Interactions Between Fish and Birds: Implications for Management

Interactions Between Fish and Birds: Implications for Management

I. G. COWX, Editor

Fishing News Books

00 11/02/2003 12:27 Page i

Interactions Between Fish and Birds: Implications for Management EDITED BY

I. G. COWX Hull International Fisheries Institute University of Hull, UK

00 11/02/2003 12:27 Page ii

© 2003 by Blackwell Science Ltd, a Blackwell Publishing Company Editorial Offices: 9600 Garsington Road, Oxford OX4 2DQ Tel: 01865 776868 Blackwell Publishing, Inc., 350 Main Street, Malden, MA 02148-5018, USA Tel: +1 781 388 8250 Iowa State Press, a Blackwell Publishing Company, 2121 State Avenue, Ames, Iowa 50014-8300, USA Tel: +1 515 292 0140 Blackwell Publishing Asia Pty Ltd, 550 Swanston Street, Carlton South, Victoria 3053, Australia Tel: +61 (0)3 9347 0300 Blackwell Wissenschafts Verlag, Kurfürstendamm 57, 10707 Berlin, Germany Tel: +49 (0)30 32 79 060 The right of the Author to be identified as the Author of this Work has been asserted in accordance with the Copyright, Designs and Patents Act 1988. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by the UK Copyright, Designs and Patents Act 1988, without the prior permission of the publisher.

First published 2003 by Blackwell Science Ltd Library of Congress Cataloging-in-Publication Data is available 0-632-06385-8 A catalogue record for this title is available from the British Library Set in Times by Gray Publishing, Tunbridge Wells, Kent Printed and bound in Great Britain by MPG Books, Bodmin, Cornwall For further information on Blackwell Publishing, visit our website: www.blackwellpublishing.com

00 11/02/2003 12:27 Page iii

Contents

Preface I 1

2

3

4

5

6

7

8

vii

Impact of birds on fisheries Bird–fisheries interactions: the complexity of managing a system of predators and preys W. Dekker and J.J. De Leeuw Key factor analysis to assess cormorant depredation on inland fisheries in the UK J.R. Britton, J.P. Harvey, I.G. Cowx, T. Holden, M.J. Feltham, B.R. Wilson and J.M. Davies The relationship between cormorant and fish populations at two fisheries in England: an overview J.M. Davies, T. Holden, M.J. Feltham, B.R. Wilson, J.R. Britton, J.P. Harvey and I.G. Cowx

3

14

28

A theoretical assessment of cormorant impact on fish stocks in Great Britain M. Diamond, M.W. Aprahamian and R. North

43

Fish stocks, commercial fishing and cormorant predation in the Vistula Lagoon, Poland L. Stempniewicz, A. Martyniak, W. Borowski and M. Goc

51

The impact of cormorants on the eel fishery in the River Havel catchment area, Germany R. Knöesche

65

Do cormorants and fishermen compete for fish resources in the Väinameri (eastern Baltic) area? R. Eschbaum, T. Veber, M. Vetemaa and T. Saat

72

Interactions in the utilisation of small fish by piscivorous fish and birds, and the fishery in IJsselmeer P.J. Mous, W. Dekker, J.J. De Leeuw, M.R. van Eerden and W.L.T. van Densen

84

00 11/02/2003 12:27 Page iv

iv 9

II 10

11

12

13

14

15

16

17

18

III 19

Contents A quantitative assessment of the impact of goosander, Mergus merganser, on salmonid populations in two upland rivers in England and Wales B.R. Wilson, M.J. Feltham, J.M. Davies, T. Holden, I.G. Cowx, J.P. Harvey and J.R. Britton

119

Bird, fish, habitat interactions Interaction between fish and colonial wading birds within reed beds of Lake Neusiedl, Austria E. Nemeth, G. Wolfram, P. Grubbauer, M. Rössler, A. Schuster, E. Mikschi and A. Herzig

139

Management of cyprinid fish populations as an important prey group for the endangered piscivorous bird, bittern, Botaurus stellaris R.A.A. Noble, J.P. Harvey and I.G. Cowx

151

On the feeding ecology of the pied kingfisher, Ceryle rudis at Lake Nokoué, Benin. Is there competition with fishermen? A. Laudelout and R. Libois

165

Seasonal and spatial variation in cormorant predation in a lowland floodplain river C. Wolter and R. Pawlizki

178

Assessing the interaction between cormorants and fisheries: the importance of fish community change A. Carpentier, J.M. Paillisson and L. Marion

187

Fish predation by great cormorants, Phalacrocorax carbo carboides, in the Gippsland Lakes, south-eastern Australia P.C. Coutin and J. Reside

196

Changes in the piscivorous bird community at a Polish submontane reservoir between 1990 and 2000 in relation to water level R. Gwiazda

211

The role of parasites in fish–bird interactions: a behavioural ecological perspective I. Barber

221

Will the explosion of Ligula intestinalis in Rastrineobola argentea lead to another shift in the fisheries of Lake Victoria? J. Marshall and I.G. Cowx

244

Mitigation measures The potential for using fish refuges to reduce damage to inland fisheries by cormorants, Phalacrocorax carbo I.C. Russell, P.J. Dare, H.V. McKay and S.J. Ives

259

00 11/02/2003 12:27 Page v

Contents 20

21

22

23

IV 24

25

26

27

28

Pilot trials to assess the efficacy of fish refuges in reducing the impact of cormorants on inland fisheries H.V. McKay, I.C. Russell, M.M. Rehfisch, M. Armitage, J. Packer and D. Parrott

v

278

Impact of cormorants on the Loch Leven trout fishery and the effectiveness of shooting as mitigation G.A. Wright

288

Enhancing stocks of European eel, Anguilla anguilla, to benefit bittern, Botaurus stellaris B. Knights

298

Experimental manipulation of great blue heron and belted kingfisher predation rates on stream fish J.A. Steinmetz, S.L. Kohler and D.A. Soluk

314

Management Managing a balance between double-crested cormorant numbers and warmwater fish abundance in the eastern basin of Lake Ontario, New York: preliminary insights from a management programme J.F. Farquhar, R.D. McCullough and A. Schiavone Management of the cormorant, Phalacrocorax carbo, and endangered whitefish, Coregonus lavaretus, populations of Haweswater, UK I.J. Winfield, D.H. Crawshaw and N.C. Durie

325

335

Turnover in a wintering cormorant population: implications for management G.A. Wright

345

The impact of fisheries on water birds in Estonia: nature protection and socio-economic perspectives M. Vetemaa, R. Eschbaum, M. Eero and T. Saat

354

Interactions between fisheries and fish-eating birds: optimising the use of shared resources I.G. Cowx

361

Subject index Species index

373 375

This page intentionally left blank

00 11/02/2003 12:27 Page vii

Preface

During a meeting of the European Inland Fisheries Advisory Commission (EIFAC) in Budapest in June 2000, the need to understand fully the key issues regarding the interactions between birds and fish and fisheries was recognised because of their importance to society in general. In view of this importance it was felt that current knowledge of the status on the interactions between birds and fish and fisheries warranted further discussion and dissemination. To this end, the International Fisheries Institute at the University of Hull, in cooperation with the EIFAC and the European Union RECAFE project, organised a symposium and workshop on Interactions between Fish and Birds: Implications for Management which took place in Hull, UK in April 2001. The objectives of the symposium were:

• • • •

to bring together experts in fisheries management and ecology and piscivorous birds to exchange knowledge among countries and to present, in reviews, an assessment of interactions between fish-eating birds and fish populations; to synthesise information on conflicts between fish-eating birds and fisheries, how conflicts arise through bird and fish population ecology and exploitation, and mechanisms for resolving these conflicts; to review current management practices in relation to fish-eating birds and fish stocks, and identify constraints and gaps in our knowledge that affect the application of fisheries management policy; to develop strategies for the conservation and management of fish-eating birds in harmony with fisheries’ interests.

The main outputs of the symposium are presented in the selected papers that make up these proceedings. They are organised around four main themes: (1) Impact of birds on fisheries – analysis of the impact of fish-eating birds on fish populations and communities; (2) Bird, fish, habitat interactions – understanding the relationships between fish and birds and the role of habitat in regulating these interactions; (3) Mitigation measures – assessment of the effectiveness of possible mechanisms for minimising the impact of fish-eating birds on freshwater fish stocks; (4) Management – methods of managing the impact of fish-eating birds on freshwater fish stocks, specifically through optimisation of shared resources. It is hoped that these proceedings will stimulate fisheries scientists, managers and academics to collaborate in further research to improve our understanding of the interaction between birds and fish and fisheries, and promote further research and collaboration

00 11/02/2003 12:27 Page viii

viii

Preface

to reduce the conflicts between these two groups, and protect these important resources worldwide. The production of these proceedings has involved considerable effort by a number of people. In particular, thanks must go to the following for their contribution to workshops held prior to the main symposium and support in reviewing the papers for the proceedings: M. Aprahamian, R. Britton, I. Barber, P.C. Coutin, J.M. Davies, C. Dieperink, M.R. van Eerden, R. Eschbaum, J. Farquhar, G. Gilbert, J. Harvey, P. Hickley, T. Keller, R. Kirk, R.D. McCullough, H. McKay, E. Nemeth, R.A.A. Noble, I.C. Russell, J. Steinmetz, L. Stempniewicz, T. Saat, M. Vetemaa, C. Wolter and I.J. Winfield. I would like to thank David Carss, Jon Harvey, Richard Noble, Andy Nunn and Emma Doy for their considerable assistance in the running of the symposium and Julia Cowx for the production of these proceedings. Finally, I would like to thank the many funding agencies and organisations for their financial support, thus ensuring truly international coverage of the issues and the success of the symposium. Ian G. Cowx University of Hull International Fisheries Institute

01 11/02/2003 12:29 Page 1

Section I Impact of birds on fisheries

This page intentionally left blank

01 11/02/2003 12:29 Page 3

Chapter 1

Bird–fisheries interactions: the complexity of managing a system of predators and preys W. DEKKER* and J.J. DE LEEUW Netherlands Institute for Fisheries Research, IJmuiden, The Netherlands

Abstract The analysis of competition between birds and fisheries typically proceeds through a series of characteristic steps. First, the state of fish and bird stocks are characterised, followed by the analysis of rate variables, including food intake by birds, diet composition of birds and fishing yield. Finally, the impact of bird predation on fishing yield is assessed, implicitly assuming density independence of yield per recruit. This assumption is considered to be neither substantiated, nor realistic. A simplified model of bird–fisheries interactions was developed. The model includes five biota (man, birds, piscivorous fish, prey fish and zooplankton) and explicitly addresses the relation between biomass and production, its density dependence (Schaeffer surplus production model), fisheries, predation and food selection. This strategic exercise enables exploration of the route for more complex and more realistic quantified analyses of interactions. Reviewing the other chapters presented here, it was concluded that state and rate variables of the processes involved are well quantified for the current situation or for the recent past. However, few of the underlying processes are addressed, disabling the assessment of the dynamics of the bird–fisheries system. Keywords: birds, density dependence, fisheries, interaction, model, predation, resilience.

1.1 Introduction Inland waters in many parts of Europe have a long tradition of commercial exploitation of fish resources (Mitchel 1965). Birds competing with fishermen for fish resources have been actively hunted for centuries. Relatively recently the intrinsic value of water bodies for nature conservation has been recognised (Wetland, Ramsar convention, Habitats Directive), resulting in a protected status for areas and birds, including piscivores such as the cormorant, Phalacrocorax spp. Consequently, nature conservationists and fishermen have conflicting interests in the management of exploited wetlands. Scientific advice in this matter requires insight into the dynamics of fish, birds and fisheries. The analysis of the interactions between birds and fisheries based on field data typically proceeds through a series of characteristic steps. First, the total food intake of the bird population is estimated, using estimates of the total number of birds and their daily food intake (Knösche, Chapter 6; Gwiazda, Chapter 16; Vetemaa, Chapter 27). Next, the *Correspondence: Dr Willem Dekker, Netherlands Institute for Fisheries Research, PO Box 68, 1970 AB IJmuiden, The Netherlands (email: [email protected]).

01 11/02/2003 12:29 Page 4

4

Impact of birds on fisheries

composition of the bird’s diet is analysed, specifically focusing on the contrast between commercial and other fish species (Davies et al., Chapter 3; Stempniewicz et al., Chapter 5; Eschbaum & Veber, Chapter 7; Wilson et al., Chapter 9; Wolter & Pawlizki, Chapter 13; Coutin & Reside, Chapter 15). Finally, the impact of bird feeding on fishing yield is evaluated. Typically, it is assumed that all predated fish can contribute to fishing yields to the same extent as the few fish that have escaped predation (Dekker 1997; see also Britton et al., Chapter 2 this volume). This latter assumption excludes density dependent growth and/or density dependent mortality, allowing only first-order approximations of the interaction between birds and fisheries. Complex simulation models (Lammens 1999) can incorporate density dependent processes, enabling quantitative simulation of scenarios for management of fisheries and bird protection, but the high complexity and data-independence mask any potential for analyses of process dynamics. In this chapter the opposite approach is taken: a simple conceptual model is developed, which identifies the fundamental processes as clearly as possible, because discussions on the interactions between birds and fisheries need to be structured before more complex or data-quantified analyses can be developed. It is not the intention to derive a realistic simulation or a data-driven quantification of the processes involved, but to explore the route to be taken in subsequent analyses. The interaction between birds and fisheries exists in marine and freshwater environments, but is most apparent in inland water systems. Inland fisheries are generally organised into small entities, in a local water body, fishing for a local fish stock (Dekker 2000). Consequently, as far as the fish and fisheries are concerned, fisheries–bird interactions are local phenomena. Most birds, however, constitute populations at larger geographical ranges (van Eerden & Gregersen 1995; Veldkamp 1997). Although the international dimension of bird management is manifest, the current analysis will disregard these aspects completely. Focus will be on local processes, in which context a realistic management option is to chase either the birds or fishermen away, wherever they may go. Concepts and processes will be illustrated by examples from Lake IJsselmeer. The fish and fisheries of Lake IJsselmeer have been described by van Densen et al. (1988), the cormorant dynamics by Zijlstra and van Eerden (1991) and a first attempt to analyse the conflict by Dekker (1997).

1.2 Model In the following, a rudimentary model of the dynamics of man–fish–bird interactions is developed. First, a basic model of fish stock dynamics (the Surplus Production Model of Schaeffer, 1954) will be derived from first principles. This model is used to represent the dynamics of fish, as well as those of birds and fishermen. Secondly, a multi-species extension is given, focusing on the interactions between biota. Although parameter values from published literature will be inserted, model calibration or validation is not pursued here. Because the prime aim of this exercise is in the conceptual construction of the model, results will not be supposed to be particularly realistic.

01 11/02/2003 12:29 Page 5

Modelling bird–fish interactions

5

1.3 Single species model 1.3.1

Surplus production model

Schaeffer (1954) developed a very simple model for fish stocks and fisheries that enables quantitative analysis of fisheries data, if and only if the impact of the fisheries is varying and if and only if it is the only varying factor. The model (and its derivatives) has been used for fisheries assessments, but its application is now largely restricted to data-poor situations. In this model, the population is represented by its biomass B, i.e. not structured by length or age. The stock is subject to individual growth, reproduction and mortality. Putting these processes together in a single parameter b, the stock will initially enable a production P proportional to its biomass, P  b  B. With only a single parameter, this model predicts the population either to become extinct or to reach infinite biomass within a few generations. As Malthus (1798) realised, some density-dependent regulation must be incorporated. Schaeffer (1954) assumed a quadratic relationship between stock and production; later derivatives have used alternative, asymmetric functions, accommodating for various assumed model departures. Here, the simplest two parameter model is used. At low biomass, the production is assumed to be directly proportional to the stock biomass, as above. At higher stock size, the ratio of production (P) to biomass (B) is reduced to P/B  b  aB, or P  bB  aB2, where b is the maximum production/ biomass ratio at zero biomass and a is the density dependent reduction in production/ biomass ratio, per unit biomass. Figure 1.1 shows that stock biomass reaches a maximum at b/a, when production P equals zero. This maximum is known in fisheries assessment as the virgin biomass, corresponding to the carrying capacity of the ecosystem for the stock. In this situation, individual growth and reproduction balance the natural mortality, resulting in an equilibrium state. A reduction in biomass (for whatever reason) results in an increased production, thus restoring the original biomass level. Note that production is highest at intermediate biomass levels. Consequently, resilience towards anthropogenic or natural impacts on the stock is high above intermediate stock levels, but low at stock sizes below this level, potentially resulting in a stock collapse.

<-- max productivity, b

1

<--max productivity, b 2*MSY

0.5

carrying capacity b/a a

0

0

50 B

Figure 1.1

V

100

P

P/B

50

MSY

25

0

0

carrying capacity b/a

50

B

100

Relation between production and stock biomass for the Schaeffer model

01 11/02/2003 12:29 Page 6

6

1.3.2

Impact of birds on fisheries

Exploitation

The analysis of fish stock dynamics now proceeds with the introduction of fisheries taking their yield (Y ) from the stock, thereby lowering the net productivity to P  aB  (b2  4a Y))/2a aB2  Y. A new equilibrium is reached, when P  0, for B  (b   2  4a Y))/2a (overfished, unstable equilibrium). (stable equilibrium) or B  (b   (b  The simple Schaeffer model allows the calculation of the maximum production, which – in a fisheries setting – conforms to the maximum sustainable yield MSY  b2/4a, attained at half the carrying capacity BMSY  b/2a. The relative productivity at MSY equals MSY/BMSY  b/2. Note that the constants 2 and 4 in these formulae are a direct consequence of the quadratic production function chosen. This same model can also be applied to predator–prey interactions, where the exploitation by the fishery is equated to the predator’s activity.

1.3.3

Parameterisation

From the parameters a and b, several key quantities were derived, including carrying capacity, MSY and BMSY. Although the parameterisation chosen yielded a simple model formulation, equivalent alternative parameterisations exist, requiring other parameters to be estimated. Foremost, density-dependent effects (a) can probably not be estimated without considering variation in carrying capacity (involving b). If fish stocks happen to reach a low biomass, apparently enabling the estimation of density independent growth (as for perch, Perca fluviatilis L., in Lake IJsselmeer, Fig. 1.2(a)) that might as well be caused by temporary decreased food availability, not resulting in increased growth at all (as for pikeperch, Stizostedion lucioperca (L.) in Lake IJsselmeer, Fig. 1.2(b)). Carrying capacity might be assessed more simply, resulting in a model parameterised by making b equal to the maximum production/biomass ratio, at zero biomass, and b/a the carrying capacity, the average biomass under pristine conditions. In the following, the model is parameterised using a and b as above, but assuming parameter values are derived from estimates of maximum production/biomass ratios and carrying capacity.

1.3.4

Scenarios

The single species model is now applied to examine several scenarios for human intervention on a stock, whether it be birds or fish (Fig. 1.3). The following scenarios are listed in arbitrary order, with an example for fish and for birds. Scenario (a): Reduction of the maximum P/B ratio (Fig. 1.3(a)). Artificial reduction of the maximum productivity reduces the production of a stock, but not the maximum biomass attained at carrying capacity. The development towards this maximum will be slowed down, that is: the resilience towards stochastic variation or other management

01 11/02/2003 12:29 Page 7

Modelling bird–fish interactions

7

measures is reduced. Application of this scenario to fish stocks is unlikely, but has been advocated for birds, culling of eggs or disruption of nesting colonies. Scenario (b): Reduction of carrying capacity (Fig. 1.3(b)). This reduces the maximum biomass as well as the maximum production, to the same degree. Supposing food is the prime factor determining the carrying capacity, this scenario can be envisaged when prey species are exploited (smelt, Osmerus eperlanus (L.), fisheries in Lake IJsselmeer), or birds are chased away from their feeding grounds. The opposite occurs when competitors are removed, freeing additional food resources.

Mean length of recruiting yearclass (cm)

10

(a) Perch

82

9

81

74 67 8

69 97 84 91 83 95 90 73 66 94 78

68

72

77

79

75

71 86 70

87 88

89

76 80 85

7 96 93 92 6 1

10

100 1000 10000 Yearclass strength (number per hour trawling)

(b) Pikeperch

97

Mean length of recruiting yearclass (cm)

20

95 92 82 94 89 83 69 7376 75 90 91 86 8196 93 70

16

68 74 12

72

100000

66 67 71

85

88

80 79 77 87

84 78

8

1

10 100 Yearclass strength (number per hour trawling)

1000

Figure 1.2 Mean length of the recruiting year class in fall as a function of the year class strength, for perch (a) and pikeperch (b) in lake IJsselmeer, the Netherlands (Buijse & Dekker 1996 and unpublished data from the authors). Plot symbols indicate the year of sampling

01 11/02/2003 12:29 Page 8

8

Impact of birds on fisheries (a)

(b)

60

50

50

40 <-reducing max P/B

30

P

40 P

60

reducing carrying capacity

30

20

20

10

10 0

0 0

50

100

0

50

B (c)

(d) 60

60

50

50

40

40

30

constant harvest

P

P

100 B

20

20

10

10

0 0

50

100 B

constant, extra mortality rate

30

0

0

50

B

100

Figure 1.3 Scenarios for human intervention with a stock, using the single species surplus production model

Scenario (c): Extra mortality, constant harvest (Fig. 1.3(c)). In the current equilibrium model this scenario results in a reduction of the maximum biomass. However, this scenario is known to result in extinction whenever stochastic variation (for instance in recruitment levels) occurs. The amount of biomass harvested, the level of variation and the likelihood of extinction are correlated. This scenario applies where the high variability in stocks and the absence of adequate stock estimates makes managers decide to rely on fixed quota. Fixed quota policies can be envisaged for fisheries and as a control measure for bird population management. It is, however, questionable whether the effort required to harvest a stock down to extinction will ever be exerted purposely. Rather, harvesting will cease when harvest success drops below a minimum value. Scenario (d): Extra mortality, constant rate (Fig. 1.3(d)). This scenario resembles scenario b, in that the maximum biomass is reduced. However, the maximum production is more strongly affected, reducing the rate at which the maximum biomass is reached. This scenario applies, where fishing effort is regulated, or limited efforts to impact bird populations are allowed.

1.4 Multi-species extensions 1.4.1

Model structure

The single species model describes the dynamics of a stock, as constrained by the maximal P/B ratio, by the carrying capacity of the ecosystem and by exploitation or

01 11/02/2003 12:29 Page 9

Modelling bird–fish interactions

9

Man

Piscivorous fish

Bird

Prey fish

Zooplankton

Figure 1.4 Schematic representation of the major ecosystem components in the interaction between birds and fisheries

predation. Extending this approach to a multi-species setting requires a single-species module for each of the components in the ecosystem. Buijse et al. (1993) quantified the biomass, production and consumption of the major biota, covering seven fish species, 12 bird species, algae, zooplankton and man. To simplify the model here, the complexity was reduced to five entities: zooplankton, prey fish, piscivorous fish, birds and man (Fig. 1.4). Predation of birds on piscivores is omitted (see below). For each pair of predator and prey, the mutual relation has two sides: the maximal production of the prey determines the carrying capacity for the predator, while the consumption by the predator harvests the actual stock of the prey.

1.4.2

Carrying capacity

For each prey, the maximum sustainable production was calculated and divided by the per capita ration of the predator, to derive an estimate of the maximal carrying capacity for the predator stock. This assumes that food availability is the factor limiting the carrying capacity for the predator. Other factors, such as shortage of nesting substrates (van Eerden & Gregersen 1991) or reproduction-disrupting pollution (van den Berg et al. 1991), might also regulate the population size. In these cases, however, competition for food would have been negligible, including inter-specific competition with fishermen. By default, the current analysis focuses on the conflict between nature conservation and fisheries, in which food limitation is inherent.

01 11/02/2003 12:29 Page 10

10

Impact of birds on fisheries

In the case of prey-fish, several predators compete for a single resource. In principle, this requires explicit modelling of the competition. In practice, however, one of the scenarios often mentioned in fisheries–birds conflicts involves the chasing away of birds from the feeding grounds. This conforms directly to a human-induced reduction in the competitive ability of the birds. Therefore, the competition problem was ignored and a range of values for the competitiveness of birds was explored. However, more realistic models of bird–fisheries interactions will have to tackle the competition problem.

1.4.3

Consumption

Knowing the carrying capacity for a predator and its maximal P/B ratio, the equilibrium biomass can be derived. Multiplication by the per capita ration yields an estimate of the consumption. For fisheries exploitation and culling of birds, a per capita ration is not related to actual mortality rates. Therefore, fixed exploitation rates were used as inputs to the model.

1.4.4

Food choice, prey switching

For birds and man, a modelling problem arises from the availability of more than one prey species. Birds do predate on prey fish and piscivorous fish, while man exploits prey fish and piscivorous fish, and might also hunt birds. Bird predation on piscivorous fish was simply ignored. For the fishery, mortality rates on the three prey items were assumed to be independent parameters, determined by a complex political and scientific process. However, any more realistic model of bird-fisheries interactions will have to tackle the food choice problem. Prey choice becomes even more complex when prey switching is considered: that is, a density dependent food choice.

1.5 Model output The model is intended to review the processes involved in the complex interactions between fish, fisheries and birds. As such, the results of the model are intended to be only indicative. It is the construction of the model that yielded points for discussion. However, because of its simplicity, the model could easily be implemented in a single spreadsheet (Fig. 1.5) and a feeling for general model behaviour obtained. Management measures intending to have an impact on the net reproductive success of birds apparently have no effect on the maximum biomass attained by the bird stock. Consequently, no effect on fishing yields is to be expected. However, lowering reproductive success has a synergistic interaction with a concurrent increase in mortality of the birds. When hunting rates exceed the rate of reproduction, the bird stock collapses. Lowering the rate of reproduction of the birds therefore lowers the hunting rate required to collapse the bird stock. Hunting birds, or decreasing their feeding rate, has a distinct effect on the yield of piscivores. However, this effect is strongly correlated with the exploitation rate exerted

01 11/02/2003 12:29 Page 11

Modelling bird–fish interactions

Man Man has only inputs: shooting fishery birds non-piscivores biomass rate

output:

f=

state:

0.000 shooting, fixed rate

0 0%

fishery piscivores

45 100%

3.0 kg.ww/ha/yr 40% %/yr

/yr

B= P=

Bird actual 0.45 Biomass 0 production

a= b=

3.4 density dep. reduction/kg/yr 150% max. P/B-ratio %/yr

potential 0.22 kg.ww/ha 0.17 kg.ww/ha/yr output: f= state:

input: b/a = d= d*B =

11

0.45 carrying capacity 71 ration per kg 32 ration of population

kg.ww/ha kg/kg/yr kg.ww/ha/yr

B= P= a= b= input: b/a = d= d*B =

35% output: f= c= state: B= P= a= b=

0.400 fishing rate

/yr

Piscivorous fish actual potential 7.5 Biomass 6.9 kg.ww/ha 3.0 production 3.0 kg.ww/ha/yr 0.063 density dep. reduction 87% max. P/B-ratio 14 carrying capacity 4.3 ration per kg 32 ration of population

/kg/yr %/yr kg.ww/ha kg/kg/yr kg.ww/ha/yr

65% 1.0 fishing rate 64 predation Prey fish actual 45 Biomass 45 production

/yr kg.ww/yr

potential 29 kg.ww/ha 91 kg.ww/ha/yr

0.11 density dep. reduction /kg/yr 720% max. P/B-ratio %/yr

input: b/a = d= d*B =

68 carrying capacity 21 ration per kg 952 ration of population

kg.ww/ha kg/kg/yr kg.ww/ha/yr

952 predation

kg.ww/yr

output: c= state:

Zooplankton B= P= a= b=

actual 143 Biomass 0 production

potential 90.5 kg.ww/ha 1431 kg.ww/ha/yr

0.17 density dep. reduction /kg/yr 3162% max. P/B-ratio %/yr

input: b/a =

181 carrying capacity

kg.ww/ha

Figure 1.5 Simplified model of fish–bird fisheries, as defined in the text, applying best guess parameters for Lake IJsselmeer

by the fisheries. The more heavily overexploited the fisheries, the bigger the impact of bird predation. In a moderately exploited fish stock, the fish production has some resilience towards added mortality, while in the overexploited case all resilience has already been exhausted.

01 11/02/2003 12:29 Page 12

12

Impact of birds on fisheries

The increasing interaction between fisheries and cormorants (whether perceived or real) provides the basis in society for this research. The cormorant population in Europe has increased considerably over the past decades. Hence, there has been no equilibrium condition in that period. In this situation, a reduction of the reproductive success of the birds might have slowed their rate of development, temporarily giving the impression of success in controlling the impact of birds on fisheries. However, in the longer run, the bird population will undoubtedly reach carrying capacity, despite moderate reductions in their reproductive success.

1.6 Discussion The model of bird–fisheries interactions developed here was intended to address the processes involved; manageability was achieved by simplification of the true entities and processes, grouping species and/or size classes into ecological groups (prey fish, piscivorous fish, birds and man) and merging separately quantifiable processes into the more fundamental processes: production, its density dependence, harvesting, food intake rate and food choice. The model was generically applied to each of the biota (prey fish, piscivorous fish and birds). The intention was to cover all processes in a simple model, at the price of reduced realism. The low realism of a model of such simplicity must be compared with the current shortcomings of field studies to address the relevant processes. For fish, most studies in this volume describe standing stock, but only a few relate standing stock to biomass production (Diamond et al., Chapter 4; Mous et al., Chapter 8). Density dependence involved in the concept of carrying capacity is rarely addressed (Mous et al., Chapter 8; Russell et al., Chapter 19). For birds, population size and productivity are relatively easily assessed, but population sizes are rarely interpreted in relation to carrying capacity or stock resilience (Farquhar et al., Chapter 24; Wright, Chapter 26; van Eerden & Gregersen 1995). Regarding the interaction between species, the actual diet of predators is commonly described, but its (discontinuous) relation to prey abundance is almost universally left untouched. When diet shifts have been observed, it is interpreted as an indicator of prey abundance, rather than viewing prey abundance as a driving force for food selection. The prime interaction between birds and fisheries is the competition for a common prey, but competition and competitive exclusion have not yet been addressed. An alternative to analytical assessment of the interaction between birds and fisheries might be found in studies comparing study areas with and without bird predation. Either study areas are selected that happen to fit in a comparison scheme (Engström 2001), or predation might be experimentally prohibited, by scaring away the predator (Russell et al., Chapter 19), or by provision of shelter to the prey (Russell et al., Chapter 19; McKay et al., Chapter 20). However, this type of experiment is by its very nature restricted to a limited number of water bodies of a relatively small size. Extrapolation to larger systems and/or other areas still requires analytical insight into the processes involved. The analysis of the processes involved in the interaction between birds and fisheries as well as tentative results of the present simplistic model exercise suggest that density dependence is of crucial importance in understanding the dynamics of the complex

01 11/02/2003 12:29 Page 13

Modelling bird–fish interactions

13

system. Without it, any bird predation on commercial fish is seen as an economic loss, and any impact on birds is seen as a remedy. A profitable fishery in the presence of birds seems to be an impossibility, but that is denied by a long tradition of commercial exploitation of inland waters in Europe. The inevitable conclusion is that the past and current state of fish and bird stocks and the rate of biomass flow from prey stocks to predators can be described, but the analysis of the complex dynamics involved in the interactions between birds and fisheries has only just begun.

References Buijse A.D., van Eerden M.R., Dekker, W. & van Densen W.L.T. (1993) A trophic model for IJsselmeer (The Netherlands), a shallow eutrophic lake. In V. Christensen & D. Pauly (eds) Trophic Models of Aquatic Ecosystems. ICLARM Conf. Proc. 26, 390 pp. Buijse A.D. & Dekker W. (1996) Uncertainty in fish stock assessment based on bottom trawl surveys in Lake IJsselmeer. In I.G. Cowx (ed.) Stock Assessment in Inland Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 260–279. Dekker W. (1997) The impact of cormorants and fykenet discards on the fish yield from lake IJsselmeer, the Netherlands. In C. van Dam & S. Asbirk (eds) Cormorants and Human Interests. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 45–52. Dekker W. (2000) The fractal geometry of the European eel stock. ICES Journal of Marine Science 57, 109–121. Engström H. (2001) Long term effects of cormorant predation on fish communities and fishery in a freshwater lake. Ecography 24, 127–138. Lammens E.H.H.R. (1999) Het voedselweb van IJsselmeer en Markermeer. Veldgegevens, hypotheses, modellen en scenario’s. RIZA Rapport 99.008, 160 pp. (in Dutch). Malthus T.R. (1798) An Essay on the Principle of Population as it Affects the Future Improvement of Society, with Remarks on the Speculations of Mr. Godwin, M. Condorcet, and Other Writers. J. Johnson in St Paul’s Churchyard, London. Mitchel N.C. (1965) Ulster Folklife 11, 1–32. Schaeffer M.B. (1954) Some aspects of the dynamics of populations important to the management of the commercial marine fisheries. Bulletin of the Inter-American Tropical Tuna Committee 1, 27–56. Van Berg M., Craane L.H.J., van Mourik, S. & Brouwer A. (1991) The (possible) impact of chlorinated dioxins (pcdds), dibenzofurans (pcdfs) and biphenyls (pcbs) on the reproduction of the Cormorant Phalacrocorax carbo – an ecotoxicological approach. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 299–313. Van Densen W.L.T., Cazemier W.G., Dekker W. & Oudelaar H.G.J. (1988) Management of the fish stocks in Lake IJssel, The Netherlands. In W.L.T. van Densen, B. Steinmetz & R.H. Hughes (eds) Management of Freshwater Fisheries. Wageningen: Pudoc, pp. 313–327. Van Eerden M.R. & Gregersen J. (1995) Long-term changes in the northwest European population of cormorants Phalacrocorax carbo sinensis. Ardea 83, 199–212. Veldkamp R. (1997) Cormorants (Phalacrocorax carbo) in Europe: population size, growth rates and results of control measures. In C. van Dam & S. Asbirk (eds) Cormorants and Human Interests. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 21–29. Zijlstra M. & van Eerden M.R. (1991) Development of the breeding population of cormorants Phalacrocorax carbo sinensis in the Netherlands till 1989. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 53–60.

02 11/02/2003 12:37 Page 14

Chapter 2

Key factor analysis to assess cormorant depredation on inland fisheries in the UK J.R. BRITTON*, J.P. HARVEY, I.G. COWX University of Hull International Fisheries Institute, Hull HU6 7RX, UK

T. HOLDEN, M.J. FELTHAM, B.R. WILSON and J.M. DAVIES Liverpool John Moores University, Byrom Street, Liverpool, UK

Abstract Studies on the impact of cormorant, Phalacrocorax sp., depredation on inland fisheries generally concentrate on the derivation of the number or biomass of fish removed by birds from the total fish density or biomass present. However, this does not provide a detailed assessment of impact as it rarely provides insight to how the actual species, size ranges and year classes of fish being depredated upon are affected. To achieve this, key factor analysis was developed to reveal the status of fish populations being affected by cormorant depredation and their status if depredation had not occurred. Key factor analysis at Holme Pierrepont Rowing Course, Nottingham, UK, showed cormorant depredation reduced the numbers of roach, Rutilus rutilus (L.), common bream, Abramis brama (L.), and perch, Perca fluvialtis L., by 62%, 51% and 65% respectively, and the biomass by 72%, 67% and 75% respectively. The cohorts of roach that dominated angler catches in 1998, i.e. the 1995 and 1996 year classes, were reduced by cormorant depredation by 75% and 65% respectively. The data highlighted that cormorant depredation reduced the availability of fish for angler exploitation in subsequent years. Keywords: angler exploitation, bream, cormorant depredation, key factor analysis, perch, roach.

2.1 Introduction The two sub-species of cormorant, Phalacrocorax carbo carbo (L.) and Phalacrocorax carbo sinensis (Blumenbach) that inhabit UK waters have been protected under UK legislation since 1981 (Wildlife and Countryside Act 1981). This contributed to the inland cormorant population increasing by 74% over a 3-year period in the late 1980s–early 1990s (Kirby et al. 1995). By 1996, there were approximately 6000 overwintering, inland cormorants present in the UK (Russell et al. 1996), although numbers are now believed to be stable (Wernham et al. 1999). As cormorants are piscivorous birds, with a daily food intake of between 340 and 520 g (Kirby et al. 1996), their

*Correspondence: Dr Rob Britton, Environment Agency, National Fisheries Laboratory, Huntingdon, UK (email: [email protected]).

02 11/02/2003 12:37 Page 15

Key factor analysis to assess cormorant depredation

15

increased presence on inland fisheries raised concern among angling groups over their deleterious impact on fish populations, with requests in the angling media that their protected status be lifted (Britton 1999). However, ornithological groups argued that there were few robust data on the impact of cormorants on inland fisheries in the UK and this void in knowledge required filling before their actual impact on inland fisheries could be ascertained and policy formulated. Although a great deal of effort has been put into estimating the impact of cormorants at freshwater fisheries (Kirby et al. 1996; Russell et al. 1996; Feltham et al. 1999), few previous studies have moved beyond estimating the number or mass of fish removed from a particular fishery by cormorants over a set period of time. For example, in salmonid river fisheries in County Mayo, Northern Ireland, cormorants consumed annually between 5 and 13.1% of salmon, Salmo salar L., and sea trout, Salmo trutta L., smolts (MacDonald 1987). Feltham & Davies (1995) estimated that cormorants removed between 2% and 38% of the stock of chub, Leuciscus cephalus (L.), roach, Rutilus rutilus (L.), and dace, Leuciscus leuciscus (L.), from the rivers Darwen and Ribble in north-west England. However, in these studies, little attention was given to the impact of cormorant depredation on the actual species, year classes and size classes of fish being predated upon. To understand the impact of the depredation, it is as important to ascertain the changes in fish community structure that occur during and after the depredation as it is to ascertain the amount of fish being removed. For example, the impact of density-independent losses to cormorants of fish species with life history strategies characterised by slow growth and a long life span may be felt for a number of years in a fishery, for example through the loss of future broodstock and fish valuable to anglers for exploitation. In this paper, key factor analysis was used to assess the impact of cormorant depredation at Holme Pierrepont Rowing Course, Nottingham, UK. Key factor analysis integrates data on fish population dynamics and cormorant depredation to provide a detailed impact assessment on specific fish species, their age structure and year classes. The methodology determines the status of the main fish populations at fisheries under the influence of cormorant depredation and predicts their status if cormorant depredation had not occurred. This enables estimates of fish losses due to cormorant depredation to be determined in a more holistic way than previously.

2.2 Materials and methods 2.2.1

Study area

The study was carried out between September 1995 and August 1998 on Holme Pierrepont Rowing Course, situated at the National Water Sports Centre, Nottingham (Fig. 2.1). The Rowing Course is 2215 m long and 135 m wide, a total water area of 28.68 ha, and a maximum depth of 3 m. Recreational, catch and release angling is practised on the Rowing Course, although the site is primarily used as a centre of excellence for other water sports. The Rowing Course is located within the locality of an

02 11/02/2003 12:37 Page 16

16

Impact of birds on fisheries

Figure 2.1 The River Trent catchment showing location of Holme Pierrepont Rowing Course and the cormorant roost at Attenborough gravel pits

over-wintering, inland cormorant colony sited at Attenborough Gravel Pits, Nottingham (Fig. 2.1). During their over-wintering period (October to March), cormorants from the Attenborough roost visited the Rowing Course daily for foraging purposes (Feltham et al. 1999).

2.2.2

Assessment of cormorant feeding ecology

Cormorant foraging rates at the Rowing Course were monitored by counting the number of birds from dawn until dusk on one day each month (Feltham et al. 1999).

02 11/02/2003 12:37 Page 17

Key factor analysis to assess cormorant depredation

17

Cormorant feeding observations were also carried out on the Rowing Course, on average three times per week. These allowed site-specific length–weight relationships to be used to estimate the numbers and mass of fish removed during each cormorant foraging bout (Britton 1999; Feltham et al. 1999). The data generated were applied to a Monte Carlo Simulation (MCS) model. The MCS output provided an estimate of the numbers of each species, separated into discrete length classes, that were depredated by cormorants during each over-wintering period (Britton 1999; Feltham et al. 1999). Integration with fish population age and growth data (Section 2.2.3) enabled determination of the numbers of fish in each age group that were removed by cormorants in each over-wintering period (Britton 1999; Feltham et al. 1999). The data were then converted into numbers per hectare taking into account the area of the Rowing Course.

2.2.3

Assessment of fish populations

Data on fish populations on the Rowing Course were collected using boom boat electric fishing (Harvey & Cowx 1995; Feltham et al. 1999). Surveys were carried out on a quarterly basis over the 3-year study period, in October, January, March and June of each study year. Species composition, fork length (mm) and weight (g) were recorded, and scale samples removed for age and growth analysis. To ascertain the response of fish populations to cormorant depredation and allow key factor analysis to predict how the fish populations would behave if cormorant depredation had not occurred, the following indices were calculated and incorporated into the model: back-calculated growth rates (Bagenal & Tesch 1978); mortality and survival rates (Ricker 1975); and year class strengths (Feltham et al. 1999).

2.2.4

Key factor analysis

Key factor analysis involved two major components: the calculation of the number and biomass of fish present in the fishery after cormorant depredation and the calculation of the number and biomass of fish that would have been present in the fishery had the cormorant depredation not occurred. The status of the fish populations after the cormorant depredation occurred was calculated by cohort analysis. The status of the fish populations had the depredation not occurred was calculated by integration of the cohort analysis data with that from the MCS (Section 2.2.2). Cohort analysis was applied to each fish species depredated by cormorants during the study period. The first fishery survey of each sampling year (October) was used as a starting point for the calculations and the individual fish lengths were assigned to their corresponding age groups, derived from growth rate analysis (Section 2.2.3). This allowed the calculation of the number of fish caught in each individual age group, and these data were used to calculate mortality rates of individual fish in the population (Ricker 1975). Owing to sampling bias of electric fishing against young fish (Zalewski & Cowx 1990), the greatest number of fish captured were often not found in the first age group,

02 11/02/2003 12:37 Page 18

18

Impact of birds on fisheries

as would be expected, and a linear decline in mortality was not found until later age groups. Therefore, the cohorts of each fish species were reconstructed to allow calculation of the number of fish in these younger age groups had sampling been non-selective. Mortality rate was converted to survival rate (Ricker 1975) and the value applied to the number of fish in the first age group where a representative decline was observed. This allowed an estimation of the number of fish in the previous age group, and this was repeated until the numbers of fish at age 1 were calculated, so removing the problem of sampling bias. Recruitment is known to vary annually in cyprinid populations (Mills & Mann 1985; Cowx et al. 1995), therefore the cohort reconstruction data required weighting for year class strength (YCS). Year class strengths were derived for each fish species in the population (Feltham et al. 1999) and divided by 100 to give values representing the YCS weighting factor. These were multiplied by the number of fish in the corresponding year class in the reconstructed cohort. The efficiency of the electric fishing gear had to be accounted for in the data (mean efficiency  0.23; Feltham et al. 1999), along with the area sampled. This resulted in a quantitative estimate of the numbers of fish per hectare and so the cohort was fully reconstructed. These data represented the number of fish by species and age group present in the fishery following one winter or more of cormorant depredation. To calculate the number of fish that would have been present in the absence of cormorant depredation, the data generated in the MCS model (Section 2.2.2) were used. The cohort reconstruction from the October 1996 fisheries survey data utilised MCS data from winter 1995/1996, which gave numbers of fish by age group in October 1996 both with and without the cormorant depredation from the previous winter. The cohort reconstruction from October 1997 fisheries survey data used MCS data from both the 1996/1997 and 1995/1996 winters’ cormorant depredation. The cohort reconstruction from June 1998 fisheries survey data used MCS cormorant depredation data from the winters of 1997/1998, 1997/1996 and 1995/1996. Natural mortality rates were applied to the historical MCS data to take account of fish that would have died from natural causes had they not been depredated by cormorants, before being applied to the model. Finally, the numbers of fish in each age group with and without cormorant depredation were converted to biomass using site-specific length–weight relationships (Feltham et al. 1999) to show minimum, mean and maximum values (kg ha1).

2.3 Results 2.3.1

Cormorant numbers and diet

Early morning cormorant counts at the Rowing Course revealed peak numbers from December to February in each winter surveyed (Fig. 2.2). Numbers decreased dramatically in April of each study year, with cormorants absent at the site between May and September. This was attributed to the breeding cycle of the birds (Britton 1999). Peak winter counts decreased over the study period, from 153 birds in February 1996, 83 in December 1996 to 79 in December 1997 (Fig. 2.2).

02 11/02/2003 12:37 Page 19

Number of cormorants

Key factor analysis to assess cormorant depredation

19

250

Peak early morning count

200

Cumulative daily count

150

100

50

A

F

D

Oct-97

A

J

A

F

D

Oct-96

A

J

A

F

D

Oct-95

0

Month

Figure 2.2 Peak early morning and estimated cumulative daily counts of cormorants feeding at Holme Pierrepont Rowing Course, 1995–1998

Roach, common bream, Abramis brama (L.), and perch, Perca fluviatilis L., were the major species depredated by cormorants at Holme Pierrepont Rowing Course, and the majority of fish ingested were below 100 mm fork length (Fig. 2.3).

2.3.2

Fish losses attributable to cormorant depredation

Cormorants foraged very efficiently at the Rowing Course, suggesting removal of a considerable quantity of roach, common bream and perch from the fishery, both in terms of numbers and biomass (Feltham et al. 1999). The MCS model estimated that the annual proportion of gross standing crop removed by cormorants ranged from 42.1 to 55.0% over the study period (Fig. 2.4). Key factor analysis revealed the abundance of roach, common bream and perch were reduced considerably following three winters of cormorant depredation (Table 2.1). There would have been a substantial increase in the number and biomass of roach, common bream and perch present in the Rowing Course if the cormorants had not foraged there (Table 2.1). Key factor analysis on the age structure of the roach, common bream and perch populations revealed that after cormorants foraged on the Rowing Course, there was a marked decrease in the abundance of young fish in the community (under 4 years old) (Figs 2.5–2.7). This decrease in fish abundance can only be attributable to cormorant depredation and to no other source of mortality. Roach were the dominant species caught by anglers in August 1998 (96% species composition), with 84% of these from the 1995 and 1996 year classes (Fig. 2.8).

02 11/02/2003 12:37 Page 20

20

Impact of birds on fisheries

Percenatge frequency

80 Roach

70 60

1995/96 (n = 135) 1996/97 (n = 59) 1997/98 (n = 103)

50 40 30 20 10 0 0–50

51–100

101–150

151–200

201–250

251–300

Size class (mm) 80 Common bream

Percenatge frequency

70 60

1995/96 (n = 12) 1996/97 (n = 70) 1997/98 (n = 110)

50 40 30 20 10 0 0–50

51–100

101–150

151–200

201–250

251–300

Size class (mm) 80 Perch

Percentage frequency

70 60

1995/96 (n = 254) 1996/97 (n = 84) 1997/98 (n = 49)

50 40 30 20 10 0 0–50

51–100

101–150

151–200

201–250

251–300

Size class (mm)

Figure 2.3

Diet composition of cormorants at Holme Pierrepont Rowing Course, 1995–1998

02 11/02/2003 12:37 Page 21

Proportion (%) standing crop removed

Key factor analysis to assess cormorant depredation

21

100

80

60

40

20

0 1995/96

1997/98

1996/97 Year

Figure 2.4 MCS estimate of proportion of standing crop removed by cormorants over the study period, 1995–1998

Table 2.1 Standing crop change attributable to cormorant depredation at Holme Pierrepont Rowing Course, 1995 to 1998 1996

1997

1998

Number (n ha1)

Biomass (g ha1)

Number (n ha1)

Biomass (g ha1)

Number (n ha1)

Biomass (g ha1)

Roach With cormorants No cormorants Percentage change

225.1 416.1 184.8

8246.9 21962.3 266.3

1173.6 1560.5 133.0

47065.3 70723.2 150.3

322.2 856.6 265.9

17762.7 62882.6 354.0

Common bream With cormorants No cormorants Percentage change

419.7 456.9 108.9

48300.1 55253.3 114.4

209.8 481.0 229.2

24150.1 59501.1 246.4

419.7 861.7 205.3

48300.1 147305.7 305.1

Perch With cormorants No cormorants Percentage change

230.8 1299.1 562.9

14967.4 48834.4 326.3

253.6 928.2 365.9

14809.5 57588.4 388.9

249.6 715.6 286.7

13200.2 53146.3 402.0

This was expected, since these were strong year classes of roach in the population and so would have comprised a high proportion of the fish population available for angler exploitation (Fig. 2.9). However, key factor analysis showed the 1995 year class of roach was reduced in number by 75% (Table 2.2) following three consecutive winters of cormorant depredation, while the 1996 year class was reduced by 65% (Table 2.3).

02 11/02/2003 12:37 Page 22

22

Impact of birds on fisheries With cormorant depredation Without cormorant depredation

With cormorant depredation Without cormorant depredation

500 Biomass (kg ha⫺1)

No. fish per hectare

600

400 300 200 100 0

2

3

4

5

6

7

8

9 10 11 12

35 30 25 20 15 10 5 0

2

3

4

5

6

7

8

9 10 11 1

Age class

Age class

500 450 400 350 300 250 200 150 100 50 0

With cormorant depredation Without cormorant depredation

With cormorant depredation Without cormorant depredation Biomass (kg ha⫺1)

No. fish per hectare

Figure 2.5 Fish density and biomass differences for the age structure of the roach population of Holme Pierrepont Rowing Course, with and without cormorant depredation, after three consecutive winters of cormorant depredation

2

3

4

5

6

7

8

9 10

140 120 100 80 60 40 20 0

2

3

4

5

Age class

6

7

8

9

10

Age class

Figure 2.6 Fish density and biomass differences for the age structure of the common bream population of Holme Pierrepont Rowing Course, with and without cormorant depredation, after three consecutive winters of cormorant depredation

With cormorant depredation Without cormorant depredation

500 400 300 200 100 0

With cormorant depredation Without cormorant depredation

60 Biomass (kg ha⫺1)

No. fish per hectare

600

2

3

4

5

6

Age class

7

8

9

50 40 30 20 10 0

2

3

4

5

6

7

8

9

Age class

Figure 2.7 Fish density and biomass differences for the age structure of the perch population of Holme Pierrepont Rowing Course, with and without cormorant depredation, after three consecutive winters of cormorant depredation

02 11/02/2003 12:37 Page 23

Key factor analysis to assess cormorant depredation

23

35 1996 year-class 1995 year-class 25 20

1997 year-class

15 10 5

290

270

250

230

210

190

170

150

130

110

90

70

50

10

0

n  216 30

Percentage frequency

30

Length (mm)

Figure 2.8 Petersen length frequency histogram of roach in angler catches during August 1998, demonstrating the dominance of the 1995 and 1996 year classes

Figure 2.9

Year class strength of roach in Holme Pierrepont Rowing Course

Table 2.2 The number of roach from the 1995 year class present in Holme Pierrepont Rowing Course with and without cormorant depredation Number per hectare Study year 1996 1997 1998

Age of 1995 year class 2 3 4

With cormorants 360 154 35

Without cormorants 486 225 141

Percentage reduction 26 32 75

02 11/02/2003 12:37 Page 24

24

Impact of birds on fisheries

Table 2.3 The number of roach from the 1996 year class present in Holme Pierrepont Rowing Course with and without cormorant depredation Number per hectare Study year

Age of 1996 year class

1997 1998

2 3

With cormorants 598 122

Without cormorants

Percentage reduction

774 350

23 65

2.4 Discussion Key factor analysis demonstrated that cormorants removed a high proportion of the fish standing crop from Holme Pierrepont Rowing Course during each winter period. The cumulative effect of this high level of depredation over three over-wintering periods was depressed levels of fish abundance in the Rowing Course due to cormorants (Table 2.1, Figs 2.5–2.7). Although the majority of the fish ingested were below 100 mm (Fig. 2.3) and anglers generally exploit fish above 100 mm in the Rowing Course (Fig. 2.8), this does not imply minimal cormorant depredation impact. A proportion of the depredated fish would have survived and grown, and may have been available for angler exploitation in future years. An example of this is shown by losses attributable to cormorant depredation in the 1995 and 1996 year classes of roach (Tables 2.2 and 2.3), fish which were targeted by anglers in August 1998 (Fig. 2.8). Key factor analysis indicated a substantial reduced abundance in these year classes of roach due to cormorant depredation. It is believed that this will have depressed the angler catch rates on the Rowing Course by limiting the number of roach available for exploitation. This is based on the premise that in stillwater catch and release fisheries, there is a strong correlation between the level of angler catch rate and the biomass of fish available for angler exploitation (North 2002). Key factor analysis not only indicated the proportion of fish that would have been available had cormorants not been present, but also an adjustment of the population structure of the target species. If cormorants had not been present, it is probable that greater numbers of roach, common bream and perch would have lived longer (Figs 2.5–2.7). Consequently, the fishery would have supported a greater number of large, older specimen fish in subsequent years, which may have improved catch rates and the overall angling experience. In 1998, recruitment from the 1995 and 1996 year classes in the fishery was strong, as expected for the climatic conditions in the first year of life of these fish (Mills & Mann 1985). Thus, the strong recruitment appeared to have been adequate to compensate the effect of the cormorant depredation (Britton et al. 2002). This resulted in sufficient fish growing beyond the optimal size for cormorant depredation in a short time, generally after 1 year (Britton et al. 2002), and ensured an adequate number of angler exploitable fish in the Rowing Course providing acceptable angling returns when

02 11/02/2003 12:37 Page 25

Key factor analysis to assess cormorant depredation

25

compared with river fisheries (Britton 1999). However, in seasons following weak recruitment of angler target species and heavy cormorant depredation, there may be too few survivors to ensure satisfactory populations of angler exploitable fish, resulting in further depressions of catch rates. Key factor analysis for determining fish losses due to cormorant depredation were based on quantitative estimates of fish population size using gear-calibrated electric fishing surveys (mean fishing efficiency of 0.23; Feltham et al. 1999). However, gear efficiency varied between 0.10 and 0.36, resulting in a range of impact values (Feltham et al. 1999). This was particularly true for common bream that has a lower susceptibility to electric fishing than roach and perch (Zalewski & Cowx 1990). Thus, the mean estimates of depredation impact on common bream may be overestimated. Notwithstanding this, the electric fishing assessment was considered to be a reasonable estimate of the fish stock abundance. This was confirmed by hydroacoustic data which provided a minimum estimate of abundance in the Rowing Course in August 1997 of 16.96 kg ha1 (range 10.79 to 25.16 kg ha1) (Lyons 1997), with cormorants removing an estimated 13 to 34 kg ha1. When compared with cormorant depredation rates from the MCS output, the hydroacoustic assessment, which provided lower fish abundance estimates than electric fishing, showed cormorant depredation removed an extremely Table 2.4

Summary of cormorant impact assessment at Holme Pierrepont Rowing Course

Measure of impact

Cormorant impact

Roach standing crop

Standing crop reduced by 62% in number and 71% by biomass after three winters of depredation. Standing crop reduced by 52% in number and 67% by biomass after three winters of depredation. Standing crop reduced by 65% in number and 75% by biomass after three winters of depredation. Roach and common bream populations, showed a shift in life history strategy from slow growing individuals with a long life span to fast growing individuals with a short life span (Britton et al. 2002). This was probably related to decreased inter- and intraspecific competition in the fish populations due to the decreased standing crop through cormorant depredation. Prior to the presence of cormorants on the Rowing Course, angler catch per unit effort was in excess of 1000 g man-h1. During the period when cormorants were known to over-winter on the Rowing Course, this declined to 50 g man-h1. There is a strong correlation between angler catch rates and fish standing crop (North 2002). In the early 1990s, prior to cormorants foraging on the Rowing Course, income from the fishery was in excess of £28 000 p.a. In 1994 and 1995, the combined income from the fishery was less than £150. The decrease in angler activity on the lake was associated with poor results in the World Angling Championships during 1994. It was believed this was due, in part, to the fish removed by cormorant depredation in winter 1993/1994.

Common bream standing crop Perch standing crop Life history strategy

Angler catch rate

Income from fishery

02 11/02/2003 12:37 Page 26

26

Impact of birds on fisheries

high proportion of fish in the lake (Feltham et al. 1999; Britton 1999), and suggests that the electric fishing studies provided a more accurate reflection of the fish stocks. The losses of fish owing to cormorants were likely to impact on other facets of the fishery and fish population dynamics. The analysis of the population dynamics suggested compensation mechanisms were capable of ensuring long-term sustainability of the fish populations by allowing the fish to grow through the size range where they are most vulnerable to predation more quickly (Britton et al. 2002). However, the advantages gained through this compensatory mechanism appear to be overridden when all the measures of cormorant impact are assessed in unison on the Rowing Course (Table 2.4). Key factor analysis enabled full integration of fish population and cormorant depredation data to provide damage estimates in a more holistic way than previously carried out. Damage estimates, used in conjunction with other measures of cormorant impact assessment, deliver a detailed insight into the impact of cormorants on specific inland fisheries, enabling future policy decisions to be based on more substantive evidence.

Acknowledgements The authors wish to thank the English Sports Council for permission to use Holme Pierrepont Rowing Course as a study site, and the staff of the National Water Sports Centre for their valuable assistance. Thanks are also extended to the Environment Agency for granting permission to undertake electric fishing surveys. The work was funded by MAFF, project contract number VC106.

References Bagenal T.B. & Tesch F.W. (1978) Age and growth. In T.B. Bagenal (ed.) Methods for the Assessment of Fish Production in Fresh Waters. IBP Handbook No. 3 (3rd edn). Oxford: Blackwell Scientific Publications, pp. 101–136. Britton J.R. (1999) The impact of cormorants (Phalacrocorax carbo carbo (L.) and Phalacrocorax carbo sinensis (Blumenbach)) on inland fisheries in the UK. PhD thesis, University of Hull, 377 pp. Britton R., Harvey J., Cowx I.G., Feltham M.J., Davies J.M., Wilson B.R. & Holden T. (2002) Key factor analysis to assess cormorant depredation on inland fisheries in the UK. In I.G. Cowx (ed.) Management and Ecology or Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 170–183. Cowx I.G., Pitts C.S., Smith K.L., Hayward P.J. & van Breukelen S.W.F. (1995) Factors Affecting Coarse Fish Populations in Lowland Rivers. National Rivers Authority R&D Report 0429/6/N&Y. Bristol: National Rivers Authority, 118 pp. Feltham M.J. & Davies J.M. (1995) Diet of cormorants, Phalacrocorax carbo L., from two fisheries in north-west England. Fisheries Management and Ecology 2, 157–159. Feltham M.J., Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999) Case Studies of the Impact of Fish-eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, contract number VC0106. London: MAAF, 207 pp. Harvey J.P. & Cowx I.G. (1995) Electric Fishing in Large Deep Waters: User Manual. National Rivers Authority R&D Report 0034/7/ST. Bristol: National Rivers Authority, 46 pp.

02 11/02/2003 12:37 Page 27

Key factor analysis to assess cormorant depredation

27

Kirby J.S., Gilburn A.S. & Sellers R.M. (1995) Status, distribution and habitat use by cormorants Phalacrocorax carbo wintering in Great Britain. Ardea 83, 93–102. Kirby J.S., Holmes J.S. & Sellers R.M. (1996) Cormorants Phalacrocorax carbo as fish predators: An appraisal of their conservation and management in Great Britain. Biological Conservation 75, 191–199. Lyons J. (1997) A Hydroacoustic Fisheries Survey on Holme Pierrepont Rowing Lake. Environment Agency, Midlands Region, Nottingham. MacDonald R.A. (1987) Cormorants and fisheries. BTO News 150, 12 pp. Mills C.A. & Mann R.H.K. (1985) Environmentally induced fluctuations in year class strength and their implications for management. Journal of Fish Biology 27, 209–226. North R. (2002) Factors affecting the performance of stillwater coarse fisheries in England and Wales. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 284–298. Ricker W.E. (1975) Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of Canada 191, 382 pp. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project number VC0104. London: MAAF, 130 pp. Wernham C.V., Armitage M., Holloway S.J., Hughes B., Hughes R., Kershaw M., Madden J.R., Marchant J.H., Peach W.J. & Rehfisch M.M. (1999) Population, Distribution, Movements and Survival of Fish-Eating Birds in Great Britain. Report to Ministry of Agriculture, Fisheries and Food. London: MAAF, 360 pp. Zalewski M. & Cowx I.G. (1990) Factors affecting the efficiency of electric fishing. In I.G. Cowx & P. Lamarque (eds) Fishing with Electricity. Oxford: Fishing News Books, Blackwell Scientific Publications, pp. 89–111.

03 11/02/2003 12:54 Page 28

Chapter 3

The relationship between cormorant and fish populations at two fisheries in England: an overview J.M. DAVIES* School of Biosciences, University of Birmingham, Birmingham, UK

T. HOLDEN, M.J. FELTHAM, B.R. WILSON School of Biological and Earth Sciences, Liverpool JMU, Liverpool, UK

J.R. BRITTON, J.P. HARVEY and I.G. COWX University of Hull, International Fisheries Institute, Hull, UK

Abstract This chapter examines the extent to which wintering cormorants, Phalacrocorax carbo (L.), are responsible for changes in fish populations, relative to other factors, at two fisheries in England. The conclusion is that the results, like many others that have estimated the impact of fish-eating birds, are equivocal. The reasons for this are: (i) the complex nature of these systems; (ii) difficulties in relating cormorant depredation to changes in fishery performance; and (iii) separating the effects of piscivorous birds from other factors affecting fish populations. The chapter concludes with recommendations for the way in which investigations examining the relationships between fish-eating birds and fisheries are made, and stresses the importance of adopting a holistic approach to such studies. Keywords: cormorants, depredation, fishery performance, impact assessment.

3.1 Introduction Many studies have examined the effect of cormorants at freshwater fisheries but most have been unable to quantify equivocally the contribution of cormorants to observed changes in fish populations (Davies 1997). This is due principally to the complex nature of most fisheries, particularly those with natural productivity, where a multitude of factors may affect the structure of fish communities (Cowx 2001). Even those studies that have looked at relatively ‘simple’ systems, such as single species fisheries, have had to make a number of assumptions when generating estimates of impact (Pawson 1986; Moerbeek et al. 1987; Draulans 1988). The spatial distribution, abundance, sizedistribution, recruitment and growth of fish populations and communities are regulated by many natural biotic and abiotic factors, of which bird depredation is but one. There

*Correspondence: Dr John Davies, School of Biosciences, The University of Birmingham, Edgbaston, Birmingham B15 2TT, UK (email: [email protected]).

03 11/02/2003 12:54 Page 29

Relationship between cormorant and fish populations

29

is, therefore, no single factor that is directly responsible for constraining the development of fish populations; rather there is a complex interaction between many factors (Engström 2001; Dekker & de Leeuw, Chapter 1). The magnitude and diversity of the biotic and abiotic factors that can affect fisheries is great. In addition, the effects that human activities have had on the freshwater environment and how they too might influence fish populations must be considered (Cowx 2001; Cowx, Chapter 28). While it seems inconceivable to many fisheries managers and anglers that cormorants and goosanders do not create problems at fisheries, conservationists are equally convinced that there is insufficient evidence in most cases to support claims that they do. Attempts to tease out the impact of fish-eating birds on fish populations from the complex of other factors known to affect fisheries is essential, but studies have tended to concentrate solely on the bird–fish relationship. The first part of this paper attempts to identify the key factors required to examine the relationship between cormorant and fish populations and the relative importance of these factors to the fishery in question. The second part presents two case studies of the impact of cormorants, drawn from a wider project which examined fish-eating birds at a range of freshwater fisheries in the UK (Feltham et al. 1999). The two case studies represent relatively ‘simple’ and ‘complex’ systems, based on the key factors outlined below. The former, in south-east England, comprises a small, singlespecies fishery at which the impact of wintering cormorants was assessed. The study took the form of an experiment, which included a control enabling natural fish mortality to be incorporated into estimates of impact. The latter was based on a multi-species, stillwater fishery in Nottingham, England which examined the relationship between cormorant and fish populations over three successive winters. These examples illustrate: (i) some of the difficulties involved with making such assessments; (ii) how a holistic approach should improve the confidence with which the impact of cormorants on fisheries can be estimated; and (iii) provide an indication of how such studies should proceed in future.

3.2 Key factors 3.2.1

Water quality

Biotic and abiotic factors play a crucial role in both the structuring of fish communities and the feeding behaviour of cormorants (Alabaster & Lloyd 1980; Jobling 1995; Cowx 2001). Water quality, habitat features and anthropogenic effects all influence the population and community structure within a fishery. Reduction of organic pollutants, such as the amount of suspended solids or the levels of dissolved oxygen levels in the water, will reduce the turbidity of the water and provide conditions more favourable for certain fish species. It has been demonstrated that water quality changes can shift the species of fish caught by anglers (Cowx 1991). On the River Trent in Nottingham, for example, water quality improvements caused a change from a predominantly roach Rutilus rutilus (L.) and dace Leuciscus leuciscus (L.) fishery to a chub Leuciscus cephalus (L.) and common bream Abramis brama (L.) fishery and an overall increase in fish species diversity (Cowx 1991). Changes in water quality may also alter the feeding behaviour of cormorants. Voslamber & van Eerden (1991), for example, observed such social fishing in the

03 11/02/2003 12:54 Page 30

30

Impact of birds on fisheries

IJsselmeer, The Netherlands, and stated that this habit only began to occur regularly in 1972. It is thought that the increasing turbidity of the water, caused by eutrophication, may have led to birds switching techniques. This was thought to improve the catchability of fish in two ways; by herding fish, hence concentrating them in a small area, and/or by pushing fish from deeper water towards the surface, making prey detection easier (Voslamber & van Eerden 1991).

3.2.2

Distribution of birds and fish

The feeding behaviour of cormorants in relation to distribution of fish populations has been studied at many types of freshwater fisheries (Richner 1991, 1995; Davies & Feltham 1996; Gremillet 1997). Information on diving behaviour and foraging success, assessed by observing feeding cormorants, highlights the relative profitability of foraging areas. If these data are collected in conjunction with information on the relative abundance of fish species, further insights into the relationship between fish populations and cormorant foraging strategies may be gained. Habitat features are also a key abiotic factor that will affect the relationship between cormorant and fish populations. Such features may act as refuges from potential predators and should be considered when examining these relationships. Preliminary studies investigating the use of refuges have demonstrated that their presence can affect the foraging behaviour of cormorants (Russell et al., Chapter 19). This can provide a greater understanding of the reasons for temporal and spatial variation in foraging success and explanations for variation in prey choice (Davies 1997). Ultimately, synthesising knowledge of fish behaviour with that of fisheating birds and the habitat features of the fishery will increase the understanding of the nature of this relationship within the context of estimating impacts.

3.2.3

Cormorant food consumption

The term consumption, i.e. the number or mass of fish removed by cormorants from a fishery, is sometimes used synonymously with impact on the grounds that any loss to a fishery is an impact. In certain circumstances, such as a put-and-take trout fishery, this may be true. At most fisheries, however, this is not strictly the case and consumption should be placed in perspective by expressing it relative to some estimate of fish stocks on the basis that it is the proportion of fish stock removed (effectively relative loss) that is most important (Marquiss & Carss 1994). It is intuitive, for example, that 100 kg of fish removed by birds from a fishery that only contained 150 kg is a greater ‘impact’ than the same removed from a fishery containing 1000 kg of fish, even though the absolute loss remains the same.

3.2.4

Changes in fishery performance

It can be further argued that consumption expressed as a proportion of biomass (as discussed above) is still not a true estimate of impact. Instead, it can be said that impact is not about the amount of fish remove by birds from a fishery per se but crucially what

03 11/02/2003 12:54 Page 31

Relationship between cormorant and fish populations

31

effect, if any, the removal of fish by birds has on that fishery. In this context this means having a pronounced effect on a fishery in terms of, for example, declining fish stocks, declining catches or adverse changes in fish community structure. Impact is, therefore, about detectable change. Most studies examine consumption by cormorants (kg of fish) in relation to an estimate of fish standing crop (kg of fish) (e.g. Dirkesen et al. 1991). Estimates of the proportion of fish biomass removed (%) can, therefore, be generated. If these are examined over time the implicit question ‘Does depredation by birds result in detectable reductions in fishery performance as indicated by declining stocks?’ can, in theory, be addressed. Collection of such biomass data, particularly at relatively large fisheries is notoriously difficult (Marquiss & Carss 1994; Feltham et al. 1999; Wilson et al. in press; Wilson et al., Chapter 9) and variation around these data often make it difficult to attribute any changes in fish biomass directly to cormorants (Davies et al. in press). Other methods of detecting the effect of cormorants on fish populations, rather than changes in biomass per se should, therefore, be considered. Changes in fish growth rates are, for example, an important tool in the assessment of fish population dynamics (Bagenal 1978). This measure enables periods of good and poor growth to be identified (Mann 1973), and growth to be compared against a national standard for a particular fish species (Hickley & Dexter 1979). If such measures can be linked to patterns of cormorant predation, they may offer a powerful tool for assessing the contribution of such birds to changes in fishery performance.

3.3 Levels of complexity 3.3.1

‘Simple’ system – Rye Meads case study

The Rye Meads lagoon complex (NGR TL3810), situated in Hertfordshire, south-east England, comprises nine shallow lagoons ranging in size from 0.6 to 4.2 ha. Two lagoons at the site were previously the focus of a study on cormorant impact (Pilcher & Feltham 1997), and these sites were again studied in 1997 (Feltham et al. 1999, McKay et al. 1999). Each lagoon was approximately 70 m  40 m (0.3 ha) with an average depth of approximately 1.5 m. One lagoon was entirely netted from the ground using 2.5 cm black nylon netting set approximately 1.5 m above water level (control lagoon) while the other lagoon was left unnetted (test lagoon). Each lagoon was stocked with approximately 400 common carp Cyprinus carpio L., equivalent to a stocking density of 140 kg ha1 (McKay et al. 1999). The experiment ran for 40 days, during which time cormorant occupancy was observed for 28 days (70% of the time). In this case study, only two of the four key factors are of importance, namely cormorant consumption and fishery performance. Biotic and abiotic factors such as water quality and habitat features were largely irrelevant since the lagoons did not undergo any dramatic changes during the study period. The lagoons were uniform in structure, so habitat features were also unimportant. After the 40-day experimental period, the biomass of fish unaccounted for in the test lagoon (i.e. fish potentially consumed by cormorants after taking into account mortality in the control lagoon) was determined. Using the mean weight of fish in each size class,

03 11/02/2003 12:54 Page 32

32

Impact of birds on fisheries

the total biomass of unaccounted fish from the test lagoon was 18.32 kg, or 57.5% of original stock. The total mass of fish observed to have been removed by cormorants over the 40-day period was 2115 g, equivalent to 76 g fish d1. This was equivalent to between 15.2 and 25.3% of daily food intake based on a theoretical range of 300–500 g d1 (Feltham & Davies 1996). Using these data, the required occupancy to account for the ‘missing’ fish from the test lagoon would be 5.4 birds d1 and 5.9 birds d1 respectively, around five times that observed during the experiment. Depredation estimates based on the observed occupancy and a theoretical daily food intake range could not, therefore account for fish ‘missing’ at the end of the experiment. The possible reasons for this are discussed by Pilcher & Feltham (1997), Feltham et al. (1999), and McKay et al. (1999). These include nocturnal depredation by herons, Ardea cinerea L., or cormorants swallowing fish underwater. This case study serves to illustrate that, with only two of the four key factors in operation, assumptions are still required to account for the ‘missing’ fish.

3.3.2

‘Complex’ system – Holme Pierrepont Rowing Course

Holme Pierrepont Rowing Course (SK 6239) is situated at the National Water Sports Centre, Nottingham and is managed by Sport England. The Rowing Course is adjacent to the River Trent, was constructed from disused gravel workings and flooded from the River Trent to form a lake 2215 m long and 135 m wide, with a total water area of 28.68 ha. Anecdotal evidence suggests cormorants were first observed feeding on the lake in winter 1993/1994 (B. Pluckrose, personal communication). The resident fish stocks of Holme Pierrepont Rowing Course comprise coarse fish species, with the dominant species captured by anglers being roach and common bream, although perch Perca fluviatilis L. has been an increasing component of catches in recent years (Feltham et al. 1999). Other species found in the lake include chub, dace, eel, Anguilla anguilla (L.), gudgeon, Gobio gobio (L.), pike, Esox lucius L., tench, Tinca tinca (L.), rudd, Scardinius erythropthalmus (L.), carp, and bleak, Alburnus alburnus (L.). This system could be described as relatively ‘complex’, with all four of the key factors identified important in determining the relationship between cormorant and fish populations.

Water quality The primary water input into Holme Pierrepont Rowing Course is from Adbolton Brook, an effluent drainage carrier stream. Supplementary water from the River Trent enters the lake through a sluice located on the north bank and inflow rate is increased manually during poor water quality conditions and periods when blooms of blue–green algae persist (T. Holden, unpublished data). Effluent discharges entering the Rowing Course via Adbolton Brook created a habitat synonymous with turbid water conditions, silty benthos and sparse macrophyte growth through the 1980s and early 1990s. Reduced inputs of organic material and suspended solids from Adbolton Brook and the River Trent during the 1990s gradually improved water clarity, lowered siltation of the gravel substrate and increased macrophyte establishment. The habitat now supports abundant populations of benthic macroinvertebrates, such as Gammarus

03 11/02/2003 12:54 Page 33

Relationship between cormorant and fish populations 3.6

Suspended Solids BOD

30

3.2

20

2.8

10

2.4

0 1992

1993

1994

1995

1996

1997

BOD (mg litre⫺1)

Suspended solids (mg litre⫺1)

40

33

2 1998

Year

Figure 3.1 Mean Suspended Solid (mg L1) and Biological Oxygen Demand (BOD) (mg L1) measures from the River Trent adjacent to Holme Pierrepont Rowing Course

spp. Changes in water quality at the Rowing Course will, in part, be reflected by those observed on the nearby River Trent, as inputs are received from this watercourse. The biological quality of the river, measured as the Biological Monitoring Working Party score (BMWP) on the River Trent increased from 1980. Measures of organic pollution, such as biological oxygen demand (BOD) and suspended solids on the stretch adjacent to the Rowing Course also improved (Fig. 3.1). The levels of suspended solids in the lower River Trent have declined markedly since the 1970s. Alabaster & Lloyd (1980) reported that the upper and lower 95% confidence limits for concentration of suspended solids were 62 mg L1 and 695 mg L1 respectively. Comparable figures for the period during which Holme Pierrepont Rowing Course was studied were 10–20 mg L1 (Environment Agency, unpublished data).

Distribution of birds and fish Although highly uniform in structure, the Rowing Course has features and structures integral to the management of the facility and support of watersports activities. A number of jetties and pontoons are situated at the western end of the lake and are used as launching facilities for rowers, canoeists and wind-surfers. Sonar scans conducted by the Environment Agency demonstrated a general migration of fish to the western end of the lake during the winter months (Feltham et al. 1999), and any assessment of the relationship between cormorant and fish populations should acknowledge the possible influence of these features. Observations of feeding cormorants highlighted a trend for birds to visit one or more of the habitat features connected with the Rowing Course. The distribution and success of cormorant foraging activity in relation to the presence of habitat features was assessed by observations on focal birds in the winters of 1996/1997 and 1997/1998. A dive bout was scored to include a habitat feature when a focal bird was seen to dive within 5 m of the feature. Conversely, a dive-bout was scored as being an ‘open’ habitat

03 11/02/2003 12:54 Page 34

34

Impact of birds on fisheries

Table 3.1 Comparison of foraging activity and foraging success of cormorants at Holme Pierrepont Rowing Course. Data are for winter 1996/1997 and 1997/1998 combined. Sample sizes refer to number of foraging bouts

Proportion (%) dive-bouts Mean (SE) proportion (%) successful dive-bouts Mean (SE) proportion (%) successful dives Mean (SE) intake rate (g fish min1)

Features (n  265)

Open habitat (n  137)

66 78  0.1 16.6  0.3 26  2.5

34 20  0.5 1.5  0.1 4  2.2

when the bird remained at least 5 m from a feature during the course of feeding activity. Combining data for both winters, cormorants were observed significantly more frequently at habitat features than in open habitat (21  40.76, P  0.001). More than 265 (66%) of cormorant dive-bouts were undertaken about features compared with only 137 (34%) in open habitat (Table 3.1). This preference for feeding in specific regions of the fishery suggests that foraging conditions varied spatially, and implies that those areas of the lake containing habitat features offered more profitable foraging than open habitats. This was confirmed by measures of foraging success on the Rowing Course. On average, 78  0.1% of cormorants dive-bouts about features resulted in the capture of prey, compared with only 20  0.5% in open habitat (t15  7.60, P  0.001) (Table 3.1). The proportion of successful dives (a dive which resulted in a prey item being caught, brought to the surface and swallowed) was greater about habitat features compared with open water. Cormorants caught fish 11 times more frequently at the former (t363  12.00, P  0.001). Similarly, intake rates (g fish min1) were greater for those birds foraging about habitat features with birds consuming 26  2.5 g fish min1 about features compared with only 4  2.2 g min1 in open water (t243  12.82, P  0.001) (Table 3.1).

Cormorant consumption The biomass of fish removed by cormorants over the three winters of the study was estimated using a Monte Carlo Simulation (MCS) model (Feltham et al. 1999; Davies et al. in press). This model incorporated data on the temporal variation in cormorant foraging dynamics and abundance at the site together with associated confidence limits (Davies et al. in press, Wilson et al. in press). MCS estimates of the mass of fish consumed by cormorants increased from 26 to 50 kg ha1 between 1995/1996 to 1997/1998 (Table 3.2). As stated earlier, consumption figures should be placed in perspective, by expressing them relative to an estimate of fish stock.

Changes in fishery performance The quantitative assessment of fish stocks in large freshwater bodies is extremely difficult (Cowx 1996), so a variety of methods were adopted to overcome these restrictions. At the Rowing Course, surveys were conducted using boat-mounted electric fishing. This equipment was constructed specifically for sampling large rivers and lakes (Harvey & Cowx 1996). The methods adopted to assess the status of the fish populations in the pres-

03 11/02/2003 12:54 Page 35

Relationship between cormorant and fish populations

35

Table 3.2 Monte Carlo Simulation (MCS) model estimates of the annual variation in the mass of fish removed by cormorants (kg ha1), the standing crop biomass of fish (kg ha1) and the proportion of standing crop biomass removed by wintering cormorants at Holme Pierrepont, 1995–1998 Mass of fish removed by cormorants (kg ha1) Year 1995/1996 1996/1997 1997/1998

Median 26 32 50

Interquartile range 22–30 27–37 43–58

Standing crop biomass (kg ha1)

Standing crop biomass removed by cormorants (%)

Interquartile Median range Median 65 65 99

7–147 3–115 11–240

42 48 55

Interquartile range 19–82 21–91 32–99

ent study were based on gear calibration and provided measures of abundance with associated confidence limits (Bayley 1985; Cowx 1996). This approach was deemed acceptable because it provides an estimate of stock size and structure against which bird depredation can be measured. The methods used to generate fisheries data and the specific sampling regimes employed are detailed elsewhere (Feltham et al. 1999). Cormorants removed, on average, between 42% and 55% of the biomass in the lake as estimated directly from electric fishing surveys and varied significantly between years (H2  33.48, P  0.001) (Table 3.2). There were, however, large variances in biomass estimates so that the interquartile ranges of consumption expressed as a proportion of these biomasses were also large (Table 3.2). In winter 1996/1997, for example, the median loss was estimated to be 48%, with lower and upper quartiles of 21% and 91% respectively. Despite these apparently high levels of depredation, as a proportion of standing crop, over the three winters of the study at Holme Pierrepont, no adverse affect on standing stock could be detected. The estimated biomass of fish removed by cormorants increased over the 3 years of the study, but no reduction in estimated fish biomass was detected during that time. This was presumably because the populations compensated for depredation by increasing growth and contribution of larger fish that had bypassed the most vulnerable predation size (50–120 mm fork length; Feltham et al. 1999). If, however, individual years are considered separately, marked declines in standing stock of the main fish species were evident (Britton et al., Chapter 2). Information on fish population structure was also available at Holme Pierrepont Rowing Course, and the methods by which these data were collected and derived are detailed elsewhere (Britton 1999; Feltham et al. 1999; Britton et al. 2002; Britton et al., Chapter 2). One such measure – relative annual growth rate – indicated that a marked change in growth of two cyprinid species, roach and common bream, had occurred at the fishery. The relative annual growth of roach was poor before 1994, compared with more recent growth from this water body (Fig. 3.2). Between 1992 and 1995, the growth rate increased by 90% and the shift was first detected in the 1994 growth year (Fig. 3.2). The historical growth rate of common bream was also poor when compared with recent data (Fig. 3.3). Between 1983 and 1990 growth was below the national standard for common bream (Britton et al. 2002). A marked increase in growth first occurred in 1993 and has continued to increase with some annual variation. In subsequent winters, the growth rate

03 11/02/2003 12:54 Page 36

36

Impact of birds on fisheries

Percenatge difference from average relative annual growth

80

40

20

0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 ⫺20

Growth year

Relative annual growth rate of Holme Pierrepont roach 1988–1997

80 Percentage difference from relative average anuual growth

Period of cormorant predation

60

⫺40

Figure 3.2

Prior to cormorant predation

Prior to cormorant predation

Period of cormorant predation

60

40

20

0 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997

⫺20

⫺40 Growth year ⫺60

Figure 3.3

Relative annual growth rate of Holme Pierrepont common bream 1983–1997

03 11/02/2003 12:54 Page 37

Relationship between cormorant and fish populations

37

exceeded the national standard for common bream, by some 100% (Britton et al. 2002). A dramatic shift in growth rate, therefore, occurred in the roach and common bream populations in the period 1993–1998. Anecdotal evidence suggests cormorants first fed on the lake in winter 1993/1994 (B. Pluckrose, personal communication), around the time of the growth rate shift. Roach and common bream were two of the major fish species depredated by cormorants at Holme Pierrepont Rowing Course during the three winters of the study (Feltham et al. 1999; Britton et al. 2002; Davies et al. in press).

3.4 Discussion Many studies have attempted to identify the precise affect of cormorants at freshwater fisheries (Kennedy & Greer 1988; Suter 1995; Warke & Day 1995; Davies 1997). The majority reached ambiguous conclusions, or had to make certain assumptions to provide quantitative assessments (Davies 1997; Feltham et al. 1999). The two case studies outlined in this chapter highlight some of these difficulties, but also attempt to provide a more holistic view of the relationship between cormorant and fish populations.

3.4.1

Rye Meads

In the first case study, at Rye Meads lagoons, only two of the four key factors were in operation. The experiment ran for only 40 days, with water quality measures constant. Absence of habitat features and the presence of only a single fish species reduced the importance of foraging strategies of birds. Natural mortality could be accounted for (approximately 4% of original stock) enabling mortality due to predation to be calculated. Despite this, assumptions still had to be made regarding cormorant predation.

3.4.2

Holme Pierrepont Rowing Course

The case study at Holme Pierrepont Rowing Course is more representative of the interactions between cormorants and fish communities than that at Rye Meads, particularly at UK fisheries. Undertaken at a multi-species fishery with natural productivity that has experienced changes in water quality, this case study demonstrated some of the generic difficulties encountered when examining these interactions. Holme Pierrepont is a fishery that has experienced large, natural fluctuations in its fish populations during the period 1992–2002 (Feltham et al. 1999; Britton et al. 2002). These fluctuations have, however, only been influenced by cormorants relatively recently.

Water quality Changes in water quality would have had profound affects on fish community structure and the distribution of fish within the lake. The precise changes that occurred were not well documented in the time before the study period. Although roach dominated electric fishing catches over the three winters of the study, the proportion of perch caught increased (Feltham et al. 1999). This was either a response to the oligotrophi-

03 11/02/2003 12:54 Page 38

38

Impact of birds on fisheries

cation processes taking place in the lake (Cowx 2002) or recovery from the disease which severely degraded perch stocks through the 1980s. Changes in community structure may not be the only important effect of changes in water conditions. The clarity of water, especially in the marginal areas, led fish to move into deeper central areas during times of clear water, exhibiting a natural avoidance to possible capture (Feltham et al. 1999). This phenomenon was observed during electric fishing surveys, when catches in the margins were low at times of clear water but were much higher when water was turbid. The possible movement of fish away from the marginal areas to deeper areas may have serious implications on angler catches. Anglers on the Rowing Course are restricted to fishing within 20 m of the bank because of the presence of underwater wires which support lane buoys. This essentially means that anglers are fishing the marginal areas and in clear water conditions fish may avoid this zone, hence reducing catches.

Distribution of birds and fish The distribution of fish and cormorants within the lake showed that cormorants were exploiting prey populations at winter refugia, a trait observed in other studies of wintering birds (Voslamber et al. 1995). These findings suggest that fish behaviour may influence the distribution of cormorants foraging within habitats. Migration of fish populations to winter refuge areas at this fishery created extensive areas of low prey density, interspersed with regions of high food availability for cormorants. Exploitation of the latter areas provided cormorants with the optimum foraging conditions of high fish densities, located in shallow water with restricted escape opportunities for prey. The reasons for fish congregating around the pontoon and jetties is unclear, although it seems likely that the shadowing effect of these features offered fish some refuge during periods when water conditions were clear. Observations of cormorant feeding behaviour revealed their ability to feed underneath these structures with great success. Electric fishing surveys under the boat pontoons found dense shoals of fish with fish in 50 cm of water and in cracks in the concrete. The fish may be using the pontoons for shelter from the elements or to avoid depredation, although apparently with limited success. The use of fish refuges under the pontoons may prove useful if the design is appropriate to allow fish access but cormorants are excluded. Indeed, the depredation may be decreased dramatically during the winter period by protecting fish shoaled under the pontoons. The use of refuges as a means of reducing cormorant predation is currently under investigation (Russell et al., Chapter 19), but the study at Holme Pierrepont demonstrated the importance of such features. Foraging behaviour was markedly different in and around these features compared with open habitat, and recognition of this heterogeneity is extremely important.

Cormorant consumption and changes in fishery performance The estimated biomass of fish removed by cormorants increased over the 3 years of the study, but no reduction in the overall fish standing crop was detected during that time. It must be emphasised that this is not necessarily an indicator of no impact. The biomass of fish in Holme Pierrepont may have been higher in the absence of cormorants

03 11/02/2003 12:54 Page 39

Relationship between cormorant and fish populations

39

(Britton et al. 2002; Britton et al., Chapter 2) and compensatory mechanisms within the fish population may have offset some of the effects of cormorant depredation. Notwithstanding this, variance about consumption estimates was, in all cases, smaller than the variance in estimates about the number, mass or density of fish present in the fishery. This variance may, therefore, mean that attempting to estimate impact using standing crop biomass may not be appropriate at large waterbodies (Davies et al. in press), especially where compensatory mechanisms are prevalent. The observed changes in growth rates of roach and common bream within Holme Pierrepont may have been a reflection of low abundance/standing crop of fish and/or high food availability (Russell et al. 1996; Feltham et al. 1999). The precise role of cormorants in this growth rate shift at Holme Pierrepont remains unclear, as no cormorant feeding data were available prior to the shift in growth rate (Feltham et al. 1999; Britton et al. 2002). Increases in growth rates, interpreted as compensatory responses, have been found in commercially exploited vendace populations in a Finnish Lake (Sarvala et al. 1994). Standing crop in Holme Pierrepont was certainly relatively low at the start of the study, and cormorants were likely to be feeding on roach and common bream during the winter of 1993/1994, when the change in these species’ relative growth rate occurred. However, in 1995/1996, the first winter that diet data were available, 86% (79–91%, 95% CI (n  215 foraging bouts)) of cormorant diet by mass comprised perch (Davies et al. in press). The proportion of roach and common bream in the diet only became significant in subsequent winters. Consequently, any interpretation of these growth rate data must remain conjectural.

3.4.3

Future studies

Future studies examining the relationship between cormorant and fish populations must take a multi-factorial approach (Dekker & de Leeuw, Chapter 1; Cowx, Chapter 28). Too frequently, studies have concentrated only on the bird–fish interaction, in only the most simple terms (mass of fish removed versus mass of fish present). Any prospective study should ask the question ‘Will this work help improve our understanding of the relationship between cormorant and fish populations?’. If the answer is no, it will often be due to the lack of available data on abiotic factors, cormorant behavioural ecology, fish population dynamics and fish productivity. These are all vital components in the relationship and consideration must be given as to how they can be incorporated into models (Dekker & de Leeuw, Chapter 1). Only a small number of the possible factors that could affect the relationship between cormorants and fish populations have been addressed in this chapter. There are many other measures, particularly relating to fish community structure, that need to be incorporated before a truly holistic approach can be claimed. Other piscivores could, for example, have profound effects on fish populations. Pike, Esox lucius L., were caught regularly during electric fishing surveys and their role in shaping community structure is unknown (Feltham et al. 1999). In addition, perch are piscivorous and are known to take fish up to 8 cm in length (Tonn et al. 1992). Studies have, moreover, demonstrated that the presence of these predators may have marked effects on fish in terms of their distri-

03 11/02/2003 12:54 Page 40

40

Impact of birds on fisheries

bution within a waterbody and lead to reduced growth (He & Kitchell 1990; Tonn et al. 1992). The value of looking at the relationship between cormorant foraging behaviour and the distribution of fish at Holme Pierrepont was demonstrated. Cormorants were clearly targeting areas of high prey density and therefore profitability and such spatial variation should be accounted for when making an assessment of cormorant impact. The experimental nature of the Rye Meads case study enabled certain factors to be controlled (natural mortality, sizes and species of fish present, abiotic and biotic factors) but still required assumptions to be made regarding cormorant predation. Experiments have a vital role to play in developing our understanding of the relationships between fish-eating birds and fish populations. They may, for example, be particularly useful for improving our understanding of behavioural ecology and predator–prey relationships (Marquiss & Carss 1994; Russell et al. 1996, Stempniewicz et al., Chapter 5; T. Strod, personal communication). Such studies could, for example, help address some of the issues relating the behavioural responses of fish to piscivores (both birds and fish) (Werner et al. 1983). Careful consideration must, however, be given to how such experiments can be applied to actual fisheries. A frequent call of those coming from an ornithological perspective is the request for ‘better data’ from fisheries researchers. In many cases, owing to limitations in the technology available, this is simply not possible (Wilson et al. in press). The case study at Holme Pierrepont suggested that, by looking at changes in biomass exclusively, the extent to which changes were driven by cormorants or natural variation could not be determined, at least over the three-year study period. Had the fishery been examined for a longer time period, it may have been possible to quantify the relative importance of these two factors. Studies must aim to collect the most robust data possible, using the most appropriate methods. The assumptions underlying these data (both for birds and fish) must be made clear and recognised in any conclusions reached. Improving the link between fish population dynamics and fish consumption by cormorants is crucial (Marquiss & Carss 1994; Russell et al. 1996). The value of examining variables such as relative growth rates of fish within Holme Pierrepont cannot be underestimated. They provided a tantalising insight into the changing structure of the population in the presence of cormorant depredation. Conclusions as to the ultimate cause of the observed changes in growth rates must be treated with a certain amount of caution as the two variables can only be related directly when these variables are temporally and spatially coincident. As Dekker & de Leeuw (Chapter 1) pointed out, ‘the analysis of the complex dynamics involved in the interactions between birds and fisheries has only just begun’. The use of such data, used in conjunction with other key factors, can only help to provide a better understanding of the complexities involved in the relationship between fish-eating birds and fisheries.

References Alabaster J.S. & Lloyd R. (1980) Water Quality Criteria for Freshwater Fish. London: Butterworths, 361 pp. Bagenal T.B. (1978) Methods for the Assessment of Fish Production in Fresh Waters. IBP Handbook No. 3 (3rd edn). Oxford: Blackwell Science, 374 pp.

03 11/02/2003 12:54 Page 41

Relationship between cormorant and fish populations

41

Bayley P.B. (1985) Sampling problems in freshwater fisheries. In Proceedings of the 4th British Freshwater Fisheries Conference. University of Liverpool, pp. 3–11. Britton J.R. (1999) The impact of cormorants (Phalacrocorax carbo and Phalacrocorax carbo sinensis) on inland fisheries in the UK. PhD thesis, University of Hull, 377 pp. Britton J.R., Harvey J.P., Cowx I.G., Holden T., Feltham M.J., Wilson B.R. & Davies J.M. (2002) Compensatory responses of fish populations in a shallow eutrophic lake to heavy depredation pressure by cormorants and the implications for management. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 170–183. Cowx I.G. (1991) The use of angler catch data to examine potential fishery management problems in the lower reaches of the River Trent, England. In I.G. Cowx (ed.) Catch Effort Sampling Strategies and their Application in Freshwater Fisheries Management. Oxford: Fishing News Books, Blackwell Science, pp. 154–165. Cowx I.G. (1996) The integration of fish stock assessment into fisheries management. In I.G. Cowx (ed.) Stock Assessment in Inland Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 495–506. Cowx I.G. (2001) Factors Influencing Coarse Fish Populations in Lowland Rivers. Literature review. R&D Report. Bristol: Environment Agency, 185 pp. Cowx I.G. (2002) Principles and approaches to the management of lake and reservoir fisheries. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 170–183. Davies J.M. (1997) The impact of cormorants (Phalacrocorax carbo L.) on angling catches on the River Ribble, Lancashire. PhD thesis, Liverpool John Moores University, 278 pp. Davies J.M. & Feltham M.J. (1996) The diet of wintering cormorants Phalacrocorax carbo L. in relation to angling catches on a coarse river fishery in north-west England. In S.P.R. Greenstreet & M.L. Tasker (eds) Aquatic Predators and Their Prey. Oxford: Blackwell Science, pp. 106–110. Davies J.M., Holden T., Feltham M.J., Wilson B.R., Britton J.R., Harvey J.P. & Cowx I.G. (in press) The use of a Monte Carlo Simulation Model to estimate the impact of cormorants at an inland fishery in England. Proceedings of the 5th International Cormorant Conference. Dirksen S., Boudewijn T.J., Slager L.K. & Mes R.G. (1991) Breeding success of cormorants Phalacrocorax carbo sinensis in relation to the contamination of their feeding grounds. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the Workshop 1989 on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 232–242. Draulans D. (1988) Effects of fish-eating birds on freshwater fish stocks: An evaluation. Biological Conservation 44, 251–263. Engström H. (2001) Long term effects of cormorant predation on fish communities and fishery in a freshwater lake. Ecography 24, 127–138. Feltham M.J. & Davies J.M. (1996) The daily food requirements of fish-eating birds: Getting the sums right. In S.P.R. Greenstreet & M.L. Tasker (eds) Aquatic Predators and Their Prey. Oxford: Fishing News Books, Blackwell Science, pp. 53–57. Feltham M.J., Davies J.M., Holden T., Wilson B.R., Cowx I.G., Harvey J.P. & Britton R.J. (1999) Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to Ministry of Agriculture Fisheries and Food, contract number VC106. London: MAAF, 207 pp. Gremillet D. (1997) Catch per unit effort, foraging efficiency, and parental investment in breeding great cormorants (Phalacrocorax carbo). ICES Journal of Marine Science 54, 635–644. Harvey J.P. & Cowx I.G. (1996) Electric fishing for assessment of fish stocks in large rivers. In I.G. Cowx (ed.) Stock Assessment in Inland Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 11–26. He X. & Kitchell J.F. (1990) Direct and indirect effects of predation on a fish community: A wholelake experiment. Transactions of the American Fisheries Society 119, 825–835.

03 11/02/2003 12:54 Page 42

42

Impact of birds on fisheries

Hickley P. & Dexter K.F. (1979) A comparative index for quantifying growth in length of fish. Fisheries Management 10, 147–151. Jobling M. (1995) Environmental Biology of Fishes. Fish and Fisheries, Series 16. London: Chapman & Hall, 455 pp. Kennedy G.J.A. & Greer J.E. (1988) Predation by cormorants Phalacrocorax carbo (L.) on the salmonid populations of an Irish river. Aquaculture and Fisheries Management 19, 159–170. Mann R.H.K. (1973) Observations on the age, growth, reproduction and food of the roach Rutilus rutilus (L.) in two rivers in southern England. Journal of Fish Biology 5, 707–736. Marquiss M. & Carss D.N. (1994) Avian piscivores: Basis for policy. National Rivers Authority R&D Project 461. Bristol: National Rivers Authority, 104 pp. McKay H., Furness R., Russell I., Parrott D., Rehfisch M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. Report to the Ministry of Agriculture Fisheries and Food, project number VC0107. London: MAAF, 127 pp. Moerbeek D.J., van Dobben W.H., Osieck E.R., Boere G.C. & Bungenberg de Jong C.M. (1987) Cormorant damage prevention at a fish farm in the Netherlands. Biological Conservation 39, 23–38. Pawson M.G. (1986) Performance of rainbow trout Salmo gairdeneri in a put-and-take fishery, and influence of anglers’ behaviour on catchability. Aquaculture and Fisheries Management 17, 59–73. Pilcher M.W. & Feltham M.J. (1997) An Assessment of Cormorant Predation on Stillwater Coarse Fish Populations in the Lea and Colne Valleys of the Thames Catchment. Environment Agency (Thames North-east Area) R&D Technical Report W101, 64 pp. Richner H. (1991) The distribution and abundance of wintering cormorants Phalacrocorax carbo. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the Workshop 1989 on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 27–29. Richner (1995) Wintering Cormorants in the Ethan Estuary, Scotland: numerical and behavioural responses to fluctuating prey availability. Ardea 83, 193–198. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture Fisheries and Food, project number VC0104. London: MAAF, 130 pp. Sarvala J., Helminen H. & Hirvonen A. (1994) The effect of intensive fishing on fish populations in Lake Pyhäjärvi, south-west Finland. In I.G. Cowx (ed.) Rehabilitation of Freshwater Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 77–89. Suler W. (1995) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: two case studies from Swiss rivers. Journal of Applied Ecology 32, 29–46. Tonn W.M., Paszkowski C.A. & Holopainen I.J. (1992) Piscivory and recruitment mechanisms structuring prey populations in small lakes. Ecology 73, 951–958. Voslamber B. & van Eerden M.R. (1991) The habit of mass flock fishing by cormorants Phalacrocorax carbo sinensis at the IJsselmeer, The Netherlands. In M.R. van Eerden & M. Zijlstra (eds). Proceedings of the Workshop 1989 on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 182–191. Voslamber B., Platteeuw M. & van Eerden M.R. (1995) Solitary feeding in sand pits by breeding cormorants Phalacrocorax carbo sinensis: does specialised knowledge about fishing sites and fish behaviour pay off? Ardea 83, 213–222. Warke G.M.A. & Day K.R. (1995) Changes in abundance of cyprinid and percid prey affect rate of predation by cormorants Phalacrocorax carbo on salmon Salmo salar smolt in Northern Ireland. Ardea 83, 157–166. Werner E.E., Gilliam D.J., Hall D.J. & Mitelbach G.G. (1983) An experimental test of the effects of predation risk on habitat use in fish. Ecology 64, 1540–1548. Wilson B.R., Feltham M.J, Davies J.M., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (in press) Increasing confidence in impact estimates – the Monte Carlo approach. Proceedings of the 5th International Cormorant Conference.

04 11/02/2003 12:57 Page 43

Chapter 4

A theoretical assessment of cormorant impact on fish stocks in Great Britain M. DIAMOND*, M.W. APRAHAMIAN Environment Agency, North West Region, Warrington, UK

R. NORTH Environment Agency, National Coarse Fisheries Centre, Kidderminster, UK

Abstract A simple empirical model was developed to estimate the impact of cormorant predation on the freshwater fish stocks in Great Britain. The study used published estimates of the surface areas of waters, standing stock, fish production and cormorant consumption rates. These data were used to estimate the weight of fish caught by cormorants on a national basis in relation to the total annual production of fish. Impact was assessed in terms of annual fish production and biomass consumed as a proportion of the standing stock. Impacts were also assessed on a local scale to provide guidelines to assess whether the number of days cormorants were feeding on a particular stillwater would be having an impact. Keywords: cormorant, fish production, Phalacrocorax carbo.

4.1 Introduction The cormorant, Phalacrocorax carbo (L.), was a relatively rare bird in Great Britain and received legal protection under the EU Birds Directive (EEC/79/409) through the Wildlife and Countryside Act 1981. In Great Britain, its population size reached a peak of approximately 20 000 birds in the early 1980s (Wernham et al. 1999). This increase in the cormorant population was pan-European and seemed to coincide with a decline in fish stocks, which led to calls from anglers across Europe to cull cormorants to protect fish stocks. The word ‘cull’ is taken to mean a deliberate reduction in a population size to a predetermined target by killing eggs, young or adults. The geographical boundaries for a population to be culled would presumably coincide with the administrative boundaries of the nation states. (It should be noted that such a cull would be illegal under European Law.) Culling should be distinguished from the licensed shooting of small numbers of birds on a local basis ‘for the purposes of preventing serious damage to … fisheries’ under a derogation allowed by the Directive.

*Correspondence: Dr Mark Diamond, Environment Agency, North West Region, Richard Fairclough House, Knutsford Road, Warrington, WA4 1HG, UK (email: [email protected]).

04 11/02/2003 12:57 Page 44

44

Impact of birds on fisheries

The aim of this chapter is to use a simple empirical model to explore what the likely impacts of cormorants might be at a national and a local scale using the best available information. The model makes a number of assumptions and these are discussed.

4.2 Methods The proportion of the biomass of freshwater fish produced annually that was eaten by cormorants in Great Britain (I) was estimated by: I  (P  B)/(C  F  R)

(4.1)

where P is the production to biomass ratio for freshwater fish, B is the biomass of freshwater fish in Great Britain, C is the population size of cormorants in Great Britain, F is the proportion of the cormorant population feeding on fresh waters and R is the biomass of fish eaten annually per bird. The potential impact of cormorants feeding intensively at a local scale (e.g. on an enclosed pond) was assessed in terms of the proportion of annual fish production eaten as follows. The number of cormorant-feeding days needed to eat a fixed proportion of the annual fish production of a fixed biomass of fish per unit area (D) was estimated by: D  (X  Y  P)/r

(4.2)

where X is the fixed proportion of fish produced, Y is the fixed biomass per unit area, P is the production to biomass ratio for freshwater fish and r is the biomass of fish eaten per bird per day. D was estimated for values of X of 1%, 2%, 5%, 10% and 20% and Y was fixed at 200 kg ha1 for illustrative purposes.

4.3 Results 4.3.1

The estimation of the production of freshwater fish in Great Britain

The annual production of freshwater fish in Great Britain was estimated as the product of the estimated biomass per unit area, the estimated P/B ratio and the estimated surface area of fresh waters in Great Britain. In a study of riverine fish survey sites in England and Wales, Mainstone et al. (1994) found that 20% of the sites had a biomass greater than 323 kg ha1, 40% were greater than 181 kg ha1, 60% were greater than 85 kg ha1 and 80% of the sites were greater than 27 kg ha1 (Fig. 4.1). The average biomass of fish in rivers was about 100 kg ha1 (and that was the 44 percentile). The biomass of fish in still waters is dependent on type, classified as: natural, improved and intensive. The average for the three types of water bodies was approximately 200 kg ha1 for natural systems, 330 kg ha1 for improved still waters, and 730 kg ha1 for intensive fisheries (North 2001) (Fig. 4.2). Therefore, for the purposes of this model, mean biomasses of 100 kg ha1 and 200 kg ha1 respectively were assumed for rivers and lakes.

04 11/02/2003 12:57 Page 45

Theoretical assessment of cormorant impact

45

Image Not Available

Figure 4.1 Cumulative frequency of fish biomass (kg ha1) in rivers of England and Wales (Mainstone et al. 1994)

Image Not Available

Figure 4.2 The median and range (25–75 percentile) in biomass of fish in three types of stillwater (North 2001)

The total surface area of water in Great Britain is estimated to be 2400 km2 (Smith & Lyle 1979), of which the vast majority is in Scotland (Table 4.1), and 80% is standing waters (Table 4.1). The total biomass of fish in Great Britain was estimated to be 43 million kg (43 000 t) of fish, the majority of which – 38 000 t (90%) – is in lakes (Fig. 4.3). Production of a number of salmonid and coarse fish species (reviewed by Mann & Penczak 1986) ranged from near zero to in excess of 30 g m2 yr1, with a mean of 6.5 g m2 yr1 (Fig. 4.4). The P/B ratio for a number of species shows a skewed

04 11/02/2003 12:57 Page 46

46

Impact of birds on fisheries

Country

Lake

River

Total

Wales England Scotland Total

74 323 1528 1924

51 353 76 480

125 675 1604 2404

50000

100

45000

90

40000

80

35000

70

30000

60

25000

50

20000

40

15000

30

10000

20

5000

10

0

0 River

Figure 4.3

Percent

Biomass (103 kg)

Table 4.1 Surface area (km2) of lakes and rivers in Great Britain (Smith & Lyle 1979)

Lake

Total

Total biomass of fish in Great Britain

50 45

Mean (95% CI) = 6.47 (2.11) [n = 159]

Percentage of sites

40 35 30 25 20 15 10

30⫹

28–30

26–28

24–26

22–24

20–22

18–20

16–18

14–16

12–14

10–12

8–10

6–8

4–6

2–4

0

0–2

5

Production (g m⫺2 yr⫺1)

Figure 4.4 Annual production for salmonids and European coarse fish species (n  number of studies) (Mann & Penczak 1986)

04 11/02/2003 12:57 Page 47

Theoretical assessment of cormorant impact 30

47

Mean (95% CI) = 1.28 (0.23) [n = 82]

Percentage of sites

25 20 15 10 5 0 0.0–0.5 0.5–1.0 1.0–1.5 1.5–2.0 2.0–2.5 2.5–3.0 3.0–3.5

3.5–4.0

Production/mean biomass

Figure 4.5 Production/mean biomass ratio for salmonids and European coarse fish species (n  number of studies) (Mann & Penczak 1986) 60000

100 90 80 70

40000

60 30000

50 40

20000

Percent

Production (103 kg yr⫺1)

50000

30 20

10000

10 0 River

Figure 4.6

Lake

Total

0

Total annual production of fish in Great Britain

distribution with a mean of 1.28 and a mode of 1.25 (Fig. 4.5). If the P/B ratio is assumed to be 1.28 then the annual production of fish in Great Britain was estimated to be 55 000 t yr1 (Fig. 4.6). The 95 percentile range is 45 200–64 900 t yr1.

4.3.2

The estimation of the biomass of freshwater fish eaten annually by cormorants

The Wetland Bird survey estimated the 5-year peak mean number of cormorants in Great Britain to be 17 200 birds, between 1990/1991 and 1994/1995 (Table 4.2). The majority (51.5%) of the birds were reported from around the coast or estuaries, followed

04 11/02/2003 12:57 Page 48

48

Impact of birds on fisheries Table 4.2 Five-year peak mean numbers (1990/1991–1994/1995) of cormorants in relation to habitat (Wernham et al. 1999)

Habitat

Five-year peak mean numbers (1990/1991–1994/1995), percentage in brackets

Coastal/estuaries Reservoir Natural waters Gravel pits Rivers/freshwater marsh Total

8 845 (51.5%) 2 848 (16.6%) 2 785 (16.2%) 1 845 (10.7%) 868 (5.0%) 17 191

Table 4.3 Range in daily food intake (g d1) for cormorant races carbo and sinensis Range in daily food intake (g d1) Soure

carbo

sinensis

Feltham & Davies (1997) Russell et al. (1996) Hughes et al. (1999) Marquiss et al. (1998)

356 –881

120 –739 300 –525 248– 415

264 –587 465–775

by reservoirs, natural still waters and gravel pits, with only 5% being reported from rivers (Table 4.2). The amount eaten per day can range from 120 to 880 g d1 (Table 4.3). Using the 5-year peak number of cormorants in Great Britain, and assuming that half the population feed at inland sites and that each bird consumes 400 g d1, then the annual consumption of fish is estimated at 1.26 million kg (1200 t).

4.3.3

The proportion of the annual production of fish eaten by cormorants in Great Britain

If cormorants consume 1200 t of fish annually, and if the total annual production of fish in Great Britain is 55 000 t, then this represents an impact of 2.5% of annual production (95 percentile: 2.0–2.9%).

4.3.4

Estimating impacts on a local scale

The impact of cormorants on a local scale is illustrated for a fishery with a biomass of 200 kg ha1 and a P/B ratio of 1.28 (Fig. 4.7). In this example, with an assumed feeding rate of 400 g d1 it would take one cormorant 640 days (or 10 cormorants 64 days) to reduce the annual production by 20% on a 5-ha lake.

