P R I M A R Y P R O D U C T I O N A N D T H E S E T T L I N G F L U X IN T W O F J O R D S OF BRITISH C O L U M B I A , C A N A D A by David Andrew Timothy M.Sc. Oceanography, U . B . C , 1994 B.Sc. Environmental Engineering, M.I.T., 1989 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF D O C T O R OF P H I L O S O P H Y in T H E FACULTY OF G R A D U A T E STUDIES D E P A R T M E N T OF EARTH A N D O C E A N SCIENCES We accept this thesis as conforming to the required standard T H E UNIVERSITY OF BRITISH C O L U M B I A July 2001 © David Andrew Timothy, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) Abst rac t A time series of primary production and sediment trap flux measurements was carried out in two fjords of British Columbia, Canada between 1983 and 1989. The fjords, pe-riodically anoxic Saanich Inlet and oxygen-replete Jervis Inlet, were chosen in order to compare organic matter formation and particle flux in these environments with largely differing redox conditions. Two sediment-trap moorings were deployed in each fjord, and each mooring had sediment traps at three depths. The moorings were serviced monthly, when primary production was also measured using the 1 4 C-uptake technique. Hydro-graphic and nutrient data were collected during portions of the experiment, and 2 1 0 P b profiling of bottom sediments allowed comparison of water-column fluxes and sedimen-tary accumulation rates. Saanich Inlet (490 g C m ~ 2 y - 1 ) was 1.7 times more productive than Jervis Inlet (290 g C m - 2 y _ 1 ) and primary production toward the mouths of both fjords was 1.4 times higher than at the heads of the fjords. The elevated rates of primary production in Saanich Inlet were probably due to exchange with the nutrient-rich surface waters of the passages leading to the Pacific Ocean, and the up-inlet gradients in both fjords reflected the relative nutrient supply. The sediment-trap material was dominated by bio-genic silica, especially in the spring and early summer but also in the late summer and fall, while organic carbon fluxes tended to peak in the summer. While winter fluxes were usually dominated by aluminosilicates, at the mouth of Jervis Inlet organic matter often comprised most of the mass flux to the 50 m sediment traps, as wintertime sources of biogenic silica and aluminosilicates were small. At the head of Saanich Inlet, the alumi-nosilicate flux closely followed the pattern of local rainfall and flow from the Cowichan ii River, a distinct difference from the other stations where turbulent resuspension from topographic boundaries and particle focusing appear to have dominated the lithogenic flux. 513C of the trapped material was heavier in the summer than in the winter, reflect-ing a higher ratio of marine to terrestrial organic matter at that time. The relationship between stable carbon isotope ratios and BSi content revealed that 70-80% of the marine O C in these fjords is diatomaceous. This relationship was furthermore used to estimate the <513C endmember of the marine organic matter and the proportion of terrigenous ma-terial to the total organic matter flux. Export ratios of organic carbon were low, likely because of solubilisation within the traps, while export ratios of biogenic silica were high. Sediment-trap fluxes at the mouth of Saanich Inlet were strongly affected by a sediment plume that extended off the nearby sill. However, compared to the other stations, this plume did not result in excess sedimentary accumulation of biogenic silica and organic carbon relative to local primary production. A t each station, similar proportions of local primary production (~5%) were buried in the sediments below, suggesting that the bulk of the marine organic matter was not preferentially preserved in the intensely anoxic sediments of Saanich Inlet. The possibility of organic matter solubilisation within the sediment traps, and the excessive water-column fluxes at the mouth of Saanich Inlet, confuse comparison of or-ganic carbon fluxes in Saanich and Jervis Inlets. However, away from the mouth of Saanich Inlet water-column fluxes of biogenic silica were proportional to local primary production. If the biogenic silica carried proportional amounts of organic matter, then the heightened primary production in Saanich Inlet resulted in a large delivery of organic matter to depth. Combined with the high primary production and export flux, low rates of vertical mixing and particle-entrapment within the fjord, factors associated with the weak estuarine circulation and weak winds of Saanich Inlet, may have also stimulated iii anoxia. Although in Jervis Inlet there is more stagnant water behind the sill and deep-water renewals were less frequent than in Saanich Inlet, the deep sill allows oxidation of a significant fraction of the sinking organic matter before the stagnant waters are reached, reducing the chances of oxygen depletion in the bottom waters. A model that estimates rates of water-column decay from sediment-trap data showing increases in flux with depth was used with the time series from Saanich and Jervis Inlets. Model results from Saanich Inlet were not conclusive, possibly because the depth interval between sediment traps was too small to resolve water-column rates of decay. However, the model fit well to the time series from Jervis Inlet, and rate constants for organic carbon and nitrogen agree well with previous estimates made from oceanic settings. The model has also allowed some of the first estimates of depth-dependent dissolution rates of sinking biogenic silica, and translation to time-dependent dissolution using a nominal sinking rate suggests the diatomaceous opal in Jervis Inlet was dissolving rapidly. Changes with depth of rate constants for organic carbon, nitrogen and biogenic silica are well described by the power function, suggesting that organic matter and biogenic silica are composed of a set of multiple components that decay at varying rates. This model for decay has been explained for organic matter, and for the biogenic silica may be caused by the presence of various diatom species or degrees of frustule fragmentation that result in a number of fractions with different dissolution rates. The model has also allowed a description of the material that causes increases in flux with depth. This sediment was depleted of organic carbon and nitrogen and thus appeared diagenetically altered, and its aluminosilicate and biogenic silica contents were characteristic of hydrodynamically sorted resuspended material. Additional material was delivered to the deepest sediment traps during deepwater renewals, but a continual process such as tidal resuspension, particle focusing, or increases in trapping efficiency with depth resulted in additional fluxes to the mid-depth sediment traps. iv T a b l e o f C o n t e n t s A b s t r a c t i i L i s t o f Tab l e s i x L i s t o f F i g u r e s x i P r e f a c e x i v A c k n o w l e d g e m e n t s x v 1 P h y s i c a l , e c o l o g i c a l a n d e n v i r o n m e n t a l s e t t i n g 1 1.1 Introduction 1 1.2 Saanich and Jervis Inlets 3 1.2.1 Physical descriptions 3 1.2.2 Plankton ecology 12 1.3 Environmental setting 14 2 P r i m a r y p r o d u c t i o n i n S a a n i c h a n d J e r v i s In l e t s 2 1 2.1 Introduction 21 2.2 Methods 23 2.3 Results 25 2.3.1 Salinity and temperature 25 2.3.2 Nutrient concentrations and uptake 31 2.3.3 Estimates of primary production 34 v 2.4 Discussion 40 2.4.1 Temporal pattern of primary production 40 2.4.2 Geographic pattern of primary production 43 2.4.3 Bottom-water oxygen in Saanich and Jervis Inlets 48 2.5 Conclusions 49 3 S e t t l i n g fluxes i n S a a n i c h a n d J e r v i s In le t s 50 3.1 Introduction 50 3.2 Materials and methods 52 3.2.1 Field design and sample preparation 52 3.2.2 Laboratory analyses 54 3.2.3 Effect of preservative treatments 55 3.2.4 Linear regressions 57 3.3 Results 58 3.3.1 Components of the mass flux 58 3.3.2 Fluxes at the head of Saanich Inlet (station SN-0.8) 61 3.3.3 Fluxes at the mouth of Saanich Inlet (station SN-9) 71 3.3.4 Fluxes in Jervis Inlet 72 3.4 Discussion 82 3.4.1 O C - B S i relationships 82 3.4.2 Marine and terrigenous O C fluxes 84 3.4.3 Export ratios of O C and BSi 95 3.4.4 The riverine source of particulates to Saanich Inlet 99 3.4.5 Deep-trap, sediment-interface and burial fluxes 103 3.5 Conclusions 114 vi 4 A m o d e l t o i n t e r p r e t increases i n flux w i t h d e p t h 117 4.1 Introduction 117 4.2 Description and solution of the model 119 4.3 Results 123 4.3.1 Sensitivity analyses and examples of the planar fits 123 4.3.2 The error term 127 4.4 Discussion 130 4.4.1 Describing depth dependence of the rate constants 130 4.4.2 The anticipated flux 133 4.4.3 The additional flux in Jervis Inlet 135 4.4.4 The additional flux in Saanich Inlet 138 4.5 Conclusions 140 5 C o n c l u s i o n s 142 B i b l i o g r a p h y 147 A p p e n d i x A S e p a r a t i n g m a r i n e f r o m t e r r i g e n o u s o r g a n i c m a t t e r 169 A p p e n d i x B U s i n g a conse rva t i ve t r a c e r w i t h t h e t r a p m o d e l 172 B . l A n alternate model for when rate constants do not converge 172 B.2 The normalisation scheme 174 A p p e n d i x C S e n s i t i v i t y ana lys i s a n d resu l t s for J e r v i s I n l e t 175 A p p e n d i x D S e n s i t i v i t y ana lys i s a n d r e su l t s for S a a n i c h I n l e t 184 A p p e n d i x E E x p o n e n t i a l fit t o ra te cons tan t s 190 vii A p p e n d i x F T a b u l a t i o n o f p r i m a r y p r o d u c t i o n d a t a A p p e n d i x G S e d i m e n t - t r a p d a t a o f S a a n i c h In l e t A p p e n d i x H S e d i m e n t - t r a p d a t a o f J e r v i s In l e t viii Lis t of Tables 2.1 Primary production in Saanich and Jervis Inlets 37 2.2 Comparison of production in Saanich Inlet with Hobson's estimates . . . 38 3.1 Station locations 53 3.2 Preservation effects of N a N 3 and brine 56 3.3 Fluxes in Saanich Inlet 70 3.4 Fluxes in Jervis Inlet 81 3.5 Terrigenous O C in the inlets 90 3.6 Total and diatom production 91 3.7 Export ratios of O C and BSi at 50 m 96 3.8 Composition of upper sediments 104 3.9 Interface and burial fluxes 108 A . l Marine 5 1 3 C endmembers and the composition of marine samples . . . . 171 C l Model results from Jervis Inlet 183 D . l Model results from Saanich Inlet 189 F . l 1 4 C uptake at station SN-9 193 F.2 1 4 C uptake at station SN-0.8 194 F.3 1 4 C uptake at station JV-3 195 F . 4 1 4 C uptake at station J V - 7 196 G . l Sediment-trap fluxes measured at station SN-9: 45 m 198 G.2 Sediment-trap fluxes measured at station SN-9: 45 m (continued) . . . . 199 G.3 Sediment-trap fluxes measured at station SN-9: 110 m • . . 200 ix G.4 Sediment-trap fluxes measured at station SN-9: 110 m (continued) . . . . 201 G.5 Sediment-trap fluxes measured at station SN-9: 150 m 202 G.6 Sediment-trap fluxes measured at station SN-9: 150 m (continued) . . . . 203 G.7 Sediment-trap fluxes measured at station SN-0.8: 50 m 204 G.8 Sediment-trap fluxes measured at station SN-0.8: 50 m (continued) . . . 205 G.9 Sediment-trap fluxes measured at station SN-0.8: 135 m 206 G.10 Sediment-trap fluxes measured at station SN-0.8: 135 m (continued) . . . 207 G . l l Sediment-trap fluxes measured at station SN-0.8: 180 m 208 G . 12 Sediment-trap fluxes measured at station SN-0.8: 180 m (continued) . . . 209 H . l Sediment-trap fluxes measured at station JV-3: 50 m 211 H.2 Sediment-trap fluxes measured at station JV-3: 300 m 212 H.3 Sediment-trap fluxes measured at station JV-3: 600 m 213 H.4 Sediment-trap fluxes measured at station JV-7: 50 m 214 H.5 Sediment-trap fluxes measured at station JV-7: 200 m 215 H.6 Sediment-trap fluxes measured at station JV-7: 450 m 216 x List of Figures 1.1 Sampling stations in Saanich and Jervis Inlets 4 1.2 Longitudinal and transverse cross-sections of Saanich Inlet 5 1.3 Longitudinal and transverse cross-sections of Jervis Inlet 6 1.4 Temperature, salinity and dissolved oxygen at station SN-9 8 1.5 Temperature, salinity and dissolved oxygen at station SN-0.8 9 1.6 Temperature, salinity and dissolved oxygen at station JV-3 10 1.7 Temperature, salinity and dissolved oxygen at station J V - 7 11 1.8 Environmental and lighthouse stations 15 1.9 Southern Oscillation Index 15 1.10 Atmospheric temperature at weather stations 16 1.11 Precipitation at weather stations 17 1.12 Hours of sunshine at weather stations 17 1.13 River flow during the study 18 1.14 Surface salinity in the southern Strait of Georgia 19 2.1 Primary production at station SN-9 in Saanich Inlet 26 2.2 Primary production at station SN-0.8 in Saanich Inlet 27 2.3 Primary production at station JV-3 in Jervis Inlet 28 2.4 Primary production at station J V - 7 in Jervis Inlet 29 2.5 Vertical profiles of T , S and nitrate 30 2.6 N:P and Si:N uptake ratios 33 2.7 Vertical profiles of primary production 35 xi 2.8 Averages of daily primary production 36 2.9 Nitrate versus phosphate in both inlets 46 3.1 Organic carbon to nitrogen ratios 59 3.2 Aluminium to lithogenic ratios 60 3.3 Total mass fluxes in Saanich Inlet 62 3.4 Biogenic silica fluxes in Saanich Inlet 63 3.5 Organic carbon fluxes in Saanich Inlet 64 3.6 Aluminium fluxes in Saanich Inlet 67 3.7 Seasonally-averaged sediment-trap fluxes in Saanich Inlet 68 3.8 Seasonally-averaged composition of settling fluxes in Saanich Inlet . . . . 69 3.9 Total mass fluxes in Jervis Inlet 73 3.10 Biogenic silica fluxes in Jervis Inlet 74 3.11 Organic carbon fluxes in Jervis Inlet 76 3.12 Aluminium fluxes in Jervis Inlet 78 3.13 Seasonally-averaged sediment-trap fluxes in Jervis Inlet 79 3.14 Seasonally-averaged composition of settling fluxes in Jervis Inlet 80 3.15 Production and O C flux residuals at station J V - 7 82 3.16 O C versus BSi in both inlets 83 3.17 5 1 3 C of the trapped organic matter of Saanich Inlet 86 3.18 <513C of the trapped organic matter of Jervis Inlet 87 3.19 5l3C versus %BSi in Saanich and Jervis Inlets 89 3.20 <513C versus O C / N in Saanich and Jervis Inlets 93 3.21 Source of A l to station SN-0.8 101 3.22 The lithogenic A l 2 0 3 : S i 0 2 weight ratio 105 3.23 Summary of O C fluxes 110 xii 3.24 Summary of BSi fluxes I l l 4.1 Diagram of the sediment-trap model 120 4.2 Plot of O C solution: station JV-3 , 300-600 m 125 4.3 Plot of BSi solution: station JV-7 , 50-450 m 126 4.4 Rate constants versus depth 129 4.5 Flux versus depth 132 4.6 Composition of the additional flux in Jervis Inlet 135 4.7 The additional flux in Jervis Inlet 137 4.8 The additional flux in Saanich Inlet 139 C l Sensitivity analysis for station JV-3 , 50-300 m 176 C.2 Sensitivity analysis for station JV-3 , 50-600 m 177 C.3 Sensitivity analysis for station JV-3 , 300-600 m 178 C.4 Sensitivity analysis for station JV-7 , 50-200 m 179 C.5 Sensitivity analysis for station JV-7 , 50-450 m 180 C.6 Sensitivity analysis for station JV-7 , 200-450 m 181 C . 7 Data removed from analyses in Jervis Inlet 182 D. l Sensitivity analysis for station SN-9, 45-110 m 186 D.2 Sensitivity analysis for station SN-9, 45-150 m 186 D.3 Sensitivity analysis for station SN-9, 110-150 m 187 D.4 Sensitivity analysis for station SN-0.8, 50-135 m 187 D.5 Sensitivity analysis for station SN-0.8, 50-180 m 188 D.6 Sensitivity analysis for station SN-0.8, 135-180 m 188 xiii Preface The results of the primary production time series have been published (Timothy and Soon, 2001); Chapter 2 is a slightly modified version to accommodate the structure of this dissertation. The descriptive portion of the sediment trap experiment (the majority of Chapter 3) will be submitted to Progress in Oceanography. The principles of the sediment-trap model used in Chapter 4 were first described by Timothy (1994) and a version of the model was published by Timothy and Pond (1997). The results of this model applied to the time series from Jervis Inlet will be submitted to the Journal of Geophysical Research or Global Biogeochemical Cycles. xiv Acknowledgements From the beginning of this project, my supervisor, Steve Calvert, has been extremely supportive intellectually and personally, and through many surprising circumstances he somehow maintained his calm. I am grateful to Steve and committee member Ken Den-man for quickly commenting on chapters of the dissertation. The rest of my committee, Paul Harrison and C.S. Wong, were always supportive and Paul gave very good advice at various stages. I am very grateful to Steve Pond who encouraged me to think critically in the early stages of my work. Collecting the data presented here was truly a Herculean endeavor, tackled almost entirely by Maureen Soon. I can't thank her enough for her commitment, and for the constructive and pleasant environment she created in the lab. R a m and Hugh, and the officers and crew of the C.S.S. Vector, provided excellent help and good humour during the field program, so P m told! Bente Nielsen, Christine Elliott, Kathy Gordon and Bert Mueller were also very helpful with laboratory analyses. I am thankful to everybody in the Annex for their help and patience. Roger Pieters, Rich Pawlowicz, Susan Allen and Lionel Pandolfo have answered questions and allowed me to work on their computers. Tom Pedersen was always available for questions and advice, and Kristin Orians helped establish methods on the I C P - M S . Yuval gave very efficient lessons in matlab particular to the modelling - thanks! Don Murray and Jesse Hoey pointed me to the solution I used in the model. Stan Stubbe at Environment Canada was helpful in providing data. Discussions with Stephanie and Markus Kienast were helpful, especially in the last days of writing. Thanks to Greg Cowie, Lou Hobson and Graham Shimmield for providing data. A n d thanks to Roger Frangois for his early involvement in the project. xv Very rarely, but every once in a while, I was not in the lab or in front of the computer. Thanks to the many friends who have made my experiences in British Columbia fantastic! I am grateful to my family for their support while I've been so far away and to Wiliam for ... refraining from tearing me limb from limb with his terrible sharp teeth and vicious, slashing claws! xvi C h a p t e r 1 P h y s i c a l , e c o l o g i c a l a n d e n v i r o n m e n t a l s e t t i n g 1.1 I n t r o d u c t i o n A n improved understanding of primary production and particle flux in the ocean re-quires the construction of reliable budgets of biogenous elements. Indeed, much work has focussed on descriptions of carbon and nutrient cycling for both open ocean (Bishop, 1989; Siegenthaler and Sarmiento, 1993; Michaels et al., 1994; Emerson et al., 1995) and coastal regions (Burrell, 1988; Wassmann, 1991; Smith and Hollibaugh, 1993; L i u et al., 2000). Budgets for carbon are of special interest because of the role that the oceans play in modulating atmospheric CO2 concentration (Siegenthaler and Sarmiento, 1993). However, descriptions for coastal regions are complicated because at the coastal bound-ary continental inputs (Berner, 1989; Smith and Hollibaugh, 1993; L i u et al., 2000) and turbulent processes causing sediment resuspension and redistribution (Hakanson et al., 1989) are significant but difficult to quantify. The coastlines are furthermore marked by their heterogeneity, so that generalisations of biogeochemical processes are difficult to make (Liu et al., 2000). Nevertheless, a description of the fluxes occurring at the oceanic boundary is crucial to our understanding of carbon and nutrient cycling on a global scale, as the ocean margins account for ~20% of global ocean primary production (Walsh, 1988; L iu et al., 2000) and most ofthe organic carbon burial in marine sediments (Berner, 1982; Hedges and Keil , 1995; Middleburg, 1997). Large, fast-sinking particles make up a small fraction of the particulate matter in sea 1 Chapter 1. Physical, ecological and environmental setting 2 water, but they are the primary vehicle by which material is transported through the water column (McCave, 1975). Therefore, an important component of our understanding of element cycling is a description of large particle sinking and decay. Combined with knowledge of the export flux of material from surface waters (Dugdale and Goering, 1967; Eppley and Petersen, 1979; Boyd and Newton, 1995), decay rates can then be used to infer, for example, the vertical distribution of oxygen and nutrients (e.g.; Suess, 1980; Martin et al., 1987) and the effectiveness of carbon delivery to the deep ocean and sediments (Suess and Miiller, 1980; Hargrave et al., 1994; Hedges, 1992). To this end, sediment-trap fluxes are often appropriate for estimates of vertical flux and particulate decay (e.g.; Bishop, 1989), but in coastal waters a number of processes can contribute excess material to deep sediment traps rendering the information they collected difficult to interpret. Fjords are semi-enclosed, overdeepened coastal features that have long been recog-nised as accessible locations where biological and geochemical processes relevant to oceanic systems can be studied (e.g.; Skei, 1983). Studies in fjords have greatly contributed to our understanding of the ecology and physiology of phytoplankton (e.g.; Sakshaug and Myklestad, 1973; Takahashi et a l , 1977, 1978; Hobson, 1981; Parsons et al., 1983; Sak-shaug and Olsen, 1986; McQuoid and Hobson, 1995; Kukert and Riebesell, 1998) and zooplankton (Huntley and Hobson, 1978; Takahashi and Hoskins, 1978; Koeller et al., 1979; Dagg et al., 1989; Buck and Newton, 1995), and the factors that control plankton community development have been tested using the environment of a fjordal basin (Ross et al., 1993). Links between the plankton community and sedimentation have been inves-tigated in a number of fjords (e.g.; Wassmann, 1984; Burrell, 1988; Sancetta and Calvert, 1988; Sancetta, 1989c; Wassmann, 1991; Overnell et al., 1995), and models of diatom aggregation have here been explored (Kiorboe et al., 1994, 1996; Hansen and Kiorboe, 1995). Measurements of biochemical markers for marine and terrigenous organic matter Chapter 1. Physical, ecological and environmental setting 3 colecting on the continental margins have been made in this environment (Hedges et al., 1988a, 1988b; Cowie and Hedges, 1992) and, as analogs of anoxic basins, the redox chemistry within fjords has been studied in great detail (e.g.; Richards, 1965; Emerson, 1980; Skei, 1983; Smethie, 1987; McKee and Skei, 1999). This thesis reports a multi-year time series of primary production and sediment-trap flux colected during the 1980s in Saanich and Jervis Inlets, two fjords of southern British Columbia, Canada. In this chapter, I provide physical and ecological descriptions of the fjords and show environmental time series from the 1980s, providing a backdrop for sub-sequent chapters. Chapter 2 reports the primary production results and discusses causes of anoxia in Saanich Inlet. Chapter 3 presents the sediment-trap data, discusses the relationship between primary production and water-column flux and compares sediment-trap fluxes with botom sediment accumulation rates. In chapter 4, a balance equation is used to determine rates of water-column remineralisation of the biogenous fluxes and describes the causes of observed increases in flux with depth in both fjords. Chapter 5 provides a summary of conclusions. 1.2 Saanich and Jervis Inlets 1.2.1 Phys i ca l descriptions Saanich and Jervis Inlets of southern British Columbia (Figures 1.1 through 1.3) are contrasting fjords contiguous with the southern Strait of Georgia. Jervis Inlet is a long (89 km), deep-siled (240 m) fjord with a maximum depth of 730 m. Moderate fresh-water input at the head of Jervis Inlet (annual mean of 180 m3 s_1; Pickard, 1961) drives weak estuarine circulation with a net movement of surface waters seaward and replacement flow at depth (Pickard, 1961; Lazier, 1963). Jervis Inlet is within the coastal western hemlock (CWH) biogeoclimatic zone that extends throughout most of coastal BC (Pojar et al., Chapter 1. Physical, ecological and environmental setting 4 128°W 126°W 124°W 122°W 24.00' Figure 1.1: Sampling stations in Saanich and Jervis Inlets. The Cowichan River flows into Cowichan Bay. Chapter 1. Physical, ecological and environmental setting SN-0.8 SN-9 0 m 100 200 300 SN-9 SN-0.8 1.0 km 1.5 Figure 1.2: Longitudinal (vertical exaggeration = 50) and transverse (to scale) cross-sections of Saanich Inlet showing mooring locations and sediment-trap depths. Sill and maximum depths of Saanich Inlet are 70 m and 235 m, respectively. 1991). The C W H zone is characterised as cool and mountainous with heavy precipitation in the fall and winter (Figure 1.11). Anoxia is not known to occur in Jervis Inlet (Lazier, 1963; Smethie, 1987; Figures 1.6 and 1.7). Located on southern Vancouver Island, Saanich Inlet (Figures 1.1 and 1.2) is 24 km long, has a maximum depth of 235 m and a broad, shallow (70-80 m) sill that runs the length of the eastern branch of Satellite Channel. Saanich Inlet is in the coastal Douglas-fir ( C D F ) zone, a low-elevation region along the southeastern coast of Vancouver Chapter 1. Physical, ecological and environmental setting 6 Figure 1.3: Longitudinal (vertical exaggeration = 50) and transverse (to scale) cross-sections of Jervis Inlet showing mooring locations and sediment-trap depths. The mooring at the mouth of Jervis Inlet was moved from JV-11.5 to JV-3 early in the time-series (see Table 3.1 dates of deployment). Sil and maximum depths of Jervis Inlet are 240 m and 730 m, respectively. This sill depth is significantly shalower than has been reported by others, including Timothy and Soon (2001), who used a value reported by Pickard (1961). Pickard's value of 385 m refers to the depth of the region where sur-face water properties are no longer afected by the outflow from Jervis Inlet, a location somewhat inside the morphological sill. Chapter 1. Physical, ecological and environmental setting 7 Island and including many islands within the southern Strait of Georgia (Nuszdorfer et al., 1991). The C D F zone is within the rainshadow of Vancouver Island and the Olympic mountains, and is drier and warmer than the C W H zone surrounding Jervis Inlet. Runoff from the Goldstream River (Figure 1.13) at the head of Saanich Inlet is very small (annual mean < 2 m 3 s - 1 ; Stucchi and Whitney, 1997) and the largest sources of fresh water, the Cowichan and Fraser Rivers (Figure 1.1), are located seaward of the fjord (Herlinveaux, 1962). Precipitation in coastal southern British Columbia tends to be highest in the fall and winter. Thus, runoff from the Cowichan River peaks in the winter while that of the Fraser River, the predominant source of freshwater to the Strait of Georgia, is highest, in June when the melt from ice and snow peaks. These external sources of freshwater cause irregular surface currents and weak or absent estuarine flow within Saanich Inlet (Herlinveaux, 1962; Stucchi and Whitney, 1997). A characteristic relatively unique among British Columbia fjords is the periodic development of anoxia in the bottom waters of Saanich Inlet (Richards, 1965; Anderson and Devol, 1978; Cowie et al., 1992; Figures 1.4 and 1.5); hydrogen sulphide periodically extends to 50-80 m off the bottom in the central basin of Saanich Inlet (Richards, 1965). Generally, neither inlet exhibits an upper mixed layer. Instead, the steepest den-sity gradient occurs at the surface in Saanich Inlet (Herlinveaux, 1962) and Jervis Inlet (Pickard, 1961) and is larger in Jervis Inlet (Sancetta, 1989a). Deepwater renewal to both fjords occurs during the summer or fall when upwelling along the coastal rim of the northeast Pacific Ocean brings dense, nutrient-rich waters through Juan de Fuca Strait and into the Strait of Georgia. Deepwater renewal appears to occur most years in Saanich Inlet (Anderson and Devol, 1973; Figures 1.4 and 1.5), but renewals occur every several years in Jervis Inlet (Lazier, 1963; Pickard, 1975; Figures 1.6 and 1.7). Chapter 1. Physical, ecological and environmental setting 8 Figure 1.4: Temperature, salinity and dissolved oxygen at station SN-9. Greater detail of the temperature and salinity structure of the upper 40 m is given in Figure 2.1. Chapter 1. Physical, ecological and environmental setting 9 Temperature f°C) I I I I I L 1 9 8 5 1 9 8 6 1 9 8 7 1 9 8 8 1 9 8 9 Figure 1.5: Temperature, salinity and dissolved oxygen at station SN-0.8. Zero dissolved oxygen concentrations were measured in waters inside the 1 pM contour. Greater detail of the temperature and salinity structure of the upper 40 m is given in Figure 2.2. Chapter 1. Physical, ecological and environmental setting 10 i i i i i i i i i i i—i— i—i—i— i— i—i—i—i—i— i— i—>—i— 1985 1986 1987 1988 1989 Figure 1.6: Temperature, salinity and dissolved oxygen at station JV-3. Greater detail of the temperature and salinity structure of the upper 40 m is given in Figure 2.3. Chapter 1. Physical, ecological and environmental setting 11 Figure 1.7: Temperature, salinity and dissolved oxygen at station JV-7 . Greater detail of the temperature and salinity structure of the upper 40 m is given in Figure 2.4. Chapter 1. Physical, ecological and environmental setting 12 1.2.2 P l a n k t o n ecology Harrison et al. (1983) provide a thorough review of plankton ecology for the Strait of Georgia and contiguous waters, for which reports from Saanich Inlet are significant. From what is known about plankton dynamics in Jervis Inlet (Stockner and Cliff, 1975; Parsons et al., 1984b; Cochlan et al., 1986; Sancetta, 1989a, 1989b, 1989c), they appear similar to the descriptions of Harrison et al. (1983). Further insight into Jervis Inlet is provided by Haigh et al. (1992) and Taylor et al. (1994) who present a two-year study of phytoplankton ecology in a connected fjord, Sechelt Inlet (Figure 1.1). Similarly, a one-year study of diatom populations outside but near Saanich Inlet (Hobson and McQuoid, 1997) is relevant. The phytoplankton succession typical of coastal, temperate seas (Margarlef, 1958; Guillard and Kilham, 1977) is observed in the southern Strait of Georgia (Harrison et al., 1983; Haigh et al., 1992) with some exceptions in Saanich and Jervis Inlets (Sancetta, 1989a). Thus, diatoms contribute most of the yearly primary production with occasion-ally significant growth of dinoflagellates and nanoflagellates, while diatom-predominance is greater in Saanich Inlet than in Jervis Inlet (Sancetta, 1989a). During much of the fall and winter in Saanich Inlet, light limits photosynthesis (Takahashi et al., 1978) and nanoflagellates are the predominant phytoplankton (Takahashi et al., 1978; Smith and Hobson, 1994). Although winter conditions may be similar in Jervis Inlet, sampling has been less frequent there. In and near both inlets, the productive season begins around Apri l with the blooming of Thalassiosira spp., while blooms of Skeletonema costatum and small Chaetoceros species occur concomitantly or shortly thereafter (Stockner and Cliff, 1975; Takahashi et al., 1977; Hobson, 1981, 1983; Sancetta, 1989a, 1989b, 1989c; Haigh et al., 1992; Hobson and McQuoid, 1997). Through frequent sampling, Takahashi et al. (1977) found the spring bloom in Saanich Inlet to last for about two weeks and, Chapter 1. Physical, ecological and environmental setting 13 throughout the summer and fall, various other diatom and flagellate blooms occured. They showed that these blooms were always preceded by high surface nitrate and silicic acid concentrations and later Parsons et al. (1983) suggested that they were caused by increased mixing across the broad sill during spring tides leading to greater influxes of nutrients. Hobson (1985) found periods of the summer when nutrients were severely depleted in Saanich Inlet and flagellates predominated. Similar flagellate-predominance was observed during nutrient-deplete, summertime conditons in Sechelt Inlet (Haigh et al., 1992; Taylor et al., 1994), a branch of the Jervis Inlet system (Figure 1.1). The zooplankton of these waters can have a significant impact on phytoplankton dynamics. Their populations sometimes comprise a large amount of surface layer biomass and can lead to mass export of the phytoplankton population (Takahashi and Hoskins, 1978; Koeler et al., 1979; Harison et al., 1983; Sanceta, 1989a, 1989b, 1989c). Calanoid copepods such as Pseudocalanus spp., Calanus pacificus and Neocalanus plumchrus have the largest efect on the phytoplankton of waters contiguous with the Strait of Georgia (Harison et al., 1983). Their greatest grazing pressure occurs at or shortly after the spring bloom, when Calanus pacificus and Neocalanus plumchrus have returned to surface waters from depth in their ontogenic cycle. In the past decade, it appears that the timing of this cycle has shifted to about one month earlier than historicaly reported, but when this shift occured relative to the 1980s when this experiment occured is unclear (Bornhold, 2000). Medusoid zooplankton have been reported to penetrate into Saanich Inlet and significantly impact phytoplankton nutrient dynamics (Huntley and Hobson, 1978). Chapter 1. Physical, ecological and environmental setting 14 1.3 Environmental seting The time series reported in this thesis began in 1983 and ended in 1989. Environmental data from this period are presented in Figures 1.9 through 1.14 in order to put the primary production and sediment-trap time series in context. A weak El Nino occured in 1986/87, and a larger El Nino occured in 1982/83 shortly before the time series began (Figure 1.9). There