The mean wind at both sides of the divergence zone was evaluated to be 

 28 knots for 30 hours. Both fetch and duration were sufficient for fully- 

 developed sea. The sea state parameter 17 from Table 5 is 203- Stability- 

 index was between 6° and 7° F. From the lower solid and dashed (diver- 

 gence) curves in Figure 17, predicted therrnocline depth at this 77 value 

 is 109 feet . 



The mean observed therrnocline depth computed from 3 BT's taken between 

 0900Z and 1600Z on 31 May was 100 feet . 



Approximation 



The variables and conditions considered in this prediction method seem 

 to be acting logically in the right directions and proportions. The method 

 is far from being exact, but the degree of approximation can be improved. 



The first source of inexactness may lie in the prediction technique. 

 The mixing factor k and the k(7i) curves were developed empirically from 

 rather limited data. Determination of 356 values of k distributed over 8 

 intervals of stability indexes does not provide good concentration of 

 points; therefore, the curves in Figures 15 through 22 may shift somewhat 

 from their actual positions. Such deviation should produce a systematic 

 error, the extent of which cannot be evaluated at present; however, this 

 error is assumed to be rather small when compared with other causes of 

 error. 



Inexactness produced by incorrect evaluation of parameters will cause 

 a random error which may often be quite large, especially if several com- 

 ponents of this error should possess the same algebraic sign. The first 

 of these erroneous components is introduced by inexact evaluation of the 

 wind field and, consequently, erroneous wave parameters. This error is 

 largest for weak winds and low stability indexes and is least with strong 

 winds and high stability indexes. 



Error Due to Wind Evaluation and Decrease of Error With Increasing Wind 



With a stability index of 9°, 15-knot winds in a normal field, and 

 fully developed sea, predicted therrnocline depth would be 86 feet . Under 

 these same conditions with l6-knot winds, the depth would be 108 feet . 

 Thus, erroneous evaluation of mean wind speed by one knot in this example 

 would produce a difference of 22 feet in the prediction. With the same 

 stability index, 26-knot winds, and fully developed sea, predicted therrno- 

 cline depth would be 139 feet ; whereas, with 27-knot winds it would be 

 1^2 feet - now the difference is only 3 feet. 



Error Due to Wind Evaluation and Decrease of Error With Increasing At 



With a stability index of 2°, 23-knot winds in a normal field, and 

 fully developed sea, predicted therrnocline depth would be 167 feet . Under 

 these same conditions with 2k- knot winds, the depth would be 182 feet . 



71 



