UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Mixing-primary production coupling in Holberg Inlet, a tidally energetic fjord Kessler, Thomas Alexander 1986

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

Item Metadata

Download

Media
831-UBC_1986_A1 K47_7.pdf [ 13.92MB ]
Metadata
JSON: 831-1.0053272.json
JSON-LD: 831-1.0053272-ld.json
RDF/XML (Pretty): 831-1.0053272-rdf.xml
RDF/JSON: 831-1.0053272-rdf.json
Turtle: 831-1.0053272-turtle.txt
N-Triples: 831-1.0053272-rdf-ntriples.txt
Original Record: 831-1.0053272-source.json
Full Text
831-1.0053272-fulltext.txt
Citation
831-1.0053272.ris

Full Text

MIXING-PRIMARY  PRODUCTION  COUPLING  IN H O L B E R G INLET, A TIDALLY E N E R G E T I C F J O R D by THOMAS ALEXANDER KESSLER Bachelor of Science, University of Alberta, 1974 A THESIS SUBMITTED IN PARTIAL F U L F I L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF DOCTOR of PHILOSOPHY in T H E FACULTY OF G R A D U A T E STUDIES Oceanography  We accept this thesis as conforming to the required standard  T H E UNIVERSITY OF BRITISH COLUMBIA J A N U A R Y 1986 © THOMAS A L E X A N D E R KESSLER, 1986  fa  In  presenting  degree freely  at  this  the  available  copying  of  department publication  of  in  partial  fulfilment  of  the  University  of  British  Columbia,  I  agree  for  this or  thesis  reference  thesis by  this  for  his thesis  and  scholarly  or for  her  Department  DE-6(3/81)  Columbia  I further  purposes  gain  be  It  is  shall  that  agree  may  representatives.  financial  permission.  T h e U n i v e r s i t y o f British 1956 M a i n Mall Vancouver, Canada V 6 T 1Y3  study.  requirements  not  that  the  Library  permission  granted  by  understood be  for  allowed  an  advanced  shall for  the that  without  head  make  it  extensive of  my  copying  or  my  written  Abstract Wind and tidal mixing effects on primary production are investigated in the HolbergRupert basin, a fjord characterized by an inflow tidal jet and a seasonal growing cycle characteristic of well mixed coastal water masses. Physical evidence is presented which suggests that the jet is more positively buoyant and hence mixes more with the surface layer of the basin than previously thought. However on the basis of mass balance calculations, the effect on primary production is suggested to be secondary to wind/tide generated diffusion effects and to a planktonic grazer component. A » 7 yr data set of environmental and primary production variables is evaluated statistically to infer the characteristic spatial and temporal scales of production in the basin, along with possible factors contributing to this distribution. The analysis indicates that persistent (i.e. orthogonal) horizontal features are generally lacking on the one month sampling time scale, except for structure shown to be evident in surface chlorophyll specific biomass distribution on the largest length scales of the basin (i.e. ca. 20 km). Orthogonal annual and inter-annual variance structure is shown to be lacking. Consistent with the mass balances, a partial correlation analysis of the same data set indicates that aside from irradiance, the dominant covariate of primary production is column stability, with the covariance of carbon uptake anti-correlated with stability. However, the covariance of biomass with stability is shown to be seasonally dependent (i.e. anti-correlated in the spring/fall and correlated in mid-summer). Short time scale (i.e. high resolution) time series of the vertical distribution of primary production variables are presented to demonstrate the evolution of vertical structure in these variables. The regular chlorophyll maximum feature in these time series is suggested to be predominantly the result of a behavioral response to high irradiance and/or low ambient nutrients, though depth dependent growth is shown to be important on occasion. A numerical model is constructed incorporating the structure and function indicated to be important from the statistical analysis of the prior data set and from the patterns and ii  inter-relationships evident in the high resolution time series. Numerical experiments are carried out evaluating model sensitivity to wind and column stability forcing. The results are interpreted to suggest that surface mixing control of primary production probably does not occur through vertical displacement effects on the near field light regime, but rather is related to phytoplankton nutrition. Non-monotonic responses by the primary production output variables in the model to monotonic changes in wind forcing in a manner that is consist with the seasonal biomassstability covariance pattern is interpreted as indicating that non-linear coupling in the governing equations could be responsible for this covariance structure.  iii  Acknowledgements I would like to acknowledge the encouragement, support, and advice of my thesis supervisor, Dr T.R. Parsons; and also the other members of my thesis committee for helpful discussions. I must thank Mr. R. Hillis and Mr. I. Home of Island Copper Mine, Rupert Inlet for kindly providing logistical field support and access to the I C M environmental database. Thanks also go to M r . D. Stucchi of IOS, Pat Bay for making available the Holberg Inlet current meter record, Dr. S. Pond for the use his window blind drogues and wind anemometer, and Dr. W. Emery for use of his Houston plotter. Finally, I wish to acknowledge that I was personally supported by a B.C. Science Council GREAT award during the first three years of my studies.  iv  Table of Contents Abstract Acknowledgements List of Tables List of Figures 1. General Introduction 2. General Features of Holberg-Rupert Basin Oceanography 1. Physical Oceanography with Respect to Tidal Mixing 1. Background 2. Re-evaluation of the Tidal Jet Buoyancy Regime 1. Methods 2. Results and discussion 3. Summary 2. Background Biological Oceanography 1. Introduction 2. EPS Data Review 1. Relevant program and methods 2. Data Summary 3. ICM Monitoring Data Set Review 1. Relevant program and methods 2. Data set summary 4. Analysis of ICM Data Set Variance 1. Statistical approach 2. ANOVA of phytoplankton biomass and production indices 3. Interpretation of the variance structure 4. Inferences About the variance structure 5. Partial Correlation Analysis of the ICM Data Set 1. Introduction 2. Statistical method 3. Interpretation of the covariance structure 4. Summary 6. General Discussion 3. Short Time Scale Primary Production Processes 1. Introduction 2. Specific Survey Objectives 3. Methodology 1. Sampling Periods 2. Drift Stations 3. Chemical/Biological Variables Measurement Procedures 4. Physical Variables Measurement Procedure 4. Results and Discussion I.May/82 Cruise 1. Wind effect on drift buoy displacement and current shear 2. M l primary production time series 3. M2 primary production time series 4. M3 primary production time series 5. P-I parameter estimates 6. Patterns in species composition V  ii iv vii viii 1 3 3 3 7 7 9 21 22 22 22 24 24 24 25 27 27 38 41 44 46 50 50 51 51 57 58 63 63 63 64 64 64 .65 68 68 69 69 78 83 83 89 94  7. Inferred grazing pressure 94 8. Inferences concerning M2 S C M formation 95 9. Summary 98 2. Aug/82 Cruise 98 1. Wind effect on drift buoy displacement and current shear 98 2. Primary production time series 104 3. Patterns in species composition 110 4. Inferences concerning S C M formation 115 5. Chla and nitrogen budgets 117 6. Summary 124 3. July/83 Cruise 125 1. Wind effect on drogue travel and current shear 125 2. Horizontal patterns in phytoplankton biomass 137 3. Primary production time series 143 4. Patterns in species composition 144 5. Chla and nitrogen budgets 152 6. Inferences about S C M formation 159 7. Summary 160 4. General Discussion 161 4. Simulation Modelling of Basin Primary Production 165 1. Introduction 165 2. The model 166 3. Forcing Functions 170 1. Time series 170 2. Vertical Diffusion (Heat Submodel) 171 4. Numerical Solution 175 5. Results and Discussion 176 1. Heat Budget-diffusion Submodel 176 2. Displacement Time Scale Attenuation of Photosynthetic Processes . . . . 178 3. July/88 standard run 185 4. Importance of depth dependent growth and sedimentation in SCM formation 190 5. Model Robustness 196 6. Wind Mixing Effects on Primary Production/Stability Covariance . . . . 200 7. Summary 202 5. Conclnsions 203 6. References 206 7. Appendices 217  vi  List of Tables II.l H.2 n.3 II.4  Analysis of variance for Chla, and S w of Chla ANOVA of PI and QI w of P\ and Q  41 42 43 43  n.5  w for <i|> and </£>/<i|>  11.6 11.7 11.8 11.9 11.10 11.11 11.12 111.1 111.2 111.3 111.4 111.5 111.6 111.7 111.8 111.9 IIL10 111.11 111.12 111.13 111.14 111.15 III. 16 111.17 111.18 111.19 111.20 111.21 m.22 111.23 m.24 111.25 111.26 m.27 UI.28 III. 29 rV.l IV. 2 IV.3 rV.4 rV.5  Group means of depth specific irradiance Station and year Chla, marginal deviations from grand mean Kendall rj, matrix of production variables Partial correlations with year fixed, of S and PI Partial correlations fixed to station, month, year Partial correlations as in Table 11.10fixedto < / £ > Partial correlations fixed to stn , mth, year of Chla May/82 drift station dates and positions May/82 Vertical current profile Production variables for May/82 cruise Least-squares estimates of P™, or, and /? Adaptation index to low light (/$) and high light (h) May/82 Observed (O) and predicted (P) carbon uptake drift sampling station dates Aug/82 vertical current profile during the A l time series Aug/82 ( A l ) production variables Total daily PAR during the Aug/82 cruise Aug/82 ( A l ) time series volume contributions . . . Aug/82 ( A l ) Chla and nitrate mass balance Aug/82 ( A l ) vertical herbivorous copepod distribution Aug/82 ( A l ) estimated grazing loss/nitrogen excretion Aug/82 ( A l ) estimated 24 h grazing pressure A l vertical gradients in Chla and NOg" Aug/92 ( A l ) 0-6 m integrated Chla and N O J defects Aug/82 Chla and NOJ" mass balance defects accounted for July/83 drift stations over the drogue time series July/83 relative current shear July/83 total daily P July/83 fixed-station transect sampling times July/83 production variables July/83 Chlo and nitrogen mass balance July/83 24 h grazing pressure July/83 volume specific sedimentation rates J1-J4 vertical gradients in Chla and total nitrate July/83 Chla and nitrogen defects July/83 Chlo and nitrogen mass balance defect accounted for Heat model fixed parameter values The temperature residual for each drift stn (2-14) Heat equivalent depth averaged residuals July/83 standard run parameter values Wind shear forcing of the Aug/82 time series simulation  B  2  z  2  g  48  2  B  z  vii  49 50 52 54 54 55 56 69 76 82 90 90 92 101 . 103 109 110 113 117 119 119 121 122 122 123 127 130 132 132 143 153 . 154 155 156 157 158 176 177 177 189 201  List of Figures 2.1 2.2 2.3 2.4 2.5a 2.5b 2.6 2.7 2.8 2.9 2.10 2.11 2.12 2.13 2.14 3.15 3.16 3.17 3.18 3.19 3.20 3.21 3.22 3.23 3.24 3.25 3.26 3.28 3.29 3.30 3.31 3.32 3.33 3.34 3.35 3.36 3.37 3.38 3.39 3.40 3.41 3.42 3.43 3.44 3.45 3.46 3.47  Location map Historic current meter moorings High positive buoyancy flow regime record 1975 basin precipitation record Summer time increasing salinity record Summer time stationary salinity record Kains Is. surface salinity 1974-1975 Kains Is. surface salinity anomaly 1974-1975 ICM and EPS station locations Mean monthly Chla, at ICM stations A - G Mean monthly S at ICM stations 1-6 Mean monthly S at ICM stations 1-6 Mean monthly PI at ICM stations 1-6 Mean monthly QI at ICM stations 1-6 ICM Pool of Chlais vs. 0-5 m Chla integrations May/82 drogue track May/82 surface wind record M l salinity and temperature profiles M2 salinity and temperature profiles M3 salinity and temperature profiles M l Chla and in vivofluorescenceprofiles M l Chla and carbon uptake profiles M l P and NO3 profiles M2 Chla and in vivofluorescenceprofiles M2 Chla and carbon uptake profiles M2 P and NO~ profiles M3 Chla and in vivo fluorescence profiles M3 P and NO3 profiles Aug/82 drogue track Aug/92 surface wind record Aug/82 salinity and temperature profiles Aug/82 Chla and in vivofluorescenceprofiles Aug/82 Chla and carbon uptake profiles Aug/82 P and N O J profiles Aug/82 size-frequence spectrum J1-J3 (July/83) 3 m depth drogue track July/83 wind record J4 (July/83) 3 m drogue track July/83 drift stn salinity and temperature isopleths July/83 drift stn <r isopleths July/83 fixed stn transect line July/83 transect 3 salinity and temperature isopleths July/83 transect 13 salinity and temperature isopleths July/83 transect 4 and transect 9-14 fluorescence isopleths J l (July/83) Chla and carbon uptake profiles J l (July/83) P and total ambient nitrogen profiles J2 (July/83) Chla and carbon uptake profiles B  p  B  B  B  B  t  B  viii  4 8 10 11 12 13 17 19 23 28 30 32 .35 36 47 70 72 74 75 77 79 80 81 84 85 86 87 88 99 100 102 105 107 108 112 126 128 129 131 133 134 135 136 138 145 147 148  3.48 3.49 3.50 4.51 4.52 4.53 4.54 4.55 4.56 4.57 4.58 4.59 4.60  J3 (July/83) Chla and carbon uptake profiles J4 (July/83) Chla and carbon uptake profiles J2-J4 (July/83) P and total ambient nitrogen profiles July/83 drift stn temperature time series and model simulation . . . . Simulation averaged Kg profile best estimate Vertical Lagrangian displacement time scale for different wind conditions Contrasting at forcing profiles Vertical Lagrangian displacement time scale for different at July/83 standard run primary production simulation Simulated July/83 carbon uptake and POC profiles (growth rate exp.) . Simulated July/83 POC time series (sinking rate exp.) Simulated July/83 phytoplankton time series (sinking rate exp.) . . . . Robust simulation of Aug/82 phytoplankton time-series B  ix  149 150 151 179 180 182 183 184 186 191 192 193 197  1. General Introduction Unlike the more mountainous British Columbia mainland, Vancouver Is. (Figure 2.1) does not accummulate a large snow pack during the winter. As a consequence, the phytoplankton growing season in fjords located there is not under the influence of a significant snow melt freshet. The Holberg-Rupert basin, Vancouver Is. (Figure 2.1), in addition to sharing this common feature is also distinguished by an energetic inflow tidal jet to which is attributed the unusual basin properties of a relatively uniform vertical salinity, temperature, and oxygen distribution. A number of studies have investigated the dynamics of this tidal jet and how it influences basin hydrography. In particular Stucchi and Farmer (1976) and Stucchi (1985) attempted to relate the relative buoyancy of the inflow to its subsequent interaction with and modification of ambient basin water, while Johnson (1974) and Hay (1981) related the tidal flow of the basin to its sedimentology. Phytoplankton are by their nature under the direct influence of water movement, and at the same time dependent upon the associated exchange for the renewal of consumed nutrient supplies. Hence hydrodynamics coupling is considered central to understanding primary production and its controlling processes in the ocean (see Legendre and Demers 1984 for a review). Nevertheless, there has to date been no effort to understand the manner in which the mixing regime in Holberg-Rupert basin (i.e. both tidally and wind generated) affects primary production there. Such improvements in this understanding and the contribution it makes to characterizing primary production in general is of added significance in the Holberg-Rupert basin owing to the public concern surrounding the effects of discharging mill tailings from a local open pit mining operation into the basin receiving waters. Accordingly, this thesis is viewed as a modest first step toward a comprehensive theory of mixing-primary production coupling in the basin. The particular approach followed can be divided into three distinct but at the same time closely related parts: 1) Chapter 2 consists of the qualitative re-analysis of a near surface current record and the 1  statistical analysis of a long term primary production data set. The purpose of the current record analysis is to find evidence for or against the presence of a tidal jet generated near surface and upwards directed entrainment flux, which by its distinctive exchange properties compared to turbulent diffusion would significantly qualify the nature of mixing-primary production coupling. The statistical analysis examines a » 7 yr low resolution data set for variance and covariance patterns from which initial inferences can be formulated concerning the nature of primary production in the basin. 2) Chapter 3 consists of the presentation and interpretation of a series of field studies addressing smaller scale spatial and temporal aspects not ammenable to investigation in the prior long term record, but implicated as important components. In particular, an attempt was made to obtain continuous records of Chla specific biomass, ambient nutrients, and growth rates with the objective of estimating mass balances so that the dominant flux terms could be identified. Also, the relative vertical distribution of these variables was evaluated in an effort to gain insight into the dominant processes controlling their distributions. 3) Chapter 4 consists of a simulation modelling excercise that strives to re-construct the essential features of the mixing-production coupled system by encorporating those aspects implicated as important components in the first two parts. The aim of this effort is not to develop a comprehensive picture of the underlying dynamical relationships, but rather to make the model as precise (i.e. realistic) as possible, and then use it in an input/output fashion to broaden the perspective otherwise limited by the finite nature of the data set.  2  2. General Features of Holberg-Rupert Basin Oceanography 2.1 Physical Oceanography with Respect to Tidal Mixing 2.1.1 Background Holberg Inlet and adjacent Rupert Inlet form a single basin of « 66 k m and mean 2  depth of » 100 m {Figure 2.1), connected to Quatsino Sound and the Pacific Ocean via Quatsino Narrows (QN), a « 3 km long channel with an 18 m sill near its northern end. Based on the calculated basin surface area and predicted tidal amplitudes, absolute transport rate through QN, averaged over a tidal cycle, ranges from » 3 10 m /s for neap 3  3  tides to » 1.2 • 10 m /s for spring tides. By comparison, the input of freshwater into the 4  3  basin from the Marble River {Figure 2.1), the dominant run-off source, ranges from » 25 m /s in mid-summer to « 150 m /s in mid-winter (Drinkwater 1973). However, it is not 3  3  clear what contribution to total run-off comes from other freshwater sources. Based on an estimated combined watershed for these secondary sources of « 600 k m and a local mean 2  precipitation of 10 mm/d (data from Drinkwater 1973), the cummulative discharge rate from these sources would not exceed » 7 m /s, or 10% of the mean Marble R. discharge. 3  Alternatively however, by directly comparing watersheds (i.e. with Marble R. watershed = 512 km ), Drinkwater (1973) estimated that these combined secondary sources should 2  represent « 50% of the total basin freshwater input. The most prominent of the secondary sources are Stevens Creek at the head of Coal Harbour, and Waukwaas Creek at the head of Rupert Inlet {Figure 2.1). Neither of these inputs have been measured for discharge rate (DFO 1982), though it would be reasonable to assume that given their relatively small watersheds, they discharge substantially less than the Marble River. However, due to the location of this discharge relative to the entrance of the basin, the input may not reflect the contribution made to a basin estuarine circulation, since the role of the Marble R. in driving such a flow is significantly altered as it mixes with the flood tidal flow through QN (Stucchi and Farmer 1976). In total, freshwater discharge over a tidal cycle ranges from 0.75% to 1.3% of a spring tidal prism and from 3% to 5.3% of a neap tidal prism, depending upon the manner in which this secondary freshwater discharged is calculated. 3  Figure 2.1 Location Map. 4  Accordingly, the salinity and temperature profiles in the basin are Type 2 (i.e. no well defined mixed layer), typical of low run-off fjords (Pickard 1963), with the thermocline / halocline extending from the surface to » 5 m.  The most obvious mixing in the basin is associated with the entrance jet on flood tides, which can visually be seen to produce regions of intense stirring in regions proximal to QN (Stucchi 1985). Stucchi and Farmer (1976) suggested that seasonal dynamics of the jet are controlled by the local run-off and offshore upwelling conditions as they affect its buoyancy relative to ambient basis water. During the winter when the near field run-off rate is highest and the offshore wind field inhibits Ekman transport, the density difference between the incoming jet and basin water is at its lowest, approaching and at times becoming positively buoyant. As a consequence, the jet is restricted to shallower regions of the water column as it mixes into the basin during this period, though it will penetrate to the bottom if the flow is fast enough and the density difference small enough (Stucchi 1985). Over the course of the winter, entrainment of underlying water into this buoyant flow results in net export of salt over a tidal cycle and an overall reduction in mean basin water column salt content. The combined effects of this basin salt removal as well as the subsequent reduction in local run-off and enhancement of offshore upwelling, act to increase the relative density of the tidal jet to the point of making it negatively buoyant during the following summer. Estimating a typical summer time A a of w 0.25 between Quatsino Sound and t  the basin, Stucchi and Farmer (1976) calculated that the gravity driven circulation during this negatively buoyant flow would sink the jet to the bottom of the basin within w 2 km of the entrance. Thus according to theory these seasonal buoyancy effects account for the large corresponding variation in bottom water properties characteristic of the basin (Stucchi 1985). On the other hand, Stucchi (1985) did attribute the appearance of a positively buoyant flow regime during Aug 29-Sept 5/75 to a short period of high local run-off stemming from a 15 mm/d precipitation rate just prior to the start of the record. The analysis also indicated that offshore wind conditions were favourable for upwelling at the time, which in combination would tend to suggest that shorter period Cuctations 5  associated with local variations in weather are superimposed on the expected seasonal pattern of jet relative buoyancy. However, it is impossible to draw any firm conclusions on the relative summer time dominance of the two flows from this one example. This is particularly the case since both the high precipitation rate and its late summer occurrance (when upwelling conditions are on average less favourable) makes it equivocal. The theory of basin jet dynamics was developed to explain the unique deep water properties of the basin (Stucchi and Farmer 1976). However, its consequences also impinge on mixing-primary production coupling due to the vertical entrainment flux associated with the jet-basin interface and the dependence of its position in the water column. Unlike vertical diffusion, vertical entrainment advects passive constituents preferentially from the less turbulent to more turbulent fluid across the entrainment interface separating the two fluids (Phillips 1975). If this interface is near or within the euphotic zone, its precise depth in relation to the nutracline determines the entrainment flux of nutrients into the near surface, while its position in relation to the light field determines the vertical length scale of the eddies compared to the euphotic depth, thereby influencing the time averaged light flux of the phytoplankton (Denman and Gargett 1983). The relative contributions by near surface entrainment and eddy diffusion in estuarine systems is often estimated to a first approximation by the ratio of the freshwater run-off to tidal prism volume (Dyer 1973). By this estimation, mixing in the Holberg-Rupert basin is primarily diffusive, since run-off represents at most 6% of the tidal prism. However, this considers only the entrainment associated with an estuarine flow. Information on the tidal jet entrainment in the basin has been obtained from near bottom current meter moorings deployed in support of sediment transport studies (Johnson 1974, Hay 1981). These data indicated a recurring phase shift between the barotropic tide and the direction of bottom currents. An internal tide contribution was assumed to be minor and hence the phase shift was attributed to mid-depth tidal jet penetration entraining underlying water, thereby generating a horizontal counter flow in the underlying layer. However for the reason already given, the deep depth of this tidal jet entrainment makes it relatively unimportant 6  (or at least its distinction from diffusive mixing relatively unimportant) in characterizing primary production related mixing. Nevertheless, assuming that the assumption of a weak internal tide is justified, it is useful in demonstrating that a tidal countercurrent (presence indicated by the current-tide phase shift) can be used as an indicator for the presence of a tidal jet generated vertical entrainment flux. This feature will be utilized below in the re-evaluation of an existing long term near surface (i.e. 30 m depth) current record in Holberg Inlet, with the purpose of determining the relative frequency of a tidal jet entrainment interface in the near surface.  2.1.2 Re-evaluation of the Tidal Jet Buoyancy Regime £.1.2.1 Methods The current data record runs from May 14/75 to Dec 8/75 (Stucchi and Farmer 1976) and was obtained with an Anderaa current-meter moored in Holberg Inlet at between 30 and 32 m depth (mooring C, Figure 2.2). The time series consisted of 2 separate deployments: May 14/75-Aug 18/75 and Aug 18/75-Dec 8/75 spanning summer time and winter time flow conditions. The phase relationship between predicted barotropic tidal current and the measured current and salinity was estimated by plotting the dominant along-basin 17-component of the velocity (essentially east-west) against predicted tidal amplitude. In this qualitative analysis it was assumed that given the 90° phase shift between the state of the tide and the barotropic currents, observed currents leading the tide by < 90° were viewed as evidence of an entrainment driven count ercurrent delaying the appearance of the normal tidal current at the current meter depth. Tidal predictions were obtained with a harmonic model of the M / , O  lt  Ki, P i , JV , 5 , M , i f , M * , and MS 2  2  2  2  4  tidal constituents estimated from a 29 day record (CHS, unpublished) at Coal Harbour, Holberg Inlet. No data processing was done to the current and salinity data other than the recursive elimination of outliers more than ± 4 and ± 3 standard deviations from a centered 3-point moving average respectively.  