does not appear to have been a strong atmospheric response to these El Nino events; however, the Fraser River freshet was perturbed from the decadal average in 1982 and 1986. Where available, the environmental data go back to 1980 in order to include the earlier El Nino and to test for teleconnections between the Equatorial Pacific and southern BC. Temperature and precipitation data (Figures 1.10 and 1.11) were colected at Victo-ria Airport, Mery Island and Malibu Rapids (Figure 1.8), and sunshine measurements (Figure 1.12) were made at Victoria Airport and Mery Island (Environment Canada). The sunshine data were colected using Campbel-Stokes Sunshine Recorders to measure cloud opacity. These instruments record the number of hours per day that light intensity is above a threshold value and are reported in units of 'actual sunlight hours', ranging be-tween zero and the number of hours between sunrise and sunset on a given day. Eforts to convert 'actual sunlight hours' into iradiance (pE/m2/s) or photosyntheticaly available radiation (PAR; pE/m2/s of wavelengths between ~400-700 nm) are generaly unsuc-cessful because the threshold value, determined as the ability to burn chemicaly coated cardboard, is afected by atmospheric humidity. Nevertheless, Figure 1.12 provides a subjective record of seasonal variability in the amount of sunshine reaching Saanich Inlet and a region near the mouth of Jervis Inlet. Sunshine at Victoria Airport and Mery Island were similar throughout the 1980s, but, because precipitation was significantly Chapter 1. Physical, ecological and environmental setting 15 Figure 1.8: Environmental (diamonds) and lighthouse (circles) stations. Temperature, precipitation and sunshine (Figures 1.10 through 1.12) were colected daily at the envi-ronmental stations; sea surface salinity (Figure 1.14) was obtained daily at the lighthouse stations. i i 80 81 82 83 84 85 86 87 88 89 Figure 1.9: The Southern Oscilation Index. The SOI is the atmospheric sea level pressure anomaly between Darwin, Australia, and Tahiti. El Nino events occur when the pressure anomaly is negative, La Ninas when it is positive. (Source: htp: /www.cgd.ucar.edu / cas / catalog/climind / soi.html.) Chapter 1. Physical, ecological and environmental setting 16 80 81 82 83 84 85 86 87 88 Figure 1.10: Atmospheric temperature at weather stations near Saanich and Jervis Inlets. Dashed lines are the decadal means. higher at Malibu Rapids (Figure 1.11) where sunshine measurements are not made, ir-radiance was certainly lower towards the head of Jervis Inlet. There is no clear evidence of an E l Nino effect on temperature, precipitation or sunshine in the southern Strait of Georgia. The Cowichan and Fraser Rivers are the major sources of fresh water to the region of Saanich Inlet, and both are continually monitored by Environment Canada (Figure 1.13). The Goldstream River at the head of Saanich Inlet has a smaller effect on Saanich Inlet due to its low flow, and is not continually monitored. However, data from 1977 and 1978 exist and are presented in Figure 1.13. The Skwawka and Hunaechin Rivers flow into the head of Jervis Inlet, but Environment Canada has no record of their flow. Adjacent to and east of the Skwawka and Hunaechin watersheds is the the larger, glaciated watershed of the Elaho River. The Elaho is a monitored river, and the record is given in Figure 1.13 as a proxy for flow into the head of Jervis Inlet. Although the Jervis Inlet watershed has only minor glaciers, it is influenced by the spring freshet (Lazier, 1963; Pickard, Chapter 1. Physical, ecological and environmental setting 17 Figure 1.12: Hours of sunshine at weather stations. Campbel-Stokes Sunshine Recorders were used to measure the number of hours per day that light intensity was above a threshold value (cloud opacity). Dashed lines are the decadal means. Chapter 1. Physical, ecological and environmental setting 18 80 81 82 83 84 85 86 87 88 89 Figure 1.13: River flow from the Fraser, Elaho, Cowichan and Goldstream Rivers. The Elaho is included as an analog for the Skwawka and Hunaechin Rivers (un-metered) that flow into the head of Jervis Inlet. Monthly flows (heavy line) are given, as are the averages for 1980-1989 (dashed lines). The Goldstream River drains into the head of Saanich Inlet and was metered during 1977 and 1978. Average flow for those years is shown. Chapter 1. Physical, ecological and environmental setting 19 Figure 1.14: Surface salinity measured daily at lighthouses throughout the southern Strait of Georgia. Monthly averages are ploted, and the dashed lines are the decadal means. Chapter 1. Physical, ecological and environmental setting 20 1961) due to melting snow from surounding mountain peaks reaching 2000 m elevation. Therefore, the seasonal patern (dominated by the freshet) and interannual variability may have been similar for the two watersheds. In the summers of 1982 and 1986, the peak of the Fraser River freshet was anomalously high, and in 1982 the Elaho freshet was large. In chapter 2 it is postulated that the 1986 freshet may have afected primary production in Saanich and Jervis Inlets, but it is unlikely the atypical riverflows of the summers of 1982 and 1986 were related to El Nino events. The freshet is the result of melting of snow and ice and therefore is controled by precipitation of the preceding winter and by spring and summer temperatures. Because the large freshets occured as the 1982/83 and 1986/87 El Nifios were beginning, it is doubtful that the 1982 and 1986 freshets were somehow a response to El Nino. The majority of the fresh water reaching the Strait of Georgia is delivered by the Fraser River, and the summer freshet results in a surface brackish layer of 5-10 m thickness extending over the southern and central portions of the Strait of Georgia (LeBlond et al., 1994). Surface salinity colected daily at lighthouse stations throughout the southern Strait of Georgia and Juan de Fuca (Figures 1.8 and 1.14) show the efect of the Fraser River freshet as low salinities around June and July. The large freshet of 1982 caused low surface salinities at lighthouse stations throughout the southern Strait of Georgia, while efects of the 1986 freshet on surface salinity are less evident (Figure 1.14). At Race Rocks in Juan de Fuca, high summertime salinities (Figure 1.14) are caused by wind-driven upweling and greater mixing of intermediate Pacific waters due to enhanced estuarine exchange during the Fraser River freshet. Chapter 2 P r i m a r y produc t ion i n Saanich and Jervis Inlets: what causes high product ion i n Saanich Inlet? 2.1 In t roduct ion The main processes of oxygen consumption in aquatic basins are by the heterotrophic ox-idation of organic mater and through bacterialy mediated oxidation of reduced chemical species (e.g.; HS~, N H 4 , CH4, Fe2+ and Mn2+) that accumulate due to organic organic mater degradation in low-oxygen regions and then mix or difuse into oxygen-replete wa-ters. The advective supply of oxygen to the deep waters of siled fjords such as Saanich and Jervis Inlets is limited to periodic renewal events so that, where oxygen consumption exceeds the difusive supply and the period between renewals is suficiently long, anoxia can result. Of the many deep- and shalow-siled fjords of BC, only a few are known to develop severe anoxia and, compared with these, renewal of the botom waters in Saanich Inlet (Anderson and Devol, 1973) is regular and relatively vigorous. Narows and Princess Louisa Inlets (Figure 1.1), smaler fjords within the Jervis Inlet system, are separated from the Strait of Georgia by multiple sils and deep-water renewal tends to be irregular and weak (Lazier, 1963; Pickard, 1975). In Nitinat Lake, a true fjord on the southwest coast of Vancouver Island where the waters are permanently anoxic below ~30 m, renewal to depth is severely restricted by a very shalow and protracted sill (Northcote et al., 1964; Richards, 1965). Efingham Inlet, located on the central west coast of Vancouver Island, 21 Chapter 2. Primary production in Saanich and Jervis Inlets 22 has only recently been discovered to be periodicaly anoxic (e.g.; Baumgartner et al., in review) and the dynamics of deep-water renewal have yet to be described. However, mixing along the approaching channel of approximately 25 km length and a set of sils leading to the inner basin of Efingham Inlet reduce the potential for deep-water renewal. Although verification is needed that Muchalat Inlet on the west coast of Vancouver Island develops strong botom-water anoxia (Pickard, 1963), a long channel and several sils lead to this fjord. The characteristics of deep-water renewal and advective oxygen supply, therefore, distinguishes Saanich Inlet from other anoxic BC fjords. Local primary production is an important factor that can lead to anoxic botom wa-ters (Richards, 1965; Calvert and Pedersen, 1992) and, indeed, conditons in Saanich Inlet are uniquely suited for phytoplankton growth. Surface stratification and relatively low vertical mixing due to weak winds, tidal curents and estuarine circulation (Stucchi and Whitney, 1997) create stable conditons for phytoplankton growth, and surface nutrient concentrations outside the sill are high year-round (Lewis, 1978; Mackas and Harrison, 1997). Intrusions of nutrient-rich, surface waters into Saanich Inlet were seen to cause phytoplankton blooms (Takahashi et al., 1977) in phase with the fortnightly, spring-neap tidal cycle (Parsons et al., 1983), and Hobson and McQuoid (in press) observed that increases in phytoplankton biomass in Saanich Inlet were related to tidaly-modulated nutrient intrusions. Herlinveaux (1962) has postulated that benthic and pelagic phyto-plankton are ultimately responsible for deep-water anoxia in Saanich Inlet, while Hobson's (1983) test of the relationship between phytoplankton biomass, bacterial metabolism and deep-water oxygen content was complicated by deep-water replacement during that study. This chapter presents the primary production time-series from Saanich and Jervis Inlets. Primary production was significantly higher in Saanich Inlet than in Jervis Inlet or the majority of the Strait of Georgia, supporting the possibilty that a large delivery of organic mater to the deep waters is partly responsible for the anoxia in Saanich Inlet. Chapter 2. Primary production in Saanich and Jervis Inlets 23 Also considered are the possibilities that weak estuarine circulation, leading to particle-retention within the fjord, and low rates of vertical mixing in Saanich Inlet stimulate deep-water anoxia. 2.2 Methods From August, 1985 to October, 1989, primary production was measured at each station (SN-9 and SN-0.8 in Saanich Inlet; JV-11.5 or JV-3 , and J V - 7 in Jervis Inlet) when the sediment-trap moorings were serviced. O n station, subsurface light was determined using a L I - C O R 185B quantum meter. Water samples were collected from depths corresponding to 56, 32, 18, 11 and 7% surface irradiance and carbon fixation was determined by the uptake of 1 4 C following the method and equation, including dark-bottle subtraction, of Parsons et al. (1984a). Unscreened seawater was transferred to two 125 m L borosilicate bottles and about 5 pCi of N a H 1 4 C 0 3 were added. One bottle from each depth was wrapped with neutral density screening to mimic in situ irradiance and the dark bottle was wrapped with electrical tape. Both were placed in a Plexiglas incubator thermally regulated by flowing surface seawater. After approximately 2 h, cells were collected by filtration with applied pressure < 12 cm Hg onto Millipore H A filters, rinsed with filtered seawater and placed in Aquasol II for analysis of radioactivity by scintillation spectrometry. The incubations may have occurred at any time during the day, depending on when stations were visited. The hourly rates of 1 4 C-uptake thus obtained can be converted to daily rates of primary production by assuming that carbon assimilation is proportional to photosyn-thetically available radiation (PAR) and scaling the incubation results by the ratio of total, daily P A R to P A R integrated throughout the incubation (e.g., Perry et al., 1989; Clifford et al., 1992). P A R was not measured during the experiment, so it was assumed Chapter 2. Primary production in Saanich and Jervis Inlets 24 that P A R throughout the day follows the first-order sine function. The incubation results were then appropriately scaled knowing the times of the beginning and end of each incu-bation relative to sunrise and sunset. Piatt et al. (1990) showed that the first-order sine function provides a good description of the curve for daily irradiance when daylength is < 20 h; at the latitude of southern British Columbia, maximum daylength is approximately 16 h. One assumption of this conversion is that the fraction of P A R attenuated by clouds during the incubation was similar to the attenuation throughout the day. Although this assumption may lead to errors in the primary production estimates for any given day, it should not result in systematically positive or negative errors. If cloud attenuation is constant throughout the day, this sine conversion will accurately estimate the ratio of daily P A R to incubation-period P A R . Areal estimates of primary production were obtained by trapezoidal integration over depth of the carbon assimilation profiles. During this procedure, the measurement of 1 4 C uptake at 56% surface irradiance was used as the rate of carbon assimilation from that depth to the surface. Also, rates of 1 4 C-uptake were extrapolated from 7 to 1% sur-face irradiance by assuming that the rates measured at 7% surface irradiance decreased with depth proportionally to light. The extinction coefficient for light was approximated as the slope of the relationship between In (PAR) and depth as measured prior to each incubation. These shallow and deep extrapolations were performed to account for the carbon assimilation that likely was occurring above the depth of 56% surface irradiance and below the depth of 7% surface irradiance. O n average, the shallow and deep extrap-olations are 38 and 12%, respectively, of the primary-production estimates integrated between the depths for which 1 4 C-uptake was measured (from 56 to 7% surface irradi-ance). Deep chlorophyll maxima between the depths of 7 and 1% surface irradiance are common in these waters, but they should not have caused the deep extrapolations to Chapter 2. Primary production in Saanich and Jervis Inlets 25 be large under-estimates because the primary production associated with them is typi-caly low (e.g., Cliford et al., 1991; Harison et al., 1991). The deep extrapolations may be over-estimates in winter, as Takahashi et al. (1978) have shown that the wintertime compensation depth in Saanich Inlet is less than the depth of 1% surface irradiance. Nev-ertheless, the deep extrapolation has been applied to the entire time-series to maintain consistency. For the data colected between November and February, the deep extrap-olations are on average 1% of the 14C-uptake profiles integrated from the surface to the depth of 7% surface irradiance. Because they are relatively low, the estimates of wintertime primary production have little efect on yearly estimates. Temperature and salinity were measured throughout the water column using a Guild-line 8705 CTD when each station was visited. Beginning January, 1988 and for the remaining two years of the program, samples for nutrient analyses (nitrate + nitrite, orthophosphate and silicic acid {Parsons et al., 1984a}) were colected. 2.3 Resul ts 2.3.1 Sal in i ty and temperature Contour plots of nitrate concentration, temperature and salinity are shown in Figures 2.1 through 2.4, and these data are compressed into average "summer" and "winter" profiles in Figure 2.5. The primary production results presented in Figures 2.1 through 2.4 are discussed in the next section. In Saanich Inlet, low surface salinities occured in the fall and winter (Figures 2.1, 2.2 and 2.5) when local precipitation (Figure 1.11) was highest. Low surface salinity anoma-lies recorded at one station were not always observed at the other (Figures 2.1 and 2.2), suggesting spatial heterogeneity of the surface water masses of Saanich Inlet, a feature that was likely caused by tidaly-modulated intrusions of water from the Cowichan and Chapter 2. Primary production in Saanich and Jervis Inlets 26 12 E 6 O I 1 0 S . 15 T3 2 0 25 30 0 „ 10 E. s 20 Q. (11 1 3 30 40 0 ~ 1 0 £ £ 20 a . " ° 3 0 40 0 ~ . 1 0 E , £ 20 a . 1 3 30 40 1985 1986 1987 1988 1989 Figure 2.1: Depth-integrated primary production at station SN-9 in Saanich Inlet (top panel) from August, 1985 to October, 1989. The doted line is the smoothed curve of daily averages from Figure 2.8, repeated annualy. The plots below show volumetric primary production (isopleths at 10, 100, 500, 1000 and 2000 mg C m~3 d-1), nitrate concentration (isopleths at 24, 12, 5 and 1 pM), temperature (isopleths at 8, 12 and 16°C) and salinity (highest isopleth at 30, the next at 29 and others decreasing by four). Points in the contour plots are sampling locations. Chapter 2. Primary production in Saanich and Jervis Inlets 27 1985 1986 1987 1988 1989 Figure 2.2: Depth-integrated primary production at station SN-0.8 in Saanich Inlet (top panel) from August, 1985 to October, 1989. The dotted line is the smoothed curve of Figure 2.8, repeated annually. The plots below show volumetric primary production (isopleths at 10, 100, 500, 1000 and 1500 mg C m - 3 d" 1), nitrate concentration (isopleths at 24, 12, 5 and 1 pM), temperature (isopleths at 8, 12 and 1 6 ° C ) and salinity (highest isopleth at 30, the next at 29 and others decreasing by four). Points in the contour plots are sampling locations. Chapter 2. Primary production in Saanich and Jervis Inlets 28 i i i i i i i i i i i i i i — i — i — i — i — i — i — 1985 1986 1987 1988 1989 Figure 2.3: Depth-integrated primary production at station JV-3 in Jervis Inlet (top panel) from August, 1985 to October, 1989. The doted line is the smoothed curve of Figure 2.8, repeated annualy. The plots below show volumetric primary production (isopleths at 1, 10, 100 and 500 mg C m~3 d-1), nitrate concentration (isopleths at 24, 12, 5 and 1 pM), temperature (isopleths at 8, 12 and 16°C) and salinity (highest isopleth at 29, others decreasing by two). Points in the contour plots are sampling locations. Chapter 2. Primary production in Saanich and Jervis Inlets 29 Figure 2.4: Depth-integrated primary production at station JV-7 in Jervis Inlet (top panel) from August, 1985 to October, 1989. The doted line is the smoothed curve of Figure 2.8, repeated annualy. The plots below show volumetric primary production (isopleths at 1, 10, 100 and 300 mg C m-3 d_1), nitrate concentration (isopleths at 24, 12, 5 and 1 pM), temperature (isopleths at 8, 12, 16 and 20°C) and salinity (highest isopleth at 29, others decreasing by four). Points in the contour plots are sampling locations. Chapter 2. Primary production in Saanich and Jervis Inlets 30 S N - 9 S N - 0 . 8 J V - 3 J V - 7 T (°C) S N 0 3 (u.M) 6 8 10 12 14 16 20 24 28 32 0 8 16 24 32 I • l o > I • l u - j • I 9 • • - " ^ 0 • 0 • om m l\ i \ — mo -- U 1 1 „, 1 1 1 1 i a , 1 , 1 ,Q 1. , 1 •6 «D \ l • \ -m \ -0 winter " • summer | * » . . . ^ ' 1 ^ b w \ -OB / I " O • / / om — m 1 1 1 1 zL ' 1 ~ > 0 • A mo — II • O " 1 1 mo w — mo — \\ mo w m 1 — • — • , u , 1 , 1 , 1 " mb mb w • 0 w -1 " •> 1 • — 1 • 1 1 1 m _ m , 1 , J , ^ 1 1 1 • O A -0 • X / ' / ° / - " 0 , • — / \ • 0 — • 6 1 , 1 , ^ 1 , 1 c c c ; • - 1 m m •--0 1 0 0 D 1 , 1 , 1 « w » 11 1 • _ 1 , k , 1 -m^Sb m '0 ~~m w ot 1 0 1 0 ~. -1 , a 1 • • (. c -* | • 1 - , 1 c9 ••— • ' _ 0 * 1 , 1 . 1 • -" " «D • w • 11 -«D 1 , 1 , 1 ,• 1 » b 0 • \ r 0 1 _ »--60 120 0 60 120 180 0 60 120 180 0 60 120 180 Q . CD T 3 Figure 2.5: Temperature, salinity and nitrate concentrations during the study. Each row of figures presents data from a different station. Profiles were created by averaging measurements collected from October to March ("winter"; open circles) and Apr i l to September ("summer"; solid circles). T and S were measured throughout the study, while nitrate was measured in 1988 and 1989 only. In Saanich Inlet, the deepest sampling depths were about 20 m from the bottom (165 m at SN-9 and 210 m at SN-0.8). Water depths are 660 m at JV-3 and 530 m at JV-7 , but water properties changed little below 200 m at these stations. Chapter 2. Primary production in Saanich and Jervis Inlets 31 Fraser Rivers (Herlinveaux, 1962). There was little discernible along-inlet salinity gradi-ent in Saanich Inlet as would be expected if typical estuarine circulation were occurring. In Jervis Inlet, decreases in surface salinity caused by fall, winter and spring precipitation were superimposed on a yearly cycle of surface salinity dominated by the freshet, causing lowest surface salinities in June and July (Figures 2.3, 2.4 and 2.5). Also, in contrast to Saanich Inlet, the surface salinity patern in Jervis Inlet was indicative of estuarine flow. Surface salinity increased seaward from JV-7 to JV-3, while the depth of the pycnocline remained similar at both stations (Figures 2.3, 2.4 and 2.5). This salinity structure is characteristic of estuarine flow where entrainment of mid-depth waters moving landward causes the surface layer to gain salt and velocity as it flows seaward. No upper mixed layer exists in Saanich Inlet (Herlinveaux, 1962) or Jervis Inlet (Pickard, 1961). Instead, the pycnocline (closely represented by the halocline) intersected the surface in both inlets and was steepest in Jervis Inlet (Figure 2.5). The vertical temperature structure of both inlets (Figures 2.1 through 2.5) reflected seasonal heating. Maximum surface temperatures were usualy in July and August, but in some years were in June (Figures 2.1 through 2.4). In Jervis Inlet, maximum surface temperatures lagged the freshet by one or two months. In the fall and winter, surface cooling destroyed the seasonal thermocline. 2.3.2 N u t r i e n t c o n c e n t r a t i o n s a n d u p t a k e Of the measured nutrients, nitrate most frequently dropped below detectable concen-trations. Beginning late April to early May and continuing until the early fall, surface nitrate concentrations were always below 5 pM at each station and concentrations less than 1 pM occured during most of this period (Figures 2.1 through 2.5). Of the four stations, nitrate depletion (< l^xM NOj) was most common at JV-7 (Figure 2.4). In Saanich Inlet, surface nitrate replenishments at SN-9 were detected on June 27, 1988, Chapter 2. Primary production in Saanich and Jervis Inlets 32 July 4, 1989 and August 28, 1989. At SN-0.8, replenishment was observed on July 4, 1989. When these nutrient replenishments occurred, primary production was relatively low, except on August 28, 1989 at SN-9. Similar lags between nutrient supply and phy-toplankton production or biomass were found by Takahashi et al. (1977) and Parsons et al. (1983). Low nitrate concentrations in the deep waters of Saanich Inlet (Figure 2.5) were caused by nitrate reduction in the oxygen-depleted basin. The slight increase in winter nitrate concentrations near the bottom at station SN-0.8 was most likely caused by up-inlet penetration of dense, nitrate-rich waters. The nutrient assimilation ratios are estimated as the slopes of nutrient-nutrient plots (Corner and Davies, 1971) for samples from waters collected at < 50 m (Figure 2.6). The nitrate:phosphate (N:P) assimilation ratios were 13.6 - 14.2 in Saanich Inlet and 12.7 - 13.1 in Jervis Inlet, compared with 12.5 as found by Smethie (1987) in Jervis Inlet. These assimilation ratios are only slightly lower than the global average of 15 -16 for the N:P concentration ratio in seawater (Redfield, 1963; Takahashi et al., 1985). The silicic acid:nitrate (Si:N) assimilation ratios were between 1.51 and 1.56 at stations SN-9, SN-0.8 and JV-3 , but at station J V - 7 the ratio was 1.22, implying that toward the head of Jervis Inlet diatoms were either more weakly silicified or they made a somewhat smaller contribution to the phytoplankton community. These Si:N assimilation ratios are high, even for phytoplankton communities composed entirely of diatoms. Brzezinski (1985) found that the ratio of the contents of Si and N for various diatoms grown in the laboratory ranged between 0.8 and 1.4, but much higher Si:C (and, assuming a less variable C : N assimilation ratio, Si:N) assimilation ratios have been measured in the Southern Ocean (Queguiner et a l , 1997). Chapter 2. Primary production in Saanich and Jervis Inlets 33 80 60 40 i~ 20 5 o CO X 60 40 20 0 0 10 20 30 40 0 10 20 30 40 NO" (uM) Figure 2.6: N:P and Si:N uptake ratios, taken as the slopes of the regression lines through points representing samples from waters < 50 m and excluding those with nutrient con-centrations less than the detection limit. The 95% confidence intervals of the slopes (in percent of the slope value) are: ± 5.1 - 7.2% for N:P, and ± 7.0 - 8.7% for Si:N. 95% confidence intervals for the intercepts (in pM) are: ± 1.4 - 2.2 pM for N:P, and ± 2.4 -3.2 pM for Si:N. Chapter 2. Primary production in Saanich and Jervis Inlets 34 2.3.3 Est imates of p r imary product ion Volumetric and areal estimates of primary production are shown in Figures 2.1 through 2.4 along with contour plots of nitrate, temperature and salinity in the upper 45 m so that comparison can be made between phytoplankton growth and the hydrographic and nu-trient conditions. The primary production results are summarized in Figures 2.7 (mean vertical profiles) and 2.8 (smoothing of the areal values to describe production through-out the year). The sampling depths for 1 4 C incubations were predicated by in situ light intensity and the deepest points in panel 2 of Figures 2.1 to 2.4 are the extrapolated depths of 1% surface irradiance (see section 2.2). O n average, the depth of 1% surface irradiance was at 12 to 15 m during the summer and 16 to 20 m in winter (Figure 2.7). Generally, the highest rates of carbon fixation occurred at the depth of 56% surface irradiance, but some subsurface maxima were observed (Figures 2.1 through 2.4 and Fig-ure 2.7). However, because surface primary production was not measured, subsurface maxima occurring above the depth of 32% surface irradiance (the second measurement in the depth profiles) would have been undetected. Excluding the 89-day interval separating the first (7-9 August, 1985) and second (4-5 November, 1985) cruises, the sampling interval was between 21 and 55 days, with an average of 32 days. In Saanich Inlet, changes in phytoplankton biomass fluctuate more frequently than this interval (Takahashi et al., 1977; Hobson and McQuoid, in press) and the same is likely true for Jervis Inlet. Because the sampling frequency may have missed important periods of high or low primary production, this time-series is not ideally suited for a comparison of primary production between years. However, 45 to 47 profiles of 1 4 C-uptake were made at each station. Assuming the sampling was random with respect to short-term fluctuations, these data provide a means to describe average primary production throughout the four-year period ofthe study (Figure 2.8 and Chapter 2. Primary production in Saanich and Jervis Inlets 35 Figure 2.7: Vertical profiles of primary production during October to March ("winter") and Apri l to September ("summer"). The plotted depths and rates of carbon uptake are the averages of all measurements taken within each time period (Figures 2.1 to 2.4; n = 21-24 for each profile); surface and deepest (1% surface irradiance) points are extrapola-tions. As these plotted depths are the average depths of 56, 32, 18, 11 and 7% surface irradiance, they can be used to estimate extinction coefficients of light at each station (section 2.4.2). In the winter, the extinction coefficients were: 0.23 m - 1 (JV-7) < 0.24 m - 1 (JV-3) < 0.28 m - 1 (SN-0.8) < 0.30 m - 1 (SN-9). In the summer, the coefficients were: 0.34 m - 1 (SN-9) < 0.35 m - 1 (SN-0.8) < 0.36 m - 1 (JV-3 and JV-7). Table 2.1). By averaging the yearly estimates at the two stations in each fjord, primary production in Saanich Inlet (1.3 g C m - 2 day - 1 ) was approximately 1.7 times higher than in Jervis Inlet (0.78 g C m - 2 day - 1 ) . Average primary production for the entire Strait of Georgia has been estimated to be 280 g C m - 2 y - 1 , or 0.77 g C m - 2 d - 1 (Harrison et al., 1983, using data of Stockner et al., 1979). Thus, Saanich Inlet appears to be significantly more productive than other local waters. Table 2.1 includes other published reports of primary production from Saanich and Jervis Inlets and, in general, there is good agreement with the results from this study. However, except for those of Takahashi et al. (1975) and Takahashi and Hoskins (1978), Chapter 2. Primary production in Saanich and Jervis Inlets 36 Figure 2.8: Averages of daily primary production. The estimates of Figures 2.1 to 2.4 were arranged by calendar day and a five-point running average was used to smooth the time-series. For each 31-day period, there are approximately five estimates of daily primary production. Chapter 2. Primary production in Saanich and Jervis Inlets 37 a n n u a l m o n t h l y a v e r a g e s location/source a v g Jan Feb M a r A p r May Jun J u l A u g Sep Oct Nov Dec (all values: g C rn" -2 d - , ) Figure 6 SN-9 1 . 6 0.097 0.079 0.31 1.6 4.1 2.9 3.4 3.1 2.2 0.89 0.16 0.078 SN-0.8 1 . 1 0.20 0.17 0.34 1.1 1.9 1.7 2.6 1.8 2.2 0.88 0.16 0.11 J V - 3 0 . 9 2 0.14 0.14 0.32 1.1 2.0 2.2 2.1 1.5 0.91 0.33 0.16 0.11 J V - 7 0 . 