Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The relationship between marine survival rates of Robertson Creek chinook salmon (oncorhynchus tshawytscha)… Tovey, Christine Phyllis 1999

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

Item Metadata

Download

Media
831-ubc_1999-0387.pdf [ 7.07MB ]
Metadata
JSON: 831-1.0053284.json
JSON-LD: 831-1.0053284-ld.json
RDF/XML (Pretty): 831-1.0053284-rdf.xml
RDF/JSON: 831-1.0053284-rdf.json
Turtle: 831-1.0053284-turtle.txt
N-Triples: 831-1.0053284-rdf-ntriples.txt
Original Record: 831-1.0053284-source.json
Full Text
831-1.0053284-fulltext.txt
Citation
831-1.0053284.ris

Full Text

THE RELATIONSHIP BETWEEN MARINE SURVIVAL RATES OF ROBERTSON CREEK CHINOOK SALMON (ONCORHYNCHUS TSHAWYTSCHA) AND THEIR FIRST MARINE YEAR LENGTHS AND GROWTH RATES by CHRISTINE PHYLLIS TOVEY B.Sc. The University of British Columbia, 1996 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Earth and Ocean Sciences, Oceanography) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA September 1999 © Christine Phyllis Tovey, 1999 UBC Special Collections - Thesis Authorisation Form 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. Department of fa/ljU, rAAjQ. OreayLj. octPAxxJa The University of British Columbia Vancouver, Canada Date Muf. ?/,. /?99 -f http://www.library.ubc.ca/spcoll/thesauth.html 6/1/99 ABSTRACT The importance of smolt size and early ocean growth to the marine survival of chinook salmon was investigated over a 10 year period for Robertson Creek hatchery Chinook salmon (Oncorhynchus tshawytscha). Comparisons of marine survival were made with smolt lengths and growth rates back-calculated from the scales of age 0.3 female chinook entering the ocean between 1982 and 1991. Marine survival was not correlated with smolt length (r* = 0.29; P = 0.139). Marine survival was significantly positively correlated with growth rates estimated for the first 12 marine circuli (r2: 0.88 to 0.93; P < 0.01). These circuli formed in July and August as estimated from the number of days taken for marine scale growth to begin following hatchery release (34 d) and the rate of deposition of marine circuli (6.8 d per circulus). During this time the majority of juvenile chinook were found to reside in Barkley Sound adjacent to the northernmost portion of the temporally dynamic Coastal Upwelling Domain. The period of high correlation was found to occur within the summer upwelling season, when predominantly north-westerly winds result in the upwelling of nutrient-rich waters and a peak in zooplankton biomass. Marine growth rates are suggested to be a linear function of the summer plankton productivity -- a dome-shaped relationship was found between July-August growth rates and the magnitude of summer upwelling (Bakun Index averaged over June, July and August; r2 = 0.68; P = 0.06) and Cury and Roy (1989) found that the productivity of Ekman-type upwelling systems is a dome-shaped function of the magnitude of the upwelling favourable winds. The strong relationship between growth and survival is suggested to result from the benefits of fast growth to survival as well as from changes in predation intensity varying concurrently with oceanographic conditions indicative of productivity. In years when large numbers of Pacific mackerel {Scomber japonicus) migrated into the upwelling region, marine survival was low independent of the conditions for growth. i i TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS iii TABLE OF TABLES vii TABLE OF FIGURES ix ACKNOWLEDGEMENTS xiii GENERAL INTRODUCTION 1 CHAPTER 1. INVESTIGATION OF INTERANNUAL VARIABILITY IN FIRST YEAR LENGTHS AND GROWTH RATES: THE MEASUREMENT OF ADULT ROBERTSON CREEK CHINOOK {ONCORHYNCHUS TSHAWYTSCHA) SCALES 4 1.1 INTRODUCTION 4 1.2 METHODS AND MATERIALS .....5 1.2.1 Description of Salmon Scales 5 1.2.2 Adult Robertson Creek Chinook Scales 5 1.2.3 Scale Measurement Procedure 10 1.2.4 Scale Measurement Difficulties 13 1.2.5. Regression Analysis „... 13 1.3 RESULTS 15 1.3.1 Interannual Variation in Scale Measurements 15 - Scale Radius and Circuli Number at Ocean Entry 15 - Interannual Variation in Scale Radius and Circuli Number for the First Year 18 - Interannual Variation in Early Marine Growth Indices 21 1.3.2 Relationship Between Scale Measures 30 1.4 DISCUSSION 31 i n CHAPTER 2. BACK-CALCULATION OF FORKLENGTHS AND DETERMINATION OF MARINE GROWTH RATES FROM ADULT CHINOOK SCALE MEASUREMENTS 34 2.1 INTRODUCTION. 34 2.2 METHODS AND MATERIALS 35 2.2.1 Regression of Forklength on Scale Radius for Back-Calculation 35 - Data Source 35 - Regression Equation 36 2.2.2 Back-calculation of Forklengths and Length Increments from Scale Radii Using GM Regression 39 2.2.3 Conversion of Length Increments into Growth Rates 41 - Determination of the Apparent Rate of Circulus Addition Using Regression 41 - Estimate of the Circulus Deposition Rate Over the Entire First Year 42 - Estimate of the Circulus Deposition Rate Over the First Marine Year 43 - Summary of Apparent Rate of Circulus Addition 43 - Equation for the Conversion of Somatic Growth Increments into Growth Rates in mmcf1 44 - Time of Year that the Somatic Lengths and Growth Rates are Completed ..45 2.3 RESULTS 46 2.3.1 Back-Calculated Forklengths 46 2.3.2 Early Marine Growth Rates 49 2.3.3 Summary of Lengths and Growth Rates 54 2.4 DISCUSSION „ 55 CHAPTER 3. THE RELATIONSHIP BETWEEN MARINE SURVIVAL AND BACK-CALCULATED FIRST YEAR LENGTHS AND MARINE GROWTH RATES 59 3.1 INTRODUCTION 59 3.2 METHODS AND MATERIALS 60 3.2.1 Survival Rate Data.. 60 3.2.2 Regression Analysis 62 IV 3.3 RESULTS 62 3.3.1. Regression of Cohort Survival Rate on Back-Calculated Somatic Measures 62 3.4 DISCUSSION 69 CHAPTER 4. INVESTIGATION INTO THE MECHANISMS DETERMING INTERANNUAL VARIABILITY IN EARLY MARINE GROWTH AND SURVIVAL OF ROBERTSON CREEK CHINOOK SALMON 75 4.1 INTRODUCTION 75 4.2 METHODS AND MATERIALS 76 4.2.1 Estimation of the Period of High Correlation between Marine Survival and Growth 76 4.2.2 Location of the Majority of RCH Chinook throughout the Highly-Correlated Period 78 - Probable Location of Chinook at the Initiation of Marine Growth 78 - Probable Location of RCH Chinook by the End of the Highly-Correlated Period 81 4.3 RESULTS AND DISCUSSION ...81 4.3.1 Time Period of High-Correlation between Marine Survival and Growth 81 4.3.2 Location of Chinook throughout the Period of High Correlation between Survival and Growth 86 - Location of Chinook at the Initiation of Marine Growth 86 - Location of Chinook at the End of the Highly-Correlated Period. — 88 4.3.3 Summary of the Location of Chinook during the Time Period of High Correlation between Survival and Growth 89 4.3.3 Factors Influencing Early Marine Growth and Survival 89 - Oceanic Productivity 89 - Predation Intensity 93 V LITERATURE CITED 98 APPENDIX I INVESTIGATION INTO POSSIBLE SEXUAL DIMORPHISM IN SCALE MEASUREMENTS FOR AGE 0.3 ROBERTSON CREEK CHINOOK SALMON 107 1.1 INTRODUCTION 107 1.2 METHODS AND MATERIALS 107 1.3 RESULTS 108 1.4 DISCUSSION 111 APPENDIX II CONVERSION OF THE MASS SCALE MEASUREMENTS FROM THE LONGEST AXIS TO THE STANDARD AXIS 20 DEGRESS VENTRAL TO THE LONGEST AXIS 113 11.1 INTRODUCTION 113 11.2 METHODS AND MATERIALS 113 11.3 RESULTS 114 11.4 DISCUSSION 114 APPENDIX III. TABLES 118 VI TABLE OF TABLES Table 1.1. Scale radius to the end of the SWE check: P two tail from two sample t-test assuming unequal variances 17 Table 1.2. The mean, minimum, and maximum average number of circuli to the end of the SWE check, the first marine annulus, and between the SWE check and the first marine annulus for 10 years (1982-83; 1985-92) 18 Table 1.3. Scale radius to the first marine year: P two tail from two sample t-test assuming unequal variances 20 Table 1.4. Intercircular distances averaged over the 3rd to 5th marine circuli: P two tail from two sample t-test assuming unequal variances ;... 29 Table 2.1. The equations for the regression of circulus number on days since first and median release date for 1989 (n = 188) and 1990 (n = 125) data separate and combined (n = 313) 42 Table 2.2. Summary of the periodicity of circulus deposition determined from several methods: regressions of circulus number on days since release using 1989 and 1990 juvenile Chinook separately and combined; and two additional calculations. .44 Table 2.3. Average and standard deviation of the back-calculated length at ocean entry (smolt length), length added from the end of the SWE check to the end of ...46 Table 2.4. The a) average and b) standard deviation of growth rates (mmd"1) for 1982-83; 85-92. The growth rates are averaged over three consecutive marine circuli from 2 to 12 50 Table 2.5. Summary of the number of circuli, the coefficient of variation (CV), and the approximate date of formation for the lengths and growth rates back-calculated from adult scales 54 Table 3.1. Equations for the linear regression of marine survival rate on the somatic lengths and growth rates back-calculated from the average scale values for 1982-83; 1985-92 (n = 10) 63 Table 3.2. Equations for the linear regression of marine survival rate on the lengths and growth rates back-calculated from average scale values for 1982-83; 1985-91 (n = 9) 68 Table 1.1. T-test results for 1983 ocean entry year first year scale features grouped by sex 99 Table I.2. T-test results for 1986 ocean entry year first year scale features grouped by sex 99 vii Table 1.3. T-test results for 1988 ocean entry year first year scale features grouped by sex 100 Table I.4. T-test results for 1989 ocean entry year first year scale features grouped by sex 100 Table 1.5. Summary of the significantly different t-test results for the 1983,1986,1988, and 1989 ocean entry years 101 Table 111.1. Descriptive statistics for the scale radius to the end of the salt water entry check 110 Table III.2. Descriptive statistics for the number of circuli to the end of the salt water entry check 110 Table III.3. Descriptive statistics for the scale radius to the end of the first marine annulus 111 Table III.4. Descriptive statistics for the number of circuli to the end of the first marine annulus 111 Table III.5. Descriptive statistics for the width of the first marine scale zone 112 Table 111.6. Descriptive statistics for the number of circuli in the first marine scale zone. 112 Table III.7. Descriptive statistics for the intercircular distance averaged over marine circuli 3 to 5 113 Table III.8. Descriptive statistics for the intercircular distance averaged over marine circuli 8 to 10 113 Table III.9. A) Conversion of calendar day to day of the month. B) Table to determine the calendar day of the first 12 marine circuli from the release dates between day 130 and 162 114 Vlll TABLE OF FIGURES Figure 1.1. Age 0.3 Robertson Creek Chinook scale diagram 6 Figure 1.2. Chart of Alberni Inlet and Barkley Sound 7 Figure 1.3. The average and standard deviation of the scale radius to the end of the salt water entry check (SWE) (10"2 mm) for 1982-83; 85-92 16 Figure 1.4. The average and standard deviation of the number of scale circuli to the end of the salt water entry check for 1982-83; 85-92 16 Figure 1.5. The average and standard deviation of the width of the first marine scale zone (10"2 mm for 1982-83; 85-92 19 Figure 1.6. The average and standard deviation of the scale radius to the end of the first marine annulus (10"2 mm) for 1982-83; 85-92 19 Figure 1.7. The average and standard deviation of the number of circuli in the first marine zone for 1982-83; 85-92 22 Figure 1.8. The average and standard deviation of the number of scale circuli present before the end of the first marine annulus for 1982-83; 85-92 22 Figure 1.9. Intercircular distances (10"2 mm) for the first 15 marine scale circuli 23 Figure 1.10. Average intercircular distance (10'2 mm) for the first 28 marine circuli for ocean entry years 1983, 85, 88, 89 and 90 23 Figure 1.11. An early marine growth index (Intercircular distance averaged over the 4th to 9th marine circuli) and an index of growth later in the first marine year (intercircular distance averaged over the 21s t to 25th marine circuli) for 1982-83; 85-92 24 Figure 1.12. Autocorrelation plot of the 2nd to 13th marine circuli up to a lag of 13 for 1985 and 1988 26 Figure 1.13. The mean and standard deviation of the intercircular distance (10"2 mm) averaged on the 3rd to 5th marine circuli (Av3-5) for 1982-3; 85-92 27 Figure 1.14. The mean and standard deviation of the intercircular distance (10"2 mm) averaged on the 8th to 10th marine circuli (Av8-10) for 1982-83; 85-92 27 Figure 1.15. The mean and standard deviation of the intercircular distance (10"2 mm) averaged over the 11th to 13th marine circuli (Av11-13)for 1982-83; 85-92 28 IX Figure 1.16. The mean and standard deviation of the intercircular distance (10"2 mm) averaged over the 13th to 15th marine circuli (Av13-15) for 1982-83; 85-92 28 Figure 2.1. Arithmetic mean (AM) regression of forklength on scale radius 37 Figure 2.2. Geometric mean (GM) regression of forklength on scale radius 37 Figure 2.3. Linear regression of the number of circuli present on the scales of juvenile Chinook caught in 1989 and 1990 and the number of days from median release date to recapture 40 Figure 2.4. Mean and standard deviation of the length at the end of the SWE check (smolt size) for 1982-83; 1985-92 47 Figure 2.5. Mean and standard deviation of length added from the SWE check to the end of the first marine annulus for 1982-83; 1985-92 47 Figure 2.6. Mean and standard deviation of length at the end of the first marine year for 1982-83; 1985-92 48 Figure 2.7. Somatic growth rates (mmd'1) for 1982-83; 85-1992 averaged over three consecutive marine circuli between 2 to 12 51 Figure 2.8. Average and standard deviation of the growth rate (mmd"1) averaged over the 3rd to 5th marine circuli (Av3-5) 52 Figure 2.9. Average and standard deviation of the growth rate (mmd"1) averaged over the 5th to 7th marine circuli (AV5-7) 52 Figure 2.10. Average and standard deviation of the growth rate (mmd"1) averaged over the 8th to 10^ marine circuli (Av8-10) 53 Figure 2.11. Average and standard deviation of the growth rate (mmd"1) averaged over the 10th to 12th marine circuli (Av10-12) 53 Figure 3.1 Marine survival rate (%) for cohorts of Robertson Creek Chinook salmon entering the ocean between 1982 and 1992 61 Figure 3.2. Linear regression and 95% confidence bands for marine survival and forklength (mm) to the end of the salt water entry check (FW length) for 1982-83; 85-92 ocean entry years 64 Figure 3.3. Linear regression and 95% confidence bands for marine survival and forklength (mm) added in the first marine zone (Marl Length) for 1982-83; 85-92 ocean entry years 64 X Figure 3.4. Linear regression and 95% confidence bands for marine survival and forklength (mm) to the end of the first marine year for 1982-83; 85-92 ocean entry years 65 Figure 3.5. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd"1) averaged over the 3rd to 5th marine circuli (Av3-5) for 1982-83; 85-92 ocean entry years 65 Figure 3.6. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd"1) averaged over the 4th to 6th marine circuli (Av4-6) for 1982-83; 85-92 ocean entry years 66 Figure 3.7. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd"1) averaged over the 8th to 10th marine circuli (Av8-10) for 1982-83; 85-92 ocean entry years 66 Figure 3.8. Line graphs of marine survival rates (%) and growth rates (mmd"1) averaged over the a) 3rd to 5th , b) 4th to 6th, and c) 8th to 10th marine circuli 67 Figure 4.1. Frequency histograms of the number of scale circuli found on juvenile Chinook salmon in Areas 1-4 77 Figure 4.2. The coefficient of determination (r2 value) for the regression of marine survival on the growth rates (mmd"1) averaged over three consecutive marine circuli from 2 to 30 79 Figure 4.3. Chart of Alberni Inlet and Barkley Sound 80 Figure 4.4. Chart of Alberni Inlet and Barkley Sound. The average number of circuli on the scales of juvenile Robertson Creek chinook salmon is presented for each capture location 82 Figure 4.5. a) Sockeye salmon b) coho salmon, and c) chinook salmon catch per unit of effort (CPUE) by date for 1987 to 1990 MASS sampling in Barkley Sound 84 Figure 4.6. The apparent migration rate for juvenile chinook from the release site (Somass Estuary) to the location of capture in Barkley Sound vs. the number of circuli that they possess on their scales ....85 Figure 4.7. Growth rate averaged over the 3rd to 10th marine circuli (mmd"1) regressed on the Bakun upwelling index averaged over June, July and August for 1982; 85-91 92 Figure 4.8. Linear regression of marine survival rate (%) on hake biomass (thousands of tonnes) for 1980-1992 92 XI Figure 4.9. Marine survival rate of chinook vs. the Bakun upwelling index averaged over April, May, June, July, and August) for 1980-1992 94 Figure 4.10. Marine survival rate of chinook vs. the Bakun upwelling index averaged over April, May, June, July, and August) for 1980-82; 85-91 94 Figure 11.1. Regression of 20 degree ventral measurements on longest axis measurements of the scale radius to the end of the SWE check (FW radius): 1986 and 1982 ocean entry years 107 Figure II.2. Regression of 20 degree ventral axis measurements on the longest axis measurements for the width of the first marine scale zone: 1986 and 1982 ocean entry years 107 Figure II.3. Regression of 20 degree ventral measurements on longest axis measurements for the scale radius to the end of the first marine year (Marl radius): 1986 and 1982 ocean entry years 108 xii ACKNOWLEDGEMENTS I would like to thank my two supervisors, Dr. Dan Ware and Dr. Al Lewis for their helpful advice and encouragement throughout the duration of this project. The many discussions were stimulating and very much appreciated. I would also like to thank Dr. Brent Hargreaves for sharing his knowledge of Chinook salmon and for kindly providing me with the juvenile Chinook database. Thank you also to the other members of my supervisory committee, Dr. Eric Taylor and Dr. Michael Healey, for many helpful and stimulating questions and comments. There were a number of people who aided my endeavors in this thesis. The fish ageing laboratory staff at the Pacific Biological Station who trained me in the methods of scale analysis and who graciously shared their laboratory. Ray Volk and Glen Rasmussen at the Robertson Creek hatchery and Susan Lehmann and Anne Martin of SEP who kindly provided me with historical databases for Robertson Creek Hatchery chinook. Support for myself was provided by the Natural Sciences and Engineering Research Council of Canada. Xlll GENERAL INTRODUCTION Mortality rates for salmonids from the time they enter the ocean as smolts to the time they return as spawning adults are very high and can account for a large portion of interannual variations in recruitment (Parker 1968). Numerous studies have provided evidence that mortality is particularly high during the first few months following entry into the ocean (Parker 1968; Matthews and Buckley 1976; Bax 1983; Fisher and Pearcy 1988). Using mark-recapture techniques, Parker (1968) quantified daily losses of pink salmon (O. gorbuscha) at 2 to 4% during the first 40 days of ocean residence. Also using mark-recapture, Bax (1983) estimated daily losses of juvenile chum salmon (O. keta) an order of magnitude higher, at 31 to 46% during their first 3 days of ocean residence. While mortality rates are high, they are also found to be highly variable interannually for Pacific salmon (Parker 1968; Bax 1971). Thus, determining whether particular morphological, physiological, or behaviour characteristics of some individuals enhance their probability of survival has valuable application for increasing the understanding and prediction of recruitment variability (Sogard 1997). In general, mortality has been found to vary inversely with body size in marine ecosystems (Peterson and Wroblewski 1984; Lorenzen 1996). Differential mortality rates for individuals may accordingly be explained in part by differences in the size at ocean entry and growth rates achieved during early marine residence. According to size-spectrum theory, larger or fast-growing individuals spend less time vulnerable to a large number of gape-limited predators than their smaller conspecifics (Cushing 1975; Sogard 1997). Large size and high growth rates have also been found to reduce the magnitude of non-predation mortality - suggested recently to be more size-dependent than predation mortality (Lorenzen 1996). Some studies have found correlations between interannual variations in recruitment and variations in average smolt length (Ward et al 1989; Holtby et al 1990; Henderson and Cass 1991). As well, several within-cohort studies have found that individual salmonids with larger size at ocean entry survive at a higher rate during early marine life than their smaller conspecifics (Parker 1971; Healey 1982; Hargreaves and LeBrasseur 1986; Ward et al 1989; l Henderson and Cass 1991). This supports the theory that mortality is size-selective. As well, the assumption that growth rates are important for survival is pervasive in the literature. However, few studies have demonstrated a clear link between marine growth rates and recruitment variability. According to Anderson (1988), the growth-mortality hypothesis provides a rational framework within which to study recruitment variation. The theory is based on bioenergetic principles and ecological theory and provides an advancement to previous hypotheses through the integration of the dynamics of feeding and predation into one testable theory. It predicts that growth is limited by available forage, and that growth rates moderated by both density-dependent and density-independent factors are important for survival. The growth-mortality hypothesis was recently tested over a 17 year period for a natural population of coho (O. kisutch) enterring the ocean on the west Coast of Vancouver Island (Holtby et al. 1990). Early marine growth rates estimated from scales were found to be highly correlated with interannual variations in survival. Holtby et al (1990) also found that the early marine growth rates and survival were positively correlated with indices of productivity in the region of early marine residence. Other studies involving coho have also found correlations between recruitment and indices of ocean productivity during the first marine year (Scamecchia 1981; Nickelson 1986). The investigators suggest that the relationship between ocean productivity and survival results from the influence of growth rates and physiological condition on the magnitude of size-selective mortality. However, for Oregon coho entering the ocean between 1981 and 1984, it was found that although the marine survival rate was correlated with indices of productivity, the growth rates, condition, and stomach fullness were as high in years of low productivity and low survival as years with high productivity and high survival (Fisher and Pearcy 1988). They suggested that increased survival in high productivity years was not directly linked to the effect of ocean productivity on growth, but likely depended on changes in predation intensity independent of size, varying concurrently with ocean conditions (Fisher and Pearcy 1988). Thus available evidence in the California Current system provides seemingly contradicting hypotheses to explain recruitment variations in coho. Further attempts are being made to differentiate between the growth-mortality hypothesis and predation intensity hypotheses. However, it is not likely possible or 2 necessarily desirable to separate out these two mechanisms as they are not inherently mutually exclusive. Such conflicting results, however, provide impetus for altering modes of thinking toward an integration of mechanisms determining recruitment variation. In this thesis, the possible mechanisms determining interannual variation in marine survival of one population of Chinook salmon (O. tshawytscha) are investigated. Specifically, scale analysis is used to determine smolt length and early marine growth rates of age 0.3 Robertson Creek chinook salmon over an eleven year period (1982-1992). Regression of marine survival rate on these measures is then used to determine whether large smolt size and/or rapid early marine growth accrue a survival advantage to those cohorts. The relationships between smolt-to-adult survival and marine growth rates are also investigated in the context of the temporal and spatial variability of the marine environment where the growth occurs. 