CHAPTER

METHODS FOR CALCULATING PROJECT FLOW

Choice of a design flood

A flood design appears in the form of a flood hydrograph or simply a flow of
point (project flow) that serves as the basis for the design and sizing of the project in question

In theory, the return period T of an event is defined by the inverse of the probability of.
exceeding F1 to observe this event: T = 1/F1.
Statistically, the occurrence for the project flow of return period T to be exceeded at least
once during n successive years is given by the hydrological risk:
Rh(%) = 1-(1-1/T) n100
The choice of the return period is dictated by the importance of the risks involved in case of default.
the work done and by economic considerations. We compare the cost of
investments in repair and damage in the event of a potential break.

The assessment of the project flow of a structure results from a hydrological study, but also
a socio-economic analysis allowing to define:
. the cost of the work and its maintenance based on its capacity
. the socio-economic consequences due to the flooding and the destruction of the structure.

the cost of the work or simply the investment increases with the level of protection, so
that, the maintenance and operating costs as well as those of the damages caused by a failure
decreases with the level of protection, that is to say with the adopted return period.

The choice of an appropriate return period thus comes down to finding an optimum in the ratio between these
two components: investment and operation on one hand and damages on the other hand.
If the cost of the work as well as the maintenance costs can be easily estimated, the other
The components of the equation are difficult to determine. This is particularly true for the costs incurred.
by the users.
Such in-depth socio-economic analyses are hardly ever conducted and the choice of the
The return period is still based on experience and tradition.

In Morocco, bridges are traditionally designed for the centennial flood, the structures
drainage, less important, for the ten-year flood The large dams for flooding
decamillennium and the little ones for the millennial flood. It is advisable to confide for more details
on the return period and its orders of magnitude in the chapter on statistical data processing
hydrological.

Prof N. SERHIR 1- Project Flow Calculation

 Operational methods for calculating project flow
The calculation of a return period flow rate involves knowledge of a certain
number of hydrological information that, unfortunately, are often partially or
totally deficient, especially for small watersheds.
Thus, it is according to the purpose assigned to the study and the available basic hydrological data that one
choose the method for calculating project flows.
In the event that the engineer has no hydrological data (rainfall and/or flows) instead of the
site studied or in the relevant watershed, there are two ways to proceed:

By analogy: to the extent that data from representative basins are available.
If this is not the case and if the deadlines are too short to proceed with the acquisition of the information
necessary, it will be necessary to resort to the use of empirical formulas.
If, however, the engineer has rainfall data in the absence of flow data, he
can use a rational type formula or any other elaborate formula involving
"rain information" in addition to the characteristic parameters of the watershed.
In cases where the engineer simultaneously has rainfall and flow data, even
in the form of a short series of measures, he will be able to analyze the "rain -
flow of recorded events and thus implement more sophisticated methods of
calculation of project flows, such as statistical methods, hydrological regressions
pluviometric and the Gradex method

2-1 / The analogical methods:

They are based on the comparison with regional watersheds for which there are
hydrological data These methods consist of extrapolating the project flow to the unmeasured site at
to take data observed in a gauged basin that is hydrologically comparable.

This comparison is based on the parameters that define geometrically and

Morphologically, the basin: shape, slopes and slope indices, parameters of the equivalent rectangle,
relief characteristics, drainage density and confluence ratio as well as parameters
climatic, geological, and pedological informing about the types of precipitation, the permeabilities
of soils, the nature of vegetation. All this information can help the engineer to
understand the hydrological behavior of the basin and consequently its flow regime.

The most commonly used methods in Morocco are the specific discharge method and the method
regional from Francou Rodier.

Method of specific flows

We list the data related to the gauged basin in which a hydrometric station is located.
provides data forming a basis for a statistical study that will allow calculating the flow
QA(T) ;

A couple of data related to the unmeasured basin for which we want to calculate the flow rate.
project for the construction of a hydraulic work.