7  8  2.1.2.2 Results and discussion In general the late summer/winter record (Aug /75-Dec./75) at mooring C was of limited quality. However, the early portion (i.e. Aug 18/75-Sept 23/75, day 229-265 in Figure 2.3) appears sufficiently continuous (with respect to the end of the apparently good quality May-Aug record (i.e. Figure 2.5a,b) to warrant interpretation. This mooring C record, which is concurrent with the mooring H record analyzed by Stucchi (1985), begins with a decline in salinity (day 238-247, Aug 17/75-Sept 4/75) from 32.1 to 31.8 ppt (Figure 2.3). The initial part of the corresponding current record appears spurious. However flood tide currents clearly lag the tide by » 90° by day 245. Interpreting this 180° phase shift according to Hay (1981), this indicates the presence of an entrainment driven counterflow throughout the tidal cycle, implying that the entrainment interface was always shallower than the current meter depth (i.e. < 30 m) over the course of a tidal cycle. The rainfall during this period (day 230-241, Figure 2.4) increased from a trace on day 235 to « 25 mm/d on day 241 and as suggested by Stucchi (1985) implicates run-off as the cause of this highly positively buoyant flow regime. The conclusion to be draw here however, is that at least under these extreme run-off conditions the current/tide phase relationship successfully indicated the presence of an entrainment driven countercurrent. This is consistent both with the previous evidence that such a flow regime existed at the time (Stucchi 1985) and with the evidence that an entrainment driven count ercurrent is a general consequence of the tidal jet-basin interaction (Hay 1981).  Applying this test to the summer time flow regime, Fig 2.5a,b presents representative segments of the May-Aug current record. Since salinity trends were used on one occasion by Stucchi and Farmer (1976) to deduce the buoyancy regime of the flow (see below), these representations include both an increasing (Figure 2.5a) and stationary (Figure 2.5b) salinity record segment. The two dominant features of both are a If-component current typically leading the predicted flood tidal current by 20 — 45° and a recurring salinity depression approximately concurrent with high tide. While a positively buoyant tidal flow interpretation of these phase relationships is consistent with Stucchi and Farmer (1976) for 0  3.6  _  UJ  o  (_>  UJ CO  sr •  CO  242  244  JULIAN DAYS  3.6  _  LU  o  CL  CE CO  238  240  242  244  246  248  JULIAN DAYS  Figure 2.S High positive buoyancy flow regime record segment during the Aug-Dec time series (i.e. spring/fall).  10  PREC  0.0  (MM/DRY)  57.5  ^  0.0  PREC  (MM/DRY)  57.5  PREC  0.0  (MM/DfiY)  57.5  JULIAN DAYS  Figure 2.5a Representative summer time increasing salinity record segment for the MayAug time series (i.e. summer time).  12  JULIAN DAYS  Figure 2.5b Representative summer time stationary salinity record segment for the May-Aug time series (i.e. summer time).  13  the stationary time series, it is inconsistent for the increasing salinity record. A possible explanation for this discrepancy is indicated by briefly reviewing the salinity trend analysis of Stucchi and Farmer (1976). In the analysis of a short (Apr 20/74-May 7/74) current, salinity, and temperature record from a mooring in place off Hankin Pt., Holberg Inlet (mooring H , Figure 2.2), an increasing trend in near surface salinity was attributed to the negatively buoyant tidal exchange, inferred independently from the corresponding vertical current distribution. Subsequently, a longer record from Mar 14/75-May 18/75 (mooring C, Figure 2.2) was also interpreted in terms of a negatively buoyant flow regime. It was the analysis of this longer record which appears to form the basis for the generalization that negatively buoyant flow dominates in the summer. However, the analysis of this second record (Stucchi and Farmer 1976) did not explicity mention an evaluation of measured tidal currents, suggesting that the flow regime interpretation was based solely on the salinity trend and an extrapolation from the covariance between salinity and relative buoyancy in the much shorter earlier record. It appears, however, that this extrapolation can be erroneous, since the representative salinity record in Figure 2.5a clearly shows that for periods of increasing salinity over the May 14/75-Aug 18/75 record analyzed here, the salt increase during each tidal inflow usually occurred within 1 h of high tide following a pronounced depression in salinity, suggesting that during this period of increasing salinity, the flow was not negatively buoyant for most of the flood tide. The dynamics of the salinity depression near high tide can be explained in terms of the control of a positively buoyant flow, characterized in a related context by Stucchi (1985) in terms of an initial densiometric Froudt number F = v{ghdpl )- l* l  d  Pv  (2.1)  where Fd scales the penetration depth of the near field buoyant flow, with v the jet speed, h the depth of the inflow, p the basin water density, and dp the density difference between w  jet and basin. Thus during the early stage of a flood when the jet is most positively 14  buoyant, Fj, is too low for the entrainment interface to penetrate to the current meter depth. As a consequence, the entraining of fluid across it, up into the buoyant flow, drives the observed counter-current (i.e. phase shift). Later in the flood as dp decreases and v increases and hence F& increases, the entrainment interface deepens. The salinity depression marks the appearance of the interface at the current meter depth followed soon after by the appearance of saltier water. It is during this final period that the overall increase in salinity occurs (Figure 2.5a). Rearranging equation (2.1), the density difference dp required for the entrainment interface h = 30 m can be estimated on the condition that the near field flow assumptions on which (2.1) are based also apply to the basin in general. Using a critical value for Fd of 1.5 estimated by Stucchi (1985, after Stolzenbach et al. 1973) and a tidal speed of 0.10 m/s (Figure 2.5a), dp = 0.03 k g / m . Assuming a temperature 3  of 10° C, this equates to a salinity difference of 0.02 ppt. This is somewhat less than the 0.05-0.1 ppt depression coincident with high slack tide in the representative record segment (Figure 2.5a), but of the correct order given the approximate nature of the calculation.  If the small current/tide phase shifts identified in the representative summer tiuie records were not spurious (e.g. prediction error in the harmonic tidal model), the consequent inconsistency must be reconciled between the positively buoyant summer time dominance inferred here from these phase shifts, and the negatively buoyant flow dominance argued using the current depth distribution of the short Hankin Pt record in 1974 (Stucchi and Farmer 1976). In a more general sense there is also the problem of reconciling a summer time positively buoyant tidal flow preponderance with observed deep water properties that are consistent with vigorous mixing of the entire water column (Pickard 1963). Based on the Kains Is. (location, Figure 2.1) daily surface salinity record, 1974 was characterized by a steeper increase in salinity over the course of the summer compared to 1975 (Figure 2.6). It would seem possible that if this salinity time derivative were large enough (compared in the 20 day flushing time of the basin, Stucchi 1985), an elevated density difference between QS and basin water could have been maintained throughout the summer in 1974 thereby diffentiating it from 1975. However, its importance on the 15  whole is questionable since the total 1 ppt difference in seasonal salinity change spread unifonnily over the day 120-240 summer time interval represents a between year difference in the time derivative of only 0.008 ppt/d. This between year contrast in the derivative may also be an exaggeration, since a comparison of the salinity anomaly time series for 1974 and 1975 (calculated on the basis of the 1935-1969 daily mean, Figure 2.7), indicate that the between year difference was due to a combined abnormally low and abnormally high early summer (i.e. w day 120) salinities in 1974 and 1975 respectively. Finally, the presence of unusual conditions in 1974 and/or 1975 does not reconcile a positively buoyant flow regime with the basin bottom water properties. In addressing the latter inconsistency, it has been calculated that a positively buoyant flow with Fd > 7 can penetrate to the bottom (Stucchi 1985). However, this is based on an assumption of steady state flow conditions and hence it is not clear under what conditions bottom penetration would actually occur. Whether the inferred limited occurrence of such a flow combined with the appearance of negatively buoyant flow is sufficient to account for both the vertically homogenous and seasonally variable bottom water properties (Stucchi 1985) is beyond the scope of this analysis. However, there remains another possible contributing process that in addition to providing deep mixing in the absence of negatively buoyant flow, could also account for the high bottom currents observed off QN (i.e. 3 m/s; Drinkwater and Osborn 1975). The mechanism is an internal hydraulic jump of the type observed by Farmer and Smith (1979) in Knight Inlet, B.C. It is a region of intense mixing associated only with stratified flows over bottom topography (Farmer and Freeland 1983). The degree of stratification of a flow over a fjord sill is suggested to be scaled by the relative size of sill entrance volume compared to the tidal prism (Farmer and Freeland 1983). If the sill volume is comparable to the tidal volume, the flow is relatively mixed to the point of inhibiting the necessary lee wave generation. This results since the speed of the flow under such conditions is always greater than the phase speed of waves generated by the bottom topography (Farmer and Freeland 1983). On the other hand, if the sill volume is sufficiently small (i.e. tidal 16  ^  o s 1111  o  g  OK ^  ars 11111  i ii ii  ass  m i ii ii  SAba Nbnnr  ozs  111  n ii ii n  ois  oos  ii ii I ii  I  11  OK  ii ii i  OK  III  11  ii  11  oti  il  i ii i h  OK 11  n  OSZ  11111 I I I I I I I I 11  iiiiiinil  11111111 1111111111 H I M [ 1111 ? 11 M | M 111 ? 1111 M 1 1 1 1 1 1 1 J 1 1 1 1 1 1 M 111111111 1 1 1 1 1 1 1 1 1 1 I I 1 1 1 1 1 1 1 1 1 1 1 I I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  3 "  oss 111 il  S  O  S  OK  S  O  OK  K  O  S  Oil  S  O  Z  OK  S  O  t  S  O  O  S  sxba N b i i n r  061  111111111111111111 n 1111111 H 1111111111111111  I  CHI  a  Oil  K  O  091  K  O  L  OSI  i  O  K  »l  O  A  K  °  OKI  111 1111111111111111 n i 1111111111111 H 1111 n 11 H i 11111111111111  3  P  l l | l l l l l l l l l | l l l l l l l l l | l l l l l l l l l | I H I I I I I I | l l l l l l l l l | l l 1l1 1l1 1l 1 l1 1lT lj 1l 1[1 1lT l1 1l1 1l 1l1 1l 1lT 1l 1 l1 1 1 J 1 1 1 1 1 1 1 T '  "  K  3  KI  OK  o n  111111 I I 11111111  OS  oai  Oil  OK  OBI  OB  n  at  Iiiiiiiiiiliiimiiiliiiniiiil  Oil  SAbO N b i i n r a  Oil  os  091  OSI  » l  OSI  »  os  K  oi  iiiliiiiiiiiiliiiimiiliiiiiiiiilimiiiiilii  I  °  mi 3  ft  !*  1111 n M 111111M 1 1 1 1 1 1 1 1 1 M 1 1 M ; 1111111 M 1111 H 1111111 > 11 Tl 1T1111T T11111»T1111T111111111T T 1 1 1 1 1 M 1 1 I T J 1 I I 1 1 Tl 1 1 1 1 1 1 1 M 1 1 1  ° si  on  ooi  •  at  at  oi  OB  »  at  at  oi  1  °  l 3  10  e  gj  '  " •  1  10  "  m  JO  90  x  to  to  100  B  110  is  1111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111111  s  »•  1975  -R  E  1  1 1 1 1 1 1 1 1 1 1 ' 1  1  1  '  1  1  10  1  1  1 ' ' 1  1  1  1  1  1  ' 1 ' '' 1  20  1  1 1  SO  ' 1 ''''''' 1 1  1  1  40  1 1  SO  ' ' 1  1 1  ' M  1 1 1 1 1 1 1 1 1 1 1 1 1 r1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 Mj1 1 1 1 1 1 1 1 1 1 T ' R  60  70  80  80  100  200  210  100  110  120  J U L I R N DAYS  ISO  140  ISO  160  170  180  220  230  2  4  o  0  11111111 i 1 1 1 1 1 1 1 1 n 111111111111H n n i n 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 n 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 n 1 1 1 1 1 1 H 1 1 I I I 1 1 I I n 1 1 1 1 1 It  -R  C  1  1 1 I I I 1 1 1 1 I I I I I I I I I I I I I 1 1 1 1 1 I 11 M 1 1 I I I I I I I I I I I 1 1 I 11 I I 1 1 1 1 I 11 1 1 1 I I I I I I 11 1 1 1 1 1 1 1 11 ISO  140  ISO  160  170  180  JULIRN  290  2B0  ; , I l l l l l l l l  C  2 7 0 2 B 0 2 8 0  111111111111111111111111  180  l l | IIII M i l l| lI 210  200  220  230  DRYS  S00  3  l  0  S  2  0  S  3  0  S  4  D  S  S  0  S  E  5  0  2  6  0  2  7  0  2  8  0  0  o  l i i i i i i m l i i i i i i i i i l i i H i i i i i l i i i i i i i i i l i i i i i i i i i l i i i i i i i i i l i i i i  I U M 1 1 1 1 1 I I H M 1 1 1 1 1 I M I I I | \ l 1 1 U 1 1 1 | M I I I 1 1 1 1 | M 111 1 1 I I | M 1 1 1 1 1 1 1 ] M l I I 1 1 1 1 p i l l I I I I l | l I I I I I I I l ] l l 1 1 1 1 1 1 1 | H I 1 1 1 1 1 1 1 1 I I 2  R  240  2  9  0  300  310  J U L I R N DRYS  Figure 2.6 (continued).  18  S  2  0  3  3  0  3  4  0  3  5  0  S  E  O  1  R  250  260  270  290  I I M I llllHill lllllllllllllllnniHlIll  1 " 250 1  ' I • •' • 11 M I 111II 260 270  290  300  290  310  u l i II  I  I 1111 I I | l l 1111 290  Ii  320 m  l  930 n  l  S« .  IniMMiil  350  360 m i n i  I I I 11 I I I I I II11 11111 M i l 1 1 I I 1 1 I I I 1111 I I I 1 1 1 l [ l 1 1 I I 1 1 1 1 1 1 1 I I 1 1 I I 11 I I I 300 310 320 330 SO 350 360  J U L I R N DAYS  Figure 2.7  i  Kains Is. surface salinity anomaly 1974-1975.  19  130  l «  ISO  160  1111111111 i i 111111111111111111111111111111  170  180  190  200  210  220  230  240  11111111111111111111111111 I < i • 1111111111111111111111 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  | I I I I I U I I | M l l l l l l l | l l l l l l l l l | I M I ) m l | l l l l l l l l l | l l l l l 1 l l l | I I I I I I M I | l l 130  140  ISO  160  170  I BO  190  J U L I A N DRYS  Figure 2.7  (continued).  20  200  210  220  230  240  prism » sill volume) a stratified flow regime can predominate, and more specifically can under appropriate conditions evolve from subcritical to critical flow over the course of the tidal cycle. During this flow transition the otherwise upstream propagating lee waves (ie. internal waves) coalesce into the hydraulic jump downstream of the topography. An upper limit of the QN sill volume is 2 • 10 m . This is smaller by a factor of 10 to 100 than the 6  3  tidal prism (6.5 • 10 to 2.6 • 10 m ), implying that this type of mixing process could be 7  8  3  an important feature of the positively buoyant tidal flow in the basin. A hydraulic jump associated with the tidal flow in Holberg-Rupert basin could also be an important contributor to primary production coupled mixing. This is due to the generated lee wave trains that propagate up current along the pycnocline (Farmer and Freeland 1983). In addition to promoting vertical mixing under wave breaking conditions, these internal waves can also modulate the near field light flux due to the wave type undulation of the water column as they propagate along the pycnocline. Since the direction of the propagation is up-current, this would be most important on ebb tides (Farmer and Freeland 1983). Both associated processes can significantly affect the growth of the phytoplankton (Denman and Gargett 1983).  2.1.2.S Summary The re-analysis of a near surface long term current record from Holberg Inlet has suggested that the importance of positively buoyant flow may have been underestimated in the past. However, in relating the consequent surface mixing regime to mixing-primary production coupling, the evidence also suggested that though the associated entrainment interface remains above 30 m depth for most of the flood tidal cycle, it appears that only during periods of high run-off, does it remain exclusively above this depth. In reconciling a predominantly positively buoyant flow regime interpretation of the shallow water current data with basin bottom water properties consistent with deep mixing, the potential significance of a hydraulic jump off QN was proposed to account for both this deep mixing and the observation of large bottom currents near QN. 21  2.2 Background Biological Oceanography 2.2.1 Introduction The Holberg-Rupert basin and adjacent Neroutsos Inlet (Figure 2.8) are sites of industrial activity. Into the former flows the mill tailings (w 4 • 10 T/d) and rock overburden 4  ( » 1 • 10 T/d) from an open pit copper-molybdenum mining operation in Rupert Inlet 5  (ICM 1979) and into the latter until 1977 (Sullivan 1979a) the mill effluent ( « 400 T / d , Waldichuk 1968) from a sulphite pulp-mill. Due primarily to these discharges and public concern for environmental quality, a number of studies of ranging scope have been undertaken with the mandate to ascertain the effects on the local environment. Inspite of this applied and to varying degrees self-serving emphasis (e.g. Goyette and Nelson 1977), some of these studies provide a useful source of information on the structure of the planktonic ecosytem in the basin and adjacent waters. The studies particular to the basin itself have consisted of a general survey assessing phytoplankton standing stock, species composition, and primary productivity from May to Oct in 1974, 1975, and 1976 (Sullivan 1979b), an on-going program (pertinent measurements begun in 1974) monitoring phytoplankton standing stock and primary productivity on a long term basis (e.g. I C M 1983); and a short-term study in Rupert Inlet from Aug 17/75-Sept 10/75 examining the covariation between phytoplankton standing stock and water column stability (Stephens and Silbert 1976). A review of these data sets is presented below. 2.2.2 EPS Data Review  22  Stroggling It.  Straggling  Is.  Figure 2.8 I C M (top) and EPS (bottom) station locations. 23  2.2.2.1 Relevant program and methods The Environmental Protection Service (EPS) maintained 8 stations in the basin and adjacent Quatsino Sound (Figure 2.8) over the course of a three year study. At these stations, carbon uptake rate, Chla specific biomass, and species enumeration estimates were made every two months in 1974 and 1976, and every month in 1975, during the May to October period (Nov in 1975). The methodology, which is basically the same as used in the ICM monitoring program (see Sec. 2.2.3.1), is outlined by Sullivan (1979b).  2.2.2.2 Data Summary Bearing in mind the sparseness of the data set, species enumeration over the three years indicated a pattern of a diatom dominated assemblage during May-June, a naked flagellate dominated assemblage during July-August, and a dinoflagellate dominated assemblage during Sept-October. The most abundant diatoms were Chaetoceros spp., Coscinodiscus spp., Skeletonema costatum, and Cyclotella sp., while the most abundant naked flagellate groups were chrysophytes and cryptophytes. No single dominant dinoflagellate was in evidence, except for a large bloom of Cochlodinium sp. in Oct 1976. Month averaged surface Chla concentrations ranged from < 1.0 mg/m in early Oct-Nov to « 10 3  mg/m in July. With two exceptions out of a total 166 samples, Chla concentration de3  creased monotonically with depth. Predictably, column primary productivity (i.e. 0-30 m integrations) reflected this seasonal biomass pattern, with a maximum month average of « 2.3 g C / m d . in July decreasing to » 0.5 g C / m d in Oct-Nov. 2  2  2.2.3 ICM Monitoring Data. Set Review  24  2.2.S.1 Relevant program and methods This data set is generated by Island Copper Mine (ICM) in compliance with its British Columbia Government agreement to discharge its mine waste into the basin (i.e. primarily Rupert Inlet). The relevant program consists of 7 sampling stations at which Chla specific phytoplankton biomass is estimated (stn A to G , Figure 2.8), and 6 sampling stations where the same biomass estimates are made in support of radiocarbon uptake experiments (stn 1 to 6, Figure 2.8). They are occupied monthly on either a year-round (stn A-G) or April through Oct (stn 1-6) basis. Chla is estimated from » 1-litre samples collected at routine depths with bottle samplers. The depths are 0 m, 5 m, 30 m, and a bottom sample « 20 m off the substrate at stn A-G and at depths of 1, 3, 5, 8 (or 10), 10 (or 15), and 20 m at stn 1-6. Samples are vacuum filtered onto cellulose acetate/nitrate filters using MgC03 buffer and stored frozen over desiccant (maximum « 2 wks) until analysis. Spectrophotometry analysis of Chla (corrected for phaeopigments by acidification) follows the general procedures and equations of Strickland and Parsons (1972). Carbon uptake rates are estimated by incubating in situ replicate 125 ml samples of seawater spiked with 5 p.C\ N a H C 0 3 . 14  Pyrex incubation jars are used with general sample handling following the procedures of Steemann-Neilsen (1952). Subsequent radioactivity of the cellulose acetate/nitrate filters is determined in Aquasol using a scintillation counter. The incubations usually encompass mid-day and last from 4-6 h. Additional unreplicated uptake measurements are made at depths of 1, 3, 5, and 20 m using opaque incubation jars under otherwise identical conditions. The calculated dark  1 4  C assimilation rate is assumed to represent non-  photosynthetic carbon fixation artifacts of all types (Taguchi and Piatt 1977; Sen Gupta and Jannasch 1973) and is therefore subtracted from the measured uptake for corresponding illuminated samples to estimate photosynthetic carbon assimilation rate (ICM 1979). It is also assumed that the totally  1 4  1 4  C technique measures net carbon uptake (i.e. respiration  C based, Harris 1980) and hence no additional respiration correction is made.  A number of environmental variables considered pertinent to the evaluation of pro25  duction are also estimated routinely at stn 1-6. The measurements are concurrent with the Chla and carbon flux estimates. They are 'total' nitrite/nitrate, orthophosphate, and silicate; as well as ambient temperature, total alkalinity, vertical light attenuation, broadband incident solar irradiance, and a% . Details concerning these variables are outlined in I C M (1979), but the procedures generally follow those outlined in Strickland and Parsons (1972). The use of the term 'total' as opposed to ambient nutrient determinations is deliberate. The I C M monitoring program has continued the practice initiated by Sullivan (1979b) of not filtering nutrient samples prior to analysis (Ian Home of ICM, pers. comm.). Information on ambient levels is germaine to an understanding of the underlying processes controlling primary production in the basin. However, since these 'total' estimates are difficult to interpret, they will not be used here to assess the ambient nutrient regime. In their place, phytoplankton nutrient status will be inferred from statistical correlations between the broadband incident irradiance measured with a Belford pyrheliometer located in Rupert Inlet (Figure 2.8), water column stability estimates based on vertical a dist  tribution, 0-15 m integrations of Chla (S ),  0-20 m integrations of carbon uptake rate  B  (S ), p  and a number of derived productivity indicies: An integral productivity index PI  (mgC/mgChla h) defined as S  normalized to S ,  p  and an integral quantum efficiency  B  index QI (mgC/mgChla •h-/jE-m -s) defined as PI normalized to the surface broadband a  irradiance averaged over the incubation period of the carbon uptake measurements < J | >; as well as their depth specific equivalents P  B  and Q • <?t at the incubation depths was z  used to calculate 0-5 m (Et), 5-25 m (Eb) and 0-25 m (E) finite difference approximations t  of the static stability index according to E = l/p(der /dz) (Pond and Pickard 1978). t  26  2.2.8.2 Data set summary The pattern in month averaged surface Chla concentration (Chla, ) during 1974-1983 at stn A-G is dominated by the cycle of annual period, consisting of an increase beginning in March and peaking in August (Figure 2.9). With the exception of a slight indication at the three Holberg Inlet stations, evidence for a spring bloom in Chla, is lacking. The annual cycle ranges from a minimum < 1 mg/m in winter to a maximum » 7 mg/m in late 3  3  summer and is consistent with the month averaged surface Chla pattern for corresponding periods in the 1974-1976 EPS study. An annual cycle is also evident in S  at stn 1-6  B  (record from 1976-1982), though a more purely monotonic summer long increase is evident (Figure 2.10). Station averaged S  B  ranges from « 25 mg/m in April to » 100 mg/m 2  in August. The annual pattern in monthly means for both Chios and S  2  is superimposed  B  on background variance (i.e. presented as a standard error, with n w 9, in Figure 2.10), which given the one year station-month sampling interval, is referred to as inter-annual variability. During the main May-Sept phytoplankton growing period, this variance is comparable in magnitude to the annual cycle. Predictably, the annual pattern in S  reflects the pattern in S  p  and Chios , with  B  uptake rates ranging from a low of « 50 m g C / m h (0.6 g C / m d ) in April and October 2  2  to a high of « 250 m g C / m h (3.0 g C / m d ) in August (Figure 2.11). These rates are in 2  2  general comparable to estimates in the 1974-1976 EPS data set. Like the Chla specific biomass variance, inter-annual variability for the more productive months is of the same order of magnitude as the amplitude in the annual cycle. Normalizing S  p  to S  B  (i.e. PI)  does not reduce the inter-annual variability noticably. However, PI does show a reduced amplitude in the annual oscillation, with some evidence of a mid-summer depression at stn 4-6 (Figure 2.12). By comparison, normalizing PI to < J | > (QI) increases the amplitude of the annual cycle compared to the inter-annual variability, primarily by amplifying the mid summer depression evident in PI (Figure 2.13).  2.2.4 Analysis of ICM Data Set Variance 27  •JPN*  _i  FEB*  i  HflR. PPR.  i  JUNE  JULY  AUG.  SEPT  OCT. NOV.  DEC.  i  Figure 2.9 Mean monthly (±S.E.) surface Chla (Chla,) at I C M stations A to G .  28  29  _ *  JAN. '  FEB. I  MAR. •  APR. 1  HAY '  JUNE •  JULY •  RUG. •  SEPT I  OCT. I  NOV. I  DEC. _ I . 85  CM X X  ha  8—  JPN.  FEB.  JAN. I  FEB. I  rifiR.riPR.fter  MAR. I  APR. I  MAY I  JUNE  JULY  due  SEPT  OCT.  NOV.  JUNE i  JULY •  AUG. I  SEPT I  OCT. I  NOV. I  DEC!  DEC. _ i , M  (VI X X  s:  h8  5»H CD  JAN.  FEB!  rtefi]  APR!  HAY  JUNE  JULY  AUG.  SEPT  (JCT.  Nov]  JAN.  FEB.  MAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  _i  i  i  i  i  i  »  •  i  i  DEC!  DEC.  i  i  _  . 8  Figure 2.10 Mean monthly (±S.E.) 0-15 m integrated Chla (S ) at I C M stations 1-6. B  30  _  JAN. 1  FEB. I  HRR. 1  APR. 1  HAY I  JUNE •  JULY •  AUG. i  SEPT •  OCT. •  NOV. t  DEC •  _ »  X X  h8  X  o  _  85-,  JAN.  FEB.  HAR.  JAN. 1  FEB. 1  MAR. 1  ripR. HAY APR. 1  HAY l  JUNE  JULY  due.  SEPT  OCT.  NOV.  DEC.  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  I  l  I  l  I  I  i  „ _ S  CM X X  5»H  h8  X  (_)  o  _  85 _  JAN.  FEB.  HAR.  APR.  HAY  JUNE  JULY  due.  SEPT  OCT.  NOV.  DEC.  JAN.  1  FEB.  MAR. 1  APR. 1  HAY  l  JUNE  JULY  AUG.  •  SEPT  l  OCT.  l  NOV.  L___8  JON.  FEE  rfc.  APR!  RAY  JUNE  JULY  due  SEPT  OCT.  NOV!  DEC!  1  l  •  l  DEC.  CM X X  (_)  Figure 2.10 (continued).  31  _  JAN.  — i  FEB.  '  MPR,  "  APR. 1  HRY 1  JUNE 1  JULY  AUG.  1  1  SEPT 1  OCT. 1  NOV. DEC. "  1— 5 r  o •—I X  o  Has  n X X  JAN.  