04 11/02/2003 12:57 Page 49

Theoretical assessment of cormorant impact

49

No. of cormorant feeding days

14000 12000 1% 2% 5% 10% 20%

10000 8000 6000 4000 2000 0 0

10

20

30

40

50

60

70

80

90

100

Size of still water (ha)

Figure 4.7 The relationship between the number of cormorant feeding days taken to consume the indicated proportion of annual productivity and the size of the water (with an assumed biomass of 200 kg ha1 and a P/B ratio of 1.28)

4.4 Discussion The simple empirical models used the best available information to estimate the average annual impacts of cormorants at a national and a local scale. The models’ estimates could be considered to be questionable as a number of uncertainties remain. For example, with regard to the number of fish eaten by cormorants, while the estimates of the number of cormorants (Table 4.2) and their daily feeding rates (Table 4.3) may be reliable, there is no reliable estimate of the proportion of the population feeding inland. There are a number of potential concerns regarding the estimate of freshwater fish production. Firstly, the estimates of biomass for both rivers and still waters (Figs 4.1 and 4.2), though reliable, were based on estimates from England and Wales and were extrapolated to Scotland where conditions might be different. Secondly, the production data were quite skewed and variable and came from a wide variety of sources across Europe and America (Mann & Penczak 1986). However, the sensitivity of the outcome to these uncertainties can be explored. To illustrate this, if all of the cormorants fed entirely on freshwater fish and if freshwater fish annual productivity was half of that estimated by the model, then predation by cormorants would still only account for approximately 10% of the annual production of freshwater fish. The outputs of the model of local impacts could be used to help develop guidelines for deciding where the control of predation by cormorants is a priority. This would require a political decision regarding what proportion of annual fish production taken by cormorants is acceptable and an estimate of the standing crop of fish in the waters where control is being considered. It is recognised that these models ignore the impact of the wounding of fish by cormorants and the impacts of size-specific feeding.

04 11/02/2003 12:57 Page 50

50

Impact of birds on fisheries

In conclusion, while some uncertainties remain, the simple models presented give no evidence that would support a national cull of cormorants and could be used to estimate the impact of cormorants on a local basis.

Acknowledgements and disclaimer The authors thank Jon Hately, two anonymous referees and the editor for comments that greatly improved an earlier version of this chapter. The views expressed are those of the authors and not necessarily those of the Environment Agency.

References Hughes B., Bevan R.M., Bowler J.M., Still L., Carss D.N., Marquiss M., Hearn R.D. & Bruce J.H. (1999) Feeding Behaviour of Fish-eating Birds in Great Britain. London: Department of the Environment, Transport and the Regions, 249 pp. Feltham M.J. & Davies J.M. (1997) Daily food intake of cormorants: A summary. Proceedings of IV European Cormorant Conference, Bolgna. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 259–268. Mainstone C.P., Barnard S. & Wyatt R. (1994) The NRA National Fisheries Classification Scheme a Guide for Users. National Rivers Authority R&D Note 206. Bristol: National Rivers Authority, 102 pp. Mann R.H.K. & Penczak T. (1986) Fish production in rivers: A review. Polskie Archiwum Hydrobiologii 33, 233–247. Marquiss M., Carss D.N., Armstrong J.D. & Gardiner R. (1998) Fish-Eating Birds and Salmonids in Scotland. Report to The Scottish Office Agriculture Environment and Fisheries Department, ITE, Banchory, and Freshwater Fisheries Laboratory, Pitlochry, 156 pp. North R. (2001) Stocking Densities for Stillwater Coarse Fisheries. R&D Technical Report W269. Bristol: Environment Agency, 93 pp. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project number VC0104. London: MAAF, 130 pp. Smith I. & Lyle A. (1979) Distribution of Freshwaters in Great Britain. Edinburgh: Institute of Terrestrial Ecology, 44 pp. Wernham C., Armitage M., Holloway S.J., Hughes B., Kershaw M., Madden J.R., Marchant J.H., Peach W.J. & Rehfisch M.M. (1999) Population, Distribution, Movements and Survival of Fisheating Birds in Great Britain. London: Department of the Environment, Transport and the Regions, 360 pp.

05 11/02/2003 13:49 Page 51

Chapter 5

Fish stocks, commercial fishing and cormorant predation in the Vistula Lagoon, Poland L. STEMPNIEWICZ* Department of Vertebrate Ecology & Zoology, University of Gdan´sk, Gdan´ sk, Poland

A. MARTYNIAK Warmia-Mazurian University, Olsztyn, Poland

W. BOROWSKI Sea Fishery Institute, Gdynia, Poland

M. GOC Department of Vertebrate Ecology & Zoology, University of Gdan´sk, Gdan´ sk, Poland

Abstract Fish community structure and cormorant Phalacrocorax carbo sinensis (L.) breeding and feeding activities were studied in the Vistula Lagoon between 1995 and 1997. Commercial catches have fluctuated considerably during the last 30 years and pike, Esox lucius L., disappeared from catches in the late 1970s. Ruffe, Gymnocephalus cernuus (L.), was preferred by cormorants, followed by roach, Rutilus rutilus (L.), perch, Perca fluviatilis (L.), and herring, Clupea harengus (L.). In one season 2200–3100 t of small fish was removed from the lagoon (bycatch: 1484–2267 t and cormorants: 712–816 t). In successive years, younger and younger age classes dominated the catches suggesting overfishing of the older individuals. However, proportions of prey species in the fish community remained the same. The most intensive predation took place in June when total energy requirements of cormorants were highest. Birds selected smaller fish than available. Due to their small size and species composition, fishes taken by cormorants were considered of negligible economic value, although their later contribution to commercial catches was not accounted for. Keywords: cormorant, fish community, Phalacrocorax carbo, Poland, Vistula Lagoon.

5.1 Introduction Cormorants, Phalacrocorax carbo sinensis (L.), are numerous and widespread top predators feeding on fish in the freshwater bodies, estuaries and coastal marine waters. Their numbers and range of distribution have expanded greatly in recent years (van Eerden & Gregersen 1995; Lindell et al. 1995), possibly as a result of increased availability of food resources (van Dobben 1995; van Eerden & Voslamber 1995; de Nie 1995; Voslamber et al. 1995). This has resulted in serious cormorant–fishery conflicts in

*Correspondence: Lech Stempniewicz, Department of Vertebrate Ecology & Zoology, University of Gdan´sk, Legionów 9, 80-441 Gdan´ sk, Poland (email: [email protected]).

05 11/02/2003 13:49 Page 52

52

Impact of birds on fisheries

most European countries (Draulans 1988; Bayer 1989; Linn & Campbell 1992; Carss 1993, Veldkamp 1995; van Dam et al. 1995; Stempniewicz et al. 1998; Stempniewicz et al. in press). As with many other regions of Europe, the number of breeding cormorants in the colony at Katy Rybackie, northern Poland increased considerably during the 1990s, reaching 5000–6000 pairs between 1995 and 1997 (Goc & Nitecki 1997; Stempniewicz et al. 1998). The consequences of cormorant predation on fisheries resources in natural freshwater ecosystems are uncertain, and results from numerous studies are ambiguous (e.g. Draulans 1988; Bayer 1989; Duffy 1995; van Eerden et al. 1995; Platteeuw & van Eerden 1995; van Rijn & Platteeuw 1996; Keller et al. 1997; Marion 1997; Stempniewicz et al. 1998). This contribution represents part of a comprehensive longterm study on the role of cormorants in the Vistula Lagoon ecosystem. This chapter presents data on seasonal and inter-seasonal changes in: (1) fish community; (2) commercial catches; (3) cormorant diet; and (4) estimation of total fish biomass taken by fishermen and cormorants from the Katy Rybackie colony. The possible impact of cormorant predation and fishery harvest in the Vistula Lagoon is discussed.

5.2 Study area The Vistula Lagoon, situated east of Vistula Mouth, is separated from the Gulf of Gdan´sk by a narrow spit. This large (838 km2) muddy water body is very shallow. Its average water depth is 2.6 m (max. 5.0 m) but in the south-west part, where cormorants feed most often (Stempniewicz et al. 1999), it does not exceed 2.0 m. The water in the lagoon is characterised by low transparency (Secchi disc values: 30–90 cm), which declines further (to 20 cm) during windy weather when wave action brings benthic material into suspension. In recent years considerable degradation of the underwater vegetation, and thus deterioration of fish spawning conditions in the lagoon, have been observed (Róz˙an´ ska & Wiecawski 1978; Stempniewicz et al. 1999; Borowski & Dabrowski, unpublished data). Katy Rybackie (54°21N, 19°14E) is situated at the base of the Vistula Spit, some 100 m from the sea and about 2.5 km from the lagoon (Fig. 5.1). Cormorants use both water bodies as feeding grounds depending on prevailing conditions. As a result, they can avoid unprofitable conditions (e.g. icing and turbulence) and make the most of temporal and local fish abundance (e.g. during the spawning period). Vistula Lagoon, however, is a much more important feeding ground, especially during spring (Goc et al. 1997). There is an intensive commercial fishery on the lagoon targeting: bream, Abramis brama (L.), pikeperch, Stizostedion lucioperca (L.), eel, Anguilla anguilla (L.), herring, Clupea harengus (L.), roach, Rutilus rutilus (L.), and perch Perca fluviatilis (L.) (Borowski & Dabrowski, unpublished data).

5.3 Material and methods Data on commercial catches between 1970 and 2000 were obtained from the Maritime Office in Gdynia. Every two weeks between April and September 1995–1997, 12 fyke

05 11/02/2003 13:49 Page 53

Fish and cormorants in the Vistula Lagoon, Poland

53

Russia Piaski Baltic Sea Vistula Spit

I – East

Cormorant colony at Katy Rybackie Frombork II – Middle

Tolkmicko

III – West

IV – South Elblag Bay

Figure 5.1 Sketch map of the Vistula Lagoon (I–IV: distinguished regions; circles denote fyke-netting sites)

nets (mesh size up to 16 mm) were set (for 48 h) in different parts of the lagoon (three in each of the four regions, Fig. 5.1) to collect fish for analysis. All fish in the samples were identified, measured (total length in mm) and grouped according to length class. All fishes from each size class were weighed (to the nearest 2 g). Age was determined on the basis of scale readings. Total fish biomass taken by fishermen as a bycatch (i.e. fish below the legal size limit) was estimated on the basis of sample catches and known fishing effort in the Vistula Lagoon. Cormorant population breeding in Katy Rybackie was estimated by direct counts of nests in the whole colony at the peak of each breeding season. The timing of cormorant breeding was studied on sampling plots comprising 185–290 nests in the particular seasons (Kopciewicz et al. in press). Cormorant food consumption was estimated using published values of daily food requirement: 350–450 g of fish for adults and 250–350 g for nestlings (van Dam et al. 1995). These values were then multiplied by the number of adults and nestlings respectively, present in the colony during the successive stages of the breeding season. The number of nestlings was calculated, basing on mean fledging success (number of fledglings per nest) studied on the sample plots (Kopciewicz et al. in press). Cormorant diet was studied on the basis of pellet analysis and the composition of regurgitated fish found at the colony. Samples (fresh pellets and fish) were collected every 3 weeks from March to August between 1995 and 1997. In 1995 the number of pellets analysed was 603 and number of regurgitated fish 1842; in 1996 the sample sizes were 702 and 1430 respectively; and in 1997, 784 and 929 respectively. Pellet

05 11/02/2003 13:49 Page 54

54

Impact of birds on fisheries

analysis followed standard methods (Hyslop 1980; Carss 1997). Regurgitated fish were measured (TL, mm) and weighed (to the nearest 2 g).

5.4 Results 5.4.1

Commercial catches

Since 1970, there has been an increase in the cormorant population in Katy Rybackie and its consequent predation on fish resources, whereas catches of most common commercial fishes in the Vistula Lagoon have fluctuated widely (Fig. 5.2). No long-term decline in the stocks of species predated on by cormorants was observed. Pike, Esox lucius L., and tench, Tinca tinca (L.), disappeared from commercial catches in the late 1970s, coupled with the

150

Roach

500

Bream

400 100

300 200

50

100 0 80

0 500

Perch

400

60

Pikeperch Eel

300

40 Catch (t)

200 20

100

0 120

0 Silver bream

4000

90

3000

60

2000

30

1000

0

0

10 Tench 8 6 4 2 0 1970 1974 1978 1982 1986 1990 1994 1998

Figure 5.2

8

Herring

Pike

6 4 2 0 1970 1974 1978 1982 1986 1990 1994 1998

Commercial catches (t) in the Vistula Lagoon during the last 30 years

05 11/02/2003 13:49 Page 55

Fish and cormorants in the Vistula Lagoon, Poland

55

loss of submerged vegetation in the lagoon. At the same time, herring arrived in the lagoon and since the early 1980s has constituted a large proportion of the catches (Fig. 5.2).

5.4.2

Fish community

Ruffe, Gymnocephalus cernuus (L.), was the most common fish species in the Vistula Lagoon (22–29% in sample catches, by numbers), followed by roach (8.2–25.5%), bream (14.0–23.6%), perch (12.6–13.8%), herring (0.9–25.5%) and pikeperch (4.1–8.8%). During three years of the study, contributions of ruffe, perch and pikeperch in the catch have changed little whereas the catches of roach and bream have fluctuated and herring have increased markedly (Table 5.1). The availability of the main cormorant prey species in the lagoon (ruffe, roach, perch), expressed as numbers and biomass in sample catches, changed considerably over the course of the season. Starting from low numbers in April, catches increased gradually, achieving the highest value in July, followed by a decline in numbers and biomass in August and September (Fig. 5.3). The age structure of many fish populations changed considerably during the study period (ANOVA: F  8892–27578, d.f.  2, P  0.001). In successive years, younger and younger classes of the most important cormorant prey species predominated in the sampled catches (Fig. 5.4). The mean length of some fish species in the sampled catches (e.g. perch and herring) tended to decrease in successive years, while for others (e.g. ruffe and roach) it remained stable (Table 5.2).

Table 5.1 Proportions of common fish species occurring in the sample catches, in cormorant diet and estimated biomass taken by cormorants in 1995–1997a In catches (%)

In diet (%)b

Species

1995

1996 1997

1995

Gymnocephalus cernua Rutilus rutilus Clupea harengus Perca fluviatilis Zoarces viviparusc Stizostedion lucioperca Blicca bjoerkna Abramis brama Anguilla anguilla Scardinus erythrophthalmus Osmerus eperlanus Total

22.3 25.5 0.9 13.6 – 7.6 0.6 23.6 1.2 0.2 0.3 95.8

29.0 8.2 25.5 13.8 – 4.1 1.0 14.0 1.6 0.6 0.2 98.0

58.0 70.6 12.1 10.0 7.7 2.2 6.9 3.5 6.0 4.0 5.4 2.9 – 2.1 2.4 1.8 1.0 1.4 – 1.1 0.5 0.3 100.0 99.9

a

27.9 18.7 1.6 12.6 – 8.8 2.6 17.2 0.8 0.6 1.0 91.8

Concerns both Vistula Lagoon and Gulf of Gdan´sk. Proportion (%) in pellets. c Does not occur in the lagoon. b

Biomass taken (t)

1996 1997 1995 1996

1997

75.5 467.2 520.3 537.6 5.3 97.3 73.7 37.7 1.2 62.1 16.2 8.5 3.7 55.3 25.8 26.3 5.3 48.1 29.5 37.7 2.6 43.4 21.4 18.5 0.7 – 15.5 5.0 1.8 19.6 13.3 12.8 1.9 8.2 10.3 13.5 0.2 – 8.1 1.4 1.4 3.9 2.2 10.0 99.6 805.0 736.3 709.0

05 11/02/2003 13:49 Page 56

56

Impact of birds on fisheries 45 Mean catches Taken by cormorants

40 35 Percentage

30 25 20 15 10 5 0 FEB

MAR

APR

MAY

JUN

JUL

AUG

SEP

Figure 5.3 Seasonal changes in the main prey (ruffe, roach and perch) availability in the Vistula Lagoon (catches) and biomass of fish taken by cormorants (t)

5.4.3

Cormorant diet

Prey proportions Generally, the most commonly occurring fish species in the lagoon constituted the bulk of cormorant prey. Ruffe was taken most commonly (58.0–75.5% by numbers), followed by roach (5.3–12.1%), perch (3.5–6.9%) and herring (1.2–7.7%). Eel pout, Zoarces viviparus (L.), which does not occur in the lagoon, constituted 4.0–6.0% of the diet. During the study period, contributions of common prey species in the cormorant diet were either constant (e.g. eel pout), increased markedly (e.g. ruffe, from 58% in 1995 to 75.5% in 1997), or tended to decrease (e.g. roach, herring, perch and pikeperch) (Table 5.1).

Prey size The great majority of prey items were 10 cm (Table 5.2), significantly below the size exploited in the commercial fishery (Mann–Whitney, P  0.001). The mean length of ruffe, roach and perch in the pellets (but not in regurgitated fish) tended to decrease in successive years of the study (Table 5.2).

Estimation of fish biomass removed from the lagoon Estimated fish biomass taken by cormorants during the whole breeding season varied between 700 and 800 t (500–580 t taken from the lagoon), and depended greatly on the number of breeding birds and chick production in the particular season. A total of 4942 pairs bred in Katy Rybackie in 1995, 5929 in 1996 and 5152 in 1997. Mean fledging success (number of fledglings per breeding pair) decreased during the study period,

05 11/02/2003 13:49 Page 57

Fish and cormorants in the Vistula Lagoon, Poland

57

Figure 5.4 Population age structure of the main fish species in the Vistula Lagoon (basing on sample catches) in the consecutive years of the study (1–10: age classes)

from 2.45 in 1995 to 2.19 in 1996 and 2.07 in 1997 (Kopciewicz et al. in press). Total fish biomass taken by cormorants during one season consisted mainly of ruffe (470–540 t), whilst the contribution of the remaining important prey types (roach, herring, perch, eel pout and pikeperch) did not exceeded 100 t (Tables 5.1 and 5.3).

Pellets

Dropped fish

Sample catches

Mean  SD (n)

Mean  SD (n)

Mean  SD (n)

Species

Season

Gymnocephalus cernua

1995 1996 1997

Rutilus rutilus

1995 1996 1997

10.8  3.2 (710) 9.9  3.2 (663) 9.3  2.7 (660)

11.8  3.1 (83) 12.6  2.6 (40) 12.9  5.0 (39)

14.1  3.1 (4879) 14.3  2.9 (6743) 14.1  2.7 (8516)

Clupea harengus

1995 1996 1997

15.1  3.0 (336) 14.8  2.6 (214) 15.0  2.9 (69)

17.4  1.0 (98) 16.4  2.8 (203) 21.8  4.1 (19)

20.5  2.9 (319) 19.5  1.8 (20964) 16.6  2.2 (737)

Perca fluviatilis

1995 1996 1997

9.0  4.7 (75) 9.5  3.5 (51) 9.5  2.2 (53)

13.8  2.8 (259) 12.3  1.9 (11357) 12.8  2.4 (5702)

Zoarces viviparus

1995 1996 1997

19.2  4.1 (180) 21.2  8.2 (144) 19.7  4.7 (305)

22.7  4.4 (169) 21.9  3.9 (52) 21.2  4.3 (45)

Not observed in the lagoon

Stizostedion lucioperca

1995 1996 1997

18.5  5.3 (92) 15.4  4.8 (247) 16.7  5.4 (72)

18.0  5.3 (50) 18.6  7.4 (15) 27.8  9.9 (6)

15.9  5.3 (1460) 17.8  5.1 (3344) 15.5  4.8 (4023)

Blicca bjoerkna

1995 1996 1997

– 12.3  5.2 (37) 10.7  3.2 (44)

– – 10.2  2.6 (9)

15.8  3.6 (107) 15.6  3.0 (802) 13.3  3.1 (1171)

8.1  1.9 (7076) 7.3  1.5 (13224) 7.4  1.5 (14975)

9.8  2.5 (528) 6.3  1.9 (1060) 6.7  1.4 (1174)

7.8  1.83 (429) 7.2  1.61 (576) 8.3  4.84 (352)

11.0  1.3 (4260) 11.0  0.8 (23854) 11.0  1.2 (12720)

Impact of birds on fisheries

Mean length (cm) of fish taken by cormorants and caught in sample catches in 1995–97

05 11/02/2003 13:49 Page 58

58

Table 5.2

15.3  3.2 (57) 13.9  3.8 (60) 15.0  3.9 (57)

14.4  4.5 (64) 16.5  4.4 (49) 11.9  3.4 (9)

15.8  3.6 (107) 15.6  3.0 (802) 13.3  3.1 (1171)

Anguilla anguilla

1995 1996 1997

50.1  8.4 (6) 35.0  10.2 (17) 47.5  5.9 (14)

41.8  8.0 (15) 36.2  4.0 (4) 45.7  10.4 (3)

47.3  13.2 (223) 47.4  9.7 (1358) 51.2  11.8 (379)

Scardinus erythrophthalmus

1995 1996 1997

– – –

– – –

15.9  2.3 (187) 12.4  1.7 (483) 13.1  1.6 (291)

Osmerus eperlanus

1995 1996 1997

10.5  4.9 (36) 8.7  1.6 (87) 8.0  0.8 (7)

17.1  5.0 (317) 14.7  4.8 (174) 12.8  4.4 (463)

Neogobius melanostomus

1995 1996 1997

– – –

– 10.2  1.3 (30) 15.2  2.3 (5)

Not observed in the lagoon

Gasterosteus aculeatus

1995 1996 1997

– – –

–

– – –

Cyprinus carpio

1995 1996 1997

– – 10.7  4.1 (6)

– – –

– – –

Tinca tinca

1995 1996 1997

– – 16.0  3.4 (5)

– 16.0 (1) 20.2 (2)

– – –

6.8  2.9 (269) 6.6  1.8 (2065) 6.9  2.5 (551)

5.3  0.4 (96) 5.9  0.2 (4)

05 11/02/2003 13:49 Page 59

1995 1996 1997

Fish and cormorants in the Vistula Lagoon, Poland

Abramis brama

59

Fish biomass taken (t)

Total number of birds Daily food (adults  juveniles) intake (g)

Young

Adults

Total

Month

1995

1996

1997

Adult Juv.

1995 1996

1997

1995

1997

1995

1996

1997

February March April May June July August Total

1.5 8.5 13.5 20.6 22.6 10.5 0.9

0.0 1.0 10.2 18.6 25.7 13.6 0.8

0.3 2.6 10.7 17.1 21.2 14.5 1.2

450 430 400 380 350 350 350

9.4 113.3 126.0 124.9 110.2 59.7 3.3 546.7

1.9 34.7 112.8 127.2 110.3 73.8 6.5 467.2

0.0 0.0 0.0 0.0 0.0 0.0 11.2 0.0 9.8 77.5 46.5 48.8 108.9 117.0 96.3 54.2 75.9 83.5 6.5 5.4 6.5 258.3 244.8 244.9

9.4 113.3 137.2 202.4 219.1 113.9 9.7 805.0

0.0 13.3 122.4 194.9 250.3 147.5 8.6 737.0

1.9 34.7 122.6 176.0 206.6 157.3 13.0 712.1

250 250 300 350 350

0.0 13.3 122.4 148.4 133.3 71.6 3.2 492.2

1996

Percent of birds Fish biomass taken feeding from the lagoon (t) in the lagoon 1995 1996 1997 90 85–90 83–87 75–80 62–63 55–65 50

8.5 98.6 119.4 151.8 135.8 62.6 4.9 581.6

0.0 12.0 101.6 155.9 157.7 95.9 4.3 527.4

1.7 31.2 104.2 137.3 130.2 94.4 6.5 505.5

Impact of birds on fisheries

Estimation of total biomass of fish taken by cormorants in the seasons, 1995–1997

05 11/02/2003 13:49 Page 60

60

Table 5.3

05 11/02/2003 13:49 Page 61

Fish and cormorants in the Vistula Lagoon, Poland

61

Table 5.4 Biomass of small fish taken by cormorants and by fishermen in the Vistula Lagoon during the study period (1995–1997)

Season 1995 1996 1997

Total fish biomass taken by cormorants (t) 805.0 737.0 712.0

Fish biomass taken by cormorants from the lagoon (t) 581.6 527.4 505.5

Total fish biomass in the bycatch (t)a

Biomass of the ruffe, perch and roach in the bycatch (t)

2267.2 2215.8 1483.6

1033.6 1129.9 771.9

Biomass in sample catches multiplied by fishing effort (N fyke-nets  n days).

a

Fish biomass consumed by cormorants during a particular season changed with the number of birds present in the colony (adults and young) and duration of their stay (feeding) at the lagoon (i.e. with their combined food requirements.). The largest biomass was taken in May–June (200–250 t, including 130–160 t taken from the lagoon) when energy requirements of cormorants (almost full-grown nestlings) were highest (Table 5.3, Fig. 5.3). Total harvest by fishermen from the lagoon was estimated to 1500–2300 t yr 1. Three important cormorant prey species (ruffe, perch and roach) constituted more than half of that figure (Table 5.4).

5.5 Discussion Seasonal shifts between the availability of the main cormorant prey species (ruffe, roach and perch) in the Vistula Lagoon (expressed as size of catches) and cormorant food requirements were observed. Sampled catches, which was assumed to approximately reflect fish availability to cormorants, showed a conspicuous peak in the warmest month (July), most probably reflecting higher fish activity associated with warmer water temperatures. The summer peak in fish catches does not reflect their spawning period, which generally occurs earlier in the season, usually in March–June, depending on ice conditions and water temperature in spring. The greatest total food requirements of cormorants was in May and June, when large nestlings were in the colony. The highest availability of fish, in July, coincided with the period when the majority of naïve young cormorants started to feed in the lagoon. This mismatch suggests that food resources in the Vistula Lagoon do not limit cormorant reproduction in the Katy Rybackie colony, except possibly at the beginning of the season. Kopciewicz et al. (in press) showed that early breeders, with peak energy requirements falling in May (when fish availability is relatively low) had considerably higher fledging success than those rearing their young during the peak of prey availability in the lagoon. Total fish biomass consumed by cormorants during the particular season depended on the duration of the breeding period (i.e. the duration of their feeding at the lagoon).

05 11/02/2003 13:49 Page 62

62

Impact of birds on fisheries

Stempniewicz et al. (2000) showed that total food consumption of the cormorants was lower during a delayed (i.e. shorter and much more synchronised) breeding season than in a season when weather conditions were more typical. However, during the period of peak energy need in June, the estimated daily food consumption of young and adult cormorants present in the colony was about 2 t higher in the synchronised season. Strong eutrophication of the Vistula Lagoon favouring mass development of small shoaling fish, coupled with overfishing of predatory species seems to be responsible for the abundance of easy, accessible prey, high breeding success of cormorants and gradual increase of the colony at Katy Rybackie. Despite strong pressure exerted by cormorants and fishermen (bycatch) on small fish in the lagoon (1280–1600 t removed per season), their contribution to the fish community remained stable. This was presumably due to the reproductive potential of the species and their ability to compensate for high mortality, even under high cormorant predation pressure. Climate warming is likely to increase the annual period suitable for cormorant reproduction. It could potentially make earlier, hence more successful, cormorant breeding possible and favour their north-eastern dispersion. Earlier breeding of cormorants, some 3–4 weeks earlier, has been noted across Central East Europe over the last 50 years (Bauer & Glutz 1966; Cramp 1977; Kopciewicz et al. in press; L. Stempniewicz, unpublished data). This is especially important in the case of large colonies, where intra-specific competition is potentially an important limiting factor. A series of severe winters delaying and shortening the time of breeding could limit cormorant population size. The combined, intensive and selective predation pressure from cormorants and fishermen seems to influence all age classes of the prey fish populations. During the successive years of the study, younger and younger age classes predominated in the sampled catches, suggesting overfishing of older individuals. As the mesh-size of the fyke nets is 16 mm, the smallest fish are under-represented, and consequently their total numbers and biomass (as well as their proportion in total catches) were underestimated. Cormorants selected fish well below the commercial size and smaller than the mean size calculated from the sampled catches, presumably exploiting the most numerous and profitable prey. Due to their small size and species composition, fishes taken by cormorants were considered of negligible economic value, although their later contribution to commercial catches (Britton et al., Chapter 2) was not accounted for.

Acknowledgements We would like to express our thanks to Henryk Dabrowski, Lech Iliszko, Czesaw Nitecki and Janusz Terlecki, as well as to all other members of the team involved in the cormorant research programme carried out in the Vistula Lagoon, for their help. The Maritime Office in Gdynia made their data on commercial catches accessible. The Maritime Office in Gdynia and Department of Environment Protection in Elblag Voivodship Office partly financed this study. Publication supported by grant no. 6PO4G08716.

05 11/02/2003 13:49 Page 63

Fish and cormorants in the Vistula Lagoon, Poland

63

References Bauer K.M. & Glutz von Blotzheim N. (1966) Handbuch der Vogel Mitteleuropas, vol. 1, Frankfurt am Main: Akademische Verlagsgesellschaft, 483 pp. Bayer R.D. (1989) The cormorant/fisherman conflict in Tillamook County, Oregon. Studies in Oregon Ornithology 6, 1–99. Carss D.N. (1993) Cormorants Phalacrocorax carbo at cage fish farms in Argyll, western Scotland. Seabird 15, 38–44. Carss D.N. (1997) Technics for assessing cormorant diet and food intake: towards a consensus view. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 197–230. Cramp S. (ed.) (1977) Handbook of the Birds of Europe the Middle East and North Africa. The Birds of Western Palearctic, vol. 1, Oxford: Oxford University Press, 483 pp. Draulans D. (1988) Effects of fish-eating birds on freshwater fish stocks: an evaluation. Biological Conservation 44, 251–263. Duffy D.C. (1995) Why is the double-crested cormorant a problem? Insights from cormorants ecology and human sociology. In D.N. Nettleship & D.C Duffy (eds) The double-crested cormorant: biology, conservation and management. Colonial Waterbirds 18 (Special Publication 1), 25–32. Goc M., Iliszko L. & Chelkowska N. (1997) Daily foraging rhythm at a cormorant Phalacrocorax carbo colony during the breeding season. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 445–451. Goc M. & Nitecki C. (1997) Human activities accelerate the expansion of the cormorant breeding colony at Katy Rybackie, NE Poland. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 453–456. Hyslop E.J. (1980) Stomach content analysis: a review of methods and their application. Journal of Fish Biology 17, 411–429. Keller T., Vordermeier T., von Lukowicz M. & Klein M. (1997) The impacts of cormorants on the fish stocks of several Bavarian water bodies with special emphasis on the ecological and economical aspects. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 295–311. Kopciewicz P., Olszewska A., Nitecki C., Bzoma S. & Stempniewicz L. (in press) Changes in the breeding success of Great Cormorants Phalacrocorax carbo sinensis in the expanding colony at Katy Rybackie (N Poland): effect of phenology and age of subcolony. Vogelwelt. Lindell L., Mellin M., Musil P., Przybysz J. & Zimmerman H. (1995) Status and population development of breeding cormorants Phalacrocorax carbo sinensis of the central European flyway. Ardea 83, 81–92. Linn I.J. & Campbell K.L.I. (1992) Interactions between white-breasted cormorants Phalacrocorax carbo (Aves: Phalacrocoracidae) and the fisheries of Lake Malawi. Journal of Applied Ecology 29, 619–634. Marion L. (1997) Comparison between the diet of breeding cormorants Phalacrocorax carbo sinensis, captures by fisheries and available fish species: the case of the largest inland colony in France, at the Lake of Grand-Lieu. Supplemento alle Ricerche di Biologia della Selvaggina XXVI, 313–322. de Nie H. (1995) Changes in the inland fish populations in Europe in relation to the increase of the cormorant Phalacrocorax carbo sinensis. Ardea 83, 115–122. Platteeuw M. & van Eerden M.R. (1995) Time and energy constraints of fishing behaviour in breeding cormorants Phalacrocorax carbo sinensis at lake Ijsselmeer, The Netherlands. Ardea 83, 223–234. Róz˙an´ska Z. & Wiecawski F. (1978) [Studies on the environmental factors in the Vistula Lagoon under the conditions of human pressure]. Studia i Materiay Oceanologiczne 21, 9–36 (in Polish). Stempniewicz L., Goc M. & Nitecki C. (1998) [The need to conduct ecological studies on the cormorant Phalacrocorax carbo in Poland]. Notatki Ornitologiczne 39, 33–45 (in Polish).

05 11/02/2003 13:49 Page 64

64

Impact of birds on fisheries

Stempniewicz L., Borowski W., Goc M., Iliszko L., Martyniak A. & Nitecki C. (1999) Impact of cormorants Phalacrocorax carbo on fish resources in the Vistula Lagoon, NE Poland. The Ring 21, 55. Stempniewicz L., Goc M., Bzoma S., Nitecki C. & Iliszko L. (2000) Can timing of breeding affect chick mortality in the great cormorant Phalacrocorax carbo? Acta Ornithologica 35, 33–39. Stempniewicz L., Martyniak A., Borowski W. & Goc M. (in press) Interrelationships between the ruffe Gymnocephalus cernuus and cormorant Phalacrocorax carbo sinensis in the Vistula Lagoon, N. Poland. Vogelwelt. Van Dam C., Buijse A.D., Dekker W., van Eerden M.R., Klein Breteler J.G.P. & Veldkamp R. (1995) Cormorants and commercial fisheries. Rapport IKC 19. Wageningen (in Dutch with English summary), 104 pp. Van Dobben W.H. (1995) The food of the cormorant: an old and new research compared. Ardea 83, 139–142. Van Eerden M.R. & Gregersen J. (1995) Long-term changes in the NW-European population of cormorants Phalacrocorax carbo sinensis. Ardea 83, 61–78. Van Eerden M.R. & Voslamber B. (1995) Mass fishing by cormorants Phalacrocorax carbo sinensis at lake Ijsselmeer, The Netherlands: the recent and successful adaptation to a turbid environment. Ardea 83, 199–212. Van Eerden M.R., Koffijberg K. & Platteeuw M. (1995) Riding on the crest of the wave: Possibilities and limitations for a thriving population of migratory cormorants Phalacrocorax carbo in mandominated wetlands. Ardea 83, 1–10. Van Rijn S. & Platteeuw. M. (1996) Remarkable fledgling mortality at the largest Great Cormorant Phalacrocorax carbo sinensis colony in The Netherlands. Cormorant Research Group Bulletin 2, 30–35. Veldkamp R. (1995) Diet of cormorants Phalacrocorax carbo sinensis at Wanneperveen, The Netherlands, with special reference to the bream. Ardea 83, 143–155. Voslamber B., Platteeuw M. & van Eerden M.R. (1995) Solitary foraging in sand pits by breeding cormorants Phalacrocorax carbo sinensis: does specialised knowledge about fishing sites and fish behaviour pay off? Ardea 83, 213–222.

06 11/02/2003 13:05 Page 65

Chapter 6

The impact of cormorants on the eel fishery in the River Havel catchment area, Germany R. KNÖESCHE* Institute of Inland Fisheries, Jaegerhof, Gross Glienicke, Potsdam-Sacrow, Germany

Abstract The amount of eel, Anguilla anguilla (L.), consumed by cormorants, Phalacrocorax carbo sinensis (L.), is the subject of considerable debate between ornithologists and fishermen in the River Havel, Germany. A long-term dataset on eel stocking intensity, eel yield and CPUE from a 32 m stow net in the River Elbe below the confluence of the River Havel was used to assess the influence of different factors (e.g. water flow, stocking rate and the number of cormorants) on fishery yield. Approximately 28% of the variability in yield of eels was explained by the number of cormorants present. Eels contributed 19% of the diet of cormorants. These results were compared with stomach contents analysed from 58 cormorants in which eels contributed 42% of the diet in midsummer but only 4.5% in winter, with a mean of 16% for the entire year. Keywords: conflict, cormorants, depredation, eel.

6.1 Introduction The amount of eel, Anguilla anguilla (L.), consumed by cormorants, Phalacrocorax carbo sinensis (L.), and the impact thereof on fishery yield, are the subject of considerable debate between ornithologists and fishermen. Fishermen consider cormorants to cause serious predation problems by depleting eel stocks. Ornithologists, on the other hand, deny any negative influence of cormorants on eel fisheries. They claim, based on faecal pellet analyses, that in inland waters, cormorants feed predominantly on small (often stunted) cyprinids, which they also consider can be a benefit to the fishery. However, van Dobben (1952, 1997) concluded that the prey spectrum of cormorants is a reflection of the fish stock in the foraging area, but that they also prefer bigger fish and species that are easy to catch. The proportion of eels in the diet of cormorants was also found to vary seasonally, with a higher contribution (up to 35%) in mid-summer. This was attributed to the more active behaviour of eel in higher temperatures during the summer, and their increased vulnerability to predation. A similar picture was found in 1998 and 1999 when the stomachs of 58 cormorants shot in the Federal State of Brandenburg, Germany, were examined (Fig. 6.1). These data on the contents of the stomachs are also supported by studies on faecal pellet analysis (e.g. Knuth & Rothe 1998; van Dam et al. 1995). *Correspondence: R. Knöesche, Institute of Inland Fisheries, Jaegerhof, D-14476 Gross Glienicke, PotsdamSacrow, Germany (email: [email protected]).

06 11/02/2003 13:05 Page 66

66

Impact of birds on fisheries

Proportion in diet (%-wt)

60 50 Summer, n = 44 Early autumn, n = 14

40 30 20 10 0 Eel

Bream/ Silver bream

Roach

Perch

Tench

Bleak

Figure 6.1 Proportion (% by weight) eel and other fish species in the diet of cormorant from northeast German lakes in 1998 and 1999

This conflict of interests and perceptions needs to be resolved. Unfortunately, because the cormorant is a protected species under the EU Birds Directive (EEC/79/409), the issue is as much a poltical argument as a scientific problem, making a solution difficult to find. This chapter attempts to evaluate, objectively, the impact of cormorant depredation on eel fisheries in Germany.

6.2 Material and methods The River Havel, a tributary of the River Elbe in Germany, drains a lake area of nearly 70 000 ha, which has a high potential eel yield. In contrast, the River Elbe upstream of the Havel confluence has a low potential eel yield. Since 1966, both areas have been exploited using two, 32 m stow nets set in the River Elbe 54 km downstream of the Havel confluence. Since the owner of the nets has recorded the catch every day, it is possible to determine the catch per unit effort (CPUE) [kg eel stow net1 night1] over a long temporal series. Because the fishing method has not changed over time these data can be considered representative of the status of the eel stocks in the river system. Three main factors, for which long-term data sets are available, were assumed to affect the Elbe stow net CPUE.

Stocking rate Stocking of glass eels, grown-on-farm eels (about 5 g) and wild eel fingerlings (about 20 g) has been carried out over the study period. The stocking rate was converted into glass eel equivalents (million glass eel equivalents yr1) according to Tesch (1983), Holmgren et al. (1997) and Anwand (1995) as follows:

• •

glass eel: mean period of growth to capture as emigrating silver eel 7–9 years; grown-on-farm eel: mean period of growth to capture 5–6 years, one advanced eel corresponds to six glass eels (estimated from assumed natural mortality between glass eel and silver eel life stages);

06 11/02/2003 13:05 Page 67

40 35

2500 CPUE Stocking rate (SR) Cormorant breeding pairs (CP) Mean annual flow (AF)

2000

30 25

1500

20 1000

15 10

500

5 0

0 1970

1980

1990

2000

67 Cormorant population [breeding pairs]; AF [m3 s⫺1]

CPUE [kg stow net⫺1 night⫺1] Stocking rate [million glass eel equiv yr⫺1]

Impact of cormorants in the River Havel eel fishery

Figure 6.2 Changes in catch per unit effort of the Elbe pound net, stocking density of glass eel equivalents, mean annual flow and number of cormorant breeding pairs on the River Havel between 1966 and 2000

•

eel fingerlings: mean period of growth to capture 4–5 years, one fingerling corresponds to 10 glass eels.

Because the capture of silver eels from one stocking age group extends over 2–4 years on average (Anwand & Valentin 1981), a group mean value from three mean annual CPUEs was calculated for every stocking year with time lags of 9, 6 and 5 years, for 8, 6 and 5 years and for 7, 5 and 4 years, respectively per type of eel stocked. The time lag of 7, 5 and 4 years gave the best correlation between stocking and catch rates. The divergence between this time lag and the time lag reported by other authors is explained by the dominance of shallow and relatively warm eutrophic and hypertrophic waters in the catchment area.

Cormorant population size Although good data are available on the breeding population size, the number of wintering birds is known only for 1994/1995 (between 3000 and 4000 birds). For this reason, and because the main depredation on eels occurs during the summer months (Fig. 6.1), the breeding population (number of breeding pairs) was considered representative of predation pressure.

Flow rate The relationships between eel CPUE and stocking rate, cormorant abundance and flow were investigated by regression analysis (winSTAT® 3.1). Annual water flow (m3 s1) in the River Elbe (Fig. 6.2), was based on data in the Deutsches Gewässerkundliches Jahrbuch.

06 11/02/2003 13:05 Page 68

68

Impact of birds on fisheries

6.3 Results A positive relationship was found between eel CPUE and stocking rate (SR) for data since 1974 (Equation 6.1), whilst a negative relationship was found between CPUE and cormorant abundance CA (Equation 6.2) (see also Fig. 6.2). Before 1974, CPUE was also influenced by the effects of natural recruitment of eel, therefore the data before 1974 were omitted from the statistical analyses. No relationship was found between annual flow and CPUE (Equation 6.3). CPUE  0.521  SR0.91

n  23

r2  0.68

P  0.001

(6.1)

CPUE  27.451  CA0.26

n  18

r2  0.82

P  0.001

(6.2)

CPUE  0.0544  AF

n  23

r  0.28

P  0.002

(6.3)

0.77

2

The multiple linear regression based on these variables was of the form CPUE  0.193 SR  0.00167 CA  0.00747 AF  0.86 n  18 r2  0.73 P  0.001 If the mean annual flow is fixed at 600 m3 s1 (the approximate mean annual flow for the River Elbe between 1996 and 2000) and the multiple regression equation applied for stocking levels of 5, 15 and 25 million glass eel equivalents (Fig. 6.3), one can conclude that an increase in the stocking rate up to the level of the 1980s could lead to an increase in CPUE from about 5 to about 7 kg night1, if the number of cormorant pairs remains around the present level of 2000. Alternatively, the impact of the 2000 cormorant pairs could be more or less compensated by a doubling of the stocking rate. To achieve the high CPUEs of the early 1980s would require a stocking rate of some 55–60 million glass eel equivalents per year.

CPUE [kg stow net⫺1 night⫺1]

12 SR = 5 mill. SR = 15 mill. SR = 25 mill.

10 8 6 4 2 0 0

500

1000 1500 2000 Cormorant population size [breeding pairs]

2500

Figure 6.3 Influence of adjusting stocking density of glass eel equivalents on CPUE at a fixed mean annual flow of 600 m3 s1 but different cormorant abundance

06 11/02/2003 13:05 Page 69

Impact of cormorants in the River Havel eel fishery Table 6.1

69

Eel yield and cormorant balance in the River Havel catchment area

Mean catch per Elbe-stow-net [kg night1 net1] Total yield of the catchment area [t] Cormorant breeding population [pairs] Total number of birds (pairs  2.7) Biomass [t] of fish eaten by the cormorants between May and August (100 days) [(number of birds  100  0.45/1000)  0.24] (assumes daily food ration of 0.45 kg and eel represents 24% of total fish consumed) Biomass of fish eaten by the wintering cormorants during 200 days [t] [(number of birds  200  0.45/1000)  0.03] (assumes daily food ration of 0.45 kg and eel represents 3% of total fish consumed) Total amount of eel consumed annually [t] Total number of eels consumed (assuming mean weight of 100 g) Expected fishery yield at a catch rate of 0.8 Yield loss due to cormorants (mean weight 300 g) [t] Theoretical yield without cormorant [t]

1980–1984

1995–1999

11.4 434 41 111

4.27 174 1632 4406

1.2

47.6

0.3 1.5

11.9 59.5

15 000 12 000 3.6 446

595 000 476 000 142.8 316.8

The trend in CPUE observed above is similarly reflected in the total eel yield (Y) for the River Havel catchment area (Table 6.1). Both the total annual yield and Elbe-stow-netCPUE have declined approximately 60% between the early 1980s and late 1990s, and a strong correlation exists between both parameters: Y  (0.002422  0.000002056 CPUE)1

n  21

r2  0.95

Consequently a strong correlation exists between yield and stocking rate, and yield and stocking rate and cormorant abundance: Y  SR (0.0272  0.00128 SR)1

n  27

r2  0.81

Y  5.535 SR  0.0804 CA  264.9

n  21

r2  0.94

P  0.001

The yield curves derived for three stocking rates (Fig. 6.4) show the same strong correlation between yield and the cormorant population as between CPUE and the cormorant population (Fig. 6.3). Another way to assess the influence of cormorants on the eel population is described in Table 6.1. Cormorant depredation on eels seems to be most important during the summer (van Dobben 1952; Fig. 6.1), thus the biomass of eel consumed during this period can be calculated as follows: the summer cormorant population size was computed by multiplying the number of the breeding pairs by 2.7 (see Koop & Kieckbusch 1995). This was multiplied by the period of heavy eel depredation of 100 days (according Van Dobben 1952) and the mean daily ration of 0.45 kg (Keller & Vordermeier 1994; Jungwirth et al. 1995) to give the total amount of fish eaten by the cormorants. According to Figure 6.1 the proportion of eel in the diet was 24% in summer and 3% for the remaining time of

06 11/02/2003 13:05 Page 70

70

Impact of birds on fisheries 450 400 SR = 5 mill. SR = 15 mill. SR = 25 mill.

350 Yield [t yr⫺1]

300 250 200 150 100 50 0 0

500

1000

1500

2000

2500

Cormorant population size [breeding pairs]

Figure 6.4 Influence of adjusting stocking density of glass eel equivalents on yield at different cormorant abundance

year, thus the consumption of eel biomass can be calculated and the number of eels eaten determined if a mean individual weight of 100 g (Knoesche 1996) is assumed as 585 000 fish. According to Dekker (2000), the expected yield is 0.8 of consumed fish. Thus, if the mean harvest weight of an eel is 300 g, 1632 cormorant breeding pairs will result in a theoretical eel yield decrease of 142 t yr1, equivalent to a 45% loss in productivity of the fishery.

6.4 Discussion This assessment suggests a strong influence of cormorant summer population abundance on eel yield. An increase in the bird population from 500 to 1000 breeding pairs resulted in a decrease in the stow-net-CPUE in the River Elbe of 11% and in a decrease in eel yield by some 13%. However, reduction in the stocking intensity of glass eels or their equivalents in the 1990s also had a profound effect on the catch rate. The problem that exists is teasing out the effects of these two variables. Furthermore, care must be taken when interpreting these results because the natural recruitment of eels declined in the late 1980s and this could equally have been responsible for the decline in CPUEs after 1986/1987. For 1988–1991, no data for CPUE are available. In addition, there could have been increased natural mortality in the River Elbe eel stocks due to parasitic infestation by Anguillicola crassus, which may have caused the decline in stocks. Thus conclusive evidence that cormorants are impacting directly on the eel fishery of the River Elbe catchment area were not forthcoming from the long-term data series. The decline in the status of the fishery could be equally due the increase in cormorant abundance or decline in recruitment (both natural and artificial). Whatever the cause there is no sign of a decline in eel fishing effort and this is likely to further impact on

06 11/02/2003 13:05 Page 71

Impact of cormorants in the River Havel eel fishery

71

the stocks (Braemick 2002). Further analysis to tease out the effects of these variables is thus required before any definitive impact of cormorant depredation can be confirmed. Nevertheless, if cormorant numbers do not increase further, the CPUE and yield for the fishery can be improved by increasing stocking density.

References Anwand K. & Valentin M. (1981) Alter und Wachstum der Aale in natürlichen Binnengewässern. Zeitschrift für Binnenfischerei 28, 139–143 (in German). Anwand K. (1995) Aalwirtschaft. Schriften VDFF 10 (Bewirtschaftung norddeutscher Seen), Offenbach/M., 19–21 (in German). Braemick U. (2002) Estimation of the fish yield potential of lakes in north-east Germany. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 26–33. Dekker W. (2000) A procrustean assessment of the European eel stock. ICES Journal of Marine Sciences 57, 938–947. Holmgren K., Wickström H. & Clevestam P. (1997) Sex-related growth of European eel Anguilla anguilla with focus on medium silver eel age. Canadian Journal of Fisheries and Aquatic Sciences 54, 2775–2781. Jungwirt A., Woschitz G., Zauner G. & Jagsch A. (1995) Einfluss des Kormorans auf die Fischerei. Österreichs Fischerei 48, 111–125 (in German). Keller T. & Vodermeier T. (1994) Einfluss des Kormorans auf die Fischbestände ausgewählter bayerischer Gewässer unter Berücksichtigung fischökologischer und fischereiökonomischer Aspekte. Forschungsabschlußbericht Bayer. Landesanstalt für Fischerei, 441 pp. (unpublished, in German). Koop B. & Kieckbusch J.J. (1995) Ornithologische Begleituntersuchungen zum Kormoran. Bericht 1995 im Auftrag des Min. f. Natur und Umwelt Schleswig-Holstein, 24 pp. (unpublished, in German). Knoesche R. (1996) Literaturstudie zu den Auswirkungen des Kormoranbefluges auf die Fischbestände und die Fischerei in Brandenburg. Project report, Min. of Agriculture, Potsdam, Germany, 51 pp. (unpublished, in German). Knuth D. & Rothe U. (1998) Ermittlung der Nahrungszusammensetzung des Kormorans Phalacrocorax carbo sinsnesis an natürlichen Gewässern durch Untersuchung von Speiballen. Project report, Min. of Environm. and Nature Protection, Potsdam, Germany, 16 pp. (unpublished, in German). Tesch F.-W. (1983) Der Aal. 2. Berlin & Hamburg: Auflage, Verlag. P. Parey, 340 pp. (in German). Van Dam C., Buijse A.D., Dekker W., van Eerden M.R., Klein Breteler J.G.P. & Veldkamp R. (1995) Aalscholvers en beroepsvisserij. Wageningen; 103 pp. (in Dutch). Van Dobben W.H. (1952) The food of the cormorants in The Netherlands. Ardea 40, 1–63. Van Dobben W.H. (1991) The food of the cormorant: 51 years later. Proceedings of a Workshop on Cormorant 1989. Lelystad: RIVO, pp. 166–169.

07 11/02/2003 13:17 Page 72

Chapter 7

Do cormorants and fishermen compete for fish resources in the Väinameri (eastern Baltic) area? R. ESCHBAUM*, T. VEBER, M. VETEMAA and T. SAAT Institute of Zoology and Hydrobiology, University of Tartu, Tartu, Estonia, and Estonian Marine Institute, Tallinn, Estonia

Abstract Cormorants, Phalacrocorax carbo sinensis (L.), established the first breeding colony in the Väinameri in 1984 and by 1998 there were 2675 pairs. In the early 1990s the intensity of the coastal fishery increased, leading to overfishing of the most important stocks and conflicts between the fisheries and fish-eating animals (cormorants and seals). Cormorants consumed 463 t of fish in 1998 which corresponded to the commercial catch (without herring). The ratio of cormorants to fishermen in removal of the most valuable species (by weight) was 0.7 for perch, Perca fluvialtis (L.), 0.5 for pike, Esox lucius L., 6.4 for pikeperch, Stizostedion lucioperca (L.), 36.2 for burbot, Lota lota (L.), 3.5 for smelt, Osmerus eperlanus (L.), and 0.7 for eel, Anguilla anguilla (L.). The length distribution of roach and perch consumed was similar to that in experimental fishing. In the case of pikeperch, pike and burbot, cormorants preyed on juvenile specimens. Keywords: Baltic Sea, coastal fishery, diet, Phalacrocorax carbo.

7.1 Introduction In recent years, the abundance of great cormorants, Phalacrocorax carbo (L.), has increased rapidly in the Estonian coastal waters (Lilleleht 1995; Lõhmus et al. 1998), as well as in the Euro-Asian region (Rusanen et al. 1998). Cormorants have been observed in Estonia since the 1960s, and the first colony was established in 1984 in the west Estonian archipelago, Väinameri (Moonsund) (Lilleleht 1995). Approximately 800 pairs nested on Tondirahu islet in 1993 (Lilleleht 1995). The number of breeding pairs in the Väinameri region increased to 2675 in 1998 and 3500 in 1999; there are now four colonies. In other coastal areas of Estonia – the gulfs of Finland and Riga – the number of breeding pairs was only 323 and 920 in 1999. Concomitant with the increase of abundance of cormorants, commercial catches of fish (Vetemaa et al. 2002) and catch per unit effort in experimental fishing (Saat & Eschbaum 2002) decreased in the Väinameri region during the 1990s. Fishermen

*Correspondence: Redik Eschbaum, Institute of Zoology and Hydriobiology, University of Tartu, 46 Vanemuise Street, EE-51014 Tartu, Estonia, and Estonian Marine Institute, Tallinn, Estonia (email: [email protected]).

07 11/02/2003 13:17 Page 73

Interactions between cormorants and fishermen in the eastern Baltic

73

blamed the cormorants for declining catches in the Väinameri, but considerable changes have taken place in fisheries since the early 1990s. More people gained access to the fish resources, and fisheries and the export of fish became highly profitable in comparison to average salaries. This resulted in increased fishing effort (Vetemaa et al. 2000). Thus the role of cormorants in the decline of the fish stocks remains questionable. The aim of this study was to estimate the amount of fish eaten by cormorants and determine the extent of depredation on fish resources in the Väinameri area.

7.2 Study area Väinameri (Moonsund) is a shallow (average depth 10 m) archipelago area off the western coast of Estonia (Fig. 7.1). It includes more than 300 islands and islets (Sepp 1970), many of which are included in protected areas. Only a few, larger, islands are populated. Salinity in the Väinameri area is low, 5–6% on the average (maximum 8%). Water transparency depends on season and hydrometeorological conditions, and varies between 0.3 and 10 m (the average is 2–3 m) (Mardiste 1970). Ice conditions of the Väinameri are the toughest along the Estonian coast. Ice cover begins to form in shallow bays in the second half of November. On average (1963–1979), the whole Väinameri is ice covered from 1 January to 11 April (Anon 1982). The sheltered and shallow parts of the Väinameri, especially Matsalu Bay, are amongst the most important fish spawning and nursing areas along the Estonian coast. About 60 fish species have been found in the region, of which more than 40% are freshwater species (Saat & Eschbaum 2002). The composition of fish assemblages of

FINLAND

SWEDEN

ESTONA

4 LATVIA

Hiiumaa 2

3 1

Saaremaa 0

20 km

Figure 7.1 Location of cormorant colonies in Väinameri: 1, Tondirahu and Sipelgarahu islets; 2, Käina bay; 3, Langekare islet; 4, Kakralaid islet

07 11/02/2003 13:17 Page 74

74

Impact of birds on fisheries

the Väinameri depends on season, especially in coastal areas and bays. Several species are almost lacking during summer but abundant in other seasons, for example, winter spawners (burbot Lota lota (L.), sculpins Cottidae), marine species spawning in coastal areas (Baltic herring Clupea harengus membras L., smelt Osmerus eperlanus (L.) garfish Belone belone (L.)) and dace Leuciscus leuciscus (L.) and some other cyprinids foraging predominantly in deeper areas of Väinameri during summer. Fishing in the Väinameri is intensive, and fishing effort increased markedly in the early 1990s. The pound–net fishery for Baltic herring in the spawning period contributes the major part of the total catch. Traditionally perch, Perca fluviatilis (L.), was the second most important species in the catch but the stock collapsed (mainly due to overfishing) in the 1990s, and the proportion of cyprinids (roach Rutilus rutilus (L.), sil-ver bream Blicca bjoerkna (L.) and ide Leuciscus idus (L.)) has increased (Saat & Eschbaum 2002).

7.3 Material and methods The number of breeding pairs in the area was determined by counting nests in colonies and was estimated to be matched in number by non-breeders (E. Mägi & A. Leito, personal communication). The number of cormorants before and after the breeding period, and the number of breeders outside the Väinameri, were also counted. Daily food consumption of fish by cormorants was taken to be 423 g d1 during the 80-day long breeding season, and 238 g d1 outside the breeding season and for nonbreeders (Grémillet et al. 1995). In the Väinameri, the mean fish weight per pellet was 325  356 g (n  105), which is close to the average daily fish consumption by breeders and non-breeders estimated by Grémillet et al. (1995). Diet was determined in August 1997 and six times during 1998 from both pellets and regurgitated stomach contents (Table 7.1). Few pellets were found at the beginning of August 1998, and most (18 of 27) comprised mucus only. No pellets were found after this time, even at big roosts, because only fresh and complete pellets were collected and analysed.

Table 7.1 Number of diet samples analysed and the sampling date Date 01.08.97 09.04.98 17.04.98 14.05.98 06.06.98 02.07.98 01.08.98

Pellets

Stomach contents 4

41 30 14 12 27

37 34 21 32

07 11/02/2003 13:17 Page 75

Interactions between cormorants and fishermen in the eastern Baltic

75

Each sample was stored frozen and analysed separately. Remains of 23 fish species were identified in the diet of cormorants. The total number of fish in diet analyses was 1981. Identifiable fish remains were sorted out and the hard structures identified using a reference collection and the literature (Härkönen 1986; Veldkamp 1995b). For noncyprinid fishes, fish weight was determined from standard relationships (Härkönen 1986; Veldkamp 1995a, b; Dirksen et al. 1995; Macer 1966; Nienhuis 1997; Komissarov unpublished data; Table 7.2) between otolith size and weight, while for cyprinids, weight was calculated from using chewing pads and pharyngeal bones. No correction was made for digestion of the hard structure. Regurgitated fishes were collected during the breeding season. As the colony was big and well fed, completely undigested stomach contents were always found. The total length (TL) of regurgitated fish was measured to the nearest millimetre. Fish weight was calculated using regression formulae (Härkonen 1986; Veldkamp 1995a,b; Dirksen, et al. 1995; Pihu 1987; Böhling et al. 1991; Macer 1966; Nienhuis 1997; P. Komissarov, unpublished data; Table 7.2). The pellet analyses underestimate the share of smaller fish and fish species with smaller otoliths. On the other hand, regurgitated fish were only found during the breeding season. Differences between regurgitates and pellets may be partly due to the difference in diet between birds rearing young and birds without chicks. Veldkamp (1995a) suggested that large, heavy and rather ‘smooth’ species are most likely to be regurgitated. Using both pellets and regurgitates was considered to characterise better fish consumption of the population, thus both types of samples were used when calculating total consumption. The amount of fish m eaten by cormorants during a certain period was expressed as: m  kps, where k is the number of cormorants, p is the number of days in the period and s is average consumption of fish per day. As the number of non-breeding cormorants

Table 7.2

Relationships used to derive fish weight (FW) from hard structure

Species

Equation

r2

N

Herring Smelt Smelt Smelt Ide Bleak Silver bream Vimba Vimba Burbot Pike Viviparous blenny Flounder

FW  0.000004 TL3.1084 FW  0.2166 OL3.1728 FW  0.000006 TL2.9909 TL  39.63 OL  11.35 FW  4  106 TL3.1935 FW  8  106 FL3.0974 FW  8  106 TL3.0867 FL  40.797 CP  3.4449 FW  0.00001 TL2.9493 TL  52.257 OL  33.998 FW  3  106 TL3.1033 FW  9  107 TL3.2968 FW  5  105 TL2.7371

0.93 0.93 0.98 0.95 0.94 0.93 0.98 0.94 0.95 0.86 0.98 0.93 0.98

198 264 266 266 28 30 36 40 582 30 46 342 302

TL: total fish length (mm); FW: fish weight (g) OL: otolith length (mm); CP: chewing pad length (mm).

07 11/02/2003 13:17 Page 76

76

Impact of birds on fisheries

exhibits a Poisson distribution, 95% confidence limits of m were calculated as m  1.96 ps k To calculate annual consumption of different fish species by cormorants, the interval between samplings was divided into two equal periods. It was assumed that the feeding pattern (species composition and relative weight of different species) during the first period corresponded to the estimated made from the first sample, and feeding pattern during the second period to the next sample. Food consumption before the sampling period was calculated using data of the first sample, and food consumption after the sampling period using data from the last sample. Confidence limits (R;   0.05) of a species proportion in the total diet were calculated as R  t(  / 2 ); n1 DR, where t(/2); n1 is the (/2)-quantil corresponding to n  1 degrees of freedom of the Student’s t-distribution. Dispersion of a weight fraction DR was calculated as: DR 

1 ( DY  R2 DX  2 R cov( X , Y )) X2

( Traat & Inno 1997)

where Y is arithmetic mean weight of a species in samples, X is mean weight of food samples, D is dispersion. The confidence interval of fish amount consumed by cormorants during a particular period was correspondingly: ps( R  t(  / 2 ); n1 DR )( k  1.96 k ) Fisheries statistics and first buyer’s prices were obtained from the Fisheries Department of the Ministry of the Environment, based on logbooks of fishermen using commercial gear.

7.4 Results In 1998 there were 2692 pairs of breeding cormorants in the Väinameri area: 2522 on the neighbouring islets of Tondirahu and Sipelgarahu, 150 in Käina Bay and 20 on Kakralaid islet. In 1999 the number increased to 3453 pairs: 2943 on Tondirahu and Sipelgarahu, 335 in Käina Bay, 43 on Kakralaid and 132 in a new colony on Langekare islet (Fig. 7.1). Cormorants arrived in the Väinameri area in late March 1998. Between 25 March and 13 April the average number of cormorants was about 5300. The spring migration was completed by mid-April and the breeding period was considered to extend from 1 May to 17 July. In 1998 all juveniles left their nests before 1 August and thereafter the number of cormorants in the area decreased rapidly. Cormorants consumed approximately 460 t of fish in 1998 (Table 7.3). The total commercial catch of fish from the Väinameri in the same year was 1292 t, including 1071.5 t of Baltic herring. Food composition in the breeding period was different in pellets and regurgitate samples. Nineteen taxa were detected in pellets and 12 in stomach contents. The main

07 11/02/2003 13:17 Page 77

Interactions between cormorants and fishermen in the eastern Baltic

77

Table 7.3 The number of cormorants and estimated amount of fish consumed during 1998 Period 25.03–13.04 14.04–30.04 01.05–25.05 26.05–19.06 20.06–17.07 18.07–30.08 01.09–30.09 01.10–31.10 Total

Number of non-breeders 5 300 10 700 5 350 5 350 5 350 7 500 4 300 1 100

Number of breeders

Amount of consumed fish (t)

5350 5350 5350

25.23  0.68 43.29  0.82 88.41  0.85 88.41  0.85 99.02  0.96 80.33  1.82 30.70  0.92 8.12  0.48 463.5

prey species was viviparous blenny Zoarces viviparus (L.) in both pellets and stomach contents (43% and 41% of the total estimated fish weight). In pellets, burbot (24%) and roach (19%) followed viviparous blenny, but the share of these two species was only 2% and 8% in stomach contents. In regurgitates the other important species were Baltic herring and sticklebacks, Gasterosteidae (both with 16%), but Baltic herring was rare (1%) and otoliths of sticklebacks were not found in pellets. The diet of cormorants varied seasonally. It was more diverse in April (Table 7.4), when the most important prey were roach (approximately 50% of the total weight of the fish eaten) followed by burbot, pike Esox lucius (L.), pikeperch, Stizostedion lucioperca (L.), and perch (Table 7.4). All these species, except for burbot, enter shallow coastal areas for spawning during this time. During the breeding season, diet of cormorants comprised mainly viviparous blenny and burbot, followed by herring and roach. Sticklebacks were also consumed in big quantities. In August, the diet of cormorants consisted mostly of roach, burbot and blenny. The length distribution of roach consumed was similar to the length distribution of roach from experimental multi-mesh size gillnetting (Fig. 7.2). The mean total length of perch consumed by cormorants was 190 mm (experimental fishing 187 mm). In the case of pikeperch, pike and burbot, cormorants preyed on juvenile specimens (well below the legal minimum size for the fishery). Blenny, roach and burbot were consumed in largest quantities; the share of these three species was about 80% of the total fish weight eaten by cormorants. Also, herring, pikeperch, sticklebacks and perch were consumed in high quantities (Table 7.5). In some cases (e.g. burbot, pikeperch and smelt) predation of commercially important species by cormorants exceeded the commercial catch several fold (Table 7.5). The monetary value of fish eaten by cormorants (2.1 million Estonian crowns, EEK, based on average first buyer’s prices; 15.6 EEK  €1) was approximately 50% of the value of the commercial catch (4.2 million EEK). However, when herring is excluded, the value of coastal fish eaten by cormorants (2.0 million EEK) exceeded the value of fish captured by fishermen (1.6 million EEK).

Species Herring Clupea harengus Smelt Osmerus eperlanus Pike Esox lucius Eel Anguilla anguilla Roach Rutilus rutilus Ide Leuciscus idus Rudd Scardinius erythrophthalmus Bleak Alburnus alburnus White bream Abramis bjoerkna Bream Abramis brama Vimba bream Abramis vimba Gibel carp Cyprinus gibelio

9 April 0.0

17 April

14 May

3.98  0.3

12.9  1.3

3.6  0.2

2.0  0.7

0.3

9.2  1.3

0.2

0.1

62.5  1.4 1.3

47.7  1.7 2.1

6 June 1.1  0.2

0.1

0.3

1.0  0.1

1.6

1.8

8.3  1.2

1 August 2.6

2.7 3.8

0.6

2.5

1.2  1.1 0.4  0.3

2 July

1.2 0.8

14.4  2.5

59.3  5.3

Impact of birds on fisheries

Diet composition of cormorants (% of the total diet weight)

07 11/02/2003 13:17 Page 78

78

Table 7.4

13.7  1.9

3.8

0.2

11.0  3.4

23.6  8.2

2.4  0.2

4.9  0.6

4.8  4.4

0.021

0.6

0.2  0.2

0.3  0.04

0.3  0.2

0.04  0.04

0.3  0.4

15.2

8.6

2.9  0.6

7.0  0.6

9.1  8.3

10.2  0.6

6.5  0.4

1.3

3.1  0.3

0.3  0.3

2.9  0.7

0.3  0.04 0.1  0.03

48.3  2.3 0.006

76.9  2.4

52.0  2.2

17.9  21.5

14.3  1.6 0.2 2.0  0.4

0.2  1.3 0.5 2.8  2.8

0.2

07 11/02/2003 13:17 Page 79

1.9  0.6

Interactions between cormorants and fishermen in the eastern Baltic

Cyprinids Cyprinidae sp. Burbot Lota lota Three-spined stickleback Gasterosteus aculeatus Nine-spined stickleback Pungitius pungitius Pikeperch Stizostedion lucioperca Perch Perca fluviatilis Ruffe Gymnocephalus cernuus Viviparous blenny Zoarces viviparus Greater sandeel Hyperoplus lanceolatus Gobiidae Bullhead Cottus gobio Flounder Platichthys flesus

79

07 11/02/2003 13:17 Page 80

Impact of birds on fisheries 18 16 14

cormorants test-fishing

12 10 8 6

310

290

270

250

230

210

190

170

150

130

110

90

70

4 2 0 50

Number of fish (%)

80

Total length (mm)

Figure 7.2 Length–frequency distribution of roach eaten by cormorants (n  208) compared with length–frequency distribution of roach in test-fishing (n  1456)

Table 7.5. Estimated consumption of different fish species by cormorants and commercial catch from the Väinameri in 1998 Species Herring Smelt Pike Eel Roach Ide Rudd Bleak Silver bream Bream Vimba Gibel carp Unidentified cyprinids Burbot 3-spined stickleback 9-spined stickleback Pikeperch Perch Ruffe Viviparous blenny Great sandeel Gobies Bullhead Flounder Other Total a

Eaten by cormorants 95 CI (t)

Commercial fishery (t)

Ratio cormorants to fisheries

25.36  9.4 1.75  3.2 4.07  4.1 3.21  0.048 125.11  63.1 3.41  0.028 0.53  0.8 0.13  0.1 0.45  0.3 1.19  0.009 0.52  0.004 0.68  0.006 4.33  0.037 72.44  55.9 10.26  4.5 0.92  1.6 12.80  7.4 6.70  3.2 3.85  6.6 179.43  36.1 0.28  0.004 2.55  7.1 0.67  0.007 2.73  1.6

1071.50 0.50 7.66 4.48 81.48 44.63

0.02 3.5 0.53 0.72 1.54 0.08

463.37

Recorded as roach in fisheries statistics.

a a

0.61 2.44

1.95 0.21

2.00

36.22

2.01 9.87 – 0.01 –

6.37 0.68 – 17940 –

– 15.04 49.64 1291.87

– 0.18 0.36

07 11/02/2003 13:17 Page 81

Interactions between cormorants and fishermen in the eastern Baltic

81

7.5 Discussion There are several estimates of daily fish consumption by cormorants in the literature. In most cases these are based on back-calculated fish weight in pellets. Several authors have criticised using pellets in daily food intake calculations (Duffy & Laurenson 1983; Grémillet & Plös 1994). In the Väinameri, the mean fish weight per pellet was 325 g, which is close to the average daily fish consumption by breeders (423 g) and non-breeders (238 g) estimated by Grémillet et al. (1995). However, the daily intakes calculated did not coincide with the logical energy requirements of cormorants, being highest in April, lowest in August and average during the breeding season. Lower values for the calculated daily intake during the breeding season could be explained by these pellets belonging mostly to non-breeders. Breeders deliver a large fraction of the fish they capture to chicks, which do not produce pellets (van Dobben 1952; Trauttmansdorff & Wassermann 1995), and the pellets of parents are more difficult to find, because they contain of fewer fish remains (Veldkamp 1995a). The food composition of cormorants varied seasonally. In general, more abundant and weak-swimming species prevailed in the diet. Before and at the beginning of the breeding period, the bulk of the diet comprised species entering shallow coastal areas for spawning and species more easily caught by social (group) fishing. Seasonal difference in diet and foraging closer to the coast in spring is probably related to water transparency; turbidity is highest in the Väinameri during the spring. During the breeding period, relatively slow-swimming, bottom-living species (viviparous blenny, burbot) and the most abundant species (roach, sticklebacks) prevailed in the diet. Length distributions of roach and perch were almost identical in the diet and in multimesh, experimental gillnetting. However, in the case of larger species (pikeperch, pike, burbot), smaller juvenile specimens (usually well below the minimum legal size for fishery) prevailed in the diet of cormorants. In several cases (including some commercially very important species) predation by cormorants was almost equal to or exceeded fishing mortality. Pikeperch is a valuable species in the region but the stock is in poor condition. Commercial catch in 1998 was approximately 4000 specimens while cormorants consumed over 100 000 juveniles. Burbot does not belong to the list of most commerciallyimportant species in the area, although about 4000 specimens were taken by fishermen in 1998. However, due to the current low abundance of other predators (pikeperch, pike, larger perch), this species is important in the fish community, but cormorants were estimated to consume over 250 000 specimens in 1998. It was concluded that in view of the present state of the commercially important stocks, cormorants compete for fish resources with fishermen in the Väinameri. However, one should keep in mind that in the early 1990s (when the number of cormorants was low) fishing effort increased substantially in the area, and stocks of the most valuable species were overexploited. Commercial catch of perch reached 639 t in 1991 and that of pikeperch 34 t in 1993. If stocks of these commercial species were in good shape, cormorants would probably have little affect because they tend to catch slow-swimming species, such as viviparous blenny and sticklebacks, which are the most important prey for predatory fish in the Väinameri (R. Eschbaum, unpublished data). Thus there

07 11/02/2003 13:17 Page 82

82

Impact of birds on fisheries

is likely to be considerable competition between cormorants and predatory fish if the latter stocks improve.

Acknowledgements We thank Eve Mägi and Aivar Leito for providing data on the abundance of cormorants, and Margus Pihlak for advice on statistics. This study was supported by the Estonian Fisheries Fund (grant to R. Eschbaum) and by the Estonian Ministry of Education (project ‘Biodiversity of the Estonian fauna’ to T. Saat).

References Anon (1982) Climatological Ice Atlas for the Baltic Sea, Kattegat, Skagerrak and Lake Vänern (1962–1979). Norrköping: SMHI, 220 pp. Böhling P., Hudd R., Lehtonen H. & Parmanne R. (1991) Fiskevården i havsområdet utanför Jakobstad. Fiskundersökningar 16. Helsingfors: Vilt-och fiskeriforskningsinstitutet, 50 pp. (in Swedish). Dirksen S., Boudewijn T.J., Noordhuis R. & Marteijn E.C.L. (1995) Cormorants Phalacrocorax carbo sinensis in shallow eutrophic freshwater lakes: prey choice and fish consumption in the non-breeding period and effects of large-scale fish removal. Ardea 83, 167–184. Duffy D.C. & Laurenson L.J.B. (1983) Pellets of Cape cormorants as indicators of diet. Condor 85, 305–307. Grémillet D., Schmid D. & Culik B. (1995) Energy requirements of breeding great cormorants Phalacrocorax carbo sinensis. Marine Ecology Progress Series 121, 1–9. Gremillet D. & Plös A. (1994) The use of stomach temperature records for the calculation of daily food intake in cormorants. Journal of Experimental Biology 189, 105–115. Härkönen T. (1986) Guide to the Otoliths of the Bony Fishes of the Northeast Atlantic. Hellerup: Danbiu ApS., 256 pp. Lilleleht V. (1995) Veel kormoranidest. Eesti Loodus 2, 44–46 (in Estonian). Lõhmus A., Kuresoo A., Leibak E., Leito A., Lilleleht V., Kose M., Leivits A., Luigujõe L. & Sellis U. (1998) Eesti lindude staatus, pesitsusaegne ja talvine arvukus. Hirundo 11, 63–83 (in Estonian). Macer C.T. (1966) Sand eels (Ammodytidae) in the southwestern North Sea; their biology and fishery. Ministry of Agriculture, Fisheries & Food, Fisheries Investigation Series 2, 24(6), 1–55. Mardiste H. (1970) Väinameri. In E. Kumari (ed.) Lääne-Eesti rannikualade loodus. Tallinn: Valgus, pp. 7–16 (in Estonian). Nienhuis J. (1997) Food choice of non-breeding cormorants in Matsalu Bay in spring: are cormorants and smews competitors? In E. Mägi & K. Kaisel (eds) Loodusvaatlusi 95/96. Tallinn: Matsalu Looduskaitseala, pp. 34–42 (in English with Estonian summary). Pihu E. (1987) Matk kalariiki. Tallinn: Valgus, 360 pp. (in Estonian). Rusanen P., Mikkola-Roos M. & Asanti T. (1998) Merimetso Phalacrocorax carbo – musta viikinki. Merimetson kannan kehitys ja siihen vaikuttavat tekijät Itämeren piirissä ja Euroopassa. Suomen ympäristo 182, 183 pp. (in Finnish). Saat T. & Eschbaum R. (2002) Väinamere kalastik ja selle muutused viimastel aastakümnetel. In T. Saat (ed.) Väinamere kalastik ja kalandus. Tartu: Tartu Ülikooli Kirjastus, pp. 9–45 (in Estonian with English summary). Sepp U. (1970) Väinamere saared. In E. Kumari (ed.) Lääne-Eesti rannikualade loodus. Tallinn: Valgus, pp. 17–26 (in Estonian).

07 11/02/2003 13:17 Page 83

Interactions between cormorants and fishermen in the eastern Baltic

83

Traat I. & Inno J. (1997) Tõenäosuslik valikuuring. Tartu: Tartu Ülikooli Kirjastus, 211 pp. (in Estonian). Trauttmansdorff J. & Wassermann G. (1995) Number of pellets produced by immature cormorants Phalacrocorax carbo sinensis. Ardea 83, 133–134. Van Dobben W.H. (1952) The food of the cormorant in The Netherlands. Ardea 40, 1–63. Veldkamp R. (1995a) Diet of cormorants Phalacrocorax carbo sinensis at Wanneperveen, The Netherlands, with special reference to bream Abramis brama. Ardea 83, 143–155. Veldkamp R. (1995b) The use of chewing pads for estimating the consumption of cyprinids by Cormotants Phalacrocorax carbo. Ardea 83, 135–138. Vetemaa M., Eero M. & Järv L. (2002) Väinamere kalandus: püügistatistika ja sotsiaalmajanduslikud aspektid. In T. Saat (ed.) Väinamere kalastik ja kalandus. Tartu: Tartu Ülikooli Kirjastus, pp. 9–45 (in Estonian with English summary). Vetemaa M., Eschbaum R., Aps R. & Saat T. (2000) Collapse of political and economical system as a cause for instability in fisheries sector: an Estonian case. Proceedings of the IIFET 2000 International Conference. Microbehaviour and Macroresults. Oregon State University July 10–14 Corvallis, USA (in press).

08 11/02/2003 13:35 Page 84

Chapter 8

Interactions in the utilisation of small fish by piscivorous fish and birds, and the fishery in IJsselmeer P.J. MOUS* Fish Culture and Fisheries Group, Wageningen Institute of Animal Science, Wageningen University, The Netherlands

W. DEKKER, J.J. DE LEEUW Netherlands Institute for Fisheries Research, IJmuiden, The Netherlands

M.R. VAN EERDEN Ministry of Transport, Public Works and Water Management, Directorate Flevoland, Lelystad, The Netherlands

W.L.T. VAN DENSEN Netherlands Institute for Fisheries Research, IJmuiden, The Netherlands

Abstract To assess the carrying capacity of IJsselmeer as a foraging area for piscivorous fish and piscivorous birds, and as a fishing ground for a commercial fishery, patterns in production and utilisation of small fish (fish 10 cm TL) over a 20-year period are analysed. Total biological production of small fish was 102 kg ha1 yr1 in the northern basin and 116 kg ha1 yr1 in the southern basin. The most important categories of small fish species, in terms of production, were smelt Osmerus eperlanus (L.), ruffe Gymnocephalus cernua (L.) and juvenile perch Perca fluviatilis (L.). Throughout the period 1970–1995, young-ofthe-year (0-group) smelt was a keystone species in IJsselmeer, contributing 72% to the total small fish production in the northern basin, and 60% to the total production of small fish of the southern basin. Inter-annual variation in production of 0-group smelt was relatively low, 67% of the annual production estimates varied within a factor of 2 around the inter-annual geometric mean. Smelt production was probably balanced by its utilisation. Most of the smelt production was utilised by piscivorous fish (mainly perch), followed by the fishery for spawning smelt and the piscivorous birds around IJsselmeer. Growth of zooplanktivorous smelt and benthivorous ruffe was density-dependent, and fast growth of smelt was associated with a large mean length of its preferred food, Daphnia spp. As production of smelt, the most important producer, was fully utilised, and since its growth was density-dependent, the capacity of IJsselmeer to sustain small fish production for utilisation by piscivorous fish, piscivorous birds and the fishery was probably reached. Management regimes that aim for a higher biological production of piscivorous fish should either include a restriction on the commercial fishery on smelt, or allow that food availability of piscivorous birds decreases. Keywords: density-dependence, fishery, perch, production, ruffe, smelt.

*Correspondence: P.J. Mous, Fish Culture and Fisheries Group, Wageningen Institute of Animal Science, Wageningen University, P.O. Box 338, 6800 AH Wageningen, The Netherlands (email: [email protected]).

08 11/02/2003 13:35 Page 85

Interaction in the utilisation of small fish by piscivorous fish and birds

85

8.1 Introduction Small-sized fish (10 cm TL) are an important component of lake ecosystems. Often, they comprise most of the fish production, and average stock biomass is high compared to that of large and older fish (Sprules et al. 1991; Downing & Plante 1993; Randall et al. 1997; Holcik 1996). Many species of small fish are zooplanktivorous, and their predation pressure can ultimately affect water quality through a cascading effect (Carpenter et al. 1985; Hanson & Butler 1994). In IJsselmeer, predatory fish, piscivorous birds and the commercial fishery depend on the production of small fish, so knowledge of the capacity of IJsselmeer to sustain production of small fish is of vital importance for the management of IJsselmeer’s natural resources. The concept of carrying capacity is closely related to production and utilisation patterns. However, there is no consensus on the meaning of carrying capacity (De Bie 1991), although Begon et al. (1990) provided a useful definition: ‘The population density where production is balanced by mortality (utilisation), in populations where production or mortality are densitydependent’. In this chapter, the carrying capacity of IJsselmeer for the production of small fish is assessed by balancing production against utilisation, and by studying density-dependency in the main categories of small fish. From a steady-state flow model that is representative for an average situation in IJsselmeer over the period 1983–1987 (Buijse et al. 1993), it is clear that the zooplanktivorous smelt, Osmerus eperlanus (L.), is the most important small fish species in terms of biological production in IJsselmeer (130 kg ha1 yr1). Other important small fish species are the benthivorous ruffe, Gymnocephalus cernuus (L.), (9 kg ha1 yr1) and juvenile perch, Perca fluviatilis (L.) (5 kg ha1 yr1). The production of smelt, ruffe and juvenile perch together contributes about 80% to the total fish production of IJsselmeer. Consumption of small fish by piscivorous fish was 81 kg ha1 yr1, of which 83% is consumed by perch, 12% by pikeperch, Stizostedion lucioperca (L.), and 4% by eel, Anguilla anguilla (L.). Small fish consumption by piscivorous birds was 48.4 kg ha1 yr1, of which 43% was consumed by cormorant, Phalocrocorax carbo (L.), 21% by blackheaded gull, Larus ridibundus (L.), and the remainder by 10 other bird species. The commercial fishery utilises small fish, as it targets spawning concentrations of smelt in early spring (catch 9 kg ha1 yr1). During the rest of the year, the fishery utilises small fish production indirectly by targeting consumers of small fish, namely large-sized eel, pikeperch and perch, for which the catch amounts to 3, 1 and 4 kg ha1 yr1 respectively. Although not targeted for, most of the discards of the summer fishery with fyke nets for eel consist of small fish (van Dam et al. 1995). Hence, a more intensive exploitation of small fish by the fishery, either directly or indirectly, may have a negative impact on the outcome of the fishery for eel, perch and pikeperch, and on the viability of bird populations. Production of small fish might strongly vary between years due to year class strength variation: in IJsselmeer, year class strength of perch varied by a factor of 400, and of pikeperch by a factor of 70 (Buijse et al. 1992b). In this study, inter-annual variation in the biological production of small fish over a period of c. 20 years is described using data from trawl surveys. Utilisation of small fish by the fishery and by fish-eating birds, and the temporal variations therein, were assessed through fish auction statistics and published reports.

08 11/02/2003 13:35 Page 86

86

Impact of birds on fisheries

For perch and pikeperch, water temperature plays an important role in determining year class strength and growth (and hence production), but the size of the parental stock does not (Buijse et al. 1992b). Density dependent effects on biological production of small fish were studied at the level of the size of the parental stock and growth. Growth and condition of smelt were studied in relation to seasonal and inter-annual fluctuations in availability of high-quality food for smelt, as indexed by the mean Daphnia spp. size in the lake.

8.2 Materials and methods Intra-annual variation in abundance and body size of small fish was assessed with a repetitive trawl survey programme between May 1994 and March 1995. Inter-annual variation was assessed with a yearly trawl survey that was conducted each autumn from 1970 to 1994. All survey hauls were made during daytime. After completion of each survey haul, the geographical position, haul duration (hr), bottom depth (m) and Secchi depth (cm) were recorded. All fish were measured (total length (TL) mm), or recalculated to total length. Allometric length–weight relationships were used to calculate body weight from total length (Table 8.1). A length–frequency distribution (LFD) was established for each species in the catch. To separate the 0-group (the young-of-the-year) from the 1 age-group, the LFDs were averaged per survey, per sampling period and per basin, and the average LFDs were subjected to Bhattacharya analysis (Sparre & Venema 1992) using the software package FiSAT (Gayanilo et al. 1995). For ruffe that were sampled during the winter half-year, separation into age groups was possible only by assuming that the left part of the LFD consisted exclusively of 0-group, because the modes of the 0-group and 1-group were not discernable. Using the fractional distribution of age classes over each length class, the number of 0-group and 1-group fish in each survey trawl haul (CPUE, number haul1), and successively fish densities (number ha1) were calculated (see Appendix).

8.2.1

Intra-annual survey

This survey was conducted during four time periods (May 1994, August–September 1994, November–December 1994 and March 1995), using a small, fine-meshed beam trawl (mouth width 3 m, mouth height 0.6 m, stretched-mesh size in cod-end 2 mm). In total, 332 hauls of 10 min each were made, during which a distance of about 900 m was covered. Hauls were randomly positioned in two depth strata (areas deeper and shallower than 5 m) and seven geographic strata (three in the northern basin, four in the southern basin; Fig. 8.1). The allocation of observations to strata was proportional to the surface area of each stratum. To ensure sufficient coverage within each stratum, only locations at full geographical minutes were allotted a probability of being sampled (cf. Jolly & Hampton 1990). See Table 8.A1 for survey details.

08 11/02/2003 13:35 Page 87

Interaction in the utilisation of small fish by piscivorous fish and birds

87

Table 8.1 Parameter values for allometric length– weight relationships (W  a TLb). W  body weight (kg), TL  total length (cm). Parameter values were estimated by regression analysis on length–weight observations (n  700, r2  0.98)

Smelt Perch Ruffe

a

b

2.30  106 4.70  106 6.48  106

3.37 3.36 3.29

Figure 8.1 Map of IJsselmeer, with stratum boundaries of the intra-annual survey and sampling locations of the inter-annual survey

08 11/02/2003 13:35 Page 88

88

Impact of birds on fisheries

During the first survey in May 1994, the bottom trawling was observed not to catch any 0-group fish, and it was suspected that 0-group fish inhabited the upper part of the water column only (Mous 2000). Therefore, 12 additional hauls were made at the surface, positioned randomly over the middle stratum of the northern basin. It was assumed that the volume swept by the surface hauls represented the 0-group density in the upper half of the water column. The March 1995 survey was conducted just before and during the smelt fishing season. To allow for near-shore spawning concentrations of smelt, the March 1995 survey was stratified in a near-shore (2 km from the shoreline) and an open-water stratum. In the open-water stratum, the hauls were randomly positioned using the same procedure as in the other intra-annual surveys. In the near-shore stratum, one haul was randomly positioned within 5-km stretches of shoreline. Towing direction was always perpendicular to the shore. Preliminary analysis showed that only in the southern basin significantly more smelt in the near shore stratum were caught (ANOVA on ln-transformed CPUE, P  0.001). Therefore, CPUE observations from the near shore and the open water stratum were pooled for the northern basin, whereas CPUE observations from the southern basin were weighted according to the surface area of each stratum in subsequent analyses. To assess the effect of the smelt fishery on smelt density, 12 sampling locations in the northern basin that were sampled before the beginning of the smelt fishing season (10 March) were sampled again on 9 April, after the closing of the fishing season. The intra-annual survey yielded estimates of variation in stock density, mortality and individual body weight of each age group of small fish over the period May 1994–March 1995.

8.2.2

Inter-annual survey

A bottom trawl was used to estimate inter-annual variation in density and size composition of small fish. The cod-end of the bottom trawl used had a 1.8 cm mesh size (stretched-mesh), mouth width of 8 m, and height of 1 m. Towing speed was about 1.5 m s1. In total, 533 hauls of varying duration (10–45 min haul1) were made. After 1982, haul duration was standardised (10 min haul1). Sampling took place at fixed locations throughout the period 1966–1994 (Fig. 8.1), but the number of observations varied between years. Each survey was completed within a 4-week period in October–November.

8.2.3

Production

Biological production (P) during the growing season was estimated per age group as – the product of average biomass (B ) during the growing season and productivity (P/B), which productivity equals the instantaneous growth rate G, when assuming exponential – growth in individual weight during the season (Ricker 1975) : P  G B [formula 1]. The instantaneous growth rate (G) was estimated from ln(We/Wb), where Wb and We are the individual weights at the beginning and end of the growing season (June–November).

08 11/02/2003 13:35 Page 89

Interaction in the utilisation of small fish by piscivorous fish and birds

89

Average biomass was approximated with NhW(Lh), where Nh is fish density (ind ha1) mid-growing season and Wh is individual weight (kg) corresponding with fish length (cm) mid-growing season: Lh  (Le  Lb)/2. Nh was reconstructed from fish density as recorded during the survey in autumn (Ne) and mortality rate Z during the growing season viz: Nh  Ne eZ/2. For 0-group fish Z was estimated from density estimates during early summer, late summer and autumn surveys. Mortality rates for 1-group fish were estimated from density estimates during the fall surveys: Z  ln(N0-group  N 1-group)yeart1  ln(N 1-group)yeart. It was assumed that all mortality took place during the growing season. For 1-group smelt in the period 1982–1994, when a fyke net fishery for spawning smelt was operated in early spring of every year, instantaneous fishing mortality (F) was subtracted from instantaneous total mortality rate (Z). This F was calculated as the difference in Z per year between the reference period 1971–1981 and the period 1982–1994 (ignoring variation through possible density dependent mortality in 1-group smelt).

8.2.4

Utilisation

Perch and pikeperch are the main fish predators in IJsselmeer, whereas fish consumption by eel is relatively low (Koenders & Mentink 1982; Buijse et al. 1993; Dekker & van Willigen 1993; Paulisse 1993). An estimate for small fish consumption by perch and pikeperch was derived from the catch statistics of the commercial fishery. Perch and pikeperch in the commercial catch are all piscivorous. If the CPUE remains constant over a period of years, biological production of the exploited stock must at least equal the commercial catch (Baranoff 1918), and a conservative estimate of consumption then equals the commercial catch multiplied by the inverse of the commonly used gross food conversion efficiency of 0.2 for fresh weight (Pauly 1986). As fishing mortality (F) is much higher than natural mortality for these intensively exploited piscivores (Dekker & Schaap 1993; Dekker 1996), the underestimation due to this approximation method will be small, but the consumption of piscivores having not yet reached commercial size should be accounted for. Perch first appear in the commercial catch at a length of about 25 cm TL (Buijse et al. 1992a), whereas perch are piscivorous at about 18 cm TL (Buijse & van Densen 1992; van Densen et al. 1996). Pikeperch is piscivorous at 10 cm TL (Buijse & Houthuijzen 1992; van Densen et al. 1996), whereas the minimum length in the catch is 37 cm TL (Buijse et al. 1992a). A simple biomass/production model (i.e. formula [1]) with Fperch  0.86 yr1, Fpikeperch  1.5 yr1 (Dekker & Schaap 1993; Dekker 1996), M  0.2 yr1 and growth parameters as in Buijse et al. (1992a), predicted that underestimation of consumption amounted to a factor 1.2 for perch and 2.9 for pikeperch. These factors were used to correct the consumption estimates. Utilisation of small fish and prey selection by piscivirous birds in IJsselmeer has been extensively studied (Doornbos 1979; Schouten 1983; van Eerden et al. 1991; Voslamber 1991; Buijse et al. 1993; van Eerden 1993; Beekman & Platteeuw 1994; Winter 1994; Prins et al. 1995; Stam 1996), and basic data from these studies are below. The commercial catch was estimated through auction statistics. Fishermen were legally obliged to trade their catch through the auction before 1975, but about 10% of

08 11/02/2003 13:35 Page 90

90

Impact of birds on fisheries

the catch was illegally sold directly to wholesalers (Nagtegaal & Snel 1984). After 1975, trading via the auction was no longer obliged, but this policy change was not reflected in a marked change in the auction statistics of the commercially-important eel. Hence, it was assumed that auction statistics continued to represent a constant proportion of the total catch. A relatively large part of smelt caught is sold directly to wholesalers, and therefore the catch volume sold directly in 1990–1993 and 1995 was estimated by interviewing the most important wholesalers. To estimate the quantity and species composition of small fish caught as discards in eel fyke nets, catch composition of experimental fyke net catches was determined in 1983 and 1987. In 1993, a creel survey was carried out to estimate the volume of discards (Dekker et al. 1993).

8.2.5

Density-dependency

As smelt and ruffe reach sexual maturity during their first year (Willemsen 1977), the stock size of smelt and ruffe in autumn is a measure of size of the spawning stock that will produce the 0-group fish of next year. To investigate whether the 0-group densities of smelt and ruffe were affected by parental stock size, 0-group fish density was plotted against size of parental stock of the previous year, using density estimates from the inter-annual survey. Population density and growth of smelt, ruffe, and perch were subjected to correlation analysis to assess possible density-dependency of production in the two basins of IJsselmeer. Also, population density and growth were correlated with average water temperatures in the second and the third quarters of the year (daily water temperature data from the Netherlands Ministry of Transport, Public Works and Water Management). The subdivision of the summer half-year in two parts was considered meaningful, since the average water temperatures of the two periods were not correlated (Pearson correlation coefficient was 0.02 and 0.03 for the northern and the southern basin respectively).

8.2.6

Food quality, growth and condition

The correlation of mean length of smelt and 0-group perch as attained in autumn, with food quality during the growing season as indexed by the average body size of Daphnia spp. in the lake, was investigated. The size distribution of Daphnia spp. was monitored in the northern and the southern basin of IJsselmeer in the years 1987, 1988, 1989 (Zuilekom 1991) and 1992 (Prins et al. 1995). Because mean length of 0-group smelt varied within a narrow range (5.9–7.9 cm), variation in the mean length was increased by including data on size and density of 0-group smelt and Daphnia spp. from Tjeukemeer, a lake close to IJsselmeer (Fig. 8.1). In IJsselmeer, the population of large-bodied Daphnia spp. is high until June, after which the population collapses due to a high predation pressure of zooplanktivorous fish (Prins et al. 1995; Veen et al. 1997; Lammens 1999). To assess whether this resulted in a decrease in condition of smelt during the growing season, the seasonal variation in

08 11/02/2003 13:35 Page 91

Interaction in the utilisation of small fish by piscivorous fish and birds

91

condition of smelt was studied. The percentage dry weight of 0-group and 1-group smelt that were sampled in the northern basin of IJsselmeer in May, June, August, September and November 1991 was measured. For each length class (0.5 cm width) available in the catch, a sample of fish weighing in total up to about 150 g was collected. The sample was homogenised using a ultrathorax. From the homogenised sample, replicates of 5–10 g material were dried during 10 hours at 70°C, and successively for 4 hours at 103°C. After being placed for 2–3 h in a dessicator, the remaining material was weighed and dry weight was calculated. If the replicates differed more than 5% from each other, the analysis was repeated. The mean of the replicate observations was used in the analysis. A statistical model that explains variation in dry weight by sampling period was constructed to describe the seasonal variation in condition. Since condition of fish varies with length (Miranda & Hubbard 1994; Liao et al. 1995), length was included as a covariate in the model.

8.3 Results 8.3.1

Intra-annual variation in density and body size

For all species and age groups studied in 1994, most growth occurred between May and September (Table 8.2); little growth occurred between December and April the following year. The 0-group smelt, having an average size of 2.2 cm TL, and perch, with an average size of 1.2 cm TL, were already present in large numbers in the survey catches of May 1994. The 0-group ruffe were probably not fully recruited at that time, as they were caught in only three of 12 surface hauls, and in none of the bottom hauls. Absolute mortality (in terms of numbers) was low during winter (November–March) compared with summer (May–September). This was probably the result of the slow growth and hence low predation pressure of piscivorous fish in winter. Also, fishing mortality due to discarding was nil during winter, since the eel fyke fishery operates only in summer. The most important cause of winter mortality must have been predation by piscivorous birds (see below). For the calculation of biological production, it was assumed that instantaneous mortality was constant during the growing season (20 May to 28 November), and that all mortality occurred in this period.

8.3.2

Inter-annual variation in body size

The 0-group smelt reached a mean length of 6–8 cm and 1-group smelt a length of 8–11 cm (Fig. 8.2). Since the building of the dyke that separates the northern basin from the southern basin (1975), 0-group smelt in the northern basin were on average 10% (0.7 cm) longer than 0-group smelt from the southern basin. The 1-group smelt differed by 15–20% (1.7 cm). The difference in length after dyke construction was due to the increase in mean length of smelt in the northern basin from, on average, 6.8 cm

N  103

CI

L

N

CI

L

N

Northern 20 May 94 1 Sep 94 28 Nov 94 23 Mar 95

(134) 132a 32 23b

(2.0) 1.6 1.4 1.3

(2.2) 5.9 6.7 6.7

430 87 53 38b

1.9 1.3 1.5 1.3

8.4 10.4 11.3 10.5

(4500) (2.5) 37 1.5 33 1.5 38 1.3

1.5 1.3 1.5

6.7 7.3 7.3

430 299 119b

1.4 1.3 1.5

8.4 9.2 9.2

Southern 6 Sep 94 13 Dec 94 23 Mar 95

34a 5.1 2.3b

CI

48 78 46

1.6 1.6 1.1

1-group ruffe

0-group ruffe

0-group perch L

N  103

CI

L

N

CI

L

(1.2) 8.3 8.0 8.4

0.00 0.56 1.1 0.93

– 1.5 1.8 1.3

1.0 7.0 7.6 8.3

331 83 83 95

1.3 1.7 2.0 1.3

10.3 11.5 11.7 11.7

6.4 7.2 7.3

0.64 2.2 2.4

1.4 1.7 1.3

8.4 8.7 8.9

61 339 431

1.4 1.6 1.3

10.1 10.4 10.6

The estimate was corrected for the water transparency-mediated vertical distribution pattern of smelt c. 7 cm TL (0-group). The estimates were corrected for fishery mortality due to the smelt fykenet fishery.

a

b

Impact of birds on fisheries

1-group smelt

0-group smelt Basin/ Date

08 11/02/2003 13:35 Page 92

92 Table 8.2 Mean density (N, ha1) and mean total length (L, cm) of the major categories of small fish in the northern and southern basin during May 1994–March 1995. The dates represent the average date of each sampling period. The 95% confidence interval of the means (CI) is presented as a factor, i.e. the upper confidence limit for mean x with CI y is xy; the lower confidence limit is x/y. Estimates between brackets were based on 12 surface trawl hauls, the other estimates were based on the bottom trawl surveys. The coefficient of variation in length was approximately 10%

08 11/02/2003 13:35 Page 93

Interaction in the utilisation of small fish by piscivorous fish and birds

Total length (cm)

12

93

Smelt

10 8 6 4 1966

1970

1974

1978

1982

1986

1990

1994

Year

Figure 8.2 Mean total length of 0-group (—) and 1-group (- - -) smelt in the northern basin () and the southern basin () in autumn. The arrow indicates the year when the dyke that separates the northern basin from the southern basin was completed (1975)

Total length (cm)

10

Perch

9 8 7 6 5 1966

1970

1974

1978

1982

1986

1990

1994

Year

Figure 8.3 Mean total length of 0-group perch in autumn. For further explanation, see caption of Figure 8.2

(1966–1975) to 7.2 cm (1976–1994). For 1-group smelt, mean length increased from 9.4 to 10.4 cm. The differences were significant for both 0-group and 1-group smelt (two-tailed t-tests, P  0.04 and P  0.01). The mean total length of 0-group perch averaged 7.5 cm (Fig. 8.3). There was no noteworthy difference in length between the northern and the southern basin. Ruffe (Fig. 8.4) reached a mean total length of 6–8 cm in their first growing season. The mean length of 1-group ruffe varied between 9 and 12 cm. After dyke construction, 0-group ruffe were on average 0.9 cm longer, and 1-group ruffe were on average 1.2 cm longer in the northern basin than in the southern basin. Also here, the difference originated from a faster growth after dyke construction in the northern basin: the difference for 0-group ruffe was 1.2 cm and for 1-group ruffe 1.4 cm (twotailed t-tests, P  0.01).

08 11/02/2003 13:35 Page 94

94

Impact of birds on fisheries Ruffe

Total length (cm)

14 12 10 8 6

4 1966

1970

1974

1978 1982 Year

1986

1990

1994

Figure 8.4 Mean total length of 0-group and 1-group ruffe in autumn. For further explanation, see caption of Figure 8.2

Number (ha⫺1)

105 104 103 102 10 1

Number (ha⫺1)

Northern basin 0-group smelt

105 104 103 102 10 1

Southern basin (a)

(b)

(c)

(d)

>=1-group smelt

70

74

78

82

86

90

94

Year (19..)

70

74

78

82

86

90

94

Year (19..)

Figure 8.5 Mean smelt density (ha1), with 95% confidence limits of the mean. A reference line is indicated at a density of 10 000 ind. ha1

8.3.3

Inter-annual variation in density

Population density in fall differed significantly between years for each species (ANOVA for each species and basin: r2  0.25; P  0.001). The models fitted the observations reasonably well, as only two residual distributions were different from normal (Wilk– Shapiro statistic 0.95, P  0.05). First, the residual distribution for 1-group smelt from the northern basin was skewed to the left. Secondly, a single outlier caused deviation from normality for 0-group ruffe from the southern basin. The density of 0-group smelt, which showed no trend over the period 1970–1994, varied by a factor of only 10 (1000–10 000 ha1) in both basins (Fig. 8.5). The density of 1-group smelt (Figs 8.5(c) and (d)) varied stronger than the density of 0-group smelt. There was no clear decreasing or increasing trend in the density of 1-group smelt, although autocor-

08 11/02/2003 13:35 Page 95

Number (ha⫺1)

Interaction in the utilisation of small fish by piscivorous fish and birds

Northern basin 0-group perch

105

95

Southern basin (a)

(b)

4

10

103 102 10 1 70

74

78

82 86 Year (19..)

90

94

70

74

78

82

86

90

94

Year (19..)

Figure 8.6 Mean 0-group perch density (ha1), with 95% confidence limits of the mean. A reference line is indicated at a density of 100 ha1

Number (ha⫺1)

105 104 103 102 10 1

Number (ha⫺1)

Northern basin 0-group ruffe

105 104 103 102 10 1

Southern basin

>=1-group ruffe

70

74

78

82 Year (19..)

86

90

(a)

(b)

(c)

(d)

94

70

74

78

82

86

90

94

Year (19..)

Figure 8.7 Mean ruffe density (ha1), with 95% confidence limits of the mean. A reference line is indicated at a density of 100 ha1

relation was high for some years (e.g. the period 1970–1975 in both basins). After 1984, density of 1-group smelt was frequently low at 100 ind ha1. Correlation between smelt density in the northern and southern basins over the period 1975–1994 was low and insignificant (0-group: Pearson  0.1; 1-group: Pearson  0.13, P  0.1). Density of 0-group perch fluctuated much more than the density of smelt, up to a factor 1000 in both basins (Fig. 8.6). As with smelt, there was no trend over the period 1970–1994. Correlation of 0-group perch density between basins was high and significant (Pearson correlation coefficient  0.68, P  0.01). The 0-group ruffe were 2.5 times more abundant, and 1-group ruffe 6.3 times more abundant in the southern basin than in the northern basin over the period 1970–1994 (Fig. 8.7). Only 1-group ruffe showed a decreasing trend over the period 1970–1994 (Figs 8.7(c) and (d)). In the northern basin, 1-group ruffe density seemed to have stabilised after 1982. Correlation between ruffe density of the northern and of the southern basin over the period 1975–1994 was low and insignificant (0-group: Pearson  0.18; 1-group: Pearson  0.04, P  0.1).

08 11/02/2003 13:35 Page 96

96

8.3.4

Impact of birds on fisheries

Production

The mortality rate (Z) of 0-group smelt during summer, calculated as ln(NMay 20)  ln(NNovember 28) (cf. Table 8.2), was low (Z  1.4 growing season1). This was caused by the low density estimate in May compared with September. It is possible that this low density estimate at the start of the growing season was caused by sampling bias, as surface hauls were only made in the middle part of the northern basin. Therefore, total mortality rate was also calculated under the assumption that the daily mortality rate during the period 20 May–28 November was the same as the daily mortality rate during the period 1 September–28 November (Z  0.0161 d1). Under this assumption, Z  3.1 growing season1 in the northern basin, and Z  4.0 growing season1 in the southern basin. Under the assumption of the high values for Z, production was about two times higher in the northern basin, and nearly four times higher in the southern basin than if the low values for Z were used for production estimates (Table 8.3). The higher production estimates were considered more realistic (Buijse et al. 1993). The 0-group smelt dominated production of small fish, contributing 72% of total small fish production in the northern basin and 60% in the southern basin. The 1-group smelt contributed little to the total production because of their marginal growth. In the southern basin, production of 0-group perch and ruffe was higher than in the northern basin. Variation in 0-group smelt production was low compared with other species groups: over the period 1976–1994, 67% of the annual production estimates varied within a factor of 2 around the geometric mean (SDG 1.9 in the northern, and 2.3 in the southern basin). Due to the high inter-annual variation in density, variation in 0-group perch production was high (SDG 5.7 in the northern, and 6.9 in the southern basin).

8.3.5

Utilisation by piscivorous fish and birds

The catch statistics for perch and pikeperch (Fig. 8.8) used to calculated total consumption were multiplied by 1.1 to allow for fish that were traded without using the services of the auction (Nagtegaal & Snel 1984). Assuming a gross food conversion ratio for wet weight of 0.2 (Pauly 1986), and a multiplication factor of 1.2 for perch and 2.9 for pikeperch to allow for fish consumption by perch and pikeperch that were smaller than the minimum legal size, fish consumption by perch and pikeperch in 1976–1994 averaged 38 kg ha1 yr1 (range 8–102 kg ha1 yr1). Cormorants contributed about 40% to the total fish consumption by birds in IJsselmeer (Table 8.4) (Buijse et al. 1993). About 75% of their food consisted of ruffe and perch. For all piscivorous bird species combined, smelt was the most important prey fish, and 40% of the smelt depredation by birds was by black-headed gull. However, black-headed gull uses a variety of terrestial and aquatic food resources, including discards of the commercial fishery (Voslamber 1991), so the species was probably not strongly dependent on the smelt stock (Mous 2000). Overwintering mergansers Mergus spp. and migrating black terns, Chlidonias niger (L.), are more directly dependent on smelt (Doornbos 1979; Schouten 1983; Beekman & Platteeuw

Smelt, 0-group, high M Smelt, 0-group, low M Smelt, 1-groupa Perch, 0-groupb Ruffe, 0-groupc Ruffe, 1-groupa

Southern basin

Mean

MeanG

SDG

Min.

Max.

Mean

MeanG

SDG

Min.

Max.

72.9 31.8 4.8 12.9 6.8 4.4

61.0 26.6 2.6 3.0 3.3 1.5

1.9 1.9 3.4 5.6 3.4 3.4

14.0 6.1 0.3 0.3 0.9 0.3

166.0 72.4 16.4 95.8 35.2 50.8

70.1 18.9 3.3 23.2 11.6 7.7

54.5 14.7 2.0 5.8 5.8 5.2

2.3 2.3 2.9 6.8 3.6 2.6

7.2 1.9 0.3 0.2 0.6 0.9

134.7 36.4 16.0 137.7 60.4 24.9

Mortality estimates for 1-group fish were calculated from the inter-annual survey (see Materials and methods section). Mortality rate was estimated as ln(N20 May 1994) – ln(N28 November 1994)  4.9 growing season1. This mortality rate was also applied to 0-group perch from the southern basin. c For 0-group ruffe, the intra-annual survey resulted in a negative estimate for summer mortality. Hence, the estimate for 1-group ruffe was adopted. a

b

08 11/02/2003 13:35 Page 97

Northern basin

Interaction in the utilisation of small fish by piscivorous fish and birds

Table 8.3 Summary statistics of yearly production (kg ha1) over the period 1976–1994 in the northern and southern basins. Mean  arithmetic mean; MeanG  geometric mean; SDG  back-transformed standard deviation of ln-transformed annual production estimates; Min., Max.  minimum and maximum production estimated

97

08 11/02/2003 13:35 Page 98

98

Impact of birds on fisheries

Catch (kg ha⫺1)

8

Pikeperch Perch

6 4 2 0

76

78

80

82

84

86

88

90

92

94

Year (19..)

Figure 8.8 Annual landings (kg ha1) of perch and pikeperch in 1976–1994 (auction statistics) Table 8.4 Fish consumption by birds (kg1 ha1 yr1). Table adjusted from Buijse et al. (1993). Data are representative for the 1990s (cormorants) and the late 1980s Prey Predator

Smelt

Cormorant, P. carbo 1.7 Black-headed gull, Larus ridibundus 8.1 Grebe, Podiceps cristatus 4.6 Goosander, Mergus merganser 2.7 Red-breasted merganser, Mergus serrator 0.3 Black tern, Chlidonias niger 1.1 Smew, Mergus albellus 0.9 Other 0.9 Total 20.3

Ruffe Roach Bream Eel

Perch

Pikeperch Total

8.4

3.1

0.0

0.2

7.3

0.0

20.7

1.0 0.6

0.0 0.1

0.0 0.1

0.0 0.0

1.0 0.4

0.0 0.0

10.1 5.7

0.6

1.5

0.0

0.0

0.8

0.0

5.6

0.0

0.0

0.8

0.0

1.1

0.0

1.1 1.0 3.0 48.3

0.0 0.1 0.4 11.1

0.7 5.4

0.5 0.6

0.0 0.9

0.5 10.0

0.0 0.0

1994; Stam 1996), although their annual consumption is about 4 times less than the annual consumption by black-headed gull (Table 8.4). Cormorant is the only fish-eating bird species that showed a clear trend in population density in the 1980s and early 1990s (Winter 1994; Prins et al. 1995). The number of breeding pairs in the three colonies around IJsselmeer increased by a factor of nearly 4 from 1978 to 1993 (Fig. 8.9), after which it dropped by one-third in 1994. The drop was probably caused by the extremely bad breeding season of 1993: cormorants had a breeding success of 0.2 young per breeding pair (van Eerden 1993), compared with 0.5–2.6 young between 1981 and 1987 (van Eerden et al. 1991). The low breeding success was caused by starvation of the chicks, as the unusually high water transparency observed during spring of 1993 negatively affected feeding efficiency of cormorants on smelt, the most important food for chicks (van Eerden 1993). Low density of smelt, as observed in autumn 1992 (Fig. 8.5) was probably also an important factor. The spring of 1993 was

08 11/02/2003 13:35 Page 99

Interaction in the utilisation of small fish by piscivorous fish and birds

99

(a) Seasonal variation in cormorant abundance (1992)

Number

60000

40000

Chicks Breeding Not breeding

20000

(b) Cormorant breeding pairs and total amount of fish consumed by cormorants 16000 15 14000 12000 10 10000 8000 6000 5 4000 2000 0 0 1978 1982 1986 1990 1994 1998 Year

Consumption (kg ha⫺1)

Breeding pairs

0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Month

Figure 8.9 Seasonal variation in number of cormorants around IJsselmeer in the year 1992 (a), and inter-annual variation in breeding pairs and fish consumption (b). The number of breeding pairs in the IJsselmeer area over the period 1978–1998 is derived from van Eerden & Gregersen (1995), Timmerman & Prins (1996) and Dutch Centre for Field Ornithology (unpublished data). The relationship between predation pressure exerted by breeding cormorant pairs on the one hand and the total cormorant predation pressure (including chicks and non-breeding cormorants) on the other hand was estimated from cormorant abundance surveys conducted in 1992. Assuming that daily fish consumption for each cormorant was 0.4 kg (cf. Platteeuw et al. 1995; Platteeuw & van Eerden 1995), the total annual predation of the cormorant population was estimated at 175 kg per breeding pair

probably the only period since 1980 during which food was limited for piscivorous birds. In 1994–1998, the number of breeding pairs of cormorant stabilised at around 11 000.

