You are on page 1of 8

PEAK OUTFLOW FROM

BREACHED EMBANKMENT DAM

By David C. Froehlich; Member, ASCE

INTRODUCTION
Hazards created by the flood resulting from a sudden, rapid, and uncontrolled release of
water through a breach that forms in a dam need to be assessed to provide adequate safety
measures in the event of such a catastrophic failure. The level of detail of hydrologic and hydraulic
analyses needed to evaluate the consequences of a dam-breach flood depends on the danger to
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

human life and the amount of property damage that would occur. If human fatalities are unlikely
and if property damage potential is small, a simple procedure might provide an adequate de-
scription of the extent and timing of downstream flooding resulting from a dam failure.
Several simple flood-routing methods requiring a small amount of data and effort have been
developed for estimating peak flows at locations downstream of a breached dam (Sakkas 1980;
Hagen 1982; Costa 1985; "Simplified" 1981). Each of these methods requires an estimate of
the peak outflow from the reservoir. Empirical formulas for estimating the peak outflow caused
by a gradual dam failure are presented by Kirkpatrick (1977), Hagen (1982), MacDonald and
Langridge-Monopolis (1984), Costa (1985), "Simplified" (1981), "Earth" (1985), "Guidelines"
(1986), and Evans (1986), while analytical formulas based on an instantaneous dam failure are
described by Cecilio and Strassburger (1974) and Price et al. (1977). Scarcity of data has resulted
in the use of flow rates that are rough estimates of the peak outflows through breached dams.
Some of the peak flows used to develop the empirical relations were measured a substantial
distance downstream of a breached dam and might be significantly less than the peak outflow
from the reservoir. Other peak flow rates were obtained from numerical simulation of a dam
breach and not measured values at all. In the case of simulated outflows, the computed peak
flow depends on the breach-formation model used to simulate the gradual failure of the dam
and assumptions regarding tailwater conditions at the breach. Considering the difficulty of
estimating parameters that define a breach (MacDonald and Langridge-Monopolis 1984), the
uncertainty of a simulated peak outflow is necessarily high.
Data from 22 embankment-dam failures were assembled from numerous published and un-
published sources. These data are used to evaluate and compare several existing empirical
equations that predict peak outflow from a breached dam. Multiple regression analysis is then
used to obtain a new empirical expression for rapidly estimating peak outflow from a breached
embankment dam. The new prediction equation uses easily obtained information and also
provides a means of computing prediction limits from which appropriate factors of safety to be
used in an evaluation of the flood-hazard potential of a dam failure can be readily determined.

EMBANKMENT-DAM FAILURE DATA


Peak outflows and other pertinent data for 22 embankment-dam failures were obtained from
a variety of sources and are summarized in Tables 1 and 2. These data include a description of
the embankment, the failure mode, reservoir characteristics at the time of failure, and the
measured peak outflow.

Embankment Description
Embankment dams are by far the most common type of dam that is constructed. The United
States Committee on Large Dams (USCOLD) estimates that 79% of all major dams in operation
in the United States are embankment dams (Lessons 1975). Embankment dams are usually
constructed of natural materials obtained from borrows or quarries, or from waste materials
obtained from mining and milling operations. Earthfill and rockfill embankments are the most
common and are classified on the basis of the predominant composition of embankment material.
Earthfill dams are composed mainly of compacted fine-grained material, and rockfill dams are
composed mostly of compacted or dumped pervious material or crushed rock. A characteristic
of an embankment that might affect the rate of breach formation and thus the peak outflow

'Ass!. Prof.. Univ. of Kentucky, Dep!. of Civ. Engrg., 212 Anderson Bldg., Lexington. Ky 40506-0046.
Note. Discussion open until July I, 1995. To extend the closing date one month. a written request must be
filed with the ASCE Manager of Journals. The manuscript for this technical note was submitted for review and
possible publication on March 23. 1993. This technical note is part of the Journal of Water Resources Planning
and Management, Vol. 121. No. I. January/February. 1995.©ASCE. ISSN 0733-9496/95/0001-0090-0097/$2.00
+ $.25 per page. Technical Note No. 5766.

