Energy 61 (2013) 541e554

Contents lists available at ScienceDirect

Energy
journal homepage: www.elsevier.com/locate/energy

Assessment of the impacts of tidal stream energy through

high-resolution numerical modeling
V. Ramos a, *, R. Carballo a, M. lvarez a, M. Snchez a, G. Iglesias b
a
University of Santiago de Compostela, EPS, Hydraulic Eng., Campus Univ. s/n, 27002 Lugo, Spain
b
University of Plymouth, School of Marine Science and Engineering, Drake Circus, Plymouth PL4 8AA, UK

a r t i c l e i n f o a b s t r a c t

Article history: When planning the installation of a tidal farm, the disturbances on the marine environment associated
Received 20 April 2013 with its operation must be studied in detail. The objectives of this paper are to assess the impacts on the
Received in revised form hydrodynamics (water level and ow velocity) and to determine how these impacts can alter the tidal
12 July 2013
resource. For this purpose, a high-resolution model of Ria de Ribadeo (NW Spain) is used to describe the
Accepted 25 August 2013
Available online 23 September 2013
potential effects resulting from the operation of two prospective tidal farms. Two different scenarios of
extracted power from the ow (high and low) are analyzed. Overall, it is found that the impact on the
water level is negligible, but that on the ow velocity is signicant. The velocity is reduced upstream and
Keywords:
High-resolution model
downstream the farm, and increased beside it. These effects are enhanced in the scenario with the higher
Tidal stream turbines power extraction. Finally, these modications in the ow pattern alter the available energy density at the
Tidal farm impacts tidal turbine, with a reduction of 21% and 12% for the high and low levels of power extraction,
Galician ria respectively.
2013 Elsevier Ltd. All rights reserved.

1. Introduction
The aim of increasing the share of renewable energy sources to
the total energy production [1e4] has brought a substantial in-
crease in the interest in marine energies over the last years [5e9].
Within them, tidal stream energy, which taps the kinetic energy of
tidal currents has been steadily gaining ground [10e14] for its ad-
vantages in terms of: (i) predictable resource derived from its as-
tronomical dependence; (ii) nonexistence of extreme ows (unlike
wind energy, for example) (iii) high load factor due to the uid
properties [15] and (iv) minimal land occupation and visual impact
[16]. On the other hand, the main disadvantages are: (i) the uc-
tuating nature of the resource (astronomical dependence), which
requires efcient control systems to optimize the power output of
the converters [17e19]; (ii) the fact that the tidal turbine technol-
ogy is still in its infancy [20] and (iii) the possible impacts on the
marine environment. On these grounds, recent research has shown
that these impacts may alter numerous environmental variables
such as: tidal regime [21], hydrodynamics [22], sediment transport
* Corresponding author.
E-mail address: josevictor.ramos@usc.es (V. Ramos).
and water quality [23e28]. These impacts may not be restricted to
the vicinity of the tidal arrays and in some cases could be noticed a
few km away from the arrays (far-eld impacts) [24]. In addition,
the operation of a tidal stream farm modies the tidal resource
varying the power density distribution and decreasing the available
energy resource [21,29]. Therefore, when planning a new tidal
stream farm all these factors need to be studied in detail.
The goal of this study is to assess the potential impacts of the
tidal energy extraction on the ow using as case study Ria de
Ribadeo, a specic type of estuary in NW Spain. With its main axis
extending 10 km along the northesouth direction, Ria de Ribadeo
has a mean width of 800 m and a total surface area of 8.59 km2
(Fig. 1). The main fresh water contribution to the ria is the River Eo,
with an average ow rate of 18.83 m3 s
1. The tidal regime is purely
semidiurnal, with a Form Factor F 0.068; the maximum tidal
range is 4.6 m, which produces a substantial tidal prism up to
20.6 Hm3 in spring-neap tides. As a result of this tidal prism and the
coastline shape, signicant tidal ows occur in certain areas of the
estuary, therefore, Ria de Ribadeo constitutes an excellent location
for installing a tidal stream farm and, consequently, for studying the
impacts that this plant would cause on the estuarine hydrody-
namics and the available tidal energy resource. For this purpose, a
numerical model of the ria, which simulates the tidal energy
extraction by adding a retarding force to the local momentum

0360-5442/$ e see front matter 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.energy.2013.08.051
542 V. Ramos et al. / Energy 61 (2013) 541e554

44
Outer Ria
100 m

43.5 43.6

50 m

43 20 m
Pt. Cruz

ADCP 1

Latitude()
Latitude()

Port of ADCP 2
Ribadeo Middle
42.5
Meteorological Ria
Station

42
43.5
Spain
Portugal

Inner Ria

River Eo

9 8.5 8 7.5 7 7.1 7.05 7

Longitude() Longitude()

Fig. 1. Ria de Ribadeo (right) in Galicia, NW Spain (left).

equations, is implemented and successfully validated based on eld
data. After validation, the model is used to illustrate the effects
associated with the operation of two tidal farms on the ow ve-
locities and water levels; and to determine how the modications
on the ow can alter the power density distribution and, therefore,
modify the available tidal energy resource.
2. Materials and methods
2.1. Numerical model
The model, Delft 3D FLOW, is a nite difference code that solves
the NaviereStokes equations along with the transport equation.