8.3.6

Utilisation of small fish production by the fishery

Smelt is the only small fish species that is targeted by the commercial fishery (Fig. 8.10). Even before the transition from a marine into a freshwater system by the construction of a dyke in 1932, a commercial fishery for smelt existed. After 1932, the fishery was practised irregularly, and only larger smelt (10 cm TL) were landed to satisfy a small domestic market. In 1982, an export market for all size classes of smelt developed. This was the beginning of an intensive fyke net fishery on spawning aggregations of smelt, which takes place for 4–6 weeks during March and April. The fyke net fishery

08 11/02/2003 13:35 Page 100

100

Impact of birds on fisheries

Smelt landings, 1900–1998

30

Wholesale Auction

Catch (kg ha⫺1)

25 20 15

Closing of dyke

10 5 0

0

10

20

30

40

50

60

70

80

90

Year (19..)

Figure 8.10 Annual catch of smelt (kg ha1 yr1) over the period 1900–1998. In the years 1990–93 and 1995, statistics on both the amount fish sold directly to wholesalers and the amount of fish traded at the auction were collected. In these years, between 47 and 62% (mean 54%) of the total amount of smelt caught were auctioned. The year when the dyke that separated IJsselmeer from the Waddensea was completed (1932) is indicated

Table 8.5 Mean total instantaneous mortality rate (yr1) of smelt from fall to fall of the next year in the periods 1970–1981 and 1982–1994. Z was estimated as ln(N0-group,yeart1  N 1-group,yeart1) ln(N 1-group,yeart), where N0-group and N 1-group represent population density estimates in fall (cf. Figs 5(a)–(d)) 1971–1981

1982–1994

Basin

Mean

SD

Mean

SD

North South

2.05 1.10

1.20 1.05

3.20 2.45

1.27 1.49

targets the entire smelt population, including the youngest age groups (Janss 1991). The fishery lands between 5 and 30 kg ha1 yr1. After development of the smelt fyke net fishery in 1982, survival of 0-group smelt dropped markedly. Instantaneous mortality rates (Z, year1) were 1.15 (northern basin) and 1.35 (southern basin) higher than before (Table 8.5; two-tailed t-test, P  0.05). Hence, fishing mortality (F) must have amounted to 1.15 year1 for the northern basin and 1.35 year1 for the southern basin. This implies that during the 4–6 weeks when the smelt fishery takes place, about 70% of the population is caught each year. The trawl surveys before and after the smelt fishing season of 1994 (see Materials and Methods) corroborated these estimates of F. The density of 0-group smelt differed significantly before and after the fishing season (ANOVA on ln-transformed catch numbers, P  0.01). Z over the fishing season (10 March–9 April) was estimated at 1.59 (95%

08 11/02/2003 13:35 Page 101

Interaction in the utilisation of small fish by piscivorous fish and birds

101

CL: 0.54–2.64). Assuming that the natural mortality was negligible, F is equal to Z. Density of 1-group smelt fell below trawl detection level during the fishing season: before the fishing season, only 2 of 13 hauls caught no fish, whereas after the fishing season, 9 of 13 hauls contained no 1-group smelt. Catch numbers of 0-group perch, 0-group ruffe and 1-group ruffe did not decrease significantly over the smelt fishing season (ANOVA on ln-transformed catch numbers; Wilcoxon’s test, P  0.10). This confirmed the observations that bycatch in the fyke net fishery for smelt was small. Before 1970, smelt and other small fish were also caught as bycatch in the trawl fishery for eel. This bycatch was sold at the auction as trash fish (Fig. 8.11). The bulk of these trash fish comprised smelt and ruffe, but substantial numbers of perch and pikeperch juveniles were also caught (De Beaufort 1954). Because of its hypothesised negative effects on the gillnet and fyke net fishery for adult perch and pikeperch, the eel trawl was banned in 1970. After 1970, the eel trawl was replaced by a fyke net fishery for eel. Similar to the eel trawl fishery, the fyke net fishery caught other species along with the target species. However, this non-target catch was not sold at the auction, but discarded. As discarded fish do not survive (Willemsen 1985), the fyke net fishery causes fishing mortality in small fish. Compared with the bycatch of the eel trawl fishery (13–62 kg ha1 year1), the bycatch of the eel fyke net fishery (4–14 kg ha1 year1) was low (Fig. 8.11). The eel fyke nets caught less smelt, more perch, and in 1993 a substantial amount of flounder, Platichthys flesus (L.), (36% of the total bycatch). The catch of smelt in the fyke nets was underestimated, since the fishermen removed gilled smelt during the hauling process, before the catch could be measured. Irrespective, the bycatch after the trawl ban in 1970 was orders of magnitude smaller than before 1970. Thus, the exploitation

Figure 8.11 Yearly bycatch of the eel trawl fishery (a) and the eel fyke net fishery (b). To estimate species composition of the eel trawl bycatch, the species composition of survey trawl catches made in summer 1946–1951 (De Beaufort 1954) and 1966–1969 (Steinmetz & Oudelaar 1971) were used. The survey gear was similar to the gear used by the commercial fishery. Total weight of the eel trawl bycatch was obtained from auction statistics. The yearly eel fyke net bycatch of 1985, 1987 and 1993 was estimated by experimental fyke nets and a creel survey among the fishermen (Dekker, et al. 1993)

08 11/02/2003 13:35 Page 102

102

Impact of birds on fisheries

pattern of smelt has changed dramatically from intensive exploitation by a trawl fishery prior to 1970, followed by low exploitation rates up to 1981, to the intensive exploitation of spawning smelt by a fyke net fishery.

8.3.6

The balance between production and utilisation

Smelt was the most important small fish species in terms of production and consumption. Although variation in the production of smelt was large, and there were uncertainties in the consumption estimates, the estimate of smelt production was low compared with its total utilisation. Thus, the major part of smelt production must have been utilised by piscivorous fish, birds and by the fishery (Table 8.6).

8.3.7

Factors affecting production

Recruitment A stock–recruitment relationship could only be proven for ruffe (Fig. 8.12). In the northern basin, the parental stock was frequently low (5 kg ha1), which resulted in a lower recruitment (linear regression of ln-transformed 0-group ruffe density on ln-transformed density of parental stock: r2  0.21, P  0.02). With only one exception, parental stock density in the southern basin was higher than 9 kg ha1, which was apparently sufficient to guarantee reproduction. Possibly, the variation in parental stock size of smelt between 1970 and 1993 was too low to detect a stock–recruitment relationship. Although the smelt fishery caught fish that had not reproduced (cf. Janss 1991), there was no evidence that this practice decreased recruitment of smelt. With an autumn smelt stock of at least 4.4 kg ha1 recruitment seemed to be guaranteed. Earlier studies on perch and pikeperch recruitment in IJsselmeer already showed that there was no significant stock–recruitment relationship (Buijse et al. 1992b) within the observed variation in the parental stock size in IJsselmeer. Table 8.6 Summary of smelt production and utilisation estimates (kg ha1 yr1). Minimum and maximum over the indicated period are between brackets

Production (1976–1994) 0-group smelt 1-group smelt Utilisation piscivorous fisha (1976–1994) piscivorous birds (1983–1987) fisheryb (1982–1994)

Northern basin

Southern basin

73 (14–166) 5 (0.34–16)

70 (7.1–135) 3 (0.32–16) 38 (8–102)  20 19 (6–34)

a Most of the fish (c. 90%) consumed by perch and pikeperch consist of smelt (Willemsen 1977; Buijse & Houthuijzen 1992; Buijse & Van Densen 1992; Buijse et al. 1993). b Corrected auction statistics, assuming that 54% of the total catch was auctioned.

08 11/02/2003 13:35 Page 103

Interaction in the utilisation of small fish by piscivorous fish and birds

103

Growth in relation to density, water temperature, and food availability Slow growth was associated with high population density for each species and in both basins, except for 0-group ruffe in the northern basin (Table 8.7). Density and mean length of smelt and ruffe were in some situations correlated with water temperature, but the cor100000

(a) Smelt

10000

Recruits (ha⫺1)

1000 Northern basin Southern basin 100 10000

(b) Ruffe

1000

100

10

1

10 Parental stock (kg ha⫺1)

100

Figure 8.12 Scatterplots of recruitment vs parental stock size over the period 1970–1993. The population stock size (weight of 0-group and 1-group fish combined) during autumn was used as a measure for the parental stock size; the density of 0-group fish (ha1) in autumn of the next year was used as a measure for recruitment Table 8.7 Pearson correlation coefficients for correlation between density (N), mean total length (TL) and water temperature in the 2nd (Q2) and 3rd (Q3) quarter. The correlation coefficients were calculated from estimates over the period 1976–1994 Basin/Species

N TL

N Q2

Northern basin Smelt, 0-group Smelt, 1-group Perch, 0-group Ruffe, 0-group Ruffe, 1-group

0.44* 0.32 0.44* 0.07 0.13

Southern basin Smelt, 0-group Smelt, 1-group Perch, 0-group Ruffe, 0-group Ruffe, 1-group

0.25 0.58** 0.28 0.17 0.55**

*0.05  P  0.1; **0.01  P  0.05.

N Q3

TL Q2

TL Q3

0.04 0.09 0.58** 0.06 0.14

0.22 0.47* 0.02 0.28 0.20

0.25 0.55 0.18 0.52** 0.03

0.04 0.10 0.53** 0.35 0.08

0.37 0.35 0.60** 0.16 0.08

0.17 0.47* 0.06 0.09 0.26

0.36 0.22 0.03 0.14 0.47*

0.24 0.29 0.47** 0.17 0.07

08 11/02/2003 13:35 Page 104

104

Impact of birds on fisheries 1.6 1:1

MLstomach (mm)

1.4 1.2 1 0.8

0-group smelt 1-group smelt 0-group perch

0.6 0.4 0.4

0.6

0.8 1 MLwater (mm)

1.2

1.4

Figure 8.13 Relationship between mean length of Daphnia spp. available in the water column (MLwater) and the mean length of Daphnia spp. encountered in stomachs of 0-group smelt (MLstomach), 1-group smelt and 0-group perch. The 1 : 1 line represents non-selectivity. Data from Houthuijzen (1989)

relations were not consistent between basins (Table 8.7), and were not considered further. A higher mean water temperature during the 2nd quarter had a positive effect on density of 0-group perch in both basins. Apparently, high temperatures during spawning and the first months of the life favours reproduction success (see also Buijse et al. 1992b). Growth of 0-group perch was positively affected by 3rd quarter mean water temperature; this correlation was consistent between the basins. The dependence of first-year growth on the mean temperature of the third quarter was to be expected, since 0-group perch gain 90% of their weight at the end of the growing season. In conclusion, 0-group perch density and growth were positively affected by temperature, whereas growth of smelt and ruffe seemed to be governed by density-dependent resource limitations only. In IJsselmeer, smelt and 0-group perch selected larger bodied Daphnia spp. (Fig. 8.13). Also in Tjeukemeer, a 2000 ha lake nearby IJsselmeer, both smelt and 0-group perch select Daphnia spp., and growth is slower in the absence of their preferred food item (van Densen 1985; Lammens et al. 1985). The relationship between growth and mean Daphnia length (Fig. 8.14(a)), and between smelt growth and density seemed to be the same for IJsselmeer and Tjeukemeer (Fig. 8.14(b)), which confirmed densitydependent growth mediated by Daphnia (Table 8.8). As production and consumption rate of zooplanktivorous fish were about the same in the northern and the southern basin (cf. Table 8.3), the lower availability of large-bodied Daphnia in the southern basin must have been caused by a lower potential production of Daphnia species. Condition, as indexed by percentage dry weight (PDW), of 0-group smelt increased with length, but was lower at the end of the growing season (November) (Fig. 8.15(a)). For 1-group smelt, the relationship between PDW and length were parabolic (Fig. 8.15(b)). Condition was maximal around 10 cm TL in June, and at the end of the growing season condition was 20% lower than in June. Growth of 1-group smelt stopped in August. An ANCOVA model with PDW of 0-group and 1-group smelt as the dependent variable, and length (TL) and month (MONTH) as independent variables (PDW  Constant  TL  TL2  MONTH) explained 98% of the variation in PDW (P  0.0001) (Table 8.9, Fig. 8.15(c)).

08 11/02/2003 13:35 Page 105

Interaction in the utilisation of small fish by piscivorous fish and birds

105

(a) Length 0-group smelt – length Daphnia spp. Mean length 0-group smelt (cm)

10 9 8 7 Tjeukemeer Northern basin Southern basin

6 5 0.5

0.6

0.7

0.8

0.9

1

1.1

Mean length Daphnia spp. (mm) (b) Length 0-group smelt – abundance 0-group smelt 10

Length (cm)

9 8 7 6 5 1

10

100

1000

10000

100000

Number (ha⫺1)

Figure 8.14 Top panel (a): Relationship between length of 0-group smelt at the end of the growing season and mean length of Daphnia spp. in August–September in IJsselmeer and Tjeukemeer. Bottom panel (b): Relationship between length of 0-group smelt at the end of the growing season and 0-group smelt density at the end of the growing season

Table 8.8 Pearson correlation coefficients for ln-transformed 0-group smelt density (ln N), smelt length (L) and Daphnia length (DL) IJsselmeer

ln N L

IJsselmeer and Tjeukemeer

L

DL

L

DL

0.31

0.77 0.83**

0.66***

0.74*** 0.82***

*0.05  P  0.1; **0.01  P  0.05; ***P  0.01.

08 11/02/2003 13:35 Page 106

106

Impact of birds on fisheries 30

(a) 0-group smelt May

27

June

24

August

21

September November

18 15 30

(b) >=1-group smelt

Dry weight (%)

27 24 21 18 15 30

(c) model estimates

27 24 21 18 15 4

6

8 10 Length (cm)

12

14

Figure 8.15 Seasonal variation in the relationship between dry weight (as % of total wet body weight) and total length of 0-group smelt (a) and 1-group smelt (b). The mean length of the age class in each month is indicated by arrows. The lower panel (c) represents the modeled relationship

In general, seasonal variation in condition of smelt reflected seasonal variation in food availability. The decrease in PDW of smelt 10 cm TL was probably caused by an increasing shortage of food, i.e. zooplankton of proper sizes. Hence, the capacity of IJsselmeer to sustain production of larger zooplanktivorous smelt must have been low. Smelt 15 cm TL, which occurred at very low density (1 ha1) in IJsselmeer, often

08 11/02/2003 13:35 Page 107

Interaction in the utilisation of small fish by piscivorous fish and birds

107

Table 8.9 ANCOVA table for PDW–length relationships in smelt. Effect mean squares are type III Source

df

MS

P

Model Error TL TL2 MONTH

6 43 1 1 4

69.9 1.09 133 106 38.0

0.0001 0.0001 0.0001 0.0001

df  degrees of freedom, MS  mean squares, P  significance level.

contained 0-group smelt in their stomachs (unpublished data). These very large smelt were probably the few survivors that succeeded in switching to piscivory.

8.4 Discussion Between 1970 and 1995, zooplanktivorous smelt was a keystone species in IJsselmeer. In the northern basin, smelt comprised 72% (cf. Table 8.3), and in the southern basin 60%, of the total production of small fish (102 kg ha1 yr1 in the northern and 116 kg ha1 yr1 in the southern basin). Total small fish production, which probably comprised the larger part of total fish production, was within the range reported for lakes of similar trophic status: according to the relationship between total phosphorus concentration (TP) and fish production of Downing et al. (1990), total fish production in IJsselmeer was estimated to range between 10 and 130 kg ha1 yr1. Production of 0-group smelt in the northern and southern basins, was relatively stable: 67% of the annual production for each species was within a factor 2 around the inter-annual mean. Such stability, here pertaining to a single species of a single age group, is found in other fisheries. For example, in Windermere (UK), 67% of the annual production estimates of the perch population varied within a factor of 1.8 (Craig 1980) around the inter-annual mean over a 12-year period, and in Danube River, this factor was 2.5 for the total fish population over a 5-year period (Holcik 1996). Variation in the production of 0-group perch in IJsselmeer was extremely high (a factor of 6), due to the high variation in year class strength (Buijse et al. 1992b). The relatively stable production of smelt formed a stable forage base for piscivorous fish and birds in IJsselmeer. In IJsselmeer, all of the smelt production was probably utilised by piscivorous fish and birds, and by the fishery. There are few studies that attempt to balance production against utilisation. Brandt et al. (1991) estimated that in Lake Michigan (USA), about 50% of alewife Alosa pseudoharengus (Wilson) production is utilised by a commercial fishery and piscivorous salmonids. Of a total fish production of 243 kg C ha1 yr1 in a Sri Lankan reservoir, 22% was utilised by avian piscivores, mainly cormorants Phalacrocorax spp., grey pelican Pelecanus philippensis (Gmelin) and storks, and

08 11/02/2003 13:35 Page 108

108

Impact of birds on fisheries (a) Total phosphate 1 0.8 0.6 0.2 0.4 Northern basin Southern basin River Rhine 0 1970

1975

0.2

1980

1985

1990

0 1995

1985

1990

1995

Concentration (mg PL1) River Rhine

Concentration (mg PL1)

0.4

Year

Concentration (mg NL1)

5

(b) Nitrate

4 3 2 1 0 1970

1975

1980 Year

Figure 8.16 Total phosphate (a) and nitrate (b) concentration in IJsselmeer and in River Rhine. Source: Netherlands Ministry of Transport, Public Works and Water Management

9% by the fishery (Piet et al. 1999). For the same Sri Lankan reservoir, Pet et al. (1996) proposed the introduction of a piscivore, barramundi, Lates calcarifer (Bloch) to utilise the excess fish production, enabling the development of a new fishery. In Lake Victoria (East Africa) the idea of introducing a large predator (Nile perch Lates niloticus (L.)) to utilise production of haplochromines was put into practice, resulting in major changes in the fish community and other components of the ecosystem (Witte et al. 1992; Ogutu-Ohwayo 1990; Wanink & Goudswaard 1994). In accordance with Parsons (1996), it is suggested that, in most situations, fish production is balanced by utilisation, and that changes in utilisation generally affect the food web structure. Growth of both zooplanktivorous smelt and benthivorous ruffe was food limited in IJsselmeer, and in both fish species growth suddenly increased in the northern basin after the closing of the dyke that separates the northern basin from the southern basin in 1975 (cf. Figs 8.2 and 8.4). The closing of the dyke separated the southern basin from the inflow of nutrient-rich water of the River IJssel, a tributary of the River Rhine (Fig. 8.16). Consequently, the average residence time of water was more than 4 times shorter in the northern basin than in the southern basin (Prins et al. 1995). The increased nutrient loading of the northern basin after 1975 was probably passed on in

08 11/02/2003 13:35 Page 109

Interaction in the utilisation of small fish by piscivorous fish and birds

109

the food chain to primary and secondary production, causing increased food availability and faster growth for smelt and ruffe. Predation pressure by smelt on zooplankton probably caused a decrease in food quality (i.e. zooplankton size), which in turn decreased smelt growth, as was suggested by the negative relationship between smelt density and smelt growth, and the positive relationship between smelt growth and Daphnia body size. The existence of this mechanism was also demonstrated for 0-group yellow perch Perca flavescens (Mitchill) in Oneida Lake, USA (Mills et al. 1989a, b). The relationships between Daphnia body size, smelt growth and smelt density were consistent over both basins of IJsselmeer and Tjeukemeer, which suggests that zooplankton body size is a suitable measure for the level of particulate zooplanktivory. This agrees with Mills & Schiavone (1982) and Mills et al. (1987), who concluded that zooplankton body size could be a useful, although rough, indicator for abundance of fish, even for comparing between systems. A possible explanation for a difference in Daphnia length between the two basins at equal predation pressure from zooplanktivorous smelt was a difference in zooplankton production between the two basins. As nutrient loading is much higher in the northern basin (Prins et al. 1995) this explanation seemed justified. In IJsselmeer, smelt reach sexual maturity in one year (Willemsen 1977). This fast rate of maturation has also been observed in other Eurasian smelt populations that inhabit shallow lakes (Belyanina 1969; Ivanova & Polovkova 1972; Ivanova 1981; Ivanova & Volodin 1982). The ability of slow-growing smelt to reach sexual maturity after its first summer is probably a key factor for its success in IJsselmeer, enabling it to withstand high predation and fishing mortality. Production of larger smelt was probably not only limited by food availability but also by mortality through starvation, as indicated by the low condition of smelt 10 cm TL. Probably, only smelt that managed to switch to piscivory by feeding on their own offspring in early summer survived, explaining the low density of larger smelt (15 cm TL) in IJsselmeer. This critical switch from zooplanktivory to piscivory was also demonstrated for pikeperch (Buijse & Houthuijzen 1992). Strong interaction between the smelt fishery and other users may occur, especially in late spring, when smelt biomass is at its minimum. This was illustrated by the food shortage experienced by cormorants in early 1993, which was primarily caused by an exceptionally weak year class of smelt, but was probably worsened by the smelt fishery in early spring. As the mechanisms that determine year class strength are mostly beyond control by management measures (Borchardt 1988; Buijse et al. 1992b; Luecke et al. 1990; Mooij 1996), an occasional conflict between the smelt fishery and cormorants cannot be avoided through wildlife or fishery management. In IJsselmeer, small fish production was probably balanced by utilisation, and growth of the main small fish producers was density-dependent. Therefore, the carrying capacity of the system for the smelt fishery, piscivorous fish and piscivorous birds was probably reached. This conclusion is supported by observations on the population dynamics of piscivorous birds and the status of the fishery. The breeding success of the cormorant population depending on IJsselmeer is now relatively low compared with other populations (van Eerden & Gregersen 1995), since the increase in the cormorant population size was apparently halted in the early 1990s (cf. Fig. 8.11) and there is strong evidence that the IJsselmeer fishing fleet suffers from overcapacity (Buijse

08 11/02/2003 13:35 Page 110

110

Impact of birds on fisheries

1992; Taal & De Wilde 1997). Management scenarios that would aim for a higher utilisation of small fish are probably not viable, as there is little scope for further increasing small fish production by management of the IJsselmeer bird populations or fish stocks. These conclusions are valid only for the IJsselmeer ecosystem in its present state. A change in the ecosystem structure could result in a change in its carrying capacity. This was illustrated by the increase in smelt growth in the northern basin, explained here by a man-induced increase in nutrient loading. There is, however, probably scope to optimise utilisation of small fish production through fishery management. Buijse et al. (1992a) showed that improved management of the gillnet fishery could increase the yield of perch and pikeperch by a factor ranging between 1 and 1.3. As this results in a higher production of piscivorous fish, predation pressure on small fish will increase. This will be at the expense of small fish utilisation by either the fishery for spawning smelt or by piscivorous birds. Consequently, improved management of the gillnet fishery is viable only if either the intensity of the smelt fishery is decreased, or if one allows less piscivorous birds around the lake.

Acknowledgements We thank Koos Vijverberg (Netherlands Institute of Ecology), Eddy Lammens (Netherlands Institute for Inland Water Management and Waste Water Treatment), Tom Buijse (Netherlands Institute for Inland Water Management and Waste Water Treatment) and Bram Huisman (Fish Culture and Fisheries Group, Wageningen University) for critically reviewing the manuscript. The field work was made possible by the RV ‘Stern’ and her crew, and by technical assistants from the Netherlands Institute of Fisheries Research. Condition measurements on smelt were done by Martin van Brakel. This research was supported by grants from the Netherlands Ministry of Agriculture, Nature and Fisheries, Directorate Fisheries, from the Netherlands Ministry of Transport, Public Works and Water Management, Rijkswaterstaat, Directorate IJsselmeergebied, and from the International Reference Centre for Nature Management of the Netherlands Ministry of Agriculture, Nature and Fisheries.

Appendix: Calculating fish density from catch-per-unit-effort observations Intra-annual survey For the estimation of fish density from CPUE observations made during intra-annual survey (Table 8.A1), CPUE observations were corrected for the swept area and expressed as numbers ha1 first. Next, ANCOVA models were constructed of ln-transformed CPUE observations, using bottom depth (DEPTH) and water transparency (SECCHI) as independent variables. Thus, the deterministic part of the model reads: ln(CPUE)  Constant  b1 DEPTH  b2 SECCHI  b3 DEPTH SECCHI

(8.A1)

08 11/02/2003 13:35 Page 111

Interaction in the utilisation of small fish by piscivorous fish and birds Table 8.A1

Sampling conditions during the intra-annual survey Depth (m)

Basin/ Survey

111

Sampling period

Secchi depth (m)

N

Mean

Min.

Max.

Mean

Min.

Max.

Northern basin A 17–26 May 94 B 29 Aug–8 Sep 94 C 21 Nov–15 Dec 94 D, NS 8–30 Mar 95 D, OW 8–30 Mar 95

39 45 57 38 57

4.6 4.5 4.9 4.2 4.9

2.8 2.7 2.7 3.0 2.9

6.2 5.8 7.5 5.6 6.6

1.09 0.66 1.48 0.77 0.67

0.6 0.5 0.7 0.4 0.4

1.6 1.7 3.0 1.3 1.3

Southern basin A B 5–7 Sep 94 C 12–14 Dec 94 D, NS 20–23 Mar 95 D, OW 21–23 Mar 95

0 29 27 21 19

– 3.9 3.7 3.5 4.1

– 2.6 3.0 2.7 3.2

– 4.5 4.3 4.2 4.6

– 0.66 0.33 0.24 0.20

– 0.4 0.3 0.2 0.2

– 0.8 0.6 0.5 0.2

N  number of hauls; NS  near shore; OW  open water.

where b1, b2 and b3 are the parameter estimates. ANCOVA statistics and parameter estimates are presented in Table 8.A2. The models were used to calculate estimates of ln(CPUE) for each observation. The density estimates were calculated as the arithmetic means of back-transformed estimated values. Because allocation of sampling locations was proportional to surface area of strata, the density estimates were unbiased with respect to water transparency- and depth-mediated distribution patterns. Using this method, it was implicitly assumed that unexplained variance was caused by sampling error rather than by random spatial variation in density. Population density of smelt was also corrected for water transparency-mediated vertical distribution (Mous 2000). Estimates of 0-group fish density in May 1994 were calculated as geometric means of surface trawl CPUE observations.

Inter-annual survey To obtain population density estimates from CPUE observations made during the inter-annual survey, four sources of bias were corrected for. First, CPUE observations were corrected for the swept area and expressed as numbers ha1. Second, CPUE observations were corrected for size selectivity of the trawl gear, which was caused by cod-end mesh penetration of small fish (Mous 2000). Third, CPUE observations were corrected for depth- and water transparency-mediated variation in spatial distribution patterns (see below). Finally, smelt CPUE was corrected for the vertical distribution pattern, assuming the vertical distribution pattern of smelt 7 cm TL as observed during summer 1992. The vertical distribution pattern was corrected for by multiplying CPUE by 0.42 times the water depth: net height ratio (Mous 2000). Correction for depth- and water transparency-mediated distribution patterns was necessary, because the sampling locations of inter-annual survey did not adequately represent the IJsselmeer habitat in respect to depth and water transparency. Most sampling

s2

MSE

F

INT

b1

b2

b3

n0

s2

MSE

F

INT

b2

1 0 0

4.50 3.67 1.95

1.98 1.70 1.52

19*** 34*** 9.9**

35.5 8.7 22.7

3.69 0.77 2.54

43.1 2.10 21.4

6.97 NS 4.26

1 0 0

1.14 0.54 1.78

1.04

3.8*

12.8

3.30

13 6 12 41

3.85 0.88 1.53 0.93

1.87 0.834 1.23 0.833

9.8** 3.4* 9.7** 4.6

19.2 3.0 0.7 3.5

3.21 0.31 0.64 0.26

24.9 NS NS 1.97

5.55 NS NS NS

1 0 1

0.68 0.36 1.47

–

NS

–

–

3 9 14

1.78 1.88 1.60

1.65 1.61 1.28

4.52 9.3** 8.4***

5.7 0.2 11.1

0.49 0.63 1.22

NS NS 14.4

NS NS 2.44

0 0 0

1.65 1.49 0.531

–

NS

–

–

0 14 6

3.11 6.87 5.36

1.67 2.96 2.17

14*** 29*** 45***

23.9 3.2 14.7

3.33 2.13 0.94

36.1 1.97 24.4

6.99 – 3.65

0 0 0

1.28 1.41 0.525

0.846

16**

9.7

6.04

2 4 25 23

4.08 3.10 4.06 2.24

2.30 2.44 2.22 1.29

15*** 6.4** 14*** 19***

5.5 4.4 0.8 15.0

0.82 0.62 1.39 1.63

4.13 5.33 2.19 22.6

NS NS NS 3.71

3 0 1

2.77 2.07 0.57

0.82

62**

11.1

n0

12.8

08 11/02/2003 13:35 Page 112

Smelt, 0-group B C D Smelt, 1-group A B C D Perch, 0-group B C D Ruffe, 0-group B C D Ruffe, 1-group A B C D

Southern basin

Impact of birds on fisheries

Northern basin Species/ Survey

112

Table 8.A2 ANCOVA statistics and parameter estimates of models which relate ln(CPUE) of the most important small fish species caught during the intra-annual survey to DEPTH (in m) and SECCHI (in m) (see text). n0  number of hauls with zero individuals in the catch, s2  total variance of ln(CPUE), MSE  mean squares of error of the model, F  F-statistic of the model. The significance level of F is coded as: *  0.05 P  0.1; **  0.01 P  0.05; ***  P  0.01; NS  not significant. Dependent variables that did not contribute significantly to the model were omitted, and the model was refitted without these variables. An independent variable was also excluded from the model if variation in the variable was low (cf. Table 8.A1). This was the case for DEPTH in the southern basin, and SECCHI in all surveys in the southern basin except intra-annual survey-b

08 11/02/2003 13:35 Page 113

Interaction in the utilisation of small fish by piscivorous fish and birds

113

Table 8.A3 Sampling conditions (mean over sampling stations, SD) during inter-annual survey in the period 1970–1994. n  number of survey hauls Northern basin Year

n

DEPTH (m)

1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

6 9 7 3 5 3 6 5 5 7 10 12 9 8 19 10 13 16 19 33 38 39 26 30 28

4.5  1.15 5.0  0.94 6.4  0.94 6.9  1.40 5.7  1.16 5.0  1.27 5.5  1.34 5.3  0.83 5.6  1.19 5.9  0.99 5.8  0.87 5.8  0.92 5.8  0.92 5.9  1.11 6.0  0.81 6.2  1.35 6.3  1.06 5.8  1.07 5.7  0.80 5.8  0.59 5.8  0.71 5.6  0.65 5.5  0.74 5.5  0.93 5.4  0.91

SECCHI depth (m) 0.27  0.08 0.65  0.42 0.58  0.30 0.50  0.26 0.49  0.15 0.42  0.16 0.42  0.07 0.41  0.12 0.84  0.21 0.50  0.12 0.59  0.27 0.58  0.20 0.56  0.34 0.60  0.23 0.78  0.24 0.63  0.24 0.55  0.17 1.12  0.46 0.98  0.27 0.83  0.43 0.64  0.29 0.80  0.33 0.61  0.15 0.84  0.32 1.01  0.54

Southern basin n

DEPTH (m)

3 4 4 1 2 3 1 4 5 3 5 7 5 6 5 5 7 3 12 15 15 15 12 14 11

3.0  0.42 2.6  0.46 3.6  0.38 3.8 3.8  0.07 3.5  0.45 3.5 3.6  0.15 3.3  0.51 3.7  0.57 3.8  0.46 3.8  0.53 3.7  0.45 4.1  0.64 3.9  0.26 4.2  0.40 3.9  0.22 3.7  0.29 3.9  0.32 4.0  0.28 3.9  0.26 3.8  0.17 3.9  0.21 4.2  0.68 4.0  0.26

SECCHI depth (m) 0.23  0.03 0.29  0.09 0.21  0.05 0.25 0.35  0.07 0.25  0.05 0.3 0.25  0.10 0.15  0.00 0.32  0.03 0.23  0.06 0.32  0.12 0.25  0.00 0.39  0.04 0.36  0.05 0.21  0.02 0.21  0.07 0.27  0.03 0.66  0.16 0.86  0.16 0.23  0.04 0.35  0.19 0.29  0.02 0.52  0.08 0.30  0.08

locations were situated in the deeper areas of the northern basin (cf. Fig. 8.1). Hence, the mean depth of all sampling locations in the northern basin was on average about 1 m deeper than the actual mean depth of the northern basin (Table 8.A3). This implies that the habitat consisting of the former tidal channels is over-represented compared with the former sandbanks. Furthermore, until 1987 the choice of sampling locations was biased towards locations, with low water transparency at the moment of sampling. The reasoning behind this strategy was that low water transparency presented better sampling conditions, because of a higher vulnerability of fish to the trawl. Indeed, Buijse et al. (1992c) show that CPUE is low if the water is clear. However, the low CPUE in clear water was probably caused by a low local density of fish, as some fish species avoid areas where water transparency is high (Ryder 1977; Mous 2000). It is therefore impossible to evaluate whether CPUE of the survey trawl was affected by variation in population density or by variation in gear vulnerability, and this sampling strategy may have caused an overestimation of population density. After 1987, sampling took place without taking water transparency into consideration. This change in sampling strategy partly explains the frequent occurrence of high mean water transparency at the sampling stations after 1987 (Table 8.A3).

08 11/02/2003 13:35 Page 114

114

Impact of birds on fisheries

To correct for depth- and water transparency-mediated distribution patterns, fish density was first described as a function of depth and water transparency, using CPUE observations from intra-annual survey-c. This survey, which was carried out during the same period in the year as inter-annual survey, was not biased in respect to depth and water transparency. With the parameter estimates (Table 8.A1), CPUE observations were scaled to values corresponding to the lakewide mean depth (4.5 m) and the lakewide mean Secchi depth (0.75 m). The values 4.5 and 0.75 were assumed to be representative for the period 1970–1994 (Prins et al. 1995). The formula used for the correction reads: ln(CPUE)  ln (CPUE)  b1(DEPTH  4.5)  b2(SECCHI  0.75)  b3(DEPTH × SECCHI  4.5 × 0.75)

(8.A2)

where ln(CPUE) is the corrected value, and b1, b2 and b3 are the parameter estimates of the spatial distribution models. If an independent variable in the model was not significant, its associated parameter value was set to zero. The back-transformed mean of ln(CPUE) was an estimate for the population density at a depth of 4.5 m and a Secchi depth of 0.75 m. This estimate represents the value most likely to occur if an CPUE observation at a randomly chosen location in fall of a particular year would have been made. Only if parameters b1, b2 and b3 all have value zero, the back-transformed mean of ln(CPUE) represents a density estimate representative for the basin as a whole. In other situations, a raising factor was applied to the back-transformed mean of ln(CPUE) to obtain a density estimate representative for the basin. The raising factor was calculated as the ratio of the arithmetic mean of predicted CPUE values to the geometric mean of predicted CPUE values of intra-annual survey-c. Because parameter estimates were unavailable for the southern basin, the CPUE observations for depth- and water transparency-mediated spatial distribution patterns could not be corrected. However, for the southern basin geometric means of CPUE observations were probably relatively unbiased estimates for population density, because depth and water transparency were almost constant during the study period (cf. Table 8.A3). Furthermore, the average depth and water transparency of sampling locations were close to the average depth and water transparency of the southern basin (cf. Prins et al. 1995). Only in 1988, 1989 and 1993 was the average water transparency of the sampling stations relatively high. Therefore, population density estimates of these years were not calculated for the southern basin. Three assumptions underpin the method to correct for spatial distribution patterns. First, the major small fish species were assumed to be distributed as during intraannual survey-c (cf. Table 8.A2). As CPUE generally increased with increasing depth and decreasing water transparency during the young fish surveys in autumn (Buijse et al. 1992), this assumption seemed valid. Second, it was assumed that inter-annual variation in depth and water transparency was caused by sampling artifacts, rather than by an inter-annual variation in the true mean depth and water transparency. For water transparency, this assumption was probably too rigid. However, because variation in annual average water transparency does not show a trend (Prins et al. 1995), this assumption probably did not obscure any significant trend in small fish density. Third, it was assumed that unexplained variation in CPUE was caused by sampling error rather than by random spatial variation in density.

08 11/02/2003 13:35 Page 115

Interaction in the utilisation of small fish by piscivorous fish and birds

115

References Baranoff T.I. (1918) On the question of the biological basis of fisheries. Institute for Scientific Ichthyological Investigations, Proceedings 1(1), 81–128. (translated by W.E. Ricker, 1945). Beekman J.H. & Platteeuw M. (1994) Het nonnetje Mergus albellus in het IJsselmeergebied. RWSDF Intern Rapport 37 LIO, 120 pp. (in Dutch). Begon M., Harper J.L. & Townsend C.R. (1990) Ecology, Individuals, Populations and Communities (2nd edn). Oxford: Blackwell Scientific Publications, 945 pp. Belyanina T.N. (1969) Synopsis of biological data on smelt Osmerus eperlanus (Linnaeus) 1758. FAO Fisheries Synopsis 78, 56 pp. Borchardt D. (1988) Long-term correlations between the abundance of smelt, Osmerus eperlanus eperlanus L., year classes and abiotic environmental conditions during the period of spawning and larval development in the Elbe River. Archiv für Fischwissenschaft 38, 191–202. Brandt S.B., Mason D.M., Patrick E.V., Argyle R.L., Wells L., Unger P.A. & Stewart D.J. (1991) Acoustic measures of the abundance and size of pelagic planktivores in Lake Michigan. Canadian Journal of Fisheries and Aquatic Sciences 48, 894–908. Buijse A.D. (1992) Dynamics and exploitation of unstable percid populations. PhD thesis, Agricultural University Wageningen, 168 pp. Buijse A.D., Pet J.S., van Densen W.L.T. & Rabbinge R. (1992a) A size- and age-structured simulation model for evaluating management strategies in a multi-species gillnet fishery. Fisheries Research 13, 95–117 Buijse A.D., van Densen W.L.T. & Schaap L.A. (1992b) Year-class strength of Eurasian perch (Perca fluviatilis L.) and pikeperch (Stizostedion lucioperca L.) in relation to stock size, water temperature and wind in Lake IJssel, The Netherlands, 1966–1989. In A.D. Buijse Dynamics and exploitation of unstable percid populations. PhD thesis, Wageningen Agricultural University, pp. 31–69. Buijse A.D., Schaap L.A. & Bult T.P. (1992c) Influence of water clarity on the catchability of six freshwater fish species in bottom trawls. Canadian Journal of Fisheries and Aquatic Sciences 49, 885–893. Buijse A.D., van Eerden M.R., Dekker W. & van Densen W.L.T. (1993) Elements of a trophic model for IJsselmeer (The Netherlands), a shallow eutrophic lake. In V. Christensen & D. Pauly (eds) Trophic Models of Aquatic Ecosystems. ICLARM Conference Proceedings 26, 90–94. Buijse A.D. & Houthuijzen R.P. (1992) Piscivory, growth and size-selective mortality of age 0 pikeperch (Stizostedion lucioperca). Canadian Journal of Fisheries and Aquatic Sciences 49, 894–902. Buijse A.D. & van Densen W.L.T. (1992) Flexibility in the onset of piscivory and growth depensation in Eurasian perch, Perca fluviatilis L. In A.D. Buijse Dynamics and exploitation of unstable percid populations. PhD thesis, Wageningen Agricultural University, pp. 89–121. Carpenter S.R., Kitchell J.F. & Hodgson J.R. (1985) Cascading trophic interactions and lake productivity. Fish predation and herbivory can regulate ecosystems. BioScience 35, 634–639. Craig J.F. (1980) Growth and production of the 1955 to 1972 cohorts of perch Perca fluviatilis L., in Windermere. Journal of Animal Ecology 49, 291–315. De Beaufort L.F. (1954) Veranderingen in de flora en fauna van de Zuiderzee (thans IJsselmeer) na de afsluiting in 1932. Uitgave van de Nederlandse Dierkundige Vereninging, 359 pp. De Bie S. (1991) The concept of carrying capacity. In S. de Bie Wildlife resources of the West African Savanna. Wageningen Agricultural University Papers 91–2, 266 pp. Dekker W. (1991) Assessment of the historical downfall of the IJsselmeer fisheries using anonymous inquiries for effort data. In I.G. Cowx (ed.) Catch Effort Sampling Strategies: Their Application in Freshwater Fisheries Management. Oxford: Fishing News Books, Blackwell Science, pp. 233–240. Dekker W. (1996) Visstand en visserij op het IJsselmeer en Markermeer: de toestand in 1995. RIVODLO Rapport C001/96, 36 pp. (in Dutch).

08 11/02/2003 13:35 Page 116

116

Impact of birds on fisheries

Dekker W. & van Willigen J.A. (1993) De aal-visserij van het IJsselmeer; evaluatie van de toestand van het visbestand tot en met 1992. RIVO Rapport 93.011, 12 pp. (in Dutch). Dekker W. & Schaap L. (1993) De nettenvisserij op baars en snoekbaars van het IJsselmeer; evaluatie van de toestand van de visbestanden tot 1992. RIVO Rapport 93.005, 11 pp. (in Dutch). Dekker W., Schaap L. & van Willigen J. (1993) Bijvangsten in de fuiken-visserij op het IJsselmeer. RIVO Rapport 93-021, 30 pp. (in Dutch). Doornbos G. (1979) Winter food habits of smew (Mergus albellus L.) on Lake IJssel, The Netherlands: species and size selection in relation to fish stocks. Ardea 67, 42–48. Downing J.A. & Plante C. (1993) Production of fish populations in lakes. Canadian Journal of Fisheries and Aquatic Sciences 50, 110–120. Downing J.A., Plante C. & Lalonde S. (1990) Fish production correlated with primary productivity, not the morphoedaphic index. Canadian Journal of Fisheries and Aquatic Sciences 47, 1929–1936. Gayanilo F.C., Sparre P. & Pauly D. (1995) The FAO-ICLARM Stock Assessment Tools (FiSAT) User’s Guide. FAO Computerized Information Series (Fisheries). No. 8. Rome, FAO.126 pp. Gerking S.D. (ed.) (1978) Ecology of Freshwater Fish Production. Oxford: Blackwell, 520 pp. Hanson M.A. & Butler M.G. (1994) Responses of plankton, turbidity, and macrophytes to biomanipulation in a shallow prairie lake. Canadian Journal of Fisheries and Aquatic Sciences 51, 1180–1188. Holcik J. (1996) Ecological fish production in the inland delta of the Middle Danube, a floodplain river. Environmental Biology of Fishes 46, 151–165. Houthuijzen R.P. (1989) Voedsel en groei van juveniele snoekbaars, baars en spiering in het IJsselmeer in 1987 en 1988. Doktoraalverslag Landbouwuniversiteit Wageningen, Vakgroep Visteelt en Visserij o.b.v, 26 pp.  bijlagen (in Dutch). Ivanova M.N. (1981) On the life span of the smelt, Osmerus eperlanus, of Lake Beloye. Journal of Ichthyology 27, 91–98 Ivanova M.N. & Polovkova S.N. (1972) Types of spawning grounds and spawning ecology of the landlocked smelt (Osmerus eperlanus (L.)) in Rybinsk Reservoir. Journal of Ichthyology 12, 625–633. Ivanova M.N. & Volodin V.M. (1982) Variability in rate of sexual maturation in freshwater populations of the smelt, Osmerus eperlanus. Journal of Ichthyology 22, 28–37. Janss G. (1991) Verandering in de spieringaanvoer (Osmerus eperlanus) door de beroepsvisserij op het IJsselmeer. Doctoraalverslag RIVO, 36 pp. (in Dutch). Jolly G.M. & Hampton I. (1990) A stratified random transect design for acoustic surveys of fish stocks. Canadian Journal of Fisheries and Aquatic Sciences 47, 1282–1291. Koenders J.Th.H. & Mentink J.W.H.M. (1982) Het voedsel van ‘rode aal’ Anguilla anguilla (L.) in het IJsselmeer, de relatie hiervan met de bodemfauna en de invloed van de voedselkeuze op de kopvorm. RIVO Rapport ZA 82-02: 57 pp.  bijlagen (in Dutch). Lammens E.H.R.R. (1999) Het voedselweb van IJsselmeer en Markermeer. Veldgegevens, hypotheses, modellen en scenario’s Rapport Directoraat-Generaal Rijkswaterstaat, Rijksinstituut voor Intergraal Zoetwaterbeheer en Afvalwaterbehandeling, 160 pp. (in Dutch). Lammens E.H.R.R., de Nie H.W., Vijverberg J. & van Densen W.L.T. (1985) Resource partitioning and niche shifts of bream (Abramis brama) and eel (Anguilla anguilla) mediated by predation of smelt (Osmerus eperlanus) on Daphnia hyalina. Canadian Journal of Fisheries and Aquatic Sciences 42, 1342–1351. Liao H.S., Pierce C.L., Wahl D.H., Rasmussen J.B. & Leggett W.C. (1995) Relative weight (Wr) as a field assessment tool: Relationships with growth, prey biomass, and environmental conditions. Transactions of the American Fisheries Society 124, 387–400. Luecke C., Rice J.A., Crowder L.B., Yeo S.E. & Binkowski F.P. (1990) Recruitment mechanisms of bloater in Lake Michigan: An analysis of the predatory gauntlet. Canadian Journal of Fisheries and Aquatic Sciences 47, 524–532. Mills E.L., Pol M.V., Sherman R.E. & Culver T.B. (1989a) Interrelationships between prey body size and growth of age-0 yellow perch. Transactions of the American Fisheries Society 118, 1–10.

08 11/02/2003 13:35 Page 117

Interaction in the utilisation of small fish by piscivorous fish and birds

117

Mills E.L., Sherman R. & Robson D.S. (1989b) Effect zooplankton abundance and body size on growth of age-0 yellow perch (Perca flavescens) in Oneida Lake, New York, 1975–1986. Canadian Journal of Fisheries and Aquatic Sciences 46, 880–886. Mills E.L. & Schiavone A. (1982) Evaluation of fish communities through assessment of zooplankton populations and measures of lake productivity. North American Journal of Fisheries Management 2, 14–27. Mills E.L., Green D.M. & Schiavone A. (1987) Use of zooplankton size to assess the community structure of fish populations in freshwater lakes. North American Journal of Fisheries Management 7, 369–378. Miranda L.E. & Hubbard W.D. (1994) Length-Dependent winter survival and lipid composition of age-0 largemouth bass in Bay Springs reservoir, Mississippi. Transactions of the American Fisheries Society 123, 80–87. Mooij W.M. (1996) Variation in abundance and survival of fish larvae in shallow eutrophic lake Tjeukemeer. Environmental Biology of Fishes 46, 265–279. Mous P.J. (2000) Interactions Between Fisheries and Birds in IJsselmeer, The Netherlands. PhD thesis, Wageningen Agricultural University, 205 pp. Nagtegaal P. & Snel H.J. (1984) De beroepsvisserij in het IJsselmeergebied. Rijksdienst voor de IJsselmeerpolders. Ministerie van Verkeer en Waterstaat. Rapport 1984 – 14 abw, 76 pp. (in Dutch). Ogutu-Ohwayo R. (1990) The decline of the native fishes of lakes Victoria and Kyoga (East Africa) and the impact of introduced species, especially Nile perch, Lates niloticus, and the Nile tilapia, Oreochromis niloticus. Environmental Biology of Fishes 27, 81–96. Parsons T.R. (1996) The impact of industrial fisheries on the trophic structure of marine ecosystems. In G.A. Polis & K.O. Winemiller (eds) Food Webs. Integration of Patterns & Dynamics. London: Chapman & Hall, pp. 352–357. Paulisse J. (1993) Het voedselgedrag van aal (Anguilla anguilla) in het IJsselmeer in relatie tot plaats, grootte, sexe en morfologie. Doktoraalverslag Vakgroep Visteelt en Visserij. Landbouwuniversiteit Wageningen. Nr. 1432, 76 pp. (in Dutch). Pauly D. (1986) A simple method for estimating the food consumption of fish populations from growth data and food conversion experiments. Fisheries Bulletin 84, 827–840. Pet J.S., Gevers G.J.M., van Densen W.L.T. & Vijverberg J. (1996) Management options for a more complete utilisation of the biological fish production in Sri Lankan reservoirs. Ecology of Freshwater Fish 5, 1–14. Piet G.J., Vijverberg J. & van Densen W.L.T. (1999) Foodweb structure of a Sri Lankan reservoir. In W.L.T. van Densen & M.J. Morris (eds) Fish and Fisheries of Lakes and Reservoirs in Southeast Asia and Africa. Otley: Westbury Academic & Scientific Publishing, pp. 187–205. Platteeuw M., Koffijberg K. & Dubbeldam W. (1995) Growth of cormorant Phalacrocorax carbo sinensis chick in relation to brood size, age ranking and parental fishing effort. Ardea 83, 235–245. Platteeuw M. & Van Eerden M.R. (1995) Time and energy constraints of fishing behaviour in breeding cormorants Phalacrocorax carbo sinensis at Lake IJsselmeer, the Netherlands. Ardea 83, 223–234. Prins K.H., Klinge M., Ligtvoet W. & de Jonge J. (1995) Biologisch monitoring zoete rijkswateren. IJsselmeer en Markermeer 1992. RIZA Nota nr. 94.060, 81 pp. (in Dutch). Randall R.G., Kelso J.R.M. & Minns C.K. (1997) Fish production in freshwaters: Are rivers more productive than lakes? Canadian Journal of Fisheries and Aquatic Sciences 52, 631–643. Ricker W.E. (1975) Computation and interpretation of biological statistics of fish populations. Department of Fisheries and Oceans, Ottawa Canada, Bulletin 191, 382 pp. Ryder R.A. (1977) Effects of ambient light variations on behaviour of yearling, subadult, and adult walleyes (Stizostedion vitreum vitreum). Journal of the Fisheries Research Board of Canada 34, 1481–1491. Schouten C. (1983) Het IJsselmeergebied als ruiplaats voor de zwarte stern (Chlidonias niger). Een onderzoek naar de conditie, rui en doortrek van de zwarte stern in het IJsselmeergebied. RIJP Rapport, 33 pp.  annexes (in Dutch).

08 11/02/2003 13:35 Page 118

118

Impact of birds on fisheries

Sparre P. & Venema S.C. (1992) Introduction to tropical fish stock assessment. Part I – Manual. FAO Fisheries Technical Paper 306/1, 376 pp. Sprules W.G., Brandt S.B., Stewart D.J., Munawar M., Jin E.H. & Love J. (1991) Biomass size spectrum of the Lake Michigan food web. Canadian Journal of Fisheries and Aquatic Sciences 48, 105–115. Stam M.A. (1996) Ruimtelijke en temporele patronen van vistetende vogels en prooivis in het IJsselmeergebied. Doctoraalscriptie Vakgroep Visteelt en Visserij 1501, 79 pp. (in Dutch). Steinmetz B. & Oudelaar H.G.J. (1971) IJsselmeer proefvisserijen met de grote kuil. Jaren 1966 t/m 1969. Hoofdafdeling Sportvisserij en Beroepsbinnenvisserij Utrecht, documentatie-rapporten no. 13, 13 pp. (in Dutch). Taal C. & de Wilde J.W. (1997) Perspectieven IJsselmeervisserij. Sociaal-economische verkenning. LEI-DLO Mededeling 597, 53 pp. (in Dutch). Timmerman J.G. & Prins K.H. (1996) Biologische monitoring zoete rijkswateren 1994. RIZA Nota nr. 96.009, 60 pp. (in Dutch). Van Dam C., Buijse A.D., Dekker W., van Eerden M.R., Klein Breteler J.G.P. & Veldkamp R. (1995) Aalscholvers en beroepsvisserij in het IJsselmeer, het Markermeer en Noordwest Overijssel. Ministerie van Lanbouw, Natuurbeheer en Visserij, Rapport IKC Natuurbeheer Nr. 19, 104 pp. (in Dutch). Van Densen W.L.T. (1985) Feeding behaviour of major 0 fish species in a shallow, eutrophic lake (Tjeukemeer, The Netherlands). Journal of Applied Ichthyology 1, 49–70. Van Densen W.L.T., Ligtvoet W. & Roozen R.W.M. (1996) Intra-cohort variation in the individual size of juvenile pikeperch, Stizostedion lucioperca, and perch, Perca fluviatilis, in relation to the size spectrum of their food items. Annales Zoologici Fennici 33, 495–506. Van Eerden M.R. (1993) Aalscholvers hebben het moeilijk in het IJsselmeer. Helderheid is ook niet alles. Van Nature 3, p. 11 (in Dutch). Van Eerden M.R., Zijlstra M. & Munsterman M.J. (1991) Factors determining the breeding success in cormorants Phalacrocorax carbo sinensis. In M.R. van Eerden & M. Zijlstra (eds) Proceedings Workshop 1989 on Cormorants Phalacrocorax carbo. Rijkswaterstaat Directorate Flevoland, pp. 67–73. Van Eerden M.R. & Gregersen J. (1995) Long-term changes in the northwest European population of cormorants Phalacrocorax carbo sinensis. Ardea 83, 61–79. Van Zuilekom W.J. (1991) Invloed van temperatuur en zooplanktonaanbod op de groei van spiering (Osmerus eperlanus) en juveniele baars (Perca fluviatilis) in het IJsselmeer. Doktoraalverslag Vakgroep Visteelt en Visserij. Landbouwuniversiteit Wageningen, 95 pp. (in Dutch). Veen A., Vijverberg J. & Mooij W.M. (1997) De populatie ontwikkeling van ‘large-bodied’ Daphnia’s in het IJsselmeer en Markermeer. Report of the Netherlands Institute of Ecology. 71 pp. (in Dutch). Voslamber B. (1991) Meeuwen in het IJsselmeergebied. Voedseloecologie van zeven op het meer voorkomende soorten. RWS-DF Intern Rapport 37, 51 pp.  appendices (in Dutch). Wanink J.H. & P.C. Goudswaard (1994) Effects of Nile perch (Lates niloticus) introduction into Lake Victoria, East Africa, on the diet of pied kingfishers (Ceryle rudis). Hydrobiologia 279/280, 367–376. Willemsen J. (1977) Population dynamics of percids in Lake IJssel and some smaller lakes in The Netherlands. Journal of the Fisheries Research Board of Canada 34, 1710–1719. Willemsen J. (1985) De invloed van de visserij met fuiken op de snoekbaars- en baarsstand in het IJsselmeer. RIVO-rapport BW85-02, 10 pp. (in Dutch). Winter H.V. (1994) Verspreiding in ruimte en tijd van vistetende vogels in het IJsselmeergebied in relatie tot de visstand. Landbouwuniversiteit Wageningen, Verslag nr. 1445, 77 pp. (in Dutch). Witte F., Goldschmidt T., Goudswaard P.C., Ligtvoet W., Oijen M.J.P. van & Wanink J.H. (1992) Species extinction and concomitant ecological changes in Lake Victoria. Netherlands Journal of Zoology 42, 84–102.

09 11/02/2003 13:39 Page 119

Chapter 9

A quantitative assessment of the impact of goosander, Mergus merganser, on salmonid populations in two upland rivers in England and Wales B.R. WILSON* Environment Agency Wales, Penyfai House, Furnace, Llanelli, UK

M.J. FELTHAM School of Biological and Earth Sciences, Liverpool JMU, Liverpool, UK

J.M. DAVIES School for Professional and Continuing Education, University of Birmingham, Birmingham, UK

T. HOLDEN School of Biological and Earth Sciences, Liverpool JMU, Liverpool, UK

I.G. COWX and J.P. HARVEY University of Hull, International Fisheries Institute, Hull, UK

J.R. BRITTON Environment Agency, National Fisheries Laboratory, Brampton, Huntingdon, UK

Abstract Fish-eating birds are frequently considered to be deleterious to fish stocks, but quantitative estimates of impact are rare, particularly for goosander, Mergus merganser L., in England and Wales. The diet and distribution of goosanders was assessed during three breeding seasons (1996–1998) on two upland river systems, the River Wye in mid-Wales and the River Hodder in north-west England. A Monte Carlo Simulation model was used to produce estimates, with confidence limits, of salmonid losses from these rivers. These values were related to fisheries biomass estimates to determine the likely proportion of standing crop removed by goosander broods. Goosander productivity showed marked annual variation in the Wye study area, but remained relatively stable on the Hodder. Stomach content analysis of birds shot on the Wye (n  11 birds in 1996 only) showed duckling diet to be dominated by salmon, Salmo salar L., (63% by mass), with the remainder comprising brown trout, Salmo trutta L., minnow, Phoxinus phoxinus (L.), eel, Anguilla anguilla (L.), bullhead Cottus gobio L., and stoneloach, Barbatula barbatula (L.). This was in marked contrast to the Hodder (n  21 birds over three years), where salmon accounted for only 14% by mass, depending on area. The total mass of salmonids removed by goosanders on the Wye study site fell from 21.3 to 4.6 kg ha1 during the course of the study. Estimates for the River Hodder were consistently lower, 3.4– 4.0 kg ha1. The proportion of the juvenile salmonid standing crop removed was as high as 683% on the Wye in a single year. Where

*Correspondence: Ben Wilson, Environment Agency Wales, Penyfai House 19 Penyfai Lane, Furnace, Llanelli SA15 4EL, UK (email: [email protected]).

09 11/02/2003 13:39 Page 120

120

Impact of birds on fisheries estimates were corrected for predation mortality, the proportions of salmonid biomass removed were between 67% and 172% on the Wye and between 29% and 39% on the Hodder. The possible reasons for these high estimates of salmonid losses, and their implications, are discussed. Keywords: diet, depredation, fish mortality, Mergus merganser, salmon.

9.1 Introduction In the UK, there has been increasing concern among anglers and fishery managers about the potential impacts of fish-eating birds on inland fisheries (Russell et al. 1996). Whereas cormorants, Phalacrocorax carbo L., are the most commonly cited cause for concern, particularly at large stillwater fisheries, the increasing abundance and distribution of goosander, Mergus merganser L., in the UK has caused concern in salmonid (Atlantic salmon, Salmo salar L., and brown trout, Salmo trutta L.) fisheries (Marquiss & Carss 1994; Russell et al. 1996). Since 1897, when goosanders were first recorded breeding in the UK, birds have moved southward to breed on upland stretches of many rivers. The current breeding population is estimated to be more than 2600 pairs (Gregory et al. 1997) and is still expanding in areas such as Wales and south-west England (Wernham et al. 1999). In the UK, goosanders are protected under the EU Birds Directive (EEC/79/409) through the Wildlife and Countryside Act 1981. Licensed shooting is allowed ‘for the purposes of preventing serious damage to … fisheries’. Current legislation does not, however, give criteria for ‘serious damage’ and so quantifying the effects of avian predation on fish populations is important, as the magnitude of any damage they might cause could have serious implications for the future management of a fishery. To date, most attempts to estimate the impact of goosanders on salmonid fisheries have been equivocal, being frequently undermined by invalid or untested assumptions (see Marquiss & Carss (1994) and Russell et al. (1996) for detailed review). Predation by goosanders on juvenile salmonids comprises two main components. Most attempts to quantify levels of goosander consumption have focused on the potential impacts of adult birds on the seaward migration (smolt run) of salmon (Shearer 1984; Carter & Evans 1986; Howell 1987; Shearer et al. 1987; Wood 1987a). Fewer studies have attempted to quantify the potential impacts of goosander broods on stream resident populations of juvenile salmonids. One of the few studies to address this issue showed that goosander broods have the potential to consume significant proportions of juvenile fish stocks. Brood predation on Pacific coho salmon, Oncorhynchus kistuch (Walbaum) in some Canadian rivers was estimated to be 82 000–131 000 fry yr1, which, assuming that predation was additive rather than compensatory, was equivalent to between 24 and 65% of the observed annual smolt production (Wood 1987b). More recent studies in Scotland showed high levels of goosander predation (up to 1.85% of salmon standing stock per day) on parts of the River Dee (Marquiss et al. 1998). No similar studies have been conducted in England and Wales. Consequently, a 3-year study was undertaken on two river systems, as part of the ‘Case studies of the impact of fish-eating birds on inland fisheries in England and Wales’ (MAFF project no. VC 0106 (Feltham et al. 1999)).

09 11/02/2003 13:39 Page 121

Impact of goosanders on salmonid populations

121

9.2 Materials and methods 9.2.1

Study sites – Rivers Wye and Hodder

Studies were conducted on the River Hodder in north-west England and the River Wye in mid-Wales (Fig. 9.1). The River Wye rises at Plynlimon, Powys, mid-Wales (SN 802 871) and flows south-east towards its confluence with the Severn estuary for an approximate distance of 250 km. The river has a catchment area of approximately 4183 km2 and supports coarse fisheries in its lower reaches and salmonid fisheries throughout. Total rod catches of salmon on the Wye have declined dramatically in recent decades, with average catches falling from 5000 fish yr1 in the 1970s to less than 1000 per annum in 2000 (Environment Agency catch returns 1960–2000). This decline in catches relates predominantly to declining runs of the large multi-sea winter (MSW) salmon, for which the Wye rod fishery is famous (EA 1999). The study detailed in this chapter focused on the upper reaches of the Wye catchment, and included 27 km of the main river between Builth Wells (SO 043513) and Rhayader (SN 968678) and a 6 km stretch of the River Irfon between its confluence with the Wye at Builth (SO 034516) to Llangammarch Well Bridge (SN 935473).

Figure 9.1

Location of study sites within the rivers Wye and Hodder catchments

09 11/02/2003 13:39 Page 122

122

Impact of birds on fisheries

The River Hodder is an important tributary of the River Ribble in north-west England. Although not as prolific a salmon rod fishery as the Wye, the Ribble catchment is a fishery of both local and national importance. Rod catches of salmon in the catchment have, moreover remained relatively stable, with a current annual average of about 620 fish. Unlike the Wye, The Ribble also supports a significant sea trout, Salmo trutta (L.), fishery, with an average catch of 1240 fish. The Hodder catchment drains the eastern side of the Bowland Fells, Lancashire and is 38 km long, 30 km of which was included in the study, from Hodder Foot (SD 711382) to Stock’s Reservoir (SD 717544).

9.2.2

Bird surveys

Both rivers were divided into count sections of approximately equal length, with six on the River Wye (including one section on the River Irfon) and five on the Hodder. Monthly or bi-monthly (Wye 1996 only) coordinated counts were conducted at each study site during the goosander breeding season (April–September) between 1996 and 1998. Counts began at first-light and were typically completed 2–3 h later. During counts, the approximate locations of goosanders were recorded, along with the time of sighting. Bird activities were assigned to one of the following categories: feeding, flying and loafing/roosting. On rivers, direction of flying birds was noted to allow adjustments to be made for birds overflying contiguous count sections which minimised the likelihood of double counting. Goosanders were assigned to one of seven cohorts: paired birds, unpaired adult males, immature males, unpaired females, females with young, the young themselves and unaged ‘redheads’. The latter refers to well grown young/females that could not be aged with confidence during counts. The numbers of ducklings in broods were also recorded and the approximate age of ducklings assessed. Ducklings were assigned to one of four categories based on size relative to adult birds: one-third grown  class 1; two-thirds grown  class 2; fully grown unfledged  class 3; and fledged  class 4 (after Underhill 1993). As juvenile goosanders fledge at 60–70 days (Cramp & Simmons 1977), age classes 1–3 approximate to 3 weeks each. Diet was assessed from the stomach contents of birds shot under licence from English Nature or the Countryside Council for Wales. On the Wye, juvenile goosanders were shot in 1996 only, and on the Hodder samples were taken each summer. Shooting was conducted late in the breeding season (July–August) when young were likely to be well grown and the greatest impact would be taking place. Prey remains were dissected from the gizzard and proventriculus. Intact fish, partially digested fish and bones resistant to digestion were used to estimate the minimal numbers of each species found within the guts of birds and the sizes of fish eaten (Feltham 1990). The mass of fish in stomachs was obtained either by direct weighing or from length–weight relationships (see Feltham et al. 1999).

9.2.3

Fishery surveys

Hand-held electric fishing surveys were conducted on both rivers to determine both the relative species compositions of the fisheries and an estimate of the biomass of

09 11/02/2003 13:39 Page 123

Impact of goosanders on salmonid populations

123

salmonids present. Surveys on the Wye took place in July 1997 and September 1998 at two sites: Rhayader (SN 969677) and Brynwern Bridge (SN 011565). On the Hodder, surveys took place at three sites: Slaidburn (SD 715524), Newton (SD 698504) and Burholme Bridge (SD 658480). Quantitative assessment of the salmonid population sizes, using a three-catch depletion method (Zippin 1959; Cowx 1983) were conducted annually at one of the Hodder sites. In each year the estimated catch efficiency at this site was used to determine the population estimates at the two other sites on the Hodder. As no quantitative estimates were derived empirically on the Wye, a mean catch efficiency of 0.41 was used, derived from the Hodder surveys.

9.2.4

Impact estimates

Although not necessarily a measure of impact per se, the first step towards producing a quantitative estimate of impact was to determine consumption of key fish species by birds. Consumption estimates have frequently been based on simple models, such as the intuitively derived equation Yi  N  d  c  pi

(9.1)

where Yi is the mass of species i removed from a fishery by birds, N is the number of birds present at a fishery, d is the number of days they are present, c is the birds’ daily food intake and pi is the proportion of species i in the diet of the birds (e.g. Suter 1995). The use of such simple equations has been criticised for assuming constants to be temporally and/or spatially constant and for failing to recognise any potential uncertainty in these values by failing to place meaningful confidence limits on final estimates (Marquiss & Carrs 1994; Feltham 1995). A different approach was therefore adopted for this study, in which Equation (9.1) was used as a building block for a Monte Carlo Simulation Model.

9.2.5

Constructing ‘building blocks’

Instead of deriving single values of N, c and pi to represent, for example, an entire study site for the course of a year, N, c and pi were estimated monthly for each study site. The building-block approach is hence a modification of Equation (9.1) of the form (9.2) Yi  yi1  yi2  yi3  …  … yin where, yi1, …, yin represent individual estimates of yi at any one study site, where n  6 (the number of months studied in a given year).

9.2.6

Monte Carlo simulations

Monte Carlo Simulations (MCS) were used to produce estimates, with confidence limits, of monthly and annual fish consumption. Instead of using single values for each

09 11/02/2003 13:39 Page 124

124

Impact of birds on fisheries

parameter of the ‘building block’ model a sample distribution was used. These distributions were based on either empirically derived data from the survey sites (i.e. count and diet data) or on estimates published elsewhere (i.e. daily food intake). These were derived as follows. N – the number of birds feeding at the fishery (normal sample distributions were used) was based on the mean of the observed densities of goosanders (birds km1) in each age group. c – the daily food intake (DFI) or daily food requirement (DFR) for adult females and fledged juveniles was based from published data derived from empirical (Feltham 1995) or theoretical (Nillson & Nillson 1976) energetics studies. So as to not place any particular importance for a given published estimate, a uniform distribution was used between cmin (0.25 kg d1) and cmax (0.45 kg d1), values that represented the range of published DFI estimates. To estimate c for goosander ducklings, it was necessary to incorporate the effects of growth. There were no published estimates, based on energetics studies, of duckling DFR. Therefore, estimates were based on published growth curves (Erskine 1971), with estimates of the DFR relative to body weight derived from observations of wild and captured birds (Wood 1987b). By combing these rate estimates it was possible to construct maximum and minimum DFR curves for ducklings (Fig. 9.2). These curves were used to estimate the minimum and maximum monthly food intake for each age class of duckling. The figures were equivalent to median DFR values of 100 g d1 for one-third-grown young, 205 g d1 for two-thirds-grown young and 255 g d1 for fullygrown, pre-fledged young. Pi, the proportion of prey type i in the diet by weight, was square root arcsine transformed and used to calculate means and standard deviations from which normal probability distributions were generated. These distribution were randomly sampled (n  1000) to produce a range of consumption estimates (total mass and salmonids) for each month. These were combined to give ranges of total consumption estimates for each season (for more detailed methodology see Feltham et al. 1999 and Wilson et al. in press).

Image Not Available

Figure 9.2 Estimated DFR (DFI) of goosander ducklings based on mean duckling masses derived from Erskine (1971) and DFR as a proportion of body mass, from Wood (1987)

09 11/02/2003 13:39 Page 125

Impact of goosanders on salmonid populations

9.2.7

125

Incorporating estimates of fish biomass

Estimates of the proportion of standing crop of salmonids removed each season were also calculated, again using MCS to produce confidence limits. Values of salmonid biomass (bi) were sampled from normal distributions, based on mean densities from electric fishing surveys for each season. These were applied to the consumption estimates (Yi) for a given season/site, using the following equation: Pbi  Yi /bi

(9.3)

Thus 1000 estimates of Pbi were produced for each river and year, from which medians and confidence limits could be derived.

9.3 Results 9.3.1

Counts

In 1996 and 1997, the first broods on the Wye were recorded during May, but they were not seen until June in 1998. Peak counts occurred in June or July of each year (nine, eight and five broods in each year, respectively) as broods produced in this part of the river were joined by young birds from sections further upstream. Similarly, peak brood counts on the Hodder occurred in either June or July (four, three and five in each year, respectively). Brood sizes showed some temporal and spatial variation. Although the median brood size on the Wye was higher in 1997 (eight young) compared with other years (four young), annual variation was not significant (H2  1.43, P  0.49). On the Hodder, however, there was a significant increase in the median brood size from five young in 1996 to eight and 10 young in the next two years (H2  6.99, P  0.03). There was, moreover, a significant difference in overall brood size between the two rivers, with a higher median on the Hodder (eight birds) compared with the Wye (five birds) (H1  8.59, P  0.003). On the River Wye, peak total counts of ducklings and adult females declined progressively during the study, from 92 birds in 1996 to 74 and 32 birds in 1997 and 1998. On the Hodder, counts were similar in the first 2 years, with peaks of 34 and 31 in 1996 and 1997, rising to 46 birds in 1998.

9.3.2

Densities

On the River Wye there were marked annual differences in the mean densities of both juvenile and adult birds (F2,99  6.19, P  0.003, F2,99  4.93, P  0.009) (Fig. 9.3). Adult densities were higher in 1996 (0.92 birds km1) than in either of the subsequent years (Fig. 9.3). Juvenile densities were significantly lower in 1998 (0.16 birds km1) compared with the two previous years (Fig. 9.3). On the Hodder there were no significant annual differences in either juvenile or adult densities (F2,40  0.5, P  0.61, F2,40  0.2, P  0.82) (Fig. 9.3).

09 11/02/2003 13:39 Page 126

126

Impact of birds on fisheries

1.4

(a) adult and post fledged juveniles

Density (birds/km)

1.2 Wye Hodder

1 0.8 0.6 0.4 0.2 0

1996

1997

1998

Year 2.5 (b) juveniles (pre-fledged)

Density (birds/km)

2

1.5 Wye Hodder 1

0.5

0

1996

1997

1998

Year

Figure 9.3 Mean (SD) densities of (a) adult and post fledged juvenile, and (b) juvenile (pre-fledged), goosanders on the Wye and Hodder 1996–1998

Overall mean adult densities were higher on the Wye (0.96 birds km1) than on the Hodder (0.22 birds km1), although the differences were not statistically significant (F1,141  7.34, P  0.08). There was little difference in the overall juvenile densities on the two rivers (mean densities 0.77 and 0.73 on the Wye and Hodder respectively, F1,141  0.03, P  0.86).

9.3.3

Diet

Eleven goosanders were shot on the Wye in July/August 1996, of which 10 contained food remains. All birds contained remains of salmonids in the 40–90 mm size range. Other prey remains included minnow, Phoxinus phoxinus (L.), eel, Anguilla anguilla (L.), bullhead, Cottus gobio (L.), and stoneloach, Barbatula barbatula (L.). The median

09 11/02/2003 13:39 Page 127

Impact of goosanders on salmonid populations

127

Proportion in sample

(a) Upper Wye 0.6

Electric fishing (n = 1099 fish)

0.5

Goosanders (n = 375 fish)

0.4 0.3 0.2 0.1 Other

Eel

Stone loach

Bullhead

Minnow

“salmonid”

Brown trout

Salmon

0

Species

Proportion in sample

0.7

(b) River Hodder Electric fishing (n = 229 fish)

0.6

Goosanders (n = 83 fish)

0.5 0.4 0.3 0.2 0.1 Chub

Eel

Stone loach

Bullhead

Minnow

Brown trout

Salmon

0

Species

Figure 9.4 Species composition of electric fishing surveys and the diet of goosander broods on: (a) the upper Wye, and (b) the River Hodder 1995–1998

number of fish recovered from each bird was five, with a maximum of 25 (16 of which were salmon fry). In total the remains of 83 fish were identified, 58 of which were salmonids (54 salmon and four trout). Of the remaining items, there were 19 minnows, two eels, two bullheads and two stoneloach (Fig. 9.4(a)). On the Hodder, a total of 22 birds were shot over the three-year period, all of which contained fish remains. Two birds contained large numbers (20 and 33) of trout fry and were thought to have fed at a nearby fish hatchery, thus these two samples were excluded from subsequent analyses. The number of prey items identified from Hodder birds (median 16.5) was higher than from those on the Wye. Of the 375 prey remains identified from these birds, the most frequently occurring species was minnow, followed by bullhead and stoneloach (Fig. 9.4(b)). Only 28 of the identified remains were

09 11/02/2003 13:39 Page 128

128

Impact of birds on fisheries

salmonid fry (17 salmon, six trout and five remains not identifiable to species level). Eel and 3-spined stickleback, Gasterosteus aculeatus L., were also identified. Impact estimates were derived for all salmonids (trout and salmon combined) using the mean proportion by weight in stomach samples. These values were 65% (SD 37%, n  10) for the Wye and 27% (SD 20%, n  22) for the Hodder.

9.3.4

Fisheries surveys/goosander diet comparison

Electric fishing catches on the Wye were numerically dominated by minnow (53%) and salmon (31%). Other species recorded included brown trout, bullhead, stoneloach and chub, Leuciscus cephalus (L.) (Fig. 9.4(a)). Goosanders took a proportionately greater number of juvenile salmon and fewer minnow compared with those caught by electric fishing (pooled sample; 2  30.998, P  0.01) (Fig. 9.4(a)). Fisheries surveys on the Hodder were dominated by minnow (37%) and bullhead (21%) (Fig. 9.4(b)). Salmon and trout accounted for only 17% and 10% of catches respectively. Goosanders took a proportionately smaller number of juvenile salmonids (salmon and trout) and more minnows, bullhead and stoneloach compared with electric fishing catches (pooled sample; 2  67.1, P  0.001) (Fig. 9.4(b)). Estimates of the total fish biomass on the Wye were higher in 1997 (4.82 kg ha1) than in 1998 (2.85 kg ha1). In both years salmonids (mostly salmon) accounted for the greatest proportion of the estimated biomass (70% and 81% in 1997 and 1998 respectively) (Table 9.1). Estimates of total fish biomass on the Hodder were higher than on the Wye in all years, as were estimates for salmonids, which increased from 5.23 kg ha1 in 1996 to 10.53 kg ha1 in 1998 (Table 9.1).

9.3.5

Impact estimates

The levels of goosander depredation on the River Hodder were highest in 1998 (median 20.2 kg ha1), although total losses in previous years were only slightly lower (medians 17.6 and 16.3 kg ha1 in 1996 and 1997 respectively). Diet comprised mostly

Table 9.1

Mean salmonid biomass estimates based on electric fishing surveys Mean (SD) Surveyed biomass of salmonids (kg ha1)

Year

Wye

Hodder

1996 1997 1998

2.83 (2.35) n  2a 3.65 (2.35) n  2 2.30 (2.28) n  2

5.23 (4.40) n  9b 5.46 (3.01) n  3 10.53 (7.75) n  3

a

Average from 1996 based on mean of 1997 and 1998. Three sites fished three times.

b

09 11/02/2003 13:39 Page 129

Impact of goosanders on salmonid populations

129

Table 9.2 MCS estimates of the annual variation in the mass of salmonids (kg ha1) and the proportion of salmonid standing crop biomass removed by goosander broods on the River Hodder, April–September 1996–1998. C  total consumption, S  median salmonid standing crop Median (interquartile range) mass of fish removed by goosanders (kg ha1)

Proportion of standing crop removed by goosanders (%)

Year

Total

Salmonids

(C/S)  100

C/(C  S)  100

1996 1997 1998

17.6 (11.2–24.0) 16.3 (12.3–21.2) 20.0 (16.9–31.1)

3.5 (2.1–5.3) 3.4 (2.4–4.7) 4.0 (2.1–6.8)

50.2 (21–106) 60.3 (37–102) 34.6 (13–70)

39 (24–58) 38 (28–52) 29 (15–47)

Table 9.3 MCS estimates of the annual variation in the mass of fish (kg ha1) removed by goosanders and the proportion of salmonid standing crop removed from the River Wye between Rhayader and Builth Wells, April–September 1996–1998. C  total consumption, S  median salmonid standing crop Median (interquartile range) mass of fish removed by goosanders (kg ha1)

1996 1997 1998

Total

Salmonids

33.3 (21.6–47.0) 27.7 (17.2–40.8) 7.8 (3.9–12.3)

21.3 (11.1–26.6) 18.1 (11.1–26.6) 4.6 (2.5–7.9)

Proportion of standing crop removed by goosanders (%) (C/S)  100

C/(C  S)  100

683 (400–1082) 607 (337–929) 152 (73–263)

172 (119–272) 116 (95–146) 67 (46–81)

coarse fish with salmonids making up between 3.4 and 4.0 kg ha1 only. The mean standing crop biomass of salmonids was 5.4 kg ha1 in 1996, 5.6 kg ha1 in 1997 and 10.9 kg ha1 in 1998. The proportion of this standing crop removed by goosanders was substantial, ranging from 34.6 to 60.3%. The proportion of pre-predation standing crop (i.e. consumption/biomass of fish consumed  estimated standing crop) removed ranged between 29 and 39% (Table 9.2). Estimates of the total mass of fish consumed by goosanders on the Wye were of comparable magnitude, but showed greater annual variation, ranging from 7.8 kg ha1 in 1998 to 33.3 kg ha1 in 1996 (Table 9.3). Salmonid consumption on the Wye was higher than on the Hodder in all years, markedly so in 1996 and 1997 (Table 9.3). Goosander predation accounted for between 152 and 683% of the salmonid standing crop based on electric fishing surveys conducted during late summer. The proportion of pre-predation standing crop removed ranged from between 67 and 112% (Table 9.3).

9.4 Discussion Estimates of goosander depredation on stream-resident juvenile salmonids differed markedly between the two river systems studied. On the River Wye, higher bird densities, combined with a greater proportion of salmonids in the diet (young of year salmon in

09 11/02/2003 13:39 Page 130

130

Impact of birds on fisheries

particular), meant that in the first 2 years of the study, consumption estimates were noticeably higher than on the River Hodder. Only in 1998, with a dramatic decrease in goosander productivity on the Wye, were consumption estimates for the two rivers similar. Goosander numbers, and hence fish consumption, appeared to vary markedly on the Wye, whereas annual variation on the Hodder was much lower. The reasons for the poor goosander productivity of the Wye in 1998 are likely to be numerous and lay outwith the scope of this study. The weather conditions during the counts conducted early in the breeding season were, however, relatively poor compared with previous years, with higher rainfall and consequently higher river levels (personal observation). This may have resulted in higher brood mortality, or may have limited the distribution of birds to smaller, unsurveyed tributaries. The relatively short nature of this study precludes any thorough investigation of the differences in goosander productivity between the two river systems. The degree to which it affects consumption (and hence impact) estimates suggests further investigations are needed. The most notable aspect of this study was that estimates of the proportion of salmonid biomass removed were high. The proportion of salmonids removed from the Hodder (34–60%) could be considered high, and may be unsustainable in terms of fish production (see later). On the Wye, however, where consumption estimates were higher and fish biomass estimates lower, the proportions of salmonid biomass removed were much higher than on the Hodder. Indeed, estimates of the proportion of standing crop removed indicate goosanders consumed up to seven times the estimated standing crop of salmonids. Even when estimates are calculated using pre-predation biomass estimates, consumption represented between 172% and 116% of the available stock in 1996 and 1997 respectively. In other words, within the breeding seasons of 1996 and 1997, goosanders consumed all of the salmonid fish present in this stretch of the River Wye. Although intuitively impossible (the Wye still produces a run of salmon) such findings are not unique to this study. Marquiss et al. (1998) produced similar results when estimating goosander consumption on the River Dee in Scotland. Using a relatively simple model to estimate the amount of salmonids removed by goosanders, they found that the highest values came from ducklings on the wider parts of the Dee in July when birds were not yet fully grown. During this time it was estimated that up to 1.85% of standing crop per day was removed by birds. In August, on the same river, when ducklings were fully grown the estimated salmon consumption rate was 60 fish per day. They pointed out that on the River Dee the levels of depredation recorded would ‘result in all fish being removed from the river within 2 months’, which, as the authors pointed out ‘clearly doesn’t happen’. Marquiss et al. (1998) suggested three possible explanations for this apparent anomaly. These may have some bearing on the observed levels of depredation found on the Wye and may also indicate why similar values were not observed on the Hodder.

9.4.1

Consumption is overestimated

For there to be an overestimate in the mass of salmonids consumed, one or more of the model parameters, or the assumptions on which they were are based, would have to be invalid. These parameters will be considered in turn.

09 11/02/2003 13:39 Page 131

Impact of goosanders on salmonid populations

131

Daily food requirement In both the current and Scottish studies, the DFR values for adults were similar (Marquiss et al. 374 g, current study a uniform distribution between 250 and 450 g). For ducklings, however, there were significant differences between the two studies. Marquiss et al. assumed the daily food requirement (DFR) to be 25% of body mass, based on published allometric equations for a range of species, and using a mean duckling weight of birds sampled in July (496 g), used a constant DFR of 117 g d1. In the current study, DFR was derived for different age classes of birds, based on published growth data (Erskine 1971) and mass related DFR estimates (Wood 1987). The average of the age-related DFR values used in the current study (200 g) was, however, higher than the constant value (117 g) used by Marquiss et al. The higher values used in the current study can be justified in that: (i) they incorporated into the DFR not only energetic requirements but also those required for somatic growth; and (ii) the resulting curve predicts an adult DFR similar to those derived separately for adults in both studies. The values used, moreover, represent the best available data relating to DFR. Also, the use of a uniform distribution between minimal and maximal DFR estimates means that neither low nor high values were likely to unduly influence the results.

Numbers of birds N, the numbers of birds counted on surveys, were probably minimum rather than maximum estimates and so might be expected to lead to an underestimate in consumption. This is because some birds could have been missed during counts, if for example, ducklings were under cover or on side streams. It is, however, difficult to see how numbers could have been overestimated. The only potential source of error of this kind could be double-counting, but as ducklings cannot fly and travel around in broods, the likelihood of this occurring is low. N could also be overestimated when applied to each month because of duckling mortality. It is implicit within the model that mortality: (i) takes place, on average, mid-way between monthly counts; and (ii) that all deaths take place at the same time. Consistent overestimates of consumption will only occur, however, if deaths always take place shortly after a count, which seems highly unlikely.

Diet The proportion of salmonids in the stomach contents of birds shot on the Wye was significantly greater than in those on the Hodder. This was one of the main contributory factors for the higher consumption estimates on the Wye. If the proportion of salmonids in the diet were overestimated then losses could be significantly overestimated. The potential biases of stomach samples in estimating the diet of piscivorous birds are well documented (e.g. Carss et al. 1997 for cormorants), and include the under-representation of smaller fish. This particular bias is unlikely to have been important, given that the same sampling and analytical methodologies were employed on both river systems, and in both cases small fish (including salmonids) were well represented. The relatively small sample size on the Wye (10 stomachs), and the reliance on only one year’s data, may be of greater importance, as low sample sizes have been shown to greatly decrease the reliability of diet estimates (Marquiss & Carrs 1997).

09 11/02/2003 13:39 Page 132

132

Impact of birds on fisheries

Despite these potential errors/biases in generating consumption estimates, they are unlikely to account alone for the degree of overestimation apparent on the Wye study site. Indeed the modelling methodology, by seeking to incorporate variability into each of the parameters, should limit the effect of any of these factors. It seems likely therefore that standing crop biomass has been underestimated.

9.4.2

Biomass underestimated

Although similar methods of fish stock assessment were used on both rivers, there were differences in the sampling methods and the nature of the river that may have resulted in biomass being underestimated, particularly on the Wye. The most notable of these differences was the lack of any fully quantitative (multiple runs with stop netting) surveys on the Wye because of the large size of the river. This meant that to convert the semi-quantitative (single run, no-stop netting) data (necessarily a minimum estimate) to an estimate of biomass, a catch efficiency value was used which was based on fully quantitative surveys on the Hodder. It was, therefore, assumed that the survey catch efficiencies on the Wye were comparable to those on the Hodder. This is unlikely to be the case due principally to the differing habitat characteristics of the to rivers. Most importantly, the Hodder is a smaller river and hence surveyed reaches were narrower than those on the Wye (mean 15 and 25 m respectively). In larger river systems, a disproportionate amount of effort is required for generally small catches of fish (Hickley & Starkie 1985), hence catch efficiencies are likely to be lower on the Wye. Other habitat differences such as flow, conductivity and available cover are also likely to lead to different catch efficiencies. Smaller sample sizes on the Wye (four surveys at two sites) compared with the Hodder (15 surveys at three sites) also mean reduced reliability of these results. The absence of survey data from the first year, moreover, meant that fisheries surveys and bird diet data were not coincident.

9.4.3

Potential for compensation

The third possibility is that neither consumption nor biomass estimates are wrong, but instead fish populations were in some way able to compensate for the observed losses. Indeed, there is some evidence to suggest that juvenile salmonids may be able to compensate for increased mortality from predation. Where fish populations are at carrying capacity, density dependent mechanisms can limit the survival of some individuals, a mechanism known as self-thinning (Grant & Kramer 1990). Where self-thinning occurs, reduced densities brought about by predation can lead to increased survival and/or growth of remaining individuals. Long-term studies have indicated population dynamics consistent with self-thinning in UK trout (Elliott 1993) and salmon populations (Gardiner & Shakley 1991), and in both cases the critical period applies only to the first summer of life. As the salmonids preyed upon in the present study were predominantly 1 year old, there may have been scope for compensation within that cohort. Indeed repeated electric fishing

09 11/02/2003 13:39 Page 133

Impact of goosanders on salmonid populations

133

surveys on the river Hodder showed the biomass of salmonids to increase during the summer of 1996 (Feltham et al. 1999). The low biomass estimates in the present study, particularly on the Wye, indicate that fish populations are currently well below carry capacity and so any further losses are unlikely to compensated.

9.5 Conclusions As a result of differences in goosander productivity (and hence density), diet and particularly fish biomass, estimates of the proportions of salmonids consumed differed markedly between the rivers Wye and Hodder. Estimated losses of fish were of sufficient magnitude to suggest that predation may have had the potential to impact on fish stocks, but the true levels of impact are still unknown. The very high estimates of salmonid losses on the River Wye indicate that estimates were in some way erroneous, either because consumption was overestimated, fish biomass was underestimated, or both. Given the limited scope for overestimating consumption (discussion in Section 9.4.1) or for fish to compensate for such losses (discussion in Section 9.4.3), it appears that fish biomass was considerably underestimated. Although the methodology used was the only practicable approach to assessing juvenile salmonid fish stocks, the efficacy of electric fishing surveys on such large (30 m wide) river systems has to be questioned. If the impact of goosander predation on the fish populations of such rivers is to be seriously quantified, there is a pressing requirement to improve our ability to assess these populations. That is not to say that consumption estimates cannot be refined still further, particularly through investigating non-destructive methods for assessing diet, but without significant improvements in estimates of fish biomass such refinements will not improve estimates of impacts. This presents a problem because, despite the potentially high levels of goosander predation observed, there is still no direct evidence of an impact on either of the salmonid fisheries studied. Catch returns, or other measures of returning adults, may provide indicators of the relative performance of these rivers (stable or improving on the Hodder and declining on the Wye), but predation is only one of many factors which may influence catches. As Davies et al. (in press) has pointed out, when considering the impact of predation on fish stocks it is essential to account for all of the key factors influencing fish stocks. Salmon stocks in particular are influenced by a wide range of factors other than predation, affecting both the freshwater and marine stages of the life cycle (Russell et al 1996; Potter & Crozier 2000). Even in simple, almost experimental, studies (Pilcher & Feltham 1997; Feltham et al. 1999), identifying predation as the main cause of reduced fish stocks can be very difficult. Therefore, to understand fully the effects of goosander depredation on salmonid fish stocks there is a need to develop and test long-term models, which incorporate not only predation, but also those other factors which affect salmonid life cycles. Until such time that robust data are available to generate and test such models, impact estimates are likely to remain disappointingly inconclusive for fisheries such as the Wye and Hodder.

09 11/02/2003 13:39 Page 134

134

Impact of birds on fisheries

Acknowledgements Thanks go to all those people who assisted with bird and fisheries surveys and to landowners, fishery owners and angling clubs for their kind permission to access their land. The work was funded by MAFF, project number VC 0106.

References Carss D.N., Bevan R.M., Bonetti A., Cherubini G., Davies J.M., Doherty D., El Hili A., Feltham M.J., Grade N., Granadiero J.P., Gremillet D., Gromadzka J., Harari Y.N.R.A., Holden T., Keller T., Lariccia G., Mantovani R., McCarthy T.M., Mellin M., Menke T., Mirowska-Ibron I., Muller W., Musil P., Nazarides T., Suter W., Trautmansdorff J.F.G., Volponi S. & Wilson B.R. (1997) Techniques for assessing cormorant diet and food intake: towards a consensus view. Supplemento Richerche di Biologia della Selvaggina 26, 197–230. Cramp S. & Simmons K.E.L. (1977) Handbook of the Birds of Europe, the Middle East and North Africa Vol. I. Oxford: Oxford University Press, 483 pp. Cowx I.G. (1983) Review of the methods for estimating fish population size from survey removal data. Fisheries Management 14, 67–82. Davies J.M., Holden T., Feltham M.J, Wilson B.R., Britton J.R., Harvey J.P. & Cowx I.G. (in press) The use of Monte Carlo models to estimate the impact of cormorants at inland fisheries. Proceedings of the 5th International Cormorant Conference, Munich, 2000. Elliot J.M. (1993) The self-thinning rule applied to juvenile sea trout, Salmo trutta. Journal of Animal Ecology 62, 371–379. Erskine A.J. (1971) Growth and annual cycles of weights, plumages and reproductive organs of Goosanders in eastern Canada. Ibis 113, 42–58. Feltham M.J. (1990) The diet of red-breasted mergansers (Mergus serrator) during the smolt run in N.E. Scotland: the importance of salmon (Salmo salar) smolts and parr. J. Zool. London 222, 285–292. Feltham M.J. (1995) Consumption of Atlantic salmon Salmo salar smolts and parr by Goosanders Mergus merganser: Estimates from doubly labelled water measurements of captive birds released on two Scottish rivers. Journal of Fish Biology 46, 273–281. Feltham M.J, Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999) Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture Fisheries and Food, project no. VC 0106. London: MAAF, 212 pp.  Annexes. Gardiner R. & Shakley P. (1991) Stock and recruitment and density-dependent growth of Salmon, Salmo salar L., in a Scottish stream. Journal of Fish Biology 38, 691–696. Grant J.W.A. & Kramer D.L. (1990) Territory size as a predictor of the upper limit to population density of juvenile salmonids in streams. Canadian Journal of Fisheries and Aquatic Sciences 47, 1724–1737. Gregory R.D., Carter S.P. & Baillie S.R. (1997) Abundance, distribution and habitat use of breeding goosanders Mergus merganser and red-breasted meransers Mergus serrator on British rivers. Bird Study 44, 1–12. Hickley P. & Starkie A.S. (1985) Cost effective sampling of fish populations in large water bodies. Journal of Fish Biology 27 (Supplement A), 151–161. Howell R. (1987) An Assessment of Goosander (Mergus merganser) Predation on the Migratory Salmonid Stocks of the A.Tywi in 1986. Welsh Water Report C/1/87 (Llanelli). Marquiss M. & Carss D.N. (1994) Avian Piscivores: Basis for policy. National Rivers Authority R&D Report 461/8/N&Y. Bristol: National Rivers Authority, 104 pp.

09 11/02/2003 13:39 Page 135

Impact of goosanders on salmonid populations

135

Marquiss M. & Carss D.N. (1997) Methods of estimating the diet of sawbill ducks Mergus spp. and cormorants Phalacrocorax carbo. Supplemento Richerche di Biologia della Selvaggina 26, 247–258. Marquiss M., Carss D.N., Armstrong J.D. & Gardiner R. (1998) Fish-Eating Birds and Salmonids in Scotland. Report to The Scottish Office Agriculture Environment and Fisheries Department, 156 pp. Nillson S.G. & Nillson I.N. (1976) Numbers, food consumption and fish predation by birds in lake Mockeln, southern Sweden. Ornis. Scandinavia 7, 61–70. Pilcher M.W. & Feltham M.J. (1997) An Assessment of Cormorant Predation on Stillwater Coarse Fish Populations in the Lea and Colne Valleys of the Thames Catchment. Environment Agency (Thames North-east Area) R&D Technical Report W101, 64 pp. Potter E.C.E. & Crozier W.W. (2000) A Perspective on the Marine Survival of Atlantic Salmon. In D.H. Mills (ed.) The Ocean Life of Salmon, Environmental and Biological Factors Influencing Survival. Oxford: Blackwell Science, pp. 19–36. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Lowestoft: Directorate of Fisheries Research, 130 pp. Shearer W.M. (1984) The Natural Mortality at Sea for North Esk Salmon. ICES CM 1984/M;23. Shearer W.M., Cook R.M., Dunkley D.A., Maclean J.C. & Shelton R.G.J. (1987) A Model to Assess the Effect of Predation by Sawbill Ducks on the Salmon Stock of the River North Esk. Scottish Fish. Res. Report 37. HMSO Edinburgh. Suter W. (1995) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: two case studies from Swiss rivers. Journal of Applied Ecology 32, 29–46. Underhill M. (1993) Numbers and Distribution of Cormorants and Goosanders on the River Wye and its Main Tributaries in 1993. Wetlands Advisory Service Report to National Rivers Authority (Wales) and Countryside Council for Wales. Wernham C.V., Armitage M., Holloway S.J., Hughes B., Hughes R., Kershaw M., Madden J.R., Marchant J.H., Peah W.J. & Rehfisch M.M. (1999) Population, Distribution, Movements and Survival of Fish-eating Birds in Great Britain. Report to the Department of the Environment, Transport and the Regions, London. Wilson B.R., Feltham M.J., Davies J.M., Holden T., Britton J.R., Harvey J.R., Harvey J.P. & Cowx I.G. (in press) Increasing Confidence in Impact Estimates – The Monte Carlo Approach. Proceedings of the 5th International Cormorant Conference, Munich, 2000. Wood C.C. (1987a) Predation of juveniles Pacific Salmon by the common Merganser (Mergus merganser) on eastern Vancouver Island. I: Predation during the seaward migration. Canadian Journal of Fisheries and Aquatic Sciences 44, 941–949. Wood C.C. (1987b) Predation of juveniles Pacific Salmon by the common Merganser (Mergus merganser) on eastern Vancouver Island. II: Predation of stream resident juvenile Salmon by merganser broods. Canadian Journal of Fisheries and Aquatic Sciences 44, 950–959. Zippin C. (1956) An evaluation of the removal method of estimating animal populations. Biometrics 12, 163–189.

This page intentionally left blank

10 11/02/2003 13:41 Page 137

Section II Bird, fish, habitat interactions

This page intentionally left blank

10 11/02/2003 13:41 Page 139

Chapter 10

Interaction between fish and colonial wading birds within reed beds of Lake Neusiedl, Austria E. NEMETH Konrad Lorenz Institute for Comparative Ethology, Vienna, Austria

G. WOLFRAM Donabaum & Wolfram OEG, Vienna, Austria

P. GRUBBAUER, M. RÖSSLER and A. SCHUSTER Konrad Lorenz Institute for Comparative Ethology, Vienna, Austria

E. MIKSCHI Museum of Natural History, Vienna, Austria

A. HERZIG Biologische Station Illmitz, Illmitz, Austria

Abstract Data from two projects carried out within the extensive (180 km2) reed belt of Lake Neusiedl, Austria – one about the fish community and the other about the colonial breeding piscivorous birds, great white egret, Casmerodius albus L., purple heron, Ardea purpurea L., grey heron, Ardea cinerea L., and spoonbill, Platalea leucorodia L. – were used to assess the impact of fish-eating birds on the fish populations. Piscivorous birds ate about 12% of total fish standing stock within the reed belt. Their food intake was estimated to account for 21% of fish production within a size range of 3–25 cm total length (potential prey size). Competition between fish-eating birds and commercial fisheries was considered to be negligible. Keywords: fish biomass, heron, piscivorous bird, predation rate, predator-prey.

10.1 Introduction The extensive reed beds (180 km2) of Lake Neusiedl on the eastern border of Austria is an internationally important site for bird species (Grimmet & Jones 1989). Of the fish-eating species, the largest population of great white egret, Casmerodius albus L., in western and central Europe, and internationally important populations of purple heron, Ardea purpurea L., and spoonbill, Platalea leucorodia L., are found there. Nearly all the heronries are located in the reed beds which form a belt up to 6.5 km wide around the open lake (Fig. 10.1). While the importance of reed beds as secure nesting habitat is evident, less is known about the importance of these areas as foraging habitat (Grüll & Ranner 1998).

10 11/02/2003 13:41 Page 140

140

Bird, fish, habitat interactions

Figure 10.1 Study area Lake Neusiedl on the border of Austria to Hungary. Roman numerals signify electric fishing areas (see text)

Information about the function of reed areas for the fish community of Lake Neusiedl is scarce. Most have concentrated on the fish fauna of the open lake (e.g. Hacker 1979), while the fisheries of the large reed belt were not examined until the mid-1990s (Wolfram-Wais et al. 1999; Wolfram et al. 2002). Data from two projects carried out within the extensive (180 km2) reed belt of Lake Neusiedl, Austria – one about the fish community and the other about the colonialbreeding piscivorous birds – were used to assess the impact of fish-eating birds on the fish population. The main objectives were to determine the interactions between fish and birds using data on fish availability within the reed beds and the utilisation of these areas by piscivorous birds as foraging habitat, and to discuss the possible implications for commercial fisheries.

10 11/02/2003 13:41 Page 141

Fish and bird interactions in the reed beds of Lake Neusiedl

141

10.2 Material and methods 10.2.1

Study site

Lake Neusiedl is a shallow alkaline lake (7.5–14.6 mval litre1, conductivity C25  1300–3200 S) with an area of 320 km2 and an average water depth of 1.1 m. More than half of the lake (180 km2) is covered by reed beds which form a belt around the lake shore, with one big reed island (6.5 km2) in the south (Fig. 10.1). The reed area consists of a mosaic of pure stands of Phragmites australis (Cavanilles) Trinius ex Steudel, channels and open water areas. Water depth in the reed beds depends on the water level of the lake and ranges from 0 to 1.7 m. While the water of the open lake is characterised by a high inorganic turbidity, water within the reed beds is clear and has a reddish-brown colour owing to a high amount of humic substances. In summer, oxygen deficiencies (1 mg litre1) frequently occur within the dense Phragmites stands. In spring 1994 the southern part of Lake Neusiedl (about 80 km2) was declared a national park by the Austrian and Hungarian governments.

10.2.2

Fish data

The majority of fish was sampled by electric fishing. Gill net fishing was carried out occasionally but produced unsatisfactory results due to the density of the Phragmites stands and the shallowness of the water (often 50 cm). Electric fishing (direct current, 8 kW, 200–300 V, one hand-held electrode) was conducted from August 1994 to November 1997 on 27 occasions lasting between 1 and 3 days (in total 68 days). Fishing concentrated on five areas of the reed belt: (i) Illmitz – Biological Station; (ii) Illmitz – south of the public baths; (iii) the southern part of the lake (National Park); (iv) near Mörbisch; and (v) near Oggau (Fig. 10.1). Sampling was undertaken along transects from the reed fringe landward. Thus the sampling site consisted of: the littoral zone of the open lake – Phragmites fringe; channels (old structures used by local fishermen and people harvesting the reed); pools within the Phragmites; and dense vegetation stands. At all sampling sites within a transect, a single fishing along a defined stretch of 30 m (standard catch) was conducted. The total number of standard catches was 599. Fishing was done by boat or at selected shallow water (50 cm deep) sites by foot. Catchability varied according to turbidity, water depth, structure and density of the reed stands but was estimated to be about 50% in terms of biomass (based on observations of fish occasionally escaping during the fishing). CPUE is always given as the original values; values for standing stock (kg ha1), however, were calculated with respect to catchability. The fish caught were identified, measured (mm TL) and returned to the water. Biomass was calculated by applying species-specific length–weight regressions defined for Lake Neusiedl during earlier studies (Herzig et al. 1994). On several sampling occasions, additional weights were taken for false harlequin, Pseudorasbora parva Temminck & Schlegel, and pumpkinseed, Lepomis gibbosus (L.), as well as for those fish sampled for subsequent gut analyses (see Wolfram-Wais et al. 1999). To calculate

10 11/02/2003 13:41 Page 142

142

Bird, fish, habitat interactions

eel biomass, four length groups (15 cm, 15–30 cm, 30–55 cm and 55 cm) were distinguished with respect to the length–frequency distribution derived during former studies from Lake Neusiedl (Herzig et al. 1994). The total biomass for all individuals in each length group was calculated from length–weight regression for eel from the lake. As all fish data were based on linear transects, they were converted to density (kg ha1) for three sample areas (Figs 10.1) in the east and south of Lake Neusiedl using digitalised aerial near-infrared photographs (1 pixel is approximately 2 m2). As the results of the calculations were similar for all three sample areas, only data for the sample area near Illmitz (34.9 ha) are presented in detail. Standing stock was estimated in three steps: (1) The reed area selected was divided into several zones with different reed cover: (a) very dense reed beds with almost no open water areas; (b) very shallow (50 cm deep), landward areas which dry up in summer; (c) open water pools 0.5 ha; and (d) remaining area characterised by a mosaic of reed stands and pools 0.5 ha. (2) Mean fish density was determined for each zone. The few, small open water areas of zone (a) were not sampled. It can be assumed, however, that due to the very dense (and often dry) reed stands they are not penetrated from larger pools even by smaller fish. Moreover, low oxygen concentrations (1 mg litre1) for long periods makes these small water bodies hostile for fish colonisation (Wolfram et al. 2002). Sampling in zone (b) was carried out occasionally at the beginning of the project, but revealed no, or only a few, scattered juvenile fish that inhabited these very shallow areas. Hence, in both zones fish standing stock was assumed to be negligible and considered zero for further calculations. Bigger water areas (zone (c)) were divided further into open water and littoral zone along the reed fringe. The CPUE along the reed fringe (kg per 30 m standard catch) was converted to kg ha1 by assuming a sampling width by electric fishing of 2 m, giving a sampling area of 60 m2. The fish density estimates calculated in this way were considered reedbed fringe fish densities (RFD). Occasional gill net fishing in the open water areas within the reed belt revealed that most fish concentrate in the fringe area and fish density in the open water areas was 10–20% of the RFD. Zone (d) within the reed belt is a dense mosaic of pools and reed stands (the majority ranging in size from a few to some dozens of square metres). CPUE based on 30 m standard catches from sites within this zone were again converted to RFD values. Taking into account the larger pools of zone (d), the RFD does not correspond, but overestimates the true fish density within this zone. Based on calculations of the proportion of pools and reed stands within zone (d) (using digitalised aerial nearinfrared photographs), about 50% of the RFD was estimated to resemble the true fish density within zone (d). (3) By calculating the proportions of the different zones in the sample area and multiplying these figures by mean fish densities for each zone, the total fish standing stock was determined. The calculation of fish standing stock within the reed belt must be considered approximate and does not consider seasonal variations or possible differences of fish densities between different years (1994–1997).

10 11/02/2003 13:41 Page 143

Fish and bird interactions in the reed beds of Lake Neusiedl

10.2.3

143

Habitat utilisation by birds and estimation of food intake

All bird data were collected during the 1998–2000 breeding seasons. Vantage point observations of birds departing from the colonies and aerial counts of foraging birds were used to estimate reed habitat usage. The undulating topography and artificial elevated points allowed observation of birds which departed from their colonies en route to their foraging grounds. In good weather conditions birds could be followed by telescope up to a distance of 12 km. Six colonies of great white egrets (87% of the breeding population), two purple heron colonies (49%), one grey heron colony (70%) and one spoonbill colony (100%) were observed. Single colonies were surveyed between late April 2000 and the beginning of July 2000. Observations concentrated on periods of main bird activity and lasted from 30 min to about 4 h per day. In 45 surveys, 1160 departing wading birds were counted (720 great white egrets, 200 purple herons, 86 grey herons and 154 spoonbills). The landing position of the birds was recorded in 72% of all departures. Birds that flew out of view were allocated to feeding grounds within or outside the reed belt according to habitat utilisation data derived from aerial counts. A combination of vantage point observations (which also comprised the Hungarian feeding habitat) and aerial counts for the great white egrets (which delivered more accurate data for the habitat utilisation on the Austrian side) were used to estimate the proportion of foraging great white egrets within the reed habitat. Aerial counts were made using a Piper PA-18 aircraft over the Austrian part of the study area, which covers about 640 km2 (Fig. 10.1). During the breeding seasons of 1998–2000, 51 flights were carried out. Each foraging bird (a total of 12 749 observations) was recorded on a map with a precision of about 100–200 m. On the ground, food intake of foraging great white egrets at selected sites was recorded. One observation lasted up to 20 minutes. Prey sizes were estimated classes in multiples of one-quarter of the average bill size (see Bayer 1985). In 1999, video surveillance of 12 nests allowed partial estimation of prey types. Species-specific food consumption rates were derived from the literature (Marion 2000). Daily food requirements were considered independently for young and adult birds. For grey herons the estimates derived by Feunteun & Marion (1994) were used. In the other cases the daily requirements of the adults were estimated according to the body mass/food consumption relationship derived by Kushlan (1976). For the food intake of young great white egrets, data from Mock (personal communication) was used. All spatial data on habitat utilisation were processed with Arcview 3.2. Statistical calculations were conducted with SPSS and Statistica.

10.3 Results 10.3.1

Species composition and dominance structure of the fish community

From August 1994 to October 1997 more than 32 000 fish from 18 species, with a total weight of about 886 kg, were caught by electric fishing within the reed belt and along

10 11/02/2003 13:41 Page 144

144

Bird, fish, habitat interactions

Table 10.1 Species list, occurrence and relative proportion of different fish species during the whole study period at all sampling stations Relative proportion (% biomass) Species

Occurrence (%)

Relative proportion (% number)

All length classes

3–25 cm excluding eel

Esox lucius L. Anguilla anguilla (L.) Abramis brama (L.) Blicca bjoerkna (L.) Alburnus alburnus (L.) Carassius auratus gibelio (Bloch) Carassius carassius (L.) Cyprinus carpio L. Pelecus cultratus (L.) Pseudorasbora parva T. & Schl. Rutilus rutilus (L.) Scardinius erythrophthalmus (L.) Tinca tinca (L.) Silurus glanis L. Gymnocephalus cernuus (L.) Perca fluviatilis L. Stizostedion lucioperca L. Lepomis gibbosus (L.)

