Documente Academic
Documente Profesional
Documente Cultură
Hydrology and dams are two fields that are obviously closely related. Yet each topic can be
largely studied without consideration to the other. The fact is that most of the analyses
developed in a hydrological study for dams are not dependent from the type of the structure
nor its characteristics. Its location is just important to determine the quantity and variability
of the water inflow available for useful use (irrigation, energy production) and the properties
of the unusual hydrological events threatening the dam (floods and their intensity). But the
hydrological studies do not address the dam structure itself.
Four bulletins have so far been published by this Committee : Selection of Design Flood –
Current methods (#82, 1992), Dams and Floods – Guidelines and cases histories (#125,
2003), Role of Dams in Flood Mitigation – A review (#131, 2006) and Integrated Flood
Management (#156, 2010). These texts have essentially addressed floods, the risks they
represent and their significance for the concerned populations. Examples were given of the
observed beneficial role of dams, as well as conceptual approaches for evaluating extreme
situations. The experiences made with integration of flood management in the reservoir
operation were described as well, along with other considerations related to climate change
and real case experiences.
The present Bulletin deviates slightly from this path, adopting a somewhat more technical
perspective. The text consists of three chapters, conceived to be accessible to the
practitioners. The first one is a useful vade mecum of usually applied "classical"
hydrological methods; these methods are briefly summarized, their key features highlighted.
The second chapter describes a promising original approach to cope with the uncertainties of
extreme floods; this method is tailored to makes use of the enormous capacity of computers.
The last one focuses on the various criteria applied throughout the world to decide which
flood severity must be considered as a reference for designing a flood evacuation system.
All Committee members contributed to the preparation of these chapters. They are thanked
for their efforts and active cooperation. A few persons ensured that the chapters would
properly take shape. They have been the spirit and the soul of this Bulletin. These authors
committed themselves with energy and perspicacity to the collection of information, its
treatment and the writing of the text. They wrote a few specific recommendations at the end
of their chapter. I am highly grateful for the quality of their work. They are:
Chapter 1 Jun Kyung Soo, Republic of Korea
Chapter 2 Roderic Nathan, Australia
Chapter 3 Douglas Sparks, Canada
Michel Tremblay, Canada
Ozcan Dalkir, Turkey
Our common wish is that the ideas developed in this Bulletin help the practitioners in their
work, be it by the use of traditional methods, the implementation of new approaches or the
proper selection of project and check floods.
Bernard Joos
Chairman, Technical Committee on
Flood Evaluation and Dam Safety
December 2015
Introduction
The Bulletin
The Chapter 1 of the present Bulletin presents a reminder of the classical methods of flood
estimation. It addresses the existing approaches and describes them succinctly. The style is
simple, focusing on the essential features of each method and avoiding dwelling on the
details. This Bulletin not being meant to be an encyclopaedia but rather a practical tool
helping the practitioner in his day-to-day activities, only the most largely recognized and
applied methods are presented. The main theories developed since the end of the 19th
century and still applied are mentioned.
Through its structure and contents, this chapter may become a very useful reference for the
practitioner, who will find in a condensed form a description of the most important
approaches and techniques traditionally applied in hydrology. Basic equations and
definitions, reference graphs, flow charts and typical values of key hydrological parameters
are logically presented. The chapter is concluded by a few recommendations that prevent the
beginner (and possibly the more experienced practitioner!) from committing common errors.
A list of references complete the presentation and invite the reader to further investigations.
The Chapter 2 discursively confronts the issue of extreme floods and new approaches
attempting to cope with the large uncertainties implied in their estimation. An extreme
precipitation causes for sure an extreme flood, but is it legitimate to pretend that an extreme
flood is necessarily the result of an extreme rainfall event? Is there a perfect coincidence
between the return period of a precipitation and that of the resulting flood? It is true that a
difference of a few years in the return period estimate of a rare event is not harmful... but
what if this difference relates to a more frequent flood? These questions indeed evade the
role of snow, which in northern countries can significantly contribute to the formation of
very large floods during the thaw season.
The uncertainties surrounding the majority of parameters involved in a hydrological process
influence considerably the quality of the determination of the flood hydrograph and the
estimation of the flood consequences. A simulation technique based on a stochastic
approach enables to efficiently apprehend this difficulty. It provides the means of
determining the return period of floods from the probability of occurrence of meteorological
events. More important, the method discussed in this Bulletin allows to determine the
distribution function of flood indicators in specific situations (e.g. resulting maximal
reservoir level, duration of spilling, etc.). Of course, it relies on particularly efficient
numerical tools, which enable generating very large quantities of random variables and
processing them. This is achieved by using the Monte-Carlo method.
A few recommendations are made for the readers interested in applying these techniques at
the cutting edge of current hydrology. A generous list of references is proposed for those
eager to reading more about this promising approach.
By its nature, the Chapter 3 is not as purely hydrological as the preceding ones; it is rather
focused on floods and dams. It is particularly important, insofar as it summarizes the current
practice followed in some thirty countries about dam safety directives related to floods. It is
indeed highly important to know how to estimate the key characteristics of an extreme
hydrological event (peak flow, duration, volume, etc.).
The current approaches can be grouped in two main categories, with a few nuances:
- The first one is an indirect method: it considers the dam and its characteristics. It
derives the potential risk and defines the magnitude of the flood to be mastered.
- The second one is a direct method: it takes into account the consequences of a dam
failure, focusing on the damages occurring in the region downstream the dam.
The presentation of the principles adopted in various countries is complemented by a
realistic example of dam and reservoir. This exercise aims at determining the critical event
that would have to be considered as design flood, as analysed through the thirty directives.
This simple comparison demonstrates clearly that, for the time being, no doctrine unity
prevails and the range of the various estimates is quite wide. A few recommendations are
also emitted this regarding. A list of references, particularly on the origin of the national
directives, round up this chapter.
To conclude
A few questions remain open at the end of this Bulletin: for instance, which approach shall
be selected to estimate the intensity of critical floods? A statistical method based only on
discharges measured in the river or a statistical method based on precipitations? Should the
latter be preferred, how can one be sure whether the return period estimated for a
meteorological event is the same as for the resulting flood (see Chapter 2)? Finally, what to
do if two types of floods do not form a consistent duo (e.g. project and check floods), if
these events have been estimated with different methods?
A theme that plays a role less negligible than usually considered, is that of the river base
flow at the beginning of a flood. Which influence can this discharge exert on the intensity of
the flood peak? Two identical rainfalls on a given watershed will indeed not lead to an
identical peak flow, if the initial filling state of the various aquifers of the watershed and the
resulting discharge in the river differ.
Another point frequently observed by the determination of extreme floods is the very high
attention put on the flood hydrograph (in particular its peak flow) and the generally much
smaller consideration attributed to the ratio of the rainfall volume to that of the resulting
flood. A re-balancing of these views can at any rate only be beneficial to the consistency of
the estimates made by the hydrologist.
A last word yet on the mastering of uncertainties, which represents a major challenge of
hydrology, as seen above. In the realm of rare events, uncertainties are particularly marked.
On one hand, there are "natural" uncertainties related to the observed floods (magnitude and
shape of the flood hydrograph, snow melt, response of the water catchment, seasonality
effect, imprecision of peak flow measurements, etc.). There are of course also the
uncertainties simply related to the measurement of the river flow. On the other, there are the
additional uncertainties of extrapolation methods used for estimating the magnitude of
extreme events (partial or strong inadequacy of the adopted distribution with respect to the
considered sample).
The random uncertainties due to the natural variability of processes combine with the
epistemic imprecision of the extrapolation; this leads in general to an important uncertainty
in the determination of the extreme events. The hydrologist faces then the delicate task of
driving safely his modest boat on the oftentimes treacherous sea of uncertainties. Here,
consistency and robustness of the approach must prevail over the illusion of precision.
In this context, these few thoughts incite to rephrase an almost bicentennial advice (*). Let
us address the following injunction to this Bulletin: "Go, little book, and choose your
audience; since to uncertain situations, who believe in them is bold, who doubt about them is
not wrong, who dives into them gets lost and who properly balances them is master".
(*)
Rodolphe Töpfer (1799-1846), inventor of comics (not a hydrologist!)
Chapter 1
Flood Hydrograph
Intensity–Duration–Frequency relationship
The most common approach of determining the design storm event involves a
relationship between precipitation intensity (or depth), duration, and the frequency (or
return period) appropriate for the facility and site location. Such designs are based on
estimates of worst-case scenarios of precipitation intensity and duration during a given
interval of time.
1. Select a design storm duration 𝐷.
2. Collect the annual maximum precipitation depth of the selected duration from n
years of historic data.
3. Determine the probability distribution of the 𝐷-hr annual maximum precipitation.
The mean and standard deviation of the 𝐷-hr annual maximum precipitation are
estimated.
4. Calculate the 𝐷-hr 𝑇-yr design precipitation depth from the probability
distribution for a number of T (return period) values.
5. Calculate the average precipitation intensity (precipitation depth divided by storm
durations)
6. Repeat steps 1 through 5 for various design storm durations.
7. Construct the I-D-F curves.
- Arithmetic mean
This simple method calculates the mean areal precipitation (MAP) by adding up
design precipitations for each station and dividing the sum by the number of
stations. It yields plausible MAP values, provided that the precipitation amounts at
each station are not different more than 10 %.
Figure 1.2 Thiessen polygon method Figure 1.3 Inverse distance method
- Isohyetal method
Isohyetal method consists of drawing isohyets or contour lines of equal
precipitation, by interpolation between the measured values at the rain gage
stations. The method can be applied with Equation (1.2).
ARF can be determined by two methods, namely storm-centered ARF and fixed-area
ARF. The storm-centered ARF are associated with rainfall intensity within the rainfall
isohyets of specific storm events, they represent the ratio of average storm depths over
an area (defined by rainfall isohyets) and the maximum rainfall depths for the storm (at
storm-centered). The fixed-area ARF relate rainfall estimation at point to the average
over catchment which is fixed in space. They are estimated by constructing from all
available rainfall data at station, the time series of catchment average rainfall,
performing the same types of extreme value analyses described above for constructing
point IDF curves, and finally relating the catchment rainfall intensities to the point
values, for the same return period and duration (Svensson and David, 2010).
Hydro-meteorological method
The hydro-meteorological method estimates the PMP using basin average rainfall,
moisture maximization ratio, storm transposition and topographic influence ratio. PMP
is estimated by the following equation.
𝑃𝑀𝑃 = 𝑂𝑃 ∙ 𝑅𝐼𝑃 ∙ 𝑅𝐻𝑇 ∙ 𝑅𝑉𝑇 ∙ 𝑅𝐺𝐹 (1.4)
where 𝑂𝑃 = basin average rainfall by extreme storm transposition;
𝑅𝐼𝑃 = moisture maximization ratio;
𝑅𝐻𝑇 = horizontality transposition ratio;
𝑅𝑉𝑇 = verticality transposition ratio;
𝑅𝐺𝐹 = topographic influence ratio.
Storm models
The application of storm model is effective where rainfall data is insufficient or non-
representative, and where uneven surface of elevation complicates rainfall phenomenon.
For instance, convective cell model is one of the storm models and it is applicable
model for thunderstorm (Wiesner, 1970). Storm models are recommend in which the
area is large.
Incremental excess for a time interval is computed as the difference between the
accumulated excess at the end of and beginning of the period. The maximum retention,
S, and watershed characteristics are related through an intermediate parameter, the curve
number (commonly abbreviated CN) as:
25400−254CN
S= (1.11)
CN
If a given one-day rainfall produces a 1-in. depth of runoff over the given drainage
area, the hydrograph showing the rates at which the runoff occurred can be considered
a unit graph for that watershed.
Thus, a unit hydrograph (UH) is the hydrograph of direct runoff (excluding base flow)
for any storm that produces exactly unit depth (1.0 cm or 1.0 inch) of net rainfall.
Assumptions in the unit hydrograph approach are follows:
(a) (b)
- S-curve method
The S-curve method overcomes restrictions imposed by the lagging method and
allows construction of UH of any duration. It assumes that a UH of duration D is
known and that we intend to generate a UH for the same watershed with duration
D′ . The first step is to generate the S-curve hydrograph by adding a series of UHs
of duration D, each lagged by time Period D (Figure 1.12a). By lagging the S-curve
in time by D′ hr and subtracting ordinates between the two S-curves, the resulting
hydrograph must be due to rainfall of 1/ D cm/hr that occurs for D′ hr. Thus, to
convert curve b to a UH, we must multiply all the hydrograph ordinates by D/
D′ (Figure 1.12b).
(a) (b)
Figure 1.13 Use of the IUH to Figure 1.14 Baseflow separation methods
generate a hydrograph
- Subjective methods
Several subjective methods are shown in Figure 1.14. The simplest one consists in
arbitrarily selecting the discharge marking the beginning of the rising limb as the
value of the base flow and assuming that this base flow discharge remains constant
throughout the storm duration.
- Area method
The area method of base flow separation consists in determining the beginning of
the base flow on the falling limb with the following empirical equation, relating the
time in days from the peak discharge, N, to the basin area, A.
Hydrograph components
The total stream flow during a storm event includes the base flow existing in the basin
prior to the storm and the runoff due to the given storm rainfall. The total stream flow
hydrographs are usually conceptualized as being composed of direct runoff and base
flow. Direct runoff is composed of contributions from surface runoff and quick
interflow. The base flow is composed of contributions from delayed interflow and
groundwater runoff (Ramirez, 2000).
1. Surface runoff includes all overland flow as well as all rainfall falling directly onto
stream channels. Surface runoff is the main contributor to the peak discharge.
2. Interflow is the portion of the stream flow contributed by infiltrated water that
moves laterally in the subsurface until it reaches a channel. Interflow is a slower
process than surface runoff. Components of interflow are quick interflow, which
contributes to direct runoff, and delayed interflow, which contributes to base flow.
3. Groundwater runoff is the flow component contributed to the channel by
groundwater. This process is extremely slow compared to surface runoff.
