Double-Crested Cormorant Feeding in Multiple Lake Environments: Intrinsic Markers Reveal Several Prey Sources and Frequent Site Switching by Breeding Birds

Home , Darter, Great cormorant, Neotropic cormorant, Sulidae

DOUBLE-CRESTED CORMORANT FEEDING IN MULTIPLE LAKE ENVIRONMENTS: INTRINSIC MARKERS REVEAL SEVERAL PREY SOURCES AND FREQUENT SITE SWITCHING BY BREEDING BIRDS

A Thesis

Submitted to the Faculty of Graduate Studies and Research

In Partial Fulfillment of the Requirements

for the Degree of

Master of Science

in Biology

University of Regina

Aleksandra Bugajski

Regina, Saskatchewan

December, 2011

UNIVERSITY OF REGINA

FACULTY OF GRADUATE STUDIES AND RESEARCH

SUPERVISORY AND EXAMINING COMMITTEE

Aleksandra Bugajski, candidate for the degree of Master of Science in Biology, has presented a thesis titled, Double-Crested Cormorant Feeding in Multiple Lake Environments: Intrinsic Markers Reveal Several Prey Sources and Frequent Site Switching by Breeding Birds, in an oral examination held on November 30, 2011. The following committee members have found the thesis acceptable in form and content, and that the candidate demonstrated satisfactory knowledge of the subject material.

External Examiner: Dr. Scott D. Wilson, Environment Canada

Supervisor: Dr. Christopher Somers, Department of Biology

Committee Member: Dr. Britt Hall, Department of Biology

Committee Member: Dr. Bjoern Wissel, Department of Biology

Chair of Defense: Dr. Scott Murphy, Department of Chemistry/Biochemistry

*Not present at defense

ABSTRACT

Conflicts between piscivorous cormorants (Phalacrocorax spp.) and humans over fisheries resources occur worldwide. To determine the level of cormorant impacts on fisheries, fish biomass removals are estimated but typically assume only one source of prey near the roost or breeding colony. Cormorants can fly long distances (>30 km) to forage, possibly resulting in fish removal being spread out over several areas within large lakes, or among other bodies of water. I examined the diet and feeding locations of double-crested cormorants (Phalacrocorax auritus; hereafter cormorants) breeding in a multiple lake environment in north-central Saskatchewan, Canada. A majority of their diet was composed of non-sport and non-commercial fish species. Yellow perch (36-196 mm) were a common prey item for cormorants, making up 30-55% of fish biomass consumed, and were therefore used as a model species to determine sources of cormorant prey. Comparison of carbon (δ13C) and nitrogen (δ15N) stable isotopes values in yellow perch collected by cormorants and those from known locations revealed several prey sources (different lakes and areas within lakes) and frequent, large-scale switching of feeding locations on a daily and seasonal basis. These findings were also substantiated by cormorant surveys throughout the study area using transect counts, and examination of flight directions to and from a major breeding colony. Prey from areas well-removed from the breeding colony lake (up to 30 km away) were an important part of cormorant diet, representing 70% of fish fed to nestlings in 2010 during the early chick rearing stage. Cormorants began to feed closer to the breeding colony during the late chick rearing stage. Linear discriminant analysis revealed classification accuracy of known location yellow perch to range from 69% to 86%, suggesting that stable isotopes of

i carbon and nitrogen performed well as intrinsic markers of fish source in my study area.

My research clearly shows that cormorant consumption of fish happens at a variety of locations, negating the value of the traditional approach of estimating biomass removal from the breeding colony lake as the guideline for making fisheries management decisions. Knowing where prey fish come from and estimating relative proportions taken from various sites will refine biomass removal estimates to help managers better understand potential interactions between cormorants and fisheries. In addition, my research shows that cormorants make decisions about foraging and feeding locations that are independent of breeding colony site selection; i.e., they often use sites well removed from the breeding colony. Factors that influence cormorant foraging locations need to be more thoroughly identified to advance our understanding of their ecology, and to aid fisheries management.

ACKNOWLEDGEMENTS

I would like to thank all of the people who offered me guidance, support and friendship through my graduate experience. Thanks to Dr. Christopher Somers, my graduate supervisor, for his priceless guidance and help throughout the last two years.

Through the discussions and communications with you, I expanded my knowledge of scientific research, but also broadened my view on the wisdom of life. I also want to give thanks to Dr. Bjoern Wissel and Dr. Britt Hall for their helpful advice and insights into my research. Special thanks to Mark Duffy from Saskatchewan Environment for providing helpful information and support during my fieldwork in La Ronge, and Dr.

Matt Reudink for guiding me through my statistical analyses.

Thanks to Saskatchewan Environment‟s Fish and Wildlife Development Fund, the

Natural Sciences and Engineering Research Council of Canada, and the Canada Research

Chairs Program for funding this research. Additional support from The Faculty of

Graduate Studies and Research, and the Fish and Wildlife Development Fund Student

Research Award was appreciated.

Many thanks to my field assistants, Patrick Barks, Nolan Hoggarth, and Alison

White, for all of their hard work. Thanks to Leanne Heisler for lending her GIS skills and

Jennifer Doucette for all of her extra guidance and useful suggestions. I am also grateful to Scott Goodwill for his support and words of encouragement through the difficulties of my graduate work.

iii

POST DEFENSE ACKNOWLEDGEMENTS

Thanks to Dr. Scott Wilson of the Canadian Wildlife Service for taking on the important role as external examiner for my M.Sc. thesis defense, and providing many thoughtful questions. I would also like to thank Dr. Scott Murphy for chairing my defense.

DEDICATION

I would like to dedicate this thesis to my mother and father, who introduced me to the great outdoors and encouraged me to pursue my love for nature. Without you, I would not have achieved as much as I have.

TABLE OF CONTENTS

ABSTRACT ...... i

ACKNOWLEDGEMENTS ...... iii

POST DEFENSE ACKNOWLEDGEMENTS...... iv

DEDICATION ...... v

TABLE OF CONTENTS ...... vi

LIST OF TABLES ...... viii

LIST OF FIGURES ...... ix

LIST OF APPENDICES ...... xi

1. INTRODUCTION ...... 1

1.1 Conflicts Between Birds and Fisheries ...... 1

1.2 Cormorants and Fisheries Management...... 2

1.3 Cormorant Foraging Habits ...... 12

1.4 Using Stable Isotopes to Track Sources of Cormorant Prey...... 16

1.5 Cormorant Populations in Saskatchewan, Canada...... 20

1.6 Purpose of Current Research ...... 22

2. METHODS ...... 23

2.1 Study Species ...... 23

2.2 Study Area ...... 24

2.3 Nestling Diet Composition ...... 29

2.4 Stable Isotope Sample Collections and Analyses ...... 30

2.5 Relative Cormorant Density and Habitat Measurements ...... 33

2.6 Cormorant Flight Paths ...... 35

2.7 Statistical Analyses ...... 36

3. RESULTS ...... 38

3.1 Nestling Diet Composition ...... 38

3.2 Stable Isotope Variability and Sources of Prey Items ...... 41

3.3 Relative Cormorant Density and Habitat Measurements ...... 57

3.4 Cormorant Flight Paths ...... 64

4. DISCUSSION ...... 66

4.1 Multiple Cormorant Foraging Locations ...... 66

4.2 Utility of Stable Isotopes to Trace Sources of Cormorant Prey...... 73

4.3 Cormorant-Fisheries Management and Future Research ...... 81

5. CONCLUSIONS...... 85

LIST OF REFERENCES ...... 87

APPENDIX A ...... 105

APPENDIX B ...... 106

APPENDIX C ...... 107

APPENDIX D ...... 108

APPENDIX E ...... 109

APPENDIX F...... 110

APPENDIX G ...... 111

vii

LIST OF TABLES

Table Page

Table 1. Average and standard deviations of δ13C and δ15N values for gill-netted yellow perch (TL = 70-130 mm) from all sites sampled in both years. Western Nemeiben Lake, Head Lake, and south Lac La Ronge were not sampled in 2009 ...... 44

Table 2. Cross-assignment matrix for gill-netted yellow perch (TL = 80-130 mm) from known locations in 2009 using LDA model 1 (two Nemeiben populations). Values on the diagonal represent the number (and proportion) of correctly assigned individuals. Lac La Ronge = LLR ...... 48

Table 3. Cross-assignment matrix for gill-netted yellow perch (TL = 70-120 mm) from known locations in 2010 using LDA model 1 (one Lac La Ronge population, three Nemeiben Lake populations). Values on the diagonal represent the number (proportion) of correctly assigned individuals ...... 50

Table 4. Variable loadings for environmental measurements on all factor components. Cormorant density (*) was used as a supplementary variable in the analysis ...... 63

Table 5. Proportion (%) and direction of inbound and outbound cormorant flights from Colony B on Egg Lake in 2010. Directions are divided by northeast, northwest, southeast, and southwest ...... 65

Table 6. Cross-assignment matrix for gill-netted yellow perch (TL = 80-130 mm) from known locations in 2009 using LDA model 2 (one Nemeiben population) and model 3 (Nemeiben Lake excluded). Values on the diagonal represent the number (and proportion) of correctly assigned individuals. LLR=Lac La Ronge ...... 105

Table 7. Cross-assignment matrix for gill-netted yellow perch (TL = 70-120 mm) from known locations in 2010 using LDA model 2 (central Nemeiben Lake excluded), model 3 (second round gill-net sampling from Egg and Bigstone Lakes), and model 4 (Bigstone Lake excluded). Values on the diagonal represent the number (and proportion) of correctly assigned individuals ...... 106

Table 8. The proportion (%) of yellow perch (TL=80-130 mm) with assignment probabilities ≥ 50% assigned to each site by day and overall in 2009 for model 2 (one Nemeiben population) and model 3 (Nemeiben Lake excluded). The largest percent proportions for each date are shaded ...... 107

viii

LIST OF FIGURES

Figure Page

Figure 1. Map of Saskatchewan, Canada (inset = North America) depicting the four major ecozones. The study area near Lac La Ronge, situated on the transition between the Boreal Plain and Boreal Shield Ecozones, is circled ...... 25

Figure 2. Detailed map of the study area showing the multiple large and small lakes that represent potential foraging locations for cormorants breeding on Egg Lake. The locations of known cormorant breeding colonies are marked with circles. Egg and Bigstone Lakes and south Lac La Ronge are situated in the Boreal Plain Ecozone, whereas Nemeiben Lake and north Lac La Ronge are situated in the Boreal Shield Ecozone. Central Lac La Ronge is the transition zone between the north and south portions of the lake. The arrows represent cormorant colonies A (~4000 breeding pairs) and B (~800 breeding pairs) which were studied in 2009 and 2010, respectively. Anecdotal observations in previous years suggested that cormorants forage extensively on the labelled lakes ...... 26

Figure 3. Locations of gill-nets (triangles) used to sample fish from known sites in the La Ronge study area. Fish from known locations were the basis for identifying the provenance of fish consumed in unknown locations by cormorants. Net sites were chosen to reflect anecdotal observations regarding important foraging locations for birds breeding on Egg Lake. Line transects used to survey for foraging cormorants in the same areas are also shown (see section 2.5 below). Nemeiben Lake and Lac La Ronge were each split into three areas due to habitat heterogeneity ...... 32

Figure 4. Diet composition of cormorant nestlings on Egg Lake in 2009 (Colony A) and 2010 (Colony B). The percent (a) abundance, (b) frequency of occurrence, and (c) biomass of each fish species observed in the collected boluses. The category „Other‟ contained trout perch, brook stickleback, Iowa darter, and crayfish ...... 39

Figure 5. Carbon and nitrogen stable isotopes values of individual yellow perch (TL = 70-130 mm) and average zooplankton values with standard error bars in (a) 2009 and (b) 2010. All fish were collected from suspected cormorant foraging locations using bottom- set gill nets, and represent „known‟ locations. In 2010 all Lac La Ronge zones were combined and Egg and Bigstone Lakes include yellow perch from a second round of gill- netting in August ...... 42

Figure 6. Carbon and nitrogen stable isotopes values of individual regurgitated yellow perch and proportions assigned to each foraging site by collection date: (a) June 18, 2009; (b) July 05, 2009. The approximate outline of stable isotopes values for yellow perch collected from known locations is shown for reference. Yellow perch were collected from a number of boluses from across the breeding colony, thereby reflecting the diet of a large number of birds ...... 53

Figure 7. (a) Carbon and nitrogen stable isotopes values of individual regurgitated yellow perch collected during early and late chick rearing stages (early = June 17 – July 15; late = July 16 – August 31). The approximate outline of stable isotopes values for yellow perch collected from known locations is shown for reference. (b) The proportion of regurgitated yellow perch assigned to each site by collection date ...... 54

Figure 8. Geometric mean cormorant densities (± SE) at each site in the La Ronge study area during the 2010 breeding season from June 08 – August 24. Data were obtained from line transect surveys. The * symbol indicates sites that are significantly different from Bigstone Lake. All other pair-wise differences were not significant individually ...58

Figure 9. Geometric mean (± SE) cormorant density at each site in the La Ronge study area during the early and late chick rearing periods in 2010. The early chick rearing period was from June 17 – July 15, and the late from July 16 – August 31. Data were obtained from line transect surveys. The * symbol indicates a significant interaction between location and chick rearing period ...... 59

Figure 10. Locations of cormorant foraging flocks (≥ 10 individuals) observed in the La Ronge area in 2010. Flocks were divided into early and late chick rearing period. Flock sizes ranged from 10 to 2500 individuals. Different sizes of circles represent the size of foraging flocks with large circles representing larger flocks (Large = 2500-1000, Large- medium = 1000-500, Medium = 500-150, Small-medium = 150-50, Small = 50-10) ...... 60

Figure 11. Principal components analysis for four environmental factors with cormorant density (DENSITY) as a supplemental variable. Transects from study sites were overlaid to illustrate how lake sites were associated with the environmental variables. BIG=Bigstone Lake, EGG=Egg Lake, NEM=Nemeiben Lake, NLLR=North Lac La Ronge, CLLR=Central Lac La Ronge, SLLR=South Lac La Ronge. Transect number is indicated after the lake code ...... 62

Figure 12. Carbon and nitrogen stable isotopes values of regurgitated yellow perch (TL = 70-120 mm) in 2010 for each day of bolus collections on Colony B ...... 108

Figure 13. Total mercury (Hg) concentrations in gill-netted yellow perch of approximately 100 mm total length from the La Ronge area in 2009 ...... 110

Figure 14. The age and total length of gill-netted yellow perch from the La Ronge area in 2009...... 111

LIST OF APPENDICES

Appendix Page

Appendix A. Cross-assignment matrix for gill-netted yellow perch (TL = 80-130 mm) from known locations in 2009 using LDA model 2 (one Nemeiben population) and model 3 (Nemeiben Lake excluded). Values on the diagonal represent the number (and proportion) of correctly assigned individuals. LLR=Lac La Ronge ...... 105

Appendix B. Cross-assignment matrix for gill-netted yellow perch (TL = 70-120 mm) from known locations in 2010 using LDA model 2 (central Nemeiben Lake excluded), model 3 (second round gill-net sampling from Egg and Bigstone Lakes), and model 4 (Bigstone Lake excluded). Values on the diagonal represent the number (and proportion) of correctly assigned individuals ...... 106

Appendix C. The proportion (%) of yellow perch (TL=80-130 mm) with assignment probabilities ≥ 50% assigned to each site by day and overall in 2009 for model 2 (one Nemeiben population) and model 3 (Nemeiben Lake excluded). The largest percent proportions for each date are shaded ...... 107

Appendix D. Carbon and nitrogen stable isotopes values of regurgitated yellow perch (TL = 70-120 mm) in 2010 for each day of bolus collections on Colony B ...... 108

Appendix E. The proportion (%) of yellow perch (TL=70-120 mm) with assignment probabilities ≥ 50% assigned to each site by day and overall in 2010 for model 1 (all sites), model 2 (central Nemeiben Lake removed), and model 3 (2nd round of gill-netting). The largest percent proportions for each date are shaded ...... 109

Appendix F. Total mercury (Hg) concentrations in gill-netted yellow perch of approximately 100 mm total length from the La Ronge area in 2009 ...... 110

Appendix G. The age and total length of gill-netted yellow perch from the La Ronge area in 2009 ...... 111

1. Introduction

1.1 Conflicts Between Birds and Fisheries

Conflicts between aquatic birds and fisheries are widespread and arise in two major ways: (1) human activities associated with fishing negatively impact bird populations; and (2) abundant piscivorous birds are perceived to compete with human harvesting of fish (Tasker et al. 2000). Industrial fishing has been shown to have negative impacts on marine birds directly through fishing gear entanglement (Murray et al. 1993,

Tasker et al. 2000, Watkins et al. 2008), and indirectly by reduction of prey availability

(Monaghan 1992, Frederiksen et al. 2004). Entanglement of marine birds in fishing lines causes direct mortality and has been shown to reduce the abundance of several bird species, contributing to their status of threatened or endangered (e.g., Murray et al. 1993,

Tasker et al. 2000, Watkins et al. 2008). Commercial fisheries have also negatively impacted marine bird species by reducing their prey base through overexploitation of fish populations, which can cause lower reproductive success as a consequence of food limitation (Monaghan 1992, Frederiksen et al. 2004). From the perspective of humans, abundant piscivorous birds are perceived to be major competitors for economically valuable fish (Draulans 1987, Feltham 1995, Tasker et al. 2000, Žydelis and Kontautas

2008). Piscivorous birds are frequently blamed for declining catch rates in commercial and recreational fisheries, and decisions to control or reduce bird populations to benefit fish species can lead to highly polarized management conflicts (Beamesderfer 2000).

The most widespread example of competition between humans and birds is with cormorant species (genus Phalacrocorax), which are widely viewed as damaging to coastal and inland fisheries resources (see Cowx 2003). These conspicuous colonial birds

1 have a long history of conflict with humans over fish in certain areas, and generate much stronger negative feelings in humans compared to other piscivorous birds (Duffy 1995).

Many studies have shown that cormorants mainly eat fish that are undesirable to humans, and have only negligible impacts on fish populations (see below). In addition, cormorants do not greatly exceed consumption rates compared to that of other piscivorous birds of their size (Madenjian and Gabrey 1995, Mous et al. 2003, Žydelis and Kontautas 2008).

Despite this, animosity towards cormorants remains high, possibly arising from socio- cultural issues related to the appearance of the birds (in addition to their fish-based diet), which has been described as diabolical (Hatch 1995). Studies on cormorant ecology encompassing diet, foraging behaviour, and fisheries impacts have been occurring for decades (e.g. Mattingley 1927, Bartholomew 1942); however, to date our understanding of cormorant ecology and their importance as a factor affecting the quality and production of fisheries remains incomplete.

1.2 Cormorants and Fisheries Management

The conflict between cormorants and humans over fisheries resources occurs on a global scale. Cormorant-fisheries conflicts exist on six of the seven continents: North

America (e.g. Milton et al. 1995, Seefelt and Gillingham 2006, Wires and Cuthbert 2006,

Dorr et al. 2010), Europe (e.g. Grémillet et al. 1995, Bearhop et al. 1999, Engström 2001,

Carss and Ekins 2002), South America (e.g. Punta et al. 1993, Quintana et al. 2004),

Africa (e.g. Linn and Campbell 1992), Asia (e.g. Takahashi et al. 2006), and Australia

(e.g. Trayler et al. 1989, Coutin and Reside 2003). Of the 39 species of cormorants worldwide, only a few are routinely implicated in fisheries management issues, such as

2 the great cormorant (Phalacrocorax carbo) in Europe, the pied cormorant

(Phalacrocorax varius) in Australia, and the double-crested cormorant (Phalacrocorax auritus) in North America (Nelson 2005). These species are abundant and widespread compared to the 11% of cormorant species that are at risk and have small endemic populations (Duffy 1992). Cormorants are being blamed for consuming too many fish and reducing the catch rate of economically valuable fish for human harvesting. The extent of the conflict depends on the measured impacts of cormorants on fisheries, which differ dramatically between areas. Studies have concluded that cormorants either have no impact or a negative impact on fisheries (reviewed by Doucette et al. in press), but the circumstances leading to economic loss of fish and conflict with humans remain uncertain and in need of further study.

Numerous studies on the population trends and status of cormorants in North

America and Europe report that during the 1960‟s, populations were nearly decimated due to human persecution and the use of organochlorine pesticides such as dichlorodiphenyl-trichloroethane (DDT) (Craven and Lev 1987, Weseloh et al. 1983,

Vermeer and Rankin 1984, Hobson et al. 1989, Carter et al. 1995, Ludwig et al. 1995,

Suter 1995, Kirby et al. 1996, Larson et al. 1996, Boertmann and Mosbech 1997, Wires et al. 2001). DDT caused reproductive failure in cormorants and other bird species by decreasing egg-shell thickness, which led to egg breakage during incubation and caused developmental abnormalities in chicks (Larson et al. 1996). The drastic decrease in the number of cormorants in the U.S.A. during this time led to the birds being added to The

Migratory Bird Treaty Act in 1972 where they are now protected under federal law

(Trapp et al. 1995). Similarly, cormorants in Britain became protected under the Wildlife

3 and Countryside Act in 1981 (Kirby et al. 1996, Parrott et al. 2003). The development of these documents appeared to play a factor in the expansion of cormorants due to less human harassment (Kirby et al. 1996, Wires et al. 2001), although illegal culling still occurs (Ewins and Weseloh 1994).

Cormorant populations have increased significantly in number and range since the

1970‟s (Craven and Lev 1987, Hobson et al. 1989, Hatch 1995, Krohn et al. 1995, Suter

1995, Weseloh et al. 1995, Frederiksen et al. 2001). DDT pesticide use was discontinued and cormorant populations began to increase as a result of higher recruitment rates

(Vermeer and Rankin 1984, Craven and Lev 1987, Weseloh and Ewins 1994, Hatch

1995, Wires et al. 2001). Other factors contributing to cormorant population growth included an increase in the abundance of available food through the creation of aquacultures ponds on over-wintering grounds (Glahn and Brugger 1995, Wires et al.

2001), fish stocking in reservoirs and lakes (Draulans 1988, Modde et al. 1996), and increasing availability of small-bodied fish due to the collapse of predatory fish populations caused by sport and commercial fishing (Gloutney and Chen 1992, Hobson et al. 1989, Post et al. 2002). Populations of cormorants have grown to impressive numbers in both Europe and North America. For example, the population of great cormorants breeding in northern Europe increased from about 5000 nesting pairs in 1970 to about 100,000 nesting pairs in the late 1990‟s (Frederiksen et al. 2001). Similarly, breeding populations of double-crested cormorants in the Great Lakes region of Canada grew from 89 to 25,842 nesting pairs between the years 1970-1991 (Weseloh et al. 1995).

Evidence shows that the rate of increase in certain areas is slowing down and populations are beginning to reach biological carrying capacity (Suter 1995, Ridgeway et al. 2006).

The abundance of cormorants is presently higher than most fishermen and wildlife managers would recall; however, evidence suggests that historical numbers were much higher (Wires and Cuthbert 2006).

As cormorants have increased in number, so has the concern over their impact on fisheries, especially given that these birds are extremely efficient aquatic predators.

Cormorants are described as being opportunistic, generalist feeders that consume mostly small benthic and pelagic fish species (Hobson et al. 1989, Johnson et al. 2002).