23.3 68.6 16.0 57.9 27.4 47.0 7.6 17.2 0.2 36.3 54.0 64.0 9.9 1.6 17.8 29.0 7.5 45.5

0.7 7.9 1.0 25.4 5.0 2.8 0.3 0.7 0.1 8.4 20.3 14.1 0.3 0.1 0.9 2.7 0.3 9.0

9.5 35.1 1.3 6.0 0.1 23.4 0.2 14.2 0.1 0.4 4.2 1.9 0.3 0.1 0.1 0.9 1.0 1.5

2.1 – 1.9 32.1 0.5 16.3 0.8 1.0 0.1 2.0 18.1 9.9 1.3 0.1 0.7 5.3 0.3 7.5

the outer edge of the Phragmites stands towards the open water zone (Table 10.1). Sixty-six per cent of all fish caught were young-of-the-year (YOY). The fish species most commonly caught were eel, Anguilla anguilla (L.), rudd Scardinius erythrophthalamus (L.), silver bream, Blicca bjoerkna (L.) and roach, Rutilus rutilus (L.). They were present in more than 50% of all samples (Table 10.1). Crucian carp, Carassius carassius (L.), pumpkinseed, false harlequin, perch, Perca fluviatilis L., bleak, Alburnus alburnus (L.) and pike, Esox lucius L., were captured less often, but still with occurrence values above 20%. In terms of abundance (including 0), silver bream, roach and rudd dominated (relative proportion 10%). Pumpkinseed, eel and false harlequin accounted for 5 to 10% of total individuals caught. Eel, crucian carp, common carp and pike prevailed in terms of biomass, whereas smaller but abundant fishes such as silver bream and roach accounted only 6.0 and 4.2% of total biomass, respectively. Considering only the length classes that are taken by piscivorous waterbirds (3–25 cm, excluding eel), silver bream, roach and crucian carp dominated in terms of biomass; all three species accounted for 67% of biomass in the 3–25 cm length class (Table 10.1).

10.3.2

Total fish standing stock

On average (median) about 30 individuals were caught within one 30 m standard catch. The maximum number was 760 individuals. Median CPUE for fish of age classes 0 was 10 individuals per 30 m, the maximum was 221 individuals per 30 m.

10 11/02/2003 13:41 Page 145

Fish and bird interactions in the reed beds of Lake Neusiedl

145

Table 10.2 Standing stock, production and consumption of fish inhabiting the reed belt of Lake Neusiedl Standing stock (t) Standing stock (kg ha1) Standing stock size class 3–25 cm (tons) Estimated P/B ratio size class 3–25 cm Fish production size class 3–25 cm (t yr1) Fate of production size class 3–25 cm): consumption by pike and zander (t yr1) consumption by birds (t yr1) other mortality including fisheries (t yr1)

592 32 326 1.0 326 251 (77%) 68 (21 %) 7 (2%)

Around Illmitz (Fig. 10.1), CPUE at 9 sites in the larger open water bodies (zone (c)) varied between 0.50 (0.32–0.78) and 1.42 (1.14–1.77) kg per 30 m (geometric mean  log transformed CL; n  6–23). These figures correspond to RFD (reed fringe fish density) values of 167 (107–261) and 473 (380–589) kg ha1, respectively. Within moderately dense reed stands (zone (d)), no or negligible fish were caught in about 30% of all samples, mainly as a result of a highly aggregated fish distribution. Mean CPUE of all catches within zone (d) was 0.14 (0.08–0.25) kg per 30 m or 47 (26–85) kg ha1 (RFD), mean CPUE excluding catches with no or negligible fish density was 0.53 (0.39–0.72) kg per 30 m or 177 (130–239) kg ha1 (RFD). Total standing stock within the sample area was 1.68 t, corresponding to 48 kg ha1. As the proportion of open water in the sample area near Illmitz was greater than the average of the whole reed belt, fish densities near Illmitz were probably greater than in the remaining study area. Thus the estimate was related to reed structure and was reduced by a factor 0.67, resulting in mean standing stock of 32 kg ha1. The factor was derived from the amount of open water in the whole reed beds estimated by remote sensing combining satellite data and aerial near-infrared photography (Nemeth et al. in press; unpublished data). The fish density along the edge of the reed belt towards the open lake is not included in the estimate of 32 kg ha1 and was calculated separately. Mean CPUE along the reed fringe was 1.40 (0.94–2.08) kg per 30 m, which corresponds to a total fish biomass of about 16 t for the whole zone, which has a shore length of 340 km. Summarising, the total standing stock of fish within the reed belt including the edge towards the open lake was about 592 t (Table 10.2).

10.3.3

Standing stock and production of 3–25 cm size class

The majority of fish consumed by the great white egrets had a total length of 3–25 cm. Prey size of the other wading birds also probably lies within this range (excluding eel, see Feunteun & Marion 1994). This size class contributed about 55% of the total fish biomass. The fish stock potentially available as food for piscivorous birds was about thus 326 t within the reed belt (exclusive of the edge towards the open water zone). Assuming a production/biomass (P/B) ratio of 1 for the size class 3–25 cm, total fish production accounts for 326 t yr1 (Table 10.2).

10 11/02/2003 13:41 Page 146

146

10.3.4

Bird, fish, habitat interactions

Consumption by piscivorous fish

Among the fish species that inhabit the reed belt only, pike is a major predator of other fish. Standing stock of pike was estimated from its contribution to the relative biomass within the reed belt (Table 10.1) at 56.2 t (3 kg ha1). Total consumption of pike is about 228 t yr1, based on estimated daily consumption rates ranging from 0.05% body weight (BW) during winter to 3.5% BW at high temperatures in summer (Diana 1991). Pikeperch, Stizostedion lucioperca (L.), inhabits mainly the open water zone and the edge of the reed bed towards the open lake (unpublished data) and is of minor importance as a predator of reed-dwelling fish. Consumption by pikeperch was estimated to about 10% of the consumption by pike (Table 10.2). The fish eaten by pike and pikeperch lie within the same size range as the prey of egrets and herons (Herzig et al. 1994).

10.3.5

Wading birds within reed beds and the amount of eaten fish

Reed beds were the most important feeding habitat for all wading bird species (Table 10.3). Fish was the food in 96% of feeding events of great white egrets bring food to their young (n  165). Ninety per cent of the food items of foraging great white egrets within reed beds (n  65) was again fish. Other prey items were frogs (4%) and insects. Most of the unidentified prey items were smaller than 3 cm and therefore contributed

Table 10.3 Foraging wading birds and amount of eaten fish from April to September 2000. Presence of wading bird species at the lake was estimated in months: grey heron 6 months, purple heron and great white egret 5 months and spoonbill 4 months. Calculations of daily food requirements of adults and young according to Feunteun & Marion (1994), Kushlan (1976) and Mock et al. (1987). An average breeding success of 1.5 young was assumed. Young birds of all three heron species were expected to stay one month after leaving the nest in the foraging area Species Breeding pairs % foraging within reed beds % foraging at reed/lake border Total consumption (tons) Total fish intake within reed beds (tons) Total fish intake at reed/lake border (tons) Fish intake kg ha1 within reed area % of total standing stock reed fish % of standing stock reed fish 3–25 cm

Great white egret

Purple heron

Grey heron

Spoonbill

Total

753 80 2 50.8 40.6

307 82 8 18.4 15.1

85 56 16 8.3 4.7

77 61 0 5.7 3.5

1222 77 23 83.2 63.8

1.0

1.5

1.3

0

3.9

2.3

0.9

0.3

0.2

3.8

7.0

2.7

1.0

12.7

5.1

1.8

1 1.0

11.7 21.3

10 11/02/2003 13:41 Page 147

Fish and bird interactions in the reed beds of Lake Neusiedl

147

little to the food intake. There remains, however, uncertainty as to how much non-fish prey is eaten. Based on observations of foraging great white egrets, approximately 10% of prey intake in reed beds was not fish (Table 10.3). Average fish total length was 5.1 cm (2.1 standard error, n  185), with a range from 3 to 25 cm. Eel was excluded as a major contribution to the diet of the great white egret since it was observed only once as prey at the nests and never observed in foraging birds. The same probably applies for spoonbill (which feeds in areas with low eel density, unpublished data). No direct information was available for the diet of grey heron and purple heron, although eel may contribute to the diet of these species. However, all herons forage in the same area and it seems reasonable that prey of grey heron and purple heron resembles that of great white egret.

10.4 Discussion The littoral zone of Lake Neusiedl forms a highly structured and complex mosaic of pools and Phragmites stands, which are responsible for the patchy distribution of fish. Large areas are characterised by very shallow open water bodies, which may dry up during the warm summer months and thus are scarcely inhabited by fish. More than half of the reed belt is formed by very dense reed stands that are not colonised by fish. Apart from these spatial constraints, low summer oxygen levels in the densely covered parts of the inner reed belt cause problems to the fish fauna (Wolfram et al. 2002). Highest fish biomass was found along the fringe zones between the Phragmites stands and open water, both in the bigger pools, within the reed belt, and along the edge of the reed towards the open water zone of Lake Neusiedl. There are very few shallow lakes within Europe with such an extended reed swamp as found at Lake Neusiedl. Lake Grand Lieu in western France is probably the most similar. In the late 1980s, Feunteun & Marion (1994) found a fish standing stock of 270 kg ha1 in open water areas within the marsh, but 30 kg ha1 in the marsh as a whole, findings which lie within the range for Lake Neusiedl. The production of fish with a total length between 3 and 25 cm within the reed belt of Lake Neusiedl was 326 t yr1. This value was derived from an estimated turnover (P/B) ratio of 1, but the ratio for fish populations is usually below 1. However, this study concentrated on the part of the population below 25 cm, which has a higher turnover ratio (Downing & Plante 1993). Also a significant proportion of total fish production is formed by the young of the year (Mathews 1971; Mann 1991). Considering only age groups 1, Hacker (1974) and Meisriemler (1974) found turnover ratios of 0.82–0.9 for silver bream and 1.01 for ruffe respectively, in the open water zone of Lake Neusiedl during the 1970s. They considered that the P/B ratio might have exceeded these values if 0 fish had been included in their calculations. Woollhead (1994) estimated P/B ratios for Lake Esrom distinguishing between non-piscivores (P/B  1.25) and piscivores (P/B  0.625). Turnover ratio of 0.6 to 1.11 were also found in several lowland cyprinid fish populations in Britain (Chapman 1978). Considering these figures, a P/B ratio of 1.0 for reed-dwelling fish between 3 and 25 cm for Lake Neusiedl seems appropriate.

10 11/02/2003 13:41 Page 148

148

Bird, fish, habitat interactions

Few studies have assessed quantitatively the impact of heron predation on fish in natural habitats (Marion 2000). Predation rates range from 6% in permanent flooded marshes in France (Feunteun & Marion 1994) to 72% of fish biomass in seasonally flooded ponds in the Everglades (Kushlan 1976). In Lake Neusiedl, piscivorous birds consume 11.7% of the total fish standing stock. This confirms the similarity between Lake Neusiedl and Lake Grand Lieu, not only in terms of fish biomass (see above), but also as regards the trophic impact of fish-eating birds. In both systems they play a minor role compared to piscivorous fish. However, in Lake Neusiedl, herons and spoonbills consume one-quarter of the total production of fish 25 cm (Table 10.2). Thus they potentially have a considerable impact on the fish community. Moreover, in restricted reed areas a marked depletion of fish stock by birds seems possible. The reed belt often acts as a trap. Falling water levels during the summer force the fish to withdraw into deeper pools without connection to the open lake. In the morning, fish often have to perform water surface respiration because of oxygen deficiencies as a result of decomposition of organic matter and nocturnal respiration of macrophytes. Such situations are exploited by large aggregations of herons, mainly great white egret. As a consequence, great white egrets were able to increase their fish intake in terms of biomass from 8 g in 10 min for birds fishing alone (n  75) to 121 g in 10 min for birds feeding in groups on fish exhibiting aquatic surface respiration (n  32). The estimates of fish consumption suggest that a small amount of the production of reed bed fish (size class 3–25 cm) is not consumed by herons or pike. These fish probably die due to parasites, diseases or physiological stress as a result of oxygen deficiencies within the reed belt. Fish mortality due to midsummer oxygen depletion would be higher without bird predation, thus similar to seasonal ponds in the Everglades (Kushlan 1976), heron predation at Lake Neusiedl might substitute for mortality from other sources. The main commercial fish species at Lake Neusiedl are eel (more than 90%), carp, pike and zander. Total catch of eel exceeded 100 t in the1980s (Herzig et al. 1994); there are no reliable data on fish yields in recent years. Direct competition between piscivorous birds and fisheries appears to be negligible because birds are confined to smaller size classes of prey. However, the indirect impact of predation on smaller sized fishes which could potentially grow to larger fish should be considered. In terms of biomass, carp, pike and zander 25 cm account for 3.5% of possible prey for fisheating birds. Eels might be eaten to some extent by grey heron and purple heron, but their food intake of these species within the reed beds will be well below the commercial yield. The fish populations of the reed beds are the most important food resource for herons at Lake Neusiedl. This dependence may explain the increase in population size of the great white egret from about 120 breeding pairs at the beginning of the 1980s to more than 700 at the end of the 1990s (Grüll & Ranner 1998). At the same time eutrophication and, hence, an improvement of food resources resulted in a ten-fold increase in biomass of cyprinid fish in the open lake (Herzig 1994; Mikschi et al. 1996), which probably corresponds to an increased abundance of fish in the reed zones.

10 11/02/2003 13:41 Page 149

Fish and bird interactions in the reed beds of Lake Neusiedl

149

Acknowledgements Anita Wolfram-Wais and Arno Hain assisted in electro-fishing. Robert Klein was pilot in the aerial surveys and helped also in fishing. Christoph Plutzar helped in the GIS analyses. Günther Brandtner was responsible for video technique. Data are from two projects, both supported by the National Park Neusiedler See.

References Bayer R.D. (1985) Bill lengths of herons and egrets as an estimator of prey size. Colonial Birds 8, 104–109. Chapman D.W. (1978) Production. In T.B. Bagenal (ed.) Methods for Assessment of Fish Production in Fresh Waters. IBP Handbook No. 3. Oxford: Blackwell Scientific Publications, pp. 202–217. Diana J.S. (1991) Energetics. In J.F. Craig (ed.) Pike. Biology and Exploitation. London: Chapman & Hall, pp. 103–124. Downing J.A. & Plante C. (1993) Production of fish populations in lakes. Canadian Journal of Fisheries and Aquatic Science 50, 110–120. Feunteun E. & Marion L. (1994) Assessment of grey heron predation on fish communities: The case of the largest European colony. Hydrobiologia 279/280, 327–344. Grimmet T.R.F.A. & Jones J.A. (1989) Important Bird Areas in Europe. ICBP Technical Publication No. 9. Cambridge: International Council for Bird Preservation, 888 pp. Grüll A. & Ranner A. (1998) Populations of the Great Egret and purple heron in relation to Ecological Factors in the Reed Belt of the Neusiedler See. Colonial Waterbirds 21, 328–334. Hacker R. (1974) Produktionsbiologische und nahrungsökologische Untersuchungen an der Güster (Blicca björkna L.) im Neusiedler See. Doctoral thesis, University of Vienna, 93 pp. (in German). Hacker R. (1979) Fishes and fisheries in Neusiedlersee. In H. Löffler (ed.) Neusiedlersee: the Limnology of a Shallow Lake in Central Europe. The Hague: Dr Junk; by publications, Monographiae Biologicae 37, 423–438. Herzig A. (1994) Predator-prey relationships within the pelagic community of Neusiedler See. Hydrobiologia 275/276, 81–96. Herzig A. & Wolfram G. (in press) Fish distribution and limiting factors in the littoral of a shallow lake. In R. Field, R.J. Warren, H. Okarma & P.R. Sievert (eds) Wildlife, Land, and People: Priorities for the 21st Century. Proceedings of the Second International Wildlife Management Congress. Bethesda: The Wildlife Society. Herzig A., Mikschi E., Auer B., Hain A., Wais A. & Wolfram G. (1994) Fischbiologische Untersuchung des Neusiedler See. BFB-Bericht 82, 1–125. (in German). Kushlan J.A. (1976) Feeding ecology of wading birds. In A. Sprunt IV, J.C. Odgen & S. Winckler (eds) Wading Birds. Research Report of the Audubon Society No. 7, pp. 249–297. Mann R.H.K. (1991) Growth and production. In I.J. Winfield & J.S. Nelson (eds) Cyprinid Fishes. Systematics, Biology and Exploitation. London: Chapman & Hall, pp. 457–482. Marion L. (2000) Aquaculture. In J.A. Kushlan & H. Hafner (eds) Heron Conservation. London: Academic Press, pp. 269–292. Mathews C.P. (1971) Contribution of young fish to total production of fish in the River Thames near Reading. Journal of Fish Biology 3, 157–180. Meisriemler P. (1974) Produktionsbiologische und nahrungsökologische Untersuchungen am Kaulbarsch (Acerina cernua (L.)) im Neusiedlersee. Doctoral thesis, University of Vienna, 110 pp. (in German).

10 11/02/2003 13:41 Page 150

150

Bird, fish, habitat interactions

Mikschi E., Wolfram G. & Wais A. (1996) Long-term changes in the fish community of Neusiedler See (Burgenland, Austria). In A. Kirchhofer & D. Hefti (eds) Conservation of Endangered Freshwater Fish in Europe. Basel: Birkhauser Verlag, pp. 111–120. Mock D.W., Lamey T.C. & Ploger B.J. (1987) Proximate and ultimate roles of food amounts in regulating egret sibling aggression. Ecology 68, 1760–1772. Nemeth E., Dvorak M., Busse K. & Rössler M. (in press) Estimating distribution and density of reedbirds by aerial infrared photography. In R. Field, R.J. Warren, H. Okarma & P.R. Sievert (eds) Wildlife, Land, and People: Priorities for the 21st Century. Proceedings of the Second International Wildlife Management Congress. Bethesda: The Wildlife Society. Wolfram G., Mikschi E. & Wolfram-Wais A. (2002) Fischökologische Untersuchung des Schilfgürtels des Neusiedler Sees. BFB-Bericht. (in press) (in German). Wolfram-Wais A., Wolfram G., Auer B., Mikschi E. & Hain A. (1999) Feeding habits of two introduced fish species (Lepomis gibbosus, Pseudorasbora parva) in Neusiedler See (Austria), with special reference to chironomid larvae (Diptera: Chironomidae). Hydrobiologia 408/409, 123–129. Woollhead J. (1994) Birds in the trophic web of Lake Esrom, Denmark. Hydrobiologia 279/280, 29–38.

11 11/02/2003 13:43 Page 151

Chapter 11

Management of cyprinid fish populations as an important prey group for the endangered piscivorous bird, bittern, Botaurus stellaris R.A.A. NOBLE*, J.P. HARVEY and I.G. COWX Hull International Fisheries Institute, University of Hull, Hull, UK

Abstract The bittern, Botaurus stellaris (L.), is a red data list bird species with currently around only 22 booming males restricted to a few major wet reed beds in the UK. The bittern, although catholic in diet, is mainly piscivorous. Eels, Anguilla anguilla (L.), and rudd, Scardinius erythrophthalmus (L.), are its main prey species. This reliance on fish in its diet makes understanding the ecology of fish in wetlands imperative if habitat and fishery management is to be beneficial to the conservation of bitterns in wetlands. Studies were undertaken to assess the dynamics of fish populations in wetlands, the factors affecting their availability to bitterns and the impact of reedbed management on fish populations. Case studies of fish populations at two reedbed sites in the UK are presented. Although eels are the key prey species, it is considered that cyprinid fish stocks may become more important in the conservation of the bittern due to the apparent decline in eel recruitment to wetland sites in Europe. A comparison is made between the fish community characteristics at a site consisting of one hydrological unit and a site split into multiple units. The impacts of current reedbed design and management practices, especially those impinging on habitat connectivity and the dynamics of cyprinid fish populations, are discussed. The potential for fishery and habitat management to enhance the availability of fish to breeding bitterns is evaluated. Keywords: Botaurus stellaris, conservation, connectivity, eel, reed edge, rudd.

11.1 Introduction The bittern, Botaurus stellaris (L.), a member of the heron family (Ardeinae, subfamily Botaurinae), is a red data list bird species with currently around 22 booming males at 14 sites in the UK (G. Gilbert, personal communication). It is also a category 3 species of European Conservation Concern (SPEC), and a species with unfavourable status in European countries (Newbery et al. 1997). The status of the species in the UK has varied over the past 250 years, with the current population in a state of decline linked to loss in the quality and quantity of its preferred wetland reedbed habitat (Smith & Tyler 1993). The key threats to the breeding status in the UK are considered to be *Correspondence: R.A.A. Noble, Hull International Fisheries Institute, University of Hull, Hull HU6 7RX, UK (email: [email protected]).

11 11/02/2003 13:43 Page 152

152

Bird, fish, habitat interactions

degradation of suitable reedbed habitat and food resources through pollution, hydroseral succession and inappropriate management (Newbery et al. 1997). The bittern is a solitary, daytime feeder, feeding mainly in flooded marginal stands of emergent vegetation (Cramp & Simmons 1977). Radio-tracking studies (G. Tyler, personal communication) suggest that the majority of feeding activity occurs within the first ten metres of wet reed bed from the reed/water interface. Bitterns forage by sight, clambering through flooded reed bed (water depth 25 cm) by clutching clumps of reed with their feet or wading through shallow water. The bittern has a catholic diet and the dietary constituents vary with locality and season (Gentz 1965). Recorded components of the bittern’s diet include fish, amphibians, insects, worms, leeches, molluscs, crustaceans, spiders, lizards, small birds and small mammals (Cramp & Simmons 1977). Studies in the UK indicate that fish is the most important component of the diet, especially for the chicks. Studies of regurgitates indicate that chick diet consists of eels, Anguilla anguilla (L.) (47%), and rudd, Scardinius erythrophthalmus (L.) (34%) (G. Gilbert, personal communication). The diet of the adult bittern was found to include small mammals (50%) and fish (40%). The fish fauna of UK wetland reed beds is generally considered to be depauperate, both in terms of species diversity and abundance, when compared to other water bodies, e.g. shallow lakes (Perrow et al. 1996). Fish communities in reed beds have been monitored by the Royal Society for the Protection of Birds (RSPB) since 1990. Fish communities in reed beds show site-specific variation, although generally comprising a combination of eels, rudd, pike, Esox lucius L., roach, Rutilus rutilus (L.), perch, Perca fluviatilus L., tench, Tinca tinca (L.) and sticklebacks, Gasterosteus aculeatus L. and Pungitius pungitius (L.). Reed beds that are part of larger aquatic systems including connection to rivers generally have a higher diversity and abundance of fish (G. Gilbert, personal communication). There is weak evidence to suggest that site occupancy by bitterns is related to fish biomass and abundance (G. Gilbert, personal communication). Recent studies also indicated that geographically discrete reserves occupied by bitterns have a lower median eel weight of 26 g per individual compared with 65 g in unoccupied reed beds, but this is also related to the level of recruitment of eels to the reserve (Tyler 1994). The most significant reedbed features linked to site occupancy and breeding status of bitterns are the total area of reed and the successional stage of the reed bed (Tyler et al. 1998). The total length of reed/water interface and the autumn water depth in wetlands was also found to be positively correlated to bittern occupancy (Tyler 1994). Reedbed management practices for conservation in the UK are primarily aimed at reducing hydroseral succession to maintain mono-dominant Phragmites reed beds. This is achieved through a combination of water level management and reed cutting. Reedcutting and water level management act to control the growth and spread of Phragmites maintaining a particular successional stage and density of reed, but also preventing reed encroaching into open waters (Hawke & José 1996). Management practices for rehabilitation purposes include lowering bed levels in dry reed beds using mechanical diggers to reinstate open waters and old dyke channels, and to re-establish wet reed beds (José 2000). Appropriate management of fish populations and aquatic habitats in UK reed beds has also been identified as a key component of conservation management for bitterns.

11 11/02/2003 13:43 Page 153

Managing fish population for an endangered piscivorous bird

153

Understanding the dynamics of fish populations in reed beds, elucidating the factors affecting the availability of fish to bitterns and identifying how current management activities affect fish populations were identified as key research activities. Research was thus initiated to provide a sound scientific basis for conservation management, rehabilitation and enhancement of fish populations for the benefit of bitterns. The key to potential enhancement of fish availability to bitterns appears to be the association of fish with the reed edge, the principle foraging habitat of the bittern. Cyprinid fish species, especially rudd, constitute an important component of the bittern’s diet, second only to eels (G. Gilbert, personal communication). The status of eel populations and their recruitment to wetland sites in the UK present unique and specific problems for stock enhancement and management. Isolation of sites from natural watercourses, an essential component of water level and quality management at many sites, has disconnected them from the natural eel recruitment pathways. This, coupled with a decline in the numbers of elvers entering fresh waters (Moriarty 1990; Knights & White 1993; Knights et al. 1996), presents a potential threat to the bittern’s food base. The problems of a decline in eel recruitment to UK wetlands and possible mitigation measures are dealt with elsewhere (Knights, Chapter 22). A decline in the density of small eels in a wetland will make it less suitable as a breeding site. Consequently this chapter concentrates on the potential for management of rudd populations to compensate for any decline in the availability of eels to bitterns and how habitat management can be used to optimise the availability of rudd to bitterns.

11.2 Methods and materials 11.2.1

Study sites

Survey work was undertaken at two key bittern breeding sites in the UK managed by the RSPB (Fig. 11.1). The two sites chosen, Leighton Moss and Minsmere, both have well established and relatively abundant fish populations (Weaver et al. 1998).

Leighton Moss Leighton Moss reserve in Silverdale, Lancashire (NGR 482 750) (Fig. 11.1) is one of the premiere RSPB reed bed reserves in the UK. It is a valley fen situated on the northwest coastline of Morecambe Bay between the Cumbrian Lake district and the Forest of Bowland. The site is classified as a Site of Special Scientific Interest (SSSI), Ramsar site and an EEC Special Protection Area (SPA). The reserve is characterised by reed swamp/fen vegetation and a number of shallow eutrophic meres essentially acting as one hydrological unit (Fig. 11.2). The site has a total area of 134.5 ha. It has five major habitat types: 79.5 ha of Phragmites reed swamp; 21.5 ha of open water bodies; 6.0 ha of dykes with marginal Juncus and Carex dominated communities; and Salix scrub communities in the drier margins of the site. Leighton Moss has averaged four to five booming male bitterns between 1995 and 2000 (G. Gilbert, personal communication).

11 11/02/2003 13:43 Page 154

154

Bird, fish, habitat interactions

Figure 11.1

The distribution of bitterns in the UK in 1995 (after Hawke & José 1996)

Figure 11.2 Leighton Moss RSPB reedbed reserve in Silverdale, Cumbria, UK. Black areas represent areas of open water and dyke systems, the white areas within the boundary represent Phragmites dominated reed swamp and the grey shaded areas represent scrub and woodland

11 11/02/2003 13:43 Page 155

Managing fish population for an endangered piscivorous bird

155

Habitat management at Leighton Moss is aimed at retaining reed quality, retaining open waters and slowing down hydroseral succession. This is achieved through rotational winter reed and scrub cutting. Reed cutting is undertaken on a 7–12 year rotation, with 4–8 ha cut annually. Open water is retained and floristic diversity encouraged through summer cutting and also spraying of reed with Roundup herbicide. Habitat management is facilitated by manipulation of water levels using the main sluice. A high water level is retained during spring before slowly dropping the level by late summer (RSPB 1995).

Minsmere Minsmere in Suffolk (NGR 460 672) (Fig. 11.1) is the RSPB’s premiere UK reserve and visitor centre. It is characterised as a coastal wetland system within an agricultural catchment. The site is classified as a SSSI, a Ramsar site, Heritage coastline, an Area of Outstanding Natural Beauty (AONB) and an EEC SPA. The site is 935 ha in area and contains eight main habitat types including reed bed, grassland, grazing marsh, woodland and scrub, arable, dune, heath and open water. The wetland component of the site contains 166 ha of reed bed, 17 ha of lagoons and large pools, 1.1 ha of ponds 0.5 ha, 10.4 ha of ponds 0.5–5 ha, and 6.1 ha of ponds 5 ha in size. The Minsmere reed bed is a complex system of hydrological units, essentially formed from three catchments (Figs 11.3 and 11.4), controlled by a complex system of bunds and

Figure 11.3 The main reed bed, lowered areas and southern level hydrological compartments at Minsmere. Black indicates areas of open water and dykes, white indicates areas dominated by Phragmites reed swamp and grey hatching indicates woodland

11 11/02/2003 13:43 Page 156

156

Bird, fish, habitat interactions

Figure 11.4 The North Marsh catchment at Minsmere. Black indicates areas of open water and dykes, white indicates areas dominated by Phragmites reed swamp and grey hatching indicates woodland

water control features. The wetland system of the main reed bed, North Marsh, the lowered areas (compartments regenerated by lowering the bed level between 1994 and 1998) and the Scrape (coastal lagoon system) are extensively controlled by a system of pipe sluices (150 sluices) and one large sluice regulating flow out from the site into the sea. These sluices isolate each compartment in terms of potential fish movement. The reedbed system is of national importance with 2 to 5 booming male bitterns recorded annually (G. Gilbert, personal communication; Welch 1998). Reedbed management at Minsmere, by rotational cutting and water level control, aims to increase the number of resident booming males to six. Rehabilitation and creation of new bittern habitat at the site aims to support another two males (Welch 1998). Since 1994, nine reedbed compartments have been rehabilitated through bed lowering and re-establishment of dykes and open waters (the lowered areas). Reed cutting at Minsmere is undertaken on a seven-year rotational programme.

11.2.2

Assessment of fish populations in wetland complexes

Fish stock assessments at the two sites (Minsmere and Leighton Moss) were undertaken between February 1999 and November 2000. Sampling activities were targeted to cover a range of areas considered representative of the water bodies present at each site. Surveys included the dyke systems, main meres and lowered compartments,

11 11/02/2003 13:43 Page 157

Managing fish population for an endangered piscivorous bird

157

together with the North Marsh reed bed at Minsmere and the main dykes and meres at Leighton Moss. Surveys were undertaken eight times between February 1999 and November 2000 to assess the seasonal abundance and distribution of fish in wetlands. Fish stock assessment was undertaken using hand-held electric fishing gear operated from a small punt. Surveys were undertaken by two persons, one operating the electric fishing gear and the other propelling the punt. Fishing operations were conducted using a single anode (40 cm diameter) powered by a 2.5 kVA generator and Millstream control box with a pulsed DC (50 or 100 Hz) output maintained between 2 and 3 A. The anode was swept through the water in front of the punt covering all available habitats as the punt was slowly manoeuvred forward. Immobilised fish were captured by the fishing operative using a dip net. Auxiliary netting was also performed by the boat operative at the stern. Fish were transferred to a tank of water in the boat until collection of data (numbers of each species, size characteristics measured (fork length to nearest mm and weight (g)) and scale samples removed). The strategy employed in fishing operations varied depending upon the systems being sampled. Narrow (2–3 m) dykes were sampled in a single run, fishing both sides of the punt, and covering both margins. Wider dykes (3 m) were sampled by fishing from one side of the punt covering only one margin. Sections of the larger dykes were sampled in two fishing runs to cover both margins. Margins of open water bodies were sampled in single runs with the fishing operative covering both sides of the punt. Transects across open water bodies were also undertaken at some sites. Catch per unit effort (CPUE) (numbers or biomass m2) was estimated for each of the surveys undertaken in dykes. CPUE in the margins of meres, where no accurate estimate of the area fished was applicable, were calculated as numbers or biomass per metre of margin fished.

11.3 Results 11.3.1

Leighton Moss

The fish community at Leighton Moss comprised eels, pike, rudd perch and sticklebacks. Catches of fish from all areas at Leighton Moss exhibited seasonal variation in distribution and abundance (Fig. 11.5). Biomass CPUE of eels in the main dykes increased from 0.51 g m2 in February to 4.94 g m2 in June. CPUE of eels then declined to 1.86 g m2 by November 1999 (Fig. 11.5). CPUE of all species from the main dykes was relatively low in both August and November 2000 (Fig. 11.5). A few large pike (between 400 and 800 mm FL) captured in the dykes contributed a major proportion of the biomass CPUE in most surveys. CPUE of rudd was low or zero for all areas surveyed until November 1999 when a large aggregation of rudd was captured in the Experimental dyke and main dyke at the downstream end of the site (Fig. 11.5). CPUE of rudd was 39.1 g m2 in the Experimental dyke (Fig. 11.5) and 0.98 g m2 in the main dyke (Fig. 11.5). In April 2000 the majority of rudd had dispersed from the Experimental dyke (CPUE  0.91 g m2) and surrounding area, but two aggregations of rudd were captured from sections of the main central dyke (CPUE  1.14 g m2). The shoals were captured in the main dyke

11 11/02/2003 13:43 Page 158

158

Bird, fish, habitat interactions

near the causeway bridge and the large willow trees adjacent to the dyke upstream of Grisedale mere (Fig. 11.2). The CPUE of fish in the margins of Island Mere showed a similar trend to catches from the main dykes (Fig. 11.5). CPUE of eels increased from 4.75 g m1 in February 1999 to 26.08 g m1 in August 1999 and then declined to 5.09 g m1 by April 2000. CPUE of perch in the margins of Island Mere increased from 0.21 to 3.96 g m1 between February 1999 and April 2000. No surveys were undertaken on Island Mere after April 2000 (Fig. 11.5).

Figure 11.5 Catch composition presented as CPUE of fish caught by electric fishing the main dykes, Experimental dyke and Island Mere at Leighton Moss, February 1999 to November 2000

11 11/02/2003 13:43 Page 159

Managing fish population for an endangered piscivorous bird

11.3.2

159

Minsmere

The fish communities in the different compartments at Minsmere comprised a combination of rudd, roach, eels, tench and sticklebacks. However, species composition varied between compartments (Fig. 11.6). No pike or perch were captured at Minsmere. Roach, having not been previously recorded at the site (Weaver et al. 1998), comprised up to 50% of biomass in some compartments, although they were not present in all areas (Fig. 11.6). Roach were only captured in the main reedbed compartment, North Marsh catchment, in only one of the lowered compartments (31 East) and some of the dykes surrounding the lowered areas (West Waltons, Compartment 38 and Old Tree Hide pool; Fig. 11.3). Tench were only captured in the North Marsh catchment. CPUE of rudd and roach in the Centre Bank ditch at Minsmere during 1999 (February to November) showed temporal variation in the abundance of fish present. CPUE was highest in the February survey (rudd  5.21 g m2, roach  4.62 g m2) and declined to 1.39 g m2 of roach and 0.44 g m2 of rudd in July. The CPUE of both species increased by August to 4.07 g m2 of roach and 3.49 g m2 of rudd (Fig. 11.7).

Figure 11.6 Catch composition by % of biomass of fish captured by electric fishing in different compartments at Minsmere

Figure 11.7 Catch composition as CPUE of fish caught by electric fishing in the Centre Bank ditch at Minsmere, February to November 1999

11 11/02/2003 13:43 Page 160

160

Bird, fish, habitat interactions

Island Mere 100

80

80

Biomass (%)

Numbers (%)

Island Mere 100

60 40 20

60 40 20 0

0 Apr. 00

Apr. 00

Aug. 00

Rudd

Roach

Aug. 00

Month of sample

Month of sample Eels

Stickleback

Tench

Figure 11.8 Catch composition by % numbers and % biomass of fish caught by electric fishing in the margins of Island Mere, Minsmere, in two months during 2000

The abundance of rudd and roach in the margins of Island Mere varied between spring and summer (Fig. 11.8). In April rudd and roach in spawning condition were found. Roach contributed 54% of the fish biomass. Very few rudd were observed in the margin during the April surveys, contributing only 20% of the biomass (Fig. 11.8). By August roach had dispersed away from the margin of Island Mere, and rudd comprised 89% of the fish biomass with roach contributing 1% (Fig. 11.8). The shift in biomass composition was linked to observed changes in size composition of rudd and roach populations associated with the margins (Fig. 11.9). Catches from the margins of Island Mere in April were dominated by roach with a modal size of 140 mm. In August only a few roach (n  7) were captured all of which had fork lengths of 100 mm. The size structure of rudd associated with the margins was broadly similar in both April and August 2000 (Fig. 11.9). Differences in the size structure of rudd populations in dykes and meres were found in surveys in Boat Mere (Fig. 11.3), a shallow mere associated with the Centre Bank ditch. Catches of rudd from Centre Bank ditch comprised individuals in the size range 20–190 mm. Strong size classes were observed with modes around 40, 90 and 125 mm with only 6% of individual 150 mm. Catches of rudd from Boat Mere comprised individuals in the size range 20–220 mm. Abundant size classes were observed with modes around 30, 40, 145, 165, 185 and 205 mm, with 43.3% of individuals 150 mm. Only 8.3% of rudd captured in Boat Mere were in the size range 65–125 mm compared with 50% of rudd captured in Centre Bank ditch (Fig. 11.10).

11.4 Discussion Fish stock assessment surveys at Leighton Moss and Minsmere revealed the dynamic nature of cyprinid fish distribution in wetlands. The seasonal variation in cyprinid distribution and habitat use has important implications for the seasonal availability of fish

11 11/02/2003 13:43 Page 161

Managing fish population for an endangered piscivorous bird

161

April 15 Rudd n = 18

10

Frequency

5

5

10

Frequency

Rudd n = 51

5

245

210

175

140

105

35

0

245

210

175

140

105

70

35

0

August

15

70

0

0

15 Roach n=7

10

5

245

210

175

140

0

245

210

175

140

105

70

35

0

105

0

0

70

Frequency

Roach n = 51

10

35

Frequency

15

Length class (5-mm classes)

Length class (5-mm classes)

Figure 11.9 Length frequency distribution of rudd and roach caught by electric fishing the margins of Island Mere, Minsmere, in April and August 2000

180

195

210

180

195

210

165

135

120

105

90

75

60

45

30

15

150

n = 134

10 8 6 4 2 0 0

Frequency (%)

Centre Bank ditch

Length class (5-mm classes)

165

135

120

105

90

75

60

45

30

15

150

n = 60

0

Frequency (%)

Boat Mere off CBD 10 8 6 4 2 0

Length class (5-mm classes)

Figure 11.10 Comparison of length frequency distribution of rudd caught by electric fishing in Centre Bank ditch and an associated section of open water, Boat Mere, at Minsmere, November 1999

11 11/02/2003 13:43 Page 162

162

Bird, fish, habitat interactions

to bitterns. The general distribution patterns at Leighton Moss and in the main reed bed at Minsmere indicate that cyprinids tend to aggregate in the deeper, more enclosed, water of dykes during the winter period as opposed to the shallower meres favoured during the summer months. Cyprinids dispersed into the meres to spawn during late spring and early summer. Roach began to disperse into open waters sooner than rudd, forming spawning shoals in marginal areas during spring. However, by the summer, larger roach (100 mm) had dispersed away from the margins with only rudd remaining associated with the reed edge. This seasonal and species-specific variation in the association of cyprinids with reed edge has important implications for their availability to bitterns. The marginal areas of open waters and dykes are therefore used only at certain times of year. Consequently, different areas/types of reed edge will only be suitable foraging zones at different times of year. The carrying capacity of a site for bitterns will in part be regulated by the availability of these feeding zones during critical periods, for example, during the breeding season. The changing distribution of fish at Leighton Moss, a site comprising a single hydrological unit, and the fragmented distribution of fish at Minsmere, a complex system of multiple hydrological units, indicate the importance of connectivity within wetlands. The use of different types of water bodies within a wetland at different times of the year highlights the varying habitat requirements of self-sustaining fish populations (Cowx & Welcomme 1998). The importance of shallow open waters for spawning and the larger size classes of rudd (the spawning stock), and of deeper enclosed dykes for overwintering mean that provision of these habitats and good connectivity between them is a prerequisite of a good fish population. Compartmentalisation of a reed bed for water management purposes, essential for the conservation of many wetlands, fragments the habitat and will disrupt connectivity between habitats. Habitat management within a compartmentalised reed bed must therefore provide sufficient, essential habitats within each compartment to maintain self-sustaining fish populations. This will be especially important for sites where the movement of fish through sluices is likely to be minimal or non-existent. The appearance of roach, a species not previously recorded at Minsmere (Weaver et al. 1998), may have important implications for food availability to bitterns. The distribution of roach in Minsmere suggests that they may have only recently entered the site. The absence of roach from all but one of the lowered areas (31 East, the most recently lowered in 1997/1998) suggests that they were not present in the site at the time that these areas were re-flooded (before 1997). Roach have not yet been identified in the diet of bitterns at Minsmere (G. Gilbert, personal communication) despite roach comprising up to 50% of the biomass in some compartments. Roach, once in a system, can quickly become the dominant species by out-competing rudd (Burrough et al. 1979). Since roach do not appear to associate with marginal habitat as much as rudd (especially during critical breeding periods), replacement of rudd with roach as the dominant species may reduce food availability to bitterns. This situation will require careful monitoring and if a shift in dominance from rudd to roach is detected to have a negative impact on bittern status then mitigation measures will have to be implemented. Options for management of habitat and fish populations to optimise conditions for rudd must be considered before such an event, so that mitigation measures can be implemented without delay.

11 11/02/2003 13:43 Page 163

Managing fish population for an endangered piscivorous bird

163

Management, creation and rehabilitation of reed beds must take account of the seasonal dynamics and habitat requirements of cyprinids. Wetland design must provide all the habitat types required for each stage of their life cycle (Cowx 2001). Only in this way will management create self-sustaining fish populations where bottlenecks to recruitment are minimised (Cowx & Welcomme 1998). Management of aquatic habitats can be used to enhance the carrying capacity of wetlands for fish, potentially enhancing the carrying capacity of the site for bitterns through enhancement of the food base. Enhancement of the fishery through management of the open water habitat must be concurrent with management of the reed edge to enhance the availability of optimal foraging areas for bitterns. Management of the reed edge habitat, the principle overlap between fish and bitterns, must in reality manage the reed/water interface from both sides – the open water habitat and the reed bed. Management of the reed bed (edge maturity and stem density) must provide optimal foraging habitat for bitterns and allow good conditions for occupation by fish. Management of the open water habitat in marginal areas is required to create habitat that will attract fish into the marginal areas during the critical period of the bittern’s breeding season. The physical features of the margin, such as water depth and edge profile, can be adjusted using existing rehabilitation techniques of bed lowering and hydrological management (Hawke & José 1996). The biological features of the reed edge, such as stem density and edge maturity, can be controlled with rotational reed-cutting and water level management regimes (Hawke & José 1996). Research is currently ongoing to determine the optimal reed edge conditions for occupation by fish. This information will be integrated with existing data about bittern foraging habitat (RSPB, unpublished data). Overlapping these two optimal habitats on a spatial and temporal scale is essential to create optimal feeding conditions for bitterns. Occupation of reed beds by bitterns has already been shown to be positively correlated with the total length of reed/water interface at a site. Management of the existing reed/water interface so that a higher proportion of it is optimal foraging habitat will potentially enhance the reproductive success and density of bitterns in UK wetlands. Although management of habitat can optimise the availability of rudd to bitterns there is little that can be done through habitat management to enhance recruitment of eel populations to wetlands. The principal factor limiting recruitment of eels to wetlands is the connectivity of the site to an elver/glass eel run and the strength of that run. Management of wetlands often requires isolation of sites from natural water systems to ensure that sites remain sufficiently wet during critical dry periods. This isolation has reduced the recruitment of small eels to these sites. This, exacerbated by the decline in elver runs in Europe (Moriarty 1990; Knights & White 1993), has important potential implications on the eel populations as the bittern’s food base. While management of cyprinid fish populations (especially rudd) can optimise the availability of alternative prey, management measures must also attempt to maintain a good supply of small eels to the site (Knights, Chapter 22).

Acknowledgements This research was undertaken by the first author as a PhD at the University of Hull International Fisheries Institute for a project sponsored by the Royal Society for the

11 11/02/2003 13:43 Page 164

164

Bird, fish, habitat interactions

Protection of Birds. The author would like to thank G. Gilbert, G. Tyler and K. Smith of the RSPB Aquatic Research department for making available the results of current bittern research.

Reference Burrough R.J., Bregazzi P.R. & Kennedy C.R. (1979) Interspecific dominance amongst three species of coarse fish in Slapton Ley, Devon. Journal of Fish Biology 15, 535–544. Cowx I.G. (2001) Factors Influencing Coarse Fish Populations in Rivers. R&D Publication 18. Bristol: Environment Agency, 146 pp. Cowx I.G. & Welcomme R.L. (eds) (1998) Rehabilitation of Rivers for Fish. Oxford: Fishing News Books, Blackwell Science, 260 pp. Cramp S. & Simmons K.E.L. (1977) Handbook of the Birds of Europe, the Middle East and North Africa: The Birds of the Western Paleartic. Vol. 1. Ostrich to Ducks. Oxford University Press. 722 pp. Gentz K. (1965) Die Grosse Dommel. Wittenberg, Lutherstadt. Cited in Cramp, S. & Simmons, K.E.L. (1977) Handbook of the Birds of Europe, the Middle East and North Africa: The Birds of the Western Paleartic. Vol. 1. Ostrich to Ducks. Oxford University Press, 722 pp. Hawke C.J. & José P.V. (1996) Reedbed Management for Commercial and Wildlife Interests. Sandy: Royal Society for the Protection of Birds, 212 pp. José P.V. (2000) Urgent Conservation Action for the Bittern Botaurus stellaris in the United Kingdom. Overall Project Report 1 July 1996–31 March 2000. Sandy: Royal Society for the Protection of Birds, 61 pp. Knights B. & White E. (1993) Anguillaphiles meet in Poland – and ask ‘Is the European eel an endangered species?’ Fish 32, 15–16. Knights B., White E. & Naismith I.A. (1996) Stock assessment of European eel Anguilla anguilla L. In I.G. Cowx (ed.) Stock Assessment in Inland Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 431–447. Moriarty C. (1990) European catches of elver 1928–1988. International Revue gestalt Hydrobiologie 75, 701–706. Newbery P., Schäffer N. & Smith K.W. (1997) European Union Bittern Botaurus stellaris Action Plan. Brussels: RSPB, Birdlife International, European Commission, 34 pp. Perrow M.R., Holve H. & Jowitt A.J.D. (1996) A review of the factors affecting the status of fish populations in the emergent plant zone of wetland habitats, particularly beds of reed (Phragmites australis), with respect to the habitat requirements of bitterns (Botaurus stellaris). Unpublished report to RSPB, Sandy, 60 pp. RSPB (1995) Management Plan: Leighton Moss and Morecambe Bay. Sandy: RSPB, 106 pp. Smith K.W. & Tyler G.A. (1993) Trends in the numbers of breeding bitterns in the UK. Britain’s Birds in 1990–1991. Sandy: RSPB, pp. 139–140. Tyler G.A. (1994) Management of reedbeds for bitterns and opportunities for reedbed creation. RSPB Conservation Review 8, 57–62. Tyler G.A., Smith K.W. & Burges D.J. (1998) Reedbed management and the breeding bittern Botaurus stellaris in the UK. Biological Conservation 86, 257–266. Weaver D.J., Williams R., Cook A. & Tyler G.A. (1998) Report on electric fishing autumn/winter 1997/98. Internal RSPB draft report, 36 pp. Welch G. (1998) Minsmere Nature Reserve Plan 1998 to 2003: Combined Management, Business and Communication Plans. Sandy: RSPB, 179 pp.

12 11/02/2003 13:45 Page 165

Chapter 12

On the feeding ecology of the pied kingfisher, Ceryle rudis at Lake Nokoué, Benin. Is there competition with fishermen? A. LAUDELOUT and R. LIBOIS* Unité de recherches zoogéographiques, Institut de Zoologie, Université de Liège, Liege, Belgium

Abstract Lake Nokoué, in southern Benin, is a heavily exploited fishery, but it is also inhabited by numerous piscivorous birds, especially kingfishers. This chapter considers the similarity between the diet of kingfishers and fish available on the local market between midFebruary to mid-May 1999, during a low water level period. Excretory pellets were collected on the top of breeding banks and inside brood chambers. The diet was determined by comparing the bones recovered from the pellets with a reference collection. Eighteen prey categories were recognised in the 1099 diagnostic items. Kingfishers preyed mostly on cichlids (Sarotherodon melanotheron Rüppell and Hemichromis fasciatus Peters), clupeids (Ethmalosa fimbriata (Bowdich)), eleotrids (Kribia sp.) and Hyporhamphus picarti (Val.). Prey size of H. fasciatus ranged from 22 to 73 mm (46.4  11.6 mm) and for S. melanotheron from 24 to 65 mm (44  9.2 mm). The composition of the diet varied depending on time and location. Overlap with marketed fish is limited to S. melanotheron. Keywords: Benin, Ceryle rudis, diet, feeding ecology, fisheries.

12.1 Introduction Lake Nokoué is a large lagoon situated near Cotonou, the economic capital of the Republic of Benin, and is crucial to the local economy. It covers an area of approximately 150 km2 in the dry season extending to 1000 km2 at the peak of the floods. It is the largest permanent lake of the country and is connected by a 5-km long channel with the Atlantic Ocean. A dense human population is established in villages built on piles. The main activities are agriculture, trade and fishing, the latter of which involves about 90 000 persons (Laleye 1995). Fish are caught by various types of nets, long lines and especially in privately-owned brush park enclosures called ‘akadjas’. These enclosures are made of immersed tree branches that allow the development of a rich plankton and provide good shelter to the fish. Twenty years ago, the annual production of this system was estimated at 4 t ha1, but this has dropped to 1–2 t ha1 (Aglinglo 1998). Local

*Correspondence: Roland Libois, Unité de recherches zoogéographiques, Institute de Zoologie, Université de Liège, Quai Van Beneden, 22, B-4020 Liege, Belgium (email: [email protected]).

12 11/02/2003 13:45 Page 166

166

Bird, fish, habitat interactions

fishermen also use other methods, such as crab pots or traps, designed to catch estuarine fish that move with the current. Some otters and many fish-eating birds, namely cormorants, egrets, herons, terns and hundreds of kingfishers also exploit this important resource (Schockert 1998). As fish is the most important animal protein source for the human population, these birds are often considered as potential competitors. This chapter examines the possible impact of one of the most common piscivorous groups – fish-eating birds – on the fish resources, using a study carried out on the pied kingfisher, Ceryle rudis (L.), the most numerous piscivorous bird species. This species was chosen because its nests or resting perches are not difficult to locate and its food-remains are conspicuous pellets, available on the banks or in the nest brood chambers. Moreover, cichlids, a target species of the fishery, are the dominant food of kingfishers (Tjomlid 1973; Douthwaite 1976; Whitfield & Blaber 1978; Reyer et al. 1988). As they are able to catch cichlids larger than 10 cm (Douthwaite 1976), pied kingfishers may be a potential competitor of fishermen, especially as the size of fish in the market has decreased due to overfishing. Pied kingfishers also occasionally prey on frogs, crustaceans, aquatic insects and even termites (Tjomlid 1973; Douthwaite 1976; Cooper 1981) and are able to survive and even to thrive, feeding mainly on small pelagic fish: clupeids, or cyprinids where cichlids are rare (Junor 1972; Jackson 1984). Their adaptability to strong changes in the ichtyocenoses was illustrated in Lake Victoria after the introduction of the Nile perch, Lates niloticus (L.) (Wanink & Goudswaard 1994), where they shifted their diet from cichlids to the small pelagic cyprinid, Rastrineobola argentea (Pellegrin). They hunt either from a perch or hovering flight. This regularly observed behaviour allows the birds to fish in pelagic waters, which is uncommon in other kingfisher species. Rough estimates of their food consumption indicate a daily intake varying from 17.5 to 26.5 g (Tjomlid 1973).

12.2 Study area Southern Benin is in a subequatorial climate zone (Fig. 12.1), with a high relative humidity (77–93%) and a high mean monthly temperature ranging from 22.4 to 32.9°C. Annual rainfall is about 1000 mm distributed into a long rainy season from March/April to July and a short season from September to mid-October (Pliya 1980). Lake Nokoué (6°23–28N, 2°22–33E) is a shallow lagoon not exceeding 2.50 m in depth. In 1990 and 1991, its mean depth ranged from 1.07 m at the end of the dry season (April) to 1.72 m during the floods (September). Its waters are relatively turbid, especially during the floods: Secchi depth varies between 50 and 120 cm in Vêki, in the vicinity of the study sites. Salinity also fluctuates widely: from 25–30 mg litre1 in April–June to 0–5 mg litre1 in August–November (Laleye 1995). The fish community comprises at least 78 species from freshwater, brackish or marine origin, but is dominated throughout the year by three families: clupeids (Ethmalosa fimbriata (Bowdich), Pellonula leonensis Boulenger, P. vorax Günther), cichlids (Sarotherodon melanotheron Rüppell, Tilapia guineensis (Günther)) and bagrids (Chrysichtys auratus Geoffroy St Hilaire, C. nigrodigitatus (Lacépède)) (Laleye 1995).

12 11/02/2003 13:45 Page 167

Pied kingfisher feeding ecology in Benin

167

Figure 12.1 Schematic map of Lake Nokoué indicating the three study sites

12.3 Methods This study was based on analysis of the excretory pellets (Doucet 1969; Douthwaite 1976; Whitfield & Blaber 1978; Hallet-Libois 1985). To identify the remains, a reference collection of the skull bones of the main fish species present in the area was made from fish bought on the local markets and identified by P. Laleye (see Fig. 12.2 for examples). Diagnostic bones were chosen in this collection to make possible the specific or generic identification of the fishes. Pellets were collected approximately every 2 days in the delta of the River Sô, northwest of Lake Nokoué (Fig. 12.1) from mid-February to mid-May 1999. They were found on the top of the breeding banks or excavated from three nest brood chambers. When recovered from the banks, they were analysed without further treatment whereas the brood chamber material was cleaned by immersion in water for a few days. Soaked pellets were then sieved under a weak water jet and dried before the characteristic skull bones were sorted, counted and some measured. In each sample, right and left bones were counted separately and the minimum number of prey belonging to a taxonomic category was considered as the maximum value of either count for this category. For S. melanotheron and Hemichromis fasciatus Peters, the standard length of the prey fish was determined from the length of the diagnostic bones using fish length–bone length relationships developed for local fishes (Fig. 12.3). The G test was used to compare the differences between the diets of different groups of birds, i.e. by time, location and size. Similarity between the diet of birds with the catch of fishermen was made using Pianka’s equation: O1

∑ pij  pik 2

where pij is the proportion of species i in the diet and pik its proportion in the marketed fish.

12 11/02/2003 13:45 Page 168

168

Bird, fish, habitat interactions

Figure 12.2 Diagnostic bones of some of the fish species present in lake Nokoué. The preopercular bones of three cichlids are shown (top left) as well as their premaxillary (top right)

12 11/02/2003 13:45 Page 169

Pied kingfisher feeding ecology in Benin

169

80

Standard length (mm)

70 60 50 40 S. melanotheron

30

H. fasciatus 20 2

4

6 8 Bone length (mm)

10

12

Figure 12.3 Relationships between the standard length and the preopercular bone length in Sarotherodon melanotheron (SL  5.731  BL  10.132; r  0.981; n  32) and in Hemichromis fasciatus (SL  7.638  BL  3.600; r  0.987; n  27)

12.4 Results The main prey of pied kingfisher at lake Nokoué were E. fimbriata (29%), S. melanotheron (24%) and H. fasciatus (22%). Other important species were the small Kribia species (10%) and Hyporhamphus picarti (Val.) (8%). Seven other fish species and arthropods accounted for the remanding 7.5% of the diet. However, the diet varied according to location, season and the age of birds. This was illustrated by comparing content of pellets taken from different parts of the lake at different times (Table 12.1). A sample taken from a nest along the River Sô 2 days before hatching (midFebruary) contained 54% and 37% of H. fasciatus and S. melanotheron respectively. The remainder comprised Kribia (4%) and Clarias sp. (6%). This is in contrast to pellets taken from a nest situated along an oxbow, where one 8-day-old nestling was present and the diet comprised mainly H. fasciatus (42%) and E. fimbriata (33%), with few S. melanotheron (4%), although Mugilidae were also important (14%). During the fledgling period (mid-March–mid-May), the diet from Vêki, along the shores of the lake, was much more diversified, comprising 15 different prey categories, of which E. fimbriata was the most important (36%). Hemichromis fasciatus was much less frequent than in the other places and than S. melanotheron. The importance of H. picarti and Kribia also increase at this site. These differences observed were highly significant (Gcorr  85.4; P  0.001; 16 d.f.). (Note, the G-statistic was computed on the basis of the species of fish, except the clupeids which were considered as one category, the three species of cichlids which were considered separately, and the Bagridae and Clariidae as a single category. Elops sp., Gerres melanopterus (Bleeker), Tilapia guineensis and the arthropods were also grouped, as were P. jubelini, Monodactylus sebae, H. picarti, Kribia sp. and Yongeichtys thomasi (Boulenger).) The partial Gs were highly significant for individual prey categories, with the exception of Y. thomasi.

12 11/02/2003 13:45 Page 170

170

Bird, fish, habitat interactions

Table 12.1 Nokoué

Local variations in the diet of the pied kingfisher in the western part of Lake Nest 1 adults mid-Feb

Ethmalosa fimbriata Unidentified Clupeidae Hyporhamphus picarti Hemichromis fasciatus Sarotherodon melanotheron Tilapia guineensis Gerres melanopterus Kribia sp. Yongeichtys thomasi Elops sp. Clarias sp. Chrysichtys nigrodigitatus Mugil sp. Unidentified fish Crustaceans Coleoptera Termites Other insects Number of prey items

Nest 2 adults (nestlings?) late Feb 37

61 42

47 4

4

2 5

6 1

114

2 16

113

Vêki adults mid-March–early May 259 13 88 110 174 6 10 95 14 4 2 1 4 2 1 1 784

To illustrate temporal variations, the samples collected in Vêki were grouped by ten-day periods, the first collected in mid-March and the last at the beginning of May (Fig. 12.4). As a whole, the temporal differences were highly significant (Gcorr  109.3; P  0.001; 25 d.f.) except for the category ‘other prey’, grouping T. guineensis, Y. thomasi, Elops sp., Chrysichtys nigrodigitatus and the arthropods. However, samples taken during March were not statistically different from each other, as were the two samples from early April and the samples from late April and early May. In March, cichlids were the dominant prey, contributing about 50% of the food items. The contribution of cichlids fell considerably in early April but recovered thereafter, whereas the proportion of the other fish remained constant with the exception of clupeids, which increased from about 10 to 60%. The contribution of E. fimbriata decreased slowly to 35% at the beginning of May. It should be noted the contribution of H. fasciatus declined progressively over the 2-month period but the importance of S. melanotheron remained relatively stable. Differences in the diet of adults and nestlings were assessed by comparing the diet of the adults from Vêki during late April with that of old nestlings found in a nest situated nearby at the same time. As adults brood their offspring until they are about 10–11 days old, the first adult nestling sample was taken on day 11 to eliminate the mixing of young and adult pellets previously excreted. The same prey categories appeared in the diets of the adults and nestlings (Table 12.2) and the differences were not significant (Gcorr  7.5; P  0.05; 7 d.f.). However, the partial Gs for the cichlids,

12 11/02/2003 13:45 Page 171

Pied kingfisher feeding ecology in Benin

171

Figure 12.4 Temporal shifts in the relative abundance of prey items in the diet of pied kingfisher at Vêki Table 12.2 Comparison of the composition of adult and nestling pied kingfishers diet at Vêki

Ethmalosa fimbriata Unidentified Clupeidae Hyporhamphus picarti Hemichromis fasciatus Sarotherodon melanotheron Tilapia guineensis Gerres melanopterus Kribia sp. Yongeichtys thomasi Elops sp. Mugil sp. Crustaceans Coleoptera Other insects Number of prey items

Nestlings

Adults

15

157 13 27 54 102 3 6 49 10 4

27 38 5 2 1

88

2 2 1 430

E. fimbriata and H. picarti were significant (P  0.01); the proportion of both cichlids being much more important (74% vs 36%) to nestlings, whereas E. fimbriata was the main prey of the adults. Hyporhamphus picarti was not found in the diet of the nestlings. There were, however, differences in the size of the two main prey species

12 11/02/2003 13:45 Page 172

172

Bird, fish, habitat interactions 40

Relative frequency (%)

35

Tilapia/pulli (n = 42) Tilapia/ad (n = 69) Jewelfish/pulli (n = 26) Jewelfish/ad (n = 52)

30 25 20 15 10 5 0

21–25 26–30 31–35 36–40 41–45 46–50 51–55 56–60 61–65 66–70 71–75 Standard length (mm)

Figure 12.5 Size distribution of two species of fish in the diet of adult and nestlings of pied kingfishers at Vêki

in the diets of adults and nestlings (Fig. 12.5). In S. melanotheron, the size of fish consumed were not statistically different (P  0.284, Kolmogorov-Smirnov test) whereas the H. fasciatus eaten by adults were much smaller than the fish brought to the nestlings (P  0.015). The catch from four akadjas (Aglinglo 1998) in the same area as the kingfishers were feeding included the same species, except Pomadasys jubelini (Cuvier), Monodactylus sebae (Cuvier) and Chrysichtys auratus, but the proportional representations were very different (Table 12.3). More than one-half of the akadjas catch was S. melanotheron, but this species comprised less that 25% of the diet of kingfishers. By contrast, H. fasciatus and E. fimbriata each contributed about 25% to the kingfisher diet but 5% to the commercial catch. Tilapia guineensis is an important (20%) commercial species but rarely preyed upon by kingfishers. Conversely, important prey of the kingfisher, Kribia sp. and H. picarti, were not found in the marketed fish. These differences were highly significant (Gcorr  157.8; P  0.001; 8 d.f.). (Note, the G-statistic was computed on the basis of the families of fish, except the three species of cichlids which were considered separately, Bagridae and Clariidae were grouped, as were P. jubelini, Monodactylus sebae, on the one hand and H. picarti, Kribia sp. and Y. thomasi on the other hand.) Overlap between the composition of the kingfisher diet and commercial catches was therefore rather limited (O  0.349). Taking into account the size distributions of H. fasciatus and S. melanotheron preyed upon by kingfishers and caught by the akadjas (Fig. 12.6), the similarity fell to O  0.082. Indeed, there was also little similarity (O  0.082) in the sizes frequency distributions of S. melanotheron and H. fasciatus, preyed upon by kingfishers and caught by the akadjas (Fig. 12.6); most cichlids harvested by the akadjas were 70 mm while they were 75 mm in the kingfisher diet.

12 11/02/2003 13:45 Page 173

Pied kingfisher feeding ecology in Benin

173

Table 12.3 Comparison of pied kingfisher diet and catch from four akadjas in the western part of Lake Nokoué (Source: Aglinglo 1998) Akadjas Ethmalosa fimbriata Unidentified Clupeidae Hyporhamphus picarti Hemichromis fasciatus Sarotherodon melanotheron Tilapia guineensis Gerres melanopterus Pomadasys jubelini Kribia sp. Yongeichthys thomasi Bagridae/Clariidae Elops lacerta Liza falcipinnis Unidentified Mugilidae Other fish Number of items

40 28 435 170

Kingfisher diet 311 13 88 245 258 6 10

47 51 10 36 7 824

106 21 11 4 17 1 1091

Figure 12.6 Size distributions of two species of fish in akadjas catches (Source: Aglinglo 1988) and in the diet of the pied kingfisher. For S. melanotheron, n  430 and 256 respectively in the akadjas and in the kingfisher diet. Corresponding numbers for H. fasciatus are 27 and 229

12 11/02/2003 13:45 Page 174

174

Bird, fish, habitat interactions

12.5 Discussion The main prey of the pied kingfisher around lake Nokoué is cichlids. These species are demersal and strictly should not be available to the bird. However, small-sized individuals of H. fasciatus and S. melanotheron are often found in shallow waters near the banks (Gosse 1963) where they are more vulnerable to predation by kingfishers. The second most important prey was Clupeidae, especially E. fimbriata. This pelagic species lives in dense shoals, and its availability is therefore limited to birds fishing offshore (hovering flight). Other pelagic species of minor importance in the diet are Elops sp. and the mullets. The third important group is Kribia (Kriba nana (Boulenger), Kribia kribensis (Boulenger), or both). These are small demersal freshwater fish, measuring less than 6 cm (total length) (Maugé 1986), occasionally found on the sandy bottom of streamlets or among aquatic vegetation in running waters (Roman 1975). These requirements may be met in the delta of the River Sô, especially where the akadjas are located. Hyporhamphus picarti is a benthic species, feeding on algae and organic debris. However, its eggs are attached to the aquatic vegetation (Collette & Parin 1990), thus during the reproductive season it may be more vulnerable to the kingfisher. Y. thomasi, G. melanopterus, Clarias sp. and C. auratus are also demersal species but their contribution to the diet is very limited. The general view of the diet reflects not only the different ways the kingfisher hunts for prey (near the banks or offshore), but also the diversity of habitats it exploits, from freshwater (presence of Kribia) to brackish areas. For example, differences were found in the diets of kingfishers between nest 1 and 2 (Table 12.1), such that S. melanotheron was less abundant in nest 2, while E. fimbriata and mullets were not present in nest 1. This was probably because nest 2 was much closer to the shores of the lakes than nest 1 (600 m vs 2600 m) and there were probably more opportunities for these breeding birds to hunt for pelagic fish. Temporal variations in diet were evident around the end of March, when the first rains were registered. At this time the lake level rose about 10 cm, strong winds were evident, the surface became turbulent and the waters became turbid. These changes probably induced some modifications in the fishing behaviour of the bird, as were reported on Lake Victoria where, in normal conditions, pied kingfishers made about 80% of dives from perches but, in unsettled weather only 14% of dives were from perches while the rest were made from a hovering position (Douthwaite 1976). The sudden shift in the diet in late March, from the cichlids to the E. fimbriata correlates with a possible change in feeding behaviour. Once the weather settled, cichlids again became relatively important numerically but the proportion of E. fimbriata remained high, a probable consequence of ongoing rains. The increased representation of Kribia at the end of the period could indicate that the kingfishers search for more sheltered places, preferably hunting along the river than near the banks of the lake. The comparison between the diet of the nestlings and of the adults suggested that adults eat smaller and thinner fish (H. picarti, Kribia and E. fimbriata) than those they bring to their offspring (S. melanotheron and H. fasciatus). This observation is partially explained by the ability of pied kingfisher nestlings to digest bones (Douthwaite

12 11/02/2003 13:45 Page 175

Pied kingfisher feeding ecology in Benin

175

1976). However, while adult birds can eat small fish on the wing (Wanink et al. 1993) or close to their fishing post, when feeding their young they have to carry prey some distance. It is probably more energy saving to carry larger than small prey. When the energy demand of the brood becomes more important, i.e. when the nestlings are 10–12 days old, the parents face an increasing feeding effort. The difference observed in prey category (slender vs stout) or size (small vs large) probably reflects a difference in the behaviour of the parents if fishing for themselves or for their offspring. Similar observations were found for C. rudis (Douthwaite 1976) and the European kingfisher, Alcedo atthis (L.) (Hallet-Libois 1985). The study suggests that kingfishers take a lot of fish that have no economic interest (Kribia sp., H. picarti, Y. thomasi) or that are of low market value (H. fasciatus). The negative economic impact of the bird seems restricted to S. melanotheron. However, the prey items are only small individuals, under the market size. Despite the possibilities that these small tilapias could grow to a marketable size, their predation by the kingfisher is likely counterbalanced by the capture of H. fasciatus in a ratio of about 1 : 1, which reduces predation pressure from this source. This cichlid is a voracious predator of small fish (Hickley & Bailey 1987), and sometimes used to control tilapias (Robins et al. 1991).

12.6 Conclusions This chapter emphasised the high degree of adaptability of the diet of the pied kingfisher. Indeed, even in a similar environment, variations were found either between sites or over very short periods of time. Around Lake Nokoué, pied kingfishers prey mainly on cichlids, E. fimbriata, H. fasciatus and Kribia. However, the importance of these prey items varies depending on the location of the nests, the age of the birds and seasonal climatic events. These modifications in the diet result in a complex decision process integrating environmental factors as well as proximal stimuli from the offspring. Competition with fishermen seems minimal because the overlap between the composition of the bird diet and the marketed fish is restricted mainly to S. melanotheron. However, the individuals taken by the birds are small, out of the range of those caught by fishermen. Nevertheless, the impact of the kingfisher on small S. melanotheron remains difficult to assess, although bird predation on one of the major fish predators of S. melanotheron could have a positive influence on the overall survival of the small tilapias.

Acknowledgements Our stay in the Republic of Benin was made possible thanks to a grant from the Commission ‘Coopération Universitaire au Développement’ (Conseil interuniversitaire de la Communauté française de Belgique). We are grateful to Dr Ph. Laleye for his friendly reception in his laboratory, his support and his scientific advice. Thanks also

12 11/02/2003 13:45 Page 176

176

Bird, fish, habitat interactions

to A. Chikou for his logistic help in the field, to Dr G. Teugels for the information about the ecology of African fishes, to V. Schockert for her kind support and to Dr I. Cowx for his invaluable help in improving the English version of the manuscript.

References Aglinglo C. (1998) Production de poissons dans les acadjas du Lac Nokoué et de la lagune de PortoNovo en république du Bénin. Situation actuelle et perspectives pour une gestion rationelle. Mémoire DES Sciences Naturelles appliquées et Écodéveloppement. Univ. Liège, 58 pp.  ann (in French). Collette B.B. & Parin N.V. (1990) Hemirhamphidae. In P.J.P. Whitehead, M.L. Bauchot, J.C. Hureau, J. Nielsen & E. Tortonese (eds) Fish of the North-East Atlantic and the Mediterranean, Vol 2. Paris: UNESCO, pp. 620–622. Cooper A.S. (1981) Pied kingfisher catches crab at sea. Cormorant 9, 135–136. Doucet J. (1969) Coup d’oeil sur le régime alimentaire du Martin-pêcheur. Aves 6, 90–99 (in French). Douthwaite R.J. (1976) Fishing techniques and foods of the pied kingfisher on Lake Victoria in Uganda. Ostrich 47, 153–160. Gosse J.P. (1963) Le milieu aquatique et l’écologie des poissons dans la région de Yangambi. Annales du Musée de l’Afrique centrale, Tervueren, N.S. 116, 113–270 (in French). Hallet–Libois C. (1985) Modulations de la stratégie alimentaire d’un prédateur: Eco-éthologie de la prédation chez le martin-pêcheur Alcedo atthis (L., 1758), en période de reproduction. Cahiers d’Éthologie Appliquée 5(4), 1–206 (in French). Hickley P. & Bailey R.G. (1987) Food and feeding relationships of fish in the Sudd swamps (river Nile, Southern Sudan). Journal of Fish Biology 30, 147–159. Jackson S. (1984) Predation by Pied kingfisher and whitebreasted cormorants on fish in the Kosi estuary system. Ostrich 55, 113–132. Junor F.J.R. (1972) Offshore fishing by the Pied Kingfisher, Ceryle rudis, at lake Kariba. Ostrich 43, 185. Laleye P. (1995) Écologie comparée de deux espèces de Chrysichtys du complexe lagunaire Lac Nokoué – Lagune de Porto Novo au Bénin. PhD thesis, Univ. Liège, 186 pp. Maugé L.A. (1986) Eleotridae. In J. Daget, J.P. Gosse & D.F.E. Thys van den Audenaerde (eds) Checklist of the Freshwater Fishes of Africa. Vol. 2. Paris, Brussels and Tervueren: ORSTOM, IRScNB and MRAC, pp. 389–398. Pliya J. (1980) La pêche dans le sud-ouest du Bénin. Étude de géographie appliquée sur la pêche continentale et maritime. Paris: ACCT, 296 pp. (in French). Reyer H.U., Migongo-Bake W. & Schmidt L. (1988) Field studies and experiments on distribution and foraging of pied and malachite kingfishers at lake Nakuru (Kenya). Journal of Animal Ecology 57, 595–610. Robins C.R., Bailey R.M., Bond C.E., Brooker J.R., Lachner E.A., Lea R.N. & Scott W.B. (1991) World fishes important to North-Americans exclusive of species from the continental waters of the United States and Canada. American Fisheries Society Special Publication 21, 243 pp. Roman B. (1975) Poissons de la Volta et de la Haute Kemal. Notes sur l’écologie des formes naines. Notes de documentation voltaïque 8, 54–67 (in French). Schockert V. (1998) Étude préliminaire de l’avifaune du lac Nokoué et des milieux humides adjacents (bas delta de la Sô et de l’Ouémé): perspective de développement d’un tourisme ornithologique? Mémoire DES Sciences naturelles appliquées et écodéveloppement. Univ. Liège: Freeman & Co., 61 pp. Tjomlid S.A. (1973) Food preferences and feeding habits of the pied kingfisher Ceryle rudis. Ornis Scandinavica 4, 145–151.

12 11/02/2003 13:45 Page 177

Pied kingfisher feeding ecology in Benin

177

Wanink J.H. & Goudswaard K.P.C. (1994) Effects of Nile perch (Lates niloticus) introduction into Lake Victoria, East Africa, on the diet of pied kingfishers (Ceryle rudis). Hydrobiologia 279/280, 367–376. Wanink J.H., Berger M.R. & Witte F. (1993) Kingfisher at fishburger queens: eating Victorian fast food on the wing. Proceedings of the VIII Pan-African Ornithological Congress, pp. 331–338. Whitfield A.K. & Blaber S.J.M. (1978) Feeding ecology of piscivorous birds at lake St Lucia Part 1: diving birds. Ostrich 49, 185–198.

13 11/02/2003 13:48 Page 178

Chapter 13

Seasonal and spatial variation in cormorant predation in a lowland floodplain river C. WOLTER* and R. PAWLIZKI Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Berlin, Germany

Abstract A total of 523 cormorant pellets were collected, before and after hatch, after fledging and before wintering in 1998. Roach, Rutilus rutilus (L.), perch, Perca fluviatilis L., and ruffe, Gymnocephalus cernuus (L.) were the dominant species (comprising 84%) in the diet. Pellets contained a mean ( standard deviation) of 17.7  13.8 individual fish. The prey fish had a mean total length of 12.5  4.8 cm (5–59 cm) and mean weight of 51.1  67.2 g (4.4–624.8 g). The mean daily food consumption of a cormorant was 490.9  260.7 g, but seasonal differences were found with a peak in June (538.8  291.9 g). Cormorant depredation increased from 4.9 t in 1993 when 20 breeding pairs were present to 228.3 t in 2000 when 930 breeding pairs occupied the colony. Feeding intensity in 2000 was estimated at 136.3 kg ha1 yr1 if the daily feeding migration was within a 20-km radius of the colony, dropping to 10.3 kg ha1 yr1 in a 50-km radius. Approximately 50% of the birds fished within the 20-km range and seasonally high cormorant predation pressure probably arose as a result of feeding behaviour in preferred waters close to the colony. Keywords: feeding ecology, fishery, pellet analysis, Phalacrocorax carbo sinensis.

13.1 Introduction A rapid increase in the numbers of cormorants Phalacrocorax carbo sinensis (Shaw & Nodder) throughout most of Europe over the past 20 years has provoked an increasing conflict with fisheries interests. While predation by piscivorous birds on fish ponds has undoubtedly resulted in substantial losses (reviewed in Suter 1991), long-term declines in fish stocks in natural freshwater habitats caused by birds has not been proven (Suter 1991, 1995, 1997; Dirksen et al. 1995; Veldkamp 1995; Carss & Marquiss 1997; Noordhuis et al. 1997). However, yields can be reduced temporarily because of compensatory balance between fisheries and bird-induced mortality (Dirksen et al. 1995; Kieckbusch & Koop 1997; van Eerden & Zijlstra 1997; Wolter & Pawlizki 2000). There is accumulating evidence for the capacity of piscivorous birds to regulate prey distribution and abundance (Crowder et al. 1997; Logerwell 1997). Consequently, piscivorous birds may be able to prey on fish beyond the threshold of

*Correspondence: C. Wolter, Leibniz-Institute of Freshwater Ecology and Inland Fisheries, Müggelseedamm 310, 12561 Berlin, Germany (email: [email protected]).

13 11/02/2003 13:48 Page 179

Seasonality in cormorant predation

179

compensatory mortality, especially at times of increased energy demand during the rearing of juveniles (D. Grémillet & D. Schmid, unpublished report). The study detailed in this chapter compares the seasonal food consumption of cormorants with fisheries population and community data (Wolter et al. 1999), to assess the temporary and spatial effects of cormorant depredation on the fisheries.

13.2 Materials and methods The cormorant breeding colony studied is situated close to the town Schwedt in northeastern Brandenburg, Germany, within the floodplain area of the Lower Oder Valley national park (Fig. 13.1). The colony was founded in 1971 by three breeding pairs, deserted in 1988, but recolonised in 1992. Counts of breeding pairs were provided by the ornithological station Buckow (T. Ryslavy, unpublished data). Fresh and intact cormorants pellets were collected at the colony on four occasions in 1998: i.e. in May (105 pellets) during incubation, in June (200) after hatch, in August (118) after fledge and in September (100) before wintering, and stored at 18°C until examination. Fish were identified from undigested remains using species-specific diagnostic bones: pharyngeal bones and chewing pads for cyprinids, otoliths and

Figure 13.1 Map of lakes and rivers around the breeding colony near Schwedt in the lower Oder valley (SQF  Schwedter Querfahrt)

13 11/02/2003 13:48 Page 180

180

Bird, fish, habitat interactions

vomer for percids and burbot, Lota lota (L.), teeth and jaws for pike, Esox lucius L., and pikeperch, Stizostedion lucioperca (L.), and otoliths and vertebrae for eel, Anguilla anguilla (L.). The number of specimens was counted as: number of chewing pads (cyprinids only), number of left or right pharyngeal bones (cyprinids only), number of jaws and vomer, and half the number of paired otoliths. The highest number of each bone type was taken as the number of specimens per pellet. Prey length was reconstructed from pharyngeal bone–total length relationships (Radke et al. 2000), from chewing pad width–total length relationships (Pawlizki 1999), from vertebrae length–total length relationships (Mehner 1990), and from otolith length–total length relationships (Suter & Morel 1996). Biomass consumed was calculated using species specific total length–weight relationships from fisheries surveys in the River Oder (Wolter et al. 1999). Two hundred and fifty feeding days per adult bird was used to determine the total annual fish consumption by cormorants, given that the birds leave the colony in October for overwintering further south and return in February. The percentage of each prey species was estimated from the total number of prey items. The number of fish, average prey fish weight and biomass per pellet were calculated using all pellets and food items. Frequency of occurrence as well as mean number of each fish species per pellet were calculated. Seasonality in food consumption was compared between sampling dates by one way ANOVA with a post hoc Tukey test or with a post hoc Dunnett-T3 test in the case of significant deviations from variance homogeneity (Levene test,   0.05). Calculations were performed using SPSS software (SPSS Inc., 1999, release 9.0.1). One pellet collected in September was excluded from the statistical analysis because it contained otoliths from an atypically high number (128) of ruffe, Gymnocephalus cernuus (L.), and therefore represented an extreme outlier for analysis of variance. To estimate the impact of cormorant depredation on the fish stocks around the colony, the area of surface waters within a 50-km radius was measured from satellite images (D-Sat 2.0, SCOUT Systems, Munich). Data on fisheries yield for these water bodies were provided by the local fishermen, and information on fish stock structure was obtained from surveys in various water bodies within the lower Oder valley (Wolter et al. 1999).

13.3 Results 13.3.1

Food consumption

The remains of 9278 fish were found in the 523 pellets examined. The relative contribution to the diet (percentage of specimens) of the nine most important species (representing 1% of the diet) was: roach Rutilus rutilus (L.) (39.8%), perch Perca fluviatilis L. (29.6%), ruffe (14.6%), rudd Scardinius erythrophthalmus (L.) (3.9%), silver bream Blicca bjoerkna (L.) (2.9%), carp Cyprinus carpio L. (2.5%), tench Tinca tinca (L.) (2.4%), Carassius (1.2%) and pike (1.2%). Prussian carp Carassius gibelio (Bloch) and crucian carp Carassius carassius (L.) could not be distinguished from undigested remains and were therefore grouped as Carassius. The importance of prey

13 11/02/2003 13:48 Page 181

Seasonality in cormorant predation

181

Image Not Available

Figure 13.2 Relative abundance (95% CL) of the most common prey items in the cormorants diet during the season. Abundance in the study area according to Wolter et al. (1999) as percentage of total catch

Image Not Available

Figure 13.3 Frequency of occurrence (95% CL) of the most common prey fish in the pellets and total frequency in the study area (Wolter et al. 1999)

species with a relative abundance 1% in the diet of cormorants was relatively stable throughout the season with respect to abundance (Fig. 13.2) and frequency of occurrence (Fig. 13.3). Most variation between sampling periods was found in ruffe, which was significantly more common in the diet in autumn (P  0.05, ANOVA). The importance of perch, ruffe and carp was considerably higher in the pellets than in the waters of the national park. Silver bream and pike were more abundant and more commonly found in the national park waters than in the pellets.

13 11/02/2003 13:48 Page 182

182

Bird, fish, habitat interactions

Figure 13.4 Length frequency distribution of all prey fish (modified after Wolter & Pawlizki 2000)

Figure 13.5 Seasonality of daily fish consumption (95% CL) by cormorants by fish numbers (left) and biomass (right). N  number of pellets

The remains of on average 17.7  13.8 specimens (mean  SD, range 1–128) from 3.5  1.4 fish species (1–8) were found in each pellet. Prey length varied between 5 and 59 cm, with a mean of 12.5  4.8 cm. Most prey fish (85.7 %) were 6–16 cm long (Fig. 13.4). The estimated fish weight consumed varied between 9.3 and 2317.8 g per pellet, with a mean of 490.9  260.7 g. The prey items had a mean weight of 51.1  67.2 g, with a range of 4.4–624.8 g. If cormorants produce one pellet per day (Trauttmansdorff & Wassermann 1995; Zijlstra & van Eerden 1995), their mean daily food intake over 1998 was about 18 fish with a total mass of 491 g. However, fish consumption by cormorants varied significantly between sampling dates (Fig. 13.5). The number of fish per pellet (mean  SD) was significantly lower in May (15.3  10.8)

13 11/02/2003 13:48 Page 183

Seasonality in cormorant predation

183

and June (15.8  11.7) than in September (22.2  15.2) (P  0.001, ANOVA). The fish mass per pellet increased from 465.7  202.0 g in May to 538.8  291.9 g in June shortly after hatch and dropped significantly to 419.8  212.5 g in September (P  0.01, ANOVA) before wintering. Correspondingly, the mean weight of prey items increased from 47.1  37.3 g in May to 60.4  72.3 g in June and then declined significantly to 35.0  54.2 in September (P  0.05, ANOVA). No significant differences in mean daily food intake were detected before and after hatch or between hatch and fledge of the chicks. By contrast, the mean daily food consumption of cormorants in September differed significantly from summer.

13.3.2

Fishing pressure

The breeding colony in the Lower Oder Valley national park is inhabited by cormorants for around 250 days per year from February to October (National Park Authority, unpublished data). The number of breeding pairs increased from 20 in 1993 to 930 in 2000. Total annual fish consumption of breeding pairs only (i.e. no juveniles), based on a daily food intake of 491 g per bird, increased from about 4.9 t in 1993 to 228.3 t in 2000 (Fig. 13.6). Since the total area of surface waters within the national park is 12.36 km2, if the cormorants fish mainly therein, their fishing pressure would equate to a removal of 184.7 kg ha1 yr1 of fish in 2000. This is considerably greater than the harvest of fish by commercial fishermen, which has remained relatively stable at around 24.7  1.3 kg ha1 yr1. Cormorants were not limited to the waters within the national park for feeding. Within a radius of 20 km around the breeding colony there is 16.75 km2 of surface water available for cormorants, within 30 km–45.47 km2, and within 40 km–148.88 km2.

Figure 13.6 Estimated total annual fish removal by cormorants as a function of the observed number of breeding pairs in comparison with the yield of the commercial fishery

13 11/02/2003 13:48 Page 184

184

Bird, fish, habitat interactions

13.4 Discussion Cormorants preyed opportunistically on a broad range of fish species and size classes, however, 84% of the diet was formed by only three species: roach, perch and ruffe. With the exception of perch and ruffe, cormorants selected for prey fishes in accordance with their availability and abundance in surrounding waters. Similar findings were reported throughout Europe (e.g. Schratter & Trauttmansdorff 1993; Dirksen et al. 1995; Keller 1995; Mann et al. 1995; Suter 1995, 1997; Veldkamp 1995; Noordhuis et al. 1997). Cormorants showed a preference for small fish (6–16 cm TL), comparable to findings elsewhere (Schratter & Trauttmansdorff 1993; Keller 1995; Mann et al. 1995; Carss & Marquiss 1997; Noordhuis et al. 1997). Smaller prey items were particularly prevalent in the pellets in September when young-of-the-year-fish (0) (mean length 6–10 cm) become available. Daily fish consumption was highest in June after hatching when cormorants have the greatest energy demands (D. Grémillet & D. Schmid, unpublished). By contrast Dirksen et al. (1995) reported higher food intake before the breeding season and before winter, i.e. October, and Veldkamp (1995) reported an increase in the size and mass of fish consumed before winter. Conflicts between fisheries and fish-eating bird result mainly from depredation on fish stocks and potential loss of catch. Based on the estimated 491 g (490.9  260.7 g) per adult cormorant per day calculated in this study, which falls within this range of 146 g and 700 g per day determined elsewhere (Worthmann & Spratte 1990; Suter 1991, 1997; Keller 1995; Dirksen et al. 1995; Grémillet 1997; Noordhuis et al. 1997; Keller & Visser 1999), the 1460 adult cormorants in the breeding colony in 1998 consumed 717 kg of fish each day. If cormorants were present for 250 feeding days in 1998, the total fish consumed in the lower Oder floodplain was of the order of 180 t, which represents a very high predation pressure (Fig. 13.6). It should also be noted that monthly variation in food intake may lead to intra-annual variation in feeding pressure. Irrespective, the impact of cormorant predation will vary according to foraging distance. T. Keller and M. Vordermeier (unpublished), and Jungwirth et al. (1995) reported daily feeding migrations of more than 40 km, but the lack of coregonids in the diet of cormorants from the Lower Oder Valley colony suggests they feed within 40 km as large lakes beyond this distance are all inhabited by coregonids. This assumption was supported by Kieckbusch & Koop (1997), who observed cormorants only flying up to 30 km to forage. Furthermore, the selection of ruffe, perch and carp indicates the cormorants probably feed at sites where these species are prevalent, i.e. the canals Hohensaaten-Friedrichsthaler-Wasserstraße and Schwedter Querfahrt for perch and ruffe (Wolter et al. 1999), plus local carp ponds that are all situated within the 20 km radius of the colony. The study detailed in this chapter indicates that cormorants from the Lower Oder Valley colony probably feed within 20 km of the roost and that they consumed 107 kg ha1 yr1 from the 16.75 km2 surface waters therein in 1998. If only half of the birds feed within this close range around the colony, their consumption in 1998 was twice the mean commercial harvest of 24.7 kg ha1 yr1 in the study area between 1993 and 2000. Whether this predation pressure will adversely affect the commercial

13 11/02/2003 13:48 Page 185

Seasonality in cormorant predation

185

fisheries has yet to be determined but the increasing number of birds is a worrying trend, which could lead to failure of the fishery.

Acknowledgements We thank Torsten Ryslavy from the ornithological station Buckow for kindly providing data on cormorants breeding pairs. The authors are much indepted to Ian Cowx and one anonymous referee for substantially improving the manuscript. The study was partially funded by Brandenburg’s Ministry of Agriculture, Environmental Protection and Spatial Planning.

References Carss D.N. & Marquiss M. (1997) The diet of cormorants Phalacrocorax carbo in Scottish fresh waters in relation to feeding habitats and fisheries. Ekologia Polska 45, 207–222. Crowder L.B., Squires D.D. & Rice J.A. (1997) Nonadditive effects of terrestrial and aquatic predators on juvenile estuarine fish. Ecology 78, 1796–1804. Dirksen S., Boudewijn T.J., Noordhuis R. & Marteijn E.C.L. (1995) Cormorants Phalacrocorax carbo sinensis in shallow eutrophic freshwater lakes: prey choice and fish consumption in the non-breeding period and effects of large-scale fish removal. Ardea 83, 167–184. Grémillet D. (1997) Stomach temperature probes in cormorants Phalacrocorax carbo: A measurement of the daily food intake in free-living individuals. Ekologia Polska 45, 233–236. Jungwirth M., Woschitz G., Zauner G. & Jagsch A. (1995) Einfluß des Kormorans auf die Fischerei. Österreichs Fischerei 48, 111–125 (in German). Keller T.M. (1995) Food of cormorants Phalacrocorax carbo sinensis wintering in Bavaria, Southern Germany. Ardea 83, 185–192. Keller T.M. & Visser G.H. (1999) Daily energy expenditure of great cormorants Phalacrocorax carbo sinensis wintering at Lake Chiemsee, Southern Germany. Ardea 87, 61–69. Kieckbusch J.J. & Koop B. (1997) Cormorant Phalacrocorax carbo and fishery in SchleswigHolstein, Germany. Ekologia Polska 45, 287–294. Logerwell E. (1997) Trophic Interactions in Pelagic Systems: Distributional and Behavioral Patterns of Seabirds and their Prey. PhD thesis, University of California, 81 pp. Mann H., Zuna-Kratky T. & Lutschinger G. (1995) Bestandsentwicklung und Nahrungsökologie des Kormorans (Phalacrocorax carbo) an der Donau östlich von Wien im Hinblick auf fischereiliche Auswirkungen. Österreichs Fischerei 48, 43–53 (in German). Mehner T. (1990) Zur Bestimmung der Beutefischarten aus Fragmenten der Wirbelsäule bei der Nahrungsanalyse (Osteichthyes, Teleostei). Zoologischer Anzeiger 225, 210–222 (in German). Noordhuis R., Marteijn E.C.L., Noordhuis R., Dirksen S. & Boudewijn T.J. (1997) The trophic role of cormorants Phalacrocorax carbo in freshwater ecosystems in the Netherlands during the nonbreeding period. Ekologia Polska 45, 249–262. Pawlizki R. (1999) Nahrungsökologie des Kormorans, Phalacrocorax carbo sinensis, im Nationalpark ‘Unteres Odertal’. Thesis, Humboldt University of Berlin, 63 pp. (in German). Radke R.J., Petzoldt T. & Wolter C. (2000) Suitability of pharyngeal bone measures commonly used for reconstruction of prey fish length. Journal of Fish Biology 57, 961–967. Schratter D. & Trauttmansdorff J. (1993) Kormorane Phalacrocorax carbo sinensis an Donau und Enns in Österreich: Analyse der Speiballen. Ornithologische Verhandlungen 25, 129–150 (in German).

13 11/02/2003 13:48 Page 186

186

Bird, fish, habitat interactions

Suter W. (1991) Der Einfluß fischfressender Vogelarten auf Süßwasserfisch-Bestände – eine Übersicht. Journal of Ornithology 132, 29–45 (in German). Suter W. (1995) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: two case studies from Swiss rivers. Journal of Applied Ecology 32, 29–46. Suter W. (1997) Roach rules: Shoaling fish are a constant factor in the diet of cormorants Phalacrocorax carbo in Switzerland. Ardea 85, 9–27. Suter W. & Morel P. (1996) Pellet analysis in the assessment of Great Cormorant Phalacrocorax carbo diet: reducing biases from otolith wear when reconstructing fish length. Colonial Waterbirds 19, 280–284. Trauttmansdorff J. & Wassermann G. (1995) Number of pellets produced by immature cormorants Phalacrocorax carbo sinensis. Ardea 83, 133–134. Van Eerden M.R. & Zijlstra M. (1997) An overview of the species composition in the diet of Dutch cormorants with reference to the possible impact on fisheries. Ekologia Polska 45, 223–232. Veldkamp R. (1995) Diet of cormorants Phalacrocorax carbo sinensis at Wanneperveen, The Netherlands, with special reference to bream Abramis brama. Ardea 83, 143–155. Wolter C. & Pawlizki R. (2000) Feeding ecology of cormorants, Phalacrocorax carbo sinensis. Berichte des IGB 10, 121–130. Wolter C., Bischoff A., Tautenhahn M. & Vilcinskas A. (1999) Die Fischfauna des unteren Odertales: Arteninventar, Abundanzen, Bestandsentwicklung und fischökologische Bedeutung der Polderflächen. In W. Dohle, R. Bornkamm & G. Weigmann (eds) Das Untere Odertal. Auswirkungen der periodischen Überschwemmungen auf Biozönosen und Arten. Stuttgart: E. Schweizerbart, pp. 369–386 (in German). Worthmann H. & Spratte S. (1990) Nahrungsuntersuchungen an Kormoranen vom Großen Plöner See. Fischer & Teichwirt 1/90, 2–8 (in German). Zijlstra M. & van Eerden M.R. (1995) Pellet production and the use of otoliths in determining the diet of cormorants Phalacrocorax carbo sinensis: trials with captive birds. Ardea 83, 123–131.

14 11/02/2003 13:52 Page 187

Chapter 14

Assessing the interaction between cormorants and fisheries: the importance of fish community change A. CARPENTIER*, J.M. PAILLISSON and L. MARION Université de Rennes 1, UMR Ecobio 6553, Av. du Gl Leclerc, Rennes, France

Abstract The inter-relationships between the fishery and the largest breeding colony of cormorants, Phalacrocorax carbo, in France on the fish community of a shallow freshwater ecosystem (Lake Grand-lieu) were examined over a 10-year period. During the study period an increase in the importance of cyprinids in the fish community was observed. This was linked to increased eutrophication of the lake and a new management regime for the spring water level, notably a rise in water depth on adjacent marsh grasslands since 1996, which favoured cyprinid recruitment. Whereas the fishery almost exclusively targeted eel, Anguilla anguilla (L.) (77% of catch by weight), cormorants fed on a variety of prey items, especially cyprinids, the contribution of which increased in their diet from 52 to 80% during the study period. The diet change reflected changes in the fish community structure and confirmed the ability of cormorants to shift their diet in response to available food resources. Keywords: commercial fishery, competition, cyprinids, diet, eutrophication, Phalacrocorax.

14.1 Introduction Interactions between fish and birds in Europe are well documented but concern mainly intensive or extensive aquaculture where fish are concentrated and easy to catch by birds (Marquiss & Carss 1994; Russell et al. 1996; Marion 1997a–c). This situation can lead to severe impact from bird depredation and direct economic losses which are relatively easy to estimate. Impact on the natural fish stocks exploited by traditional fisheries or recreational angling on large water bodies are less well understood because of the difficulties in assessing the status of fish stocks and the economic value of fisheries in such ecosystems (Feltham et al. 1999). It is for this reason that avian impact on freshwater fish communities is often only highlighted when the fishery declines or important changes occur in the fish community (Craven & Lev 1987; Carss et al. 1997; Keller et al. 1997; van Dam 1997). Furthermore, in the case of generalist fish-eating bird species such as the cormorant, predation on the whole fish community presents

*Correspondence: Alexandre Carpentier, Université de Rennes 1, UMR Ecobio 6553, Av. du General Leclerc, 35 042 Rennes Cedex, France (email: [email protected]).

14 11/02/2003 13:52 Page 188

188

Bird, fish, habitat interactions

major difficulties in obtaining an accurate estimate of the impact on the fish community (van Eerden & Zijlstra 1997). The study reported in this chapter was based on the largest breeding cormorant colony in France, at Lake Grand-Lieu, between 1990 and 2000, a period sufficiently long to assess both possible modifications of the fish community structure and diet adaptations of birds. This breeding population was established at the beginning of the 1980s following the creation of a Nature Reserve and the implementation of European legislation to protect the species in 1979. Conflicts with the professional fishery rapidly appeared due to a large increase in the number of breeding pairs (from seven in 1981 to 505 in 2000), and a consequent depredation of the stocks from an estimated 9 t of fish in 1989 to 68 t in 1998 (Marion 1997a; Carpentier 1999). In the study reported in this chapter the change in diet of cormorants in relation to the fish community structure and fishery yields were examined through a multivariate statistical approach and potential predator–prey interaction, and possible competition between the cormorants and the commercial fishery are discussed.

14.2 Materials and methods 14.2.1

Study site

Lake Grand-Lieu is a shallow (less than 1.2 m in summer), turbid, eutrophic, natural freshwater ecosystem in western France (47°05N, 1°39W). This lake constitutes one of the most important bird areas in France, supporting large breeding populations of Ciconiiformes (herons) and cormorant. The wetland is designated a Special Protection Area under the EC Birds Directive and a Ramsar Site, and contains a nature reserve. This status offers protection to breeding birds and prevents recreational angling. The wetland is of value because of its diversity of habitats with four major complementary functional units. The lake covers 4000 ha in summer, which increases to 6300 ha in winter through the flooding of adjacent wet peaty grasslands. About 20 km of the lake margin, covering an area of 2000 ha, is covered by a floating peat fen, which becomes progressively exposed in summer. This is a habitat frequently used by many colonial breeding bird species such as cormorants. Between April and October, most (about 1400 ha) of the permanently flooded central area of the lake is covered by floatingleaved macrophytes. The central open water region without floating or emergent plants, except small patches of submerged macrophytes, covers about 600 ha (see Marion & Brient (1998) for a more detailed description of the habitats and the hydrological regime).

14.2.2

Fish sampling

Assessment of the fish composition and stock was realised during two periods: 1990/1991 and 1999/2000. Different sampling methods appropriate for the littoral and offshore regions were used. Sampling was conducted using 10-mm fyke nets (423 net visits)

14 11/02/2003 13:52 Page 189

Interaction between cormorants and fisheries

189

and electric fishing (point abundance sampling; Nelva et al. 1979) totalling 380 spot samples. The second technique was chosen because it is well adapted to shallow waters and is effective for all species and life stages (Randall et al. 1996; Huatagalung et al. 1997; Copp & Penaz 1988). Fish were identified to species, measured (nearest mm total length, TL) and immediately returned to the water. Blicca bjoerkna (L.) and Abramis brama (L.) were grouped as breams as it was difficult to identify juveniles in the field. During the first sampling period, each individual was weighed (nearest g) to determine length–weight relationships for each species (Adam & Elie 1993). These relationships were used for the second period to calculate the individual weights from each length measured. Relative frequencies of each species were defined and considered to be good descriptors of the fish community structure in the lake.

14.2.3

Fishery

The fishery essentially operates in spring and summer in the central open water area and floating macrophyte areas. It is controlled by the central authorities who limit the number of professional fishermen to eight, as well as the types and number of fishing gears (n  120) and types of boats used. Passive fixed fyke nets are generally used and occasionally bosselle traps and long-lines. Data on annual fishery yields, by species, were based on official statistics from the Departmental Prefecture. This information provided trends in the fishery yield between 1990 and 1999. Fishery data did not discriminate roach, Rutilus rutilus (L.), and rudd, Scardinius erythrophthalmus (L.), so they were considered as one item.

14.2.4

Cormorant diet

Cormorant diet was based on two types of data. Firstly, fish regurgitates were collected from the colony during the breeding season and secondly the observations were made on fish caught from the lake by cormorants during their mass feeding bouts. Fish remains were identified, measured and the biomass was assessed using the derived length–weight relations. Note that the numbers of birds feeding on the lake fluctuated widely from almost 600 (31.5%) birds in April 1996 to 1800 (60.0%) in July 1996. Most of the foraging areas used by cormorants had a similar fish community (Feunteun & Marion 1994) so it was assumed that diet did not vary between feeding sites (Marion 1997a).

14.2.5

Data analysis

Changes in the utilisation of the fish stocks by the fishery and the cormorants, were examined by comparing shifts in exploitation patterns between 1989 and 1993 and 1997 and 2000. Principal components analysis (PCA) was carried out on the fish community, fisheries and cormorant diet, with data transformed into frequencies. This multivariate

14 11/02/2003 13:52 Page 190

190

Bird, fish, habitat interactions

procedure, derived from classical principal components analyses, was recently applied to diet composition data (de Crespin et al. 2000) to highlight foraging strategies. For more details of the principle and the interpretation of this multivariate procedure, see de Crespin et al. (2000). The use of proportions removes the unequal weight among years and this procedure is shown as an alternative to traditional statistical tests (chisquare, likelihood ratio chi-square and rank correlation) for detecting interactions between several components and possible competition between cormorants and fishermen. The biplot of annual information on fish species captured shows dominant versus rare fish species represented by long versus tiny arrows, and predicts changes in the fishery exploitation and cormorant diet from a specialist strategy (plots situated near to the end of one arrow) to a more generalist strategy (plots noted between several arrows). Multivariate analyses and graphics were performed with ADE software (version 4, Thioulouse et al. 1997).

14.3 Results 14.3.1

The fish community

Breams were the predominant species in the fish community (Fig. 14.1(a)) (42  10%) followed by eel (19  6.5%) and rudd/roach (15  5%). The proportion of cyprinids in the community increased from 57% to 79% between 1990 and 1991 and 1999 and 2000, mainly due to an increase in the contribution of rudd, roach and breams. Rarer species, such as pike, Esox lucius L., tended to decline in abundance towards the end of the study period (1999–2000) whereas perch, Perca fluviatilis (L.) increased. Carp, Cyprinus carpio L., was well represented in 1999 mainly due to a few large individuals being caught. Sunbleak, Leucaspius delineatus (L.), appeared in 1999 and 2000 while catfish, Ictalurus melas (Rafinesque) disappeared.

14.3.2

Cormorant diet

Cormorants had a diverse diet comprising 13 of the 15 species sampled in the lake, confirming its generalist diet. Cyprinids (breams, rudd, roach and tench) dominated the diet, contributing between 52 and 80% of the food intake (Fig. 14.1(b)). Pike, catfish and eel were also important species in the diet.

14.3.3

The fishery

The fishery harvested (Fig. 14.1.(c)) an average of 41 250  5860 kg year1 between 1990 and 1999. It focused almost exclusively on eel accounting for 77  3% of the harvest throughout the period. The mean biomass of other species caught varied from 5% for catfish and pike to 3% for tench, Tinca tinca (L.), and zander, Stizostedion lucioperca (L.). Carp, roach, rudd and bream were only occasionally caught.

14 11/02/2003 13:52 Page 191

Interaction between cormorants and fisheries

191

Figure 14.1 Comparison of the fish community structure of Lake Grand-Lieu (a) and the diet of cormorants (b) between the early and late 1990s, and the composition of the catch from the commercial fishery (c)

14.3.4

Interaction between fish, birds and fishermen

The first two axes of the PCA (Fig. 14.2) accounted for 97% of the total variance (0.169) in the fish composition. The major part of the variability was accounted for the first factor (89.50%) (Fig. 14.2(a)). The dominant fish species (cyprinids and eels) contributed to the partitioning of the three components (Fig. 14.2(b)): cormorant diet, fishery catches and fish community composition. The output supported marginal changes in the fish community structure over the time period, which was reflected in the cormorant diet, and the targetting of eels by the fishery. Relationships were strong for breams and eel, but less obvious for secondary fish species such as tench, pike and catfish.

14 11/02/2003 13:52 Page 192

192

Bird, fish, habitat interactions

Figure 14.2 Principle Component Analysis (PCA) on relative biomass frequency between the fishery catch composition, cormorant diet and fish community structure according to species and years. (a) Histogram of eigenvalues identifying the relative contribution of the first two axes that defined the average structure. (b) Biplots of annual components (•) on the factorial plane according to the fish species (→). Ninety-nine percent confidence ellipses were used for making groups according to the fishery, cormorant diet and fish community structure

14.4 Discussion 14.4.1

Interactions between cormorants, the fishery and the fish community

Direct comparison between the three components under study (fish stocks, cormorant diet and fishery harvest) showed little evidence of any inter-relationships. However the PCA identified a good relationship between cormorant diet and the fish community structure, as has been found elsewhere (Hald-Mortensen 1997; Suter 1991; Leopold et al. 1998). By contrast, the commercial fishery was almost exclusively focused on eel, suggesting little competition between cormorants and the fishery. Similar scenarios

14 11/02/2003 13:52 Page 193

Interaction between cormorants and fisheries

193

were found elsewhere in Europe, especially in eutrophic freshwater lakes (Marteijn & Dirksen 1991; Suter 1991). As already shown by Marion (1997a), cormorants appear to select positively for the scarcer species, such as pike, tench and catfish, which are of relatively low commercial value (except for pike). However, there was little or no change in the yield of pike to the fishery before or after the establishment of the cormorant colony.

14.4.2

Changes in cormorant diet and the fish community structure

The diet of cormorants in Lake Grand-Lieu appeared to shift in response to the availability of food. The diet changed from generalist, with an even distribution of the seven major fish in the diet in the late 1980s early 1990s, to more specialised on breams (31% in 1998 and 57% in 1999) at the end of the 1990s (except for 2000). By contrast, Marion (1997a) concluded, for the period 1989–1994, that cormorants selected strongly for tench, despite its scarcity in stock. A similar active selection for a particular species was found by Dirksen et al. (1995), but the majority of studies found cormorant diet tended to reflect availability of food in the foraging area (Boldreghini et al. 1997; Carss & Marquiss 1997; Noordhuis et al. 1997; van Eerden & Zijlstra 1997). This shift in the diet of cormorants over the study period can be explained by two major changes within the lake. First, the lake has become more eutrophic (Marion & Brient 1998, 2000), which has resulted in a progressive shift in the fish community structure towards a dominance of cyprinids, notably breams and roach. This phenomenon is a common trophic shift found elsewhere in Europe (de Nie 1995) and more particularly in Swedish (e.g. Jeppesen et al. 2000) and in Dutch lakes (Dirksen et al. 1995; van Dobben 1995; Veldkamp 1995). Second, low water levels in the early 1990s left the adjacent peaty marsh grasslands exposed (May–June) whereas in the mid- to late-1990s the water level has been increased to flood these peat wetlands, and this is considered to have favoured cyprinids recruitment (Crivelli et al. 1995). As a consequence of both effects, cyprinids became more prevalent in the fishery and cormorants shifted their foraging behaviour to exploit the changes. Why the diet of cormorants in 2000 should become more specialised on pike and catfish is unclear, but it may be linked to strong 1998 and 1999 year classes which arose because of optimal spawning conditions in these years. These fish were more prevalent in the marginal areas of the lake where the cormorants also feed. Their relative paucity in the stock assessment was because it focused more on the main waters of the lake. In conclusion, the cormorant’s ability to adapt to prey availability and to environmental changes (e.g. Leopold et al. 1998) was corroborated by this study. The lack of competition for the main fish resources (eel) with the commercial fishery is important, but it is possible that there will be conflict with secondary species (pike, tench and zander).

References Adam G. & Elie P. (1993) Etude de la faune ichtyologique et de l’exploitation halieutique professionnelle du Lac de Grand-Lieu, Loire-Atlantique. Rapport CEMAGREF Bordeaux, 171 pp. (in French).

14 11/02/2003 13:52 Page 194

194

Bird, fish, habitat interactions

Boldreghini P., Santolini R. & Pandolfi M. (1997) Abundance and frequency of occurrence of preyfish in the diet of cormorants Phalacrocorax carbo in the Po River delta (Northern Italy) during the wintering period. Ekologia Polska 45, 191–196. Carpentier A. (1999) Approche méthodologique de la modélisation de l’impact du grand cormoran sur les peuplements piscicoles et la pêche professionnelle en eaux continentales: cas du lac de Grand-Lieu. Rapport de DEA, Université d’Orléans, MNHN, 45 pp. (in French). Carss D.N. & Marquiss M. (1997) The diet of cormorants Phalacrocorax carbo in Scottish freshwaters in relation to feeding habitats and fisheries. Ekologia Polska 35, 207–222. Carss D.N., Marquiss M. & Lauder A.W. (1997) Cormorant Phalacrocorax carbo carbo predation at a major trout fishery in Scotland. Supplemento alle Ricerche di Biologia della Selvaggina 26, 281–294. Copp G.H. & Penaz M. (1988) Ecology of fish spawning and nursery zones in the flood plain, using a new sampling approach. Hydrobiologia 169, 209–224. Craven S.R. & Lev E. (1987) Double-crested cormorants in the Apostle Island, Wisconsin, USA: Population trends, food habits, and fishery depredations. Colonial Waterbirds 19, 64–71. Crivelli A.J., Grillas P., Jerrentrup H. & Nazirides T. (1995) Effects on fisheries and waterbirds of raising water levels at Kerkini reservoir, a Ramsar site in Northern Greece. Environmental Management 19, 431–443. de Crespin V., Chessel S. & Dolédec D. (2000) Biplot presentation of diet composition data: an alternative for fish stomach contents analysis. Journal of Fish Biology 56, 961–973. de Nie H. (1995) Changes in the inland fish populations in Europe in relation to the increase of the cormorant Phalacrocorax carbo sinensis. Ardea 83, 115–122. Dirksen S., Boudewijn T.J., Noordhuis R. & Marteijn E.C.L. (1995) Cormorants Phalacrocorax carbo sinensis in shallow eutrophic freshwater lakes: prey choice and fish consumption in the non-breeding period and effects of large-scale fish removal. Ardea 83, 167–184. Feltham M.J., Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999) Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, contract number VC 0106. London: MAFF, 207 pp. Feunteun E. & Marion L. (1994) Assessment of Grey Heron predation on fish communities: the case of the largest European colony. Hydrobiologia 279–280, 327–344. Hald-Mortensen P. (1997) Does cormorant food tell more about fish than cormorants? Supplemento alle Ricerche di Biologia della Selvaggina 26, 295–311. Huatagalung R.A., Lim P., Belaud A. & Lagarigue T. (1997). Effets globaux d’une agglomération sur la typologie ichtyenne d’un fleuve: cas de la Garonne à Toulouse (France). Annales de Limnologie 33, 263–279 (in French). Jeppesen E., Jensen J.P., Sondergaard M., Lauridsen T. & Landkildehus F. (2000) Trophic structure, species richness and biodiversity in Danish lakes: changes along a phosphorus gradient. Freshwater biology 45, 201–218. Keller T., Vordermeier T., Von Lukowicz M. & Klein M. (1997) The impact of cormorants on the fish stocks of several Bavarian water bodies with special emphasis on the ecological and economical aspects. Supplemento alle Ricerche di Biologia della Selvaggina 26, 295–311. Leopold M.F., Van Damme C.J.G. & Van der Veer H.W. (1998) Diet of cormorants and the impact of cormorant predation on juvenile flatfish in the Dutch Wadden Sea. Journal of Sea Research 40, 93–107. Marion L. (1997a) Comparison between the diet of breeding cormorants Phalacrocorax carbo sinensis, captures by fisheries and available fish species: the case of the largest inland colony in France, at the Lake of Grand-Lieu. Supplemento alle Ricerche di Biologia della Selvaggina 26, 313–322. Marion L. (1997b) Les populations de Hérons cendrés en Europe et leur impact sur l’activité piscicole. In P. Clergeau (ed.) Oiseaux à risques en ville et en campagne. Paris: INRA Collection, pp. 101–132 (in French).

