2. Discussion of Basic Data 



The radio refractive index of the atmosphere, 

 n, exceeds unity by at most 450 parts per mil- 

 lion ; it is therefore customary to utilize the 

 radio ref ractivity, N, given by 2 



N = (n-1) x 10 6 . (1) 



The radio refractivity of air for frequencies up 

 to 30,000 Mc/s is given by Smith and Wein- 

 traub [1953] : 



N = 77.6^ + 3.73 X 10 s —' < 2) 



where P is the total atmospheric pressure in 

 millibars (mb), T is the absolute temperature 

 in degrees Kelvin (K) , and e is the partial pres- 

 sure of water vapor in millibars. [For the de- 

 velopment of this equation from theory, c.f. 

 Bean, 1962]. The P/T term in (2) is frequently 

 referred to as the "dry term" (even though 

 there is a small water vapor component in the 

 total pressure) and the e/T 2 term, as the "wet 

 term." 



The radiosonde is in general use throughout 

 the world to measure the pressure, temperature, 

 and relative humidity of the upper air. These 

 data can be used to obtain the corresponding 

 vertical profile of radio refractivity. However, 

 there are a number of disadvantages in the use 

 of radiosonde data for the purpose of obtaining 

 iV-profiles ; perhaps the most important of these 

 is the relatively slow response (large lag con- 

 stants) of the radiosonde temperature and hu- 

 midity sensors [Bunker, 1953 ; Wagner, 1960, 

 1961; Bean and Dutton, 1961]. Also of some 

 importance is the method of selecting levels for 

 which data are reported. The procedure fol- 

 lowed by most meteorological services consists 

 of reporting temperature, humidity, and height 

 at certain fixed pressure levels, called "manda- 

 tory levels" (e.g., 850 mb), plus a sufficient 

 number of additional "significant" levels to 

 provide a profile of temperature and relative 

 humidity such that linear segments between 

 levels will not depart from the original data at 

 any point by more than 1°C or 10 percent rela- 

 tive humidity. Such tolerances, although ac- 

 ceptable for most meteorological purposes, may 

 result in errors of as much as 30 N-units (under 



2 Throughout this atlas, the term atmosphere should be understood 

 to mean the nonionized atmosphere, i.e., excluding the ionosphere. 



3 A list of these stations is included in appendix A. 



extreme conditions) in a linear-segment N-pro- 

 file constructed from radiosonde data. Some 

 punched-card refractivity data are available 

 which were calculated from significant levels 

 chosen with an even wider tolerance, 2°C and 

 30 percent relative humidity [Bean and Cahoon, 

 1961a]. In spite of these deficiencies, this atlas 

 is based entirely upon radiosonde data, since no 

 other worldwide, long-term upper-air data are 

 available. (More detailed measurements, ob- 

 tained primarily from wiresonde and aircraft 

 refractometer flights, are available only for a 

 very few locations and for very limited periods 

 of time.) 



The first step toward obtaining a broad 

 sample of upper-air refractivity data was the 

 selection (by geographic-climatic considerations 

 and period of record available) of 112 radio- 

 sonde stations from the worldwide network. 3 

 Wherever possible, a total of 5 years of data was 

 obtained for each of the 4 representative "sea- 

 sonal" months of February, May, August, and 

 November. 



Five-year means were selected for use in the 

 preparation of this atlas for two reasons : (a) a 

 large number of stations, including all of those 

 in the U.S.S.R., have available records dating 

 back only to the International Geophysical Year 

 (IGY), 1958; (b) 5 years seemed to represent 

 the best compromise between the number of 

 stations and the length of record, since the total 

 amount of data which could be processed was 

 naturally limited. A large number of stations 

 is desirable for mapping purposes (better cov- 

 erage), while longer periods of record yield 

 more stable (accurate) estimates of long-term 

 means (of climatic variables). For each radio- 

 sonde ascent, the reported values of pressure, 

 temperature, and humidity at each mandatory 

 or significant level were converted by means of 

 (2) to radio refractivity values (by the Nation- 

 al Weather Records Center, Asheville, N. C). 

 These data, when received at ITSA, were used 

 to obtain four monthly mean N-profiles for the 

 available period of record for each of the 112 

 stations. The procedure followed was to obtain 

 for each profile the values of N at a number of 

 standard height levels by assuming separate 

 exponentials for the dry and wet terms between 

 each reported level. The mean A'-profile for the 



