SECT. 2] SMALL-SCALE INTERACTIONS 69 



rather than recorded continuously for the whole period, it is virtually essential 

 to employ the difference method in order to avoid excessive sampling errors. 



Measurements of temperature and/or humidity profiles, employing sampling 

 at individual heights, have been published by Wiist (1937), Montgomery 

 (1940), Bruch (1940), Sverdrup (1946) and Takahashi (1958). Some of these 

 results have been collated and discussed by Brocks (1955). Measurements using 

 a pair of thermistors to record differences are given by Deacon et al. (1956), and 

 further measurements in Port Phillip Bay and Bass Strait in November, 1956, 

 and May, 1958, are as yet unpublished. In the last-mentioned paper are repro- 

 duced also some earlier results of Johnson and Meredith, who measured dif- 

 ferences with the aspirated platinum resistance thermometers described by 

 Johnson (1927). In all the work referred to above, the observations were made 

 from shipboard. Other recent investigations are those of Fleagle, Deardorff and 

 Badgle}" (1958) using a raft-mounted mast, and of K. Brooks (results not yet 

 published) using the buoy-mast illustrated in Fig. 4. 



F. Nature of Observed Results 



Tjrpical forms of temperature profile in lapse and inversion conditions are 

 illustrated in Fig. 11. In (a) the plot against a linear height scale shows the 

 marked height-dependence of the gradient. In (b) the plot against a logarithmic 

 height scale is more nearly linear, but is concave towards the height axis in 

 lapse conditions and convex in inversion conditions. These characteristics are, 

 of course, broadly the same as observed over land surfaces. The potential 

 temperature difference between the surface and a height of a few metres 

 comprises, roughly speaking, about one-fifth across the layer of molecular 

 transfer about a millimetre thick adjacent to the water surface, and the other 

 four-fifths across the overlying region of turbulent transfer ; this is estimated 

 from the theoretical approach to be dealt with later. 



Profiles of specific humidity (or vapour pressure) are qualitatively similar in 

 form, having the same rule of dependence on thermal stratification as stated 

 above. However, the sign of the humidity gradient is, of course, normally 

 negative, the greatest humidity being right at the water surface and corre- 

 sponding at least approximately to saturation at the temperature there. 



One might expect as a first approximation that the potential temperature 

 gradient at a given height, or difference between two given heights, would be 

 proportional to the difference of temperature between sea and air ; that is, that 

 Fh would be approximately constant. Broadly speaking, this is found to be 

 true at the lower levels, as illustrated by the measurements plotted in Fig. 12. 

 At higher levels, say above 4 m, the effect of thermal stratification illustrated 

 by the curvature in Fig. lib becomes important. This can be seen in Fig. 13, 

 where measurements of ^12.6 — ^4 are plotted. It is apparent that the greater 

 the sea-air temperature difference and the lighter the wind, the greater is the 

 deviation from the linear relationship (though these measurements unfortu- 

 nately do not include light winds on the inversion side, Ta—Ts> 0). The 