3 CHAPTER 1. INTERANNUAL VARIABILITY IN FIRST YEAR LENGTH AND GROWTH INDICES: THE MEASUREMENT OF ADULT ROBERTSON CREEK CHINOOK (ONCORHYNCHUS TSHAWYTSCHA) SCALES 1.1 INTRODUCTION The focus of this thesis is to determine whether relationships exist between interannual variability in marine survival rates and the first year lengths and marine growth rates of Robertson Creek Chinook salmon (Oncorhynchus tshawytscha). Although it is difficult to determine the lengths and growth rates of fish accurately in the marine environment, they can be inferred from the analysis of the growth history recorded on hard body structures. Several structures have been used for the determination of length and growth including scales, otoliths, opercular bones, vertebrae, and fin rays or spines (reviewed by Francis 1990). The methods for the use of otoliths (Neilson et al. 1985; Neilson and Geen 1986; West and Larkin 1987) and scales (Clutter and Whitesel 1956; Ward et al. 1989; Fisher and Pearcy 1990; Matthews and Ishida 1989; Holtby et al. 1990; Henderson and Cass 1991; and many others) to infer growth information in salmonids {Oncorhynchus spp.) have been well documented. Scales have been found to provide the most convenient means (Fisher and Pearcy 1990). Strong linear relationships have been found to exist between scale radius and forklength following the time of scale formation in Pacific salmon species (Clutter and Whitesel 1956; Fisher and Pearcy 1990). As well, several studies have shown that the spacing between individual circuli is significantly positively correlated with somatic growth rate (Bilton 1975; Doyle et al. 1987; Fisher and Pearcy 1990; Fukuwaka and Kaeriyama 1997). Thus indices of forklength at age can be determined from the measurement of the scale radius to annular rings and other groups of closely-spaced and irregular ridges (circuli) which form as a result of known changes in the environment (Clutter and Whitesel 1956; Fisher and Pearcy 1990). Indices of growth rate can be determined from the spacing between successive scale circuli (Bilton 1975; Doyle et al. 1987; Fisher and Pearcy 1990; Fukuwaka and Kaeriyama 1997). In this chapter, first year length and growth indices are estimated for Robertson Creek Chinook salmon 4 entering the ocean between 1982 to 1992. An assessment is made to determine whether these features vary significantly between years and whether relationships exist between the length and growth indices. 1.2 METHODS AND MATERIALS 1.2.1 Description of Salmon Scales Salmon scales are flat, calcified structures which lie in pockets within the skin of the fish (Figure 1.1). The anterior portion of the scales have concentric ridges called circuli. The innermost circuli form during fresh water growth, and the outer circuli form in the ocean. The transition from freshwater to saltwater growth is marked by the formation of salt water entry checks (Figure 1.1). Checks are groups of thin, closely spaced ridges that often cannot be differentiated as individual circuli (Clutter and Whitesel 1956). Circuli that are formed during rapid somatic growth are thick and far apart, and those that are formed during slower growth are thinner and closer together. During the winter, when growth is especially slow, broad checks called annuli are formed (Figure 1.1). Fish age is determined from the total number of winter annuli present on the scale (Clutter and Whitesel 1956; Bilton 1982). 1.2.2 Adult Robertson Creek Chinook Scales For this analysis, an archive of Robertson Creek Hatchery Chinook scales stored at the Pacific Biological Station (FOC Science Branch, Nanaimo) was utilized. The Robertson Creek hatchery (RCH) is located on the Somass River near the head of Alberni Inlet on the west coast of Vancouver Island (Figure 1.2). Scales were collected when the Chinook returned to spawn by staff at the hatchery. They were removed from a conventional location on the fish, approximately 2 scales up from the lateral line along an axis extending from the posterior margin of the dorsal fin and the anterior margin of the anal fin (the "preferred area"). Following collection, the fish age corresponding to each scale was determined by staff in the fish ageing laboratory at the Pacific Biological 5 POSTERIOR SCALE MARGIN Figure 1.1. Age 0.3 Robertson Creek chinook scale diagram. Concentric rings are called circuli. The distance between two consecutive circuli is an intercircular distance. Features marked include: 1) scale focus or origin; 2) salt water entry (SWE) check; 3) end of the first marine annulus; 4) 20° ventral axis used for measuring; and 5) uneven scale margin indicating resorption; 6) longest scale axis. Along the 20° ventral axis, the a) freshwater radius from the origin to the SWE check, and b) width of the first marine zone are marked. The radius to the end of the first marine year is equal to a) plus b). 6 Figure 1.2. Chart of Alberni Inlet and Barkley Sound. Robertson Creek hatchery (RCH) Chinook are released into the Somass River and migrate through Alberni Inlet on their way to Barkley Sound. Triangles show the standard location where juvenile Chinook salmon were recaptured during the MASS Program. Inset is the west coast of British Columbia, Canada. (Figure adapted from Hargreaves et al. 1991). 7 Station (FOC Science Branch, Nanaimo, BC). Age designations for RCH chinook ranged from 0.1 to 0.6 (the numbers before and after the decimal place indicating the number of years spent in freshwater and in the ocean, respectively). Only age 0.3 chinook scales were used in the present study. The purpose was to prevent the possible effect of early marine growth rate on spawning age from confusing the identification of interannual growth differences (Peterman 1987; Healey 1991). The scales that were measured were from age 0.3 chinook that spawned between 1985 and 1995. As age 0.3 chinook enter the ocean three years before they return to spawn, the analyses include the ocean entry years between 1982 and 1992. It was not possible to include the 1984 ocean entry year in this study, however, as too few age 0.3 female chinook from that year returned to spawn (or too few scales were collected) to obtain a significant sample size. The minimum sample size for each year was 45 for a total sample size of 572 scales. Only the scales of female chinook salmon were measured for this analysis. This was to prevent possible sexual dimorphism in early marine growth rates and age at return from confounding interannual differences in length and growth measurements. For some years, however, the scales of both males and females were measured as the sex was not indicated in the scale samples. The sex data for these sampling years was later located allowing for blind tests for sexual dimorphism in the scale features of age 0.3 Robertson Creek chinook salmon to be performed. The results, which are presented and discussed in Appendix I, confirmed the presence of sexual dimorphism in some of the scale features in some years. Thus, the male samples were removed from the database. Spawners with high or low early marine growth rates may return predominantly early or late in the spawning season (Hilborn and Walters 1992). Therefore, a representative proportion of scales collected throughout the spawning period were measured. During each spawning year, chinook returning to the hatchery were held in ponds until they were spawned. Each spawning day a proportion of the chinook were sampled and their scales were entered into scale books (5 fish per book). Chinook returning to the 8 hatchery at earlier (later) dates were sampled generally earlier (later) in the spawning season (generally October to mid-November). The total number of age 0.3 Chinook sampled in each spawning year was determined, and the proportion of scales measured from each date was then weighted by the number sampled on that date. Thus, measurements were obtained from only one age class of female Chinook, and only represent survivors returning at that age. These Chinook are to be used as an index of the growth conditions experienced by the entire cohort during the first year of marine life. However, the 0.3 age class is the first group of females to return to the hatchery, as few return after only 2 years in the ocean (age 0.2). If age at return is affected by first year growth rates (as has been found for the relationship between early marine growth and jack incidence), then the age 0.3 class of females may represent a disproportionate portion of the faster growing females of the cohorts. Neilson and Geen (1986) found a negative correlation between size at the end of the first marine year and age at return. However, age at return has a genetic component, and the size at the end of the 3rd marine year is likely influenced more by recent growth than by growth experienced during early marine life. A study on Kamchatka River stream-type Chinook found that the final length was most highly correlated with the length added in the final ocean year. This suggests that growth conditions in the final year at sea were important in determining the size of each age class (reviewed by Healey 1991). However, size biased mortality or variation in the proportion of ages at return may result in the scale measurements not being representative of the true growth of the cohorts during early life. Disproportionately high mortality rates for certain size classes of fish can result in directional selection, biasing the scale measurements. However, it has been shown that a very high intensity of size-selective mortality is required before the effects become evident in the survivors (Sogard 1997). As well, Lee's Phenomenon which has been found in some studies, is characterized by back-calculated lengths at age being smaller the older the fish that the measurements are taken from (Ricker 1992). It is suggested to result from the higher mortality rates for the larger fish of a cohort due to early maturation and senility. The pooling of male and 9 female measurements has also been invoked, due to sexual differences in growth rate and natural mortality rate (Ricker 1975). However, Lee's Phenomenon should not influence this study since only the scales from only female chinook spawning at roughly the first spawning age were used for this analysis. 1.2.3 Scale Measurement Procedure A projection microscope in combination with a digitizer was used to measure the adult chinook scales. An acetate impression of a scale was placed on the stage of a Neopromar microscope which projected the scale image onto a digitizing tablet at 100x magnification. The digitizer was used to first mark the origin, and then mark each subsequent scale feature. The digitized data was received by a FORTRAN program which provided output on the number of scale zones measured, the width of each zone, the incremental distance between circuli digitized within each zone, and the number of circuli in each zone. The final units of the output data were in 10"2 mm. The scale radius to the end of the freshwater zone, the radius to the end of each of the marine annuli, as well as the distance between each circulus within the freshwater (FW) and first marine (MAR1) zones were digitized for this study (Figure 1.1). The end of the freshwater zone was determined by the location of a salt water entry (SWE) check consisting of 2-4 thin and irregular circuli. The last circulus that was measured within the freshwater zone was the circulus before the last circulus within the SWE check (Figure 1.1). This was consistent with the belief that the SWE check is formed during the transition from fresh to salt water (Darlene Gillespie pers. comm., FOC Canada, Pacific Biological Station, Nanaimo, BC). The end of each marine annulus (winter check) was measured at the beginning of the last narrow circulus that precedes the first wide intercircular space following the winter check (Bilton and Ludwig 1966). All measurements were made along the radial line that extends ventro-anteriorly from the scale focus at a 20° angle to the longitudinal axis (Figure 1.1). This line, approximately perpendicular to the anterior edge of the posterior portion of the scale that is lacking circuli, has proven to provide the most consistent and representative results for circuli 10 counts and radii measures (Clutter and Whitesel 1956). In practice this axis was disclosed by overlaying an acetate guide on the projected scale image. The strict criteria that were followed to determine the end of the salt water entry check ensured a high degree of precision in the data presented. This is important as small changes in the criteria were found to have a relatively large effect on the measurements. After the majority of the scales were digitized, the data was graphed by the date that the measurements were taken. It became apparent that the determination of the end of the SWE check had changed over the study period as a result of increased experience. This produced appreciable changes in the scale radius to the end of the SWE check, the first intercircular distance, and the intercircular distances for the very early marine growth period. An investigation indicated that all measurements made from February 5, 1998 forward were consistent. By that time experience had been gained in the identification of the SWE check and precise criteria for digitizing had been clearly established. All scales digitized prior to February 5, 1998 (approximately 200) were remeasured. A precision test was also conducted to quantify the repeatability of the measurements to the end of the SWE check. 50 scales from the 1988 and 1990 ocean entry years were digitized twice. Precision was assessed by similarity in the number of circuli digitized to the end of the SWE check. 56% of the scales were found to have the same circulus count to the end of the SWE check. 32% of the scales were found to be different by one circulus count, and 12% were found to differ by two circuli. No scales were found to differ by three or more circuli. Since intercircular distances in the region of the check aregenerally between 1.3 and 2.3 X 10"2 mm, the increase (reduction) in the back-calculated forklength estimated for ocean entry resulting from including one more (one less) circulus before the end of the SWE check would be approximately 2.7 to 4.7 mm (Chapter 2). Thus inconsistently assessing the precise circulus for the end of the SWE check by 1 circulus changes the estimated smolt size by approximately 2.5-5%. Over- (under-) estimating the number of circuli by 2 results in an overestimation (underestimation) of smolt size by 5-10%. However, equal numbers of scales were found to be over- and under-estimated in the second count. Since at least 45 scales were measured for each ocean n entry year, the over- and under- estimated scales should negate each other and produce no significant effect on the average scale radius. Furthermore, different circulus counts within the freshwater zone can also arise from using a slightly different axis for measurement. Since circuli within the check are irregular and often broken or branched, the circulus count could differ within the FW zone, yet have no influence on radius to the end of the SWE check. Determining the accuracy of the correspondence of the end of scale zones with shifts in the life history stage of the Chinook was beyond the scope of this study. However, in Chapters 2 and 4 the time of the year and the location of Chinook when the SWE check is completed is discussed in detail. The conclusions support the belief that the SWE check forms as a result (direct or indirect) of the transition from predominantly freshwater and estuarine habitat to predominantly marine habitat. As well, studies have shown that the winter annulus, which marks the end of the first marine year scale growth, may be completed in January, during a period of decreasing water temperature (Barber and Walker 1988). It is suggested that the annulus is completed in response to increasing photoperiod as well as food availability (Barber and Walker 1988). Since the amount of light and food available at a given time of the year may differ regionally as well as latitudinally, there may be variations in the absolute date that the winter annulus is completed between groups of Chinook migrating varying distances to the north in their first marine year. Therefore, variability around the mean time of formation of the first marine annulus is expected. Between years, changes may also occur in the time for formation of the winter annulus. However, it appears that the scale region of most interest in this study, the SWE check which forms as a result of the transition to the marine environment, is a good determinant of the size after final ocean entry (smolt size). Variability will also likely occur in the marine growth rates estimated from the scales. However, as greater than 45 scales for each ocean entry year were measured, the mean which is used for back-calculation, should provide a reasonable assessment of the central tendency of these measures for a given year. 12 1.2.4 Scale Measurement Difficulties To ensure consistent results, roughly 10% of the scales eligible according to age and sex were not measured. Scale measuring difficulties were associated with uncharacteristic morphology and improper mounting. Those scales with visibly distorted shapes were not measured as their odd shape distorts intercircular distance measurements and may indicate that the scale was removed from a location on the fish far from the preferred sampling area (Clutter and Whitesel 1956). Improper mounting resulted from scale books becoming wet during sampling, and the glue covering the face of the scale, obscuring fine details. The number of scales in the archive that were unreadable decreased through the 11 year period as scale sampling procedures were improved. For some scales that were measured, the exact location of the first marine annulus (Figure 1.1) was not determined with confidence. The two criteria that are typically used to determine the end of the annulus: the point that the circuli become thicker and more widely-spaced; and the point where circuli stop "pinching inwards" at the posterior scale margin (Darlene Gillespie pers. comm., FOC Canada, Pacific Biological Station, Nanaimo, BC) were not concordant for some scales. This was especially true for many scales from the 1983, 1987, 1988, 1991, and 1992 ocean entry years. Therefore, the measurements of the radius to the end of the first marine annulus may not be accurate indications of the length at the end of winter growth for some years. The SWE check, however, was identified in 100% of the scales and thus the FW radius and marine intercircular distances were measured with a high degree of accuracy and precision. 1.2.5. Regression Analysis The relationships between the survival and length or growth rate measures were determined by arithmetic mean regression. Significance of an individual regression performed within multiple simultaneous regression tests was assessed using the 13 sequential Bonferroni technique (Rice 1989; Sokal and Rohlf 1995). When conducting multiple, simultaneous tests, the experiment-wise (or group-wide) error rate (the probability of making at least one type I error) is equal to 1-(1-cc)k, where k is the number of tests that are conducted (Sokal and Rohlf 1995). According to this formula, the probability of declaring at least one test incorrectly significant (making a type I error) at P = 0.05 when 10 simultaneous tests are conducted is 40%. Thus, when conducting multiple tests, the significance level for each test must be decreased so that the probability of making a type I error in any of the k tests does not exceed a (Sokal and Rohlf 1995). The nonparametric sequential Bonferroni technique effectively controls for type I errors while simultaneously maintaining substantial power in detecting false null hypotheses. It has been found to be valid in virtually all applications and does not require independence of the component tests (Rice 1989). The procedure used is to determine the P values for the test statistics of each component test, and then to order them from the smallest (Pi) to largest (Pk). The formula Pj < a (1 + k - i) _1 , with the chosen significance level (a) and the total number of simultaneous tests conducted (k) is then used to determine whether each individual test P value (Pj) is less than the adjusted, or "table-wide" a that would result in significance at the chosen P level ( here P = 0.05). For the first test it is determined whether the smallest P value (Pi) is less than or equal to ock "1. If P is not less than this new, "table-wide" a level, then all tests are declared non-significant. If the inequality is met, then the first test is declared significant and the next test is performed: P2 < a (k -1) ~1. The test is continued iteratively until the inequality is not met. At that point, the current test, and all subsequent test are declared non-significant at the chosen P level (0.05) (Rice 1989). 14 1.3 RESULTS 1.3.1 Interannual Variation in Scale Measurements Scale Radius and Circuli Number at Ocean Entry The scale radius to the end of the SWE check (FW radius: Figure 1.1) was determined for the adult chinook scales. As entry into the marine environment is assumed to occur during check formation, the radius is an index of length at final ocean entry (smolt size). The average and standard deviation of the freshwater radius between 1982 and 1992 is presented in Figure 1.3. Regression analysis indicates that there was a significant linear increase in the average FW radius during this period (r2 = 0.57; P = 0.01). As the data were normally distributed, two tail t-tests assuming unequal variances were performed on each pair of years to determine which years differed significantly from each other. 16 of the 45 comparisons were found to be statistically significant at the a level adjusted for multiple tests using the sequential Bonferroni method (Table 1.1). The detailed descriptive statistics for the FW radius are shown in Table 111.1 (Appendix III). The number of circuli present on the scales before the end of the SWE check averaged 13.6 ± 1.3 (Table 1.2) and individual counts ranged from 10 to 19 over the eleven year study period. The averages and standard deviations are presented in Figure 1.4. The average number of circuli present before the end of the SWE check increased significantly between 1982 and 1992 (r2 = 0.65; P = 0.005). The detailed descriptive statistics can be found in Table III.2 (Appendix III). 15 (0 Q < CO UJ _ i < o 48 46 44 42 40 38 36 34 32 I ~A Y T 1 Mean+SD Mean-SD —D— M e a n A 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.3. The average and standard deviation of the scale radius to the end of the salt water entry check (SWE) (10"2 mm) for 1982-83; 85-92. 16.5 10.5 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.4. The average and standard deviation of the number of scale circuli to the end of the salt water entry check for 1982-83; 85-92. 16 Table 1.1. Scale Radius to the end of the SWE check: P Two Tail from Two Sample T-test Assuming Unequal Variances. Nonsignificant differences are underlined (P: 0.05) Year 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 1982 9.91 E-03 2.63E-03 9.47E-07 1.37E-11 1.86E-08 1.33E-05 1.57E-04 4.43E-07 9.77E-13 1983 9.10E-01 2.27E-02 2.02E-05 2.64E-03 7.05E-02 3.42E-01 1.31E-02 2.19E-06 1985 2.71 E-03 1.28E-09 5.73E-05 1.96E-02 1.74E-01 1.31 E-03 7.86E-11 1986 2.72E-02 4.47E-01 6.83E-01 1.14E-01 7.75E-01 3.98E-03 1987 1.34E-01 1.34E-02 4.77E-05 6.96E-02 3.20E-01 1988 2.57E-01 1.47E-02 6.61 E-01 2.36E-02 1989 2.88E-01 5.05E-01 2.07E-03 1990 6.77E-02 3.65E-06 1991 1.26E-02 1992 Table 1.2. The mean, minimum, and maximum average number of circuli to the end of the SWE check, the first marine annulus, and between the SWE check and the first marine annulus as well as standard deviations (SD) for 10 years (1982-83; 1985-92). Scale Zone 1.SWE Check 2. SWE Check to End Mar 1 3. End Marl Mean (All yrs.) 13.6 30.8 44.4 SD 1.3 2.8 3.8 Min.Av. 12.7 29.6 43.3 SD 1.2 2.3 2.5 MaxAv. 