6 4 0.075 0.093 0.38 1.0 1.4 1.1 1.0 1.2 0.81 0.29 0.15 0.057 Saanich Inlet Takahashi et al. (1975) 1 Takahashi and Hoskins (1978) 2 Parsons et al . (1983) 3 Parsons et al . (1983) 4 Parsons et al . (1983) 5 0.061 0.080 2.1 1.8 4.0 1.6 0.053 Jervis Inlet Stockner and Cli f f (1975) 6 Stockner and Cli f f (1975) 7 Parsons et al . (1984b) 8 Cochlan et al. (1986) 9 Cochlan et al. (1986) 1 0 0.31 0.29 2.8 2.1 1.8 0.74 1.8 1.1 1.7 3.1 0.83 0.77 1.5 1 Pat r ic ia Bay; n = 15 2 mouth; n = 2-4 m o n t h - 1 3 Prevost Passage; n = 3 11 Satellite Channel; n = 5 5 mouth; n = 3 6 northern Hotham Sound; n = 1 m o n t h - 1 7 southern Hotham Sound; n = 1 m o n t h - 1 8 various stations near mouth; n = 6 9 ~11 km up-inlet from J V - 3 ; n = 1 1 0 ~11 km seaward of J V - 3 ; n = 1 Table 2.1: Annual and monthly estimates of primary production at each station derived by integrating the curves of Figure 6 over the appropriate time intervals. The annual averages are based on 45-47 profiles of primary production measured at each station, while each monthly estimate represents the results of approximately four profiles. For comparison, other estimates of primary production in Saanich and Jervis Inlets are given. Where only hourly estimates of primary production were available (Takahashi et al., 1975; Cochlan et al., 1986), we converted to daily rates using a best approximation of their incubation periods and the model of daily sunshine described in section 2.2. Chapter 2. Primary production in Saanich and Jervis Inlets 38 head central mouth Period (SN-0.8) basin (SN-9) 08 Jan - 13 Aug 31 Mar - 30 Sep 1.2 1.8 0.95 2.2 2.0 2.9 Table 2.2: Comparison of the curves of Figure 2.8 for Saanich Inlet with estimates of primary production made by L. A. Hobson (Department of Biology, University of Victo-ria, Victoria, BC, pers. comm.) in the central basin several km south of SN-9. Hobson's sampling interval was, on average, 18 days. His estimates are based on in situ 24 h 14C incubations usualy carried out at the depths of 0, 1, 2, 3, 4, 5, 6 and 7 m. His 08 Jan - 13 Aug (n = 12) and 31 Mar - 30 Sep (n = 11) estimates are from 1975 and 1976, respectively, while the estimates for stations SN-0.8 and SN-9 are from integrating the curves of Figure 2.8 over the appropriate time periods (08 Jan - 13 Aug and 31 Mar - 30 Sep). Hobson's estimates are his revisions of those presented in Harison et al. (1983). all of the estimates of primary production in Table 2.1 are the result of << 24-hour 14C incubations and therefore do not account for autotrophic respiration at night which can consume a significant fraction of gross carbon assimilation (Langdon, 1993; Sakshaug, 1993). For instance, by comparing 24-h and shorter 14C incubations made throughout the year, Pery et al. (1989) estimated that 20% of daytime net primary production was respired at night by the phytoplankton communites of the continental shelf and slope of Washington State. The thorough investigation carried out in the center of Saanich Inlet during 1975 and 1976 (L. A. Hobson, Department of Biology, University of Victoria, Victoria, BC, pers. comm.) is therefore a pertinent comparison because his results (Table 2.2) are based on 24 h 14C incubations and therefore should account for respiration at night. Considering year-to-year variability and spatial heterogeneity, Hobson's estimates for the center of Saanich Inlet and those from SN-0.8 are similar, while the mouth of Saanich Inlet (SN-9) appears to be significantly more productive (Table 2.2). Chapter 2. Primary production in Saanich and Jervis Inlets 39 Other work has found similar along-inlet gradients in phytoplankton biomass (Hobson and McQuoid, in press) and sinking particle flux (Sancetta and Calvert, 1988; Chapter 3). The winter estimates of primary production at stations SN-9 and SN-0.8, however, are significantly higher than those made by Takahashi and Hoskins (1978) using 24 h 1 4 C incubations. This discrepancy is unlikely the result of our deep extrapolations (see section 2.2) because the differences are too large, but may be due to their methodological consideration of autotrophic respiration at night. Assigning the difference in our winter estimates to respiration at night, up to 60% of the carbon assimilated during the day may have been respired at night by phytoplankton. It is also possible that the estimates of Takahashi and Hoskins (1978) are low due to grazing by the large microzooplankton populations during the 24 h incubations. Because the winter values are small, uncertainty in them has little effect on average yearly primary production. Stockner et al. (1979) present a map of primary production for the Strait of Georgia and connecting waters. Their classification of the mouth of Jervis Inlet (JV-3) agrees with the estimate presented here, while their estimates for the upper reaches of the fjord (JV-7) may be too low. They classify the body and head of Saanich Inlet as a region with rates of primary production between 300 and 400 g C m ~ 2 y _ 1 . The results from SN-0.8 located within this area place it at the top of that range. However, their designation of primary production at the mouth of Saanich Inlet^and in Satellite Channel (200 to 300 g C m ~ 2 y _ 1 ) is half the value of the estimate at station SN-9. The results of Parsons et al. (1983, see our Table 1) suggest that Satellite Channel might also be considered a region with rates > 400 g C m ~ 2 y _ 1 . Chapter 2. Primary production in Saanich and Jervis Inlets 40 2.4 D i s c u s s i o n 2.4.1 T e m p o r a l p a t t e r n o f p r i m a r y p r o d u c t i o n Seasona l v a r i a b i l i t y At each station, daily primary production began to increase around Apr i l and by late September or October decreased to near-winter levels (Figures 2.1 through 2.4 and Fig-ure 2.8), reflecting the availability of light with the exception of notable decreases occur-ring near the summer solstice. Except at station SN-0.8, the lull in production in early summer is especially noticeable in the curves of Figure 2.8, but it is also apparent for at least one year at all of the stations (Figures 2.1 through 2.4). Although these lulls may have been due to sampling artifacts, early to mid-summer minima in primary production or phytoplankton biomass are commonly observed in waters contiguous with the Strait of Georgia (e.g., Gilmartin, 1964; Huntley and Hobson, 1978; Stockner et al., 1979; Har-rison et al., 1983; Hobson, 1983; Haigh et al., 1992) and may be related to the transition from the blooms of several diatom species in the spring (Thalassiosira spp., Skeletonema costatum and Chaetoceros spp.) to summer phytoplankton assemblages. This transition may be facilitated by a number of factors, including decreased upward mixing of sub-surface nutrients due to stratification caused by surface-water warming and weak winds, variations in grazing pressure (Harrison et al., 1983; Bornhold, 2000) and photoinhibi-tion (Takahashi et al., 1973; Harrison et al., 1983). Warm water may further enhance nutrient limitation by causing greater photosynthetic nutrient demands (Hobson, 1981). Aggregate formation (e.g.; Ki0rboe et al., 1994; see also Ki0rboe et al., 1996) at the end of the spring bloom and the summertime freshet may also influence primary production throughout the Strait of Georgia at this time of year. Sancetta (1989a) found that, in the spring and fall, many of the diatom frustules caught by the sediment traps deployed Chapter 2. Primary production in Saanich and Jervis Inlets 41 at each station during the experiment were intact, while during July and August they were mostly fragmented, suggesting that grazing may have been the cause of the early summer minima in production. Low phytoplankton biomass was likely associated with these mid-summer luls, which probably occured when chlorophyl-normalised rates of photosynthesis were maximum (Hobson, 1981). A somewhat surprising feature of this time-series is the lack of a pronounced spring-time peak in production. For a station in the Strait of Georgia, Parsons (1979) shows a large peak in primary production in March, and similar peaks have been observed in Boca de Quadra, southeast Alaska (Burrel, 1983) and Balsfjorden, northern Norway (Hopkins, 1981). However, the production cycles of Figures 2.1 through 2.4 and Fig-ure 2.8 show little evidence for a spring bloom more productive than subsequent months. The lack of this signal may be the result of the long period between sampling (~32 days). Assuming a winter-accumulated nitrate concentration of 30 pM in the upper 10 m, com-plete utilisation would result in approximately 20 to 30 g C m~2 of production which in these waters can be realised in 2 to 10 days. Thus, the sampling may have missed the growth associated with the consumption of winter-accumulated nutrients. This analysis also demonstrates that, in productive waters with a shalow euphotic zone, the portion of yearly primary production occuring as a result of winter-accumulated nutrients is minimal compared with the production driven by nutrients mixed into and regenerated within the euphotic zone; less than 10% of yearly primary production in Saanich and Jervis Inlets is generated by winter-accumulated nutrients. Similar seasonal paterns occur in some shalow fjords and pols of western Norway. Although Korsfjord and Kvi-turdvikpol show highest rates of primary production during the spring, Nordasvann and Vagsbopol do not (Wassmann, 1991). Also, a model based on chlorophyl a concentra-tions, light availability and temperature does not predict peaks in primary production in the spring for fjords and the coast of western Sweden near Norway (Soderostrom, 1996). Chapter 2. Primary production in Saanich and Jervis Inlets 42 Interannual var iabi l i ty Because of the long period (~32 day) between sampling, year-to-year variability in this time-series is not emphasised. Nevertheless, during the E l Nino year of 1986 (Figures 2.1 through 2.4), primary production was 1.8, 1.3, 1.5 and 1.5 times higher than the annual averages (Table 2.1) at stations SN-9, SN-0.8, JV-3 and JV-7 , respectively. Also in 1986, the Fraser River freshet was delayed and the June peak was amplified. Compared to the averages for 1987-1989, flow of the Fraser River at the beginning of the freshet in May, 1986, was about 25% reduced, in June and July at the peak of the freshet was approximately 30% greater and total flow for the year of 1986 was 12% greater. The Fraser River freshet has a significant impact on the physical and biological oceanography of the Strait of Georgia (Harrison et al., 1983; LeBlond et al., 1994) and Saanich Inlet (Herlinveaux, 1962; Stucchi and Whitney, 1997). Estuarine exchange be-tween the Strait of Georgia and the northeast Pacific Ocean is driven by the Fraser River and occurs primarily through Juan de Fuca Strait (Griffin and LeBlond, 1990). As well as wind-induced upwelling occurring at the seaward end of Juan de Fuca Strait during the summer and fall (Mackas et al., 1987), modelling and data show that the estuar-ine circulation within Juan de Fuca Strait also forces upwelling (Masson and Cummins, 1999). Intense mixing caused by strong tidal currents and estuarine exchange through Juan de Fuca and Haro Straits (eg; Pawlowicz and Farmer, 1998) maintains high surface nutrient concentrations year-round throughout the southern Strait of Georgia (Mackas and Harrison, 1997). Perhaps, from the delay of the Fraser River freshet later into the summer when nutrients normally would be most limiting to phytoplankton in the south-ern Strait of Georgia, increased estuarine flow in June and July of 1986 provided a greater supply of nutrients to surface waters by enhancing upwelling and mixing at this time of year. Although the effect on Jervis Inlet is less certain, it is possible that the imprint Chapter 2. Primary production in Saanich and Jervis Inlets 43 of Fraser River discharge on the density structure of waters of the Strait of Georgia affects exchange with Jervis Inlet and therefore has an influence on the phytoplankton populations within the fjord. It is possible that the freshet from the various rivers (all un-metered) entering Jervis Inlet was also anomalous in 1986. 2.4.2 G e o g r a p h i c p a t t e r n o f p r i m a r y p r o d u c t i o n Primary production in Saanich Inlet was significantly higher than in Jervis Inlet and the seaward stations of both fjords were more productive than their landward counterparts (Table 2.1). This pattern was reflected in the flux recorded by the sediment traps moored at 50 m at each station during the study (Chapter 3). Of the sediment-trap material, biogenic silica (BSi) is a good tracer of local primary production (Sancetta and Calvert, 1988; Sancetta, 1989a) as diatoms are its principal source. Organic carbon (OC) is not as good a tracer of primary production due to the presence of terrestrial OC and the high degree of recycling of OC in surface and sub-surface waters (Sancetta and Calvert, 1988; Timothy and Pond, 1997; Chapters 3 and 4). The measured flux of OC might also be affected by live zooplankton entering the traps; swimmers would more likely modify measured fluxes of OC than those of BSi. Annually averaged, the fluxes of BSi to the 50 m traps (g BSi m~ 2 d - 1 ) were: 1.8, 0.67, 0.46 and 0.31 at SN-9, SN-0.8, JV-3 and JV-7, respectively (Chapter 3). Other than at station SN-9 where material washing into Saanich Inlet or resuspended off the sill may have reached the 50 m traps (Sancetta and Calvert, 1988; Sancetta, 1989a; Chapter 3), the molar ratio of the 50 m BSi flux to local primary production (each averaged over the entire study period) was similar at each station, ranging between 0.087 and 0.11. Assuming that the average Si:C ratio of cultured marine diatoms (0.13±0.04, Brzezinski, 1985) can be applied to the diatoms of Saanich and Jervis Inlets, more than half of assimilated Si sank out of the euphotic zone and was caught in the 50 m traps. Chapter 2. Primary production in Saanich and Jervis Inlets 44 This large BSi flux supports the suppositon that diatoms were the predominant primary producers in these fjords in spring and summer when most of the growth occured, even though flagellates are present throughout the year and can be abundant or predominant during periods of the summer, fall and winter (e.g.; Harison et al., 1983 and references therein). The large BSi flux also implies that much ofthe Si deposited as diatom frustules was neither recycled in the euphotic zone nor dissolved in waters less than 50 m deep. The recycling of organic carbon in surface waters appears to have been much greater than that of BSi, as was its spatial and temporal variability (Sanceta and Calvert, 1988; Chapter 3). The ratios of the 50 m OC flux to primary production at each station (each averaged over the entire study) ranged between 0.10 and 0.16 (see Chapter 3). These low export ratios, especialy considering that terrestrial OC is included in the sediment-trap fluxes, may have been due to the degradation of OC within the sediment traps. Kumar et al. (1996) found that as much as 70% of particulate organic carbon caught by sediment traps was released into solution during deployment, even when the deployments were short. The factors afecting the OC fluxes wil be discussed further in Chapter 3. The spatial patern of primary production may have been the result of the distribution of nutrient supply to surface waters throughout the southern Strait of Georgia. Due to upweling into Juan de Fuca Strait and mixing in Juan de Fuca and Haro Straits as discussed above, surface waters with year-round nitrate concentrations in excess of 10 pM reach into Satelite Channel at the mouth of Saanich Inlet but do not extend as far north as the entrance to Jervis Inlet (Mackas and Harrison, 1997). Waters above the sill in Saanich Inlet are in exchange with those of Satelite Channel both tidaly and through isopycnal mixing (Herlinveaux, 1962; Hobson, 1985; Stucchi and Whitney, 1997), and Hobson (1985) showed that lateral mixing at depths of 5-20 m, then vertical mixing to the surface, was a major source of nutrients to the phytoplankton in Saanich Inlet. The surface stratification and generaly weak winds and tides of Saanich Inlet (Herlinveaux, Chapter 2. Primary production in Saanich and Jervis Inlets 45 1962; Stucchi and Whitney, 1997) are furthermore conducive to phytoplankton growth, and intrusions of nutrient- and phytoplankton-laden waters into the inlet during spring tides result in high chlorophyl biomass throughout much of the inlet, but especialy towards the mouth (Takahashi, 1977; Parsons et al., 1983; Hobson and McQuoid, in press). Although the biological front at the mouth of Saanich Inlet is wel documented (Par-sons et al., 1983; Hobson and McQuoid, in press), there is less evidence that tidal mixing through constrictions in the vicinity of the mouth of Jervis Inlet (JV-3) causes heightened primary production. Cochlan et al. (1986) found no evidence of a biological front at the mouth of Jervis Inlet, while Parsons et al. (1984b) did observe high chlorophyl a concen-trations and high rates of primary production ~6 km southeast of JV-3. They ascribed the heightened biological activity to enhanced nutrient supply by intense mixing through Skookumchuk Narows at the mouth of Sechelt Inlet (Figure 1.1). The finding that pri-mary production at the mouths of Saanich and Jervis Inlets was approximately 1.4 times greater than at the inland stations is surprising considering the relative weakness of the front at the mouth of Jervis Inlet. However, unpublished work by Ann Garget (Old Dominion University, Norfolk, VA, pers. comm.) suggests an eficient supply of surface nutrients throughout the length of Saanich Inlet during transitions from neap to spring tides. During neap tides, water from the Cowichan and Fraser Rivers causes low surface salinity of Satelite Channel and Saanich Inlet. Intensified mixing during spring tides increases surface salinity in Satelite Channel and a density-driven surface outflow from Saanich Inlet results. This outflow is accompanied by rapid, sub-surface inflow, providing much of the length of Saanich Inlet with a new supply of nutrients. Other evidence that nutrients were supplied to the phytoplankton of Saanich Inlet lat-eraly from Satelite Channel is the high nitrate to phosphate (N:P) ratios at mid-depths Chapter 2. Primary production in Saanich and Jervis Inlets 46 1 2 3 0 1 2 3 4 5 6 7 HP0 4 2" OiM) Figure 2.9: Nitrate versus phosphate in both inlets. The average N:P ratio at 30-50 m was 10.7 in Saanich Inlet and 11.4 in Jervis Inlet. Most of the samples from 30-50 m in Saanich Inlet with PO3- concentrations greater than 3 M^ and low N:P ratios were colected in the summer and fall during deep-water renewals. Otherwise, the N:P ratios at 30-50 m in Saanich Inlet suggest little mixing between deep and mid-depth waters. of the inlet (Figure 2.9). When oxygen depletion and subsequent nitrate reduction (Fig-ure 2.5) occured in the deep waters of Saanich Inlet, N:P ratios decreased substantialy (Figure 2.9). If vertical mixing between deep and shalow waters were a significant source of nutrients to the phytoplankton, then, assuming a significant fraction of nitrate reduc-tion were to completion (yielding N2 gas), low N:P ratios would have been found at intermediate depths (~30-50 m). Although some mixing into shalower regions did occur during deep-water renewal (see caption to Figure 2.9), for most ofthe year there is little evidence that deep nutrients mixed upward. Finding that (515N of the sediment-trap ma-terial caught in Saanich Inlet was similar to that from Jervis Inlet, Calvert et al. (2001) have come to the same conclusion; mixing of nitrate from the deep waters of Saanich Inlet (where nitrate reduction would cause the remaining nitrate to be isotopicaly heavy) into the euphotic zone must have been smal compared with the lateral nutrient supply from outside the fjord. Also, Ward et al. (1989) showed that upward methane fluxes across Chapter 2. Primary production in Saanich and Jervis Inlets 47 30 m depth in Saanich Inlet were too smal to balance the evasive flux to the atmosphere and Liley et al. (1982) concluded that lateral exchange between Saanich Inlet, Satelite Channel and the Strait of Georgia, combined with the isolation of the deepwater methane of Saanich Inlet, caused regional similarity in methane profiles at depths above the sill. Thus, enhanced primary production at SN-0.8 relative to JV-7 may have been the result of a greater landward nutrient flux in Saanich Inlet and could explain the similar-ity in the seaward:landward ratios of primary production between the fjords despite the productive biological front at the mouth of Saanich Inlet. Indeed, of the four stations, surface nitrate depletion during the growing seasons was most severe at JV-7 (Figures 2.1 through 2.4) and it appears that at JV-7 diatoms were a smaler fraction of the phyto-plankton compared with the other stations (Figure 2.6). A higher relative abundance of flagellates at the landward station in Jervis Inlet may have been a response to a lower nutrient supply. It should be noted that variations in light atenuation due to riverine silt is an unlikely cause of the observed geographic patern of primary production. Within Howe Sound south of the Jervis Inlet system, primary production is afected by the turbidity caused by riverine silt and clay (Stockner et al., 1977; Parsons et al., 1981), raising the possi-bility that the more glacialy influenced riverwater of Jervis Inlet may contain suficient silt to afect light transmission, especialy at station JV-7. However, a comparison of average winter and summer extinction coeficients at each station is not consistent with this possibilty (Figure 2.7). Indeed, Captain Vancouver (1798, from Pickard {1961}) observed silty waters in some of the British Columbian fjords, but not in Jervis Inlet and, during the cruises of this experiment, silt was not noticed in the surface waters of Saanich or Jervis Inlet (Maureen Soon, UBC, pers. comm.) Chapter 2. Primary production in Saanich and Jervis Inlets 48 2.4.3 B o t t o m - w a t e r o x y g e n i n S a a n i c h a n d J e r v i s In le t s The oxygen dynamics of fjords are afected by factors including fjord morphology (e.g.; sill and fjord dimensions), water circulation and mixing, the frequency and oxygen supply of deep-water renewal and the delivery of autochthonous and alochthonous organic mater to the deep waters. Of the BC fjords, Saanich Inlet is unique, as the other anoxic basins are severely restricted by their sils from deep-water renewals (section 2.1). The results of this chapter show that primary production and export flux were high during the study in Saanich Inlet and the principal cause has been argued to be the fertility of adjoining waters. The weakness of estuarine flow may have further enhanced particle flux to botom waters. Advection is recognised as an important factor in the transport of particles into and out of fjords (e.g.; Sakshaug and Myklestad, 1973; Lewis and Thomas, 1986; Wassmann, 1996) and Gilmartin (1964) estimated that 25% of local community production was lost from a nearby fjord (Indian Arm) due to estuarine circulation. Such losses from Saanich Inlet would be minimal and, although the source of the very high sediment-trap fluxes at 50 m at station SN-9 is uncertain (whether resuspended of the sill or washed into the fjord; see Chapter 3), it is possible that because of the unusual circulation there was a net flux of organic material into Saanich Inlet. These sources of organic mater to the deep waters are likely to result in an unusualy large oxygen demand behind the sill of Saanich Inlet. Also, low rates of vertical mixing wil exacerbate oxygen depletion in deep basins. Although some evidence exists that mixing rates in Saanich Inlet are not anomalous (Smethie, 1981), weak estuarine circulation, winds and tides provide little energy to mix the deep waters of Saanich Inlet (Ann Garget, Old Dominion University, Norfolk, VA, pers. comm.) Indeed, DeYoung and Pond (1988) found vertical eddy difusion is about an order of magnitude lower in Saanich Inlet than in Indian Arm. In Jervis Inlet, botom waters remain oxygenated despite infrequent (< one year-1) Chapter 2. Primary production in Saanich and Jervis Inlets 49 deep-water renewals. Primary production was not exceptionaly high in Jervis Inlet and the estuarine flow, albeit weak, would cary a portion of surface biomass out of the fjord. The deep (385 m) sill must also play an important role in the oxygen balance of the deep waters of Jervis Inlet, as much of the sinking organic mater wil be oxidised before botom waters are reached. 2.5 Conclusions A four-year time-series of monthly primary-production determinations in Saanich and Jervis Inlets, British Columbia, Canada shows that Saanich Inlet was significantly more productive than Jervis Inlet or the Strait of Georgia, while the landward stations within each fjord were less productive than those seaward. Surface nutrient-supply from outside the fjords was likely higher to Saanich Inlet and may have controled the diferences in primary production between stations. Fluxes of biogenic silica at 50 m reflected the geographical patern of primary production and the oxygen demand caused by the setling flux of organic mater probably contributed significantly to the periodic, deep-water anoxia of Saanich Inlet. Retainment of organic particles due to weak estuarine exchange and low rates of vertical mixing in Saanich Inlet would have further induced anoxia. Chapter 3 Sources and patterns of sett l ing fluxes i n Saanich and Jervis Inlets 3.1 In t roduct ion Coastal oceans constiute approximately 10% of the area of the global oceans, yet these margins account for ~20% of global ocean primary production (Walsh, 1988; Liu et al., 2000) and most of the organic carbon burial in marine sediments (Berner, 1982; Hedges and Keil, 1995; Middleburg, 1997). Fluvial delivery (Meybeck, 1982) is the prin-cipal source of organic mater to coastal waters and sediments (Berner, 1989; Hedges and Keil, 1995; Liu, 2000). However, much of the primary production of these nutrient-replete boundaries is by diatoms (Nelson et al., 1995) and, because of their high growth rates and unique physiological adaptations to turbulent and nutrient-rich waters (Smetacek, 1985), diatoms are also important vectors of carbon flux in coastal setings (e.g.; Pitcher, 1986; Sanceta and Calvert, 1988; Sanceta, 1989a, 1989b, 1989c; Waite et. al., 1992; Conley and Malone, 1992; Kiorboe et al., 1994; Ki0rboe et al., 1996; Tiselius and Kuylenstierna, 1996; Del Amo et al., 1997; Pike and Kemp, 1999). The relatively shalow waters of continental margins and the high rates of sediment accumulation and burial furthermore result in short diagenetic residence times and enhanced organic mater preservation (Mid-dleburg, 1997). Thus, these factors (large terrestrial and marine inputs, enhanced preser-vation) result in the disproportionate amount of organic mater burial in coastal marine regions. However, although coastal regions play an important role in global carbon fluxes, they are characteristicaly heterogeneous and generalisations of biogeochemical processes 50 Chapter 3. Settling fluxes in Saanich and Jervis Inlets 51 are difficult to make (Smith and Hollibaugh, 1993; L iu et al., 2000). Glacially-carved fjords are unique coastal features that are typically over-deepened and hydrographically restricted by sills (Syvitski et al., 1987; Burrell, 1988; Wassmann, 1991; Wassmann et al., 1996). Fjords are catchment basins for high-latitude mountainous regions and collect a disproportionate amount of weathered products delivered from land to the much as one quarter of the fluvial sediment entering the sea in the past 100,000 years is trapped in fjord basins (Syvitski et al., 1987). Furthermore, fjords are semi-enclosed mesocosms mimicking ocean basins (Skei, 1983; Burrell, 1988) and have long been recognised as accessible locations where physical, biological and geochemical processes relevant to oceanic systems can be studied. In this chapter, I examine a unique time-series of particle fluxes in two British Columbian fjords in an attempt to gain further understanding of the physical and ecolog-ical processes leading to water-column sedimentation and, ultimately, bottom sediment burial. Sediment traps were moored at stations in Saanich Inlet and Jervis Inlet, two contrasting fjords of British Columbia, Canada, for about five years during the 1980s; some of the information on the time series has been published. The 1983-84 sediment-trap fluxes from Saanich Inlet were used to aid interpretation of the geochemistry of the bottom sediments (Frangois, 1987; 1988), and the fluxes at 50 m from a 1985 station were included in a study of marine and terrigenous biomarkers in Saanich Inlet (Cowie et al., 1992). Fluxes of total dry weight, carbon, biogenous silica and diatom valves from Saanich Inlet during 1985 have been described (Sancetta and Calvert, 1988), as well as diatom valve flux in relation to atmospheric conditions for both inlets from 1985 to 1987 (Sancetta, 1989a; 1989c). Sancetta (1989b) reported fecal pellet flux for 1986 at the landward station in each fjord. Based on microscopic examination of the sediment-trap samples, Sancetta's observations provide a very useful descriptive foundation for this work. Primary production for most of the time-series has also been reported (Chapter Chapter 3. Settling fluxes in Saanich and Jervis Inlets 52 2; Timothy and Soon, 2001). The total mass, biogenic silica, organic carbon, nitrogen and aluminium fluxes are presented and discussed in this chapter. A relationship between 513C and biogenic silica content is used to separate marine and terrestrial organic carbon in a way that alows estimates of the relative contribution of diatoms to the sinking flux of marine organic mater. Export ratios for organic carbon and biogenic silica are computed using the primary production results of Chapter 2 and the sediment trap fluxes. Finaly, the sediment-trap fluxes are compared to mass accumulation rates of the botom sediments, alowing a top (primary production) to botom (sediment accumulation) description of the organic carbon and biogenic silica fluxes in two fjords with difering redox conditons. 3.2 Mater ia l s and methods 3.2.1 F i e l d design and sample preparat ion The sediment-trap study began in the fall and winter of 1983-84 in Saanich Inlet and expanded into Jervis Inlet in spring-summer of 1985 (Figures 1.1, 1.2 and 1.3, and Ta-ble 3.1). The initial mooring site (JV-11.5) at the mouth of Jervis Inlet was in an active shipping lane, so the mooring was moved to station JV-3 after several deployments. Throughout this work, the data colected from JV-11.5 are treated with those colected at JV-3. Sediment traps were placed at three depths on each mooring and were serviced and redeployed approximately monthly. The sediment traps were made from PVC cylinders with an inside diameter of 12.7 cm and a height of 48 cm (aspect ratio = 3.8). Bafle grids (1.5 cm square) were placed in the opening and toward the base of each trap in order to decrease mixing (Gardner, 1980a). Prior to deployment, traps were filled with seawater and 500 mL of a 30% brine solution were funneled to the botom of the traps. The sediment traps were always Chapter 3. Settling fluxes in Saanich and Jervis Inlets 53 station location water depth (m) trap depths (m) sediment-trap time series (yymmdd) l ° - p r o d u c t i o n time series (yymmdd) Saanich mouth 45 830809 850807 SN-9 48°40.2'N 165 110 to to 123°30.2'W 150 891215 891010 Saanich head 50 840112 850809 SN-0.8 48°33.0'N 210 135 to to 123°32.7'W 180 891215 891010 Jervis mid-inlet 50 850328 850808 J V - 7 50°03.4'N 530 200 to to 123°48.9'W 450 891219 891011 Jervis mouth 50 850808 850808 JV-11.5 49°48.6'N 660 300 to to 123°56.1'W 600 860128 860128 Jervis mouth 50 860311 860311 JV-3 49°48.3'N 660 300 to to 124°02.4'W 600 890829 891011 Table 3.1: Station locations, sediment trap depths, and duration of the sediment-trap and primary-production time series. deployed in pairs with 0.5% sodium azide (NaN3) as a bactericide in one trap of each pair. NaN3 inhibits aerobic but not anaerobic bacterial respiration and acts as a poison to zooplankton that may be atracted to material caught by traps. Upon retrieval, the trapped debris was filtered through 0.47 mm Nitex monofilament bolting cloth to remove large zooplankton and other swimmers and in the laboratory the sediment was washed free of salt by repetitive centrifugation with deionised water. The solid phase was freeze-dried, weighed and ground to a fine powder. From 1985 to 1987, the sample was split upon retrieval and analysed microscopicaly (Sanceta and Calvert, 1988; Sanceta, 1989a, 1989b, 1989c). Chapter 3. Settling fluxes in Saanich and Jervis Inlets 54 3.2.2 Labora tory analyses Total carbon and nitrogen were determined by gas chromatography on a model 1106 Carlo-Erba CHN analyser with an analytical precision of ±1.3% for carbon and ±2% for nitrogen (Verardo et al., 1990). Inorganic carbon was determined with a Coulometrics Inc. CO2 coulometer (precision ±2%) and converted to CaCOs. Organic carbon (OC) was determined by subtracting inorganic carbon from total carbon. Biogenic silica (BSi) was measured folowing the method and equations of Mort-lock and Froelich (1989). The procedure involves extracting amorphous silica with 2 M Na2C03 and then measuring the dissolved silcon concentration in the extract by molybdenum-blue spectrophotometry. Conversion from biogenic silcon (%Si) to bio-genic silica (%Si02 ' nH20) assumed biogenic silica contains 12% water by weight. The precision for replicates of in-house sediment standards with relatively high BSi content (> 10% BSi) was ±4%. Aluminium (Al) and other metals (Ti, Fe, Ba and Zr) were measured by inductively-coupled plasma mass spectrometry (ICP-MS). Approximately 5 mg sediment were added to a ~0.5 mL cocktail of concentrated H N O 3 , HC1 and HF in 1.5 mL Teflon vials with screw caps. The vials were placed in larger Teflon 'bombs' and microwaved for 30 min at ~40 psi. The vials were then placed on a hotplate, with their caps removed, until dryness. ~0.5 mL 2 N H N O 3 was weighed and added to the residue and after a second 30 min, 40 psi microwave digestion, the 2 N H N O 3 solution was diluted ~400-fold with 0.1 N HN03 using a precision balance. Further dilutions were performed by volume. Al, Ti, Ba, Fe and Zr were analysed on a VG Elemental 'PlasmaQuad Turbo Plus' ICP-MS equipped with an SX300 quadrupole mass analyser and Galieo 4870-V channel electron multiplier. Liquid argon served as nebuliser, auxilary and cooling gases. External standards were used to determine metal concentrations in the diluted samples. The ICP-MS was de-tuned Chapter 3. Settling fluxes in Saanich and Jervis Inlets 55 so that the response was linear to 100 ppb A l in order to minimise required dilutions. Indium and scandium were used as internal standards to correct for plasma instabilities and sensitivity changes during the analyses. A number of in-house (bottom sediments from Saanich and Jervis Inlets) and certified sediment (MESS-2) standards were digested and analysed with the sediment-trap samples; the accuracy of the A l analyses on these standards was 3% or better, while the precision was ± 6 % or better. The isotopic composition of organic carbon was determined on decarbonated (10% HC1) subsamples using a V G P R I S M isotope ratio mass spectrometer, with a Carlo Erba C H N analyser fitted in-line as a gas preparation device. Values are reported in S notation relative to the P D B reference: r5 1 3C = 1000 ( [ 1 3 C : 1 2 C ] s a m p i e / [ 1 3 C : 1 2 C ] p D B - 1). The precision was ±0 .2° /oo -3.2.3 Effect of preservative treatments To test the effectiveness of sodium azide as a preservative, fluxes measured by the sedi-ment traps with and without N a N 3 from all depths at each station were averaged. Fluxes were considered instead of percentages because variable preservation of one constituent could affect the content of another constituent. For the total mass, O C and N fluxes, there was no clear preservative effect for N a N 3 (Table 3.2). Therefore, in order to re-duce the number of analyses, BSi and A l were measured on sediments collected from the NaN 3 -treated traps and <513C was measured on samples from the 'brine only' traps. A number of cross-analyses found little or no effect of preservative treatment on BS i fluxes (Table 3.2). Where analyses were performed on samples from each trap in a pair, the values presented in this chapter are the average of the two treatments. C a C 0 3 fluxes were 16 to 28% greater for the sediment traps treated with N a N 3 (Ta-ble 3.2). C 0 3 _ is undersaturated with respect to C a C 0 3 in coastal waters, but N a N 3 buffers C a C 0 3 dissolution (Knauer and Asper, 1989); Timothy and Pond (1997) also Chapter 3. Settling fluxes in Saanich and Jervis Inlets 56 SN-9 SN-0.8 JV-3 J V - 7 mass flux (g m - 2 d _ 1 ) brine 9.66 2.16 1.62 1.27 brine + N a N 3 9.74 2.19 1.66 1.31 %A (n) 0.79 (173) 1.3 (189) 2.2 (122) 3.3 (144) O C flux (mg m - 2 d"1) brine 468 176 144 130 brine + N a N 3 457 173 140 135 %A (n) -2.4 (173) -1.5 (189) -2.7 (122) 4.0 (144) N flux (mg m~ 2 d - 1 ) brine 57.6 22.7 16.8 15.7 brine + N a N 3 55.2 22.2 16.4 16.3 %A (n) -4.2 (173) -2.2 (189) -2.4 (122) 3.9 (144) C a C O s flux (mg m" 2 d"1) brine 66.3 36.0 22.8 19.1 brine + N a N 3 78.9 45.4 30.4 26.6 %A (n) 16 (173) 21 (189) 25 (122) 28 (144) BSi flux (mg m" 2 d _ 1 ) brine 2750 1450 703 668 brine + N a N 3 2840 1400 686 677 %A (n) 3.4 (9) -3.5 (7) -2.4 (13) 1.5 (6) Table 3.2: Average fluxes to both sets of traps (with and without N a N 3 ) in each deploy-ment-pair. Negative differences occur where fluxes to NaN 3 -treated traps were smaller than fluxes to traps without N a N 3 . Chapter 3. Settling fluxes in Saanich and Jervis Inlets 57 found that C a C 0 3 was higher in sediments from NaN 3 -treated traps in a shorter exper-iment in Sechelt Inlet (Figure 1.1). Because of the N a N 3 effect on C a C 0 3 preservation, the lack of an effect on organic preservation cannot be attributed to outwash of the preservative during each deployment. Although these results might therefore suggest that sodium azide is an ineffective bactericide, Lee et al. (1992) found it to be a better preservative than brine alone, though not as good a bactericide as other preservatives. Other possible explanations for the lack of a preservative effect on the organics include: O M loss was minimal in both treatments or the organic degradation was not aerobic; cell lysis and other physical processes affected solubilisation of the most labile organics while the sediment traps were moored; the treatment of samples after they were retrieved (rinsing and centrifugation) removed the labile organics that may have been preserved differently in the traps. 