Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The physical oceanography of British Columbia's inside passage with respect to the return migration of… TerHart, Bert Adrian 1990

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

Item Metadata

Download

Media
831-UBC_1990_A6_7 T47.pdf [ 6.92MB ]
Metadata
JSON: 831-1.0053104.json
JSON-LD: 831-1.0053104-ld.json
RDF/XML (Pretty): 831-1.0053104-rdf.xml
RDF/JSON: 831-1.0053104-rdf.json
Turtle: 831-1.0053104-turtle.txt
N-Triples: 831-1.0053104-rdf-ntriples.txt
Original Record: 831-1.0053104-source.json
Full Text
831-1.0053104-fulltext.txt
Citation
831-1.0053104.ris

Full Text

T H E P H Y S I C A L O C E A N O G R A P H Y O F B R I T I S H C O L U M B I A ' S I N S I D E P A S S A G E W I T H R E S P E C T T O T H E R E T U R N M I G R A T I O N O F O N C O R H Y N C H U S N E R K A by Bert A d r i a n terHart .Sc. (Physics and Physical Oceanography) Royal Roads M i l i t a r y College A T H E S I S S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E C R E E O F M A S T E R ; O F S C I E N C E in T H E F A C U L T Y O F G R A D U A T E S T U D I E S D E P A R T M E N T O F O C E A N O G R A P H Y We accept this thesis as conforming to the required standard University of B r i t i s h Columbia M a r c h 1990 © Bert A d r i a n terHart , 1990 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 QCd/U4)6&9cW1/ The University of British Columbia Vancouver, Canada Date ^ > APte/C J ??0 DE-6 (2/88) A B S T R A C T D a t a f rom five conductivity-temperature-depth ( C T D ) surveys collected dur-ing 1985 and 1986 in support of project M O I S T -Meteorological and Oceanographic Influences on Sockeye Tracks- are used to describe the salient oceanographic fea-tures of the waters ly ing between Vancouver Island and the B r i t i s h C o l u m b i a main-land coast. Using these data, four oceanographic regimes are clearly denned on the basis of salinity structure. Temperature-Salinity diagrams are used to discuss water types and mix ing ratios in these regimes. Vigorous t idal mix ing over shallow sills and/or in narrow channels produces t idally mixed fronts that separate oceano-graphic regimes. The t idal evolution of two of these fronts located near Weynton Passage and Cape Mudge are discussed by means of 24-hour C T D stations. Sea-sonal variabil ity of the residual estuarine circulation is examined and estimates of the seaward flow in the upper layer of a very simple two-layer geostrophic model were found to be in reasonable agreement w i t h the few direct measurements made in this region. Seasonal variabil ity of the general hydrography is described. Ultrasonic telemetry provided horizontal and vertical distr ibution time se-ries data for return migrat ing sockeye salmon (Oncorhynchus nerka). Concurrent high spatial resolution C T D data was used to specify the ambient temperature and salinity fields in the immediate vic ini ty of the tagged sockeye. Spectral analysis of the depth and ambient oceanographic data t ime series revealed periodic vertical movements at approximately 15 and 33 minutes per cycle for fish tracked in the slightly stratified regimes of Queen Charlotte Strai t , western Johnstone Strait and the Strait of Georgia. H i g h frequency large amplitude periodic vertical movements were characteristic of fish that did not make significant progress towards the Fraser i i River : low frequency small amplitude vertical movements were characteristic of well oriented fish. Aspect rat io, defined as the horizontal distance travelled divided by the vertical distance travelled, gave an indication of the relative degree of homeward orientation. Vert ical distr ibution and orientation were also related to the frequency and durat ion of successive vertical excursions. F i s h depth and vertical swimming ve-locity were found to be positively correlated in regions of weak stratification and/or for well oriented fish. Ambient density gradients were not found to inhibit verti-cal movements as the rate of doing work against hydrodynamic drag was several orders of magnitude greater than that of doing work against a varying buoyancy force. In the presence of strong temperature and salinity gradients, tracked sockeye were most often observed at depths not associated wi th the m a x i m u m gradients. In stratif ication regimes where temperature and salinity gradients were nearly uni-form w i t h depth, tracked sockeye were observed at depths uniformally distributed throughout the thermo- and haloclines. M i n i m u m vertical swimming velocities were generally associated w i t h min imum vertical gradients. These observations suggest that the tracked sockeye frequently swam through but did not reside in the region of the m a x i m u m gradients. Dimensional analysis suggested that physical variables alone are insufficient to specify the vertical distr ibut ion of the tracked sockeye. iii T A B L E O F C O N T E N T S Abstract ii List of Tables v i List of Figures vii i Acknowledgements xv 1 H y d r o g r a p h y of Bri t ish Columbia's Inside Passage 1 1.1 Introduct ion 1 1.2 D a t a Col lect ion and Processing 4 1.3 Discussion 6 1.3.1 The Inside Passage 7 1.3.2 Queen Charlotte Sound 15 1.3.3 Queen Charlotte Strait 22 1.3.4 The Strait of Georgia 38 1.4 T - S Characteristics 61 1.5 Rota t iona l Effects 62 1.6 Summary 67 2 Ultrasonic Tracking: Horizontal and Vertical Distributions 69 2.1 Introduct ion 69 2.2 D a t a Col lect ion and Processing 71 2.3 Hor izonta l Movements 74 2.3.1 General Overview 74 2.4 Ver t i ca l Movements 77 iv 2.4.1 General Overview 77 2.4.2 Ambient Oceanographic Variables 84 2.4.3 Energy Expenditure 107 2.4.4 Swimming Velocity and A O V Relationships 119 (a) Horizontal and Vert ical Velocity 119 (b) Depth Anomaly and Normal ized Vert ical Velocity 120 (c) Depth and Normalized Vert ica l Velocity 122 (d) Depth and Normalized Temperature Gradient 126 (e) Normalized Temperature Gradient and Vert ica l Velocity 129 (f) Dive /Ascent Characteristics 129 2.4.5 Dimensional Analysis 132 2.5 Summary 137 References 144 V Tables L I ST O F T A B L E S Table 1 : Details of the C T D sections. Figures 2-5 show the cruise tracks of the research ships and section numbers. Figure 6 shows the locations of the two sets 24-hour stations. Sections and 24-hour stations are grouped geographically and are summarized in terms of the number and station spacing 8 Table 2 : Details of the observed fronts. The fronts are numbered in order of their first mention in the text. Summary includes location, Sections observed i n , type of front (sfc implies surface, ssfc subsurface, M T mixed tidal and E estuarine), horizontal gradients and the state of the tide 14 Table 3: Details of the set of 24-hour stations that were conducted in the vicin-ity of Weynton Passage. The summarized data includes station location, the number of C T D casts, the spacing-in hours-between the casts and the C T D cast numbers. Note that the C T D cast numbers are not sequential. . . . 36 Table 4: Details of the set of 24-hour stations that were conducted in the vicinity of Cape Mudge Shoals. The summarized data includes station location, the number of C T D casts, the spacing- in hours-between the casts and the C T D cast numbers. Note that the C T D cast numbers are not sequential. . . . 55 Table 5: Parameters and results of the T S analysis made in §1.5. Summary includes the water mass, source water masses (subscripts 'sfc' and ' d p ' refer to surface and deep layers respectively). T S properties of the source water masses and the derived m i x i n g co-efficients 64 Table 6: Geostrophic shear calculations using a two layer model . Channel width (km), internal Rossby radius of deformation, r-i (km), interfacial isopycnal and surface layer depths are shown for each across-channel section where subsurface isopycnals were obviously sloped 66 Table 7: Loca t ion , durat ion, direction, speed of movement and depth of travel of adult sockeye salmon tracked in 1985 and 1986. Net direction refers to the compass direction from point of release to the point where the fish was abandoned, lost or recaptured. Net distance travelled is determined in an analogous manner to that of net direction 75 VI Tables Table 8: Results of the run test performed on the c*t and 'y,- t ime series data for each fish. The number of runs of 0 2 0 , (the standard deviation of sequential 20 minute segments of the track) about omeci (the median standard deviation of al l the cr2o.) is given for D , T and S. The number of runs is shown as Nrun. Stationarity may be assumed at the 95% significance level if the number of runs, N r u n satisfies 6 < N r u n < 15. Note that 6 and V imply, respectively, that stationarity can or can not be assumed 99 Table 9: The average work rate calculated for a complete track. Units are Joules • m i n - 1 . The rate of energy expenditure swimming vertically, swim-ming horizontally and the total rate is given for each fish 116 Table 10 : Normal ized aspect ratios, A R , , , for each fish. Total distances travelled (m) in the vertical and the horizontal as well as the raw aspect rat io, A R , is given 118 Table 11: Linear least squares fit to the dive duration data for each fish. Intercept, b , and slope, m , are shown for each fish. The standard error is given for b and m . F i sh order is by orientation by regime and year. Oriented and disoriented fish are represented by the subscripts 'o ' a n d ' d ' respectively 133 vi i L IST O F F I G U R E S Figures Figure 1. M a p of Vancouver Island and the Br i t i sh C o l u m b i a mainland coast. Locations referred to in the text are shown as well as important land masses and waterways 3 Figure 2 . M a p showing the location of the two sets of 24 hour stations. Stations in the vicinity of Weynton Passage were conducted 28-19 August 1985 and those in the vicinity of Cape Mudge were conducted 24-15 June 1986 5 Figure 3 . Cruise track of the research vessel C S S Vector (section 1: 24-25 June 1985). Station locations and selected station numbers are superimposed over the vessel's track 10 Figure 4. Cruise track of the research vessel C S S Vector (section 2: 25-26 August 1985). Station locations and selected station numbers are superimposed over the vessel's track 11 Figure 5 . Vertical section of at for section 1. Stations extend from central SG to Q C S D (right to left) and were occupied during the period 25-26 June 1985. Station locations are shown in Figure 2. Contouring interval is 0.5 ot. . 12 Figure 6. Vert ical section of Ot for section 2. Stations extend from central S G to Q C S T and were occupied during the period 26-27 August 1985. Station locations are shown in Figure 3. Contouring interval is 0.5 o~t 16 Figure 7. Cruise track of the research vessel CSS Parizeau (sections 4 and 5: 3 September 1986). Station locations are superimposed over their respective tracks. Section numbers are shown adjacent to their respective tracks. 17 Figure 8. Vert ical section of ot for section 3. Stations extend from Q C S D to JS and were occupied during 26 June 1985. Contouring interval is 0.5 cr t. 19 Figure 9. Vert ical section of cf< for section 4. Stations locations (shown in Figure 4) extend from western JS to Q C S D and were occupied during 3 September 1986. Contour ing interval is 0.5 Ot 20 Figure 10. Vert ical section of o~t for section 5. Stations locations (shown in Figure 4) extend west to east through Goletas Channel and were occupied 3 September 1986. Contouring interval is 0.5 ot 21 vi i i Figures F i g u r e 11. Vert ical section of Ot for section 6. Stations extend meridionally (true north to south) in Q C S D and were occupied during 25 June 1985. Contouring interval is 0.5 Ot 23 F i g u r e 12. Vertical section of at for section 7. Stations extend meridionally in Q C S D and were occupied during 3 September 1986. Contouring interval is 0.5 ot 24 F i g u r e 13. Vertical section of at for section 9. Stations extend from western JS to Q C S T and were occupied 29 M a y 1986. Contouring interval is 0.2 ot. . 25 F i g u r e 14. Vertical section of at for section 10. Stations extend north to south through Weynton Passage and were occupied 3 September 1986. Contouring interval is 0.4 at 26 F i g u r e 15. Vert ical section of at for section 8. Stations extend east to west Broughton Channel and were occupied 26 June 1985. Contour ing interval is 0.5 ot 27 F i g u r e 16. Vert ical section of at for section 11. Stations extend across-channel in the northern port ion of Q C S T and were occupied 26 June 1985. Contouring interval is 0.5 at 30 F i g u r e 17. Vertical section of ot for section 12. Stations extend across-channel in northern Q C S T and were occupied 3 September 1986. Contour ing interval is 0.5 ot : 31 F i g u r e 18. Vert ical section of at for section 13. Stations extend across-channel in the central port ion of Q C S T and were occupied 27 August 1985. Contouring interval is 0.5 at 33 F i g u r e 19. Vertical section of at for section 14. Stations extend across-channel in southern Q C S T and were occupied 26 June 1985. Contouring interval is 0.5 at 34 F i g u r e 20. Vert ical section of Ot for section 15. Stations extend across-channel in southern Q C S T and were occupied 3 September 1986. Contour ing interval is 0.5 Ot 35 IX Figures Figure 21. T ime-depth contour plot of at for Station W P located i n Weynton Passage. This station was sampled recursively over the 24 hour per iod 28-29 August 1985. Contouring interval is 0.5 a*. The occurence of Higher high water ( H H W ) and lower low water ( L L W ) is shown from which the general duration of the flood and ebb tides may be inferred 37 Figure 22. Cruise tracks of the research vessels CSS Vector (section 16: 23-24 June 1986 and section 18: 27-28 Jun . 1985) and CSS Parizeau (sections 17, 19 and 22: 5-6 September 1986) Station locations are superimposed over their respective tracks. Section numbers are shown adjacent to their respective tracks 39 Figure 23. Vert ical section of at for section 16. Station locations (shown in Figure 5) extend axially in the S G and were occupied 23-24 June 1986. Contouring interval is 1.0 at 40 Figure 24. Vert ical section of at for section 17. Station locations (shown in Figure 5) extend north f rom central S G to D P and were occupied 4 September 1986. Contouring interval is 1.0 at 42 Figure 25. Vert ical section of at for section 18. Station locations (shown in Figure 5) extend north to south in M a l a s p i n a Strait and were occupied 27-28 June 1985. Contouring interval is 0.5 a< 43 Figure 26. Vert ical section of at for section 19. Station locations (shown in Figure 5) extend north to south in M a l a s p i n a Strait and were occupied 5 September 1986. Contouring interval is 1.0 at 44 Figure 27. Vert ical profiles of T , S and at for C T D 200-001 and 200-155 taken in central S G (sections 1 and 18, Figures 2 and 5 respectively). Stat ion positions are identical but were occupied during different phases of the tide 46 Figure 28. Vert ical section of at for section 20. Stations extend north to south from D P to Cape Lazo and were occupied 27 June 1985. Tides at Campbel l River were flooding. Contouring interval is 0.5 at 47 Figure 29. Vert ical section of at for section 21. Stations extend north to south from D P to Cape Lazo and were occupied 27 June 1985. Tides at Campbel l River were ebbing. Contouring interval is 0.5 at 48 x Figures Figure 3 0 . Ver t ica l section of Ot for section 22. Station locations (shown in Figure 5) extend north to south from Cape Mudge to Cape Lazo and were occupied 5 September 1986. Tides at Campbell River wer flooding. Contouring interval is 0.5 ot 50 Figure 3 1 . Vert ical section of Ot for section 23. Stations extend south to north f rom Cape Lazo to Cape Sutil and were occupied 27 June 1985. Contouring interval is 0.5 Ot 51 Figure 3 2 . Vert ical section of ot for section 24. Stations extend across-channel in the extreme north of the SG and were occupied 27 June 1985. Contouring interval is 0.5 ot 52 Figure 3 3 . Vert ica l section of ot for section 25. Stations extend across-channel in the extreme north of the S G and were occupied 5 September 1986. Contouring interval is 0.5 ot 54 Figure 3 4 . Vert ical section of ot for section 26. Stations extend across-channel in central S G (Cape Lazo to Powell River) and were occupied 27 June 1985. Contouring interval is 0.5 at 56 Figure 3 5 . Vert ical section of Ot for section 27. Stations extend across-channel in central S G (Cape L a z o to Powell River) and were occupied 5 September 1986. Contour ing interval is 0.5 Ot 57 Figure 3 6 . Vert ical section of ot for section 28. Stations extend east to west across-channel in southern S G and were occupied 5 September 1986. Contouring interval is 1.0 ot 58 Figure 3 7 . T ime-depth contour plot of ot for Station D P located at the southern entrance to D P . This station was sampled recursively over the 24 hour period 24-25 June 1986. Contouring interval is 0.5 Ot- Times of higher high water ( H H W ) and lower low water ( L L W ) are shown from which the general duration of the flood and ebb tides may be inferred 60 Figure 3 8 . Composite T S curves for the entire study area for both 1985 and 1986. Water masses ( l = S G s f c , 2 = S G d P , 3 = D P , 4 = J S s f c , 5 = J S a p , 6 = Q C S T s f c and 7 = QCSTdp) are shown adjacent to their characteristic T S curves 63 Figure 3 9 . Oriented and disoriented vertical movements for 8516. Depth is shown on the y-axis and t ime on the x-axis 80 xi Figures Figure 4 0 . Composite fish depth histogram and T and S profiles for 8507, 8508, 8509 and 8510. Tracks were conducted in the homogeneous waters of eastern JS and D P 81 Figure 4 1 . Composite fish depth histogram and T and S profiles for 8503, 8504 and 8505. Tracks were conducted in the slightly stratified waters of Q C S T and western JS 82 Figure 4 2 . Composite fish depth histogram and T and S profiles for fish tracked in 1985 and 1986 ( (a) and (b) respectively) in Q C S T / J S and the S G . Dashed lines represent ± 1 standard deviation. There were 3,175 (7,414) observations of fish depth and 128(74) C T D casts in 1985(1986) 85 Figure 4 3 . Depth time series A O V data for 8512 tracked in the S G near the northern t ip of Texada Island. The dashed line represents the least-squares estimate of the linear trend 90 Figure 4 4 . Coherence and phase differences for raw T / S A O V data for fish 8503 tracked in the western extremity of JS. The t ime series are highly coherent and vir tual ly 180° out of phase at all times 93 Figure 4 5 . R u n test of standard deviation (020) for consecutive intervals of 20 observations about the median for all intervals. A single run consists of a se-quence of consecutive observations of 0 2 0 either greater or less than the median value. Data is for 8504 tracked in Q C S T near M a l c o l m Island 98 Figure 4 6 . Autocorrelat ion R^1 vs lag for 8509 and 8510 tracked in the homoge-neous water of D P 101 Figure 4 7 . Autocorrelat ion R^1 vs lag for 8511 tracked in the S G near the northern t ip of Texada Island. Periodicity is implied by statistically significant zero-crossings 103 Figure 4 8 . Autospect rum C for all fish tracked in both years (upper frame). Ga (T data) for 8516 (tracked in the SG) is shown in the lower frame. . . . 104 Figure 4 9 . Total work rate considering rate of energy expenditure by swimming vertically and horizontally. Uni ts of work rate are J o u l e s - m i n - 1 . Elapsed track time is shown on the x-axis. Normalized area represents the average work rate over the entire track in Joules • m i n - 1 . 8516 was tracked in the central S G and 8603 central Q C S T 114 xii Figures Figure 50. Scatter plots of horizontal and vertical velocities for 8516, tracked in central SG near the B C mainland coast, and 8610, tracked in northern SG south of Cape Mudge 121 Figure 51. Scatter plots of depth anomaly and normalized vertical velocity for 8503, tracked in the western extremity of JS, and 8513, tracked in the S G near the southern entrance to Sabine Channel . The depth anomaly represents the absolute value of the distance away from the mean depth 123 Figure 52. Scatter plots of depth and normalized vertical velocity for the composite data 85.3-5 (fish tracked in Q C S T and western JS) and 85.7-10 (fish tracked in the SG) 124 Figure 53. Scatter plots of depth and normalized vertical velocity for the composite data 85.11-16 and 86.10 -13 . These data are for all fish tracked in the SG for both years 125 Figure 54. Scatter plots of normalized temperature derivative, [dT/dz)^ and depth for 8511, tracked in the S G near the northern t ip of Texada Island, and 8514, tracked in cenral S G near the entrance to Sabine Channel 128 Figure 55. T i m e between successive dives/ascents (min) versus depth of dive/ascent (m) for 8511, tracked in the S G near the northern t ip of Tex-ada Island, and 8516, tracked in central S G near the B C mainland coast. Data shown for 8511 is characteristic of disoriented fish; there are frequent deep dives and no periods of relatively constant depth. These data are in sharp contrast to those for oriented fish represented here by 8516 130 Figure 56. T ime between successive dives/ascents (min) versus depth of dive/ascent (m) for 8603 and 8606 (both tracked in Q C S T ) . These data are characteristic of oriented fish in that there are relatively infrequent dives and long periods spent at constant depths 131 Figure 57. Dive durat ion (min) vs depth of dive (m) for 8511 and 8512. Linear , least-squares best fit estimates are shown as solid straight lines through the data. B o t h fish were tracked in the S G near the northern t ip of Texada Island and were considered disoriented 134 xiii Figures F i g u r e 58. Dive durat ion (min) vs depth of dive (m) for 8516, tracked in central S G near the B C mainland coast, and 8610, tracked in northern SG south of Cape M u d g e . Linear , least-squares best fit estimates are shown as solid straight lines through the data. B o t h tracks are considered characteristic of oriented fish 135 F i g u r e 59. Scatter plots of non-dimensional parameterization of ultrasonically telemetred fish tracking and physical oceanographic data . D a t a shown is for 8603 which was tracked in central Q C S T 138 F i g u r e 60. Scatter plot of dimensional relationships determined by parameteriza-tion of ultrasonically telemetred fish tracking and physical oceanographic data for 8603. T h i s fish was tracked in central Q C S T 139 xiv A C K N O W L E D G E M E N T S Funding for Project M O I S T was provided by N S E R C Strategic Grant G-1485 to Drs . L . M y s a k , K . Hamil ton and C . Groot and by the Department of Fisheries and Oceans of Canada . Funding was provided to the author in the form of an N S E R C Postgraduate Scholarship and by research grants to D r s . L. Mysak and W . W . Hsieh. The many animated and informative discussions w i t h Drs . T . P . Quinn and K . A . Thomson are gratefully acknowledged as is the unfai l ing support of D r . W . W . Hsieh, who became my supervisor upon the departure of D r . Mysak from the University of Br i t i sh Columbia . The generosity of A . T . and A . J . Weaver, and O . Walsh knows no bounds and is gratefully acknowledged. The many hours my sister, L . t e rHar t , contributed in helping prepare the figures is gratefully acknowledged. T h i s thesis is dedicated to my wife, M . J . t e rHar t , whose undying confidence was instrumental in seeing this work to its conclusion, and to my father, J . D . terHart , whose knowledge of the sea far exceeds that gained by academic experience alone. XV Chapter 1: Hydrography C H A P T E R 1 H Y D R O G R A P H Y O F BR IT I SH C O L U M B I A ' S INS IDE P A S S A G E 1.1: Introduction A n ambitious mult i -discipl inary study was undertaken in the spring and sum-mer of 1985 and 1985 to address the general relationships between an adult sockeye salmon's (Oncorhynchus nerka) immediate environment and its return migration route and t i m i n g . One aspect of this study was an intensive field program designed to combine high resolution data of the sockeye's vertical and horizontal movements w i t h similar resolution physical oceanographic data. Thus , high resolution temper-ature (T) , salinity (S) and depth data were obtained along the return migration route of ultrasonically tagged sockeye. It became clear if the T and S data collected during tracking were to be used to relate the fish's immediate environment to its horizontal and vertical movements, then a detailed hydrography of the study area would be required. To this end, data collected from five conductivity-temperature-depth surveys conducted during the spring and summers of 1985 and 1986 in support of project M O I S T - M e t e o r l o g i c a l and Oceanographic Influences on Sockeye Tracks- are used to describe the salient oceanographic features of the waters ly ing between Vancouver Island and the Br i t i sh C o l u m b i a mainland coast. A total of 28 conductivity-temperature-depth ( C T D ) sections were completed and seven 24-hour stations occupied during the cruises of the research vessels CSS Vector (24-28 June 1985, 26-29 August 1985, 29 M a y 1986 and 23-25 June 1986) 1 •Chapter 1: Hydrography and C S S Parizeau (3-6 September 1986). The study region was l imited to those waters lying between Vancouver Island and the Bri t ish C o l u m b i a ( B C ) mainland coast; an area known as the "inside passage". Special attention was given to those regions wi th in the study area where strong t idally mixed and estuarine fronts were known to exist (Thomson et al., 1985). The purpose of al l the cruises was to provide baseline hydrographic data for the 1985 and 1986 sockeye tracking program of project M O I S T (Quinn and terHart , 1987, Quinn et al, 1989). To this end, the different stratification regimes and fronts that lay in the return migratory path of the sockeye salmon were examined. The southern portion of the inside passage consists of Queen Charlotte Strait ( Q C S T ) , Johnstone Strait (JS), Discovery Passage (DP) and the Strait of Georgia (SG) . The general geographic region is i l lustrated in Figure 1. Table 1 summarizes the 28 C T D sections, Table 2 the observed fronts and Tables 3 and 4 the two sets of 24-hour stations in Weynton Passage and Discovery Passage, respectively. Using these data , four oceanographic regimes are clearly defined on the basis of salinity structure. Temperature-salinity (TS) diagrams are used to discuss water types and mix ing ratios in these regimes. Table 5 summarizes T S characteristics ob-served in this region. Vigorous t idal m i x i n g over constricted sills produces t idally mixed fronts throughout the inside passage which serve to separate the oceano-graphic regimes. The isopleths of the observed fronts are substantial ly modified by the residual estuarine circulation and as such, the isopleth topology differs from that of a classically defined t idal ly mixed front. This different form of a t idally mixed front is defined as a hybr id t idal ly mixed front (Bowman and Essais, 1978). The t idal evolution of the fronts located near Weynton Passage (F7) and Cape 2 Chapter 1: Hydrography 1. Texada Island 2. Sabine Channel 3. Malaspina Strait 4. Powell River 5. Cape Lazo 6. Cape Sutil 7. Cape Mudge 8. Willow Point 9. Seymour Narrows 10. Campell River 11. Malcolm Island 12. Weynton Passage 13. Blackney Passage 14. Kelsey Bay 15. Nahiwitti Bar 16. Gordon Channel 17. Salmon Passage 18. Port Hardy 19. UBC 20. PBS Figure 1 . Map of Vancouver Island and the B.C. mainland coast. Locations referred to in the text are shown as well as important land masses and waterways. 