14.8 32.4 45.7 SD 1.2 2.5 3.3 Interannual Variation in Scale Radius and Circuli Number for the First Year The average and standard deviation of the width of the scale from the end of the SWE check to the end of the first marine annulus (Figure 1.1) for 1982-83; 1985-92 is presented in Figure 1.5. No significant linear trend was found between 1982 and 1992 (r2 = 0.02; P = 0.70). The descriptive statistics can be found in Table III.5 (Appendix III). The total scale radius at the end of the first marine year (Figure 1.1) was determined by adding the width of the first marine zone to the FW radius. Since the first marine zone ends following the first winter annulus (Figure 1.1), this measurement is an index of the length of Chinook at the end of their first winter in the ocean. A plot of the means and standard deviations is presented in Figure 1.6. A significant linear relationship was not found over the study period (r2 = 0.23; P = 0.16). Using pair-wise, two tail t-tests assuming unequal variances, only 7 of the 45 comparisons differed significantly at the P = 0.05 significance level (Table 1.3: sequential Bonferroni technique). The descriptive statistics can be found in Table III.3 (Appendix III). The number of circuli counted from the end of freshwater growth (SWE check) to the end of the first marine year (winter annulus) averaged 30.78 ± 2.79 (Table 1.2) and individual counts ranged from 22 to 40 over the 10 years sampled (Appendix III: Table III.6). The total number of circuli from the scale origin to the end of the first marine year (winter annulus) averaged 44.4 + 3.8 (Table 1.2) and individual counts ranged from 35 18 Ill z o N 111 _l < o </) Ul z H u. O X 1-o «s 1 1 5 1 1 0 1 0 5 1 0 0 9 5 9 0 8 5 8 0 7 5 * f A Y_ I Mean+SD Mean-SD —O— Mean ] I > - - • • i k \ - / y < 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.5. The average and standard deviation of the width of the first marine scale zone (10"2 mm for 1982-83; 85-92. 160 150 (0 2 140 Q < 111 ^ 130 o V) 120 110 I "2 { T Mean+SD Mean-SD —D— Mean i 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.6. The average and standard deviation of the scale radius to the end of the first marine annulus (10"2 mm) for 1982-83; 85-92. 19 Table 1.3. Scale radius to the end of the first marine year: P Two Tail from Two Sample T-test Assuming Unequal Variances. Nonsignificant differences are underlined (P: 0.05). Year 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 1982 5.89E-04 2.54E-02 4.20E-03 1.67E-01 1.86E-04 5.57E-04 1.30E-01 5.15E-04 4.86E-05 1983 6.60E-02 4.64E-01 3.86E-03 5.44E-01 8.40E-01 3.15E-02 8.42E-01 7.06E-01 1985 2.88E-01 2.05E-01 2.06E-02 5.34E-02 5.39E-01 7.37E-02 1.06E-02 1986 3.15E-02 1.98E-01 3.71 E-01 1.40E-01 5.60E-01 2.30E-01 1987 1.20E-03 3.81 E-03 6.84E-01 2.91 E-03 1.21 E-04 1988 6.96E-01 1.01E-02 4.10E-01 7.49E-01 1989 2.58E-02 6.88E-01 8.96E-01 1990 3.48E-02 5.81 E-03 1991 5.26E-01 to 55 (Appendix III: Table 111.4). Plots of the averages and standard deviations of circulus counts for the first marine zone and to the end of the first winter annulus are presented in Figures 1.7 and 1.8. Significant linear time-dependent trends were not found for the average number of circuli in these zones (r2 = 0.15; P = 0.26 and r2 = 0.04; P = 0.58, respectively). Interannual Variation in Early Marine Growth Indices The spacing between individual scale circuli is significantly positively correlated with scale growth rate and somatic growth rate (Bilton 1975; Doyle et al. 1987; Fisher and Pearcy 1990; Fukuwaka and Kaeriyama 1997). Thus intercircular distances are a measure of the growth rate of the scale as well as an index of somatic growth. Figure 1.9 presents the intercircular distances for the first half of the first marine year between 1982 and 1992. These data indicate that the intercircular distances increase rapidly from the 1st to the 4th marine circulus suggesting a rapid increase in the scale (somatic) growth rate following ocean entry. It can be seen that interannual variation in the scale growth rate is high, particularly between the 4th and 9th marine circuli. Figure 1.10 presents the average intercircular distance for the first 28 marine circuli for selected years (1983, 1985, 1988, 1989, and 1990). It can be seen that interannual variation in growth rates decreases in the latter half of the first marine year. As well it is evident that cohorts that grew relatively rapidly by the 10th marine circulus (1988, 1989, and 1990) exhibited smaller growth increments by roughly the 20th marine circulus. Those that grew relatively slowly early on (1983 and 1985) had the highest growth increments near the end of the year (approximately marine circuli 20-26). Regression of a later growth index (average intercircular distance for the 21s t to 25th marine circuli) on an index of early marine growth (average intercircular distance for the 4th to 9th marine circuli) for all years (1982-83; 1985-92) was not found to be significant (Y = -0.237x + 3.72, r2 = 0.086; P = 0.414). However, line graphs of the two growth measures (Figure 1.11) suggests an inverse relationship between early and later marine growth indices for several years. Years in which the scale growth was moderate during early marine life 21 DC UJ ffl 2 3 Z ZJ 3 O o 3 6 3 4 3 2 3 0 2 8 2 6 Mean+SD Mean-SD Mean 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.7. The average and standard deviation of the number of circuli in the first marine zone for 1982-83; 85-92. 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.8. The average and standard deviation of the number of scale circuli present before the end of the first marine annulus for 1982-83; 85-92. 22 5 6 7 8 9 10 11 12 13 14 15 Marine Circulus Number Figure 1.9. Intercircular distances (10 mm) for the first 15 marine scale circuli. 3.7 3.5 3.3 £ E ,— o B-CD o £Z 1 Q *-co 3 o .!= o 1 — I 3.1 2.9 2.7 2.5 2.3 2.1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 Marine Circulus Number Figure 1.10. Average intercircular distance (10"2 mm) for the first 28 marine circuli for ocean entry years 1983, 85, 88, 89 and 90. 23 3.6 3.5 -F E <&> T— * • — * * CD (3 c: ffl to Q 1 _ CD 3 C3 O L . CD +-* _C 3.4 3.3 3.2 3.1 3.0 ?.9 2.8 2.7 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Ocean Entry Year Figure 1.11. An early marine growth index (Intercircular distance averaged over the 4th to 9th marine circuli) and an index of growth later in the first marine year (intercircular distance averaged over the 21s t to 25th marine circuli) for 1982-83; 85-92. 24 (between roughly 2.9 and 3.1 X 10'2 mm) exhibited the smallest change in growth by the end of the first marine year (Figure 1.11). The intercircular distances for the 2nd to 13th marine circuli were tested for autocorrelation. The measures for 1983 and 1985 were found to have significant autocorrelation at a lag of 1 ciruclus (r = 0.6; P < 0.05). Three other years (1986, 1989, and 1991) had nearly significant autocorrelation at lag 1 (r = 0.5 to 0.55: P > 0.05), and six did not. Figure 1.12 presents the autocorrelation plots for a year with significant autocorrelation at lag 1 (1985) and for a year with no significant autocorrelation (1988). To account for the autocorrelation and to smooth to any short term variability in the scale growth data, the intercircular distances used in this study have been averaged over three consecutive marine circuli. The notation, Av3-5, Av4-6, etc. indicates which marine circuli (i.e. those added after the end of the SWE check) are included in the average growth index. The average and standard deviations for 1982-83; 1985-92 of the intercircular distance averaged over the 3rd and 5th (Av3-5), 8th and 10th (Av8-10), 11th and 13th (Av11-13), and 13th and 15th (Av13-15) marine circuli are presented in Figures 1.13, 1.14, 1.15, and 1.16. These growth indices were chosen to illustrate that the between year differences in the growth indices is lower for later in the first marine year (AV11-13 and Av13-15) than for earlier in the first marine year (Av3-5 and Av8-10). Thus, while investigating interannual variation in the growth rates, focus will be made on the early marine period. Two tail t-tests assuming unequal variances were performed on the very early marine growth index, Av3-5, for each pair of years. It was found that 14 of the 45 comparisons were statistically significant at the P two-tail of 0.05 (Table 1.4: sequential Bonferroni method). For the early marine growth indices (Av3-5: Figure 1.13 and Av8-10: Figure 1.14) the highest growth occurred between 1988 and 1990, and the lowest between 1982 and 1985. The detailed descriptive statistics for the Av3-5 and Av8-10 can be found in Tables III.7 and III.8 (Appendix III). 25 Autocorrelat ion Plot c o 05 <D i _ t— O O c o <D I— i_ o O 1.0 Lag Autocorrelation Plot Lag Figure 1.12. Autocorrelation plot of the 2nd to 13th marine circuli up to a lag of 13 circuli for 1985 (top) and 1988 (bottom). 1985 has a significant autocorrelation at lag 1 as indicated by the overlap of the 95% confidence line for significance with the Pearson correlation coefficient (bars). 1988 does not have significant autocorrelation. 26 UJ o z < w Q DC < _J r» o cc o cc UJ o . o 3 .4 3 .0 2 .6 o o I J— ~~T" Mean+SD Mean-SD —CJ— M e a n 3 T" "" "J~ T 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.13. The mean and standard deviation of the intercircular distance (10"2 mm) averaged on the 3rd to 5th marine circuli (Av3-5) for 1982-3; 85-92. UJ O z < </) Q CC < -J O DC O CC UJ h -z 3.8 3.4 3.0 2.6 o o L I Mean+SD Mean-SD —O— Mean b C p " • • " ^ ^ ^ • ' t P ^ 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.14. The mean and standard deviation of the intercircular distance (10"2 mm) averaged on the 8th to 10th marine circuli (Av8-10) for 1982-83; 85-92. 27 UJ o 4.2 3.8 r^  3.4 3.0 W Q DC < _l D O cc o cc HI Z 2.6 2.2 1 "T~ Mean+SD Mean-SD —Q— Mean - ^ y* 3 r — 1 J 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.15. The mean and standard deviation of the intercircular distance (10"2 mm) averaged over the 11th to 13th marine circuli (Av11-13) for 1982-83; 85-92. UJ o W Q CC < _l Z3 O CC o cc Hi ^-4.2 3.8 3.4 3.0 2.6 2.2 " [ / / 1 ~"T~ Mean+SD Mean-SD —D— Mean . 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 1.16. The mean and standard deviation of the intercircular distance (10"2 mm) averaged over the 13th to 15th marine circuli (Av13-15) for 1982-83; 85-92. 28 Table 1.4. Intercircular distance averaged over the 3rd to 5th marine circuli: P Two Tail from Two Sample T-test Assuming Unequal Variances. Nonsignificant differences are underlined (P: 0.05) Year 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 1982 8.68E-01 2.61 E-01 2.34E-01 2.53E-02 2.91 E-04 6.61 E-08 1.36E-03 7.76E-02 1.50E-03 7583 3.18E-01 2.80E-01 3.13E-02 3.40E-04 4.99E-08 1.63E-03 9.61 E-02 1.76E-03 1985 7.88E-01 1.50E-01 1.41E-03 8.12E-09 7.37E-03 3.99E-01 7.54E-03 1986 3.86E-01 1.27E-02 1.65E-05 4.21 E-02 6.79E-01 5.95E-02 1987 3.67E-02 4.54E-06 1.26E-01 5.89E-01 1.85E-01 1988 9.64E-02 6.06E-01 1.39E-02 3.38E-01 1989 2.24E-02 1.24E-06 1.61 E-03 1990 5.37E-02 6.97E-01 1991 7.41 E-02 1992 1.3.2 Relationship Between Scale Measures To assess whether relationships exist between the length and growth indices as well as the number of circuli deposited in particular scale regions, linear regressions were performed. The results are presented in Table 1.5. Table 1.5. Equations for the linear regressions of key scale features on other scale features. Regressions include 44 samples from each of ten years (1982-83; 1985-1992) for a pooled sample size of 440. The "table-wide" a levels that the test P values must not exceed for signif icance at P = 0.05 using sequential Bonferroni tests are included. Significant tests are marked with an asterisk (*). Y variable X variable Equation J2 P value Table-Wide a 1. 2. 3. 4. 5. Av3-5 MAR 1 WID MAR1WID FWCIRC MAR1 FWRAD FWRAD Av3-5 FWRAD MAR 1 WID Y = 0.045(x) +1.192 Y = 0.531 (x) + 72.781 Y = 7.529(x) +71.686 Y = 0.173(x) +6.542 Y = 0.175(x) + 14.513 0.132 0.0352 0.108 0.335 0.548 3.89E-15* 7.47E-05* 1.54E-12* 1E-40* 1.7E-77* 0.0167 0.05 0.025 0.0125 0.01 CIRC A significant relationship (P < 0.05) was found for all five regressions (Table 1.5). Only a small proportion of the variability in the early marine growth index (AV3-5) was accounted for by the regression on scale radius at the end of the SWE check (FW RAD) (Regression 1: r2 = 0.13; P < 0.017). As well, weak positive relationships were found for the regressions of the width of the first marine zone (MAR1 WID) on the scale radius at the end of the SWE check (FW RAD) (Regression 2: r2 = 0.035; P <0.05) and on the early marine growth index (Av3-5) (Regression 3: r2 = 0.11; P < 0.025). A moderately strong relationship was found between the number of scale circuli to the end of the SWE check (FW CIRC) and the scale radius at the end of the SWE check (FW RAD) 30 (Regression 4: r2 = 0.34; P < 0.0125). A large portion of the variability in the number of scale circuli added from the end of the SWE check to the end of the first marine annulus (MAR1 CIRC) was accounted for by linear regression on the scale width added in the first marine zone (MAR1 WID) (Table 1.5, Regression 5: r2 = 0.55; P < 0.01). 1.4 DISCUSSION Interannual variation was found in the index of smolt size at ocean entry (SWE Radius) as 16 of the 45 year-by-year comparisons were found to be significant (Table 1.1: t-tests assuming unequal variances: sequential Bonferroni P two-tail < 0.05). A significant increase in the average index of smolt size from 1982 to 1992 (r2 = 0.57; P = 0.01) was found. This may result partially from an increase in the average weight at hatchery release - a weak relationship was found between the weight of Chinook at hatchery release and smolt size, although it was not significant (r2 = 0.27; P = 0.123). This may have resulted from changes in the temperature of the water at the hatchery accelerating growth. The number of circuli before the end of the SWE check was also found to increase significantly between 1982 and 1992 (r2 = 0.65; P = 0.005). The regularity of circulus deposition, suggests that Chinook in more recent years may have made the transition to the marine environment at a later age. However, as will be discussed in Chapter 2, the rate of deposition of scale circuli can vary under different conditions of food and light. Improvement in the feeding regime for Chinook at the hatchery may have accelerated the rate of deposition slightly in more recent years. Only a small amount of the variability in the early marine growth index may be explained by the index of smolt size (FW RAD) (Table 1.5, Regression 1: r2 = 0.13; P < 0.017). The length at ocean entry may be partially dependent on the size at release, as well as the growth conditions experienced during estuarine residence. After entry into the marine environment during SWE check formation, the growth rates are likely influenced more by the interannual variation in marine productivity. Early marine growth, as indexed by the average intercircular distance for the 3rd to 5th marine circuli, differed significantly in 14 of 45 year by year comparisons (Table 1.4: t-test assuming 31 unequal variances: sequential Bonferroni P two-tail < 0.05). If the between-year differences in early marine growth indices continued throughout the first marine year, then years with high growth rates would be expected to have greater length indices at the end of the year. However this did not occur. A very weak relationship was found between the length added during the first marine year and the early marine growth index (Table 1.5, Regression 3: r2 = 0.108; P < 0.025). Furthermore, Figure 1.11 indicates that some cohorts that grew relatively rapidly (slowly) between the 4th and 9th marine circuli exhibited smaller (larger) growth increments between the 21s t and 25th marine circuli. Generally it is found that variability in individual fish length within a cohort increases during early life (growth depensation). However, due to resulting shifts in position on the growth curve, growth compensation results from smaller fish having larger growth increments and thus catching up to the larger ones. This effectively narrows the size variation in a cohort towards the end of the first year, or into the second year (Ricker 1975). Growth compensation can also result from the plasticity in fish growth rates. Studies have shown that growth rates are usually sub-maximal, and can be increased when an individual has fallen below its expected growth trajectory (Nicieza and Metcalfe 1997). Fish subjected to periods of growth restriction or starvation have an amazing ability to increase their growth rates. The mechanism is suggested to result from increased intake rates. However, intake rates are often increased at the expense of increased predation risk. Atlantic salmon increased their growth rates competitively through increased aggressive interactions making them more conspicuous and less vigilant (Nicieza and Metcalfe 1997). Individuals increasing their growth rates in this way would likely suffer greater mortality decreasing the benefit of larger size achieved by the end of the first year. Furthermore, evidence suggests that fish under reduced ration compromise growth in weight over growth in length (skeletal growth) (Nicieza and Metcalfe 1997). Therefore, the growth in length data investigated in this thesis do not indicate the full extent of the growth reduction that may occur under poor feeding conditions. Those cohorts with slow growth in length may be further compromised by poor physiological condition and decreased energy reserves increasing their susceptibility to non-predation and predation mortality. Regardless of the mechanism of growth compensation, the narrowing of the length variation during 32 the latter half of the first marine year between cohorts suggests that the relationship between marine survival and length may only be evident before the end of the first year of life for Robertson Creek hatchery chinook. Therefore focus will be made on early marine lengths and growth rates for correlation studies with marine survival. 33 CHAPTER 2. BACK-CALCULATION OF FORKLENGTHS AND DETERMINATION OF MARINE GROWTH RATES FROM ADULT CHINOOK SCALE MEASUREMENTS 2.1 INTRODUCTION For Pacific salmon (Oncorhynchus spp.) strong relationships have been found between body length and the radius of their scales after scale formation (Clutter and Whitesel 1956; Fisher and Pearcy 1990). Therefore it is possible to determine the body length at an earlier age from the scales of older fish. Back-calculation of length-at-age has a long history - work beginning as early as 1909 (reviewed by Francis 1990). Several statistical methods have been developed over the years to decrease possible back-calculation errors based on bias that has been found (reviewed by Francis 1990). The recently recommended proportional methods require the final length and scale radius of the fish for their calculation. Because it was not possible to determine the final scale radii for spawning Robertson Creek Chinook salmon due to resorption at the scale margin (Figure 1.1), these methods were not possible for the current study. In such cases, the geometric mean regression, which provides a symmetrical axis based on the ratio of the dispersion of length on the dispersion of scale radius is preferred over ordinary regression for back-calculation (Ricker 1992). In this chapter a regression of forklength on scale radius is derived using juvenile Chinook salmon captured at various times after release from the Robertson Creek hatchery. The resulting equation is then used to back-calculate first year lengths from adult scale radii, as well as somatic growth increments from the distance between consecutive circuli. The approximate rate of circulus deposition is estimated for the conversion of back-calculated length increments into somatic growth rates in mmd"1. Converting the scale measurements into measures of body length and growth rate enables interannual variations to be assessed in a life history context. The approximate time of the year that the first year lengths and growth rates are achieved is also presented for further investigations. 34 2.2 METHODS AND MATERIALS 2.2.1 Regression of Forklength on Scale Radius for Back-Calculation Data Source As a component of the Marine Survival of Salmon (MASS) Program, juvenile salmon were captured by purse seine in Alberni Inlet and Barkley Sound (Hargreaves et al. 1991). During each year of the program (1987-1992) a large number of young-of-the-year chinook salmon were captured at sampling sites indicated in Figure 1.2 (Chapter 1). Of the chinook that were captured, some possessed fin clips indicating coded wire tagging (CWT). The tags of the marked fish were analyzed, and the Chinook's rearing history was determined from the code. The information included the hatchery of origin, the release site, and the time period during which chinook from each tag group were released. This information, as well as that obtained at capture (date of capture, location of capture, length, and scale number) was entered into a database kindly provided by Dr. Brent Hargreaves (FOC, Pacific Biological Station, Nanaimo). Scales taken from juvenile Robertson Creek hatchery (RCH) chinook captured in 1989 and 1990 were digitized by staff in the ageing laboratory at the Pacific Biological Station (FOC, Science Branch, Nanaimo). The scales were measured using the longest scale axis. Radius measurements had to be converted to the corresponding 20° ventral axis value to be consistent with the adults scale measurements (for conversion see Appendix II). Once converted, the scale data (the total number of circuli and total scale radius) was added to the appropriate individual in the database by matching up the scale and book numbers. Thus, the final database used for this analysis included the release and capture information as well as the scale measurements for each marked juvenile RCH chinook caught in 1989 and 1990. 35 Regression Equation The juvenile Chinook database was used to construct a regression of forklength on scale radius for the back-calculation of length at age. Ordinary, arithmetic mean (AM) regression is not appropriate for this study as it requires that the independent variable not be subject to natural variability or be measured with error. Geometric Mean (GM) regression is the appropriate method as the dimensions of the forklength and scale radius data differ (Ricker 1975; Sokal and Rohlf 1995). Both the ordinary (AM) and geometric mean (GM) regressions were estimated for comparison. A linear relationship was found between the forklengths (FL) and scale radii (SR) of juvenile Chinook caught in 1989 and 1990 (Figure 2.1). The arithmetic mean (AM) regression (Figure 2.1) equation is presented below. FL = 1.6695 (SR) + 32.726 (r2 = 0.67; P = 7.02 E-76). For geometric mean (GM) regression, the regression coefficient (v) was found by taking the square root of the ratio of the sum of the squared deviations from the mean of Y and X (or the ratio of the standard deviations of Y and X). The equation for the linear regression using GM regression (Figure 2.2) was found to be: FL = 2.042 (SR) + 18.536 (r2 = 0.67) For the AM regression, the null hypothesis, parametric slope (B) = 0 was tested by calculating the F statistic at the one tail significance level of 0.05. The slope was found to be significantly different from 0 (P = 7.02 E-76). This test is not appropriate for the GM regression, as the slope (v) cannot be 0 or undefined unless the standard deviation of Y or X is equal to 0 (Sokal and Rohlf 1995). However, the GM regression slope (v) is always steeper than the AM regression slope (b). This is because the GM slope (v) is 36 I 40 o 20 LL 0 0 FL= 1.6605 (SR)+32.726 r squared = 0.67 20 40 Scale Fladius (10"2 mm) 60 80 Figure 2.1. Arithmetic mean (AM) regression of forklength on scale radius. 180 160 o LL 40 20 0 0 FL = 2.042 (SR)+18.536 r squared = 0.67 20 40 Scale Radius (10~2 mm) 60 80 Figure 2.2. Geometric mean (GM) regression of forklength on scale radius. 37 defined as: v = (b)(r"1) (Ricker 1973; 1975), where b is the AM slope and r, the correlation coefficient is always less than unity. Since the slope for the AM regression was found to be significantly different from 0 at P = 0.05, then the GM regression slope is also significant. The 95% confidence intervals for the GM slope (v) calculated from the standard error of the regression slope and the t statistic at the 0.05 significance level (Ricker 1973; Sokal and Rohlf 1995) were estimated as: L1 = 1.907 and L2 = 2.176. Theoretically, the Y-intercept of the regressions describes the forklength at which the scale begins to form, assuming that the same relationship exists for Chinook at a size smaller than those included in the construction of the regression. The AM regression resulted in a Y-intercept (and, thus theoretical length at which the SR is 0) of 33 mm and the GM regression in a Y-intercept of 19 mm. Chinook fry collected at the RCH in 1998 with an average forklength of 36 mm were found to have a significant portion of their scales formed. As the scale radius increases linearly with forklength following scale formation (Clutter and Whitesel 1956; Bilton 1982), the AM regression Y intercept of 33 mm may be too high: the scale radius that was observed on 36 mm fish would have had to form over a 3 mm increase in forklength. Thus, assessment of the Y intercept based on biological data supports the statistical theory that the GM regression equation is more appropriate for the back-calculation of forklength from scale radius (Ricker 1975; Ricker 1992). This is the regression that has been used. The regression may be used to back-calculate to lengths of RCH chinook included in construction of the regression (70 to 147 mm). Extrapolation to lengths greater than 147 mm may result in inaccurate forklength estimates due to a possible change in the relationship between scale radius and forklength for larger fish (Francis 1990; Ricker 1992). 