SECT. 2] LARGE-SCALE INTERACTIONS 219 



sea and air, and thus might well affect subsequent exchange, storm formation, 

 etc., as well as the salinity, convective structure and biological processes in 

 the thermocline region of the sea. However, rather than indulge in the intriguing 

 speculation to which this whole subject tempts the unwary, it is less news- 

 worthy but more useful to lay a quantitative foundation for a series of specific 

 questions. What actually happens to processes in the moist layer when the 

 trade strength oscillates about its mean? 



The description of trade-wind structure and transports in the preceding sub- 

 section was based largely upon the April, 1946, (Wyman-Woodcock) Caribbean 

 expedition, which investigated a strong trade regime (average wind from 

 090°, 9.1 m/sec). Fortunately, a succeeding Woods Hole expedition in April, 

 1953, found in the same area and season a situation of low index and poorly 

 developed trade (average wind from 106°, 5.7 m/sec). Although the data were 

 not adequate to resolve definitively many of the vital questions, a framework 

 of inquiry was set up in a monograph by Malkus (1958) contrasting the two 

 sets of results. 



The most obvious visual differences between the weak and strong trade 

 situations lay in the sea state and cloud structure. In the strong trade case, 

 white caps and rough seas prevailed on all observing days, while they were 

 absent 80% of the time in the low index situation. Apparently correlated was 

 the development of trade cumulus clouds, which were plentiful and vigorous 

 during the first expedition and suppressed, feeble and sparse during the second. 

 On the days of weakest and most southerly flow, glassy seas with unprecedented 

 clear spaces, hundreds of miles across, frustrated the aircraft observers, who 

 sought to duplicate the cloud penetrations of Figs. 54 and 55. The sounding data 

 showed that the reason lay in the disrupted function of the lowest air layers and 

 suggested the stopping down of the exchange valve. 



While the properties of the cloud layer differed little from those shown in 

 Fig. 52, in the low index case the mixed layer was abnormally shallow. On the 

 poorest days for trade cumuli, its top fell about 200 m below the condensation 

 level. In addition, it was far from homogeneous. While the mean temperature 

 and specific humidity of the lowest kilometer were nearly identical to those of 

 Fig. 52, their vertical stratification showed the consequences of reduced 

 turbulence : the temperature lapse rate was slightly less stable than normal and, 

 most important, the upward rate of moisture decrease was 3.5 times greater 

 than in the strong trade situation. It appears that the water vapor was just not 

 getting pumped up through the sub-cloud layer and it, therefore, accumulated 

 in the lowest levels. This may in turn have inhibited further evaporation. 

 Reduced upward transport resulted from weak and southerly flow. The former 

 gave less surface roughness and diminished forced stirring. The latter, perhaps, 

 was responsible for weakened thermal turbulence at low levels due to reduction 

 of the sea-air temperature difference. The aircraft-measured momentum fluxes 

 averaged about 40% of normal. 



In striking contrast to the suppression of trade cumuli, the April, 1953, low- 

 index period provided one of the best displays of tropical hot towers which 



