A frequency of 2Tr(10)/96 would still be present a,t the poiirl oi' obser- 

 vation if it had been produced by a 19 knot wind 150 NM upwind of the 

 point of observation 6 hours prior to the time of observatio:;... Wtre tlie 

 fetch a little shorter, this would result in the low frequency cutoff and 

 the sharp steep forward fact of the observed spectrum. 



The spectrum which was observed could easily have resulted from a 

 combination of these effects, although the sparsity of oceanographic obser- 

 vations makes it difficult to demonstrate the exact disposition of the generat- 

 ing area which would lead to the observed spectrumo 



The figures given in Part 5 which describe the wind field were the 

 operational maps for the project. Additional ship reports for the area 

 were obtained by checking back through the data, and weather maps chewing 

 these reports are plotted in figures 10.3 a through 10.3 h. The wind ac :. e- 

 ported in knots is inserted in the feather of the arrow showing the v/ind 

 direction. 



From a study of the winds upwind of the ATLANTIS, it is possiMs Lo 

 conclude that the above argument is quite plausible and that therefoi e vari- 

 ations in wind speed from place to place explain the variation in the ob- 

 served spectrum. 



Comparison with the C. C. S. curves 



If the sum of the AE values for [J. = ZiTk/96 to 2tt 60/96 is computed, 



an estimate of the point on the co-cumulative spectrum curve for 



|JL = 2Tr(k -4)/96 is obtained. The estimates of the spectral curve can be 



141 



