286 ATTENUATION OF RADIO WAVES 



7.6. Derivation of Absorption Estimate for Other Areas 



The values of total path absorption given above are for two specific 

 locations. For this particular study to be of practical use, a means 

 should be provided for arriving at estimates of geographic and annual 

 variations of total path absorption for various surface distances and 

 frequencies. The method chosen utilizes the correlation between total 

 path absorption and the surface value of the absolute humidity, expressed 

 in grams of water vapor per cubic meter, which appears explicity in 

 Van Vleck's water vapor absorption formulas [4,5]. The basis for expect- 

 ing a correlation to exist between the absolute humidity and the total 

 path absorption is that the absorption at those frequencies for which 

 water vapor absorption is dominant (approximately 10 to 32 Gc/s) varies 

 directly as the absolute humidity while, for those frequencies at which 

 oxygen absorption is dominant, it varies inversely as the absolute humid- 

 ity due to the inverse relationship of oxygen density and water vapor 

 density. That is, during the warm seasons of the year the total atmos- 

 pheric pressure tends towards its yearly minimum (as does the oxygen 

 absorption), while the absolute humidity tends towards its yearly maxi- 

 mum (as does the water vapor absorption). Conversely, during the 

 colder seasons of the year, the pressure tends towards its maximum value 

 while the absolute humidity tends towards its minimum value [3]. 



As an example of the correlation method, the surface absorption was 

 calculated at a water vapor-dominated frequency (22.2 Gc/s) and 

 oxygen-dominated frequency (50 Gc/s) for each month throughout the 

 year for both Washington and Bismarck. These values are plotted on 

 figure 7.13. The term surface absorption is used for the values of ab- 

 sorption calculated from standard ground level weather observations. 

 The water-dominated 22.2 Gc/s data fall on a smooth curve despite the 

 pressure and temperature differences of the two stations. The oxygen- 

 dominated 50 Gc/s data, however, display an interesting separation of 

 points for each station, although the distribution of points at the two 

 locations display similar slopes. The 50 Gc/s absorption is more sensi- 

 tive to the atmospheric density difference between the stations. If the 

 pressure differences were taken into account, the Bismarck data would 

 increase about 12 percent and the two curves would be distributed along 

 a common line w^ith the same slope as the original two curves. This 

 figure, then, indicates that the absorption is correlated with the absolute 

 humidity. The above illustration is for surface values rather than for 

 integrated propagation path values. Variations in the upper air meteor- 

 ology that are not reflected in the surface values will tend to diminish 

 the correlation. (For comparison of percentage absorption over a 300-km 

 path, the first few hundred feet contribute about 5 percent at 100 Mc/s 

 increasing to 42 percent at 10 Gc/s and remaining constant to 50 Gc/s.) 

 Keeping these reservations in mind, one may utilize the method of least 



