REFRACTIVE INDEX PROFILE SAMPLE 341 



and 



ARe = a. (Re, ht) + 62 [Re, ht) Ns ± S.E.2 (Re, ht), (8.24) 



where e is the elevation angle error, AR e is the radio range error, ^0 is the 

 apparent elevation angle, hi is the target height, Re is the apparent radio 

 range, and S.E. is the standard error of estimate about the regression 

 line of € or ARe on N s- Values of the coefficients may be obtained by 

 performing linear regressions of e or ARe, as ray traced for an appropriate 

 sample of radio refractive index profiles, ujjon A^^, for a large number of 

 target positions. As a byproduct of these calculations, one also obtains, 

 for each target position, a value of the residual error (the standard error of 

 estimate) to be expected for the particular type of profile sample used. 



In order to obtain a general set of eciuations to be useful under arbi- 

 trary conditions of location, climate, and weather, a large sample of 

 N profiles has been assembled which is believed to be representative of 

 both mean climatic and geographic trends and the larger synoptic varia- 

 tions which may be encountered. This was done by choosing 13 radio- 

 sonde stations representative of the major geographic and climatic types 

 of the world, and then choosing from each station six A'' profiles of particu- 

 lar types, two of which are typical of the extremes of monthly mean 

 conditions for that location, and the other four of which are typical of 

 some of the variations which are found at that location [18]. The result 

 is a sample of 77 N profiles,^ which has been found over a period of years 

 to be a sound cross section of general refractive conditions and has thus 

 been named the CRPL Standard Atmospheric Radio Refractive Index 

 Profile Sample, hereafter referred to as the CRPL Standard Sample. 

 Although the locations chosen for this sample are heavily weighted towards 

 the United States, it has been found that the general behavior of the 

 refractive index structure as inferred from the standard sample is typical 

 of conditions experienced in most parts of the world [5]. 



The remainder of this section will be devoted to some comimrisons of 

 observed radio refraction data with the predictions supplied by the 

 CRPL Standard Sample, as derived from the linear regressions men- 

 tioned above. 



Since the refraction measurements re[3orted here consist of samples 

 taken at particular locations over comparatively short i)eriods of time, 

 they should provide a test for the general set of coefficients derived from 

 the Standard Sample; not only is the general theoretical api)roach tested 

 against measured values, but the measurements coming from places of 

 more or less homogeneous nature, they provide a check as to whether 01 

 not coefficients derived for a large heterogeneous sample of data are 

 applicable also to individual places and times; i.e., they should reveal 



One of the types could not be found for one of the stations used. 



