Such a wave peak does not occur in either of the fixed-thermistor 

 temperature spectra for the November 12 (moderate winds) data (Fig. 7), 

 Only the temperature spectrum of the float -mounted thermistor data 

 recorded during the first half of the November 12 run shows a peak 

 at the frequency where the wave spectrum peaks (Table I). 



The wave spectra are typical narrow-band wave spectra but are not 

 completely analogous to ocean wave spectra. The wave energy contribu- 

 ting to the high frequency end of the wave spectrum is beginning to 

 come from capillary waves. 



In almost all cases there is at least a decade where the tempera- 

 ture fluctuation spectra follow the expected -5/3 power law. The fall- 

 off from the -5/3 slope beyond 10 Hz is due to a combination of low 

 pass filters and thermistor response. 



CONCLUSIONS 



These data indicate that there was a large temperature gradient 

 near the lake surface and possibly just below the lake surface. Just 

 how applicable these results are over the ocean remains to be seen, 

 but certainly there are appreciable portions of the ocean where strong 

 temperature gradients can exist: either cold air over warm water (such 

 as was experienced in the Pacific Northwest in the winter of 1968) or 

 warm air over cold water (such as off the California coast). 



The heat flux is considered in general to be invariant with height. 

 Heat flux estimates are often based on the wind velocity and air tem- 

 perature at 10 m. and the sea surface temperature. This would appear 

 to be insufficient since the heat flux according to these results is 

 going to be determined in the first few centimeters above the surface. 

 This being the case, any parameterization should include in some form 

 the temperature gradient just above and just below the sea surface. 

 The effect of waves on the transport of sensible heat would appear 

 to be small. 



The large temperature fluctuations were at first surprising. 

 Extensive laboratory studies (Ramzy and Young, 1969) were made on 

 the thermistors to test their response to velocity and humidity fluc- 

 tuations and wetting. The linearity of the circuit was also rechecked. 

 None of these features could account for the large fluctuations observed. 



The data are being reworked so as to improve the spectra and clarify 

 questions concerning the relation between the wave spectrum peak and the 

 temperature spectra. Whether or not such a peak occurs in the spectrum 

 derived from the fixed thermistor and how close it must be to the sur- 

 face for the waves to influence its spectrum has some bearing on wave 

 generation theories (Stewart, 1967). 



381 