FEB.  HRR.  APR  RAY  JUNE  JULY  due  SEPT  OCT.  NOV. DEC.  JHi  FEB!  HRR.  APR.  RAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV! DEC!"  JAN!  FEB!  HAR.  A"PR. HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV!  3ECT  Figure 2.11 Mean monthly (±S.E.) 0-20 m integrations of carbon uptake rate (S ) I C M stations 1-6. p  82  at  JAN. JJ-i  '  FEB. '  RPR.  MAR. '  '  MAY  JUNE  •  JULY •  AUG. '  SEPT 1  OCT. '  NOV. '  OEC. «  r-3  CO X X  (_)  o  JAN.  FEB.  HAR.  APR. "MAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  JAN.  FEB.  MAR.  APR.  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  -I  I  I  I  MAY l  •  •  I  I  I  l  '  O X  —fM X X  (_)  TAN!  JAN. $  n  1  FEB!  HAR.  dpR.  MAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV!  DEC!  FEB.  MAR,  APR.  HAY  JUNE  JULY 1  AUG.  SEPT  OCT.  NOV.  DEC.  JULY  due  1  1  1  1  L_  1  1  1  1  1  o X  X C M .  en  x x  (_) CO  JON.  Fra!  HAR.  APR.  HAY  JUNE  Figure 2.11 (continued).  33  SEPT  OCT.  NOV.  SC!  .  3  JAN.  FEB.  MAR. APR,  JUNE  JULY  1  —I  AUG.  1  SEPT  OCT.  I  NOV.  '  DEC.  •  •  m o E  JAN.  FEB.  MAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  JAN.  FEB.  MAR.  APR.  MAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  —I  1  1  1  i  i  i  i  •  i  i  i  O  in  E V.  O  E  JAN.  FEB.  HAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC  JAN.  FEB.  MAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  —I  1  1  1  l  l  I  I  i  I  i  i  <? -.--I  in  o E  JAN.  FEB.  HAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  Figure 2.12 Mean monthly (±S.E.) 0-15 m depth averaged productivity index (PI) at I C M stations 1-6.  34  JAN. _ l  FEB. I  HAR. I  APR, 1  HA l _  JUNE -I  JULY I  RUG. I  SEPT I  OCT. I  NOV. DEC. I  L  _  o  E  ^  o E  JON.  FEB.  riRR.  dpR,  HRY  JUNE  JULY  due  SEPT  OCT.  NOV.  JAN.  FEB. 1  m a 1  RPR. 1  MAY  JUNE  JULY  AUG. SEPT  OCT.  NOV. DEC.  JAN.  FEB!  HAR!  APR]  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV!  3EZ  JAN.  FEB.  MAR.  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  JAN.  FEB.  HAR.  dpR.  HAY  JUNE  JULY  due  SEPT  OCT.  NOV.  DEC.  —i  I  I  I  I  i  I  I  DEC.  I  \ o o E  »i  o o> E  Figure 2.12 (continued).  35  JG  _l  JAN.  FEB. 1  HAR. 1  RPR. HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  JAN.  FEB.  HAR.  rtPR.  HAY  JUNE  JULY  RUG.  SEPT  OCT.  NOV. DEC.  JAN.  FEB. 1  HAR. 1  APR. 1  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  JON.  FEB!  HAR.  APR.  HAY  JUNE  JULY  RUG.  SEPT  OCT.  NOV! DEC!  JAN!  FEB!  HAR.  RPR.  HAY  JUNE  JULY  due  SEPT  OCT.  NOV.  i  l  i  i  I  i  l  i  O E C JO  i_  E  o o> E v. u  f:  —I  I  I  I  »  •  •  •  DEC.  »  •  E  v. -«=<M V. — D o  -  -C  O  o> E  \ o o> E °  OECT  Figure 2.13 Mean monthly (±S.E.) 0-15 m depth averaged PI mormalized to surface irradiance </f > (QI) at I C M stations 1-6.  36  R  JAN.  FEB.  -J  1  HAR. 1  APR. 1  NAY 1  JUNE I  JULY I  AUG. I  SEPT •  OCT. •  NOV. I  DEC l  E  v. o  (_) E o en E  o  JAN.  FEB.  HAR.  rfpR.  HAY  JUNE  JULY  RUG.  SEPT  OCT.  NOV.  DEC  JAN.  FEB.  MAR.  APR.  MAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  OEC  —1  '  1  1  1  1  1  1  1  I  I  l  E  N.  -  O  =»'  U  o> E  o E  5  Figure 2.IS  JAN.  FEB.  HAR,  APR.  HAY  JUNE  JULY  AUG.  SEPT  OCT.  NOV.  DEC.  JAN.  FEB.  HAR.  APR.  HAY  JUNE  JULY  due  SEPT  OCT.  NOV.  DEC  (continued).  37  £.2.4-1 Statistical approach Assigning data variance components to the time and length scales of the sampling regime (i.e. order km, month, year in the case of the I C M data set), is useful in objectively characterizing the underlying structure in primary production (Lewis 1978). Two possible approaches to the variance partitioning are analysis of variance (ANOVA) and analysis of covariance (ANCOVA). The two differ primarily in that the latter can compensate for the confounding effect of one or more independent covariate (Sokal and Rohlf 1981, p. 509). For example, an ANCOVA of S  p  and covariate S , and of PI and covariate <Ig > would be B  alternatives to an ANOVA of PI and QI respectively. An important point to note however, is that ANCOVA accomplishes this covariate correction using standard Model I regression procedures. In other words, the independent variable (i.e. covariate) is assumed to be a non-random variate without measurement error (Sokal and Rohlf 1981, p.509). Since none of the covariates in the I C M data set meet this requirement, the ANOVA approach will be used. Accordingly, a number of specific ANOVA models are available, ranging from the completely fixed effect to the completely random effect type. The important distinction between the two is that unlike the fixed-effect model, the expected mean squares for main effects in a random model do not include interactive variance elements (Sokal and Rolhf 1981, p. 372). This is pertinent to the randomness interpretation of sampling effects and results in lower significance levels (i.e. lower variance component estimates in a random effects model) and hence reduced statistical power (Zar 1975, p. 44). From a conceptual standpoint, the model most appropriate for the ICM data set is a mixed type consisting of two fixed effects (spatial position and time of year) and one random effect (year of sample). The argument for the contrasting treatment of month of sample from year of sample is that there is no prior reason for a particular year to represent a treatment or sampling effect different from any other year. By comparison, the seasonality of temperate climate primary production engenders the month of year with a predictable sampling effect. The same reasoning can be applied to the sampling position in space effect. That is, the entrance jet is perceived to be an important mixing source. It follows, therefore,  38  that in a variance analysis aimed at revealing mixing related structure in phytoplankton production, a determinant sampling effect assigned to a station on the basis of, say in this case, distance from QN would be more powerful. Nevertheless, inspite of the advantage of tailoring the variance model with an eye to the specific goal of the analysis, examples in the literature typically use the pure random design (e.g. Lewis 1978, Tett and Grantham 1979). It is worth noting that a useful feature of this latter approach is an approximate significance test that can be performed on the main sampling effects using compounded degrees of freedom (df ) (Zar 1974, p. 343), that does not require an error term free of sampling effects (Sokal and Rohlf 1981, p. 347). Hence, both to simplify comparisons with efforts elsewhere, as well as to take advantage of this main effect significance testing feature, the strictly random approach is adopted here as well. No transformations were performed on the I C M data inspite of the strong dependence of cell variance on its mean for most variables (Appendix A). The justification is that with the exception of data interdependence, ANOVA is relatively robust with respect to distribution requirements, while the transformation necessary to ameliorate violations can bias the added variance component estimate for interactive sampling effects (Lindman 1974, p. 105). Data interdependency in the ICM data set can stem from an autocorrelation between successive sampling times (Lewis 1978). This is not explicity tested for. However, the factors most likely to scale the temporal autocorrelation function (e.g. local phytoplankton growth rate, diffusive and advective fluxes) have characteristic e-folding times much shorter than the « 30 day sampling interval and hence it is assumed that data autocorrelation is not prevalent. The implementation ANOVA algorithm used is BMDP:8V described in B M D P (1979, p. 598). This program requires a balanced data matrix. Since there are no true data replicates in the data set this means empty cells are not permitted. Where there are only one or two missing data points as in Chla, , the iterative approximation technique of Li 39  (1964, p. 228) was used to obtain approximate cell values. However, for variables where more than two missing cells were involved (e.g. S , PI, and QI)> all unbalanced treatment B  classes were removed prior to analysis. To estimate the maximum possible bias in the F-statistic due to untransformed data distribution violations, both the numerator and denominator df were divided by the numerator df to generate the necessary df for a conservative F-test (Box 1953). This alternate significance level is given in parentheses in the summary tables. When both estimates indicate the variance is significantly greater than zero (P<0.05), it is referred to as a robust variance component, as opposed to significant for one indication and non-significant if neither F-ratio indicates a variance component significantly greater than zero. Strictly speaking, variance component estimates should be quoted only when the effect is statistically non-zero, while in lieu of cell replicates, confidence limits could not be estimated for any of these variance components (Sokal and Rohlf 1981, p. 217). However, to facilitate comparisons, the approach of Therriault and Piatt (1978) is followed of including component estimates for those effects where the F-statistic is greater than one, even though it may not be significant. In addition to the 3 main sampling effects, 3 interactions were evaluated in the ANOVA; year-station (y-s), year-month (y-m), and station-month (s-m). These can be interpreted as follows: Y-s is the variability of year averaged estimates between stations and is a measure of inter-annual variability in the inter-station contrast, y-m is the variability of year averaged estimates between months and is a measure of inter-annual variability in the inter-month contrast, and s-m is the variability of station averaged estimates between months and is a measure of inter-station varibility in the inter-month contrast.  40  £.2.4.8 ANOVA of phytoplankton biomass and production indices The ANOVA for Chla, and S  B  are summarized in Tabie I L L For both variables the  two major variance components are the month of sampling main effect and its interaction with year of sample. The proportion of the total variance these components account for are 26% and 19% for the main and interactive effect respectively for Chla, , and likewise 20% and 33% for  S. B  Analysis of variance for Chla, And S . Probability levels in parentheses denote lower bounds (see text for details). Added variance components (VC) and % accounted for variance (w ) are also listed. Synthetic df were used for main effect significance testing (Zar 1974). Variance components not reported when F<1.0. T a b l e II.1  B  2  Source Chla, Year (y) stn (s) Mth. (m) y-s y-m s-m y-s-m S Year (v) stn (s) Mth. (m) y-s y-m s-m y-s-m  Err. Term  SS  ys+ym-ysm ys+ms-ysm ym+sm-ysm ysm ysm ysm  180 156 3065 577 2420 573 4061  8 6 11 48 88 66 528  MS  F  P  23 26 279 12 28 8.7 7.7  0.71 2.01 9.78 1.56 3.58 1.13  >0.50 0.20 0.001 .05) 0.01 .50) 0.001 .20) 0.24  6617 2477 18627 884 4592 909 814  1.21 1.38 3.39 1.09 5.64 1.12  >0.50 >0.50 0.02(.50) 0.3 6 0.001(.05) 0.32  VC  -  -  0.12 3.97 0.36 2.83 0.11 7.69  0.8 26.3 2.4 18.8 0.7 51.0  47 36 387 10 630 16 814  2.4 1.9 20.0 0.5 32.5 0.8 42.0  B  ys+ym-ysm ys+ms-ysm ym+sm-ysm ysm ysm ysm ysm  33083 5 12383 5 111764 6 22101 25 137750 30 27664 30 122077 150  In both cases these two components represent almost all the accounted variance. For Chla, the month main effect is robust and the y-m interaction is significant, and for S  B  it  is the reverse. This contrast is due primarily to the year-round sampling effort for Chla, , since the y-m added variance component is also the robust dominant for Chla, when the data record is truncated to the May-October interval (Tabie TJ.2). Unlike 5 , however, B  the month main effect remains non-significant for this shortened record. Both complete and seasonally truncated Chla, have a significant y-s interaction, though in each case the variance contribution is minor. Calculating the variance components for Chla, (vertically integrated to obtain S ) indicates that the y-m interaction is the dominant variance comB  ponent from 1 m to 5 m (Tabie II.2). There is also a suggestion of an increase in w  2  over  this depth range. However, this latter trend is obscured when the seasonally truncated 41  Table I I . J to of Chla at stn 1-6, as well as tor a May-October subset of (he Chlo, data at stn A-G. Asterisks indicate number of signiBcant F-ratios (see text for details). VC not reported if F<1.0. 9  z  Source  Chlo,  Chlai  Chlo*  Chloj  Chi flu  Year (y) stn (s) Mth. (m) y-s y-m s-m y-s-m  3.9 3.0* 25.0" 1.3 65.9  4.9 2.6 8.8 17.8" 65.8  0.4 2.6 11.3 32.1" 53.5  2.6 2.6 11.9 38.1" 2.4 42.4  1.6 7.8 13.6 0.1 3.3 10.0* 63.7  Chlog data is included in the comparison (Table II.2). The summary of the ANOVA for PI and QI is tabulated in Tabie 2.3. The significant y-m interaction is the dominant component for both PI and QI (20% and 24% respectively). For QI this represents almost all the variance accounted for, with the difference in w  2  between the two indices attributable mostly to the small variance associated with  the year and month main effects in QI. With w  equal to 35% and 25% for PI and QI  2  respectively, the variance accounted for is somewhat less than for either Chla, or ANOVA of P  B  z  2  B  B  and Q indicates that a robust y-m component dominates at depths <  10 m (<tw > « 2 8 % and 35% for P that P  S.  B  and Q respectively, Tabie II.4). It should be noted z  was normalized to <I > rather than the near-field irradiance <J* > so as to h  s  maximize df , since estimates of the vertical attenuation profiles were frequently absent from the ICM data set. However, an analysis of the irradiance record later on will show significant spatial effects for <I > only at z > 5 m (i.e. Table H.5), whereas these variance Z  components appear in P  B  only for z < 5 m (Tabie H.4). The major  components for Qio are the non-significant month and station main effects, and nonsignificant s-m interaction, while for P  B 0  they are the non-significant year and month main  effects and the non-significant y-m interaction, w  2  P  B Q  is 50% for P  B  and Qi and 25% for  and Qio . However, the distribution of variance among components differs for the  two indices. P  B  has a significant s-m component (11%) and Qi does not; while at 10  m the relative importance of the non-significant y-m and s-m interactions, and the nonsignificant year and station main effects are reversed between the two indices. Thus year 42  Table II.3 ANOVA of PI and QI. Parentheses as in Table 2.1. Err. Term  SS  dj  MS  PI Year (v) stn (s) Mth. (m) y-s y-m s-m y-s-m  ys+ym-ysm ys+ms-ysm ms+ym-ysm ysm ysm ysm  77.08 6.70 60.56 41.29 190.4 90.06 343.5  5 5 6 25 30 30 150  15.42 1.34 10.09 1.65 6.35 3.00 2.29  2.70 0.1 0.57 >0.5 1.43 0.5 0.72 0.08 2.77 0.001(.2) 1.31 0.15  QI Year (v) stn (s) Mth. (m) y-s y-m s-m y-s-m  ys+ym-ysm ys+ms-ysm ym+sm-ysm ysm ysm ysm  0.112 0.026 0.021 0.061 0.295 0.112 0.495  4 5 5 20 20 25 100  0.028 0.005 0.004 0.003 0.015 0.004 0.005  2.17 2.03 0.30 0.62 2.98 0.90  Source  F  P  VC  0.50 0.50 >0.50 >0.50 0.001(.2) >0.50  0.23 0.00 0.08 0.00 0.68 0.12 2.29  6.8 0.0 2.4 0.0 20.0 3.5 67.4  0.0001 0.0000 0.0000 0.0000 0.0016 0.0000 0.0050  1.5 0.0  -  23.9  -  74.6  and month main effects and y-m interaction are dominant for Pf , while station and Q  month main effects and s-m interaction are dominant for Q Q . At 3 m and 5 m the X  respective total variances are for Pf w 60% and 50% respectively; and for Q w 50% and z  40% respectively. Table II.4 w* of P* and Q . Asterbks and dash as in Table 2.2. z  Source Year (y) stn (s) Mth. (m) y-s y-m s-m y-s-m  Pf 4.9 0.0 6.3 0.0 26.1" 10.8* 51.9  P? 7.6 0.0 5.9 3.9* 28.8" 11.9* 41.9  9.7 0.0 9.6 1.6 28.0" 2.6 51.4  pB 10 r  7.8 2.5 7.9  -  6.8  73.9  Qi  Q»  Qs  Qio  7.4 0.0 0.0 0.0 37.3" 2.5 53.1  4.3 0.8 0.0 0.0 41.7" 0.0 53.1  11.5 0.7 0.0 1.7 25.3" 0.0 60.7  2.9 9.0 9.9  0.6 5.3 72.2  Again, the relatively constant proportionality over depth belies the underlying differences in components distribution. For Pf there is a significant s-m interaction (12%) which partly offsets the larger y-m variance component in Q (28% vs. 42% respectively). The 3  remainder of the 10% greater total variance accounted for in Pf is due to a significant y-s interaction (4%) and a non-significant month main effect (6%) not present in Qz . At 5 m, by comparison, most of the total variance contrast resides in a larger non-significant month main effect component in Pf (10% vs 0% respectively).  43  £.£.4-3  Interpretation of the variance structure  The ANOVA of Chla and carbon uptake distribution can be summarized as follows: 1) The y-m interaction is the dominant variance component for all indices during the May-October growing period. This dominance is independent of depth to 10 m. 2) Spatial effects are absent in Chla specific biomass except for a significant s-m interaction in Chi ax 5 and a significant y-s interaction in untruncated Chla, . 3) Significant spatialtemporal interactions, are present in P  B  and P  , but absent when P  B  B  is normalized  to <Ig >. 4) The only significant main temporal effect in any variable is the significant month main effect in S . B  The general lack of orthogonal (i.e. main) variance components in the ICM data set can be interpreted to mean that persistent spatial and temporal structure in phytoplankton biomass and carbon assimilation distribution are small compared to their fluctuations across these main sampling effect boundaries (Lewis and Piatt 1982). Though this is not evidence that such persistent features do not exist, it indicates that this structure is weak compared to their interactions. This is cogently referred to by Lewis (1978) as ephemeral spatial structure. The nature of this interaction is illustrated by comparing the sampling scales to the characteristic scales of the basin phytoplankton. This is begun by characterizing the critical length scale L for phytoplankton biomass, defined as L = e  c  ir(Khfp) ^ , where Kh is horizontal diffusivity and p. the phytoplankton specific growth 1  2  rate. This is a scaling argument defining the horizontal patch size at which growth and dispersive effects are equal, allowing the patch to remain in steady state (Okubo 1978). Substituting the known range in p. (1.2 • 1 0  - 5  to 1.0-10 s ) and a typical surface mixed -6  - 1  layer value of Kh (10 cm /s) gives estimates of L ranging from 2-50 km. This critical 5  2  c  range delineates a spatial domain where neither growth nor mixing processes particularly dominate and hence the relative scaling is sensitive to small changes in either one. Legendre and Demers (1984) conclude that as a consequence spatial heterogeneity is at a maximum within this region, with mixing acting to homogenize the phytoplankton distribution on shorter length scales and phytoplankton growth compensating for the effects of turbulent 44  mixing on longer length scales in each case regardless of the combined variations in growth and mixing rate. The reason for this is that on length scales shorter than the transition zone the time-scale of kinetic energy transfer between adjacent eddies in wave number space is short compared to the e-folding time of the phytoplankton, while on longer length scales, it is the phytoplankton generation time scale which is short compared to the time-scale of the turbulent energy cascade (Denman and Piatt 1976). The 2-50 km estimate for this transition zone was determined by substituting the range in sensible growth rates. As an alternate approach, the empirical analysis of Chla variance over wave number space indicates that a transition from - 5 / 3 to —1 in the slope of this dependence (i.e. a region analogous to L ) occurs at length scales ranging from » 1-10 km e  (Denman and Piatt 1975, Fasham and Pugh 1976, Lekan and Wilson 1978). Most of the length scales in the I C M sampling program (i.e. linear separations between stations) are within this latter range (Figure 2.8) and hence the protocol is in this respect suited to sampling the dominant variance structure of the basin phytoplankton. However, since length and time scales are closely coupled (Lewis and Piatt 1982), the sampling time interval must also be appropriate to ensure adequate identification of the individual components. Okubo (1971) obtained an empirical relation (i.e. least-squares fit) between the diffusion time-scale and the horizontal patch size for a conservative contaminant (rhodamine dye), wherein the horizontal patch size expressed as a standard deviation cr = 0.6-t  1-17  . Substi-  tuting the lower (1 km) and upper (10 km) bound of the critical region and solving for t gives diffusion times from a point source and hence characterisitic persistence time-scales of » 2 and 9 days respectively. The implication is that if the sampling protocol is to separate spatial structure from fluctuations in L stemming from time-dependent changes c  in n and/or Kh (i.e. orthogonal spatial and temporal components), its sampling interval should be comparable to this range. Consequently, it would seem that the within station sampling interval of the I C M sampling protocol ( « 30 days) is up to an order of magnitude too long to allow the necessary separation.  45  2.8.4-4 Inferences About the variance structure The presence of significant temporal structure in S  B  and its absence in the depth  specific biomass indices (i.e. seasonally truncated Chla, and Chlo, ) suggests that the annual cycle is obscured in these latter variables by the inter-annual variability in the vertical biomass distribution. Furthermore, the precise nature of this variability may be depth-dependent, with the dominant spatial-temporal interaction being y-m for 1 < z < 5 m and s-m for z = 15 m. A scatter plot of Chlais vs. the 0-5 m integration of Chla,, indicates positive covariance (Figure 2.14). This covariance is to be expected with the vertical fluxes redistributing the surface concentrations in Chla specific biomass. What is more interesting however, is explaining by what mechanism the significant shallow y-m variance components map onto the significant deep s-m component. One possibility is that the one or more processes which affect the vertical distribution of phytoplankton, are spatially dependent in the horizontal. Potential processes for this effect are a vertical mixing regime dependence on the QN entrance jet as a turbulent kinetic energy point source, and a horizontally dependent static stability field which modifies the dependence of the vertical mixing regime in the horizontal.  Spatial effects in < i | > are small, restricted only to the s-m interaction (presumably attributable to the typical 3 day time interval required to complete the carbon uptake experiment at all 6 sampling stations, I C M 1979) and hence could not have forced the y-s component in Chlog . Significant spatial-temporal components (i.e. in particular the y-s interaction) do appear in the </£ > record normalized to <Ig >. However, this is restricted to depths > 3 m (Tabie II.5) suggesting that variance in the phytoplankton biomass and hence in the phytoplankton self-shading effect is more likely to be the cause of the component variance structure in the column irradiance field rather than the converse cause and effect relationship. On the other hand, the possibility of this switch in cause and effect (i.e. fluctuation in the irradiance field forcing the structure in phytoplankton biomass) remains a possibility, since variance in the depth integrated column light attenuation must increase with depth, and hence pattern not detectable near the surface could emerge deeper in the 46  cn  O O CD  1  0.04  '15M  CHLA  1 (MG C H L R / M X X 3 )  Figure 2.14 Scatter plot of Chlo / vs. near surface (i.e. 0-5 m) integrated Chla ; pool of ICM stations 1-6. 5  12.54  water column even if the pattern itself were not depth dependent. This would then explain equally well the lack of detectable variance structure in near surface irradiance.  Table II.5 u>*for</|> and <J*>/</|>. Particulars as in Table U.2. Source Year (y) stn (s) Mth. (m) y-s y-m s-m ysm  < J* >  </*>/</*>  </»>/</»>  </»>/</»>  </J > / < . / » >  10.1  4.5 0.1 10.1*  28.4*  24.1*  39.3" 4.8 45.0  10.6* 74.5  15.r 9.4 15.1* 59.8  6.7* 16.7" 8.1" 40.1  9.8* 13.8* 4.0 48.3  0  The s-m variance component in Pf and Pf can be at least partly attributed to the presence of the finite (though statistically zero) s-m component in <I > (Tabie JTJ.5), h  s  since normalizing Pf and Pf to < i | > eliminates the effect (Tabie H.4), though again this must be interpreted as an artifact of non-coincidental carbon uptake measurements, since as was mentioned in regard to the Chla, variance structure, spatial-temporal variance in <J* >/</$ > is restricted to z > 3 m and cannot therefore have contributed to Pf and Pf variance structure. However, <i* > / < / | > does have a significant year effect at depths > 1 m. For the reason presented earlier, appearance of this component structure higher in the water column compared to the spatial-temporal variance structure implies that the inter-annual variation in column light attenuation is comparatively larger. A tabulation of the depth specific yearly means in <J* >/<Ig > suggests a periodic pattern with the maximum in column attenuation occurring in 1979 (Tabie H.6). Interestingly, this pattern is coherent with the observed rate of sediment deposition near Hankin Pt (Tabie II.8) inferred from bathymetry measurements: Net deposition was negative from 1977 to 1980 and positive from 1980 to 1982 (ICM 1983). The 1977-1980 erosion regime is attributed to tidal scouring of these sediments (ICM 1979). Since this results in an increased surface layer sediment load, periodicity in the deposition regime could impart a similar pattern on the general basin surface layer light extinction field. 48  Table II.6 Group means of depth specific irradiance during C uptake experiments normalized to inc BuxandofPf. 14  Depth  1976  <J* > / < / » > 0.67 lm 3m 0.32 5m 0.15 10m 0.03 Pf Pf Pf Pf Pfo  -  -  1977  1978  1979  1980  —  0.59 0.23 0.09 0.01  0.56 0.18 0.06 0.003  0.71 0.27 0.08 0.008  3.73 2.67 1.93 0.67  4.74 4.19 3.31 2.95  3.39 2.60 2.01 1.50  3.29 2.26 1.80 1.39  -  -  1S81  1982  -  0.88 0.44 0.19 0.04  —  3.80 2.70 1.57 1.03  4.64 3.64 2.83  -  The amplitude of the oscillation in <i* > / < / | > was sufficient to reduce irradiance at 10 m below 1% of surface irradiance in 1979, which is traditionally viewed as the net photosynthesis compensation depth (Parsons et al. 1984, p. 75). However, the year effect in Pf variance (and Pf in general) was not significant (TaWe 2.4). This implies either that Q  the inter-annual variation in the irradiance regime is within the photo-adaptation capabilities of the phytoplankton, either on a cellular (Prezelin (1980) or community (Raymont 1982, p. 109) level, or other processes are of greater important in controlling primary production. There was a significant y-s interaction in Pf (Tabie II.4). However, since this variance component was negligible in 0-3 m <i|  >/</* >, there is no reason to attribute  this to irradiance forcing. Turning to the biomass variance structure, one possible argument explaining the y-s interaction effect in Chlag compared to its lack in S  B  and Chla, is the greater spatial  sampling domain of the Chlag stations (locations, Figure 2.8). This is demonstrated by comparing treatment class deviations from the grand mean for stn A-G (TaWe H.7), with stn G at the western end of Holberg Inlet and stn F at the eastern end of Rupert Inlet (Figure 2.8) having the two largest deviations. Since both these stations are outside the stn 1-6 sampling domain, one interpretation is that these stations represent unique regions of the basin. An alternate (or complimentary) explanation for the y-s component is related to the 49  Table II.7 Station and year Chla, marginal deviations from grand mean (4.64 mgChh /m ). 