Fig. 2. Computational grid.

 V. Ramos et al. / Energy 61 (2013) 541e554 543

As in previous assessments of the tidal resource [30] the depth- Table 1

averaged equations were used in their baroclinic form so the Tidal constituents at the ocean boundary of the ria grid.

model takes into account the ow driven by horizontal density Constituent Amplitude (m) Phase ( )
gradients, which can be relevant in a semienclosed bodies of water M2 1.2513 91.40
where fresh and ocean waters mix. Thus, the equations solved by S2 0.4396 122.17
the model are: N2 0.2647 71.45
K2 0.1227 120.23

vz vd zU vd zV
K1 0.0714 73.27
Q; (1) O1 0.0623 324.20
vt vx vy P1 0.0216 64.82

8 9
>
>
> Zx 0 >
>
>
>
> vU U vU V vV
f $V
g vx
g vr ssx
sbx 2 >
>
>
> dz y V U>
> stress acting on the sea surface, and sbx and sby are the shear stress
> vt vx vy vx r vx ro $d z h >
>
<
o
>
= components at the bottom; nally, in the transport equation c

d
; stands for salinity or temperature, Dh is the horizontal eddy diffu-
>
> Zx 0 >
>
>
> ssy
sbx >
> sivity, ld represents the rst order decay process, and R is the source
>
> vx vr >
vV vV vV g
> vt U vx V vy
f $U
g vx
ro
> dz y V2 V> >
> term per unit area.
>
: vy ro $d z h >
;

d For the spatial discretization the model uses the Arakawa-C grid,
a staggered grid in which water levels z are computed at grid cell
(2)
centers, whereas ow velocity components (U and V) are dened at
the mid-points of the grid cell faces to which they are perpendic-
vz dc vd zUc vd zVc ular. The discretization of the horizontal advection terms in the

vt vx vy momentum and transport equations is carried out by means of the
Dh V2 c
ld d xc R; (3) Cyclic method. For the temporal discretization the model resorts to
a semi-implicit ADI algorithm with two time levels per iteration.
Equations (1) and (2) express the conservation of mass and Three computational grids were combined in the implementa-
momentum, respectively. Equation (3) is the transport equation, tion of the model (Fig. 2): (i) in the inner ria a coarse grid, with a
which is solved both for salinity and temperature. In these equa- resolution of 50  50 m, spanning an area of approximately 5 km2
tions, d is the local water depth relative to a reference plane; U and was used, (ii) in the middle ria, where the strongest tidal ows
V stand for the vertically integrated eastward and northward occur, a ne grid with a much higher resolution (5  5 m) was used
components of the velocity, respectively; Q represents the intensity in order to properly simulate the operation of the tidal farm, and
of mass sources per unit area; f is the Coriolis parameter, yh is the nally (iii) a coarse grid covering the outer ria and its approaches
kinematic horizontal eddy viscosity, ro is the reference density, r0 is (10 km2) was used with an initial resolution of 50  50 m, which is
the anomaly density, ssx and ssy are the components of the wind gradually increased northward up to the outer boundary, where it
reaches a cell size of 50  150 m. The bathymetry (Fig. 3) was
digitized from the nautical chart #4071# of the Spanish Hydro-
graphic Institute. In the inner ria the chart was supplemented with
maps from the 1:5000 cartography of the Galician Regional Gov-
ernment (Xunta de Galicia).
Along the ocean boundary a Dirichlet boundary condition was
imposed with the sea level prescribed as a function of time using
the seven major tidal harmonics (Table 1) obtained from global
ocean tide model TPXO 7.2. The boundary conditions at the land
margins were null ow and free sleep (zero shear-stress). At the
sea-bed the shear stress was computed from:

1 p 1 p
sbx 2
ro gU U2 V 2; sby 2
r o gV U2 V 2; (4)
C2D C2D

where C2D is the 2D Chzy coefcient, which depends on the water

depth and bottom roughness. Its dependency on bottom roughness
may be written in terms of the Manning coefcient, n:

p
6
d z
C2D : (5)
n
The value of the Manning coefcient was determined as a
function of the water depth following Dias and Lopes [31].

Table 2
Salinity and temperature at the ocean boundary of the ria grid.