14 11/02/2003 13:52 Page 195

Interaction between cormorants and fisheries

195

Marion L. (1997c) Le grand coromron en Europe: dynamique des populations et impacts. In P. Clergeau (ed.) Oireaux à risques en ville et en campagne. Paris: INRA collection, pp. 133–178. (in French). Marion L. & Brient L. (1998) Wetland effects on water quality: input–output studies of suspended particulate matter, nitrogen (N) and phosphorus (P) in Grand-Lieu, a natural plain lake. Hydrobiologia 373/374, 217–235. Marion L. & Brient L. (2000) Effect of wetlands on water quality of rivers: the case of the important natural French plain lake, Grand-Lieu. Verhandlungen International Vereinigung Limnologie 27, 368–371. Marquiss M. & Carss D.N. (1994) Avian Piscivores: Basis for Policy. National Rivers Authority R&D Report 461/8/N&Y. Bristol: National Rivers Authority, 104 pp. Marteijn E.C.L. & Dirksen S. (1991) Cormorants Phalacrocorax carbo sinensis feeding in shallow eutrophic freshwater lakes in The Netherlands in the non-breeding period: prey choice and fish consumption. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 135–155. Nelva A., Persat H. & Chessel D. (1979) Une nouvelle méthode d’étude des peuplements ichtyologiques dans les grands cours d’eau par échantillonnage ponctuel d’abondance. Comptes Rendus de l’Académie des Sciences 289(D), 1295–1298 (in French). Noordhuis R., Marteijn E.C.L., Noordhuis R., Dirksen S. & Boudewijn T.J. (1997) The trophic role of cormorants Phalacrocorax carbo in freshwater ecosystems in the Netherlands during the nonbreeding period. Ekologia Polska 45, 249–262. Randall R.G., Minns C.K., Cairns V.W. & Moore J.E. (1996) The relationship between an index of fish production and submerged macrophytes and other habitat features at three littoral areas in the Great Lakes. Canadian Journal of Fish and Aquatic Sciences 53 (Suppl. 1), 35–44. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report of the Directorate of Fisheries Research, Lowestoft, 130 pp. Suter W. (1991) Food and feeding of cormorants Phalacrocorax carbo wintering in Switzerland. In M.R. van Eerden & M. Zijlstra (eds) Proceedings workshop 1989 on cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat, pp. 156–165. Thioulouse J., Chessel D., Dolédec S. & Olivier J.-M. (1997) ADE-4: a multivariate analysis and graphical display software. Statistic Computer 7, 75–83. Van Dam C. (1997) Cormorants and commercial fisheries in The Netherlands. Supplemento alle Ricerche di Biologia della Selvaggina 26, 333–341. Van Dobben W.H. (1995) The food of the cormorant Phalacrocorax carbo sinensis: old and new research compared. Ardea 83, 139–142. Van Eerden M. & Zijlstra M. (1997) An overview of the species composition in the diet of Dutch cormorants with reference to the possible impact on fisheries. Ekologia Polska 45, 223–232. Veldkamp R. (1995) The use of chewing pads for estimating the consumption of cyprinids by cormorants Phalacrocorax carbo. Ardea 83, 135–138.

15 11/02/2003 13:53 Page 196

Chapter 15

Fish predation by great cormorants, Phalacrocorax carbo carboides, in the Gippsland Lakes, south-eastern Australia P.C. COUTIN* Marine and Freshwater Resources Institute, Queenscliff, Victoria, Australia

J. RESIDE Wildlife Unlimited, Bairnsdale, Victoria, Australia

Abstract Roost counts and aerial surveys showed that the population of great cormorants, Phalacrocorax carbo carboides Gould, around the Gippsland Lakes increased 10-fold between 1987 and 1992, but then halved to 3470 in 1998. Similar quantities, sizes and species of prey were found from analyses of otoliths in regurgitated pellets and stomach contents of captured wild birds. The predominant prey was black bream, Acanthopagrus butcheri Munro (Sparidae), which occurred in 87% of the pellets and 86% of stomach contents. Based on two estimation methods, P. carbo consumed between 261 and 531 t of prey in 1998, exceeding the commercial and recreational catch. Incorporation of bird depredation on black bream stocks would improve stock assessment models because it is so high, selective and effects pre-recruit mortality. Management strategies for commercial and recreational fisheries, aquaculture and stock enhancement need to take cormorant population levels into account because stock abundance and catches are likely to be affected by predation. Keywords: Acanthopagrus butcheri, Australia, cormorants, diet, management, otoliths.

15.1 Introduction In Australia, great cormorants, Phalacrocorax carbo carboides Gould inhabit permanent coastal and inland waters all year round, but ephemeral wetlands and temporary lakes are used opportunistically (Pollard 1971; Barlow & Bock 1984; Baxter 1985; Braithwaite et al. 1985). Large aggregations breed along the Murray River where colonies range in size from less than 10 pairs to many thousands of birds. Breeding also takes place around large inland lakes in central Australia, such as Lake Eyre and the Menindee Lakes, after flooding, and this is followed by dispersal as the water recedes (Llewellyn 1983). *Correspondence: Patrick Coutin, Marine and Freshwater Resources Institute, PO Box 114, Queenscliff, Victoria 3225, Australia (email: [email protected]).

15 11/02/2003 13:53 Page 197

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

197

Periodically, large numbers of great cormorants inhabit the Gippsland Lakes, a large brackish system in south-eastern Australia. Nocturnal roosts are usually close to open water in large gum trees (Eucalyptus sp.) or amongst inundated thickets of swamp paperbark, Melaleuca ericifolia Smith, and these have been used consistently for several decades (D. Young, personal communication). There is very little breeding by P. carbo carboides in the Gippsland Lakes as virtually the whole population consists of immature birds, so increases in the population are mostly due to immigration from other locations. There is some evidence of this from an influx of up to 3000 P. carbo carboides into the Gippsland Lakes between August and November in 1954 (McNally 1957). A similar event occurred in 1992, when the population of P. carbo carboides reached 7801 birds to become the most abundant of the five cormorant species that inhabit the Gippsland Lakes (Reside & Coutin 2001). The other four species are the little black cormorant, P. sulcirostris Brand, the little pied cormorant, P. melanoleucos Viellot, the large pied cormorant, P. varius Gmelin and the black-faced shag P. fuscescens Viellot. This increase in the cormorant population has led to concerns about their effect on the fisheries and fish stocks, particularly black bream, Acanthopagrus butcheri. Although previous studies of P. carbo carboides in the Gippsland Lakes found that their diet comprised a wide variety of prey consisting of 33 estuarine species (Mack 1941; McNally 1957), fishermen have recently observed large flocks feeding daily almost entirely on schools of black bream on their fishing grounds. Representatives of both commercial and recreational fishing sectors have called for a resumption of culling of cormorants to preserve the fish stocks in the Gippsland Lakes, which support the most important black bream fishery in Victoria (Coutin 2000). However, culling would be inconsistent with current wildlife policies as cormorants are protected and the Gippsland Lakes are listed as internationally significant under the Ramsar Agreement, the Japan Australia Migratory Bird Agreement (JAMBA), and the China Australia Migratory Bird Agreement (CAMBA). To provide advice for both fisheries and wildlife management, an assessment of the level of fish predation from the Gippsland Lakes by P. carbo carboides was carried out. This chapter describes the assessment procedures and examines the degree of interaction between cormorants and the stocks of black bream.

15.2 Methods 15.2.1

Study area

The Gippsland Lakes extend over an area of 400 km2 and are separated from Bass Strait by vegetated sand dunes. The interconnected series of lakes, lagoons and wetlands are brackish and there are tidal flows through a narrow, artificial channel that was opened in 1889 and is kept open by dredging. The major components of the system where cormorant roosts are located are Lake Wellington, Lake Victoria and Lake King (Fig. 15.1). The lakes are fed by several rivers, the largest of which are the LaTrobe and Avon that flow into Lake Wellington, and the Mitchell, Nicholson and Tambo that flow into Lake King.

15 11/02/2003 13:53 Page 198

198

Bird, fish, habitat interactions

Nicholson River Jones Bay

Gippsland Lakes

Tambo River Salt Creek Roost Reeves Channel North Boxes Creek Arm

Mitchell River Newlands Arm

Lake King

Entrance

Lake Victoria Blond Bay Avon River Betsys Lagoon

Bass Strait

Lake Wellington

Latrobe River

Tucker Swamp

N 0

5

10 15 km

Figure 15.1 Map of the Gippsland Lakes showing the location of cormorant roosts

15.2.2

Cormorant population estimation

To estimate the size of the P. carbo carboides population inhabiting the Gippsland Lakes during 1998, their daily flocking behaviour and seasonal migrations were taken into account. Monthly counts of birds at their roosts were made using binoculars after 5 p.m. when all birds had returned from their feeding grounds. This was supported by summer aerial surveys after the post-breeding dispersal from inland areas when the number of cormorants at the Gippsland Lakes is highest, and by winter aerial surveys after the seasonal emigration period when some birds return inland to breed. During the aerial surveys, birds were counted from aircraft on calm days after 10 a.m. when the birds had left their roosts and were actively feeding on the lakes or were resting on sandbanks after feeding. The 4-h flights took place on 2 and 3 March 1998 using a small helicopter and the 13 and 14 August 1998 using a fixed-wing aeroplane. The aircraft flew at an altitude of 80–150 m at a cruising speed of between 40 and 70 knots. The flights followed a path 100 m from the shoreline over a distance of 300 km between Lakes Entrance and the mouth of the River LaTrobe in Lake Wellington. Two observers counted, photographed and recorded the number of birds feeding, flying or resting. All large flocks were photographed and the estimates of birds were verified from enlarged prints. In September 1997, based on information from local residents, all the major roosts that had been used in the past by P. carbo carboides around the Gippsland Lakes were

15 11/02/2003 13:53 Page 199

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

199

located. During 1998, only the Salt Creek roost was being used permanently. Counts were made in eight months between September 1997 and September 1998 at the roost which consisted of 80 trees, each holding between 12 and 49 birds. Shoreline counts of all aquatic water bodies in the Gippsland Lakes, including cormorant roosts were also made by the Birds Australia organisation between 1987 and 1991 (L. Turner, personal communication) and combined with the 1992 survey (Reside & Coutin 2001) to provide a time series to show the population trends.

15.2.3

Dietary study

Diet was determined from analyses of regurgitated pellets that were verified by examinations of stomach contents of birds captured in the wild. A total of 180 pellets were collected from beneath the roost trees at Salt Creek during summer (n  85) in February 1998 and during winter (n  95) in July 1998. Fish otoliths were removed from each pellet under a microscope and then identified, matched into pairs, counted and measured. To determine the species and size of fish from the otoliths, samples of the most common local fish species were collected. The total length (TL, mm) of fish and otolith length and weight measurements were recorded. Relationships between otolith length and fish weight (g) and total length were derived and used to estimate the lengths and weights of prey from the otoliths in the pellets (Reside & Coutin 2001). To verify the results of the pellet analysis, seven cormorants were captured at Salt Creek on 28 January 1998 using three cannon-nets made of heavy mesh-net, measuring 30 m by 8 m, with solid steel projectiles that were fired from four black-powder charged cannon. Some of the birds escaped from the net, some regurgitated their stomach contents and others died whilst being captured. Captured birds were each placed in a plastic bin and induced to regurgitate by administering an emetic liquid (10 ml of solution, 30% Epicac emetic and 70% water) through a tube inserted into the stomach (Ford et al. 1982). The birds were then immediately released and the prey from all the samples of stomach contents were identified, weighed, and measured.

15.2.4

Fish consumption estimation

The following equation was used to estimate the annual fish consumption: Number of fish eaten  N1  t  Fp/N2 Weight of fish eaten  N1  t  Wp/N2 where N1 is the mean number of birds counted each day in the aerial survey for each season; N2 is the number of pellets sampled in each season or number of birds captured; t the number of days in each season; Fp the number of otolith pairs for each species in pellets or number of fish regurgitated or in stomachs of captured birds; and Wp the weight of each species estimated from otoliths in pellets or weight of fish regurgitated or found in stomachs of captured birds.

15 11/02/2003 13:53 Page 200

200

Bird, fish, habitat interactions

The method assumed that: the pellets and birds collected were representative of the entire season and the population; each cormorant fed every day; one pellet was produced by each bird on each day at the roost; all fish consumed in a day were represented as otolith pairs in the pellets; all otoliths in the pellets were found and species were identified correctly; otolith decomposition had no effect on estimation of fish length and weight; birds captured at midday had completed feeding for the day; birds with full and empty stomachs were equally vulnerable to capture; all fish eaten on the day were still in the stomach of captured birds and were emitted; partially digested fish were the same size and weight as whole fish.

15.3 Results 15.3.1

Aerial surveys

The two aerial surveys conducted in summer produced very similar population estimates for P. carbo carboides with a mean of 3617 around the Gippsland Lakes. Phalacrocorax carbo carboides was the most abundant species and contributed 51% of the total cormorant population of 7091 birds in summer (Table 15.1). The two aerial surveys conducted in winter produced more variable population estimates of P. carbo carboides with a mean of 3322 around the Gippsland Lakes. The relative abundance of the different species of cormorants was similar in summer and winter of 1998. P. carbo carboides was the most abundant species and contributed 59% of the total cormorant population of 5654 birds in winter. Because there was little seasonal change in the population of P. carbo carboides, the annual estimate of 3470 birds for 1998 was based on the mean count from the two surveys.

15.3.2

Roost census

A lower population estimate of 2088 birds was obtained from roost counts compared to the aerial surveys (Fig. 15.2). During summer, there was a stable population with numbers fluctuating between 2240 and 2578 birds (median count  2409). During Table 15.1

Species

Aerial counts of cormorants on the Gippsland Lakes during 1998 Summer

Winter

1998

2/03/98 3/03/98

13/8/98 14/8/98

Mean %

Great cormorant, P. carbo carboides 3 581 Little black cormorant, P. sulcirostris 2 230 Little pied cormorant, P. melanoleucos 836 Large pied cormorant, P. varius 54

3 653 2 923 837 68

3 037 1 137 544 66

3 607 2 135 719 62

3 470 2 106 734 63

All species

7 481

4 784

6 523

6 373

6 701

54.5 33.0 11.5 1.0 100

15 11/02/2003 13:53 Page 201

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

201

winter, there was widespread flooding which caused the colony to disperse and the population at the roost decreased to 983 in June 1998 and 780 in July 1998. However, within a few weeks of the flood subsiding, the colony had reassembled and the highest count of 2880 birds was made in September 1998.

15.3.3

Trends in cormorant populations

The annual trends in the population sizes of the three species of cormorant were indicated by the maximum number of birds observed during the Birds Australia census, the 1992 aerial survey and this study (Fig. 15.3). Since 1987, the population of P. carbo carboides has increased more than threefold, and fluctuated during the 1990s. Following the first estimate in 1987, there was a decline in the P. carbo carboides population from 1000 to 200–300 birds in 1990 and 1991, when the other cormorant species were predominant and P. carbo carboides constituted only 14–20% of all cormorant species in the Gippsland Lakes. The large influx of P. carbo carboides to the

Roost census

Number of birds

4000

Aerial survey

3000 2000 1000

Sep-98

Aug-98

Jul-98

Jun-98

Apr-98

May-98

Mar-98

Feb-98

Jan-98

Dec-97

Nov-97

Oct-97

Sep-97

0

12 10

Little Pied Little Black

8 6 4

Great

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

1988

2 0 1987

Bird numbers (x 1000)

Figure 15.2 Monthly counts of great cormorants P. carbo carboides at the Salt Creek roost and during aerial surveys of the Gippsland Lakes between September 1997 and September 1998

Figure 15.3 Population estimates of the three species of cormorants on the Gippsland Lakes between 1987 and 1998

15 11/02/2003 13:53 Page 202

202

Bird, fish, habitat interactions

Gippsland Lakes in 1992 caused an increase in their population to 7801 birds. However, this large number of birds had partially dispersed by October 1997 and the roost counts showed that the population had dropped to 2578 birds. The roost counts and aerial surveys conducted in March 1998 showed that the population of between 2419 and 3563 birds was lower than in 1992.

15.3.4

Prey

80

Summer Winter

60

Annual

40

Others

Garfish

Luderick

Hardyhead

Yellow-eyed mullet

0

Anchovy

20

Black bream

Number of fish (%)

Black bream was found in 87% of the pellets and was the predominant prey of P. carbo carboides by occurrence, number and weight in summer and winter (Figs 15.4 and 15.5).

Figure 15.4 The relative proportion of prey species by number found in the pellets of P. carbo carboides in the Gippsland Lakes during summer and winter 1998 and averaged across seasons Summer Winter Annual

60 40

Others

Luderick

Hardyhead

Yellow-eyed mullet

0

Anchovy

20

Black bream

Weight of fish (%)

80

Figure 15.5 The relative proportion of prey species by weight found in the pellets of P. carbo carboides in the Gippsland Lakes during summer and winter 1998 and averaged across seasons

15 11/02/2003 13:53 Page 203

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

203

Ten other species were identified in the pellets: yellow-eyed mullet, Aldrichetta forsteri Val., anchovy, Engraulis australis White, luderick, Girella tricuspidata Quoy & Gaimard, river garfish, Hyporhamphus regularis Günther, tupong, Pseudaphritis urvillii Val., tailor, Pomatomus saltatrix L., skipjack trevally, Pseudocaranx wrighti Whitley, tommy rough, Arripis georgiana Val., small-mouthed hardyhead, Atherina microstoma Günther and flat-headed gudgeon, Philypnodon grandiceps Krefft. Three pellets (1.7%) contained only the exoskeletons of prawns and five (2.8%) were empty. The remains of invertebrate fish lice, Ceratothoa imbricatus Fabricus, were also found in pellets, but these were probably consumed while attached to fish.

15.3.5

Daily fish consumption and size of prey

The number of fish of all species found in the pellets ranged between 0 and 86 with a mean of 5.3 fish per pellet. The number of black bream found in the pellets ranged between 0 and 8 with a mean of 2.6 fish per pellet. The total weight of all species found in the pellets ranged between 0 and 1437 g with a mean of 419 g. The total weight of all the black bream found in the pellets ranged between 0 and 741 g per pellet with a mean of 266 g. The estimated size of prey found in pellets (Table 15.2) ranged from 8 to 305 mm TL and 0 to 379 g with means of 149 mm TL (95% CI  4 mm) and 79 g (95% CI  5 g). The size of black bream found in pellets ranged from 33–292 mm TL with a mean of 178 mm and fish weight ranged from 1–379 g with a mean of 102 g (Fig. 15.6).

15.3.6

Stomach flushing of captured cormorants

The contents of the pellets were consistent with the species and prey types that were recovered from captured wild birds. The size of all species of fish found in stomachs

Table 15.2

Size of fish prey estimated from otoliths found in pellets in 1998

Species

Season

Mean TL (mm)

Range TL (mm)

Mean weight (g)

Range weight (g)

Anchovy Black bream Black bream Black bream Garfish Gudgeon Hardyhead Luderick Tailor Tupong Yellow-eyed mullet All species

Winter 1998 Summer Winter Winter Winter Winter 1998 Winter Winter 1998 1998

88 178 186 170 186 46 33 181 106 121 191 149

49–118 33–292 33–292 53–288 33–292 13–82 10–152 8–278 60–175 85–166 50–305 8–305

3.5 102 113 92 113 1 0.9 179 15 20 100 79

0.6–8 1–379 1–378 2–364 6–208 0–5 0–33 3–304 4–36 6–48 1–329 0–379

15 11/02/2003 13:53 Page 204

204

Bird, fish, habitat interactions

50

Black Bream (Winter) Black Bream (Summer)

40

Frequency (n)

Mean TL = 17.8 cm Mean Wt = 102 gm 30

20

10

0 0

2

4

6

8

10

12 14 16 18 20 22 Total length (cm)

24 26

28 30

Figure 15.6 Size of black bream consumed by P. carbo carboides in the Gippsland Lakes in 1998 determined from pellets

ranged between 40 and 533 mm TL, with a mean of 196 mm. The weight of prey consumed ranged between 5 and 235 g with a mean of 100 g. Black bream was found in most of the stomach contents (86%) and was also the predominant prey by number (44%) and weight (48%) of fish in the stomachs. The size of the partially digested black bream ranged between 40 and 221 mm with a mean of 150 mm and the weight ranged between 14 and 192 g with a mean of 85 g. Luderick occurred in more than half of the birds in the sample (57%) and was also important by number (33%) and weight (46%). The size of luderick in stomachs ranged between 176 and 215 mm with a mean of 194 mm and the weight ranged between 78 and 172 g with a mean of 120 g. A large short-finned eel, Anguilla australis Richardson, that measured 532 mm TL and 234 g in weight was found in captured birds, but otoliths of this species were not found in the pellets.

15.3.7

Annual fish consumption

A large quantity of fish from the Gippsland Lakes were consumed by P. carbo carboides in 1998 and this was estimated from pellets to be 6.5 million fish weighing 531 t (Table 15.3). The majority of their prey consisted of black bream (3.3 million fish, 340 t), yellow-eyed mullet (0.8 million fish, 82 t), luderick (0.3 million fish, 55 t) and river garfish (0.3 million fish, 28 t). Although the methods based on stomach contents produced lower estimates of predation (3.3 million fish weighing 261 t for all species), the birds were still found to take a large quantity of black bream (1.5 million fish, 126 t), and luderick (1.0 million fish, 121 t).

15 11/02/2003 13:53 Page 205

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

205

Table 15.3 Fish consumption estimates of the P. carbo carboides population in the Gippsland Lakes based on pellets collected in 1998 Summer

Winter

1998

Species

Number of fish  1000

Weight of fish (t)

Number of fish  1000

Weight of fish (t)

Number of fish  1000

Weight of fish (t)

Anchovy Australian salmon Black bream Gudgeon Hardyhead Luderick Prawns River garfish Tailor Tupong Yellow-eyed mullet Others Total

0 0 1 755 0 0 39 31 0 0 0 450 78 2 376

0 0 198 0 0 7 2 0 0 0 61 9 278

957 6 1 544 185 300 268 13 249 128 38 345 83 4 129

3 1 142 0 0 48 1 28 2 1 21 5 253

957 6 3 299 185 300 307 44 249 128 38 795 161 6 505

3 1 340 0 0 55 3 28 2 1 82 14 531

15.4 Discussion 15.4.1

Cormorant population trends

Since 1987, there have been substantial shifts in the cormorant populations in the Gippsland Lakes with different trends in each species, which confirmed the observations of the local commercial fishermen, anglers and bird watchers. The large fluctuations in the populations of P. carbo carboides and P. sulcirostris are in contrast to the relatively low and stable numbers of P. melanoleucos and P. varius. Many studies in other countries have also shown that populations of cormorants can periodically change and disperse, but the causes and population trends are often uncertain (Debout et al. 1995; De Nie 1995; Ewins et al. 1995; van Eerden & Grehgersen 1995; Yesou 1995). In the Gippsland Lakes, it is still unclear whether the recent increase in P. carbo carboides reflects a long term population trend. The observations in 1998 indicated that there had only been a slight rise in the population since 1954 compared with the upper estimate of 3000 P. carbo carboides made by McNally (1957). However, the population in 1992 (Reside & Coutin 2001) was higher than the 5744 P. carbo carboides recorded by Braithwaite et al. (1985) from aerial surveys across eastern Australia after the droughts in the mid 1980s. More recent aerial surveys covering 332 360 km2 of eastern Australia indicated that the great cormorant population reached a peak of just less than 20 000 birds during the period 1991–1993, but had declined to between 659 and 2210 birds during the period 1996–1999 (Kingsford et al. 2000). Llewellyn (1983) suggested that the cormorant population in New South Wales was determined to a great extent by the water levels in Lake Eyre in Central Australia and

15 11/02/2003 13:53 Page 206

206

Bird, fish, habitat interactions

the flood and drought cycles. If this is also the case in Victoria, as indicated for waterbirds by Briggs et al. (1997), it seems likely that the increase in the Gippsland Lakes population in 1992 was associated with a coastal redistribution of P. carbo carboides from inland areas following a period of low rainfall in the region in the early 1990s. Since 1998 intensive breeding by P. carbo carboides has been observed around Lake Eyre after heavy rainfall in the catchment and water levels in 2000 were higher than in the last flood in 1974 (Llewellyn 1983). If the fledglings survive and then migrate eastwards, it seems likely that another influx of P. carbo carboides will occur in the Gippsland Lakes, particularly if there is another drought.

15.4.2

Validation of pellet analyses

There is some doubt about the validity of estimates based on pellets, although many similar studies have assumed that the pairs of otoliths from fish consumed each day are all regurgitated once per day as a single pellet at the nocturnal roost (e.g. Craven & Lev 1987). Some dietary studies on other cormorant species suggested that some fish species were under-represented numerically by otoliths recovered from pellets (e.g. Brugger 1993; Russell et al. 1995). However, sub-adult great cormorants (P. carbo sinensis Blumenbach) have been shown to produce a single pellet per day containing only the undigested remains from the previous day (Trauttmansdorff & Wassermann 1995) and it was considered that their pellets could be used to estimate the size and weight of prey consumed (Jobling & Breiby 1986; Zijlstra & van Eerden 1995). In this study, similar sizes and species of prey were obtained from analyses of pellets and stomachs of captured birds. The weight of fish in the small sample of stomachs was well within the range estimated from pellets. However, the mean total weight of all species of fish in each pellet (419 g) was more than twice the mean weight of fish found in the stomach contents of captured birds (206 g). The maximum total weights of all species of fish in each pellet were also higher (1023 g) compared with the stomach contents (402 g). The weights of stomach contents were lower because the prey was partially digested and some of the cormorants may have regurgitated their stomach contents during capture. Assuming that fish regurgitated during capture had been consumed by the captured birds and not by others in the flock, the mean weight of their stomach contents would have been higher (312 g) and closer to the mean weight estimated from pellet analyses. A further allowance for the partial digestion of the prey brings the results of the two methods even closer together. The estimates from pellets were also consistent with the daily energy requirements of P. carbo carboides and indicated a daily food intake for P. carbo carboides of 18.6% for a body weight of 2.25 kg. This is similar to food consumption estimates of 16.8% of body weight (1.22 kg) for P. capensis Sparrman (Laugksch & Duffy 1984), and 14.5% for P. auritus Brisson (Madenjian & Gabrey 1995). The estimates of average daily consumption were lower than the 31% of body weight derived from a captive feeding study of P. atriceps King from a colder climate (Casaux et al. 1995) and estimates for breeding cormorants raising chicks (Furness 1978). The stomach contents of birds captured in the wild and their daily energy requirements support the assumption

15 11/02/2003 13:54 Page 207

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

207

that P. carbo carboides produced one pellet per day containing the prey consumed on that day. This indicates that the estimates of cormorant predation from pellets are reasonably accurate and are of the right order of magnitude.

15.4.3

Prey

Phalacrocorax carbo carboides fed on similar prey species in 1998 as in 1954 (McNally 1957), but larger quantities of anchovy, river garfish and gobies were taken in 1939 (Mack 1941). Prey consumption from pellets collected during February and July 1998 was estimated, but monthly sampling would have improved the estimates and allowed for seasonal variation in the diet. Seasonal and annual changes in the abundance or distribution of fish in the vicinity of the roost are likely to influence prey selection by great cormorants. Large size, spines or benthic behaviour may also influence prey selection and partly account for low predation of certain species. Some common species of fish in the Gippsland Lakes were not found in the diet such as carp, Cyprinus carpio L, leather jackets, Meuschenia freycineti Quoy and Gairnard and sand flathead, Platycephalus bassensis Cuvier. Although the range of prey of P. carbo carboides in the Gippsland Lakes included 36 estuarine species of fish, the predominant prey in 1998 was small black bream both in summer and winter. Recruitment of black bream is known to be episodic (Morison et al. 1998), and following periods of high recruitment after successful spawning between 1994 and 1996 (Coutin 2000), it seems likely that P. carbo carboides was feeding selectively and targeting the schools of small black bream that were abundant at this time. It is quite possible that the relative abundance of black bream and anchovy in the Gippsland Lakes may have changed since the 1930s and affected prey selection by P. carbo carboides. This could account for the differences with the earlier study by Mack (1941).

15.4.5

Impact of Phalacrocorax carbo carboides depredation on fish stocks

The aerial surveys and roost counts showed that the population of P. carbo carboides in the Gippsland Lakes fluctuated considerably from year to year. When their population is high, these birds are capable of consuming large quantities of prey. Based on pellets, the annual fish predation by P. carbo carboides in the Gippsland Lakes in 1998 was 531 t for all species of prey of which 340 t was black bream. This was more than twice the amount of these fish species that was taken by the commercial fishery (229 t of all prey species and 145 t of black bream) from the Gippsland Lakes in 1997/1998 and exceeded the combined recreational and commercial catch (Coutin 2000). The effects of fish predation by P. carbo carboides on the black bream stocks are likely to have been far greater than the fishing mortality in 1992 when the population was very high. Because the birds feed on small, pre-recruit black bream below the legal minimum length of 26 cm TL, the number taken by predation would have been far greater than the

15 11/02/2003 13:54 Page 208

208

Bird, fish, habitat interactions

number of fish retained by fishermen. This would have caused unusually high levels of natural mortality to the black bream in the Gippsland Lakes and may have reduced the abundance of the younger age classes and affected the age structure of the stock. Incorporation of the impact of cormorant predation on fish stocks into fisheries models (see Dekker et al., Chapter 1) would improve stock assessments as it may be the most important factor affecting stock abundance in some years and could be predicted with further monitoring and modelling. Forecasts of black bream year class abundance based on surveys of juvenile stocks may need to be markedly reduced when there are high cormorant populations. This type of interaction between fish stock dynamics, fisheries and wildlife species is an important consideration (Furness 1982; Crawford & Dyer 1995; Waneless et al. 1998), and multispecies or ecosystem models are needed to adequately describe the complexity of this brackish system. There are implications for fisheries and wildlife management particularly when there is an influx of cormorants into the Gippsland Lakes. For wildlife conservation and fisheries management, the fish resources of the Gippsland Lakes need to be allocated appropriately in management strategies. Predation by cormorants is likely to be an important consideration for aquaculture development and stock enhancement in the vicinity of P. carbo carboides roosts because expected yields and returns may be reduced by this large aquatic predator. Fisheries management plans should allow for sudden increases in predation of pre-recruits that may reduce the biomass of black bream and impact commercial and recreational catches.

Acknowledgements We are indebted to Ken Stephenson for his permission to access the Salt Creek roost and to the commercial fishermen who provided fish samples. We are grateful to Lynn Turner for permission to use the 1987–1992 Birds Australia survey data and we would like to thank the Victorian Wader Study Group for their assistance. This research was funded by the Department of Natural Resources and Environment.

References Barlow C.G. & Bock K. (1984) Predation of fish in farm dams by cormorants, Phalacrocorax spp. Australian Wildlife Research 11, 559–566. Baxter A.F. (1985) An Analysis of the Stomach Contents of Cormorants Collected from Trout Waters in Victoria between 1978 and 1982. Arthur Rylah Institute for Environmental Research, Department of Conservation, Forests and Lands Technical Report 13, 26 pp. Braithwaite L.W., Maher M.T. & Parker B.S. (1985) Aerial surveys of wetland bird fauna in eastern Australia, October 1984. Division of Wildlife and Rangelands Research, CSIRO Technical Report 23, 92 pp. Briggs S.V., Thornton S.A. & Lawler W.G. (1997) Relationships between hydrological control of river red gum wetlands and waterbird breeding. EMU 97, 31–42. Brugger K.E. (1993) Digestibility of three fish species by double-crested cormorants. The Condor 95, 25–32.

15 11/02/2003 13:54 Page 209

Fish predation by Phalacrocorax carbo in the Gippsland Lakes

209

Casaux R.J., Favero M., Barrera-Oro E.R. & Silva P. (1995) Feeding trials on an imperial cormorant Phalacrocorax atriceps: Preliminary results on fish intake and otolith digestion. Marine Ornithology 23, 101–106. Coutin P.C. (2000) Black bream – 1997. Compiled by the Bay & Inlet Fisheries and Stock Assessment Group. Fisheries Victoria Report 18, 52 pp. Craven S.R. & Lev E. (1987) Double-crested cormorants in the Apostle Islands, Wisconsin, USA: Population trends, food habits and fishery depredation. Colonial Waterbirds 10, 64–71. Crawford R.J.M. & Dyer B.M. (1995) Responses by four seabird species to a fluctuating availability of Cape anchovy Engraulis capensis off South Africa. Ibis 137, 329–339. De Nie H. (1995) Changes in the inland fish populations in Europe in relation to the increase of the cormorant. Ardea 83, 115–122. Debout G., Rov N. & Sellers R.M. (1995) Status and population development of cormorants Phalacrocorax carbo carbo breeding on the Atlantic coast of Europe. Ardea 83, 47–59. Ewins P.J., Weseloh D.V. & Blokpoel H. (1995) Within-season variation in nest numbers of doublecrested cormorants (Phalacrocorax auritus) on the Great Lakes: Implications for censusing. Colonial Waterbirds 18, 179–192. Ford H.A., Forde N. & Harrington S. (1982) Non-destructive methods to determine the diets of birds. Corella 6, 6–10. Furness R.W. (1978) Energy requirements of seabird communities: a bio-energenetics model. Journal of Animal Ecology 47, 39–53. Furness R.W. (1982) Competition between fisheries and seabird communities. Advances in Marine Biology 20, 225–307. Jobling M. & Breiby A. (1986) The use and abuse of fish otoliths in studies of feeding habits of marine piscivores. Sarsia 71, 265–274. Kingsford R.T., Porter, J.L., Ahern A.D. & Davis, S.T. (2000) Aerial Surveys of Wetland Birds in Eastern Australia – October 1996–1999. National Parks and Wildlife Service. N.S.W. Occasional Paper No.31, 70 pp. Laugksch R.C. & Duffy D.C. (1984) Energetics equations and food consumption of seabirds in two marine upwellings areas: comparisons and the need for standardization. South African Journal of Marine Science 2, 145–148. Llewellyn L.C. (1983) Movements of cormorants in south-eastern Australia and the influence of floods on breeding. Australian Wildlife Research 10, 149–167. Mack G. (1941) Cormorants of the Gippsland Lakes fishery. Memoirs of the Natural History Museum of Victoria 12, 95–117. Madenijan C.P. & Gabrey S.W. (1995) Waterbird predation on fish in Western Lake Erie: a bioenergetics model application. The Condor 97, 141–153. McNally J. (1957) The feeding habits of cormorants in Victoria. Victorian Fish and Game Department Fauna 6, 36 pp. Morison A.K., Coutin P.C. & Robertson S.G. (1998) Age determination of black bream, Acanthopagrus butcheri (Sparidae), from the Gippsland Lakes of south-eastern Australia indicates slow growth and episodic recruitment. Marine and Freshwater Research 49, 491–498. Pollard D.A. (1971) Faunistic and environmental studies on Lake Modewarre, a slightly saline athalassic lake in southwestern Victoria. Bulletin of the Australian Society of Limnology 4, 25–42. Reside J. & Coutin P.C. (2001) Preliminary Estimates of the Population, Diet and Fish Consumption of the Great Cormorant, (Phalacrocorax carbo carboides, (Gould 1838)) in the Gippsland Lakes, Victoria during 1998. Marine and Freshwater Resources Institute Report No. 27, 53 pp. Russell A.F., Wanless S. & Harris M.P. (1995) Factors affecting the production of pellets by shags Phalacrocorax aristotelis. Seabird 17, 44–49. Trauttmansdorff J. & Wassermann G. (1995) Number of pellets produced by immature cormorants Phalacrocorax carbo sinensis. Ardea 83, 133–134.

15 11/02/2003 13:54 Page 210

210

Bird, fish, habitat interactions

Van Eerden M.R. & Grehgersen J. (1995) Long-term changes in the northwest European population of cormorants, Phalacrocorax carbo sinensis. Ardea 83, 61–79. Waneless S., Harris M.P. & Greenstreet S.P.R. (1998) Summer sandeel consumption by seabirds breeding in the Firth of Forth, south-east Scotland. ICES Journal of Marine Sciences 55, 1141–1151. Yesou P. (1995) Individual migration strategies in cormorants Phalacrocorax carbo passing through or wintering in western France. Ardea 83, 267–274. Zijlstra M. & van Eerden M.R. (1995) Pellet production and the use of otoliths in determining the diet of cormorants Phalacrocorax carbo sinensis: trials with captive birds. Ardea 83, 123–131.

16 11/02/2003 14:00 Page 211

Chapter 16

Changes in the piscivorous bird community at a Polish submontane reservoir between 1990 and 2000 in relation to water level R. GWIAZDA* Karol Starmach Institute of Freshwater Biology, Polish Academy of Sciences, Kraków, Poland

Abstract Changes in the community structure and food consumption of fish-eating birds at a submontane reservoir in Poland were estimated. Peak bird numbers were recorded in autumn each year. In the early 1990s, great crested grebe, Podiceps cristatus (L.), predominated, while great crested grebe, cormorant, Phalacrocorax carbo (L.), and grey heron, Ardea cinerea L, were dominant in the late 1990s. Increases in cormorant, plus herring/Caspian gulls, were due to population trends of these species while the greater numbers of grey heron was caused by low water level in autumn and development of a breeding colony. The decline in number of great crested grebes over the 1990s was probably due to disturbance by fisherman at night. The highest density of fish-eating birds was observed in the shallow bay. A relationship between the number of cormorants in backwaters of the reservoir and the water level was found. The three main species of fish-eating birds ate about 10.5 kg ha1 of fish in 1990 and 13.7 kg ha1 in 1999. About 54% of the fish was eaten between August and October in 1990, rising to 84% for the same months in 1999. The change in fish-eating bird predation was poorly related to fish density but more to habitat (occurrence of resting and refuge sites). Keywords: consumption, density, environment, food, piscivorous birds, structure.

16.1 Introduction Predation pressure from fish-eating birds on fish stocks depends upon the number of birds and bird community structure. The water bird assemblage at Dobczyce Reservoir (southern Poland) has a high proportion of fish-eating species but the community structure has changed between 1990 and 2000 as a result of environmental changes of the reservoir (Gwiazda 1989, 1990, 1996). The study detailed in this chapter examines the changes in predation pressure and food selectivity caused by the shift in the fish-eating bird community structure in different parts of Dobczyce Reservoir.

*Correspondence: R. Gwiazda, Karol Starmach Institute of Freshwater Biology, Polish Academy of Sciences, ul. Slawkowska 17, 31-016 Kraków, Poland (email: [email protected]).

16 11/02/2003 14:00 Page 212

212

Bird, fish, habitat interactions

Figure 16.1 Location of different zones in Dobczyce reservoir: (1) deep zone; (2) main basin; (3) shallow bay

16.2 Study area The study was carried out on the Dobczyce Reservoir (49°52N, 20°02E) built in 1986 to supply drinking water (Fig. 16.1). It is a submontane, eutrophic reservoir (normal water level 269.9 m above sea level) in southern Poland (25 km south of Cracow) with an area of about 970 ha and mean depth of 11 m (max. 30 m) and a capacity of 99.2  106 m3. The littoral zone is steep, with shallow water 2 m forming only about 10% of the total area. Land-use is meadow and forest. Macrophytes (Phragmites australis, Typha latifolia, Polygonum amphibium) are restricted to the edge of a shallow bay in the northern part of reservoir. In the 1980s a relatively large biomass of perch, Perca fluviatilis (L.), and rheophilic cyprinids was present. In the 1990s limnophilic cyprinids (mainly roach, Rutilus rutilus (L.), and bream, Abramis brama (L.)) were dominant (Godlewska & Jelonek 2000). The reservoir is usually ice covered in January and February. The water level was relatively stable in 1990 and 1991 (annual fluctuation 0.6–0.9 m) and unstable in 1999 and 2000 (2.4–3.9 m, with decreasing water level in autumn). Fishing effort and fish farming activities were low in the early 1990s but have become more intensive since 1993 when new methods and gears were introduced.

16.3 Materials and methods Only birds that consume mainly fish were studied. For example, black-headed gull, Larus ridibundus L., common gull Larus canus L., or black tern, Chlidonias niger (L.)

16 11/02/2003 14:00 Page 213

Piscivorous bird community of a Polish reservoir

213

were not included because their diet comprises mainly insects (Cramp & Simmons 1983, 1985; Gwiazda 2000). Counting was carried out from the shore in 1990 and 1991 (stable water level period), and again in 1999 and 2000 (decreasing of water level, nocturnal fishing in the shallow bay). Counting was carried out once per month, between 09.00 and 15.00 h, from March to December (no counting in July 2000), using 10  50 binoculars and 40  64 telescopes. An additional 29 days of counting in the backwaters of the reservoir were undertaken from August to October in 1998 and 1999. Data of water level were provided by the Regional Board of Water Management in Cracow. The Dobczyce Reservoir was divided in three zones: (i) deep zone (narrow littoral area, steep banks except on the west shore); (ii) the main basin (narrow littoral, mainly steep banks, large changes of area in backwaters); and (iii) the shallow bay (wide littoral with macrophytes, flat banks). The area of these parts was determined from data provided by the Regional Board of Water Management in Cracow. A Friedman test was used to compare the number of piscivorous bird species between 1990, 1991, 1999 and 2000. Data for herring gull, Larus argentatus Pont., and Caspian gull, Larus cachinnas Pall., were not presented separately because identification of those species was not always possible. Spearman rank correlation was used to determine the relationship between the water level and the number of fish-eating birds in backwaters in autumn (August–September) 1998 and 1999. The program Statistica 5.0 was used for statistical analyses. The quantity of fish consumed annually by the three dominant bird species (great crested grebe, Podiceps cristatus (L.), cormorant, Phalacrocorax carbo (L.), and grey heron, Ardea cinerea L., was calculated as the product of the daily food demand, the number of birds present in each month and the number of days in a month. Daily food consumption rates of these species was based on literature data (Bauer & Glutz von Blotzheim 1966; Cramp & Simmons 1977; EIFAC 1988) as: great crested grebe – 200 g, cormorant – 450 g, grey heron – 330 g.

16.4 Results Twenty-one species of fish-eating birds were observed: black-necked diver, Gavia arctica (L.), little grebe, Tachybaptus ruficollis (Pall.), great crested grebe, red-necked grebe, Podiceps grisegena Bodd., horned grebe, Podiceps auritus (L.), black-necked grebe, Podiceps nigricollis Brehm, cormorant, grey heron, great white egret, Egretta alba (L.), night heron, Nycticorax nycticorax (L.), red-breasted merganser, Mergus serrator L., goosander, Mergus merganser L., smew, Mergus albellus L., osprey, Pandion haliaetus (L)., herring gull, Caspian gull, lesser black-backed gull, Larus fuscus L., common tern, Sterna hirundo L., sandwich tern, Sterna sandvicensis Lath., whiskered tern, Chlidonias hybridus (Pall.), kingfisher, Alcedo atthis (L.). The highest relative abundance and maximum number were recorded for great crested grebe (Table 16.1). Fish-eating birds constituted 20–28% of water bird community in 1990, 1991, 1999 and 2000. The peak number of fish-eating birds (236–668) was recorded in late summer

16 11/02/2003 14:00 Page 214

214

Bird, fish, habitat interactions Table 16.1 Relative abundance, maximum number and frequency of visits of fisheating birds on the Dobczyce reservoir in 1990, 1991, 1999, and 2000 Species

Relative abundance

Maximum number

Frequency

Podiceps cristatus Phalacrocorax carbo Ardea cinerea Larus argentatus and L. cachinnans Sterna hirundo Mergus merganser Tachybaptus ruficollis Alcedo atthis Gavia arctica Podiceps nigricollis Mergus serrator Egretta alba Sterna hybridus Podiceps grisegena Pandion haliaetus Nycticorax nycticorax Podiceps auritus Sterna sandvicensis Mergus albellus Larus fuscus

94.7 28.7 18.8 13.6 2.2 1.1 0.5 0.5 0.3 0.3 0.3 0.2 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1

350 224 85 193 25 21 5 6 7 3 10 8 5 1 2 2 2 1 1 1

39 24 38 17 14 5 11 13 4 6 2 1 1 2 2 1 1 1 1 1

(August, September) each year (Fig. 16.2). Only three species were classified as nesting: great crested grebe (breeding every year), common tern (breeding irregularly) and grey heron (breeding since 1998). In the early 1990s, great crested grebe dominated, constituting from 70 to 97% of the fish-eating bird community. In the late 1990s, great crested grebe, grey heron and cormorant dominated, constituting 54–93% of the fisheating bird community in 1999 (Fig. 16.2). Mergansers were observed only five times in small numbers. The numbers of grey heron, cormorant and herring/Caspian gull in particular years were significantly different (P  0.003, P  0.001, and P  0.001, n  10, respectively) and was higher in 1999 and 2000 than 1990 and 1991. The median number of grey herons was four individuals (Q1  3.5, Q3  6.5, n  10) in 1990 and 21 (Q1  6, Q3  23.8, n  10) in 1999. Cormorants and herring/Caspian gulls were rarely observed in 1990 but the median number of cormorants and herring/Caspian gulls were 3.5 (Q1  2.5, Q3  40.6, n  10) and 3.5 (Q1  0.5, Q3  14, n  10) respectively in 1999. The number of great crested grebe in particular years was different (P  0.02, n  10), and highest in 1990 (median: 152.5, Q1  71.2, Q3  206, n  10). The highest density of fish-eating birds was observed in the shallow bay. In 1990, the median was 3.8 ind. 10 ha1 (range: 1.12–6.64, n  10), primarily because the breeding population of great crested grebe (28–29 pairs) was concentrated in this area, and in 1999,

16 11/02/2003 14:00 Page 215

Piscivorous bird community of a Polish reservoir

215

Figure 16.2 Annual change in numbers and structure of fish-eating birds at Dobczyce reservoir in 1990, 1991, 1999 and 2000

the median was 2.0 ind. 10 ha1 (range: 0.89–13.24, n  10) (Fig. 16.3). The breeding population of great crested grebes in the shallow bay area had declined to 5–6 pairs. Greater numbers of cormorants were observed in backwaters when the water level was lower (r  0.7176, P  0.01) (Fig. 16.4). No such correlation was found for grey heron and herring/Caspian gull.

16 11/02/2003 14:00 Page 216

216

Bird, fish, habitat interactions

Figure 16.3 Annual changes in density and community structure of birds in different parts of the reservoir in 1990 and 1999. (a) shallow bay, (b) deep zone, (c) main basin

Figure 16.4 Correlation between the number of cormorants in backwaters of the reservoir and water level, represented by altitude (P  0.01, n  32)

16 11/02/2003 14:00 Page 217

Quantity of fish consumed (kg ha⫺1)

Piscivorous bird community of a Polish reservoir

217

25 20 15 1990 1999

10 5 0 Deep zone

Main basin

Shallow bay

Parts of reservoir

Figure 16.5 Quantity of fish consumed by fish-eating birds in different parts of Dobczyce reservoir

Foraging activity of piscivorous birds was observed in the all parts of the reservoir. The greatest depredation by fish-eating birds (20 kg ha1) was in the shallow bay in 1990 (by breeding great crested grebes) but in the deep part of the reservoir in 1999 (by migrating cormorants) (Fig. 16.5). The three dominant species of fish-eating birds ate 10.5 kg ha1 of fish in 1990 and 13.7 kg ha1 in 1999. About 54% of total fish consumption was between August and October in 1990, rising to 84% in the same months in 1999.

16.5 Discussion The proportion of fish-eating birds in the bird communities of large, shallow lowland reservoirs in Poland is generally low, for example, Sulejowski and Zegrzy´nski (mean 5.4 and 3.0% respectively) (Markowski 1982; Dombrowski et al. 1990) and dominated by phytophagous and benthophagous species. By contrast, Dobczyce Reservoir is mainly an open water body with only a small area of macrophytes and a poor benthos, which cannot support an abundant phytophagous and benthophagous bird community. The great crested grebe was the dominant piscivorous bird species at Dobczyce Reservoir from its filling in the late 1980s to the early 1990s (Gwiazda 1989, 1990). It is also the dominant bird species on other submontane, for example, Orava (Feriancová-Masárová 1962), Otmuchów (Dyrcz 1981), and lowland, for example, Wocawek (Nowysz-Wesoowska 1976), Sulejów (Markowski 1982), Zegrzy´nski (Dombrowski et al. 1990), Jeziorsko (Janiszewski et al. 1998), reservoirs. The average number of great crested grebe at Dobczyce Reservoir was higher than at Zegrzy´nski Reservoir (54.1 ind., Dombrowski et al. 1990) which is shallower or Otmuchów Reservoir (39.2 ind., Dyrcz 1981) where the water level is unstable. The large area of open water without emergent vegetation at Dobczyce Reservoir was used by great crested grebes as a good foraging location during migration and as a safe place for moulting. However, the maximum number of great crested grebes on Dobczyce Reservoir was lower than on Goczakowice Reservoir (850 ind.; Dyrcz et al. 1991) or Jeziorsko Reservoir (900 ind.; Janiszewski et al. 1998), although this was probably

16 11/02/2003 14:00 Page 218

218

Bird, fish, habitat interactions

because the areas of these latter reservoirs are 3–4 times greater than Dobczyce Reservoir. Greater numbers of grey herons and mergansers were recorded on Orava (Feriancová-Masárová 1962) and Otmuchów reservoirs (Dyrcz 1981), presumably because of the greater size of the breeding populations of grey heron at these reservoirs and the greater occurrence of mergansers during their migration. The greater numbers of mergansers recorded at Janiszewski (Dombrowski et al. 1990) and Jeziorsko reservoirs (Janiszewski et al. 1998) were because the species winters in the area and larger flocks are seen during the spring migration (Tomiaojc´ 1990). Birds usually did not winter on Dobczyce Reservoir as it is ice covered. Cormorants were only occasionally observed in the shallow bay at Dobczyce Reservoir, possibly because of the lack of safe resting/roosting sites (islands, flat open banks) in that part of the reservoir. The increase in cormorant and herring/Caspian gull numbers was related to population trends of these species in Europe and their dispersion to new breeding areas (Dubois et al. 1990; Gruber 1995; Lindell et al. 1995; Van Eerden & Gregersen 1995). The breeding population of cormorant in Poland has increased from about 1500 pairs in 1981 to 5200 pairs in 1988, and to more than 8000 pairs in 1992 (Lindel et al. 1995, Przybysz et al. 1997). Herring gulls have been observed more often in Poland since the early 1980s (mainly on reservoirs in southern Poland) but the Caspian gull was rarely observed in that time (Dyrcz et al. 1991). In the late 1990s, breeding Caspian gulls were recorded in southern Poland and their numbers have increased on new sites (Faber et al. 2001). The increase in grey heron numbers on Dobczyce Reservoir can be explained by the good foraging conditions, the occurrence of small pools in backwaters when water levels are low and the establishment of a breeding colony. Nowysz-Wesoowska (1976) claimed that the larger number of this species on Wocawek Reservoir (Central Poland) in autumn was the result of good food conditions and minimal disturbance. The fall in numbers of great crested grebe can partially be explained by smaller numbers of breeding and moulting birds on reservoirs compared with the 1980s (Gwiazda 1989), but it may also be due to disturbance by fishermen who regularly fish at night in the shallows during the breeding season. These activities were undertaken within the biomanipulation programme to reduce planktivorous fish numbers, thus decreasing their pressure on zooplankton and improving the water quality. The marked decline in numbers of common tern was because their nesting places at the time of inundation of the reservoir (on an island with little vegetation) were flooded. The increase in the numbers of cormorants in the backwaters when water levels were lower can be explained by the availability of more roosting sites for the birds (remains of branches and bushes on the bottom and emerging from the water). When the water level was high, cormorants and grey heron occupied the flat shore of the deep part of reservoir, which was protected from the wind. The availability of suitable perches (e.g. poles) and human disturbance (e.g. by fisherman) were also shown to be important factors affecting the distribution of the wintering cormorants in the Po Delta (Boldreghini et al. 1997). The greater consumption of fish in the shallow bay in 1990 was linked to higher densities of fish-eating birds in this part of the reservoir, and the greater consumption of fish in the deep part of the reservoir in 1999 was the result of the presence of high

16 11/02/2003 14:00 Page 219

Piscivorous bird community of a Polish reservoir

219

number of fish-eating birds (mainly cormorants) in autumn. The large breeding and moulting population of great crested grebe in 1990 resulted in the large scale removal of fish biomass from the shallow bay and the middle part of the reservoir. In 1999 the greatest depredation pressure was in the deep part of the reservoir and backwaters by migrating cormorants, despite the greatest fish density (determined by acoustic methods) being in the shallow bay area (50 compared with 400 ind. 1000 m3) (Godlewska & Jelonek 2000). Thus it appears that fish-eating bird depredation was less connected with fish density and more with habitat factors (occurrence of resting and refuge sites). Water level fluctuations appears to be an important factor influencing for the number and distribution of fish-eating birds on Dobczyce Reservoir. The change in bird species resulted in a change in predation pressure on the fish in the reservoir. The fish-eating birds are able to forage in different ecological zones of the reservoir: grey heron in the littoral, cormorant and great crested grebe in pelagic zone and herring/Caspian gull in the epilimnion. The impact of birds in the shallows increased because of foraging by grey heron. Larger fish are a greater potential prey of fish-eating birds because cormorant and grey heron are able to catch bigger fish (40 cm TL) than great crested grebe (21 cm TL) (Cramp & Simmons 1977, Del Hoyo et al. 1992). According to the concept of biomanipulation (improving water quality by reducing the numbers of planktivorous fish to decrease pressure on zooplankton as a natural grazer of phytoplankton) the fish-eating birds may be helpful in this process.

Acknowledgements I thank Micha Baran for providing the information about cormorant, grey heron and herring/Caspian gull in the backwaters of the reservoir in autumn 1998 and 1999.

References Bauer K.M. & Glutz von Blotzheim U.N. (1966) Handbuch der Vogel Mitteleuropas. Frankfurt am Main: Akademische Verlagsgesellschaft, 1, 483 pp. Boldreghini P., Santolini R., Volponi S., Casini L., Montanari F.L., Tinarelli R. (1997) Variations in the use of foraging areas by a cormorant Phalacrocorax carbo wintering population: a case study in the Po Delta (northern Italy). Ekologia Polska 45, 197–200. Cramp S. & Simmons K.E.L. (1977) Handbook of the Birds of Europe, the Middle East and North Africa, Vol. 1: The Birds of Western Palearctic. Oxford: Oxford University Press, 722 pp. Cramp S. & Simmons K.E.L. (1983) Handbook of the Birds of Europe, the Middle East and North Africa, Vol. 3: The Birds of Western Palearctic. Oxford: Oxford University Press, 913 pp. Cramp S. & Simmons K.E.L. (1985) Handbook of the Birds of Europe, the Middle East and North Africa, Vol. 4: The Birds of Western Palearctic. Oxford: Oxford University Press, 960 pp. Del Hoyo J., Elliott A., Sargatal J. (1992) Handbook of the Birds of the World, Vol. 1. Barcelona: Lynx Edicions, 696 pp. Dombrowski A., Kot H. & Rzepaa M. (1990) Birds assemblages of Zegrzy´nski Reservoir. In Z. Kajak (ed.) Functioning of Water Ecosystems and Their Protection and Reclamations. I. Ecology of Water Dam Reservoirs and Rivers. Warsaw: SGGW-AR Warsaw, pp. 163–180 (in Polish).

16 11/02/2003 14:00 Page 220

220

Bird, fish, habitat interactions

Dubois P.J., Skakuj M. & Stawarczyk T. (1990) Occurrence of yellow-legged gull in Poland. Dutch Birding 12, 14–17. Dyrcz A. (1981) Birds of the Otmuchów Water Reservoir. Acta Zoologica Cracoviensia 25, 69–102 (in Polish with English summary). Dyrcz A., Grabn´ski W., Stawarczyk T. & Witkowski J. (1991) Birds of Silesia – Faunistic Monograph. Wrocaw: Uniwersytet Wrocawski, 526 pp. (in Polish with English summary). EIFAC (1988) Report of the EIFAC Working Party on Prevention and Control of Bird Predation in Aquaculture and Fisheries Operations. EIFAC Technical Paper 51, 79 pp. Faber M., Betleja J., Gwiazda R., Malczyk P. (2001) Mixed colonies of large white-headed gulls in southern Poland. British Birds 94, 529–534. Feriancová-Masárová Z. (1962) Vy´znam Oravskej Priehrady pre tah a hniezdenie vodného vtáctva. Biológia, Bratislava 17, 340–354 (in Czech with German summary). Godlewska M. & Jelonek M. (2000) Ichthyofauna. In J. Starmach & G. Mazurkiewicz-Boro (eds) Dobczyce Dam Reservoir. Ecology – Eutrophication – Protection. Kraków: ZBW PAN, pp. 137–148 (in Polish). Gruber D. (1995) Die kennzeichen und das vorkommen der Weiskopasmöwe Larus cachinnans in Europe. Limicola 9, 121–165. Gwiazda R. (1989) Initial stage of bird settlement on the Dobczyce dam reservoir (Vistula basin, southern Poland). Acta Hydrobiologica 31, 373–384. Gwiazda R. (1990) An attempt at estimating the trophic role of birds during formation of the ecosystem of the Dobczyce Reservoir (basin of the River Vistula, southern Poland). Acta Hydrobiologica 32, 457–467. Gwiazda R. (1996) Birds assemblages and the diet of waterfowl at the Dobczyce Dam reservoir in the first years of its existence. Folia Zoolica 45, 161–169. Gwiazda R. (2000) The trophic role of water birds. In J. Starmach & G. Mazurkiewicz-Boro (eds) Dobczyce Dam Reservoir. Ecology – Eutrophication – Protection. Kraków: ZBW PAN, pp. 185–192 (in Polish). Janiszewski T., Wodarczyk R., Bargiel R., Grzybek J., Kali´nski A., Lesner B. & Mielczarek S. (1998) Birds of the Jeziorsko reservoir in 1986–1996. Notatki Ornitologiczne 9, 121–150 (in Polish with English summary). Lindell L., Mellin M., Musil P., Przybysz J. & Zimmerman H. (1995) Status and population of development of breeding Cormorants Phalacrocorax carbo sinensis of the central Europen flyway. Ardea 83, 81–92. Markowski J. (1982) Birds of Pilica Valley. Ochrona Przyrody 44, 163–217 (in Polish with English summary). Nowysz-Wesoowska W. (1976) Observations on the water and marsh birds of the storage-reservoir on the Vistula near Wocawek during migration seasons. Acta Zoologica. Cracoviensia 21, 501–525 (in Polish with English summary). Przybysz J., Mellin M., Mirowska-Ibron I., Przybysz A. & Gromadzka J. (1997) Recent development of the cormorant Phalacrocorac carbo sinensis population in Poland. Ekologia Polska 45, 111–115. Tomiaojc´ L. (1990) The Birds of Poland – Their Distribution and Abundance. Warsaw: PWN, 462 pp. (in Polish with English summary). Van Eerden M.R. & Gregersen J. (1995) Long-term changes in the northwest European population of Cormorants Phalacrocorac carbo sinensis. Ardea 83, 61–80.

17 11/02/2003 15:00 Page 221

Chapter 17

The role of parasites in fish–bird interactions: a behavioural ecological perspective I. BARBER* Institute of Biological Sciences, University of Wales Aberystwyth, Aberystwyth, UK

Abstract Many parasites in aquatic ecosystems have evolved life cycles that utilise both fish and birds as hosts, with transmission typically being achieved when piscivorous birds ingest infected fish. As a consequence, such parasites are intrinsically involved in fish–bird interactions, and many appear to have evolved strategies of host exploitation to facilitate this route of transmission by altering aspects of fish biology – behaviour, sensory performance or morphology – in ways that are likely to increase avian predation pressure. Other parasites, which do not have a requirement to be transmitted to birds, may also influence the availability of infected fish to avian predators as a side effect of general pathology associated with infection. As a result, parasitised fish may become more available to avian piscivores. The literature on infection-associated changes in fish biology that are likely to influence fish–bird interactions is reviewed, and studies of avian predation on infected fish populations are examined to synthesise knowledge and determine important areas for future study. Keywords: avian ecology, diseases, parasites, predation, piscivory, trophic interaction.

17.1 Introduction: why parasites can influence fish–bird interactions Many parasites have evolved complex life cycles that require transmission between a number of separate hosts before they can reproduce. Amongst the cestodes, trematodes, nematodes and acanthocephalans, transmission between intermediate and definitive (adult-harbouring) hosts in the life cycle is typically achieved via the food chain, with parasites relying on the ingestion of the current host by a susceptible predator. In aquatic systems, many parasite taxa have evolved life cycles that utilise fish as intermediate hosts and birds as definitive hosts, relying on their fish hosts being ingested by avian predators to be transmitted. Parasites whose fish hosts are ingested by susceptible avian predators therefore have a high probability of producing large numbers of progeny, whereas those infecting fish that are not eaten by birds effectively have a fitness of zero. There is therefore intense selection pressure on parasites to evolve mechanisms of host ‘manipulation’ to facilitate this route of transmission (Poulin 1998).

*Correspondence: Iain Barber, Institute of Biological Sciences, Edward Llwyd Building, Penglais Campus, University of Wales Aberystwyth, Aberystwyth SY23 3DA, UK (email: [email protected]).

17 11/02/2003 15:00 Page 222

222

Bird, fish, habitat interactions

Observed changes in the behaviour and morphology of fish infected with such parasites often appear to enhance the likelihood of contact between prey fish and their avian predators, and in some cases a disproportionately high predation pressure by birds on infected fish has been recorded (e.g. van Dobben, 1952; Lafferty & Morris, 1996). Other fish parasites may not require ingestion by birds to complete their life cycle, but may still be capable of surviving passage through the avian intestine and could thereby benefit from the dispersal mechanism birds offer. Even parasites that do not benefit from transmission to birds may influence the availability of host fish to birds if infection has a significant impact on behaviour and/or mortality of host fish. Parasites therefore, through a variety of mechanisms, have the potential to dramatically influence predator–prey relationships between fish and birds, both in natural fish populations and in those managed for sport or aquaculture. In addition, because fish become infected with many parasites following the invasion of infective stages that emanate directly or indirectly from eggs passed into the water with the faeces of infected birds, it is clear that piscivorous birds play a central role in the spread of certain types of parasite infections between fish populations. This paper reviews the literature on the effects of parasites on aspects of the behaviour and ecology of fish hosts that are likely to have considerable influence on fish–bird interactions.

17.1.1

The range of parasites involving fish and birds in their life cycle

Fish are parasitised by a range of taxa, but the precise nature of the role of the fish as hosts depends on the parasite in question. Whereas some fish parasites can complete their entire life cycle living on or inside a single fish (e.g. monogenean trematodes), or can be transmitted directly between individuals of the same host fish species (e.g. branchiuran lice), others exhibit complex, indirect life cycles, requiring at least one other host species to complete the life cycle (e.g. pseudophyllidean cestodes).

17.1.2

Parasites that use fish and birds as obligate hosts

A wide variety of parasites exhibit life cycles that have an absolute requirement to utilise both fish and birds in their life cycles. In aquatic ecosystems, avian piscivores typically occupy the highest trophic levels, and consequently many parasites have evolved life cycles that rely on transmission between fish intermediate hosts and bird definitive hosts. Transmission is typically achieved when a piscivorous bird ingests a fish harbouring the infective larval form of the parasite. Many of the digenean trematodes, pseudophyllidean cestodes, nematodes and acanthocephalans utilise fish as intermediate hosts, in which they undertake significant levels of growth and/or development, before achieving the reproductively active adult form in the alimentary tract of piscivorous birds. Birds are thought to be particularly attractive definitive hosts for parasites because their mobility between water bodies, catchments and even continents,

17 11/02/2003 15:00 Page 223

The role of parasites in fish–bird interactions

223

offers an otherwise unavailable mechanism for dispersal and exploitation of remote host populations (Bourne 1989). Although these parasites often exhibit narrow host specificity in terms of the fish utilised as intermediate hosts, specificity with respect to the definitive host is less common, and a wide variety of avian taxa have been recorded as harbouring many of these parasites – effectively in most cases any fish-eating bird will suffice. Predictably, the majority of studies examining the role of parasites in fish–bird interactions deal with these types of parasites.

17.1.3

Fish parasites that do not infect birds

It should be recognised that parasites do not need to have a requirement to mature within an avian host to influence fish–bird interactions. Any parasite infection that has a significant impact on some aspect of host biology (e.g. growth, body condition, sensory performance or behaviour) could influence the susceptibility of the fish host to avian predation. In addition, those parasites that bring about the premature death of fish hosts – either through direct pathology or indirectly by facilitating the colonisation by pathogenic secondary infections – also increase the availability of dead or dying fish, which may be taken by either truly piscivorous, or scavenging, birds. Where the parasites die as a result of ingestion by the bird (e.g. directly transmitted ectoparasites, fish-dwelling adult worms, or the larval stages of helminths that require transmission to non-avian taxa) increased levels of predation are usually interpreted as side-effects of infection (e.g. Smith & Margolis 1970), but if the parasites are capable of surviving transmission through the avian gut (e.g. myxozoan infections such as Myxobolus cerebralis, the causative agent of whirling disease; Taylor & Lott 1978) they potentially benefit from dispersal to remote fish populations. In the latter case, parasite mechanisms that lead to increased avian predation on infected fish may also be actively selected.

17.2 Changes in the behaviour of parasitised fish expected to increase susceptibility to avian predators Many studies have compared the behaviours of parasitised fish with those of nonparasitised individuals in an attempt to determine whether infection is associated with behavioural variation that could serve to facilitate parasite transmission to susceptible avian predators (see review by Barber et al. 2000). The various behaviours affected by parasite infections that could influence fish–bird interactions are reviewed in this section, and key empirical studies are detailed in Table 17.1.

17.2.1

Conspicuous behaviours of infected fish

Although difficult to quantify, many of the behaviours reported to be associated with parasite infections appear to increase the visibility of infected fish to human observers and, presumably, to avian predators. Conspicuous ‘tumbling’ at the water surface

Parasite

Vision

3-spined stickleback

Diplostomum spathaceum (Trematode) Diplostomum spathaceum (Trematode) Schistocephalus solidus (Cestode) Diplostomum spathaceum (Trematode) Diplostomum spathaceum (Trematode) Diplostomum phoxini (Trematode) Myxobolus cerebralis (Myxosporidea) Ligula intestinalis (Cestode) Diplostomum mordax (Trematode) Pseudoterranova decipiens (Nematode) Myxobolus arcticus (Myxozoan) Nanophyetus salmonicola (Trematode) Ascocotyle pachycystis (Trematode) Eubothrium salvelini (Cestode) Schistocephalus solidus (Cestode)

Dace Time budget

3-spined stickleback Dace

Activity

Guppy

Swimming performance

European minnow Various salmonids Common shiner Pejerrey Smelt, eels Sockeye Coho, steelhead Sheepshead minnow

Antipredator behaviour

Sockeye 3-spined stickleback

Next host Bird Bird Bird Bird Bird Bird (None) Bird Bird Mammal (None) Bird Bird (None) Bird

Notes

References

Decreased reactive distance to prey Decreased attack success on prey Increased time spent foraging Increased time spent foraging Reduced activity levels

Owen et al. (1993)

Intermittent swimming movements Erratic circular swimming at water surface ‘Jerky’ and ‘sluggish’ swimming movements Impaired movement

Crowden & Broom (1980) Giles (1987b) Crowden & Broom (1980) Brassard et al. (1982) Rees (1955) Markiw (1992) Dence (1958) Szidat (1969)

Reduced maximum swimming speed Reduced swimming speed

Sprengel & Luchtenberg (1991) Moles & Heifetz (1998)

Reduced burst swimming speed Reduced ability to sustain maximum velocity Decreased fatigue distance Feed closer to predatory fish

Butler & Milleman (1971) Coleman (1993) Smith & Margolis (1970) Milinski (1990)

17 11/02/2003 15:00 Page 224

Host fish

Bird, fish, habitat interactions

Host trait

224

Table 17.1 A summary of documented changes in the behaviour of fish infected with the intermediate stages of parasites that potentially alter susceptibility to avian predation

Bird

3-spined stickleback

Schistocephalus solidus (Cestode)

Bird

Upland bully

Telogaster opisthorchis (Trematode)

Bird

3-spined stickleback

Schistocephalus solidus (Cestode)

Bird

European minnow

Ligula intestinalis (Cestode)

Bird

Killifish

Crassiphiala bulboglossa (Trematode) Ligula intestinalis (Cestode) Schistocephalus solidus (Cestode)

Bird

Gudgeon, roach 3-spined stickleback 9-spined stickleback Fathead minnows Dace 3-spined stickleback Pollack, whiting menhaden

Bird Bird Bird Bird Bird (None) (None)

Godin & Sproul (1988) Tierney et al. (1993) Poulin (1993) Barber et al. (1995) Barber & Huntingford (1996) Krause & Godin (1994) Bean & Winfield (1992) LoBue & Bell (1993) Smith & Kramer (1987) Radaburgh (1980) Crowden & Broom (1980) Jakobsen et al. (1988) Sproston & Hartley (1941) Guthrie & Kroger (1974)

225

Schistocephalus solidus (Cestode) Ornithodiplostomum ptychocheilus (Trematode) Diplostomum spathaceum (Trematode) Schistocephalus solidus (Cestode) Lernaeocerca branchialis (Copepod) Lernaeenicus radiatus, (Copepod) Olencira praegustator (Isopod)

Bird

Returned quicker to food patch after being disturbed by model heron Recovered more quickly following disturbance by a model heron Reduction in antipredator behaviour increased with host age Infected fish left shoals quicker when food deprived Increased nearest neighbour distance in schools; occupied peripheral positions Occupied peripheral positions in schools Infected fish swam higher in the water column Infected fish more buoyant and swam higher in water column Swam higher in water column Infected fish swam closer to water surface Infected fish swam closer to water surface Infected fish found more distant from vegetation Delayed seaward spring migration Infected adults remained with juvenile schools in estuaries

The role of parasites in fish–bird interactions

Habitat use

Schistocephalus solidus (Cestode)

17 11/02/2003 15:00 Page 225

Shoaling / Schooling behaviour

3-spined stickleback

17 11/02/2003 15:00 Page 226

226

Bird, fish, habitat interactions

(Kreirer & Baker 1987), the erratic swimming movements of fish infected with braindwelling trematodes (Rees 1955), ‘whirling’ in salmonids induced by myxozoan parasites (Markiw 1992), the excessive lateral amplitude exhibited by swimming fish infected heavily with cestode plerocercoids (Harris & Wheeler 1974) and the ‘flashing’ behaviour of fish attempting to rid themselves of ectoparasites (Wyman & Walters-Wyman 1985; Urawa 1992), could all attract the attention of avian piscivores. Since the proportion of time predators spend actively locating normally cryptic prey (‘search time’) is generally considerable (Stephens & Krebs 1986), parasitised fish that exhibit behaviours that appear to enhance their visibility and ‘detectability’ could lead to significant predation pressure from birds attempting to maximise food intake rates.

17.2.2

Surface swimming in infected fish

Even when infection is not associated with highly conspicuous erratic swimming movements such as those described above, changes in the position of infected fish in the water column are expected to have significant consequences for the susceptibility of host fish to predation by the many piscivorous birds (including gulls, terns, herons and kingfishers) that are limited to foraging in the upper water column (Kramer et al. 1983). Such effects are some of the most commonly reported changes in the behaviour of fish hosts infected with the larval stages of bird parasites (Barber et al. 2000; Table 17.1). Various mechanisms appear to be responsible for the observed patterns of vertical distribution of infected fish. Experimental investigations suggest that fish infected with cestode plerocercoids may be more buoyant than uninfected conspecifics (LoBue & Bell 1993). In addition, the increased metabolic rate and the greater energetic costs of locomotion suffered by fish infected with nutritionally-demanding parasites may lead them to exploit favourable oxygen tensions close to the water surface (Walkey & Meakins 1970; Lester 1971; Giles 1987a). Alternatively, the reduced foraging performance associated with infections that impair vision (see Section 17.3.1) may force fish into spending prolonged periods feeding on abundant surface prey items (Crowden & Broom 1980). Laboratory experiments using green heron, Butroides striatus virescens L., as a predator confirm that, for this type of piscivorous bird at least, the risk of predation does increase for fish swimming nearer to the surface (Kramer et al. 1983).

17.2.3

Habitat use of infected fish

As well as deviant vertical positioning, other observed changes in the spatial distribution of parasitised fish may have consequences for their relative susceptibility to avian predation. A trapping programme determined that three-spined sticklebacks, Gasterosteus aculeatus L., harbouring plerocercoids of the cestode Schistocephalus solidus (Müller) were more likely to be found further away from vegetation than uninfected fish (Jakobsen et al. 1988). The same authors reported selective predation on infected fish by predatory salmonids, but an increase in predation by piscivorous birds is also a likely consequence of straying from cover.

17 11/02/2003 15:00 Page 227

The role of parasites in fish–bird interactions

227

Parasites also appear to interfere with the seasonal migrations undertaken by both freshwater and marine fish into and out of shallow inshore waters, which offer favourable growing conditions but are also associated with an elevated predation risk from certain types of piscivorous birds (e.g. cormorants, Phalacrocorax sp.; Adams et al. 1994). Fish infected with nutritionally-demanding parasites, such as cestode plerocercoids (Bean & Winfield 1989) or ectoparasitic copepods (Sproston & Hartley 1941), appear to delay offshore migrations, presumably because infections either oblige hosts to take advantage of equitable foraging conditions, or render them less responsive to environmental stimuli that naturally control such movements. However, attributing causality is frequently a problem in studies of infection-associated variation in habitat use. Balling & Pfeiffer (1997) suggested that the increased occurrence of the eyefluke Tylodelphis clavata (Nordmann) at certain sites within one lake may be explained by the local abundance of herons, Ardea cinerea L., the final host of the parasite. A similar correlation has been proposed between the distribution of piscivorous birds and local levels of infection with Schistocephalus solidus among a lacustrine three-spined stickleback population (Gilbertson 1979). If the occupation of certain habitats or regions of a water body exposes fish to a higher risk of infection, and the fish do not disperse to any great extent, then observational studies alone will be insufficient to demonstrate parasite-induced shifts in habitat preference.

17.2.4

Antipredator behaviour of infected fish

Fish have evolved a wide variety of antipredator behaviours to cope with the equally wide array of potential predators they must cope with in aquatic ecosystems (Godin 1997). Although habitat use may be described as one form of antipredator behaviour, the term is more usually applied to reaction time, active escape behaviour, swimming performance and aggregative behaviours such as shoaling and schooling (which are the primary antipredator behaviours of small fish occupying open waters). Because of the consequences of predation for the transmission of many parasites, various aspects of the antipredator behaviour of infected fish have received considerable attention.

Increased ‘boldness’ of infected fish Infections that impose significant nutritional drains on fish hosts, or reduce their competitive ability, have been shown to force infected fish to feed closer to potential predators (Milinski 1985), or to return to a feeding area sooner following disturbance by a predator (Giles 1987b; Godin & Sproul 1988; Tierney et al. 1993). Presumably the function of this behaviour from the fish’s perspective is to maximise food intake rate by taking advantage of the absence of non-infected competitors, yet both behaviours are likely to increase predation risk. By returning quickly to recently disturbed areas, infected fish could place themselves at risk from avian predators that may make multiple successive strikes in one place (e.g. herons, Ardea spp., and kingfishers, Alcedo spp.). Parasites that negatively influence the competitive ability of their host fish may additionally force infected fish to leave shoals – which offer protection against predators but which expose their competitive disabilities – in an attempt to maximise food intake

17 11/02/2003 15:00 Page 228

228

Bird, fish, habitat interactions

rates (Barber et al. 1995; Barber & Ruxton 1998), with important consequences for predation risk.

Reduced escape performance Once attacked, fish must attempt to evade the attacking predator. This relies to a great extent on their ability to detect and respond to the stimulus and swim quickly from the point of impact. Burst-swimming performance, which is likely to affect the outcome of predatory strikes by birds such as terns, Sterna spp., kingfishers and herons that attack fish from above, can be markedly reduced by trematode infections (Butler & Millemann 1971), whereas reduced swimming stamina, such as that associated with some cestode infections (Smith & Margolis 1970), may increase the predation susceptibility of fish to birds that engage in underwater pursuit of prey, such as cormorants, grebes, Podiceps spp., and divers, Gavia spp. (Ehrlich et al. 1994). The mechanisms that bring about reduced swimming performance are varied: atrophy of the musculature associated with plerocercoid infections (Sweeting 1977); the anaesthetic effects of secretions from muscle-dwelling nematodes (McClelland 1995); and pathology of the nervous system associated with infections (Markiw 1992); are all likely to interfere with normal swimming movements of parasitised fish. Other pathology, such as pulmonary obstruction, may significantly decrease the time infected fish are able to swim at their maximum sustainable velocity before becoming exhausted (Coleman 1993).

17.3 Parasite case studies Research into the behavioural and ecological effects of avian-transmitted parasites of fish has tended to be limited to a small number of ‘model’ systems. Here, the potential of three well-studied parasites to influence fish–bird interactions is examined in detail. The parasites have been selected because they infect fish of value in aquaculture, in sport fisheries and/or because they are major parasites of fish that are of importance in the diet of piscivorous birds.

17.3.1 Diplostomum spp. (Trematoda: Diplostomatidae) Diplostomatid trematodes infect a wide range of freshwater fish, when free-swimming cercariae released from parasitised freshwater molluscs penetrate the skin and migrate through the host fish’s body to a parasite species-specific infection site (Erasmus 1959). Piscivorous birds, particularly gulls, acquire the parasite after feeding on infected fish. Metacercariae of the trematode Diplostomum spathaceum (Rudolphi) are common parasites of a wide variety of host fish taxa, including commerciallyimportant salmonids, which migrate to the eye of host fish and occupy the lens causing exophthalmia (Chappell et al. 1994) and impair vision to the extent that host foraging success is reduced (Crowden 1976; Owen et al. 1993). High levels of infestation cause losses in aquaculture (Field & Irwin 1994) and lead to a reduced catch rate by anglers

17 11/02/2003 15:00 Page 229

The role of parasites in fish–bird interactions

229

in high-value salmonid sport fisheries (Moody & Gaten 1982). The impaired visual performance of infected fish has both an indirect and a probable direct consequence for their susceptibility to avian predation. Diplostomum spathaceum-infected dace, Leuciscus leuciscus (L.) (stream dwelling cyprinids) spend more time at the water surface, where they forage on drift items – but would be particularly susceptible to avian predation (Kramer et al. 1983) – to compensate for their reduced foraging performance resulting from lower visual acuity (Crowden & Broom 1980). Also, infection-induced visual impairment (Owen et al. 1993) potentially reduces the ability of infected fish to respond to an attacking predator, although experiments would be needed to confirm this. For other species, such as Diplostomum phoxini (Faust) and the closely related Ornithodiplostomum ptychocheilus (Faust), the fish brain is the site of infection, and the parasites aggregate in and damage lobes of the brain that are associated with host swimming, postural control and spatial awareness (Barber & Crompton 1997). Behavioural traits associated with infection are consistent with these observations. Schools of fathead minnows, Pimephales promelas Rafinesque, infected with these parasites form less compact, less cohesive shoals than those formed by uninfected fish (Radabaugh 1980), which potentially reduces inter-individual communication and the ability to perform coordinated group manoeuvres (Gray & Denton 1991). Shoals of infected fish would be expected to be less able to successfully evade attacks from avian (and other) predators. The conspicuous surface swimming behaviour exhibited by minnows harbouring large numbers of D. phoxini (Rees 1955) would also be expected to increase both the visibility and the predation susceptibility of host fish to avian predators.

17.3.2 Ligula intestinalis (L.) (Cestoda: Pseudophyllidea) Ligula plerocercoids are widespread and common parasites of cyprinid fish in the northern hemisphere and have been recorded from the body cavity of nine British cyprinids (Kennedy 1974). Ligula is also a common parasite of galaxid and eleotrid fish, such as Galaxias maculatus (Jenyns) and the common bully, Gobiomorphus cotidianus McDowell, in the southern hemisphere (Pollard 1974; Weekes & Penlington 1986) and of catostomid fish in North America (Hoffman 1999). Ligula exhibits a typical pseudophyllidean cestode life cycle, requiring transmission between three different hosts – a cyclopoid copepod, a fish and a piscivorous bird – before it can reproduce (Dubinina 1966). Cormorants are commonly infected with Ligula in the UK (R. Kirk, personal communication). Ligula plerocercoids grow to a large size in the body cavity of infected fish (see Fig. 17.1) and heavy infections can be fatal for host fish, resulting in occasional mass mortality in some situations (e.g. Burrough & Kennedy 1979) and potentially increasing localised food availability for scavenging birds (Wyatt & Kennedy 1988). Ligula also has various sub-lethal effects on the biology of host fish (Arme & Owen 1968; Sweeting 1977; Hoole & Arme 1983) that may facilitate its transmission to susceptible bird predators. One of the most obvious external symptoms of ligulosis is host body distension (e.g. Smith 1973; Barber 1997). The body wall becomes

17 11/02/2003 15:00 Page 230

230

Bird, fish, habitat interactions

Figure 17.1 Two Ligula intestinalis plerocercoids recovered from the body cavity of an infected minnow from Loch Maragan, Scotland

extremely thin in infected individuals due to muscular atrophy (Sweeting 1977) and stretches to accommodate the rapidly growing plerocercoids, even causing the scales to become separated in extreme cases (Arme & Owen 1968). The swelling increases cross-sectional area, flow resistance and frictional drag (Rodewald & Foster 1998), changes that reduce the body flexibility necessary for fast starts and are costly in terms of locomotory speed and efficiency (Blake 1983; Videler 1993). The ‘jerky’ swimming movements, characterised by increased lateral amplitude relative to swimming speed, reported from infected bleak, Alburnus alburnus (L.) (Harris & Wheeler 1974), and the ‘sluggish’ movements of infected common shiners, Notropis cornutus (Mitchill), (Dence 1958) probably result from such mechanisms, and are likely to increase the visibility and susceptibility of infected fish to avian predation. Ligula infection is also associated with a change in the schooling behaviour of minnow hosts. Minnows and other small cyprinid fish rely on schooling behaviour to a great extent to combat predation, and an increase in schooling behaviour is typically observed under threat of avian attack (Litvak 1993). Ligula-infected minnows in schools of otherwise uninfected conspecifics occupied peripheral positions (thought to offer the lowest protection against predation) more often than expected and exhibited increased nearest neighbour distances (Barber & Huntingford 1996), potentially reducing their ability to participate in coordinated manoeuvres (Barber & Folstad 2000). Ligula infections are also associated with increase in swimming height in the water column (Bean & Winfield 1992; Marshall & Cowx, Chapter 18) and a delay in the timing of offshore migrations of cyprinid hosts, potentially increasing the time they are in contact with avian predators (Bean & Winfield 1989). Yet despite these indications that infected fish are likely to be more susceptible to predation by definitive avian hosts, few studies have examined the actual predation rate on Ligula-infected and uninfected individuals (but see a notable study by van Dobben 1952 detailed in section 17.5).

17 11/02/2003 15:00 Page 231

The role of parasites in fish–bird interactions

231

17.3.3 Schistocephalus solidus (Müller) (Cestoda: Pseudophyllidea) Schistocephalus is closely related to Ligula and has a similar life cycle involving a copepod, a fish and a bird definitive host. The fish host is typically the three-spined stickleback, but nine-spined sticklebacks, Pungitius pungitius (L.), can also harbour growing plerocercoids. The Gasterosteus–Schistocephalus system has been used as a model system for the study of host–parasite relationships and consequently a good deal is known about infection-induced behaviour change in infected sticklebacks (reviewed by Milinski 1990; Barber & Huntingford 1995). Infected fish exhibit a host of morphological and behavioural characteristics that are expected to maximise the likelihood of their eventually succumbing to avian predation. As well as having disrupted patterns of camouflage (see Fig. 17.2(a) and section 17.4) infected fish spend more time swimming at the surface to exploit oxygen tensions (Lester 1971), will feed closer to predators (Milinski 1985) and will return more quickly to an area following disturbance by a simulated avian piscivore (Giles 1987b). In addition, the swimming behaviour of infected sticklebacks is impaired and infected fish are less likely to respond to avian predators (Ness & Foster 1999). Studies by Tierney et al. (1993) demonstrated that behaviour changes in stickleback hosts that would be expected to enhance transmission by increasing susceptibility to avian predators are not observed until the parasite reaches a size at which it is infective to piscivorous birds. In this case pre-infective worms (those 50 mg in weight; Tierney & Crompton 1992) appeared to enhance the antipredator behaviour of host fish, a strategy that would be predicted if parasites were manipulating host behaviour in an adaptive fashion. Again, however, no studies have examined the avian predation pressure on infected compared with non-infected sticklebacks in natural populations, although anecdotal accounts suggest that this may well occur (G. Gilbert, personal communication).

17.4 Changes in host appearance Changes in the appearance, as well as in the behaviour, of host fish may also have significant consequences for the relative susceptibility of infected fish to avian predation, although this is an area for which very little experimental evidence exists. Although highly likely to increase visibility to, and predation pressure from, avian piscivores, the role of morphological changes alone (in isolation from concomitant behavioural manipulation) has not been ascertained. Such changes take a variety of forms but typically disrupt the usual appearance of the fish, which is expected to have evolved to be optimal for local conditions. Swellings associated with heavy pseudophyllidean burdens disrupt the natural patterns of countershading, since the paler coloured ventral surface is visible from above (Barber 1997), effectively highlighting infected fish against a darker substratum (see Fig. 17.2(a)). In some populations of sticklebacks harbouring Schistocephalus, heavy infection is coupled with demelanisation of the skin on the dorsal surface, leaving the fish appearing white in colour (LoBue & Bell 1993; Ness & Foster 1999). Other changes include the black melanised skin spots that appear as a host reaction to

17 11/02/2003 15:00 Page 232

232

Bird, fish, habitat interactions

Figure 17.2 Schistocephalus solidus. (a) A juvenile three-spined stickleback Gasterosteus aculeatus harbouring a heavy infection of Schistocephalus solidus. The disruption to natural patterns of host countershading is obvious when viewed from above (lower photograph). (b) The sexually mature, egg-producing adult form of Schistocephalus solidus, the form found in the intestine of piscivorous birds

encysted metacercariae of Cryptocotyle, Neascus, Crassiaphiala and related trematode genera. On pale-bodied fish such as sand gobies, Pomatoschistus minutus (Pallas), or on the fins and pale areas of pipefish (Sygnathus spp.), killifish (Fundulus spp.) and sticklebacks, such spots are highly visible and are associated with avoidance by conspecifics (Krause & Godin 1994; Rosenqvist & Johansson 1995). The external

17 11/02/2003 15:00 Page 233

The role of parasites in fish–bird interactions

233

cysts created by myxozoan infections, such as those of Glugea anomala (Moniez) in sticklebacks, increase host visibility to human observers and presumably to other visual predators such as birds; these parasites do not require avian predation for transmission, but the parasite may benefit in two ways if a bird consumes its fish host. Firstly, it is possible that handling by the bird ruptures the parasite cysts, which is necessary before the parasite can be released, and secondly, some of these parasites are known to survive passage through the avian intestine and therefore could benefit from dispersal to remote fish populations (Taylor & Lott 1978). For fish that naturally form shoals, individuals infected with parasites that change host appearance may be at a higher risk of attack from avian predators, since their altered appearance makes them ‘stand out from the crowd’. This potentially increases their susceptibility to visually hunting predators, which are known to target ‘odd’ individuals in groups to overcome the confusion effect associated with foraging on aggregated prey (Landeau & Terborgh 1986).

17.5 Are infected fish more susceptible to predation? A number of studies have reported increased predation by definitive hosts (‘parasite increased trophic transmission’ or PITT; Lafferty 1999) on infected invertebrates (e.g. Holmes & Bethel 1972; Moore 1983) and mammals (e.g. Hoogenboom & Dijkstra 1987; Rau & Caron 1979). However, although fish hosts exhibit a number of behaviour changes that would be expected to increase predation risk (see above and Table 17.1), few studies have demonstrated unambiguously the increased susceptibility of fish intermediate hosts to susceptible avian predators. Lafferty and Morris (1996) presented the best evidence that parasitised fish are subject to greater avian predation, in their study of California killifish, Fundulus parvipinnis Girard, infected with the brain-encysting trematode, Euhaplorchis californiensis Martin. In common with other brain-dwelling infections, Euhaplorchis is known to increase the frequency of conspicuous behaviours that host fish perform. Enclosed pens in a natural lagoon were stocked with mixed groups of experimentally parasitised and unparasitised fish and some were covered with netting to prevent avian predation whilst others were left open to allow predation by birds (herons and egrets, Egretta spp.; definitive hosts of the parasite). After twenty days, the numbers of infected and non-infected fish remaining in the two types of enclosures were counted. Infected fish in enclosures left open to allow avian predation were found to be thirty times more likely to have succumbed to avian predation than non-infected individuals. Another important study is that of van Dobben (1952) who examined the diet of the cormorant in the Netherlands. He reported that although only 6.5% of roach, Rutilus rutilus (L.), in the population at large were infected with Ligula intestinalis, 30% of the roach captured by cormorants where infected, strongly suggesting that parasitised fish were selectively predated by this definitive host under natural conditions. However, enhanced predation on infected fish has not only been reported for piscivorous birds; non-host predators also appear to selectively ingest parasitised fish. Pike, Esox lucius L., selectively predated roach infected with Ligula (Sweeting 1976), as did trout feeding on an infected minnow population (Museth 2001), suggesting that the transmission

17 11/02/2003 15:00 Page 234

234

Bird, fish, habitat interactions

consequences of host behaviour change are complex and are likely to depend on the relative abundance of susceptible and non-susceptible predators.

17.6 Negative impacts of parasites on bird hosts: why don’t birds avoid infected prey? Although it has proven a difficult area for research – and as a result there are few published studies – it seems likely that birds take infected fish disproportionately from host populations. If this is indeed the case then it is important that we ask what effects these parasites have for the biology of the avian host. There is very little known about the detrimental effects of these parasites on their avian hosts, with most studies of bird disease and their effects on hosts focussing on bacterial, viral or arthropod parasites. The assumption is that parasites that rely on fish hosts to be ingested by avian predators should have evolved to have only minimum impact on their definitive hosts, thereby preventing the evolution of bird strategies to avoid infected fish. Many of the parasites are small (e.g. digenean trematodes are typically only a few mm in length) and/or are short lived in the avian host. Schistocephalus solidus, for example, undergoes all of its growth and most of its development in the fish host, allowing it to mature rapidly in the avian intestine, producing eggs within 36 hours and living only a week or so inside the bird (Smyth 1994). Unusually for a tapeworm, the adult has no invasive connection with the host gut, nor does it appear to utilise host-derived nutrients to any marked extent. Holmes & Zohar (1990) stated that a host’s ability to compensate for sublethal effects of parasites on, for example, energy balance is largely dependent on body size and general condition. It is therefore likely that these parasites are not much of a cost to healthy adult birds, but young hosts could be at a higher risk. In addition, Hudson & Dobson (1997, p. 137) stated that ‘the impact of parasites on both host reproduction and survival is often related to the nutritional status of the host. Poorly nourished hosts are more vulnerable to infection because nutrients are diverted by the parasites and because poorly nourished hosts are less able to launch a strong immunological response’. Also, if birds ingest large numbers of infective parasites the sheer physical size of some parasites (e.g. Schistocephalus solidus, see Fig. 17.2(b)) may cause a problem since blockage of the intestine is a real possibility, especially in juveniles (Reece 1989). Other studies suggest that such infections are often associated with poor health and fitness prospects of avian hosts. Locke et al. (1964) reported large merganser, Mergus spp., kills associated with infection with larval nematode parasites. The same types of parasites were associated with gross lesions in the proventriculus and fibrinous haemorrhages over the intestinal serosa of Nankeen night herons, Nycticorax caledonicus Mathews, found dead and in poor nutritional condition (McOrist 1989). Obendorf & McColl (1980) reported Tetrabothrius sp. (a cestode of marine fish-eating birds) to be one of the parasites found in little penguins dying from multiple parasite infections whereas McNeil et al. (1994) found that greater yellowlegs, Tringa melanoleuca Gmelin, heavily infected with trematode parasites, were unable to complete northerly spring migrations due to a shortage of body fat.

17 11/02/2003 15:00 Page 235

The role of parasites in fish–bird interactions

235

Figure 17.3 Fitness consequences of predation of fish by birds: (a) for fish that do not harbour parasites infective to birds; (b) for fish harbouring parasites infective to birds. P: Predation; T: transmission. Although fitness consequences for the fish and the parasite are obvious, the consequences to birds of feeding on infected fish are unclear

Observational evidence therefore suggests that infections acquired from the ingestion of parasitised fish should be avoided by birds, since even if they are not particularly harmful to healthy adult birds, any infections are likely to have more severe consequences for young, sick or senescent individuals. So why do birds feed on infected fish? There are several reasons why birds may not discriminate against infected prey. If prey fish harbouring infective parasites are recognisable, then birds should have evolved selection strategies to avoid them, but only providing that the fitness benefit of avoiding infection outweighs the fitness cost of adopting of the strategy (Fig. 17.3). Therefore, if infected fish offer only a minimal health risk but are, for example, much easier to locate and/or catch (thereby reducing the energy expended in foraging) then there may be no pressure on birds to select against parasitised prey, even if they can be distinguished as such. Indeed, Lafferty (1992) has even suggested that in some circumstances predators may prefer infected prey. In addition, fish infected with large plerocercoids often appear swollen and show similar behavioural characteristics to gravid females (personal observation); even if infection has a negative impact on bird fitness a strategy that discriminated against infected fish may lead them also to select

17 11/02/2003 15:00 Page 236

236

Bird, fish, habitat interactions

against gravid females, which may lead to them ignoring the most energetically profitable fish in the prey population.

17.7 Future challenges Although it seems, from the studies available, that birds are likely to have a higher foraging success on fish infected with certain types of parasites, there is a need for more data on the relative avian predation pressure on infected fish. Apart from this gap in knowledge, there are other areas that would benefit from greater research. Here three such areas are highlighted.

17.7.1

Parasites and ecosystem change: consequences for fish–bird interactions

When examining the distribution of parasites of a particular species within host populations, the majority of parasite infections exhibit some degree of overdispersion, with the majority of fish harbouring low levels of infections and a small proportion of fish harbouring heavy burdens. In many cases there is likely to be a threshold effect of infection on host behaviour, with parasites having to attain a high enough intensity before they significantly influence host susceptibility to predation. Those individuals comprising this small proportion of very heavily infected fish are likely to be those that exhibit behavioural change and would therefore be expected to be exposed to the highest risk of avian predation. However, if anthropogenic impacts on aquatic habitats lead to altered parasite distributions, which in turn result in a larger number of individuals harbouring heavier infections, a larger proportion of the fish population are likely to attain the threshold level for behavioural change (see Fig. 17.4 for diagrammatic explanation). Changes in parasite distributions could result from chronic pollution, such as the increase in trematode transmission to fish in the warm water effluents of power stations (Höglund & Thulin 1990), or the creation of artificially high densities in aquaculture or sport fisheries, which serve to maximise the inter-fish transmission of parasites with direct life cycles. The predicted increased predation pressure on the greater proportion of heavily infected fish would have significant consequences for the population dynamics of the fish population, but also could affect piscivorous birds that had evolved under conditions where heavily infected fish were more scarce.

17.7.2

Foraging strategies of avian piscivores: are infected hosts selected actively or passively?

The small numbers of studies that examined avian predation on infected fish found that birds removed a disproportionate number of infected fish from populations (section 17.5). Whether the increased predation risk of parasitised fish is a result of selective targeting of infected individuals (pre-attack mechanisms arising from increased conspicuousness – as a result of infection-associated changes in movement, colouration

17 11/02/2003 15:00 Page 237

The role of parasites in fish–bird interactions

237

Figure 17.4 Histograms showing the overdispersed pattern of distributions of: (a) Cryptocotyle lingua metacercariae on the fins of three-spined sticklebacks (1 year class) from the Gullmarsfjord, Sweden, September 1998; and (b) Ligula intestinalis in the body cavity of minnows Phoxinus phoxinus from Loch Maragan, Scotland 1993 (c) shows a conceptual model by which changes in parasite transmission patterns, brought about as a result of, for example, anthropogenic activity, could affect interactions between fish and bird populations. A typical overdispersed parasite distribution amongst host fish is shown (bold solid line) alongside two other types of distribution (narrow dotted and narrow solid lines) that may be observed following changes to patterns of parasite transmission resulting from anthropogenic disturbance. The bold dotted vertical line identifies a notional threshold level (Icrit) that infections must attain before fish–bird interactions are affected. Changes in patterns of parasite transmission could lead to a larger proportion of the fish population with parasite intensities Icrit, potentially increasing the availability of infected fish to avian predators

17 11/02/2003 15:00 Page 238

238

Bird, fish, habitat interactions

or size – or altered habitat occupancy), or a result of the reduced escape performance of infected fish (post-attack mechanisms resulting from physiological or physical disability associated with parasite infection) is unclear. Although it is likely that some or all of these mechanisms are likely to influence the outcome of predator–prey interactions, there is also likely to be variation in the importance of each depending on the fish, the parasite and the avian predator in question. The foraging strategies employed by different bird species, and the various effects parasites have on host fish behaviour, are likely to expose differentially different avian taxa to particular parasites. Detailed research into the strategies used by piscivorous birds to select prey fish, carried out in conjunction with studies examining changes in fish behaviour associated with specific infections would give significant insight into the development of natural predator-prey communities and the role of parasites in shaping fish–bird interactions in aquatic ecosystems.

17.7.3

Fitness consequences of avian parasites acquired from fish prey

As outlined in Section 17.6, very little is known about the detrimental effects of these parasites on their avian hosts, with most studies of bird diseases focusing on bacterial, viral or arthropod parasites (Clayton & Moore 1997). The few studies available do suggest that the impact of fish-transmitted parasite infections on bird hosts may be considerable, and there is a need to gain a greater understanding of how infections might impact on the growth, development and survival of host birds and in particular juveniles, which may be more susceptible to infections. This is particularly important when considering the implementation of conservation programmes designed to maximise the success of breeding attempts. In many cases, the availability of suitable prey limits breeding success and even chicks that survive are likely to be under nutritional stress. Low prey availability could force adult birds to capture the most easily located prey, which may be those infected fish in the population, leading to problems when nutritionally stressed chicks are fed heavily infected fish. Detailed experimental studies of the effects of infection alone, and under concomitant nutritional stress, on the growth and survival of juvenile avian piscivores would provide much needed data. Such studies should be carried out alongside field research on the breeding success of wild piscivorous birds, which in turn should incorporate indirect measurements of parasite load (e.g. by counting eggs in faeces typically discharged during handling) of all chicks and adults to examine the link between infection status, body condition, growth and survival. From a management perspective, for particularly important breeding populations of birds regular screening of prey fish availability and infection status could also be undertaken to monitor the likely risk of avian piscivores ingesting significant numbers of parasites.

References Adams C.E., Brown D.W. & Keay L. (1994) Elevated predation pressure associated with inshore migrations of fish in a large lake, Loch Lomond, Scotland. Hydrobiologia 290, 135–138.

17 11/02/2003 15:00 Page 239

The role of parasites in fish–bird interactions

239

Arme C. & Owen R.W. (1968) Occurrence and pathology of Ligula intestinalis infections in British fishes. Journal of Parasitology 54, 272–280. Balling T.E. & Pfeiffer W. (1997) Location-dependent infection of fish parasites in Lake Constance. Journal of Fish Biology 51, 1025–1032. Barber I. & Crompton D.W.T. (1997) The distribution of Diplostomum phoxini metacercariae (Trematoda) in the brain of minnows (Phoxinus phoxinus). Folia Parasitologica 44, 19–25. Barber I. (1997) A non-invasive morphometric technique to estimate the cestode plerocercoid load of small freshwater fishes. Journal of Fish Biology 51, 654–658. Barber I. & Folstad I. (2000) Schooling, dusk flight and dance: social organisations as amplifiers of individual quality? Oikos 89, 191–194. Barber I. & Huntingford F.A. (1995) The effect of Schistocephalus solidus (Cestoda: Pseudophyllidea) on the foraging and shoaling behaviour of three-spined sticklebacks, Gasterosteus aculeatus. Behaviour 132, 1223–1240. Barber I. & Huntingford F.A. (1996) Parasite infection impairs schooling behaviour: spatial deviance of helminth-infected minnows in conspecific groups. Proceedings of the Royal Society of London Series B 263, 1095–1102. Barber I. & Ruxton G.D. (1998) Temporal prey distribution affects the competitive ability of parasitised sticklebacks. Animal Behaviour 58, 1477–1483. Barber I., Huntingford F.A. & Crompton D.W.T. (1995) The effect of hunger and cestode parasitism on the shoaling decisions of small freshwater fish. Journal of Fish Biology 47, 524–536. Barber I., Hoare D. & Krause J. (2000) The effects of parasites on fish behaviour: an evolutionary perspective and review. Reviews in Fish Biology and Fisheries 10, 131–165. Bean C.W. & Winfield I.J. (1989) Biological and ecological effects of a Ligula intestinalis (L.) infestation of the gudgeon, Gobio gobio (L.), in Lough Neagh, Northern Ireland. Journal of Fish Biology 34, 135–147. Bean C.W. & Winfield I.J. (1992) Influences of the tapeworm Ligula intestinalis (L.) on the spatial distributions of juvenile roach Rutilus rutilus (L.) and gudgeon Gobio gobio (L.) in Lough Neagh, Northern Ireland. Netherlands Journal of Zoology 42, 416–423. Blake R.W. (1983) Fish Locomotion. Cambridge: Cambridge University Press, 208 pp. Bourne R.P (1989) The role of birds in the long-distance dispersal of disease. In J.E. Cooper (ed.) Disease and Threatened Birds. International Council for Bird Conservation Technical Report, No. 10. Cambridge, pp. 121–128. Brassard P., Rau M.E. & Curtis M.A. (1982) Parasite-induced susceptibility to predation in diplostomoiasis. Parasitology 85, 495–501. Burrough R.J. & Kennedy C.R. (1979) The occurrence and natural alleviation of stunting in a population of roach, Rutilus rutilus (L.). Journal of Fish Biology 15, 93–109. Butler J.A. & Millemann R.E. (1971) Effect of the ‘salmon poisoning’ trematode, Nanophyetus salmonicola, on the swimming ability of juvenile salmonid fishes. Journal of Parasitology 57, 860–865. Chappell L.H., Hardie L.J. & Secombes C.J. (1994) Diplostomiasis: the disease and host-parasite interactions. In A.W. Pike & J.W. Lewis (eds) Parasitic Diseases of Fish. Dyfed: Samara Publishing, pp. 59–86. Clayton D.H. & J. Moore (1997) Host-parasite Evolution: General Principles and Avian Models. Oxford: Oxford University Press, 473 pp. Coleman F.C. (1993) Morphological and physiological consequences of parasites encysted in the bulbus arteriosus of an estuarine fish, the sheepshead minnow, Cyprinodon variegatus. Journal of Parasitology 79, 247–254. Crowden A.E. (1976) Diplostomum spathaceum in the Thames; occurrence and effects on fish behaviour. Parasitology 73, vii. Crowden A.E. & Broom D.M. (1980) Effects of the eyefluke, Diplostomum spathaceum, on the behaviour of dace (Leuciscus leuciscus). Animal Behaviour 28, 287–294.

17 11/02/2003 15:00 Page 240

240

Bird, fish, habitat interactions

Dence W.A. (1958) Studies on Ligula-infected common shiners (Notropis cornutus frontinalis Agassiz) in the Adirondacks. Journal of Parasitology 44, 334–338. Dubinina M.N. (1966) Tapeworms (Cestoda, Ligulidae) of the Fauna of the USSR. Moscow & Leningrad: Akademii Nauk SSSR, 262 pp. (in Russian). English translation, 1980, New Delhi: Amerind Publishing Co. Ehrlich P.R., Dobkin D.S., Pimm S.L. & Wheye D. (1994) The Birdwatcher’s Handbook: a Field Guide to the Natural History of the Birds of Britain and Europe. Oxford: Oxford University Press, 700 pp. Erasmus D.A. (1959) The migration of Cercaria X Baylis (Strigeida) within the fish intermediate host. Parasitology 49, 173–190. Field J.S. & Irwin S.W.B. (1994) The epidemiology, treatment and control of diplostomiasis on a fish farm in Northern Ireland. In: A.W. Pike & J.W. Lewis (eds) Parasitic Diseases of Fish. Dyfed: Samara Publishing, pp. 87–100. Gilbertson L.G. (1979) Variation and natural selection in an Alaskan population of the threespine stickleback (Gasterosteus aculeatus L.). PhD thesis, University of Washington. Giles N. (1987a) A comparison of the behavioural responses of parasitised and non-parasitised threespined sticklebacks, Gasterosteus aculeatus L., to progressive hypoxia. Journal of Fish Biology 30, 631–638. Giles N. (1987b) Predation risk and reduced foraging activity in fish: experiments with parasitised and non-parasitised three-spined sticklebacks, Gasterosteus aculeatus. Journal of Fish Biology 31, 37–44. Godin J.-G.J. (1997) Behavioral Ecology of Teleost Fishes. Oxford: Oxford University Press, 384 pp. Godin J.-G.J. & Sproul C.D. (1988) Risk taking in parasitised sticklebacks under threat of predation: effects of energetic need and food availability. Canadian Journal of Zoology 66, 2360–2367. Gray J.A.B. & Denton E.J. (1991) Fast pressure pulses and communication between fish. Journal of the Marine Biological Association 71, 83–106. Guthrie J.F. & Kroger R.L. (1974) Schooling habits of injured and parasitised menhaden. Ecology 55, 208–210. Harris M.T. & Wheeler A. (1974) Ligula infestation of bleak Alburnus alburnus (L.) in the tidal Thames. Journal of Fish Biology 6, 181–188. Hoffman G.L. (1999) Parasites of North American Freshwater Fishes. Berkeley, CA: Cornell University Press, 539 pp. Höglund J. & Thulin J. (1990) The epidemiology of the metacercariae of Diplostomum baeri and D. spathaceum in perch (Perca fluviatilis) from the warm water effluent of a nuclear power station. Journal of Helminthology 64, 139–150. Holmes J.C. & Bethel W.M. (1972) Modification of intermediate host behaviour by parasites. In E.U. Canning & C.A Wright (eds) Behavioural Aspects of Parasite Transmission. London: Academic Press, pp. 123–149. Holmes J.C. & Zohar S. (1990) Pathology and host behaviour. In C.J. Barnard & J.M. Behnke (eds) Parasitism and Host Behaviour. London: Taylor & Francis, pp. 193–229. Hoogenboom I. & Dijkstra C. (1987) Sarcocystis cernae: A parasite increasing the risk of predation of its intermediate host, Microtus arvalis. Oecologia 74, 86–92. Hoole D. & Arme C. (1983) Ultrastructural studies on the cellular responses of fish hosts following experimental infections with the plerocercoid of Ligula intestinalis. Parasitology 87, 139–149. Hudson P.J. & Dobson A.P. (1997) Host-parasite processes and demographic consequences. In D.H. Clayton & J. Moore (eds) Host-parasite Evolution: General Principles and Avian Models. Oxford: Oxford University Press, pp. 128–154. Jakobsen P.J., Johnsen G.H. & Larsson P. (1988) Effects of predation risk and parasitism on the feeding ecology, habitat use, and abundance of lacustrine threespine stickleback (Gasterosteus aculeatus). Canadian Journal of Fisheries and Aquatic Science 45, 426–431. Kennedy C.R. (1974) A checklist of British and Irish freshwater fish parasites with notes on their distribution. Journal of Fish Biology 6, 613–644.