3  Station  Deviation  Year  Deviation  0.18 0.15 0.42 -0.18 -0.37 -0.95 0.74  1974 1975 1976 1977 1978 1979 1980 1981 1982  1.09 -0.72 -0.84 0.83 0.69 -0.07 0.66 -0.46 -1.18  A B C D  E  F G  observation by Piatt et ol. (1970) in St. Margaret's Bay, Nova Scotia that between station variance is dependent upon the total area sampled . This dependence was attributed to the interaction of phytoplankton patch and inter-patch spacing with between station spacing, since it disappeared as the sampling domain was increased (i.e. by increasing the interstation gap) beyond « 1 k m (tested up to « 10 km ). Direct extrapolation of this 2  2  sampling effect to the Holberg-Rupert basin is confounded by the more complex shape and greater length to width ratio of the latter. Nevertheless, the surface area domains of stn 1-6 and stn A-G (i.e. « 8 k m and « 10 k m respectively) are well beyond the area2  2  dependent part of the St. Margaret's Bay station gap sampling effect and can probably be discounted as a possible basis for the y-m component in Chla, .  2.2.5 Partial Correlation Analysis of the ICM Data Set 2.2.5.1 Introduction In the foregoing, analysis of variance was used to identify the dominant spatial and temporal structure in phytoplankton biomass and growth indices. Irradiance is bound to be an important forcing variable of primary production in the basin. Consequently, structure in the light field was also evaluated mainly tc suggest sources for its variation, though inferences were also drawn from comparisons with the corresponding structure in the phytoplankton indices. For the remainder of this chapter relationships between the indices of phytoplankton biomass and carbon uptake, and the potential forcing variables of irradiance and column stability will be more rigorously inferred from their covariance patterns. 50  £.2.5.2 Statistical method Both parametric (partial correlations) and non-parametric (Kendall's Tf, of concordance) statistics are used in the analysis. Kendall's TJ, amounts to a simple correlation of ranks (Sokal and Rholf 1981, p. 601), while partial correlations can be viewed as representing the product-moment correlation between the residuals from separate least-squares fits of the dependent variables against one or more common covariate. This linear defending with respect to the covariates estimates the direct covariance between the dependent variables (Sokal and Rholf 1981, p. 656), and thus differs from Kendall's T& which cannot differentiate between direct and confounded effects. As a general approach to the correlation analysis, a preliminary matrix of TJ, coefficients was calculated, since this statistic makes no underlying assumptions about data distribution (Sokal and Rohlf 1981, p. 601). This initial analysis therefore allowed the identification of strong interactions prior to embarking on the more laborious parametric treatment. The disadvantage of such an approach is that the non-parametric test suffers from a loss of statistical power (i.e. detecting a false H ). To partially compensate for 0  this, significance of the Tf, statistic was tested at the P<0.10 level. However, due to the more rigorous parametric treatment which always followed, this reduced significance level did not jepardize the overall Type I error level of the analysis (Zar 1974, p. 44).  2.2.5.8 Interpretation of the covariance structure Underlying statistical relationships that could possibly contribute to the annual cycle in S  B  are evaluated in Tabie II.8 which presents the T& matrix for S ,  PI, < / | > and  B  the 0-5 m depth averaged ambient temperature <T>; as well as the column stability parameters E, Et , and Eb described in Sec. 2.2.3.1. To maximize the covariance, the data set was divided into a basin (stn 1-4) and QS (stn 5 & 6) data subset. Significance levels were calculated according to Sokal and Rohlf (1981, p. 606). The only significant covariance (P < 0.10) in S  B  and PI in QS is the correlation  between <Ig > and PI, and the anti-correlation between year and S . B  51  In the basin  Table II.8 Kendall jfc matrix of production variables tor the basin and QS subsets of the ICM data set. I addition to the 0-5 m depth averaged ambient temperature <T>, the matrix consists of S , PI, </£ >, E, S , and Et described in Sec. 2.2.3.1. Asterisk denotes significant concordance (P<0.10 ). B  B  S  <T>  PI  B  basin (:n=165) PI -0.052 <T> 0.037 0.019 </| > 0.235* 0.234* -0.173* -0.131* £ -0.015 -0.109* E -0.223* -0.119* E Mth -0.136* -0.106* Stn -0.071 -0.003 -0.161* 0.063 Yr Quatsino Sound (n=77) PI -0.107 <T> 0.104 -0.008 0.136* <I s > 0.061 -0.067 -0.064 E -0.009 -0.087 E -0.038 -0.049 E Mth -0.014 -0.070 0.041 Stn 0.112 Yr -0.220* 0.077 b t  b  b t  by comparison, both S  E  £6  Et  0.020 -0.019 -0.065 -0.013 0.500* -0.061 -0.037  -0.124* -0.053 -0.126* -0.148* 0.045 -0.184*  0.505* 0.847* -0.121* -0.049 0.056  0.354* -0.222* -0.027 0.040  -0.065 -0.049 0.067  -0.016 0.102 -0.008 0.152* 0.608* 0.149* -0.042  0.001 0.012 -0.019 0.008 -0.089 -0.199*  0.692* 0.631* 0.027 0.250* -0.045  0.323* -0.134* 0.318* -0.060  0.108 0.113 -0.035  B  and PI are anti-correlated with at least one of the stability  parameters and correlated with < J | >. In the basin, both S with month, while S  B  and PI are also correlated  is also anti-correlated with year.  B  <Ig >, E Et, , S , and PI are all anti-correlated with month. Hence it is not clear B  t  what covariance can be attributed to direct interaction as opposed to a common interaction between one or more covariates. Since the object is to clarify the month effect variance component in terms of the environmental forcing variables, the month effect per se must be removed. To a lesser extent this also applies to the year effect since it covaries with S and < / | >. However, due to the size of the month variance component in S  B  B  (Tabie H.l),  these two effects were removed differently. While the correlation matrix was detrended with respect to the year effect by calculating partial correlations fixed relative to year, the month effect was removed by calculating this partial correlation matrix on a per month basis. The partial correlation coefficient is a parametric statistic, and hence the data were first tested for departure from normality, with details of the test and the data transformation procedure outlined in Appendix B . Significance levels for the partial correlations were 52  determined according to Sokal and Rohlf (1981, p. 583) using the z* transformation of the coefficients. Unlike the correlation coefficient itself, this transformation is normally distributed for n > 10. The standard deviation o~' used in the associated t-test was s  estimated by a' = (n — 3 — m ) / , where m is the number of variables fixed. - 1  3  t  The partial correlations between S  B  and PI and the forcing variables are tabulated  in Tabie II.9, while the entire matrix is presented in Appendix C l . Based on Tabie II.9, S  B  is anti-correlated with column stability except in April (note that n < 10), July in  QS and in June, July, August in the basin; of which the July correlation and Oct anticorrelation in QS; and the June, July correlations and May, Oct anti-correlations in the basin are signficant (P<0.05). PI is generally correlated with < i | > throughout (the exceptions are Aug in the basin and May, Oct in QS). However, only May, Oct in the basin and June, July in QS are significant (P<0.05). Except for Sept, PI is generally anti-correlated with stability in the basin, while in QS it is divided into a distinct AprilJuly period when it is correlated, and an Aug-Oct period when it is anti-correlated with stability. Only the anti-correlations for May, Oct and the correlation for Sept in the basin, and the anti-correlation for April and July in QS are significant (P<0.05). Since the covariance between < J | > and one or more of the stability indices could potentially confound interpretation of the correlation between PI and <Ig > and between PI and column stability, a partial correlation matrix was also calculated fixed with respect to <Ig > as well as station and year. However, the only change of any consequence resulting from this adjustment are a loss of the significant anti-correlation between Ek and PI for the month of May in the basin (Appendix C2). The correlation between S  B  and < i | >  in Tabie II.9 must be assumed due mostly to the covariance between PI and <7| >. Nevertheless, this does not compromise the interpretation of the other covariances, and hence the appropriate correction was not made. In general then, the covariations between the phytoplankton indices and the environmental variables are not seriously confounded as originally presented in Table 2.9. To explore the relationship between carbon assimilation and environmental factors in 53  Table 11.0 Partial correlations with yearfixed,of S and PI to the environmental forcing variables < J E, Et , and E for monthly groupings of the within basin and QS data subsets. Asterisks denote significa correlations (P<0.05). Sample size in parentheses. B  b  April  S  basin  B  pj  S  B  May  pj  S  B  June pj  B  S  July pi  B  S  Aug  pj  S  B  Sept  pi  S  B  (12) -.313 .069 .164 -.043 -.533 -.013-.191 -.276 -.546  (23) .138 .198 .591* -.521* -.403 -.518* .040 -.509* -.434*  (24) -.281 -.557* .142 .178 -.209 .480* -.282 .099 -.150  (22) -.452* -.096 .052 .094 -.198 .773* -.161 -.160 -.163  (28) (28) -.273 -.271 .421* -.1 37 -.443* .137 .031 -.206 -.253 .603* .157 -.195 -.323 .193 -.054 -.077 -.150 .581*  (6) PI -.541 <4 > --281 .115 E .313 -.952* E .207 -.935* E .270 .427  (12) -.334 .836* -.163 -.038 -.285 -.077 .003 .164 -.555  (12) .119 .250 .649* -.297 -.483 -.280 -.375 -.210-.516  (10) -.872 -.802* .778* .816* -.868* .869* -.820* .681 -.798*  (13) -.223 .360 .415 -.313 .109 -.424 .020 -.08 5 .309  PI </' > E E E QS b t  b t  (13) -.544 -.622* .407 -.539 .230 -.537 .138 -.355 .355  Oct  pj  (28) -.001 -.187 .433* -.479* -.459* -.292-.573* -.565* -.379 (11) .415 -.076 -.547 -.276 .514 -.737* .122 -.039 .329  more detail, a partial correlation matrix of Pf , </£ >, depth specific ambient temperature T , and E, Et , Eb was calculated fixed relative to station, month, and year. The z  correlations specific to Pf are presented in Tabie 11.10, while the complete matrix is in Appendix C3.  Table 11.10 Partial correlationsfixedto station, month, year of Pf , depth specif c temperature irradiance <I >, and E, Et , and S . Astensk denotes statistical significance at P<0.05. Dash deno nonsensical comparisons. B  Z  P* basin n=89 0.073 r, T T Tis 0.207 </i > </ > -  b  0.204* 0.304* 0.314* -0.459* -0.383* -0.337*  s  0.302  A s  s  </ > 5  </io > E -0.172 -0.166 E -0.142 E QS n= 41 r, 0.641* r T* Tn 0.155 </i > </s > t  </ > 5  </io > E E E t  b  -  0.211 0.188 0.225  pB  ' 10  Pf  -  -0.191 -0.244 -  -  0.232 0.214 -0.001  -0.164 -0.117 -0.091  0.129 0.173 0.010 -0.032 -0.138 -0.057 -0.050 -0.050  0.452* 0.136 0.160 0.110 -0.018 -0.048 -0.013  0.013 0.017 -0.155 -0.228 0.162 -0.302 -0.251 -0.252  0.248* 0.176 0.090 0.112  .  54  In the basin, Pf is significantly correlated (P<0.05) with T at all depths except 10 m, x  while it is significantly correlated with irradiance and anti-correlated with stability only at 3 m. In QS by comparison, the only significant co-variation is a correlation between Ti and Pf and between T and Pf . 6  Temperature and irradiance are correlated (Appendix C3), presumably due to the local climatic pattern whereby the annual increase in total daily heat flux increases surface water temperature. To correct for this confounding effect on the estimated covariance described above, a correlation matrix was calculated fixed with respect <I > as well as station, $  month, and year. The correlation coefficients specific to Pf are presented in Tabie 11.11, with the complete matrix in Appendix C4. Table 11.11 Partial correlations as in Table 11.10fixedto </|> as well as station, month, year. Pf basin n=89 Ti 0.020 T T r E -0.205 -0.171 E, -0.246* E„ QS n =39 Tx 0.428' r, s s  1 5  r E E E  1 5  t b  -  0.234 0.230 0.243  Pf  Pf  pB  -  -  -  -0.032  -  0.017  •MO  -0.484* -0.441* -0.380*  -0.151 -0.111 -0.145  -  -0.120 -0.015 0.058 -0.146  -  -  -  -  0.183 -0.330 -0.291 -0.309  0.285  -  0.231 0.199 0.011  0.149 -0.004 -0.014 -0.029  With the matrix detrended with respect to < / | >, T is no longer significant (i.e. z  P > 0.05) except for a correlation with Pf in QS. This is consistent with the interpretation of the irradiance temperature covariance suggested above. By comparison the anti-correlation between stability and Pf in the basin remains significant in Tabie 11.11, indicating indirect effects such as weather pattern modification of both irradiance (i.e. cloud cover and rainfall/run-off effects on column light attenuation) and column stability (i.e. rainfall/run-off buoyant effect) are not likely to be significant processes underlying this covariance. Finally, the partial correlation matrix (fixed with respect to month, station, 55  and year) of Chla, and the stability indices for the basin and QS data subsets, indicates significant anti-correlation (P < 0.05) between column stability and Chios ,Chlo in both 15  the basin and in QS (Tabie 11.12). Thus though horizontal variation in turbulent kinetic energy may still be an important factor in the appearance of the spatial-temporal biomass variance structure at depth (see Table II.2), the anti-correlation with column stability suggests that horizontal variation in E contributed as well. However since 2? is also dependent to a degree on vertical mixing, this covariance pattern is equivocal. Table 11.12 Partial correlations Axed to stn , mtb, year of Chla to E, E , and S . statistical signiBcance at P<0.05. B  z  Chla, basln=129 E 0.060 0.015 E 0.165 E QS n=59 E -0.113 -0.037 E -0.164 E t  b  t  b  Chla,  Chios  -0.137 -0.169 0.016  -0.249* -0.272* -0.092  -0.202* -0.195* -0.154  -0.265* -0.189 -0.249  -0.294* -0.229 -0.251  -0.363* -0.261 -0.342*  Chla  t  Asterisk denote  15  The following summarizes the structure that emerges from the variance and covariance analysis: 1) S is generally correlated with column stability during mid-summer and anti-correlated B  during spring and fall, with the slight indication that when they are anti-correlated (i.e. spring/fall) the strongest covariate is Et as opposed to Eb when they are correlated (midsummer). Neither this pattern in E and Eb nor the cyclical transition from correlated to t  anti-correlated covariance is apparent in QS. 2) PI both in the basin and in QS lack the distinct annual pattern in covariance with column stability (particularly QS), being generally anti-correlated except for the fall. However, there is a slight pattern of degree in the E% and Eb anti-correlation with PI, with the E -PI anti-correlation dominating in the spring/fall and both stability parameters of t  equal importance during the mid-summer. This latter shift in dominance is not associated with the < / | > covariate (Tabie H . l l ) . 56  3) PI and Pf are generally not correlated with temperature when corrected for irradiance. 4) Based on the depth specific pattern in co-variance, the PJ-stability anti-correlation in the basin is centered at 3 m and 5 m, and the PI-<Ig > covariation is centered at 3 m. In other words the correlation between Chlo specific carbon uptake and both these forcing variables coincides approximately with the base of the pycnocline. 5) With the exception of the strong correlation with temperature at 1 m, no environmental variables co-vary with Pf in QS. 6) Horizontal variation in column stability appears to be implicated in Chloj variance structure at z > 5 m.  2.2.5.4 Summary Qualitative examination of a prior data set from the Holberg-Rupert basin indicated a phytoplankton growing period consisting predominantly of a season long (i.e. monotonic) increase in Chla specific biomass and primary productivity, though the growth efficiency indices PI and QI showed evidence of a mid-summer depression. Superimposed on this annual cycle was an inter-annual variation of comparable magnitude. The variance analysis of this data set could not identify any persistent horizontal structure (i.e. on a 30 d time scale) except for surface Chla specific biomass in headward regions of the basin. Significant temporal variance structure during the Apr-Oct growing period was restricted to the annual pattern in surface and a 0-15 m integration of Chla specific biomass. Analysis of the covariance between the primary production and forcing variables available indicated a strong relationship between carbon uptake rate, irradiance, and column stability. The biomass-stability covariance showed evidence of a seasonal switch in sign that could be related to seasonal changes in the tidal flow buoyancy regime. No seasonal pattern was evident in the uptake-stability covariance, being mostly anti-correlated and 57  strongest near the pycnocline. In general, this analysis of the prior data set demonstrated a significant coupling between primary production and the mixing regime of the basin. 2.2.6 General Discussion The seasonal pattern of primary production in the basin differs significantly from the pattern in a run-off dominated coastal water body such as Strait of Georgia, British Columbia. Generally speaking, the pattern in the latter is one of a pronounced bloom in May-June coinciding with the start of the summer freshet, followed by up to a 2-3 fold reduction in ambient nutrients levels, Chla specific carbon uptake, and phytoplankton standing stock for much of the summer (Stephens et al. 1969, Harrison et al. 1983). Some of this characteristic pattern in biomass can be attributed to changes in grazing pressure (Harrison et al. 1983). However, the concurrent reduction in carbon uptake efficiency must be attributed to an associated reduction in ambient nutrients levels, particularly nitrogen. Exceptions to this generalization appear to be more distal regions (i.e. relative to the point of dominant riverine discharge) and regions within an island archipelago. These regions of Strait of Georgia, but in particular those in the island system, demonstrate a seasonal phytoplankton biomass pattern resembling that of the Holberg-Rupert basin, with phytoplankton standing stock increasing (and presumably carbon uptake efficiency remaining high) throughout much of the summer (Stephens et al. 1969). The common hydrological feature of these atypical regions of Strait of Georgia is that they are relatively free of the buoyancy effect of the summer freshet, due either to distance or to pronounced vertical mixing in connecting island passages. Studies in Saanich Inlet (i.e. in the island system) by Takahashi et. al. (1977), Parsons et al. 1981, and by Parslow (Dept. of Oceanography, University of British Columbia, British Columbia Canada; pers. comm.), indicate that phytoplankton standing stock and specific carbon uptake there are coherent with surface layer nutrient replacement, which appears itself to be forced by the periodic variation in amplitude of the tidal excursion 58  through nearby Juan de Fuca and Haro Straits. It is suggested that this high frequency input of nutrients maintains elevated biomass and production levels throughout the summer. By comparison, the surface layer nutrient flux in the Strait of Georgia in general is insufficient to support a comparable rate of primary production following spring bloom depletion of the in situ ambient nutrient supply (Harrison et al. 1983). Hence, nutrientlimited growing conditions predominate until the fall reduction in river discharge, with its concomitant increase in the vertical nutrient flux. The relatively small amplitude summer time oscillation in phytoplankton productivity (i.e. of the same order as the inter-annual variability) in Holberg-Rupert basin is consistent with a high nutrient flux regime (or at least a euphotic zone that is comparatively nutrient replete throughout the year). Furthermore, it is compelling to attribute this elevated nutrient regime to the forcing by the mixing regime (Tett and Grantham 1979). An evaluation of this hypothesis, as well as related effects such as the effect of vertical motion on the near field light flux of the phytoplankton, requires a characterization of the general nutrient and light status of the phytoplankton. The only historical research effort with this specific aim was the study by Stevens and Silbert (1976) assessing the dependence of phytoplankton biomass on column static stability. The data were obtained over an intensive sampling period (Aug 17-Sept 10 1975), with Chla levels compared to the 0-20 m finite difference approximation of E. The covariance indicated a correlation between biomass and stability from which it was inferred, that given the general absence of nutrient depletion, primary production was limited by the surface layer mixed depth (i.e. light limited). The timing of this study corresponds to a transition in the ICM data set from when S  B  was correlated with stability to when it was anti-correlated (Tabie II.9). Hence a com-  parison of the two data sets is inconclusive. Nevertheless, if a cause and effect relationship can be assumed responsible for the covariance relationships in the ICM data set, then the picture is more complex than presented by Stephens and Silbert (1976), with mixing apparently both promoting carbon uptake rate (i.e. particularly near the base of the pycn59  ocline) as well as dispersing the phytoplankton, with the dispersion effect appearing to be important only during the mid-summer months. In other words, during the spring and fall the promotion of carbon uptake (i.e. growth) through vertical mixing (by whatever specific mechanism) is large in terms of primary production compared to concurrent mixing loss of biomass, while during the mid-summer months, the dispersive loss of biomass is as equally importance as this carbon uptake enhancement effect. The anti-correlation between stability and carbon uptake can be conventionally rationalized in terms of nutrient-limited growth, with the vertical nutrient flux (or flux of some other growth stimulant, Piatt et al. 1977) limiting growth. An alternate explanation which has received recent attention and which adds support to the interpretation of Stevens and Silbert (1976) is based on a mixing regime-dependent modification of the time averaged irradiance flux, acting to ameliorate the effects of supersaturating light conditions (Harris 1973; Marra 1978; Piatt 1980; Marra and Heinemann 1982). An analysis of the Chla, -column stability covariance was not done on a monthly basis. However, the covariance linearly detrended with respect to month indicates that significant co-variation occurred only at z > 3 m as opposed to 3 m for carbon uptake. In other words, it gives some suggestion that the concurrent mixing effects of uptake promotion and biomass dispersion may be displaced in the vertical, with the uptake enhancing effect higher in the water column than the biomass dissipation effect. On the other hand, this apparent displacement could be an artifact of the inherently low signal-to-noise ratio in measured carbon uptake rate estimate at depth. Accordingly, the counter-intuitive anticorrelation between Chla specific biomass and column stability has several possible causes. Elevated deep Chla concentrations could simply reflect primary production higher in the water column, which is in turn anti-correlated with stability. However, this places constraints on the nature, magnitude, and depth dependence of the mixing regime due to the simultaneous dispersion of the phytoplankton. That is, an increase in mixing that raises the nutrient flux (and hence elevates carbon uptake under the assumption that production is nutrient-limited) would ordinarily equally increase the dispersion of the phytoplankton, 60  since the eddy diffusivity for these two properties are comparable. Hence an increase in phytoplankton biomass following an increase in mixing requires either that mixing is associated with an upward directed entrainment flux into a predominantly euphotic surface layer, mixing is primarily wind generated and hence stability effects on mixing are restricted to the top few meters of the water column, or vertical transport either via an active behavioral response by motile phytoplankters or gravitational sedimentation of sufficient magnitude (see below) ameliorates turbulent dispersion. Analysis of a near surface current record indicated that buoyant tidal flow may predominate at least regionally in the basin during the summer. If this flow is subcritical within the euphotic zone it can effect an upward transport of nutrient rich water from below without seriously dispersing the phytoplankton in the vertical. However, for such a flow regime to occur (i.e. entrainment interface above the net photosynthesis compensation depth, see Parsons et al. 1984, p. 71) for a significant part of a tidal cycle, appears to require positively buoyant flow conditions that are associated only with periods of high run-off. This is not likely to be a common condition during the summer months, but can occur during the spring/fall (Sec. 2.1.2.2). Interestingly, the seasonal switch in the r run-off regime (e.g. Figure 2.4) corresponds with the switch in the 5 -stability covariance fl  pattern, though the associated sign of the covariance is inconsistent with the foregoing argument. Alternatively, a wind generated surface turbulent layer could also entrain limiting nutrient into the euphotic zone. Again, if the entrainment interface remains shallow, concommitant vertical dissipation effects in the productive surface layer and on biomass accumulations deeper in the water column would be minimized. The depth of this interface would be determined by the strength of the wind and as in the case of tidal mixing, by the buoyancy regime. There is no historical data with which to directly evaluate the contribution by seasonal patterns in wind speed. Though a role of wind generated mixing is implicated by the slight suggestion of a Et -PI anti-correlation dominance during spring/fall when stability and S  B  are also anti-correlated, compared to the mid-summer 61  Et -PI and Eb -PI equality (implying that PI is also forced by bottom mixing) when stability and S  B  are correlated; it is not possible to determine whether this suggested  shift is due to seasonal patterns in the wind field, buoyancy flux (i.e. both wind and tidal mixing consequences) or both. Finally, there remains the possibility that an upwards directed entrainment flux and/or a depth dependent mixing regime are not important in characterizing biomass dispersion and mixing enhancement of growth. Riley et al. (1949) in analyzing the dynamical interaction between cell sinking, vertical mixing, and depth-dependent phytoplankton growth, were able to show that the combined action of vertical smearing and growth could under appropriate combinations of these three terms counteract the loss due to a depth-independent sedimentation, resulting in a steady state subsurface Chlo maximum (SCM). Furthermore, Parslow (1981) was able to show that this steady state was not an outcome of their simplifying depth-dependent growth assumptions. This type of interaction could contribute to the observed Chla anti-correlation with column stability in two ways. The first is that it could allow an elevated vertical mixing regime to stimulate near surface primary production and at the same time be attenuated in its direct dispersive effect on phytoplankton distribution. The second is a direct biomass dependence on the mixing regime independent of primary production, whereby it increases with decreasing column stability not as a result of elevated production, but rather from the direct ameliorating effect of increased vertical mixing on sedimentation loss.  62  3. Short Time Scale Primary Production Processes 3.1 Introduction Evaluation of mixing-primary production coupling suggested to be important through statistical inference in chapter 2 requires an analysis of the nutrient and light status of the phytoplankton and its dependence upon the surface layer mixing regime. Historical data on ambient nutrients and photosynthesis vs light characteristics (P-I) of the phytoplankton and its dependence on physical conditions are either not a component of, or not readily extractable from, either the I C M data set, Sullivan (1979b), or Stevens and Silbert (1976). In light of this, a number of field studies were undertaken in 1982 and 1983 designed to provide this information as well as to generally better characterize basin primary production on time and vertical space scales inferred from the analysis of the ICM data set to be more appropriate in discerning the nature of underlying processes.  