Ocean boundary

Salinity (ppt) 35.642

Temperature ( C) 15.635
Fig. 3. Bathymetry of Ria de Ribadeo.
544 V. Ramos et al. / Energy 61 (2013) 541e554

ADCP 1
4.6

Water level (m)

2.3

0
10/15 10/18 10/21 10/24 10/27 10/30 11/02 11/05

ADCP 2
4.6
Water level (m)

2.3

0
10/15 10/18 10/21 10/24 10/27 10/30 11/02 11/05

Fig. 4. Computed (D) and observed (
) time series of sea level during the validation period.

Wind stress on the surface was determined from

 
!  ! !
ss CD ra U10 U10 ; (6) Table 3
Correlation coefcient between observed and simulated time series of current ve-
locity and water level. [RU, correlation coefcient of eastward current velocity
!
where U10 is the wind velocity vector at a height of 10 m above the component; RV, correlation coefcient of northward current velocity component; Rx,
sea surface, ra is the air density, and CD is the dimensionless drag correlation coefcient of water levels].

coefcient. This was modeled following Smith [32] for wind ve- RU RV Rx
locities below 6 m s
1,
ADCP 1 0.9451 0.9556 0.9919
ADCP 2 0.9449 0.9710 0.9938
CD 1:1  10
3 ; (7)

ADCP 1

0.25
U (ms1)

0.25

10/15 18/15 10/21 10/24 10/27 10/30 11/02 11/05

ADCP 1

0.75
V (ms1)

0.25
0
0.25

10/15 18/15 10/21 10/24 10/27 10/30 11/02 11/05

ADCP 2

0.1
U (ms1)

0.1

10/15 10/18 10/21 10/24 10/27 10/30 11/02 11/05

ADCP 2

1.0
V (ms1)

0.0

1.0

10/15 10/18 10/21 10/24 10/27 10/30 11/02 11/05

Fig. 5. Computed (D) and observed (
) time series of sea level during the validation period.
 V. Ramos et al. / Energy 61 (2013) 541e554 545

and following Yelland [33] for wind velocities over 6 m s
1 2.2. Numerical modeling of tidal stream power extraction

With the purpose of modeling the operation of a tidal farm, the

CD 0:50 0:071U10  10
3 ;
Salinity and temperature were prescribed at the ocean boundary
(Table 2) based on the data from the numerical model ROMS
(Regional Ocean Modelling System) run by the Spanish Oceanog-
raphy Institute [34].
(8) 2DH water momentum equations (Eq. (2)) have been modied by
adding the terms (Mx, My), which represent the contribution of
momentum generated by the external forces (i.e. tidal stream tur-
bine) in the x and y directions, respectively [21,24]. Consequently,
the momentum equations in the x and y directions can be
expressed as:

8 Zx 9
>
> vx vr0 ssx
sbx >
vU vU vV g
> vt U vx V vy
f $V
g vx
r
> dz yh V U Mx>
2
>
>
>
>
<
o vx ro $d z >
>
=

d
: (9)
>
> Z x >
>
>
> vr0 ssy
sbx >
>
>
> vx >
: vV vV vV g
vt U vx V vy
f $U
g vx
ro dz y V2 V My >
;
vy ro $d z h

d

Fig. 6. Mid-ood and mid-ebb velocity in Ria de Ribadeo computed with the numerical model.
546 V. Ramos et al. / Energy 61 (2013) 541e554

The added momentum terms can be written as: takes into account both the ow and water depth information,
8 9 which makes it particularly appropriate for rias or estuaries where
< Mx Frx = the water depth constitutes a limiting factor [11]. Moreover, the TSE
; (10) takes into consideration both the mid-ebb and mid-ood velocities,
: M Fy ;
y r which is important in estuaries and rias where a tidal asymmetry
exists as is the case of Ria de Ribadeo.
where Fx and Fy, are, respectively, the components in the x and y The TSE index is expressed [35] as:
directions of the ow retarding force per unit volume due to the
presence of a TST (tidal stream turbine). According to Newtons  
x
third law these components are equal and opposite to the thrust TSE Vf3 Ve3 h; (12)
force exerted by the ow on the TST [29]. Therefore the momentum 2V03 h0
contribution can be written as;
  where, Vf and Ve are respectively the mid-ood and mid-ebb depth-
8 9
> C A  > mean velocities, h is the mid-tide water depth, V0 is a characteristic
< Mx
12 VTc U U =
  ; (11) ow velocity, h0 is a characteristic water depth, and x is a penalty
>
: My
1 CT AU V >
; function, which is given by:
2 Vc

where A is the cross-sectional area of the turbine rotor, r is the x 0 if h
D2h  h1 ;

water density, jUj is the magnitude of the ow velocity, Vc is the
volume of control, U and V are the magnitudes of the x and y x 1
h2
h1
if h1 < h
D2h  h2 ; and (13)
components of the velocity, and CT is the thrust coefcient.
x 1 if h
D2h  h2 ;

2.3. Tidal site selection: the TSE approach where Dh is the maximum tidal range, h1 and h2 are, respectively,
the lower and upper boundaries of the penalty range. According to
The selection of the site for a prospective tidal farm was carried Ref. [35], h1 and h2 were set at 2 m and 5 m, respectively, h0 at 2 m,
out using the TSE (Tidal Stream Exploitability) index [35]. The TSE and V0 at 1.5 m s
1.

