rance v severn

Delivered by ICEVirtualLibrary.com to:

IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

Proceedings of the Institution ofCivil EngineersMaritime Engineering 162March 2009 Issue MA1Pages 11–26doi: 10.1680/maen.2009.162 .1.11

Paper 800007Received 8/05/2008Accepted 17/07/2008

Keywords:dams, barrages & reservoirs/environment/renewable energy

Robert KirbyManaging Director,Ravensrodd Consultants Ltd,Taunton, Somerset, UK

Christian RetiereDirector, Laboratoire Maritime,Museum National d’HistoireNaturelle, Dinard, France

Comparing environmental effects of Rance and Severn barrages

R. Kirby PhD, CGeol., FGS and C. Retiere Dr. es Sc

This paper examines the similarities and contrasts

between environmental and water quality changes caused

by the 1966 tidal power barrage at La Rance in North

Brittany, France and those expected to result from

construction of a tidal power barrage in the Severn

estuary in the UK. Over the 40-year period since the

opening of La Rance, a great deal of knowledge has been

accumulated concerning the operation of such schemes

and the way water quality and ecosystems have altered.

Knowledge gained from experiences at Rance is conse-

quently of great benefit in predicting the effects of the

proposed Severn barrage. The nature of the changes

anticipated in the Severn estuary is thus unambiguous,

but whether this would make the estuary better or worse

is a matter of perception. As at Rance, it would be

different.

1. INTRODUCTION

Tidal power schemes relying on impoundment are confined to

hypertidal systems [. 6 m mean tidal range (MTR)] which are

rare around the world. A tidal power barrage was built between

1961 and 1966 in the Rance estuary near St Malo in Brittany,

France. It was formally opened on 26 November 1966 and the

final turbine installed on 4 December 1967. On several occasions

during the past 30 years, evaluations of potential schemes for

tidal power barrages in the much larger Severn estuary in the

south-west of the UK have been carried out. A new phase of

feasibility is envisaged, now focused on a single basin scheme

located on the so-called ‘Cardiff–Weston’ line (see Figure 1).

Table 1 compares the scale of the two schemes.

There are contrasts in the relative sizes and manner of

construction – actual and envisaged – and in the location of

barrages within as well as in the types of these estuaries.

Equally, many of their effects on ecosystems will be comparable.

The La Rance barrage (Tidal barrage in Figures 2 and 3) is sited

close to the estuary entrance where the equinoctial spring tidal

range was formerly 13?5 m (44 ft) (13?9 m, 45?6 ft highest

astronomical tide (HAT) range). La Rance is a relatively small

(20 km length above the barrage) ria-type estuary (definition: a

flooded, steep-sided, funnel-shaped river valley branching

infrequently and deepening seaward) lacking a broad marginal

alluvial plain, although formerly having a wide intertidal zone.

Before closure La Rance was predominantly sandy and had a low

suspended solids load. It was built by a then-favoured method of

blocking the entire estuary with two cofferdam walls, pumping

out the water and building the structure between them as if on

land. Means to discharge the low river flow were provided, but for

5 years this meant a non-tidal, more or less stagnant lake was left.

Indeed, it changed status from tidal to non-tidal and back,

brackish to fresh and back. Such an approach would not be

favoured today. It was not considered necessary at the design

stage to provide fish passes in the barrage.

The Severn has a HAT range of 14?7 m (47 ft) and is an estuary

of mixed type. It is principally a coastal plain estuary, having

wide reclaimed alluvial flats on the Gwent Levels along the

Welsh coast (Peterstone–Wentlooge and Caldicot Levels), Vale

of Berkeley, Gordano Valley and Yeo–Kenn Levels on the

English coast (Figure 1). However, between Cardiff and

Lavernock Point on the Welsh coast and from Portishead

virtually continuously through to Brean Down on the English

coast the estuary has a high, often cliffed coast. To seaward of

the proposed barrage line, and other than for the Somerset

Levels, both sides of the Bristol Channel have a cliffed coast.

Whereas La Rance was predominantly sandy, the Bristol

Channel and much of the outer Severn has a rocky bed. Along

with the scarcity of unconsolidated sediment, the other

dominating feature is its muddiness. Powerful but variable

strength currents constantly remobilise and redistribute a high

proportion of the fine sediment in the system. This places an ‘all

pervasive’ fingerprint on the ecosystems of the Severn from

microscopic (internal floc micro-climates) up to a regional scale.

By comparison, the relatively localised unconsolidated sand

fraction has a less dominant impact on subtidal ecosystems, and

has little or no effect on the waterbody and intertidal zone. The

intertidal zone of the Severn is exceptionally wide (up to 3 km).

Other than locally, principally along its axial zone and a few

‘perched’ beaches at Sand Bay, Weston Bay and Brean to

Burnham-on-Sea, where it is sandy, the intertidal zone is

muddy. Salt marshes are absent or often limited to a narrow

fringe. At their surface the mud flats exhibit Flandrian fossil

forest beds, peats, at times over-consolidated terraced clay

exposures, and with (at times and in places) contrasted, under-

consolidated fluid mud veneers. Normally-consolidated mud

deposits, the typical host substrates for invertebrates, are

uncommon. These various physical attributes have a strong

controlling influence on water quality and ecosystems.

A Severn barrage would extend 16 km across a narrows from

Maritime Engineering 162 Issue MA1 Comparing environmental effects of Rance and Severn barrages Kirby N Retiere 11


IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

Lavernock Point on the Welsh side to just south of Brean Down on

the English side (Figure 1). To be comparably sited to that at La

Rance, a barrage would need to cross the Bristol Channel between

Ilfracombe in Devon and The Gower in Sourth Wales (see inset on

Figure 1). Studies have shown the Cardiff–Weston line to optimise

output versus construction cost. Should such a structure be decided

upon, it would be built ‘in the wet’ in the manner perfected by the

offshore oil and gas industry, using methodologies closely

analogous to those used successfully in the Mulberry harbours

installed on the coast of occupied Europe in 1944. Extensive use

would be made of prefabricated concrete caissons built at a number

of distant sites, then floated and towed into position. This

construction method, plus the alignment and installation sequence,

has already been largely optimised during earlier phases of the

feasibility study. About 7 years will be required for the completion

of these works plus up to a further 2 years to finalise construction,

installation and achieve maximum output from all turbines. By

maintaining permeability across the estuary until an advanced

stage, such a method has the added bonus of preventing significant

reworking of sediment and associated water quality and ecological

changes during this prolonged period. By adopting such methods,

‘closure’ by installing sluice gates and turbine doors, could be more

or less instantaneous. Sudden closure – that is, during one neap

tide – presently seems to offer a number of benefits.

La Rance Severn

Length: km 0?75 16Enclosed area: ha 2200 48 000Sluices 6 166Turbines 24 216Turbine diameter: m 5?35 variable pitch (25˚ to +35 ) 9?0 (variable pitch)Turbine rpm 94 50Generators: MW 10 40Installed capacity: MW 240 8640 (66 a pressurised water reactor)Annual output 600 million kW h/year ,17 TW h/year (17 000 million kW h/year)Crossings Road (25–30 000 vehicle crossings/day, saving 30 km

each)Road and rail

Vessel facilities Recreational boat lock (17 000 boat movements/year) Large commercial and two recreational shiplocks

Supplying 100 000 persons (8% of electricity consumed in Brittany– equivalent to city of Rennes)

5% UK electrical consumption

Visitors: number/year 350 000 –Other outputs – Produce ‘green’ hydrogen?

Table 1. Scale comparison

WALES

Pembroke CarmarthenBay

SwanseaBay

BridgwaterBay

Bristol Channel

Severnestuary

The Gower

Lundy lsland

BarnstapleBay

Ilfracombe

CelticSea

HartlandPoint

WALES

R Usk

R WyeBerkeley

Gloucester

Netheridge(Hempsted)

Vale of Berkeley

OldburyCaldicotLevels

Gwent LevelsPeterstoneWentlooge

Barry LavemockPoint

CardiffBayBarrage

Nash Point

BRISTOL CHANNEL

HurlstonePoint

Minehead

WatchetSomerset Levels

Burnham-on-Sea

ENGLAND

ENGLAND

Weston-super-MareWeston Bay

R Axe

Flat HolmBreanDownSteep Holm

Cutver Sand

BRIDGWATER BAY

HinkleyPoint

Avonmouth

R AvonPortishead

GordanoValley

Yeo KennLevels

Sand Bay

Bristol

Newpo

rt Dee

p

Deep

Middle Ground

SEVERN ESTUARYThe ShootsUskmouth

N

km0 8 16

Figure 1. Locality map of Severn estuary and Bristol Channel

12 Maritime Engineering 162 Issue MA1 Comparing environmental effects of Rance and Severn barrages Kirby N Retiere


IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

2. RANCE ESTUARY

2.1. Preconstruction physical and ecological regimes

The Rance river discharge at the tidal limit at Le Chatelier lock

(Figure 2) is relatively minor. The average flow is 7 m3/s, with a

low summer discharge of 0?5 m3/s and a decadal statistical

average maximum of 80 m3/s. By comparison, the neap tide

input with the barrage in place is 9000 m3/s and on springs it

rises to 18 000 m3/s.1

This disparity in scale, which was obviously greater prior to the

construction of the barrage, led to the brackish zone being

limited to 5 km in winter and only 2 km in summer down-

estuary from the tidal limit (winters tend to be wetter than

summers).2 Other than for the inner reaches, the disparity in

scale of the fluvial and marine water inputs led to the salinity

structure being well mixed and fully saline. Figure 3 shows pre-

and post-closure salinity.