38 2.2.2 Back-Calculation of Forklengths and Length Increments from Scale Radii Using GM Regression The GM regression equation (Figure 2.2: FL = 2.042 (SR) + 18.436; r2 = 0.67 P < 0.05) was used to back-calculate length at age from adult Robertson Creek Chinook scale measurements. This was done by substituting the scale radii corresponding to the desired length into the equation for SR. Ricker (1975) indicates that accurate back-calculation of individual forklengths requires the final scale radius for each individual. When these data are not available, he suggests that the average of the scale measurements be first estimated, and then back-calculation performed for the averages. This is what has been done. Somatic length Increments (FLI) corresponding to the distance between consecutive scale circuli (intercircular distances) were also back-calculated. This was done by subtracting the length estimated for the beginning of the intercircular interval from the length estimated for the end of the intercircular interval. A sample calculation is presented below for an intercircular distance that begins at a scale radius of 40 X 10"2 mm (SR1) and ends at a scale radius of 43 X 10"2 mm (SR2). FLI =((2.042(SR2)) + 18.536) - (2.042(SR1) + 18.536) FLI =((2.042(43) + 18.536) - (2.042(40) + 18.536) FLI = 106.34-100.22 FLI = 6.12 mm somatic length increase for the intercircular interval 39 25 0 0 y = 0.1543x +8.3511 R2 = 0.5969 10 20 30 40 50 Days from Median Release to Capture 60 Figure 2.3. Linear regression of the number of circuli present on the scales of juvenile chinook caught in 1989 and 1990 and the number of days from median RCH hatchery release date to recapture. 40 2.2.3 Conversion of Length Increments into Growth Rates Determination of the Apparent Rate of Circulus Addition Using Regression To convert the back-calculated length increments into somatic growth rate (mmd"1) it was necessary to estimate the typical number of days that lapse between the formation of circuli within the first marine zone. Apparent rates of circulus deposition and inversely, the number of days that it takes to form a circulus, were determined using data obtained from juvenile Robertson Creek Chinook salmon caught in 1989 and 1990. The rate of deposition was estimated as the slope of the regression of total scale circulus number on the number of days that the scales had been growing since hatchery release ("post release age"). "Post release age" was determined as the number of days from hatchery release to recapture during the MASS program. For some tag groups, this could be determined precisely by subtracting the calendar day of release from the calendar day of recapture. However, most of the tag groups were not released on a single day, but were released over a period of time that ranged from 2 to 20 days. Therefore, the number of days since release ("post release age") was estimated both as the capture date minus the first release date, and the capture date minus the median release date. These two measures of "post release age" are justified as it is not possible to determine which of the recaptured individuals had been released on the first day and which had been released later in the release period. The equations for the regressions of circulus number on days since first and median release dates using 1989 and 1990 juvenile Chinook data separate and combined are presented in Table 2.1. The slopes (circ d"1) calculated for the two measures of "post release age" have been averaged as neither is known to be more accurate than the other. The number of days that it takes to form a circulus (d circ"1) during early marine life is determined as the inverse of the averaged regression slope. The regression for the number of circuli versus days since median release date for the 1989 and 1990 juvenile Chinook data combined is presented in Figure 2.3. 41 Table 2.1. The equations for the arithmetic mean regression of circulus number on days since first and median release date for 1989 (n = 188) and 1990 (n = 125) data separate and combined (n = 313). Year Circ # vs. Days Since Circ # vs. Days Since Circ d-1 D circ-1 Mid-Release First Release 1989 C = 0.146d +7.797 C = 0.146d + 7.534 0.146 6~1J5 ?;P 0.665; 4.97 E-46 0.691; 2.81 E-49 1990 C = 0.154d +9.480 C = 0.162d + 9.00 0.158 6.33 A* ;P 0.620; 1.29 E-27 0.685; 1.14 E-32 1989+1990 C = 0.154d + 8.351 C = 0.157d + 8.005 0.156 6.42 r2; P 0.597; 2.52 E-63 0.634; 6.84 E-70 The circulus formation rates were determined using juvenile chinook scales that had deposited circuli before, during, and after salt water entry (SWE) check formation. As circuli that are laid down during formation of the check are irregular (Clutter and Whitesel 1956: Bilton 1982), it is possible that the rate of deposition is variable during this time. However, as scales with 6 to 19 circuli were included in the regression, the effect of the check circuli (approximately circuli 12 to 15) may simply be to increase variability in the regression. The periodicities of circulus deposition estimated using regression are tested using two mathematical calculations. Estimate of the Circulus Deposition Rate Over the Entire First Year The rate of addition of circuli over the entire first year of growth was determined from the number of circuli formed during this period and the total number of days taken to form the circuli. In Chapter 1 it was estimated that an average of 44.5 circuli were deposited on the scales of chinook during the first year of life (Table 1.2). Assuming 42 that both hatching and winter annulus completion occurs on approximately January 30 (Bilton and Ludwig 1966; Barber and Walker 1988), there is roughly 1 year (365 days) between hatching and the end of the first year's growth. However, as RCH Chinook do not possess circuli on their scales until approximately 60 days after hatching (~ end of March), the total number of days to lay down all circuli in the first year is roughly 305 (365-60). Dividing the number of days to form the circuli (305) by the total number of circuli formed (44.5 circuli) provided an average value of 6.85 days to form each circulus during the first year. Estimate of the Circulus Deposition Rate Over the First Marine Year Because the freshwater zone and the region of the SWE check may be part of different growth stanzas than the first marine period (Ricker 1992), a mathematical determination of the circulus deposition rate for the first marine period, following SWE check completion was also performed. This circulus deposition period was determined by dividing the number of days between ocean entry and the end of the first marine year by the average number of circuli formed in the first marine year. In Chapter 1 it was found that RCH Chinook deposited an average of 31 circuli in the first marine zone of their scales (Table 1.2). The calendar day at the end of the SWE check (179) was determined as the release date (approximately day 145) plus about 34 days to complete the check following hatchery release (estimated in this chapter). Subtracting this value (179) from 365, it was found that there are roughly 186 days from SWE check completion to the end of the calendar year. Adding 30 days for the end of the winter annulus completion (~ January 30) provided a formation time of 216 days (186+30). Dividing the formation time by the number of circuli formed in the first marine year (216 d/31 circ.) resulted in an estimate of 6.97 days per circulus. Summary of Apparent Rate of Circulus Addition The number of days to form a scale circulus estimated by regression (Table 2.1) and the two additional calculations are presented in Table 2.2. 43 Table 2.2. Summary of the periodicity of circulus deposition determined from several methods: regressions of circulus number on days since release using 1989 and 1990 juvenile Chinook separately and combined; and two additional calculations. Method of Determination Circulus Deposition Rate (dcirc1; 1. Regression: 1989 only 6~1) 2. Regression: 1990 only 6.3 3. Regression: 1989 and 1990 combined 6.4 4. Number of days to form all 1st year circ 6.9 5. Number of days to form 1st marine circ 7.0 Average of methods 3-5 6.8 It can be seen that there is little variation between the results from the two mathematical calculations and from the juvenile scale regressions (Table 2.2). The final value that will be used to approximate the rate of deposition, 6.8 days, was determined as the average of the results for the 1989 and 1990 combined regression (Table 2.1, Figure 2.3: r2 = 0.60; P = 2.52 E-63) and the two additional calculations (Table 2.2). Equation for the Conversion of Somatic Growth Increments into Growth Rates in mmd1 Growth rates in mmd"1 (GR) were calculated by dividing the back-calculated length increments by the number of days for the growth over one scale increment (Table 2.2: 6.8 d). An example calculation is provided below to illustrate how a forklength increment (FLI) was back-calculated from the distance between the scale radii 40 X 10"2 mm (SR1) and 43 X 10~2 mm (SR2), and the resulting value converted into growth rate in mmd"1. 44 GR = GR = GR = GR = GR = = FLI/6.8 d = ((2.042(SR2) = ((2.042(43) + = (6.12)/6.8d + 18.536) 18.536)-- (2.042(SR1) + 18.536))/6.8 d (2.042(40)+ 18.536))/6.8d = 0.90 mmd"1 over the intercircular interval (6.8 days) Time of Year that the Somatic Lengths and Growth Rates are Completed From the information determined thus far, it was possible to estimate the average time of the year that the Chinook attain each of the first year forklengths and experience the marine growth rates. Smolt size has been back-calculated from the radius of the scale to the end of the SWE check. To determine the time of the year that smolt size is achieved, it is necessary to determine the approximate time for completion of the SWE check following hatchery release. It was determined in Chapter 1 that adult RCH chinook have an average of 13.6 ± 1.3 circuli on their scales at SWE check completion (Table 1.2). Thus, the time for completion of the SWE check can be estimated from the number of days it takes for the scales to have approximately 13.6 circuli on their scales following release. This was estimated by back-calculation from the regression of the number of circuli present on the scales of juvenile chinook caught in 1989 and 1990 on the number of days from median release to recapture (Figure 2.3: Y = 0.1543x + 8.3511; r2 = 0.60; P = 2.52 E-63). From this regression, it was estimated that it takes on average, 34 days to complete the SWE check from the time of hatchery release. Using the median release date for 1982 to 1992 (19 May; day 140), the SWE check is completed (and smolt size is established) on average by 23 June (day 174). To determine the approximate date of formation of all circuli in the first marine year, the number of days to form a circulus (6.8 dcirc"1) was added for each circulus following the time of SWE check completion. An example calculation is provided below. The release date (RD) of day 140 (19 May) and the rate of deposition of scale circuli of 6.8 dcirc'1 are used to determine the date that the 4th marine circulus (Circ #) is completed. 45 Calendar Day = Calendar Day = Calendar Day = = RD + 34 d + = 140d+34d = 201 d, or 20 (Circ #)(6.8 dcirc"1) + (Circ 4)(6.8 dcirc"1) July for completion of the 4th marine circulus The approximate date for the completion of the winter annulus determined from the average number of circuli added in the first marine year (Table 1.2: 31) was estimated as 20 January (day 20). 2.3 RESULTS 2.3.1 Back-Calculated Forklengths Table 2.3. Average and standard deviation (SD) of the back-calculated length at ocean entry (smolt length), length added from the end of the SWE check to the end of the first marine annulus (Marl increase), and the length at the end of the first marine year for 1982-83; 1985-92. Lengths are in units of mm. Year 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Smolt Length 93.7 98.6 98.4 102.6 105.6 103.7 101.9 100.2 103.1 106.7 SD 8.6 8.9 6.2 7.7 6.3 6.9 8.4 8.0 9.0 7.7 Marl increase 203.3 218.8 210.1 210.8 198.3 217.1 216.6 205.3 213.3 212.5 SD 23.1 24.2 19.6 21.1 18.1 28.4 26.0 27.6 23.3 24.4 Length End Yearl 278.5 298.8 290.0 294.9 285.3 302.4 300.0 287.1 297.8 300.7 SD 28.2 25.8 21.5 25.5 20.5 30.7 28.7 29.1 25.8 27.5 Av. 101.5 7.8 210.6 23.6 293.6 26.3 118 112 E 106 E o 100 z UJ CC 9 4 O U. 88 • 82 mm MF.AN • ... : ... «i-r I ' 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.4. Mean and standard deviation of the length at the end of the SWE check (smolt size) for 1982-83; 1985-92. 250 240 230 E E 220 a 210 z UJ ^ 200 CC O U. 190 180 170 M E A N 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.5. Mean and standard deviation of length added from the SWE check to the end of the first marine annulus for 1982-83; 1985-92. 47 E E, x (5 Z HI _l *: cc o 340 320 300 280 260 240 ma MEAN • w t f • • * w —J— • • 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.6. Mean and standard deviation of length at the end of the first marine year for 1982-83; 1985-92. 48 The average length of chinook at the end of the SWE check (smolt length) was found to range between 93.7 ± 8.6 mm (1982) and 106.7 ± 7.7 mm (1992). It averaged 101.5 ± 7.8 mm over all years (Table 2.3). The averages and standard deviations for these years are presented in Figure 2.4. The average length increase from the end of the FW zone to the end of the first marine year (Figure 2.5) ranged between 198.3 ± 18.1 mm (1987) and 217.1 ±28.4 mm (1988). It averaged 210.6 + 23.6 mm over all years considered (Table 2.3). The average length at the end of the first marine year was found to range between 278.5 ± 28.2 mm (1982) and 302.4 ± 30.7 mm (1988) (Table 2.3, Figure 2.6). The average value for all years was found to be 293.6 ± 26.3 mm (Table 2.3). 2.3.2 Early Marine Growth Rates The average and the standard deviation of the growth rates averaged over three consecutive marine circuli between the 2nd to the 4th (Av2-4) and the 10th to the 12th (Av10-12) are presented in Table 2.4. The highest growth rate in this series (1.08 mmd1) was found for Av5-7 in 1989, and the lowest (0.82 mmd"1) was found for Av2-4 in 1982 (Table 2.4). There was a 0.26 mmd'1 difference between the lowest and the highest average growth rates. This corresponds to a difference in the gain in forklength over an intercircular interval (approx. 6.8 d) of 1.76 mm between chinook with the highest and those with the lowest growth increments. Similar trends between 1982 and 1992 were found for growth rates between Av2-4 and Av10-12 (Figure 2.7). The highest growth rates were found in 1988 to 1990, and the lowest in 1982 to 1986. To visualize the growth rates estimated, the averages and standard deviations for Av3-5 (Figure 2.8), Av5-7 (Figure 2.9), Av8-10 (Figure 2.10), and Av10-12 (Figure 2.11) are presented. 49 a) YEAR Av2-4 Av3-5 Av4-6 Av5-7 Av6-8 Av7-9 Av8-10 Av9-11 Av10-12 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Ave rape 0.82 0.84 0.86 0.83 0.89 0.92 0.93 0.90 0.88 0.90 0.88 0.84 0.85 0.88 0.88 0.91 0.97 1.01 0.95 0.90 0.94 0.91 0.87 0.83 0.89 0.90 0.92 0.99 1.06 0.97 0.93 0.93 0.93 0.90 0.83 0.91 0.90 0.94 1.02 1.08 0.96 0.95 0.93 0.94 0.91 0.85 0.91 0.90 0.95 1.02 1.07 0.97 0.95 0.93 0.94 0.92 0.89 0.93 0.91 0.95 1.03 1.04 0.97 0.93 0.96 0.95 0.90 0.90 0.92 0.90 0.94 1.00 1.02 0.98 0.91 0.96 0.95 0.90 0.92 0.93 0.90 0.94 1.01 0.99 0.99 0.92 0.96 0.94 0.90 0.91 0.93 0.91 0.93 1.00 0.98 1.01 0.94 0.95 0.95 b) YEAR Av2-4 Av3-5 Av4-6 Av5-7 Av6-8 Av7-9 Av8-10 Av9-11 Av10-12 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Ave rape 0.15 0.11 0.13 0.16 0.12 0.14 0.11 0.15 0.15 0.16 0.14 0.16 0.15 0.13 0.17 0.13 0.16 0.10 0.17 0.13 0.15 0.15 0.16 0.15 0.11 0.15 0.14 0.17 0.13 0.16 0.14 0.16 0.15 0.15 0.14 0.13 0.14 0.15 0.17 0.15 0.15 0.16 0.17 0.15 0.15 0.15 0.13 0.15 0.16 0.16 0.15 0.14 0.16 0.18 0.15 0.17 0.15 0.13 0.15 0.14 0.17 0.15 0.14 0.15 0.18 0.15 0.16 0.15 0.14 0.16 0.14 0.17 0.16 0.15 0.14 0.18 0.15 0.15 0.13 0.15 0.15 0.12 0.17 0.16 0.15 0.14 0.17 0.15 0.16 0.12 0.14 0.13 0.13 0.15 0.14 0.17 0.14 0.17 0.15 Table 2.4. The a) average and b) standard deviation of growth rates (mmd"1) for 1982-83; 85-92. The growth rates are averaged over three consecutive marine circuli from 2 to 12. 50 - • •— Av2-4 - A - -Av3-5 —N— Av4-6 —3R-— * --e--Av5-7 -Av6-8 -Av7-9 -Av8-10 ~Av9-11 -Av10-12 # <fi N# # / ^ N# N# # # Year j-1x Figure 2.7. Somatic growth rates (mmd ) for 1982-83; 85-1992 averaged over three consecutive marine circuli between 2 to 12. 51 1.15 1.05 E .§. 0.95 UJ I-< DC I 0.85 O DC ° 0.75 V. 0.65 — [ -a- MEAN ] ' ' 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.8. Average and standard deviation of the growth rate (mmd"1) averaged over the 3rd to 5th marine circuli (Av3-5). •a E E, UJ < DC 1.3 1.2 1.1 1.0 0.9 O 0.8 DC O 0.7 [ 0.6 -a- MEAN 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.9. Average and standard deviation of the growth rate (mmd"1) averaged over the 5th to 7th marine circuli (AV5-7). 52 Figure 2.10. Average and standard deviation of the growth rate (mmd"1) averaged over the 8th to 10th marine circuli (Av8-10). 1.2 1.1 •D £ E, 1.0 UJ l-< CO X 0.9 o DC C3 0.8 0.7 — — I - o - MEAN 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 YEAR Figure 2.11. Average and standard deviation of the growth rate (mmd"1) averaged over the 10th to 12th marine circuli (Av10-12). 53 2.3.3 Summary of Lengths and Growth Rates The following table (Table 2.5) is a summary of the average lengths and growth rates estimated for the first marine year. The mean coefficient of variation (CV) for the lengths and growth rates is included for a comparison of the average within cohort variability between these measures. The average number of scale circuli in the corresponding scale zones (Chapter 1), and the approximate time of the year that the lengths and growth rates were completed are also presented. Table 2.5. Summary of the number of circuli, the coefficient of variation (CV), and the approximate date of formation for the lengths and growth rates back-calculated from adult scales. The date provided for the growth rates is estimated for the last circulus included in the average. Measure 1.SWE Length 2. Av2-4 3. Av3-5 4. Av4-6 5. Av5-7 6. Av6-8 7. Av7-9 8. Av8-10 9. Av9-11 10. AvlO-12 11. Mar 1 Width 12. End Mar 1 Length Av. All Yrs. 101.5 mm 0.88 mmd'1 0.91 mmd"1 0.93 mmd"1 0.94 mmd"1 0.94 mmd"1 0.95 mmd"1 0.95 mmd"1 0.94 mmd"1 0.95 mmd"1 210.6 mm 293.6 mm Min. Av. 93.7 mm 0.82 mmd"1 0.84 mmd"1 0.83 mmd"1 0.83 mmd"1 0.85 mmd"1 0.89 mmd"1 0.90 mmd"1 0.90 mmd"1 0.90 mmd"1 203.3 mm 278.5 mm Max. Av. 106.7 mm 0.93 mmd"1 1.01 mmd"1 1.06 mmd"1 1.08 mmd"1 1.07 mmd"1 1.04 mmd"1 1.02 mmd"1 1.01 mmd"1 1.01 mmd"1 218.8 mm 302.4 mm Av. Circuli # 13.6 15.6 to 17.6 16.6 to 18.6 17.6 to 19.6 18.6 to 20.6 19.6 to 21.6 20.6 to 22.6 21.6 to 23.6 22.6 to 24.6 23.6 to 25.6 30.8 44.5 CV 7.7% 16% 16% 16% 16% 16% 16% 16% 16% 15% 11% 9.0% Date 23 June 20 July 27 July 3 Aug. 10 Aug. 16 Aug. 23 Aug. 30 Aug. 6 Sept. 13 Sept. 23 June-20 Jan. 20 Jan. 54 2.4 DISCUSSION Although it has been shown that some juvenile chinook can tolerate immediate entry into the marine environment, they tend to spend varying amounts of time in the estuary. The length of time spent in the estuary varies between rivers and years, however, fry from Nitinat Lake and the Nanaimo River spent an average of 17-25 days (Healey 1991). Mark and recapture studies have found that some chinook remain in the estuary for as long as two months (reviewed by Healey 1991). The amount of time that passed from release to completion of the SWE check for RCH chinook was estimated at 34 d. However, entry into the ocean likely occurs over the period of time during which the 2 to 4 circuli of the SWE check are formed. The beginning of the transition to salt water can be estimated from an approximation of the time following release to the beginning of SWE check formation. Since the number of circuli to the end of the SWE check (Table 1.2: 13.6 ± 1.3) was counted to the second to last check circulus (Chapter 1), and SWE checks are generally comprised of 2-4 circuli, the SWE check begins when the chinook scales have an average of 11.6 circuli on their scales (13.6+1-2). The amount of time that the chinook reside in the estuary before laying down their first check circulus (11.6) back-calculated from the regression equation (Figure 2.3: Y = 0.154(x) + 8.3511; r2 = 0.60; P = 2.52 E-63) is 21 d. Therefore, emigration from the estuary may occur on average between 21 to 34 days following release. From the median release date of 19 May, the transition to salt water is estimated to occur between 10 June and 23 June. This is consistent with the juvenile chinook catch per unit of effort data (Chapter 4, Figure 4.5) indicating that the numbers of chinook that emigrated to Barkley Sound had begun to increase by 6 June, and peaked by 21 June. Smolt size for RCH chinook was estimated by back-calculation from the scale radius to the end of the salt water entry check (Figure 1.1). Average lengths over the ten years (1982-83; 85-92) ranged from 93.7 ± 8.6 mm to 106.7 ± 7.7 mm (Tables 2.3; 2.5). This is consistent with observations that chinook fry begin emigration from the estuary once they reach about 70 mm forklength (Healey 1991). However, as mentioned above, the smolt size back-calculated from the scale radius to the end of the SWE 55 check may be an overestimate of the actual size of Chinook upon entry into the ocean. It is possible to obtain a rough estimate of the length at the beginning of the SWE check. The intercircular distances for the final check circuli were generally found to range between 1.3 and 2.3 X 10"2 mm. By back-calculation, the amount of somatic length added for each of these circuli is thus 2.7 to 4.7 mm. If the length of the Chinook salmon to the end of the SWE check is decreased by 2 circuli (2X2.7 mm or 2X4.7 mm), then the range of smolt lengths is reduced by 5.4 mm (10.4 mm) to range between 88.3 mm (84.3 mm) and 101.3 mm (97.3 mm) forklength. The transition to salt water probably occurs at lengths between those calculated for the beginning and the end of the SWE check. The within-cohort variation (CV) for smolt length was found to be low (Table 2.5, Measure 1: CV = 7.7%). This suggests that substantial depensation in Chinook forklengths may not have occurred by the time that their SWE checks were completed. However, the relatively high variation in the growth rates (Table 2.5, Measures 2-10; CV = 15-16%) suggests that depensation may be occurring during the early marine period, resulting in an increase in the within-cohort variability in fish length over time. Growth compensation during the latter half of the first marine year (discussed in Chapter 1) may produce the reduction in the within cohort variation observed for the length at the end of the first marine year (Table 2.5, Measure 12: CV = 9.0%). Lengths at that time were estimated to average 293.6 ± 26.3 mm over all years (Table 2.5), and averages ranged from 278.5 ± 28.2 to 302.4 ± 30.7 mm (Tables 2.3). The final first year length was estimated to be completed by approximately 20 January (Table 2.5). These data are fairly consistent with observations that two races of ocean-type chinook caught in roughly February to March of their second year (i.e. after the first winter) were approximately 300 mm forklength (Healey 1991: see Figure 31). Thus, although the regression of forklength on scale radius was constructed with chinook much smaller than 300 mm (70-146 mm), it appears to provide fairly accurate estimates for these forklengths. 