obtainedby Yates and Taylor, which are available from the original Naval Research Labora- 

 tory Report, reference 4, and are reproduced in Holter, et al, reference 3, have been used to 

 plot the average transmission in the 8^ to 14/i band used by the Barnes Infrared Therm- 

 ometer. The average transmission versus precipitable water vapor present is shown in 

 Figure 7. Given the flying altitude and absolute humidity (or the relative humidity plus air 

 temperature), we can determine the quantitity of precipitable water vapor in the path be- 

 tween the airborne Infrared Thermometer and the ocean; and from Figure 7 we can de- 

 termine the trans mis sivity of the atmosphere seen by the Infrared Thermometer. This 

 can be substituted into an equation which will be shortly developed, along with air tempera- 

 ture to give the error which is introduced, so that our raw data can be corrected. Although 

 strictly empirical, this technique should normally provide an absolute accuracy of better 

 than 1°F. 



To the extent that the atmosphere absorbs the ocean signal, it also emits a signal of 

 its own, which cause an additional error, but which is also corrected out in the same 

 equation. 



One further signal remains to be considered. This is the signal which is reflected 

 from the ocean. On days in which considerable atmospheric absorption exists, this re- 

 flected signal is merely more atmosphere. On a day in which no atmospheric absorption 

 whatever is present, this signal would be cold sky, or outer space. On intermediate days 

 in which small atmospheric absorption is present, the reflected signal is neither lower 

 atmosphere nor cold outer space, but something in between. Normally, the solid angle of 

 "sky" seen by the radiometer is nearly a hemisphere, since the wave slopes on the surface 

 of the ocean reflect much "horizon". Certainly much of the 6.28 steradian viewed by re- 

 flection consists of sufficient air paths to be relatively opaque. In the extreme case of a 

 perfectly smooth, flat, 27°C ocean surface and a 100% transmissive atmosphere, the "sky 

 error" would be 1.5°C. Ewing and McAlister'^ found that a clear sky at night produced an 

 apparent ocean surface temperature a few tenths of a °C colder than actual. 



PREClPlTftSLE VATER.CH 



Figure 7. Average Transmission of the Atmosphere vs. Precipitable Water, 1000 



Feet Path, 8-14 Microns 



-31- 