17 11/02/2003 15:00 Page 241

The role of parasites in fish–bird interactions

241

Kramer D.L., Manley D. & Bourgeois R. (1983) The effect of respiratory mode and oxygen concentration on the risk of aerial predation in fishes. Canadian Journal of Zoology 61, 653–665. Krause J. & Godin J.-G.J. (1994) Influence of parasitism on the shoaling behaviour of banded killifish, Fundulus diaphanus. Canadian Journal of Zoology 72, 1775–1779. Kreirer J.P. & Baker J.R. (1987) Parasitic Protozoa. Boston: Allen & Unwin, 241 pp. Lafferty K.D. (1992) Foraging on prey that are modified by parasites. American Naturalist 140, 854–867. Lafferty K.D. (1999) The evolution of trophic transmission. Parasitology Today 15, 111–115. Lafferty K.D. & Morris A.K. (1996) Altered behaviour of parasitised killifish increases susceptibility to predation by bird final hosts. Ecology 77, 1390–1397. Landeau L. & Terborgh J. (1986) Oddity and the ‘confusion effect’ in predation. Animal Behaviour 34, 1372–1380. Lester R.J.G. (1971) The influence of Schistocephalus plerocercoids on the respiration of Gasterosteus and a possible resulting effect on the behavior of the fish. Canadian Journal of Zoology 49, 361–366. Litvak M.K. (1993) Response of shoaling fish to the threat of aerial predation. Environmental Biology of Fishes 36, 183–192. LoBue C.P. & Bell M.A. (1993) Phenotypic manipulation by the cestode parasite Schistocephalus solidus of its intermediate host, Gasterosteus aculeatus, the threespine stickleback. American Naturalist 142, 725–735. Locke L.N., DeWitt J.B., Menzie C.M. & Kerwin J.A. (1964) A merganser die-off associated with larval Eustrongyloides. Avian Disease 8, 420–427. Markiw M.E. (1992) Salmonid Whirling Disease. US Fisheries and Wildlife Service Leaflet 17. Washington, DC, 11 pp. McClelland G. (1995) Experimental infection of fish with larval sealworm, Pseudoterranova decipiens (Nematoda, Aniskinae), transmitted by amphipods. Canadian Journal of Fisheries and Aquatic Science 52, 140–155. McNeil R., Diaz M.T. & Villeneuve A. (1994) Trematode infestation and the over summering of shore birds. Journal für Ornithologie 135, 25. McOrist S. (1989) Some diseases of free-living Australian birds. In J.E. Cooper (ed.) Disease and Threatened Birds. International Council for Bird Conservation Technical report No. 10, pp. 63–68. Milinski M. (1985) Risk of predation of parasitised sticklebacks (Gasterosteus aculeatus L.) under competition for food. Behaviour 93, 203–216. Milinski M. (1990) Parasites and host decision-making. In C.J. Barnard & J.M. Behnke (eds) Parasitism and Host Behaviour. London: Taylor & Francis, pp. 95–116. Moles A. & Heifetz J. (1998) Effects of the brain parasite Myxobolus arcticus on sockeye salmon. Journal of Fish Biology 52, 146–151. Moody J. & Gaten E. (1982) The population dynamics of eyeflukes Diplostomum spathaceum and Tylodelphys clavata (Digenea: Diplostomatidae) in rainbow and brown trout in Rutland Water: 1974–1978. Hydrobiologia 88, 207–209. Moore J. (1983) Responses of an avian predator and its isopod prey to an acanthocephalan parasite. Ecology 64, 1000–1015. Museth J. (2001) Effects of Ligula intestinalis on habitat use, predation risk and catchability in European minnows. Journal of Fish Biology 59, 1070–1080. Ness J.H. & Foster S.A. (1999) Parasite-associated phenotype modifications in threespine stickleback. Oikos 85, 127–134. Obendorf D. & McColl K. (1980) Mortality in little penguins (Eudyptula minor) along the coast of Victoria, Australia. Journal of Wildlife Diseases 16, 251–259. Owen S.F., Barber I. & Hart P.J.B. (1993) Low level infection by eyefluke, Diplostomum spp. affects the vision of three-spined sticklebacks, Gasterosteus aculeatus. Journal of Fish Biology 42, 803–806.

17 11/02/2003 15:00 Page 242

242

Bird, fish, habitat interactions

Pollard D.A. (1974) The biology of a landlocked form of the normally catadromous Salmoniform fish Galaxias maculatus (Jenyns) VI. Effects of cestode and nematode parasites. Australian Journal of Marine and Freshwater Research 25, 105–120. Poulin R. (1993) Age-dependent effects of parasites on anti-predator responses in two New Zealand freshwater fish. Oecologia 96, 431–438. Poulin R. (1998) Evolutionary Ecology of Parasites: from Individuals to Communities. London: Chapman & Hall, 212 pp. Radabaugh D.C. (1980) Changes in minnow, Pimephales promelas Rafinesque, schooling behaviour associated with infections of brain-encysted larvae of the fluke, Ornithodiplostomum ptychocheilus. Journal of Fish Biology 16, 621–628. Rau M.E. & Caron, F.R. (1979) Parasite-induced susceptibility of moose to hunting. Canadian Journal of Zoology 57, 2466–2468. Reece R.L. (1989) Avian pathogens: their biology and methods of spread. In J.E. Cooper (ed.) Disease and Threatened Birds. International Council for Bird Conservation Technical Report No. 10, pp. 1–24. Rees G. (1955) The adult and Diplostomulum stage (Diplostomulum phoxini (Faust)) of Diplostomum pelmatoides Dubois and an experimental demonstration of part of the life cycle. Parasitology 45, 295–311. Rodewald A.D. & Foster S.A. (1998) Effects of gravidity on habitat use and anti-predator behaviour in three-spined sticklebacks. Journal of Fish Biology 52, 973–984. Rosenqvist G. & Johansson K. (1995) Male avoidance of parasitized females explained by direct benefits in a pipefish. Animal Behaviour 49, 1039–1040. Smith H.D. (1973) Observations on the cestode Eubothrium salvelini in juvenile sockeye salmon (Oncorhynchus nerka) at Babine Lake, British Columbia. Journal of the Fisheries Research Board of Canada 30, 947–964. Smith R.S. & Kramer D.C. (1987) Effects of a cestode (Schistocephalus sp.) on the response of ninespine sticklebacks (Pungitius pungitius) to aquatic hypoxia. Canadian Journal of Zoology 65, 1862–1865. Smith H.D. & Margolis L. (1970) Some effects of Eubothrium salvelini (Schrank, 1790) on sockeye salmon, Oncorhynchus nerka (Walbaum), in Babine Lake, British Columbia. Journal of Parasitology 56, 321–322. Smyth J.D. (1994) An Introduction to Animal Parasitology (3rd edition). Cambridge: Cambridge University Press, 549 pp. Sprengel G. & Lüchtenberg H. (1991) Infection by endoparasites reduces maximum swimming speed of European smelt Osmerus eperlanus and European eel Anguilla anguilla. Diseases of Aquatic Organisms 11, 31–35. Sproston N.G. & Hartley P.H.T. (1941) The ecology of some parasitic copepods of gadoids and other fishes. Journal of the Marine Biological Association 25, 361–392. Stephens D.W. & Krebs J.R. (1986) Foraging Theory. Princeton, CT: Princeton University Press, 261 pp. Sweeting R.A. (1976) Studies on Ligula intestinalis (L.) effects on a roach population in a gravel pit. Journal of Fish Biology 9, 515–522. Sweeting R.A. (1977) Studies on Ligula intestinalis: some aspects of the pathology in the second intermediate host. Journal of Fish Biology 10, 43–50. Szidat L. (1969) Structure, development and behaviour of new strigeatoid metacercariae from subtropical fishes of South America. Journal of the Fisheries Research Board of Canada 26, 753–786. Taylor R.E.L. & Lott M. (1978) Transmission of salmonid whirling disease by birds fed trout infected with Myxosoma cerebralis. Journal of Protozoology 25, 105–106. Tierney J.F. & Crompton D.W.T. (1992) Infectivity of plerocercoids of Schistocephalus solidus (Cestoda: Ligulidae) and fecundity of the adults in an experimental definitive host, Gallus gallus. Journal of Parasitology 78, 1049–1054.

17 11/02/2003 15:00 Page 243

The role of parasites in fish–bird interactions

243

Tierney J.F., Huntingford F.A. & Crompton D.W.T. (1993) The relationship between infectivity of Schistocephalus solidus (Cestoda) and anti-predator behaviour of its intermediate host, the threespined stickleback, Gasterosteus aculeatus. Animal Behaviour 46, 603–605. Urawa S. (1992) Trichodina truttae Mueller 1937 (Ciliophora, Peritrichida) on juvenile chum salmon (Oncorhynchus keta) – pathogenicity and host–parasite interactions. Fish Pathology 27, 29–37. Van Dobben W.H. (1952) The food of the cormorant in the Netherlands. Ardea 40, 1–63. Videler J.J. (1993) Fish Swimming. London: Chapman & Hall, 260 pp. Walkey M. & Meakins R.H. (1970) An attempt to balance the energy budget of a host–parasite system. Journal of Fish Biology 2, 361–372. Weekes P.J. & Penlington B. (1986) First records of Ligula intestinalis (Cestoda) in rainbow trout, Salmo gairdneri, and common bully, Gobiomorphus cotidianus, in New Zealand. Journal of Fish Biology 28, 183–190. Wyatt R.J. & Kennedy C.R. (1988) The effect of a change in the growth rate of roach, Rutilus rutilus (L.), on the biology of the fish tapeworm, Ligula intestinalis (L.). Journal of Fish Biology 33, 45–57. Wyman R.L. & Walters-Wyman M.F. (1985) Chafing in fishes: occurrence, ontogeny, function and evolution. Environmental Biology of Fishes 12, 281–289.

18 11/02/2003 15:03 Page 244

Chapter 18

Will the explosion of Ligula intestinalis in Rastrineobola argentea lead to another shift in the fisheries of Lake Victoria? J. MARSHALL and I.G. COWX* Hull International Fisheries Institute, University of Hull, Hull, UK

Abstract The fisheries of Lake Victoria have changed dramatically in the past 30 years primarily as a result of overfishing, introduction of Nile perch, Lates niloticus (L.), and cultural eutrophication. The fishery was once dominated by hundreds of species of haplochomines but it is now based on three species: Nile perch, Rastrineobola argentea (Pellegrin) and Oreochromis niloticus (L.). Rastrineobola argentea exploded in the lake as a result of the reduced predation pressure previously imposed by the haplochromines. It is now a major food resource of cormorants, Phalacrocorax spp and the pied kingfisher, Ceryle rudis (L.). In recent years there has been a noticeable increase in the prevalence of the cestode Ligula intestinalis (L.) in R. argentea. The parasite has reduced the reproductive abilities of R. argentea and potentially made the stocks susceptible to collapse. This chapter examines the dynamics of L. intestinalis in Lake Victoria, especially in relation to R. argentea and the piscivorous bird populations, and evaluates whether it is likely to contribute to a demise of the R. argentea stocks. Keywords: behaviour, fisheries, Ligula, Rastrineobola, reproduction.

18.1 Introduction Lake Victoria, at 68 800 km2, is the largest tropical lake in the world, but despite its size, is only approximately 80 m maximum depth, with a mean depth of 40 m (LoweMcConnell 1993). Since the 1950s, Lake Victoria has undergone changes (Bugenyi & Magumba 1996) that have irreversibly altered its trophic structure (Goldschmidt et al. 1993). The fisheries have changed as a result of overfishing, introduction of Nile perch, Lates niloticus (L.), in the late 1950s/early 1960s (Ogutu-Ohwayo 1988; Goldschmidt et al. 1993), and cultural eutrophication. By the 1980s the numbers of Nile perch increased so dramatically that by 1990 it contributed 90% of the total catch (Ligtvoet & Mkumbo 1991). By contrast the endemic species flock of haplochromines collapsed (Witte et al. 1992), with an estimated two-thirds of the haplochromine cichlid taxa disappearing (Lowe-McConnell 1993). The diminishing haplochromine populations, combined with the presence of a new large predator, disrupted the food chains within the lake and one species to prosper was the small pelagic cyprinid Rastrineobola argentea (Pellegrin) *Correspondence: Dr Ian G. Cowx, Hull International Fisheries Institute, University of Hull, Hull HU6 7RX, UK (email: [email protected]).

18 11/02/2003 15:03 Page 245

Impact of Ligula intestinalis on Rastrineobola argentea

245

(Witte et al. 1992). Rastrineobola argentea has become the second most important species in the commercial fishery and the main food of several avian piscivores, including the pied kingfisher, Ceryle rudis (L.), (Wanink et al. 1993; Wanink & Goudswaard 1994) and cormorants, Phalacrocorax spp., which previously fed on the haplochromines. The increased importance of R. argentea to the fishery resulted in considerable attention focusing on the species. In the 1990s, Wanink (1998) reported on the presence of the cestode parasite, Ligula intestinalis (L.), in R. argentea, which he considered was linked through its life cycle to the feeding habits of cormorants, black-backed gull, Larus fuscus, and possibly pied kingfisher. More recently, there have been concerns about the increased prevalence of Ligula in R. argentea, which may have serious implications for the fishery, considering the potential for the cestode to lead to fish population collapse (e.g. Burrough & Kennedy 1979; Wyatt & Kennedy 1988). The aim of this chapter is to examine the biology of R. argentea within Lake Victoria and assess the potential impact of L. intestinalis on the population dynamics of the species, and how it could affect the future of the fishery.

18.2 Material and methods 18.2.1

Data collection

Samples of Rastrineobola argentea (Pellegrin) were collected during August 1999 and February 2000 by the 17.1 m long research vessel MV Ibis. Samples were collected, using a mid-water frame trawl, at locations all over the lake, including open waters, coastal waters, deep and shallow trawls in waters of all the riparian states. The exact position of each site was determined using a hand-held, global positioning system (GPS) and the distance to shore determined from maps. In 1999 a total of 21 sites was sampled, compared with 23 in 2000 (Fig. 18.1). The 2000 survey included more sites from the centre of the lake.

Figure 18.1 Maps to show the location of trawl samples for Rastrineobola in 1999 and 2000

18 11/02/2003 15:03 Page 246

246

Birds, Fish, Habitat interactions

The frame trawl consisted of a funnel of graded mesh of 152 mm to 51 mm, leading to a cod-end of polyamide twine of 5 mm mesh net. Attached to the open end of the netting was a steel frame of 3.5 m  3.5 m, which was towed by the vessel at the required water depth. The depth of the net was set at the position of greatest density of fish in the water column, and controlled by a netsonde detector. The duration of each trawl varied from 10 to 30 min, and trawling was only carried out in daylight hours. The total catch was initially weighed, after which individual species were separated and weighed again to evaluate CPUE per species. Sub-samples of approximately 200–400 g of R. argentea were taken, if the samples were large enough to do so, and preserved in 4% formalin for laboratory analysis. In the laboratory, fish were measured (standard length (SL) to the nearest mm), and sorted into 1 mm length intervals. Individual fish weights could not be determined accurately owing to the imprecise weighing equipment available during the course of this study. Only fish above approximately 40 mm in length could be weighed individually, and for this reason a group of fish of each length class was weighed to the nearest 0.1 g, and average weight per fish was then calculated. If possible, ten random specimens were taken from each 1 mm length class. These were dissected to establish the sex and maturity of each individual. Maturity was based on the status of the gonad following the criteria of Wandera (1999). Fecundity of all mature females was determined by removing and flattening their ovaries on a glass grid to a one egg thick surface layer, and manipulated to form a square. Sub-samples were counted based on the grid squares and extrapolated to the total area cover to obtain an estimate of the total number of eggs per individual. Generally, only females of maturity stages five and six had eggs which could be separated to count them. Every fish from all of the samples was checked for L. intestinalis. Fish that were not already dissected to determine their sex and stage of maturity were cut open along their ventral surface and inspected for the presence of the parasite in the visceral cavity. When L. intestinalis was found, the length of the fish host was noted, as well as its maturity and sex, and, where appropriate, fecundity was estimated. The lengths of the parasite were recorded (nearest mm), and because fish often harboured more than one parasite, the number of L. intestinalis in each fish was recorded.

18.2.2

Data analysis

From the length data, Petersen length–frequency histograms were constructed to assess the population size structure at the time of sampling for each sample site. Mean lengths were also calculated for each site so that the relationship between mean population length and depth and distance from the shore could be determined. The modes for each population were identified using Battacharya’s method on FiSAT, and the mean for each mode established. The catch rate per unit trawl for R. argentea at different stations was correlated against depth and distance from the shore to establish if any disparities existed between population densities in inshore and offshore locations. The length, sex and maturity data were used to determine the length of maturity (LM50), i.e. when 50% of fish are mature, of the populations at the different sample locations.

18 11/02/2003 15:03 Page 247

Impact of Ligula intestinalis on Rastrineobola argentea

247

LM50 was determined by Solver, an iterative function in the Excel program which fits a sigmoid curve through a data set ranging from no mature fish to all mature fish. Mean LM50 values (males and females) for all R. argentea populations in 1999 and 2000 were then compared with the depth and the distance from the shore to determine if any relationship exists. The prevalence (% of fish in each population with L. intestinalis) of L. intestinalis in relation to the size of fish, location and depth was established. These data were analysed to give an indication of the effect of the parasite on the fish populations with respect to maturity status and fecundity. This was achieved by comparing maturity status of parasitised and uninfected R. argentea, within the population. The effect of L. intestinalis on the maturity of R. argentea was only analysed for sites which had a high enough prevalence of L. intestinalis to make the findings meaningful (above 10% prevalence). The relationship between fecundity and length in the presence or absence of L. intestinalis was also examined. The prevalence of L. intestinalis in fish over 40 mm in length was correlated with water depth and the distance from the shore. This was to establish whether a populations’ proximity to shore, and hence its accessibility to avian predators, (some of which are the definitive hosts of L. intestinalis), affects the prevalence of the parasite.

18.3 Results 18.3.1

Distribution of Rastrineobola argentea

Data from both the 1999 and 2000 surveys were combined to show the distribution of fish (g trawl hr1) in relation to water depth and distance from the shore (Fig. 18.2). A wide range of CPUEs was found (15–8800 g trawl hr1), but there was no correlation between abundance and water depth or distance from the shore, suggesting R. argentea is distributed throughout the pelagic zone of the lake. However, variation was found in the size distribution of R. argentea between different locations around the lake (Fig. 18.3). There appears to be a tendency for larger fish to inhabit deeper, or less sheltered, open water environments. For example, at one of the shallowest areas in the CPUE = ⫺1176.3 ln(distance) ⫹ 4285

CPUE (g trawl hr⫺1)

25000

r 2 = 0.0893

CPUE = ⫺106.14 depth ⫹ 6988.4

25000

20000

20000

15000

15000

10000

10000

5000

5000

r 2 = 0.1725

0

0 0

10 20 30 Distance from shore (km)

40

0

20

40 Depth (m)

60

80

Figure 18.2 Relationships between abundance of Rastrineobola argentea and water depth (m) and distance from the shore (km)

18 11/02/2003 15:03 Page 248

248

Birds, Fish, Habitat interactions 20 15 10 5 0

Site U2 n = 117

100

Site U3 n = 320

50 0 15

Site U4 n = 117

10

Site T7 n = 15

20 15 10 5 0

Site T9 n = 114

30 Site T11 n = 374

20 10

5

Frequency

4 3 2 1 0

0

0

40 30 20 10 0

40 30 20 10 0

Site U5 n = 150

30

Site U6 n = 149

20

60

20

0

0 Site T3 n = 376

40

Site T12b n = 421

40

10

60

Site T12a n = 344

20

Site T13 n = 199

15 10

20

5

0

0

40 30 20 10 0

Site T4 n = 226

30

Site K1 n = 201

20 10 0

3

Site T5 n=4

2 1

Site K2 n = 283

1 7 13 19 25 31 37 43 49 55 61 67 73

0

50 40 30 20 10 0

6

Site T6 n = 27

4 2

71

64

57

50

43

36

29

22

8

15

1

0 Standard length (mm)

Figure 18.3 Length frequency distribution of Rastrineobola caught in trawl samples in different region of Lake Victoria in 1999. Code numbers refer to sites in Figure 18.1

18 11/02/2003 15:03 Page 249

Impact of Ligula intestinalis on Rastrineobola argentea

249

north-east sector of the Ugandan part of the lake (30 m deep) two length modes were found at 25 and 32 mm (Fig. 18.3). As water depth increased there was a shift towards larger sizes of fish in the population with fewer smaller individuals (Fig. 18.3). Inshore shallow sites appear to have a greater contribution of juvenile fish, for example, Site T3 in 1999 where the modal mean lengths were 22 and 31 mm (Fig. 18.3).

18.3.2

Reproductive biology

Rastrineobola of both sexes matured (LM50) at around 39 mm, although females showed a tendency to mature at a slightly larger size than males (Table 18.1), but the differences were not significant (P  0.05). No differences were found between the size of maturity in the different survey years or in the three countries (Table 18.1). There was some indication to suggest that LM50 of R. argentea increased with increasing depth of the water column up to 40 m deep, although this was based on few observations (r2  0.49 and 0.68 for 1999 and 2000 respectively). No such trend was found for distance from the shore. A high proportion of the uninfected Rastrineobola in the samples was at maturity stages 4, 5 and 6, indicating they were in advanced stages of maturation prior to spawning (Fig. 18.4). Although ripe fish are found throughout the year (R. Tumwebaze, Table 18.1 Mean size ( variance) at first maturity in different regions of Lake Victoria in 1999 and 2000 1999 survey

2000 survey

Site

Female (LM50)

Male (LM50)

Female (LM50)

Male (LM50)

Uganda Tanzania Kenya Whole lake

– 39.9  0.7 35.5  2.4 38.8  5.0

38.7  2.6 39.3  1.4 35.1  0.1 38.4  3.8

39.0  1.7 39.6  2.9 38.3  1.1 39.1  1.6

38.5  1.7 38.8  1.8 38.5  2.0 38.7  1.5

Figure 18.4 Effect of Ligula presence on maturation of Rastrineobola argentea

18 11/02/2003 15:03 Page 250

250

Birds, Fish, Habitat interactions

personal communication), these fish were probably preparing for spawning in the peak breeding season around October/November. Data for fecundity were pooled with respect to year and site, as a comparison between sites would have proved difficult due to limited data for individual sites. The main point to note was that fecundity increased with fish length (Fig. 18.5).

18.3.3

Ligula infestation

The prevalence of L. intestinalis increased exponentially for fish greater than 40 mm SL (Fig. 18.6). Although fish 40 mm SL were infected, prevalence was low partly

Fecundity (No. eggs per fish)

2000

1500

1000

With Ligula Without Ligula

500

0 30

35

40

45

50

55

60

65

Fish length (mm)

Figure 18.5 Fecundity of Rastrineobola argentea with and without Ligula 100 Prevalence (% of fish infested)

90 80 70 60 50 40 30 20 10 0 0

10

20

30

40

50

60

70

Length (mm)

Figure 18.6 Relationship between fish length and prevalence of Ligula

80

18 11/02/2003 15:03 Page 251

Impact of Ligula intestinalis on Rastrineobola argentea

251

because the parasite was difficult to locate due to its small size. Marked increases in Ligula size and number were observed with increasing fish length (Figs 18.7 and 18.8). There was no relationship between the prevalence of L. intestinalis in fish greater than 40 mm in length, and either depth or distance from the shore, in either 1999 or 2000. Ligula intestinalis had a marked affect on the breeding cycle of R. argentea (Fig. 18.4). The proportion of the population in breeding state (stage 4, 5 and 6) was considerably reduced, indeed only the occasional fish was in breeding condition compared with uninfected fish of the same size. The percentage males and females

L. intestinalis length (mm)

80

Ligula length = 1.093 fish length  17.358 r 2 = 0.6328

60

40

20

0 0

10

20

30

40

50

60

70

80

Fish length (mm)

Figure 18.7 Relationship between size of Ligula intestinalis and fish length

Intensity (No. L. intestinalis per fish)

9

Intensity = 0.2874 exp (0.0333) fish length r 2 = 0.4582

8 7 6 5 4 3 2 1 30

35

40

45

50

55

60

65

Length (mm)

Figure 18.8 Relationship between number of Ligula intestinalis and fish length

70

18 11/02/2003 15:03 Page 252

252

Birds, Fish, Habitat interactions

breeding dropped from 25 and 33% in uninfected fish to 11 and 10% respectively for parasitised fish. Infestation by L. intestinalis reduced the fecundity of R. argentea in all sizes (Fig. 18.5), although the differences were not significant, largely due to the wide variation in fecundity of Rastrineobola with size.

18.4 Discussion 18.4.1

Distribution of Rastrineobola argentea

Based on catch per unit effort (CPUE) data, R. argentea appears to be a species widely distributed in the pelagic waters of Lake Victoria. Abundance decreased slightly with distance from the shore or depth, but this was not statistically significant. The location of spawning and nursery areas for the species is somewhat unclear. Larger individuals were generally found in offshore, deeper waters but juveniles were found in both inshore and offshore, although their relative abundance at the latter was lower. Mature females were also found in both inshore and offshore waters. By contrast, Wanink et al. (1998) found the abundance of mature females and juveniles was higher in inshore waters, 2 km from the shore, and suggested that the species spawns in shallow inshore waters. The differences observed between the studies were the result of sampling strategy because the present study used a mid-water trawl and did not fish in waters less than 14 m deep within 2 km of shoreline, and Wanink et al. (1998) only sampled up to 5 km from the shore line. These data suggest that R. argentea is an inshore spawning species, as opposed to the earlier belief that the fish spawns in a pelagic environment (Graham 1929; Greenwood 1966; Wandera 1993), but there is either active or passive dispersion offshore. Nevertheless, offshore spawning cannot be ruled out. The presence of an inshore spawning stock is consistent with optimal nursery conditions for growth of juveniles in an area where primary productivity and zooplankton abundance (Cowx & Tweddle 1999) were highest. By moving offshore the species may be exploiting reduced predation pressure because abundance of Nile perch and other predators is lower in this zone.

18.4.2

Prevalence of L. intestinalis in R. argentea

The prevalence of L. intestinalis in R. argentea increases exponentially in individuals 40 mm. One of the major explanations for this upsurge may be a change in diet from small to large zooplankton as the fish grows. Ligula intestinalis larvae arguably have numerous intermediate hosts, and one is likely to be a zooplankton copepod (Wyatt & Kennedy 1988). Infection in larger individuals is probably linked to only larger R. argentea being able to ingest the primary host zooplankton. The absence of any discernible trend between L. intestinalis infestation and depth or distance from the shore is possibly linked to the dispersal mechanism of R. argentea,

18 11/02/2003 15:03 Page 253

Impact of Ligula intestinalis on Rastrineobola argentea

253

distribution of the intermediate zooplankton host and/or foraging behaviour of the birds. Prevalence of L. intestinalis was expected to be at its highest levels close to the shore since this is where its definitive avian hosts, cormorants and black-backed gull (Wanink 1998), most actively feed (Dence 1958). As R. argentea exploded in Lake Victoria (Wanink 1998), avian predators, such as the cormorants, gulls and pied kingfisher, switched their diet from haplochromines to R. argentea (Wanink et al. 1993; Wanink & Goudswaard 1994). However, R. argentea is smaller than most haplochromine species, thus offering a lower reward to the avian piscivores (Wanink & Goudswaard 1994), possibly forcing birds to fly further afield in search of larger R. argentea. The assumption that birds feed on inshore fish is therefore possibly invalid, as it is necessary for the predator to gain as much benefit from its prey whilst minimising search and handling effort (Holmes 1976). In short, avian piscivores probably fly further to feed on larger individuals, as long as the trade-off between effort expended and benefit gained is in its favour, and the parasite life cycle will, therefore, be completed both inshore and offshore throughout the bird’s predator range. Alternatively, the lack of any relation between the parasite prevalence and the location of the fish can be attributed to the assumption that R. argentea’s nursery areas are located in the littoral areas around the shoreline (Wanink 1999) and thus disperse offshore as they get larger. It is possible that the fish host is infected in the inshore areas (Dence 1958; Holmes 1976) when it is a juvenile, or when moving towards a more pelagic habitat. In addition, R. argentea are constantly moving to and from the shore line to breed, so L. intestinalis prevalence and distance from the shore is likely to be constantly changing. Infestation by Ligula also exhibited a subtle effect of the vertical distribution of R. argentea. The normal diel behaviour of R. argentea is to feed near the surface at night and aggregate in the lower water column around the oxycline during the day. However, Wanink (1993, 1998) observed that parasitised fish were commonly found in the surface waters throughout the day and night making them more vulnerable to predation from avian piscivores. Holmes and Bethel (1972), and Kramer et al. (1983) indicated that parasites can change the behaviour of their intermediate hosts to increase their vulnerability to predation by the final host, but often the mechanism by which it is achieved is unknown. This change in behaviour in R. argentea is probably brought about by an increased oxygen demand imposed on the fish by the parasite (Barber, Chapter 17).

18.4.3

Impact of L. intestinalis infestation on R. argentea stock dynamics

Infestation by Ligula has important implications for reproduction in R. argentea because the size at which infestation becomes apparent, possibly because the cestode starts to grow and becomes visible, and the size of first maturity coincide. In the presence of L. intestinalis there is a marked retardation in the maturation of the ovary and fecundity declines. Numerous explanations have been proposed as to why the presence of L. intestinalis effects maturation, including the suggestion that the parasite produces a non-steroid

18 11/02/2003 15:03 Page 254

254

Birds, Fish, Habitat interactions

substance which inhibits gonadal development or it may produce a stress response in the fish host (Arme et al. 1982). In addition, L. intestinalis absorbs its nutrients from its host, thus reducing the energy that R. argentea can afford for reproduction (Campbell 1996). Furthermore, L. intestinalis parasitism incurs a reduction in the space available in the visceral cavity (Dence 1958). As the fish length increases so does the average number of L. intestinalis present in each fish host, as well as the average size of the parasite. Subsequently, the space available for maturation of the gonads must be reduced, which is again detrimental to R. argentea reproductive potential. The increase in the intensity and average length of L. intestinalis must also logically exacerbate any effect it has, other than space availability, on its fish host. Importantly, it is the bigger fish which are crucial for reproductive output that are most affected by L. intestinalis, so an epidemic of the parasite may seriously affect the R. argentea fishery. Although there is no one proven cause, whatever mechanism L. intestinalis uses to inhibit reproduction in the fish host, there is an apparent consensus, which is supported by this study, suggesting that L. intestinalis has a very real effect on R. argentea reproduction. If the effect on reproduction by L. intestinalis is as severe as suggested, the implications for the future viability of the R. argentea fishery are serious. The species supports an extremely important fishery, which landed some 150 000 t in Lake Victoria in 2000. The exploitation rate is very high, with a fishing mortality (F) between 3 and 4 yr1. The species has compensated for the heavy fishing pressure in the past by its propensity for rapid reproduction. A drop in the reproductive capacity could adversely affect the stock dynamics resulting, in the worst case scenario, of a collapse in the fishery. This has been observed in other Ligula-infected fish populations, e.g. the roach, Rutilus rutilus (L.), population at Slapton Ley (Burrough & Kennedy 1979). This collapse allowed the expansion of a sympatric rudd, Scardinius erythrophthalamus (L.), population, which had previously been suppressed by the roach population. If such a scenario occurs in Lake Victoria, the impact on the estimated one million people dependent on the fishery could be catastrophic. It also raises the question of whether the demise of the R. argentea stock would lead to a further recovery of the pelagic haplochromine species flock, thus resulting in further shifts in the Lake Victoria ecosystem, perhaps towards its original complexity. Whatever the outcome, there is no management intervention that could address the potential impact of L. intestinalis in the R. argentea stock. There is no practical mechanism for breaking the Ligula life cycle as a reduction in the numbers of the definitive hosts (cormorants and gulls) is impossible given the tens of thousands of birds that inhabit the lake. Consequently, the natural cycle of events that transpires will have to be followed through and any changes in the fishery for R. argentea will have to be managed on an adaptive basis.

References Arme C., Griffiths D.V. & Sumpter J.P. (1982) Evidence against the hypothesis that the plerocercoid larva of Ligula intestinalis (Cestoda: Pseudophyllidea) produces a sex steroid that interferes with host reproduction. Journal of Parasitology 68, 169–171.

18 11/02/2003 15:03 Page 255

Impact of Ligula intestinalis on Rastrineobola argentea

255

Bugenyi F.W.B. & Magumba K.M. (1996) The present physicochemical ecology of Lake Victoria, Uganda. In T.C. Jonson & E.O. Odada (eds) The Limnology, Climatology and Paleoclimatology of the East African Lakes. Amsterdam: Gordon & Breach, pp. 141–153. Burrough R.J. & Kennedy C.R. (1979) The occurrence of natural alleviation of stunting in a population of roach, Rutilus rutilus (L.). Journal of Fish Biology 15, 93–109. Campbell N.A. (1996) Biology (4th edn). California: Benjamin/Cummings Publishing Company, Inc. Cowx I.G. & Tweddle D. (eds) (1999) Report on Fourth FIDAWOG Workshop held at Kisumu, 16 to 26 August, 1999. Lake Victoria Fisheries Research Project, Phase II, Technical Document No. 7. Jinja: LVFRP, 211 pp. Dence W.A. (1958) Studies on Ligula-infected common shiners (Notropis cornutus frontalis Agassiz) in the Adirondacks. Journal of Parasitology 44, 334–338. Goldschmidt T., Witte J. & Wanink J. (1993) Cascading effects of the introduced Nile perch on the detritivorous/phytoplanktivorous species in the sublittoral areas of Lake Victoria. Conservation Biology 7, 686–700. Graham M. (1929) The Victoria Nyanza and its Fisheries. London: Waterlow & Sons. Greenwood P.H. (1966) The Fishes of Uganda (2nd edn). Kampala: Uganda Society. Holmes J.C. (1976) Host selection and its consequence. In C.R. Kennedy (ed.) Ecological Aspects of Parasitology. Amsterdam: North Holland, pp. 21–39. Holmes J.C. & Bethel W.M. (1972) Modification of intermediate host behaviour by parasites. In E.U. Canning & C.A. Wright (eds) Behavioural Aspects of Parasite Transmission. London: Academic Press, pp. 123–149. Kramer D.L., Maley D. & Bourgeois R. (1983) The effect of respiratory mode and oxygen concentration on the risk of aerial predation in fishes. Canadian Journal of Zoology 61, 53–65. Ligtvoet W. & Mkumbo O.C. (1991) A pilot sampling survey for monitoring the artisanal Nile perch (Lates niloticus) fishery in southern Lake Victoria (East Africa). In I.G. Cowx (ed.) Catch Effort Sampling Strategies. Their Application in Freshwater Fisheries Management. Oxford: Fishing News Books, Blackwell Science, pp. 349–360. Lowe-McConnell R. (1993) Fish faunas of the African Great Lakes: Origins, diversity, and vulnerability. Conservation Biology 7, 634–643. Ogutu-Ohwayo R. (1988) Reproductive potential of Nile perch, Lates niloticus L., and the establishment of the species in Lakes Kyoga and Victoria (East Africa). Hydrobiologia 162, 193–200. Wandera S.B. (1993) Seasonal abundance, vertical migration and the fishery of dagaa Rastrineobola argentea in the Ugandan waters of Lake Victoria. In B.E. Marshall & R. Mubamba (eds) Symposium on Biology, Stock Assessment and Exploitation of Small Pelagic Fish Species in the African Great Lakes Region. FAO, CIFA Occasional Paper 19, pp. 257–270. Wandera S.B. (1999) Reproductive biology of Rastrineobola argentea (Pellegrin) in the northern waters of Lake Victoria. In I.G. Cowx & D. Tweddle (eds) Report on fourth FIDAWOG workshop held at Kisumu, 16 to 26 August 1999. Lake Victoria Fisheries Research Project, Phase II, Technical Document No. 7. Wanink J.H. (1999) Prospects for the fishery on the small pelagic Rastrineobola argentea in Lake Victoria. Hydrobiologia 407, 183–189. Wanink J.H., Berger M.R. & Witte F. (1993) Kingfishers at fishburger queens: Eating Victorian fast food on the wing. Annales Musée Royal de l’Afrique Centrale (Zoologie) 268, 319–326. Wanink J.H. & Goudswaard K. (1994) Effects of Nile perch (Lates niloticus) introduction into Lake Victoria, East Africa, on the diet of pied kingfishers (Ceryle rudis). Hydrobiologia 280, 367–376. Wanink J.H, Goudswaard K. & Berger M.R. (1998) The sustainable role of the small pelagic cyprinid Rastrineobola argentea as a major resource in the ecosystem and the fishery of Lake Victoria. In W.L.T. van Densen & M.J. Morris (eds) Lacustrine Fish Communities in SE-Asia and Africa: Ecology and Exploitation. Cardigan: Samara Publishing, pp. 295–310.

18 11/02/2003 15:03 Page 256

256

Birds, Fish, Habitat interactions

Witte F., Goldschmidt T., Goudswaard P.C., Ligtvoet W., van Oijen M.J.P. & Wanink J.H. (1992) Species extinction and concomitant ecological changes in Lake Victoria. Netherlands Journal of Zoology 42, 214–232. Wyatt R.J. & Kennedy C.R. (1988) The effects of a change in the growth rate of roach, Rutilus rutilus (L.), on the population biology of the fish tapeworm Ligula intestinalis (L.). Journal of Fish Biology 33, 45–57.

19 11/02/2003 15:05 Page 257

Section III Mitigation measures

This page intentionally left blank

19 11/02/2003 15:05 Page 259

Chapter 19

The potential for using fish refuges to reduce damage to inland fisheries by cormorants, Phalacrocorax carbo I.C. RUSSELL* and P.J. DARE Centre for Environment, Fisheries and Aquaculture Science, Lowestoft, Suffolk, UK

H.V. McKAY Central Science Laboratory, Sand Hutton, York, UK

S.J. IVES Centre for Environment, Fisheries and Aquaculture Science, Lowestoft, UK

Abstract Recent research in the UK has concluded that cormorants, Phalacrocorax carbo (L.) can have a major impact on some freshwater fisheries. There is thus a need for effective management measures that will reduce the interaction between these predators and their prey, and reduce the level of impact at such sites. A recent comprehensive review of possible management measures indicated that underwater refuges could influence cormorant foraging behaviour, and reduce losses and levels of damage to fish. It also recommended further studies to test the effects of refuge design on the behaviour of different species and sizes of freshwater fish, and the number and behaviour of cormorants visiting a site. This chapter reviews the information on cormorant and fish behaviour, the response of fish to predators, and the design and application of artificial structures in fishery management, as part of a new investigation to assess the efficacy, practicality and acceptability of using fish refuges to limit predation by cormorants. Keywords: cormorant, fish refuge, foraging, freshwater fish, predation.

19.1 Introduction Interactions between fish and cormorants, Phalacrocorax carbo (L.), currently command a high profile throughout many parts of Europe due to rapidly increasing populations of these birds in many countries. In England and Wales, cormorants have extended their range inland from coastal areas, and now over-winter and feed at many inland water sites (Marquiss & Carss 1994; Russell et al. 1996). The population of cormorants wintering in Britain has increased steeply over the last 25 years (Wernham et al. 1999). In addition, over the last 18 years cormorants have started to develop inland breeding *Correspondence: Ian Russell, Centre for Environment, Fisheries and Aquaculture Science, Lowestoft Laboratory, Pakefield Road, Lowestoft, Suffolk NR33 0HT, UK (email: [email protected]). © British Crown Copyright, 2003.

19 11/02/2003 15:05 Page 260

260

Mitigation measures

colonies partly, it is thought, through colonisation from the continent of the typically tree-nesting sub-species Phalacrocorax carbo sinensis (L.) (Sellers et al. 1997). Numbers of inland breeding colonies have risen sharply over this time (Hughes et al. 2000). This range extension and increase in numbers of birds has increasingly brought cormorants into conflict with inland fisheries. As a result, the UK Government commissioned a number of studies into bird predation at fisheries in England and Wales (Russell et al. 1996; Feltham et al. 1999; Hughes et al. 1999; McKay et al. 1999; Wernham et al. 1999). Feltham et al. (1999) concluded that the impact of cormorants can be significant at some sites, although on a national scale the impact appears to be less serious (Diamond, Aprahamian & North, Chapter 4). It is also recognised that where stocks of catchable fish are significantly reduced by predation this can have potentially serious economic implications (Russell et al. 1996). In such circumstances, management action may be appropriate to limit the impact that birds might have on stocks of fish and associated fisheries. Effective management measures that will reduce the interaction between these predators and their prey, and reduce the level of impact at such sites, are thus required. In England and Wales, cormorants are protected under the Wildlife and Countryside Act (WCA) 1981, which implements the 1979 European Community Directive on the Conservation of Wild Birds (EEC/79/409). The WCA makes provision for killing or taking birds under licence for the purpose of preventing serious damage to fisheries. Licences only allow limited numbers of birds to be killed to reinforce the effects of other scaring methods, and where there are no other effective and practical alternatives. A wide range of potential alternative methods exists. A recent comprehensive review of possible management measures assessed a range of potential techniques for their effectiveness, practicality, cost and public acceptability (McKay et al. 1999). As part of this investigation, a preliminary field study indicated that underwater refuges may influence cormorant foraging behaviour and reduce levels of damage to fish (McKay et al. 1999; McKay et al., Chapter 20 in this volume). Results from the above project, although not conclusive, were sufficiently encouraging that a further study was recommended to test the effects of refuge design on the behaviour of different species of freshwater fish, and the number and behaviour of cormorants visiting a site. A further investigation to assess the efficacy, practicality and acceptability of using such refuges is now being carried out. In an unmanaged situation, the interaction between a predator and its prey will tend to oscillate around an equilibrium. For refuge structures to work, these will effectively have to shift the predator–prey balance in favour of the fish, either making fish less vulnerable to capture or making the foraging conditions less favourable for cormorants. Efforts to engineer such an effect will need to be based on an understanding of the foraging behaviour and prey preferences of cormorants, and the ecology and behaviour of the appropriate fish species. This chapter reviews relevant information on the behaviour of cormorants and freshwater fish, the response of fish to predators, and the application of artificial reef and refuge structures in fisheries as a basis for assessing whether similar structures may be appropriate for use in freshwater systems to reduce levels of predation by cormorants on freshwater fish.

19 11/02/2003 15:05 Page 261

Fish refuges to reduce cormorant damage to fisheries

261

19.2 Cormorant behaviour 19.2.1

Foraging behaviour

Cormorants are agile feet-propelled pursuit-divers (Cramp & Simmons 1977), which surface between dives to breathe. Fish are grasped with the hooked bill, usually from underneath, and just behind the gills (van Dobben 1952). Fish-eating birds generally feed during intensive bouts of diving, typically comprising a series of dives interspersed by brief periods of resting on the surface (inter-dives). At the end of each divebout, birds may spend time in feather maintenance before loafing, either on the water or on shore (Hughes et al. 1999). Cormorants have previously been regarded mainly as bottom-feeders in shallow waters (10 m depth) (van Dobben 1952; Cramp & Simmons 1977). However, more recent studies indicate that bottom-living species have been replaced in the diet by pelagic shoaling species in many areas (Voslamber 1988; Platteeuw et al. 1992; Veldkamp 1994; Dirksen et al. 1995; van Eerden & Voslamber 1995; Suter 1997) suggesting a change in fishing behaviour as a result of changes in fish stocks (Suter 1997). Pelagic shoaling species roach and perch also tend to predominate in the diet of birds shot at inland fisheries in England and Wales (CEFAS, unpublished data). Nonetheless, it is known that birds are capable of feeding at depths well in excess of 10 m (Suter 1991; Callaghan et al. 1998; Kato et al. 1998; Grémillet et al. 1999), as well as in shallow riverine ‘glides’ of 30 cm in depth (Kennedy & Greer 1988). Birds have also been shown to execute both benthic and pelagic dives during foraging bouts (Grémillet et al. 1998). Cormorants thus appear to be opportunists capable of foraging at various depths; since cormorants are almost neutrally buoyant, they may prefer to hunt at shallow depths given equal prey density (van Dobben 1952). Wilson & Wilson (1988) also reported that the time spent searching for prey is independent of water depth. Cormorants mainly feed independently of each other, but may also feed socially on mid-water shoaling fish (Marquis & Carss 1994). Suter (1991) suggested that flockfeeding is a recent adaptation to enable birds to exploit rich stocks of shoaling cyprinids in increasingly eutrophic water bodies in Europe. This behaviour is also common among cormorants feeding on the IJsselmeer, The Netherlands (Voslamber & van Eerden 1991). Hughes et al. (1999) observed flock-feeding on almost a daily basis at stillwater sites in England (in groups ranging from 3 to 480 birds) but never at river sites. They reported that the proportion of birds flock-feeding, as opposed to solitary feeding, increased with the number of feeding birds present and occurred mainly at periods of intense feeding activity in the early morning, and on overcast days. Hughes et al. (1999) also investigated the features that make habitats attractive as cormorant feeding sites, based on a stratified random sample of 80 inland stillwaters in UK. The results suggested that there were more cormorants at larger sites with less waterborne disturbance in eastern and south-central England and on sites close to night roosts. Further analysis indicated that numbers were greater at more open sites with lower surrounding land gradients and more indented shorelines. Similar conclusions were reached in a previous study of stillwater fisheries in England and Wales

19 11/02/2003 15:05 Page 262

262

Mitigation measures

(Callaghan et al. 1994; Callaghan et al. 1998). On the River Wye, England, feeding activity of cormorants was greatest when water levels were low, suggesting that the ability of cormorants to locate and catch prey was enhanced with reduced water speed and lower turbidity (Hughes et al. 1999). Birds are reported to hunt successfully in both clear and turbid waters. However, cormorants have been observed to abandon feeding sites when flow rates and turbidity increase (Hughes et al. 1999; Feltham et al. 1999). In addition, diving behaviour may be affected by weather. Hughes et al. (1999) found dive-bout duration increased with cloud cover, suggesting that reduced light levels made the detection of fish more difficult, increasing the time required to locate and catch prey. It is generally accepted that cormorants detect prey visually in most circumstances. Laboratory experiments in Israel indicated that cormorants can detect fish visually at a distance of at least 1.4 m in clear water (Strod et al. 1999). However, it is not known how birds locate fish resting on muddy seabeds in very low visibility. King et al. (1998) also reported an observation of nocturnal flock-feeding, under conditions of no moonlight, in the double-crested cormorant, Phalacrocorax auritus (L.). These observations suggest that prey may also be detected by touch. The swimming capabilities of individual cormorants are variously estimated at between 1 and 4 m s1. In the Netherlands, hunting cormorants were reported to be capable of swimming speeds of 1.4 m s1 (Voslamber & van Eerden 1991). Schmid et al. (1995) found that swimming speeds of four captive cormorants varied from 0.9 to 2.2 m s1, averaging 1.51 m s1. Junor (1969, cited in Cramp and Simmons 1977) measured a mean swimming speed of 1.5 m s1 for Phalacrocorax carbo carbo and stated that Phalacrocorax carbo sinensis could theoretically swim faster (1.8–2.0 m s1). Wilson and Wilson (1988) found that cormorants had a mean dive speed of about 1 m s1 and foraged along the seabed at speeds of up to 4 m s1; foraging speed was dependent on bottom topography and was positively correlated with depth. Dive duration is related to water depth (Wilson & Wilson 1988; Traylor et al. 1989; Wanless et al. 1991), although there is wide variation between species and individuals (Stonehouse 1967). Sea-foraging Phalacrocorax carbo dives lasted for 16–152 s, averaging 40 s (Grémillet et al. 1999). Diving behaviour may also be affected by factors such as prey behaviour and weather. Hughes et al. (1999) showed that dive duration was longer in the winter than the summer. This was thought to reflect seasonal changes in the distribution of fish, which tend to move into deeper water during the winter. Dive-bout duration was also affected by cloud cover, with the duration of bouts increasing with increased cloud cover, presumably as a result of reduced light levels making the detection of fish more difficult. The number of dives executed per foraging bout is also highly variable. Hughes et al. (1999) reported a range of 1 to about 80 dives per bout, and indicated that dive-bouts on rivers lasted longer and contained more dives than those at still waters. For radio-tagged Phalacrocorax carbo foraging in the sea, 2–320 dives per trip, average 42, were recorded (Grémillet et al. 1999). The number of dives is thought to depend on foraging success rate. The extent to which cormorants swallow fish underwater (as opposed to surfacing prior to swallowing) is uncertain, although there is little doubt that birds are capable of this. This is important when assessing, on the basis of bird observations, how many

19 11/02/2003 15:05 Page 263

Fish refuges to reduce cormorant damage to fisheries

263

fish a foraging cormorant consumes, and may also be an important consideration with regard to the potential use of fish refuges. Grémillet et al. (1998) reported that most prey items of both cormorants and shags, Phalacrocorax aristotelis (L.), are quickly handled and swallowed underwater, with only larger items being brought to the surface. Hughes et al. (1999) estimated that the percentage of cormorant prey swallowed underwater may be as high as 74%, on the basis of predictions from bioenergetics estimates (from detailed observations of the behaviour of individual birds). In addition, captive P. c. sinensis in China were observed to ingest several fish at once underwater at depths of 1 m during a single dive (P. Sayers & S. Axford, personal communication, reported in McKay et al. 1999). However, during observations of feeding birds in north-west England, Davies & Feltham (1994, 1995, 1996) reported that, in shallow water, even small cyprinid fry are regularly ingested at the surface. They also indicated that observations of the number of fish caught by cormorants in a river closely corresponded to the number of fish found in their stomachs when they were shot. It thus appears likely that the extent to which underwater feeding occurs will vary for different species and sizes of fish, and between sites. The extent of such variation and the factors affecting it are poorly understood. Few studies have considered underwater habitat complexity, for example, vegetation cover, in relation to foraging by cormorants. However, it is likely that underwater cover and visibility affects the ease by which prey are located and consumed. On fresh waters, cormorants are said to avoid widespread floating vegetation or obstructive emergent growth (Cramp & Simmons 1977). It is not clear whether weeds hinder birds underwater, at the surface, or during landing and take-off, or whether their presence reduces foraging efficiency. All these effects might apply. However, Hebshi (1998) reported that cormorants (in marine habitats) prefer rocky substrates with good kelp cover, presumably because fish density is higher than in more open sites. It is therefore possible that some trade-off between visibility and prey availability might apply – if fish aggregate in weedy areas, it may still be worth foraging there despite longer prey location times. Cormorants are normally diurnal feeders and forage especially in the early morning (van Dobben 1952; Feare 1988; Martucci & Consiglio 1991; Davies & Feltham 1994, 1995, 1996; McKay et al. 1998). However, at some sites, peaks of activity occur twice a day: once in the morning and again in the afternoon (Builles et al. 1986; Im & Hafner 1984). Hughes et al. (1999) also reported a second peak in foraging activity during summer evenings.

19.2.2

Cormorant diet

The cormorant is an opportunistic feeder and most studies on prey selection have found that different species of fish are consumed in relation to their abundance (van Dobben 1952; Pearson 1968; Pilon et al. 1983; Im & Hafner 1984). However, in some situations they appear to select the most abundant prey types (Platteeuw et al. 1992; Marteijn & Dirksen 1991), or shoaling species (Suter 1997). Cormorant diet can also vary spatially and temporally (e.g. Marquiss et al. 1998) and with habitat ‘type’ (Suter 1997; Russell et al. in press). In Sweden, the diet of cormorants varied between months

19 11/02/2003 15:05 Page 264

264

Mitigation measures

within the same area and between different areas (Lindell 1997). Furthermore, samples of pellets collected only five days apart showed large variation in the relative importance of prey species. A wide range of fish species, marine, brackish and freshwater, have been recorded in the diet of cormorants feeding in north-western Europe (see review in Russell et al. 1996). For birds feeding at inland sites in the UK, more than 20 species, almost exclusively fresh water, have been recorded (Russell et al. 1996). However, in most instances diet is dominated by a few species: commonly shoaling pelagic species such as roach, Rutilus rutilus (L.), perch, Perca fluviatilis (L.), and, to a lesser extent, bream, Abramis brama (L.), in stillwater fisheries and lowland rivers; stocked rainbow trout, Oncorhynchus mykiss (Walbaum), and brown trout, Salmo trutta (L.), together with roach and perch, in stillwater put-and-take trout fisheries; and wild salmonid species in more upland rivers (Russell et al. 1996; Marquiss et al. 1998; Hughes et al. 1999; Feltham et al. 1999; Russell et al., in press; CEFAS, unpublished data). Cormorants are also capable of consuming a wide size range of fish, although prey items are commonly in the range 10–30 cm for cyprinids and perch, and 5–40 cm for trout species (see review in Russell et al. 1996). More recent field observations at study sites in lowland England (Feltham et al. 1999) were in broad agreement: Cyprinids Perch Rainbow trout Cyprinids Cyprinids

5–20 cm 5–25 cm 20–45 cm 10–20 cm 5–40 cm

River Trent River Trent Nottingham. Lake River Ribble Lancashire. Reservoir

At Grafham Water, central England, the main prey was rainbow trout, with some brown trout and roach, and a few other coarse fish species (Hughes et al. 1999). Of fish brought to the surface and eaten, 65% were estimated to be 7.5 cm, 25% were 7.5–15 cm, and 10% were 15 cm. Overall, therefore, cormorants at many inland fisheries in the UK tend to take predominantly 5–25 cm coarse fish, but take larger salmonids (often up to 45 cm) where these are available at put-and-take fisheries. The larger size of trout is, at least in part, thought to result from many of these fish being stocked from hatcheries at lengths of 30 cm; these fish are often regarded as being relatively naïve and more vulnerable to predation than wild fish. There is indirect evidence that cormorant density is mainly determined by fish abundance, although the difficulty in accurately assessing fish stocks means that this relationship is not always apparent. In Switzerland, cormorant density on lakes was strongly positively correlated with trophic status (and hence productivity) and with the commercial landings of percids and cyprinids (Suter 1994, 1995). Also, cormorant numbers declined following both natural (Suter 1995) and experimental (Dirksen et al. 1995) reductions in fish numbers. Grémillet & Wilson (1999) modelled the influence of water temperature, dive depth, foraging techniques and prey availability on the energetic costs of diving, prey search time, daily food intake and survival in foraging, non-breeding cormorants. Their model predicts that cormorant foraging parameters are most strongly influenced by

19 11/02/2003 15:05 Page 265

Fish refuges to reduce cormorant damage to fisheries

265

prey availability, so that even a limited reduction in prey density makes birds unable to balance energy needs and may thus limit their influence on prey stocks. Their results suggest that the success of the cormorant’s foraging strategy is dependent upon finding feeding sites with high densities of prey, and that cormorants will abandon a site when the prey density falls below a threshold level. For fish refuges to be effective, therefore, it is anticipated that these will need to reduce the ‘available fish density’ of a site, thus making it less attractive to cormorants.

19.3 Factors likely to affect the vulnerability of freshwater fish to predation by cormorants The distribution, life history features and behaviour of UK freshwater fish species vary markedly. These factors have been reported in numerous reference books and other commissioned reports (e.g. Winfield & Nelson 1991; Maitland & Campbell 1992; Mann & Winfield 1992; Giles 1994; Lucas et al. 1998; Cowx 2001a). A detailed review of these factors is beyond the scope of this chapter, but certain key factors are summarised below where these appear to be particularly relevant to assessing the mechanisms of interaction between cormorants and fish, and the possible application of artificial refuges. In this regard, factors such as fish distribution, the extent to which different species use habitat features and natural ‘structure’, social behaviour, the extent of diel movements, swimming speed, and the behaviour of fish in relation to other fish predators, may be the most relevant. The distribution of freshwater fish and the level of overlap with cormorants will affect which fish species are most likely to be subject to cormorant predation. There is marked variation in the fish species present in the river systems in England and Wales, with species diversity increasing with decreasing latitude and decreasing altitude. Thus, fish communities are dominated by relatively small numbers of species (primarily salmonids) in many of the rivers in northern England and in those draining more upland areas on the west coast and in Wales. By contrast, more diverse fish communities (predominantly cyprinids) are present in most of the rivers towards the south and east of England and in the lower reaches of some rivers elsewhere. Still waters occur throughout the country and are estimated to account for almost 80% of the freshwater habitat area in Great Britain overall (Smith & Lyle 1979). Stillwater sites support both coarse fisheries and trout (game) fisheries (including put-and-take), and many sites are managed and have elevated fish densities due to stocking (North 2002). Cormorants are present throughout the country at various habitat types, but the highest numbers of birds are observed in south and east England (Kirby et al. 1995; Hughes et al. 1999) at stillwaters rather than riverine sites (Wernham et al. 1999), and at sites with high prey densities (Grémillet & Wilson 1999). As a consequence, most conflicts between cormorants and fisheries relate to coarse fisheries, both still water and riverine, and stillwater put-and-take trout fisheries. Fish vary in their susceptibility to predators (e.g. Matkowski 1989) due to factors ranging from their size and shape (de Nie 1995) to their ‘use’ of aquatic habitat features (e.g. Lovvorn et al. 1999). Many species commonly use underwater cover as a

19 11/02/2003 15:05 Page 266

266

Mitigation measures

defence against predators (e.g. Tabor & Wurtsbaugh 1991; Gotceitas & Brown 1993) and differences in affinity for cover affect a fish’s vulnerability to predation (Copp 1992; Mikheev et al. 1994). The use of weed cover and other submerged structures by many coarse fish species is widely regarded as an adaptation to reduce the constant risk of predation by piscivorous fish (e.g. Savino & Stein 1989a; Tonn et al. 1992; Persson 1993; Persson & Eklöv 1995; Eklöv & Persson 1996; Jacobsen & Perrow 1998). However, such refuging behaviour tends to restrict the spatial use a species makes of the available food and other resources, and represents a trade-off between the benefits of predator avoidance and the cost of lost feeding opportunities (Diehl 1988; Dill 1987; Dill & Gillett 1991; Sih 1992, 1997; Krause et al. 1999). Various authors (e.g. Bickerstaff et al. 1984; Rozas & Odum 1988; Savino & Stein 1982, 1989a & b; Persson & Eklöv 1995) noted that predatory success (by piscivorous fish) in aquatic systems decreased as vegetation density increased. The consequences include both improved survival of the prey and reduced growth rates in the predator (Persson & Eklöv 1995). However, Savino & Stein (1989b) also noted that increasing structural complexity did not reduce capture success in any simple way. Different responses were noted by different predator and prey species, and structural complexity alone did not always provide refuge for prey; prey had to use the available structure to avoid predators. This is a key consideration in the design of possible fish refuge structures. Fish size also affects refuge use (Krause et al. 1998), as larger fish are less vulnerable to predation and are less affected by food deprivation. However, while this suggests that larger fish will use refuges less, it is not clear whether any intra-specific competition might operate, for example exclusion of smaller fish by larger fish. Shoaling behaviour is another method by which certain fish species limit the impact of predation. Shoaling fish use a wide variety of coordinated evasive tactics against fish predators (e.g. Pitcher 1986; Magurran 1990). Much less is known about the use of such evasive responses to attack by aerial predators, although Litvak (1993) showed that shoals of golden shiner, Notemigonus crysoleucas (Mitchill), did elicit a startle response and adjust their shoaling behaviour in response to the threat of aerial predation from a model kingfisher. The extent to which UK freshwater fish species might utilise either shoaling and/or use of structure as evasive strategies in response to attack by pursuit divers, such as cormorants, is not entirely clear. In UK freshwater ecosystems, predation risks from such fish-eating birds have been relatively insignificant compared to those from piscivorous fish such as pike, Esox lucius L., perch and brown trout. Only in recent decades have numbers of diving birds, such as cormorant, great crested grebe, Podiceps cristatus (L.), and sawbill ducks, Mergus spp., increased rapidly due to relaxation of persecution combined with expanding geographical ranges (Cramp & Simmons 1977; Russell et al. 1996). It may therefore be possible that prey fish have not evolved specific anti-predator behaviour strategies against attack by these birds, to the extent that such responses are an adaptive part of their life histories against piscivorous fish. Diel movements are a feature in the behaviour of many common prey species, such as roach and perch (Allen 1935; Bohl 1980; Goldspink 1990; Gliwicz & Jachner 1992; Garner 1996), with peaks of activity at dawn and dusk (Cowx 2001a). The fish commonly move from more inshore, vegetated areas during the day to offshore waters at

19 11/02/2003 15:05 Page 267

Fish refuges to reduce cormorant damage to fisheries

267

night, often to feed on zooplankton, although opposite movements also occur (Cowx 2001a). Such movements are thought to be influenced by the need to optimise feeding opportunities, mediated by the risk of predation. Movement by prey species is believed to carry a high risk of attracting predators (Ware 1973) and fish migrations can affect levels of predation by cormorants (Adams et al. 1994; Neuman et al. 1997). The precise timing and extent of these diel movements are likely to influence the level of interaction between cormorants and different fish species (Gliwicz & Jachner 1992). This suggests that reducing the extent of diel movements might be one possible mechanism by which fish refuges might be used to limit levels of predation. It has been shown that shoaling cyprinids are particularly vulnerable to stalking (fish) predators at low light intensities, i.e. around dawn and dusk (e.g. Cerri 1983; Pitcher & Turner 1986); referred to as the ‘twilight hypothesis’. During daylight, approaching predators are readily detected by sight, and prey then take appropriate avoidance reactions (such as shoaling or diving into cover). In twilight, however, stalking predators have small but significant visual and tactical advantages (such as striking from below when the target is silhouetted against the sky) which enable them to attack prey more successfully than would be possible in full daylight. Such effects are likely to be enhanced where these coincide with peaks in activity and diel movements. The extent to which pursuit predators may also benefit from foraging at low light intensities is less clear. However, it is likely that the tendency for cormorants to forage at dawn also reflects the improved availability and catchability of prey species at this time. The maximum swimming speed of a fish, and its endurance time at that speed, are both determined by its body length and by ambient water temperature (Beach 1984; Environment Agency, unpublished data). The maximum speed increases considerably with rising water temperature, but less so with larger size. For endurance at maximum speed, however, the converse relationships apply. Thus, for a given length of fish, increasing temperature results in a dramatic reduction in endurance; while, for a given temperature, a large increase in endurance results from an increase in length. The critical burst swimming speeds of various freshwater fish species, over the size range 8–20 cm (i.e. coinciding with sizes most commonly consumed by cormorants), ranges from 0.9–1.5 m s1 at temperatures below 10°C, to 1.2–1.9 m s1 at temperatures above 15°C (Environment Agency, unpublished data). Thus, the reported 1–4 m s1 underwater swimming speed of cormorants is likely to exceed that of, at least smaller, freshwater fish. This will apply particularly in the winter, although swimming endurance will be greater at this time. A cormorant that can get within 2–3 m of its prey undetected in open water would thus seem to have a reasonable chance of making a catch. However, the manoeuvrability of fish and cormorants may be other critical elements in the predator–prey equation, but neither capability can be quantified.

19.4 Features of artificial reefs and refuges The introduction of artificial reefs and fish attraction devices (FADs) into aquatic environments to enhance fisheries is a long-established practice (e.g. FAO 1989), and

19 11/02/2003 15:05 Page 268

268

Mitigation measures

continues to be an area of rapidly expanding development (Bolding et al. 2001). There are now a plethora of such systems for both benthic and pelagic fish species (e.g. reviews and conference proceedings edited by D’Itri 1985; Seaman Jr & Sprague 1991; Anon 1994; Jensen et al. 2000). However, much of this work has been carried out in the marine environment, with relatively little in freshwater habitats. The concept of providing artificial refuges, other than net enclosures, to protect fish from avian predators in natural and semi-natural freshwater habitats appears to be a relatively recent idea (Mott & Boyd 1995), but has not been investigated in any detail. Key features of existing artificial structures, the mechanisms by which these operate, and the results of investigations, particularly those in freshwater systems, are examined below to assess what possible factors might be identified in the context of designing appropriate refuges for freshwater fish.

19.4.1

Fish attraction devices (FADs)

The development of FADs was originally inspired by the widespread observation that many fish species are attracted to surface-floating objects (natural or artificial flotsam) in the ocean. Simple floating FADs have long been used by native peoples around the western Pacific islands and the south-central Mediterranean, for example, to congregate fish and allow more efficient capture by nets. These basic, but highly effective, methods have now spawned manufacture of a wide variety of increasingly complex surface and mid-water constructions using synthetic materials (see D’Itri 1985, for example). Studies have shown that marine fish respond within 5 days to the appearance of a new FAD, with attracted shoals occurring at distances of up to 200 m from a FAD in the open sea (Mottet 1985). To date, adaptations of this approach to fresh waters have been more limited in scale, although appear to be an area of growing interest (Löffler 1997; Cowx 2001b), particularly in North America (Bolding et al. 2001). In one trial, plywood structures were suspended in mid-water in an oligotrophic reservoir (Reeves et al. 1977). Diver observations showed that significant numbers of spotted bass, Micropterus punctulatum henshallis (Rafinesque), and bluegill sunfish, Lepomis macrochirus Rafinesque, were attracted to, and spawned on, the structure and their fry stayed close by, feeding on plankton. In a further trial at the same site (Smith et al. 1980), devices constructed from plastic, fibreglass and old tyres attracted fish to spawn and became good nursery grounds. Thus, after two months, four of the units were estimated to host 50 000 fry of four species. Helfman (1979) also reported significantly higher densities of freshwater fish under floating structures, and that fish densities were positively correlated with float surface area. He further noted that the number of fish attracted to such structures was reduced during overcast conditions. The size and vertical profile of a FAD were also shown to be important factors controlling fish abundance (Rountree 1989). Other experiments to enhance stillwater fisheries in the south-eastern United States, using mini-FADs, have been reported by Myatt (1985), and the use of artificial structure to enhance freshwater fisheries in still waters has recently been reviewed by Bolding et al. (2001). The reasons why fish gather beneath FADs and other flotsam is not entirely clear and various hypotheses have been put forward to explain the association (summary in

19 11/02/2003 15:05 Page 269

Fish refuges to reduce cormorant damage to fisheries

269

Rountree 1989). For example, such features may act as some form of visual reference in an otherwise unstructured environment or may concentrate food supply (Helfman 1979). However, the most widely held view (Rountree 1989) is that ‘fishes utilise floating materials in some manner which gives them protection from predators’ (i.e. other fish). This might be achieved through: direct shelter, camouflage, or by some mechanism of interference with a predator’s ability to capture prey (perhaps by making prey silhouettes difficult to distinguish, or by protecting a shoal’s blind zone from surprise approach). Among lake fishes, it has been suggested (Helfman 1981) that shadows from rafts provide protective advantage for fishes within the shadow. It was noted that a shaded observer can see a sunlit target at more than 2.5 times the distance that a sunlit observer could see a shaded target. It was suggested that this was due to a combination of a sunlit viewer having a raised contrast perception and difficulty in responding to a shaded target, and by the veiling-brightness effect, in which particles scatter relatively bright light into the observer’s eyes, further reducing the target’s visual contrast.

19.4.2

Benthic reefs and refuges

As with FADs, the use of benthic structures to enhance freshwater fisheries has previously received relatively little attention compared with marine fisheries, however, this appears to be changing (Bolding et al. 2001). Much of the emphasis hitherto has been on providing benthic reefs on ‘featureless’ lake beds to increase fish populations for sport-fishing. The dumping of rock and other debris to create reefs in some larger North American waters appears to have had relatively little success. However, more promising results have been noted from constructing simple shelters, mainly from natural (biodegradable) materials, that simulate naturally-occurring vegetational features (e.g. Seaman & Sprague 1991). Log cribs (Stone et al. 1991), which simulate natural piles of tree trunks commonly found in North American lakes, were shown to attract benthic fishes. Yellow perch, Perca flavescens (Mitchill), for example, preferentially use areas of lake bed littered with conifer logs at certain times of year, possibly for shelter (Moring et al. 1989). Log cribs filled with brushwood proved more effective at holding fish than those without, and several cribs placed close together were more effective than individual cribs (Bassett 1994). Brush bundles and vitrified clay pipes scattered on a lake bed also proved very attractive to species such as bluegill, Lepomis macrochirus, largemouth bass, Micropterus salmoides (Lacépède) and white catfish, Ictalurus catus (L.) (Wilbur 1978; Wege & Anderson 1979; Prince & Maughan 1979a). Other structures reported to have beneficial effects for fish include weighted conifer trees (Rold et al. 1996), cinder blocks (Moring & Nicholson 1994), tyre bundles (Prince & Maughan 1979b; Paxton & Stevenson 1979; Moring & Nicholson 1994), screened gravel and cobble (Bassett 1994), artificial vegetation (Mottet 1985), and wire cages (McKay et al. 1999, Chapter 20 in this volume). In a number of studies, the introduction of such structures resulted in significantly higher fishing success and substantial angler benefit (e.g. Wilbur 1978; Paxton & Stevenson 1979; Wege & Anderson 1979; Prince & Maughan 1979b; Bolding et al. 2001). The use of artificial structures in freshwater fisheries to enhance the spawning

19 11/02/2003 15:05 Page 270

270

Mitigation measures

success of fish has also been a growing area of interest (Nash et al. 1999; Cowx 2001b; Winfield et al. 2001; Zalewski & Frankiewicz 2001).

19.4.3

Overview

The results of studies with FADs and benthic reefs indicate that such artificial structures can have beneficial effects on freshwater fish stocks and fisheries, and can be used to attract and aggregate fish species. There are no specific examples for freshwater fish species native to the UK, but the evidence for related North American species suggests that similar findings may apply. This may enable structures to be used to increase levels of fish biomass per unit area, but these structures in themselves are unlikely to afford fish much, if any, protection from diving birds such as cormorants. Further, simply aggregating fish may serve as an attraction to predators. Designs for refuge structures will therefore need to incorporate additional features which will provide fish with some level of protection. To be effective, such structures will either need to admit fish and exclude cormorants, or, by their presence, render a site less attractive to cormorants, possibly as a result of increasing the foraging ‘costs’, so that birds choose to forage elsewhere.

19.5 Synthesis and conclusions Cormorants are versatile predators capable of foraging in a variety of aquatic habitats and of consuming a variety of fish species across a range of sizes. The birds are mainly visual feeders and foraging behaviour can be affected by factors such as water turbidity and cloud cover. However, that birds are still able to feed in quite turbid water suggests that other prey detection mechanisms may also operate. Foraging bouts occur mostly at dawn. This coincides with peaks in activity and diel movements in certain key prey species and may indicate a causative relationship; cyprinid species are also reported to be particularly vulnerable to attack by other predatory fish (stalking predators) at this time. Published estimates of the swimming speed of both cormorants and fish suggest that cormorants can swim faster and have greater endurance than most coarse fish and smaller trout, particularly in the winter months. However, the relative manoeuvrability of the two is unknown. Aquatic plants and other underwater structures are important features for many native fish species as a refuge against attack by predatory fish. Evidence indicates better survival of prey species and reduced growth of fish predators where there is greater vegetation structure acting as a refuge. The impact of vegetation structure on cormorant foraging behaviour is not entirely clear. There is some suggestion that cormorants avoid such areas, possibly due to higher foraging costs, although where prey densities are higher, a trade-off between increased foraging costs and increased prey availability might apply. However, cormorant numbers tend to be highest on inland waters in England and Wales in the winter months when vegetation dies back and is at its lowest level. There thus appears to be a reasonable basis for attempting to reduce levels of predation by the creation of summer-type habitat in the winter months, through the provision of artificial alternatives.

19 11/02/2003 15:05 Page 271

Fish refuges to reduce cormorant damage to fisheries

271

It is evident that the inclusion of artificial structures in freshwater systems can be used to attract fish and to enhance fish stocks and fishery performance. This will be a key requirement in respect of possible refuge design. Since cormorants have an apparent advantage over fish in terms of swimming speed, refuges are unlikely to operate as fish ‘bolt holes’. Instead, such structures will need to be able to attract fish and serve as secure holding areas where fish are able to shelter and find protection from cormorants. One possible mechanism by which fish refuges might operate would be for these structures to reduce the extent of diel movements by fish, through providing cover in more open areas and closer to preferred foraging localities. The appropriate siting of refuges relative to other habitat features, and each other, is likely to be a key requirement in this regard. The features of artificial habitats that appear to be most attractive to UK freshwater fish are the presence of ‘structure’, to mimic natural habitat features (i.e. refugia), and/or overhead cover to provide shading and an enhanced ability to detect oncoming predators. The types of ‘structure’ most suitable as attractants might include artificial weed (either floating, mid-water, or benthic), brushwood bundles or branches, and submerged pipes. Responses to different habitat features are likely to vary between species, and structures may need to be tailored for different species of fish. However, to be effective as refuges from cormorants, structures to attract fish will also need to alter the pattern of predator–prey interaction in such a way that the fish are conferred some advantage and are provided with additional protection from diving birds. This artificially induced ‘mismatch’ is likely to necessitate either excluding the predators (e.g. by surrounding fish attractants with appropriate sized netting to make them ‘cormorant proof’), or obstructing the birds in such a way that their foraging efficiency is reduced or their movements are restricted relative to the fish. Recent results suggest that cormorants require feeding sites with high densities of prey, and will abandon a site when prey density falls below a threshold level. Reducing the ‘available fish density’ of a site by deploying fish refuges might therefore be expected to increase the costs of foraging to such an extent that birds choose to forage elsewhere. Further work is required to assess whether such refuge structures might be effective and to evaluate design features to suit the particular requirements of different fish species.

Acknowledgement This work was funded by the Ministry of Agriculture, Fisheries and Food (now the Department for Environment, Food and Rural Affairs), UK.

References Adams C.E., Brown D.W. & Keay L. (1994) Elevated predation risk associated with inshore migrations of fish in a large lake, Loch Lomond, Scotland. Hydrobiologia 290, 135–138. Allen K.R. (1935) The food and migration of the perch in Windermere. Journal of Animal Ecology 4, 264–273.

19 11/02/2003 15:05 Page 272

272

Mitigation measures

Anon. (1994) Fifth International Conference on Aquatic Habitat Enhancement, Long Beach, California, November 1991. Bulletin of Marine Science 55 (2-3), 265–1360. Bassett C.E. (1994) Use and evaluation of fish habitat structures in lakes of the eastern United States by the USDA forest service. Bulletin of Marine Science 55 (2-3), 1137–1148. Beach M.H. (1984) Fish Pass Design: Criteria for the Design and Approval of Fish Passes and other Structures to Facilitate the Passage of Migratory Fish in Rivers. MAFF Directorate of Fisheries Research, Lowestoft. Technical Report No. 78, 46 pp. Bickerstaff W.B., Ziebell C.D. & Matter W.J. (1984) Vulnerability of redbelly tilapia fry to bluegill predation with changes in cover availability. North American Journal of Fisheries Management 4, 120–125. Bohl E. (1980) Diel pattern of pelagic distribution and feeding in planktivorous fish. Oecologia 44, 368–375. Bolding B., Bonar S. & Divens M. (2001) Use of Artificial Structure to Enhance Freshwater Sportfish Communities in Lakes, Ponds and Reservoirs: A Literature Review. Washington Department of Fish and Wildlife, Technical Report Number FPT01-06, 35 pp. Builles A., Julian J.M., Yesou P. & Girard O. (1986) Rythme d’activite et occupation de l’espace par le Grand Cormoran (Phalacrocorax carbo) sur un site d’hivernage: l’exemple de la Region d’Olone (Vendee). Gibier Faune Sauvage 3, 43–65. (in French). Callaghan D.A., Kirby J.S. & Bell M.C. (1994) An Assessment of Cormorant Phalacrocorax carbo Occupancy and Impact at Stillwater Game Fisheries in England and Wales. Report to the Association of Stillwater Game Fishery Managers, WWT, Slimbridge, 29 pp. Callaghan D.A., Kirby J.S., Bell M.C. & Spray C.J. (1998) Cormorant Phalacrocorax carbo occupancy and impact at stillwater game fisheries in England and Wales. Bird Study 45, 1–17. Cerri R.D. (1983) The effect of light intensity on predator and prey behaviour in cyprinid fish: factors that influence prey risk. Animal Behaviour 31, 736–742. Copp G. (1992) Comparative microhabitat use of cyprinid larvae and juveniles in a lotic floodplain channel. Environmental Biology of Fishes 33, 181–193. Cowx I.G. (2001a) Coarse Fish in Lowland Rivers. R & D Note 459, Bristol: Environment Agency, 129 pp. Cowx I.G. (2001b) Principles and approaches to the management of lake and reservoir fisheries. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 376–393. Cramp S. & Simmons K.E.L. (eds) (1977) Handbook of the Birds of Europe, the Middle East and North Africa. Vol. 1. The Birds of the Western Palearctic. Oxford: Oxford University Press. Davies J.M. & Feltham M.J. (1994) The Impact of Cormorants (Phalacrocorax carbo carbo) on Angling Catches on the River Ribble, Lancashire. Interim Report to the National Rivers Authority. Davies J.M. & Feltham M.J. (1995) Do Cormorants and Anglers Compete for the Same Resources? Proceedings of the 25th Annual Institute of Fisheries Management Study Course 1994, Lancaster, pp. 167–188. Davies J. M. & Feltham M.J. (1996) The diet of wintering cormorants Phalacrocorax carbo L. in relation to angling catches on a coarse river fishery in north-west England. In S.P.R. Greenstreet & M.L. Tasker (eds) Aquatic Predators and their Prey. Oxford: Fishing News Books, Blackwell Science, pp. 106–110. De Nie H.W. (1995) Changes in the inland fish populations in Europe in relation to the increase of the cormorant Phalacrocorax carbo sinensis. Ardea 83, 115–122. Diehl S. (1988) Foraging efficiency of three freshwater fish: effects of structural complexity and light. Oikos 53, 207–214. Dill L.M. (1987) Animal decision making and its ecological consequences: the future of aquatic ecology and behaviour. Canadian Journal of Zoology 65, 803–811. Dill L.M. & Gillett J.F. (1991) The economic logic of the barnacle Balanus glandula (Darwin) hiding behaviour. Journal of Experimental Marine Biology and Ecology 153, 115–127.

19 11/02/2003 15:05 Page 273

Fish refuges to reduce cormorant damage to fisheries

273

Dirksen S., Boudewijn T.J., Noordhuis R. & Marteijn E.C.L. (1995) Cormorants Phalacrocorax carbo sinensis in shallow eutrophic freshwater lakes: prey choice and fish consumption in the non-breeding period and effects of large-scale fish removal. Ardea 83, 167–184. D’Itri F.M. (ed.) (1985) Artificial Reefs: Marine and Freshwater Applications. Michigan: Lewis Publishers, 589 pp. Eklöv P. & Persson L. (1996) The response of prey to the risk of predation: proximate cues for refuging juvenile fish. Animal Behaviour 51, 105–115. Food and Agriculture Organisation of the United Nations (1989) How to Make and Set FADs (Fish Aggregating Devices). Compiled by M. Ben-Yami, FAO Training Series, Rome, Number 15, 68 pp. Feare C.J. (1988) Cormorants as predators at freshwater fisheries. Institute of Fisheries Management Annual Study Course 18, 18–42. Feltham J.M., Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999) Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0106. London: MAAF, 406 pp. Garner P. (1996) Diel behaviour of juvenile 0-group fishes in a regulated river: the Great Ouse, England. Ecology of Freshwater Fish 5, 175–182. Giles N. (1994) Freshwater Fish of the British Isles. Shrewsbury: Swan Hill Press, 192 pp. Gliwicz Z.M. & Jachner A.J. (1992) Diel migrations of juvenile fish: a ghost of predation past or present? Archiv für Hydrobiologie 124, 385–410. Goldspink C.R. (1990) The distribution and abundance of young (I–II) perch, Perca fluviatilis L., in a deep eutrophic lake, England. Journal of Fish Biology 36, 439–447. Gotceitas V. & Brown J.A. (1993) Substrate selection by juvenile Atlantic cod (Gadus morhua): effects of predation risk. Oecologia 93, 31–37. Grémillet D., Argentin G., Schulte B. & Culik B.M. (1998) Flexible foraging techniques in breeding cormorants Phalacrocorax carbo and shags Phalacrocorax aristotelis: Benthic or pelagic feeding? Ibis 140, 113–119. Grémillet D. & Wilson R.P. (1999) A life in the fast lane: energetics and foraging strategies of the great cormorant. Behavioural Ecology 10, 516–524. Grémillet D., Wilson R.P., Storch S. & Gary Y. (1999) Three-dimentional space utilization by a marine predator. Marine Ecology Progress Series 183, 263–273. Hebshi A. (1998) Foraging site preferences of Brandt’s cormorants off the Santa Cruz, California, coast. Colonial Waterbirds 21, 245–250. Helfman G.S. (1979) Fish attraction to floating objects in lakes. In D.L. Johnson & R.A. Stein (eds) Response of Fish to Habitat Structure in Standing Water. North Central Division American Fisheries Society Special Publication No. 6, 49–57. Helfman G.S. (1981) The advantage to fishes of hovering in shade. Copeia 1981, 392–400. Hughes B., Bevan R.M., Bowler J.M., Still L., Carss D.N., Marquiss M., Hearn R.D. & Bruce J. (1999) Feeding Behaviour of Fish-Eating Birds in Great Britain. Report to the Department of the Environment, Transport and Regions, 249 pp. Hughes B., Bruce J., Ekins G. & Newson S. (2000) Movements and Distribution of Inland Breeding Cormorants in England. English Nature Research Report No. 360, 95 pp. Im B.H. & Hafner H. (1984) Impact des oiseaux piscivores et plus particulierement du grand cormorant Phalacrocorax carbo sinensis sur les exploitations piscicoles en Camargue, France. Le Sambuc: Station Biologique de la Tour du Valat. Jacobsen L. & Perrow M.R. (1998) Predation risk from piscivorous fish influencing the diel use of macrophytes by planktivorous fish in experimental ponds. Ecology of Freshwater Fish 7, 78–86. Jensen A.C., Collins K.J. & Lockwood A.P.M. (eds) (2000) Artificial Reefs in European Seas. Dordrecht: Kluwer, 508 pp. Junor F.J.R. (1969) Aspects of the biology of piscivorous birds of Lake Kyle, Rhodesia. MSc thesis, Rhodes University, Grahamstown, South Africa.

19 11/02/2003 15:05 Page 274

274

Mitigation measures

Kato A., Watanuki Y. & Naito Y. (1998) Benthic and pelagic foraging of two Japanese cormorants, determined by simultaneous recording of location and diving activity. Journal of the Yamashina Institute for Ornithology 30, 101–108. Kennedy G.J.A. & Greer J.E. (1988) Predation by cormorants Phalacrocorax carbo L on the salmonid populations of an Irish river. Aquaculture and Fisheries Management 19, 159–170. King D.T., Harrel J., Brent J., Dorr B. & Reinhold D. (1998) Observations of nocturnal foraging in the double-crested cormorant. Colonial Waterbirds 21, 234–235. Kirby J.S., Gilburn A.S. & Sellers R.M. (1995) Status, distribution and habitat use by cormorants Phalacrocorax carbo wintering in Great Britain. Ardea 83, 93–102. Krause J., Loader S.P., McDermott J. & Ruxton G.D. (1998) Refuge use by fish as a function of body length-related metabolic expenditure and predation risks. Proceedings Royal Society of London Series B 265, 2373–2379. Krause J., Loader S.P., Kirkman E. & Ruxton G.D. (1999) Refuge use by fish as a function of body weight changes. Acta ethol 2, 29–34. Lindell L. (1997) Recent population development of the cormorant in Sweden. Ekologia Polska 45, 79–82. Litvak M.K. (1993) Response of shoaling fish to the threat of aerial predation. Environmental Biology of Fishes 36, 183–192. Löffler H. (1997) Artificial habitats for fishes in Lake Constance (Bodensee): observation of fish aggregating devices with a remotely operated vehicle. Fisheries Management and Ecology 4, 419–420. Lovvorn J.R., Yule D. & Derby C.E. (1999) Greater predation by double-crested cormorants on cutthroat versus rainbow trout fingerlings stocked in a Wyoming river. Canadian Journal of Zoology 77, 1984–1990. Lucas M.C., Thom T.J., Duncan A. & Slavik O. (1998) Coarse Fish Migration, Occurrence, Causes and Implications. R & D Technical Report W152, 161 pp. Bristol: Environment Agency. Magurran A.E. (1990) The adaptive significance of schooling as anti-predator defence in fish. Annales Zoologici Fennici 27, 51–66. Maitland P.S. & Campbell R.N. (1992) Freshwater Fishes of the British Isles. London: Harper Collins, 368 pp. Mann R.H.K. & Winfield I.J. (1992) Restoration of Riverine Fisheries Habitats. R & D Note 105, 79 pp. Bristol: National Rivers Authority. Marquiss M. & Carss D.N. (1994) Avian Piscivores: Basis for policy. R & D Report 461/8/NandY. Bristol: National Rivers Authority, 104 pp. Marquiss M., Carss D.N., Armstrong J.D. & Gardiner R. (1998) Fish-eating birds and salmonids in Scotland. Report to The Scottish Office Agriculture Environment and Fisheries Department, 156 pp. Marteijn E.C.L. & Dirksen S. (1991) Cormorants Phalacrocorax carbo sinensis feeding in shallow eutrophic freshwater lakes in The Netherlands in the non-breeding period: prey choice and fish consumption. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 135–155. Martucci O. & Consiglio C. (1991) Activity rhythm and food choice of cormorants (Phalacrocorax carbo sinensis) wintering near Rome, Italy. Le Gerfaut 81, 151–160. Matkowski S.M.D. (1989) Differential susceptibility of three species of stocked trout to bird predation. North American Journal of Fisheries Management 9, 184–187. McKay H.V., Russell I.C., Parrott D. & Watola G. (1998) Study of Fish-Eating Birds on the Rivers Ouse and Ure, Yorkshire. Report to the Environment Agency, 51 pp. McKay H.V., Furness R.W., Russell I.C., Parrott D., Rehfisch M.M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0107. London: MAAF, 254 pp.

19 11/02/2003 15:05 Page 275

Fish refuges to reduce cormorant damage to fisheries

275

Mikheev V.N., Metcalfe N.B., Huntingford F.A. & Thorpe J.E. (1994) Size-related differences in behaviour and spatial distribution of juvenile Atlantic salmon in a novel environment. Journal of Fish Biology 45, 379–386. Moring J.R., Negus M.T., McCullough R.D. & Herke S.W. (1989) Large concentrations of submerged pulpwood logs as fish attraction structures in a reservoir. Bulletin of Marine Science 44, 609–615. Moring J.R. & Nicolson P.H. (1994) Evaluation of three types of artificial habitats for fishes in a freshwater pond in Maine, USA. Bulletin of Marine Science 55 (2–3), 1149–1159. Mott D.F. & Boyd F.L. (1995) A review of techniques for preventing cormorant depredations at aquaculture facilities in the south-eastern United States. In D.N. Nettleship and D.C. Duffy (eds) The double-crested cormorant: Biology, conservation and management. Colonial Waterbirds 18 (Special Publication 1), pp. 176–180. Mottet M.G. (1985) Enhancement of the marine environment for fisheries and aquaculture in Japan. In F.M. D’Itri (ed.) Artificial Reefs: Marine and Freshwater Applications. Michigan: Lewis Publishers Inc., pp. 13–112. Myatt D.O. (1985) Midwater fish attractors. In D’Itri, F.M. (ed.) Artificial Reefs: Marine and Freshwater Applications. Michigan: Lewis Publishers, pp. 303–315. Nash K.T., Hendry K. & Cragg-Hine D. (1999) The use of brushwood bundles as fish spawning media. Fisheries Management and Ecology 6, 349–355. Neuman D.L., Pearl P.J., Ewins R., Black D.V., Pike M. & Karwowski K. (1997) Spatial and temporal variation in the diet of double-crested cormorants (Phalacrocorax auritis) breeding on the lower Great Lakes in the early 1990s. Canadian Journal of Fish and Aquatic Sciences 54, 1569–1584. North R. (2002) Factors affecting the performance of stillwater coarse fisheries in England and Wales. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 284–298. Paxton K.O. & Stevenson F. (1979) Influence of artificial structures on angler harvest from Killdeer Reservoir, Ohio. In D.L. Johnson & R.A. Stein (eds) Response of Fish to Habitat Structure in Standing Water. North Central Division, American Fisheries Society Special Publication Number 6. Bethesda: American Fisheries Society, pp. 70–76. Pearson T.H. (1968) The feeding biology of sea-bird species breeding on the Farne Islands, Northumberland. Journal of Animal Ecology 37, 521–552. Persson L. (1993) Predator-mediated competition in prey refuges: the importance of habitat dependent prey resources. Oikos 68, 12–22. Persson L. & Eklöv P. (1995) Prey refuges affecting interactions between piscivorous perch and juvenile perch and roach. Ecology 76, 70–81. Pilon C., Burton R. & McNeil R. (1983) Summer food of the great and double-crested cormorants on the Magdalen Islands, Quebec. Canadian Journal of Zoology 61, 2733–2739. Pitcher T.J. (1986) Functions of shoaling behaviour in teleosts. In T.J. Pitcher (ed.) The Behaviour of Teleost Fishes. Baltimore: Johns Hopkins University Press, pp. 294–337. Pitcher T.J. & Turner J.R. (1986) Danger at dawn: experimental support for the twilight hypothesis in shoaling minnows. Journal of Fish Biology 29 (Suppl. A), 59–70. Platteeuw M., Beekman J.H., Boudewijn T.J. & Marteijn C.L. (1992) Cormorants Phalacrocorax carbo in Lake Ketelmeer outside the breeding season: numbers, food choice and food availability. Limosa 65, 93–102. Prince E.D. & Maughan O.E. (1979a) Telemetric observations of largemouth bass near underwater structures in Smith Mountain Lake, Virginia. In D.L. Johnson & R.A. Stein (eds) Response of Fish to Habitat Structure in Standing Water. North Central Division, American Fisheries Society Special Publication Number 6. Bethesda: American Fisheries Society, pp. 26–32. Prince E.D. & Maughan O.E. (1979b) Attraction of fishes to artificial tire reefs in Smith Mountain lake, Virginia. In D.L. Johnson & R.A. Stein (eds) Response of Fish to Habitat Structure in Standing Water. Bethesda, USA: North Central Division, American Fisheries Society Special Publication Number 6. Bethesda: American Fisheries Society, pp. 19–25.

19 11/02/2003 15:05 Page 276

276

Mitigation measures

Reeves W.C., Hooper G.R. & Smith B.W. (1977) Preliminary observations of fish attraction to artificial midwater structures in freshwater. Proceedings Annual Conference S.E. Association of Fish and Wildlife Agencies 31, 471–476. Rold R.E., McComish T.S. & van Meter D.E. (1996) A comparison of cedar trees and fabricated polypropylene modules as fish attractors in a strip mine impoundment. North American Journal of Fisheries Management 16, 223–227. Rountree R.A. (1989) Association of fishes with fish aggregation devices: effects of structure and size on fish abundance. Bulletin Marine Science 44, 960–972. Rozas L.P. & Odum W.E. (1988) Occupation of submerged aquatic vegetation by fishes: testing the roles of food and refuge. Oceologia 77, 101–106. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.A. (1996) Assessment of the problem of fish-eating birds in inland fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0104. London: MAAF, 130 pp. Russell I.C., Cook A., Kinsman D.A., Ives M.J. & Lower N.J. (in press). Stomach contents analysis of cormorants at some different fishery types in England and Wales. Proceedings 5th European Cormorant Conference, Freising, Germany, December 2000. Savino J.F. & Stein R.A. (1982) Predator–prey interaction between largemouth bass and bluegills as influenced by simulated, submerged vegetation. Transactions of the American Fisheries Society 111, 255–266. Savino J.F. & Stein R.A. (1989a) Behaviour of fish predators and their prey: habitat choice between open water and dense vegetation. Environmental Biology of Fishes 24, 287–293. Savino J.F. & Stein R.A. (1989b) Behavioural interactions between fish predators and their prey: effects of plant density. Animal Behaviour 37, 311–321. Schmid D., Grémillet D. & Culik B.M. (1995) Energetics of underwater swimming in the Great Cormorant (Phalacrocorax carbo sinensis). Marine Biology 123, 875–881. Seaman W. Jr & Sprague L.M. (eds) (1991) Artificial Habitats for Marine and Freshwater Applications. San Diego: Academic Press Inc., 285 pp. Sellers R.M., Ekins G.R., Hughes B. & Kirby J.S. (1997) Population developments of inland breeding Cormorants in Great Britain. Supplement. Ric. Biol. Sevaggina XXVI, 11–21. Sih A. (1992) Prey uncertainty and the balancing of anti-predator and feeding needs. American Naturalist 139, 1052–1069. Sih A. (1997) To hide or not to hide? Refuge use in a fluctuating environment. Tree 12, 375–376. Smith B.W., Hooper G.R. & Lawson C.S. (1980) Observation of fish attraction to improved artificial midwater structures in freshwater. Proceedings Annual Conference S.E. Association of Fish and Wildlife Agencies 34, 404–409. Smith I. & Lyle A. (1979) Distribution of Freshwaters in Great Britain. Report by the Institute of Terrestrial Ecology, Cambridge, 44 pp. Stone R.B., McGurrin J.M., Sprague L.M. & Seaman W. Jr (1991) Artificial habitats of the world: synopsis and major trends. In W. Seaman Jr & L.M. Sprague (eds) Artificial Habitats for Marine and Freshwater Applications. San Diego: Academic Press, Inc., pp. 31–60. Stonehouse B. (1967) Feeding behaviour and diving rhythms of some New Zealand shags Phalacrocoracidae. Ibis 109, 600–605. Strod T., Izhaki I., Arad Z. & Katzir G. (1999) Prey Detection and Prey Preference in Diving Cormorants. Waterbird Society 23rd Annual Meeting and Workshops, November 1999, Grado, Italy (abstract, p. 65). Suter W. (1991) Food and feeding of cormorants Phalacrocorax carbo wintering in Switzerland. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 156–165. Suter W. (1994) Overwintering waterfowl on Swiss lakes: how are abundance and species richness influenced by trophic status and lake morphology. Hydrobiologia 279/280, 1–14.

19 11/02/2003 15:05 Page 277

Fish refuges to reduce cormorant damage to fisheries

277

Suter W. (1995) Are cormorants Phalocrocorax carbo wintering in Switzerland approaching carrying capacity? An analysis of increase patterns and habitat choice. Ardea 83, 255–266. Suter W. (1997) Roach rules: shoaling fish are a constant factor in the diet of cormorants Phalocrocorax carbo in Switzerland. Ardea 85, 9–27. Tabor R.A. & Wurtsbaugh W.A. (1991) Predation risk and the importance of cover for juvenile rainbow trout in lentic systems. Transactions of the American Fisheries Society 120, 728–738. Tonn W.M., Paszkowski C.A. & Holopainen I.J. (1992) Piscivory and recruitment: mechanisms structuring prey populations in small lakes. Ecology 73, 951–958. Traylor K.M., Brother D.J., Wooler R.D. & Potter I.C. (1989) Opportunistic foraging by three species of cormorants in an Australian Estuary. Journal of Zoology London 218, 87–98. Van Dobben W.H. (1952) The food of the cormorant in the Netherlands. Ardea 40, 1–63. Van Eerden M.R. & Voslamber B. (1995) Mass fishing by cormorants Phalacrocorax carbo sinensis at Lake Ijsselmeer, The Netherlands: a recent and successful adaptation to a turbid environment. Ardea 83, 199–212. Veldkamp R. (1994) Voedselkus van Aalscholvers Phalacrocorax carbo sinensis in NordwestOverijssel. Opdrachtgever, Rijkswaterstaat RIZA. Voslamber B. (1988) Visplaatskeuze, fourageerwijze en voedselkeuze van aalscholvers Phalacrocorax carbo sinensis in het Ijsselmeerebied in 1982. Flevobericht, 286. Lelystad: Rijkdientst voor de Ijsselmeerpolders. Voslamber B. & van Eerden M.R. (1991) The habit of mass flocking fishing by cormorants Phalacrocorax carbo sinensis at the IJsselmeer, The Netherlands. In M.R. van Eerden & M. Zijlstra (eds) Proceedings of the 1989 Workshop on Cormorants Phalacrocorax carbo. Lelystad: Rijkswaterstaat Directorate Flevoland, pp. 182–191. Wanless S., Burger A.E. & Harris M.P. (1991) Diving depths of shags Phalacrocorax aristotelis breeding on the Isle of May. Ibis 133, 37–42. Ware D.M. (1973) Risk of epibenthic prey to predation by rainbow trout (Salmo gairdneri). Journal of the Fisheries Research Board of Canada 30, 787–797. Wege G.J. & Anderson R.O. (1979) Influence of artificial structures on largemouth bass and bluegills in small ponds. In D.L. Johnson & R.A. Stein (eds) Response of Fish to Habitat Structure in Standing Water. North Central Division, American Fisheries Society Special Publication Number 6. Bethesda: American Fisheries Society, pp. 59–69. Wernham C.V., Armitage M., Holloway S.J., Hughes B., Hughes R., Kershaw M., Madden J.R., Marchant J.H., Peach W.J. & Rehfisch M.M. (1999) Population, Distribution, Movements and Survival of Fish-Eating Birds in Great Britain. Report to the Department of the Environment, Transport and Regions. London: DETR, 360 pp. Wilbur R. (1978) Two types of fish attractors compared in Lake Tohopekaliga, Florida. Transactions of the American Fisheries Society 107, 689–695. Wilson R.P. & Wilson M-P.T. (1988) Foraging behaviour of four sympatric cormorants. Journal of Animal Ecology 57, 943–953. Winfield I.J. & Nelson J.S. (eds) (1991) Cyprinid Fishes: Systematics, Biology and Exploitation. London: Fish and Fisheries Series 3, Chapman & Hall, 667 pp. Winfield I.J., Fletcher J.M. & Winfield D.K. (2001) Conservation of the endangered whitefish, Coregonus lavaretus, population of Haweswater, UK. In: I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 232–241. Zalewski M. & Frankiewicz P. (2001) The potential to control fish community structure using preference for different spawning substrates in a temperate reservoir. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 217–222.

20 11/02/2003 15:06 Page 278

Chapter 20

Pilot trials to assess the efficacy of fish refuges in reducing the impact of cormorants on inland fisheries H.V. McKAY* Central Science Laboratory, Sand Hutton, York, UK

I.C. RUSSELL Centre for Environment, Fisheries and Aquaculture Science, Lowestoft, UK

M.M. REHFISCH and M. ARMITAGE The British Trust for Ornithology, The Nunnery, Thetford, Norfolk, UK

J. PACKER ADAS, Burghill Road, Westbury-on-Trym, Bristol, UK

D. PARROTT Central Science Laboratory, Sand Hutton, York, UK

Abstract Fish refuges are artificial structures designed to protect fish from predators; their efficacy, however, is untested for UK freshwater species. Wire mesh refuges were evaluated in two trials undertaken at Rye Meads, England, during autumn/winter 1997/1998. Two adjacent drainable ponds (0.3 ha) were stocked with three sizes of carp, Cyprinus carpio L. One pond was netted over to exclude piscivorous birds, the other left open. Refuges were placed in both ponds prior to the second trial, and cormorant numbers, foraging behaviour, and losses and wounding rates of fish compared between trials. A 60% difference in mortality between the netted and open ponds was assumed to be due to cormorant predation. Cormorant dive duration was longer, and the percentage of large fish recovered with wounds was lower, when refuges were present. These results suggest that refuges may reduce the availability of carp to cormorants, particularly the larger size classes more sought after by anglers (28 cm), but further research is recommended. Keywords: carp, cormorant, fish refuge, foraging, predation, wounding.

20.1 Introduction The number of cormorants, Phalacrocorax carbo L., overwintering and breeding inland in the UK has increased over the past 20 years, leading to concerns about the impact on recreational fisheries (Russell et al. 1996; Hughes et al. 1999; Rehfisch et al. 1999). In 1995, the UK Government commissioned a programme of research on the problem *Correspondence: Helen McKay, Central Science Laboratory, Sand Hutton, York YO41 1LZ, UK (email: [email protected]).

20 11/02/2003 15:06 Page 279

Fish refuges and cormorant predation

279

of fish-eating birds at inland fisheries in England and Wales. As part of this programme, McKay et al. (1999) reviewed management measures to control damage by fish-eating birds. They concluded that fish refuges were potentially a low-cost, environmentally-friendly measure which warranted further investigation. The efficacy of refuges has not been experimentally tested for freshwater fish species in the UK. This chapter describes a preliminary investigation of the technique. Fish refuges are artificial structures introduced into the water body to provide cover for fish from predators. Studies indicate that cormorants are opportunist feeders that exploit locally abundant fish stocks (e.g. Suter 1995), and that reductions in fish availability result in a shift in foraging location. Grémillet and Wilson (1999) modelled the effect of prey availability on the foraging behaviour of cormorants, and concluded that even a limited reduction in prey density makes birds unable to balance their energy needs. Habitat complexity and the availability of natural refugia are known to affect the behaviour of many fish, such as juvenile Atlantic salmon, Salmo salar L. (Mikheev et al. 1994), smelt, Osmerus eperlanus L., perch, Perca fluviatilis (L.) and roach, Rutilus rutilus (L.) (Gliwicz & Jachner 1992). Matkowski (1989) suggested that differential susceptibility of three species of stocked trout to bird predation was partly due to habitat preference; the less susceptible species were thought to have well-developed escape reactions and frequented cover. If fish refuges are likely to reduce the impact of cormorants on fisheries they must: (a) be utilised by the species and sizes of fish of value to anglers; (b) reduce the availability (and hence consumption) of these fish to cormorants, and/or reduce wounding rates; and (c) not hinder the activity of anglers. The pilot trials described in this chapter were mainly designed to investigate the second of these factors. A full account is given in McKay et al. (1999). Two identical trials were carried out in succession; the first without refuges, and the second with refuges; a comparison of the trials indicating the effect of refuge presence.

20.2 Materials and methods 20.2.1

Study site

The study was carried out using two adjacent ponds that could be drained (0.3 ha each in area and approximately 1.5 m deep) at Rye Meads in Hertfordshire (National Grid Reference: TL 3810). The ponds formed part of a 17 ha complex of shallow water bodies. One pond was covered to ground level, by 2.5-cm-mesh nylon netting set approximately 1.5 m above water level. The other pond was left uncovered. The first trial began on 21 October 1997 and lasted 42 days, the second, on 11 December and lasted 47 days.

20.2.2

Refuge design

Five refuges were placed in each pond on 4 December 1997, prior to the second trial. Each refuge comprised three cylinders of galvanised steel wire mesh (Rylock

20 11/02/2003 15:06 Page 280

280

Mitigation measures

Figure 20.1 Refuge design: three open-ended cylinders of deer fence tied together in a triangular arrangement. Mesh size varied between 15 cm  6 cm and 15 cm  18 cm

high tensile deer fence) tied together by cable ties in a triangular arrangement (Fig. 20.1). The refuges measured approximately 1.8 m  1.8 m, and were anchored to the bottom of the pond by 1.5 m iron stakes.

20.2.3

Fish stocking

All fish were measured (fork length to the nearest mm) prior to stocking in the two ponds. Three size classes of carp, Cyprinus carpio L., were used: large (mean length 27–29 cm), medium (12–14 cm) and small (7–8 cm). Any fish showing previous damage were stocked to the covered pond. Numbers and mean length of fish are given in Table 20.1. The mean weight and stocking density of fish in the ponds was estimated from length/weight regressions for common and mirror carp – regression equations supplied courtesy of the Environment Agency (B. Byatt, personal communication). The estimated stocking densities (135–161 kg ha1) were within the range typical of stillwater recreational fisheries in the Thames catchment (Pilcher & Feltham 1997).

20.2.4

Fish recovery

At the end of each trial the ponds were drained and all remaining fish recovered and examined. Most fish were recovered with the aid of backpack electric fishing gear. Recovery procedures and effort were consistent between ponds and trials. Details of any dead fish or any fish remains were also recorded. Recovered fish were measured and assessed for recent bird damage (Carss 1993). Fish were further classified according to the level of damage. Any dead fish recovered with evidence of bird damage or any fish remains were recorded separately.

20 11/02/2003 15:06 Page 281

Fish refuges and cormorant predation Table 20.1

Details of carp stocked in trial 1 (without refuges) and trial 2 (with refuges)

Pond

Fish size

Number stocked

Mean length (cm)

Estimated mean weight (g)

Trial 1 (without refuges) Covered Large Medium Small Total Open Large Medium Small Total

51 104 256 411 54 103 257 414

27.3 11.7 7.4

638 47 13

27.6 11.8 7.4

663 45 13

Trial 2 (with refuges) Covered Large Medium Small Total Open Large Medium Small Total

50 100 287 437 50 100 287 437

28.3 13.7 7.4

711.1 75.9 13.0

28.5 13.9 7.5

727.6 79.9 13.4

20.2.5

281

Estimated stocking density (kg ha1)

135.6

145.8

156.3

160.7

Bird observations

Monitoring of cormorants was undertaken between 2 and 7 days per week, generally from dawn to dusk. Scan sampling was conducted at 15-min intervals, which involved noting the number and behaviour of all cormorants present (e.g. flying overhead or feeding). Nocturnal observations were conducted using night-viewing equipment, to detect herons and other predators (described in McKay et al. 1999). Periodically, remote video cameras were used to record activity at the ponds. Video surveillance of the covered pond confirmed no piscivorous bird presence. Focal observations on the foraging behaviour of individual cormorants were collected concurrently with the scan sampling. The following data were recorded: (i) total time on the water (bout length); (ii) number of dives; (iii) dive duration; (iv) success of each dive – judged by whether prey was brought to the surface; and (v) prey size – estimated by comparison with bill length and classified as either small, medium or large based on the three sizes of fish stocked.

20.3 Results 20.3.1

Numbers of fish recovered

The numbers of fish of each size category recovered following each trial are shown in Table 20.2. In the covered pond, there was an overall recovery of 96% for the first trial,

20 11/02/2003 15:06 Page 282

282

Mitigation measures

Table 20.2 Recovery of carp following trial 1 (without refuges) and trial 2 (with refuges) Pond

Fish size

Number stocked

Number recovered % recovered (live and dead)

Trial 1 (without refuges) Covered Large Medium Small Total Open Large Medium Small Total

51 104 256 411 54 103 257 414

51 103 240 394 32a 31b 93 156

100.0 99.0 93.8 95.9 59.3 30.1 36.2 37.7

Trial 2 (with refuges) Covered Large Medium Small Total Open Large Medium Small Total

50 100 287 437 50 100 287 437

50 97 214 361 33 27 29 89

100.0 97.0 74.6 82.6 66.0 27.0 10.1 20.4

Three dead fish (two large, one medium) removed from the open pond during the trial without refuges are not included above: aincludes four recovered dead; bincludes three recovered dead.

with 4% of the fish unaccounted for. As predators were excluded, this loss was probably due to post-stocking mortality or failure to recover. For the second trial, in the covered pond there was an overall recovery of 83% with 17% of the fish stocked unaccounted for. The lower recovery in the second trial was probably due to the poorer quality of fish stocked. In the open pond, 38% of fish stocked were recovered (36% excluding fish recovered dead) for the first trial and 20% for the second. The differences in recovery rates between the open and covered ponds (58% and 62% for the two trials respectively) was assumed to be due to predation. These differences were significant (trial 1: 12  314, P  0.001, trial 2: 12  339, P  0.001). The differences between the number of fish stocked to the open pond and the number subsequently recovered indicate the number of fish unaccounted for, and presumably lost to predation, over the course of each trial. In the first trial, 238 of such fish were lost (20 large, 70 medium and 148 small), representing an estimated weight of 18.3 kg. In the second trial, 272 fish were lost (17 large, 70 medium and 185 small), representing an estimated weight of 20.4 kg.

20.3.2

Wounding rates of recovered fish

The frequency of damage to the recovered fish in each trial is shown in Table 20.3. In both trials a small number of fish with evidence of bird damage was recovered from the

20 11/02/2003 15:06 Page 283

Table 20.3 Levels of damage to recovered fish: trial 1 (without refuges) and trial 2 (with refuges). S  small, M  medium, L  large

Condition of fish

S

Trial 2 (with refuges)

Fish grade

Fish grade

M

100

102

99

49

96

0 0 0 0 0

1 0 0 0 0

1 0 0 0 0

0 2 0 0 0

0 4 0 0 0

100

103

100

51

92

99

25

80

1 0 0 0 0

1 0 0 0 0

2 0 1 0 3

93

100

31

%

L

%

S

%

M

%

L

%

212 2 0 0 0

99 1 0 0 0

95 1 1 0 0

98 1 1 0 0

38 8 4 0 0

76 16 8 0 0

100

0 214

0 100

0 97

0 100

0 50

0 100

7

22

28

97

25

93

19

58

6 0 3 0 10

2 0 19 2 2

6 0 59 6 6

100

32

100

1 0 0 0 0 29

3 0 0 0 0 100

2 0 0 0 0 27

7 0 0 0 0 100

12 1 0 0 1 33

36 3 0 0 3 100

Fish refuges and cormorant predation

%

Covered pond 240 No evidence of bird damage Evidence of bird damage: Scale loss (symmetrical) – single beak marks 0 Old beak marks and scars 0 Multi-beak marks, wounds and punctures 0 0 Dead – with wounds/beak marks 0 Dead – scavenged (remains only) Remains only – bones and scales 240 Total Open pond No evidence of bird damage Evidence of bird damage: Scale loss (symmetrical) – single beak marks Old beak marks and scars Multi-beak marks, wounds and punctures Dead – with wounds/beak marks Dead – scavenged (remains only) Remains only – bones and scales Total

Trial 1 (without refuges)

283

20 11/02/2003 15:06 Page 284

284

Mitigation measures

covered pond consistent with the stocking of all previously damaged fish to this pond. In the open pond, 27 (18%) of the fish recovered in the first trial had evidence of recent bird damage, compared with 15 (17%) in the second trial. The extent of bird damage was related to fish size: in the first trial 77% of the large fish had wounds (excluding scavenged fish), compared with 11% of the medium and 1% of the small fish. The difference in wounding rate between the large fish and the medium and small fish combined, was significant (12  88, P  0.001). Most of the large fish had puncture and flesh wounds characteristic of cormorant damage; no fish was recorded with stab wounds (characteristic of herons). In the trial with refuges, however, only 38% of large fish had wounds, and the difference between the trials was significant (12  5.6, P  0.05). Also, the severity of damage to individual fish was lower in the trial with refuges, with no fish having multiple wounds (82% of large fish had such damage in the earlier trial) and fewer dead fish recovered.