56 The growth rates during early marine life were also estimated for RCH chinook salmon. To obtain somatic growth rates in mmd"1 from the somatic length increments corresponding to the distance between two scale circuli, it was necessary to estimate the periodicity of circulus deposition. Numerous counts in the literature suggest that scale circuli are deposited at fairly regular intervals under both laboratory and natural conditions (Bilton 1982; Healey 1982; Holtby et al. 1990; Fukuwaka and Kaeriyama 1997). This suggests that there may be inherent controls on circulus deposition rate (Bilton 1975, 1982). However, it has also been shown that changes in environmental conditions such as food availability or light period can influence the rate of deposition (Bilton 1982). Furthermore, it was found that sockeye salmon did not produce any circuli during periods of starvation (Bilton and Robins 1971; Bilton 1982). Therefore, although there appears to be an inherent periodicity in circulus deposition, the rate can vary under certain conditions. However, in Chapter 1, it was found that the number of circuli added before the end of the SWE check and during the first marine year was relatively constant between years for RCH chinook (Table 1.2). Assuming that the time to form these circuli is approximately equal between years, the periodicity of circulus deposition appears to be relatively stable between years for these chinook salmon. The number of days for RCH chinook scales to form a circulus during the first year of life was estimated as 6.8 days (Table 2.2). Using this value, the average growth rates estimated for the 2nd to 12th marine circuli ranged between 0.82 ±0.15 and 1.08 ± 0.15 mmd"1 (Table 2.4). Growth rates within 1 standard deviation of these averages ranged from 0.67 to 1.23 mmd"1. Although information on the marine growth rate of chinook salmon is sparse, Healey (1980) estimated values between 0.75 and 0.85 mmd"1 for Strait of Georgia chinook (1976) and Miller et al. (1983) presented growth rates of 1.74 mmd"1 for the Washington/Oregon coast (1980). Thus, the growth rates that have been estimated are well within the range of reported values. However, the literature values were estimated by determining the change in length over time for chinook caught within a given region. Size-biased outmigration from the sampling area, as well as the addition of new individuals to the region tends to result in large errors in these growth rate estimates. Therefore, obtaining growth rates from scales provides an advancement in accuracy, assuming that size-selective mortality has not influenced the 57 estimates greatly. The highest growth rates during early marine life were found for the 1988 to 1990 ocean entry years and the lowest for 1982 to 1985 (Figure 2.7). Similar trends in growth rates were found from the 2nd to the 12th marine circulus, occurring between approximately 6 July and 12 September. This suggests that interannual variability in factors affecting growth rate are relatively constant over this time period. However, the Av2-4 did not match the other growth values as well. This may result from the inclusion of the 2nd marine circulus in the average. This circulus is formed soon after the irregular circuli of the SWE check and the growth rate of the fish may be influenced by physiological changes resulting from the stress of habitat change and salt water adaptation. 58 CHAPTER 3. THE RELATIONSHIP BETWEEN MARINE SURVIVAL AND BACK-CALCULATED FIRST YEAR LENGTHS AND MARINE GROWTH RATES 3.1 INTRODUCTION The mortality rate of salmonids has been found to be particularly high during the first few months of ocean residence (Parker 1968; Matthews and Buckley 1976; Bax 1983; Fisher and Pearcy 1988). Parker (1968) estimated daily mortality rates for pink salmon of 2 to 4 % for the first 40 days after ocean entry. He found that the mortality rate decreased substantially to 0.4 to 0.8% per day for the following 410 days studied (Parker 1968). High interannual variability in mortality rates is characteristic of salmonid species. To increase our understanding of recruitment variations considerable effort has been made to determining whether certain characteristics of individuals allow them to survive at a higher rate. Size at ocean entry (Parker 1971; Healey 1982; Ward et al. 1989; Henderson and Cass 1991) and the growth rates realized during early marine life (Holtby et al. 1990) have been found to be important for the survival of some stocks of salmonids. According to size-spectrum theory, individuals entering the ocean at a smaller size are susceptible to predation by a larger number of marine predators than their larger conspecifics. Growth rates achieved following ocean entry modify the amount of time spent vulnerable to successive predator fields, each of which is less numerous than the previous (reviewed by Sogard 1997). Small size and low growth pose an additional threat due to increases in non-predation mortality, including decreased resistance to starvation, lower tolerance to environmental extremes (Webb 1995; Sogard 1997), and an increase in the susceptibility to disease (Fagerlund et al. 1995). Interannual variations in early marine growth rates are likely a density-dependent function of the interannual variability in the overall productivity of the nearshore nursery feeding areas in which the fish reside. In years of low growth and low survival, mortality can also increase due a reduction in the abundance of alternate prey such as zooplankton for predators or changes in the predator abundance varying concurrently with oceanographic productivity. 59 In this chapter, regression analysis is used to assess possible links between the marine survival rates of Robertson Creek chinook salmon entering the ocean between 1982 and 1992 and their first year lengths and marine growth rates. 3.2 METHODS AND MATERIALS 3.2.1 Survival Rate Data The survival rates used for this analysis were obtained from the Robertson Creek hatchery (Glen Rasmussen, unpublished data). They were determined as the proportion of marked chinook fry released between 1982 and 1992 that were either recovered in sport and commercial fisheries or returned to the hatchery to spawn. Thus, the survival rates represent the proportion of chinook from each smolt year that did not die from natural causes in the marine environment. The values obtained are cohort survival rates as they incorporate all of the various ages-at-retum (Figure 3.1). Although there was found to be less than 5% survival in all years, significant interannual variability in survival rates exists as the cohort survival rate for the highest survival year (1989: 4.4%) was found to be 147 times higher than for the lowest (1984: 0.03%). The survival rates do not take account of catch and release mortality, or those fish that stray from their home spawning stream. However, the catch and release mortality may not vary substantially between years, and thus may represent a somewhat constant additional means of mortality included within the natural mortality component. As well, the number of chinook that stray is likely to be small, as ocean-type chinook have been found to show very strong fidelity to their release site -estimates over 4 brood years were found to average 98.6% fidelity. However, it has also been found that straying increases for older chinook, or for those from poor survival years (reviewed by Heaiey 1991). Therefore, the importance of straying may increase for some age groups or smolt years. 60 5.00 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 Ocean Entry Year Figure 3.1 Marine survival rate (%) for cohorts of Robertson Creek Chinook salmon entering the ocean between 1982 and 1992. 61 3.2.2 Regression Analysis The relationships between survival and length or growth rate measures were determined by ordinary (AM) linear regression. Significance of an individual regression was assessed by a comparison of the test P value to the "table-wide" a level adjusted for multiple, simultaneous comparisons bias using the sequential Bonferroni technique discussed in Chapter 1 (Rice 1989; Sokal and Rohlf 1995). Early marine growth rates used in the regressions were obtained by averaging the growth rates over three consecutive marine circuli. This was done because evidence suggests that there may be a time lag in the response of circulus growth to changes in somatic growth rates (Major and Craddock 1962; Bilton and Robins 1971; Bilton 1982). As well, prior growth history may influence the growth of circuli as significant autocorrelation was found in the marine growth rates between the 2nd and 13th circulus at lag 1 for some years (Chapter 1: Figure 1.12). Averaging three consecutive growth rates should also smooth out any short-term variability in scale growth rates. The notation, Av3-5, Av4-6, etc. indicates which marine circuli (those that follow the end of the SWE check) are used to calculate the average growth rate. 3.3 RESULTS 3.3.1. Regression of Cohort Survival Rate on Back-Calculated Somatic Measures The survival rates for ten years between 1982 and 1992 were regressed on the length at ocean entry, the length added during the first marine year, the length at the end of the first year, and growth rates averaged over the 3-5th, 4-6th, and 8-1 Oth marine circuli. The results are given in Table 3.1. The associated figure number is provided in the last column of the table. 62 Table 3.1. Equations for the linear regression of marine survival rate on the somatic lengths and growth rates back-calculated from the average scale values for 1982-83; 1985-92 (n = 10). The "table-wide" a levels that the test P values must not exceed to be significant at P = 0.05 using sequential Bonferroni tests are included. Significant P values are marked with an asterisk (*). Survival Regressed on: Equation T P value Table-Wide a Fig. # 1. Length at SWE check Y = 0.0008x-0.067 0.0496 0.536 0.0167 3.2 2. Length Added in First Y = 2E-05x+0.014 9E-05 0.979 0.05 3.3 Marine Year 3. Length at End First Year 4. Av3-5 5. Av4-6 6. Av8-10 Y = 0.0002X-0.045 Y = 0.2118x-0.175 Y = 0.0569x-0.156 Y = 0.2775X-0.244 0.0142 0.6218 0.7183 0.668 0.743 0.00672* 0.00196* 0.00388* 0.025 0.0125 0.0083 0.01 3.4 3.5 3.6 3.7 From these results it can be seen that the only measures that are significantly positively correlated with survival between 1982 and 1992 are the early marine growth rates: Av3-5 (Regression 4.: r2 = 62; P < 0.0125); Av4-6 (Regression 5.: r2 = 0.72; P < 0.008); and Av8-10 (Regression 6.: r2 = 0.67; P < 0.01)) (Table 3.1). The annual growth and marine survival rates are shown in Figures 3.8 a-c. It is evident that the 1992 ocean entry year had a very high growth rate relative to survival. 1992 was found to be a significant outlier in all three regressions (Regression 4.: Studentized residual = -6.378; Regression 5.: Studentized residual = -3.675; Regression 6: Studentized residual = -5.967). Regressions were run again with 1992 excluded. The results are presented in the following table (Table 3.2). 63 0.055 0.045 0.035 < g 0.025 CO 0.015 0.005 -0.005 '*•-.. " » „ • • •• "82" O • * t o 83 O . .-••" 90 89 o 88 ..•"' o. ••* . * * o o 86 .. -O... ' • • 92. o •• 9 2 94 96 98 1 0 0 102 FW LENGTH 1 0 4 1 0 6 108 Figure 3.2. Linear regression and 95% confidence bands for marine survival and forklength (mm) to the end of the salt water entry check (FW length) for 1982-83; 85-92 ocean entry years. 0.055 0.045 0.035 < ^ 0.025 DC Z> to 0 . 0 1 5 0.005 -0.005 196 2 0 0 • . • • 87 O * ""90-... O 82 O . • • * 85 086 O 89 O 88 ..o--o 92 *• • . O O '•. . • • • • 204 208 212 216 MAR1 LENGTH INCREASE 2 2 0 2 2 4 Figure 3.3. Linear regression and 95% confidence bands for marine survival and forklength (mm) added in the first marine zone (Marl Length) for 1982-83; 85-92 ocean entry years. 64 0.055 0.045 0.035 < g 0.025 cc r> 0.015 0.005 -0.005 - . 82 O * % , 87 O -aa. o 86 O 89 O 88* 6* -92--. o • • 276 282 288 294 300 LENGTH AT THE END OF THE FIRST YEAR 306 Figure 3.4. Linear regression and 95% confidence bands for marine survival and forklength (mm) to the end of the first marine year for 1982-83; 85-92 ocean entry years. < > > DC 3 u.uoo 0.045 0.035 0.025 0.015 0.005 _n nnn - " * " 82 Q3 .••' . c > . * 87 o .• • • 91 .-• 85 .Or ' Q86' .'"" -' o . •* „ ' 89 O . -8*8 .•' O .-'90 o .•' " * • " " » * * ^ • * 92 O , . • • " ' # • - " ' 0.82 0.86 0.90 0.94 AV3 5 0.98 1.02 1.06 Figure 3.5. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd1) averaged over the 3rd to 5th marine circuli (Av3-5) for 1982-83; 85-92 ocean entry years. 65 0.055 0.045 0.035 < ^ 0.025 h DC V) 0.015 0.005 -0.005 0.80 83 O . 0.84 82 Q 85 PB6 " O 87 O •91 O 92 O 90-O 88 O 0.88 0.92 AV4 6 0.96 1.00 89 1.04 1.08 Figure 3.6. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd"1) averaged over the 4th to 6th marine circuli (Av4-6) for 1982-83; 85-92 ocean entry years. 0.055 0.045 0 . 0 3 5 ^ 0 . 0 2 5 DC to 0.015 • 0.005 -0.005 0.88 0.90 0.92 0.94 0.96 AV8 10 0.98 1.00 1.02 1.04 Figure 3.7. Linear regression and 95% confidence bands for marine survival and the growth rate (mmd"1) averaged over the 8th to 10th marine circuli (Av8-10) for 1982-83; 85-92 ocean entry years. 66 1.05 CD CO CC % o (5 CD CC J 0.95 --0.9 --0.85 0.8 5.00% 4.50% 4.00% 3.50% 3.00% 2.50% 2.00% 1.50% 1.00% 0.50% 0.00% CD •s LL > £ 3 to 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 5 4 4 3 3 2 2 1 1 0 0 0 0% 5 0% .0 0% . 5 0 % .0 0% . 5 0 % .0 0% .5 0% .0 0% 5 0% 0 0% CD CO CC co & 3 CO 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Year <D +-* co CC "co > CO Figure 3.8. Line graphs of marine survival rates (%) and growth rates (mmd"1) averaged over the a) 3rd to 5 th, b) 4th to 6th, and c) 8th to 10th marine circuli. 67 Table 3.2. Equations for the linear regression of marine survival rate on the lengths and growth rates back-calculated from average scale values for 1982-83; 1985-91 (n = 9). The "table-wide" a levels that the test P values must not exceed to be significant at P = 0.05 using sequential Bonferroni tests are included. Significant P values are marked with an asterisk (*). Survival Regressed on: Equation Is P value Table-wide a 1. Length at SWE check Y = 0.0021 x - 0.1907 0.285 0.139 0.0167 2. Length added in First Y = 0.0001 x - 0.0042 0.0034 0.881 0.05 Marine Year 3. Length at End First Y = 0.0005x - 0.129 0.0855 0.445 0.025 Year 4.Av3-5 Y = 0.2393x-0.1972 0.9335 2.27E-05* 0.0083 5. Av4-6 Y = 0.1893x-0.1552 0.8843 0.000161* 0.0125 6. Av8-10 Y = 0.3000x-0.2625 0.9332 2.31 E-05* 0.01 The results indicate that the length added during the first marine year and the total length at the end of the first year were still not significantly correlated with survival (Regression 2.: r2 = 0.003; P > 0.05 and Regression 3.: r2 = 0.086; P > 0.025, respectively). The relationship between survival and smolt size (length determined from the scale radius to the end of the SWE check) was also not found to be significantly correlated with survival (Regression 1.: r2 = 0.29; P > 0.017). However, removing the 1992 ocean entry year from the regressions resulted in substantial improvement in the relationship between the early marine growth measures and survival: statistically significant relationships were found for Av3-5 (Regression 4.: r2 = 0.93; P < 0.008), Av4-6 (Regression 5.: r2 = 0.88; P < 0.0125), and Av8-10 (Regression 6.: r2 = 0.93; P < 0.01) (Table 3.2). The growth rate determined for later in the first marine year (Av8-10) was found to be as highly correlated with survival as that estimated for very early marine growth (Av3-5). Survival was also found to be significantly correlated with the length at the end of the 10th marine circulus {r2 = 0.73; P = 0.0035). However, the 68 correlation results from the strong significant correlation between survival and the growth rates realized over the first 10 marine circuli. 3.4 DISCUSSION Smolt size was found to be variable interannually (Chapter 2), however, a significant relationship was not found with the interannual variations in marine survival (Table 3.2, Regression 1: r2 = 0.29; P = 0.139). Although many within-cohort studies have indicated that individual salmonids with larger size at ocean entry survive at a higher rate during early marine life than their smaller conspecifics (Parker 1971; Healey 1982; Hargreaves and LeBrasseur 1986; Ward et al. 1989; Henderson and Cass 1991), fewer studies have found correlations between interannual variations in recruitment and between-year variations in average smolt length. For the majority of salmonid populations studied: Carnation Creek coho (O. kisutch: Holtby et al. 1990); Fraser River sockeye (O. nerka: Henderson and Cass 1991); and Fraser River pink salmon (O. gorbuscha: Henderson et al. 1995), no consistent between year survival advantage has been found for large smolt size. However, for Keogh River steelhead (O. mykiss) a very good relationship between observed and back-calculated smolt size at ocean entry and marine survival has been found (Ward and Slaney 1988; Ward et al. 1989). Mean steelhead smolt lengths were larger and more variable (140 to 187 mm) than for the RCH Chinook and many of the other salmonids studied, possibly providing more of a contrast for size-selective mortality processes to act. Whether interannual variations in smolt size affect the smolt-to-adult survival rate likely depends on the characteristics of the particular prey and their main predators. The main predator of RCH chinook in the first few months following release was found to be the migratory stock of Pacific hake (Merluccius productus) (Hargreaves et al. 1991). These predators are generally found to reach high numbers in Alberni Inlet during June and July. The arrival time of hake coincides with the residence of juvenile chinook in the region of Alberni Inlet since the chinook are released between late May 69 and early June and were estimated to have emigrated from Alberni Inlet between primarily 6 June and 21 June (Chapter 2). Analysis of stomach content data indicates that the highest incidence of predation on juvenile salmon occurred in Alberni Inlet. Throughout June in 1990, between 35 and 45% of the hake captured there had consumed juvenile salmon (Hargreaves etal. 1991). However, the intensity of predation by hake was found to decline upon entry of Chinook into Barkley Sound. Only 3.5% had been found to prey on juvenile salmon in the areas adjacent the mouth of Alberni Inlet. No potential predators captured in outer Barkley Sound had consumed juvenile salmon during 4 years of sampling (1987-1990). The migratory stocks of hake range in length between 400 and 650 mm (Tanasichuk et al. 1991). They have been known to consume Pacific herring (Clupea harengus pallasf) predominantly between 120 and 200 mm, and are able to consume those up to 240 mm standard length (Tanasichuk et al. 1991). However, a comparison of the length frequency of juvenile chinook consumed by hake with that in the surrounding region suggests that hake predation was selective for smaller chinook (Hargreaves pers. comm.). If a large portion of marine mortality of juvenile RCH chinook occurs by hake soon after release, and hake are selective for smaller individuals, then it would be expected that cohorts with larger average size would suffer a smaller predation loss during their residence in Alberni Inlet. However, the length of these chinook at entry into Alberni Inlet is determined largely by the length at release from the hatchery. Release length was fairly constant between 1982 and 1992, at approximately 80 mm. Although smolt size was found to vary interannually (Chapter 1), this length was estimated to be achieved by roughly 23 June, once the chinook had migrated into Barkley Sound. Thus, the majority of hake predation appears to occur when the chinook are at lengths intermediate between release length and smolt length - before substantial growth depensation had increased the interannual variability in forklength. Therefore, interannual variability in length at the time of high intensity of size-selective predation may not have been sufficient to confer a noticeable survival advantage to cohorts with greater mean forklength. As well, since smolt size is determined after a large amount of mortality has probably occurred, the reduction in intraspecific competition that would result from a higher than average mortality in some years may allow the survivors to 70 attain larger size at the time of ocean entry. This would further reduce the likelihood of finding a relationship between survival and smolt size. A significant relationship between marine survival and the length of the Chinook estimated for the end of the first marine year was also not found (Table 3.2, Regression 3.: r2 = 0.09; P = 0.45). A similar result was found for back-calculated length at the end of the first marine year and marine survival for Barkley Sound sockeye salmon (Hyatt et al. 1989). For RCH chinook, the lack of relationship may result from the reduction in the length differential between cohorts towards the end of the first marine year discussed in Chapter 1. In contrast to the results for smolt length and length at the end of the first marine year, the early marine growth rates investigated were found to be very highly correlated with survival (Table 3.2, Regressions 4-6: r2 = 0.88 to 0.93; P < 0.05). Growth rates in the sea are considered to be a density-dependent function of the food availability. Factors such as temperature and interspecific interactions can also moderate growth (Weatherley and Gill 1995). Growth rate is important for survival as it governs or influences many processes: the length of time that fish are small enough to be preyed on by a large number of predators; the size and quality of prey that can be consumed; the intensity of intra- and interspecific competition for resources; physiological condition (fatness) (Weatherley and Gill 1995); and the ontogenetic development of the immune system (Lorenzen 1996). Low growth rates may lead to smaller size and a reduction in available energy reserves. Overwinter mortality in freshwater has been found to be higher for smaller fish due to metabolic allometry - smaller fish with higher metabolic rate per unit weight than larger fish, use up smaller fat stores at a higher rate (Post and Evans 1989). Susceptibility to parasitism and disease have also been found to increase for smaller and slower-growing fish (Fagerlund et al. 1995), and the mechanism may be through suboptimal body composition compromising the immune system. Whether slow growth results in increased non-predation mortality, or fish under poor growth conditions are more susceptible to predation, the effect of growth rate on mortality is important. Generally it is found that mortality decreases for 71 increasing sizes of marine organisms (Peterson and Wroblewski 1984). The same effect is suggested to be true for fish growing through successive life history stages. Theory states that fish grow through successive predator fields, each of which is less numerous than the previous (Cushing 1975). Thus, the faster that individual fish grow the higher their survival rate due to a reduction in the time spent vulnerable to a large number of predators. Inherent in this theory is that the predator field is stationary. However, the abundance and type of predators can change over different time scales, adding another degree of complexity, confounding the relationship between productive capacity, growth, and survival. Growth rates of juvenile salmon may index the productivity of the local environment, however, predation intensity may be somewhat unlinked to the dynamics of the pelagic community. This can occur when exotic species enter the nursery areas of the Chinook in an unpredictable fashion, such that evolutionary response are not able to develop to minimize mortality. The 1992 ocean entry year was found to be a significant outlier in the regressions of survival on early marine growth rate (Table 3.1: Regressions 4-6.: Studentized residuals = -6.378; -3.675; and -5.967, respectively). Inspection of the time series of three early marine growth rates and survival between 1982 and 1992 (Figure 3.8) indicates that extremely high growth rates relative to survival occurred in the 1992 ocean entry year. Since growth rates were determined from the scales of survivors from each cohort, the measures may not be representative of the actual cohort growth. If mortality in 1992 was highly biased toward smaller or slower-growing individuals, then the measurements taken from survivors may be inflated relative to that of the entire cohort (reviewed by Sogard 1997). The 1992 cohort entered the ocean in a summer characterized by strong El Nino conditions. A large biomass of mackerel (mainly Pacific mackerel (Scomber japonicus)) was found in the region of Alberni Inlet and Barkley Sound (Chapter 2: Figure 2.1).