3.2.4 L i n e a r regress ions A l l linear regressions are model II (functional) geometric mean regressions ( G M R ) , also known as the reduced major axis regression. The 95% confidence intervals of the slopes and intercepts are calculated according to Ricker (1984). Functional regressions are preferred over the more commonly applied model I (predictive) regression when describing relationships of field data such as reported here (Ricker, 1973, 1984; Mark and Church, 1977; Laws and Archie, 1981; Sokal and Rohlf, 1995). A property of the G M R line that makes it better for describing the relationship between two variables with similar degrees of natural variability is that, when y is plotted against x, the slope is the inverse of the line obtained when x is plotted against y. For practical purposes, it is useful to know that the slope of the G M R line (mn) can be calculated from the slope of the model I regression (mi): mn = m i / | r | , where r is the correlation coefficient. Thus, the slope of the G M R line is always steeper than the slope of the model I regression line; the Chapter 3. Settling fluxes in Saanich and Jervis Inlets 58 importance of choosing the proper regression increases as the correlation between two variables decreases. Although the use of G M R is strongly advised for the description of naturally varying data (references above), there remains controversy about the application of model II regression for predictive purposes. Ricker (1973, 1975, 1984) recommends that, in some cases where there is natural variability in a data set, G M regression should be used for both description and prediction, while Sokal and Rohlf (1995, p. 543) discourage the use of model II regressions for any predictive purpose. In the one case where regression is used for predictive purposes (Figures 3.19 and 3.20), the G M R is used. 3.3 R e s u l t s 3.3.1 C o m p o n e n t s o f t he mass flux To satisfy the mass balance of the total flux, factors to convert organic carbon (OC) to total organic matter and A l to total lithogenic material are estimated. Following Timothy (1994) and Timothy and Pond (1997), the O M : O C ratio is approximated by assuming a molar O C P ratio of 106:1 (Redfield et al., 1963) and writing the model organic molecule as: (CH 2 0) io6(NH3) x (H 3 P04) , where x = 106 N : O C For the range of O C N ratios observed during the study (7.1 to 15; Figure 3.1), the O M : O C ratio was 2.67 to 2.78. The use of 106:1 for the O C P ratio is an overestimate of the amount of phosphorus in the sediment-trap samples (unpublished data), but these estimates are not sensitive to the O C P ratio. Figure 3.2 shows the relationship between A l and total lithogenic material (LM) . According to these plots, the lithogenic flux was 8.5 - 10% A l . The tendency for negative intercepts on the A l axes of Figure 3.2, especially at station SN-0.8, may be caused by the presence of quartz in the samples. Gucluer and Gross (1964) reported quartz in the Chapter 3. Settling fluxes in Saanich and Jervis Inlets 59 Figure 3.1: O C versus N for the entire data set, showing the range in O O N ratios caught in the traps. Dashed lines in the upper panels are the slopes of the regression lines of the lower figures. Filled circles are summer data, open circles winter data. By multiplying weight percent N by 12/14, the slope of the regression line gives the mean molar ratio. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 60 Figure 3.2: Al versus lithogenic material at each station. The lithogenic percentage is calculated as 100 - %BSi - (OM:OC) %OC - %CaC03. The slopes are interpreted as being the average Al content of the setling aluminosilcates. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 61 sands and fine laminae of the bottom sediments of Saanich Inlet and X-ray diffraction analyses (unpublished) indicate the presence of quartz in both sediment-trap and bottom sediments of Saanich and Jervis Inlets. CaCO"3 made only small contributions to the sediment-trap material (on average 2% of the mass flux at stations JV-3 , J V - 7 and SN-0.8, and 0.8% of the mass flux at station SN-9), as foraminifera and coccolithophorids are rare in waters connecting to the Strait of Georgia. Other sources of C a C 0 3 may have been resuspended fragments of mollusk shells, eroded limestone exposures within the watersheds and, in Saanich Inlet, a cement plant at mid-fjord on the west side (Gucluer and Gross, 1964). Higher carbonate fluxes to the deep trap at SN-0.8 (average of 2.2% and occasionally surpassing 4% of the mass flux) might have been the result of sediment transport from the area of the cement plant. 3.3.2 F l u x e s at t he h e a d o f S a a n i c h In le t ( s t a t i o n S N - 0 . 8 ) The magnitude and the character of the settling fluxes at the two stations in Saanich Inlet differred dramatically, as the sediment traps at station SN-9 were in a nepheloid region or depocenter affected by resuspended sediments most likely originating from the broad sill at the mouth ofthe inlet (Figure 1.1). Therefore, discussed first are the settling fluxes measured at station SN-0.8 toward the head of Saanich Inlet where the influence of the autochthonous phytoplankton was more apparent. A t station SN-0.8, interannual variability of the mass, BSi , O C and A l fluxes was small during the study (Figures 3.3 through 3.6), so when referring to the entire time series, most of the attention will be given to seasonal patterns and depth variations of fluxes (Figures 3.7 and 3.8, and Table 3.3). B i o g e n i c fluxes Mass fluxes to 50 m at station SN-0.8 (Figures 3.3 and 3.7; Table 3.3) were highest in the spring, decreased in the summer and were low in the fall and winter. O f the primary Chapter 3. Settling fluxes in Saanich and Jervis Inlets 62 Figure 3.3: Total mass fluxes in Saanich Inlet. The gray lines are the curves of Figure 3.7, repeated annually. Pooling both stations and all depths, lost samples represented 9% of the time series and on 1% of the collected samples not all laboratory analyses were performed. Missing data can be inferred from gaps in the bar graphs. Note that the range of the ordinate is four times larger for station SN-9 than for station SN-0.8. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 63 C O C M T - O C V J T - O C M T - O L _ , , I . , i , , L _ , , i , , i , , L _ , , i , , i , , r 0 5 ( 0 C 0 O C 0 ' C 0 O C 0 C 0 O L./fep z LU jsa 6 Figure 3.4: Biogenic silica fluxes in Saanich Inlet. The gray lines are the curves of Figure 3.7, repeated annualy. Note the diference in the scale of the y-axes for the two stations. Figure 3.5: Organic carbon fluxes in Saanich Inlet. The gray lines are the curves of Figure 3.7, repeated annualy. There is a three-fold scale change between stations for the OC flux. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 65 components, this patern resembled most closely that of BSi (Figures 3.4 and 3.7). The main source of amorphous silica to Saanich and Jervis Inlets is diatom frustules, with mi-nor contributions made by silcoflagelates (Sanceta and Calvert, 1988; Sanceta, 1989a), while organic mater can have terigenous and other marine (e.g.; flagellates and zoo-plankton) sources. Even with the terrestrial contributions (section 2.4.1), OC fluxes at 50 m showed a similar seasonal patern to those of BSi (Figures 3.5 and 3.7) and both roughly reflected the yearly cycle of primary production (Figure 3.7) by increasing in March-April and decreasing in September-October. Although the deployment period (~1 month) was too long to measure changes in plankton dynamics shorter than seasons (e.g.; Deuser, 1996), the averaged curves of Figure 3.7 show a pronounced June-July drop in the flux of BSi, folowed by a smal rebound in August. The early summer drop in BSi fluxes occured while OC fluxes decreased slightly but remained near peak levels, supporting the possibilty that, due to limited nutrient supply, flagellates played an in-creasingly important role in the ecology of the plankton after the spring bloom towards the head of Saanich Inlet (Takahashi et al., 1977; Hobson, 1981; Parsons et al., 1983; Francois, 1987; Sanceta and Calvert, 1988; Hobson and McQuoid, in press; Timothy and Soon, 2001). Winter fluxes of BSi to 50 m at station SN-0.8 were a factor of 10 smaler than their spring peaks, while the contribution of BSi to the mass flux fel from a maximum of more than 50% to about 15% in the winter (Figures 3.4, 3.7 and 3.8). Seasonal variations in OC fluxes were smaler; the winter fluxes were about 30% of the spring and summer maxima (Figures 3.5 and 3.7). The proportion of OC in the setling flux peaked in the summer at about 15% in the upper traps, and dropped to 8% in the winter and early spring (Figure 3.8). OC contents were low in the spring when fluxes of organic mater were highest at station SN-0.8, the result of dilution by BSi. Changes in flux with depth were minimal for BSi and OC at station SN-0.8. In Chapter 3. Settling fluxes in Saanich and Jervis Inlets 66 the fall and winter, both tended to increased with depth, especialy from the shalow (50 m) to the mid-depth (135 m) traps (Figure 3.7). In the spring and summer, OC decreased slightly with depth, as did BSi in the summer. However, springtime fluxes of BSi were highest to the mid-depth traps (Figure 3.7). As at the other stations, the mass flux generaly increased with depth at station SN-0.8, largely due to increasing fluxes of lithogenic material in deep waters. Thus, while at 50 m the mass flux was determined mostly by BSi and OC fluxes, at depth the mass flux was more heavily influence by lithogenic debris (Figure 3.7). Contents of BSi and OC in the setling material usualy decreased with depth at station SN-0.8 (Figure 3.8). A l u m i n i u m fluxes With a remarkably regular patern and unlike at any other station, Al fluxes at station SN-0.8 were high in the winter and low in the summer (Figures 3.6 and 3.7, and Table 3.3). This seasonality occured at all depths, while Al fluxes increased by a factor of two from 50 m to 135 m and were nearly the same at 135 m and 180 m. As the Al flux showed an opposite seasonal trend to the mass flux, the contribution to the mass flux by aluminosilcates was particularly high in the winter. The proportion of aluminosilcates in the sinking particles at station SN-0.8 was 10-20% in the spring and summer and rose to more than 50% in the winter (Figure 3.8). The pronounced seasonality of the Al flux at the head of Saanich Inlet folowed that of local rainfal and flows from the Cowichan and Goldstream Rivers, and efectively rules out the Fraser River as a potential source of lithogenic particulate mater at station SN-0.8 because flow from the Fraser River peaks during the June freshet. The sources of Al to Saanich Inlet are discussed in section 3.4.4. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 67 C D C O O C D CO C O O ^ -»— CO L O C\J C D C D C O O O O O O O O O C\ j C \ i •»—' -r^ T~ O O O O ^Aepg.oi iv 6 Figure 3.6: Aluminium fluxes in Saanich Inlet. The gray lines are the curves of Figure 3.7, repeated annualy. There is a six-fold scale change between stations for the aluminium flux. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 68 ra E 3 fe Q. 2 o ra ra E " ° CM 5 4 3 2 1 0 20 15 10 v 4.5 E 3.0 w CQ o) 1.5 0.0 0.9 \ •a 0.6 CM E O ° 0.3 o.o 1.2 •o 0.8 E 3 0.4 0.0 i 1 1 1 1~ i 1 r . ~ i 1 1 1 1 1 1 1 1 r n 1 1 1 1 1 1 1 1 i i i r J F M A M J J A S O N D J F M A M J J A S O N D 5 4 3 2 1 0 4 3 2 1 0 1.5 1.0 0.5 0.0 0.3 0.2 0.1 0.0 0.15 0.10 0.05 0.00 Figure 3.7: Seasonaly-averaged sediment-rap fluxes.(total mass, BSi, OC and Al) in Saanich Inlet. The seasonal patern of primary production (Figure 2.8) at each station is included for comparison. The curves describing the fluxes at diferent depths were made by giving each day during the time series (Figures 3.3 through 3.6) the flux value measured on that day. The series was then sorted by Julian day and a seven-day smoothing was applied. Note the diference in scale between stations. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 69 8.0 h 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 J F M A M J J A S O N D J F M A M J J A S O N D Figure 3.8: Seasonaly-averaged compositonal characteristics of setling fluxes in Saanich Inlet. The curves were created as described for Figure 3.7. <513C and OC/N are flux-weighted averages. Multiplication of %OC and %A1 by 2.7 and 9.2, respectively, give estimates of %OM and % aluminosilcates (section 3.3.1). Chapter 3. Settling fluxes in Saanich and Jervis Inlets 70 S N - 9 S N - 0 . 8 sp su au wi annual sp su au wi annua l mass flux (g m - 2 d _ 1 ) Z l 6.12 7.80 5.19 3.59 5.67 2.75 1.79 0.936 1.04 1.63 Z2 11.3 13.4 10.1 9.50 11.1 3.51 2.11 1.41 2.31 2.33 Z3 11.0 18.7 11.5 9.81 12.7 3.19 1.95 2.00 2.45 2.40 B S i flux (mg m ~ 2 d _ 1 ) Z l 2400 2940 1420 608 1840 1430 757 284 208 669 Z2 3310 4030 2110 1360 2700 1530 756 324 390 750 Z3 3090 4500 2240 1380 2800 1320 736 324 351 680 O C flux (mg m " 2 d" 1 ) Z l 385 517 312 189 351 250 228 106 89.9 168 Z2 551 737 518 366 543 244 211 121 132 177 Z3 451 744 468 330 498 215 204 162 131 178 A l flux (mg m - 2 d _ 1 ) Zl 220 296 240 220 244 35.7 18.9 26.1 53.7 33.6 Z2 560 646 556 626 597 86.2 37.7 48.6 128 75.1 Z3 576 1060 659 649 735 81.8 39.1 62.1 126 77.3 <513C Z l -19.8 -19.6 -20.7 -22.0 -20.2 -19.6 -19.5 -20.6 -21.9 -20.0 Z2 -20.0 -19.9 -21.1 -22.2 -20.5 z 3 -20.6 -20.5 -21.2 -22.2 -21.0 -19.9 -19.7 -20.9 -22.0 -20.5 O C / N (molar) Zl 8.77 8.82 10.0 9.83 9.29 8.73 8.90 9.71 9.58 9.20 Z2 9.18 9.21 10.5 10.4 9.73 8.94 8.76 9.45 9.74 9.16 Z3 9.60 9.54 10.1 10.8 9.97 8.85 8.75 9.32 9.76 9.09 Table 3.3: Fluxes in Saanich Inlet at each sediment-trap depth (zi, z2 and z3; see Table 3.1 for depths of traps). Values were obtained by averaging the curves of Figure 3.7 over each seasonal period. 513C and OC/N were flux-weighted, sp = spring, su = summer, au = autumn and wi = winter. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 71 3.3.3 F luxes at the mouth of Saanich Inlet (station SN-9) Compared to those at station SN-0.8, The settling fluxes at station SN-9 inside the sill of Saanich Inlet were many times larger (Figures 3.3 through 3.7; Table 3.3) and the trapped material was rich in A l , depleted in BSi and especially OC-poor (Figure 3.8). Fluxes of biogenic silica and organic carbon generally followed autochthonous primary production (high in spring and summer, decreasing in fall and winter; Figures 3.4, 3.5 and 3.7), but the A l fluxes show that resuspension was significant throughout the water column and especially at the deep traps when replacement waters flowed over the sill and into the fjord basin in the summer and fall (Chapter 4). Year-round, fluxes of BSi and O C increased significantly from 45 m to 110 m. From 110 m to 150 m, however, changes with depth were less pronounced (Figure 3.7); on average, BS i fluxes increased slightly, and O C fluxes decreased between these depth intervals (Table 3.3). Aluminium and mass fluxes increased ~two-fold from shallow to mid-depth traps, but changed little between 110 m and 150 m, except during renewal periods when peaks in A l and mass fluxes were recorded at 150 m (Figure 3.7 and Table 3.3). Thus, it appears that turbulence across the sill (approximately 70 m) regularly delivered sediment to the water column at station SN-9 such that fluxes at 110 m and 150 m were similar while, during renewal events, fluxes to the deep traps were most affected. %BSi and % O C decreased with depth year-round at station SN-9 while %AI increased with depth, a sign of water column organic remineralisation and a preponderance of resuspended, reworked sediments in deep sediment traps. Deepwater renewal often overlapped with the period of high primary production at station SN-9 and, therefore, resuspended fluxes of BSi and O C clouded the sediment-trap signal of surface ecology. The resuspended fluxes tended to peak during the second half of the production season, broadening the biogenic flux maxima into the late summer and Chapter 3. Settling fluxes in Saanich and Jervis Inlets 72 fall. Upward transmission of the resuspension signal was common and examples of the efect on BSi and OC fluxes are the summer-fal periods of 1984, 1985 and 1986. During one or more deployments within this period of each year, Al fluxes were especialy high at 150 m and decreased upward (Figure 3.6). The BSi and OC fluxes (Figures 3.4 and 3.5) behaved similarly, but their resuspension signals were nearly lost within the seasonal cycle of surface export delivered to the mid-depth and shalow traps. Despite the combined and large influences of both local production and resuspension, the yearly cycles of BSi and OC fluxes (Figures 3.4 and 3.5) were very regular at all depths throughout the time series at station SN-9. Thus, without recognising the extent of resuspension at this location, the BSi and OC flux paterns could be interpreted as primary production signals. Chapter 4 provides a more detailed description of the resuspended fluxes in Saanich and Jervis Inlets. 3.3.4 F l u x e s i n J e r v i s In le t The nature of the setling fluxes at the two mooring sites in Jervis Inlet (JV-3 and JV-7) were suficiently similar that they are discussed together. B i o g e n i c fluxes As in Saanich Inlet, mass, BSi and OC fluxes to the 50 m sediment traps in Jervis Inlet folowed local primary production, increasing in March-April and decreasing in September-October (Figures 3.9, 3.10, 3.11 and 3.13). BSi fluxes varied by a factor of ~20 at the 50 m traps in Jervis Inlet, while OC fluxes at stations JV-3 and JV-7 varied by factors of ~3 and ~5, respectively (Figure 3.13). Again, the smaler seasonal variations in OC flux (compared with BSi fluxes) were due to the wintertime presence of terrestial OC (see section 3.4.2) and to the likelihood that flagellates were larger contributors to the marine OC flux in the winter. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 73 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure 3.9: Total mass fluxes in Jervis Inlet. The gray lines are the curves of Figure 3.13, repeated annually. Pooling both stations and all depths, lost samples from all or part of a mooring represented 11% of the time series in Jervis Inlet and all analyses were performed on all collected samples. Missing data can be inferred from gaps in these bar graphs. At both stations in Jervis Inlet, biogenic silica, organic matter and lithogenic debris all comprised >50% ofthe mass flux at certain times and depths (Figure 3.14). Organic matter often comprised a significant fraction of the settling debris at 50 m, but the seasonal pattern of O C content did not follow the pattern of O C flux (Figures 3.13 and 3.14). In fact, at station JV-3 , O C content was inversely related to O C flux at the 50 m traps. The A l fluxes at this location were very low year-round (Figure 3.12), as were the BSi fluxes in the winter. Thus, in the fall and winter, organic matter remained the Chapter 3. Settling fluxes in Saanich and Jervis Inlets 74 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure 3.10: Biogenic silica fluxes in Jervis Inlet. The gray lines are the curves of Figure 3.13, repeated annualy. dominant constiuent of the mass flux even though OM fluxes were minimal. Indeed, wintertime mass fluxes at the shalow traps of station JV-3 were the lowest of the time series (Tables 3.3 and 3.4), reflecting the smal lithogenic and diatomaceous contributions. In general, the magnitude of the BSi fluxes was similar from year to year and the seasonality was clear at al depths (Figure 3.10). Although spring and summer fluxes were relatively constant with depth, in the autumn and winter BSi fluxes increased with depth (Figure 3.13), possibly due to resuspension of BSi-rich sediments. During spring and summer, there were periods when fluxes of biogenic silica decreased and then rebounded (Figure 3.10). These fluctuations lagged the early-summer luls in primary production (Chapter 2) by about two to four weeks (Figure 3.13). If the summertime drops in the two time series are not caused by sample biasing and are not an artifact Chapter 3. Settling fluxes in Saanich and Jervis Inlets 75 of the averaging schemes, then this lag implies that the early-summer lull in production resulted in a decrease in the BSi flux (and OC flux; see Figure 3.13) several weeks later. Sanceta (1989a) found evidence of increased grazing in July and August in Saanich and Jervis Inlets, while Sanceta (1989b) showed that fecal pelets and BSi fluxes were reduced at this time; if grazing was the cause of the mid-summer production luls, it may have resulted in smaler fluxes of BSi to the sediment traps. Fecal pelets are important vectors of downward transport, but grazing can decrease the amount of material exported from the euphotic zone as fecal pelets are largely decomposed in the water column (Smetacek, 1980b; Jumars et al., 1989; Sanceta, 1989b; Noji et al., 1991) and grazing by zooplankton (e.g.; Harison et al., 1983; Bornhold, 2000) and heterotrophic dinoflagelates (Jacobson and Anderson, 1986; Buck and Newton, 1995; Tiselius and Kuylenstierna, 1996) can control diatom growth, biomass and, ultimately, export flux. The seasonality of the OC fluxes dampened significantly with depth (Figures 3.11 and 3.13; Table 3.4). Although for most depth intervals in Jervis Inlet OC fluxes re-mained relatively constant, from 50 m to 200 m at station JV-7, the summertime OC fluxes decreased substantialy, while at 450 m the seasonal cycle was absent (Figures 3.11 and 3.13). In 1985 and 1986 at station JV-7, the spring and summer fluxes of OC at 50 m were significantly higher than the seasonal mean, although the deeper OC fluxes were similar to other years (Figure 3.11). Primary production was also high in 1986 (section 2.4.1), but the high OC fluxes at 50 m were not accompanied by BSi (Figures 3.10). The high OC fluxes, therefore, may have been caused by the depositon of non-diatomaceous phytoplankton caught at 50 m and largely remineralised or lateraly transported before sinking to the depth of 200 m. If lateral transport were to have caused these changes in flux between 50 m and 200 m, then the horizontal gradients in the export flux must have been very large (e.g.; Siegel and Deuser, 1997; Timothy and Pond, 1997). Because high primary production in 1986 was recognised at the other stations, it is likely the Chapter 3. Settling fluxes in Saanich and Jervis Inlets 76 CO •o CvJ E o o 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure 3.11: Organic carbon fluxes in Jervis Inlet. The gray lines are the curves of Figure 3.13, repeated annualy. production signal of 1986 was a regional event and horizontal gradients in the export flux were not significantly diferent than in other years. Lateral transport, therefore, probably did not cause the large depth changes in OC fluxes between 50 m and 200 m. Figure 2.6 suggests that flagellates were relatively more abundant at station JV-7 than at the other mooring sites; thus, blooms of phytoplankton other than diatoms could explain the OC fluxes to the shalow traps in 1985 and 1986. However, another possibilty for the high OC fluxes is that 'swimmers' were atracted to the 50 m sediment traps during the sum-mers of 1985 and 1986. Blooms of pelagic polychaetes were sometimes observed during the experiment and they were occasionaly caught within the grids of the sediment traps. Perhaps the high OC fluxes were in fact polychaete remains. Corelations between resid-uals of primary production and of sediment-trap flux have been determined and it was Chapter 3. Settling fluxes in Saanich and Jervis Inlets 77 found that at all stations, BSi flux residuals were not strongly correlated with the pro-duction residuals. At stations SN-9, SN-0.8 and JV-7 , O C flux residuals and production residuals also were not significantly correlated but, at station J V - 7 , there was a positive relationship between these residuals (Figure 3.15), lending support to the possibility that the high 50 m O C fluxes of 1985 and 1986 were the result of flagellate blooms. A l u m i n i u m fluxes At both stations in Jervis Inlet, A l flux (Figures 3.12 and 3.13) and A l content (Fig-ure 3.14) increased with depth. A l fluxes showed little seasonality, except that A l as-sociated with turbulent resuspension during deepwater renewals in the fall was caught in the deeper traps (Chapter 4). 50-m A l fluxes were "smallest at station JV-3 , as the mooring was within a large basin and was farther from riverine sources than station J V - 7 (Figures 1.1 and 1.3). A l fluxes in Jervis Inlet were a good tracer of 'additional' fluxes, which are discussed in more detail in Chapter 4. The high A l fluxes in the fall of 1985 and recorded for station JV-3 (Figure 3.12) occurred when the traps at the mouth of Jervis Inlet were moored at station JV-11.5 (Figures 1.1 and 1.3; Table 3.1). At 50 m and 300 m, the total flux did not deviate substantially from the seasonal averages but, at 600 m, the traps at station JV-11.5 collected a large amount of Al-rich material (Figures 3.9 and 3.12). Although there was a deepwater renewal occurring at the time (Chapter 4) and similarly high fluxes might have been recorded at station JV-3 , the high fluxes at 600 m at station JV-11.5 may have been caused by localised slumping or resuspension and, therefore, may not represent the fluxes that were occurring at station JV-3 . Chapter 3. Settling fluxes in Saanich and Jervis Inlets 78 0.15 0.10 0.05 0.00 0.10 0.05 is o.oo < 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 600 m 450 m J 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure 3.12: Aluminium fluxes in Jervis Inlet. The gray lines are the curves of Figure 3.13, repeated annualy. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 79 1 o 0.9 0.6 0 .3 0.0 0.3 0.2 0.1 0.0 0 .18 " ° 0 .12 CM E < cn 0 .06 0 .00 JV-7 1 1 1 r ~i 1 r - 50 m • 200 m - 450 m i i i r "T 1 r i i 1 1 1 1 1 1 1 1 1 r—i 1 1 1 1 1 i i i i i i r J F M A M J J A S O N D J F M A M J J A S O N D Figure 3.13: Seasonaly-averaged sediment-trap fluxes (total mass, BSi, OC and Al) in Jervis Inlet. The seasonal patern of primary production (Figure 2.8) at each station is included for comparison. The curves describing the fluxes at diferent depths were made by giving each day during the time series (Figures 3.9 through 3.12) the flux value measured on that day. The series was then sorted by Julian day and a seven-day smoothing was applied. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 80 Figure 3.14: Seasonaly-averaged compositonal characteristics of setling fluxes in Jervis Inlet. The curves were created as described for Figure 3.13. 513C and OC/N are flux-weighted averages. Multiplication of %OC and %A1 by 2.7 and 8.7, respectively, give estimates of %OM and % aluminosilcates (section 3.3.1). Chapter 3. Settling fluxes in Saanich and Jervis Inlets 81 JV-3 J V - 7 sp su au wi annual sp su au wi annual mass flux (g m" 2 d" Zl 1.61 1.27 0.743 0.521 1.04 1.56 1.21 0.762 0.688 1.05 Z2 1.90 1.71 1.03 1.02 1.42 1.37 1.24 0.955 0.894 1.11 Z3 2.36 2.49 2.77 1.90 2.38 1.78 1.61 1.83 1.66 1.72 BSi flux (mg m - 2 d"1) Zl 875 598 224 147 460 535 450 181 78.1 311 Z2 769 647 273 196 471 474 456 231 90.3 313 Z3 800 725 580 337 611 519 472 349 227 392 O C flux (mg m" 2 d"1) Z l 174 173 150 80.6 144 268 222 110 93.3 173 Z2 141 157 107 99.5 126 120 127 87.8 80.3 104 Z3 136 162 161 121 145 113 116 122 111 116 A l flux (mg m- 2 d" Z l 16.3 10.8 14.5 16.0 14.4 25.8 12.5 27.1 31.9 24.3 z 2 54.1 46.0 41.4 51.2 48.2 41.7 26.8 40.4 51.7 40.2 z 3 85.3 99.0 141 100 107 71.7 53.3 85.5 85.4 74.0 613C i Z l -21.7 -21.0 -22.5 -22.5 -21.8 -22.4 -21.6 -23.7 -23.2 -22.5 Z2 z 3 -21.7 -21.0 -21.8 -22.1 -21.6 -22.3 -21.7 -22.6 -23.1 -22.4 O C / N (molar) Zl 9.53 9.66 10.2 11.0 10.1 8.83 9.19 10.1 9.52 9.36 Z2 10.0 10.0 10.3 10.4 10.2 9.63 9.82 10.5 10.9 10.2 Z3 10.2 10.3 9.93 10.4 10.2 10.3 10.1 10.2 10.6 10.3 Table 3.4: Fluxes in Jervis Inlet at each sediment-trap depth (zi, z 2 and z 3 ; see Table 3.1 for depths of traps). Values were obtained by averaging the curves of Figure 3.7 over each seasonal period. <513C and O C / N were flux-weighted, sp = spring, su = summer, au = autumn and wi = winter. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 82 1.5 1.0 nj •g E 0.5 I • • • 1 1 1 • • • • • • • • • • • * -1* • * * m ^ m i • r = 0.47 1 1 1 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 production residual Figure 3.15: Production and 50 m O C flux residuals at station JV-7 . Production resid-uals are the average of the production measurements at the beginning and end of each deployment minus the average for the deployment period from the curve of Figure 2.8. Flux residuals are values of Figure 3.11 minus the appropriate portion of the curve of Figure 3.13. Residuals are normalised to the mean production or O C fluxes for each deployment period. 3.4 Discussion 3.4.1 O C - B S i relationships Figure 3.16 shows settling fluxes of organic carbon plotted against fluxes of BSi . The data are separated into 'summer' and 'winter' regimes. For the landward stations (SN-0.8 and JV-7) in 1986, Sancetta (1989b) found that fecal pellets in the traps deployed between Apri l and September were pale green, while the pellets trapped between October and March were brown. She suggested the reason was not only changes in the zooplankton community, but also a shift from organic-rich to detrital sources of energy for the graz-ers. These time intervals (April-September, October-March) were used for the seasonal separation of the data in Figure 3.16. The slopes of Figure 3.16 are interpreted as average ratios ofthe diatomaceous O C and Chapter 3. Settling fluxes in Saanich and Jervis Inlets 83 • shallow • middle • deep filled = summer hollow = winter =0.18 60 50 40 30 20 10 J V - 7 R o o n , , ' w: 1^ =0.59 OC = 1.2 BSi + 4.6 s: OC = 0.84 BSi + 4.8; r*=0.25 i i i ' ' I i i i 10 15 20 25 0 10 15 20 120 100 80 60 40 20 0 • 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i 1 1 1 i s: OC = 0.85 BSi + 4.5; 1^ =0.72 w: OC = 1.3 BSi + 3.0; 1^ =0.75 I i i i I i i i I 40 0 20 40 60 80 100 120 ' 1 1 1 i 1 1 1 1 i 1 1 1 1 i 1 1 1 s: OC = 0.63 BSi + 8.2; r*=0.61 w: OC = 1.7 BSi + 2.5; ^=0.53 \ T 30 40 mmol BSi m d 2 ^ 1 - 1 Figure 3.16: OC fluxes versus BSi fluxes. Data from summer (s) and winter (w) were separated. Points within doted regions were not included in regressions. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 84 BSi (OCd:BSi ratios) for the material reaching the sediment traps. These molar ratios (approximately 1) are much lower than the O C d : B S i ratio of about 7.7 for living diatoms (Brzezinski, 1985) and are indicative of intensive recycling of diatomaceous carbon rela-tive to BSi . Glycine and serine are concentrated in the cell walls of diatoms (Hecky et a l , 1973), and studies (Burdige and Martens, 1988; Cowie and Hedges, 1992) have shown these amino acids to have a longer residence time in the marine environment than intra-cellular amino acids, suggesting preferential preservation of the cell wall proteins. Not only will zooplankton grazing and cell lysis cause preferential loss of O C relative to BSi , dinoflagellates that phagocytise (Buck and Newton, 1995; Tiselius and Kuylenstierna, 1996) or extracellularly digest (Jacobson and Anderson, 1986) diatom chains while leav-ing the frustules intact also provide an efficient means to produce low O C d i B S i ratios in sinking material. The intercepts of Figure 3.16 indicate that a significant portion of the settling organic material in Saanich and Jervis Inlets was non-diatomaceous. The non-diatomaceous O C would have been composed of both terrestrial O C and marine O C from sources such as nanoflagellates, dinoflagellates, heterotrophic plankton including bacteria, and transparent exopolymer particles ( T E P ; Alldredge et al., 1993). 