II. WEIBULL DISTRIBUTION FUNCTION 
A general model which can be used to approximate the empirical distribution 
of significant wave height is 
a o 
" st sc Heo nin] 
> = 
5 FB = He S Hse } ie) 
where 
Heya = significant wave height divided by mean significant height 
Hee SespeetitedsvallucvorssHor 
lige > Hla = probability that H,. is greater than or equal to a speci- 
fied ratio eS 
floes ain = minimum expected value of H,. 
Heat, a = other parameters in distribution function. 
Equation (1) is a form of the Weibull distribution function with three param- 
eters (H H eG ©) 
sc min® “sc? 
IIL. COMPILATION OF EMPIRICAL DISTRIBUTION 
The parameters in equation (1) must be evaluated for each site by opti- 
mizing the agreement between equation (1) and the empirical, dimensionless 
distribution of significant height at the site using the following procedure: 
(a) Assemble all significant heights obtained by reliable, consist- 
ent analysis methods at a particular site; 
(b) delete significant heights from months in which more than 50 
percent of the possible observations are missing; 
(c) compute mean significant height for each remaining month; 
(d) divide each significant height by the appropriate monthly mean 
Significant height; and 
(e) combine the dimensionless significant heights in step (d) from 
all months into one distribution. 
The implicit assumption in step (e) is that monthly variations in wave condi- 
tions can be completely represented by variations in monthly mean significant 
height but that variations relative to the mean are consistent from month to 
month. The assumption may not be valid at sites strongly affected by hurri- 
cane waves unless hurricane waves are treated separately. 
