AN ANALYTICAL INDEX FOR FLOOD ATTENUATION DUE TO

RESERVOIRS
F. Miotto, P. Claps, F. Laio and D. Poggi
1
(1) Dipartimento di Idraulica, Trasporti ed Infrastrutture Civili, Politecnico di Torino - Torino (IT)
e-mail: [email protected]
Keywords: Flood attenuation, articial reservoirs, synthetic index.
ABSTRACT
Progress on practical problems such as quantifying ood impact on piedmont regions demands increased under-
standing of how articial and natural reservoirs modulates the properties of wave propagation. In particular, how
dams embedded into the river network impact the peak attenuation of the ood wave remains a problem that re-
ceived surprisingly little attention. In fact, although an appropriate setting and management of the dams network
may offer one of the most promising tools to control ood waves propagation on mountain regions, progress on
this topic is hindered by the lack of hydrology data and adequate management policy.
Here, we investigate analytically how the main features of a single dam, such as the lake surface and the width
of the spillways, along with the basic characteristics of the river basin upstream and of the meteorological event,
impact the ood peak attenuation. From this analysis a new synthetic index for ood attenuation, SFA, is derived.
To test the reliability and the robustness of the new index, we apply the SFA to a number of articial reservoirs
constituting a complex river network system in the Toce basin (North Italy).
1 INTRODUCTION
Despite the progresses in engineering practice and management approach, ood mitigation continues to constitute
a major challenge in both advanced and developing countries. In particular, current climate change predictions and
increasing exploitation of oodplains suggest that extended areas in the future will experience a greater frequency
of ooding.
Along with other ood management strategies, natural and articial ood storage is now being considered as a
reliable method for ood risk mitigation. Although it is surely not a panacea, ood storage could be a useful
complement to conventional ood defences, particularly in the piedmont regions of the Alps. Along this line,
there is now increasing interest in multipurpose use of existing reservoirs, based on the understanding that these
reservoirs may provide sustainable attenuation of ood wave peak. In fact, regardless of the individual functions,
they cause important changes in the downstream river regime. In particular, articial reservoirs usually exert a
partial and positive effect protecting downstream areas from ood damage.
Despite the relevant role of articial reservoirs in ood risk mitigation, it is not surprising that a low proportion
of dams are reported as having ood attenuation as their original purpose. This implies that the efciency of
existing dams in damping the ood wave peak has seldom been optimized and is often not even known. In general,
2 F. Miotto, P. Claps, F. Laio and D. Poggi
1
in understanding and quantifying the effectiveness of a single dam in reducing ood wave peak the simple and
intuitive continuity equation is the only needed mathematical tool,
q
i
(t) q
o
(H(t)) =
dV (H(t))
dt
. (1)
where q
i
is the incoming ow discharge, q
o
is the outlet ow discharge, V and H are the lake volume and the
water surface level upstream the dam structure, respectively. Despite the simplicity of the hydraulic formulation,
the ood attenuation problem is complicated by a plethora of variables that have to be taken into account in
solving the continuity equation. These parameters are both typical of the dam structure (i.e. the link between
H(t) and both V (t) and qo(t)), and intrinsic in the hydrology of the basin upstream the dam (i.e. q
i
(t)). The
overall impact of articial reservoirs on individual oods depends on both the storage capacity of the dam (relative
to the ood volume) and the way the dam is operated. There are several problems in operating a dam for ood
storage purposes: a) rstly, dams are rarely used for ood storage alone and the optimum strategy for ood storage
and other purposes can conict. In general, the ideal is to keep the reservoir full and, to provide the maximum
efciency, to put it down in a drawn state in case of ooding; b) generally, water can be passed through a dam to
the river downstream by at least two means: by an outlet structure at a low level and by a spillway at a higher level.
The latter type of hydraulic structure may be gated to allow full control of the discharge. How to manage both the
outlet structure and the spillway during an extreme event remains an indeterminate, yet crucial, problem in ood
risk management.
In literature there are several interesting attempts to develop and apply complex models to a number of real cases
of ood propagation (Lpez-Moreno et al., 2002; Montaldo et al., 2004). All these models were developed to take
into account a number of hydrological and hydraulic variable as large as possible. This approach, based on a wider
and wider comprehension of the ood wave propagation, is surely valid and promising. However, the time and the
length scales required for modeling wide and complex rivers-dams systems can exceed the computational ability of
these type of models. In this sense, empirical and analytical models, although much simpler and more rudimental,
can be more suitable to quantify the impact of a large number of reservoirs on modulating ood waves.
Some attempts to build synthetic ood attenuation indexes, both qualitative and quantitative, have been carried out
in the past. In particular, it is worth citing the FARL index (Flood attenuation by reservoirs and lakes) proposed
by Scarrott et al. (1999). Being qualitative, such an index can not be considered a ood attenuation coefcient.
More recently, Piga et al. (2000) proposed an empirical index based on a multi-regressive technique taking into
account few variables, such as the area of the reservoir and of the upstream basin, and the length of the spillway
gates. Because of the empirical derivation, this index is strictly applicable to basins having features comparable to
those of the basins used in the calibration study.
Having in mind the necessity of improving the preliminary, large-scale, analysis on the reservoirs effects on oods,
in this paper we analytically investigate how the main features of a single dam, (e.g. its total area, the width of the
spillways, the basic river basin characteristics and the intensity of meteorological events), impact the ood peak
attenuation. To accomplish this goal, a number of assumptions regarding the way the dam is operated have been
adopted: a) we use a conservative and prudential approach that considers the reservoir always completely full at the
spillway crest level; b) for safety reasons and analytical manageability, we assume the outlet structures closed and
all the spillways ungated, so that no gate management must be handled. We are aware that these are very restrictive
assumptions; nevertheless we are interested in building a rather general and robust index, that can describe some
objective ood attenuation potentiality of existing reservoirs under real (i.e. non-optimal) conditions. As stated
before, the management optimization of existing dams is a much broader topic and is beyond the scope of this
paper.
3 AN ANALYTICAL INDEX FOR FLOOD ATTENUATION DUE TO RESERVOIRS
2 THE STATE OF ART
In literature, synthetic indexes for ood attenuation are build upon simple hypothesis about ood waves shape and
spillway typology. These indexes can be classied in qualitative and quantitative. Qualitative synthetic indexes
provide information about ood attenuation effects for basins without difference between articial or natural lakes,
to allowing comparison between different reservoirs. Quantitative synthetic indexes provide coefcients based on
few signicant parameter as reservoir storage volume, ood volume, ood peak time and duration. A qualitative
synthetic index is the Flood Attenuation by Reservoirs and Lakes (FARL) (Scarrott et al., 1999), which is a relation
between reservoir water surface A
L
, basin surface A
B
and catchment area A
C
. In this index the attenuation is
= (1