Fig. 7. Power density in Ria de Ribadeo at mid-ebb and mid-ood of a mean spring tide.
 V. Ramos et al. / Energy 61 (2013) 541e554 547

3. Results and discussions

3.1. Model validation

The model was successfully validated based on eld data of tidal

level and ow velocity. The period of validation extended from 15
October 2010 to 5 November 2010, during which two ADCPs
(Acoustic Doppler Current Prolers) were deployed in the ria. The
ADCPs (Sontek Argonaut-XR) measured the tidal level and the ve-
locity and direction of the ow at two points in the ria. ADCP 1 was
set up in the inner ria in front of the Port of Ribadeo, whereas the
ADCP 2 was installed in the ria mouth, south of Pt. Cruz (Fig. 1). In
addition to the tidal level and ow velocity, the model was forced
with wind and river discharges. Wind data, (10 min measurements
of wind velocity and direction), were collected from a meteoro-
logical station, located in an open area near the shoreline (Fig. 1).
The hourly discharge of the river Eo was obtained from a gauging
station at the river mouth.
The initial hydrodynamic conditions were null velocity and
surface elevation (cold start). The simulation extended from 16
September 2010 to 5 November 2010. The purpose of the rst 30
days, the spin-up period, is to adjust dynamically the ow eld so
that the initial conditions (the cold start) do not affect the nu-
merical results during the period of interest (15 Octobere5
November) [36].
An excellent agreement was found between the observed and
computed time series of tidal level (Fig. 4). In the case of the ow
velocity (Fig. 5), the agreement in general is very good, with only
minor differences in the maximum velocities of some tidal cycles.
Overall, the model reproduces properly the behavior of the
Fig. 8. TSE (Tidal Stream Exploitability) index in Ria de Ribadeo.

Fig. 9. Location and conguration of the proposed tidal farms in Ria de Ribadeo.
548 V. Ramos et al. / Energy 61 (2013) 541e554

hydrodynamics, as conrmed by the high correlation coefcients not present a high TSE values in spite of its high tidal ow, because
obtained (Table 3). of its limited water depth, which makes it inappropriate for the
installation of a tidal farm. Therefore, areas II and III were selected
as the sites to study the operation of a tidal farm and its impacts on
3.2. Resource assessment and site selection
the hydrodynamics and the available tidal energy resource. The
prospective tidal farms are conformed by eight TST (tidal stream
Upon validation, the model was used to calculate the unaltered
turbines) with a 5 m diameter. The selection of the TST diameter
tidal ow velocities over the ria (baseline case). A spring-neap cycle
was based on some prototypes of horizontal axis TSTs, like the
was studied, from 16 March 2012 to 30 March 2012. As mentioned,
Evopod Turbine [38], which are especially appropriate in depth
a spin-up period of 30 days was used. Fig. 6 shows the ow ve-
limited areas, like Ria de Ribadeo. A triangular conguration for the
locities computed by the model at mid-ood and mid-ebb of a
tidal farm was chosen, with the TSTs separated 15 m (three-di-
mean spring tide. The strongest ows occur in the main channel
ameters) from each other. The location and conguration of the two
(area I) and in front of the Port of Ribadeo (areas II and III), with
tidal farms are shown in Fig. 9. Finally, to simulate the operation of
peak values of 1.92 m s
1, 1.56 m s
1 and 1.89 m s
1, respectively. It
the TSTs of the farm the momentum equations were modied as
is also noteworthy that a tidal asymmetry exists, with the ood
indicated in Section 2.2 for the cells of the computational domain
stronger than the ebb. This ood dominant asymmetry can be
where the TSTs are located. The high resolution of the model
caused by the different layout of the estuary at different tidal levels,
(5  5 m) in the area of interest allows the effect of each TST to be
or by the distortion of the tidal wave as it propagates over the
simulated individually, which represents an upgrade in this kind of
coastal shelf [37]. In addition, the power density associated with
studies. Two values of the thrust coefcient (CT) were introduced in
the tidal ow is represented in Fig. 7 at mid-ebb and mid-ood of a
the model, representing two different cases of power extraction
mean spring tide. On the ooding tide the highest power densities,
(high and low). Following [23,39,40] the values of the CT used were
above 1400 W m
2, were found in areas I, II, and III. During the ebb,
0.8 (case 1) and 0.4 (case 2).
the power density is considerably lower due to the pronounced
tidal asymmetry, only area I stands out with a power density close
to 1350 W m
2. 3.3. Effects of energy extraction: hydrodynamics
Based on the ow computed by the model, the TSE index was
determined according to Equation (7) (Fig. 8). Areas II and III, stand In order to assess the impacts on the water levels resulting from
out with values of TSE exceeding 3 over roughly circular regions the operation of the tidal farms, the differences between the water
with surface areas of 1800 and 2000 m2, respectively. Both areas levels with and without the farms were calculated, at mid-ebb and
constitute promising tidal stream sites. On the contrary, area I, does mid-ood of a mean spring tide, and the results obtained are