3 Chapter 1: Hydrography M u d g e ( F l ) are described by means of twenty-four hour stations. Figure 2 gives the locations of the 7 sets of 24-hour stations. A description is given of the seasonal changes in the general hydrography where applicable. Thomson et al. (1985) and terHart (1988) contain the complete set of T , S and sigma-t ( ot) sections for each of the 28 sections, the 7 24-hour stations and plots of all 393 C T D profiles. These surveys provide the greatest spatial resolution of the hydrography of this region to date. Detailed examinations of the physical oceanography of the region may be found in Waldichuck (1957) and in the excellent reviews of Thomson (1981) and L e B l o n d (1983). 1.2: Data Collection and Processing The research vessels C S S Vector and C S S Parizeau of the Institute of Ocean Sciences, B C were used to collect 393 C T D profiles. W i t h the exception of the 24-28 June 1985 cruise of the CSS Vector and the 3-6 September 1986 cruise of the CSS Parizeau, ship time aboard the C S S Vector was shared w i t h other investigators. A t the request of the author, the data collected on 29 M a y 1986 presented herein was acquired while the CSS Vector was transi t t ing the study area. A Gui ld l ine Instruments M o d e l 8705 Digi ta l C T D Probe and a M o d e l 87102 Control U n i t were used in conjunction w i t h a Hewlett Packard (HP) 93101 personal computer and plotter. The manufacturer's reported accuracies are ± 0 . 0 0 5 ° C , ± 0.01 parts per thousand (ppt) and ± 0.15% ful l scale pressure. The use of the 1000 and 1500 db pressure transducer resulted in depth accuracies of ± 1 . 5 0 and ± 2 . 2 5 m respectively. This system transmitted a triplet of data to the control unit at 40 ms intervals that was recorded on a stereo casette tape recorder. Data were recorded 4 Chapter 1: Hydrography Figure 2 . Map showing the location of the two sets of 24 hour stations. Stations in the vicinity of Weynton Passage were conducted 28-19 August 1985 and those in the vicinity of Cape Mudge were conducted 24-15 June 1986. 5 Chapte r 1: Hydrography only on the downcasts. A rate of descent of 0.5-1.0 m/s generated 12 to 25 triplets of data per meter of depth. Conduct iv i ty and pressure were converted to salinity and depth respectively by the H P and plotted in real t ime at .5 m intervals. The 87102 control unit was later used to play back the stereo cassette tape to a 9-track magnetic tape for processing on U B C ' s A m d a h l computer. The data were despiked and averaged over one meter depth intervals and crt calculated using the International Equat ion of State, 1980 (Pond and Pickard, 1983). Header data for each cast was added consisting of cruise designation number, consecutive C T D cast number for a part icular cruise, time ( P D T ) , date and position to one one-hundredth of a minute. A l l graphical output was produced on U B C ' s Q M S Lazergrafix 1200 laser printer. A n intermittant electronics failure w i t h i n the C T D probe itself probe pro-duced an intermittant positive spike in the T readings. This fault occurred gen-erally at depths greater than 70 m and only during the in i t ia l cruise of the CSS Vector (24-28 June 1985). This was not considered a catastrophic problem and rather than smooth the C T D casts affected, the bad data were l iberally removed from the affected profiles. 1.3: Discussion The inside passage (Figure l ) consists of two large basins ( Q C S T in the north and the S G in the south) connected by a narrow, sinuous channel (JS and D P ) . Sections obtained during this multi-year survey include sixteen along-channel and twelve across-channel transects. The seven 24-hour stations consist of data obtained from a set of three stations in Weynton Passage and four in the vic ini ty of Cape 6 Chapter 1: Hydrography M u d g e (Figure 2). It should be noted that more effort was expended in areas of more physical oceanographic interest. The following discussion w i l l be formatted as follows: an examination of the along-channel sections through the entire inside passage, of the along and across-channel sections and 24-hour stations obtained in Q C S D and Q C S T and finally, of the along and across-channel sections and 24-hour stations obtained in the S G . Table 1 summarizes all sections for the 5 surveys. 1.3.1: The Inside Passage T w o continuous along-channel sections were conducted throughout the inside passage during 24-25 June 1985 and 26-27 August 1985 (sections 1 and 2 - Figures 3 and 4 - respectively). The general physical oceanographic characteristics of the inside passage result f r o m the intrusion, at depth, of relatively co ld , salty Pacific Ocean water through Q C S D and Q C S T , the relatively w a r m , brackish water of the S G moving seaward on the surface and the vigorous t idal m i x i n g in the constricted channels of JS and D P (Thomson, 1981; Thomson et al., 1985; terHart , 1988). Four distinct oceanographic regimes can be identified in section 1 (Figure 5): the slightly stratified vertical structures of the S G , JS and Q C S T w i t h surface to bot tom salinity differences of 3 ppt , .5 p p t , and 1.5 ppt respectively and the vertically well-mixed waters of D P . The slightly stratified vertical salinity structure of the S G , w i t h surface waters of over 12°C and less than 29 ppt , is largely due to the fresh water run-off of the Fraser River . Run-off from several large inlets, most notably Bute and Toba in northern S G and Howe Sound in the south, amounts to only a fract ion of that of the 7 Chapter 1: Hydrography Table 1 Details of the CTD sections. Figures 2-6 show the cruise tracks of the research ships and section numbers. Sections and 24-hour stations are grouped geographically and are summarized in terms of the number and station spacing. Section Region N o . of C T D ' s Stn. spacing 1 Centra l S G to Q C S D (25-26 June 1985) 46 3-20 k m 2 Central S G to Q C S T (26-27 A u g 1985) 30 4-20 k m 3 Q C S D to JS (26 June 1985) 24 3-10 k m 4 Western JS to Q C S D (3 Sep 1986) 14 7.5-25 k m 5 Goletas Channel (3 Sep 1986) 7 7.5-14 k m 6 M e r i d i o n a l section of Q C S D (25 June 1985) 5 7 k m 7 M e r i d i o n a l section of Q C S D (3 September 1986) 5 7-10 k m 8 Broughton Strait (26 June 1985) 9 3-8 k m 9 Western JS to Q C S T (29 M a y 1986) 7 4-9 k m 10 Weynton Passage (3 September 1986) 3 5.5-8 k m 11 Transverse section of northern Q C S T (26 June 1985) 7 3 k m 12 Transverse section of northern Q C S T (3 September 1986) 7 2-3 k m 13 Transverse section of central Q C S T (27 August 1985) 9 3 k m 14 Transverse section of south Q C S T 26 June 1985 5 3 k m 15 Transverse section of southern Q C S T (3 September 1986) 5 2-3 k m 16 S G (23-24 June 1986) 15 7.5-14 k m 8 Chapte r 1: Hydrography Table 1 (Cont.) Details of the CTD sections. Figures 2-6 show the cruise tracks of the research ships and section numbers. Sections and 24-hour stations are grouped geographically and are summarized in terms of the number and station spacing. Section Region N o . of C T D ' s S tn . spacing 17 Centra l S G (4 Sep 1986) 8 15 k m 18 M a l a s p i n a Strait (27-28 June 1985) 6 13 k m 19 M a l a s p i n a Strait 5 September 1986 5 7-15 k m 20 D P to Cape Lazo (flood) (27 June 1985) 14 3 k m 21 D P to Cape Lazo (ebb) (27 June 1985) 17 3 k m 22 Cape Mudge to Cape Lazo (4 September 1986) 5 7.5 k m 23 Transverse section of northern S G (4 Sep 1986) 5 2.5-3.5 k m 24 Cape Lazo to Cape Suti l (27 Jun 1985) 10 3-5 k m 25 Cape Sut i l to W i l l o w Point (27 June 1985) 7 3 k m 26 Transverse section of central S G (4 Sep 1986) 8 4 k m 27 Cape Lazo to Powell River (27 Jun 1985) 10 3 k m 28 Transverse section of southern S G (4 Sep 1986) 7 4-6 k m Fraser River and as such, accounts for l itt le of the observed salinity structure. The vertically well-mixed waters of D P are produced by rigorous t ida l agitat ion through Seymour Narrows. To the east of Kelsey Bay, the relatively shallow and slightly stratified waters of JS have temperatures and salinities that vary only slightly from 9 I29U I28U 127U 126U 125U 124U 123U Figure 3. Cruise track of the research vessel CSS Vector (section 1: 24-25 June 1985). Station locations and selected station numbers are superimposed over the vessel's track. bC Mainland ffl << a. -i 9» XI cr Figure 5. Vertical section of ot for section 1. Stations extend from central SG to QCSD (right to left) and were occupied during the period 25-26 June 1985. Station locations are shown in Figure 2. Contouring interval is 0.5 at. Chapte r 1: Hydrography 10°C and 30 ppt . West of Kelsey Bay, the slightly stratified but relatively deep waters of this portion of JS have values of T and S in the range of 9 - 1 0 ° C and 31-32 ppt respectively. The greater degreee of stratification in the deep western portion of JS is a result of the intrusion of cold, salty Pacific Ocean water v i a Q C S T and Q C S D . Last ly , the slightly stratified waters of Q C S T show bot tom waters of Pacific Ocean origin wi th temperatures near 7°C and salinities greater than 33 ppt . Surface temperatures are near 10°C and surface salinities are less than 30 ppt . T i d a l forcing provides the kinetic energy necessary to mix and in extreme cases, homogenize the water column over shallow sills and in constricted channels (Farmer and Freeland, 1983). Indeed, the general estuarine (gravitational) cir-culat ion is greatly dependent on the rate at which m i x i n g occurs and its spatial extent w i t h i n these topographic features (Geyer and C a n n o n , 1982). Thus , the four oceanographic regimes are separated by three t idal ly mixed fronts (see Figures 5 and 6): a subsurface/surface front at the southern entrance to D P ( F l ) , a surface front at Kelsey Bay (F2) and subsurface fronts at the western entrance to JS in Blackney Pass (F3) and Weynton Passage (F7: Figure 8). Table 2 summarizes all fronts observed in the study area. The strongest and most dynamically significant feature anywhere in the study area is surface/subsurface front ( F l ) near Cape M u d g e where the vertical ly well-mixed waters of D P debouch into the S G to meet a buoyancy layer formed by fresh water run-off and solar heating. T h i s surface layer is less than 30 m deep and is bounded by the 10.5°C isotherm, the 29.5 ppt isohaline or the 22.5 at isopycnal. Surface gradients in the region of the front are 4°C and 1.5 ppt per 10 k m . The surface front at Kelsey Bay (F2) separates the vertically wel l -mixed waters of D P 13 Chapte r 1: Hydrography Table 2 Details of the observed fronts. The fronts are numbered in order of their first mention in the text. Summary includes location, Sections observed in, type of front ('sfc' implies surface, 'ssfc' subsurface, M T mixed tidal and E estuarine), horizontal gradients and the state of the tide. Designation Locat ion Section T y p e Hor . G r a d . ° C / 1 0 k m ppt/10 km F l (Ebb) Cape M u d g e 1 sfc/ssfc, M T 2.7 1.0 F l (Flood) Cape M u d g e 20 sfc/ssfc, M T 12.5 3.3 F l E b b ) Cape M u d g e 21 sfc/ssfc, M T 1.4 1.2 F l (Ebb) Cape M u d g e 2 sfc/ssfc, M T 4.0 1.5 F l (Ebb) Cape Mudge 22 sfc/ssfc M T 2.5 1.0 F2 E b b ) Kelsey Bay 1 sfc/ssfc, M T .1/.2 .1/.3 F2 (Ebb) Kelsey Bay 2 sfc, M T 0.3 0.5 F 3 (Flood) Blackney Pass 1 ssfc, M T 1.5 0.8 F3 (Slack) Blackney Pass 2 sfc, M T 1.5 1.0 F 3 (Ebb) Blackney Pass 4 sfc, M T 1.0 1.0 F4 F lood) Q C S T 4 sfc, E 1.0 0.3 F 5 F lood) N a h w i t t i B a r (west) 1 sfc M T 0.8 0.7 F 5 (Ebb) N a h w i t t i B a r (west) 5 sfc, M T 3.0 1.0 F 6 F lood) N a h w i t t i Bar (east) 1 sfc, M T 1.5 0.8 F 6 (Ebb) Nahwit t i B a r (east) 5 sfc, M T 1.0 1.0 F7 (Flood) Weynton Passage 3 ssfc, M T 1.0 0.5 F7 (Flood) Weynton Passage 9 sfc/ssfc, M T A/.2 .8/.8 F7 (Flood) Weynton Passage 10 sfc/ssfc M 1.0 0.4 F8 (Ebb) Broughton Strait 8 ssfc, M T 1.9 1.1 F 9 (Ebb) Malaspina Strait 26 sfc, E 0.3 1.9 and eastern JS from the slightly stratified waters of western JS. Surface gradients associated wi th F 2 are .3°C and .5 ppt per 10 k m . There is a subsurface front (F3) in Blackney Pass where the intrusion of dense Pacific water into the deep basin of JS is evidenced by waters less than 9°C and greater than 32 ppt . The intrusion is deeper than 25 m and its upper extent is delineated by the 9°C isotherm, the 32 ppt isohaline or the 25 ot isopycnal. Subsurface gradients are 1.5°C and 1.0 ppt per 10 k m . There is a m i n i m a l surface signature. T i d a l agitation over the relatively long shallow si l l at the seaward entrance to Goletas Channel produces the hybrid t idal ly mixed fronts F 5 and F 6 at the western and eastern extremities, respectively, 14 Chapter 1: Hydrography of the s i l l . The major difference between sections 1 and 2 (Figure 6) is the presence of a warmer, more buoyant surface layer in the S G during the latter part of the summer. Differences in ot w i t h i n the surface layer (< .5) were largely due to increased solar radiat ion. Increased surface salinities wi th in the central S G were not evident despite significantly less freshwater input from the Fraser River during late summer. Royer and Emery (1982) discuss the response of the Fraser River plume to discharge of the river itself and to wind and t idal forcing. Their results show that the northern port ion of the central S G is not as strongly influenced by the plume as the southern port ion . Observations made during this study support Royer and Emery's (1982) conclusions. 1 . 3 . 2 : Queen Charlotte Sound D a t a collected in Q C S D was logistically l imi ted to the area to the immediate nor th of the northernmost t ip of Vancouver Island. The general physical oceanogra-phy is characterized by the outflow of surface waters and the inflow of dense Pacific Ocean waters caused part ia l ly by continuity constraints and by the wind-induced upwell ing of deep waters in Q C S D and offshore (Dodimead, 1980; Danie l , 1985). Three along-channel sections (Figures 8-10) and two meridional section (Fig-ures 11 and 12) were made during the C S S Vector and Parizeau cruises of 25-28 June 1985 and 3-6 September 1986 respectively. Locations for sections 4 and 5 are shown in Figure 7. Sections 3 and 4 (Figures 8 and 9), f rom the western entrance of JS through Blackney Pass or Weynton Passage and into Q C S T and the open waters of Q C S D 15 Chapter 1: Hydrography 4 1 1 1 1 1 1 1 I 1 1 1 • — • >—-H—>—I » I ' § Flange: 1 tick=15 km F i g u r e 6 . Ver t ica l section of ot for section 2. Stations extend from central S G to Q C S T and were occupied during the period 26-27 August 1985. Station locations are shown in Figure 3. Contour ing interval is 0.5 at. 16 Chapter 1: Hydrography Figure 7 . Cruise track of the research vessel CSS Parizeau (sections 4 and 5: 3 September 1986). Station locations are superimposed over their respective tracks. Section numbers are shown adjacent to their respective tracks. 17 Chapter 1: Hydrography show the subsurface t idal ly mixed fronts (F3 and F7) , in Blackney Pass and Weyn-ton Passage respectively, and the progressively more saline water column as the section progresses seaward. Subsurface gradients are slightly greater than 1.5°C and 0.8 ppt peT 10 km at 100 m (F3) and 1.0°C and 0.5 ppt per 10 k m at 100 m (F7). A weaker front (F4) , delineated by the surfacing of the 24.5 at isopycnal, separating the greater degree of stratif ication in Q C S D from that of the weaker stratified waters of Q C S T is apparent in both sections in the region north of Hope and Nigel Islands. Surface gradients associated w i t h F4 are slightly greater than 1.0°C and .3 ppt per 10 k m . Section 3 (26 June 1985) has a cooler ( l ° C ) less saline (1 ppt) surface layer in Q C S D t h a n section 4 (3 September 1986). Surface layer variabi l i ty in sections 3 and 4 (Figures 8 and 9) is caused by increased solar heating and much reduced freshwater input . Section 5 (Figure 10), f rom Q C S D to Q C S T v i a Goletas Channel , shows the strong t ida l ly mixed fronts (F5 and F 6 , fig 10) produced by t idal flow over N a h w i t t i B a r (see Figure 1 for location). Gradients are 3.0°C and 1.0 ppt and 1.0°C and 1.0 ppt for F 5 and F6 respectively. The meridional sections 6 and 7, Figures 11 and 12, made in the open waters of Q C S D shows the Pacific Ocean waters outside of Q C S T . Most of the water column deeper than 20 m is relatively cold (< 8.5°C) and saline (> 32.5 ppt ) . A s opposed to Figure 12 (section 7), subsurface isopycnals are clearly sloped down to the nor th in Figure 11 (section 6).f Southward sloping subsurface isopycnals on t G i v e n that the Rossby internal deformation radius is less than the channel w i d t h , sloping isopycnals may, but are not necesssarily indicative of rotational effects (LeBlond and M y s a k , 1982). The applicabil i ty of simple geostrophy to these data 18 129U 52N 128U 5 IN 50N 129U I27U ( it I28U 127U 126U 52N 5IN 50N 126U Range: 1 tick=20 km Figure 8. Vertical section of at for section 3. Stations extend from QCSD to JS and were occuoied during 26 June 1985. Contouring interval is 0.5 at. t Chapter 1: Hydrography Range: 1 tick=7.5 km Figure 9 . Vertical section of at for section 4. Stations locations (shown in F i g u r e 4) extend from western JS to QCSD and were occupied during 3 September 19S6. Contouring interval is 0.5 <rt. 20 Chap te r 1: Hydrography Figure 10 . Vertical section of <?, for section 5. Stations locations (shown in F i g u r e 4) extend west to east through Goletas Channel and were occupied 3 September 1986. Contouring interval is 0.5 ct. 21 Chapter 1: Hydrography the Vancouver Island side of Figure 12 (section 7) indicate a relatively strong influx of cold (< 8 ° C ) , saline (> 33 ppt) Pacific Ocean water into Q C S T v i a Goletas Channel . T h e weakening of the residual seaward flowing current velocities in the upper layer, as evidenced by the change in the across-channel slopes of subsurface isopycnals i n Figures 11 and 12, is largely due to decreased freshwater input during summer months. The influx of co ld , saline Pacific Ocean water at depth during the summer is mainly due to increased wind-induced upwell ing in Q C S D (Pickard, 1975; T h o m s o n , 1981) 1.3.3: Q u e e n C h a r l o t t e S t r a i t D a t a f rom five across-channel sections, three along-channel sections and a set of three 24-hour stations were collected during the early and late summer of 1985 and the late spring and summer of 1986. The across-channel sections conducted in the northern, central and southern regions of Q C S T (26 June 1985, 27 August 1985 and 3 September 1986) were made to determine the across-channel structure of Q C S T . T w o along-channel sections through Weynton Passage (29 M a y 1986 and 4 September 1986) were made to observe the intrannual variabi l i ty of the subsurface t idally mixed front (F7) in this region. A single section through Broughton Passage (26 June 1985) was conducted to observe the t idal m i x i n g that occurs over the shallow sil l (< 50 m) separating Q C S T from western JS and the resultant t idal ly mixed front (F8). Three 24-hour stations were occupied dur ing 28-29 August 1985 to observe the strength of the t ida l signal wi th in Weynton Passage. The across and along-channel structures of Q C S T are similar in that there is generally weak vertical stratification and significant seasonal variability. Section 9 22 Figure 11. Vertical section of ot for section 6. Stations extend meridionally in QCSD and were occupied during 25 June 1985. Contouring interval is 0.5 ot. RANGE : ONE TICK = 5.0 KM Figure 12. Vertical section of ot for section 7. Stations extend meridionally in QCSD and were occupied during 3 September 1986. Contouring interval is 0.5 at. W «< CL o XI F7 Figure 13. Vertical section of ot for section 9. Stations extend from western JS to QCST and were occupied 29 May 1986. Contouring interval is 0.2 ot. F i g u r e 14. Vertical section of rr, for section 10. Stations extend north to south through Weynton Pn.ss.Tgc and were occupied 3 September 1086. Contouring interval is 0.4 rr,. s Chapter 1: Hydrography (Figure 13) made during 29 M a y 1986 contrasts w i t h Section 10 (Figure 14) made during 4 September 1986. D a t a f rom section 9 show colder surface waters (< 9.0°C) in the upper buoyancy layer owing to relatively weak solar insolation in early sum-mer. Increased freshwater run-off accounts for decreased surface salinities (< 30.5 ppt) in Section 9 as compared to Section 10. Addi t iona l ly , the t idally mixed front (F7) associated w i t h the intrusion of waters of Pacific Ocean origin into JS in Sec-tion 9 (Figure 13) shows a significant surface signature in sharp contrast to the same feature in Section 10 (Figure 14). Surface gradients in Section 9 associated wi th F7 are .4°C and .8 ppt per 10 k m ; subsurface gradients are .2°C and .8 ppt per 10 k m at 100 m . The upper l imit of the dense intruding waters is denoted by the 7.9°C isotherm, the 31.3 ppt isohaline or the 24.3 at isopycnal (Figure 13). Subsurface gradients for this same feature (F7) in Section 9 are 1.0°C and .4 ppt per 10 k m at 100 m . Section 8 (Figure 15), conducted on 26 June 1985 form JS to Q C S T , clearly shows the presence at depth of co ld , saline waters originating in the Pacific. T i d a l mix ing over the shallow sill in Broughton Passage results in the formation of the t idal ly mixed front F8 and the product ion of relatively dense (at > 25.0) water that eventually spills into the deep trough in western JS (Figure 15). In fact, seasonal changes in water properties in b o t h the upper and lower layers are communicated to varying depths v i a this mechanism ( L e B l o n d , 1983). It should be noted that the horizontal extent of the si l l significantly affects the manner in which t idal energy is propagated into the basin beyond the si l l itself (Farmer and Freeland, 1983). Sills in Weynton Passage and Blackney Pass play significantly different dynamical roles in redistr ibuting tidal energies. D u r i n g summer, when lower layer salinities are at a m a x i m u m (Thomson, 1981), the negatively buoyant plume produced by t idal agi-28 Chap te r 1: Hydrography tat ion may penetrate the complete water column, enhancing vertical structure and abyssal dissolved O 2 . The along-channel two layer, weakly stratified structure of Q C S T was ob-served to vary seasonally as was the across-channel structure. The depth of this layer is fairly uniform along-channel due to t idal mix ing in the north over shoals and through narrow channels in the south. Along-channel variations in the depth of the buoyancy layer may be caused by entrainment of dense, off-shore waters, especially i n summer when wind-induced upwelling is at a seasonal m a x i m u m . This effect, however, was not obvious in the sections obtained during the summer months and is likely obscured by the t idal m i x i n g in the north and the south extremeties of Q C S T . A s expected, the across-channel structure is stongly affected by varying freshwater input albeit freshwater sources are significantly less in Q C S T than in the S G . A transverse section made in northern Q C S T during 26 June 1985 shows slop-ing subsurface isopycnals throughout most of the water co lumn (section 11, Figure 16). These slopes are indicative of a relatively strong seaward flow in the upper layer. In sharp contrast to this section is section 12, obtained on vir tual ly the same transect during 3 September 1986, which shows a nearly uni form across-channel structure. Surface temperatures and salinities are > 11°C and < 30.5 ppt. re-spectively. C o l d (7.0°C) and saline (33.0 ppt) Pacific Ocean water is present in the deep basin near the Vancouver Island coast. Figure 17 shows the isopycnals gently sloping down to the north on the mainland side of the channel indicating a weak geostrophic flow to the N W at section 12. In addit ion, isopycnals (> 25.5) slope down to the south on the Vancouver Island side of the channel indicating a 29 o 8 Range: 1 tick=5.0km I29U 128U 127U 50N 129U 0 IP f 1 1 1 126U 52N 5IN 128U 127U 50N 126U Figure 10. Vertical section of ot for section 11. Stations extend across-channel in northern QCST and were occupied 26 June 1985. Contouring interval is 0.5 at. O •8 w ex. o •a sr Chapter 1: Hydrography Figure 17 . Vertical section of er, for section 12. Stations extend across-channel in northern QCST and were occupied 3 September 1986. Contouring interval is 0.5 a,. 