r)
w
where r = A
L
/A
B
and w = A
B
/A
C
. In basins containing more reservoirs, the FARL index is the
product of different estimated for every lake. This index has been built as it was a geomorphoclimatic parameter
having an inuence on the index ow, which in the FEH ((AA.VV., 1999)) is calculated as
Q
index
= a

X
b
i
, (2)
where X
i
are geomorphoclimatic parameters and a, b are coefcients to be estimated by regression.
For the design of a reservoir storage capacity, so with ood attenuation know a priori, a quantitative index proposed
by Mays (????). This index is based on the shape of the inowhydrograph and on a given maximumwater elevation
in reservoir. A promising way to evaluate ood attenuation indexes is based on multiregressive approaches. For
example, Piga et al. (2000) calculated a ood attenuation index using multiregressive analysis based on parameters
as L[m] (spillway width), A
L
[km
2
] (lake area) and A
B
[km
2
] (basin area). The authors obtained the following form
of a ood attenuation index
= 1 + exp[0.119 L
0.225
A
0.583
L
A
0.405
B
] (3)
valid for high return periods and for basins with climatical and hydrological characteristics similar to the Sardinia
basins, used to estimate the coefcients in equation 3.
3 THEORY BEHIND A NEW INDEX
The overall efciency of a reservoir in attenuating ood peaks is usually expressed through the ratio = Q
o
/Q
i
,
where Q
o
= Max(q
o
(t)) and Q
i
= Max(q
i
(t)). Hence, the core of the efciency problem is to evaluate the
behavior of q
o
and q
i
. This can be done, as stated before, through the continuity equation once q
i
(i.e. the
incoming hydrograph) and the mathematical relations q
o
(H) and V (H) are known. While the evaluation of q
i
is
an hydrological problem, q
o
(H) and V (H) are of easier evaluation. In particular, V (H) simply depends on the
storage volume geometry and is usually parameterized through a power law:
V (H) = V
o
+A
L
H
m
(4)
where V
o
and A
L
are the storage volume and the lake area when the water level is at the spillway crest, respectively,
H is the water depth over the spillway crest, and m is a tting coefcient. The value q
o
(H) depends both on the
typology of outlet structures and on the way they are operated. As stated before, we assume: a) the initial water
level coinciding with the lowest crest of the available spillways, (i.e. the ood storage volume is only the surcharge
storage), b) outlet structures at a low level are considered closed; only the spillway gates are used to evacuate the
4 F. Miotto, P. Claps, F. Laio and D. Poggi
1
incoming ood, and c) the reservoir is not "actively" governed: it passively impact the ood wave through ungated
spillways. Under these conditions q
o
(H) can be written as
q
o
(H) = c
f
L H
r
(5)
where c
f
is a coefcient of discharge, L is the total length of the spillways, and r = 3/2. From equations (4) and
(5) we get
V (t) = V
o
+A
L