3.4.2 M a r i n e a n d t e r r i g e n o u s O C fluxes Figures 3.17 and 3.18 show the time series of <513C at each station during the study (only at SN-0.8 were the mid-depth samples analysed for 1 3 C / 1 2 C ) . The isotopic signatures, summarised in Figures 3.8 and 3.14, and in Tables 3.3 and 3.4, were heavy in spring and summer and light in autumn and winter. In their seminal work, Sackett and Thompson (1963) noted that terrestrial plants are enriched in 1 2 C (isotopically light or 513C values that are more negative) when compared to marine plants, and suggested that the progres-sively heavier isotopic composition of sediments away from the mouth of the Mississippi River is the result of a decreasing presence of terrigenous O M . Although these data have Chapter 3. Settling fluxes in Saanich and Jervis Inlets 85 since been reinterpreted based on contributions to marine sediments by degraded, iso-topically heavy, terrestrial C 4 grasses (Gohi et al., 1997; Gofii et al., 1998), in the Pacific Northwest, C 4 plants are rare (Teeri and Stowe, 1976). Indeed, changes in particulate f5 1 3C due to variable proportions of marine and terrestrial organics are consistent with the trends of marine and terrestrial biomarkers in Saanich Inlet (Cowie et al., 1992) and the Washington margin (Prahl et al., 1994). However, marine and terrestrial organic (5 1 3C endmembers vary regionally (Prahl et ab, 1994), so that 8l3C endmembers appropriate for Saanich and Jervis Inlets should be determined if the <513C data are to be used to separate terrestrial from marine organics. Cowie (unpublished) has measured r513C on soils collected from the banks of Saanich Inlet and Jervis Inlet. From his data, a best estimate for the organic matter of soils is -25.1°/oo in Saanich Inlet and -26.5°/oo in Jervis Inlet. These values are within the general range expected for terrestrial C 3 plants (Deines, 1980) and are saddled by the terrestrial (51 3C endmember (-25.7°/oo) used by Prahl et al. (1994) to describe Columbia River basin sediments. The different endmembers observed for the fjords might be explained by the biogeoclimatic zones they occupy. Jervis Inlet is located in the coastal western hemlock (CWH) zone which covers most of the coast of British Columbia, penetrating into Alaska to the north and along the coasts of Washington and Oregon to the south (Pojar et al., 1991). The C W H zone is the rainiest biogeoclimatic zone of British Columbia with a cool, mesothermal climate where nutrient leaching from the mineral soils is rapid (Pojar et al., 1991). Saanich Inlet and portions of the Cowichan River watershed are located in the coastal Douglas-fir ( C D F ) zone, a stretch of low-elevation terrain covering the southeastern coast of Vancouver Island and many islands of the southern Strait of Georgia (Nuszdorfer et al., 1991). The C D F zone is in the rainshadow of Vancouver Island and the Olympic mountains, and is warmer and drier than the C W H zone (Nuszdorfer et al., 1991). Given the differences in the biogeoclimatic zones of these fjords, it is reasonable Chapter 3. Settling fiuxes in Saanich and Jervis Inlets 86 Figure 3.17: 8l3C of the trapped organic matter of Saanich Inlet. Carbon isotopes were not measured on sediments trapped at mid-depth at station SN-9. The gray lines are the curves of Figure 3.8, repeated annually. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 87 ' " i i i — i 1 1 i 1 " 1 i 1 . JV-3 __ _ -_ JV-7 50 m 50 m - _ -I I I I I ' I I"1 ' " i 1 . __ : 4 5 0 m - '. 600 m ; "- : 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure 3.18: r5 1 3C of the trapped organic matter of Jervis Inlet. Carbon isotopes were not measured on samples collected at the mid-depth traps. The gray lines are the curves of Figure 3.14, repeated annually. to expect that the residence time of the O M in the soils surrounding Saanich Inlet is longer than that for the O M of Jervis Inlet soils. The potential of a longer soil residence time and the warmer temperatures of the C D F zone may promote a greater degree of O M recycling within the Saanich Inlet/Cowichan Valley watershed, resulting in heavier terrestrial S13C as 1 2 C is preferentially respired. (51 3C of marine O C varies significantly, but a median value at temperate latitudes is about -20°/oo (Deines, 1980). During photoautotrophic carbon fixation, the resulting isotopic composition of the phytoplankton is affected by <513C of the dissolved inorganic carbon pool and a number of physiological and environmental factors that affect biological fractionation (for a summary, see Kukert and Riebesell, 1998). Indeed, Hedges et al. (1988a) measured S13C on plankton tows from Dabob Bay and found values ranging between -26.0°/oo and -19.5° /oo, Kukert and Riebesell (1998) found that <513C of suspended P O C increased from -25%o to -20°/oo throughout the spring bloom of predominantly Skeletonema costatum in the Norwegian fjord Lindaspollene, and Rau et al. (2001) Chapter 3. Settling fluxes in Saanich and Jervis Inlets 88 atributed fluctuations in 513C of particulate OC between about -28°/oo and -16°/oo in Monterey Bay, California to variable photosynthetic 13C fractionation. For the sediment-trap samples from Saanich and Jervis Inlets, the corelations between 513C and BSi content (Figure 3.19) were used to estimate the marine <513C endmembers of -17.3°/oo for Saanich Inlet, and -19.6°/oo in Jervis Inlet (Appendix A). During carbon assimilation, phytoplankton fractionate 12C from 13C less efectively (becoming isotopicaly heavier) as growth rate increases and [C02]aq decreases (Laws et al., 1995; Bidigare et al., 1997; Burkhardt et al., 1999; Tortel et al., 2000). Furthermore, diatoms are isotopicaly heavier than other phytoplankton (Fry and Wainright, 1991; Pancost et al., 1997; Kukert and Riebesel, 1998; Reinfelder et al., 2000), as are large cels when compared to smal cels (Pancost et al., 1997; Popp et al., 1998). Thus, when the phytoplankton of Saanich Inlet are compared to those of Jervis Inlet, higher growth rates (as inferred from higher nutrient supply and primary production {Chapter 2; Timothy and Soon, 2001}), a greater predominance of diatoms (Sanceta, 1989a; 1989b; Chapter 2; Timothy and Soon, 2001) and larger diatoms (Sanceta, 1989a) could explain the heavier marine <513C endmember in Saanich Inlet. The large diference between these marine <513C endmembers for Saanich and Jervis Inlets (-19.6%o to -17.3°/oo) and <513C of the net tow samples from Dabob Bay (-26.0°/oo to -19.5%o; Hedges et al, 1988a) and of filtered POC of Lindaspolene (-25%o to -20°/oo; Kukert and Riebesel, 1998) must be addressed, as it is difficult to invoke large regional diferences in plankton 513C for these similar environments. The size-fractionated samples of Hedges et al. (> 64 pm) and Kukert and Riebesel (< 20 pm and > 20 pm) represent suspended material, while this study sampled the sinking material. Kukert and Riebe-sel (1998) found the > 20 pm size fraction (predominately diatoms) to be several per mil heavier than the < 20 pm size fraction (mostly flagellates) and, furthermore, they found heaviest values when particulate OC concentrations (i.e.; Skeletonema biomass) Chapter 3. Settling fluxes in Saanich and Jervis Inlets 89 O CO "to shift of 5 1 3 Cdue to non-diatomaceous, marine OM o co ""to shift of 8 1 3 Cdue to non-diatomaceous, marine OM 10 20 30 40 50 60 70 80 90 100 %BSi Figure 3.19: <513C versus %BSi in Saanich and Jervis Inlets. With estimates of the terrige-nous <513C endmembers, these corelations are used to find the marine 513C endmember in each fjord. The equations of the regressions and the way they are used in this exercise is explained in Appendix A. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 90 terr igenous O C % of total O C at shallow trap sp su au wi a n n u a l SN-9 SN-0.8 J V - 3 J V - 7 32 30 43 61 29 28 42 59 31 20 43 43 40 29 59 53 37 35 32 41 Table 3.5: Terrigenous O C in Saanich and Jervis Inlet. These estimates were made using <513C of Tables 3.3 and 3.4, and the marine and terrigenous r5 1 3C endmembers of -17.3°/oo and - 2 5 . 1 ° / 0 0 in Saanich Inlet, and -19.6%o and -26.5°/oo in Jervis Inlet. were highest. The isotopically heavy values associated with the large size fraction at the peak and end of the bloom (-21°/oo to -20°/oo; Kukert and Riebesell, 1998) are the most likely signals to be transmitted to the settling flux, significantly decreasing the appar-ent discrepancy between the marine <513C endmembers identified here (Table A . l ) and the time-course of particulate O C <513C in Lindaspollene. The sediment-trap samples also represent material that was likely recycled to some degree, and therefore potentially frac-tionated towards heavier values, as 1 3 C accumulates in higher trophic levels (DeNiro and Epstein, 1978; Wada et al., 1987). Finally, just as terrigenous organic debris contributes to the settling fluxes in coastal environments, it is likely to be suspended throughout the water column so that the records of Kukert and Riebesell (1998) and Hedges et al. (1988a) may be somewhat lighter than if only marine O M had been sampled. The net tow samples from Dabob Bay did not contain microscopically visible vascular plant debris (Hedges et al., 1988a), but flocculated terrestrial humic substances (Sholkovitz, 1976) are likely present in all these fjords. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 91 total primary production (diatom production) mg C m" 2 d - 1 sp su au wi annual SN-9 2740 (2160) 2880 (2280) 609 (481) 119 (94) 1580 (1250) SN-0.8 1390 (1100) 2160 (1710) 621 (491) 207 (164) 1090 (861) JV-3 1600 (1120) 1650 (1160) 280 (196) 157 (110) 921 (645) J V - 7 1150 (805) 1060 (742) 233 (163) 128 (90) 641 (449) Table 3.6: Total and diatom production at each station. Total production is from aver-aging Table 2.1 over each season. Diatom production is estimated as 79% and 70% of the total production in Saanich and Jervis Inlets, respectively (Table A.l). The marine and terrestrial 513C endmembers have been used to estimate the ter-rigenous contributions to the setling OC in the shalow sediment traps (Table 3.5). Considering the various erors of the estimates, the annualy-averaged terigenous con-tributions were similar at each station (30-40%). In spring and summer, terrestrial OM comprised 20-40% of total OM, while in the fall and winter it was 40-60% of the total. Ancilary information from the exercise in Appendix A is the compositon of the 'typical' marine sample from each fjord (Table A.l). In Saanich Inlet, this marine sample was 6% BSi, 27% diatomaceous OM and 7% non-diatomaceous OM. In Jervis Inlet, it was 65% BSi, 24% diatomaceous OM and 1% non-diatomaceous OM. Comparing the pro-portions of diatomaceous and non-diatomaceous OM, 79% of the marine organic mater was diatomaceous in Saanich Inlet and 70% in Jervis Inlet. These estimates are not sig-nificantly diferent (Table A.l), although they are consistent with the suppositon that Chapter 3. Settling fluxes in Saanich and Jervis Inlets 92 diatoms were more prevalent in Saanich Inlet. This result supports the generality made by Nelson et al. (1995) that 75% of primary production in nutrient-replete waters is by diatoms, and allows estimates of diatomaceous carbon assimilation in Saanich and Jervis Inlets (Table 3.6). However, because translation from the settling flux of diatomaceous carbon to diatom primary production assumes similarity in the recycling of diatoma-ceous and non-diatomaceous marine organic matter, the estimates of diatom production of Table 3.6 should be viewed as rough approximations. The exercise in Appendix A has been carried out for various sub-sets of the time series. When the entire data set (all four stations) or some portion of it (one or two stations) was separated by season ('winter' and 'summer' as defined section 3.4.1), <513C of the marine endmember was not different for the two periods because the correlation between <513C and BSi content was low. Different marine S13C endmembers at any two stations (e.g.; SN-9 and SN-0.8) were not observed for the same reason. The O C : N ratio of marine phytoplankton of 6.6 (Fleming, 1940; Redfield et al., 1963) is a broad average that can be used over large spatial scales, but locally varies because of factors including species composition and nutrient availability (Parsons, 1961; Redfield et al., 1963; Sakshaug, 1977; Sakshaug et al., 1983; Sakshaug and Olsen, 1986; Sakshaug et al., 1989; Wong et al., 1999). Nevertheless, the O C : N ratio of marine O M tends to be lower than the O C : N ratio of terrestrial O M (e.g.; Hedges et al., 1986), and can sometimes be used as an indicator of marine and terrigenous O M fractions (e.g.; Hedges et al., 1988b; Prahl et al., 1994; Ruttenberg and Goni, 1997). Indeed, the O C : N ratio of the settling material in Saanich and Jervis Inlets was high in the fall and winter and lower in the spring and summer (Figures 3.8 and 3.14), correlating with the relative abundance of marine and terrigenous O M based on SnC endmembers (Table 3.5). However, the poor correlation between r5 1 3C and O C / N (Figure 3.20), especially for Jervis Inlet, indicates that one or both of these compositional traits was affected by factors other than the Chapter 3. Settling fluxes in Saanich and Jervis Inlets 93 -27 -26 -25 -24 -23 -22 -21 -20 -19 -18 -17 -16 5 1 3 C Figure 3.20: SUC versus OC/N in Saanich and Jervis Inlets. A point (OC/N = 14.7, <513C = -21.0) from Jervis Inlet is out of range and not included in the regression. The choice of Model I regression or Model II regression (used here) is crucial for descriptions of data with low r values. The common Model I regression (|r|_1 less steep than these Model II regressions; section 3.2.4) would result in marine and terigenous OC/N endmembers with little physical meaning and, statisticaly speaking, is interpreted as a lack of confidence in predicting OC/N endmembers from 513C endmembers. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 94 marine:terrigenous O M ratio. Although the low correlation between r5 1 3C and %BSi in Jervis Inlet (Figure 3.19 and Appendix A) may have been caused by larger variability in the relative contributions of flagellates and other non-diatomaceous marine plankton to the settling flux than occurred in Saanich Inlet, it is also likely that the terrigenous and marine O C / N and 513C endmembers were not temporally constant. The exercise to estimate <513C of the marine O M endmember (Figure 3.19 and A p -pendix A) could not be performed using a correlation between O C / N and %BSi because the O C : N ratio of terrigenous O M is not known. (In the exercise of Appendix A , a mea-sure of <513C in the terrigenous O M fraction was required to estimate the marine <513C endmember.) However, using the <513C endmembers that have been described for each fjord (Figure 3.19), endmember values for the O C : N ratio of marine and terrigenous O M can be estimated (Figure 3.20). The overall agreement in the O C / N endmembers of Figure 3.20 with expected values (e.g.; Prahl et al., 1994) suggests that r5 1 3C and O C / N can be used as rough guides of marine and terrigenous O M in the sedimentary record of these fjords (e.g.; Tunnicliffe, 2000), while they should be used cautiously for sampling that represents short time periods. Figure 3.20 also demonstrates that these proxies of marine and terrigenous contributions are better suited for some environments (e.g.; Saanich Inlet) than others (e.g.; Jervis Inlet). Indeed, off the west coast of Vancouver Island, settling fluxes followed primary production and were highest in the spring and summer, but O C : N ratios of the sinking material were lowest in the winter (Pefia et al., 1999). Pefia et al. (1999) attributed this signal to inorganic nitrogen adsorbed onto clays, and to the possibility of reduced organic degradation caused by ballast-mediated, rapid sinking (e.g.; Ittekkot, 1993). Chapter 3. Settling fluxes in Saanich and Jervis Inlets 95 3.4.3 Export ratios of O C and B S i The e-ratio (the ratio of the setling flux of an element to the photoautotrophic assim-ilation of that element; defined for organic carbon by Downs, 1989) is a measure of the degree of recycling and remineralisation that has occured since the element was assim-ilated. In sections 3.3.2 through 3.3.4, OC and BSi fluxes are presented and, in section 3.4.2, terigenous and marine OC are separated and estimates of the diatomaceous con-tribution to total primary production are made. From these data, Table 3.7 gives e-ratios for both total and marine OC. OCto e-ratios have less oceanographic signifcance than marine OC e-ratios, but they can be compared to other data sets where terigenous OC has not been subtracted. The BSi e-ratios of Table 3.7 are relative to both total and diatom production. While (BSi flux/total production) has ecological and biogeochemical significance, (BSi flux/diatom production) is a beter measure ofthe degree of recycling and dissolution of reactive silicon. Because the OC e-ratio is a C:C ratio, its hypothetical maximum is one. E-ratios for biogenic silica are Si:C molar ratios and the maximum is set by the assimilation of these elements by phytoplankton. Brzezinski (1985) found Si:C ratios of various species of laboratory diatoms ranging between 0.04 and 0.42. BSi:POC molar ratios as high as 1.75 in Antarctic surface waters were due to the presence of heavily silicified diatoms, and possibly to more rapid recycling of POC (Queguiner et al., 1997). For both OC and BSi, the annual e-ratios of Table 3.7 are flux weighted, so they most closely reflect the ratios of the spring and summer when fluxes were highest. The spring and summer OCmar e-ratios of about 0.1 appear low when compared to the f-ratios that might be expected in these highly productive waters with large nutrient supply (Eppley and Peterson, 1979; Piatt and Harrison, 1985; Harison et al., 1987; Wassmann, 1991). The f-ratio is the ratio of new nutrient assimilation to total nutrient assimilation (sensu Chapter 3. Settling fluxes in Saanich and Jervis Inlets 96 O C * . e - r a t i o — o c w t 50 m flux w v ^ t o t e ratio t n t n r f ) d (OC m a r e-ratio tot prod O C m a r 50 m fluX tot prod ) C:C ratio BSi e-ratio = B ^ 5 t ° ™ o d f l u x (BSi e-ratio = BSi:C molar ratio sp su au wi annual sp su au wi annual SN-9 0.14 (0.096) 0.18 (0.13) 0.51 (0.29) 1.6 (0.62) 0.22 (0.14) 0.16 (0.20) 0.18 (0.23) 0.42 ( 0.53) 0.91 (1.1) 0.21 (0.26) SN-0.8 0.18 (0.13) 0.11 (0.076) 0.17 (0.098) 0.43 (0.18) 0.15 (0.10) 0.18 (0.23) 0.062 (0.079) 0.081 (0.10) 0.18 (0.23) 0.11 (0.14) JV-3 0.11 (0.075) 0.10 (0.084) 0.53 (0.31) 0.51 (0.29) 0.16 (0.11) 0.097 (0.14) 0.065 (0.092) 0.14 (0.20) 0.17 (0.24) 0.089 (0.13) J V - 7 0.23 (0.14) 0.21 (0.15) 0.47 (0.19) 0.73 (0.34) 0.27 (0.16) 0.083 (0.12) 0.076 (0.11) 0.14 (0.20) 0.11 (0.16) 0.086 (0.12) Table 3.7: Export ratios of O C and BSi at 50 m at each station. 50 m fluxes are from Tables 3.3 and 3.4. Total and diatom production are from Table 3.6. While O C t o t e-ratios consider the total O C flux, O C m a r e-ratios subtract terrigenous O C contributions (Ta-ble 3.5). While the first values may be compared to other studies, values in parentheses are ecologically more meaningful. Dugdale and Goering, 1967; Eppley and Peterson, 1979), and in steady state conditions is a predictor for the export of that nutrient in organic (dissolved and particulate) form. Therefore, for steady state and assuming lateral transport does not export significant amounts of organic material, f-ratios should be greater than e-ratios by the amount of dissolved and suspended matter that is exported, and by the amount of remineralisation occurring between the base of the euphotic zone and the depth of the sediment traps. A number of factors can affect* the amount of organic matter caught by sediment traps. D O C held interstitially within sinking particles can be a significant fraction of the total O C flux (Noji et al., 1999) and was not quantified in this experiment, and swimmers can unpredictably affect measured fluxes of organic matter (Lee et al., 1988; K a r l and Chapter 3. Settling fluxes in Saanich and Jervis Inlets 97 Knauer, 1989; Michaels et al., 1990). Based on considerations of fluid flow (Hargrave and Burns, 1979; Lau, 1979; Gardner, 1980b; Butman et al., 1986; Hawley, 1988), variable trapping efficiency has been documented in controlled experiments (Hargrave and Burns, 1979; Gardner, 1980b; Gardner, 1985; Butman, 1986; Gardner and Zhang, 1997) and in the field (Gardner, 1980a; Blomqvist and K 0 f o e d , 1981; Baker et al., 1988; Laws et al, 1989; Buesseler, 1991; Honjo et al., 1992; Gust et al., 1992; Gust et al., 1994; Nodder and Alexander, 1999; Buesseler et al., 2000; Yu et al., 2001). Although over-collection has been observed (e.g.; Buesseler et al., 1994), low trapping efficiency is more commonly observed. Poor trapping efficiency is especially common in surface waters, partly because the sinking debris is unconsolidated. As particulates sink, they are biologically and physically repackaged with the result that sinking rates tend to increase with depth. Syvitski et al. (1985) found that sinking rates increase with depth, and several authors have suggested trapping efficiency improves in deep waters at least partly because of the increased fall velocities and greater consolidation ofthe settling debris (Smetacek et al., 1978; Timothy and Pond, 1997; Yu et al., 2001). Some of the explanations for the lack of a difference in measured fluxes of organic matter by the sediment traps with and without sodium azide as a preservative (section 3.2.3) involve the loss of labile organic matter between the time of interception and measurement in the laboratory. Iseki et al. (1980) described a sediment-trap experiment in Patricia Bay of Saanich Inlet. The absolute OC fluxes they reported for traps moored at 50 m (2 to 3 times lower than the fluxes collected at a similar season and depth and given in Table 3.3 for stations SN-0.8 and SN-9) are not comparable to the fluxes of this study because their experiment occurred in a relatively isolated portion of Saanich Inlet. However, Iseki et al. (1980) determined rate constants of decay for the OC collected at 50 m to be between 0.014 and 0.029 day - 1 . They also derived an equation to correct for in situ loss knowing the rate constant of decay and the length of a deployment. Using their Chapter 3. Settling fluxes in Saanich and Jervis Inlets 98 rate constants for 50 m (0.014 0.029 d a y - 1 ) and their equation 6, measured O C fluxes should be multiplied by a factor of 1.2 to 1.5 to obtain the true flux caught by sediment traps moored for 30 days. Others have also shown organic matter to leach into the surrounding sediment-trap solution (Knauer et al., 1984; Kar l et al., 1988) and Knauer et al. (1990) found that more than half of the organic nitrogen contained within sediment traps was dissolved. Furthermore, Kumar et al. (1996) found that as much as 70% of organic matter was rapidly lost during even short sediment trap deployments. Thus, the apparently low e-ratios for organic carbon (Table 3.7) are likely caused by the presence of interstitial D O M , low trapping efficiency and organic leaching after interception. O f these possibilities, trapping efficiency may not have been a serious problem, as it appears the majority of sinking diatom frustules was trapped (discussed below). To correct for possible O C leaching, multiplication of the spring and summer carbon fluxes by a factor of two to four (Knauer et al., 1990; Kumar et al., 1996) would make the O C e-ratios of Table 3.7 closer to the expected values for these highly productive fjords. In light of these complications, it is difficult to interpret the higher e-ratios in the fall and winter for both marine and total O C (Table 3.7). They may reflect less O M recycling and remineralisation or better trapping efficiency due to a shift from uncon-solidated diatom aggregates to a greater proportion of fecal pellets (Sancetta, 1989a, 1989b, 1989c). Lithogenic ballast in the fall and winter (Ittekkot, 1993) might facili-tate decreased water-column remineralisation and increased trapping efficiency, although the flux of aluminosilicates was not notably seasonal, except at station SN-0.8 (Fig-ures 3.6, 3.7, 3.12 and 3.13). Organic matter in fecal pellets might also be less likely leached while in the sediment traps or during rinsing and centrifugation after retrieval. It is also possible that a large proportion of resuspended O C reached the shallow traps during the fall and winter. The range of O C t o t e-ratios (0.24 to 0.55) presented by Wass-mann (1990) for coastal waters of the north Atlantic including many fjords and bays is Chapter 3. Settling fluxes in Saanich and Jervis Inlets 99 generally higher than the spring and summer values from Saanich and Jervis Inlets, but similar to, or lower than, the autumn and winter values (Table 3.7). The Si:C ratio (molar) for a suite of laboratory diatoms ranged between 0.04 and 0.42, with a mean of 0.13 ± 0 . 0 4 (Brzezinski, 1985). If this ratio is applicable to the diatoms of Saanich and Jervis Inlets, it appears that most of the silicon assimilated by diatoms was eventually trapped at 50 m, as the BSi e-ratios with respect to diatom production are very close to Brzezinski's average (Table 3.7). It should be noted, however, that Kumar et al. (1996) found 25% of biogenic silica was lost from sediment traps and Scharek et al. (1999) estimated 10-60% of captured biogenic silica dissolved in sediment traps. Thus, unlike the case for organic carbon, the BSi e-ratios appear high, but could be the result of the combined effects of minimal BSi recycling, heavily silicified diatoms (Si:C ratio > 0.13) and the presence of a BSi-rich, resuspended fluxes reaching the shallow traps. This latter possibility seems especially likely in the fall and winter throughout the study area (except in the fall at station SN-0.8; Table 3.7) and year-round at station SN-9. The water-column fluxes at the mouth of Saanich Inlet were much higher than elsewhere (section 3.3.3) and were likely caused by sediment transport from the broad sill at the entrance to Saanich Inlet. 3.4.4 The riverine source of particulates to Saanich Inlet One of the most interesting and certainly the most unexpected feature of this time series is the aluminium record from station SN-0.8 (Figure 3.6). Unlike at the other stations (SN-9, JV-3 and JV-7) where A l fluxes were largely affected by the physical processes leading to turbulent resuspension, at station SN-0.8 fluxes of A l appear to have been determined by the environmental factors that delivered A l to surface waters. A l fluxes peaked in the late fall and winter when local rainfall (Figure 1.11) and flow of the Cowichan and Goldstream Rivers (Figure 1.13) were highest; local runoff in the vicinity of station Chapter 3. Settling fluxes in Saanich and Jervis Inlets 100 SN-0.8 and/or the sediment load of the Cowichan River could have delivered the Al to the head of Saanich Inlet. The correlation (r = 0.72) between Cowichan River flow and 50 m Al flux at station SN-0.8 is beter than the correlation (r = 0.49) between precipitation at Victoria Airport and the Al flux, but these correlations may be caused by coincidence of the seasonal cycle in each time series. Comparison of the residuals of each time series (Figure 3.21), however, suggests that the Cowichan River had a greater influence on the flux of aluminosilcates at the head of Saanich Inlet than did local runof. Because daily riverflow is available, it was also possible to test whether a time lag occured between Cowichan River flow and Al flux at the head of Saanich Inlet. As the residuals of Cowichan River flow were alowed to precede the flux residuals, the correlation did not improve, but it did remain constant for 4-6 days and decreased as the separation lengthened. Noting that the 30 day flux and riverflow averages (see caption to Figure 3.21) are not wel suited to explore time-lags less than about one averaging period, or a month, one interpretation of these results is that sediments reach the head of Saanich Inlet within a week of discharge from the Cowichan River. Aluminium appears to have been delivered from the surface at station SN-0.8 and the seasonal patern of flux was nearly identical at each depth (Figures 3.6 and 3.7). However, Al fluxes increased by a factor of ~2 between 50 m and 135 m and were similar at 135 m and 180 m (Table 3.3). The increase in flux with depth between 50 m and 135 m was probably not caused by turbulent resuspension, as no high-energy events (e.g., similar to the Al fluxes associated with deepwater renewals) were recorded. For a shorter time series from Sechelt Inlet, fluxes similarly increased with depth most dramaticaly from shalow to mid-depth traps, increasing less towards the botom. Timothy (1994) and Timothy and Pond (1997) concluded that, although resuspension events were probably recorded, particle focusing in the U-shaped fjord (e.g., Wassmann, 1984) and increases in trapping eficiency with depth were largely responsible for observed changes in flux Chapter 3. Settling fluxes in Saanich and Jervis Inlets 101 Figure 3.21: Residuals of 50 m Al flux from station SN-0.8 ploted against residuals of Cowichan River flow and residuals of rainfal at the Victoria Airport. Daily riverflow and rainfal data were used to create average values for time periods coresponding to the sediment trap deployment periods. Points in parentheses were not included in regressions. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 102 with depth. At station SN-0.8, the channel width (Figure 1.2) is 1.3 times greater at 50 m than at 135 m and, therefore, particle focusing cannot have accounted for all of the Al flux changes between these depths. Although no data on curents exist from the study in Saanich and Jervis Inlets, curent speed probably decreased with depth, and particles aggregate and consolidate as residence time (depth) increases. This flocculation wil cause sinking rates to increase (e.g.; Syvitski et al., 1985) and therefore trapping eficiency to improve (Butman, 1986; Yu et al., 2001). Fraser River flow peaks in the spring and summer (Figure 1.13) and thus cannot have been the source of aluminosilcates to station SN-0.8 at the head of Saanich Inlet. But, was sediment from the Fraser River plume deposited at station SN-9 inside the mouth of Saanich Inlet? Because the Cowichan and Fraser Rivers are both seaward of Saanich Inlet, the ratio of the concentrations of particulate mater from the rivers occuring in surface waters at the mouth ofthe fjord cannot vary within the fjord except by diferential sinking of the particles from the two sources. As a plume leaves the mouth of a river, the particulate load decreases due to mixing with seawater and setling (Hill et al., 2000) and the size fractions of the particulates shift from predominantly coarse silts to fine silts and clays (e.g.; Syvitski et al, 1988). The Stokesian setling velocity of the particulate load of a plume, therefore, decreases as the distance from a river mouth increases; however, a number of studies (e.g.; Syvitski et al., 1985; Gibbs and Konwar, 1986; Kineke et al., 1986; Miligan, 1995; Sternberg et al., 1999; Hill et al., 2000) show that flocculation has a significant efect on predicted setling rates of the smalest size fractions. Nevertheless, it is reasonable to model the sinking rate of plume particles as either constant (e.g.; as found by Hill et al. {2000} for the Eel River plume) or as decreasing (e.g.; as found by Syvitski et al., {1985} for the Homathko River of Bute Inlet) with distance from the source. The Fraser River is farther from Saanich Inlet (~ 50 km) than is the Cowichan River (~ 14 km). At the mouth of Saanich Inlet, Fraser River sediments wil be more fine-grained Chapter 3. Settling fluxes in Saanich and Jervis Inlets 103 with similar or smaler sinking rates than Cowichan River sediments and thus the Fraser River sediments wil disperse throughout Saanich Inlet as wel or beter than Cowichan River sediments. Because a Fraser-River signal was not observed at station SN-0.8 where discharge of the Cowichan River appears to have afected Al fluxes, it is unlikely fine silt and clay from the Fraser River reached parts of Saanich Inlet seaward of station SN-0.8. 