190 BELL SYSTEM TECHNICAL JOURNAL 



Atmospheric Refraction 



As in the case of ground reflection, the refractive effect of the atmosphere 

 has been found to play a somewhat varying role, depending upon the wave- 

 length employed. In some of the early work on ultra-short-wave propaga- 

 tion,^' 2 the concept of average atmospheric refraction was found to bring 

 about better agreement between observed and calculated results. Due to 

 the variation of temperature and water vapor content of the atmosphere 

 with height above ground, the dielectric constant of the atmosphere nor- 

 mally decreases with height. The effect of this negative dielectric constant 

 gradient is to cause the path of a radio wave to be bent slightly downward 

 toward the earth, thus effectively increasing the horizon distance. It 

 has been suggested that a good approximation for average refraction was to 

 assume the radius of curvature of the ray to be four times that of the earth.^ 

 This condition is used at the present time to describe a ''standard atmos- 

 phere." 



It was soon found, however, that atmospheric refraction could vary be- 

 tween rather wide limits depending chiefly on the gradient of water vapor 

 with height.^ Refraction effects were found to be greater in summer than 

 in winter since the air contains a higher percentage of water vapor in the 

 summertime. A diurnal variation in refraction was also observed on over- 

 land transmission paths. During the day, rising convection currents and 

 surface winds, caused by surface heating of the earth, usually produce a 

 well mixed atmosphere near the earth so that "standard" atmospheric 

 conditions prevail. On clear nights, however, particularly if the wind veloc- 

 ity is low, radiation cooling of the earth may cause a temperature inversion 

 in the lower atmosphere; if, also, the water vapor decreases with height, the 

 combined temperature and water vapor effects may add to produce a steep 

 negative gradient in the dielectric constant. Stormy weather and over- 

 cast skies usually result in standard atmospheric conditions. 



Most of the signal variations observed during a two-year study of propaga- 

 tion of two and four-meter waves over the 39-mile over-land optical path 

 between Beer's Hill, N. J. and Lebanon, N. J.^ could be explained satis- 

 factorily on the basis of wave interference between direct and ground- 

 reflected radiations; the relative path lengths, and hence the phases, of these 

 two components of the received field varied with the refractivity of the 

 atmosphere. The fading on the two wavelengths was usually similar in 

 major detail as might be expected from the geometry of the path. On the 



»'»Loc. cit. 



* Englund, Crawford and Mumford, "Further Studies of Ultra-Short-Wave Transmis- 

 sion Phenomena", B. S. T. J., vol. 14, pp 369-387; July 1935. 



' Englund, Crawford and Mumford, "Ultra-Short-Wave Transmission over a 39 mile 

 'Optical' Path", Proc, I. K. E., vol. 28, pp. 360-369; August 1940. 



