i;indi3 and large sea-air temperature differences , such as might 

 be found nvorth of a strong polar front, Qg is extremely large 

 and mi-dit easily be greater in magnitude tnan Qp. When this 

 occui'S the net flux of heat per day is negative with respect 

 to the ocean; therefore, to arrive at the new thermal structure 

 it would be necessary to consider the process of convectional 

 cooling to relatively great depths. If, however, the flux 

 of heat is positive with respect to the oceans, the convection 

 resulting fcom exchange with the atmosphere will take place 

 vrithin a relatively shf.llow layer because of marked surface 

 stability. This is the reason why winter temperature traces 

 are essentially isotherpial to great depths whereas a shallow 

 thermocline exists in summer. Superimposed upon this effect 

 is the effect of turbulent mixing, which will further modify 

 the thermal structure. The processes of convection and 

 turbulent mixing will be discussed in the following paragraphs. 

 l\-. Mixing processes: Convection 



So far we have developed the following constituents of 

 a predicted thermal structure: (1) an amended temperature 

 distribution, Tp(z), which would result from a process of 

 radiation alone (discussed in 2 above), and (2) an amount of 

 heat, Qe, which is to be lost to the v/ater because of exchange 

 with the atmosphere during the forecast period (discussed in 

 3 above). 



Our problem novj is to aetermine the modifications that 

 v/ill occur in Tj,(z) due to the loss of heat Qg at the surface. 

 When one considers tills problem, it is necessary to consider 



12 