Cormorants are foot-propelled divers that forage for their prey in depths of ≤10 metres

(Custer and Bunck 1992, Hebshi 1998, Coleman et al. 2005, Enstipp et al. 2006a), although in some cases cormorants may dive up to 34 metres (Wanless et al. 1991). Dive durations among cormorant species differ but most are 20-40 seconds with larger species submerging for longer periods (Cooper 1985, Sapoznikow and Quintana 2003). Grémillet

(1997) determined the catch per unit effort of prey caught by great cormorants to be 9 to

15 g min-1, which is much higher than other diving piscivorous bird species such as the

Adélie penguin (Pygoscelis adeliae) (1.55 g min-1; Chappell et al. 1993) and the African penguin (Spheniscus demersus) (2.1 g min-1; Nagy et al. 1984). Cormorants are less buoyant compared to other aquatic birds, which helps them to submerge and pursue prey

(Lovvorn and Jones 1991). Partially wettable plumage facilitates the reduced buoyancy of cormorants; however, it also results in increased thermoregulatory costs during dives

(Ribak et al. 2005, Grémillet et al. 2005a, Enstipp et al. 2006b). To deal with cold water environments, aquatic mammals develop layers of insulating subcutaneous fat, while most seabirds have dense waterproof feathers (Grémillet et al. 2005a). Cormorants have neither of these characteristics, yet they are still found to forage in cold polar waters

(Grémillet et al. 2005b). Recent evidence suggests that cormorants do not take larger quantities of fish to compensate for high heat loss during dive bouts; rather, consumption rates are comparable to other aquatic birds of similar size (Grémillet et al. 1995,

Grémillet et al. 2003). Cormorants will feed solitarily but a majority of the time they forage in flocks, which may herd schools of fish to increase the efficiency of prey capture

(Bartholomew 1942, Van Eerden and Voslamber 1995). A flock of 5000 individual cormorants can cover up to 3 ha of surface water while foraging (Van Eerden and

Voslamber 1995). Mass fishing by highly efficient cormorants can potentially deplete prey abundance in an area relatively quickly, further contributing to perceived conflicts with fishermen.

Cormorant diet in relation to fisheries has been well studied and problems appear to arise mainly when these birds forage in heavily modified systems that humans use

(Doucette et al. in press). Diet studies on cormorants worldwide are numerous, so I have only considered studies from North America here. Concern mainly stems from the potential for cormorants to consume economically valuable fish, such as walleye

(Stizostedion vitreum; Rudstam et al. 2004) and smallmouth bass (Micropterus dolomieui;

Lantry et al. 2002). In some cases, cormorants have been shown to consume significant quantities of sport fish, ranging from 7-81% of their diet by biomass (Ottenbacher et al.

1994, Modde et al. 1996, VanDeValk et al. 2002, Rudstam et al. 2004). Although cormorants are capable of consuming large fish up to 40 cm in total length (Bur et al.

1999, Ottenbacher et al. 1994), most often they are reported to consume sub-adults (e.g. age 0-3) of sport fish species, thereby reducing future recruitment into fisheries (Modde et al. 1996, Burnett et al. 2002, VanDeValk et al. 2002, Rudstam et al. 2004). In these

6 situations, cormorants tend to be part of aquatic ecosystems that have been heavily modified by humans, including heavily stocked lakes with super-abundant sport fish

(e.g., Oneida Lake; VanDeValk et al. 2002, Rudstam et al. 2004), or highly altered food webs (e.g., Lake Ontario; Lantry et al. 2002). Aquaculture facilities located in the wintering range of cormorants in the southern United States also provide piscivorous birds with large amounts of accessible fish. Predation of channel catfish (Ictalurus punctatus) at aquaculture facilities can be high but quite variable, ranging from 0-97% biomass of catfish consumed in the 10-20 cm size range (Glahn and Brugger 1995, Glahn et al. 1995). Accurate damage assessments to fish farms are difficult because cormorants vary their catfish consumption significantly by location and time of year (Glahn et al.

1995, Stickley et al. 1992).

In systems with no stocked prey, or intact food webs, cormorants mainly consume fish that are of little to no economic value, thereby having negligible impacts on fisheries

(Robertson 1974, Craven and Lev 1987, Hobson et al. 1989, Blackwell et al. 1995, Bur et al. 1999, Neuman et al. 1997, Johnson et al. 2002). Benthic prey such as sculpins

(Cottidae) and suckers (Catostomidae), and pelagic species like herring (Clupeidae) and shiners (Cyprinidae), are consumed extensively by cormorants; economically valuable fish comprise less than 4% of diet biomass (Hobson et al. 1989, Bur et al. 1999, Johnson et al. 2002). However, there is some concern that cormorants may impact sport and commercial fish populations indirectly through changes to the fish community and competition for similar prey. Some studies have shown evidence of dietary overlap between cormorants and predatory fish, such as walleye, but the extent of food competition between the two species is still uncertain (Bur et al. 1999, Stapanian et al.

2002, Doucette et al. in press). Circumstances leading to significant economic loss of fish by cormorants likely differ by system (e.g. stocked lakes, aquacultures, natural lakes), making it difficult to generalize. In addition, cormorants can show spatial and temporal changes in their diet (Craven and Lev 1987, Blackwell et al. 1995, Bur et al. 1999,

Neuman et al. 1997), further complicating quantitative analyses.

Despite complications with diet and fisheries impact assessments, it is clear based on bioenergetics modeling that cormorants can consume large amounts of fish (Grémillet et al. 2003, Ridgeway 2010). A single double-crested cormorant weighing 2.0 kg is likely to consume 500 g fish•day -1, based on the assumption that the species consumes 25% of its body weight daily (Dunn 1975, Glahn and Brugger 1995). The amount of fish consumed by individuals differs slightly based on parameters such as species, body mass, assimilation efficiency of food, energy expenditure, and time of year (Glahn and Brugger

1995, Grémillet et al. 1995, Ridgeway 2010). For example, Grémillet et al. (1995) reported that fish consumption by breeding great cormorants increased from 238 g•bird-

1•day-1 to 588 g•bird-1•day-1 from incubation to rearing of fledglings. A variety of studies have used per capita consumption rates like those above to estimate the amount of fish that a cormorant population can consume during the breeding season (Johnson et al.

2002, VanDeValk et al. 2002, Rudstam et al. 2004, Diana et al. 2006, Ridgeway 2010).

These analyses also include the daily food intake requirements of nestlings. For double- crested cormorant nestlings, fish consumption has ranged from 327 g•bird-1•day-1 (Fowle

1997) to 434 g•bird-1•day-1 (Dunn 1975). Total fish consumption by cormorants depends on colony size and residence time; for example, a small colony of 730 adult double- crested cormorants and 111 fledged chicks was estimated by Rudstam et al. (2004) to

8 consume 47,300 kg (484 000 prey items) of fish, whereas Johnson et al. (2002) estimated that a much larger colony of 16,820 adults and 2.25 chicks per nest consumed 2.1 million kg (50 million prey items) of fish during one breeding season.

Cormorant consumption of fish is often considered problematic because large colonies can actually remove more fish biomass than sport or commercial fisheries. For example, in the eastern basin of Lake Ontario, cormorants took 76,000 kg and 174,400 kg of smallmouth bass and yellow perch, respectively (Johnson et al. 2002). In comparison, the sport fishery took 24,700 kg of smallmouth bass, and the commercial and sport fisheries combined took 35,100 kg of yellow perch (Johnson et al. 2002). However, modeling fish stock availability in Lake Ontario indicated that 28,000 cormorants consumed only 0.5% of available fish biomass, which was less than the 5% that predatory fish were estimated to consume (Weseloh et al. 1995). Although cormorants eat a large amount of fish, it is often difficult to compare consumption estimates to fish stock size available in lakes. Information on stock size is usually unavailable or unreliable

(Madenjian and Gabrey 1995, Johnson et al. 2002), which limits the usefulness of cormorant consumption data for making management decisions.

In addition, small inaccuracies in estimating per capita consumption by cormorants can lead to large discrepancies in estimates of colony-level consumption

(Ridgeway 2010). Large variance is introduced into calculations of daily food intake from variation in individual body mass, length of residence in an area, and fledging rate per nest (Hatch and Weseloh 1999). Ridgeway (2010) reviewed the various methods and estimates of daily food intake that are used in cormorant-fisheries studies and concluded that unless the study uses site-specific diet and energetics data, biologists and managers

9 must use a consistent approach to estimating cormorant fish consumption. One apparent problem with this suggestion is that the daily food intake and energy expenditure of cormorants can differ widely among areas. For example, cormorants may have different body masses, travel further to obtain prey, or forage in colder or more turbid waters, changing their daily energy budget. Also, cormorants often forage in multiple different habitats or bodies of water, so the fish they consume may be from multiple sources. If this is the case, the approach of estimating biomass consumed by cormorants solely from areas near breeding or roosting areas will be inaccurate for assessing impacts on fish populations. Few studies have considered sources of prey as a factor affecting their consumption models.

Concern over the state of commercial and sport fisheries has resulted in the establishment of cormorant population control programs across much of the northern hemisphere (Krohn et al. 1995, Kirby et al. 1996). Control programs include lethal control by shooting adults and suppression of reproduction by egg oiling, which asphyxiates developing embryos (Bédard et al. 1995). Management plans have been developed to control cormorant populations on large scales in both North America for the double-crested cormorant (U.S. Fish and Wildlife Service 2003) and Europe for the great cormorant (Carss 2003). Millions of dollars are spent on cormorant management annually in order to increase the quality of fisheries (Shwiff et al. 2009). Thousands of cormorants are killed at various locations each year. In France alone, about 10,000 cormorants were culled in 1998-99, not including illegal shooting (Frederiksen et al. 2001). However, there is little evidence of these programs being efficient in removing cormorants and improving fishing quality. Studies have reported decreases in the abundance of

10 cormorants at local colonies after the first few years of establishing a control program; however, results were inconclusive as to whether the program was successful in future reductions of cormorants and increased fishing quality (Bédard et al. 1995, Brian et al.

2010, Dorr et al. 2010). Even if culling is feasible for reducing cormorant populations on the short term, there are many economic and ethical problems with this method used for improving fishing quality for humans.

A variety of non-lethal deterrent methods have been implemented on local scales, particularly for fish farms and stocked lakes, to reduce cormorant predation pressure on fish. Methods used to deter cormorants from these areas include visual and acoustic signals like shooting (Parrott et al. 2003) and pyrotechnics (Brian et al. 2010), as well as mechanical barriers such as overhead netting (Moerbeek et al. 1987). However, cormorants have been reported to become accustomed to noises meant to frighten them and have changed their foraging behaviour if barriers were set up (Moerbeek et al. 1987).

Other deterring methods manipulate fish habitat and community. For example, Russell et al. (2008) tested the effectiveness of artificial fish refuges and concluded that cormorants experienced significantly lower foraging success in refuge ponds, where fish could escape predation, than in control ponds. Mott and Boyd (1995) suggested the use of buffer prey in aquaculture facilities where non-commercial fish are used to detract cormorants from valuable fish species to protect aquaculture stocks. These methods provide alternatives to the current lethal control methods of shooting and egg oiling.

However, before mitigation measures are taken, impact assessments should be made to determine if cormorants are a threat to the fishery.

1.3 Cormorant Foraging Habits

Colonial birds moving to and from a central location (breeding colony or roosting site) to forage should travel distances that provide the greatest net energy yield (Hamilton et al. 1967, Horn 1968, Evans 1982, Gorke and Brandl 1986, Ydenberg et al. 1986,

Brandl and Gorke 1988, Brown et al. 1992, Obst et al. 1995, Elliott et al. 2009). For central place foragers, energy costs consist of both transit time to a foraging area and search and handling time of prey (Elliott et al. 2009). Ashmole (1963) proposed that seabird communities in the tropics should forage within the immediate vicinity of breeding colonies to decrease energy expended for travel. When prey is depleted in the area immediately surrounding the colony, foraging birds will move farther away in an expanding ring around the colony. This phenomenon, referred to as Ashmole‟s Halo, has been extended to include colonial birds at higher latitudes. Elliott et al. (2009) concluded in their study that populations of thick-billed murres (Uria lomvia) in the Arctic were regulated by prey densities near the breeding colony, and that larger prey were depleted near the colony as the season progressed. Birt et al. (1987) found supporting evidence for the “halo” specifically with double-crested cormorants, whereby fish densities were significantly lower near colonies than outside the known foraging range. The general concept of Ashmole‟s Halo suggests that colonial birds should obtain their prey from one source near the colony; thus, bioenergetics models estimate cormorant consumption of fish by assuming one source of prey, which is generally the lake that they nest on.

Many factors modify the relationship between distance traveled and food acquired by piscivorous birds foraging from a central place (Gaston et al. 2007, Elliott et al. 2009).

These include prey size and distribution, wing-loading, flight style, and colony size

(Ashmole 1963, Gaston et al. 2007). Heavy wing-loaders such as auks (Uria spp.) are found to forage closer to the colony compared to low wing-loading petrels (Pterodroma spp.) due to greater flight costs (Gaston et al. 2007). Also, the larger the colony, the further the distance birds travel due to intra-specific competition for prey (Ashmole

1963). During the breeding season, cormorants display central place foraging where they obtain prey from distant locations and bring it back to their chicks on the nesting colony.

Cormorants are expected to travel as little as possible to foraging areas from their colony due to the high energetic demands of flight and thermoregulation (Owre 1967, Grémillet and Wilson 1999). Cormorants have high wing loading and must constantly flap during flight resulting in considerable energy costs (Owre 1967). Cormorants are also poorly insulated from the water due to their partially wettable plumage so they incur higher thermoregulatory costs than most other waterbirds (Grémillet and Wilson 1999, Enstipp et al. 2006a). Thus, from a strictly optimal foraging perspective based on the energy demands associated with flight, cormorants should travel very short distances and concentrate their activities in the immediate vicinity of breeding colonies.

Despite these predictions, cormorants have shown a large range in foraging distances from their breeding colonies. Multiple radio-telemetry studies have supported predictions based on central place foraging theory, with individuals using sites within 3 km of colonies on average (Hebshi 1998, Wanless and Harris 1993, Custer and Bunck

1992, Quintana 2001, Sapoznikow and Quintana 2003, Coleman et al. 2005, Gandini et al. 2005, Seefelt and Gillingham 2006). However, in apparent contrast to central place foraging theory, cormorants also consistently use much more distant foraging sites

(Anderson et al. 2004). For example, several studies have shown that cormorants

13 routinely forage within distances of 7-20 km from breeding colonies (Wanless et al.

1991, Grémillet et al. 1995, Neuman et al. 1997, Kotzerka et al. 2011), and maximum flight distances of up to 30 km are commonly recorded (Hobson et al. 1989, Duffy 1995,

Grémillet et al. 1995, Lantry et al. 2002, Stapanian et al. 2002). Differences between short and long flights among studies are likely due to geographical (e.g. water and land) and ecological (e.g. fish abundance) constraints, which modify the balance between energy expended and gained during foraging bouts (Stapanian et al. 2002, Sapoznikow and Quintana 2003).

Cormorants are capable of long foraging flights, and apparently choose to make them under certain circumstances, so they are potentially able to forage in multiple habitats within a large body of water, or travel between multiple bodies of water.

Although cormorant species have been shown to have high site fidelity for feeding areas in some instances (Wanless and Harris 1993, Sapoznikow and Quintana 2003, Coleman et al. 2005, Kotzerka et al. 2011), they are also known to switch locations due to changes in prey abundance (Birt et al. 1987, Wanless et al. 1991). Cormorants exhibit temporal and spatial variability in their diet throughout the breeding season in response to changes in prey density (Blackwell and Krohn 1997, Neuman et al. 1997). Thus, prey availability appears to be the main driver for cormorant foraging locations (Grémillet and Wilson

1999, Kotzerka et al. 2011). Cormorants are reported to forage in both marine and inland freshwater habitats when they are located close to both sources on the coast (Grémillet et al. 1995, Warke and Day 1995, Bearhop et al. 1999, Carss and Ekins 2002, Gandini et al.

2005, Liordos and Goutner 2008). Cormorants breeding inland have also been reported to fly and forage regularly on nearby lakes other than the ones that contain their breeding

14 colonies (Childress and Bennun 2001, Engström 2001). Bioenergetics modeling has shown that prey availability has a greater influence on daily food intake than does water temperature and dive depth (Grémillet and Wilson 1999). This would suggest that cormorants will base their foraging areas more heavily on biotic factors such as prey density than abiotic factors.

Although prey availability is proposed to be the main driver of foraging site selection, cormorants are also constrained by abiotic habitat characteristics. Water depth, substrate type, wind speed, and water clarity all factor into where cormorants select to forage. Cormorant species can dive considerable depths but often uniformly prefer to forage in shallow water less than 10 m deep (Hebshi 1998, Quintana 2001, Stapanian et al. 2002, Gandini et al. 2005, Seefelt and Gillingham 2006). The bottom substrate type that cormorants will forage over varies by species, particularly when there are cormorants feeding sympatrically (Ainley et al. 1981, Childress et al. 2002). For example, Ainley et al. (1981) found that double-crested cormorants foraged in pelagic zones above flat bottoms, whereas pelagic cormorants (Phalacrocorax pelagicus) foraged in rocky substrates, and Brandt‟s cormorants (Phalacrocorax penicillatus) fed in rocky as well as flat sandy or muddy areas. Coleman et al. (2005) observed double-crested cormorants foraging most often in lake habitats with cobble substrate, followed by those with rubble, and least often in areas of silt or clay. According to flight models, flight costs to and from a central place increase as wind force increases (Pennycuick 1978), thus affecting flight directions and distances to foraging locations. At high wind speeds over 6 m•s-1 cormorants can be wind drifted to their foraging areas (Van Eerden and Voslamber

1995). Wind speed can also affect water turbidity and sunlight penetration in shallower

15 lakes where waves can cause upwelling of bottom material (Moerbeek et al. 1987, Van

Eerden and Voslamber 1995). Cormorants detect their prey visually and so have better foraging efficiency in clearer waters (Moerbeek et al. 1987, Van Eerden and Voslamber

1995). Even though cormorants rely heavily on prey availability as a factor in selecting foraging locations, cormorants will not forage in areas where prey density is high if the habitat characteristics are not suitable (Stapanian et al. 2002).

In cases where cormorants have the opportunity to forage in different areas, wildlife and fisheries managers must take into account multiple sources of prey in order to gain a more precise estimate of cormorant impacts on fisheries. Cormorants flying up to 30 km from the colony to forage at other locations may be spreading their fish biomass removal between the lake the breeding colony is located on, and other bodies of water.

Cormorants consume a lot of fish, but if the large amounts of fish are taken from a variety of sources, as opposed to only within the vicinity of the colony, this alone may mitigate the potential negative impacts of these birds on fish populations. In order to determine if this is the case, the foraging ecology of cormorants must be studied and their feeding locations must be known.

1.4 Using Stable Isotopes to Track Sources of Cormorant Prey

Radio-telemetry has been traditionally used to track the movements of animals

(Mech 1983). This technique, however, has some difficulties associated with it. For example, the equipment needed to obtain the data can be costly, prompting researchers to use small sample sizes. Telemetry studies for cormorants have used 6 to 34 tagged individuals from colonies containing hundreds to thousands of breeding pairs (e.g.

Wanless et al. 1991, Sapoznikow and Quintana 2003, Gandini et al. 2005). Small sample sizes like these could pose problems for interpreting individual foraging patterns and applying them to colony wide foraging habits. In addition, radio-telemetry can provide information on the spatial movement of cormorants but may not be able to provide information on whether cormorants are successfully obtaining prey. For example, cormorants that are diving may not always be coming up to the surface with fish. As well, cormorants use certain distant sites as loafing areas between foraging bouts (Coleman et al. 2005). Advances in the use of stable isotopes as markers make it possible to track the movements of animals based on what they consume and do not rely on re-capture or re- sighting of tagged individuals as is done with telemetry (Rubenstein and Hobson 2004).

Stable isotopes ratios in foodwebs are determined by the external environment and incorporated into animal tissue through diet; therefore, they are useful for inferring the geographic origin of individuals by examining where the animals obtained their diet

(DeNiro and Epstein 1978, Peterson and Fry 1987, Rubenstein and Hobson 2004, Bowen et al. 2005). Such differences in stable isotopes ratios occur spatially based on various biogeochemical processes that fractionate the light and heavy isotopes at different rates

(Peterson and Fry 1987). Continental gradients of hydrogen (δD) and oxygen (δ18O) in water, for example, are used to track animal migrations on large geographic scales

(Rubenstein and Hobson 2004, Bowen et al. 2005). In North America, precipitation shows a continent-wide pattern of enriched values (more heavy isotope) in the southeast and relatively depleted values (less heavy isotope) in the northwest (Taylor 1974).

Hydrogen and oxygen stable isotopes are not useful for small-scale resolution; however, studies have been able to track animal use of, for example, inshore versus offshore,

17 marine versus freshwater, and C3 plant versus C4 plant foodwebs, based on stable isotopes of carbon (δ13C), nitrogen (δ15N), sulphur (δ34S), and strontium (δ87Sr; reviewed by

Hobson 1999). Numerous studies have used this natural variation in isotopic ratios to determine the geographical origin of various species like arthropods (e.g. Hobson et al.

1999), mammals (e.g. Cerling et al. 2006), birds (e.g. Chamberlain et al. 1997, Hebert et al. 2008), and fish (e.g. Hesslein et al. 1991).

Stable isotopes have been used to link the breeding and wintering grounds of birds including cormorants, and have been used to determine the extent of feeding in different environments (Bearhop et al. 1999, Hebert et al. 2008). The following cormorant isotope studies used tissue from the birds to determine their foraging habits.

Hebert et al. (2008) used sulphur stable isotopes in feathers to determine if there was a relationship between winter habitat use in southern U.S.A. and body condition of cormorants breeding on the Great Lakes. They concluded that cormorants with stable isotope ratios indicative of freshwater foraging during winter, and thereby use of aquaculture facilities containing abundant winter prey, had better body condition for breeding relative to cormorants with marine diet (Hebert et al. 2008). Similarly, Bearhop et al. (1999) used carbon stable isotopes in feathers to determine the extent of freshwater feeding by great cormorants wintering at inland fisheries in England. They found that cormorants fed almost exclusively in freshwater during the winter, but switched to a marine diet during the breeding season. These cormorant studies showed the utility of using stable isotopes to determine the diet and foraging areas of the birds; however, they do not tell us the extent to which cormorants are feeding on economically valuable fish species and whether the birds fed on one particular waterbody or several (Bearhop et al.

1999). Small scale variation in stable isotopes is needed to answer these types of questions. Instead of using cormorant tissue to determine specific foraging locations, stable isotopes of prey may be of use. In general, fish move on a smaller spatial scale compared to the cormorants themselves, and may therefore have the potential to show greater habitat specific stable isotopes values that can differentiate populations on a smaller scale.

Differences in baseline δ13C and δ15N values in aquatic foodwebs, which may result in different values in consumers by site, result from various sources of input including terrestrial plant and rock material, atmospheric deposits, respiratory processes, and precipitation (Peterson and Fry 1987, del Giorgio and France 1996). Habitat specific variation in baseline δ13C and δ15N values of primary producers, zooplankton, and invertebrates in food webs can exist both within and among lakes (Kling et al. 1992,

Cabana and Rasmussen 1996, Vander Zanden and Rasmussen 1999, Post 2002), which contributes to variation within the same species of fish from different habitats (Kiriluk et al. 1995, Beaudoin et al. 2001, Lake et al. 2001, Overman and Parrish 2001, Murchie and

Power 2004, Barks et al. 2010). For example, Overman and Parrish (2001) found that

δ13C values of walleye captured in different areas within Lake Champlain, Vermont, decreased from north to south by about 4‰, which was evident in all of the biota analysed. Murchie and Power (2004) collected adult yellow perch from three distinct sites on an oil-sands mining lease in northern Alberta, and found the average δ13C values among sites to differ by -2‰ and -4‰, and average δ15N values to differ by 3‰.