3.2 Specific Survey Objectives A qualitative survey of Holberg Inlet in 1981 indicated the frequent presence of a subsurface chlorophyll maximum (SCM) at 2-4 m depth in Holberg Inlet. This coincides approximately with the depth of maximum covariance between column stability and carbon uptake in the I C M data set (Sec. 2.2.5.3). Hence it was concluded that understanding the processes underlying this feature is germaine to a more general understanding of the coupling between the vertical mixing and primary production and hence was selected as the primary focus of the surveys. There were two considerations in designing the appropriate sampling scheme: 1) Cullen (1982) emphasized the importance of measuring the small scale vertical structure in SCM if the aim is to glean the mechanisms of formation and maintenance. Hence the vertical sampling protocol was tailored to maximize this aspect within practical limits. 2) The chapter 2 statistical analysis of the I C M data set indicated a general lack of stable horizontal structure (albiet on the ICM sampling time scale) in either phytoplankton biomass or carbon uptake, while at the same time a scaling argument relating mixing and phytoplankton growth time scales (Sec. 2.2.4.3) suggested that up 63  to a time period of several days, coastal water masses on the scale of Holberg-Rupert basin respond to external forcing in a spatially non-uniform manner (Lewis and Piatt 1982). Therefore it was felt that the basin could only be adequately sampled efficiently by adopting a horizontal Lagrahgian (i.e. quasi-Lagrangian) sample reference frame, wherein a surface parcel of water is marked with a drift buoy and the time evolution of pertinent variables measured. On the time scale of a few days and to the degree to which the drift station is truely quasi-Lagrangian, this evolution is attributable mostly to vertical and local processes. Furthermore, given the lack of permanent horizontal structure, this evolution is potentially representative of short time scale dynamics in the basin generally. Incidentally this sampling approach also provided a verification data set of the correct type for the one dimension (i.e. z,t) simulation modelling in chapter 4. To test the validity of the horizontal homogeneity assumption a synoptic survey of Chlo specific biomass, salinity, and temperature concurrent with the drift stations was carried out at fixed stations along the length of Holberg Inlet during the July/83 cruise. 3.3 Methodology 3.3. J Sampling Periods The field program consisted of 3 main sampling periods: May 10/82 - May 20/82, Aug 9/82 - Aug 18/82, and July 12/83 - July 20/83. Sampling during these cruises was confined to Holberg Inlet as it and Rupert Inlet are somewhat symmetrical with respect to QN. Therefore, each was presumed to respond similarly to the physical forcing associated with Q N .  3.3.2 Drift Stations The radar reflector equipped drifter used to mark the water parcel consisted of a 2 m by 5 m canvas sail of the window blind type described by Buckley (1977). This design is noted for its ease of deployment under adverse conditions and its ability to align itself 90° ± 10° to the current within a distance of « 10 m. The sail could be positioned at depths between 1 m and 20 m. 64  3.3.3 Chemical/'Biological Variables Measurement Procedures Chla estimates were based on « 1 litre diaphram pump samples vacuum filtered ( < 15 mm Hg) onto either Whatman 984AH or G F / C (43 mm) glass-fibre filters. Filters were stored frozen over desiccant until analysis ( » 2 wks), with the filtrate stored in Whirlpac bags for nutrient analysis. Acetone extraction of chlorophyll and phaeopigment correction followed the procedures of Strickland and Parsons (1972). In 1982 determinations were done spectrophotometrically using the equations of Jeffery and Humphrey (1975), while in 1983 determinations were done with i n vitrofluorescenceusing a Turner Designs 10-000R filter fluorometer calibrated with pure Chla . To ensure that no major features in the vertical Chla structure were inadvertently missed by the discrete depth sampling protocol, continuous vertical in vivo fluorescence traces were also recorded spanning the sampling domain, using a Turner 111 filter fluorometer. Specifically, the inlet of a 2 cm flow-through fluorometer door was connected to the outlet from a small volume diaphram pump (i.e. 5 litres/min) located « 0.5 m above the waterline. The pump-hose intake was raised or lower by hydrowire winch at ea 4 m/min, with the depths indicated on the meter wheel noted as a function of time. The depth vs. time information was then used to convert the strip chart fluorescence vs. time record to fluorescence vs. depth, correcting for the hose/pump residence time (as well as first-order smearing for the July/83 cruise data) according to Anderson and Okubo (1982). Admittedly, it appeared that this simple recording approach was in the end susceptible to a systematic translation error. This was indicated by the lack of depth correspondence between discretely sampled (i.e. pump-hose intake held at a fixed depth for several minutes) Chla features and features in the continuous fluorescence profiles. However, the effect of internal wave displacement cannot be totally discounted (Sec. 2.1.2.2). In any case, the discrepancy did not interfere with the primary function of the fluorescence trace which was to ensure that the vertical structure was adequately sampled. In 1982 the only nutrient measured was N O J , using the manual cadmium reduction column procedure outlined by Strickland and Parsons (1972). This analysis was carried 65  out on refrigerated samples collected within the previous 10 h. In 1983 N O J , NH^ , P O f , and S i O j  4  were measured using an Autoanalyzer (Strickland and Parsons 1972).  In this case, samples were stored frozen for « 3 wks prior to analysis. Size-frequency distribution of particulates of refrigerated samples was determined with a model B Coulter counter from 1 to 10 h after collection. Either a 100 /*m or 200 /xm aperture tube was used. Carbon uptake rate determination followed the procedure outlined in Sec. 2.2.3.1, with dark carbon assimilation measured at 0 m and 14 m in May/82 and Aug/82, as well as the depth of the daily P-I experiment in May/82. In July/83 a dark carbon uptake estimate was made at each incubation depth. A filter blank was also estimated as part of the May/82 P-I experiment. Since this blank is among other things a function of the phytoplankton concentration of the sample, a linear regression was fit to the Chla vs. blank count data. This function was used to correct all May/82 dark bottle counts based on Chla concentration, with the subsequent light bottle count correction done using either the blank corrected dark count or the blank count, whichever was greater. As in the case of the I C M data set, the dark count was assumed to estimate respiration and  1 4  C -uptake  artifacts. The P-I experiment was used to estimate light curve parameters at a single depth, varying between 3 m and 6 m. Aliquots from a 6 1 sample were incubated in duplicate in flowing seawater (pumped from 0.5 m depth into a clear Plexiglas tank on board deck) under irradiances ranging from 100% to » 1% I , with the radio-carbon procedure duplis  cating that used in the in situ carbon uptake measurements. Since it was assumed that net photosynthetic uptake was being measured by the  1 4  C technique, a respiration parameter  was not incorporated in the associated P-I model function. A photo-inhibitory parameter was included however, since one of the suggested possible mechanisms underlying the covariance between column stability and carbon uptake invokes the photo-inhibitory supersaturating irradiance effect. The particular three parameter model fit to the P-I data 66  follows Piatt et al. (1980)  where P  m  is the theoretical maximum Chla specific carbon uptake rate in the absence  of photo-inhibition, a is the initial slope of the light dependent uptake rate, (3 is the photo-inhibition parameter, and I is PAR irradiance (see below). The fitting scheme was p  a grid-step approach that allowed estimation of the 95% confidence limits according to Silvert (1979). To estimate ambient available CO2, total alkalinity was determined according to Strickland and Parsons (1972) in 1982, while in 1983 a linear least squares fit of alkalinity (A) to salinity (5) was used to make the estimation (i.e. A = 0.063-5 + 0.183). The necessary salinity and total alkalinity data (n=776) was obtained from the I C M data set. Incident photosynthetically active radiation (PAR) was measured with a cosine corrected 2JT pyrheliometer probe equipped LiCor LI185B radiometer and integrated over 1 4  C incubation intervals by digitizing the acquired analog chart record on an electronic dig-  itizing tablet. The broadband response of this sensor was converted to PAR (i.e. 400-700 nm) using the conversion factor recommended in LiCor publication 8004-03 (I = 0.37-/*). p  Irradiance at depth was measured either with an LI185B (equipped with a cosine corrected 2n PAR quantum sensor) or a Biospherical Instruments MER1000 PAR integrating spectroradiometer (cosine corrected 2JT sensor). Total daily PAR was obtained with the Belford broadband pyrheliometer (cosine corrected 2TT sensor) located in Rupert Inlet (Figure 2.8). Since the frequency response of the Belford pyrheliometer was not available, it was assumed that the LiCor broadband to PAR conversion factor was a reasonable approximation for converting daily irradiance to daily PAR.  67  3.3.4 Physical Variables Measurement Procedure Sedimentation flux of particulates was estimated on board deck in 1983 using the SETCOL technique of Bienfang (1981) on volume specific seston. An effort was also made to measure Chla specific and phaeopigment specific fluxes. However, the small volumes ( » 25 ml) afforded by the SETCOL apparatus used, precluded measurements of sufficient accuracy to enable useful estimates. In addition, the assumption of a comparatively small positively buoyant fraction required by the procedure was usually not satisfied (see Sec. 3.4.3.5). Hence the coupled equations for positively and negatively buoyant fluxes were also evaluated iteratively. In this way the bias in the sinking rate estimate due to a nontrivial positively buoyant particle flux was minimized. However, since the convergence also resulted in overestimates of both fluxes, these were reported along with the once-through calculation to provide approximate lower and upper bounds respectively for the sinking and floating particle settling velocities. Wind velocity was recorded with a NOAA W102-P Sky vane I anemometer located at A (Figure 3.15) » 2-4 m above mean sea level. Current shear was estimated in 1982 by attaching an InterOcean model 135 internally recording current meter (with its own surface float) to the drift marker. Since the marker was effectively anchored to the 2-4 m stratum, the velocity profile relative to this stratum was measured as the current meter was raised and lowered to discrete depths at « 20 min intervals. In 1983, in place of the current meter, two window blind drogues, one anchored at 1 m and the other at 20 m were deployed near the 3 m drift marker and their rates of separation over time (as measured with ship's radar) used to calculate a mean 0-20 m current shear. Temperature and salinity profiles were determined in May/82 using a Bissett-Berman STD, in Aug/82 using an Autolab model 602 Wheatstone bridge temperature-salinity probe, and in July/83 using a Guildline model 8705 C T D .  3.4 Results and Discussion 68  3.4.1 May/82 Cruise S.4-1.1 Wind effect on drift buoy displacement and current shear The drogue track for the May/82 cruise is presented in Figure 3.15, while map positions in relation to drift station sampling times are presented in Tabie ILLl.  It should be  noted that drogue position was only estimated in day light during both the May/82 and Aug/82 cruises. Hence the plotted tracks include net estimates of the actual nocturnal displacement, given the oscillating tidal excursion as well as the effects of shifting wind direction. These interpolated tracks are represented as a dashed line in Figure 3.15. Table III. 1 May/82 drift station dates and positions tor Figure 3.15. Sample Date  Sample Time(h)  Time series  Position  May 11 12 13 16 17 18 19  0700-1545 0850-1420 0945-1230 1245-1740 1215-1640 1240-1815 1315-1540  Ml Ml Ml M2 M2 M3 -  1-4 5-8 9-14 22-27 28-32 34-38 39-44  Due to the narrow basin width, continuous tracks were limited to 2-3 days in duration before the drogue beached in shallow water. In total, one 3-day (Ml) and two 2-day (M2 and M3) time series were obtained. M l (May 11-13, position 1-14 in Figure 3.15) began in mid-channel » 12 km east of QN and travelled a net distance of w 3 km eastward over the 3 days. The wind field during M l was dominated by NW winds averaging w 3-4 m/s, and ranging up to 6-7 m/s (Figure 3.16). Over this time interval, the net drogue travel was essentially from west to east, consistent with the expected momentum transfer from the surface wind shear to the surface layer. The wind during M2 (position 22-32) was the most energetic during the May/82 cruise, particularly on May 16 when the predominant SSE winds were up to 10 m/s. Inspite of these prevailing winds, however, the drogue travelled mainly to the southwest (position 22-27), suggesting that possible physiographic effects associated with the compass orientation of the Coal Harbour-QN bend in Holberg Inlet modified the SSE 69  Figure 3.15 May/82 drogue track. Position numbers correspond to equivalents in Table 3.1. Dashed lines indicate over night interpolations. (A) denotes wind anemometer location.  Table 3.1. Dashed lines indicate over night interpolations. (A) denotes wind generated surface current. Consistent with this, when the winds were light and from the SW on May 17, surface currents were essentially parallel to the wind stress (position 28-32). During M3 (position 34-44), the wind changed from predominantly the NW on May 18 to variable winds from the S on May 19. Surface currents appeared to be correlated with the wind shear on both days.  Tabie III.2 tabulates the relative current profile for the May/82 cruise. The depth specific relative currents measured were quasi-synoptic since a 15-20 min sampling interval was required at each depth to obtain a sufficiently long record (i.e. total profiling time of » 100 min). This could somewhat confound inferences relating the relative current shear to the wind and tide. However a potentially more serious problem concerns the type of current meter used in the profiling. The InterOcean current meter used is not of the vector averaging type, making it a relatively poor choice for measuring currents in the wave zone (Saunders 1976). To make matters worse this particular current meter is characterized by a very long response time difference between the slowly responding 0.5 m x 0.5 m directional vane and the much faster responding savonious speed rotor. Wave action in the Holberg-Rupert basin is usually confined to amplitudes somewhat < 20 cm and periods < 1 s, hopefully reducing this noise corruption to an acceptable level in most cases. However, without detailed information on the frequency response function of rotor and vane, the effect of this wave regime cannot be adequately evaluated. Consequently it was assumed that sufficiently unbiased orthogonal components could not be calculated from the direction and speed data available. Nevertheless, the calculated velocities are presented though only as a qualitative indicator augmenting the directional information (i.e. sign), which is itself free from the effects of the contrasting sensor responses and except for periods of tidal reversal immune to aliasing as well. In regard to the latter, Tabie III.2 also tabulates tidal state, showing that though tide reversal did occur during profiling in M l , it occurred at depths other than where most of the shear was evident. Combining this with a lack of significant shear at these depths (i.e. 10-20 m) for the 71  A  NORTH  MAY 9  I  MRY 10  '  MRY 11  '  MRY 12  '  MAY 13  I I I I | I I I I I jI I I I I | I I I I I I I I I I I | I T I I I J I I I I I I I I I I I I I I I I I | I I I I I J I I I I I | I I I I I I I I I ''| I I I I I J I 'I I I | I  MRY 19  Figure 3.10  I  I1RY 20  May/82 surface wind record.  I  MRY 21  I  MRY 22  '  MAY 23  profiles where this aliasing was not a problem (i.e. May 16-18), leads to the inference that this non-synopticity in the data record is not a serious hinderance to interpretation. There is one final point regarding the interpretation of these data. Most of the current speeds measured during the May/82 cruise were near the 2.5 cm/s detection limit of the instrument. Subsequent digitizing of the analog records and averaging over the sampling interval often resulted in mean speeds below this detection limit. Since these interval averages are underestimates due to the detection threshold cutoff, it was felt that to treat values below 2.5 cm/s as spurious was too severe. At the same time, however, it was important to ensure a conservative current estimate, since a minimum current is required for accurate current direction detection by the directional sensor. As a consequence of these contrasting requirements, an ad hoc minimum speed threshold of 1.0 cm/s was chosen that hopefully maximized the available information in the current meter record. Current shear during M l (predominantly the 17-component) was mostly confined to the 1-5 m depth interval (Table IH.2) corresponding to the thermocline/halocline (Figure 3.17). For M2 the dominant shear again spanned this 1-5 m depth range, but was of a greater magnitude with U and V-component shears on May 16 of « 0.02 s  - 1  compared to 0.01 s  - 1  for M l . Though the drogue travel was generally not correlated with the SSE wind shear on May 16, the corresponding lmU  and V component velocities were consistent with the  observed drogue movement in relation to the wind shear. Since the drogue was anchored from 2-4 m, this requires that the wind generated surface current was approximately confined to the top 1 or 2 m of the water column over the time interval of the May 16 drift station (Tabie III.l). The May 16 temperature profile was inconsistent with this interpretation as it showed evidence of mixing down to 5 m (Figure 3.18). However, the STD profile was taken at the beginning of the May 16 sampling period (Table IH.l), shortly after a 3 h SSE wind burst (10 m/s compared to the mean of » 5 m/s during the station occupation, Figure 3.16), and could therefore reflect these somewhat more severe wind shear conditions. On May 17, when the prevailing winds were from the SW, and the drogue track was parallel the wind shear, the 1-5 m shear was again reduced in 73  Figure 3.17 M l salinity (dash) and temperature (line) profiles.  Figure 3.18  M2 salinity (dash) and temperature (line) profiles.  T a b l e I I I . 2 May/82 Vertical current proBle. U (east-west) and V (portb-soutb) components are pre wben current speed exceeded 1.0 cm/s. + and - denote Bood and ebb tide respectively. Date  Depth (m)  Tide  1 5 10 15 20 1 5 10 15 20 1 5 10 15 20 1 5 10 15 20 1 5 10 15 20 1 5 10 15 20  •+ + + + + -  M a y 11  M a y 12  M a y 13  M a y 16  M a y 17  M a y 18  + +  --  + + + + +  -  -  --  -  Speed (cm/s) 2.8 1.4 1.1 0.7 0.2 1.6 1.6 0.3 2.0  -  2.9 0.2 0.4 0.5 1.6 10.2 4.2 2.4 1.9 1.0 4.8 2.5 2.9 3.8 2.2 7.5 3.6 3.1 1.8 1.0  U  V  (cm/s)  (cm/s  2.1 1.3  -1.8 -0.3  -  -  --  1.6 -1.6  -0.2 -0.4  -1.7  -1.0  2.8 -  -  -1.6 -5.9 -4.2 2.2 1.9 0.4 -3.0 -2.4 -2.6 -3.7 -2.2 7.3 -2.0 -3.1 1.7 -1.0  -  -0.7  -  0.0 8.4 -0.5 -0.2 0.2 0.9 -3.8 0.8 1.3 0.8 0.0 -1.8 -2.9 0.2 -0.6 0.0  magnitude (0.01 s ) . On the only day the shear was estimated during M3 (i.e. May 18), a strong - 1  shear of 0.02 s  - 1  was observed over the 1-5 m interval, while the wind was essentially NW  3-5 m/s. As on May 16, since the drogue remained essentially stationary (Figure 3.15), this shear appears to have been confined to the top 1-2 m. Unlike May 16, the corresponding May 18 STD profile does not argue against this interpretation of the shear distribution, as it showed evidence of a mixed layer that was also, in a consistent fashion, confined to the top few meters of the surface layer (Figure 3.19). The relative current profile on May 18 also showed evidence of a 2-layer flow since the 5-20 m counterflow was essentially uniform with depth (Table D1.2). Based on the relative current shear and salinity / temperature distribution during May/82, it appears that if the 2-layer flow suggested in chapter 2 was present during 76  Figure 3.19 M3 salinity (dash) and temperature (line) profiles.  May/82 it was restricted to regions of Holberg Inlet east of Coal Harbour (i.e. proximal to QN) and/or the associated vertical exchange was small compared to the buoyancy flux and wind generated mixing effects (both acting to impose structure on an otherwise homogeneous turbulent surface layer). On the other hand, as discussed in Sec. 2.1.2.2 it is also possible that the small phase shifts measured between the barotropic tide and observed currents were in large measure an artifact of tidal prediction error. In either case, however it can be concluded that during the May/82 cruise period in the basin generally, surface mixing associated with 2-layer tidal flow was of minor importance compared to surface generated effects. Mean surface wind speed and precipitation were » 3 m/s and » 1.4 mm/d respectively (ICM 1983), and helps characterize surface conditions under which this mixing regime dominance occurred. However, regional differences in the strength of the 2-layer flow were present. The May 18 drift stn, which came within w 2 km of the long term current meter mooring C (Figure 2.2) analyzed in chapter 2, showed some evidence of a 2-layer flow regime (Tabie HI.2).  S.4-1.2 Ml primary production time series Surface irradiance during the May/82 cruise (Tabie ni.3) was generally « 60% of the climatological mean for May as estimated by <I^> in the ICM data set (Appendix J). Over the course of M l , rainfall averaged w 3 mm/d (ICM 1983). Hence the relative constancy of the salinity profile (Figure 3.17) can be interpreted to indicate an absence of significant admixture of water masses through horizontal advection/diffusion and/or an absence of significant drift buoy slippage, since the concurrent buoyancy flux which would otherwise mask such effects, was relatively small. Unlike salinity, the temperature profile time series, which is potentially a more sensitive indicator of water-mass mixing since it makes only a small contribution to density, does not reflect this relative lack of horizontal mixing (i.e. varying w 1°C over the 3 days of M l ) . However, this discrepancy is inconclusive since the reliability of temperature as such an indicator is limited near the surface by its non-conservative nature there. 78  CHLfl 0.0 I 0  REL.  -a  (MG/MXX3) 1.1 I  2.3 I  FLUORESCENCE 25  50  CHLR 0.0 I 0  REL.  L  22.37  SIGMA-T  21.23  (MG/MXX3)  CHLfl  1.5 I  3.0 I  0.0 I  25  50 Lo  0  FLUORESCENCE J  22.17  23.11  REL.  21.17  SIGMA-T  (MG/MXX3) 0.8 I  1.7 I  25  50  FLUORESCENCE  22.19  SIGMA-T  Figure 3.20 M l Chla (squares) and in vivofluorescence(line) in relation to a (dots). t  23.21  Figure 3.21 M l Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to at (circles).  Figure 3.22 at (dots).  Ml P  B  (diamonds ± S.D.) and N O 3  (squares) profiles in relation to  \  The pycnocline gradient was w 0.15 at / m during M l , with little evidence of a defined surface mixed layer. The vertical distributions of Chla and fluorescence in relation to the at distribution are presented in Figure 3.20, Chla and carbon uptake in relation to a in t  Figure 3.21, and the vertical distributions of N O J and P  B  in relation to a in Figure 3.22. t  M l was characterized by vertically uniform Chla and fluorescence distribution inspite of the distinct pycnocline, with carbon uptake on both a total and Chla specific basis decreasing monotonically with depth. N O J remained > 9 / i M throughout, while 0-6 m integrated carbon uptake was low ranging from 6 m g C / m h on May 12 to 13 m g C / m h on May 13 2  2  (Tabie III.3). Corresponding incident PAR, averaged over the incubation interval ( < / £ >), ranged from 309 to 750 /iE/m -s, with total daily PAR ( £ / £ ) ranging from 21 to 43 E / m 2  2  (Tabie III.3). Based on the calculation procedure outlined in Appendix D, carbon uptake during M l ranged from 51 m g C / m d on May 11 to 208 m g C / m d on May 13, while 2  2  Chla (0-6 m integration) increased from 5.3 mg/m to 6.8 mg/m and the 0-6 m depth 2  2  averaged ambient N O J concentration decreased from 10.7 to 10.4 fiM over the course of the time series (Tabie IJJ.3). Phaeopigments remained consistently below the detection limit (Appendix E).  