Rivers 2011
6th – 9th December 2011, Penang, Malaysia
Environmental Hydraulics: Integrated Water Management Solutions from
Cloud to Coast
ROGER A. FALCONER, Halcrow Professor of Water Management, Hydro-environmental Research Centre (HRC),
Cardiff School of Engineering, Cardiff University, The Parade, Cardiff, CF24 3AA, United Kingdom. Email:
falconerra@cardiff.ac.uk
KUNLE AKANDE, Associate Director and Leading Specialist, Halcrow Group Ltd., Burderop Park, Swindon, Wiltshire,
SN4 0QD, United Kingdom
BRIAN A. BOYE, Research Student, Hydro-environmental Research Centre, Cardiff School of Engineering, Cardiff
University, The Parade, Cardiff, CF24 3AA, United Kingdom
ABSTRACT
This paper reviews the traditional approach of linking models to cover integrated water management from the upper reaches of
catchments through river basins, into estuaries and then into the marine environment. It highlights some of the deficiencies in the
approaches currently being adopted in many of non-integrated studies, where artificial boundaries are included in the system and then
highlights the need for a more integrated conceptual approach. Two example studies are discussed, namely the non-compliance of
bathing waters along a UK coastal site from riverine inputs, with the inputs arising from agricultural and urban runoff, and the
proposed Severn barrage project where a large barrage is proposed across the Severn estuary to provide 5% of the UK’s electricity
needs. In this latter case study the barrage will have a major impact on the turbidity levels upstream, which will significantly impact on
the levels of light penetration through the water column and potentially faecal indicator organisms. An integrated approach shows that
the lower turbidity levels leads to improved water quality, through increased light penetration.
Keywords: Environmental hydraulics, computational hydraulics, integrated water management, water quality, turbidity
1
river or in the estuary etc. as depicted in Figure 1. It is
the modeller who creates the boundary and we need to
ensure that such boundaries are as smooth and as
transparent as possible. For example, it is imperative
that when a 1-D model is linked to 2-D model there
should not be a sudden change in the bed friction or the
diffusion/dispersion values included in the integrated
model, unless the sudden change can be justified. Thus
the challenge is to have a truly integrated water
management system for accurate and effective solutions
and this paper seeks to develop appropriate solution
methods for implementation of a Cloud to Coast C2C
solutions approach.
Introduction
In the current environment where nations increasingly
accept that climate change is happening and more
countries are experiencing the impact of climate change,
in the form of increased floods and more intense storms
or more droughts etc., there is a growing interest worldwide in adopting a more integrated approach to river
basin management, particularly when rivers cross state
boundaries and countries. For example, within Europe
there is the EU Water Framework Directive, wherein
member states are required to set up river basin
management plans.
Integrated water management is not new; many
research teams and companies have been modelling the
system from the cloud to the coast for many years, but
in many of these instances there have been different
teams or individuals modelling and monitoring
environmental hydraulics parameters through the
catchment, river, estuary and coastal basin. However,
all too often different parameters and formulations have
been used in these studies for the various processes.
When models are linked from catchments to rivers, and
particularly rivers to estuaries and coastal basins,
artificial boundaries are created in a natural system
where, in general, no boundary exists in nature. When
the raindrop falls from the cloud to the catchment, and
moves from the stream to the river, to the estuary and to
the sea, it does not know at any stage whether it is in the
Figure 1 Water Cycle (Royal Academy of Engineering 2010)
4
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
resilience and the associated economies of scope and
scale. The C2C solutions approach embraces the
findings of the US National Science Foundation report
on revolutionising science and engineering through
CyberInfrastructure” (see Atkins et.al., 2003), which
identifies opportunities in integrated approaches linked
to
new
developments
in
computing
and
communications technology. The purpose of this paper
is to introduce an innovation in integrated water
management for engineers and scientists with
responsibility for water ecosystem and catchment and
river basin management.
Observation of movements and patterns between
features in the water environment often suggest an
underlying conceptual relationship. Within a C2C
system, a component of a complex system is
indeterminate until that component can be associated
with other features of the system. The starting point for
the application of the C2C solutions philosophy is the
definition of the C2C boundary and its
conceptualization in space and time, within the context
of the C2C Decision Matrix shown in Figure 2.
Within the defined C2C boundary, prediction of
system behaviour, interaction between system
components is achieved through the use of combined
models facilitated by an OpenMI interface. Promotion
of the C2C solutions concept to-date, has involved the
development of a protocol for implementation, as a
system based approach. The protocol generally involves
Understanding integration requires an appreciation
of systems. Systems thinking helps to understand the
‘big picture’ whilst still retaining the complexity of
smaller systems. Considering a catchment (or other
elements of the environment) as a system means taking
account of multiple perspectives, stakes, purposes,
boundaries, interdependencies, controversies and
practices within it. Systems are best understood locally,
however there is a risk that localism can threaten
strategic priorities.
Water sector needs, driven by climate variability,
urbanisation and energy considerations, coupled with
the challenges of investment planning, require
innovative solutions. Halcrow and the HRC at Cardiff
University in a strategic collaboration have developed
such an innovative approach for integrated water
management. This result, promoted as Cloud to Coast
or C2C solutions, is described herein through two case
studies. In the first study the HRC work started with a
basic research of the C2C concept linking bathing water
compliance of an estuarine basin. However, the
requirements for practical application of the C2C
approach are now being developed further through an
on-going Halcrow supported Natural Environment
Research Council funded collaboration with industry
CASE research studentship.
Although the term “C2C solutions” is new, as
stated previously the concept of integrated water
management is not new. There are three current
Figure 2 Cloud to Coast Application Matrix
a modelling component and a stakeholder participatory
component, both dependent upon one another.
