Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Trophic phasing of juvenile chum salmon (Oncorhynchus keta Walbaum) and harpacticoid copepods in the… D'Amours, Denis 1987

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1987_A1 D35.pdf [ 8.71MB ]
Metadata
JSON: 831-1.0053321.json
JSON-LD: 831-1.0053321-ld.json
RDF/XML (Pretty): 831-1.0053321-rdf.xml
RDF/JSON: 831-1.0053321-rdf.json
Turtle: 831-1.0053321-turtle.txt
N-Triples: 831-1.0053321-rdf-ntriples.txt
Original Record: 831-1.0053321-source.json
Full Text
831-1.0053321-fulltext.txt
Citation
831-1.0053321.ris

Full Text

TROPHIC PHASING OF JUVENILE CHUM SALMON [ONCORHYNCHUS KETA WALBAUMI AND HARPACTICOID COPEPODS IN THE FRASER RIVER ESTUARY, BRITISH COLUMBIA by DENIS D'AMOURS B.Sp.Sc. Universite du Quebec a Rimouski A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY m THE FACULTY OF GRADUATE STUDIES The Department of Oceanography We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA 7 October 1987 c Denis D'Amours, 1987 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. I 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. The University of British Columbia 1956 Main Mall Vancouver, Canada Department of V6T 1Y3 Abstract Within the environmental approach for the study of fluctuations in fish population abundance, factors that may regulate the overlap in time and space (phasing) of fishes, food supply, and predators, are sought. This trophic phasing analysis is based on the recognition that production of food is a process at least partially independent of the production of consumers. Trophic phasing analysis was applied in investigating production of chum salmon in the Fraser River estuary. Juvenile chum salmon were captured near a tidal flat; the abundance of salmon near the flat was highest in late May in 1985 and in early June in 1986. These salmon relied heavily on harpacticoid copepods as a food source. Individual taxa as well as the assemblage of main prey harpacticoids also had periods of highest abundance in the water column. The blooming period of the prey harpacticoid assemblage coincided with the appearance in the sediment of warming episodes. These warming episodes result from interactions between the daily heat cycle and specific tide patterns. Variations in the degree of overlap of the periods of highest abundance of salmon and harpacticoids could affect the survival of the fishes. The degree of overlap of those periods was hindcasted using indices for the temporal patterns of abundance of salmon and harpacticoids on the flat. The median date of downstream migration at a counting station upstream was used for the salmon; the timing of the second annual occurrence of tide conditions giving rise to a warming event in the sediment was used for harpacticoids. Difference in time between the two events was taken as a phasing index accounting for two degrees of freedom in the process of fish production. There is suggestion of a non-monotonic relationship between the index of survival of even broodyear chum salmon and the hindcasted phasing index. ii Contents A b s t r a c t i i Table of Contents i i i L i s t of Fi g u r e s iv L i s t of Tables v Acknowledgements v i 1 G e n e r a l I n t r o d u c t i o n 1 2 F i s h P o p u l a t i o n R e g u l a t i o n 7 2.1 Introduction 7 2.2 The stock-recruitment connection 8 2.3 The Environmental Thesis 15 2.3.1 A Review of Selected Studies 16 2.3.2 Trophic Phasing Analysis 20 2.4 Chapter Summary 22 3 E a r l y Sea Lif e of C h u m Sal m o n 23 3.1 Introduction 23 iii 3.2 Near-Shore Life of Juvenile Chum Salmon 24 3.3 Study Site and Methods 25 3.4 Results and Discussion . 31 3.5 Chapter Summary 43 4 Spat i o Temporal Distribution of Natant Harpacticoid Copepods 44 4.1 Introduction 44 4.2 Vertical Distribution of Selected Species of Natant Harpacticoid Copepods 46 4.2.1 Methods 47 4.2.2 Results and Discussion 48 4.2.3 Discussion 64 4.3 Thermal Regime and Seasonal Abundance of Selected Species of Natant Harpacticoids 66 4.3.1 Material and Methods 68 4.3.2 Results and Discussion 71 4.4 Chapter Summary 97 5 E a r l y Near-Shore Life and Survival of C h u m Salmon 98 5.1 Introduction 98 5.2 Summary 124 6 General Discussion and Conclusion 125 6.1 Chum Salmon Abundance and Early Near-Shore Life 125 6.2 Trophic Phasing Analysis and Fisheries Oceanography 129 6.3 Conclusion 133 iv References 135 A 143 B 153 v List of Figures 1.1 Density related attributes of a general stock recruitment relationship 4 2.1 Density-dependent population regulation 10 2.2 Stock-recruitment curve and replacement line 12 3.1 Map of study area 27 3.2 Photograph of sled sampler 29 3.3 Catches of juvenile chum salmon on Roberts Bank in 1985 and 1986 32 3.4 Chum salmon gut contents in 1985 and 1986 36 3.5 Relative proportions of dominant harpacticoid prey taxa in 1985 and 1986 (gut contents) 37 3.6 Zooplankton composition in the water column in 1985 and 1986 . . . 39 3.7 Relative proportions in the water column of dominant harpacticoid prey taxa in 1985 and 1986 40 3.8 Electivity index for dominant harpacticoid prey species in 1985 and 1986 42 4.1 Vertical distribution of Zaus aurtlii 49 4.2 Vertical distribution of Harpacticus uniremis 50 4.3 Vertical distribution of Tisbe spp 51 4.4 Mean number of harpacticoids (selected species) per transect in 1986 52 vi 4.5 Specific composition of natant harpacticoid assemblage 54 4.6 Volume intersected through a density surface described by dN{z)fdz = -kN[z) 59 4.7 Absolute abundance and index of abundance of Zaus aurelii in the water column 60 4.8 Absolute abundance and index of abundance of Harpacticus uniremis in the water column 61 4.9 Absolute abundance (biased and corrected) and index of abundance of Tisbe spp. in the water column 62 4.10 Seasonal Abundance of Natant Harpacticoids 72 4.11 Temperature in the sediment in 1985 (recording depth: 5 cm) . . . . 75 4.12 Temperature in the sediment in 1986 (recording depth: 1 cm) . . . . 76 4.13 Temperature in the sediment in 1986 (recording depth: 5 cm) . . . . 77 4.14 Pigments concentrations in the sediment 79 4.15 Abundance of harpacticoid ovisacs in 1986 80 4.16 Water level over station H at flowing tide 81 4.17 Potential effect of time of high water on the magnitude and direction of change in sediment surface temperature 84 4.18 High frequency temperature fluctuations in the sediment in 1985 . . 87 4.19 High frequency temperature fluctuations in the sediment in 1986 (recording depth: 5 cm) 88 4.20 High frequency temperature fluctuations in the sediment in 1986 (depth of recording: 1 cm) 89 4.21 Timing and duration of harpacticoid blooms and warming episodes in the sediment 91 vii 5.1 Variation of time to grow through a given size interval for various growth coefficients 101 5.2 The match or mismatch of larval production to that of their larval food . . 104 5.3 Daily abundance of outmigrant Fraser River chum salmon at Mission City in 1972 106 5.4 Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 odd &: even broodyears I l l 5.5 Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 odd broodyears 115 5.6 Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 even broodyears 116 5.7 Juvenile-to-adult survival of chum salmon versus median date of downstream migration at Mission (redrawn from Beacham & Starr (1982)) 119 5.8 Juvenile-to-adult index of survival of chum salmon versus median date of downstream migration at Mission (even - f odd broodyears). 120 5.9 Juvenile-to-adult index of survival of chum salmon versus median date of downstream migration at Mission (odd broodyears) 121 5.10 Juvenile-to-adult survival of chum salmon versus median date of downstream migration at Mission (even broodyears) 122 viii List of Tables 3.1 Length (fork) of juvenile chum salmon on Roberts Bank in 1985 and 1986 33 3.2 Comparison of chum salmon fork length before and after fixation . . 34 4.1 Abundance of Zaus aurelii at various sampling levels 55 4.2 Mean number of Harpacticus uniremis at various sampling levels . . 55 4.3 Fitting the distribution of Tisbe spp. with negative exponentials . . 58 4.4 Total abundance of Tisbe spp., Zaus aurelii, and Harpacticus uniremis in the water column at station H during the sampling period in 1986. 58 4.5 Relative abundance of the three harpacticoid taxa 73 5.1 Median date of chum salmon downstream migration, hindcasted date of natant harpacticoid bloom onset, phasing index, and juvenile-to-adult survival of chum salmon 109 5.2 Comparison of predicted and observed tides at Point Atkinson (1980)117 ix Acknowledgements I thank my research supervisor, Dr. T.R. Parsons, for his encouragement, support, and advice during the course of this study. I also thank the other members of my research committee, Drs. P.G. Harrison, C D . Levings, T .F . Pedersen, and F.J.R. Taylor, for their guidance. I was personally supported during the time of this study by the Department of Fisheries and Oceans, by Mr. Gordon Melbourne (through TRP) , and by Mr. John (late) and Mrs. Nerina Bene. Mr. D. Dobson (BIO) kindly provided the thermographs; Mr. M . Farwell (DFO) provided data on salmon survival and migration. Finally, I thank Dr. J . Petkau and Mr. B. Leroux (Statistical Consulting and Research Laboratory), Dr. P. LeBlond, Mr. D. Webb, Ms. D. Masson, and Mr. D.K. Lee, for their invaluable contributions to various aspects of this work. x C hapter 1 Genera l Introduction A seemingly inherent characteristic of fishery resources is their variability on a wide range of temporal and spatial scales (Csirke & Sharp 1984; Steele 1984). To in-vestigate this variability, two "radically different approaches" (Legendre & Demers 1984) have been developed. The difference between the two approaches can be ap-preciated by comparing the frameworks through which each addresses the issue of fish population variability. One approach proposes to investigate variability in fish abundance as the result of natural environmental fluctuations; the other proposes to investigate fish population variations as the result of anthropogenic activity, namely fishing. In 1914, Hjort proposed that the magnitude of a fish year-class was set by the success of the fishes in colonizing their environment during early life stages. This success was assumed to vary in accordance with some environmental factor(s) in such a manner that the size of the parental fish population would be of little importance, if any, in the determination of the filial year-class strength. A well known study along Hjort's proposition is that of Burkenroad (1948) on the Pacific halibut. Burkenroad submitted that major fluctuations in abundance of the halibut could be attributed to natural causes and that the relevance of the then current fish population regulation theories remained to be demonstrated. How-1 ever, Burkenroad's thesis was perhaps simply balancing that of Thompson (1937) who, after studying the same fish, had concluded that fishing was responsible for the fluctuations in abundance of the Pacific halibut. In 1954, Ricker proposed a model, better known as "stock-recruitment model", by which the effect of fishing on subsequent fish stock sizes could be formally investigated. This model was based on the principle of population homeostasis through density-dependence, which states that the rate of increase of an animal population decreases as its size increases. The apparent soundness of the foundations of this model and its timeliness in a pe-riod when the reported world fish catch was sharply increasing, lead to substantial efforts in the investigation and modeling of the effects of fishing on the dynamics of fish populations. At the same time, investigation and modeling of the effects of environmental fluctuations on the dynamics of fish populations were disregarded (Larkin 1978). However, it later became obvious that even when many years' data were available, stock-recruitment models accounted for only a fraction, sometimes small, of the observed variability in typical fish abundance data (Hunter 1976; Sissenwine 1984). The interest in Hjort's thesis was reactivated since much of the blurring in stock-recruitment data was considered to be the result of environmental variability (Anonymous 1980) and since advances in biological oceanography were providing a much better understanding of the complex mechanisms hitherto lumped by fisheries biology into the single word "environment" (Larkin 1978). However, as pointed out by various sources (e.g. Ricker 1975; Gulland 1983; Sissenwine 1984; Larkin 1973,1984) recruitment models relating fish population fluctuations and environmental factors generally fail to predict post-publication events probably because of the often misleading significance of the supportive correlations. Applications of either of the two radically different approaches for the investiga-2 tion of fluctuations in fish abundance lead to inconclusive results when the nature and function of fish population regulation mechanisms are investigated. Sophisti-cation of statistical and computational techniques aside, the situation today is not much different than it was 40 or 50 years ago: whether the size of a fish population is determined by that of the parental stock size or by some environmental factor(s) remains conjectural. The recent literature on fish population regulation is still characterized by a dichotomy of approaches between fisheries biology, honoring the parental stock hypothesis, and fisheries oceanography, finding in Hjort's thesis an attractive rationale for its concern for the environment of the fishes. This schism has resulted in an "apparent immiscibility of fisheries biology and oceanography" (Wooster 1983). The difficulty in reconciling the two approaches is apparent in the following semantic conflict. When the different regions of a curve relating parental and filial abundance of a fish population are qualified in fisheries biology and fisheries oceanography, totally opposite attributes are used for the same re-gions (Fig. l.l). One who undertakes a study on fish population regulation can only be in a quandary when trying to select an appropriate approach. Clearly then, the development of a conceptual framework upon which to organize the research effort should be the initial step in a study aimed at gaining an understanding on the nature and function of the mechanisms regulating fish population abundance (Anonymous 1980; Csirke & Sharp 1984). In this thesis, a study on the variability in abundance of the chum salmon (Oncorhynchus keta Walbaum) of the Fraser River is presented. For reasons men-tioned above, chapter 2 consists of a review of the two current paradigms on the nature and function of the mechanisms regulating fish abundance. In this review, a framework is proposed that is termed "trophic phasing analysis". This framework 3 Spawners(S) Figure 1.1: Density related attributes of a general stock recruitment relationship. Region "A" can be considered density-independent since dR/dS s= constant and region B can be considered density-dependent since dR/dS ^ constant (parental filial thesis); alternatively, region u A r can be considered density-dependent since R appears to be of first-order with respect to S, and region "B" can be considered density-independent since R appears to be of O-order w.r.t. S. (environment thesis). 4 requires the description of the spatial and temporal distributions of the fishes and knowledge of their trophodynamics. It is then possible to identify environmen-tal factors that modulate the realized overlap of the fishes, their predators, and their food items, and which may regulate survival and abundance of the fishes. In subsequent chapters, forming the core of the experimental and analytical ef-fort of this thesis, this proposed "trophic phasing analysis" is applied in a case study on the chum salmon. In chapter 3, the life-cycle of the chum salmon is re-viewed with emphasis on the initial near-shore stage; the diet of the fish at that stage is described relative to food samples collected using an innovative technique and the main dietary items are identified. The spatial and temporal distributions of an assemblage of species of harpacticoid copepods are described in chapter 4. The assemblage is made up of those species preferentially preyed upon by the chum salmon. Effort is made to resolve some difficulties in locating harpacticoid copepods effectively consumed by the fish. The temporal variability of the same assemblage of harpacticoids is linked with specific patterns in the thermal regime in their environment. In chapter 5, potential limitation of chum salmon production in the Fraser River estuary is discussed in light of the information exposed in the two previous chapters. Available data on chum salmon survival from outmigrant juveniles to returning adults is related to a hindcasted index of phasing of the fish and harpacticoid copepods. In chapter 6, the results of this research effort are discussed at two different, but nested, levels. Applicability of the results in the management of the chum salmon are discussed; then, the use of trophic phasing analysis in the study of fish population regulation is appraised. The orientation of this research project is based on the view that production of food organisms is a process at least partially independent of the feeding of predators 5 (Parsons et al. 1984 a). This thesis was undertaken with the aims of providing a contribution to the on-going debate on the nature and action of the mechanisms regulating fish abundance, and to contribute to our understanding of the ecology of harpacticoid copepods, which might help clarify their role in the diet of juvenile chum salmon. Temperature data collected during the course of this study were deposited at the Data Library of the University of British Columbia in a machine-readable format (see D'Amours 1986). 6 C h a p t e r 2 Fish Populat ion Regulation 2.1 Introduction After reviewing the population dynamics of the Pacific sardine, Clark & Marr (1956) concluded that for any particular fish species, the year-class strength could be determined by three possible scenarios: 1. year-class strength is a function of the parental stock size (parent-stock thesis) 2. year-class strength is determined by environmental factor(s) (environmental thesis) 3. year-class strength is determined by a mixture of 1. and 2. (dual thesis). More than 30 years later, it seems that Clark & Marr's third scenario has gained wide acceptance. The strength of a year-class of fishes is generally considered as the result of density-dependent homeostatic balance operating on stages set by environmental factors (Cushing 1984; Wooster 1983; Larkin 1978; Beddington & May 1977). Considering scenario 1. or 2. independently of each other would be an "unfortunate dichotomy" (Gulland 1977). However, there is reason to believe that the wide acceptance of the "dual thesis" could be detrimental to progress in 7 understanding fish recruitment mechanisms. In this chapter, the foundations of the stock-recruitment theory (parent-stock hypothesis) are reviewed; the use of a stock-recruitment model in fisheries management is briefly discussed. A distinc-tion is established between the operational usefulness of stock-recruitment models and the need for empirical evidence supporting either the parental-stock or the environmental thesis. 2.2 The stock-recruitment connection A most fundamental observation in population biology is that animal populations do not increase without limits (Krebs 1978). In fisheries biology, which is but a special case of population biology, the equivalent statement is made when it is con-sidered basic that fish populations are limited in size even when not fished (Ricker 1954; Beverton &; Holt 1957). As a result of this observation, it must be concluded that some relation must exist between the abundance of the reproductive stock and the number of recruits produced (Ricker 1985). In fisheries biology, the relation-ship between the parental stock size and the number of recruits produced can be derived as follows. Let it be first mentioned that a stock-recruitment relationship can be established between any two points in the life cycle of a fish. Assume an initial number of fish eggs N0 that decline in time following ^P-F(JVW). (2-D The functionality between dN(t)/dt and N(t) could be established in various ways. For a population to stop increasing when a certain size is reached, either • mortality increases • fecundity decreases • mortality increases and fecundity decreases. The change of intensity in mortality and/or fecundity as the population increases is what sets a limit to the size a population can reach and is known as density-dependent population regulation, a concept developed by Nicholson (1933) (see Fig. 2.1). In fisheries biology, it has generally been assumed correct and probably not misleading to incorporate the density related effects on the growth rate, in mortality (Z) (Gulland 1983). Eq. 2.1 can then be re-written as ^ = -Z(t)N(t). (2.2) With the simple case of linear density-dependent effect in Z, that is when Z(t) = (M 1 + M 2AT 0), (where Mi and M2 are mortality coefficient components) eq. 2.2 can re-written as d N ^ = -(Mi + M2N0)N(t). (2.3) Integrating 2.3, At t = 0, dt \oge N(t) = -Mit - M2N0t + const. l°g e N0 = const. , then, Let N0 = Sf, and N(tr) = R, where 5=number of spawners, and /= number of eggs per spawner, <r=time to recruitment, and .Renumber of recruits, POPULATION DENSITY Figure 2.1: Density-dependent population regulation. As the size (density) of a population increases, either fecundity decreases and mortality increases (a), or fecundity decreases (b), or mortality increases (c) (redrawn from Krebs 1978). 10 then, R = Sfe -M1tre-M2Sft, (2.4) With a — f e — const. and 0 = M2ftr — const. , eq. 2.4 reduces to R = aSe~0S. (2.5) Eq. 2.5 is currently known as the Ricker type stock recruitment curve. In fish-eries management, this relationship can be used to forecast the number of recruits, and determine what fraction of the recruits should be made available to fisheries. First, the relationship is parametrized with historical data of spawners and recruits abundance. A plot of this parametrized relationship is more or less dome-shaped, and the fraction of the stock available for harvest is that which lies above the re-placement line (Fig. 2.2). To locate the maximum recruitment, R is differentiated with respect to 5, i.e. dR dS = ae cxS0e-f)S = ae'fiS{l-0S). The maximum recruitment is obtained when dR d~S = 0 that is, when 11 Figure 2.2: Ricker type stock-recruitment curve for even broodyears 1956-1976 of Fraser River chum salmon (from Beacham 1984). The fraction of the the stock above the replacement line (surplus recruitment) is available for harvest. 12 and will be equal to a _ 0.3679a We 0 ' The maximum surplus recruitment (MSR) will be obtained when dR __ dS ~ 1 ' that is, when the production curve is at greatest distance from the replacement line (Hilborn &; Peterman 1977). The solution for dR/dS = 1 is to be found by iteration, or can be approximated directly from the values of a and /? (Hilborn 1985). For a fish like the salmon that is fished just before its once in a lifetime spawning, the MSR is practically equal to the maximum sustainable yield (MSY) (Ricker 1985). The MSY is that level of exploitation that can yield the highest return on a continuous basis. Since the only variable of the whole fishing process under the control of fisheries managers is escapement (Hilborn 1983), establishment of quotas is a basic tool in fisheries management. Expanded as stock-recruitment models, the parent-stock hypothesis is one effective way to meet the requirements of fisheries managers of establishing quotas with consistency and objectivity. In that sense, the parent-stock hypothesis is no longer one among other hypotheses on the nature of fish population regulation mechanisms; it is a tool of proven operational usefulness for the determination of quotas by which escapement can be controlled. As pointed out earlier, the fit of typical stock-recruitment data on the model is often very poor. Larkin (1978) summarized the situation as follows It is common gossip that if you didn't have the certain knowledge that zero adults produce zero offspring, we could fit a Ricker model, or a Beverton-Holt model, or a straight line, or a circle, with equal satisfac-tion. 13 Why, then, despite their poor empirical verification, are stock-recruitment mod-els widely used, and more importantly in the present discussion, why must any hypothesis on the nature of fish population regulation (e.g. environmental thesis) be merged with the parent-stock hypothesis to avoid an "unfortunate dichotomy" ? The possible answer to these two questions is the same: because of the opera-tional usefulness of the formal extensions of the parent-stock hypothesis. While operational usefulness appears a reasonable criterion to adopt the parent-stock hy-pothesis for the business of management, operational usefulness does not prove the relevance of Nicholsonian stock-recruitment models as the best model for fish population regulation mechanisms. It is quite probable, then, that Clark & Marr's third scenario (dual thesis) owes its popularity to the difficulty of divorcing the idea of operational usefulness within the actual scheme of fisheries management. Merging the parent-stock hypothesis and the environment thesis in the study of fish population regulation mechanisms could be premature. From an intuitive point of view, it may appear necessary that any animal population would become limited from increasing by its very own size. However, the nature and ways of operation of the mechanisms responsible for this self-regulation are still conjectural (as discussed by Krebs (1978)). As pointed out by this author, the conceptual frameworks for the study of animal population regulation must be validated against real populations. The parent-stock hypothesis seems easily applied to real populations. While all must agree that zero spawners produce zero offspring, there is no agreement on the validity of assuming that the observed maximal stock size is in any way related to the theoretical maximal size. This means that the parameters computed for the stock-recruitment relationship could be nothing but fitting parameters with only faint nominal statistical support (see Parsons et al. 1984 a). 