Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Water properties, currents and zooplankton distribution over a submarine canyon under upwelling-favorable… Vindeirinho, Carine 1998

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

Item Metadata

Download

Media
831-ubc_1998-0653.pdf [ 7.9MB ]
Metadata
JSON: 831-1.0053340.json
JSON-LD: 831-1.0053340-ld.json
RDF/XML (Pretty): 831-1.0053340-rdf.xml
RDF/JSON: 831-1.0053340-rdf.json
Turtle: 831-1.0053340-turtle.txt
N-Triples: 831-1.0053340-rdf-ntriples.txt
Original Record: 831-1.0053340-source.json
Full Text
831-1.0053340-fulltext.txt
Citation
831-1.0053340.ris

Full Text

W A T E R PROPERTIES, C U R R E N T S A N D Z O O P L A N K T O N DISTRIBUTION O V E R A S U B M A R I N E C A N Y O N U N D E R U P W E L L I N G - F A V O R A B L E CONDITIONS by Carine Vindeirinho ENSTA School of Engineering, Paris, France A THESIS 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 G 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 STUDIES E A R T H A N D O C E A N SCIENCES ( O C E A N O G R A P H Y ) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F BRITISH C O L U M B I A July 1998 © Carine Vindeirinho, 1998 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. Earth and Ocean Sciences (Oceanography) The University of British Columbia 2075 Wesbrook Mall Vancouver, B C Canada V6T 1Z1 Abstract Two canyons (Juan de Fuca and Barkley) off the west coast of Vancouver Island show enhanced upwelling during the summer. Barkley Canyon is also a region of zooplankton aggregation. An intensive survey was conducted to characterize their circulation, water properties and, for Barkley, zooplankton distribution. CTD casts and zooplankton tows were performed in 1997 around Barkley Canyon. Current meter moorings were positioned within both canyons (Juan de Fuca in 1993 and Barkley in 1997). The circulation in Juan de Fuca Canyon is characterized by a southeastward flow at 30 and 75 m but a northward, in-canyon flow at 200 m. The deep flow supports a strong and consistent upwelling from May to October with a strong decrease in temperature equiv-alent to a 200-m vertical displacement. The northward-flowing California Undercurrent is found to have no significant influence on the upwelling. In Barkley Canyon, there is a weaker upwelling signal that nonetheless explains an important variance in the CTD data. Isopleths upwell towards the head of the canyon and dome over it near the surface. This implies that the influence of the canyon reaches shallower depths than in Astoria Canyon, for example. The circulation of the canyon is assessed from the mooring data: the flow bends over the canyon at 150 m, with a possible cyclonic eddy at the head, while the current is mainly cyclonic at depth. Some in-canyon flow events are observed, which result in moderate upwelling. Zooplankton are found to aggregate along the walls near the head and mouth of Barkley Canyon with different distribution patterns according to the species' vertical and spatial preferences. Interaction with the circulation, especially the deep cyclonic flow, seems to play an important role in this aggregation. ii Table of Contents Abstract ii List of Tables vi List of Figures vii Acknowledgments x 1 Introduction 1 1.1 Submarine canyons 1 1.1.1 General features 1 1.1.2 Circulation around canyons 1 1.1.3 Important parameters 4 1.2 Biological implications on zooplankton and fisheries 6 1.2.1 Advection and upwelling 6 1.2.2 Diel and ontogenetic migration 7 1.3 Oceanographic setting 10 1.3.1 Oceanic circulation 10 1.3.2 Hydrography 12 1.4 Objectives 14 2 Juan de Fuca Canyon 16 2.1 Study area 16 2.2 Winds and currents 18 iii 2.3 Water properties and upwelling 26 2.4 The northward-flowing California Undercurrent 31 2.5 Summary 38 3 Physical Properties of Barkley Canyon: the C T D data 39 3.1 The geographical setting of Barkley Canyon 39 3.2 Processing of the data 44 3.3 Upwelling within Barkley Canyon 45 3.3.1 Variations of the water properties 46 3.3.2 Upwelling in the canyon . . 50 3.4 Comparison with Astoria Canyon 58 3.5 Influence of the internal tide 62 3.6 Summary 67 4 Physical Properties of Barkley Canyon: the mooring data 68 4:1 Winds around Barkley Canyon 68 4.2 Currents in the vicinity of Barkley Canyon 70 4.2.1 Theory and model predictions 70 4.2.2 Currents at mooring locations in Barkley Canyon 73 4.2.3 Currents at mooring A l 81 4.3 Upwelling episodes in Barkley Canyon 83 4.3.1 Temperature and salinity 83 4.3.2 Three upwelling events 88 4.4 Summary 92 5 Circulation patterns and zooplankton transport 93 5.1 Introduction 93 iv 5.1.1 Zooplankton species on the west coast of Canada . 94 5.1.2 Zooplankton vertical patterns 95 5.2 Zooplankton spatial aggregation 96 5.3 Links between physics and biology 101 5.4 Error analysis 103 6 Conclusions 105 6.1 Juan de Fuca Canyon 105 6.2 Barkley Canyon 106 6.3 Zooplankton aggregation 107 6.4 Further work 108 Bibliography 110 A Empirical Orthogonal Function Analysis 115 A . l Theoretical purpose of E O F analysis 115 A.2 Practical use of E O F analysis 115 A.2.1 The E O F vectors 115 A.2.2 The E O F contributions 116 A.3 Application of the E O F analysis for this thesis 117 B Zooplankton data from the Barkley Canyon study 118 v List of Tables 2.1 Values of temperature and salinity at two different stations outside Juan de Fuca Canyon 30 3.1 Location of the stations sampled around Barkley Canyon 43 3.2 Location of the moorings in Barkley Canyon 44 vi List of Figures 1.1 Schematic of differential transport of adult and larval euphausiids . . . . 8 1.2 Geography and depth contours for the south coast of British Columbia and the north coast of Washington 11 1.3 Circulation off the British Columbia/Washington coast 13 2.1 Geographical area around Juan de Fuca Canyon 17 2.2 Daily-averages of the wind velocity at Destruction Island 19 2.3 Daily-averages of the currents from JF03 at 30, 75 and 200 m 21 2.4 Time-series of the alongshore current from JF03 at 30 m and alongshore wind at Destruction Island 22 2.5 Vertical profiles of salinity for three stations on the shelf 23 2.6 Time-series of the cross-canyon current at 75 m and up-canyon current at 200 m from JF03 25 2.7 Daily-averages of the temperature from JF03 at 30, 75 and 200 m . . . . 26 2.8 Daily-averages of the salinity from JF03 at 30, 75 and 200 m 27 2.9 Daily-averages of the pressure illustrating the depth from which water at 200 m was upwelled 29 2.10 Monthly-averaged currents from JF03 at 30, 75 and 200 m for May 1993 32 2.11 Monthly-averaged currents from JF03 at 30, 75 and 200 m for August 1993 33 2.12 Monthly-averaged currents from JF03 at 30, 75 and 200 m for September 1993 34 2.13 Daily-averages of the current from A l at 400 m 36 vii 2.14 Time-series of the cross-canyon current at 75 m and temperature at 200 m from JF03 37 3.1 Geographical area around Barkley Canyon 40 3.2 Stations sampled around Barkley Canyon and mooring locations 42 3.3 Vertical profile of salinity at station BCE4 47 3.4 Vertical profiles of temperature and density at station BCE4 48 3.5 Temperature/Salinity diagrams for all stations 49 3.6 Cross-section of salinity along line BCB 52 3.7 Cross-section of salinity along line BC3 53 3.8 Cross-section of salinity along line B of La Perouse Grid 54 3.9 Contours of the depth of the 31.5 isohaline 56 3.10 Contours of the depth of the 32.5 and 33 isohalines 57 3.11 Astoria Canyon off the Columbia River on the Washington Shelf 58 3.12 Contoured section of temperature at the mouth across Astoria Canyon . 60 3.13 Vertical profiles of salinity at stations BCB4 and BCD4 on July 25th, 26th and 27th, 1997 63 3.14 Contours of the contribution of the first EOF to stations around Barkley Canyon 65 4.1 Daily-averages of the wind velocity from buoy 46206 69 4.2 Model output for the horizontal velocity at 150 m 71 4.3 Model output for the horizontal velocity at 350 m 72 4.4 Daily-averages of the currents from BC01 at 157, 257 and 357 m 74 4.5 Daily-averages of the currents from BC02 at 250 and 350 m 75 4.6 Daily-averages of the currents from BC03 at 142, 242 and 342 m . . . . . 76 4.7 Daily-averages of the current from BC04 at 230 m 77 viii 48 Monthly-averaged currents at 150, 250 and 350 m in July 1997 80 4.9 Daily-averages of the currents from A l at 32, 97, 172 and 397 m 82 4.10 Daily-averages of the temperature from BC01 at 157, 257 and 357 m . . 84 4.11 Daily-averages of the temperature from BC02 at 250 and 350 m 85 4.12 Daily-averages of the temperature from BC03 at 142, 242 and 342 m . . 86 4.13 Daily-averages of the temperature from BC04 at 230 m 87 4.14 Contours of temperature from the thermistor chain at A l 90 5.1 Chart of the expected species' region preferences and of the canyon study results 98 5.2 Relative abundance of Acartia longiremis and Euphausia calyptopis in the 50-m tows during the day . 99 ix Acknowledgments I would like to say "thank-you" to everyone who helped me, each in their own way, to achieve my goals and write this thesis. First of all, I would especially like to thank my supervisor, Susan Allen, who offered me this project and helped me constantly with her advice and scientific insight. Thanks also to Rick Thomson for allowing me to use the data from his cruises and for offering me his wisdom and incommensurable intelligence. Thank-you to Mike Foreman and Dave Mackas who provided me with useful feedback on physics and biology, respectively. Thanks to Robert Goldblatt for helping me interpret the biological data associated with this thesis, to Moira Galbraith for counting the animals and to Shannon Harris for the nutrient analysis. A special thanks goes to my friends in France who let me leave for that far-away place called Canada and kept me posted on what was going on back home: Isabelle et Jean-Yves, Lydia, Nathalie et Laurent, Estelle et Romain. Thanks also to my friends here who supported me during those two years and helped me believe in myself: David, Erica, Alison, Kyle, Jason, Pal, Christine, Rina and especially Bruno with all my love. Finally, from the deepest part of my heart, I want to thank my parents for all they did for me: for their love, care and support, although they were far away. You were always present in my heart and never doubted me. I love you immensely and I hope you are proud of what I have achieved. Je vous aime enormement et j'espere que vous etes tiers de ce que j'ai accompli. x Chapter 1 Introduction 1.1 Submarine canyons 1.1.1 General features Submarine canyons are important features of the coastal regions in the world's oceans. They are steep, V-shaped troughs cutting the slope of many continental shelves, much like canyons cut by rivers on land. These ubiquitous features along the west coast of North America (Shepard et al., 1979) cut the continental shelf at what is called the head and open at the deep-sea fan on the abyssal plains at the mouth. They exist only when the slope is greater than 3° (Thurman, 1994) and are thought to be erosional features formed by the action of turbidity currents. In 1936, Richard Daly first presented the idea that highly erosive turbidity flows of sediment-laden water moving down-slope would carve the canyons. Francis P. Shepard later conducted many experiments to study the formation of canyons and to support this theory. 1.1.2 Circulation around canyons Several studies have been conducted in the past few years to show that submarine canyons influence the local circulation. Most of these are modelling studies but field studies were also done. One of the models Allen (1996) uses is a homogeneous, linear model with a layered stratification and the Ekman layer effect represented by a sink at the coast and no 1 Chapter 1. Introduction 2 friction. She shows that, in the vicinity of a canyon, the onshore flux across the shelf break is reduced and redirected through the canyon. The response of a stratified fluid to a horizontal pressure gradient imposed at the top of a canyon system is studied by Freeland and Denman (1982). They point out that Juan de Fuca Canyon induces an enhanced upwelling with a baroclinic pressure gradient opposing the barotropic gradient (see Section 1.3.2). Chen and Allen (1996) use a homogeneous, inviscid, linear model and show that an infinitely long, flat-bottom canyon acts as a complete barrier to an approaching geostrophic shelf flow. However, an infinitely long, stepped-bottom canyon is not a complete barrier and the flow crosses the isobaths. They forecast that the results would be different if viscosity or advection were added. A two-level (ocean-canyon), homogeneous, inviscid, rotationally modified model used by Klinck (1988) also shows that submarine canyons in steady state act as barriers to the barotropic geostrophic flow. Freeland and Mcintosh (1989) show that the variations in vorticity with time on the shelf south of British Columbia are being driven by the vertical velocities in the canyon. These in turn are driven by a deep northward flow along the axis of the canyon, closely related to the southward shelf-break flow. The overall circulation pattern in canyons is that, in the upper water column, the flow directly crosses the canyon most of the time (see the next section for exceptions). Within about 50 m of depth of the rim, the flow crosses the isobaths and leads to a cyclonic vortex on the upstream side and an anticyclonic vortex on the downstream side. This is associated with the change in vorticity as the water column stretches when falling into the canyon (Hickey, 1995). Finally, below the rim, the flow follows the direction of the local isobaths and is cyclonic (Allen, 1996; Hickey, 1997). More precisely, Klinck (1996) models the circulation around a canyon and finds good agreement with the observed flows. His model is non-linear and includes linear vertical density stratification and weak friction. He divides his study into four sections and finds the following: Chapter 1. Introduction 3 • Weak stratification, downwelling-favorable winds (right-bound) - The flow is to the right over the shelf (except over the canyon where it is cyclonic) and cyclonic inside the canyon below the shelf break. The current speed around the rim is double the speed offshore. - There is downwelling at the upstream side and weaker upwelling on the down-stream side. • . Strong stratification, downwelling - The horizontal flow above the shelf is very uniform and there is a slight amount of cyclonicity over the canyon while within the canyon, the deep circulation is weak. - The vertical velocity is the same as before but weaker and less extended. • Weak stratification, upwelling-favorable winds (left-bound) - Over the shelf, the flow is to the left directed onshore as a strong jet along the coastal walls. Below the shelf break, the water turns towards the coast within the canyon and there is a weak cyclone downstream of the canyon. - The vertical velocity is upward everywhere at the top of the canyon except along the upstream rim where there is downwelling. • Strong stratification, upwelling - Below the shelf, there is a strong turning of the flow into the canyon at the upstream side. The water entering the canyon splits in two, one branch con-tinuing the cyclonic turn towards the head, the other turning anticyclonically and exiting the canyon. Chapter 1. Introduction 4 - There is a strong upwelling over the entire top of the canyon and two small down-wellings at the upstream corner and immediately downstream of the canyon. The offshore distance over which the density field is disturbed is larger than in the previous case. These studies show that during upwelling-favorable conditions, the flow near the sur-face passes over the canyon without being disturbed, crosses the isobaths above the rim and turns cyclonically into the canyon below the rim. 1.1.3 Important parameters Three numbers are important when studying canyons. They are the Rossby number, temporal Rossby number and Burger number. The Rossby number1 is Ro = u/ fL (where u is the velocity, / the Coriolis parameter and L a horizontal lengthscale) and measures the importance of the advection (and thus of the non-linearities) compared to the Coriolis force in the momentum equation. As Hickey (1995, 1997) shows, the higher the Rossby number, the less the flow over the canyon "feels" the canyon and the less it turns to follow the isobaths. If the Rossby number is smaller, the flow turns to follow the isobaths on the upstream side. Perenne et al. (1997) assess the dependence of a homogeneous flow on the values of both the Rossby and temporal Rossby numbers. The temporal Rossby number1 Rot = cj/f (where OJ is the frequency) measures the importance of the time-dependence compared to the Coriolis force. They use an oscillatory current in laboratory experiments (tanks) in addition to a numerical model. They find two regimes in the laboratory exper-iments: one for small Rot (~ < 0.5) and all Ro investigated (0.03 < Ro < 0.25) where cyclones form on the canyon walls (the cyclone regime); one for large Rot (~ > 0.5) and 1 For more detail, see Pedlosky (1987) or Gill (1982). Chapter 1. Introduction 5 all Ro where closed eddies do not have time to form because the background flow does not separate from the canyon boundaries (the cyclone-free regime). Their model uses the rotating, incompressible, hydrostatic and Boussinesq approximations with forcing and dissipation (in the form of lateral but not vertical friction). The normalized excursion amplitude X is the parameter used in the model. It is the ratio of the tidal excursion to the width of the canyon (X = (u/oS)/W = Ro/Rot). Perenne et al. (1997) then plot the normalized maximum mean flow versus the normalized excursion amplitude. For X > 0.4, the normalized maximum mean flow is found to be independent of X and of the same order as the background flow (in the cyclone regime). For X < 0.4, the plot separates into 1) a lower branch with small mean flow increasing with X (cyclone-free regime) and 2) an upper branch with the mean flow decreasing with X (cyclone regime). Finally, the Burger number1 is B = fL/NH, with N the Brunt-Vaisala frequency. It quantifies the flow stratification, which controls the interaction of the coastal flow with canyons. Klinck (1996) states that it also controls the magnitude of the response of the canyon and can limit the influence of the canyon on the overlying flow, i.e. as stratification increases, the effect of the canyon decreases. A canyon is considered narrow if its width is less than about half of the smaller of the radius of deformation (R = ^JgH/ f = NH/ f) or the spatial scale of the current (Klinck, 1988). In a narrow canyon, the circulation is not geostrophic since the pressure gradient is not balanced by a cross-canyon circulation. If the canyon is narrow enough, this cross-channel flow is highly reduced. Klinck (1989) uses a stratified, rotating, linear, inviscid model and includes the feedback of the upwelled dense water on the cross-shelf pressure gradient. He shows that the width of the canyon determines the strength of the interaction between the canyon and the shelf flow, with the interaction decreasing as the canyon becomes narrower than the radius of deformation. Yet, the density is always redistributed even in very narrow canyons, which means that 1 For more detail, see Pedlosky (1987) or Gill (1982). Chapter 1. Introduction 6 there is always a density signal in the proximity of a canyon. These parameters (Ro, Rot and B) are used to assess the extent of the canyon's influence on the flow, especially near the surface. 