An introductory case study used for the proof of the
C2C concept, based on an investment planning project,
is one of the conceptual examples being used as part of
the on-going C2C proof of concept activities. The
project is aimed at developing a sustainable water
integrative approaches in the literature, each with
different implementation problems. The approaches are:
watershed or river basin management, adaptive
management, and water resources management in the
context of global environmental change. C2C solutions
are aimed at developing water infrastructure and
maintenance programmes that incorporate system
5
Rivers 2011
6th – 9th December 2011, Penang, Malaysia
management strategy for a riverine special area of
ecological conservation. In this application, whilst
acknowledging that the C2C system is a continuum, the
system was divided into two contrasting parts so that
assessment and investments could be defined for the
upper and lower parts of the system and for the system
as a whole (see Figure 3). For the case study a
comparison of the investment proposal with or without
C2C solutions is given in Table 1.
The river water highlighted in the above schematic
is used for potable and non-potable water supplies. Two
water treatment plants identified as WTW(1) and
WTW(2) provide potable water. Wastewater collection
and treatment is achieved through combined and
separate sewer systems. The central wastewater
treatment plant (WwTW) discharges treated flows to a
depleted reach of the river, which is eutrophic with
wildlife interests. Although the discharge of treated
effluent upstream of WTW(2) provided a boost to the
local water resource, it has also introduced a range of
water quality problems.
With the constraints of land space, the capacity of
the storm tanks at the WwTW is inadequate under
flooding conditions. Prior to the introduction of C2C
solutions, the area marked A in the case study
schematic was analysed. Without C2C solutions, the
investment strategy called for additional and expensive
treatment processes to be added to WTW (2), at a unit
cost of 64p/m3, to address the water quality problems.
With the introduction of C2C solutions, the boundary of
the system was re-defined and conceptualised.
Application of the C2C solutions approach resulted in
the following:
a) Development of control rules for upstream
groundwater flow augmentation as a replacement
water resource for WTW (2).
b) Relocation of the WwTW outfall downstream of
the raw water intake for WTW(2), with a large
reduction in unit treatment costs at WTW(2) from
64p/m3 to 30p/m3. Lomborg (2007) in the
publication titled “Solutions to the World’s Biggest
Problems” stated that putting prices on such
solutions make investment decisions better
informed. This is one of the strengths of the
innovative C2C solutions approach for the design
and operation of water systems.
The above example highlights that a C2C solutions
concept is feasible. It ensures that the wider
implications of interventions are understood for
investment planning. New knowledge on when and
where C2C solutions would be cost effective has been
developed to increase visibility of risks. There are
integrated benefits from the application of the C2C
solutions philosophy for people and wildlife.
Table 1 A comparison of aspects of Investment Proposal for
the Case Study with or without C2C solutions
Aspect
4
Approach to
sustainable water
management
Control
philosophy for
operations
Evaluation of
climate change
adaptation
strategies
Long term cost
5
Short term cost
1
2
3
With C2C
Solutions
Without C2C
Solutions
Holistic
Reductionist
Demand led
(flexible)
Yield led
(rigid)
Robust
Weak
Low
Low to High
depending on
system
complexity
High
Probably High
2
In applying the C2C approach to practical river and
coastal basin studies, involving environmental and ecohydraulics problems, details are first given herein of the
integrated modelling formulations used to simulate the
hydrodynamic and solute transport processes in the
river, estuarine and coastal system. The flow and solute
inputs from the catchments are not included herein, but
were provided from use of the field data and land-use
management models as outlined later in the paper.
The hydro-environmental models used by river and
coastal basin engineers and environmental managers to
predict the flow, water quality and sediment transport
Figure 3 Conceptual basis of case study and flow profile
6
Integrated Modelling Details
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
where Y = speed of earth’s rotation, I = earth’s latitude
and g = gravitational acceleration. The main effects of
the earth’s rotation, giving rise to the Coriolis
acceleration, are to set up transverse water surface
slopes across the river and to enhance the effect of
secondary currents and meandering.
For three-dimensional flow predictions, either the
full three-dimensional governing equations are solved,
which leads to a complex numerical formulation to
evaluate the pressure P, or, more usually, a hydrostatic
pressure distribution is assumed to occur in the vertical
(z) direction, and leading to an expression for P of the
following form:
processes in river and coastal systems are based on first
solving the governing hydrodynamic and solute
transport equations. For a Cartesian co-ordinate system,
with the main body of the flow in the x-direction, the
corresponding 3-D Reynolds averaged equations for
mass, momentum and solute in the flow direction can
be written in a general conservative form as (Falconer
and Lin, 1997):
wu wv ww
wx wy wz
wu wu 2 wuv wuw
wy
wz
wt
wx
,
1
(1)
0
X
,
3
1 wP
U wx
P( z )
wu cu c wu cvc wucwc (2)
wx
wy
wz
where u, v, w = velocity components in x, y, z co-
wP
wx
wM wuM wvM wwM w
w
w
u cM c vcM c wcM c
wt
wx
wy
wz
wx
wy
wz
,
2
(3)
4
ordinate directions respectively, t = time, X = body
force in x direction, U = fluid density, P = fluid
pressure, u cu c , u cv c , u cw c = Reynolds (or apparent)
stresses in the x direction on the x, y, z planes
respectively. Similar equations to the momentum
equation (2) can be written to evaluate the velocity
components v and w in the y and z directions
respectively. For the numbered terms in the momentum
equation (2), these terms refer to: local acceleration
(term 1), advective (or convective) acceleration(2),
body force (3), pressure gradient (4) and turbulent shear
stresses (5).
