APPENDIX 
METHOD FOR ESTIMATING PARAMETERS IN THE WEIBULL DISTRIBUTION FUNCTION 
The Weibull distribution function, equation (1), can be transformed into a 
form suitable for linear regression analysis. First, the natural logarithm of 
equation (1) is taken 
H.. - H __\% 
On Pe _ (is i sc si (A-1) 
Both sides of equation (A-1) are multiplied by -1, and again natural logarithms 
are taken 
on (- ’n F) = on (isc = Bsc nin) (A-2) 
Hc 
Equation (A-2) can be rewritten as 
fim (oe > Hye atin) = 8 Hse $5 Qn (- £n F) (A-3) 
Equation (A-3) is in the form 
Yeatp (A-4) 
where 
i ie in(Hs. = lee ‘ita 
a = &n Hee 
ie! 
X = Mn (© Qn F). 
An initial value of the parameter Hgce min was obtained from Table 1 of 
Thompson and Harris (1972) as the "minimum significant height" divided by the 
observed mean significant height. The value for Nags Head was 0.31. Alterna- 
tively, Hsc min could be estimated initially as 0.38 from equation 4-8 in the 
SPM. The estimated value of Hge min and empirical tabulation of F asa 
function of Hg, are used to compute a table of X and Y values. Linear 
regression analysis is then used to estimate optimum values of a and b in 
4THOMPSON, E.F., and HARRIS, D.L., "A Wave Climatology for U.S. Coastal 
Waters," Proceedings of Offshore Technology Conference, May 1972, pp. 675-688 
(also Reprint 1-72, U.S. Army, Corps of Engineers, Coastal Engineering Research 
Center, Fort Belvoir, Va., NTIS AD 746 365). 
15 