Baseline δ13C and δ15N values at those sites showed similar patterns (Murchie and Power

2004). Regardless of the sources of variation, whether due to anthropogenic effects or

19 natural factors, as long as fish are isotopically distinct, it may be possible to trace them back to their lake of origin (Rubenstein and Hobson 2004). The use of such intrinsic markers in fish may be useful for determining sources of cormorant prey, eliminating lakes as potential sources of prey, or both.

1.5 Cormorant Populations in Saskatchewan, Canada

In Saskatchewan, Canada, the abundance and distribution of the double-crested cormorant population is changing. The population has grown substantially in recent decades, resulting in animosity towards these birds that parallels cormorant fisheries conflicts worldwide. The estimated cormorant population in Saskatchewan grew from

1080 breeding pairs in 1968 to 34,057 breeding pairs in 2006. Cormorant numbers increased significantly in both the Prairie Ecozone in the south and Boreal Plain Ecozone in central Saskatchewan, with the latter containing 91% of the breeding population in the most recent counts. The largest nesting colonies in the province are located on mesotrophic lakes in the Boreal Plain Ecozone, some of which have grown more than 30- fold over the past 40 years, and are as large as the biggest colonies on the Great Lakes, a current hotspot for cormorant management issues (Somers et al. 2010). Within the past decade, cormorants have also been expanding the northern limit of their breeding range onto the Boreal Shield Ecozone (Doucette et al. 2010). The Boreal Shield contains thousands of oligotrophic lakes, presenting cormorants with a new environment to forage in. Oligotrophic lakes are less productive and therefore potentially more sensitive to ecosystem perturbations than the mesotrophic lakes cormorants have traditionally nested on (Downing and Plante 1993). Consequently, the overall impact of a new aquatic top

20 predator (cormorants) in these previously unoccupied areas is unknown. In addition, the environment provides cormorants with multiple lakes to obtain prey, and sources of these prey items are also unknown.

Fishing in Saskatchewan is an important economy with commercial fishing and recreational angling generating $4.8 million and $63.9 million in revenue for the province, respectively (Derek Murray Consulting Associates 2006); therefore, it is important to understand how cormorants might impact fish populations in this province.

From an economic perspective, Saskatchewan could lose millions of dollars in revenue if out-of-province anglers no longer visit due to decreased fishing quality and negative media surrounding cormorants. Anglers not only bring in revenue from buying fishing licences, but they must pay for accommodations and goods and services, which helps a variety of businesses (Shwiff et al. 2009). Lac La Ronge is Saskatchewan‟s fourth largest lake and is used for multiple economically and culturally important fisheries and has experienced significant declines in fisheries quality since 1992 (Duffy 2007). Anglers from all over Canada and the U.S.A. come to fish on Lac La Ronge and the surrounding area; however, the number of endorsements issued to Saskatchewan and U.S. residents has declined since 1997 by 20% and 50%, respectively, due to lower fishing quality

(Duffy 2007). Cormorant abundance began to increase in the area shortly after fishing quality began to decline, but the relationship (cause and effect) between cormorant abundance and fisheries decline is unknown. Very few studies have been published on cormorant ecology in Saskatchewan (e.g. Kuiken et al. 1999, Doucette et al. 2008,

Doucette et al. in press), and more research needs to be done in order to understand how cormorants interact with fish populations and fisheries in the province.

1.6 Purpose of Current Research

There is conflict over fisheries resources between cormorants and humans in the

La Ronge area of north-central Saskatchewan, which is situated on the border of the

Boreal Plain and Boreal Shield Ecozones. Cormorants are expanding their breeding range onto the oligotrophic lakes of the Boreal Shield, and stakeholders are concerned about the recovery of the sport and commercial fishery in Lac La Ronge. There is a critical need to understand how newly abundant cormorants use available fish resources in areas that may be more sensitive to food web perturbations. Nothing is known about cormorant foraging ecology in the La Ronge area. Basic knowledge on what kinds of fish cormorants eat and where they get them from does not exist, and this information is needed to link cormorant fish consumption and fisheries quality.

The transition from the Boreal Plain to the Boreal Shield Ecozone in

Saskatchewan shows a natural gradient in chemically different lakes, thereby providing the opportunity for fish populations to be isotopically distinct between lakes in this area.

The change in habitat and water chemistry may provide an advantage for tracking sources of cormorant prey in multiple lake environments by comparing fish collected by adults at breeding colonies to fish sampled from potential foraging areas.

The overall purpose of my research was to develop an understanding of cormorant foraging ecology in an area of recent range expansion and conflict with humans. In particular, there are several lakes and areas within lakes that are easily accessible to cormorants in the La Ronge area, and my objective was to determine whether they routinely target specific foraging locations, or move among several sites. My intent was to use this information to aid understanding of cormorant and fisheries management

22 issues in the area, and to better inform assessments of cormorant interactions with local fish populations.

The specific objectives of my research were to:

1) Characterize the diet of double-crested cormorants in the La Ronge area to

develop some basic information on their feeding.

2) Assess the utility of carbon and nitrogen stable isotopes as intrinsic markers in

fish to track cormorant feeding locations and determine whether breeding

cormorants feed on one or multiple sources of prey.

3) Assess lake use by cormorants in the area and describe what environmental

factors may play a role in the selection of foraging areas.

2. Methods

2.1 Study Species

Double-crested cormorants (hereafter referred to as cormorants) are one of six cormorant species in North America; they have the widest geographic range and are the only cormorant species to be found in the interior part of the continent (Hatch 1995).

Cormorants are common inhabitants of both seacoasts and inland freshwaters (Hatch and

Weseloh 1999). Cormorants are dark brown to black waterbirds with an orange gular sack and moderately long neck and bill. Lengths of these birds range from 70-90 cm and they weigh between 1.2-2.5 kg (Wires et al. 2001). They are named for the feather crests on their head, which develop, and are briefly present, early in the breeding season

(Nelson 2005). During the summer months, cormorants breed across much of North

America, with particularly large populations in southern Canada and the northern U.S.A.

Cormorants usually nest on small islands ranging from 0.5-10 ha in size in fairly remote areas (Wires and Cuthbert 2010). Egg laying first occurs within 2-3 weeks of spring arrival (which varies by location), with clutch size averaging four eggs (Hatch and

Weseloh 1999). After about 30 days of incubation, both parents brood and feed the highly altricial chicks. Chicks remain in their nests until 3-4 weeks old, and then begin to leave the nest and form large crèches with other chicks in the breeding colony (Hatch and

Weseloh 1999). Parents continue to feed their young until they fledge and are able to fly at 6-8 weeks old (Hatch and Weseloh 1999). Cormorants reach sexual maturity at approximately age 3 and have a potential lifespan of 15+ years (Nelson 2005).

Cormorants eat mainly fish but they will also less frequently consume amphibians and crustaceans. These birds are condemned by fishermen and other stakeholders in North

America and are subject to widespread population control measures due to fisheries conflicts. In Canada, the cormorant is managed and protected provincially and not under federal law (Keith 1995).

2.2 Study Area

I conducted my research from June to August of 2009 and 2010 in the La Ronge area (55° 6′ N, 105° 18′ W) of north-central Saskatchewan, Canada (Fig. 1). There are a variety of lakes and habitats for breeding cormorants to forage on in this area, which has several large and many small lakes. Cormorants in the La Ronge area are at the northern limit of their breeding range in Saskatchewan (Doucette et al. 2010). Five new cormorant colonies, ranging from 50 to 200 breeding pairs, have been discovered on Lac La Ronge since 2006 (Fig. 2). The largest known nesting colony (~4000 breeding pairs) in the La

Nemeiben Lake

North Lac La Ronge A B

Bigstone Lake Central Lac La Ronge Egg Lake

South Lac La Ronge

Ronge area established sometime >15 years ago and is located on Egg Lake, 17 km west of Lac La Ronge (Fig. 2). Two smaller breeding colonies (~800 breeding pairs) are also present on Egg Lake (Fig. 2). In 2009, I focused my study on the largest nesting colony on Egg Lake (Colony A = 55° 7′ 14.88” N, 105° 33′ 17.05” W). However, this colony was abandoned in 2010 so I focused on one of the smaller nesting colonies on Egg Lake

(Colony B = 55° 7′ 21.05” N, 105° 30′ 20.11” W), approximately two kilometres east of

Colony A. In both years, I was working on the largest known breeding colony in the area.

Egg, Bigstone, and Nemeiben Lakes and Lac La Ronge are suspected major foraging sites for breeding cormorants in the study area (Fig. 2). These four sites were identified based on anecdotal observations by other researchers and local residents, as well as personal observations early in my research. These foraging locations also fall within a 20-km flying radius to the east around the major breeding colonies on Egg Lake, which makes them accessible to foraging cormorants (Custer and Bunck 1992, Neuman et al. 1997). I did not consider potential foraging sites to the west of Egg Lake because there are almost no bodies of water in that direction, cormorants were almost never observed moving to or from the west, and fisheries conflicts with cormorants are currently concentrated east of Egg Lake on Lac La Ronge and Nemieben Lake. Egg and

Bigstone Lakes are situated on the Boreal Plain Ecozone and are characterized as shallow eutrophic lakes. Egg Lake is 10,690 ha (Cushing 1964) with an average depth of 3 m.

Bigstone Lake is smaller at 1,980 ha (Cushing 1964) with an average depth of 1.5 m.

Large macrophytes are abundant in Bigstone Lake during the summer, essentially filling the lake from top to bottom. Nemeiben Lake (15,483 ha) is an oligotrophic lake situated on the Boreal Shield Ecozone 20 km north of the town of La Ronge, and has three

27 distinct basins: North Bay, Central, and Outlet Region (Duffy 2008). The average depths of these basins are 30 m, 7 m, and 3 m, respectively. Due to these different basins,

Nemeiben Lake is heterogenous in terms of aquatic habitat. Lac La Ronge is the largest study lake at 130,700 ha (Duffy 2007). Lac La Ronge can be divided into three distinct zones: north, central, and south Lac La Ronge. North Lac La Ronge is situated on the

Boreal Shield Ecozone and is characterized by deep (≥ 25 m), clear waters; south Lac La

Ronge is situated in the Boreal Plain Ecozone and is characterized by relatively shallow

(≤ 10 m), productive waters. Central Lac La Ronge is a transition zone between the two other areas and is intermediate in terms of habitat characteristics.

Nemeiben Lake and Lac La Ronge have been used for commercial, sport, and subsistence fishing for decades (Gloutney and Chen 1992, Wallace 1993). Commercial fishing on Nemeiben Lake began prior to the 1940‟s, and sport fishing increased notably after 1958 when road access was developed (Wallace 1993). The commercial and recreational fisheries on Lac La Ronge were established in the 1920‟s and 1950‟s, respectively (Gloutney and Chen 1992). Human development (i.e. cottages) on these two lakes has increased substantially since the 1980‟s (Gloutney and Chen 1992, Wallace

1993). The fisheries in both lakes have declined in quality since the 1960‟s (Gloutney and

Chen 1992, Wallace 1993), and stakeholders are currently concerned that the recent arrival of cormorants as new predators in the area may cause further decline in fish populations, or prevent recovery. Of particular interest is the Lac La Ronge fishery, which has continued to decline since 1992 (Duffy 2007), coincident to the establishment of a large cormorant breeding colony nearby on Egg Lake. There is very little human activity on Egg and Bigstone Lakes, with the exception of occasional subsistence fishing

28 and one commercial fishing license on Egg Lake (M.Duffy, personal communication).

Thus, concern over cormorants feeding on Egg and Bigstone Lakes is limited. However, if cormorants on Egg Lake regularly fly to distant sites on Nemeiben Lake and Lac La

Ronge to consume fish, then the fisheries conflict is potentially more serious.

2.3 Nestling Diet Composition

I collected spontaneously regurgitated fish (boluses) from cormorant nestlings on

Colony A on June 18 and July 5, 2009, and on Colony B from June 14 – July 22, 2010 on ten separate days. Bolus collections were done in the mid-morning after adults presumably fed their chicks and during fair weather to avoid exposure of nestlings to extreme temperatures. Boluses were collected throughout the entire colony to represent as many breeding pairs as possible; all boluses were placed individually in plastic bags for further analysis.

Regurgitated fish were identified to species and measured (total length) to the nearest five millimetres. Using this information I calculated the percent composition, frequency of occurrence, and biomass of fish species observed in the boluses. The biomass of each fish species consumed, except for those with negligible percent composition and frequency of occurrence (< 1%), was estimated by first determining the log-length log-mass regression equation. Growth curve parameters for yellow perch were calculated; equations for other fish species were taken from the literature (Coregonid, sucker, trout perch - Carlander 1969; Pike - Willis 1989; Shiner - Hartman and Margraf

1992; Burbut - Fisher 1996; Stickleback - Ward et al. 2008). All fish species were divided into size categories by 10-mm length increments. For each size category, the

29 regression equations and frequency of occurrence of fish of each size class were used to determine the mass. All size category masses were then summed for total biomass.

Nestling diet characterized via regurgitations in colonies may not be an entirely accurate reflection of what adult birds eat (Doucette et al. in press). However, because chick rearing is energetically demanding (Dunn 1980), the fish that adults bring back to their chicks represent a substantial amount of biomass removed from lakes, and are an important consideration for fisheries management. In addition, I can gain considerable insight into cormorant foraging ecology using nestling diet because adults must still make decisions about where to forage, and will collect fish from important feeding sites.

2.4 Stable Isotope Sample Collections and Analyses

I used carbon and nitrogen stable isotopes as markers in fish to determine sources of cormorant prey; these stable isotopes were selected because they are known to vary by lake and habitat. My overall approach was to compare fish collected by cormorants

(unknown origin) to those sampled at potential foraging sites (known location). To generate stable isotopes values for fish from known locations, I sampled fish from lakes, and different areas within large lakes, in the La Ronge area that were identified as potential foraging locations for cormorants. I used bottom-set, multi-panel gill-nets (mesh sizes 0.5, 1.0, 1.5, 2.0, 2.5, 3.0 inches; 3 nets) in depths ranging from 2-10 m overnight for 12-16 hours. This sampling method was approved by the President‟s Committee on

Animal Care at the University of Regina (AUP #09-06, Food Web Structure of

Freshwater Fish in Saskatchewan). I set out gill-nets at five sites in 2009 from July 4 to

July 13: Egg Lake, Bigstone Lake, north Lac La Ronge, central Lac La Ronge, and

Nemeiben Lake (Fig. 3). In Nemeiben Lake in 2009, gill-nets were set out in the central and the eastern areas. In 2010, gill-nets were set out from June 23 to July 10 at all of the same sites plus three additional ones: south Lac La Ronge, the western part of Nemeiben

Lake, and Head Lake, a small inlet that flows into Nemeiben Lake from the west (Fig. 3).

Gill-nets were also set on August 1 and 2, 2010, on Egg and Bigstone Lakes to increase sample size and determine whether stable isotopes values changed in fish within the season. All captured fish were identified to species and measured (total length) to the nearest five millimetres and frozen for further analysis. I chose yellow perch (Perca flavescens) in the 70-130 mm size range as the focus for stable isotopes analysis because they were the most abundant prey item in cormorant nestling diet (see section 3.1 below).

I randomly chose up to 30 gill-netted yellow perch from the specified size range each year from all sampled locations. The minimum sample size of gill-netted (known location) yellow perch was 6 from eastern Nemeiben Lake in 2009, which was due to very low capture rates. Thus, from known locations I analyzed δ13C and δ15N values in a total of 122 gill-netted yellow perch in 2009, and 146 in 2010.

Samples of whole and partially digested regurgitated yellow perch (70-130mm

TL) from each day of nestling bolus sampling (section 2.3) were collected for stable isotopes analysis to determine the origin of fish prey in cormorant diet (unknown fish).

Samples were collected on two days from Colony A in 2009 as a pilot study, and 10 days from Colony B in 2010. Fifteen regurgitated yellow perch (unknown location) were analysed for June 18, and 159 were analysed for July 05, for a total of 186 prey items in

2009. In 2010, I expanded the study and analyzed up to 50 yellow perch for each of 10 sampling days. The smallest sample size was 13 (July 13), for a total of 391 regurgitated

West East Nemeiben Nemeiben

Head Lake Central Nemeiben

North Lac La Ronge

Egg Lake Central Lac La Ronge Bigstone Lake

South Lac La Ronge

32 prey items. Yellow perch were chosen randomly for analysis out of the total pool collected from the colony of each sampling date. This selection method ensured that analyzed fish came from a large number of nests, and therefore foraging adult birds (that subsequently fed these fish to their young).

I removed dorsal muscle tissue from both gill-netted and regurgitated yellow perch and rinsed the tissue in deionized water. Muscle samples were dried at 48˚C for

≥48 hours and ground into a fine powder using an amalgamator. Zooplankton for baseline stable isotopes values were collected from July 04 – 12 in 2009 and August 10 – 12 in

2010, using an 80 μm mesh size Wisconsin net. Zooplankton samples were dried and processed whole. All samples for stable isotopes analysis were weighed into tin capsules and analysed using a Finnigan Delta Plus isotope ratio mass spectrometer at the

Environmental Quality Analysis Laboratory (Faculty of Science, University of Regina,

Regina, SK). Stable isotopes ratios are in delta (δ) notation denoting the difference between sample and international standards in parts per thousand (‰):

H δ X = [(Rsample/Rstandard) - 1] x 1000 where δHX is δ15N or δ13C and R is the ratio of heavy to light isotopes (i.e. 15N/14N). The international standards for carbon and nitrogen are PeeDee Belemnite (PDB) and atmospheric air (N2), respectively (Peterson and Fry 1987).

2.5 Relative Cormorant Density and Habitat Measurements

I surveyed for foraging cormorants in the La Ronge area using line transects in the potential foraging areas identified in previous sections (Fig. 3). Placement of transects was determined by numbering 1 km x 1 km grids in each area of interest. I randomly

33 selected points in the middle of each grid and connected them to create lines. Some transects had to be moved slightly to avoid boating obstacles. Transects were 300 m wide on each side of the boat for a total width of 600 m; observers were trained (prior to surveying) to work with this distance using spherical buoys placed 300 m away from the boat (Gould and Forsell 1989). The number and length of transects, and thus area searched, was decided based on the size of the site. For example, Bigstone Lake has the smallest surface area and needed only one transect, whereas central Lac La Ronge required a larger search area and, therefore, more transects. Search area was kept as proportional as possible. Approximately 3-7% of the total site was covered within these transects, with the exception of Bigstone Lake at 13% due to its relatively small area. The number of transects per site and the distance and area covered by them differed for each:

Egg Lake = 2 transects, 7.3 km, 4.4 km2; Bigstone Lake = 1 transect, 4.3 km, 2.6 km2;

Nemeiben Lake = 4 transects, 11.8 km, 7.1 km2; North Lac La Ronge = 4 transects, 9.2 km, 5.5 km2; Central Lac La Ronge = 7 transects, 23.0 km, 13.8 km2; South Lac La

Ronge = 2 transects, 10.1 km, 6.0 km2.

Line transects were driven at 15 km/h with two observers in the bow counting all cormorants within 180˚ field of view that were swimming inside the boundaries of the transect, or that flushed as a result of the approaching boat. Cormorants that landed in the water behind the boat were not included. We used binoculars to aid counting and identification of cormorants. Counts were projected up to 600 m ahead of the boat as long as the location of the cormorant was known to be within the transect. Observers were switched regularly to decrease detection bias. I surveyed the complete transect route on all lakes 23 times between June 8 and August 24, 2010. Most surveys were completed

34 within two days unless delayed by inclement weather. Transects were surveyed when visibility was >5 km, average wave height was <1 m, and wind speed was <20 km h-1. I measured water clarity with a Secchi-disk, and water surface temperature from July 11 to

August 17, 2010, once per week during transect counts at each site. These measurements were conducted when the waves were ≤ 0.5 m to get accurate Secchi-disk and surface water temperature readings. The average water depth for each transect was determined using a depth finder.

During transect counts I also observed many foraging groups of cormorants that were outside of the transect boundaries. To take advantage of this information, which is less standardized, I also recorded the GPS coordinates of cormorant foraging flocks (≥ 10 individuals) during both transect counts and normal boat travel to and from sites from

June 9 to August 22, 2010. Flock size and water depth were recorded. The Geographic

Information System (GIS) program, ArcGIS Desktop 10, was used to overlay GPS locations and sizes of foraging flocks with a map of the La Ronge area to determine spatial and temporal patterns in cormorant foraging sites.

2.6 Cormorant Flight Paths

In 2010, I recorded the number and direction of inbound and outbound flying cormorants on Colony B on eleven separate days from June 17 to July 31. Seven of these days coincided with bolus collections; flights were observed immediately prior to disturbing the colony for bolus collection. Flights were observed from an anchored boat for 30 minutes at varied times of day from approximately 600 m south of the colony where the view in all directions around the colony was unobstructed. Flight directions

35 were estimated by two observers and recorded on a paper compass rose with 0˚ facing magnetic north. The proportion of flying cormorants, both inbound and outbound, was determined for four categories of direction: Northwest, Northeast, Southeast, and

Southwest.

2.7 Statistical Analyses

I applied a general linear model (GLM) to δ13C and δ15N values from 2009 and

2010 to examine spatial and temporal differences in gill-netted yellow perch from known locations. In general, δ13C and δ15N values were normally distributed and had homogenous variances. I performed a two-factor ANOVA with site and year as fixed effects for δ13C and δ15N, separately. South Lac La Ronge, western Nemeiben Lake, and

Head Lake were excluded from the two-factor analysis because the sites were not sampled in 2009; a one-way ANOVA was performed on these sites to determine the effect of site on stable isotopes values. A Tukey HSD post-hoc test was used for multiple comparisons of significant results. The analysis was done in the statistical software package Statistica 8.

I used Linear Discriminant Analysis (LDA) with leave-one-out cross-validation to classify yellow perch from known sampling locations (gill-netted fish) based on δ13C and

δ15N variation. LDA is an effective tool for chemical pattern recognition (Li et al. 1999).

The leave-one-out method removes an individual case (fish) from the analysis and re- classifies the individual based on the discriminant function of the remaining cases (n-1).

The prior probability was set to be proportional to sample size (n=6 to 30) for each location. Following LDA and classification of known location fish, I used multivariate

36 density functions (MDF) to assign locations to prey of unknown origin collected by cormorants (regurgitated fish) based on their stable isotopes ratios. The MDF analysis classifies unknowns to the group for which the distance between a case and the centroid for each known location group is smallest (Mahalanobis distance). Regurgitated yellow perch with assignment probabilities of ≥ 50% to a site were used for determining the proportion of yellow perch assigned to a site, unless otherwise stated. Weighted average proportions (reflecting bolus sampling effort) are reported for overall assignments of all regurgitated yellow perch analysed for the season. I used various models in the LDA and

MDF analyses by pooling and removing sampled sites. MDF assignment probabilities were analysed by collection date as well as chick rearing stage (early vs. late). Early chick rearing stage was denoted by the majority of nestlings remaining in their nests

(June 17 – July 15) during disturbance, and late chick rearing stage was denoted by fledglings that formed crèches (July 16 – August 31). I conducted the LDA and MDF in the statistical software, R 2.12.1, using the rrcov, mvtnorm, and mvnmle packages.

