Refractive index of air: 3. The roles of CO2, H2O, and refractivity virials.

The author's recent studies of the refractive index of air are extended, and several assumptions made therein are further examined. It is shown that the alternative dispersion equations for CO2, which are due to Edlen [Metrologia 2, 71 (1966)] and Old et al. [J. Opt. Soc. Am. 61, 89 (1971)] result in differences of less than 2 x 10(-9) in the phase refractive index and less than 3 x 10(-9) in the group refractive index for current and predicted concentrations of CO2. However, because the dispersion equation given by Old et al. is consistent with experimental data in the near infrared, it is preferable to the equation used by Edlen, which is valid only in the ultraviolet and the visible. The classical measurement by Barrell and Sears [Philos. Trans. R. Soc. London Ser. A 238, 1 (1939)] on the refractivity of moist air is shown to have some procedural errors in addition to the one discussed by Birch and Downs [Metrologia 30, 155 (1993)]. It is shown that for normal atmospheric conditions the higher refractivity virial coefficients related to the Lorentz-Lorenz relation are adequately incorporated into the empirically determined first refractivity virial. As a guide to users the practical limits to the calculation of the refractive index of the atmosphere that result from the uncertainties in the measurement of the various atmospheric parameters are summarized.