FISH ECOLOGY IN OOTSA LAKE, BRITISH COLUMBIA IN RELATION TO SUBMERGED TIMBER HARVESTING by Linda Brooks BEd., University of Western Ontario, 1994 BSc., University of Western Ontario, 1993 BMus., University of Western Ontario, 1991 THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in BIOLOGY UNIVERSITY OF NORTHERN BRITISH COLUMBIA LIBRARY © Linda Brooks, 2000 Prince George, BC THE UNIVERSITY OF NORTHERN BRITISH COLUMBIA October, 2000 All rights reserved. This work may not be reproduced in whole or in part, by photocopy or other mean, without the permission of the author. ABSTRACT Ootsa Lake is part of a 48-year-old northwestern British Columbia reservoir (Nechako Reservoir) that contains extensive amounts of flooded forests. Recently, logging companies have begun harvesting this standing underwater timber because of its potential as an alternative fibre source. This project was initiated because of the general lack of information on fishes in the reservoir and because the potential impact of submerged timber removal on the fish community had not previously been explored. Catches with experimental gill nets and live traps were monitored between July 7 and October 22, 1998 to estimate the abundance, diversity, size, and condition of fishes in Ootsa Lake. Catch per unit effort (CPUE) was quantified and ranked to evaluate the impact of submerged timber harvesting on the abundance of the dominant fish species. Mean ranks were compared across species; across three near-shore habitats with different levels of structural heterogeneity (treed, harvested, or open); and across the summer and fall seasons. Shannon-Wiener diversity indices were calculated and compared across habitats. Size and condition of the main fish species were compared across habitats. Fish abundance was positively associated with habitat structural complexity. Overall abundance was highest in the treed habitat. In the summer, rainbow trout (Oncorhynchus mykiss) abundance was highest in the treed habitat whereas northern pikeminnow (Pfychochei/us oregonensis) were abundant in both treed and harvested habitats. Rainbow trout and northern pikeminnow abundance in near-shore areas ii decreased in the fall, but numbers of kokanee (Oncorhynchus nerka) increased . Shannon-Wiener diversity indices did not differ significantly among habitats overall, but in the summer the index was significantly higher in the open than the treed and harvested habitats, whereas it was significantly lower in the fall. Sizes of rainbow trout, kokanee, and northern pikeminnow were related to habitat structural complexity with the smallest fish occupying the treed habitat and the largest rainbow trout and northern pikeminnow occurring in the open habitat. Fulton's condition factor was not consistently related to habitat structural complexity but differed among sites within habitat. Kokanee were found to be exceptionally small (mean fork length =180 mm ± 0.95) with over 85% of individuals being age 2+. They also exhibited characteristics similar to "residual" sockeye salmon (Oncorhynchus nerka) including a significantly male-biased sex ratio and olive-black spawning colouration. I hypothesize that kokanee are exhibiting adaptive life history patterns in response to the cold, oligotrophic conditions in this large reservoir. Juvenile fish often seek complex underwater structure for protection from predators, and submerged structure has been shown to increase invertebrate production . .Therefore, it is likely that small rainbow trout and northern pikeminnow inhabit areas with submerged timber because it provides refuge from predators and/or because of a high abundance of food. The use of benthic harvested areas by small northern pikeminnow may be for similar reasons because overturned root wads and woody debris remain post-harvest. Diversity of fish species was not positively associated with habitat structural complexity. iii Taken together, these results suggest that alteration of structural heterogeneity in nearshore habitats caused by the harvest of standing submerged timber may result in decreased abundance of juvenile rainbow trout and northern pikeminnow, and that fall harvesting may influence the spawning migration of kokanee. Because this study was based on only one year of sampling, further research is needed before concrete conclusions can be drawn. In the meantime, management strategies should include retaining large areas of submerged timber for juvenile fish habitat and avoiding nearshore harvest activities in or near kokanee spawning areas. Future research should include a focus on kokanee life history patterns in relation to lake productivity, and should include assessment of changes in the fish communities over time as harvesting continues. iv TABLE OF CONTENTS ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ii TABLE OF CONTENTS v . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii .LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x BACKGROUND INFORMATION . . .. . . . . . . . . . . . . . .. . .. . .. . .. . .. . . . . . . . . . . .. . . . . . . . . 1 CHAPTER1 ................................................................ .............. Spatial and temporal patterns of fish abundance and diversity in relation to structural complexity of three near-shore habitats of a northern British Columbia reservoir (Ootsa Lake). 3 ABSTRACT ..................................... ·................................ 3 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 5 MATERIALS AND METHODS Site Description and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Sampling ......................... ................................... .... 9 Data Analysis Abundance (Catch per Unit Effort) . . . . . . . . . . . . . . . . . .. 12 Diversity . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 13 RESULTS Site Location and Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 14 Abundance (CPUE) Abundance Across Species . . . . . . . . . . . . . . . . . . . . . . . . . . .. 17 Abundance Across Habitats . . . . . . . . . . . . . . . . . . . . . . . . . . .. 17 Abundance Across Seasons . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 23 DISCUSSION Abundance ......................... ... . .. . .. . .. ................... .. . .. 23 Diversity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . 29 v CHAPTER 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 34 Differences in size and condition of fish in Ootsa Lake, BC in relation to changes in habitat structural complexity caused by submerged timber harvesting ABSTRACT ....................................................................... 34 INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 MATERIALS AND METHODS Sampling ................................................................ 41 Data Collection and Analysis - Descriptive Baseline Information Length-Weight Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Length at Age .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 42 Diet .......................................................... ..... 43 Kokanee Age at Maturity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Kokanee Fecundity, Egg Size and Sex Ratio . . . . . . . . .. 46 Water Quality .............. ... .................................. 4 7 Statistical Analysis and Hypothesis Testing Size and Fulton's Condition Factor .................... 49 RESULTS Descriptive Baseline Information ............................... 51 Length-Weight Relationships .............. .... ............ 51 Length at Age .. .. .. .. .. .. . .. .. .. . . . . . . . . .. .. .. .. .. .. .. . .. .. .. .. 52 Diet ............................................................... 52 Kokanee Age at Maturity, Fecundity, & Egg Size ................................................ 54 Kokanee Sex Ratio, Spawning Colour, & Parasitic Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Water Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 55 Statistical Results - Size and Condition Factor . . . . . . . . . . . . 56 Rainbow Trout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 56 Kokanee ....................................... ..... ............. 58 Northern Pikeminnow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 64 DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 78 FINAL CONCLUSIONS AND MANAGEMENT RECOMMENDATIONS . . .. . .. . .. . .. .... .. .... .. . .. . .. . .. . .. . .. .... .. 89 APPENDIX . . .. . . . . . . . . . . .. . . . . . . . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . .. . . . . . . . .. . .. . .. . 92 REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 115 vi LIST OF FIGURES Figure 1.1: Map of Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 Figure 2.1: Fulton Condition Factors for Rainbow Trout in Each Site . . . . . . . . . . . . . . . . . . . . . . . ... 59 Figure 2.2: Interaction between Habitat and Seasons for a) Length and b) Weight of Kokanee . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 62 Figure 2.3: Fulton Condition Factor for Kokanee in Each Site .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ... 63 Figure 2.4: Fulton Condition Factor for Kokanee Across Seasons .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 65 Figure 2.5: Fulton Condition Factor among Sexes for Kokanee . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66 Figure 2.6: Interaction between Habitat and Sex for Condition Factor of Kokanee ..... .... 67 Figure 2.7: Interaction between Season and Sex for Condition Factor of Kokanee .. .. .. ... 68 Figure 2.8 : Interaction Between Habitat and Sex for a) Length and b) Weight of Northern Pikeminnow .............................. ...... ............ .. ......... 71 Figure 2.9: Interaction Between Sex and Season for a) Length and b) Weight of Northern Pikeminnow ........................................................... 72 Figure 2.10: Interaction Between Sex and Depth for a) Length and b) Weight of Northern Pikeminnow ........................................................... 73 .Figure 2.11 : Fulton Condition Factor for Northern Pikeminnow in Each Site .. .. .. .. .. .. .. ... 75 Figure 2.12: Fulton Condition Factor for Northern Pikeminnow Across Seasons . . . . . . . . . .. 76 Figure 2.13: Interaction between Habitat and Depth for Condition Factor of Northern Pikeminnow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Figure A.1 : Relationship between Length and Weight of Rainbow Trout in each Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Figure A.2 : Relationship between Length and Weight of All Rainbow Trout .... .. .... .. ..... 94 Figure A.3: Relationship between Length and Weight of Kokanee in each Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Figure A.4 : Relationship between Length and Weight of All Kokanee .. .. .. .. .. .. .. .. .. .. .. .. . 96 Figure A.5: Relationship between Length and Weight of Northern Pikeminnow in each Habitat .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 97 Figure A.6: Relationship between Length and Weight of All Northern Pikeminnow ... ... ... 98 Figure A.7 : Length at Age for Rainbow Trout in Each Habitat .................................... 99 Figure A.8 : Length at Age for Kokanee in Each Habitat .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. 100 Figure A.9: Length at Age for Northern Pikeminnow in Each Habitat .. .. .. .. .. .. .. .. .. .. .. .. .. 101 Figure A.1 0: Length at Age for Northern Pikeminnow Age 12 and Under . . . . . . . . . . . . . . . . . . .. 102 vii Figure A.11 : Frequency of Occurrence of Food Items for Each Species in Each Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 Figure A.12 : Percent Composition by Number of Food Items for Each Species in Each Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 Figure A.13 : Percent Composition by Number of Aquatic Insects and Terrestrial Spiders for Each Species in Each Habitat ......... ... .... ...... .... ......... . 105 Figure A.14: Temperature Profiles for Each Site ...... ... ............. .. ... ...... .. .... .. .. ... .... ... 106 Figure A.15: Oxygen Profiles for Each Site .. ........ ...... ........ .......... .. ... .... ..... ............ 107 Figure A.16 : Oxygen Profiles of the Harvested Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 108 Figure A.17: Secchi Disk Readings in Each Habitat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 Figure A.18: Suspended Sediment in Each Site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 110 Figure A.19: Percent Organic Matter in the Substrate of Each Habitat . . . . . . . . . . . . . . . . . . . . . . . . 111 viii LIST OF TABLES Table 1-1 : UTM Site Coordinates ... ....... ........ ... .... ............................ ..... .... ....... .. 15 Table 1-2: Size and Description of Sampling Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Table 1-3: Catch per Unit Effort Among Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 Table 1-4: Catch per Unit Effort Within Habitat Among Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... 19 Table 1-5. Catch per Unit Effort Among Habitats .. ... ... .. . ....... .. ... ... ........ . ... . . . .......... 20 Table 1-6. Catch per Unit Effort Across Seasons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 22 Table 1-7: Shannon-Wiener Indices of Diversity ......... ............ ........ ........ ........ .... .. 24 Table 2-1: Diet Categories for Stomach Content Analysis .. ... .. .... ............ ...... ...... .... 45 Table 2-2: Results of Nested ANOVA for Rainbow Trout .. .. .... ..... ...... ... .. ........ ... .. .... 57 Table 2-3: Results of Nested ANOVA for Kokanee .......... .................... .......... .. ....... 60 Table 2-4: Results of Nested ANOVA for Northern Pikeminnow .... .. .... ...... ...... .. .... .... 69 Table A-1 : Percentage of Kokanee in Each Age Class ........................................... 112 Table A-2: Fecundity of Kokanee .. .. .. .. .. .. .. .. .. . .. .. . .. .. .. . .. .. . .. .. .. .. .. .. .. .. .. . .. .. .. .. .. .. .. .. 113 Table A-3: Sex Ratio of Kokanee 114 ix Acknowledgements I would like to thank the BC Ministry of Forests (Lakes District), Canadian Forest Products Ltd. (CANFOR), and the Natural Sciences and Engineering Research Council of Canada (NSERC) for their financial support of this project. I would also like to thank Jason Hwang at the Federal Department of Fisheries and Oceans (DFO), Prince George and Dana Atagi at the Ministry of Environment, Lands and Parks (MELP), Smithers for their encouragement and support. My industrial supervisors from CAN FOR, Sandy Long and Dennis Anderson, were invaluable resources, and they provided me with essential information and guidance that allowed the project to occur. Special thanks to my UNBC supervisor (Max Blouw) who gave me gentle guidance and moral support, and to my other committee members (Dan Heath, Keith Egger, and Kathy Parker) for their helpful suggestions throughout all stages of the project. Also, thanks to Dan Cadwaladr, Jason Kasper, Linda Rankin, and Trent and Shelley Hoover for their assistance in the field and lab, and to Dieter Ayers for his statistical expertise. Finally, a very special thank you to David and Liam Rutledge for their sacrifices that allowed me to see this project through to the end. X BACKGROUND INFORMATION The Nechako Reservoir, located approximately 300 km west of Prince George, British Columbia (BC) was created in 1952 by the Aluminum Company of Canada (ALCAN) in order to provide hydroelectric power for their aluminum smelters in Kitimat. As a result, 334 km 2 of terrestrial vegetation was inundated (Northcote and Atagi 1997) including vast areas of white spruce (Picea glauca), lodgepole pine (Pinus contorta) and subalpine fir (Abies lasiocarpa) forest. In 1996, the Ministry of Forests granted two 10-year licenses for the commercial harvest of submerged timber in the Nechako Reservoir, one to Canadian Forest Products Limited (CANFOR) in conjunction with the Cheslatta Development Corporation (CDC), and one to Fibrecon Management Limited and the Cheslatta Carrier Nation. Both companies anticipated the use of this alternative fibre source for structural studs and pulp for paper production. CANFOR conducted nearshore and deep water logging in Ootsa Lake, the northern-most lake depression in the reservoir, between 1996 and 1998, whereas Fibrecon removed near-shore timber mostly in Whitesail Reach, west of Ootsa Lake, and continues to do so. The Nechako Reservoir is different from many of the reservoirs in BC in that the water is controlled by ALCAN and not BC Hydro (the primary supplier of power to the province). The province has made efforts to monitor the status of fisheries in most of BC's reservoirs, but because no provisions for management of fisheries in the Nechako Reservoir were made under the original"1950 Agreement" between ALCAN and the provincial government or under the "1987 Settlement Agreement" between ALCAN and the provincial and federal governments, or the "BC/Aican 1997 Agreement", the status of the fish community in the reservoir has been largely unknown since impoundment. Because of this general lack of information regarding the fish community in the reservoir, the Department of Fisheries and Oceans (DFO) (Prince George) along with the Ministry of ·Environment, Lands, and Parks (MELP) (Smithers) expressed concern over the potential long-term impacts that submerged timber removal could have on the fish community. The purpose of this study was to provide baseline information which will facilitate studies of the long-term impacts of submerged timber harvesting on the fish community in Ootsa Lake. To accomplish this I compared the abundance, biodiversity, and ecological characteristics of fishes among three near-shore habitats with different levels of structural heterogeneity. These habitats included treed sites that contained a substantial amount of inundated timber (high heterogeneity), harvested sites where inundated timber had been recently removed (medium heterogeneity), and open sites that were flooded but never contained standing timber (i.e. previous farmland) (low heterogeneity). For clarity, this thesis is divided into two chapters. The first addresses issues of fish abundance and biodiversity; the second examines ecological characteristics of the fish community. 