TABLE 2 



Change in Temperature Differentials from Fitted Cxirves (° p) 

 Hour (Z) 0-20 ft 20-40 ft 40-60 ft 60-8O ft 



1200-1300 



.022 



.026 



.008 



.005 



1300-1400 



.027 



.032 



.012 



.006 



1400-1500 



.028 



.035 



.018 



.004 



1500-1600 



.026 



.033 



.022 



.004 



1600-1700 



.021 



.028 



.022 



.004 



1700-1800 



.012 



.018 



.020 



.001 



1800-1900 



.002 



.006 



.014 



.000 



1900-2000 



-,008 



-.007 



»005 



-.002 



2000-2100 



-.017 



-.017 



-.004 



-.006 



2100-2200 



-.022 



-.022 



-.011 



-.007 



2200-2300 



-.024 



-.028 



-.021 



-.010 



2300-0000 



-.023 



-.027 



-.023 



-.002 



If the rate of phase difference is estimated to be 20 feet per 

 hour, the amount of change may be estimated. The average absolute tem- 

 perature differential change within the 20- to 40-foot layer in any 

 given hour and vithin the 4o- to 60-foot layer an hour later is ,0073° F. 

 Using a specific heat of .93 cal/gm/° C, the thermal conductivity co- 

 efficient is .00065 cal/cm/sec/° C. The figure of .OOO65 is considerably 

 lower than that given in the Smithsonian Tables (.00143). Thus some 

 fifty percent of thermal conductivity appears to be counterbalanced by 

 turbulent mixing due to wave action. It should be noted that Qxible 2 

 includes only daylight hours during which convective mixing was absent, 

 so that tiirbulent or mechanical mixing constitutes the only process for 

 vertical heat transfer that is unaccounted for by this analysis. 



VARIABILITT 



In the discussion above, conclusions have been confined to the ex- 

 planation of mean temrperature and mean salinity changes in the upper 

 layer of the ocean. However, as with all observations, there is much 

 variability in the temperature readings from which the mean values are 

 computed. The major sources of variation may be summarized as follows: 

 (1) Instrumental error; that is, errors caused by faulty temperature and 

 pressure elements in a bathythermograph. Such errors may be assvuned 

 constant for temperature and to increase slightly with depth as pressure 

 increases. (2) Reading error, defined as the average difference between 

 readings of the same BT trace by two different persons. (3) Interdiurnal 

 variability; that is, day-to-day difference in temperature owing to either 

 short- or long-period fluctuations such as tidal periods and seasonal ef- 

 fects, plus variability introduced by grouping the observations by hours. 

 (4) Advection or horizontal movements of water of differing temperatures. 



17 