90 JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT

J. Water Resour. Plann. Manage. 1995.121:90-97.


TABLE 1. Description of Failed Embankment Dams and References

Dam name and Year Year


location completed failed Embankment type References
(1 ) (2) (3) (4) (5)
Apishapa, Colo. 1920 1923 Homogeneous earthfill "Failure" (1923); Field et al. (1923)
Baldwin Hills, Calif. 1951 1963 Homogeneous earthfill "Investigation" (1964); Jessup (1964);
Warne (1964)
Butler, Ariz. - " 1982 Homogeneous earthfill Aldridge (personal communication,
1987)
Castlewood, Colo. 1890 1933 Stonewall-earth "Castlewood" (1933); Field (1933);
Follansbee and Sawyer (1948); Houk
(1933)
Fred Burr, Mont. - 1948 Homogeneous earthfill Ulrick (personal communication. 1992)
French Landing, Mich. 1925 1925 Homogeneous earthfill "Undermining" (1925)
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

Frenchman Creek, 1952 1952 Homogeneous earthfill Babb and Mermel (1963); Oltman
Mont. (1955); Ulrick (personal communi-
cation, 1992)
Hatchtown, Utah 1908 1914 Zoned earthfill "Ninth" (1915); Sterling (1914); Ster-
ling (1915)
Hell Hole, Calif. 1964 1964 Rockfill "Hell Hole" (1965); Scott and Grav-
lee (1968); "Failure" (1965)
Ireland No.5. Colo. - " 1984 Homogeneous earthfill DeGrave (personal communication,
1986)
Kelly Barnes, Ga. 1948 1977 Homogeneous earthfill Crisp et al. (1977); "Dam" (1977);
Sanders and Sauer (1979)
Laurel Run, Penn.. - " 1977 Homogeneous earthfill Chen and Armbruster (1980); "Johns-
town" (1977); Hoxit et al. (1982)
Lily Lake, Colo. 1913 1951 Homogeneous earthfill Graham (personal communication.
1987)
Little Deer Creek, 1962 1963 Homogeneous earthfill Collins (1964); Rostvedt (1968)
Utah
Lower Latham, Colo. - " 1973 Homogeneous earthfill DeGrave (personal communication.
1986)
Lower Two Medicine, - " 1964 Homogeneous earthfill Boner and Stermitz (1967); "Guide-
Mont. lines" (1986), "Schedule" (1905)
Oros, Brazil 1960 1960" Zoned earthfill and rockfill Lessons (1974); Jansen (1983)
Prospect, Colo. 1914 1980 Homogeneous earthfill DeGrave (personal communication,
1986)
Puddingstone, Calif. 1926 1926 Homogeneous earthfill "Flood" (1926); Lessons (1974)
Quail Creek, Utah 1986 1989 Homogeneous earthfill "Investigation" (1989); Hall (personal
communication, 1989)
South Fork, Penn. 1853 1889 Homogeneous earthfill Francis et al. (1891); Jansen (1983);
McCullough (1987)
Teton, Idaho 1975 1976 Zoned earthfill Brown and Rogers (1977); Jansen
( 1983)
"Information not available.
"Dam failed during construction.

rate is the average width of the embankment from the bottom of the final breach to the top of
the dam. Average embankment width was available for most ofthe dams studies and is presented
in Table 2.

Failure Mode
There are several causes of embankment-dam failure and in some instances it is difficult to
determine the exact failure mode, especially if there are no eyewitness accounts of the failure.
The International Commission on Large Dams (Lessons 1974) reports that about one-third of
all embankment-dam failures are caused by inadequate spillway capacity that result in overtop-
ping of the dam, Roughly another third of embankment-dam failures have been attributed to
piping caused by concentrated seepage that erodes soil particles along the path of the leakage,
gradually enlarging the flow passage until a failure occurs, The remaining third of the failures
are caused by sliding of the embankment, settlement of the foundation, or inadequate protection
against wave action. On the basis of this information, failure modes presented in Table 2 are
classified as either overtopping; piping or seepage; or foundation defects.