Prof N. SERHIR 2- Project Flow Calculation

 The specific discharge method involves assessing the flow rate [Link] from QAwhile holding
only accounts for the respective influence of the watershed areas.
This amounts to neglecting the other factors that contribute to the hydrological behavior of
each of the basins: ,nature soil and vegetation , topography, ....
She assumes that if these factors are the same for both basins, (which is not always true.)
So the specific discharges for the same return period will be identical.
What allows to write:
QA(T)QB(T)
SA SB
The regional method of Francou–Rodier

The method uses a regional empirical formula. The most used in Morocco is that of Francou -
Rodier: K
S (1 )
Q 10 6( ) 10
10 8
What is the parameter of Francou Rodier. The method
consists of transferring this estimated regional parameter from a gauged basin to the concerned ungauged basin
by determining the project flow.

The first step in using the regional method is to calculate the parameter K.
knowing the flood flowAcalculated for a return period T, in the gauged basin of
surfacesA.
QAcan be calculated by a statistical study applied to the data observed in the basin, or by
the rational method. The choice of method depends on the available data in the basin
gauge and the return period T of the forecast.

The second step of the calculation is to use the same value of K to calculate the flood flow.
in the unmeasured surface basinB. The Francou Rodier formula is therefore used for this
extrapolation.

The calculation of K can be done easily by performing a logarithmic change of variable.

on the Rodier formula:
K(T)
log(QA(T)) 6  (1 ) (logSA 8) K(T)
10
then the extrapolation to basin B gives:
K(T)
S 1
QB(T) 10 6( B) 10
10 8
A good knowledge is acquired on the variation of this
factor according to Moroccan regions: 4.0 k 5.0

2-2 / Empirical methods of predetermination

Several empirical formulas exist in the literature. We will mention the following relationships:

Prof N. SERHIR 3- Project Flow Calculation

 Empirical formulas based on the surface area of the basin
In general, the specific flow of a flood decreases when the surface of the basin
increases. This observation has led many authors to link the maximum flood flow to the
surface of the basin.
Most of the formulas giving the maximum flow rates Q (m3depending on the area of the watershed
S (km²) is related to Myer's formula: Q=C.S 
Where C = the 'Myer' rating of the basin, depending on the characteristics of the basin
Exponent generally taken in Morocco equal to 0.5 (ranging from 0.4 to 0.8 depending on the regions)

Also belonging to this first group are the formulas of M Hazan and Lazarevic, determined
particularly in different regions of Morocco. The peak flow of the thousand-year flood is
given by a relation of the type Q1000= a Sbwhere the coefficients vary depending on the area
of the basin, geographic areas, and ranges of annual rainfall. Table 1 in
present the summary:

Table Geographical zone Relation3/s) Rainfall (mm)

to 1: Central Rif Q1000= 15.55 S0.776 1000-1300
formula Western Rif Q10009.78 S0.793 800-1000
is from Eastern Rif Q1000= 7.58 S0.808 600-800
Hazen Saharan High Atlas Q1000= 9.38 S0.742 200-400
and Middle Atlas Q1000= 14.94 S0.636 700-900
Lazarus Middle Atlas Q100013.51 S0.613 500-700
vic Middle Atlas (Karsts) Q1000= 13.47 S0.587 400-700

Regional formulas have been determined in the Sebou basin:

These formulas are of the form: Q(T) = a(T)Sb.

QTpeak flow in m3/s of the flood with a return period T.

S: watershed area (km2).
b: regional coefficient equal to 0.752.
a(T) variable parameter depending on the area considered in the Sebou basin and depends on T
The parameter a(T) thus regionalized in the Sebou basin integrates the influence of various factors
of the genesis of floods

Formulas for the flood return period function

These methods are generally encountered in the bibliography related to 'the study
the hydrological design of structures for crossing watercourses in Morocco.

Fuller formula

The following relationship is proposed to be used in Morocco:

Qp = q (1 + a log (T) )(1 + 2.66/S0.3)

With:
Prof N. SERHIR 4- Project Flow Calculation
 Qp= maximum flow for the return period T in m3/s ,
S = watershed area (km2)
q = average of the maximum flows for each year, calculated based on the data
available in m3Where

a = 0.8 to 1.2 for the Rif valleys

a = 3.0 to 3.5 for the Saharan wadis
Mallet and Gauthier formula:
Q(T) 2Klog(1 a * P ) *anS *  1 4log(T) log(S)1/2 
L1/2
Q(T) : débit max en m/s 3 return period T.
S = Watershed area in km2
Panaverage annual rainfall in m
L = Length of the main watershed in km
K = variant coefficient from 0.5 to 6 (0.5 for large basins and 6 for small basins)
a = coefficient variant from 20 to 30

In Morocco, the values usually used are k = 2 and a = 20.

