{"http:\/\/dx.doi.org\/10.14288\/1.0087773":{"http:\/\/vivoweb.org\/ontology\/core#departmentOrSchool":[{"value":"Science, Faculty of","type":"literal","lang":"en"},{"value":"Zoology, Department of","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/dataProvider":[{"value":"DSpace","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeCampus":[{"value":"UBCV","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/creator":[{"value":"Ghan, David","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/issued":[{"value":"2009-03-30T20:54:02Z","type":"literal","lang":"en"},{"value":"1997","type":"literal","lang":"en"}],"http:\/\/vivoweb.org\/ontology\/core#relatedDegree":[{"value":"Doctor of Philosophy - PhD","type":"literal","lang":"en"}],"https:\/\/open.library.ubc.ca\/terms#degreeGrantor":[{"value":"University of British Columbia","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/description":[{"value":"I have studied the adaptive significance of vertical\r\nmigration of zooplankton in 2 populations of the copepod\r\nSkistodiaptomus oregonensis that migrate and 2 populations that\r\ndo not migrate. Vertical migration of the copepods is associated\r\nwith the presence of pelagic sticklebacks. This observation is\r\nconsistent with the hypothesis that the adaptive benefit of\r\nvertical migration by S. oregonensis is to avoid stickleback\r\npredators. The alternative hypotheses including avoidance of\r\njuvenile sockeye predators, foraging efficiency, bioenergetic\r\nefficiency, or combined foraging\/bioenergetic efficiency are not\r\nsupported by the comparisons of migratory behaviour of S.\r\noregonensis in the 4 lakes.\r\nBoth the depth and timing of S. oregonensis migration are\r\nconsistent with the hypothesis that copepods are avoiding\r\npredation by sticklebacks. Light intensities at the depth at\r\nwhich S. oregonensis reside during the day are sufficiently low\r\nto reduce predation risk from visual foraging stickleback and the\r\ntiming of ascent at dusk and descent at dawn are such that S.\r\noregonensis remain at light intensities that reduce risk from\r\nstickleback. S. oregonensis are at the surface at dusk and dawn\r\nat the time that juvenile sockeye feed in the surface habitat.\r\nVertical migration appears to be a trade-off with resource\r\nacquisition. Phytoplankton are less concentrated in the deep\r\nhabitat where S. oregonensis reside during the day. Furthermore,\r\nmigrating copepods contain less phytoplankton as food in their\r\nguts than do non-migrating individuals.\r\nIn vertical columns in the laboratory, the presence or\r\nabsence of sticklebacks does not influence the vertical\r\ndistributions of S. oregonensis collected from lakes with either\r\nmigratory or non-migratory populations. This indicates that the\r\nmigration phenotype is fixed rather than being a flexible\r\nbehaviour induced by environmental cues. I developed a dynamic\r\noptimization model to predict the optimal depth decisions for S.\r\noregonensis based on depth dependent lake food and temperature\r\nconditions, fish abundance and predation rates, and S.\r\noregonensis bioenergetics. The model predicts that vertical\r\nmigration should occur to avoid sticklebacks under a broad range\r\nof modelled conditions, but with a fitness cost due to feeding\r\nopportunity costs. This demonstrates quantitatively that it is\r\ntenable to hypothesize that vertical migration involves a tradeoff\r\nbetween stickleback avoidance and feeding opportunity.\r\nTaken together, these results are consistent with the view\r\nthat different migration behaviours in these populations are a\r\nresult of divergent evolution driven by environmental variation\r\naffecting the optimal solution to the predation risk\/resource\r\nacquisition trade-off.","type":"literal","lang":"en"}],"http:\/\/www.europeana.eu\/schemas\/edm\/aggregatedCHO":[{"value":"https:\/\/circle.library.ubc.ca\/rest\/handle\/2429\/6631?expand=metadata","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/terms\/extent":[{"value":"6540981 bytes","type":"literal","lang":"en"}],"http:\/\/purl.org\/dc\/elements\/1.1\/format":[{"value":"application\/pdf","type":"literal","lang":"en"}],"http:\/\/www.w3.org\/2009\/08\/skos-reference\/skos.html#note":[{"value":"ADAPTIVE SIGNIFICANCE OF VERTICAL MIGRATION BEHAVIOUR OF SKISTODIAPTOMUS OREGONENSIS by DAVID GHAN B.Sc, M c G i l l U n i v e r s i t y , 1987 M.Sc, The U n i v e r s i t y of Toronto, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY i n THE FACULTY OF GRADUATE STUDIES (Department of Zoology) We accept t h i s t h e s i s as conforming to the r e q u i r e d standard THE UNIVERSITY OF BRITISH COLUMBIA March 1997 David Ghan, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. 1 further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of piD J\\ Uy \u2014 zru~ The University of British Columbia Vancouver, Canada Date lm\/7srA .7 I'\/fl\"? DE-6 (2\/88) 1 1 Abstract I have studied the adaptive significance of v e r t i c a l migration of zooplankton i n 2 populations of the copepod Skistodiaptomus oregonensis that migrate and 2 populations that do not migrate. V e r t i c a l migration of the copepods i s associated with the presence of pelagic sticklebacks. This observation i s consistent with the hypothesis that the adaptive benefit of v e r t i c a l migration by S. oregonensis i s to avoid stickleback predators. The alte r n a t i v e hypotheses including avoidance of juvenile sockeye predators, foraging e f f i c i e n c y , bioenergetic e f f i c i e n c y , or combined foraging\/bioenergetic e f f i c i e n c y are not supported by the comparisons of migratory behaviour of S. oregonensis i n the 4 lakes. Both the depth and timing of S. oregonensis migration are consistent with the hypothesis that copepods are avoiding predation by sticklebacks. Light i n t e n s i t i e s at the depth at which S. oregonensis reside during the day are s u f f i c i e n t l y low to.reduce predation r i s k from v i s u a l foraging stickleback and the timing of ascent at dusk and descent at dawn are such that S. oregonensis remain at l i g h t i n t e n s i t i e s that reduce r i s k from stickleback. S. oregonensis are at the surface at dusk and dawn at the time that juvenile sockeye feed i n the surface habitat. V e r t i c a l migration appears to be a trade-off with resource acquisition. Phytoplankton are less concentrated i n the deep habitat where S. oregonensis reside during the day. Furthermore, migrating copepods contain less phytoplankton as food i n the i r guts than do non-migrating individuals. I l l In v e r t i c a l columns i n the laboratory, the presence or absence of sticklebacks does not influence the v e r t i c a l d i s t r i b u t i o n s of S. oregonensis c o l l e c t e d from lakes with either migratory or non-migratory populations. This indicates that the migration phenotype i s fixed rather than being a f l e x i b l e behaviour induced by environmental cues. I developed a dynamic optimization model to predict the optimal depth decisions for S. oregonensis based on depth dependent lake food and temperature conditions, f i s h abundance and predation rates, and S. oregonensis bioenergetics. The model predicts that v e r t i c a l migration should occur to avoid sticklebacks under a broad range of modelled conditions, but with a fitness cost due to feeding opportunity costs. This demonstrates q u a n t i t a t i v e l y that i t i s tenable to hypothesize that v e r t i c a l migration involves a trade-off between stickleback avoidance and feeding opportunity. Taken together, these r e s u l t s are consistent with the view that d i f f e r e n t migration behaviours i n these populations are a r e s u l t of divergent evolution driven by environmental v a r i a t i o n a f f e c t i n g the optimal solution to the predation risk\/resource a c q u i s i t i o n trade-off. i v Table of Contents A b s t r a c t i i Table of Contents i v L i s t of Tables v L i s t of Figures v i Acknowledgements x i i Chapter 1: General I n t r o d u c t i o n 1 Chapter 2: Testing hypotheses on the b e n e f i t s and costs of v e r t i c a l m i g r a t i o n by Skistodiaptomus oregonensis by comparing populations 9 I n t r o d u c t i o n . 9 Methods 10 Results 19 Di s c u s s i o n 48 Chapter 3: The Timing and extent of v e r t i c a l m i g r a t i o n by Skistodiaptomus oregonensis r e l a t i v e to the temporal-s p a t i a l d i s t r i b u t i o n of p r e d a t i o n r i s k 58 I n t r o d u c t i o n 58 Methods 59 Results 63 Dis c u s s i o n . 98 Chapter 4: V e r t i c a l m i g r a t i o n behavior of Skistodiaptomus oregonensis: c o n s t i t u t i v e or induced? . 103 In t r o d u c t i o n 103 Methods 104 Results 108 Di s c u s s i o n 120 Chapter 5: A dynamic o p t i m i z a t i o n model of Skistodiaptomus oregonensis v e r t i c a l m i g r a t i o n 128 In t r o d u c t i o n 128 Model D e s c r i p t i o n 129 Model Results 140 Summary 151 Chapter 6: Thesis Summary 153 References 158 V L i s t of Tables Table 2.1: P h y s i c a l c o n d i t i o n s i n the four study la k e s . . . . 37 Table 2.2: B i o l o g i c a l c o n d i t i o n s i n the four study la k e s . . . 3 8 Table 3.1: ANOVA r e s u l t s t e s t i n g H0 that i n i t i a l zooplankton d e n s i t i e s i n experimental feeding tanks d i d not d i f f e r among l i g h t i n t e n s i t y treatments 64 Table 3.2: ANOVA r e s u l t s f o r each prey type t e s t i n g Ho that i n i t i a l d e n s i t i e s of the prey types d i d not d i f f e r among treatments 65 Table 3.3: ANOVA r e s u l t s t e s t i n g Ho that i n i t i a l zooplankton d e n s i t i e s i n experimental feeding tanks d i d not d i f f e r among days 69 Table 3.4: Regression r e s u l t s f o r the e f f e c t of l i g h t (X) on four Y v a r i a b l e s where In(Y+l)=a*ln(X)+b, except f o r the c a p t u r e \/ s t r i k e r a t i o where a r c s i n ( s q r t (Y+l) ) =a*ln(X) +b 78 Table 4.1: Experiment 1 - u n i v a r i a t e repeated measures analyses of variance f o r A) the e f f e c t of fishwater [+\/-] treatments on the p r o p o r t i o n of S. oregonensis below 100 cm during the day and B) the e f f e c t of fish w a t e r [+\/-] treatments on the change i n the pr o p o r t i o n of S. oregonensis below 100 cm from night to day 112 Table 4.2: Experiment 1 - nested analyses of variance f o r depths of i n d i v i d u a l S. oregonensis: t e s t s of s i g n i f i c a n c e f o r fi s h w a t e r [+\/-] treatment e f f e c t s and tube w i t h i n treatment e f f e c t s 114 Table 4.3: Experiment 2 - u n i v a r i a t e repeated measures analyses of variance f o r A) the e f f e c t of fishw a t e r [+\/-] treatments on the p r o p o r t i o n of S. oregonensis below 100 cm during the day and B) the e f f e c t of fishwa t e r [+\/-] treatments on the change i n the pr o p o r t i o n of S. oregonensis below 100 cm from night to day 118 Table 4.4: Experiment 2 - nested analyses of variance f o r depths of i n d i v i d u a l S. oregonensis: t e s t s of s i g n i f i c a n c e f o r fishw a t e r (+\/-) treatment e f f e c t s and tube w i t h i n treatment e f f e c t s 119 Table 4.5: Experiment 3 - nested analyses of variance f o r depths of i n d i v i d u a l S. oregonensis: t e s t s of s i g n i f i c a n c e f o r f i s h (+\/-) treatment e f f e c t s and tube w i t h i n treatment e f f e c t s 123 v i L i s t of Figures Fig. 2.1: Location of the four study lakes i n southwestern B r i t i s h Columbia 11 Fig. 2.2: Map of Great Central Lake showing sample location (X) and lake contours i n meters. Depth contours taken from Rutherford et al. (1986) 12 Fig. 2.3: Map of Hobiton Lake showing location of sample s i t e (X) and lake contours i n meters 13 Fig. 2.4: Map of Kennedy Lake showing location of sample s i t e (X) and depth contours i n meters 14 Fig. 2.5: Map of Paxton Lake showing location of sample s i t e (X) and depth contours i n meters 15 Fig. 2.6: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Kennedy Lake on August 25, 1993 20 Fig. 2.7: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Kennedy Lake on November 12, 1992. . . . 21 Fig. 2.8: A) Mean day and night depth of 750-850 um metasome length S. oregonensis i n Kennedy Lake on 13 sample dates from 1992-1994. B) Mean day and night depth of 3-50 um size f r a c t i o n of phytoplankton i n Kennedy Lake from 13 sample dates from 1992 to 1994. C) Data i n part A and B redrawn to compare mean day and night depth of S. oregonensis with mean day and night depth of 3-50 um size f r a c t i o n of phytoplankton 22 Fig. 2.9: V e r t i c a l d i s t r i b u t i o n of chlorophyll a from the 3-50 um size f r a c t i o n of phytoplankton on 12 sample dates i n Kennedy Lake 23 Fig. 2.10: Depth d i s t r i b u t i o n of temperature (open c i r c l e s ) and of chlorophyll a from the 3-50 um size f r a c t i o n of phytoplankton day (open squares) and night ( s o l i d squares) on s i x sample dates 25 Fig. 2.11: Changes i n density of S. oregonensis i n three depth strata of Paxton Lake during dawn of August 19, 1994. . 26 Fig. 2.12: Changes i n density of S. oregonensis i n three depth strata of Paxton Lake during dusk of August 19, 1994 27 Fig. 2.13: Day and night v e r t i c a l d i s t r i b u t i o n of chlorophyll a from the 3-50 um size f r a c t i o n of phytoplankton i n Paxton Lake, August, 1994 28 v i i Fig. 2.14: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Great Central Lake on July 7, 1992. . 30 Fig. 2.15: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Great Central Lake on August 18, 1992. . \u2022 31 Fig. 2.16: A) Mean day and night depth of 750-850 um metasome length S. oregonensis i n Great Central Lake on 7 sample dates i n 1992 and 1993. B) Mean day and night depth of 3-50 um size f r a c t i o n of phytoplankton i n Great Central Lake from 7 sample dates from 1992 and 1993. C) Data i n part A and B redrawn to compare mean day and night depth of S. oregonensis with mean day and night depth of 3-50 um phytoplankton 32 Fig. 2.17: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Hobiton Lake on August 5, 1992. . . . 33 Fig. 2.18: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 metasome length classes of S. oregonensis i n Hobiton Lake on June 25, 1992 34 Fig. 2.19: Day (open c i r c l e s ) and night ( s o l i d squares) mean depth of 660-795 um metasome length class S. oregonensis and the 3-50 um size f r a c t i o n of phytoplankton on June 25 and August 25, 1992 35 Fig. 2.20: Mean concentration of chlorophyll a from the 3-50 um size f r a c t i o n of phytoplankton i n the 4 study lakes 39 Fig. 2.21: Top: S. oregonensis weighted mean daytime depth versus metasome length classes i n Kennedy Lake. Bottom: Difference between weighted mean daytime depth and weighted mean night time depth for each sample date 41 Fig. 2.22: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 small metasome length classes of S. oregonensis i n Kennedy Lake on A p r i l 14, 1993 42 Fig. 2.23: Day (hatched bars) and night ( s o l i d bars) v e r t i c a l d i s t r i b u t i o n for 5 small metasome length classes of S. oregonensis i n Kennedy Lake on November 12, 1992 43 Fig. 2.24: Day versus night gut pigments i n S. oregonensis i n Kennedy.Lake ( s o l i d square) and Great Central Lake (open squares) on May 11-13, 1993 (top), June 22-23, 1993 (middle), and July 15-16, 1993 (bottom) 44 v i i i Fig. 2.25: Daytime S. oregonensis gut pigments versus depth on June 23, 1993 (top) and July 16, 1993 (bottom). . . . 45 Fig. 2.26: S. oregonensis gut pigment changes at dawn and dusk i n Great Central Lake and Kennedy Lake, August 25-27, 1993 46 Fig. 2.27: S. oregonensis gut pigment changes at dawn and dusk i n Great Central Lake and Kennedy Lake, October 17-20, 1993. . 47 Fig. 2.28: Weighted mean daytime depth of S. oregonensis i n Kennedy Lake versus mean chlorophyll a concentration of the 3-50 um phytoplankton size f r a c t i o n i n the top 10 m 49 Fig. 3.1: Total zooplankton densities broken down by taxa or size categories i n each of 11 treatments for experiment date June 5 67 Fig. 3.2: Total zooplankton densities broken down by taxa or size categories i n each of 11 treatments for experiment date June 7 68 Fig. 3.3: Mean zooplankton density across treatments for each experiment date. V e r t i c a l bars represent entire range of densities for each day 70 Fig. 3.4: A comparison of the r e l a t i v e density of prey types i n one of the experimental tanks ( s o l i d bar) and i n Kennedy Lake (open bar) on May 20, 1995 71 Fig. 3.5: A comparison of the size of prey types i n one of the experimental tanks ( s o l i d square) and i n Kennedy Lake (open squares) on May 20, 1995 72 Fig. 3.6: Number of S. oregonensis i n stomachs of stickleback fed for 15 minutes at 11 l i g h t i n t e n s i t i e s . Top: Light i n t e n s i t y range 0.1-15.5 uE \u2022 s~x-m~2. Bottom: Enlargement of l i g h t i n t e n s i t y range from 0.1-1.6 uE-s~ 1-m-2 (bottom l e f t corner of top figure) 73 Fig. 3.7: Number of prey i n stomachs of stickleback fed for 15 minutes at 11 l i g h t i n t e n s i t i e s . Top: Light . i n t e n s i t y range 0.1-15.5 uE-s^-m\"2. Bottom: Enlargement of l i g h t i n t e n s i t y range from 0.1-1.6 uE-s~ 1-m\"2 (bottom l e f t corner of top figure) 74 Fig. 3.8: Number of prey s t r i k e s taken by stickleback fed for 15 minutes at 11 l i g h t i n t e n s i t i e s . Top: Light i n t e n s i t y range 0.1-15.5 uE \u2022 s\"1 -m\"2. Bottom: Enlargement of l i g h t i n t e n s i t y range from 0.1-1.6 uE-s\" 1'm\"2 (bottom l e f t corner of top figure) 75 i x Fig. 3.9: Ratio of captures\/strikes for stickleback fed for 15 minutes at 11 l i g h t i n t e n s i t i e s . Top: Light i n t e n s i t y range 0.1-15.5 uE-s _ 1 - r r f 2 . Bottom: Enlargement of l i g h t i n t e n s i t y range from 0.1-1.6 uE\u00abs~ 1-m\"2 (bottom l e f t corner of top figure) 76 Fig. 3.10: Vanderploeg-Scavia index of e l e c t i v i t y for S. oregonensis by stickleback fed for 15 minutes at 11 l i g h t i n t e n s i t i e s . Top: Light i n t e n s i t y range 0.1-15.5 uE-s^-nT2. Bottom: Enlargement of l i g h t i n t e n s i t y range from 0.1-1.6 uE-s^-m\"2 (bottom l e f t corner of top figure) 79 Fig. 3.11: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Kennedy Lake during dawn on May 20, 1995 80 Fig. 3.12: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Kennedy Lake during dusk on May 20, 1995 81 Fig. 3.13: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Kennedy Lake during dawn on June 24, 1994 82 Fig. 3.14: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Kennedy Lake during dusk on June 24, 1994 83 Fig. 3.15: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Paxton Lake during dusk on August 19, 1994 84 Fig. 3.16: V e r t i c a l migration of S. oregonensis r e l a t i v e to stickleback feeding rate thresholds i n Paxton Lake during dawn on August 19, . 1994. . 