The Three-dimensional Temperature Distribution and its Variation in Time 115 



This effect may still be noted down to 50 or 100 m, but the heating process is inter- 

 rupted in June and only reappears later at 50 m. The reason for this remarkable 

 phenomena can be seen in the circumstance that turbulence due to the wind affects 

 the greater depths in spring, while the water mass processes an indifferent or weakly 

 stable stratification, so that surface heat can penetrate still to depths below 50 m. 

 The rapid temperature rise of the surface layers soon builds up such a strong tempera- 

 ture gradient thai turbulence is unable to prove a match for the created strong vertical 

 stability of the water masses and the turbulent transport of heat therefore ceases. The 

 upper layer is heated further by continued incoming radiation, and because of mixing 

 becomes almost isothermal while the lower layers remain cold. The temperature in 

 these layers rises again. Only when the density gradient is destroyed in the autumn 

 can the effect of mixing and convection again extend to deeper layers. Only then can 

 further heat be carried to the layer beneath the thermocline (Defant, 1936fl). 



In places where the ocean currents are subjected to considerable displacements, in 

 both direction and strength during the year, the annual temperature variation can 

 be considerably affected down to great depths by these current displacements. A 

 typical example for this is the annual temperature variation in Monterey Bay, Cali- 

 fornia (Skogsberg, 1936). Here there are three different periods in the annual variation: 

 the period of the Davidson Current from the middle of November to the middle of 

 February, when the temperature varies only slightly with depth down to almost 100 m; 

 then follows a period of upwelling water from the middle of February to the end of 

 July with low temperatures and stronger stratification ; while from the middle of July 

 to the middle of November the Californian Current prevails and the temperature 

 variation shows normal oceanic conditions. 



On the other hand, the temperature variation in the Kuroshio south of Japan 

 (KoENUMA, 1939) shows almost exactly the same conditions as in the Bay of Biscay 

 which was mentioned above. 



Fjeldstad (1933) has attempted to use the observations of Helland-Hansen in 

 area B to calculate the eddy conductivity coefficient from the changes in the annual 

 temperatuie variation with depth. He developed the annual temperature variation in 

 the individual depths into harmonic series, and obtained in that way the values c„ 

 and a„ as the amplitude and phase of the «th term of the series. Fjeldstad then showed 

 that 



A na 



dz. 



p K (^««/^^) 



P" 



where a = l-n-jT, T is the annual period and h is the depth at which the ampUtude 

 vanishes. A better representation of the observations can only be achieved by assum- 

 ing a seasonal variation of the eddy conductivity coefficient. The mean value at the 

 surface is 16gcm-isec-\ at 25 m it is 3gcm-^sec-^ and at 100 m the annual 

 mean is only 3, in summer 0-5 and in winter 5-5. 



The same method has been applied by Sverdrup (1940) to values for the Kuroshio 

 which in this case appears to be permissible, since the advective effects are outweighed 

 by radiation and the eddy conductivity. In the Kuroshio area, where the strength of 

 the current is large and the turbulence correspondingly high, the annual temperature 