Reservoir Characteristics
Easily measured reservoir characteristics that influence peak outflow from a breached dam
include the volume of water contained in the reservoir at the start of breach formation and the
height of water in the reservoir at the start of breach formation, both quantities being measured
from the elevation of the final breach bottom. Inflows to a reservoir during failure might also

JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT 91

J. Water Resour. Plann. Manage. 1995.121:90-97.


TABLE 2. Embankment and Reservoir Characteristics, and Measured Peak Outflows

Volume
of water Height
Average above of water
width of breach above Height Peak
embank- bottom Vw breach of dam outflow
Dam name and Failure ment W (millions bottom H w breach H rate Q p Method used to calculate
location mode" (m) of m3 ) (m) (m) (m 3 /s) peak outflow rate
(1) (2) (3) (4) (5) (6) (7) (8)
Apishapa, Colo. P 82.4 22.2 28.0 31.1 6,850 Reservoir volume change
(IS min period)
Baldwin Hills, P 59.6 0.910 12.2 21.3 1,130 Reservoir volume change
Calif. (IS min period)
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

Butler, Ariz. 0 9.63 2.38 7.16 7.16 810 Slope-area measurement


600 m downstream of
dam
Castlewood, 0 47.4 6.17 21.6 21.3 3,570 Reservoir volume change
Colo. (IS min period)
Fred Burr, P 30.8 0.750 10.2 10.4 654 Slope-area measurement
Mont. "'a short distance down-
stream"' of dam
French Landing, P 34.3 3.87 8.53 14.2 929 Reservoir volume change
Mich. (I h period)
Frenchman P 37.3 16.0 10.8 12.5 1.420 - "
Creek, Mont.
Hatchtown, P or F 44.8 14.8 16.8 18.3 3,080 Reservoir volume change
Utah (I h period)
Hell Hole, Calif. P 103 30.6 35.1 56.4 7.360 Reservoir volume change
(I h period)
Ireland No.5, r 18.0 0.160 3.81 5.18 110 Slope-area measurement
Colo. "a short distance"
downstream of dam
Kelly Barnes, P 19.4 0.777 11.3 12.8 680 Slope-area measurement
Ga. 250 m downstream of
dam
Laurel Run, 0 40.6 0.555 14.1 13.7 1,050 Slope-area measurement
Penn. 1.6 km downstream of
dam
Lily Lake, Colo. r - h
0.0925 3.35 3.66 71 Slope-area measurement
at unknown location
Little Deer P 63.1 1.36 22.9 27.1 1,330 Slope-area measurement
Creek, Utah at unknown location
Lower Latham. P 25.7 7.08 5.79 7.01 340 Slope-area measurement
Colo. at unknown location
Lower Two r - h
29.6 11.3 11.3 I,SOO Slope-area measurement 4
Medicine, km downstream of dam
Mont.
Oros, Brazil 0 110 660 35.8 35.5 9,630 - "
Prospect, Colo. P 13.1 3.54 1.68 4.42 116 Reservoir volume change
(unknown time period)
Puddingstone, 0 - h
0.617 15.2< 15.2 4S0 Reservoir volume change
Calif. (15 min period)
Quail Creek, P 56.6 30.8 16.7 21.3 3,110 Reservoir volume change
Utah (IS min period)
South Fork, 0 64.0 18.9 24.6 24.4 8,500 Reservoir volume change
Penn. (30 min period)
Teton, Idaho P 250 310 77.4 86.9 65,120 Slope-area measurement 4
km downstream of dam
"Failure mode abbreviations are as follows: 0 = overtopping; r = piping or seepage; and F = foundation
defects.
"Information not available.
<Height of water was assumed to equal height of dam breach but might have been greater.

affect the peak outflow, especially during large floods that cause the dam to be overtopped.
However, difficulty of estimating reservoir inflow hydrographs for the reported dam failures
precludes an evaluation of inflow effects.