q
o
(t)
c
f
L
m
r
(6)
Here we introduce the additional condition of linear reservoir, that assumes the equality between the two coef-
cients m and r. To this end it is to consider that m varies, at least for alpine reservoirs, between 1.2 and 1.6, then
making the linearity a plausible assumption. Based on these parametrization, the continuity equation becomes
q
i
(t) q
o
(t) = I
dq
o
(t)
dt
. (7)
that is an ODE having a general solution of the type:
q
o
(t) = e

t
I

C +

t
e
t
d
I
q
i
(t
d
)
I
dt
d

, (8)
where I = A
L
/(c
f
L), t
d
is a dummy variable, and C is the integration constant. Note that this solution does
not constrain the incoming ow to have a prescribed form. Nevertheless, to obtain a simple solution of equation
(8), as the base to build a viable attenuation index, we prefer to simplify q
i
(t) as much as possible while retaining
the fundamental information of a real ow hydrograph (i.e. the volume and the duration of the ood). In order to
clearly describe the procedure used to evaluate q
i
(t), a brief review on synthetic hydrographs is provided below.
4 PARAMETERIZATION OF THE INCOMING HYDROGRAPH
Plausible and robust assumptions on the incoming hydrograph represents perhaps the most sensitive part of a
reliable ood attenuation analysis. In fact, even though the volume and the peak are given, the shape variability
of q
i
(t) may markedly impact the evaluation of reservoirs efciency in attenuating a ood. A possible general
hydrograph shape, hereafter called PR hydrograph, can be written as q
i
(t) = Q
c
(
D
+t
D

D
), where Q
c
and

D
are the value and the reduction ratio of the ood peak discharge. The reduction ratio, as a function of a generic
hydrograph duration D, is dened as

D
= Q
D
/Q
c
where Q
D
= max

1
D

t+D
t
q
i
()d

. (9)
and it should characterise synthetic ood waves analogous to observed ones.
This approach, exemplied in gure (1) is very useful, being
D
parameterizable both analytically (Bacchi et al.,
1992; Fiorentino et al., 1987) and empirically AA.VV. (1975) once the hydrological features of the basin are known.
For example, Pianese & Rossi (1986) suggested
5 AN ANALYTICAL INDEX FOR FLOOD ATTENUATION DUE TO RESERVOIRS

D
= (1 +b D)
c
, (10)
where b = 1/(2 t
R
), t
R
is the basin delay time, c = 1 n. Based on this denition for
D
the expression of PR
becomes
q
cen
(t) = Q
c
[(1 + 2 b |t t
p
|)
c
2 b c |t t
p
| (1 + 2 b |t t
p
|)
c1
], (11)
where Q
c
and t
p
are value and position of the ood peak . Note that the hydrograph parameterized using equation
(10) is symmetric with respect to the peak.
A simplied synthetic hydrograph: The PR hydrograph can be simplied neglecting the time dependence as
q
i
(t, D) = Q
c

D
, where the incoming ow is now also variable with D. This simplication, produce rectangular
form that, while making q
i
easier to be treated mathematically, allow to carry out a maximizing approach on the
efciency by varying the duration D. This is done in the next section.
5 THE SYNTHETIC FLOOD ATTENUATION INDEX
Having dened the simplied hydrograph, the outgoing hydrograph (i.e. equation (8)) can now be written, as a
function of t and D, as
q
o
(t, D) = Q
c
e