Hydrographs can also be described in terms of the time characteristics. The main timing
aspects of the hydrograph can be characterized by the following parameters:
Method Formula
𝐿 = length of channel, ft
Kirpich 𝑡𝑐 = 0.0078𝐿0.77 𝑆 −0.385 𝑆 = average watershed slope, ft/ft
L = length of longest watercourse,
USBR Design 11.9𝐿3 0.385 mi
of Small Dams 𝑡𝑐 = 60( ) H = elevation difference between
𝐻
divide and outlet, ft
i= rainfall intensity, in/hr
41.025(0.0007𝑖 + 𝑐)𝐿 0.33 c= retardance coefficient
Izzard 𝑡𝑐 = L= length of flow path, ft
𝑆 0.333 𝑖 0.667
S= slope of flow path, ft/ft
C = rational method runoff
Federal Aviation coefficient
𝑡𝑐 = 1.8(1.1. −C)𝐿0.5 𝑆 −0.333
Administration L = length of overland flow, ft
Lag time
Delaying time of which a flow caused by excess rainfall over a watershed reaches point
of maximum peak of runoff. Conceptual way to explain lag time comprise of bands
which divide area into several subarea. Lag time is a weighted time of each travel time
from the centroid of individual bands to outlet of main watershed.
t L = 0.6t C (1.15)
where t L is the lag time and t c is the time of concentration.
Table 1.3 Typical Snyder’s coefficients for U.S locations (Viessman and Lewis, 2003)
North/mid-Atlantic
- 0.6/√𝑆 - -
States
Corps of Engineers
0.4-8.0 0.3-0.9 - -
training course
t B = 3 + t L /8 (1.18)
where t B is time base of the hydrograph (days) and t L is lag time (hr).
The duration (D) of rainfall excess for Snyder’s synthetic UH development is a function of
lag time:
D = t L / 5.5 (1.19)
For other rainfall excess duration D', an adjusted formula for t L becomes
t ′L = t L + 0.25(D′ − D) (1.20)
where t ′L = adjusted lag time (hr) for duration D′ (hr).
(a) DEM (b) Hill slope (c) Flow accumulation (d) Flow distances
Figure 1. 20 Feature of watershed (Loesch, 2000)
If the storage is plotted against outflow for a river reach, the resulting curve will generally
take the form of a loop. This loop effect implies greater storage for given outflow during
falling stages than during rising stages. If one considers water surface profile at various
times during passage of the flood wave, the concept of prism and wedge storage is useful.
Figure 1.21 Storage in a river reach Figure 1.22 Prism and wedge storage
concepts
Rational method
Rational method is one of the simplest and best-known methods routinely applied in
urban hydrology, although it contains subtleties that are not always appreciated. Peak
flows are predicted by the simple product
Qp = C · I · A (I = in/hr, A = acres) (1.33 a)
Flow-based method
For design locations where records of stream flows are available, or where flows from
another basin can be transposed to the design location, a design flood magnitude can be
estimated directly from the stream flows by any of the following methods:
1. Frequency analysis of flood flows at the design location or from a similar basin in
the region.
2. Use of regional flood frequency equations normally developed from regression
analysis of gauged flood data.
3. Examination of the stream and floodplain for signs of highest historical floods and
estimation of the flow rates using measurements of the cross section and slope of
the stream.
Precipitation-based method
Where stream-gauging records are unavailable or inadequate for streamflow estimation,
design floods can be estimated by evaluating the precipitation that would produce the
flood, and then the precipitation is converted into runoff. Typical methods include:
1. Design using the greatest storm of record at the site, by converting the
precipitation to runoff.
2. Transposition of a severe historical storm another similar watershed in the region.
3. Frequency analysis of precipitation and conversion of the design storm to runoff.
4. Use of a theoretical probable maximum precipitation (PMP), or fraction of PMP,
based on meteorological analyses.
Critical-event method
Due to the high risk to lives or properties below major structures, the design of these
structures generally includes precisions for a flood caused by combination of the most
sever meteorologic and hydrologic conditions that are possible. Instead of a design for
some frequency or the least expected total cost, flood-handling facilities for the
Watershe
Overland
d Minimal Evapotran Channel Pond
Model name flow routing
conceptua time step -spiration routing routing
method
l-ization
Plan M, M-C,
User-
HEC-HMS 1 min UH, KW mod Puls, Mod Puls
/ channel supplied
lag, KW
Lumped Constant
AnnAGNPS Lumped 1 day Penman Manning’s
travel time outflow
Thomthwai
Grid or sub KW or te, H-Q, KW, DW, KW, DW,
HYDROTEL 1 hour
watershed GIUH Linacre, P- Mod Puls Mod Puls
M, P-T
Table
Rational /
P8 Lumped 1 hour Hamon’s Crude interpolati
NRCS
on
Grid or Daily
Mod. P-M, P-T, Muskingum
SWAT subbasin / 1 day water
Rational Hargreaves , Manning’s
channel balance
1.7 Recommendations
The application of the methods and approaches presented in this chapter is in general
relatively straightforward. The reader is referred to the literature listed below to clarify
points that still need more thoughts. A few recommendations are nevertheless formulated,
which are worth keeping in mind while dealing with hydrology :
Hydrology cannot be properly managed without consideration to all its uncertainties and
biases. While solving an hydrological problem, it is recommended to consider various
approaches and mobilize several, if possible clearly different methods.
The hydrologist should aim at reaching a reasonable consistency between the various
applied approaches rather than focusing on just one method and painstakingly
attempting to apply it down to its tiniest details.
There are well known inherent difficulties to precisely quantify hydrological processes,
as well as additional uncertainties introduced by extrapolation methods (for flood
estimation in particular). The key characteristics of floods (peak flow, duration, volume)
are inevitably imbedded in a wide range of uncertainties. A minimum number of
significant digits is sufficient to express these quantities.
The existing methods presented in this chapter are not the alpha and omega of
hydrology. New methods appear regularly, some adopting simple, original approaches,
others relying on heavily numerical solutions. These new procedures are possibly not as
intuitive as the classic ones, but offer interesting perspectives worth the hydrologist's
attention.
References
Akan, A. O., and Houghtalen, R. J. (2003). Urban hydrology, hydraulics, and stormwater
quality: engineering applications and computer modeling. John Wiley & Sons.
Asquith, W.H. (1999). “Areal-reduction factors for the precipitation of the 1-day design
storm in Texas.” Report, USGS, Austin, USA.
Bedient, P.B., and Huber, W. C. (2002). Hydrology & floodplain analysis. Prentive Hall,
USA.
Chow, W.T., Maidment D.R., and Mays L.W. (1988). Applied Hydrology. McGraw-Hill,
USA.
Clark, C.O. (1945). “Storage and the unit hydrograph.” Transactions, ASCE, 110, 1419-
1446.
Creutin, J. D., & Obled, C. (1982). “Objective analyses and mapping techniques for
rainfall fields: an objective comparison.” Water resources research, 18(2), 413-431.
2.1 Introduction
This chapter provides an overview of current trends in the estimation of extreme floods
relevant to dam safety. Given this context, emphasis is given to the estimation of floods with
annual exceedance probabilities rarer than 1 in 100, up to the Probable Maximum Flood.
Such estimates are required to assess the adequacy of the flood capacity of dams, and for
assessing the (incremental) impacts of dam outflows on downstream areas. The severity of
floods relevant to dam safety may be one or more orders of magnitude more extreme than
has been observed in the historic record. As such, the estimation of this range of floods
presents special difficulties that are not generally encountered when dealing with floodplain
management and planning.
The hydrologic criteria relevant to the design of flood mitigation dams is generally based on
two design conditions (ICOLD, 2003), namely 1) the maximum inflow flood that can be
safely passed by the dam, and 2) the optimum degree of flood mitigation achieved under
controlled operations.
Traditionally, standards-based approaches have been used to define these design conditions,
where the upper limit of interest is defined by the Probable Maximum Flood (PMF). The
details of how PMFs should be estimated vary with region, but a common attribute has been
to consider the most severe combination of conditions that are reasonably likely to occur (eg
ICOLD, 1992). The focus here is on the magnitude of the flood, where consideration is
given to both the peak and the volume of the hydrograph, as these two characteristics
influence the degree of flood mitigation achievable by the dam.
More recently, attention is being given to risk-based approaches to design criteria, in which
it is necessary to characterise the relationship between both the magnitude of the flood and
its probability of exceedance. Thus, rather than consider a single extreme storm event (such
as the Probable Maximum Precipitation) it is necessary to derive a distribution of floods,
ranging over the whole probability domain of interest.
The two different approaches to specifying design criteria lend themselves to different
estimation methodologies. For the derivation of many standards-based criteria it is sufficient
to use deterministic approaches, in which the floods of interest are derived using a fixed
combination of hydrometeorological inputs. By contrast, the derivation of risk-based criteria
are best undertaken using stochastic approaches, in which the selected factors controlling the
conversion of rainfall into floods are sampled from their associated probability distributions.
In practice, however, there are a range of methods applicable to both risk-based and
standards-based assessments. The next section categorises the various approaches that are
applicable, and this is followed by more detailed descriptions of the main approaches
applicable to dam safety. These are followed by a brief description of how the impacts of
climate change can be incorporated into the required design estimates.
Design rainfall event of specified Stochastic sample of rainfall Time series of daily (or shorter)
Rainfall events (based on rainfall
annual exceedance probability historic (or synthetic) rainfalls
input frequency curve)
Flood
characteristic Resulting flood hydrograph Distribution of flood events, Time series of daily (or shorter)
of known assumed to have same annual where total probability theorem streamflows, where statistical
annual exceedance probability as input used to derive flood distribution is fitted to extracted
exceedance design rainfall characteristic of specific annual maxima
probability exceedance probability
Figure 2.1 Different simulation frameworks used for the estimation of design floods
When used to estimate floods with a specified annual exceedance probability, deterministic
techniques make the assumption that the exceedance probability of the derived flood is the
same as the input rainfall. This assumption is made on the basis that the hydrologic factors
that control runoff production are set to be “probability-neutral”.
In practice this means that factors related to the temporal and spatial distribution of rainfall,
antecedent conditions and losses, are set to “typical” values that are associated with the input
rainfall. Factors related to formation of the hydrograph are generally assumed to be invariant
with rainfall. The assumption of probability-neutrality is not relevant when estimating the
Probable Maximum Flood, as here the design objective is to derive a flood that is the
maximum “reasonably” possible, where no account is given to its exceedance probability.
Stochastic techniques offer an alternative to the design event method (eg Rahman et al,
2002; Weinmann et al, 2002; Nathan et al, 2002). These techniques recognise that any
design flood characteristics (e.g. peakflow) could result from a variety of combinations of
flood producing factors, rather than from a single combination. For example, the same peak
flood could result from a moderate storm on a saturated basin, or a large storm on a dry
basin.
An illustration of this is provided in Figure 2.2, where the grey points represent the flood
peaks arising from the stochastic combination of key flood-producing factors for a
catchment with an area of 7000 km2. The vertical range of these flood peaks for a given
annual exceedance probability reflects the sensitivity of the floods to stochastic factors for a
given rainfall depth (in this case, the spatial and temporal patterns of rainfall, and initial and
continuing losses).
Such approaches attempt to mimic “mother nature” in that the influence of all important
stochastic inputs are explicitly considered, thereby providing a more realistic representation
Figure 2.2 Illustration of the variability in flood peaks arising from combination
of stochastic factors and their relationship to the derived frequency curve
using the total probability theorem
Further details regarding the practicalities of this approach are described in a later section,
but it is worth noting here that this approach can use exactly the same types of hydrologic
inputs and models that are used in deterministic approaches. The deterministic and
stochastic approaches are well suited to event-based models, where it is assumed that the
probability of storm rainfall is the primary variable that controls the exceedance probability
of the outflow floods. In some situations that are more volume-dependent – such as a
cascade of dams or seasonal snowmelt conditions – this may not be an appropriate
assumption, and it may be better to adopt a continuous simulation, or hybrid approach.
With the comprehensive approach (see review by Boughton and Droop, 2003), continuous
simulation techniques are used to convert input time series of rainfall and evaporation into
an output time series of streamflow; the flood events of interest are extracted from the
In essence, Monte Carlo simulation provides the means to transform a frequency curve of
rainfall into a corresponding frequency curve of some desired flood characteristic (Figure
2.5). The outputs of such an analysis can be compared directly to independent information
on flood frequencies, and be used to derive frequency curves over the full probability
domain of interest (Figure 2.6). The fundamental rationale for this approach is that there is
considerably more data available on rainfall (especially regionally) than for streamflows,
and thus we are able to estimate the probability of floods that are much rarer than can be
found in the historic streamflow record, including consideration of the variations in other
flood-producing characteristics that can be included in the Monte Carlo simulation.
It should be noted that the Monte Carlo approach provides information on flood likelihood
or AEPs that is not available from a deterministic estimate of the Probable Maximum Flood
(PMF) based on the Probable Maximum Precipitation (PMP). Thus, the PMP is treated as
merely one point on the rainfall frequency curve, and Monte Carlo simulation allows the
estimation of annual exceedance probabilities for floods with magnitudes between those that
may have been observed in the historic record and those represented by, or even exceeding,
the PMF.
Adapting models to run in a stochastic environment allows the practitioner to take advantage
of existing design information and experience. If industry-accepted models are used, the vast
body of experience and empirical relationships available concerning model configuration
and parameterisation are still largely applicable. Less important parameters can be fixed at
values deemed appropriate from previous experience and available design guidance, and
average values of stochastically varying inputs can be expected to fall within a range
previously considered for use in deterministic design.
The information available for specifying input distributions is often the same as that used to
derive typical or average values of fixed inputs, and many inputs can easily be non-
Another method is the SHYREG Method. The SHYREG method is a version of the
SHYPRE (Simulated HYdrographs for flood PRobability Estimation) method (Arnaud and
Lavabre, 2002), adapted for the purposes of regional flood studies. SHYPRE was first
developed to simulate catchment flood scenarios. It couples a stochastic hourly rainfall
generator (Arnaud and Lavabre, 1999; Arnaud et al., 2006; Cantet et al., 2010; Cernesson et
2500
2000
1500
1000
500
0
1 10 100 1,000 10,000 100,000
Return Period (years)
References
Arnaud, P., and J. Lavabre (1999): Using a stochastic model for generating hourly
hyetographs to study extreme rainfalls., Hydrological Sciences Journal, 44(3), 433-446.
Arnaud, P., and J. Lavabre (2002): Coupled rainfall model and discharge model for flood
frequency estimation. Water Resources Research, 38(6).
