Maximum Gradient-Gradient (3) 



y=il-.8l - 1.328 sin x - I.O68 cos x + 0.^56 sin 2x - COS^^ cos 2x 



Maximum Gradient-Thickness {k) 



y = 91.0 - kh.9 sin x - 35.2 cos x + 18.3 sin 2x + 9.0 cos 2x 



Maximum Gradient -Temperature of the Upper Bound (5) 



y-71.90 - 5.36 sin X - 2.12 cos x + l.63 sin 2x - 0.20 cos 2x 



First Gradient-Thickness (6) 



y = 107,0 - 80.6 sin X - 66.2 cos x + 22.7 sin 2x + 8.0 cos 2x 



where y = the desired parameter mean for a given month, 

 x=15° for January , 

 = 45° for February, 

 =75° for March, etc. 



Minimum and maximum values could he determined from these ctirves. 

 These values usually occur 6 months apart; however, month of occiurrence 

 varies from parameter to parameter. Minimums and maximiims of some pa- 

 rameters may persist for a period of 2 months, as shown in columns 2 

 and 3 of table 25 (appendix B) . This indicates that thermal structure 

 change occurs at a relatively slow rate during maximiim and minimum pe- 

 riods. The minimiOTs and maxlrnvms of certain parameters, however, do 

 not consistently occur in the same month from year to year. 



Owing to use of a 450-foot BT to obtain the historical data, minimums 

 and maximums of certain parameters could not be computed with any degree 

 of reliability during January, February, March, and April. 



Maximiam surface temperature can be expected during August or Sep- 

 tember. The minimum occurs during March or April. The layer depth and 

 depths of the upper bound of the maximum gradient and first gradient are 

 maximum in January or February and are minimum about 6 months later. The 

 minimum gradients occur in winter — usually a month or two after the 

 maximum depths. The gradients and their thicknesses are usually maximum 

 during August. 



The maximum standard deviations for surface temperature, as recorded 

 in appendix B, were compared month by month for each parameter. This compar- 

 ison showed that November surface temperatures yielded the largest standard 

 deviations, indicating that temperature changes were greatest during this 

 month. April had the next largest standard deviations ; however, these were 

 not nearly as large as those for November, indicating that spring warming 

 occurs at a slower rate than autumnal cooling. Months of minimum surface 

 temperature change are January, February, Aiigust, and September. 



Although the probability values given in tables 11 through 20 are 

 only approximate, the indicated trends probably are valid. A study of the 

 individual probabilities indicates considerable variation. This may be 



8 