The Rance river drains a small catchment area, its coast is often

high and rocky, and the sea floor immediately to seaward is

mainly exposed bedrock, sand and gravel.3 Roa Morales2

undertook a sedimentological and oceanographic study of the

Rance Estuary. In the course of research into the turbidity

regime, focused on the proposed barrage line and mainly

undertaken in 1955 and 1956, he made 19 visits to 11 anchor

stations, taking 259 water samples for gravimetric analyses of

the fine fraction.

Concurrent velocity measurements were made. These covered a

range of seasons, all phases of the spring–neap cycle, and

spanned the entire estuary. He was unable to measure complete

ST MALO

ST JOUAN

St SULIAC

Mont GARO

Port St JEAN

Port St HUBERT

ST SAMSON

DINAN

0 1 2 3 km

LEHON

LANVALLAY

LA SOUHAITIER

PLESLIN

PLOUERLA

VILLEGER

MORDREUCPLEUDIHEN

CHATELIERlock

Pont deLESSARD

Pointe GAREL

CHATEAUNEUF(I & V)

DINARD

LA RICHARDAIS

PLEURTUIT

JOUVENTE

Anse desRIVIÈRES

Tidalbarrage

Open sea

Maritime basin

Fluvial Rance

LA RANCE

Figure 2. Locality map of Rance estuary (after Reference 1)



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

13 h 30 min tidal cycles but managed 11 complete and two

partial flood tide cycles. His initial measurements were all made

with the bottle sampler and current meter located at 1 m above

the bed, although later he collected surface and bed, flood and

ebb samples. A map of the pre-barrage near-bed suspended

sediment regime from the mean flood tide results has been

compiled (Figure 4), which embraces most of the total samples

obtained. In the down-estuary reaches on spring tides, large

eddy systems at times brought sand off adjacent banks into the

bottle sampler. This occasional sand component was not

considered and Figure 4 shows only the fine fraction. The Rance

Maritime down-estuary of the Port St Jean/Port St Hubert

narrows has extremely low turbidities, mainly less than 10 mg/l.

Turbidity only rises from Mordreuc and up-estuary, forming a

well-developed maximum turbidity zone extending to the Le

Chatelier lock. In this zone mean values in the range 180–

210 mg/l occurred. This is comparable with measurements in

this and other French estuaries reported by Allen and

Klingebiel,4 and is further consistent with later post-barrage

measurements reported by Bonnot-Courtois et al.1

Consistent with this background, bed sediment maps for 1883

(published 1889) and 1956 are very similar, showing a narrow,

axial deep-water channel with rocky outcrops and gravel

(Figure 5). Sand covered the greatest area of the bed in both the

intertidal and subtidal zones, whereas tributary branches and

embayments along both shorelines were occupied by either

sandy or muddy deposits, the highest zones of which were

colonised by salt marsh vegetation. The Rance has a small rural

catchment and there were no significant industrial discharges.

Sewage and agricultural inputs were probably at a level

determined by the lowly populated hinterland. Water quality in

the Rance has traditionally been high. Major man-induced water

quality crises occurred over time periods consistent with initial

closure and reopening of the estuary, and in recent years excess

nutrient inputs are inducing large-scale plankton blooming in

the course of hot dry summers. These latter are no worse than

those occurring elsewhere in small estuaries of the north coast of

Brittany. The former crises, although not the latter, are related to

the presence of a barrage.

Until the latter part of the 20th century, marine biologists tended

to study species as opposed to habitats, communities or

ecosystems.5 Arising directly from this there are no such data for

the preconstruction era. However, the Museum National

d’Histoire Naturelle was set up on the bank of the Rance at

Dinard in 1882 and a great deal of information was obtained in

the 80-year period prior to commencement of the works. Species

richness was already very high prior to the 1960s, with no less

than 110 species of polychaete worms, 47 species of decapod

crustaceans and 70 species of vertebrate fish.5 The lack of

benthic community studies prior to construction can be side-

stepped and their nature can be back-predicted in two ways.

First, benthic invertebrate communities are characterised by

‘type or indicator species’. If past records of these exist, local

zoogeographical knowledge can be used to anticipate their

associated communities. Second, in confined water bodies there

is a ‘cascade’ of interactions – ria bathymetry determines the

tidal current regime, which dictates the bed sediment distribu-

tion and thus the substrate-dependent invertebrate benthic

faunal communities. Consequently, using the large biological

database, the 1889 and 1956 bed sediment distribution maps are

Tidal barrage

SAINT SULIAC

Salinity

30 to 34‰

20 to 30‰

< 20‰

PORT SAINT JEAN

MORDREUC

CHATELIER lock

2 km

Before 1963 After 1966

ˆ

Figure 3. Distribution of salinity in La Rance prior to and after construction of the tidal powerbarrage (after Reference 1)



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

a ‘surrogate’ for benthic faunal communities. Prior to con-

struction, Rance exhibited four prolific intertidal communities:

rocky hard grounds, gravel and coarse sand, coarse and medium

sand, and mud/saltmarsh over 75% of the total wetted area. By

comparison, the subtidal zone was less important, being limited

to 25% of the estuary and dominated by highly mobile, poorly

populated sands, with the impoverished benthic community

showing very limited biodiversity with subsidiary, equally

impoverished, gravel invertebrate communities. The vertebrate

fish population was typical of an estuarine community, whereas

the size and variety of habitats in the intertidal zone ensured

good waterbird numbers, in keeping with adjacent embayments

along the Brittany coast.

2.2. Construction phase physical and ecological regimes

The barrage was built by the (then) current practice of

constructing cofferdams above and below the site, pumping out

the water, and building the structure ‘in the dry’. Over a three year

period (1963–1966), water level was maintained at 8?5 ¡ 1 m,

leaving the foreshore permanently subaerial. Much of the former

intertidal zone became defaunated and there was a mass mortality

of most marine species, which were unable to adapt to these

extreme conditions. Water level was maintained by weekly

flushing of ‘sanitary’ water to balance the river input. Salinity

reduced progressively to 5–10ø in the zone immediately inshore

of the enclosure, becoming completely fresh in the up-basin

reaches. Vertical mixing must have diminished, a large propor-

tion of the suspended sediment settled onto the bed and there was

no exchange or input of sediment from seaward. In the euryhaline

reaches close to the works, highly tolerant species such as Nereis

diversicolors (ragworm), Mytilus edulis (mussel), and the verte-

brate fish Promatoschistus sp. (a goby) invaded and abundantly

filled the ecological niche created by mortality of the marine

species. Other highly tolerant fish species, such as eels, which do

not migrate annually, also probably continued to occupy the

water body. Freshwater invertebrate species invaded the reaches

of the enclosure adjacent to the river input. This method of

construction has never been considered for the Cardiff–Weston

scheme and the only lesson to be learned from it is how aquatic

communities recover from such gross interference.

2.3. Post-closure physical and ecological regime

Following reopening on 26 November 1966, marine waters were

once more able to enter the Rance. The new hydrodynamic

regime has led mean water level to be raised by 2?5 m, reduced

the tidal range by 40% and the water volumes exchanged with

the sea by 30%. Tidal range is now 7—8 m on springs and 2?5 m

on neaps. The large increase in basin water volume has led to the

relative fresh water component of input decreasing still further,

such that the saline zone of the ‘Rance maritime’ has become

more extensive (Figure 3), the brackish region close to the tidal

limit reducing in area. Similarly, the relative intertidal/subtidal

areas have become 50:50. The foreshore became intertidal again,

experiencing rising and falling tides. The revised tidal current

regime, other than in the immediate vicinity of the barrage,

became more moderate and benign, 0?7 m/s maximum neap and

1?0 m/s maximum spring, reducing in an up-basin direction.

Arising from this, the formerly minor levels of turbidity have

reduced still further, never rising above 10 mg/l, except in the

immediate vicinity of the Le Chatelier ship lock, where fluid mud

is alleged to occur.

