A New Method for Computing Differential Color Refraction

Abstract

Listen

Translate article

ABSTRACTA new method for computing differential color refraction (DCR) is presented in this paper that uses observations taken through a conventional filter as well as through a narrowband interference filter that serves as a template for the reductions. The method is tested with observations taken through Johnson BV and Cousins RI filters, and, in general, there is good agreement (excellent in all but one case) between observed and predicted values of DCR. A full description of the method is given as well as a discussion of the refraction model used to make the predictions. As expected, DCR is strongest in the B passband and weakest in the I band because of the wavelength dependence of atmospheric refraction. Furthermore, the common practice of calibrating DCR against a stellar color is investigated, finding that the procedure is justified, albeit the intrinsic scatter in the calibration can vary widely with the choice for the color. This scatter can be minimized with a judicious choice for the color system, and examples are given for the BVRI passbands considered in the paper. Nonetheless, there will always be some remaining error caused by the calibrating process, which will affect mostly star positions determined at large zenith distances. Further gains in positional accuracy are possible by choosing a passband for which the DCR is very small. Two such passbands are the Cousins I band (FWHM = 1900 Å) and the passband (FWHM = 630 Å) chosen for the USNO CCD Astrograph Catalog project. The ranges in the DCR for these passbands are, respectively, only 19 and 12 mas at a zenith distance of ZD = 60°. When calibrated against stellar color, remaining errors in the DCR calibrations for both of these passbands are expected to be under 5 mas. In the final section of the paper, numerical results are given for other possible passbands with small ranges (≤20 mas) in DCR.