D
K

1 e

t
I

, (12)
where K and I = A
L
/(c
f
L) take into account the main hydrological and hydraulic variables. In this expression,
instead of equation (10), a simpler parametrization of the reduction ratio,
DS
= exp(D/K), has been used. In
this formulation, the parameter K = 1/(b c) can be derived equating the linear series expansion of equation (10)
and
DS
.
To evaluate the efciency of the reservoir in attenuating the ood peak, the maximum ow discharge and the
relative ood duration have to be dened from equation (12). In gure 5 both q
i
(t, D) and q
o
(t, D) are shown for
several duration, D. Note that, for each q
i
(t, D) the associated q
o
(t, D) is a monotonic growing function and its
maximum is always at t = D. Figure 5 also shows the function obtained merging the maximum value of q
o
(t, D)
(i.e. the so-called q
o
(D, D)). Note that for small D the inow is very large but the ood volume is limited. On the
other hand, for large D the peak is modest and the volume is very large. In the rst case, being the ood volume
modest, the reservoir efciency is very high but, because of the large peak, the maximum is still considerable.
On the contrary, in the second case the efciency is scarce but, being the peak already very small, the maximum
outow is also very small. This behavior implies that the maximum outlet ow is for an intermediate duration,
D
C
, for which both the ood volume and the wave peak are still relevant. From a quantitative perspective, we
analytically get the absolute maximum of q
o
(D, D) = max[q
o
(t, D)] as
q
o
(D
c
) = max[q
o
(D, D)] = Q
c
e

D
c
K

1 e

D
c
I

= Q
c
K
I +K

I +K
I
K
I
, (13)
where D
c
= I log[(I + k)/I] is the critical ood duration for which the ood attenuation is minimized. In
comparison to the solutions obtained by Pianese & Rossi (1986) and by Fiorentino & Margiotta (1997) relation
13 allows one to obtain a next simplied attenuation index (SFA) by
6 F. Miotto, P. Claps, F. Laio and D. Poggi
1

Q
D1
Q
D2
Q
Di
Q
Dn
D1
D2
Di
Dc
q
i
Q
Dc q
o
(Dc) Max |q
o
(D, D)|
t, D
t

Q

Q

q
o
(D, D)

q
o
(t, D)

q
i
(t, D)

