The results of the comparative flights are shown in Figures 2 and 3. The first day's 

 flight, including transects B-E, was somewhat hampered by navigational errors and instru- 

 ment problems. The second day's flight, including transects F-J, was better coordinated 

 and provided a more controlled test. A general consistency can be seen in gradient patterns 

 along all transects. For transects F-J, the more comparable set, the differences vary from 

 1° to 4°F. The average difference of about 2.5° F. lower for the Navy instrument is a matter 

 of instrument calibration. The real variation then was ±1.5°F., this is 0.5° F. higher than 

 the nominal accuracy level of ±1.0° F. required to meet our program objectives. In order to 

 reach the higher level of accuracy required, a completely new instrument ensemble will be 

 utilized; power supply, recorder, and IRT will all be replaced with improved equipment. 



It is apparent that subsurface temperatures can be estimated from IRT surface read- 

 ings if all significant parameters of variation are known and their effects precisely determined 

 There is at hand neither the theory nor the empirical data with which to do this. However, the 

 major effect appears to be produced by the annual cycle of solar radiation. This is demon- 

 strated in Table 2 where we have listed micro-surface/bucket-depth temperature difference 

 for a series of monthly flights extending over one year, all of which were made in daylight 

 hours between 8:30 A.M. and 4:00 P.M. The micro-surface tends to be cooler than the im- 

 mediate subsurface in Autumn and Winter, nearly the same in Spring, and apparently some- 

 what warmer in the Summer. The "average" condition is for the micro-surface to be about 

 1.0° F. cooler than the immediate subsurface. The data are plotted in Figure 4. 



NoyllC«l MI1«a 



Figure 2. Temperatures from Sandy Hook - Navy comparative 

 flight for transects B-E, September 30, 1963 



■138- 