85 Fig. 3.17: Summary of available data on time of the beginning of movement of juvenile sockeye toward the surface (X), the period of migration toward the surface (dotted lines) and the period of feeding near the surface ( s o l i d line) r e l a t i v e to sunset 89 Fig. 3.18: Summary of available data for juvenile sockeye on the time of juvenile sockeye feeding near the surface ( s o l i d line) and the time of the beginning of movement toward depth (X) r e l a t i v e to sunrise 90 Fig. 3.19: Maximum (1.0 f t . - c ) , half maximum (0.01 f t . - c ) , and minimum (0.0001 ft. - c . ) feeding rate isolumes for juvenile sockeye (from A l i 1959) for clear sky conditions during dawn (top panel) and dusk (bottom panel) on June 23, 1994 91 X Fig. 3.20: Changes i n v e r t i c a l d i s t r i b u t i o n of f i s h targets (top) and S. oregonensis (bottom) i n Kennedy Lake at dusk on October 19, 1993 93 Fig. 3.21: Changes i n v e r t i c a l d i s t r i b u t i o n of f i s h targets (top) and S. oregonensis (bottom) i n Kennedy Lake at dusk on June 24, 1994 94 Fig. 3.22: Changes i n v e r t i c a l d i s t r i b u t i o n of f i s h targets (top) and S. oregonensis (bottom) i n Kennedy Lake at dusk on May 19, 1995 95 Fig. 3.23: Changes i n v e r t i c a l d i s t r i b u t i o n of f i s h targets (top) and S. oregonensis (bottom) i n Kennedy Lake at dawn on June 24, 1994 96 Fig. 3.24: Changes i n v e r t i c a l d i s t r i b u t i o n of f i s h targets (top) and S. oregonensis (bottom) i n Kennedy Lake at dawn on May 19, 1995 97 Fig. 3.25: Changes i n v e r t i c a l d i s t r i b u t i o n of N. mercedis (top) and S. oregonensis (bottom) at dusk on October 20, 1993 99 Fig. 4.1: Light i n t e n s i t y depth p r o f i l e i n experimental tubes (primary axis) and corresponding depth of sim i l a r l i g h t i n t e n s i t i e s i n Kennedy Lake at surface l i g h t i n t e n s i t y of 1000 uE-s^-m\"2 and k=0.401 (secondary axis) 105 Fig. 4.2: Experiment 1 - t y p i c a l day and night d i s t r i b u t i o n s of Kennedy Lake S. oregonensis i n cyclinders i n fishwater [ +\/-] treatments 109 Fig. 4.3: Proportion of Kennedy Lake S. oregonensis below 100 cm during the day i n experiment 1. S o l i d squares indicate fishwater [-] treatments and open squares indicate fishwater [ + ] treatments 110 Fig. 4.4: Experiment 1 - change i n the proportion of Kennedy Lake S. oregonensis below 100 cm from night to day i n fishwater [ + \/-] treatments I l l Fig. 4.5: Experiment 2 - t y p i c a l day and night d i s t r i b u t i o n s of Great Central Lake S. oregonensis i n cyclinders i n fishwater [ + \/-] treatments 115 Fig. 4.6: Experiment 2- change i n the proportion of Great Central Lake S. oregonensis below 100 cm from night to day i n fishwater [ + \/-] treatments 116 Fig. 4.7: Experiment 2 - proportion of Great Central Lake S. oregonensis below 100 cm during day 117 X I F i g . 4.8: Experiment 3 - change i n p r o p o r t i o n o f Kennedy-Lake S. oregonensis below 100 cm from n i g h t t o day. S o l i d squares i n d i c a t e f i s h [-] t r e a t m e n t s and open squares i n d i c a t e f i s h [ + ] t r e a t m e n t s 121 F i g . 4.9: I n c r e a s e i n t h e p r o p o r t i o n o f Kennedy Lake S. oregonensis below 100 cm from n i g h t t o day i n f i s h w a t e r [ + \/-] t r e a t m e n t s o f e x p e r i m e n t 3 122 F i g . 4.10: Day and n i g h t w e i g h t e d mean depth of S. oregonensis i n Kennedy Lake (top) and G r e a t C e n t r a l Lake (bottom) f o r m u l t i p l e sample d a t e s 124 F i g . 5-1: Cohort a n a l y s i s o f S. oregonensis i n l a t e w i n t e r \/ s p r i n g of 1993 (Kennedy Lake) 131 F i g . 5-2: Copepod f i l t e r i n g r a t e s as a f u n c t i o n body w e i g h t . G e n e r a l e q u a t i o n f o r copepods t a k e n from P e t e r s and Downing (1984), t h e v a l u e s f o r mature S. oregonensis from Richman (1966) 135 F i g . 5-3: Model r e a c t i v e d i s t a n c e s o f j u v e n i l e sockeye a t n i g h t (top) and s t i c k l e b a c k d u r i n g the day (bottom) a t each depth f o r 0.25 mm, 0.50, 0.75 mm and 1.00 mm s i z e p r e y 139 F i g . 5-4: Model e f f e c t o f s t i c k l e b a c k and sockeye f e e d i n g on day (open squares) and n i g h t ( a s t e r i s k s ) o p t i m a l growth and depths of S. oregonensis 141 F i g . 5-5: Model e f f e c t o f s t i c k l e b a c k and j u v e n i l e sockeye f e e d i n g on f i n a l body s i z e and l i f e t i m e s u r v i v a l p r o b a b i l i t y o f S. oregonensis a t 1700 ( s o l i d s quares) and 1900 (open squares) p h y t o p l a n k t o n c e l l s p e r u l . . . 143 F i g . 5.6: Model e f f e c t o f p h y t o p l a n k t o n c e l l d e n s i t y on (A) day and n i g h t o p t i m a l depths and growth o f S. oregonensis and on (B) t h e f i n a l body s i z e and l i f e t i m e s u r v i v a l p r o b a b i l i t y 144 F i g . 5.7: Model e f f e c t o f changes i n S. oregonensis m e t a b o l i c r a t e p a r a m e t e r s a t f o u r l e v e l s o f p h y t o p l a n k t o n f o o d d e n s i t y . The we i g h t c o e f f i c i e n t and i n t e r c e p t were v a r i e d \u00b1 10% and the temp e r a t u r e c o e f f i c i e n t \u00b1 20% 147 F i g . 5.8: Model e f f e c t o f v a r y i n g S. oregonensis f i l t e r i n g r a t e parameters between 90-110% of d e f a u l t v a l u e s f o r f o u r l e v e l s of p h y t o p l a n k t o n f o o d d e n s i t y 148 F i g . 5.9: Model parameter v a l u e s t h a t p r e d i c t n o n - m i g r a t i o n of s m a l l e r S. oregonensis 150 X l l Acknowledgement s I g r a t e f u l l y acknowledge the support and patience of Paula Gongalves, my wife, during the extended period of time i t took to complete this thesis. Dr. J. D. McPhail supervised the thesis project with competence and wisdom. Drs. K. Hyatt, J. Myers, and W.E. N e i l l provided useful input and guidance as committee members. Many early conversations with K. Hyatt were useful i n the genesis and i n i t i a l stages of the thesis project. Assistance with equipment and f i e l d surveys was provided by James Baxter, Leonard Ghan, Paula Gongalves, Leonardo Huato, Chantal Ouimet, D.P. Rankin, Lynda Ritchie, Jordan Rosenfeld, Lome Rothman, Ron Saimoto, Peter Troffe, and Olfe Zimmermann. Research funding was provided by NSERC research grants to Dr. J.D. McPhail, FOC Salmonid Enhancement Program funding to Dr. K.D. Hyatt, a Provincial Fisheries grant, and NSERC postgraduate scholarships. Dr. K.D. Hyatt provided for the use of some FOC f i e l d equipment. 1 Chapter 1 Genera l I n t r o d u c t i o n Many pelagic organisms migrate v e r t i c a l l y between shallower water at night and deep water during the day. Reverse v e r t i c a l migrations (between deep water at night and shallower water during the day) are also observed. The l i t e r a t u r e on such v e r t i c a l migrations i s extensive and dates back more than a century (Ringelberg 1 9 9 3 ) . In thi s thesis I investigate three questions concerning v e r t i c a l migration as an adaptive behaviour: what i s the benefit of v e r t i c a l migration, i s there a cost to this behaviour, and i s migration behaviour of individuals fixed or a f l e x i b l e response to environmental factors that a f f e c t the benefit or cost? I have studied the v e r t i c a l migrations of a freshwater lacustrine copepod, Skistodiaptomus oregonensis (Pennak - formerly Diaptomus oregonensis). In thi s chapter, I review the background l i t e r a t u r e , outline the study system, and describe the research strategy. Hypotheses concerning the adaptive nature of vertical migration There are several hypotheses about the benefits of v e r t i c a l migration (for reviews see Kerfoot 1 9 8 5 , Bayly 1 9 8 6 , Haney 1 9 8 8 , Lampert 1 9 8 9 ) , and di f f e r e n t s e l e c t i v e factors may operate i n di f f e r e n t situations (eg. Bayly 1 9 8 6 ). Three hypotheses that apply to lake dwelling zooplankton are described below. One hypothesis, the foraging e f f i c i e n c y hypothesis, argues that v e r t i c a l migration maximizes feeding rates. Alewives (Janssen and Brandt 1980 ) and freshwater sardines, Limnothrissa miodon (Begg 1976) track the movements of t h e i r v e r t i c a l l y migrating zooplankton prey. C a r r i l l o et al. (1991) suggest that phytoflagellates migrate to follow nutrient patches generated by the excretions of s i m i l a r l y migrating herbivorous zooplankton. Heuch et al. (1995) propose that the p a r a s i t i c copepod salmon louse, Lepeophtheirus salmonis, undertakes reverse migrations that are the opposite of t h e i r salmonid hosts to increase encounter rates with the host at dawn and dusk. A second hypothesis i s the metabolic e f f i c i e n c y hypothesis. This hypothesis contends that growth benefits are attained by spending part of the day i n deep water where cold temperature lowers the metabolic rate. Brett (1983), modelled the metabolism of sockeye salmon and suggested that v e r t i c a l migration allows a larger f r a c t i o n of ingested energy to be allocated to growth. McLaren (1963) proposed that migrating zooplankton move down into colder water to process food acquired at the surface and that t h i s results i n a decease i n t h e i r mean metabolic rate. This, i n turn, causes an upward s h i f t i n the equilibrium size at which intake rate equals metabolic rate and, thus, a larger body size and greater fitness through higher fecundity. Neither of these two hypotheses excludes the other, and a v e r t i c a l migration strategy could be selected through growth maximization r e s u l t i n g from the combined effects of both depth-dependent foraging and metabolic rates. Bevelhimer and Adams (1993) describe a growth model for kokanee salmon i n which v e r t i c a l migration produces the highest growth through such combined effe c t s . Enright (1977) and Enright and Honegger (1977) propose that zooplankton gain the greatest net energy through feeding at the surface at night when algae are more abundant, and 3 of higher quality, which increases feeding e f f i c i e n c y , and then moving to depth during the day to minimize metabolic costs. The t h i r d hypothesis i s the predator avoidance hypothesis. This hypothesis argues that v e r t i c a l migration i s undertaken to decrease predation r i s k near the surface during daylight when di u r n a l l y foraging v i s u a l predators are active (Lampert 1993). Si m i l a r l y , the reverse v e r t i c a l migration could be to avoid nocturnal, surface-dwelling predators. Recent f i e l d comparisons and experiments provide support for the predator avoidance hypothesis for many species (Luecke 1986, Bollens and Frost 1989a, Bollens and Frost 1989b, Levy 1990, N e i l l 1990, Ohman 1990, S t i r l i n g et al. 1990, Tjossem 1990, Wright and Shapiro 1990, Bollens and Frost 1991, Dini and Carpenter 1991, Ringelberg 1991a, Ringelberg 1991b). Opportunity cost of predator avoidance Food resources are often concentrated near the lake surface and v e r t i c a l migration away from the surface to avoid predators or high temperature may thus r e s u l t i n a cost of l o s t feeding opportunity (Huntley and Brooks 1982, Dagg 1985, Gliwicz 1985, Johnsen and Jakobsen 1987, Pijanowska and Dawidowicz 1987, Gabriel and Thomas 1988, Mangel and Clarke 1988, Lampert 1989, Guisande et al. 1991, Dini and Carpenter 1992). There i s some ind i r e c t evidence for such a cost. F i e l d observations of Neocalanus plumchrus (Dagg 1985) and observations on Calanus pacificus i n large deep marine water tanks (Huntley and Brooks 1982) and Daphnia i n enclosures (Johnsen and Jakobsen 1987) a l l indicate that low food abundance i s correlated with a decrease, or cessation of, downward migration during the day. These findings are consistent with the interpretation that v e r t i c a l migration to depth result s i n a feeding opportunity cost and that as food becomes scarce, the cost becomes too high r e l a t i v e to the benefits and the organisms remain near the surface. Students of the r e l a t i o n s h i p between diurnal feeding rhythms and diurnal v e r t i c a l migration have variously argued both for and against a feeding opportunity cost. Some authors (eg., Gauld 1953) found that zooplankton do not feed i f they v e r t i c a l l y migrate to deep water during the day, but do feed over the entire diurnal period i f they remain at the surface. Others f i n d that the diurnal feeding rhythms p e r s i s t whether migration occurs or not (eg., Stearns 1986). This suggests that migration and feeding rhythms are independent behaviours. Fixed or flexible vertical migration V e r t i c a l migration patterns can vary among populations of a species ( N e i l l 1992, Stewart and Sutherland 1993), over time within a population (Ohman 1990, Frost and Bollens 1992, Hays et al. 1995), and, at a given time, among individuals within a population ( S t i r l i n g et al. 1990). The migration pattern of each in d i v i d u a l may be fixed and differences among individuals may be due to di f f e r e n t genotypes. I f so, the various genotypes e x i s t due to selective pressures that vary among habitats or that fluctuate over time i n the same habitat. A l t e r n a t i v e l y , the va r i a t i o n i n migration patterns could be due to f l e x i b l e behaviour of individuals i n response to di f f e r e n t conditions both int e r n a l (hunger or energy) and external (predation r i s k ) that 5 vary across time and place. Whether f l e x i b l e or fixed v e r t i c a l migration behaviour evolves may depend on the p a r t i c u l a r environmental conditions. Inducible defenses should evolve when predation r i s k varies unpredictably, when the time necessary to acquire the defensive t r a i t i s b r i e f r e l a t i v e to the fl u c t u a t i o n i n threat, when the fitness cost of maintaining the machinery to produce the induced defence does not exceed the benefits of the induced defence, when r e l i a b l e and non-fatal cues are available, and\/or when the fitness costs of defence offset some of the benefits of the defence (Harvell 1990, Pijanowska 1993). Recent experiments indicate that zooplankton migration i s a f l e x i b l e behaviour induced by the presence of predators, or predator cues, and suppressed when predators or t h e i r cues are absent. Examples include Acartia hudsonica avoiding stickleback predation (Bollens and Frost 1991), Daphnia avoiding f i s h predators (Dini and Carpenter 1992), Daphnia avoiding Chaoborus (Ramcharan et al. 1992), Diaptomus kenai avoiding Chaoborus ( N e i l l 1990), and Chaoborus avoiding f i s h predators (Tjossem 1990 and Dawidowicz et al. 1990). Decreases i n migration associated with decreases i n food abundance (Huntley and Brooks 1982, Dagg 1985, Johnsen and Jakobsen 1987) suggest f l e x i b i l i t y i n migration behaviour as a response to foraging opportunities. F l e x i b l e migratory responses to predator abundance and food a v a i l a b i l i t y would be esp e c i a l l y advantageous r e l a t i v e to fi x e d behaviour when there are costs associated with the behaviour. There i s evidence for within population genetic differences i n the migration of zooplankton. For example, protein gel electrophoresis indicates f i v e common genotypes i n a Daphnia population ( S t i r l i n g et al. 1990). One genotype was much more abundant i n the hypolimnion during the day. This genotype was also most abundant i n the winter and may be adapted to both cold water and low oxygen conditions. Another example are Daphnia magna clonal l i n e s . Lines i s o l a t e d from d i f f e r e n t populations showed g e n e t i c a l l y d i s t i n c t phototactic responses (De Meester and Dumont 1988) . Some clones were photopositive, others photonegative, and a t h i r d type were \"gypsies\" that migrate continuously between low and high l i g h t i n t e n s i t y environments. Although these three types were i s o l a t e d from d i f f e r e n t populations, clonal lines i s o l a t e d from within a population also showed d i f f e r e n t v e r t i c a l d i s t r i b u t i o n s i n a v e r t i c a l l i g h t gradient set up i n aquaria (De Meester and Dumont 1989). These same authors demontrated combined genetic (fixed) and environmental (flexible) effects on v e r t i c a l d i s t r i b u t i o n i n Daphnia. The amount of food supplied to each clon a l l i n e during culture affected t h e i r v e r t i c a l d i s t r i b u t i o n , and there was a s i g n i f i c a n t i n t e r a c t i o n between clonal l i n e and food supply e f f e c t s . One clonal l i n e i s o l a t e d from a d i f f e r e n t population remained photonegative at a l l times, and never migrated upward at any feeding l e v e l . Similarly, Dodson (1990) showed that depth d i s t r i b u t i o n responses to predators varied for d i f f e r e n t species of Daphnia c o l l e c t e d from d i f f e r e n t lakes. Study system My investigation of v e r t i c a l migration behaviour focusses on migratory populations of S. oregonensis i n Kennedy and Paxton 7 lakes and on non-migratory populations from Great Central and Hobiton lakes. These lakes provide a system to consider alternative hypotheses concerning the benefits of v e r t i c a l migration, and to investigate the p o t e n t i a l costs of t h i s behaviour. Because a l l four lakes are located i n south coastal B r i t i s h Columbia they are easy to access, and are subject to si m i l a r climatic conditions. This f a c i l i t a t e s comparison. The r e l a t i v e l y simple aquatic communities and trophic structure i n these lakes also s i m p l i f i e s analysis of p o t e n t i a l species interactions. Low productivity i n these lakes may select for v e r t i c a l migration to maximize foraging and\/or metabolic e f f i c i e n c y and i f there are feeding opportunity costs associated with v e r t i c a l migration they should be detectable. Alternatively, predation pressure i s p o t e n t i a l l y high on zooplankton i n these lakes from feeding by juvenile sockeye Oncorynchus nerka, stickleback Gasterosteus aculeatus, Neomysis mercedis, and phantom midge Chaoborus. High predation may select for v e r t i c a l migration as a predator avoidance strategy. Another advantage of studying these lakes i s that Fisheries and Oceans Canada (FOC) research on the limnology and f i s h community composition i n Kennedy, Great Central, and Hobiton lakes provides the background information necessary for interlake comparisons and modelling. Research strategy In Chapter 2, I use a comparative approach to evaluate the three hypotheses for the benefits and costs of v e r t i c a l migration. I attempt to correlate interlake differences i n 8 migration patterns with the dens i ty of predators , food abundance and d i s t r i b u t i o n , temperature, and other environmental factors predic ted by the various hypotheses to a f fec t the benef i ts of v e r t i c a l migrat ion. I also compare the d i u r n a l feeding rhythms of migratory and non-migratory S. oregonensis to test pred ic t ions made by the hypothesis that v e r t i c a l migrat ion resu l t s i n a feeding opportunity cost . In Chapter 3, I test the c o r o l l a r y of the predator avoidance hypothesis that v e r t i c a l migrat ion re su l t s i n a decrease i n the r i s k of predation by three p o t e n t i a l predators (s t ickleback, juveni l e sockeye, and Neomysis mercedis). This i s accomplished by r e l a t i n g the timing and extent of v e r t i c a l migration of S. oregonensis i n Kennedy and Paxton lakes to estimated s p a t i a l -temporal d i s t r i b u t i o n s of predat ion r i s k from each predator. In Chapter 4, I use laboratory experimental manipulations and f i e l d data to test whether i n d i v i d u a l S. oregonensis exh ib i t changes i n v e r t i c a l migration behaviour i n response to changes i n predat ion r i s k . In Chapter 5 I use a t h e o r e t i c a l dynamic opt imizat ion model, incorporat ing both a v i s u a l predat ion r i s k model and a bioenerget ic growth model, to explore whether S. oregonensis v e r t i c a l migration i n Kennedy Lake can be represented q u a n t i t a t i v e l y and r e a l i s t i c a l l y as a trade-of f between maximizing feeding and minimizing predator r i s k . In Chapter 6, I summarize my work, attempt to place i t i n context, and out l ine further research opportuni t i es . 