MICROWAVE REPEATER RESEARCH 191 



few occasions when the fading could not be accounted for in this simple 

 fashion, it was assumed that signal components were arriving from above 

 by virtue of reflections from small, relatively abrupt changes in the dielectric 

 constant of the atmosphere. The existence of such reflections was demon- 

 strated by a frequency-sweep method during propagation studies on the 

 70 mile over-water path between Highlands, N. J. and East Moriches, Long 

 Island.^ 



With microwaves, where, as stated previously, ground reflections are 

 usually small or absent, one might surmise that changes in atmospheric 

 refraction would have a smaller effect on transmission than at ultra-short- 

 wavelengths where strong ground reflections are present, and that fading 

 should, therefore, be less. Actually the opposite is observed. Fading is 

 found to be more frequent, faster, and deeper as the wavelength is decreased. 

 This frequency effect may be explained in a qualitative fashion by a con- 

 sideration of the relative sizes of Fresnel zones at ultra-short waves and at 

 microwaves. It is know^n that the dielectric constant of the atmosphere 

 usually does not vary with height in a smooth linear manner; on calm nights, 

 particularly, very steep gradients in the dielectric constant may exist over 

 small vertical ranges measuring only tens of feet. The effectiveness of these 

 steep gradients would be expected to depend on their extent relative to the 

 size of a Fresnel zone. Thus on a path such as that in Fig. 1-4, a steep grad- 

 ient extending over only a hundred feet would include practically the whole 

 first Fresnel zone at 3 centimeters while it would cover only a small part of 

 a zone at 3 meters wavelength; the effective gradient, therefore, would be 

 considerably less at 3 meters than at 3 centimeters. Analyses based on 

 wave theory show^ that atmospheric layers, in which the dielectric constant 

 has a steep negative gradient, tend to confine or guide the radiation in much 

 the same way as a waveguide, and that this ^'trapping" phenomenon, for a 

 given layer thickness, becomes more pronounced as the wavelength is de- 

 creased.' 



The mechanism of microwave propagation is certainly a complicated one, 

 and a considerable amount of experimental work in the fields of radio and 

 meteorology will be required to unravel it. However, it is very difficult to 

 interpret the radio measurements in terms of meteorological data. The 

 chief difficulty is that meteorological measurements often do not give an 

 accurate picture of the atmosphere, particularly at those times when micro- 

 wave fading indicates that rapid changes of some sort are occurring in the 



^ Englund, Crawford and Mumford, "Ultra-Short-Wave Transmission and Atmos- 

 pheric Irregularities", B. S. T. /., vol. 17, pp. 489-519; October 1938. 



^ H. G. Booker, in England, was the first to call attention to this phenomenon. For 

 more recent work see: C. L. Pekeris, "Wave Theoretical Interpretation of Propagation in 

 Low-Level Ocean Ducts, Proc. I. R. £., vol. 35, pp. 453-462; May, 1947. This paper 

 gives references to other work in this field. 



