Some Neglected Pixel Problems

We have developed several improved techniques because standard reduction software was inadequate for the Treasury Project on η Carinae. Though adapted specifically to HST/STIS/CCD data, some of these methods can be useful for images obtained with other instruments. For instance, our subpixel modeling technique reduces some of the evil effects of marginal sampling, i.e., in images where the pixels are moderately too large to take full advantage of the instrument’s basic point spread function (PSF). Marginal sampling causes the apparent PSF to depend on precise location in the pixel grid, a subtle effect that perturbs measurements. Our method considerably reduces the amplitude of this phenomenon. 1. Preliminary Comment: Not Just Spectroscopy The techniques sketched here were motivated by intensive STIS spectroscopy of one object. But their usefulness is far more general; they can be applied to a broad class of images obtained with various instruments. Whenever the PSF’s FWHM is between about 1 and 3 pixel widths, one should consider “subpixel modeling” as outlined in Section 4. 2. The Role of η Carinae Concerning HST Data in General Before getting into data processing techniques, let me frankly indulge in some advertising which serves a valid purpose here. Most astronomers have a nodding acquaintance with η Car: a very massive star with intense, ostentatiously bipolar ejecta. The degree of structural complexity, however, and the surprisingly basic unsolved problems lurking there, are less well known. Together these characteristics have made Eta unique in HST’s record: • The ejecta are spatially complex on size scales of 0.01 to 10 arcsec; • They and the star are spectrally complex too, with thousands of identified emission lines; • At least six distinct types of spectra are found there, relevant to diverse problems; • And some of those spectra vary systematically in time. • Thus we have needed excellent spatial and spectral resolution, appreciable spatial coverage, broad wavelength coverage, and repeated observations, all together. • Both the star and its ejecta are bright enough to make such observations feasible without requiring huge amounts of telescope time. Scarcely any other major HST target shares this combination of attributes. As a result, η Car has consistently provided unusual illustrations of HST’s capabilities. (Why this matters to us here, will become evident in the next paragraph.) Since 1991 this object has held the record for angular resolution in spectroscopy of a complex target, first with For general information about this object, see Davidson & Humphreys (1997) and three proceedings volumes: Humphreys & Stanek (2005), Gull et al. (2001), and Morse et al. (1999). The HST Treasury Program for Eta has produced numerous scientific results, but here we’re concerned with data techniques.