2 CHAPTER 1 Spatial and temporal patterns of fish abundance and diversity in relation to structural complexity of three near-shore habitats of a northern British Columbia reservoir (Ootsa Lake). ABSTRACT Catches with experimental gill nets and live traps (blunder, Plexiglas, minnow) were monitored between July 7 and October 22, 1998 to estimate the abundance and diversity of fishes in Ootsa Lake, BC. Catch per unit effort (CPUE) was quantified and ranked to evaluate the impact of submerged timber harvesting on the predominant fish species. Mean ranks were compared across species; across three near-shore habitats with different levels of structural heterogeneity (treed, harvested, or open); and across the summer and fall seasons. ShannonWiener diversity indices were compared across habitats. Fish abundance was positively associated with habitat structural complexity. Overall abundance was highest in treed habitat. In summer, rainbow trout (Oncorhynchus mykiss) abundance was highest in the treed habitat whereas northern pikeminnow (Pytchochei/us oregonensis) were abundant in both treed and harvested habitats. Rainbow trout and northern pikeminnow abundance in near-shore areas decreased in the fall but kokanee (Oncorhynchus nerka) abundance increased. Abundance of benthic northern pikeminnow was highest in 3 the harvested habitat. The temporal changes observed in fish abundance are consistent with seasonal migrations that occur in natural lakes. Shannon-Wiener diversity indices did not differ significantly among habitats overall , but in the summer the index was significantly higher in the open than the treed and harvested habitats, whereas it was significantly lower in the fall. Fish species ·diversity is probably also under the influence of other environmental and behavioural characteristics. Evidence from exploratory sampling with nets on the bottom of open sites suggests that trap sampling may have been species biased, resulting in biased estimates of diversity. These results suggest that the alteration of structural heterogeneity in near-shore habitats caused by harvesting of standing submerged timber may result in decreased abundance of rainbow trout, and that fall harvesting may influence the spawning migration of kokanee. Removal of inundated timber may introduce favourable benthic habitat for northern pikeminnow. To accurately estimate the long-term impacts of harvesting on fishes, yearly monitoring is recommended to determine how the fish communities change in different habitats over time. 4 INTRODUCTION In the past, almost nothing has been done to manage fish populations in British Columbia's aging, oligotrophic reservoirs (Ken Ashle/, personal communication). Authorities in the United States (US) have purposely retained standing timber in many reservoirs for enhancement of fish habitat postimpoundment (Laufle and Cassidy 1988). The effects of flooded standing timber on sport fish harvest, however, continue to be poorly understood, especially in the cold oligotrophic systems of northern latitudes. The importance to fish of standing timber in reservoirs has been identified as a deficiency in fisheries management investigations in North America (Northcote and Atagi, 1997). In British Columbia vast areas of inundated forest have recently sparked the interest of several logging companies because of the economic potential of drowned timber as an alternative fibre source. Harvesting of partially and fully submerged coniferous trees began in Ootsa Lake, BC in 1996 and continues to date. Scientific research regarding the impact of inundated timber harvest on fish is essentially non-existent. Therefore, information on how fish interact with structurally complex underwater forests and how harvesting alters fish community structure is crucial to researchers, fisheries managers, and to the forest industry. 1 Fisheries Research and Development Section, Fisheries Branch, Ministry of Environment, Lands and Parks, Province of BC, 2204 Main Mall, University of British Columbia, Vancouver, BC, Canada, V6T 1Z4 5 Submerged trees contribute to structural complexity of the habitat, and the importance of structural complexity in fish habitat is well documented. For example, the presence of aquatic vegetation or artificial structures results in higher growth rates in fish (Crowder and Cooper 1982; Wege and Anderson 1979), as well as habitat partitioning among species (Werner et al. 1977; Stang and Hubert 1984 ). Successful fisheries management strategies have involved introducing artificial reefs to attract fish in marine (Wickam et al. 1973), lacustrine (Reeves et al. 1977; Prince and Maughan 1979; Helfman 1979), and reservoir environments (Paxton and Stevenson 1979). Underwater spatial structure is particularly important for juvenile fish. For example, persistence and amount of flooded terrestrial vegetation is correlated with abundance of young-of-the-year largemouth bass (Micropterus sa/moides) '(Summerfelt, 1993). Similarly, artificial and natural wood structures in streams provide juvenile Oncorhynchus spp. with refuge from predators (Gulp et al. 1996) and relief from high water velocities (Shirvell 1990). Young rainbow trout in Utah reservoirs select near-shore habitats that are structurally complex in order to avoid predators (Tabor and Wurtsbaugh 1991 ). In addition, survival of juvenile coho salmon (Oncorhynchus kitsutch) is strongly correlated with abundance of woody debris (Quinn and Peterson 1996), and submerged pulpwood logs from previous log-driving transport appear to serve a protective role for female and immature fish (Moring et al. 1989). 6 Despite the abundance of information on the importance of habitat structural heterogeneity as refuge for fishes, there is limited information available on the role that flooded standing timber plays for fish communities in reservoirs. Available data suggest that submerged forests increase the food base for fish (Poddubnyy and Fortunatov 1961; Mclachlin 1970). In addition, Davis and Hughes (1971) found that largemouth bass and bluegill (Lepomis macrochirus) were more abundant in areas of submerged standing timber, whereas Willis and Jones (1986) found that standing fish crops (kg/hectare) of several species including sunfishes (Lepomis spp.), crappies (Pomoxis spp.), and largemouth bass were higher in areas of Kansas reservoirs that had standing timber. Inundated timber may also serve as spawning substrate for shad (Dorosoma spp.) and improve shelter and food availability for larval shad (VanDenAvyle and Petering 1988). Habitat heterogeneity has been positively associated with species diversity in both terrestrial (MacArthur and MacArthur 1961; Pianka 1967; Rosenzweig and Winakur 1969) and aquatic (Abele 1974, Bronmark 1985) systems. More specifically, within-lake fish species diversity has been correlated with several factors of habitat heterogeneity including depth gradient (Benson and Magnuson 1992), diversity of invertebrate prey and substrate, and vegetation complexity (Eadie and Keast 1984; Tonn and Magnuson 1982). In addition, greater habitat diversity may support higher spatial heterogeneity of fish communities (Benson and Magnuson 1992). I know of no literature that examines the relationship 7 between fish species diversity and the habitat heterogeneity associated with submerged standing timber. Flooded standing timber provides a more structurally complex habitat than either harvested areas, which contain root wads and woody debris, or open areas with .non-vegetated substrate. Based on the literature, I predicted that structural heterogeneity of treed areas would be positively related to species diversity. If fish respond quickly to harvesting, it was expected that the harvested habitat should have an intermediate level of diversity. I also predicted that presence of submerged standing timber would influence spatial segregation within and among species. To determine the validity of these predictions and to assess the potential impacts of harvesting on the fish community in Ootsa Lake, I examined differences in fish species abundance and diversity among habitats with different levels of structural complexity (treed, harvested, or open). Because fishes are known to migrate into and out of near-shore areas seasonally, I also examined seasonal variation (summer and fall). The overall null hypothesis was that there is no difference in abundance or diversity of fish species among habitats with varying degrees of structural heterogeneity, or among seasons. 8 MATERIALS AND METHODS Site Description and Characterization Three treed, three harvested, and three open sites were identified for this study and UTM coordinates were recorded for each site. Sites were chosen on the basis of structural heterogeneity of underwater fish habitat as determined by depth sounder information (treed - high heterogeneity, harvested - medium heterogeneity, open- low heterogeneity) with special consideration for similarity in all other aspects (e.g. depth, gradient, distance from shore) to minimize uncontrolled variables. A further consideration was that open sites be large enough (at least 20 hectares) to minimize potential edge effects (Wilcove et al. 1986). Site locations are shown in Figure 1.1. Various site characteristics were estimated from a submerged timber assessment conducted by Canadian Forest Products Limited (CAN FOR) prior to harvesting; these included size (ha), tree species composition, tree age, tree height, and wood volume. Sampling All sites were sampled on the surface with 48 m monofilament gill nets (Redden Nets, Campbell River, BC) arranged in the experimental gang configuration (2 m depth, 8 m long panels of 19, 25, 38, 51, 64, 77 mm stretch mesh (Hubert 1996)) and on the bottom with various traps (blunder (hand-made), Plexiglas (handmade), and standard minnow) baited with dog or cat food. Blunder traps had 9 0 Scale 2km 0. o- Barker../? lsland v 01 0.'\1 0 GV' 6 "" 0 ~ Knox & Island 02 H2 Ootsa Lake • Schoolhouse CANFOR Forestry Bay Field Station T2 0 Figure 1.1. Map of study area in Ootsa Lake, BC showing 1997/98 fish and water quality sampling sites. T=treed habitat; H=harvested habitat; O=open habitat. 0 ' Oo. 03 .,. openings of approximately 20, 40, 64, and 76 mm diameter, Plexiglas traps had diameter openings of approximately 65 mm, and minnow traps had diameter openings of about 25 mm. Both nets and traps were set approximately 100 - 200 m from shore in the evening near dusk and retrieved in the morning near dawn. Within each site nets were set perpendicular to shore, and traps were placed approximately 10- 20 m apart. Sampling occurred between July 9 and October 22, 1998. Summer sampling with nets took place between July 9- August 16, 1998 and with traps from August 19 -August 28, 1998. Net sampling in the fall occurred from September 17 - October 16, 1998 and with traps from October 20 - October 22, 1998. Sampling effort with gill nets was identical for each habitat (12 sets total: 6 in the summer and 6 in the fall). Bottom sampling was also identical for each habitat (7 sets: 5 in the summer and 2 in the fall). However, certain sites of each habitat type were not accessible during hazardous weather conditions so sampling effort among sites differed. The treed and open sites each had equal net sampling effort (2 in the summer and 2 in the fall). Harvested site H1 (Figure 1.1) was sampled 3 times in the summer and once in the fall, H2 was sampled twice in the summer and twice in the fall, and H3 was sampled once in the summer and 3 times in the fall. For bottom sampling with traps, T1, T2, H2, H3, 01, and 02 were sampled twice in the summer and once in the fall whereas T3, H1 and 03 .were sampled once in the summer and not sampled in the fall. 11 Between August 16 and September 17, 1998 part of the treed site T1 was unexpectedly harvested. The precise spot where the net was normally set was cleared of timber but the treed areas directly surrounding the specific site remained. The size of the harvested patch was estimated at approximately 0.51.0 ha. Post-harvest, I continued to set the net in the same location. To determine possible effects on abundance of this partial harvest, AN OVA on rank transformed CPUE was used to compare post-harvest abundance of all species across treed sites, as well as abundance within each species across treed sites. No significant differences were found so all data acquired from T1 samples were included in the main data analysis. Data Analysis Abundance (Catch per Unit Effort) Only rainbow trout (Oncorhynchus mykiss), kokanee (Oncorhynchus nerka), and northern pikeminnow (Ptychocheilus oregonensis) were caught in sufficient numbers in surface nets and only northern pikeminnow were caught in sufficient numbers in bottom traps to warrant statistical analysis. Other species caught during surface net sampling were longnose sucker (Catostomus catostomus) (n=12), largescale sucker (Catostomus macrocheilus) (n=2), and Rocky Mountain whitefish (Prosopium williamsom) (n=1 ); other species caught during bottom trap sampling were burbot (Lata Iota) (n=2) and prickly sculpin (Cottus 12 asper) (n=2). Only summer sampling with blunder traps produced enough data for analysis. Catch per unit effort (CPUE) (#fish caught per hour of soak time) for each sampling event and sampling method (gill net or blunder trap) was calculated for each species. CPUE estimates rather than counts were used because CPUE accounts for differences in sampling effort. A CPUE of zero was scored if no captures were made. Less than 11% of all data points (surface and bottom) were zero. ANOVA on rank transformed CPUE (Conover 1980; Bruno Zumbo2 , personal communication) was used to estimate differences in overall (among species) and spatial (among habitats) patterns of fish abundance; t-tests on ranked CPUE were used to determine differences in temporal (among seasons) patterns (Conover 1980; Bruno Zumbo, personal communication). This method alleviates the problem of non-normal CPUE data, and also allowed multiple comparisons with Bonferroni post-hoc tests. Tests were done separately for surface and benthic sampling with SPSS statistical software (SPSS Inc., Chicago, IL). Diversity Due to unequal sampling efforts with traps, the Shannon-Wiener diversity index calculation uses adjusted numbers of fish to reflect the maximum number of traps set (4 blunder, 3 minnow, 1 Plexiglas) for each sampling episode. The Shannon- 13 Wiener (H' =-Lpilog1o(Pi)) index of diversity (Pielou 1975), where H' is species diversity and Pi is the fractional abundance of the i1h species, was calculated for each habitat in each season and over both seasons. An adapted t-test (Zar 1984) was used to test for significant differences between Shannon-Wiener indices among habitats. Diversity indices were calculated in EXCEL 97 (Microsoft Corporation) and t-tests were calculated by hand. RESULTS Site Location and Characterization The location of each site is shown in Figure 1.1 and UTM co-ordinates for all site ·locations are presented in Table 1-1. Sites were similar in depth (8 - 15 m ), bottom gradient (Table 1-2), and distance from shore (-100- 200m). Treed and harvested sites had similar tree species composition and submerged wood volume (prior to harvesting). All sites were > 20 ha in size (Table 1-2). Therefore, aside from the heterogeneity among sites in the amount of standing timber, they were assumed to be qualitatively and quantitatively similar. 