31 Chapte r 1: Hydrography weak geostrophic flow into Q C S T . Section 13 (Figure 18) reveals the across channel structure of central Q C S T . There is somewhat less vertical stratif ication present in the central strait than in the northern portions. There is a broad thermocline and halocline and cold (< 8.5°C) and saline (> 32.5 ppt) Pacific Ocean water is present in the deep port ion of the basin. Surface temperatures and salinities are approxi-mately 9.5°C and 31.5 ppt respectively. The isopycnals slope down to the north, indicative of a geostrophic flow out of Q C S T to the northwest. Similar transverse sections of the southern entrance to Q C S T , made during 26 June 1985 (section 14, Figure 19) and 3 September 1986 (section 15, Figure 20) show l i t t le difference in temperature, salinity or Ot values. There is, however, less slope to the subsurface isopycnals in section 15 (September 1986) due to much reduced freshwater input during the summer months. Consequently, the upper layer estuarine flow out of Q C S T to the northwest (seaward) diminishes. Table 6 presents geostrophic calcu-lations based on a simple two layer model as well as summariz ing the transverse sections made in Q C S T . Stations north of M a l c o l m Island (Station W I ) , in Weynton Passage (Station W P ) and off of B l inkhorn Light (Station B L ) were sampled recursively during the period 28-29 August 1985. Table 3 provides details on this set of 24-hour stations. The location of Station W P and the times of higher-high water ( H H W ) and lower-low water ( L L W ) during the period when this station was occupied are shown in Figure 21. D a t a obtained from Station M I (north of Weynton Passage) shows highly cor-related vertical excursions of isotherms, isohalines and isopycnals of approximately 20 m below 15 m indicating the intrusion of cold (< 9 .0°C) , saline (> 32.0 ppt) rel-32 Chapter 1: Hydrography o « n n t i n co r - co rn n n n n n n n n RANGE t CHE TICK = 5.0 KM Figure 18 . Vertical section of ct for section 13. Stations extend across-channel in central QCST and were occupied 27 August 1985. Contouring interval is 0.5 at. 33 129U 128U 52N 5IN Range: 1 tick=5.0 km 50N 129U 127U Hi /I 0 0 126U 52N 5!N 50N I28U 127U 126U Figure 19. Vertical section of ot for section 14. Stations extend across-channel in southern QCST and were occupied 26 June 1985. Contouring interval is 0.5 ot. t Chapte r 1: Hydrography F i g u r e 20 . Vert ical profile of ot for section 15. Stations extend across-channel in southern Q C S T and were occupied 3 September 1986. Contour ing interval is 0.5 35 Chapte r 1: Hydrography T a b l e 3 Details of the set of 24-hour stations that were conducted in the vicinity of Weynton Passage. The summarized data includes station location, the number of CTD casts, the spacing-in hours-between the casts and the CTD cast numbers. Note that the CTD cast numbers are not sequential. Station location N o . of C T D ' s C T D spacing C T D cast no.'s M a l c o l m Island 12 2 hr 010056-089 (Station M I ) Weynton Passage 12 2 hr 010057-090 (Station W P ) B l i n k h o r n Light 12 2 hr 010058-091 (Station B L ) atively dense (> 25.0 ot) water into Weynton Passage. A t Station W P in Weynton Passage itself, this same water mass is seen to advance and retreat w i t h the flood and ebb tide. Figure 21 shows the core of this dense tongue of water (> 25.5 ot) to completely appear and disappear indicat ing the strength and variabi l i ty of the t idal forcing. The strength of the t idal signal, that is to say the volume of water entering JS , is evidenced by the depression of the 9.0°C isotherm by approximately 100 m at Station B L , located to the south and east of Weynton Passage. The surface waters at Stat ion B L , delineated by the 9 .5°C, the 32.5 ppt isohaline or the 24.0 ot isopycnal remained relatively undisturbed during the one t ida l cycle observed. The depression of the isopleths is also evidence of the rapid s inking of the negatively buoyant plume debouching into JS f rom Weynton Passage. The net result of these t idal motions is the formation of strong subsurface hy-br id t idal ly mixed front (F7, Figures 8, 13 and 14) and a negatively buoyant plume. The plume is pulsed over the si l l in Weynton Pass and flows down-slope into the deep basin in western JS. Thomson (1981) describes current measurements made in the deep port ion of western JS that indicate dense waters formed and forced into JS 36 T ime: l t i c k = 3 hours Figure 2 1 . Time-depth contour plot of ot for Station WP located in Weynton Passage. This station was sampled recursively over the 24 hour period 28-29 August 1985. Contouring interval is 0.5 ot. The occurence of Higher high water (H11W) and lower low water (L1AV) is shown from which the general duration of the flood and ebb tides may be inferred. Chap te r 1: Hydrography through Weynton Pass result in a strong bottom jet that sweeps up-channel (east-ward) at m a x i m u m velocities in excess of 150 cm/s . Similar deep water formation occurs through Blackney Passage and to a lesser extent, Broughton Passage where the horizontal exent of the si l l serves to appreciably alter the dynamics (Geyer and C a n n o n , 1982; Farmer and Freeland, 1983). 1.3.4: Strait of Georgia During the 25-27 June 1985 and the 23-26 June 1986 cruises of the CSS Vector and the 3-5 September cruise of the CSS Parizeau, eight along-channel and five across-channel sections and four 24-hour stations were conducted within the central and northern port ion of the S G . Part icular attention was paid to the hybrid t ida l mix ing front ( F l ) observed in sections 1 and 2, (Figures 5 and 6). Locations for sections 16-19 are shown in Figure 22. Section 16, taken along-channel f rom central S G to just south of Cape Mudge (see Figure 22 for section and station location), shows the early summer vertical structure of the S G . Surface temperatures observed were in excess of 16°C in the south decreasing to less than 12°C in the north while surface salinities increased f rom < 16 ppt in the south to > 26 ppt in the north. There is no surface mixed layer owing to weak and variable winds during summer and a brackish, buoyant layer bounded by the 20 ot isopycnal (Figure 23) caused by solar heating and r u n -off. In contrast to section 2, Figure 6, early summer conditions in the central S G show a pronounced decrease in surface salinities (16 ppt as compared to 28.5 ppt) and similar surface temperatures of « 15°C. Section 17 (Figure 24), taken on 4 September 1986 from A c t i v e pass to Sabine 38 Chapter 1: Hydrography Figure 22 . Cruise tracks of the research vessels CSS Vector (section 16: 23-24 June 1986 and section 18: 27-28 Jun. 1985) and CSS Parizeau (sections 17, 19 and 22: 5-6 September 1986) Station locations are superimposed over their respective tracks. Section numbers are shown adjacent to their respective tracks. 39 Range: 1 tick=5.0 km Figure 23. Vertical section of at for section 16. Station locations (shown in Figure 5) extend axially in the SG and were occupied 23-24 June 1986. Contouring interval is 1.0 ot. Chapte r 1: Hydrography Channel , again contrasts section 16 and may also be compared to section 2, con-ducted approximately one year earlier. In Section 17, a warm brackish buoyancy layer, bounded by the 22.5 ot isopycnal near 25 m is present (Figure 24). Sur-face temperatures of 16 — 18°C are generally 2 — 3°C greater in September 1986 as compared to June 85, but surface salinities are 1.0-1.5 ppt less. In contrast to the early summer conditions (Section 16, F igure 23) of the same year, Section 17 shows a. much warmer, much more saline and less buoyant surface layer. The southern S G is more strongly stratified, in terms of salinity and density structure, in early summer due to increased run-off f rom the Fraser River. Temperature stratification increases as increased solar heating, reduced cloud cover and weak wind-induced mix ing combine to heat the surface layer as summer wears on. However, in temper-ate regions, density is much more strongly influenced by changes i n salinity than changes in temperature. Thus the density stratification decreases, despite increased surface layer temperatures, due to the increased surface layer salinities. Sections 18 and 19 (Figures 25 and 26), conducted during 27-28 June 1985 and 4 September 1986, clearly show the effects of variable fresh-water input and heating. In general, waters in M a l a s p i n a Strait are very similar in vertical structure to those in the central port ion of the S G . However, residence times of brackish waters originating in Howe Sound and near the mouth of the Fraser River that are advected into M a l a s p i n a Strait are prolonged by generally weak, variable t idal currents and winds. Hence, the effects of surface heating and di lut ion are magnified in M a l a s p i n a Strait . In early summer, surface T and S ranged f r o m 10-16.5°C and the surface buoyancy layer was bounded by the 22.5 ct isopycnal at 20 m . Section 19 (Figure 26), conducted during 4 September 1986, shows a s imi-41 RfWCE I ONE TICK = 8.0 KM Figure 24. Vertical section of ot for section 17. Station locations (shown in Figure 5) extend north 3 from central SG to DP and were occupied 4 September 1986. Contouring interval is 1.0 ot. Z H 1 1 H RflNCE i ONE TICK s 3 KM f i g u r e 25 . Vertical section of a, for section 18. Station locations (shown in Figure 5) extend north to south in Malaspina Strait and were occupied 27 28 June 1985. Contouring interval is 0.5 ot. RANGE I ONE TICK = 2.5 KM «< a. o Figure 26. Vertical section of at for section 19. Station locations (shown in Figure 5) extend north S to south in Malaspina Strait and were occupied 5 September 1986. Contouring interval is 1.0 at. Chapte r 1: Hydrography larily stratified water column to that of Figure 25. There is a strong buoyancy layer in both sections and a marked decrease in surfacce salinites. Surface layer temperatures and salinities of 10-15.5°C and 29-28 ppt bounded at 20 m by the 22.5 at isopycnal were observed. There is no surface mixed layer. A general trend to warmer (16°C) and less saline (27 ppt) waters towards the southern entrance to Malasp ina Strait is evidence in both sections of the presence of brackish waters originating in the central portions of the S G . The strength of the t idal signal near the southern entrance to M a l a s p i n a Strait may be examined by considering two C T D casts made at different times at the same location. Figure 27 shows surface layer T and S to be 12.8°C and 27 ppt respectively one hour into a strong ebb after a weak flood. D a t a f rom the same location but collected at the end of a relatively strong flood, shows surface layer T and S to be 16.8°C and 23 ppt . Three sections were made along the axis of the t idal jet exi t ing D P in the v ic in i ty of Cape M u d g e in the northern S G . Sections 20 and 21 (Figures 28 and 29) were made during 27 June 1985 on the flood and ebb respectively; section 22 (Figure 30) was made during 4 September 1986 on the ebb tide (the location for this section is shown in Figure 22). The surface hybr id t idally mixed front is strongest on the flood w i t h surface gradients of 12 .5°C, 3.3 ppt and 1.5 ot units per 10 km (Figure 28). Section 21 shows surface layer temperatures to increase southward from 10.0-16.4°C and surface layer salinities to decrease southward from 29.5-28.0 ppt . Dur ing flood tides, waters at the entrance to D P may be expected to be vertical ly homo-geneous. Section 21, conducted on 27 June 1985 during an ebb t ide, shows surface gradients of 1.4°C, 1.2 ppt and 0.5 ot units per 10 k m (Figure 29) and surface T(S) 45 S!GMR-t 17 18 19 20 21 22 29 24 25 28 27 SOLIN1TY10/00) 29 24 2 5 2 8 27 2 8 2 9 3 0 91 9 2 9 3 9 4 TEMPERATUREI#CI 7 8 9 10 1.1 12 19 14 15 l« 17 V 18 19 5IGMR-T 20 2,1 22 29 24 25 26 27 24 25 SRLIM1TYI0/00) 28 27 28 29 30 91 92 99 94 6 3- — 7 • TEMPERATUREI'C) 9 10 l . | 12 19 14 IS 18 17 Figure 2 7 . Vertical profiles of T. S and o, for CTD 200-001 and 200-155 taken in central SG (sections 1 and 18, Figures 2 and 5 respectively). Station positions are identical but were occupied during different phases of the tide. O tr p> X) m '< o. o <M •-i 0> XI a-«< Figure 28. Vertical section of ot for section 20. Stations extend north to south from DP to Cape Lazo and were occupied 27 June 1985. Tides at Campbell River were flooding. Contouring interval is 0.5 ot. Range: 1 tick=5.0 km SIN I26U I25U 124U 3 50N 49N 126U u > hi I25U 124U I23U 5IN SON I23U o •8 Figure 29. Vertical section of o, for section 21. Stations extend north to south from DP to Cape Lazo and were occupied 27 June 1985. Tides at Campbell River were ebbing. Contouring interval IS U.O (Tfm w •g Chapter 1: Hydrography increase(decrease) southward from 10 .0-12 .5°C(29 .5-28 .0 ppt) . Section 22, con-ducted on 4 September 1986 during an ebb tide shows a marked increase (decrease) in surface T(S) from < 1 0 . 5 ° C ( > 28.5 ppt) at the entrance to D P to > 16"C(< 27 ppt) off Cape Lazo. Surface gradients are of the order 2 .5°C, 1.0 ppt and 1.0 ot units per 10 k m (Figure 30). D u r i n g ebb tides, the surface T and S gradients are significantly weaker as the buoyancy layer is pulled back into the S G and D P by the southward and north-ward retreating tides respectively (LeBlond , 1983). Hence, the ebb tides lead to a rarifaction of isopleths and , consequently, weaker horizontal gradients. Considering the 22.5 Ot isopycnal to be the leading edge of the buoyancy layer, Figures 28 and 29 show the surface front to be 7 k m closer to D P on the flood than on the ebb. Section 23 (27 June 1985) from Cape Lazo to Cape Sut i l is roughly perpen-dicular to the horizontal gradients in the northern S G . Thus , there are no apparent horizontal gradients (Figure 31). F ive across-channel sections were obtained to help specify the across-channel structure of the majority of the S G . Sections 24-28 (Figures 32-36) reveal the cross channel structure of the entire S G to be s imilar : there is a relatively shallow, strong buoyancy layer on the surface w i t h no surface mixed layer and very sharp, co-incidental thermo-, halo- and pycnoclines. The bot tom boundary of this layer, taken as the depth of the 22 ot isopycnal increases southward f rom < 15 m in the north and central S G to approximately 20 m i n the south. In the extreme northern port ion of the S G , section 24 (27 June 1985, Figure 32) shows the influence of the dense well-m i x e d water t idally pumped into the S G . Surface T(S) increases (decreases) eastward from 10 .0 -13 .0°C(29 .5 -28 .9 p p t ) . Figure 32 shows the subsurface isopycnals to slope 49 Figure 30. Vertical section of ot for section 22. Station locations (shown in Figure 5) extend north to south from Cape Mudge to Cape Lazo and were occupied 5 September 1986. Tides at Campbell River wer flooding. Contouring interval is 0.5 a%. 126U 51N 125U 124U 50N 49N w > 123U 51N 50N 126U 125W 124U 49N 123U Range: 1 tick=5.0 km Figure 31. Vertical section of at for section 23. Stations extend south to north from Cape Lazo to Cape Sutil and were occupied 27 June 1985. Contouring interval is 0.5 at. 126U 51N 125U 124U 50N 49N Mi > efi to lO Range: lnck=5.0 km 126W 12SU I24U 123U 51N 123U Figure 32. Vertical section of at for section 24. Stations extend across-channel in northern SG and were occupied 27 June 1985. Contouring interval is 0.5 ot. Chapte r 1: Hydrography steeply down to the north signifying a strong residual flow out of the S G in the upper layer. Section 25 (4 September 1986, F igure 33) is very similar to section 24 made one year earlier during early summer. Surface temperatures increase and surface salinities decrease westward (< 11.5 — 13.5°C and 28.5 - 27.5 p p t ) . Decreased across-channel subsurface isopycnal slopes are evidence of a weaker residual outflow i n Figure 33 (September 1986) than i n Figure 32 (June 1985). In the central S G , Section 26 and 27 (27 June 1985, Figure 34, and 4 September 1986, Figure 35) are essentially the same. Temperature data f rom sections 26 and 27 (Thomson et al., 1985; terHart , 1988) both show generally w a r m water in the western port ion of the section. The presence of this water is likely due to the summer warming of the shallow waters contained in Comox Bay and on the banks off Cape Lazo and the subsequent t idal flushing of these waters into the S G . Temperatures decrease eastwards f rom approximately 17°C in the west to 14°C in the east. Surface salinities s imilari ly increase eastwards f rom < 27 ppt to > 28 ppt . Section 26 (Figure 34) does show surface ot values to be < 19.0) near the mainland. A g a i n , this is due to run-off max ima in early summer. A surface estuarine front exists where the buoyancy layer formed in M a l a s p i n a Strait meets the surface waters of the S G proper (F9: Figure 34). The surface salinity gradient associated w i t h this feature is 1.9 ppt per 10 k m . There is a minimal surface temperature gradient. In Figure 34 there is an obvious sloping of the subsurface isopycnals down to the north indicative of a geostrophic flow out of the channel to the northwest. The downward northerly sloping isopycnals are not evident in Figure 35 (section 27). Once again, a decrease across-channel isopycnal slopes are likely caused by reduced run-off dur ing late summer. B o t h sections, however, display a broadening of the 53 Chapter 1: Hydrography Figure 33 . Vertical section of at for section 25. Stations extend across-channel in northern SG and were occupied 5 September 1986. Contouring interval is 0.5 ct-5 4 Chapte r 1: Hydrography 9.0-10 .0°C isotherm and 29.5-30.0 ppt isohaline on the western (Vancouver Island) side of the transect. This feature is the result of the southern subsurface movement of a cold, saline negatively buoyant created by t idal mixing in D P . Figure 34 shows the deep isopycnals sloping down to the west on the island side of the S G indicating a deep geostrophic flow into the S G . This feature is not observed in section 27. Surface temperatures in the south (section 28) are approximately > 15°C and fairly uniform across the Strait . Surface salinities, however, are directly affected by freshwater input from the Fraser River and they decrease from 20.5 ppt near the eastern shore to 25.5 ppt near the western shore. However, the depth of the buoyancy layer, bounded by the 22.5 at isopycnal, is uniform across the channel (Figure 36) despite the significant across-channel variations in the upper layer itself. The set of four 24-hour stations conducted in the vic ini ty of Cape Mudge Shoals (Figure 2) depicts the strength and horizontal variabil i ty of the t idal jet exi t ing Discovery Passage. Table 4 provides the details on this set of 24-hour sta-tions. Table 4 Details of the set of 24-hour stations that were conducted in the vicintity of Cape Mudge Shoals. The summarized data includes station location, the number of CTD casts, the spacing-in hours-between the casts and the CTD cast numbers. Note that the CTD cast numbers are not sequential. Station location N o . of C T D ' s C T D spacing C T D cast no.'s Discovery Passage 13 2 hr 100101-113 (Station D P ) W i l l o w Point 25 1 hr 100201-225 (Station W P T ) Oyster Bay 13 2 hr 100301-313 (Station O B ) Cape Mudge Shoals 12 2 hr 100401-412 (Station C M S ) 55 126U 51N 125U 50N 49N 124U J, 1 if. > 126U 125U 124U 123U 5IN 50N 49N I23U Range: 1 tick=5.0 km Figure 34. Vertical section of at for section 26. Stations extend across-channel in central SG (Cape Lazo to Powell River) and were occupied 27 June 1985. Contouring interval is 0.5 ot. Chapter 1: Hydrography RfiNGE I (ME TICK c 5.0 KM F i g u r e 35 . Vert ical section of at for section 27. Stations extend across-channel i n central S G (Cape Lazo to Powell River) and were occupied 5 September 1986. Contouring interval is 0.5 o~t. 57 Chapter 1: Hydrography Figure 36 . Vertical section of <r« for section 28. Stations extend east to west across-channel in southern SG and were occupied 5 September 1986. Contouring interval is 1.0 ct. 58 Chapter 1: Hydrography Stat ion D P (Figure 37), occupied at the extreme southern entrance to D P clearly shows the vertically well-mixed waters produced by t idal flows through Sey-mour Narrows being pumped past the station location as the tide floodsjind ebbs (Figure 37). Of further interest is the presence of cold (8 .5°C) , saline (30 ppt) , dense (23.5 Ot) waters underlying a warm (13 .5°C) , brackish (25.5 ppt) buoyancy layer (< 19.0 ot) between the periods when homogeneous water occupies the station site. These observations indicate a cold, saline negatively buoyant plume of water of intermediate density debouching into the S G on the flood tide. These dense waters flowing into the SG represent the formation of a density current that may play a significant role in the formation of intermediate and deep water in the northern S G (Waldichuck, 1957; T h o m s o n , 1981; LeBlond, 1983). These same features were in data collected during the same 24-hour period at Station W P T ( C T D 100201-225), just offshore of Wi l low Point and further south, at Station O B ( C T D 100301-313). The negatively buoyant plume is obvious, but becomes slightly warmer (9°C as op-posed to 8.5°C) and less saline (29.5 ppt as opposed to 30.0 ppt) due to entrainment of upper layer fluid and m i x i n g . It is interesting to note that the appearance of the most dense bottom water occurs when surface stratif ication is strongest (Figure 37). It appears, then, that there was a t idally modulated inflow of this negatively buoyant plume over this one t ida l period and that the greatest propagation speed of the gravity current thus formed (Benjamin, 1968) w i l l occursduring the m i n i m u m in t ida l flow. Geyer and Cannon (1982) report a similar result for a gravity current in Puget Sound where "the nature of t idal forcing is curious in that the low frequency c irculat ion is maximized during min imum t idal flow.". These conclusions were at least part ia l ly substantiated by conductivity-temperature-depth-velocity ( C T D V ) 59 126' 18* \ K uadra •land \ y 8 U t l o n DP * ) \ ( V a n c o u v e r M V . Bbr»I t of G«orgi« 1 ^ 138* 00* BO* 08* 80*00* •40*64' TIME i ONE TICK = 2.0 HR ft Figure 37. Time-depth contour plot of ot for Station DP located at the southern entrance to DP. This station was sampled recursively over the 24 hour period 24-25 June 1986. Contouring interval is 0.5 at. Times of higher high water (HHW) and lower low water (LLW) are shown from which the general duration of the flood and ebb tides may be inferred. Chapter 1: Hydrography data collected during 26 Nov 1986. E b b / f l o o d current velocities (directed parallel to the Vancouver Island shore) were obtained from several C T D V casts stationed axially in the jet. The flood currents were clearly not uniform in direction w i t h depth and generally showed an inflow at depth during all phases of the tide. These (unpublished) data at least qualitatively support the existence of a density current emanating f rom D P . Observations at Station C M S ( C T D 100401-412) off Cape M u d g e Shoals, are in marked contrast to those for Stations D P , W P T and O P . There is an advance and retreat of the dense bot tom waters, but relatively little t ida l movement in the upper layer w i t h the exception of the period during the 12 — 15th hours corresponding to the change from lower-low water to higher-high water. The negatively buoyant plume is found below 20 m at all four stations. 