20.3.3

Bird observations

During the first trial, observations were made on 28 days totalling 269 h (247 h excluding video surveillance). During the second trial, observations were made on 12 days totalling 84 h (65 h excluding video surveillance). During the first trial, a total of 395 cormorants were counted flying overhead (1.6 birds h1 of observation) and 29 cormorants foraged in the pond (0.11 birds h1). Of these, focal observations were made on 22 birds. During the second trial, 159 cormorants were counted flying overhead (2.4 birds h1) and 11 cormorants foraged in the pond (0.13 birds h1). Of these, focal observations were made on nine birds. Focal observations could not be made using the video recordings. Analysis of the foraging parameters indicates (Table 20.4) that mean dive duration was significantly longer in the trial with refuges (17 s) compared with the trial without refuges (12 s, t  6.7, d.f.  28, P  0.001). Bout length, dives per bout and success rate did not differ significantly between trials (P  0.05).

Table 20.4 refuges)

Cormorant foraging parameters during trial 1 (without refuges) and trial 2 (with Mean  SE

Foraging parameter Sample size (number of focal observations) Bout length (minutes) Dives per bout Dive duration (seconds) Success (fish caught and swallowed per minute)

With refuges 9 12.00  1.69 23.88  3.42 17.09  0.78 0.08  0.04

Without refuges

d.f.

t

P

29 28 28 29

0.40 0.43 6.70 1.35

NS NS 0.001 NS

22 10.75  1.88 26.50  3.46 11.99  0.36 0.04  0.01

20 11/02/2003 15:06 Page 285

Fish refuges and cormorant predation

285

During the first trial, 13 herons, Ardea cinerea L., and five kingfishers, Alcedo atthis, were observed in the vicinity of the open pond. During the second trial, seven herons and two kingfishers were observed. No fish were seen to be caught by birds other than cormorants in either trial.

20.4 Discussion The effect of wire mesh fish refuges on cormorant foraging was investigated. Cormorant numbers, foraging behaviour and fish losses and wounding were compared between two sequential 6-week trials (one without refuges and one with). Mean dive duration was significantly longer when refuges were present (17 s) than when absent (12 s). In addition, a smaller proportion (38%) of the large fish recovered from the open pond (with refuges) had injuries consistent with recent bird damage compared to the trial without refuges (77%). These differences between the two trials might be explained if the refuges were reducing the availability of fish, particularly the large size class, to foraging cormorants. By contrast, estimated fish losses were similar between the trials: 272 fish (62%) with refuges and 238 fish (58%) without. However, the rate at which cormorants visited the open pond was higher in the second trial: 0.13 birds h1 of observation with refuges compared with 0.11 birds h1 without (although the difference was not statistically significant: McKay et al. 1999). This suggests that cormorant foraging may have been less effective in the trial with refuges, in which an estimated 5.7 fish (426 g) were ‘lost’ per cormorant visit, compared with 6.1 fish (470 g) without refuges (details in McKay et al. 1999). However, focal sampling revealed no significant difference in observed capture rates between trials (although sample sizes were low). Variability in the effectiveness of a refuge might arise from the design or differences in micro-habitat selection between size classes of fish. Larger juvenile Atlantic salmon, Salmo salar L., remained in refuges significantly longer after being frightened than smaller fish (Mikheev et al. 1994). This was interpreted as being due to the greater caution and higher energy reserves of larger fish. Also, when shelters were limited, young chub, Leuciscus cephalus (L.), did not seek refuge, whereas adults aggregated in the shelters (Fraser 1983). It is likely that the optimum design of fish refuges will differ between size classes and species, and more work is required to investigate refuge design in relation to fish behaviour. The experimental design used in the present study (two sequential trials) was not ideal, as the results were confounded by time. Therefore, factors other than the presence of refuges may explain the differences between the trials. The quality of the fish stocked was poorer in the trial with refuges, which was reflected by a higher poststocking mortality. Also, observations suggested that more birds were present at the study site during the second trial (with refuges). This is consistent with inland numbers of cormorants tending to peak in December/January (Rehfisch et al. 1999). However, it is difficult to see how these factors could account for the observed differences in cormorant dive duration or fish injury rates. Variations in environmental conditions were not thought to be a confounding factor; the weather was generally mild (3–14°C) throughout both trials and water temperatures also remained similar.

20 11/02/2003 15:06 Page 286

286

Mitigation measures

In summary, this study has provided some evidence that refuges may reduce the incidence of wounding to larger size classes of carp (which are more valuable and sought after by anglers), and decrease the availability of carp to birds. Due to confounding factors, results could not be used to test the effect of refuges on fish stocks or number of foraging cormorants. Further study, therefore, is recommended to test these effects, and also to investigate refuge design in relation to fish behaviour, for different size classes and species.

Acknowledgements Thanks are due to the Environment Agency and particularly Mark Pilcher and Tom Cousins for help and advice throughout the study. Karl Evans and John Davies helped with the bird observations. Dan Kinsman, Mark Ives, Stuart Ives and Rosemarie Gaines helped with fish stocking and recovery. Thanks also to Kevin Roberts and Graham Wilton-Jones for their onsite assistance. Steve Langton gave advice on statistics and Bob Furness, Steve Hunter and Pete Robertson helped revise previous versions of the manuscript.

References Carss D.N. (1993) Cormorants Phalacrocorax carbo at cage fish farms in Argyll, W. Scotland. Seabird 15, 38–44. Fraser D.F. (1983) An experimental investigation of refuging behaviour in a minnow. Canadian Journal of Zoology 61, 666–672. Gliwicz Z.M. & Jachner A. (1992) Diel migrations of juvenile fish: a ghost of predation past or present? Archiv für Hydrobiologie 124, 385–410. Grémillet D. & Wilson R.P. (1999) A life in the fast lane: energetics and foraging strategies of the great cormorant. Behavioral Ecology 10, 516–524. Hughes B., Kirby J. & Rowcliffe J.M. (1999) Waterbird conflicts in Britain and Ireland: Ruddy Ducks Oxyura jamaicensis, Canada Geese Branta canadensis, and Cormorants Phalacrocorax carbo. Wildfowl 50, 77–99. Matkowski S.M.D. (1989) Differential susceptibility of three species of stocked trout to bird predation. North American Journal of Fisheries Management 9, 184–187. McKay H., Furness R., Russell I., Parrott D., Rehfisch M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. Report to the UK Ministry of Agriculture, Fisheries and Food, London. 256 pp. Mikheev V.N., Metcalfe N.B., Huntingford F.A. & Thorpe J.E. (1994) Size-related differences in behaviour and spatial distribution of juvenile Atlantic salmon in a novel environment. Journal of Fish Biology 45, 379–386. Pilcher M.W. & Feltham M.J. (1997) An Assessment of Cormorant Depredation on Stillwater Fish Populations in the Lea and Colne Valleys of the Thames Catchment. R&D Technical Report W101. Bristol: Environment Agency, 64 pp. Rehfisch M., Wernham C.V. & Marchant J.H. (eds) (1999) Population, Distribution, Movements and Survival of Fish-Eating Birds in Great Britain. Report to the Department of the Environment, Transport and the Regions. London: DETR, 360 pp.

20 11/02/2003 15:06 Page 287

Fish refuges and cormorant predation

287

Russell I.C., Dare P.J., Eaton D.R. & Armstrong D.A. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the UK Ministry of Agriculture, Fisheries and Food, Project VCO104. London: MAAF, 130 pp. Suter W. (1995) Are cormorants Phalacrocorax carbo wintering in Switzerland approaching carrying capacity? An analysis of increase patterns and habitat choice. Ardea 83, 255–266.

21 11/02/2003 15:08 Page 288

Chapter 21

Impact of cormorants on the Loch Leven trout fishery and the effectiveness of shooting as mitigation G.A. WRIGHT* Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, UK

Abstract Owing to a perceived conflict with the Loch Leven trout fishery, cormorants have been shot in large numbers. This chapter summarises changes in cormorant numbers, fish populations, angling catches and angling effort over 32 years, and seeks evidence of detrimental impact of cormorants on the fishery and of beneficial effects of large-scale cormorant shooting. There was no evidence of a reduction in the brown trout, Salmo trutta L., population as a result of increased wintering cormorant numbers, or evidence of any adverse impact on angling catch. Catch per unit effort remained relatively stable despite the cormorant increase, and the principal determinant of angling catch was angling effort. The proportion of brown trout found to be wounded by cormorants was low. There was no evidence of a reduction in wintering cormorant numbers, or of an increase in angling catches, as a consequence of shooting large numbers of cormorants. Keywords: licensing, pest control, piscivorous birds.

21.1 Introduction For centuries, fishery managers have persecuted cormorants, Phalacrocorax species, and other fish-eating birds because it is perceived they consume large quantities of exploitable fish, wound fish that may then be unmarketable or die, and cause stress and abnormal behaviour in fish resulting in reduced availability to anglers (Russell et al. 1996). However, to date there is little empirical evidence to support the damage done by cormorants on fish populations in natural fresh waters (e.g. Carss 1997). Furthermore, there is little evidence of any beneficial effects of shooting cormorants for fishery protection purposes (e.g. Kirby et al. 1996; Russell et al. 1996), and research which aims to address these issues is therefore of particular importance in helping to differentiate between perception and reality. Increases in cormorant numbers throughout Europe are well documented (e.g. Blanco et al. 1994; Kirby et al. 1995; van Eerden & Gregersen 1995). In Scotland, cormorants are given statutory protection under the Wildlife and Countryside Act, 1981, *Correspondence: Gordon Wright, Corran Bhan, Portincaple, By Garalochhead, Argyll G84 0ET, UK (email: [email protected]).

21 11/02/2003 15:08 Page 289

Cormorant impact and effectiveness of shooting

289

which implements European Community Directive 79/409/EEC on the Conservation of Wild Birds (The Birds Directive). Additionally, some Scottish cormorant breeding sites are designated as Special Protection Areas (SPAs) under European Community Directive 92/43/EEC on the Conservation of Natural Habitats and of Wild Flora and Fauna (The Habitats Directive). In November 1994, Loch Leven was notified as a potential SPA given its populations of breeding, wintering and migratory wildfowl, which includes a nationally-important wintering population of cormorants, and this led to the cessation of cormorant shooting. Cormorants were recorded on the loch as far back as 1791. In the 1940s they were present in very large numbers with reports of over 200 birds, and in 1974 they were present throughout the year, with up to 60 birds from January to March (Allison et al. 1974).

21.2 The Loch Leven trout fishery Loch Leven has a surface area of 13.3 km2 and an average depth of 3.9 m. It has supported a commercial fishery for at least 680 years and in 1873 angling replaced netting as the means of exploitation (Thorpe 1974a). As a result of concerns over declining catches, brown trout Salmo trutta (L.) of Loch Leven origin have been stocked into the loch since 1983. Each spring, 120-mm juveniles are released into the loch, with a view to their achieving a catchable size by the following spring. In addition, stocking of rainbow trout, Oncorhynchus mykiss (Walbaum), commenced in 1993. The fishery has detailed records of catches, fishing effort and stocking, and together these factors make Loch Leven a suitable site to study the impact of cormorants on a recreational fishery. The size of the brown trout population has been estimated on several occasions: viz. capture/recapture (Thorpe 1974b), gill netting (O’Grady et al. 1993) and a combination of gill netting, trawling and hydroacoustics (Alexander et al. 1999), although obtaining an accurate estimate on a water body as large as Loch Leven is difficult. Over recent decades nutrient enrichment from point-source and diffuse pollution has resulted in a deterioration of water quality and episodes of dense blooms of bluegreen algae (Bailey-Watts et al. 1994). This has had direct consequences for the fishery by making the loch less attractive to fishermen, and on occasions by raising fears of the possible toxic effects of the algal blooms. It has also resulted in wide fluctuations in pH, which may influence fish survival. Analysis of the stomach contents of cormorants shot at Loch Leven for fishery protection purposes showed that they do take brown trout. Carss & Marquiss (1992, 1994) showed a general shift from a perch, Perca fluviatilis L., dominated diet in the 1970s to a brown trout dominated diet in the 1980s and 1990s. Alexander et al. (1999) reported a marked change in the loch’s fish population with the resurgence of perch, and in the netting samples perch outnumbered brown trout by 2.3 : 1, but with many perch showing signs of cormorant wounding. This supported the evidence of regurgitates and pellet samples collected from 1997 to 1999, which indicated a high proportion of perch in the cormorant diet (unpublished data).

21 11/02/2003 15:08 Page 290

290

Mitigation measures

Carss & Marquiss (1992, 1994) showed that brown trout taken by cormorants were predominately in the 16 to 35 cm length range, with about half being of a catchable size for anglers. Rainbow trout were predominately in the 26–40 cm range, and most were of catchable size. Although the potential for economic loss to the fishery from cormorant predation was identified, Carss & Marquiss (1992) concluded that this was far from established. Carss et al. (1997) surmised that bird predation had no detectable effect on angling harvest, but this was not demonstrated quantitatively. This chapter reviews data gathered between 1968 and 2000, and tests the following hypotheses that (a) increased cormorant numbers have damaged the Loch Leven trout fishery by depleting the trout population and reducing angling catches; and (b) shooting cormorants on Loch Leven is an effective mitigation measure, resulting in reduced cormorant numbers and increased trout catches.

21.3 Material and methods Data on numbers of cormorants wintering on Loch Leven from 1968/1969 to 1999/2000 were extracted from field notebooks (A. Allison, unpublished data; G. Wright, unpublished data), and from monthly National Waterfowl Census records. Loch Leven has elevated observation points to enable the whole site to be overlooked, and roosting and loafing sites are well known. Counts were conducted by professional field staff once per month from September to March. In addition, multiple daily counts were conducted (as described in Chapter 26), and no significant diurnal variation was found. Variation between individual counters was also tested (as described in Chapter 26), and although there was some scope for error due to unobserved bird movements, the counts were assumed to have a reasonably high degree of consistency. Annual indices of cormorant abundance from 1986/1987 to 1999/2000 were obtained from the Wildfowl and Wetlands Trust. Details of fishing effort (angler days), fish catches and stocking from 1968/1969 to 1999/2000, and records of cormorants shot on Loch Leven from 1981/1982 to 1999/2000, were obtained from Loch Leven Fisheries Company. Between June and August 1998 inclusive, 246 brown trout were caught with multimesh gill nets, 30 m long with twelve panels from 5 to 55 mm knot to knot mesh size, set at 35 sites in a range of depths. Five pelagic zones were also sampled using verticallyset nets. All fish were examined for signs of cormorant damage, as described by Russell et al. (1996), and the lengths (FL, mm) and weights (g) of all fish were recorded. The netting was repeated in February and March 1999 to detect changes in the trout population over the 1998/1999 winter. Twenty-four sites were sampled using multi-mesh gill nets, 60 m long with twelve panels from 8 mm to 50 mm half mesh size, set on the bottom in a range of depths. Three pelagic zones were also sampled with vertically-set nets. A total of 424 brown trout were caught and examined as previously. Damage was classed as fresh where the wound had been inflicted recently, or old where signs of healing were evident.

21 11/02/2003 15:08 Page 291

Cormorant impact and effectiveness of shooting

291

21.4 Results 21.4.1

Wintering cormorant population

A substantial increase in the number of cormorant was found between 1968/1969 and 1999/2000 (Fig. 21.1), equivalent to an annual growth rate of 12.5% up to the peak in 1994/1995, or an annual rate of 11% taken over the whole period. Between 1987/1988 and 1994/1995 the annual increase was 23%, compared with average annual increases of 12.6% for estuaries and 24.8% for gravel pits (Russell et al. 1996). This information, together with the reference made earlier to a large number of cormorants present during the 1940s (Allison 1974), demonstrates that cormorant numbers have long been subject to fluctuation, and the current large wintering population is not an entirely new phenomenon.

21.4.2

Cormorant impact on the trout population since 1968

The relationship between winter trout population estimates and catches the following summer is shown in Table 21.1. Angling catches were principally drawn from age 3 fish. It should be noted that survey methods yielded contrasting results in 1998. In addition, the use of the 260mm length class in 1998 excluded the smaller sizes of the age 3 class (Thorpe 1974b), so when compared with earlier totals, the 1998 results should be regarded as underestimates. It is apparent that trout populations in 1993 and 1998 were no smaller than they were in 1968–1971, and there is no evidence of an impact on the brown trout population of vastly increased wintering cormorant numbers. It is also 500

Mean monthly cormorant count

450 400 350 300 250 200 150 100 50

68 /6 9 70 /7 1 72 /7 3 74 /7 5 76 /7 7 78 /7 9 80 /8 1 82 /8 3 84 /8 5 86 /8 7 88 /8 9 90 /9 1 92 /9 3 94 /9 5 96 /9 7 98 /9 9

0

Winter

Figure 21.1 Mean ( standard errors) monthly cormorant count for the winters 1968/1969 to 1999/2000

21 11/02/2003 15:08 Page 292

292

Mitigation measures

Table 21.1

Loch Leven brown trout population estimates and angling catches Population Confidence estimate limits

Year Method 1968 1969 1970 1971 1993 1998 1998 1998 1998

Mark/recapture Mark/recapture Mark/recapture Mark/recapture Netting Netting/sonar Netting/sonar Mark/recapture Mark/recapture

126 665 103 497 114 526 52 737 186 000 217 000 48 000 555 000 157 000

Age, size range

Age 3 68 901–172 363 Age 3 82 438–171 938 Age 3 39 546–73 969 Age 3 Age 3 162 mm 260 mm 192 200 162–260 mm 260 mm

Percent 3 Angling & 260 mm catch caught 37 796 20 605 20 331 9 571 13 150 4 122 4 122 4 122 4 122

29.8 19.9 17.8 18.1 7.1 8.6 2.6

Table 21.2 Brown trout gill net catch per unit effort before and after winter 1998/1999 (mean cormorant count September to March  270)

Catch date Jun–Aug 1998 Feb–Mar 1999

Net area set (m2)

Mean time set (h : m)

No. brown trout caught

1800 2160

16 : 53 23 : 15

246 424

No. brown trout (m2 d1) 0.194 0.203

No. perch caught

No. perch (m2 d1)

567 59

0.448 0.028

apparent that as a percentage of the population, the trout catch has declined markedly from 20 to 30% of the 3 trout stock, to less than 10%, possibly as low as 2.6%.

21.4.3

Cormorant impact on the trout population during the 1998/1999 winter

Catch per unit effort (CPUE) data from gill net sampling in June–August 1998, before wintering cormorants arrived, was compared with data from February to March 1999 when a mean of 270 cormorants had been present for five months. Netting effort was higher in February/March to compensate for anticipated fish stock depletion by cormorants. However, CPUE was very similar for both brown trout samples (Table 21.2), providing no direct evidence of depletion of the population over the winter. By contrast, the perch CPUE declined by over 90% during the same period. The methods were not identical, and there may be changes in trout mobility or habitat specificity, which could influence the CPUE.

21.4.4

Angling catch per unit effort

Figure 21.2 shows fluctuations in the brown trout CPUE for the 1974–1991 angling seasons with a five-year moving mean, and the mean monthly cormorant count for the preceding winters. Data from 1992 were omitted as the loch was badly affected by a

21 11/02/2003 15:08 Page 293

Cormorant impact and effectiveness of shooting

293

Figure 21.2 Brown trout catch per unit effort and mean monthly cormorant count for preceding winter, 1974–1991

severe algal bloom causing the closure of the fishery, and from 1993 it changed to a mixed brown trout and rainbow trout fishery. While the mean monthly cormorant count increased from less than 20 to over 200, CPUE fluctuated around 1.8 fish boat1 throughout the period. Although the lowest CPUE coincided with the highest preceding mean monthly cormorant count in 1991, the highest CPUE recorded in 1989 followed the second highest preceding mean monthly cormorant count. A significant relationship was found between angling effort and brown trout catch in the years before rainbow trout stocking began (rs  0.716, P  0.001), thus angling effort appears to be the principal determinant in the size of fish catches.

21.4.5

Cormorant impact on total catch

No correlation was found between cormorant numbers and the number or weight of all fish caught in all years, or the number or weight of brown trout caught before rainbow trout stocking began (Table 21.3). There was no correlation between cormorant numbers and the mean weight of brown trout caught, or the CPUE number or weight of all fish caught, or the CPUE number or weight of brown trout caught before rainbow trout stocking began.

21.4.6

Wounding of fish by cormorants

Of the 246 brown trout caught in July to August 1998, only one fish (0.4%) was wounded. This contrasts with a sample of 567 perch caught at the same time of which

21 11/02/2003 15:08 Page 294

294

Mitigation measures

Table 21.3 Spearman rank correlations: number of cormorants shot and mean monthly cormorant count during same and subsequent winters (all correlations were not significant at   0.05) Analysis

Years

n

rs

P (1-tailed)

Mean cormorant count during same winter Mean cormorant count during second winter Mean cormorant count during third winter Mean cormorant count during fourth winter

1981/82–1998/99 1981/82–1998/99 1981/82–1998/99 1981/82–1998/99

16 16 16 15

0.215 0.012 0.015 0.163

0.211 0.483 0.476 0.281

Figure 21.3 Mean cormorant count and number of cormorants shot between 1981/1982 and 1999/2000

21.5% showed signs of cormorant wounding. At the time there were 20–30 cormorants summering on the loch. Of the 424 brown trout caught in February and March 1999, 11 had old wounds and nine had fresh wounds, giving an overall proportion of 4.7% wounded fish. At the time there were around 200 cormorants wintering on the loch.

21.4.7

Effect of shooting

Over 1400 cormorants have been killed since 1981/1982, with over 700 killed during the three winters 1988/1989 to 1990/1991 (Fig. 21.3). There was no indication of a reduction in cormorant numbers during the same winter or any of the three subsequent winters as a result of shooting (Table 21.3), nor was there a relationship between the

21 11/02/2003 15:08 Page 295

Cormorant impact and effectiveness of shooting

295

Table 21.4 Spearman rank correlations: cormorants shot and subsequent angling catch. (all correlations were not significant at   0.05) Analysis

Years

n

rs

P (1-tailed)

Number of all fish caught during summer after shooting Weight of all fish caught during summer after shooting CPUE number of fish during summer after shooting CPUE weight of fish during summer after shooting Number of brown trout caught during summer after shooting Weight of brown trout caught during summer after shooting Brown trout mean weight during summer after shooting CPUE number of brown trout during summer after shooting CPUE weight of brown trout during summer after shooting

1981/82–1998/99

16

0.068

0.401

1981/82–1998/99

16

0.001

0.498

1981/82–1998/99

16

0.034

0.450

1981/82–1998/99

16

0.063

0.408

1981/82–1991/92

11

0.364

0.135

1981/82–1991/92

11

0.223

0.255

1981/82–1991/92

11

0.187

0.291

1981/82–1991/92

11

0.223

0.255

1981/82–1991/92

11

0.196

0.282

number of cormorants shot and the total fish catch, CPUE or mean brown trout weight during the following summer, either for all years or for only those years before rainbow trout stocking (Table 21.4).

21.5 Discussion Despite the general protection afforded by the Wildlife and Countryside Act, cormorants may be shot for the purpose of preventing serious damage to fisheries, and to this end licences may be issued by the appropriate government department. This raises the possibility of birds that are protected on breeding and wintering grounds by UK and European legislation, being legally shot elsewhere. This is a particular issue in Scotland, as declines in breeding cormorants in north-west Scotland, including on some Special Protection Areas, may reflect the impact of shooting (Russell et al. 1996). Applicants for licences to shoot cormorants are required to provide evidence to support their claim that serious damage is occurring. Marquiss & Carss (1994) argued that damage cannot be considered serious if it cannot be measured. They also argued that to justify the issue of a licence to kill birds, losses have to be shown to be the result of bird predation. In practice, hard evidence of serious damage attributable to cormorants has proved elusive, and in effect, the presence of cormorants in the vicinity of a fishery is deemed synonymous with serious damage occurring. This general assumption may be based on inappropriate comparisons with the situations at fish farms or intensively stocked fisheries, or on the interpretation of results from poorly planned experiments lacking in scientific rigour (e.g. Marquiss & Carss 1994; Russell et al. 1996).

21 11/02/2003 15:08 Page 296

296

Mitigation measures

The Loch Leven data on wintering cormorant numbers, angling catches, fish stocking, angling effort and cormorant shooting offered a unique opportunity to explore longterm, inter-relationships between these factors on an internationally important wetland and commercial trout fishery. No evidence of any detrimental effect of wintering cormorants on fish stocks or angling catches, and thus no evidence to support the assumption that they are causing serious damage, was found for this fishery. Furthermore, CPUE was relatively stable despite the increase in cormorant numbers and there was strong evidence that the principal factor determining angling catch was angling effort, which is likely to be influenced by many external factors, including increased competition from other fisheries. Hypothesis (a) is therefore rejected. However, it should be recognised that cormorant predation is a site-specific issue and more prevalent on small, still waters (Feltham et al. 1999), than on large systems such as Loch Leven. The 4.7% of gill-netted brown trout showing signs of cormorant damage in late winter – an increase from 0.4% at the start of the winter – demonstrated that cormorants are wounding trout of catchable in respect of angling. The large sample size and low percentage wounded suggest that wounding is unlikely to have a major impact on the fishery. However, wounded fish, even those that show evidence of healing, are considered undesirable to anglers, so this must be considered as a loss to the fishery. Shooting is widely used in Britain as a mitigation measure (e.g. McKay et al. 1999). The Loch Leven data show no beneficial effects in terms of reduced cormorant numbers or improved fish catches as a result of large-scale shooting, which raises questions as to the validity of issuing licences for cormorant control. Hypothesis (b) is therefore rejected. However, whether the results are transferable to other sites where cormorants are perceived to have an impact is unclear. McKay et al. (1999) showed that site size may be an important determinant in this debate. Much of the problem relates to whether shooting markedly reduces cormorant density in a given location, or whether the local cormorant population size remains relatively stable because it is replaced from elsewhere, being within the foraging range of a much larger population. Wright (Chapter 26) showed that the latter scenario occurred at Loch Leven, so response to shooting was not effective. The debate about the effectiveness of shooting to control cormorant density is likely to continue until a marked reduction in cormorant numbers can be achieved on a regional-scale fishery, and this is unlikely under current legislation. Consequently, it is considered that the use of shooting as a control method is of marginal value and serves principally to appease the concerns of anglers.

Acknowledgements My thanks go to Bob Furness for his helpful comments on the draft paper, and to the University of Glasgow, the Open University Crowther Fund, Scottish Natural Heritage and Scottish Office Agriculture, Environment and Fisheries Department for financial and material support. This study involved many hours of fieldwork in all weathers, and I am most grateful for the efforts of Alan Lauder and Paul Brooks who made it possible. Additional data were supplied by Willie Wilson of Loch Leven Fisheries and by The Wildfowl and Wetlands Trust, and Alan Allison gave access to his field notes and diaries.

21 11/02/2003 15:08 Page 297

Cormorant impact and effectiveness of shooting

297

References Allison A., Newton I. & Campbell C. (1974) Loch Leven National Nature Reserve; A Study of Waterfowl Biology. WAGBI Conservation Publication, Sevenoaks: The Caxton & Holmesdale Press, 124 pp. Alexander G., Adams C.E., Devine J., Drummond J. & Bean C.W. (1999) Loch Leven: Hydroacoustic Technique Development and Fish Population Estimation. University of Glasgow Report to Scottish Natural Heritage. Edinburgh: Scottish Natural Heritage, 58 pp. Bailey-Watts A.E., Kirika A., Gunn I.D.M., Bryant C.L. & Wiltshire N.J. (1994) The environment of fish: physics, phosphorus, phytoplankton and fleas. In P. Hutchinson & A. Walker (eds) The Loch Leven Trout Fishery, its Future. Pitlochry: Institute of Fisheries Management, pp. 7–15. Blanco G., Velasco T., Grijalbo J. & Ollero J. (1994) Great cormorant settlement of a new wintering area in Spain. Colonial Waterbirds 17, 173–180. Carss D. & Marquiss M. (1992) Cormorants and the Loch Leven Trout Fishery. Institute of Terrestrial Ecology Project T135c1 Report to Scottish Natural Heritage. Edinburgh: Scottish Natural Heritage, 36 pp. Carss D. & Marquiss M. (1994) The Stomach Contents of Cormorants from Loch Leven, 1992–94. Institute of Terrestrial Ecology Project T02073c1 Report. Banchory Research Station, 21 pp. Carss D., Marquiss M. & Lauder A. (1997) Cormorant Phalacrocorax carbo carbo predation at a major trout fishery in Scotland. Supplemento alle Ricerche Biologia della Selvaggina XXVI, 281–294. Carss D. (ed.) (1997) Techniques for assessing cormorant diet and food intake: towards a consensus view. Supplemento alle Ricerche Biologia della Selvaggina XXVI, 197–230. Feltham J.M., Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999) Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 1006. London: MAAF, 406 pp. Kirby J.S., Gilburn A. & Sellers R.M. (1995) Status distribution and habitat use by cormorants Phalacrocorax carbo wintering in Britain. Ardea 83, 93–102. Kirby J.S., Holmes J.S. & Sellers R.M. (1996) Cormorants Phalacrocorax carbo as fish predators: an appraisal of their conservation and management in Great Britain. Biological Conservation 75, 191–199. Marquiss M. & Carss D.N. (1994) Avian Piscivores: Basis for Policy. Institute of Terrestrial Ecology report 461/8/N&Y to National Rivers Authority. Bristol: National Rivers Authority, 104 pp. McKay H.V., Furness R.W., Russell I.C., Parrott D., Rehfisch M.M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. London: Report to the Ministry of Agriculture, Fisheries and Food, project VC 0107. London: MAAF, 254 pp. O’Grady M.F., Gargan P. & Roche W. (1993) A Fish Stock Survey of Loch Leven and Management Proposals for this Resource as a Trout Fishery. Edinburgh: Scottish Natural Heritage, 37 pp. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0104. London: MAAF, 130 pp. Staub E. (1997) Cormorant Phalacrocorax carbo predation and conflicts with species conservation and fisheries in Switzerland. Ekologia Polska 45, 309–310. Thorpe J.E. (1974a) Trout and perch populations at Loch Leven, Kinross. Proceedings of the Royal Society of Edinburgh, B. 74, 295–314. Thorpe J.E. (1974b) Estimation of the number of brown trout Salmo trutta (L.) in Loch Leven, Kinross, Scotland. Journal of Fish Biology 6, 135–152. Van Eerden M.R. & Gregersen J. (1995) Long-term changes in the northwest European population of cormorants Phalacrocorax carbo sinensis. Ardea 83, 61–79.

22 11/02/2003 15:10 Page 298

Chapter 22

Enhancing stocks of European eel, Anguilla anguilla, to benefit bittern, Botaurus stellaris BRIAN KNIGHTS* Applied Ecology Research Group, University of Westminster, London, UK

Abstract The European eel Anguilla anguilla (L.) often comprises a major component of the diet of bittern, Botaurus stellaris (L.), and other endangered piscivores. The life cycle, physiological and behavioural characteristics of the eel, which potentially make it an ideal prey species, especially in reed bed habitats, are reviewed. Eel recruitment to Europe is, however, currently relatively low, possibly posing problems for the recovery of bittern in Britain. Impacts are greatest in habitats on southern North Sea coasts, the Baltic and Mediterranean areas and on the Iberian Peninsula, these being the most distant from the Atlantic migration pathways. In rivers, the density of eels tends to decline with distance from tidal limits. Biomass, however, tends to be maintained because the average size of individuals increases, with a preponderance of eels maturing longer and emigrating as large females. Commercial glass eel fisheries and migration barriers (at tidal limits and due to water-level management structures inland) may further reduce recruitment. This leads to relatively low densities and a paucity of smaller eels. Site-specific studies are essential to assess such recruitment problems and possible solutions. As examples, important UK bittern reed bed reserves on the west coast (Leighton Moss, Lancashire) and east coast (Minsmere, Suffolk) of England are discussed. The cost–benefits of fishery controls, eel passes, trapping, stocking, habitat management and the values of partnership research, funding and management are reviewed. Keywords: Anguilla, Botauris stellaris, commercial fishing, management, migration barriers, recruitment.

22.1 Introduction Bittern, Botauris stellaris (L.), and mammals such as otters, Lutra lutra (L.), are currently restricted in numbers and distribution in England and Wales and many other European countries. Loss of habitat (e.g. of reed beds and productive marginal littoral fringes due to land drainage for agriculture) has had a major impact. For bittern, cessation of the management of wetlands for commercial reed cutting and subsequent habitat changes due to hydroseral succession have been important (Tyler et al. 1998). Changes in water quality, especially those due to pesticides and nutrient enrichment, have impacted aquatic macrophytes and faunal communities, including fish, in many areas. *Correspondence: Brian Knights, Applied Ecology Research Group, University of Westminster, London W1M 8JS, UK (email: [email protected]).

22 11/02/2003 15:10 Page 299

Enhancing stocks of eel to benefit bittern

299

In certain locations, eels have been found to be potentially important prey items for bittern, especially as food for fledglings (Noble et al., Chapter 11). This chapter reviews the biological bases for such preferences, contrasting this to the possibly deleterious impacts of declines in recruitment of eels to Europe, especially to the areas of eastern and southern Britain most remote from the Atlantic migration pathways. Commercial fisheries for glass eels emigrating from the sea could exacerbate recruitment declines, as could physical barriers to migration. Such problems are reviewed, with emphasis on bittern and important reed bed habitats on the west and east coasts of England. The potential benefits of fishery controls, eel passes, stocking and habitat management are discussed.

22.2 Assessing predator selectivity and prey abundance Selectivity for particular prey often appears to be related to prey abundance. However, selectivity is not always easy to prove and the relative abundance of different prey is often difficult to measure accurately. Direct long-term observational studies of feeding are rarely feasible and are impossible during any nocturnal foraging. Behaviour of captive animals cannot be considered as natural. Killing endangered species for postmortem studies of gut contents is also not an option. Therefore, quantitative data have to be gained indirectly and validated as far as possible against observational and other evidence. Prey items can be identified and size and weight measured if fresh or only partially digested regurgitates can be collected. Where this is not possible, a commonly used approach is to identify bones or otoliths in regurgitations or faeces (Wolter & Pawlizki, Chapter 13). Tagging fish with external or PIT-tags and counting retrieved tags in the vicinity of bird roosts has also been used (H. Engström and H. Wickström, personal communications). Such studies, despite difficulties in accurate assessment of actual numbers and sizes of prey consumed, do generally agree with other sources of evidence. They are therefore sufficient to indicate the importance of eels in diets in different locations. Basic information on fish community structures can be gained via electric fishing. Semi-quantitative data such as catch per unit effort (CPUE) can be gained from singlerun surveys, but for accurate assessments of density per unit area, multiple-runs and catch-depletion calculations are required (Naismith & Knights 1990). There are a number of inherent problems in such methods. Surveys can only be conducted in daylight when eels are hiding in burrows or crevices. Eels are generally not efficiently sampled during multi-species electric fishing surveys, especially in saline waters (Naismith & Knights 1990; Knights et al. 1996). Eels are also relatively difficult to see and capture, especially in dense submerged or emergent vegetation and in turbid waters 1 m deep. Catches therefore often increase on successive runs. Using relatively high voltages and focusing on netting eels alone can increase capture efficiencies three to four fold (Knights et al. 2001). Thus the densities and biomasses at many sites discussed in later sections may well be underestimated. Sampling is especially difficult in dense reed beds, thus point abundance sampling methods have to be used (Perrow et al. 1996; Noble et al., Chapter 11).

22 11/02/2003 15:10 Page 300

300

Mitigation measures

Another complicating factor is that densities of eels can vary enormously between sites, even those very close together. Overall accessibility and habitat quality may be important, but Knights et al. (2001) concluded that the main determining factor was the availability of suitable places to hide during the daytime, when sampling occurs. Thus, soft substrates to burrow in are favoured and in other areas the availability of crevices, root masses or dense vegetation (such as reeds) may be critical. For example, exceptionally high densities of 100 eels m2 were found in soft muddy areas near to the outflow of a sewage-treatment works in a tributary of the River Dart (Devon) in SW England (Knights et al. 2001). This gives a distorted idea of the effective population density, as these eels can disperse at night to feed in the main river and estuary just downstream, where electric fishing appeared to show eels were scarce. All these drawbacks must be borne in mind when interpreting data on eel abundance.

22.3 Eels as prey for bittern 22.3.1

Abundance and availability

Eels are ubiquitous in their distribution, being found in all types of fresh, brackish and coastal waters. They have been found, for example, in 13 out of 14 widely separated reed bed nature reserves in England and Wales (Perrow et al. 1996). They are catholic feeders, utilising a wide range of small invertebrate and vertebrate prey, although some larger eels (40–45 cm) may become primarily piscivorous (Tesch 1977). Furthermore, their body form means they can pass between the stems of reeds, and hence deep into the dense reed beds favoured by bittern. Eels can hide between the stems and in the root masses, but can forage in-between the reeds or out into the littoral margins. This provides protection from fish (including larger eels) and other predators. The restricted spaces between the reeds means, however, that when they grow too large, they have to move into the littoral zone or into deeper waters. Densities (mainly of smaller eels) therefore tend to be highest in the reed bed fringes and adjacent littoral zone. This is illustrated by electric fishing surveys of Leighton Moss, an SSSI (Site of Special Scientific Interest), an SPA (Special Protection Area, EC Birds Directive) and Ramsar Site managed by the RSPB (Royal Society for the Protection of Birds) Reserve in Lancashire on the north-west coast of England (Perrow et al. 1996). This 125 ha wetland is of particular interest because it is one of the few extensive shallow Phragmites wetlands in Lancashire and is now one of the few known bittern breeding sites outside of East Anglia (Noble et al., Chapter 11). Perrow et al. (1996) and Noble et al. (Chapter 11) showed that few fish (other than sticklebacks) could penetrate far into the reed beds but the density of eels was still about five eels per 100 m2 at 12 m from the edge (Fig. 22.1). Densities peaked in the productive and ecologically diverse littoral fringe, but other species became dominant in more open waters. These comprised mainly juvenile rudd, Scardinus erythrophthalmus (L.), (100 mm, especially numerous in the spring–summer) and perch. Perca fluviatilis (L.), (numerous in late summer–autumn), plus 10-spined sticklebacks, Pungitius pungitius (L.), and small pike, Esox lucius (L.). The great majority of

22 11/02/2003 15:10 Page 301

Enhancing stocks of eel to benefit bittern

301

Density (fish 100 m⫺2)

25

20

15

10

5

0 Open

Littoral

3m

6m

9m

12 m

Location

Figure 22.1 Leighton Moss, Lancashire: density (per 100 m2  SE) of all fish species () and of eels () in open water, the littoral zone and at different distances (m) into the reed beds. No fish were found 12 m into the reed beds. Adapted from Perrow et al. (1996)

eels electric fished were 25 cm (30 g and maximum girth diameter about 10 mm; Knights 1982). The availability of eels is therefore relatively high for bittern, given that the birds feed while climbing among the stems of reeds, or wading in the shallows of the littoral zone in water 0.25 m deep. Other features that promote the availability of eels in reed beds and littoral zones are their tolerance of the high temperatures and low oxygen levels that can prevail during the summer. Indeed, it is primarily a warm-water species, being most active at temperatures 14–16°C (Tesch 1977; White & Knights 1997a). Conversely, eels are relatively inactive when temperatures fall below 5–10°C, commonly remaining in burrows during winter, often in deeper water where conditions remain more constant. This means eels are most abundant from late spring to late autumn, i.e. throughout the bittern breeding season.

22.3.2

Capture, ingestion and energy value

Eels can be caught when hiding in reed stands during the day and when showing activity at dusk and dawn (Baras et al. 1998). They commonly swim near to the bottom when foraging, but are relatively slow swimmers (generally 2 m s1; Knights & White 1998). They are also relatively easy to swallow once grasped in the beak, being smooth and round-bodied without abrasive scales or spiny fins. Small eels are probably ideal as food items for fledglings. Eels also have a relatively high energy value, especially as they build up fat reserves for over-wintering and as they mature ready for migration back across the Atlantic to the Sargasso Sea (Tesch 1977; Knights 1982). Males typically mature in England and Wales at 30–45 cm after 4–6 years, females at 45 cm after 6 to 15 years.

22 11/02/2003 15:10 Page 302

302

Mitigation measures

22.4 Problems of declining eel recruitment and geographical variations The available evidence suggests that mature (silver) adult European eels migrate to breed in the Sargasso Sea. The genetically separate American eel, Anguilla rostrata (Le Sueur), also appears to breed in the same area, but interbreeding is avoided in some way. Recent studies suggest that there is some genetic separation between populations of A. anguilla from more northerly and more southerly continental growth-stage areas. These possibly result from differences in migration pathways and distances to the Sargasso Sea, and some form of mating separation (Wirth & Bernatchez 2001). Fecundity is very high, to compensate for very high levels of natural mortality, with even the smallest of females producing about one million eggs. The small leptocephalus larvae that hatch migrate via ocean currents to northern Europe, the Mediterranean and N. Africa – a journey that could take one to three or more years (Knights 2002). Leptocephali metamorphose into unpigmented, glass eels as they pass onto the continental shelf and are then attracted to freshwater output from rivers. They enter estuaries when temperatures are 4–6°C, i.e. in the winter in the Biscay areas of France, but not until spring in Britain (Naismith & Knights 1988; White & Knights 1997a). Glass eels use selective tidal-stream transport to move up estuaries, and metamorphose around the point of tidal reversal into pigmented elvers. These start feeding and may begin active upstream migration. Some eels may migrate long distances in their first year in fresh water, but some stay in estuaries or coastal waters. Other eels show step-wise upstream migrations in successive years, with peaks occurring as water temperatures exceed 14–16°C (Naismith & Knights 1988; White & Knights 1997a, b). Populations in the lower reaches of rivers achieve high densities but as eels grow, relative biomass and hence competition for food and space increase. Agonistic encounters may then act as a stimulus for further upstream migrations (Knights 1987). This would explain the general tendencies for population densities to decrease with distance from tidal limits, but for relatively high biomasses to be maintained (Knights et al. 2001). This correlates with an increasingly higher proportion of females that grow for a longer period of time and to a larger size than males before emigrating (Fig. 22.2). In recent decades, recruitment of glass eels to Europe has declined markedly from peaks in the late 1970s to the early 1980s. Commercial catch data indicate recruitment in England and Wales has declined by 75–90% since the 1980s, tending to stabilise in the late 1990s (Knights et al. 2001). These time-trends correlate with those seen elsewhere in Europe, although declines appear to have been less severe and have occurred later than in more northerly Scandinavian and southerly Mediterranean areas. This could be because of the more favourable position of Britain (and the Biscay coast of France) relative to Atlantic oceanic migration pathways, especially SW and NW England and NE Scotland (Knights 2002). Factors such as overfishing and pollution may have had negative impacts. Conversely, causes of declines may largely be natural, for example, owing to changes in Atlantic currents affecting the transoceanic migration of leptocephali. From the few robust long-term data series available, recruitment to Europe was as low in the 1920s–1960s as at present and the peaks of the 1970–1980s may have been rela-tively unusual. Oceanic factors are also suggested because recruitment of A. rostrata to

22 11/02/2003 15:10 Page 303

Enhancing stocks of eel to benefit bittern

303

Density (eels 100 m⫺2 )

40 S WEST N = 138 r 2 = 0.2082

35 30 25 20 15

EAST N = 171 r 2 = 0.1424

10 5 0 0

10

20

30

40

50

60

70

80

Biomass (g 100 m⫺2)

700 600 500 400 S WEST N = 138 r 2 = 0.0405

300 200 EAST N = 171 r 2 = 0.1623

100 0 0

10

20

30

40

50

60

70

80

80 EAST N = 171 r 2 = 0.3275

70

Length (cm)

60 50 40 30

S WEST N = 138 r 2 = 0.2358

20 10 0 0

10

20

30

40

50

60

70

80

Distance from tidal limit (km)

Figure 22.2 Relationships between density biomass and average length of eels with distance (km) from the tidal limits for 11 south-west and 4 east coast rivers in England. See text for further explanation

22 11/02/2003 15:10 Page 304

304

Mitigation measures

North America in recent decades declined along a similar trend to A. anguilla in Europe (Knights 2002). Whatever the causes, falls in recruitment could be leading to shortages of eels as prey for piscivores. Knights et al. (2001) found no evidence of significant declines in stocks. Eel-specific electric fishing surveys detected high intra- and inter-annual variations but no significant long-term trends in stocks in the rivers Severn (despite the major glass eel fishery based there) or Dee (N. Wales). There are, however, a few catchments where migration barriers affect recruitment. Also, inland regions of a few catchments appear to have lower population densities than might be expected. This is possibly because an abundance of suitable downstream habitat(s) can accommodate most of the reduced numbers of recruits. The rivers Piddle and Frome (running into Poole Harbour in Dorset) and the Thames (with a large and productive estuary) fall into this category. It is possible that recruitment is still generally in excess of the carrying capacity of many rivers in the south-west of England but that this is not so for southerly and easterly catchments facing the English Channel and North Sea, most distant from the Atlantic Ocean. Such differences are illustrated in Figure 22.2, drawing on routine electric fishing surveys carried out by the Environment Agency and its predecessor organisations. Data for density, biomass and body length with distance from the tidal limit were combined for 11 rivers in south-west England (Fowey, Tamar, Tavy, Teign, Otter, Exe, Culm, Creedy, Avon, Dart and Erme) and are compared with data for four east coast rivers (Stour, Colne, Blackwater and Chelmer). For clarity, the graphs are replotted from Knights et al. (2001) to show only lines of best fit. The wide scatter is due to the difficulties of efficiently sampling eels. Wide variations in apparent densities between sites at similar distances from the tidal limits also occur because of differences in the availability of daytime refuges (see Section 22.2). Given these provisos, average densities and sizes of eels are much larger in south-westerly than in easterly rivers (Fig. 22.2). In both cases, density declines with distance upstream, but this trend is particularly noticeable in the south-west. The relatively high levels of recruitment to southwest and westerly coasts, and low densities of larger eels in the east, correlate with the distribution of commercial fisheries, i.e. the former targeting glass eels caught by hand-netting, the latter targeting yellow and silver eels using fyke nets and traps. The geographical differences in recruitment, and hence stocks, have important implications for bittern and other piscivores in different regions. Thus, the relatively high abundances of small (20–25 cm) eels found in reed bed habitats in the northwest at Leighton Moss, Lancashire (Fig. 22.1) contrast strongly with the situation seen in Strumpshaw Fen on the east coast in Suffolk (Fig. 22.3). Leighton Moss is a valley fen with streams running into it and a short connection to the sea in Morecombe Bay – a large and productive area of tidal flats. In Strumpshaw, densities were relatively much lower, eels showed a wider distribution of sizes and they were mainly restricted to littoral zones (Perrow et al. 1996). Recruitment is probably further limited at Strumpshaw Fen because it is about 20 km from the sea and although lying next to the tidal River Yare, bunds probably inhibit the entry of immigrants. Similarly low densities and a bias towards larger females 40 cm has been found further upstream in the main River Yare and other east coast rivers (Knights et al. 2001). Strumpshaw was

22 11/02/2003 15:10 Page 305

Enhancing stocks of eel to benefit bittern

305

50

Density (fish 100 m⫺2)

45 40 35 30 25 20 15 10 5 0

Open

Littoral

3m

6m

9m

12 m

Location

Figure 22.3 Strumpshaw Fen, Suffolk: density (per 100 m2  SE) of all fish species () and of eels () in open water, the littoral zone and at different distances (m) into the reed beds. No fish were found 12 m into the reed beds. Adapted from Perrow et al. (1996)

stocked with 3000–6000 and 10 000 elvers in 1994 and 1995 respectively and therefore higher numbers of small eels might have been expected in 1995. However, it is possible that stocking rates were too low to produce significant enhancements of stocks. Possible solutions to the problems of low recruitment and geographical imbalances in eel supply (including stocking) are discussed in more detail later, with special reference to the Minsmere and Leighton Moss reed bed systems.

22.5 Enhancing the availability of eels Options to maintain or augment eel stocks include the use of passes, control of glass eel fisheries (if these could threaten local recruitment) and stocking. Each option requires site-specific studies and field surveys to assess their biological effectiveness and cost–benefits.

22.5.1

The use of eel passes

Natural barriers, such as waterfalls and rapids, can inhibit the initial entry of immigrants into fresh waters and then their dispersion throughout catchments. Problems are exacerbated in many reedbed systems and other low-lying wetlands by man-made barriers. For example, complex sluice systems have to be used in the Somerset Levels in SW England to prevent tidal flooding and to control water levels in different compartments (B. Knights, unpublished report). Glass eels use selective tidal-stream transport to carry them inland, but tidal incursion is often limited by passive tidal flaps or gates, or by controllable sluices. Elvers and older juveniles can swim and climb actively, but

22 11/02/2003 15:10 Page 306

306

Mitigation measures

they show maximal upstream migration at 4–16°C during the summer – a time when rainfall is lowest and sluices have to be closed to maintain water levels (White & Knights 1997a,b). Eels cannot jump and although small eels of 12 cm can climb vertical surfaces using mucus for adhesion, vertical steps 50–60% of the body length are impassable to larger eels. Small eels cannot swim against currents which are faster than 0.2–0.5 m s1, or larger fish against 2 m s1. Eels of all sizes can, however, easily climb very steep slopes via surface irregularities or vegetation. They can also move across land in damp conditions. Site-specific studies are required to assess the impact of particular barriers and to predict the additional upstream recruitment that might be achieved by providing passes (Knights & White 1998). To quantify impact, fishery surveys should be conducted to assess local stock densities and population structures for comparison with data for similar waters and geographical locations. Information can also be gained from eel traps attached to passes. Pass designs need to be chosen according to local conditions. The most important requirement is to provide a means of climbing over or around obstacles. By-pass channels of gentle slope and with low water velocities could be constructed, but the simplest and cheapest solution is to provide rough or weedy surfaces. Alternatively, eel ladders can be used, consisting of troughs or pipes attached to barriers, and provided with suitable climbing material, for example, plant material (such as reeds or heather), geotextiles, synthetic brushes or horticultural netting. Additional important requirements are that a flow of water is needed to attract eels towards a pass, and that entrances and exits need to be sited to facilitate entry and aid escape upstream. The simplest means of maintaining a flow down a pass is via an overflow from a higher level to a lower one, with a direct exit from the pass into the upper level. High flows may, however, inhibit exit from passes. Conversely, overflows may have to be curtailed to maintain upstream water levels, especially during the dry summer months. Pumps can be used, but this involves extra costs and a power supply, together with extra maintenance and servicing needs. Simple flow-fed passes can be cheap to construct, but even these need regular attendance and maintenance to check water flows, remove blockages and to repair them after spates. Fitting traps to passes will provide eels for stocking (Section 22.5.3). These will also provide useful information on eel migration, recruitment, population distributions and dynamics. Workloads will, however, be increased as traps need regular emptying and maintenance.

The value of eel passes at Leighton Moss As discussed earlier, eel stocks appear relatively healthy in Leighton Moss Reserve but there were originally concerns that the drop-board sluice on the main exit stream connecting to the sea inhibited recruitment. This led to the provision of a simple ladder pass from 1997, with a trap to assess recruitment. Capture efficiency of the trap was estimated at about 40% by comparing catches with visual observations of eels passing directly over the pass (D. Mower, unpublished data). Very few glass eels have used the pass, although they have been commonly observed and hand-netted just below the sluice. Instead, immigrants mainly comprise pigmented elvers and small juveniles up to

22 11/02/2003 15:10 Page 307

Enhancing stocks of eel to benefit bittern

307

Table 22.1 Estimates of eel population densities and biomasses in different habitat compartments in Leighton Moss Reserve, Lancashire, UK (see text for further explanation)

Habitat type Reed bed margins Littoral zones Dykes Openwater meres Total Mean

Surface area (ha) 28.8 7.2 6 12.5 54.5

Eel biomass/ compartment (g per 100 m2) 212 327 126 65 – 660

Eel total biomass (kg) 343 1324 76 813 632

Eel density/ Estimated compartment total number (eels per 100 m2) of eels 6 15 7.5 3 – 6.2

17 280 36 330 4 500 3 750 36 330

25 cm. Consequently, peaks of immigration tend to occur from late spring into the summer as water in the shallow reed bed systems warms up. The estimated number of immigrants per year varies widely, from 15 000 to 45 000. To determine how significant this is in relation to electric fishing information on standing stocks in the reserve, the following were used to arrive at the data shown in Table 22.1: (a) eel densities and biomass in different habitat compartments (Perrow et al. 1996); (b) estimates of the areas of open water compartments (from Environment Agency Water Level Management Plans and RSPB maps); and (c) estimates of overall lengths of reed bed margins (G. Gilbert, personal communication). Areas of different habitat types were thus:

• • • •

surface area of inhabited reed bed margins  approximate total length of margins (24 000 m)  maximum depth of penetration into reed beds (12 m, after Perrow et al. 1996)  28.8 ha; surface area of littoral zone habitat  approximate total length of margins (24 000 m)  width of littoral zone beyond the reed beds surveyed by electric fishing (assumed to be 3 m after Perrow et al. 1996)  7.2 ha; surface area of open water in meres  12.5 ha; surface area of dykes  6 ha.

Although eel populations may commonly be underestimated by electric fishing, eels comprised 60–90% of the fish biomass in late spring–early autumn (Perrow et al. 1996). Small rudd were much the more numerous species (and hence more important in the diet of bittern) in the littoral zones in March and November. The other main contributors to biomass (especially in the dykes and open waters in late summer–early autumn) were low numbers of relatively large pike. The estimated total population of approximately 36 000 eels compares with the annual recruitment estimate of between 15 000 and 45 000. If recruitment averages about 30 000 yr1 over an average freshwater residence time of five to ten years, overall natural mortality could be in the region of 75–90%. This compares with estimates from the literature of between 40 and 90% in a range of freshwater habitats, with a value of 70–75% being accepted as an approximate norm by the EIFAC/ICES Working Group on Eel (Moriarty & Dekker 1997). The presence of large eels, that must have entered the reserve before addition of the pass, suggests that the sluice is not

22 11/02/2003 15:10 Page 308

308

Mitigation measures

a total barrier to immigrants or that there are other, lesser routes for entry. However, the pass is obviously an important aid for immigrants. Eel stocks in Leighton Moss should be sustainable at current levels of recruitment of elvers and small juveniles. Population levels and structures appear healthy, providing relatively large amounts of food for bittern (and other piscivores). Further enhancements could only be achieved by stocking. This might benefit bittern but could be offset by increases in predation by larger eels, pike and other piscivores. Other possible negative impacts of overstocking are discussed in Section 22.5.3.

The potential benefits of eel passes at Minsmere Minsmere RSPB Reserve is a large coastal wetland system in Suffolk on the east coast of England. It is the major national site for breeding bittern. Eels are present but at very much lower densities and of larger mean size than in Leighton Moss (Noble et al., Chapter 11). These characteristics are predictable as Minsmere is located on the southwestern coast of the North Sea where recruitment currently is relatively low. It is possible that passes might aid immigration, but this is highly unlikely because of the nature of the potential entry point for eels and compartmentalisation of the wetlands. Minsmere is located behind a high shingle ridge facing directly out to sea. There is no tidal estuary or shallow bay to seawards and for flood protection purposes, the drainage from the wetlands and from other catchments falls into a common exit chamber, passes through a long pipe running under the shingle ridge and discharges offshore. Outflows are therefore rapidly diluted and dispersed, making it difficult for eels to detect the freshwater flows and to locate the pipe entrance. Even if they enter, there is a lack of tidal flow transport and eels would have to swim against the freshwater flows. Other barriers are imposed by separate sluices, used to prevent tidal incursions and to control water levels in the different catchments. Even after entry, dispersal of eels throughout the wetlands is difficult as older and drier regions of the reserve have had to be deepened to maintain reed stands. This has produced a complex system of bunded compartments with water levels that vary relative to one another throughout the year, controlled by many interconnecting adjustable pipe sluices. Despite these obstacles, elver traps have shown that some eels manage to enter the Reserve (G. Gilbert, personal communication). While it is always worth using passes to enhance eel stocks if at all possible, it appears that their effectiveness is limited at Minsmere. This example illustrates the need for site-specific surveys of the potential efficacy of passes. Where these are likely to be ineffective, stocking must be considered.

22.5.2

Controls on fishing mortality

Declines in glass eel recruitment throughout Europe have led to calls for the international application of the precautionary approach and controls on commercial fisheries. In England and Wales, the Environment Agency has followed the recommendations of Knights et al. (2001) and is planning to ban glass eel fishing in all but the traditional

22 11/02/2003 15:10 Page 309

Enhancing stocks of eel to benefit bittern

309

fishing areas, located in south and west coast estuaries that receive large numbers of immigrants. The remaining fisheries could potentially impact recruitment to wetland sites and such concerns have been raised at Leighton Moss. Commercial fishing has taken place for a number of years, catching glass eels as they accumulate about 1 mile downstream below the tidal flap sluice at Red Barn Dyke connecting to Morecombe Bay. Fishermen have negotiated exclusive fishing rights from the local landowner. The minimum commercial catch on the main February–April spring tides in 1999 was about 30 kg (87 550 eels) (D. Mower, unpublished data). This compares with an estimated recruitment to the Reserve of about 900 eels during this period, i.e. an apparently very low escapement of 5% during the main glass eel runs. Overall recruitment from trapping between February and October was estimated to be between 15 000 and 45 000 eels, i.e. relying mainly on later-arriving pigmented elvers, plus juveniles up to 15 cm long moving upstream from Morecombe Bay. The overall annual fishing mortality was therefore estimated at the time to be about 50–85%. The RSPB then purchased the fishing rights from the landowner and the fishermen cooperated in studies of recruitment. They voluntarily stopped fishing on a number of key high tides in the spring of 2000, but this did not result in any marked increases in eels in the sluice trap. This means that the initial assumptions that the apparent lack of glass eel immigrants in spring was due to fishing mortality is not tenable. As discussed earlier, later-moving pigmented elvers are the predominant recruits into fresh waters and not glass eels per se. Elvers and juveniles indeed comprised the bulk of the total annual recruitment. Similar conclusions have been reached from other studies on, for example, the Rivers Severn and Avon (White & Knights 1997b; Knights et al. 2001). From the calculations in Section 22.5.1, elver/juvenile recruitment via the sluice pass appears to be sufficient to sustain eel populations in Leighton Moss, independent of glass eel fishing mortality. Further studies are in progress to assess the relationships between fishing effort and total annual recruitment, for example, by marking glass eels in the early spring and assessing the proportion that survive to recruit later as elvers or juveniles. No other similarly specific conflicts are known between fishing for glass eels (or yellow and silver eels) and conservation of endangered species. However, the Leighton Moss studies will provide useful information if such conflicts should come to light at a future date. Another potentially important conflict with commercial fishing is that diving birds and otters are known to have been trapped and drowned in eel fyke nets (Jefferies et al. 1988). Waders and surface-swimming chicks and adult birds have also been trapped when nets have been exposed above the water surface. The only detailed data on bycatches in England and Wales come from extensive fyke netting (total of 1593 end days) in rivers, still waters and the estuary in the Thames catchment (Naismith & Knights 1994). This showed that bycatches and mortalities are not generally of great significance in regularly attended nets. Occasionally, however, large numbers of schooling juvenile fish were gill netted, especially in the leaders. Fyke nets and eel traps should be excluded from sensitive waters or at key times, for example when fledglings first leave the nest or during the migration periods of salmonids. Where they are allowed, they should be submerged at all times and should not be left unattended for more than one day or a tidal cycle, and by-catches released alive if possible.

22 11/02/2003 15:10 Page 310

310

22.5.3

Mitigation measures

Stocking of eels

Stocking can be a cost-effective means of restoring or maintaining yields in eel fisheries and enhancing spawner escape (Knights & White 1997). It could benefit stocks in isolated waters suitable for eels. It could also be important in wetlands and catchments with migration barriers where passes are ineffective (e.g. Minsmere). Glass eels can be purchased from dealers or trapped at the tidal limits of well-recruiting catchments and transferred to other waters. Site-specific surveys are needed to assess needs and cost–benefit analyses should be carried out. Stocking densities need to be adjusted to ensure that optimum survival, growth rates and sex ratios are achieved. Higher densities would augment the supply of small eels for species such as bittern while lower densities would help enhance the development of relatively more later maturing and larger females of high fecundity. Knights & White (1997) concluded that in warmer and more productive still waters, suitable intermediate densities would be about 0.1 kg ha1 (i.e. about 300 glass eels/elvers ha1 or the equivalent weight of juveniles). The potential yield is 20 kg ha1 at 40–50 g recruit1. In colder and less productive still waters, stocking rates should be reduced to 150–200 eels ha1. In both cases, stocking density can be reduced if larger yellow eels can be obtained, for example, using eels grown-on in culture or wild-caught. In rivers and dykes, Knights & White (1997) recommended that eels should be scatter-stocked (to minimise density-dependent mortality) in the summer, when temperatures are high enough to encourage dispersal. Typical target densities used in previous stocking programmes were 1–2 juvenile yellow eels m2 in low productivity waters, rising to 4–5 m2 in warmer waters with plenty of bottom cover and/or marginal vegetation and high productivity of macroinvertebrate prey. The example of Minsmere can be used to illustrate some of the estimates that need to be made when considering stocking, as well as financial and biological implications. The first assumption is that the densities of eels found in different habitats at Leighton Moss (Table 22.1) are optimal for bittern. The areas of different habitats inhabited by eels at Minsmere are (estimated as for Leighton Moss, Section 22.5.1), 42 ha of reed bed margin (at a margin length of 35 000 m), 12.2 ha of littoral zone, 5.4 ha of dykes/rivers and 17.6 ha of meres/ponds, i.e. an overall total of 59 ha. This leads to estimates of the desirable numbers of eels per compartment of 25 200, 18 300, 4050 and 5280 respectively: a total of 53 000 eels, compared with 36 000 eels at Leighton Moss. If the recruitment–stock relationship was similar to that at Leighton Moss, the optimal annual recruitment required would be about 22 000–65 000 glass eels/elvers. At 3000 individuals kg1, this is equivalent to between 7 and 22 kg, i.e. weights equivalent to those caught commercially at Leighton Moss. The purchase of such commercial catches could therefore meet the stocking requirements at Minsmere. The price of glass eels/elvers on the commercial market has been high in recent years. These were driven up during the late 1990s by a high demand for seed stock for eel farms, especially in China. Prices peaked at £200 kg1 in the late 1990s and although they have declined since, they vary throughout the season and were still about £50–80 kg1 on the European market in 2000. (Knights et al. 2001). Given that prices might be lower within Britain, the costs of stocking Minsmere could be in the region of £1000–1800, plus transport and stocking costs.

22 11/02/2003 15:10 Page 311

Enhancing stocks of eel to benefit bittern

311

Even if stocking is considered economically viable at sites like Minsmere where eels have naturally been at low density for some time, such options need careful consideration on biological grounds. While stocking might benefit bittern and other piscivores in the short term, ecological imbalances could result in the longer term. For example, increased density could affect the growth rates, sex ratios and the production of silver eels. Diseases and parasites might also be introduced. There could also be negative impacts on wider aquatic communities, for example, owing to increased competition or predation, especially if these involved rare species of other prey species that are also valuable food sources for bittern. Ideally, before any major stocking programmes are initiated in ecologically fragile wetland sites, small-scale (but long-term) pilot studies are needed in enclosed compartments.

22.5.4

Habitat management

Knights et al. (2001) concluded that the habitat requirements of eels are so generalised that specific habitat management programmes would be of little benefit to overall stocks and to spawner production and escape in Britain. The only specific limiting factors might be migration barriers, as discussed above, and the availability of substrates, vegetation, etc. where eels can hide during the day. The supply of eel (especially smaller ones) as prey for bittern in reed beds could, however, be enhanced by various management measures. From preceding discussions, it would be beneficial to maximise the extent of reed bed fringe (to 12 m) and adjacent littoral zones. Selective reed cutting and ditch and pond creation are also beneficial. The growth of aquatic macrophytes and successional encroachment by carr scrub may also need controlling, for example, by bed lowering and adjusting water levels. Such measures have been shown to have enhanced bittern numbers and breeding at Minsmere (Tyler et al. 1998; Smith et al. 2000).

22.6 Conclusions This chapter illustrates how eels can be made more available as prey for piscivores such as bittern. However, it appears that other fish species and prey can be important, particularly when eels are in short supply, for example, during the colder months. Minsmere supports the largest number of breeding bittern in Britain, despite the paucity of eels compared with sites such as Leighton Moss. If it is considered that augmenting eel populations might be beneficial at a particular site, careful site-specific studies are essential to check that this is indeed the case. Similarly, careful consideration needs to be given to the biological short- and long-term benefits (and disbenefits) of stocking, the provision of passes and other management approaches. Cost–benefit analyses are also needed, as are post-project studies to assess the efficacy of the measures taken. Costs could be a disincentive to non-government funded bodies like the RSPB. It is possible, however, that joint initiatives and shared funding with conservation bodies such as English Nature and other organisations such as the Environment Agency could be fruitful. Because of the important conservation and biodiversity

22 11/02/2003 15:10 Page 312

312

Mitigation measures

issues involved, additional funding could be sought from government and EU sources. All these possibilities should be pursued to achieve the best and most cost-effective means of protecting and enhancing eel stocks for the benefit of endangered piscivores.

References Baras E., Jeandrain D., Serouge B. & Philippart J.C. (1998) Seasonal variations in time and space utilisation by radio-tagged yellow eels Anguilla anguilla (L.) in a small stream. Hydrobiologia 371–372, 187–198. Jefferies D.J., Cripps S.J., Gorrod L.A., Harrison B., Johnstone J.S. & Potter E.C.E. (1988) The Effect of Otter Guards on the Efficiency of Eel Fyke Nets. London: Vincent Wildlife Trust, pp. 47. Knights B. (1982) Body dimensions of farmed eels (Anguilla anguilla L.) in relation to condition factor, grading, sex and feeding. Aquacultural Engineering 1, 297–310. Knights B. (1987) Agonistic behaviour and growth of eels in warm-water aquaculture. Journal of Fish Biology 31, 265–276. Knights B. (2002) A review of the possible impacts of long-term oceanic and climate changes and fishing mortality on recruitment of anguillid eels of the Northern Hemisphere. The Science of the Total Environment (in Press). Knights B. & White E.M. (1997) An appraisal of stocking strategies for the European eel, Anguilla anguilla L. In I.G. Cowx (ed.) Stocking and Introduction of Fish. Oxford: Fishing News Books, Blackwell Science, pp. 121–140. Knights B. & White E.M. (1998) Enhancing immigration and recruitment of eels: the use of passes and associated trapping systems. Fisheries Management & Ecology 4, 311–324. Knights B., Bark A., Ball M., Williams E., Winter E. & Dunn S. (2001) Eel and Elver Stocks in England and Wales – Status and Management Options. R&D Technical Report No. W248. Bristol: The Environment Agency, 294 pp. Knights B., White E. & Naismith I.A. (1996) Stock assessment of European eel, Anguilla anguilla L. In I.G.Cowx (ed.) Stock Assessment in Inland Fisheries Oxford: Fishing News Books, Blackwell Science, pp. 431–447. Moriarty C. & Dekker W. (eds) (1997) Management of the European eel. (Second Report of the EU Concerted Action AIR A94-1939). Fisheries Bulletin No. 15. Dublin: The Marine Institute, 52 pp. Naismith I.A. & Knights B. (1988) Migrations of elvers and juvenile European eels, Anguilla anguilla L., in the River Thames. Journal of Fish Biology 33 (Suppl. A), 161–175. Naismith I.A. & Knights B. (1990) Studies of sampling methods and of techniques for estimating populations of eels, Anguilla anguilla L. Aquaculture & Fisheries Management 21, 357–367. Naismith I.A. & Knights B. (1994) Fyke-netting and coarse fisheries in lowland Britain: Practical advice for fishery owners and managers. Fisheries Management and Ecology 1, 107–116. Naismith I.A. & Knights B. (1988) Migration of elvers and juvenile eel in the River Thames. Journal of Fish Biology 33, 161–175. Perrow M, Howe H. & Jowitt A. (1996) A Review of Factors Affecting the Status of Fish Populations in the Emergent Plant Zone of Wetland Habitats, Particularly Beds of Reeds (Phragmites australis), with Respect to the Habitat Requirements of Bittern (Botaurus stellaris). Unpublished Report by ECON, University of Essex, for the Royal Society for the Protection of Birds, England. Smith K.W., Welch G., Tyler G.A., Gilbert G., Hawkins I. & Hirons G. (2000) Recent management of the reed beds on the RSPB Minsmere Reserve and its impact on breeding bittern, Botaurus stellaris. British Wildlife (August 2000), 3–8. Tesch F.-W. (1977) The Eel: Biology and Management of Anguillid Eels. London: Chapman & Hall, 434 pp.

22 11/02/2003 15:10 Page 313

Enhancing stocks of eel to benefit bittern

313

Tyler G.A., Smith K.W. & Burges D.J. (1998) Reed bed management and breeding bitterns, Botaurus stellaris in the UK. Biological Conservation 86, 257–266. White E.M. & Knights B. (1997a) Environmental factors affecting the dynamics of upstream migration of the European eel, Anguilla anguilla L., in the Rivers Severn and Avon. Journal of Fish Biology 50, 1104–1116. White E.M. & Knights B. (1997b) Dynamics of upstream migration of the European eel, Anguilla anguilla L., with special reference to the effects of man-made barriers. Fisheries Management & Ecology 4, 311–324. Wirth T. & Bernatchez L. (2001) Genetic evidence against panmixia in the European eel. Nature 409, 1037–1040.

23 11/02/2003 15:12 Page 314

Chapter 23

Experimental manipulation of great blue heron and belted kingfisher predation rates on stream fish J. STEINMETZ* University of Illinois and Illinois Natural History Survey, Center for Aquatic Ecology, Champaign, Illinois, USA

S.L. KOHLER Environmental Research Institute, Western Michigan University, Kalamazoo, Michigan, USA

D.A. SOLUK Illinois Natural History Survey, Champaign, Illinois, USA

Abstract The objective of this study was to determine if bird predation on a natural stream fish assemblage could be manipulated on a relatively large spatial scale. The study was conducted in two streams at Midewin National Tallgrass Prairie, Illinois, USA. The experimental design consisted of one experimental factor, predation intensity, at three levels: reduced, ambient and elevated. The treatments successfully manipulated avian predation pressure. All three treatments were significantly different in the total number of kills by herons and kingfishers combined. Assuming the controls represent natural predation rates, a total of 0.34 prey per 100 m2 d1 were consumed. The technique reported could be used elsewhere in conjunction with avian diet studies to assess better both the impact of avian predation on fisheries and the effectiveness of various mitigation measures. Keywords: bird, fish, manipulation, predation, predation rate, stream.

23.1 Introduction There has been recurring interest in bird–fish interactions over the last century (White 1939; Elson 1962; Alexander 1979; Wood 1987; Suter 1995; Kirby et al. 1996), usually spurred-on by concerns about game fish losses to avian predators. Despite the economic importance of the question, whether or not birds impact fish populations remains unclear. Most of the research has been done on game fish, with some authors reporting significant impacts, (Elson 1962; Wood 1987; Kennedy & Greer 1988), while others found no effect (Salyer & Lagler 1946; Glahn et al. 1998). Sometimes different conclusions were reached from the same data. Staub et al. (1992, cited in Suter 1995, 1998) concluded that wintering great cormorants, Phalacrocorax *Correspondence: J. Steinmetz, University of Illinois, IL 61801, USA and Illinois Natural History Survey, Center for Aquatic Ecology, 607 E. Peadbody Dr., Champaign, IL 61820, USA (email: [email protected]).

23 11/02/2003 15:12 Page 315

Manipulated avian predation rates

315

carbo (L.) were a significant source of mortality for grayling, Thymallus thymallus (L.) in Swiss rivers, whereas Suter (1995, 1998), using the same data, concluded that birds were not affecting grayling population dynamics. Most of these studies link fish population estimates with gut contents and abundance of avian predators. Some authors (e.g. Kirby et al. 1996), however, noted that assessing consumption rates based on gut contents or pellet analysis has limitations. Differences in digestibility of prey and error in measuring gut contents, predator abundance, or prey abundance can lead to over- or underestimation of the importance of predation. Ideally, diet studies could be conducted in conjunction with experiments. The problem is that predator manipulation on a scale large enough to be biologically meaningful is difficult. The objective of the study detailed in this chapter was to determine if bird predation on a natural stream fish assemblage could be manipulated on a relatively large spatial scale.

23.2 Materials and methods The study was conducted from mid-July to mid-September 2000 at Midewin National Tallgrass Prairie, Illinois, USA. A number of fish-eating birds occur at this site, including great blue herons, Ardea herodias L., green herons, Butorides virescens L., blackcrowned night herons, Nycticorax nycticorax L., great egrets, Ardea alba L., belted kingfishers, Ceryle alcyon L., and hooded mergansers, Lophodytes cucullatus L. Two streams were used for the study: Prairie Creek and Jackson Creek. The most common avian predators on these sites are great blue herons and belted kingfishers. Prairie Creek is a low gradient third order stream fed principally by surface run-off, shallow groundwater and field tile discharge. Substrate is mostly cobble and gravel; areas with predominately fine sediments are found at several locations. Thirty-five fish species have been recorded in Prairie Creek; common species are smallmouth bass, Micropterus dolomieu L., striped shiner, Luxilis chrysocephalus R., bluntnose minnow, Pimephales notatus R., central stoneroller, Campostoma anomalum R., and green sunfish, Lepomis cyanellus R. Dominant riparian vegetation comprises reed canary grass, Phalaris arundinacea L., sedges, willows, Salix spp., cottonwood, Populus deltoides M., Osage orange, Maclura pomifera S., ash, Fraxinus spp., and elm (Ulmus spp.). Jackson Creek is also a low gradient third-order stream fed primarily by surface run-off, shallow groundwater and field tile discharge. The streambed comprises mostly cobble and boulder; bedrock is exposed in a few areas. Twenty-two species of fish are present in Jackson Creek; common species include smallmouth bass, striped shiner, bluntnose minnow, central stoneroller, spotfin shiner, Cyprinella spiloptera C., and green sunfish. Main riparian vegetation consists of reed canary grass, sedges, willows, cottonwood, Osage orange, black walnut, Juglans nigra L., ash, and elm. The experimental design was a randomised block with one experimental factor – predation intensity – at three levels: reduced, ambient and elevated. Reduced treatments were designed to prevent predation by all piscivorous birds, ambient treatments served as controls to estimate natural predation rates, and elevated treatments were designed to

23 11/02/2003 15:12 Page 316

316

Mitigation measures

increase kingfisher predation. Each of these levels was replicated across four different sites (two sites per stream), chosen for their similarity in habitat characteristics, trophic structure and accessibility to avian predators. Within each site, reaches were selected based on the similarity of both within-stream and riparian characteristics. Average depths ranged from 29 to 41 cm, with a minimum of 12 cm in riffles to a maximum of 102 cm in pools. Each treatment area was approximately 12 m wide by 60 m long (720 m2). Reaches with reduced predation were covered with 3.81 cm plastic bird netting to exclude avian predators. Netting was suspended in a tent-like manner over the stream for two reasons: (1) to allow access to manipulated sections of stream for sampling; and (2) to decrease tendency for kingfishers to fly over the stream and elicit behavioural responses in prey. Control (ambient) reaches were simply unmanipulated sections of stream. Reaches with elevated predation had two types of perches added: wire stretched across the stream and wooden dowel rods attached perpendicularly to fence posts. Four wires and wooden dowel rods were added per reach. Treatments within each reach were separated by at least 120 m. Blocks on Jackson and Prairie Creek were separated by 2.5 and 11.5 km respectively, with a number of shallow riffles between each block, making fish movement between blocks unlikely. All sites were dominated by bluntnose minnows, striped shiners, central stonerollers, hornyhead chubs, and sand shiners. There were no significant differences in total fish numbers, or in fish assemblages among the sites at the start of the experiment (Steinmetz, in press). Details of the fish assemblage and its response to the manipulation are presented elsewhere (Steinmetz, in press). Observations of each treatment on each stream were conducted throughout the course of the experiment. These observations provided information on the time spent foraging, the number of prey taken and the predation pressure on prey populations. Behavioural information was collected using a combination of observers and timelapse video. Live observations were made in three observational periods: morning (07:00–12:00), afternoon (12:00–17:00), and evening (17:00–20:00). Each site was observed at least once in each time block, on at least three different days, so that each of the twelve sections was observed for one full day. Video was conducted by placing a camera connected to a time-lapse VCR in the field for one to two days. Tapes were collected at the end of each observation period, and the cameras rotated, so that each of the twelve sections was video taped for at least one full day. For each observation the following were recorded: (1) bird species observed; (2) total time spent at the experimental unit; (3) activities at the site (fly-through, foraging, etc.); (4) number of attacks; (5) number of kills; and (6) prey type consumed, if it could be determined. If the exclusions worked as intended, kingfisher activity should have been greatest in perch sections, intermediate in control sections and zero in the bird exclusion areas. Heron predation should have been equal in the perch and control sections, and zero in the bird exclusion areas. Attack rates were recorded by dividing the total number of attacks by the total time each section was observed. This number was extrapolated to the number of attacks in 14 hours (daylight during the study period was from approximately 06:00 to 20:00) to obtain a daily attack rate. Daily rates were used for all analyses. The data were analysed by ANOVA, blocking by site. Post hoc multiple comparisons were conducted with Fisher’s LSD. All statistics were calculated using SYSTAT 7.0.

23 11/02/2003 15:12 Page 317

Manipulated avian predation rates

317

23.3 Results The total time that each section was observed ranged from 8 to 35 h, with an average time of 19 h. Several sections were not filmed due to camera problems. On two occasions condensation inside the camera housings made tapes unusable, and on one occasion a battery ran out of power during taping. Despite these problems, there were no significant differences between treatments in total time watched (Kruskal–Wallis nonparametric test statistic  2.208, d.f.  2, P  0.332). The treatments successfully manipulated avian predation pressure (Table 23.1; Figs 23.1 and 23.2). The total number of attacks by herons and kingfishers combined was significantly greater in both the control and perch treatments compared with the bird exclusion treatments (Fisher’s LSD, P  0.05 in both cases). There was also a strong trend for a greater attack rate in perch compared with control treatments, but this was not significant (Fisher’s LSD, P  0.07). All three treatments were significantly different in the total number of kills by herons and kingfishers combined (Fisher’s LSD, P  0.05 in all cases). Breaking the observations down by species, herons foraged equally in control and perch sections, but only rarely in bird exclusion areas (Figs 23.1 and 23.2). Occasionally, herons ducked under the netting either when it came loose along the sides, or when the stream level dropped markedly. Kingfisher attacks and kills were non-existent in the exclusion areas, intermediate in the controls, and greatest in the perch areas (Figs 23.1 and 23.2). Herons would forage by slowly moving through the study reach. They spent an average of 22 min in each study reach per visit, with visits ranging in duration from 14 to 40 min. Kingfishers were much more variable in the time spent in the study reaches. The average time spent in each reach per visit was 46 min, but ranged from 14 to 180 min. Assuming the controls represent natural predation rates, herons consumed 0.14 prey per 100 m2 d1, kingfishers consumed 0.20 prey per 100 m2 d1, and a total of 0.34 prey per 100 m2 d1 were consumed (Fig. 23.2). Extrapolating, herons consumed 13.6 prey per ha d1, kingfishers consumed 20.8 prey per ha d1, and a total of 34.4 prey per

Table 23.1

ANOVAs

Source

for combined heron and kingfisher attack and kill rates

Sum-of-squares

d.f.

Mean-square

F-ratio

P

Total attack rate Stream Treatment Error

8.547 94.165 24.312

3 2 6

2.849 47.082 4.052

0.703 11.619

0.584 0.009

Total kill rate Stream Treatment Error

1.458 39.692 7.754

3 2 6

0.486 19.846 1.292

0.376 15.357

0.774 0.004

23 11/02/2003 15:12 Page 318

318

Mitigation measures

1.4

Attacks/Day/100 m2

1.2

Heron Attack Rate Kingfisher Attack Rate Total Attack Rate

1.0 0.8 0.6 0.4 0.2 0.0 Netted

Control

Perch

Treatment

Figure 23.1 Number of attacks per 100 m2 d1 by herons, kingfishers, and herons and kingfishers combined. Number of attacks is based on live and video-taped observations. Error bars represent 1 SE

1.4

Kills/Day/100 m2

1.2

Heron Kill Rate Kingfisher Kill Rate Total Kill Rate

1.0 0.8 0.6 0.4 0.2 0.0 Netted

Control

Perch

Treatment

Figure 23.2 Number of kills per 100 m2 d1 by herons, kingfishers, and herons and kingfishers combined. Number of kills is based on live and video-taped observations. Error bars represent 1 SE

ha d1 were consumed. Owing to their feeding behaviours, it was usually impossible to identify heron prey. Herons usually captured prey underwater, and when successful would raise their heads, extend their necks and swallow the prey. Thus successful strikes could readily be differentiated from misses but prey identity could usually not be determined. Kingfishers catch prey, return to a perch, then beat the prey against the

23 11/02/2003 15:12 Page 319

Manipulated avian predation rates

319

perch to make it easier to manoeuvre it for swallowing. Thus, when close enough, it was fairly easy to at least differentiate crayfish from fish. Eight of the 44 kills (18%) where kingfisher prey could be identified were crayfish, while 36 of the 44 kills (82%) were fish.

23.4 Discussion The experiment was successful in manipulating avian predation rates: the numbers of attacks and kills by kingfishers and great blue herons was lowest in the netted areas, intermediate in the controls, and greatest in the sections with perches. The natural kingfisher and heron predation rate in the control areas was 0.34 prey per 100 m2 d1. Some of the kingfisher prey, however, were crayfish. Other studies have found fewer crayfish in adult kingfisher diets (Hamas 1994). However, the high densities of crayfish in these streams (up to 8 per m2; J. Steinmetz, unpublished data), probably led to a higher encounter rate, and hence a higher percentage of the diet than in other streams. If crayfish are removed from the total, 0.30 fish were consumed per 100 m2 d1 during the late summer. While there are numerous diet studies of both belted kingfishers and great blue herons, no other studies that we are aware of report predation rates per unit area per unit time. A few other studies used focal animal observations and report predation rates per unit time. Predation rates of belted kingfishers and great blue herons at aquaculture facilities in the north-eastern United States were estimated at 1.7 and 2.2 trout h1 respectively (Glahn et al. 1999). Other estimates for great blue herons are 0.14 catfish h1 at catfish ponds at the National Wildlife Research Center in Mississippi, USA (Glahn et al. 2000) and 0.8 catfish h1 at catfish farms in Mississippi, USA (Stickley et al. 1995). By dividing the total number of kills observed by the total time watched for each species, numbers were comparable to these other studies obtained. Lower predation rates were found in this study than in the previously reported studies: 0.10 fish h1 for kingfishers and 0.07 fish h1 for great blue herons. The higher predation rates reported in other studies are probably because they were conducted at hatcheries which have higher fish densities than those typically found in natural systems, thus making foraging easier for avian predators. The predation rates reported for Jackson and Prairie Creek are undoubtedly lower in the winter, as the herons migrate south (Butler 1992). Kingfishers remain as long as there is open water (Hamas 1994). One other study found that predation rates of avian predators at aquaculture facilities to be minimal in winter (Glahn et al. 1999). It is unclear whether the numbers would be higher or lower in the spring and early summer. Birds are breeding at this time and young will be fledging (Butler 1992; Hamas 1994), thus energetic demands should be higher. However, the streams generally have higher flows and are more turbid during this time, which should create more refuge habitat (in terms of both depth and turbidity), and decrease bird foraging efficiency. Future experiments should examine seasonal changes in predation rates to establish better the amount of prey consumed on a yearly basis. The experiment showed that it is possible to manipulate avian predation at a relatively large spatial scale. This technique may be useful in examining avian impacts on

23 11/02/2003 15:12 Page 320

320

Mitigation measures

fish populations in a number of field situations, especially on small- to medium-sized streams, ponds and hatchery facilities. Stretching netting across large rivers or lakes, however, would be more difficult, if not impossible. The netting was effective in excluding both wading birds and kingfishers. Netting would probably not be effective against diving predators such as cormorants, mergansers and loons along stream reaches. It would be impossible to extend the netting to the bottom of the stream to exclude these predators as debris would quickly clog the netting and eventually tear it down. Netting could, however, effectively exclude diving predators along ponds and hatcheries, where flow is not a problem and netting could be attached directly to the ground or fences. This technique could be used along appropriate waterways in conjunction with diet studies of regional avian predators. Diet studies could be conducted first to see if there is the potential for avian impacts. These studies could then be followed with a manipulative field experiment to confirm the results from the diet studies. Refuge habitats within the stream (e.g. woody debris) could also be manipulated to examine the effectiveness of adding cover in reducing avian impacts on important game species (see Russell et al., Chapter 19; McKay et al., Chapter 20). This would provide an alternative to the current control method of shooting birds (Kirby et al. 1996).

Acknowledgements We would like to thank Heather Vance for helpful comments on an earlier version of this manuscript, and Jason Cashmore, Robb Diehl, Eric Gittenger and Barry Williams for help with field work. This work was funded by a Program in Ecology and Evolutionary Biology Summer Research Fellowship, a Sigma Xi Grant-in-Aid of research, and the Philip W. Smith Memorial Fund.

References Alexander G.R. (1979) Predators of fish in coldwater streams. In H.R. Stroud & H. Clepper (eds) Predator–Prey Systems in Fisheries Management. Washington, DC: Sport Fisheries Institute, pp. 153–170. Butler R.W. (1992) Great Blue Heron. In A. Poole & F. Gill (eds) The Birds of North America, No. 84. Philadelphia: The Academy of Natural Sciences (Washington, DC: The American Ornithologists Union, 20 pp.). Elson P.F. (1962) Predator-prey relationships between fish-eating birds and Atlantic salmon. Bulletin of the Fisheries Research Board of Canada 133, 87 pp. Glahn J.F., Harrel J.B. & Vyles C. (1998) The diet of wintering double-crested cormorants feeding at lakes in the Southeastern United States. Colonial Waterbirds 21, 431–437. Glahn J.F., Dorr B. & Tobin M.E. (2000) Captive great blue heron predation on farmed channel catfish fingerlings. North American Journal of Aquaculture 62, 149–156. Glahn J.F., Rasmussen E.S., Tomsa T. & Preusser K.J. (1999) Distributions and relative impact of avian predators at aquaculture facilities in the Northeastern United States. North American Journal of Aquaculture 61, 340–348.

23 11/02/2003 15:12 Page 321

Manipulated avian predation rates

321

Hamas M.J. (1994) Belted Kingfisher (Ceryle alcyon) In A. Poole & F. Fill (eds) The Birds of North America, No. 25. Philadelphia: The Academy of Natural Sciences (Washington, DC: The American Ornithologists Union, 16 pp.). Kennedy G.J.A. & Greer J.E. (1988) Predation by cormorants, Phalacrocorax carbo (L.), on the salmonid poulations of an Irish river. Aquaculture and Fisheries Management 19, 159–170. Kirby J.S., Holmes J.S. & Sellers R.M. (1996) Cormorants Phalacrocorax carbo as fish predators: an appraisal of their conservation and management in Great Britain. Biological Conservation 75, 191–199. Power M.E., Dudley T.L. & Cooper S.D. (1989) Grazing catfish, fishing birds, and attached algae in a Panamanian stream. Environmental Biology of Fishes 26, 285–294. Salyer J.C. & Lagler K.F. (1946) The Eastern Belted Kingfisher, Megaceryle alcyon alcyon (Linnaeus) in relation to fish management. Transactions of the American Fisheries Society 76, 97–117. Staub E., Krämer A., Müller R., Ruhlé Ch. & Walker J. (1992) Einfluss des Kororans (Phalacrocorax carbo) auf Fischbestrände und Fangerträge in der Schweiz. Schriftenreihe Fis´cherei 50, 1–138 (in German). Staub E., Egloff K., Krämer A. & Walter J. (1998) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: two case studies from Swiss rivers. Comment. Journal of Applied Ecology 35, 607–610. Stickley A.R. Jr, Glahn J.F., King J.O. & King D.T. (1995) Impact of Great Blue Heron depredations on channel catfish farms. Journal of the World Aquaculture Society 26, 194–199. Steinmetz J., Kohler S.L. & Soluk D.A. (in press) Birds are overlooked top predators in aquatic foodwebs. Ecology. Suter W. (1995) The effect of predation by wintering cormorants Phalacroxorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: two case studies from Swiss rivers. Journal of Applied Ecology 32, 29–46. Suter W. (1998) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trouth (Salmonidae) populations: two case studies from Swiss rivers. Reply. Journal of Applied Ecology 35, 611–616. White H.C. (1939) Bird control to increase the Margaree River Salmon. Bulletin Fisheries Research Board of Canada 58, 30 pp. Wood C.C. (1987) Predation of juvenille pacific salmon by the common merganser (Mergus merganser) on eastern Vancouver Island. II: Predation of stream-resident juvenile salmon by merganser broods. Canadian Journal of Fisheries and Aquatic Science 44, 950–959.

This page intentionally left blank

24 11/02/2003 15:14 Page 323

Section IV Management

This page intentionally left blank

24 11/02/2003 15:14 Page 325

Chapter 24

Managing a balance between double-crested cormorant numbers and warmwater fish abundance in the eastern basin of Lake Ontario, New York: preliminary insights from a management programme J.F. FARQUHAR, R.D. McCULLOUGH* and A. SCHIAVONE New York State Department of Environmental Conservation, Division of Fish Wildlife and Marine Resources, Watertown, New York, USA

Abstract Double-crested cormorants, Phalacrocorax auritus Lesson, began nesting on the United States side of Lake Ontario in 1974 and have increased from 22 pairs to over 5000 pairs. Studies conducted by the New York State Department of Environmental Conservation (NYSDEC) and the United States Geological Survey, established a link between cormorants and the reduced abundance of smallmouth bass, Micropterus dolomieu Lacépède, an important sport fish and keystone predator. New York State Department of Environmental Conservation decided to reduce cormorant abundance to improve the public benefits from fishing while maintaining the ecological integrity of the lake ecosystem. In 1999, a cormorant populationreduction programme began in the eastern basin of Lake Ontario. On Little Galloo Island, the largest cormorant colony in the basin, reproductive suppression via egg oiling resulted in a 98% reduction in reproductive success. Smallmouth bass depredation fell more than 28% relative to a no-management condition. Cormorant management efforts on Lake Ontario suggest that cost-effective techniques can reduce problems on a local level. Keywords: double-crested cormorants, egg oiling, nest removal, recreational fishery, smallmouth bass.

24.1 Introduction The double-crested cormorant, Phalacrocorax auritus Lesson, is the most abundant of six cormorant species nesting in North America (Hatch & Weseloh 1999). Doublecrested cormorants nest across the continent in numbers estimated between one and two million birds (Hatch 1995). Historical records suggest a widespread distribution, but the first documented nesting on the Great Lakes did not occur until 1913 on Lake Superior. Cormorant nesting colonies subsequently spread through the Great Lakes, *Correspondence: Russ McCullough, New York State Department of Environmental Conservation, Division of Fish Wildlife and Marine Resources, 317 Washington Street Watertown, NY 13601, USA (email: rdmccull@ gw.dec.state.ny.us).

24 11/02/2003 15:14 Page 326

326

Management

reaching Lake Ontario by 1938 (Weseloh et al. 1995). Cormorants disappeared from Lake Ontario in the 1960s, probably due to chemical contamination (Weseloh et al. 1995). With improving environmental conditions, cormorants established a colony of 22 nests on Little Galloo Island, Lake Ontario in 1974. From 1974, cormorant numbers increased rapidly on Little Galloo Island, reaching 8410 nests in 1996. Concerns related to the impact of cormorants on eastern basin sportfish began to emerge during the late 1980s, when local anglers observed large flocks of cormorants feeding on recently stocked salmonids and within productive fishing areas. Due to angler and management concerns, cormorant diet studies were initiated in 1992. By 1997 sufficient information was acquired to suggest the need for more intensive study. These studies, initiated in 1998, focused on determining the relationship between cormorant predation and decline in smallmouth bass, Micropterus dolomieu Lacépède, abundance, with consideration of other influences. Through this series of studies, a probable link between increasing cormorant numbers and declining bass populations was established within the eastern basin of Lake Ontario. More specifically, decline in the relative survival of 3–5 year-old smallmouth bass in the eastern basin was correlated with cormorant populations exceeding levels associated with 1200 nesting pairs. Based on these results and a strong public desire for a viable fishery, the New York State Department of Environmental Conservation (NYSDEC) established a management programme balancing cormorant numbers with an improved fishery. Through a series of public outreach activities, the NYSDEC solicited input on a set of management objectives for the eastern basin ecosystem. The overriding goal adopted for the programme was to improve the benefits that people derive from Lake Ontario’s eastern basin ecosystem from both fishery and colonial waterbird resources. This chapter describes the process of developing, implementing and evaluating a cost-effective programme to manage the impacts of a fish-eating bird, the doublecrested cormorant, on the complex aquatic ecosystem of eastern Lake Ontario.

24.2 Study area Lake Ontario is the lowermost of the Great Lakes, with a surface area of over 19 400 km2 and a maximum depth of over 245 m. The eastern basin (Fig. 24.1), often called the Kingston Basin in Canada, is a relatively small (2100 km2), shallow (less than 65 m) area north and east of the Main Duck Sill, which runs from Stony Point, New York to Prince Edward Point, Ontario (43°54N). Approximately half of the area is under US jurisdiction. The area contains a complex series of rocky points, islands, shoals and bays. Cormorant management activity in the eastern basin of Lake Ontario relative to fish community objectives has focused on Little Galloo Island. Three other islands (Bass, Calf and Gull) are managed to prevent colony establishment by cormorants in relation to private property and waterbird objectives. Little Galloo and Gull Islands are owned by NYSDEC. Bass and Calf Islands are privately owned. Little Galloo Island is a 17.4 ha, tilted limestone shelf located within the eastern basin 14.5 km west of Henderson Harbor, New York. The island is a major colonial waterbird resource, with 53 000 pairs of ring-billed gulls, Larus delawarensis Ord, 200 pairs of

24 11/02/2003 15:14 Page 327

Managing a balance between cormorant numbers and warmwater fish

327

Ontario

Eastern Basin of Lake Ontario New York

Figure 24.1 Location of eastern Lake Ontario study area

herring gulls, Larus argentatus Pontoppidan, 11 pairs of great black-backed gulls, Larus marinus L., and 1200 pairs of Caspian terns, Sterna caspia Pallas, in addition to over 5000 pairs of double-crested cormorants in 1999–2000. Bass (2 ha) and Gull (1 ha) Islands lie at the mouth of Henderson Bay. Gull Island held 45 herring gull nests and 20 black-crowned night heron, Nycticorax nycticorax L., nests, while 2300 ring-billed gull pairs, 10 herring gull pairs and 36 night-heron pairs nested on Bass Island in 2000. Night herons nest sporadically at Calf Island (13 ha), 2 km east of Little Galloo Island. Smallmouth bass remains the most abundant and widespread eastern basin sportfish despite significant declines in recent years (Chrisman & Eckert 1999). It is still the most sought-after sportfish, attracting over 35 000 directed angler trips in 1998 (McCullough & Einhouse 1999). Yellow perch, Perca flavescens Mitchill, is the most commonly harvested pan fish, although harvests are currently well below historical levels. Alewife, Alosa pseudoharengus (Wilson), is present in the Basin from late spring to autumn and remains an important food source for most piscivores (Lantry & Shaner 1998).

24.3 Methods 24.3.1

Decision

Information from public meetings, a citizen task group and opinion surveys was used to determine the desired benefits from the eastern basin resource and acceptable management techniques that could be used to achieve them. Formal public contact began in early 1998 with a open meeting where information regarding Lake Ontario fisheries and cormorants was presented and an intensive research programme announced. To assist NYSDEC in development of the proposed course of action, the final research findings, experimental goals, objectives and management alternatives were presented and public input was requested at three environmental or sporting events (totalling over 10 000 participants) and through press releases distributed throughout New York State. Information was also provided and input requested through the NYSDEC website.

24 11/02/2003 15:14 Page 328

328

Management

Smallmouth bass population trends and survival/mortality rates were examined using data collected through the NYSDEC Warmwater Fish Stock Assessment Program. Annual assessments of the warmwater fish community in the New York waters of Lake Ontario’s eastern basin have been conducted each August from 1976 to 2000. Standard sinking gillnet gangs were set overnight at randomly selected sites. Each net gang consisted of nine 15 m panels, 2.4 m deep ranging in size from 51 to 152 mm stretch mesh. Net sites were stratified geographically and by depth. Examination of smallmouth bass population trends was based on catch-per-unit-effort (CPUE). The relative mortality of young smallmouth bass (not fully recruited to the gear) was calculated as CPUE age-3 at time t divided by CPUE age-6 at time t  3 (Lantry et al. 1999). Cormorant diet composition and fish consumption were based on pellet analysis and a food consumption/feeding day model. Diagnostic prey remains recovered in regurgitated pellets were used to describe the diet of double-crested cormorants on Little Galloo Island. Approximately 150 pellets were collected at 2-week intervals beginning in late April and ending in late September. In the laboratory, diagnostic bones, all otoliths, and representative scales were removed from the pellets and identified under magnification. Eye lenses were also enumerated since, although they could not be used in species identification, their total number (i.e. number of lenses/2) generated fish counts that exceeded those based on bones or otoliths in some pellets (Johnson et al. 2001). To estimate the number of fish consumed by cormorants from the Little Galloo colony, a model similar to that of Weseloh and Casselman (unpublished report: Fish consumption by double-crested cormorants on Lake Ontario, Burlington, Ontario) was used. This model incorporated cormorant age-class population size and seasonal residence time (time spent feeding in area), mean daily fish ingestion rates, and a faecal pathway correction factor for fish not detected in pellets (Johnson et al. 2001). Angling quality was determined through the Lake Ontario–Eastern Basin Creel Survey. The creel survey consisted of two independent samples. Angler interviews were conducted by direct contact at representative access sites to provide data such as catch and harvest rates, and angler characteristics. Effort data were collected by instantaneous aerial counts of fishing boats (McCullough & Einhouse 1999). This method is very similar to that regularly used on the open waters of Lake Erie (Culligan et al. 1997). The survey was conducted from May to October 1998. Evaluation of other factors that may have contributed to the decline in smallmouth bass abundance involved examination of lake-wide fishery trends (based on an annual creel survey), environmental conditions, invading species (dreissenid mussels) (Eckert 1999), and alternative predators such as walleye (Schneider et al. 1999).

24.3.2

Operations

Nest removal activities on Gull and Bass Islands began in 1994 to protect private property interests and reduce competition with other waterbirds. All ground nests were removed by hand while tree nests were removed with a telescoping pole. Each nest removed was scattered as much as possible to discourage rebuilding. In 1997, Calf Island was included in nest removal activities following an attempt by cormorants to establish a colony. Nest

24 11/02/2003 15:14 Page 329

Managing a balance between cormorant numbers and warmwater fish

329

removal activities on these islands since 1999 have also supported the eastern basin fish community objective of 1500 pairs of nesting cormorants. Visits to Gull and Bass Islands were generally made weekly beginning in mid-May and continued until mid-July. Egg oiling to reduce reproductive success began on Little Galloo Island in May and June 1999 and was repeated during the same season in 2000. NYSDEC staff treated all cormorant nests accessible from the ground with pure food grade vegetable oil. The oiling process was conducted five times in 1999 and four times in 2000, at 2-week intervals. Oil was applied from a backpack sprayer unit in sufficient volume to cover the exposed surface of each egg, approximately 6 mL egg1. Oil use for all visits totalled 295 litres in 1999 and 208 litres in 2000. Each nest or group of nests treated was marked with spray paint to ensure the treatment of all nests accessible from the ground. Two or three teams of two to three persons each completed the spraying in three hours or less (not including travel time). Each team could effectively oil 500 to 700 nests h1, depending on nest density. Application of oil at two-week intervals ensured that each nest would be treated at least twice during the incubation period.

24.3.3

Evaluation

Effectiveness of reproductive suppression was examined through on-site nest and chick counts, which were conducted during each site visit, to determine the short-term effectiveness of oiling on reducing chick (fledgling) production and to examine longerterm change in the breeding population at this site. In addition to applying oils the teams recorded the number of nests treated, the number of eggs per nest, the number of chicks observed and the number of nests not treated (generally tree nests). Disturbance to other bird species was evaluated during each site visit. Localised VHF telemetry tracking projects were implemented in 2000 to assess within-season nest site fidelity or displacement. Satellite telemetry was used to examine within- and between-season fidelity. Smallmouth bass population trends and survival/mortality rates were examined using data collected through the NYSDEC Warmwater Fish Stock Assessment Program. Cormorant diet composition and fish consumption were based on pellet analysis and a food consumption/feeding day model. Smallmouth bass angling quality (catch and harvest rates) will be determined through a detailed follow-up creel survey of the eastern basin fishery which is scheduled for 2003 to examine changes in fishing quality. Public consultation procedures are planned for 2003 to determine if stakeholders have considered management to be successful.

24.4 Results 24.4.1

Decision

Public consultation provided NYSDEC with 896 responses (directly received) providing input on management alternatives. Most respondents (90.1%) preferred the most

24 11/02/2003 15:14 Page 330

330

Management

aggressive management, 4.6% preferred less aggressive management, 2.5% preferred no lethal control of adult cormorants, and 2.8% preferred current management or no active management. An additional 2558 management alternative responses were received through the Lake Ontario Fisheries Coalition and State Senator Wright’s office: of these 97.3% preferred the most aggressive possible management. The preferred colonial waterbird/fishery management alternative, which involved egg oiling and phased-in culling of adults (culling was not implemented), was presented to the public through press releases and a public meeting attended by 57 people, in March 1999. Thirty-nine participants provided written comments. All felt that they understood the proposal, 53% supported the proposed action and 18% disagreed, calling for more aggressive cormorant reduction. Another 18% did not specifically agree or disagree with the proposal but suggested more aggressive cormorant management. There were two responses (5%) neither agreeing nor disagreeing, but calling for less aggressive management, and two (5%) stating that management should be guided by scientific information. Smallmouth bass population, angling quality and cormorant diet analyses led to the conclusion that there was reasonable evidence that cormorant predation on smallmouth bass in the New York waters of the Lake Ontario eastern basin was excessive from a fishery viewpoint. Angling quality for smallmouth bass had deteriorated in the eastern basin, with catch rates declining from 43% to 51% compared with pre-cormorant levels (McCullough & Einhouse 1999). Catch rates were unchanged or improved in other areas of Lake Ontario that were outside the feeding range of most cormorants (Eckert 1999). This suggested that broad ecosystem changes (e.g. weather, reduced phosphorus, lower zooplankton densities, expanding dreissenid mussels) were not the primary factors responsible for the eastern basin smallmouth bass decline, since these factors operated throughout the lake (Eckert 1999). The discovery that nearly one million age-3 to age-5 bass were consumed by cormorants annually, just prior to entering the fishery, suggested that mortality of these fish could have increased substantially (Johnson et al. 1999). The mean relative mortality of young smallmouth bass increased from 0.4 to 2.3 – a significant (at the 95% level) increase in mortality – which was significantly (99% level) associated with increasing cormorant abundance (Lantry et al. 1999). Furthermore this increased mortality threatened to deplete the 1995 year class, the largest year class to be detected since 1988.

24.4.2

Operations

On Little Galloo Island, cormorant reproduction was controlled to protect the declining smallmouth bass fishery and maintain nesting opportunities for other colonial waterbird species. Gull, Calf and Bass Islands were managed to prevent the establishment of new cormorant colonies and to protect black-crowned night heron colonies. The maximum number of eggs oiled per trip to Little Galloo Island declined from 16 310 in 1999 to 10 917 in 2000. Peak nest count was 5627 in 1999 and 5119 in 2000 (Table 24.1). Nesting attempts at Bass and Gull Islands (including re-nests), and therefore nest removals, varied from year to year with no nesting attempts in 1995 and with a peak of

24 11/02/2003 15:14 Page 331

Managing a balance between cormorant numbers and warmwater fish Table 24.1

331

Cormorant management (egg oiling) activity on Little Galloo Island

Date

Nests treated

Eggs treated

2 856 5 627 4 283 3 889 2 844 2 613 4 301 3 907 2 929

7 119 16 310 9 718 7 568 5 015 6 953 10 917 8 742 5 509

6 May 1999 20 May 1999 3 June 1999 17 June 1999 8 July 1999 16 May 2000 30 May 2000 14 June 2000 28 June 2000

Treatment duration (h) 2.0 4.0 2.3 2.5 1.5 2.0 2.5 2.6 2.0

1500

Nests

1000

500

0 1994

1995

1996

1997

1998

1999

2000

Figure 24.2 Cormorant nests removed from Bass and Gull Islands, Lake Ontario

345 nests in 1997. Nesting attempts and removals on Bass and Gull Islands increased markedly to 1368 (including re-nests) in 2000 (Fig. 24.2). Only sporadic attempts at nesting were observed at Calf Island.

24.4.3

Evaluation

Reproductive suppression resulted in a hatching success (number of chicks hatched per egg) for oiled nests of less than 1% annually since 1999. This meets the objective set in the NYSDEC five-year management plan to reduce the number of successful cormorant nests on Little Galloo Island by 90%. It was estimated that fewer than 300 cormorant chicks have hatched on Little Galloo Island annually since 1999, mostly in untreated tree nests, i.e. about 5900 fewer cormorant chicks were produced as a result of oiling. Since the nest removal programme began in 1994, there has been no successful cormorant reproduction on Gull, Bass and Calf Islands. The preliminary results of the satellite and radio telemetry studies of cormorant movements indicated that cormorants on Little Galloo Island and at two Canadian colonies show nest site fidelity during the breeding season. Disturbance observations indicate that flushing levels ranged from 20% for ring-billed gulls to 100% for cormorants (Fig. 24.3). Disturbance

24 11/02/2003 15:14 Page 332

332

Management

Ring-billed Gull Caspian Tern Herring Gull Great Black-backed Gull Black-crowned Night Heron Double-crested Cormorant 0

20

40

60

80

100

Percentage flushed per visit

Figure 24.3 Bird disturbance during reproductive suppression activities at Bass, Gull and Little Galloo Island, Lake Ontario

to non-target species is an undesirable effect, but to date, overall colony viability does not appear to be reduced for any species. Diet and fish consumption information revealed that decreased chick production reduced total fish consumption by the Little Galloo Island colony by an estimated 4.8 million fish, a 24% reduction in biomass consumed and a 19% reduction in total fish consumed (Johnson et al. 2000). Because of the seasonal variation in the diet of cormorants from the Little Galloo colony, the oiling of eggs may provide the greatest protection for those fish species that are proportionally more abundant in the diet during the chick feeding and post-chick feeding periods. In this regard, alewife and smallmouth bass may benefit the most since their contribution to the diet is substantially greater during those periods. Consumption of smallmouth bass was reduced by an estimated 400 000 fish as a result of egg oiling. Smallmouth bass abundance in recent years has been at a record low level, with abundance in the past 5 years being the lowest recorded during the 25-year period of record. Abundance of yellow perch, an important recreational species and major cormorant forage, was also at record low levels during the past 5 years (Eckert 2001).

24.5 Discussion Governmental fish and wildlife management agencies in the United States are typically charged with the stewardship of fish and wildlife resources held in public trust. Responsibilities to protect and conserve resources are often balanced with mandates to promote beneficial use of these resources and to consider social and economic factors in their management. Frequently, the issues at hand involve complex biological systems and equally complex social, economic and political considerations. Scientific research and biological knowledge are integral to the decision-making process, but managers rarely have all of the biological or socio-economic information desirable to make optimal decisions. Decisions, including ‘no action’ decisions, are still necessary,

24 11/02/2003 15:14 Page 333

Managing a balance between cormorant numbers and warmwater fish

333

and adaptive management, including ongoing evaluation and adjustment over the long term, is often required in lieu of definitive research results. Management directed at reducing populations of non-game species, such as cormorants, is inherently controversial in modern society. Although initiation of cormorant management was well received locally, several national environmental and animal rights organisations expressed concern about the current programme and may initiate legal action to prevent the lethal control of adult cormorants. Within the scientific community, opinions vary about the adequacy of existing fisheries impact data, appropriate thresholds for initiating management and the methodology being used. Nonetheless, the public has generally accepted, if not uniformly endorsed, the current management programme. The experimental egg oiling programme was effective in reducing fish consumption in eastern Lake Ontario; however, it is too early to determine what effect this will have on the fish community or individual species. Commitment to the programme is necessarily long term because the fishery depends on the presence of strong year classes, which occur at unpredictable intervals. Additionally, models suggest that, because of low adult mortality rates, reduction of cormorant numbers to desired levels via reproductive suppression alone will take time. Monitoring and evaluation will continue, including new insights from telemetry research, to evaluate management effectiveness, results and consequences, and to validate or refute previous conclusions. Thus far, cormorant management in eastern Lake Ontario has been operationally effective and publicly tolerated. Few negative biological or social consequences have been observed. While the ultimate measure of success – an improved fishery and viable cormorant population – may not be realised for several years, cormorant management within the Lake Ontario eastern basin offers a solid example of an integrated fish and wildlife management programme in progress.