9 Chapter 2 Testing hypotheses on the benefits and costs of ver t i ca l migration by Skistodiaptomus oregonensis by comparing populations Introduction Several hypotheses have been advanced to explain the adaptive si g n i f i c a n c e of v e r t i c a l migration by zooplankton (reviewed by Kerfoot 1985, Bayly 1986, Lampert 1989). One, the predator avoidance hypothesis, proposes that migration to deeper, darker water during the day reduces v u l n e r a b i l i t y to d i u r n a l l y foraging v i s u a l predators (Lampert 1993). Assuming food i s less concentrated i n deeper water, there may be a feeding opportunity cost associated with v e r t i c a l migration to avoid predators. An alter n a t i v e hypothesis, the bioenergetic hypothesis, proposes that cooler temperature i n deeper water reduces metabolic costs and, thus, increases growth. A t h i r d hypothesis, the foraging e f f i c i e n c y hypothesis, argues that v e r t i c a l migration increases food intake by following a migrating prey resource. If the metabolic advantage of residing i n cool water during the day i s coupled with higher energy intake rates when feeding at night (or during the crepuscular period when prey are more concentrated or of higher q u a l i t y ) , there may be a combined foraging\/bioenergetic benefit. In this chapter I evaluate these hypotheses concerning the benefits and costs of v e r t i c a l migration of the zooplankter Skistodiaptomus oregonensis by comparing populations i n four coastal lakes i n B r i t i s h Columbia. Such interpopulation comparisons of copepods i n lakes with r e l a t i v e l y simple 10 vertebrate and invertebrate communities provide an opportunity, within the natural environment, to look for associations between v e r t i c a l migration and factors hypothesized to a f f e c t benefits and costs. Methods Zooplankton Field Collections The four study lakes are located i n southwestern British-Columbia (Fig. 2.1). Samples were c o l l e c t e d at deep offshore locations i n each lake (Fig. 2.2-2.5). To determine the v e r t i c a l d i s t r i b u t i o n of zooplankton i n Kennedy, Great Central, and Hobiton lakes zooplankton samples were c o l l e c t e d simultaneously i n each lake at 7 sample depths by towing conical 100 um nitex mesh plankton nets attached at in t e r v a l s along a single rope. Each net was 108 cm long with a 15 cm diameter opening. Tows were 20 minutes i n duration at a speed of 50 cm per second. During a tow, an angle of 50 degrees from the horizontal was maintained. A maximum of 40 seconds passed from the time the f i r s t net was released into the water u n t i l a l l nets were i n place at the towing depth. Retrieval of the nets required a maximum of 1 minute. Thus, r e l a t i v e to the amount of water f i l t e r e d from the sample depths, nets f i l t e r e d only small volumes of water during descent and r e t r i e v a l . Trigonometric calculations were used to estimate sample depths of 1, 3, 5, 7, 10, 17, and 24 m, assuming no curvature i n the tow l i n e . Weighted mean depths (WMD) for S. oregonensis were calculated Fig. 2.1: Location of the four study lakes in southwestern British Columbia. Fig. 2.2: Map of Great Central Lake showing sample location (X) and lake contours in meters. Depth contours taken Rutherford etal. (1986). Fig. 2.3: Map of Hobiton Lake showing location of sample site (X) and lake contours in meters. Other details as in Fig. 2.2. Fig. 2.5: Map of Paxton Lake showing location of sample site (X) and depth contours in meters. Other details as in Fig. 2.2. 16 from the s t r a t i f i e d samples as WMD= ^ n \u00b1 L i d \u00b1 , (2.1) where nL i s the abundance (individuals per m3) i n the depth i n t e r v a l A i with midpoint d i (Osgood and Frost 1994). Borders between depth s t r a t a were defined as midway between subsequent sample depths. The t o t a l number of individuals within each stratum (TJ i s expressed as the number under a i m 2 surface area and was estimated as 2 i = i 2 i A i (2.2) A depth sounder was employed from a second boat during one deployment of the sample nets to determine the exact depth of each net. Analysis of the echograms indicates a s l i g h t overestimate (1-2 m) of the depth of the deepest two sample nets. The r e s u l t i n g bias i n the calculations i s too small to a f f e c t the ov e r a l l patterns observed or conclusions drawn. To track changes i n S. oregonensis v e r t i c a l d i s t r i b u t i o n during dawn and dusk i n Paxton Lake a series of v e r t i c a l l y s t r a t i f i e d zooplankton hauls were c o l l e c t e d for the 9-5, 5-3, and 3-0 m strata. The s t r a t a were sampled from deepest to shallowest i n rapid succession over approximately 8-12 minutes with a closable double-ringed Wisconsin-style 100 um nitex mesh plankton net 292 cm i n length with a 55 cm diameter opening. A 20 minute i n t e r v a l separated the s t a r t of one series of s t r a t i f i e d samples 17 and the star t of the next series. Upon r e t r i e v a l , samples were preserved i n a solution of 4% sugared formaldehyde. In the laboratory, a subsample of each sample was counted under a Wild M-5 microscope. Species were i d e n t i f i e d using the key i n Pennak (1989). Each sample was sub-sampled u n t i l either 100 S. oregonensis were counted or u n t i l the entire sample was counted. For some sample dates, subsampling ceased before counting 100 S. oregonensis. This happened when the subsample volume exceeded 10 times the volume required to count 100 S. oregonensis from samples at the other depths from the same sample tow. When lengths were measured these were recorded with an automated micro-computer based d i g i t i z e r system described i n Roff and Hopcroft (1986). Metasome lengths are used i n most size class analyses because metasome length can be measured with greater accuracy than t o t a l length. Measuring food available to SL_ oregonensis To estimate the depth d i s t r i b u t i o n of food available to S. oregonensis i n the study lakes, I determined the concentration of chlorophyll a extracted from the 3-50 um size f r a c t i o n of phytoplankton i n lake water from discrete sample depths. This size f r a c t i o n approximates the size of p a r t i c l e s selected by S. oregonensis i n feeding experiments on lake seston (Vanderploeg 1990) and lake phytoplankton (McQueen 1970). In my samples, lake water was f i r s t passed through a 50 um nitex mesh to remove larger p a r t i c l e s . A measured volume (3 00-1000 ml) of the remaining f i l t r a t e was passed through a Poretics Corporation 47 mm diameter, 3 um polycarbonate membrane f i l t e r . The membrane 18 was placed i n a p e t r i dish, frozen on dry ice, and retained i n the dark. Within one week of c o l l e c t i o n , chlorophyll a and phaeopigment were extracted and measured with a Turner model 10-AU spectrofluorometer and following the procedure described by Parsons et al. (1984). Weighted mean depth of chlorophyll a depth d i s t r i b u t i o n s were calculated using the same method used for zooplankton (equation 2.1). This estimate of available food does not include non-fluores'cing alternative food sources (e.g., c i l i a t e p r o t i s t s , heterotrophic f l a g e l l a t e s , and fine p a r t i c u l a t e s ; Stoecker and Capuzzo 1990, G i f f o r d and Dagg 1991, Hartmann et al. 1993, Ohman and Jeffrey 1994, Cervetto et al. 1995). Presumably, i n freshwater, these alternative prey are more abundant near the surface than below 10 m since marine c i l i a t e s are found almost exclusively within 5 m of the surface (Jonsson 1989). S. oregonensis gut pigment analysis To estimate feeding a c t i v i t y at the time of c o l l e c t i o n , the quantity of chlorophyll a and phaeopigment i n the guts of f i e l d caught S. oregonensis was determined i n Kennedy and Great Central lakes. To compare gut pigments across depth, zooplankton were collected from discrete depths with the horizontal tow method described e a r l i e r . To measure changes i n gut pigment during dawn and dusk and, to compare day versus night gut pigments, v e r t i c a l hauls from 24 m to the surface were made with the Wisconsin-style plankton net described above. On capture, S. oregonensis were collected onto 8 X 8 cm pieces of nitex mesh, placed i n p e t r i dishes and frozen on dry i c e . In the laboratory, 19 samples were thawed i n t a p wat e r and 10 i n d i v i d u a l a d u l t S. oregonensis were randomly s e l e c t e d , r i n s e d i n d i s t i l l e d w a t e r , and, f o r c h l o r o p h y l l a e x t r a c t i o n , p l a c e d i n t o 5 ml o f 90% acetone and s t o r e d o v e r n i g h t i n the d a r k a t 5\u00b0. C h l o r o p h y l l a and phaeopigment c o n c e n t r a t i o n i n the e x t r a c t were d e t e r m i n e d u s i n g the same method used f o r l a k e water c h l o r o p h y l l a samples. Results Vertical distributions of S_^_ oregonensis and phytoplankton food resources in the study lakes-Kennedy Lake and Pa x t o n Lake b o t h c o n t a i n m i g r a t o r y p o p u l a t i o n s o f S. oregonensis. I n Kennedy Lake a l l s i z e c l a s s e s o f S. oregonensis l a r g e r t h a n 450 um were c o n s i s t e n t v e r t i c a l m i g r a t o r s on a l l s u r v e y d a t e s . T y p i c a l day and n i g h t v e r t i c a l d i s t r i b u t i o n s f o r d i s c r e t e s i z e c l a s s e s a r e d i s p l a y e d i n F i g u r e 2.6 and F i g u r e 2.7. Mean d e p t h of S. oregonensis i n t h e 750-850 Vim metasome l e n g t h c l a s s ( F i g . 2.8A) was s i g n i f i c a n t l y deeper d u r i n g t h e day tha n a t n i g h t i n samples from 1992 t o 1994 ( p a i r e d sample t - t e s t : t=12.805, P<0.001, n=13). The mean depth f o r t h e d i s t r i b u t i o n o f the 3-50 um s i z e f r a c t i o n of p h y t o p l a n k t o n i n Kennedy Lake i s c o n s i s t e n t l y l o c a t e d n ear the s u r f a c e ( F i g . 2.8B) and does n ot d i f f e r s i g n i f i c a n t l y from day t o n i g h t ( p a i r e d sample t - t e s t : t=2.127, 0.10>P>0.05, n=8). Except i n w i n t e r , when abundance i s an o r d e r o f magnitude lower and t h e p h y t o p l a n k t o n a r e homogenously d i s t r i b u t e d down t o 24 m ( F i g . 2.9), the 3-50 um s i z e f r a c t i o n i s c o n s i s t e n t l y most abundant above 10. meters. D u r i n g summer s t r a t i f i c a t i o n , p h y t o p l a n k t o n d e n s i t i e s d e c l i n e above t h e base o f t h e 20 Fig. 2.6: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Kennedy Lake on August 25, 1993. Densities are point estimates at the depths indicated on the vertical axis. Fig. 2.7: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Kennedy Lake on November 12, 1992. 22 0 5 10\" Q. 0) 15 Q 20 25 Nigh: 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 10fc 8\" 151 Q 201 25 S. oregonensij O P o o o o o o oo o o o 1 sz Q. 0) 15 Q Chlorophyll a c j | O 20 25\" 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 15-Jul-92 31-Jan-93 19-Aug-93 07-Mar-94 23-Sep-94 O mean day depth \u2022 mean night depth Fig. 2.8: A) Mean day and night depth of 750-850 um metasome length S. oregonensis in Kennedy Lake on 13 sample dates from 1992-1994. B) Mean day and night depth of 3-50 um size fraction of phytoplankton in Kennedy Lake from 13 sample dates from 1992 to 1994. C) Data in part A and B redrawn to compare mean day and night depth of S. oregonensis with mean day and night depth of 3-50 um size fraction of phytoplankton. Error bars in A and B indicate standard deviations for each depth distribution. CO CD c Q. CD Q 23-Jul-92 1 3 5 7 10 17 24 0 12-NOV-92 24-Feb-93 \u2014I 0.4 0 1111111 1 0.31 0 14-Apr-93 13-May-93 m 1 0.071 o 1 0.45 0 1 23-Jun-93 1 0.41 0 0.5 cti CD \u20224\u2014\u00bb C Q. CD Q 16-Jul-93 1 3 5 7 10 17 24 0 26-Aug-93 1 i$Sm!fl$HtHtt! 20-Oct-93 5-May-94 11-Jun-94 12-Jul-94 0.3 0 0.51 0 0.38 0 Chlorophyll a (|jg\/l) 1.1 0 \u2014 i 0.62 0 1.0 Fig. 2.9: Vertical distribution of chlorophyll a from the 3-50 urn size fraction of phytoplankton on 12 sample dates in Kennedy Lake. All samples shown were taken during daytime except for July 12, 1994. 24 thermocline, which l i e s close to, or s l i g h t l y below, 10 meters (Fig. 2.10). In spring, despite the absence of a strong thermocline, phytoplankton densities decrease at s i m i l a r depths. S. oregonensis migrate from a daytime depth below the thermocline where phytoplankton are less abundant to a night depth that i s at or above the thermocline which brings them into close proximity to t h e i r phytoplankton food resource (Fig. 2.8C). Across sample dates from 1992-1994, mean daytime depth of S. oregonensis i s s i g n i f i c a n t l y below the mean daytime depth of phytoplankton (paired sample t - t e s t : t=9.757, P<0.001 ,n=ll); while at night, mean depth of S. oregonensis r i s e s and does not d i f f e r from mean depth of phytoplankton (paired sample t - t e s t : t=1.206, 0.5
CD c 4.0-9.0 CL CD 9.0-12.0 0 15 30 I I I 3 2 # Diaptomus*10\/ m I 3:10 i i 1 i - 1 1 1 1 3:50 4:30 Local Apparent Time 5:10 5:50 Fig. 2.11: Changes in density of S. oregonensis in three depth strata of Paxton Lake during dawn of August 19, 1994. tSJ en 0.0-4.0 CO CD \u2022\u00bb\u2014\u2022 c Q _ 0 Q 4.0-9.0 9.0-12.0 0 15 30 1 I I 3 2 # Diaptomus*10\/m I l l l l I I I I I P I J J I Ji 18:20 1 I I 19:00 ~ i r 19:40 - 1 r 20:20 Local Apparent Time - I \u2014 21:00 Fig. 2.12: C h a n g e s in density of S. oregonensis in three depth strata of Paxton Lake during dusk of August 19, 1994. M \u2014I 28 4 3 2 1 0 1 2 3 4 Chlorophyll a (ug\/l) Fig. 2.13: Day and night vertical distribution of chlorophyll a from the 3-50 pm size fraction of phytoplankton in Paxton Lake, August, 1994. 29 lakes are not. Typical d i s t r i b u t i o n s of 5 size classes are shown for two sample dates i n Figure 2.14 and Figure 2.15. In Great Central Lake, the means of depth d i s t r i b u t i o n s of S. oregonensis i n the 750-850 um metasome length class did not d i f f e r day versus night (Fig. 2.16A; paired sample t - t e s t : t=0.552, P>.5, n=6). V e r t i c a l d i s t r i b u t i o n s i n Great Central (Fig. 2.16A) were more variable both between and across dates, compared to Kennedy Lake (Fig. 2.8A). During February, A p r i l , and May sampling i n 1993, when S. oregonensis densities were extremely low, a s i g n i f i c a n t proportion of the population were found i n the 17 and 24 m str a t a either during the day, during the night, or at both times. These results are not an a r t i f a c t of low sample si z e : at least 40 S. oregonensis were counted at the depth of maximum density for each of these d i s t r i b u t i o n s . Neither the day nor night d i s t r i b u t i o n s were-consistently deeper. The 3-50 um size f r a c t i o n of phytoplankton was also at the surface with no s i g n i f i c a n t difference i n mean depths between day and night (Fig. 2.16B; paired sample t - t e s t : t=2.086, 0.2>P>0.1, n=5). Figure 2.16C i l l u s t r a t e s the si m i l a r mean depth of S. oregonensis and the 3-50 um size class of phytoplankton both day and night. In Hobiton Lake a l l sizes of S. oregonensis were non-migratory. Most individuals i n a l l size classes remained at the surface both day and night (Fig. 2.17 and Fig. 2.18). The depth d i s t r i b u t i o n of both large S. oregonensis and the 3-50 um size f r a c t i o n of phytoplankton were near the surface day and night (Fig. 2.19). 30 1 3 , _ 5 sz 7 Q . CD 10 Q 17 24 # individuals\/m 450-550 um 40 20 0 t # individuals\/m' 20 40 60 650-750 um \u2022 899999999993 B99999999] 60 0 # individuals\/m 60 # individuals\/m Fig. 2.14: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Great Central Lake on July 7, 1992. 31 # individuals\/m # individuals\/m 250 125 0 125 250 300 150 0 150 300 # individuals\/m # individuals\/m 40 20 0 20 40 # individuals\/m Fig. 2.15: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Great Central Lake on August 18, 1992. 32 o-5' 10-x: a. 15-CD Q 20' 25' Day 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 5 \u2022 E, 10 \u2022 Q. 15 \u2022 CD Q 20 \u2022 25- Night 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 0 5 ^ 10 j= 15 H g 20 25 H Night 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 E. n 03 Q 0' 5' 10-15-20-25-S. oregonensis O \" 1 O o \" \u00b0 \u2022 \u2022 o E. Q. CD Q 25 Chlorophyll a O mean day depth \u2022 mean night depth 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 26-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 Fig. 2.16: A) Mean day and night depth of 750-850\/vm metasome length S. oregonensis in Great Central Lake on 7 sample dates in 1992 and 1993. B) Mean day and night depth of 3-50 um size fraction of phytoplankton in Great Central Lake from 7 sample dates from 1992 and 1993. C) Data in part A and B redrawn to compare mean day and night depth of S. oregonensis with mean day and night depth of 3-50 um phytoplankton. Error bars in A and B indicate standard deviations for each depth distribution. 33 500 250 0 250 500 400 200 0 200 400 # individuals\/m # individuals\/m 500 250 0 250 500 2000 1000 0 1000 2000 # individuals\/m # individuals\/m 250 125 0 125 250 # individuals\/m Fig. 2.17: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Hobiton Lake on August 5, 1992. 34 # individuals\/m # individuals\/m 1 3 E 5 E 10 CD Q 1000 500 0 # individuals\/m 500 2 1000 350 0 # individuals\/m 350 2 700 17 24 I 660-795 um 3500 1750 0 1750 3500 2 # individuals\/m Fig. 2.18: Day (hatched bars) and night (solid bars) vertical distribution for 5 metasome length classes of S. oregonensis in Hobiton Lake on June 25, 1992. 35 Diaptomus Chlorophyll a June 25 August 5 June 25 August 5 Fig. 2.19: Day (open circles) and night (solid squares) mean depth of 660-795 um metasome length class S. oregonensis and the 3-50 um size fraction of phytoplankton on June 25 and August 25, 1992. Bars indicate standard deviations of the population distribution. 36 Biological and physical comparison of study lakes The physical conditions do not d i f f e r i n any way that i s consistent with the differences i n migration behaviour of S. oregonensis (Table 2.1). Light e x t i n c t i o n values for Great Central and Hobiton (no v e r t i c a l migration) f a l l between those for Paxton and Kennedy Lake ( v e r t i c a l migration). Also, temperature and pH values do not d i s t i n g u i s h lakes where S. oregonensis migrate from lakes where they do not migrate. S. oregonensis migrate i n large and deep Kennedy lake, but also i n small and shallow Paxton Lake. Non-migratory S. oregonensis are also found i n both large and small lakes. In Great Central Lake, where juvenile sockeye density i s the highest among the lakes (Table 2.2), S. oregonensis do not migrate. In Hobiton Lake, where juvenile sockeye are present they also do not migrate; however, i n Kennedy Lake (where sockeye and stickleback are present) they migrate, and i n Paxton Lake (which contains stickleback but no juvenile sockeye) they also migrate. Zooplankton biomass and primary productivity are sim i l a r i n Kennedy, Hobiton, and Great Central Lake. The density of the 3-50 um size f r a c t i o n of phytoplankton i s also s i m i l a r for Kennedy, Hobiton and Great Central lakes, although i t i s higher i n Paxton Lake (Fig. 2.20). Among b i o l o g i c a l factors (Table 2.2), only the presence of pelagic stickleback d i s t i n g u i s h the two lakes where S. oregonensis migrate from the two lakes where they do not. Table 2.1: Physical conditions i n the four study lakes. lake elevation (m) area (km2) depth l i g h t compensation (m) ext. depth (m) pH no migration: Great Central Hobiton 82 15 51 3.6 2121 361 .44 0.44 18.5 10.4 7.0 6.8 migration: Kennedy Paxton 4 61 64 0.162 331 162 .38 501 12 .3 unknown 7.1 unknown -\"\u2022mean depth 2maximum depth Table 2.2: B i o l o g i c a l conditions i n the four study lakes. lake temp.1 limnetic 2 sockeye density (103\/ha) limnetic 2 stickleback density (103\/ha) zooplankton biomass (mg dry wt\/m3) annual primary production (g C\/m2) no migration: Great Central Hobiton 18.4 17 .8 1.356 0.775 03 0 9.0 11.0 18.9 28.7 migration: Kennedy - 18.1 Clayoquot Arm Paxton max=23 0.555 0.957 present 4 5.4 unknown 16.5 unknown xmean temp. June 1 and Sept. 3 0 at 1 m 2mean values for 10 annual estimates from 1986-1995 v i a hydroacoustic and trawl surveys i n late f a l l to early winter (K. Hyatt, FOC, unpublished data). Hyatt et al. 1984 and Hyatt and Stockner 1985 may be consulted for d e t a i l s of survey procedures and a n a l y t i c a l methods, respectively. 