1 22 The Three-dimensional Temperature Distribution and its Variation in Time 



equalization of temperature thus forming an upper isothermal top layer. Beneath this, 

 at a definite constant level, lies the thermocline, which acts as a barrier for all turbulent 

 processes. The important point in the explanation of the formation of tropical and 

 subtropical thermoclines is the exclusion of turbulence and their consequences in a 

 fixed depth due to the increase of the vertical density gradient above a critical value. 

 The condition for the reduction and final elimination of the turbulence in a non- 

 laminar flow is that the dimensionless quantity (Richardson number): 



{glp){hpidz) Ab 

 {dujdzY -^ At' 



In this relation u is the basic velocity of the turbulent flow, /Ib is the exchange coefficient 

 for the flow momentum (apparent viscosity) and At is the exchange coefficient for 

 density differences (temperature and salinity). In the ocean the ratio between these 

 two quantities is between about 5 and 20 (see p. 103). In the thermocline of the equa- 

 torial region of the Atlantic the quantity Spjdz is of the order of 5 X 10"* for a vertical 

 interval of 20 m. In drift currents dujdz can be taken as about 10 cm/sec for every 

 20 m. The left-hand side of the above inequality is thus 100, which is considerably 

 more than the value of the right-hand side. With such a stratification the turbulence 

 in a current cannot be maintained (Defant, 1936), 



The basis of the theory for the formation of the thermocline has been given by 

 MuNK and Anderson (1948). They have shown that the sharp transition between the 

 top layer with mixing and the thermocline can be explained theoretically on the as- 

 sumption that the eddy coefficients are a function of the vertical stability and of the 

 wind shear. This theory gives a value for the depth of the thermocline that is some- 

 what too sm.all but it is of the correct order of magnitude. This depth depends on the 

 wind velocity, on the latitude, on the heat flux and on the [r^Sl-relation in that order. 

 This theory undoubtedly penetrates deeply into the important processes that control 

 this phenomenon but it does not yet completely satisfy all points. Experimental 

 investigation and systematically planned observations would very probably improve 

 the basis of the theory. 



(b) The Oceanic Stratosphere 



The vertical temperature differences in the very deep layer of the oceanic strato- 

 sphere are small. Here also the distribution is almost everywhere anothermic; however, 

 the temperature gradient at depths below 1000 m falls rapidly to values less than 

 0-4°C per 100 m, at 2000 m it is at the most O-TC and at 3000 m and below it is 

 barely 0-05 °C/ 100 m. Departures from this anothermic distribution are found only 

 in the Western Atlantic (Brazilian and Argentinian basins) and in the south-western 

 Indian Ocean where at a depth of 1300-1600 m there is a very weakly marked tem- 

 perature inversion, a phenomenon of particular importance for the oceanic circulation 

 of these oceanic spaces. Table 51 shows particularly well-developed inversions at some 

 "Meteor" stations. Inversions such as these occur only rarely in the eastern half of the 

 South Atlantic and are very weak. They appear to be due to long-term changes asso- 

 ciated with aperiodic variations in intensity of the deep-sea circulation (Merz, 1922; 

 WiJST, 1936, 1948). 