Fig. 10. Water level difference between altered and unaltered situation. Case 1.
 V. Ramos et al. / Energy 61 (2013) 541e554
represented in Figs. 10 and 11 for cases 1 and 2, respectively.
Overall, it can be observed that the operation of the two tidal farms
results in a higher and lower water level upstream and downstream
the TSTs, respectively. Comparing the two cases of power extrac-
tion, the maximum difference in the water levels is close to 0.025 m
and 0.015 m for Case 1 and 2, respectively. Therefore, the effects on
the water level are not signicant since these variations are barely
noticeable in an area where the mean water depth exceeds 8 m.
Finally, in accordance with other studies [21,24], it can be observed
that the biggest differences in the water levels occur for the sce-
nario in which both farms are operating with the larger power
extraction (case 1).
On the other hand, to measure the effects of the farms on the
tidal currents, the differences in the velocities with and without the
farms were computed at mid-ebb and mid-ood of a mean spring
tide. The results obtained for the areas of interest are shown in
Figs. 12 and 13 for cases 1 and 2, respectively. In view of both gures
it can be concluded that the velocities are reduced both upstream
and downstream of the farms. These reductions are more signi-
cant downstream of the farms reaching values close to 0.25 m s
1
and 0.15 m s
1 for the cases 1 and 2, respectively. Furthermore,
downstream of the farm, the size of the area affected by the
reduction of the velocity is much larger than the area affected
upstream of the farm. It is important to note that there are increases
in the ow velocity on both sides of the farms, ranging between
0.12 m s
1 and 0.06 m s
1 for cases 1 and 2, respectively, since the
ow tries to avoid the farms and is constrained on both sides of the
farms, resulting in a rise in the ow velocity. Comparing the im-
pacts for cases 1 and 2, it can be observed that the effects on the
Fig. 11. Water level difference between altered and unaltered situation. Case 2.
549
ow velocity prevail in a larger area (1.7 km away from the farm) for
case 1, in which a larger power extraction is simulated.
As a novelty in this type of studies, a high resolution model
mesh was used (grid size of 5  5 m) in the areas of interest,
allowing the assessment of the behavior of the ow within the farm
and in each TST. The variations of the velocities inside the farm
follow the same trend, upstream and especially downstream of a
single TST the velocity is reduced. The velocity experiences its
biggest drop behind each individual TST and gradually recovers as
the distance from the TST grows. Moreover, the ow between the
TSTs placed side-by-side increases its velocity slightly.
Finally, comparing the two farms it can be observed that farm 2
presents a bigger impact on the ow than farm 1. This may be due
to the fact that the area where farm 2 is located presents stronger
tidal ow velocities, so the presence of the tidal farm causes larger
disturbances in the ow, both in the velocity drop and the size of
the affected area. For the same reason, the impacts at mid-ood are
larger than at mid-ebb for both farms, since a tidal asymmetry with
ood dominance is present in the ria.
3.4. Effects of energy extraction: power density distribution and
tidal resource
As indicated in the previous section, the presence of the tidal
farms alters the ow, which affects the power density distributions.
The differences in the tidal stream power density at mid-ood and
mid-ebb of a mean spring-tide with and without the tidal farms are
shown in Figs. 14 and 15 for cases 1 and 2, respectively. Overall,
there is a signicant impact upstream and especially downstream
550 V. Ramos et al. / Energy 61 (2013) 541e554
of the farm, with a substantial decrease in the power density. Due to
the tidal asymmetry with ood dominance present in the ria, at
mid-ood, the effects are still apparent a few kilometers away from
the farms. On the contrary, at mid-ebb the reductions in the power
density are less signicant and so is the size of the affected area.
There is also an increase in the power density on both sides of the
farms, which can be attributed to the fact that the ow that is not
harnessed by the farm is forced to circulate through a constricted
area, increasing its velocity and, consequently, its power density.
Comparing the effects for both farms, the drop in the power
density for farm 2 at mid-ood reaches values close to 600 W m
2
and 350 W m
2, for cases 1 and 2, respectively. On the other hand,
at mid-ebb the impacts are less signicant, with reductions in the
power density of 400 W m
2 and 250 W m
2 for cases 1 and 2
respectively. With regard to farm 1, at mid-ood, the reduction in
the power density is close to 500 W m
2 and 350 W m
2 for cases 1
and 2 respectively. At mid-ebb the impacts are less signicant with
a decrease in the power density of 350 W m
2 and 200 W m
2 for
farms 1 and 2, respectively.
In order to quantify the reduction of the available energy on the
incoming ow of a single TST, the time distribution of the power
density on the incoming ow at the TST was computed for the two
cases of power extraction (cases 1 and 2) and the baseline case
(without TSTs). For the 14-day period considered (spring-neap cy-
cle), the energy density available in the incoming ow of the TST is
the area under the curve in Fig. 16. A numerical integration yields
1.56 105 Wh m
2 and 1.73 105 Wh m
2 for cases 1 and 2, respec-
tively. Comparing these values with the obtained for the baseline
case, 1.96 105 Wh m
2, a reduction of 21% and 12% can be observed
in the energy density of the incoming ow at the TST during a
Fig. 12. Velocity difference between altered and unaltered situation. Case 1.
spring-neap cycle, for cases 1 and 2 respectively. Therefore, the
presence of a tidal farm involves signicant variations in the tidal
resource, since the tidal ow is highly affected slowing down up-
stream of the farms, which involves a reduction of the available
energy in the incoming ow.
4. Conclusions
The impacts caused by the operation of a tidal stream farm on
the marine environment are one of the biggest issues that must be
addressed if the tidal stream energy is to become a fully-edged
renewable. This work investigates the potential impacts on the
ow derived from the tidal energy extraction by modeling the
operation of two tidal stream farms placed in an estuary in NW
Spain, Ria de Ribadeo.
A numerical model of the ria hydrodynamics was implemented
and validated based on observations of tidal level and current ve-
locity. A substantial tidal stream resource was found, with some
areas (I, II and III) presenting power density values above 1400 and
1350 W m
2 at mid-ood and mid-ebb of a mean spring tide,
respectively. Then, the TSE (Tidal Stream Exploitability) index was
used to dene the best location for tidal stream energy exploitation.
It was found that areas II and III represent promising sites for the
operation of a tidal plant; therefore it was in these two areas where
the TST (tidal stream turbines) were modeled.
The TSTs were simulated by implementing a retarding force in
the local momentum equations of the model. In addition, the
high resolution of the model in the area of interest allows the
effect of each TST to be simulated individually. The effects
derived from the operation of the TSTs were investigated for two
 V. Ramos et al. / Energy 61 (2013) 541e554 551