Arnaud, P., J.-A. Fine, and J. Lavabre (2006): An hourly rainfall generation model
adapted to all types of climate. Atmospheric Research, 85(2), 230-242.
Baker, V. (2008): Paleoflood hydrology: Origin, progress, prospects. Geomorphology
101, pp 1–13.
Barsugli, J., Anderson, C., Smith, J.B., and Vogel, J.M. (2009). Options for Improving
Climate Modeling to Assist Water Utility Planning for Climate Change. Water Utility
Climate Alliance. USA.
Bergström, S., Hellström, S., Lindström, G., and Wern, L. (2008). Follow-Up of the
Swedish Guidelines for Design Flood Determination for Dams, Svenska Kraftnät Report
No.1. 2008, BE90.
Boughton, W and Droop, O, (2003): Continuous simulation for design flood estimation -
a review, Environmental Modelling and Software, 18:309-318.
Cantet, P., J.-N. Bacro, and P. Arnaud (2010): Using a rainfall stochastic generator to
detect trends in extreme rainfall, Stoch Environ Res Risk Assess, 13 p.
Castellarin, A., Vogel, R., Matalas, N.C. (2007): Multivariate probabilistic regional
envelopes of extreme floods. J. Hydrol. 336, 376– 390
Castellarin (2007): probabilistic envelope curves for design flood estimation at ungauged
sites. Water Resour. Res. 43, W04406, 12 pp.
Castellarin, A., Kohnová, S, Gaál, L., Fleig, A., Salinas, J.L., Toumazis, A., Kjeldsen,
T.R., Macdonald, N. (2012): Review of Applied European Flood Frequency Analysis
Methods, Centre for Ecology & Hydrology, Natural Environment Research Council, UK.
Cernesson, F., J. Lavabre, and J.-M. Masson (1996): Stochastic model for generating
hourly hyetographs. , Atmospheric Research, 42, 149-161
Chetty, K and Smithers, J, (2005): Continuous simulation modelling for design flood
estimation in South Africa, Physics and Chemistry of the Earth, 30:634-638.
Darch, G.J.C. and P.D.Jones, P.D. (2010): From sensitivity to scenarios: assessing future
flood risk through the application of climate model output in operational hydrological-
hydraulic models. Proc. British Hydrological Society, Third International Symposium,
Managing Consequences of a Changing Global Environment, Newcastle 2010.
3.1 Introduction
3.1.1 Background
Most countries have defined criteria and methods to select the design flood for dams to
protect their population. Approaches to select the IDF varied considerably along time and
along countries. In the United States, for example, the evaluation of the IDF by the Bureau
of Reclamation has evolved this way1:
For dams constructed before the early 1940s, the IDFs were based on a combination of flood
frequency curve extrapolations, envelope curves, and maximum flood ratios (e.g., 50 percent
greater than the flood of record);
In the 1940s, the Bureau of Reclamation has developed deterministic approaches to estimate
design floods based on the PMF or some variants;
Prior to 1980, an equivalent deterministic approach was used, where the Maximum Probable
Flood (MPF) was typically selected as the IDF for most storage dams. The MPF is roughly
equivalent to the PMF, except that site-specific PMP rainfall information was used, rather
than PMP estimates from the HMR series.
Subsequently adopted PMF nomenclature in the early 1980s, along with most Federal
agencies.
During the early 1980s and into the mid-1990s, IDF selection criteria were used which were
based on downstream hazard classification and potential loss of project operations.
The methods to select the IDF have evolved over time and continue to evolve with the
availability and development of new tools and new approaches which give a better
evaluation and understanding of the possible consequences of a dam failure.
3.1.2 Context
ICOLD Bulletin 125 on Dams and Floods: Guidelines and case histories (ICOLD 2003)
define three generations of approaches for specifying or selecting design floods, as follows:
First generation – based on empirical and general considerations, and applicable to
any dam and in any situation, without taking into account size or type of the dam,
volume of reservoir, nor downstream consequences hazard2.
1
USDI, Bureau of Reclamation, Design Standards no 14 – Appurtenant Structures for Dams (Spillway and
Outlet Works) - Chapter 2: Hydrologic Considerations - Draft: Phase 3 (Public Review), December 2012.
2
In English-speaking countries the terms hazard or potential hazard are widely used in dam engineering
circles to describe the consequences of a dam failure. This use is at odds with the use of the term in the
field risk assessment where a hazard is defined as a source of potential harm or a situation with a potential
to cause loss (ICOLD 2005).
3.1.3 Terminology
Inflow Design Flood (IDF): The flood used to design and/or modify a specific dam and
its appurtenant works; particularly for sizing the spillway and outlet works, and for
determining surcharge storage and height of dam requirements. (USBR)
Hazard Potential Classification: The hazard potential is determined through an
assessment of the greatest incremental losses that could result from an uncontrolled
release of the reservoir due to the failure of a dam or its appurtenances. Potential
incremental losses are to be assessed with respect to life, property, the environment and
heritage sites at the dam site, upstream, downstream, or at other areas influenced by the
dam. (Canada - Ontario)
Incremental: Under the same conditions (e.g., flood, earthquake, or other event), the
difference in impact that would occur with or without failure or mis-operation of the
dam. (FEMA);
Loss of Life (LOL): The number of persons who could die if the dam, or the proposed
dam after its construction, fails. The evaluation of the LOL is based on theoretical
approaches;
Maximum Experience (Historical) Flood: Maximum flood observed at the project site
or maximum flood that could be observed based on flood experienced in the basin;
Population at Risk (PAR): The number of persons whose safety will be at risk if the
dam, or the proposed dam after its construction, fails;
Probable Maximum Flood (PMF): The largest flood that may occur at a given point on a
drainage area from the most severe combination of critical meteorological and
hydrological conditions reasonably possible on a particular watershed.
Probable Maximum Precipitation (PMP): The greatest depth of precipitation for a given
duration that is physically possible over a given storm area at a particular geographical
location at a certain time of the year.
Safety Check Flood (SCF): This is an extreme flood condition that a dam must be
capable of withstanding while continuing to operate safely, accepting some damage and
a reduction in safety factors but without causing dam failure (Canada-Quebec).
3.2.1 Generalities
The consequences following a dam failure can take various forms, such as:
Possible loss of life, injuries to persons at risk;
Economic impacts
o Loss of property, transportation infrastructure, energy production, water supply,
navigation, irrigation, other economic activities, …
Social impacts
o Consequences on “human” life after a dam failure;
o Cultural – built heritage;
o Fear of such event occurring in the future;
o Relocation of the population;
Environmental impacts
o Impact on living organisms in general and wildlife particularly,
o Modification of the topography/bathymetry along the river area;
o Pollution related to human activities (fuel, chemical, tailings, …).
The determination of the IDF depends in most countries on a Dam Hazard Classification
System, which examines the risk related to the dam and/or structures; the greater the risk,
the greater the requirement to reduce, contain or manage this risk.
Dam height
The dam height is one of the main parameters consider for the dam classification since the
energy released during a dam failure is directly related to the dam height or more
specifically to the water head upstream from the dam
Volume of reservoir
Another important parameter is the volume of the reservoir which has a direct impact on the
duration of the flood as well as the energy released ant the downstream water depth after a
dam failure. Sometimes, a relation between the dam height and the reservoir volume is
considered to classify the dam. For example, ICOLD considers as a “Large Dam” any dam
over 15 m high or between 5 m and 15 m high and with a reservoir capacity of 3 000 000 m3
or more.
Dam type
The dam type can also affect the dam hazard classification, since the mode of failure will be
different.
For a concrete and/or masonry dam it is generally assumed that the dam can withstand
overtopping. Failure will occur when the water depth over the crest of the dam would cause
instability or after erosion of the foundation3.
For an embankment dam (including rockfill dams, earth dams, …), it is generally assumed
that the dam failure will occur when the water level exceeds the crest of the dam. It should
also be noted that most dam failures have been observed on embankment dams (80% - ref
ICOLD Bull. 99).
Other parameters are considered for the dam hazard classification. These parameters are not
always contributing in the determination of the IDF, but are considered for the frequency
and type of inspection required for existing dams, such as :
The age of the dam (or time since the last major rehabilitation);
The type of dam foundation;
The general dam and structures conditions;
The seismicity in the dam area;
The reliability of the discharge facilities.
3
If a concrete dam is located on an erodible foundation, it should not be designed to be overtopped.
4
Autoridad nacional de los servicios Publicos - Republica de Panama. NORMAS PARA LA SEGURIDAD
DE PRESAS, Septiembre de 2010
I > 1 000 Very > 333 > 133.3 > 100 Very > 750
important important
Environmental consequences
Environmental consequences are sometimes considered in the dam classification. Most of
the time, the consequences are measured related to the loss of fish, wildlife and their habitat;
the hazard potential is estimated based on the potential damages and the possibility to
regenerate these habitats. However, other aspects are also considered, such as the potential
contaminated area and the impact on the potential loss of cultural and historical heritage
sites, as well as pollution created by the inundation and damage of hazardous waste
detention sites.
Hazard
Environmental losses
potential
To evaluate the consequences of a dam failure for a specific flood, a complete dam break
study is normally required, which includes the evaluation of the cascade effects when a river
or drainage basin has a series of dams. The incremental or the total consequences related to
a dam failure can be considered for the dam classification. The incremental consequences
for a specific flood are estimated considering the difference of damages occurring with and
without the dam failure; normally the larger the flood, the lower the incremental damages
related to a dam failure (also depends on the storage volume of upstream reservoirs).
5
Ontario Ministry of Natural Resources, Classification and Inflow Design Flood Criteria – Technical
Bulletin, August 2011, Canada.
3.2.4.4 Freeboard
The freeboard reduces the risk that the structures will be submerged and can be considered
as a safety factor against the uncertainty related to the flood. It is normally higher for
embankment dams than for concrete dam since the latter can often be submerged without
major damages. A minimum freeboard is defined in the country guidelines of some
countries, such as Italy, Japan and Switzerland.
6
Central Water & Power Commission – Ministry of Irrigation & Power – Government of India,
Estimation of Design Flood – Recommended procedures.
Table 3.3 – France – Dam classification based on height and reservoir volume
IDF
Dam Safety
Classification criteria
class Embankment check flood
Concrete dams
dams
A H ≥ 20m 1 000 to 3 000 dams
10 000 100 000
B H≥10m and H2√V ≥ 200 1 000 3 000 30 000
C H≥5m and H2√V ≥ 20 300 1 000 10 000
D H≥2m 100 300 1 000
In Russia, dam classification is based on the dam height, dam type, reservoir volume and the
use of the dam (which is representative of the economical impact following a dam failure).
The IDF can vary from the 1:-20-yr flood to the 1:1 000-yr flood and the safety check flood
can be the 1:10 000-yr flood. However, Russian design practice considers a guarantee
correction to the maximum flow of exceedance probability of 1:10 000-yr; the shorter the
hydrological record series is, the larger the guarantee correction will be.
Safety check
Dam class Classification criteria Design flood flood
H < 2m, V < 10 000 m3 NA
0 1:200-yr
Minimal consequence
Low consequence PMF or
1 1:500-yr 1,5 * 1:500-yr
(no permanent dwelling)
Medium consequence PMF or
2 1:1 000-yr 1,5 * 1:1 000-yr
(1 to 20 dwellings)
High consequence PMF
3 1:1 000-yr
(21 to 150 dwellings)
Very high consequence
4 (more than 150 1:1 000-yr PMF
dwellings)
Norwegian rules specified also the minimum freeboard acceptable for concrete dam and
embankment dam depending of the type of dam and the number of gates to be considered
not available for the routing of the design flood.
Economical Damages
The most significant economical damages would consist of the crops that would be
temporarily submerged and probably lost, plus the value of the inundated houses. No
significant industrial or commercial facilities or transportation infrastructure are located in
the incremental area. However, the loss of energy production will have a regional impact.
The value of the incremental damages is therefore considered low.
Environmental Damages
No significant wildlife habitat would suffer permanent damage from the temporary
submergence by the dam break flood wave.
7
Graham, W.J., September 1999, “A Procedure for Estimating Loss of Life Caused by Dam Failure”,
DSO-99-06, US Bureau of Reclamation
Canada –
1 000-yr 1 000-yr 1 000-yr 1 000-yr 1 000-yr 1 000-yr
Quebec
New
10 000-yr 10 000-yr 10 000-yr 10 000-yr 10 000-yr 10 000-yr
Zealand
South
100-yr 100-yr 100-yr 200-yr 100-yr 100-yr
Africa
USA -
PMF PMF PMF PMF PMF PMF
FEMA
USA -
PMF PMF PMF PMF PMF PMF
USACE
Table 3.5 – Example – Project β – Evaluation of the IDF – Sensitivity analysis (cont’d)
1/3 between
Canada - CDA 1 000-yr 1 000-yr PMF
and PMF
Canada -
1 000-yr 100-yr PMF
Quebec
Czech
1 000-yr 100-yr 10 000-yr
Republic
3.5 Recommendations
Based on the overview of the guidelines from different countries and the results obtained for
the proposed example, a few recommendations are proposed. They shall serve for the
determination of the inflow designs floods (IDF) :
The determination of the IDF depends on the consequences of a failure of the dam in a
specific area (persons at risk or loss of life, economical, environmental, social, ...). The
determination of the IDF based on the system characteristics takes implicitly into account
the consequences of a failure, but the evaluation of the IDF based directly on the
consequences appears to be a better approach. An alternative option could be to use a
conservative value of the IDF (high value) based on the system characteristics and perform
an evaluation of the consequences, should the owner wants to reduce the value of the IDF.
The large majority of all current national selection rules of the IDF consider a return period
larger than or equal to 1 000 years. Only a few are more severe than 10 000 years. A
reasonable return period for the IDF would be located in this range. More stringent IDF
may be required, depending on local conditions and risks associated to each site.
To increase the safety, structural or operational constraints may be imposed independently
from the IDF, as proposed in various countries. For instance, a N-1 rule (or N-x rule) for the
gates operation and/or a minimum freeboard to be guaranteed at all time can be imposed.
Particular local, legal, technical or administrative conditions may require other types of
constraints.