1.4: T-S Characteristics Construct ing T-S diagrams has proven to be an invaluable aid in identifying source water masses and mixing ratios (Neumann and Pierson, 1966; Tolmazin , 1985). A l t h o u g h variabil i ty in the upper layers of the ocean due to direct atmo-spheric forcing makes identification of a surface layer water mass difficult, T-S curves for all C T D data obtained enabled an envelope of possible surface layer T-S curves to be determined. A reasonable average from w i t h i n these envelopes was selected and considered as the surface water mass for S G surface ( S G s f c ) , JS surface ( JS s f c ) and Q C S T surface ( Q C S T s f c ) waters. The T-S curves clearly show the two layer structure of the entire inside pas-sage. Salient features of the physical oceanography are also easily identified. Surface 61 Chapte r 1: Hydrography water becomes progressively colder and more saline seaward; deep lower layer wa-ters become warmer and less saline as they progress up-channel. In addition, the penetration at depth of oceanic waters into the S G is evidence of the major impact continental shelf processes can have on the general estuarine circulation and the ventilation of deep waters throughout the study area (P ickard , 1975; Geyer and C a n n o n , 1982; Farmer and Freeland, 1983). Differences in the T-S characteristics and two layer structure between the two years were negligible (Figure 38). M i x i n g ratios were estimated for the waters of D P , JS and Q C S T . These esti-mates are shown in Table 5 as well as T -S characteristics of the source water masses for each area. M i x i n g ratio estimates were calculated by solving three simultaneous algebraic equations given as: 1 = 7l + 1 2 + 73 Tv = iiTi + i 2 T 2 + I3T3 ST, = iiSt + 7 2 S 2 + I 3 S 3 where T^ and 5,, represent the T-S characteristics of a part icular water mass, and 71,2,3) ^ 1 , 2 , 3 and S i j 2 )3 represent, respectively, the m i x i n g ratios and T-S charac-teristics associated wi th the three source water masses. Errors i n 71,2,3 are approx-imately ± 1 0 % and arise due to ambiguities in the choices of T and S for specific water masses. 1.5: Rotational Effects Sloping subsurface isopycnals were a feature of many of the across -channel sections made in both years (Figures 12 and 13, 17-21 and 32-36). When this 62 Figure 38. Composite TS curves for the entire study area for both 1985 and 1986. Water masses (1 - SG s f c , 2 = SG d p , 3 = DP, 4 = JS s f c, 5 = JS d p , 6 = QCST s f c and 7 = QCST d $ are shown adjacent to their characteristic TS curves. Chapte r 1: Hydrography Table 5 Parameters and results of the T-S analysis made in §1.5. Summary includes the water mass, source water masses (subscripts 'sfc' and 'dp' refer to surface and deep layers respectively), T-S properties of the source water masses and the derived mixing co-efficients. Mixing co-efficients are approximately ± 1 0 % . Year Water Mass ° C / p p t Source Water Masses T-S properties °C/ppt M i x i n g ratios % 1985 D P 9.8/29.8 SG s f c SGdp JS sf c 10.5/29.2 8.6/30.2 9.0/31.0 56 23 21 JS sf c 9.0/31.0 D P QCSTsfc Q C S T d p 9.8/29.8 8.8/31.55 7.0/33.1 41 49 10 1986 D P 10.4/29.6 SGgfc SGdp J Ssfc 11.5/29.0 8.5/30.5 9.4/31.4 62 34 4 J Ssfc 9.4/31.4 D P QCST s fc Q C S T d p 10.4/29.6 10.7/32.1 7.4/33.1 42 22 36 phenomenon is present in space-varying (along-channel) two-dimensional data, such as that presented here, it indicates the presence of three-dimensional processes. Oceanic length and time scales associated w i t h the residual c irculat ion of the inside passage imply that the three-dimensional characteristics of this circulation may be part ly associated w i t h rotation. However, in narrow channels, three-dimensional effects may arise due to channel curvature, lateral f r ic t ion , turbulence, topography or time-dependent effects (Young and Hay, 1987). For the impl ied motions to be considered geostrophic, the f luid must be assumed to be invisced, time invariant and hydrostatic . The importance of rotation can be demonstrated by comparing the channel w i d t h w i t h the internal Rossby radius of deformation associated w i t h the baroclinic 64 Chapter 1: Hydrography modes of mot ion. Given the two layer structure of Q C S D , Q C S T and the S G , the internal Rossby radius, r 2 , for a two-layer system is (LeBlond and M y s a k , 1978): where gt, the reduced gravity is: 9 , = 9 ( £ i _ ^ i i . ( 1 . 2 ) P2 The subscripts 1 and 2 refer to the upper and lower layers respectively, H to the layer depth and p to the layer density. Table 6 shows the calculated r 2 for all across-channel sections. In a l l cases, the local channel width is greater than the local r 2 and rotational effects may be considered of some importance. However, it is dangerous to conclude that the implied motions are strictly geostrophic! Nonetheless, when rotation can be considered important , simple geostrophic calculations can give some, albeit l imi ted , insight into the implied motions. Geostrophic shear in a two layer model is calculated as: <"> where dHi/dy is the across-channel slope of the interfacial isopycnal and the lower layer is considered at rest. Thus , (1.3) gives an estimate of the residual flow in the surface layer if the assumption that lower layer velocities are negligible is reasonable. Table 6 shows the calculated upper layer velocities, V, for across-channel sections w i t h identifiable sloping interfacial isopycnals. Direct current measurements made wi th in the study area are in reasonable agreement w i t h estimates of V f rom (1.3). A mooring in the vic ini ty of section 11 (Figure 17) gave mean currents at 15 m of 13 cm/s seaward (Thomson, 1976), 65 Chapte r 1: Hydrography Table 6 Geostrophic shear calculations using a two layer model. Channel width (km), inter-nal Rossby radius of deformation, r 2 (km), interfacial isopycnal (ot) and geostrophic shear (cm • s _ 1 ) are shown. Only those sections where subsurface isopycnals were obviously sloping across-channel were considered. Section Channel T2 Interfacial V ( c m - s - 1 W i d t h isopycnal 6 27.5f 5.7 24.5 18.7 7 16.0| 5.0 25.5 -18.5 11 17.0 4.2 24.5 12.6 12 17.0 5.5/8.4 24.5/25.5 14.2/-18.7 13 23.0 4.3 24.5 9.8 14 15.0 2.8 24.5 7.2 15 16.0 3.6 24.6 10.5 24 16.0 4.5 23.0 40.0 25 16.0 4.4 22.5 16.1 26 25.0 13.4 22.5 13.4 t This distance represents the length of the section itself as the channel width of Q C S D is in the order of hundreds of kms. % The w i d t h of the channel in this instance is given by the distance separating the 60 m isobaths. compared to 12.6 cm/s as per Table 6. D a t a f rom section 26 (Figure 34) yielded surface currents of 13.4 cm/s compared to 10 cm/s obtained f rom drift bottles (Waldichuck, 1957). It should be noted that the net volume transport is very difficult to quantify as it is subject to large errors (Thomson, 1977; Godin et al., 1981). The upper and lower layer volume transports are neary equal and of opposite sign (LeBlond, 1983). Thus , the net volume transport is small and of the order of the errors associated w i t h the measurements. Furthermore, the residual circulation is sensitive to me-teorological conditions and shelf processes thereby introducing errors into residual transport measurements (Frisch et al., 1981; G o d i n et al., 1981). Calculations of geostrophic shear f rom single across-channel sections are, therefore, not expected 66 to portray completely the residual c irculat ion. Chapte r 1: Hydrography 1.6: Summary D a t a collected during the cruises of the research vessels C S S Vector (26-29 August 1985, 29 M a y 1986 and 23-25 June 1986) and Parizeau (3 September 1986) were used to compile a general hydrography of the waters ly ing between Vancouver Island and B C ' s mainland coast. This hydrography was in turn used in an analysis of the return migrat ion routes of adult homing sockeye salmon (Oncorhynchus nerka) through these waters. Positive estuarine circulation dominates the physical oceanography of this region. W a r m , relatively fresh water moves seaward on the surface and cold, saline water intrudes at depth. This circulation was observed to be stronger in spring and early summer, when fresh water input into the system was greater, than in later summer. The four oceanographic regimes identified in Thomson et al. (1985) and terHart (1988) remained distinct: the sl ightly stratified central and northern portions of the S G resulting from a buoyancy layer composed of fresh water run-off principal ly f rom the Fraser River, and to a much lesser extent f rom mainland inlets; the vertically well -mixed waters of D P due to vigorous t idal agitation; the weakly stratified waters of JS ; and the less weakly stratified waters of Q C S T caused by the presence of JS water on the surface and the intrusion of co ld , saline Pacific Ocean water at depth. These four regimes are separated by t idal ly mixed fronts created by vigorous m i x i n g of the entire water column by rapid t i d a l streams in the shallow and constricted regions of Seymour Narrows, Blackney Pass and Weynton 67 Chapte r 1: Hydrography Passage. S imi lar fronts were found east and west of Nahwit t i Bar . A surface front separating the well mixed water of eastern JS from the weakly stratified water of western JS was observed near Kelsey Bay. A surface estuarine front was observed in northern Q C S T where the buoyancy layer exiting Q C S T meets the upwelling enhanced oceanic waters of Q C S D . A s imilar estuarine front was found in the S G where the buoyancy layer formed in M a l a s p i n a Strait and enhanced by restricted lateral spreading debouches into the S G . A set of three 24-hour stations in the vic in i ty of Weynton Passage and a set of four 24-hour stations in the vicinity of Cape M u d g e show the strength and t idal variabi l i ty of the flows and m i x i n g processes over a diurnal t i d a l cycle. Dense water masses formed in these regions are seen to exude into western JS and northern SG respectively during all phases of the tide. These processes may play an important role in the formation and renewal of intermediate and deep water in both basins. The most dense waters were present during minimal t ida l flows and the resultant density currents are expected to be strongest during these periods. T-S diagrams clearly show the two layer structure of the entire inside passage as well as the relative variabi l i ty in the upper and lower layers. The upper layer is clearly seen to become more saline seaward; s imilari ly , bottom waters become progressively fresher as they progress up-channel. Calculated mixing ratios give rough estimates of the proportions of S G surface and deep water present in JS . 68 Chapte r 2: Ultrasonic Track ing C H A P T E R 2 U L T R A S O N I C T R A C K I N G : H O R I Z O N T A L A N D V E R T I C A L D I STR IBUT IONS 2.1: Introduction The return migration of Pacific salmon (Oncorhynchus spp.) is a complex, fascinating and much debated phenomenon. The return migration of adult sockeye salmon, which may traverse distances as great as 6000 k m (French et al., 1976) and ends w i t h remarkable spatial and temporal precision (Royce et al., 1968) in the salmon's natal stream, is a classic example of animal migrat ion. Quite apart from the academic interest associated w i t h this migration, the returning adult sockeye have a tremendous worth as a commercial fishery. Knowing the sockeye's return migrat ion mechanism and understanding its choice of routes is clearly of great value to both the scientific commmunity and the commercial fishery. The return migration itself may be thought of in three phases: oceanic, coastal or estuarine and riverine. Oceanic migrat ion mechanisms are, as yet, undetermined. Riverine migratory behaviour is generally considered to be directed by olfactory information learned years earlier in freshwater (Johnson and Hasler, 1980; Hasler and Scholz, 1983). The mechanisms responsible for the estuarine phase of the return migrat ion must necessarily be a combination of both the oceanic and riverine mechanisms as estuarine environments represent the interface between oceanic and riverine environments. A persistent, unavoidable large scale change in the sockeye's immediate environment during the return migration must require a similar scale change i n guidance mechanisms. Moreover, the returning sockeye begin to undergo 6 9 Chapte r 2: Ult rasonic Tracking profound changes in morphology, physiology and behaviour that prepare them to enter freshwater and spawn. Given the changing oceanographic conditions and the changing condition of the animal itself, it is not surprising that the coastal phase of the return migration is poorly understood. To further complicate issues, how migrat ing salmon integrate diverse orientation cues, such as celestial (Groot, 1965; Brannon , 1972) and magnetic (Quinn , 1980; Q u i n n and B r a n n o n , 1982) is not known. Nonetheless, as migratory cues abound in the salmon's immediate hydrographic environment (Tully et al., 1960; Royal and Tul ly , 1961; Favorite, 1961; Royce et al., 1968; Leggett, 1977; M y s a k et al., 1982; Q u i n n , 1982; Groot and Q u i n n , 1986), it is only natural to consider the horizontal and vert ical movements during migration to the hydrography. To this end, oceanic temperatures (Mysak et al., 1982, I P F S C , 1984; H a m i l -ton , 1984; Burgner, 1980; B lackbourn , 1987; Groot and Q u i n n , 1987) and Fraser River run-off (Favorite, 1961; Wicket t , 1977) have been posit ively correlated with sockeye return migration routes and t imings. These correlations, in part , motivated M O I S T . To study the factors guiding migrations in coastal waters, ultra-sonically tagged sockeye were tracked w i t h a directional hydrophone while the physical oceanographic hydrography was specified by recording vertical T and S profiles. Currents , which may significantly affect the fish's horizontal and vertical distr ibu-t ion (Brett , 1983; A r n o l d and Cook , 1984; Dodson and Dohse, 1984), were measured along the telemetred track by a Lagrangian drifter. The collection of these data was intended to help illustrate the relationship between the migrat ing salmon's horizontal and vertical movement and distr ibut ion and its immediate environment. 70 Chapter 2: Ul t rason ic Tracking 2.2: Data Collection and Processing terHart et al. (1987) and terHart and Q u i n n (1989) provide compilations of the complete data set for 1985 and 1986 respectively. A complete description of the capture and tagging of individual salmon as well as the logic behind the chosen methods is provided in Quinn and terHart (1987). Thus , methodology w i l l not be discussed in detai l here. The ultrasonic tag and receiver/decoder unit were manufactured by Vemco, L t d . , Nova Scotia. The transmitters were 74 m m x 16 m m , weighed 28 grams in air (13 grams in water) and, for fish tracked in 1985, emitted pulses at one of five frequencies (50, 60, 65.5, 69 or 76.8 khz) . Interference with commercial ly available depth sounders operating at a frequency of 50 khz resulted in only one transmitt ing frequency (76.8 khz) being employed in 1986. The pulse rate was proportional to depth and prior cal ibration to a depth of 50 m resulted in accuracies of ± 1 m. The M e l i b e , an 11 m workboat, was used as the tracking vessel and as such, uti l ized an externally mounted directional hydrophone (Vemco, L t d . ) that could be swung into or out of position approximately 1 m below the water's surface and manually rotated through approximately 300° . The receiver/decoder was mounted inside the cabin of the Melibe and provided a continuous digi tal readout of depth. F i s h depth was recorded manually every minute, provided the unit was receiving, and ship's posi t ion, determined by a L O R A N C (Raytheon, L td . ) receiver, was recorded manually every 5 minutes. The L O R A N C unit was cal ibrated periodically using radar fixes and local topographic features. The accuracy of the position finding system was dependent on the location of the vessel and was generally found to be better in open water than in constricted channels. The range of the transmitter 71 Chapter 2: U l t rasonic Tracking was such that the posit ion of the tracking vessel closely approximated the position of the fish. The N i t i n a t Queen, a 12 m self-propelled steel barge, was the platform from which the physical oceanographic measurements were made. The l imited manuever-abil i ty and speed of this vessel, as well as logistic constraints, made this a somewhat less than ideal choice. A s a result, the physical oceanography was conducted aboard the K e t a , a 14 m converted purse seiner, during 1986. The same physical oceanographic instrumentation used during the baseline cruises was employed during the salmon tracking. C T D casts were made at 15-45 minute intervals along the fish's track and extended from the surface to 60 m depth or to wi th in 10-15 m of the bot tom if the depth was less than 60 m . The C T D data were despiked and averaged over 1 m depth intervals. Sigma-T (OT) was calculated using the International Equat ion of State, 1980 (Pond and P i c k a r d , 1983). Header data consisting of fish number, consecutive C T D cast num-ber, local time ( P D T ) , date (day, month , year) and position (lat, long) were added for each track. D a t a that were contaminated by an intermittent failure in the tem-perature c ircuitry dur ing 1985 were removed manually. This problem did not occur during 1986. Current measurements were made using surface drifters w i t h passive window bl ind drogues. Cross-sectional area of the drogues was approximately 6 m 2 . In 1985, these drogues were rigged at depths of 7.5 m and 30.0 m (measured from the surface to the center of the drogue) and deployed or recovered from a 12 ft Zodiac inflatable boat. A s many as 4 drifters were deployed at a time and placed so that the fish's most likely track would pass through or at least near the current drifter. 72 Chapter 2: U l t rason ic Tracking T h e ini t ia l distance separating drifters rigged at the same depth was approximately .2-.5 nautical miles. Drifters were recovered when their position no longer approx-imated the fish's location or when the distance from the Nit inat Queen precluded an accurate determination of position. Radar was used to determine the posit ion of the Nit inat Queen and conse-quently, the current drifters. Thus, to determine a drifter's posi t ion, the ship's posit ion would be fixed by taking radar bearings and ranges from the local topog-raphy and then a radar range and bearing would be determined from the ship to the drogue. The posit ion of the C T D stations were s imply those of the ship. Accuracies i n the posit ion of the ship were likely ± 5 0 - ± 1 0 0 m , depending on the distance from the shore and the nature of the topography. Accuracies in the drogue's position, then would be of the order ± 5 0 - ± 2 0 0 m , depending on the distance from the ship and the care taken to produce the fix. In 1986, current drifters were rigged for a depth of 7.5 m and deployed in arrays consisting of 3 drifters only. The drifters were deployed from and recovered to the K e t a . Dri f ter positions were established by positioning the ship alongside the drifter and using the ship's L O R A N C (Raytheon, L td . ) receiver. T h i s resulted i n greater accuracies in the drifter positions as well as a savings in manpower and increased safety margins during deployment and recovery. C T D station positions were also determined by L O R A N C in 1986. The aerodynamic and hydrodynamic drag on the component parts of the current drifter is negligible compared to that of the drogue (Buckley and P o n d , 1976). Thus , the drogue can be said to be experiencing only the current at whatever depth it has been rigged for and is integrating over. There was, therefore, no 73 Chapter 2: Ultrasonic Tracking correction made to the data for the depth of the drogue. 2.3: Horizontal Movements 2.3.1: General Overview A n analysis of the tagged sockeye's horizontal movements in relation to the observed T , S and currents for the two years of tracking data is presented i n Quinn and terHart (1987) and Q u i n n et al. (1989). A brief overview only w i l l be provided here in order that the vert ical and horizontal movements may be compared and contrasted. A total of 33 tracks were obtained during the summers of 1985 and 1986. Twenty-eight sockeye were tracked with fully functioning ultrasonic tags for suffi-cient t ime to characterize the fish's vertical and horizontal distributions. C T D data were not collected for 3 fish tracked in 1986 (8607t, 8608 and 8609) due to logistic constraints. Table 7 (adapted from Quinn et al., 1989) summarizes the tracking data for both years. W h i l e correlations of sockeye migratory routes (Groot and Q u i n n , 1987; X i e and Hsieh, 1989) and t i m i n g (Burgner, 1980; B lackbourn , 1987) w i t h oceanographic variables helped motivate M O I S T , a relation between horizontal movement and dis-t r ibut ion and T a n d / o r S has, to date, proved very elusive. In general, horizontal movements were extremely variable, far more so than the genera] patterns of ver-t ical movements, and were apparently controlled by different factors. It was clear, t The convention used throughout when referring to a particular fish track is such that 8607 refers to the 7 t h fish (07) tracked in 1986 (86). Thus , 8516 refers to the 1 5 t h fish tracked in 1985 and so on. 74 Chapter 2: U l t rasonic Tracking Table 7 Location, duration, direction, speed of movement and depth of travel of adult sock-eye salmon tracked in 1985 and 1986. Net direction refers to the compass direction from point of release to the point where the fish was abandoned, lost or recaptured. Net distance travelled is determined in an analogous manner to that of net direction. Track Region Durat ion Net Water S p e e d y Net Average Depth (hours) direction (km/h) distance (meters) 8503 JS 14.83 101° 2.47 13.50 6.1 8504 Q C S T 9.45 63° 2.85 0.47 9.1 8505 JS 6.02 20° 2.00 2.47 8.6 8507 D P 3.57 136° 2.15 1.50 20.3 8508 D P 2.93 2° 2.66 1.99 19.9 8509 D P 12.33 212° 3.28 1.11 11.8 8510 D P 4.98 195° 1.38 4.43 11.7 8511 S G 7.83 299° 2.13 12.45 18.0 8512 S G 10.00 111° 1.31 1.20 21.3 8513 S G 2.93 125° 1.16 6.42 6.7 8514 S G 9.82 112° N . A . 13.16 11.3 8515 S G 7.02 110° 3.42 16.15 N . A . 8516 S G 12.50 108° 3.82 34.00 14.9 8517 S G 6.75 274° N . A . 7.76 N . A . 8601 Q C S T 21.38 294° 1.84 5.52 14.6 8602 Q C S T 9.33 127° 3.89 28.04 15.2 8603 Q C S T 10.00 303° 1.18 10.32 19.9 8604 Q C S T 6.50 145° 1.52 2.15 9.1 8605 Q C S T 6.00 277° 1.20 3.61 16.0 8606 Q C S T 33.00 105° 2.43 43.24 7.5 8607- JS 8.00 111° N . A . 16.89 17.8 8608* Q C S T 18.00 125° N . A . 10.98 14.3 8609* JS 13.08 302° N . A . 10.83 13.2 8610 S G 13.07 113° 3.69 28.72 11.2 8611 S G 9.25 351° 1.15 5.87 17.8 8612 S G 7.50 128° 2.51 11.95 17.8 8613 S G 22.08 35° 1.78 12.45 19.5 8614 S G 26.05 126° 1.42 24.63 28.6 8615 S G 9.00 102° 2.64 12.57 16.7 8616 S G 8.33 354° 1.87 4.70 13.4 * Concurrent C T D data were not obtained for this track. t Water speeds are estimated swimming speeds of the fish in the water, derived from ground speeds and estimates of current speeds. X N . A . implies that data were not available. 