Fig. 13. Velocity difference between altered and unaltered situation. Case 2.

Fig. 14. Changes in the power density distribution between altered and unaltered situation. Case 1.
552
Fig. 15. Changes in the power density distribution between altered and unaltered situation. Case 2.
levels of power extraction, high (case 1) and low (case 2), in
terms of the change in the tidal velocities, water levels and po-
wer density distribution with respect to the undisturbed ow
(i.e. without TSTs).
Overall, the effect on the water level can be considered negli-
gible. The maximum variations occur during the ebb, up to 0.025 m.
These variations are concentrated in the vicinity of the TSTs, with
increases and decreases on the water level upstream and down-
stream the TSTs, respectively. With regards to the ow velocities,
the presence of the farm alters the ow patterns reducing the ow
velocity upstream and downstream and increasing it beside the
farm. The maximum variations occur during the ood with differ-
ences up to 0.25 m s
1 (more than 10% of the unaltered ow
magnitude). Contrary to the case of water level, the disturbances on
the ow velocity are not restricted to the vicinity of the farms and
can be noticed as far as 1.7 km away from the farm. In addition,
Power density (W/m2)
1.250
750
250
16/03 18/03
Fig. 16. Available energy density in the incoming ow of the turbine.
V. Ramos et al. / Energy 61 (2013) 541e554
important asymmetry effects are also observed with larger impacts
at mid-ood, which may be caused by the ood dominance present
in the ria. Finally, these modications in the ow alter the power
density distribution and, therefore, the available energy density at
the farm, which was quantied during a spring-neap cycle for a
single TST, giving as a result a reduction of 21% and 12% for cases 1
and 2, respectively.
In summary, the installation of a tidal farm in an estuary may
cause important disturbances in the water levels, ow velocities
and the available tidal resource. Therefore these issues must be
studied in detail when planning the installation of a tidal farm.
Finally, the operation of a tidal stream power plant may produce
environmental effects which are out of the scope of this work, such
as modications to the sediment transport patterns or the residual
circulation, which will be dealt with as a continuation of this line of
research.
Available power density (Baseline case)
Available power density (Ct = 0.4)
Available power density (Ct = 0.8)
20/03 22/03 24/03 26/03 28/03 30/03
Time (Days)
 V. Ramos et al. / Energy 61 (2013) 541e554 553