Likewise for the solute transport equation (3) M =
time averaged solute concentration, Ms = source or sink
solute input (e.g. an outfall), Md = solute decay or
growth term, and Mk = total kinetic transformation rate
for solute, with the individual terms in equation
generally referring to: local effects (term 1), transport
by advection (2), turbulence effects (3), and source (or
sink), decay (or growth) and kinetic transformation
effects (4).
In modelling estuarine flows in two and three
dimensions then the effects of the earth’s rotation needs
to be included giving, for the body force components:
Y
Z
2vY sin I ½
°
2uY sin I ¾
°
g
¿
wH
wP
Ug sin T a
wx
wx
ª wu wu º ½
u cu c Q t « » °
¬ wx wx ¼ °
ª wu wv º °°
u cvc Q t « » ¾
¬ wy wx ¼ °
ª wu ww º °
u cwc Q t «
»°
¬ wz wx ¼ °¿
(6)
(7)
where typically for a depth averaged flow the viscosity
Q t can be estimated from field data of the vertical
velocity profile or, assuming bed-generated turbulence
dominates over free shear layer turbulence, then a
logarithmic velocity profile can be assumed, together
with field and laboratory data giving (Fischer et al.,
1979):
Q t 0.167U * H
(8)
where U* = shear velocity. For most practical riverine
and coastal systems, even this value is low compared to
measured data recorded in rivers, with values for
Q t (U * H ) typically ranging from 0.42 to 1.61.
For the bed friction, this term is also generally
represented in the form of a quadratic friction law, as
given by:
(4)
7
Ug
where T = angle of channel slope. Likewise, a similar
representation can be written for the pressure gradient
in the y direction. Apart from modelling flows in long
rivers, the effects of the atmospheric pressure gradient
are generally small and are neglected in the riverine part
of the system.
The only unknown terms remaining in equation (2)
are the Reynolds stresses, which need to be related to
the 3-D velocity field before solving for the water levels
and the three-dimensional velocity components and are
generally represented in a diffusive manner, giving:
3
Ms Md Mk
X
(5)
where H = total depth of flow, z = elevation above the
bed and Pa = atmospheric pressure. The corresponding
derivative of equation (5), for inclusion in equation (2),
gives:
5
1
Ug ( H z ) Pa
4
2
Rivers 2011
6th – 9th December 2011, Penang, Malaysia
(ii) Turbidity – important in governing the flow
conditions and biological processes for many water
quality constituents.
(iii) Temperature – important in governing the flow
conditions and biological processes for many water
quality constituents.
(iv) Other physical indicators – such as colour,
conductivity and radio activity.
Vs
(9)
C 2H
where Vs = depth average fluid speed, qx = discharge
per unit width in the x-direction, and C = de Chezy
roughness coefficient, which can be expressed in terms
of the Manning roughness formulation as:
W xb
C
Ugq x
H 1/ 6
n
(10)
2. Chemical indicators of water quality:
(i) Dissolved oxygen – which is important for water
quality, since it is essential for supporting most
forms of aquatic life.
(ii) Biochemical oxygen demand – which is a measure
of the consumption of bacteria during aerobic
degredation of organic matter.
(iii) Nitrogen – which is an essential nutrient for
biological growth and a major constituent of
domestic effluents. Nitrogen may be present in a
variety of chemical forms, including: organic,
ammoniacal, nitrite and nitrate nitrogen.
(iv) Phosphorous – which is also an essential nutrient
for biological growth and which appears
exclusively as phosphate in aquatic environments.
Phosphate may be present in several forms,
including: orthophosphate, condensed phosphates
(i.e. pyro-, meta- and poly-phosphates) and
organically based phosphates.
(v) Chlorides – commonly occurring in the form of
salinity in estuarine reaches and an indication of
sewage pollution in rivers.
(vi) Metals – particularly those which are easily
dissolved in water and are toxic, e.g. arsenic,
cadmium, chromium, lead and mercury. These
metals occur mainly as a result of industrial
discharges, mine water tailings and agriculture.
Although these contaminants are usually only
present in low quantities, they can accumulate in
the food chain and cause toxicity problems.
where H = total depth of flow in a wide estuary or open
coastal basin, or the hydraulic radius in a relatively
narrow river and where side-wall effects are crucial and
n = Manning roughness coefficient, with typical values
of n being in the range from 0.012 for smooth lined
channelised rivers to 0.04 or more for meandering
rivers with vegetation etc. Although this approach is
appropriate for most rivers, the Manning representation
assumes that the flow is rough turbulent flow and that
the local headloss is dependent only on the size and
characteristics of the bed roughness, i.e. form drag
dominates totally. However, for low velocity flows on
shallow floodplains and wetlands, Reynolds number
effects may be significant, reflecting the increased
influence of skin friction. This complex hydrodynamic
phenomenon can be represented using the more
comprehensive friction formulation given by
Colebrook-White (Henderson, 1966).
For the solute transport equation The crossproduced terms u cMc etc. represent the mass flux of
solute due to the turbulent fluctuations and, by analogy
with Fick’s law of diffusion, it can be assumed that this
flux is proportional to the mean concentration gradient
and is in the direction of decreasing concentration.
Hence, the terms can be written as:
wM ½
°
wx °
wM °
vcM c Dty
¾
wy °
wM °
wcM c Dtz
°
wz ¿
ucM c Dtx
3. Biological indicators of water quality:
(i) Pathogens – including the faecal coliform group,
with the principal bacteria being Escherichia Coli.
(ii) Other biological organisms – such as algae.