2 Psychology Department, University of Northern BC, 3333 University Way, Prince George, BC, Canada, V2N 4Z9 14 ___.. CJ1 09/11/98 09/17/98 5957805.567 44 5961047.37216 5959993.98909 5959904.83408 5959756.88302 5966783.75056 5965535.96020 5967144.93441 T2 T3 H1 H2 H3 01 02 04 09/10/98 09/17/98 09/18/98 09/10/98 09/10/98 09/10/98 09/11/98 5959633.72987 T1 16:53:28 3:01:44 2:43:10 19:33:24 23:40:44 22:37:25 0:12:33 0:55:33 15:39:14 Table 1-1. UTM site coordinates for 1998 sampling sites in Ootsa Lake, BC . Values are not differentially corrected; accuracy is within 25- 100m. ...... C1l 80% P; 20% S 70% P; 20% S; 10% A 90% P; 10% S 60% P; 30% S; 10% A 90% P; 10% S 11.0- 13.0 10.6- 12.4 8.0 - 14.0 10.0 - 12.0 11 .8-14.9 10.9 - 14.3 10.6 - 15.0 10.4- 13.7 177.6 52.1 25.0 27 .1 20.1 55.0 522.2 107.0 T2 T3 H1 H2 H3 01 02 04 Non-forested Non-forested Non-forested 90% P; 10% S 9.1 - 11 .1 29.5 T1 Species Composition Sampling Depth (m) Area (ha) Site 268 201 24 20 N/A N/A N/A N/A N/A 22 22 N/A N/A N/A 234 224 216 234 21 3 Wood Volume (m /ha) 22 Tree Height (m) N/A 90 90 90 90 90 90 Tree Age (y) Timber Description Table 1-2. Size and description of 1998 fish sampling sites in Ootsa Lake, BC (July- October). Sampling depth describes the range of depths over which nets and traps were set and is a reflection of bottom gradient. Species composition for harvested sites is a pre-harvest description. P = lodgepole pine (Pinus contorta ); S =white spruce (Picea glauca ); A= trembling aspen (Populus tremuloides ). Abundance (CPUE) Abundance across species Overall abundance was significantly different among species (p<0.05) (Table 13). Post-hoc tests indicated that kokanee were more abundant than rainbow trout. Although northern pikeminnow had the highest estimated mean raw CPUE (i.e. not rank transformed), there was a high variance in the data so that mean .rank CPUE was not significantly higher than other species. Differences in species abundance were evident in the open habitat (p<0.01) (Table 1-4) where kokanee were more abundant than both rainbow trout and northern pikeminnow. Although at the margin of significance (p=0.06), species abundance also differed in the harvested habitat (Table 1-4) where a trend of more northern pikeminnow than rainbow trout was noted. In the summer, species abundance differed in both the treed (p<0.05) and harvested habitats (p<0.001) (Table 1-4 ). Northern pikeminnow were much more abundant than the other two species in both habitats. Fall species abundance differed in all three habitats (p<0.05) (Table 14 ). Kokanee were more numerous than rainbow trout in all three habitats, and they were also more abundant than northern pikeminnow in the open. Abundance across habitats (spatial variation) Total abundance of fishes varied significantly among habitats (p<0.05) (Table 15). There was a trend in abundance (treed > harvested >open), but the only significant difference was between the extremes of the treatment groups. Northern pikeminnow abundance varied spatially (p<0.01) (Table 1-5); they were 17 (X) 36 36 0.10 0.28 0.58 0.99 Kokanee Northern Pikeminnow 36 0.03 0.23 Rainbow Trout I I n (#net sets) SEM Species Mean CPUE Table 1-3. Catch per unit effort (CPUE) among fish species in Ootsa Lake, BC from 1998 sampling (July - October). Analysis of differences in abundance is based on ANOVA on rank transformed CPUE data. Dots on vertical lines connect those species among which abundance cannot be shown to differ (p>0.05). co Open Harvested Treed Habitat 1.68 NSC 0.82 0.13 KO NSC 0.17 1.17 NSC RB 0.44 KO I 0.20 0.50 KO RB 0.34 RB Species Mean CPUE 0.04 0.25 0.05 0.47 0.13 0.04 0.65 0.10 0.08 SEM 12 12 12 12 12 12 12 12 12 n Overall p<0.01 n.s. (p=0.06) n.s. p I I 0.13 0.36 0.23 2.15 0.22 0.24 3.02 0.40 0.56 Mean CPUE 0.06 0.14 0.08 0.77 0.06 0.07 1.06 0.11 0.09 SEM 6 6 6 6 6 6 6 6 6 n Summer n.s. p<0.001 p<0 .05 p II II I 0.13 1.27 0.10 0.19 0.67 0.16 0.33 0.59 0.12 Mean CPUE 0.54 0.42 0.03 0.05 0.23 0.03 0.10 0.17 0.03 SEM Fall 6 6 6 6 6 6 6 6 6 n p<0.05 p<0.05 p<0.05 p Table 1-4. Catch per unit effort (CPUE) within habitat among fish species in Ootsa lake, BC. Significant differences in abundance are based on ANOVA on rank transformed CPUE data collected July- October 1998. Where significant differences occur, dots on vertical lines connect those species among which abundance cannot be shown to differ (p>0.05). n.s. = not significant; RB = rainbow trout; KO = kokanee; NSC = northern pikeminnow; n = # net sets. N 0 0.17 Open NSC All Species NSC 0.06 0.11 0.03 Harvested Open 0.37 0.60 0.84 0.13 1.17 Treed Open Harvested Treed Open Harvested I I 0.82 Open 1.68 0.44 Harvested Treed 0.50 Treed I 0.20 Harvested KO 0.34 Treed RB Mean CPUE Habitat Species 36 36 0.24 0.17 21 21 0.03 21 0.03 0.03 36 12 0.04 0.10 12 12 0.65 0.47 12 12 12 12 12 12 n 0.25 0.13 0.10 0.05 0.04 0.08 SEM Overall p<0.001 p<0.05 p<0.01 n.s. n.s. p Surface 0.13 2.15 3.02 0.05 0.11 0.07 Bottom I I 0.36 0.22 0.40 0.04 0.04 0.04 0.06 0.77 1.06 0.14 0.06 0.11 0.08 0.07 •I 0.24 • 0.23 0.09 SEM Summer 0.56 I Mean CPUE = 15 15 15 6 6 6 6 6 6 6 6 6 n p<0.001 p<0.001 n.s. p<0.05 p 0.13 0.19 0.33 1.27 0.67 0.59 0.10 0.16 0.12 Mean CPUE 0.05 0.05 0.10 0.42 0.23 0.18 0.03 0.03 0.03 SEM Fall = 6 6 6 6 6 6 6 6 6 n n.s. n.s. n.s. p Table 1-5. Catch per unit effort (CPUE) offish species among habitats in Ootsa Lake, BC. Significant differences in abundance are based on ANOVA on rank transformed CPUE data collected July- October 1998. Where significant differences occur, dots on vertical lines connect those habitats among which abundance cannot be shown to differ (p>0 .05). Dots separated by a dashed line indicate that post-hoc tests were at the margin of significance. Surface values reflect net sampling and bottom values reflect trap sampling . n.s. not significant; dash indicates that comparison was not done; RB rainbow trout; KO =kokanee; NSC =northern pikeminnow. more abundant in both treed and harvested habitats than in the open habitat. In the summer, rainbow trout and northern pikeminnow abundance differed among habitats (p<0.05 and p<0.001, respectively) (Table 1-5). Rainbow trout were more numerous in the treed habitat than the harvested and open habitats, but post-hoc tests indicated only a trend of higher abundance in the treed habitat compared to the harvested (p=0.07) or open (p=0.09) habitats. Northern pikeminnow at the surface in summer were more plentiful (p<0.001) (Table 1-5) ·in the treed and harvested habitats than the open. No species showed spatial differences in abundance in the fall (Table 1-5). Northern pikeminnow in benthic areas differed in abundance among habitats (p<0.001) (Table 1-5). Their abundance was highest in the harvested habitat and lowest in the open habitat with each group being significantly different from the other. In summer their abundance was higher in the treed and harvested habitats than the open (p<0.001) (Table 1-5). Abundance across seasons (temporal variation) Total abundance of fishes did not differ across seasons (p=0.55) (Table 1-6). Within each species, however, temporal variation in abundance was evident. Abundance of rainbow trout and northern pikeminnow decreased from summer to fall (p<0.01) whereas kokanee abundance increased (p<0.05) (Table 1-6). Within habitats, temporal differences were significant only for numbers of rainbow trout in the treed habitat (p<0.01) and for northern pikeminnow in both treed (p<0.05) 21 N N All Species NSC KO RB Species 0.39 0.22 Fall Fall 1.77 Summer 0.81 0.51 0.84 Fall Summer 0.18 0.33 Summer 0.07 0.19 0.45 0.06 0.17 0.13 Fall 0.06 0.34 SEM Summer Season Mean CPUE Overall 54 54 18 n.s. 0.33 3.02 18 0.40 0.59 p<0.01 p<0.05 0.56 p<0.01 0.12 Mean CPUE p 18 18 18 18 n 0.10 1.06 0.18 0.11 0.34 0.09 SEM Treed 6 6 6 6 6 6 n p<0.05 n.s. p<0.01 p 0.19 2.15 0.67 0.22 0.16 0.24 Mean CPUE 0.05 0.77 0.23 0.06 0.03 0.07 SEM 6 6 6 6 6 6 n Harvested p<0.01 n.s. (p=0.09) n.s. p 0.13 0.13 1.27 0.36 0.10 0.23 Mean CPUE 0.05 0.06 0.42 0.14 0.03 0.08 SEM Open Table 1-6. Catch per unit effort (CPUE) of fish species among seasons in Ootsa Lake, BC from 1998 sampling. Significant differences in abundance are based on t-test on rank transformed CPUE data collected July - October 1998. n.s. =not significant; RB =rainbow trout; KO = kokanee; NSC =northern pikeminnow. 6 6 6 6 6 6 n n.s. n.s. n.s. p and harvested (p<0.01) habitats (Table 1-6). There was no seasonal variation in abundance of benthic northern pikeminnow. Diversity Fish species diversity was not consistently related to habitat structural heterogeneity. Overall, fish diversity among habitats did not vary significantly (Table 1-7). Diversity in the summer, however, was higher in the open habitat than in the treed and harvested habitats (p<0.05), and in the fall, it was lower in the open habitat compared to treed and harvested habitats (p<0.05) (Table 1-7). Differences in diversity between the treed and harvested habitats were not statistically significant, either overall or within season. DISCUSSION Abundance The collective findings of this study generally support the prediction that fish abundance in the limnetic zone of Ootsa Lake varies according to habitat structural complexity and season. Treed areas supported the highest abundance of fishes. In the summer, rainbow trout were captured more often in treed areas, whereas northern pikeminnow inhabited both treed and harvested habitats. In the 23 """" N 0.261 0.559 Harvested H' Open H' 0.335 Treed H' I Summer Habitat 0.247 0.435 0 .403 I Fall 0.399 0.382 • 0.401 I Overall Table 1-7. Shannon-Wiener indices of diversity (H') for fish among habitats of Ootsa Lake, BC in each season and overall. Dots on vertical lines connect those habitats among which index values cannot be shown to differ (p>0.05). fall both species moved out of near-shore areas in contrast to kokanee that migrated onshore. Benthic areas of harvested habitat were readily inhabited by northern pikeminnow. Therefore, I rejected the null hypothesis that fishes in the near-shore areas of Ootsa Lake show no spatial or temporal variation in abundance. Gill nets are known to be size selective (see Hamley 1975 for a review of gill net selectivity), but the use of experimental nets in this study should reduce the problem of size selectivity. A related concern is that species with a wider size range will be more vulnerable, and interspecific comparisons may be suspect. However, the majority of the abundance comparisons in this study are within species. The majority of rainbow trout (87.1 %) were caught in 3 mesh sizes (25, 38, and 51 mm) as were the majority of northern pikeminnow (91.7%) (19, 25, and 38 mm). Kokanee were primarily (91.6%) caught in 19 and 25 mm mesh. So, while abundance comparisons among species are potentially problematic, the fishing effort for rainbow trout and northern pikeminnow is similar and comparisons of abundance between these two species are less dubious. While the word "catch" or "CPUE" may be more appropriate than "abundance" when referring to among species comparisons, I have used CPUE as an estimator of abundance and therefore I use the words "abundance" or "numbers" throughout for the sake of consistency and clarity. 25 Evidence from this study suggests that flooded standing timber in Ootsa Lake provides important habitat for rainbow trout and northern pikeminnow, particularly in the summer months. Increased underwater structure has three main possible benefits for fish in lentic systems: increased spawning substrate (VanDenAvyle and Petering 1988), enhancement of aquatic invertebrate production (Pardue and Neilsen 1979), and protection from predation (Tabor and Wurtsbaugh 1991 ). Submerged timber provides increased spawning substrate for species such as shad (VanDenAvyle and Petering 1988) that spawn near the surface and have adhesive eggs that attach to underwater structure (Scott and Crossman 1998). However, certain life history features of rainbow trout and northern pikeminnow make it highly unlikely that treed sites would provide appropriate habitat for successful spawning of either species. Both rainbow trout and northern pikeminnow spawn in gravelly benthic areas, and rainbow trout eggs are not adhesive (Scott and Crossman 1998). Therefore, increased pelagic structure would not improve spawning success of these species. Absence of fish in spawning condition and lack of benthic gravel in the substrate of treed sites (see Chapter 2) provide further evidence that these species are not congregating in the underwater trees of Ootsa Lake to spawn during the summer. In reservoirs, flooded terrestrial vegetation is considered important for providing food and protection for fishes (Aggus and Elliott 1975) and is directly correlated to fish abundance and growth (Shirley and Andrews 1977). Whether food items 26 are more abundant in areas of Ootsa Lake where submerged trees are present was not determined in this study but the literature suggests this is a possibility (VanDenAvyle and Petering 1988; Poddubnyy and Fortunatov 1961; Mclachlan 1970; Pardue and Nielsen 1979). Direct comparisons with studies related to submerged timber and food availability are difficult due to the very different composition of fish species examined. The most relevant studies address impacts on fish species different from those found in Ootsa Lake (e.g. Centrarchids (largemouth bass), crappie (Pomoxis sp. ), and Clupeiformes (shad)). Valuable information is available, however, on benthic invertebrate fauna and the potential of submerged trees as attachment substrate. Mclachlin (1970) concluded that bark surfaces support a community of fauna in which immature chironomids were predominant with a notable abundance of oligochaetes (larger worms), Trichoptera (caddisflies) and Ephemeroptera (mayflies). Fish from this study consumed all of these invertebrate species with the exception of oligochaetes (see Chapter 2). This evidence implies that submerged trees and, possibly, large woody debris may provide added attachment surface for invertebrate prey. However, experiments designed to specifically test this hypothesis are required to make concrete conclusions related to invertebrate use of submerged timber in Ootsa Lake. Treed areas may serve as shelter for young rainbow trout and northern pikeminnow for which the main predators (excluding humans) in this system are osprey (Pandion haliaetus) (Lloyd 1999) and large northern pikeminnow that 27 cannibalize young of their own species (L. Brooks, unpublished data). Given their opportunistic feeding behaviour and their reputation as a salmonid predator in lakes (Brown and Mayle 1981 ), it is likely that northern pikeminnow also prey on small rainbow trout in Ootsa Lake. Complex underwater structure is often chosen by young rainbow trout and other Oncorhynchus spp. as refuge from predators (Culp et al. 1996; Tabor and Wurtsbaugh 1991; Shirvell 1990) and as a result, juvenile mortality is reduced (Quinn and Peterson 1996). Dense structure also reduces predator efficiency (Crowder and Cooper 1982). Harvesting does not appear to affect pelagic northern pikeminnow abundance, but it does result in a high number of that species in benthic areas. Possible reasons for high abundance of northern pikeminnow in benthic harvested areas include increased food availability due to suspended benthos, presence of previously unavailable predator refuge (overturned root wads), or physicochemical changes associated with harvest activities. Fish abundance patterns in near-shore areas both among and within lakes can be influenced by environmental factors including temperature (Brandt et al. 1980), dissolved oxygen (Tonn and Magnuson 1982), and transparency (Marshall and Ryan 1987). Based on my water quality observations (see Chapter 2) and those by Perrin et al. (1997), effects of harvesting on water quality are short-lived. Therefore, introduction of an unexploited niche with previously unavailable food and shelter resources may be an influential factor responsible for high abundance of northern pikeminnow in benthic harvested sites. 