EM 1110-2-1100 (Part II)
30 Apr 02
L = desired time period, in years
(3) Values of Pe expressed in percent for some typical coastal engineering concerns are given in
Percent Chance for Hs Equaling or Exceeding Return Period Hs
Desired Time Period (year)
g. Extrapolation of data
. The main objective
an appropriate extremal distribution
function is to get the best possible estimates of extreme conditions at desired return periods. Often the data
must be extrapolated to probabilities beyond the record length to match design return periods. Extrapolation
beyond 2-3 times the data record length should be avoided if possible. For example, 10 years of data should
be used for estimating return values of 20-30 years or less.
II-8-6. Analysis of Key Meteorological and Hydrodynamic Processes in Design
(1) Key processes in design are reviewed in this chapter. Applicable chapters of Part II are cited. The
design importance of each process is briefly stated. Critical design concerns are discussed, including two
(2) Statistical information about the processes is summarized in Table II-8-4. For each process (as
applicable and available), the table includes representative statistical distribution functions and general
expressions for parameters of the distribution function.
(3) It is important to bear in mind that the most extreme event of record may not merely be an intensified
version of lesser extreme events. Most experienced coastal and ocean engineers and scientists can remember
at least one catastrophic event that was distinctly different from typical storm events. Often the catastrophic
event arises from an unusual interaction between several major weather features. The "Halloween Storm"
that occurred in the northwestern Atlantic Ocean in October 1991 is a good example (U.S. Department of
Commerce 1992). Three significant meteorological systems, including a hurricane and an intense winter
storm, combined to create very strong winds over an extremely long fetch, which lasted for a period of days.
This type of event is difficult to anticipate, but it should be recognized that such things can occur. They may
appear as outliers in extreme data distributions.
Hydrodynamic Analysis and Design Conditions