3although stickleback are present i n both GCL and Hobiton, FOC surveys indicate t h e i r v i r t u a l absence from open water. 4see McPhail 1993 00 39 Great Central Kennedy Hobiton Paxton Fig. 2.20: Mean concentration of chlorophyll a from the 3-50 um size fraction of phytoplankton in the 4 study lakes. Each point represents the mean all samples in the top 10 m averaged across sample dates. Bars represent 95% confidence intervals. 40 Body size and vertical distribution of S\\_ oregonensis in Kennedy Lake Both mean daytime depth of S. oregonensis i n Kennedy Lake (Fig. 2.21A), and the change i n mean depth from night to day (Fig. 2.21B), increase with metasome length up to approximately 450 um. Beyond th i s size there i s l i t t l e , i f any, further e f f e c t . S. oregonensis i n size classes 375 and 425 um are highly variable i n mean daytime depth between sample dates and show considerable change i n mean depth from night to day on some dates and not on others, although the change i s not as great as i t i s for the larger size classes. V e r t i c a l d i s t r i b u t i o n s both day and night of small metasome length classes of S. oregonensis i n Kennedy Lake for two dates are displayed i n Figure 2.22 and Figure 2.23. Gut pigments and diurnal feeding patterns of migratory and non-migratory oregonensis Migrating S. oregonensis from Kennedy Lake contain more chlorophyll a and phaeopigment i n t h e i r guts at night than during the day, while the non-migrating Great Central Lake population showed no diurnal feeding rhythm (Fig. 2.24). Within Kennedy Lake, daytime gut pigments decreased with the depth at which S. oregonensis were captured (Fig. 2.25). Gut pigments of the migrating Kennedy Lake population increase at dusk, the time of migration toward the surface, and decrease again at dawn at the time of descent. Non-migratory Great Central Lake S. oregonensis show no such pattern (Fig. 2.26 and Fig. 2.27) . 41 Q . CD \" D CD E CO \" D C CO CD 25 20 -15 -10 0 \u2022 \u2022 \u2022 C P \u2022 C P \u2022 \u00ab=*=\u2022 cP a \u2022 \u2022 \u2022 C O c5 \u2022 E P \u2022 \u2022 Cl \u2022 \u2022 CP \u2022 cP \u2022 \u2022 cF Cl \u2022 \u2022 cF C D \u2022 c P \u2022 \u2022 \u2022 \u2022 Z P \"=tb C P \u2022 Cl \u2022 200 300 400 500 600 700 800 900 1000 c CD E CD O _cd C L w .c \u2022\u2022\u2014> C L CD T J c CO CD 200 300 400 500 600 700 800 Length class midpoint (pm) 900 1000 Fig. 2.21: Top: S. oregonensis weighted mean daytime depth versus metasome length classes in Kennedy Lake. Bottom: Difference between weighted mean daytime depth and weighted mean night time depth for each sample date. Positive differences indicate greater daytime mean depth. Depth (m) Depth (m) Depth (m) 43 CL CD Q 1 3 5 7 10 17 24 <250 urn S66666666&J466666666664J >66663 'yyyyyyyyvyi 30 15 0 f # individuals\/m' 15 Q. CD Q 30 1 3 5 7 10 17 24 250-300 u m 200 100 0 100 # individuals\/m 200 CL CD Q 1 3 5 7 10 17 24 xxxxxxxxxxxxxxxxxxxxxxxi 300-350 Lim 320 160 0 # individuals\/m' 160 1 ^ 3 \u00a3 5 o 10 Q 17 24 320 350-400 u m \u2022 \u2022 A A A A A A A A A \/ y V v l S A A A A A A A J > W ^ A J t X X J l XXXXXXXXXXXJ 5553 360 180 0 180 # individuals\/m 360 Q. CD Q 1 3 5 7 10 17 24 340 XXXXXXXXJ VVVVVVVAVVVVS\/VVVWWVS \u2022 400-450 u m .AAXAAAAAAAAAAAAAAAAAAA^ 170 0 170 340 # individuals\/m' Fig. 2.23: Day (hatched bars) and night (solid bars) vertical distribution for 5 small metasome length classes of S. oregonensis in Kennedy Lake on November 12, 1992. 44 0.8 0.7-Night Day Fig. 2.24: Day versus night gut pigments in S. oregonensis in Kennedy Lake (solid square) and Great Central Lake (open squares) on May 11-13, 1993 (top), June 22-23, 1993 (middle), and July 15-16, 1993 (bottom). Each point represents the mean of 10 mature individuals. Lines join grand mean at each time within each lake. 45 0 10 15 Depth (m) 20 25 10 15 Depth (m) Fig. 2.25: Daytime S. oregonensis gut pigments versus depth on June 23, 1993 (top) and July 16, 1993 (bottom). Each point represents the mean of 10 individuals. Fig. 2.26: S. oregonensis gut pigment changes at dawn and dusk in Great Central Lake and Kennedy Lake, August 25-27, 1993. The dashed vertical lines indicate time of sunrise and sunset. Each point represents the mean of 10 individuals. 47 Fig. 2.27: S. oregonensis gut pigment changes at dawn and dusk in Great Central Lake and Kennedy Lake, October 17-20, 1993. Other details as in Fig. 2.26. 48 Relationship between weighted mean depth of oregonensis and chlorophyll a concentration in Kennedy Lake The mean daytime depth of S. oregonensis i n Kennedy Lake tends to decrease with decreasing mean chlorophyll a concentration from the 1,3, 5, and 7 meter depth (Fig. 2.28). I f the two o u t l i e r s (Jn93 and My94) are removed the mean chlorophyll a concentration i n the 1, 3,5, and 7 m samples explains 61% of the v a r i a t i o n i n the mean depth of S. oregonensis and the least-squares l i n e a r regression i s s i g n i f i c a n t (P < .02). Discussion The adaptive benefit of vertical migration The only factor examined that distinguishes the two lakes i n which S. oregonensis v e r t i c a l l y migrate from the two i n which they do not migrate i s the presence of pelagic sticklebacks i n the former. This i s consistent with the hypothesis that the ultimate selective force for S. oregonensis v e r t i c a l migration i s avoidance of stickleback predation. I f juvenile sockeye predation was dri v i n g v e r t i c a l migration, migrations would be expected i n Great Central Lake, where juvenile sockeye density i s more than twice that of Kennedy Lake, and migrations would not be expected i n Paxton Lake where there are no juvenile sockeye (Table 2.1). If foraging e f f i c i e n c y were the sel e c t i v e advantage for v e r t i c a l migration, food abundance or food d i s t r i b u t i o n should d i f f e r between the lakes where migration occurs and those where migration does not occur. However, neither food abundance nor food d i s t r i b u t i o n separate lakes where S. oregonensis migrate from lakes where they do not migrate. In the same vein, under 49 Q. CD \u2022o CD E -I\u2014\u00bb co ~o c co CD -12 -14 -j -16 -18 -20 -22 -24 Jn93 JI92NV92 p L g o My93 JI93 JI94 My94 Au93 Oc93 Ap93 Jn94 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 Mean chl a in top 10 m (pg\/l) Fig. 2.28: Weighted mean daytime depth of S. oregonensis in Kennedy Lake versus mean chlorophyll a concentration of the 3-50 um phytoplankton size fraction in the top 10 m. Dates (month\/year) indicated for each sample point. 50 the b i o e n e r g e t i c hypothesis, temperature p r o f i l e s i n the two cl a s s e s of lakes would need to d i f f e r i n such a way as to provide a temperature-dependent b i o e n e r g e t i c advantage f o r v e r t i c a l m i g r a t i o n i n the lakes where m i g r a t i o n occurs. However, temperature regimes are s i m i l a r i n a l l four l a k e s . Recently, evidence has accumulated f o r predator avoidance as the s e l e c t i v e f o r c e d r i v i n g v e r t i c a l m i g r a t i o n i n many zooplankton species (eg. Ohman 1990, D i n i and Carpenter 1991, Bol l e n s et al. 1992, N e i l l 1992, B o l l e n s et al. 1994). I n t e r l a k e comparisons demonstrating a p o s i t i v e r e l a t i o n s h i p between mig r a t i o n and pr e d a t i o n r i s k e x i s t f o r Cyclops abyssorum ( G l i w i c z 1986) and Diaptomus kenai ( N e i l l 1992). Schmidt et al. (1994) found greater m i g r a t i o n of Cyclops sp. and Diaptomus sp. i n lakes w i t h higher predation r i s k , although the species of zooplankton i n t h i s study were not reported and they may not have been the same i n d i f f e r e n t l a k e s . Stewart and Sutherland (1993) rep o r t a strong v e r t i c a l m i g r a t i o n of S. oregonensis i n McCargo lake i n co n t r a s t to weak or absent m i g r a t i o n of S. oregonensis i n Lime Lake i n western New York State. Data on pr e d a t i o n r i s k i n these lakes i s l a c k i n g . W i thin populations, many st u d i e s show seasonal onset, or changes, i n the i n t e n s i t y of zooplankton m i g r a t i o n that correspond to changes i n p r e d a t i o n r i s k . Examples i n c l u d e Acartia hudsonica (Bollens et al. 1992), Calanus pacificus (Bollens and Frost 1989b), Diaptomus sanguineus (De S t a s i o 1993), Pseudocalanus newmani (Ohman 1990), Cyclops abyssorum, Mesocyclops sp., Eudiaptomus gracilis, Daphnia hyalina, and Bosmina sp. ( S t i c h 1989). 51 Stickleback predation pressure The hypothesis that stickleback predation drives v e r t i c a l migration by S. oregenensis i n Kennedy and Paxton lakes requires that pelagic stickleback present a s i g n i f i c a n t predation r i s k . O'Neill and Hyatt (1987) demonstrated that sticklebacks i n Kennedy Lake prefer copepods, including S. oregonensis. Stickleback and juvenile sockeye, either i n a l l o p a t r y or sympatry, i n th e i r lake enclosures were strongly s i z e - s e l e c t i v e and decreased the mean size of zooplankton i n the enclosures from approximately 450-550 um (tot a l length) to between 190-330 um i n less than one month. Mature S. oregonensis range i n length from approximately 800 to 1200 um and are the largest zooplankton species i n Kennedy and Paxton Lake, with the exception of Neomysis mercedis (in Kennedy) and Chaoborus sp. (in Paxton). N. mercedis and Chaoborus are strong v e r t i c a l migrators themselves that are deep during the day and, thus, remain unavailable to stickleback. Such si z e - s e l e c t i v e predation may cause v e r t i c a l migration i n the larger individuals within a prey species but not i n the smaller individuals (eg. N e i l l 1992, Osgood and Frost 1994, Angeli et al. 1995). My data indicate that the t o t a l distance of v e r t i c a l migration by S. oregonensis does, indeed, decrease with size (Fig. 2.21). This size effect i s apparent for metasome length classes less than about 450 um (625 um t o t a l length). In comparison, O'Neill and Hyatt (1987) found that s i z e - s e l e c t i v e stickleback predation cropped the mean t o t a l length of zooplankton from 450-550 um to 190-33 0 um i n enclosure experiments i n Kennedy Lake. Thus, we might expect that below 52 t h i s size range (190-330 um) the strength of the v e r t i c a l migration i n S. oregonensis w i l l decline. Juvenile sockeye predation pressure Juvenile sockeye are also s i z e - s e l e c t i v e zooplanktivores, capable of exerting considerable predation pressure. O'Neill and Hyatt (1987) report that the s i z e - s e l e c t i v e cropping effect of juvenile sockeye i n Kennedy Lake enclosures was s i m i l a r to that exerted by stickleback. In Alaskan lakes Schmidt et al. (1994) observed v e r t i c a l migration of diaptomids where there were unusually high densities of juvenile sockeye, and no v e r t i c a l migration i n lakes where densities remained at lower, h i s t o r i c a l l e v e l s . Thus, i t i s surprising that juvenile sockeye, unlike the stickleback, do not seem to be dri v i n g v e r t i c a l migration i n my study lakes. This apparent paradox may be related to differences i n the timing of the predation r i s k from the two predators. The available evidence indicates that stickleback are diurnal predators (Manzer 1976) and that juvenile sockeye are primarily crepuscular feeders (Narver'1970, Barraclough and Robinson 1972, McDonald 1973). In Chapter 3, I examine i n d e t a i l the temporal-s p a t i a l pattern of predation by both sockeye and stickleback r e l a t i v e to the timing of v e r t i c a l migration. Vertical migration, feeding rhythms, and the cost of lost feeding opportunity In zooplankton, diurnal feeding rhythms showing higher food intake at night often correspond with v e r t i c a l migration from a deep daytime habitat (low food concentration) to a night-time 53 surface habitat (high food concentration). Examples include Calanus finmarchicus (Gauld 1953 and Simard et al. 1985), Calanus pacificus (Hassett and Landry 1988), and Calanus hyperboreus (Head et al. 1985). This suggests a d i r e c t cause and effect relationship between v e r t i c a l migration and feeding - a c o r o l l a r y of the hypothesis that v e r t i c a l migration results i n a cost of l o s t feeding opportunity. Contrary evidence suggests that DVM and feeding are controlled separately. Feeding rhythms i n the absence of v e r t i c a l migration are reported for Acartia tonsa (Durbin et al. 1990), for Calanus hyperboreus and Calanus glacialis (Head et al. 1985), for marine copepods (Morales et al. 1993), for Pseudocalanus sp. and Centropages hamatus (Nicolajsen et al. 1983), and for Paracalanus parvus, P. crassirostris, Acartia erythraea, and Eucalanus subcrassus (Tang et al. 1994). Also, the timing of feeding rhythms does not always correspond with v e r t i c a l migration movements (Dagg et al. 1989, Atkinson et al. 1992a, Tang et al. 1994). Feeding rhythms are known where migration depth ranges within a homogenous food d i s t r i b u t i o n (Haney 1985, Stearns 1986, Mourelatos et al. 1989, Atkinson et al. 1992a, Bollens and Stearns 1992, Cervetto et al. 1995), although a lack of feeding rhythm under these same conditions also occurs (Haney 1985, Lampert and Taylor 1985, Mourelatos et al. 1989) . In Kennedy Lake, S. oregonensis feeding rhythms correspond to movement between a surface habitat where food i s abundant and a deep habitat where food i s scarce, and the timing of v e r t i c a l migration and feeding rhythm are the same. The contrasting 54 absence of a feeding rhythm i n a non-migratory Great Central S. oregonensis population, and the fact that on some dates some Kennedy Lake S. oregonensis remained at the surface and fed during the day, provides stong evidence that v e r t i c a l migration i s , i n fact, the cause of the feeding rhythm i n S. oregonensis. S i m i l a r l y , Angeli et al. (1995) found that large Daphnia migrated across a v e r t i c a l food gradient and displayed a feeding rhythm, while small Daphnia and mature' Eudiaptomus migrated only within a zone of homogenous food and did not display a feeding rhythm. Gibbons (1992) found that an offshore population of Sagitta serratodentata tasmanica migrating between a food r i c h surface zone and a food poor depth displayed a feeding rhythm, whereas an inshore population of the same species migrating within a homogenous food zone fed continuously. Perhaps, both v e r t i c a l migration and the cessation of daytime feeding are separate, complementary strategies that reduce daytime predation r i s k and co-occur i n environments where predation r i s k i s high. The cessation of daytime feeding may decrease predation r i s k by reducing movement and gut pigmentation that a t t r a c t predators (Bollens and Stearns 1992) . Further observations are required to test t h i s hypothesis. Migrators that feed less during the day may compensate with increased night time food intake i n such a way that t o t a l intake over the diurnal period i s equal i n migrators and non-migrators (Angeli et al. 1995). Higher levels of gut pigment at night i n Kennedy lake r e l a t i v e to both night and day values i n Great Central i n May, June, July and August of 1993 (Figs. 2.24 and Fig. 2.26) suggest that migrating S. oregonensis may compensate 55 for lower daytime feeding during the day by feeding more at night. However, th i s c l e a r l y does not occur i n October (Fig. 2.27). Lower gut fluorescence i n Great Central Lake S. oregonensis may also r e s u l t from the ingestion of non-fluorescing food types or may indicate less available food. Knowledge of the temporal scale of cycles of hunger and sa t i a t i o n i s necessary to evaluate whether spending part of the diurnal period away from food decreases o v e r a l l food intake. Bimodal feeding rhythms with dusk and dawn feeding maxima and low rates of feeding i n the middle of the night are known to occur i n other zooplankton (e.g., Starkweather 1983, Baars and Oosterhuis 1984, Atkinson et al. 1992a, Atkinson et al. 1992b), and t h i s suggests s a t i a t i o n a f t e r dusk feeding may l a s t u n t i l dawn. In Kennedy Lake, however, S. oregonensis gut fluorescence measured near midnight on October 17, 1993 (Fig. 2.27) i s nearly as high as the highest mean value recorded post-dusk on that date. Large differences i n gut fluorescence are also detectable i n samples taken near midnight i n comparison to daytime samples (Fig. 2.24). Furthermore, reported gut evacuation rates for zooplankton are short: 38 minutes for Acartia tonsa (Cervetto et al. 1995), 24 minutes for Calanus acutus (Atkinson et al. 1992b), 2-3 hours for Calanus finmarchicus (Simard et al. 1985), 100 minutes for Metridia spp. and Pleuromamma spp. (Morales et al. 1993), 24.1 minutes for Acartia tonsa, (Durbin et al. 1990). Given a high gut evacuation rate, S. oregonensis l i k e l y need to feed continuously to maintain high gut pigments through the night. This suggests they do not remain satiated during the long day period away from surface food resources. 