Planetary Lunch Talk

How to Improve the S/N of Your Existing Data by 30%

Monday, 12:20 - 1:20, 22 September 2003, 622 Space Sciences Building, bring your lunch.

Optimal Spectral Extraction Figure
PNG version. Postscript version.

In a departure from the usual Planetary Lunch format, this talk discusses several data analysis techniques that all observers should know but few actually use. These include optimal estimation, which, if a PSF can be estimated, improves the S/N of point-source spectroscopy and photometry by up to 30% over "sky box" and aperture methods. Also covered (time permitting) will be a new method for non-linear interpolation of bad pixels. My students and I have written several general IDL packages that implement these methods, so there's no excuse, and big rewards! No special preparation will be needed to understand the talk, but those wishing to come away with the fullest understanding of optimal estimation can read Horne (1986, PASP 98: 609--617) and/or Chapters 3 and 4 of Bevington. Copies of Horne are available free online via ADS and will also be available in hardcopy at the talk.

Last revised: 2003 September 16 - J. Harrington