75 Chapter 2: Ultrasonic Tracking however, that the coastal phase of the homeward migration was less well oriented than the oceanic phase (Quinn and terHart , 1987). The majority of tracked fish d id display a strong directional preference, either southeastward or northwestward, dur ing some parts of the observed track. Exceptional fish displayed apparently ran-dom orientation (8509) or almost no movement whatsoever (8612). It is noteworthy that no sockeye made significant progress to the northeast or southwest. Current data obtained in the vicinity of the telemetred track helped resolve a highly significant, asymmetrical bimodal orientation ( S E / N W ) in the tracking data. W h e n the sockeye's progress over the ground was considered the resultant of an advecting current and the derived fish's speed and direction, it was shown that the tracked sockeye maintained some directional preference and did not compensate for current displacement. S imilar current analysis also showed a weak but non-random fish orientation wi th respect to the observed currents. Biased bimodal orientation was evident when southeastward moving fish en-countered topography that obstructed the fish's path (8503, 8606, 8610). The ob-struct ion tended to cause the sockeye to swim aimlessly in the near vicinity of the obstruction, then reverse direction and eventually, by swinging around to the nor th , resume a southeastward heading. M a n y fish caught and released near ob-structions headed to the northwest immediately (8504, 8511, 8613). Biased bimodal orientation has been documented for schools of sockeye smolts (Groot , 1965; 1972), laboratory studies of juvenile sockeye (Groot , 1965; Quinn and B r a n n o n , 1982), C h u m salmon (Oncorhynchus ktta)(Quinn and G r o o t , 1984b) as well as many other animals (e.g. reptiles: L a n d o t h , 1973; birds: B i n g m a n , 1981). Analysis of the horizontal movements also showed that there was a much 76 Chapter 2: U l t rasonic Tracking stronger southeastward bias in fish direction in the S G than in Q C S T , JS or D P . G i v e n the bimodal orientation of fish tracked in S G , Q C S T and JS , it is noteworthy that the average speeds were faster in the homeward (southeastward) direction than in the seaward (northwestward) direction. N o significant difference in overall average swimming speed was found in different oceanographic regimes. In summary, it appeared that the increased stratification in the S G may have aided orientation towards the Fraser River and that a general asymmetrical bimodal orientation (southeastward/northwestward) in Q C S T , JS , and D P gave way to a unimodal southeastward orientation in the S G . A t some point, olfactory stimuli and rheotactic responses characteristic of upriver migration (Johnson and Hasler, 1980) must supercede the mechanism responsible for the observed orientation in these data. A l t h o u g h near surface waters, especially in the southern S G , contain relatively large volumes of water originating from the Fraser River , no clear evidence of this change in migratory mechanisms was observed. 2.4: Vertical Movements 2.4.1: General Overview Vert ical distr ibution of sockeye tagged in 1985 is discussed in Q u i n n and ter-H a r t (1987). A brief summary of vertical distr ibut ion for both the 1985 and 1986 tagging studies is presented in Q u i n n et al. (1989). Both studies deal quantita-tively w i t h the sockeye's depth distr ibution but only qualitatively w i t h the relation between depth and T or S. In order for subsequent, more quantitative observations of the sockeye's depth dis tr ibut ion to be made w i t h respect to the vertical oceano-graphic structure, the results obtained in Q u i n n and terHart , (1987) and Quinn et 1 1 Chapte r 2: Ultrasonic Tracking al. (1989) must be introduced. In comparison, patterns of vertical movement for 1985 and 1986 in all 4 oceanographic regimes were similar and much less complicated than the correspond-ing horizontal patterns of movement. In general, many salmon dived abruptly im-mediately upon release, stayed at considerable depths (30-60 m) for periods up to 80 minutes and then ascended to the surface. Of those sockeye that made deep ini t ia l dives, subsequent vertical movement differed substantially and was easily distinguishable from this init ial dive. This type of ini t ia l vertical distribution was observed more often in Q C S T , JS and D P than in the S G . Subsequent patterns of vertical movement were characterized by long periods spent w i t h i n a narrow range of depths, interspersed w i t h brief, shallow dives or ascents. Only a few sockeye repeatedly dove and ascended and only one, 8511 did so continuously. Die l patterns of movement were obtained for 5 fish tracked in 1986. Estimated swimming speeds were generally slower during the night than during the day and during certain night-t ime periods, the fish were apparently dri f t ing ( L O R A N C positions indicated that the tracking boat was moving but no propulsive power was required to keep w i t h i n range of the transmitter) . In addi t ion, diel patterns of vertical dis tr ibut ion showed that during night-time periods, tracked fish were observed at depths significantly closer to the surface than during day-time periods (Quinn et al., 1989). Qual i tat ive observations suggested that orientation and depth distribution were related. A disoriented fish is defined as one that did not make significant progress homeward (in all cases assumed to be the Fraser River ) . Invariably, fre-quent large ampli tude vertical excursions were associated w i t h disoriented horizon-78 Chapter 2: U l t rasonic Tracking ta l movements. The term disoriented shall be used to refer to either poor horizontal orientation or vertical distr ibutions characterized by frequent large amplitude ex-cursions. The term oriented shall be used to refer to fish that made significant progress homeward during al l or part of their track or to vertical distributions characterized by infrequent small amplitude vertical excursions. Figure 39 (from Q u i n n and terHart , 1987) clearly shows this difference in oriented versus disoriented behaviour. Sockeye tracking conducted in the vertically and horizontally well-mixed wa-ters of D P was considered the experimental control. The absence of even weak vertical structure leaves the fish with very l imited choices of T or S and virtual ly no choice of T or S gradient. The fish's vertical distribution in this type of oceano-graphic regime can therefore be considered independent of a preferred T or S or gradient. Tracking observations in D P and eastern JS clearly showed a general preference for near surface waters (Figure 40). Sockeye tracked in the northern regime ( Q C S T , JS and D P ) spent significantly more time with in 10 m of the surface than did sockeye in the S G . F i sh in the north also displayed a greater tendency to dive not only immediately after release, but also throughout the duration of the track. The fish in this region were rarely observed at depths greater than 30 m and as the thermoclines and haloclines extended to depths of 20-30 m (Figure 41), sockeye occupied depths in the vicinity of the largest gradients in T and S. The more complex T and S stratification in the SG seemed to elicit more complex patterns of vert ical distr ibution and movement. Minute by minute analysis of the depth data revealed that sockeye tracked in the S G spent less t ime in the 79 Vertical movomonts of 8316, trockid in Gtorojia Strait ntar th« Frator Rivtr on August 22,1983. A. STEAOY 0 PROGRESS £S£ H36-I333h B. EXCURSION INTO HOWE SOUND I336-I530h a. u i a C. RESUMPTION OF PROGRESS £ I33l-I730h 10 20 0 K> 20 30 0 fO-20-Figure 39. Oriented and disoriented vertical movements for 8516. Depth is shown on the y-axis and time on the x-axis. Chapter 2: Ultrasonic Tracking NUMBER OF OBSERVATIONS 9 10 ?o 30 40 so SAUNITT 10/00) x n z» * v v TEMPERATURE tDEG.C) 10 1,1 I ? 18 \* © co riSH T W O B 0 7 NUMBER OF OBSERVATIONS 0 10 <B 30 40 90 SAUNITT (0/001 X V V 9 * V TEMPERATURE IDEG.C) 1,0 1,1 » 1,3 14 •5 & Q 8* o to riSM TwcK eo8 NUMBER Of OBSERVATIONS « « 0 1 » 180 |40 300 SAUNITT t0/00l » 7 7 39 30 3? 33 TEMPERATURE IDEG.CI 10 \\ 12 13 14 ri3M T M C X era NUMBER OF OBSERVATIONS 0 K 40 to K 100 SRLlNlTT 10/00) I' » 8 8 i i 8 TEMPERATURE 10EC.C) » 10 U 1? 13 rw B 1 0 Figure 40 . Composite fish depth histogram and T and S profiles for 8507, 8508, 8509 and 8510. Tracks were conducted in the homogeneous waters of eastern JS and D P . 81 Chapter 2: Ultrasonic Tracking 2.1 NUMBER OF OBSERVATIONS 75 150 S i 300 3 7 5 SALINITY ( 0 / 0 0 1 2 6 2 9 3 0 31 3 2 TEMPERAT'JRE IPEG.C) NUMBER OF OBSERVATIONS 0 3 5 7 0 105 1 « 175 SALINITY ( 0 / 0 0 1 2 4 26 26 30 3 2 3 4 TEMPERATURE IDEG.C) NUMBER OF OBSERVRTIONS P SO 1.00 1,50 200 250 SALINITY (0/001 2 7 28 2 9 30 31 3 2 TEMPERATURE (DEG.C) in 1* 9 J2_ 11 12 1.3 8-F1SH T W O 8 5 0 5 F i g u r e 4 1 . Composite fish depth histogram and T and S profiles for 8503, 8504 and 8505. Tracks were conducted i n the slightly stratified waters of QCST and western JS. 82 Chapter 2: Ultrasonic Tracking top 10 m of the water co lumn than d id their northern counterparts even though the average depth of travel d i d not statistically differ. This result is explained by the frequent deep dives and ascents made by sockeye tracked in the north which served to increase the average depths of several fish tracked for short periods of t ime. Sockeye tracked in the S G tended to avoid the warmer, less saline (< 27 ppt) near surface waters (Thomson et al.,1986) and preferred depths directly above or below the m a x i m u m gradients. A l t h o u g h there was a general avoidance of near surface waters, frequent, brief excursions (tens of minutes) were made into these waters. T h e only fish that spent a significant portion of time in waters > 16°C did so as part of a sequence of 28 dives and ascents. It should be noted that in this case, salinities in the surface mixed layer were > 27 ppt . In general, vertical profiles of T and S in the S G were characterized by th in thermo/halocl ines and relatively strong T and S gradients (Figure 42). There were several exceptions, however, and the vertical distr ibution of sockeye tracked in these exceptional regimes w i t h i n the S G resembled the vertical distribution of sockeye tracked in Q C S T and JS in that there was a relatively broad range of observed depths that extended throughout the thermo- and haloclines. The overall results generally supported Westerberg (1982; 1984) and Ichihara and Nakamura (1982) who proposed that vertical distr ibut ion and movement was related to a well defined T and S gradient. The near surface orientation of sockeye tracked in 1985 in all oceanographic regimes was not as evident i n 1986. T and S structure was essentially similar for both years and was not likely the cause for this difference (Figure 42). It has been proposed ( T . P . Q u i n n personal communication) that the genetic differences inherent 83 Chapte r 2: Ul trasonic Tracking in sockeye returning to different river systems may be responsible for the observed differences in depth of travel. 2.4.2: A m b i e n t O c e a n o g r a p h i c V a r i a b l e s The remainder of this study w i l l deal wi th those tracks for which concurrent C T D data were obtained. Twenty-five of the 28 tracks considered by Quinn and terHart (1987) and Q u i n n et al. (1989) meet this criteria (Table 7). In addition, data from 8614, 8615 and 8616 were not considered. Hence, the remaining analysis is based on data from 22 separate fish tracks. The following discussion briefly addresses the importance, possible use and mechanism whereby the physical oceanographic structure may aid orientation by homeward migrat ing sockeye. As such, it w i l l also serve to motivate the remainder ,of this analysis. Orientat ion requires that the fish observe some directional quantity in its en-vironment. The hydrography can provide directional cues through horizontal gra-dients of scalar properties sensible to the fish. Some of the hydrographic dependent orientation mechanisms that have been proposed are (Brannon, 1982): i) Opt imiza t ion of physiological conditions; ii) Populat ion specific pheromones; iii) Selective tidal-stream transport ; iv) Olfaction-mediated t ida l rheotaxis; and v) Olfact ion at stratified depths. 84 Chapter 2: Ultrasonic Tracking 0 ) Salinity (0/00) 18-0 21-0 24-0 27*0 300 330 • * Temperoture (*C) 50 8-0 IK> 14-0 17-0 20-0 (b) Solinity(0/00) 18 0 21-0 240 270 30-0 330 i • * * . . . . . . Temperature (*C) 50 80 IIO 140 170 20O Figure 42 . Composite fish depth histogram and T and S profiles for fish tracked in 1985 and 1986 ( (a) and (b) respectively) in QCST/JS and the SG. Dashed lines represent ±1 standard deviation. There were 3,175 (7,414) observations of fish depth and 128(74) CTD casts in 1985(1986) . 85 Chapter 2: Ultrasonic Tracking Detai led, quantitave analysis of the relationship between T , S and D and the tracked sockeye is necessary in order to determine how the hydrographic structure affects orientation and, ult imately, migrat ion. The large horizontal scale and long time scale hydrographic features that are of the same order of magnitude as the length and time scales of the complete return migrat ion cannot, of course, be directly sensed by the migrat ing sockeye. In fact, the hydrographic near field of the fish is dominated by fine- and micro-structure features that are at least an order of magnitude larger than the mean vertical gra-dients, which are themselves at least an order magnitude of greater than the mean horizontal gradients. In many cases, vertical profiles of hydrographic parameters, such as T and S, contain 'steps': nearly homogeneous layers are separated by large gradients (Muench et al., 1990). the distr ibution of other solutes, and in part icu-lar directional olfactants, are concomitant to that of temperature and/or salinity (Westerberg, 1984). The magnitude of the solute gradient at a particular T or S step need not necessarily be related to the magnitude of the T or S gradient. Moreover, as solutes diffuse more slowly than heat, olfactant micro-structure can be present after the temperature structure has diffused to homogeneity. Vert ical variability in the ocean's upper layers is the result of a large number of complex processes. Near surface irreversible fine- and micro-structure can be related to atmospheric forcing, shear-induced turbulence and double-diffusive salt-fingering and layering. Reversible fine- and micro-structure, associated w i t h the modulat ion of the density field by internal wave fields, is always present in an externally and/or internally forced stratified fluid (LeBlond and M y s a k , 1978). A well known feature of the world's oceans is the aspect ratio of hydrographic 86 Chapte r 2: Ultrasonic Tracking features; the horizontal extent of these features are S> than their vertical extent ( G i l l , 1982). Gravity and stratif ication are resposible for greatly extending some oceanographic phenomeon, such as specific water masses, in the horizontal and com-pressing them in the vertical . Hence, surfaces of equal density and other properties tend to be almost horizontal . In many cases, the thinness of these layers relative to their horizontal extent is such that the vertical stratification can mirror the hor-izontal distr ibut ion of hydrographic parameters. In part icular , a single dive can take a fish through hydrographic regimes corresponding to a basin-wide horizontal dis tr ibut ion. The spreading and diffusion of a passive tracer in the oceans is also strongly affected by the presence of fine- and micro-structure features (Ewart and Bendiner, 1981). Tracer diffusion experiments show that a point source in a homogeneous layer is ini t ia l ly spread by weak turbulence in the layer. Spreading wi th in fine-and micro-structure gradients above and below the homogeneous layer is by slow vertical diffusion and shear dispersion. Filaments of the tracer trail down-shear w i t h intermittent turbulence leaking the tracer into other homogeneous regions. A complex regime develops w i t h several levels delineated by the tracer. The net result is that the extent of the t ra i l ing filaments is much larger than the bulk of the tracer. The analogous oceanic aspect ratio is preserved. The vertical distr ibution of odour f r o m a large-scale, diffuse and continuous source, like the Fraser River , results in homogeneous layers that are filled w i t h odour of differing concentrations. The vertical distribution is concomitant w i t h that of T and S. Vert ical current shear has been observed to be concentrated in the density gradient layers (Woods, 1968; S impson, 1975 ; V a n Leer and Rooth , 1975) and 87 Chap te r 2: Ultrasonic Track ing therefore strongly affects fine- and micro-structure features present there. Any hor-izontal variat ion in the layer w i l l be stretched in the direction of the shear. This effect is integrated in time so that even a weak shear w i l l have a noticeable effect in elongating these structures in the down-stream direction. S imilar i ly , vertical varia-tions w i t h i n the gradient layers w i l l be t i l ted and elongated in the horizontal . Hence, vertical gradients can become part of the horizontal fine- and micro-structure. In estuarine oceanographic regimes, the residual estuarine circulation may in-directly provide orientation cues. A homeward migrating salmon could orient to the river mouth by swimming aainst the shear as mirrored in fine- and micro-structure gradient layers. The detection of these structures could be by olfaction or by d i -rectly sensing the temperature field. However, it is not clear how the sign of the shear, or more directly, the T , S or odour source, can be determined. In a two-layer estuarine model , which is appropriate for most of the southern inside passage, the thermo- and haloclines as well as the greatest shear in the residual circulation are co-incident. Hence, fine- and micro-structure features w i t h i n the thermocline should have a good correlation to the direction of the source. In addit ion, the buoyancy layer in the S G contains a greater percentage of Fraser River water by volume than other surface layers w i t h i n the study area. Thus , the concentration of home stream odors is enhanced i n conjunction wi th increased stratification; orienta-tion by olfaction or T sensing in shear modulated fine- and micro-structure may be more easily observed in this regime than others. In addi t ion , the differing oceano-graphic regimes wi th in the southern inside passage may ellicit different orientation mechanisms. In this way, the role of the general oceanography on orientation and migrat ion may be determined. 88 Chapter 2: Ultrasonic Tracking D u r i n g the tracking itself and the subsequent qualitative analysis, the depth time series of the fish appeared periodic in nature (Figure 43). Observations of fish depth at 1 m i n . intervals for periods up to 33 hours provided detailed time series of depth and these data suggested an underlying periodicity to the fish's vertical movements regardless of the T or S structure. Therefore, I undertook to quantify this apparent pattern and to use that information to help determine if the fish sought or avoided a particular value of T or S or a particular gradient of T or S. The premise was that knowing the fish's inherent depth distribution independent of the vertical structure in the water column and that this distribution would be manifested in other oceanographic regimes, one could filter the two depth distributions and be left w i t h the fish's preferred choice of T or S or their gradients. B u l l (1936) and Bardach and B jork lund (1957) found conditioned responses in fishes to T changes as little as . 03 - .05°C. Westerberg (1982; 1984) gave evidence supporting the hypothesis that At lant ic salmon were orienting to microstructure T gradients. Thus, it is not considered unreasonable that fish tracked in this study were sensitive and would react to smal l changes in T and S. The vert ical and horizontal homogeneity of the water masses in D P and east-ern JS provided depth time series data independent of relatively large variations in T and S. Thus , any depth preference would be most obvious in these oceanographic regimes. Fourier analysis might resolve the frequency components in the JS and D P depth data and therefore provide an appropriate filter to determine if a relationship between the fish and T or S existed. To this end, the C T D data obtained during the fish tracking program was combined w i t h the fish depth t ime series data to produce an "ambient" oceano-89 Chapter 2: Ultrasonic Tracking Figure 43 . Depth time series AOV data for 8512 tracked in the SG near the northern tip of Texada Island. The dashed line represents the least-squares estimate of the linear trend. 90 Chapter 2: Ultrasonic Tracking graphic variable ( A O V ) dataset. It is considered "ambient" in the sense that the time series produced is a mathematical representation of the T and S that the fish actually experienced at any specific t ime during the track (Thomson et al., 1986). The A O V ' s were determined by a simple algorithm that assumes that: dT dS dT dT dS dS T h i s is to say that the vertical gradients of T and S are much greater than the horizontal slopes of the isotherms or isohalines. This is considered a reasonable assumption as: AT* — « 0(4 .0 x K T 2 ) °C - m - \ oZ — « 0(5 .0 x l O - ^ ^ C - m - 1 , d S — « 0(2 .0 x 1 0 - 2 ) ppt • m _ 1 , dS dS - , dz dS dx'tw^ 0 ( 2 - 5 X 1 0 ^ P P * ' 1 " " This assumption breaks down in two regions of concern. F irs t ly , in the v ic in -ity of strong hybr id tidally mixed fronts, like those observed at Cape Mudge in the Strait of Georgia , and Blackney and Weynton Passages in Johnstone Strai t , the horizontal and vertical gradients may well be of the same order of magnitude. However, no successful tracks were completed i n the vic ini ty of these features and there is therefore no A O V data computed for these regions. Secondly, the water masses present in Discovery Passage and eastern Johnstone Strait were vertical ly and horizontal ly very well-mixed and observations showed that the vertical and horizontal gradients, though weak, were of the same order of magnitude. A l t h o u g h the assumption (2.1) breaks down, this poses no problem insofar as the analysis is concerned as T and S are very nearly constant in both the vertical and horizontal directions and the salmon therefore has but one choice of T and/or S. 91 Chapte r 2: Ultrasonic Tracking Ambient T(S) was calculated by linearly interpolating the T(S) provided by the C T D casts immediately before and after observation of fish depth at a given time. The ambient depth is simply the observed depth of the fish. Thomson et al. (1986) provides a sampling of A O V data products for fish tracked dur ing 1985. The work presented in this manuscript is a graphical repre-sentation of the frequency distributions and correlations of the A O V ' s . Other than stating that the fish tended to avoid water w i t h salinities < 27 ppt , no attempt was made by the authors to interpret their results. In order to determine the possible existence of a preferred T or S, the ex-tremely strong cross-correlation between depth (D) , T and S in the A O V data due to purely physical constraints must be considered. Figure 44 shows the coherence and phase for T and S. The strength of this and other cross-correlations (ie. T / D or S / D ) would clearly mask any significantly weaker cross-correlations, like those assumed to exist between the fish's depth of migration and T or S. Thus, filtering out the strong cross-correlation in T / D and S / D in the raw data due solely to the physical oceanography of the region is necessary to reveal any other weaker correla-tions between these variables and the fish. However, the nature of the relationships between T , S or D and the sockeye is not known and any form of filtering may, in fact, remove the sought after signal. A n acceptable form for this type of filter was not found. Considering this weakness, cross-correlations between the T(S) and D time series due to the tracked sockeye's response to its ambient environment cannot be determined w i t h any degree of certainty. A simple a lgori thm was used to smooth the T and S t ime series A O V data to reduce time dependent effects. The C T D cast data was moving-averaged in t ime over a period (approximately 3 hours) in which 92 Chapter 2: Ultrasonic Tracking Figure 44. Coherence and phase differences for raw T/S AOV data for fish 8503 tracked in the western extremity of JS. The time series are highly coherent and virtually 180° out of phase at all times. 