I used a GLM to compare cormorant densities among my study sites based on transect surveys. Prior to the analysis, I determined that the number of transects per site

(R2 = 0.006, df = 450, p = 0.091) and transect length (R2 = 3.54E-6, df = 450, p = 0.968) did not significantly influence count data. Daily cormorant density was therefore calculated for each site by dividing the numbers of birds by the area searched. I performed a two-factor ANOVA on cormorant density with site and chick rearing period

(early vs. late) as fixed effects. Cormorant densities were log-transformed to meet the assumptions of ANOVA; a value of one was added to all densities to accommodate zeros in the transformation. The log transformation resulted in a normal distribution, but it did

37 not correct for equal variances among sites. ANOVA is robust to minor deviations from equal variance, so I proceeded with the parametric version of the GLM (Box 1954). The

Games-Howell post-hoc test for unequal variances was performed following a significant outcome of ANOVA. Transect count data are reported as geometric means due to the log-transformation; GLM statistical tests were performed in SPSS 11.0.

I used principal components analysis (PCA) to examine environmental factors related to relative cormorant density. The variables I used included water clarity (a proxy for productivity), water surface temperature, water depth, and distance from the Egg Lake colony to the foraging site. Although cormorants observed on transects may also be from other colonies, distance from the Egg Lake colony was used because it was the largest known breeding population of cormorants in the area. Cormorant density was used as a supplementary variable in the PCA. Water clarity and temperature were measured starting in July; therefore, cormorant density in the PCA was for the months of July and

August only. Data were analysed by transect within sites to capture a greater range in environmental variation. PCA was conducted using the statistical programs Statistica 8 and Canoco 4.5.

3. Results

3.1 Nestling Diet Composition

I collected 147 and 701 boluses from Colony A (2009) and Colony B (2010), respectively, and was able to identify 94% (Colony A = 765/816) and 99% (Colony B =

5204/5252) of the fish present. Small yellow perch, ranging from 36-196 mm, were the most important prey item of cormorant nestlings on Egg Lake colonies (Fig. 4). Yellow

90 a) 80 2009 70 2010 60 50 40

% Number % 30 20 10 0

80 b) 70 60 50 40 30

20 % Occurrence % 10 0

60 c) 50 40 30

20 % Biomass % 10 0

39 perch occurred in approximately half of the boluses collected from Colony A (45%) and three-quarters of the boluses collected from Colony B (76%). In 2009, 66% of the fish identified were yellow perch with an average total length (TL) of 92.5 ± 29.1 mm. These yellow perch ranged in size from 45-190 mm and made up 30% of the total fish biomass fed to nestlings. Boluses from Colony B in 2010 contained a similar size range of yellow perch (36-196 mm; average TL = 78.9 ± 21.5 mm), which composed 83% of all fish found in boluses and made up 55% of the total fish biomass fed to nestlings on Colony B.

Coregonids (Coregonus clupeaformis and Coregonus artedi) and Catostomids

(Catostomus commersoni and Catostomus catostomus) were the most prevalent fish in boluses after yellow perch (Fig. 4). In 2009, Coregonids ranging from 109-228 mm

(average TL = 186.2 ± 29.0 mm) in size were the second-most abundant prey items from

Colony A, followed by Catostomids that ranged from 116-275 mm (average TL = 207.0

± 46.7 mm). Coregonids and Catostomids occurred in over half (66%) of the boluses collected from Colony A in 2009 and made up approximately 70% of fish biomass consumed. In 2010, Coregonids contributed 36% less to fish biomass, resulting in

Catostomids being the second-most abundant prey item from Colony B. Together, these fish made up approximately 28% of fish biomass consumed. Size ranges (TL) of

Coregonids and Catostomids in 2010 were similar to those measured in 2009 (Coregonids

= 80-195 mm, 133.3 ± 30.9 mm; Catostomids = 100-290 mm, 155.6 ± 41.9 mm).

Other prey consumed included ninespine sticklebacks (Pungitius pungitius; TL =

51.6 ± 5.2 mm), spottail shiners (Notropis hudsonius; TL = 66.9 ± 12.1 mm), trout perch

(Percopsis omiscomaycus; TL = 91.3 ± 25.8 mm), burbot (Lota lota; TL = 225.6 ± 73.7 mm), and crayfish (~50 mm). Sport fish were not detected in any of the boluses in 2009.

In 2010, northern pike (Esox lucius) represented 1% of all fish identified and occurred in

3% of the boluses. The estimated size ranges of pike consumed by nestlings were 58-64 mm (61.8 ± 2.5 mm, n = 21) and 100-275 mm (198.6 ± 45.7 mm, n = 21). Pike in the 58-

64 mm size range were found once in one bolus collected on June 22, 2010. Northern pike made up 4% of the total fish biomass consumed by nestlings. Walleye (Sander vitreus) were not detected in any of the 2010 boluses.

3.2 Stable Isotope Variability and Sources of Prey Items

Stable Isotope Variability of ‘Known Location’ Prey

In 2009, yellow perch showed varying degrees of overlap in δ13C and overall spanned 11.8‰ (-30.2‰ to -18.5‰), spanning about 3-4‰ within each site with the exception of Nemeiben Lake (Fig 5a). Carbon stable isotope values for Nemeiben Lake varied the most, spanning 11.7‰ (-30.2‰ to -18.5‰). It appears there are two isotopically distinct populations of yellow perch, corresponding with the central and eastern areas of Nemeiben Lake sampled. Yellow perch sampled from the central area showed large variability spanning 6.9‰ (-30.2‰ to -23.4‰) in δ13C, and those sampled from the east spanned a smaller range with 2.4‰ (-20.8‰ to -18.5‰). In 2010, overall

δ13C values spanned 14.6‰ (-33.1‰ to -18.5‰) (Fig. 5b). Interestingly, all three Lac La

Ronge sites in 2010 showed similar δ13C values ranging from -29.2‰ to -24.6‰ in δ13C.

Yellow perch from Nemeiben Lake again showed the greatest range in δ13C values.

Carbon stable isotope values from western, central, and eastern Nemeiben Lake were:

6.4‰ (-33.1‰ to -26.7‰), 9.5‰ (-31.4‰ to -21.8‰), and 4.9 (-23.4‰ to -18.5‰), respectively. Carbon stable isotope values in gill-netted yellow perch (TL = 70-130 mm)

14 a) Bigstone Egg 12 Nemeiben Central LLR North LLR

N (‰) N 15

δ 8 Nemeiben Bigstone 6

Lac La Ronge Egg 4 -35 -30 -25 -20 -15

13 δ C (‰)

b) 14 Nemeiben Central Nemeiben East Nemeiben West 12 Head Lac La Ronge Bigstone 10 Egg

Egg

N (‰) N 15 δ 8 Central All Lac La Nem. Ronge 6 West Nemeiben Bigstone

4 -35 -30 -25 -20 -15 13 δ C (‰) Figure 5. Carbon and nitrogen stable isotopes values of individual yellow perch (TL = 70-130 mm) and average zooplankton values with standard error bars in (a) 2009 and (b) 2010. All fish were collected from suspected cormorant foraging locations using bottom- set gill nets, and represent „known‟ locations. In 2010 all Lac La Ronge zones were combined and Egg and Bigstone Lakes include yellow perch from a second round of gill- netting in August.

42 varied significantly among sites (F=103.3, df=5, error df=208, p<0.001), but not between years (F=0.8, df=1, error df=208, p=0.373; Table 1). Post-hoc testing showed that

Bigstone Lake and eastern Nemeiben Lake were significantly different from all other sites. Homogenous groups that were significantly different from all other comparisons were Egg Lake and central Lac La Ronge, and central Nemeiben Lake and north Lac La

Ronge. For sites sampled only in 2010 (south Lac La Ronge, Head Lake, western

Nemeiben Lake), significant differences were found between sites (F=131.4, df=2, error df=45, p<0.001). All three sites were significantly different from each other. There was a significant interaction between lake site and year (F=2.7, df=5, error df=208, p=0.022) indicating that δ13C values in yellow perch from different sites depended on the year they were sampled. This resulted from north Lac La Ronge showing higher (more littoral)

δ13C values in 2010 compared to 2009, whereas δ13C values for all other sites in 2010 were lower.

On average, δ15N values in yellow perch were higher in 2009 (10.1 ±1.5‰) than

2010 (9.7 ± 1.4‰). In 2009, yellow perch overall spanned 6.3‰ (7.1‰ to 13.3‰) in

δ15N, approximately 2-3‰ within each site, with the exception of Nemeiben Lake (5.1‰;

7.1‰ to 12.2‰) (Fig. 5a). Nitrogen stable isotope values for yellow perch from central and eastern Nemeiben Lake spanned 3.3‰ (8.9‰ to 12.2‰) and 0.9‰ (7.1‰ to 8.0‰), respectively. In 2010, yellow perch overall spanned 6.4‰ (5.9‰ to 12.2‰; Fig.5b).

Nitrogen stable isotope values for all three Lac La Ronge sites were similar (10.0‰ to

11.9‰). In comparison, all three areas in Nemeiben Lake showed different δ15N values.

Nitrogen stable isotope values for the western, central, and eastern areas ranged: 2.4‰

(9.8‰ to 12.2‰), 3.8‰ (8.4‰ to 12.2‰), and 1.2‰ (7.2‰ to 8.4‰), respectively.

2009 Location N Mean δ13C Std. Dev. Mean δ15N Std. Dev. Bigstone 29 -22.5 1.2 10.1 1.5 Egg 30 -25.5 0.4 10.4 0.4 Central Nemeiben 14 -26.9 2.0 11.4 0.5 East Nemeiben 6 -20.6 2.3 7.8 0.7 North LLR 22 -27.7 1.1 10.7 0.4 Central LLR 21 -25.9 0.9 11.8 0.8 2010 N Mean δ13C Std. Dev. Mean δ15N Std. Dev. Bigstone 20 -23.3 1.2 9.7 1.4 Egg 16 -26.4 0.7 10.0 0.8 Central Nemeiben 17 -27.1 2.7 10.4 1.2 East Nemeiben 13 -20.5 1.4 7.8 0.4 North LLR 19 -26.7 1.0 10.8 0.4 Central LLR 15 -26.3 0.8 10.9 0.6 West Nemeiben 24 -28.9 1.6 11.3 0.6 Head Lake 9 -21.4 0.7 6.7 0.4 South LLR 15 -26.3 0.4 11.0 0.3

Significant differences were found among sites (F=151.8, df=5, error df=208, p<0.001) and between years (F=11.7, df=1, error df=208, p=0.001; Table 1). Egg Lake and central

Lac La Ronge were significantly different from all other sites. Homogenous groups that were significantly different from all other comparisons were central Nemeiben Lake and north Lac La Ronge, and Bigstone Lake and eastern Nemeiben Lake. For sites sampled only in 2010 (south Lac La Ronge, Head Lake, western Nemeiben Lake), significant differences were found between sites (F=304.2, df=2, error df=45, p<0.001), with only

Head Lake being significantly different from the two other sites. There was a significant interaction between site and year (F=4.9, df=5, error df=208, p=0.0003), indicating that

δ15N values in gill-netted yellow perch differed among sites based on the year they were sampled. This resulted from central Lac La Ronge and central Nemeiben Lake having significantly higher δ15N values in 2009 compared to in 2010.

Zooplankton differed among lakes in 2009, showing baseline differences that paralleled findings for yellow perch collected from the same locations. The average δ13C and δ15N values for zooplankton spanned -4.1‰ (-29.4‰ to -25.3‰) in δ13C and 1.9‰

(5.3‰ to 7.2‰) in δ15N across sites (Fig. 5a). Considering average δ13C values, zooplankton from Egg, Bigstone, and Nemeiben Lakes differed from each other by approximately 2‰, whereas Lac La Ronge and Nemeiben Lake differed by only 0.4‰.

In comparison, average δ15N values differed by about 1‰ among sites, with Bigstone and

Nemeiben Lakes showing similar results. Average δ13C and δ15N values for 2010 zooplankton showed some differences by site, except for the three Lac La Ronge zones, which overlapped, a similar situation to gill-netted yellow perch. With the exception of

Egg Lake, average δ13C and δ15N of zooplankton spanned -1.5‰ (-28.7‰ to -27.2‰) in

δ13C and 1.6‰ (5.9‰ to 7.5‰) in δ15N across sites (Fig. 5b). Egg Lake had an average

δ13C value of -30.3‰, but a notably higher δ15N value of 9.1‰ compared to the other sites.

Temporal differences in stable isotopes in yellow perch and zooplankton were evident between years, but there may be some differences within the season as well. In

2010, Egg and Bigstone Lakes were sampled twice during the season in late June and early August. Sample sizes for Egg and Bigstone Lakes increased from 6 to 16 and 3 to

20, respectively, as a result of the two sampling periods. Carbon stable isotope ratios in yellow perch from Egg Lake averaged -25.9 ± 0.8‰ (-26.7‰ to -24.4‰) in late June and

-26.8 ± 0.4‰ (-27.3‰ to -26.2‰) in early August. Nitrogen stable isotope ratios were more similar between sampling periods: 9.7 ± 0.7‰ (8.5‰ to 10.3‰) in late June; 10.2 ±

0.8‰ (8.8‰ to 11.2‰) in early August. In Bigstone Lake, carbon stable isotope ratios in yellow perch collected in late June averaged -21.5 ± 0.6‰ (-22.2‰ to -21.0‰) and in early August averaged -23.6 ± 1.0‰ (-25.8‰ to -21.2‰). Similar to Egg Lake, nitrogen stable isotope ratios of yellow perch from Bigstone Lake did not appear to change between sampling periods: 9.7 ± 0.7‰ (8.5‰ to 10.3‰) in late June; 8.3 ± 0.7‰ (7.2‰ to 9.8‰) in early August. In general, there was a shift in δ13C to slightly lower (more pelagic) values by 0.9‰ and 2.1‰ for Egg Lake and Bigstone Lake, respectively.

Considering that δ15N values did not appear to change within the season, the change in

δ13C could be due to the fish moving into, and obtaining more of their diet, from the pelagic zones of the lakes, as opposed to a within season change in the baseline.

Classification of ‘Known Location’ Prey

I used various LDA models for both years to determine cross-validation success rates of gill-netted yellow perch. Three LDA models were analysed for gill-netted yellow perch in 2009. The first model assumed distinct populations for Egg and Bigstone Lakes, two populations in Lac La Ronge (north and central), and two distinct populations of yellow perch from Nemeiben Lake (central and east). This analysis resulted in 77%

(94/122) correct classification of yellow perch to their lake of origin (Table 2). Overall, this classification system worked well for all sites except Nemeiben Lake. For example,

26 of the 29 (90%) yellow perch from Bigstone Lake were correctly re-assigned back to

Bigstone Lake, whereas the other three yellow perch were incorrectly classified to Egg

Lake. In contrast, 11 of 14 (79%) yellow perch from central Nemeiben Lake were re- assigned to other sites, reducing the overall cross-validation success rate.

The second model assumed only one yellow perch population from Nemeiben

Lake, and in the third Nemeiben Lake was removed altogether. The cross-validation success rates were 71% (87/122) and 86% (88/102) for models 2 and 3, respectively.

Cross-assignment success rates within most sites in both analyses were the same as in the first model in Table 2 (Appendix A). A notable change in the second model was that none of the yellow perch sampled from Nemeiben Lake were classified correctly, suggesting that the assumption of a single isotopic population is invalid. The third model increased overall classification success for the remaining four sites by eliminating the uncertainty introduced via the wide range of isotopic values for yellow perch sampled from Nemeiben Lake. Notably, in model 3 there was an increase in correct re-assignment by 5% for yellow perch from north Lac La Ronge, the nearest and most similar site to

Predicted Central Eastern North Central Actual Bigstone Egg Nemeiben Nemeiben LLR LLR

Bigstone (n=29) 26 (0.90) 3 (0.10) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00)

Egg (n=30) 0 (0.00) 30 (1.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00)

Central Nemeiben (n=14) 0 (0.00) 3 (0.21) 3 (0.21) 0 (0.00) 5 (0.36) 3 (0.21)

East Nemeiben (n=6) 2 (0.33) 0 (0.00) 0 (0.00) 4 (0.67) 0 (0.00) 0 (0.00)

North LLR (n=22) 0 (0.00) 5 (0.23) 1 (0.05) 0 (0.00) 15 (0.68) 1 (0.05)

Central LLR (n=21) 0 (0.00) 4 (0.19) 1 (0.05) 0 (0.00) 0 (0.00) 16 (0.76)

Nemeiben Lake. However, the third model suffers from the problem of eliminating a major potential foraging site for cormorants in the study area.

Carbon and nitrogen stable isotopes ratios in yellow perch from 2010 were not as clearly distinguishable among sites as in 2009, resulting in lower overall success rates for cross-validations. Yellow perch sampled from the three Lac La Ronge zones were combined for all analyses in 2010 because they had very similar and highly overlapping values. For 2010 data, I analysed four different models using LDA. The first analysis assumed distinct populations for Egg Lake, Bigstone Lake, Head Lake, and Lac La

Ronge, and three distinct populations for Nemeiben Lake (west, central, and east). Only yellow perch from the first round of gill-net sampling were included in the analysis, thus yellow perch from Egg and Bigstone Lakes collected in August were excluded to eliminate possible within season variation. Overall cross-validation success for correct re- assignment was 69% (82/119), but varied widely by site (Table 3). The highest re- assignment rates were for eastern Nemeiben Lake (100%; 13/13), Head Lake (100%;

9/9), and Lac La Ronge (94%; 44/47). The lowest re-assignment rates were for Egg Lake

(17%; 1/6), central Nemeiben Lake (6%; 1/17), and Bigstone Lake (0%; 0/3), where yellow perch were most often re-assigned to other sites.

The highest cross-validation success generated (85%; 87/102) was for model 2 when central Nemeiben Lake, the site with the largest variability in stable isotopes ratios in yellow perch, was removed from the analysis. With central Nemeiben Lake removed, correct assignment rates for Egg Lake and western Nemeiben Lake increased by 50%

(from 17% to 68%) and 13% (from 58% to 71%), respectively. All other sites had similar re-assignment rates to model 1 in Table 3 (Appendix B). For the third model, I added

Predicted Lac La Central East West Actual Bigstone Egg Ronge Nemeiben Nemeiben Nemeiben Head

Bigstone (n=3) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 3 (1.00) 0 (0.00) 0 (0.00)

Egg (n=6) 0 (0.00) 1 (0.17) 3 (0.50) 2 (0.33) 0 (0.00) 0 (0.00) 0 (0.00)

Lac La Ronge (n=47) 0 (0.00) 0 (0.00) 44 (0.94) 2 (0.04) 0 (0.00) 1 (0.02) 0 (0.00)

Central Nemeiben (n=17) 0 (0.00) 4 (0.24) 5 (0.29) 1 (0.06) 2 (0.12) 5 (0.29) 0 (0.00)

East Nemeiben (n=13) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 13 (1.00) 0 (0.00) 0 (0.00)

West Nemeiben (n=24) 0 (0.00) 0 (0.00) 10 (0.42) 0 (0.00) 0 (0.00) 14 (0.58) 0 (0.00)

Head (n=9) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 9 (1.00)

50 yellow perch gill-netted in August from the second round of sampling to increase sample sizes for Egg (n= 6 to 16) and Bigstone (n= 3 to 20) Lakes. Central Nemeiben Lake was placed back into the analysis. Cross-validation success rate for correct re-assignment was

71% (103/146; Appendix B). Eighty percent (16/20) of Bigstone yellow perch were correctly re-assigned compared to 0% in the previous two models. Re-assignment rates for Egg and western Nemeiben Lake were low at 38% (6/16) and 63% (15/24), respectively. Eastern Nemeiben Lake, Lac La Ronge, and Head Lake had high success rates of 85% (11/13), 96% (45/47), and 100% (9/9), respectively.

In the fourth model, Bigstone Lake was removed as cormorants rarely foraged on

Bigstone Lake during the 2010 breeding season (see section 3.3 below). I used yellow perch from the first round of gill-net sampling in Egg Lake. Overall cross-validation success rate was 71% (82/116; Appendix B). The highest re-assignment rates belonged to

Lac La Ronge (94%; 44/47), eastern Nemeiben Lake (100%; 13/13), and Head Lake

(100%; 9/9). Approximately half (58%; 14/24) of the yellow perch from western

Nemeiben Lake were correctly re-assigned back to their source of origin; the other half

(42%) were re-assigned to Lac La Ronge. Only 17% (1/6) of yellow perch were correctly re-assigned to Egg Lake. The other Egg Lake perch were incorrectly re-assigned to Lac

La Ronge (50%) and central Nemeiben Lake (33%). In the next section, I describe my results for the subsequent MDF analyses on unknown origin yellow perch using model 1 for 2009 and model 4 for 2010. These models offer the best combination of resolution, confidence in assignments, and coverage of potential foraging areas.

Assignment of ‘Unknown Origin’ Prey

Yellow perch regurgitated by cormorant nestlings had a similar range in δ13C and

δ15N to those of known location perch collected from potential foraging sites. In 2009, regurgitated yellow perch spanned 9.7‰ (-30.9‰ to -21.2‰) in δ13C and 5.5‰ (7.5‰ to

13.0‰) in δ15N (Fig. 6). In 2010, there was a 1.5 fold increase in the range of δ13C values measured in regurgitated yellow perch compared to in 2009, possibly due to a larger sample size and collection from more areas. Regurgitated yellow perch from 2010 spanned 14.2‰ (-31.6‰ to -17.2‰) in δ13C and 7.4‰ (6.4‰ to 13.8‰) in δ15N (Fig.

7a). When comparing the δ13C and δ15N values of regurgitated yellow perch to those of gill-netted yellow perch, it appeared that cormorants were collecting the majority of yellow perch from sources that were sampled. However, there are some stable isotopes values for regurgitated yellow perch that fell outside the boundary of gill-netted yellow perch values, indicating that cormorants may have taken these fish either from different lakes or different habitats within lakes that were not sampled.

Of the 186 regurgitated yellow perch analysed in 2009, 174 had assignment probabilities of ≥ 50% to a site. When I used 2009 model 1 (two distinct populations of yellow perch from Nemeiben Lake) for the MDF analysis, considering all regurgitated yellow perch analysed using weighted averages (by sample size each collection day),

46% were assigned to Bigstone Lake, followed by 24% for north Lac La Ronge, 12% for

Egg Lake, 10% for central Nemeiben Lake, 7% for central Lac La Ronge, and 2% for eastern Nemeiben Lake. A similar trend was observed when I examined individual yellow perch with assignment probabilities of ≥ 80% to a site (n=64). Forty-five percent of regurgitated yellow perch were assigned to Bigstone Lake, 34% to north Lac La

a) 15 Central LLR 14

12 North LLR

N (‰) N 10

15 Bigstone δ 9 Egg 8

6 -33 -28 -23 -18 δ13C (‰)

b) 15 14 Central LLR

13 North LLR 12

N (‰) N 10

15 Bigstone δ 9 Egg 8

6 -33 -28 -23 -18 13 δ C (‰) Figure 6. Carbon and nitrogen stable isotopes values of individual regurgitated yellow perch and proportions assigned to each foraging site by collection date: (a) June 18, 2009; (b) July 05, 2009. The approximate outline of stable isotopes values for yellow perch collected from known locations is shown for reference. Yellow perch were collected from a number of boluses from across the breeding colony, thereby reflecting the diet of a large number of birds.

a) 14 Early Chick Rearing West Nemeiben 13 Late Chick Rearing 12 Lac La Ronge 11 Bigstone

10 N (‰) N

15 9

δ East Nemeiben 8 Egg 7 6 Head 5 -33 -30 -27 -24 -21 -18 -15 δ13C (‰)

100 b) Lac La Ronge 90 Head 80 West Nemeiben 70

60 East Nemeiben

50 Central Nemeiben

40 Egg 30

Proportion Assigned (%) 10

3 7

- -

19 11 15

29 17 21 25

- - -

- - - -

Jul Jul

Jul Jul Jul

Jun Jun Jun Jun

Bolus Collection Dates Figure 7. (a) Carbon and nitrogen stable isotopes values of individual regurgitated yellow perch collected during early and late chick rearing stages (early = June 17 – July 15; late = July 16 – August 31). The approximate outline of stable isotopes values for yellow perch collected from known locations is shown for reference. (b) The proportion of regurgitated yellow perch assigned to each site by collection date.