(11)
In modelling these various water quality indicator
parameters in river and coastal basin systems, each
process includes a range of different source terms and
kinetic interactions. Details of these interactions are
given in the US Environmental Protection Agency
manual Bowie et.al. (1985).
where Dtx, Dty, Dtz = turbulent diffusion coefficients in
x, y, z directions. For riverine flows it is common to
assume isotropic turbulence and to approximate the
horizontal diffusion terms to the depth mean
coefficients as given by Fischer (1973), whereby in the
absence of field data, these terms are often equated to:
Dtx
Dty
0.15U * H
3
(12)
3.1 General
In modelling water quality processes in rivers and
coastal basins a range of indicator organisms are often
modelled, including (Falconer and Chen, 1996):
1. Physical indicators of water quality:
(i) Suspended solids – including inorganics, e.g. sand,
and organics, e.g. sewage solids.
The differential flow and water quality equations
outlined above can be solved using a wide range of
commercial or research orientated hydro-environmental
software tools. The first author has been involved in
8
Integrated Model Applications
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
the development of 3-D (TRIVAST), 2-D (DIVAST)
and 1-D (FASTER) flow, water quality and sediment
transport research based models and the details are
given herein of the application of these models to flow
in compound channels and, in particular, using a C2C
approach to two recent studies in the UK, namely faecal
indicator level predictions in the Ribble Estuary and
sediment and bacteria levels in the Severn Estuary as a
result of a barrage across the estuary to generate
renewable energy.
occasionally failed to comply with the EU mandatory
water quality standards of the Bathing Framework
Directive.
A hydro-environmental modelling study was
undertaken to establish the water quality of the EU
designated bathing waters located at the mouth of the
Ribble estuary. In order to reduce the possible
inaccuracies caused by setting up the boundary
conditions required by the numerical models, the
upstream boundaries were set up at the tidal limits of
the rivers Ribble, Darwen and Douglas (see Figure 4)
and the downstream boundary was located around the
25 m depth contour in the Irish Sea. The length of the
seaward boundary was 41.2 km, with the width of the
river boundaries being generally less than 10 m. Such a
great variation in the modelling dimensions made it
impractical to use either a 1-D or a 2-D model alone. In
applying either of these two models individually or
linking the models statically would have introduced
inaccuracies in the model prediction results and would
also have required a considerable amount of effort in
exchanging the data. Thus, in this study a 1-D and 2-D
model (namely FASTER and DIVAST) were linked
dynamically to create a single model, in which the
numerical simulations of the hydrodynamic variables
and bacterial indicators were undertaken simultaneously
across the entire modelling region. More details
regarding this linked model may be found in
Kashefipour et al (2002).
An extensive programme of data collection was
also undertaken by the U.K. Environment Agency to
provide data for model calibration and verification.
Hydrodynamic and water quality data were collected at
6 sites during the winter period of 1998 and the summer
period of 1999, to include a combination of different
weather and tide conditions. During each survey
measurements were taken at all discharge locations and
upstream river boundaries (i.e. tidal limits) for two
consecutive days. Comprehensive data sets were
collected, including: water depths, current speed and
directions, salinity levels and concentrations of
suspended solids, faecal and total coliforms, and faecal
streptococci. In total there were 34 input sources
identified that contributed to the pollution loads of the
estuary. These included: direct discharges of treated
wastewater from treatment plants, inputs from the
upstream boundary of the 3 major rivers, and inputs
from several smaller rivers and combined sewer
overflows. Four calibration points were chosen along
the main channel of the Ribble, with these sites being
referred to as: 11milepost, 7milepost, 3milepost and
Bullnose (see Figure 4). Measurements at these
locations were taken only for the second day of a
survey.
The 2-D area was represented horizontally using a
mesh of 618 u 454 uniform grid squares, each with a
length of 66.7m. A total of 1075 cross-sections was
designed for the 1-D area with a minimum and
maximum distances between two consecutive crosssections being 10m and 50m respectively.
3.2 Bathing Water Quality in the Ribble Basin
The Ribble estuary is located along the north-west coast
of England. At the mouth of the estuary there are two
well known seaside resorts, namely Lytham St Annes
and Southport, with both being designated EU
(European Union) bathing waters. The Fylde Coast,
which is bounded between Fleetwood in the north and
the Ribble estuary in the south, includes one of the most
famous beaches in England for tourism, namely
Blackpool, with an average of more than 17 million
visitors per annum. The area has four main centres of
population, namely Blackpool and Lytham St Annes to
the north of the Ribble Estuary, Southport to the south
of the estuary, and the town of Preston, which is inland
and straddles the river Ribble at the tidal limit (see
Figure 4).
BlackPool
433000
431000
429000
Fylde Coast
St Anne’s
North
Preston
Ribble Boundary
Bullnose
Lytham
,
St Annes
St Anne’s
Pier
427000
7milepost
11milepost
Blue Bridge
3milepost
Penwortham
Darwen Boundary
Upstream 2-D Boundary
425000
Downstream 1-D Boundary
423000
Measuring Water Elevation
Tarleton Lock
421000
Tide Survey Point
Measuring Discharge
Douglas River
419000
Southport
Southport
Upstream 1-D Boundary
Bathing Water Point
417000
415000
413000
Wanes Blades Bridge
Douglas Boundary
411000
326000
330000
334000
338000
342000
346000
350000
354000
358000
362000
Figure 4 Fylde Coast, showing the Ribble Estuary and its
Tributaries
In order to enhance the bathing water quality along
the Fylde coast, a major civil engineering investment
programme was undertaken in recent years to reduce
the bacterial input to the estuary. About £600 million
was invested over the past 20 years in new sewerage
works and treatment plants along the Fylde Coast and
Ribble Estuary. Examples include:- upgrading the
wastewater treatment works at Clifton Marsh from
primary treatment to include UV disinfection; reducing
storm water discharges from the wastewater network by
constructing 260,000 m3 of additional storage.