It is difficult to use due to the uncertainty surrounding the estimation of
coefficients k and a

Mac-Math formula:

This formula is based on observations made in the southern region of California where the lands are
arides.
Q(T) = K x P(24h,T) x (S0.58) x (I0.42)

Q(T): Maximum flow rate of return period T in (m3/s).

S surface watershed (in km2).
I slope of the longest watercourse (in m/m).
P(24,T): Maximum precipitation over a duration of 24 hours with a return period T (in mm).
K Coefficient that depends on the vegetation cover and the topography of
watershed, it varies from 0.11 to 0.43:

K = 0.11 large watershed and vegetation.

K = 0.22: cultivated area and vacant land in suburban areas
K = 0.32: unprepared terrain, non-rocky, of medium slope
K = 0.43: unmanaged rocky terrain with a steep slope.

Function formula of precipitation

Among the most commonly used formulas involving precipitation and time
concentration, we will specifically mention that of Turazza: (Italian formula):

Qmax (F) c
3.6t c
Prof N. SERHIR 5- Project Flow Calculation
 Q = maximum flood discharge (m3/s),
C = flow coefficient of the basin for the considered flood,
H = Precipitation recorded over a duration equal to the concentration time tc of the basin, in mm
tcconcentration time (h),
S = area of the watershed (km²).

This formula, which is nothing more than a rational type formula, is very suitable in the case
studies of small basins or urban wastewater networks.
This formula implicitly introduces the notion of the flood discharge frequency. It is assumed, in
indeed, in the first approximation, the frequency of the determined flow is equal to that of the intensity
maximum, which is equal to the maximum rainfall over the concentration time, that is H / tcexpressed in
mm/hour.

Important note:

Empirical formulas are numerous and one must be aware of the fragility and narrowness of the
scope of each of them. Let us not lose sight of the fact that these relationships are based on
the analysis of data collected in basins under climatological and topographic conditions
particular, and that one cannot extrapolate them to other basins without verification.
conditions different from those for which they were established. The dangers of extrapolation
unconsidered of these formulas are certain. One just needs to be convinced by comparing the flows of
results obtained by applying these different formulas for the same value of the parameters taken
en compte.
Therefore, they should be used with great caution.

One way to optimize the use of such formulas is to compare them with the data from the
hydrological station of the neighboring basin or the closest similar basin, in order to identify at
better choice of parameters. The results of flood studies must also be compared to
data collected in the field of the study. This is done by conducting flood surveys with the
riverside inhabitants.
However, it is always important to remember that:
A bad measurement is better than a long calculation

2-3 / Probabilistic method

(refer to the chapter on the statistical (frequency) analysis of hydrological data)

Probabilistic methods involve fitting probability distributions to the observed floods and to
extrapolate the best law that represents the empirical distribution for return periods
data. Two major problems arise:

The adjustment:

In most cases, it is possible to find several probability laws that fit correctly.
according to the available data; Depending on the objective of the established frequency analysis (interpolation or
extrapolation, flood or drought predictions) we should rather say that the sample fits well to
this law in an interval a-b. The confirmation with the KHI2 adequacy test and the calculation of
confidence intervals around the observations will prove necessary in some cases for better
qualify the final choice of law

Extrapolation:
Prof N. SERHIR 6- Project Flow Calculation
 The extrapolation following the adjustments made at low frequencies leads us to suppose that the
Rare frequency floods are merely an extension of the common floods observed. Let us also note that
that all
extrapolation can sometimes give, depending on the adopted law, results that can differ by 50 to
100% for very rare floods.