References
1. ASARIN, A.E. Estimation of the Design Flood in the Design of Russian dams,
Hydropower & Dams, Issue Three, 2007
2. Autoridad nacional de los servicios Publicos - Republica de Panama. NORMAS
PARA LA SEGURIDAD DE PRESAS, Septiembre de 2010
3. Bradlow, D. D., Palmieri, A., and Salman, S. M. A. (2002). Regulatory frameworks
for dam safety , The World Bank, Washington DC.
4. Bureau of Indian Standards (BIS), Guidelines for Fixing Spillway Capacity – IS
11223-1985, Govt. of India, 10 pages
5. Canadian Dam Association, Dam Safety Guidelines, 2007
6. CEATI International inc., COMPARISON OF FLOOD HAZARD ESTIMATION
METHODS FOR DAM SAFETY – Phase 1: TASK 1, Draft, 20 December 2012
7. Central Water & Power Commission – Ministry of Irrigation & Power – Government
of India, Estimation of Design Flood – Recommended procedures.
8. Central Water Commission – Ministry of Water Resources – Govt. of India,
Development of Hydrological Design Aids (Surface Water) under Hydrology Project –
II – State of the Art Report, July 2010, 362 pages
9. Central Water Commission – Ministry of Water Resources – Govt. of India,
Guidelines for Safety Inspection of Dams, June 1987 (revised), 59 pages
10. ELETROBRAS, Critérios de projeto de usinas hidrelétricas, Outubro/2003
11. France, G DEGOUTTE, P ROYET, P CRUCHON, P LE DELLIOU, N MONIE
(France), « Nouvelle réglementation française concernant la sécurité des barrages et
des digues », 23rd ICOLD Congress, Brasilia (2009),
12. France, « recommandations pour le dimensionnement des évacuateurs de crues de
barrages », Juin 2012, rapport du groupe de travail « dimensionnement des
évacuateurs de crues de barrages » du Comité Français des Barrages et Retenues
13. Fridolf, T, Design flood for dams - analysis of the Swedish guidelines, Dams in a
European Context, 2001
14. Gouvernement du Québec, Loi sur la sécurité des barrages, 2000
15. Gouvernement du Québec, Règlement sur la sécurité des barrages, 2002, Décret D.
300-2002, 2002, G.O. 2, 2043
16. Graham, W.J., September 1999, “A Procedure for Estimating Loss of Life Caused by
Dam Failure”, DSO-99-06, US Bureau of Reclamation
17. ICOLD, Selection of Design Flood – Current Method, Bulletin 82, 1982, 233 pages
18. ICOLD, Dams and Floods - Guidelines and cases histories, Bulletin 82, 2003, 229
pages
19. Liu, J. (2002). "Selection of Design Floods in Southeast Asia." In: 5th International
Conference on Hydro -Science & -Engineering (ICHE-2002) , Warsaw.
APPENDIX A
Note
The Appendix A presents a summary of the guidelines in use in different countries to
determine the Inflow Design Flood (IDF) of a project. The objective of this appendix
consists in showing the various approaches considered over the world to establish the IDF.
It must be understood that the information presented hereafter is based on guidelines, papers
and personal contributions made available to the Committee during the preparation of the
Bulletin. Since this time, the information may have been updated; it may also have been
misinterpreted. It must therefore be verified before being used for any specific purpose.
1. AUSTRALIA 8 9 10
In 2000 the new Guidelines on Selection of an Acceptable Flood Capacity for Dams were
published (ANCOLD 2000). In these new guidelines risk assessment is integrated in the
determination of design flood. The Acceptable Flood Capacity, AFC, for a specific dam is
defined as "the overall flood capacity, including freeboard as relevant, which provides an
appropriate level of safety against a flood initiated dam failure to protect the community and
environment, to acceptable overall risk levels, within the total context of overall dam safety
from all load cases".
The risk process requires the owner, or other decision-maker, to take the responsibility to set
the risk management criteria and then to make the decisions on the overall management,
community and environment and political and legal issues, using the information provided
by the risk study.
According to the AFC Guidelines the hydrologic safety should be assessed within the total
load context, and not as a separate case, in order to achieve optimum safety and economy
and not just concentrate on flood safety. In cases where a detailed risk process is too costly
and not practical the guidelines include an alternative based on a hazard classification based
on the population at risk and the severity of damage and loss.
The population at risk is defined as the number of persons whose safety will be at risk if the
dam, if the proposed dam after its construction, fails. Unless otherwise indicated, PAR is the
incremental PAR due to the failure event, that is, the difference in the PAR for the same
event with dam failure relative to the event without dam failure.
Three approaches are defined to evaluate the AFC.
Small Dams
The following approach can be used for dams for :
a zoned or relatively homogeneous earthen embankment less than 12 meters high and
a PAR of 15 or less and
uncontrolled spillways and
depths of flooding of PAR of less than three meters and the product of the depth of
flooding and the average flow velocity is less than 4.6 m2/sec.
For a dam under this situation, the annual exceedance probability (AEP) is defined by the
following relation:
8
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European
Context, Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
9
Dam Safety Committee, DSC 13 – Consequences Categories for Dams, March 2002
10
Queensland Government, Guidelines on Acceptable Flood Capacity for Water Dams, January
2013
Hazard Category
The second approach is based on the hazard category rating of a dam is done, which will be
determine in accordance with table A1-1 :
Note 1 It is unlikely that the severity of damage and loss will be ‘negligible’ where one
or more houses are damaged.
Note 2 Minor damage and loss would be unlikely when PAR exceeds 10.
Note 3 Medium damage and loss would be unlikely when the PAR exceeds 1000.
Note 4 Not used.
Note 5 Change to High C where there is the potential for one or more lives being lost.
Note 6 See section 2.7 and 1.6 in the ANCOLD Guidelines on Assessment of the
Consequences of Dam Failure (ANCOLD, 2000) for an explanation of the range
of high hazard categories
The corresponding AEP can be determined from table A1-2.
Table A1-2 Australia - Acceptable flood capacity based on PAR and damages
11
2. AUSTRIA
Since most of the dams in Austria pose a risk of loss of life, the large dams have not been
classified based on the consequences of a dam failure, but depending of their characteristics.
For dams higher than 15 m above the foundation level and/or reservoir with capacity of
500 000 m³ or more, the IDF should be the 1:5000-yr flood. The reservoir must be assumed
to be at the maximum operation level at the beginning of the event and the discharge
capacity of the bottom outlets and the power plant are not taken into account in the flood
routing. For embankment dams with gated spillways, the (n-1)-gates condition should be
respected. 12
11
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European
Context, Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
12
R. Melbinger, The Austrian approach to dam safety: A symbiosis of rules and engineering
judgement, Dam Safety, Berga (ed.)© 1998 Balkema, Rotterdam, ISBN 90 5410 974 2
3. BRAZIL
Ref : ELETROBRAS, Critérios de projeto de usinas hidrelétricas, Outubro/2003
Bypass Flows
The bypass flow rates for each stage of the management of the River, will be defined by
recurrence time resulting from a risk analysis, comparing the cost of diversion works with
the expected value of the cost of the damage resulting from its floods. In the calculation of
the damages will be considered local, costs damages due to delay in the schedule and any
upstream and downstream damage.
where:
r = risk taken;
Tr = time to recurrence of the flood outflow in years (inverse of annual risk);
n = duration of phase deviation, in years.
Bypass operations and final closure for filling the tank will be the flow characteristics of the
time scheduled for its operations.
4. BULGARIA
In Bulgaria, the criteria to determine the “design non-exceedance of the maximal water
flow-rate” are described in an ordinance published by the Ministry of Energy and Energy
Resources13. This ordinance is valid for hydro project with capacity higher than 5 MW, but
is not valid for the hydropower projects on the Danube River and on the Black Sea Coast.
The design non-exceedance of the maximal water flow-rates to define the maximal water
levels and to size the evacuation (spillway) structures is defined depending on the category
of the structure according to Table A4-1, considering the flood routing in the reservoir.
The design non-exceedance of the maximum water flow-rates for the river diversion
structures for construction of hydropower projects is adopted as 5%, considering the flood
routing. It is acceptable to adopt higher non-exceedance (e.g. 10%) based on technical and
economical justification and an assessment of the consequences.
According to their purpose, the components of the hydropower projects are classified as
permanent or temporary and the permanent components are classified as main or secondary.
The main components are the ones used to retain, keep, transmit and utilise the hydropower
resource, such as :
- the dams;
- the spillway, outlet and intake structures;
- the intakes;
- the derivation channel and hydropower tunnels;
- the pressure basins and daily regulating reservoirs;
- the surge tanks;
- the penstocks;
- the powerhouses of HPP and SHPP.
- the structures and the equipment whose failure leads to cease of the operation of the
components mentioned above;
- the structures whose failure or cease of operation can cause important damage to the
environment or threat to human life.
The secondary components are all permanent components not corresponding to the
definition of the main components.
13
Republic of Bulgaria – Ministry of Regional Development – Ministry of Energy and Energy
Resources, Ordinance no 14 - Technical rules and norms for design, construction and utilisation
of the projects and the facilities for generation, transformation, transmission and distribution of
electrical energy, June 15, 2005
Installed capacity of
>100 25-100 10-25 5-10 <5
the hydropower project, MW
Permanent Main structures І ІІ ІІІ ІV V
Category
structures
(class) of the Secondary structures ІІІ ІІІ ІV ІV V
structure
Temporary structures V V V V V
The category of the dams is defined as the higher one of the following :
- the category defined according to Table A4-2 as a structure (object, element) of a
hydropower project;
- the category defined according to Table A4-3 depending on the foundation conditions.
If the component is part of a scheme of complex purpose (hydropower, water supply,
irrigation, etc.), its category is defined depending on the purpose of the project requiring the
highest category of the respective structure.
The class defined can be increased, if the damage or the failure of the structure can threaten
people’s life or can cause catastrophic consequences to the populated areas or infrastructure
and/or are related to trans-border impact.
Category of dam
Dam type Foundation type І ІІ ІІІ ІV
Height of Structure, m
Rock >100 50-100 15-50 <15
Sand, stones, clays in hard or
> 70 35-70 15-35 <15
Embankment dams semi-hard condition
Saturated clays in plastic
>40 20-40 10-20 <10
state, fine sands
Concrete and reinforced Rock >100 50-100 20-50 <20
concrete dams, underwater
Sand, stones, clays in hard or
structures of the HPP/SHPP >40 20-40 10-20 <10
semi-hard condition
powerhouses, fish passages,
retaining and other Saturated clays in plastic
>20 15-20 10-15 <10
structures that create head state, fine sands
14 15
5. CANADA
In Canada, the dams are classified using the Hazard Potential Classification (HPC) system.
The HPC system categorizes dams according to the potential hazards presented by the dam.
The hazard potential is determined through an assessment of the greatest incremental losses
that could result from an uncontrolled release of the reservoir due to the failure of a dam or
its appurtenances. Potential incremental losses are to be assessed with respect to the
population at risk, loss of life, the environment and cultural values and economic aspect at
the dam site, upstream, downstream, or at other areas influenced by the dam.
Table A5-1 outlines the criteria for determining the dam classification.
14
Canadian Dam Association, Dam Safety Guidelines, 2007
15
Ontario Ministry of Natural Resources, Classification and Inflow Design Flood Criteria,
Technical Bulletin, August 2011
Note: where several dams are situated along the same watercourse, consideration must be
given to the cascade effect of failures when classifying the structures, such that if failure of
an upstream dam could contribute to failure of a downstream dam, then the classification of
the upstream dam must be the same as or greater than that of the downstream structure.
Flood Analysis
Statistical analysis is required for estimating the flood peaks and volumes associated with a
range of annual exceedance probabilities. In addition to the peaks, the volumes and the
associated hydrographs for the floods of interest are usually required for reservoir routing or
dam breach and downstream channel routing. This analysis is done on a seasonal basis and
is of greater significance for storage reservoirs that have large fluctuations in water levels
and are designed to capture spring runoff. For run-of-the-river facilities, only the peak
annual flood is usually required.
Flood statistics are subject to a wide margin of uncertainty, which should be taken into
account in decision-making. In particular, the following should be noted:
Results obtained may vary significantly from one statistical distribution to another, and
no reliable method exists for the selection of the most appropriate distribution. This task
relies entirely on the hydrologist's judgment.
The evaluation of the highest floods on record often depends on the extrapolation of the
rating curve at the station, which may be subject to a large degree of uncertainty.
Beyond an AEP in the order of 1/500 year, statistics give only an "order of magnitude"
estimate. For the purpose of flood evaluation in the dam safety process, extrapolation
beyond the 1/1000 year flood is discouraged.
Evaluation of the confidence limits on the statistical estimate is recommended.
Some canadian provinces, such as Ontario, as based their dam classification system and the
determination of the design flood on the CDA guidelines by adapting the approach to their
particluar needsd and concerns.
6. CANADA – Quebec 16 17
In the province of Quebec, the Dam Safety Regulation (2002) applies to all dams governed
by the Dam Safety Act (2000).
16
Gouvernement du Québec, Loi sur la sécurité des barrages, 2000
17
Gouvernement du Québec, Règlement sur la sécurité des barrages, 2002, Décret D. 300-2002,
2002, G.O. 2, 2043
If the dam failure consequences were assessed on the basis of a dam failure analysis, the
category to be considered for the purposes of the table is the highest consequence category
resulting from the examination of dam failure scenarios in flood conditions.
The safety check flood corresponding to the consequence category of the dam can be
determined based on Table A6-2.
Table A6-2 Canada-Quebec - Safety check flood vs. consequence category of the dam
Safety check flood
Consequence
(recurrence
category
interval)
Very low 1:100-yr
Low 1:100-yr
Moderate 1:1 000-yr
High 1:1 000-yr
1:10 000-yr or ½
Very high PMF
(highest value)
Severe PMF
The crest of an “erodible dam”18 at its lowest point must not be less than one meter above
the safety check flood level, unless the owner demonstrates to the Minister’s satisfaction that
all hydrologic and hydraulic uncertainties and flood management uncertainties have been
taken into account in the determination of the safety check flood. Any impervious
component of an erodible dam must be at least as high as the safety check flood level.
For a dam of which at least half the inflow is controlled by another dam operated upstream,
the safety check flood is the greater of the safety check flood determined or the lesser of the
10 000-year flood and the inflow equivalent to the discharge capacity of the upstream dam
together with the local inflows.