14 Since the 1950's, most ecologists have chosen to abandon the tiresome argu-ments about density-dependence by choosing an empirical approach to the study of animal population; dogmatic theories of animal population regulation are re-placed by more "fruitful arguments about empirical relationships" (Krebs 1978). Perhaps this evolution in the approach of ecologists studying population regulation would have been difficult had the notion of operational usefulness been kept in the forefront of research. Once the issues of operational usefulness and empirical ver-ification are disentangled, it appears that the study of fish population regulation aiming at understanding recruitment mechanisms has at its disposition two the-ses, the parent-stock thesis and the environment thesis, both in want of empirical verification. In the case of general population biology, debates on whether the Nicholsonian principle of population regulation through density dependence is op-erative, have been replaced by a search for empirical arguments on real populations. Oceanography is providing an increasingly detailed knowledge of the environment of marine organisms; it appears reasonable to consider the study of fish population regulation through the environment thesis as a practical concern of oceanography (see Parsons et al. 1984 a). 2.3 The Environmental Thesis Once it has been chosen to replace debates on density-dependence by "arguments about empirical relationships", environments of fishes become overwhelmingly rich aggregates of factors which might be considered as determinants of fish population abundance. The term environment is a broad one indeed, with the result that one is in a quandary when having to choose one particular environmental factor against which to analyse fluctuations in fish abundance. Nonetheless, a careful 15 selection is possible if knowledge is available on the mechanisms regulating growth and mortality of fishes. Such analysis based on environmental factors selected for sound biological reasons is considered to be less likely to result in spurious relationships (Ricker 1975). To make a sensible choice of environmental factor(s), some sort of conceptual framework with which to theorize and organize the research effort is required. By reviewing selected studies that yield new insights in fish recruitment mechanisms, recurring themes appear that could form the core of such a conceptual framework. 2 . 3 . 1 A Review of Selected Studies Hjort (1914) was the first to suggest that hydrological and biological conditions could determine the "subsequent wealth or poverty of a [fish] year class". Today, this proposition is referred to as Hjort's hypothesis. Legendre &: Demers (1984) have given the following interpretation of Hjort's hypothesis: recruitment depends on the success or failure of the annual colonization of their environment by the fish larvae, irrespective of the initial number of eggs produced. This interpretation emphasizes the role of the environment in determining fish year-class strength in early-life stages. A well articulated hypothesis in that sense was provided by Cushing (1975). Cushing submitted that fish abundance variability could result from differential matching of the fish larvae and their food resource. Larvae were assumed to be transported by currents to feeding grounds after fixed spawning periods. The onset of the production cycle on the feeding grounds was reported to be determined by the vertical stability of the water column. The timing of this onset could vary from year to year depending on fluctuations in meteorological conditions regulating the vertical stability of the water column. Any variation in 16 the relative timing of the arrival of the larvae on the feeding grounds and the onset of the production cycle could alter the food quantity available per fish. Fluctuations in food supply could then cause variability in survival by starvation. The whole hypothesis was named the match/mismatch hypothesis. However, as pointed out by Sinclair & Tremblay (1984), the match/mismatch hypothesis may owe some of its popularity to the attractiveness of its formulation rather than to its empirical support. Sinclair & Tremblay studied the distribution of the Atlantic herring and found no evidence in support of the coupling of the timing of spawning and the spring bloom onset central in Hjort's hypothesis. Instead, Sinclair & Tremblay proposed that the variability in the timing of spawning of herring is an adaptation to specific oceanographic retention characteristics. Herring retained in a region where a spring bloom allows a fast growth can be spawned in spring and still hatch before the next winter. Herring retained in a region with a depressed primary production would grow more slowly as a result of a lower food supply and would need to be spawned the previous fall to metamorphose in the same time envelope as those spawned in spring. The concept of larval retention was put forward by lies & Sinclair (1982). These authors proposed that hydrodynamics (tidal mixing) and fish larvae behaviour could interact so as to maintain an aggregated population of Atlantic herring larvae for many months. The hitherto unexplained variability of size between stocks could then be related to the extent of the respective retention areas: the larger the retention area, the larger the stock. Although the theory falls short of detailing the retentive mechanisms, it proposes an original stock concept. As for the within stock interannual variability in size, lies & Sinclair submitted that it need not be associated in whole or in part with fluctuations in mean food supply since observed 17 interanrmal stock size fluctuations are far higher than those in mean food supply. Rather, interannual size fluctuations within a stock could result from variations in retentive characteristics. (It should be mentioned that this argumentation assumes that mean food supply is an appropriate factor to be related to fish abundance variability, which is surely arguable). If variations in retentive characteristics are to cause variations in survival of fish larvae, the next logical step is to ask what causes the "non-retained" larvae to be lost from recruitment? A possible cause of death for non-retained fishes, if food supply is not regarded, could be sudden exposure to predators in "ecologically dangerous" sites (Frank & Leggett 1985). Frank & Leggett have reviewed evidence challenging the view that reciprocal oscillations of predators and prey should indicate trophic interactions. Rather, reciprocal oscillations observed at one point in the ocean could be the result of the passing of different water masses in which predators and prey are segregated for adaptive reasons. This would suppose that the concerned organisms could maintain their positions, actively or passively, in various water masses. For the prey, the adaptive value of being retained in an "ecologically safe site" where predators are absent seems obvious; it is not so clear that there would be any adaptive reason for the predators to stay away from their potential prey. Any relaxation of the retention of the prey in their "safe sites" could make them exposed to predators in "dangerous sites". This interruption of segregation could produce sudden predator-prey interaction more of the all-or-none type rather than of the continuous Ivlev type. Another possible, albeit very speculative, reason for the non-retained fishes to be lost from recruitment is that young fishes could need a minimum time of retention in order to be able to come back as spawners to the spawning site (parental stream hypothesis). The application of the concept 18 of the parental stream hypothesis to open ocean fishes could be an extension of the already noted analogy between the life-cycles of the salmon and the herring (Cushing 1973). Another well developed hypothesis as to how environmental factors may affect fish survival is Lasker's (1975) stability hypothesis. Lasker showed that storms could dissipate food aggregates in the upper layers of the waters where anchovy (Engraulis mordax) feed. This dissipation would result in a reduced ration for the fishes; the following depressed growth and weakened condition of the fishes could increase physiological and predation mortality. Peterman &; Bradford (1987) pro-vided some evidence that indeed wind-driven turbulent mixing affects variability in survival of larval anchovy. These authors found the daily mortality rate of lar-val anchovy to be significantly correlated to an index of wind speed: the higher the wind index, the lower the survival. However, Peterman & Bradford's findings do not necessarily indicate that variability in larval mortality due to wind-driven turbulent mixing sets the level of recruitment: variability in post-larval stages unaccounted for by their model could obliterate any relationship between larval survival and final recruitment. Nonetheless, Peterman & Bradford's results sup-port Lasker's proposed mechanism of variability in larval mortality of the anchovy. Vertical stability of the water column is the distal environmental factor regulating fish mortality via its effect on the food supply. Common to those reviewed studies that yield new insights in fish recruitment mechanisms are the following two points: • the spatial and temporal distribution of juvenile fishes is described • the realized trophodynamics of the fishes is assumed to be determined by their degree of overlap with their food supply and/or their predators. 19 It has been suggested by Legendre & Demers (1984) that food supply be consid-ered as the proxi through which the distal effects of environmental factors may be conducted up the food web to the fishes. Since predation could also be modulated by environmental factors (e.g. "safe sites" and "dangerous sites"), Legendre & Demers' proposition could be expanded as follows: trophodynamics is the proxi through which distal environmental variability may be conducted up or down the food web to the fishes. The co-occurence of fishes, their food supply and their predators appears to be determined by various environmental factors. Variability of those factors can affect the realized overlap (i.e. phasing) of the components of the food web; such shifts in phase can greatly affect the time course of the system. Factors causing phase shifts between trophically related components of an ecosystem have been termed "phasing functions" (Parsons & Kessler 1986). In brief, there is a distinct research avenue within the environmental thesis that seeks to identify phasing functions by detailing the temporal and spatial distributions of organisms mutually involved in trophic relationships; for these reasons, this avenue could be termed "trophic phasing analysis". 2.3.2 Trophic Phasing Analysis In order to represent formally the concept of trophic phasing analysis, the basic equation to describe the distribution of a non-conservative variable in the ocean may be used (see Wroblewski 1983). In a Lagrangian reference frame, the variation of a biological variable (P) can be written as : d2P dP dP dP dP rjr — + u— + v—- + w— = KH at ox dy dz d2P d2P dx2 dy2 20 where u,v, and w are velocity components; KH and Ky are vertical (V) and hori-zontal (H) diffusion coefficients; and B is the sum of sources and sinks. From time t to time t + i within one fish generation, there can be many sinks. In the reviewed studies, predation and starvation have come out to be the most often mentioned probable causes of mortality of fishes. However, there are no sources properly speaking. Only food-supply can be considered source-like in that fluctuations in food supply will cause fluctuations in mortality. Equation 2.6 can be written for the fishes, their food items, and their preda-tors. The time course of the phasing of the fishes and either or both of their food supply and their predators can then be investigated along combinations of spatial dimensions. Lasker's stability hypothesis is a 1-dimensional model (z;t), while the Match-Mismatch hypothesis is a 2-dimensional model (x,y;t). This allows clear identification of the functional space in which trophic interactions occur separate from the interactions themselves. In other words, the geometry of a trophic system is recognized as a distinct component that could separate species, or assemblages of species. As it is, equation 2.6 provides a grid to assess systematically the adequacy of the knowledge on the life history of the fish species under study. This grid favours a structured approach towards the identification of phasing functions that may largely contribute to the determination of fish abundance by modulating the overlap of the fishes, their food supply (sink) and their predators (sink). In practice, this grid requires a thorough documentation on the trophodynamics of the fishes as well as a detailed description of the temporal and spatial distributions of the various components involved in it. Trophic phasing analysis may then be defined as the search for environmental factors that regulate the spatio-temporal overlap 21 of the c o m p o n e n t s of a t r o p h i c s t ruc tu re . 2.4 Chapter Summary T w o theses p reva i l for address ing the issue of fish p o p u l a t i o n v a r i a b i l i t y . T h e p a r e n t a l - f i l i a l thesis pos tu la tes tha t regula t ive m e c h a n i s m s c a n leap genera t ions ; t h i s is an a t t r a c t i v e thesis w h e n the secur ing of a cons tan t benefit f r o m an ex-p l o i t e d fish p o p u l a t i o n is sought . T h e e n v i r o n m e n t thesis pos tu la tes t ha t r ea l fish p o p u l a t i o n a b u n d a n c e is d e t e r m i n e d by e n v i r o n m e n t a l fac tor (s ) , i r respec t ive of the p a r e n t a l s tock size. E n v i r o n m e n t a l v a r i a b i l i t y is a ssumed to be c o n d u c t e d to fish p o p u l a t i o n s t h r o u g h the vec tor of food up take i n ea r ly life stages. 22 C h a p t e r 3 E a r l y S e a L i f e o f C h u m S a l m o n 3.1 Introduction In the previous chapter, the need to describe the spatial and temporal distributions of fishes relative to their food supply and predators was emphasized. This descrip-tion could eventually allow the identification of where and when in the life cycle the effects of environmental fluctuations may be transmitted to the fish through the vector of food uptake. To deal with the complexity of the life cycle of a fish, it is useful to identify distinct stages within this cycle, and investigate those stages the most likely to comprise determinants of year-class strength. The life history of the Pacific salmon can be realistically and conveniently divided into a series of stages, each characterized by risks of particular kinds of mortality (Larkin 1977). For the chum salmon Oncorhynchus keta, a more thorough understanding of the biological and physical processes that affect individual growth and production has been recognized as a central goal of fisheries research (Neave 1953; Wickett 1958; Volk et al. 1984). In this chapter, the life cycle of the chum salmon is briefly reviewed. The early near-shore life of the fish is identified as a distinct ecological stage and its diet at that stage is described. 23 3.2 Near-Shore Life of Juvenile Chum Salmon The chum salmon is one of the five species of Pacific salmon (Oncorhynchus spp.) that occur naturally in the North American side of the Pacific ocean (as reviewed by Larkin 1977). All five species are anadromous and die after spawning. The spawning period is a convenient point to start and end a description of the life cycle of the Pacific salmon. In the fall of the year, females deposit their eggs in redds in rivers; the eggs are simultaneously fertilized by the male. After two to three months of incubation, eggs hatch and alevins remain in the gravel for a few more weeks, while gradually migrating to the overlaying waters. In spring, rising temperature or other environmental factors, in conjunction with their increased ability to swim, will cause the juvenile fishes to leave the gravel beds for a free-swimming life. At this time, chum salmon immediately start swimming and drifting to sea (Hoar 1951). In small rivers, the sea can be reached within days; in larger rivers, the sea may be reached only after weeks of fresh water life. Many chum salmon will remain in coastal waters until mid- or late summer before dispersing to offshore regions. By the end of their second season at sea, chum salmon originating from British Columbia are found over a wide area of the Gulf of Alaska. After two to three summers at sea, most chum salmon return to their parental streams to spawn. Chum salmon will then pass through estuaries twice during their life: once as juvenile outmigrants, and once as redd-bound mature adults. As juvenile outmigrants, chum salmon may reside in estuarine near-shore areas for variable periods. In estuarine near-shore areas of many rivers, juvenile chum salmon exhibit clear stenophagic feeding patterns (Kaczynski et al. 1973; Feller & Kaczynski 1975; Healey 1979; Sibert 1979). Despite a broad variety of available food, chum salmon rely heavily on harpacticoid copepods as a food source. In the Fraser River 24 estuary, intertidal habitats, vegetated and non-vegetated, are considered as having a substantia] capacity for fish rearing; juvenile chum salmon are also known to occur seasonally in close proximity to these intertidal habitats (Gordon & Levings 1984; see also the review by Levings et al. 1983). In order t o discuss the potential effects of the feeding patterns of the chum salmon on its growth and production in the Fraser River estuary, its diet in this particular area had to be described and compared with the results of other similar studies. The particular site selected for the study was chosen because juvenile salmon have been reported in its immediate vicinity (Gordon & Levings 1984) and it was vegetated with eelgrass (Zostera marina L.). Although no detailed description of the meiofauna on the study site was available, eelgrass beds are known to harbour populations of harpacticoid copepods (as reviewed by Hicks &; Coull 1983); Levings et al. (1983) further report the presence of harpacticoid copepods on a vegetated section of Roberts Bank, of which the selected site is part. 3.3 Study Site and Methods The sampling station for chum salmon was located near the Westshore Terminal Causeway on Roberts Bank, British Columbia (Fig. 3.1). A detailed biosedimen-tological description of the area is provided by Swinbanks & Luternauer (1987). Swinbanks & Luternauer's description is based on aerial photographs taken in 1977, prior to expansion of the causeway. Harrison (1987) documents changes in the eelgrass cover on the same section of Roberts Bank from 1969 to 1984. The eelgrass cover expanded from ca 250 ha in 1969, the date of the construction of the causeway, to ca 430 ha in 1984 (from sources cited in Harrison 1987). In this region of the Strait of Georgia, the character of the tide is mixed, mainly semi-diurnal 25 (as reviewed by LeBlond 1983). The range of the tide at the port of reference (Point Atkinson) is 3.3 m for mean tide and 4.9 m for large tide. The eelgrass cover at sampling station H was up to 100 shoots/m2 in mid-spring (D. Webb, U.B.C. Oceanography, unpublished data), which is similar to the values reported by Harrison (1987). Chum salmon were caught with a beach seine (length: 15 m; height: 1.85 m; bunt: 3 m) with 1 cm mesh (stretched) in the wings and 3 mm mesh (stretched) in the bunt. At low tide (ca 5 hours after the onset of the ebb at the port of reference, Point Atkinson), the beach seine was dragged from the beach with a motorized boat to ca. 10 m from the shoreline; the seine was then stretched parallel to the beach and retrieved with lines from the shore. This operation was repeated 3 times at each sampling. The number of chum salmon in each of the hauls was recorded; a random sample of fishes was kept for gut-content analysis after fixation in 4% formaldehyde-sea water solution. Chum salmon were identified according to McConnel & Snyder (1972) and Phillips (1977). On May 13, 1986, 30 chum salmon were measured immediately after capture and prior to fixation in formaldehyde; the same fishes were re-measured subsequently to assess the effect of fixation on length measurements. Fish lengths reported are in all cases fork lengths (FL) (from the tip of the snout to the inside of the tail notch). The sum of the chum salmon in the 3 hauls was reported as the catch per unit effort (C.P.U.E.). Beach seining was repeated fortnightly during the sampling season. The sampling season extended from March 22 to July 5 in 1985, and from March 5 to July 10 in 1986. In the laboratory, fish were measured to the nearest mm (FL); fish length measurements were made at least a few weeks after fixation in formaldehyde. The mid-guts of the fishes were extracted and their contents identified and counted. At each sampling date, the gut contents of at least 10 fishes were determined. Zoo-26 Figure 3.1: Map of study area. Chum salmon were fished at station F; harpacticoid copepods were collected at station H (the dotted line indicates the limit of the tidal flat). 27 plankton samples against which to compare the fish gut contents were collected amidst an extensive eelgrass bed on Roberts Bank, close to the salmon fishing sta-tion (Fig. 3.1). Zooplankton samples were collected bi-weekly from March 1 to July 5 in 1985, and from February 7 to July 10 in 1986. Samples were collected with a stratified sled-sampler (Fig. 3.2) operated by a SCUBA diver. The sled-sampler consisted of 5 butyrate cylinders (internal diameter: 83 mm; length: 150 mm) serially mounted on a stake fixed on a skid-pad. The butts of the cylinders were closed with 64 fim Nitex© gauze, thus forming filtering baskets. The centered height of each of the 5 cylindrical baskets above the sediment was 56, 156, 256, 356, and 456 mm respectively. At low tide, transect line(s) (1 in 1985; 3 in 1986) were spread on the sediment surface at station H. In 1985, the single transect was always oriented towards the east; in 1986, the 3 transects had a common origin and were respectively oriented towards the south, east, and north, which was also the order of their execution. For both years, sampling of the zooplankton with the sled-sampler was initiated ca. 1 hour after slack high tide (2.5 to 3 m of water above the station). The sled-sampler was opened at the surface prior to descending on the transect (s). This allowed wetting of the 64 fim mesh of the filtering baskets that otherwise seemed to resist flow by surface tension. The sled-sampler was then capped and the diver descended to the transect(s). After properly orienting the sled-sampler, the diver uncapped the sled-sampler and pushed it along a transect line always keeping the skid-pad in contact with the sediment; this assured a con-stant height above sediment for the filtering baskets. At the end of the transect, the sled-sampler was capped and brought back to the tending boat. Baskets were removed from the sled by the tender, rinsed, and their contents stored in 4% filtered (40 /xm) sea water-formaldehyde. The baskets were re-installed on the sled-sampler 28 Figure 3.2: Photograph of diver-operated sled-sampler (scale bar=15 cm) (see text for description). 