1.2 Biological implications on zooplankton and fisheries Zooplankton (especially euphausiids) are the most important prey for many harvested fish off British Columbia (hake, salmon, herring) (Simard and Mackas, 1989; Mackas, 1992). Current patterns can have an effect on zooplankton biomass, which in turn can have implications on the feeding pattern of larval fish. Zooplankton need to interact and adapt to the prevailing mean and seasonal currents in order to remain abundant (Peterson et al, 1979; Wroblewski, 1982; Roughgarden et al, 1988). Mackas and Sefton (1982) find that distinct species' assemblages are coincident with large scale patterns of currents and water masses. The timing of the life cycle of various zooplankton species with the Spring Transition (when winds reverse from their winter pattern to their summer pattern) is also known to have an effect on zooplankton growth, as does the summer upwelling which affects the productivity over the shelf break. Finally, canyons can be considered regions where zooplankton may be able to escape advection off the shelf. 1.2.1 Advection and upwelling Mackas (1992) shows that zooplankton biomass is determined primarily by transport (advection west into the deep ocean and alongshore) and predation. Loss due to advection is at least comparable to loss caused by local predation. He also reports that the greatest aggregation of zooplankton is at the shelf break and in upwelled water forming at the inner margin of the Shelf-Break Current. The zooplankton are found to migrate up canyons and aggregate along the canyon walls where they can maintain their position Chapter 1. Introduction 7 against the prevailing advection off- and alongshore. Advection can be a major agent for loss of zooplankton since currents can be large and able to transport particles at a rate of up to 50 km/day, as reported in the vicinity of Quinault Canyon (Hickey et al, 1986). The interaction with the local circulation is two-fold: upwelling provides nutrients to the phytoplankton on which the zooplankton graze, and advection of the animals by the Shelf-Break Current occurs whenever they are in the upper layer. Spatial aggregation is usually associated with bathymetric edges and is governed by three mechanisms (Mackas et al, 1997): 1) enhanced reproduction and/or survivorship, 2) semi-passive accumula-tion/retention by convergent or divergent flow fields and 3) active migration to preferred sites or habitats. On the one hand, advection can cause the loss of animals to the deep ocean and seems to be reduced in the vicinity of canyons. On the other hand, the advective segregation of adult and larval stages minimizes the risk of cannibalism. It appears then that submarine canyons act as traps for the zooplankton (reducing advective losses) while also reducing cannibalism and predation (see Figure 1.1). 1.2.2 Diel and ontogenetic migration Aggregation occurs through the coupling of advection and vertical swimming be-haviour of the animals. In particular, diel2 and ontogenetic3 migrations are of prime importance. As part of the diel migration, the animals are at the surface feeding at dusk and descend to their day-depth before dawn (Simard and Mackas, 1989). During the daytime, the euphausiids are generally at depths of around 150-200 m at the shelf break and thus in the California Undercurrent waters (see Section 1.3.2). This vertical migra-tion transports the animals in different currents during the day and can trap them in a 2 Diel migration is the daily vertical migration. 3 Ontogenetic migration is the vertical migration associated with the life cycle of the animal. Chapter 1. Introduction 8 Om 50 m 100 m 150 m 200 m Larvae © equatorward © ® Adults and y late juveniles poleward (x) L Figure 1.1: Schematic of differential transport of adult and larval euphausiids along the shelf break under summer conditions (after Mackas et al, 1997). Chapter 1. Introduction 9 particular circulation; for example, in the canyons of the Mediterranean Sea, Macquart-Moulin and Patriti (1996) notice that some species accumulate on the upper slope and at the canyon heads where they have been trapped after their morning descent, having travelled over the shelf during their nocturnal migration. Diel migration also helps the adult copepods arriving near the coast to maintain their position in newly upwelled water by migrating vertically down into the onshore bottom flow (Wroblewski, 1982). As far as ontogenetic migration is concerned, Pillar et al. (1989) explain that a cross-shore separation of life stages develops as a consequence of different depth preferences. For most oceanic species, the eggs sink then hatch into nauplii at depth and ascend to the upper layers as feeding larvae. However, in neritic and shelf-dwelling species, the eggs and nauplii have been shown to reside in the upper layers while the larvae and older stages migrate through different depth ranges. This shows that the currents acting on eggs and nauplii are different from those acting on older larvae, juvenile and adult stages. Pillar et al. (1989) illustrate how the deep vertical migration of older larvae interacts with the poleward deep countercurrent to act as a return mechanism into the upwelling Benguela System: the eggs and nauplii are advected northward in the surface layer but the older stages can return southward at depth. Peterson et al. (1979) also report that a favorable interaction of the life history and vertical distribution preferences with the local circulation pattern enables the animals to remain off Oregon, an upwelling-favorable region in summer. Finally, Wroblewski (1982) states that the vertical migration of the adults could interact with upwelling currents to increase the residence time of egg-laying copepods in the nearshore zone. Ontogenetic migration then increases the reproduction rates and allows the zooplankton to remain in a favorable environment. Chapter 1. Introduction 10 1.3 Oceanographic setting The western margin of Vancouver Island, southwest of British Columbia, is char-acterized by an extensive shelf, steep slope and broad continental rise. The maximum width of the continental shelf is 65 km westward of Juan de Fuca Strait and narrows northward to 5 km off Brooks Peninsula. The region is both wider and topographically more complex than the coastal regions to the north and south. The southern end of the shelf is wide, deep and cut by many submarine canyons that extend from Juan de Fuca Strait to Queen Charlotte Islands (Juan de Fuca, Nitinat, Barkley, Clayoquot Canyons). These canyons are separated by a series of shallower banks (70-100 m, e.g. La Perouse Bank) and semi-enclosed basins (150-200 m). Juan de Fuca Strait provides an important outflow of fresh water mainly from the Fraser River through the Strait of Georgia (see Figure 1.2). With the exception of Chapter 2 which deals with Juan de Fuca Canyon at the entrance of Juan de Fuca Strait, the data supporting this thesis were collected around Barkley Canyon situated off Barkley Sound, on the west coast of Vancouver Island. This canyon is known to be highly "visited" by fishermen, suggesting that fish and hence zooplankton are in great abundance there. 1.3.1 Oceanic circulation The west coast of British Columbia is in the North Pacific Gyre. The eastward-flowing Subarctic Current bifurcates at the latitude of Vancouver Island to form the northward Alaska Current and the southward California Current. In the winter, the Alaska Current is intensified and, in response to the prevailing southeasterly winds, the flow along Vancouver Island is northward parallel to the shore. In the summer, the California Current is shifted shoreward and the flow is southward in response to the Chapter 1. Introduction 11 i Figure 1.2: Geography and depth contours (m) for the south coast of British Columbia and the north coast of Washington (after Foreman and Thomson, 1997). Chapter 1. Introduction 12 wind forcing. Along Vancouver Island, the equatorward flow is near the shelf break and it is common to call it the Shelf-Break Current (Freeland et al, 1984; Thomson et al, 1989; see Figure 1.3). In the coastal region, the Davidson Current (flowing northward) appears in the winter while the subsurface, poleward-flowing California Undercurrent develops principally in the summer extending from 200 to more than 450 m deep (Freeland and Denman, 1982). The outflow from Juan de Fuca Strait also generates a current, the Vancouver Island Coastal Current, flowing northward even against the prevailing winds. This current is found by Freeland et al. (1984) and Thomson et al. (1989) to be buoyancy-driven by the freshwater outflow. Thomson et al. (1989) also show that this current can act as a barrier for cross-shore advection of animals. 1.3.2 Hydrography On the west coast of Vancouver Island, surface temperatures are usually around 10°C, salinities around 31 and density around 23 sigma-t (ot)A but the California Undercurrent waters have different properties: 6-7°C, 33.75-34, 26.5 ot (Freeland and Denman, 1982). The southern end of the shelf off Vancouver Island is a region of high upwelling during the summer. The winds are indeed upwelling-favorable since they blow to the southeast. The Ekman transport is therefore offshore in the upper Ekman layer, allowing water to be brought from depth to the surface. This upwelled water is generally cold, salty and very rich in nutrients, which is important for the growth of phytoplankton (Mackas and Sefton, 1982). Freeland and Denman (1982) point out that the upwelled water brought to the continental shelf in the Juan de Fuca region has the properties of the California Undercurrent. They show that only interaction with nearby Juan de Fuca Canyon can 4 at is density p-1000+0.025 (Gill, 1982). Chapter 1. Introduction 13 Figure 1.3: Circulation off the British Columbia/Washington coast in A) winter and B) summer. The numbers give the speeds in cra.s - 1. The Vancouver Island Coastal Current (VICC), the Shelf-Break Current (SBC) and the California Undercurrent (CU) are added (after Thomson, 1981). Chapter 1. Introduction 14 bring water from 450 m (depth of the California Undercurrent) onto the shelf, since wind-induced upwelling is too weak to account for it (see also Cannon, 1972). The interaction between canyons and upwelling comes from the fact that in canyons, the cross-shelf pressure gradient is no longer balanced by the Coriolis force (Freeland and Denman, 1982; Kinsella et al, 1987; Klinck, 1988). This pressure gradient brings denser water up to and over the rim of the canyon, water which tends to come up on the southern (downstream) side of the canyon. On the northern (upstream) side, the water flowing over the canyon tends to fall into it (Freeland and Denman, 1982; see also Allen, 1996). Upwelling is indeed an important feature in the vicinity of canyons because of their particular physical behaviour. Since surface and deep waters have different properties, they will be used in this study to "trace" upwelling. 1.4 Objectives In the past few years, many studies have been conducted by the oceanographic com-munity to better understand the physical and biological processes (and their interactions) at work in the coastal regions. This has been urged since coastal waters are intensively used for commercial fishing, transportation, mineral extraction, and recreation. In this light, the GLOBEC (Global Ocean Ecosystem Dynamics) program was created to im-prove the cooperation of scientists and "to advance our understanding of the structure and functioning of the global ocean ecosystem, its major subsystems, and its response to physical forcing so that a capability can be developed to forecast the response of the marine ecosystem to global change" (Roger Harris, Global Change NewsLetter, 1996). There are three major gaps in our knowledge of the marine ecosystem: 1) the dynam-ics of zooplankton population relative to phytoplankton and to their major predators, 2) the influence of physical forcing on these dynamics and 3) the estimation of biological Chapter 1. Introduction 15 and physical parameters associated with the dynamics of zooplankton relative to phyto-plankton. In order to improve our understanding of these issues, the Pacific component of the Canadian-GLOBEC program decided to emphasize six elements: the modelling and observation of the Central Northeast Pacific; the modelling of its continental margin; and time-series observations, process studies and modelling of the continental shelf. De-termining the flow patterns around submarine canyons off the west coast of Vancouver Island and their significance in inhibiting offshore and alongshore zooplankton advec-tion is, as part of the process study of the continental shelf, one of our contributions to GLOBEC. More precisely, several factors influencing the aggregation of zooplankton near sub-marine canyons are studied, such as currents that can trap zooplankton and enhanced upwelling which brings up nutrient-rich water. Barkley Canyon was chosen as the re-gion of an extensive survey over a two-year period. The important issue was to find out whether submarine canyons impact on the local circulation and are regions of enhanced upwelling and aggregation of zooplankton. The circulation was assessed through moor-ings Aanderaa RCM4's (Recording Current Meters) along the north and south walls, and the water properties through Conductivity-Temperature-Depth (CTD) surveys. Bongo tows were performed in the canyon to assess the zooplankton aggregation patterns. In this thesis, Chapter 2 studies Juan de Fuca Canyon (situated at the mouth of Juan de Fuca Strait) which displays a very intense upwelling and particular circulation pattern. The data from the CTD casts in Barkley Canyon are interpreted in Chapter 3 and used to investigate upwelling. These results are also compared with similar observations from Astoria Canyon (at the mouth of the Columbia River on the Oregon/Washington shelf). In Chapter 4, we analyse the mooring data (currents, temperature, salinity) and draw a picture of the circulation pattern. The zooplankton data are presented in Chapter 5 and the regions of zooplankton abundance are interpreted in relation to the current patterns and upwelling. Finally, conclusions and further remarks are presented in Chapter 6. Chapter 2 Juan de Fuca Canyon This chapter presents results on Juan de Fuca Canyon based principally on mooring data acquired from May to October 1993, and to a lesser extent on CTD data. It is intended as an introduction to the importance of submarine canyons on the dynamics. Although Juan de Fuca Canyon is peculiar due to its shape and location, many of its properties are identical to those of other submarine canyons. The patterns of temperature, salinity and circulation found in the submarine canyon are studied. This chapter first presents the geographical setting of Juan de Fuca Canyon, then the wind and current systems around and within it, and the water properties associated with the enhanced upwelling that the canyon induces. Finally, we discuss the importance of the California Undercurrent on the upwelling in Juan de Fuca Canyon. 2.1 Study area Juan de Fuca Canyon is situated south of Vancouver Island (British Columbia). It is a narrow, deep canyon thought to intersect a glacial trough (Cannon, 1972). It is 135 km long and 8 km wide, cutting into a 90-km shelf and reaching depths of about 300 m (Allen, 1996). Its most important peculiarity is that it extends across the shelf into Juan de Fuca Strait (see Figure 2.1). This strait is very important for the local circulation since it provides the southern shelf with a large freshwater outflow from the Strait of Georgia. The canyon has several sharp bends along its path and from one of these a second canyon extends northward, shallowing until it disappears on the continental shelf 16 Chapter 2. Juan de Fuca Canyon 17 (Freeland and Denman, 1982). This canyon, a spur off the side of Juan de Fuca Canyon, has been nicknamed "Spur" Canyon but now bears the official name Tully Canyon. 1 2 6 ° 2 0 ' W 1 2 6 ° 0 0 ' W 1 2 5 ° 4 0 ' W 125'20'W 1 2 5 ° 0 0 ' W 124 o 40"W 1 2 4 ° 2 0 ' W Figure 2.1: Geographical area around Juan de Fuca Canyon. The 200-m isobath is shown as well as the position of moorings JF03 and A l , and of Destruction Island. The Juan de Fuca region is very important to regional biological productivity. The system formed by the canyon and the local circulation induces an enhanced upwelling in the summer months (see Section 2.3). The high nutrient content of the upwelled water is of great importance for the productivity of nearby La Perouse Bank (Mackas and Sefton, 1982). High dissolved nutrients, primary productivity and phytoplankton biomass characterize the region. Mackas and Sefton (1982) report that an inshore maximum of Chapter 2. Juan de Fuca Canyon 18 zooplankton concentration occasionally occurs at the mouth of Juan de Fuca Strait and off Barkley Sound (north of La Perouse Bank, see Figure 1.2). The data on which most of this chapter is based were gathered in Tully Canyon at mooring JF03 near the head of the canyon (48° 15' N, 125°12' W) during the La Perouse Project conducted by Richard E. Thomson (Institute of Ocean Sciences (IOS), Sidney, BC). The instruments were Aanderaa RCM4's which recorded the salinity, temperature, conductivity ratio, pressure, direction and speed of the current and were moored at 30, 75 and 200 m deep in 220-m depth. The data were recorded between May 9th and October 21st, 1993. Another mooring, JF01 (48°03' N, 125° 19' W), recorded the same physical properties but only until May 15th, 1993 when it was hit by fishermen. The data were too scarce to be analysed. The following study uses the data from JF03, as well as wind data from the U.S. National Data Buoy Office1. We also examine CTD data from the La Perouse Project and current meter data from mooring A l (48°32' N, 126°12' W) (courtesy of Richard E. Thomson, IOS). 2.2 Winds and currents Among others, Freeland et al. (1984) and Simard and Mackas (1989) describe the detailed circulation on the southern continental shelf off British Columbia. In the winter, the flow is to the northwest parallel to the coast. Greater speeds (30-40 cm.s~l) in the first 20 km from the shore characterize of the Vancouver Island Coastal Current. In the summer, the flow remains to the northwest over the inner shelf with speeds of 10-15 cm.s~l but a southeast current develops near the shelf break in response to the seasonal shift to northerly winds. The flow in the core of the underlying California Undercurrent (200-250 m) is generally to the northwest at 5 to 10 cm.s"1 (Freeland and Denman, 1982). 