Although the reduction in input bacterial loads has
resulted in a marked decrease in the concentration of
bacterial indicators in the coastal receiving waters,
occasional elevated faecal coliform counts were still
measured. As a result, the bathing waters still
9
Rivers 2011
6th – 9th December 2011, Penang, Malaysia
June 1999 were compared in Figure 6, using the
original model. Due to the large variation in the values
of faecal coliform concentrations, the results are plotted
using a logarithmic scale. Relatively good agreement
between both sets of data were obtained from this
figure, with an average error of 35000cfu/100ml, i.e.
30.5%.
Figure 7 shows an example of application of the
original linked model application to the Ribble estuary,
comparing the predicted and measured faecal coliform
concentrations at 3milepost for the 19th May 1999
survey, which was carried out for a dry event and a
spring tidal range. As can be seen from this figure both
sets of data agreed well, with an average error of
477cfu/100ml, i.e. 27.6%. For this dry event the T90
Measured discharges at the upstream tidal limits of
the rivers Ribble, Darwen and Douglas were used as the
upstream open boundary conditions. The water levels
for the seaward boundary conditions were acquired
from the Proudman Oceanographic Laboratory (POL),
predicted using their Irish Sea model. The main
parameter used for calibration of the hydrodynamic
model was the bed roughness. By changing the values
of the bed roughness and comparing the model
predictions with measured data, the model was
calibrated by choosing the best fit between the predicted
results and measured data. Different roughness values
were used along different reaches of the model domain
to reflect the local conditions. For the open coastal
waters it was found that the most appropriate value for
the Nikuradse equivalent sand grain roughness was 20
mm. For the 1-D model region the Manning roughness
coefficient was used, with the optimum values ranging
from 0.021 for the lower part of the river to 0.028 for
the upper part of the river.
M ode l
7mile p o st
be 20.8hrs. All of these values were for a water
temperature of 20qC.
Wate r Ele v atio n
M e asu r e d
4
2
0
(a)
1.0E+07
Concentration (cfu/100ml)
Elevation (m )
6
values for the 2-D area were found to be 14.7 and
20.2hrs for day- and night-time periods respectively.
The corresponding values for the 1-D area were
obtained and found to be 21.4 and 26.9hrs respectively.
For this event an average T90 value was considered to
-2
3
Cu r r e n t Sp e e d
Speed (m/s)
2.5
2
1.5
1
Model
Measured
Faecal Coliform at 7milepost
03/06/1999
1.0E+06
1.0E+05
1.0E+04
1.0E+03
1.0E+02
1.0E+01
0.5
(b)
0
1.0E+00
30
32
34
36
38
40
42
44
Simulation time (hr)
46
48
50
-0.5
Figure 6 Comparison of predicted and measured faecal
coliform concentrations at 7milepost, on 3 June 1999
C urrent D irection
400
300
1.0E+06
200
100
Concentration (cfu/100ml)
Direction (deg)
500
(c)
0
30
32
34
36
38
40
42
44
46
48
50
S im u la tion tim e (hr)
Figure 5 Comparison of predicted and measured: (a) water
levels, (b) current speeds and (c) current directions at
7milepost, on 3 June 1999.
Measured
1.0E+04
1.0E+03
1.0E+02
1.0E+01
1.0E+00
24
26
28
30
32
34
36
38
40
42
44
46
48
50
52
Simulation time (hr)
Figure 7 Comparison of predicted and measured faecal
coliform concentrations at 3 milepost, on 19 May 1999
As an example of the simulations for the original
study are given herein where the results obtained for the
survey on the 3rd June 1999 are discussed (Kashefipour
et.al. 2002). This survey was carried out during a wet
period with a mean tidal range. Figures 5a to 5c show
comparisons at 7milepost between the predicted water
elevations and current speeds and directions
respectively. Figure 5a shows that good agreement was
achieved between the predicted and measured water
levels at this site, with an average error of 0.11m, i.e.
2.1% error, when compared with the measured tidal
range. Figures 5b and 5c illustrate relatively good
agreement between the predicted current speeds and
directions with the corresponding measured data, with
an average error of 0.13m/s (i.e. 8.8% error relative to
the maximum measured speed) and 8.1q respectively.
Predicted and measured faecal coliform
concentrations at 7milepost for the survey on the 3rd
In reviewing the original study, directed by the first
author (Kashefipour et.al., 2002), in the context of a
C2C solutions approach, the authors found a number of
anomalies with the original study and some key lessons
learnt for the future, particularly in the context of the
principle that the raindrop – possible with attached
pollution (either in solution or in particulate form) does
not know whether it is in a catchment stream, river,
estuary or coastal basin. Taking account of these
anomalies the following observations were recently
made with regard to the original studies:
(i) different values of the decay rate were used in the
1-D and 2-D parts of the linked model (see Table 2),
with the decay rate being higher in the 2-D part of
the 2-D part of the linked model in comparison
with the 1-D part. The length of day-time was
assumed to be 12hr from sunset to sunrise;
10
Model
Faecal Coliform at 3milepost
19/05/1999
1.0E+05
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
(ii)
(iii)
model was nearly two orders of magnitude greater than
that in the 1-D model at the same site. A similar
comparison again showed that the peak concentration
was not reached quite so closely as for the old model
but that the lower values were much more closely
predicted when an integrated C2C approach was
adopted and a unified value of 100 was used for the
dispersion constant J in the dispersion equation.
different values of the dispersion coefficient were
used in the 1-D and 2-D parts of the linked model
(see Table 3);
different values and formulations were used for the
roughness coefficient in the 1-D (Ks=20mm) and 2D (n=0.021) linked models.
