226 AIALKUS [CHAP. 4 



While Qe varied little throughout, we see a marked enhancement of sensible 

 heat input in the disturbed period, and a marked suppression in the fair period, 

 in inverse variation with the wind strength. As the table shows, the explanation 

 lies in the changes in the sea-air temperature difference which was very large 

 in disturbed conditions and averaged vanishingly small in fair periods. 



Figs. 59-61 provide increased mechanistic insight. In these figures, the more 

 detailed breakdown into periods is as follows : 



Period A: 17-19 August. Ship station continually in disturbed convergent 

 zone. Total rainfall 41.9 mm, cloud cover always greater than 75%. Average 

 wind speed 4.5 m/sec, direction variable, sometimes westerly. Directional 

 steadiness low. 



Period B: 25-26 August. Fairer than average, only three light showers. 

 Cloud cover less than 50%. Average wind 4.8 m/sec from east-northeast. 



Period C : 1-2 September. Very fair period with only one brief shower. 

 Cloud cover well below 50%. Average wind 7.2 m/sec from east-northeast. 



The ocean-surface temperature varies little between the periods, although it 

 is slightly cooler in the strong-wind fair period C, than during the weaker-wind 

 fair period B, indicating the effect of stirring. However, the dominant control 

 upon To— Ta is produced by air temperature variations. Disturbed Period A is 

 consistently colder (up to 3.0°F) than any other: the fairest period, C, is the 

 warmest. Reduced insolation due to cloud cover, direct cooling by rain showers, 

 as well as the descent from aloft of cool air associated with the condensation- 

 precipitation cycle, all contribute to these differences which are demonstrably 

 not advective in origin. 



Figs. 59-61 in conjunction with Figs. 62-65 i)ermit for the first time a 

 quantitative picture of the diurnal cycle in exchange and transport over the 

 tropical oceans. The average sea-surface temperature (solid curve. Fig. 60) 

 shows a well defined diurnal variation with a spread of 0.6°C. The diurnal air- 

 temperature range is about 1.5°C, in contrast with the 7-8°C common over 

 islands and land-masses in the same latitudes and season. Comparison of 

 Figs. 59 and 60 shows that there is considerable lag in the diurnal heating and 

 cooling of the sea surface relative to the air. Whereas the minimum air tempera- 

 ture is registered at 0500 LST, the minimum sea-surface temperature occurs at 

 0700 LST ; similarly, the maximum air temperature is reached at 1300 LST, while 

 the sea-surface temperature reaches its maximum between 1600 and 1700 LST. 

 We may deduce from this an important consequence to exchange, namely, a 

 minimum in Tq— Ta around midday. Fig. 61 shows that the minimum consists 

 of an actual reversal in sign, so that the sensible heat flow is from air to sea in 

 the noon hours. Thermal turbulence should then reach its minimum. Figs. 63-65 

 confirm this deduction and illustrate its consequences upon cumulus convection. 

 As astute observers in the tropics have long siis])ected. there was an oceanic 

 cloudiness minimum around midday and maxima in the dawn and midnight 

 hours. Our suggestion in the preceding sub-section connecting trade cumuli with 

 active transports in a turbulent mixed layer has received direct support, since 