93 Chapter 2: Ultrasonic Tracking the tides did not dramatical ly alter the vertical structure of the water column. The fish's observed depth and interpolated ( A O V ) T(S) at that depth and time was subtracted from the mean state created by the moving average. Thus , we have: ^ a n o m i . k | z where xk(z) and t 3 k Note that N represents the total number of observations, M represents the weight of the moving average and P represents the total number of weighted averages formed. Note that for the majority of Xk{z), M — 4 except in cases where there were fewer than 4 casts made. Once l a n o m , had been calculated from the A O V T and S data , some pre-processing and qualification procedures were carried out to ensure that the correct analytical techniques and therefore accurate results would be obtained. Pre-processing included the removal or correction of obvious outliers in the data, caused by errors in the recording or processing of the original tracking and C T D data, and trend removal. Trend removal, required so as not to mask low frequency spectral content in the data and to unduly distort the processing of correlation quantities (Bendat and Piersol , 1971), was carried out by fitting a least-squares linear curve to the data. In fact, the linear trend may actually be a low-94 !a.oVi. k Xk 1 ~ o ' * o b i : (2.2) 1 M (2.3) 1 ,2,3, - iV 1,2,3, • M 1 , 2 , 3 , - . - P Chapter 2: Ultrasonic Tracking frequency component in the data. No evidence was found to indicate that the calculated linear trend was actually a true trend in the data. The linear f i t , x a n o m > to x a n 0 m , i s given as: Z a n o m , - + Z = 1, 2, 3, • • • JV where N N 12 £ - 6 ( i V + l) £ n = l n = l a = hN(N - l)(N + l) 2(2N + 1) £ - 6 £ nx a n o m , . n+1 n=l ~~ N{N - 1) ' N is the total number of observations and h, the sampling interval is 1 m i n . De-trended time series were then calculated as: ^ a n o m , - ^ a n o m , \ z ^ a n o m ; *• = 1> 2, 3, • • • N for T(S) and ^ a n o n i j ^ o b s ; ^ a n o m , = 1^2,3, • • • N for D . Lastly, to s implify later calculations, the de-trended time series were trans-formed so that the arithmetic means of the transformed time series were equal to zero. The zero-mean, de-trended T(S) and depth anomaly time series were tr ivial ly calculated as: 1 N Cti = X a n o m , — — ^ ^ X a n o m , i = 1,2, 3, • • • N (2-4) n = l for T(S) and k - ^ Y , ^ i= 1 ,2 ,3 , - - - JV (2.5) TV 1 l i = Z: N 95 Chapter 2: Ultrasonic Tracking for depth. Qualification of the data included tests for stationarity and a search for pe-riodicity. Determination of stationarity is v i ta l as the analytical procedures for stationary and non- stationary data are different. Stationarity implies that the sta-t ist ical properties of the time series are independent of t ime. A m i n i m u m require-ment for this to be true is that the probabil i ty density function be t ime invariant and hence a constant mean and variance exists (Jenkins and Watts , 1968). A n estimate of the probabil i ty density function for the raw depth time series data is provided by the depth histograms for each fish (Figures 40 and 41). The unimodal distr ibut ion of the majority of the data suggests that an assumption of stationarity is not unreasonable. However, as the physical processes which were responsible for the raw A O V data were not time invariant, a more rigorous test of stationarity is necessary.f If stationarity is to be determined, it must be assumed that any given sample record wi l l properly reflect the non-stationary character of the process in question and that the sample record itself is long enough to differentiate random fluctuations f rom non-stationary trends. A simple, non-parametric procedure which does not t Note that the t ime series derived herein do not contain data f rom observations made in darkness. Q u i n n et al. (1989) shows that the vertical distributions and movement patterns for sockeye tracked during darkness are substantially different f r o m those tracked dur ing the day. Thus , stationarity cannot be assumed for time series combining tracking data acquired during day and night. S imilar i ly , time series produced by combining data from different oceanographic regimes cannot be considered stationary. 96 Chapter 2: Ultrasonic Tracking require a knowledge of the sampling distributions of any data parameter (ie. if they are normally distributed or otherwise), is the run test. Consider a sequence of observations where each observed value may be classified into one of two mutually exclusive categories: either a '+ ' observation or a ' - ' observation. A ' r u n ' is denned as a sequence of observations in one category that is preceded and followed by a sequence of observations in the other category. The number of runs in the complete dataset is an indication as to whether or not the results are independent observations of the sample parameter chosen. The existence of a trend can be determined by hypothesizing that no trend exists and hence a sequence of N observations are independent realizations of the same random variable. A n y population estimator, or parameter, wi l l work equally well (Bendat and Piersol , 1971). To test for stationarity in <*{ and 7,, values of standard deviation, a, calculated as: " = [ £ ^ 7 ] .- = 1 ,2 ,3 , . i= i for Xt = cti and 7, were computed from consecutive 20 m i n . segments ( M = 20) for the entire record length. The consecutive a values were then categorized as either greater than (ie. a '+ ' observation) or less than (ie. a ' —' observation) the median value. Figure 45 is a representative result of this test. The data can be considered stationary w i t h 95% confidence if the number of runs in the sequence of o relative to the median value was > 6 but < 15 (Bendat and Piersol , 1971). Table 8 shows that the majority of derived time series can be considered stationary at the 95% confidence level. Tests for periodicity in ct, and 7, were conducted for the D P data (8507, 8508, 8509 and 8510) first to determine if an intrinsic depth distr ibution independent 97 Chapter 2: Ultrasonic Tracking Figure 45 . Run test of standard deviation (<T20) for consecutive intervals of 20 observations about the median for all intervals. A single run consists of a sequence of consecutive observations of a2o either greater or less than the median value. Data is for 8504 tracked in QCST near Malcolm Island. 9 8 Chapte r 2: U l t rasonic Tracking T a b l e 8 Results of the run test performed on the a t and 7; time series data for each fish. The number of runs of c^o, (the standard deviation of sequential 20 minute segments of the track) about Omed (the median standard deviation of all the 020J is given for D, T and S. The number of runs is shown as N r u n . Stationarity may be assumed at the 95% significance level if the number of runs, N r u n satisfies 6 < N r u n < 15. Note that 4 and 9 imply, respectively, that stationarity can or can not be assumed. Fish &med D / T / S N r u n Stationarity D / T / S 8503 8504 8505 8509 8510 8511 8512 8514 8516 8601 8602 8603 8604 8605 8606 8610 8611 8612 8613 .291/, .686/ .532/ 1.090/ 1.299/ 2.045/ 1.009/ 1.187/ .395/. 1.907/ 1.450/ 1.548/ .743/, 1.227/ .525/, .492/ .944/ 1.318/ 1.251/ .045/.016 .051/.024 .043/.021 .010/.018 .003/.002 .172/.095 .116/.057 .306/.235 .072/.022 .023/.003 .072/.028 .027/.015 .026/.019 .050/.403 .270/.015 .026/.037 .099/.020 .135/.404 .160/.050 7/ 6/ 4 8/10/ 9 4/ 8/ 8 9/ 3 / 6 3/ 2/ 2 6/ 9 / 3 5/ 2 / 5 5/10/10 7/ 6/ 7 7/ 6/ 8 6/ 8/ 7 4/ 7/ 7 10/ 4/ 8 5/ 6/ 3/ 7/ 3 10/ 6/ 8 12/ 7/11 5 / 2 / 8/ 6/ 6 4 / 4 / 9 4 / 4 / 4 9 / 4 / 4 9 / 9 / 9 • 9 / 9 / 9 9 /4W* 4 / 4 / 4 4 / 4 / 4 4 / 4 / * 9 / 4 / 4 4 / 9 / 4 9 / 4 / 9 / 4 / 9 4 / 4 / 4 4 / 4 / 4 9 / 9 / 4 / 4 / 4 of T(S) or T(S) gradients existed. Est imates of the normalized autocorrelation function, R ^ r , given by: N - r (JV - r ) a 2 r = 1,2,3, . m (2.6) for T(S) and, when a i is replaced by 7 t, D . iV is the total number of observations of a , or 7,-. The lag number, r , corresponds to the time displacement and the max imu m lag number, m=100, was chosen as larger values produced negligible differences in the results. 99 Chapter 2: Ultrasonic Tracking and R ^ r (hereafter defined as R " ^ for convenience) results show much greater correlations between the T and S t ime series than between the T (S) and depth time series. The form of R^f shown in Figure 46 suggests that periodicity is not a feature of these data. for 8510, though different in appearance from R^r for 8509, is not statistically different given the 95% confidence interval . The general form of R^r indicates that 7; corresponds to a O ( l ) auto-regressive process (Jenkins and Watts , 1968). This result is in agreement w i t h the qualitative observation that fish tracked in homogeneous regimes were unimodally surface oriented. A quantitative estimate of this behaviour could be obtained according to: X , = a x » - i + - a) 0 < a < 1 where X i is the predicted depth, /J, is the observed mean depth and a is an empirically determined parameterization of the O ( l ) process. However, the error associated w i t h a predictive estimate of depth using this type of process and these data would be large enough to render the exercise meaningless. A s no periodici ty was found in 7,- or a , for D P , R^f estimates for al l regimes were calculated without filtering for the hypothesized intrinsic depth distr ibution. The general f o r m of for all t ime series data was not sufficiently similar to conclude that - R ^ ' 7 was independent of vertical T(S) structure. The general form of R^'1 for data in stratified regimes was such that periodicity appeared to be characteristic of these data . One notable exception, 8511, in the R^r data was found. In this instance, periodicity was clearly present. There is a statistically significant cycle of approx-imately 18 lags evident in the autocorrelogram (Figure 47). T h i s translates into a period of approximately 18 minutes in 7,-. In fact, 8511 is the only fish that continu-100 o 00 o o © (N o d CO d 95 % (T, S and Depth) Dash=T Dot=S Maxlag=100 Figure 46. Autocorrelation R*? vs lag for 8509 and 8510 tracked in the homogeneous water of DP. o fl> *-t to C 8 B" Chapter 2: Ultrasonic Tracking ously dove and ascended throughout the durat ion of its track. It made a series of 28 dives and ascents dur ing the 7.83 hour track duration. Considering one cycle to be a dive or ascent, we have approximately .28 hrs/cycle or about 17 mins/cycle. This agrees reasonably well with the results obtained v ia the R^r analysis and served to motivate addit ional analysis of the time series datas. Based on the R^r result that some evidence of periodicity was present, esti-mates of the autospectrum, G 0 ' 1 were calculated using an F F T algori thm available i n U B C ' s computing library ( U B C F O U R T , 1986). Specific data pre-processing for the calculation of the F F T consisted of applying a cosine taper to the first and last 10% of the record to be analyzed to reduce the effect of side-lobe leakage inherent i n finite length records (Bendat and Piersol , 1971). The cosine taper, C y , was of the form: ,90.0°™, C T n = c o s ( — — ) n = 1,2,3,-•• N where N equals 10% of the total record length. Raw estimates of G a n were smoothed by band averaging. The choice for the number of bands, Ng, to smooth over was determined by inspection. Ng = 12 provided the best resolution w i t h reasonable confidence intervals. The 95% confidence interval was calculated using Bendat and Piersol , (1971). The effective resolution bandwidth , Be, was calculated as (Bendat and Piersol , 1971): v Be = —. v — 2NB, N = total record length For a constant v = 2Ng = 24, Be varies as the inverse of the record length. Most fish tracked in both years provided JV large enough to yield acceptable confidence intervals and resolution, Be. 102 Chapter 2: Ultrasonic Tracking Figure 47 . Autocorrelation R*>p vs lag for 8511 tracked in the SG near the northern tip of Texada Island. Periodicity is implied by statistically significant zero-crossings. 103 Chapter 2: Ultrasonic Tracking Figure 48 . Autospectrum G 7 for all fish tracked in both years (upper frame). G° (T data) for 8516 (tracked in the SG) is shown in the lower frame. 104 Chapte r 2: Ultrasonic Track ing G a n results are in general agreement with those derived from R^1. There are no spectral peaks evident in cti or -7, for fish tracked in D P . The 18 m i n . cycle that was detected in R^r for 8511 is evident in G * ' 7 for the majority of fish tracked in stratified regimes as a .07 ± .01 cyc les /min peak ( p e r i o d s 15 min) . Figure 48 shows a plot of a calculation of Ga for 8516 and G1 for all fish tracked in both years. Note that there is evidence of an additional peak at approximately .03 ± .01 cyc les /min , corresponding to a period ofes 33 mins. , in both (a) and (b) of Figure 48. T h i s peak was observed in fish tracked in stratified regimes in the majority of T , S and D data . These results were verified by calculating variance preserving forms of G a n . However, large variances at low frequencies d id indicate the existence of trends in a t and 7,-. S imilar i ly , Table 8 shows that trends exist in some of the linearly de-trended data. Thus , this form of trend removal was not successful in all cases. However, in cases where linear de-trending was appropriate significant spectral contributions are evident at the same frequencies as those cases where trend removal was not completely successful. Thus, trend removal was not necessary to resolve these peaks. G i v e n that 95% confidence implies that noise alone w i l l produce a s imilar result in 1 out of 20 records and that the peaks in G a n are not large compared to the error bars, these periodicities are considered significant for two physical reasons: 1.) Spectral peaks in the tracking data for the Q C S T / J S and S G oceanographic regimes could arise due to the internal scales of mot ion inherent i n the density structure. The internal wave cut-off frequency (the B r u n t - V a i s a l a frequency, i V 2 ) in these regimes is at least an order of magnitude greater than the observed fre-105 Chap te r 2: Ultrasonic Track ing quencies in the tracking data. In the homogeneous waters of D P , N2 is of the same order of magnitude as the observed frequencies and may therefore explain why no spectral contributions were de-tected in this regime. However, it is extremely unlikely that the moving platform from which the C T D measurements were made could resolve the same frequencies in different density stratifica-t ion . Moreover, it is not known what external or internal forcing would produce motions at these periods. The conclusion, then, is that these frequencies do not arise due to purely dynamic oceano-graphic causes. 2.) They occur in fish that appeared both oriented and disoriented. 8511 and 8603 were disoriented (swam continuously northwest-ward away from the Fraser River) throughout the duration of their tracks. 8514, 8516, 8606, 8611, and 8613 all displayed portions of oriented (swam southeastward towards the Fraser River) and dis-oriented behaviour. It is somewhat co-incedental that one frequency is twice the other. If the d iv ing velocity (depth/min) were constant for oriented and disoriented fish, and the depth of successive dives was different, it would appear to the F F T a lgor i thm that there are two separate frequencies when, in fact, there is only one. However, the d iv ing velocity, as is discussed in §2.4.4: (f), is not the same for oriented and disoriented fish and the F F T algori thm does resolve two true, distinct peaks. It is hypothesized that the two spectral peaks correspond to two dist inct modes of behaviour in stratified regimes. The higher frequency (ca. .07 cyles/min) 106 Chapter 2: Ultrasonic Track ing corresponds to searching behaviour (Westerberg, 1982; 1985; D0ving et al., 1985) in disoriented fish. T h e lower of two frequencies (ca. .03 cycles/min) represents depth distr ibutions associated w i t h oriented fish. These fish were observed to swim at relatively constant depths and to make infrequent but periodic sudden vert ical excursions. T i m e series data for completely disoriented fish (ie. 8511 and 8512) do not have significant spectral contributions at this lower frequency while data for fish displaying oriented and disoriented swimming behaviour (ie. 8516) contain both spectral peaks. Composi te spectrums were produced in order to decrease the 95% l imits and more completely resolve the observed frequencies. G1 is shown in Figure 48 for al l fish tracked in 1985 and 1986 i n Q C S T , JS and the S G . Al though the 95% confidence interval is reduced, the peaks remain only marginally significant. 2.4.3: Energy Expenditure Periodic i ty in the data deduced from spectral analysis did not directly reveal any relations between the physical oceanography and the fish's vertical d is t r ibut ion . Periods of approximately 15 and 33 minutes were found in oriented and disoriented fish tracked in Q C S T , JS and the SG but not in D P . These periods were also more apparent in fish tracked in regions where the T gradients were larger (SG and Q C S T ) . A s an intrinsic depth distr ibution was not apparent and the calculated frequencies, in general, appeared related to searching for migratory cues, a more direct approach to determining any relationship between the hydrography and the fish was considered. It was observed that sockeye swimming i n regions of little or no strat i f icat ion 107 Chapter 2: Ultrasonic Track ing frequently made a series of steep dives and ascents. Moreover, it was also observed, at least qualitatively, that sockeye swimming in the homeward direction usually swam w i t h i n a relatively narrow range of depths as opposed to those fish who appeared disoriented. A quantitative estimate of this behaviour was considered in terms of the energy expended swimming vertically and horizontally. The energy expended by migrat ion over long distances dictates that econom-ical choices be made between foraging for food, searching for migratory cues and clues, s w i m m i n g speed and direction ( M c K e o w n , 1984; Weihs, 1987). Energy bud-gets for fishes have evolved to the generally accepted form (Brafield, 1985): C = P + R + U + F (2.7) where C is the energy input (ingested food), P is the energy accumulated as growth of the somatic and reproductive tissues, U + F is the energy lost as waste ( urine and feces respectively) and not available for growth or metabol ism, R. M e t a b o l i s m , R , reflects the energy lost as heat. This heat loss is composed of the fish's basal metabol i sm, the energy expended in a resting, non-stressed 'basic' state, and the energy expended in specialized activities such as s w i m m i n g , feeding, aggression, m i g r a t i o n , spawning, etc. The vast majority of energy requirements for fish are incurred i n these specialized activities ( M c K e o w n , 1984). Brett (1983) concludes that for sockeye migrat ing in offshore waters, 44% of the total available energy is required for metabolic purposes and that up to 75% of the total available energy is required for riverine migrat ion. A s a considerable por t ion of the energy budget for migrat ing sockeye is ex-pended s w i m m i n g , a quantitat ive estimate of the role the ambient density gradient may play i n the vert ical d is t r ibut ion of these fishes was attempted. 108 Chapter 2: U l t rason ic Tracking The basic premise is that , as in all of physics, there is a finite amount of en-ergy available. Furthermore, as in all of nature, any animal must optimize energy expended in order to survive. In this case, the migrat ing salmon must reach its nata l stream w i t h sufficient energy reserves to spawn and thereby ensure the sur-v i v a l of the species. Considering the relationship between vertical and horizontal s w i m m i n g , it is reasonable to expect that an excess of energy spent swimming ver-t ical ly , for whatever purpose, w i l l leave insufficient energy available for horizontal movement and vice versa. B o t h types of swimming behaviour are necessary; vertical movements may provide migratory clues (Brannon, 1982), orientation information (Westerberg, 1982; 1984; D 0 v i n g , 1985), danger avoidance and access to food while horizontal movement, at the very least, is required to successfully complete the migrat ion . A n approximat ion of R was attempted by assuming that all the energy lost as heat was expended by s w i m m i n g and that R » P + U - r F a t least for the durat ion of any part icular track. T h u s , R can be estimated by the work done by the fish against the fluid by swimming horizontally and vert ical ly and the work done against the buoyancy force by s w i m m i n g vertically. Energy expended swimming in the vertical is clearly related to the vert ical density gradient and therefore, environmental factors may impose some form of control on the fish's vert ical dis tr ibut ion. To calculate the rate of work (the energy expended) by a fish swimming ver-t ical ly , we first consider the vertical buoyancy force F ; . It is given as: N X > ( p / - p t ) x V O L t=2 where g is the acceleration due to gravity, A r is the number of observations made d u r i n g all or part of a track (s), p / is the density of the fish, Pi is the ambient 109 Chapter 2: U l t rasonic Tracking density and V O L is the approximate volume of the fish, estimated as .012 m 3 . T h u s , if pj ^ pi, F > 0 and the tracked sockeye must expent energy to compensate for this additional force if constant depth is to be maintained. Sockeye salmon are capable of mainta ining neutral buoyancy (ie. pj = pi => F = 0) in the water column by means of a swim bladder connected to the esophagus (Gee, 1983). The fish's density can be matched to the ambient den-sity by forcing gases into or extracting gases f rom the swim bladder either directly by gulping or eructing gases , or directly by osmosis (Gee, 1983). The time lag between change of depth and achieving neutral buoyancy by the above mentioned means is not precisely known. In addit ion, numerous fishes have been observed to compensate for non-neutral buoyancy by coupling swimming speed and angle of attack (the longitudinal angle of the fish relative to the horizontal) . A n increase or decrease in hydrodynamic lift can therefore be generated (Magnuson, 1978; Webb and Weihs, 1983) to compensate for the hydrostatic buoyancy forces. For simplicity, it was assumed that the change in depth of the fish and the resulting change in the buoyancy force was compensated for immediately at each new depth by ut i l iz ing the s w i m bladder and/or changing swimming speed and angle of attack. Thus , pj = Pi and the vertical buoyancy force F g . is given as: N F B , = X > ( P i - * - i ) x V O L t=2 T h e energy expended to achieve neutral buoyancy was assumed to exactly balance the change in buoyancy forces between observed depths. (2.8) implies that if p, = Pi-i, then F ^ = 0. Hence, no energy is expended by the fish compensating for non-neutral buoyancy w i t h i n vertically homogeneous density fields. The rate of doing work against the buoyancy force W g , was calculated as 110 (2.8) Chapter 2: Ultrasonic Tracking (Webb, 1978): W ^ F ^ x V f »- = 2 , 3 - - - J V - l (2.9) where the absolute value of vertical velocity, V ^ , was calculated using a centered second order finite difference scheme. This is given by: ( D i + 1 - D i _ ! ) v7 = i = 2 ,3 , - - -N - 1. (2.10) 2 x 60 has units of m s " 1 as observations of depth, D , were made every minute. The rate of work done by a swimming fish against a fluid is related to the hydrodynamic drag, D R , (Webb, 1978) where: D R , = .5 x C D , x \pi\ x S x |V;| x V< (2.11) Some simplif ications of (2.11) are possible as: |V,| 2 - ( V f ) 2 + ( V t v ) 2 . (2.12) where Vf is the horizontal velocity and V t H » V t v . (2.12) may be re-written as: |V tf « ( V f ) 2 . (2.13) T h e vert ical D R j ' , and horizontal , D R ^ , components of (2.11) using (2.13) are: D R ] 7 = .5CD|'|p t -|SV?V i r ' ; (2.14) D R f = . 5 C D f | p t - | S ( V f ) 2 . (2.15) S, the wetted surface area of the fish, is taken to be 0.10 m 2 . In (2.14), the mean density, p, is s i m p l y : P= (pi + Pi-i)/2 t = 1,2,3 111 Chapte r 2: Ul t rasonic Tracking and in (2.15), p is taken to be the average ambient density of the entire track. Thus, for (2.15), we have: T h e vertical drag co-efficient, C D , , may be expressed as a function of the vertical Reynolds number, R ^ . For laminar boundary layer flow, C D ^ is calculated as: C D t v = (4.2)(1.33) ( R D - 1 7 2 (2.16) T h e factor 4.2 arises as a result of a correction for pressure drag (Bainbridge, 1961) and for the undulatory motion associated w i t h propulsion (L ighth i l l , 1971). The average swimming speed for all fish was calculated to be approximately 1.0 m/s (Quinn et al, 1989). A t these velocities, the boundary layer may be expected to be laminar (Webb, 1975; Blake , 1983). However, (2.16) is recognized as a simplification and w i l l only give reasonable estimates of the drag co-efficient for streamlined bodies at small flow speeds. A t higher flow speeds, (2.16) underestimates the the drag co-efficient. R v defined as: V V T v\ where L , defined as the characteristic fish length, is taken to be .60 m . Variations in temperature can cause significant variations in the kinematic vis-cosity , v, and therefore R ^ (Webb and Weihs, 1983). Considering these variations, V{ was calculated f rom the empirical formula (Weast, 1974): 10 a ; * = (2-17) p 1301 3.30233 998.333 + 8.1855(T-, - 20) + 0.00585(T ; - 20) 2 112 x 1 0 - 1 . Chap te r 2: U l t rasonic Tracking The rate of work done by the fish s w i m m i n g vertically against the fluid is calculated as (Webb, 1978): W£. = D R ^ x V t v . (2.18) F ina l ly , the total rate of work done by the fish swimming vertically, W^f, can now be calculated using (2.10) and (2.17). We have: W¥=E(W£+W£). (2-19) It now remains to estimate the work done by the fish s w i m m i n g horizontally. T h e reader w i l l recall that we have considered horizontal gradients of density negli-gible in the regions where tracking was carried out. This assumption in conjunction w i t h those used to construct the estimate for R imply that we need only consider the work expended by swimming against the f lu id . The horizontal drag co-efficient, C D f is analogous to (2.16) and is given as: C D f = (4.2)(1.33) ( R f ) ~ 1 / 2 and R f , the horizontal Reynolds Number is taken as: R f = — . Vi The total work expended by swimming horizontally is analogous to eq.