Figure 1: Left panel: Rectangular synthetic hydrographs as dened in equation 9. Right panel: Representation
of incoming hydrographs, q
i
(t, D), along with the associated outlet ow, q
o
(t, D). The maximum outow and the
relative critical duration, D
C
, are also shown.
SFA =
1
R + 1
, where R = I/K =
A
L
(1 n)
2 t
R
c
f
L
. (14)
Analyzing qualitatively the SFA index one can note that the larger the storage area (i.e. A
L
) the smaller the
SFA. Similarly, having assumed the reservoir unlimited, the narrower the spillway gate the smaller the SFA.
This is coherent with what expected; in fact, small spillways and wide storage volumes have the effect of retaining
upstream larger amount of water, then reducing the ood peak. On the contrary, large coefcient n and wide basin
area involve great ood volumes, hence reducing the efciency of a given reservoir. In the next section a more
quantitative analysis is carried out.
6 SFA VALIDATION VIA A CASE STUDY
To check the reliability of the SFA index, a sensitivity analysis has been carried out using a number of different
synthetic hydrographs and sixteen articial reservoirs existing in the Toce basin, as described below.
The Toce basin is an alpine basin located in the North of the Piemonte region, in Italy. It has also the 10% of
its area in the southern Switzerland. The total area is about 1780 km
2
, and the main stream length is of 80 km,
estimated from its sources near San Giacomo Pass, to the Maggiore Lake. This basin is characterized by steep
narrow valleys with glacial origin. Elevations within the basin range from 4623.5 m a.s.l. (Rosa Mountain) to 193
m a.s.l. of the Maggiore Lake, with an average of about 1500 m s.l.m.. Climate within the basin is inuenced by
the high elevations and by heavy precipitations. The Toce basin contains sixteen dams, whose main features are
shown in Table (1). Given that these reservoirs are not a signicant group to constitute a robust case study, we
have considered a much higher number of virtual reservoirs, changing at random the six features shown in Table 1,
and removing only the least reasonable resulting congurations. These virtual reservoirs, along with the real ones,
have been used to check the capability of the SFA index to evaluate the reservoir efciency in attenuating oods.
7 AN ANALYTICAL INDEX FOR FLOOD ATTENUATION DUE TO RESERVOIRS
Table 1: Principal characteristic of the sixteen articial reservoirs in the Toce basin.
RESERVOIR A
B
[km
2
] t
R
[ore] n[] Q
m
[m
3
/s] A
L
[km
2
] L[m]
Agaro 10.60 0.556 0.51 19.55 0.650 36.00
Agrasina 17.83 1.128 0.50 31.66 0.0135 12.30
Alpe Cavalli 23.74 1.253 0.56 29.03 0.467 18.00
Alpe Larecchio 3.02 0.464 0.50 6.90 0.183 25.00
Busin Inferiore 2.54 0.357 0.46 6.13 0.336 10.00
Campliccioli 35.19 1.401 0.56 38.40 0.324 13.50
Camposecco 4.08 0.497 0.56 7.28 0.383 105.85
Ceppo Morelli 121.00 2.099 0.56 97.248 0.040 84.00
Devero 25.38 0.866 0.51 36.40 0.96 25.50
Lago Avino 5.32 0.497 0.54 9.04 0.43 45.25
Lago Cingino 3.06 0.431 0.56 5.78 0.14 65.00
Morasco 35.26 1.305 0.46 42.59 0.65 45.00
Quarazza 25.81 1.002 0.55 35.26 0.028 31.20
Sabbione 14.37 0.818 0.46 22.93 1.21 10.00
Val Toggia 10.32 0.720 0.44 17.71 0.81 28.00
Vannino 11.94 0.727 0.47 20.75 0.48 22.80
To carry out the sensitivity analysis, a number of different synthetic inow hydrographs are considered, to which
numerically implement the continuity equation. Here we use both triangular and PR hydrographs. For each of
these forms we can use the canonical expression (i.e. with a central peak) or similar hydrographs having the peak
at t = 0 and t = D
max
, where D
max
= 2 t
p
is the ood duration. The PR hydrograph can be respectively written
as
q
in
(t) = Q
c
[(1 +b t)
c
c b t (1 +b t)
c1
], (15)
q
fin
(t) = Q
c
[(1 +b |t t
max
|)
c
b c |t D
max
| (1 +b |t D
max
|)
c1
]. (16)
In Figure (6) the three PR forms and the three triangular ones are shown. Note that both the volume and the
value of the peak discharge are constant for the six hydrographs. Using the continuity equation (i.e. equation
1) six values of
i
have been calculated for each reservoir:
1
,
2
,
3
are for the triangular hydrographs having,
respectively, right, central, and left peak. Similarly,
4
,
5
,
6
are the efciencies for the PR hydrographs.
The efciency value obtained,
i
, along with the SFA index, shown in Figure (6) as a function of the mean
efciency (obtained from the six
i
). From, Figure (6) it is apparent that the maximum and the minimum value of
are very different for each reservoir. This nding demonstrates the sensitivity of the reservoir efciency to the
hydrograph shape. Moreover, for each synthetic hydrograph, the ood wave attenuation presents a well dened
behavior. Not surprisingly, while the worst efciency is always for the triangular hydrograph having the peak
at t = t
p
(
1
in gure (6)) the best efciency is for the RP hydrograph with the initial peak (
6
in gure (6)).
This strong sensitivity of on the ood shape conrms the difculty in dening an unique level of efciency of
reservoirs in attenuating oods. Nevertheless, it is encouraging that the SFA index, both for the real and the
virtual reservoirs, is well included into the variability due to the hydrograph shape, without any bias. Furthermore,
note that most of the SFA values are bounded by
2
and
5
that represent the triangular and the PR hydrograph
8 F. Miotto, P. Claps, F. Laio and D. Poggi
1
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
0.0
0.2
0.+
0.6
0.8
1.0
0.0 0.2 0.+ 0.6 0.8 1.0
tJD
max
[-|
q
i
n
{
t
)

J

Q
T

[
-
|
Figure 2: The six hydrographs used to carry out the sensitivity analysis.
for the central peak case. This nding is particularly encouraging being the central peak cases surely the closer
description of real hydrographs. Moreover, the triangular hydrograph with a central peak is often used by the
professional engineers to design or verify the volume of articial reservoirs.
6.1 An expeditious SFA index
When one is interested in roughly evaluating the efciency of a number of articial reservoirs in attenuating the
ood peak without knowing the specic features of both the basin and the reservoir, such as n, c
f
and t
R
, an
expeditious index can be apply instead of the complete SFA index. Here an expeditious SFA index, hereafter
called SFA
e
, is dened based on the following considerations. a) The discharge coefcient can to assume a
constant value; for example, we assume c
f
= 0.31, given that most of the existing spillways have a standard
Creager weir. b) For every reservoir, a standard coefcients n can be assumed (i.e. in this paper the mean value
computed over the whole Toce basin has been used). c) The lag time can be taken as t
R
= A
0.5
B
. Using these
simplications the parameter R in equation (14) becomes
R =
s A
L

