3. Treatment of Data 



As the volume of data made the plotting of every flight impracticable 

 the following method of treatment was adopted. The temperature and 

 humidity were extracted at each standard isobaric surface (following 

 the general plan presented on the W.B.A.N. adiabatic charts), and the 

 average virtual temperatures at these levels were computed. The mean 

 values for each station were then replotted on an adiabatic chart, after 

 which the heights of the standard surfaces were computed and a pressure 

 height curve drawn. The temperature values at standard geopotential 

 were then read directly from the pressure height curve. In some cases 

 only material coded in the 1943 Radiosonde Code was available, and 

 in this case pressure temperature and humidity at the standard heights 

 were averaged and the same procedure used. 



This method of treatment of the data can be justified, both theoreti- 

 cally and practically, but a practical test is more likely to be convincing. 

 This is shown in Table 2, which gives the heights of the standard isobaric 

 surfaces at Norfolk Island as computed by the above method (A) and 

 also by the more usual method of averaging the heights of the pressure 

 surfaces as computed from each flight (B). The data are for February^ 

 1944-1946. 



Pressure 

 Surface, Mb. 



1,000 

 900 

 800 

 700 

 600 

 500 

 400 

 300 

 200 



Where possible data for the height of the tropopause were obtained 

 by both methods, and good agreement was obtained as long as sufficient 

 observations were available to yield reasonably reliable means. 



4. The temperature cross-sections are shown in Figs. 1 and 2 and 

 show the principal features common to all such diagrams. There is no 

 need to describe these well-known features here, but attention may be 

 directed to a few details that are not generally recognized. 



Although the mean tropopause is drawn as a continuous line it is 

 very probable that there are two tropopauses — tropical and polar — 

 and that there is a discontinuity at a certain latitude. This is clearly 

 seen in the mean sounding for Auckland (lat. 38° S.) in winter, where 

 both tropopauses appear with sufficient regularity to appear in the mean 

 temperature distribution. 



Another fact that has not hitherto been sufficiently emphasized is 

 that in the higher parts of the troposphere there exists in sub-tropical 

 and tropical latitudes a reversed temperature gradient in which the 

 isothermal surfaces slope upward from the Equator towards higher 

 latitudes. Willett(2) has described a similar effect in the North American 

 region in summer as being due to the heating effect of the American 

 Continent, but such an explanation cannot be applied to a purely oceanic 

 region, and the explanation must therefore be considered doubtful. 



64 