3.4.5 D e e p w a t e r - c o l u m n , s ed imen t - in t e r f ace a n d b u r i a l f luxes Comparison of sediment-trap fluxes and botom accumulation rates has been used to gain information on the processes that control sediment delivery to the seafloor and burial (e.g.; Dymond, 1984; Anderson et al., 1994) and also to judge the fidelity of sediment traps (e.g.; Dymond, 1984). This section uses estimates of mass accumulation rate (MAR) calculated from 210Pb profiles to evaluate the accuracy of the sediment-trap fluxes and to address the causes and consequences of the depth-dependent fluxes in the two fjords. Because of the unique sedimentary environment at the mouth of Saanich Inlet, the high water column fluxes at station SN-9 are considered separately. M A R f r o m 2 1 0 P b prof i les In the summer of 1988, sediment cores were extracted from each station using a Peder-sen corer (Pedersen et al., 1985), and the top 25-30 cm were sampled at 1 cm intervals. OC, N and S13C were measured in the sediment samples as described in section 3.2.2. A I 2 O 3 , Si02 and other metalswere measured by X-ray fluorescence (XRF) spectrometry. BSi was not measured on the core samples, but was estimated by subtracting lithogenic Si02 from total Si02. Lithogenic Si02 was calculated using the regression equation of Figure 3.22 and measured A1203 concentrations. The slope of the regression line of Fig-ure 3.22 (3.4) is a best estimate of the Al203:Si02 ratio for the aluminosilcate fraction caught by the sediment traps and is within the range expected for upper continental Chapter 3. Settling fluxes in Saanich and Jervis Inlets 104 SN-9 SN-0.8 JV-3 JV-7 linear AR (cm yr l) 1.5 0.60 0.60 0.27 sediment composition % BSi 19.9 24.5 15.0 18.7 % o c 2.91 4.84 3.96 3.59 % N 0.292 0.461 0.400 0.266 % Al 6.73 5.03 6.38 6.84 5nC -21.7 -21.9 -21.7 -22.9 Table 3.8: Compositonal properties of the upper 25-30 cm of cores extracted at each station. (See Table 3.1 for station locations.) crust (Taylor and McClennan, 1985). Linear sedimentation rates for the upper 25-30 cm were estimated from measures of excess 210Pb and converted to depth-averaged mass ac-cumulation rates (MAR) using measured porosity values and estimated dry bulk density (Shimmield, unpublished data). 210Pb has a half life of 22.26 yr and is wel suited for dating rapidly accumulating, laminated sediments such as those in Saanich Inlet. With more uncertainty, 210Pb can be used to date bioturbated sediments if linear accumulation rates are suficiently fast that excess 210Pb accumulates below the bioturbated layer (Bru-land, 1974; Skei and Paus, 1979; Skei, 1981; Crusius and Anderson, 1995), as is the case in Jervis Inlet. Linear accumulation rates and average sedimentary compositon for the upper 25-30 cm of each core are given in Table 3.8, while within Table 3.9 are the mass accumulation rates (therein termed burial of the mass flux). The linear sedimentation rate at station JV-7 is half those of stations JV-3 and SN-0.8, but the porosity at station JV-7 was proportionaly lower so that MAR at the three stations was similar. Table 3.9 compares fluxes to the deep sediment traps, to the sediment-water interface Chapter 3. Settling fluxes in Saanich and Jervis Inlets 105 70 60 \-o 20 50 \-40 30 10 0 0 2 4 6 8 10 12 14 %AI203 Figure 3.22: The A l 2 0 3 : S i 0 2 ratio for the lithogenic debris in Saanich and Jervis Inlets using analyses on the sediment trap material. A I 2 O 3 and S i 0 2 were measured by X R F on 104 sediment-trap samples from Saanich Inlet and nine samples from Jervis Inlet (Frangois, 1989). Biogenic silica has been measured on all samples. Open circles are total metal concentrations ( X R F ) . Closed circles are lithogenic S i 0 2 plotted against total A l 2 0 3 . Lithogenic S i 0 2 is calculated as total S i 0 2 minus biogenic S i 0 2 . Two points with parentheses were not included in the regression. The nine samples from Jervis Inlet fall close to the regression line. and those representing permanent burial. The sediment-trap fluxes are the annual aver-ages of Tables 3.3 and 3.4 and the burial fluxes are from multiplying the 2 1 0 Pb-derived M A R s by the sediment concentration of each constituent (Table 3.8). The differences between interface and burial fluxes of Table 3.9 provides the percent of the interface flux that was remineralised. The flux of constituent j reaching the sediment-water interface was estimated using the focusing factor for the change in flux of A l between the depths of the deep sediment trap and the bottom sediments (Al^uriai/Aldeeptrap). Equation 3.1 assumes A l behaves conservatively during diagenesis and that the material (3.1) Chapter 3. Settling fluxes in Saanich and Jervis Inlets 106 causing the difference in flux between sediment traps and the sediment-water interface is compositionally similar. The first assumption is reasonable for these recently deposited sediments and the second is supported by work suggesting that resuspended material is "rebound" in nature (Walsh et al., 1988a) and is often similar in composition to the bulk material caught in deep traps (Bloesch, 1982; Walsh and Gardner, 1992; Timothy and Pond, 1997; Lampitt et al., 2000), so that debris transported laterally underneath the deep traps may have been compositionally similar to the sediments reaching the deep traps. If diagenetically old sediment with high A l content is a significant fraction of the material causing larger fluxes to the interface, however, Equation 3.1 will overestimate constituent fluxes to the interface. For a station in the center of Saanich Inlet nine km south of SN-9 and using 2 1 0 P b profiling in the manner described above, Bruland (1974) estimated a linear accumulation rate of 0.8 cm y r - 1 and a M A R of 1,300 g m ~ 2 y r - 1 . Again from 2 1 0 P b profiles, Sage (1994) measured a M A R of 2,600 g m - 2 y r - 1 two km south of SN-9. Thus a gradient exists in the M A R of bottom sediments in Saanich Inlet, from 1,900 to 2,600 g m ~ 2 y r - 1 at the mouth (Table 3.9; Sage, 1994), to 1,300 g m - 2 y r - 1 in the center of the fjord (Bruland, 1974) and 770 g m ~ 2 y r - 1 towards the head at station SN-0.8 (Table 3.9). Based on chlorophyll a concentrations (Hobson and McQuoid, in press), rates of primary production (Chapter 2; Timothy and Soon, 2001) and the sediment trap fluxes at stations SN-9 and SN-0.8, these decreasing M A R s towards the head of Saanich Inlet can be explained by a gradient in primary production and the introduction of sediments into the fjord from the region of the sill. A t a station one km south of JV-3 and in waters slightly deeper (685 m versus 660 m at JV-3) , Sage (1994) estimated a M A R of 1,500 g m - 2 y r - 1 , more than twice the value of 650 g m - 2 y r - 1 reported for J V - 3 in Table 3.9. This difference may reflect inherent errors in 2 1 0 P b dating, but may also be the result of heterogeneous sedimention rates near Chapter 3. Settling fluxes in Saanich and Jervis Inlets 107 the mouth of Jervis Inlet; the core extracted in somewhat deeper waters (Sage, 1994) may have been from a localised depocenter. Stations SN-0.8, JV-3 and JV-7 Apart from station SN-9, the accumulation rate of Al was higher in the sediments than in the deep sediment traps by factors ranging from 1.1 to 1.7 (Table 3.9). Focusing in these U-shaped fjords (Wassmann, 1984; Timothy and Pond, 1997), as wel as lateral transport of material from topographic rises and slumping of sediment of the side-wals (Hedges et al., 1988b; Cowie et al., 1992; Bornhold et al., 1994) is likely to have caused the Al flux increases from the deep sediment traps to the fjord floors. Noting that some increase in flux with depth is expected between the deep sediment traps and the seafloor, the magnitude of these focusing factors qualitatively suggests the deep sediment traps were accurately trapping the setling flux of Al. The largest increases in flux between the depths of the middle and deep sediment traps (Tables 3.3 and 3.4) were recorded at station JV-3. Therefore, the smal focusing factor there (Table 3.9) needs explanation. Much of the diference in flux between 300 and 600 m at station JV-3 is the result of turbulent resuspension during deepwater renewals (Chapter 4). Otherwise, station JV-3 is in the largest basin of the four sediment-trap mooring sites. The shalowest part of the Jervis Inlet sill is at ~280 m and, apparently, during deepwater renewal a sediment plume propagates at depths principaly below 300 m. Because station JV-3 is otherwise in a large basin, changes in flux with depth below 600 m may be smal. The water-column fluxes at station SN-9 demonstrate the efects of a plume propagating from a sill. The largest increases in flux with depth occured between the shalow (45 m) and mid-depth (110 m) traps, where the lip of the sill is located, but fluxes between the mid-depth (110 m) and deep (150 m) sediment traps changed little. Nevertheless, it may be that the smal focusing factor at station JV-3 reflects eror in the 210Pb dating, as Sage (1994) Chapter 3. Settling fluxes in Saanich and Jervis Inlets 108 SN-9 SN-0.8 JV-3 J V - 7 : A l flux (mol m~ 2 yr" deep trap 9.94 1.05 1.44 1.00 interface 4.77 1.43 1.54 1.72 burial 4.77 1.43 1.54 1.72 focusing factor 0.48 1.4 1.1 1.7 mass flux (g m - 2 yr" deep trap 4650 875 870 628 interface 2230 1200 928 1080 burial ( M A R ) 1910 770 650 680 % loss 14 36 30 37 BSi flux (mol m~ 2 yr deep trap 15.2 3.69 3.31 2.12 interface 7.27 5.06 3.53 3.65 burial 5.65 2.80 1.45 1.89 % loss 22 45 59 48 oc to t flux (mol m 2 yr *) deep trap 15.1 5.41 4.41 3.52 interface 7.26 7.42 4.70 6.06 burial 4.63 3.10 2.14 2.03 % loss 36 58 54 66 O C m a r flux (mol m 2 yr *) deep trap 8.05 3.21 3.11 2.09 interface 3.86 4.40 3.31 3.60 burial 2.03 1.28 1.49 1.07 % loss 47 71 55 70 oc ter flux (mol m 2 yr 1) deep trap 7.10 2.20 1.31 1.43 interface 3.40 3.02 1.39 2.46 burial 2.60 1.82 0.66 0.97 % loss 24 40 53 61 N flux (mol m~ 2 yr~ deep trap 1.53 0.595 0.432' 0.341 interface 0.734 0.816 0.461 0.587 burial 0.398 0.253 0.186 0.129 % loss 46 69 60 78 Table 3.9: Deep sediment-trap (Tables 3.3 and 3.4), sediment-water interface (Equa-tion 3.1) and burial (Shimmield, unpublished) fluxes. The focusing factor is of Equa-tion 3.1. The percent lost is the fraction of the interface flux remineralised during dia-genesis in the upper sediment core. Marine and terrestrial O C fluxes are estimated from <513C as for section 3.4.2. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 109 obtained a MAR for a core colected near station JV-3 of more than twice the estimate in Table 3.9. As much as 54 to 6% of the OC deposited at the sediment-water interface at stations SN-0.8, JV-3 and JV-7 was remineralised in the upper 30 cm of sediment, and 45 to 59% of the biogenic silica dissolved (Table 3.9); most of this loss probably occured in the uppermost portion of the core (e.g.; Hedges et al., 1988b). These OC losses agree with other estimates made in Saanich Inlet (Cowie et al., 1992) and other fjords (Burrel, 1983). In Jervis Inlet, roughly equal fractions of the marine and terigenous organic mater reaching the sediments were remineralised, while at station SN-0.8 significantly more of the marine OC was degraded (71%) than of the terigenous OC (40%; Table 3.9). This diference between fjords may be due to a greater degree of water-column remineralisation ofthe most labile marine OM in Jervis Inlet due either to its greater depth, or the periodic anoxia in Saanich Inlet. However, anoxia does not appear to afect the decompositon of bulk OM, as similar amounts of the carbon flux to the interface are lost during sediment diagenesis. In Saanich Inlet, low oxygen conditons may indirectly efect opal preservation by decreasing the rate by which organic coatings are degraded (e.g.; Lewin, 1961; Bidle and Azam, 1999), but this factor does not appear to be significant; at station SN-0.8 45% of the biogenic silica dissolved, while in Jervis Inlet 48-59% was lost from the sediment. Station SN-9 Although the OC e-ratios at station SN-9 were not higher than for the other stations, the BSi e-ratios were about double those observed at the head of Saanich Inlet and in Jervis Inlet and, at all depths, the material intercepted by sediment traps at station SN-9 was more refractory than at the other stations; %OC and %BSi were lowest, while %A1 was highest at this site (Figures 3.8 and 3.14). Refractory material was likely resuspended of the sill of Saanich Inlet and perhaps also traveled from the mouth of Chapter 3. Settling fluxes in Saanich and Jervis Inlets 110 mol C m" 2 yr"1 10 20 30 40 50 yy 14(22)% ^ 8.0(15)% 4.2 (9.6)% SN-9 0 0.33 0.66 0.99 1.32 1.65 g C m"2 day"1 mol C m" 2 yr"1 0 10 20 30 40 50 6 /; 16(27)% 18 (31)% 5.5(10)% i i_ production 50 m flux interface flux burial flux terrigenous flux 0 0.33 0.66 0.99 1.32 1.65 ,-2 ^o„-1 g C m day" Figure 3.23: Summary of O C fluxes. Primary production is from Table 2.1, 50 m fluxes are from Tables 3.3 and 3.4 and sediment-water interface and burial fluxes are from Table 3.9. The sediment-trap and benthic fluxes are separated into marine (filled part of each bar) and terrigenous (hatched portion of bar) components, so that their sum is the total O C flux at each interval. Numbers associated with each bar are the percent of primary production accounted for by each flux. Numbers in parentheses are for total O C , while numbers before parentheses are for marine O C . Chapter 3. Settling fluxes in Saanich and Jervis Inlets 111 mol Si m"2 yr"1 0 2 4 6 8 10 12 200% 150% 110% SN-0.8 110% 150% 82% 0 0.37 0.74 1.11 1.48 1.85 2.21 g BSi m"2 day"1 mol Si m"2 yr"1 0 2 4 6 8 10 12 JV-7 96% 210% 110% I I production I I 50 m flux interface flux burial flux 0 0.37 0.74 1.11 1.48 1.85 2.21 g BSi m"2 day"1 Figure 3.24: Summary of BSi fluxes. Si production is calculated as 0.13 x diatoma-ceous carbon assimilation (Table 3.6). 50 m fluxes are from Tables 3.3 and 3.4, and sediment-water interface and burial fluxes are from Table 3.9. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 112 the Cowichan River to the vicinity of station SN-9. Nevertheless, the water-column fluxes of A l were higher than the M A R of A l in the sediments, resulting in a focusing factor less than one (Table 3.9). In fact, there is no evidence of enhanced sediment deposition at station SN-9 when comparing sedimentary fluxes of O C and BS i to local primary production (Figures 3.23 and 3.24). According to Gucluer and Gross (1964) and Frangois (1987), station SN-9 is located at the northern edge of sediments characteristic of the central basin, and the large linear accumulation rates measured there (Table 3.8) argue against this as a location of winnowing. Therefore, it appears that a nepheloid layer or sediment plume extends from the sill of Saanich Inlet to station SN-9. This plume, furthermore, does not appear to deposit an amount of biogenic material to the sediments in excess of the focusing occurring at other stations. Remineralisation within the sediments at station SN-9, however, is less than at the other stations (Table 3.9). The enhanced preservation inside the sill of Saanich Inlet is probably caused by the very high accumulation rate, but may also be the result of particularly refractory sediments being deposited here. At all stations, the BSi interface flux is 1.4 to 2.1 times greater than estimated Si assimilation (Figures 3.23 and 3.24). While at stations JV-3 , J V - 7 and SN-0.8 these additional fluxes may represent relatively young, autochthonous debris focused toward the bottom, at station SN-9 the additional sediments coming from outside the fjord may be more diagenetically altered. Oxygen and carbon budgets for the deep waters of Jervis Inlet For the stagnant waters behind the sill of a fjord, the decay of dissolved oxygen over time allows estimation of the water-column and sediment oxygen demand in the deep basin. This exercise is complicated for Saanich Inlet because dissolved oxygen concentrations did not decay in a gradual manner (see Figures 1.4, 1.5 and 4.8), likely due to the oxidation of reduced chemical species during and shortly after deepwater renewals. However, in Chapter 3. Settling fluxes in Saanich and Jervis Inlets 113 Jervis Inlet, dissolved oxygen decayed in such a manner that this exercise is possible (Figures 1.6, 1.7 and 4.7). A box model is used to compare the observed loss of oxygen behind the sill (240 m) of Jervis Inlet and the predicted oxygen demand based on organic carbon remineralisation within the deep waters and sediments. Oxygen contours (Figures 1.6 and 1.7) show that dissolved oxygen was often homogeneously distributed below about 300 m, while above this depth mid-water advection seasonally affected oxygen concentration. Thus, 300 m is taken as the top of the box within which advective sources of oxygen should be negligible and diffusion effects are ignored, noting that vertical gradients in dissolved oxygen concentration at 300 m at stations JV-3 and J V - 7 were small. The mean depth of the center of Jervis Inlet is 495 m (Pickard, 1961). Accounting for the sloping sides below 300 m (see transects of Figure 1.3), the average depth of the box is 160 m. For sampling locations within the box (Figures 1.6 and 1.7), the rate of dissolved oxygen decay during periods between deepwater renewals was approximately 0.10 pM d _ 1 (see Figure 4.7 for an example), converting to an oxygen consumption rate of 5.8 moi m ~ 2 y r - 1 within the box. In Chapter 4, water-column remineralisation rates of organic carbon are estimated. From those results and using a mean flux of 130 mg O C m ~ 2 d _ 1 (mean annual O C flux for sediment traps moored within the box; Table 3.4) at the top of the box, 0.64 moi O C m - 2 y r - 1 were lost within the water-column of the box. Finally, from Table 3.9, 2.6 to 4.0 moi O C m ~ 2 y r - 1 were lost within the sediments at stations J V - 7 and JV-3 , respectively, for an average sedimentary O C loss of 3.3 moi O C m - 2 y r - 1 . These water-column and sedimentary losses of organic matter can be converted to an oxygen demand using the stoichiometry outlined in section 3.3.1. For the O C : N ratios observed in Jervis Inlet, oxygen and organic carbon should be lost at a molar ratio of 126:106 during organic matter degradation (Timothy, 1994). Thus, the water-column and sediment oxygen demands are predicted to have been 0.76 and 3.9 moi 0 2 m - 2 y r - 1 , Chapter 3. Settling fluxes in Saanich and Jervis Inlets 114 respectively. Considering the errors, these independent estimates of the oxygen demand for the bottom 160 m of Jervis Inlet (5.8 mol m ~ 2 y r - 1 based on the time-course of dissolved oxygen concentration and 4.7 mol m - 2 y r - 1 determined from remineralisation of organic matter) agree very well. The largest sources of error in the estimate based on dissolved oxygen concentrations is determining the upper boundary of the box, which could be off by about 50 m ( ± 3 0 % ) , and the rate of oxygen decay, which ranged between about 0.06 and 0.16 pM d _ 1 for different periods, depths, and stations of the time series. The average of 0.10 pM d _ 1 is probably accurate to within 20%. The predicted oxygen demand based on organic matter degradation is most sensitive to the sedimentary oxygen demand, as it accounted for 84% of predicted organic matter remineralisation within the box. It is difficult to put an error on the sediment oxygen demand, which is based on 2 1 0 P b profiles and sediment-trap fluxes, but ± 5 0 % is probably a conservative estimate. 3.5 C o n c l u s i o n s 1. Water-column fluxes of organic carbon and biogenic silica follow local primary pro-duction in Saanich and Jervis Inlets. However, physical processes of resuspension and particle focusing in the vertically-narrowing channels, and possibly increased trapping efficiency with depth due to weak deepwater currents and accelerated sinking, cause increases in flux with depth. At station SN-9, the effect of mid and deepwater renewals on particle flux throughout the water column deceptively suggest late summer plankton blooms. 2. Although resuspension largely affects the A l fluxes at stations SN-9, JV-3 and JV-7 , at station SN-0.8 the rain of aluminosilicates is controlled by terrigenous runoff, as A l fluxes closely follow local precipitation and flow of the Cowichan and Goldstream Chapter 3. Settling fluxes in Saanich and Jervis Inlets 115 Rivers. There is some evidence that the Cowichan River has a greater efect on Al fluxes than local runof. The Fraser River as a significant source of aluminosilcates to Saanich Inlet is unlikely. 3. The relationship between stable carbon isotopes and BSi content reveals that 70-80% of the marine OC in these fjords is diatomaceous. This relationship was used to make estimates of the mean 513C signal of setling marine organic mater. In Saanich Inlet, an estimate of the mean 513C of marine organic mater is -17.3°/oo and, in Jervis Inlet, is -19.6°/oo- The diference between fjords may be due to a greater predominance of fast-growing diatoms in Saanich Inlet. In the spring and summer, 20-40% of the sinking organic mater at 50 m is terigenous while, in the fall and winter, 40-60% is terigenous. 4. The sediment-trap export ratios of OC were very low in both fjords during this study, and may have been afected by solubilsation of organic material within the sediment traps and during laboratory processing. Export ratios of BSi were high, suggesting an eficient transfer of biogenic silica out of the euphotic zone. The observed BSi e-ratios may have furthermore been afected by resuspended or focused sediments reaching the shalow sediment traps. Exceptionaly high fluxes of biogenic silica at station SN-9 mark a sediment plume or nepheloid layer associated with the sill. 5. Although a sediment plume extends from the sill into Saanich Inlet, when com-paring to local primary production, biogenic material is not transmited to the sediments in excess of the biogenic delivery to sediments at the head of Saanich Inlet or in Jervis Inlet. However, at station SN-9 organic mater and biogenic silica is beter preserved in the sediments, probably because of the high accumulation rates and the refractory nature of the sedimenting material. Chapter 3. Settling fluxes in Saanich and Jervis Inlets 116 6. Comparing accumulation rates in the upper 30 cm of the sediments with local primary production, there is little evidence that either organic mater or biogenic silica is preferentialy preserved in the periodicaly anoxic environment of Saanich Inlet; in both Saanich and Jervis Inlets, about 5% of the carbon and 60-110% of the Si assimilated in the euphotic zone is buried in the sediments beneath. In Chapter 4 it is estimated that water-column dissolution in Jervis Inlet causes about a 20% loss of biogenic silica as it sinks from 100 m to 200 m, and a 35% loss for sinking from 100 m to 600 m. Considering water-column dissolution yet the high proportion of assimilated Si buried in the sediments, there is a large amount of particle focusing occuring in these fjords and/or the Si:C uptake ratio of 0.13 used to estimate Si assimilation is too low. Away from station SN-9 where accumulation rates are highest and the sediment is of a more refractory nature, similar amounts of organic carbon (54-66%), nitrogen (60-78%) and biogenic silica (45-59%) are lost between the sediment-water interface and deeper in the sediment core at stations SN-0.8, JV-3 and JV-7. Chapter 4 A model to interpret increases i n flux w i t h depth: remineral isa t ion rate constants and the addi t ional flux to deep sediment traps 4.1 In t roduct ion In studies of vertical particle flux in the ocean using moored sediment traps, increases in flux with depth are often observed, especialy in coastal environments. For these situations, deducing particulate supply from surface waters as wel as the alteration of material as it sinks is not straightforward without a means of separating primary fluxes from the additional material caught by deep traps. Quantiative methods to interpret sediment-trap fluxes that increase with depth have considered either the decay of primary fluxes or the compositon and amount of addi-tional fluxes, but not both simultaneously. For instance, Noriki et al. (1985) determined regeneration rates for biogenic silica and organic mater in a shalow bay in Japan by normalising fluxes to aluminium (Al). Later, Noriki and Tsunogai (1986), for trap fluxes from the Pacific and Southern Oceans, and Walsh et al. (1988b), for the Equatorial North Pacific, also normalised fluxes of particulate organic carbon, calcium carbonate and biogenic silica to Al. The oceanic estimates of remineralisation give similar amounts of biogenous flux decay over comparable depth intervals. However, the normalisation procedure assumes compositonal similarity between primary and additional fluxes, a constraint that is not everywhere appropriate. For example, artificialy high decay rates are obtained if Al-rich material, such as refractory botom sediment, is the cause of an 117 Chapter 4. A model to interpret increases in flux with depth 118 increase in flux with depth (Walsh et al., 1988a). Normalisation to Al also assumes that the Al flux is conservative, though dissolved Al is known to be scavenged by sinking particles (Orians and Bruland, 1986). In regions where Al fluxes are smal, scavenging may cause over-estimates of decay. Bloesch (1982) applied a method to quantify the amount of resuspended material that reached near-botom traps in the shalow and tur-bulent waters of Lake Erie. The method was able to detect that there were both localy resuspended sediments and material of a more organic-rich nature within the additional flux to hypolimnetic traps. Walsh and Gardner (1992) described a similar model and found that the compositon of the additional flux to deeper traps moored in the Gulf of Mexico was more similar to the primary flux than to botom sediments. However, these two techniques were not general because water-column decay was either not considered (Bloesch, 1982) or had to be estimated independently using the normalisation scheme (Walsh and Gardner, 1992). Three common methods to unravel coastal sediment-trap fluxes that increase with depth were reviewed by Hakanson et al. (1989), and a combination of two of those meth-ods (the base-line approach and the burial approach) was used by Pejrup et al. (1996) to separate primary from resuspended fluxes in a shalow, coastal environment. Techniques similar to the label approach of Hakanson et al. (1989) have been used to infer much about processes afecting particles as they sink. For instance, Blomqvist and Larsson (1994) used the Al concentration of primary and resuspended sediments to estimate the proportion of primary setling material to sediment traps moored at a single depth at two stations during a five year time series in the Baltic Sea. A greater abundance and difer-ent assemblages of intact phytoplankton cels in deep sediment traps relative to shalow traps was used to evaluate the degree of lateral transport from the shelf to the slope of the Middle Atlantic Bight (Falkowski et al., 1994). Also, the compositon of sediment-trap material and underlying sediments has provided insight into the biochemical changes that Chapter 4. A model to interpret increases in flux with depth 119 occur to particles as they sink through the water column and become incorporated into the sediments in Dabob Bay, WA (Hedges et al., 1988a; 1988b). However, none of the approaches outlined by Hakanson et al. (1989) describes remineralisation of the primary flux, a term that may be relatively large where water depths are greater than roughly 10 to 100 m (Pejrup et al., 1996). The primary concern of this work is to estimate remineralisation rates of setling particulate material where observed fluxes increase with depth so that more accurate elemental budgets can be described for coastal regions. A general balance equation that treats as unknowns both water-column decay and the compositon of additional material caught by a deeper trap was developed by Timothy (1994) and Timothy and Pond (1997). That model is applied to fluxes of organic carbon, nitrogen, biogenic silica and aluminium during the multi-year sediment-trap time-series from Saanich and Jervis Inlets presented in Chapter 3. 4.2 Descr ip t ion and solution of the model Timothy (1994) and Timothy and Pond (1997) described a model to estimate rates of water-column decay where measured fluxes increase with depth for the general case where the rain to the deep sediment trap is a mixture of debris from diferent sources and with diferent compositonal properties. The model (Figure 4.1) partitions the flux to a deep sediment trap into two components. The anticipated flux is the representative material expected to reach the deep sediment trap knowing the flux at the shalow sediment trap, and the additional flux is the material caught by a deep sediment trap in excess of the anticipated flux. The observed flux to a deep sediment trap, therefore, is the sum of anticipated and additional fluxes. In general, anticipated and additional fluxes might be equated with primary and resuspended fluxes, respectively, but the operational terms are Chapter 4. A model to interpret increases in flux with depth 120 upper trap c lower trap anticipated flux Figure 4.1: Schematic of the sediment-rap model. Anticipated fluxes are alowed to decay as they sink and there is no constraint on the source of additional fluxes. used to emphasise that the model can be applied to a data set regardless of the sources of material reaching upper and lower sediment traps. Fluxes are recorded by sediment traps at depths z\ and z2 (z is positive downward). j is a component of the mass flux, J. Anticipated and additional fluxes are identified by subscripts n and d, respectively, so that: The model is designed largely to find rates of remineralisation of the sinking flux of j between depths z\ and z2 bounded by two sediment traps. Using the results of the modeling on the six depth intervals in Jervis Inlet, it wil be shown in section 4.4.1 that rates of remineralisation of OC, N and BSi decreased with depth during the study in Jervis Inlet. However, because the depth distribution of loss cannot be determined from fluxes (4.1a) J2 = jn + 3d • (4.16) Chapter 4. A model to interpret increases in flux with depth 121 measured at two depths, here it is assumed that the rate of remineralisation of constiuent j between the depths of two sediment traps is constant, so that anticipated fluxes decrease exponentialy with depth as they sink (e.g.; Walsh et al., 1988b). Equation 4.16 becomes: J2 = jie-k'^+jd. (4.2) kj (m-1) is the rate constant for component j. In order to estimate the rate constant, jd must be quantifed and in the model is writen as the product of Jd and the fraction of j in Jd, (j/J)d- Equation 4.2 is re-writen as: . 3 2 = h e - k ^ z + ( ^ j J d . (4.3a) If an estimate of Jd can be made for each deployment period, then Equation 4.3a is linear with the measured (or estimated) variables j2, j\ and Jd- For data from n deployment periods, these variables can be ploted on three orthogonal axes. If a plane is fitted through the data points, the slope of the line representing intersection of the fitted plane with the J2 • ji plane is a best-estimate of e~k>Az, and the slope of the line where the fitted plane intersects the j2 : Jd plane is an estimate of (j/'J)d. The degree of fit of the plane to the data is a measure of the extent to which e~kjAz and {j/J}d behaved as constants over the period and depth interval considered. As Equation 4.3a is writen, the solution plane should pass through the origin and therefore should have a zero-intercept on the j2-axis. However, there may be erors associated with the trapping experiment or with the assumptions of the model that might cause a non-zero intercept on the j2-axis. A term (e, the intercept on the j2-axis) is included in Equation 4.3a to quantify these possible errors. h = he-k'^+(^j Jd + e (4.36) This eror term is discussed further in section 4.3.2. Chapter 4. A model to interpret increases in flux with depth 122 In order to solve Equation 4.36, the total additional flux, Jd, must be estimated for each deployment period. If fluxes setle conservatively, Jd can be estimated as the diference in flux between depths z i and z 2 : Jd — J2 — J\- However, where water-column decay occurs, Jd wil be greater than J2 — J\ by the amount of J\ that is lost while it sinks to depth z 2 . The sinking flux is composed of organic mater, biogenic silica, calcium carbonate and lithogenic debris. Lithogenous fluxes are expected to be conservative and CaCO"3 made up only about 2 % of the observed fluxes in Saanich and Jervis Inlets. Only the degradation of POM and the dissolution of biogenic silica are therefore included in the expression for Jd in this study: Jd = J2 - Jl + 2.7Ci (1 - e-*c**) + Sh (1 - e -ke.. Az (4.4) C\ and Si\ are fluxes of organic carbon and biogenic silica at the depth of the upper sediment trap, kc and kSi are decay constants for organic carbon and biogenic silica, and the factor 2.7 is used to convert POC to POM (section 3.3.1). In equation 4.4, the terms with C\ and Si\ account for the portion of J\ lost to the water column between depths z\ and zi- Replacing Jd of Equation 4.36 with Equation 4.4, 32 = 3i e ~-kjAz + (1) (J2 -J1 + 2.7C7i {1 - e~kcAz} + Sh {1 - e-ks*Az}) + e . (4.5) For data colected over n deployment periods, Equation 4.36 is a set of n simultaneous equations. 3n Jdi 1 3l2 Jd2 1 e-kjAz jln Jdn 1 Writing Equation 4.6a in abbreviated matrix notation: 321 J22 32n (4.6a) X a = Y (4.66) Chapter 4. A model to interpret increases in flux with depth 123 - l and matrix multiplication of both sides of Equation 4.6b by (XTXj XT gives: periods represented. In applying Equation 4.7, ( J 2 — J i ) is used as a first approximation of Jrf. First estimates of kc and kSi are made by solving Equation 4.7 for j — O C and j = BSi , respectively. These rate constants are used in Equation 4.4 to make improved estimates of Jd which, again applying Equation 4.7, give new estimates of kc and kSi. This iterative procedure is continued until the rate constants used to calculate Jd converge with kc and kSi given back when solving Equation 4.7 for O C and BS i fluxes. W i t h final estimates of kc and kSi for the depth interval and deployment periods being addressed, best estimates of Jd (Equation 4.4) can be made, and kj and (j/ J)d of constituent fluxes (e.g.; N, Al) can be solved without iteration. For the case where the rate constants do not converge, or where computational re-sources do not allow the iteration described here, an alternate solution to the model is described in Appendix B . 