FIGURE 17 Power spectra of five 90-minute grovimeter records taken during 

 5 and 6 July 1961 — Spectra are corrected for instrument 

 response. 



record. This long-period buildup can easily be seen by examining the first 

 tvo records. The long-period, high amplitude vaves can be seen clearly on the 

 2237 to 0007Z record (figure 4a). Five ho\xrs later, between 03^5 and 0515Z, 

 this long-period energy has disappeared, and energy between 20- and 4o-second 

 periods has increased considerably. Later, between 08^5 and 1015Z and well 

 after the storm's peak, energy of about 25-second period has increased 2 to 3 

 times in magnitude; the long-period energy has essentially disappeared in this 

 record. Ten hours after the storm's peak, the overall energy has been substan- 

 tially reduced. 



(2) Through examination of energy with periods less than I6 seconds, 

 the direct dependency of wave energy on local surface windspeed is readily 

 observed. This dependence was previously reported by Sytinskiy and Tripol'nikov 

 (196^) and Hunkins (1962). Hunkins noted that there appeared to be some thresh- 

 old velocity above which the wind excited the ice and suggested a value between 

 20 and 2^ knots. It appears from figure 17 that such a threshold velocity in 

 this case is at or below 12 or 13 knots. ARLIS II, however, is a special case, 

 since it is considerably thicker than the pack on which the above two studies 

 were conducted. Thus, the short-period energy that appears so dependent upon 

 the wind is probably the natural bobbing of the ice island as computed 



31 