Table III.3 Production variables for May/82 crube: including the mean PAR irradiance <Ig > inciden over the course of C inritv.and P-I incubations, daily PAR (E^s ), the sample collection depth Z i fo the P-I experiment, the depth of estimated saturating light intensity Z » (200 uE/m -s) based on </£ 0-6 m integrated hourly (Ph) and daily (Pi) carbon uptake (mgC/m ), 0-6 m depth averaged ambien N O 7 (tiM), Chla speciBc biomass S (mgChla /m ), and the 0-6 m (May 11-13) or 0-7 m (May 15-16, May 18) depth averaged productivity index PI (mgC/mgChla b). i4  t a m p  7  s  2  B  Date May May May Mav May May May  <I's > 11 12 13 16 17 18 19  309 750 393 322 530 952  -  Zt»mple 21 43  28 39 33 51  4 4 4 6 6 6  -  7  Ztat Ph 2.2 6.6 3.4 2.3 5.6 7.4 -  6 13 13 39 84 39  -  Pi  N0 "  51 148 208 1185 1160 425 -  10.5 10.7 10.3 9.1 4.9 5.9 4.6  82  3  S  PI  10.3 9.8 8.2  1.7 1.3 1.6 1.8 1.7 1.2  B  21.1  49.5 25.9 35.2  -  t  e  9.+.1.8 MS primary production time series M2 (position 22-34, Figure 3.15) began at approximately the same location in the basin as M l , with evidence that surface mixing from SSE winds (Figure 3.16) resulted in a relatively homogeneous near surface vertical temperature distribution, particularly on May 16 (Figure 3.18). Furthermore, average salinity in the surface layer decreased significantly over the 2 days (Figure 3.18), inspite of only 0.4 mm of precipitation over this period (ICM 1983). Since this interpretation is inconsistent with the quasi-Lagrangian time series assumption, significant drogue slippage and/or horizontal exchange associated with the large surface layer shear on May 16 must have occurred. Nevertheless, given its unique vertical phytoplankton structure (see below), this time series is useful in the general characterization of basin primary production. The Chlo , fluorescence, carbon uptake, P  B  , and N O J profiles in relation to a are t  plotted in Figure 3.23 to Figure 3.25 respectively. The a profile remained essentially unt  changed from M l (i.e. « 0.15 a /m). However, Chla and fluorescence were not uniformly t  distributed with depth, showing a monotonic decrease on May 16 and pronounced subsurface maxima on May 17. On the other hand, the distribution of P  B  was similar to M l ,  essentially decreasing in a monotonic fashion with depth. N O J remained > 4 /xM except at 6 m on May 17 coincident with the SCM, where it declined to < 1 pM . Daily carbon uptake rate according to Appendix D was 1185 mg/m -d on May 16 and 1160 mg/m -d 2  2  on May 17 (Tabie III.3). Corresponding £ i | was 39 and 33 E / m respectively and the 2  0-7 m integrated Chla increased from 11.5 mg/m on May 16 to 38.6 mg/m on May 17. 2  2  8.4.I.4 MS primary production time series The M3 time series is of particular interest since it represents the only situation of near NO J depletion at the surface layer during the May/82 cruise. Both the temperature and salinity profiles for M3 are consistent with a quasi-Lagrangian interpretation of the time series (Figure 3.19). Chla , fluorescence, carbon uptake, P  B  , and N O 3  profiles in  relation to the at profiles are plotted in Figure 3.26 to 3.28. The pycnocline gradient was 83  CHLfl 0.0 I  REL.  0  (MG/MXX3)'  1.9 3.9 1 I FLUORESCENCE 38 75  (MG/MXX3)  CHLfl 0.0 ; 10.0 20.0 i : 1 i R E L . FLUORESCENCE 0 38 75  oe  21.36  22.3 SIGMA-T  23.23  20.38  21.87 SIGMfl-T  Figure 3.23 M2 Chla (squares) and in vivofluorescenceprofiles in relation to <r (dots). t  23.37  UPTAKE 0.0 I 0.0 oj  CHLA  (MGC/M3/HR)  UPTAKE  5.2 I  10.5 I  2.0 I  4.0 Lo  (MG/M3)  0.0 I  21.0 I  CHLA 0.0 o.i  (MGC/M3/HR)  UPTAKE  42.0 I  (MG/M3) 10.0 l  0.0 I  5.2 I  CHLA 20.0 l o  0.0  (MGC/M3/HR)  (MG/M3) 4.5  10.5 I  9.0  00  21.12  22.28  SIGMA-T  23.44  20.13  21.86  SIGMA-T  23.58  19.14  Figure 3.24 M2 Chla (squares) and carbon uptake (dots ± S . D . ) profiles in relation to at (circles).  21.26  SIGMA-T  23.38  Figure 3.25 <r (dots). t  M2 P  B  (diamonds ±S.D.) and NO  (squares) profiles in relation to  Figure 3.26 at (dots).  M3 Chlo (squares) and in vivo fluorescence (line) profiles in relation to  Figure 3.28 M3 P  B  (dots ±S.D.) and N O J (squares) profiles in relation to a (c'rcles). t  somewhat steeper during M3 ( » 0.35 o% /m) than during M l or M2. The 0-7 m integrated Chla increased from 24.9 to 33.5 mg/m over the 2 days, while the daily carbon uptake 3  rate on May 18 was 469 m g C / m d (not measured on May 19). Both fluorescence and 2  Chla demonstrate a subsurface maxima on May 18, with onlyfluorescenceindicating this phytoplankton distribution on May 19. On both days the 0-2 m surface layer was almost depleted of N O 3  ( < 0.4 pM. ). Correspondingly, the maximum in pB was subsurface for  the first time, though total carbon uptake on May 18 continued to decrease monotonically with depth. 8.4-1.5 P-I parameter estimates Parameter estimates for M l , M2, and the single May 18 P-I incubation during M3; as well as their approximate 95% confidence limits are tabulated in Table III.4, while the derived adaptation parameters Is = Pjp/  a  along with  Is >  >«  = Z  s a n  , i p  e  —  znd If, = P™//? are presented in Tabie 111.5 Z t, s a  and estimated P-I sample in situ carbon  specific growth rate /x. The specific growth rate calculation assumed a nutrient replete surface C/Chla value of 20. Since this approaches the lower limit of C/Chla in phytoplankton, these estimates of p. constitute an upper limit. Is is analogous to the Tailing parameter Ik and is an index of low light adaptation, while If, is a comparable index for high light adaptation (Piatt et al. 1980). The fit of the model curves to the data points are presented in Appendix F. Best estimates for P™ (mgC/mgChla h), a and /5 (mgC/mgChla h / x E m s ) ranged 2  from 2.8-8.0, 0.02-0.10, and 0.005-0.010 respectively during the May/82 cruise. The large confidence limits associated with these estimates makes a discussion of their timedependency somewhat speculative. With this caveat in mind, the only discernable timedependent feature among the P-I parameter estimates was a persistent increase in P  m  dur-  ing M l . However, this is not associated with a monotonic increase in daily surface irradiance, since the highest flux was recorded on May 12 (Tabie HI.5). It is unlikely that a lag effect in the adaptation time course to high light could be responsible 89  Table III.4 Least-squares estimates ot P™, a, and /? and their approximate 95% CI tor the P-I incubation experiments. pm  Date  Time series  est.  95% CI  May 11  Ml  3.3  May 12  Ml  4.6  May 13  Ml  5.5  May 16  M2  2.8  May 17  M2  5.4  May 18  M3  8.0  1.5 14.0 1.7 14.3 2.8 16.1 2.9 14.8 3.7 16.3 4.4 17.0  a est.  95% CI  est.  95% CI  0.01 0.05 0.01 0.20 0.05 0.24 0.03 0.05 0.04 0.24 0.03 0.06  0.008  0.000 0.060 0.000 0.060 0.000 0.080 0.000 0.080 0.000 0.080 0.000 0.056  0.02 0.02 0.10 0.04 0.09 0.05  for the monotonic increase in P  0.005 0.006 0.005 0.010 0.015  , since it would be inconsistent with the lack of such an  m  effect on May 11, which also followed a period of high irradiance (TaWeIII.5). Furthermore, it can be argued that it is unjustified to rationalize observed shifts in these light curve parameters in terms of the previous day light history, since they have been found to vary in magnitude on a diel basis by a factor of > 4 (Piatt et al. 1980, Harding et al. 1982).  Table III.5 Adaptation index to low light (I ) and high light (It) tor the P-I incubation experiments, well as . <I >- * = Sample - Z. , and /i. s  P  z  S  Date  11 12 13 16 17 18  z (m)  (E/m d)  <Is > (pE/ra's)  45 21 43 28 39 33 51  .  .  309 750 393 329 530 810  1.8 -2.6 0.6 3.7 0.4 -1.4  2  M a y 10  al  d  The periodic fluctuations in P  m  tt (h"») _  0.06 0.20 0.19 0.06 0.12 0.07  Is (*iE/m»s)  . 165 230 55 70 60 160  h (f«E/m*-s) _  413 920 917 560 540 530  and a tend to be in phase and of similar magnitude  (Harding et al. 1982, Cote and Piatt 1983). Hence, fluctuations in the derived parameters Is and lb are less pronounced, making them more suitable indices of true photo-adaptation (Piatt et al. 1980). The pattern in these indices was a tendency to co-vary with the sign of Zd- This covariance was strongest with Is- If Zd was positive and hence on the basis of < / | > the light field at the P-I sample depth was subsaturating, then Is was small indicative of a  shade adapted population. Conversely, if Z& was negative, Is was large indicative of a sun adapted population (Piatt et al. 1980). The exception to this pattern was May 11, when though Zd > 0, Is was large. There are two possible explanations for this discrepancy. The first concerns the quality of the P-I data. Though data scatter is high for most of the P-I incubations, the scatter at the lower light levels on May 11 is particularly large (Appendix F). It seems reasonable to expect that this adds significantly to the uncertainty of the estimate of a, though it is not clear why this is not reflected in the corresponding estimates of the 95% CI (Table JJI.4). It is speculated that this may be due to the general linearity of the P-I relationship on May 11 perhaps in a relative sense constraining the least squares estimation of the simultaneously calculated 3-dimensional confidence limit space. The second possibility concerns the growth rate of the P-I sample population on May 11. Assuming a C/Chlo of 20, the carbon specific growth rate was « 0.06 h  - 1  . This  comparatively low growth rate (Tabie IU.5) may have significantly increased the photoadaptation response time to the large change in the light flux between May 10 and May 11. However inconsistent with this, Prezelin and Matlick (1980) found that the dinoflagellate Glenodinium sp. adapts on going from high (2500 w/m ) to low (500 w/m ) light in ca 2  2  0.1 of a generation time, which if applied to the May 11 P-I population is equal to a few hours. On the other hand much of the initial (i.e. rapid) response by Glenodinium sp. is apparently mediated by a rapid increase in the cell peridinin concentration. This auxiliary pigment response may not have a parallel in the diatom-flagellate assemblage of May/82, though a rapid change in cell Chla quota is known to occur in some diatoms (Skeletonema costatum , Riper et al. 1979 and Lauderia borealis, Marra 1980).  The covariance between 7 and Zd was not as strong, with the sign contrary to ex0  pectation on May 13 and 18 (Tabie HI.5). Unlike Is, h conformed to theory on May 11. However, this was undoubtedly an artifact of the truncated range in P-I incubation irradiance levels available on that day (Appendix F). In addition to being times of transition in 1% , May 13 and 18 (along with May 11) were also characterized by low phytoplankton specific growth rates (Tabie HI.5). Thus the unexpected lb — Zd correlation on these dates 91  could possibly be explained as in the case of the Is — Zd correlation on May 11 in terms of growth rate limitation of the photo-adaptation time course. If this is the case, then the fact that Is adjusted according to theory implies that the underlying mechanisms for the two parameters have time scales that are coupled differently to growth rate. A second estimate of depth dependent carbon uptake during the May/82 cruise can be inferred by modelling uptake on the basis of equation (3.1) and these parameter estimates. The success of the predictions is limited largely by the degree to which the euphotic zone phytoplankton can be described as a vertically homogeneous population with common light curve characteristics, which in turn is a measure of the degree to which vertical displacement has modified the near field light flux of the phytoplankton on a relevant time scale (Falkowski 1983). Hence the success of the fit can be used to gauge the depth dependency of the P-I parameters. Due to light meter failure on May 16, an estimate of irradiance distribution with depth required for the light dependent uptake prediction was not always available. However, from May 11-15 the mean 0-20 m water column light attenuation coefficient varied by only ± « 20%, ranging from 0.21 m  _ 1  on May 11 to 0.23 m  - 1  on May 15. Thus the mean of these  extinctions (0.20 m ) was used as a common parameterization of depth dependent light _ 1  flux for all sampling stations. Tabie in.6 tabulates the observed and predicted carbon uptake for the May/82 cruise. The predictions are based on the light curve parameter values tabulated in Tabie ni.4.  Table III.6 May/82 Observed (O) and predicted (P) carbon uptake in (mgC/m b), and % residual (R) relative to O. Predited uptake based on P-I parameters tabulated in Table IIIA. Z, npU of the P-I incu sample in parentheses. 7  aT  Time series Ml Depth May 11 (4) O P R (m) 0 4 5 6  1.4 1.3 -7% 0.8 0.8 0%  Ml May 12 (4) O P R  Ml May 13 (4) O P R  M2 May 16 (6) O P R  1.3 1.2 -8% 1.8 1.2 -37%  2.9 2.0 -31% 1.5 1.3 -13%  -  -  8.2 4.2 -49% 9.2 4.8 -48% 6.0 3.1 -48% 7.1 7.6 7% -  _  -  -  -  M2 May 17 (6) O P R  -  M3 May 18 (6) O P R 5. 3 5.9 11%  -  2.8 3.3 1S%  In comparing the observed and predicted estimates it is important to bear in mind the 92  inherent imprecision in the C l4  incubation method, with the relative error particularly  large at the low uptake rates characteristic of M l . From this, it is reasonable to interpret the seemingly random residual pattern during May 11-13 as noise related and hence that the observed and predicted uptake estimates during M l are essentially indistinguishable. For M2 by comparison, there is evidence of a pattern in the residuals, with the prediction consistently underestimating uptake by « 50% at the near surface. The implied depth dependency in P-I properties is consistent with the distribution in Chla and N O J indicating that M2 was a relatively mature water parcel. With M3 the physical evidence suggested a relatively stable water column, yet the pattern in the residuals most closely approximated M l . However, the associated near depletion of the 0-2 m surface layer N O J suggests that this agreement between M l and M3 is hot due to similar depth independent P-I characteristics but rather stems from the near surface nitrogen limited growth during M3 masking a depth dependency in P-I. On the general assumption that a photo-adaptation interpretation of the pattern in residuals is valid and further that this pattern is in steady state, an estimate can be made of the relative change in vertical mixing between M l and M2. This is based on the analysis by Lewis et al. (1984), who scaled the relative importance of photo-adaptation responses which tend to produce depth dependent distributions in P-I properties on the one hand, and vertical mixing which smears the effect on the other. The analysis derived two nondimensional parameter groups that defined this relative dominance: fc,7 and m  where k{ is the light attenuation coefficient, l  m  K /fi l , z  a m  the mixing depth, K the eddy diffusivity z  coefficient, and \i the c-folding time scale for the photo-adaptation response. On May 17 a  surface uptake was underestimated by « 50%. Assuming this reflects a 2-fold difference in the P-I response in the vertical (i.e. over / ) and that this difference was zero during M l , m  K lnJm  18  s e e n  z  *° range from » 0.05 to 1.6 (Figure 3 in Lewis et al. 1984). Assuming  H and lm are constant, this represents a decrease in water column K by a factor of « 30 a  g  from M l to M2. Summarizing within the limits of interpretation given the large confidence limits of the 93  P-I parameter estimates, the pattern in the Is /h — Zd covariance and the vertical distribution of residuals in the carbon uptake predictions indicates a significant evolution in the depth dependency of P-I characteristics over the course of the May/82 cruise period, dependent upon the combined action of the observed range in surface irradiance and column static stability (i.e. mixing). However, the lagged appearance of surface nitrogen depletion tended to counteract the effect of P-I parameter depth dependency with the result that a constant parameter P-I growth model was on the whole relatively successful at predicting depth dependent growth over the May/82 cruise period.  3.4 1-6 Patterns in species composition The species composition information for the May/82 cruise is sparse, with no information related directly to the M l or M2 time series. However, enumeration samples collected at 0 m on May 18, 0 and 3 m on May 19, and 3 m on May 20 indicated 2 dominant species in the assemblage. The diatom Thalassiosira decipiens (vol « 5200 /xm ) and an 3  unidentified photoautotrophic cryptomonad (vol « 700 /xm ) were present in all 4 samples 3  (Appendix G), with T. decipiens the dominant species in 3 of the 4 samples, ranging from a high of 91% of the 3 m sample on May 19 to a low of 66% of the 0 m sample on May 19, while the cryptomonad was dominant in the fourth sample (0 m on May 18), comprising 54% of the sample (compared to 4% for T. decipiens ). Three of the four samples also contained the photosynthetic ciliate Mesodinium rubrum, though its contribution to the total on a volume basis was small in all cases.  S.4-1-7 Inferred grazing pressure Planktonic herbivore distribution was not estimated during the May/82 cruise. The lack of detectable phaeopigments during M l implies low grazing pressure, though the detection threshold is itself limited by the low Chla concentration (Strickland and Parsons 1972). However, phaeopigments were detectable during M2 (Appendix E). The 0-7 m £ P h a e o p i g m e n t / X ) C h l a index indicative of grazing pressure (Lorenzen 1967) was 0.2 94  on May 16 and 0.1 on May 17. Accordingly, this is equivalent to a copepod density in the range of » 5 • 10 copepods/m . Grazer community composition responsible for this 3  3  grazing effect can be estimated from daytime and nighttime zooplankton tows carried out by Island Copper Mine at ICM stations A - D (Figure 2.8). The daytime quantitative tows were done at 5 m, 30 m, 80 or 130 m (depending on station), and vertically throught the entire water column, using Clarke-Bumpus nets (80 cm aperture). The nighttime 2  qualitative tows were done only near the surface (2-5 m) using a large diameter plankton net ( « 8 • 10 cm aperture). The nearest I C M sample date (June/82) was dominated 3  2  numerically during the day by Acartia clausii, copepod nauplii, and copepodites I-II and IJI-V, and nocturnally on a wet weight basis by Calanus marshallae, Euphausid nauplii, A. clausii, and Pseudocalanus minutus (ICM 1983). Assuming the maximum clearance rate for A. clausii is representative of this grazer community (i.e. 3.0 ml/d, Raymont 1983, p. 547), the  Phaeopigment/  Chlo estimated copepod density gives a 24 h grazing  loss of « 0.3 mgChlo / m . This inferred biomass sink is insignificant compared to the 2  biomass deficit of 30 mgChlo / m over the course of M2 (defect based on the estimated 2  daily carbon uptake rate and a C/Chla of 20).  3.4'l-8 Inferences concerning M2 SCM formation As already noted, the anomalous salinity change during M2 makes a Lagrangian interpretation of the pronounced evolution in Chlo distribution somewhat dubious. However, if for the sake of argument M2 were assumed to sufficiently approximate a quasi-Lagrangian time series, the 6 m N O 3  minimum coinciding with the SCM on May 17 would repre-  sent a 10 /xM drop in ambient N O J concentration over 24 h. Such a rate of decline is likely only if the nitrogen demand responsible did not arise solely from in situ growth: The mean incubation irradiance at 6 m was w 180 /xE/m -s on May 17 which is typically 2  near saturating (Chan 1978, Harris 1978). Nevertheless, if in situ growth were the only nitrogen sink as opposed to an allochthanous biomass contribution adding to the nitrogen demand, it would require a Chla specific growth rate at this depth of 2.6 d 95  - 1  . The small-  est phytoplankter observed during the May/82 cruise were Chaetoceros spp. with a cell volume of « 200 jim . According to Mullin et al. (1966), this represents a cell carbon 3  content of « 40 pg, which corresponds to a maximum specific growth rate at 21°C of « 2 d  _ 1  Banse (1976). Assuming a Q  1 0  of 2 and a mean euphotic zone temperature of 12° C,  the maximum possible specific growth rate during May/82 would have been only « 1 d  - 1  .  Hence even a theoretical maximum growth rate is more than a factor of two too small to satisfactorily account for the decline in ambient NO3  on the basis of in situ growth.  Furthermore, even if the necessary growth rate were attained, the strong depth dependent distribution of the Chla specific biomass is not adequately explained. Ambient NO3" was high enough at the surface to be non-limiting over the time series (i.e. > 4 z*M ) and as a consequence nutrient related depth dependent sedimentation and/or in situ growth can be ruled out as possible mechanisms. This leaves only near surface photo-inhibition of carbon uptake (i.e. growth) as a possible cause. < i | > was only moderately higher on May 17 compared to May 16, while total irradiance was actually lower (TaWe ni.5), making a strictly photo-inhibition explanation of the SCM somewhat problematical. This is not to say that photo-inhibitory effects were not important during May/82. The P  B  maximum  on May 18 (i.e. M3) occurred several meters below the nitracline (Figure 3.28), implying that this subsurface peak in P  B  was due not just to N O J depletion in the near surface,  but to a photo-inhibitory effect as well. However, assuming again for sake of argument that the May 17 light field was photo-inhibitory, it remains to be explained how a contrast in depth dependent growth could result that would be sufficient to produce the observed Chla distribution.  Assuming on the other hand that an allochthanous input contributed to the May 17 6 m SCM and hence to the in situ nitrogen demand, then the necessary (vertical) biomass flux is attributable either to a behavioral response by a motile phytoplankter and/or to passive gravitational sedimentation. The required sedimentation rate and importance of a dependence on depth will be deferred until chapter 4. In any case however, it is unlikely that the dominant phytoplankter (T. decipiens) could have contributed significantly to 96  the flux given its characteristic range in sinking rate (i.e. from 1 — 2 m/d, Smayda 1970). A significant depth dependent sinking rate component in a sedimentation flux is equally unlikely due to the high surface N O J level throughout M2 and the small dynamic range in nitrogen dependent sinking rate (Bienfang and Harrison 1984). Hence in the probable absence of appropriate circumstances for the necessary sedimentation flux, it would appear that a behavioral explanation for the SCM, based on a motile phytoplankter, is the most likely explanation. Of the two motile species observed in any numbers, the cryptomond was by far the most important numerically. However, M. rubrum is recognized as a strong vertical migrator (Packard et al. 1978) and given the coarseness of the enumeration sample distribution and the lack of direct species enumeration data for the M2 time series, was potentially also an important contributor to the observed pattern in Chla distribution. Since P  B  was monotonic with depth on May 17, there also remains the matter of  explaining the cause of the vertical migration in light of this distribution in P  B  . < / £ > was  supersaturating on May 17 (Tabie IH.5), and hence one possibility is that the monotonic P  B  observed was an artifact of using glass  1 4  C incubation bottles which selectively filter  out photo-inhibitory UV radiation (Lorenzen 1979). Hence if true in situ optimal growing conditions were in fact subsurface, then the vertical migration could be a response by the phytoplankton that optimizes the availability of light and nutrient (Cullen and Holligan 1981). On the other hand, if the observed P  B  distribution was not an artifact, then its  distribution must be explained in terms of strongly depth dependent P-I characteristics, while as a direct consequence the Chla specific biomass distribution would appear to require a depth dependent respiration rate, though a mechanism for this depth dependence is not known.  ©7  8.4-1-9 Summary The STD and relative current profiles indicated that the 0-20 m surface layer was generally well stratified with relative currents decreasing with depth, indicating an an absence of the 2-layer flow inferred in chapter 2. Assuming that the small phase shifts evaluated in chapter 2 were real, then this general absence of complimentary evidence in May/82 indicates either that the consequences of this flow were masked by wind and buoyancy flux effects and/or it was restricted to regions of Holberg Inlet proximal to QN. Under the prevailing mixing regime vertical phytoplankton structure evolved over the course of the time series, both in the form of light adaptation types and in biomass distribution. Growth was primarily light limited, though near surface photo-inhibition and nutrient limitation also occurred. The lagged appearance of the nutrient limited growing conditions after the appearance of depth dependent P-I characteristics acted to mask the effect on depth dependent growth, thereby improving the performance of a constant parameter P-I model prediction. One SCM was observed during May/82. A behavioral response to high irradiance appeared to be its main mechanism of formation, though the apparent non-Lagrangian nature of the associated time series makes it equivocal. Based on an ambient phaeopigment grazing index, herbivore grazing accounted for only « 1% of the observed Chla mass defect.  3.4.2 Aug/82 Cruise 8.4.2.1 Wind effect on drift buoy displacement and current shear Two overlapping time series ( A l and A2) were obtained during the Aug 9-18/82 cruise by alternating on a daily basis the sampling effort between two drift stations. The marker drogues used were identical to the one deployed on the May/82 cruise, with the anchor again rigged at 2-4 m depth. After daily completion of the drift station sampling program for the day, the alternate drift station was occupied if time permitted and STD and/or fluorescence profiles taken. Otherwise, daily switching back and forth between A l and A2 08  Figure 3.29 Aug./82 drogue track. Position numbers correspond to equivalent in Table 3.7. Dashed lines indicate over night interpolations.  A NORTH  i  ,  •  I  i  I  RUG 9  |  I  AUG 10  1  I  I  I ' " "  A U G 11  1  l ^ k . ^  I  12  AUG  1  ^  ^  ^  ^  8 ,,,,  • 1  •  i  '  1  i 1  i  RUG 1 5  I " ' " ' 1  i  ,  AUG 16  1  17  AUG  5 n/s  1 1 1 1 1 1 1 1 1 1 1  AUG  1111111  18  I  '  1 1 1 1 1 '  [  1 1 1 1 1 1 1 1 1 1 1 1 1  AUG 19  Figure S.SO Aug./82 surface wind record.  1  I  , 1 1 1 ! 