4.3 Resul ts 4.3.1 Sensi t ivi ty analyses and examples of the planar fits The model of Equation 4.5 and solved using Equation 4.7 has been applied to the sediment-trap fluxes from Saanich and Jervis Inlets reported in Chapter 3. A sensi-tivity analysis of the data from each depth interval (three at each station) was performed in order to find and remove records that heavily biased model solution (Appendices C and D). The model solutions presented in Tables C . l and D . l are the results after out-liers have been removed. The rate constants estimated for Saanich Inlet (Table D . l , Figure 4.4) are spatially inconsistent, and the differences between constituents are not (4.7) Chapter 4. A model to interpret increases in flux with depth 124 clearly meaningful. Possible causes of these results are discussed in Appendix D; one is that the depth intervals separating the sediment traps may have been too small to resolve the decay of the settling flux. Rate constants from Saanich Inlet are plotted in Figure 4.4. Otherwise, this chapter will consider only the results for Jervis Inlet. In Jervis Inlet, 11 of the surface records biased the model results for the surface-mid and/or the surface-deep intervals. Of these 11 records, eight were from station J V - 7 and most were from periods of anomalously high fluxes of O C to the 50 m sediment traps (Figure C.7). The majority of these periods occurred in the spring and summer of 1985 and 1986, and were generally characterised by high O C : B S i ratios (compare Figures 3.11 and 3.10). Primary production was not measured throughout 1985, but in 1986 was higher than average in Jervis Inlet. It is possible that these high O C fluxes were associated with blooms of thecate dinoflagellates or nanoflagellates as described for Sechelt Inlet (Haigh et a l , 1992; Taylor et al., 1994), adjoining Jervis Inlet near station JV-3 (Figure 1.1). Nutrient data from station J V - 7 (Figure 2.6) further suggest that towards the head of Jervis Inlet autotrophic flagellates were more prevalent than at station JV-3 or in Saanich Inlet. The high O C fluxes with low O C : B S i ratios also may have represented a response by the zooplankton to blooms of diatoms or flagellates. While grazing heterotrophs may have died and sank to the depth of the 50 m sediment traps, it is also possible they were attracted to the debris in the 50 m traps and contaminated the samples (e.g.; Michaels et al., 1990). This latter explanation for the high O C fluxes is supported by the observation of polychaetes occasionally caught within the grids of the sediment traps (Chapter 3). Furthermore, for these fluxes to have resulted from naturally sinking phytoplankton or heterotrophs, remineralisation between the 50 m and the mid-depth traps would have been exceptionally high. Without a protective shell, sinking flagellates may remineralise quickly, but if so they would not be expected to sink in abundance to 50 m, a depth well below the euphotic zone in Jervis Inlet (Figure 2.7). Chapter 4. A model to interpret increases in flux with depth 125 J V - 3 : 300-600 m. O C solution 0 . 5 ^ 0 . 4 v 0 . 3 ^ Figure 4.2: Plot of OC solution: station JV-3, 300-600 m. The upper plot depicts the plane relative to the OC2, OC\ and Jd axes. Below, the plot has been rotated to an angle where the plane intersects the data to show the goodness of fit of the model solution. Open circles are plotted data, filled circle is the OCVintercept. Chapter 4. A model to interpret increases in flux with depth 126 J V - 7 : 50-450 m. BSi solution Figure 4.3: Plot of BSi solution: station JV-7, 50-450 m. The upper plot shows the plane relative to the BSi2, BSi\ and Jd axes and, below, is rotated to show the goodness of fit. Open circles are ploted data, filled circle is the BS^-intercept. Chapter 4. A model to interpret increases in flux with depth 127 Swimmer contamination was therefore a likely cause of the anomalously high O C fluxes to the 50 m sediment traps in Jervis Inlet. Figures 4.2 and 4.3 show examples of the solution plane fit to the sediment-trap data from Jervis Inlet. These examples were chosen because they represent the range of goodness-of-fit of the model to this data set (Table C . l ) . The fit to O C fluxes of 300-600 m at station JV-3 (Figures 4.2) was very good, while the fit to BS i flux for 50-450 m at station J V - 7 was not as good, and a significant intercept on the .BS^-axis occurred (Figure 4.3). 4.3.2 The error t e rm The error term included in the model and depicted as the intercept on the ji axis (e.g.; Figure 4.3) was small for the deep sediment-trap depth intervals in Jervis Inlet, but, for many cases from the more shallow intervals, the error was non-zero (Table C . l ) . The model assumes that the material representing the anticipated flux reaches the deep sediment trap, but this assumption may be at least partially invalid for a number of reasons including changes in trapping efficiency with depth and horizontal advection across a horizontal gradient in the settling flux. Thus, the anticipated flux might be over- or under-represented in the deep sediment trap, giving physical significance to e. The modelled fluxes to a deep sediment trap, including the error term (Equation 4.36), can be equated with the true fluxes in a way that allows for the possibility that the deep sediment trap imperfectly captured the anticipated flux. y is the fraction of jn that reaches the deep sediment trap. Rearranging Equation 4.8: jn +jd + £ = yjn+ jd • (4.8) e = (V - 1) Jn • (4.9) Chapter 4. A model to interpret increases in flux with depth 128 • e > 0: over-colection of jn at z2. \ • • e < 0: under-colection of jn at z2. For the depth intervals associated with the surface sediment traps, positive erors (the intercepts reported in Table C.l) were common for OC, N and Al, and could be caused by increased trapping eficiency with depth. Improved trapping eficiency in deep water can be caused by reduced curent speeds at depth, and by accelerated sinking of the setling debris (e.g.; Smetacek et al., 1978; Timothy and Pond, 1997; Yu et al., 2001). However, if improved trapping eficiency were causing these intercepts for the OC, N and Al solutions, it is expected the BSi solutions would also have positive intercepts, but they do not. The negative BSi intercepts could be caused by horizontal advection carying the anticipated flux away from the deep sediment trap and replacing it with material from a region of smal export flux (Deuser et al., 1988; Siegel et al., 1990; Yu et al., 2001). Timothy and Pond (1997) presented a scale analysis for Sechelt Inlet to determine whether increases in flux with depth could have been caused by the advection of sediment from a high export region. They concluded that horizontal gradients in the export flux, relative to local curent speeds, were not suficient to cause increases in flux with depth unless sinking rates were quite low (< 20 m d_1). Currently, the cause of the diference in sign of the BSi erors and those of OC, N and Al is unresolved. It remains to be seen whether sample artifacts, such as diferential preservation and swimmer contamination, are afecting these results. Chapter 4. A model to interpret increases in flux with depth 129 Figure 4.4: The relationship between rate constants (kc, k^ and kst) and depth. Top row: rate constants ploted against depth. Middle row: semi-ln plots describing an exponential dependence (Equation E.2) of rate constants on depth. Botom row: ln-ln plots showing the power function relationship (Equation 4.11). For the power functionality, z0 is set at 20 m, the approximate base of the euphotic zone in Jervis Inlet. Solid and dashed lines are for semi-ln and ln-ln treatments, respectively. kWOm, ct and (3 are from the regression equations of the semi-ln and ln-ln plots. See Appendix E for the exponential (semi-ln) treatment, and why it was rejected. The model results from Saanich Inlet are viewed skepticaly (Appendix D ) , but the rate constants are ploted here for comparison. Chapter 4. A model to interpret increases in flux with depth 130 4.4 D i s c u s s i o n 4.4.1 D e s c r i b i n g d e p t h dependence o f t he r a t e cons t an t s At Equation 4.2, it was noted that the depth dependence of the rate constants cannot be determined from data collected at two sediment traps, so it was assumed that the rate parameters were constant with depth. However, results from the six depth intervals in Jervis Inlet (Figure 4.4) show that the rate parameters decreased with depth non-linearly. Organic matter is made of a range of constituents that differ in their susceptibility to degradation; the most labile compounds are oxidised quickly, while refractory components require more time (or depth) to decay (Toth and Lerman, 1977; Westrich and Berner, 1984; Harvey et al., 1995). Diatom frustules also exhibit a large range in their dissolution rate. Temperature plays an important role in silica dissolution (Lewin, 1961; Lawson et al., 1978; Kamatini and Riley, 1979), as does ambient silicic acid concentration (Hurd and Birdwhistell, 1983; Van Cappellen and Qiu, 1997). However, below 50 m in Jervis Inlet, these factors did not vary significantly (T = 9° ± 1°; [Si ( O H ) J = 54 - 6SpM). Susceptibility to dissolution varies widely between diatom species depending on the spe-cific surface area and morphology of their frustules (e.g.; Lewin, 1961; Kamatini and Riley, 1979; Kamatini , 1980). Furthermore, organic coatings play an important role in protecting biogenic silica from dissolution (e.g.; Lewin, 1961; Bidle and Azam, 1999) and diatom fragmentation by grazing can also greatly accelerate BS i dissolution (Sancetta, 1989a, 1989b; Ragueneau et al., 2000). Indeed, Sancetta (1989c) found that the small and weakly silicified cells in Jervis Inlet were preferentially lost in the water column, while the dense taxa and diatom cysts better survived transit to the deep sediment traps. The scenario where organic matter is composed of various fractions that decay with different reactivities has been described with the mult i -G model (J0rgensen, 1978; Westrich and Berner, 1984), which predicts that changes in the rate constant with time (depth) are Chapter 4. A model to interpret increases in flux with depth 131 most rapid near to (zo). Middleburg (1989) found that the power function describes rate constants of organic carbon ranging eight orders of magnitude in age, including labora-tory data that had previously been characterised by the quantum-G model (a variant of the multi-G model that considers a finite number of reactive components; Westrich and Berner, 1984). Because the diatom assemblage in Jervis Inlet encompasses diverse taxa with variable susceptibilty to dissolution, the power function may also be appropriate to describe changes with depth of biogenic silica rate constants. In constructing the power function model, changes in kj with depth are a function of depth: ^ = - / 3 f c o ^ - D . (4.10) dz P is a dimensionless constant that describes the dependence of kj on z. Equation 4.10 is writen as such so that integration gives the power function: k = koz-P . (4.11) The natural logarithm of Equation 4.11 is: ln(ifc) = -Pln{z) + ln(A;o) . (4.12) P and k0 can thus be determined as the slope and the intercept, respectively, of the regression line of ln(fc) ploted against \n(z). P is sensitve to z0, the depth where ma-terial begins sinking and equivalent to t0 of Middleburg (1989). For the curve fitting of Figure 4.4, z0 was set at 20 m, the approximate depth of 1% surface irradiance during the study (Figure 2.7). Chapter 4. A model to interpret increases in flux with depth 132 OC:BSi (molar) 0 .5 1.0 1.5 normalized flux normalized flux OC:N (molar) 0.0 0 .5 1.0 0.0 0 .5 1.0 9 .0 9 .5 10 .0 1 0 . 5 11 .0 Figure 4.5: Flux versus depth in Jervis Inlet, a. Decrease with depth of the anticipated fluxes of P O C and BSi for Jervis Inlet. The curves extending to ~300 m were generated from results from nearby Sechelt Inlet (Timothy and Pond, 1997). The longer curves are from the results of Figure 4.4; heavy lines are from the power-function description of rate constants (Equation 4.14) and light lines are from the exponential description of the rate constants (Equation E.5; the two curves for BSi flux are nearly identical and difficult to differentiate), b. Comparison of the Jervis P O C curve with curves published by Martin et al. (1987; central dashed curve), Betzer et al. (1984; right-hand dashed curve) and Suess (1980; left-hand dashed curve). Other oceanic curves (Bishop, 1989) are within the envelope of those presented here. c. Flux ratios with depth in Jervis Inlet. Chapter 4. A model to interpret increases in flux with depth 133 4.4.2 T h e a n t i c i p a t e d flux Having fitted the power function of Equation 4.11 to the rate constants of Table C.l, a description of changes in flux with depth can be made. I = -kj (4.13) Substiuting Equation 4.11 into Equation 4.13 and integrating: {(5^1). (4.14) Equation 4.14 is a description of the flux that would be observed if additional material did not reach deep sediment traps. From k0 and 3 of Figure 4.4, the anticipated fluxes of OC and BSi in Jervis Inlet are given in Figure 4.5. OC:N and OC:BSi ratios of the particulate fraction are also given in Figure 4.5. The OC:BSi ratio of the setling flux decreases with depth, as remineralisation of OC exceeds that of BSi. In general, there is an increase in the particulate OC:N ratio with depth, as N remineralisation occurs slightly more rapidly than carbon, except for the surface-mid and surface-deep depth intervals at station JV-7 (Table C.l). Here, kc is larger than kN, though the diference was not significant. Therefore, the surface wiggle in the OC:N-ratio profile is not certain. Nevertheless, Timothy and Pond (1997) found kc > ftjv for a portion of their study, and Harvey et al. (1995) observed that phytoplankton carbohydrates were lost more rapidly than proteins or lipids under oxic conditons during laboratory experiments of organic decay. The power function is commonly used to describe water-column fluxes of setling material (review by Bishop, 1989) but to my knowledge a treatment such as presented here (using the power function to describe changes in k with depth, and Equation 4.14 for changes in flux with depth) has not been made previously. Although the shape is somewhat diferent between the curve for organic carbon flux in Jervis Inlet and oceanic Chapter 4. A model to interpret increases in flux with depth 134 results, the two sets of curves give very similar amounts of decay over a 500 m depth interval. The faster fal-off of decay in Jervis Inlet, compared to the oceanic curves, is due to the smal rate constants in the deep waters of Jervis Inlet and not the fitting procedures performed here, as the deepwater rate constants were very wel fit by the power function. These smal rate constants for OC and N in the deep waters of Jervis Inlet, compared to the oceanic rate constants implied from descriptions of flux (e.g.; Martin et al., 1987), are unlikely a model artifact because the model best described deepwater sediment-trap flux. In order to evaluate the biogenic silica rate constants from Jervis Inlet, they can be converted to time-dependent dissolution rates (Vdiss; h-1) by multiplying by the sinking rate. Using a nominal sinking rate of 100 m d_1 (Suess, 1980; Fowler and Knauer, 1986; Aldredge and Gostchalk, 1988), the kSi values of Table C.l were 0.02 h"1 at 110 m and decreased to 0.002 h_1 at 450 m. These estimates are similar to those measured in the coastal waters of Northwest Africa (0.004-0.023 h_1; Nelson and Goering, 1977), which were the highest dissolution rates reported in a compilation by Ragueneau et al. (2000). The conversion from kSi to Vdiss is dependent on the sinking rate and an interesting possibilty is that sinking was less than 100 m d_1 in the upper water column and accelerated into deeper waters. However, it is likely that much of the curvature in the depth profile of kSi (Figure 4.4) was in fact caused by decreasing Vdiss with depth, as Sanceta (1989c) found that weakly silicified diatoms did not survive the fall into deep waters as wel as did the larger diatoms. As noted above, this shift to more heavily silicified taxa occured in waters with relatively homogeneous temperature and ambient silicic acid concentration. Chapter 4. A model to interpret increases in flux with depth 135 composition of additional flux (relative to mass flux at z2) -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 100 [• depletion in additional flux excess in additional flux Al BSi 200 [• E 300 o. Az + -£-(Al2 - Ah) . (B.4) Equation B.4 simplifies to: < a 5 > Walsh et al. (1988b) used Equation B.5 to estimate decay rates of organic carbon, biogenic silica and CaC03 for the deep Equatorial North Pacific, and Noriki and Tsunogai (1986) used a form of this model on the same fluxes from data colected in the Pacific and Southern Oceans. Walsh et al. (1988b) noted that an assumption of Equation B.5 is compositonal similarity between additional fluxes and the bulk material reaching the deep sediment traps. However, they also demonstrated (Walsh et al., 1988a) that the increase in flux with depth as the botom boundary layer of the ocean is approached was caused by material that was indeed more similar to the material sinking from above than to botom sediments. They postulated that a 'rebound' flux of relatively young sediments caused observed increases in flux with depth, and others have found additional fluxes less like botom sediment than expected if resuspended material were causing increases in flux with depth (Smetacek, 1978; Walsh and Gardner, 1992; Timothy and Pond, 1997). An alternative to the rebound hypothesis is that upper sediment traps catch material less eficiently than deep traps, causing an apparent, but not a real, increase in flux with depth (Smetacek, 1978; Timothy and Pond, 1997; Yu et al., 2001). A p p e n d i x C Sensi t ivi ty analysis and results for Jervis Inlet Data from some deployment periods heavily weighted the model results for Jervis Inlet. In general, the decay terms (e~kiAz) were more sensitve to data removal than were the terms describing the compositon of the additional flux ({j/J}d) and, furthermore, kc (and kN) were more sensitve than ksi- Also, while (j/J)d has localy useful information about sedimentation in Jervis Inlet, the decay constants are more important for elemental budgets and can be applied to other temperate coastal setings. Therefore, solutions for e-kcAz ^ n e oc decay terms) were used to find records from deployment periods that largely afected model results. The method used was to remove data from one deployment period at a time, succes-sively from the first to the last deployment, and to re-solve the model of Equation 4.6a. If the removal of a certain deployment period resulted in a significantly diferent solution, that deployment was removed from the data set and the exercise was repeated. The crite-rion used to determine whether to reject a deployment period's data was whether removal of those data caused a change in e~kcAz of more than 30% (dashed lines in the folowing figures). Figures C.l through C.6 show the sensitivity analysis, and Figure C.7 shows the deployment periods that were permanently removed from the modeling exercise in Jervis Inlet. 175 Appendix C Sensitivity analysis and results for Jervis Inlet 176 0.35 * 0.25 l < u 0.2 0.15 b1. JV3: 50-300 m-> OC data, final solution (n=37) -i—i—i—r~i—i—i—i—i—i—i—i—i—i—i—r—i—i—i—i—i—i—i—i—i—i—m—i—i—i—i—i—i—i—i r~ 0.8 0.6 5 10 15 20 25 30 35 b2. JV3: 50-300 m-> BSi data, final solution (n=37) -I—I—I—I—l—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I—I I I I T" 0.4 h -10 15 20 25 30 35 number of deployments (n) Figure Cl: Sensitivity analysis for station JV-3, 50-300 m. al: e~kcAz when successive deployment periods are removed from the entire data set. a2: One period has been permanently removed and the sensitivity analysis repeated (and again for a3). bl and b2: Sensitivity analysis for OC and BSi on the final subset used for this depth interval. Appendix C Sensitivity analysis and results for Jervis Inlet 177 a. JV3: 50-600 m-> all OC data (n=40) number of deployments (n) b1. JV3: 50-600 m-> OC data, final solution (n=37) -i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i i i i—l—i—i—i—i—i—i—i—i—i—(—i—i—i—i—i—r-0.4 0.3 \ CD 0.2 h : 5 10 15 20 25 30 35 b2. JV3: 50-600 m-> BSi data, final solution (n=37) ~1 I I I 1 I I I I 1 I i i 0.8 h : ~1—I—I—I—I—I—I—I—1—1 0.6 0.4 h : _ l I I I I I I U . __l 1 I I I I l _ 10 15 20 25 30 35 number of deployments (n) Figure C.2: Sensitivity analysis for station JV-3, 50-600 m. a: e~fccAz when successive deployment periods are removed from the entire data set. bl and b2: Sensitivity analysis for OC and BSi after removing the three deployment periods that were taken out of the 50-300 m depth interval (Figure C.l). Appendix C Sensitivity analysis and results for Jervis Inlet 178 a1. JV3: 300-600 m—> all O C data, final solution (n=40) i—i—i—i—i—i—r- I—i—i—i—i—i—i—i—i—r" 0.8 0.6 0.4 5 10 15 20 25 30 35 40 a2. JV3: 300-600 m—> all BSi data, final solution (n=40) -i—i—i—i—i—i—i—r- -l—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r 1.2f < 1 V 0.8 0.6 10 15 20 25 30 35 40 number of deployments (n) Figure C.3: Sensitivity analysis for station JV-3, 300-600 m. The entire data set is used for this depth interval. Appendix C Sensitivity analysis and results for Jervis Inlet 179 a1. JV7: 50-200 m-> all OC data (n=47) 5 10 15 20 25 30 35 40 45 a2. JV7: 50-200 m ^ O C data (n=45) 5 10 15 20 25 30 35 40 45 a3. JV7: 50-200 OC data (n=44) 5 10 15 20 25 30 35 40 number of deployments (n) b1. JV7: 50-200 m-> OC data, final solution (n=40) —i—i—i—i—i—i—i i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—l—i—i—i—i—i—i—i—i—i i i l r 0.2 0.1 5 10 15 20 25 30 35 40 b2. JV7: 50-200 r r w BSi data, final solution (n=40) 0.8 0.6 0.4 i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 10 15 20 25 30 35 40 number of deployments (n) Figure C.4: Sensitivity analysis for station JV-7, 50-200 m. al: e~kcAz when succes-sive deployment periods are removed from the entire data set. a2: Two periods have been permanently removed and the sensitivity analysis repeated (and, again, one period removed for a3). bl and b2: Sensitivity analysis for OC and BSi on the final data subset. Appendix C Sensitivity analysis and results for Jervis Inlet 180 a. JV7: 50-450 m-> all OC data (n=48) 5 10 15 20 25 30 35 40 45 number of deployments (n) b1. JV7:50-450 m->OC data, final solution (n=40) 0.14 N 0.12 * 0.1 l < u 0.08 0.06 I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I i i i i I i i i i t i i i i i i i i i i i t i i i 5 10 15 20 25 30 35 40 b2. JV7: 50-450 m-> BSi data, final solution (n=40) i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i < 0.8 \ 0.6 0.4 [ • 5 10 15 20 25 30 35 40 number of deployments (n) Figure C.5: Sensitivity analysis for station JV-7 , 50-450 m. a: e~kcAz when successive deployment periods are removed from the entire data set. b l and b2: Sensitivity analysis for O C and BS i after the seven deployment periods that were taken out of the 50-200 m depth interval (Figure C.4), plus one more, have been removed from this depth interval. Appendix C Sensitivity analysis and results for Jervis Inlet 181 a1. JV7: 200-450 m -> all O C data, final solution (n=47) i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 0.8 h 0.6 0.4 1.2 1 0.8 0.6 5 10 15 20 25 30 35 40 45 a2. JV7: 200-450 m - » all BSi data, final solution (n=47) i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 5 10 15 20 25 30 35 40 45 number of deployments (n) Figure C.6: Sensitivity analysis for station JV-7, 200-450 m. The entire data set is used for this depth interval. Appendix C Sensitivity analysis and results for Jervis Inlet 182 1985 1986 1987 1988 1989 1985 1986 1987 1988 1989 Figure C.7: Data removed from shalow-mid and shalow-deep analyses (dark bars). Most of these deployment periods are characterised by high OC fluxes to the shalow sediment traps and large decreases with depth of the OC rain rate. In Sept/Oct, 1985, while the 50-200 m data from station JV-7 did not stand out in the sensitivity analysis, the data from the 50-450 m depth interval did. The entire time series was used for the mid-deep depth interval at each stations. Appendix C Sensitivity analysis and results for Jervis Inlet 183 50-200 m 50-300 m 50-450 m 50-600 m 200-450 m 300-600 m JV-7 JV-3 JV-7 JV-3 JV-7 JV-3 n = 40 n = 37 n = 40 n = 37 n = 47 n = 40 organic carbon solution r 2 0.76 0.76 0.72 0.85 0.85 0.97 z l / 2 94 134 154 248 311 442 *c 0.012 (12) 0.0056 (22) 0.0058 (15) 0.0022 (15) 0.0018 (19) 0.00075 (21) % O C d 7.9 (12) 7.8 (16) 5.7 (10) 4.4 (7.7) 5.5 (7.7) 4.5 (3.2) . O C int 37 (17) 27 (40) 35 (26) 30 (34) 6.5f(M00) -5.8f(M00) biogenic silica solution r 2 0.92 0.95 0.79 0.90 0.86 0.97 Zl/2 114 158 235 293 320 445 0.0040 (12) 0.0022 (11) 0.00076 (30) 0.00084 (14) 0.00069 (35) 0.00046 (22) % B S i d 40 (12) 45 (11) 30 (14) 22 (9.0) 22 (15) 20 (4.4) B S i int -81 (32) -190 (21) -160 (38) -70f(66) -38f(95) -16f(M00) nitrogen solution r 2 0.79 0.79 0.65 0.80 0.72 0.94 z l / 2 98 133 156 242 307 439 0.011 (8.4) 0.0058 (19) 0.0056 (13) 0.0024 (15) 0.0024 (21) 0.00099 (20) % N d 0.91 (12) 1.0 (14) Q.61 (12) 0.46 (9.1) 0.66 (11) 0.49 (4.3) N int 3.3 (22) 2.2f(54) 4.3 (26) O.Of(MOO) 1.6f(73) 0.20f(MOO) a lumin ium solution r 2 0.89 0.56 0.88 0.95 0.94 0.98 z l / 2 124 171 253 333 326 451 0.0004 (96) 0.0006 (MOO) -0.0002 (MOO) -0.0002 (M00) -0.0001 (MOO) -0.00008 (>10C % A l d 0.96 (44) 1.7 (31) 3.5 (14) 5.8 (5.1) 4.4 (7.9) 6.1 (2.4) A l int 8.1 (35) 22 (24) 4.9f(M00) -10 (52) 0.30f(M00) -9.4f(57) Table C . l : Results of the model applied to each depth interval in Jervis Inlet after the sensitivity analysis was used to remove some deployment periods. Standard errors (percent of the value represented) are in parentheses and results that are not significantly different from zero (P > 0.05) are tagged (t; for a rate constant of zero, e~kjAz of the model is one and, therefore, significant). zx/2 is the depth between z\ and z 2 at which half of the decay of the anticipated flux has occurred (where jn = 0.5{1 -f- e~k^Z2~Zl^}) and is used as the reference depth for which decay constants are characteristic. Rate constants (kj) have units m - 1 , intercepts are mg m - 2 d - 1 . A p p e n d i x D Sensi t ivi ty analysis and results for Saanich Inlet This Appendix shows the results of the sensitivity analyses for the data from Saanich Inlet. For a description of the procedure, see Appendix C. On the whole, the data of only one deployment period and one depth interval (station SN-0.8, 135-180 m; Figure D.6) were removed from final analyses. From the sensitivity analyses and the regression coeficients (compare r2 of Tables C.l and D.l), it appears that the model described the data from Saanich Inlet relatively wel, yet the rate constants are spatialy inconsistent and diferences between constiuents are not easily explained. For instance, for the 45-110 m depth interval at station SN-9, rate constants are negative and, comparing the 45-150 and the 110-150 m depth intervals, the rate constants increase with depth (Table D.l). At station SN-0.8, although kc and kN are spatialy reasonable (decreasing with depth) and of similar magnitude, ksi increases with depth and, for the 135-180 m depth interval, the magnitude of the rate constants (kM > kSi > kc > kN) is the reverse of that expected. Other evidence that the model results from Saanich Inlet are suspect is the compositon of the additional flux. Although not shown explicitly (e.g.; as is for Jervis Inlet at Figure 4.6), for each depth interval in Saanich Inlet, OC, N, BSi and Al are all depleted in the additional flux when compared to the total mass flux reaching the deep sediment traps. In the real case, constiuents that are depleted must be balanced by others in excess. Thus, the model appears to be partitioning too much material into the anticipated flux to depth z2 in Saanich Inlet. If this is the case, the rate constants are too smal (resulting in a large anticipated flux), 184 Appendix D Sensitivity analysis and results for Saanich Inlet 185 and the size of the additional flux (parameterised by {]/ J}d) is too smal. A number of factors might have contributed to these peculiar results from Saanich Inlet. The depth intervals between sediment traps was much less than in Jervis Inlet and possibly too smal to resolve decay of the setling flux. At station SN-9, extremely high additional fluxes occured for the two depth intervals anchored above by the 45 m sediment trap. Determining the size of the anticipated flux for these intervals may be inaccurate because it is the diference between two large numbers (Equations 4.1a and 4.16). Horizontal gradients in the export flux, combined with horizontal currents, might have occured such that the construction ofthe model (Figure 4.1) was not valid, especialy at station SN-9 where large horizontal gradients in surface biomass (e.g.; Hob-son and McQuoid, in press) and significant horizontal transport of resuspended material occured. However, Equations 4.8 and 4.9 suggest that horizontal advection, enhancing or diminishing the downward flux from one depth to the next, would result in a positive (advection from high export region) or a negative (advection from regions of low export) eror term without necessarily compromising estimates of rate constants. It is unlikely that seasonal anoxia afected the sinking flux; only the 180 m sediment trap at station SN-0.8 would have been exposed to anoxic waters for extended periods of time. Indeed, Thunel et al. (2000) found that decreases with depth of POC fluxes in the anoxic waters of the Cariaco Trench were not diferent than those predicted from oxygenated setings, and certain results (high rate constants for Al, depletion of all constiuents in the additional fluxes) cannot be explained by low redox conditons. Appendix D Sensitivity analysis and results for Saanich Inlet 186 a1. SN9: 45-110 m -> all O C data, final solution (n=61) •I I M I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I II I I I I I I I I I 1 I I I I I I I I 5 10 15 20 25 30 35 40 45 50 55 60 a2. SN9: 45-110 m -> all BSi data, final solution (n=61) 1 6 -1 4 -1 2 -1 -0 8 -0 6 -5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) Figure D.l: Sensitivity analysis for station SN-9, 45-110 m. No data were removed from final analysis. a1. SN9: 45-150 m-> all O C data, final solution (n=62) i i i i i i i 1 i i i i i i i i i i i i i i i i i 111 i i 1 i i i i i 1 i i i i i i i i 1 i i i i i i i i i • • i • • i • • '• i • • i i • i i i i i • • • • • • • ' • • • ' ' ' • 5 10 15 20 25 30 35 40 45 50 55 60 a2. SN9: 45-150 m-» all BSi data, final solution (n=62) I I I I 1 1 1 I 1 I T I I I 1 I I 1 1 I "I I I M I I 1 II I I 1 I I I I I I l~T niTTI 1 I 1 I I 1 I I I I I i i i i i i i i i j j i j i i i i i i 1 i i i i i i i i i i i I I I I I t i i I I I i 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) l . D 1.4 1.2 1 0.8 | 0.6 1 -0 8 -0 6 -0 4L 1 4 -1 2 -1 -0 8 -0 6 -Figure D.2: Sensitivity analysis for station SN-9, 45-150 m. No data were removed from final analysis. Appendix D Sensitivity analysis and results for Saanich Inlet 187 0.8 h 0.6 h 0.4 a1. S N 9 : 1 1 0 - 1 5 0 m - > all OC data, final solution (n=61) I I l I I I I I I I I I I I I I I I I I I I ) I I i i i i i T I T I 1 II t t I t I T I I I 1 I 1 I I I I I I I I ' ' 5 10 15 20 25 30 35 40 45 50 55 60 a2. SN9: 110-150 m-> all BSi data, final solution (n=61) 1.2 1 0.8 0.6 i i i i i t i i i i I I i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) Figure D.3: Sensitivity analysis for station SN-9, 110-150 m. No data were removed from final analysis. a1. SN0.8: 50-135 m-> all O C data, final solution (n=62) i i i 11 i i i i i i i i i 11 i i i i i i i 11 i i i 11 11 i i i i i i i 11 i i i i i i i i i M 0.8 0.6 0.4 1 • ' ' • • i i , , , j 1.4 1.2 1 0.8 0.6 5 10 15 20 25 30 35 40 45 50 55 60 a2. SN0.8: 50-135 m-> all BSi data, final solution (n=62) j i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i I J 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) Figure D.4: Sensitivity analysis for station SN-0.8, 50-135 m. No data were removed from final analysis. Appendix D Sensitivity analysis and results for Saanich Inlet 188 N 0 8 < 0.6 0.4 a1. SN0.8: 50-180 all OC data, final solution (n=63) l ; l l l l l l I I I M l l l l l l l l l ) l ) l l l l I I I l l l l l l l l l l l l l l l l I I I l l l l l l l ) 5 10 15 20 25 30 35 40 45 50 55 60 a2. SN0.8: 50-180 m-> all BSi data, final solution (n=63) 1.2 1 r 0.8 0.6 i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i I J 1 • • ' • • • • * 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) Figure D.5: Sensitivity analysis for station SN-0.8, 50-180 m. were removed from final analysis. a. SN0.8: 135-180 m-> all OC data (n=63) i i i i i i i i i i i i i i i i i i i i i i i t i i i i i i i i i i i i i 1.2 1 0.8 0.6 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) b1. SN0.8: 1 35-180m-> OC data, final solution (n=62) 1.4 1.2 0.8 0.6 1.2 1 0.8 0.6 I I I I 1 I I I I rrr-r-i-T-i r i i i i i 1 i i i i i i i ' 5 10 15 20 25 30 35 40 45 50 55 60 