PROGRAMS AND PLANS--Guidelines on Publishing Flood-Frequency Estimates

In Reply Refer To:                               July 10, 1989
WGS-Mail Stop 415


Subject:  PROGRAMS AND PLANS--Guidelines on Publishing Flood-
                              Frequency Estimates

This memorandum revises limitations on the publication and
distribution of estimated flood-frequency information.  In
particular, this memorandum revises some of the recommendations
given in Surface Water Branch (SWB) Technical Memorandum No.

Recommendations in SWB Technical Memorandum No. 77.17 relative to
the publishing of recurrence intervals for observed flood dis-
charges are not revised.  The purpose of this memorandum is to
revise guidelines for publishing flood-frequency estimates in
regional flood reports.  SWB Technical Memorandum No. 77.17

Discharges may be estimated and published for any recur-
rence interval up to 100 years.  For recurrence intervals
greater than 100 years, the discharge may be presented
only in response to a specific request.  For example,
only those discharges for recurrence intervals up to and
including 100 years would be estimated and published in a
regional flood-frequency report but discharges having
recurrence intervals of up to 500 years would be pre-
sented in a Flood Insurance Administration study.

This statement is now revised to permit estimation and publication
of estimated discharges having recurrence intervals up to and
including 500 years.  Both at-site estimates of the 500-year flood
and the regional regression equation for the 500-year flood may
now be published in a regional flood-frequency report.

The motivation or justification for this revision is as follows:

1.  Estimates of 500-year flood discharges are needed for
flood-insurance studies of the Federal Emergency
Management Agency (FEMA).

2.  Other Federal agencies, in addition to FEMA, need esti-
mates of flood discharges with recurrence intervals in
excess of 100 years for floodplain management purposes,
and the design of dams and bridges.  For example, the
Federal Highway Administration recommends the use of
500-year flood discharges in evaluating possible scour at

3.  At-site station frequency estimates are based on both
station and generalized skew and therefore should be more
accurate than estimates based on only station skew.

4.  The recommended at-site estimate of the 500-year flood (as
well as other T-year floods) is a weighted average of the
station estimate and the regional regression estimate (see
SWB Technical Memorandum Nos. 76.07 and 76.12).  This pro-
cedure improves the accuracy of the station 500-year
estimate by weighting with the regional estimate.  SWB
Technical Memorandum No. 76.07 recommends that both sta-
tion and weighted estimates be presented in a regional
flood report.

5.  A methodology has been developed for extrapolating the
regional regression equations to obtain estimates of the
500-year flood discharge for those states not having 500-
year equations.  These estimates will be provided in the
U.S. Geological Survey (USGS) National Flood Frequency
program which is under development (see Office of Surface
Water Technical Memorandum No. 88.14).

The uncertainty in the T-year flood discharges for the station and
regional and weighted estimates should be documented in the
regional flood report.  It is particularly important to recognize
the large uncertainty in estimates of extreme flood discharges,
such as the 500-year flood.  The methodology for defining the
standard error of station T-year flood discharges estimated from a
Pearson Type III distribution is described in USGS Professional
Paper 750-C, pages C228-C236, and in other publications.  The
standard error of the T-year regional regression estimate is
available from the regression analysis and should be reported as
well.  This standard error can be used to weight the regional T-
year estimate with the station T-year estimate.  Appendix 8 of
Bulletin 17B provides guidance on weighting independent estimates
and guidance for determining the standard error or variance of a
weighted estimate.

The publication of 500-year flood estimates and regression equa-
tions for estimating 500-year floods at ungaged sites will provide
the needed flood-frequency information for floodplain management
and design of major water-resources structures.  The necessary
constraint is that the analyst properly qualify the uncertainties
in the 500-year estimates so that data users are aware of the
large uncertainty associated with such an extreme flood.

                                       Charles W. Boning
                                       Chief, Office of Surface Water

The memorandum partially supercedes SWB Technical Memorandum
No. 77.17.

WRD Distribution:  A, B, S, FO, PO