Acknowledgments sby Y component of the shear stress at the bottom [N m
2]

yh kinematic horizontal eddy viscosity [m2 s
1]
This work is part of the project Assessment of Renewable Energy
Resources (DPI2009-14546-CO2-O2) supported by Spains Ministry
References
of Science and Innovation (Ministerio de Ciencia e Innovacin). Its
authors wish to thank P. Garca and M. Ouro for their contribution to [1] Lund H. Renewable energy strategies for sustainable development. Energy
the eld campaign, as well as the Spanish Oceanographic Institute 2007;32(6):912e9.
(Instituto Espaol de Oceanografa) for the salinity and temperature [2] Lund H, Mathiesen BV. Energy system analysis of 100% renewable energy sys-
tems: the case of Denmark in years 2030 and 2050. Energy 2009;34(5):524e31.
data. During this work V. Ramos has been supported by the I2C grant  
[3] Cosi c B, Kraja
ci
c G, Dui
c N. A 100% renewable energy system in the year 2050:
PRE/2011/236 (Plan Galego de Investigacin Innovacin e Crece- the case of Macedonia. Energy 2012;48(1):80e7.
mento 2011e2015) of the Xunta de Galicia. [4] Gmez A, Zubizarreta J, Dopazo C, Fueyo N. Spanish energy roadmap to 2020:
socioeconomic implications of renewable targets. Energy 2011;36(4):1973e85.
[5] Bahaj AS. Generating electricity from the oceans. Renew Sustain Energy Rev
Nomenclature 2011;15(7):3399e416.
[6] Voss A. Waves, currents, tides: problems and prospects. Energy 1979;4(5):
823e31.
[7] Published in the framework of the Coordinated Action on Ocean Energy EU
Roman project under FP6 Priority: 6.1.3.2.3; Renewable Energy Technologies with the
c salinity or temperature (transported substance) support of the European Commission directorate-General for Research under
CT thrust coefcient contract SES6-CT2004-502701Ocean energy conversion in europe e recent
advancements and prospects. CRES; 2006.
C2D Chzy coefcient [8] Iglesias G, Carballo R. Choosing the site for the rst wave farm in a region: a
d local water depth [m] case study in the Galician Southwest (Spain). Energy 2011;36(9):5525e31.
Dh horizontal eddy diffusivity [m2 s
1] [9] Carballo R, Iglesias G. Wave farm impact based on realistic waveeWEC
interaction. Energy 2013;51(0):216e29.
Fx x component of the ow retarding force per unit volume [10] Bryden IG, Macfarlane DM. The utilisation of short term energy storage with
[N] tidal current generation systems. Energy 2000;25(9):893e907.
Fy y component of the ow retarding force per unit volume [11] Bryden IG, Naik S, Fraenkel P, Bullen CR. Matching tidal current plants to local
ow conditions. Energy 1998;23(9):699e709.
[N] [12] Batten WMJ, Bahaj AS, Molland AF, Chaplin JR. The prediction of the hydrodynamic
f Coriolis parameter performance of marine current turbines. Renew Energy 2008;33(5):1085e96.
g gravitational acceleration [m s
2] [13] Ramos V, Iglesias G. Performance assessment of tidal stream turbines: a
parametric approach. Energy Convers Manage 2013;69(0):49e57.
h mid-tide water depth [m]
[14] Snchez M, Iglesias G, Carballo R, Fraguela JA. Power peaks against installed
h0 characteristic water depth [m] capacity in tidal stream energy. IET Renew Power Gen 2013;7(6):246e53.
h1 lower bound of the penalty range [m] [15] Bahaj AS, Myers L. Analytical estimates of the energy yield potential from the
h2 upper bound of the penalty range [m] Alderney Race (Channel Islands) using marine current energy converters.
Renew Energy 2004;29(12):1931e45.
Mx x component of the momentum generated by an external [16] O Rourke F, Boyle F, Reynolds A. Tidal energy update 2009. Appl Energy
force [Nm] 2010;87(2):398e409.
My y component of the momentum generated by an external [17] Garrido AJ, Garrido I, Amundarain M, Alberdi M, De la Sen M. Sliding-mode
control of wave power generation plants. IEEE Trans Ind Appl 2012;48(6):
force [Nm] 2372e81.
n Manning coefcient [18] Alberdi M, Amundarain M, Garrido AJ, Garrido I, Casquero O, De La Sen M.
Q intensity of mass sources per unit area [m2 s
1] Complementary control of oscillating water column-based wave energy
conversion plants to improve the instantaneous power output. IEEE Trans
R source term per unit area Energy Convers 2011;26(4):1021e32.
U vertically integrated eastward component of the ow [19] Medina J, Gonzalez-Escriva J, Garrido J, De Rouck J. Overtopping analysis using
