Mean values of At at all stations follow an annual wave pattern rather 

 consistently. Except for temperatures at the UOO-foot level at station 

 DELTA, the mean monthly temperatures also generally fit the annual wave 

 pattern. Mass advection apparently plays an important part in temperature 

 distribution of area DELTA. It should be noted that the number of obser- 

 vations for this station is considerably less than those available for 

 other areas. 



During winter months, the frequency distribution of At is strongly 

 skewed, and the range of At is quite limited. Such a distribution is 

 understandable because the mixed-layer thickness is generally greater than 

 U00 feet during these months. When heating of surface water increases, 

 the frequency distribution becomes more symmetrical and often approaches 

 the normal distribution. Double maxima patterns, similar to those occur- 

 ring in September and November in Figure 30 or in July in Figure 31, are 

 due to anomalous data of one or a few years for these months. The anoma- 

 lies may be caused by erroneous BT's or by mass advection which may per- 

 sist for several weeks. Some BT data are questionable, owing to prolonged 

 use of unreliable instruments. 



Temperature values at the ^00- foot level generally follow the trend 

 of values at the surface during the seasonal thermocline. If the surface 

 temperature is a few degrees higher than the mean during a given month, 

 the mean temperature at 400 feet can also be expected to be higher than 

 the mean at the ^00-foot level. The mean temperature at the 400-foot 

 level during any month is rather stable; therefore, the range of At should 

 be small. The range appears rather large, however, during the seasonal 

 thermocline, because instantaneous temperatures taken at ^00 feet often 

 show large oscillations due to internal waves at the bottom of the thermo- 

 cline . 



At values that lie approximately within the limits of standard devia- 

 tions in frequency distributions can be termed "real"; values near the 

 flanks, however, should be called "apparent", because the majority of them 

 are produced by short-term temperature oscillations at the 400- foot level. 



Statistical values are computed from the collective data for a 5-d-egree 

 square around the center of each of the 3 stations. The combined areas of 

 these squares represent only a small fraction of the North Atlantic; there- 

 fore, mean values of temperatures and stability indexes can serve only as 

 general information and for comparison with other information. 



Prediction of Mixed-Layer Decay and Associated Thermocline Depth 



Conditions are uncertain after active mixing ceases and decay of the 

 mixed layer begins. This situation may be partially relieved by surface 

 conditions which might exert a mixing force on the mixed layer. The ques- 

 tion is: How deep will the force penetrate? If it penetrates to the bottom 

 of the mixed layer, decay will be prevented; if it does not extend to the 

 bottom of the mixed layer, decay may advance upward to the level of effec- 

 tive mixing. In either case mixing always takes plaee in the mixed layer 



