EM 1110-2-1100 (Part II)
30 Apr 02
Ur ' Um % 0.78 σm [ ln (12Tr) & 0.577 ]
(II-8-3)
where
Ur = wind speed with r-year return period
Um = mean value of maximum monthly wind speeds
σm = standard deviation of maximum monthly wind speeds
and
σm
0.49 % 0.89 ln (12Um) % 0.67 [ln (12Um) & 0.577]2
σrm '
(II-8-4)
Nm
where
σrm = standard deviation of the sampling error in estimating Ur
Nm = number of months of data
(b) The parameter σrm can be related to confidence intervals using Table II-8-5. An integral number of
years of data covering at least 3 years (36 months) is needed for this simple approach. Only populations of
extreme events that are well-represented in the data can be effectively included in the long-term extreme
estimates (e.g., if hurricane events do not appear or are sparsely represented in the data sample, they will not
be effectively represented in the extreme estimates).
Table II-8-5
Standard Deviation and Confidence Interval Relationships
Confidence Level, %
Confidence Interval Bounds Around Ur
80
1.28σrm
10.0
85
1.44σrm
7.5
5.0
90
1.65σrm
95
1.96σrm
2.5
99
2.58σrm
0.5
(c) If data are unavailable, extreme wind speed can be estimated for various return periods with
Figures II-8-7 through II-8-9. These figures were developed to estimate maximum wind loads for building
design and are expected to be conservative for coastal engineering applications. Figure II-8-7 gives extreme
fastest mile wind velocity data with a 50-year return period (annual probability of 0.02 that the wind speed
is exceeded). Wind speed information was prepared from data collected at 129 weather stations (Simiu,
Changery, and Filliben 1979), representing a 10-m elevation. Data were statistically reduced using extreme
value analysis based on Fisher-Tippett Type-I distributions. Wind speed contours of Alaska are based
primarily on data collected in open areas (Thom 1968-69). Wind speed contours in the hurricane-prone
region, Atlantic and Gulf of Mexico coastlines, are based on Monte Carlo simulations of hurricane storms
striking the coastal region (Batts, Russell, and Simiu 1980). Recurrence intervals for 25 and 100 years may
be estimated by multiplying wind speed from Figure II-8-7 with the appropriate adjustment factor in
Table II-8-6.
II-8-16
Hydrodynamic Analysis and Design Conditions