in the low frequency bands is high. On the other hand, if the energy 

 in the high frequency bands is large, the rapid energy decay with 

 depth will allow the lower frequency bands to dominate with depth. 

 In particular, if the lower frequency energy is in a different direction, 

 the resultant will swing towards this direction with depth. In a situ- 

 ation where swell is running against a local wind sea at the surface, 

 an indication of resultant direction reversal can be seen with in- 

 creasing depth. Actually, these are simply consequences of the fact 

 that long waves (low frequency) reach greater depths than short 

 waves (high frequency). 



SURFACE DRIFT RESULTS 



Surface fields of mean Stokes' velocity have been computed for 

 both 12Z, 27 February 1967 and 18Z, 4 March 1967. Figures 2 and 

 3 show the wind field at 19.5 meters above the sea surface and sur- 

 face mass transport (or drift) velocity, respectively, for 4 March 

 1967. The 519 arrows indicate the direction of flow at each grid 

 point and the isotach analysis indicates the speed" distribution. The 

 arrows on these figures show a general tendency of the current to 

 flow in the wind direction. The current flow is mainly cyclonic 

 under the intense low in the north (centered at approximately 

 62°N, 22°W) and roughly anticyclonic under the subtropical high 

 situated further south (centered at approximately 35°N, 30°W). The 

 isotach analyses clearly indicate the tendency for strong drift velo- 

 cities to be situated in the vicinity of strong winds and weak or vari- 

 able drift velocities to be situated near lesser winds. 



In order to analyze these results, two conventional oceanographic 

 statistics were employed. The first was a drift-speed to wind-speed 

 ratio expressed as a percentage. The second was an angular measure 

 of drift direction relative to the wind direction. If the current flowed 

 to the right of the wind, the angle between the two vectors was taken 

 as positive and if the current flowed to the left, the angle was taken 

 as negative. 



The amount of information derived from the first statistic proved 

 to be only of limited significance. The reason for this, however, 

 proved to be of considerable significance. If the ratio statistic were 

 completely significant, this would imply a direct causal link between 

 wind and waves. This, of course, is not true since there is a lag 

 between wind changes and sea surface response. The isotach analyses 

 in Figures 2 and 3 give an indication of the lag. A good example of 

 this appears at grid point 276 (35.2°N, 25.2°W) in the 27 February 

 analysis. In the hours prior to 12Z, brisk northerly winds raised 

 appreciable energy in the high frequency bands. This, in turn, 

 produced an appreciable Stokes' velocity. By 12Z, the wind had 

 shifted to easterly and had decreased to about 1 knot. The spectrum, 

 however, did not respond as rapidly and could still maintain a drift 

 magnitude of 6.2 cm/sec. The computed ratio was then 12.22% , 

 which is extremely high. Empirical studies usually set this ratio at 



402 



1 