For average and neap tides, a single-action (ebb-only) genera-

tion is employed. For springs higher than 10?5 m CD (Chart

Datum) a double-action (flood and ebb phase generation) is

resorted to. In 1995 and 1996, double action (flood and ebb) was

used 22% of the time. This was the mode for which the turbine

N5

5 5

5

ABC D10

10

10

15J

K

K

"L""L"

200

210

0 1 2 3 km

180

2525

100125

150175

5075

EFHG

Figure 4. Pre-barrage near-bed mean, mainly flood tidesuspended sediment concentration (mg/l) for La Rance. A–J,K and L are taken from Bonnot-Courtois et al.2 – the lattertwo not given station letters in reir thesis. Compare withFigure 8 for Severn

1889 1956

DINARD

St MALO N

GravelSandSandy mudMudSalt marshRock

GravelSandMudRock 2 km

Figure 5. Bed sediment distribution maps for 1883 (published1889) and 1956 of Rance estuary. The similarities between thetwo, coupled with extensive nature of clean sand deposits, arenoteworthy (after Reference 1)



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

blades were designed, being a compromise shape between one

ideal and the other. Slack water periods were formerly never

more than 15 min in duration, but now extend almost to

120 min on 10% of tides, when flood phase pumping is used to

amplify tide height inside the basin. Notably, ebb tides are

especially reduced in strength and are not now able to evacuate

suspended sediment brought in by the rising tide.

Three main ecological consequences, one site specific and two

generic, have occurred. First, the varied intertidal substrates of

the former lower shore have become permanently subtidal, so

increasing the subtidal habitat variety and increasing biodi-

versity and abundance of organisms in this formerly impover-

ished zone. Second, both the intertidal and subtidal bed have

become muddier. This might, in part, perpetuate the fine

sediment settlement incurred during the construction era, but

this has been accentuated by the prolonged stillstands and

reduced capacity of ebb-phase tides. Arising from this siltation,

broad zones of former sands have been converted into muddy

sands or sandy muds. Both these mixtures are more hydrauli-

cally stable than clean sands and attract diagnostic and more

prolific benthic invertebrate communities. Post-closure bed

sediment maps (1982 and 1994), plus benthic community maps

(1971, 1976 and 1995) are shown in Figure 6.

The most important point to emphasise is that bed sediment and

benthic community distributions are more or less indistin-

guishable, so confirming the substrate dependency of the

benthos. A third factor not directly evident from maps is that the

carrying capacity of the foreshore has increased, compensating

for, or more than compensating for, the 33% of area lost.

Increased carrying capacity may arise from the more benign

tidal current regime, increased sediment stability due to

incorporation of mud, more prolonged slack waters, etc. The

factor of shelter is likely to be neutral here. The estuary was

already very sheltered from wind and waves. Whereas the

barrage itself might have some additional, local wave sheltering

effect, the greater wind fetch on the broader water body at high

water, plus prolonged high-water stillstand inducing wind–wave

focusing at particular elevations on the 10% of tides when this

occurs, are likely to counterbalance this. These generic factors,

especially the increased muddiness of the bed and elevated

productivity of the foreshore, might be termed a ‘Rance effect’

and, all things being equal, should be expected at tidal power

schemes elsewhere. Possibly the poorer hydraulic fractionation

of grain sizes of bed sediment is another. These may reflect, too,

the greater uniformity in the hydraulic regime imposed onto the

system by barrage operations. In the case of Rance, an estuary

has become more of an embayment of the sea.

Invertebrate benthic communities were already rich but have

become even richer. The improvement has been much greater in

the formerly impoverished subtidal zone. Studies have shown

how marine species richness developed from the initial start-

point in 1966, taking the first 10 years for stable benthic

ecosystems to be re-established (Figure 7). Since this time,

species richness has increased above the pre-closure condition

(250 benthic species by 1995), and has now stabilised. To be

specific, long-term study (1972–1982) of the principal compo-

nents of the Melinna palmata, Abra alba and Corbula gibba

invertebrate communities shows this rise in species richness and

abundance of individuals. Primary production in the system is

now higher than in the adjacent, unenclosed Bay of Mont-St-

Michel. The relative average annual biomass of Abra alba and

Corbula gibba communities in the muddy sediments of La Rance

is approximately double that of tidal flat and subtidal

homologous communities in the mouth of the otherwise similar

Morlaix river, 135 km to the west on the north Brittany coast

near Roscoff.6

In respect of higher organisms, La Rance is now more important

than before in respect of its waterbird numbers and is designated

as a Ramsar wetland of international importance, the basin

playing host to thousands of overwintering birds. This has been

attributed to a number of factors. Together, the very high

1971 1976 1982 1994 1995

N

2 km

Gravel andhard groundSandFine sandymud

Gravel androckSand

Mud

Fine sandymud

Gravel androckSand

Mud

Fine sandymud

GravelMediumcourse sand

MudHard ground

Sandy mud

GravelSand

MudMuddy sand

Salt marshRock

Sandy mud

Figure 6. Bed sediment (1982, 1994) and benthic invertebrate community maps (1971, 1976, 1995) for the post-construction regimeat Rance. The same grey tones representing sand and mud or muddy benthic communities are used as in Figure 5 (after Reference 1)

Num

ber

of s

peci

es

200

100

0

205

114

240

1966 1971 1976 1995

Figure 7. Graph of species abundance plotted against time afterreopening at Rance



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

productivity of prey species of the intertidal zone, plus the

prolonged tidal flat feeding time occasioned by the differences

in water level outside and within the basin caused by normal

operations, have induced this improvement. Arising from these,

feeding opportunities are further enhanced. The bird densities

on its tidal flats are higher than those on the adjacent Bay of

Mont-St-Michel. The Rance is also important for fish-eating

birds. The feeding opportunities for gulls, cormorants and auks

(razerbills and guillemots) are likely to have been improved, if

only in degree, by increased clarity of the waters. These factors

are also favourable for diving ducks and grebes. Since the 1970s

the numbers of wintering ducks, especially shelduck (Tadorna

tadorna) and Brent geese (Branta bernicula) has continued to

rise. Similarly, the concentration of small-sized waders remains

at a high level for these species. According to Le Mao,7

‘Harnessing tidal power has thus not had any disastrous

consequences for the avifauna’.

Similarly, the species diversity of vertebrate fish has risen above

its previous level. The estuary is used regularly by at least 70

fish species and 30 species use it for spawning or as a nursery.

La Rance is important for catadromous fish (those living in

freshwater and breeding in the sea) – that is, they need to

traverse the barrage several times to complete their life cycle.8

The barrage is not equipped with fish passes but this is not a

barrier to fish migration. Fish and cephalopods are claimed to

make their way through the turbines without damage, due to the

inherent characteristics of the latter (large diameter, 5?35 m, and

slow rotation speed, 94 rpm). Le Mao8 claims that Rance

functions identically with other unenclosed estuaries in respect

of fish access. Mechanical damage by blade strikes leading to

wounds or abrasion of scales has not been recorded. Le Mao8

reports that passage through turbines appears, at times, to

disrupt or disorientate shoals of fish, occasionally leading to

individuals surfacing and becoming prey to fish-eating birds. It

is notable that even fragile species such as the cuttlefish (Sepia

officinalis) move to and fro through the turbines without

damage.

There are no marine cetaceans or mammals in the vicinity of

Rance, and reef-building worms, which colonise hard ground

elsewhere, are not reported. The productivity of the system has

become very high. The productivity enhancement, due to factors

induced by barrage construction, has been further increased by a

rise in nutrient runoff from sewage discharge and allochthonous

inputs from agricultural fertiliser release. Productivity of green

algae is high, and includes the intertidal macrophyte

(Entermorpha sp). This detrimental enhanced algal productivity

is not a consequence of reduced flushing, elevated temperatures

or prolonged stillstands due to the barrage, as adjacent

embayments and estuaries along the north Brittany coast are

identically affected. The Rance basin is a robust and productive

area used at times as a nursery from which to reseed adjacent

zones that, for one reason or another, have themselves become

depleted of organisms as opposed to being an area impoverished

in any way by its tidal power plant.

3. SEVERN ESTUARY

3.1. Preconstruction physical, water quality and

ecological regimes

3.1.1. Present regime. The pre-closure Rance suspended

sediment regime was about 1000 times lower than that of the

Severn today. A review of large data sets specifying the

physical, chemical and biological regimes of the estuary has

been undertaken.9 The nature of the ecology of the Severn is

now agreed with conservationists.10 The Severn exhibits eco-

systems which are naturally suppressed by its physical regime,

in some cases to the extent of total barrenness. Estuarine flora

and fauna are among the most adaptable to fluctuations in

external physical conditions. However, the Severn is dynamic

and turbid to such an extreme degree as to put it beyond the

tolerance of all but the hardiest of common species. This

inhibition arises from the coupled high suspended sediment load

and the repeated cycling of the bulk of this between the water

body and bed (Figures 8 and 9).11

There is an unknown amount more than 30 Mt of fine sediment

in suspension between Watchet and The Shoots on a spring

tide, which reduces to less than 4 Mt 14 days later on a

succeeding neap, the balance settling to form temporary fluid

mud pools.11

On top of this, the severity of the physical regime is believed to

be increasing steadily. Sea level is rising, storm surges are

increasing in height, and their return period is falling sharply.

Storminess is becoming more severe. The tidal range is also

suspected to be rising slowly. These factors, together, are

expected to be steadily raising regional turbidity. The biggest

contributor to the fine sediment budget is foreshore to subtidal

transfer of mud, which is thought to lie in the region 5–

10 million m3/year. Thus, the ability of the estuary to support

many varieties of living organisms is exceptionally low and very

likely diminishing.