Measured Peak Outflows


Reported peak outflows at each of the failed dams are determined from either stage recordings
of reservoir levels or by slope-area measurements. Reservoir levels are used to determine the
reservoir volume change during a short time period from which an average outflow rate is
computed. If the time period used to estimate the average outflow is long relative to the time

92 JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT

J. Water Resour. Plann. Manage. 1995.121:90-97.


needed for the reservoir to drain, the computed outflow might be significantly less than the
instantaneous peak outflow. Slope-area measurements are made at a channel location a short
distance downstream of a dam and rely on measured cross-sectional geometry, water-surface
slopes, and estimates of roughness coefficients to calculate the peak flow rate using Manning's
equation (Dalrymple and Benson 1984). The method used to estimate the reported peak outflow
is known for al1 but one dam (Oros).

OTHER FORMULAS
Some empirical formulas for estimating the peak outflow caused by a gradual dam failure are
evaluated and compared using the assembled data. The evaluation identifies possible deficiencies
and the comparison points out differences between methods.
The Soil Conservation Service (SCS) provides the fol1owing procedure for estimating the peak
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

outflow from a breached earth dam ("Earth" 1985). For H", 2: 31.4 m, where H", = the reservoir
depth at the dam at the time of failure measured from the bottom of the final breach to the
reservoir surface, peak outflow through the breach, in m'/s is computed as
Q" = 16.6H~"5 (1)
For H < 31.4 m, peak outflow is computed as
Q" = O.000421B~35 (2)
where B r is a breach factor defined as

B = V",H".
r A (3)

and V", = reservoir storage volume at the time of failure (m'); and A cross-sectional area
of the embankment at the location of the breach, m 2 . Cross-sectional area of the embankment
can be computed as A = W x H, where W = average width of the embankment from the
bottom of the final breach to the top of the embankment; and H = distance from the bottom
of the final breach to the top of the embankment. Therefore, peak outflow for H", < 3\.4 m,
is given by

Q = 0.000421 [V~H,,] us (4)


" WH
However, the peak outflow given by (4) is not to exceed the value given by (1) nor be less than
Q" = 1.77H~5 (5)
Peak outflows computed using the SCS procedure for the 22 historical embankment-dam failures
are compared to the measured peak outflows in Fig. 1. The method seems to provide good
estimates for al1 except the smal1est of measured peak outflows. However, the SCS procedure
does not provide a means of determining a peak outflow that provides a factor of safety when
evaluating downstream flooding.
Several studies have used the product of reservoir volume and depth at the time of failure

.0
:;=2
00
....I U
II. UJ
I-Ul 10000
::la:
°UJ
:ll:11.
<Ul
UJa:
I1.UJ
01-
1000 •
UJUJ
a::I
::lu
Ul-
<ID
UJ::l Line of perfect agreement
:lU
~

1000 10000 100000

COMPUTED PEAK OUTFLOW,


IN CUBIC METERS PER SECOND
FIG. 1. Comparison of Measured Peak Outflows from Embankment-Dam Failures and Values Computed
Using SCS Equation ("Earth" 1985)

JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT 93

J. Water Resour. Plann. Manage. 1995.121:90-97.


. 0 •
3: z
00
...1 0
u. W
1-1/)
:::la:
Ow
:o::D.
<I/)
wa:
D.w
01-
WW
a:::i!
:::lo
1/)-
<Ill
w:::l
::i!0
~
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

DAM FACTOR

FIG. 2. Comparison of Measured Peak Outflows from Embankment-Dam Failures and Values Given by
Equations Presented by Costa (1985) and MacDonald and Langridge-Monopolls (1984) Based on Dam
Factory (VW x H w )

to predict peak outflow from a breached dam. Costa (1985) refers to this produce (V", x H",)
as the "dam factor," which is described as a crude index of the energy expenditure at the dam
when it fails. On the basis of measured peak outflows from failed constructed dams, which
include both embankment dams and concrete dams, Costa (1985) develops the folJowing equation
to predict peak outflow from a breached dam:
Q" = 0.763(V"H w )OA2 (6)
Peak outflows from failed concrete dams are generalJy larger than peak outflows from failed
embankment dams having similar reservoir volumes because of faster enlargement of the breach
during failure. Hence, (6) likely overpredicts peak outflows from embankment dams. MacDonald
and Langridge-Monopolis (1984) present a graphical relation for peak outflow from breached
dams as a function of the dam factor, which is approximated by the following equation:
Q" = 1. 175(V"H,,)O 41 (7)
Eq. (7) is also based on peak outflows from both failed concrete and embankment dams and
thus might overpredict peak outflows from embankment dams. The peak-outflow prediction
relations presented by Costa (1985) and MacDonald and Langridge-Monopolis (1984) are dis-
played in Fig. 2 along with the assembled measured peak outflows. The equations give somewhat
different solutions for the range of dam-factor shown. In addition, both equations tend to
overpredict the smalJer peak outflows from the assembled failures.