28 Seasonal patterns of fish distribution have been documented in both natural lakes (Tonn and Magnuson 1982) and man-made reservoirs (Gelwick and Matthews 1990). The temporal variation in abundance of northern pikeminnow and kokanee in the limnetic zone of Ootsa Lake may be explained by typical seasonal migration patterns observed in reservoirs and natural lakes, respectively. In the fall, northern pikeminnow move offshore into deeper water (Martinelli and Shively 1997) where fish become a major food item (Scott and Crossman 1998), whereas kokanee move onshore toward spawning areas (Lorz and Northcote 1965). Reasons for rainbow trout migration out of near-shore areas in the fall remain unclear. Although they are typically spring spawners (Scott and Crossman 1998), movement by rainbow trout upstream for fall spawning is known to occur in eastern North American lakes (So reman 1981 ). However, factors related to temperature and food changes associated with winter survival are a more likely explanation for the decrease in rainbow trout catches during the fall in Ootsa Lake. Diversity There was no consistent relationship between fish species diversity and habitat structural complexity. Fall values were positively associated with structural complexity of habitat, as predicted, but summer indices were opposite to those predicted . Therefore, the extent to which higher structural heterogeneity supports 29 higher species diversity is limited in this system and appears to be dependent on seasonal changes. It is generally agreed that species diversity will be higher in more heterogeneous habitats because similar species are able to coexist by utilizing different microhabitats (Pielou 1975). Shannon-Wiener's index of diversity accounts for both species abundance and evenness. Communities that have equal numbers of individuals in each species will have a higher index (H') value than those that have highly variable abundance across species (Pielou 1975). Therefore, higher summer diversity in open areas of Ootsa Lake can be explained by the more even representation of the predominant species observed (rainbow trout, kokanee, and northern pikeminnow). In contrast, treed and harvested habitats had a very high number of northern pikeminnow (i.e. less even representation of species), which would cause a lower diversity index. Differences in the fall can be similarly explained in that the open habitat had a very high number of kokanee compared to a more equal abundance of the three main species in the treed and harvested habitats. Because there is a shift in evenness of species in the treed and harvested habitats from uneven in the summer to more even in the fall, and a shift in the open habitat from even in the summer to less even in the fall, overall abundance of the three predominant species across habitats is fairly equal. Therefore, it is not surprising that differences in overall diversity among habitats were not significant. Burbot, Rocky Mountain whitefish, prickly sculpin, largescale 30 suckers, and longnose suckers were caught in such low numbers in all habitats that their abundance is close to equal in all habitats in both seasons. The importance of specific habitat features to fish is variable seasonally and latitudinally. For example, Tonn and Magnuson (1982) reported that winter fish assemblages were related to oxygen concentrations and refuge availability, whereas vegetation diversity was the main factor related to species richness in the summer. Eadie and Keast (1984) noted that fish diversity in northern Ontario lakes was correlated to diversity of benthic prey and not physical heterogeneity of habitats; fish species diversity in southern lakes was correlated to several factors of habitat complexity including substrate diversity and vertical complexity of vegetation. Mean depth, lake size, and latitude have been associated with fish species diversity among lakes (Marshall and Ryan 1987; Barbour and Brown 1974). Within-lakes, thermal regimes are sometimes responsible for habitat partitioning (Brandt et al. 1980), and variation in depth gradient can be responsible for spatial heterogeneity of fish communities (Benson and Magnuson 1992). Of the within-lake possibilities, aquatic vegetation diversity, mean depth, and depth gradient can probably be excluded as contributing factors to differences in species richness among habitats in my study because these factors were all similar across habitats. Temperature, oxygen, and substrate differences, however, may have influenced species diversity in this study. It is possible, 31 therefore, that these factors combined with seasonal migration behaviour and structural heterogeneity all contribute at some level to the differences in species diversity observed in this study. Further, because the benthic sampling with traps may have influenced abundance estimates of the actual fish assemblage, it is difficult to formulate concrete conclusions related to diversity. Two observations suggest that trap sampling may have been species-biased. First, SCUBA divers hired for this project in 1997 noted several sculpins in the ·roots of trees, whereas only two sculpins were captured over the entire sampling period in 1998. Second, exploratory net sampling on the bottom of the open habitat resulted in a much higher abundance of Rocky Mountain whitefish and suckers than the numbers represented by our sampling with traps. A more complete picture of species richness in benthic areas of each habitat may provide very different diversity results, although benthic net sampling in treed and harvested sites remains an unsolved problem. Several methods were attempted but each presented different challenges. Placement of 8 m sections of net required SCUBA divers for retrieval due to submerged branches and snags. The cost of divers under current Workers Compensation Board (WCB) regulations made this an unreasonable option. Hydroacoustic assessment was essentially impossible in the treed sites and did not provide species information. Set lines and angling were somewhat successful but very inefficient. Rotenone sampling was not attempted due to ethical considerations. Therefore, the use of these 32 diversity results for predicting associations of fish species richness and habitat structural heterogeneity is limited. 33 CHAPTER2 Differences in size and condition of fish in Ootsa Lake, BC in relation to changes in habitat structural complexity caused by submerged timber harvesting. ABSTRACT Fish were sampled in three habitats with varying levels of structural heterogeneity (treed =high heterogeneity; harvested =medium heterogeneity; open = low heterogeneity) to determine the size distribution and condition of fishes in relation to submerged timber and its removal in Ootsa Lake, BC (Nechako Reservoir). Ages and diet of fish and water quality parameters were also recorded to provide baseline descriptive information. Sizes of rainbow trout (Oncorhynchus mykiss), kokanee (Oncorhynchus nerka), and northern pikeminnow (Ptychochei/us oregonensis) were related to habitat structural complexity. The smallest fish occupied the treed habitat and the largest occurred in the open habitat. It is likely that smaller fish utilize treed areas for refuge from predators and/or because of high prey availability. Fulton's condition factor was not associated with habitat structural complexity, but differed among sites within habitat. Kokanee were exceptionally small (mean fork length =180 mm ±0.95) with over 85% of individuals being age 2+. They also exhibited characteristics similar to residual sockeye (Oncorhynchus nerka) including olive-black spawning colouration. I hypothesize that kokanee are exhibiting adaptive life history 34 patterns in response to the cold, oligotrophic conditions in the large reservoir. Future research should focus on locating kokanee spawning sites and on kokanee life history patterns in relation to lake productivity. Management strategies should include maintenance of areas of submerged timber for juvenile fish habitat, and the potential for a fertilization project to improve fish production should be examined. 35 INTRODUCTION Damming of streams to create impoundments results in a typical productivity cycle of "boom and bust" within the resultant reservoir (Ney 1996). Initially, productivity is high due to the influx of nutrients such as phosphorus from inundated land ("boom"). After approximately 5 years, those newly introduced nutrients are slowly depleted with little source of renewal and the waters return to their pre-impoundment productivity status. Productivity in reservoirs continues to decline due to sediment retention, phosphorous loss from discharge, and decreased carbon production from the littoral zone due to drawdown effects. Eight to fifteen years post-impoundment, reservoirs typically reach an ultraoligotrophic "bust" in which productivity is at a minimum (Stockner et al. 2000). BC's reservoirs, including the Nechako Reservoir, are currently in the "bust" part of the productivity cycle (Ken Ashley, personal communication). In 1951, prior to flooding, Lyons and Larkin (1952) described Ootsa Lake as "moderately productive with excellent conditions for producing trout". In 1992, 40 years after impoundment, Perrin (1992) reported that Ootsa Lake was ultra-oligotrophic with an extremely low supply of nutrients and plankton biomass. Five years later, the reservoir as a whole was characterized as oligotrophic with extremely low nutrient levels (Perrin et al. 1997). 36 .As phosphorus levels decline in reservoirs, primary production decreases and the result is manifested through the food chain (Stockner 2000). At the top of the food chain, fish growth and production decrease in response to the trophic effects of oligotrophication (Ellis 1941, Ney 1996, Lindstrom 1973, Stockner 1987). Notable declines in fish production have been observed in several of BC's oligotrophic reservoirs including Kootenay Lake (Ashley et al. 1997) and Arrow Reservoir (Pieters et al. 1998). Kokanee populations in particular have been decreasing dramatically in reservoirs, in large part because of extremely low productivity caused by impoundment, but also because of introduction of an exotic mysid shrimp (Mysis re/icta) (Ashley et al. 1997, Ashley and Shepherd 1996). Monitoring productivity at multiple trophic levels over time is an important research tool that provides essential information to fisheries managers who depend on such information to initiate effective management strategies. For example, knowledge of oligotrophication and consequent decline of kokanee in Kootenay Lake, BC prompted managers to undertake a fertilization project to counteract the decrease in productivity caused by previous human intervention (Ashley et al. 1997). The results have been successful enough that similar projects have been initiated in several other oligotrophic reservoirs and lakes in BC, particularly those where kokanee are thought to be a keystone species (Paine 1969) (Ken Ashley, personal communication). 37 Because the Nechako Reservoir is in a fairly remote northern location, it has not been influenced by human impacts to the same degree as southern BC reservoirs. For example, residential development around the reservoir is minimal as is recreational fishing and boating. Also, exotic species have not been introduced into the Nechako Reservoir, unlike those reservoirs in the south. However, submerged timber harvesting and the possible construction of a coldwater release facility at the Kenney Dam create potential for further major perturbations to this ecosystem. Therefore, information regarding its biological status is required. These elements make the Nechako Reservoir an important ecosystem for research in the fisheries and aquatic sciences. Alcan owns the water rights to the Nechako Reservoir, and because there were no management requirements for the reservoir outlined in the agreements between Alcan and the provincial government (see Background Information), information regarding the status of fish and productivity in the Nechako Reservoir is minimal. For this reason, the Department of Fisheries and Oceans (DFO) and the Ministry of Environment, Lands and Parks (MELP) had concerns about the impact that submerged timber removal would have on the fish community in the reservoir. Therefore, baseline information regarding both fish ecological characteristics and water quality were required. The purpose of this study was threefold. First, I wanted to provide baseline data on age and diet of fishes in Ootsa Lake as well as to describe baseline water 38 quality. Statistical analysis was not conducted on these descriptive data because age, diet, and water quality were not the prime foci of this study. Rather, descriptive variation among habitats for these parameters is presented. Because almost no information was available on the Nechako Reservoir, my results will provide a basis for the development of future research projects on the reservoir. They also will provide information that will enable researchers to monitor changes in fish growth and production over time and to track productivity changes in the reservoir. Second, I wanted to evaluate whether or not fish exhibit habitat segregation according to body size and whether fish condition is related to habitat structural complexity. More specifically, I wanted to evaluate the effects that harvesting of submerged timber in Ootsa Lake might have on fishes by examining differences in size and condition factor of fish found in habitats with different levels of habitat structural complexity (treed, harvested, open- See Chapter 1 for further habitat descriptions). Underwater structure is known to provide refuge from predators for small fishes. For example, Shirvell (1990) found that juvenile steelhead trout (Oncorhynchus mykiss) and coho salmon (0. kisutch) selected areas with structurally complex rootwads because they provided protection from predators. Density and biomass of young rainbow trout (0. mykiss) increases at sites where simulated complex woody debris is experimentally added (Culp et al. 1996). Juvenile rainbow trout are abundant in structurally complex inshore habitats where presence of cover decreases predation rates (Tabor and Wurtsbaugh 39 1991 ). Therefore, I predicted that fish in the treed habitat would be smaller than those in the open habitat. Condition factor is a ratio of weight to length. It is considered to reflect fatness with higher condition factors reflecting greater weight for a given length. Condition factor has also been used as a reflection of growth (Filbert and Hawkins 1995), and growth (or fatness) is correlated with food availability (Ensign and Strange 1990). Higher growth rates in fish have been linked to the presence of aquatic vegetation (Crowder and Cooper 1982) and artificial structure (Wege and Anderson 1979) that provides added attachment substrate for aquatic invertebrates (Mclachlin 1970). In addition, submerged timber is thought to increase the food base for fish (Poddubnyy and Fortunatov 1961; VanDenAvyle and Petering, 1988, Mclachlin 1970). Because of these relationships, I predicted that fish would have a higher condition factor in the treed than open habitats. If fish respond quickly to harvesting, I predicted intermediate levels of size and condition in the harvested habitat. The null hypothesis was that there is no difference in size or condition factor of fishes among habitats (treed, harvested, open). Third, I report unusual characteristics of the kokanee population that were discovered incidentally, but which will provide future researchers with basic information on this important species. 40 MATERIALS AND METHODS Sampling 1997 Sampling Three treed (T1, T2, and T3) and two open (01 and 02) sites were identified in August 1997 as suitable sampling sites (See Chapter 1, Figure 1.1 ). Monofilament gill nets in the experimental gang configuration (Hubert 1996) were set on the surface at these near-shore sites ( 100 - 200 m from shore) between September 4 and November 2, 1997. Nets were set at dusk and retrieved at dawn. With the exception of net panel length, all other aspects of fish sampling were the same as detailed for 1998 sampling described in Chapter 1. Only data related to kokanee age, fecundity, egg size and sex ratio were analyzed from this sampling year. Due to the limited field season and inconsistent sampling regimes, other data from 1997 did not provide useful analytical information so it was, therefore, excluded from any analysis . .1998 Sampling Sampling methods for 1998 are described in Chapter 1 Material and Methods. All analyses included data from this sampling year with the exception of kokanee egg size. 41 Data Collection and Analysis- Descriptive Baseline Information Length-weight relationships Fork length (mm) and weight (g) were measured for each fish in the field with digital scales (CT200 and CT1200, Ohaus Corporation, Florham Park, NJ) (up to 1200 g) or with a spring scale (T-20, Accu-Weigh, USA) (over 1200 g). Length and weight values were log-transformed to produce a normal distribution and then plotted against each other in a regression analysis with SYSTAT statistical software (SPSS Inc., Chicago, IL) to estimate the relationship between length ·and weight of each species in each habitat and overall. The regression coefficient (b) (i.e. the slope of the regression line) was used as an indicator of condition. Values of b that are greater than 3.0 represent fish that are fatter at a given length than those fish with b values less than 3.0, and a b value equal to 3.0 represents isometric growth. This index can also be a robust predictor of fecundity, reproduction, and growth (Anderson and Neumann 1996). Analysis of covariance (ANCOVA) in SYSTAT was used to test for homogeneity of slopes among habitats for each species. Length at Age For all fish, scales were removed from the area of skin between the lateral line and the dorsal fin on the left side of each fish; otoliths were also removed. Age estimates were made by Birkenhead Scale Analysis3 with standard techniques (Mackay et al. 1990). For analysis, I generally used the ages that were 3 C.41 McMillan RR 1, Lone Butte, BC, Canada, VOK 1XO, (250) 395-3880, birksc@bcinternet.net 42 determined from otoliths because scales can sometimes be resorbed or regenerated and, therefore, inaccurate. In those few cases where otoliths were damaged or unavailable the scale age was used, but only if the scale was in good condition. Mean length at age was plotted for each species in each habitat to provide insight into age structure related to habitat. Diet Stomachs were removed from each fish and preserved in 70% ethanol. Up to 20 fish of each species from each sampling event were analyzed for stomach contents. Contents were removed with a probe and squirt bottle and the contents were diluted to a known volume. An appropriate subsample was examined (4 ml for full stomachs, 10 ml for- half full, or 20 ml if- less than half full). In cases where there was little evidence of food items, the entire contents of the stomach were examined in a dilution of 100 mi. Food items were counted and total numbers were estimated with the following formula: T=V;xN Vs where T =total # of food items; V; = the initial dilution volume; Vs =the subsample volume; and N = the number food items counted in the subsample (Bowen 1996). 