The adjustment and extrapolation of flood flows must be handled with great care.
precautions regarding the choice of the probabilistic law

2-4 / The hydrometeorological methods

The principle of these methods is to relate flood flows to precipitation. It involves using
the rainfall information to calculate the flood discharge. The recommended rational method
for small catchment areas allows for the establishment of acceptable forecasts for return periods
ranging from 2 to 20-25 years and the developed Gradex method can be used for large basins
and gives very secure values for forecasting very rare events where T varies from 100 to
10000 years

2-4-1 / Calculation by the rational method:

This method is well suited for small watersheds whose area does not exceed 150 km.2.
It assumes that the peak flow rate of the runoff cannot be observed at the outlet of a basin.
since when the entire area contributes, that is, if the duration of the widespread downpour is at
less equal to the concentration time tcof the basin.

She admits, among other things, that the frequency or return period of the determined maximum flow
is equal to that of the average maximum rainfall observed in the basin over the duration tc
The practical form of the relationship is written as:
C P(tc,T)S
Qmax (T)
3.6tc
which is actually equivalent to:
CI(tc,T)S
Qmomx (T)
3.6
I (tc(T) represents the intensity of the average maximum rainfall that fell over the period of [Link] period of
return T.
It is expressed in mm/h
S is the surface area of the basin in km2
It is the average flow coefficient
tcis the concentration time of the basin, expressed in hours
Qmaxis calculated in m3/s

The use of the rational method requires the calculation of the concentration time.
Chapter watershed) and the calculation of the average rainfall of given frequency and equal duration
At the concentration time. The calculation is done according to the available basic data.
We can either:

Prof N. SERHIR 7- Project Flow Calculation

 1. Use already defined IDF curves in the region (refer to the chapter dealing with the study of
detailed hydrology assessments). The regional parameters of Montana or Talbot known,
we calculate the intensity of frequency F and durationc

2. Otherwise, it will be necessary to statistically analyze the maximum daily rainfall per step to calculate the maximum rainfall.
day of frequency F per station and subsequently deduce the maximum average daily rainfall for the basin
frequency F. We then calculate an average rainfall duration of concentration time,
frequency F, based on regional formulas (see Gradex method). This allows for achieving
at the maximum average intensity of duration tcand frequency F

2-4-2 / The Gradex method

This method is attributed to MM Guillot and Duband from EDF Grenoble (1970), who designed it.
for relatively impermeable basins in which exceptional floods are
provoked primarily by rain and not by snowmelt.

It is of interest to take into account the 'rain' information to complement the 'flow' information.
which is generally shorter. Usually, the flow samples are much less detailed
that the rain samples. This means that if we base ourselves solely on the flow sample
it will be difficult to extrapolate return periods of the order of 10 or 20 years. At
on the contrary, the rain samples being generally longer, it will be easier to determine with
sufficient precision the centennial value, millennial for example. The method then allows to
address the weakness of the 'flow' sample by using the 'rain' sample, better known because
longer.

Limits of the method

According to the classic scheme of transforming rain into flow, we will focus on rain with a duration of t.c
(concentration time) that involves the entire watershed in the runoff. If the rain has a
duration greater than or equal to tcAn equilibrium is established and the flow at the outlet reaches a maximum.
We then extrapolate the maximum flows in the basin for rainfall of duration [Link] must have
of several years of daily average rainfall measurements in the basin (> 10 years) of
good quality. The flows may not be available or exist for a very short duration.

In general, it is recommended to apply this method to watersheds that are quite

waterproof (which can reach saturation for a received rainfall threshold), with an area
going up to 5000 km2and whose concentration time is between 1 hour and 4 days.

foundations of the method

Starting from the verified hypothesis of relative permeability, the basin must have an absorption capacity.
limited. It is therefore capable of reaching saturation for a threshold of rainfall received which we will denote as P.0 .
The Gradex method is based on two assumptions:

For any rain P received by the basin exceeding P0the flow would be complete.