Ronge, 13% to Egg Lake, 8% to central Lac La Ronge, and 0% to both central and eastern Nemeiben Lake.

Regurgitated yellow perch collected on June 18, 2009 from Colony A came from isotopically different source populations than the ones collected on July 05, 2009 (Fig. 6).

On June 18 (n=15), half of the regurgitated yellow perch were assigned to north Lac La

Ronge (Fig. 6a). This was followed by yellow perch assigned to central Nemeiben Lake, central Lac La Ronge, and Bigstone Lake, with the fewest yellow perch assigned to Egg

Lake. On July 05 (n=159), the majority of yellow perch collected by cormorants was assigned to Bigstone Lake, followed by Egg Lake (Fig. 6b), suggesting a switch in foraging locations compared to June 18, 2009. I observed similar trends in the proportions of regurgitated yellow perch assigned to each potential foraging site using other 2009 models as well (Appendix C).

Here, I present 2010 MDF results using model 4 (Bigstone Lake removed from the analysis) because cormorants were rarely seen foraging on Bigstone Lake during the

2010 breeding season (see section 3.3 below). Of the 391 yellow perch analysed, 367 had assignment probabilities of ≥ 50%. Considering all regurgitated yellow perch using weighted averages by sample size on each collection date, approximately 26% were assigned to Egg Lake, followed by 21% for western Nemeiben Lake and Lac La Ronge,

16% for eastern Nemeiben Lake, 12% for central Nemeiben Lake, and 5% for Head

Lake. Overall, proportions of yellow perch with assignment probabilities of ≥ 80% were similar (Egg Lake = 34%; eastern Nemeiben Lake = 22%; western Nemeiben Lake =

21%; central Nemeiben Lake = 14%; Head Lake = 7%; Lac La Ronge = 3%).

Carbon and nitrogen stable isotopes ratios of regurgitated yellow perch differed based on chick rearing stage as well as sample date, indicating frequent changes in cormorant foraging locations (Fig. 7). Overall, during the early chick rearing stage

(n=311), a majority of the yellow perch were assigned to Nemeiben Lake (western

Nemeiben Lake = 25%; eastern Nemeiben Lake = 21%; and central Nemeiben Lake =

14%) using weighted averages. This was followed by yellow perch assigned to Egg Lake at 18%, Lac La Ronge at 17%, and Head Lake at 6%. During the late chick rearing stage

(n=56), a majority of the yellow perch were assigned to Egg Lake at 56%, followed by

Lac La Ronge at 36%, central Nemeiben Lake at 5%, and western Nemeiben Lake at 3%.

None (0%) were assigned to eastern Nemeiben Lake and Head Lake. Cormorants appeared to switch main foraging locations by date as well (Fig. 7b; Appendix D).

Yellow perch collected on June 17, 2010, were assigned mostly to eastern Nemeiben

Lake with 63%, and on June 22 to the same area with 52%. On the next collection date of

June 25, 2010, yellow perch were assigned to various sites almost equally (about 20% to each of four sites). On the next dates, yellow perch were assigned mostly to western

Nemeiben Lake on June 28 (82%) and July 01 (80%), and then back again to eastern

Nemeiben Lake on July 04 (53%). During the rest of the sampling period (July 08, 13,

20, 22) yellow perch were assigned mostly to Egg Lake (47% - 66%) and Lac La Ronge

(27% - 42%). The 2010 MDF results for models 1, 2, and 3, for chick rearing stage and collection date were similar to model 4, where changes to proportions varied ± 10% depending on the site and model (Appendix E).

3.3 Relative Cormorant Density and Habitat Measurements

Relative cormorant density varied significantly by location (F=5.2, df=5, 125, p=0.0002; Fig. 8). In general, Egg Lake had the highest density of cormorants ranging from 0 to 1000 individuals per count. Nemeiben Lake had the second highest density of cormorants ranging from 0 to 600 individuals, and Lac La Ronge had counts ranging from 0 to 11 in the north, 0 to 40 in the central, and 0 to 331 in the southern portions surveyed. Very few cormorants were observed on Bigstone Lake, with the largest number of individuals counted being four. Post-hoc testing showed significant pair-wise differences between Bigstone Lake and both Egg Lake and central Lac La Ronge.

Although Nemeiben Lake had a higher relative cormorant density than central Lac La

Ronge and lower relative cormorant density than Egg Lake, it was not considered significant from Bigstone Lake, most likely due to its high variance in cormorant numbers, which affects the sensitivity of the post-hoc test.

Cormorants foraged at different sites during the early and late chick rearing stages

(Fig. 9). Relative cormorant density in the La Ronge area did not significantly differ between chick rearing stages (F=0.116, df=1, 125, p=0.734); however, the interaction between chick rearing stage and foraging site was significant (F=5.596, df=5, 125, p<0.001), indicating that foraging location was dependent on the stage of nesting during the breeding season. The most significant driver for this was that relative cormorant density on Nemeiben Lake was highest during the early chick rearing stage, and much reduced in the late chick rearing stage. In comparison, cormorant density on Egg Lake and south Lac La Ronge increased into the late chick rearing stage. Use of the different areas is further supported by observations of foraging flocks observed outside of

Nemeiben Lake

Egg Lake Bigstone Lake

Lac La Ronge

60 transects. In the early chick-rearing part of the season, most flocks were concentrated on

Nemeiben Lake (Fig. 10). However, during the late chick rearing stage, almost all cormorant flocks were observed on Egg Lake (Fig. 10). In general, cormorant flock sizes ranged from 10 to 2500 individuals (385.4 ± 56.3) in water depths ranging from 1.22 m to 13.11 m (4.54 m ± 0.31 m).

Water clarity and temperature varied among potential foraging sites. Secchi disk depths ranged from 1 to 5 m in the La Ronge area with Egg (1.1 ± 0.06 m) and Bigstone

Lakes (1.3 ± 0.3 m) having relatively turbid waters, followed by eastern (2.0 ± 0.2 m), central (2.4 ± 0.3), and western Nemeiben Lake (2.5 ± 0.3 m). Lac La Ronge showed the greatest transparency with the south (3.1 ± 0.3 m) appearing relatively more turbid compared to the central (3.5 ± 0.4 m) and north (3.5 ± 0.6 m) zones. Surface water temperature varied between 17.2˚C and 23.1˚C within and among sites.

Potential cormorant foraging sites showed variation in all four environmental variables measured (Fig. 11). The first axis of the PCA is comprised of water depth, clarity, and temperature. In decreasing order of depth, and thereby increasing productivity and surface water temperature, the sites were: north, central, and south Lac La Ronge,

Nemeiben Lake, Egg Lake, and Bigstone Lake. It is worth noting that one of the transects on central Lac La Ronge, indicated as CLLR 5 in Figure 11, showed similar environmental characteristics to Egg Lake. This transect was located near the town of La

Ronge close to the beach and lakeshore homes, and most cormorant flocks were observed there for the central Lac La Ronge site. In terms of distance on the second axis,

Nemeiben Lake was the furthest site from the Egg Lake colony. Particularly, transect numbers 3 and 2 were located in the eastern area at about 30 km away (refer to Fig. 3

Table 4. Variable loadings for environmental measurements on all factor components. Cormorant density (*) was used as a supplementary variable in the analysis.

Components Variable Factor 1 Factor 2 Factor 3 Factor 4 depth -0.889 -0.004 0.435 -0.145 clarity -0.965 -0.001 0.002 0.262 temperature 0.542 0.789 0.276 0.087 distance -0.691 0.625 -0.346 -0.111 density * 0.464 0.161 -0.131 -0.316 eigenvalue 2.49 1.01 0.38 0.11 variance % 62.32 25.34 9.61 2.73 cumulative variance % 62.32 87.66 97.27 100.00

63 for transect locations). Relative cormorant density could be partly explained by some of these environmental factors. The first component of the PCA explained 62% of the variance in relative cormorant density in the La Ronge area (Table 4), which was driven primarily by water depth and clarity. Relative cormorant density was negatively correlated with water depth and clarity, suggesting that the birds avoided deeper and less productive waters, or selected shallower, more turbid waters. Distance from the Egg Lake colony to the foraging site had a weaker effect; however, sites farther from Egg Lake did tend to have fewer cormorants. Surface water temperature was positively correlated with cormorant density in the second principal component, suggesting that cormorants chose relatively warmer waters.

3.4 Cormorant Flight Paths

Cormorants flew to and from their colony on Egg Lake (Colony B) in various directions during the season; however, there appeared to be a non-random pattern in flight direction for most days (Table 5). On certain days over the season, more than 50% of cormorants flew to and from the north towards Nemeiben Lake and north Lac La

Ronge, and on two days in July, over 50% of cormorants flew to and from a foraging flock on Egg Lake in the south-west. There were seven days in 2010 when both flight observations and bolus collections were done: June 17, 25, 28, and July 01, 04, 08, and

20. Cormorant flight paths showed high correspondences with the MDF results (see section 3.2; model 4) showing major foraging locations. For example, on June 17, about

77% of cormorants flew to and from a northern direction towards Nemeiben Lake when

63% of regurgitated yellow perch were assigned to eastern Nemeiben Lake. Another

Table 5. Proportion (%) and direction of inbound and outbound cormorant flights from Colony B on Egg Lake in 2010. Directions are divided by northeast, northwest, southeast, and southwest.

Date Sample Size NE NW SE SW 17-Jun 197 27.4 50.8 3.6 18.3 25-Jun 104 26.5 60.9 6.6 6.0 28-Jun 151 79.1 18.7 0.7 1.4 1-Jul 153 51.6 23.5 7.2 17.6 4-Jul 354 82.5 5.1 2.5 9.9 8-Jul 229 33.2 25.8 28.4 12.7 20-Jul 276 26.8 4.7 14.1 54.3 24-Jul 132 77.3 0.8 8.3 13.6 25-Jul 255 16.5 10.2 12.2 61.2 28-Jul 131 37.4 31.3 19.1 12.2 31-Jul 55 74.5 10.9 7.3 7.3

65 example shows that on July 20, about 54% of cormorants flew to and from a large foraging flock observed southwest of the colony on Egg Lake, and 66% of regurgitated yellow perch were assigned to Egg Lake that day. On July 28, cormorants that flew in a northern direction did not fly above the tree line as if they were heading to Nemeiben

Lake; instead, they landed on Egg Lake north of the colony.

4. Discussion

4.1 Multiple Cormorant Foraging Locations

Several lines of evidence from my study suggest that cormorants foraged widely over much of the La Ronge area, including different lakes and regions within lakes.

Cormorants also focused much of their feeding effort at sites that were well removed from the breeding colony lake, and switched locations regularly. Carbon and nitrogen stable isotopes in regurgitated yellow perch, relative density of cormorants, and flight paths, all provide support for this pattern of foraging by cormorants. This finding is important because it shows that cormorants are not foraging solely on the lake containing their breeding colony, or even those closest to it. Rather, cormorants are spreading their feeding effort out over at least several lakes, and thereby interacting with a variety of prey fish populations from different sources. Within these lakes, relative cormorant density was highest in waters ≤13 m deep within 3 km of shore, and lowest in open water, similar to other observations (e.g., Stapanian et al. 2002, Stapanian and Bur 2002).

Acentric colony locations on the perimeter of the foraging range, like the one in this study, have been observed in other colonial waterbirds and can lead to increased flight distances to foraging areas (Wittenberger and Dollinger 1984). Interestingly, cormorants

66 nesting on Egg Lake were collecting prey fish from areas over 30 km away from their breeding colony in totally different habitats and ecoregions. Use of these distant areas was not minor (70% of fish in 2010), indicating a relatively weak spatial connection between the location of the breeding colonies and foraging sites used in our study area.

Multiple foraging lakes and habitats may be a more common feature of cormorant ecology than previously thought. My findings are supported by other studies on great cormorants in Europe, where the birds did not consistently forage on the lake they nested on. For example, great cormorants nesting on Lake Ymsen, Sweden, spent 20% of the time foraging at other nearby lakes, 7 km and 11 km away (Engström 2001). Similarly, great cormorants nesting on Lake Selent, Germany, foraged 50% of the time on the lake itself and the other 50% of the time on two different fishing grounds in marine habitat, located an average 30 km away (Grémillet et al. 1995). In an extreme case, great cormorants regularly commuted 60 km from their coastal breeding colony on Sheep

Island in Northern Ireland, to Lough Neagh, the largest freshwater lake in the British

Isles, which supports a wide range of fish species, during the 1980‟s (Warke et al. 1994,

Warke and Day 1995). The distances travelled by cormorants in my study are comparable to those documented in other areas. Thus, cormorants might commonly move large distances, 10-60 km away from their breeding colony, if the reward of prey is sufficient.

To my knowledge there are currently no published North American studies on double- crested cormorants that specifically address multiple foraging sites, despite evidence that cormorant movement may be a common phenomenon. For example, Eisenhower and

Parrish (2009) reported that the number of foraging cormorants in a particular basin of

Lake Champlain, Vermont, declined after July 16-31 in all three study years, concluding

67 that cormorants switched to forage in other, more distant, areas. It is critical that cormorant foraging ecology, specifically where they consume prey, and how sites are selected, be more completely understood to inform management decisions.

Cormorants switched foraging areas as a colony, whereby large numbers of individuals across colonies fed on yellow perch from the same sources on the same days.

Early in the breeding season cormorant prey tended to come from distant Nemeiben

Lake, indicating that cormorants were making long foraging flights. However, later in the season the majority of prey sampled originated from sites much closer to the breeding colony, indicating that foraging trips were much shorter. Evidence for these patterns mostly arises from carbon and nitrogen stable isotopes values in regurgitated yellow perch and MDF analyses, but is also seen in analyses of relative cormorant density, and foraging flock locations. It is possible that cormorants were somehow sharing information regarding successful foraging locations (directly or indirectly), resulting in the large shifts in location that affected the whole colony. Group foraging and information transfer is common in cormorants and other colonial birds. Breeding colonies may be used as “information centres” for food supply (Horn 1968, Ward and Zahavi

1973). Colonial birds feed on patchy and unpredictable prey resources and need to find new feeding sites frequently (Bayer 1982). Information on the availability of new foraging sites is gathered through various cues (e.g. flight directions) and by following known successful foragers to feeding areas (Ward and Zahavi 1973). Waltz (1987) tested the Information-Centre Hypothesis on two common tern (Sterna hirundo) colonies and found that individuals that brought back fish from one foraging trip were more likely to be followed by others on the next trip. The Information-Centre Hypothesis is an

68 interesting ecological feature of colonial birds but is not well-studied in cormorants (but see Doucette et al. 2008). However, our findings suggest that large-scale changes in cormorant foraging locations may be common.

Surprisingly, cormorants were still targeting small yellow perch (TL = 70-130 mm) despite changes in the sources of prey items. It is currently unclear why cormorants switched locations to capture the same prey species and size; however, prey availability likely played a role in their decisions. Yellow perch abundance in Egg Lake was low in late June, where catch per unit effort (CPUE) was about 0.34 yellow perch•net-1•hour-1.

In early August, CPUE had increased to about 18.30 yellow perch•net-1•hour-1, a 54x increase in the number of yellow perch caught. Cormorants switched from foraging on

Nemeiben Lake to Egg Lake later in July, perhaps due to this increased availability of prey in Egg Lake. Other evidence in the literature also suggests that prey availability drives cormorant decisions regarding foraging locations (e.g. Warke et al. 1994, Dirksen et al. 1995, Eisenhower and Parrish 2009). For example, cormorants foraged on the freshwater lake, Lough Neagh, in Northern Ireland, for several years until the roach and perch populations began to decline sharply due to parasitic infections and commercial over-harvesting, respectively, resulting in a noticeable shift in cormorant diet to marine fish species (Warke et al. 1994, Warke and Day 1995). Dirksen et al. (1995) reported that at one point, the proportion of foraging cormorants on Lake Wolderwijd, Netherlands, increased as fish biomass increased. The fact that cormorant capture rate has been shown to relate more to prey density as opposed to visual acuity (Hao 2008), suggests that cormorants will select to forage in locations that provide the most efficiently available prey. In the context of optimal foraging theory, moving around to capture the same prey

69 item from different sources must have been a rewarding strategy for cormorants nesting on Egg Lake, despite some of the long distances travelled.

Prey availability can be affected by several factors such as season, water temperature, and habitat. In large aquatic systems like the Great Lakes region, seasonal fish migrations can directly influence prey availability for cormorants (Neuman et al.

1997, Johnson et al. 2002). Alewife (Alosa pseudohurengus), for example, come inshore to spawn during the spring (O‟Gorman et al. 1991) and become more susceptible to cormorant predation at this time. In general, the movement of fish can be attributed to changes in water temperature (O‟Gorman et al. 1991). Yellow perch often spawn early in the spring in shallow, littoral waters and then move into deeper waters once the temperature has reached 15˚C (Whiteside et al. 1985). This may have contributed to cormorants in the La Ronge area switching to forage on Egg Lake late in the season. As aquatic top predators, cormorants can induce anti-predator behaviour in fish, and submerged aquatic vegetation can provide refuge for small maneuverable fish, causing them to become less available to cormorants (van Eerden and Voslamber 1995, Engström

2001). Complex aquatic habitats with submerged aquatic vegetation can physically inhibit the swimming abilities and visual detection of cormorants (Eisenhower and

Parrish 2009). As the biomass of vegetation increases in a lake as the season progresses, cormorants will change foraging locations to areas where fish can be caught more efficiently (Grémillet and Wilson 1999, Eisenhower and Parrish 2009). The large aquatic macrophytes that grow in Bigstone Lake, may have been responsible for why cormorants appeared to not forage on the lake in 2010.

Other factors that can affect cormorant foraging locations include wind speed and direction, water clarity, and group versus single foraging. Strong winds over 6 m•s-1 can drift flying cormorants in a certain direction, or force them to forage closer to the colony rather than expend extra energy against the wind (van Eerden and Voslamber 1995).

Strong winds can in turn affect water turbidity through upwelling of bottom substrates

(van Eerden and Voslamber 1995). Cormorants rely on their vision for capturing prey, so turbid waters will reduce the visual acuity of cormorants and negatively influence their reactive distance, decreasing their rate of fish capture (Hao 2008, Strod et al. 2008).

Cormorants in general will avoid foraging in water that is naturally turbid due to large algal biomass, and use sites that are relatively clearer (Radomski and Zimba 2010).

However, group foraging in turbid waters can be highly effective for capturing prey

(Dirksen et al. 1995, van Eerden and Voslamber 1995). During social fishing, birds form a line at right angles to the direction of the flock‟s movement, where the front serves to drive schools of fish in front of the foraging cormorants, and the length of that line keeps the fish from escaping around the edges (Bartholomew 1942). Mass fishing effectively exploits turbid waters by having cormorants push the fish up to the clear top waters where they are then consumed (van Eerden and Voslamber 1995).

The stage of the reproductive cycle for cormorants can also contribute to decisions on foraging location. Adult cormorants may be seeking particular species and sizes of prey to feed juveniles, and may therefore be forced to travel widely to find these items. The energy requirements of chicks increase throughout the season and parents respond by supplying either more food or higher quality food (Grémillet 1997). The energetic costs of reaching foraging sites can factor into preferred feeding areas (Neuman

71 et al. 1997); long distance flights should be to areas supplying higher prey quantity or quality (Obst et al. 1995, Platteeuw et al. 1995). Great cormorants have travelled 35 km and even 60 km to obtain high quality fish for their young that were caught with the same catch per unit effort as lower quality fish around the breeding colony (Warke and Day

1995, Grémillet 1997). Long distance flights can be energetically demanding but breeding cormorants will undertake this when fish availability at distant locations is high

(Warke and Day 1995). Considering that cormorants can fly long distances during the energetically demanding chick rearing period (Dunn 1975), it is likely that swimming time during foraging is more critical than flight time in terms of energetic costs

(Grémillet 1997), so that sites with high prey density result in shorter swim times for cormorants.

Understanding the motivation for cormorants to target particular sites over others may provide insight into fisheries management issues. For example, the most important prey item of cormorants on Egg Lake was small yellow perch, which is often fished for recreationally once a minimum size of 250 mm is achieved (Barks et al. 2010). Initially, there could be concern for cormorants reducing the future recruitment of yellow perch populations into the fishery in a particular lake. However, cormorants are unlikely to deplete fish populations to the point where compensatory mechanisms no longer work and will forage in an area for as long as prey density is high enough to satisfy their energetic requirements, otherwise they will move on to other locations (Grémillet and

Wilson 1999).

4.2 Utility of Stable Isotopes to Trace Sources of Cormorant Prey

Although carbon and nitrogen stable isotopes values in yellow perch from known locations were not 100% distinct among sites, our findings indicate the potential usefulness of this technique to determine cormorant foraging locations. In general, 2009 data performed better in terms of classification rates for gill-netted yellow perch of known location than 2010 data due to less overlap in stable isotopes values at certain sites. Classification rates based on my data were similar to those that used LDA as a tool to distinguish fish populations of known location among estuaries and wetlands using multiple (4-5) trace elements in otoliths (Thorrold et al. 1998, Gillanders and Kingsford

2000, Brazner et al. 2004). Overall success rates ranged from 63% to 76% with individual sites having success rates ranging from 50% to 100% (Thorrold et al. 1998,

Gillanders and Kingsford 2000, Brazner et al. 2004). Similarly, habitat fingerprinting using trace elements showed that sites had significant variation both within and among sites (Thorrold et al. 1998), and there was overlap with some elements at some sites, which contributed to loss of power in the LDA (Campana et al. 1995, Campana et al.

2000, Gillanders and Kingsford 2000, Brazner et al. 2004). Considering that my study was done on a smaller spatial scale (less potential variance by site) with only two markers

(δ13C and δ15N), the LDA performed reasonably well.

Although my data showed some overlap in δ13C and δ15N values in yellow perch between locations, they typically remained separated from other sites. How well the intrinsic marker technique works for assigning unknown origin fish collected by cormorants to their source lakes depends on the amount of overlap in values among sites.

All of the sites, with the exception of central Nemeiben Lake, showed little variation and

73 overlap, creating distinct sites that regurgitated yellow perch could be assigned to with high confidence. However, the high variability in stable isotopes values of yellow perch from central Nemeiben Lake is noteworthy and accounted for the poor classification rates in the LDA for this study. Carbon and nitrogen stable isotopes values for yellow perch from central Nemeiben Lake overlapped with yellow perch from western Nemeiben Lake and up to three other lakes in the area: Bigstone Lake, Egg Lake, and Lac La Ronge. The large variability and overlap in stable isotopes of central Nemeiben Lake decreases the overall confidence in assignment probabilities of regurgitated yellow perch for the above mentioned sites. Murchie and Power (2004) observed within-lake variability of δ13C and

δ15N values in young-of-the-year yellow perch from Mildred Lake, Alberta, Canada, and suggested these differences were either due to undocumented anthropogenic effects or natural variability. The high variability in stable isotopes of yellow perch from Nemeiben

Lake may hence be due to habitat heterogeneity.

