method may be applied to obtain a 3-hour average; however, the assumption of 

 constant wind again may not be valid. 



If thunderstorms or other brief, severe winds are included in the data, 

 the method may overestimate results; but often there are no other data 

 available. 



c. Stability Correction . If the air-sea temperature difference 

 AT = T -T is zero, the boundary layer has neutral stability and no 



3,S 3. S J ^ 



windspeed correction is necessary. If AT is negative, the boundary layer 

 is unstable and the windspeed is more effective in causing wave growth. If 

 AT is positive, the boundary layer is stable and the windspeed is less 

 effective. The correction factor, R^ , is a function of AT and was 

 defined by Resio and Vincent (1977b) to account for this effect. An effective 

 windspeed is obtained by 



U = R^ U(10) (3-27) 



where R-r. is read from Figure 3-14. This correction can be substantial. For 

 example, if the winds are estimated for a AT of +10° C and AT is 

 actually -10° C, the error in U* is 50 percent. AT may vary season- 

 ally. In the fall a lake's water may still be warm, but cold winds may blow 

 across it; in the spring, the reverse may be true. Investigation of the 

 values of T is usually warranted, and the a priori assumption of a 

 neutrally staole boundary layer should be questioned. In the absence of 

 temperature information, R^ = 1.1 should be assumed. 



d. Location Effects . Often overwater wind data are not available, but 

 data from nearby land sites are. It is possible to translate overland winds 

 to overwater winds if they are the result of the same pressure gradient and 

 the only major difference is the surface roughness (Resio and Vincent, 

 1977b). For first-order airport weather stations around the Great Lakes, the 

 relationship between overwater winds to nearby overland winds is given for 

 neutral stability by R, in Figure 3-15; this can be used as an approximation 

 for other areas unless the landscape roughness characteristics are markedly 

 different. The land anemometer site should be close enough to the body of 

 water so that the winds are caused by the same atmospheric pressure 

 gradient. Thunderstorms and squall lines are small-scale phenomena and 

 violate the assumption that overland winds and overwater winds are from the 

 same pressure gradient. If the anemometer site is adjacent to shore, winds 

 blowing off the water require no adjustment for location effects; i.e., Rj^ = 

 1 . A stability adjustment Rq. should be used, however. 



e. Coefficient of Drag . The wave growth formulas and nomograms are 

 expressed in terms of wind-stress factor Ua (adjusted windspeed) . After the 

 appropriate windspeed conversions are made, the windspeed is converted to a 

 wind-stress factor by either of the following formulas: 



U^ = 0.71 U ^'^^ (U in m/s) (3-28a) 



U. = 0.589 U ^'^^ (U in mph) (3-28b) 



A 



3-30 