1 Internet site http://seaboard.ndbc.noaa.gov Chapter 2. Juan de Fuca Canyon 19 From the U.S. National Data Buoy Office, hourly values of the direction and speed of the winds near Destruction Island (47°41' N, 124°29' W; see Figure 2.1) were obtained from May 9th to October 21st, 1993. Daily-averages of the horizontal wind velocities are calculated and plotted against time in Figure 2.2. Wind velocity at Destruction Island _ _ i i i , i i Jun Jul Aug Sep Oct Figure 2.2: Daily-averages of the wind velocity at Destruction Island, from May 9th until October 21st, 1993. The largest velocity is 8.69 m.s~l. Chapter 2. Juan de Fuca Canyon 20 The data mainly reflect the summer season with an insight into the Fall Transition (when winds and currents reverse from their summer pattern to their winter one). There is much variability in the wind signal, at a relatively high frequency. Yet, winds are found to be mainly from the northwest following the coast. In late May and early June, the winds vary around a northwesterly direction but only around day 160 (October 8th) is a shift from northwesterly to southeasterly noticed: this marks the beginning of the Fall Transition. Thomson and Ware (1996) pinpoint the Spring and Fall Transitions with the use of a "velocity index" /„, different from the Bakun index. Their index characterizes the low-frequency baroclinic variability and is linked to instability processes that effect the transfer of potential energy stored in the mean cross-slope density field to the kinetic energy of mesoscale meanders and eddies. The transition seasons are characterized by Iv fa 0, upwelling by Iv 0, and downwelling by /„ <C 0. They studied velocity fields just north of Juan de Fuca Canyon and found October 18th to be the beginning of the Fall Transition in 1993, which is close to the date suggested by Figure 2.2. Daily-averages of the currents from JF03 at 30, 75 and 200 m are computed from the hourly data and presented in Figure 2.3. Even at 30 m, the current from JF03 is only weakly correlated to the winds. Cannon (1972) shows that in the Juan de Fuca region the current reversals are linked to the wind reversals, except occasionally when the currents remain against the winds due to the presence of a strong density gradient isolating the currents. At 30 m, the current is mainly cross-canyon, southeastward after May and slightly northeastward after September. Its intensity decreases after August. From Figures 2.2 and 2.3, the directions of the wind and the current at 30 m seem decoupled. Figure 2.4 presents the time-series of the alongshore 30-m current at JF03 and alongshore wind at Destruction Island, which appear to be out of phase. This suggests that there is not a very good correlation between winds and currents at 30 m. After low-pass filtering each signal and removing the mean, the correlation between the Chapter 2. Juan de Fuca Canyon 21 Currents at mooring JF03 Jun Jul Aug Sep Oct F i g u r e 2.3: Dai ly-averages of the currents f rom J F 0 3 at 30, 75 a n d 200 m in the upper , m i d d l e a n d lower panels (respectively) f rom M a y 9th unt i l O c t o b e r 21st, 1993. T h e largest velocity is 0.338 m . s - 1 . Chapter 2. Juan de Fuca Canyon 22 alongshore components of the 30-m current and the wind at Destruction Island is less than 0.5 (r=0.3815). The correlation between the cross-shore components is even lower (r=0.2401). Alongshore current from JF03 at 30 m (filtered) 0.21 1 1 1 1 -i 1 TJ CD Jun Jul Aug Sep Oct Alongshore wind velocity at Destruction Island (filtered) Jun Jul Aug Sep Oct Figure 2.4: Time-series of the alongshore current (m.s l) from JF03 at 30 m (upper panel) and alongshore wind (m.s -1) at Destruction Island (lower panel) from May 9th until October 21st, 1993. Chapter 2. Juan de Fuca Canyon 23 In order to understand the reason for this decoupling, the location of the 30-m moor-ing is investigated. CTD data (especially salinity) from different stations near JF03 on the shelf are used for May, June and August 1993. Figure 2.5 shows the vertical pro-files of salinity. Station LB10 (48°18' N, 125°41' W) was sampled on May 7th, station LA06 (48° 12' N, 125° 17' W) was sampled on May 8th and June 29th, and station LB14 (48°08' N, 125°60' W) was sampled on August 20th, 1993. LB10 (May 7th) LA06 (May 8th) D ' ' ' ' 1 ' ' ' 1 30 31 32 33 34 31 32 33 34 35 Salinity Salinity Figure 2.5: Vertical profiles of salinity for three stations on the shelf in May, June and August 1993. Chapter 2. Juan de Fuca Canyon 24 Figure 2.5 shows that the 30-m mooring is in the upper layer at LB10 in May, below the halocline at LA06 in May and June, and on the halocline at LB14 in August. It is therefore isolated from the surface for some periods of time and unable to represent the surface current. It is expected that the surface current would be at a small angle with the wind direction, but data to support this statement are lacking. As expected, there is no correlation with the winds at greater depths. At 75 m, the current is cross-canyon, mainly south to southeastward. It also gradually weakens after August and turns northeastward. At 200 m, the current is northward (in-canyon) sup-porting the expected summer upwelling due to northwesterly winds. It is straightforward to think that the 30- and 75-m cross-canyon flows induce the 200-m in-canyon flow. In effect, the pressure gradient in the canyon is no longer balanced by the Coriolis force, compelling the flow to move down the pressure gradient, i.e. inshore (Freeland and Den-man, 1982). A correlation study on the fluctuations of the low-pass filtered currents was performed in order to test this assertion. The cross-shore components of the shallower currents are not coupled with the alongshore component of the deeper current (r=0.3005 for the correlation at 30-200 m, and r=0.3652 for the correlation at 75-200 m). However, the alongshore (cross-canyon) components of the 30- and 75-m currents are relatively well correlated with the cross-shore (up-canyon) component of the 200-m current, as seen in Figure 2.6 for the 75-200 m components. The correlation coefficient is r=0.6292 for 75-200 m (it is r=0.4615 for 30-200 m). This suggests that the up-canyon 200-m current is strongly related to the cross-canyon 75-m current. Figures 2.10, 2.11 and 2.12 in Section 2.3 also illustrate these cross- and in-canyon flows. In summary, the winds are mainly from the northwest (upwelling-favorable) in the summer. The 30-m current is below the halocline, thus isolated from the near-surface flow, and does not correlate with the wind direction. Both the 30- and 75-m currents mainly flow across-canyon. In response, the 200-m current is up-canyon. Chapter 2. Juan de Fuca Canyon 25 Figure 2.6: Time-series of the cross-canyon current (m.s *) at 75 m (upper panel) and up-canyon current at 200 m (lower panel) from JF03, from May 9th until October 21st, 1993. Chapter 2. Juan de Fuca Canyon 26 2.3 Water properties and upwelling From May until October, the temperature at 30 m is around 10°C and the salinity around 31. At 75 m, the temperature decreases to around 8-9°C while the salinity increases to 32, reflecting the increase of density with depth. At 200 m, the temperature is around 6°C and the salinity around 33-34. Yet, as can be seen in Figures 2.7 and 2.8, these quantities fluctuate during the summer season. Temperature at mooring JF03 30 m 14 i 1 1 1 r -0) JF03 200 m 7.51 : 1 1 1 Jun Jul Aug Sep Oct Figure 2.7: Daily-averages of the temperature (°C) from JF03 at 30, 75 and 200 m in the upper, middle and lower panels (respectively) from May 9th until October 21st, 1993. Chapter 2. Juan de Fuca Canyon 27 Figure 2.8: Daily-averages of the salinity from JF03 at 30, 75 and 200 m in the upper, middle and lower panels (respectively) from May 9th until October 21st, 1993. Chapter 2. Juan de Fuca Canyon 28 More precisely, both the temperature and salinity at 30 m decrease after July. At 75 m, the signal of upwelling can be found after July as temperature decreases while salinity increases. At 200 m, the same trend can be noticed for the temperature and salinity, with the salinity starting to decrease after August. As Hickey et al. (1986) state, these patterns demonstrate that upwelling is present, i.e. lower temperature is associated with higher salinity. If the changes were to be caused by alongshore advection, low temperature would be associated with lower rather than higher salinity, due to river run-offs near the coast. To support this idea, the values are more closely examined. After June 15th, the temperature at 200 m is ~ 6.5°C and the salinity ~ 33.9 (note: this is close to the California Undercurrent values, see Section 1.3.2). After July 1st, the temperature at 75 m is ~ 7.5°C and the salinity ~ 34. After August 1st, the temperature at 30 m is ~ 9°C and the salinity ~ 31-32. Using CTD data from station LB 14 (48° 08' N , 125°60' W), the 200-m canyon water is found to have characteristics of the shelf-break water at 250-350 m deep (on May 7th) or at 240-380 m deep (on August 20th). Thus, the shelf-break water is upwelled from these depths to 200 m through the canyon. Likewise, the 200-m water in the canyon is upwelled to 75-m depth between June and July and the water at 75 m is itself upwelled to 30 m between July and August. It may be interesting to recall Hickey's (1986) remark that a change in water temperature of ~ 0.5-l°C represents a vertical movement of ~ 100-200 m. To see these results more clearly, the CTD data on May 7th are used to link the value of the salinity at one date to the depth the water had been upwelled from. For each day, the value of the salinity in the time-series is compared to the values in the May CTD file and the pressure associated with that salinity is plotted. Figure 2.9 shows the time-series of the depth from which the water at 200 m originated. This figure shows the existence and the strength of the upwelling. For instance in the middle of July, water at 200 m is upwelled from as deep as 440 m. Although its strength Chapter 2. Juan de Fuca Canyon 29 Pressure associated with the salinity from JF03 at 200 m 1501 1 1 1 1 4 5 01 1 1 1 1 1 Jun Jul Aug Sep Oct Figure 2.9: Daily-averages of the pressure (dbar) illustrating the depth from which water at 200 m was upwelled, from May 9th until October 21st, 1993. Chapter 2. Juan de Fuca Canyon 30 fluctuates, upwelling appears to be especially active in June and July, relaxing at the end of July. Freeland and Denman (1982) as well as Cannon (1972) show that Juan de Fuca Canyon can enhance upwelling in the region, raising water from great depths (450 m, bottom depth of the California Undercurrent). To show that upwelling through the canyon is stronger than that occurring on the shelf away from the canyon, the temperature and salinity at two stations on the shelf at different times are compared at 30, 75 and ~ 150 m. The results are listed in Table 2.1. Table 2.1: Values of temperature and salinity at two different stations outside Juan de Fuca Canyon. Station LA06 May 8th June 29th Depth Temperature Salinity Depth Temperature Salinity 30 m 9°C 31.6 30 m 11.9°C 31.8 75 m 9.3°C 32.7 75 m 8.6°C 32.9 100 m 8.5°C 33.2 120 m 7.5°C 33.6 Station L B 10 May 7th August 18th Depth Temperature Salinity Depth Temperature Salinity 30 m 10.1°C 32.3 30 m 9.2°C 32.6 75 m 9.1°C 32.7 75 m 7.5°C 33.4 150 m 6.8°C 33.9 140 m 6.8°C 33.9 When comparing the values in Table 2.1 with the ones in Figures 2.7 and 2.8 on the same dates, the values appear more extreme (lower temperature and higher salinity) inside the canyon than outside. This supports the idea that the upwelling is stronger and more developed inside the canyon than directly outside at the slope. Chapter 2. Juan de Fuca Canyon 31 Finally, monthly-averages of the currents at 30, 75 and 200 m are plotted to see the upper cross-canyon and lower in-canyon flows. In-canyon flow occurs at 200 m from May to September, confirming the existence of upwelling. As an example, Figures 2.10, 2.11 and 2.12 show that upwelling occurs in the deep canyon in May, August and September, respectively. The strength of the up-canyon flow is similar at 200 m in May and in August. In summary, the temperature and salinity signals record the upwelling, which is strong and persistent from May through October. Water is upwelled from as deep as 440 m to 200 m, and the signal propagates towards the surface. Upwelling is stronger inside the canyon than outside. 2.4 The northward-flowing California Undercurrent Freeland and Mcintosh (1989) state that the deep southward flow at the shelf edge correlates with a northward (along-canyon) flow within Tully Canyon with a three-day lag, i.e. the southward flow crossing the canyon drives an up-canyon flow. Upwelling in the Juan de Fuca region occurs in the summer due to northwesterly winds (with the coast to the left), as predicted by the Ekman theory. It may then be interesting to look at the possible effects the California Undercurrent might have on upwelling since it flows northward, but at depth. The idea is that upwelling is not affected by the direction of the California Undercurrent but only by the Shelf-Break Current, which determines the pressure gradient at the top of the canyon. CTD data gathered during La Perouse Project by Richard E. Thomson (IOS) from mooring A l (48°32' N, 126°12' W; see Figure 2.1) at 400 m are studied. The current meter was in the California Undercurrent waters and spanned from May 6th to Octo-ber 21st, 1993. An Aanderaa RCM4 recorded the salinity, temperature, conductivity Chapter 2. Juan de Fuca Canyon 32 1 2 5 ° 4 0 ' W 1 2 5 ° 2 0 ' W 125° OO'W 1 2 4 ° 4 0 ' W Figure 2.10: Monthly-averaged currents from JF03 at 30, 75 and 200 m for May 1993 (the 200-m isobath is shown). Chapter 2. Juan de Fuca Canyon 33 1 2 5 ° 4 0 ' W 1 2 5 ° 2 0 ' W 125° 00'W 1 2 4 ° 4 0 ' W Figure 2.11: Monthly-averaged currents from JF03 at 30, 75 and 200 m for August 1993 (the 200-m isobath is shown). Chapter 2. Juan de Fuca Canyon 34 1 2 5 ° 4 0 ' W 1 2 5 ° 2 0 ' W 1 2 5 ° 0 0 ' W 1 2 4 ° 4 0 ' W F i g u r e 2.12: Month ly -averaged currents f rom J F 0 3 at 30, 75 a n d 200 m for September 1993 (the 200-m isobath is shown). Chapter 2. Juan de Fuca Canyon 35 ratio, pressure, east-west and north-south current speeds. The hourly values were daily-averaged and the time-series of the 400-m current plotted (Figure 2.13). Some data are lacking due to the failure of the instrument. Basically, the current at 400 m is southward until the beginning of July, when it turns to be mainly northward, with some variability around August. This signals the onset of the California Undercurrent in the summer. The current at depth is in the same direction as the surface flow in the spring, and opposes it in the summer when the undercurrent forms. As stated before, upwelling occurs as soon as May and persists (even strengthens) during the summer. Yet, looking back at Figure 2.9, upwelling appears to start relaxing around the end of July, i.e. some days after the onset of the undercurrent. This could indicate that the undercurrent has an effect on the upwelling. When comparing the time-series of the undercurrent from A l to the temperature from JF03 at 200 m (not shown), there does not seem to be a correlation: temperature continues to decrease even after the undercurrent turns to the north. On the contrary, Figure 2.14 shows that there is an anti-correlation between the low-pass filtered signal of the cross-canyon flow at 75 m (blue) and the temperature signal at 200 m (red, the axis has been reversed to show the anti-correlation). When the current starts decreasing in intensity, the temperature starts increasing and thus upwelling starts relaxing. The tem-perature actually lags the current by some days (except perhaps in May). The correlation between the fluctuations of the temperature and of the low-pass filtered current is rela-tively high, especially if the time-series is separated into two parts at day 80 (r=0.7536 before day 80 and r=0.6142 after day 80). Day 80 (July 28th) appears to be close to the day when the current at 75 m begins to decrease. These correlations show that the cross-canyon current drives the pressure gradient, which in turn drives the upwelling, hence the temperature signal. Chapter 2. Juan de Fuca Canyon 36 Figure 2.13: Daily-averages of the current from A l at 400 m, from May 6th until October 21st, 1993. The largest velocity is 0.0983 ra.s-1. Chapter 2. Juan de Fuca Canyon 37 15 c 0.05[ o >. c 03 1 0.04 H Cross-canyon current at 75 m (filtered) and temperature at 200 m from JF03 Jun Jul Aug Sep Oct F i g u r e 2.14: Time-ser ies of the cross-canyon current (m.s x ) at 75 m (blue) a n d t e m -perature (°C) at 200 m (red, reversed) from J F 0 3 , f rom M a y 9th unt i l O c t o b e r 21st, 1993. Chapter 2. Juan de Fuca Canyon 38 The formation of a northward-flowing undercurrent does not appear to dramatically affect upwelling, which appears to be only caused by a southward current, namely the Shelf-Break Current. If the California Undercurrent were to perturb upwelling, its onset in July would have been accompanied by weaker or cessation of upwelling. In fact, it is the relaxation of the southward current that seems to drive the relaxation of upwelling. 2.5 Summary From the data gathered in Juan de Fuca, it can be seen that the winds are mainly from the northwest in the summer. The 30- and 75-m currents are across-canyon while the 200-m current is up-canyon. The shallower flows are shown to induce the deep up-canyon flow. As recorded in the temperature and salinity, there is an enhanced upwelling within the canyon (stronger than on the shelf), with water being upwelled over as much as 200 m. The temperature signal at 200 m deep is shown to be correlated with the 75-m across-canyon velocity. The California Undercurrent does not appear to affect the upwelling which is mainly linked to the direction of the Shelf-Break Current. Chapter 3 Physical Properties of Barkley Canyon: the C T D data This chapter describes Barkley Canyon and its area as well as data gathered in 1997 (CTD, moorings and winds). While the next chapter deals with the time-series provided by the mooring data, this chapter focuses on the results from the CTD data to give a picture of the water properties in Barkley Canyon. In particular, it presents the varia-tions of the properties with depth and space and illustrates the upwelling enhanced by the canyon. A comparison of this canyon and its physics with Astoria Canyon on the Oregon/Washington shelf is also presented. Finally, the influence of the internal tide on the data is investigated. 3.1 The geographical setting of Barkley Canyon Barkley Canyon is a 15-km-wide, 26-km-long canyon situated off Barkley Sound, on the west coast of Vancouver Island (see Figure 3.1). It is located 60 km from the coast and cuts into the continental shelf reaching depths of a thousand meters. The water inshore from the canyon is 100 to 150 m deep. Other canyons (e.g. Nitinat and Loudoun) surround Barkley Canyon and help make this area a highly productive fishery zone. The summer northerly winds on the west coast of Vancouver Island are upwelling-favorable and cause coastal upwelling (Freeland et al, 1984). The interaction of the canyon with the southward-flowing Shelf-Break Current induces an enhanced upwelling within the canyon and in its vicinity. 39 Chapter 3. Physical Properties of Barkley Canyon: the CTD data 40 1 2 6 ° 2 0 ' W 1 2 6 ° 0 0 ' W 125° 40'W 1 2 5 ° 2 0 ' W 1 2 5 ° 0 0 ' W Figure 3.1: Geographical area around Barkley Canyon. The locations of the wind buoy, mooring A l and lines B and C of La Perouse Grid are indicated. The 150-, 300- and 800-m isobaths are shown. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 41 The data used in this thesis are CTD and zooplankton data acquired during a cruise on July 25th, 26th and 27th, 1997 as well as mooring and wind data collected from April to October 1997. The sixteen stations are plotted in Figure 3.2 and the location of each station is listed in Table 3.1. CTD casts were performed at each of the stations with two extra records for BCB4 and BCC4, and three extra records for BCD4 (for time-series considerations). In addition, zooplankton bongos were towed at five of these locations, marked by green stars in Figure 3.2. Oxygen samples were taken at BCB3, BCC2, BCC3 and BCD3. A tow-yo CTD cast was also performed on line 4 (back and forth between BCB4 and BCD4): the CTD was towed by the boat at a speed of approximately two knots, while alternatively lowered and lifted. Moorings (shown as red dots in Figure 3.2) were also positioned in the canyon. One special mooring (BC04) was an upward-looking, bottom-mounted (230 m deep), backscat-ter acoustic instrument at 150 kHz to measure zooplankton but it collected only six days worth of data. The other three moorings had current meters at 150, 250 and 350 m deep and BC03 had a thermistor chain immediately below the 150-m instrument. The locations of these moorings are listed in Table 3.2. Finally, the wind data were recorded at a permanent AES(Atmospheric Environment Service)-maintained buoy near Barkley Sound (buoy 46206: 48°50' N, 126° W; see Figure 3.1) from April 1st to October 31st, 1997. These data are used to link the near-surface circulation to the wind direction and speed. The wind and mooring data are discussed in Chapter 4. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 42 126° 20'W 126° OO'W 125° 40'W Figure 3.2: Stations sampled around Barkley Canyon (black crosses and green stars), zooplankton tow stations (green stars) and mooring locations (red dots) (the 200- and 300-m isobaths are shown). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 43 Table 3.1: Location of the stations sampled around Barkley Canyon. Station Latitude Longitude Depth Zooplankton tow BCA1 48°30.1' N 125°59.9' W 140 m BCA3 48°27.9' N 126°04.8' W 200 m -BCA4 48°25.2' N 126°10.1' W 400 m -BCB2 48°26.0' N 125°55.8' W 170 m yes BCB3 48°24.9' N 125°59.4' W 190 m -BCB4 48°22.8' N 126°04.3' W 350 m yes BCC1 48°25.8' N 125°50.6' W 140 m _ BCC2 48°24.5' N 125°53.5' W 175 m -BCC3 48°22.6' N 125°56.1' W 600 m yes BCC4 48°20.6' N 125°59.2' W 700 m -BCD2 48°22.3' N 125°50.1' W 150 m yes BCD3 48°20.2' N 125°53.1' W 400 m -BCD4 48°18.7' N 125°54.9' W 300 m yes BCE1 48°22.3' N 125°43.0' W 140 m _ BCE2 48°19.8' N 125M6.8' W 160 m -BCE4 48°16.8' N 125°51.0' W 400 m -Chapter 3. Physical Properties of Barkley Canyon: the CTD data 44 Table 3.2: Location of the moorings in Barkley Canyon. Mooring Latitude Longitude Depth BC01 48°23.9' N 125°54.7' W 400 m BC02 48°22.3' N 125°58.5' W 550 m BC03 48°21.1' N 125°56.1' W 550 m BC04 48°24.9' N 125°56.8' W 250 m 3.2 Processing of the data The CTD data were delivered in a raw format which gave the pressure, temperature (primary and secondary) and salinity (primary and secondary). The secondary channels were only used when the primary channels had failed. Density was computed using temperature and salinity with the formula from Gill (1982, appendix 3). The data are processed using a combination of tools in Matlab and Fortran. First, they are filtered with a forward and reverse digital filter following the equation: y(n) = b(l) x x(n) + 6(2) x x(n - 1) + ... + b(nb + 1) x x(n - nb) — a(2) x y[n — 1) — ... — a(na + 1) x y(n — na) where a and b are the filter vectors. After filtering in the forward direction, the filtered sequence is reversed and run back through the filter. The cut-off frequency of this filter is 0.0104 seconds (the frequency of sampling was 0.0416667 seconds). This removes most of the high-frequency wiggles. The data are then re-ordered to be monotonic (to suppress the small ups and downs of the pressure) and finally, they are despiked using a "cleaning" Chapter 3. Physical Properties of Barkley Canyon: the CTD data 45 function: outliers that are too far from the current median value are replaced with this value. The second step is to bin-average the data with the pressure as the bin channel. Two types of files are created: one with the real bin-averaged pressure and one with the pressure equally spaced to be used in certain programs that require a regular pressure pattern. This is done for the data acquired in Barkley Canyon as well as for lines B and C of La Perouse Grid (see Figure 3.1). Line B extends from (48°40' N, 124°59' W) to (48° N, 126°17' W). Line C extends from (48°50' N, 125°28' W) to (48°18' N, 126°27' W). The tow-yo data are a particular set of CTD data. Basically, they are a succession of up and down CTD casts. The files are segmented into up- and down-casts then filtered and bin-averaged using the same programs as before, after being adapted to handle up-and down-casts (CTD's are usually down-casts). The wind data from the buoy reported hourly values of the wind direction and vector-averaged wind speed. They were daily-averaged to give a record of the wind velocities for the period April-October 1997. The mooring data were semi-hourly values of current direction and speed, and were processed at the Institute of Ocean Sciences (Sidney, BC; courtesy of Richard E. Thomson). 3.3 Upwelling within Barkley Canyon The CTD data give synoptic information on the strength and extent of upwelling in Barkley Canyon. There is a doming of the isopleths over the canyon axis and upwelling towards the head. Yet, the upwelling seems to extend much closer to the surface than what theory predicts. This also differs from other canyons, like Astoria (see Section 3.4). Before presenting cross-section plots of the water properties, their vertical and spatial distribution is analysed. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 46 3.3.1 Variations of the water properties All sixteen stations show similar vertical profiles of temperature, salinity and density. The temperature decreases with depth and ranges from 15-16°C at the surface to 6°C around 400 m (at BCE4 for instance) and 4°C around 800 m (at BCC4). The salinity increases with depth and ranges from 30.5 at the surface to 34 between 140 and 400 m (it appears to be almost constant in that depth range). From 600 to 800 m, it increases to above 34. The density, whose profile is similar to that for salinity, ranges from 22.5 ot at the surface to around 26.5, increasing to above 27 at 700-800 m. As an example, the profiles of salinity, temperature and density at BCE4 are shown in Figures 3.3 and 3.4. The mixed layer appears to be situated in the first 10 m throughout the spatial domain. This layer is a zone of well-mixed water where the water properties do not vary much with depth. On the contrary, below the mixed layer lies a zone where the quantities vary rapidly with depth (halocline for the salinity, pycnocline for the density and thermocline for the temperature). The salinity plot shows the presence of two haloclines, a seasonal and a permanent. In winter, strong winds and surface cooling produce a deeper mixed layer which can extend to the permanent halocline. In summer, as the temperature rises and the wind mixing reduces, a seasonal halocline develops above the permanent one. Here, the seasonal halocline is situated around 20-30 m and is steeper than the permanent halocline found around 100 m (see Figure 3.3). Profiles of temperature versus salinity (T/S diagrams) are also plotted for each station. Most of them show a separation between the surface waters (well mixed, high temperature and low salinity) and the deep Pacific waters (low temperature and high salinity). In Figure 3.5, the T/S diagrams for all stations are plotted together. There is a clear separation between the warmer and saltier offshore surface waters (top cluster of mainly dashed lines) and the inshore waters. Most of the stations have the same behaviour Chapter 3. Physical Properties of Barkley Canyon: the CTD data 47 Figure 3.3: Vertical profile of salinity at station BCE4 (July 25th, 1997). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 48 Temperature at station BCE4 (July 25th) Density Figure 3.4: Vertical profiles of temperature (°C, upper panel) and density (ot, lower panel) at station BCE4 (July 25th, 1997). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 49 Temperature/Salinity diagrams _ _ _ BCA3 BCA4 BCB3 BCB4 BCC4 BCD4 BCE2 BCE4 30.5 31 31.5 32 32.5 Salinity 33 33.5 34 34.5 Figure 3.5: Temperature/Salinity diagrams for all stations (refer to the legend for the colour codes). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 50 around a nodal point of salinity 32.5 and temperature 8.5°C, which separates the surface waters from a second water mass. The position of this nodal point is found around 60 m deep for the offshore stations but 50 m deep for the inshore ones, suggesting that the water from the mouth of the canyon may have upwelled towards the head. The offshore waters tend to propagate towards the shore and the solid blue line (for BCA1) shows some intrusion of offshore water (around 10°C and salinity 32). Beyond that nodal point, there is water which tends to be a mix between the water mass mentioned earlier (present around 50-60 m) and another water mass (6.9°C, 33.9) originating from 200 m in the deeper parts of the canyon. 3.3.2 Upwelling in the canyon As discussed in the previous chapters (see Sections 1.3.2 and 2.3), the wind-induced summer upwelling is enhanced in the vicinity of canyons. The temperature inside the canyon reflects this upwelling: it becomes colder towards the head as the water makes its way up from depth offshore. At 100 m, it is 8.25°C at BCC4 (offshore), 8.14°C at BCC3 and 8.04°C at BCC2 (inshore). It is also colder at the head of the canyon (BCC2) than on the shelf (for instance, station LC08 (48°29' N, 126°07' W)), which implies that the upwelling is stronger inside the canyon than on the shelf. The following figures illustrate the upwelling of isopleths in the canyon as well as north and south of it, on the shelf. From the CTD data, contour plots of salinity and temperature are drawn along the major lines of the canyon (lines BCA to BCE and BC1 to BC4). A program similar to the objective mapping concept (see Denman and Freeland, 1985) gives the best results, given the scarcity of the data (only three or four stations per line). Some plots are expanded to more clearly show the pattern of the isopleths associated with upwelling. Although the isotherms display the same pattern as the isohalines, the salinity plots are clearer. On the across-shore lines (BCA to BCE), the isopleths dome in the center of the Chapter 3. Physical Properties of Barkley Canyon: the CTD data 51 canyon, close to the surface at around 20 m deep. The upwelling is stronger towards the head, at depths between 40 and 120 m, as Figure 3.6 shows for line BCB. The upwelling is also stronger within the canyon: the vertical displacement of the isohalines is larger at the lines within the canyon than at line BCA (upstream) or BCE (downstream) (20-40 m vs 10-20 m). Also, the values are more extreme, showing lower temperatures and higher salinities at the same depth for stations within the canyon. Considering the alongshore lines (BC1 to BC4), the isopleths dome at the center of the canyon over the axis in the upper 30 m. On line BC1 close to the head, the 8°C isotherm is upwelling from the northern, upstream side to the southern, downstream side. In the middle of the canyon (line BC3 in Figure 3.7), the isopleths tend to bulge down between BCB3 and BCC3 (where the depth plunges from 200 to 600 m). Klinck's (1996) numerical simulations show some downwelling at the upstream wall and upwelling towards the downstream wall (see Section 1.1.2), as seen here. Though, the fact that the water column is affected by the canyon so close to the surface is in contradiction with Klinck's (1996) and Hickey's (1997) findings that the flow usually directly crosses the canyon in the upper water column. Contour plots of temperature and salinity for lines B and C of La Perouse Grid (see Figure 3.1) are generated to consider the conditions up- and downstream. Upwelling is noticeable around 100 m, deeper than inside the canyon (e.g. at line B in Figure 3.8). This supports the idea that upwelling is stronger inside the canyon and that water is upwelling from deeper and to shallower depths than outside the canyon (also Freeland and Denman, 1982). During the CTD casts, water samples were taken at different depths and titrated to obtain their amount of oxygen. 0 2 values also support upwelling towards the shelf since low values are found at shallower depths towards the shore. Values as low as 2 mL/L are found as shallow as 100 m on the shelf while the same values are found at 300 m over the shelf-break. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 52 Salinity at Line B in Barkley Canyon 0 6 12 Distance (km) Figure 3.6: Cross-section of salinity from the surface to 130 m deep along line BCB. The triangles show the position of the stations: BCB2 (170 m deep) at the left, BCB3 (190 m) and BCB4 (350 m). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 53 Salinity at Line 3 in Barkley Canyon J L 10 Distance (km) Figure 3.7: Cross-section of salinity from the surface to 130 m deep along line BC3. The triangles show the position of the stations: BCA3 (200 m deep) at the left, BCB3 (190 m), BCC3 (600 m) and BCD3 (400 m). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 54 Salinity at Line B (La Perouse Grid) 0 20 40 60 80 Distance (km) Figure 3.8: Cross-section of salinity from the surface to 200 m deep along line B of La Perouse Grid. The triangles show the position of the stations: LB01 (41 m deep) at the left, LB02 (60 m), LB06 (150 m), LB08 (145 m), LB09 (150 m), LB10 (155 m) and LB12 (510 m). The topography is outlined. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 55 The depths of various isohalines are also analysed. In Figure 3.9, the 31.5 isohaline is found between 10 and 14 m. The most important feature is the presence of a dome above the head of the canyon particularly on the downstream side, which signals an enhanced upwelling triggered by the canyon. There is also an upwelling signature along the shelf break. It is important to note that the upwelling signal reaches very shallow depths compared with the extent of the canyon's influence expected from the theory. Upwelling is noticeable as well in the 32.5 isohaline (upper panel of Figure 3.10). This isohaline is between 44 and 54 m and also becomes shallower on the downstream side, at the head of the canyon. Upwelling is visible on the shelf as well. Allen's (1996) results when running non-linear simulations with a three-layer representation of Astoria Canyon are consistent with this picture. She reports that the interface height between the upper two layers (the upper layer being 40 m deep) shows a trough at the upstream side and a ridge at the downstream side, very much like what is seen here. Klinck's (1996) numerical findings on the circulation around canyons (see Section 1.1.2) also partially agree with these data, except for the extent of the canyon's influence to very shallow depths. Finally, in the lower panel of Figure 3.10, the 33 isohaline shows upwelling at the head to a depth as shallow as 70 m. It is deepest (82-84 m) on the upstream side by the mouth, which Allen (1996) also reports for the lower interface height. In summary, the salinity isolines from the CTD data dome over the canyon near the surface and upwell towards the head and the downstream side of the canyon at depth. The upwelling signal is stronger inside the canyon than outside. The depths of the plotted isohalines also support enhanced upwelling over Barkley. The canyon's influence seems to reach shallower depths than what is expected from the theory and than what is seen in Astoria Canyon. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 56 126° 10'W 126° OO'W 125° 50'W 125° 40'W Figure 3.9: Contours of the depth (m) of the 31.5 isohaline around the canyon (the 300-m isobath is shown in bold and the red dots show the CTD stations). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 57 126° 10'W 126° OO'W 125° 50'W 125° 40'W Figure 3.10: Contours of the depth (m) of the 32.5 (upper panel) and 33 (lower panel) isohalines around the canyon (the 300-m isobath is shown in bold and the red dots show the CTD stations). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 58 3.4 Comparison with Astoria Canyon Astoria Canyon is situated 16 km off the Oregon/Washington coast, just offshore of the Columbia River (see Figure 3.11). It is 7 km wide and 20 km long with depths down to 800 m (the depth at the axis is 600 m). It is thus narrower than, and as deep as, Barkley Canyon. The depth below its rim is roughly three times the depth of the incident flow (~ 450 m vs 150 m). Astoria is also very steep, its walls approaching 45° in some locations. As with Barkley, Astoria Canyon's axis is roughly perpendicular to the coast and to the currents (Hickey, 1997). 46° 20'N 46° 10'N 124° 40'W 124° 20'W 124° OO'W Figure 3.11: Astoria Canyon off the Columbia River on the Oregon/Washington Shelf (the 200-m isobath is shown). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 59 The currents are mainly to the south during spring and summer and to the north in fall and winter. In summer, the California Undercurrent forms and flows northward at depth. Astoria Canyon intersects the shelf around the depth of the undercurrent velocity maximum. Typical currents near the shelf break are around 20 cm.s"1 but can increase to 50 and even 100 cm.s"1 in the winter (Hickey, 1997). Similarities and differences exist between Barkley Canyon and Astoria. For example, Hickey (1997) finds vertical excursions of the isopleths due to the semi-diurnal internal tide of about 25 m just above the rim of the canyon (125 m deep). In Barkley Canyon, only a 14-m variation is found at 100 m (see Section 3.5) but this difference with Astoria could be due to the fact that the cruise in July 1997 was during neap tides. Episodic upwelling is also reported in Astoria Canyon but it does not seem to reach the mixed layer. Over Barkley, the isopleths dome over the centre of the canyon, even close to the surface (up to 10 m from the surface, as seen in Figure 3.9). In contrast, the near-surface flow passes directly over Astoria: Hickey (1997) reports that the currents more than 60 m above the canyon are decoupled with the canyon topography. Figure 3.12 shows a section of temperature along the farthest-offshore across-canyon line. The near-surface isopleths appear almost undisturbed but a small dome can be seen close to the surface on the upstream side. This is a faint signal and the main effect of the canyon is clearly seen only below 100 m. The 6.5-7°C isotherm rises around 180 m at the upstream side to 150 m towards the downstream side, while the 6°C isotherm plunges down from 180 m. at the upstream side to 250 m. The 5°C isotherm domes around 500 m. Despite the absence of influence of Astoria on the near-surface flow, the isotherms • appear to show a disturbance close to the surface but fainter than in Barkley. The parameters that were introduced in Section 1.1.3 are different and thus cause different behaviours in the canyons. For example, the Rossby number Ro = ujjL is between 0.3 and 0.8 in Astoria (for flows between 10 and 30 cm.s"1). In wider Barkley Canyon, flows Chapter 3. Physical Properties of Barkley Canyon: the CTD data 60 Figure 3.12: Contoured section of temperature (°C) at the mouth across Astoria Canyon (line 1) on May 21st, 1983 (from Hickey, 1997). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 61 are between 10 and 40 cm.s - 1 during the upwelling event, giving values of Ro between 0.12 and 0.5 (smaller than in Astoria). This causes the flow to "feel" the canyon to shallower depths in Barkley and produces the dome seen in Figure 3.6. The Burger number B = fL/NH (or stratification parameter S = 1/B) also quantifies the amount of stratification and the extent of the canyon effect on the flow. In Astoria, the depth H above the rim is 150 m and the half-width L is 3.5 km. N varies between 5 x 10~3 and 10 x 10 - 3 s - 1 , thus S ranges from around 2 to 4. In Barkley, #=150 m as well, N is between 5 x 10 - 3 and 12 x 10 - 3 s _ 1, but L=7.5 km, giving values of S between 1 and 2.5. A lower value of S in Barkley Canyon means a lower stratification and thus a larger effect of the canyon on the flow. Hickey (1989) also studies Quinault Canyon, which is wider (30 km wide) than Barkley. She states again that the near-surface flow is undisturbed over the canyon but she does not present results on the temperature or salinity. One hypothesis is that they would also show a dome over the canyon close to the surface. In summary, Astoria and Barkley Canyons differ mainly because of their geometries. Although the near-surface flow may remain undisturbed, the isohalines appear to be perturbed to shallower depths and to a larger extent in Barkley but some influence can be seen in the isotherms from Astoria as well. The fact that the stratification parameter is smaller in wider Barkley Canyon could explain this result. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 62 3.5 Influence of the internal tide The internal tide signal (semi-diurnal tide) is typically very important in data sets collected in the ocean near topographic features. If it appears to distort the upwelling signal found in the CTD data, it should be removed. Time-series of vertical profiles are available at BCB4, BCC4 and BCD4 where several CTD's were taken at hourly intervals. The second sample was usually taken after about 20 hours and the subsequent ones at three-hour intervals. Three different CTD casts were taken at BCB4 and BCC4, and four were taken at BCD4. After plotting the different casts, small variations in time can be noticed: the position of the different segments varies with depth. The following vertical profiles of salinity are considered for the different casts performed at BCB4 and BCD4 (top and bottom panels of Figure 3.13). The first cast (black, on July 25th) is taken as a reference. At station BCB4, the seasonal halocline is shallower on July 26th (blue, some 26 hours after the first cast) and is deeper on July 27th (red, some 3 hours after the preceding cast). At depth, the contrary occurs and the blue permanent halocline lies below the black one, while the red lies above it. The vertical displacement of the halocline is 7 m between the extremes at 20-m depth and 14 m at 100-m depth. At station BCD4, the blue seasonal halocline (22 hours after the first cast) is deeper than the black one. The green halocline (3 hours after the preceding cast) is shallower than the blue but deeper than the black. Finally, the red halocline (3 hours after) is the shallowest. At depth, this pattern is inverted, although the green and red permanent haloclines remain respectively deeper and shallower than the black one. The vertical displacement here is around 8 m at 20-m depth and 11m around 100-m depth. There seems to be a periodicity in the see-saw pattern of the depth of the haloclines, which can be attributed to the internal tide. The "succession" of the haloclines (shallow Chapter 3. Physical Properties of Barkley Canyon: the CTD data 63 Salinity at station B C B 4 250 l 300 k 50 I 1 1 1 1 1 1 1 1 30.5 31 31.5 32 32.5 33 33.5 34 34.5 Salinity Salinity at station B C D 4 Salinity Figure 3.13: Vertical profiles of salinity at station BCB4 on July 25th (20:22), July 26th (22:20) and July 27th (01:11) in the top panel and at station BCD4 on July 25th (22:52), July 26th (20:30), July 26th (23:25) and July 27th (02:16) in the bottom panel (refer to the legends for the colour codes). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 64 to deep) on the graph depends on the position of the first cast in relation to the internal tide pattern. Then the other casts are at other phases of the pattern, which produces this see-saw phenomenon. At BCD4, the blue and red curves are in opposition of phase because there are approximately 6 hours between them (the half-period of the internal tide). The amplitude of the internal tide is larger at depth, where the density difference is smaller, than close to the surface. There is almost a two-fold increase in the vertical displacements at the "surface" and at 100 m. Empirical Orthogonal Function (EOF) analysis1 is performed on the density field (from the sixteen stations plus the extra time-series at BCB4, BCC4 and BCD4) to try to isolate the signal of the internal tide, and find the vertical modes that explain most of the variability between the different stations. The first three modes explain 62, 17 and 7% of the spatial variance of the field, respectively. The first EOF is not a first baroclinic mode, as would be expected, but mostly the upwelling signal: it represents the up-and-down movement of both the seasonal and permanent pycnoclines due to upwelling of waters through the canyon. It is possible from the value of the first EOF to derive its contribution to each station (see Appendix A). Figure 3.14 is a contour plot of the values obtained. This figure is almost identical to Figure 3.9 on the depth of the 31.5 isohaline over the canyon and shows the presence of a denser patch over the downstream side of the canyon. This is consistent with an upwelling signal which therefore explains 62% of the vertical variability of the data set. The second EOF (not shown) represents an inshore/offshore spatial variability associated with the tilting of the isopycnals (higher density over the shelf) and explains 17% of the total variance. The internal tide signal is present in neither of the first two EOFs, which explain most of the variance. Therefore, the data are not significantly affected by the internal tide. See Appendix A for more details on E O F analysis. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 65 126° 10'W 126° 00'W 125° 50'W 125° 40'W Figure 3.14: Contours of the contribution of the first EOF to the stations sampled around Barkley Canyon (the 300-m isobath is shown in bold). Chapter 3. Physical Properties of Barkley Canyon: the CTD data 66 Although it is not represented by the EOFs, the internal tide is an important frequency in the data set. A Fast Fourier Transform (FFT) analysis is performed on the semi-hourly values of the alongshore current from mooring BC01 at 157 m to find the most important frequencies. The Fourier Transform projects a temporal signal f(t) into a frequency space F(u) using the following equation: F(u) = f{n)e-j2™n'N N 71=0 Since the FFT is complex, its magnitude is plotted. The frequencies associated with the larger peaks are then derived from the indices. If i is the index associated with the peak, N is the record length and T is the sampling period (here T = 1/2 hr), then the frequency is F = j^. The semi-diurnal frequency is the main one (F = 2.24 x 10 - 5 s - 1 , i.e. T = 12.4 hr), followed by the diurnal frequency (F = 1.17 x 10 _ s s - 1 , i.e. T = 23.8 hr) and the Coriolis parameter (/ = 1.1 x 10 - 4s - 1). There is also an important magnitude associated with a period of three days, as well as with the lunar period (T = 29.7 days). Although the internal tide frequency is dominant in the spectrum, there is more energy2 in the three-day period (associated with upwelling) than in the semi-diurnal frequency. This explains why the upwelling signal contributes to most of the variance of the entire data set, while the internal tide can be considered not to introduce a slant in the analysis. 2 The energy is taken as the area under the spectrum peak divided by the square of the frequency. Chapter 3. Physical Properties of Barkley Canyon: the CTD data 6 7 3.6 Summary Vertical profiles from the CTD's for the temperature, salinity and density in Barkley Canyon show variations with depth. There is also a separation between the warmer, saltier offshore surface waters and the colder, fresher inshore ones. Contours of salinity show a dome over the canyon axis and upwelling towards the head, both along- and across-shore. The upwelling is stronger inside the canyon than outside. Depths of isohalines also show a near-surface dome and a "trough upstream/ridge downstream" pattern deeper. When compared to Astoria Canyon, the influence of Barkley reaches shallower depths mainly because of differences in the canyon geometries. The stratification parameter is smaller in wider Barkley, thus the flow "feels" the canyon to a greater extent and the upwelling signal reaches closer to the surface. Finally, the internal tide signal is seen in the time-series of haloclines and is the most important frequency, but it does not appear to distort the CTD data. Most of the variance is in fact explained by the upwelling signal. Chapter 4 Physical Properties of Barkley Canyon: the mooring data This chapter analyses the data from the moorings deployed within Barkley Canyon from April to October 1997. Time-series of currents are discussed to characterize the circu-lation in the canyon and compare it to the theory and model results. Upwelling is also investigated through the temperature and salinity time-series. Three main upwelling episodes that occurred in June, July and late August-early September are discussed in relation to the temperature and currents. 4.1 Winds around Barkley Canyon The wind data were obtained from a permanent AES (Atmospheric Environment Service) buoy near Barkley Sound (48°50' N, 126° W; see Figure 3.1). Buoy 46206 is situated near the coast and therefore tends to experience bi-directional winds. It recorded hourly data from April 1st to October 31st, 1997. After daily-averaging the north-south and east-west components of the wind, the time-series of the wind velocity is plotted. Figure 4.1 shows that during the Spring Transition (April-May), the wind direction exhibits great variability, changing from mainly southeasterly (winter pattern) to northwesterly (summer pattern). From July 10th to August 20th, the wind becomes less variable and aligns mainly from the northwest, as expected, with increasing intensity. This is the only clear upwelling-favorable period but it seems that the bottom-shelf current, more than the wind, is the main acting mechanism for upwelling through the canyon. Upwelling may therefore occur, even during southeasterlies. After August 20th, 68 Chapter 4. Physical Properties of Barkley Canyon: the mooring data 69 there is an abrupt and clear reversal to southeasterlies which lasts until the beginning of September. The period September-October is characteristic of the Fall Transition, with the wind showing increased variability and turning back to southeasterly. Figure 4.1: Daily-averages of the wind velocity from buoy 46206, from April 1st until October 31st, 1997. The largest velocity is 8.89 m.s - 1. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 70 4.2 Currents in the vicinity of Barkley Canyon 4.2.1 Theory and model predictions Section 1.1.2 presents the theory associated with the circulation within canyons. The near-surface flow is expected to cross the canyon without being disturbed. Around 150 m (just above the rim), the flow bends and crosses the isobaths, while below the rim it follows the direction of the local isobaths and is cyclonic (Allen, 1996; Hickey, 1997). Klinck (1996) also reports downwelling at the upstream wall and upwelling everywhere else in the canyon. The interaction with the bathymetry is very important and although the theory does not predict the formation of eddies, some topographic eddies may be created (Denman and Freeland, 1985) especially at the head of the canyon. Geostrophic currents consistent with Barkley's CTD observations, using bottom friction and assuming zero flow at the bottom at the upstream boundary (450 km upstream of the canyon) were calculated by M . Foreman (Institute of Ocean Sciences, Sidney, BC) using a high-resolution, diagnostic, finite-element method similar to that described in Foreman and Thomson (1997) and Naimie et al. (1994). At 30 m, the calculated horizontal current is following the Shelf-Break Current in a southeastward direction. Some interaction with the bathymetry is visible as the current tends to bend to cross the isobaths and turns cyclonically over the canyon heads. The influence of the canyons is therefore visible close to the surface. The model output for the horizontal velocity at 150 m (see Figure 4.2) shows a fully-developed cyclonic eddy downstream over the head of Barkley Canyon, which suggests particle trapping. The current is slightly in-canyon at the mouth. Figure 4.3 shows the circulation at 350 m: the California Undercurrent flows northwestward at the shelf break, closely following the isobaths, but the flow is southeastward offshore in the vicinity of the canyons, bending to Chapter 4. Physical Properties of Barkley Canyon: the mooring data 71 F i g u r e 4.2: M o d e l output for the horizonta l velocity at 150 m . T h e 50-m, 150-m, 500-m, a n d 1000-m isobaths are inc luded for reference (from M . F o r e m a n , pers. c o m m . ) . Chapter 4. Physical Properties of Barkley Canyon: the mooring data 72 F i g u r e 4.3: M o d e l output for the horizonta l velocity at 350 m . T h e 50-m, 150-m, 500-m, a n d 1000-m isobaths are inc luded for reference (from M . F o r e m a n , pers. c o m m . ) . Chapter 4. Physical Properties of Barkley Canyon: the mooring data 73 enter them with strong in-canyon flows up to the head. Further offshore, the current in the open ocean is northwestward again. As far as the vertical velocities are concerned, the model outputs show downwelling on the upstream side and upwelling at the downstream side at all depths, as expected from the theory. 4.2.2 Currents at mooring locations in Barkley Canyon The moorings, shown in Figure 3.2 with positions listed in Table 3.2, were deployed on April 19th, 1997 for BC02, BC03 and BC04 and on June 4th for BC01. They were recovered on October 2nd. BC01 and BC03 recorded time-series of currents, temperature and salinity at three different depths (approximately 150, 250 and 350 m), BC02 at 250 and 350 m, and BC04 only at 250 m. The semi-hourly data (every half-hour) are averaged to obtain daily values and the currents are low-pass filtered with a three-day period for the plots only. The time-series of temperature and salinity will be discussed in the next section. Figures 4.4, 4.5, 4.6 and 4.7 present the time-series of the currents at BC01, BC02, BC03 and BC04, respectively. They appear to be very variable: • At 150 m, the current at the southern side of the mouth (BC03) is mainly across-canyon (northwestward) in April and May. Starting in June, the currents downstream show a similar pattern at the head (BC01) and at the mouth (BC03): they are both ini-tially across-canyon (eastward) then turn up-canyon (northeastward). At the beginning of July, they are both eastward again but from the middle of July until the middle of August, the flow is northwestward at the head and southeastward at the mouth. Then, both currents turn back to northwestward until the end of the time-series. • At 250 m, the current at the head (BC04) is constantly down-canyon (southwest-ward). At the southern side of the head (BC01), it is mainly across-canyon (south-eastward) from June to August then it turns up-canyon (northward). At the northern Chapter 4. Physical Properties of Barkley Canyon: the mooring data 74 May Currents at mooring BC01 ^\\vsJ®mm^ 157 m -»- v s—•••w«^-w—^•iih. l / t^.\uii^^,..^.\\\i/^.„.,^ 257 m 357 m ^ 357 m magnified Jun Jul Aug Sep Oct Figure 4.4: Daily-averages of the currents from BC01 at 157, 257 and 357 m in the first three panels (respectively) from June 4th until October 1st, 1997. The largest velocity is 0.173 m.s~l. The lower panel is the current at 357 m magnified by a factor of 5. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 75 Currents at mooring BC02 i i i i i i May Jun Jul Aug Sep Oct Figure 4.5: Daily-averages of the currents from BC02 at 250 and 350 m in the upper and lower panels (respectively) from April 19th until October 1st, 1997. The largest velocity is 0.152 m.s - 1. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 76 Currents at mooring BC03 i i i i i i May Jun Jul Aug Sep Oct Figure 4.6: Daily-averages of the currents from BC03 at 142, 242 and 342 m in the upper, middle and lower panels (respectively) from April 19th until October 1st, 1997. The largest velocity is 0.181 m.s - 1. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 77 Current at mooring BC04 May Jun Jul Aug Sep Oct Figure 4.7: Daily-averages of the current from BC04 at 230 m, from April 19th until October 1st, 1997. The largest velocity is 0.089 m.s~l. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 78 side of the mouth (BC02), the current is much stronger and very variable from April to June (between NW and SE during a short period in June) then it turns down-canyon (southwestward) in June-July and finally across-canyon (northwestward) from August until September. At the southern side of the mouth (BC03), the same across-canyon (southeastward) period can be noticed in June, but after that point the current remains relatively up-canyon (northward). In general, the currents are north to northwestward past mid-August, except at the head (BC04). There is an upwelling period at the end of June (which will be presented in Section 4.3) characterized by a short and defined southeastward episode at all moorings, except the head (BC04). • At 350 m, the current is very weak at the head (BC01) and oscillates between up-canyon (northeast) and across-canyon (southeast). The current at the northern side of the mouth (BC02) is relatively consistent with a down-canyon (southwestward) direction while at the southern side of the mouth (BC03), it is more variable but displays a period of up-canyon (northward) flow between June and the beginning of August. The two currents are up-canyon for several instances that coincide with periods of temperature decrease (upwelling). The velocities at 150 m are comparable at the head and at the mouth: they are around 0.05 m.s - 1, with a maximum of 0.181 m.s - 1 at the mouth. At 250 m, they are larger at the mouth than at the head (0.035 m.s - 1 vs. 0.018 m.s - 1). The largest velocity is 0.174 m.s - 1 at the mouth downstream (BC03), which is comparable to the largest velocities at 150 m. At 350 m, the velocities are also larger at the mouth (0.025 m.s - 1) than at the head (0.008 m.s - 1). This implies that particles will be advected at a greater speed at the mouth than at the head which, combined with different current directions, may result in a differential net advection between the two locations. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 79 Since the currents are so variable, their monthly averages are studied, particularly in July when the cruise coincides with the time-series and when the currents are the most consistent. Figure 4.8 presents the colour-coded monthly-averaged currents for July: the red currents are at 150 m, the green at 250 m and the blue at 350 m. At 150 m, the cur-rent at the mouth supports the pattern that Klinck (1996) presents: originating from the northwest, it curves as it flows over the canyon and turns to exit on the southeast. Yet, the current at the head turns too drastically: it is north/northeastward. Looking closely at the bathymetry, the canyon appears to split in two branches close to where BC01 was. The presence of a spur may explain why the currents do not seem to follow the expected pattern in that region. Denman and Freeland (1985) also state that the prevailing cur-rents over the continental slope can be baroclinically unstable, with eddies and meanders often distorting the flow. The modelled flow (Figure 4.2 presented in Section 4.2.1) agrees with this hypothesis. At 250 m, the currents turn cyclonically inside the canyon and exit to the south at the mouth, upstream. During the month, the current at the head (BC01) shows much variability, turning from mainly southeastward to northwestward, but it is weak. Finally, the currents at 350 m are also cyclonic (north/northwestward at BC03, southwestward at BC02) but there are several days during July (as well as in May, June and September) when both currents are up-canyon, accompanying a decrease in temper-ature and therefore supporting upwelling. The cyclonic pattern is consistent with the water column being stretched as it flows over the canyon. This flow will plunge into the canyon on the upstream side and turn cyclonically to come up again at the downstream side and exit the canyon (see Sections 1.1.2 or 4.2.1). Chapter 4. Physical Properties of Barkley Canyon: the mooring data 80 126° 10'W 126° OO'W 125° 50*W Figure 4.8: Monthly-averaged currents at 150, 250 and 350 m (red, green and blue respectively) from BC01, BC02, BC03 and BC04 for July 1997. The 200-, 300- and 400-m isobaths are shown. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 81 4.2.3 Currents at mooring A l Data from mooring A l (48°32' N, 126°12' W; see Figure 3.1) (courtesy of Richard E. Thomson, IOS) were used in Chapter 2 to assess the conditions upstream of the canyon. Here as well, it is interesting to investigate the current conditions at A l . The semi-hourly values obtained from the mooring are daily-averaged and low-pass filtered with a three-day period for the plots. The currents were recorded at 32, 97, 172 and 397 m and are shown in Figure 4.9. The current at 32 m recorded the near-surface flow until the beginning of July. It is mainly southward but there is a two-week period of northward flow at the end of May. At 97 m, the current is also mainly southward, especially in July, and can be considered to represent the Shelf-Break Current. There is also a two-week period of northward current in June, which is also visible from BC02 at 250 m and BC03 at 142 m. A shift to northward flow, characteristic of the winter pattern, is visible as early as mid-August and persists until the end of the time-series. At 172 m, the current is mainly northward with two weak periods of southward flow in June and July. The directions are similar to those from BC01 at 157 m. Finally, the flow at 397 m is mostly northward, probably influenced by the California Undercurrent. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 82 Currents at mooring A1 May Jun Jul Aug Sep Oct Figure 4.9: Daily-averages of the currents from A l at 32, 97, 172 and 397 m, from April 19th until September 30th, 1997. The largest velocity is 0.337 m.s~l. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 83 4.3 Upwelling episodes in Barkley Canyon The moorings also provided time-series of temperature and salinity. There are only a few instances when the temperature decreases or the salinity increases consistently, which makes Barkley Canyon very different from Juan de Fuca Canyon as far as up-welling is concerned. Chapter 2 shows how the temperature constantly decreases in Juan de Fuca Canyon, accompanying a strong and constant upwelling. In Barkley Canyon, upwelling events are fewer and weaker. In the next figures which show the time-series of temperature, three "major" upwelling events are observed. There is one in June (sharp decrease in temperature), one in July (weaker, during the cruise period) and one in late August-early September. 4.3.1 Temperature and salinity The following Figures 4.10, 4.11, 4.12 and 4.13 present the time-series of temperature from April 19th (June 4th for BC01) to October 1st, 1997 for the four moorings. Data were lost for BC02 at 250 m after one month and for BC03 at 242 m after July. In general, the time-series show an almost constant increase in temperature and decrease in salinity (not shown) from spring to summer. This is an anomalously large and spread-out warming in the water that may be attributed to El Nino. Compared with the time-series gathered at Juan de Fuca Canyon (see Section 2.3), upwelling is not as strong and consistent in Barkley. The most important upwelling events have been marked with a red line on the tem-perature plots. At BC01, the temperature mainly increases throughout the time-series, with some periods of decrease at 157 and 257 m (very little at 357 m). There are larger decrease events around the middle of June, in July and at the end of August. At BC02, the pattern at 350 m is also mainly an increase with some indications of upwelling at the Chapter 4. Physical Properties of Barkley Canyon: the mooring data 84 Figure 4.10: Daily-averages of the temperature (°C) from BC01 at 157, 257 and 357 m in the upper, middle and lower panels (respectively) from June 4th until October 1st, 1997. The red lines denote important upwelling events. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 85 Temperature at mooring BC02 at 250 m May Jun Jul Aug Sep Figure 4.11: Daily-averages of the temperature (°C) from BC02 at 250 and 350 m in the upper and lower panels (respectively) from April 19th until October 1st, 1997. The red lines denote important upwelling events. Chapter 4. Physical Properties of Barkley Canyon: the mooring data Temperature at mooring BC03 142 m BC03 242 m 7.51 1 1 1 CD _ U I I I I l _ l I I I I I May Jun Jul Aug Sep Oct Figure 4.12: Daily-averages of the temperature (°C) from BC03 at 142, 242 and 342 m in the upper, middle and lower panels (respectively) from April 19th until October 1st, 1997. The red lines denote important upwelling events. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 87 Temperature at mooring BC04 at 230 m May Jun Jul Aug Sep Oct Figure 4.13: Daily-averages of the temperature (°C) from BC04 at 230 m, from Apri l 19th until October 1st, 1997. The red lines denote important upwelling events. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 88 end of April, in mid-June and in September. At BC03, periods of decreasing temperature at 142 m are at the end of April, in May, June and July. At 242 m, there is a sudden peak at the beginning of June when the temperature anomalously rises from 6 to 7°C. At 342 m, there is much variability with large ups and downs. Finally at BC04, the temperature at 230 m oscillates between 6 and 7.6°C with significant drops in May, June and September. The signals of temperature at BC02 and BC03 are very similar at 350 m, which may imply that they are in the same water mass. The few events of temperature decrease can be associated with upwelling, which is therefore not continuous throughout the time-series, as it is in Juan de Fuca Canyon. These episodic events start first at depth since there is a time lag (some days) before the signal initiated at depth reaches shallower waters (for instance, see Figure 4.10). Finally, salinity at all moorings has very few fluctuations and gradually decreases with time. 4.3.2 Three upwelling events During the time Barkley Canyon was monitored, four major upwelling events occurred. These are several short events (in the order of days) that occur typically over two-week periods. They are chosen for being the most important decreases in the temperature plots. They appear in May (strong decrease of temperature around the 20th, only visible' at BC03), in June (between the 4th and the 24th), in July (between the 8th and the 30th) and in August-September (between August 25th and September 10th). Since the May event is only visible in the time-series from BC03, the focus is on the three others. The strength of the upwelling (as assessed by the temperature decrease) differs from date to date, depth to depth, and station to station. The strongest event is at BC04 at 230 m between June 3rd and 12th (the temperature gradually drops by 1.1°C). During this period, the average temperature decrease across the moorings is around 0.4°C at Chapter 4. Physical Properties of Barkley Canyon: the mooring data 89 150 m, 0.9°C at 250 m and 0.7°C at 350 m. It is much smaller in July when the upwelling event (coincident with the cruise) is really a succession of small upwellings (the temperature decrease is around 0.1°C at 350 m). Between approximately August 25th and September 10th, the temperature decreases by 0.2°C at 150 m, 0.4°C at 250 m and 0.6°C at 350 m. Usually the head moorings display the largest ranges in temperature (BC04, followed by BC01). These temperature decreases correspond to a displacement of water over 20 to 50 m at 150 m deep, and over 100 to 130 m at 250-350 m deep. The large drop in June visible at BC04 is equivalent to a vertical displacement of 150 m. This is smaller than the vertical extent of the upwelling in Juan de Fuca (200 m) but still important. Hickey (1989) reports that water in Astoria Canyon can be upwelled from 200-300 m deep up over the rim of the canyon, while she finds a vertical displacement of as much as 100 m in Quinault Canyon. The vertical extent of upwelling is also larger within Barkley Canyon than on the shelf: for instance at A l , the maximum displacement is 20 m at 97 m deep. A thermistor chain was placed on the downstream mouth mooring (BC03) under the 150-m instrument and recorded time-series of temperature at 144, 154, 164, 174, 179, 184, 189 and 194 m deep. The major upwelling events cannot be easily seen but a weak upwelling brings water from 190 to 160 m in May-June. There is also a pool of colder water around 180 m at the same period. Mooring A l provides a clearer time-series from a thermistor chain placed under the 99-m instrument (courtesy of Richard E. Thomson, IOS). It recorded the temperature at 99, 104, 109, 114, 119, 124, 129, 134, 139, 144 and 149 m deep. Figure 4.14 is a contour plot of the temperature from this thermistor chain. Upwelling events are seen more clearly in this figure, especially in May, but also in June, July and September (they correspond to the upwelling events isolated from the mooring data). A strong warming of the surface waters beginning in July is visible in both time-series. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 90 Figure 4.14: Contours of temperature (°C) from the thermistor chain at A l over time and depth, from April 19th until September 30th, 1997. Chapter 4. Physical Properties of Barkley Canyon: the mooring data 91 During these events, neither the wind nor the currents are in a consistent upwelling-favorable direction. Yet, as far as the wind is concerned, there is always a small period of northwesterlies before the onset of the upwelling. As already stated in Section 4.1, the only strong northwesterly period is in July but it does not result in a stronger upwelling. This supports the hypothesis that it is the currents (especially the Shelf-Break Current), more than the wind, which act as the driving mechanism for upwelling. The only surface-water mooring available to assess the near-surface current is A l but it is not in the direct vicinity of the canyon and the time-series stops at the beginning of July. The current at 97 m, although not in the surface waters, still represents the Shelf-Break Current as stated in Section 4.2.3 (Figure 4.9). It varies sharply in direction and in intensity around the three upwelling periods described. It turns from northwestward to southeastward at the beginning of June coincident with the June upwelling. It decreases in intensity in the middle of July and then strengthens again at the time of the July upwelling. After turning to the northwest in August, it briefly returns to the south during the September upwelling period. These changes to the south of the Shelf-Break Current are well correlated with the observed upwelling events in Barkley Canyon and may influence their onset. During the June event, the temperature drop at 250-350 m is equivalent to an upwelling of 110 m. There is a clear southeastward flow at 250 m at all moorings except upstream at the head (BC04). At 150 m, it is also southeastward at the mouth downstream (BC03) but east/northeastward at the head (BC01). At depth, the flow is cyclonic: southwestward at the mouth upstream (BC02) and northwestward downstream (BC03), but northeastward at the head (BC01). There is not a consistent deep up-canyon flow, as seen in Juan de Fuca Canyon, but only instances when the currents at the mouth are up-canyon during several days. Some up-canyon flow during upwelling is also reported by Hickey (1997) in Astoria Canyon at her two slope moorings (north and south). Chapter 4. Physical Properties of Barkley Canyon: the mooring data 92 During the weaker July upwelling, the water is displaced over 40 m at 250-350 m. The flow downstream at the mouth (BC03) is southeastward above the rim and northeastward below it. It is southwestward upstream (BC02) but occasionally, the currents from both moorings at 350 m align to the north (up-canyon). This deep up-canyon flow is associated with upwelling, as it accompanies the temperature drops. At the head downstream (BC01), the flow is north/northeastward above the rim and southeastward below it. 4.4 Summary The winds around Barkley Canyon are mainly southeasterlies but variable. There is only one northeasterly (upwelling-favorable) event around July. The currents fluctuate but partly support the theory and the modelled flows. The 150-m current bends across the isobaths and exits to the southeast. The head mooring does not follow this pattern, due to the bathymetry or a possible eddy (present in the modelled flow). The eddy may be due to distortions in the flow caused by the local topography which makes it unstable. At depth, the currents are mainly cyclonic. Some up-canyon flow (also found in Astoria) supports upwelling. Despite variable currents, a pattern for July can be drawn which will help understand the zooplankton distribution in Chapter 5. As a pre-El Nino year, the 1997 upwelling season is relatively weak: the temperature consistently increases throughout the record but there are still some instances when upwelling occurs (in May, June, July and September). The shift to the south of the Shelf-Break Current, as recorded by mooring A l , is well correlated with the upwelling events. The upwelling range is around 40 m at 150 m deep and between 100 and 150 m at 250-350 m deep. It is enhanced compared to the shelf (Al). It is similar in Astoria and Quinault Canyons but much larger in Juan de Fuca Canyon. Hopefully, the year 1998 will demonstrate stronger upwelling events and perhaps clearer current patterns. Chapter 5 Circulation patterns and zooplankton transport This study presents the interaction between Barkley Canyon's circulation and zooplank-ton aggregation. It is believed that some migratory species combine vertical migration and current patterns to remain in areas where food availability is high and advection is low. The head of canyons is a favorite spot but there are other places where the canyon creates a favorable environment for zooplankton development. Zooplankton species com-mon off Vancouver Island are described and their aggregation patterns are discussed with respect to the circulation around Barkley Canyon. 5.1 Introduction In the marine ecosystem, the first trophic level is the phytoplankton, the plant com-ponent of the plankton. The second level is the zooplankton (the animal component) which graze on phytoplankton (herbivores), on dead organic material (detritivores) or on other zooplankton (carnivores). Zooplankton are the primary food source for many fish on the west coast of Vancouver Island and are therefore important in the global marine ecosystem. Several studies have granted much importance to upwelling zones and regions of steep bathymetry as sites where zooplankton aggregate (Mackas, 1992; Mackas et al., 1997; Peterson et al, 1979; Wroblewski, 1982). 93 Chapter 5. Circulation patterns and zooplankton transport 94 5.1.1 Zooplankton species on the west coast of Canada On the west coast of Vancouver Island, some species are found abundantly. The most abundant are the calanoid copepods (Acartia longiremis, Paracalanus parvus, Calanus marshallae, Pseudocalanus spp., Metridia pacifica, Neocalanus plumchrus). They usually make up 70% or more of all net-collected plankton. As part of the Crustacea subphylum, they have a segmented body with three distinctive body regions. Some tend to feed by capturing phytoplankton in currents generated by movements of the swimming legs and mouthparts. Development involves twelve different stages: the first six are nauplius stages, the last six are copepodite stages. The cyclopoid copepods (Oithona similis, Oithona atlantica) are also very abundant. The euphausiids (Euphausia pacifica, Thysa-noessa spinifera) are another order of crustaceans and are an important food source for many harvested fish off British Columbia (Mackas, 1992; Simard and Mackas, 1989). The eggs hatch into nauplii (a larval stage), which change into calyptopis stages (also larval). Those change into furcilia stages (juvenile), which finally become adults. They are generally omnivorous, feeding on detritus, phytoplankton and smaller zooplankton. A fourth group is the chaetognaths or arrow worms (Sagitta elegans, Sagitta scrippsae, Eukrohnia hamata). They are often motionless in the water but can pursue a prey with swift darting motions. Their food source is generally smaller zooplankton but they are not selective and adapt to the local relative abundance of food. Finally, the larvaceans (Oikopleura sp.) look like tadpoles. Most species secrete a spherical balloon of mucus in which they reside. Water is filtered through this balloon which provides the animals with food. When the filter becomes clogged, the house of mucus is abandoned. Larvaceans are especially abundant in coastal waters (Lalli and Parsons, 1993). Chapter 5. Circulation patterns and zooplankton transport 95 5.1.2 Zooplankton vertical patterns Most zooplankton species are epipelagic, i.e. they permanently inhabit the region extending from below the surface to 200 or 300 m deep, even during the daytime (e.g. copepods, larvaceans). Others migrate into this region at night from deeper depths while during the day, they live in the mesopelagic zone that lies between the bottom of the epipelagic region and around 1000 m deep (e.g. euphausiids). They move upward at night to feed on phytoplankton. This vertical migration occurring with a 24-hour period (upward to the surface at night and downward to deeper waters during the day) is called diel vertical migration (see Section 1.2.2). Changes in ambient light intensity may be one of the primary stimuli for this migration. Animals are thought to migrate daily so that they remain in near-darkness over 24 hours and are therefore less vulnerable to visual predators. They also migrate upward to be in the surface zone where food is more abundant. Finally, some research has shown that this vertical migration puts the animals in currents moving in different directions and speeds, which ensures retention within an appropriate habitat or within a productive upwelling area (Peterson et al., 1979; Wroblewski, 1982). The vertical migration of some species can also be seasonal, associated with different life stages: ontogenetic migration reflects the different depth preferences of eggs, nauplii and adults. Dominant copepods show changes in their depth patterns in inshore and offshore waters, and also in winter and summer. They stay at different depths whether they are adults laying eggs, eggs, nauplii or maturing copepodite stages. For example, in the inshore waters, Neocalanus plumchrus adults are usually at depth (300-450 m) laying eggs in the winter (December-April). The nauplii are in the near-surface waters from February to April and mature to the copepodite V stage during March to June. In the offshore waters, the spawning takes place for a more prolonged period of time, from July to February (Lalli and Parsons, 1993). Chapter 5. Circulation patterns and zooplankton transport 96 5.2 Zooplankton spatial aggregation Bongo tows were performed during the cruise in July 1997 to assess the different spatial patterns of zooplankton around the canyon. 