Dam classification
Every dam must also be classified on the basis of the degree of risk it poses to persons and
property, measured by multiplying the numerical value of its vulnerability (V) by the
numerical value of the potential consequences of a dam failure (C), to which “P” is the
assigned value in the formula
P=VxC
The vulnerability (V) of a dam is measured by multiplying the arithmetic mean value of the
constant physical parameters by the arithmetic mean value of the variable parameters. The
constant physical parameters to be considered are the dam height, dam type, impounding
capacity and dam foundation type.
The variable parameters to be considered are :
(1) the dam age, which is the number of years since its construction or, as the case may
be, as determined by the engineer in charge of the safety review on the basis of the
useful life of the dam ;
(2) the seismic zone in which the dam is located according to the seismic zone map;
(3) the dam condition, which is assessed by considering the physical state and structural
condition of the dam, the quality and effectiveness of maintenance, aging, possible
effects of external factors such as frost or earthquakes and any dam design or
structural defects. At the completion of the assessment, the dam condition is rated
“very good”, “good”, “acceptable” or “poor or unknown”; and
(4) the reliability of the discharge facilities of the dam, which must be capable of passing
the inflow design flood. The reliability is assessed on the basis of the design of the
discharge facilities and the procedures established by the owner to ensure that they
operate effectively during floods. At the completion of the assessment, the reliability
of the discharge facilities is rated “satisfactory”, “acceptable” or “unsatisfactory or
unknown”.
18
“erodible dam” means a dam with an earthfill or rockfill component that is not designed for
overflow and the erosion of which would cause a dam failure in a flood
7. CHINA
In China, dam classification is based on the main characteristics of the system and on the
potential consequences on the regional and national economy. Tables A7-1 to A7-3
summarize the elements considered for the dam classification in China. (Liu, 2002).
Notes
- The storage capacity of reservoir means the storage of reservoir below check flood level
- The irrigation and waterlogged areas refer to design areas
- The rank of tide prevention projects may be defined referring to the stipulations for
flood prevention. Where disasters of tide are very serious, the rank may be raised
properly
- The importance of water supply works are defined according to their scale, economic
and social benefits.
Notes
1. Permanent structures are the structures used for operation of the project, and are
divided into two categories in accordance with their importance : Main structures that
Table A7-3 China - Design flood and check flood criteria for permanent structures
Notes
• The standards of powerhouse and irrigation structures (Classes 4 and 5) may be lowered
according to actual situations
• For Class 1 embankment dam, PMF should be considered if its failure will cause
catastrophe in downstream area, and for Classes 2 to 4 embankment dams, the check
design floods may be raised by one grade
• For concrete dams that cause serious damage in case of overtopping, 1000-yr flood may
be adopted as check design flood after examination and approval by competent
authorities
• For low water head structures and the structures that do not cause serious damage,
check design flood may be lowered by one grade after examination and approval by
competent authorities
19 20
8. CZECH REPUBLIC
The issue of dams‘ safety in Czech Republic is projected on many legal acts of several
levels, from laws, government orders and decrees, through technical standards and
methodical instructions, up to specific developed documents, like flood plans, programmes
for dam safety supervision, operational rules and others.
Four classes of dams are defined. The class of dams and the corresponding design flood
depends of the following consequences related to the dam failure :
Number of persons at risk;
Possibility of loss of life;
Expected damages to residential, industrial and public structures;
Economic losses;
Environmental damages.
The design flood can vary for the 1:10 000-yr flood for a class I dam to a 1:100-yr flood for
a class IV dam and even a 1:20-yr flood if the consequences are negligible. Table A8-1
presents a summary of the criteria considered for the dam classification.
Damages
Persons
Class Loss of life Economical Environment
at risk
aspect consequence
Large Extensive
I > 1 000 Whole country
amount damages
Damages to
II > 100 Anticipated essential Autonomous district
structures
Important
III > 10 Possible damages but Autonomous region
could be repair
Limited material
IV < 10 Improbable Negligible
damages
The following rules deal with specific issues of dam supervision related to hydrological
conditions, and mainly concerning flood load:
Act No. 254/2001 Coll. – The Water Act
The Water Act deals with all kind of water issues. Its purpose is to protect superficial water
and groundwater, to make a cost-effective use of water sources, to ensure the safety of dams,
19
R. Kucera – Personal communication - 2012
20
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European
Context, Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
Technical standard No. 75 2340 Designing Dams – Main parameters and equipment
This is a basic technical standard in the field of dam construction. The standard is valid for
design of dams (weirs, safety spillways, water outlets, hydropower constructions, etc.) and
their equipment. The standard defines basic general requirements for a project, and deals
with the main parameters of the dam and the safety premises. The requirements for
hydraulic dimensioning and the principles for design of functional constructions and
equipment, concern to safety spillways (whose capacity is determined by the water
management project and by fulfilling the conditions of the Technical standard No. 75 2935),
as well as to bottom outlets and to intake equipment. The Czech State Standard No. 75 2340
determines, that the basic role of a dam design, is to ensure its safety at supposed limit flood
situations, according to Technical standard No. 75 2935, including the determination of
emergency measures in case of extreme situation.
21
9. FINLAND
In Finland, three classes of dams are considered, based on the possible consequences of the
dam failure on the danger on human life or their health and/or the danger on environment
and properties. These dam classes are presented hereafter :
Class P If in the event of an accident, the dam may endanger human life or health or
seriously endanger the environment or property. The dam classification is
based on a hazard risk assessment.
Class N If in the event of an accident, the dam represents a smaller risk than a class P
dam, but cannot be considered an O dam.
Class O If in the event of an accident, the dam constitutes only a minor hazard risk.
The design flood can vary between the 1:5 000-yr to the 1:10 000-yr flood for a class P dam,
from the 1:500-yr to the 1:1 000-yr flood for a class N dam and from the 100-yr flood to the
500-yr flood for a class O dam.
21
E.Loukola & T.Maijala, New dam safety guidelines in Finland, Dam Safety, Berga (ed.)© 1998
Balkema, Rotterdam, ISBN 90 5410 974 2
22
10. FRANCE
Until 2007
Until 2007 the design flood standard was set according to the factor H2√V where H (m) is
the height of the dam and V (hm3) is its storage capacity.
The parameter H2√V does not appear to have any particular theoretical basis but expresses
the need to consider store volume as well as impounded water depth as a contributor to the
overall hazard posed by the dam. Table A10-1 shows the design flood related to the height-
volume parameter.
Table A10-1 France - Design flood based on the H2√V parameter – Until December 2007
H2√V <5 5 to 30 30 to 100 100 to 700 > 700
Probability of the design flood [%] 1 0,5 0,1 0,05 0,01
Flood return period (years) 100 200 1 000 2 000 10 000
When frequency analysis is required, the Gradex method is used, i.e. and apply on
precipitation data instead of runoff data.
Revised approach
Dam classification
According a decree issued in December 11th 2007, a revised approach is considered for the
classification of the dam and the determination of the design flood.
The revised approach for the dam classification is based on the height of the dam and on the
H2√V parameters as shown in the following table.
Table A10-2 France - Design flood based on the H2√V parameter – From December 2007
Class A B C D
H (m) and V H ≥ 10 and H ≥ 5 and
H ≥ 20 H≥2
(hm3) H2√V ≥ 200 H2√V ≥ 20
The parameter H2√V is considered by the regulation as an indicator of the general potential
of risk.
Design flood
Recent recommendations (June 2013) propose a double approach for the design of spillways.
These recommendations insist on the fact that they are proposed for French dams, taking
into account the general regulations, the existing dams, the general organization. For class D
dams, the recommendations are limited to dams with a capacity greater than 50 000 m3.
22
Radzicki, K., Szczesny, J., Tourment, R., Comparison of laws, procedures, organizations and
technical rules for dams and dikes safety in Poland and France, Selected problems of water
engineering, Politecknika Krakowska Cemagref: results of cooperation, Krakow, POL, 9-11
October 2003
In some cases, it is possible to consider that, for a given hydrological situation, the
additional downstream impact on persons and property due to dam failure is negligible.
Provided that this assumption is demonstrated, this can allow an increased target probability
for the dam in question (see 5.6).
Flood estimates for an overrun probability lower than 10-4 are considered very uncertain. At
this stage it is recommended to raise the flood likely to occur every 10 000 years by a
multiplying factor on the flows: of the order of 1.3 for Class A dams and of 1.15 for Class B
dams.
Audits are conducted by taking account of rolling by the reservoir and the upstream
structures and for an initial retention level equivalent to the maximum level under a normal
operation situation (NWL).
Auditing is to be carried out on the basis of two sets of assumptions:
23
Class A comprises in particular all the dams with a height greater than 20 metres above the
natural ground, hence with very different hazard potentials. Reducing the target probability for
the highest of these should not be ruled out.
11. GERMANY
Classification of dams
According to the German Technical Standard dams are classified in two classes
Class 1 Large dams : damming structure higher than 15 m or reservoir’s storage capacity is
greater than 1 hm3;
Class 2 Middle/small dams that do not conform with class 1 dam conditions.
Design Floods
According to the German Technical Standard three design floods are specified:
- Design flood 1 (BHQ1) is used for spillway design and safety against spillway
overloading. In case of this flood event no disturbance of structural safety, service
ability or durability are allowed. The flood peak discharge of BHQ1 has an exceeding
probability of 10-3 (return period: T = 1 000 years) for dams of class 1 and 2x10-3 (T =
500 years) for dams of the second category.
- Design flood 2 (BHQ2) is used to ensure that no disturbance of structural safety exists
caused by extreme floods. Here damages of structural parts, operating or measuring
equipments are accepted. The discharge of BHQ2 has an exceeding probability of 10-4
(return period: T = 10 000 years) for class 1 (class 2: exceeding probability of 2x10-4 (T
= 5 000 years).
- Design flood 3 (BHQ3) is applied to design the regular flood storage capacity (see figure
1). It characterizes the flood protection ensured by the dam. It has to be selected
according to downstream protection requirements downstream.
It is possible but not mandatory to consider flood routing within reservoirs. For design
problems without flood routing, the flood peaks are sufficient to specify the design floods.
The Technical Standard recommends to consider flood hydrographs, whose shapes and
volumes result in maximizing the reservoir’s storage content during floods, even if the peak
stays below the thresholds mentioned above.
To consider residual risks the Probable Maximum Flood (PMF) can be used.
The initial and boundary conditions of flood design are specified in Table A11-1. Figure
A11-1 shows the storage segments and the thresholds of water levels which are used to
specify them.
Table A11-1 - Germany - Initial and boundary conditions for flood design reservoirs
Impact start /
Design flood 1 and 3 Design flood 2
Boundary conditions
Design flood discharge BHQ1 and 3 BHQ2
Effect of retention To be considered, in general, in calculations
Water level at beginning Water level in general equal to design water level
of flood event (Zs, see figure 1); seasonal differences of design water levels
have to be taken into account
Freeboard Compliance of freeboard f1 Compliance of freeboard f2
ahead of the flood storage ahead of the flood storage
level 1 (ZH1) level 2 (ZH2)
Pre-release (before Allowed by bottom-outlets if Allowed by bottom-outlets and
reaching top of joint use (n-1)-rule is adhered and through flood spillways
capacity ZV, see figure 1) through flood spillway under operating-outlets can be
usually after starting to fill compliance with (n-1) and considered
the exclusive flood control (n-a) rule If pre-release is taken into
If pre-release is taken into account, it has to be verify that
account, it has to be verify the resulting discharge remains
that the resulting discharge below the maximum allowed
remains below the maximum downstream discharge
allowed downstream
discharge
Simultaneous release Allowed by bottom-outlets if Allowed by bottom-outlets and
(after reaching top of joint (n-1)-rule is adhered and through flood-spillways
use capacity ZV) through flood-spillway under operating-outlets can be
compliance with (n-1)- and considered
(n-a)-rule
Flood alleviation Flood spillway operative Flood spillway operative
multi-gate spillways which (without constraints)
are under water level
constantly have to be
considered with compliance
to the (n-1)- and (n-a)-rule
Emergency alleviation No emergency alleviation Allowed (under specific
allowed conditions and if existent)
Resulting water level Flood storage level 1 (ZH1) Flood storage level 2 (ZH2)
IF2 ZH2 f2
IF1 f1
ZH1
IAHR2 7
IAHR1
ZV
ZS IGHR
6
5 1
IBR
2
ZA
ZT IR 3
IT
Legend
1 damming structure ZV top of joint use capacity
2 operating-outlet ZS design water level
3 bottom-outlet ZA drawdown elevation
4 crest elevation ZT deepest drawdown elevation
5 flood storage capacity 1 (BHQ1) IF2 head space BHQ2
6 flood storage capacity 2 (BHQ2) IF1 head space BHQ1
7 spillway crest IAHR2 surcharge flood storage 2 (BHQ2)
f1 freeboard 1 (BHQ1) IAHR1 surcharge flood storage 1 (BHQ1)
f2 freeboard 2 (BHQ2) IGHR flood absorption storage
ZK crest water level IBR operating storage
ZH2 flood storage level 2 IR reserve storage
ZH1 flood storage level 1 IT dead storage
24
12. INDIA
According to the “Guidelines for fixing Spillway Capacity” (IS: 11223:1985), the
hydrologic design criteria for fixing the spillway capacity, depends of the classification of
the dams based on gross storage behind the dam and the hydraulic head at the minimum tail
water level. The classification of the dams and the inflow design flood depends of the most
critical of these two parameters:
24
Mathur, SH N.K., & Singh B., HYDROLOGIC SAFETY OF DAMS IN INDIA, India Water
Week 2012 – Water, Energy and Food Security : Call for Solutions, 10-14 April 2012, New Delhi
25
Central Water & Power Commission – Ministry of Irrigation & Power – Government of
India, Estimation of Design Flood – Recommended procedures.
13. IRELAND
In Ireland, there is no legislation or national standards in relation to design floods for dams.
Owners of dams in Ireland have applied different design flood standards to their dams. The
Electricity Supply Board (ESB), a state utility, owns and operates most of the large dams in
Ireland. The following describes the standard that has been applied to Electricity Supply
Board’s dams.