29 and the operation repeated for the subsequent transects. When 3 transects were done (1986), the sampling operation was completed in less than 1 hour, and accord-ingly less when only 1 transect was done (1985). In the laboratory, zooplankton samples were treated according to the procedures recommended by Uhlig et al. (1973) for meiofauna. Certain taxa of harpacticoid copepods were identified to the species according to the procedures recommended by Coull (1977). Other taxa were identified to major groups. In all cases, the total number of animals in each basket of each transect was counted. Principal keys consulted for the identification of harpacticoid copepods were those of Lang (1948 ; 1965) and of Wells (1976). The gut contents of the salmon were compared to the zooplankton samples using Ivlev's (1961) electivity index. This electivity index (E.I.) is defined as: BJ. = ^ , r, + Pi where is the numerical proportion of a prey in the ration, and p< is the numerical proportion of the same animal in the reference food sample. This electivity index varies between -1 and +1; a negative value indicates avoid-ance, while a positive value indicates preference. The unidentified copepods in the sled samples were not included in the computation of the electivity index; their number in the water column was such that the electivity index for the prey zoo-plankters would have been highly distorted. Those unidentified copepods (reported as UCOP in Appendix B) were mostly early copepodites, and were very rarely, if ever, observed in gut contents. The unidentified copepods in the gut contents were included, since they represented adult individuals of various species (reported as UNCO in Appendix A). 30 3.4 Results and Discussion During both sampling years, a pulse of juvenile chum salmon was observed on Roberts Bank ( F i g . 3.3). In 1985, chum salmon were first observed on March 29 (day 88); in 1986, chum salmon were first observed on April 1 (day 91). Subsequent temporal patterns in the abundance of salmon are similar for both years. First, a low peak was observed (on April 10 (day 100) in 1985 and on April 15 (day 105) in 1986). There followed a second and more important peak which rapidly declined until no more juvenile salmon were observed. The significance of the peaks can not be assessed for a lack of variance estimation. Because of limited logistic support, only a very limited number of beach seine hauls were done. However, the similarity in the temporal patterns of abundance of fish on the flat between years suggests that the two peaks may reasonably be assumed to represent meaningful features of abundance. The first peak of salmon may represent a pulse of fish from stocks from nearer spawning areas, or earlier outmigrants, or hatchery reared fishes. As to the last possibility, no fish with clipped adipose fin, which identifies some of the hatchery reared fishes, was ever observed despite close observation. The second and highest peak of abundance occurred on May 27 (day 147) in 1986, and on June 6 (day 157) in 1985. The proportion of the Fraser chum stock represented by those peaks is unknown. It can only be assumed, as in Valiela and Kistritz (1980) who discussed the abundance of chum salmon in Fraser river marsh ecosystems, that at least a proportion of the whole stock utilizes tidal flats for a period of time during their outmigration. The size of the juvenile chum salmon captured in 1985 ranged from 47.5 to 62.1 mm (FL), and from 31.3 to 54.2 mm (FL) in 1986 (Table 3.1). Why the upper limit of the size range was lower in 1986 than in 1985 is unclear. In 1986, several chum salmon were observed with yolk-sac 31 Q_ O 600-i 500 A 400-300 H 200 H 100 H 60 Legend > 1985 80 100 120 140 160 180 200 Time (calendar days) Figure 3.3: Catches of juvenile chum salmon on Roberts Bank in 1985 and 1986 (C.P.U.E.=catch per unit effort; 1 unit effort=3 beach seine hauls). 32 CHUM SALMON FORK LENGTH 1985 1986 Date (day) Length (mm) mean+/- 1 S.E. Date (day) Length (mm) mean +/- 1 S.E. March 29 (88) 47.5 +/- 1.0 March 19 (78) 31.3 +/- 1.8 April 10 (100) 53.6 +/- 0.9 April 1 (91) 44.3 +/- 1.2 April 23 (113) 49.2 + /- 2.5 April 15 (105) 48.8 +/- 2.5 May 9 (129) 53.8 +/- 2.1 April 29 (119) 37.0 +/- 0.5 May 23 (143) 55.8 +/- 3.2 May 13 (132) 44.7 +/- 1.1 June 7 (157) 55.0 +/- 1.1 May 27 (147) 47.6 +/- 1.3 June 21 (172) 62.1 +/- 2.8 June 11 (162) 54.2 + /- 1.4 July 5 (188) 53.0 +/- 1.9 Table 3.1: Length (fork) of juvenile chum salmon on Roberts Bank in 1985 and 1986 (length measured on fishes preserved in formaldehyde). 33 FIXATION SHRINKAGE Fish # Length (mm) Fish # Length (mm) May 13 May 20 May 13 May 20 1 46 43 16 43 40 2 45 42 17 40 38 3 40 38 18 42 40 4 39 35 19 42 40 5 40 38 20 55 52 6 37 34 21 43 40 7 39 37 22 49 46 8 34 31 23 47 44 9 42 38 24 33 33 10 38 36 25 35 33 11 35 33 26 40 37 12 45 44 27 40 38 13 40 36 28 40 40 14 36 34 29 39 38 15 44 41 30 35 34 Table 3.2: Comparison of chum salmon fork length before and after fixation (1 week in 10% sea water-formaldehyde solution). scar still present, which were not present in 1985. The fixation in formaldehyde had an effect on the measurements of fish fork length. The mean size of recently caught fresh fishes was 40.8 + /- 0.9 (S.E.) mm; after a week of fixation, the mean length of the same group of fishes had dropped to 38.4 +/- 0.8 (S.E.) mm, which is significant (one-tailed f-test; p<0.05) (Table 3.2). Globally, fish shrank by 5.9 % because of the fixative within 1 week. Subsequent measurements indicated no further appreciable shrinking. Chum salmon size corrected for fixative shrinking ranged from 50.5 to 66.0 mm in 1985, and from 33.2 to 57.6 mm in 1986. Gut contents analysis showed that for both sampling years, juvenile chum salmon relied heavily on harpacticoid copepods as a food source (Fig. 3.4). In 34 1985, the digestive tracts of 92 fishes (10-15 fishes per fishing effort) were ana-lyzed; in 1986, the digestive tracts of 66 fishes (at least 10 per fishing effort) were analyzed; gut contents analysis results are summarized in Appendix A. In 1985 and 1986, 77.2% and 71.8% (numerical abundance) of the prey of the juvenile salmon were either Harpacticus uniremis Kroyer 1842, Zaus aurelii Poppe 1884, or Tisbe spp. (this taxa assemblage is subsequently referred to as the H + Z+T+ assem-blage). Other unidentified adult harpacticoids (UNCO in Fig. 3.4) accounted for 8.0% and 2.5% of the fish's diet in 1985 and 1986 respectively. For both years, calanoids accounted for ca 8% of the food items. Other food items (amphipods, cypris, fish larvae, isopods, ostracods, cumaceans) each accounted for less than 5% of the diet, except in 1986 when cumaceans accounted for 11.7% of the diet. Other food items were observed (euphausids, insect larvae, isopods) that each accounted for less than 0.1% of the diet in both or single years (not visible on Fig. 3.4). Within the H+Z+T assemblage, Harpacticus uniremis was the dominant taxon (64.2% in 1985 and 46.3% in 1986); respectively for 1985 and 1986, Zaus aurelii accounted for 26.7% and 35.8% of the assemblage while Tisbe spp. accounted for 9.1% and 17.9% of the assemblage (Fig. 3.5). In 1985, the zooplankton taxonomic composition of the combined 5 sampling levels of the sled-sampler were analyzed; in 1986, the sum (3 transects) of the combined 5 sampling levels of the sampler were analyzed; the results of these taxonomic descriptions are detailed in Appendix B. The composition of the samples collected with the sled-samplers differed from that observed in juvenile salmon gut contents for both years (Fig. 3.6). The H+Z-rT assemblage accounted for 58.7% and 76.5% of the total numerical abundance of the animals in 1985 and 1986 respectively; the proportion of calanoids was ca 4 times higher in 1985 compared 35 H+Z+T Figure 3.4: Chum salmon gut contents in 1985 and 1986. 36 Harpacticus 64.2% ^ 1985 Tisbe 9.1% Figure 3.5: Relative proportions of Harpacticus uniremis. Zaus aurtlii and Tisbe spp. in H + Z-rT assemblage in 1985 and 1986 (gut contents). 37 to 1986 (32.7% vs 7.8%). In 1986, adult harpacticoids (UNCO) other than those of the H+Z+T+ assemblage were observed in appreciable amount (3.2%); Mesochra sp. accounted for 9.6% of the total amount in the sled-samples. Within the H+Z+T assemblage in the sled-samples, Tisbe spp. was by far the dominant taxon (98.9% in 1985 and 80.7 in 1986); for 1985 and 1986 respectively, the proportion of Zaus aurelii was 0.6% and 14.9%, and that of Harpacticus uniremis, 0.6% and 4.5%. (Fig. 3.7). Between years, the proportion of the H+Z+T assemblage in the water column appeared to fluctuate principally because of changes in the abundance of calanoids. The dominance of the H+Z+T assemblage in the gut contents was similar for both years. Globally, then, the H+Z+T assemblage was the dominant taxonomic fea-ture in the sled-samples as well as in the gut contents. However, there is a striking difference in the relative proportion of Tisbe spp., Zaus aurelii, and Harpacticus uniremis in the H+Z+T assemblage when sled samples and gut contents are com-pared (Figs. 3.7 & 3.8). The dominant species (H. uniremis) of the assemblage in the gut contents was the least abundant in the sled-samples for both years (64.2% & 46.3% vs 0.6% &: 4.5%), and vice versa. Similarly for Z. aurelii, the proportion in the gut contents (26.7% & 35.8%) was higher than that in the water column (0.6% &i 14.9%). This is interpreted as an indication of the high preference of juvenile chum salmon for H. uniremis and Z. aurelii. This evidence for highly selective feeding of juvenile chum salmon must however be considered with caution. Some of the difference in specific proportion in the H + Z+T assemblage between gut contents and sled-samples could be caused by the fishes feeding on areas with faunal composition different from that on the sled sampling station. Horizontal heterogeneity (zonation or patchiness) is a well 38 1986 H+Z+T 58.7% \ 1985 Am^hipods Ostracods 0.2% Calanoids 7.8% Caprellids 0.4% mrjhipodj UNCO 3.2% Calanoids 32.7% Mesochra 9.6% Ostracods 0.4% Cyprls 2.5% Figure 3.6: Zooplankton composition in the water column in 1985 and 1986 (unidentified copepodites are not included). 39 1985 Figure 3.7: Relative proportion of Tisbe spp., Zaus aurelii, and Harpacticus uniremis in the H-!-Z-rT assemblage in 1985 and 1986 (water column). 40 recognized characteristic of harpacticoid copepods on tidal flats (as reviewed by Hicks & Coull 1983). Also, it could be that harpacticoid species preferentially preyed upon by juvenile chum salmon could be more easily caught by the fish than by the sled-sampler. As it is, H. uniremis and Z. aurelii are members of the Harpacticidae family, which is characterized by the presence of strongly prehensile legs for clinging on substrate, especially on plant surfaces. Contrary to Tisbe spp., H. uniremis and Z. aurelii were often observed on the surface of eelgrass blades. With those cautions in mind, the electivity index was computed for each taxon of the H + Z+T assemblage (Fig. 3.8). In 1985, the electivity index for H. uniremis and Z. aurelii was always positive, and constantly very near 1. This suggests a very strong preference of the fish for those species. At the opposite, the electivity index for Tisbe spp. in 1985 was always negative, and strongly so most of the time. In 1986, the electivity indices were somewhat different. The electivity index of H. uniremis was still always positive, although not as strongly as in the previous year. The electivity index of Tisbe spp. was also negative throughout, albeit less strongly than in 1985. The sharpest difference in electivity index between 1985 and 1986 was observed with Z. aurelii: in 1986, the electivity index was negative (very close to -1) in early season and then shifted to positive (as throughout 1985) in later season. Globally, the electivity index for the 3 taxa over the 2 sampling periods indicated • a strong preference for H. uniremis; • a strong avoidance of Tisbe spp.; • a least a period of strong preference for Z. aurelii in both years with a period of avoidance in 1986. 41 Legend A Tisbe  O Zaus  • Harpocticus 1985 1-i • B a> > 100 120 140 160 180 Time (calendar days) | i & i i |Ai i i I 140 160 180 100 120 Time (calendar days) Figure 3.8: Electivity index for Tisbe spp., Harpacticus uniremis, and Zaus aurelii in 1985 and 1986. 42 Harpacticoid copepods can the be considered as the main food items as well as the most important food items of juvenile chum salmon (sensu Berg 1979). The shift from avoidance to preference for Z. aurelii in 1986 and its absence in 1985 could be interpreted as follows. In 1985. the abundance of H. uniremis was always very low (data in Appendix B) and the electivity index for Z. aurelii was always high. In 1986, the period for the negative electivity index of Z. aurelii is coincident with an abundance of H. uniremis much higher than the previous year. It might be that when H. uniremis is available in quantity, anything else is disregarded by the juvenile chum salmon. The reasons for such a possibly high preference for H. uniremis over any other food item might be its dark coloration (green to dark green), its large size (up to 1.5 mm), or any other yet to be identified behavioral idiosyncrasies (e.g. Marcotte 1984). 3.5 Chapter Summary Juvenile chum salmon were most abundant in the vicinity of Roberts Bank in late May or early June. Juvenile chum salmon captured ranged in size from 33 to 60 mm (means of a sample of fish from each fishing effort). The chief food items of the juvenile chum salmon were harpacticoids, mostly H. uniremis, Z. aurelii, and to a lesser extent, Tisbe spp.; the electivity index for the first species was always positive and high; that for the second species varied from low negative to high positive perhaps as a result of fluctuations in absolute abundance of H. uniremis; the electivity index for Tisbe spp. was always negative, and generally strongly so. Horizontal heterogeneity and/or particular harpacticoid behavior need further investigation to state more clearly why such patterns of electivity were observed. 43 C hapter 4 Spat io -Tempora l Distr ibut ion of Natant Harpact ico id Copepods 4.1 Introduction In chapter 2, the description of the spatio-temporal distribution of each compo-nent of a trophic structure was recognized as a prerequisite for the identification of environment-coupled mechanisms that may regulate fish population abundance. In chapter 3, the spatio-temporal distribution of juvenile chum salmon was discussed and the preferred food items of the fish were identified as being harpacticoid cope-pods. In this chapter, the spatial (vertical) and temporal (seasonal) distributions of harpacticoid copepods on which juvenile chum salmon heavily rely as a food source, are described. The following summary of the ecology of harpacticoid copepods is extracted from the comprehensive review provided by Hicks & Coull (1983). Harpacticoida is one of seven Orders of the subclass Copepoda; it contains small copepods ranging in size from 200 to 2500 fim. Harpacticoid copepods occur in almost all aquatic habitats and are a major component of the meiobenthos, i.e. those benthic organisms passing through a 2000 pm mesh, but retained on a mesh of 40-100 /Am. Harpacticoida is primarily a benthic free-living Order. Those 44 harpacticoids closely associated with the sea bed can be separated into three groups: • interstitial • burrowing • epibenthic. Interstitial forms are usually vermiform, while burrowers are either broadened anteriorly to push sediment particles or equipped with flattened appendages to dig in muddier sediment. Epibenthic forms often have the ability to swim, and have various body shapes. Other species of harpacticoid copepods are common on aquatic macroalgae and angiosperms, and are equipped with strongly prehen-sile appendages; these epiphytic species are often adapted also for free-swimming. Globally, harpacticoid copepods comprise from 4 to 95% of the total sediment meiobenthos, and from 11 to 60% of the phytal meiobenthos. Collectively, benthic harpacticoid copepods feed on epipelic or epiphytic diatoms, phytoflagellates, bac-teria, either as aggregated cells or detritus associates, fungi and yeasts, blue green algae, mucoid substances, and ciliates. There is growing evidence that harpacti-coids may well be the most important fraction of the meiofauna as a food source for larval and juvenile fishes. Whether the reproductive activities of harpacticoid copepods are continuous or protracted as a general rule is not clearly established. Some studies have shown clear periodicity in reproduction, while others suggest a rather continuous reproductive activity. However, for particular species, or species assemblages in given environmental conditions, temporal patterns in reproductive periodicity have been identified. The female benthic harpacticoid carries its eggs in an external egg sac (ovisac) which is paired in some species. Harpacticoids possess naupliar larvae which hatch from the ovisac directly in the habitat of the adult af-45 ter an incubation period of generally 1 to 8 days. The post-embryonic development includes six naupliar stages and six copepodite stages; the sixth copepodite stage corresponds is the adult. Development rates are known to be greatly influenced by temperature and food supply. In the laboratory, egg to adult time is known to vary from 20 to 70 days, in inverse proportion to temperature. The life-span of a harpacticoid copepod appears to range from 4 to 13 months in nature. Many field studies have shown one to three annual generations to be the usual occurrence. 4.2 Vertical Distribution of Selected Species of Natant Harpacticoid Copepods Harpacticoid copepods are known to occur in the water column on vegetated tidal flats (Sibert 1981). An early attempt at quantifying the abundance of water-borne (natant) harpacticoids on a vegetated tidal flat has shown their density varies from 9 to 94 individuals 1_1 within 10 cm of the sediment surface (Sibert et al. 1977). Further, there is evidence that the vertical distribution of natant harpacticoids exhibits persistent patterns. Sibert (1981) sampled the water column in a vegetated tidal channel in the Nanaimo estuary, British Columbia. Water samples were pumped from within 5 and 30 cm of the sediment surface, and the density of animals nearer to the bottom always exceeded that at the higher level by a factor of 2 to 20. Clearly then, and as recommended by Bell et al. (1984) for tidal seagrass systems, sampling programs aimed at quantifying the abundance of harpacticoids on vegetated tidal flats should be designed with consideration of the three possible sub-habitats of the animal: • plant surface • sediment 46 • water column. In order to correctly estimate the abundance of harpacticoids in the water column, a detailed description of the animals' distribution therein is required. This section reports the results of a study on the vertical distribution of an assemblage of species of natant harpacticoids on a vegetated tidal flat. The taxa in the assemblage are Tisbe spp., Zaus aurelii, and Harpacticus uniremis. Those species were previously reported as the dominant food items in the diet of the juvenile chum salmon caught near Roberts Bank (chapter 3). 4.2.1 Methods The location of the study site and the methods of collecting natant harpacticoids have previously been described (chapter 3). The salient points to recall are that natant harpacticoids were collected with a stratified sled-sampler and that the cen-tered height above the sediment at each sampling level was 56, 156, 256, 356, and 456 mm respectively. In 1986, 3 transects were done on each sampling trip, while in 1985, only 1 transect was done on each sampling trip. All sampling was done at station H (see Fig. 3.1). The total number of adults, or near-adults (copepodites 4 and 5) of each of the 3 selected species was determined at each sampling level of each transect for every sampling efforts. The detailed taxonomic composition of the water-borne zooplankton samples is reported in Appendix B. All unidentified harpacticoid copepods (sub-adults or early copepodites) are reported as uniden-tified copepodites. Only the adult or near-adult copepodites were enumerated to species since juvenile chum salmon have been shown to feed almost strictly on adult or near-adult stages. 47 4.2.2 Results and Discussion Specific and persistent patterns were observed in the vertical distribution of the harpacticoid taxa under study (Figs. 4.1, 4.2, and 4.3). The maximum abundance of Zaus aurelii was observed at the intermediate sampling level (156-256 mm) (Fig. 4.1). On either side of this maximum, abundance decreased in more or less linear fashion. The distribution curve of Zaus aurelii in the water column was roughly pyramidal in shape throughout most of the season. The maximum abundance of Harpacticus uniremis was observed at the lowest sampling level (56 mm) for half of the sampling period (days 91, 105, and 119) and at the intermediate level (156 mm) for the remainder of the time (Fig. 4.2). The abundance of Harpacticus uniremis was not consistently decreasing with height above sediment after the level of maximum abundance had been reached (days 78, 91, and 147). The maximum abundance of Tisbe spp. was observed at the lowest sampling level (56 mm) except for the last sampling efforts of the season when their absolute abundance was low (days 132 and on) (Fig. 4.3). When the abundance of Tisbe spp. was maximal at the lowest sampling level (days 55 to 119), the abundance of animals at each suc-cessive level was always lower than that at the previous. The decline in abundance from one sampling level to the next was faster at lower levels. Throughout most of the sampling period, the distribution curve of Tisbe spp. in the water column appeared to have the characteristics of a negative exponential function. The number of adults of the 3 selected harpacticoid taxa per transect (sum of animals in the 5 sampling baskets) varied little from one transect to another at any sampling date (Fig. .4.4). This low variance in the number of harpacticoids per transect for the assemblage as a whole is assumed to be representative of the variability in number between corresponding levels of transect. The specific com-48 D D NI 3 o -Q E Z J 100 H 90 80H 70 60-50-40-30-20-10-0 Legend A 678 O d91 0. O d1105 O d_119_ ^ • d_132_ \ A \ \ " * \ 0 ^ \ ^ \ o 11 pTn 11 \ i i"p 1 1 1 j 1 1 1 1 1 1 1 1 1 1 1 I * T I ~ P r f t 1 1 1 1 0 50 100 150 200 250 300 350 400 450 500 Centered height above sediment (mm) Figure 4.1: Vertical Distribution of Zaus aurelii at station H in 1986. 49 30 - i 3 25 "o D Q- 20 D D D 0) E 15 H 10-5H Legend \ \ A d 78 O d 91 • d_105 O d 119 • 1111111111111111111111111111111111111111 iTTft 1111 0 50 100 150 200 250 300 350 400 450 500 Centered height above sediment (mm) Figure 4.2: Vertical distribution of Harpacticus uniremis at station H in 1986. 50 J Q CO D D E <D 3 O 0) -Q E 700 n 600 -j 500-400 300-200 » 100-0 0 700 q 600 -f 500-j 400-j 300 200 -j 100-0 Legend A d_55 O d64 \ \ \ \ G - \ \ O dI 78 O d_91_ • d105 1111 i 1 1 1 1 I 1 1 " I 1 1 1 '7""""!^  l I r | i i i~i p iT ~fi*l TTJBr - r - rq 50 100 150 200 250 300 350 400 450 500 Legend • _dJ19_ O d_132^  _ O d 147 V O d162 \ *~^-~-B i » i i P i i i | i i i i p i i i | i i i i p u n i i i i pTi,, rrr ; i i i i i p i i i | i i i i p i i i | i i i i p i M | 0 50 100 150 200 250 300 350 400 450 500 Centered height of basket above sediment (mm) Figure 4.3: Vertical distribution of Tisbe spp. at station H in 1986. 51 2000 - i Time (calendar days) Figure 4.4: Mean number of the harpacticoid species assemblage H + Z + T per transect at station H in 1986 (mean of 3). Error bars are +/- 1 S.E. o f mean; n=3. 52 position of the harpacticoid assemblage also varied very little between transects at any sampling date (Fig. 4.5). For the 3 harpacticoid taxa considered, a maximum of abundance in the water column was bracketed with the sled-sampler. This sug-gests that the sled-sampler extended sufficiently in the water column. The patterns observed in the vertical distribution of the selected taxa of harpacticoids are then reasonably assumed to be representative of the bulk of the population. Once their vertical distribution in the water column is described, it is possible to estimate the total abundance of natant harpacticoids by computing the area under their vertical distribution curves. The distribution curves of Zaus aurelii and Harpacticus uniremis (Figs. 4.1 & 4.2) were closed by extending straight lines fitted through the first two and last two observed data points respectively. Intercepts (on the vertical or horizontal axis, depending) were derived from those fitted lines to add closing intervals on each distribution curve (Tables 4.1 & 4.2). The abundance of Zaus aurelii and Harpacticus uniremis in the water column was computed by numerically integrating the closed distribution curves as follows: T^TV.-L, (4.1) t=i where i = interval number, TV, = mean number of animals in interval i, and L, = length of interval (mm). For Tisbe spp., the rate of change of the animal abundance in the water column was assumed to be described by ^ = - * * M (4.2) where N =number of animals, k =decay coefficient, and z =height above sediment. Equation 4.2 proved to be a reasonable assumption. The abundance of Tisbe spp. 53 100-90-80--c o 70-^ -60-o CL E 50-o o JO 40--• o 30-Q) Q. 20-10-o4 Legend A Tisbe • Zaus ^ ° H.9rP.9.c.f]?.u_? f A* A ffirl^" i f f i | i i i | i i i | i i . i . . • 40 60 80 100 120 140 160 180 200 Time (calendar days) Figure 4.5: Specific composition of natant harpacticoids in the H+Z+T assemblage at station H on Roberts Bank during sampling period in 1986. Error bars are +/-1 S.E. of the mean; n=3. 