If we denote T0the payback period corresponding to the threshold P0for which the soil equilibrium saturation
is then reached the flow generated by P0(T0) will be a flow Q(T0) the same return period. It
is called "the hinge or saturation flow rate".
Prof N. SERHIR 8- Project Flow Calculation
 The authors recommend that the probabilistic law of basin behavior be the law of
extreme runoff can be deduced from the probabilistic law of rainfall beyond Q(T0)

They also relied on several frequency studies on daily precipitation.

maximums, based on the statistical principle that the Gumbel law fits well to this hydrological variable.
Thus, if the strongest rains follow this probabilistic law, the graphical representation of the
Gumbel law, on the paper Gumbel will be linear with a slope of 1/a (based on the reduced variable
of Gumbel and )

The slope 1/a of the rain line is what we call the rain gradient, denoted as G.p.
It is a climatological factor that actually characterizes the risks of heavy rain in the basin.

From these two hypotheses follows the fact that the distribution function of extreme flows is extrapolated,
beyond the threshold flood flow, alongside that of extreme rainfall, namely the flow Q
it will also be linear with respect to y from the return period T0, since beyond T0,
the flow is integral.

Thus, for all T > T0the flow curve will tend towards a Gumbel distribution (thus a straight line)
parallel to the line of extreme rainfall. With unit changes, the Gradex of flows
deduced from the 24-hour rainfall Gradex by:
Gradex of flows (m3Gradex of precipitation (mm) * S ( k m 2)
3.6*24(h)
Figure 2 graphically illustrates these interpretations.

The 24-hour rainfall Gradex is calculated based on the average maximum daily rainfall data.
We remind you that the calculation of the rain Gradex is done using the formulas studied within the framework of
Chapter on the statistical analysis of hydrological data: see Gumbel's law.

Figure 2:

Method you Gradex :

Extrapolation of the law of
debts

Calculation of the Gradex of duration t flowsc:

As we are interested in calculating the maximum flow rate, with a return period T during the time of
Concentration therefore we must calculate the Gradex of the flows from the Gradex of the rainfall of duration t.c.
We must then:
1. Calculate the concentration time as tc = N hours
2. Calculate the rainfall Gradex for a duration equal to tc:
Prof N. SERHIR 9- Project Flow Calculation
 We are basing ourselves on the regional relationship experienced in Morocco.

K
N
P(N hours)P(24h) 24
And it is applied on the average rainfall Gradex over a duration of 24 hours.
A regional parameter ranging from 0.3 to 0.5. A value of 0.3 is often adopted.

3. Calculate the flow Gradex for a duration equal to t.c:

Sbas sIn
Gd(tc ) G p(tc )
tc3.6

Where Gpis
expressed in mm
and Gdis expressed in m3/ s et tcin hours
The surface area of the basin in km2

Calculation of the law on maximum flows in return period T:

The Gumbel probability law on maximum flows is valid from the return period T.0.
It is written according to the aforementioned hypotheses by:
Q( T ) = Gd( tc) . y (T) + Qo

Where y(T) is the reduced Gumbel variable for the return period T.
Thus the calculation of Qo takes place from the saturation flow rate Q (T0who must be known and calculated
in the basin.
The method recommends a return period of ten years: T0= 10 years, except recommendation
different.
The calculation of Q (T0) can be done by an analogical method, or by the rational method or by
an empirical formula, methods that remain valid for a return period of 10 years.

If we consider Ts = 10 years, as recommended by the method, we can write that:

Q(10) = Gd( tc) . y (10) + Qo

y(10) = - Ln-Ln (0.90 )= 2.25

Qo = Q(10) - Gd( tc) . y (10)

The final Ainsila relation is written:

Q(T) G(t)* y(T) 2.25 Q(10)
d c
It is the Gumbel probability law that will allow the calculation of the maximum flow by Gradex for
the payback period T

A safety factor is often applied to the value calculated by the Gradex method for
determine the peak flow of the project flood that we will use for the design of a structure
important, with a return period exceeding 100 years.
An average coefficient called the passage of the order of: rp 1.5 is recommended.

Prof N. SERHIR 10 Project Flow Calculation

 We can also rely on the results of the table below. It was found through studies conducted at
rp varies between:
Morocco and in France that

1.2 and 1.7 for French watersheds and

1.2 to 3 for Moroccan watersheds.