Table 2 Differences in Decay Rate for 1-D and 2-D Model
Domains
Faecal Coliform Decay Rates
1-D
2-D
Day time Decay Rate (per day)
0.85
1.0
Equivalent Day time t90 value(hr)
65.0
55.3
Night Time Decay Rate (per day)
0.5
0.68
110.5
81.3
Equivalent Night Time t90 value (hr)
Table 3 Differences in Dispersion cCoefficients for 1-D and
2-D Model Domains
Location
Original Model
Milepost 7 (m2/s)
1.44 m2/s
Milepost 11 (m2/s)
1.63 m2/s
Node 1 (m2/s)
256 m2/s
Node 51 (m2/s)
240 m2/s
Figure 8 Comparison of Faecal Coliform Levels with Time
for the New and Old Models after Changes in Decay Rates
In adopting a C2C solutions approach to this study
site, the integrated catchment to coast model was re-run
using the same formulations for the momentum
correction coefficient, eddy diffusivity, Manning
roughness coefficient, dispersion and diffusion
coefficients and the kinetic decay rates (specified in the
form of T90 values) along the whole river and coastal
basin and with the various parameters only being
changed where there were sound technical reasons for
doing so. This differed from the previous study where
the value of the parameter was very much based on
whether the ‘water package’ was in the 1-D, 2-D or 3-D
model domains. In the new model, the T90 value was set
as 20hr for day time and 100hr for night time in both
the 2-D and 1-D models. The equivalent daytime rate
(2.76 per day) was therefore higher than the values in
the old model (Kashefipour et.al., 2002). The equivalent
night time decay rate (0.55 per day) is lower than that in
the 2-D part of the previous model. A comparison of the
resulting faecal coliform predictions is shown in Figure
8, both for the old and new models, with these values
for the day and night time decay rates being found to
give closest agreement with the field data.
The results of this comparison show that whilst the
old model, which was calibrated to a maximum in each
reach, gave best agreement with the peak faecal
coliform results, there is no doubt that the lower levels
were much more accurately predicted using the more
holistic systems based C2C solutions approach and
hence provides a more accurate tool for decision
planning for the future.
Likewise similar results were obtained for the
dispersion and diffusion coefficients, where in the old
modelling study the dispersion coefficient in the 2-D
Finally the resulting predictions for the faecal
coliform concentration distributions are shown in figure
9 for the estuary using the old and new models, where a
comparison of the C2C approach to modelling
highlights the different concentrations predicted and the
corresponding implications for investment strategies
appropriate for the various treatment options and
solutions in the river basin upstream.
3.3 Severn Tidal Power Water Quality Studies
In recent years there has been growing international
public concern about climate change, global warming,
reducing the carbon footprint, increasing oil and gas
prices and the rapid depletion of fossil fuel reserves. In
addition to these challenges there is a growing global
demand for energy. The UK is also committed to the
EU renewable energy targets, with the UK being
expected to produce 15% of its total energy from
renewable resources by 2020, which corresponds to
about 35% of the UK’s electricity demand. At present
only about 5% of the UK’s electricity comes from
renewable resources. More recently, the UK Committee
on Climate Change recommended in October 2008 that
the UK should aim to reduce greenhouse gas emissions
by at least 80% by 2050 compared to 1990 levels. In
the same timeframe, global emissions will need to fall
by at least 50%. These issues, among others, have led to
enthusiasm in the UK to look to increasing its
generation of renewable energy from tidal sources, with
tides having the advantage over wind and waves in that
they are predictable.
11
Rivers 2011
6th – 9th December 2011, Penang, Malaysia
because of its unique characteristics, the estuary is
protected under a number of European and international
legislative directives.
The most attractive means of harnessing the tidal
power from the Severn Estuary is therefore to build a
barrage across the estuary. The UK Government has
investigated this option and the most favourable
alternative is the so called ‘Severn Barrage’. This
barrage would span from Lavernock Point just south
west of Cardiff, to Brean Down just south of Westonsuper-Mare (see Figure 10).
(a) Old model
Figure 10 Location of Severn Barrage (picture courtesy
of Severn Tidal Power Group)
(b) New model
Figure 9 Simulations of Faecal Coliform in Ribble Estuary
using (a) Old and (b) New (C2C) Approach to linked
Modelling
The traditional proposal for operating the barrage
has been for the structure to be operated under ebb tide
generation conditions only, wherein the sluice gates
would be opened on the incoming tide and the turbines
closed. The impounded area, of approximately 500km2,
would be allowed to flood to high tide elevation, at
which point during the tidal cycle the turbines and
sluice gates would be closed. The impounded water
would then be held for just over 2.5h to create sufficient
head for generation, with the turbines then being
opened and generating power over the next 5h.
The main effects of the barrage would be to reduce
the peak upstream tidal range from 14m to 7m. This
would result in a significant loss of the intertidal habit
areas, estimated to be a loss of 14,000h. The currents
would reduce significantly upstream of the barrage,
leading to reduced levels of suspended sediments and
turbidity. There would be an increase in light
penetration through the water column and an increase
and change in the primary productivity of the benthic
flora and fauna.
