the first layer to the third layer. In summation, the maximum heat 

 differential between the first and second layers and between the 

 second and third layers did not occiir xmtil I80OZ and 2000Z respec- 

 tively, or nearly at s\mset, although the maximum svirface tempera- 

 ture was reached at I70OZ or 2.3 hours after local noon. 



Figure 12 continues the study of differential heating of layers 

 and shows that the differences of heat content between the third and 

 fourth layers and between the fourth and fifth layers do not show 

 much regxilarity. This indicates that the heating process is confined 

 mainly to the upper hO feet, although convection extends to 60 feet 

 if differential heating alone is considered. In contrast, reversed 

 logic leads to the conclusion that appreciable solar heating act\aally 

 extends well below 100 feet as shown by the peaks in Figure 2. How- 

 ever, the amount of radiation absorbed in each layer is more constant 

 than wovG-d be assumed from normal extinction coefficients (Sverdrup 

 et al., 19'<-2). As explained above, it must be remembered that Figure 

 2 is complicated by apparent internal wave action below a depth of 60 

 feet, 



COEFFICIENT OF THERMAL COlTDUCTIVITr 



An inherent property of substances is equalization of temperature 

 within the substance (assuming absence of outside influence). This 

 property is thermal conductivity and occurs when thermal gradients ex- 

 ist in the substance. The coefficient of themnal conductivity for sea 

 water is slightly smaller than that for fresh water, according to the 

 Smithsonian Tables. The BT data described above may be used to esti- 

 mate this coefficient. 



Since the coefficient of thermal conductivity is given in units 

 of calories per centimeter per second per degree centigrade, it is evi- 

 dent that the dimensions of these units are equivalent to the rate of 

 change of the temperature gradient. If this rate of change can be ap- 

 proximated, a value of the coefficient can be determined. In terms of 

 finite differences, the temperature difference between layers is pro- 

 portional to the temperature gradient. The time change of the differ- 

 ence is the first differential, and the rate of change is the coeffi- 

 cient of thermal conductivity. 



Sampling errors are minimized by use of fitted curves. Therefore 

 a Fourier analysis (for the second hamionic) of the differences plotted 

 in Figure 3 was performed by digital computer. The fitted curves are 

 shown in Figure I3, Gradients within the layers are shown in Figuire l4. 

 Table 2 lists changes in temperature differences, from which the coeffi- 

 cient may be estimated. In Table 2, phase changes from positive to neg- 

 ative or differences of maximal differentials are about one hour between 

 the 20- to ifO-foot layer and the ko- to 60-foot layer. The 0- to 20- 

 foot layer must be omitted because of wave action. As shown in Figrire 

 l4, temperatxire gradients in the upper two layers (0-20 and 20-Uo feet) 

 are in phase; however, wind mixing reduces the gradients in the 0- to 

 20-foot layer so that eddy conductivity and thermal conductivity cannot 

 be distinguished. 