ANALYSIS OF DATA
Multiple linear regression analysis was used to develop an equation for predicting peak outflow
from a breached embankment dam. Logarithmic transformation of alJ variables was found to
provide the best linear relation as follows:
In QI' = -0.499 + 0.295 In V" + 1.24 In H" (8)
where QI' = predicted peak outflow, (mJ/s); V" = reservoir volume at the time of failure (m J);
and H" = height of water in the reservoir at the time of failure above the final bottom elevation
of the breach (m). Failure mode and average embankment width did not improve the regression
equation significantly. The coefficient of determination of (8) is 0.934, and the standard error
of the predicted natural logarithm of QI' is 0.4198. Taking the exponential of each side of (8)
gives
(9)
Measured values of Qp are compared to values computed using (9) in Fig. 3. The equation
provides good agreement between measured and computed peak outflows over the entire range
of values used in the analysis.

Hidden Extrapolation
A quantitative way of detecting an instance where the prediction equation would provide an
extrapolation beyond the range of data used in the regression analysis, that is, a hidden ex-
trapolation. compares the quantity

94 JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT

J. Water Resour. Plann. Manage. 1995.121:90-97.


100000
• 0
Z

3:
00
..J(')
u. W
I-lII
;:)a:
10000

Ow
~l1.
c(lII
Wa:
l1.W 1000
01-
WW
a::::E
;:)(,)
lII-
••
c(1II
W;:) 100 Line of perfect agreement
:::E(,)
~
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

10 ~~"""",~"""'""",,- .........~"---~~~oL-~~""""
10 100 1000 10000 100000

COMPUTED PEAK OUTFLOW,


IN CUBIC METERS PER SECOND

FIG. 3. Comparison of Measured Peak Outflows from Embankment-Dam Failures and Values Given by
Regression Equation Based on Measured Values

h"" = x;,(X'X) 'x" (10)


to the largest diagonal element of the "hat matrix" H = X(X'X) - IX' denoted as h n,,'" where
Xo = vector of regressor variables; X = (n x p) matrix of (p - 1) independent variables
augmented by a column of ones, X' = the transpose of X; n = number of observed values;
and p = number of regression model parameters (Montgomery and Peck 1982). If h oo exceeds
h Illax the prediction is certainly an extrapolation and needs to be carefully considered before
acceptance. If h ,K , "S h llla " the prediction might still be an extrapolation; however, the smaller
the value of hoo, the closer the point X o lies to the centroid of the variable space covered by the
data and the less likely it is an extrapolated value. For the regression model given in (1), h lllax
= 0.4630.

Prediction Limits
A 100 x (1 - (X) percent prediction interval for a future estimate of logarithm of peak outflow
is given by
x;,(3 - ("npVa~(1 + x;,(X'X) IX,,] :<:: In Qp :<:: x;,(3 + (".n pVa~[1 + x;,(X'X) IX,,] (11)
where Xo = {I, In V w , In HJ is the transpose of the (p x 1) vector of predictor variables; 13'
== [0.597, 0.305, 1.22] is the transpose of the (p x 1) vector of estimated regression coefficients;
(".n -" = the one-sided (-distribution statistic corresponding to an exceedance probability (X and
n - p degrees of freedom; a == 0.4899 is the standard error of estimate of the regression model,
and

0.577627 0.043853 0.236078]


(X'X)-I = 0.043853 0.016231 0.027455
[
0.236078 0.027455 0.109801
is the inverse of the X'X matrix.