54 D a t e S a m p l i n g level height (number of Zaus aurelii) 78 10.5(0) 56(5) 156(16) 256(17) 356(10) 456(1.3) 471(0) 91 19.6(0) 56(12) 156(45) 256(17) 356(10) 456(1.3) 567(0) 105 0(2) 56(16) 156(41) 256(73) 356(40) 456(13) 504(0) 119 0(70) 56(78) 156(41) 256(82) 356(59) 456(22) 515(0) 132 0(2) 56(2) 156(2) 256(2) 356(1) 456(1) 556(0) Table 4 .1 : Sampling level heights and corresponding abundance of Zaus aurelii (in brackets) for various sampling efforts (date in calendar days). Date Sampling level height (number of Harpacticus uniremis 78 1.9(0) 56(2.3) 156(3) 256(2.6) 356(2) 456(0.2) 467(0) 91 0(4.7) 56(5) 156(5.6) 256(2) 356(3.6) 456(2) 581(0) 105 0(19.7) 56(15) 156(19.3) 256(12) 356(12) 456(5.6) 543(0) 132 0(22.7) 56(20.3) 156(16) 256(12.3) 356(4) 456(2) 556(0) 147 0(0.21) 56(4.3) 156(8.3) 256(5) 356(3.3) 456(1.6) 550(0) Table 4 .2 : Sampling level heights and corresponding abundance of Harpacticus uniremis (in brackets) for various sampling efforts in 1986 (date in calendar days). 55 at each sampling level was regressed versus the corresponding centered height of the basket (Z) for various sampling efforts according to to this model, the total abundance of Tisbe spp. in the water column is equal to: TV However, due to the finite size of the opening of the sampling baskets on the sled-sampler, the intercept iV 0 in eq. 4.3 is somewhat biased. This bias introduced in the computation by relating the number of harpacticoids in a basket to the corresponding centered height must be evaluated to assess the accuracy of the estimates of abundance of the animals. The volume intersected by a cylinder (centered at height Z and with radius a) through a density surface described by dN(z)/dz = — kN with true intercept JVJ, is equal to ( Fig. 4.6): \oge N{Z) = \ogeN- kZ which provided a good fit for most of the sampling efforts (Table 4.3). According 56 iV(Z)7ra 2 { l + (fca)2 + •••} 8 which finally reduces to NX {1 + [ka)2 •+•••} , for area na2 . 8 With consideration of the two leading terms in the above expansion, the relation-ship between the biased intercept iV0 (eq. 4.5) and the true intercept N* can be written as The corrected expression for the computation of the abundance of animals in the water column is then written as The absolute number of harpacticoid copepods in the water column was re-shaded by the sled-sampler (= length of transect (1960 cm) x diameter of baskets (8.3 cm) =16848 cm2) (Table 4.4). The mean number of each harpacticoid taxon per transect was determined for each sampling effort, which provided an index of abundance by which to determine the temporal patterns of abundance of the an-imals. The temporal patterns observed with the index of abundance series were then compared to those observed with the absolute abundance series determined by integration of the distribution curves; for Tisbe spp., the biased and corrected estimates of absolute abundance were also compared (Figs. 4.7, 4.8, and 4.9). The highest abundance of Zaus aurelii in the water column (1.96 individuals /cm2) was observed on on April 29 (day 119) (Fig. 4.7). The early and late season N0 1 + M ! * ported in individuals/m2 by dividing the computed amount by the bottom area 57 Date N0 k R2 42 112 -0.0099 0.96 55 73 -0.0098 0.87 64 249 -0.0075 0.97 78 782 -0.0072 0.96 91 426 -0.0035 0.96 105 1196 -0.0069 0.97 119 390 -0.0049 0.96 132 49 -0.0017 0.64 Table 4.3: Intercept (N0), decay coefficient (k), and coefficient of determination (it!2; n — 5) of the exponential distribution curves of Tisbe spp. in the water column (date in calendar days). Date Tisbe spp. Z. aurelii H. uniremis N0/k #/cm2 r E i #/cm2 E . 41cm2 42 1145 0.7 55 7521 0.5 64 33240 2.0 78 107850 6.6 4731 0.28 949 0.05 91 120807 7.4 20183 1.2 1867 0.11 105 174186 10.7 17666 1.05 3289 0.19 119 79312 4.9 33053 1.96 7839 0.46 132 28383 1.7 912 0.05 5539 0.33 147 2453 0.14 Table 4.4: Total abundance of Tisbe spp., Zaus aurelii, and Harpacticus uniremis in the water column at station H during the sampling period in 1986. 58 N Figure 4.6: Volume intersected by a cylinder of radius a through a density surface described by dN(z)/dz = -kN(z). Note that the horizontal axis has the dimension of N, not of length. 59 Legend O TOTAL ABUNDANCE 70 90 110 130 150 Time (calendar days) Figure 4.7: Absolute abundance and index of abundance of Zaus aurelii in the water column during the sampling period in 1986 at station H. 60 Legend O TOTAL ABUNDANCE Figure 4.8: Absolute abundance and index of abundance of Harpacticus uniremis in the water column during the sampling period in 1986 at station H . 61 Legend O BIASED ABUNDANCE • INDEX OF ABUNDANCE 40 60 80 100 120 140 160 180 Time (calendar days) Figure 4.9: Absolute abundance (biased and corrected) and index of abundance of Tisbe spp. in the water column during the sampling period in 1986 at station H. 62 minima of abundance were 0.28 individuals/cm2 on March 19 (day 78) and 0.05 individuals/cm2 on May 13 (day 134). The temporal patterns in the abundance of animals in the water column observed in the computed absolute abundance series closely paralleled those observed in the index of abundance series. The max-imal abundance of Harpacticus uniremis (0.46 individuals/cm2) was also observed on April 29 (day 119) (Fig. 4.8). The early and late season minima were 0.05 individuals/cm2 on March 19 (day 78) and 0.14 individuals/cm2 on May 27 (day 147). Again, there was a close match between the temporal patterns in the total abundance series and in the index of abundance series. The maximal corrected abundance of Tisbe spp. (10.6 individuals/cm2) was observed on April 15 (day 105) (Fig. 4.9). The early and late season minima were 0.5 individuals/cm2 on February 24 (day 55) and 1.7 individuals/cm2 on May 13 (day 132). The biased absolute abundance was only slightly different (less than 1%) from the corrected absolute abundance (biased maximum abundance: 10.7 individuals/cm2; corrected maximum abundance: 10.6 individuals/cm2). This indicates that the size of the filtering baskets on the sled-sampler (8.3 cm internal diameter) was appropriate for the decay coefficients of harpacticoids in the water column (k = —0.006 to —0.011) (Table 4.3). By relating the number of animals in a basket to the correspond-ing centered height above sediment to parametrize the exponential decay curve, it was implicitly assumed that within the opening of the basket, the decay curve was linear. This assumption was required to obtain an analytical expression of the decay curve. The small size of the truncation error introduced in the computations suggest that this assumption is reasonable. Similar patterns in the temporal abun-dance of Tisbe spp. in the water column were observed in the computed absolute abundance series and in the index of abundance series. For the 3 harpacticoid taxa 63 under study, a simple index of abundance (number of animals/transect) appears lo be sufficient to track the temporal patterns in the abundance of the animals. However, as illustrated in the case of Tisbe spp., small differences between the index of abundance and the absolute abundance series can lead to substantially different speculations. Between March 19 (day 78) and April 1 (day 91) (Fig. 4.9) the index of abundance of Tisbe spp. remains the same, suggesting a plateau. One might propose that this plateau could result from a sudden predation pulse on the harpacticoid population. However, the observed temporal patterns of abundance leading to this proposition are less obvious when considering the computed abso-lute abundance data: during the same period extending from days 78 to 91, the computed absolute abundance does not indicate such a flat plateau as the index of abundance does. As it is, the salient feature in the absolute abundance during this period is a decrease in the decay coefficient k (from -0.0072 to -0.0035) (Table 4.3). 4.2.3 Discussion In various estuaries, harpacticoid copepods have been recognized as a dominant food item for juveniles of commercially important fishes (Feller & Kaczynski 1975; Sibert 1979). However, some difficulties arise when interpreting the relationships between predators and prey due in part to a lack of knowledge of the local prey distribution (Sibert 1981). Patterns in the distribution of natant harpacticoids such as those demonstrated here could help in the investigation of the trophic relationships between predatory juvenile fishes and harpacticoid copepods. Fur-ther, the impact on the global dynamics of estuaries of such trophic relationships should only be assessed with due consideration of the total abundance of natant 64 harpacticoids. Healey (1979) submitted that the salmonid-harpacticoid trophic re-lationships had little effect on the dynamics of the Nanaimo estuary as a whole; this conclusion was reached with considering only the abundance of harpacticoids in the sediment, and many of the harpacticoid prey species were those that have shown to be substantially abundant in the water column in this study. The sampling technique used to collect harpacticoids in this study appears to be appropriate. The simplicity of use of the sled-sampler makes it a convenient field tool; importantly, information obtained on the localized distribution of natant harpacticoids can provide estimations of absolute abundance. Studies reporting ab-solute abundances of animals rather than index of abundance could be easier to compare, and provide a more detailed basis for the understanding of the tropho-dynamics of estuaries. One technical point remains to be clarified with respect to the sled-sampler. As pointed out earlier, the flow of water in each sampling basket seemed to be less than ideal probably because of the high aspect-ratio of the sieve and the elementary design of the basket. For this reason, the estimations of absolute abundance of natant harpacticoids are probably biased towards lower values. Probably balancing this underestimation of abundance is the contamina-tion of the natant harpacticoid samples by epiphytic animals scraped off eelgrass blades by the sled-sampler. This contamination problem remains to be investigated with a micro-scale sampling approach. Despite these difficulties, evidence for the existence of specific localized patterns in the distribution of natant harpacticoids is presented. A theoretical basis for the assessment of the derived computations of total abundance is developed. This also provides an objective basis for choosing an appropriate sampling basket opening. Indeed, as pointed out earlier, studies on harpacticoid copepods on vegetated tidal flats should take into account the 3 65 potential sub-habitats of the animals; further, such studies should as well take into account the specificity of the distribution patterns of each species or assemblages of species of harpacticoid copepods in the water column. 4.3 Thermal Regime and Seasonal Abundance of Selected Species of Natant Harpacticoids The reproductive cycle of most temperate and boreal invertebrates is decisively affected by temperature (Kinne 1970). A relationship between the seasonal abun-dance of marine invertebrates and the annual temperature cycle is then expected. Evidence for the existence of such a link between temperature and invertebrate abundance has been reported for meiofauna on various tidal flats, especially for some groups of harpacticoid copepods (Muus 1967; Ito 1971; Harris 1972; Jewett & Feder 1977; Coull 1985). In a most general sense, these studies report that the density of certain species or assemblages of species of harpacticoid copepods tend to fluctuate in relation to temperature. Whether the link between animal abundance and temperature has a physiological or trophic basis is yet to be es-tablished. For harpacticoids, the mechanism responsible for the apparent coupling between thermal regime and seasonal abundance is thought to be the effect of temperature variations on development time (Heip & Smol 1976). However, a con-comitant increase in food supply has been suggested as a possible alternative or complementary mechanism (Hicks 1979; Coull 1985). Irrespective of the nature of the mechanism(s) involved, the observed relationship between temperature and abundance of some harpacticoid copepods is described in the broadest of terms. Possible reasons for this broadness are the spottiness of the temperature data against which the seasonal abundance of the animal is compared, and the variation 66 in observational methodologies. Any identification of the fine temporal patterns in the thermal regime is difficult-only the broad patterns are obvious, from which an equally broad definition of the link between animal abundance and thermal regime is derived. Variability of temperature on a wide range of temporal scales is a well de-scribed environmental characteristic of temperate tidal flats (Vugts & Zimmerman 1975,1985; Harrison 1984,1985; Harrison & Phizacklea 1987). These detailed stud-ies of temperature on tidal flats indicate that the thermal regime of such areas is characterized by periodic and aperiodic fluctuations as a result of interactions between radiative heating and cooling, and sensible and latent heat exchange with overlaying air and water. It is suggested in the same studies that temperature could be a prime environmental factor in the ecology of the fauna associated with the tidal flats. Ecological studies detailing the seasonal abundance of harpacticoid copepods in relation to temperature on tidal flats could benefit from standardized thermometric techniques providing a more detailed description of the thermal regime. Conversely, detailed studies of the thermal regime on tidal flats merely suggest the potential relevance of such knowledge for an understanding of the ecology of those areas. In this section, the results are reported of a study of the seasonal abundance of an assemblage of harpacticoid copepod species in relation to the thermal regime on a tidal flat. The target assemblage of harpacticoids is composed of 3 taxa (Tisbe spp., Zaus aurelii, Harpacticus uniremis) that have been recognized as dominant food items in the diet of juvenile chum salmon. The abundance of harpacticoid copepods was monitored on a time-scale recognized as appropriate in such studies seeking to identify seasonal patterns of abundance (Montagna et al. 1983). The 67 thermal regime in the sediment was monitored on a high resolution time-scale. 4.3.1 Material and Methods The study site and the technique for collecting natant harpacticoids have previously been described in chapter 2. A time-series of the abundance of the selected species of harpacticoids in the water column was obtained by taking the mean number of adult harpacticoids per transect (n=3 in 1986). As discussed in the previous section, the index of abundance (number of animals/transect) appears appropriate to identify seasonal patterns of abundance: the temporal patterns are the same as those identified in the absolute abundance series. The reason for choosing the index of abundance to be analyzed in relation to temperature instead of the computed absolute abundance is that the computation of a standard error is straightforward with the index of abundance. Temperature in the sediment was recorded with Peabody- Ryan© thermographs from February 15 to July 8 in 1985, and from February 19 to July 9 in 1986. Each thermograph was encased in a protective PVC© tube; the tube was cribbled with holes to allow water percolation. The upper extremity of the protective casing did not extend beyond the tip of the thermograph sensor; a 1 kg lead weight was attached to the lower end of the casing. At low tide, the thermographs were buried in the sediment, sensor up. The thickness of the sediment above the tip of the probe was recorded after leveling the replaced sediment with the surrounding sur-face. Further checks during the season proved that the thickness of sediment above the probe remained constant. In 1985, 1 thermograph was successfully moored at station H; the tip of the probe was covered by 5 cm of sediment. In 1986, two Peabody-Ryan 402-6 6th Street South P.O. Box 599 Kirkland, WA 98033, U.S.A. 68 thermographs were successively moored at station H; the probe of one was covered by 1 cm of sediment, and that of the other, by 5. Upon retrieval of the thermo-graphs, temperature data were visually extracted from the recording chart paper. Data were read at the nearer 0.5 °C at every 4 hours of recording. Temperature time-series were then subjected to a first-difference filter: where T = temperature and i=time. This filter removes the lower frequency components from the signal, emphasizing the higher frequency components (Jenkins & Watts 1968). Due to reading uncer-tainty in each temperature datum, and to emphasize the warming of the sediment, only temperature differences (A ( ) greater than or equal to 0 were retained. Sediment pigment concentrations were measured as recommended by Parsons et al. (1984 b). Sediment samples were collected at station H at high tide by a SCUBA diver: 9 cores were taken at random in a 100-cell 0.25 m 2 quadrat. Cores were collected with syringes (5.31 cm 2 opening). Loaded coring devices were brought to the tending boat, and sediment was extruded so as to retain only the top cm; the retained sediment fraction was transferred to a plastic bag and kept on ice in the dark. In the laboratory, samples were frozen for at most 10 days before processing. Frozen samples were transferred to a mortar and ground for 5 minutes in 30 ml 90% (v/v) acetone with a few drops of a suspension of M g C 0 3 . The sediment-acetone slurry was transferred to a graduated plastic centrifuge tube and the total volume made up to 50 ml with additional 90% acetone. The samples were left for two At = Tt with A t if A t > 1 0 otherwise, 69 hours in the dark at 5 °C. After this period, the samples were centrifuged @ 800 g for 10 minutes. The volume of the sediment plug was recorded after centrifugation. A fraction of the extract was transferred to a 1 cm path cuvette. Extinction was measured against a 90% acetone blank with a spectrophotometer (Bausch & Lomb Spec tron ic© 2000) at wavelengths of 6650 and 7500 A. For the determination of phaeophytin a concentrations, the samples were acidified by adding a few drops of 10% HC1; the cuvette was stopped with a tin foil and lightly shaken for 30 s, and the extinction remeasured at 6650 and 7500 A. The concentrations of chlorophyll a and phaeophytin a were calculated by the equations of Parsons et al. (1984): • ,, / / 3x 26.7(6650o - 6650a) x v Chlorophyll a[mg/m ) — — 26.7(l.7[6650o] - 6650o) x v Phaeophytin a{mg/m ) = V~l where 6650o and 6650a are the extinctions at 6650 A before and after acidification (corrected for extinction at 7500 A), v is the volume of acetone extract (ml), V is the volume of the sediment plug (1), and / is the path length of the cuvette (cm). The pigment concentrations were reported in g/m 2 following: Pigment(g/m2) - mg/m3(1883 x V) , where 1883 is the ratio between 1 m 2 and the surface opening of a coring device (5.31 cm 2). The number of ovisacs in the bottom level (centered height above sediment: 56 mm) of the sled-sampler was determined for each transect of each sampling effort in 1986. The lowest sampling level was assumed to be representative of the population as a whole since the study on vertical distribution had revealed that 70 the bulk of the animals were concentrated near the bottom. It was recognized that some species are not concentrated near the bottom (e.g. Zaus aurelii), but the dominant taxon was bottom-concentrated (Tisbe spp.), and the egg masses were not identified to the species. The amount of water covering sampling station H relative to the predicted tide height at the port of reference (Point Atkinson) was measured during flowing tide on May 24, 1986. The water level was measured by observing a graduated stake installed at station H, from the shoreline. 4.3.2 Results and Discussion Natant Harpacticoid Abundance The abundance of natant harpacticoid copepods showed a pronounced seasonality (Fig. 4.10). In 1985, the abundance of natant harpacticoids (number/transect) ranged from a winter low of 254 to a spring high of 1131. Shortly after the spring high, the assemblage nearly disappeared; secondary small peaks were then ob-served. In 1986, the natant harpacticoid abundance ranged from a winter low of 77 to a spring high of 1765; secondary small peaks were also observed. The relative abundance of the 3 harpacticoid taxa in the assemblage under study was deter-mined for both sampling years (Table 4.5). In 1985, Tisbe spp. was the dominant taxon for most of the sampling period; Zaus aurelii and Harpacticus uniremis only made up to a few percent of the assemblage. In late spring, Tisbe spp. declined and the proportion of Zaus aurelii increased, but the total abundance was low (Fig. 4.10). The relative abundance of each taxon was different in 1986 (Table 4.5): Tisbe spp. dominated by far in early spring; Zaus aurelii became the domi-nant species later in the season as its absolute abundance increased when that of Tisbe spp. declined. In mid-season, Harpacticus uniremis became relatively abun-71 2000 40 60 80 100 120 140 160 180 200 Time (calendar days) Figure 4.10: Seasonal abundance (individuals/transect) of natant harpacticoids at station H in 1985 and 1986 (for 1986: mean +/- 1 S.E., n=3). 72 Time 1985 1986 T Z H T Z H 42 96 4 0 55 99 1 0 60 99 1 P 64 98 2 P 72 98 1 1 78 94 5 1 88 99 P P 91 81 18 1 100 98 1 1 105 88 10 2 113 99 0 1 119 62 32 6 129 98 1 1 132 73 3 24 143 99 0 P 147 79 9 12 158 94 6 P 162 46 27 27 172 65 10 25 176 5 94 1 186 P 66 33 Table 4.5: Relative abundance (%) of Tisbe spp. (T), Zaus aurelii (Z), and Harpacticus uniremis (H) in the natant harpacticoid assemblage in 1985 and 1986 at station H (for 1986: mean of 3 transects). Time in calendar days; p=present, but less than 1 %. 73 dant. The variance in the number of natant harpacticoids among the 3 transects for each sampling effort was very low in 1986 (Fig. 4.10). It was assumed that this variance was similarily low in 1985, for which the data of only one transect for each sampling were available. The natant harpacticoid time-series of abundance for both sampling years are both characterized by a period of most rapid increase in March (calendar days 60 to 90). In 1985, the steepest increasing section of the natant harpacticoid abundance curve is located between March 13 and 29 (days 72-88) and between March 5 and 19 (days 64-78) in 1985. Sediment Temperature The thermal regime in the sediment at the study site exhibited temporal patterns on various time scales (Figs. 4.11, 4.12, and 4.13). In 1985, the temperature in the sediment ranged from 4 °C in winter to 15 °C in early summer (Fig. 4.11); in 1986, the temperature in the sediment ranged from 3 °C in winter to 15 "C in early summer (Figs. 4.12 & 4.13).The range in temperature for both years closely matches the reported annual temperature range of the surface waters in the Strait of Georgia (5-18 °C) (Harrison et al. 1983). Superimposed on this seasonal increasing trend in mean temperature are sharp daily fluctuations apparent for both years. The amplitude of those daily fluctuations increased throughout the season. When comparing the two temperature time-series obtained in 1986 (Fig. 4.12, recording depth: 1 cm; Fig. 4.13, recording depth: 5 cm), a dampening of features is apparent. The temperature data collected deeper in the sediment showed consistently lower amplitudes in daily fluctuations: around March 1 (day 60), temperature fluctuations were up to 3 °C at recording depth of 1 cm, but no more than 1 °C at recording depth of 5 cm. When comparing time-series for 74 25-, 0 { i i i | i i i | i i i | i i i | i i i | i i i | i i i | i i i | 40 60 80 100 120 140 160 180 200 Time (calendar days) Figure 4.11: Temperature in the sediment at station H during sampling period in 1985. Depth of recording: 5 cm. 75 25-, P 20 A 15 Q. E E TJ Q) CO 40 60 i—r i | i i i | i i i | i i i j i i i | i 80 100 120 140 160 Time (calendar days) • i • • • i 180 200 Figure 4.12: Temperature in the sediment at station H during sampling period in 1986. Depth of recording: 1 cm. 76 25 -i 0 "j | i ' ' | i ' i | i i i | i i i | i i i | i i i | i i i | 40 60 80 100 120 140 160 180 200 Time (calendar days) Figure 4.13: Temperature in the sediment at station H during sampling period in 1986. Depth of recording: 5 cm. 77 different years, but for similar recording depth (Figs. 4.11 & 4.13: recording depth: 1 cm) daily temperature fluctuations are of similar amplitude (ca 1 °C for the first two weeks of recording). The depression in the mean temperature appearing between March 2 (day 60) and March 12 (day 70) in 1985 (Fig. 4.11) is coincident with a colder period: during those days, minimum daily air temperature was 2 to 5 °C lower than for the same days in 1986 (Environment Canada Weather Records). Pigments and Ovisacs The chlorophyll a concentration (mg/m2) in the sediment in 1985 went from a winter low of 27 (March 2 (day 60)) to a spring high of 57 (April 11 (day 100)); it then reached a low of 29.9 (May 10 (day 129))and leveled off around ca 45 for the remainder of the sampling period. The phaeophytin a concentration (mg/m2) went from a winter low of 54 (March 15 (day 73)) to a spring high of 112.4 (May 28 (day 157)) and remained in this range for the remainder of the sampling season (Fig. 4.14). The number of ovisacs at level A of the sled-sampler (centered height above sediment: 56 mm) ranged from a winter low of 15 (March 20 (day 78)) to a spring high of 282 (May 3 (day 132)) (Fig. 4.15). The distribution of ovisacs was bi-modal, with a first peak of 217 on April 15 (day 105) and a second one of 282 on May 12 (day 132), with an in-between low of 133 on April 29 (day 119). Tide level At the date of recording (May 24, 1986), the level of water on the station started to rise at 1445 H (PST) (Fig. 4.16). For this time, the predicted tide level at the port of reference was calculated to be 1.6 m. 78 130-i 120 A 20-10-J I I I I I — I — I — I — I — I — I — | — i — | — i 50 60 70 80 90100110120130140150160170180190200 Time (calendar days) Figure 4.14: Chlorophyll a and phaeophytin a concentrations in the sediment at station H in 1985 (mean +/- 1 S.E., n=3). 