During floods the top priority is the proper management of the flood to avoid any risk to
dam safety. ESB has categorized its dams in a similar manner to the Institution of Civil
Engineers (UK) guide; “Floods and Reservoir Safety”. Dams are categorised based on the
consequences of a breach.
Two categories are considered. The first one, “Category A”, includes dams where a breach
could endanger lives in a community. The following are the design flood standards used for
Category A dams:
Ability to pass the 10 000 year flood without overtopping the dam crest when all
spillway gates are in operation and
Ability to pass the 1 000 year flood with one spillway gate unavailable and with a
freeboard allowance for wave run-up
For lower categories, i.e. for negligible risk to human life in the case of a dam breach, the
following design flood standard is used:
Ability to pass the 1 000 year flood with one spillway gate unavailable and with a
freeboard allowance for wave run-up
The above standards and their associated design flood levels are also applied to any
embankments, including headrace canals, associated with each dam.
Where necessary to safely pass the design floods, reservoir or spillway capacities were
increased. In addition, downstream channel protection and improvement works were carried
out as required to ensure the safe passage of extreme floods.
26
14. ITALY
The spillways of the dam have to be dimensioned for the wave with peak flood discharge
corresponding to the return period of 1 000 years for concrete dams and 3 000 years for the
embankment dams, taking into account the regulation effect developed by the reservoir.
The flood discharge will have to be evaluated with probabilistic methods founded on the
complete rainfall and hydrometric information, deducible by the historical series and their
updates, for the watershed subtended by the dam section.
In absence or lack of information, it will be possible to make reference to data of adjacent
watersheds with homogeneous hydrological characteristics, using appropriate and justified
calculation methodologies.
The verifications have to include also the estimate of the flood discharge with return period
of 50, 100, 200 and 500 years.
The spillways of the dam have to be dimensioned so that the net freeboard respects the rules
presented on Table A14-1.
For the intermediate values of the height of the dam, the net freeboard is determined by
interpolation.
For every dam, the return period of the flood event corresponding to a net freeboard of 0 m
has to be estimated.
For concrete dams, the spillway can consist of free sills or gated sills. For embankment
dams, the spillway has to consist of free sills or partially gated sills; in the second case, the
free sills will have to be dimensioned to discharge, at the maximum water level, at least 50%
of the design flow, taking into account the regulation effect developed by the reservoir.
In case of gated sills, the hypothesis of non-operation will have to be considered for at least
50 % of the gates, in case of embankment dams, and of at least 20 % of the gates, in case of
concrete dams, verifying that in such condition the net freeboard comes down, at worst, to
one third of the above mentioned values. The calculation can consider the contribution of the
discharge which flows over the closed gates, if these can be overtopped.
26
Email from E. Baldovin, 2013-04-08
27
15. JAPAN
In Japan, the inflow design floods for dams higher than 15 m are stipulated in the Structural
Standards for River Protective Facilities (Cabinet Order), which was drawn up on the basis
of River Law.
For a concrete dam, the inflow design floods must be taken as the largest value among the
following three cases:
The 200-year flood at the dam site;
The maximum flood discharge observed at the dam site, and
The maximum flood discharge that can be expected at the dam site based on the
maximum experienced flood discharge in the basins with similar hydrological
conditions or climate. The discharge is estimated based on the Creager equation as
shown in Figure A15-1.
For an embankment dam, the design flood will be 1.2 times the discharge values for a
concrete dam. The return period of the design flood for an embankment dam is not estimated
but it is believed that it will be around 1:1000 years. Figure A15-1 presents the specific
discharge to be considered for each area of Japan and Figure A15-2 illustrates the location of
each area.
The minimum freeboard of the dam is estimated taking into account the type of dam, the
wave height caused by wind and/or earthquake and an allowance for gate operation. Table
A15-1 presents the approach considered to estimate the freeboard of dams in Japan.
Hf Freeboard (m)
hw Wave height due to wind (m)
he Wave height due to earthquake (m)
ha Allowance for gate operation (m)
(with gate: 0.5m, without gate: 0m)
27
Mathur, SH N.K., & Singh B., HYDROLOGIC SAFETY OF DAMS IN INDIA, India Water
Week 2012 – Water, Energy and Food Security : Call for Solutions, 10-14 April 2012, New Delhi
q C A A
0 . 05
Area No. 1
Where,
q : Spesific discharge (m3 /sec/km2)
A : Catchment area (km2)
C : Coefficient for each area
Boundary of areas
1 Hokkaido
2 Tohoku
4 Hokuriku
2
Boundary of areas
8 Sanin
3 Kanto
11
9
6
5 Chubu
10 7 South Kii
South Shikoku
6 Kinki
9 Setiuchi
11 Kyushu, Okinawa
Reference
(1) Description of the law of Ministry of Construction Ordinance for Structural Standard
for River Administration Facilities, 2007
28
16. NEW ZEALAND
The New Zealand guidelines are intended for dams where the potential impacts of failure
include loss of life and damages beyond the owners property. Dams which would be
classified in the very low category are generally outside the scope of these Guidelines. The
dam classification is based on the potential incremental consequence of a dam failure, i.e on
the number of fatalities and the socio-economic, financial and environmental impact.
The dam height and reservoir volume parameters while useful for an initial screening of
potential impact classification should not control the potential impact classification where
the consequences of a dam failure are not consistent with such an initial screening. For
example, a 10 - 15 metre high dam whose failure can lead to fatalities should be classified
with a high potential impact. Similarly, a 25 metre high dam whose failure would not cause
fatalities and where damages are moderate can be classified as Low Potential Impact.
Very Low Potential Impact dams are likely to be those that are exempt under the Building
Act, being those dams that retain not more than 3 metres depth, and not more than 20 000
cubic metres volume of water. The dam height would generally be less than 4 metres.
Low Potential Impact dams are likely to have dam heights less than 10 metres, and less than
6 metres if the storage exceeds 50 000 cubic metres. For this category of dams the minimum
IDF is usually between a 1 in 100 and 1 in 1 000 AEP.
Medium Potential Impact dams are likely to have dam heights in the 10 - 20 metre range,
but not exceeding 15 metres if the stored volume of water exceeds 1 000 000 cubic metres.
For medium potential impact dams the minimum IDF is usually between a 1 in 1 000 and 1
in 10 000 AEP.
If the dam height and storage exceeds those for other categories then a high potential impact
classification is probable. For this category of dams, the minimum IDF is usually between 1
in 10 000 AEP and the PMF. The PMF is usually selected if a large number of fatalities
would result from failure of the dam. If collapse of the dam has little incremental damage
potential the design flood may be reduced.
28
The New Zealand Society on Large Dams, New Zealand Dam Safety Guidelines, 2000
17. NORWAY 29
The Norwegian Dam Safety Regulation considered five classes of dams based on the
possible consequences after a dam failure. Based on this classification, a design flood and a
safety check flood is defined. The approach for the PMF is to use hydrological modeling
based on pessimistic assumptions of extreme precipitation, heavy snowmelt and saturated
soils to generate the PMF. Table A17-1 presents a summary of the dam classification and
the corresponding design flood and safety check flood.
Safety check
Dam class Classification criteria Design flood
flood
H < 2m;
0 V < 10 000 m3 1:200-yr NA
minimal consequence
Low consequence (no PMF or
1 1:500-yr
permanent dwelling) 1,5x1:500-yr
PMF or
Medium consequence
2 1:1 000-yr 1,5x1:1 000-
(1 to 20 dwellings)
yr
High consequence
3 1:1 000-yr PMF
(21 to 150 dwellings)
Very high consequence
4 (more than 150 1:1 000-yr PMF
dwellings)
The application of the design flood and the safety check flood depends of the type of dams.
For concrete dams :
Class 0 to 4 no overtopping is allowed for the design flood;
Class 1 to 4 overtopping is allowed for the safety check flood, subject to verification
of the resistance to erosion of the foundation
For embankment dams with central core :
Class 0 no overtopping is allowed for the design flood;
Class 1 to 3 the freeboard must be at least equal to the water level during the design
flood, plus the wave height and a safety margin of 0.5 m. For the safety flood, the core
level could be slightly exceeded as well as the crest of the dam if the downstream face
can resist to erosion;
Class 4 same criteria than the previous case, but minimal difference of 6 m between
the maximal normal operating level and the crest of the dam
29
Norwegian Dam Safety Regulation, Water Ressources Act, Jan 2010
Spillway availability
For gated spillway, the previous criteria defined for the safety flood are applied to the design
flood with the following operating constraints:
• One gate non operational for a system with one to three gates;
• Two gates non operational for a system with four to six gates;
• Three gates non operational for a system with seven gates or more; ;
Clogging of the spillway by floating debris shall be considered in the evaluation if the
watershed is heavily forested. Under these conditions, a reduction of 25% of the discharge
capacity of the spillway is used for the design flood.
30
18. PANAMA
In Panama, the dam classification is based on risk evaluation of the incremental losses due
to a of dam failure (LOL, economic losses and environmental losses). Three categories are
considered :
High Potential Risk: These dams are classified as category "A" and are those where failure
or misoperation will most probably cause loss of human life and/or major economic or
environmental damages.
Significant potential risk: These dams are classified as category "B" and are those where
failure or misoperation results in an unlikely loss of life, but can cause economic loss,
localized damage to the environment, interrupt and communications facilities. These dams
may be located in rural areas but with population and associated infrastructure.
These Low hazard potential: dams are classified as category "C" and are those where
failure or misoperation results in an unlikely loss of human life and low economic and
environmental losses. Significant losses are limited to structural closure.
Table A18-1 shows a summary of these criteria and the corresponding design flood:
Category A B C
Risk High Significant Low
Certain (in one or Uncertain (rural Not expected (due to
more residential, location with few the rural location
Direct loss of lives commercial or homes and only without housing)
industrial) transient or industrial
development)
Interruption of Interruption of Any break in
Loss of essential essential facilities essential facilities service, repairs of
services and roads to critical and roads damage is simple or
levels quickly repairable
Extensive on public Greater public Private agricultural
Loss properties and private facilities involvement and land, equipment and
private facilities buildings isolated
High cost of Important mitigation Damage minimal
Environmental mitigation or is required incremental
losses impossible to
mitigate
Qd = %PMF or Qd = % CMP or Qd = Q100
Design flood Q1 000 to Q5 000 Q 500 to Q1 000 or
ERA
30
Autoridad nacional de los servicios Publicos - Republica de Panama. NORMAS PARA LA
SEGURIDAD DE PRESAS, Septiembre de 2010
31 32
19. POLAND
In Poland, dams and dikes are classified into one of four classes of hydraulic structures. A
structure belongs to the class which has the lowest indicator pointed by any of the nine
indicators. The five most important indicators are related to public safety and are presented
hereafter on Table A19-1.
The design standards are set by law as a base flood (design flood) which is used to set the
spillway capacity for normal operation without damage and a control flood (safety flood)
which must be passed without failure but some damage may be experienced. Table A19-2
presents the base flood and the control flood for which of the four dam classes depending if
the dam will be destroyed in case of overtopping
31
Radzicki, K., Szczesny, J., Tourment, R., Comparison of laws, procedures, organizations and
technical rules for dams and dikes safety in Poland and France, Selected problems of water
engineering, Politecknika Krakowska Cemagref: results of cooperation, Krakow, POL, 9-11
October 2003
32
Central Water Commission – Ministry of Water Resources – Govt. of India, Development of
Hydrological Design Aids (Surface Water) under Hydrology Project – II – State of the Art
Report, July 2010, 362 pages
Table A19-2 - Poland – Flood return period vs. type of hydraulic structure
33 34
20. PORTUGAL
The Portuguese Dam Safety Regulation (PDSR) applies to large dams (H ≥ 15 m following
the ICOLD definition) and to smaller dams with reservoir capacity larger than 100 000 m3
independently of the height of the dam 35. A revised version published in October, 2007
replaced the previous regulation published in 1990.
According to the PDSR, the dams hazard classification considered three classes of dams :
Class I(high hazard) -dams whose failure will affect 25 or more residents.
Class II(significant hazard) -dams whose failure will affect at least one resident or
existence of significant infrastructures, environmental patrimony difficult to recover
destruction caused by dam-break wave or storage installations for hazardous substances.
Class III(low hazard) -dams whose failure will not affect any human live nor significant
infrastructures, environmental patrimony difficult to recover destruction caused by dam-
break wave or storage installations for hazardous substances.
The national regulation considers downstream risk as the basis for the selection of the design
flood as well as the dam type and size as shown on Table A20-1. The return period for the
design flood varies between 1:500 and 1:10 000-yr. A comparison with the PMF value is
also recommended when the return period exceeds the 1 000-yr flood.
Hydrologic studies shall be completed with the evaluation of the inundated areas and the
downstream flow propagation based on dam-break scenarios.. For the two major risk level
dams (A and B), a dam-break flood downstream wave propagation study shall be done.
The potential risk definition grades used on those regulations are similar, based on loss of
life and downstream material damages.
The regulation states that the spillway flood design shall attend to flood attenuation due to
the reservoir and the ones upstream. It also states that the spillway capacity shall be
sufficient to discharge the flood, without any other auxiliary discharge appurtenances.
33
Pinheiro A. N., Portuguese Dam Safety Legislation - Existing framework and practical aspects of
its implementation, Workshop on Regulatory Framework on Dam Safety, June, 2011
34
Maia, R. & Ribeiro, A.A., Dam safety criteria on international rivers: Portugal and Spain case,
Dam Safety, Berga (ed.) © 1998 Ba/kerna, Rotterdam, /SBN 90 5410 9742
35
The dam projects that are not covered by the Portuguese regulation shall refer to a specific
regulation on small dams.
36
21. ROMANIA
In Romania the standard STAS 4068/2 – 87 establishes the maximum flows whose
evacuation must be carried out in normal operation conditions (design flows) and in special
operation conditions (check – up flows) that take into account the classification of hydraulic
structures in classes of importance (standard STAS 4273 – 83).
The class of importance of a dam is established depending on dam height, stored volume
and the output for which it is used (economic measurement unit expressed as installed power
in case of dams for hydroelectricity and flows for water supply, irrigated area etc., in case of
dams for other uses) working duration and degree of importance within the hydro-technical
scheme
Table A21-1 shows the characterization of constructions and hydraulic structures, the
corresponding annual probabilities of exceeding the maximum flows and their IDF
equivalent.