43 The items were then compiled into the 6 categories listed in Table 2-1. The mean frequency of occurrence (the proportion of fish that contained one or more of a given food type) and mean percent composition by number (the number of items of a given food type expressed as a percentage of the total number of food items counted) (Bowen 1996) were calculated for each species in each habitat to evaluate differences in the proportion of fish eating certain food items and in the abundance of the different groups of food , respectively. For mean percent composition by number calculations, only items that could be ·considered individual organisms were included (i.e. plant and fish fragments were not included). For insects, one head , 6 legs, or 4 wings was considered as one individual, and for molluscs 50 mollusc fragments were conservatively estimated to be one individual (based on visual observation). Kokanee Age at Maturity Upon dissection, fish were identified as male or female and examined for maturity. Each fish was ranked from 1 to 4 where 1 =immature (either unable to determine sex or small stringy gonad); 2 =maturing (eggs or testes obvious but not well developed); 3 = mature (well-developed gonads but not expressing eggs or milt); and 4 = ripe (males expressing milt; eggs distinct and separated) (Resource Inventory Committee (RIC) 1997). Age class percentages and percentage of each age class for those fish considered mature or ripe were calculated only for kokanee because the age composition of this population was 44 ~ 01 Zooplankton Cladocera Cladoceran Egg Case Copepoda Plant Debris Algae Plant Debris Woody Debris Fragments Plecoptera (stoneflies) Unidentified aquatic and terrestrial insects Hymenoptera (wasps) Hemiptera (true bugs, corixidae, homoptera) Coleoptera larvae and adults Diptera (flies and midges) larvae, pupae, adults Partially Aquatic Insects Trichoptera (caddisflies) larvae, pupae, adults Bones Scales Whole fish Fish Odonata (dragon and damselflies) Bivalvia (clams) Gastropoda (snails) Aquatic Insects Ephemeroptera (mayflies) Molluscs Aquatic Insects Table 2-1. Diet categories for stomach content analysis of fish from Ootsa Lake, BC (July- October, 1998). Bryozoa (statoblasts) Hydrachnidia (water mites) Arachnidia (spiders) Other noted as unusual. Observations on internal and external parasites were also noted during dissection. Kokanee Fecundity, Egg Size and Sex Ratio During data collection in 1997, fish were not assigned a maturity ranking (see above) based on visual inspection of gonad development as they were in 1998. To enable a common estimator of maturity to be applied, a post-dissection ranking was assigned to the 1997 kokanee based on data from 1998 sampling. This ranking was determined as follows. Gonadosomatic index (GSI) (gonad mass I total body mass) was calculated for both 1997 and 1998 samples. Minimum GSI values were determined for each of the maturity ranks assigned in 1998. These values were then used to divide the 1997 GSI values into maturity rank categories that were similar to the 1998 ran kings (i.e. each category would have the same minimum GSI value for both 1997 and 1998 samples). Specifically, males with GSI < 0.023 and females with GSI < 0.019 were assigned a "maturing" (2) ranking; males were designated "mature" (3) if their GSI was between 0.023 and 0.043 and females were considered mature for GSI between 0.019 and 0.077; and males were classified as "ripe" (4) for GSI > 0.043 and females were given this classification if their GSI was > 0.077. Individuals from 1997 whose sex was unknown due to unsubstantial gonad development were assigned an "immature" (1) ranking. 46 In both 1997 and 1998, eggs of "ripe" females were counted to determine fecundity. Fecundity for "mature" females (where the eggs did not separate easily) was estimated in 1997 by weighing 20 eggs, and in 1998 by weighing about 10% of the total gonad mass and counting the number of eggs in that 10%. Atrophied eggs were not counted. Fecundity was then estimated for mature females in both years with the following equation: F =(a +b)xc where F =fecundity (total number of eggs), a= the subsample number of eggs weighed, b = mass of the subsample of eggs, and c =total gonad mass (Grim and Glebe 1990). Average egg diameter for ripe kokanee was determined in 1997 by measuring the total distance occupied by 10 - 20 eggs arranged side by side in a row and then dividing the length by the total number of eggs measured. Overall mean egg diameter was then calculated. Egg diameter was not measured in 1998 samples. Overall sex ratio of males to females was also calculated. Sex ratios for both summer and fall seasons were calculated for 1998 samples, but summer sampling did not occur in 1997 so only fall values were available. Deviation from a 1:1 sex ratio was tested by x2 analysis (Sokal and Rohlf 1995). Unknowns were ignored in the analysis, but they accounted for less than 10% of the total sample size and were usually juveniles with no clear gonad development. All analyses 47 were performed with Statistical Package for Social Sciences (SPSS Inc., Chicago, IL). Water Quality Temperature and oxygen profiles of each site were taken on August 15, 1998 between 10:00 and 15:00. Values were recorded every metre as the probe (Handy Mk Ill, OxyGuard, Canada) was submerged and again as it was retrieved. Values at each depth were averaged and graphed. Secchi disk transparency readings were recorded (Wetzel 1983) on August 14, 1998 between 11:15 AM and 12:50 PM. Two water samples were taken from each site. The first was taken at a 1-m depth from the surface and the second was taken one metre from the bottom with a VanDorn water sampler (Wildco, USA). In the lab, each sample was mixed well and then filtered to determine the amount of suspended solids (mg/L) in each sample according to Greenberg et al. (1992). Two substrate samples were retrieved from each site with a standard Eckman .dredge (Wildco, USA). Samples were combined and frozen for later analysis in the lab. After thawing, samples were described qualitatively and then three separate sub-samples from each site were weighed (wet weight) and dried at 104°C. Dried samples were weighed (dry weight), then placed in an ashing oven at 550°C, and removed once all material had turned powdery white (Greenberg et al. 1992). Ashed samples were weighed (ash weight) and percent organic matter was calculated with the following formula: 48 OM= DW -AW x lOO DW where OM= percent organic matter; DW = dry weight (g) of sample; AW = ash weight (g) of sample (Greenberg et al. 1992) Means were calculated for each site to provide descriptive baseline information. Statistical Analysis & Hypothesis Testing -Size and Fulton's Condition Factor Fulton's condition factor (K) was calculated for each fish with the following formula: w K = - 3 X 100,000 L where W= weight (g) and L = length (mm) (Anderson and Neumann 1996). Within each species, condition factors were only compared among fish of similar length because longer fish tend to have a higher condition factor due to increasing rate of weight gain compared to length as the fish ages (Anderson and Neumann 1996). The following size ranges were analyzed: rainbow trout (230- 49 400 mm), kokanee (120- 240 mm), and northern pikeminnow (1 00- 300 mm). These ranges were determined for each species separately. The length range chosen for statistical comparisons was that where scatter around the condition factor (K) was equally distributed as determined from examining plots of length vs. condition factor (Dieter Ayers4 , personal communication). A nested AN OVA design was used to test the null hypothesis that there is no difference in length, weight or condition factor of fish among sexes, habitats, or seasons. For northern pikeminnow, differences among capture depths (surface vs. bottom) were also compared. Second order interactions (between habitat, season, sex, and capture depth) were tested, but third order interactions were not readily interpretable and were incorporated into the error term. The nested design was used because it enabled evaluation of site effects. Variables in the model were log1 0 length, log1 0weight, and condition factor (K). The main factor in the nested ANOVA design was habitat (treed, harvested, open), which was fixed. There were three replicate samples (sites) in each habitat. Sites were random factors and were nested within habitat. Other factors included in the model were sex and season for each species (rainbow trout, kokanee, and northern pikeminnow). Depth of capture (surface or bottom) was included for northern pikeminnow only because it was the only species caught in sufficient numbers (n=137) during bottom sampling. These factors were included 4 Statistical Consultant, UNBC, 3333 University Way, Prince George BC V2N 4Z9 50 to provide insights into spatial and temporal segregation of different sizes of fish ~ relation to each habitat. The model statement for the AN OVA was: y =Jl + ai + {Jj(l) + Ak+ &+ Tm+ ( aA.)ik + ( a5)il + ( a-r)im + (r5Jkl + (y-r)km + (8-r)Jm + 8j(J)klmn where Jl is the parametric mean of the population, ai is habitat, pj(l) is the effect of site within habitat, y is season, 8 is sex, 't is capture depth, and Ej(JJklmn is the error term. For analysis of rainbow trout and kokanee, the model was the same except that all terms with 't (capture depth) were excluded. RESULTS Descriptive Baseline Information For clarity, all tables and figures related to descriptive baseline information are included in the Appendix. Length-weight Relationships There was a significant relationship between log1 0length and log1 0weight in each habitat and overall for rainbow trout (p<0.001) (Figures A.1 and A.2), kokanee (p<0.001) (Figure A.3 and A.4 ), and northern pikeminnow (p<0.001) (Figure A.5 and A.6). Using the regression coefficient b as an indicator of condition, tests for 51 homogeneity of slopes indicated that the empirical condition factors for kokanee and northern pikeminnow varied among habitats (p<0.05; p<0.001, respectively). Among habitat differences for rainbow trout were at the margin of significance (p Q) .....J 15 / _ _L __ / 61 54 117 27 ------ Treed ____. - Harvested -----A- Open 160 155 150 Fall Summer 80 75 70 65 -- 55 .0> 60 ..c 0> Q) 5: 50 45 40 35 30 b) 24 / --7- / 61 54 117 / 15 27 ------ Treed ____. - Harvested -----A- Open Summer Fall Figure 2.2. Interactions between habitat and season for a) lengths and b) weights of kokanee. Analysis was done on log-transformed data. Interactions are significant for both length and weight (p<0.05). Numbers indicate sample sizes. Error bars represent + two SE. 62 ~ ~ - ~ ._ '- .9 (..) co u.. 33 1.1 26 45 58 62 14 21 17 c HABITAT 0 E "0 c 0 (.) c 0 1.0 I II • .::!:: Treed I :::::1 u.. e Harvested .9 ~ I ~ T1 ~ T2 ~ T3 H1 ~ H2 H3 01 02 03 ~ .A Open Site Figure 2.3. Comparison of mean Fulton condition factor of kokanee in each site sampled in Ootsa Lake, BC (July - October, 1998). T =treed habitat; H =harvested habitat; 0 =open habitat; and numbers represent different sites within each habitat. Condition varies among sites within all three habitats (p<0.001 ). Numbers above error bars indicate sample sizes . Error bars represent+ 2 SE. 63 There was also a significant increase in condition factor of kokanee in the fall from the summer (p<0.001) (Figure 2.4 ), again presumably due to increased weight associated with fall gonad development. Condition factor of kokanee differed among sexes (p<0.001) with males having higher condition than both females and unknowns, and females having higher condition than the unknowns {Figure 2.5). There was no interaction between habitat and season for kokanee condition factor (Table 2-3). There was an interaction for condition factor between habitat and sex (p<0 .01 ), where unknowns had a much higher condition factor in the open habitat compared to the other habitats (Figure 2.6). The interaction between season and sex indicates that males across seasons showed almost no change in condition whereas both females and unknowns showed a dramatic increase over time (Figure 2. 7). Northern pikeminnow Size of northern pikeminnow varied among habitats (p<0.05) (Table 2-4 ). They were significantly smaller in length and weight in the treed (197.9 ± 2.6 mm; 85.9 ± 5.5 g; n=268) and harvested (202.3 ± 3.6 mm; 111.4 ± 9.5 g; n=247) habitats than the open (274.3 ± 17.4 mm; 351.1 ± 68.3 g; n=33). There was no difference in length of northern pikeminnow within habitats among sites (Table 2-4), but there was a difference in weight among sites within the harvested habitat (p<0 .05) where the mean weight in H2 (130.9 ± 24.3 g; n=47) was higher than 64 1.10 -- 1.08 co 1.04 ~ 232 1.06 I.... 0 ....... (.) LL c 0 E 65 1.02 "'0 c 0 () c 0 ~ ::J LL 1.00 II .98 .96 .94 Summer Fall Season Figure 2.4. Comparison of mean Fulton condition factor for kokanee across seasons in Ootsa Lake, BC (July - October, 1998). Differences are significant (p<0.001 ). Numbers indicate sample sizes. Error bars represent+ 2 SE. 65 - 1.2 -~ 165 I- 0 +-' (.) m 1.1 LL c + 0 :;::. ""0 c 0 0 c 0 1.0 I 104 .::::! 28 II :::J LL .9 Male Female Unknown Sex Figure 2.5. Comparison of mean Fulton condition factor of kokanee among sexes in Ootsa Lake, BC (July- October, 1998). Condition varies among sexes (p<0.001 ). Numbers indicate samples sizes. Error bars represent+ 2 SE. 66 -----Treed -II - Harvested __.___Open 1.16 ........... 1.12 30 1.08 87 48 ~ I.... 0 t5 ro LL c 0 ......._,_ ~ 8 ~ 1.04 ~ 1.00 ~ ~ "'C c 0 u c 0 ~ ~ 0.96 9 12 0.92 ::I LL 0.88 0.84 0.80 Male Female Unknown SEX Figure 2.6. Interaction between habitat and sex for mean Fulton condition factor of kokanee in Ootsa Lake, BC (July- October, 1998). The interaction is significant (p<0.01 ). Numbers indicate sample sizes. Error bars represent +two SE. 67 1.15 - -::::.::: I.... 0 ....... 1.10 - 21 • .__ ~ ~ 74 1.05 - (.) ctl LL c 1.00 - 0 +J "'C c 0 u c 0 :!:: ::J LL 0.95 0.90 - / / 30 .r / 15 J ( / / / / / / / ~~ 14 / ------ Male __. _ Female __.__ Unknown 0.85 0.80 144 ~~ ~~ Summer Fall SEASON Figure 2.7. Interaction between season and sex for mean Fulton condition factor of kokanee in Ootsa Lake, BC (July- October, 1998). The interaction is significant (p<0.001 ). Numbers indicate samples sizes. Error bars represent +two SE. 68 ~ ~ c.o Ol • p=0.0001 p=0.101 p=0.134 • p=0.002 • p=0.017 • p=0.0001 • p=0.027 p=0.954 condition factor • p=0.0001 log 10 weight p=0.583 p=0.081 • p=0.026 log 10 1ength Sex Site Habitat Northern Pikeminnow Season p=0.774 • p=0.005 • p=0.008 p=0.792 p=0.496 p=0.542 p=0.094 • p=0.0001 • p=0.0003 p=0.329 • p=0.0001 • p=0.0002 • p=0.048 p=0.397 p=0.566 Depth (Surf/Bot) Habitat • Season Habitat • Sex Season • Sex Habitat • Depth p=0.060 p=0.752 p=0.848 Season • Depth p=0.424 • p=0.0001 • p=0.0001 Sex•Depth Table 2-4. Results of nested ANOVA for northern pikeminnow from Ootsa Lake, BC (July- October, 1998). *indicates where differences are significant. Bonferroni post-hoc tests were used to determine where differences occurred. that of H3 (97 .0 ±15.1 g; n=71 ), and those from the open were intermediate in weight (112.2 ± 13.5 g; n=129). Northern pikeminnow did not differ in length or weight across seasons but they did differ among sexes (p<0.001) (Table 2-4). Males (213.0 ± 2 .8 mm; 99.7 ± 4.4 g; n=117) and females (285.3 ± 8.5 mm; 330.4 ± 33.2 g; n=89) were larger in length and weight than the unknowns (180.5 ± 1.6 mm; 61.6 ± 2.2 g; n=342), and females were longer and heavier than males. Individuals caught on the bottom were significantly smaller in both length (191.3 ± 3.1 mm; n=137) and weight (78.2 ± 4.3 g; n=137) than those caught on the surface (208.8 ± 3.0 mm; 125.1 ± 9.1 g; n=411) (p<0.001) (Table 2-4 ). There was no interaction between habitat and depth or between season and depth (Table 2-4). There was an interaction for both length (Figure 2.8a) and weight (Figure 2.8b) between habitat and sex which indicates that males and unknowns show similar changes in size across habitats but that females are much larger in the open habitat. The interaction between season and sex (p<0.001) (Table 2-4) indicates that both males and unknowns caught in the summer were somewhat larger in length and weight than those caught in the fall, but the female trend was the opposite (Figures 2.9a,b ). Finally, the interaction between depth and sex (p<0.001) (Table 2-4) indicates that males and unknowns are slightly larger in length (Figure 2.1 Oa) and weight (Figure 2.1 Ob) on the surface, but the females are dramatically larger on the surface than at the bottom (Figures 2.10a,b). 70 400 a) -e- Treed 350 .- - E 300 E I £. ........ C) c Q) _J ____. - Harvested ___.__ Open 17 250 57 54 6 200 I I I \ \ \ 37 \ I \ 1/ ~ \ \ 1/ 35 z ~ \\ \ 150 Male 900 800 Female 10 179 153 Unknown b) -e- Treed 700 ---- - Harvested ___.__ Open 600 .- - 500 C) ........ £. C) 400 s: 300 Q) 200 100 0 57 (harvested) 1o(open) 179(treed) 153(harvested) 54(treed) 6(open) Male Female Unknown Figure 2.8. Interactions between habitat and sex for a) length and b) weight of northern pikeminnow in Ootsa Lake, BC (July-October, 1998). Analysis was done on log-transformed data. Interactions are significant for both length and weight (p<0.001 ). Numbers indicate sample sizes. Error bars represent+ two SE. 71 400 - - a) 350 - E 300 E ......... .c C) c Q) _J r -- 250 - 58 200 - 88 • ---- ---- ----t 31 ---e- Male ! 29 ___. - Female --.t.- Unknown 312 ._ _ _ _ _ _ -130 150 100 700 - -C) ........ ..c C) Q) ~ I I I Summer Fall b) 600 500 400 300 200 100 0 sar / / / / •• 31 ---e- Male ___. - Female --.t.- Unknown ! 88 • 29 312 ._ - - - - - ~ 30 I I Summer Fall I Figure 2.9. Interactions between sex and season for a) length and b) weight of northern pikeminnow in Ootsa Lake, BC (July- October, 1998). Analysis was done on log-transformed data. Interaction is significant for both length and weight (p<0.001 ). Numbers indicate sample sizes. Error bars represent+ two SE. 72 340 - a) 320 - ~ 300 .......... 