Variation in stable isotopes values between and within lakes could be due to natural and anthropogenic factors. Carbon and nitrogen stable isotopes values in zooplankton collected from my study sites varied, suggesting that differences in food web baselines led to the differences observed in yellow perch. Factors affecting baseline δ13C values in lakes on a small scale include respiratory processes and photo-assimilation of

13 13 C depleted CO2 by phytoplankton and the input of detritus that is more enriched in C

(del Giorgio and France 1996). The proportion of lighter CO2 that is assimilated by phytoplankton within a lake increases with water-column depth (del Giorgio and France

1996). Thus, vertically migrating zooplankton in deeper lakes will have relatively depleted δ13C values by incorporating 13C depleted phytoplankton in their diet. In

74 addition, eutrophic lakes tend to be more enriched in 13C where the proportion of available organic carbon is largely from terrestrial and littoral matter (del Giorgio and

France 1996). Other factors that can simultaneously affect baseline δ13C are inputs from atmospheric CO2 and dissolved inorganic carbon weathering of carbonates in the watershed (Wachniew and Rόżański 1997).

Differences in δ15N values of yellow perch can be attributed to differences in the food web baseline where food sources in one lake are more naturally enriched than in another, the fish occupy higher trophic positions, or a combination of these factors

(Vander Zanden and Rasmussen 1999, Lake et al. 2001, Quevedo et al. 2009). Small scale variation in baseline δ15N is dependent on the input of organic and inorganic nitrogen entering lakes and the variability in fractionation of different phytoplankton and microbial communities during transformations of available nitrogen (Jones et al. 2004,

Vuorio et al. 2006). Sources of nitrogen can come from the excretion of ammonium

+ (NH4 ) by zooplankton (Gu and Alexander 1993) and atmospheric deposition of isotopically light nitrogen (Vuorio et al. 2006). The balance between nitrogen supply and demand will cause baseline values to be highest for nitrogen limited lakes (Jones et al.

2004). When nitrogen supply is high, aquatic macrophytes and organic matter will discriminate against the heavy 15N isotope, becoming isotopically lighter with lower values (Evans 2001). Anthropogenic inputs can cause enrichment of δ15N baselines from sewage, atmospheric pollution, and fertilizers, which can be attributed to the high trophic level of humans and loss of the light 14N isotope through fractionation during ammonification of nitrogenous waste (Heaton 1986, Cabana and Rasmussen 1999,

Vander Zanden and Rasmussen 1999, Lake et al. 2001). Fish have shown consistency in

75 trophic position among several lakes once baseline δ15N values have been accounted for

(Lake et al. 2001); however, community structure can also affect the trophic position and thus δ15N values of organisms (Vander Zanden et al. 1999). Yellow perch, for example, have been shown to vary in lake-specific trophic positioning between secondary consumers (invertebrate feeding) and tertiary consumers (piscivorous) due to differences in food chain length (Cabana and Rasmussen 1996).

The difference in δ15N values of yellow perch sampled from the western and eastern areas of Nemeiben Lake could be due to a combination of differences in community structure and food web baseline. The eastern area of Nemeiben Lake is where most of the predatory fish (e.g. pike and walleye) are found, which could result in yellow perch appearing lower on the food chain. Unfortunately, an inadequate sample size of zooplankton was collected from eastern Nemeiben Lake for stable isotope analysis to determine if the baseline differed from the western area of Nemeiben Lake. In general, abundance of zooplankton from Nemeiben Lake appeared lower than that of other lakes.

However, I examined δ13C and δ15N values of white suckers from Nemeiben Lake and they too showed a similar trend in isotope ratios as yellow perch. The diet of white suckers consists of insect larvae and detrital matter and they are shown to be consistent in trophic level (Hesslein et al. 1991); thus, the differences in δ13C and δ15N ratios in white suckers likely indicate differences in baseline δ13C and δ15N values. It is important to note that lack of understanding of the mechanisms behind distinct stable isotopes in yellow perch populations does not detract from their utility as intrinsic markers. Rather, I am interested in detecting isotopic differences in gill-netted yellow perch and assigning regurgitated fish collected by cormorants to potential foraging sites.

Annual and seasonal changes in δ13C and δ15N were observed. Yellow perch from the three Lac La Ronge zones had similar stable isotopes values in 2010, whereas in

2009, the north and central zones were distinguishable. This finding was surprising because the northern and southern areas of Lac La Ronge are divergent in a number of important features. The central zone was considered a transition zone so it was expected to show more variable or intermediate values between the north and south. These annual changes suggest that fish samples from gill-nets need to be obtained each year if the intrinsic marker technique is to be used for determining cormorant foraging locations.

Changes within season to lower (more pelagic) δ13C values was observed in yellow perch of similar size from Egg and Bigstone Lakes in 2010 where a month went by between sampling periods. The turn-over rate for fish muscle tissue is about a month (Hesslein et al. 1993) and the observation of more yellow perch appearing in the pelagic zones of Egg and Bigstone Lakes support the finding of lower δ13C values (France 1995). The seasonal variability complicates field sampling and poses problems with timing of sampling of known location prey; however, there was overlap observed in the δ13C values of yellow perch sampled at the end of June and beginning of August. Sampling may be done more than once to obtain different sets of isotope values for analysis.

It is important to consider that the sites that I sampled were assumed to be the only sites that cormorants foraged on. However, the possibility of there being a cormorant foraging site that was not sampled with yellow perch that have similar stable isotopes values as another sampled site could exist. Hives Lake is a shallow lake 874 ha in size and similar in characteristics to Bigstone Lake. It is located approximately 9 km northeast from the Egg Lake colonies, south of Nemeiben Lake. It is accessible only by

77 all-terrain vehicle. Cormorants may have obtained yellow perch from this lake; however, the lake is not considered an important fishery and attracts little angler attention

(Saskatchewan Ministry of Environment 2003); thus, conflicts regarding cormorants foraging on Hives Lake would be limited. Although, the intrinsic marker technique could not accurately determine specifically which lake a regurgitated yellow perch originated from, this technique showed that cormorants were certainly taking yellow perch from a variety of sources, opposed to just one.

Despite some lack of precision on its own, the intrinsic marker analysis was a good indicator of cormorant foraging locations. In 2009, 78% of regurgitated yellow perch collected on July 05 were assigned to Bigstone Lake. This result is supported by the observation of a foraging flock of about 300 cormorants on Bigstone Lake, only two hours after bolus collections. For 2010, relative cormorant density and flight paths typically showed trends in foraging locations in accordance with the carbon and nitrogen stable isotopes intrinsic marker analysis. Using model 4 (Bigstone excluded), the MDF assigned 60% of regurgitated yellow perch to Nemeiben Lake during the early chick rearing stage, and then 56% of regurgitated yellow perch to Egg Lake into the late chick rearing stage, suggesting a switch in the main sources of cormorant prey. Relative cormorant densities and the GPS locations of large foraging flocks uphold this pattern, where Nemeiben Lake and Egg Lake had the highest cormorant density in the early and late chick rearing stage, respectively. In addition, cormorant flight paths from the colony to foraging locations are typically direct (Wanless et al. 1991, Kotzerka et al. 2011), and regurgitated fish were assigned to sites that were located in the flight direction that a majority of cormorants took to and from the colony. In one case, on June 17, 2010, about

500 cormorants flew in from the north to the colony, steadily for two hours in the morning (9:00-11:00). Boluses were collected right after and the MDF results showed that 63% of regurgitated yellow perch were assigned to eastern Nemeiben Lake, followed by 16% to central Nemeiben Lake, which is located north of the colony. In general, the combined approach of using all of the indicators of cormorant foraging locations and sources of prey aligned well together.

Although the weakness of the intrinsic marker analysis lies in the temporal variation of δ13C and δ15N values, this technique provides a good resolution of colony wide feeding locations and foraging habits, which can be difficult to assess using telemetry. Telemetry studies on cormorant foraging locations have used small sample sizes representing typically <5% of the colony (Wanless et al. 1991, Sapoznikow and

Quintana 2003, Coleman et al. 2005, Gandini et al. 2005, Seefelt and Gillingham 2006,

Kotzerka et al. 2011). This can be problematic when foraging patterns for a few individuals are generalized to the entire colony, especially if certain individuals or groups of birds from a colony specialize in their feeding habits (Warke and Day 1995,

Voslamber et al. 1995, Doucette et al. unpublished data). Using the intrinsic marker technique, I was able to effectively increase sample size by collecting boluses throughout the entire colony to represent as many breeding pairs as possible, and provide insight into coordinated group foraging behaviour by cormorants. In addition, telemetry suffers from logistical (e.g. transmitter size, tracking difficulty due to geography, re-captures or re- sightings) and financial (e.g. cost of equipment) constraints, leading to emerging interest in using intrinsic markers for studying animal movements (Hobson 1999, Rubenstein and

Hobson 2004). Telemetry studies are still important, however, because there is the need

79 to understand the spatial foraging patterns of cormorants, and when they are actively foraging so that their fishing efforts can be accurately assessed (Coleman and Richmond

2007).

Intrinsic marker analysis can be useful when cormorants feed in several locations on the same prey item. When waterbodies differ in fish species and size compositions, the most likely feeding grounds can be proposed through observations of prey type and sizes consumed by cormorants (Pūtys and Zarankaite 2010). For example, Dalton et al. (2009) found that white perch (Morone americana) were a dominant species in the diet of roosting cormorants. White perch were common in the Connecticut River located near the roost but were not found in Bride Lake, another foraging area used by cormorants

(Dalton et al. 2009). Similarly, marine and freshwater fish species can be distinguished in boluses from breeding colonies located on the coast to infer relative use of these two habitats by cormorants (Carss and Ekins 2002, Pūtys and Zarankaite 2010). However, if cormorants appear to target the same prey type and size class from several locations, as was found in my study, it is more difficult to infer from which location the prey came.

The advantage of using intrinsic markers on prey with similar external appearances is that different source populations can be detected chemically, and thus be used as tracers of geographic origin to determine feeding locations of cormorants.

Stable isotopes as biogeochemical intrinsic markers in fish provide a unique application to track sources of cormorant prey. The La Ronge area is a multiple lake environment that has a natural habitat gradient, creating potential isotopic differences among fish, thereby providing the opportunity to test this method. The broad utility of the stable isotopes approach using prey is currently uncertain, but has some promise.

Variation in lake chemistry is controlled by both regional (e.g. climate and geology) and local factors (e.g. input/retention of organic matter, lake morphometry; Stendera and

Johnson 2006), and it is these differences in lake properties that may create isotopically distinct fish populations, allowing for this technique to work. However, it should be recognized that as with any stable isotopes study, if fish reside in isotopically indistinguishable habitats, then their geographic origin may not be identified in this manner. A multivariate approach using several biogeochemical markers (i.e. other stable isotopes and trace elements) can greatly amplify and increase the precision in detecting small scale variation among fish populations. During preliminary analyses of intrinsic markers for this study in 2009, individual gill-netted yellow perch were screened for mercury (Hg) to determine if there were any differences in concentrations among study sites. Unfortunately, mercury in yellow perch did not show distinct variation among sites, except for Egg Lake (Appendix F), and was therefore not pursued in 2010. However, Hg might be used as a marker at other locations. Dissolved organic matter, pH, sulfate (SO4) concentrations, and body condition can all influence Hg accumulation in fish and cause small scale variation among lakes (Gilmour et al. 1992, Greenfield et al. 2001).

Biological intrinsic markers, like life history characteristics or genetics, may also be useful; however, in my study, fish length at age was not a promising marker due to considerable overlap among sites (Appendix G).

4.3 Cormorant-Fisheries Management and Future Research

The presence of cormorants combined with declines in fishing quality has generated stakeholder concern over the recovery of fisheries in Lac La Ronge, and

81 possibly Nemeiben Lake. In view of the fact that cormorants are fairly new to this system, cormorant foraging ecology and its relevance for fisheries management is still poorly understood. Aquatic conditions have become more favourable for the increased expansion of double-crested cormorants onto Lac La Ronge. Non-sport and non- commercial fish species in Lac La Ronge and Nemeiben Lake have increased in abundance since the 1960‟s (Gloutney and Chen 1992, Duffy 2008), likely due to changes in food webs through overharvest of predatory fish by humans (Hobson et al.

1989, Post et al. 2002). There have been no significant changes in the abiotic characteristics and productivity of Lac La Ronge for decades, suggesting that the observed changes have a management origin (Gloutney and Chen 1992). Regardless, changes to Lac La Ronge and other lakes in the area have created an adequate food supply for breeding cormorants in the area. The widespread increase and expansion of cormorants in North America began when fisheries were already declining (Wires et al.

2001, Diana et al. 2006), demonstrating that cormorants are likely a symptom of fisheries issues associated with declining predatory fish populations, rather than the cause of fisheries problems.

Cormorants nesting on Egg Lake targeted a wide variety of prey, but the three most important fish species consumed by nestlings were small yellow perch, coregonids

(cisco and lake whitefish), and suckers, providing additional support that double-crested cormorants seldom take economically valuable fish (Craven and Lev 1987, Hobson et al.

1989, Blackwell et al. 1995, Bur et al. 1999, Neuman et al. 1997). However, it must be noted that I only examined nestling diet for a limited part of the season. In the La Ronge area of Saskatchewan, Canada, economically important fish species include lake trout

(Salvelinus namaycush), walleye, northern pike, and lake whitefish. Lake trout and walleye were not observed in the diet of nestlings, and northern pike made up a negligible portion of the biomass consumed. Lake whitefish are commercially fished in Saskatchewan but they are difficult to distinguish in boluses from cisco, a non-commercial and non-sport fish, making the proportion of lake whitefish consumed by cormorants difficult to estimate. However, lake whitefish are at an acceptable population level in the La Ronge area (M. Duffy, personal communication) and have increased in relative abundance in

Nemeiben Lake from 14% in 1992 to 20% in 2003 (Duffy 2008), suggesting that cormorants have a negligible effect on local lake whitefish populations.

Bioenergetics modeling of cormorant fish consumption is used widely as an indicator of the level of impact cormorants have on fish, and it has been shown that cormorants consume a large amount (Glahn and Brugger 1995, Derby and Lovvorn 1997,

Engström 2001, Johnson et al. 2002, VanDeValk et al. 2002, Rudstam et al. 2004, Diana et al. 2006). Typically, these assessments tend to assume that all fish biomass removal from a breeding colony occurs on the lake they nest on (e.g. Rudstam et al. 2004), or particular area nearby (e.g. Diana et al. 2006), but clearly this is not always the case. Our findings show that cormorants obtained prey from multiple distant (~30 km) locations, which negates the approach of simply assuming one prey source and complicates fisheries management assessments. Cormorants are thus spreading their foraging efforts among various lakes and areas, and this could mitigate the potential negative impacts of these birds on fish populations. In addition, given that the expected foraging range of cormorants can be 20-30 km (Custer and Bunck 1992, Neuman et al. 1997), the daily food intake and energy expenditure of cormorants can differ widely among areas and

83 thereby affect consumption models (Ridgeway 2010). It is important to identify source populations of prey fish and resource managers must consider this when assessing cormorant impacts on a fishery of concern. Knowing where the prey fish come from and estimating relative proportions taken from various sites will refine biomass removal estimates to help understand the potential impacts of cormorants on fisheries.

Cormorant prey in my study system was obtained from lakes that are heavily used by humans for fishing and recreational activities (Lac La Ronge and Nemeiben Lake), as well as from lakes with little human activity (Egg and Bigstone Lakes). The conflict between cormorants and humans in the La Ronge area stems from the fact that some cormorant prey on the Egg Lake colonies came from lakes with economically important fisheries that have declined in production over time. However, sources of cormorant prey also came from lakes with no real fisheries concerns; cormorants are therefore spreading their consumption out over a variety of sites with different food webs and degrees of human influence. This should somewhat reduce the conflict between cormorants, humans, and fisheries because it suggests that no single site is going to be highly impacted. Habitat usage by humans and cormorants can overlap, although not completely, further adding to the perception that cormorants may be a threat to fish populations (Stapanian and Bur 2002). In some areas, the return of cormorants after a long period of absence from a local or regional area where cormorants used to be historically, has led to policies and management plans because humans use these areas now more intensely and are not used to seeing large numbers of cormorants (Wires et al.

2001, Wires and Cuthbert 2006).

Most cormorant radio-telemetry studies have been done on colonies situated in marine habitats (e.g. Hebshi 1998, Sapoznikow and Quintana 2003, Kotzerka et al. 2011) or large freshwater lakes (i.e. Great Lakes; e.g. Custer and Bunck 1992, Stapanian et al.

2002, Seefelt and Gillingham 2006), where cormorants are essentially surrounded by water for long distances. Few studies have investigated the movement patterns of cormorants on smaller, inland freshwater lakes, especially in North America. This is unfortunate, especially because the effects of cormorants on fisheries are insinuated to be greater on inland lakes than coastal or Great Lakes systems (Rudstam et al. 2004). In large bodies of water, fish migration and cormorant foraging over a larger area can dilute the effect of cormorant predation on fish (Rudstam et al. 2004). In addition, some studies have reported localized and minimal effects of cormorants on fish stocks in larger bodies of water (Madenjian and Gabrey 1995, Bur et al. 1999, Stapanian et al. 2002). It is suggested that cormorant impacts may be greater on inland lakes because there is little to no fish migration to offset the loss of fish (Rudstam et al. 2004). However, results like the ones from this study demonstrate that cormorants can forage on multiple lakes, which include ones that are distant from the breeding colony. Therefore, cormorants are essentially spreading their feeding efforts among different lakes and taking fish from a variety of sources, which may reduce their impact on any single site.

5. Conclusions

1) I aimed to characterize the diet of nestling double-crested cormorants in the La

Ronge area.

Fish of little economic value composed the majority of cormorant nestling diet,

providing additional support that double-crested cormorants seldom take

economically valuable fish. Cormorants nesting on Egg Lake targeted a wide

variety of prey, but the three most important fish species consumed by nestlings

were small yellow perch, coregonids (cisco and lake whitefish), and suckers.

2) I assessed the utility of carbon and nitrogen stable isotopes as intrinsic markers in

fish to track cormorant feeding locations and determined whether cormorants fed

on one or multiple sources of prey.

Linear discriminant analysis revealed classification accuracy of known location

gill-netted yellow perch to range from 69% to 86%. Although I could not assign

regurgitated yellow perch of unknown origin to a specific site 100% of the time, I

was able to determine that cormorants obtained prey from multiple locations on a

daily and seasonal basis, and they switched sites frequently on a colony-wide

basis. This indicates that biologists and managers must take into account multiple

sources of prey for more accurate fisheries impact assessments.

3) I examined lake use by cormorants in the area and described what environmental

factors may play a role in selection of foraging areas.

Cormorants foraged at all sites in the La Ronge area but mainly used areas with

relatively shallow and productive waters ≤13 m deep.

LIST OF REFERENCES

Ainley, D.G., D.W. Anderson and P.R. Kelly. 1981. Feeding ecology of marine cormorants in southwestern North America. The Condor 83: 120-131.

Anderson, C.D., D.D. Roby and K. Collis. 2004. Foraging patterns of male and female double-crested cormorants nesting in the Columbia River estuary. Canadian Journal of Zoology 82: 541-554.

Ashmole, N.P. 1963. The regulation of numbers of tropical oceanic birds. Ibis 103b: 458- 473.

Barks, P.M., J.L. Doucette and C.M. Somers. 2010. Lack of angling-sized yellow perch in a Canadian boreal lake: potential influences of growth rate, diet, and predation by double-crested cormorants. Transactions of the American Fisheries Society 139: 1029-1040.

Bartholomew, G.A. 1942. The fishing activities of double-crested cormorants on San Francisco Bay. The Condor 44: 13-21.

Bayer, R.D. 1982. How important are bird colonies as information centers? The Auk 99: 31-40.

Beamesderfer, R.C.P. 2000. Managing fish predators and competitors: deciding when intervention is effective and appropriate. Fisheries 25: 18-23.

Bearhop, S., D.R. Thompson, S. Waldron, I.C. Russell, G. Alexander and R.W. Furness. 1999. Stable isotopes indicate the extent of freshwater feeding by cormorants Phalacrocorax carbo shot at inland fisheries in England. Journal of Applied Ecology 36: 75-84.

Beaudoin, C.P., E.E. Prepas, W.M. Tonn, L.I. Wassenaar and B.G. Kotak. 2001. A stable carbon and nitrogen isotope study of lake food webs in Canada‟s Boreal Plain. Freshwater Biology 46: 465-477.

Bédard, J., A. Nadeau and M. Lepage. 1995. Double-crested cormorant culling in the St.Lawrence River estuary. Colonial Waterbirds 18 (Special Publication 1): 78-85.

Birt, V.L., T.P. Birt, D. Goulet, D.K. Cairns and W.A. Montevecchi. 1987. Ashmole‟s halo: direct evidence for prey depletion by a seabird. Marine Ecology Progress Series 40: 205-208.

Blackwell, B.F., W.B. Krohn and R.B. Allen. 1995. Foods of nestling double-crested cormorants in Penobscot Bay, Maine, USA: temporal and spatial comparisons. Colonial Waterbirds 18: 199-208.

Blackwell, B.F. and W.B. Krohn. 1997. Spring foraging distribution and habitat selection by double-crested cormorants on the Penobscot River, Maine USA. Colonial Waterbirds 20: 66-76.

Boertmann, D. and A. Mosbech. 1997. Breeding distribution and abundance of the great cormorant Phalacrocorax carbo carbo in Greenland. Polar Research 16: 93-100.

Bowen, G.J., L.I. Wassenaar and K.A. Hobson. 2005. Global application of stable hydrogen and oxygen isotopes to wildlife forensics. Oecologia 143: 337-348.

Box, G.E.P. 1954. Some theorems on quadratic forms applied in the study of analysis of variance problems, II. Effects of inequality of variance and of correlation between errors in the two-way classification. The Annals of Mathematical Statistics 25: 484- 498.

Brandl, R. and M. Gorke. 1988. How to live in colonies: foraging range and patterns of density around a colony of black-headed gulls Larus ridibundus in relation to the gulls‟ energy budget. Ornis Scandinavica 19: 305-308.

Brazner, J.C., S.E. Campana and D.K. Tanner. 2004. Habitat fingerprints for Lake Superior coastal wetlands derived from elemental analysis of yellow perch otoliths. Transactions of the American Fisheries Society 133: 692-704.

Brian, S. D., A. Moerke, M. Bur, C. Bassett, T. Aderman, D. Traynor, R.D. Singleton, P.H. Butchko and J.D. Taylor II. 2010. Evaluation of harassment of migrating double-crested cormorants to limit depredation on selected sport fisheries in Michigan. Journal of Great Lakes Research 36: 215-223.

Brown, C.R., M.B. Brown and A.R. Ives. 1992. Nest placement relative to food and its influence on the evolution of avian coloniality. The American Naturalist 139: 205- 217.

Bur, M.T., S.L. Tinnirello, C.D. Lovell and J.T. Tyson. 1999. Diet of the double-crested cormorant in western Lake Erie. Pages 73-85 in Symposium on Double-Crested Cormorants: Population Status and Management Issues in the Midwest (M.E. Tobin, Ed). Unites States Department of Agriculture Technical Bulletin 1879.

Burnett, J.A.D., N.H. Ringler, B.F. Lantry and J.H. Johnson. 2002. Double-crested cormorant predation on yellow perch in the eastern basin of Lake Ontario. Journal of Great Lakes Research 28: 202-211.

Cabana, G. and J.B. Rasmussen. 1996. Comparison of aquatic food chains using nitrogen isotopes. Proceedings of the National Academy of Sciences USA 93: 10844-10847.

