404 CHARTS, GRAPHS, TABLES, AND COMPUTATIONS 



9.4. Glimatological Data of the Refractive Index 

 for the United States 



Charts of 8-year mean A''o values for 0200 and 1400 local time are given 

 for the four seasons of the year on figures 9.1 to 9.8. 



Selected stations illustrate the range of seasonal and diurnal variations 

 of N s in differing climatic regions across the country. These are plotted 

 on figures 9.9 to 9.14. Note the large annual and diurnal ranges of means 

 for humid subtropical Washington, D.C. (fig. 9.9), compared with the 

 modest variations exhibited by maritime-dominated Tatoosh Island 

 (fig. 9.14). The results of these analyses are consistent with those ob- 

 tained from climatic studies of the refractive index structure over North 

 America, described in section 4 of chapter 4. That is, large seasonal and 

 diurnal ranges are exhibited by continental-climate stations (Colorado 

 Springs) and small ranges by maritime-climate stations (Tatoosh Island). 



Cumulative distribution curves for A^^, useful in radio ray bending pre- 

 dictions, are given in figures 9.15 to 9.20 for the example stations above. 



Figures 9.21 and 9.22, standard deviation of N s, provide a measure of 

 the accuracy of the A^o charts of figures 9.1 to 9.8. As has been noted 

 in chapter 4, a number of climatic features are apparent on the standard 

 deviation maps. The climatic stability of various regions of the country, 

 for example, is reflected in these charts. Small standard deviations 

 characterize the maritime climate of the west coast. By comparison, the 

 strong air mass changes of wintertime synoptic patterns sweeping across 

 the southeastern United States are indicated by large standard devia- 

 tions in that region (fig. 9.21). 



9.5. Statistical Prediction of Elevation Angle Error 



The elevation angle error, e, as defined in chapter 3, can, like bending, 

 be predicted from surface refractivity, N s, by an equation of the form 



e = mNs + i. 



In tables 9.36 to 9.44 for a given height h, the coefficients m and t have 

 been determined by a statistical regression performed on the Standard 

 CRPL Sample, each having a unique N s- 



This process is analogous to the statistical bending prediction method 

 described in chapter 3. 