280 - ~ E E 260 ..._. - ..c C) 240 - Q) _J 220 - c -Male _. - Female -----6.- Unknown ~ ~ 76 200 180 - f 41 ! 76 266 ~ 160 ) • 20 -- I I I Surface Bottom DEPTH 500 450 400 350.......... C) 300 ..._. ..c 250 .Q> Q) 200 ~ 150 100 50 0 b) -Male _. - Female -----6.- Unknown 69 I ~ - '\ '\ '\ '\ 76 I 120 41 266 &--- - - - - - & 76 I I Surface Bottom I DEPTH Figure 2.1 0. Interactions between sex and depth for a) length and b) weight of northern pikeminnow in Ootsa Lake, BC (July- October, 1998). Analysis was done on log-transformed data. Interactions are significant for both length and weight (p<0 .001 ). Numbers indicate sample sizes. Error bars represent +two SE. 73 There was no difference in northern pikeminnow condition factor among habitats (Table 2-4 ). Site differences in condition factor within habitats occurred (p<0.001) in the treed habitat where fish in T1 had significantly lower condition factor than those in T2 and T3, and in the harvested habitat where fish in H2 had significantly lower condition factor than those in H1 and H3 (Figure 2.11 ). Northern pikeminnow had a significantly higher condition factor in the fall than in the summer (p<0 .01) (Figure 2.12). There was no difference in condition factor among sexes or among depths (Table 2-4). There was only one second order interaction, between habitat and depth (p<0.05) (Table 2-4) where fish in the treed habitat had a much higher condition factor on the bottom than the surface (Figure 2.13). In summary, all three species varied in size among habitats with smaller fish occupying the treed habitat. The smallest northern pikeminnow occurred in benthic areas. Condition factor did not differ among habitats, but was site specific and varied among sexes for kokanee and among seasons for kokanee and northern pikeminnow. Second order interactions suggest that the change in length and weight of kokanee across seasons was not independent of habitat, and that condition factor for each sex of this species varied among habitats and seasons. In addition, the variation in length and weight of the different sexes of northern pikeminnow was dependent on habitat, season and depth of capture, and the variation in condition factor on the surface or bottom was dependent on habitat. 74 - ~ -- 24 76 ::::c::: '- 0 -(.) co •• 69 7 12 120 43 1.0 u.. c HABITAT 0 :e "0 c 161 0 u I - • c 0 Treed I e Harvested .8 ~ I ~ T1 ~ T2 ~ T3 ~ H1 ~ H2 ~ H3 ~ 01 ~ 02 03 ~~ .A. Open Site Figure 2.11. Comparison of mean Fulton condition factor for Northern pikeminnow within habitat among sites in Ootsa Lake, BC (July- October, 1998}. Condition varies among sites within the treed and harvested habitats (p<0.001 ). Numbers indicate sample sizes . Error bars represent+ 2 SE. 75 ~ - ~ __. I.... .....0(.) ro u. c 0 ;e "0 c 0 ~ 1.06 I 1.04 1.02 74 1.00 () c 0 :: ::J .98 u. .96 .94 I 439 Summer Fall Season Figure 2.12. Comparison of mean Fulton condition factor for northern pikeminnow across seasons in Ootsa Lake, BC (July- October, 1998). Condition varies across seasons (p<0.01 ). Numbers indicate sample sizes. Error bars represent+ 2 SE. 76 - 1.08 - ~ ...._ -------Treed - Harvested ----A- Open 1.04- I.... 0 ....... u co LL c 1.00 - c 0 () 0.96- c 235 4 0 ....... ::::J LL --- •~ 911333 ---- - - - - _.1 20 ~ 0 :.;::; ""0 156 I 0.920.88 ~ ~ Surface Bottom DEPTH Figure 2.13. Interaction between habitat and depth for mean Fulton condition factor of northern pikeminnow in Ootsa Lake, BC (July- October, 1998). The interaction is significant (p<0.05). Numbers indicate sample sizes. Error bars represent + two SE. 77 ·DISCUSSION The baseline data presented here provide important indicators of current reservoir productivity. They also provide a reference for future estimates of biological productivity in Ootsa Lake. Analysis of the distribution of fish sizes revealed that small juvenile fishes utilized near-shore areas that have inundated timber and that very small northern pikeminnow utilized benthic harvested areas. Variation in condition factor, however, was not related to habitat structural complexity, but rather was more site-specific. Low productivity levels in lakes are indicated by comparatively small fish size in upper age classes compared to other populations (Stockner 1987; Lindstrom 1973; Ellis 1941 ), and by high transparency readings that suggest a low volume of phytoplankton (Wetzel 1983). The decrease in size of rainbow trout in Ootsa Lake since impoundment provides further evidence that fish productivity has decreased over the past 45- 50 years. In 1951, Lyons and Larkin (1952) noted that rainbow trout 2 years and older were of "legal size" which at that time was 8 inches (203 mm) (British Columbia Sport-Fishing Regulations 1952), whereas the mean length for age 2 fish in this study was 181 mm with only one age 2 rainbow over 200 mm. They also reported that fishing in Ootsa Lake was excellent and that rainbow trout had moderate growth with a relatively high condition factor (although no estimates were given). In contrast, the mean condition factor for rainbow trout in this study was relatively low for this species (Carlander 1969). 78 The mean length of mature kokanee in this study (145- 225 mm fork length; mean= 184 ± 0.79 mm; n=238) was also low compared to other northern BC populations (Chris Foote 5 , personal communication), in which individuals generally average between 203- 229 mm (Carlander 1969). Kokanee from Takla Lake, BC (near Ootsa Lake) have been known to mature between 170-220 mm (fork length) (Wood and Foote 1996). Upper age classes of northern pikeminnow are also small in size compared to other BC populations (Carlander 1969). Extremely low phosphorus levels (5.7 1-lg L- 1) (Perrin et al. 1997) and high Secchi disk transparency (8 - 9 m), provide further evidence of the highly oligotrophic state of Ootsa Lake. Evidence from other studies indicates that submerged timber increases food available to fish (VanDenAvyle and Petering 1988; Poddubnyy and Fortunatov 1961) and can provide increased attachment substrate for some aquatic insects (McLachlin 1970). My study was not designed to specifically address differences in fish diet, and the sampling regime that I used dictates against a detailed analysis. Fish could have been retained in the nets anywhere from a few minutes to several hours, possibly resulting in wide variation in the degree of stomach fullness (since fish do not eat while captured in gill nets but digestion continues). Because length of capture time in nets was not standardized (for example by retrieving nets every 30 minutes), results from statistical analyses of diet would have been misleading. Although the ability to draw confident conclusions 5 Instructor (Fisheries and Aquaculture), Malaspina University-College, 900 Fifth Street, Nanaimo, British Columbia, V9R 5S5 79 ·regarding diet is limited, nevertheless the descriptive results suggest that fish in harvested areas have fewer aquatic insects in their stomachs, which may reflect lower numbers of aquatic insects available in that habitat. This may be due to the decreased volume of attachment substrate for insect larvae (Mclachlin 1970) especially for clingers (e.g. some Ephemeroptera, Trichoptera, Diptera) and climbers (e.g. Odonata) that require attachment to surfaces (Cummins and Merritt 1996). However, Trichoptera and molluscs were prevalent in the diet of rainbow trout and northern pikeminnow from both treed and harvested habitats. The particularly high abundance of molluscs and Trichoptera in fish from harvested areas suggests that harvesting operations may increase the availability of these food items. Dense submerged timber may provide refuge for invertebrate prey and reduce foraging efficiency of predators (Crowder and Cooper 1982). Therefore, removal of trees may make these prey items more vulnerable to fish, but the abundance would be temporary if the presence of submerged timber positively influenced their proliferation. Research directed at determining the importance of inundated trees to aquatic invertebrates, which are the basis of rainbow trout and northern pikeminnow diets in Ootsa Lake, would provide useful information. In relation to submerged timber harvesting, my results on the distribution of different sizes of fish are important because they establish which size classes of fish might be affected by harvest activities and whether removal of submerged timber might alter the seasonal growth patterns of fish. Rainbow trout and 80 northern pikeminnow caught in the treed areas were smaller than those in the open, but not in the harvested habitat. Only kokanee were smaller in treed than in harvested areas. Long-term impacts of harvesting are unclear from these data, but evidence from this study and others suggests that juvenile fishes benefit from complex underwater structure because it provides refuge from predators (Gulp et al. 1996; Tabor and Wurtsbaugh 1991; Shirvell 1990) and it enhances aquatic invertebrate production (Pardue and Neilsen 1979). Other impacts related to habitat segregation among different sizes of fish may occur, as indicated by the occurrence of smaller northern pikeminnow in the most recently harvested site (harvested for approximately 2 months) compared to the one that had been harvested one year previously. Because northern pikeminnow were very small and highly abundant in benthic areas of the harvested habitat (see Chapter 1), the smallest individuals of that species may benefit from harvesting activities, perhaps due to small refuge areas provided by root wads and woody debris. Although the biological implications of differential condition factors is hotly debated, some biologists believe that change in condition reflects growth patterns in fish (Filbert and Hawkins 1995). If this is true, then it is somewhat surprising that condition factors were not linked to habitat structural complexity since submerged timber is thought to improve the abundance of some invertebrates (Mclachlin 1970), and growth is generally correlated with food availability (Ensign and Strange 1990). In addition, presence of aquatic vegetation and artificial structures has been linked to higher growth rates in fish 81 (Crowder and Cooper 1982; Wege and Anderson 1979). Instead, condition factor varied among sites within habitats. Rainbow trout and northern pikeminnow condition factor was significantly lower in the site that had been harvested one year previously than in two that had been harvested only a few months previously. Kokanee also had a higher mean condition factor in the most recently harvested site compared to the one harvested one year prior to sampling. This evidence suggests that fish may initially benefit from harvest activities, possibly due to increased prey availability, but those benefits are short-lived. Therefore, long-term negative impacts of harvesting on fish growth and condition may occur. Site effects on kokanee condition factor were also evident in other habitats. Because kokanee spawn in the fall, however, condition factors may be influenced by gonad development and therefore the prediction of harvesting effects on the condition of this species must be made with caution. The low number of site replicates makes it difficult to draw firm conclusions regarding the long-term impacts of tree removal on fish condition. That northern pikeminnow exhibited lower mean condition factor in T1 than the other two treed sites and that site effects were also apparent for kokanee condition factor in all three habitats suggests that differences in condition may be attributed to factors other than habitat structural complexity. Research with multiple replicates of harvested sites through time would allow a more accurate estimate of long-term impacts of harvesting on fish growth and condition. Because there is controversy 82 over the biological interpretation of condition factor, direct measurements and comparisons of fish growth would allow more concrete conclusions regarding the impact of harvesting on fish well-being. I do not necessarily expect that fish caught in a particular habitat are long-term residents of those habitats, as seasonal movement of fish in temperate regions is well known. Mark-recapture efforts to determine movement patterns were unsuccessful due to difficulty catching sufficient numbers of undamaged fish from treed and harvested areas. However, lake resident rainbow trout generally do not utilize extensive territory (Scott and Crossman 1998) and Martinelli and Shively (1997) found that the majority of northern pikeminnow movement in Columbia River reservoirs were short-term movements of< 1 km, likely in response to foraging opportunities. Therefore, it is likely that these species reside, at least partially, in the habitat in which they were caught. Kokanee are known to make extensive daily vertical and onshore-offshore migrations in the summer, most likely associated with temperature and food availability (Scott and Crossman 1998), and fall migrations toward spawning streams (Lorz and Northcote 1965). Specific information on movement patterns by all three species in Ootsa Lake would help to clarify whether fish utilize a particular habitat for extended periods. Concern for kokanee in the reservoir should be at the forefront of management strategies. They are potentially a keystone species (Paine 1969) in ecosystems like Ootsa Lake (Ken Ashley, personal communication) where they forage largely 83 on zooplankton and, in turn, provide a main source of food for large rainbow trout (Lyons and Larkin 1952) and northern pikeminnow (Ricker 1941 ). Kokanee populations in southern BC have come under close investigation recently due to their declining abundance in oligotrophic reservoirs and lakes. Much of the decline is associated with human activities including industrial runoff, impoundment for agriculture or hydroelectric power, and introduction of exotic species (Ashley and Shepherd 1996). However, the decline has been stemmed in Kootenay Lake by the success of lake fertilization projects (Ashley et al. 1997). In contrast, northern populations have not been studied as extensively and, in the case of the Nechako Reservoir, they have received no serious attention. Kokanee in Ootsa Lake have the characteristics of residual sockeye in Cultus Lake, BC described by Ricker (1938). Residual sockeye are progeny of anadromous sockeye salmon (Oncorhynchus nerka) that do not migrate to the ocean, but instead remain in fresh water their entire life. Residuals in Cultus Lake are small in size, they mature at age 2 (in their third year), they spawn late in the season (October- December), they are heavily infested with Sa/mineola parasites, they exhibit an olive-black spawning colour, and they have a malebiased sex-ratio. The male biased sex-ratio is characteristic of residuals that exist sympatrically with anadromous sockeye (Foote, personal communication) and is not necessarily true of all residual populations (see Ricker 1959). 