Annual IDF
Class of Characterization of constructions and
probability of equivalent
importance hydraulic structures
exceeding (%) (yr)
Hydraulic structures of exceptional
importance whose damaging has
I 0.01 10 000
catastrophic consequences or where the
interruptions in working are unacceptable
Hydraulic structures of special importance
whose damaging has serious effects or
II 0.1 1 000
whose functioning can be only interrupted
exceptionally, for a short time
Hydraulic structures of medium
III importance whose damaging endangers 0.5 200
social – economic objectives
Hydraulic structures whose damaging has
IV little influence on other social – economic 1* 100
objectives
* Based on a technical and economic justification approved at the same time with the
project, the check–up at special conditions of operation can be dropped
Romanian standards take also into account the dam height and the degree of inhabitance and
economic development, but neglect the vulnerability and the accepted risk
For a number of years there is a discussion among specialists for the updating of the
standards but it was not yet finalized and implemented. So the design activity continues to
use the standards in force but will adopt the new regulations as soon as the legislation will
be updated.
36
Communication from O. Gabor
37
22. RUSSIA
In Russia, the Building Code 33-01-2003 "Hydraulic Structures. Basic Provisions" defined
the design flood for four classes of dams. The dam classification is based on the main
characteristics of the reservoir and structures, such as the reservoir volume and utilization,
the dam of dam and the type of foundation, ..., as well as their utilization. Tables A22-1 and
A22-2 are used for the classification of the dams
Class of structures
Hydraulic structures and their
parameters
1 2 3 4
Impoundments with storage volume of 200 to 50 to
>1000 <50
(106 m3) 1000 200
Hydraulic structures for conventional 300 to 10 to
>1000 <10
HPPs, pumped storage plants, tidal 1000 300
37
Asarin. A.E. Estimation of the design flood in the design of Russian Dams, Hydropower & Dams,
Issue Three, 2007
If a dam failure could cause human casualties, the category of the hydraulic structure is
classified as "first class". Dams where a failure could cause damage to more than 3000
people living permanently downstream are also classified as 'first class' structures.
The most important factor affecting the quality of the engineering decisions taken for the
design of a spillway is the accuracy of hydrological data, which depends on the duration of
the available hydrological records. In Russian design practice, to consider this factor in
developing the maximum inflow to a high risk structure, the guarantee correction is
introduced to the maximum flow of exceedance probability of 1:10 000-yr. The shorter the
hydrological record series is, the larger this correction will be.
For dams in cascade, the design flow for the proposed dam projects should be specified
taking into account its structural category but should not be lower than values equal to the
sum of the discharge capacities of the upstream dam and the maximum lateral inflows at the
river stretch.
38 39
23. SOUTH AFRICA
Dam classification
In South Africa, dams at least five meters high are classified depending of their height and
their potential risk. The size classification of the dams, as defined in the SANCOLD
guidelines, is presented in Table A23-1 :
The hazard classification of a dam is based on the potential loss of life and the economical
damages following the dam failure as shown in the following table :
Based on these two types of classification, the dam the category of each dam is determined
according to Table A23-3.
Design Flood
In South Africa, envelope curves are used for the determination of design floods. Maximum
flood peaks observed in a hydrologically homogeneous region are plotted against catchment
38
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European
Context, Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
39
Cullis, J. Gorgens A. Lyons S. Review of the Selection of Acceptable Flood Capacity for Dams in
South Africa in the Context of Dam Safety, WRC Report no 1420/1/07
The SED is the peak discharge which is used to evaluate the adequacy of the spillway
system of a new or existing dam under extreme flood conditions. Substantial damages may
result from the occurrence of the SED, but the design must be such that the dam will not fail.
The SED is defined in terms of the RMF, which is calculated according DWA Technical
Report 137 (Kovacs, 1988). The SED for different categories of dams is determined by
scaling the RMF one step up or down depending of the K-value region in which the dam site
is located.
The recurrence for the RMF related floods depends of the probability distribution used.
Table A23-6 shows an estimate of the recurrence.
Table A23-6 South Africa - Estimated recurrence intervals (RI) for the RMF related floods
Probability Median RI Lower 95%
Design flood
distribution (years) RI (years)
RMF 6 000 100
Log-Pearson
RMF+∆ 18 800 200
type III
RMF-∆ 1 200 50
RMF 3 000 200
GEVpwrn RMF+∆ 6 000 400
RMF-∆ 900 50
40 41
24. SPAIN
In Spain the dams are classified according to size, potential hazard and type of dam. Two
different floods are used, the design Flood and the extreme Flood or safety check Flood. For
dams in the highest hazard class the 1000-year flood is used as design flood. The extreme
Flood for dams in the highest hazard class varies between the 5 000 and 10 000 year flood,
where the higher values should be used for embankment dams.
Table A24-1 presents the dam classification criteria in Spain (Maia & Ribeiro) and Table
A24-2 the IDF corresponding to each category.
40
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European
Context, Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
41
Maia, R. & Ribeiro, A.A., Dam safety criteria on international rivers: Portugal and Spain case,
Dam Safety, Berga (ed.) © 1998 Ba/kerna, Rotterdam, /SBN 90 5410 9742
42
25. SWEDEN
In Sweden, the determination of design floods is based on statistical methods or on
simulations, using hydrological models. Both methods contain elements of uncertainty
which should be accounted for when the results are evaluated. The selection of a time period
providing the basis for the calculations is most important and should be taken into special
consideration.
When determining design floods for dams, a classification according to flood design
categories is applied. This classification is made according to which consequences a dam
failure could produce as a result of large floods (Table A25-1). The guidelines are not
applicable to dams that in case of a dam failure would not cause damage to anyone, except
the dam owner.
Table A25-1 Sweden - Flood design categories when determining design floods
(*) The approach is to use hydrological modeling based on pessimistic assumptions of extreme
precipitation, heavy snowmelt and saturated soils. Comparisons with frequency analysis
indicate that floods calculated this way have return periods of more than 10 000 years, at an
average.
42
Swedenergy, Svenska Kraftnät and SveMin - Swedish Guidelines for Design Flood Determination for
Dams New edition 2015, ISBN 978-91-7622-198-3
43
26. SWITZERLAND
In Switzerland, the design flood (Qb(t)) is the 1:1 000-yr flood, The design flood should be
released without the capacity of the powerplant (if any) and the gate with the highest
capacity which is considered not operational (N-1 gates). The minimum freeboard of the
dam is defined based on the type of dam (concrete or embankment) and the height of the
dam; the minimum freeboard can vary from 0.50 m to 3.0 m as shown on Table A26-1.
Safety flood
The safety flood safety corresponds to the discharge that the system must be able to manage
(to retain or to spill) in extreme conditions. During such event, the water level in the
reservoir must not exceed the danger (or critical) elevation of the dam. The danger elevation
corresponds to the water level for which the safety of the dam (or structure) could be
compromised due to a lack of stability or because of the possible damages).
In the case of homogeneous embankment dams, the danger elevation correspond to the crest
of the dam, in the case of the other embankment dams it is at the top of the impervious level
of the dam. For concrete dams, the danger elevation corresponds to the crest of the dam
The following criteria apply for the safety flood:
• The danger elevation must not be exceeded based on the hydrograph of the safety flood
Qs(t);
• Limited damages to the dam are acceptable, but any partial or total rupture of the dam is
excluded;
• For concrete dams, all hydraulic passages are considered operational during the flood,
but the powerplant (if any) is not operational.
• For embankment dams, the gate with the highest capacity is considered not operational
("n-1" rule) and the powerplant (if any) is also not operational.
• The water supply flows are taken into account in accordance with paragraph 3.1.
In Switzerland, the safety flood (Qs(t)) is corresponding to 1.5 times Qb(t) for existing dams
or 1.5 times the Qb(2/3t) for new dams. It is also acceptable to consider the PMF, which are
based on the Probable Maximum Precipitation (PMP) already defined for any region in
Switzerland.
43
Confédération suisse - Département fédéral de l’environnement, des transports de l’énergie et de
la communication DETEC - Office fédéral de l’énergie OFEN - Section Barrages, Sécurité des
ouvrages d’accumulation – Documentation de base relative à la vérification de la sécurité en cas
de crue, juin 2008.
44
27. TURKEY
Proper dimensioning of spillways must take into account both the security and economic
point of view. Different design floods shall be selected for the embankment and concrete
dams. The basic principles that have to be followed are summarized in the following.
1- In the selection of the Project Design Flood, the type of dam shall always be taken into
account. For large dams, i.e. dam higher than 25 m :
a. For embankment dams, the dimensions of the spillways shall be determined by using
Probable Maximum Flood (PMF) and by performing flood routing for all the
spillways, gated or ungated.
During the calculations of flood routing the entire freeboard shall be used, i.e. it is
accepted that the reservoir water level may rise up to the crest level of the dam.
At the first stage of the flood routing, the preliminarily determined freeboard will be
used. Afterwards, an optimization shall be performed between the spillway
dimensions and the dam height. The most economical spillway dimensions shall be
decided after the optimization.
In case the hydrological data does not cover a long enough period and, as a result one
is not confident enough in the flood calculations, the engineer, according to his
engineering judgment, may not use the entire freeboard during the flood routing and
may allow the reservoir water level to rise only until 1.0 m below the dam crest or,
alternatively, up to the crest level of the impervious core.
b. For concrete dams, the spillway capacity shall be equal to the peak discharge of the
1 000-year flood with the Full Supply Level. 10 000-year flood shall be used as the
Safety Check Flood with routing. During the occurrences of the 10 000-year flood
and Probable Maximum Flood the overtopping of concrete dams may be tolerated.
In some projects the engineer may choose a different flood as the Project Design
Flood, with the condition that he or she justifies his or her selection.
In Turkey IDF for concrete dams higher than 25 m is:
• 10 000-year flood with flood routing, or
• 1 000-year flood without flood routing, whichever is greater.
2- In small dams, the height of which is less than 25 m from thalweg, the Project Design
Flood shall be selected as follows:
a) For embankment dams, depending on the judgment of the engineer, 1 000-year flood
with routing.
b) For concrete dams, depending on the judgment of the engineer, 500-year or smaller
flood;
However, if the reservoir volume is larger than 10 million m3 or downstream risks are
important, depending on the judgment of the engineer, larger floods may be selected.
3- In the determination of the capacities of spillways, the attenuation of floods through the
reservoirs of the upstream dams (existing, under construction and projected in the short
term) shall be taken into account.
44
T. R. Ministry of Forestry and Water Resources, State Hydraulic Works, Ankara, Turkey, Official
Circular - 2005
45 46
28. UNITED KINGDOM
In the UK, all dams with normal impounded volume exceeding 25 000 m3 are regulated
under the 1975 Reservoirs Act. However, proposals to reduce this limit were considered in
2010-11 following incidents caused by intense rainfall on drainage areas controlled by small
dams.
The present dam classification is based on the consequences of a dam failure and the dam
are classified in four categories; for each of these categories a normal design flood and a
minimum design flood, if overtopping is permitted, are defined as shown in Table A28-1 :
Minimum
Normal Initial Wind speed and
Categ Consequence of standard if
design reservoir minimum wave
ory a dam breach overtopping
standard condition surcharge
tolerable
Endangers lives Spilling Mean annual
in a community 10 000-yr long-term maximum wind speed
A PMF
(more than 10 flood average Minimum 0.6 m wave
persons) inflow surcharge
Endangers lives Full to
of individuals or 10 000- 1 000-yr spillway
B As Category A
Causes extensive yr flood flood crest (no
damage spill)
Full to Mean annual
Negligible risk to maximum wind speed
1 000-yr 150-yr spillway
C life and limited
flood flood crest (no Minimum 0.4 m wave
damage
spill) surcharge
No risk to life and Spilling Average annual
very limited 150-yr 150-yr long-term maximum wind speed
D
additional flood flood flood average Minimum 0.3 m wave
damage inflow surcharge
A gated spillway should have at least two gates and if one of them is out-of-order, the
remaining gate(s) shall release at least the 150-yr flood. If the dam is from Category A, the
gates must be automated.
45
Fridolf, T. Design flood for dams - analysis of the Swedish guidelines, Dams in a European Context,
Midttomme et al (eds), ©2001 Swets & Zeitiinger, Lisse, ISBN 90 5809 196 1
46
Central Water Commission – Ministry of Water Resources – Govt. of India, Development of
Hydrological Design Aids (Surface Water) under Hydrology Project – II – State of the Art Report, July 2010,
362 pages
47 48 49
29. USA
Generalities
In the US, the different dam owners, including different US agencies, use different
guidelines for the determination of the design flood. Most of these approaches are based on
the probability of loss of life and the PMF (or a percentage of the PMF) is often considered
as the maximum value of the IDF. A review of some of these guidelines is presented
hereafter.
FEMA
The Federal Emergency Management Agency (FEMA) published in 2004 the federal
guidelines for dam safety. One of the four documents, Guidelines for Selecting and
Accommodating Inflow Design Floods for Dams, covers the selection of the Inflow Design
Floods (IDF) for dams. FEMA published an updated version of the federal guidelines on
inflow design flood for dams in 2013, in which it recommends alternative approaches for
selecting and accommodating IDF, recognizing that the wide variety of dams and
watersheds requires a variety of approaches to achieve a reasonable balance of public
protection, efficiency of evaluation, and efficiency of project operation.
According to FEMA, the goal of selecting the IDF should be to balance the risks of a
hydrologic failure of a dam with the potential downstream consequences and the benefits
derived from the dam.
Selection of an IDF can involve tradeoffs in trying to satisfy multiple objectives including:
1. Providing acceptable safety to the public.
2. Effectively applying the resources of the dam owner.
3. Maintaining the credibility of the regulator in representing the interest of the public.
4. Assessing the desire of the public for the benefits of a dam in exchange for the inherent
risks that come from living downstream of a dam.
The following alternative approaches to defining the IDF are recommended to accommodate
the wide variety of situations, available resources, and conditions which might be
encountered in practice:
Prescriptive Approach – In this initial phase, a planned dam is designed or an existing dam
is evaluated for a prescribed standard based on the hazard potential classification of the dam.
This approach is intended to be conservative to allow for efficiency of resource utilization
while providing reasonable assurance of the safety of the public. It is not intended to assure
that there is an economical marginal benefit from designing for a conservative IDF. Table
A29-1 summarises the IDF requirements for the three hazard potential classes.