(2.19) and wri t ten as: TV =3H ( D R ? x V ? ) . (2.20) t = 2 A representative plot of the total work rates vs t ime is shown in Figure 49 for 8516 and 8603. It can be seen that intermittency is common to a l l these figures. 113 Chapter 2: Ultrasonic Tracking Figure 49 Total work rate considering rate of energy expenditure by swimming vertically and horizontally. Units of work rate are Joule*.min-*. Elapsed track time is shown on the x-axis. Normalized area represents the average work rate over the QCST t r a C k m J ° u l e S m i n ' 1 - 8 5 1 6 w a s t r a c k e d 'm t h e SG and 8603 central 114 Chapter 2: Ultrasonic Track ing A l t h o u g h it appeared during tracking that most of the fish were generally swimming in a steady fashion, calculations of work rate show that energy expended swimming varied a great deal over the durat ion of the track. Laboratory studies on sockeye metabolic rates (Brett, 1963) give estimates of energy expenditure of approximately 45 c a l / k g / d a y for fish migrat ing in riverine conditions. Considering a 3 kg fish and Bret t ' s (1983) figure of 44% of total energy required for metabolic processes in offshore waters, approximately 79.2 ca l /day are expended metabolically. The area under the curve shown in Figure 49 was numerically determined and this calculat ion represents the total energy expended by the fish swimming vertically and horizontally. The total energy expended divided by the track durat ion gives an average rate of energy expenditure. Table 9 gives the average energy expenditure (Joules • m i n - 1 ) for indiv idual fish. The extreme variabil i ty i n the total work rate cannot be completely accounted for. Calculat ions of horizontal swimming velocity are not corrected for advection by near surface currents and this error does account for some of the observed v a r i -ability. 8510 was tracked just north of Seymour Narrows when t idal currents were approaching a m a x i m u m . In this instance, advection accounts for the very large rate of energy expenditure shown in Table 9. The variabi l i ty in the total rate of work implies that different fish proport ioned differing amounts of energy to swimming. T h i s is in general agreement w i t h Bret t (1983) in that only 44% of available energy is used for migratory s w i m m i n g and therefore the majority of energy available is being used in other parts of the energy budget. Table 9 also shows that the rate of work done against the buoyancy force is much less than that expended against the fluid alone. Considering the disparity 115 Chapte r 2: Ul t rasonic Track ing Table 9 The average work rate calculated for a complete track. Units are Joules • m i n - 1 . The rate of energy expenditure swimming horizontally, the total rate considering vertical and horizontal swimming and the standard error associated with the total work rate are given for each fish. F i s h Horizonta l Work Rate Total Work Rate Standard E r r o r ( ± ) 8503 0.125 0.127 .004 8504 3.070 3.153 .103 8505 0.043 0.043 .002 8507 0.046 0.046 .002 8508 0.115 0.120 .006 8509 0.634 0.640 .006 8510 10.374 11.618 .284 8511 0.034 0.038 .002 8512 0.022 0.022 .001 8513 0.010 0.013 .005 8514 0.018 0.018 .003 8516 11.582 11.530 .175 8601 0.066 0.066 .001 8602 3.430 3.385 .045 8603 0.050 0.051 .002 8604 0.039 0.039 .002 8605 0.075 0.077 .003 8606 0.156 0.155 .014 8610 1.235 1.264 .003 8611 0.195 0.204 .003 8612 0.106 0.104 .003 8613 0.195 0.204 .005 in these work rates, it appears that the vert ical difference in buoyancy force, and therefore the T and S gradients, do not inhibi t the fish's vertical excursions. T h e aspect ratio, A R , of a part icular track is calculated as: _ D I S T H  A R " D I S T V where D I S T H and D I S T V represent the horizontal and vertical distances travelled. T h e aspect ratio gives a general indication of the fish's overall energy expendi ture ! w i t h regards to either vert ical or horizontal s w i m m i n g as well as a measure of t It should be noted that overall distance travelled gives no indicat ion of the 116 Chapter 2: Ul trasonic Track ing orientat ion (Quinn and terHart , 1987). It is necessary, however, to determine some numeric standard that relates orientation and A R . The choice, while not completely arbi trary, is somewhat subjective. The choices made for the purposes of selecting an 'or iented' and 'disoriented' A R follow Quinn and terHar t (1987), and are: i . ) T h e entire track of 8512 is representative of an extremely disori-ented fish as very l i tt le progress over the the ground was made as well as frequent deep vertical excursions. The A R for this track was the m i n i m u m observed for both years and as such, was chosen as A R m j n . i i . ) T h e port ion of 8516 between 1531 hrs and 1730 hrs represents a very well oriented fish. In this case, the A R for this portion of this track was considered to be A R m a x . A s A R min — 9.765 and A R m a x — 75.141, the calculated A R ' s for each track were normal ized as: . A.R, ~~ .AR-miri A R " = A R r - A l T T 0 < A R , < 1. T h e normalized aspect rat io, A R , , , represents the relative degree of southeastward orientat ion where A R , , = 1 implies 'perfect' homeward orientation and A R , , = 0 implies complete disorientation. acceleration. Swimming at m a x i m u m velocity (burst speed) incurs significant en-ergy expenditures (Soofiani and Hawkins , 1985). Bre t t and Groves (1979) state that the metabolic cost of 20 seconds at burst ing speeds equals 15 minutes of active metabol ism (approximately 1-2 body lengths/sec for adults) or 3 hours of basal metabol i sm. 117 Chapter 2: Ultrasonic Tracking Table 10 gives both A R and A R , , . Clearly, orientation varies cons iderab ly and is poor overall , but s imilar for all fish in all regimes. The parameter A R , , , when compared to the actual horizontal track of the tagged fish, yields a reasonable quanti tat ive estimate of the relative degree of homeward orientation. Table 10 Normalized aspect ratios, A R , , , for each fish. Total distances travelled (m) in the vertical and the horizontal as well as the raw aspect ratio, AR, is given. Regime F ish Vert ica l distance travelled Horizonta l distance travelled A R AR„ Q C S T / J S 8503 816.00 24599.50 30.146 0.312 8504 846.00 18654.50 22.050 0.188 8505 526.00 8573.41 16.299 0.100 D P 8507 624.00 8486.52 13.600 0.059 8508 218.00 5594.66 25.664 0.243 8509 2232.00 65348.06 29.278 0.298 8510 760.00 15319.98 20.158 0.159 S G 8511 1163.00 11886.15 10.220 0.007 8512 1330.00 12987.63 9.765 0.000 8513 479.00 6989.42 14.592 0.074 8514 1839.00 36883.71 20.056 0.157 8516 1278.00 47996.96 37.556 0.425 Q C S T / J S 8601 1220.00 16283.64 13.347 0.055 8602 699.00 30006.11 42.927 0.507 8603 1197.00 19400.81 16.208 0.099 8604 566.00 9999.15 17.666 0.121 8605 596.00 14075.13 23.616 0.212 8606 1015.00 51329.75 50.571 0.624 S G 8610 1034.00 54890.05 53.085 0.663 8611 1237.00 23269.78 18.811 0.138 8612 588.00 20948.08 35.626 0.396 8613 743.00 16198.73 21.802 0.184 118 Chapte r 2: Ultrasonic Track ing 2.4.4: Swimming Velocity and A O V Relationships A s spectral analysis and A R , , gave some evidence for a possible relation be-tween vert ical /horizontal distr ibution and the sockeye's physical environment, a number of hypotheses were tested in an attempt to further quantify the results of previous sections. A R , , implied that vertical and horizontal s w i m m i n g velocities are negatively correlated as oriented and disoriented behaviour could be characterized by the depth and frequency of vertical excursions. W i t h o u t an acceptable model for relating vert ical and horizontal swimming behaviour, or swimming behaviour and physical environment, scatter plots of sev-eral hypothesized relations for individual and composite fish tracks were produced. These plots were then examined for any discernable relations between the chosen variables. The intent here was to sacrifice detailed causal knowledge for general relationships. Where applicable, all data were normalized using the following transforma-t i o n : * , = J " ~ _ X m i ° , 0 < Z „ < 1 ; (2.21) where xn represents the nth data point. This transformation provided a convenient scale w i t h which to compare the different data subsets. (a) Horizontal and Vert ica l Velocity Figure 50 is an example of the scatter plot produced in an attempt to relate horizontal and vertical velocity. N o correlation between horizontal and vert ical velocity was evident i n any of the scatter plots produced. The tendency for the 119 Chap te r 2: Ul t rasonic Track ing data values to lie against either axis is an artifact of the normalizat ion procedure when a relatively large maximum value exists in the non-normalized data. F i sh tracked in the same oceanographic regimes were grouped together in or-der to reduce the statistical error inherent in small data sets. The resulting plots showed no clear evidence of any correlat ion. There is also no clear evidence indi -cating that fish tracked in a particular oceanographic regime were better oriented than in others. It was clear, however, that most data points were concentrated near the origin and that there is less chance of finding data points near the m a x i m u m in the horizontal or vertical swimming velocities. It is also interesting to note that the m a x i m u m swimming speed observed is well above the anaerobic threshold for sockeye salmon (> 3 body lengths/sec or « 1.8 m/sec; Bre t t , 1983). A s for horizontal and vertical velocities, simple correlations were sought be-tween depth and vertical velocity. G i v e n a statistically significant mean depth of travel and the qualitative observation that well oriented fish tended to swim w i t h infrequent shallow vertical excursions, the data were analysed to see if a tendency to return to a preferred depth (the mean) would manifest itself as an increase in ver-t ical velocity the greater the deviation f r o m the mean. If this were true, a positive correlation between depth and vertical velocity would exist. Representative scatter plots of depth anomaly, D a n o m . , calculated as: (b) Depth Anomaly and N o r m a l i z e d Vert ical Velocity and normalized vertical velocity, are shown in Figure 51. It was evident f r o m these calculations for al l tracks in both years that D amom: .. and V , j for fish tracked 120 Chapter 2: Ultrasonic Tracking > CD > u o > CD > 6 o CO © to d CN O • s o ° a a e V: 0.2 0.4 0.6 0.8 1.0 CO d co d d CN d Hor. Vel. Ftk8516 a • V SD e Bgo a a o MP . a*e a • 9. 1 a F t k 8 6 1 0 -I-I 0.2 0.4 0.6 Hor. Vel. 0.8 1.0 Figure 5 0 . Scatter plots of horizontal and vertical velocities for 8516, tracked in central SG near the BC mainland coast, and 8610, tracked in northern SG south of Cape Mudge. 121 Chap te r 2: Ultrasonic Track ing in JS , Q C S T and D P were more positively correlated than those tracked in the S G for both years. Correlations in the S G were, in general, sl ightly negative while those for Q C S T , JS and D P were never negative. ( c ) Depth and Normalized Vert ica l Velocity Correlat ions bewteen D a n o m i and V ^ suggested that the mean and therefore D a n o m . d i d not accurately represent the sockeye's depth distr ibution. Thus , unf i l -tered observations of depth and V ^ were considered. Scatter plots of depth and normalized vertical velocity generally show a weak positive correlation. Given the scatter of the data, it is probably more appropriate to conclude that there is no clear evidence of a negative correlation. Nonetheless, there are several conclusions that can be made based on these plots. F i r s t l y , the mean depth is easily seen to be different for each oceanographic regime and for different years in the same regime. Secondly, as vertical velocity and depth are positively correlated, the greater the depth , the greater the vert ical velocity. T h i s correlation is stronger in the less stratified regimes of JS, Q C S T and D P than in the more strongly stratified S G (cf. Figures 52 and 53). The two tracking years, 1985 and 1986, are similar i n that weaker stratification impl ied a stronger positive correlation between depth and vert ical velocity. In terms of orientation, the observation that well oriented fish tended to s w i m at a specific depth for relatively long periods of t ime is part ia l ly evidenced in these plots. W e l l oriented fish exhibited stronger positive correlations than those fish which were poorly oriented. Wi thout prior knowledge of the fish's actual track, however, the difference in correlation coefficients alone is not sufficient to differen-tiate oriented from disoriented fish. 122 Chapter 2: Ultrasonic Tracking oo © > E © o © Ver. Vel. Max.=0.083 Ver. Vel. Min.=0.000 Lin. Cor. Cof.=0.418 • a F t k 8 5 0 3 10 20 30 Depth (m) > > 6 o co d CD d d d Ver. Vel . Max.=0.100 Ver. Vel . Min.=0.000 L i n . Cor. Cof.=-0.352 F t k 8 5 1 3 10 Depth (m) 20 30 Figure 51 . Scatter plots of depth anomaly and normalized vertical velocity for 8503, tracked in the western extremity of JS, and 8513, tracked in the SG near the southern entrance to Sabine Channel. The depth anomaly represents the absolute value of the distance away from the mean depth. 123 Chapter 2: Ultrasonic Tracking co d to © > u CD > U O o • a o f f • . oo a *A l« ' *o , o • B o Ver. Vel . Max.=0.125 Ver. Vel . Min.=0.000 L i n . Cor. Cof.=0.604 Ftk 85.3-5 10 20 30 Depth (m) 40 50 CD > u CD > o CO d co d d <N d •a - 8 a . Ver.- Vel . Max.=0.183 Ver. Vel . Min.=0.000 L i n . Cor. Cof.=-0.567 a o Ftk 85.7-10 10 ~20 30 40 50 Depth (m) Figure 52 . Scatter plots of depth and normalized vertical velocity for the composite data 85.3-5 (fish tracked in QCST/JS) and 85.7-10 (fish tracked in the SG). 124 Chapter 2: Ultrasonic Tracking > a> > 6 o > > s o 00 © CO d d CN d • a . B • B * * •a Mg, " J - F t k 85 .11-16 Qnoo B 10 20 30 40 50 Ver. Vel . Max.=0.167 D e p t h (m) Ver. Vel . Min.=0.000 oo d d d d e fto F t k 86 .10-13 20 30 40 50 V e r . V e l . M a x . = 0 . 1 3 0 D e p t h ( m ) V e r . V e l . M i n . = 0 . 0 0 0 V V Figure 53 . Scatter plots of depth and normalized vertical velocity for the composite data 85.11-16 and 86.10 -13. These data are for all fish tracked in the SG for both years. 125 Chapte r 2: Ultrasonic Track ing (d) D e p t h and Normalized T Gradient Q u i n n et al. (1987) and Q u i n n and terHart (1988) could neither conclu-sively support nor refute, at least qualitatively, Westerberg (1982; 1984) and D0ving (1985) who concluded that migrat ing At lant i c salmon used microscale temperature gradients as a migratory clue. It is well documented that salmonids possess highly sensitive olfactory systems (Bertmar and Toft , 1969; D 0 v i n g , 1985) capable of dif-ferentiating fine and micro-scale hydrographic features. Specifically, the olfactory abil ity of many salmonids to differentiate water masses of different origins, as well as different pheromones, has been observed (Brannon, 1982). It does seem reasonable, then, to assume that such a highly evolved faculty has some specialized purpose. T gradients demark transitions from one T regime to another and wi th in B C ' s inside passage, as the thermoclines and haloclines are coincidental, they demark different water masses of different origins. Given these abilities, the most efficient way to gather T or olfactory information is to remain in or near the region of m a x i m u m gradients in T or S. Thus , very l i tt le energy is expended in detecting the greatest possible difference. Scatter plots of normalized T gradient (dT/dz)^ and depth showed that fish tracked in this study tended to occupy depths associated wi th the weakest T gradi-ents. There were clearly more data points in the vic ini ty of the m i n i m u m gradient for the majority of fish tracked. In a d d i t i o n , sockeye that were, in general, well oriented had greater concentrations of data points in the region of the weakest T gradient. In fact, where C T D vertical profiles of T and S showed a thin distinct thermocline and halocline, there is a bi furcat ion of the data showing a concentration of data points corresponding to different relatively well -mixed layers separated by 126 Chapte r 2: U l t rason ic Tracking the strong T and S gradients. Figure 54, showing a 8511 and 8514 tracked in the S G clearly illustrates this bifurcation. In regions of little or no stratification (eastern JS and D P ) , i n d i v i d u a l scatter plots showed little tendency towards this phenomenen. Similarity, in the presence of a weak T gradient encompassing the observed range of fish depths, da ta points were equally distributed throughout the observed values for the T gradient. Considering the T derivative and depth scatter plots, 8514 (Figure 54) pro-vides an interesting composite of both types of observed behaviour. The C T D data shows a strong t idal influence present in the surface waters of the central S G . Warm, less saline water f rom Howe Sound and the Fraser River is seen to be advected by a strong flood tide into the region where 8514 was tracked . The presence of a surface m i x e d layer and a strong thermocline is eradicated by the advancing warm, comparatively fresh water originating in the south-central S G . The t ime evolution of the C T D data clearly shows the gradual weakening and deepening of the thermo-cline and the displacement of the surface mixed layer (terHart and Q u i n n , 1989). In conjunct ion w i t h the above mentioned observations, 8514 does not exhibit any ten-dency toward the m i n i m u m T gradient when there are weak T gradients extending throughout the observed range of depths, as there were during the latter portions of the track. Dur ing the early portions of the track, when a th in thermocline and strong T gradients existed, 8514 exhibited a tendency toward the wel l -mixed regions separated by the thermocline and there was a bifurcat ion of the data points . The scatter plot of depth and (dT/dz)^(Figure 54) is the sum of these behaviours; there is no clear tendency towards the m a x i m u m gradient and there is a bifurcation of the data points. 127 Chapter 2: Ultrasonic Tracking T der. Max.=0.471 T der. Min.=0.024 O • O f a CJ Q a a a a 0.2 0^ 4 0.6 N o r m . T d e r . ft O a a a D a • o •a F t k 8511 0.8 a 8 0 _ a " "o.aa^ _a • • a a a 1.0 T der. Max.=0.538 T der. Min.=0.041 a a a F t k 8514 o 4-02 0.4 0.6 N o r m . T d e r . 0.8 1.0 Figure 54 . Scatter plots of normalized temperature derivative, (dT/dz)^ and depth for 8511, tracked in the SG near the northern tip of Texada Island, and 8514, tracked in cenral SG near the entrance to Sabine Channel. 128 Chap te r 2: U l t rasonic Tracking (e) (dT/dz)^ and Scatter plots of (dT/dz)^ and were produced in an attempt to further elucidate the relationship between T gradient and depth. [dTjdz)n and V , , were generally weakly correlated. Frequent, slow dives/ascents of short duration wi l l produce plots s imilar to that of Figure 54 and discussed in §(d) above. Thus, ob-served depth and w i l l not relate to the maximums i n (dT/dz)^. The conclusion, then, is that sockeye did not remain w i t h i n the greatest gradient, but frequently swam through i t . Scatter plots of (dT/dz)^ and D support this observation. This conclusion is in concert wi th those made by Westerberg (1984). (f) D i v e / A s c e n t Characteristics Considering the time between successive dives/ascents and depth, Figures 55 and 56 show that these data correspond to two distinct modes of behaviour: frequent deep diving and infrequent deep, but frequent shallow dives and long durations at nearly constant depths. Disoriented fish (Figure 55: 8511) do not display the latter grouping of da ta points while oriented fish (Figure 55: 8516 and Figure 56: 8603, 8606), even if only for a portion of their track do. Dive characteristics were further examined by considering dive durat ion. F i g -ures 57 and 58 show dive depth and dive durat ion for disoriented (Figure 57: 8511, 8512) and oriented (Figure 58: 8516, 8610) fish in the S G . There is a tendency in oriented fish to make infrequent deep dives of long durat ion and frequent shallow dives of short durat ion . Hence, the depth dis tr ibut ion of an oriented fish w i l l be characterized by relatively long periods spent w i t h i n a narrow range of depths. C o n -versely, disoriented fish tended to make frequent, deep dives of short duration and 129 Chapter 2: Ultrasonic Tracking o 9 8 > CD > c o < © 1 © B "Bos ° Ftk 8511 i i '4* e OS 25 e < i • i i i i i i i i i i 50 75 T i m e B e t w e e n D i v e / A s c e n t 9 e s Ftk 8516 e T • i I • 1 1 r a _ 25 • * e i i B • • 50 i • i t i 75 100 § 91" o co o •9 T i m e B e t w e e n D i v e / A s c e n t 100 Figure 55 . Time between successive dives/ascents (min) versus depth of dive/ascent (m) for 8511, tracked in the SG near the northern tip of Texada Island, and 8516, tracked in central SG near the BC mainland coast. Data shown for 8511 is charac-teristic of disoriented fish; there are frequent deep dives and no periods of relatively constant depth. These data are in sharp contrast to those for oriented fish represented here by 8516. 130 Chapter 2: Ultrasonic Tracking o _ » 8' > 9 C e l l <D CJ CO < © in Ftk 8603 25 50 i i i i i i 75 100 T i m e B e t w e e n D i v e / A s c e n t •a Ftk 8606 8, #occo ^ # ! ' 1 ' I I l • l t 1 i » 50 75 o 100 co < T i m e B e t w e e n D i v e / A s c e n t o. Figure 56 . Time between successive dives/ascents (min) versus depth of dive /ascent (m) for 8603 and 8606 (both tracked in QCST). These data are charac-teristic of oriented fish in that there are relatively infrequent dives and long periods spent at constant depths. 131 Chapte r 2: Ultrasonic Track ing spent l i t t le t ime at relatively constant depths. These results are in good agreement w i t h qualitative observations of depth distr ibutions made earlier. Table 11 shows the least-squares linear fit to the dive and duration data for each fish. The slopes of the linear fit to these data are different for oriented and disoriented fish in different oceanographic regimes (cf. fish tracked in D P where all fish were disoriented and those tracked in any other regime) imply ing the diving characteristics for oriented and disoriented fish differ depending on stratification. Note that the mean slopes for these lines are different for different years in the same regimes. In addit ion, the magnitude of the slope is inversely proportional to orientation. 2.4.5: Dimensional Analysis Investigators have had success correlating physical variables such as freshwater run-off, sea-surface T , S, sea-surface height, etc (Favorite, 1961; Wickett , 1977; Burgner , 1980; M y s a k et al., 1982; I P S F C , 1984; H a m i l t o n , 1984; B lackbourn, 1987; Groot and Q u i n n , 1987, X i e and Hsieh, 1989) w i t h run t iming and/or horizontal dis t r ibut ion. It must be remembered, however, that correlation is not a necessary or sufficient condit ion for cause. To date, a significant correlation between some physical parameter and a salmon's vert ical distr ibut ion has not been found, let alone a causal relationship. In an effort to identify any possible relationship between the salmon's physical environment and its vert ical d is t r ibut ion , dimensional analysis was employed. Dimensional analysis techniques allow one to determine the functional rela-t ionship between groups of variables (Binder , 1973) without exactly knowing the 132 Chap te r 2: U l t rasonic Tracking Table 11 Linear least squares fit to the dive duration data for each fish. Intercept, b, and slope, m, are shown for each fish. The standard error is given for b and m. Fish order is by orientation by regime and year. Oriented and disoriented fish are represented by the subscripts 'o ' and ' d ' respectively F i s h Regime b (±) m ( ± ) 8503 o JS(w) 2.96 0.82 2.56 0.34 8504 d Q C S T 3.14 0.59 2.08 0.15 8505 d JS(w) 4.01 2.35 4.29 1.19 m = 2.98 ± .42 8507 d D P 1.70 1.88 6.99 0.96 8509 d D P 6.17 1.04 3.44 0.49 8508 d D P 4.86 2.06 2.46 0.86 8510 d D P 4.05 1.32 4.14 0.58 m = 4.26 ± .38 8516 0 S G 5.40 0.85 1.21 0.34 8513 0 S G 5.19 1.06 1.29 0.50 8514 0 S G 2.30 0.78 3.09 0.28 8511 d S G 0.75 1.25 4.11 0.35 8512 d S G 2.67 0.66 2.53 0.17 m = 2.47 ± .15 8602 o Q C S T 3.45 1.31 2.17 0.53 8606^ Q C S T 2.69 0.58 2.56 0.26 8603 d Q C S T 2.45 0.84 2.25 0.28 8605 d Q C S T 7.19 1.35 0.26 0.55 8604 d Q C S T 3.92 1.65 1.36 0.65 8601 d Q C S T 5.55 0.71 1.07 0.28 m = 1.61 ± .18 8612 0 S G 4.44 1.04 0.91 0.39 8610 o S G 6.96 1.30 0.68 0.66 8613 d S G 6.38 1.19 0.81 0.36 8611 d S G 2.83 0.73 2.57 0.20 m = 1.24 ± .22 under ly ing physics. The functional form of the relationship is provided at the ex-pense of the detai l . The fish's vertical distr ibut ion may be represented by the absolute value of its vert ical velocity, V . If V is large, the fish's vert ical d is t r ibut ion is characterized 133 Chapter 2: Ultrasonic Tracking S i CD > Ftk 8511 Slope: 4.11 Intcpt: 0.75 10 i i 20 30 D i v e D u r a t i o n ( m i n ) S i 8 P 8 Ftk 8512 Slope: 2.53 Intcpt: 2.67 1 0 2 0 D i v e D u r a t i o n ( m i n ) 30 Figure 57 . Dive duration (min) vs depth of dive (m) for 8511 and 8512. Linear, least-squares best fit estimates are shown as solid straight lines through the data. Both fish were tracked in the SG near the northern tip of Texada Island and were considered disoriented. 