79 340-320-300-280-260^ 240 : 220 : 200^ 180-160 : 140-120-100 : 80 : 60-40 : 20-60 T—r T " 80 i—i—r T i—i—r T T—i—r T T—i—r T 100 120 140 160 Time (calendar days) • i 1 » « i 180 200 Figure 4.15: Number of harpacticoid ovisacs at level A of the sled-sampler in 1986 (mean +/- 1 S.E. n=3) (n.b. in some cases the error bars are contained within the size of the symbol). 80 Figure 4.16: Water level over station H during flowing tide on May 24, 1986. 81 Discussion The aim of this section is to describe the seasonal abundance of natant harpacti-coid copepods in relation to the thermal r eg ime in the sed iment . B y comparing the d a t a of a n i m a l abundance and of the t h e r m a l regime on the t i d a l flat (F ig s . 4.10, 4.11. 4.12, and 4.13), the b r o a d conc lus ion can be made tha t the a n i m a l abundance and the t e m p e r a t u r e w i t h i n the sed iment t end to f luc tua te s i m i l a r l y , at least for the period of increase in animal abundance. However, such a description of the link between animal abundance and thermal regime is of limited use for further ad-vancement due to the broadness of its formulation. While this description suggests a potential physiological or trophic basis for this link, it would be difficult to verify its validity in the field. The reason for this is the asymmetry in the time-scales of the two events related: the spring increase in abundance of the harpacticoid assem-blage occurs over a brief period while the seasonal increase in temperature in the sediment is a long, progressive event. If the harpacticoid blooming period happens anywhere between late-winter and mid-summer, it will always be coincidental with an increase in mean temperature in the sediment. Defining the harpacticoid in-crease in relation to the increase in mean temperature is then potentially heuristic on physiological or trophic grounds, but of little use in term of prediction. In this study, the thermal regime in the sediment was monitored on a fine temporal scale (1 measurement every 4 hours). A description of the natant harpacticoid seasonal abundance in relation to fine-scale temporal patterns in the thermal regime in the sediment is then possible. Daily temperature fluctuations are apparent throughout the sampling period for both years (Figs. 4.11, 4.12, and 4.13). Only for their amplitudes, all 3 temperature time-series appear to have the same variability structure. The following detailed 82 description of high frequency variations of temperature in the sediment is derived from the data recorded nearer to the surface in 1986 (Fig. 4.12), which for reasons to be discussed, show the highest amplitude of variation. From February 19 to 24 (days 50 to 55), daily downwards peaks were observed, while from March 1 and on, daily upwards peaks were observed. From February 19 to 24, lower daily tide occurred during night time; thermograph records indicate that the lower temperature values in the downwards peaks were reached when the flat was exposed during the winter night. From March 1 and on, low tides had started to occur during day time and the higher temperature values in the daily upwards peaks were reached at maximal exposure of the flat during daylight hours. This is in agreement with the results of previous studies on the thermal regime of tidal flats by Johnson (1965) and Harrison (1984). Both authors showed that when the tide falls below the level at which the flat becomes exposed during daytime, there is a heat gain in the sediment from solar radiation; when this exposure happens during night, the magnitude and direction of temperature changes in the sediment is dependent upon the atmospheric conditions (air temperature and wind) (Fig. 4.17). Another characteristic of the temperature records at station H is that the amplitude of the daily fluctuations itself has a periodicity. This is best illustrated by comparing the series of temperature peaks around March 1 (day 60), March 16 (day 75), and March 31 (day 90) (Fig. 4.12): a maximum in amplitude of the daily temperature variations is reached ca every 15 days. This fortnightly cycle has been described in detail by Vugts <k Zimmerman (1975,1985) and Harrison (1985). These authors showed theoretically and experimentally that on M2-dominated temperate tidal flats, the precession of the timing of the tide relative to the solar day gives rise to a beat in both the mean and amplitude of 83 Figure 4.17: Potential effect of time of high water on the magnitude and direction of change in sediment surface temperature. The arrows indicate the direction and magnitude of temperature changes in the sediment (redrawn from Harrison (1984)). 84 temperature fluctuations in the water and in the sediment: this beat has a period of 14.72 d. One crucial observation in the temperature data from station H is that this fortnightly beat in mean and amplitude of temperature fluctuations only starts to become visible as sharp daily peaks at a specific time during spring. As the season progresses, the beat becomes more and more obvious, although later in the season, aperiodic fluctuations probably related to local atmospheric conditions tend to blur the signal. This increase in the daily temperature peaks can be related to the gradual increase in coincidence of the lower low waters and the daily solar zenith. In other words, the precession of the timing of the tide relative to the solar day, which is continuous, appears as an apparent fortnightly temperature beat only when lower low waters start to occur during day time. This beat in temperature, then, is a seasonal phenomenon and has a specific timing of onset. The shift of timing of lower low tides from night in winter to day in spring was recognized in the Strait of Georgia by Waldichuk (1957). This author also pointed out that the actual pattern of the shift itself is a transitory feature in the Strait: this shift varies with its own period of slightly more than 900 y. The above description of temperature variability is directly applicable to the temperature records from 1985 (Fig. 4.11) and 1986 (deeper recording level)(Fig. 4.13) that only differ by degrees of intensity. The amplitude of temperature vari-ations recorded deeper in the sediment (Figs. 4.11 & 4.13) is smaller than that of temperature variations recorded nearer to the surface (Fig. 4.12). This dissipation of features in temperature records with increasing depth of recording in the sedi-ment was recognized in land soil temperature analysis (Carson 1961) as well as in previous tidal sediment temperature analysis (Johnson 1965; de Wilde & Berghuis 1979). Differences in absolute values in temperature records between the two years, 85 for same recording level, are attributed to differences in meteorological conditions. All temperature time-series recorded in the sediment at station H, despite differ-ences in magnitude of amplitude caused by varying depth of recording or between year meteorological fluctuations, exhibit seasonal, daily, and intermediate-scale cy-cles; all temperature data sets also indicate the seasonality and the specificity of onset time of recurrent strong warming episodes in the sediment related to the interaction of the tide cycle and the daily heat cycle. Temperature time-series were filtered according to the previously described pro-cedures (see Methods) to emphasize the higher frequency components of variability (Figs. 4.18, 4.19, and 4.20). In 1985, 3 small individual peaks appear between February 17 (day 48) and February 26 (day 57) (Fig. 4.18). Thermograph records indicate that these peaks, although positive, do not correspond to warming episodes of the sediment. Rather, these peaks seem to be caused by the flooding of the flat with water warmer than the cooled sediment surface. A similar series of peaks is seen in 1986 filtered temperature data from February 19 (day 50) to February 24 (day 24) (Figs. 4.19 & 4.20). In both cases, the flat had been exposed during night time and accordingly cooled off; the return of relatively warmer water later in the night or early in the morning is recorded as temperature increases in the sediment. Those temperature increases not related to warming of the sediment by the sun, but instead to a return to overlaying water temperature after cooling, correspond to the daily inverted peaks already mentioned in the description of the non-filtered temperature data. When these temperature increases are over-looked, the specific timing of appearance and gradual amplification of warming events in the sediment become more obvious. In 1985 (Fig. 4.18) the first of the warming episodes in the sediment is seen as a series of pulses from March 10 (day 69) to 86 8~1 7 -6 5 4 3-2-H r**r—T 40 60 80 100 120 140 160 Time (calendar days) 180 200 Figure 4.18: High frequency temperature fluctuations in the sediment at station H in 1985 (depth of recording: 5 cm). 87 8 - i 7-Time (calendar days) Figure 4.19: High frequency temperature fluctuations in the sediment at station H in 1986 (depth of recording: 1 cm). 88 8-i 7-6--Time (calendar days) Figure 4.20: High frequency temperature fluctuations in the sediment at station H in 1986 (depth of recording: 5 Cm). 89 16 (day 75); the second warming episode, of higher amplitude, follows from March 25 (day 84) to 29 (day 88); the third warming episode, of even higher amplitude, appears from April 4 (day 94) to 12 (day 102). In 1986 (Figs. 4.19 & 4.20) similar patterns are observed: the first warming episode occurs between February 28 (day 59) and March 2 (day 61): the second one from March 13 (day 72) to 21 (day 80); and the third one from March 27 (day 86) to April 4 (day 94). After the third warming episodes of both years, apparently aperiodic pulses tend to mask the signal, although a periodicity is still apparent. The attenuation of temperature variability with depth of recording is more apparent in the filtered data time-series: temperature pulses recorded at 5 cm in 1986 (Fig. 4.20) are clearly smaller than those recorded at 1 cm during the same year (Fig. 4.19). The filter has emphasized the higher frequency components of temperature variability and those components are the most dampened with increasing depth of record as a result of the thermal inertia of the sediment. Sediment warming episodes and harpacticoid bloom The natant harpacticoid assemblage under study went through its period of maxi-mal increase from March 13 (day 72) to 29 (day 88) in 1985, and from March 5 (day 64) to 19 (78) in 1986. For both years, these blooming periods coincide with the appearance of warming episodes in the sediment (Fig. 4.21). The warming episodes have variable durations within and between years. Tide tables indicate that the observed daily temperature pulses, that make up warming episodes, only started to appear and kept appearing as long as the tide level at the port of reference (Point Atkinson) fell below 1.6 m during daytime (from 0800 H PST to 1630 H PST). This critical level is precisely the elevation of sampling station H above chart datum as 90 Legend > 1985 WE 1986 WE 1985 BLOOM 1986 BLOOM I—i—i—i—j—i—i—i—|—i—i—i—j—i—i i—j 40 60 80 100 120 Time (calendar days) Figure 4.21: Timing and duration of initial warming episodes (WE) in the sediment and of natant harpacticoid blooming periods (BLOOM) at station H in 1985 and 1986. 91 de te rmined ear l ier ( F i g . 4 .16) . T h i s suggests t ha t a cons t r a in t must be added in addition to the shift in the pattern of exposure of the flat ( lower low tide during night in winter; lower low tides during daytime in summer) in order to define the appearance and recurrence of w a r m i n g episodes in the sed iment : the lower low waters m u s t fa l l d u r i n g d a y t i m e and be low 1.6 m at the po r t of reference. T h i s c r i t i c a l level is equal to the e levat ion of the s t a t i on ; th i s suggests tha t indeed the thermal i n e r t i a of wa t e r is great and that the station has to be minimally covered to record warming episodes in the sediment. However, contrary to what the above data might suggest, station H was never observed to be totally dry, even during the lowest tides of the years; a residual slick of ca. 5 cm was always remaining before the onset of the following flood. This slick was apparently retained by the eelgrass mat. In 1985 and 1986, the warming episodes, once initiated, seemed to recur on a fortnightly basis. For both years, the first low waters below 1.6 m (port of reference) during daytime were followed by similar low waters every 2 weeks for at least the next 4 weeks. This would explain the consistent coincidence of the fortnightly lower low waters and the warming episodes. However, the warming episodes in the sediment need not be fortnightly themselves. Inspection of tide tables for years other than 1986 and 1985 indicate that the first lower low waters to fall below 1.6 m (port of reference) during daytime (0800 H - 1630 H PST) are not always followed by similar lower low waters two weeks later. Early warming episodes in the sediment could then be separated by more than 2 weeks in early spring in some years. Warming episodes in the sediment started to occur 10 days earlier in 1986 than in 1985 (Fig. 4 .21) . The onset of the natant harpacticoid blooming periods also appear to be phased by a similar period of time between the two years. The 92 temperature data were collected every 4 hours; those for the natant harpacticoid assemblage, every two weeks. A phasing of 10 days between corresponding features in the temperature time-series can reasonably be considered as significant; a similar phasing between the harpacticoid bloom onsets can not be considered significant. Nonetheless, it may suggest that the natant harpacticoid bloom could be linked with a specific (the second) warming episode of the early spring ones. The above results suggest that the link between natant harpacticoid abundance and the thermal regime in the sediment be defined in terms of the coincidence of the animal spring bloom and the onset of recurrent, although not necessarily periodic, warming episodes in the sediment. In 1986, the blooming period extended from day 64 to 78, while the first and second warming events were respectively observed from day 59 to 61, and from day 72 to 80. In 1985, the blooming period extended from day 72 to 88, while the first and second warming events were respectively observed from day 69 to 75, and from day 84 to 88. Because of the overlap of the blooming periods during both years with the second warming event, it may be suggested, albeit quite tentatively, that the harpacticoid bloom could coincide with the second of the early warming episodes. This definition of the link between harpacticoid abundance and temperature is based on variability components of the thermal regime with periods much shorter than the seasonal mean temperature trend. The similarity of the time scales of the initial warming episodes and of the harpacticoid blooming period makes the above definition more robust than one linking harpacticoid abundance with the seasonal trend in mean temperature: its applicability to other harpacticoid species or assemblages of species, other tidal flats, other times, is readily verifiable. Future research effort in this direction should aim at improving this proposed definition of the link between harpacticoid 93 abundance and thermal regime by increasing the correspondence of the sampling time scale for the harpacticoids with that for the reference signature in the thermal regime. To achieve this, the abundance of the natant harpacticoid assemblage would need to be monitored on a time-scale allowing the identification of small between year differences in blooming period onset; the co-variance between the harpacticoid blooming period and any specific warming episodes could then be more rigorously evaluated. The pattern of abundance of the natant harpacticoid assemblage after the spring bloom is not addressed in this study. As seen above (Fig. 4.10), the spring max-imum of abundance is followed by a period of most rapid decline. The factors responsible for this decline are probably varied, of both biological and environmen-tal origin, and coupled in intricate fashions. The identification and study of these factors is likely to be eased if some of the variability in the seasonal abundance of natant harpacticoids can be accounted for by specific environmental signals such as warming episodes in the sediment as suggested in this study. As pointed out earlier, the reason(s) for the existence of a link between harpacti-coid abundance and temperature in the sediment remains rather speculative. Re-sults of the study of chlorophyll a in the sediment show that the pigment concen-tration (in mg/m2) ranged from 27 to 54. Those values are intermediate between those reported by Harrison (1981) on Sturgeon Bank (32 to 186 mg/m2, calculated with 5 cm deep cores) and those reported by Bawden et al. (1973) on Roberts Bank (62.7 to 423.7 mg/m2, calculated with 1 to 2 cm deep cores). In 1985, the harpacticoid bloom occurred between March 13 (day 72) and 29 (day 88) (Fig. 4.10). In the same year, the chlorophyll a concentration in the sediment doubled between March 1 (day 60) and April 10 (day 100) (Fig. 4.14). This overlap could 94 be taken as suggesting a trophic basis for the link between temperature and an-imal abundance. However, the lack of a clear lag between pigment increase and animal increase requires a careful consideration of this overlap. A possible reason for the lack of a distinct relationship between pigment and animal abundance is that biomass of pigment could be an inappropriate measurement of the availability of food for harpacticoids: instead, rates of production could have indicated a sharp increase in productivity without being necessarily accompanied by an equivalent increase of concentration of pigments. Also, harpacticoids may feed on various food sources, and a monitoring of chlorophyll a does not account for the variability in time of other potential food sources. As summarized by Fenchel (1978), temporal patterns of abundance of meiofauna are probably largely controlled by the increase in photosynthetic production in the spring, reproductive potential of the different groups, and by predation. Obtaining more than an elementary description of the covariability of those factors and the temporal variability of harpacticoid cope-pod abundance would require an independent and very involved research program. As for the phaeopigments, a gradual increase in concentration was observed dur-ing the sampling period in 1985. Harrison (1981) observed a similar increase in phaeopigments and suggested that it might be associated with the decay of fila-mentous diatoms accumulating on the sediment surface. Similar accumulation of filamentous diatoms was observed on the sediment at station H during the course of this study, and it is proposed this might account for the increase in phaeopig-ments. However, settling of chlorophyllous material from the water column could also account for some of this accumulation of phaeopigments. In 1986, the number of ovisacs in level A of the sled-sampler increased from 15 on March 19 (day 78) to 217 on April 15 (day 105) (Fig. 4.15). In the same 95 year, the harpacticoid blooming period occurred between March 5 (day 64) and March 19 (day 78) (Fig. 4.10). This increase in the abundance of ovisacs follows the bloom and must be interpreted as the offsprings of the year-class, and not as its origin. However, the animals appearing during the blooming period must come from ovisacs. As it is, ovisacs were observed before the blooming period, albeit is small quantity. It may then be suggested that the low abundance of eggs in winter is at the origin of the spring bloom animals, and that later increase in egg number is the year-class production. Obviously then, a year-round tracking of the abundance of ovisacs would be required to state more clearly on the potential effect of temperature fluctuations on the harpacticoid development schedule. One aspect of the sampling must be discussed at this point, namely the sam-pling of harpacticoids in the water column and the measurement of temperature in the sediment. The initial reason for doing this was technical: it was obvious that a thermograph left at the surface of the sediment would collect drifting plants and become heavily fouled, which could cause serious measurements errors. Inserting the thermographs in the sediment seemed the ideal simple solution to alleviate this problem. The second reason for considering as relevant the measurement of temperature in the sediment is that the bulk of the natant harpacticoid population is very near the bottom and definitely close to the sediment surface at low tide. Although natant, the harpactocoid taxa considered may also spend some time on the sediment (as reviewed by Hicks & Coull 1983), which only makes the temper-ature in the sediment more relevent to them. Lastly, the sediment may actually be acting as a filter which removes aperiodic rapid temperature fluctuations; these fluctuations could otherwise blur the periodic components and make difficult the identification of a specific signature to be related to the animal blooming period. 96 4.4 Chapter Summary In this chapter, the temporal and the spatial variability of an assemblage of species of harpacticoid copepods was investigated. By sampling the animals on a fine vertical scale, persistent and specific patterns of distribution were observed in the water column. Importantly, knowledge of the shape of these distributions allowed a quantitative estimation of the natant harpacticoid abundance. Comparing absolute abundances and indices of abundance justified the use of an index of abundance to track the temporal patterns in the abundance of harpacticoids. Once it was realized that natant harpacticoid copepods were abundant in the water column, and that their abundance could be quantitatively expressed, it became possible to identify the vernal blooming period; the vernal harpacticoid bloom was defined as that period of most rapid increase of abundance (as identified with the sampling time-scale used). This blooming period was related with the appearance of warming episodes in the sediment. The occurrence of those warming episodes in the sediment is seasonal; the beginning of the period during which those warming episodes will be observed is not constant between years. The timing of onset of warming episodes of the sediment varied by 10 days between the two sampling years. There is indication that the onset of the blooming periods of harpacticoids may have varied by a similar amount of time between the two years. The association of the harpacticoid bloom with the appearance of warming episodes seems very plausible; it also provides a more robust basis for the definition of the link between harpacticoid abundance and thermal regime. The association of the harpacticoid bloom with any specific warming episode (the second?) is much more tentative owing to the difference in time-scales used to describe both. However, the harpacticoid bloom coinciding with the second warming episode could be possible. 97 C hapter 5 E a r l y Near-Shore Life and Surviva l of C h u m Salmon In the previous chapters, the spatial and temporal distributions of chum salmon and harpacticoid copepods were described. Those descriptions were considered prerequisites for the identification of environment-coupled mechanisms potentially regulating the abundance of the fish. In this chapter, the results of selected studies on some aspects of the ecology of juvenile chum salmon are briefly reviewed. Adding information on the spatio-temporal distribution of natant harpacticoid copepods from the present study to those previous results, provides a basis for discussing regulation of salmon abundance in estuaries in an original manner. 