50-m and 250-m tows were performed at five stations (green stars in Figure 3.2) during day and night. The 50-m tows are integrated from 50 m deep to the surface and the 250-m tows from 250 m (or the bottom depth) to the surface. Bongo nets with a 0.56-m diameter and 236-/zm mesh size were used. The animals were preserved in buffered 5% formalin solution, then counted at the Institute of Ocean Sciences (Sidney, BC). Two of the stations were close to the walls at the mouth of the canyon (BCB4, BCD4), one was in the center of the canyon (BCC3) and the last two were at the head of the canyon (BCB2, BCD2) on the shelf. The 50-m night tow at station BCD2 was missed. The following species were selected in the study: large copepods were represented by the genera Neocalanus and Eucalanus; small copepods by Paracalanus, Pseudocalanus, Calanus, Acartia and Oithona; chaetognaths by Sagitta and Eukrohnia; euphausiids by Euphausia and Thysanoessa; and larvaceans by Oikopleura. Since the data come from one point in time, it is difficult to generalize and therefore some results may not agree with what is expected. In particular, when comparing night and day tows, the migratory species do not appear to be abundant in the 50-m night tows and 250-m day tows. Most species are more abundant in the 50-m tows during the day (which is somehow strange since it is believed that animals are able to avoid the nets during the day). Finally, all values at BCC3 at the canyon axis are much lower than at the other stations. Appendix B presents the zooplankton data from the Barkley Canyon study. Differentiating species according to their swimming abilities and their natural habitat proves useful in assessing their response to the canyon's circulation. Figure 5.1 shows Chapter 5. Circulation patterns and zooplankton transport 97 the classification of different species whether they are expected to be migrators or non-migrators, and expected to be living on the shelf, the shelf-edge or the open ocean (D. Mackas, pers. comm.). The abundance of zooplankton is compared at all stations and the results of the canyon study are gathered between stations BCB2 and BCD2 as the "head" and between BCB4 and BCD4 as the "mouth". There are little differences between the upstream and downstream stations, the largest differences are between the "head" and the "mouth" stations. The results of the canyon study are presented in Figure 5.1: species that are mainly at the head of the canyon are coded in red dots, those that are more abundant at the mouth are in blue, and the species that are found in similar abundance at the head and mouth are in green. Calanus is the only genus for which a diel migration is clearly evident. It is most abundant in the 50-m night tows and the 250-m day tows, and especially at the head. There appears to be a segregation between the different life stages since the copepodites (stage III or lower) who do not have swimming ability are most abundant at the mouth. Metridia (both the adults and the copepodites) also appear to experience a diel migration but it is less clear. They are generally present at the mouth in the 50-m tows during the night and partially in the 250-m tows during the day. E. pacifica and T. spinifera are most present at the head during the night in the 50-m tows but only partially in the 250-m during the day, while Euphausia furcilia are at the mouth and Euphausia calyptopis are equally abundant at the head and at the mouth. P. parvus is the most abundant species in all tows especially in the 50-m day tows at the mouth. The genus Sagitta does not appear to migrate (it is abundant in the 50-m day tows) and is present at the mouth (S. scrippsae) as well as at the head (S. elegans). E. hamata (supposedly a migrator) is abundant at the mouth in the 250-m tows during the day but scarce in the 50-m tows even at night. A. longiremis is most present at the head in the 50-m tows but the copepodites (stage IV or lower), like in the case of Calanus, are abundant at the Chapter 5. Circulation patterns and zooplankton transport 98 Oceanic Shelf-edge Shelf Oikopleura sp. Non-migrator Neocalanus plumchrus O Acartia longiremis Oithona spp. • Euphausia calyptopis Calanus marshallae Paracalanus parvus Calanus marshallae Euphausia pacifica Migrator Thysanoessa spinifera Metridia pacifica Eukrohnia hamata Sagitta scrippsae Sagitta elegans # Inshore species # Intermediate species # Offshore species Figure 5.1: Chart of the expected species' region preferences (from D . Mackas, pers. comm.) and of the canyon study results (dots). Species found at the head are marked with a red dot, those at the mouth with a blue dot, and species with a green dot are equally abundant at both locations. Chapter 5. Circulation patterns and zooplankton transport 99 126°20'W 126'00'W 125° 40'W Figure 5.2: Relative abundance of Acartia longiremis and Euphausia calyptopis in the 50-m tows during the day (the 300-m isobath is shown in bold). Chapter 5. Circulation patterns and zooplankton transport 100 mouth. Oithona spp. (0. similis, 0. atlantica) are more abundant at the mouth, mostly in the 50-m tows during the day (they are usually found in both coastal and oceanic waters). Oikopleura sp., although mainly a shelf species, is most found offshore in the first 50 m. Finally, Pseudocalanus spp. are particularly present in the first 50 m during the day at the canyon axis (BCC3) and at the head (BCB2). As an example, Figure 5.2 illustrates the relative abundance of A. longiremis and Euphausia calyptopis in the 50-m tows during the day. From these results, summarized in Figure 5.1, it appears that most non-migratory species are aggregating at the mouth along the walls. Their cross-shelf distribution in the canyon is relatively similar to that outside the canyon. On the contrary, migratory species have a different distribution: they are found closer to the head. Some species (e.g. E. pacifica) that are usually farther offshore are present closer inshore within the canyon, which may mean that they are advected and trapped along the canyon's head wall. This suggests that there may be an interaction between the species' vertical diel migration and the canyon's deep flow. During their diurnal descent, the animals may be caught in the deep up-canon flow and be advected towards the head where they may be trapped against the walls or within a possible eddy. A separation of stages is apparent with nauplii sometimes aggregating at different places than their adult counterparts. These results agree to some extent with the literature (e.g. Mackas, 1995) and show that the canyon, by its current patterns and enhanced food supply from upwelled waters, is a zone of zooplankton aggregation especially along the canyon walls. Chapter 5. Circulation patterns and zooplankton transport 101 5.3 Links between physics and biology Some types of mixing (upwelling) result in elevated surface nutrient concentrations, high primary production and increased numbers of zooplankton (Lalli and Parsons, 1993). Currents can trap animals and reduce their advection offshore, thus reducing one of the greatest causes for zooplankton loss (see Section 1.2.1). Canyons as regions of enhanced upwelling and steep bathymetry provide both a favorable food environment and a favorable current system. The circulation pattern tends to trap zooplankton against the canyon walls (Macquart-Moulin and Patriti, 1996; Mackas et al, 1997). In Barkley Canyon, the nutrient data gathered during the cruise in July 1997 show that surface nitrate (NOs~) levels increase towards the shore (S. Harris, pers. comm.), probably due to upwelling of deep waters onto the shelf. There is a particularly sharp increase between the station sampled just before the shelf break and the station sampled on the shelf close to the canyon's head. This higher supply of nutrients within and in the vicinity of the canyon may have a positive impact on phytoplankton productivity and therefore on zooplankton growth for species able to inhabit this region. Many species are believed to have different depth preferences. In order to remain in a favorable environment, some can vertically migrate on a daily basis (diel migration) or vertically migrate during their life cycle (ontogenetic migration) so that they enter different current systems. Other species are passive animals and do not migrate vertically. In addition, zooplankton have a cross-shelf distribution whereby oceanic species are found offshore, shelf-edge species are found at the shelf break and shelf species are found more inshore. In Barkley Canyon, the current system varies with the depth and the location, and therefore places the animals in different environments. At 30 m from the surface, the flow is mainly southeastward except at the head where a cyclonic eddy may form (see Section 4.2.1). Animals present in this region in the first Chapter 5. Circulation patterns and zooplankton transport 102 50 m may therefore be trapped at the head where upwelling also brings nutrients to phytoplankton. Animals present at the mouth (either upstream or downstream) must be carried away southward by the current. Since the current direction (southeastward) is aligned with Line BC4 (BCB4-BCD4), the fact that zooplankton are abundant at these stations may only reflect the flow direction, i.e. they are not just abundant at the walls but all along the line aligned with the shelf break across the canyon mouth. At depth, the currents are mainly advecting animals inside the canyon towards the walls, especially downstream at the mouth. Upstream at mooring BC02, the flow is out-canyon (southwestward) but the animals may then be advected by the in-canyon flow right at the mouth and be trapped in the cyclonic eddy at the head (fully-developed around 150 m). In conclusion, the current systems are different at 30, 150 and 300 m, thus the animals at these depths are advected differently. Accumulation of zooplankton along the walls near the head or the mouth of the canyon illustrates the different influences the currents have on the animals depending on whether they migrate vertically. The study hypoth-esis was that zooplankton aggregated at the head of the canyon, either naturally or by migrating in order to be caught in the deep in-canyon flow. Yet, there are many species that are more abundant at the mouth. In order to fully understand their interaction with the canyon circulation, it would be interesting (but unrealistic) to follow a group of animals and observe if specimens abundantly present at the mouth close to the surface managed to reach the productive region of the canyon's head by migrating to depth to be caught in the up-canyon flow. This group would probably have suffered subsequent losses, which would explain why the species is not as abundant at the head. Chapter 5. Circulation patterns and zooplankton transport 103 5.4 Error analysis Statistical errors are typically associated with the patchiness characteristic of zoo-plankton distributions and the sampling techniques. The individuals of most species are often distributed in a patchy, or non-random manner. Patchiness can be due to interaction between zooplankton and their food, or to reproduction patterns. From a physical point of view, gyres, upwelling zones, turbulence and the Langmuir circulation may also induce patchiness. As a result, small-scale differences in the horizontal zooplankton distribution are difficult to detect. Nets are usually towed through the water for distances on the order of tens of meters or more, so that the numbers of collected animals are averaged over these distances. The small-scale patchiness is then masked (Lalli and Parsons, 1993). The only way to estimate the error due to patchiness is by replicating samples, which was not possible during the cruise. Cassie (1963) finds that the coefficient of variation of a single plankton sample due to patchiness is in the order of 22-44%, although larger coefficients are not uncommon. Errors due to sampling techniques can arise from flowmeter readings or from animal counts. When the net is towed, a flowmeter records the net distance travelled, which is multiplied by the area of net opening to get the volume of filtered water. This is then used to calculate the number of animals per cubic meter. Occasionally, the flowmeter does not work properly or the towing cable forms a large angle with the vertical. This happens when the net is towed too fast or when there is too much wind. The flowmeter will then overestimate the volume filtered and the reading will induce an error on the numbers. Visually, most volumes appeared coherent in this study. Finally, the errors due to counting depend on the number of animals. The more animals are counted, the smaller is the range of error. Edmondson and Winberg (1971) say that the precision of a count can be calculated as Vm = -T= , where Vm is the coefficient of variation, m is the Chapter 5. Circulation patterns and zooplankton transport 104 number of animals counted and n is the number of subsamples. Since n is always 1 (only one subsample is counted), Vm — The coefficient of variance is then 10% for 100 animals counted or 50% if only 4 animals are counted. Errors also arise from the fact that most samples are split to be counted. Each splitter is unique but typically they are precise to ± 2%, so each time a sample is split another 2% error is introduced. A thick sample may be split five times. There is no way to precisely estimate the amount of error present in these samples other than by replicating them, which is unfortunately time- and money-consuming. A 20-40% error on the numbers from this study does not mask the differences between head and mouth, and thus does not introduce a significant error on the principal conclusions. The numbers look reasonable when compared to the literature and are therefore accepted as representative of the zooplankton population in the region of Barkley Canyon, at the time the tows were performed. Chapter 6 Conclusions Initially, the purpose of this thesis was to focus on the influence of submarine canyons on zooplankton aggregation. Some conclusions have been drawn but the results from the next year of sampling will allow a more comprehensive assessment. The focus was then to study the circulation, water properties and especially upwelling in two canyons. Juan de Fuca Canyon was interesting because of its location and geometry, and data from 1993 were analysed. Barkley Canyon was surveyed in 1997 and the data were interpreted to characterize the current system, hydrography and zooplankton spatial accumulation. 6.1 Juan de Fuca Canyon Juan de Fuca is a narrow and long canyon, situated at the mouth of Juan de Fuca Strait, in the path of the estuarine flow from the Fraser River. These characteristics confer upon it a particular physical behaviour. The currents were found to generally flow southeastward, typical of the summer wind-driven Shelf-Break Current (0.08 m.s - 1 at 30 m and 0.05 m.s"1 at 75 m). But at depth (200 m), the current turned strongly up-canyon with velocities of around 0.03 m.s~l. This resulted from the pressure gradient set by the cross-canyon Shelf-Break Current not being balanced by the Coriolis force within the canyon. The consequence was an enhanced upwelling in the canyon compared to the coastal upwelling occurring on the shelf. Values of temperature inside the canyon showed a constant decrease from May to October 1993, equivalent to water being upwelled over 200 m. The temperatures at 200 m were highly correlated with the cross-canyon flow at 105 Chapter 6. Conclusions 106 75 m. The California Undercurrent at 400 m was recorded by mooring A l . It was found to have no influence on the upwelling since its onset in July did not perturb the strength of the upwelling. 