* Observations of the feeding behaviour of mackerel indicate that they are voracious predators, consuming large quantities of a single prey type at a time, and thus will tend to focus on prey organisms that are locally * Large numbers of mackerel were also found in 1984 (Ashton et al. 1985). However, survival in that year was so poor (0.03%) that insufficient numbers of scales were collected to include it in this study. 72 abundant (Fry 1936). Robertson Creek chinook salmon are locally abundant each year in Alberni Inlet and Trevor Channel. Approximately 8 million are released at the head of Alberni Inlet (Chapter 2: Figure 2.1) over a two week period between May and June. In 1984 and 1992 it was found that migratory mackerel preyed heavily on juvenile salmonids (Ashton et al. 1985; Hargreaves and Hungar 1994). Examination of the stomach contents of mackerel caught within Alberni Inlet and Trevor Channel in 1984 showed that some had consumed between 8 and 10 juvenile chinook each (Ashton et al. 1985), and in 1992, observations of 5 chinook consumed per mackerel were common (Hargreaves and Hungar 1994). The numbers of juvenile chinook salmon caught during MASS sampling in 1992 declined to about zero by early June (Hargreaves unpubl. data). In previous years chinook abundance remained high at least until the cessation of sampling (end of July). Chinook were released between 9-26 May in 1992. This suggests that the predation by mackerel may have been extremely intense in the first month after hatchery release (Hargreaves and Hungar 1994). It was estimated in Chapter 2 that the length at the end of the SWE check for the survivors in 1992 averaged 107 + 26 mm. Since SWE was found to occur roughly 34 days after release (Chapter 2), this length is achieved by 12 June - 2 July, after the juvenile catch had declined to ~ zero (Hargreaves and Hungar 1994). Since Pacific mackerel caught in 1984 ranged between 29 and 33 cm total length (Ashton et al. 1985), they would not have been physically limited in their ability to consume all sizes of juvenile chinook salmon during this time. However, the better predator evasion abilities of larger chinook may have enabled them to survive at a higher rate. Mackerel were also found to migrate into BC waters in another year included in the analyses (1983). However, the growth rates determined from the scales of survivors in 1983 were not noticeably elevated relative to survival (Figure 3.8). Observations suggest that the biomass of mackerel was lower in 1983 than 1992 (Ashton et al. 1985; Hargreaves and Hungar 1994). As well, the cohort survival rate was 4 times higher in 1983 than in 1992. The growth rates for both of years is expected to be low due to the low productivity associated with El Nino events. The fact that directional selection is apparent in the survivors from the very low survival year 73 (1992) and not the relatively higher survival year (1983) may support the theory that a very high intensity of size-biased mortality is necessary for the effects to become evident in the length frequencies of survivors (Sogard 1997). However, possible differences in the arrival timing of the migratory mackerel with respect to the distribution of RCH chinook in 1983 and 1992, as well as possible differences in the abundance of alternative prey for the mackerel in these years may provide alternative explanations for the observations. Differentiating alternative hypotheses to explain these data is beyond the scope of this thesis. The fact that only 1992 (and perhaps 1984) adult scales exhibit substantial growth elevation suggests that the growth rates determined from adult scales in 1982-1991 may be fairly representative of the true growth of age 0.3 RCH chinook. 74 CHAPTER 4. INVESTIGATION INTO THE MECHANISMS DETERMING INTERANNUAL VARIABILITY IN EARLY MARINE GROWTH AND SURVIVAL OF ROBERTSON CREEK CHINOOK SALMON 4.1 INTRODUCTION Mortality rates of salmon are generally very high during the first few months of marine life (Parker 1968; Matthews and Buckley 1976; Bax 1983; Fisher and Pearcy 1988). However the mechanisms determining recruitment variations in salmon are poorly understood as little is known about the time and location where the main determinants of marine mortality act. Correlational studies have indicated that stocks of salmon entering the ocean in regions of similar marine influence can at times experience similar interannual variations in smolt-to-adult survival (Scarnecchia 1981; Hyatt et al. 1989). This suggests that the environmental conditions shared by these fish during their ontogenetic migrations following ocean entry may have great influence on subsequent marine survival. Several studies have demonstrated that marine survival of salmonids is correlated with ocean conditions indicative of productivity in the first year following ocean entry (Scarnecchia 1981; Nickelson 1986; Holtby et al. 1990). However, the particular oceanographic indices used in these studies have been chosen predominantly through exploratory correlations or assumptions about migration timing, with little empirical evidence (Scarnecchia 1981; Nickelson 1986; Fisher and Pearcy 1988; Holtby et al. 1990). Evidence that differences in the time (Parker 1971; Bilton et al. 1982; Matthews and Ishida 1989; Henderson 1995) and location of ocean entry for salmonids can have large impacts on subsequent rates of marine survival points to the importance of obtaining a more precise knowledge of the time period over which the marine climate influences survival for increasing our understanding and eventual forecasting of recruitment variability. In the previous chapter, it was established that marine survival rates of RCH chinook entering the ocean between 1982 and 1991 were highly correlated with early marine growth rates. In this chapter, the time of the year during which growth rates are 75 highly correlated with survival is estimated, and the probable location of the juvenile Chinook during this time is then investigated. The timing and location are then analyzed in the context of the temporally dynamic oceanographic environment in an attempt to provide possible explanations for the interannual variability observed for marine growth and survival rates. 4.2 METHODS AND MATERIALS 4.2.1 Estimation of the Period of High Correlation between Marine Survival and Growth Marine growth rates that have been investigated occur following completion of the salt water entry (SWE) check. It was determined in Chapter 1 that adult RCH Chinook have an average of 13.6 ± 1.3 circuli on their scales at check completion. It was estimated in Chapter 2 that it takes on average, 34 days to form the SWE check from the time of hatchery release back-calculated using the regression of the number of circuli on the scales of juvenile Chinook caught in 1989 and 1990 on the number of days from release to recapture (Figure 2.3: Y = 0.1543x + 8.3511; r2 = 0.60; P = 2.52 E-63). RCH chinook caught near the release site were found to have between 8 or 9 scale circuli (Figure 4.1: Area 1). From the time of hatchery release to the completion of the SWE check, the scales add approximately 5 circuli (13.6 - 8.5) over approximately 34 days (6.8 d circ"1 *5 circ). Assuming the median release date between 1982 and 1992 of 19 May (day 140), the SWE check was completed on average by 23 June (day 174). The approximate date of formation of all circuli in the first marine year was estimated by adding the number of days to form a circulus (6.8 d circ'1: Chapter 2) for each circulus following the time of SWE check completion. An example calculation is provided in Chapter 2. To 76 1 2 30 20 -10 r i - r - i 10 15 CIRCTOTAL 0.3 - 0 . 2 - 0 . 1 03 Si 20 0.0 15 10 -1 r i - •• 5- i— 10 15 CIRCTOTAL - 0 . 2 5 o o 3. 0 1 o° 20 0.0 10 8 Count 4 2 0 c 1 1 — -i 10 15 2 CIRC TO! "AL 0.3 - 0 . 2 0 . 1 03 0.0 7 6 5 c 4 3 O O 3 2 1 rv I 1 ---0.3 0.2 0.1 B> 10 15 CIRCTOTAL 20 0.0 Figure 4.1. Frequency histograms of the number of scale circuli found on juvenile chinook salmon in Areas 1-4. Area 1 is Alberni Inlet, Areas 2-4: Trevor Channel; Imperial Eagle Channel; and Loudon Channel in Barkley Sound are in order of increasing distance from the point of hatchery release (Somass Estuary). 77 determine which of the first marine year circuli are most highly correlated with survival, regressions of marine survival on marine growth rates averaged over 3 consecutive circuli from the 2nd to the 30th marine circuli were performed. The coefficient of determination (r2) of survival on these growth rates was then plotted versus the calendar day of completion of last circulus included in each average growth rate (Figure 4.2). The release date of 140 (19 May) was estimated as the median release date found for the 1982 to 1992 hatchery releases (Robertson Creek Hatchery unpubl. data). A table (Appendix III: Table 111.9) is also provided to estimate the date for each of the marine circuli to the end of the highly-correlated period (circulus 12) for release dates between 130 and 162 (10 May - 11 June). This enables the approximation of the highly-correlated time for particular years in which the release date differed from the estimate of 19 May (day 140). 4.2.2 Location of the Majority of RCH Chinook throughout the Highly-Correlated Period Probable Location of Chinook at the Initiation of Marine Growth The probable location of chinook at the initiation of marine growth was determined by comparing the number of circuli present on the scales of juvenile chinook captured in Alberni Inlet and Barkley Sound with the number present on adult scales at ocean entry. In Chapter 1 it was found that the number of circuli laid down before the end of the SWE check on adult scales averaged 13.60 ± 1.3 over the ten years studied (Chapter 1). Thus, chinook with more than 12.3 (13.6-1.3) circuli on their scales may have a completed SWE check and have thus begun marine growth. In the MASS juvenile chinook database, a code was given for the general region in the study area that each of the juvenile chinook was captured (Dr. Brent Hargreaves unpublished data). The code ranges from 1 to 4 with generally increasing distance from the Somass Estuary: Area 1 is Alberni Inlet; Area 2 is Trevor Channel; Area 3 is Imperial Eagle Channel; and Area 4 is Loudon Channel (Figure 4.3). Frequency histograms of the number of circuli present on the scales of juvenile chinook captured in each of the 4 areas were plotted (Figure 4.1). The proportion of chinook with greater than 12.3 scale circuli in 78 1.20 g CO c "E <D •4—» a) c CD "o "o o O <f ^ # ^ # ^ ^ <vj v <$" <sj «sj Calendar Day of Last Circulus in Average Figure 4.2. The coefficient of determination (revalue) for the regression of marine survival on the growth rates (mmd"1) averaged over three consecutive marine circuli from 2 to 30. These values are plotted by the calendar date for the formation of the last circulus included in the average growth rate. 79 Figure 4.3. Chart of Alberni Inlet and Barkley Sound. Robertson Creek chinook salmon enter the region at the Somass Estuary. Areas 1 to 4 indicate the general region of capture for juvenile chinook salmon during MASS 1989 and 1990 sampling. 80 each area was then investigated. To investigate the probable location of SWE check completion on a finer scale, the mean number of circuli on the scales of juvenile chinook caught at 19 sampling sites within Alberni Inlet and Barkley Sound was determined. The mean number at each location was then charted, excluding the sampling sites where less than three chinook were captured (Figure 4.4). Probable Location of RCH Chinook by the End of the Highly-Correlated Period To determine the location of the majority of RCH chinook at the end of the highly-correlated period, the catch per unit of effort (CPUE) of chinook in Barkley Sound by sampling date was analyzed (Figure 4.5). Further evidence that juvenile chinook reside in Barkley Sound long after release was found through a comparison of the apparent migration rates of chinook caught in Barkley Sound soon after release and those caught a longer period of time following release (Figure 4.6). Apparent migration rates were determined by dividing the distance from the Somass Estuary to the capture location by the number of days from median release date to recapture. An index of the time following release was estimated as the number of circuli on the juvenile chinook scales -- chinook possess 8 or 9 circuli on their scales at release, and circulus deposition is largely a function of time (Figure 2.3). Length at the capture locations was not used as an estimate of time since release since current length of a fish is a function of time as well as prior growth over time. 4.3 RESULTS AND DISCUSSION 4.3.1 Time Period of High-Correlation between Marine Survival and Growth Using data obtained from the recapture of juvenile RCH in 1989 and 1990 it was found that the SWE check is completed in the majority of fish 34 days following release from the hatchery. At the median release date of 19 May (day 140), the SWE check is 81 Figure 4.4. Chart of Alberni Inlet and Barkley Sound. The average number of circuli on the scales of juvenile Robertson Creek chinook salmon is presented for each capture location. Marine growth begins following completion of a SWE check found to occur at approximately 13.6 ± 1.3 scale circuli. It can be seen that the number of circuli is less than 12.3 (13.6-1.3) in Alberni Inlet. In Barkley Sound, chinook caught at the majority of locations had greater than 12.3 circuli on their scales. Somass Estuary 10.0 11.2 11.5 11.2 12.8 12.3 11.5 Albemi Inlet 1 15 2 ^ 12.4 12.8 , , , Amphitrite Pt. l o ' * V 1 3 - 3 m «? 11.4 14.8 mm 1 U 0 Loudon Channel B a r k l e y S o u n d 1 2 ° 11Q Trevor Channel Imperial Eagle Channel 14 3 10.8 3 5 0 BARKLEY SOUND SOCKEYE SALMOM 4/22 6 /7 6/22 6/6 6/21 DATE 7 /6 — 1 7/21 8/6 A V E R A G E C P u E 35 30 26 20 16 10 6 B / " / / A 1988 XI987 Y \ 1990 / " ^ ^ i . 1— BARKLEY SOUND COHO SALMON \ ^ ^ 0 4/22 6/7 6/22 6/6 6/21 DATE 7 /6 7/21 8/S A V E R A Q E C P U E i z u -100-8 0 -60 40 20-0 ' c / BARKLEY SOUND / CHINOOK SALMON / 1 9 - — " ^ I T 1 r \ n 1 9 9 0 / \ j ( 1 9 8 9 , . 1 9 8 8 / / 1 9 8 7 4/22 6/7 5/22 6/6 6/21 DATE 7/6 7/21 8/5 Figure 4.5. a) Sockeye salmon b) coho salmon, and c) chinook salmon catch per unit of effort (CPUE) by date for 1987 to 1990 MASS sampling in Barkley Sound. Note that only the chinook CPUE remains high to the end of sampling (5 August). (Figure modified from Hargreaves et al. 1991). 84 5.00 ^ 4.00 3.00 •o E CC § 2.00 aj & _ 2 1.00 0.00 • 1 • t • * t 1 — • • t • $ • i : A — T A • • ~f-• — 1 — y = 9.4016e'°-1257x R2 = 0.5937 • : i 8 10 12 14 16 18 20 Number of Circuli on Scales Figure 4.6. The apparent migration rate for juvenile Chinook from the release site (Somass Estuary) to the location of capture in Barkley Sound vs. the number of circuli that they possess on their scales. Chinook with more circuli on their scales, and thus longer time since hatchery release, have slower apparent migration rates than those with few circuli added since release. 85 completed, and the marine growth begins, on average by 23 June. Using the circulus deposition rate of 6.8 dcirc"1 (Chapter 2), the approximate date at which each circulus within the first marine zone formed was estimated. The coefficient of determination (r2) between marine survival and growth rate averaged over 3 consecutive marine circuli plotted by the date of formation of the last circulus in the average is shown in Figure 4.2. Growth rates occurring between 6 July and 12 September were found to be highly correlated with survival. Within this period, extremely high significant correlation was estimated for 6 July (Av2-4: r2 = 0.93; sequential Bonferroni: P < 0.05) to 30 August (Av8-10: r2 = 0.93; sequential Bonferroni: P < 0.05), and a period of high, yet declining regression strength was found to occur between 30 August and 12 September(Av10-12: r2 = 70; P < 0.1). However, following 30 August, the relationship declined very rapidly to result in an r2 of 0.36 by 19 September, and further decreased to 0.009 by 16 October (17th marine circulus). 4.3.2 Location of Chinook throughout the Period of High Correlation between Survival and Growth Location of Chinook at the Initiation of Marine Growth The frequency histograms in Figure 4.1 indicate the number of juvenile Chinook caught in Areas 1-4 (Figure 4.3) with a given number of circuli on their scales. It can be seen that in Area 1 (Alberni Inlet), the mode is at approximately 9 circuli, and very few individuals had greater than 12.3 circuli. In Areas 2 to 4, some scales also have less than 10 circuli, however the proportion of individuals with greater than 12.3 circuli has increased. In Areas 3 and 4 a large proportion of Chinook had greater than 17 scale circuli. From these observations, it appears that the SWE check was completed in few chinook caught in Alberni Inlet (Area 1). The majority may have formed the SWE check in Areas 2-4 (Barkley Sound). In Areas 3-4, the mode of 17 suggests that several chinook caught within these regions had already laid down approximately 3 to 4 circuli of their early marine growth period. 86 To investigate the location of SWE check completion on a finer scale, the mean number of circuli present on the scales of juvenile RCH chinook captured at each sampling site within Alberni Inlet and Barkley Sound was charted (Figure 4.4). It can be seen that mean circuli numbers were less than 11.5 in Alberni Inlet, and therefore the check had begun to form on few of those chinook. Near the small inlets in Barkley Sound, near the entrance to Alberni Inlet, the averages ranged from 12.0 to 12.8. As the average number of circuli present before the end of a SWE check minus one standard deviation was found to be 12.3 (Chapter 1) it is likely that these chinook were completing or had completed formation of the SWE check. Chinook with the greatest number of circuli were caught away from the coast, towards the Broken Group Islands, as well as in Area 2 on the south bank of Trevor Channel (Figure 4.4). In summary, the majority of RCH chinook caught in Barkley Sound were either forming or completed formation of the SWE check. However, check formation had not been completed prior to leaving Alberni Inlet for the majority of chinook. Some chinook caught within Barkley Sound also had between 9 and 10 circuli on their scales (Figure 4.1). These fish may have migrated through Alberni Inlet and into Barkley Sound before initiating formation of their SWE checks. The majority were captured approximately 9 to 20 days following release, in the lower salinity waters near the mouth of Alberni Inlet and along the nearshore areas in the northeast portion of the sound (Figure 4.4). Their presence in predominantly lower salinity waters in the north-eastern portion of Barkley Sound is consistent with evidence suggesting that smaller chinook inhabit the low salinity intertidal regions of most estuaries and feed on riverine or estuarine prey (reviewed by Healey 1991; Fisher and Pearcy 1996). Therefore, prior to entering the ocean (forming their SWE check), these chinooks' growth and survival may be more closely linked to the estuarine productivity and predators than those found in the marine areas of outer Barkley Sound. However, chinook consume more profitable prey once they migrate to the more saline areas of the estuary (Fisher and Pearcy 1996). Since check formation has been found to occur following an increase in feeding (Bilton 1982), the period of check formation may occur during the transition from the estuarine to marine environment. In Chapter 2 it was estimated that the check 87 forms on average 21 to 34 days following release -- generally between 10 June and 23 June. The correspondence of the time of SWE check formation with increase in abundance of Chinook in Barkley Sound from 6 June to peak abundance by 21 June (Figure 4.5), suggests that the SWE check is formed mainly upon emigration from Alberni Inlet. Therefore, the majority of RCH Chinook were found to be in Barkley Sound at the initiation of marine growth. Those that arrived prior to SWE check formation were not likely to be influenced by the marine environment until they were forming their SWE checks. Location of Chinook at the End of the Highly-Correlated Period The MASS catch per unit effort (CPUE) data (Figure 4.5) indicate that while other salmonids reach peak abundance in Barkley Sound between May (month 5) and June (month 6), and decline in numbers rapidly to very low levels by the end of June, chinook salmon reach high abundance at a later time (by late June), and maintain these levels at least until the end of sampling (5 August). Since the abundance appears to be relatively stable at the end of the MASS sampling (5 August), it is likely that a large number of chinook will still be in Barkley Sound at least until the end of the most highly-correlated period (3Q August). Whether they will remain in Barkley Sound for two more weeks, until the end of the period of high yet declining regression strength (11 September), is still unknown. However, evidence from coded wire tag recoveries suggests that RCH chinook migrate to their final locations from southwest Vancouver Island to southeast Alaska by the end of their first winter at sea (Healey and Groot 1987). Further evidence that RCH chinook do not migrate continuously through Barkley Sound, but remain to feed for an extended period of time comes from an investigation of their apparent migration rates (Figure 4.6). Barkley Sound sockeye salmon which tend to exhibit a continuous north to northwest orientation upon ocean entry (Hartt and Dell 1986) were estimated to migrate at roughly 3 kmd"1 out of Barkley Sound (Groot and Cooke 1989). From Figure 4.6 it was found that RCH chinook captured in Barkley 88 Sound soon after release exhibited similar migration rates to the sockeye (3.5 kmd"1). However, those captured a longer time after release exhibited much lower apparent migration rates (less than 1 kmd"1). This suggests that although RCH chinook may migrate rapidly to Barkley Sound after release, once they enter Barkley Sound they do not immediately continue migration northward, but remain and feed in the region for a significant amount of time. This is consistent with evidence suggesting that juvenile chinook may have directed movements immediately upon entry into the ocean, however, their movements later in the summer appear to be less directed (Healey 1991). 4.3.3 Summary of the Location of Chinook during the Time Period of High Correlation between Survival and Growth The relationship between marine survival and growth rates was found to be very high throughout July and August (6 July and 30 August). By the time that marine growth began the majority of chinook were in Barkley Sound. RCH chinook likely remain in Barkley Sound at least until 30 August. They are found to migrate to their final ocean locations by the end of their first winter at sea (Healey and Groot 1987), however it is unknown when they initiate emigration from Barkley Sound. Since ontogenetic migration patterns tend to evolve in part due to changes in feeding opportunities (Mackas 1992), it is likely that the chinook will remain in the region of Barkley Sound until there is a significant reduction in feeding opportunity. 