COMPARISON OF SOME PARAMETERS 243 



6.1.8. Comparison of Some Parameters 



It is probably fair to say that, at present, the greatest interest is con- 

 centrated on the relative merits of N s, AN, and Qe, for these quantities 

 have received more detailed attention than any others. (It might be 

 argued that, since the initial gradient of refractive index with height has a 

 strong influence on refraction, this quantity should be highly correlated 

 with signal level. However, the errors of measuring the initial gradient 

 by radiosonde techniques effectively mask the correlation with the signal.) 

 In a comparison of N s, AN, and Qe, the monthly median values of trans- 

 mission loss, L, were determined for the hours of radiosonde ascents for 

 20 radio paths in various parts of the United States. 



The equivalent gradient, Qe, was then calculated using standard ray- 

 tracing techniques [15] assuming: 



(a) the actual antenna heights, 



(b) a smooth earth with the sea-level radius, 



(c) horizontal stratification of A'^ with a vertical distribution the same 

 as that of the monthly mean. 



Standard statistical methods were then employed to determine the cor- 

 relation between monthly median values of (a) E and Qe, (b) E and N s, 

 and (c) E and AA'^ (surface to 1 km height). The average values of these 

 correlation coefficients for the 20 paths are given in table 6.2. The paths 

 studied in this comparison were the same as those listed by Bean and 

 Cahoon [2] in their analysis of N ^ —AN correlation. The data used were 

 representative of climatic conditions ranging from those of New England 

 and the Great Lakes area to Texas and the Pacific coast. This investi- 

 gation, at least, would seem to justify existing prediction procedures 

 based on A^^ and AA^. It is valuable, however, to consider some of the 

 results for specific paths in more detail, since they illustrate some features 

 of interest directly relevant to the arguments concerning the value of 

 N s and AA'^. 



Table 6.2. Average correlations, r(z), of refraction variables with field strength 

 (20 paths, 130 to 446 km long,/=92 to 106 Mc/s) 



6.1.9. Some Exceptions and Anomalies 



In so complex a matter as field-strength prediction, it would be un- 

 reasonable to expect any radio-meteorological parameter to be of world- 

 wide application to a uniform degree of reliability, and some conditions 

 have already been quoted in which N s may not be a reliable parameter. 

 The significance of these examples is really the issue on which views 