1 1 1 1 1 1 1  AUG 2 0  1 1 1 1 1 1  j  1111  AUG 21  • i  I •••••• i 1  |  AUG 2 2  continued until Aug 14 after which sampling effort was concentrated at A2. Only the A l time series will be presented. The drogue track is presented in Figure 3.29, while positions in relation to drift station sampling times are presented in Tabie III.7. A l was deployed on Aug 9 at the western end of the Straggling Is. group (position 1) from which it tracked in a generally eastward direction until beaching on the north shore of Holberg Inlet on Aug 18 (position 18). The prevailing winds during this period were Table III.7 drift sampling station dates in relation to drogue positions in Figure 3.26. Sampling Date  Sampling Time (h)  Position  Aug 9 10 11 12 13 17  0800-1230 1440 0915-1410 0945 1100-1420 1045-1708  1-2 3 4-7 8 9-11 16-17  predominantly from the NNW interspersed with NW and W winds (Figure 3.30). Wind speed did not exceed 6 m/s, though given the westerly component in the wind field, wind generated surface currents presumably contributed to the eastward drogue displacement. Superimposed on this net eastward travel was an east-west oscillation that in the absence of an easterly wind component is attributed to tidal advection, or at least to a combination of surface tidal current and the wind-shear. Figure 3.31 presents the A l salinity and temperature profiles. Aug 9-13 is characterized by a persistent erosion of the halocline/thermocline due to a deepening surface mixed layer though the pycnocline increased from » 0.3 o% / m to 0.4 o~t / m (Figure 3.32). The erosion is inferred from the near surface decreasing and increasing trends in temperature and salinity respectively, consistent with turbulent mixing of underlying colder, saltier water up into the mixed layer. Thus by Aug 13 the profiles were characterized by a distinct thermocline/halocline » 2-3 m in thickness separating two homogeneous layers. A STD profile was not taken again until Aug 17. Hence, the immediate effect of the large Aug 14 displacement (position 11-13) on the salinity/temperature profiles could not be 101  a.o I  10.0  i  i  TEMP ( C ) tie i3.s  32.0 28.0 _l I 13.0 10.0  iBh  i p p  a  J  TEMP ( C ) 11.B 13.3 i_  32.0 28.0 -I I 15.0 10.0  TEMP ( C ) is.3  32.0 28.0 —I I  II.e  l I I  t A  o  flUC 13 DROGUE I  Figure 3.31 Aug./82 (Al) salinity (triangles) and temperature (circles).  TEMP ( C ) It.8 13.3  32.0 28.0 -I I 13.0 10.0  TEMP ( C ) 11.8 13.3 _l U  32.0 _l 19.0  determined. However, the Aug 17 salinity profile closely resembles the one on Aug 13, even though the cummulative precipitation in the basin for Aug 13-17 was only 10 mm (ICM 1983). By comparison, during Aug 9-13, surface salinity increased 1.5 ppt inspite of a 2-fold greater surface fresh-water flux (i.e. cummulative precipitation of 22 mm, ICM 1983). Relative vertical current profiles determined as in May/82 are presented in TaWe HI.8.  Table III.8 Aug/82 vertical current proBle during the Al time series. U (east-west) and V (northcomponents are presented where the current speed exceeded 1.0 cm/s. + and - denotefloodand eb respectively. Date  Depth (m)  Tide  Spd (cm/s)  1 5 10 15 20  +  7.3 4.3 4.0 5.6 5.4  7.0 -4.3 -4.0 -4.8 -4.9  -1.9 0.4 -0.3 3.0 2.3  4.3 3.3 2.8 1.0 0.8  4.1 -3.3 -2.7  1.4 -0. 0.7  10.8 4.6 5.0 5.6 3.7  6.6 -4.6 -4.9 -4.6 -2.9  -8.5 0.3 1.0 3.2 2.3  1 5 10 15 20  A u g 11  1  A u g 13  5 10 15 20  -  -  U (cm/s) V(cm/s)  As in the case of May/82, the estimated shear was predominantly in the U direction and spanned the 1-5 m depth interval with magnitudes ranging from 0.02 s 0.05 s  - 1  - 1  on Aug 11 to  on Aug 13. Also as in May/82, the magnitude and relative direction of the 1  m relative current was approximately correlated with the wind shear, with the strongest surface current (Aug 13) occurring in conjunction with the » 5 m/s NW wind at the time. Unlike May/82 however, there was the suggestion of a 2-layer flow indicated by the apparent depth independent counterflow from 5 to 20 m (TaWe HI.8). It appears then that though confounded somewhat by a greater surface radiant heat flux during Aug 14-17 compared to Aug 9-13 (Table IH.9), the expansion of the thermocline between Aug 13-18 (particularly pronounced between Aug 17-18) reflected a switch 103  from a relatively high to low energy surface mixing regime. A corresponding reduction in wind speed implicates wind stress as a significant source of this mixing energy as in May/82 (Aug/82 cruise period average wind speed < 2 m/s). The run-off regime also declined by a factor of two between the two time periods, though it remained seasonally nominal for both periods (compare with Figure 2.4). This concommitent decrease in surface buoyancy flux is inconsistent with the apparent reduction in mixing between the two periods, indicating either that the corresponding decline in the wind shear more than compensated for the reduction in buoyancy flux and/or that the decline in run-off affected the relative buoyancy of the tidal jet in a significant way. The contribution of a run-off induced positively buoyant tidal exchange to the elevated Aug 9-13 mixing regime hinges on the equivocal relative summer time dominance of positively and negatively buoyant tidal flow under moderate run-off conditions (i.e. precipitation « 2 mm/d). On the other hand, the Aug 9-13 relative current profiles and the STD profiles after Aug 13 were consistent with the presence of a 2-layer flow (i.e. well mixed surface layer depth independent 5-20 m counterflow respectively ), though the associated entrainment interface associated with such a flow would be located at only 1 — 5 m depth. According to the Fronde number argument in chapter 2, this required a density difference between the jet and the basin surface layer of w 0.09 ppt. This is comparable to the density difference present in the high positive buoyancy regime current record analyzed in chapter 2 (i.e. Figure 2.3). However, the precipitation rate (i.e. run-off) was a factor of 5 greater for this latter period.  8.4.2.2 Primary production time series On the basis of the I C M <Ig > for Aug (Appendix J), surface irradiance during A l was nominal. Figure 3.32 presents the Chla ,fluorescence,and a profiles, Figure 3.33 the t  vertical distribution of carbon uptake rate in relation to Chla and a , and Figure 3.34 t  the P  B  and N O J profiles in relation to tr . On Aug 9, ambient N O 3 t  surface (> 4 pM ) with P  B  was high at the  maximal there, indicative of light limited growth at depth.  There was the suggestion of a S C M at » 2 m depth, though this was not confirmed by 104  CHLH 0.0 I  REL.  0  22.24  (MG/MXX3)  CHLR  5.8 11.6 I I FLUORESCENCE 25 SO  23.13 SIGMR-T  24.02  0.0  REL.  FLUORESCENCE 25.0 50.0 _L  0.0 I  REL.  0  '<k~*  22.82  Figure 3.32 Aug./82 (Al) Chlo (squares) and in vivo fluorescence (line) profiles in relation to a (circles). t  (MG/MXX3)  10.0 20.0 1 I FLUORESCENCE 25 50 L  23.40 SIGMA-T  23.99  CHLfl  0.0 I 0.0  REL.  FLUORESCENCE 25.0  50.0  0  REL.  >4  (MG/MXX3) 5.5 I  11.0 1  FLUORESCENCE 25  0.0  50  REL.  FLUORESCENCE 40.0  80.0  ex...  m 4(0 •  (b en  RUG 1 3 DROGUE 1 23.26  Figure 3.32 (continued).  23.63  SIGMA-T  23.97  u>  22.9  PUG 1 7 DROGUE 1 1 23.54  SIGMA-T  •ID  24.19  0.0 I  UPTAKE  CHLA 0.0  (MGC/M3/HR) 12.8 25.6 I I (MGCHLA/M3) 3.2  1  - T - O  22.0  23.25 SIGMA-T  o  0.0 I  UPTAKE  CHLA 6.3  0.4  r ° 24.5  -<p 22.0  (MGC/M3/HR) 38.6 77.2 I I (MGCHLA/M3) 8.3  1  23.25 SIGMA-T  o  0.0 I  UPTAKE  CHLA 16.2  po  24.5  0.0  o  T^I  22.0  Figure 3.33 Aug./82 ( A l ) Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to cr (circles). t  (MGC/M3/HR) 30.1 60.2 I I (MGCHLA/M3) 4.6  1  23.25 SIGMA-T  9.1  po  24.5  0.0 I  PB  -j  22.0  (MGC/MGCHLA/HR) 2.6 1  1  23.25  o—<_>  PB (MGC/MGCHLR/HR)  S.2 1  p a  3.3 1  22.0  23.25  a  24.5  SIGMA-T  0.0 I  |  »  u  8.6 1  J —' 1  24.5  PB (MGC/MGCHLA/HR)  0.0 I  O -J  22.0  SIGMA-T  F i g u r e 3.34 Aug./82 ( A l ) P c*t (circles).  B  (dots ±S.D.) and N O J (squares) profiles in relation to  3.5 I  1  23.25  SIGMA-T  7.0 I  1- O  24.5  the coincidental fluorescence profile. On Aug 10 however, the fluorescence profile did indicate a SCM at 3-4 m, subsequently confirmed as a Chla feature near the base of the pycnocline on Aug 11. The Aug 11 0-2 m depth averaged N O J was reduced significantly from Aug 9, but was still > 1 / i M , with the peak in carbon uptake slightly shallower than the SCM. This subsurface primary productivity maximum (SPM) was due mostly to biomass distribution since corresponding P  B  decreased essentially monotonically with  depth. Comparing the Aug 11 and 13 fluorescence, Chla , and o profiles indicates that t  relative Chla distribution remained constant in relation to the pycnocline after Aug 11, though the SCM did decrease in amplitude. The 0-4 m Chla distribution on Aug 13 was uniform with depth and of higher concentration than on Aug 11. However, the peak in carbon uptake had continued to increase in depth and became coincident with the SCM on Aug 13 (Figure 3.33). The 0-6 m integrated Chla specific biomass decreased from 40.5 mgChla / m on Aug 2  9 to 35.8 mg/m on Aug 13, though increasing first to 54 mg/m on Aug 11. The 0-6 m 2  3  integrated carbon uptake rate ranged from 850 m g C / m d on Aug 9 to 1200 m g C / m d on 2  2  Aug 11 and 13, while 0-6 m depth averaged N O J increased from 6.3 /xM to 6.7 /xM over the 5 days (Tabie in.9). Total daily irradiance over A l ranged from 14 E / m on Aug 9 2  and 11 to 46 E / m on Aug 13 (Table i n . 10). 2  T a b l e III.9 Aug/82 (Al) production v&riables < 7 | >, Z  t a t  (m) as in Tab/e 1/7.3, 0-6 m P (mgC/nPh)  and P (mgC/m -d), 0-6 m depth averaged NOJ concentration f/iM ), 0-6 m S m PI (mgC/mgChla h). 7  d  Date Aug 9 A u g 11 A u g 13  <I's >  Z lt 5  Ph  Pd  NO "  311 1050 656  2.0 8.3 5.6  82 193 91  853 1200 1200  6.3 3.1 6.7  s  S  PI  34.2 47.5 31.4  2.9 4.1 2.9  B  109  h  B  (mgChla /ni ), and 0-6 2  T a b l e 111.10 Total daily PAR during the Aug/82 cruise. Date  E n  (E/m d) 2  14 46 29 14 36 38 36 29 35 46  Aug 9 Aug 10 Aug 11 Aug 12 Aug 13 Aug 14 Aug 15 Aug 16 Aug 17 Aug 18  8.4.2.8 Patterns in species composition The species composition data for A l is presented in Appendix G . To summarize, the diatom Thalassiosira polychorda, was dominant at the surface on a volume basis (29%) on Aug 9, with secondary contributions (i.e. 10-20%) by two autotrophic dinoflagelates, Peridinium sp. and Gymnodinium sp., two unidentified autotrophic flagellates (i.e. a cryptophyte and a chrysophyte; the latter perhaps Isochyrsis galbana, Dr. F J R Taylor, Dept. Oceanography UBC, pers. comm.), and the autotrophic ciliate Mesodinium rubrum. At depth, the assemblage was dominated by a chrysophyte (38%) at 10 m, and by Peridinium sp. (45%) and a cryptophyte (33%) at 20 m. Total cell volume ranged from 1.7 • 10  6  p.m / m l at the surface to 8.6 • 10 / 1 m / m l at 20 m. 3  4  3  On Aug 11, the surface was equally represented (i.e. 20-25%) by Peridinium sp. and M. rubrum . This shift in distribution compared to Aug 9 resulted from a 10-fold decline in T. polychorda and 5-fold increase in M. rubrum . At 2.5 m, corresponding to the upper limit of the SCM (Figure 3.33), the chrysophyte was dominant (47%) with a 24% contribution from Chaetoceros spp. (individual cell vol = 1000 tun ) and 10% contribution 3  from M. rubrum . Deeper samples were not collected. Total cell volume increased from 1.3 • 10 / i m / m l at the surface to 2.1 • 10 p.m / m l at 2.5 m. fl  3  6  3  The surface on Aug 13 was dominated by the chrysophyte (33%) and Chaetoceros spp. (19%), with 10% contributions by Peridinium sp., Dinophysis sp., the cryptomonad, and M. rubrum. The shift in relative abundance from Aug 11 was due to a 2-fold reduction in 110  Peridinium sp. and M . rubrum , a 3-fold increase in the chrysophyte and a 40-fold increase in Chaetoceros spp. At 2.5 m, corresponding as on Aug 11 to the upper limit of the SCM, the composition was dominated by the chrysophyte (49%), with 10-15% contributions by Peridinium sp., M. rubrum, and T. polychorda. In this case the shift in relative abundance was due to a 5-fold reduction in Chaetoceros spp. At 3 m, corresponding to the SCM, the chrysophyte was again dominant (75%), with small contributions (i.e. 5-10%) from Chaetoceros spp., Af. rubrum and Peridinium sp. Total cell volume ranged from 1.2 • 10 /xm / m l at the surface to 3.5 • 10 3  6  6  /xm / m l at 3 m. Again deeper samples were not 3  collected.  This sparse species enumeration data can be augmented by the spatially more comprehensive seston size-frequency distribution data. Unfortunately, this collation is obfuscated somewhat by the apparent lack of a volume specific correspondence between certain features in the species enumeration record (volumes estimated by eyepiece micrometry) and the Coulter counter seston data. Notably, the essentially mono specific (chrysophyte) assemblage at 3 m on Aug 13, should have had a distinctive peak in the size-frequency spectrum at w 1000 /xm . Instead, the peak extended from 900 to 5 • 10 3  /xm (Fig-  3  3  ure 3.35). On the other hand, the T. polychorda and Peridinium sp. dominated surface sample on Aug 9 demonstrated a size-frequency peak that corresponds fairly closely with the 3 • 10  4  /xm mean cell volume estimate for these two species (Figure 3.35). Thus 3  this volume discrepancy does not appear to be a consistent offset, and in fact may be limited to the chrysophyte. This conjecture is supported by the fact that this latter phytoplankter preserved so poorly in Lugol's solution, that only the chloroplast structures were recognizable. Significant cell shrinkage could be associated with this fixation effect. Consequently, it was assumed that a volume correspondence existed between the Coulter counter estimation and all the major phytoplankton contributors except the chrysophyte, which instead was assumed to be the primary contributor of the entire 900 — 5 • 10 /xm volume 3  3  range. The calculated volume contributions for those phytoplankters identified as important components in the species enumeration data and that could also be so correlated ill  en o '  CO  RUG 9 DROGUE 1 0.0 n  o  O —t  1—°'  CD a»  <  "3  T 110  55.  28.  14.  5.0  VOLUME  nn.nl l.nn.nn. 3.4  BINS  1.7  0.9  .43  .22  .11  (10 pm ) 3  3  cn o •  cn  RUG 13 DROGUE I 3.0 n  O  —I I D co i—o  CD at  3  T 110  55.  28.  14.  5.0  VOLUME  3.4  BINS  JZl T  T  r  1.7  0.9  .43  o o  .22  .11  (10 /jm ) 3  3  Figure 3 . 3 5 Aug./82 ( A l ) seston size-frequency spectrum expressed on a per cent total seston volume (left) and a volume basis (right).  112  with a feature in the seston size-frequency spectrum are tabulated in Tabie III. 11. Since Chaetoceros spp. could not be identified with a feature in the size-frequency spectrum it is not represented in these estimates. It should also be noted in connection with the chrysophyte fixation artifact that the identification of this phytoplankter is speculative, since it is largely based on the shape of the discernable chloroplast structure and the single sighting of a bi-flagellated individual. However, since the aim of the classification is to differentiate components according to trophic regime (i.e. photo-autotrophic vs heterotrophic) and presence or absence of active locomotion, this was deemed sufficient.  Table III.ll Aug/82 (Al) time series volume contributions by pbytoplankters identified to be consiste dominant in the Lugol s solution preserved samples. Vbiumes based on total volume specific seston presented in units of 10 ftm /ml. Percent of total seston volume in parentheses. s  Date  Depth (m)  Aug 9  0 2 4 6 10 14 20  2158 830 498 664 166 830 166  (28) (14) (12) (17) (9) (12) (21)  A u g 11  0 1.5 2.0 2.5 3.0 3.5 4 6 8 12 20  0 1 2.5 3 3.5 4 6 8 12 16 20 * In units of 10  Peridinivm sp.  (55-110* )  A u g 13  s  3  M. rubrum  (28-55* )  T. polychorda chrysophyte  (14-28* )  (0.9-5* )  1411 (19) 1743 (29) 498 (12 664 17 83 (5) 166 (6 83 (11  1176 (16) 1092 (18) 672 (16 378 (10 84 5) 126 (16 126 (16  1161 1105 1359 1388 922 847 311  830 (19) 1328 (16) 1660 (17) 2324 (19) 2656 (19 1494 (16) 1162 (13) 498 (30) 332 23) 830 41) 166 (41)  913 (21) 1743 (21) 2739 (29) 2324 (19) 2905 (21) 1660 (17) 1992 ( 23) 83 (5) 415 (29) 166 (8) 0(0)  546 (12) 1344 (16) 1344 (14) 1764 (14) 1470 (11) 1386 (15 1134 (13) 84 (5) 84 6) 294 (15) 84 (21)  919 (21) 2360 (29) 2481 (26) 3623 (30) 4481 (33) 3399 (36 2897 (33) 579 34) 360 25) 397 20) 52 (13)  98 (10) 996 (14) 1C60 (20 2158 15 664 (9) 498 (11 166 (14 498 (40 166 (23 166 (20 332 (28  581 (12) 1660 (24' 1328 16 1660 (11, 830 (11 664 (15 83 (7) 166 (13) 83 (12) 166 (20' 249 (21  420 (9) 420 (6) 1008 (12) 1764 (12) 756 12 462 10 84 7 K 126 10 126 18 84 (10) 84 (7)  1934 (40 1501 (34 1587 (32 6120 (42 2080 (43 1620 (37 452 (41 240 (19 170 (30 230 (35 211 (81  15) 18) 33) 36) (47 22) (20)  Aim /ml 8  In general the species specific seston volume fractions were in approximate agreement with the microscopic enumeration data on a percent total volume basis. The exception 113  is the chrysophyte fraction, which did not show the same contrast either spatially or temporally as did the enumeration data. This is presumably attributable to this lack of a good volume correspondence between these components in the two data sets. On Aug 9 the Peridinium sp. and T. polychorda fractions decreased monotonically with depth, while the M. rubrum and the chrysophyte fractions showed subsurface maxima at 2 and 4-6 m respectively. On Aug 11 all four fractions increased monotonically with depth to a maximum at 2.5-3 m, except for M. rubrum which again showed a secondary maximum at 2 m. The depth of maximum volume varied between the fractions. For Peridinium sp., M. rubrum , and the chrysophyte the peak depth was 3 m, while for the T. polychorda fraction the peak was 2.5 m. However, T. polychorda was not an important species at 2.5 m and hence this peak must be attributed to some other seston component. Chaetoceros spp. were important species, present in the form of chained colonies. To what degree this clumping of cells shifted the apparent volume of this 1400 /xm species in the 3  Coulter counter estimation is not known. On Aug 13, the M. rubrum fraction again had a secondary volume peak, though this time at 1 m. However, all four fractions maximized at 3 m. In drawing inferences from this seston volume pattern, it must be kept in mind that the particle counts on which the volume estimates were based are frequently < 20 for the > 14 • 10 /xm volume bins. Assuming the counts have a Poisson distribution, the 3  3  95% CI for a single count of 20 is ± » 9 (Elliott 1977, p. 85, with a more precise estimate of the confidence limits not justified due to the uncertainty associated with ignoring non-phytoplankton contributions to the seston volume fractions). Thus volume estimates for these large size fractions have confidence limits of at least ± 50%. Hence the subsurface maxima for the chrysophyte and M. rubrum fractions on Aug 9 and the secondary M. rubrum maxima on Aug 11 and 13 are not significant statistically. The same also applies to the depth differences in the volume maxima between fractions on Aug 11.  114  S.4-2-4 Inferences concerning SCM formation The appearance of the S C M on Aug 11 was not associated with a subsurface peak in P  B  . Hence as for M2 in May/82, assuming this P  B  distribution was not an artifact (see  Sec 3.4.1.8), then at first glance depth dependent in situ growth and/or vertical migration/sedimentation combined with depth dependent respiration and P-I characteristics are required to account for this Chlo distribution. The subsequent descent of the SPM on Aug 13 (Figure 3.34) however, was associated both with a subsurface peak in P  B  and a nutrient  depleted (i.e. < 1 (JLM) 0-2 m surface layer, thereby implicating an eventual involvement of nutrition in the SCM dynamics as well. Furthermore, photo-inhibition is known to increase with increased nutrient deprivation (Belay 1981). Combined with Z t > 5 m on Ba  Aug 13, this suggests that photo-inhibition may also have contributed at least by way of this interaction in the appearance of the subsurface peak in P  B  .  Another mechanism associated with SCM situated at the nutracline is the response of cellular C/Chla to the elevated near field nutrient and decreased near field light regimes (Cullen 1982). However, the seston volume data indicated a close correspondence in most cases between Chla and inferred carbon distribution (Tabie HI.ll) indicating a relatively constant C/Chla with depth. The main exception was the unusually large total seston volume for the chrysophyte fraction at 3 m (i.e. SCM) on Aug 13. This represents a » 50% increase from Aug 11 (Tabie in. 11), though corresponding Chlo decreased by about the same amount (Figure 3.33). Furthermore, P  B  at 3 m on Aug 13 was « 6 mgC/mgChla hr,  a factor of 2 higher than on Aug 11 and a factor of 3 higher than the depth interpolated estimate for Aug 9 (Figure 3.34). Hence, this suggests either that the Chla concentration was underestimated on Aug 13 at 3 m or the chrysophyte dominated assemblage had an inherently low Chla content. However, the latter effect cannot be attributed to cell size, since in general volume specific Chla content is an inverse function of cell volume, with cell chlorophyll proportional to surface area and cell carbon proportional to volume (Eppley and Sloan 1966). Also, errors associated with the acetone extraction of Chla appear to be largely restricted to members of the Chlorophyceae, Cyanophyceae, and Dinophyceae 115  (Holm-Hansen and Riemann 1978), and hence is an equally improbable explanation. The only reasonable possibility remaining is some form of analytical error. This is supported by the fluorescence profiles during A l .  The fluorometer was not re-calibrated during  the Aug/82 cruise, though it was turned off at the end of each sampling day. On the assumption that sufficient warm-up time was allowed to stabilize the instrument each day, the profiles were essentially comparable between days. Hence the 2-fold increase in SCM fluorescence between Aug 11 and 13 (Figure 3.33) suggests a trend in inferred Chlo opposite to that of the extracted Chla , implying that the latter is spurious. If this is the case then it can be concluded that within the limit of detection associated with estimating phytoplankton carbon from seston volume data (i.e. non-phytoplankton particulate background), C/Chla remained uniform with depth over the course of A l . The homogeneity of the 0-4 m Chla distribution and the slight shoaling of the nitracline on Aug 13 is consistent with the increase in surface mixing interpretation of the Aug 9-13 salinity profile time series. It is speculated that since the wind record from Aug 12-13 did not differ significantly from that of Aug 9-11, this increase may have been due to abrupt mixing following a cummulative reduction in column static stability, which in turn could have resulted from the continuous erosion of the pycnocline over the 5 day interval. Much of the reduction in S C M Chla on Aug 13 coincident with this mixing event may have been an analytical artifact. However, the species enumeration and seston size frequency data did also indicate a concurrent qualitative shift in the phytoplankton composition of the SCM following this mixing event. In particular, motile forms appeared to be retained preferentially over non-motile forms, with the possibility being that a locomotory capability enabled these motile forms to counteract the effects of the increase in mixing. The last fluorescence profile on Aug 17 showed evidence of vertical smearing and a reduction in the amplitude of the SCM. However, it is unlikely that this was due to wind mixing, since the winds were comparatively low on Aug 16-17. This contention is supported by the otherwise inconsistent increase in surface temperature from Aug 13 to Aug 17 (Figure 3.31). However, the shallow bathymetry of the Aug 17 station ( « 35 m) 116  compared to > 90 m at the other stations suggests that an increase in tidal mixing due to the shallower water column may have contributed. 8.4.2.5 Chla and nitrogen budgets Chla and nitrogen budgets can be calculated from the A l time series as a basis to estimating the importance of the various sources/sinks for the precusor nutrient and Chla specific biomass. For this calculation nutrient replete C / N and C/Chla of 7 and 20 were assumed respectively (ratios taken from Laws and Bannister 1980). Since carbon uptake, Chla specific biomass, and ambient N O 3  were determined every other day, it  was necessary to interpolate between days. In keeping with the approximate nature of these calculations, simple linear interpolation between day specific estimates was done. An additional simplification was that N O 3  was assumed to be the sole source of ambient  nitrogen. Finally, two estimates of Chla specific production for the Aug 11-13 interval were calculated due to the equivocal Chla time series, one using the change in extracted Chla concentration at the 3 m SCM, the other using the observed increase of the fluorescence yield, with this yield converted to equivalent Chla concentration by calibrating with the Aug 11 Chla concentration. Table III. 12 presents predicted daily NO3  demand and  Chla specific production, as well as the observed change in these constituents during the A l time series.  Table III. 12 Aug/82 (Al) Chla and nitrate mass balance. Predictions based on the measured carb uptake rate (Table til.9) and an assumed nutrient replete C/Chla and C/N of 20 and 7 respectively. Predicted N O , demand (mgNOj / m )  Date  2  Aug 9 A u g 10 Total  557 670 1227  A u g 11 A u g 12 Total  784 784 1567  Measured dNO, (mgNO^ / m ) 2  Predicted C h l a production (mgChla / m ) 2  Measured dChlo (mgChla/m ) 2  43  -600  800  Total  51 94  12  Total  60 60 120  -15 (-4)*  * Based on change In S C M in vivo fluorescence  Both Chla and N 0 timates.  3  show large discrepancies between observed and predicted es-  From Aug 9-11, when 0-6 m integrated production should have been » 94 117  mgChla / m based on daily carbon uptake, the actual measured increase was only w 2  12 mg/m . Over the same period 0-6 m N O 3 2  demand was predicted to be 1.2 • 10  3  mgNOJ / m , whereas the measured decrease in ambient NO3 was only 600 mg/m . For 2  2  Aug 11-13 these defects were even larger, with predicted production exceeding observed production by « 130 mgChlo / m and the 0-6 m ambient N O 3 content increasing by 800 2  mgNOj / m whereas it was predicted to decrease by 1.6 • 10 mg/m . The ambiguous 2  3  2  nature of the Aug 13 SCM magnitude does not detract seriously from these calculations (Tabie III. 12). In the absence of significant drogue slippage and/or horizontal advection/diffusion effects violating the quasi-Lagrangian assumption underlying the time series interpretation, these discrepancies must be attributed to one or more of several possible sinks and sources for N O J and Chlo not included in the initial carbon uptake-based calculations. One such factor is herbivore grazing which, though through indirect evidence was indicated not to be important during the May/82 cruise, still represents potentially both an important sink for gross Chlo production and an alternate nitrogen source. Unlike the May/82 cruise, near surface diurnal grazer abundance was estimated during Aug/82. Tabie HI.13 presents these Clarke-Bumpus plankton tow estimates of vertical zooplankton concentration in the euphotic zone. The herbivorous zooplankton component was dominated by the copepod Acartia spp., while the major carnivorous contributors were the siphonophore Muggiaea atlantica, decapod zoea, and a Cancer sp. megalops larvae (Appendix H). Over the course of A l , the copepod component of zooplankton biomass increased from a depth averaged 8% on Aug 9 to 41% on Aug 13 with this increase concentrated at 2-3 m below the SCM. This distribution is somewhat surprising since observations elsewhere (Longhurst and Herman 1981) suggest that the herbivore maximum when present is typically displaced above the SCM in approximate coincidence with the P  B  maximum. However, since a vertical concentra-  tion of zooplankton was not in evidence on Aug 11, even though a pronounced S C M was present, the non-uniform Aug 13 distribution may have been ephemeral and not generally 118  Table III.IS Aug/82 (Al) vertical herbivorous copepod distribution (primarily Acartia sp.), as well a cent of total sample by dry wt. Dry wt., based on average wt. from all samples, estimated to be ug/animal. Total voJume sampled by the C-B tows also tabulated. Date  Depth (m)  # /m3  dry wt. (mg)  2 8 10 16 20 1 3 5 20 1 2 3  458 104 238 158 37 189 192 253 203 120 243 77 313 3794 1202 243  5.3 1.2 2.7 1.8 0.4 2.2 2.2 2.9 2.3 1.4 2.8 0.9 3.6 46.3 13.8 2.8  Aug 9  Aug 11  Aug 13  4  5 8 20  %T Towed Volume (»')  12 1 100 2 2 5 5 17 5 56 67 44 15 50 42 13  29 31 33 33 32 40 13 34 31 20 22 23 24 45 32 31  representative. In any event, Tabie 111.14 presents the depth specific daily grazing loss calculated on the basis of these observed copepod densities, assuming a constant clearance rate of 3.0 ml/d.  