5.1 Introduction Stocks of chum salmon in British Columbia are known to fluctuate considerably in abundance (Hoar 1951; Wickett 1958). Between 1965 and 1969, the estimated number of Fraser River chum salmon caught in Johnstone Strait ranged from 0 to 228 000, while 10 000 to 196 000 chum salmon were caught in the Fraser River proper; these marked fluctuations in abundance have not been satisfactorily ac-counted for (Beacham & Starr 1982). To investigate environment coupled mech-98 anisms that may regulate fish abundance, it is convenient to divide the fish's life cycle in stages, and identify the particular risks of mortality at each stage. Early life stage in nearshore regions is regarded by many as a critical time in the life of the chum salmon when the strength of the year-class may substantially be altered. For these reasons, a more thorough understanding of the biological and physical processes that affect individual growth and production of chum salmon in estuaries and near-shore nursery grounds is a central goal in fisheries research (Neave 1953; Wickett 1958; Volk et al. 1984). Central in any discussion on the ecology of chum salmon is the fact that mortal-ity tends to be high in early life (reviewed by Ricker 1976), and that this mortality appears to be caused by predation and is size-selective (Parker 1966, 1971; Harg-reaves & LeBrasseur 1986). The latter authors showed that yearling coho salmon O. kisuich (112-130 mm FL) consumed the smaller fishes within an available array of prey ranging in size from 43 to 63 mm (FL). Parker (1966, 1971) had already suggested that predatory mortality of chum salmon was size-dependent, and that outgrowing their predators would reduce chum salmon mortality. The ponderal growth of a juvenile fish can be represented by where W=ponderal size, t=time, G=growth coefficient. From eq. 5.1, the time (T) to grow through a given size interval Wx — W2 can be derived: G For a fixed size interval, T will vary with G (Fig. 5.1) that is, the faster the growth of the fish, the briefer the time it spends in this size range where size-selective 99 predation is acting. This is commonly referred to as predation field escapement (e.g. Cushing 1973). Since G can be written as (Parsons et al. 1984 a) G = aR-C, (where a=assimilation efficiency; R=ration; C=sum of metabolic costs) it can be concluded that the time to grow through a given size interval will be minimal when aR is maximal and C is minimal. The mortality rate of fish can be represented as ^ = -ZN(t) , (5.2) where N=number of fishes; t=time; Z=mortality coefficient. Integrating eq. 5.2, Nt = N0e-Zt , (5.3) where 7V0=initial number of fishes. If mortality of fishes is only, or mainly, effective over a certain size range, and if the fishes are within the limits of this range for variable time, survival will be variable. Variations of the growth coefficient resulting from variations in ration, assimilation efficiency, or metabolic costs, can then cause variations in fish survival as they result in variable time of exposure to size-selective predation within a certain size range. There is evidence that harpacticoid copepods provide chum salmon with a high growth coefficient. Volk et al. (1984) fed the harpacticoid copepod Tigriopus californicus to juvenile chum salmon and food conversion efficiency K (K — G/R x 100, G & R both in % body weight/day (b.w./d)) was much higher than for other food taxa. When fed the amphipod Paramorea mohri and the calanoid copepod 100 G1 G2 G3 t1 t2 t3 Time(t) Figure 5.1: Variation of time to grow through a given size interval Wj - W2 for various growth coefficients (G). 101 Pseudocalanus minutus, the salmons' maximum food conversion efficiencies were 16.3 and 20% for rations of 9.8 and 9.5% b.w./d respectively. When the fish were fed the harpacticoid, the maximum food conversion efficiency was 40.1% for a ration of 5.7% b.w./d. The authors concluded that a high assimilation efficiency and a low cost of capture could explain the high food conversion efficiency of chum salmon feeding on harpacticoids. Similar high food conversion efficiencies for other species of harpacticoid copepods could justify why harpacticoid copepods have been considered a better nutritional source than many other meiobenthic organisms (Hicks & Coull 1983). For the remainder of this chapter, the life cycle of the chum salmon can be simplified by making the following assumptions, based on the aforementioned ev-idence. First, most of the mortality of the fish will be assumed to occur in early near-shore life, be caused by predation (assumed constant), and be size-selective. It will also be assumed that a substantial fraction of the Fraser River chum salmon stock does feed on harpacticoid copepods in near-shore nursery grounds, and that each fish spends a substantial amount of time feeding on harpacticoid copepods (as discussed in chapter 3). As discussed in chapter 4, the assemblage of natant harpacticoid copepods on which chum salmon preferentially feeds has a clear seasonality of abundance. The abundance of chum salmon on the tidal flats appeared to be equally seasonal. If the relative timing in the seasonal cycle of abundance of the fish and the harpacti-coid copepods varies, the ratio of abundance of both will also vary. The overlap of harpacticoid and chum salmon populations on the flat is determined by extrinsic and intrinsic factors. The extrinsic factors are those that determine the timing of the downstream migration of the fishes and the onset of the harpacticoid bloom. 102 These extrinsic factors have nothing to do with fact that there exists a trophic relationship between salmon and harpacticoids. The intrinsic factors are the re-sults of the trophic interactions that determine the growth and/or death rates of both populations. Extrinsic and intrinsic factors are intimately linked. If extrinsic factors are such that the bulk of the outmigrating chum salmon starts feeding on a poorly developed harpacticoid population, the prey population density may never be high as it is grazed faster than it can increase. The amount of food available per fish would be low. Similarly, if the extrinsic factors are such that the bulk of the salmon population can feed on the harpacticoid population only after it has sharply declined, the ration available per fish will be low. The maximum ration of harpacticoids available for the maximum number of fishes will then occur at some intermediate degree of overlapping of the 2 pulses. This variation in overlapping of predator and prey with resulting variation in ration available has been described by Cushing (1975) (Fig. 5.2). Although developed for herring, Cushing's description of the effect of varying degree of overlap of predator and prey on feeding success of predator, the concept of variation in the timing of phasing of predator and prey seems relevant for any species. As it is, the proposition that variation in overlap of predators and harpacticoid copepods could affect the time course of the prey pop-ulation abundance was put forward by Muus (1967). Muus suggested that an early warming of the shallow waters allows the harpacticoid to take an early lead over the consumers. Variations in the degree of phasing of chum salmon and harpacti-coid copepods could then result in variations of survival of the fish for reasons previously discussed. Since the abundance of both chum salmon and harpacticoid copepods on the flat have been shown to be transitory, an index of the degree of phasing of both could be obtained by measuring the time between well identified 103 Time Figure 5.2: The match or mismatch of larval production to that of their larval food. The number of nauplii/larva represents the degree of feeding success. The three curves represent three conditions of copepod nauplii production and hence three conditions of feeding success, a < b < c (from Cushing 1975). 104 and consistent signatures in respective abundance time-series. Ideally, this index of abundance would be available for many years with concurrent records of juvenile salmon survival. In the Fraser River, the abundance of outmigrating chum salmon has been monitored at Mission City. 80 km upstream from Roberts Bank. Daily sampling of the number of outmigrant chum salmon during the duration of the mi-gration season provides an estimate of the fish's migration pattern (Fig. 5.3). From the daily patterns of outmigrant salmon migration, the median date of downstream migration at Mission City was obtained (Table 5.1). The median date of down-stream migration at Mission was taken as an index of abundance of chum salmon on Roberts Bank. This assumes the existence of a consistent relationship between the temporal patterns of salmon abundance at Mission and on Roberts Bank. In chapter 3, evidence was provided that the abundance of chum salmon on Roberts Bank is characterized by a period of highest density. The salmon counts at Mission City indicate that the downstream migration also has a period of highest activity. Assuming that the periods of highest downstream migration activity at Mission City and of highest density on Roberts Bank could be related, albeit with a certain lag, is probably reasonable due to the already mentioned preference of chum salmon for salt water which translates into rapid downstream migration, and to the relative proximity of Mission City and Roberts Bank. However, no statistical verification of this assumption is possible with the extant data of abundance of chum salmon on Roberts Bank. If data on salmon abundance on Roberts Bank from various sources all indicate the transitory nature of the salmon temporal distribution on the flat (this study; Gordon & Levings 1984; Conlin et al. 1981), grouping them to obtain sufficient data for statistical analysis would not be reasonable because of the variations in sampling techniques and frequency between the different studies. 105 4000 - i Time (calendar days) Figure 5.3: Daily abundance of outmigrant Fraser River chum salmon at Mission City in 1972 (provided by the Department of Fisheries & Oceans, Canada). 106 Then, the median date of downstream migration at Mission City is considered only as an index of abundance of chum salmon on Roberts Bank, and the particular relationship between this index and actual maximum of abundance of salmon on the flat remains to be further investigated along the following lines. The variability in this lag is likely to increase as the lag itself increases. However, a low variability in this lag must be assumed for the relationship between specific patterns in the abundance of salmon at Mission City and on Roberts Bank to be meaningful in the model. With a lag variability assumed small, it is likely that the feature in the temporal abundance of fish on the flat to be linked with the median date of down-stream migration at Mission, should be the first peak of abundance instead of the second (Fig. 3.3). In 1985 and 1986, the first salmon peaks on Roberts Bank were respectively observed on days 100 and 105, while the second peaks occured more than 40 days later. Since the median date of downstream migration at Mission ranged from days 96 to 119 (Table 5.1) in the years for which data are available, the shortest lag is obtained with the first peak. An historical index of abundance of natant harpacticoid copepods on Roberts Bank was then required to match the historical index of abundance of chum salmon on Roberts Bank provided by the historical records of downstream migration at Mission. As discussed in chapter 4, a pattern in the seasonal abundance of natant harpacticoid which seemed to represent a consistent signature was the timing of blooming period onset. This blooming period onset has been related to the ap-pearance of warming episodes in the sediment, and there was indication that it may have been linked with the second of the warming events. Warming events in the sediment on Robert Bank have been proved to occur when the tide level on the flat would fall below 1.6 m at the port of reference between 0800H PST and 107 1630H PST. The date at which the tide level at the port of reference (Point Atkin-son) fell to or below 1.6 m for the second time of the year between 0800H PST and 1630H PST was determined by inspection of the Canadian Hydrographic Tide Tables. This date was then taken to represent the onset of the natant harpacticoid blooming period. One assumption must be made here, namely that the various taxa in the natant harpacticoid assemblage appear in a consistent temporal succes-sion. Only then can the timing of the pulse of the natant harpacticoid assemblage (mostly composed of Tisbe spp.) be considered as an indication of the timing of appearance of H. uniremis and Z. aurelii. The difference in days between the me-dian date of downstream migration of chum salmon at Mission and the date of the hindcasted natant harpacticoid blooming period onset was defined as a phasing index (AG) which gives a measure of the overlap of salmons and harpacticoids on Roberts Bank (Table 5.1). The median date of downstream migration varied little over the years (range: 23 days; coefficient of variation: 5.2) when compared to the hindcasted bloom onset date (range: 36 days; coefficient of variation: 17.4). The extremes of median date of downstream migration of chum salmon at Mission did not correspond to the extremes of phasing index. The earliest date of median downstream migration (April 6, 1976)) and the latest date of downstream migra-tion (April 28, 1971) have corresponding phasing indices of 29 days and 55 days respectively; the smallest phasing index is for 1980 (17 days), and the highest, for 1974 (65 days). The earliest hindcasted bloom onset occurred in 1970 (February 16), and the latest, in 1980 (March 24). The year with the latest hindcasted bloom onset is also the year with the smallest phasing index (1980). During the year with the highest phasing index (1974), the hindcasted bloom occured only one day after the earliest hindcasted bloom of all (1970) (day 48 in 1974; day 47 in 1970). 108 Broodyear Mission50 Bloom AG % Survival (DFO) Date Day # Date Day # 1965 April 17 107 February 24 55 52 31.0 66 April 24 114 March 14 73 41 31.4 67 April 20 111 March 3 62 49 14.3 68 April 17 107 February 21 52 55 52.0 69 April 18 108 February 26 57 51 33.0 1970 April 17 107 February 16 47 60 32.4 71 April 28 119 March 4 64 55 27.0 72 April 21 111 February 21 52 59 24.5 73 April 16 106 February 26 57 49 22.5 74 April 23 113 February 17 48 65 34.8 75 April 11 102 March 5 65 37 36.8 76 April 6 96 March 8 67 29 12.8 77 April 13 103 March 12 71 32 28.2 78 April 22 112 March 17 76 36 35.4 79 April 13 104 March 18 78 26 29.7 1980 April 10 100 March 24 83 17 15.1 81 April 16 106 March 16 75 31 19.4 Table 5.1: Median date of chum salmon downstream migration at Mission City (Missionso), date of hindcasted natant harpacticoid bloom onset at station H on Roberts Bank (Bloom), phasing index (A0), and juvenile-to-adult chum salmon survival (% Survival (DFO)). 109 This indicates that the variations in the phasing index are more closely associated with the variations in the hindcasted dates of harpacticoid bloom onset than with variations in median dates of downstream migration of salmon. This was expected, since the coefficient of variation of hindcasted blooming dates proved to be larger than that of median date of downstream migration. Also, Table 5.1 indicates that the actual value of the phasing index for each year is a bivariate parameter, in that it results from the difference in timing of two events with independent degrees of variability. Data on juvenile-to-adult survival of Fraser River chum salmon were obtained to be related to the hindcasted phasing index of chum salmon and natant harpacti-coid copepods on Roberts Bank (Table 5.1). These data on survival represent the ratio of returning adults to the actual count of outmigrant juvenile salmons dur-ing a fixed counting effort, and as such are an index of survival. The index of survival was plotted versus the hindcasted phasing index for years 1965 to 1981 (Figs. 5.4, 5.5, and 5.6). When the odd and even broodyears are grouped (Fig. 5.4), no particular structure is evident in the plot, except perhaps for the sugges-tion of a monotonic increase in survival index with an increase of phasing index. However, this is not what would be expected if the survival were to optimize at some intermediate value of index of phasing. One reason why no optimization of survival is observed at intermediate value of phasing index could be that the available data set does not bracket all possible values of phasing index and con-sequently, the possible decline in survival with increasing phasing could simply be absent from the data. Another possible reason is that grouping data for odd and even broodyears could blur possible relationships between phasing and survival in-dices because of the presence of pink salmon (O. gorbuscha) every other year. In 110 55-i 50-45-X 40H £ 35H D 30 25^ 20 -15-10 80 75 78 79 66 77 ft 76 68 74 69 • • 65 70 71 72 73 67 ~i I 1 1 1 1 1 1 1 1 1 10 15 20 25 30 35 40 45 50 55 60 65 Phasing Index Figure 5.4: Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 (odd & even broodyears). I l l the Fraser River, outmigrant pink salmons only occur in even years. This means that even-year outmigrant juvenile chum salmon (that were spawned the previous spring, and thus are part of an odd broodyear) will share their migration route with pinks. Pink salmon are known to feed on harpacticoid copepods (e.g. Levy et al. 1979); gut-contents of pink salmon caught on Roberts Bank in this study did not seem to differ in composition from those of chum salmon (D. Webb, Depart, of Oceanography, U.B.C, pers. comm.). It may well be, then, that pink and chum salmon feed equally on harpacticoids. In the Fraser River, the average number of juvenile outmigrants of pinks is about twice that of chum (DFO). Although there is indication that pink salmon residence in estuaries and near-shore environment is shorter than that of the chum, juvenile outmigrants of chum and pink may then be trophically equivalent to a certain point. If this is the phasing index computed with the median date of downstream migration of chum salmon could be invalid when applied to years when pinks are present. If pink and chum are trophically equivalent with respect to harpacticoids, this phasing index would have to be based on the median date of downstream migration of the combined total of pink and chum. For these reasons, the odd- and even broodyear data were plotted separately (Figs. 5.5 & 5.6). For odd broodyears (Fig. 5.5), the plot of index of survival versus phasing index suggests no pattern. As discussed above, a reason for this might be that the median date of downsteam migration for chum salmon only is inappropriate for those broodyears that will share the migration route with pink salmons. For even broodyears (Fig. 5.6), there is indication that indeed, survival could optimize at some intermediate value of phasing index. Broodyears 1980 and 1976 had survival indices of 15.1 and 12.8 with phasing indices of 17 and 29. Broodyears 1970, -72, &-74 had survival indices of between 24.5 and 34.8 with 112 phasing indices of between 55 and 65. Broodyears 1966, -68, &; -78 had survival indices of between 31.4 to 52.0 with phasing indices of between 36 and 55. For an approximation of the location of the apex (A0 o p ( ) on the plot of sur-vival index (S.I.) versus phasing index for even broodyears, two straight lines were respectively fitted in the ascending and decreasing sections of the data points (Fig. 5.6): S.I. = -7.69 + 1.O4A0 (i?2 = 0.83, n - 5) (5.4) for the ascending section, and 5.7. = 123.72 - 1.47A0 {R2 = 0.27, n = 4) (5.5) for the descending section. The slope of Eq. 5.4 was significantly different from zero (p=0.03; F=15.2) while that of Eq. 5.5 not. For these reasons, the increase in survival with an increase of phasing is considered significant up to a phasing index value of 55 days; after this value of phasing, the apparent decline in survival can only be considered as a suggestion of a reversal of the trend in the relationship between survival and phasing indices. Keeping those qualifications in mind, the apex is found by equating Eq. 5.4 and Eq. 5.5: -7.69 + l.O4A0 o p t = 123.72 - 1.47A0op( , which indicates a survival index (S.I.) of 46.76 for A0opt=52.36 days. The survival vs. phasing relationship can then be described by: f -7.69+1.O4A0 if A 0 < 52.36 , . j 123.72 - 1.47A0 otherwise. ^ ' It must also be pointed out that the indication of optimization of survival at inter-mediate values of phasing index is heavily dependent on broodyear 1968, and that 113 further data are needed to assess the significance of the seemingly highest survival of chum salmon at intermediate values of the phasing index. The data point for broodyear 1976 also may contribute unduly to the appearance of a hump in the survival curve of even broodyear chum salmon versus hindcasted phasing index. In 1977, the year during which the bloom onset must be hindcasted for the 1976 fish broodyear, the first hindcasted warming episode in the sediment was not followed by a similar warming episode 15 days later. The first occurrence of a low tide falling below 1.6 m during daytime (as previously defined in chapter 4) was on February 9; ca 15 days later, when another similar tide is expected, the tide level did not fall lower than 1.7 m during daytime. Only 4 weeks after the first warming episode (on March 8) did the tide fall below 1.6 m during daytime. This requires the second warming episode, to which the harpacticoid bloom was tentatively associated, to be hindcasted 4 weeks after the first one, contrary to 2 weeks if the tide conditions required for the occurrence of a warming episode in the sediment had been defined with a lower critical tide height. This causes the phasing index for 1976 to be small (29 d) instead of 43 d if the second warming episode had been hindcasted 15 days after the first one. If the survival datum for broodyear 1976 was moved to a corresponding A0=43 d (Fig. 5.4), suggestion of the presence of a hump in the survival vs. phasing indices plot would be somewhat diminished. However, if one chooses to consider the quasi-warming episode happening 15 days after the first one in 1976 as the one to which the harpacticoid bloom should be related, the definition of the tide conditions required for the occurrence of a warming episode in the sediment has to be modified. Attempts at modifying this definition only resulted in further ambiguous situations. Since the definition for the occurrence of warming events in the sediment had been a priori defined with 1.6 m as the critical 114 55- i 50-45-X 40-<D TJ f 35H D £ 30 H ^ 25-1 20 15 H 10 75 79 I77 181 69 I65 171 I73 67 I I I 1 1 1 1 1 1 1 1 10 15 20 25 30 35 40 45 50 55 60 65 Phasing Index Figure 5.5: Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 (odd broodyears). 115 55-i 50-45-40 35 30 25 20 15 10 6 8 / / \ / \ 7 8 / / 6 6 / / / / 80 / / L6 \ 7 4 * # \ 7 2 I I I I I I 1 1 1 1 1 10 15 20 25 30 35 40 45 50 55 60 65 Phasing Index Figure 5.6: Index of survival versus phasing index for Fraser River chum salmon from 1965 to 1981 (even broodyears). The increasing dashed line was fitted with the data points from 1966, 1968, 1976, 1978 & 1980; the decreasing dashed line was fitted with the data points from 1968, 1970, 1972 & 1974. 116 Point Atkinson (1980) % A T AH 48.8 <5 min 78.7 <10 min 92.2 <15 min 31.2 <5 cm 56.4 <10 cm 73.2 <15 cm Table 5.2: Comparison of predicted and observed tides at Point Atkinson (1980). Percent (%) of the time when observed timing of tide deviated from predicted (timing deviation (+ or -)= AT), and percent of the time when observed tide height deviated from predicted (height deviation (+ or -)=Ai7) (data provided by the Canadian Hydrographic Service, IOS). tide level following empirical evidence, and to be as consistent as possible, it was chosen to consider as the second warming episode the later date in 1977 (March 8), keeping in mind all the potential implications of this choice. To assess the validity of the hindcasted date of harpacticoid bloom onset based on the hindcasted second warming event, the accuracy of predicted tide levels must be evaluated. The observed level of tide usually differed little from the predicted level (Table 5.2). In 1980, which is used as a paradigm, the observed and predicted timings of tide (high or low) never differed for more than 15 min for 92 % of the time at Point Atkinson (port of reference). The predicted and observed tide levels never differed by more than 15 cm for 73 % of the time. It can be concluded that predicted tide height and timing are accurate, and that the definition of tide conditions (level and timing) to generate a warming event should be reliable in fore- or hindcasting. Beacham & Starr (1982) analysed various environmental factors that could reg-117 ulate the abundance of chum salmon in early life stages. The median date of down-stream migration of the fish (also at Mission) was one of the factors investigated (Fig. 5.7). What must be emphasized here is that in that study, the median date of downstream migration was used directly in the analysis as a univariate parameter. Beacham &; Starr (1982) suggested that there seemed to be a decrease of survival of the fish with later median dates of downstream migration. The conclusions of Beacham & Starr on the survival of chum salmon in relation to median date of downstream migration can not be compared to those of the present study. The reason for this is that they used migration and survival data based on estimates of total abundance of downmigrant frys. Today, the techniques to estimate total abundance of frys are no longer used, and migration dates and survival rates are computed directly from the actual count of frys (T. Beacham, Pacific Biological Station, pers. comm.). However, it is possible to see if conclusions derived from analysing survival index versus phasing index (bivariate) will differ from those de-rived from analysing survival index versus median date of downstream migration per se (univariate). Data on chum salmon juvenile-to-adult index of survival were plotted versus median dates of downstream migration for even & odd years, odd years, and even years (Figs. 5.8, 5.9, and 5.10). The general suggested trends in the dispersion patterns of the chum salmon survival index data in relation to median date of downstream migration (Figs. 5.8, 5.9, & 5.10) differ from those of the same data in relation to the hindcasted phasing index of the fish and natant harpacticoids (Figs. 5.4, 5.5, & 5.6). For all broodyears (Fig. 5.8), the data are very dispersed and no trend is suggested. For odd broodyears alone (Fig. 5.9), the index of survival generally decreases with later median dates of downstream migration. For even broodyears alone (Fig. 5.10), the index of survival of chum 118 3-i 2.5 D > 3 CO CD a CD C L 2 -1.5-1-0.5-65 68 64 69 66 • • 70 72 • • 74 75 67##73 71 I 1 1 1 1 100 105 110 115 120 125 Median date of downstream migration Figure 5.7: Juvenile-to-adult survival of chum salmon versus median date of down-stream migration at Mission (redrawn from Beacham & Starr (1982)). 