56 Effect of food availability on S \\ . oregonensis daytime vertical distribution Johnsen and Jakobsen (1987) showed a s i g n i f i c a n t decrease i n the daytime depth of v e r t i c a l l y migrating Daphnia longispina i n response to an experimentally depleted food supply i n enclosures, and argued that t h i s was a r e f l e c t i o n of the optimal strategy to balance a trade-off between feeding and predator avoidance. The mean daytime depth of S. oregonensis i n Kennedy Lake also shows a tendency to decrease as food supply i n the top 10 m decreased. Although these results must be interpreted with caution, as sample size i s low and o u t l i e r s were removed, they do provide some evidence for a feeding opportunity cost associated with v e r t i c a l migration. Fitness consequences of the feeding opportunity cost of vertical migration Direct measures of a feeding opportunity cost as a resu l t of v e r t i c a l migration, expressed as changes i n fi t n e s s (survival, growth, and\/or reproduction), have not been demonstrated for zooplankton. Johnston (1990) showed that migrating kokanee salmon f r y grew more slowly than non-migratory f r y i n experimental enclosures. In chapter 5 I present a model that explores potential effects of v e r t i c a l migration on growth and reproduction of S. oregonensis. Summary The hypothesis that v e r t i c a l migration by S. oregonensis i s driven by stickleback predation i s supported by the comparison of 57 lakes. Alternative hypotheses, including the avoidance of juvenile sockeye predation, foraging e f f i c i e n c y , bioenergetic e f f i c i e n c y , and combined foraging\/bioenergetic e f f i c i e n c y are not consistent with the between lake observations. The stickleback avoidance hypothesis i s also supported by enclosure experiments which indicate that stickleback depredate the same size classes of S. oregonensis that v e r t i c a l l y migrate i n Kennedy Lake. Juvenile sockeye also exert strong predation pressure, but may not drive v e r t i c a l migration because they are crepuscular feeders rather than daytime feeders. V e r t i c a l migration appears to re s u l t i n a feeding opportunity cost. Phytoplankton are less concentrated i n the deep habitat where S. oregonensis reside during the day. Gut pigments are less during the day than at night i n migratory Kennedy Lake S. oregonensis, while gut pigments i n non-migratory S. oregonensis from Great Central Lake do not d i f f e r between day and night. Within Kennedy Lake, S. oregonensis near the surface during the day contain more gut pigments than individuals i n deep water. The decline i n mean depth of S. oregonensis as food abundance decreases also suggests a feeding opportunity cost. 58 Chapter 3 The T iming and ex ten t of v e r t i c a l m i g r a t i o n by Skistodiaptomus oregonensis r e l a t i v e to the t e m p o r a l - s p a t i a l d i s t r i b u t i o n of p r e d a t i o n r i s k I n t r o d u c t i o n Daytime v e r t i c a l migration of zooplankton from the bright, near-surface habitat has been proposed as a strategy to decrease the r i s k from l i g h t - l i m i t e d v i s u a l predators (Mangel and Clarke 1988, Bollens and Frost 1989a, Angeli et al. 1995). The interlake comparison presented i n chapter 2 supports the hypothesis that v e r t i c a l migration by S. oregonensis i n the study lakes i s an adaptation to avoid predation by v i s u a l l y feeding sticklebacks. In t h i s chapter, I test predictions concerning the pattern and timing of diurnal v e r t i c a l migration that are c o r o l l a r i e s of the stickleback avoidance hypothesis. S p e c i f i c a l l y , the l i g h t i n t e n s i t y at the depth where S. oregonensis are located during the day should reduce the feeding rate of sticklebacks r e l a t i v e to t h e i r feeding rate i n surface waters. Secondly, the timing of the v e r t i c a l movements of S. oregonensis at dawn and dusk should keep them at a l i g h t i n t e n s i t y which reduces stickleback feeding rate. Juvenile sockeye are present i n three of the four study lakes and, although the data presented i n chapter 2 do not support the hypothesis that juvenile sockeye predation drives v e r t i c a l migration i n S. oregonensis, Schmidt et al. (1994) suggest that unusual abundance levels of juvenile sockeye i n Alaskan lakes produced a s h i f t i n copepod behaviour from no 59 migration to migration. Furthermore, an enclosure study i n Kennedy Lake (O'Neill and Hyatt 1987) demonstrated that juvenile sockeye and stickleback are both size s e l e c t i v e planktivores that can strongly impact the zooplankton community. Consequently, I also test predictions that are c o r o l l a r i e s of a juvenile sockeye avoidance hypothesis. Namely, that the l i g h t i n t e n s i t y at the depth where S. oregonensis are located during the day results i n a decreased feeding rate by juvenile sockeye ( r e l a t i v e to th e i r feeding rate i n the surface waters), and that the v e r t i c a l movement of S. oregonensis at dawn and dusk are timed to avoid juvenile sockeye predation. Methods Field measures of zooplankton vertical movements at dawn and dusk To track the timing of v e r t i c a l migration by S. oregonensis during dawn and dusk, time series of v e r t i c a l l y s t r a t i f i e d zooplankton tows were collected at 24-12 m, 12-5 m and 5-0 m depth intervals with a closable double-ringed Wisconsin-style 100 um nitex mesh plankton net (292 cm i n length and with a 55 cm diameter opening). The three depth strata were sampled i n rapid succession, from deepest to shallowest, over approximately 8-12 minutes. A 20 minute i n t e r v a l was maintained between the star t of one set of s t r a t i f i e d samples and the s t a r t of the next set. Immediately upon r e t r i e v a l , samples were preserved i n 4% sugared formaldehyde. In the laboratory, a l l S. oregonensis larger than 800 um t o t a l length were counted under a Wild M-5 microscope. The sub-sampling methodology was described i n chapter 2. In each 60 stratum, abundance was estimated as the number of S. oregonensis occurring under a i m surface area (see chapter 2 methods). In October 1993, a number of Neomysis mercedis were captured i n Kennedy Lake and these were ennumerated i n each stratum as we l l . Field measures of vertical distribution of fish at dawn and dusk In the intervals between each set of v e r t i c a l zooplankton hauls described above, hydroacoustic measurements of the depth d i s t r i b u t i o n of f i s h targets were determined with a Furuno FM-22 200 KHz sounder with 100 watts power. The sounder was mounted on a hydroplane lashed into p o s i t i o n at 1 m depth alongside a boat moving at approximately 0.5-1.0 m-s\"1 (see Hyatt et al. 1984 for further d e t a i l s ) . The time varied gain c i r c u i t , which controls for spreading and attenuation losses of the acoustic signal with distance, was inoperative on the sounder. This l i k e l y resulted i n a depth-dependent bias i n the estimated density of targets but the data s t i l l provide a measure of the movement of targets across depth over time. Fish targets on the echosounder traces were counted to determine the mean number of f i s h targets per second i n each stratum during each time i n t e r v a l . Light measurements Light extinction was determined near mid-day on each sample date by measuring l i g h t i n t e n s i t y just below the surface and at 1 m depth intervals with a Li-cor 186A l i g h t meter. Light 61 i n t e n s i t y at depth i s given by: Iz=I0*e-kz (3.1) where Iz=light extinction at depth z, Io=light i n t e n s i t y at the surface, and k=light e x t i n c t i o n c o e f f i c i e n t . This equation was rearranged to: l n l 0 - l n l z = k z (3.2) and a least squares regression was f i t to the f i e l d measurements of Io, Iz, and z to determine the l i g h t e xtinction c o e f f i c i e n t , k. This estimate of k was used to calculate l i g h t i n t e n s i t y depth p r o f i l e s during dawn and dusk sampling from measurements of l i g h t i n t e n s i t y at the surface (I D) . This method assumes that the l i g h t extinction c o e f f i c i e n t does not change as l i g h t i n t e n s i t y changes over the dawn and dusk period. Levy (1989 -his Figure 5-5) demonstrated that k did not change during crepuscular periods i n Cultus Lake and Great Central Lake. Light i n t e n s i t y i s measured i n quantum units of microEinsteins \u2022 second\"1-meter-2 (uE-s-1-m\"2) . One Einstein i s equivalent to one mole of photons. Stickleback feeding rate experiment Kennedy Lake sticklebacks were col l e c t e d i n minnow traps on May 21, 1995, transported to the laboratory, and held i n a 208 1 holding tank illuminated by two 40 watt V i t a - L i t e f u l l spectrum fluorescent l i g h t s placed 40 cm above the tank. An automatic switch turned the l i g h t s off at sunset and on at sunrise each 62 day. Once per day, the stickleback were fed l i v e zooplankton collected from Shirley Lake and\/or Marion Lake i n the UBC Research Forest. Every 2 to 3 days the stickleback were fed thawed Tubifex worms. Shirley Lake zooplankton consisted mainly of Hesperodiaptomus kenai, fourth i n s t a r Chaoborus trivittatus and Aglaodiaptomus leptopus. Marion Lake zooplankton consisted mainly of mature S. oregonensis (779 um mean metasome length), S. oregonensis copepodites (410 um mean metasome length) and a cyclopoid copepod (440 um mean metasome length). Twenty-four hours before the beginning of experimental t r i a l s for each day, 11 f i s h were transferred from the holding tank to a 56 1 starvation tank. The experimental feeding unit consisted of an outer clear plexiglass tank measuring 30 x 58 x 38 cm into which was placed a 27 x 51 x 32 cm inner tank which consisted of clear plexiglass v e r t i c a l walls and a porous 100 um nitex mesh bottom. Lake water f i l t e r e d through a 50 um mesh was added to the experimental feeding unit to a volume of 40 1 i n the inner tank. A piece of non-reflective black a r a l d i t e was placed over the mesh bottom during the experiment to reduce r e f l e c t i o n and to reduce contrast of prey against the bottom background. The exterior surface of the bottom and 3 of the 4 v e r t i c a l walls of the outer tank were covered with brown packing paper over which was placed l i g h t impermeable black p l a s t i c . For observation, one of the larger v e r t i c a l walls of the outer tank was l e f t open, but covered from the outside by a large black clo t h b l i n d attached along the top edge of the open v e r t i c a l wall by velcro\u00ae. The observer, wearing dark clothing, sat motionless between the b l i n d and the open w a l l . This allowed the observer 63 to see into the tank and prevented background l i g h t from entering the tank through the open side w a l l . Each day, for 6 days, 11 in d i v i d u a l f i s h were fed at 11 dif f e r e n t l i g h t i n t e n s i t y treatments. For each t r i a l , Marion Lake zooplankton were added to the experimental feeding unit and b r i e f l y mixed. Following t h i s , a single f i s h was transferred from the starvation tank into the, inner tank of the experimental feeding unit. A sheet of plexiglass was placed over the top of the experimental chamber and, depending on the desired l i g h t i n t e n s i t y , 1 to 11 sheets of grey, p l a s t i c Permascreen l i g h t f i l t e r s were placed on top of t h i s plexiglass sheet. A l l f i s h commenced feeding within 2-10 minutes of placement i n the experimental feeding unit. Once feeding behaviour had begun, the number of feeding s t r i k e s were recorded for 15 minutes. The f i s h was then removed, anaesthetized and k i l l e d i n 95% ethanol, and then fixe d i n 4% formaldehyde. Zooplankton remaining i n the tank were coll e c t e d by removing the a r a l d i t e sheet from the bottom of the inner tank, l i f t i n g the mesh-bottomed inner tank out of the outer tank, and ri n s i n g the contents of the inner tank into a j a r of 4% sugared formaldehyde. R e s u l t s Stickleback feeding rate experiments The t o t a l density of zooplankton did not d i f f e r among l i g h t i n t e n s i t y treatments (Table 3.1), nor did the i n i t i a l densities of each of s i x prey categories (Table 3.2). Within each day, the t o t a l density as well as the proportion of each prey type i n the experimental tanks were approximately equal across treatments Table 3.1: ANOVA results testing HQ that i n i t i a l zooplankton densities i n experimental feeding tanks did not d i f f e r among l i g h t i n t e n s i t y treatments. Source DF Sum of Squares Mean Square F Value P>F Treatment 10 53.096 5.310 0.19 0.9967 Error 55 1577.650 28.685 Corrected Total 65 1630.746 Table 3.2: ANOVA r e s u l t s f o r each prey type t e s t i n g Ho that i n i t i a l d e n s i t i e s of the prey types d i d not d i f f e r among treatments. For prey type Treatment DF=10 and E r r o r DF=55. Prey Type Treatment Mean Square E r r o r Mean Square F P>F cyclopoids 0.930 6.471 0 .14 0 .9989 larg e S. oregonensis 0 .419 3 . 076 0 .14 0 .9991 small S. oregonensis 0.292 9.029 0 .03 1 .0000 Bosmina 1.896 23 .523 0 . 08 0 .9999 larg e cladocerans 0.024 0.068 0 .35 0 .9629 small cladocerans 0.006 0.020 0 .29 0 .9817 66 (values for 2 days shown i n Fi g . 3.1 and 3 .2). There were s i g n i f i c a n t differences among days (Table 3 .3 , and Fi g . 3 . 3 ) . Zooplankton density i n the laboratory experimental tank treatments ranged from 18.8 to 3 6.7 per l i t e r with a mean of 26 .0 . The d a i l y means ranged from 21 .6-31 .8 per l i t e r . These values are on the order of 4-5 times higher than l o c a l Kennedy Lake zooplankton concentrations of 6.2 per l i t e r (May 1995) and 6.4 per l i t e r (October 1993) estimated from the highest densities i n v e r t i c a l l y s t r a t i f e d zooplankton samples. The Marion Lake zooplankton prey used i n the experimental treatments are comparable to the prey available i n Kennedy Lake. A l l zooplankton present i n a sample c o l l e c t e d from Kennedy Lake i n May of 1995 are represented by four groups that include bosminids, cyclopoids, large S. oregonensis (>800 um t o t a l length), and small S. oregonensis (<800 um t o t a l length). Marion Lake contained these same groups i n s i m i l a r proportions as well as only a small number of various other cladocerans (Fig. 3 . 4 ) . Furthermore, the sizes of the prey taxa were very s i m i l a r for a l l prey types (Fig. 3 . 5 ) . In the experiment, the number of prey per stickleback stomach, the number of S. oregonensis per stomach, the number of stri k e s , and the capture\/strike r a t i o a l l increase over a l i g h t i n t e n s i t y range of 0.1 to 1.6 uE-s^-m\"2. After a l i g h t i n t e n s i t y of 1.6 uE-s^-m\"2 i s reached there i s no further l i g h t e f f e c t (Figs. 3 . 6 - 3 . 9 ) . Natural log transformations were performed for each variable (with the exception of the capture\/strike r a t i o , for which the arcsine square-root transformation was used) and then separate least-squares l i n e a r regressions were performed 67 Fig. 3.1: Total zooplankton densities broken down by taxa or size categories in each of 11 treatments for experiment date June 5. 68 25 Treatments cyclopoids i large S. oregonensis 1 1 small S. oregonensis II11II bosminids c ladocerans F ig . 3.2: Total zoop lank ton densi t ies b roken down by taxa or s ize ca tegor ies in e a c h of 11 t reatments for exper iment date J u n e 7. 69 Table 3.3: ANOVA r e s u l t s t e s t i n g Ho that i n i t i a l zooplankton d e n s i t i e s i n experimental feeding tanks d i d not d i f f e r among days. Sum of Mean Source DF Squares Square F Value P>F days 5 1340.017 268.003 55.31 0.0001 e r r o r 60 290.729 4.846 co r r e c t e d t o t a l 65 1630.746 70 June 4 June 5 June 6 June 7 June 10 June 11 Experiment Date Fig. 3.3: Mean zooplankton density across treatments for each experiment date. Vertical bars represent entire range of densities for each day. 71 0.7 F ig . 3.4: A compar i son of the relative density of prey types in one of the exper imenta l tanks (solid bar) and in Kennedy L a k e (open bar) on M a y 20 , 1995. 900 CD N '(fi C 03 CD 800 -700 \" 600 -500 400 H 300 200 -100 -0 ii bosminids cyclopoids large S. oregonensis 1 small S. oregonensis r other cladocerans Fig. 3.5: A comparison of the size of prey types in one of the experimental tanks (solid square) and in Kennedy Lake (open squares) on May 20, 1995. Vertical bars represent standard deviations for each size distribution. to 73 o CO E o co i_ CD CL C\/> 13 E o Q . CO b O i CD E 3 o CO E o CO t CD C L c\/> E o \u00bb Q . CO b \u00a3 is 118 9 8 -78 58 38 18 \u2022 \u2022 n \u2022 : \u2022 \u2022 \u2022 \u2022 i5b \u2022 \u2022 jo! B B \u2022 1 3 i 5 \u2022 \u2022 \u2022 \u2022 B 7 9 11 Light Intensity (uE-s 1-m 13 15 17 Light Intensity (uE-s -m Fig. 3.6: Number of S. oregonensis in s tomachs of st ickleback fed for 15 minutes at 11 light intensities. Top: Light intensity range 0.1-15.5 uE-s \" 1 m \" 2 . Bottom: Enlargement of light intensity range from 0.1-1.6 uE-s~ 1 -m\" 2 (bottom left corner of top figure). 235 195 155 115 115 Light Intensity (uE-s '^m 2 ) Light Intensity (uE-s ^m\" 2) Fig. 3.7: Number of prey in s tomachs of st ickleback fed for 15 minutes at 11 light intensities. Top: Light intensity range 0.1-15.5 uE-s \" 1 - m \" 2 . Bottom: Enlargement of light intensity range from 0.1-1.6 u E s \" 1 - n r 2 (bottom left corner of top figure). 235 205 H 175 U5 115 85 55 25 -5 -1 n P \u2022 \u2022 \u2022 a \u2022 \u2022 \u2022 \u2022 \u2022 B \u2022 \u2022 \u2022 \u2022 \u2022 B \u2022 3 5 7 9 11 13 15 17 Light Intensity (uE-s 1-m ^ Light Intensity (uE-s -m Fig. 3.8: Number of prey strikes taken by st ickleback fed for 15 minutes at 11 light intensities Top: Light intensity range 0.1-15.5 u E s - m \" 2 . Bottom: Enlargement of light intensity range i t it from 0.1-1.6 uE-s\" 1 -m\" 2 (bot tom left corner of top figure). 1.2 1.1 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 -0.1 -1 \u2022 \u2022 \u2022 \u2022 B -r-3 5 7 9 11 Light Intensity (uE-s 1-m 13 \u2022 \u2022 \u2022 15 17 -1 Light Intensity (uE-s -m \u20225 Fig. 3.9: Ratio of captures\/str ikes for st ickleback fed for 15 minutes at 11 light intensities. Top: Light intensity range 0.1-15.5 uE-s\" 1 m \" 2 . Bottom: Enlargement of light intensity range from 0.1-1.6 uE-s\" 1 -m\" 2 (bot tom left corner of top figure). 77 using each variable as a dependent variable against the natural log of l i g h t i n t e n s i t y as an independent variable. These transformations were used to reduce heteroscedasticity. Regressions were performed separately for l i g h t ranges less than and greater than 1.6 uE ' S _ 1*nf 2 (Table 3.4). For a l l four dependent variables, the regressions were s t a t i s t i c a l l y s i g n i f i c a n t for l i g h t i n t e n s i t y less than 1.6 uE-s _ 1 - ir f 2 and non-s i g n i f i c a n t for l i g h t i n t e n s i t y greater than 1.6 uE-s^'m\"2. At a l l l i g h t l e v e l s , large S. oregonensis are selected by most of the stickleback i n these t r i a l s (Fig. 3.10). The rate at which S. oregonensis are eaten i s lowest at the two lowest l i g h t l e v e l treatments used i n t h i s experiment. Thus, the mean number of S. oregonensis per f i s h at the two lowest l i g h t levels was 1.45 (n=ll) compared to 23.42 (n=23) at l i g h t l e v e l s above 1.6 uE. Timing of SL oregonensis v e r t i c a l migration i n r e l a t i o n to stickleback feeding rate Figures 3.11-3.16 plot the v e r t i c a l d i s t r i b u t i o n of S. oregonensis over the periods of rapid l i g h t change (dawn and dusk) i n Kennedy and Paxton lakes r e l a t i v e to the 1.6 uE isolume and the 0.1 uE isolume. Based on my experiments, these l i g h t levels are defined as the maximum (MFRT) and the near-zero feeding rate thresholds (FRT) for stickleback. At dawn on May 20, 1995 (Fig. 3.11) l i g h t levels were below FRT at the surface before sunrise and S. oregonensis were mainly located between 5-12 m. As the FRT isolume s h i f t e d downward into the 5-12 m stratum, S. oregonensis also moved downward, although some S. 78 Table 3 .4 : Regression result s for the effect of l i g h t (X) on four Y variables where ln(Y+l)=a*ln(X)+b, except for the capture\/strike r a t i o where arcsin(sqrt(Y+l))=a*ln(X)+b. A l i g h t i n t e n s i t i e s ranging from 0 1 to 1. 6 uE -1 -? \u2022 s -m : Dependent Variable a b n r 2 F P>F number of prey i n stomach 1.06 3 .54 42 0 .37 23 .98 0 .0001 number Diaptomus i n stomach 0.86 2 . 50 42 0 .40 26.18 0 . 0001 number of strik e s 1.05 4.33 42 0 .33 19.60 0 . 0001 capture\/strike r a t i o 0.12 0.77 - 36 0 .20 8.48 0 . 02 B l i g h t i n t e n s i t i e s ranging from 1 6 uE tc > 15. 5 uE-s'1 \u2022m\" \u20222 . Dependent Variable a b n r 2 F P>F number of prey i n stomach 0.34 3 .44 24 0 .11 2.64 0 .1186 number Diaptomus i n stomach 0.03 2.80 24 0 .00 0.02 0 .9033 number of str i k e s 0.15 4.42 24 0 .04 0.91 0 .3510 capture\/strike r a t i o 0.083 0.078 24 0 .20 1.51 0.50 7 9 11 Light Intensity (uE-s 1-m - 1 17 0.6 0.9 1.2 Light Intensity (uE-s 1-m ^ 1.8 Fig. 3.10: Vanderp loeg-Scav ia index of electivity for S. oregonensis by st ickleback fed for 15 minutes at 11 light intensities. Top: Light intensity range 0.1-15.5 u E s \" 1 m \" 2 . Bottom: Enlargement of light intensity range from 0.1-1.6 uE-s~ 1 -m \" 2 (bottom left corner of top figure). Values above dotted horizontal line represent positive electivity and value below this line represent negative electivity. 0 5.5 11 I I I 3 2 # Diaptomus*10\/ m sunrise 0.0-5.0 5.0-12.0 M F R T CO CD \u2022*\u2014> C 12.0-24.0 Q_ CD Q 3:00 3:40 4:20 5:00 5:40 Local Apparent Time Fig. 3.11: Vertical migration of S. oregonensis relative to stickleback feeding thresholds in Kennedy Lake during dawn on May 20, 1995. Horizontal bar graphs represent total number of S. oregonensis under 1 m square surface in three depth strata. Curves represent depth isolumes for stickleback maximum feeding rate threshold (MFRT) and near-zero stickleback feeding rate threshold (FRT). 0 5.5 11 I I I 3 2 # Diaptomus* 10 An sunset 0.0-5.0 5 .0 -12 .0 CC 0 Q_ CD Q 12 .0 -24 .0 18 :40 Local Apparent Time Fig. 3.12: vertical migration of S. oregonensis relative to stickleback feeding thresholds in Kennedy Lake during dusk on May 20, 1995. Detai ls as in F ig . 3.11. 2:40 3:20 4:00 4:40 5:00 Local Apparent Time Fig. 3.13: Vertical migration of S. oregonensis relative to stickleback feeding thresholds in Kennedy Lake during dawn on June 24, 1994. Details as in Fig. 3.11. I I I I I I I I I 18:55 19;35 20:15 20:55 21:35 Local Apparent Time Fig. 3.14: Vertical migration of S. oregonensis relative to stickleback feeding thresholds in Kennedy Lake during dusk on June 24, 1994. Details as in F ig . 3.11. 