The current and alternative proposals for the
Severn Barrage have been studied extensively using
hydro-environmental models, developed within the
HRC at Cardiff University, to predict the hydrodynamic,
sediment transport and water quality processes in the
In considering the development of tidal renewable
energy in the UK then the Severn Estuary and Bristol
Channel is an ideal site. The estuary, including the
Bristol Channel, is located some 240km (or 150miles)
west of London and spans the South Wales coast from
Milford Haven to Gloucester and the northern reach of
the South West coast of England from Hartland Point,
also to Gloucester. There is a weir at Gloucester and for
most of the spring-neap cycle this defines the tidal limit
of the estuary. The estuary, including the Bristol
Channel, is approximately 200km long and has the
second highest rise and fall of tide in the world, with
typical spring and neap tidal ranges peaking at over
14m and 7m respectively at Avonmouth, which is just
further upstream of the Port of Bristol. The spring tidal
currents in the estuary are well in excess of 2m/s and
there is a large area of intertidal mudflats. Suspended
sediment levels are large, with the spring and neap
loads being estimated at 30 Mt and 4Mt respectively.
Light penetration through the water column is very
limited and dissolved oxygen saturation levels are
reduced. The estuarine regime is harsh and there is
limited aquatic life in the water column. However,
12
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
wastewater treatment works along the estuary and river
(Environment Agency, 2007). Typical comparisons
between measured and predicted water elevations and
current speeds are given in Yang et.al. (2008), where
good agreement has been obtained both sets of results.
In particular, the results have been improved further
following a C2C approach in terms of enhanced
integration of the 1-D/2-D linking of the models.
The model was then extended to refine the prediction
of bacteria in coastal waters as part of an extensive
research programme of field monitoring and numerical
modelling to predict faecal contamination in the Bristol
Channel and Severn Estuary where, despite extensive
levels of wastewater treatment (including UV
disinfection), high levels of enteric bacterial
contamination were still found in some of the bathing
waters along the estuary. In this study a new enteric
bacterial–sediment transport interaction conceptual
model was developed and measured data used to
estimate the adsorption and desorption rates between
the bacteria and the sediments. The data acquired
through the field data monitoring studies showed that
the bacteria can remain trapped within the sediments for
considerably long periods, only to be released and
transported potentially to bathing waters during storm
or spring tide conditions. When the bacteria were
adsorbed onto the sediments and deposited on the bed
then the decay rate was found to be very low and the
bacteria were regarded as almost conservative.
Relatively good agreement was obtained between the
measured and predicted levels of faecal coliform along
the estuary, with the results highlighting considerable
differences between the concentration levels measured
and predicted when the interactions between the
bacteria and the sediments were included in the model,
see Gao et.al. (2011). Typical model predictions of
bacterial levels in the estuary both with and without the
barrage are shown in Figure 11.
estuary, both without and with a barrage. The models
used were similar to those used for the Ribble study and
included both a 1-D/2-D linked regular grid model and
an unstructured grid model, with the models being
verified with field data and more recently tested against
a small scale 1:25,000 physical model of the estuary.
As part of the studies to refine the bacteriological model
predictions for the estuary, experiments were
undertaken from field data of faecal coliform levels to
determine the T90 decay rate as a function of the
suspended sediment levels and with another set of
experiments being undertaken to establish the
adsorption rates for the coliform levels on the sediments.
The 2-D DIVAST and 1-D FASTER models were
modified and set up to investigate the far-field impacts
of the proposed barrage in the Bristol Channel and the
Severn Estuary. Modelling studies were undertaken
with the boundaries set from Milford Haven and
Hartland Head along the western boundary and to
Gloucester at the eastern extremity (Lin and Falconer,
1995). At Gloucester a flow rate varying between
60m3/s to 106m3/s was used for normal operating
conditions. The 1-D model consisted of four reaches
which were modelled using 351 cross-sections with an
average distance of 240m between two consecutive
cross-sections. The inflows for the four major rivers in
the modelled area, namely the rivers Wye, Usk, Frome
and Little Avon, were treated as lateral inflows. The 1D boundary was specified as a water level boundary
and located close to the Severn Bridge, where water
level data were used to drive the model, obtained from
the 2-D model. To ensure stability in data transfer
between the 1-D and 2-D models, and to ensure
momentum conservation, the 1-D and 2-D models
overlapped for a reach of over 4.5km between the old
and new Severn Bridge crossings (Yang et al, 2008). In
addition to flow and water elevation data being included
at the open boundaries, discharges and enterococci
inputs were included for all the main rivers and
(a)
(b)
Figure 11 Predicted Coliform Levels in the Severn Estuary and Bristol Channel (a) Without and (b) With a Severn Barrage
13
Rivers 2011
6th – 9th December 2011, Penang, Malaysia
agreement with the field data was obtained across a
wider range of spatial and temporal measured data
values, although the peak values were slightly under
predicted. These model results could then be better used
to give more informed and improved management
decisions for more effective investment strategies for
improved water quality in the river and coastal basin for
the future.
Finally, the authors are extending their C2C
numerical modelling studies of the hydrodynamic,
morphological, water quality and contaminant transport
processes to the Severn estuary to predict the impacts of
a proposed barrage across the estuary, for tidal energy
provision, on the hydro-environmental characteristics of
the basin. There is no doubt that such a barrage will
have an adverse impact on a range of environmental
and ecological aspects, and particularly upstream of the
barrage, the model studies undertaken have shown that
there will also be some positive hydro-environmental
benefits associated with a barrage. In particular, there
will be a general reduction in the magnitude of the tidal
currents, leading to a significant reduction in the
suspended sediment concentrations and turbidity levels.
There will be a corresponding increase in the light
penetration levels and reduced bacterial levels in the
water column, due to the reduced input from bacteria
adsorbed onto the sediments and the increased decay
rate as a result of increased light penetration.
The model predictions also show that there will be
reduced peak water elevations, both upstream and
downstream of the barrage, with the level upstream
reducing by up to nearly 2 m at Gloucester. Using the
C2C Solutions approach Halcrow and the HRC are
currently investigating a range of options for this
project, including two-way generation and low head
turbine alternatives.