EXAMPLE APPLICATION

The peak outflow from a hypothetical dam failure is computed to illustrate application of the
prediction equation. A lO-m high embankment dam is postulated to fail when the embankment
is overtopped by a depth of 0.5 m, at which level the reservoir contains a volume of 500,000
m 3 of water. The breach is assumed to erode downward through the entire height of the em-
bankment and terminate at the base of the embankment. With V w = 500,000 m 3 and H w =
10.5 m, (8) yields an expected peak outflow Q" = 538 m 3 /s. The transpose of the vector of
predictor variables in this case is X o == {I, 13.1224, 2.3026}, giving h(K) = 0.1110. Since h lHl «
h lllax = 0.4630, the prediction is likely not an extrapolated value. Upper and lower prediction
limits for several exceedance probabilities given by (11) are presented in Table 3 and provide
a measure of reliability of the peak outflow estimate. Because the peak outflow is only an
estimate, a larger value might be used to provide a factor of safety in a flood-hazard analysis.
Choosing a 0.10 exceedance probability, an upper prediction estimate of 1,068 m 3 /s is computed.

JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT 95

J. Water Resour. Plann. Manage. 1995.121:90-97.


TABLE 3. Lower and Upper Peak Outflow Prediction Limits for Several Exceedance Probabilities
Peak Outflow Rate
(m 3 /s)
Exceedance probability t-distribution statistica Lower limit Upper limit
(1 ) (2) (3) (4)
0.0005 3.883 72 3,994
0.005 2.861 123 2,357
0.010 2.539 145 1,996
0.025 2.093 183 1,585
0.050 1.729 220 1,313
0.10 1.328 271 1,068
0.25 0.688 377 767
OAO 0.257 471 614
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

0.50 0.000 538 538


"Standard normal deviate for 19 degrees of freedom (i.e., the number of degrees of freedom corresponding
to the regression model error).

SUMMARY AND CONCLUSIONS


If human fatalities are unlikely and if property damage potential is small, procedures requiring
a small amount of data and computational effort can provide an adequate description of the
extent and timing of downstream flooding resulting from a dam failure. A simple empirical
method for quickly estimating the peak outflow from a breached embankment dam is presented.
The procedure is based on a regression analysis of measured peak outflows from 22 embankment
dam failures and provides good agreement between measured and computed peak outflows over
the entire range of values used in the analysis. Upper and lower prediction limits given by the
analysis provide a means of establishing a factor of safety needed in making a rapid assessment
of flood hazards created by a dam failure.