84 Winsby et al. (1998) found 87% of kokanee captured in late summer in Ootsa Lake to be age 2 with the remainder being three years of age. They also noted a male-biased (63%) sex ratio and that most kokanee had advanced gonad development with some in spawning condition (Winsby et al. 1997). Although gill nets can be selective for male kokanee due to development of secondary sex characteristics (e.g. hooked lower jaw and dorsal hump) (Foote, personal communication), these changes were not readily visible in kokanee from Ootsa Lake. Nevertheless, it is possible that only slight development (not detectable without specific measurements) of male sex characteristics contributed to the male-biased sex ratio. In addition, different behaviours between males and females may have been a contributing factor. For example, shore spawners in Okanagan Lake, BC have male-biased ratios on the spawning grounds because males tend to occupy the spawning area longer than females, making them more likely to be caught (Cannings et al. 1998). Therefore, it is possible that the malebiased sex ratio of kokanee from our study is not truly representative of the population as a whole. On a superficial level, these data suggest that residual sockeye exist in Ootsa Lake. However, sockeye salmon were not known to spawn historically above the confluence of the Nautley River (International Pacific Salmon Fisheries Commission, 1953) due to the high gradient and water velocity in the Nechako Canyon (Don Cadden 6 , personal communication). If these fish are not residual 6 Fish , Wildlife and Habitat Protection, Ministry of Environment, Lands and Parks, 1011-41h Avenue , Prince George, BC, V2L 3H9 85 sockeye but long-term resident kokanee, then the question remains as to why they exhibit similar life history characteristics to residuals. Kokanee populations in BC are known to have varying life history traits within and among populations. Differences in fecundity and egg size occur among spawners from two different tributaries of Upper Arrow Lake (Murray et al. 1989). Kokanee within the Okanagan Lake system have developed two separate spawning strategies; some spawn in inlet streams while others spawn on the rocky shoreline within the lake (Ashley and Shepherd 1996). Some kokanee populations spawn deep in lakes and turn black instead of red at spawning in response to lack of light penetration at greater depths (Chris Foote, personal communication). Environmental pressures are known to cause various changes in populations. In fishes, environmental factors can influence size (Beamish and Tandler 1990; Wooton 1984 ), sex determination (Conover et al. 1992; Craig et al. 1996), and reproductive characteristics including age and size at maturation (Wootton 1984; Stearns and Crandall 1984 ). Temperature is one of the most influential environmental variables because it is fundamental to poikilotherms and is often correlated with fish growth rate (Mayle and Cech 2000). Age and size at maturity are both genetically and environmentally determined (Stearns and Crandall 1984 ). In addition, it is predicted that organisms that experience lower growth due to environmental stress will experience a change in their age and size at 86 maturity such that any reduction in fitness associated with slower growth and smaller size will be minimized (Stearns and Crandall 1984 ). I hypothesize that factors associated with large, cold, oligotrophic systems are driving the early maturation of kokanee in Ootsa Lake. Cold temperatures and low food availability caused by low productivity result in slower growth (Wurtsbaugh and Cech 1980) and lower fecundity due to smaller size (Roff 1992). Therefore, there is potential for reduced fitness under these conditions (Stearns 1992). Early maturation is beneficial if it increases an organism's chances of surviving to reproductive age and also results in offspring being born and reproducing earlier. These factors result in higher fitness in an environment where reduced growth occurs (Stearns 1992). Although this study provides no direct evidence that early maturation of kokanee in Ootsa Lake results in improved fitness, indirect evidence suggests that it is likely an adaptive characteristic. The potential for somatic growth in upper age classes is limited due to low lake productivity and cold temperatures, and therefore, the reproductive benefits associated with larger size (i.e. higher fecundity) would, presumably, be relatively small compared to those gained from maturing at an early age (i.e. improved chance of survival to reproductive age, and offspring born and reproducing earlier). In other words, body volume of kokanee would not increase substantially during their fourth and fifth year of life, and as a result, the volume available for gonad development would be limited. 87 Consequently, fecundity (and therefore their fitness) would not increase significantly if maturity was delayed. Experiments designed to test this hypothesis are needed, however, to conclude with any degree of certainty that this trait is indeed adaptive. It is unlikely (though not impossible) that these kokanee are residual sockeye. A more probable explanation is that they are exhibiting similar characteristics to residual sockeye due to environmental pressures that cause slow growth and reduced fecundity. Early maturation has a profound effect on reproductive output (Roff 1992; Stearns 1992) and generation time, which in turn have consequences for evolutionary processes. The preliminary results presented here suggest that this population of kokanee could provide an excellent opportunity to investigate the relationship of lake productivity and life history adaptations. Kokanee spawning grounds in Ootsa Lake are currently unknown. The oliveblack colouration of some individuals leads me to believe that kokanee may be spawning deep in the lake on submerged river or lake basins or possibly near inundated springs. Shore spawning is unlikely to occur because of the effects that reservoir drawdown would have on incubating eggs, but this possibility should also be explored. Ecological and genetic comparisons with stream spawners from Andrews Creek on the northwest shore of Ootsa Lake may also provide answers regarding adaptive life history strategies. Once spawning areas 88 are identified, harvesting activities can be regulated to provide protection of vulnerable habitats. It is clear that productivity in the Nechako Reservoir is low. Future habitat alterations in the reservoir may produce short-term productivity benefits (as suggested by the results of this study). However, they may also have negative effects on productivity. For example, if a cold-water release facility that draws reservoir water at depth is constructed, it may have the effect of further depleting phosphorus availability. Other effects may be similar to those seen in southern BC reservoirs. The physical, chemical and biological status of the reservoir should be consistently monitored so that informed decisions regarding submerged timber harvesting, construction of a cold-water release facility, and fisheries management can be made. ·c oNCLUSIONS AND MANAGEMENT RECOMMENDATIONS Fish abundance and the distribution of different sizes of fish are related to habitat structural heterogeneity in Ootsa Lake. Submerged standing timber in near-shore areas influences spatial patterns of fish abundance and size. My results, combined with a review of the literature, make it a reasonable assumption that juvenile rainbow trout and northern pikeminnow are utilizing treed areas as 89 predator refuge and/or because of high prey availability. Very small northern pikeminnow may also be using benthic areas of harvested habitat for similar reasons. Because there was a high abundance of young rainbow trout and northern pikeminnow in the treed areas, and because submerged timber may potentially provide added attachment surface for invertebrates, retention of large areas of submerged timber for juvenile fish habitat is recommended. Retention of root wads and woody debris in harvested areas is also recommended since artificial structures are often added to aquatic environments to improve fish habitat (Wickam et al. 1973; Reeves et al. 1977; Prince and Maughan 1979; Helfman 1979; Paxton and Stevenson 1979). In addition, these areas may provide refuge for small northern pikeminnow. All woody debris from harvesting, however, should be well below the drawdown zone of 15ft (4.57 m) to avoid becoming a boating hazard. The most important issue that should be addressed is the status of the kokanee populations in the Nechako Reservoir. This species may have developed an alternative reproductive strategy that could provide an opportunity for researchers to investigate the relationship of lake productivity and life history adaptations. It is unknown if kokanee populations in the Nechako Reservoir are decreasing, but because kokanee are considered a keystone species by some biologists, their preservation is important to the ecosystem as a whole (Meffe and 90 Carroll 1994 ). Therefore, any disruption of kokanee spawning migrations or habitat caused by harvesting should be prevented. More specific information is needed on kokanee spawning sites and patterns within the reservoir, but based on my preliminary results, near-shore harvest activities should decrease between September and December to avoid disruption of kokanee migrations, and harvesting near potential kokanee spawning areas should be prohibited between September and December. If further management strategies for kokanee are to be employed on the Nechako Reservoir, such as a fertilization project, it should be well researched before proceeding. An important question that needs to be addressed is the ability of the fishes in the Nechako Reservoir to survive and adapt to further severe changes to their ecosystem. Extensive harvesting of the lakes and construction of a cold water release facility at Kenney Dam are actions that may put fish under severe stress due to drastic ecosystem change. Any further perturbations to this system should be approached with caution and potential impacts to the reservoir's fish population should be considered. This study has provided a basis of knowledge on which to build, and several possibilities for future research have been suggested throughout the text. Any conclusive assessment of long-term impacts of harvesting on the fish distribution patterns in Ootsa Lake will require consistent monitoring over a longer period of time. 91 APPENDIX 92 3.0 2.8 a) Treed; slope=2.82 2.6 2.4 :§ :E 2.2 01 ~ 2.0 8' 1.8 ~ ..J 1.6 1.4 1.2 1.0 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.5 2.6 2.5 2.6 Rsq = 0.9769 Log 10 Length (mm) 3.0 2.8 b) Harvested; slope= 2.61 2.6 2.4 :§ :E 2.2 • • 01 ~ 2.0 8' 1.8 0 • ..J 1.6 1.4 1.2 1.0 2.0 2.1 2.2 2.3 2.4 Rsq = 0.9463 Log 10 Length (mm) 3.0 2.8 c) Open; slope=3.02 2.6 -9 :E: Cl ·a; 2.4 2.2 :s: 2.0 0 Cl 0 ....J 1.8 1.6 1.4 1.2 1.0 2.0 Rsq = 0.9573 2.1 2.2 2.3 2.4 Log 10 Length (mm) Figure A.1. Relationship between log1olength and log10weight of rainbow trout in a) treed b) harvested and c) open habitats in Ootsa Lake, BC (JulyOctober, 1998). p<0.001 for all habitats. 93 3.0 2.8 --C) ....... £. C) "(i) 0 ~ C) .....J • 2.6 2.4 2.2 2.0 ~ 0 slope= 2.77 .. 1.8 1.6 1.4 • 1.2 1.0 2.0 2.1 2.2 2.3 2.4 2.5 2.6 Rsq = 0.9664 Log 10 Length (mm) Figure A.2. Relationship between log1olength and log1oweight for rainbow trout in all habitats sampled in Ootsa Lake, BC (July - October, 1998). p<0.001 94 ~ a) Treed ; slope=3.36 2.2 2.0 § ~ 1.8 ~ ... 1.4 1.2 2.0 ~ 2.1 2.2 ~ 2.3 2.4 Rsq = 0.9085 Log 10 Length (mm) 2.4 , - - - - - - - - - - - - - - - - - - - - , b) Harvested; slope= 3.21 2.2 2.0 § ~ 1.8 ~ ~ 3 • 1.6 1.4 • • 1.2 2.0 ~ ~ 2.1 Rsq = 0.8691 2.2 2.3 2.4 Log 10 Length (mm) ~ c) Open ; slope=3.01 2.2 2.0 § :c ~ Cl ~ 8' 1.8 1.6 -' 1.4 1.2 2.0 2.1 ~ ~ 2.2 ~ Rsq = 0.9247 2.3 2.4 Log 10 Length (mm) Figure A.3. Relationship between log1olength and log1oweight of kokanee in the a) treed b) harvested and c) open habitats in Ootsa Lake, BC (JulyOctober, 1998). p<0.001 for all habitats. 95 ~ slope= 3.20 2.2 • 2.0 -- .......... Cl ..c Cl 1.8 "Q) ~ 0 ...Cl •• 1.6 •• 1.4 1.2 1.0 ~ 2.0 • •• •• • , •• 0 .....J • •• • ~ 2.1 2.2 2.3 ~ 2.4 Rsq =0.9034 Log 10 Length (mm) Figure A.4. Relationship between log1olength and log1oweight for kokanee in all habitats sampled in Ootsa Lake, BC (July - October, 1998). p<0 .001 96 3.5 a) Treed; slope=2 .95 3.0 ~ § 2.5 :E 01 ~ 2.0 ~ 8' ...J 1.5 1.0 .5 1.8 Rsq = 0.9458 1.9 2.0 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.6 2.7 2.8 2.6 2.7 2.8 Log 10 Length {mm) 3.5 b) Harvested; slope = 3.13 • 3.0 § 2.5 :E :;:"' ·a; 2.0 ~ "' 1.5 0 ...J 1.0 .5 1.8 Rsq = 0.9809 1.9 2.0 2.1 2.2 2.3 2.4 2.5 Log 10 Length {mm) 3.5 c) Open; slope=3.30 3.0 s 2.5 :E C> ·a; :;: 2.0 ~ "' 1.5 0 ...J 1.0 .5 1.8 Rsq = 0.9949 1.9 2.0 2.1 2.2 2.3 2.4 2.5 Log 10 Length {mm) Figure A.5. Relationship between log10length and log1oweight for northern pikeminnow in the a) treed b) harvested and c) open habitats in Ootsa Lake, BC (July- October, 1998). p<0.001 for all habitats. 97 3.5 slope= 3.12 • 3.0 -- 0) ...... • • 2.5 ..c 0> "(i) ~ 2.0 0 ...... 0) 0 _J 1.5 1.0 1.8 ~~~ 1.9 2.0 ~ 2.1 ~ 2.2 ~ 2.3 ~ 2.4 ~ 2.5 ~ 2.6 ~~ 2.7 Rsq = 0.9702 2.8 Log 10 Length (mm) Figure A.6. Relationship between log1olength and log1oweight for northern pikeminnow in all habitats sampled in Ootsa Lake, BC (July - October, 1998). p<0.001 98 500 450 -EE ............ £. ...... 0> c Q) _J 400 -.- •• 350 300 250 200 150 .... I !! I. • Treed I • 100 N= HABITAT 7 1 5 5 5 2311 6 8 7 10 3 4 2 1 1 2+ 3+ 4+ 5+ 6+ 7+ 1 Harvested I .... Open 8+ Age (years) Figure A.7. Mean fork length at age for rainbow trout in each habitat of Ootsa Lake, BC (July - October, 1998). Sample sizes are shown above each age class . Error bars represent + 2 SE. 99 240 220 - -,..- 200 ..c 0> (],) _J ~ I II E E 180 ............ c ~ HABITAT I 160 140 120 I Treed I • A 4 2 1+ Harvested I -- 100 N= • -- 71 68 126 2+ 5 7 3+ A Open 11 4+ Age (years) Figure A.8. Mean fork length at age for kokanee in each habitat of Ootsa Lake, BC (July- October, 1998). Sample sizes are shown above each age class . Error bars represent+ 2 SE. 100 500 , - - - - - - - - - - - - - - - - - - - - - - - - , a) Treed 400 .se 3oo .t:: c, :5 ....J 200 100 0 ~~~~~~ ~ ~~~ ~~~~~ ~~~ 3+ 5+ 7+ 6+ 4+ 9+ 8+ 11+ 10+ 13+ 12+ 15+ 14+ 17+ 16+ 23+ 21+ 26+ Age (years) 500 , - - - - - - - - - - - - - - - - - - - - - , b) Harvested 400 .se 3oo £C) :5 ....J 200 100 :X: • I :X: =-= ~~~~~~~~~~~~~~ N = 1 3 20 28 57 63 24 11 10 8 2 2+ 4+ 3+ 6+ 5+ 8+ 7+ 1 4 ~ 3 4 1 2 2 2 10+ 12+ 15+ 18+ 21+ 23+ 9+ 11+ 14+ 16+ 19+ 22+ Age (years) 500 , - - - - - - - - - - - - - - - - - - - - , c) Open 400 .se 3oo .t:: c, c: .3 200 100 ~~ N= 1 1 2+ 3 3 6+ 5+ 7 4 8+ 7+ ~~~~~~~~~~~~ 1 2 10+ 9+ 2 2 14+ 12+ 2 1 17+ 15+ 1 1 19+ 18+ 1 1 ~ 22+ 21+ 24+ Age (years) Figure A.9. Mean fork length at age for northern pikeminnow in the a) treed b) harvested and c) open habitats of Ootsa Lake, BC (JulyOctober, 1998). Sample sizes are shown above each age class. Error bars represent + 2 SE. 101 ~ ~ 400 - E E 300 HABITAT ..... 0) a5 200 _J ..!: I • Treed I 100 • Harvested I 4+ 2+ 3+ 6+ 5+ 8+ 7+ .... Open 10+ 9+ 12+ 11+ Age (years) Figure A.1 0. Comparison of mean fork length at age for northern pikeminnow aged 12 and under in each habitat of Ootsa Lake, BC (July - October, 1998). Sample sizes are equivalent to those shown in Figure A.9a-c. Error bars represent + 2 SE. 102 II Treed II Harvested DOpen 1.2 a) Rainbow Trout .,u .,c 0.8 t: "uu 0 Ci,., 0.6 ." u c cr !!? 0.4 IL 0.2 0 Plants 1.2 .. . u c t: " Zooplankton Aquatic Insects Molluscs Fish Other Zooplankton Aquatic lnsecls Molluscs Fish Other Aquatic Insects Molluscs Fish Other b) Kokanee 0.8 u u 0 Ci,., 0.6 ." u c cr 0.4 !!? IL 0.2 0 Plants 1.2 . u c ~ c) Northern Pikeminnow 0.8 " u u 0 Ci,., 0.6 ." u 0:: cr 0.4 I!! IL 0.2 0 Plants Zooplankton Food Item Figure A.11. Comparison of frequency of occurrence of food items for a) rainbow trout b) kokanee and c) northern pikeminnow in each habitat sampled in Ootsa Lake, BC (July - October, 1998). Values represent proportion of fish that had at least one category item present in the stomach contents. 103 II Treed DOpen • Harvested 100 a) Rainbow Trout 90 ... 80 E 70 >c 60 Cll .t:l :I z .t:l 0 :;::; 50 0 40 E 0 30 ~ 0 20 ·u; Q. (.) 10 0 Zooplankton Aquatic Insects Molluscs 100 ...Cll .t:l E :I z >.t:l c ~ ·u; Other Fish Other 50 30 ~ 0 Fish 70 60 E 0 b) Kokanee 80 40 (.) Other 90 0 Q. Fish 20 10 0 Zooplankton 100 80 E 70 :I z >c .t:l ~ ·u; 60 50 0 40 E 30 Q. 0 (.) ';/!. Molluscs c) Northern Pikeminnow 90 ...Cll .t:l Aquatic Insects 20 10 0 Zooplankton Aquatic Insects Molluscs Food Item Figure A.12. Comparison of mean percent composition by number of food items for a) rainbow trout b) kokanee and c) northern pikeminnow in each habitat sampled in Ootsa Lake, BC (July- October, 1998). 104 II Treed ... 16 ::I 14 >.c c 12 Q) .c E 16 z II Harvested DOpen 10 ~ ·c;; 8 0 6 c. E 4 0 2 0 0 0 :::e '"' ,cf<'""' §'"' 'vq:s:-"' 18 ... b) "VI\0 11" " 16 Q) .c E 14 ::I z >.c c 12 :;:::; 8 0 c. 6 0 4 10 0 ·c;; E 0 :::e - 2 0 0 ... I ....__n ..r-1 .n 18 Q) .c 16 ::I 14 >.c c 12 ·c;; 8 E z 10 g 0 c. E 6 0 2 4 0 :::e 0 0 -t;"' cf<" ~ ~ "'(:' ,;? 0 b #" cfi ~ '"' .,._.,,'I> ,f:>q'""' oq ~ e"' Food Item Figure A.13. Comparison of mean percent composition by number of aquatic insects and terrestrial spiders in diet of a) rainbow trout b) kokanee and c) northern pikeminnow in each of the habitats sampled in Ootsa Lake, BC (July - October, 1998). 105 0> 0 T1 e T2 .. T3 --D-H1 --o-H2 -6-H3 - o- 01 - o- 02 - -6- 03 Figure A.14. Comparison of temperature profiles for each site in Ootsa Lake, BC (August 15, 1998) . T =treed habitat; H = harvested habitat; 0 = open habitat; numbers represent specific sites within each habitat. • 0 CXl 9 I c VI VI > 0 Q) "C 0 )( >- Cl Q) c: -+- H3 ·1 • 2 .,._ 3 ~ __. B<::::: 4 lk::::: 5 • 6 ::::;::::::: """"""= 7 ~ Depth (m) 8 9 10 ~ 11 12 13 ~ 14 15 ~ I Figure A.16. Comparison of oxygen profiles from the harvested sites of Ootsa Lake, BC (August 15, 1998). H3 was the most r cently harvested (- 2 months); H2 had been harvested for the longest time(- 1 year); and H1 was intermediate to the other two(- 4 months). 7.5 8 .§. 8.5 Cil ::::; ..... H2 9.5.----------------------------------------------------------------------------------------------, ....,... H1 0 CD c Cl) Q. ; g Treed Habitat Harvested Open Figure A.17. Comparison of mean Secchi disk readings in each habitat of Ootsa Lake, BC. Values were recorded on August 14, 1998 etween 11 :15 and 12:50 from the shaded side of the boat. Error bars represent one SE. 0 +-------' 2 +---------1 4 +---------1 6 -o- - - - - -- 8 +---------1 ~ ~ 0 0.000003 E '6 Q) U) :::s Ul c. Q) c: 'C Q) 'C U) 0 Surface Site • Bottom Figure A.18. Comparison of suspended sediment in each site of Ootsa Lake, BC (August 14, 1998). T =treed habitat; H = harvested habitat; 0 = open habitat; numbers represent individual sites. Where no bar is present, the value for that sample is zero. -0.000002 -0.000001 0.000000 0.000001 0.000002 0.000004 0 .§. 'E Q) ::::r 0.000005 0.000006 0.000007 I I Q) ... 