References Chrisman J.R. & Eckert T.H. (1999) Population trends among smallmouth bass in the eastern basin. In Final Report: To assess the impact of double-crested cormorant predation on the smallmouth bass and other fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, February 1, 141 pp. Culligan W.J., Cornelius F.C., Einhouse D.W., Zeller D.L., Zimar R.C., Beckwith B.J. & Wilkinson M.A. (1997) 1997 Annual Report to the Lake Erie Committee. Albany: NYS Department of Environmental Conservation, 68 pp. Eckert T.H. (1999) Trends in Lake Ontario smallmouth bass sport fishery, 1985–98. In Final Report: To Assess the Impact of Double-Crested Cormorant Predation on the Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, Albany, New York, 141 pp. Eckert T.H. (2001) Summary of 1976–2000 Warmwater Assessment. Annual Report to the Great Lakes Fishery Commission, Lake Ontario Committee. Albany: NYS Department of Environmental Conservation, 261 pp. Hatch J.J. (1995) Changing populations of double-crested cormorants. Colonial Waterbirds 18 (Special Publication 1), 8–24. Hatch J.J. & Weseloh D.V. (1999) Double-crested cormorant. In A. Poole & F. Gill (eds) The Birds of North America, No. 441: 1–36. Ithaca: Cornell Laboratory of Ornithology and Philadelphia: Academy of Natural Sciences.

24 11/02/2003 15:14 Page 334

334

Management

Johnson, J.H., Ross R.M. & Adams C.M. (1999) Diet composition and fish consumption of doublecrested cormorants in eastern Lake Ontario. In Final Report: To Assess the Impact of DoubleCrested Cormorant Predation on the Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, Albany, New York, 141 pp. Johnson J.H., Ross R.M., McCullough R.D. & Edmonds B. (2001) Diet Composition and Fish Consumption of Double-crested Cormorants from the Little Galloo Island Colony of Eastern Lake Ontario in 2000. Double-Crested Cormorant predation on Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. Summary of 2000 studies. NYSDEC Special Report, Albany, New York, 173 pp. Lantry B.F., Eckert T.H. & Schneider C.P. (1999) The relationship between the abundance of smallmouth bass and Double-crested Cormorants in the eastern basin of Lake Ontario. In Final Report: To Assess the Impact of Double-Crested Cormorant Predation on the Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, Albany, New York, 141 pp. Lantry B.F. & Shaner T. (1998) The Status of the Pelagic Prey Fish in Lake Ontario 1997. Annual Report to the Great Lakes Fishery Commission, Lake Ontario Committee. Albany: NYS Department of Environmental Conservation, 173 pp. McCullough R.D. & Einhouse D.W. (1999) Lake Ontario-eastern basin creel survey. In Final Report: To Assess the Impact of Double-Crested Cormorant Predation on the Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, Albany, New York, 141 pp. Schneider C.P., Eckert T.H. & McCullough R.D. (1999) Predation on smallmouth bass by walleye in the eastern basin of Lake Ontario. In Final Report: To Assess the Impact of Double-Crested Cormorant Predation on the Smallmouth Bass and Other Fishes of the Eastern Basin of Lake Ontario. NYSDEC Special Report, Albany, New York, 141 pp. Weseloh D.V., Ewins P.J., Struger J., Mineau P., Bishop C.A., Postupalsky S. & Ludwig J.P. (1995) Double-crested cormorants of the Great Lakes: Changes in population size, breeding distribution and reproductive output between 1913 and 1991. Colonial Waterbirds 18 (Special Publication 1), 48–59.

25 11/02/2003 15:15 Page 335

Chapter 25

Management of the cormorant, Phalacrocorax carbo, and endangered whitefish, Coregonus lavaretus, populations of Haweswater, UK I.J. WINFIELD* CEH Windermere, The Ferry House, Far Sawrey, Ambleside, Cumbria, UK

D.H. CRAWSHAW United Utilities, Liverpool Road, Great Sankey, Warrington, UK

N.C. DURIE Environment Agency, Gillan Way, Penrith Business Park, Penrith, Cumbria, UK

Abstract Within the UK, the whitefish, Coregonus lavaretus (L.), is protected under nature conservation legislation. The status of one population in Haweswater, a reservoir in the English Lake District, has declined since the 1960s due to increased variations in lake level during the spawning period. A substantial reduction in leakage from the water distribution system has reduced overall demand for water, and this development, aided by recent rainfall patterns, has resulted in a series of years with good spawning conditions. Nevertheless, recruitment has continued to be poor and studies have indicated that foraging by a local cormorant, Phalacrocorax carbo L., colony, which has grown rapidly from its establishment in 1992, is likely to be responsible. In this paper management measures undertaken on the cormorant colony in 1999 and 2000, together with those proposed for the future, are described. The context of this conflict of fish and bird conservation issues is also explored with respect to future potential developments within the wider English Lake District. Keywords: breeding colony, conservation, Coregonus lavaretus, Haweswater, impact, Phalacrocorax carbo.

25.1 Introduction The whitefish, Coregonus lavaretus (L.), occurs in just seven UK lakes and is accordingly considered to be of national conservation importance with protection under Schedule 5 of the Wildlife and Countryside Act, 1981. All of the English populations of this coregonid occur in the north-west of the country within the English Lake District. The present study is concerned with the population of Haweswater and the potential impact upon it by a recently established breeding colony of cormorants, Phalacrocorax carbo L. Although increased use of freshwater habitats by this piscivorous bird has been *Correspondence: I.J. Winfield, CEH Windermere, The Ferry House, Far Sawrey, Ambleside, Cumbria LA22 0LP, UK (email: [email protected]).

25 11/02/2003 15:15 Page 336

336

Management

widespread in the UK in recent years, with attendant concerns over impacts on many fish populations (see Russell et al. 1996), the establishment of inland breeding colonies has been limited to just seven sites (Newson 2000). During the 1990s, the whitefish population of Haweswater was extensively studied (Beaumont et al. 1995; Winfield et al. 1996; Winfield, Fletcher & Cubby 1998; Winfield et al. 2002) and monitored (Winfield et al. 2001) to provide a sound scientific basis for its conservation management. Corresponding, but less extensive, investigations (Winfield, Winfield & Fletcher 1998; Newson 2000; Winfield et al. 2002) have also been undertaken on the cormorant population of this lake in the late 1990s. The Haweswater lake fish and cormorant system is thus one of the best-studied examples of its kind in the UK, and so has the potential to make a significant contribution to the development of a national policy towards resolving this controversial issue. Note that at Haweswater, the conflict is simply between fish and bird conservation as there are no significant commercial or recreational fisheries interests on the lake. Here, the ecology, status and management of the whitefish population are described, followed by an account of the development, ecology and impact of the cormorant population and its management in 1999 and 2000. Finally, this conflict of fish and bird conservation issues is considered in a wider context, including future potential developments within the English Lake District.

25.2 Study site Situated in the north-west of England (National Grid Reference NY 480 140), Haweswater lies at an altitude of 240 m, with a surface area of 391 ha and a maximum depth of 57 m. The present surface area and maximum depth were increased by 252 ha and 29 m, respectively, over those of the original lake by the construction of a dam in 1939, which facilitates the lake’s current operation as a potable water reservoir by United Utilities and which incidentally created the lake’s only island, Wood Howe (National Grid Reference NY 477 119). In addition to whitefish, the fish community of this oligotrophic lake comprises Arctic charr, Salvelinus alpinus (L.), brown trout, Salmo trutta L., eel, Anguilla anguilla (L.), perch, Perca fluviatilis L., minnow, Phoxinus phoxinus (L.), and three-spined stickleback, Gasterosteus aculeatus L. (Winfield et al. 1996). The only fishery activity is very limited recreational angling for trout.

25.3 Ecology, status and management of the whitefish population In the early 1990s, a gill-netting survey of all whitefish populations in England and Wales revealed the Haweswater population to be relatively slow growing and have restricted length and age frequency distributions, indicating extremely inconsistent recruitment in the late 1980s and early 1990s (Winfield et al. 1996). Subsequent comparisons of whitefish entrapped in the water abstraction system between 1992 and 1994 with specimens of the same origin between 1965 and 1967 reported by Bagenal (1970) showed that there had been a decline in age class equitability and individual growth (Winfield et al. 1998a). In the 1970s, Broughton (1972) found entrapped whitefish

25 11/02/2003 15:15 Page 337

Management of cormorants at Haweswater

337

from a 1972 sample to include a wide range of individual lengths, including young individuals, but a depressed growth pattern similar to that recorded in the 1990s. Specimens entrapped in 1983 and examined but not aged by Maitland (1985) included only large individuals. Furthermore, the numbers of whitefish entrapped from 1973 to 2000 (Fig. 25.1(a)) showed a marked decrease in the early 1980s between the sampling periods of Broughton (1972) and Maitland (1985). Since 1992, the whitefish population of Haweswater has been monitored by quantitative echo sounding. Following isolated surveys in May 1992 and May 1996, the fish community has been monitored in May, July and September each year between 1997 and 2000 by echo sounding using a Simrad EY200P echo sounder with a 200 kHz singlebeam transducer and data analysis using HADAS (for full details see Winfield et al. 1998a). While the numbers of recorded small (40–100 mm in length, Fig. 25.1(b)) and medium (100–250 mm in length, Fig. 25.1(c)) sized fish are probably a combination of whitefish, Arctic charr, brown trout and perch, the numbers of large (greater than 250 mm in length, Fig. 25.1(d)) fish are probably mainly adult whitefish. Variation within the abundance of such large fish from 1992 to 2000 was significant (ANOVA, F5,54  6.882, P  0.001), with post-hoc comparisons by the Tukey HSD test revealing that the 1992 abundance was significantly higher (P  0.05) than in all subsequent years. The abundances of small- and medium-sized fish will be returned to below. All of the above population data are consistent with an interpretation of the whitefish population having a good status in the mid-1960s, with a decrease in growth rate by the early 1970s, followed by a decrease in the consistency of recruitment by the 1980s. These reductions in growth and recruitment have been shown to be attributable to increases in water level variability of this reservoir between the 1960s and 1990s (Winfield et al. 1998a). Whitefish recruitment is primarily influenced by water level variations during the first 90 days of the year, during which spawning occurs largely within days 20 to 40, followed by incubation to beyond day 90. Water levels during days 1 to 90 of the year are shown in Figure 25.2 for three example years from the 1960s and 1980s, and for 1998, 1999 and 2000, which are the most recent years with available data. Between 1965 and 1967 (Fig. 25.2(a)) water levels were generally very high and stable, while 20 years later, from 1985 to 1987 (Fig. 25.2(b)), falls in water level in excess of 4 m were not unusual. During the mid to late 1990s, significant reductions in leakage from the water distribution network by the owners of Haweswater (United Utilities, formerly North West Water Limited), leading to a reduction of overall demand from the reservoir, contributed to a water level regime more sympathetic to the requirements of whitefish spawning. Thus, between 1998 and 2000 (Fig. 25.2(c)) variations in water level during the early part of the year were markedly reduced. In addition to the more sympathetic water level management described above, management measures instigated in recent years for the benefit of the whitefish population included the development of a mobile artificial spawning substrata system and an attempt to establish refuge populations in two smaller lakes within the Haweswater catchment. Both of these actions are described in more detail by Winfield et al. (2002). Despite the above management actions and conducive spawning conditions in the mid to late 1990s, the decline of adult whitefish which began in the early 1980s (Fig. 25.1(a)) has not been reversed or even stopped (Fig. 25.1(d)). By contrast, the abun-

25 11/02/2003 15:15 Page 338

338

Management

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

600 500 400 300 200 100 0

1973

Number of individuals

(a)

Year (b)

1500000 1000000

1997

1998

1999

2000

1997

1998

1999

2000

1997

1998

1999

2000

1996 1996

0

1996

500000 1992

Population size

2000000

Year 14000 12000 10000 8000 6000 4000 2000 0 1992

Population size

(c)

Year 7000 6000 5000 4000 3000 2000 1000 0 1992

Population size

(d)

Year

Figure 25.1 (a) Numbers of whitefish entrapped from 1973 to 2000, and estimated population sizes (geometric means with 95% confidence limits) of (b) small (40 to 100 mm in length), (c) medium (100 to 250 mm in length), and (d) large (greater than 250 mm in length) fish in May 1992 and 1996 to 2000. Small and medium fish are probably a combination of whitefish, Arctic charr, brown trout and perch, while large fish are probably mainly adult whitefish

25 11/02/2003 15:15 Page 339

Management of cormorants at Haweswater

339

(a)

Lake level (m)

0 1965 1966 1967

⫺2 ⫺4 ⫺6 0

10

20

30

40

50

60

70

80

90

Day of year (b)

Lake level (m)

0 1985 1986 1987

⫺2 ⫺4 ⫺6 0

10

20

30

40 50 Day of year

60

70

80

90

(c)

Lake level (m)

0

1998 1999 2000

⫺2

⫺4

⫺6 0

10

20

30

40

50

60

70

80

90

Day of year

Figure 25.2 (a) Lake levels during the first 90 days of the year from (a) 1965 to 1967, (b) 1985 to 1987, and (c) 1998 to 2000. Whitefish spawning occurs largely within days 20 to 40, followed by incubation to beyond day 90

dance of small fish in Haweswater (Fig. 25.1(b)), which may include young whitefish among several other species, has shown some positive signs in the 1990s with significant variation (ANOVA, F5,54 3.358, P  0.01) and post-hoc comparisons by the Tukey HSD test revealing that the 1999 abundance was significantly higher than equivalent figures from 1996, 1997 and 1998. Variations in the abundance of medium fish, how-

25 11/02/2003 15:15 Page 340

340

Management

ever, have shown no significant variation (ANOVA, F5,54  2.264, 0.05  P  0.10). Thus, although there have been indications of successful whitefish recruitment from the egg to juvenile stages in the 1990s, particularly by the 1995 year class (Winfield et al. 2002), few individuals have survived to adulthood and the status of the whitefish population remains extremely poor. A potential cause for this whitefish recruitment failure in the late 1990s was the foraging activities of local cormorants.

25.4 Development, ecology and impact of the cormorant population Although great increases in the numbers of overwintering cormorants have been seen in recent years on many lakes and reservoirs in the UK, such a trend has not occurred on Haweswater (Fig. 25.3(a)). By contrast, between the early 1970s and the late 1990s, overwintering numbers have tended to decrease at this site. It should be noted that before the whitefish decline of the early 1980s, the numbers of overwintering cormorants ranged up to 25 individuals, but since the mid 1980s they have not exceeded 13 individuals. However, the number of cormorants frequenting Haweswater during the spring and summer months has increased greatly in recent years due to the development of a breeding colony on the island of Wood Howe. Following the first record of breeding by one pair in 1992, the colony rapidly increased, such that by 1998 there were 48 breeding pairs including both P. c. carbo and P. c. sinensis (Fig. 25.3(b)). A comparative study of cormorant breeding performance at a range of inland and coastal colonies in 1997 and 1998 by Newson (2000) showed that the Haweswater colony had the highest mean clutch size in both years (ranked first among eight sites), with mean brood size also relatively high in 1997 (ranked second among 10 sites) but lower in 1998 (ranked equal sixth among 10 sites). Given the above increase in the numbers of cormorants at Haweswater and their potential for impacting on the whitefish population, a project was undertaken to determine the numbers and local feeding behaviour of cormorants from November 1996 to December 1997 (Winfield et al. 1998b; Winfield et al. 2002). During the breeding season (Fig. 25.3c), cormorant numbers increased markedly in April and peaked in July at 84 individuals, including young. Numbers subsequently decreased in the late summer as young and adults dispersed from the lake. Using data on the fish standing stock from the May 1997 echo-sounding survey, the above cormorant counts, and an assumed daily consumption rate of 0.6 kg cormorant d1, the total annual consumption by cormorants was expressed as a percentage of the fish community standing stock and ranged from 943 to 3867% for minimum and maximum daily cormorant counts, respectively (Winfield et al. 2002). Clearly, such a local predation rate cannot be sustained by the Haweswater ecosystem, but the solution to this apparent paradox is that although cormorants do feed at Haweswater, where whitefish have been observed to be eaten and identified in regurgitated pellets (Winfield et al. 1998b), they also make extensive movements off the lake to feed elsewhere (Winfield et al. 2002). Nevertheless, even the limited amount of feeding observed at Haweswater itself in 1997 may account for 44% of the adult whitefish standing stock of that year. Adopting the precautionary principle (see below), management of the cormorant population is warranted.

25 11/02/2003 15:15 Page 341

341

30 25 20 15 10 5 1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

0 1970

(a)

Count (number of individuals)

Management of cormorants at Haweswater

Year Colony size (number of pairs)

(b)

50 40 30 20 10 0 1992

1993

1994

1995

1996

1997

1998

Year Count (number of individuals)

(c)

90 80 70 60 50

1997 2000

40 30 20 10 0 Mar

Apr

May

Jun

Jul

Aug

Month

Figure 25.3 (a) Numbers of overwintering cormorants present in January from 1970 to 1997 (data supplied by Kathleen Atkinson, The Wildfowl & Wetlands Trust and the first author), and (b) development of the breeding colony from the first record of breeding (1992) to the last year before scaring (1998) (data supplied by the Royal Society for the Protection of Birds and The Wildfowl & Wetlands Trust), and (c) the numbers of cormorants present at the breeding colony from March to August in 1997 (when no scaring was undertaken, data from Winfield et al. (1998b)) and in 2000 (during scaring, data supplied by Tony Benson)

25 11/02/2003 15:15 Page 342

342

Management

25.5 Management of the cormorant population Following a meeting in 1998, attended by stakeholders including English Nature, the Environment Agency, the Lake District National Park Authority and the Royal Society for the Protection of Birds, steps were taken by United Utilities to discourage cormorants from nesting in 1999. This was attempted by initially deploying a variety of scaring activities and devices (human visits to the island, ballons, reflective plates, a stuffed eagle owl, Bubo bubo L., and a loud radio) on the nesting island from April to July, although the nearby presence of England’s only nesting pair of golden eagles, Aquila chrysaetos L., precluded the use of gas cannons. All scaring activities and devices were approved by English Nature. The only effective scaring technique was found to be frequent disturbance by human visits to the island, which were made during 24 persondays from 19 April to 17 August 1999. In this way, nesting was successfully prevented, although considerable numbers of adult cormorants still used the lake for roosting, loafing and some feeding. A second meeting of stakeholders in late 1999 recommended the repetition of these scaring measures in the 2000 breeding season, during which island visits totalling 19 person-days were made between 3 March and 23 August 2000, together with limited roost management to cut down some trees used for roosting and nesting. Nesting was again successfully prevented, although considerable numbers of adult cormorants continued to frequent the island from April to early July: the numbers counted on the island in 2000 exceeded those seen in 1997 when no scaring was undertaken (Fig. 25.3(c)). Foraging pressure on the whitefish population is thus unlikely to have been reduced markedly by scaring alone. Stakeholders agreed a continuation of the above scaring measures for the cormorant breeding season of 2001, although local access restrictions due to foot-and-mouth disease subsequently prevented any cormorant management during the spring or summer. Irrespective of this practical problem, given that the extensive scaring measures employed in 1999 and 2000 only prevented nesting and did not reduce the numbers of adult cormorants at the lake, a cull of adult birds is recommended as the only means of further reducing the impact of foraging on the whitefish population. This contentious issue is returned to below.

25.6 Discussion The management of any vertebrate predator has a great potential to become a controversial issue, and in this context cormorants foraging on inland waters have proved to be no exception. Moreover, the science of impacts on freshwater fish populations by this species, although being far from fully understood, is highly variable as evident from recent UK case studies (Feltham et al. 1999). This site-specificity makes it extremely dangerous to transfer scientific conclusions from one lake or river to another, meaning in effect that interactions at the majority of inland sites will never be understood on a robust scientific basis. Such scientific uncertainty besets the management of a range of other environmental problems and prompted the agreement of Principle 15 in the 1992 Rio Declaration On

25 11/02/2003 15:15 Page 343

Management of cormorants at Haweswater

343

Environment and Development which states ‘In order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.’ This approach is particularly appropriate for the situation at Haweswater where scientific evidence strongly suggests, but does not prove beyond any doubt, that a population of a native rare fish is threatened with extinction by the foraging activities of an alien species which has established a breeding population at the site only within the past 10 years, and even then only on an artificial island. In this situation, the precautionary principle should be adopted to the benefit of the whitefish over the cormorant. Taking management action against the cormorant colony at Haweswater can only be robustly defended if other adverse environmental factors that impact on the whitefish population have themselves been addressed to the fullest feasible extent. This is the case at Haweswater, where water level fluctuations are now more sympathetic to whitefish spawning requirements and a mobile artificial spawning substrata system has been developed for possible use in extreme conditions. There are no eutrophication, acidification or species introductions problems at this lake, leaving the cormorant issue as the only area where further practicable management action can be taken. Nevertheless, the scaring activities and roost management conducted in 1999 and 2000 have only prevented breeding and have not significantly reduced the numbers of adult birds frequenting the lake in the spring and summer months. Among the range of options suggested for the control of cormorants in a guidance leaflet issued by the Ministry of Agriculture, Fisheries and Food (Ministry of Agriculture, Fisheries and Food 2000), considerations or trials at Haweswater showed that noise generating scarers, stocking control, ‘buffer’ species and fish refuges are inappropriate, visual scarers are ineffective, and roost management and human disturbance had only limited success. This leaves culling of the adult birds by shooting as the only realistic management option that may reduce the predation impact on the whitefish to levels that may allow the population to recover. The cormorant is a relatively long-lived bird which, given the high natal fidelity found by Newson (2000), means that adults originating from the Haweswater colony during the 1990s are likely to keep returning for some considerable time to come. This further reinforces the conclusion that stopping current and future breeding is not in itself sufficient to safeguard the conservation of the whitefish. Moreover, on a national perspective, cormorant population modelling by Newson (2000) indicated that there will be considerable inland population development of this species in the near future. For the English Lake District, where adult cormorants in breeding plumage are already frequently seen at a number of the major lakes (IJW, personal observation), this is likely to mean the establishment of additional colonies. An examination of maps indicates that a further nine lakes within the area offer potential island nest sites. Even the establishment of cormorant colonies on only some of these additional nine lakes would lead to increased predation pressure on other whitefish populations, together with those of Arctic charr and the UK’s rarest freshwater fish, the vendace, Coregonus albula (L.). Returning to the specific case of Haweswater, unless appropriate management measures are taken very soon there is a serious risk of the local extinction of the endangered whitefish.

25 11/02/2003 15:15 Page 344

344

Management

Acknowledgements We thank Craig Denny, John Sanders and colleagues at United Utilities for their provision of entrapment specimens and lake level data. We are also indebted to Tony Benson for giving us access to cormorant counts made during the scaring in 2000, while Bill Kenmir of the Royal Society for the Protection of Birds, Baz Hughes of The Wildfowl & Wetlands Trust and Kathleen Atkinson kindly made available unpublished data and background information on cormorants. This work was funded by United Utilities.

References Bagenal T.B. (1970) Notes on the biology of the schelly Coregonus lavaretus (L.) in Haweswater and Ullswater. Journal of Fish Biology 2, 137–154. Beaumont A.R., Bray J., Murphy J.M & Winfield I.J. (1995) Genetics of whitefish and vendace in England and Wales. Journal of Fish Biology 46, 880–890. Broughton N.M. (1972) Taxonomy of some British coregonids. Unpublished dissertation, Queen’s University, Belfast. 51 pp. Feltham M.J., Davies J.M., Wilson B.R., Holden T., Cowx I.G., Harvey J.P. & Britton J.R. (1999). Case Studies of the Impact of Fish-Eating Birds on Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0106. London: MAAF, 212 pp. Ministry of Agriculture, Fisheries and Food (2000). Advisory Leaflet WM14 – Fisheries and the presence of cormorants, goosanders and herons. Maitland P.S. (1985) Monitoring Arctic charr using a water supply screening system. Information Series International Society of Arctic Char Fanatics No. 3, pp. 83–88. Newson S.E. (2000) Colonisation and range expansion of inland breeding great cormorants in England. Unpublished PhD thesis, University of Bristol, 61 pp. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.D. (1996) Assessment of the Problem of Fish-Eating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food project VC 0104. London: MAAF, 130 pp. Winfield I.J., Cragg-Hine D., Fletcher J.M. & Cubby P.R. (1996) The conservation ecology of Coregonus albula and C. lavaretus in England and Wales, U.K. In D. Hefti & A. Kirchhofer (eds) Conservation of Endangered Freshwater Fish in Europe. Basel: Birkhauser Verlag, pp. 213–223. Winfield I.J., Fletcher J.M. & Cubby P.R. (1998a) The impact on the whitefish (Coregonus lavaretus (L.)) of reservoir operations at Haweswater, U.K. Advances in Limnology 50, 185–195. Winfield I.J., Winfield D.K. & Fletcher J.M. (1998b) The Feeding Behaviour of Cormorants at Haweswater. Final Report. Unpublished Report by the Institute of Freshwater Ecology to North West Water Limited. WI/T11068J7/2, 46 pp. Winfield I.J., Fletcher J.M. & James J.B. (2001) Monitoring of the Schelly of Haweswater, April 2000 to March 2001. Final Report. Unpublished Report by CEH Windermere to North West Water Ltd. WI/C01512/2, 30 pp. Winfield I.J., Fletcher J.M. & Winfield D.K. (2002) Conservation of the endangered whitefish (Coregonus lavaretus) population of Haweswater, UK. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, 232–241 pp.

26 11/02/2003 16:17 Page 345

Chapter 26

Turnover in a wintering cormorant population: implications for management G.A. WRIGHT* Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, UK

Abstract Earlier work on Loch Leven showed that shooting was ineffective in reducing the wintering cormorant population. This chapter reviews data from multiple counts, radio tracking, satellite telemetry and counts from other sites, to investigate cormorant foraging strategy and turnover within the loch’s population. Multiple cormorant counts revealed large-scale arrivals and departures, and radio tracking revealed intermittent absences, with individuals present for 51% of the time. Satellite telemetry indicated that birds mostly ranged within 45 km of Loch Leven, with occasional journeys further afield. The loch’s wintering cormorant population comprised 10% of the population within 45 km, and there was evidence of movement between sites during the winter. High turnover within the population reduces its amenability to control, and would account for the ineffectiveness of shooting as a mitigation measure. Keywords: fishery management, mitigation, predator control, shooting.

26.1 Introduction The cormorant, Phalacrocorax carbo carbo (L.), is a top predator and the increase in numbers wintering in Britain is well documented (e.g. Kirby et al. 1995), and has resulted in concern over the possible impact of higher cormorant numbers on commercial fishery interests (e.g. Carss & Marquiss 1997; van Eerden & Zijlstra 1997). Attempts to control cormorant numbers by shooting on open water sites to reduce their potential impact have met with little success (e.g. Marquiss & Carss 1994; Mellin & Mirowska-Ibron 1997; McKay et al. 1999). If shooting is to be effective, it requires the population to be sufficiently discrete to be amenable to such a control measure. Cramp & Simmons (1977) considered cormorants to be individually nomadic outside the breeding season, but Sellers & Sutcliffe (1987) believed that they showed a fair degree of winter site fidelity both within and between seasons. Yésou (1995) estimated that the number of individuals using his study site was 3.9 to 6.2 times higher than the highest mid-month count. He found a small number of long-staying birds to be markedly site-faithful, with very

*Correspondence: Gordon Wright, Corran Bhan, Portincaple, By Garelochhead, Argyll G84 0ET, UK (email: [email protected]).

26 11/02/2003 15:17 Page 346

346

Management

little evidence of intermittent attendance. Buchheim (1997) found a large number of short-staying birds, but strong site fidelity among the long-stayers. From an analysis of ringing recoveries, Coulson & Brazendale (1968) concluded that dispersal of cormorants from breeding to wintering grounds was colony-specific. They showed that birds wintering in south-eastern Scotland were drawn principally from south-west and north-west Scotland, the Orkney Islands and the Farne Islands. This was supported by ringing recoveries at Loch Leven, which included birds ringed in southwest and western Scotland, northern Scotland and Orkney, and the Forth and Farne Islands. Coulson & Brazendale (1968) also showed that the logarithm of the number of birds wintering within a particular distance of the colony was linearly related to that distance. This suggests the possibility that if shooting depletes the number in a particular area, other birds may move in from shorter or longer ranges to maintain that distribution pattern. Loch Leven (NGR: NO 140020) is located in east-central Scotland, between the Forth and Tay Estuaries. It is an important trout fishery as well as a wetland of international importance for breeding and wintering waterfowl. It is a National Nature Reserve, Ramsar Site and Special Protection Area, and long-term fishery and bird records facilitate the study of fishery/cormorant interactions. Until 1995, cormorants were shot in large numbers for fishery protection purposes, despite the lack of evidence of impacts on the fishery, or of any beneficial effect of shooting (Allison et al. 1974; Carss & Marquiss 1992). This chapter reviews evidence of turnover within the Loch Leven cormorant population, and considers its amenability to control measures and the implications for shooting as a mitigation measure.

26.2 Materials and methods Multiple cormorant counts were conducted by four principal observers, three times each day for 106 days in October and November and February to April during the winters of 1996/1997 and 1997/1998. In addition, two counts per day were completed for 55 days during December and January. Loch Leven is not a difficult place to count cormorants, with sufficient elevated observation points to enable the whole site to be overlooked, and cormorant roost and loafing sites are well known. Each count was conducted by one person and took up to 2 h. There was therefore some scope for error due to unobserved bird movements during a count, but these counts are assumed to have a generally high degree of accuracy. Individual observer performance was analysed by comparing mean counts of pairs of observers conducted during overlapping periods. Nine cormorants were captured and fitted with TW3 short-range radio transmitters supplied by Biotrack. Three transmitters were fitted on each of 17 March, 7 October and 21 December 1997. They were tracked manually with a Mariner M57 receiver and Yagi antenna, and all birds were checked for presence or absence at Loch Leven two or more times each day. Nine cormorants were fitted with PTT 100 satellite-tracked, radio transmitters, supplied by Microwave Telemetry, and were tracked by the Argos satellite system. Six transmitters were fitted on 7 October 1997, two on 21 December 1997 and one on 4 February 1997. They were set to transmit for 8 h per day, stepping forward 1 h each day so that over 24

26 11/02/2003 15:17 Page 347

Cormorant turnover and management implications

347

days the full 24-hour period would be equally covered. They transmitted a series of identification signals at intervals, and it required the receipt by the satellite of four of these signals to enable an accurate fix. Once the area over which Loch Leven birds ranged was established, cormorant counts obtained from the Wildfowl and Wetlands Trust were collated for 179 marine, estuary, stillwater and river sites within that area. Twenty-nine major sites (including Loch Leven) held 40% of all cormorants and had adequate data sets to enable monthly variation in cormorant distribution between habitat types to be explored. Marine sites comprised rocky shore areas on the Fife and Lothian coastlines. Estuary areas included the Forth upstream of the Forth Road Bridge, the Tay upstream of the railway bridge, and smaller areas including Tentsmuir Point, Eden Estuary and Aberlady Bay. Rivers included the Forth upstream of Fallin, the Avon, Carron, Teith and Devon tributaries, the Tay upstream of Earn mouth and the Earn.

26.3 Results 26.3.1

Multiple counts

Multiple counts over a 31-day period illustrated large-scale movements, e.g. on 2 March 120 birds arrived and on 6 March 100 departed, and periods when numbers were relatively stable, such as from 21 to 25 February (Fig. 26.1). The arrival and departure of cormorants was confirmed by casual observations during fieldwork, although there was no systematic collection of such data. Most birds travelled as individuals or small groups, but larger parties of over 100 were occasionally observed.

Figure 26.1 Within-day and between-day fluctuations in cormorant counts between 11 February 1998 and 13 March 1998

26 11/02/2003 15:17 Page 348

348

Management

Mean numbers of cormorants were calculated for each of the morning, mid-day and afternoon counts on the three-count days, and for each of the morning and afternoon counts on the two-count days. Although there was some variation between the means, Z tests for matched pairs showed that these differences were not significant and there was no evidence of a diurnal pattern in the fluctuations. Counts conducted by four individual counters during overlapping periods were compared and found to be reasonable close. Differences ranged from 1.1 to 7.1% of the lower count, and were not statistically significant. This suggests that day-to-day differences in excess of this were likely to reflect true changes in numbers present.

26.3.2

Short-range radio tracking

Of the three birds fitted with radio transmitters in March 1997, one bird disappeared the following day, and was not located again. The second bird stayed for 6 days but was absent for 1 day during that time. The third left after 6 days before returning 2 days later, and stayed for between 5 and 8 days, the uncertainty being due to a receiver problem. Of the three birds fitted with radio transmitters in October 1997, one transmitter was relatively weak and tracking was insufficiently reliable for proving presence/absence. The other two birds were tracked until December, and were present for 71% of 76 days, and 41% of 82 days (Fig. 26.2).

Presence/Absence of cormorant 6.7 – 8 October to 31 December 1997 Present/Absent

1

22-Dec

27-Dec

22-Dec

27-Dec

17-Dec

12-Dec

7-Dec

2-Dec

27-Nov

22-Nov

17-Nov

12-Nov

7-Nov

2-Nov

28-Oct

23-Oct

18-Oct

13-Oct

8-Oct

0

Date

Presence/Absence of cormorant 9.1 – 8 October to 31 December 1997 Present/Absent

1

17-Dec

12-Dec

7-Dec

2-Dec

27-Nov

22-Nov

17-Nov

12-Nov

7-Nov

2-Nov

28-Oct

23-Oct

18-Oct

13-Oct

8-Oct

0

Date

Figure 26.2 Presence (blocked areas) and absence of radio-tracked cormorants 6.7 and 9.1 at Loch Leven from 8 October to 31 December 1997

26 11/02/2003 15:17 Page 349

Cormorant turnover and management implications

349

22-Mar

17-Mar

12-Mar

7-Mar

2-Mar

25-Feb

20-Feb

15-Feb

10-Feb

5-Feb

31-Jan

26-Jan

21-Jan

16-Jan

11-Jan

6-Jan

1-Jan

27-Dec

0 22-Dec

Presence/Absence

Presence/Absence of cormorant 8.2 – 22 DEC 1997 to 25 MAR 1998 1

Date

23-Mar

16-Mar

9-Mar

2-Mar

23-Feb

16-Feb

9-Feb

2-Feb

26-Jan

19-Jan

12-Jan

5-Jan

29-Dec

0 22-Dec

PResence/Absence

Presence/Absence of cormorant 5.7 – 22 DEC 1997 to 25 MAR 1998 1

Date

Figure 26.3 Presence (blocked areas) and absence of radio-tracked cormorants 8.2 and 5.7 at Loch Leven from 22 December 1997 to 25 March 1998

Of the three birds fitted with radio transmitters on 21 December 1997, one left after four days and was not seen again. The others were present for 76% of 29 days and 28% of 94 days (Fig. 26.3). Although there was no concerted effort to track departing birds away from Loch Leven, Cormorant 5.7 was located several times on the Forth Estuary. On two occasions it was also tracked departing with other birds in the early morning and flying to feed on another freshwater site before returning to the loch later the same day. When presence and absence data are compiled for all birds, the total time present represents 51% of 303 days. This suggests the number of individuals using the loch is at least double the mean population counted. It was not possible to determine the length of stay of individual birds, as time spent on the loch before the radio transmitters were fitted was not known and birds may have remained beyond the date when radio transmitters failed. However, the period over which transmissions were received may be used to indicate a minimum length of stay. The mean minimum length of stay for all nine birds was 43 days (SE  12.99). Allowing for 51% attendance, over the seven months from September to March this suggests that the number of individuals using the loch may be up to 9.7 times the mean count, which over the past ten winters was 233. These figures allow an estimate of the mean number of individual cormorants using the loch of between 466 and 2260.

26.3.3

Satellite telemetry

The nine transmitters lasted from 13 to 56 days, with a mean of 27 days. A total of 564 fixes were recorded, but their accuracy depended on the number of sequential trans-

26 11/02/2003 15:17 Page 350

350

Management

Figure 26.4 Diurnal variation in recording of all telemetry fixes and accurate fixes

missions received by the satellite. Thus, 198 Class B fixes were based on only two transmissions received and were relatively inaccurate. Of increasing accuracy were 153 Class A fixes and 150 Class 0 fixes, based on three and four fixes received respectively, but the accuracy could only be quantified as 1000 m or less for 63 Class 1, 2 and 3 fixes. Diurnal variation in the timing of different classes of fix appears to follow similar profiles (Fig. 26.4). The daily peaks represent times when the two satellites were approaching the overhead and were thus best placed to receive transmissions. The early winter mornings, when cormorants were most active, fall within one of the least suitable time for obtaining fixes, so there is likely to be some bias towards recording them at roost sites rather than when feeding. Furthermore, when actively diving, the transmission cycle is subject to repeated interruption and fixes are much less likely to be obtained. Calculation of apparent flight speed between consecutive fixes indicated speeds of up to 985 m s1, which were unachievable, and 59 fixes where the speed exceeded 20 m s1. Consequently, 41 fixes, which were considered biologically implausible, were discounted. Of these, 34 were Class B, five were Class A and two were Class 0, which confirmed that Class B fixes were much less accurate that Classes A or 0. The satellite tracking record of one bird, cormorant 470, is shown in Figure 26.5 by way of an example. It was tracked for 44 days and produced 110 useable fixes. Most were centred on Loch Leven, the Forth and Tay estuaries, but on 19 and 20 October three consecutive fixes placed the bird on the west coast. Cormorant 467 was tracked for 56 days and produced 157 useable fixes. Although most are in south-east Scotland, for two days the bird was located 280 km north at Orkney, and two fixes were located on the west coast near Oban. Cormorant 471 spent most of the time in south-east Scotland, but the last two fixes placed the bird near the west coast in the vicinity of Loch Lomond.

26 11/02/2003 15:17 Page 351

Cormorant turnover and management implications

351

) )

)

) ) )

OBAN

)

)

) )

) )

SCOTTISH WESTCOAST LOCH LOMOND

)

TAY ESTUARY ) )

) ) ) )) ) ) ))) ) ) ) )))) )) LOCH LEVEN ))))))) ) ))))))))) ) ) ) )))))) )) ))))) )) )) ) ) ) )) FORTH ESTUARY ) ) ) ) ) ) ) ) ) ) ) ) )

50 kilometres

29470

Figure 26.5 Satellite tracking record of cormorant 470 between 7 October and 19 November 1997

Cormorant 463 spent most of the time on the Forth estuary, and also visited the west coast, but was only recorded in the vicinity of Loch Leven on one occasion following capture. The 63 fixes with errors of 1 km or less gave a maximum range from Loch Leven of 44 km. A circle of radius 45 km from the centre of Loch Leven would thus incorporate all the accurate fixes, plus 92% of Class 0 fixes, 89% of Class A fixes and 70% of Class B fixes. Such a circle may reasonably be regarded as encompassing the normal home range of Loch Leven birds, although they do travel further at times.

26.3.4

The wider cormorant population

Mean winter cormorant counts from 179 sites within 45 km of Loch Leven totalled 1635. However, not all sites in the area were counted and coverage was 72% for marine sites, 61% for estuaries, 42% for major rivers and 60% for still waters. Densities were calculated for each habitat and applied to uncounted areas to provide a correction to the count total. Cormorant density on Loch Leven was judged to be atypical so it was omitted from the density calculation for still waters. When corrections were applied, a total area cormorant population of 2317 was estimated. This figure lacks confidence limits, and its calculation assumes that uncounted areas are broadly as attractive to cormorants as counted areas. It represents a crude approximation to the numbers which may be present in the wider area, and should therefore be treated with caution. The distribution of cormorant counts with respect to habitat types showed the majority of birds were found on saltwater sites (35% around estuaries and 34% around the coast),

26 11/02/2003 15:17 Page 352

352

Management

and still waters (27%) and rivers hold relatively few birds (4%). Analysis of monthly data across the winter showed that though there was considerable variation between sites, there was a decline of 50% at marine sites during the winter and 70% at estuaries. This was accompanied by a slight decline in total numbers in the area and by a shift to freshwater sites.

26.4 Discussion The multiple counts provided evidence of large fluctuations in cormorant numbers with the population halving or doubling on occasions. Short-range radio tracking revealed intermittent rather than constant presence of individual birds, a different situation from that reported by Yésou (1995). In addition, tracking of one individual indicated short-term absences of a few hours as well as longer absences, and the overall presence/absence results may underestimate such short-term absences. It is apparent from satellite tracking that Loch Leven birds travel through central Scotland and as far as the west and north coasts, but activities are concentrated in eastern Scotland, particularly within 45 km of the loch. This range incorporates the marine and estuarine areas of the Firths of Forth and Tay, lowland river systems and upland and lowland freshwater lochs. The actual number of cormorants passing through is probably nearer the upper end of the estimate from radio tracking of between 466 and 2260, as the population was not obviously depleted by the shooting of up to 370 birds each winter. The estimate of the wintering population within 45 km of Loch Leven of 2317 is 10 times the Loch Leven mean, and close to the upper end of the estimate from radio tracking. Cormorant distribution between sites changes markedly during the winter with a shift from salt water to fresh water. These results suggest that birds leave to sample other sites, and birds visit to sample Loch Leven do not support the concept of a discrete Loch Leven population, amenable to manipulation and control, which raises questions regarding the justification for killing birds on Loch Leven. Given the scale of turnover, shooting as a mitigation measure is unlikely to be effective. Shot birds would have left anyway, and new arrivals compensate for those killed. To be effective, control measures would have to apply to the wider population, and the scale of shooting required to deplete it to a level where cormorants no longer winter at Loch Leven is likely to prove unacceptable. Furthermore, this wider population may not itself be a discrete entity, and there may be interaction with birds wintering beyond 45 km, so depletion through large-scale shooting may simply be compensated by immigration. The evidence suggests that due to their foraging behaviour, cormorants are not sufficiently sedentary to enable the effective use of shooting as a site-specific mitigation measure.

Acknowledgements This study involved many hours of fieldwork in all weathers, and I am most grateful for the efforts of the team who made it possible. In particular Alan Lauder for catching

26 11/02/2003 15:17 Page 353

Cormorant turnover and management implications

353

cormorants and fitting the transmitters, and Paul Brooks, Ben Wright, Ian Parkinson and Jim Walls for counting and tracking. Additional data were supplied by Willie Wilson of Loch Leven Fisheries, and by The Wildfowl and Wetlands Trust. My thanks go to Bob Furness for his helpful comments on the draft chapter, and to the University of Glasgow, the Open University Crowther Fund, Scottish Natural Heritage and Scottish Office Rural Affairs Department for financial and material support.

References Allison A., Newton I. & Campbell C. (1974) Loch Leven National Nature Reserve; A Study of Waterfowl Biology. WAGBI Conservation Publication, Sevenoaks: The Caxton & Holmesdale Press, 124 pp. Buchheim A. (1997) Temporal limits of overwintering in migratory Cormorants and the influence of frost-periods on wintering individuals. Supplemento alle Ricerche Biologia della Selvaggina XXVI, 111–118. Carss D. & Marquiss M. (1992) Cormorants and the Loch Leven Trout Fishery. Institute of Terrestrial Ecology Project T135c1 report to Scottish Natural Heritage. Edinburgh: Scottish Natural Heritage, 36 pp. Carss D. & Marquiss M. (1997) The diet of cormorants Phalacrocorax carbo in Scottish freshwaters in relation to feeding habitats and fisheries. Ekologia Polska 45, 207–222. Coulson J.C. & Brazendale M.G. (1968) Movements of cormorants ringed in the British Isles and evidence of colony specific dispersal. British Birds 61, 1–21. Cramp S. & Simmons K. (1977) The Birds of the Western Palaearctic, Vol.1. Oxford: RSPB/Oxford University Press, pp. 200–207. Kirby J.S., Gilburn A. & Sellers R.M. (1995) Status distribution and habitat use by Cormorants Phalacrocorax carbo wintering in Britain. Ardea 83, 93–102. McKay H., Furness R., Russell I., Parrott D., Rehfisch M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. Report to the UK Ministry of Agriculture, Fisheries and Food, London, 256 pp. Marquiss M. & Carss D.N. (1994) Avian Piscivores: Basis for Policy. Institute of Terrestrial Ecology Report 461/8/N&Y to National Rivers Authority, Water Research Centre: Medenham, 104 pp. Mellin M. & Mirowska-Ibron I. (1997) Results of cormorant Phalacrocorax carbo control in northeastern Poland in 1987–1992. Ekologia Polska 45, 305–308. Sellers R.M. & Sutcliffe S. (1987) Colour-ringing cormorants. B.T.O. News 150, 13. Van Eerden M. & Zijlstra M. (1997) An overview of the species composition in the diet of Dutch cormorants with reference to the possible impact on fisheries. Ekologia Polska 45, 223–232. Yésou P. (1995) Individual migration strategies in cormorants passing through or wintering in western France. Ardea 83, 267–274.

27 11/02/2003 15:18 Page 354

Chapter 27

The impact of fisheries on water birds in Estonia: nature protection and socio-economic perspectives M. VETEMAA*, R. ESCHBAUM, M. EERO and T. SAAT Estonian Marine Institute, University of Tartu, Tartu, Estonia

Abstract The share and value of the catch from various fishing activities in Estonia in 1998 and 1999 was evaluated against the risk of capture of rare and endangered birds. Important Bird Areas cover nearly half of the west Estonian coastline, but the value of high risk (gillnet) fishing is 1% of the total value of catch and 3% of the value generated by coastal and inland fishing, corresponding to annual revenues of not more than €100 000–150 000. Despite the low revenues, around 700 coastal fishermen are active in west Estonian counties, and changes in fishing patterns will be not easy to achieve. Keywords: bird conservation, Estonia, fishery, protected areas, waterfowl.

27.1 Introduction Fisheries have historically played a very important role in the economy of the Estonian coastal regions. In many coastal and some inland areas, fisheries still constitute one of the main employment sectors. As a candidate country for membership of the European Union, Estonia is obliged to adjust its nature protection administration and legislation to EU requirements. Among these are the Birds Directive (79/409), the Habitats Directive (92/403) and the Natura 2000 network. There are 111 Species of European Conservation Concern (SPEC) (this and subsequent criteria are given according to Tucker & Health 1994), which regularly breed in Estonia (Kalamees 2000). Further, 58 of these species are classified as having unfavourable conservation status in Europe (Tucker & Health 1994; Kalamees 2000). Many of them, as well as some other species wintering or migrating through Estonia, are tightly connected to water habitats, such as the globally threatened (SPEC 1) Steller’s eider, Polysticta stelleri (Pall.), and lesser white-fronted goose, Anser erythropus (L.). Some species like the sandwich tern, Sterna sandvicensis Lath., the black guillemot, Cepphus grylle (L.), the common gull, Larus canus L., and the redshank, Tringa totanus (L.), are

*Correspondence: M. Vetemaa, Estonian Marine Institute, Vanemuise 46, University of Tartu, Tartu 51014, Estonia (email: [email protected]).

27 11/02/2003 15:18 Page 355

Fisheries and water birds in Estonia

355

listed in SPEC 2, and many species such as the black-throated diver, Gavia arctica (L.), the gadwall, Anas strepera L., the little gull, Larus minutus Pall., and the bittern, Botaurus stellaris (L.) are in SPEC 3. Sixty-two Estonian bird species are listed in the EU Birds Directive Annex 1 (BDA 1). In Estonia, 57 Important Bird Areas (IBA) have been defined (Kalamees 2000), which in west Estonia cover nearly half of the coastline. The bird species listed in BDA 1, SPEC 1 and SPEC 2 – hereafter referred to as ‘birds of special concern’ – need strict protection regimes, the most important component of which is habitat protection through the establishment of special protected areas (SPAs). Furthermore, the Birds Directive states that, in addition to killing or chasing, the ‘deliberate disturbance of these birds particularly during the period of breeding and rearing’ is also prohibited. This also concerns SPEC 3 species. As some BDA 1 species like the Arctic tern, Sterna paradisaea Pont., and the common tern, Sterna hirundo L., and SPEC 2 species, e.g. redshank and common gull, are very common in Estonian coastal areas, problems associated with bird conservation and the establishment of SPAs have generated concern in Estonian fishing communities. The establishment of new protected areas and the implementation of new and stricter rules for nature conservation on already existing areas may have negative consequences for fisheries. It is highly likely that some fishermen will be unable to continue their historical pattern of fishing activities as new closed areas and closed time periods appear. Owing to the very rapid character of adoption of necessary legislative acts concerning nature protection, there is a danger that fishermen will not be included in the decision-making process or be informed about all the effects of the new system. The impact of fisheries on birds can be divided into two components: (a) drowning (bycatch) of water-birds in fishing gears, and (b) fishing activities carried out in the vicinity of nesting sites may disturb birds. The present study aims to evaluate the share and value of the catch obtained in Estonia in 1998–1999 from fishing activities which potentially have a high risk of catching birds of special concern.

27.2 Material and methods Estonian fisheries were divided on the basis of gear, fishing area (170 in the coastal sea and 100 on Lake Peipsi-Pihkva, Fig. 27.1) and fishing month into segments or fishing activities. It was hypothesised that the effect of such separate fishing activities to birds remains stable from year to year. All fishing activities were classified as ‘safe’ or ‘high risk’ according to the risk of bird bycatch. Only those fishing activities with a risk of bycatches of birds of special concern were classified as high risk. The data on Estonian total catches by different segments originate from the Department of Fish Resources, Ministry of the Environment of Estonia. Catches by gears in different segments were obtained from fishermen’s logbooks. Average first-buyer prices were obtained from the Estonian Environmental Inspectorate. Since there is no central database of fishermen in Estonia, the number of fishermen used in the study was based on the number of fishing licences issued by counties in 1998–1999. Consequently, the data reflect the total number of people participating in commercial fishing and not the number

27 11/02/2003 15:18 Page 356

356

Management

Figure 27.1 Research area. The names of the western counties are given in italics

of full-time fishermen. The main source of ornithological information (abundance of species, importance of different areas) was the publication Important Bird Areas in Estonia (Kalamees 2000). Estonian fishermen are obliged to note bird bycatches in their monthly reports. However, since fishermen are afraid that this information may affect their fishing possibilities in the future, reliable data on bycatches of birds (as well as grey seal Halichoerus grypus Fabricius and ringed seal Pusa hispida Schreber ) are lacking. According to the logbook data and interviews with fishermen, only gillnet fisheries produced bycatch of waterfowl, therefore fishing with other gears (trawling, traps, fykes, pound nets, seines and longlines) was classified as safe. Gillnet fishing in areas not important for birds of special concern, or during the period when they are not present in Estonia, was also classified as safe. When the impact was unknown or disputable, the fishing was classified as high risk.

27.3 Results 27.3.1

The value created by different sectors of Estonian fishing, number of fishermen

The Estonian Baltic fishery has two distinctive sectors: trawl fishery and coastal artisanal fishery. The open-sea trawl fishery targets herring (Clupea harengus membras L.) and sprat, (Sprattus sprattus (L.)) and, to a lesser extent (and mostly outside of the Estonian Exclusion Zone), cod (Gadus morhua L.). Cod and salmon (Salmo salar L.) fisheries

27 11/02/2003 15:18 Page 357

Fisheries and water birds in Estonia

357

Table 27.1 Value of catch (€’000s) and share (%) of fishing not endangering birds of special concern in Estonian fishing (Atlantic fishing excluded) in 1998–1999 Value of catch Area

Fishing segment

1998

1999

Baltic Sea

Open sea fishing Coastal fishing Lake Peipsi-Pihkva Lake Võrtsjärv Other

9 044 2 430 2 008 272 57 13 811

6 738 1 631 1 901 289 38 10 597

Inland waters

Total

Share (%) of safe fishing 1998–1999 100 95 99 99 95 99

also use drift nets, but there no such licences were issued for Estonian waters in 1998– 2000. Trawling is allowed at depths over 20 m. Baltic coastal fisheries which target 15–20 species, use trap and fyke nets (including pound nets for herring), gillnets and to a lesser extent seines and longlines. Economically, the most important fishery is the herring pound net fishery on spawning grounds in spring and early summer. Gillnet fisheries, both legal and illegal, are common in all coastal areas. On Lake Peipsi-Pihkva (Fig. 27.1), Danish seining is employed, which is economically the most profitable fishing method in inland waters and generates the highest revenues (Vetemaa et al. 2001). Trawling is prohibited. Most of the catches are taken by seines, fyke nets and gillnets. Gillnets are mostly used in winter for under-ice fishing, but also in spring and autumn. The main target species of the gillnet fisheries are pikeperch (Stizostedion lucioperca (L.)), perch (Perca fluviatilis L.), bream (Abramis brama (L.)), roach (Rutilus rutilus (L.)), pike (Esox lucius L.) and whitefish (Coregonus laveratus maraenoides Poljakow). The value of catches by the main Estonian fisheries in 1998–1999 is presented in Table 27.1. The number of fishermen in 1998–1999 was approximately 800 in the Baltic open-sea fishery, 1500 in the Baltic coastal fishery and 500 in the inland fishery. There are about 700 coastal fishermen operating in the territory of three west Estonian counties (Saare, Hiiu, Lääne) (Fig. 27.1), where the bulk of the high-risk fishing activities take place and also where most all important IBAs for water birds are situated.

27.3.2

High risk fishing activities

Waterfowl seldom get stuck in fyke, trap or pound nets. Gillnets, by contrast, are high risk, especially to diving ducks, eiders and grebes. One of most critical species is Steller’s eider wintering in coastal areas of west Estonia, where they spend the majority of the time in shallow, near-shore areas where it is common for gillnet fishing to target sea trout (Salmo trutta L.) and whitefish. Safe gears in coastal fisheries not responsible for bycatch of birds are pound nets (targeting herring, garfish Belone belone (L.)) and fyke nets (targeting eel, Anguilla

27 11/02/2003 15:18 Page 358

358

Management

anguilla (L.)). In total these gears were responsible for 71% and 54% of total revenues from Estonian coastal fishing in 1998 and 1999, respectively. The most important area for coastal fisheries by value is Pärnu county (Fig. 27.1). A substantial part of the catches are taken on the territory of IBA 052, which aims to protect migrating geese (Kalamees 2000). However, as geese are not diving birds, they are not endangered by fyke nets and gillnets. This area is also important for migrating blackthroated divers, which normally pass during the second half of May and from the second half of September to the first half of October (A. Kalamees, unpublished data). During the rest of the year the fishing in this area can be considered to be of no danger to the birds of special concern. If the catches taken in those critical periods are excluded, it appears that the value of safe catches taken in Pärnu county (excluding herring, garfish and eel) corresponded to 14 and 20% to the total value of Estonian coastal fishery in 1998 and 1999, respectively. If the value of safe fishing carried out in other counties is added, at least 95% of the Estonian coastal fishery revenues in 1998 and 1999 were obtained through fishing activities not endangering BDA 1 and SPEC 1–3 bird species. Although the value of the catches in inland waterbodies is approximately the same as in the Baltic coastal fisheries, the risk of bycatch of birds of special concern is much smaller. None of five IBAs on Lake Peipsi-Pihkva is important for diving birds. The value of the catch by gillnets during the ice-free period in 1998–1999 was less than 10% of the total value of the catch. On Lake Võrtsjärv, there are no nominated, important bird areas. IBAs on the River Emajõgi are mainly aimed at protecting wading and raptorial birds (Kalamees 2000), and the rather small-scale fishery (mostly fyke net fishing in spring) does not produce a bycatch of birds. Numerous small lakes across Estonia are not extensively fished, and mostly by recreational fishermen, therefore data on catches and bycatches are lacking. However, no IBA on those lakes has been established due to their unimportance for diving birds. The annual value of the official catch from fishing activities of high risk to birds of special concern in Estonia did not exceed €100 000–150 000 in 1998–1999. Over 80% of the catch was taken in the west Estonian coastal sea.

27.4 Discussion The impact of fishing on water birds has been widely discussed. According to Stempniewicz (1994), in the southern Baltic up to 20% of the wintering population of long-tailed ducks are killed annually in gillnets. This species is also very abundant (35% of drowned seabirds) in the Gulf of Riga (Urtans & Priednieks 2000). However, this study also reported a high share of red-throated divers (Gavia stellata (Pont.)) and black-throated divers (BDA 1 species) in bycatches, totalling 20% of drowned birds in the Gulf of Riga. The use of gillnets, mostly monofilament nets which produce the largest bycatches (Lien et al. 1989), has doubled in Estonia in 1990s (Vetemaa et al. 2000). Thus a substantial increase of bycatches of birds is likely in the Estonian fishery. However, Tasker et al. (2000), who compiled data on many issues including both direct and indirect effects, concluded that fishing effects on the population abundance of birds are hard to demonstrate, even if they do exist.

27 11/02/2003 15:18 Page 359

Fisheries and water birds in Estonia

359

The value of catches obtained by fishing activities endangering BDA 1 and SPEC 1–3 birds is not high in Estonia. Despite the high risk of certain fishing activities to birds, both Estonian and EU regulations, in general, do not legislate against these methods. It may, however, be possible to adjust fishing patterns to the needs of nature conservation without decreasing revenues. In this context, some further restriction of gillnet fisheries is recommended as it is not only of high risk to birds but is also responsible for overfishing of pikeperch, pike and perch stocks in the coastal sea (Saat & Eschbaum, in press). One way to illustrate the importance of annual revenues equalling between €100 000 and 150 000 to the Estonian fisheries sector is to calculate the number of jobs this could create. The Estonian average annual salary, including social tax, is currently around €5500. If fishing costs correspond to roughly half of revenues, the value created through critical fishing activities in Estonian small-scale fisheries equates to 10–15 working places on an annual basis, which is less than 1% of the people engaged in small-scale fishing. Calculations of both total monetary value and employment value of high-risk fishing activities revealed that the conflict between fisheries and bird protection interests is marginal in Estonia. However, people’s expectations and attitudes are seldom based on monetary calculations alone. The role of coastal fishery as a source of livelihood has decreased markedly during the last decade, due to the rapid growth of other sectors of the economy, causing a more than 10-fold increase in average salaries in the past 8 years, and to overfishing during the first half of the 1990s (Vetemaa et al. 2000, 2001). This has caused a serious decline in the standard of living in fishing communities in comparison to people employed in other sectors of the national economy. Further, a mail survey of Hiiu County fishermen in 2000 (M. Vetemaa, unpublished data) showed that almost half of the respondents connected the decline of fish stocks and the fall of coastal fisheries to the effects of nature protection in the form of an increased number of cormorants, Phalacrocorax spp., and seals. In several coastal areas, such an understanding has already given birth to a hostile attitude against nature protection in general, as well as to illegal actions such as culling of cormorant nestlings. Most of the commercially important coastal fish stocks have been seriously overfished in Estonia during the 1990s. As an example, the perch catch in the west Estonian Archipelago Sea in 1999 (ICES sub-area 29-4) was only 1% of the average of the last 30 years (Vetemaa et al. 2000). Since perch is extensively fished by gillnets, any improvement in the perch stock abundance in future will also potentially increase both the gillnet fishing effort and the value of the catch obtained through fishing activities endangering birds. Even if catches are small now, the number of fishermen in west Estonia performing potentially high-risk fishing activities is still around 700. Despite the low financial value of the catch through fishing activities causing a high risk of bycatch of birds of special concern in 1998–1999, the whole problem is complex and should be taken very seriously. At the present time, parallel to the rapid establishment of Natura 2000 sites, the most important tasks of fisheries and nature conservation scientists and managers must be both the detailed analysis of unavoidable changes facing the Estonian fisheries sector after adoption of EU legislation and the dissemination of information to stakeholders.

27 11/02/2003 15:18 Page 360

360

Management

27.5 Conclusions The present study revealed that the value of catch obtained through the use of fishing activities with a high bycatch risk of birds of special concern is very small in Estonian fisheries, i.e. 1% of the total value of the Estonian catch and 3% of the value generated by small-scale fishing activities (coastal and inland fishing). These figures correspond to annual revenues of not more than €100 000 to 150 000. However, despite the low revenues, around 700 coastal fishermen are active in west Estonian counties – areas that are most important for the birds of special concern. Since the success of attempts to ban dangerous fishing activities depends not only on the value of the catch produced, but also on the number of people involved, changes in fishing patterns will be not easy to achieve.

Acknowledgements This research was supported by the Estonian state-financed project ‘Biodiversity of the Estonian fauna’ (0180515s98) and the German Federal Ministry of Environment, Nature Conservation and Nuclear Safety through the Baltic Environmental Forum.

References Kalamees A. (ed.) (2000) Important Bird Areas in Estonia. Tartu: Eesti Loodusfoto, 114 pp. Lien J., Stenson G.B. & Ni I.H. (1989) A review of incidental entrapment of seabirds, seals and whales in inshore fishing gear in Newfoundland and Labrador: a problem for fishermen and fishing gear designers. Proceedings of a World Symposium on Fishing Gear and Fishing Vessel Design (St John’s) Newfoundland and Labrador Inst. of Fisheries and Marine Technology, St John’s, NF (Canada), pp. 67–71. Saat T. & Eschbaum R. (in press) The Fishery of Väinameri and its changes during last decades (In Estonian with English summary). In T. Saat (ed.) The Fishes and Fishery of Väinameri. Tartu: Tartu University Press. Stempniewicz L. (1994) Marine birds drowning in fishing nets in the Gulf of Gdansk (southern Baltic): numbers, species composition, age and sex structure. Ornis Svecica 4, 123–132. Tasker M.L., Camphuysen M.C.J., Cooper J., Garthe S., Montevecchi W.A. & Blaber S.J.M. (2000) The impacts of fishing on marine birds. ICES Journal of Marine Science 57, 531–547. Tucker G.M. & Health M.F. (1994) Birds in Europe: Their Conservation Status. Cambridge: Birdlife International (BirdLife Conservation Series no. 3), 600 pp. Urtans E. & Priednieks J. (2000) The Present Status of Seabirds By-catch in Latvian Coastal Fishery of the Baltic Sea. International Council For the Exploration of the Sea, CM 2000/J171, 8 pp. Vetemaa M., Eschbaum R., Aps R. & Saat T. (2000) Collapse of political and economical system as a cause for instability in fisheries sector: an Estonian case. In Proceedings of the IIFET 2000 International Conference. Microbehaviour and Macroresults. Oregon State University, July 10–14, Corvallis, USA. Available at: http://www.orst.edu/Dept/IIFET/2000/papers/vetemaa.pdf Vetemaa M., Vaino V., Saat T. & Kuldin S. (2001) Co-operative fisheries management of the cross border Lake Peipsi-Pihkva. Fisheries Management and Ecology 8, 443–452.

28last 11/02/2003 15:21 Page 361

Chapter 28

Interactions between fisheries and fish-eating birds: optimising the use of shared resources I.G. COWX* Hull International Fisheries Institute, University of Hull, Hull, UK

Abstract The upsurge of fish-eating birds in inland waters in Europe and elsewhere in the past decade has created considerable conflict between conservationists and fisheries managers. Despite considerable research no pragmatic solution to the problem has been forthcoming. This is largely because of the protected status given to the birds and the ineffectiveness of control measures tried to date. However, if one steps back and considers why fish-eating birds are perceived as a problem, one can start to understand the underlying issues. This chapter examines the reasons why fish-eating birds have moved inland and examines the types of fisheries that are mostly heavily exploited. This provides a key to the root cause, i.e. it is linked to loss of food resources in the preferred habitat ranges, high stocking densities of fish in inland waters and degradation of habitat with reduced cover for fishes. The chapter moves on to examine how lessons can be learnt from the management of overexploitation of fisheries and ameliorating environmental degradation to support the enhancement of fisheries on a sustainable basis, and in harmony with fish-eating birds. Keywords: conflict resolution, cormorants, fisheries, fish stocks, rehabilitation.

28.1 Introduction The relationships between birds and fish have long been recognised, especially as functional units within both marine and freshwater ecosystems. In recent years, however, there has been increasing awareness of the effect of fisheries activities, mainly exploitation, on bird populations and vice versa, i.e. the impact of expanding populations of fish-eating birds on fish stocks. Both interactions have led to growing concerns about the conservation of birds on the one hand and, on the other, the sustainability of the fisheries resources for both commercial and recreational exploitation. The information in the literature relating to these interactions is extremely fragmented and highly contradictory. Much of the contradiction arises because of the apparent conflicting objectives of those managing the resources, i.e. conservation of birds versus maintaining and improving fishery exploitation. There is thus an urgent need to bring together experts on both aspects in an attempt to identify and resolve the areas of conflicts. The symposium on

*Correspondence: I.G. Cowx, Hull International Fisheries Institute, University of Hull, Hull HU6 7RX, UK (email: [email protected]).

28last 11/02/2003 15:21 Page 362

362

Management

Interactions between Fish and Birds: Implications for Management, was convened to address this need. The objectives of the symposium were:

• • • •

to assess the interactions between fish-eating birds and fisheries from both the biological and socio-economic perspectives; to synthesise information on conflicts between fish-eating birds and fisheries, how conflicts arise through bird and fish population ecology and exploitation, and mechanisms for resolving these conflicts; to review current management practices in relation to fish-eating birds and fish stocks and identify constraints and gaps in our knowledge that affect the application of fisheries management policy; to develop strategies for the conservation and management of fish-eating birds in harmony with fisheries interests.

To draw conclusions from such a wide coverage of information is complex. Many aspects need to be brought together to highlight limitations in our knowledge of the interactions between fish-eating birds and fisheries, and promote the positive aspects for sustainability of both resources for future generations. This chapter synthesises the main outputs of the symposium, including papers not submitted for presentation within these proceedings. It is broken down into three key areas: the impacts of fish-eating birds on fisheries and current actions on mitigation of these impacts; the interactions between fisheries activities and birds; and concludes with possible options for the management of fisheries interests in harmony with the conservation of fish-eating birds.

28.2 Bird–fisheries interactions The upsurge of fish-eating birds, especially cormorants, in inland waters in Europe (e.g. Dekker & De Leeuw, Chapter 1; Davies et al., Chapter 3; Stempniewicz et al., Chapter 5; Knösche, Chapter 6; Eschbaum et al., Chapter 7) and elsewhere in the world (see, for example, Coutin & Reside, Chapter 15, and Farquhar et al., Chapter 24) since the late 1980s has created considerable conflict between conservationists and fisheries managers. Reasons behind the birds’ upsurge are complex and varied but can be related to a number of events. Perhaps the most significant reason in western Europe was the promulgation of the EU Birds Directive (EEC/79/409) which afforded protected status to many species of birds, especially cormorants. As a consequence, the species can no longer be persecuted in the name of protecting fisheries from damage. One upshot of this reduced persecution has been the apparent colonisation and adaptation of Phalacrocorax species to new habitats and food sources associated with inland waters. Whether this is a new phenomenon is debatable because cormorants existed around inland waters in the past (see Wright, Chapter 26), but their numbers were reduced by shooting or other control measures. In the UK there is also debate as to whether the influx of cormorants is of the European inland species Phalacrocorax sinensis or the endemic P. carbo species. Notwithstanding these arguments, the cormorant’s increased presence, and that of several other species, e.g. goosanders, Mergus merganser L. (see Wilson et al., Chapter 9), is perceived as detrimental to inland fisheries

28last 11/02/2003 15:21 Page 363

Interactions between fisheries and fish-eating birds

363

through depredation of the stocks, and thus is the cause of much conflict. It should be noted that similar protection has been afforded fish-eating birds in other parts of the world leading to similar scenarios.

28.2.1

Assessing impact of piscivorous birds on fish stocks

The impact of bird depredation on fisheries is varied both in real and perceived terms. Figure 28.1 illustrates the main issues relating to fish–bird interactions based on expert opinion (including both bird and fisheries representatives) of the problems found in some 50 fisheries across Europe. (Note that the exercise was carried out as part of a workshop prior to the main symposium.) Numerous studies have also reported on detrimental impacts of birds on fisheries but these are often refuted by the bird lobby [see Russell et al. (1996) for review, and Suter (1995) versus Staub et al. (1998) for an example of the conflicting debate]. Much of the problem in assessing the impact of birds on fisheries arises because the spatial distribution, abundance, size distribution, recruitment and growth of fish populations/communities are regulated/controlled by many biotic and abiotic factors (Fig. 28.2), of which bird predation is but one. There is, therefore, no single factor that is directly responsible for constraining the development of fish populations, rather there are complex interactions between many factors. In addition, one must consider the effects that human activities have had on the freshwater reduced catch reduced stock – lowered production reduced value of catch (damage) loss of stocked fish reduced revenue effects on popn dynamics/comm structure loss of juvenile fish – lowered recruitment loss of spawners reduced catchability (stress/behav) shooting disturbance threats to endangered fishes increased recurrent costs reduced capital values of fisheries removal of fish from nets loss of aquaculture stock damage to vegetation loss of employment damage to fishing gear reduced fishing tackle sales vectors of diseases/parasites drowning in fishing gear landscape alteration interactions with other birds eutrophication lead poisoning 0

10

20

30

40

50

60

Number of responses

Figure 28.1 Principal impacts of fish-eating birds on fisheries and ecosystems based on expert judgement

28last 11/02/2003 15:22 Page 364

364

Management

OTHER PISCIVOROUS PISCIVOROUS FISH PREDATORS BIRDS

PISCIVOROUS BIRDS

PREDATION

PREDATION

CLIMATIC EVENTS STARVATION

PISCIVOROUS FISH

NATURAL MORTALITY

NATURAL MORTALITY

OTHER PREDATORS ENVIRONMENTAL AND HABITAT FACTORS

FISHING MORTALITY

DISEASE AND PARASITISM MIGRATION DISPERSION

EGGS, LARVAE & JUVENILE FISH

GROWTH & MATURATION

FISH YIELD

ADULT FISH & SPAWNING STOCK

REPRODUCTION

STOCKING

FOOD RESOURCES

BIOTIC AND ABIOTIC FACTORS

FOOD RESOURCES

BIOTIC AND ABIOTIC FACTORS

ANGLER SUCCESS

FISHERY VALUE

Figure 28.2 Model to illustrate the complexity of factors to be addressed to identify the factors influencing coarse fish populations

environment and how these might influence fish populations/communities (see Cowx 2002a for review). Notwithstanding the conflicts between parties and complexity of identifying direct evidence of impact of fish-eating birds on fish stocks, some indicators of potential impact are available.

• • • • •

Cormorants consume 400–800 g per day per individual bird (Carss et al. 1997); Total loss of stock – usually on fish farms and intensively stocked fisheries (e.g. Pilcher & Feltham 1997); Predation on part of the stock – usually natural fisheries (see Russell et al. 1996 for examples); Wounding/damage to fish – mainly where the stock comprises large individuals (Russell et al. 1996); Scaring of fish, especially on natural water bodies (Russell et al. 1996).

These impacts can all lead to direct economic loss to the fishery. The net loss of fish through increased depredation, and wounding and scaring of fish, all potentially reduce the fishery performance and this has economic implications for the fishery owners. Where the loss is realised in a decline in the catch, the usual response is to replace the fish through stocking which has financial implications. In addition, reduced fishery performance tends to discourage anglers and this leads to a reduction in income. Occasionally, however, depredation can result in a positive outcome realised through accelerated growth rates (Britton et al. 2002). Basically, the reduced stock density creates great scope for growth, and fish pass the critical stage of intense predation pressure quicker, thus making more larger fish available to anglers. This latter aspect serves to

28last 11/02/2003 15:22 Page 365

Interactions between fisheries and fish-eating birds

365

illustrate the problems with assessing impact in real terms. These problems led Feltham et al. (1999) to come to the conclusion that assessing the ‘impact by fish-eating birds is a problem for specific fisheries rather than a general one’. It appears that at some sites depredation levels may be high enough to cause a decline in the fishery and at others they may not. There is no single level of depredation in terms of, for example, the proportion of standing crop removed by birds and no single estimate of impact (e.g. percentage annual decline in catches) that can be taken as the threshold above which loss is considered detrimental. Each fishery appears to have its own threshold set by the complex interaction between bird depredation and fish population dynamics, between consumption and production.

28.2.2

Current actions to resolve the impact of piscivorous birds on fish stocks

McKay et al. (1999) reviewed the effectiveness of various management measures to control damage by fish-eating birds to inland fisheries. They examined the effectiveness of a number of measures, the main ones being:

• • • • • • •

shooting; egg destruction; human disturbance; laser guns; conditioned taste aversion; stocking control; fish refuges/habitat management.

McKay et al. (1999) concluded that no one single management intervention was effective at mitigating the problems created by cormorants. Shooting does not appear to be a viable option (Wright, Chapter 21) unless the numbers are reduced to the levels of the past. Continuous dispersal and turnover of birds appears to result in a more or less stable population size, particularly fisheries. Furthermore, legislation and public reactions would prevent such an action. Controlling of the bird population density by destroying nesting areas, oiling eggs, etc., is again only likely to have a localised effect and be short term. Similarly, scaring methods (human disturbance, laser guns and taste aversion) do not appear to be effective because they must be carried out on a continuous basis, birds become accustomed to the methods employed, and the problem is probably dissipated to other fisheries. Large-scale exclusion devices are not feasible as they restrict or prohibit fishing activities. However, some success has been achieved with fish refuge devices (Russell et al., Chapter 19; McKay et al., Chapter 20). These features included artificial reefs or underwater fenced-off zones that do not allow access to fish-eating birds. The problem is the refuge structures often interfere with angling either by concentrating fish, thus making them more easily caught, or by snagging gears. The solution to the problem of bird depredation is thus complex and multi-faceted. It is unlikely legislation promulgated to protect birds will be relaxed in the future and scientific evidence/advice seems unable to provide any easy solutions. Furthermore,

28last 11/02/2003 15:22 Page 366

366

Management

irrespective of the physical measures necessary to reduce the problems, the conflicts that now exist are deep-rooted, societal issues and will not be resolved unless all stakeholders are involved in the debate. Part of this latter problem arises because the fisheries and bird antagonists have in the past often operated in isolation, and thus the prospects for resolving the issues are compromised.

28.3 Fisheries–bird interactions Although not covered explicitly in these proceedings, the converse situation exists between fish abundance, fisheries exploitation and fisheries management activities on the dynamics of fish-eating bird populations. These interactions are the result of direct fisheries activities and changes in the ecosystem dynamics which are driven by environmental change.

28.3.1

Impact of fisheries on birds

The effects of fishing on birds may be direct or indirect (Tasker et al. 2000). Most direct effects involve killing by fishing gear such as drowning in nets or mortality due to hook damage. For example, high seas drift nets have had, prior to the banning of their use, a considerable impact on seabirds in the northern Pacific, as have gillnets in south-weal Greenland, eastern Canada (Piatt & Nettleship 1987), and elsewhere, and albatrosses and petrels are a bycatch of long-line fisheries in the north Pacific and in the Southern Ocean (Brothers 1991). On a lesser scale some fishing activities also disturb birds. Indirect effects mostly work through the alteration in food supplies. Many activities increase the food supply to scavengers by providing large quantities of discarded fish and wastes, particularly those from large, demersal species that are inaccessible to seabirds, from fishing vessels (Blaber et al. 1995; Regehr & Montevecchi 1997; Stenhouse & Montevecchi 1999). Also, fishing has changed the structure of marine communities. Fishing activities have led to depletion of some fish species fed upon by seabirds, but may also lead to an increase in small fish prey by reducing numbers of larger fish that may compete with birds. Both direct and indirect effects are likely to have operated at the global population level on some species. Proving the scale of fisheries effects is difficult because of confounding and interacting combinations with other anthropogenic effects (pollution, hunting, disturbance) and oceanographic factors (Tasker et al. 2000).

28.3.2

Trophic dynamics

In many cases the contribution of a particular fish species in the diet of bird species is significantly related to both its abundance in the wild and the commercial catches (Frank et al. 1996; Garthe 1997). However, major alterations in marine ecosystems, in many cases as a result of climatic change, have resulted in shifts in the abundance and distribution of many fish species and these have impacted on the diet, and ultimately, population

28last 11/02/2003 15:22 Page 367

Interactions between fisheries and fish-eating birds

367

dynamics of the foraging bird species (Carscadden et al. 2001). For example, in 1991 a major cold-water event occurred in the north-west Atlantic and it has continued to have pervasive effects on the pelagic food webs. Large migratory warm-water pelagics, such as mackerel, short-finned squid and Atlantic saury, were replaced by cold-water species, such as capelin, Mallotus villosus, and to a lesser extent Atlantic salmon, Salmo salar L., and this was reflected in the diets of gannets, Sula bassana (Montevecchi & Myers 1996, 1997). Although the sea surface water temperatures returned to “normal” from the mid1990s, the seabird diets remained essentially unchanged, because the migratory warmwater pelagics have not returned to former levels because of a reduction in the zooplankton abundance. Concomitantly the commercial fisheries have continued to fail. Birds may also act as indicators of biological productivity at lower trophic levels and can provide a means of assessing how oceanographic conditions may influence mortality and condition of fish stocks, and ultimately the availability of stocks for exploitation. For example, Pacific herring, Clupea pallasi, is of commercial importance in California and assessment of the status of the herring population, and establishment of quotas, is based mainly on spawning biomass in the previous season. However, this metric may not reflect stock size for the following season, resulting in excessive quotas. Seabirds are good indicators of biological production at lower trophic levels and can thus provide a means of assessing how oceanographic conditions may influence the mortality and condition of adult herring (K.L. Mills, personal communication).

28.4 Options for the future The evidence reviewed in the previous sections, and elsewhere in these proceedings, suggests considerable interaction between birds and fisheries. On the one side fisheries exploitation is degrading the stocks and this is having knock-on effects on the survival of bird populations. The problem is further exacerbated by global climatic changes that are altering the ecosystem dynamics resulting in shifts in food availability for birds. While it is difficult to see solutions to the climate change problems in the foreseeable future, management of the fish stocks must consider the impacts on bird populations, and trophic ecology as a whole, and allocate resources to ensure the sustainability of avian biodiversity. Conversely, piscivorous birds do impact on fisheries, although the extent depends upon the locality and predation pressure. The common strategies to ameliorate the problems caused by fish-eating birds, e.g. scaring, shooting, seem to be ineffective, thus alternative approaches need to be developed. Many of the problems that exist with respect to the conflicts between fish and birds arise because of the social and economic importance of inland fisheries in developed countries. In these countries freshwater fisheries are not typically heavily exploited for commercial purposes and management is orientated towards recreation and conservation (Welcomme 2001; Arlinghaus et al. 2002). The economic value of recreational fisheries in these regions is huge (Peirson et al. 2002; Cowx 2002b; Arlinghaus et al. 2002). Consequently any problems that affect the recreational resources are deemed unacceptable by the angling fraternity, and become the subject of intense discussion and demands for control. The cormorant debate is one such interaction. Fisheries owners and

28last 11/02/2003 15:22 Page 368

368

Management

managers consider that the presence of cormorants has a direct impact on their stocks and livelihoods. However, this remains subject to debate (Feltham et al. 1999) and it appears that the problems are site-specific and not necessarily always detrimental (Britton et al., Chapter 2). Part of the problem arises because of a lack of awareness by all stakeholders. Fisheries managers and owners perceive the removal of fish by birds as detrimental to the stocks and are not always conversant with the positive aspects that natural mortality/ predation can have on regulating the dynamics of fish populations (see Fig. 28.2 and Britton et al., Chapter 2). Managers and owners are concerned mainly with providing a quality fishing experience and thus removal of fish is seen as reducing this prospect. This is not to say that in certain fisheries, bird predation does not reduce stocks because evidence does exist to show birds can reduce stocks dramatically to very low levels (Pilcher & Feltham 1997). Conversely, the bird conservation lobbies are often negligent of the resource depletion and economic impact that birds can have on inland fisheries. They often consider that birds are a component of the ecosystem and that they have an equal right to exploit the resources. One of the fundamental causes of this conflict is that many inland waters are intensively stocked, at great expense, to ensure good angling performance, thus cormorant depredation is directly impinging on this objective. Further problems exist with the current legislation governing fish-eating birds. Many species are protected under national and EU (Birds Directive) legislation. This prevents large-scale culling of the birds to benefit fisheries. However, if legislation is amended to allow selective hunting for cormorant, this is likely to have little effect in controlling bird numbers because of the need for large-scale culling over a wide area (see Wright, Chapter 21). This would be socially unacceptable and contrary to wider conservation and biodiversity arguments. The interaction between birds and fisheries also has serious conservation issues. Many fisheries managers and practitioners are polarised in their views about the impact of piscivorous birds on fish stocks because of the problems perceived with cormorants. However, there are many other piscivorous birds that are reliant on fish stocks in inland waters for their continued survival. Perhaps the most notable in Europe is the bittern, Botaurus stellaris L. (see Noble et al., Chapter 11), which is critically endangered and the subject of large-scale conservation initiatives (Newbury et al. 1997; José 2000). Care must be taken to ensure that actions to mitigate problems caused by cormorants do not compromise initiatives to conserve and enhance endangered piscivorous bird populations. In this respect, breeding and feeding areas of endangered birds, such as the bittern, tend to be in Special Protected Areas or nature reserves, so the birds and their food resources are afforded some protection against indiscriminate actions against birds. To resolve the perceived problems generated by cormorants moving inland, there is a need to examine the reason for the colonisation and increased abundance of inland waters. As indicated, protection of the birds under the EU Birds Directive is almost certainly a prime driver. They have benefited from the reduced persecution and adopted a strategy to exploit available food resources, both inland and presumably in coastal waters. The question that must be asked, however, is why the prevalence of cormorants has increased so dramatically in inland regions since the mid to late 1990s. The reasons possibly lie in the interactions between fisheries and birds described in

28last 11/02/2003 15:22 Page 369

Interactions between fisheries and fish-eating birds

369

Resource allocation

BIRDS

FISH STOCKS

MAN

Stock enhancement

Figure 28.3 Diagrammatic representation of resource allocation in fisheries

Section 28.3. This intimates that bird numbers and distribution are very much dictated by the food resources available. The over-harvesting of the fish resources around the coastal waters of Europe potentially mean that there are few resources for cormorants in these waters and they have moved inland for more lucrative feeding opportunities. The recent practice to intensively stock fisheries, especially still waters, for the benefit of anglers, and the increased prevalence of fish farms, have offered an ideal opportunity for cormorants. This scenario is, in part, supported by anecdotal information that suggests it is the more intensively stocked waters that are subjected to greatest predation pressure (Feltham et al. 1999). The question still remains: How can we minimise the potential conflicts between fishermen and bird conservationists? The solution probably lies in the optimisation of resource allocation of the fish stocks to satisfy both groups (Fig. 28.3). Essentially, sufficient fish must be available to satisfy the demands of the anglers/fishermen in terms of catching success while allowing the birds to co-exist. The principal mechanism adopted by fisheries managers to enhance fish stocks is stocking. While stocking has potentially benefits it is not an overarching mechanism to achieve optimal resource allocation because stocked fish tend to be naïve and prone to predation, although tactics can be employed to minimise depredation by birds on stocked fish (see below). A better strategy is to improve the status of the stocks by addressing the bottleneck in natural recruitment and enhancing the prospects of survival to an exploitable size. Several tactics can be employed to achieve these objectives. (1) Address any overexploitation problem that exists. This is only relevant if the fishery is exploited in a manner that depletes the stocks, such as is found on Lake IJsselmeer (Dekker & De Leeuw, Chapter 1). Few recreational fisheries fall into this category (mainly where the fish are removed for consumption, e.g. Scandinavian countries) and it is unlikely to be a viable option in such circumstances. (2) Rehabilitation or habitat improvement to reinstate spawning and nursery areas, and provide optimal conditions for growth and survival. The mechanisms by which this is carried out are varied and should target the bottlenecks to recruitment and survival within the specific waters (Cowx & Welcomme 1998). To ensure the expected outcome of the rehabilitation exercise requires careful assessment and planning. Invariably the actions are inadequate and fail to enhance the target populations.

28last 11/02/2003 15:22 Page 370

370

Management

(3) The final tactic is to reduce foraging opportunities for the birds. This can be achieved in several ways. Some success has already been achieved with refuges that allow access to fish but exclude cormorants (Russell et al., Chapter 19; McKay et al., Chapter 20). However, these will only function on a small scale and do have the disadvantage of interfering with angling practices, e.g. snagging hooks. A better strategy might be to examine why cormorants do not forage on all available waters. It may be that there are habitat features which make certain sites unattractive to cormorants (possibly because of water depth, poor access, poor loafing sites or similar), and if these could be recreated in other waters the problems may be reduced by habitat manipulation. Consideration should also be given to the stocking strategies adopted to enhance the fish populations. The latter include stocking with larger individual fish that are greater than the size range preferred by birds, and stocking at times when predation pressure is at its lowest, i.e. in the summer when birds are feeding at sea. Although this may impose additional costs they should be offset by improved survival of stocked fish. Finally, and perhaps must importantly, it must be recognised that fisheries management is today more a multidimensional approach that has to balance human requirements against protection of the environment and biodiversity (Cowx 2000). Modern conservation challenges for fisheries management encompass all aquatic resources within the whole ecosystem, but also the fishery per se. One of the major challenges is to make sound management decisions to ensure viable commercial and recreational fisheries are compatible with aesthetic and nature conservation values in the twenty-first century (Arlinghaus et al. 2002). However, this requires harmonisation of philosophical views of rather biocentric (e.g. environmentalists) and anthropocentric (e.g. inland fishermen) oriented stakeholders, which resembles a socio-cultural and political issue. Consequently, a strategy to resolve the conflicts between conservation and fisheries protagonists must be to apply the stakeholder approach to decision making. The key to success involves building up relationships and sharing in the decision-making process based on sound science or factual evidence (Fig. 28.4). It will also include: (1) expanding the manager’s view of

Effective resource utilisation Relationships Trust/Values Networks Institutions

Figure 28.4 Conceptual model of the elements required for successful resource allocation comprising substantive knowledge, effective and legitimate process and relationships based on trust and common values all working towards a common goal (adapted from Meffe 2002)

28last 11/02/2003 15:22 Page 371

Interactions between fisheries and fish-eating birds

371

who is substantially affected by fish and wildlife management (stakeholder); (2) identifying and understanding stakeholder views; (3) seeking compromise between competing and conflicting demands when appropriate; and (4) improving communication between stakeholders. Ultimately, due to the expanded notion of values such as responsibility (for the fisheries resources), fairness, justice, and long-term concern for the sustainability of resources, the stakeholder approach forces fisheries managers and conservationists to consider ethical questions in decision making, which can only be to the benefit of all parties concerned (Decker & Enck 1996; Williams 1997).

References Arlinghaus R., Mehner T. & Cowx I.G. (2002) Reconciling traditional inland fisheries management and sustainability in industrialised countries, with emphasis on Europe. Fish and Fisheries 3, 261–316. Blaber S.J.M., Milton D.A., Smith G.C. & Farmer M.J. (1995) Trawl discards in the diets of tropical seabirds of the northern Great-Barrier-Reef, Australia. Marine Ecology – Progress Series 127, 1–13. Britton J.R., Harvey J.P., Cowx I.G., Holden T., Feltham M.J., Wilson B.R. & Davies J.M. (2002) Compensatory responses of fish populations in a shallow eutrophic lake to heavy depredation pressure by cormorants and the implications for management. In I.G. Cowx (ed.) Management and Ecology of Lake and Reservoir Fisheries. Oxford: Fishing News Books, Blackwell Science, pp. 170–183. Brothers N. (1991) Albatross mortality and associated bait loss in the Japanese longline fishery in the Southern Ocean. Biological Conservation 55, 255–268. Carscadden J.E., Frank K.T. & Leggett W.C. (2001) Ecosystem changes and the effects on capelin (Mallotus villosus), a major forage species. Canadian Journal of Fisheries and Aquatic Sciences 58, 73–85. Carss D.N., Bevan R.M., Bonetti A., Cherubini G., Davies J.M., Doherty D., El Hili A., Feltham M.J., Grade N., Granadiero J.P., Gremillet D., Gromadzka J., Harari Y.N.R.A., Holden T., Keller T., Lariccia G., Mantovani R., McCarthy T.M., Mellin M., Menke T., Mirowska-Ibron I., Muller W., Musil P., Nazarides T., Suter W., Trautmansdorff J.F.G., Volponi S. & Wilson B.R. (1997) Techniques for assessing cormorant diet and food intake: Towards a consensus view. Supplemento Richerche di Biologia della Selvaggina 26, 197–230. Cowx I.G. (2000) Management and Ecology of River Fisheries. Oxford: Fishing News Books, Blackwell Science, 444 pp. Cowx I.G. (2002a) Analysis of threats to freshwater fish conservation: Past and present challenges. In M.J. Collares-Pereira, I.G. Cowx & M.M. Coelho (eds) Conservation of Freshwater Fish: Options of the Future. Oxford: Fishing News Books, Blackwell Science, pp. 201–220. Cowx I.G. (2002b) Recreational fishing. In P.J.B. Hart & J.S. Reynolds (eds) Handbook of Fish Biology and Fisheries Volume II. Oxford: Fishing News Books, Blackwell Science, pp. 367–390. Cowx I.G. & Welcomme R.L. (1998) Rehabilitation of Rivers for Fish. Oxford: Fishing News Books, Blackwell Science, 260 pp. Decker D.J. & Enck J.W. (1996) Human dimensions of wildlife management: knowledge for agency survival in the 21st century. Human Dimension of Wildlife 1, 60–71. Frank K.T., Carscadden J.E. & Simon J.E. (1996) Recent excursions of capelin (Mallotus villosus) to the Scotian shelf and Flemish cap during anomalous hydrographic conditions. Canadian Journal Of Fisheries And Aquatic Sciences 53, 1473–1486. Garthe S. (1997) Influence of hydrography, fishing activity, and colony location on summer seabird distribution in the south-eastern North Sea. ICES Journal Of Marine Science 54, 566–577.

28last 11/02/2003 16:18 Page 372

372

Management

José P.V. (2000) Urgent Conservation Action for the Bittern Botaurus stellaris in the United Kingdom. Overall Project Report 1 July 1996–31 March 2000. Sandy: Royal Society for the Protection of Birds, 61 pp. McKay H.V., Furness R.W., Russell I.C., Parrott D., Rehfisch M.M., Watola G., Packer J., Armitage M., Gill E. & Robertson P. (1999) The Assessment of the Effectiveness of Management Measures to Control Damage by Fish-Eating Birds to Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0107. London: MAAF, 254 pp. Meffe G.K. (2002) Connecting science to management and policy in freshwater fish conservation. In M.J. Collares-Pereira, I.G. Cowx & M.M Coelho (eds). Conservation of Freshwater Fish: Options for the Future. Oxford: Fishing News Books, Blackwell Science, pp. 363–372. Montevecchi W.A. & Myers R.A. (1996) Dietary changes of seabirds indicate shifts in pelagic food webs. Sarsia 80, 313–322. Montevecchi W.A. & Myers R.A. (1997) Centurial and decadal oceanographic influences on changes in northern gannet populations and diets in the north-west Atlantic: Implications for climate change. ICES Journal of Marine Science 54, 608–614. Newbery P., Schäffer N. & Smith K.W. (1997) European Union Bittern Botaurus stellaris Action Plan. Brussels: RSPB, Birdlife International, European Commission, 34 pp. Peirson G., Tingley D., Spurgeon J. & Radford A. (2001) Economic evaluation of inland fisheries in England and Wales. Fisheries Management and Ecology 8, 415–424. Piatt J.F. & Nettleship D.N. (1987) Incidental catch of marine birds and mammals in fishing nets off Newfoundland, Canada. Marine Pollution Bulletin 18 (6B), 344–349. Pilcher M.W. & Feltham M.J. (1997) An Assessment of Cormorant Predation on Stillwater Coarse Fish Populations in the Lea and Colne Valleys of the Thames Catchment. Environment Agency (Thames North-east Area) R&D Technical Report W101, 64 pp. Regehr H.M. & Montevecchi W.A. (1997) Interactive effects of food shortage and predation on breeding failure of black-legged kittiwakes: Indirect effects of fisheries activities and implications for indicator species. Marine Ecology-Progress Series 155, 249–260. Russell I.C., Dare P.J., Eaton D.R. & Armstrong J.A. (1996) Assessment of the Problem of FishEating Birds in Inland Fisheries in England and Wales. Report to the Ministry of Agriculture, Fisheries and Food, project VC 0104. London: MAAF, 130 pp. Staub E., Egloff K., Kramer A. & Walter J. (1998) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: Two case studies from Swiss rivers. Comment. Journal of Applied Ecology 35, 607–610. Stenhouse I.J. & Montevecchi W.A. (1999) Indirect effects of the availability of capelin and fishery discards: Gull predation on breeding storm-petrels. Marine Ecology – Progress Series 84, 303–307. Suter W. (1995) The effect of predation by wintering cormorants Phalacrocorax carbo on grayling Thymallus thymallus and trout (Salmonidae) populations: Two case studies from Swiss rivers. Journal of Applied Ecology 32, 29–46. Tasker M.L., Camphuysen C.J., Cooper J., Garthe S., Montevecchi W.A. & Blaber S.J.M. (2000) The impacts of fishing on marine birds. ICES Journal of Marine Science 57, 531–547. Welcomme R.L. (2001) Inland Fisheries: Ecology and Management. Oxford: Fishing News Books, Blackwell Science, 358 pp. Williams C.D. (1997) Sustainable fisheries: economics, ecology, and ethics. Fisheries 22(2), 6–11.

28last 11/02/2003 15:22 Page 373

Subject index

Aerial surveys 196, 200, 202, 205, 207 Aquaculture 187, 197, 208, 222, 228, 236, 319 Artificial reefs 267–269 Baltic Sea 72 Behaviour 222, 223, 230, 240, 244, 259, 260 Bioenergetics 263 Biomanipulation 218, 219 Biomass 3, 5, 8, 18, 19, 52, 53, 56, 57, 61, 85, 141, 157, 180, 298, 299, 302, 307, 332 Bycatch 101, 309, 356–358,366 Carrying capacity 6–12, 85, 109, 110, 132, 133, 162, 165, 304 Catch and release 15, 24 Catch statistics 90, 96 Commercial fisheries 52, 56, 72, 84, 85, 89, 96, 99, 139, 140, 183, 187, 192, 196, 298, 304, 308, 336, 367, 370 Compensatory mechanism 35, 38 Competition 3, 9, 10, 12, 62, 82, 139, 148, 165, 175, 188, 190, 192, 266, 311, 328 Conflict resolution 361 Conservation 9, 151–153, 162, 208, 238, 309, 335, 336, 343, 354, 359–362, 368, 370 Counts 125, 340, 347, 348 Culling 7, 10, 43, 197, 330, 342, 343, 359. 368

FADs, see Fish attraction devices Fecundity 246, 247, 250–253, 302 Feeding behaviour 29, 30, 38, 174, 178, 261, 340 Fish attraction devices 267–272 Fishery management 151, 197, 259, 345, 366 Fishery performance 28–31, 38, 364 Fishing effort 8, 73, 74, 81, 212, 289, 290, 359 Fishing mortality 81, 89, 91, 100, 101, 109, 207, 254, 308, 309 Flock feeding 261 Food consumption 62, 165, 180 Foraging behaviour 260–263 270, 278, 279, 352 Gippsland lakes 196–208 Growth rate 5, 17, 31, 35, 36, 40, 88 Habitat management 298, 299, 311 Habitat preferences 227, 279 Haweswater 335–343 Holme Pierrepont 14–26, 32–40 Human disturbance 343, 365 Key factor analysis 14–21

Daily food intake 3, 14, 32, 49, 74, 81, 124, 178, 182, 183, 264 Depredation 14, 15, 17, 24, 28, 32, 35, 38, 178, 180, 187, 219, 325, 363–369 Diet 3, 4, 56, 122, 126, 131, 174, 187, 189, 263–265, 298–309, 314, 330 Disease 222, 234, 238, 311 Dobczyce reservoir 211–219

Lake Grand-lieu 187–193 Lake IJseelmeer 4–11, 30, 84–110, 261 Lake Neusiedl 139–148 Lake Nokoué 165–175 Lake Ontario 325–333 Lake Peipsi Pihkva 355 Lake Victoria 244–254 Legislation 14, 120, 188, 295, 335, 359, 365, 368 Leighton Moss 153–163, 298–311 Licensing 288 Loch Leven 288–296, 345–352

Economic loss 13, 187, 290, 364 Economic value 51, 52 Eel passes 298, 299, 305–308 Egg oiling 325, 329–333 Electric fishing 17, 25, 34, 141, 143, 157, 189, 299, 300 Eutrophication 30, 62, 148, 187, 244 Exploitation 3, 6, 10, 14, 15, 21, 24

Management 4, 6, 109, 110 Maturity 90, 109, 246–249, 253 Mesh size Migration 198, 298, 299, 302, 306 Minsmere 155–163, 298–311 Mortality 4–7, 10, 11, 17, 18, 28, 31, 37, 40, 62, 66, 84, 88–91, 96, 100, 130–132, 145, 148, 178, 22, 229, 282, 285, 328–330, 367, 368

28last 11/02/2003 15:22 Page 374

374

Subject index

Natural mortality 101, 207, 302, 307 Nature reserve 188 Nest removal 325, 328–331 Netting 316–320 Oder Valley 178–184 Optimal foraging 38, 163 Overexploitation 359, 361, 369 Overfishing 51, 62, 72, 165, 244 Parasite transmission 222, 223, 231, 235 Parasites 221–238, 311 Pellet analysis 53, 54, 65, 74, 168, 178, 206, 315, 328, 329 Pest control 288 Pollution 152, 236, 302 Precautionary principle 340, 343 Predation 3, 9, 11, 49, 62, 109, 120, 129–133, 148, 196, 197, 204, 229–231, 259–262, 266, 278, 282, 290, 295, 296, 311–315, 364 Predation pressure 85, 90, 175, 178, 184, 211, 219, 221–224, 231, 236, 244, 252, 317, 364, 367–370 Predation risk 226–228, 233, 236 Predator control 345 Predator–prey 6, 40, 188, 238, 271 Prey switching 10 Production 3, 5, 43–45, 84–88, 91, 96, 99, 102, 107, 139, 145, 147 Protected areas 354 Public consultation 329–330 Put-and-take fisheries 164, 165 Radio tracking 348 Recreational fisheries 196, 278, 280, 289, 325, 336, 367–370 Recruitment 18, 24, 25, 70, 102, 152, 153, 163, 187, 193, 207, 298, 302, 304, 306, 309, 310, 337, 340, 363, 369 Reed beds 139–148, 151–164 Reed cutting 152, 155, 156, 163, 298, 311

Refuges 30, 38, 259–272, 278–286, 343, 365, 370 Regurgitates 189, 196, 198 Rehabilitation 152, 153, 156, 163, 369 Risk 354–359 River Dee 130 River Elbe 65–70 River Havel 65–69 River Hodder 119–133 River Trent 29, 32 River Wye 119–133, 262 Roost counts 196, 202, 207 Rye Meads 31, 37, 40, 279 Scaring 12, 260, 342, 343, 361, 364, 367 Shooting 288–296, 320, 343–346, 352, 362, 365, 367 Sport fisheries 228, 229, 236, 269 Standing crop 19, 31, 35, 38, 49, 119, 129, 130, 132, 340, 365 Stock enhancement 153, 196, 208 Stocking 65–70, 281–284, 289, 290, 293, 298, 299, 305, 308, 310, 311, 343, 365, 369, 370 Stomach flushing 203 Telemetry 329–33, 345–350 Trophic dynamics 366, 367 Turbidity 29, 30, 81 Väinameri 72–81 Vantage point observations 143 Video surveillance 281, 284 Vistula lagoon 51–62 Water level management 152, 153, 162, 163, 336, 337 Water quality 29–33, 37, 85, 298 Wounding 278–285, 293–296, 364 Year class strength 17, 18, 85, 109 Yield 3, 5, 6, 10, 65, 70, 148, 178, 180, 183, 188, 193, 208

28last 11/02/2003 15:22 Page 375

Species index

Abramis brama 14, 18–22, 24, 29, 32, 35, 36, 38, 52, 55, 189, 212, 264 Acanthopagrus burcheri 196–208 Acanthopagrus butcheri 196–208 Alburnus alburnus 32, 144, 230 Alcedo spp 227, 228, 285 Aldrichetta forsteri 202, 203, 204 Alewife, see Alosa pseudoharengus Alosa pseudoharengus 327, 332 Anguilla anguilla 32, 52, 65–70, 72, 85, 90, 101, 119, 126, 127, 144–153, 157–160, 163, 180, 187, 190, 298–311, 336, 357 Anguilla australis 204 Anguillicola crassus 70 Arctic charr, see Salvelinus alpinus Ardea cinerea 32, 139, 143, 147, 148, 211–219, 227, 285 Ardea herodias 316–319 Ardea purpurea 139, 143, 147, 148 Arripis georgiana 202, 203 Asian topmouth gudgeon, see Pseudorasbora parva Atherina microstoma 202, 203 Atlantic salmon, see Salmo salar Barbatula barbatula 119, 126–28 Belted kingfisher, see Ceryle alcyon Bittern, see Botaurus stellaris Black bream, see Acanthopagrus butcheri Blackbass, see Micropterus salmoides Blackheaded gull, see Larus ridibundus Bleak, see Alburnus alburnus Blicca bjoerkna 74, 144, 180, 181, 189 Botaurus stellaris 151–164, 298–311, 355, 368 Brown trout, see Salmo trutta Bullheads, see Cottus gobio Burbot, see Lota lota Butroides striatus virescens 226 Campostoma anomalum 316, 317 Carassius auratus 180, 181 Carassius carassius 144 180, 181 Casmerodius albus 139, 143–148 Ceryle alcyon 316–319 Ceryle rudis 165–175, 244, 245, 253 Chlidonias niger 96, 98, 212, 214 Chrysichtys auratus 166, 170–174 Chrysichtys nigrodigitatus 166, 170–172 Chub, see Leuciscus cephalus Clarias sp. 169, 174

Clupea harengus 51, 52, 55–57, 74 Common bream, see Abramis brama Common carp, see Cyprinus carpio Coregonus albula 343 Coregonus lavaretus 335–343, 357 Cormorant 3, 12, 14–26, 28–40, 43–50, 51–62, 65–82, 85, 96, 98, 107, 109, 187–193, 229, 245, 253, 259–271, 278–286, 288–296, 320, 335–343, 345–352, 359, 361, 362, 368–370 Cottus gobio 119, 126–128 Cryptocotyle lingua 237 Cyprinus carpio 31, 144, 148, 180, 207, 208, 280 Dace, see Leuciscus leuciscus Diplostomum phoxini 229 Diplostomum spathaceum 228, 229 Eel, see Anguilla anguilla Engraulis australis 202, 203 Esox lucius 39, 51, 54, 72, 77, 81, 144, 146, 148, 152, 153, 180, 181, 190, 191, 193, 233, 266, 300, 357, 359 Ethmalosa fimbriata 165, 166–174 Euhaplorchis californiensis 233 Fundulus parvipinnis 233 Galaxis maculatus 229 Gasterosteus aculeatus 128, 152, 157–159, 226, 227, 231, 232, 237, 336 Gavia spp 228 Girella tricuspidata 202, 203, 204 Glugea anomala 233 Gobiomorphus cotidianus 229 Goosanders 29, 119–133, 266 Great blue heron, see Ardea herodias Grey heron, see Ardea cinerea Gymnocephalus cernuus 51, 55, 56, 61, 84, 84, 90–96, 101, 109, 178–181, 184 Hemichromis fasciatus 165–175 Hyporhamphus picarti 165, 169–175 Hyporhamphus regularis 202, 203, 204 Ictalurus catus 269 Ictalurus melas 190, 191 Kingfishers, see Alcedo sp Kribia sp 165, 169–175

28last 11/02/2003 15:22 Page 376

376

Species index

Largemouth bass, see Micropterus salmoides Larus fuscus 245 Larus ridibundus 85, 96, 98, 212, 214 Lates calcifer 108 Lates niloticus 108, 165, 244, 252 Lepomis gibbosus 141, 144 Lepomis macrochirus 268, 269 Leucaspius delineatus 190 Leuciscus cephalus 15, 29, 32, 128, 285 Leuciscus idus 74 Leuciscus leuciscus 15, 29, 32, 74, 230 Ligula intestinalis 229, 230, 233, 237, 244–254 Lota lota 72, 74, 76, 81, 180 Luxilis chrysocephalus 316, 317 Mallotus villosus 367 Mergus mergansus 96, 19–133, 213, 214, 362 Mergus sp. 234 Micropterus dolomieu 325–332 Micropterus punctulatum hensallis 268 Micropterus salmoides 269 Minnow, see Phoxinus phoxinus Myxobolus cerebralis 223 Notemigonus crysoleucus 266 Notropis cornutus 230 Nycticorax caldonicus 234 Oncorhynchus kisutch 120 Oncorhynchus mykiss 264, 289–295 Ornithdiplostomum ptychocheilus 229 Osmerus eperlanus 7, 72, 74, 77, 84–91, 85–109, 279 Pelecanus philippensis Pellonula leonensis 166,170–172 Pellonula vorax 166, 170–172 Perca flavescens 109, 269, 327, 332 Perca fluviatilis 6, 14, 19–22, 24, 32, 38, 39, 51, 52, 55–57, 61, 72, 74, 77, 81, 84–86, 89, 93, 95, 96, 101–03, 110, 144, 152, 157–165, 178, 184, 192, 212, 264, 266, 279, 289, 292, 293, 300, 336–338, 357, 359 Perch, see Perca fluviatilis Phalacrocorax aristotelis 263 Phalacrocorax auritus 262, 325–333 Phalacrocorax carbo 14, 28, 43, 85, 120, 187, 211–219, 259, 262, 278, 314, 335, 340, 345, 362 Phalacrocorax carbo carboides 196–208 Phalacrocorax carbo sinensis 14, 15, 55, 72, 174, 206, 260, 262, 263, 340, 362 Phalacrocorax fuscescens 197, 201 Phalacrocorax melanoleucos 197, 201, 205 Phalacrocorax sulcirostris 197, 201, 205 Phalacrocorax varius 197, 201, 205 Philypnodon grandiceps 202, 203 Phoxinus phoxinus 119, 126–128, 230, 237, 336 Pied kingfisher, see Ceryle rudis

Pied kingfisher, see Ceryle rudis Pike, see Esox lucius Pikeperch, see Stizostedion lucioperca Pimephales notatus 316, 317 Pimephales promelas 229 Platalea lencorodia 139, 143, 146 Podiceps cristatus 211–216, 219, 266 Pomatomus saltatrix 202, 203 Procambarus clarkii Pseudaphritis urvillii 202, 203 Pseudocaranx wrighti 202, 203 Pseudorasbora parva 141, 144 Pumpkinseed sunfish, see Lepomis gibbosus Pungitius pungitius 152, 231, 300 Purple heron, see Ardea purpurea Rainbow trout, see Oncorhynchus mykiss Rastrineobola argentea 165, 244–254 Roach, see Rutilus rutilus Rudd, see Scardinius erythrophthalmus Ruffe, see Gymnocephalus cernuus Rutilus rutilus 14, 15, 19–25, 29, 32, 35–38, 51, 52, 55–57, 61, 74, 77, 81, 144, 152, 159–162, 178, 180, 181, 184, 189, 190, 212, 233, 264, 266, 279, 351 Salmo salar 15, 119–121, 127, 128, 132, 133, 279, 288, 356 Salmo trutta 15, 119–122, 127, 128, 132, 264, 289–296, 336–338, 357, 367 Salvelinus alpinus 336–338, 343 Sarotherodon melanotheron 165–175 Scardinius erythrophthalmus 32, 144, 151–153, 157–163, 180, 189, 190, 254, 300 Schistocephalus solidus 226, 227, 231, 232, 234 Silver bream, see Blicca bjoerkna Smallmouth bass, see Micropterus dolomieu Smelt, see Osmerus eperlanus Spoonbill, see Platalea lencorodia Stizostedion lucioperca 6, 52, 55–57, 72, 77, 81, 85, 86, 89, 96, 101–103, 109, 110, 145, 148, 180, 190, 193, 357, 359 Sula bassana 367 Tench, see Tinca tinca Tetrabothrius sp 234 Three-spined stickleback, see Gasterosteus aculeatus Thymallus thymallus 315 Tilapia guineenis 166, 169–172 Tinca tinca 32, 54, 152, 159–161, 190–193 Tringa melanoleuca 234 Tylodelphis clavata 227 White egret, see Casmerodius albus Whitefish, see Coregonus laveratus Yongeichtys thomasi 169–175 Zoares viviparus 77, 81