3.1.2. Water quality and fauna. A review is provided below of

water quality followed by the higher birds and fish in the

Severn. The water body of the estuary is sufficiently hostile as to

preclude photosynthesis – the basic building block of oxygen-

based food chains.9,12 This inability arises from lack of daylight

penetration, regionally exacerbated, to the most extreme degree

on particular occasions by oxygen depletion induced by

entraining suspended fine sediment in the inner estuary. Such

extreme natural ecosystem suppression is acknowledged as

unusual.

Several factors contribute to water quality; for example, as here,

attributes of the physical regime plus the contaminant inputs.

These are reviewed briefly below. Overall water quality has a

direct bearing on primary production. Comprehensive surveys of

zoo- and phytoplankton in the area shown in Figure 10 were

undertaken between June 1971 and October 1980. This involved

1579 plankton net hauls in the course of 52 separate shipboard

surveys involving 58 stations in every month of the year, other

than December, most months during spring and summer seasons

being sampled in 5 or more years.

The seasonal variability of chlorophyll-a and biomass (mg car-

bon/m3) in the six sub-regions specified in Figure 10 for all

omnivorous and carnivorous zooplankton between November

1973 and February 1975 is shown in Figure 11.

Values in the inner channel are negligible but in the outer Severn

(Figure 11), still down-estuary of the zone of maximum turbidity,9



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

1000

1000

1000

10 000

5000

5000

500

250

500

5 m

5 m

5 m

5000

250

15 000

Figure 8. Mean bed neap suspended sediment concentration (mg/l) between Watchet and TheShoots, Severn estuary, UK. Map shows exceptionally high regional turbidity, often reachingmore than 10 g/l. During this phase, regional turbidity at the surface is lower than on springsand bed concentrations reach their peak

29.3.73

51º30'N

Avonmouth

Low water line

5 m contour

Track plot

Acoustically detectable stationary0 1000 2000

m

45' 40'2º50'W55'3º00'

Bristol Deep

Deep

Newpo

rt

3000 suspensions(fluid mud pools)

Figure 9. Neap tide fluid mud pools in Bristol and Newport Deep, Severn estuary, UK. Fluid mud concentrations reach beyond therange of detection even of high-range turbidity meters (see neap phase bed concentrations in Figure 8) and echo soundings revealtheir distribution



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

there is no primary production. Consequently, there can be even

less prospect of primary production occurring up-estuary of

station 1 off Portishead on Figure 10, from whence mean turbidity

rises steeply into the river estuary above The Shoots. Similarly,

Joint and Pomroy13 estimated annual primary production in the

same outer channel region to be 164?9 g C/m2/year, in the central

channel region, 48?5 g C/m2/year, and in the Inner channel region

6?8 g C/m2/year. It must decline to zero in the outer Severn.

Joint14 explains that the euphotic zone is so shallow and mixing

so intense that any phytoplankton are unable to spend sufficient

time in the surficial layer to attain their minimum maintenance

energy for growth. Dong et al.15 reported that, for the above

reasons, the normal correlations between nutrient loads and

chlorophyll-a values do not exist in the Severn.

Turning to contaminant history, for the most part there is a good

record of medium timescale changes in anthropogenic contami-

nant levels in the Severn, that of Abdullah and Royal16 being

followed by Owens17 for the Severn Estuary Joint Committee.

Reductions in cadmium levels and many other metals were

documented by Owens17 who compared levels to earlier results of

Abdullah and Royal.16 More recent overviews have been under-

5º00' 4º00' 3º00'

Newport

WALES

51º30'

InnerEstuary

HolmIslands

Cardiff

BarryAvo

nmou

th

Outer Estu

ary

INNER CHANNEL

ENGLAND

51º00'

OUTER CHANNEL SOUTHCeltic Sea

LundyN

0

km

25

56 51 45

52 46

47 40

41

37

36

31 27 23 18 12 10 8 6

23

1

4

5

5557 50 44

5458

53 49 43

48 42

35

34

33

29 25 21

20 1415

16

32 28 24 19

30 26 22 17

13 11 9 738

39

OUTER CHANNEL NORTH

CENTRAL CHANNEL SOUTH

CENTRAL CHANNELNORTH

SwanseaBay

Carmarthen Bay

Milford Haven

Figure 10. Grid of 58 sample positions in six zones of the Severn and Bristol Channel monitored in 52 separate surveys betweenJune 1971 and October 1980 involving every calendar month except December. Most spring and summer months sampled in fiveor more years. Institute for Marine Environmental Research, now Plymouth Marine Laboratory

mg

C/m

3

mg

Chl

orop

hyll-

a/m

3

North OuterChannel

South OuterChannel

South CentralChannel

Outer Estuary

Inner Channel

3

0

3

0

9

6

30

20

10

0

20

10

0

North CentralChannel

Omnivores

Carnivores

Chlorophyll

N D J F M A M J J A S O N D J F

1973 1974 1975

N D J F M A M J J A S O N D J F N D J F M A M J J A S O N D J F

Figure 11. Seasonal variability of chlorophyll-a and biomass (mg carbon/m3) in the six sub-regions specified in Fig. 10 for allomnivorous and carnivorous zooplankton between November 1973 and February 1975. Up-estuary of Station 1 (PortisheadPoint) turbidity continues to rise. Institute for Marine Environmental Research, now Plymouth Marine Laboratory



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

taken by Ellis18 and by Langston et al.19 Ellis reviewed water and

sediment quality data for the Environment Agency helicopter

monitoring programme. More than 30 determinants were

measured repeatedly at up to 43 sites over a period 1977–1997.

Since the early 1980s, average concentrations of some dissolved

metals appear to have decreased further; for example, lead (Pb)

and cadmium by a factor of 2. Somewhat limited data indicate

that copper, chromium, nickel and zinc in fine sediments has

decreased by as much as 25–50% since the 1970s, perhaps due to

the contraction of extractive industries over this period. Two

discharges of mercury (Hg) described by Owens, have also been

reduced since this time. The rivers introduce the largest loadings

of organic chemicals. Quite separately, the EU Urban Wastewater

Treatment Directive20 (UWTD) required an order of magnitude

decrease in discharge of particulate organic waste to coastal

waters. The water industry’s ‘Clean Sweep’ programme of

investment for the estuary was embarked upon in 1989. Dumping

of sewage sludge to seaward of the barrage site off Swansea Bay

has been discontinued and most primary discharges were stopped

during the period 1998–2004. In addition to the nine major

improvements mentioned,9 an expensive programme of

improvement has been completed at Netheridge near Hempsted

west of Gloucester in the Inner Severn. Cole et al.21 produced a

comparison of the nutrient status of 33 selected English estuaries

using general quality assessment based on total inorganic

nitrogen (TIN) (5 bio-available N) and total reactive phosphate

(TRP) for P. The Severn rates B on a scale A–D for TIN, being lower

than the mean (C), and C for TRP, being much lower than the

mean (D) of the 33. Taking the large English systems, the Severn is

better than the Thames and Wash and comparable with the

Humber and Mersey.

Arising from review of these large data sets, it is clear that

reductions in industrial effluent and sewage inputs have been

very significant. These are coupled with at least two other

factors. First, this is a large and highly dynamic system with, in

most circumstances, an unusually rapid dispersion and dilution

rate for point source contaminant inputs. The second is the

ability of fine-grained sediment to ‘self-cleanse’ itself of

sediment-adsorbed contaminants. In the Severn this arises from

the fortnightly cycling of fines between aerobic and anaerobic

chemical climates.22 This self-cleansing capability in the Severn,

in this case of cadmium and zinc, was remarked on by Ellis,18

but must also apply to other sediment-adsorbed contaminants.

Arising from these various factors, environmental quality

standards (EQS) for anthropogenic contaminants are rarely, if

ever, exceeded in the Severn.

This gives rise to the anomalous situation in which reductions in

industrial discharges do not translate into whole system

improvements in water quality and biological productivity, as

the all-pervasive factor of the suspended solids levels and

cycling remains unchanged. Equally and as above, the large-

scale improvements at Netheridge are not manifest as an

unconditional benefit, as high temperatures and low summer

flows might still lead to whole reaches of the estuary

experiencing severe oxygen depletion, no longer from the

sewage component but still from the neap–spring entrainment

phase of the organic-rich fine sediment population. Except for

the fine sediment, underlying water quality and dispersion

characteristics are mostly now good. Bearing in mind the effects

of climate change on higher organisms (see later), order of

magnitude reductions in organic input from implementation of

the UWTD can be expected to have induced a further

degradation in biological productivity. Such degradation has

evidently been reported from other coastal systems where

‘improvement’ in water quality has been undertaken.

Exceptionally for a large temperate estuary, the main carbon

sources in the Severn are allochthanous (mainly from rivers)

rather than the normal autochthanous dominance.23 This is

another manifestation of the unusual natural attributes of this

hypertidal turbid system. The implication for susceptibility to

fish kills due to summer de-oxygenation in the Inner Severn is

that these could only be eradicated if both the sewage input and

the large-scale anaerobic fine sediment cycling were to be

eliminated. There is a link between the impaired water quality,

the sediment instability and the fish community.