0 15 30 I I I 3 2 # Diaptomus*10\/ m 0.0-4.0 CO Z CD 4.0-9.0 Q . CD Q 9.0-12.0 sunset 18:20 MFRT 19:00 19:40 Local Apparent Time 20:20 21:00 Fig. 3.15: Vertical migration of S. oregonensis relative to stickleback feeding thresholds in Paxton Lake during dusk on August 19, 1994. Details as in Fig. 3.11. Fig. 3.16: Vertical migration of S. oregonensis relative to stickleback feeding thresholds in Paxton Lake during dawn on August 1 9 , 1 9 9 4 . Detai ls as in F ig . 3.11. oo 86 oregonensis temporarily remained i n the 5-12 m stratum. The FRT isolume l e v e l l e d o f f at about 17 m during the day and very few S. oregonensis remained above 12 m. Because of the sampling depth i n t e r v a l , t h e i r d i s t r i b u t i o n w i t h i n the 12-24 m stratum cannot be defined. I t i s c l e a r , however, that most i n d i v i d u a l s are below the MFRT isolume. At dusk on the same date ( F i g . 3.12), almost a l l S. oregonensis remained below 12 m u n t i l the FRT isolume rose above t h i s depth and, again, most i n d i v i d u a l s remained below MFRT throughout the sample p e r i o d . On June 24, 1994 the p a t t e r n was s i m i l a r . Before f i r s t l i g h t , l i g h t l e v e l s i n the lake remained below FRT and S. oregonensis were l o c a t e d near the surface ( F i g . 3.13). As the FRT isolume moved downward at dawn, S. oregonensis moved downward out of the 5-12 m stratum. Almost a l l i n d i v i d u a l s reached the 12-24 m stratum before the FRT isolume. At dusk ( F i g . 3.14), most i n d i v i d u a l s remained below the MFRT isolume at a l l times and migr a t i o n i n t o the 5-12 m stratum corresponded c l o s e l y to the time of movement of the FRT through t h i s stratum. I f we assume s i m i l a r v i s u a l feeding thresholds apply to Paxton Lake s t i c k l e b a c k s , v e r t i c a l m i g r a t i o n i n Paxton Lake S. oregonensis a l s o appears to be timed to remain below the MFRT, and the data are c o n s i s t e n t w i t h the n o t i o n that most i n d i v i d u a l s remain below the FRT isolume. Since the maximum depth of Paxton Lake i s l e s s than 15 m, i t was not p o s s i b l e to sample below 12 m. During d a y l i g h t hours, the FRT isolume was l o c a t e d at or j u s t below 12 m. Thus, most S. oregonensis were below 12 m and i n a c c e s s i b l e to the net samples. This e x p l a i n s both the low'' numbers of S. oregonensis above 12 m at 18:29 and the sudden 87 increase i n numbers i n the 9-12 m stratum at 18:49 just as the FRT isolume r i s e s above 12 m (Fig. 3.15), and the decrease i n S. oregonensis caught a f t e r dawn as the FRT isolume s h i f t s to below the 9-12 m stratum (Fig. 3.16). The daytime depth d i s t r i b u t i o n of S. oregonensis i n Kennedy Lake approximates the expectation of the stickleback avoidance hypothesis that they are at depths where l i g h t i n t e n s i t i e s reduce stickleback feeding rate. Discrete samples at 1, 3, 5, 7, 10, 17, and 24 m in t e r v a l s indicate few S. oregonensis at 10 m or above, and peak abundances at 17 or 24 m (chapter 2). The v e r t i c a l l y s t r a t i f i e d samples indicate S. oregonensis are below the 12 m depth as dusk begins and dawn ends. In Kennedy Lake, during f u l l sunlight (2000 uE-s^-nT2) the MFRT l i g h t i n t e n s i t y i s estimated to be at 16.9 m and FRT l i g h t i n t e n s i t y at 23.8 m (using the extinction c o e f f i c i e n t of 0.401 measured i n June of 1994). Under cloudy conditions (1000 uE-s^-m-2) , MFRT i s estimated to be at 15.2 m and FRT at 22.1 m. Timing of \u00a3\\_ oregonensis vertical migration relative to sockeye vertical migrations With the exception of g l a c i a l l y turbid lakes (Hyatt et al. 1989), information obtained from diurnal hydroacoustic sampling of f i s h target depth d i s t r i b u t i o n , analyses of gut contents of juvenile sockeye sampled by open water trawl, and v i s u a l observations of surface feeding (Narver 1970, Barraclough and Robinson 1972, McDonald 1973, Levy 1990) indicate that juvenile sockeye t y p i c a l l y remain i n deep water during the day and that, at this depth, t h e i r feeding rate i s low. Ascent toward the 88 surface occurs at dusk and feeding may occur as they ascend (Barraclough and Robinson 1972). At the surface, active feeding continues for a period before the t w i l i g h t fades. The juvenile sockeye then migrate down to the thermocline where they spend the night. Dense schooling, and the r a r i t y of fresh food i n gut samples at night, indicate that the feeding rate i s low. As l i g h t i n t e n s i t y increases at dawn, sockeye may approach the surface and feed again before descending to deep water for the daylight period. The timing of migration and feeding i s consistent r e l a t i v e to sunrise and sunset (Fig. 3.17). In 11 of 13 instances, the dusk ascent begins within 3 0 minutes of sunset, and by about 15-2 0 minutes after sunset the juvenile sockeye a r r i v e near the surface. They feed here for about 20-40 minutes before again descending to near the thermocline. Less data are available for the dawn period and the r i s e from the thermocline to the surface for feeding i s harder to define than from the hypolimnion at dusk. Some juvenile sockeye do not feed at dawn. This suggests that at the time that some sockeye r i s e to feed at the surface, other sockeye descend to day time depths. This disperses the f i s h targets i n echo sounder graphs and obscures the timing of these events. Nonetheless, the available information indicates that feeding s t a r t s and ends before sunrise (Fig. 3.18). Depth isolumes i n Kennedy Lake for maximum, half maximum, and minimum feeding rate l i g h t i n t e n s i t i e s of juvenile sockeye over dawn and dusk periods (Fig. 3.19) indicate that juvenile sockeye have the v i s u a l capacity to feed near the surface up to an hour after sunset and an hour before N1 CO CD TJ C\/) X ! t -X \" r X x- -x-N2 N3 N4 N5 ? B1 - - h M1 \u2022150 \u2022120 I -90 -60 Minutes from sunset Fig . 3.17: Summary of available data on time of the beginning of movement of juvenile sockeye toward the surface (X), the period of migration toward the surface (dotted lines) and the period of feeding near the surface (solid line) relative to sunset. Uncertainty about t ime that feeding begins or ends is represented by question marks. N1-N5 from Narver (1970) at Babine Lake, B1 from Barraclough (1972) in Great Central Lake , M1 from McDona ld (1973) in Babine Lake, L1-L3 from Levy (1989) in Shuswap Lake, Babine Lake, and Quesne l Lake , and G 1 - G 3 my data from Kennedy Lake. N1 N5 X CO CD T 3 Z5 \u20224\u2014' CO N4 B1 M1 L1 12 G1 G2 -165 -140 -115 T T T i \u2014 h - r -90 -65 -40 Minutes from sunrise 1 \u202215 sunrise 10 35 Fig. 3.18: Summary of available data for juvenile sockeye on the time of juvenile sockeye feeding near the surface (solid line) and the time of the beginning of movement toward depth (X) relative to sunrise. Other details as in Fig. 3.17. 91 -60 -45 -31 -16 sunrise 12 27 Minutes from sunrise -10 sunset 10 20 30 40 50 60 70 80 Minutes from sunset Fig . 3.19: Max imum (1.0 f t . -c) , half max imum (0.01 f t . -c) , and min imum (0.0001 ft.-c.) feeding rate i so lumes for juveni le sockeye (from AN 1959) for c l ea r sky condi t ions during dawn (top panel) and dusk (bottom panel) on J u n e 2 3 , 1 9 9 4 . Ca lcu la t ions b a s e d on average c lea r sky instantaneous sur face light intensities on J u n e 2 3 at 50 deg rees Lat. (United S ta tes Navy B u r e a u of* Sh ips 1952) and an est imated light extinction coeff icient of 0 .401 . 92 sunrise. Figs. 3.20-3.24 show timing of v e r t i c a l movements of S. oregonensis and f i s h at dawn and dusk i n Kennedy Lake. The interpretation of v e r t i c a l movement of juvenile sockeye by hydroacoustics i s complicated i n Kennedy Lake by the presence of large numbers of stickleback. Stickleback do not v e r t i c a l l y migrate and generally remain at the surface regardless of l i g h t l e v e l . Thus stickleback account for most of the targets near the surface during the day (Kim Hyatt, FOC, unpublished data). Therefore, sockeye movements can be inferred from the movement of f i s h targets from deeper s t r a t a toward the surface at dusk and from the surface stratum downward at dawn. At dusk on October 19, 1993 (Fig. 3.20), June 24, 1994 (Fig. 3.21), and May 19, 1995 (Fig. 3.22) some deep water f i s h targets, assumed to be juvenile sockeye, rose to < 12 m within 10-15 minutes after sunset. S. oregonensis arrived i n the surface stratum shortly before juvenile sockeye on October 19, 1993 (Fig. 3.20) and June 1994 (Fig. 3.21). In May, S. oregonensis rose to the surface shortly a f t e r dusk and at about the same time as the juvenile sockeye (Fig. 3.22). Thus, S. oregonensis were vulnerable to post-sunset surface feeding by juvenile sockeye on a l l of these dates. I f sockeye feed as they ascend, as reported for Great Central Lake sockeye (Barraclough and Robinson 1972), then on a l l three sample dates S. oregonensis were also exposed to sockeye predation from 10 minutes before sunset to 15 minutes aft e r sunset. This was the time period when both S. oregonensis and juvenile sockeye were moving upward and present i n the 5-12 meter stratum. CO CD c Q . CD Q CO CD CL CD Q 0.0-5.0 5.0-8.5 8.5-12.0 12.0-15.5 15.5-19.0 19.0-22.5 22.5-26.0 26.0+ 0.0-5.0 5.0-12.0 12.0-24.0 0 7 14 I I I Fish targels\/trin b i sunset \u2022 F P 16:20 17:00 \u2014n 17:40 18:20 1 I 19:00 0 18.5 37 I I 1 3 2 # Diaptomi\u00a3*107m i i r r r r 16:20 17:00 17:40 I 18:20 19:00 Local Apparent Time Fig. 3.20: Changes in vertical distribution of fish targets (top) and S. oregonensis (bottom) in Kennedy Lake at dusk on October 19, 1993. Upper and lower dashed lines represent approximate timing of vertical movement of juvenile sockeye targets. 94 0 6 12 I I I Fish taigets\/min 0.0-5.0 5.0-8.5 8.5-12.0 \"CO ^ 12.0-15.5 CD -*\u20141 \u2014 15.5-19.0 SZ gj\" 19.0-22.5 Q 22.5-26.0 26.0+ 0.0-5.0 5.0-12.0 CC CD -*\u2014< C 12.0-24.0 CL CD Q sunset 18:55 I 19:35 I 20:15 0 13 26 1 I I 3 2 # Diaptomus'10\/ m \u2022 I I 18:55 I I i I 19:35 20:15 Local Apparent Time I I I i 20:55 I r 20:55 I 21:35 ILL! f r I 21:35 Fig. 3.21: Changes in vertical distribution of fish targets (top) and S. oregonensis (bottom) in Kennedy Lake at dusk on June 24, 1994. Other details as in F ig. 3.20. 95 0 3 6 I I I Fish targets\/min sunset 0.0-5.0 5.0-8.5 J ^ , 8.5-12.0 \"CO ^ 12.0-15.5 CD 15.5-19.0 19.0-22.5 22.5-26.0 26.0+ 18:40 19:20 Ti 1\u2014 20:00 1 20:40 1 21:20 0 5.5 11 1 \\ \\ # DiaptomiB*107 m 0.0-5.0 5.0-12.0 12.0-24.0 I 18:40 I I I I I | I I I 19:20 I 20:00 I I 20:40 I 21:20 Local Apparent Time Fig . 3.22: Changes in vertical distribution of fish targets (top) and S. oregonensis (bottom) in Kennedy Lake at dusk on May 19, 1995. Other details as in F ig . 3.20. 96 0.0-5.0 5.0-8.5 8.5-12.0 12.0-15.5 15.5-19.0 19.0-22.5 22.5-26.0 26.0+ I I I Fish targets\/rrin sunrise 1 2:40 3:20 4:00 4:40 \"I 1\u2014 5:00 0.0-5.0 5.0-12.0 12.0-24.0 0 13 26 I I I 3 2 # Diaptomus*10\/ m 2:40 I 3:20 4:00 4:40 i i ~ 5:00 Local Apparent Time Fig. 3.23: Changes in vertical distribution of fish targets (top) and S. oregonensis (bottom) in Kennedy Lake at dawn on June 24, 1994. Other details as in F ig . 3.20. 97 0 3 6 , sunrise Fish targets\/rrin I 1 1 1 1\u2014i 1 1 3:00 3:40 4:20 5:00 0 5.5 11 i i i 3 2 # Diaptomus*lO\/ m ; 0.0-5.0 ; -g- 5.0-12.0 1 \u2022 \u2022 \u2022 \u2022 ! 1 1 i terval i 12.0-24.0 C L CD Q 1 i 1 1 1 1 - -\u2022 ; - -i i i i i i i i i 3:00 3:40 4:20 5:00 Local Apparent Time Fig . 3.24: Changes in vertical distribution of fish targets (top) and S. oregonensis (bottom) in Kennedy Lake at dawn on May 19, 1995. Other details as in F ig. 3.20. 98 At dawn on June 24, 1994 (Fig. 3.23) and May 19, 1995 (Fig. 3.24), juvenile sockeye began descending within 20 minutes of sunrise. S. oregonensis began to s h i f t downward from the 5-12 to 12-24 m depth i n t e r v a l at about the same time, or l a t e r , and therefore a f t e r the juvenile sockeye feeding period. Timing of \u00a3\\. oregonensis vertical migration relative to N. mercedis vertical migrations N. mercedis, an invertebrate zooplankton predator, also occurs i n Kennedy Lake (Cooper et al. 1992) . N. mercedis i s also a strong v e r t i c a l migrator, and i t was captured i n abundance during the dusk sampling of October 1993. On th i s date, S. oregonensis ascended to the 0-5 m stratum about 20-40 minutes e a r l i e r than N. mercedis (Fig. 3.25). Since S. oregonensis move upward well before the N. mercedis begin feeding near the surface, the timing of S. oregonensis movements has no s i g n i f i c a n t effect on the amount of time they are vulnerable to the N. mercedis. D i s c u s s i o n Timing of S\\_ oregonensis v e r t i c a l migration i n r e l a t i o n to stickleback feeding rate As predicted by the stickleback avoidance hypothesis, as S. oregonensis move up at dusk and down at dawn they cl o s e l y track the threshold l i g h t i n t e n s i t y for stickleback feeding. I f there i s a daytime feeding opportunity cost associated with v e r t i c a l migration, the optimal strategy may be to tolerate predation r i s k greater than 0. This may explain the p o s i t i o n of some S. 99 0 40 80 I I I #NeomySB \/ m sunset 0.0-5.0 5.0-12.0 12.0-24.0 16:20 17:00 0 18.5 37 I I I 3 2 # Diaptomus*! 0 \/ m 0.0-5.0 5.0-12.0 12.0-24.0 16:20 \u2014 r ~ 17:00 [ rrc I I i 17:40 1 I 18:20 1 1 19:00 i i r r r r I I I I I l 17:40 I 18:20 1 I 19:00 Local Apparent Time Fig . 3.25: Changes in vertical distribution of N. mercedies (top) and S. oregonensis (bottom) at dusk on October 20, 1993. 100. oregonensis at l i g h t i n t e n s i t i e s above FRT but below MFRT. The posi t i o n of some S. oregonensis at l i g h t i n t e n s i t i e s above FRT also may be an a r t i f a c t of underestimating FRT because of the higher prey densities i n the feeding experiments than i n Kennedy Lake. These higher densities elevate encounter rates and, thus, may cause feeding to be p r o f i t a b l e at lower l i g h t i n t e n s i t y than i t i s i n the lake. In turn, p o t e n t i a l l y , t h i s underestimates feeding rate l i g h t thresholds and the extent to which S. oregonensis avoid threshold l i g h t levels for stickleback feeding. While FRT may be underestimated somewhat due to high experimental prey densities, the estimate of MFRT obtained from the laboratory experiments i s l i k e l y to be accurate because prey densities were low enough that non-zero time i n t e r v a l s e x i s t between prey encounters. Therefore, trends of feeding rate as a function of l i g h t i n t e n s i t y for each prey density should be i d e n t i c a l i n shape and d i f f e r only i n elevation. Above a threshold prey density, one or more prey are always within view and feeding rates are determined only by light-independent pursuit, capture, and handling times. Below t h i s threshold, feeding rate i s affected by the average time increment between encounters which are determined by l i g h t dependent reactive distances (Ware 1973, Aksnes and Giske 1993). Under these conditions the reactive distance for each prey item, as a function of l i g h t i n t e n s i t y , w i l l be independent of prey density and r e l a t i v e rates of feeding at di f f e r e n t l i g h t i n t e n s i t y w i l l , therefore, be independent of prey density as w e l l . D i s t i n c t time i n t e r v a l s between prey encounters were observed i n the lab feeding experiments. This can be confirmed 101 with simple time a l l o c a t i o n calculations. With one exception, a l l f i s h i n the experiment pursued less than 180 prey over a 15 minute i n t e r v a l (Fig. 3.8), and a l l prey encountered appeared to be pursued. I conservatively estimate a mean combined pursuit, capture, and handling time per prey item of 3 seconds. Even at 180 encounters, t h i s leaves an average 2 second time i n t e r v a l between encounters. Timing of oregonensis v e r t i c a l migration r e l a t i v e to sockeye v e r t i c a l migrations Clearly, S. oregonensis do not decrease daytime r i s k from sockeye predation by migration: sockeye are crepuscular feeders and S. oregonensis are at the surface at dawn and dusk when juvenile sockeye are feeding. In fact, v e r t i c a l migration may increase r i s k from juvenile sockeye predation since descent of S. oregonensis to deep water during the day brings them closer to juvenile sockeye that also migrate downward. In Babine Lake, freshly eaten individuals of a v e r t i c a l l y migrating copepod, Heterocope septentrional is, were found i n the stomachs of juvenile sockeye that occupied deep water during the day (Narver 1970) . The conclusion that v e r t i c a l migration by S. oregonensis i s neither driven by, nor results i n , juvenile sockeye avoidance contradicts the conclusions of Schmidt et al. (1994) for turbid Alaskan Lakes. Possibly, the turbid conditions i n these lakes force juvenile sockeye to feed i n daylight near the surface, rather than crepuscularly.and, thus, provide a daytime selective pressure for downward migration of zooplankton s i m i l a r to that 102 provided by stickleback i n the lakes we studied. Hydroacoustic and trawl surveys indicate that juvenile sockeye remain near the surface throughout the d i e l period i n turbid Owikeno Lake (Hyatt et al. 1989). From hydroacoustic data, i t has been inferred that juvenile sockeye i n Nimpkish Lake undertake a nocturnal v e r t i c a l migration rather than a t y p i c a l diurnal migration and that daytime predation threat from these sockeye drove the observed diurnal v e r t i c a l migration of zooplankton i n that lake (Levy 1990) . However, pelagic stickleback are.also present i n Nimpkish Lake and these may account for the zooplankton v e r t i c a l migration. The sticklebacks may also account for the large number of surface f i s h targets i n daytime hydroacoustic data that were interpreted as juvenile sockeye. S. oregonensis vertical migration and separate juvenile sockeye and stickleback resource niches Stickleback and juvenile sockeye both u t i l i z e zooplankton resources that appear to be l i m i t e d i n many coastal lakes (Hyatt and Stockner 1985, O'Neill and Hyatt 1987) and are the only two common planktivores i n many of these lakes (Hyatt and Stockner 1985). In lake enclosure experiments i n Kennedy Lake the diets of both f i s h species were s i m i l a r . Both preferred larger prey (including S. oregonensis) and strongly impacted the size d i s t r i b u t i o n and species composition of the zooplankton community (O'Neill and Hyatt 1987). In the lake, however, juvenile sockeye are able to exploit v e r t i c a l l y migrating S. oregonensis, but stickleback are not. 103 Chapter 4 Vert ica l migration behavior of Skistodiaptomus oregonensis: constitutive or induced? Introduction I presented evidence i n chapters 2 and 3 to argue that interlake differences i n the v e r t i c a l migration behaviour by S. oregonensis are associated with the presence, or absence, of pelagic sticklebacks. In th i s chapter I examine whether the migration phenotypes of individuals are fixed genetic (constitutive) characters or environmentally induced ( f l e x i b l e ) behaviours. Many recent publications demonstrate that v e r t i c a l migration i n ind i v i d u a l planktors can be induced by environmental cues. In most cases, exudates released by predators provide t h i s cue (Dodson 1988, Dawidowicz et aT. 1990, N e i l l 1990, Tjossem 1990, Bollens and Frost 1991, Ringelberg 1991a, Ringelberg 1991b, N e i l l 1992, De Meester 1993, Forward and Rittschof 1993, Loose and Dawidowicz 1994), although i n some cases di r e c t mechanical or vi s u a l cues from the predator are necessary (Bollens et al. 1994). Also, genotype differences for d i f f e r e n t migration or v e r t i c a l d i s t r i b u t i o n phenotypes within populations of zooplankton are well known (Weider 1984, Weider 1985, De Meester and Dumont 1988). Interaction between genotype and induced effects on migration has been documented for. g e n e t i c a l l y d i s t i n c t clonal groups of Daphnia magna (De Meester 1993) and for di f f e r e n t species of Daphnia c o l l e c t e d from d i f f e r e n t lakes (Dodson 1990) . Here, I present laboratory experiments designed to d i r e c t l y 104 test the hypothesis that f i s h presence or absence affects the v e r t i c a l d i s t r i b u t i o n of indi v i d u a l S. oregonensis. F i e l d evidence for v a r i a t i o n i n migration behaviour c o l l e c t e d over three years of f i e l d sampling i s also investigated. Individuals from both non-migratory (Great Central Lake) and migratory (Kennedy Lake) populations are considered. Methods Laboratory experiments Four c y l i n d r i c a l plexiglass columns 195 cm t a l l and 15 cm i n diameter were used as microcosms to investigate the effect of stickleback presence or absence on the v e r t i c a l migration behaviour of S. oregonensis. Each column was placed within a chamber 245 cm i n height, 150 cm i n length, and 120 cm i n width. The c e i l i n g and walls of each chamber consisted of l i g h t -impermeable black p l a s t i c . Twenty-five cm diameter photographer lamps f i t t e d with standard incandescent 25 Watt l i g h t bulbs were suspended 22 cm d i r e c t l y above each tube. The lamps were also f i t t e d with blue Westsun t h e a t r i c a l l i g h t f i l t e r s to remove a large portion of the red l i g h t which i s normally absorbed within the f i r s t few meters i n lakes (Wetzel 1975) . The l i g h t i n t e n s i t y p r o f i l e within the tubes i s shown i n (Fig. 4.1). An automated l i g h t switch turned the l i g h t s on at dawn and off at dusk each day. S.