b2. SN0.8: 135-180 m-> BSi data, final solution (n=62) i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i 5 10 15 20 25 30 35 40 45 50 55 60 number of deployments (n) Figure D.6: Sensitivity analysis for station SN-0.8, 135-180 m. Although it passed the 30% criteria used in Jervis Inlet (Appendix C), One deployment period's data (a) stood out and was removed from the final analysis (bl and b2). Appendix D Sensitivity analysis and results for Saanich Inlet 189 45-110 m SN-9 n = 61 50-135 m SN-0.8 n = 62 45-150 m SN-9 n = 62 50-180 m SN-0.8 n = 63 110-150 m SN-9 n = 61 135-180 m SN-0.8 n = 62 organic carbon solution Zl/2 %ocd O C int Zl/2 %BSi d BSi int 0.86 78 -0.0011 (78) 2.9 (12) 12t(M00) 0.97 78 -0.0015 (25) 16 (7.5) -180f(52) 0.89 90 0.0028 (19) 5.1 (12) 1.2f(>100) 0.96 92 0.00036 (83) 26 (11) -120 (31) 0.96 94 0.0024 (19) 2.7 (4.0) 29f(54) 0.71 108 0.0032 (20) 4.8 (16) 17f(95) biogenic silica solution 0.98 97 0.00017 (MOO) 15 (4.0) -84f(80) 0.89 112 0.0014 (25) 15 (24) -33f(>100) 0.94 128 0.011 (15) 2.7 (6.5) 79 (28) 0.99 129 0.0030 (18) 14 (4.0) 64f(82) 0.91 157 0.00082 (>100) 6.0 (7.5) -0.30f(M00) 0.90 157 0.0027 (37) 20 (20) 1.8f(>100) nitrogen solution z l / 2 % N d N int r % A l d A l int 0.88 78 -0.0014 (54) 0.32 (14) -O.40f(M00) 0.94 77 0.00080 (>100) 6.68 (3.5) 16f(>100) 0.89 90 0.0033 (16) 0.58 (14) 1.2t(M00) 0.85 97 -0.0052 (19) 3.2 (14) -3.3f(>100) 0.94 94 0.0023 (21) 0.31 (5.2) 1.4f(>100) 0.71 108 0.0033 (19) 0.57 (18) 2.6f(78) aluminium solution 0.96 94 0.0023 (62) 6.8 (2.6) 47f(64) 0.70 120 -0.0023 (43) 2.5 (20) 4.0t(M00) 0.93 128 0.011 (16) 0.32 (7.9) 8.8 (32) 0.99 130 0.00075 (>100) 7.0 (2.1) -19f(>100) 0.91 157 0.00052 (MOO) 0.79 (8.3) -0.30f(M00) 0.77 156 0.0047 (35) 2.3 (20) 10f(53) Table D.l: Results of the model applied to each depth interval in Saanich Inlet. From the sensitivity analyses, one deployment period (135-180 m depth interval at station SN-0.8) was removed. Standard erors (percent of the value represented) are in parentheses and results that are not significantly diferent from zero (P > 0.05) are tagged (f; for a rate constant of zero, e~kjAz of the model is one and, therefore, significant, so rate constants not diferent than zero are not tagged). z i / 2 is the depth between zi and z2 at which half of the decay of the anticipated flux has occured (where jn = 0.5{1 + e~k^Z2~Zlty). Rate constants (kj) have units m-1 and intercepts are in mg m~2 d-1. A p p e n d i x E E x p o n e n t i a l fit t o r a t e cons t an t s Originaly I had intended to fit an exponential function to the depth profiles of k. The reasoning was simply that the curvature of these profiles might be wel explained by an exponential function. However, the power function discussed in Chapter 4 describes the profiles beter, especialy the plot of k$i versus depth. The power function, furthermore, has more physical meaning than the exponential function, as it describes the realistic case where the decaying material (OM or BSi) is made up of a suite of components, each decaying at a diferent rate (Middleburg, 1989). Although the power function is the beter model with which to describe variations in k with depth, the formulation of the exponential fit is given here. Figure 4.4 shows the exponential fits to the rate constants, and Figure 4.5 uses Equation E.5 to compare the description of flux with depth arrived at using the power function and the exponential model of k versus depth. If the decay parameter k is modeled to decrease with depth at a constant rate, a dk ^- = -ak. (E.l) dz Integrating Equation E.l gives the exponential relationship: k = k0 e~az , (E.2) where k0 is the value of & at a reference depth, which can be chosen arbitrarily because the shape of this first-order exponential function does not change with depth. Taking the logarithm of Equation E.2 provides a means to find ko and a of the rate constants 190 Appendix E Exponential At to rate constants 191 presented in Figure 4.4. \n(k) = -OLZ + ln{k0) (E.3) Thus, for a plot of \n(kj) versus z, the slope of the regression line is —a and the intercept at z = 0 is \n(k0) (Figure 4.4). Change in the flux of j with depth is writen as: f = ~kj . (E.4) dz Substiuting k of Equation E.2 into Equation E.4 and integrating gives an expression for the change with depth of j for the case where k decreases exponentialy with depth: (E.5) 3=j\ exp [e-^-e-^Yj A p p e n d i x F Tabula t ion of p r imary product ion data This appendix presents the primary production data collected during the study (see section 2.2 for methods). Tables F . l through F.4 give 1 4 C-uptake rates (mg C m " 3 d _ 1 ) at the depths of sampling (z: meters); section 2.2 explains how hourly rates were converted to the daily rates presented here. The header at the top of each block of data in Tables F . l through F.4 is the date of sampling (yymmdd); depth and 1 4 C-uptake rates are given in the left and right columns of each block. The first 5 depths in each block are those that corresponded to 56, 32, 18 and 7% surface irradiance. The last depth in each group is an extrapolated estimate of the depth of 1% surface irradiance, and the corresponding rate of 1 4 C uptake is also an extrapolation (see section 2.2 for details). The 1 4 C-uptake value under the line at the end of each block is the vertically integrated estimate of primary production (mg C m - 2 d _ 1 ) , obtained by trapezoidal integration from the surface to the depth of 7% surface irradiance (where the 1 4 C-uptake rate at the depth of 56% surface irradiance was extrapolated to the surface) and by assuming that primary production decreased exponentially with depth in proportion to light between 7% and 1% surface irradiance. See section 2.2 for the manner in which extinction coefficients for light were estimated. 192 Appendix F Tabulation of primary production data 193 z 1 4 c Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C 850807 851104 851216 860127 860310 860414 860512 860605 1 284 1 31.2 2 18.7 1 7.18 1 45.5 2 23.7 1 2791 1.5 870 3 162 3.5 6.77 4.5 6.59 2 10.2 2 91.9 4.5 9.67 1.5 2677 2.5 831 4.5 197 5.5 10.2 7.5 10.4 4 10.8 4.5 73.0 7 1.06 2.5 2026 4 717 7.5 20.6 9 6.10 12 8.24 6 16.1 6.5 46.3 9.5 0 3 1651 5.5 392 14 30.9 16 6.10 20 6.59 12 13.8 9 35.3 12 0 4 791 6.5 229 23.6 4.41 27.3 0.87 33.8 0.94 20.2 1.96 15.9 5.05 20.9 0 6.6 113 11.3 32.8 1627 197 237 204 648 104 9610 4942 860714 860805 860908 861014 861112 861208 870120 870216 2.5 858 1.5 2094 1 883 1 230 1.5 12.4 1 13.1 1.5 18.0 1.5 19.0 4 384 2.5 1705 3 182 2 206 3 9.32 2 4.72 3 7.94 3 8.84 5.5 291 3.5 1778 4 26.8 3.5 205 5 9.32 3.5 9.10 5 2.81 4.5 1.90 7.5 129 5 1860 5 34.4 4.5 152 7 7.46 4.5 5.69 7 1.22 6 0 9 74.3 6 694 6.5 4.69 6 57.9 10 6.84 5.5 3.01 8 0.0 7.5 0 15.1 10.6 10.2 99.1 11.2 0.67 10.3 8.27 16.9 0.98 9.7 0.43 14.4 0.0 12.9 0 4362 12094 2121 1202 113 49.6 61.8 58.9 870309 870504 870615 870713 870826 870921 871019 871116 1 17.1 1.5 17.8 1.5 173 2 254 1.5 73.5 1.5 949 1 26.5 1 98.0 2 8.90 3.5 17.8 3 193 3 144 4 101 3 843 2.5 22.1 2 103 3.5 7.12 6 11.9 4.5 98.6 4.5 124 6 95.0 4 497 5 9.58 3 63.7 5 0 8.5 8.89 6.5 141 6 58.7 7.5 53.8 4.5 1555 6.5 11.8 4 51.9 7 0 11 5.93 7 113 7 28.8 10.5 15.2 5.5 379 9 5.90 5.5 15.3 12.0 0 19.3 0.85 12.7 16.2 11.9 4.12 17.9 2.18 9.3 54.1 15.8 0.84 9.2 2.19 47.4 165 1329 1153 790 5569 158 416 871214 880105 880201 880229 880328 880425 880524 880627 1.5 6.06 1.5 22.4 1.5 8.19 1.5 5.86 1 36.9 1 602 1 1130 1 365 3 0 3.5 12.3 3 7.45 3 3.73 2.5 38.7 2 852 2 917 2 300 4.5 0.46 5 5.16 5 3.36 5.5 4.27 4 41.4 3.5 1417 3 822 3 179 6 0 7 0 7 1.16 7.5 9.06 6 15.2 4.5 674 4 384 4 136 8 0 10 0 9 0 10.5 6.66 7.5 3.77 5 228 5 533 5 79.8 13.5 0 16.8 0 15.6 0 18.0 0.95 13.3 0.54 9.1 32.6 8.6 76.1 8.6 11.4 14.3 86.5 40.5 85.2 234 4698 4932 1329 880808 880916 881017 881115 881212 890104 890130 890306 1 321 1 109 2 17.5 1.5 7.87 1.5 4.19 1.5 5.13 2 5.22 1 19.7 2.5 261 3.5 73.9 4 7.71 3 3.31 4 1.07 3 1.43 4 5.22 2.5 16.6 4 196 5 31.8 6 3.44 5 0.79 6.5 1.40 5 0 6 4.31 4 6.83 5.5 113 8 18.0 9 0 7 0 9 0 7 3.80 7.5 3.82 6 0.17 6.5 72.3 10 10.3 13 0 9 0 11 0 9 2.54 9 2.31 7 0 11.7 10.3 17.9 1.47 21.8 0 15.6 0 19.7 0 15.6 0.36 15.6 0.33 12.7 0 1583 555 76.6 25.1 17.7 31.6 47.8 71.5 890403 890501 890704 890828 891010 1.5 56.5 1.5 303 1.5 26.9 1 1436 2 26.6 3 53.3 3 275 3 109 2.5 103 4 45.6 5 39.2 4.5 140 5 29.0 4 72.7 6 57.2 7 29.1 6.5 75.3 7 0 5.5 75.8 9 25.8 8.5 13.1 9 9.37 9.5 0 8.5 43.7 12.5 5.03 15.0 1.87 15.2 1.34 16.3 0 14.3 6.24 21.1 0.72 396 1546 310 3129 426 Table F.l: Volumetric (mg C m~3 d_1) and areal (mg C m 2 d l) rates of UC uptake versus depth (z: meters) at station SN-9. See text of this appendix for details. Appendix F Tabulation of primary production data 194 z 14C Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C 850809 851104 851216 860127 860310 860414 860512 860605 1 583 1.5 3.97 1.5 40.7 1 47.3 1 195 1 21.0 1 795 1 316 2 504 3 9.92 3.5 32.2 2.5 43.9 2 108 3.5 6.67 2.5 593 2 189 3 290 4 3.97 6 20.3 4 54.1 4 71.1 6 0 4.5 265 3.5 143 6 24.2 6 3.97 11 18.6 6 32.1 6 62.6 9 0 6 117 5 71.5 15 0 13 1.98 20 i:69 12 28.7 10 53.0 12 0 7 96.6 6 23.1 25.4 0 21.1 0.28 34.2 0.24 20.0 4.10 17.0 7.57 21.4 0 12.8 13.8 10.7 3.30 2104 59.6 399 563 1058 63.8 3323 1073 860714 860805 860908 861014 861112 861208 870120 870216 1.5 699 1.5 448 1 900 1.5 233 1 8.42 1 26.3 1 15.8 1.5 27.3 3 161 3 320 2 1965 2.5 203 2 0 2 11.1 2.5 1.16 3 20.1 4.5 152 5 224 3 591 4 131 3 0 3.5 11.6 4 0.44 5.5 9.65 5.5 38.4 6.5 331 3.5 182 5.5 91.6 5 0 5 4.56 6 2.25 7.5 10.5 7 0 8 92.0 4.5 182 6.5 8.28 6.5 0 6.5 0 8 1.09 9.5 7.16 11.9 0 14.0 13.1 7.6 26.1 11.3 1.18 11.3 0 11.3 0 14.0 0.16 16.7 1.02 2052 2768 4244 1051 12.6 77.5 38.5 174 870309 870406 870504 870615 870713 870826 870921 871019 0.5 25.3 1 26.8 1.5 39.6 1 136 2.5 529 1.5 192 1 988 1.5 101 1.5 13.7 2 36.1 4 18.2 3 53.7 4.5 1081 2.5 57.6 2 1014 3 94.4 3 10.6 3 141 7 20.3 4.5 45.5 6 1149 4.5 35.2 3 1013 5 77.2 4.5 2.64 4.5 78.6 10 7:48 5.5 57.9 8 145 6 6.40 4 398 7 21.3 5.5 0.32 7 35.1 13 3.21 6.5 57.9 9 106 8 0 5 371 9 8.21 10.2 0 11.6 5.02 23.1 0.46 11.7 8.27 15.6 15.1 13.7 0 8.6 53.1 15.6 1.17 62.4 528 261 639 6328 544 4684 621 871116 871214 880105 880201 880229 880328 880425 880524 1 137 2 0 1.5 55.7 1.5 43.0 1.5 7.48 1.5 46.0 1.5 447 1 407 2 59.6 6 0 3 29.3 3 32.6 3 9.05 3 32.0 3 811 3 365 3 49.4 10 0 4.5 12.6 4.5 21.0 5 6.91 5 19.2 4.5 582 4 347 5 33.0 13.5 0 7 2.04 6.5 12.1 7 7.48 7 0.27 5.5 365 6 309 7 5.83 17 0 9.5 0:40 8 4.07 9.5 0.33 9.5 0 7 98.6 7 255 12.0 0.83 30.4 0 16.2 0.06 13.9 0.58 16.3 0 16.3 0 11.9 14.1 12.6 36.4 424 0 201 217 64.7 198 3696 3078 880627 880808 880916 881017 881115 881212 890104 890130 1 202 1.5 213 1.5 38.3 1 43.0 1.5 0 1 0 1.5 1.49 1.5 0.41 2.5 157 3 187 3 27.4 2 28.2 3 0 4 0 3.5 1.10 3.5 0.82 4 85.6 4 81.2 5 15.1 3 29.5 4.5 0 5 0 4.5 1.17 6 0 5.5 141 6 71.3 7 17.1 5 13.5 6.5 0 7 0 7 0 9 0.26 7 44.6 7 7.24 9.5 0 8 2.70 8.5 0 11 0 11.5 0 14 0.29 12.3 6.37 12.2 1.03 16.3 0 13.4 0.39 14.5 0 18.5 0 19.0 0 23.8 0 1064 961 203 181 0 0 7.42 5.95 890306 890403 890501 890704 890828 891010 2 16.3 1.5 23.8 1.5 94.6 2 69.1 1 684 1 286 4 17.7 3.5 30.1 3.5 93.9 4 67.0 1.5 948 2 185 5.5 16.9 7 16.1 6 55.5 6.5 45.4 2 282 3 10.9 7 7.70 9 5.19 8.5 79.6 8.5 34.6 3 0 4 0 10.5 5.56 11 0 11 71.5 11 29.2 4.5 201 5 0 17.1 0.79 20.0 0 19.3 10.2 18.9 4.18 7.3 28.7 8.6 0 151 197 1133 677 1953 625 Table F.2: Volumetric (mg C m~3 d_1) and areal (mg C m 2 d x) rates of 14C uptake versus depth (z: meters) at station SN-0.8. See text of this appendix for details. Appendix F Tabulation of primary production data 195 z 1 4 c Z 1 4 C Z 1 4 C Z 1 4 C Z 1 4 C Z 1 4 C Z 1 4 C Z 1 4 C 850808 851105 851217 860128 860311 860415 860513 860604 1.5 80.1 2.5 7.94 2 24.3 1.5 29.8 1 59.1 1.5 336 1 567 0.75 670 4 127 6 5.95 4 18.3 3 26.5 3 44.6 3 264 2 438 2.5 526 5 101 9.5 13.9 8 0 5 6.63 5.5 29.0 5 179 3.5 353 3.5 653 7 59.1 13 6.61 13.5 0 8 13.3 7.5 21.8 7 99.2 4.5 186 4.5 241 13 30.0 20 6.61 25 2.03 18 19.9 10 12.4 9 59.5 5.5 108 5.5 140 21.1 4.29 33.6 0.94 42.6 0.29 30.0 2.84 17.8 1.78 15.6 8.50 9.7 15.4 9.8 20.0 1035 202 155 425 390 2007 2274 3035 860715 860806 860909 861015 861113 861209 870121 870217 1 731 1 464 1.5 362 1 70.7 1.5 26.8 1 9.80 1 9.01 1.5 19.3 2.5 469 2 547 3 359 2 62.9 3 13.0 3.5 4.60 2 9.01 3.5 16.3 4 616 3 249 5 131 4 62.9 5 8.93 6 3.24 3.5 6.43 5.5 13.4 5 227 4.5 259 7.5 56.9 5.5 12.4 7 2.68 8 2.56 6 5.02 8 8.17 6.5 48.1 5.5 111 9 45.0 7 14.8 9 0 11.5 2.19 8 3.67 10.5 5.79 11.3 6.88 9.6 15.9 16.0 6.43 12.5 2.11 15.6 0 20.1 0.31 14.1 0.52 18.2 0.83 3175 2134 2021 375 106 59.9 62.3 158 870310 870407 870505 870616 870714 870827 870922 871020 1 8.02 1 159 1 96.6 1 244 1 218 1.5 247 1 261 0.5 170 2.5 7.23 2 110 2 62.1 2 419 2 322 2.5 220 2 124 2 85.2 5 2.96 3 107 3 71.8 4 317 4 246 4 198 3.5 103 3 89.5 7 0 5 59.1 4.5 41.4 5 248 6 677 5 105 5 57.9 4 58.2 9 0 6.5 30.1 6 35.9 6.5 189 9 193 6.5 50.2 6 36.7 5 30.4 16.2 0 11.3 4.29 10.2 5.13 11.5 27.0 15.5 27.6 10.9 7.17 10.7 5.25 9.0 4.34 35.2 697 454 2330 3841 1285 865 533 871117 871215 880106 880202 880301 880329 880426 880525 1.5 22.6 1 11.0 2 3.74 2 0 2 41.9 2 72.1 1.5 128 2 111 3 17.0 3 8.75 5 3.91 4.5 0 4 23.3 3 58.3 3.5 188 3 704 4.5 13.2 4.5 8.12 7 1.69 8.5 0 6 9.84 5 36.1 5.5 99.5 4 513 6.5 6.91 7 5.87 9.5 0 12.5 0 8 6.46 7 22.1 8 70.1 5 373 8 4.65 9 7.37 14 0 16 0 10 0 9 13.0 9 37.8 6 199 13.9 0.66 15.9 1.05 23.4 0 28.5 0 17.2 0 15.3 1.86 16.4 5.41 9.9 28.5 127 96.2 26.7 0 205 434 1178 2327 880628 880809 880917 881018 881116 881213 890105 890131 1 96.2 1.5 198 1.5 104 1.5 0 2 24.5 2 8.75 1.5 13.7 1.5 8.83 3 115 3 190 3.5 57.7 2.5 0 3.5 15.7 4.5 5.56 4 11.4 3.5 8.90 4 139 5 180 5.5 51.9 4 0 5 12.7 7 4.22 6.5 11.3 5.5 5.26 5.5 132 6.5 119 7.5 37.9 6 0 7 9.49 9 3.91 9 8.30 7 1.07 7 40.1 8 56.7 9 24.7 7.5 0 8 7.04 12 2.16 12 8.68 9.5 0 12.2 5.73 14.0 8.10 16.0 3.53 12.9 0 13.9 1.01 20.5 0.31 20.9 1.24 16.2 0 856 1463 638 0 149 73.0 164 51.2 890307 890404 890502 890606 890705 890829 891011 2 16.5 1 69.4 1.5 336 2 608 1 94.5 1.5 53.8 1 58.6 4 16.1 2.5 69.9 3 511 3.5 531 2 69.3 3 65.3 2 32.6 6.5 11.2 4.5 32.3 4 453 4.5 267 4 33.8 4 53.0 4 9.41 9.5 12.2 6.5 10.6 5 128 6 78.6 5 12.6 5.5 41.2 5.5 2.54 12.5 4.92 8.5 4.01 6 70.7 7.5 13.8 7.5 5.01 7 0 8 0 21.6 0.70 15.1 0.57 10.2 10.1 12.4 1.97 12.8 0.72 11.8 0 13.8 0 180 345 2144 2829 337 331 158 Table F.3: Volumetric (mg C m~3 d"1) and areal (mg C m"2 d_1) rates of 14C uptake versus depth (z: meters) at station JV-3. Samples for the first four periods (850808 to 860128) were colected at station JV-11.5. See text of this appendix for details. Appendix F Tabulation of primary production data 196 z 14C Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C Z 14C 850808 851105 851217 860128 860311 860415 860513 860604 1.5 79.8 1.5 20.8 2.5 9.50 2.5 0 1 67.1 1 373 1 485 1 96.0 3.5 56.3 3 9.02 5.5 0 5 11.7 3 50.5 2.5 235 2 503 2 81.5 6 83.9 5.5 6.94 9.5 6.71 9 14.2 5 31.7 5 118 3.5 434 3 374 8.5 61.4 9 7.63 15 0 14 6.79 8 34.7 7 64.5 4.5 188 4.5 788 13 22.5 18 5.55 25 0 20 4.32 10 24.1 9 37.2 6 58.4 5.5 282 22.1 3.22 30.3 0.79 42.3 0 34.5 0.62 18.1 3.45 16.2 5.32 10.3 8.34 9.6 40.3 893 189 69.9 180 508 1669 2290 2324 860715 860806 860909 861015 861113 861209 870121 870217 1 216 1 344 1.5 204 1.5 40.7 1.5 61.8 0.5 15.5 1.5 11.5 1.5 13.9 3 182 2 232 3 93.5 2.5 46.5 3 27.3 2 9.78 3 7.14 3 10.2 4.5 218 3 371 5 60.0 4.5 26.8 5 18.0 4.5 5.29 5 3.61 5 3.70 6 102 4.5 313 6.5 54.4 6.5 9.08 7 6.54 6.5 5.76 7.5 1.31 8 0 7.5 0 5.5 141 8 29.3 8 1.63 9 1.36 8 4.89 11 0.92 10 0 13.3 0 9.6 20.2 14.0 4.19 14.0 0.23 15.6 0.19 15.0 0.70 18.7 0.13 17.7 0 1232 1926 907 226 241 78.8 55.2 58.3 870310 870407 870505 870616 870714 870827 870922 871020 0.5 27.3 1 227 1 137 1 205 2 96.3 2 102 1.5 152 1 163 2 11.1 2 151 2 190 2 198 3 227 3.5 111 3 69.1 2 82.3 4 9.55 3 110 4 60.1 3 142 4.5 150 5 127 4.5 38.5 3.5 20.6 6 5.12 4.5 46.6 6 45.8 4 81.6 6.5 45.9 7 84.0 6.5 14.9 5 10.7 8.5 3.41 6 10.6 7 43.9 6 54.4 8.5 17.9 8.5 37.9 8.5 7.87 6 5.10 15.2 0.49 10.2 1.51 12.9 6.27 9.9 7.78 14.2 2.56 14.5 5.41 14.5 1.12 10.7 0.73 98.2 727 810 923 944 944 572 405 871117 871215 880106 880202 880301 880329 880426 880525 1.5 17.6 1.5 10.6 2 5.10 2 0 1 31.3 1.5 95.3 1 171 1 63.9 3 16.8 3.5 7.68 4.5 6.93 4 0 2.5 17.2 3 85.6 2 161 2.5 52.7 5 3.92 6 3.14 7 3.19 7.5 0 4 7.88 5 69.2 3 149 4 58.3 7.5 2.72 8.5 0.0 10 3.19 11 0 6 4.13 7 21.8 4 174 5 71.8 9.5 1.36 11 0.0 14 3.82 15 0 8 0 9 6.32 5 98.4 6.5 46.0 16.6 0.19 19.3 0.0 23.8 0.55 26.1 0 14.0 0 15.6 0.90 8.6 14.1 11.3 6.57 89.6 51.6 78.3 0 103 571 947 485 880628 880809 880917 881018 881116 881213 890105 890131 1.5 69.6 2 137 1 138 1 26.3 2 10.9 1.5 3.87 2 3.73 1.5 5.81 3 51.3 4 102 2 161 2 20.8 3.5 7.25 3 2.52 4 2.02 3.5 0 4.5 30.5 6 130 4 127 4 7.62 5.5 6.64 4.5 4.19 6.5 0.73 5.5 3.95 6 25.6 7.5 213 5.5 109 6 5.26 8 1.21 6 4.73 10 2.02 7.5 5.16 7 9.89 9 81.5 7 42.4 8.5 1.52 10 0 8 1.93 13 1.35 10 3.73 12.3 1.41 15.6 11.6 12.5 6.06 14.8 0.22 17.2 0 13.5 0.28 22.6 0.19 17.2 0.53 339 1459 965 104 60.2 33.8 32.2 50.6 890307 890404 890502 890606 890705 890829 891011 1.5 29.5 1.5 108 1.5 273 2 89.5 1.5 76.5 1 146 0.5 114 4 29.3 3 71.2 2.5 272 3.5 91.4 3 45.4 2 175 1.5 32.9 6.5 25.4 5 21.3 3.5 229 5 245 5.5 75.2 3.5 269 2.5 6.68 9 14.9 7 14.7 4.5 167 7 212 7 44.3 5 269 4 2.71 10.5 3.46 9 9.15 6 82.9 9 93.4 8 17.5 6 43.0 5.5 0 19.1 0.49 15.6 1.31 9.8 11.8 15.1 13.3 14.6 2.49 10.7 6.14 9.7 0 263 476 1466 1588 526 1286 160 Table F.4: Volumetric (mg C m~3 d"1) and areal (mg C m 2 d l) rates of 14C uptake versus depth (z: meters) at station JV-7. See text of this appendix for details. A p p e n d i x G Sediment-trap data of Saanich Inlet This appendix presents the sediment-trap data collected from Saanich Inlet during the study (see section 3.2 for methods). Sediment traps were moored in pairs with a brine solution at the base of each. In addition, N a N 3 was used in one sediment trap of each pair while no preservative was used in the other (section 3.2). The total mass, O C and N fluxes presented in these tables are the averages of the fluxes to the two traps in each pair. BSi , A l and T i were measured on samples collected by the NaN 3 -treated sediment traps and stable isotope ratios were determined for samples collected by sediment traps without N a N 3 . The C a C 0 3 fluxes presented in these tables are those collected by the sediment traps treated with sodium azide, as N a N 3 buffers C a C 0 3 dissolution, "start" and "end" are the beginning and end of each deployment period. 197 endix G Sediment-trap data of Saanich Inlet 198 start end mass O C N BSi C a C 0 3 A l T i <513C 5 1 5 N ddmmyy mg m 2 day - 1 830809 830912 10197 531 64.2 3751 203 428 28.1 -19.9 -830912 830930 8135 498 59.3 3715 176 265 16.5 -19.1 -830930 831031 5511 309 34.2 1873 79.3 243 16.5 -20.2 -831031 831128 6812 313 32.7 1312 26.4 - - - -831128 840112 4300 163 17.0 528 44.1 267 17.8 -22.4 -840112 840209 4311 206 24.5 569 61.7 273 17.3 - -840209 840305 3226 165 21.9 527 0 204 13.1 - -840305 840409 3999 233 35.2 1499 0 174 11.4 -19.9 -840409 840510 5802 337 47.9 1920 0 275 18.1 -19.5 -840510 840618 6982 413 60.1 2943 70.5 246 16.5 -19.6 -840618 840716 9039 554 80.1 3349 132 335 21.1 -19.2 -840716 840824 8690 570 82.2 3510 52.9 321 19.8 -19.1 -840824 840919 7415 543 73.9 2095 52.9 365 22.6 -20.3 -840919 841108 . 7954 402 46.5 1803 177 391 24.1 -20.9 -841108 841213 4230 174 19.7 580 59.8 264 17.2 -22.8 -841213 850117 3495 163 18.2 504 80.0 220 13.8 -22.6 8.9 850117 850218 2550 142 16.8 391 63.1 156 10.2 -22.3 9.4 850218 850328 2656 134 14.7 525 53.2 153 9.42 -23.0 7.3 850328 850425 5891 331 45.7 2392 93.4 264 17.0 -21.1 7.3 850425 850521 4206 329 45.6 2201 69.3 102 6.56 -19.3 7.3 850521 850703 6148 419 56.3 2532 168 220 14.2 -19.4 6.9 850807 850917 9775 574 71.2 4709 165 378 24.4 -19.3 6.1 850917 851008 5571 388 49.9 2105 26.8 229 14.3 -19.9 7.1 851008 851104 4500 229 27.9 991 63.9 280 17.7 -21.1 6.9 851104 851216 3302 175 20.1 560 50.3 191 11.9 -21.9 7.5 860310 860414 4172 258 32.6 1578 56.0 195 12.3 -20.6 7.3 860414 860512 3360 211 28.9 892 48.7 182 11.5 -21.2 7.5 860512 860605 5127 363 50.4 2174 79.4 187 12.2 -20.5 7.3 860605 860714 7234 485 64.2 2075 97.2 345 21.4 -20.5 7.6 860714 860805 4067 356 48.7 1215 64.5 172 11.1 -19.8 8.4 860805 860908 7180 466 57.6 3086 70.5 242 15.5 -19.3 7.3 860908 861014 6893 542 65.0 2536 85.4 266 16.3 -20.1 7.7 861014 861112 5405 490 57.9 1527 55.8 246 15.7 -20.5 8.6 861112 861208 5437 316 37.2 793 59.9 361 23.0 -21.4 8.7 861208 870120 4451 221 25.6 594 50.1 310 20.0 -21.9 8.9 Table G.l: Sediment-rap fluxes measured at station SN-9: 45 m. endix G Sediment-trap data of Saanich Inlet start end mass O C N BSi C a C 0 3 A l T i <513C (515N ddmmyy mg m - 2 day - 1 870120 870216 3128 176 20.2 432 34.0 206 12.4 -22.0 9.4 870216 870309 4368 236 28.8 735 53.6 286 17.2 -22.9 9.4 870309 870406 3266 206 26.6 872 36.1 179 11.2 -22.4 8.5 870406 870504 6808 431 59.9 3457 34.7 193 12.3 -19.6 7.1 870504 870615 7427 420 56.4 2830 89.5 303 18.8 -19.5 5.9 870615 870713 10333 697 91.4 3748 193 362 22.4 -19.3 8.0 870713 870826 9008 577 73.8 3531 159 328 19.2 -19.7 6.3 870826 870921 7417 558 83.5 2933 116 241 15.5 -19.4 7.4 870921 871019 7053 536 61.0 3037 59.2 189 12.0 -19.7 7.0 871019 871116 4207 258 31.3 1442 49.8 175 11.2 -21.5 7.1 871116 871214 2885 147 16.9 849 45.0' 152 10.3 -22.7 -871214 880105 2647 158 18.7 352 28.1 166 10.4 -21.6 -880229 880328 3788 215 26.1 802 59.1 209 12.4 -22.2 8.6 880328 880425 7614 420 52.9 2816 51.1 282 18.0 -20.6 6.4 880425 880524 7146 497 59.2 3009 22.5 140 9.62 -19.3 6.5 880524 880627 10534 525 61.1 3452 65.0 418 26.8 -19.7 6.3 880627 880808 6166 446 58.0 1797 99.6 255 15.4 -20.0 7.0 880808 880916 6130 452 59.1 2284 110 196 12.6 -20.1 6.7 880916 881017 5754 449 54.9 2411 95.6 183 11.4 -20.1 7.7 881212 890104 3065 172 20.8 443 47.1 171 11.7 -21.4 8.8 890104 890130 3368 207 24.4 476 51.5 206 13.0 -21.8 8.8 890130 890306 4530 230 25.3 758 95.2 283 18.0 -22.3 7.2 890306 890403 2852 161 19.2 517 39.7 160 9.91 -19.8 -890403 890501 7778 488 65.6 3650 43.0 193 12.3 -19.3 6.0 890501 890605 5336 448 55.5 2292 28.7 92.9 6.45 -18.7 6.4 890605 890704 6430 508 66.2 2681 37.6 135 9.00 -18.7 6.9 891010 891215 4504 300 33.8 1431 41.9 180 12.3 -19.8 6.4 Table G.2: Sediment-rap fluxes measured at station SN-9: 45 m (continued). Appendix G Sediment-trap data of Saanich Inlet 200 start end mass O C N BSi C a C 0 3 A l T i <513C 5 1 5 N ddmmyy mg m - 2 day - 1 830809 830912 17369 765 86.8 5184 247 835 50.9 -830912 830930 13688 669 73.9 4625 344 642 40.2 -830930 831031 10451 446 49.1 2852 61.7 606 39.6 -831031 831128 - - - - - - - -831128 840112 9182 280 28.9 1134 0 590 36.7 -840112 840209 5998 211 22.2 743 - 389 24.4 -840209 840305 13064 386 39.8 1690 0 902 58.4 -840305 840409 9060 316 42.1 2447 0 528 34.6 -840409 840510 17655 608 79.4 4142 88.2 1108 73.5 -840510 840618 12376 615 86.1 4058 106 608 41.3 -840618 840716 20146 891 126 5290 167 1082 69.7 -840716 840824 16259 844 111 5124 52.9 733 49.2 -840824 840919 9504 560 76.6 2380 52.9 507 34.3 -840919 841108 12516 547 58.5 2328 162 669 42.8 -841108 841213 9567 331 39.7 1223 121 593 38.9 -841213 850117 7873 282 30.3 1036 104 495 33.4 7.6 850117 850218 11517 392 43.2 1422 129 855 51.3 7.7 850218 850328 13742 480 52.2 1862 157 1043 67.8 7.7 850328 850425 8323 408 52.0 2492 88.2 462 27.4 8.6 850425 850521 10255 551 73.9 3796 82.4 487 30.6 8.8 850521 850703 10173 611 79.8 3636 84.7 484 32.2 8.3 850807 850917 15572 765 94.1 5930 167 765 45.6 6.0 850917 851008 9320 527 65.2 2898 35.7 474 30.5 7.1 851008 851104 8960 424 48.4 1733 74.2 570 36.7 6.3 851104 851216 11499 794 79.3 1686 121 644 43.0 8.2 860310 860414 6739 352 42.1 1880 53.9 367 23.3 8.1 860414 860512 9744 395 48.7 1851 75.1 575 38.5 8.0 860512 860605 10847 594 79.1 3183 96.7 555 34.8 8.9 860605 860714 12360 707 95.4 2972 91.8 641 41.3 9.1 860714 860805 9260 571 75.9 2374 70.7 511 31.5 8.3 860805 860908 13175 768 93.5 4290 106 578 37.6 7.8 860908 861014 9845 789 92.6 2679 79.6 457 29.0 8.6 861014 861112 8196 668 79.2 1921 64.0 456 29.0 9.4 861112 861208 12913 551 62.0 1786 112 852 54.9 8.4 861208 870120 7720 343 39.0 1084 71.1 476 31.9 8.8 Table G.3: Sediment-rap fluxes measured at station SN-9: 110 m. Appendix G Sediment-trap data of Saanich Inlet 201 start end mass O C N BSi CaCOs A l T i <513C 5 1 5 N ddmmyy mg m - 2 day - 1 870120 870216 7686 378 43.8 1028 65.2 513 31.1 9.7 870216 870309 8057 423 49.9 1317 77.3 523 33.5 10.4 870309 870406 5425 327 42.6 1174 48.6 332 20.2 10.0 870406 870504 12222 595 79.4 4058 66.9 545 35.0 -870504 870615 14779 677 87.2 4316 96.9 744 48.7 7.1 870615 870713 16751 1000 136 4799 118 685 46.9 8.8 870713 870826 13005 734 90.9 4291 104 629 38.8 7.0 870826 870921 12099 745 92.4 3495 89.4 571 34.6 7.6 870921 871019 10699 640 72.7 3415 50.4 456 27.5 6.7 871019 871116 7797 369 43.7 1967 53.0 414 26.5 6.9 871116 871214 15130 516 54.5 2567 109 931 61.3 6.3 871214 880105 7300 342 37.2 1049 41.1 464 30.4 -880229 880328 13348 501 58.1 2073 65.2 852 55.4 -880328 880425 12419 502 59.1 3641 54.3 644 40.3 7.1 880425 880524 12359 689 76.4 3961 49.4 441 29.9 8.2 880524 880627 11439 633 74.2 3801 77.7 452 29.0 7.8 880627 880808 10847 618 76.5 2858 74.3 600 37.7 6.6 880808 880916 11783 764 92.4 3464 91.4 504 32.4 7.6 880916 881017 10169 618 75.7 2880 65.4 445 29.1 6.9 881212 890104 9248 335 36.6 1177 64.2 561 37.8 7.6 890104 890130 7573 306 33.0 1075 55.9 491 32.1 7.3 890130 890306 12955 492 54.9 1770 113 804 56.0 7.6 890306 890403 6927 341 41.6 1173 63.4 357 24.3 8.8 890403 890501 9776 468 59.1 3266 36.6 377 23.3 7.6 890501 890605 9754 532 65.3 3197 57.7 402 25.9 -890605 890704 13117 711 87.8 3856 53.7 453 30.9 8.0 891010 891215 7330 449 47.3 1914 49.0 335 22.7 7.2 Table G.4: Sediment-rap fluxes measured at station SN-9: 110 m (continued). Ux G Sediment-trap data of Saanich Inlet 202 start end mass O C N BSi C a C 0 3 A l T i 5 1 3 C <51BN ddmmyy mg m - 2 day - 1 830809 830912 18459 724 83.0 4854 414 960 62.5 -20.5 -830912 830930 12662 592 64.5 4151 212 593 37.9 -19.7 -830930 831031 11456 446 51.5 2791 70.5 626 41.8 -20.5 -831031 831128 7087 289 34.7 1308 0 461 30.1 -20.9 -831128 840112 8658 260 25.6 1075 52.9 554 37.2 -22.4 -840112 840209 6008 209 21.1 788 0 393 24.9 -22.0 -840209 840305 9965 291 30.4 1292 0 652 42.6 -22.1 -840305 840409 7066 264 34.6 1842 0 417 26.4 -21.0 -840409 840510 13905 468 58.4 3114 79.3 821 54.5 -20.7 -840510 840618 9626 429 57.2 2973 150 473 31.8 -20.0 -840618 840716 39631 1273 161 7821 194 2437 162 -19.7 -840716 840824 30306 1069 139 6625 79.3 1894 119 -20.5 -840824 840919 10208 521 69.9 2244 0 565 38.3 -20.9 -840919 841108 14308 530 58.5 2466 212 794 50.8 -21.4 -841108 841213 9750 317 39.0 1241 117 587 37.1 -21.8 -841213 850117 8041 277 30.2 1045 119 507 33.0 -22.7 7.2 850117 850218 9886 320 34.1 1277 119 714 44.8 -22.7 6.4 850218 850328 23082 631 64.6 3007 264 1737 to -22.8 5.9 850328 850425 8289 333 41.0 2197 89.9 512 30.9 -22.1 7.4 850425 850521 13145 536 69.7 4595 100 645 45.7 -20.6 7.1 850521 850703 12460 563 72.9 3842 113 638 43.1 -20.2 -850807 850917 18396 752 92.8 5619 204 1040 65.0 -20.3 6.7 850917 851008 15026 665 78.9 3669 148 874 54.1 -20.8 -851008 851104 16156 580 64.6 2490 157 1132 70.1 -21.6 6.4 851104 851216 14742 633 71.7 2110 142 859 56.7 -21.4 7.4 860310 860414 6924 299 34.6 1912 49.7 387 24.1 -21.3 6.6 860414 860512 9462 337 38.3 1802 57.8 638 41.2 -21.8 5.9 860512 860605 9191 385 49.2 2732 75.5 526 32.4 -21.1 6.4 860605 860714 11909 545 70.4 2719 85.4 645 40.6 -21.2 9.1 860714 860805 11138 498 61.7 2503 79.7 670 41.8 -20.9 7.9 860805 860908 26615 896 101 5564 169 1591 101 -21.0 6.3 860908 861014 14135 594 69.3 3482 135 803 50.0 -20.9 6.4 861014 861112 11665 508 60.1 2382 88.1 692 43.8 -21.2 7.1 861112 861208 12947 474 53.7 1748 114 868 58.4 -22.7 6.5 861208 870120 8205 322 36.9 1099 64.9 567 34.5 -22.3 8.0 Table G.5: Sediment-rap fluxes measured at station SN-9: 150 m. Appendix G Sediment-trap data of Saanich Inlet start end mass O C N BSi C a C 0 3 A l T i 5 1 3 C <515N ddmmyy mg m~ 2 day - 1 870120 870216 7943 292 31.0 1016 52.6 542 33.1 -22.6 7.2 870216 870309 7457 277 27.6 1102 43.1 437 28.1 -23.0 5.6 870309 870406 5080 204 22.1 1055 28.8 311 19.4 -22.8 6.5 870406 870504 10614 400 48.8 3704 36.5 496 30.6 -20.7 5.8 870504 870615 13316 489 58.0 3760 61.5 772 50.5 -20.5 5.7 870615 870713 16842 711 85.1 4825 81.4 886 56.0 -20.3 7.1 870713 870826 19208 765 96.0 5025 144 1099 67.5 -20.5 7.1 870826 870921 16580 732 90.3 4153 138 769 51.1 -20.7 7.2 870921 871019 13173 650 75.8 3800 80.7 596 36.5 -20.3 6.4 871019 871116 9030 462 54.2 2134 62.5 481 31.4 -21.4 7.4 871116 871214 13688 463 51.5 2308 93.1 844 56.1 -22.4 6.0 871214 880105 8239 302 33.0 1146 37.4 527 32.5 -22.1 -880229 880328 14590 491 52.6 2341 92.7 922 55.7 -21.6 6.1 880328 880425 16218 604 67.3 4020 90.6 895 54.9 -21.2 -880425 880524 11491 543 64.9 3846 59.2 495 33.1 -19.6 6.8 880524 880627 6931 464 56.1 2837 45.1 173 11.0 -19.0 7.3 880627 880808 12062 579 69.9 2768 92.9 664 43.1 -20.5 7.7 880808 880916 15643 718 80.6 3865 124 808 53.5 -20.6 7.0 880916 881017 11594 558 66.7 2994 110 553 36.0 -20.4 6.7 881212 890104 10241 361 38.4 1304 84.6 634 42.9 -21.5 6.9 890104 890130 6655 257 28.3 926 44.6 425 27.7 -21.5 6.9 890130 890306 13411 456 49.0 1798 119 813 54.8 -22.1 6.7 890306 890403 5554 230 25.5 920 34.8 318 21.9 -21.8 6.9 890403 890501 9699 416 49.5 3278 40.5 399 27.4 -20.7 6.4 890501 890605 8734 424 49.8 2801 43.9 330 22.2 -19.9 6.5 890605 890704 18274 729 84.0 4729 63.3 838 56.8 -19.9 7.4 891010 891215 7298 350 40.8 1851 49.7 339 22.6 -20.4 6.4 Table G.6: Sediment-rap fluxes measured at station SN -9: 150 m (continued). Appendix G Sediment-trap data of Saanich Inlet 204 start end mass O C N BSi C a C 0 3 A l T i 5 1 3 C <515N ddmmyy mg m~ 2 day - 1 840112 840209 1042 60.8 7.09 _ 17.6 - - - -840209 840305 682 50.2 6.35 111 44.1 39.5 2.34 -22.5 -840305 840409 2142 226 37.1 1229 159 36.1 2.19 -19.6 -840409 840510 3787 305 42.7 1920 0 50.3 3.02 -18.4 -840510 840618 2898 266 37.4 1615 0 25.2 1.59 -18.4 -840618 840716 2459 300 47.1 794 26.4 27.2 1.67 -19.2 -840716 840824 2385 317 47.0 996 44.1 27.2 1.50 -18.8 -840824 841006 894 122 16.8 294 35.3 11.9 0.795 -19.5 -841006 841108 1102 110 12.8 294 53.7 47.5 3.02 -21.7 -841108 841213 981 101 14.5 135 38.4 44.5 2.96 -22.4 -841213 850117 969 75.4 8.24 127 36.2 62.5 3.75 -22.9 9.3 850117 850218 973 65.6 8.38 170 43.6 56.0 3.18 -22.1 9.5 850218 850328 1027 76.1 10.4 227 44.9 54.8 3.22 -22.3 8.7 850328 850425 2896 213 29.5 1669 43.4 51.5 3.23 -20.8 6.6 850425 850521 2917 298 41.9 1885 40.8 23.2 1.52 -19.4 7.8 850521 850703 2960 326 44.0 1591 81.7 36.7 2.06 -18.9 8.9 850703 850807 1898 283 38.3 883 25.7 23.0 1.34 -19.5 8.7 850807 850917 1584 176 21.6 809 57.3 15.0 1.11 -19.0 9.2 850917 851008 1716 183 22.7 791 5.20 14.5 0.989 -19.0 8.1 851008 851104 673 79.2 10.3 177 23.9 23.5 1.52 -20.2 9.4 851104 851216 720 79.3 10.9 140 3.18 32.0 2.07 -21.2 10.6 851216 860127 1333 112 12.5 216 35.1 76.9 4.41 -21.8 8.5 860127 860310 1275 106 13.2 131 19.8 85.0 5.12 -22.2 8.8 860310 860414 1629 145 19.4 821 25.2 35.6 2.14 -20.2 9.0 860414 860512 849 116 16.0 289 25.1 23.9 1.48 -20.9 8.4 860512 860605 3035 323 40.1 1650 84.2 39.7 2.42 -20.3 8.8 860605 860714 1305 201 26.8 500 25.2 - - -20.1 8.7 860714 860805 1386 217 27.6 680 13.9 12.7 0.861 -19.3 9.0 860805 860908 2468 268 34.4 1378 53.3 19.0 1.49 -18.8 8.2 860908 861014 746 103 12.6 322 28.1 11.2 0.756 -19.3 8.5 861014 861112 969 131 16.8 404 21.5 10.1 0.735 -19.4 8.7 861112 861208 1088 113 12.9 177 24.9 47.0 3.20 -21.2 8.9 861208 870120 947 69.7 8.24 151 15.7 56.2 3.26 -22.4 9.8 Table G.7: Sediment-rap fluxes measured at station SN-0.8: 50 m. Appendix G Sediment-trap data of Saanich Inlet 205 start end mass O C N BSi CaCOa A l T i <513C <515N ddmmyy mg m - 2 day - 1 870120 870216 1042 97.6 11.6 160 18.9 62.9 3.65 -21.4 10.7 870216 870309 1325 117 14.6 233 33.8 70.1 4.26 -22.4 9.3 870309 870406 1274 124 16.0 431 ' 7.99 49.9 2.91 -22.7 8.6 870406 870504 3807 310 42.1 2312 18.4 43.4 2.49 -19.5 7.9 870504 870615 2537 249 32.7 1442 27.7 31.0 1.60 -19.2 8.0 870615 870713 2866 291 36.9 1319 42.8 29.0 2.33 -19.1 -870713 870826 2630 292 37.6 1253 58.7 27.8 1.97 -19.4 8.0 870826 870921 622 108 14.3 176 20.9 9.74 0.720 -20.9 8.6 870921 871019 1175 142 17.8 510 6.95 6.36 0.514 -19.1 7.8 871019 871116 353 58.1 6.33 70.4 22.0 10.3 0.736 - -871116 871214 1110 88.6 10.2 355 29.4 35.6 2.42 -23.0 7.9 871214 880105 1036 105 12.2 125 23.3 59.5 3.70 -22.2 6.3 880105 880201 688 65.9 7.21 113 29.0 33.0 1.91 -22.4 6.5 880201 880229 645 67.5 7.76 131 17.6 29.9 1.86 -23.1 7.3 880229 880328 694 69.6 8.71 182 39.9 29.8 1.84 -20.3 7.1 880328 880425 2610 191 22.6 1174 11.9 45.2 2.63 -19.8 5.8 880425 880524 4266 368 46.1 2404 29.1 26.4 1.59 -19.0 7.1 880524 880627 3262 232 28.2 1702 25.1 45.0 2.86 -20.8 8.2 880627 880808 836 173 23.1 206 36.6 12.1 0.786 -21.9 6.4 881017 881115 988 117 11.1 350 30.0 17.2 1.23 -21.8 6.4 881115 881212 805 93.3 10.4 158 25.6 31.9 2.17 -20.4 8.4 881212 890104 513 56.3 6.23 75.3 19.2 25.1 1.62 -22.3 8.9 890104 890130 925 96.3 10.6 128 27.0 49.2 2.93 -21.7 9.0 890130 890306 1305 117 14.4 187 87.1 60.6 3.87 -21.9 9.0 890306 890403 481 56.0 7.82 104 12.2 23.2 1.35 -22.9 9.3 890403 890501 4120 296 40.0 2104 24.7 58.4 3.33 -20.9 6.7 890501 890605 3077 300 36.7 1246 22.2 14.6 1.10 -19.1 7.5 890605 890704 2963 327 40.9 1160 33.2 25.4 1.61 -19.1 7.4 890704 890828 1790 243 29.3 507 33.0 12.1 0.906 -19.7 7.9 890828 891010 863 142 16.8 146 7.24 5.95 0.488 -20.4 7.1 891010 891215 961 105 11.4 414 17.1 19.2 1.31 -19.9 7.1 Table G.8: Sediment-rap fluxes measured at station SN-0.8: 50 m (continued). Appendix G Sediment-trap data of Saanich Inlet 206 start end mass O C N BSi C a C 0 3 A l T i <513C