119 X D > > 3 to 55n 5 0 -45 40 35 30 25 20 15 H 10 68 75 78 >74 771 7 0 ^ 6 9 72 731 181 76 so 67 171 1 1 1 1 1 1 90 95 100 105 110 115 120 Median Date of Downstream Migration Figure 5.8: Juvenile-to-adult index of survival of chum salmon versus median date of downstream migration al Mission (even -f odd broodyears). 120 55 n 50 H 45 H X 40H <D £ 35H D 30 H W 25-20-15-10 75 69 79 65 71 77 173 81 67 1 1 1 1 1 1 90 95 100 105 110 115 120 Median Date of Downstream Migration Figure 5.9: Juvenile-to-adult index of survival of chum salmon versus median date of downstream migration at Mission (odd broodyears). 121 55-i 50-45-X 40-0) £ 35 H D > > c 30-<S> 25-1 20-15-10 80 76 68 70< 7 8 « # 7 4 • 66 • 7 2 90 95 I 1 1 -100 105 110 1 1 115 120 Median Date of Downstream Migration Figure 5.10: Juvenile-to-adult survival of chum salmon versus median date of down-stream migration at Mission (even broodyears). 122 salmon generally increases with later median dates of downstream migration. None of those trends is statistically significant (a = 0.05), and even if they are considered as indicative of some patterns, those patterns would be monotonic. However, as indicated by Parsons & Kessler (1987), survival of juvenile salmonids should vary non-monotonically with monotonic changes in forcing functions. Those authors modeled the survival of pink salmon (O. keta) under various conditions of environ-mental factors. Their main conclusion was that the salmon biomass increase was maximal at an optimum (intermediate) level of each environmental factor studied. This maximum was attained in each case largely through the phasing of zooplank-ton and phytoplankton production in such a way as to maximize the standing stock of the former, and hence, the growth rate of the salmon. In the same study, it was also concluded that the timing of salmon arrival after the initiation of the spring bloom was a relatively inconsequential event for the survival of the fish under the most favourable growing conditions for zooplankton. However, this does not dis-connect the conclusions of their study from the present. In the present study, it is not solely the availability of food which is assumed to regulate salmon growth and survival, but the quality of harpacticoid copepods specific food items that provide the fish with the highest growth rates. It was shown earlier that this provision of specific food resource was an ephemeral event. If the bulk of the fish population overshoots this event, food is still available, but with less growth effectiveness. For the fishes, this would be equivalent to grazing on zooplankton growing under less than favourable conditions, which could reduce the amount of food available. In these conditions, Parsons & Kessler (1987) suggested that timing of salmon arrival could be consequential for the fish survival. Survival of chum salmon increased with increasing phasing index for low phasing values; at higher phasing values, there is 123 at least no evidence of further increasing of survival, and there is suggestion of decrease in survival, which appears to be consistent with the results of modeling investigations by Parsons & Kessler of the survival of juvenile salmonids. For all the above reasons, it can be concluded that the phasing index (A0) ap-pears as an appropriate covariate of chum salmon survival since a non-monotonical change in the survival of the fish is suggested with monotonic changes in their index of phasing with harpacticoid copepods. 5.2 Summary Growth of chum salmon has been considered as a way for the fish to escape their size-selective predators by outgrowing them. The food quality of natant harpacti-coid copepods could promote the fastest growth in chum salmon. However, the period of availability of natant harpacticoid copepods is brief. The overlap of the salmon pulse and the harpacticoid pulse can vary so as to affect individual growth of fish, and hence affect survival. The survival of the fish should optimize at some intermediate overlap (phasing). It appears that an index of phasing measuring the overlap of chum salmon and harpacticoid (using the median date of downstream migration as proxi for the fish and a predictable tide pattern signature for the copepods) is an appropriate covariate of the chum salmon survival index. It is concluded that using the median date of downstream migration within a bivariate phasing index as the co-variate of salmon survival is more appropriate than using it per se since the latter approach indicates a continuous trend in the survival vs. phasing relationship, while the former indicates a discontinuous trend. 124 C hapter 6 Genera l Discussion and C onclusion The initial aims of this research project were to contribute to the on-going debate on the nature and action of mechanisms regulating fish population abundance, and to elucidate some aspects of the ecology of harpacticoid copepods to increase our understanding of their role in the diet of juvenile chum salmon. To meet those aims, a conceptual framework (trophic phasing analysis) was proposed to address the issue of fish abundance variability, and was subsequently applied in a case study. In this chapter, a brief evaluation of accomplishments in the case study on chum salmon is presented. Potential usefulness of trophic phasing analysis in fisheries oceanography is then discussed in light of those accomplishments. 6.1 Chum Salmon Abundance and Early Near-Shore Life Discussions on potential limitation of production of chum salmon during the fish's early near-shore life are generally very open-ended. After studying the feeding habits of juvenile chum salmon in the Nanaimo estuary, Healey (1979) concluded that "it is probably common in estuaries for continuous production of salmon to 125 depend upon the conservation of specific food resources and the habitat charac-teristics that make these resources available to the salmon". In their study of the feeding of juvenile chum salmon in the Puget Sound, Kaczynski et al. (1973) ex-plicitly stated that the results of the "trophic analysis... was to be used in the predictive models of... [chum salmon]...return". The main conclusion reached by Kaczynski et al. in that study was that the onshore stages of development appeared to be a distinct ecological stage in the life cycle of chum salmon. Despite such substantial gains of insight in the ecology of juvenile chum salmon in previous trophic analyses, the assessment of potential limitation of production of chum salmon in estuaries is yet to be completed. A possible reason for this is that hitherto, trophic analyses of chum salmon have focused on describing the diet of the fish without considering the production dynamics of the prey items. Harpacticoid copepods have been recognized as the chief food item of juvenile salmon; the period of reliance on harpacticoids by the fish has even been called, perhaps excessively so, as "obligate" (Chandler 1986). While it is generally recognized that changes in availability of harpacticoids could cause fluctuations in growth and survival, very little, if any, of the spatio-temporal variability of harpacticoids is taken into account in previous trophic analysis. As emphasized earlier, the production of food is a process at least partially independent of the effects of predators (Parsons et al. 1984 a). Attempting to assess limitation of salmon production in estuaries with trophic analysis and not including the dynamics of prey production is likely to be inconclusive since it disregards one of the degrees of freedom of the fish production process. In this study, some aspects of the spatio-temporal variability of harpacticoid copepods were considered; an essential characteristic of the production dynamics 126 of some harpacticoid copepods, namely the association of the spring bloom with a distinct signature in the thermal regime, was investigated in detail. This allowed assessement of the potential limitation of salmon production in estuaries with con-sideration of two degrees of freedom of the fish's production process-the time course of salmon abundance on the flat, and the time course of harpacticoid abundance on the flat. The results of this study could have some practical applications. It is one of the stated objectives of the Salmon Enhancement Program (SEP) to identify the most appropriate time to release hatchery produced chum salmon to maximize their chances of survival (SEP Annual Report 1983). The results of this research could provide the basis to establish this date and orient future research in this direction. As indicated in the previous chapter, the apex of the survival-phasing relationship appears to correspond to a phasing index of 52 days; although based on limited amount of data, this figure provides the first quantitative expression of the potential effect on salmon survival of the fish's migration schedule, based on a structural analysis. Also, the environmental characteristics that make harpacticoid copepods available to juvenile chum salmon could be defined precisely and moni-tored if need be. These environmental conditions could be defined in terms of the link between harpacticoid abundance and thermal regime on the tidal flats. Any modification in the thermal regime could be readily monitored (and at low cost). Factors potentially affecting the thermal regime on tidal flats are those affecting the water balance on the flat (slope, plant cover, porosity). The results of this study could provide some objective criteria to define and monitor some aspects of the environment quality on tidal flats. Another indirect practical aspect of this research is the effect it could have on the annual decision that has to be taken by 127 the Department of Fisheries & Oceans as to whether or not continue the moni-toring of outmigrant juvenile salmon at Mission City (CD. Levings, DFO, pers. comm.). The fact that historical data on juvenile chum salmon migration could indeed be useful in assessing potential limitation of fish production in estuaries could contribute to securing the practice of salmon outmigration monitoring. A most desirable practical application of such research would be establishing a relationship from which returns of chum salmon could be confidently predicted. Although it is suggested that the results of this study are consistent with recent advances in modeling of survival of juvenile salmonids, it would be premature to attempt building predictive models with these results. The three principal reasons for this are the following: 1. no estimation of confidence limits around the index of survival of chum salmon are available; 2. evidence of optimization of chum survival at intermediate level of phasing is heavily reliant on a limited number of data points; 3. the proportion of the Fraser River stock of chum salmon actually using tidal flats as feeding grounds remains speculative. Nonetheless, this study has clarified some aspects of the ecology of harpacticoids, which allowed discussion of their role in the diet of chum salmon in an original manner. This led the way to a discussion on the potential limitation of production of chum salmon in estuaries which is more robust, as much for the synthesis of extant information this research provides, as for the clearer identification of future research priorities it also provides. 128 The application of trophic phasing analysis to a case study yielded new infor-mation on the mechanisms that may regulate fish abundance, and this information could have practical applications. It may be suggested that trophic phasing analysis could be applied to several fish species for further insights in environment coupled fish recruitment mechanisms, which is the object of fisheries oceanography. 6.2 Trophic Phasing Analysis and Fisheries Oceanog-raphy The study of animal population regulation is a central effort in ecology, irrespec-tive of the particular species concerned. In ecology, it has been recognized that the dogma of density dependence should be replaced by more fruitful arguments about observations on real animal populations. In that sense, the goal of studies on fish populations should be the elucidation of the mechanisms regulating animal abundance; only then should concerns about exploitation be addressed. However, one fundamental characteristic of fish populations makes the issues of operational usefulness of population regulation models paramount in fisheries research. Collec-tively, fisheries resources represent a major source of animal protein ( ca 75 106 t/y (landed)) and yet, of all the intensively exploited animal populations, fishes (sensu largo) are the animals on whose environment little, if any, control is possible. This, in conjunction with the Malthusian belief that an animal population should reach equilibrium through density-dependent regulation, has led the rationale to secure a lasting benefit from this resource to be developed around the parental-filial theme. Sustainable yield strategies depend on quantifying density dependence in stock re-cruitment relationships (Sissenwine 1984). The whole theory of fishing depends on the existence of compensation: an unexploited fish population is in near equilib-129 rium with no surplus production (recruitment and growth are balanced by natural mortality); fishing reduces population size which responds (compensates) with a surplus production available for harvest (Sissenwine 1984). Parental-filial relation-ships have gained wide acceptance perhaps more for their usefulness in establishing quotas, for which at least consistency is required, than for their validation as mod-els for the mechanisms of fish population regulation. If it is chosen to further investigate fish population regulation mechanisms, empirical verification of models must be separated from operational usefulness. When the operational usefulness of stock recruitment models is set aside, the rel-evance of the parental-filial theory must be addressed at the level of its foundations. The most fundamental tenet in the parental-filial theory is that the abundance of an animal population (N) should vary in time (t) until an equilibrium level (K) is reached, that is (see Morey 1980) lim Nit) = K. Without such an assumption on the time course of the fish population abundance, the functionality between parental and filial generations would be impossible to be formulated, let alone quantified. There is indication that the assumption by fisheries management of a natural persistence at equilibrium level in fish stocks is questionable (Anonymous 1980; Steele 1984; Steele & Henderson 1984). The reason for the widespread belief in equilibrium of non-exploited fish stocks is per-haps the ease with which the undeniable fact that animal populations do not grow beyond certain limits can be equated with the assumption that these limits are the levels at which populations would settle if left to themselves. Equating the latter assumption with the former observation provides the only justification of stock recruitment theory as a model for fish abundance regulation. However, if 130 doubts can be raised on the existence of "equilibrium levels" of virgin populations, doubts can also be raised on the parental-filial theory for animal population reg-ulation. If parental-filial theory is only one approach for the investigation of fish abundance variability, the environmental thesis can itself be considered as a dis-tinct and complete approach, especially since advances in oceanography are making the hitherto overlooked environment of the fishes increasingly better understood. For its concern for the environment of fishes, fisheries oceanography appears as a comprehensive field of research on fish population regulation, and trophic phasing analysis appears an appropriate methodology. Besides its use to describe the methodology of fisheries oceanography, the con-cept of trophic phasing analysis permits the avoidance of confusion in the terms previously criticized by Sinclair & Tremblay (1984). Those authors pointed out that the expression "Match-Mismatch" was often used to refer to various oceano-graphic or meteorological phenomena that may cause fish abundance variability through fluctuations in food supply for early life stages. As it is, the concept of "Match-Mismatch" was developed to describe the variability in overlap of the spring phytoplankton bloom due to seasonal stratification changes in the water col-umn in the drifting environment of larval fishes, which is a very specific definition. It can be suggested that the "Match-Mismatch" is but one case of variable trophic phasing that may influence fish abundance, as are the "Stability Hypothesis" or the "Larval Retention Area Hypothesis". This would help recognize that the en-vironment of fishes is perhaps much more reticulated than perceived, and that it is unlikely that a single environment coupled mechanism would account for the variability in abundance of every fish species. The goal of fisheries oceanography is to identify those environmental conditions 131 that may cause drastic changes in the abundance of fish populations. However, it has been suggested that this goal was not achievable. Walters (1984) pointed out that some stock recruitment data are log-normally distributed about the fitted stock recruitment curve; this is what would be expected if variations in recruitment are caused by many independent small variations throughout the life of the fish. In that case, it is unlikely that a single environmental factor could be identified as the main source of variability in fish abundance. Walters submits that if many environmental factors are acting independently on the survival of fishes, the task of determining the contribution of each in the variations of fish abundance would be impossible. This argument could temper research efforts in fisheries oceanography if it was not using a stock recruitment curve to determine the type of distribution of the data points. In doing so, environmentally induced variability is perceived as noise on the a prion'parental filial curve. However, Walters' (1984) argument that prediction of fish stock size is but an option for management must be stressed. This emphasizes that much of the the concerns of fisheries management is not physiocratic in nature, and that the argument of better predictability of fish stock abundance to justify research on fish population variability should be kept in that perspective. As reviewed by Alverson & Paulik (1973): If one had to classify the basis for regulation governing the uses of marine living resources... the status of the stock would be the rationale most frequently given for a management action. However, allocation of the resources and economic status of the user are obviously involved. Although the key word [in] international management is also "conser-vation", more frequently the contentious issue involves allocation of a resource among multiple user groups... 132 6.3 Conclusion This study was undertaken with the view that noise in fish stock recruitment curves could be reduced by studying the environment of the animal. The main thesis as to how environment may influence animal abundance was derived from Hjort's (1914) seminal paper. Hjort submitted that fish abundance may be regulated by environ-mental factors during early life stages, irrespective of the parental population size. However, after trying to apply this view to a case study, it became apparent that research on fish population regulation had become dominated by the imperative of operational requirements for fisheries management. The environmental thesis and the parental-filial thesis should be compared at the level of their foundations, and substantiated by empirical verification. Rather, a common orientation in research is to argue strongly in favour of the role of the environment in the determination of fish abundance, and then to try to amalgamate this argument with parental-filial theory. This persistent return to parental-filial theory is probably the result of dif-ficulties in divorcing the idea of operational usefulness of the parental-filial theory in the actual scheme of fisheries management. My greatest gain in this research was the realization that research on fish abun-dance variability is but one special case of the study of animal population regula-tion. The theory of animal population regulation is yet to be completed, and the dogma of density dependence seems tenacious because of the counter-intuitiveness of proposing non-equilibrium ecological tenets. Perhaps the dogma of density de-pendence is as tenacious as our drive to see, and impose if need be, order in Nature. No one else but Hjort, in the same paper already referenced (1914) could have sum-marized the situation better. Wrote Hjort: In the interest of biological research, also, it must be of the greatest 133 i m p o r t a n c e to ascer ta in the na tu re of the laws w h i c h govern the renewal of the a n i m a l p o p u l a t i o n s , espec ia l ly w h e n t a k i n g in to c o n s i d e r a t i o n a l l t ha t has been w r i t t e n f r o m t i m e to t i m e a b o u t o v e r p o p u l a t i o n a n d the s t ruggle for existence i n the a n i m a l w o r l d , a l l f r o m the p o i n t o f v i e w of h u m a n c o n d i t i o n s . 134 R E F E R E N C E S Alverson, D.L. & G.J. Paulik. 1973. Objectives and problems of managing aquatic resources. Journal of the Fisheries Research Board of Canada 30:1936-1947. Anonymous 1980. Fisheries ecology-Some constraints that impede advances in our understanding. Ocean Science Board, National Academy of Sciences. 16p. Bawden, C.A., W.A. Heath, and A.B. Norton. 1973. A preliminary baseline study of Roberts and Sturgeon Banks. Westwater Research Center Technical Report 1:54 p. Beacham T.D. 1984. Catch, escapement, and exploitation of chum salmon in British Columbia, 1951-1981. Canadian Technical Report of Fisheries and Aquatic Sciences No. 1270. 13 p.+184. Beacham, T.D., and P. Starr. 1982. Population biology of chum salmon, On-corhynchus keta, from the Fraser river, British Columbia. Fishery Bulletin 80:813-825. Beddington, J.R. and R.M. May. 1977. Harvesting natural populations in a randomly fluctuating environment. Science 197:463-465. Bell, S.S., K. Walters, and J.C. Kern. 1984. Meiofauna from seagrass habitats: A review and prospectus for future research. Estuaries 7:331-338. Berg, J. 1979. Discussion of methods of investigating the food of fishes, with refer-ence to a preliminary study of the prey of Gobiusculus flavescens (Gobiidae). Marine Biology 50:263-273. Beverton, R.J.H. and S.J. Holt. 1957. On the dynamics of exploited fish popula-tions. U.K. Min. Agric. Fish., Fish. Invest. (Ser II) 19:533 p. Burkenroad, M.D. 1948. Fluctuations in abundance of Pacific halibut. Bulletin of the Bingham Oceanographic Collection 11:81-129. Carson, J.E. 1961. Analysis of soil and air temperature by Fourier techniques. Journal of Geophysical Research 68:2217-2232. Chandler, G.T. 1986. High-density culture of meiobenthic harpacticoid cope-pods within a muddy sediment substrate. Canadian Journal of Fisheries and Aquatic Sciences 43:53-59. 135 Clark, F.N. and J.C. Marr. 1956. Population dynamics of the Pacific sardine. Calif. Coop. Oceanic Fish. Invest. Rep. July 1, 1953-March 31, 1955. p. 11-48. Conlin, K.B., P. Lawley, P. Futer, M. Kotyk, L. Jantz, B. Hillaby, R. Elvidge, B. Piercey, D. Gordon, C. Levings, K. Hutton, and R. Maclndoe. 1982. Fraser estuary comparative habitat study: 1. Beach seine catches, water characteristics, and geomorphology, March 1980 to July 1981. Can. Data Rep. Fish. Aquat. Sci. 340:125p. Coull, B.C. 1977. Marine flora and fauna of the northeastern United States. Copepoda: Harpacticoida. U.S. Dept. Commer. NOAA Tech. Rep., NMFS Cir. 399. 