4.3.4 Factors Influencing Early Marine Growth and Survival Oceanic Productivity Barkley Sound is adjacent the northernmost portion of the temporally dynamic Coastal Upwelling Domain (Ware and McFarlane 1989). The period of highest regression strength between survival and growth (throughout July and August), occurs entirely within the summer upwelling period (Ware and McFarlane 1989). This period, 89 lasting from roughly May through September is the most productive time of the year. Prevailing northwesterly winds result in an Ekman transport of continental shelf surface waters offshore, and the subsequent upwelling of cold, nutrient-rich intermediate waters onto the continental margin. Especially strong upwelling occurs off the northwest tip of Vancouver Island, and the upwelled water is transported southward by a shelf break current between Cape Scott and about 48°N (Cape Flattery) (Thomson et al. 1989). The Vancouver Island Coastal Current, composed of the tidally-mixed waters of the Juan de Fuca Strait and water upwelled from the California Undercurrent (Mackas 1992), forms a confused circulation over the southern banks, further accumulating nutrient-laden water there. Under optimal conditions, these nutrient-rich waters spill into the coastal basins of the west coast of Vancouver Island, including Barkley Sound (Thomson et al. 1989). The input of these high nutrient waters to the continental shelf region adjacent to Barkley Sound results in high productivity in this region throughout the summer. Under average conditions, the total zooplankton biomass peaks in May-June (Mackas 1992). Euphausiid biomass (primarily Euphausia pacifica and Thysanoessa spinifera) peaks in May-June, declines in July and peaks again in August-September (Mackas 1992; Robinson and Ware 1994). Zooplankton are a major component in the diet of juvenile Chinook salmon. The diet composition of juvenile Chinook in relation to available prey organisms was investigated for the Washington and Oregon coast between 1980 and 1985 (Brodeur 1989). The dominant neustonic fauna were found to be euphausiids, decapod larvae, hyperiid amphipods, and larval fishes. These prey were mainly consumed according to availability, although the juvenile Chinook appeared to have a preference for larval and juvenile fishes. The period of declining regression strength between survival and growth occurs between roughly 5 September and 16 October, and thus occurs during the Fall transition from predominantly upwelling to predominantly downwelling conditions (Thomson et al. 1989). During this time, both total zooplankton and euphausiid biomass decline and reach very low levels by October-November (Mackas 1992). Since production in the Coastal Upwelling Domain decreases dramatically by the end of 90 the Fall Transition, and very low productivity is found in the winter downwelling period, the chinook may begin their northward migration in the fall. The productivity of the summer upwelling period can vary interannually as a result of many factors: the timing of the Spring Transition from predominantly down-welling to upwelling conditions; the length of the upwelling season; and the magnitude of the upwelling-favourable winds (Robinson 1994). An approximation of the amount of nutrient-rich intermediate waters upwelled onto the continental shelf is given by the Bakun Upwelling Index based on wind speed calculated from atmospheric pressure data. The index represents the number of cubic meters of water upwelled through the bottom of the Ekman layer per second per 100 m of coastline to replace surface waters displaced offshore through Ekman transport (Bakun 1973). Regression of the growth rate averaged over the highly correlated time period (July and August) on the rate of summer upwelling (Bakun index for 48° N 125° W averaged over June, July, and August) resulted in a strong dome-shaped relationship for 1982; 1985-91 (Figure 4.7: r2 = 0.68; P = 0.06). The highest growth rates were found at an intermediate rate of upwelling (36 metric tons or m3 per second per 100 m of coastline). Cury and Roy (1989) observed that recruitment of several pelagic fish stocks in Ekman type (wind-induced) upwelling regions also appeared to be a dome-shaped function of the amount of upwelling. They propose that primary production increases as the magnitude of the upwelling favourable winds increase due to the inputs of new nutrient to the euphotic zone. However, as wind velocity increases over roughly 5-6 ms"1 productivity is reduced. The reduction in productivity is proposed to result from a combination of the advection of surface production offshore and disaggregation of plankton patches as well as deepening of the mixed layer due to the increased turbulence with high winds. Thus, in years of relatively high or low levels of upwelling, the productivity of the upwelling zone is greatly reduced, resulting in a reduction in the growth rate of juvenile chinook. The link between marine survival and growth rate may result from trophic interactions governed largely by the productivity of the marine environment in the region of Barkley Sound. In low productivity years, slow growth may result in juvenile chinook spending a longer time vulnerable to a large number of predators, or becoming more vulnerable to predation due to poor physiological condition or increase in risk prone behaviour to obtain food. Slow growth also delays the time at which the juvenile chinook can switch 91 1 .\J<* -1.02 -1 -TJ 0.98 -E E, 0.96 -CD CO 0-94 -CC sz 0.92 -2 0.9 -0.88 -0.86 -• 91 y = i i • 89 • 88 ^ ^ • 9 0 ^ S -0.0003X2 + 0.0214x + 0.5966 r2 = 0.675 • sK • 82 i 86 V • \ • 85 1 15 20 25 30 35 40 45 50 Upwelling Rate (metric ton s"1100m"1) 55 60 Figure 4.7. Growth rate averaged over the 3rd to 10th marine circuli (mmd"1) regressed on the Bakun upwelling index averaged over June, July and August for 1982; 85-91. CD "co cc "co > CD 5.00 4.50 -4.00 -3.50 -3.00 2.50 H 2.00 1.50 1.00 -0.50 -0.00 -1 y = -0.0061 x + 3.3957 R2 = 0.3162 115 215 315 415 515 Hake Biomass (thousands of tonnes) 615 Figure 4.8. Linear regression of marine survival rate (%) of Chinook on hake biomass (thousands of tonnes) for 1980-1992. 92 to a primarily piscivorous diet, further reducing growth rates. However, growth rates appear to be an index of the productivity of the marine environment of chinook nursery areas. Survival may be affected by changes in other factors varying with ocean productivity. Predation Intensity During low productivity years, mortality may increase due to changes in predation intensity and predator abundance varying concurrently with ocean conditions linked to productivity. The highest amount of predation on chinook shortly after release from the RCH has been found to be due to Pacific hake when mackerel are not present (Hargreaves et al. 1991). Although Pacific hake consume herring and chinook as well as other fish species, euphausiids normally make up the majority of their diet (Tanasichuk et al. 1991; Ware and McFarlane 1995). Interannual variability in the proportion of herring and euphausiids in the diet of Pacific hake has been suggested to result from changes in the availability of euphausiids (Tanasichuk et al. 1991). In low productivity years, when euphausiids are relatively scarce, herring as well as juvenile chinook may experience higher predation intensity by hake. Furthermore, it has been found that the absolute biomass of Pacific hake migrating into the upwelling region adjacent Vancouver Island may be linked to ocean conditions varying with the intensity of upwelling. During poor upwelling years, the volume of cold intermediate water upwelled onto the continental shelf is reduced, resulting in higher than average sea surface temperatures. Following a 1 degree Celsius increase above the long term mean June sea surface temperature at Amphitrite Point (Figure 4.4) it was found that the biomass of migratory Pacific hake in the upwelling region adjacent Vancouver Island doubled from 179 0001 to 353 5001 (Ware and McFarlane 1995). A significant linear relationship between chinook survival and hake biomass was found for 1980-1992 (Figure 4.8: r2 = 0.32; P = 0.045), suggesting that the absolute biomass of hake may also influence survival of RCH chinook. However, the biomass of hake may not be a particularly good index of the intensity of predation on juvenile chinook as the arrival time (Hargreaves et al. 1991) and the location of the main body of the migratory stock of 93 o.uu -4.50-4.00-^ 3.50 -£ 3.00 -g. 2.50 -g 2.00 -E 1-50-J3 1.00-0.50-0.00-• 88 • 90 .*8?1 84 «83 *__ 92 • 89 • 87 • 80 #85 • 86 • 82 , , 15 20 25 30 35 Upwelling Rate (m3 s"1 (100m)*1) 40 Figure 4.9. Marine survival rate of Chinook vs. the Bakun upwelling index averaged over April, May, June, July, and August) for 1980-1992. 5.00 4.50 _ 4"0 0 " g, 3.50 -i§ 3.00 -\ CO ? 2.50 CO > 2.oo ^ I 1 - 5 0 1.00 0.50 H 0.00 15.0 89 y = -0.0306X2 + 1.5752x r2 = 0.5975 16.908 82 20.0 25.0 30.0 35.0 Upwelling Rate (m3 s-1 (100m)-1) 40.0 Figure 4.10. Marine survival rate of chinook vs. the Bakun upwelling index averaged over April, May, June, July, and August) for 1980-82; 85-91. 94 hake also vary interannually - thus providing varying degrees of overlap, both temporally and spatially, with the juvenile Chinook. The abundance and predation intensity by hake varies with ocean conditions linked to productivity. Since hake biomass increases with increased SST, and SST decreases with the magnitude of upwelling, in years when growth rates are low due to lower than optimal summer upwelling, the abundance of hake predators is high. In low upwelling (low productivity) years, the larger biomass of hake is competing for a very limited food supply, thus likely exerting more predation pressure on the juvenile salmonids. High temperature, which increases the standard metabolic rate of poikilotherms, further stimulates predation in warmer years (Ware 1975). In high upwelling years, the growth is again limited by oceanic productivity (Figure 4.7), however, the colder temperatures suggest that the biomass of hake and predation intensity by hake should be low. Thus, there would be expected to be a slower decline in survival as the upwelling increased beyond the optimal level. Using a slightly longer time series of data (1980-1992), the marine survival rate of Chinook salmon was plotted against the rate of upwelling averaged over the entire summer (Average of April, May, June, July, August, and September Bakun Indices) (Figure 4.9). A dome-shaped relationship is apparent, although a lower than expected survival rate is found for ocean entry years in which a large number of mackerel are known to have migrated into the region of Barkley Sound (1983, 1984, and 1992: discussed in Chapter 3). Removing these years from the regression produced a strong dome-shaped relationship between survival and upwelling rate (Figure 4.10: r2 = 0.60; P = 0.069). The poor fit in Figure 4.9 for certain years (1983, 1984, 1992) can possibly be explained by predation from mackerel migrating intermittently into the upwelling region. In these years, low survival rate is found independent of conditions for growth. Holtby et al. (1990) conducted a similar study on wild coho salmon originating in Carnation Creek over a 17 year period between 1971 and 1987. These coho enter the ocean within Barkley Sound near the mouth of Alberni Inlet (Figure 4.4). The relationships found between marine survival and smolt lengths and growth rates are 95 qualitatively similar to the results for RCH chinook. Holtby et al. (1990) found that no consistent survival advantage was conferred by large smolt size. However, they did find that larger smolts survived at a higher rate in years of low survival. Marine survival of age 1+ coho was strongly correlated with the average growth rate estimated from survivors' scales for the first 60 days following ocean entry. Since the migration of smolts from the estuary into Barkley Sound peaked on average between late April and early May, the highly-correlated growth period occurred during roughly May to early July. Coho abundance is known to decline to low numbers in Barkley Sound by late June (Figure 4.5). Following emigration from Barkley Sound, the coho are observed to migrate northward, reaching the northwest tip of Vancouver Island by roughly August (Holtby et al. 1990). Therefore, during the period of high correlation between marine survival and growth rates (May through early July), the coho were migrating through Barkley Sound and northward along the west coast of Vancouver Island between Barkley Sound and the northern tip of Vancouver Island (Figure 4.3 inset). Both growth rates and marine survival were found to be correlated with oceanographic conditions indicative of productivity at the north end of Vancouver Island. Highest correlations were found with the average sea surface salinity (SSS) recorded at Kains Island Light (51 °N 131°W) for June, July, and August (Holtby et al. 1990). Regressions involving the Bakun upwelling index were not found to be significant. However, this may result from Holtby et al.'s (1990) restriction to testing linear functions - a quadratic function best described the relationship between survival and upwelling for RCH chinook (Figures 4.7 and 4.10). Although Carnation Creek coho reside in Barkley Sound for a shorter period of time than RCH chinook, the growth rates found to be highly-correlated with marine survival also occur within a region influenced by the summer productivity of the Coastal Upwelling Domain (extending northwards to roughly the tip of Vancouver Island). Highly-correlated growth rates for coho occurred on average 2 months before that estimated for chinook (May-early July compared to July-August). The summer upwelling season begins on average by May, and is terminated by early September (Dr. D. Ware, FOC, Pacific Biological Station, Nanaimo, BC, pers. comm.). Thus, 96 growth rates correlated with marine survival for both chinook and coho occur within the summer upwelling season. Processes occurring within the summer upwelling season may also help explain the covariance in the marine survival of several stocks of sockeye originating from the region of Barkley Sound observed by Hyatt et al. (1989). These salmonids enter the ocean at a size similar to both chinook and coho, and migrate through Barkley Sound at roughly the same time as coho (Figure 4.5). 97 LITERATURE CITED Anderson, J.T. 1988. A review of size dependent survival during the pre-recruit stages of fishes in relation to recruitment. J. Northw. Atl. Fish. Sci. 8: 55-66. Ashton, H. J., V. Haist, and D.M. Ware. 1985. Observations on abundance and diet of Pacific mackerel {Scomber japonicus) caught off the west coast of Vancouver Island, September 1984. Can Tech. Rep. Fish. Aquat. Sci. No. 1394 Bakun, A. 1973. Coastal upwelling indices, west coast of North America, 1964-71. U.S. Dept. Commer., NOAATech. Rep. NMFSS SSSRF-671, 103 p. Barber, W.E., and R.J. Walker. 1988. Circuli spacing and annulus formation: is there more than meets the eye? The case for sockeye salmon, Oncorhynchus nerka. J. Fish Biol. 32: 237-245. Bax, N.J. 1983. Early marine mortality of marked juvenile chum salmon (Oncorhynchus keta) released into Hood Canal, Puget Sound, Washington, in 1980. Can. J. Fish. Aquat. Sci. 40: 426-435. Bilton, H.T. 1975. Factors influencing the formation of scale characters .Int.North Pac.Fish.Comm.Bull. 32:102-108. Bilton, R.R., and S.A.M. Ludwig. 1966. Times of annulus formation on scales of sockeye, pink, and chum salmon in the Gulf of Alaska. J. Fish. Res. Board Can. 23: 1403-1410. Bilton, H.T, and G.L Robins. 1971. Response of young sockeye salmon (Oncorhynchus nerka) to prolonged periods of starvation. J. Fish. Res. Bd. Can. 28: 1757-1761. 98 Bilton, H.T., D.F. Alderdice, and J.T. Schnute. 1982. Influence of time and size at release of juvenile coho salmon {Oncorhynchus kisutch) on returns at maturity. Can. J. Fish. Aquat. Sci. 39: 426-447. Brodeur, R.D. 1989. Neustonic feeding by juvenile salmonids in coastal waters of the Northeast Pacific. C. J. Zool. 67: 1995-2007. Clutter, R.I., and L.E. Whitesel. 1956. Collection and interpretation of sockeye salmon scales. Int. Pac. Salmon Fish. Comm. Bull. 9: 159 p. Cury, P., and C. Roy. 1989. Optimal environmental window for pelagic fish recruitment success in upwelling areas. Can. J. Fish. Aquat. Sci. 46: 670-680. Cushing, D.H. 1975. Biology of fishes in the pelagic community pp. 317-340 In Cushing and Walsh [eds.] The ecology of the seas. W.B. Saunders Co., Toronto, Canada. Doyle, R.W., A.J. Talbot, and R.R. Nicholas. 1987. Statistical interrelation of length, growth, and scale circulus spacing: appraisal of a growth rate estimator for fish. Can. J. Fish. Aquat. Sci. 44: 1520-1528. Fagerlund, U.H.M., J.R. McBride, and I.V. Williams. 1995. Stress and tolerance. In Physiological ecology of Pacific salmon. C. Groot et al. [eds]. UBC Press, Vancouver, pp. 461-503. Fisher, J.P. and W.G. Pearcy. 1990. Spacing of scale circuli versus growth rate in young coho salmon. Fish. Bull. 88: 637-643. 99 Fisher, J.P. and W.G. Pearcy. 1988. Growth of juvenile coho salmon (Oncorhynchus kisutch) off Oregon and Washington, USA, in years of differing coastal upwelling. Can. J. Fish. Aquat. Sci. 45: 1036-1044. Fisher, J.P., and W.G. Pearcy. 1996. Dietary overlap of juvenile fall- and spring- run chinook salmon, Oncorhynchus tshawytscha, in Coos Bay, Oregon. Fish. Bull. 95: 25-38. Francis, R.I.C.C. 1990. Back-calculation offish lengths: a critical review. J. Fish Biol. 36: 883-902. Fry, D.H. Jr. 1936. A preliminary summary of the life history of the Pacific mackerel {Pneumatophorus diego). Calif. Fish. Game 22(1): 30-36. Fukuwaka, M., and M. Kaeriyama. 1997. Scale analyses to estimate somatic growth in sockeye salmon, Oncorhynchus nerka. Can. J. Fish. Aquat. Sci. 54: 631-636. Groot, C , and K. Cooke. 1989. Sockeye smolt migration in Barkley Sound pp. 3-6 In The marine survival of salmon program, annual progress report, 1988. Available from T.D. Beacham (program coordinator), Biological Sciences Branch, Department of Fisheries and Oceans, Pacific Biological Station, Nanaimo, BC V9R 5K6. Hargreaves, N.B. and R.M. Hungar. 1994. Robertson Creek chinook assessment and forecast for 1994. Part B: Early marine mortality. PSARC Working Paper S94 -07. 51p. Hargreaves, N.B., and R.J. LeBrasseur. 1986. Size selectivity of coho {Oncorhynchus kisutch) on juvenile chum (O. keta) salmon. Can. J. Fish. Aquat Sci. 43: 581-586. 100 Hargreaves, B., B. Patten, B. Hungar, and T. Carter. 1991. Abundance, migrations, and predation mortality of juvenile salmon in Alberni Inlet and Barkley Sound, B.C. in 1990. In T.D. Beacham [Program coordinator] The marine survival of salmon program: Program outline and investigators summaries for 1990/91. 63 p. Hartt, A.C. and M.B. Dell. 1986. Early oceanic migrations and growth of juvenile pacific salmon and steelhead trout. Int. North Pac. Comm. Bull. 46: 105 p. Healey, M.C. 1980. The ecology of juvenile salmon in Georgia Strait, British Columbia, p. 203-229. In: W.J. McNeil and D.C. Himsworth [eds]. Salmonid ecosystems of the North Pacific. Oregon State University Press, Corvallis, OR. Healey, M.C. 1982. Timing and relative intensity of size-selective mortality of juvenile chum salmon {Oncorhynchus keta) during early sea life. Can. J. Fish. Aquat. Sci 36: 952-957. Healey, M.C. 1991. Life history of chinook salmon {Oncorhynchus tshawytscha). pp 311 393. In C. Groot and L. Margolis [eds.] Pacific salmon life histories. UBC Press, Vancouver: Healey, M.C. and C. Groot. 1987. Marine migration and orientation of ocean-type chinook and sockeye salmon. Am. Fish. Soc. Symp. 1: 298-312. Henderson, M.A., and A.J. Cass. 1991. Effects of smolt size on smolt-to-adult survival for Chilko Lake sockeye salmon {Oncorhynchus nerka). Can. J. Fish. Aquat. Sci. 48: 988-994. Holtby, L.B., B.C. Anderson, and R.K. Kadowaki. 1990. Importance of smolt size and early ocean growth to interannual variability in marine survival of coho salmon {Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 47(11): 2181-2194. 101 Hyatt, K.D., G.J. Steer, DP. Rankin, R. Traber, and I. Miki. 1989. Sockeye salmon recruitment variations. Pp. 14-21 In The marine survival of salmon program, annual progress report, 1988. Available from T.D. Beacham (program coordinator), Biological Sciences Branch, Department of Fisheries and Oceans, Pacific Biological Station, Nanaimo, BC V9R 5K6. Lorenzen, K. 1996. The relationship between body weight and natural mortality in juvenile and adult fish: a comparison of natural ecosystems and aquaculture. J. Fish. Biol. 49: 627-647. Mackas, D.L. 1992. Seasonal cycle of zooplankton off southwestern British Columbia: 1979-89. Can. J. Fish. Aquat. Sci. 49: 903-921. Major, R.L., and D.R. Craddock. 1962. Marking sockeye salmon scales by short periods of starvation. U.S. Fish Wildl. Serv. Spec. Sci. Rep. Fish. 416: 12 p. Matthews, S.B., and R. Buckley. 1976. Marine mortality of Puget Sound coho salmon (Oncorhynchus kisutch). J. Fish. Res. Board. Can. 33: 1677-1684. Matthews, S.B., and Y. Ishida. 1989. Survival, ocean growth, and ocean distribution of differentially timed releases of hatchery coho salmon (Oncorhynchus kisutch). Can. J. Fish. Aquat. Sci. 46: 1216-1226. Mysak, LA. 1986. El Nino, interannual variability and fisheries in the northeast Pacific Ocean. Can. J. Fish. Aquat. Sci. 43: 464-497. Nicieza, A.G., and N.B. Metcalfe. 1997. Growth compensation in juvenile Atlantic salmon: responses to depressed temperature and food availability. Ecology. 78(8): 2385-2400. 102 Nickelson, T.E. 1986. Influence of upwelling, ocean temperature, and smolt abundance on marine survival of coho salmon (Oncorhynchus kisutch) in the Oregon Production Area. Can. J. Fish. Aquat. Sci. 43: 527-535. Neilson, J.D. and G.H. Geen. 1986. First-year growth rate of Sixes River Chinook salmon as inferred from otoliths: Effects on mortality and age at maturity. Trans. Am. Fish. Soc. 115(1): 28-33. Neilson, J.D., G.H. Geen, and D. Bottom. 1985. Estuarine growth of Chinook salmon (Oncorhynchus tshawytscha) as inferred from otolith microstructure. Can. J. Fish. Aquat. Sci. 42: 899-908. Parker, R.R. 1968. Marine mortality schedules of pink salmon of the Bella Coola River central British Columbia. J. Fish. Res. Board Can. 25: 757-794. Parker, R.R. 1971 Size selective predation among juvenile salmonid fishes in a British Columbia inlet. J. Fish. Res. Board Can. 28: 1503-1510. Peterman, R.M. 1987. Review of the components of recruitment of Pacific salmon. Am. Fish. Soc. Symp. 1: 417-429. Peterson. I., and J.S. Wroblewski. 1984. Mortality rate of fishes in the pelagic ecosystem. Can. J. Fish. Aquat. Sci. 41: 1117-1120. Post, J.R., and D.O. Evans. 1989. Size-dependent overwinter mortality of young-of-the-year yellow perch (Perca flavescens): laboratory, In situ enclosure, and field experiments. Can. J. Fish. Aquat. Sci. 46: 1958-1968. Ricker, W.E. 1973. Linear regression in fishery research. J. Fish. Res. Board Can. 30: 409-434. 103 Ricker, W.E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board Can. 191: 382 p. Ricker, W.E. 1992. Back-calculation of fish lengths based on proportionality between scale and length increments. Ca. J. Fish. Aquat. Sci. 49: 1018-1026. Rice, W.R. 1988. Analyzing tables of statistical tests. Evolution. 43: 223-225. Robinson, C.L.K. 1994. The influence of ocean climate on coastal plankton and fish production. Fish. Oceanogr. 3(3): 59-171. Robinson, C.L.K., and D.M. Ware. 1994 Modelling pelagic fish and plankton trophodynamics off southwestern Vancouver Island, British Columbia. Can. J. Fish. Aquat. Sci. 51: 1737-1751. Scarnecchia, D.L. 1981. Effects of streamflow and upwelling on yield of wild coho salmon {Oncorhynchus kisutch) in Oregon. Can. J. Fish. Aquat. Sci. 38: 471-475. Sogard, S.M. 1997. Size-selective mortality in the juvenile stage of teleost fishes: A review. Bull. Mar. Sci. 60(3): 1129-1157 Sokal, R.R., and F.J. Rohlf. 1995. Biometry Third Edition. W.H. Freeman and Company, New York. Tanasichuk, R.W., D.M. Ware, W. Shaw, and G.A McFarlane. 1991. Variations in diet, daily ration, and feeding periodicity of Pacific hake {Merluccius productus) and spiny dogfish {Squalus acanthias) off the lower west coast of Vancouver Island. Can. J. Fish. Aquat. Sci. 48:2118-2128. 104 Thomson, R.E., B.M. Hickey, and RH. LeBlond. 1989. The Vancouver Island Coastal Current: fisheries barrier and conduit, pp. 265-296. In R.J. Beamish and G.A. McFarlane [eds.] Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models. Can. Spec. Publ. Fish. Aquat. Sci. 108. Ward, B.R., and Slaney, A.R. 1988. Life history and smolt-to-adult survival of Keogh River steelhead trout (Salmon gairdneri) and the relationship to smolt size. Can. J. Fish. Aquat. Sci. 45: 1110-1145. Ward, B.R., RA. Slaney, A.R. Facchin, and R.W. Land. 1989.Size-biased survival in steelhead trout (Oncorhynchus mykiss): back-calculated lengths from adults' scales compared to migrating smolts at the Keogh River, British Columbia. Can. J. Fish. Aquat. Sci. 46: 1853-1858. Ware, D.M. 1975. Relation between egg size, growth, and natural mortality of larval fish. J. Fish. Res. Board Can. 32: 2503-2512. Ware, D.M., and G.A. McFarlane. 1986. Relative impact of Pacific hake, sablefish, and Pacific cod on west coast Vancouver Island herring stocks. INPFC No. 47. 67-78. Ware, D.M., and G.A. McFarlane. 1989. Fisheries production domains in the Northeast Pacific Ocean, p.359-379. In R.J. Beamish and G.A. McFarlane [ed.] Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models. Can. Spec. Publ. Fish. Aquat. Sci. 108. Ware, D.M., and McFarlane. 