Campana, S.E., J.A. Gagné and J.W. McLaren. 1995. Elemental fingerprinting of fish otoliths using ID-ICPMS. Marine Ecology Progress Series 122: 115-120.

Campana, S.E., G.A. Chouinard, J.M. Hanson, A. Fréchet and J. Brattey. 2000. Otolith elemental fingerprints as biological tracers of fish stocks. Fisheries Research 46: 343-357.

Carlander, K.D. 1969. Handbook of freshwater fishery biology, Volume 1. Iowa State University Press, Ames, Iowa, USA.

Carss, D.N. and G.R. Ekins. 2002. Further European integration: mixed sub-species colonies of great cormorants Phalacrocorax carbo in Britain – colony establishment, diet, and implications for fisheries management. Ardea 90: 23-41.

Carss, D.N. 2003. Reducing the conflict between cormorants and fisheries on a pan- European scale (REDCAFE). Vol. 1, pan-European overview. Report of a Concerted Action of European Union. Centre for Ecology and Hydrology, Banchory, UK.

Carter, H.R., A.L. Sowls, M.S. Rodway, U.W. Wilson, R.W. Lowe, G.J. McChesney, F. Gress and D.W. Anderson. 1995. Population size, trends, and conservation problems of the double-crested cormorant on the Pacific Coast of North America. Colonial Waterbirds 18 (Special Publication 1): 189-215.

Cerling, T.E., G. Wittemyer, H.B. Rasmussen, F. Vollrath, C.E. Cerling, T.J. Robinson and I. D.-Hamilton. 2006. Stable isotopes in elephant hair document migration patterns and diet changes. Proceedings of the National Academy of Sciences 103: 371-373.

Chamberlain, C.P., J.D. Blum, R.T. Holmes, X. Feng, T.W. Sherry and G.R. Graves. 1997. The use of isotope tracers for identifying populations of migratory birds. Oecologia 109: 132-141.

Chappell, M.A., V.H. Shoemaker, D.N. Janes, S.K. Maloney and T.L. Bucher. 1993. Energetics of foraging in breeding Adélie penguins. Ecology 74: 2450-2461.

Childress, R.B. and L.A. Bennun. 2001. White-breasted cormorant, Phalacrocorax carbo, breeding season timing and its control factors at a seasonally constant Kenyan lake. Ostrich 72: 149-158.

Childress, R.B., L.A. Bennun and D.M. Harper. 2002. Population changes in sympatric great and long-tailed cormorants (Phalacrocorax carbo and P. africanus): the effects of niche overlap or environmental change? Hydrobiologia 488: 163-170.

Coleman, J.T.H., M.E. Richmond, L.G. Rudstam and P.M. Mattison. 2005. Foraging location and site fidelity of the double-crested cormorant on Oneida Lake, New York. Waterbirds 28: 498-510.

Coleman, J.T.H. and M.E. Richmond. 2007. Daily foraging patterns of adult double- crested cormorants during the breeding season. Waterbirds 30: 189-198.

Cooper, J. 1985. Diving patterns of cormorants Phalacrocoracidae. Ibis 128: 562-570.

Coutin, P.C. and J. Reside. 2003. Fish predation by great cormorants, Phalacrocorax carbo carboides, in the Gippsland Lakes, south-eastern Australia. Pages 196-210 in Interactions Between Fish and Birds: Implications for Management (I.G. Cowx, Ed). Fishing News Books, Blackwell, Oxford, UK.

Cowx, I.G. (Ed) 2003. Interactions Between Fish and Birds: Implications for Management. Fishing News Books, Blackwell, Oxford, UK.

Craven, S.R. and E. Lev. 1987. Double-crested cormorants in the Apostle Islands, Wisconsin, USA: population trends, food habits, and fishery depredations. Colonial Waterbirds 10: 64-71.

Cushing, C.E., Jr. 1964. Plankton and water chemistry in the Montreal River lake-stream system, Saskatchewan. Ecology 45: 306-313.

Custer, T.W. and C. Bunck. 1992. Feeding flights of breeding double-crested cormorants at two Wisconsin colonies. Journal of Field Ornithology 63: 203-211.

Dalton, C.M., D. Ellis and D.M. Post. 2009. The impact of double-crested cormorant (Phalacrocorax auritus) predation on anadromous alewife (Alosa pseudoharengus) in south-central Connecticut, USA. Canadian Journal of Fisheries and Aquatic Sciences 66: 177-186. del Giorgio, P.A. and R.L. France. 1996. Ecosystem-specific patterns in the relationship between zooplankton and POM or microplankton δ13C. Limnology and Oceanography 41: 359-365.

DeNiro, M.J. and S. Epstein. 1978. Influence of diet on the distribution of carbon isotopes in animals. Geochimica et Cosmochimica Acta 42: 495-506.

Derby, C.E. and J.R. Lovvorn. 1997. Predation on fish by cormorants and pelicans in a cold-water river: a field and modeling study. Canadian Journal of Fisheries and Aquatic Sciences 54: 1480-1493.

Derek Murray Consulting Associates. 2006. Economic evaluation of Saskatchewan‟s commercial and non-outfitted sport fishing. Saskatchewan Ministry of Environment, Regina, Saskatchewan.

Diana, J.S., S. Maruca and B. Low. 2006. Do increasing cormorant populations threaten sportfishes in the Great Lakes? A case study in Lake Huron. Journal of Great Lakes Research 32: 306-320.

Dirksen, S., T.J. Boudewijn, R. Noordhuis and E.C.L. Marteijn. 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.

Dorr, B.S., T. Aderman, P.H. Butchko and S.C. Barras. 2010. Management effects on breeding and foraging numbers and movements of double-crested cormorants in the Les Cheneaux Islands, Lake Huron, Michigan. Journal of Great Lakes Research 36: 224-231.

Doucette, J.L., V.A. Kjoss and C.M. Somers. 2008. Size and composition of foraging flights in two species of piscivorous colonial birds: limited evidence for intra- or interspecific information transfer. The Open Ornithology 1: 48-56.

Doucette, J.L., A. Bugajski and M. Duffy. 2010. Double-crested cormorants expand their breeding range onto the Boreal Shield in Saskatchewan. Blue Jay 68: 71-74.

Doucette, J.L., B. Wissel, and C.M. Somers. In Press. Cormorant-fisheries conflicts: stable isotopes reveal a consistent niche for avian piscivores in diverse food webs. Ecological Applications. [doi:10.1890/10-2384.1]

Downing, J.A. and C. Plante. 1993. Production of fish populations in lakes. Canadian Journal of Fisheries and Aquatic Sciences 50: 110-120.

Draulans, D. 1987. The effectiveness of attempts to reduce predation by fish-eating birds: a review. Biological Conservation 41: 219-232.

Draulans, D. 1988. Effects of fish-eating birds on freshwater fish stocks: an evaluation. Biological Conservation 44: 251-263.

Duffy, D.C. 1992. Biodiversity and research on seabirds. Colonial Waterbirds 15: 155- 158.

Duffy, D.C. 1995. Why is the double-crested cormorant a problem? Insights from cormorant ecology and human sociology. Colonial Waterbirds 18 (Special Publication 1): 25-32.

Duffy, M. 2007. Lac La Ronge endorsement surveys 1997-2005. Fish and Wildlife Technical Report 2007-3. Saskatchewan Ministry of Environment, Regina, Saskatchewan. 33p.

Duffy, M. 2008. Assessment of Nemeiben Lake fish stocks, 2003. Fish and Wildlife Technical Report 2008-2. Saskatchewan Ministry of Environment, Regina, Saskatchewan. 39p.

Dunn, E.H. 1975. Caloric intake of nestling double-crested cormorants. The Auk 92: 553- 565.

Elliott, K.H., K.J. Woo, A.J. Gaston, S. Benvenuti, L. Dall‟Antonia and G.K. Davoren. 2009. Central-place foraging in an Arctic seabird provides evidence for Storer- Ashmole‟s halo. The Auk 126: 613-625.

Engström, H. 2001. Long term effects of cormorant predation on fish communities and fishery in a freshwater lake. Ecography 24: 127-138.

Enstipp, M.R., D. Grémillet and D.R. Jones. 2006a. The effects of depth, temperature and food ingestion on the foraging energetics of a diving endotherm, the double-crested cormorant (Phalacrocorax auritus). The Journal of Experimental Biology 209: 845- 859.

Enstipp, M.R., D. Grémillet and D.R. Jones. 2006b. Heat increment of feeding in double- crested cormorants (Phalacrocorax auritus) and its potential for thermal substitution. The Journal of Experimental Biology 211: 49-57.

Eisenhower, M.D. and D.L. Parrish. 2009. Double-crested cormorant and fish interactions in a shallow basin of Lake Champlain. Waterbirds 32: 388-399.

Evans, R.M. 1982. Efficient use of food patches at different distances from a breeding colony in black-billed gulls. Behaviour 79: 28-38.

Evans, R.D. 2001. Physiological mechanisms influencing plant nitrogen isotope composition. Trends in Plant Science 6: 121-126.

Ewins, P.J. and D.V. Weseloh. 1994. Effects on productivity of shooting of double- crested cormorants (Phalacrocorax auritus) on Pigeon Island, Lake Ontario, in 1993. Journal of Great Lakes Research 20: 761-767.

Feltham, M.J. 1995. Predation of Atlantic salmon, Salmo salar L., smolts and parr by red-breasted mergansers, Mergus serrator L., on two Scottish rivers. Fisheries Management and Ecology 2: 289-298.

Fisher, S.J., D.W. Willis and K.L. Pope. 1996. An assessment of burbot (Lota lota) weight-length data from North American populations. Canadian Journal of Zoology 74: 570-575.

Fowle, M.R. 1997. Population dynamics, food habits, and bioenergetics of double-crested cormorants in Lake Champlain. M.Sc. Thesis, University of Vermont, Burlington, Vermont.

France, R.L. 1995. Differentiation between littoral and pelagic food webs in lakes using stable carbon isotopes. Limnology and Oceanography 40: 1310-1313. Frederiksen, M., J.-D. Lebreton and T. Bregnballe. 2001. The interplay between culling and density-dependence in the great cormorant: a modeling approach. Journal of Applied Ecology 38: 617-627.

Frederiksen, M., S. Wanless, M.P. Harris, P. Rothery and L.J. Wilson. 2004. The role of industrial fisheries and oceanographic change in the decline of North Sea black- legged kittiwakes. Journal of Applied Ecology 41: 1129-1139.

Gandini, P., E. Frere and F. Quintana. 2005. Feeding performance and foraging area of the red-legged cormorant. Waterbirds 28: 41-45.

Gaston, A.J., R.C. Ydenberg and G.E.J. Smith. 2007. Ashmole‟s halo and population regulation in seabirds. Marine Ornithology 35: 119-126.

Gillanders, B.M. and M.J. Kingsford. 2000. Elemental fingerprints of otoliths of fish may distinguish estuarine „nursery‟ habitats. Marine Ecology Progress Series 201: 273- 286.

Gilmour, C.C., E.A. Henry and R. Mitchell. 1992. Sulfate stimulation of mercury methylation in freshwater sediments. Environmental Science and Technology 26: 2281-2287.

Glahn, J.F. and K.E. Brugger. 1995. The impact of double-crested cormorants on the Mississippi Delta catfish industry: a bioenergetics model. Colonial Waterbirds 18 (Special Publication 1): 168-175.

Glahn, J.F., P.J. Dixson, G.A. Littauer and R.B. McCoy. 1995. Food habits of double- crested cormorants wintering in the delta region of Mississippi. Colonial Waterbirds 18 (Special Publication 1): 158-167.

Gloutney, B.L. and M.Y. Chen. 1992. Status of the Lac La Ronge fishery 1989 and 1990. Fisheries Technical Report 92-2. Saskatchewan Ministry of Environment, Regina, Saskatchewan. 88p.

Gorke, M. and R. Brandl. 1986. How to live in colonies: spatial foraging strategies of the black-headed gull. Oecologia 70: 288-290.

Gould, P.J. and D.J. Forsell. 1989. Techniques for shipboard surveys of marine birds. Fish and Wildlife Technical Report 25. United States Department of the Interior Fish and Wildlife Service, Washington, DC. 22p.

Greenfield, B.K., T.R. Hrabik, C.J. Harvey and S.R. Carpenter. 2001. Predicting mercury levels in yellow perch: use of water chemistry, trophic ecology, and spatial traits. Canadian Journal of Fisheries and Aquatic Sciences 58: 1419-1429.

Grémillet, D., D. Schmid and B. Culik. 1995. Energy requirements of breeding great cormorants Phalacrocorax carbo sinensis. Marine Ecology Progress Series 121: 1-9.

Grémillet, D. 1997. Catch per unit effort, foraging efficiency, and parental investment in breeding great cormorants (Phalacrocorax carbo carbo). ICES Journal of Marine Science 54: 635-644.

Grémillet, D. and R.P. Wilson. 1999. A life in the fast lane: energetics and foraging strategies of the great cormorant. Behavioral Ecology 10: 516-524.

Grémillet, D., G. Wright, A. Lauder, D.N. Carss and S. Wanless. 2003. Modelling the daily food requirements of wintering great cormorants: a bioenergetics tool for wildlife management. Journal of Applied Ecology 40: 266-277.

Grémillet, D., C. Chauvin, R.P. Wilson, Y. le Maho and S. Wanless. 2005a. Unusual feather structure allows partial plumage wettability in diving great cormorants Phalacrocorax carbo. Journal of Avian Biology 36: 57-63.

Grémillet, D., G. Kuntz, C. Gilbert, A. J. Woakes, P.J. Butler and Y. le Maho. 2005b. Cormorants dive through the Polar night. Biology Letters 1: 469-471.

Gu, B. and V. Alexander. 1993. Dissolved nitrogen uptake by cyanobacterial bloom (Anabaena flos-aquae) in a subarctic lake. Applied and Environmental Microbiology 59: 422-430.

Hamilton, W.J., W.M. Gilbert, F.H. Heppner and R.J. Planck. 1967. Starling roost dispersal and a hypothetical mechanism regulating rhythmical animal movement to and from dispersal centers. Ecology 48: 825-833.

Hao, W. 2008. Spatial analysis of the impact of dredging induced increased water turbidity on the visual acuity of great cormorants in Poyang Lake, China. M.Sc. Thesis, Wuhan University, Wuhan, China.

Hartman, K.J. and F.J. Margraf. 1992. Evidence of predatory control of yellow perch (Perca flavescens) recruitment in Lake Erie USA. Journal of Fish Biology 43: 109- 119.

Hatch, J.J. 1995. Changing populations of double-crested cormorants. Colonial Waterbirds 18 (Special Publication 1): 8-24.

Hatch, J.J. and D.V. Weseloh. 1999. Double-crested cormorant (Phalacrocorax auritus). In The Birds of North America Online (A. Poole, Ed). Ithaca: Cornell Lab of Ornithology. http://bna.birds.cornell.edu/bna/species/441doi:10.2173/bna.441

Heaton, T.H.E. 1986. Isotopic studies of nitrogen pollution in the hydrosphere and atmosphere: a review. Chemical Geology (Isotope Geoscience Section) 59: 87-102.

Hebert, C.E., M. Bur, D. Sherman and J.L. Shutt. 2008. Sulfur isotopes link overwinter habitat use and breeding condition in double-crested cormorants. Ecological Applications 18: 561-567.

Hebshi, A. 1998. Foraging site preference of Brandt‟s cormorants off the Santa Cruz, California, Coast. Colonial Waterbirds 21: 245-250.

Hesslein, R.H., M.J. Capel, D.E. Fox and K.A. Hallard. 1991. Stable isotopes of sulfur, carbon, and nitrogen as indicators of trophic level and fish migration in the lower Mackenzie River basin, Canada. Canadian Journal of Fisheries and Aquatic Sciences 48: 2258-2265.

Hesslein, R.H., K.A. Hallard and P. Ramlal. 1993. Replacement of sulfur, carbon, and nitrogen in tissue of growing broad whitefish (Coregonus nasus) in response to a change in diet traced by δ34S, δ13C, and δ15N. Canadian Journal of Fisheries and Aquatic Sciences 50: 2071-2076.

Hobson, K.A., R.W. Knapton and W. Lysack. 1989. Population, diet and reproductive success of double-crested cormorants breeding on Lake Winnipegosis, Manitoba, in 1987. Colonial Waterbirds 12: 191-197.

Hobson, K.A. 1999. Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia 120: 314-326.

Hobson, K.A., L.I. Wassenaar and O.R. Taylor. 1999. Stable isotopes (δD and δ13C) are geographical indicators of natal origins of monarch butterflies in eastern North America. Oecologia 120: 397-404.

Horn, H.S. 1968. The adaptive significance of colonial nesting in the Brewer‟s blackbird (Euphagus cyanocephalus). Ecology 49: 682-694.

Johnson, J.H., R.M. Ross and R.D. McCullough. 2002. Little Galloo Island, Lake Ontario: a review of nine years of double-crested cormorant diet and fish consumption information. Journal of Great Lakes Research 28: 182-192.

Jones, R.I., L. King, M.M. Dent, S.C. Maberly and C.E. Gibson. 2004. Nitrogen stable isotope ratios in surface sediments, epilithon and macrophytes from upland lakes with differing nutrient status. Freshwater Biology 49: 382-391.

Keith, J.A. 1995. Management policies for cormorants in Canada. Colonial Waterbirds 18 (Special Publication 1): 234-237.

Kirby, J.S., J.S. Holmes and R.M. Sellers. 1996. Cormorants Phalacrocorax carbo as fish predators: an appraisal of their conservation and management in Great Britain. Biological Conservation 75: 191-199.

Kiriluk, R.M., M.R. Servos, D.M. Whittle, G. Cabana and J.B. Rasmussen. 1995. Using ratios of stable nitrogen and carbon isotopes to characterize the biomagnifications of DDE, mirex, and PCB in a Lake Ontario pelagic food web. Canadian Journal of Fisheries and Aquatic Sciences 52: 2660-2674.

Kling, G.W., B. Fry and W.J. O‟Brien. 1992. Stable isotopes and planktonic trophic structure in Arctic lakes. Ecology 73: 561-566.

Kotzerka, J., S.A. Hatch and S. Garthe. 2011. Evidence for foraging-site fidelity and individual foraging behavior of pelagic cormorants rearing chicks in the Gulf of Alaska. The Condor 113: 80-88.

Krohn, W.B., R.B. Allen, J.R. Moring and A.E. Hutchinson. 1995. Double-crested cormorants in New England: population and management histories. Colonial Waterbirds 18 (Special Publication 1): 99-109.

Kuiken, T., G. Wobeser, F.A. Leighton, D.M. Haines, B. Chelack, J. Bogdan, L. Hassard, R.A. Heckert and J. Riva. 1999. Pathology of Newcastle disease in double-crested cormorants from Saskatchewan, with comparison of diagnostic methods. Journal of Wildlife Disease 35: 8-23.

Lake, J.L., R.A. McKinney, F.A. Osterman, R.J. Pruell, J. Kiddon, S.A. Ryba and A.D. Libby. 2001. Stable nitrogen isotopes as indicators of anthropogenic activities in small freshwater systems. Canadian Journal of Fisheries and Aquatic Sciences 58: 870-878.

Lantry, B.F., T.H. Eckert, and C.P. Schneider and J.R. Chrisman. 2002. The relationship between the abundance of smallmouth bass and double-crested cormorants in the eastern basin of Lake Ontario. Journal of Great Lakes Research 28: 193-201.

Larson, J.M., W.H. Karasov, L. Sileo, K.L. Stromborg, B.A. Hanbidge, J.P. Giesy, P.D. Jones, D.E. Tillitt and D.A. Verbrugge. 1996. Reproductive success, developmental anomalies, and environmental contaminants in double-crested cormorants (Phalacrocorax auritus). Environmental Toxicology and Chemistry 15: 553-559.

Li, Y., J-H. Jiang, Z-P. Chen, C-J. Xu and R-Q. Yu. 1999. Robust linear discriminant analysis for chemical pattern recognition. Journal of Chemometrics 13: 3-13.

Linn, I.J. and K.L.I. Campbell. 1992. Interactions between white-breasted cormorants Phalacrocorax carbo (Aves: Phalacrocoracidae) and the fisheries of Lake Malawi. Journal of Applied Ecology 29: 619-634.

Liordos, V. and V. Goutner. 2008. Habitat and temporal variation in diet of great cormorant nestlings in Greek colonies. Waterbirds 31: 424-437.

Lovvorn, J.R. and D.R. Jones. 1991. Body mass, volume, and buoyancy of some aquatic birds, and their relation to locomotor strategies. Canadian Journal of Zoology 69: 2888-2892.

Ludwig, J.P., H.J. Auman, D.V. Weseloh, G.A. Fox, J.P. Giesy and M.E. Ludwig. 1995. Evaluation of the effects of toxic chemicals in Great Lakes cormorants: has causality been established? Colonial Waterbirds 18 (Special Publication 1): 60-69.

Madenjian, C.P. and S.W. Gabrey. 1995. Waterbird predation on fish in western Lake Erie: a bioenergetics model application. The Condor 97: 141-153.

Mattingley, H.E. 1927. Cormorants in relation to fisheries. The Condor 29: 182-187.

Mech, L.D. 1983. Handbook of animal radio-tracking. University of Minnesota Press, Minneapolis, Minnesota.

Milton, G.R., P.J. Austin-Smith and G.J. Farmer. 1995. Shouting at shags: a case study of cormorant management in Nova Scotia. Colonial Waterbirds 18 (Special Publication 1): 91-98.

Modde, T., A.F. Wasowicz and D.K. Hepworth. 1996. Cormorant and grebe predation on rainbow trout stocked in a southern Utah reservoir. North American Journal of Fisheries Management 16: 388-394.

Moerbeek, D.J., W.H. van Dobben, E.R. Osieck, G.C. Boere and C.M. Bungenberg de Jong. 1987. Cormorant damage prevention at a fish farm in the Netherlands. Biological Conservation 39: 23-38.

Monaghan, P. 1992. Seabirds and sandeels: the conflict between exploitation and conservation in the northern North Sea. Biodiversity and Conservation 1: 98-111.

Mott, D.F. and F.L. Boyd. 1995. A review of techniques for preventing cormorant depredations at aquaculture facilities in the southeastern United States. Colonial Waterbirds 18 (Special Publication 1): 176-180.

Mous, P.J., W. Dekker, J.J. De Leeuw, M.R. Van Eerden and W.L.T. Van Densen. 2003. Interactions in the utilisation of small fish by piscivorous fish and birds, and the fishery in IJsselmeer. Pages 84-118 in Interactions Between Fish and Birds: Implications for Management (I.G. Cowx, Ed). Fishing News Books, Blackwell, Oxford, UK.

Murchie, K.J. and M. Power. 2004. Growth- and feeding-related isotopic dilution and enrichment patterns in young-of-the-year yellow perch (Perca flavescens). Freshwater Biology 49: 41-54.

Murray, T.E., J.A. Bartle, S.R. Kalish and P.R. Taylor. 1993. Incidental capture of seabirds by Japanese southern bluefin tuna longline vessels in New Zealand waters, 1988-1992. Bird Conservation International 3: 181-210.

Nagy, K.A., W.R. Siegfried and R.P. Wilson. 1984. Energy utilization by free-ranging jackass penguins, Spheniscus demersus. Ecology 65: 1648-1655.

Nelson, J.B. 2005. Pelicans, cormorants, and their relatives (Pelecanidae, Sulidae, Phalacrocoracidae, Anhingidae, Fregatidae, Phaethontidae). Oxford University Press, Oxford, UK.