48 p. Coull, B.C. 1985. Long-term variability of estuarine meiobenthos: an 11 year study. Marine Ecology Progress Series 24:205-218. Csirke, J. and G.D. Sharp. (ED.) 1984. Reports of the Expert Consultation to examine changes in abundance and species composition of neritic fish re-sources. A preparatory meeting for the FAO World Conference on fisheries management and development. FAO Fish. Rep. 29l(l):102 p. Cushing, D.H. 1973. Dependence of recruitment on parent stock. Journal of the Fisheries Research Board of Canada 30:1965-1976. Cushing, D.H. 1975. Marine Ecology and Fisheries. Cambridge University Press (Cambridge). 278 p. Cushing, D.H. 1984. The dependence of recruitment on parent stock derived from the average recruitment. Journal du Conseil 41:276-279. D'Amours, D. 1986. Sediment temperature on Roberts Bank [computer file]. De-nis D'Amours, Department of Oceanography, University of British Columbia [producer]; University of British Columbia Data Library [distributor]. 4 com-puter files (number of logical records varies). de Wilde, P.A.W.J. and E.M. Berghuis. 1979. Cyclic temperature fluctuations in a tidal mud-fiat, in R.G. Naylor (ed.) p. 435-441. Cyclic phenomena in Marine Plants and Animals. Pergamon Press. Feller, R.J. and V. Kaczynski. 1975. Size selective predation by juvenile chum salmon (Oncorhynchus keta) on epibenthic prey in Puget Sound. Journal of the Fisheries Research Board of Canada 32:1419-1429. Fenchel, T.M. 1978. The ecology of micro- and meiobenthos. Annual Review of Ecology and Systematics 9:99-121. Frank, K.T. and Leggett. 1985. Reciprocal oscillations in densities of larval fish and potential predators: a reflection of present or past predation? Canadian Journal of Fisheries and Aquatic Sciences 42:1841-1849. 136 Gordon, D.K. and C D . Levings. 1984. Seasonal changes of inshore fish popula-tions on Sturgeon and Roberts Bank, Fraser River estuary, British Columbia. Canadian Technical Report of Fisheries and Aquatic Sciences No. 1240, 81p. Gulland, J.A. 1977. Fish catches and environment. Journal of the Fisheries Research Board of Canada 34:2221. Gulland, J.A. 1983. Fish Stock Assessement FAO/Wiley Series on Food and Agriculture, Vol. 1. 233 p. Hargreaves, N.B. and R.J. LeBrasseur. 1986. Size selectivity of coho (On-corhynchus kisutch) preying on juvenile chum salmon (O. keta). Canadian Journal of Fisheries and Aquatic Sciences 43:581-586. Harris, R.P. 1972. Reproductive activity of the interstitial copepods of a sandy beach. Journal of Marine Biology of the Association of the United Kingdom 52:507-524. Harrison, B.J. 1981. The biological determinants of the structure of harpacti-coid copepod communities on an estuarine intertidal flat. Ph.D. Thesis, The University of British Columbia, Department of Zoology. 394p-t-49. Harrison, P.G. 1987. Natural expansion and experimental manipulation of sea-grass (Zostera spp.) abundances and the response of infaunal invertebrates. Estuarine, Coastal and Shelf Science 25 (in press). Harrison, P.J., J.D. Fulton, F.J.R. Taylor, and T.R. Parsons. 1983. Review of the biological oceanography of the Strait of Georgia: pelagic environment. Canadian Journal of Fisheries and Aquatic Sciences 40:1064-1094. Harrison, S.J. 1984. Temperature fluctuations in Chichester Harbour, Southern England. Weather 39:166-173. Harrison, S.J. 1985. Heat exchange in muddy intertidal sediments: Chichester Harbour, West Sussex, England. Estuarine, Coastal and Shelf Science 20:477-490. Harrison, S.J. and A.P. Phizacklea. 1987. Temperature fluctuation in muddy intertidal sediments, Forth Estuary, Scotland. Estuarine, Coastal and Shelf Science 24:279-288. Healey, M.C. 1979. Detritus and juvenile salmon production in the Nanaimo estu-ary: I.Production and feeding rates of juvenile chum salmon (Oncorhynchus keta). Journal of the Fisheries Research Board of Canada 36:488-496. Heip, C. and N. Smol. 1976. Influence of temperature on the reproductive po-tential of two brackish-water harpacticoids (Crustacea:Copepoda). Marine Biology 35:327-334. Hicks, G.R.F. 1979. Pattern in the reproductive cycles of benthic harpacticoid copepods. mR.G. Naylor (ed.) p. 139-147. Cyclic Phenomena in Marine Plants and Animals. Pergamon Press. 137 Hicks, G.R.F. and B.C. Coull. 1983. The ecology of marine meiobenthic harpacti-coid copepods. Oceanography annd Marine Biology Annual Review 21:67-175. Hilborn, R. 1983. Stock recruitment manual. MS. Institute of Animal Resource Ecology, The University of British Columbia. 34p-(-III. Hilborn, R. 1985. Simplified calculation of optimum spawning stock size from Ricker's stock recruitment curve. Canadian Journal of Fisheries and Aquatic Sciences 42:1833-1834. Hilborn, R. and R.M. Peterman. 1977. Changing management objectives, in D.V. Ellis (ed.) p. 68-98. Pacific Salmon Management for People. Western Geographical Series, Vol. 13. The University of Victoria, Victoria. Hjort, J. 1914. Fluctuations in the great fisheries of northern Europe viewed in the light of biological research. Rapports et Proces- Verbaux 20:1-228. Hoar, W.S. 1951. The behavior of chum, pink and coho salmon in relation to their seaward migration. Journal of the Fisheries Research Board of Canada 8:241-263. Hunter, J.R. (ED.) 1976. Report of a colloquium on larval fish mortality studies and their relation to fishery research, January 1975. NOAA Technical Report CIRC-395, 5p. lies, T.D. and M. Sinclair. 1982. Atlantic herring: stock discreteness and abun-dance. Science 215:627-633. Ito, T. 1971. The biology of a harpacticoid copepod, Harpacticus uniremis Kroyer. Journal of the Faculty of Science of the Hokkaido University, Ser.6, Zoology 18:235-255. Ivlev, V.S. 1961. Experimental Ecology of the Feeding of Fishes, translated by D. Scott. Yale University Press, New Haven. 302 p. Jenkins, G.M. and D.G. Watts. 1968. Spectral analysis and its applications. Hoiden-Day series in Time series Analysis. San Francisco. 525 p. Jewett, S.C. and H.M. Feder. 1977. Biology of the harpacticoid copepod, Harpacti-cus uniremis Kroyer, on Dayville Flats, Port Valdez, Alaska. Ophelia 16:111-129. Johnson, R.G. 1965. Temperature variation in the infaunal environment on a sand fiat. Limnology and Oceanography 10:114-120. Kaczynski, V.W., R.J. Feller, J. Clayton, and R.J. Gerke. 1973. Trophic analysis of juvenile pink and chum salmon (Oncorhynchus gorbuscha and O. keta) in Puget Sound. Journal of the Fisheries Research Board of Canada 30:1003-1008. Kinne, O. 1970. Temperature, in O. Kinne (ed.) p. 321-346. Marine Biology, vol. 1, part 1. Wiley-Interscience. 138 Krebs, C.J. 1978. A review of the Chitty hypothesis of population regulation. Canadian Journal of Zoology 56:2463-2480. Lang, K. 1948. Monographie der Harpacticiden. 2 vols. H. Ohlssons, Lund. 1882 p. Lang, K . 1965. Copepoda Harpacticoidea from the Californian Pacific Coast. Kungl. Svenska Vetenskapsakademiens Hanlingar 10:1-560. Larkin, P.A. 1973. Some observations on models of stock and recruitment rela-tionships for fishes. Rapports et Proces Verbaux 164:316-324. Larkin, P.A. 1977. Pacific salmon, in J.A. Gulland (ed.) p. 157-186. Fish Population Dynamics. John Wiley &; Sons. Larkin, P.A. 1978. Fisheries management-An essay for ecologists. Annual Review of Ecology and Systematics 9:57-73. Larkin, P.A. (Rapporteur) 1984. Strategies for multi-species management, in R.M. May (ed.) p. 287-301. Exploitation of Marine Communities. Dahlem Konferenzen 1984. Springer-Verlag. Lasker, R. 1975. Field criteria for survival of anchovy larvae: the relation between inshore chlorophyll maximum layers and successful first feeding. Fishery Bulletin 73:453-462. LeBlond, P.H. 1983 The Strait of Georgia: functional anatomy of a coastal sea. Canadian Journal of Fisheries and Aquatic Sciences 40:1033-1063. Legendre, L. and S. Demers. 1984. Towards dynamic biological oceanography and limnology. Canadian Journal of Fisheries and Aquatic Sciences 41:2-19. Levings, C D . , R.E. Foreman, and V.J. Tunnicliffe. 1983. Review of the benthos of the Strait of Georgia and contiguous fjords. Canadian Journal of Fisheries and Aquatic Sciences 40:1120-1141. Levy, D.A., T.G. Northcote, and G.J. Birch. 1979. Juvenile salmon utilization of tidal channels in the Fraser river estuary, British Columbia. Westwater Research Center Technical Report No. 23 70 p. Marcotte, B.M. 1984. Behaviourally defined ecological resources and speciation in Tisbe (Copepoda: Harpacticoida). Journal of Crustacean Biology 4:404-416. McConnell, R.J. and G.R. Snyder. 1972. Key to field identification of anadromous juvenile salmonids in the Pacific Northwest. NOAA Technical Report NMFS Circ-366. 6p. Montagna, P.A., B.C. Coull, T.L. Herring, and B.W. Dudley. 1983. The relation-ship between abundances of meiofauna and their suspected microbial food (diatoms and bacteria). Estuarine, Coastal and Shelf Science 17:381-394. Morey, E.R. 1980. Fisheries resources: an introduction and review. Natural Resources Journal 20:827-851. 139 Muus, B.J. 1967. The fauna of Danish estuaries and lagoons. Meddr Danm. Fisk. -og Havenders.,N.S. 5:1-316. Neave, F. 1953. Principles affecting the size of pink and chum salmon populations in British Columbia. Journal of the Fisheries Research Board of Canada 9:459-491. Nicholson, A.J. 1933. The balance of animal populations. Journal of Animal Ecology 2:132-178. Parker, R.R. 1966. Estimation of sea mortality rates for the 1961 brood-year pink salmon of the Bella Coola area, B.C. Journal of the Fisheries Research Board of Canada 22:1523-1554. Parker, R.R. 1971. Size selective predation among juvenile salmonid fishes in a British Columbia inlet. Journal of the Fisheries Research Board of Canada 28:1503-1510. Parsons, T.R., M. Takahashi, B. Hargrave. 1984 a. Biological Oceanographic Processes. Third Edition. Pergamon Press. 330 p. Parsons, T.R., Y. Maita, and C M . Lalli. 1984 b. A Manual of Chemical and Biological Methods for Seawater Analysis. Pergamon Press. 173 p. Parsons, T.R. and T.A. Kessler. 1986. Computer model analysis of pelagic ecosys-tems in estuarine waters, in S. Skreslet (ed.) p. 161-181. The Role of Freshwater Outflow in Coastal Marine Ecosystems. Springer-Verlag. Parsons, T.R. and T.A. Kessler. 1987. An ecosystem model for the assessment of plankton production in relation to the survival of young fish. Journal of Plankton Research 9:125-137. Peterman, R.M. and M.J. Bradford. 1987. Wind speed and mortality rate of a marine fish, the Northern Anchovy (Engraulis mordax). Science 235:354-356. Phillips, A.C. 1977. Key field characters of use in identifying young marine Pacific salmon. Fishery and Marine Service Technical Report No. 746. 13 p. Ricker, W.E. 1954. Stock and recruitment. Journal of the Fisheries Research Board of Canada 11:559-623. Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bulletin of the Fisheries Research Board of Canada 191:382 p. Ricker, W.E. 1976. Review of the rate of growth and mortality of Pacific salmon in salt-water, and non-catch mortality caused by fishing. Journal of Fisheries Research Board of Canada 33:1483-1524. Ricker, W.E. 1985. Varieties and properties of three recruitment curves. La Mer 23:153-157. 140 SEP-Annual Report Summary 1983. Salmonid Enhancement Program. Depart-ment of Fisheries and Oceans, 1090 West Pender Street, Vancouver, B.C. 16 P Sibert, J.R. 1979. Detritus and juvenile salmon production in the Nanaimo estu-ary.: II. Meiofauna available as food to juvenile chum salmon (Oncorhynchus keta). Journal of the Fisheries Research Board of Canada 36:497-503. Sibert, J.R. 1981. Intertidal hyperbenthic populations in the Nanaimo estuary. Marine Biology 64:259-365. Sibert, J.R., B.A. Kask, and T.J. Brown. 1977. A diver-operated sled for sampling the epibenthos. Fisheries and Marine Service Technical Report No. 738. 19 P-Sinclair, M. and M.J. Tremblay. 1984. Timing of spawning of Atlantic her-ring (Clupea harengus harengus) populations and the match-mismatch theory. Canadian Journal of Fisheries and Aquatic Sciences 41:1055-1065. Sissenwine, M.P. 1984. Why do fish populations vary? in R.M. May (ed.) Ex-ploitation of marine communities. Dahlem Konferenzen 1984. Springer-Verlag. Steele, J.H. 1984. Kinds of uncertainty affecting fisheries, in R.M. May (ed.) p. 245-262. Exploitation of Marine Communities. Dahlem Konferenzen 1984. Springer-Verlag. Steele, J.H. and E.W. Henderson. 1984. Modelling long-term fluctuations in fish stocks. Science 224:985-987. Swinbanks, D.D. and J.L. Luternauer. 1987. Burrow distribution of Thalassinidean shrimp on a Fraser Delta tidal flat, British Columbia. Journal of Paleontology 61:315-332. Thompson, W.F. 1937. Theory of the effect of fishing on the stoch of halibut. Rep. Intern. Fish. Comm. 12:5-22. Uhlig, G.H., H. Thiel, and J.S. Gray. 1973. The quantitative separation of meio-fauna. Helgolander wissenschaftliche Meeresuntersuchungen 25:173-195. Valiela, D. and R.U. Kistritz. 1980. Dependence of salmon on Fraser estuar-ine marsh ecosystems: a simulation analysis. Westwater Research Center Technical Report No. 23 70 p. Volk, E . C , R . C Wissmar, C.A. Simenstad, and D.M. Eggers. 1984. Relationship between otolith microstructure and the growth of juvenile chum salmon (On-corhynchus keta) under different prey rations. Canadian Journal of Fisheries and Aquatic Sciences 41:126-133. Vugts, H.F. and J.T.F. Zimmerman. 1975. Interactions between the daily heat balance and the tidal cycle. Nature 255:113-117. 141 Vugts, H . F . and J.T.F. Zimmerman. 1985. The heat balance of a tidal flat area. Netherlands Journal of Sea Research 19:1-14. Waldichuck, M. 1957. Physical oceanography of the Strait of Georgia, British Columbia. Journal of Fisheries Research Board of Canada 14:321-486. Walters, C . J . 1984. Managing fisheries under biological uncertainty, tn R.M. May (ed.) p. 263-274. Exploitation of Marine Communities. Dahlem Konferenzen 1984. Springer-Verlag. Wells, J.B.J. 1976. Keys to Aid in the Identification of Marine Harpacti-coid Copepods. Department of Zoology, University of Aberdeen. 215 p. Wickett, W.P. 1958. Review of certain environmental factors affecting the pro-duction of pink and chum salmon. Journal of the Fisheries Research Board of Canada 15:1103-1126. Wooster, W.S. (ED.) 1983. From year to year. Washington Sea Grant, Seattle, WA. 208 p. Wroblewski, J.S. 1983. The role of modeling in biological oceanography. Ocean Science and Engineering 8:245-285. 142 Appendix A KEY CODE FOR 1985 GUT CONTENTS #=fish number l=fish fork length (mm) \=Harpacticus uniremis 2= Zaus aurelii 3=Tisbe spp. 4=Calanoids 5=Amphipods 6=UNC0 l=Scutellidium spp. 8=Fish larvae 9=Cumaceans 10=Cypris l l = Ostracods 143 M a r c h 29, 1985 (day 88) # 1 1 2 3 4 5 6 7 8 9 10 11 1 48 18 9 4 13 7 1 1 17 2 42 74 34 182 3 14 2 3 85 62 8 4 3 8 4 51 12 8 34 6 1 2 1 5 111 27 60 2 36 7 6 43 19 1 3 3 1 3 7 44 11 53 10 4 1 8 54 15 54 56 12 1 1 9 43 4 1 5 12 10 46 36 10 8 3 2 11 57 13 27 30 26 1 3 12 41 16 14 3 2 13 49 26 2 1 5 14 46 1 2 4 4 1 4 15 45 13 1 1 1 1 1 16 51 6 3 2 2 5 17 40 24 8 6 1 18 40 36 5 8 3 1 19 49 41 11 22 20 54 45 22 6 22 6 A p r i l 10, 1985 (day 100) # 1 1 2 3 4 5 6 7 8 9 10 11 1 51 46 4 3 1 2 57 102 157 1 1 1 3 45 72 170 2 2 4 52 104 202 1 1 2 1 5 52 122 88 2 8 5 6 56 113 30 1 7 53 122 65 1 2 8 51 68 12 2 4 1 9 57 74 7 1 3 3 7 10 58 135 116 8 3 4 2 11 55 143 92 3 1 2 1 1 12 54 146 124 1 4 2 13 55 46 55 1 2 4 10 14 63 5 244 5 1 2 2 4 15 57 101 50 10 3 10 1 16 51 55 3 1 17 58 64 2 2 3 18 42 54 41 2 4 19 51 53 60 4 5 1 20 47 40 34 4 145 A p r i l 23, 1986 (day 113) # 1 1 2 3 4 5 6 7 8 9 10 11 1 46 1 1 1 2 36 19 3 13 1 6 4 2 3 56 50 48 3 11 11 8 4 48 24 2 1 1 1 5 48 56 5 2 6 43 29 1 5 7 58 13 1 19 1 15 1 13 8 67 11 1 3 5 .7 9 50 14 4 9 2 4 10 45 137 2 4 14 20 11 44 16 3 1 2 M a y 9, 1985 (day 129) # 1 1 2 3 4 5 6 7 8 9 10 11 1 51 221 6 48 11 2 50 155 1 23 160 1 1 3 57 328 78 43 12 9 2 4 47 413 14 2 6 2 5 46 346 1 4 13 1 6 65 252 12 100 16 1 8 7 64 339 27 3 24 3 34 8 65 362 3 28 4 9 9 44 285 7 13 1 10 51 411 2 2 1 3 11 59 711 1 34 4 10 12 52 455 1 1 5 2 4 13 52 455 1 1 5 2 4 146 M a y 23, 1985 (day 143) # 1 1 2 3 4 5 6 7 8 9 10 11 1 78 3 2 52 11 1 10 7 30 7 3 47 34 5 22 2 9 2 4 61 1 2 1 1 5 62 12 72 14 1 28 10 7 27 6 50 37 69 18 3 9 84 4 7 48 23 2 2 22 1 8 49 16 41 4 23 9 48 6 1 53 6 25 10 63 3 3 J u n e 7, 1985 (day 157) # 1 1 2 3 4 5 6 7 8 9 10 11 1 53 6 12 16 5 50 80 2 57 2 6 66 3 7 2 3 51 6 8 113 7 4 35 11 4 56 1 3 3 150 117 4 18 5 50 2 20 10 17 25 4 30 6 61 1 6 6 25 11 6 39 7 51 3 8 9 11 1 7 8 57 12 9 120 12 20 9 57 2 7 12 27 17 12 60 10 57 2 136 1 147 June 21, 1985 (day 172) # 1 1 2 3 4 5 6 7 8 9 10 11 1 66 3 39 2 69 2 2 1 o o 50 3 260 6 6 6 4 81 8 2 236 5 62 6 288 6 60 6 224 3 25 2 145 7 62 325 8 56 2 15 5 4 2 6 169 9 51 12 444 3 2 12 40 1 10 64 4 2 July 5, 1985 (day 188) # 1 1 2 3 4 5 6 7 8 9 10 11 1 48 21 10 7 4 2 50 33 10 4 12 3 1 1 3 60 23 4 19 29 5 4 58 1 3 2 39 5 51 18 4 1 3 8 3 12 148 KEY CODE FOR 1986 GUT CONTENTS #=fish number l=fish fork length (mm) 1 = Harpacticus uniremis 2=Zaus aurelii 3=Tisbe spp. 4 = Calanoids 5 = Amphipods 6=Longipedia sp. 7=Orthopsyllus sp. 8=UNC0 9=Scutellidium sp. 10=Isopods ll=Fish larvae 12=Diptera 13=Dacty-lopodia sp. 14=Cumaceans 15=Cypris 16=Ostracods 17' = Diarthrodes sp. 149 A p r i l 1, 1986 (day 91) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 43 53 4 96 7 56 1 6 2 46 1 6 3 45 16 90 1 4 46 5 40 3 5 5 45 31 3 5 2 10 6 43 1 2 1 7 41 15 8 1 11 7 1 8 44 15 10 1 2 9 41 80 11 8 15 16 10 39 120 4 25 5 10 11 54 70 1 3 23 3 3 A p r i l 15, 1986 (day 132) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 46 32 9 3 3 2 43 1 1 1 3 59 8 3 3 7 4 64 4 2 12 5 38 18 3 6 1 6 50 3 3 3 1 7 51 3 5 1 6 8 37 17 3 1 1 3 9 43 28 11 1 3 10 59 2 2 14 2 11 39 3 7 1 3 12 38 4 3 1 13 42 34 5 5 1 150 A] pri l 29, 1986 (day 119) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 36 100 5 1 4 14 2 37 66 1 47 1 1 3 36 3 6 1 9 2 4 35 79 2 69 4 5 38 100 2 13 7 7 6 40 41 3 96 1 7 7 35 110 7 2 9 8 39 129 2 18 6 1 5 9 38 31 76 9 1 10 36 52 1 20 4 1 M a y 13, 1986 (day 132) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 47 24 3 22 15 5 10 4 2 2 48 15 10 22 4 1 6 3 47 1 1 3 4 48 14 63 6 2 1 5 45 65 1 3 9 1 8 6 44 45 2 37 52 8 1 7 43 43 30 37 3 1 2 2 8 39 20 17 5 15 2 2 9 47 22 7 11 1 5 2 10 39 5 2 60 2 2 2 1 151 M a y 27, 1986 (day 147) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 45 15 35 2 6 1 2 50 20 60 1 5 3 1 3 42 1 10 2 4 48 21 55 2 5 49 60 150 1 2 10 6 53 33 50 7 10 3 7 48 25 92 8 45 15 12 13 6 9 54 8 80 2 8 10 42 e m p t y s t o m a c h J u n e 11 , 1986 (day 162) # 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 1 52 8 120 2 4 12 8 2 54 5 26 45 6 2 2 3 64 6 2 1 14 4 44 15 2 4 5 54 8 22 5 17 6 57 2 210 4 10 6 7 54 56 65 1 7 2 17 8 52 28 40 3 4 6 9 58 6 42 8 4 37 10 56 2 90 6 26 11 54 4 110 9 7 12 51 1 17 1 6 5 21 152 Appendix B KEY CODE FOR 1985 AND 1986 SLED SAMPLES CONTENTS AMPH=amphipods CALA=calanoids CAPR=caprellids CUMA=cumaceans CYPR=cypris DACT= Dactylopodia sp. DIAR= Diarthrodes sp. DIOS=Diosaccus sp. FLAR=fish larvae GAST=gastropods ORTH= Orthopsyllus sp. HARP=Harpacticus uniremis ISOP=isopods MESO=Mesoc/irasp. OSTR=ostracods SCUT=Scutellidium sp. TISB= Tisbe spp. UCOP=unidentified copepodites ZAUS=Zaus aurelii 153 March 1, 1985 (day 60) Level A B C D E TISB 135 101 24 8 2 MESO 4 4 5 D1AR 4 ZAUS 2 1 DACT 1 1 1 AMPH 3 CALA 2 48 44 64 82 OSTR 2 UCOP 126 56 29 10 7 March 13, 1985 (day 72) Level A B C D E TISB 183 59 5 1 1 MESO 8 2 1 ZAUS 1 DACT 7 4 1 HARP 1 SCUT 1 AMPH 8 5 CALA 6 19 18 9 12 FLAR 1 UCOP 711 153 12 10 8 154 March 29, 1985 (day 88) Level A B C D E TISB 481 418 147 63 20 MESO 1 3 4 3 7 HARP 2 DACT 1 2 2 AMPH 2 CALA 2 2 2 5 3 UCOP 515 440 208 113 43 April 10, 1985 (day 100) Level A B C D E TISB 514 223 48 19 6 MESO 1 2 1 1 DIAR 1 ZAUS 2 2 DACT 3 1 1 HARP 1 1 2 AMPH 5 CALA 10 11 11 1 4 CAPR 1 CYPR 9 21 25 24 UCOP 449 182 45 13 10 April 23, 1985 day 113) Level A B C D E TISB 14 9 12 5 DACT 1 HARP 1 AMPH 3 2 1 1 CALA 1 CYPR 2 CAPR 1 UCOP 37 12 14 5 3 May 9,1985 (day 129) Level A B C D E TISB 47 43 27 6 9 ZAUS 1 1 HARP 1 AMPH 40 18 10 6 CALA 147 78 61 14 7 CYPR 2 2 1 UCOP 175 129 76 45 20 May 23, 1985 (day 143) Level A B C D E TISB 7 3 DIAR 1 AMPH 19 2 4 9 CAPR 1 CYPR 1 6 1 1 UCOP 173 183 176 117 80 June 7, 1985 (day 157) Level A B C D E TISB 6 9 14 11 12 MESO 3 6 3 1 2 DIAR 2 1 ZAUS 1 1 1 AMPH 8 2 1 1 1 CALA 107 112 105 84 77 UCOP 161 151 93 105 95 156 • June 21, 1985(day 172) Level A B C D E TISB 9 2 2 MESO 12 7 4 DIAR 1 1 1 HARP o o 2 ZAUS 2 DACT 4 1 2 AMPH 18 10 2 CALA 50 39 41 28 16 CAPR 8 6 CYPR 1 UCOP 282 276 85 55 31 July 5, 1985 (day 188) Level A B C D E TISB 1 MESO 2 DIAR 2 DACT 2 1 HARP 2 AMPH 20 31 CALA 20 6 14 19 13 CUMA 1 UCOP 1172 361 216 106 67 157 February 11, 1986 (day 42) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 35 13 7 4 1 49 27 14 2 1 54 48 20 3 1 MESO 35 25 14 5 1 31 19 17 8 5 134 42 24 6 4 DACT 1 1 1 1 4 1 2 2 3 2 ZAUS 2 1 1 1 3 1 4 1 1 DIOS 1 AMPH 2 1 GAST 2 1 1 4 9 4 5 OSTR 4 1 7 3 3 1 13 5 1 1 CAPR 1 1 1 CALA 2 3 2 5 3 1 8 UCOP 1 3 3 1 1 1 3 2 2 1 February 24, 1986 (day 55) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 43 21 2 1 1 44 38 8 1 3 40 17 4 1 1 MESO 18 14 3 1 27 11 7 4 39 20 6 2 4 DACT 1 3 2 1 1 2 1 4 2 ZAUS 1 1 DIAR 1 DIOS 2 1 2 2 2 OSTR 3 3 3 3 CALA 4 2 1 2 1 1 6 7 14 1 AMPH 1 1 1 1 1 GAST 1 1 5 UCOP 5 3 2 1 4 9 1 3 4 2 1 158 March 5, 1986 (day 64) Transect. 1 2 3 Level A B C D E A B C D E A B C D E TISB 136 60 22 12 7 138 91 42 27 1 130 108 60 27 11 H A R P 1 1 M E S O 63 34 26 7 1 50 25 26 10 6 62 29 13 8 10 D A C T 10 3 2 2 7 7 3 1 6 5 1 ZAUS 2 2 2 1 1 3 1 1 DIOS 1 4 1 11 12 7 4 2 11 14 4 2 OSTR 11 1 1 7 2 8 8 1 2 A M P H 9 5 2 1 11 5 2 2 4 4 2 C A L A 3 1 2 5 1 3 4 3 3 1 5 G A S T 1 ISOP 5 C Y P R 3 O R T H U C O P 23 11 3 3 9 . 4 1 6 3 22 10 2 3 March 19, 1986 (day 78) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 486 321 150 90 28 490 279 160 76 44 474 205 74 17 10 M E S O 83 43 34 51 33 23 81 25 19 18 54 35 20 12 2 D A C T 3 1 1 2 1 1 2 2 2 H A R P 2 2 4 5 1 2 4 2 1 1 3 3 2 ZAUS 3 6 21 11 2 8 25 21 18 1 4 17 8 2 1 DIOS 6 2 7 5 2 1 2 4 2 1 1 S C U T 3 1 4 1 DIAR 18 5 1 4 5 8 6 1 5 3 10 7 3 4 A M P H 10 13 10 3 1 23 11 2 3 11 8 1 C A L A 3 3 1 6 2 2 3 1 O S T R G A S T 3 8 2 8 1 9 3 1 1 6 U C O P 1453 647 276 201 101 2070 630 370 194 138 1435 719 224 79 31 159 I April 1, 1986 (day 91) Transect 1 2 3 Level A B C D E A B C D E A B C D | E TISB 347 294 183 124 75 312 265 208 102 93 273 277 186 104 86 M E S O 32 40 5 14 16 27 10 14 9 25 17 13 11 16 11 DIAR 3 2 1 1 3 1 1 5 4 1 D A C T 2 1 1 3 1 1 5 4 1 H A R P 6 6 4 1 3 7 8 2 6 4 2 3 4 2 ZAUS 8 47 58 36 15 20 42 44 76 50 7 45 79 60 25 S C U T 1 4 8 3 1 2 5 3 1 A M P H 16 6 5 2 7 19 10 3 9 15 7 2 C A L A 5 5 1 1 1 4 C A P R 1 2 1 2 3 1 U C O P 1454 1018 538 215 164 2269 1128 633 276 210 I 1996 1496 630 255 194 April 15, 1986 (day 105) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 690 514 165 88 32 668 561 244 119 63 613 459 253 123 42 M E S O 58 37 10 9 6 43 9 2 6 5 35 13 11 8 3 DIAR 2 1 H A R P 18 13 3 5 2 10 4 7 2 3 17 6 6 3 2 ZAUS 20 28 28 16 13 14 48 72 32 5 14 48 118 72 21 D A C T 9 5 1 4 1 1 7 2 S C U T 3 3 A M P H 7 5 2 1 3 5 3 1 2 2 4 1 1 C A L A 13 10 14 11 13 13 16 10 6 28 10 12 11 8 C A P R 2 1 3 U C O P 1345 533 272 196 126 1070 552 403 206 152 1167 678 395 244 130 1 6 0 April 29, 1986 (day 119) Transect. 1 3 Level A | B | C D E A B C D E A B C | D E TISB 416 136 103 82 28 256 134 105 66 54 387 150 121 74 43 M E S O 82 70 69 55 18 54 41 35 60 39 102 34 47 38 20 DIAR 1 3 D A C T 23 26 5 9 1 22 14 11 19 11 37 11 18 8 3 H A R P 22 19 11 13 4 16 20 7 20 12 43 19 18 2 1 ZAUS 81 93 62 45 8 62 78 117 73 34 90 105 194 58 24 S C U T 2 2 4 1 A M P H 11 9 5 5 3 8 46 10 1 6 6 12 2 7 7 C A L A 37 19 5 9 8 16 9 7 9 4 14 5 10 3 3 C A P R 2 1 3 U C O P 2775 2173 1297 1551 643 1802 1239 1036 1364 849 2982 1170 941 1031 807 May 13, 1986 (day 132) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 95 30 37 27 13 20 26 29 15 10 4 49 69 37 35 M E S O 25 10 8 5 2 9 5 3 4 9 11 7 3 2 D A C T 11 3 3 7 1 3 1 2 1 2 5 4 1 1 H A R P 30 10 11 4 6 12 12 2 4 19 26 24 5 3 ZAUS 2 2 2 4 1 1 2 1 1 DIOS 1 1 A M P H 7 4 2 2 1 3 1 1 2 4 7 2 2 C A L A 79 11 6 4 4 14 14 6 5 2 14 24 14 10 C A P R 2 1 2 1 3 2 C U M A 1 U C O P 933 317 364 319 154 197 234 369 173 183 138 629 653 481 373 161 May 27, 1986 (day 147) Transect 1 2 3 Level A B C D E A B C D E A B . C D E TISB 14 21 31 25 10 47 39 41 22 5 41 49 47 32 16 M E S O 11 28 13 8 . 3 21 21 22 14 3 12 12 7 10 4 D A C T 7 13 15 11 3 5 4 5 7 8 6 12 14 8 6 H A R P 3 16 5 3 6 3 3 3 3 4 6 7 1 2 ZAUS 1 1 2 5 7 3 1 4 1 3 5 2 7 2 DIAR 2 1 1 1 SCUT 1 2 A M P H 1 4 8 5 1 7 3 3 2 2 5 4 1 1 C A L A 20 30 25 48 22 43 27 26 17 7 55 64 59 62 33 C A P R 7 7 7 3 6 15 6 C U M A 2 1 1 1 U C O P 46 93 96 164 88 130 130 156 115 59 162 197 159 138 73 June 11, 1986 (day 162) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 3 2 2 3 2 2 2 2 1 M E S O 2 1 2 3 1 DIAR 3 3 1 1 1 1 1 2 1 H A R P 1 1 2 2 3 1 1 S C U T 2 2 1 D A C T 2 1 5 3 2 2 A M P H 5 2 1 1 2 C A P R 6 4 1 2 U C O P 49 22 19 12 9 24 16 25 23 22 43 29 36 18 15 162 June 25, 1986 (day 176) Transect 1 2 3 Level A B C D E A B C D E A B C D E TISB 6 3 10 4 1 4 1 1 4 M E S O 3 9 2 6 2 2 10 5 D A C T 28 21 12 27 12 6 7 15 22 11 9 DIAR 3 1 6 2 1 1 1 1 1 1 4 2 3 2 H A R P 1 1 1 2 ZAUS 10 16 66 73 72 36 32 64 83 61 15 15 49 63 28 SCUT 2 3 2 4 1 2 1 3 5 2 1 A M P H 8 2 2 1 1 5 1 2 C A L A 165 154 95 83 62 146 155 71 36 40 210 147 121 105 23 C A P R 1 1 1 C U M A 1 U C O P 76 83 81 47 39 48 45 41 30 16 81 71 55 60 21 163 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053321/manifest

Comment

Related Items