1995. Climate-induced changes in Pacific Hake (Merluccius productus) abundance and pelagic community interactions in the 105 Vancouver Island upwelling system. In R.J. Beamish [ed.] Climate change and northern fish populations. Can. Spec. Publ. Fish Aquat. Sci. 121. Weatherly, A.H., and H.S. Gill. 1995. Growth InC. Groot et al. [eds.] Physiological ecology of Pacific salmon. UBC Press, Vancouver. West, C.J., and P.A. Larkin. 1987. Evidence for size-selective mortality of juvenile sockeye salmon (Oncorhynchus nerka) in Babine Lake, British Columbia. Can. J. Fish. Aquat. Sci. 44: 712-721. 106 APPENDIX I. INVESTIGATION INTO POSSIBLE SEXUAL DIMORPHISM IN SCALE MEASUREMENTS FOR AGE 0.3 ROBERTSON CREEK CHINOOK SALMON 1.1 INTRODUCTION There is great plasticity in the age at return of chinook to the Robertson Creek hatchery: observed ages range from 0.1 to 0.6. It has been suggested that many factors including early ocean growth rates may affect age at return for individual chinook (Healey 1991). However, factors such as early ocean growth rate may affect males and females differently. There is stronger selection for female chinook to grow to a large size before spawning due to fecundity- and egg size-size relations (Healey 1991), and they therefore tend to represent a larger proportion of the older age classes. As well, it has been determined that only male chinook with larger size at ocean entry and faster growth rates tend to return as jacks after 1 year in the ocean (Healey 1991). Thus, it is possible that back-calculated early marine growth rates and lengths may differ for the two sexes returning after three years in the ocean. Statistical analyses were performed to determined whether the first year scale measures of age 0.3 RCH male and female chinook differ significantly. I.2 METHODS AND MATERIALS Female scales were digitized preferentially for this study. However, the sex associated with individual scale samples was not indicated on the scale cards for some years (1983, 1986, 1988, 1989). The data for these years was grouped as "sex unspecified" or U. Once the sex data for these scales was obtained, it was possible to perform blind tests for sexual dimorphism in first year scale features. T-tests on scale features estimated for age 0.3 chinook from the 1983,1986, 1988, and 1989 ocean entry years grouped by sex were performed. The sequential Bonferroni method for multiple, simultaneous tests was used to assess significance at P = 0.05. 107 1.3 RESULTS The results for each year are presented in the following tables (Tables 1.1 to I.4). Significant differences (P = 0.05) are indicated by an asterisk (*). A table that summarizes the significant different features by year is provided at the end (Table I.5). Table 1.1. T-test results for 1983 ocean entry year first year scale features grouped by sex. 45 female and 37 male samples were included in the tests. Scale feature FWRAD FWCIRC MAR1 WIDTH MAR 1 CIRC CIRC1 MAR1RAD Av. 2to6 Av.2to10 Av. 3to4 CIRC30 Female Av. 39.22 12.89 98.05 32.40 2.34 137.27 2.74 2.83 2.84 2.88 Female SD 4.38 1.25 11.85 2.46 0.36 12.65 0.36 0.34 0.52 0.83 Male Av. 38.85 13.16 90.71 32.08 2.27 129.57 2.56 2.58 2.63 2.66 Male SD 4.41 1.52 11.25 3.17 0.45 11.92 0.36 0.31 0.51 0.62 Separate Var. Prob. 0.709 0.383 0.005* 0.618 0.506 0.006* 0.024* 0.001* 0.077 0.212 Pooled Var. Prob. 0.709 0.374 0.005* 0.609 0.496 0.006* 0.024* 0.001* 0.078 0.236 Table I.2. T-test results for 1986 ocean entry year first year scale features grouped by sex. 34 female and 19 male samples were included in the tests. Scale feature FWRAD FWCIRC MAR1 WIDTH MAR1CIRC CIRC1 MAR1RAD Av. 2to6 Av.2to10 Av. 3to4 CIRC30 Female Av. 40.43 13.91 93.08 30.84 2.43 133.48 2.82 2.89 2.83 2.99 Female SD 3.34 1.08 9.96 1.80 0.55 11.52 0.50 0.41 0.67 0.82 Male Av. 41.56 14.21 90.89 30.44 2.18 132.47 2.68 2.75 2.79 2.95 Male SD 4.24 0.78 8.08 2.28 0.33 10.85 0.53 0.46 0.71 0.66 Separate Var. Prob. 0.325 0.255 0.405 0.528 0.043* 0.760 0.354 0.263 0.831 0.882 Pooled Var. Prob. 0.289 0.297 0.432 0.498 0.078* 0.763 0.342 0.246 0.828 0.891 108 Table 1.3. T-test results for 1988 ocean entry year first year scale features grouped by sex. 36 female and 21 male samples were included in the tests. Scale feature Female Female Male Av. Male SD Separate Pooled Av. SD Var. Prob. Var. Prob. FWRAD FWCIRC MAR1 WIDTH MAR 1 CIRC CIRC1 MAR 1 RAD Av. 2to6 Av. 2to10 Av. 3to4 CIRC30 40.70 13.17 98.26 30.84 2.45 136.05 3.16 3.21 3.15 3.64 3.24 0.94 15.09 2.71 0.74 23.48 0.48 0.43 0.56 1.19 40.29 13.71 93.46 30.32 2.37 134.03 3.00 3.08 3.05 3.17 3.22 0.96 17.85 4.37 0.59 19.30 0.36 0.31 0.39 0.73 0.646 0.042* 0.333 0.639 0.672 0.740 0.172 0.174 0.417 0.153 0.646 0.040* 0.310 0.596 0.689 0.753 0.203 0.212 0.460 0.220 Table I.4. T-test results for 1989 ocean entry year first year scale features grouped by sex. 46 female and 18 male samples were included Scale feature FWRAD FWCIRC MAR1 WIDTH MAR 1 CIRC CIRC1 MAR 1 RAD Av. 2to6 Av. 2to10 Av. 3to4 CIRC30 Female Female Av. SD 40.85 4.09 13.48 0.96 96.98 12.73 30.76 3.37 2.24 0.49 137.84 14.07 3.29 0.28 3.34 0.30 3.20 0.36 2.93 0.83 Male Av. 39.32 13.56 89.03 30.22 2.34 128.35 3.13 3.14 3.25 2.70 in the tests. Male SD 3.60 1.15 10.37 2.98 0.75 10.01 0.45 0.38 0.77 0.91 Separate Var. Prob. 0.151 0.802 0.014* 0.541 0.604 0.004* 0.171 0.060* 0.798 0.469 Pooled Var. Prob. 0.171 0.487 0.022* 0.560 0.531 0.012* 0.090 0.033* 0.728 0.443 109 Table 1.5. Summary of the significantly different t-test results for the 1983, 1986,1988, and 1989 ocean entry years. The sex with the significantly greater value is provided. Year 1983 1986 1988 1989 Significantly Different Feature MAR1 Width MAR1 Radius Av.2to6 Av2to10 CIRC1 FWCIRC MAR1 Width MAR1 Radius Av2to10 Separate Var. Prob. 0.005* 0.006* 0.024* 0.001* 0.043* 0.042* 0.014* 0.004* 0.060* Pooled Var. Prob. 0.005* 0.006* 0.024* 0.001* 0.078 0.040* 0.022* 0.012* 0.033* Sex with the Higher Value F F F F F M F F F For the 1983 ocean entry year, the most significant difference between sexes was found for the average intercircular distance between the 2nd and 10th marine circulus (P 0.05) (Table 1.1). This suggests that early ocean growth rates differed significantly between females and males for this year. However, as the length of time in the marine environment considered is reduced, the degree of difference decreases: the average of the 2nd to 6th was still significantly different (P = 0.024), however, the average of the 3rd to 4th was not (P = 0.078). The width of the first marine zone (P = 0.005) and the scale radius to the end of the first marine annulus (P = 0.006) were also found to differ significantly in 1983. In contrast to the 1983 results, the t-test results indicate that there was no significant sexual dimorphism in any scale characteristics except the first marine circulus width for the 1986 ocean entry year (Table I.2). And for the 1988 ocean entry year, only the number of circuli present before the end of the salt water entry check was found to be significantly different between the sexes (Table I.3). For the 1989 ocean entry year, however it can be seen that 3 of the 10 measures differed significantly between the sexes (to P = 0.05) (Table I.4). The largest difference was seen for the radius to the end of the first marine annulus (P = 0.004). The other two features include the width of the first marine zone (0.014), and the intercircular distance of the 2nd to 10th marine circuli (P = 0.033). The results are similar for the 1983 and 1989 ocean entry years. This is interesting since 1983 was an extremely low no growth rate and low survival year, and 1989 was an extremely high growth and survival year. I.4 DISCUSSION The results indicate that for some years, the first year length and growth indices differed between the sexes. The same characteristics were not found to differ significantly for the years considered (Table 1.5). For all years, however, there was found to be significant differences in either the early marine growth indices or the indices describing the region before the end of the SWE check. Growth differences showed up as early as the 6th marine circulus, however, it was found that the intercircular distance averaged over the 3rd to 4th marine circuli did not show significant differences between females and males in any of the years considered (Table 1.5). The results indicate that there are significant differences in scale measurements between males and females. However, this does not necessarily indicate that sexual dimorphism exists in the first year lengths and growth rates for Robertson Creek Chinook salmon. What has been determined in this investigation is that some indices of length and growth differ significantly between males and females that returned to spawn at age 0.3. It has been found that the fastest-growing males of a cohort tend to return to the hatchery as jacks or age 0.2 (reviewed by Healey 1991) whereas female RCH chinook predominantly delay spawning until age 0.3 or 0.4. Since the fastest-growing males of a cohort are removed from the population by age 0.3 and few female chinook have returned to spawn by that time, the measures determined from the scales of age 0.3 males would likely be smaller than for the females. This is supported by the results in the summary table (Table 1.5). 8 out of 9 of the significantly different features were found to be larger for females than for males. Ward et al (1989) found similar results as discussed here. The mean back-calculated smolt lengths determined for all age classes of a cohort combined did not differ between males and females. However, the mean back-calculated smolt length determined for separate age classes of steelhead did differ between the sexes. This i n suggests that the phenomenon may apply to other anadromous salmonid species. Studies using back-calculation to determine lengths or growth rates of salmonids at an earlier age may benefit from separating measurements made for females and males. 112 APPENDIX II. CONVERSION OF JUVENILE CHINOOK SCALE MEASUREMENTS FROM THE LONGEST AXIS TO THE STANDARD AXIS 20° VENTRAL TO THE LONGEST AXIS 11.1 INTRODUCTION During the MASS program, juvenile Robertson Creek chinook salmon were captured at sampling sites throughout Barkley Sound and Alberni Inlet (Chapter 1: Figure 1.2). Scales taken from chinook captured in 1988 and 1989 were digitized by staff in the fish ageing lab at the Pacific Biological Station (FOC, Science Branch, Nanaimo, BC). The radius measurements obtained from these scales are necessary for the construction of a regression of forklength on scale radius required for the back-calculation of length at age from adult chinook scales (Chapter 2). However, the juvenile chinook scales were measured along an axis different from the one used to measure the adult chinook scales. According to convention, adult chinook scales were measured along the radial line that extends 20 ° ventro-anteriorly from the longest axis of the scale (Chapter 1: Figure 1.1). However, as the juvenile chinook scales were approximately symmetrical about the longest axis, it is not possible to determine precisely which is the ventral and which is the dorsal portion of the scale. Thus, the juvenile scales were measured along the longest scale axis. To make use of the juvenile scale measurements, they had to be converted to the corresponding 20° ventral axis values. The early scale growth does not change once it is formed, and thus later, asymmetrical scale growth on adult scales allows one to identify the ventral axis of the nearly symmetrical early growth region. The relationship between the longest axis measurement and 20° ventral measurement taken from several adult chinook scales provided the conversion factor. II.2 METHODS AND MATERIALS Assuming that the shape of the scale during early life would not differ between years, sexes or spawning age, the conversion was constructed using scale measurements taken from the scales of age 0.3, sex unspecified Robertson Creek chinook salmon that 113 entered the ocean in 1986. The distance from the scale origin to the end of the salt water entry check and the to end of the first marine annulus were measured (Chapter 1: Figure 1.1) along both the longest and 20° ventral axes of 36 scales. The measurements for the two axes were then plotted and the relationship between them was determined by least squares regression. The slope of the regression equation provided the conversion from longest axis measurements to those, 20° ventral to the longest axis. II.3 RESULTS Three graphs are presented which explore the conversion using the different zones measured. The first regression (Figure 11.1) using the FW radius measurements (n = 35) resulted in a slope of 0.906 (Y = 0.9062(x); r2 = 0.66; P < 0.001). This indicates that the 20° ventral axis (20DVA) measurements are about 90.6 % of the long axis (LA) measurements. The second graph (Figure II.2) indicates that in the first marine zone (n = 29), the 20DVA measures are 94.4 % of the LA measures (Y = 0.9437(x); r2 = 0.71; P < 0.001). The final graph (Figure II.3) indicates that when the fresh water radius and the first marine zone width are added together to form the radius to the end of the first marine annulus (n = 29), the 20DVA is approximately 93.29 % of the LA measures (Y = 0.9329(x); r2 = 0.75; P < 0.001). The 20DVA measures are a smaller proportion of the LA measures for early growth of the scale (i.e. the FW zone: 90.6 %) than towards the end of the first year (first marine width: 94.4 %). This results from the change in shape from oval to more round circuli as the scale grows (Figure 1.1). II.4 DISCUSSION The conversion was to be used for measurements taken from juvenile Chinook salmon caught in the nearshore waters of Barkley Sound and Alberni Inlet shortly after hatchery release (<3 months). Juvenile chinook caught in 1989 and 1990 had an average of 14.9 circuli on their scales. The minimum number of circuli was 7 and the maximum was 19. According to the number of circuli present on the scales of adult 114 Chinook at the end of the SWE check (Table 1.2: average = 13.6 + 1.3), the juvenile Chinook scales possessed a few less or a few more circuli than the number present to the end of the check. Therefore, the conversion would thus be required for the fresh water zone of the scale, and a few circuli into the saltwater region. The conversion factor for the FW zone alone, 0.9062 (Figure 11.1: ? = 0.66; P < 0.001), was found to be the most appropriate conversion factor to use. The juvenile Chinook radius measurements were converted from the longest axis measurement to the 20° ventral measurement by multiplying by the conversion factor (0.9062). 115 30 35 40 45 50 55 Long Axis Figure 11.1. Regression of 20 degree ventral measurements on longest axis measurements of the scale radius to the end of the SWE check (FW radius) for 1982 and 1986 ocean entry years. 70 80 90 100 110 120 Long Axis Figure II.2. Regression of 20 degree ventral axis measurements on the longest axis measurements for the width of the first marine scale zone for 1982 and 1986 ocean entry years. 116 CO *—• c CD > CD CD i _ CO CD Q o CM 155-145-135-125-115-105-y = 0.9329x r2 = 0.7506 0 0 — i - — i 1 r -O „ ^ ^ 0 — 1 110 120 130 140 Long Axis 150 160 170 Figure II.3. Regression of 20 degree ventral measurements on longest axis measurements for the scale radius to the end of the first marine year (Marl radius), for 1982 and 1986 ocean entry years. 117 APPENDIX III. TABLES 118 Table 111.1. Descriptive statistics for the scale radius to the end of the salt water entry check (units: 0.01 mm). STAT Mean Standard Error Median Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) Table III.2. Descriptive STAT Mean Standard Error Median Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 1982 36.83 0.63 36.83 4.20 17.67 -0.63 0.09 15.93 28.84 44.77 1657.53 45 1.263 1983 39.22 0.65 38.40 4.38 19.14 -0.42 0.32 18.56 30.78 49.34 1764.82 '45 1.315 1985 39.13 0.39 39.59 3.03 9.20 -0.29 -0.05 13.79 32.08 45.87 2347.91 60 0.784 1986 41.18 0.54 41.05 3.78 14.31 0.06 0.34 17.74 33.21 50.95 2058.76 50 1.075 1987 42.62 0.36 42.22 3.07 9.40 0.06 -0.21 15.40 34.61 50.01 3068.99 72 0.720 1988 41.72 0.48 41.55 3.40 11.56 0.56 -0.56 15.45 33.03 48.48 2127.80 51 0.956 1989 40.84 0.60 40.74 4.09 16.75 0.71 0.59 20.07 32.82 52.89 1878.86 46 1.215 statistics for the number of scale circuli to the end of the salt water entry check. 1982 12.67 0.185 12 1.243 1.545 0.143 0.825 5 11 16 570 45 0.373 1983 12.89 0.186 13 1.247 1.556 -0.375 0.219 5 11 16 580 45 0.375 1985 13.15 0.121 13 0.936 0.875 -0.410 0.205 4 11 15 789 60 0.242 1986 14.02 0.147 14 1.040 1.081 0.506 -0.608 5 11 16 701 50 0.296 1987 13.69 0.120 14 1.016 1.032 0.035 0.155 5 11 16 986 72 0.239 1988 13.31 0.139 13 0.990 0.980 -0.234 -0.294 4 11 15 679 51 0.278 1989 13.48 0.142 14 0.960 0.922 -0.919 -0.329 3 12 15 620 46 0.285 1990 40.00 0.50 39.85 3.93 15.45 -0.06 0.15 19.71 30.71 50.42 2440.28 61 1.007 1990 14.03 0.144 14 1.125 1.266 2.469 0.587 7 11 18 856 61 0.288 1991 41.40 0.56 40.69 4.40 19.37 -0.02 0.45 20.62 33.35 53.97 2525.26 61 1.127 1991 13.64 0.175 14 1.367 1.868 0.729 -0.242 7 10 17 832 61 0.350 1992 43.18 0.42 43.17 3.78 14.26 -0.40 -0.19 17.39 34.58 51.97 3497.38 81 0.835 1992 14.78 0.138 15 1.245 1.550 0.817 0.435 7 12 19 1197 81 0.275 119 Table 111.3. Descriptive statistics for the scale radius to the end of the first marine annulus (units: 0.01 mm). STAT 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Mean Standard Error Median Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 127.32 2.06 129.87 13.79 190.04 -0.54 -0.15 58.91 96.46 155.37 5729.44 45 4.142 137.27 1.89 137.50 12.65 160.09 2.52 -0.63 75.31 93.71 169.02 6177.22 45 3.801 132.94 1.36 131.42 10.55 111.38 0.01 • -0.12 50.55 105.60 156.15 7976.30 60 2.726 135.35 1.80 134.05 12.49 155.92 -0.48 0.14 49.43 109.90 159.33 6496.99 48 3.626 130.64 1.19 130.08 10.06 101.28 -0.11 0.32 47.09 109.89 156.98 9405.79 72 2.365 139.03 2.19 139.79 15.02 225.61 0.22 0.39 67.86 110.36 178.22 6534.52 47 4.410 137.84 2.10 136.27 14.07 198.05 0.06 0.06 63.88 105.34 169.22 6202.93 45 4.228 131.53 1.84 129.29 14.23 202.39 0.12 0.24 66.93 98.22 165.15 7891.74 60 3.675 136.77 1.63 137.56 12.63 159.54 -0.52 0.29 52.75 109.87 162.62 8206.41 60 3.263 138.18 1.50 138.22 13.48 181.65 1.41 -0.38 81.22 95.14 176.36 11192.68 81 2.980 Table 111.4. Descriptive statistics for the number of scale circuli to the end of the first marine annulus. STAT 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Mean Standard Error Median Mode Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 31.51 0.34 31.00 31.00 2.28 5.21 -0.02 0.29 10.00 27.00 37.00 1418 45 0.686 32.40 0.37 32.00 32.00 2.46 6.06 0.47 0.37 11.00 27.00 38.00 1458 45 0.740 30.48 0.29 31.00 31.00 2.27 5.14 0.80 0.22 12.00 25.00 37.00 1829 60 0.585 31.15 0.30 31.00 30.00 2.09 4.38 -0.22 0.38 9.00 27.00 36.00 1495 48 0.608 29.64 0.27 30.00 30.00 2.27 5.14 0.04 -0.31 10.00 24.00 34.00 2134 72 0.533 30.43 0.40 31.00 32.00 2.72 7.42 0.04 0.11 12.00 25.00 37.00 1430 47 0.800 30.76 0.50 30.00 30.00 3.37 11.33 0.45 -0.07 17.00 22.00 39.00 1384 45 1.011 29.78 0.41 30.00 29.00 3.14 9.87 0.03 0.11 15.00 22.00 37.00 1787 60 0.811 31.52 0.41 31.00 30.00 3.14 9.85 -0.40 0.04 14.00 24.00 38.00 1891 60 0.811 . 30.93 0.31 31.00 30.00 2.83 7.99 0.68 0.14 16.00 24.00 40.00 2505 81 0.625 120 Table 111.5. Descriptive statistics for the width of the first marine scale zone (units: 0.01 mm) STAT 1982 1983 1985 1986 1987 1988 1989 1990 W91 1992 Mean Standard Error Median Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 90.49 1.69 90.27 11.32 128.20 -0.41 -0.19 46.84 65.67 112.51 4071.91 45 3.402 98.05 1.77 97.01 11.85 140.42 1.32 -0.13 63.14 62.93 126.07 .4412.40 45 3.560 93.81 1.24 93.49 9.62 92.61 0.18 0.05 46.63 71.67 118.30 5628.39 60 2.486 94.17 1.49 94.95 10.34 106.96 -0.71 -0.01 41.26 72.28 113.54 4520.00 48 3.003 88.01 1.05 87.91 8.88 78.79 0.28 0.28 45.78 68.85 114.63 6336.80 72 2.086 97.24 2.03 96.93 13.93 194.04 0.48 0.57 65.01 70.59 135.60 4570.05 47 4.090 96.98 1.90 97.02 12.73 161.99 0.48 0.17 63.59 67.31 130.90 4364.16 45 3.824 91.48 1.74 89.71 13.51 182.49 0.31 0.41 65.67 61.02 126.69 5489.00 60 3.490 95.39 1.47 93.11 11.39 129.68 0.06 0.32 55.67 66.62 122.29 5723.33 60 2.942 95.00 1.33 95.91 11.96 143.04 1.40 -0.36 70.69 57.48 128.17 7695.30 81 2.645 Table III.6. The descriptive statistics for the number of circuli in the first marine scale zone. STAT 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Mean Standard Error Median Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 44.18 0.40 44.00 2.66 7.06 -0.29 -0.10 12.00 38.00 50.00 1988.00 45 0.798 45.29 0.41 45.00 2.72 7.39 1.38 0.18 13.00 39.00 52.00 2038.00 45 0.817 43.63 0.31 43.00 2.37 5.63 0.47 0.30 12.00 38.00 50.00 2618.00 60 0.613 45.17 0.36 45.00 2.48 6.14 -0.36 0.53 10.00 41.00 51.00 2168.00 48 0.720 43.33 0.30 43.00 2.52 6.34 0.19 0.01 12.00 37.00 49.00 3120.00 72 0.592 43.70 0.46 44.00 3.16 10.00 0.39 0.07 15.00 36.00 51.00 2054.00 47 0.928 44.24 0.56 45.00 3.75 14.10 0.08 -0.14 18.00 35.00 53.00 1991.00 45 1.128 43.82 0.40 44.00 3.12 9.75 -0.16 0.17 14.00 38.00 52.00 2629.00 60 0.806 45.15 0.41 45.00 3.20 10.23 -0.28 0.00 15.00 37.00 52.00 2709.00 60 0.826 45.71 0.36 45.00 3.26 10.64 0.09 0.23 17.00 38.00 55.00 3657.00 80 0.726 121 Table 111.7. Descriptive statistics for the intercircular distance averaged over marine circuli 3 to 5 (0.01 mm). STAT 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Mean Standard E Median rror Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level(95.0%) 2.79 0.08 2.79 0.53 0.28 -0.15 0.13 2.46 1.46 3.92 125.52 45 0.160 2.80 0.08 2.78 0.51 0.26 -0.34 0.19 2.19 1.64 3.84 126.21 45 0.152 2.90 0.05 2.89 0.41 0.17 0.14 0.05 2.01 1.85 3.86 173.85 60 0.107 2.92 0.08 2.86 0.56 0.31 0.07 0.15 2.56 1.57 4.13 146.16 50 0.158 3.01 0.05 3.03 0.44 0.20 -0.20 -0.05 2.21 1.86 4.07 216.41 72 0.104 3.20 0.07 3.15 0.53 0.28 -0.76 0.26 2.01 2.34 4.35 163.07 51 0.148 3.34 0.05 3.41 0.32 0.10 0.00 -0.48 1.43 2.47 3.90 153.86 46 0.095 3.14 0.07 3.09 0.56 0.32 -0.12 0.34 2.60 1.92 4.52 188.63 60 0.145 2.96 0.06 3.00 0.45 0.20 -0.43 -0.15 1.82 1.98 3.80 180.79 61 0.114 3.11 0.06 3.16 0.50 0.25 -0.30 -0.13 2.35 1.91 4.26 248.63 80 0.112 Table III.8. Descriptive statistics for the intercircular distance averaged over marine circuli 8 to 10 (10'2 mm). STAT 1982 1983 1985 1986 1987 1988 1989 1990 1991 1992 Mean Standard E Median rror Standard Deviation Sample Variance Kurtosis Skewness Range Minimum Maximum Sum Count Confidence Level (95.0%) 2.99 0.08 2.98 0.54 0.29 -0.11 0.02 2.41 1.70 4.12 134.40 45 0.162 2.99 0.07 2.98 0.49 0.24 -1.25 0.13 1.79 2.14 3.93 134.61 45 0.149 3.05 0.06 2.99 0.46 0.21 -0.11 0.40 2.12 2.13 4.26 183.06 60 0.119 2.99 0.08 2.94 0.53 0.28 -0.45 0.39 2.12 2.07 4.19 149.45 50 0.151 3.11 0.05 3.10 0.45 0.21 0.00 0.29 2.16 2.26 4.42 224.13 72 0.107 3.30 0.08 3.29 0.55 0.30 -0.59 0.21 2.13 2.27 4.40 168.31 51 0.154 3.37 0.08 3.38 0.52 0.27 -0.55 -0.27 2.11 2.22 4.33 155.21 46 0.155 3.24 0.06 3.20 0.48 0.23 0.01 0.48 2.04 2.33 4.37 194.41 60 0.125 3.02 0.06 3.03 0.46 0.21 -0.39 -0.04 2.07 1.85 3.92 184.46 61 0.117 3.17 0.07 3.10 0.58 0.34 1.25 0.67 3.23 2.02 5.26 253.97 80 0.130 122 a) Month JAN FEB MAR APR MAY JUN JUL AUG SEP OCT NOV DEC Day Zero of Month 0 31 59 90 120 151 181 212 243 273 304 334 b) Release Date of SWE Marine Circulus Number I 130 132 134 136 138 140 142 144 146 148 150 152 154 156 158 160 162 check end 164 166 168 170 172 174 176 178 180 182 184 186 188 190 192 194 196 1 171 173 175 177 179 181 183 185 187 189 191 193 195 197 199 201 203 2 178 180 182 184 186 188 190 192 194 196 198 200 202 204 206 208 210 3 184 186 188 190 192 194 196 198 200 202 204 206 208 210 212 214 216 4 191 193 195 197 199 201 203 205 207 209 211 213 215 217 219 221 223 5 198 200 202 204 206 208 210 212 214 216 218 220 222 224 226 228 230 6 205 207 209 211 213 215 217 219 221 223 225 227 229 231 233 235 237 7 211 213 215 217 219 221 223 225 227 229 231 233 235 237 239 241 243 8 218 220 222 224 226 228 230 232 234 236 238 240 242 244 246 248 250 9 225 227 229 231 233 235 237 239 241 243 245 247 249 251 253 255 257 10 232 234 236 238 240 242 244 246 248 250 252 254 256 258 260 262 264 11 238 240 242 244 246 248 250 252 254 256 258 260 262 264 266 268 270 12 245 247 249 251 253 255 257 259 261 263 265 267 269 271 273 275 277 Table III.9. a) Conversion of calendar day to day of the month. b) Table to determine the calendar day of the first 12 marine circuli from release dates between day 130 and 162. The median release date between 1982 and 1992 was found to be day 140. 123 

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:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0053284/manifest

Comment

Related Items