APPENDIX I. REFERENCES
Babb, A. 0., and Mermel, T. W. (1963). Catalog of dam disasters, failures and accidents. U.S. Department of
the Interior, Bureau of Reclamation, Denver, Colo.
Boner, F. c., and Stermitz, F. (1967). "Floods of June 1964 in Northwestern Montana." US. Geological Survey
Water-Supply Paper 1840-B, U.S. Government Printing Office, Washington, D.C.
Brown, R. J., and Rogers, D. C. (1977). "A simulation of the hydraulic events during and following the Teton
Dam failure." Proc., Dam-Break Flood Routing Model Workshop, Bethesda, Maryland, U.S. Water Resources
Council, Hydrology Committee, 131-163.
"Castlewood Dam failure floods Denver." (1933). Engrg. News-Record, 101(32), 174-176.
Cecilio, C. B., and Strassburger, A. G. (1977). "Downstream hydrograph from dam failures." Proc., Engrg.
Found. ConI, The Evaluation of Dam Safety, Pacific Grove, Calif.. 437 -460.
Chen, C. L., and Armbruster, J. T. (1980). "Dam-break wave model-formulation and verification." 1. Hydr.
Div., ASCE, 106(5),747-767.
Collins, W. E. (1964). "Report on the cause of failure of the Little Deer Creek Dam, Provo River Project, Utah."
U.S. Department of the Interior, Bureau of Reclamation., Denver, Colo.
Costa, J. E. (1985). "Floods from dam failures." US. Geological Survey Open-File Rep. 85-560, Denver, Colo.
Crisp, R. L., Fox, W. E., Robinson, R. c., and Sauer, V. B. (1977). "Report of failure of Kelly Barnes Dan,
Toccoa, Georgia." Federal1nvestigative Board Rep., U.S. Army Corps of Engineers, Atlanta, Ga.
Dalrymple, Tate, and Benson, M. A. (1984). "Measurement of peak discharge by the slope-area method."
Techniques of water-resources investigations of the United States geological survey, Book 3, chapter A2, U.S.
Government Printing Office, Washington, D.C.
"Dam collapses in southeast spur federal inspection plan." (1977). Engrg. News Record, 199(19), 13.
"Earth dams and reservoirs." (1985). US. Department of Agriculture Tech. Release No. 60, U.S. Department of
Agriculture, Soil Conservation Service, Washington, D.C.
Evans, S. G. (1986). 'The maximum discharge of outburst floods caused by the breaching of man-made and
natural dams." Can. Geotech. J., 23(3), 385-387.
"Failure of Apishapa earth dam in southern Colorado." (1923). Engrg. News-Record, 91(9), 357-358.
"The failure of Hell Hole Dam." (1965). Western Canst., 40(4), 65-70.
Field, J. E. (1933). "Data on Castlewood Dam failure and flood." Engrg. News-Record, 101(36),279-280.
Field, J. E., Jones, B. M., and Floyd, O. N. (1923). "Failure of Apishapa earth dam in Colorado-II." Engrg.
News-Record, 91(11), 418-424.
"Flood damage to Puddingstone earth-fill dam." (1926). Engrg. News-Record, 96(22), 913.
Follansbee, R., and Sawyer, L. R. (1948). "Floods in Colorado." U.S. Geological Survey Water-Supply Paper
997, U.S. Government Printing Office, Washington, D.C.
Francis, J. B., Worthen, W. E., Becker, M. J., and Fteley, A. (1891). "Report of the committee on the cause
of the failure of the South Fork Dam-with discussion" Trans., ASCE, 24, 432-469.
"Guidelines to decision analysis" (1986). ACER Tech. Memo. No.7, U.S. Department of the Interior, Bureau
of Reclamation, Denver, Colo.
Hagen, V. K. (1982). "Re-evaluation of design floods and dam safety." Proc., 14th Int. Commissioll Oil Large
Dams ConI, Rio de Janeiro, Brazil.
"Hell Hole Dam isn't a complete washout." (1965). Engrg. News-Record, 174(10),28-29.
Houk, I. E. (1933). "Failure of Castlewood rock-fill dam." Western Constr. News and Hwy. Builders, 8(16), 373-
375.

96 JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT

J. Water Resour. Plann. Manage. 1995.121:90-97.


Hoxit, L. R., Maddox, R. A., Chappel, C. F., and Brua, S. A. (1982). "Johnstown-western Pennsylvania storm
and floods of July 19-20, 1977." U.S. Geological Survey Profl. Paper 12/l, U.S. Government Printing Office,
Washington, D.C.
"Investigation of the case of Quail Creek dike failure." (1989). Rep. of Independent Review Team, prepared for
governor of Utah, Salt Lake City, Utah.
"Investigation of failure of Baldwain Hills Reservoir." (1964). California Department of Water Resources, Sac-
ramento, Calif.
Jansen, R. B. (1983). Dams and public safety. u.s. Department of the Interior, Bureau of Reclamation, Denver.
Colo., 166-167, 184-213.
Jessup, W. E. (1964). "Baldwin Hills Dam failure." Civ. Engrg. 34(2),62-64.
"Johnstown is inundated again by a record 500-year flood." (1977). Engrg. News-Record, 199(4),9.
Justin, J. D. (1932). Earth dam projects, 2nd Ed., John Wiley and Sons, New York, N.Y.
Kirkpatrick, G. W. (1977). "Guidelines for evaluating spillway capacity." Water Power & Dam Constr., 29(8),
29-33.
Lessons from dam incidents. (1974). Complete Ed., International Commission on Large Dams, Paris, France,
Downloaded from ascelibrary.org by GEORGE MASON UNIVERSITY on 06/22/13. Copyright ASCE. For personal use only; all rights reserved.

