SECT. 2] LARGE-SCALE INTERACTIONS 201 



and/or the less the radiation loss. The latter becomes the predominant factor 

 as the size and time scale of the considered circulation branch becomes large. 

 Efficiency of condensation conversion only might be considered on the synoptic 

 scale, as done by Riehl and Byers (1958) in seeking upper limits for tropical 

 storm development, while in the 2500 km-long trade section we would probably 

 be unjustified in ignoring variations in Ra. Kraus' large-scale study considered 

 efficiency variations entirely in terms of alteration in radiation sink ; in the 

 entire tropical belt, as we saw, nearly all latent heat is converted before export. 

 Clearly the dependence of fluid heat-engine efficiency upon scale of motion is a 

 frontier facet of stability studies as yet barely touched by geophysicists (Spiegel, 

 1958; Goody, 1956). 



A fortunate and vastly important feature of the lower trades is their two 

 dimensionality, which had to be treated as given in the limited dynamic model 

 described. Also related intimately to their stability against synoptic-scale 

 disturbances, it has so far received only a physical explanation (Riehl, 1951). 

 To roll up the currents into waves or vortices, with lateral and tangential 

 velocity components varying both with s and n, hydrostatic pressure fields to 

 sustain these flow configurations must be produced; that is to say, similarly 

 convoluted horizontal temperature fields must be maintainable in the atmos- 

 phere, with synoptic-scale gradients in the s and n directions. In the lower 

 trades, this is just what sea-air interaction prevents. The close coupling by 

 exchange and vertical convection to a vast homogeneous oceanic reservoir 

 precludes development of the necessary sharp air temperature gradients. Thus 

 lateral eddies must recede in importance : because of lack of thermal stability 

 over millions of square miles, the extreme smallness of oceanic temperature gradients, 

 and the effectiveness of vertical turbulence, horizontal circulations are held to a 

 minimum. Flow in the low levels is, therefore, quasi-uniform, particularly 

 under trade inversions ; the air below is both shielded from any disturbing 

 influence from high levels and contains internal mechanisms for restoring 

 departures which may be imposed. Above the convective layer, as in most of 

 the middle-latitudes, these restoring mechanisms and the vertical coupling to a 

 uniform sea no longer exist and the steadiness is absent. Controlled rotating 

 dishpan or channel experiments could profitably be undertaken to explore the 

 range of validity of this two-dimensional property of bottom-heated currents. 



As the trades themselves flow equatorward, they sow the seeds of their own 

 destruction as a stable system, by weakening the inversion lid and deepening 

 the moist layer by convection. When, finally, "hot towers" manage to shoot 

 through to the tropopause, deep wave- and vortex-sustaining pressure fields can 

 be organized (Malkus and Riehl, 1960). Then synoptic-scale disturbances be- 

 come common through the whole depth of the troposphere. In these, air-sea 

 fluxes and their mechanisms are violently stepped up and, paradoxically, play 

 a dominant role in disrupting the very machinery whose smooth operation they 

 formerly maintained. It is with the exchange mechanisms themselves and 

 their fluctuations on the synoptic and longer scales that we shall be concerned 

 in the final sections. 