velocity [m s
1] neural networks. World Scientic; 2003.
[20] Liu P. A computational hydrodynamics method for horizontal axis turbine e
jUj vertically integrated magnitude of the ow velocity
panel method modeling migration from propulsion to turbine energy. Energy
[m s
1] 2010;35(7):2843e51.
V vertically integrated northward component of the ow [21] Defne Z, Haas KA, Fritz HM. Numerical modeling of tidal currents and the
velocity [m s
1] effects of power extraction on estuarine hydrodynamics along the Georgia
coast, USA. Renew Energy 2011;36(12):3461e71.
Vf mid-ood depth-mean velocity [m s
1] [22] Plew DR, Stevens CL. Numerical modelling of the effect of turbines on currents in a
Ve mid-ebb depth-mean velocities [m s
1] tidal channel e Tory Channel, New Zealand. Renew Energy 2013;57(0):269e82.
V0 characteristic ow velocity [m s
1] [23] Ahmadian R, Falconer RA. Assessment of array shape of tidal stream turbines
on hydro-environmental impacts and power output. Renew Energy
Vc volume of control [m3] 2012;44(0):318e27.
[24] Ahmadian R, Falconer R, Bockelmann-Evans B. Far-eld modelling of the
hydro-environmental impact of tidal stream turbines. Renew Energy
2012;38(1):107e16.
Greek [25] Kadiri M, Ahmadian R, Bockelmann-Evans B, Rauen W, Falconer R. A review of
Dh maximum tidal range [m] the potential water quality impacts of tidal renewable energy systems. Renew
z water level [m] Sustain Energy Rev 2012;16(1):329e41.
[26] Ahmadian R, Falconer R, Lin B. Hydro-environmental modelling of proposed
ld rst order decay process severn barrage, UK. Proc Inst of Civ Eng Energy 2010;163(EN3):107e17.
x TSE penalty function [27] Neill SP, Jordan JR, Couch SJ. Impact of tidal energy converter (TEC) arrays on
ro water reference density [kg m
3] the dynamics of headland sand banks. Renew Energy 2012;37(1):387e97.
anomaly density [kg m
3]
[28] Neill SP, Litt EJ, Couch SJ, Davies AG. The impact of tidal stream turbines on
r0 large-scale sediment dynamics. Renew Energy 2009;34(12):2803e12.
r seawater density [kg m
3] [29] Shapiro GI. Effect of tidal stream power generation on the region-wide cir-
ra air density [kg m
3] culation in a shallow area. Ocean Sci 2011;7:165e74.
!
ss wind stress on the sea surface [N m
2] [30] Carballo R, Iglesias G, Castro A. Numerical model evaluation of tidal stream
energy resources in the Ra de Muros (NW Spain). Renew Energy 2009;34(6):
ssx X component of the wind stress on the sea surface 1517e24.
[N m
2] [31] Dias JM, Lopes JF. Implementation and assessment of hydrodynamic, salt and
ssy Y component of the wind stress on the sea surface heat transport models: the case of Ria de Aveiro Lagoon (Portugal). Environ
Model Softw 2006;21(1):1e15.
[N m
2] [32] Smith SD. Wind stress and heat ux over the ocean in gale force winds. J Phys
sbx X component of the shear stress at the bottom [N m
2] Oceanogr May 1980;10(5):709e26.
554 V. Ramos et al. / Energy 61 (2013) 541e554
[33] Yelland MJ, Moat BI, Taylor PK, Pascal RW, Hutchings J, Cornell VC. Wind
stress measurements from the open ocean corrected for airow distortion by
the ship. J Phys Oceanogr 1998;28(7):1511e26.
[34] Otero P, Ruiz-Villarreal M, Peliz A. Variability of river plumes off Northwest
Iberia in response to wind events. J Marine Syst 2008;72(1e4):238e55.
[35] Iglesias G, Snchez M, Carballo R, Fernndez H. The TSE index e a new tool for
selecting tidal stream sites in depth-limited regions. Renew Energy
2012;48(0):350e7.
[36] Carballo R, Iglesias G, Castro A. Residual circulation in the Ra de Muros (NW
Spain): a 3D numerical model study. J Marine Syst 2009;75(1e2):130.
[37] Blanton JO, Lin G, Elston SA. Tidal current asymmetry in shallow estuaries and
tidal creeks. Cont Shelf Res 2002;22(11e13):1731e43.
[38] Ocean ow energy. www.oceanowenergy.com [accessed 10.07.12].
[39] Batten WMJ, Bahaj AS, Molland AF, Chaplin JR. Experimentally validated nu-
merical method for the hydrodynamic design of horizontal axis tidal turbines.
Ocean Eng 2007;34(7):1013e20.
[40] Bahaj AS, Molland AF, Chaplin JR, Batten WMJ. Power and thrust measure-
ments of marine current turbines under various hydrodynamic ow condi-
tions in a cavitation tunnel and a towing tank. Renew Energy 2007;32(3):
407e26.