236 TRANSHORIZON PARAMETERS 



from randomly dispersed eddies; on occasions there will also be a semi- 

 coherent field arising from partial reflection at elevated stable layers. 

 Some aspects of this complex situation are discussed in section 6.2. 



To sum up other important conclusions reached by the authors men- 

 tioned above, the correlation between E and N s 



(1) increases with increasing variation of E or N s, 



(2) is greater for seasonal cycles of night-time values of the variables 

 than for the midday values, and 



(3) is greater for summer diurnal cycles than for winter ones. 

 Conclusions (1) and (2) are particularly important if we try to assess the 

 utility of A^s in prediction work in terms of signal data from areas in which 

 the variation of E is small (e.g., Western Europe), or in terms of data for 

 afternoon periods only. 



Another parameter, closely related to AN, is the "equivalent gradient," 

 Qe, proposed by Misme [19]. This is defined as that linear decrease of n 

 with height which produces the same amount of bending as the actual 

 inhomogeneous atmosphere over a given transmission path. The prob- 

 lem is illustrated in figure 6.7, where the dotted line represents the actual 

 ray path between T and P above an earth of radius a. It is required to 

 determine the curvature, p, of the circular, full-line path which is tangen- 

 tial to the real path at P and which corresponds to a constant value of 

 dn/dh in a fictitious atmosphere. Boithias and Misme [20] have described 

 a graphical method for calculating Qe ( = l/p), for example, when P is 

 located in the center of the common volume of the antenna beams. 

 Monthly median values of ^^ are obtained from the corresponding monthly 

 median values of dn/dh, and Misme [19] has given tables of Qe for various 

 path lengths for different months. 



It is evident that Qe, like AA^", expresses the amount of refraction pro- 

 duced by the atmosphere and one might expect a high correlation between 

 the two parameters. This is known to be the case in some results quoted 

 by Misme [19] and which are illustrated in figure 6.8. Here the monthly 

 median values of AN and Qe (for a 300 km path) are compared, together 

 with the variation in N s, for Leopoldville in the Congo area. The corre- 

 lation between Qe and AA^ is high, but the maximum values in Ns in 

 December and January are accompanied by a local minimum in the 

 values of Qe and AA^. These results, and similar ones from Dakar (W. 

 Africa), have been quoted in support of arguments that N s is of limited 

 value in predicting seasonal and geographical variations in field strength 

 [19]. It is essential, Misme claims, to investigate the nature of the 

 N ,-AN correlation in separate climatic areas; the correlation seems poor 

 for some equatorial climates, probably because of the presence of semi- 

 permanent elevated layers. In these conditions, Alisme feels that an 

 exponential reference atmosphere and a correlation between N s and AA^ 

 are not to be expected. 