The vertebrate fish species for Hinkley Point currently includes

82 species.24 The fish fauna of the estuary have been known

since the 1880s. After remaining stable for 100 years and

presumably being, similarly, stable back through time prior to

this, since the 1970s there has been a general increase in both

numbers and in species richness for both fish and crustaceans.

This rise has been attributed to an increase in seawater

temperature, reduced salinity and variations in the index of the

North Atlantic Oscillations,24 that is, due to climate change. New

species, almost all occasional visitors, are captured on fish

screens at Hinkley at the rate of 0?83 species/year; that is, one

every 15 months. A further 2 C rise in sea water temperature can

be expected to increase total fish species richness in Bridgwater

Bay by a further 10%. A notable feature of the vertebrate fish

fauna of the estuary which is not changing is the absence from

the community of exclusively benthic feeding species. This

absence is induced by the fine sediment regime. The migratory

species, on the other hand, are in poor shape and declining. One

factor, among several, affecting salmon is the 7 C rise in mean

winter sea water temperatures due to climate change.

3.1.3. Subtidal and intertidal bed. The subtidal estuary bed has

rocky, sandy and muddy substrates (Figure 12). In Bridgwater

Bay, Newport Deep and Bristol Deep, subtidal mud patches are

barren and have been since they first began to be deposited. The

inability to support living organisms extends back at least

1900 years in Bridgwater Bay.25 The absence of bottom-

dwelling organisms is shown by seven separate lines of

evidence.9

(a) Large numbers of thick slab (3 mm) axial sections of mud

cores have been X-radiographed and show the internal

primary depositional fabric. There is no sign of post-

depositional secondary homogenisation due to mixing by

animals (bioturbation).26,27

(b) Artificial radioactive tracers in core samples such as 210Pb,210Po and 137Cs from weapons testing, power station

discharges and reprocessing are not ‘smeared out’ by the

biological mixing normally universal in the materials.

Industrial Revolution heavy metal discharge signatures are

similarly ‘unsmeared’.26

(c) Palaeomagnetic signatures preserving variation of the

position of the Earth’s magnetic pole in tiny magnetic grains

can be detected extending back to 1900 BP in cores. This is

made possible by the strength of the palaeomagnetic

signature due to very large numbers of very small particles,



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

coupled with absence of subsequent ‘randomisation’ by

bioturbation.26

(d) Magnetic property analysis of mud cores (measuring the

degree of alignment of microscopic, but larger, magnetic

particles to the prevailing tidal currents) shows a degree of

alignment as strong as glacial varves and deep ocean fines.

These are among the strongest natural signatures ever

detected. This testifies both to the strength and rectilinear

nature of the depositing current, as well as to the absence of

subsequent biological mixing.26

(e) Tidal mud flats of the Severn show a long-term net

erosional trend. Eroding mud flats throughout the world are

characterised by accumulation at their free surface and

upshore migration of shelly material (cheniers) winnowed

from the degrading surface. Absence of cheniers from tidal

flats of the Severn and Bridgwater Bay testifies to their

long-term abiotic nature.

(f) British Geological Survey (BGS) maps of the percentage of

biogenic carbonate in the Bristol Channel and Severn estuary

show a diminution to zero or virtually zero in an up-estuary

direction into the turbid reaches.28 Absence of biogenic

carbonate reflects long-term absence of shelly benthic

infaunas. One would not expect invertebrates lacking an

exoskeleton to survive in places where those having such a

growth form cannot. Hence, barrenness is inevitable.

(g) In places in Bridgwater Bay the subtidal mud surface is

veneered by a thin brittle ferro-manganese crust, further

testifying to absence of any biota.26

All of the above items are a reflection of the anaerobic advecting

near-bed mobile layers and stationary fluid mud pools which

preclude colonisation of the subtidal mud patches and much of

the deep water channels.

Rock exposures support localised, ephemeral, and depauperate

patches of the reef-building worm Sabellaria alveolata.

Exceptionally, Sabellaria, although it only ever achieves a

tentative foothold in the Severn, is the only filter-feeding

invertebrate known to cope to any degree with the exceptional

turbidities and is the only invertebrate known to reproduce in

the estuary, all others being recruited by up-estuary advection

of spat following spawning in the Bristol Channel. Much of the

subtidal sandy substrate areas are barren or verging on barren.29

Where found at all, invertebrate benthic faunas are typified at

times by large numbers of recently-recruited juveniles, by dwarf

adults and a total absence of the filter-feeding component of

‘normal’ estuarine communities. All these characteristics are

attributed to various aspects of the exceptionally stressful

suspended fine sediment regime. The estuary exhibits poorly

developed examples of typical estuarine communities found

better developed elsewhere. No species typical of exceptionally

stressed ecosystems occur and there are no threatened or rare

so-called ‘red-book’ species in this zone.

The intertidal zone is less impoverished than the subtidal zone.

When water covered, migratory fish species use this and the

immediate marginal subtidal zone as a least-hostile route

through the estuary. The indigenous fish fauna of this periphery

is exceptionally sparse,30,31 and it is likely that fish avoid the

strong currents and elevated turbidities of the estuary axis

mS(g)S

gSgmS

MsMgM

msGmG

sG

G

SMuddy sandSlightly gravelly sandGravelly muddy sandGravelly sand

MudSandy mudGravelly mud

Muddy sandy gravelMuddy gravel

Sandy gravel

Areas with extensive outcropsof bedrock at seabed, oftencovered by a thin, discontinuous,ephemeral sediment veneer

Gravel

Sand

mSmS

mSmS

mSmS

S sMsM

sGsG sGsG

sGsG

sGsG

sGsG

msGmsG

sGsG

gSgS

gSgS

G

G

G

G

G

G

G

SS

S

S

gSgS gSgS

gSgSgSgS

gSgSgSgS

gSgS gSgS

gSgS

mSmS

S

S

gSgS

(g)S(g)S

(g)S(g)S(g)S(g)S

gSgS

gSgS

gSgS

SS

gSgSSG

G G

G

S

mSmS

SS SM

M

M

M

N

M

M

M

MsMsM

M

M

MsMsM

sMsM

sMsM

gMgMmSmS

mSmS

mGmG

mSmS

mGmGmsGmsG

msGmsG

msGmsG

msGmsG

sGsG

sGsG

sGsG

sGsG

sGsG

sGsGsGsG

sGsG

sGsG

sGsG

sGsG

sGsG

sGsG G

G

S

SS

S

SSG

G

G

sMsM

G

GS

S

M

sGsG

sGsG

sGsG

sGsGsGsG

sGsGsGsG

sGsGsGsG

sGsG

G

GG

mGmGgSgSgSgS

gSgS

S

S

mS

mS

mS

S sM

sG sG

sG

sG

sG

msG

sG

gS

gS

G

G

G

G

G

G

G

SS

S

S

gS gS

gSgS

gSgS

gS gS

gS

mS

S

S

gS

(g)S

(g)S(g)S

gS

gS

gS

SS

gSSG

G G

G

S

mS

SS SM

M

M

M

N

M

M

M

MsM

M

M

MsM

sM

sM

gMmS

mS

mG

mS

mGmsG

msG

msG

msG

sG

sG

sG

sG

sG

sGsG

sG

sG

sG

sG

sG

sG G

G

S

SS

S

SSG

G

G

sM

G

GS

S

M

sG

sG

sG

sGsG

sGsG

sGsG

sG

G

GG

mGgSgS

gS

S

S

Figure 12. Distribution of outcrops of rock (shaded zone) at the sea bed in the Bristol Channel and Severn Estuary based upon Lundyand Bristol Channel, 1:250 000 Series ‘Sea Bed Sediments’ with the permission of the British Geological Survey28



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

completely. When this marginal zone is exposed at low tide it is

utilised by shorebirds. The carrying capacity (numbers per area)

of Severn estuary mudshores for shorebirds is lower than any

other in the UK (Table 2).

The Severn also has the highest ratio of potential energy production

to waterbird carrying capacity of any UK estuary32 (Figure 13).

This is a consequence of its hypertidal nature and directly induced

ecosystem suppression. It is notable that the carrying capacity for

waterbirds is more than 20 times lower than that of certain other

UK tidal power prospects such as the Duddon. As with the fish

population, bird abundance is well documented to be system-

atically changing, in this case reducing. Austin and Rehfisch33

report that during recent warmer winters smaller populations of

seven species of common waders wintered in south-west Britain,

including the Severn, with the smallest species, such as Dunlin,

showing the greatest decline. The relative importance in the UK of

the Severn has declined steadily (Table 3).

Austin and Rehfisch33 report this long-term decline in Dunlin on

the Severn (55 000 in the early 1970s to below 14 000 in the

late 1990s). The 14 000 value is important, being the number of

this species the estuary must support to be maintained in a

‘favourable conservation status’ under the terms of the EU

Special Protection Area. In order to qualify, an area must

support more than 1% of the national population. Figure 14

from the British Trust for Ornithology shows the situation to

2008. Climate change has thus diminished the importance of the

estuary for shorebirds in this dramatic manner. Continued

climate change, possibly accompanied by the recent reduction

in organic input subsequent to implementation of the Urban

Wastewater Treatment Directive,20 could see the trend continuing.

