fixed in calstis 2.9. (OPR 42774)