100 



MALKUS 



3. Determination of Air-Sea Fluxes 



[chap. 4 



The classical works on global air-sea fluxes and the overall climatology of 

 exchange have been a series of heroic researches by W. C. Jacobs beginning in 

 the early 1940's (Jacobs, 1942, 1943, 1949, 1950, 1951, 1951a). Recently, a 

 series of Russian studies, carrying out and extending the same type of computa- 

 tion with a more modern and complete set of data, has become available ; these 

 data have been summarized in a monumental work by Budyko (1956) which 

 will be drawn upon heavily in this chapter. All in all it may be said that both 

 the methods and results of Jacobs and his predecessors (Wiist, 1936; Mosby, 



40° 

 North 



30° 

 Lafi tude 



Fig. 4. Mean annual evaporation in North Atlantic and North Pacific together as function 

 of latitude. Comparison of determinations of several workers: Mosby (1936), Wiist 

 (1936), Jacobs (1951a) and Budyko (1956). 



1936) have borne the test of time remarkably well. First the computational 

 procedures and assumptions have been subject to considerable test by use in a 

 variety of situations (cf. Sections 5-7 in this chapter) and some (still limited) 

 direct verification in actual aircraft flux measurements (Bunker, 1960). 

 Secondly, the actual figures arrived at by the Russians and other recent 

 workers differ very little indeed from those presented earlier (see Fig. 4). In 

 short, air-sea flux computations appear to be stable and reproducible, and to 

 fit in consistently with mechanistic, budget and dynamic requirements on 

 numerous scales. 



Direct measurements of evaporation and sensible heat flow from sea to air 

 are difficult, doubtful and scanty. What is wanted is a method of obtaining 

 reliable exchange values at numerous points over the oceans as a function of 

 time, from measurements cheaply and routinely obtainable. At present, this 

 ideal is realized only approximately. Sensible heat and water-vapor fluxes 

 between ocean and atmospliere may be computed indirectly in two independent 