It is clear from the above review that the ecosystems of the

estuary are exceptionally suppressed and that, other than for

indigenous fish and crustaceans, this suppression is increasing,

migratory fish being similarly diminished. Some sub-environ-

ments are already barren, others are heading that way. The

degradational trend and the improved fish habitat for the group

specified are both manifestations of climate change.

3.2. Post-closure physical, water quality and

ecological regimes

3.2.1 Water and bed regime. As noted above, the construction

phase leading to closure would be organised in a manner such

that physical change will be telescoped into the last few tides or

tide. At the moment of closure mean sea level within the basin

will rise about 3 m, current velocities will diminish permanently

and suspended sediment concentration will reduce by at least an

order of magnitude. Raising mean sea level would only

influence the main Severn estuary; the tidal bore, a feature of

the tidal river section with its inclined thalweg, would still

occur. The sequence of changes to the fine sediment regime have

been long predicted and are here endorsed for the first time by

our knowledge of the same effects having happened at Rance

after its reopening in 1966.35–44 As stated above, regional-scale

turbidities in the Severn are perhaps 1000 times higher than

they ever were in Rance, inducing a specific ‘Severn Effect’ on

closure. Fluid mud, deposited preferentially in the main

channels, will progressively dewater to give rise to muddy bed

deposits beyond the present settlement areas. Loss of a

significant fraction of the fine sediment from the water body

will admit daylight and raise dissolved oxygen levels, triggering

the onset of photosynthesis, leading to primary production

being set in motion for the first time in millennia. Phyto- and

zooplankton production will begin, leading to the onset of

meiofaunal and higher organism invasion and production. In

the more benign post-closure hydrodynamic and fine sediment

regime a surge in colonisation of a wide variety of invertebrates

will follow, initially recruited into the basin from the Bristol

Channel. Instead of being restricted to solely muddy and sandy

faunal communities, progressive mixing of substrates will lead

to muddy sand and sandy mud faunas developing. The new

faunal communities will include the full range of estuarine

LocationArea of intertidal

zone: km2Wader months

(Nov–Mar)Winter wader months

per km2 Shore substratum

Moray Firth 23?9 28 992 1210 SandyForth 49?8 131 251 2636 Muddy and sandyWash 270?0 744 981 2759 Sandy and muddyMedway 19?9 96 914 4870 MuddySouthampton Water 13?4 43 397 3239 MuddySevern 196?0 218 222 1114 MuddyDee 81?0 280 788 3466 SandyMersey 45?4 92 413 2036 Muddy and sandyRibble 91?0 209 117 2298 SandyMorecambe Bay 303?0 554 520 1830 Sandy and muddySolway 205?0 252 363 1231 Sandy

Table 2. Winter wader data for months November–March 1981/82–1985/86. Some data are more approximate than others. Datafrom the British Trust for Ornithology34

Severn

Solway

Morecambe Bay

Mersey

Cleddau

Dee

Burry Inlet

Duddon

0 20 40 60 80

Birds per GWh

100 120

Figure 13. Comparison of numbers of overwintering wildfowl inselected estuaries having tidal power potential plotted againstprospective electrical output in GWh (after Clark32) Withacknowledgement to the British Ornithologists Union



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

invertebrates supplemented by the presently absent filter-

feeders. Unlike at present, organisms will be able to grow to

maturity and local reproduction will replace the existing

recruitment of spat and juveniles from seaward. A normal

community structure of juveniles, adults and older organisms,

involving each year-class, will develop. Dwarfism will

disappear. Little need be said in faunistic terms regarding the

zone seaward of the barrage. Figure 12 shows that, other than

for Bridgwater Bay, the Inner Bristol Channel is largely devoid

of unconsolidated sediment. As a result, there will not be

significant alteration to the bed and associated faunal

communities to seaward. Bridgwater Bay will remain a high

turbidity regime.

The indigenous vertebrate fish faunal community is likely to

change less than any other. Species abundance will increase due

to invasion by the seven exclusively benthic-feeding estuarine

fish currently excluded by several aspects of the severity of the

fine sediment regime. Based on the 40 years of experience at La

Rance, there is no reason to suppose that present indigenous

species will find difficulty traversing the barrage line. Saving

the migratory fish fauna represents a challenge unrelated to

barrage construction.

The impact of a barrage on the muddy foreshore, specifically on

shorebirds, remains one of the few questions to which

unequivocal answers are not yet possible. With a rise of about

3?0 m in mean water level, somewhere in the region of 62% of

the intertidal zone will become subtidal, to the benefit of

exclusively subtidal organisms. Predicting how the remaining

38% will respond is a key question. It is inevitable that the

‘Rance Effect’ of increased bed muddiness coupled with elevated

intertidal carrying capacity will combine with the ‘Severn Effect’

arising from gross-scale fines deposition and raised bed

stability. What seems almost inevitable, although some changes

will take longer than others to accomplish, is that the mudshores

in the enclosed basin will trend strongly in the direction of

‘look-alikes’ to the Cardiff Bay foreshores prior to construction

of the Cardiff Bay Barrage.40,45 This had high and convex-

shaped tidal flats, a normally-consolidated substrate, and

supported higher invertebrate and bird numbers than the main

estuary. Both the extensive over- and under-consolidated mud

substrates will disappear. Several factors – enhanced still-

stands, when a greater suspended sediment fraction can settle,

increased up-shore fine sediment migration due to ‘settling lag’,

improved local shelter and greater substrate stability due to

algal binding – will all encourage normally-consolidated mud

beds, the ideal host-sediment for organisms, to develop. Given

time, presently pure sandy intertidal areas (such as Middle

Ground) will, in places, tend to become more stable and muddy,

broadening the 38% of area amenable to sandy mud and muddy

sand invertebrate colonisation. We know from Ferns46 that

maximum invertebrate biomass tends to be sited in the upper-

mid shore. This vital zone might well expand in comparison

with the present situation. Increased (i.e. autochthanous)

primary production on the tidal flats might replace the organic

input lost by cleaning up sewage discharges. A Cardiff–Weston

barrage could not be in place before about 2018. Quite whether

all these fairly inevitable modifications will diminish or reverse

the strong trend of shorebirds to forsake western UK estuaries in

favour of eastern UK estuaries, due to the driving force of

climate amelioration in the east, is unclear. It is inevitable that

relative proportions of bird species using the estuary will

change. Dunlin, a bird that is exceptionally small at adulthood,

are still relatively abundant numerically in the Severn arising

from the preponderance of small-size juvenile and dwarf adult

invertebrate prey species. It seems inevitable that the abundance

of the various bird species will change in a post-closure

intertidal regime. There is no immediately obvious reason why

bird-carrying capacity should not increase to at least the levels

encountered in more westerly UK muddy sites such as the

Mersey and Southampton Water. In such an event, bird

abundance would actually rise rather than fall. Scientists

advising the conservation bodies, similarly acknowledge this.

‘By reducing turbidity a future Severn Barrage, if constructed,

would theoretically increase primary productivity and the

diversity of bottom fauna’ (Reference 19, p. 160).

It is demonstrated above that the constraint on water quality in

the Severn is the suspended sediment regime. This can be simply

confirmed by considering the fate of water, sediment and spat

routinely admitted to the impounded dock systems. When

sediment settles in the quiescent dock water, prolific colonies –

for example, of mussels – colonise these water bodies. This

process would also occur at places up-estuary of the barrage

following closure. Building a tidal power barrage of the type

envisaged would convert the zone up-estuary to a high

macrotidal system (4?5–5?0 m plus an unknown additional

amount due to pumping), not a stagnant lake. Reduced tides

created in the basin would still be higher than is naturally

experienced around most of the UK coast. Flushing rates can do

no other than remain high. Modelling indicates that the salinity

structure of the estuary would be only imperceptibly altered,

confirming the high flushing rate post-closure. From long-term

monitoring at Hinkley Point it seems that freshwater inputs are

increasing. With permanent reduction in the sediment load a

long-standing constraint on fish migration and a site of

Date Ranking

1990–91 9th1998–99 14th2003–04 19th

Table 3. Relative importance of Severn among UK foreshores

71/72 76/77 81/82 86/87 91/92 96/97 01/02 06/07

% c

hang

e

400

360

320

280

240

200

160

120

80

40

0

Figure 14. Average winter (Dec–Feb) numbers of dunlin on theSevern Estuary SPA 1971/2 to 2006/7 from the British Trustfor Ornithology34



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

persistent fish kills in summers with high temperatures and low

river flows in the Inner Severn would be removed. Improvement

works at Netheridge greatly reduced biochemical oxygen

demand induced by primary sewage discharge. These improve-

ments could do nothing to prevent natural sediment-induced

dissolved oxygen (DO) sags under comparable climatic conditions.

In a post-closure regime the sediment-induced DO sags would also

disappear. Whether this provides significant support for already

threatened migratory fish populations remains to be seen.