47
FEMA. "Federal Guidelines for Dam Safety - Hazard Potential Classification System for
Dams ", Federal Emergency Management Agency. April 2004
48
FEMA. " Federal Guidelines for Dam Safety - Selecting and Accommodating Inflow
Design Floods for Dams", Federal Emergency Management Agency, August 2013
49
Central Water & Power Commission – Ministry of Irrigation & Power – Government of
India, Estimation of Design Flood – Recommended procedures.
Table A29-1 USA – IDF requirements for dams using a prescriptive approach
50
USDI – Bureau of Reclamation, Design Standards No. 14 - Appurtenant Structures for Dams
(Spillways and Outlet Works) - Chapter 2: Hydrologic Considerations - Draft: Phase 3 (Public
Review), December 2012.
Figure A29-1 USA – Reclamation dam safety – Risk guidelines Reclamation’s f-N chart
US Department of Agriculture
Earth Dams and Reservoirs, Natural Resources Conservation Services TR–60/ July 2005
NRCS Technical Release 60 (TR-60) specifies the minimum requirements for planning and
designing earth dams and associated spillways for USDA dams that are classified as Low
Hazard Class dams with a product of storage times the effective height of the dam of 3,000
acre-feet2 or more, those more than 35 feet (10.7 m) in effective height, and all Significant
1
Precipitation amounts by return period in years.
2
Applies to irrigation dams on ephemeral streams in areas where the annual rainfall is less the
25 inches (635 mm).
3
The minimum criteria are to be increased from P25 to P100 for a ramp spillway.
4
Low Hazard Class dams involving industrial or municipal water are to be designed with a
minimum criteria equivalent to that of Significant Hazard Class.
5
Applies when the upstream dam is located so that its failure could endanger the lower dam.
1
P100 = Precipitation for 100-year return period. PMP = Probable maximum precipitation
2
Dams involving industrial or municipal water are to use minimum criteria equivalent to that of
Significant Hazard Class.
3
Applies when the upstream dam is located so that its failure could endanger the lower dam
State Level
The vast majority of dams in the USA are regulated by state dam safety programs. Alabama
is the only state that lacks a dam safety program; all other states, plus Puerto Rico, have
legislative authority to administer dam safety regulatory programs. The Association of State
Dam Safety Officials (ASDSO) compiles the dam safety policies and standards for selected
states, which include the inflow design flood requirements for high and significant hazard
potential dams, and spillway design requirements for low hazard potentials dams of selected
states. Majority of the states requires an IDF equal to the PMF or a percentage of it for high
hazard potential dams.
TABLE A-1 - COMPARISON OF THE CHARARCTERISTICS CONSIDERED TO EVALUTE THE DESIGN FLOOD
APPENDIX B
EXAMPLE – PROJECT β
RESULTS FOR EACH COUNTRY
AUSTRALIA
Class name Definition Risk IDF
High LOL expected Extreme Damage 10 000-yr to PMF
Significant LOL possible Appreciable losses 1 000-yr to
10 000-yr
Low No expected LOL Minimal losses 100-yr to 1 000-yr
IDF Criteria for conditions 10 000-yr
equivalent to Project β
AUSTRIA
Class name Definition Risk IDF
IDF Criteria for conditions 5 000-yr
equivalent to Project β
BRAZIL 51
Class name Definition Risk IDF
Class A More than 30 m and/or PMF
storage exceeding 50 hm³ or
dam involving risk of loss of
life
Class B Otherwise 1 000-yr or more
IDF Criteria for conditions PMF
equivalent to Project β
BULGARIA
Class name Definition Risk IDF
I Project more than 100 MW 10 000-yr
Depending on dam type and
foundation
II Project more than 25 MW 1 000-yr
Depending on dam type and
foundation
III Project more than 10 MW 200-yr
Depending on dam type and
foundation
51
In Brazil, it appears there is no legislation or national standards in relation to design floods for
dams. This criteria is the one used by Eletrobras, the state owned electricity utility.
BULGARIA (cont'd)
IV Project more than 5 MW 100-yr
Depending on dam type and
foundation
V Project less than 5 MW 33,33-yr
Depending on dam type and
foundation
IDF Criteria for conditions 1 000-yr
equivalent to Project β
CHINA
Class name Definition Risk IDF
I Storage > 1 000 hm³ (Embankment) 10 000-yr to PMF
(Concrete) 5 000-yr
II Storage > 100 hm³ (Embankment) 2 000-yr
(Concrete) 1 000-yr
III Storage > 10 hm³ (Embankment) 1 000-yr
(Concrete) 500-yr
IV Storage > 1 hm³ (Embankment) 1 000-yr
(Concrete) 500-yr
V Storage < 1 hm³ (Embankment) 200-yr
(Concrete) 100-yr
IDF Criteria for conditions 2 000-yr
equivalent to Project β
CZECH REPUBLIC
Class name Definition Risk IDF
I More than 1 000 persons at 10 000-yr
risk – Large amount of
LOL anticipated – Extensive
damage – Environmental
consequence for the country
II More than 100 persons at 2 000-yr if losses of 2 000-yr to
risk – LOL anticipated human life are 10 000-yr
Extensive damage – unlikely
Environmental consequences
higher than the autonomous
district
III More than 10 persons at risk 200-yr if losses of 200-yr to 1 000-yr
– LOL could result – human life are
Important damage – unlikely
Environmental consequences
for the autonomous district
IV LOL Improbable – Some 20-yr to 100-yr
damages – Environmental
consequences are small
IDF Criteria for conditions 1 000-yr
equivalent to Project β
FINLAND
Class name Definition Risk IDF
P Endanger life or 5 000-yr to
environment 10 000-yr
N In between 500-yr to 1 000-yr
O Cannot endanger life or 100-yr to 500-yr
environment
IDF Criteria for conditions 10 000-yr
equivalent to Project β
FRANCE
Class name Definition Risk IDF
A H ≥ 20 m Exceptional
H: Height (m) situation:
Concrete dams 1 000 - 3 000 yr
Embankment 10 000 yr
Extreme situation 100 000 yr
B H(m) ≥ 10 m and Exceptional
H2√V ≥ 200 situation:
V: volume of reservoir Concrete dams 1 000 yr
(hm³) Embankment 3 000 yr
Extreme situation 30 000 yr
C H ≥ 5 m and H2√V ≥ 20 Exceptional
situation:
Concrete dams 300 yr
Embankment 1 000 yr
Extreme situation 10 000 yr
D H ≥ 2m Exceptional
situation:
Concrete dams 100 yr
Embankment 300 yr
Extreme situation 1 000 yr
IDF Criteria for conditions 3 000-yr
equivalent to Project β
GERMANY
Class name Definition Risk IDF
Large dams Storage > 1 hm³ or 10 000-yr
H >15 m
Medium and small Storage < 1 hm³ and 5 000-yr
H < 15 m
Very Small 1 000-yr
IDF Criteria for conditions 10 000-yr
equivalent to Project β
INDIA
Class name Definition Risk IDF
Large dams Storage > 60 hm³ or PMF
Height > 30 m
Intermediate Storage > 10 hm³ or 1 000-yr or SPF 52
Height > 12 m
Small Storage < 10 hm³ or 100-yr
Height < 12 m
IDF Criteria for conditions PMF
equivalent to Project β
IRELAND 53
Class name Definition Risk Class Name
Class A Can endanger life 10 000-yr or 1 000-
yr with N-1 gates
Otherwise Negligible risk to human life 1 000-yr with N-1
gates
IDF Criteria for conditions 10 000-yr
equivalent to Project β
ITALY
Class name Definition Risk IDF
Large dams H > 15 m or V > 1hm³
- Embankment 3 000-yr
- Concrete 1 000-yr
IDF Criteria for conditions 3 000-yr
equivalent to Project β
52
Standard Project Flood
53
In Ireland, it appears there is no legislation or national standards in relation to design floods for
dams. This criteria is the one used by Electricity Supply Board (ESB), the state owned electricity
utility
JAPAN
Class name Definition Risk IDF
Concrete dam 200-yr flood or 200-yr
Maximum experienced flood
or
Maximum that can occur
based on max. flood
experienced on the basin
Embankment dam 1.2 times of the discharge About 1 000-yr
for concrete dam
Return period of design
flood for embankment dam
is equivalent to 1 000 years
or more (CWC, 2010)
IDF Criteria for conditions 1 000-yr
equivalent to Project β
NEW ZEALAND
Class name Definition Risk IDF
High LOL Catastrophic 10 000-yr to PMF
damages
Medium A few LOL Major damages 1 000-yr to
10 000-yr
Low No LOL expected Moderate damages 100-yr to 1 000-yr
Very low No LOL Minimal damages
IDF Criteria for conditions 10 000-yr to PMF
equivalent to Project β
NORWAY
Class name Definition Risk IDF
Very High More than 150 houses Inflow Design Flood 1 000-yr
Safety Check Flood
PMF
High More than 20 houses Inflow Design Flood 1 000-yr
Safety Check Flood
PMF
Medium 1 to 20 houses Inflow Design Flood 1 000-yr
Safety Check Flood
1.5x1 000-yr or
PMF
Low 0 houses Inflow Design Flood 500-yr
Safety Check Flood 1.5x1 000-yr or
PMF
IDF Criteria for conditions Inflow Design Flood 1 000-yr
equivalent to Project β
PANAMA
Class name Definition Risk IDF
High (A) Risk of LOL 1 000-yr to
5000-yr
Significant (B) Unlikely LOL but economic 500-yr to 1 000-yr
loss
Low (C) Unlikely LOL or economic 100-yr
or environmental loss
IDF Criteria for conditions
equivalent to Project β
POLAND
Class name Definition Risk IDF
I H > 30 m, or; 1 000-yr
V > 50 hm³, or;
Submersion area > 50 km²;
Number of people > 300
II H > 15 m, or; 500-yr
V > 20 hm³, or;
Submersion area > 10 km²;
Number of people > 80
III H > 5 m, or; 333-yr
V > 5 hm³, or;
Submersion area > 1 km²;
Number of people > 10
IV H > 2 m, or; 200-yr
V > 0.2 hm³, or;
Submersion area <= 1 km²;
Number of people <= 10
IDF Criteria for conditions 1 000-yr
equivalent to Project β
PORTUGAL
Class name Definition Risk IDF
High Dam failure will affect 25 or Depends of the type 1 000-yr to
more residents of dam and dam 10 000-yr
height
Significant Dam failure affect at least Depends of the type 500-yr to 5 000-yr
one resident and have of dam and dam
significant impact on height
infrastructure
Low No loss of life little damage NA
ROMANIA
Class name Definition Risk IDF
I Hydraulic structures of Catastrophic 10 000-yr
exceptional importance damages and/or
unacceptable
interruption
II Hydraulic structures of Serious damages 1 000-yr
special importance and/or interruptions
acceptable for short
period
III Medium importance Social – economic 200-yr
damages
IV Low importance Little influence on 100-yr
other social –
economic aspects
IDF Criteria for conditions 1 000-yr
equivalent to Project β
RUSSIA
Class name Definition Risk IDF
I V > 1 000 hm³ Inflow Design 1 000-yr
Flood
Safety Check Flood 10 000-yr
II V > 200 hm³ Inflow Design 100-yr
Flood
Safety Check Flood 1 000-yr
III V > 50 hm³ Inflow Design 33.3-yr
Flood
Safety Check Flood 100-yr
IV V < 50 hm³ Inflow Design 20-yr
Flood
Safety Check Flood N/A
IDF Criteria for conditions Inflow Design 1 000-yr
equivalent to Project β Flood
SOUTH AFRICA
Class name Definition Risk IDF
High More than 10 LOL Severe economical 200-yr (IDF)
losses >10 000-yr(Safety)
Medium 10 or less LOL Significant 100-yr (IDF)
economical losses >6 000-yr(Safety)
Low No LOL Minimal 100-yr (IDF)
economical losses >1,200-yr(Safety)
IDF Criteria for conditions 100-yr (IDF)
equivalent to Project β >6 000-yr(Safety)
SPAIN
Class name Definition Risk IDF
Inflow Design 1 000-yr
A More than 5 houses Flood
Safety Check Flood 10 000-yr
Inflow Design 500-yr
B 1 to 5 houses Flood
Safety Check Flood 5 000-yr
Inflow Design 100-yr
C Incidental loss of life Flood
Safety Check Flood 500-yr
IDF Criteria for conditions Inflow Design 1 000-yr
equivalent to Project β Flood
SWEDEN
Class name Definition Risk IDF
High Significant risk Human Life, environment, SDF 2
economic
Low 100-yr
IDF Criteria for conditions SDF
equivalent to Project β
SWITZERLAND
Class name Definition Risk IDF
Design Flood 1 000-yr
Check Flood 1.5x1 000-yr or
PMF
IDF Criteria for conditions 1 000-yr
equivalent to Project β
TURKEY
Class name Definition Risk IDF
High – Embankment H > 25 m PMF
High – Concrete H > 25 m 10 000-yr with 1 000 yr
flood routing or
1 000-yr without
Low – Embankment H < 25 m 10 000-yr with 1 000 yr
Storage < 10 000 m³ flood routing or
1 000-yr without
flood routing
Low – Concrete H < 25 m 500 yr
IDF Criteria for conditions 1 000 yr
equivalent to Project β
UK
Class name Definition Risk IDF
A More than 10 persons at risk PMF
(community)
B Limited LOL Extensive damages 10 000-yr
C Negligible risk of LOL Limited damage 1 000-yr
D No LOL Very limited 150-yr
damage
IDF Criteria for conditions PMF
equivalent to Project β
USA / FEMA
Class name Definition Risk IDF
High Clear danger of LOL Serious danger to PMF
industries and
important activities
Significant Possible danger of LOL Significant danger 0,75 PMF
to industries and
important activities
Low No expected LOL Some danger to 100-yr
agricultural
activities
IDF Criteria for conditions PMF
equivalent to Project β
USA / USACE
Class name Definition Risk IDF
High More than few LOL Excessive losses PMF
Significant Few LOL Appreciable losses PMF
Low No expected LOL Minimal losses 0,5 to PMF
IDF Criteria for conditions PMF
equivalent to Project β