\u2022 oregonensis were captured from the lake with v e r t i c a l tows at depths from 3 0 m to the surface using a double-ringed Wisconsin-style 100 um nitex mesh plankton net 292 cm i n length with a 55 cm diameter opening. Zooplankton were transported i n 105 0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Light intensity (uE) Fig. 4.1: Light intensity depth profile in experimental tubes (primary axis) and corresponding depth of similar light intensities in Kennedy Lake at surface light intensity of 1000 uE and k=0.401 (secondary axis). 106 25 1 carboys to the laboratory where they were maintained under a diurnal l i g h t rhythm. Light was switched on at sunrise and turned off at sunset by an automated timer. Within 48 hours of capture, equal numbers of S. oregonensis were placed i n each experimental tube. For a period of up to eight days and nights, the number of individuals i n each of 11 depth i n t e r v a l s was recorded (once near noon and once near midnight). At night the tubes were surveyed using a narrow-beam f l a s h l i g h t f i t t e d with a f i l t e r that allowed only far red l i g h t to pass. Diaptomus are reported to be insen s i t i v e to far red l i g h t (Ramcharan and Sprules 1989). Evasive hopping behaviour, on i n i t i a l exposure to the l i g h t beam, occurred less than 5% of the time. Three separate experiments were performed. In experiment 1 Kennedy Lake S. oregonensis were placed into tubes f i l l e d with Kennedy Lake water. For f i s h water addition treatments (fishwater [+]) two Kennedy Lake sticklebacks, previously fed Kennedy Lake zooplankton, were placed for 24 hours i n a dish containing 500 ml of Kennedy Lake water. The sticklebacks were then removed and the water added to the two fishwater [+] treatment tubes. I added equal amounts of Kennedy Lake water that had not held sticklebacks to the two f i s h water absent (fishwater [-]) treatment tubes. Experiment 2 was conducted using S. oregonensis from Great Central Lake i n tubes f i l l e d with Great Central Lake water. Fishwater [+\/-] treatments were prepared as i n experiment 1, except that Great Central Lake zooplankton were fed to the sticklebacks p r i o r to the experiment. Experiment 3 was conducted i n Kennedy Lake water with S. 107 oregonensis from Kennedy Lake i n Kennedy Lake water. In these f i s h [+] treatments two sticklebacks were placed into a nitex mesh cage and suspended i n the top 10 cm of water i n the tube. The mesh was flush with the i n t e r i o r walls of the tube and f l a t along the bottom. Similar cages, but without stickleback, were placed i n the f i s h [-] treatments. S t a t i s t i c a l analyses were performed using the SAS System software ( L i t t e l et al. 1991). S t a t i s t i c a l tests on proportional data were performed for l o g i t transformations of the proportion values (p'=ln[p\/[1-p]], where p=proportion). Univariate repeated measures analyses of variance for the proportion of individuals below 100 cm during the day, and for the change i n the proportion below t h i s depth from night to day, were performed for the dependent time series measurements i n each tube. This procedure performs an analysis of variance based on a single average measure for each time series ( L i t t e l et al. 1991). Because the low number of replicates (two tubes per treatment) yielded i n s u f f i c i e n t error degrees of freedom, i t was not possible to use multivariate repeated measures analyses of variance to test for a time eff e c t . For the univariate repeated measures analyses, the low number of replicates (only 1 error degree of freedom) l i m i t the power of the test to detect treatment a f f e c t s . To increase error degrees of freedom and the r e s u l t i n g s t a t i s t i c a l power, I also performed mixed-model nested analyses of variance for the depths of indi v i d u a l S. oregonensis within each sample time. These analyses increased error degrees of freedom and allowed for tests of both fixed treatment effects ( f i s h or fishwater [+\/-]) and random tube within treatment ef f e c t s . 108 Field collections Methods for determining zooplankton depth d i s t r i b u t i o n from horizontal tows and for extraction of chlorophyll a from the 3-50 um phytoplankton size f r a c t i o n as a measure of the food available to S. oregonensis are described i n Chapter 2. Results Experiment 1 S. oregonensis collected from the migratory population i n Kennedy Lake (see chapter 2) also migrated i n the experimental tubes. Typical day and night d i s t r i b u t i o n s i n a l l four tubes are shown i n Figure 4.2. At night S. oregonensis i n a l l four tubes were d i s t r i b u t e d r e l a t i v e l y evenly across depths but, during the day, a large proportion of S. oregonensis were concentrated near the bottom. These d i s t r i b u t i o n s are comparable to t y p i c a l day and night v e r t i c a l d i s t r i b u t i o n s of S. oregonensis i n Kennedy Lake (Fig. 2.3 and Fig. 2.4). Although a large proportion of S. oregonensis remained below 100 cm i n the columns on a l l days (Fig. 4.3), the proportion of individuals below 100 cm increased from night to day on nearly a l l dates (Fig. 4.4). This suggests a net downward movement of individuals during the day. Neither the daytime v e r t i c a l d i s t r i b u t i o n nor the change i n d i s t r i b u t i o n from day to night were affected by fishwater [+\/-] treatments. Univariate repeated measures analysis of variance indicates no s i g n i f i c a n t fishwater [+\/-] treatment effect for either the proportion of individuals below 100 cm during the day or the change i n the proportion below t h i s depth from night to day (Table 4.1). Mixed model nested analyses of variance for the 109 E o Q . CD a 0-10 10-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 Night no fish 15 - I \u2014 -10 fish 0 To\" E o CL CD Q 0-10 10-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 Day no fish -1\u2014 -10 ZZJ tube 1 tube 2 IWW1 fish mm,\u201e -5 0 5 Number of individuals 10 tube 3 tube 4 Fig. 4.2: Experiment 1 - typical day and night distributions of Kennedy Lake S. oregonensis in cycl inders in f ishwater [ +\/-] treatments. 110 0.9 CL 0.3 - * 1 1 0.2 J , , , , 18-Jul 20-Jul 22-Jul 24-Jul 26-Jul 28-Jul Fig. 4.3: Proportion of Kennedy Lake S. oregonensis below 100 cm during the day in experiment 1. Solid squares indicate fishwater [-] treatments and open squares indicate fishwater [+] treatments. Lines join values from the same cyclinder through time. I l l Fig. 4.4: Experiment 1 - change in the proportion of Kennedy Lake S. oregonensis below 100 cm from night to day in fishwater [+\/-] treatments. Zero line indicates no change. Other details as in Fig. 4.3. 112 Table 4.1: Experiment 1 - univariate repeated measures analyses of variance for A) the effe c t of fishwater [+\/-] treatments on the proportion of S. oregonensis below 100 cm during the day and B) the effect of fishwater [+\/-] treatments on the change i n the proportion of S. oregonensis below 100 cm from night to day. Analyses were performed using the l o g i t transformation of the proportion data. A Source DF Type I I I SS Mean Square F Value Pr > F TREAT 1 0.63984722 0.63984722 0.97 0 .4281 Error 2 1.31666008 0*65833004 B Source DF Type I I I SS Mean Square F Value Pr > F TREAT 1 0.00026310 0.00026310 0.00 0 .9876 Error 2 1.70411574 0.85205787 113 depths of individuals, performed separately for each sample time, also indicate neither treatment effects nor effects of tubes nested within treatments for any day or night during the course of the experiment (Table 4.2) Experiment 2 S. oregonensis from the non-migratory Great Central Lake population (see chapter 2) did not migrate as strongly i n the experimental tubes as the Kennedy Lake S. oregonensis i n experiment 1. Typical day and night d i s t r i b u t i o n s i n a l l four tubes are shown i n Figure 4.5 (compare Fig. 4.2). r Compared to Kennedy Lake S. oregonensis (experiment 1), both the change i n the proportion below 100 cm (Fig. 4.6) as well as the proportion below 100 cm during the daytime (Fig. 4.7) i s , on average, much smaller and the change i n proportion below 100 cm from day to night i s as often negative as p o s i t i v e . The fishwater [+] treatment appears to increase the proportion of S. oregonensis below 100 cm during the day (Fig. 4.7), although a univariate repeated measures analysis of variance detects no s i g n i f i c a n t difference (Table 4.3). Despite s the apparent increase i n the proportion of S. oregonensis i n deep strata during the day i n fishwater [+] treatments, there i s no clear effect of fishwater [+\/-] treatments on the change i n v e r t i c a l d i s t r i b u t i o n from day to night (Fig. 4.6 and Table 4.3). Nested analyses of variance for depths of in d i v i d u a l S. oregonensis show no s i g n i f i c a n t treatment effects except on the second night and no s i g n i f i c a n t tube nested within treatment effects (Table 4.4). Table 4.2: Experiment 1 - nested analyses of variance for depths of i n d i v i d u a l S. oregonensis: tests of significance for fishwater [+\/-] treatment effects and tube within treatment effects. For a l l times, treatment df=l and tube within treatment df=2. tube treatment within effect F t r e a t m e n t P>F t r e a t m e n t treatment F t u b e P>Ftube time error df error df day 1 2 12 0 212 0 6878 31 0 466 0 6319 day 2 2 00 0 267 0 6565 33 0 278 0 7593 day 3 2 06 0 115 0 7659 35 0 894 0 4181 day 4 2 13 0 475 0 5582 33 0 326 0 7240 day 5 2 07 1 022 0 4151 33 0 247 0 7826 day 6 2 10 0 068 0 8179 33 0 653 0 5271 day 7 2 02 0 352 0 6127 32 1 329 0 2791 day 8 2 07 0 179 0 7118 33 1 659 0 2059 night 1 2 08 1 465 0 3458 34 0 368 0 6946 night 2 2 03 0 416 0 5840 33 0 905 0 4144 night 3 3 31 8 141 0 0578 33 0 011 0 9890 night 4 2 09 1 452 0 3470 34 0 588 0 5612 night 5 2 22 2 081 0 2740 33 0 400 0 6736 night 6 2 00 0 061 0 8282 35 1 286 0 2891 night 7 2 14 3 746 0 1841 34 0 789 0 4623 115 0-10 10-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 Night no fish fish \u20221 0 5 5 A 3 0 2 4 6 8 10 0-10 10-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 -10 Day W W W W j i p \" \" ' \" \" \" \" tube 1 \" \" \" \" tubegR-^sl [HH tube 3 \u2022ttt(\/f\/\/\/t\/\\ V777A tube 4 no fish fish -4 -2 0 2 4 Number of individuals 6 8 10 Fig . 4.5: Experiment 2 - typical day and night distributions of Great Central Lake S. oregonensis in cycl inders in fishwater [+\/-] treatments. 116 0.4 30-Aug 31-Aug 01-Sep 02-Sep 03-Sep F ig . 4 .6 : Exper iment 2- c h a n g e in the proport ion of Grea t Cen t ra l L a k e S . oregonensis be low 100 c m f rom night to day in f ishwater [+\/-] t reatments. O the r detai ls a s in F ig . 4.3 117 0.35 0-1 , . . 1 30-Aug 31-Aug 01-Sep 02-Sep 03-Sep Fig. 4 .7 : Exper iment 2 - proport ion of Grea t Cent ra l L a k e S . oregonensis be low 100 c m during day. Other detai ls a s in F ig . 4 .3 . 118 Table 4.3: Experiment 2 - univariate repeated measures analyses of variance for A) the effect of fishwater [+\/-] treatments on the proportion of S. oregonensis below 100 cm during the day and B) the effect of fishwater [+\/-] treatments on the change i n the proportion of S. oregonensis below 100 cm from night to day. Analyses were performed using the l o g i t transformation of the proportion data. A Source DF Type I I I SS Mean Square F Value Pr > F TREAT 1 0.10654630 0.10654630 3 .61 0.1977 Error 2 0.05468853 0.02734426 B Source DF Type I I I SS Mean Square F Value Pr > F TREAT 1 0.02890008 0.02890008 0 . 01 0 .9322 Error 2 6.25834208 3 .12917103 Table 4.4: Experiment 2 - nested analyses of variance for depths of in d i v i d u a l S. oregonensis: tests of significance for fishwater (+\/-) treatment effects and tube within treatment effects. For a l l times, treatment df=l and tube within treatment df=2. tube within treatment F t u b e P>Ftube error df day 1 2 02 0 . 640 0 5068 31 0 398 0 6750 day 2 2 31 0.530 0 5334 29 0 635 0 5370 day 3 2 01 7.913 0 1059 34 0 449 0 6419 day 4 2 00 2 .550 0 2512 32 0 404 0 6713 night 1 2 09 0.040 0 8598 32 0 360 0 7006 night 2 2 01 20.21 0 0459 30 0 128 0 8807 night 3 2 01 0.014 0 9157 28 0 981 0 3875 treatment effect F t r e a t m e n t P > - - ; ' t r e a t m e n t time error df 120 Experiment 3 The addition of f i s h d i r e c t l y into f i s h treatment cyclinders did not r e s u l t i n a higher proportion of S. oregonensis below 100 cm r e l a t i v e to either the f i s h [-] treatment (Fig. 4.8) nor the fishwater [+] treatments i n experiment 1. Likewise, the change i n the proportion of S. oregonensis below 100 cm from night to day did not increase with f i s h d i r e c t l y i n the tank rather than fishwater treatment of experiment 1 (Fig. 4.9). Nested analyses of variance for depths of ind i v i d u a l S. oregonensis indicate no s i g n i f i c a n t treatment or tube within treatment effects (Table 4.5). Temporal variation in vertical distribution in the lakes The means of the daytime depth d i s t r i b u t i o n s i n Kennedy Lake are consistently deep and, through time, the d i s t r i b u t i o n s have a sim i l a r variance. At night, when the d i s t r i b u t i o n s are shallower, the variance of the d i s t r i b u t i o n s are greater and the mean depths more variable between sample dates. In Great Central Lake, mean depth (both day and night) are comparable to mean depth i n Kennedy Lake at night, except that on some dates mean depths are deeper. The variances of each depth d i s t r i b u t i o n are high for both day and night (Fig. 4.10). D i s c u s s i o n The absence of an effect of f i s h treated water or caged f i s h on v e r t i c a l migration behaviour of S. oregonensis suggests that i n these lakes t h i s species exhibits c o n s t i t u t i v e v e r t i c a l migration behaviour. N e i l l (1992) also reported c o n s t i t u t i v e 121 Fig. 4.8: Experiment 3 - change in proportion of Kennedy Lake S. oregonensis below 100 cm from night to day. Solid squares indicate fish [-] treatments and open squares indicate fish [+] treatments. 122 E o o o o 0 _Q c o \"\u2022c o CL o 0.78 0.76 0.74 0.72 0.70 H 0.68-0.66-0.64 24-Oct 25-Oct 26-Oct Fig. 4.9: Increase in the proportion of Kennedy Lake S. oregonensis below 100 cm from night to day in fishwater [+\/-] treatments of experiment 3. Other details as in Fig. 4.8. Table 4.5: Experiment 3 - nested analyses of variance for depths of i n d i v i d u a l S. oregonensis: tests of significance for f i s h (+\/-) treatment effects and tube within treatment effects. For a l l times, treatment df=l and tube within treatment df=2. tube within treatment F t u b e P>Ftube error df day 1 2 00 0 156 0 .7308 26 0 436 0 6514 night 1 2 02 0 756 0 .4753 27 0 199 0 8210 night 2 2 03 1 154 0 .3937 31 0 983 0 3856 treatment effect ^treatment \u2022P->Ftreatmen time error df t to Day Night 15-Jul-92 31-Jan-93 19-Aug-93 7-Mar-94 23-Sep-94 15-Jul-92 31-Jan-93 19-Aug-93 7-Mar-94 23-Sep-94 0 5 10 \u2022 15 \u2022 20 \" 25 \u2022 -i r-i i r 6-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 6-May-92 3-Sep-92 12-Dec-92 22-Mar-93 30-Jun-93 Fig. 4.10: Day and night weighted mean depth of S. oregonensis in Kennedy Lake (top) and Great Central Lake (bottom) for multiple sample dates. Vertical bars represent variance of each depth distribution for each date. 125 v e r t i c a l migration i n Diaptomus kenai exposed to cutthroat trout (Onchorhynchus clarki) predation. In contrast, most of the l i t e r a t u r e reporting experiments designed to test for inducible effects show such effects. These include studies of freshwater copepods ( N e i l l 1990, Neill 1992), Acartia hudsonica (Bollens and Frost 1991), cladocerans (Dodson 1988, Ringelberg 1991a, Ringelberg 1991b, De Meester 1993, Loose and Dawidowicz 1994), Artemia larvae (Forward and Rittschof 1993), and chaoborid larvae (Dawidowicz et al. 1990, Tjossem 1990). The absence of seasonal changes i n v e r t i c a l migration behaviour of S. oregonensis provides further evidence that the behaviour i s fixed. In winter, stickleback move to deep water (Dr. K.D. Hyatt, FOC, unpublished data). They l i k e l y reduce t h e i r feeding a c t i v i t y to a low l e v e l . Temperature throughout the water column i s below 6o c, and, i n laboratory tanks, stickleback at t h i s temperature are nearly inactive (Dr. J.D. McPhail, Department of Zoology, University of B r i t i s h Columbia pers. comm. and Dr. K.D. Hyatt, FOC, pers. comm.). Despite the apparent decreased predation r i s k from stickleback, S. oregenensis i n Kennedy Lake continue to migrate i n the winter. In contrast, many zooplankton species exhibit decreased migration, or a complete absence of migration, during seasons when predators are reduced i n numbers or inactive (Bollens and Frost 1989b, Ohman 1990, Frost and Bollens 1992, De Stasio 1993, Huang et al. 1993, Brancelj and Blejec 1994). Inducible versus c o n s t i t u t i v e v e r t i c a l migration behaviours are expected to be selected i n d i f f e r e n t kinds of environments (Harvell 1990, N e i l l 1992, Pijanowska 1993), and i n Kennedy Lake 126 three factors may favour the evolution of the l a t t e r . F i r s t , predation r i s k from stickleback may be invariant from year to year. In the absence of flu c t u a t i n g r i s k , f l e x i b l e behaviour would not be selected. Second, the absence of a cost to v e r t i c a l migration also favours a fixed behaviour. Although evidence from the analysis of feeding rhythms i n chapter 2 suggests a cost to v e r t i c a l migration, t h i s cost may be absent i n winter when food abundance i s low and i t s d i s t r i b u t i o n i s homogenous down to 24 m (Fig. 2.7 - February). This may explain continued v e r t i c a l migration i n winter when stickleback are r e l a t i v e l y inactive. Third, environmental cues from stickleback may be too unreliable, making f l e x i b l e v e r t i c a l migration too r i s k y . During the long summer breeding season, stickleback numbers offshore can fluctuate rapidly as they migrate i n synchrony between l i t t o r a l breeding zones and the offshore habitat (Hyatt and Ringler 1989). Cues that induce migration might disappear offshore when the l i t t o r a l migrations occur. I f so, when stickleback return to the pelagic habitat i n large numbers, S. oregonensis dependent on such a cue, i n i t i a l l y would not be migrating and thus ( u n t i l the cue took effect) they would be vulnerable to predation. The lack of a treatment ef f e c t i n my experiments i s possibly because of i n s u f f i c i e n t cues for the induction of migration. Although the caged f i s h were placed d i r e c t l y into the tubes to ensure a high, and constant, l e v e l of exudate, Bollens et al. (1994) showed that only d i r e c t mechanical or v i s u a l s t i m u l i induce v e r t i c a l migration i n Acartia hudsonica. Free swimming f i s h i n enclosures trigger v e r t i c a l migration of Acartia hudsonica while caged f i s h are inadequate (Bollens and Frost 127 1989a). Nonetheless, i n the experiments involving Kennedy Lake S. oregonensis, lack of a treatment ef f e c t appears to be due to strong migration of most individuals both with and without f i s h , rather than the lack of s u f f i c i e n t stimulus to cause greater v e r t i c a l migration r e l a t i v e to the non-fish treatment. Although the experimental resul t s support the hypothesis that v e r t i c a l migration behaviour by S. oregonensis i s fixed at maturity, my experimental design did not test for developmental determination of the migration phenotype. Further experiments with laboratory reared individuals are required to determine i f exposure to p a r t i c u l a r predation regimes during development determines the migration phenotype of mature ind i v i d u a l s . N e i l l (1992) showed that Diaptomus kenai from a lake containing predatory cutthroat trout displayed invariant, obligate v e r t i c a l migration that was unaffected by exposure to the trout during ontogeny. 