Arising from the reductions in inorganic and organic inputs,

coupled with mixing in the high macrotidal regime, plus the

benefit gained by prolonged ‘self-cleansing’ of the fine

sediment, finally emphasised by the much more benign fine

sediment regime, water quality will improve by a large amount.

Although exotic and noxious microbiological species from

ballast water flushing, etc. are becoming a problem in UK

waters, there is no reason to suppose that they would be any

more problematic in a post-barrage, high macrotidal regime

than in other estuaries.

4. CONCLUSIONS

The Rance Barrage reached 40 years of age on 26 November

2006. It is presently considered to have a long and indefinable

future; its original bulb turbines remain serviceable with no plan

for their replacement. It produces 600 MWh/year with high

serviceability.

In the past, relatively little information has been available

concerning the ecology of La Rance before, during and

subsequent to construction. This has permitted some to maintain

that lack of such information, plus the intrinsic differences

between Rance and Severn, means little can be learned from the

former which helps to anticipate consequences of a barrage in

the latter. This has been addressed by an extensive review of

long-standing data sets recently made available.

Prior to construction of the Rance barrage the ria was an

ecosystem rich in invertebrates, fish and birds with a low turbidity

and high water quality. Bed sediment maps from 1889 and 1956,

which are more or less indistinguishable, permit the substrate-

dependent benthic invertebrate community to be ‘back-predicted’

in detail and with confidence. After almost total destruction of the

fauna and flora during the 3-year barrage closure period, the

revised hydrodynamic regime on reopening set in motion two

major sedimentological changes with far-reaching environmental

consequences. The estuary subtidal bed changed from predomi-

nantly clean, mobile sand to stable, muddy sand and sandy mud,

inducing more bio-diverse and prolific benthic communities. A

number of hydrodynamic and sediment changes on the foreshore

raised the carrying capacity of this zone such that it at least

compensates for the 33% of area lost by raising low water. The

richness and abundance of this zone is now higher, in one quoted

case double, that of equivalent adjacent sites. This might be called

a generic ‘Rance (Barrage) Effect’. The barrage lacks fish passes,

but comparison with adjacent sites found no evidence that this

has impaired passage of any organisms. Water quality, restored

following reopening, has remained unimpaired.

In the, similarly, pre-barrage hypertidal Severn, but unlike in the

former Rance, the extremity of the fine sediment load plus its

constant cycling between water column and bed severely impair

both the ecosystems of the estuary and the water quality. The

estuary is degrading physically due to enhanced rates of

foreshore erosion, smothering of the subtidal bed by expanding

abiotic mud deposits, and conceptually by raised turbidities due

to increasing tide range and velocities. The water body is unable

to sustain any primary production and fish avoid all but its

shallow margin. In addition to barren subtidal mud patches,

subtidal sand and rock outcrops are virtually so, and the over/

under-consolidated muds of the foreshore have the lowest bird-

and perhaps fish-carrying capacity per area of any in the UK.

Since the 1970s, fish and birds have started to be strongly

affected by climate change in opposite ways: the fish species

abundance except for migratory species is rising, whereas

shorebird numbers have crashed. Strenuous efforts have been

made to improve water quality by cutting industrial discharges

and reducing sewage input by a factor of 10. Underlying water

quality has thus been improved, but the constraining effect of

the fine sediment load prevents this being of other than

superficial benefit to ecosystems.

A Cardiff–Weston barrage built ‘in the wet’ could not prevent a

tidal bore forming in the inner estuary. The barrage would

address both ecological and water quality issues. Hydrodynamic

amelioration of comparable scale to that at Rance would

inevitably induce a ‘Rance Effect’, increasing muddiness in the

subtidal bed zone and improving intertidal substrate ‘quality’ to

compensate for intertidal area lost by raising low water. This has

been predicted repeatedly for the Severn since 1975. However,

the Rance never exhibited ecological and water quality

inhibition. It was originally comparable in species richness and

carrying capacity with the Duddon, Burry Inlet and Dee (see

Figure 13). An order of magnitude decrease in suspended

sediment concentration, plus the induced increased bed sedi-

ment stability and mixing, would be a site-specific ‘Severn

Effect’. Unlike at Rance, which has experienced more limited

changes, it is anticipated that more substantial modifications to

the tidal flats will occur. Reduced currents and increased shelter

are expected to induce a long-term change in shape from low

and concave to high and convex in cross-section. Both under-

and over-consolidated substrates will be replaced by normally-

consolidated deposits. The substrate changes will occur rapidly.

The shape change will evolve over decades. All of these will

favour high-density benthic invertebrate colonisation.

The ‘Rance and Severn effects’ would sum, inevitably leading to

a major increase in biodiversity and abundance of individuals.

There can be few major civil engineering projects in the world

whose construction would induce such an increase, but this is

unavoidable. Equally unavoidable, the ‘Severn Effect’ would

drastically improve water quality, for example jump-starting

photosynthesis and phytoplankton production in the water

column. The estuary would remain high macrotidal with a tide

range and velocities still above other UK macrotidal muddy

estuaries. Consequently, the underlying improvements in water

quality, currently masked by high sediment loads and cycling,

would become manifest. The lower pollutant discharges could

not build up in the basin. Fine sediment stabilisation in the inner

estuary, induced by permanent settlement of most of the

currently-entrained material, would prevent any tendency for

summer fish kills. Removal of a significant fraction of this all-

pervasive suspended material might encourage nutrient-induced

plankton blooming, although the reduced organic burden



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

arising from UWTD20 would contribute to minimising such

effects. Current velocities and flushing rates would prevent this

becoming a problem that was worse than elsewhere in the UK.

What is new for Rance is that there is copious data and a good

understanding of the effects of a barrage; this does increase

‘comfort’ that our 30 years of predictions of substrate,

environmental and water quality improvement with a Cardiff–

Weston barrage in the Severn are robust. For the Severn in

recent years there is a much more comprehensive environmental

database than before, coupled with more sophisticated under-

standing. These latter further strengthen confidence in predic-

tion. It is not implied that a new, more prolific ecosystem in the

Severn would be good or bad. Clearly, as at Rance, it will be

unavoidably different. Which of the examples of biodiversity

does our society prefer: the uniquely depauperate suppressed

ecosystem without a barrage or the proliferation in fauna and

flora with a barrage? In the latter situation, an example of

natural ecosystem suppression would remain in Bridgwater Bay

located seawards of a Cardiff–Weston barrage. It is anticipated

that there remain important ‘issues’ such as navigation, flood

defence, and so on, still to be addressed in the Severn; however,

these are not potential ‘show-stoppers’ in the same way as

ecological and water quality issues are regarded by our society.

REFERENCES

1. BONNOT-COURTOIS C., CALINE B., L’HOMER A. and LE VOT M.

(eds). The Bay of Mont-Saint-Michel and the Rance Estuary.

Recent Development and Evolution of Depositional

Environments. Editions Technip, Paris, 2002.

2. ROA MORALES P. La Maree de Salinite, la Turbidite et les

Sediments de la Rance Maritime. Doctoral thesis, University

of Paris, 1956 (in French).

3. VASLET D., LARSONNEUR C. and AIFFRET J.-P. Carte des

sediments meubles superficiels de la Manche a 1/500,000.

Bureau de Recherches Geologiques et Mineres and Institut

Francais de Recherche pour l’Exploitation de la Mer co-

edition, Orleans, France, 1979, geological map of the

continental shelf.

4. ALLEN G. P. and KLINGEBIEL A. La sedimentation estuarine:

example de la Gironde. Bulletin Centre de Recherche Pau–

SNPA, 1974, 8, 263–293 (in French).

5. RETIERE C. and DESROY N. Ecological impact of the Rance tidal

power station. In The Bay of Mont-Saint-Michel and the

Rance Estuary. Recent Development and Evolution of

Depositional Environments (BONNOT-COURTOIS C., CALINE B.,

L’HOMER A. and LE VOT M. (eds)). Editions Technip, Paris,

2002, pp. 236–239.

6. CLAVIER J. Ecologie Descriptive et Fonctionnelle du

Peuplements des Sables Fins Vaseux dans la Bassin

Maritime de la Rance. Doctoral thesis, University of Paris,

1981, VI, pp. 1–232 (in French).

7. LE MAO P. L’avifaune aquatique de la ria de la Rance. Penn ar

Bed, 1996, 160/161, 55–64.

8. LE MAO P. Peuplements Pisciole et Teuthologique du Bassin

Maritime de la Rance: Impact de l’Amenagement Mare-

moteur. Doctoral thesis, Ecole Nationale Superieur

d’Agriculture, Rennes, 1985.

9. KIRBY R., HENDERSON P. A. and WARWICK R. M. The Severn

Estuary, UK: Why is the estuary different? Proceedings of

the Institute of Marine Engineering, Science and Technology,

Journal of Marine Science and Environment, 2004, Part C2,

3–17.

10. ENGLISH NATURE/THE BRISTOL PORT COMPANY. Severn Estuary/

Mor Hafren pSAC. Agreed scientific statement, pp. 1–38

(1998, unpublished).

11. KIRBY R. Suspended Fine Cohesive Sediment in the Severn

Estuary and Inner Bristol Channel, UK. Report to ETSU,

report no. ETSU-STP-4042, 1986.