50.0 tf!. Treed Habitat Harvested Open Figure A.19. Comparison of percent organic matter in the substrate of each habitat sampled in Ootsa Lake, BC (August 14, 1998). Values represent means triplicate samples from each of the three sites within each habitat (n=9 for each habitat). Error bars represent one SE. 0.0 10.0 20 .0 30.0 0Cl 40.0 ftl "2 (J :::!: ftl :t: 60.0 70.0 80.0 90.0 N Year 1997 1997 m/r 1998 1998 m/r n 112 95 298 239 1+ 6.3 0 2 0 2+ 88.4 93.7 88 .9 92.5 4.5 5.3 7.7 6.2 3+ 4+ 0.9 1.1 0.3 0.4 Table A-1. Percentage of kokanee in each age class in both sampling years from Ootsa Lake, BC. m/r denotes percentages of fish that were considered mature (substantial gonad development) or ripe (ready to spawn). m/r is a subset of the total number of fish sampled in each year. (..oJ 30 57 203 + 7.19 210+5.71 1998 n 1997 Mean+ SE # of Eggs (Age 2+) Year 2 5 234 + 32.3 n 240 + 38.0 #of Eggs (Age 3+) Mean +SE 212 + 5.81 205 + 7.13 # of Eggs (Age 2-3) Mean+ SE 62 32 n Table A-2. Mean fecundity of mature/ripe female kokanee in each sampling year from Ootsa Lake, BC. """ Year 1997 1998 1998 1998 Season Fall Summer Fall Overall n 105 51 218 269 %male 61.9* 41 .2 ns 66.1** 61.3** Table A-3. Percentage of kokanee caught in Ootsa Lake, BC that were male. * p<0.05; ** p<0.001; ns =not significant. Fish that could not be sexed (unknowns) were not included in the analysis . REFERENCES Abele,L.G . 1974. Species diversity of decapod crustaceans in marine habitats. Ecology 55: 156-161 . Aggus,L.R. and Elliott,G.V. 1975. Efforts of cover and food on year class strength of largemouth bass . In Black Bass Biology and Managment. Edited by H.Ciepper. Sport Fishing Institute, Washington, D.C . pp. 317-322 . Anderson,R.O. and Neumann,R.M. 1996. Length, weight, and associated structural indices. In Fisheries Techniques. Edited by B.R.Murphy and D.W.Willis. American Fisheries Society, Bethesda, Maryland pp . 447-482. Ashley, K. and Shepherd, B. 1996. Okanagan Lake workshop report and action plan. Prepared for the Ministry of Environment, Lands and Parks, Fisheries Branch , Province of British Columbia No. Fisheries Project Report No. RD 45. Ashley,K., Thompson,L.C., Lasenby,D.C., McEachern,L., Smokorowski,K.E., and Sebastian,D. 1997. Restoration of an interior lake ecosystem: the Kootenay Lake fertilization experiment. Water Quai.Res.J .Canada 32: 295-323. Barbour,C.D. and Brown,J .H. 1974. Fish species diversity in lakes . The American Naturalist 108: 473-489. Beamish,F.W.H. and Tandler,A. 1990. Ambient ammonia, diet and growth in lake trout. Aquatic Toxicology 17: 155-166. Benson,B.J. and Magnuson,J.J. 1992. Spatial heterogeneity of littoral fish assemblages in lakes: relation to species diversity and habitat structure. Canadian Journal of Fisheries and Aquatic Science 49: 1493-1500. Boreman,J . 1981 . Characteristics of fall and spring migrant rainbow trout in Cayuga Inlet, New York. New York Fish and Game Journal28: 100-107. Bowen,S .H. 1996. Quantitative Description of the Diet. In Fisheries Techniques. Edited by B.R.Murphy and D.W.Willis. American Fisheries Society, Bethesda, Maryland pp. 513-532 . 115 Brandt,S.B. , Magnuson ,J.J., and Crowder,L.B . 1980. Thermal habitat partitioning by fishes in Lake Michigan. Canadian Journal of Fisheries and Aquatic Science 37: 1557-1564. Brown ,L.R. and Moyle,P.B. 1981 . The impact of squawfish on salmonid populations: a review. North American Journal of Fisheries Management 1: 104-111 . Bronmark,C. 1985. Freshwater snail diversity: effects of pond area, habitat heterogeneity and isolation. Oecologia 67: 127-131 . Cannings,S ., Cannings,R., and Cannings,R. 1998. British Columbia: A natural history: The world of fresh water. Greystone Books, Vancouver. Carlander,K.D . 1969. Handbook of freshwater fishery biology: life history data on freshwater fishes of the United States and Canada, exclusive of the Perciformes. The Iowa State University Press, Ames, Iowa. Conover,D.O., Voorhees,D.A., and Ehtisham,A. 1992. Sex ratio selection and the evolution of environmental sex determination in laboratory populations of Menidia menidia. Evolution 46: 1722-1730. Conover,W.J. 1980. Practical Nonparametric Statistics. John Wiley and Sons, New York. Craig,J.K., Foote,C.J ., and Wood ,C.C. 1996. Evidence for temperature-dependent sex determination in sockeye salmon (Oncorhynchus nerka). Canadian Journal of Fisheries and Aquatic Science 53: 141-147. Crim,L.W. and Glebe,B.D . 1990. Reproduction . In Methods for Fish Biology. Edited by C.B.Schreck and P.B.Moyle. American Fisheries Society, Bethesda, Maryland pp. 529-553. Culp,J .M., Scrimgeour,G .J., and Townsend,G.D. 1996. Simulated fine woody debris accumulations in a stream increase rainbow trout fry abundance. Transactions of the American Fisheries Society 125: 472-479 . Cummins,K.W. and Merritt,R.W. 1996. Ecology and distribution of aquatic insects. In An Introduction to the Aquatic Insects of North America. Edited by R.W.Merritt and K.W .Cummins. Kendall I Hunt Publishing Company, Dubuqu, Iowa pp. 74-86. 116 Davis,J.T. and Hughes,J.S. 1971 . Effects of standing timber on fish populations and fisherman success in Bussey Lake, Louisiana . Edited by G.E.Hall. American Fisheries Society, Washington, D.C. pp. 255-264. Eadie,J.M. and Keast,A. 1984. Resource heterogeneity and fish species diversity in lakes. Canadian Journal of Zoology 62: 1689-1695. Ellis,M .M. 1941 . Fresh-water impoundments. Transactions of the American Fisheries Society 71 : 80-93. Ensign,W.E. and Strange,R.J . 1990. Summer food limitation reduces brook and rainbow trout biomass in a southern Appalachian stream . Transactions of the American Fisheries Society 119: 894-901. Filbert,R.B . and Hawkins,C.P. 1995. Variation in condition of rainbow trout in relation to food, temperature, and individual length in the Green River, Utah. Transactions of the American Fisheries Society 124: 824-835. Gelwick,F.P. and Matthews,W.J. 1990. Temporal and spatial patterns in littoral-zone fish assemblages of a reservoir (Lake Texoma, Oklahoma-Texas, USA). Environmental Biology of Fishes 27: 107-120. Greenberg,A.E. , Clesceri,L.S., and Eaton ,A.D. 1992. Standard Methods for the Examination of Water and Wastewater. American Public Health Association; American Water Works Association; Water Environment Federation, Washington . Hamley,J.M. 1975. Review of gillnet selectivity. Journal of the Fisheries Research Board of Canada 32 : 1943-1969. Helfman,G.S. 1979. Fish attraction to floating object in lakes. In Response of Fish to Habitat Structure in Standing Water. Edited by D.L.Johnson and R.A.Stein . American Fisheries Society Special Publication 6, North Central Division pp. 49-57. Hubert,W.A. 1996. Passive capture techniques . In Fisheries Techniques. Edited by B.R.Murphy and D.W .Willis. American Fisheries Society, Bethesda, Maryland pp. 193-220. International Pacific Salmon Fisheries Commission 1953. A review of the sockeye salmon problems created by the Alcan project in the Nechako River watershed . 117 Laufle,J .C. and Cassidy,R.A. 1988. Retention of standing timber for fish and wildlife habitat in reservoirs. Lake and Reservoir Management 4: 271-279. Lindstrom T. 1973. Life in a lake reservoir: fewer options, decreased production. Ambia 2: 145-153. Lloyd, R. A. 1999. Foraging Locations and Habits of Ospreys in the Nechako Reservoir. Prepared for BC Environment, Wildlife Branch , Smithers, BC . Lorz,H.W. and Northcote,T.G. 1965. Factors affecting stream location, and timing and intensity of entry by spawning kokanee (Oncorhynchus nerka) into an inlet of Nicola Lake, British Columbia. Journal of the Fisheries Research Board of Canada 22 : 665-687 . Lyons, J. C. and Larkin, P. A. 1952. The effects on sport fisheries of the Aluminum Company of Canada Limited development in the Nechako drainage. Prepared for the Ministry of Environment, Lands and Parks Fisheries Branch , Province of British Columbia No. Fisheries Management Report No. 10 . . MacArthur,R.H. and MacArthur,J.W. 1961 . On bird species diversity. Ecology 42: 594-598. Mackay,W.C., Ash,G .R., and Norris,H.J. 1990. Fish ageing methods for Alberta. R.L.&L. Environmental Services Ltd . I Alberta Fish and Wildlife Division I University of Alberta, Edmonton . Marshaii,T.R. and Ryan,P.A. 1987. Abundance patterns and community attributes of fishes relative to environmental gradients . Canadian Journal of Fisheries and Aquatic Science 44: 198-215. Martinelli,T.L. and Shively,R.S . 1997. Seasonal distribution , movements and habitat associations of northern squawfish in two lower Columbia river reservoirs. Regulated Rivers : Research and Management 13: 543-556 . McLachlan,A.J. 1970. Submerged trees as a substrate for benthic fauna in the recently created Lake Kariba (central Africa). Journal of Applied Ecology 7: 253-266. Meffe,G.K. and Carroii,C.R. 1994. Principles of Conservation Biology. Sinauer Associates, Inc., Sunderland, Massachusetts. 118 Moring ,J.R. , Negus,M.T., McCullough,R.D. , and Herke,S.W. 1989. Large concentrations of submerged pulpwood logs as fish attraction structures in a reservoir. Bulletin of Marine Science 44: 609-615. Moyle,P.B. and Cech,Jr.J.J. 2000. Fishes: An Introduction to Ichthyology. 4th Edition . Prentic Hall, Upper Saddle River, NJ. Murray,C.B., McPhaii,J.D., and Rosenau ,M.L. 1989. Reproductive and developmental biology of kokanee from Upper Arrow Lake, British Columbia. Transactions of the American Fisheries Society 118: 503-509 . Ney,J .J. 1996. Oligotrophication and its discontents: effects of reduced nutrient loading on reservoir fisheries . American Fisheries Society Symposium 16: 285-295. Northcote, T. G and Atagi, D. Y. 1997. Ecological interactions in the flooded littoral zone of reservoirs : the importance and role of submerged terrestrial vegetation with special reference to fish, fish habitat and fisheries in the Nechako reservoir of British Columbia, Canada. BC Ministry of Environment, Lands, and Parks No. SK-111 . Paine,R.T. 1969. A note on trophic complexity and community stability. American Naturalist 103: 91-93. Pardue,G.B. and Nielsen,L.A. 1979. Invertebrate biomass and fish production in ponds with added attachment surface. In Response of Fish to Habitat Structure in Standing Water. Edited by D.L.Johnson and R.A.Stein. American Fisheries Society Special Publication Number 6, North Central Division pp. 34-37. Paxton ,K.O. and Stevenson ,F. 1979. Influence of artificial structure on angler harvest from Killdeer Reservoir, Ohio. In Response of Fish to Habitat Structure in Standing Water. Edited by D.L.Johnson and R.A.Stein. American Fisheries Society Special Publication Number 6, North Central Division pp. 70-76. Perrin , C. J. 1992. Pilot fertilization of the Nechako River IV: monitoring to improve precision . Limnotek Research and Development Inc. for Nechako Fisheries Conservation Program cited in Perrin et al 1997. Perrin, C. J., McDevitt, C. A. , Macisaac, E. A. , and Kashino, R. 1997. Water quality impact assessment for Nechako Reservoir submerged timber salvage operations: baseline water quality. Ministry of Environment, Lands and Parks, Environmental Protection Skeena Region , Smithers, BC . 119 Pianka,E.R. 1967. On lizard species diversity: North American flatland deserts. Ecology 48: 333-351. Pielou,E.C . 1975. Ecological Diversity. John Wiley & Sons, New York. Pieters, R., Thompson, L. C., Vidmanic, L., Pond, S., Stockner, J . G., Hamblin, P., Young, M., Ashley, K., Lindsay, B., Lawrence, G., Sebastian, D, Scholten, G ., and Lombard, D. L. 1998. Arrow Reservoir limnology and troph ic status - Year 1 ( 1997/98) report. Prepared for Ministry of Environment, Lands and Parks, Province of British Columbia No. Fisheries Project Report No. RD 67. Poddubnyy,A.G . and Fortunatov,H.A. 1961. Problem of reservoir use by fisheries in various geographical zones. Voprosy lkhtiologii (Problems of Ichthyology) Moscow 1: 599-611. Cited in Northcote and Atagi (1997). Prince,E.D . and Maughan,O.E. 1979. Attraction of fishes to tire reefs in Smith Mountain Lake, Virginia. In Response of Fish to Habitat Structure in Standing Water. Edited by D.L.Johnson and R.A.Stein. American Fisheries Soceity Special Publication Number 6, North Central Division pp. 19-25. Quinn ,T .P. and Peterson,N.P. 1996. The influence of habitat complexity and fish size on over-winter survival and growth of individually marked juvenile coho salmon (Oncorhynchus kisutch) in Big Beef Creek, Washington. Canadian Journal of Fisheries and Aquatic Science 53: 1555-1564. Reeves,W.C., Hooper,G .R., and Smith,B.W . 1977. Preliminary observations of fish attraction to artificial mid-water structures in fresh water. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 31 : 471-476. Ricker,W .E. 1938. "Residual" and kokanee salmon in Cultus lake. Journal of the Fisheries Research Board of Canada 4: 192-218. Ricker,W.E. 1941. The consumption of young sockeye salmon by predaceous fish . Journal of the Fisheries Research Board of Canada 5: 293-313 . Ricker,W.E. 1959. Additional observations concerning residual sockeye salmon and koknaee (Oncorhynchus nerka). Journal of the Fisheries Research Board of Canada 16: 897-902 . 120 Roff,D.A. 1992. The Evolution of Life Histories: Theory and Analysis. Routledge, Chapman and Hall, Inc., New York. Rosenzweig,M.L. and Winakur,J. 1969. Population ecology of desert rodent communities : habitats and environmental complexity. Ecology 50: 558-572. Scott,W.B. and Crossman,E.J . 1998. Freshwater Fishes of Canada. Galt House Publications Ltd., Oakville, Ontario, Canada. Shirley,K.E. and Andrews,A.K. 1977. Growth, production, and mortality of largemouth bass during the first year of life in Lake Carl Blackwell, Oklahoma. Transactions of the American Fisheries Society 106: 590-595. Shirveii,C.S. 1990. Role of instream rootwads as juvenile coho salmon (Oncorhynchus kisutch) and steelhead trout (0. mykiss) cover habitat under varying streamflows. Canadian Journal of Fisheries and Aquatic Science 47: 852-861. Sokai,R.R. and Rohlf,F.J. 1995. Biometry. W.H. Freeman and Company, New York. Stang,D.L. and Hubert,W.A. 1984. Spatial separation of fishes captured in passive gear in a turbid prairie lake. Environmental Biology of Fishes 11:309-314. Stearns,S .C. and Crandaii,R.E. 1984. Plasticity for age and size at sexual maturity: a life-history response to unavoidable stress . In Fish Reproduction . Edited by G.W.Potts and R.J.Wootton. Academic Press, London pp. 13-33. Stearns,S.C. 1992. The Evolution of Life Histories. Oxford University Press, New York. Stockner,J.G. 1987. Lake fertilization : the enrichment cycle and lake sockeye salmon (Oncorhynchus nerka) production. In Sockeye salmon (Oncorhynchus nerka) population biology and future management. Edited by H.D.Smith, L.Margolis, and C.C.Wood . Canadian Special Publication of Fisheries and Aquatic Sciences 96., pp. 198-215. Stockner,J.G., Rydin,E., and Hyenstrand,P. 2000. Cultural oligotrophication: causes and consequences for fisheries resources . Fisheries 25: 7-14. 121 Summerfelt,R.C . 1993. Lake and reservoir habitat management. In Inland Fisheries Management in North America. Edited by C.C.Kohler and W.A.Hubert. American Fisheries Society, Bethesda pp. 231-261 . Tabor,R.A. and Wurtsbaugh,W.A. 1991. Predation risk and the importance of cover for juvenile rainbow trout in lentic systems. Transactions of the American Fisheries Society 120: 728-738. Tonn,W.M. and Magnuson,J.J. 1982. Patterns in the species composition and richness of fish assemblages in northern Wisconsin lakes . Ecology 63: 1149-1166. VanDenAvyle,M.J. and Petering,R.W. 1988. Inundated timber as nursery habitat for larval gizzard and threadfin shad in a new pumped storage reservoir. Transactions of the American Fisheries Society 117: 84-89. Wege,G .J. and Anderson,R.O. 1979. Influence of artificial structure on largemouth bass and bluegills in small ponds. In Response of Fisht to Habitat Structure in Standing Water. American Fisheries Society Special Publication Number 6, North Central Division pp . 59-69. Werner,E.E., Haii,D.J., Laughling,D .R., Wagner,D .J., Wilsmann,L.A., and Funk,F.C. 1977. Habitat partitioning in a freshwater fish community. Journal of the Fisheries Research Board of Canada 34: 360-370 . Wetzei,R.G . 1983. Limnology. 2nd Edition. Saunders College Publishing, Orlando. Wickam,D.A., Watson,Jr.J .W., and Ogren,L.H. 1973. The efficacy of midwater artificial structures for attracting pelagic sport fish. Transactions of the American Fisheries Society 102: 563-572 . Wilcove,D .S., Mclellan,C.H., and Dobson,A.P . 1986. Habitat fragmentation in the temperate zone. In Conservation Biology: The Science of Scarcity and Diversity. Edited by M.E.Soule. Sinauer Associates, Inc., Sunderland, Massachusetts pp . 237-256. Willis,D.W. and Jones,L.D. 1986. Fish standing crops in wooded and nonwooded coves of Kansas reservoirs. North American Journal of Fisheries Management 6: 105-108. Winsby, M. B., Taylor, G. C., Young, L. U., Munday, D. R., and Stables, B. 1997. Nechako Reservoir: Impact of timber salvage on fish and fish habitat - 1996 studies. Prepared by 122 Hatfield Consultants Ltd. for the Ministry of Environment, Lands and Parks, Skeena Region . Winsby, M. B., Taylor, G. C., and Munday, D. R. 1998. Nechako Reservoir: Impacts of timber salvage on fish and fish habitat- 1997 studies. Prepared by Hatfield Consultants Ltd . for the Ministry of Environment, Lands and Parks, Skeena Region, BC. Wood,C .C. and Foote,C .J. 1996. Evidence for sympatric genetic divergence of anadromous and nonanadromous morphs of sockeye salmon (Oncorhynchus nerka). Evolution 50: 1265-1279. Wootton,J.T. 1984. A Functional Biology of Sticklebacks. Crrom Helm Ltd. , London . Wurtsbaugh,W.A. and Cech,Jr.J.J. 1983. Growth and activity of juvenile mosquitofish: temperature and ration effects. Transactions of the American Fisheries Society 112: 653-660. Zar,J .H. 1984. Biostatistical Analysis. Prentic Hall, Englewood Cliffs, New Jersey. 123