SECT. 2] LARGE-SCALE INTERACTIONS 211 



The problem is a difficult one to treat theoretically since, in fully turbulent 

 flow, the applicability of the classical laminar stability treatments (Rayleigh, 

 1916; Jeffreys, 1926; Avsec, 1939; see also review by Stommel, 1947) is 

 doubtful, and the new turbulent models have not yet been extended to such 

 complex geophysical situations. When the cloud groups are more in random 

 bunches, it is likely that this scale of instability is not developed. In these 

 cases, there is evidence that the spatial variations in the ' thickness of the 

 mixed layer are associated with, or possibly forced by, thermal inhomogeneities 

 in the sea surface below. A theoretical treatment (Malkus, 1957) shows that 

 the observed ocean-surface temperature anomalies of 0.05-0.3°C are adequate 

 to produce the necessary deformations of the air flow ; the warm spots weakly 

 simulate the effects of small heated "islands" or "mountains" in the sea! The 

 interaction on the meso-, or 10-100 km, scale of oceanic and atmospheric 

 processes is a frontier of marine science just barely opened in the first decade 

 of aircraft research. 



Thus, trade-wind cloud groups are born in those areas where the homo- 

 geneous layer extends to the level of water-vapor condensation. At first, small 

 cloudlets 100-300 m across break out there in great numbers, from the wetter 

 eddies swirled upward by the turbulent trade. The larger clouds appear to be 

 built up by aggregations of several of these cloudlets, and the group as a whole 

 may have a lifetime of many hours. 



c. The cloud layer and its transports 



Growing day and night over tens of millions of square miles, the trade-wind 

 cumulus is the most common, and the most persistently studied, cloud form. 

 The cloud layer is defined as the vertical stratum extending from the condensa- 

 tion level to inversion base ; its thickness ranges from a few hundred meters in 

 the eastern and poleward fringes of the trades, to about 3 km as the equatorial 

 zone is approached. Here the cumuli themselves are the transporting elements. 

 In the intervening clear spaces, the air is weakly descending and small-scale 

 turbulence is entirely suppressed. Since 1946, the Woods Hole expeditions have 

 been building up a picture of the structure of these clouds and how they 

 interact with their surroundings (see reviews by Malkus, 1952; Malkus and 

 Witt, 1959). 



An aircraft profile through a large, active trade cumulus is shown in Fig. 54 

 (Malkus, 1954). Photographs of the cloud are shown in Fig. 55. The buoyant 

 updraft portion, with rising speeds of 1-4 m/sec, is restricted to small regions 

 within the cloud much of which is descending even in the active phase shown. 

 This phase lasts only 3-10 min, or a short fraction of the cloud's visible lifetime, 

 so that most cumuli seen or photographed on a given occasion are in an in- 

 active or decaying stage. We wish to inquire first whether and how these 

 evanescent and undistinguished-looking chimneys can be capable of the 

 enormous vertical moisture transports deduced in the budget studies (Section 5, 

 pages 144-147). 