12. WILLIAMS R. Zooplankton of the Bristol Channel & Severn

Estuary. Marine Pollution Bulletin, 1984, 15, No. 2, 66–70.

13. JOINT I. R. and POMROY A. J. Primary production in a turbid

estuary. Estuarine and Coastal Marine Science, 1981, 13, No.

3, 303–316.

14. JOINT I. R. The microbial ecology of the British Channel.

Marine Pollution Bulletin, 1984, 15, No. 2, 62–66.

15. DONG L. F., NEDWELL D. B., UNDERWOOD G. J. C. and SAGE A. S.

Environmental Limitations of Phytoplankton in Estuaries.

University of Essex, Colchester, 2000 (DETR Contract No.

CW0694, NE Atlantic Marine Res. Prog.), Final Report, No.

97, pp. 1–51 + appendix.

16. ABDULLAH M. I. and ROYAL L. G. A study of the dissolved and

particulate trace elements in the Bristol Channel. Journal of

the Marine Biology Association of the UK, 1974, 54, 551–

597.

17. OWENS M. Severn estuary – an appraisal of water quality.

Marine Pollution Bulletin, 1984, 15, No 12, 41–47.

18. ELLIS J. C. Water Quality Trends in the Severn Estuary. Water

Research Centre, Swindon, 2004, R & D Technical Report to

UK Environment Agency. No E133, pp. 1–38.

19. LANGSTON W. J., CHESHAN B. S., BURT G. R., HAWKINS S. J.,

REASMAN J. and WORSFOLD P. Site Characterization of the SW

European Marine Sites: Severn Estuary pSAC. Marine

Biological Association, Plymouth, 2003, occasional pub-

lication no. 13.

20. EUROPEAN PARLIAMENT and COUNCIL OF THE EUROPEAN UNION.

Directive 91/271/EEC of the European Union and of the

Council of 21 May 1991 concerning urban waste-water

treatment. Official Journal of the European Communities,

1991, L135.

21. COLE S., CODLING I. D., PARR W. and ZABEL T. Guidelines for

Managing Water Quality Impacts within UK European

Marine Sites. Water Research Centre, Swindon, 1999.

22. KIRBY R., WURPTS R. and GREISER N. Emerging concepts for

managing fine cohesive sediment. In Sediment and

Ecohydraulics: Intercoh ’05 (KUSUDA T., YAMANISHI H.,

SPEARMAN J. and GAILANI J. Z. (eds)). Elsevier, Amsterdam, the

Netherlands, pp. 1–16.

23. MANTOURA R. F. C. and MANN S. V. Dissolved organic carbon

in estuaries. In Thirteenth Symposium of the Colston

Research Society (SEVERN R. T., DINELY D. and HAWKER L. E.

(eds)). Scientechnica, Bristol, 1978, pp. 279–286.

24. HENDERSON P. A. Discrete and continuous change in the fish

community of the Bristol Channel in response to climate

change. Journal of the Marine Biology Association of the UK,

2007, 87, No. 2, 589–598.

25. LONG A. J., DIX J. K., KIRBY R., LLOYD JONES D., ROBERTS D. H.,

CROUDACE I. W., CUNDY A. B., ROBERTS A. and SHENNAN I.

Holocene & Recent Evolution of Bridgwater Bay & The

Somerset Levels. Universities of Durham & Southampton,

Durham, 2002, Report to the North Devon & Somerset

Shoreline Management Group.

26. KIRBY R. and PARKER W. R. Settled Mud Deposits in



IP: 128.240.229.3

On: Sat, 20 Nov 2010 16:35:54

Bridgwater Bay, Bristol Channel. Institute of Oceanographic

Sciences, Taunton, 1980, Report No 107, pp. 1–65.

27. KIRBY R. Distinguishing features of layered muds deposited

from shallow water high concentration suspensions. In The

Micro-structure of Fine-grained Sediments – From Mud to

Shale (BENNETT R. H., BRYANT W. R., HULBERT M. H., CHIOU

W. A., FAAS R. W., KASPROWICZ J., LI H., LOMENICK T., O’BRIEN

N. R., PAMUKCU S., SMART P., WEAVER C. E. and YAMAMOTO T.

(eds)). Springer-Verlag, Berlin, 1991, pp. 167–173.

28. BRITISH GEOLOGICAL SURVEY. Bristol Channel Sheet 51 N–

04˚W, 1:250 000 Series ‘Sea Bed Sediments and Quaternary

Geology’ (undated).

29. METTAM C. Analysis of Subtidal Sand Samples. Short report

on the biotopes identified in the context of the Severn

estuary and the UK, including their nature conservation

interest to English Nature; Contract No. H505/01, 1998

(unpublished).

30. CEFAS EXPLORATORY BIOLOGY SURVEY. Report of Surveys

Carried Out to Evaluate the Likelihood that the North Middle

Ground is a Fish Nursery Area. Crossaven Report No.

CO102026, 1999 (unpublished).

31. CEFAS BIOLOGICAL MONITORING SURVEY. Report of the Second

Annual Trawl Survey of the North Middle Ground in Relation

to its Role as a Fish Nursery Area. Crossaven Report No.

XO353, 2000 (unpublished).

32. CLARK N. A. Tidal barrages and birds. Ibis, 2006, 148, s1,

152–157.

33. AUSTIN G. E. and REHFISCH M. M. Shifting non-breeding

distributions of migratory fauna in relation to climate

change. Global Change Biology, 2005, 11, No. 1, 31–38.

34. BRITISH TRUST FOR ORNITHOLOGY. 2008 Alerts Status for Severn

Estuary. See www.bto.org/WEBS/alerts/index.htm. Accessed

03/11/2008.

35. KIRBY R. and PARKER W. R. Sediment dynamics in the Severn

Estuary: a background for studies of the effects of a Severn

Barrage. In Environmental Effects of Constructing a Severn

Barrage (SHAW T. (ed.)). Universty of Bristol Press, Bristol,

1975, pp. 35–46.

36. KIRBY R. An Assessment of Changes in the Fine Sediment

Regime Associated with Construction of a Tidal Power

Barrage near the Holms Islands in the Severn Estuary.

Ravensrodd Consultants Ltd report to Severn Tidal Power

Group, 1984 (unpublished).

37. KIRBY R. Changes to the fine sediment regime in the Severn

Estuary arising from two current barrage schemes. In ICE/

IEE Symposium on Tidal Power 30/31.10.1986. Thomas

Telford, London, 1986, Paper 11, pp. 173–186.

38. KIRBY R. The Ecological Implications of Possible changes in

the Sedimentological Regime caused by a Proposed Severn

Barrage. Report to Severn Tidal Power Group, 1987

(unpublished).

39. KIRBY R. Prediction of Post-Barrage Densities of Invertebrates

and Shorebirds: Sediments. Report to ETSU, 1988, report no.

ETSU-TID-4054, pp. 1–34.

40. KIRBY R. Development of a Methodology to Predict Changes to

Tidal Flat Shapes arising from Tidal Power Schemes. Report

to Severn Tidal Power Group, 1988, Ref. STPG Tasks

S1(iii)i1, pp. 1–92.

41. KIRBY R. Sediment problems arising from barrage construc-

tion in high energy regimes: an example from the Severn

Estuary. In Developments in Tidal Energy – Proceedings of

the Third Institution of Civil Engineers Conference on Tidal

Power. Thomas Telford, London, 1989, pp. 233–244.

42. KIRBY R. Environmental consequences of tidal power in a

hypertidal muddy regime. Ecological problems associated

with tidal power management. 30th anniversary of La Rance

Tidal Power Station, Dinard, France. La Houille Blanche,

Revue Internationale de l’Eau, 1997, 341, No. 3, 50–56.

43. KIRBY R. and SHAW T. L. Severn barrage – environmental

reappraisal. Proceedings of the Institution of Civil Engineers,

Engineering Sustainability, 2004, 158, No. 1, 31–39.

44. WARWICK R. M., GOSS-CUSTARD J. D., KIRBY R., GEORGE C. L.,

POPE N. D. and ROWDEN A. A. Static and dynamic factors

determining the community structure of estuarine macro-

benthos in SW Britain: why is the Severn estuary different?

Journal of Applied Ecology, 1991, 28, 329–345.

45. KIRBY R. Effects of sea level rise on muddy coastal margins.

In Coastal and Estuarine Studies, Vol 40, Dynamics and

Exchanges in Estuaries and the Coastal Zone (PRANDLE D.

(ed.)). Springer Verlag, Berlin, 1992, pp. 313–334.

46. FERNS P. N. Birds of the Bristol Channel and Severn estuary.

Marine Pollution Bulletin, 1984, 15, No. 2, 76–81.

What do you think?To comment on this paper, please email up to 500 words to the editor at [email protected]

Proceedings journals rely entirely on contributions sent in by civil engineers and related professionals, academics and students. Papersshould be 2000–5000 words long, with adequate illustrations and references. Please visit www.thomastelford.com/journals for authorguidelines and further details.


rance v severn

Documents