128 Chapter 5 A dynamic optimization model of Skistodiaptomus oregonensis ver t i ca l migration Introduction In chapters 2 and 3 I present evidence consistent with two hypotheses: 1) that S. oregonensis v e r t i c a l migration i n Kennedy and Paxton lakes i s driven by stickleback predation, and, 2) that t h i s avoidance behaviour has a cost associated with l o s t feeding opportunities i n surface waters. In th i s chapter I construct an optimal depth decision model for S. oregonensis i n Kennedy Lake based on the physiological energetics of S. oregonensis, the temperature environment, the abundance and v e r t i c a l d i s t r i b u t i o n of food, and estimates of predation r i s k . I then use the model to examine the effect of predation on optimal depths, and the consequent feeding opportunity costs. This provides an assessment of whether the hypotheses (1 and 2, above) are tenable on the basis of the present knowledge of the biology of S. oregonensis and relevant factors of the Kennedy Lake environment. To b u i l d the model, I use a dynamic optimization algorithm. This provides a technique to determine sequential behavioural decisions that maximize fi t n e s s over the entire l i f e - h i s t o r y of an organism. Multiple factors a f f e c t i n g fit n e s s are expressed i n a common currency. This makes i t possible to explore f i t n e s s trade-offs that may exist among factors (Mangel and Clarke 1988). In the model presented here, expected l i f e t i m e fecundity i s used as a measure of fitness which, i n turn, depends on both growth to maturity and the p r o b a b i l i t y of surviving predation. I then 129 estimate the sequential depth choices necessary to optimize a combination of growth and the p r o b a b i l i t y of survi v a l through a l i f e t i m e . Although the dynamic optimization technique has been used previously to predict the occurrence of v e r t i c a l migration as a trade-off between predator avoidance and food a q u i s i t i o n (Clarke and Levy 1988, Mangel and Clarke 1988), my model i s the f i r s t to use r e a l i s t i c representations of growth and predation r i s k over the entire l i f e - h i s t o r y of an organism. Model D e s c r i p t i o n Overview My model predicts l i f e t i m e fitness-maximizing depth choices for discrete size classes of S. oregonensis for each day and night period for l i f e durations ranging from 30-70 days. For each depth there are associated growth benefits determined by food intake (depends on food abundance which varies with depth) and metabolic rate (depends on temperature which varies with depth). Also, for each depth there are associated costs due to the r i s k of being eaten by either of two f i s h predators: juvenile sockeye or stickleback. Since these f i s h are v i s u a l predators, r i s k not only depends on depth but also time of day. This i s because l i g h t i n t e n s i t y decreases with depth and varies with the time of day. Fitness i s defined as the product of reproductive output at the f i n a l time period, T, m u l t i p l i e d by the p r o b a b i l i t y of survival to the f i n a l reproductive stage. Animals must grow to reach a minimum reproductive size above which reproductive output 130 increases l i n e a r l y with weight. However, increased fitness due to greater growth may trade off with s u r v i v a l i f feeding near the surface increases predation r i s k . The algorithm f i r s t calculates the optimal choice at terminal time, T, then works recursively through time to calculate the optimal combination of choices to arrive at th i s f i n a l outcome. The res u l t i n g decisions y i e l d the maximum l i f e t i m e fitness which maximizes the p r o b a b i l i t y of survi v a l from time = 1 to T mul t i p l i e d by reproductive success at time T (Mangel and Clarke 1988). Once optimal depth choices for each size at each time are calculated, the growth trajectory from time t=l to t=T can be modelled for any size i n d i v i d u a l at t=l. F i r s t , growth at t=l i s calculated for an indi v i d u a l S. oregonensis of size(t=l) residing at the optimal depth. This growth i s added to the size( t ) to become s i z e ( t + l ) . This process i s repeated for each time step and the size at each time i s stored i n a growth trajectory array. Then l i f e t i m e survival i s calculated as the product of the p r o b a b i l i t i e s of survival at each time step from t=l to T. Biological Components of the Model Cohort analysis of samples from lat e winter\/spring of 1993 indicate that growth from egg stage to maturity takes approximately 50 to 71 days (Fig. 5-1). The default model l i f e h i s t o r y duration was set at 50 days, although a range of 30-70 days was modelled. In the model, v e r t i c a l migration and growth of S. oregonensis are simulated each day and night, beginning A p r i l 15. One night period and one daytime period were modelled 24-02-93 eggs <.4 4.25 4.75 5.25 5.75 6.25 6.75 7.25 7.75 8.25 >8.5 Size class (10~1 mm) F ig . 5 -1 : Cohor t analys is of S . oregonensis in late winter\/spring of 1993 (Kennedy Lake) . 132 for each day. For s i m p l i c i t y , I assume S. oregonensis to be semelparous i n the model, although a f t e r maturity a number of egg broods may be produced before death. Metabolism Specific rates of oxygen consumption per i n d i v i d u a l (ug 02\/hr) for 5 species i n the genus Diaptomus (including S. oregonensis) were determined by Comita (1968) as function of temperature (T, \u00b0C) and weight (W, ug) : l o g 1 0 s p e c i f i c 02 = a*T-b* l o g 1 0 W + c (5.1) with a=0.0364, b=0.3418, and c=0.6182. This i s converted to t o t a l consumption per i n d i v i d u a l by taking the a n t i l o g and multiplying by weight: t o t a l 0 2 = 10 1 0 9 1 0 s * e c i f i c \u00b02 * w (5.2) Total 02 i s converted to carbon consumed i n r e s p i r a t i o n using the conversion of Parsons et al. (1984): Cresp = Total 0 2 * - ^ - j * RQ (5.3) using an RQ (respiratory quotient) equal to 1. Temperature for each time and depth are determined from f i e l d temperature depth p r o f i l e s for three sample dates during the period modelled (15-04-93, 13-05-93, and 23-06-93). Values for each day between the sample dates were estimated by li n e a r interpolation between dates for each depth. 133 Food Intake I measured the concentration of chlorophyll a i n the 3-50 um size f r a c t i o n (see chapter 2 for d e t a i l s of method) on three sample dates at Kennedy Lake (15-04-93, 13-05-93, and 23-06-93). The sample depths were 1, 3, 5, 7, 10, 17, and 24 m. Values at 1 m inte r v a l s to 24 m were estimated by li n e a r interpolation. These were used to r e l a t i v i z e the estimate by Stockner et al. (1980) of t o t a l epilimnetic phytoplankton c e l l density i n Kennedy Lake (mean = 0.91 c e l l s per ul) across depths on each sample date: Pz = relative chl az * Pepllimn * 24 (5.4) where Pz = phytoplankton c e l l concentration at depth z and P e p i i i m n = phytoplankton c e l l concentration i n epilimnion. Phytoplankton c e l l densities for each day between the three sample dates were then estimated from l i n e a r i n t e r p o l a t i o n for each depth between dates. Based on the l i s t of phytoplankton c e l l types, c e l l sizes, and q u a l i t a t i v e abundance i n Kennedy Lake (Stockner et al. 1980), I assumed an average c e l l volume of 750 um3. From Strathmann (1967) I estimated carbon content per c e l l (C) from c e l l volume (V) as: lo g 1 0 C = 0.866 * l o g 1 0 V - 0.46 (5.5) The amount of carbon per ml of lake water at each depth i s 134 determined as: C ml cell c (5.6) Carbon intake can then be calculated as: intake = F * \u2014 ml (5.7) where F = f i l t e r i n g rate (ml\/animal\/day). Peters and Downing (1984) used an extensive l i t e r a t u r e survey to determine a predictive equation for copepod f i l t e r i n g rates: where R = food p a r t i c l e volume, C = volume of the experimental container, and M = experiment duration. Temperature was not a si g n i f i c a n t factor according to t h e i r analysis. I set R=750 um3, and C and M at the median values reported by Peters and Downing (1984) for t h e i r l i t e r a t u r e survey (C = 500 ml, M = 1440 min). The equation thus s i m p l i f i e s to: Reported maximum f i l t e r i n g rate for mature S. oregonensis (Richman 1966) f a l l p r e c i s e l y on the curve for t h i s equation (Fig. 5-2). l o g 1 0 F = -1.245 + 0.534 * l o g 1 0 W + 0.683 * l o g 1 0 R - 0.067 * (log 1 0 R) 2 + 0.0001 * C - 0.0002 * M (5.8) l o g 1 0 F = -0 . 07315 + l o g 1 0 W * 0 . 534 (5.9) Growth Carbon available for growth i s the difference between t o t a l carbon intake and carbon used i n r e s p i r a t i o n . Carbon available 135 Fig. 5-2: Copepod filtering rates as a function body weight. General equation for copepods taken from Peters and Downing (1984), the values for mature S. oregonensis from Richman (1966). 136 for growth i s m u l t i p l i e d by 2 (Peters 1984) to estimate t o t a l growth: growth = (Clntake - Cresp) * 2 (5.10) Fecundity The metasome lengths of the smallest egg-bearing female S. oregonensis i n Kennedy Lake were approximately 800 um which corresponds to dry weight of 6.33 ug (length-weight conversion for S. oregonensis from Culver et al. 1985). In the model, fecundity of S. oregonensis less than 6.33 ug was set at 0 at time T. Above 6.33 ug, the number of eggs produced increases with weight according to the fecundity\/weight r e l a t i o n s h i p established for Lake Erie S. oregonensis (Davis 1961). ln{eggs) = 0.193059 * dry weight + 0.592806 (5.11) Survival Survival i s modelled as the p r o b a b i l i t y of not being eaten by sockeye and stickleback. I use the v i s u a l predation model described by Aksnes and Giske (1993) and Clarke and Levy (1988). Feeding rate of one f i s h predator (f) measured as number of prey taken per unit time i s determined by: f = T\u00b1 * n * [RD * s i n Q ) 2 * v * Nprey where T 1=search time, T 2=handling time, RD=reactive distance, 8=reactive f i e l d angle, v=fish swimming speed, and N p r e y=prey density. Handling time (T2) was assumed to be 0, reactive f i e l d 137 angle to be 45 s, v was set at 6 cm^s\"1 for juvenile sockeye and 5 cm*s_1 for stickleback, and N p r e y was set at 2000 per m3. The p r o b a b i l i t y of being eaten for each S. oregonensis i s then determined as: \" p r e y P^obmottality=^ * f (5.13) where N f i s h = f i s h density. For juvenile sockeye, N f i s h was set at 93 0 per hectare and for stickleback at 1400 per hectare. These values correspond to the means of 10 annual point estimates for Kennedy Lake (Kim Hyatt, Department of Fisheries and Oceans, unpublished data). Reactive distance i s a function of several components which include ambient l i g h t i n t e n s i t y . Aksnes and Giske (1993) derive a theoretical function for RD as: RD Iz * Tl * L j r e y * CQ (5.14! N is where I z = l i g h t l e v e l at depth z, C0 = inherent prey contrast, LP r e y = prey length, and *S = s e n s i t i v i t y threshold. Values for C0 and *S are not known. Juvenile sockeye begin feeding at a l i g h t l e v e l of about 9.3 * 10\"7 lux and feeding rate reaches a plateau at about 0.093 lux ( A l i 1959, his Fig. 11). Constants of the above equation were adjusted to provide RD values that increase over these same range of l i g h t l e v e l s . *S was set at 4.5 * 10\"7, and C0 was set at 0.5. In chapter 3, evidence was presented that juvenile sockeye begin feeding shortly a f t e r sunset and continue for 40 minutes to 138 an hour after sunset. Thus, i n the model, juvenile sockeye feed during the night period and not during the day. The surface l i g h t i n t e n s i t y during feeding was set at 0.224 lux, the approximate l i g h t i n t e n s i t y 22 minutes af t e r sunset on a clear night. Light i n t e n s i t y at each depth i s determined according to equation 3.1 using an e x t i n c t i o n c o e f f i c i e n t of 0.401. The reactive distances of sockeye at each depth r e s u l t i n g from t h i s model are shown i n Figure 5-3 (top). In contrast to juvenile sockeye, stickleback i n the model feed during the day only. Chapter 3 provided evidence that stickleback do not feed at the l i g h t levels that occur a f t e r sunset and before sunrise. As i n my laboratory experiments (Chapter 3), I model the effect of l i g h t i n t e n s i t y on stickleback feeding rate using quantum units, rather than the photometric units used for the juvenile sockeye model ( A l i 1959), except that daytime surface l i g h t i n t e n s i t y was set at 1000 lux. Maximum l i g h t i n t e n s i t y at the lake surface at mid-day on a clear day i s approximately 2000 lux. However, i n t e n s i t i e s are lower i n the morning and afternoon and are approximately halved by t h i n clouds, and cut by a t h i r d i n average cloudy conditions (United States Navy Bureau of Ships 1952). Default \u00b1S was set at 1.0 * 10\"5 and C0 at 0.5. Experimental data (chapter 3) show that feeding rate i s maximum at 1.6 uE-s^-rrr 2 and nearly 0 at 0.1 i^E ' S \" 1-m'2. Thus, where l i g h t i n t e n s i t y at depth exceeded 1.6 uE, I set l i g h t i n t e n s i t y at these depths to 1.6 jaE-s^-m\"2, and at depths at which l i g h t i n t e n s i t y was < 0.1 laE-s^-m\"2, l i g h t i n t e n s i t y was set to 0. The modelled reactive distance of the stickleback at each depth during the day are shown i n Figure 5-3 (bottom) . 139 Fig . 5-3: Mode l reactive d is tances of juveni le s o c k e y e at night (top) and s t ick leback dur ing the day (bottom) at e a c h depth for 0.25 m m , 0.50, 0.75 m m and 1.00 m m s ize prey. 140 Predation i n t e n s i t y due to stickleback and juvenile sockeye was altered i n the model by specifying the t o t a l number of hours each day or night that each f i s h a c t i v e l y feed. Model R e s u l t s The model results predict v e r t i c a l migration only when f i s h predation occurs (Fig. 5-4). When neither f i s h species feed (panel a), maximum fitness i s achieved when S. oregonensis remains near the surface both day and night over i t s entire l i f e . This strategy maximizes growth, r e s u l t i n g i n higher fecundity. When stickleback feeding i s introduced into the model (panel b and c), maximum fitness i s achieved when S. oregonensis of a l l sizes are at depth during the day and near the surface at night. Increasing the duration of stickleback feeding from 2 hours up to 10 or more has v i r t u a l l y no effect on optimal depth choice. When juvenile sockeye feed for one hour, and stickleback r i s k i s not included i n the model (panel d), maximum fitness i s achieved when S. oregonensis remain deeper at night than during the day. This effect i s due to the night time r i s k of juvenile sockeye predation. Under these conditions, S. oregonensis remain quite deep during the day u n t i l about 25-30 days, af t e r which they r i s e to the surface during the day and, thereafter, grow rapid l y to a large mature si z e . However, i f stickleback feeding i s included i n the model along with sockeye feeding, S. oregonensis migrate to deep water during the day to avoid the stickleback and r i s e to the surface at night to feed, although they do not r i s e as high at night as they do when sockeye predation i s absent (panel b and c). no stickleback stickleback feed 2 hours stickleback feed 4 hours 0 >> 0 o o CO o c 3 . cd n CO q. cd Q 0 -5 -10 -15 -20 -25 I I I I I i J I L 5 10 15 20 25 30 35 40 45 50 5 10 15 20 25 30 35 40 45 50 5 10 15 20 25 30 35 40 45 50 \"D 0 & J 0 - L-^ i\u2014 O o CO O) cd n CO q. cd Q 5 10 15 20 25 30 35 40 45 50 5 10 15 20 25 30 35 40 45 50 5 10 15 20 25 30 35 40 45 50 Days Fig. 5-4: Model effect of stickleback and sockeye feeding on day (open squares) and night (asterisks) optimal growth and depths of S. oregonensis. Dashed horizontal line indicates the minimum size of reproduction. Phytoplankton food cell density was set to 1900 * 10 phytoplankton cells per ml. Default values were used for other model parameters. 142 While the degree of stickleback predation r i s k has l i t t l e e ffect on the optimal depths chosen, i t does decrease fi t n e s s by decreasing f i n a l body size (and the r e s u l t i n g fecundity) and li f e t i m e s u r v i v a l p r o b a b i l i t y (Fig. 5-5). This decrease i s greater i f sockeye predation i s included i n the model. Interaction exists between predation r i s k and food density effe c t s . At 1700 phytoplankton c e l l s per u l , p r o b a b i l i t y of surviv a l for S. oregonensis approaches 0 as predation r i s k increases, while at 1900 phytoplankton c e l l s per u l , l i f e t i m e s u r v i v a l p r o b a b i l i t y remains near 0.4. In order to reach the minimum size of maturity at the lower food density, S. oregonensis i s required to increase predation r i s k i n surface waters. Like predation r i s k , food density does not effect the predicted optimal depths (Fig. 5-6A). Optimal daytime depth i s below 20 m for a l l sizes and night depth i s near the surface. At 1500 phytoplankton c e l l s \/ u l , S. oregonensis i s unable to grow to maturity, at 1700 c e l l s \/ u l the minimum size for maturity i s reached but the p r o b a b i l i t y of survival i s very low (Fig. 5-6B). Between 1700-2500 phytoplankton c e l l s per u l both f i n a l size and pro b a b i l i t y of survi v a l increase. The lack of an effect of food density on mean daytime depth i n the model i s inconsistent with observations of such an effect by Huntley and Brooks (1982), Dagg (1985), and Johnsen and Jakobsen (1987) and the suggestion of t h i s effect for S. oregonensis (Fig. 2.28). In the model, the optimal strategy as food density decreases i s to stay deep during the day and s a c r i f i c e growth, rather than to r i s e toward the surface and no sockeye sockeye feed 1 hour Hours per day stickleback feed Hours per day stickleback feed Fig. 5-5: Model effect of stickleback and juvenile sockeye feeding on final body size and lifetime survival probability of S. oregonensis at 1700 (solid squares) and 1900 (open squares) phytoplankton cells per pi. Dashed horizontal line indicates the minimum size of reproduction. Default values were used for other model parameters. A CD N CO Q_