In summarising the results from this study,
including a C2C solutions approach, the key findings
were as follows: (i) the peak high tide levels upstream
of the barrage were approximately 1 m lower with a
barrage and the high tide levels reduced typically of the
order of 50 cm for a region up to about 15 km seawards
of the barrage; (ii) the peak tidal currents exceeding 2
m/s were mainly in the main channel from a transect
just seawards of Minehead to Aberthaw, to a transect
upstream of the old Severn Bridge; (iii) the region
upstream of the barrage is a region of high turbidity and
high suspended sediment concentrations in excess of
1200mg/l, comparisons showed that if a barrage were
built then the suspended sediment concentrations would
be reduced significantly to about 200 mg/l, leading to
increased light penetration in the water column and
more rapid decay of enteric bacteria; (iv) the novel
bacteria model showed that there was a strong
interaction between adsorption and desorption of enteric
bacteria with sediments, and particularly those
sediments known to be high in organic content, with the
resulting predictions showing that a barrage would lead
to reduced bacterial levels in the water column in the
region; and (v) the predictions of the peak water
elevations have shown that the risk of flooding would
be reduced both upstream and downstream if a barrage
were built.
4
Conclusions
The authors have introduced the concept of a systems
based solutions approach to integrated water
management in the form of Cloud to Coast C2C
Solutions. The approach involves considering the water
system in its entirety, rather than the current common
practise of de-compartmentalising the basin into a
catchment, river, estuary and coastal basin and treating
each separately. Halcrow are currently applying a C2C
approach to water projects and this approach is leading
to improved and/or more efficient design solutions. In
the context of modelling for C2C solutions, it is
necessary to consider the raindrop as a water package,
possibly with pollution, moving through the system (i.e.
the catchment, river, estuary and coast), with the
roughness, turbulence, diffusion, dispersion, decay etc,
only changing if physically and/or bio-chemically
appropriate and not based on the artificial boundaries
imposed through modelling.
In applying a C2C Solutions approach to hydroenvironmental modelling the authors have repeated a
previous study (previously undertaken by the first
author) using this concept, namely modelling the
hydrodynamic and faecal indicator organism processes
in the Ribble river and coastal basin in northern
England. When applying the C2C approach the authors
found inherent big changes in the roughness, dispersion
and, to a lesser extent, the decay rates when transferring
from the 1-D to the 2-D model in the original study.
When removing these changes and re-calibrating the
integrated linked models for unified parameters, better
References
1.
2.
3.
4.
5.
14
The Royal Academy of Engineering (2010). Global
Water Security: An Engineering Perspective.
Published on line at www.raeng.org.uk/.
Atkins, D.E., Droegemeir, K.K., Fieldman, S.I.,
Garcia-Molina, H., Klein, M.L., Messerschmitt,
D.G., Messina, P., Ostriker, J. P. and Wright, M.H.
(2003). Revolutionising Science and Engineering
through CyberInfrastructure: Report of the
National Science
Foundation
Blue-Ribbon
Advisory Panel on CyberInfrastructure.
Lomborg, B. (2007). Solutions for the World’s
Biggest Problems. Costs and Benefits. ISBN-10 0511-36423-7. Cambridge University Press.
Falconer, R. A. and Lin, B. (1997). Threedimensional modelling of water quality in the
Humber Estuary. Water Research. 31(5), 10921102.
Fischer, H.B., List, E.J., Koh, R.C.J., Imberger, J.
and Brooks, N.H. (1979). Mixing in inland coastal
waters. Academic Press Inc., San Diego, 483 pp.
3rd International Conference on Managing Rivers in the 21st Century:
Sustainable Solutions for Global Crisis of Flooding, Pollution and Water Scarcity
6.
7.
8.
9.
10.
11. Lin, B. and Falconer, R. A., (1995). Modelling
sediment fluxes in estuarine waters using a
curvilinear co-ordinate grid system, Estuarine,
Coastal and Shelf Science, 41(4), 413-428.
12. Yang, L., Lin, B. and Falconer, R. A. (2008).
Modelling enteric bacteria levels in coastal and
estuarine waters, Proceedings of Institution of Civil
Engineers, Engineering and Computational
Mechanics, 161(4), 179-186.
13. Environment Agency. (2007). Fate and transport
of particles in estuaries, Vol. II: Estimation of
enterococci inputs to the Severn Estuary from point
and
diffuse
sources,
Science
Report
SC000002/SR2, Environment Agency, 55pp.
14. Gao, G., Falconer, R. A. and Lin B. (2011)
.Numerical modelling of sediment-bacteria
interaction processes in surface waters. Water
Research. 45(5), 1951-1960.
Henderson, F.M. (1966). Open Channel Flow.
Collier-Macmillan Publishers, London, 522 pp.
Fischer, H.B. (1973) Longitudinal dispersion and
turbulent mixing in open channel flow. Annual
Review of Fluid Mechanics, Vol. 5, 59-78.
Falconer, R.A. and Chen, Y. (1996). Modelling
sediment transport and water quality processes on
tidal floodplains. In: Floodplain Processes, M.G.
Anderson, D.E. Walling and P.D.Bates (eds), John
Wiley and Sons Ltd., Chichester, Chapter 11, 361398.
Bowie, G.L. et al (1985). Rates, Constants and
Kinetics Formulations in Water Quality Modelling,
Environmental Research Laboratory, US EPA,
Athens, GA, Report No. EPA/600/3-85/040, 475
pp.
Kashefipour, S.M., Lin, B., Harris, E.L. and
Falconer, R.A. (2002). Hydro-environmental
modelling for bathing water compliance of an
estuarine basin., Water Research, IWA, Vol. 36,
pp.1854-1868.
15
View publication stats