A
B
L
. (17)
where s is a coefcient that takes into account the previous simplications for n and c
f
(i.e. s 100 in our case
study). In gure 6 b the SFA
e
is shown along with the actual reservoir efciency for the six synthetic hydrographs.
Note that the evaluation of is comparable with that of the more complex SFA index. In particular, the SFA
e
index is still bounded by
2
and
5
and does not show any particular bias.
CONCLUSIONS
Flood control through dams storage volumes represents a challenge for the application of statistical procedures
related to ood risk planning. In fact, the high specic knowledge of the deterministic (hydraulic) part of the phe-
nomenon contrasts with the lack of methods able to grossly but systematically assess performances and attitudes
9 AN ANALYTICAL INDEX FOR FLOOD ATTENUATION DUE TO RESERVOIRS
Figure 3: The real and virtual reservoir efciency;
1
,
2
, are for the triangular hydrograph with right and central
peak;
5
and
6
are for the PR hydrograph with central and left peak.
of systems of dams towards ood risk reduction. As a contribution toward this goal, a new synthetic index for
ood attenuation, SFA, has been proposed here. Such an index takes into account both the main features of arti-
cial reservoirs, such as the lake surface and the spillways width, and the basic characteristics of river basins and
rainfall forcing. Despite the fact that several assumptions were used in the analytical development of the index, the
directness of the approach, makes the SFA index an apparently robust tool to preliminarily sort out complex dam
networks.
The index performances were in fact tested by means of a sensitivity analysis, in which we apply the SFA to a
considerable number of real and hypothetical dams, built around the complex system of reservoirs in the Toce
basin (Northern Italy). Each one of this reservoirs has been subjected to a number of synthetic hydrographs,
characterized by different shapes but identical volume and peak, computing the reservoir efciency in reducing the
hydrograph peak, . These actual values of were compared with the SFA index, which is intended as a ood
shape-independent value. The comparison showed that SFA values fall in the middle part of the distribution
range, and present no apparent bias. An equally good description of the overall ood attenuation capability of the
considered dams was also demonstrated by an expeditious index, SFA
e
, merely based on basin and lake area and
on spillway length.
REFERENCES
AA.VV., Natural Environmental Research Council, Estimation of ood volumes over different duration, Flood
Studies Report, volume 1. London (1975).
AA.VV., Flood Estimation Handbook: procedures for ood frequency estimation., volume 5 volumes. Institute of
Hydrology, Wallingford (1999).
Bacchi B., Brath A., Kottegoda N., Analysis of the relationships between ood peaks and volumes based crossing
properties of river ow processes. Water Resources Research (1992), 28, pp. 27732782.
10 F. Miotto, P. Claps, F. Laio and D. Poggi
1
Fiorentino M., Margiotta M., La valutazione dei volumi di piena e il calcolo semplicato delleffetto di laminazione
di grandi invasi, Tecniche per la difesa dallinquinamento. Ed.Bios (1997).
Fiorentino M., Rossi F., Villani P., Effects of the basin geomorphoclimatic characteristics on the mean ood reduc-
tion curve, Proc. Of 18th Annual Pittsburgh modelling and Simulation Conference, volume 5 (1987).
Lpez-Moreno J., Beguera S., Garca-Ruiz J., Inuence of the yesa reservoir on oods of the aragn river. Hydrol-
ogy and Earth System Sciences (2002), 64, pp. 753762.
Mays L., Hydraulic design handbook. McGraw-Hill (1999).
Montaldo N., Mancini M., Rosso R., Flood hydrograph attenuation induced by a reservoir system: analysis with a
distributed rainfall-runoff model. Hydrological Processes (2004), 18, pp. 545563.
Pianese D., Rossi F., Curve di possibilit di laminazione delle piene. Giornale del Genio Civile (1986), 4,5,6,
pp. 131148.
Piga E., Saba A., Salis F., Sechi G., Distribuzione probabilistica delle portate massime annue laminate da un
invaso con soratore superciale, XXVII Convegno di Idraulica e Costruzioni Idrauliche, volume 3. Genova
(2000).
Scarrott R., Reed D., Bayliss A., Indexing the attenuation effect attributable to reservoirs and lakes., Flood Esti-
mation Handbook, volume 45. Institute of Hydrology, Wallingford (1999).