Zone Climate Crues Rp Order of magnitude of rp

FRANCE weak weak from 1.2 to 1.7
BLACK AFRICA violent fort from 2.0 to 2.5
MOROCCO Mediterranean violent fort from 2 to 2.5
MOROCCO Atlantic weak weak from 1.2 to 1.5
MOROCCO Saharan extremes grand from 2.5 to 3

Summary of the steps of the GRADEX method

This method is based on the principle that extrapolating the flow curve towards the
Strong values corresponding to low probabilities of exceedance cannot be reasonably achieved.
that parallel to the precipitation curve, since it cannot runoff more water than it
does not fall and that soil retention is limited. This means that the outflow coefficient tends
after soil saturation.
This method adheres to the following principles:

Daily precipitation adjusts according to laws with exponential behavior.

simple. So, the maximum monthly or annual values of precipitation
adjust at least in the extreme values according to the Gumbel law.

Beyond the pivotal flood, the most reasonable extrapolation of the flow law
it consists of drawing a parallel to the law of precipitation.

The law of instantaneous flow rates is obtained by an affinity applied to the law of flow rates.
daily.
In other words, the daily flow will be adjusted via a peak coefficient. The choice of this
The coefficient will depend on the analysis of maximum instantaneous flows and maximum daily flows.
The conduct of the Gradex method is as follows:

Study the random variable 'rainfall received by the watersheds in 24 hours';

adjust it according to the Gumbel law and calculate its average Gradex;

Convert this daily value into a value corresponding to the time of

concentration

Prof N. SERHIR 11 - Project Flow Calculation

 Consider the ten-year flow for each basin;

Extrapolate the distribution function of flows beyond 10 years (F = 0.9) using a straight line
of slope equal to the rainfall Gradex converted into flow rate value using the surfaces
watersheds.

Consider a peak coefficient to convert average flows into peak flows

point.

3 - Hydraulic method for calculating a historical maximum flow rate

This complementary method allows for the calculation of the maximum instantaneous flow rate (Qmax) without
frequency. Since it is based on the testimonies of local residents, we can be certain that the
The frequency of the strongest flood experienced in a human lifetime is close to a centennial.
The information gathered from local residents or the traces left by the exceptional flood,
would allow to raise the necessary topographical profiles. It is essential to have a
profile across the watercourse near the site in question (up to the shore reached by the most
high waters Hmax) and a longitudinal profile developed over at least 200m, distributed fairly
equally between the upstream and downstream of the site (in the absence of the traces left by the flood along the
banks, follow the thalweg of the riverbed.
The calculation of the historical Qmax will be done through the well-known relationships in hydraulics:

Qmax = A x V (m3Where V = C R hI and C = 1/n Rh 1/6

In which:

A = wetted section for the coast Hmax, in m2

V = average speed, in m/s
C = Chezy coefficient
Rh Hydraulic radius, in m, is equal to the ratio of the wetted section to the wetted perimeter.
which is measured with a curvimeter on the cross profile
I = slope of the riverbed in mm; it is extracted from the longitudinal profile;
n = Manning's roughness coefficient, tabulated according to the nature of the riverbed and its
banks

Prof N. SERHIR 12 - Project Flow Calculation

 Conclusion :
Several methods have also been used to assess peak flows. We
notably, we will cite the S.C.S (Soil Conservation Service) method, the unit hydrograph.
The downside of these methods is that they must first be calibrated on experimental basins.
you have sufficient data.

Finally, we will emphasize that all the results of the flood studies must be
confronted with the data collected in the field. Hence the need to conduct flood surveys.
among the residents. Experience has shown that this valuable information allows for understanding
with precision the peak flows, at least up to the centennial frequency. This is therefore an operation
indispensable regardless of the layout, big or small.

Any value on the obtained flow must be placed in its regional context: hence the interest of
compare specific flow rates among themselves. This approach is very essential to ensure
the hydrological homogeneity.

To conclude, we will simply say that it is not recommended to use this or that.
method. Indeed, the choice is dictated both by the available database, common sense and
the initiative of the engineer in charge of the study.

Prof N. SERHIR 13 - Project Flow Calculation