6.2 Barkley Canyon Barkley Canyon is wider and shorter than Juan de Fuca Canyon and situated off Barkley Sound. From the CTD data, the mixed layer was found to be 10 m deep, below which the water properties were no longer well-mixed but varied with depth. Temper-ature/Salinity diagrams showed a separation between warmer, saltier offshore surface waters and colder, fresher inshore ones. An internal tide signal (equivalent to a vertical displacement of 7 m around 20 m deep and 14 m around 100 m deep) was seen in the CTD casts but was difficult to isolate. EOF analysis showed that the upwelling, and not the internal tide, was responsible for most of the variance. Although the semi-diurnal frequency was the main frequency in the spectrum analysis, there was more energy in the three-day period associated with upwelling. Based on temperature differences, upwelling was found to be stronger inside the canyon than on the shelf. Across-shore, the isopleths domed over the center of the canyon close to the surface and upwelled towards the head. Alongshore, the dome was still visible close to the surface but the isopleths bulged down deeper, in response to the depth plunging from 200 to 600 m over the canyon axis. The isohalines contoured over the canyon also showed a doming over the canyon around 10 m deep, and a "trough upstream/ridge downstream" pattern at deeper depths. The fact that the isopleths were perturbed so close to the surface means that the influence of the canyon's circulation was extending to shallower depths than other canyons (like Astoria for instance), i.e. the stratification parameter was weaker in wider Barkley Canyon. This was supported by the physical Chapter 6. Conclusions 107 parameters derived from its geometry and currents (Rossby and Burger numbers). The data from the moorings provided time-series of currents, temperature and salinity. The currents were highly variable at all depths. At 150 m, the flow at the head may have been embedded in an eddy or affected by the local bathymetry (there appears to be a split, i.e. a spur at the head of the canyon). At the mouth, the flow followed the expected pattern: it bent across the isobaths and exited the canyon to the southeast. At 250 m, the flow was mainly cyclonic (northwestward downstream, southwestward upstream). The flow was also cyclonic at 350 m with some instances when it was up-canyon both up- and downstream, accompanying periods of temperature decrease (upwelling). Throughout the time-series, the temperature continued to increase and did not show a strong and constant upwelling signal. There were several periods of temperature decrease (upwelling) during the record, with three major upwelling events in June, July and at the end of August-beginning of September. They were equivalent to a maximum 50-m vertical displacement at 150 m deep and 150-m displacement at 250-350 m deep. June displayed the largest range of temperature decrease. The upwelling range from the mooring records was comparable to the CTD observations. Generally, the flow patterns in Barkley Canyon were very variable and the water properties did not support a very strong upwelling. It was more intense within the canyon than on the shelf, and comparable to the events in Astoria Canyon, but it was weaker than in Juan de Fuca Canyon. 6.3 Zooplankton aggregation Zooplankton tows were performed in Barkley Canyon in July 1997. Species had different aggregation patterns whether they were migrating or not and whether their natural habitat was the shelf, the shelf-break or the open ocean. Most species were Chapter 6. Conclusions 108 found abundantly in the 50-m day tows. In general, migratory species were found closer to the head while passive species were more abundant at the mouth. Swimming species are believed to combine vertical migration with current patterns to remain in favorable environments. With a high supply of nutrients and phytoplankton, the head of the canyon would be the most favorable environment for them. Also the deep circulation (up-canyon) would facilitate a return to the head. The non-swimming species are dependent on the current direction. The fact that they were accumulating at the mouth stations along the walls could 1) indicate the existence of a convergence zone on each side of the canyon, 2) be a result of the small-scale circulation on the day the tows were performed or 3) imply that they were abundant not only at the walls, but all along the line between the two mouth stations, as a result of the southeastward direction of the Shelf-Break Current. There were a few exceptions in the distribution pattern. Some migratory species whose natural habitat is the open ocean were still found at the mouth, rather than at the head. Also, most of them did not seem to experience a diel migration. This can arise from the fact that only a few tows were performed, which may not be representative of the species' normal behaviour. 6.4 Further work Barkley Canyon will be the place of an extensive survey for another year. In preparing the plan for the data collection, it would be important to visit more stations (bearing in mind that ship time is valuable) so that statistical analyses could be performed. It was difficult to plot sections of quantities because of the few stations that had been sampled. Also, time-series from the tow-yo cast were sometimes too far apart to be able to spot an internal tide signal. More stations should also be sampled for the zooplankton data and care should be taken in choosing these stations and the time of sampling. Chapter 6. Conclusions 109 The other initial purpose of the thesis was to help develop a bio-physical model that would account for the circulation within the canyon and the zooplankton distribution. If this project were carried out, the data collected during the cruises would be important for correctly choosing the parameters of the model. Oguz et al. (1996) developed a one-dimensional physical-biological model to simulate the annual plankton productivity cycle. Their main conclusion relevant to this thesis was that extensive surveys were needed in order to select accurate parameters. I believe that a two-year time-series of current, temperature and salinity observations will help improve our knowledge of submarine canyons. They are places that can still raise more issues of interest and it is therefore important to keep investigating them. Studying the close interaction with zooplankton and other trophic levels may also enable fisheries to more efficiently manage our resources. Bibliography A L L E N , S . E . 1996: Topographically generated, subinertial flows within a finite length canyon. J. Phys. Oceanogr. 26(8): 1608-1632. C A N N O N , G . A . 1972: Wind effects on currents observed in Juan de Fuca submarine canyon. J. Phys. Oceanogr. 2(3):281-285. C A S S I E , R . M . 1963: Microdistribution of plankton. Oceanography and marine biology: an annual review 1:223-252. C H E N , X . and S . E . A L L E N , 1996: The influence of canyons on shelf currents: a theo-retical study. J. Geophys. Res. 101 (C8): 18,043-18,059. D E N M A N , K . L . and H . J . F R E E L A N D , 1985: Correlation scales, objective mapping and a statistical test of geostrophy over the continental shelf. J. Mar. Res. 43:517-539. EDMONDSON, W . T . and G . G . WINBERG, 1971: A manual on methods for the assess-ment of secondary productivity in fresh waters. Blackwell Scientific Publications, Oxford, 358 pp. F O R E M A N , M . G . G . and R . E . THOMSON, 1997: Three-dimensional model simulations of tides and buoyancy currents along the west coast of Vancouver Island. J. Phys. Oceanogr. 27(7): 1300-1325. F R E E L A N D , H . J . , W . R . CRAWFORD, and R . E . THOMSON, 1984: Currents along the Pacific coast of Canada. Atmosphere-Ocean 22(2):151-172. 110 BIBLIOGRAPHY 111 F R E E L A N D , H . J . and K . L . D E N M A N , 1982: A topographically controlled upwelling center off southern Vancouver Island. J. Mar. Res. 40(4):1069-1093. F R E E L A N D , H . J . and P. MCINTOSH, 1989: The vorticity balance on the southern British Columbia continental shelf. Atmosphere-Ocean 27(4):643-657. G I L L , A . E . 1982: Atmosphere-Ocean Dynamics. Academic Press, San Diego, 662 pp. H I C K E Y , B . M . 1989: Patterns and processes of circulation over the Washington conti-nental shelf and slope. In LANDRY, M . and B. H I C K E Y (eds), Coastal Oceanography of Washington and Oregon, pages 41-116. Elsevier Science, Amsterdam. H I C K E Y , B . M . 1995: Coastal submarine canyons. In M U L L E R , P. and D. HENDERSON (eds), 'Aha Huliko'a Hawaiian Winter Workshop on topographic effects in the ocean, pages 95-110. SOEST Special Publication, University of Hawaii at Manoa. H I C K E Y , B . M . 1997: The response of a steep-sided, narrow canyon to time-variable wind forcing. J. Phys. Oceanogr. 27(5):697-726. H I C K E Y , B . M . , E . B A K E R , and N . K A C H E L , 1986: Suspended particle movement in and around Quinault Submarine Canyon. Marine Geology 71:35-83. KINSELLA, E .D. , A . E . H A Y , and W.W. DENNER, 1987: Wind and topographic effects on the Labrador current at Carson Canyon. J. Geophys. Res. 92(C10):10,853-10,869. K L I N C K , J . M . 1988: The influence of a narrow transverse canyon on initially geostrophic flow. J. Geophys. Res. 93(C1):509-515. K L I N C K , J . M . 1989: Geostrophic adjustment over submarine canyons. J. Geophys. Res. 94(C5):6133-6144. BIBLIOGRAPHY 112 K L I N C K , J . M . 1996: Circulation near submarine canyons: a modeling study. J. Geophys. Res. 101(C1):1211-1223. L A L L I , C M . and T . R . PARSONS, 1993: Biological oceanography: an introduction. Perg-amon Press, Oxford, 301 pp. M A C K A S , D . L . 1992: Seasonal cycle of zooplankton off southwestern British Columbia. Can. J. Fish. Aquat. Sci. 49:903-921. M A C K A S , D . L . , R . KIESER, M . SAUNDERS, D . R . Y E L L A N D , R . M . B R O W N , and D . F . M O O R E , 1997: Aggregation of euphausiids and Pacific Hake (Merluccius productus) along the outer continental shelf off Vancouver Island. Can. J. Fish. Aquat. Sci. 54:2080-2096. M A C K A S , D . L . and H. S E F T O N , 1982: Plankton species assemblages off southern Vancouver Island: Geographic pattern and temporal variability. J. Mar. Res. 40(4):1173-1200. M A C Q U A R T - M O U L I N , C. and G. PATRITI, 1996: Accumulation of migratory micronek-ton crustaceans over the upper slope and submarine canyons of the northwestern Mediterranean. Deep-Sea Res. 43(5):579-601. NAIMIE, C . E . , J . W . L O D E R , and D . R . L Y N C H , 1994: Seasonal variation of the three-dimensional residual circulation on Georges Bank. J. Geophys. Res. 99(C8):15,967-15,989. O G U Z , T . , H. D U C K L O W , P. M A L A N O T T E - R I Z Z O L I , S. T U G R U L , N . P . NEZLIN, and U . UNLUATA, 1996: Simulation of annual plankton productivity cycle in the Black Sea by a one-dimensional physical-biological model. J. Geophys. Res. 101(C7):16,585-16,599. BIBLIOGRAPHY 113 P E D L O S K Y , J . 1987: Geophysical Fluid Dynamics. Springer-Verlag, New York, 710 pp. P E R E N N E , N., J . V E R R O N , D. RENOUARD, D . L . B O Y E R , and X . Z H A N G , 1997: Recti-fied barotropic flow over a submarine canyon. J. Phys. Oceanogr. 27(9): 1868-1893. P E T E R S O N , W . T . , C.B. M I L L E R , and A. HUTCHINSON, 1979: Zonation and main-tenance of copepod populations in the Oregon upwelling zone. Deep-Sea Res. 26(A):467-494. PILLAR, S.C., D.A. ARMSTRONG, and L . HUTCHINGS, 1989: Vertical migration, dis-persal and transport of Euphausia lucens in the southern Benguela current. Mar. Ecol. Prog. Ser. 53:179-190. R O U G H G A R D E N , J . , S. GAINES, and H . POSSINGHAM, 1988: Recruitment dynamics in complex life cycles. Science 241:1460-1466. SHEPARD, F.P. , N.F. M A R S H A L L , P.A. M C L O U G H L I N , and G . G . SULLIVAN, 1979: Currents in submarine canyons and other seavalleys. The American Association of Petroleum Geologists, Tulsa, Oklahoma, 173 pp. SIMARD, Y . and D . L . M A C K A S , 1989: Mesoscale aggregations of euphausiid sound scattering layers on the continental shelf of Vancouver Island. Can. J. Fish. Aquat. Sci. 46:1238-1249. T H O M S O N , R . E . 1981: Oceanography of the British Columbia coast. Can. Spec. Publ. Fish. Aquat. Sci. 56, Department of Fisheries and Oceans, Ottawa, 291 pp. BIBLIOGRAPHY 114 T H O M S O N , R . E . , B.M. H I C K E Y , and P . H . L E B L O N D , 1989: The Vancouver Island Coastal Current : Fisheries barrier and conduit. In BEAMISH, R . J . and G.A. McFARLANE (eds), Effects of ocean variability on recruitment and an evaluation of parameters used in stock assessment models, volume 108, pages 265-296. Can. Spec. Publ. Fish. Aquat. Sci. T H O M S O N , R . E . and D.M. W A R E , 1996: A current velocity index of ocean variability. J. Geophys. Res. 101(C6):14,297-14,310. T H U R M A N , H . V . 1994: Introductory oceanography. Macmillan Publishing Company, New York, 550 pp. WROBLEWSKI , J.S. 1982: Interaction of currents and vertical migration in maintaining Calanus marshallae in the Oregon upwelling zone - a simulation. Deep-Sea Res. 29(6A):665-686. Appendix A Empirical Orthogonal Function Analysis A . l Theoretical purpose of E O F analysis EOF analysis is used with a large set of data measured both on a discrete spatial grid and at discrete points in time, when the user wants to reduce the data. The set of data is Si(tn) where i G {1,...,N} labels the station realisations and n € {1,...,T} labels the observation times. If there is a correlation between the Si(tn) at different times tn, these vectors will then tend to be aligned along certain axes. The purpose of the EOF analysis is to define a basis of vectors {ei)f=l that characterizes the cluster of data by maximizing for each i the time-mean of the squares of the projections of the data vectors on the basis vectors (< (s • e;)2 >) (where < A >= ^ J2T=i A(tn)). A.2 Practical use of E O F analysis EOF analysis involves the construction of the matrix of fluctuations Css: Css =< (s- < s >)(s- < s >)T > (if < s >^ 0). It is easy to see that (Css)ij = (Css)ji, so that Css is a symmetric matrix, hence orthog-onally diagonalisable, and whose eigenvectors form an orthogonal basis. A.2.1 The E O F vectors We are looking for the eigenvectors and eigenvalues u-k such that: Css Ck = £k • 115 Appendix A. Empirical Orthogonal Function Analysis 116 The (e*;) are then denoted as the empirical orthogonal functions (EOF). Since the (e^ ) form a basis, we can express each data vector at tn: N s(tn) = J2ak(tn) • ek k=l and a>k(tn) = s(tn) • ek are the principal components and constitute a time series for each of the associate EOF. The pair of an EOF and its principal component is called a mode. Since the trace of an orthogonally diagonalisable matrix is equal to the sum of its eigenvalues, we have: N N ^2(css)jj — H i=i j=i where the diagonal elements of Css are the variance of the signal at each station. Thus, the trace of Css is the total variance in the data set and ^ k is the fraction of the variance explained by the kth mode. The final step in EOF analysis is to reduce the size of the data set by retaining only a small number of the leading EOFs. Usually, only the first few modes explain significant variance. A.2.2 The E O F contributions The contribution of an EOF to a certain time index is derived by projecting the raw data onto this EOF mode. That is, if Sj(t„) is the value of the data set at the ith station for time tn, and ek(i) is the value at the iih station of the kth mode, then the contribution of the kth mode to the nth time is: a(n,k) = Si(tn) • ek(i) . Appendix A. Empirical Orthogonal Function Analysis 117 If this projection is higher at one time than at another, then the mode in question (k) is more important at the first time. A.3 Application of the E O F analysis for this thesis In Chapter 3, a variation of the EOF analysis is used to find a way to express the density structure in the canyon. Instead of space and time, depths and stations were used. Depths replaced the station realisations while each station represented a time observation. By using this method, we were able to find EOFs that accounted for most of the vertical distribution of density fluctuations within the domain. Only the first three EOFs were relevant here, explaining between 7 and 62% of the variance. EOF analysis was also performed with the casts done at BCB4, BCC4 and BCD4. In this time-series was obtained but the depths still replaced the station realisations. This helped identify which was the most important signal for the vertical variations of the isopleths: upwelling or internal tide. Appendix B Zooplankton data from the Barkley Canyon study This appendix presents the number of animals per cubic meter that were found in the bongo tows performed during the cruise over Barkley Canyon in July 1997. Only the most significant species were retained: Acartia sp.<YV, Acartia longiremis, Calanus sp.<III, Calanus marshallae, Neocalanus plumchrus, Metridia sp.<lV, Metridia paci-fica, Oithona similis, Oithona atlantica, Paracalanus parvus, Pseudocalanus spp., Thysa-noessa spinifera, Euphausia pacifica, Euphausia furcilia, Euphausia calyptopis, Eukrohnia hamata, Sagitta elegans, Sagitta scrippsae and Oikopleura <5mm. The following table presents the different stations, depths and times of the tows, and the number of animals per cubic meter (# per m3). 118 Appendix B. Zooplankton data from the Barkley Canyon study 119 Station BCB2 BCB2 BCB2 BCB2 BCB4 BCB4 BCB4 BCB4 Depth (m) 51.7 167.1 50.0 170.0 49.8 250.2 50.0 254.3 Time 0:00 0:12 19:04 19:16 22:40 22:57 13:49 14:05 night night day day night night day day Species (# per m3) Acartia sp.<YV 0.00 0.00 0.00 0.00 0.00 0.00 18.71 0.00 A. longiremis 92.83 40.48 139.25 28.84 17.78 3.54 68.59 4.08 Calanus sp.<III 0.00 0.00 0.00 0.00 0.00 0.44 6.24 0.00 C. marshallae 65.23 32.89 19.89 43.25 13.33 3.10 0.00 0.00 N. plumchrus 0.00 0.00 0.00 0.00 0.00 0.00 6.24 3.27 Metridia sp.<TV 30.11 26.56 19.89 30.28 48.89 3.10 87.29 10.61 M. pacifica 0.00 1.26 0.00 0.00 22.22 0.89 0.00 0.00 0. similis 25.09 39.21 248.66 54.79 155.55 5.76 399.05 20.41 0. atlantica 7.53 6.32 49.73 14.42 40.00 2.66 93.53 9.80 P. parvus 524.39 223.90 805.67 148.50 866.62 109.36 1907.95 142.04 Pseudoc. spp. 122.94 94.87 343.16 90.83 142.21 9.30 193.29 24.49 E. pacifica 5.65 3.72 0.00 4.05 1.60 0.53 0.58 0.28 T. spinifera 0.20 2.13 0.00 0.36 0.07 0.03 0.00 0.00 E. furcilia 0.00 0.00 4.97 0.00 8.89 0.00 6.24 0.00 E. calyptopis 17.56 5.06 109.41 17.30 26.67 2.66 56.12 2.45 E. hamata 0.00 0.08 0.62 0.27 0.00 0.14 0.10 0.48 S. elegans 1.80 2.85 4.35 2.16 2.01 0.22 4.58 0.54 S. scrippsae 0.04 0.16 0.00 0.45 0.56 0.06 0.39 0.23 Oikopleura 17.56 15.18 94.49 37.49 146.66 5.76 106.00 6.53 Appendix B. Zooplankton data from the Barkley Canyon study 120 S t a t i o n BCC3 BCC3 BCC3 BCC3 BCD2 BCD2 BCD2 D e p t h ( m ) 50.2 250.0 49.8 250.0 126.9 50.0 155.1 T i m e 1:38 1:53 17:55 18:09 2:48 8:07 8:18 night night day day night day day S p e c i e s (# per m3) Acartia sp.<YV 0.00 0.00 0.00 10.31 0.00 0.00 0.00 A. longiremis 2.70 1.05 51.95 7.22 11.02 68.01 6.77 Calanus sp.<III 0.00 0.00 4.33 0.00 1.22 0.00 1.13 C. marshallae 30.99 3.67 4.33 5.16 40.40 32.38 55.29 N. plumchrus 0.00 0.00 0.00 0.00 3.67 6.48 0.00 Metridia sp.<YV 9.43 1.05 4.33 5.16 15.91 16.19 9.03 M. pacifica 5.39 2.10 0.00 1.03 0.00 0.00 1.13 0. similis 12.13 12.59 190.48 35.07 12.24 136.01 11.28 0. atlantica 2.70 3.67 8.66 4.13 3.67 0.00 10.15 P. parvus 265.48 118.01 532.47 106.24 166.49 576.44 130.88 Pseudocalanus spp. 12.13 6.82 380.96 47.45 50.19 236.40 82.37 E. pacifica 2.86 1.21 0.14 0.52 2.14 0.25 0.12 T. spinifera 0.00 0.00 0.00 0.00 0.46 0.00 0.02 Euphausia furcilia 0.00 0.52 4.33 0.00 1.22 0.00 0.00 Euphausia calyptopis 0.00 2.62 12.99 3.09 22.04 9.72 5.64 E. hamata 0.00 0.33 0.00 0.64 0.31 0.00 0.02 S. elegans 2.44 0.59 2.50 0.84 1.53 2.18 0.32 S. scrippsae 0.51 0.13 0.00 0.52 0.38 0.25 0.16 Oikopleura 0.00 2.62 95.24 19.63 14.69 55.05 0.00 Appendix B. Zooplankton data from the Barkley Canyon study 121 Station BCD4 BCD4 BCD4 BCD4 Depth (m) 49.8 250.0 51.6 250.5 Time 0:12 0:29 16:16 16:32 night night day day Species (# per m3) Acartia sp.<IV 0.00 0.00 0.00 0.00 A. longiremis 20.67 7.12 48.09 0.00 Calarms sp.<III 3.45 2.37 24.04 0.00 C. marshallae 6.89 10.67 8.01 0.94 N. plumchrus 0.00 2.37 0.00 2.81 Metridia sp.<TV 12.06 2.37 128.23 12.16 M. pacifica 3.45 14.23 0.00 1.87 0. similis 60.29 24.91 512.93 52.37 0. atlantica 18.95 7.12 56.10 17.77 P. parvus 540.93 269.21 1578.85 170.19 Pseudocalanus spp. 75.80 23.72 200.36 20.57 E. pacifica 1.27 1.07 1.31 0.50 T. spinifera 0.03 0.04 0.00 0.00 Euphausia furcilia 1.72 0.00 0.00 0.00 Euphausia calyptopis 3.45 3.56 64.12 9.35 E. hamata 0.03 0.78 0.13 1.14 S. elegans 0.62 0.93 2.07 0.94 S. scrippsae 0.13 0.37 0.19 0.29 Oikopleura 10.34 3.56 208.38 36.47 

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.831.1-0053340/manifest

Comment

Related Items