Neuman, J., D.L. Pearl, P.J. Ewins, R. Black, D.V. Weseloh, M. Pike and K. Karwowski. 1997. Spatial and temporal variation in the diet of double-crested cormorants (Phalacrocorax auritus) breeding on the lower Great Lakes in the early 1990s. Canadian Journal of Fisheries and Aquatic Sciences 54: 1569-1584.

Obst, B.S., R.W. Russell, G.L. Hunt Jr., Z.A. Eppley and N.M. Harrison. 1995. Foraging radii and energetics of least auklets (Aethia pusilla) breeding on three Bering Sea islands. Physiological Zoology 68: 647-672.

O‟Gorman, R., E.L. Mills and J. DeGisi. 1991. Use of zooplankton to assess the movement and distribution of alewife (Alosa pseudoharengus) in south-central Lake Ontario in spring. Canadian Journal of Fisheries and Aquatic Sciences 48: 2250-2257.

Ottenbacher, M.J., D.K. Hepworth and L.N. Berg. 1994. Observations on double-crested cormorants (Phalacrocorax auritus) at sportfishing waters in southwestern Utah. Great Basin Naturalist 54: 272-286.

Overman, N.C. and D.L. Parrish. 2001. Stable isotope composition of walleye: 15N accumulation with age and area-specific differences in δ13C. Canadian Journal of Fisheries and Aquatic Sciences 58: 1253-1260.

Owre, O.T. 1967. Adaptations for locomotion and feeding in the anhinga and the double- crested cormorant. Ornithological Monographs 6: 1-138.

Parrott, D., H.V. McKay, G.V. Watola, J.D. Bishop and S. Langton. 2003. Effects of short-term shooting program on nonbreeding cormorants at inland fisheries. Wildlife Society Bulletin 31: 1092-1098.

Pennycuick, C.J. 1978. Fifteen testable predictions about bird flight. Oikos 30 (Proceedings of a Symposium at Falsterbo, Sweden, 3-8 October): 165-176.

Peterson, B.J. and B. Fry. 1987. Stable isotopes in ecosystem studies. Annual Review of Ecological Systems 18: 293-320.

Platteeuw, M., and K. Koffijberg and W. Dubbeldam. 1995. Growth of cormorant Phalacrocorax carbo sinensis chicks in relation to brood size, age ranking and parental fishing effort. Ardea 83: 235-245.

Post, D.M. 2002. Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83: 703-718.

Post, J.R., M. Sullivan, S. Cox, N.P. Lester, C.J. Walters, E.A. Parkinson, A.J. Paul, L. Jackson and B.J. Shuter. 2002. Canada‟s recreational fisheries: the invisible collapse? Fisheries 27: 6-17.

Punta, G., P. Yorio and G. Herrera. 2003. Temporal patterns in the diet and food partitioning in imperial cormorants (Phalacrocorax atriceps) and rock shags (P. magellanicus) breeding at Bahia Bustamante, Argentina. The Wilson Bulletin 115: 307-315.

Pūtys, Ž. and J. Zarankaitė. 2010. Diet of the great cormorant (Phalacrocorax carbo sinensis) at the Juodkrantė colony, Lithuania. Acta Zoologica Lituanica 20: 179- 189.

Quevedo, M., R. Svanbäck and P. Eklöv. 2009. Intrapopulation niche partitioning in a generalist predator limits food web connectivity. Ecology 90: 2263-2274.

Quintana, F. 2001. Foraging behaviour and feeding locations of rock shags Phalacrocorax magellanicus from a colony in Patagonia, Argentina. Ibis 143: 547-553.

Quintana, F., P. Yorio, N. Lisnizer, A. Gatto and G. Soria. 2004. Diving behavior and foraging areas of the neotropic cormorant at a marine colony in Patagonia, Argentina. The Wilson Bulletin 116: 83-88.

Radomski, A.A. and P.V. Zimba. 2010. Does pond water reflectance influence double- crested cormorant selection of aquaculture ponds? Journal of the World Aquaculture Society 41: 430-437.

Ribak, G., D. Weihs and Z. Arad. 2005. Water retention in the plumage of diving great cormorants Phalacrocorax carbo sinensis. Journal of Avian Biology 36: 89-95.

Ridgeway, M.S., J.B. Pollard and D.V. Weseloh. 2006. Density-dependent growth of double-crested cormorant colonies on Lake Huron. Canadian Journal of Zoology 84: 1409-1420.

Ridgeway, M.S. 2010. A review of estimates of daily energy expenditure and food intake in cormorants (Phalacrocorax spp.). Journal of Great Lakes Research 36: 93-99.

Robertson, I. 1974. The food of nestling double-crested and pelagic cormorants at Mandarte Island, British Columbia, with notes on feeding ecology. The Condor 76: 346-348.

Rubenstein, D.R. and K.A. Hobson. 2004. From birds to butterflies: animal movement patterns and stable isotopes. Trends in Ecology and Evolution 19: 256-263.

Rudstam, L.G., A.J. VanDeValk, C.M. Adams, J.T.H. Coleman, J.L. Forney and M.E. Richmond. 2004. Cormorant predation and the population dynamics of walleye and yellow perch in Oneida Lake. Ecological Applications 14: 149-163.

Russell, I., D. Parrott, M. Ives, D. Goldsmith, S. Fox, D. Clifton-Dey, A. Prickett and T. Drew. 2008. Reducing fish losses to cormorants using artificial fish refuges: an experimental study. Fisheries Management and Ecology 15: 189-198.

Sapoznikow, A. and F. Quintana. 2003. Foraging behavior and feeding locations of imperial cormorants and rock shags breeding sympatrically in Patagonia, Argentina. Waterbirds 26: 184-191.

Saskatchewan Ministry of Environment. 2003. La Ronge integrated land use management plan, background document. Regina, Saskatchewan.

Seefelt, N.E. and J.C. Gillingham. 2006. Foraging locations of double-crested cormorants in the Beaver Archipelago of northern Lake Michigan: potential for impacts on smallmouth bass. Waterbirds 29: 473-480.

Shwiff, S., K. Kirkpatric and T. DeVault. 2009. The economic impact of double-crested cormorants to central New York. New York Wildlife Services Report. National Wildlife Research Center, Colorado.

Somers, C.M., V.A. Kjoss, F.A. Leighton and D. Fransden. 2010. American white pelicans and double-crested cormorants in Saskatchewan: population trends over five decades. Blue Jay 68: 75-86.

100

Stapanian, M.A. and M.T. Bur. 2002. Overlap in offshore habitat use by double-crested cormorants and boaters in western Lake Erie. Journal of Great Lakes Research 28: 172-181.

Stapanian, M.A., M.T. Bur, J.T. Tyson, T.W. Seamans and B.F. Blackwell. 2002. Foraging locations of double-crested cormorants on western Lake Erie: site characteristics and spatial associations with prey fish densities. Journal of Great Lakes Research 28: 155-171.

Stendera, S. and R.K. Johnson. 2006. Multiscale drivers of water chemistry of boreal lakes and streams. Environmental Management 38: 760-770.

Stickley, A.R., Jr., G.L. Warrick and J.F. Glahn. 1992. Impact of double-crested cormorant depredations on channel catfish farms. Journal of the World Aquaculture Society 23: 192-198.

Strod, T., I. Izhaki, Z. Arad and G. Katzir. 2008. Prey detection by great cormorant (Phalacrocorax carbo sinensis) in clear and in turbid water. The Journal of Experimental Biology 211: 866-872.

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.

Takahashi, T., K. Kameda, M. Kawamura and T. Nakajima. 2006. Food habits of great cormorant Phalacrocorax carbo hanedae at Lake Biwa, Japan, with special reference to ayu Plecoglossus altivelis altivelis. Fisheries Science 72: 477-484.

Tasker, M.L., C. J. Camphuysen, J. Cooper, S. Garthe, W.A. Montevecchi and S.J.M. Blaber. 2000. The impacts of fishing on marine birds. ICES Journal of Marine Science 57: 531-547.

Taylor, H.P, Jr. 1974. The application of oxygen and hydrogen isotope studies to problems of hydrothermal alteration and ore deposition. Economic Geology 69: 843-888.

Thorrold, S.R., C.M. Jones, P.K. Swart and T.E. Targett. 1998. Accurate classification of juvenile weakfish Cynoscion regalis to estuarine nursery areas based on chemical signatures in otoliths. Marine Ecology Progress Series 173: 253-265.

Trapp, J.L., T.J. Dwyer, J.J. Doggett and J.G. Nickum. 1995. Management responsibilities and policies for cormorants: United States Fish and Wildlife Service. Colonial Waterbirds 18 (Special Publication 1): 226-230.

101

Trayler, K.M., D.J. Brothers, R.D. Wooller and I.C. Potter. 1989. Opportunistic foraging by three species of cormorants in an Australian estuary. Journal of Zoology, London 218: 87-98.

U.S. Fish and Wildlife Service. 2003. Final environmental impact statement (DEIS) on double-crested cormorant management. Washington, D.C.

Vander Zanden, M.J. and J.B. Rasmussen. 1999. Primary consumer δ13C and δ15N and the trophic position of aquatic consumers. Ecology 80: 1395-1404.

VanDeValk, A.J., C.M. Adams, L.G. Rudstam, J.L. Forney, T.E. Brooking, M.A. Gerken, B.P. Young and J.T. Hooper. 2002. Comparison of angler and cormorant harvest of walleye and yellow perch in Oneida Lake, New York. Transactions of the American Fisheries Society 131: 27-39.

Van Eerden, M.R. and B. Voslamber. 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.

Vermeer, K. and L. Rankin. 1984. Population trends in nesting double-crested and pelagic cormorants in Canada. The Murrelet 65: 1-9.

Voslamber, B., M. Platteeuw and M.R. Van Eerden. 1995. Cormorants Phalacrocorax carbo sinensis: does specialized knowledge about fishing sites and fish behaviour pay off? Ardea 83: 213-222.

Vuorio, K., M. Meili and J. Sarvala. 2006. Taxon-specific variation in the stable isotopic signatures (δ13C and δ15N) of lake phytoplankton. Freshwater Biology 51: 807- 822.

Wachniew, P. and K. Rόżański. 1997. Carbon budget of a mid-latitude, groundwater- controlled lake: isotopic evidence for the importance of dissolved inorganic carbon recycling. Geochimica et Cosmochimica Acta 61: 2453-2465.

Wallace, R.G. 1993. Nemeiben Lake: walleye, pike, and trout fisheries, 1978-1992. Fisheries Technical Report 93-5. Saskatchewan Ministry of Environment, Regina, Saskatchewan. 79p.

Waltz, E.C. 1987. A test of the information-centre hypothesis in two colonies of common terns, Sterna hirundo. Animal Behaviour 35: 48-59.

Wanless, S., A.E. Burger and M.P. Harris. 1991. Diving depths of shags Phalacrocorax aristotelis breeding on the Isle of May. Ibis 133: 37-42.

102

Wanless, S. and M.P. Harris. 1993. Use of mutually exclusive foraging areas by adjacent colonies of blue-eyed shags (Phalacrocorax atriceps) at South Georgia. Colonial Waterbirds 16: 176-182.

Ward, M.C., D.W. Willis, B.R. Herwig, S.R. Chipps, B.G. Parsons, J.R. Reed and M.A. Hanson. 2008. Consumption estimates of walleye stocked as fry to suppress fathead minnow populations in west-central Minnesota wetlands. Ecology of Freshwater fish 17: 59-70.

Ward, P. and A. Zahavi. 1973. The importance of certain assemblages of birds as “information-centres” for food-finding. Ibis 115: 517-534.

Warke, G.M.A., K.R. Day, J.E. Greer and R.D. Davidson. 1994. Cormorant (Phalacrocorax carbo [L]) populations and patterns of abundance at breeding and feeding sites in Northern Ireland, with particular reference to Lough Neagh. Hydrobiologia 279/280: 91-100.

Warke, G.M.A. and K.R. Day. 1995. Changes in abundance of cyprinid and percid prey affect rate of predation by cormorants Phalacrocorax carbo carbo on salmon Salmo salar smolt in Northern Ireland. Ardea 83: 157-166.

Watkins, B.P., S.L. Petersen and P.G. Ryan. 2008. Interactions between seabirds and deep-water hake trawl gear: an assessment of impacts in South African waters. Animal Conservation 11: 247-254.

Weseloh D.V., S.M. Teeple and M. Gilbertson. 1983. Double-crested cormorants of the Great Lakes: egg-laying parameters, reproductive failure, and contaminant residues in eggs, Lake Huron 1972-1973. Canadian Journal of Zoology 61: 427- 436.

Weseloh, D.V. and P.J. Ewins. 1994. Characteristics of a rapidly increasing colony of double-crested cormorants (Phalacrocorax auritus) in Lake Ontario: population size, reproductive parameters and band recoveries. Journal of Great Lakes Research 20: 443-456.

Weseloh, D.V., P.J. Ewins, J. Struger, P. Mineau, C.A. Bishop, S. Postupalsky and J.P. Ludwig. 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.

Whiteside, M.C., C.M. Swindoll and W.L. Doolittle. 1985. Factors affecting the early life history of yellow perch, Perca flavescens. Environmental Biology of Fishes 12: 47-56.

103

Willis, D.W. 1989. Proposed standard length-weight equation for Northern Pike. North American Journal of Fisheries Management 9: 203-208.

Wires, L.R., F.J. Cuthbert, D.R. Trexel and A.R. Joshi. 2001. Status of the double-crested cormorant (Phalacrocorax auritus) in North America. Final Report to U.S. Fish and Wildlife Service. University of Minnesota, St. Paul, Minnesota.

Wires, L.R. and F.J. Cuthbert. 2006. Historic populations of the double-crested cormorant (Phalacrocorax auritus): implications for conservation and management in the 21st century. Waterbirds 29: 9-37.

Wires, L.R. and F.J. Cuthbert. 2010. Characteristics of double-crested cormorant colonies in the U.S. Great Lakes island landscape. Journal of Great Lakes Research 36: 232-241.

Wittenberger, J.F. and M.B. Dollinger. 1984. The effect of acentric colony location on the energetic of avian coloniality. The American Naturalist 124: 189-204.

Ydenberg, R.C., L-A. Giraldeau and D.L. Kramer. 1986. Interference competition, payoff asymmetries, and the social relationships of central place foragers. Theoretical Population Biology 30: 26-44.

Žydelis, R. and A. Kontautas. 2008. Piscivorous birds as top predators and fishery competitors in the lagoon ecosystem. Hydrobiologia 611: 45-54.

104

APPENDIX A

PREDICTED Bigstone Egg Nemeiben North LLR Central LLR ACTUAL Model 2 Bigstone (n=29) 26 (0.90) 2 (0.07) 1 (0.03) 0 (0.00) 0 (0.00) Egg (n=30) 0 (0.00) 30 (1.00) 0 (0.00) 0 (0.00) 0 (0.00) Nemeiben (n=20) 6 (0.30) 5 (0.25) 0 (0.00) 4 (0.20) 5 (0.25) North LLR (n=22) 0 (0.00) 7 (0.32) 0 (0.00) 15 (0.68) 0 (0.00) Central LLR (n=21) 0 (0.00) 4 (0.19) 0 (0.00) 1 (0.05) 16 (0.76) Model 3 Bigstone (n=29) 26 (0.90) 3 (0.10) NA 0 (0.00) 0 (0.00) Egg (n=30) 0 (0.00) 30 (1.00) NA 0 (0.00) 0 (0.00) Nemeiben (n=20) NA NA NA NA NA North LLR (n=22) 0 (0.00) 5 (0.23) NA 16 (0.73) 1 (0.05) Central LLR (n=21) 0 (0.00) 4 (0.19) NA 1 (0.05) 16 (0.76)

105

APPENDIX B

PREDICTED Lac La Central East West Bigstone Egg Ronge Nemeiben Nemeiben Nemeiben Head ACTUAL Model 2 Bigstone (n=3) 0 (0.00) 0 (0.00) 0 (0.00) NA 3 (1.00) 0 (0.00) 0 (0.00) Egg (n=6) 0 (0.00) 4 (0.68) 2 (0.33) NA 0 (0.00) 0 (0.00) 0 (0.00) Lac La Ronge (n=47) 0 (0.00) 0 (0.00) 45 (0.96) NA 0 (0.00) 2 (0.04) 0 (0.00) Central Nemeiben (n=17) NA NA NA NA NA NA NA East Nemeiben (n=13) 1 (0.08) 0 (0.00) 0 (0.00) NA 12 (0.92) 0 (0.00) 0 (0.00) West Nemeiben (n=24) 0 (0.00) 0 (0.00) 7 (0.29) NA 0 (0.00) 17 (0.71) 0 (0.00) Head (n=9) 0 (0.00) 0 (0.00) 0 (0.00) NA 0 (0.00) 0 (0.00) 9 (1.00) Model 3 Bigstone (n=20) 16 (0.80) 1 (0.05) 0 (0.00) 0 (0.00) 3 (0.15) 0 (0.00) 0 (0.00) Egg (n=16) 1 (0.06) 6 (0.38) 9 (0.56) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) Lac La Ronge (n=47) 0 (0.00) 1 (0.02) 45 (0.96) 0 (0.00) 0 (0.00) 1 (0.02) 0 (0.00) Central Nemeiben (n=17) 4 (0.24) 2 (0.12) 5 (0.29) 1 (0.06) 0 (0.00) 5 (0.29) 0 (0.00) East Nemeiben (n=13) 2 (0.15) 0 (0.00) 0 (0.00) 0 (0.00) 11 (0.85) 0 (0.00) 0 (0.00) West Nemeiben (n=24) 0 (0.00) 0 (0.00) 9 (0.38) 0 (0.00) 0 (0.00) 15 (0.63) 0 (0.00) Head (n=9) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 9 (1.00) Model 4 Bigstone (n=3) NA NA NA NA NA NA NA Egg (n=6) NA 1 (0.17) 3 (0.50) 2 (0.33) 0 (0.00) 0 (0.00) 0 (0.00) Lac La Ronge (n=47) NA 0 (0.00) 44 (0.94) 2 (0.04) 0 (0.00) 1 (0.02) 0 (0.00) Central Nemeiben (n=17) NA 4 (0.24) 5 (0.29) 1 (0.06) 2 (0.12) 5 (0.29) 0 (0.00) East Nemeiben (n=13) NA 0 (0.00) 0 (0.00) 0 (0.00) 13 (1.00) 0 (0.00) 0 (0.00) West Nemeiben (n=24) NA 0 (0.00) 10 (0.42) 0 (0.00) 0 (0.00) 14 (0.58) 0 (0.00) Head (n=9) NA 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 0 (0.00) 9 (1.00)

106

APPENDIX C

2009 Bigstone Egg Nemeiben North LLR Central LLR Date Model 2 June 18 (n=17) 5.9 5.9 11.8 23.5 52.9 July 05 (n=160) 75.0 15.6 7.5 0.6 0.6 Overall (n=177) 40.4 10.8 9.6 12.1 26.8 Model 3 June 18 (n=17) 11.8 5.9 NA 29.4 52.9 July 05 (n=164) 81.1 17.7 NA 0.6 0.6 Overall (n=181) 46.4 11.8 NA 15.0 26.8

107

APPENDIX D

West Nemeiben Lac La Ronge

East Egg Nemeiben Bigstone Head

Figure 12. Carbon and nitrogen stable isotopes values of regurgitated yellow perch (TL = 70-120 mm) in 2010 for each day of bolus collections on Colony B.

108

APPENDIX E

Table 9. The proportion (%) of yellow perch (TL=70-120 mm) with assignment probabilities ≥ 50% assigned to each site by day and overall in 2010 for model 1 (all sites), model 2 (central Nemeiben Lake removed), and model 3 (2nd round of gill-netting). The largest percent proportions for each date are shaded. Central East West Lac La 2010 Bigstone Egg Nemeiben Nemeiben Nemeiben Head Ronge Date Model 1 Jun 17 (n=45) 26.7 4.4 13.3 40.0 4.4 11.1 0.0 Jun 22 (n=45) 17.8 6.7 20.0 33.3 6.7 2.2 13.3 Jun 25 (n=46) 6.5 26.1 21.7 4.3 2.2 19.6 19.6 Jun 28 (n=43) 0.0 2.3 7.0 0.0 81.4 0.0 9.3 Jul 01 (n=35) 0.0 2.9 8.6 0.0 80.0 0.0 8.6 Jul 04 (n=44) 29.5 0.0 25.0 15.9 9.1 4.5 15.9 Jul 08 (n=35) 5.7 51.4 8.6 0.0 8.6 0.0 25.7 Jul 13 (n=12) 0.0 50.0 0.0 0.0 8.3 0.0 41.7 Jul 20 (n=41) 0.0 65.9 2.4 0.0 0.0 0.0 31.7 Jul 22 (n=15) 0.0 46.7 6.7 0.0 6.7 0.0 40.0 Overall (n=361) 8.6 25.6 11.3 9.4 20.7 3.7 20.6 Model 2

Jun 17 (n=50) 32.0 8.0 NA 44.0 4.0 10.0 2.0 Jun 22 (n=47) 21.3 14.9 NA 34.0 12.8 2.1 14.9 Jun 25 (n=49) 10.2 34.7 NA 6.1 6.1 18.4 24.5 Jun 28 (n=43) 0.0 7.0 NA 0.0 81.4 0.0 11.6 Jul 01 (n=35) 0.0 5.7 NA 0.0 85.7 0.0 8.6 Jul 04 (n=44) 38.6 9.1 NA 22.7 9.1 4.5 15.9 Jul 08 (n=38) 7.9 52.6 NA 0.0 10.5 0.0 28.9 Jul 13 (n=13) 0.0 53.8 NA 0.0 7.7 0.0 38.5 Jul 20 (n=45) 0.0 66.7 NA 0.0 0.0 0.0 33.3 Jul 22 (n=18) 0.0 50.0 NA 0.0 5.6 0.0 44.4 Overall (n=382) 11.0 30.3 NA 10.7 22.3 3.5 22.3 Model 3 Jun 17 (n=47) 27.7 2.1 0.0 53.2 4.3 12.8 0.0 Jun 22 (n=44) 13.6 4.5 9.1 52.3 6.8 2.3 11.4 Jun 25 (n=41) 19.5 19.5 14.6 2.4 0.0 24.4 19.5 Jun 28 (n=40) 0.0 2.5 7.5 0.0 80.0 0.0 10.0 Jul 01 (n=32) 3.1 0.0 6.3 0.0 84.4 0.0 6.3 Jul 04 (n=44) 27.3 0.0 4.5 34.1 9.1 9.1 15.9 Jul 08 (n=33) 6.1 51.5 6.1 3.0 9.1 0.0 24.2 Jul 13 (n=13) 0.0 46.2 7.7 0.0 7.7 0.0 38.5 Jul 20 (n=37) 0.0 59.5 10.8 0.0 0.0 0.0 29.7 Jul 22 (n=14) 7.1 50.0 0.0 0.0 7.1 0.0 35.7 Overall (n=345) 10.4 23.6 6.7 14.5 20.8 4.9 19.1

109

APPENDIX F

Hg (ng/g) Hg 30

0 Egg Bigstone Nemeiben Central LLR North LLR

Lake Figure 13. Total mercury (Hg) concentrations in gill-netted yellow perch of approximately 100 mm total length from the La Ronge area in 2009.

110

APPENDIX G

300 Bigstone Lake Egg Lake 250 Nemeiben Lake North Lac La Ronge Central Lac La Ronge 200

150

100 Total Length (mm) Length Total

0 0 1 2 3 4 5 6 Age

Figure 14. The age and total length of gill-netted yellow perch from the La Ronge area in 2009.

111