68-70, 510-512, 867-869.


Lessons from dam incidents, USA. (1975). U.S. Committee on Large Dams, ASCE, New York, N.Y.
MacDonald, T. c., and Langridge-Monopolis, J. (1984). "Breaching characteristics of dam failures." 1. Hydr.
Engrg., ASCE, 110(5), 567-586.
Montgomery, D. c., and Peck, E. A. (1982). Introduction to linear regression analysis, John Wiley and Sons,
New York, N.Y., 143.
McCullough, D. (1987). The Johnstown flood, 2nd Ed., Simon and Schuster, New York, N.Y.
"Ninth biennial report of the state engineer to the governor of Utah-1913-1914." (1915). The Arrow Press,
Salt Lake City, Utah, 13-14.
Oltman, R. E. (1955). "Floods of April 1952 in the Missouri River basin." U.S. Geological Survey Water-Supply
Paper 1260-B, U.S. Government Printing Office, Washington, D.C.
Price, J. T., Lowe, G. W., and Garrison, J. M. (1977). "Unsteady flow modeling of dam-break waves." Proc..
U.S. Water Resources Council. Hydro. Committee Dam-Break Flood Routing Model Workshop, National Tech-
nical Information Service, Springfield, Va., PB-275-437.
Rostvedt, J. O. (1968). "Summary of floods in the United States during 1963." U.S. Geological Water-Supply
Paper 1830-B, U.S. Government Printing Office, Washington, D.C. B84-B87.
Sakkas, J. G. (1980). "Dimensionless graphs of floods from ruptured dams." Res. Note No.8, U.S. Army Corps
of Engineers, Hydrologic Engineering Center, Davis, Calif.
Sanders, C. L., and Sauer, V. B. (1979). "Kelly Barnes Dam failure flood of November 6, 1977." U.S. Geological
Survey Hydrologic Investigations Atlas HA-613, U.S. Government Printing Office, Washington, D.C.
"Schedule, general provisions, specifications, and drawings-Lower Two Medicine Dam-Blackfeet Indian Ir-
rigation Project, Montanta." (1965). Specification No. DC-6320, U.S. Department of the Interior, Bureau of
Reclamation, Denver, Colo.
Scott, K. M. and Gravlee, G. C. Jr. (1968). "F1ood surge on the Rubicon River, California-Hydrology, Hy-
draulics, and Boulder Transport." U.S. Geological Survey Profl. paper 422-M, U.S. Government Printing
Office, Washington, D.C.
"Simplified dam-breach routing procedure." (1981). U.S. Department of Agriculture Tech. Release No. 66, U.S.
Department of Agriculture, Soil Conservation Service, Washington, D.C.
Sterling, G. (1914). "Failure of the dam of the Hatchtown Reservoir, Utah." Engrg. News, 71(23), 1274-1275.
Sterling, G. (1916). "Analysis of the failure of an earthfill dam." Engrg. News, 75(2), 56-61.
"Undermining causes failure of French landing dam." (1925). Engrg. News-Record, 94(18), 735-736.
Warne, W. E. (1964). "The Baldwin Hills dam failure." Western Constr., 39(2), 78-80.

APPENDIX II. NOTATION


The following symbols are used in Ihis paper:

A cross-sectional area of dam embankment;


Br breach factor;
H height of embankment measured from bottom of final breach to top of embankment;
H "hat matrix";
H.. reservoir depth at time of failure measured from bottom of final breach to reservoir surface;
h max largest diagonal element of H;
n number of observed values;
p number of regression model parameters;
Q" peak outflow from breached embankment dam;
I I-distribution statistic;
V.. reservoir storage volume at time of failure;
W average width of embankment from bottom of final breach to top of embankment;
X augmented matrix of independent variables;
Xt, vector of regressor variables;
f3 vector of estimated regression coefficients; and
(J" standard error of estimate of regression model.

JOURNAL OF WATER RESOURCES PLANNING AND MANAGEMENT 97

J. Water Resour. Plann. Manage. 1995.121:90-97.

You might also like