134 Chapter 2: Ultrasonic Tracking 8' 3 OJ > p a-p. Ftk 8516 Slope: 1.21 Intcpt: 5.40 10 20 D i v e D u r a t i o n ( m i n ) • i 30 s 8 > P 8 Ftk 8610 Slope: 0.68 Intcpt: 6.96 Hr 10 20 D i v e D u r a t i o n ( m i n ) 30 Figure 58 . Dive duration (min) vs depth of dive (m) for 8516, tracked in central SG near the B C mainland coast, and 8610, tracked in northern SG south of Cape Mudge. Linear, least-squares best fit estimates are shown as solid straight lines through the data. Both tracks are considered characteristic of oriented fish. 135 Chapter 2: Ultrasonic Track ing by frequent, large ampli tude ascents and/or descents. If V is smal l , then depth changes tend to be infrequent and of small ampli tude. Relevant physical variables were taken to be the temperature dependent fluid viscosity, i/, the fish's mass, M , the earth's gravitat ional force, G , and the buoyancy force, B , due to vertical density changes in the water co lumn. In terms of their dimensions, V , v, B , G and M are: T / L L 2 L M L r ' u = Y' T2' 5 = Y 2" where L , M and T are the dimensions of length, mass and time respectively. We seek a dimensionally homogeneous relationship of the form: V = f(B,M,v,G), (2.21) In terms of the appropriate dimensions, we have: E q u a t i o n (2.22) yields three equations for the four variables a, b , c and d. These three equations may be solved to give a, c and d in terms of b . The choice (2.21) yields the relat ionship: V oc B - h M b u 1 / z G b + l / 3 which is re-written in dimensional form as: V «{vG)V*f{¥£). (2.23) R e - w r i t i n g (2.23) w i l l give a convenient set of non-dimensional coordinates: 136 Chapter 2: Ultrasonic Tracking The choice (2.21) is not unique in that other equally valid relationships may be determined depending on the choice of variables. To test other possible solutions for any evidence of functional dependence, the fol lowing relations were also considered: V = f(M,u,G,dp/dz) (2.25) and V = f(M,v,dp/dz) (2.26) w h i c h yielded, respectively, the non-dimensional co-ordinates: V (dP/dz)»*/* v ( d P i d z \ x / A , , Provided an appropriate choice of variables has been made, scatter plots of (2.24), (2.27) and (2.28) w i l l reveal the form of the functional relationship between the non-dimensional variables. Figures 59 and 60 are representative plots of (2.24) and (2.27) and (2.28) respectively for one fish. A l t h o u g h all possible variables were not tested, the most reasonable choices were. These results suggest that physical variables alone are not sufficient to specify the sockeye's vertical d is t r ibut ion . Other undetermined behavioural factors, such as physiological requirements, may play an important role in motivat ing the fish's vert ical movements. 2.5: Summary Phys ica l oceanographic influences on the estuarine phase of the return m i -grat ion of sockeye salmon (Oncorhynchus nerka) were examined using data col-137 Chapter 2: Ultrasonic Tracking CD © 00, d CD d d oft °. • CN d 1 • a a • oo i • OJ _o_ » I I I I I I • » I • I l l 02 0.4 0.6 0.8 L O CD d d CN o • a "oT 0.4 1 1 • • • 1 1 . • • • i • • • • i • i • • i 0.6 0.8 L 0 (MG/B) Figure 59 . Scatter plots of non-dimensional parameterization of ultrasonically telemetred fish tracking and physical oceanographic data. Data shown is for 8603 which was tracked in central Q C S T . 138 Chapter 2: Ultrasonic Tracking co d d d CN o* e e a a • 8 " ff 8° " * 1 a " e e • 8 " - V 8 % 8 O O 8 8 _ 8 8 8 I D _ o _a • e a 8 8 s s _ 8 8 ° 8 ° 883 8 eo o^ a 8 8 1 I 1 ' ' • I I • I 0.2 0.4 0.6 0.8 L0 vi (Bp/dz)1'4 F i g u r e 6 0 . Scatter plot of dimensional relationships determined by parameterization of ultrasonically telemetred fish tracking and physical oceanographic data for 8603. This fish was tracked in central QCST. 139 Chapte r 2: U l t rason ic Tracking lected during two summers of ultrasonic telemetry and concurrent conductivity-temperature-depth casts. Twenty-two separate tracks were examined and an at-tempt to quantify the environmental factors influencing vertical and horizontal dis-t r ibut ion was made. The wel l -mixed, nearly homogeneous water of Discovery Passage was consid-ered an experimental control in that sockeye tracked in this oceanographic regime must necessarily exhibit a depth dis t r ibut ion independent of a preferred temper-ature/sal ini ty or temperature/sal inity gradient. Al though frequent vertical excur-sions qualitat ively appeared to be a characteristic feature of fish tracked in verti-cally wel l -mixed waters, spectral analysis revealed no periodicity i n the tracking data for Discovery Passage. Similar analysis of tracking data collected in oceano-graphic regimes w i t h varying degrees of stratif ication showed statistically significant periodic vert ical excursions of approximately 15.4 and 33.3 mins . /cyc le . Periodic movements quanti tat ively distinguished oriented or disoriented tracks, or portions thereof, in stratif ied regimes. The nature of cyc l ica l vertical excursions is believed to be an orientat ion mechanism whereby disoriented fish 'sample' the available water co lumn for orientation clues. This type of behaviour has been observed for other salmonids (Ichihara and Nakamura, 1982; Westerberg, 1982, 1984). Calculat ions of work rate considering vertical and horizontal swimming and vertical density gradients show that energy expenditure was highly intermittent for al l fish tracked. The high degree of variabi l i ty in energy expenditure is partial ly caused by t ime varying swimming speeds and frequent burst s w i m m i n g . Horizontal work rate was several orders of magnitude greater than the vert ical work rate. Work done against vert ical buoyancy differences was negligible compared to the work done 140 Chapte r 2: Ultrasonic Tracking against the hydrodynamic drag. Thus , stratification did not l imit or inhibit ver-tical excursions. Calculations of average energy expenditure (cal/track) also agree favourably w i t h Brett (1983). However, several tracks show caloric expenditures beyond what may reasonably be assumed to be available. The normalized aspect rat io , a linear transformation of the non-dimensional ratio of horizontal to vertical distance travelled, gave a clear indication of homeward orientation. However, scatter plots for individual and composite fish tracks revealed no relation between horizontal and vertical velocity. Scatter plots of depth and normalized vertical velocity, V ^ , were more pos-itively correlated in regimes that were more weakly stratified. T h i s result is in concert w i t h the qualitative observation of the near-surface depth distr ibution of fish tracked in Discovery Passage and eastern Johnstone Strait . Depth and Vv were more strongly correlated in oriented than disoriented fish and this is an indication that well oriented fish remain at relatively constant depths for extended periods of time. Observations of depth and normalized T gradient, (dT/dz)n showed that sockeye were more often observed at depths associated w i t h the m i n i m u m in (dT/dz)v, especially in regimes characterized by strong stratif ication. In regimes of weak or nearly uni form vert ical gradients, observations of fish depth spanned the range of observed (dTjdz)^. Ver t i ca l distr ibution relative to (dT/dz)^ was similar i n different oceanographic regimes regardless of the magnitude of the the gradients if the form of the stratif ication was similar. Scatter plots of (dT/dz)^ and V ^ were very weakly correlated and a consistent conclusion regarding depth, Vv and (dT/dz)r,\s such that sockeye d i d not remain at depths associated w i t h the 141 Chapter 2: Ultrasonic Tracking m a x i m u m gradients, but frequently swam through them. Dive/ascent characteristics revealed two dist inct modes of vertical movements. T h i s result is in concert wi th those obtained from the spectral analysis. Time between successive dives showed that disoriented fish made frequent, relatively deep dives and rarely remained at constant depth while oriented fish made infrequent deep, but frequent shallow dives and often remained at constant depths. T h e slope of the least-squares linear fit to dive durat ion versus depth data shows that the rate of change of depth is, in general, greater for disoriented fish in all oceanographic regimes. In addi t ion, the rate of change of depth for oriented fish is greater in regimes that are more weakly stratified. Future field work in this area should be implemented so that the vert ical , hor-izontal and spatial resolution of the physical oceanographic measurements and the vert ica l resolution of the tracking data be improved. The vertical and horizontal scales associated w i t h the hydrographic parameters measured in this study were in the order of one and hundreds of meters respectively. In the future, measurements of these parameters and current shears should be made at scales as close as pos-sible to the estimated sensing capabilities of the fish. Free-falling instrumentation and mult iple vessels could provide the much more highly resolved data required. These data would not only provide a far more complete qualitative description of the tracked fish's vertical d is t r ibut ion , but also temperature, salinity and depth at scales commensurate wi th those of hydrographical ly borne orientation cues. C o m -plete quantif ication of the relationship between the physical oceanography and the tracked sockeye's vertical dis tr ibut inon must await improved field and theoretical studies. 142 Chapter 2: Ultrasonic Tracking A s no clear quantitative relationship was found between physical oceano-graphic variables and vertical distr ibutions, dimensional analysis was employed in an at tempt to reveal the functional form of any relationship that might exist in the data . T h i s attempt was unsuccessful and it can be concluded that physical variables alone were insufficient to specify the vertical d is t r ibut ion of the tracked sockeye. T h e numerous other attempts discussed previously would seem to support this generalization. A n effective quantification of the sockeye's vertical distr ibution must somehow include the physiological and motivational state of the fish. 143 References References Arnold, G.P. and P.H. Cook, 1984: Fich migration by selective tidal stream transport: first results with a computer simulation model for the European continental shelf. In: McCleave, J.D., G.P. Arnold, J.J. Dodson and W.H. Neill (eds.), Mechanisms of Migration in Fishes. Plenum Press, New York, New York. p227-261. Bainbridge, R., 1961: Problems of fish locomotion. Vertebrate Locomotion Sympo-sium of the Zoological Society of London. 5: 13-32. Bardach, J.E. and R.G. Bjorklund, 1957: The temperature sensitivity of some Amer-ican freshwater fishes. American naturalist, 91: 233-251. Blake, R.W., 1983: Fish Locomotion. University Press, Cambridge, England. 203pp. Bendat, J.J and A.G. Piersol, 1971: Random Data: Analysis and Measurement Pro-cedures. Wiley Interscience, New York, New York. 407pp. Benjamin, T.P., 1968: Gravity currents and related phenomena. J. Fluid Mech., 31: 209-248. Binder, R., 1973: Fluid Mechanics. Prentice-Hall, Inc. Englewood cliffs, N.J.. 401pp. Bertmar, G. and R. Toft, 1969: Sensory mechanisms of homing in salmonid fish. I. Introductory experiments on the olfactory sense in grilse of Baltic salmon (Salmo salar). Behaviour, 35: 235-241. Bingman, C, M.D. Godfrey and J.W. Tukey, 1967: Modern techniques of power spectral analysis. Trans. IEEE Audio and Electroacoustics, AU-15. Bingman, V.P., 1981: Savannah sparrows have a magnetic compass. Anim. Behav., 29 : 962-963. Blackbourn, D.J., 1987: Sea surface temperature and preseason prediction of return timing in Fraser River sockeye salmon Oncorhynchus nerka. Can. Sp. Publ. Fish. Aquat. Sci., 96: 296-306. Bowman, M.J. and W.E. Essais, (eds), 1978: Oceanic Fronts in Coastal Processes. Proceedings of a Workshop Held at the Marine Sciences Research Center, May 25-27, 1977. Springer-Verlag, Berlin Heidelberg New York. 114pp. 144 References Brafield, A.E., 1985: Laboratory studies of energy budgets. In: Tytler, P. and P. Cal-low (eds.), Fish Energetics. John Hopkins University Press. Baltimore, Mary-land. p283-307. Brannon, E.L., 1972: Mechanisms controlling migration of sockeye salmon fry. Int. Pacific Salmon Fish. Comm. Bull., 21: 1-86. Brannon, E.L., 1982: Orientation mechanisms of homing salmonids. In: Brannon, E.L. and E.M. Salo (eds.), Proceedings of the Salmon and Trout Migratory Be-haviour Symosium. School of Fisheries, University of Washington, Seattle, Wa.. p219-227. Brett, J.R., 1965: The swimming energetics of salmon. Scientific American, 213: 80-85. Brett, J.R., 1983: Life energetics of sockeye salmon Oncorhynchus nerka. In: Osprey, W.P. and S.I. Lustick (eds.), Behavioural Energetics: The cost of Survival in Vertebrates. Ohio State University Press, Columbia, Ohio. p29-63. Brett, J.R. and N.R. Glass, 1973: Metabolic rates and critical swimming speed of sockeye salmon Oncorhynchus nerka in relation to size and temperature. J. Fish. Res. Board Can., 30: 379-387. Brett, J.R. and T.P. Groves, 1979: Physiological energetics. In: Hoar, W.P., D.J. Randall and J.R. Brett (eds.), Fish Physiology, Vol. VIII. New York Academic Press, New York, New York. p279-352. Buckley, J.P. and S. Pond, 1976: Wind and surface circulation of a fjord. J. Fish. Res. Bd. Can., 10: 2265-2271. Bull, H.D., 1936: Studies on conditined responses in fishes. Part VII. Temperature perception in teleosts. Journal of the Marine Biological Association of the United Kingdom, 21: 1-27. Burgner, R.L., 1980: Some features of ocean migrations and timing of Pacific salmon. In: McNeill, W.J. and D.C. Himsworth (eds.). Salmonid Ecosystems of the North Pacific. Oregon State University Prees, Corvallis, Oregon. pl53-164. Daniel, F.N., 1985: Summer circulation of the waters in Queen Charlotte Sound. Atmos. Ocean, 23: 393-413. 145 References Dodimead, A.J. 1980: A general review of the oceanography of the Queen Char-lotte Sound-Hectate Strait-Dixon Entrance region. Can. Manuscr. Rep. Fish. Aquat. Sci. No. 1574- Dep. Fish, and Oceans, Pac. Biol. Stn., Nanaimo, B.C.. 248pp. Dodson, J.J. and L.A. Dohse, 1984: A model of olfactory mediated conditining of directional bias in fish migrating in reversing tidal currents based on the homing migratino of American Shad (Alosa sapidisisima). In: McCleave, J.D., G.P. Arnold and W.H. Neill (eds.). Mechanisms of Migration in Fishes. Plenum Press, New York, New York. p263-281. D0ving , K.B., H. Westerberg and P.B. Johnson, 1985: Role of olfaction in the behav-ioral and neuronal responses of Atlantic salmon, Salmo salar, to hydrographic stratification. Can. J. Fish. Aqua. Sci., 42: 1658-1667. Farmer, D.M. and H.J. Freeland, 1983: The physical oceanorgraphy of fjords. Progress in Oceanography, 12: 147-219. Ewart, T.E. and W.P. Bendiner, 1981: An observation of the horizontal and vertical diffusion of a passive tracer in the deep ocean. JPO, 5: 95-107. Favorite, F., 1961: Surface temperatures and salinity off the Washington and British Columbia coast, August 1958 and 1959. J. Fish. Res. Bd. Can., 18: 311-319. French, R., H. Bilton, M. Osako and A. Hartt, 1976: Distribution and origin of sockeye salmon (Oncorhynchus nerka) in offshore waters of the North Pacific Ocean. IPSFC Bull. 34, 113p. Frisch, A.S., J. Holbrook and A.B. Ages, 1981: Observations of a summertime reversal in the circulation in the Strait of Juan de Fuca. J. Geophys. Res., 86j 2044-2048. Gee, J.H., 1983: Ecologic implications of buoyancy control in fish. In: Webb, P. and D. Weihs (eds.), Fish Biomechanics. Praeger, New York, New York. 398pp. Geyer, W.R. and G.A. Cannon, 1982: Sill processes related to deep water renewal in a fjord, J. Geo. Res., 87: 7985-7996. Gill, A.E., 1982: Atmosphere-Ocean Dynamics. Academic Press, Inc., New York, New York. 661pp. 146 References Godin, G.J., J. Candelar and R. de-la-Paz-Vela, 1981: On the feasability of detecting net transports in and out of Georgia Strait with an array of current meters. Atmos. Ocean, 19: 148-157. Groot, C., 1965: Onthe orientation of young sockeye salmon (Oncorhynchus nerka) during their migration out of lakes. Behav. Suppl., 14: 1-198. Groot, C., 1972: Migration of yearling sockeye salmon (Oncorhynchus nerka)a.s de-termined by time-lapse photography of sonar observations. J. Fish. Res. Bd. Can., 29: 1431-1444. Groot, C. and T.P. Quinn, 1987: The homing migration of sockeye salmon, On-corhynchus nerka, to the Fraser River. Fish. Bull., 85: 455-469. Hamilton, K., 1985: A study of the variability of the return migration route of Fraser River sockeye salmon (Oncorhynchus nerka). Can. J. Zool, 63: 1930-1943. Hasler, A.D. and A.T. Scholz, 1983: Olfactory Imprinting and Homing in Salmon. Springer-Verlag, Berlin. 134pp. Ichihara, T. and A. Nakamura, 1982: Vertical movement of mature chum salmon contributing to the improvement of net structure on the Hokkaido coast. In: Melteff, B.R. and R.A. Neve (eds.), Proceedings of the North Pacific aquaculture symposium. Alaska Sea Grant, Fairbanks, AK. p39-49. International Pacific Salmon Fisheries Commission (IPFSC), 1984: Annual Report, 1983. New Westminister, B.C.. Jenkins, G.M. and D.G. Watts, 1968: Spectral Analysis and its Applications. Holden-Day, San Francisco, Ca. 525pp. Johnson, P.B. and A.D. Hasler, 1980: The use of chemical cues in the upstream migration of coho salmon, Oncorhynchus kisutch, Walbaum. J. Fish. Biol., 17: 67-73. Landreth, J.F., 1973: Orientation and behaviour of the rattlesnake, Crotalus atrox. Copeia, 1973: 26-31. LeBlond, P.H., 1983: The Strait of Georgia: functional anatomy of a coastal sea. Can. J. Fish. Aq. Sci., 40: 1033-1063. LeBlond, P.H. and L.A. Mysak, 1978: Waves in the Ocean. Elsevier Scientific Pub-lishing Co., Amsterdam. 602pp. 147 References Leggett, W.C., 1977: The ecology of fish migrations. A. Rev. Ecol. Syst.,8: 285-308. Lighthill, M.J., 1971: Large-amplitude elongated body theory of fish locomotion. Proc. Roy. Soc. (B), 179: 125-138. Magnuson, J.J., 1978: Locomtion by scombroid fishes: Hydromechanics, morphology and behaviour. Fish Physiology, 7: 239-313. McKeown, B.A., 1984: Fish Migration. Finiker Press, Beaverton, MC. 217pp. Muench, R.D., H.J.S. Fernando and G.R. Stegen, 1990: Temperature and salinity staircases in the northwestern Weddell Sea. JPO, 20: 295-306. Moore, C, 1986: Discrete Fourier Transforms with Application to Fourier Series and Convolution Integrals. UBC Computing Services Publication Mysak, L.A., C. Groot and K. Hamilton: 1986: A study of climate and fisheries: Interannual variability of the northeast Pacific ocean and its influence o homing migration routes of sockeye salmon. Climatological Bulletin, 20: 26-35. Mysak, L.A., H.Hsieh and T. Parsons, 1982: On the relationship between interan-nual baroclinic waves and fish populations in the northeast Pacific. Biological Oceanography, 2: 63-103. Neumann, G. and W.J. Pierson: 1966. Principles of Physical Oceanography. Prentice-Hall, Inc. Englewood Cliffs, N.J.. 544pp. Pickard, G.L., 1975: Annual and longterm variations of deepwater properties in the coastal waters of southern British Columbia. J. Fish. Res. Bd. Can., 9: 1561-1587. Pond, G.S. and G.L. Pickard, 1983: Introductory Dynamic Oceanography. 2 n d ed.. Pergammon Press, Oxford. 329pp. Prandtl, L. and O.C. Tietjen, 1957: Applied Hydro and Aeromechanics (2nd ed.). Dover Press, New York, New York. Quinn, T.P., 1980: Evidence for celestial and magnetic compass orientation in lake migrating sockeye salmon fry. J. comp. Physiol., 137: 243-248. 148 References Quinn, T.P., 1982: A model for salmon navigation on the high seas. In: Brannon, E.L. and E.O. Salo (eds.), Proceedings of the salmon and trout migratory behaviour symosium. College of Fisheries, University of Washington, Seattle, Wa. p229-237. Quinn, T.P. and E.L. Brannon, 1982: The use of celestial and magnetic cues by orienting sockeye slamon smolts. J. comp. Physiol., 147: 547-552. Quinn, T.P. and B. terHart, 1987 Movements of adult sockeye salmon (Oncorhynchus nerka) in British Columbia coastal waters in relation to temperature and salinity stratification: Ultrasonic telemetry results. Can. Spec. Pub. Fish. Aquat. Sci. 96, p61-77. Quinn, T.P. and C. Groot, 1984: The effect of water flow rate onbimodal orientation of juvenile chum salmon, Oncorhynchus keta. Anim. Behav., 32: 628-629. Quinn, T.P., B. terHart and C. Groot, 1989: Migratory orientation and vertical movements of homing adult sockeye salmon, Oncorhynchus nerka, in coastal waters. Anim. Behav., 37: 587-599. Royal, L.A. and J.P. Tully, 1961: Relationship of variable oceanographic factors to migration and survival of Fraser River salmon. Cal. Coop. Ocean Fish. Invest. Rep., 8: 65-68. Royce, W.F., L.S. Smith and A.C. Hartt, 1968: Models of oceanic migrations of Pacific salmon and comments on guidance mechanisms. Fish. Bull., 66: 441-462. Royer, L. and W.J. Emery, 1982: Variations of the Fraser River plume and their relationship to forcing by tide, wind and discharge. Atmos. Ocean, 20: 357-372. Schumacher, J.D., CD. Perason, R.L. Charnell and N.P. Laird, 1978: Regional re-sponse to forcing in southern Strait of Georgia. Est. Coast. Mar. Sci., 7: 79-91. Simpson, J.H., 1975: Observations of small scale vertical shear in the ocean. Deep-Sea Research, 22: 619-627. Soofiani, N.M. and A.D. Hawkins, 1985: Field studies of energy budgets. In: Tytler, P. and P. Callow (eds.), Fish Energetics, John Hopkins University Press, Balti-more, Maryland. p283-307. 149 References terHart, B.A., 1988: A multi-year hydrography of British Columbia's inside passage. Dep. of Ocean. UBC Manuscr. Rep.. No. 55 terHart, B.A., A.T. Weaver and T.P. Quinn, 1987: Vertical and horizontal distribu-tions of homeward migrating sockeye salmon in relation to temperature, salinity and current observations in coastal waters: Moist 1985. Dep. of Ocean. UBC Manuscr. Rep.. No. 49. terHart, B.A.and T.P. Quinn, 1989: Vertical and horizontal distributions of home-ward migrating sockeye salmon in relation to temperature, salinity and current observations in coastal waters: Moist 1986. Dep. of Ocean. UBC Manuscr. Rep.. No. 56. Thomson, K.P., B.A. terHart and P. Welch, 1985: A CTD survey of the inside passage of British Columbia 24-28 June 1985. Dept. of Ocean. UBC Manuscr. Rep.. No. 45. Thomson, K.P., B.A. terHart and T.P. Quinn, 1986: Ambient oceanographic variables of Oncorhynchus nerka (MOIST 1985). Dept. of Ocean. UBC Manuscr. Rep.. No. 50. Thomson, R.E., 1976: Tidal currents and estuarine circulation in Johnstone Strait, British Columbia. J. Fish. Res. Board Can., 33: 2242-2264. Thomson, R.E., 1977: Currents in Johnstone Strait, British Columbia: supplement data on the Vancouver Island site. J. Fish. Res. Board Can., 34: 697-703. Thomson, R.E., 1981: Oceanography of the British Columbia coast. Can. Spec. Pub. Fish. Aquat. Sci. 56. 291p. Tolmazin, D., 1985: Elements of Dynamic Oceanography. Allen & Gunnison, Inc., Winchester, Mass. 181pp. Tully, J.P., A.J. Dodimead and S. Tabata, 1960: An anomalous increase of tempera-ture in the ocean off the Pacific coast of Canada through 1957 and 1958. J. Fish. Res. Bd. Can., U: 61-80. Van Leer, J.C. and C.G. Rooth, 1975: Shear observations in the deep thermocline. Deep-Sea Research, 22: 831-836. Weast, R.C. (ed.), 1974: Handbook of Chemistry and Physics. CRC Press, Cleveland, Ohio. 150 References Webb, P.W., 1978: Hydromechanics: Non-scombroid fish. Fish Physiology, 3: 189— 237. Webb, P.W. and D. Weihs, 1983: Fish Biomechanics. Praeger, New York, New York. 381pp. Weihs, D., 1987: Hydromechanics of fish migration in variable environments. Amer-ican Fisheries Symposium, p254-261. Westerberg, H., 1982: Ultrasonic tracking of Atlantic Salmon (Salmon Salar L.) II. Swimming depth and temperature stratification. Inst. Freshwater Res. Drot-tningholm. Rep., 60: 102-120. Westerberg, H., 1984: The orientation of fish and the vertical stratification at fine and microstructure scales. In: McCleave, J.D., G.I. Arnold and J.J. Dodson (eds.), Mechanisms of Migration in Fishes. Plenum Press, New York, New York. P179-203. Wickett, W.P., 1975: Relationship of coastal oceanographic factors to the migration of Fraser River sockeye salmon (Oncorhynchus nerka). Int. Comm. Explor. Sea. Cm., 1977/M.26: pl8. Woods, J.D., 1968: Wave-induced shear instability in the summer thermocline. JFM, 32 : 791-800. Xie, 1. and W.W. Hsieh, 1989: Predicting the return migration routes of the Fraser River sockeye salmon (Oncorhynchus nerka). Cdn. J. Fish. Aqua. Sci., 46: 1287-1292. Young, B. and A.E. Hay. 1987: Density current flow into Fortune Bay, Newfoundland. J.P.O., 17: 1066-1071. 151 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053104/manifest

Comment

Related Items