Discussion of “Atmospheric refraction at College, Alaska, during the winter, 1947–1948”

Refraction by the Earth's atmosphere of the light from celestial bodies is simple to explain qualitatively, but extremely difficult to treat quantitatively, as was shown by Willis [1941], The statement cited by the authors from Humphreys, that the total refraction depends only on the initial angle of incidence and the speeds in the first and last media, applies only to monochromatic light traversing plane layers; for concentric spherical shells the problem is far more complex, as Humphreys went on to show. Basically, any computation of the amount of refraction involves some assumption about the rate of decrease of air density with height, in which the all‐important factor is the vertical temperature distribution, or lapse rate. This is particularly important for rays emanating a few degrees above the horizon; at higher altitudes, formulas based on almost any atmospheric model can be adjusted to agree with observed refractions. Various theories have used various lapse rates; almost all United States tables, including the 1885 Doolittle tables which the authors used as standard, are based on those of Bessel, whose development assumed a lapse rate of around 0.6° F per 1000 ft, decreasing very slightly with height.