106 MALKUS [chap. 4 



Therefore we have for an annual average 



Q, = LE = {R-Qro)j{\+r). (8) 



It is worth noting that, since climatologically r is generally small compared to 

 one, particularly in the tropics where it averages less than 0.1, uncertainties in 

 its determinations are not serious in computing evaporation. Jacobs used 

 equation (8) to obtain evaporation over portions of the oceans where Qvo is 

 known to be small. He then used the resulting Qe to obtain the proportionality 

 constant in the transfer formula ; the latter was applied, using extensively 

 available meteorological data, to assess the distribution of evaporation over all 

 the Northern Hemisphere oceans (summarized by the solid curve in Fig. 4). As 

 a retrospective check, he found that in those regions where Qvo should be 

 significant, the departures of the results of (8) from those of the transfer 

 formula were logically explainable in terms of major current transports. 



Ideally, air-sea budget studies could be completed by independent determina- 

 tion of all terms in (8) ; however, due primarily to inadequate data and still, to 

 some extent, to imperfect or quasi-empirical physical laws of restricted applica- 

 bility, this is rarely possible. Most existing studies of air-sea exchange and 

 budget energetics are, therefore, forced to restrict their range of validity by 

 neglecting terms, by approximating, and often by computing one term as a 

 residual, to be checked by fragmentary observations and consistency. Despite 

 these obstacles, studies of this sort stand today as one of the largest single 

 contributions to meteorology and oceanography, and have been paramount in 

 the enormous increase in our knowledge of the tropical fuelling region, as will 

 be illustrated in the remainder of this chapter. The major step in enabling these 

 studies, and giving confidence in the general soundness of their results, is the 

 development of the exchange formulas and their placement upon a sound 

 foundation, which has progressed greatly since the days of their first courageous 

 use by Jacobs twenty years ago. 



B. Exchange Computations by the Transfer Formulas 



The classical works in this area were carried out in the decade 1930-1940 by 

 Rossby and Montgomery (1935; 1936) and by Montgomery (1940). Their aims 

 were a quantitative formulation of the turbulent structure at the air-sea inter- 

 face and a method of deducing thereby the fluxes of heat, moisture and momen- 

 tum from the ocean surface through the lowest decameters of the air above. In 

 the intervening years, this approach has been refined, examined critically, 

 extended and tested, particularly by British and Australian workers (Sheppard, 

 1958; Deacon, Sheppard and Webb, 1956; Priestley, 1959), some of whom 

 report on it in detail elsewhere in this volume. For present purposes, we shall be 

 interested mainly in the application of the results to the larger-scale air-sea 

 interaction problem ; it is, however, necessary in doing this briefly to re- 

 capitulate the basic foundations and assumptions of the formulations in order 

 to delimit their range of applicability to our problem. Fortunately, as we shall 