Table III.14 Aug/82 (Al) estimated grazing loss/nitrogen excretion . Excretion presented as NO, eq alent to facilitate comparison. Date  Depth (m)  Grazing Loss (mgChla /m -d) xl0-  N-excretion (mgNO^ /m d)  7.4 0.6 1.3 0.6 0.05 38.8 4.3 8.0 4.3 0.3 28.9 1.6 3.2 1.8 4.4 37.4 0.07 66.7  0.059 0.005 0.010 0.005 0.000 0.310 0.034 0.064 0.034 0.002 0.230 0.013 0.025 0.014 0.035 0.298 0.000 0.530  s  s  s  Aug 9  0-6 m Total Aug 11  0-6 m Total Aug 13  2 8 10 16 20 1 3 5 20 1 2 3 4 5 20  0-6 m Total  The estimated 0-6 m integrated grazing losses ranged from 0.04 mgC/m on Aug 9 3  to 0.07 mgC/m on Aug 13, indicating that daytime grazing, as in the case of May/82, 2  119  was an insignificant component in the total Chla flux. Table 111.14 also tabulates the calculated nitrogen excretion assuming a 60% ration elimination (Raymont 1983, p. 665) and 100% recovery of fecal pellet nitrogen. Acartia spp. biomass was often a small fraction of total zooplankton biomass (Tabie HI.13). However, the calculated nitrogen flux (i.e. presented as equivalent NO3" for comparison purposes) was sufficiently small compared to the estimated N O 3  defect that the omission of higher trophic level contributions can not  alter the conclusion that as in the case of its impact on the Chla budget, diurnal grazing was not an important term in the nitrogen budget.  The  Phaeopigment/ ]jT) Chla index can again be used to estimate combined diurnal  and nocturnal grazing pressure, with the Sept/82 ICM nighttime zooplankton survey used to infer grazer community structure during Aug/82. In regard to drawing inferences from the ICM zooplankton record, it is worth mentioning that the Sept/82 I C M daytime 3-5 m tows were dominated as in the case of the Aug/82 cruise, by Acartia spp. (ICM 1983). This adds validity to the tacit assumption that the zooplankton community did not differ significantly between Sept/82 and Aug/82. The I C M nighttime tows in Sept/82 were dominated on a wet weight basis by Calanus pacificus, Metridia pacifica, and Euphausia pacifica (ICM 1983). Due to the larger size of these grazers compared to Acartia spp., the 3.0 ml/d clearance rate used to calculate diurnal grazing pressure for May/82 and daytime Aug/82 is not appropriate. In a statistical analysis of clearance rate dependence on body size, Peters and Downing (1984) estimated a mean filtering capacity of w 10 ml/d for copepods of the size of C. pacificus and M. pacifica (i.e. 2-3 mm in length) and » 100 ml/d for E. pacifica (15-20 mm body length, I C M 1983). The 100 ml/d clearance rate of E. pacifica is somewhat contentious, however, since the corresponding approximate body wt. (1 mg) is near the upper limit in the range of, and is present only occasionally in, the data set used by Peters and Downing (1984) to establish their clearance rate vs. body size regression. Hence it does not contribute much to the least-squares fit, with the result that the residual error over this body size range is large. The extent of this bias is illustrated by Ohman (1984) who measured a maximum clearance rate of 1.8 1/day 120  for E. pacifica of the same approximate body size feeding on the diatom Thalassiosira angstii. In any event, since the grazing index is based on a copepod dominated herbivore community (Lorenzen 1967), the euphausiid component was ignored in this initial grazing estimate and a clearance rate of 10 ml/d adopted. Tabie III. 15 tabulates the calculated 0-6 m integrated copepod content based on the phaeopigment grazing index linearly interpolated between sampling days, as well as the calculated daily grazing loss and nitrogen regeneration (again assuming 60% ration elimination and 100% fecal pellet nitrogen release).  Table III.15 Aug/82 (Al) estimated 24 h grazing pressure on the basis of the phaeopigment grazing in integrated from 0-6 m. Nitrogen excretion presented as NO, equivalent to facilitate comparison. Date  ]P Phaeopigment/£ Chla  J^copepod (xlO )  Total grazing (mgChlo/m )  NO equiv. (mgNOr/m )  4  2  s  2  Aug 9 10 Total  0.06 0.06  1.7 1.8  1.0 1.2 2.2  7 8 15  Aug 11 12 Total  0.07 0.05  6.2 4.3  3.4 2.7 7.6  23 19 42  Based on this phaeopigment-inferred grazing pressure, the grazing contribution to the Chla and nitrogen budgets is substantially larger than the estimate based solely on observed daytime grazer density. However, this contribution to both budgets remains small, accounting for, as in the case of May/82, « 3% of the estimated defect in each. Another sink/source of Chla and N O J is turbulent diffusion across the pycnocline. This flux F can be estimated if the vertical eddy diffusivity and vertical gradient in the scalar 5 are known, with F = K -dS/dz. z  K can be estimated by scaling the eddy viscosity z  A to the gradient Richardson number Ri wherein K = A (l + Ri)~*. This particular z  t  z  z  approach will be used in chapter 4, where £ will be empirically derived by tuning a heat budget model predicting a temperature profile time series of Holberg Inlet. However, for the present preliminary calculation in comparing the potential contribution by sinks and sources pertinent to the Chla and nitrogen budgets, K was expressed as a function of z  121  the kinetic energy disipation c and the Brunt-Vaisalla frequency N, where K = aeN~ . 2  x  a is an empirical parameter related to the onset of turbulence in a stratified fluid and is equal to « 0.25 (Denman and Gargett 1983). N was calculated from the measured density profiles, while an appropriate value for e in the basin surface layer was unknown and hence was treated as a free parameter. Beginning with estimates for surface mixed layers taken from the literature (Denman and Gargett 1983), c was adjusted to generally minimize the N O J and Chla defects. In this way an equivalent range for K  z  was determined  which brackets an upper limit of K that remains consistent with an essentially balanced z  NO3" and Chla budget. Tabie 111.16 tabulates the gradients in Chla , and NOJ" , as well as the buoyancy frequencies; while TaWe 1H.17 summarizes the calculated fluxes over the derived range in K of 5 • 10~ m /s to 5 • 10~ m /s. 5  2  4  2  z  Table III.16 A l vertical gradients in Chla and NO, ,as well as the Brunt-Vaisalla frequency used estimated diffusivefluxesof Chla and NOJ . Linear interpolation of gradients between sampling dat Date  dChlo /dz (mg/m )  dNOJ /dz (mg/m )  0.7 2.5 4.4 3.3  94 112 130 193  4  Aug 9 10 11 12  N (10-  4  2  S- ) 1  4.9 4.7 4.5 5.2  Table 111.17 Aug/92 (Al) 0-6 m integrated Chla and NOJ defects (Ce and Ne), and Chla and NOJ diffusion Buxes (Cf and Nf) over the optimal mass conserving range in e. (-) indicates missing mass. Interval  ~C~e mg/m  1 = 5-10Aug 9-11 Aug 11-13 Total e = 5-10" Aug 9-11 Aug 11-13 Total  7  6  m /s -82 -130* -212 m /s -82 -130* -212 2  2  s  s  2  Cf mg/m  2  Ce+Cf mg/m  Ne mg/m  2  [K = 5 10- m /s) 15 -67 627 37 -93 2367 52 -160 2994 (A\ = 5•lO- m /s) 153 71 627 373 243 2367 526 314 2994 s  Nf mg/m  2  Ne-Nf mg/m  2  2  2  t  4  979 1492 2471  -352 875 523  2  1-10 1.5 • 10 2.5 • 10 4  4  4  -9.2 • 10 -1.3-10 -2.2 • 10  s  4 4  * in vivofluorescencebased estimate  From Tabie III. 17 it appears that the upper limit of K is probably somewhere between z  5 •10  - 5  and 5 • 1 0  - 4  m /s and recognizing the small inferred contributions by the other 2  flux terms is most likely closer to 1 • 1 0  - 4  m /s. This estimate is typical for a weakly  122  2  stratified mixed layer under low wind conditions ( « 5 m/s, Denman and Gargett 1983) or for a thermocline (N = 7 • 1 0  s ) under high wind conditions ( » 15 m/s, Dillon and  - 3  - 1  Caldwell 1980). Low wind conditions and a surface layer significantly more stratified than the thermocline on which Dillon and Caldwell (1980) based their estimate of K = 2 • 10~  5  z  m /s (i.e. AT was > 4 • 10~ s ) prevailed during A l . Hence another mixing source is 2  3  - 1  indicated, which implicates tidal mixing as an important source of mixing across the basin pycnocline during Aug/82. An additional sink for Chla is the sedimentation flux, resulting from the sinking of negatively buoyant phytoplankton cells. According to Stoke's Law, the sinking speed of a spherical particle is proportional to the square of its radius (Smayda 1970). Hence to a first approximation in the absence of active locomotion and/or buoyancy control, the shift in the seston size-frequency spectrum during A l should have affected the Chla specific sedimentation flux. The dominant particle size ranged from a volume bin mean of 82 • 10 / i m on Aug 9 to 3 • 10 / i m on Aug 13 (TaWe H L l l ) , with the corresponding 3  3  3  3  size-dependent sinking rate decreasing from « 1 m/d to w 0.3 m/d (Bienfang and Harrison 1984). Assuming a mean of 0.7 m/d and multiplying by the 0-6 m depth averaged Chla concentration over the course of A l , gives a total sedimentation of 19 mgChla / m  2  or a loss term of about the same magnitude as the phaeopigment inferred grazing loss. The estimated sink and source terms calculated above along with the mass balance defects are summarized in TaWe III.18.  Table III.18 Aug/82 Chla and NOJ mass balance defects accounted for by each sink and source term Flux Chla (mg/m )  Source/sink  2  daytime grazing 24 h grazing Sedimentation Diffusion 5- 10- m /s 5 10- m /s 1- 10- m /s 5  2  4  2  4  2  Flux  % Defect Chla  % Defect NO -  0 4 9  0 2  s  (mg/m ) 2  0.1 8 19  1 57  52 526  2.5 • 10 2.5 • 10  25 248  84 835  104  5-10'  50  168  -  s  4  -  123  Though diel grazing and sedimentation are significant components of each mass balance, they are clearly small compared to the diffusion term. However, since diffusivity was treated as a free parameter, this inference is inconclusive due in particular to the large uncertainty in the grazing term. Large grazers such as E. pacifica , while clearly important elements of the grazer community (ICM 1983) were not included in the calculation. The presence of these zooplankters has two effects on the grazing estimate. The first is an obvious underestimation of grazer specific clearance rate. The second is the biasing of the 53 P h a e o p i g m e n t / C h l a grazing index since proportionally more phaeopigment is lost from the euphotic zone through the flux of large (i.e. fast sinking) fecal pellets. Though an entrainment flux was not explicitly included in the mass balance calculation, a rough estimate of its importance can be inferred from the relative Chla and NO J defects. Due to the sign of the N O J and Chla depth distributions, the net diffusion flux will be upward and downward directed respectively. Hence, any (upward) entrainment component in this mixing contribution to the mass balance will be represented in the NO3 but not in the Chla balance. As a consequence, NO3  balance  will conserve with a lower mixing  parameter (i.e. diffusion parameter) estimate than will Chla . This will reflect in a difference in mass defects at a given diffusion rate. As TaWe IH. 17 demonstrates, NO3" does indeed conserve at the lower diffusion rate, though the imprecision in the N O J and Chla gradient estimates may also have contributed to the difference.  S.4-2.6 Summary The Aug 9-13 salinity time series and the mass balance inferred eddy diffusivity both suggested that unlike May/82, tidal mixing may have contributed to surface mixing, with the physical evidence further suggesting that tidal mixing was at least partly in the form of the positively buoyant tidal jet driven two layer flow suggested in chapter 2. The average wind speed during the Aug/82 cruise was slightly lower and the precipitation slight higher than during May/82. Combined with the more homogeneous surface layer during Aug/82 inspite of the greater buoyancy flux this suggested that surface mixing 124  was slightly greater during Aug/82, with the mixing attributable to tide effects. However, inspite of this, a SCM developed in association with a nutrient depleted surface layer. As in May/82 the initial development of this feature appeared to be behaviorally based. However the subsequent appearance of an unequivocal subsurface peak in P  B  (associated  with the nitracline) also contributed to its further development and maintenance, though preferential retention of motile phytoplankton contributors following a wind mixing event indicated that the behavioral component was still important. The development of the vertical structure in Chla and N O J inspite of the presence of the 2-layer flow (with its implied entrainment regime), as well as the wind generated mixing, indicated that the verticalfluxesof biomass and nutrient remained small compared to the local source and sink terms. Comparing the source and sink terms for Chla and N O J on the basis of their mass balance calculations, mixing loss over the course of the A l time series accounted for from « 50 — 75% of the Chla and NO3 defects, while diel grazing and sedimentation accounted for « 4% and 9% respectively. A mass conserving time averaged vertical diffusivity of 1 • 10~"*m /s was calculated. 2  N O J conserved for a lower estimate of the diffusion parameter, which was interpreted as indicating an entrainment component to the Chla and N O 3  mass balance.  3.4.3 July/83 Cruise 8.4-8.1 Wind effect on drogue travel and current shear Four quasi-Lagrangian time-series (J1-J4) were obtained during the July/83 cruise. The drift stations in relation to drogue position (Figure 3.36) are tabulated in TaWe HI.19.  J l was deployed on July 13 » 6 km east of Coal Harbour (position 3), and travelled « 3 km due east over the next 24 h (position 6). The wind during this time period was variable with the strongest winds (10 m/s) blowing from the S and SW (Figure 3.37). 125  Figure 3.36 J1-J3 (July/83) 3 m depth drogue track (i.e. drift stn. 2-11).  Table 111.19 July/83 drift stations over the drogue time series Sample Date  Sample Time(h)  drift stn  Time Series  Drogue Position  July 13 14 14 15 15  1220-1730 0300-1000 1030-1830 0500-0630 1400-1930  2 3 4 5 6  Jl Jl Jl Jl Jl  3-4 5-6 6 7-8 8-9  16 17  2030-0130 0950-1315  8 9  J2 J2  13-14 14-15  17 18  2120-0145 1012-1630  10 11  J3 J3  17-18 21-22  19 20  1145-1850 0130-0400  13 14  J4 J4  37-38 42-43  Over the second 24 h period, the wind was predominantly W and SW at < 5 m/s, while the drogue travelled an additional 1 km eastward. J2 was deployed on July 16 w 7 km west of the J l starting point (position 13). From there it travelled east « 3 km over the next 12 h (position 15). The wind was predominantly W (3-4 m/s) during this period. J3 was deployed « 1 km west of the J l starting point (position 17) and remained essentially stationary over the next 14 h. The changing direction of the 2-4 m/s wind during this period (i.e. N E to SW) may have contributed to this apparent lack of drogue travel. J4 was deployed « 3 km west of Coal Harbour (position 37, Figure 3.38) and travelled a net distance of « 1 km east over the next 12 h (position 43). The wind was generally SW and W (2-4 m/s) except for a short period of E and N E (2 m/s). Lagrangian estimates of the relative current shear at various times during J1-J4 are tabulated in TaWe 111.20. As for the May/82 and Aug/82 Eulerian estimates of the relative current profile, the shear was predominantly east-west and spanned the near surface depth interval (i.e. 1-3 m in this case). However, during J l and J4 the magnitude of the shear was typicially higher than the 1-5 m Eulerian estimates by a factor of 10. How much of this can be attributed to differences in technique is not clear. However, in maximizing near surface vertical resolution by making the window blind sea anchor for the 1 m drogue only 1 m deep as opposed to 2 m for the 3 m and 20 m drifters, the direct wind effect on this 127  Figure 3.37 July/83 wind record.  Figure 3.38 J4 July/83 drogue track (i.e. drift stn. 12-15).  Table 111.20 July/83 relative current shear. Sl/3, S3/20, and Sl/20 are shear estimates from 1-3 m, 3 m, and 1-20 m respectively. Start of the Lagrangian measurement and state of the tide also tabulated (+) flood and (-) denoting ebb. Date  drift stn  Start Time  July 13 13 13 14  2 2 3  0445 1153 1600 0240  JS July 17 17 17 18 18  10 10 10 10 11  2153 2245 2342 0045 1010  J4 July 19 19 19  13 13 13  1100 1230 1330  Jl  Tide  Sl/3 (10~ s ) U V s  -  + -  +  -1  4 11 46 11 59 9 133 44  S3/20 (10~ s" ) U V 5  2 3 0 2  1  Sl/20 (10"* s" ) U V 1  0 1 0 0  3 10 7 16  1 0 1 5 1 2 0 1 0  -  8 5  1 1  2 0 2 3 1  -  26 17 31 3 30 18  4 0 3 2 2 0  3 1 8 3 6 2  -  -  -  drogue may have been significantly increased compared to the 3 and 20 m drogues, thereby overestimating the wind generated surface current. The 3-20 m shear (Tabie 111.20) was an order of magnitude less than the Aug/82 1-5 m shear, indicating that the maximum shear was embedded in the biased 1-3 m shear estimate, and hence if a 2-layer flow regime was important during July/83, the associated entrainment interface would have been located between 0 and 3 m. Contour plots of salinity and temperature for the drift stations comprising time series J1-J4 (i.e. stn 2-14) are presented in Figure 3.39. J l (i.e. drift stn 2-6) was characterized by a shallow (1-2 m), but strong (3 ppt/m) halocline and weak thermocline (0.05°C/m) throughout, while J2 to J4 (i.e. drift stn 8-9, 10-11, and 13-14 respectively) were characterized by a constant though weakened halocline (1 ppt/m) and a time dependent stengthening of the thermocline (i.e. from » 0.2°C/m for J2 to l ° C / m for J4). The shallow but strong halocline during J l is attributable to the high precipitation from July 11-13 (total of 36 mm, ICM 1984), while the strengthening thermocline during J2, J3, and J4 is attributable to the general increase in surface irradiance after July 15 (Table 3.21). The resultant pycnocline time course was a persistent weakening of the pycnocline from 130  Figure 3.30 July/83 drift stn salinity (ppt) and temperature (°C) time series isopleths. 131  Table 111.21 July/83 total daily I" Date July 13 July 14 July 15 July 16 July 17 July 18 July 19 July 20  Is (E/m ) J  16 14 26 44 41 41 41 28  July 14 to July 15 (i.e. drift stn 4 to 6, Figure 3.40), followed by a vertically more distributed strengthening after the increase in surface heating on July 15. This distributed strengthening can be attributed to the vertical penetration and absorption of the radiant energy buoyancy flux. It appears that the observed evolution of the temperature and salinity profile was predominantly a response to meteorological forcing and hence is generally devoid of a spatial component artifact resulting from drift station location. This was clearly demonstrated in the evolution of temperature and salinity along a fixed station transect line (Figure 3.41), where C T D profiles were taken between successive drift station occupations (Tabie III.22). Figure 3.42 and 3.43 plot the temperature and salinity contours for transects 3 and 13, respectively. These transects show the effects of high precipitation/low irradiance forcing (i.e. transect 3) and low precipitation/high irradiance forcing (i.e. transect 13) on the Holberg Inlet temperature and salinity fields. In comparing Figure 3.42 and 3.43 to Figure 3.39, it is further evident that under both forcing regimes, spatial gradients along the inlet were small throughout the July/83 cruise period compared to the time dependent changes associated with the drift station time series. Table 111.22 July/83 Sxed-station transect sampling times. transect 2 4 9 10 11 12 13 14 15  Date  Sampling Interval  July 13 July 14 July 17 July 17 July 18 July 18 July 19 July 19 July 20  0554-1000 1930-0230 0200-0920 1400-1718 0217-0930 1715-2133 0426-1016 1630-2352 0800-1300 132  Figure 3.41  July/83 fixed sampling station transect line.  FIXED  STATIONS  ' J  .,  11  12.0  11.4  TEMPERATURE TRfiNSECT: 3 12  10  g  FIXED 10  STATIONS  9  7  30.6  SALINITY TRfiNSECT: 3 14  12  10  B  Figure S.42 July/83 transect 3 (fixed station) salinity (ppt) and temperataure isopleths. 135  FIXED 14  12  10  9  STATIONS 8  7  6  S  4  3  2  F i g u r e 3.43 July/83 transect 13 (fixed station) salinity (ppt) and temperature isopleths. 136  8.4$-2 Horizontal patterns in phytoplankton biomass Though the hydrography of the basin surface layer appeared to respond uniformily to meterological forcing, the concurrent response by the phytoplankton was decidedly nonuniform, as indicated by the in vivo fluorescence also measured along the fixed station transect line. The fluorescence contour plots for transect 4, and 9 to 14 are presented in Figure 3.44. Transect 4 sampled on July 14, indicated a relatively uniform distribution in the horizontal while transect 9, sampled on July 16 shortly after the abrupt increased in Ig, showed evidence of elevatedfluorescenceat fixed stn 2-6. On subsequent transects, this phytoplankton patch appeared to be more dispersed in the horizontal with some suggestion of a less clearly pronounced patch at stn 7-14 during nighttime-early morning transects 11, 13, and 15; only for the original pattern to re-appear during daytime transects 10, 12, and 14 (i.e. transect 14 stn 1-6 were occupied from 1630-1845h).  With the exception of transect 4, which was initiated at the mid-point of an ebbing tide (all transects started at fixed stn 2), the transects were all initiated within 2 h of the turn to flood tide. Thus it is not likely that this response can be attributable simply to the tidal excursion relocating the patch within the sampling domain. An interaction between the state of the tide and the wind field that could impart a diel periodicity on the advection/diffusion of the phytoplankton patch is also not apparent (i.e. compare Figure 3.37 and Tabie III.22, recalling that most transects were begun within » 2 h of the turn to flood). The only factor that does consistently co-vary with the appearance and disappearance of the Chla patch is the photoperiod. Transects that immediately followed a diurna.1 ebb tide had the patch present at stn 1-6, while those that followed a noctural ebb tide did not. As already mentioned there was also the suggestion that during these latter times in vivofluorescencewas slight elevated in the headward region of the basin. Physiological processes modifying the fluorescence yield of the phytoplankton seemingly cannot be invoked to account for the overall pattern since dark adaptation has been shown to increase the fluorescence yield, not decrease it (Vincent 1980), while the endogenous rhythm in yield time dependency is generally in phase with the photoperiod (Prezelin and 137  Figure 3.44 July/83 transect 4 and transect9-14fluorescenceisopleths.  138  139  Figure 3.44 (continued).  140  l  S  U  I  I  »  STRTXONS f t  7  «  3  4  I  I  Ley 1980). Furthermore, if some physiological adjustment was responsible, the decline in fluorescence would presumably be spatially uniform, which it clearly was not. A more likely explanation for the coherence of the patch with the photoperiod is that it is a direct result of horizontal/vertical mixing and advection acting to disperse the phytoplankton. During diurnal tides, this is compensatated for by in situ growth, while during nocturnal tides it is not. Though the importance of photosynthetic growth in this periodicity appears clear, the relative contributions by exchange processes is not. Three basic exchange models come to mind. They can be referred to as the residual flow, horizontal mixing, and vertical mixing models of patch response. In the residual flow model the patch, having presumably developed by in situ growth somewhat to the west of transect stn 2-6, was advected down inlet by a residual surface current. Turbulent dissipation of the patch was assumed to have been unimportant, with the disappearance of the patch after every second ebb tide simply reflecting the magnitude of the residual flow; requiring two tidal cycles to advect the patch out of the basin. Subsequently, with each return to daylight a new patch evolved through renewed growth. The horizontal diffusion model on the other hand assumes that mean tidal transport is unimportant, since the time scale of patch dissipation due to potential turbulent mixing of the inflow into the basin surface layer was short compared to this mean transport rate. In the vertical diffusion model, horizontal mixing is unimportant. Rather, it is the vertical entrainment flux associated with horizontal transport that is responsible for patch dispersion. Drogue displacements east of Coal Harbour during the Aug/82 cruise (Appendix I) indicated that large surface layer transport associated with the tidal excursion did occur in this region of the basin. At the same time significant horizontal mixing near QN was not indicated due to the sharp eastward most discontinuity (i.e. front) in the fluorescence isopleths in T12 (Figure 3.44). For these reasons the vertical mixing and residual circulation models are the most consistent with the evidence.  142  S-4-S.S Primary production time series Summary statistics for the July/83 cruise are tabulated in Tabie 3.23. During J l (drift stn 2-6), <Jg > was » 30% of the I C M data set July mean (Appendix J). In addition to the low surface irradiance, the water column light attenuation was high compared to May/82 and Aug/82 (i.e. jfe, > 0.50 compared to 0.20 and 0.25 respectively).  July/83 production variables: 0-6 m integrated Chla and carbon uptake rate as well a 0-6 m depth averaged N O , ,NHf ,POjf ,and Si07* levels tor J1-J4. C incubation interval-average PAR irradiance </| >, the depth of saturating irradiance Z »t, a n d tie depth averaged productivity in PI are also presented. T a b l e III.23  14  B  Station Jl 2 3 4 5 6  <I> >  PH  Pi  NO "  NH+  POf  SiO^*  S  D  PI  6.2 5.2 6.1 7.3 6.3  2.1 1.9 1.9 2.5 2.0  1.6 0.7 1.1 1.2 1.3  17.6 13.1 17.8 23.2 24.3  6.4 6.5 8.1 6.8 6.4  4.2  99  -  7.1 6.7  1.8 1.5  1.1 1.1  19.9 19.3  11.0 7.6  1.2  260  0.6  27.1  417  370  1.4  47.4  498  470  1.6  60.6  931  -  -  -  -  -  s  5.9  -  9.5  J2 8 9  1520  JS 10 11  1370  4.0  50.9  347  -  5.9 5.0  1.3 1.1  0.8 1.0  17.8 17.1  19.1 23.9  2.1  J4 13 14  1140 -  3.6  130.8  1307  4.5 5.1  1.1 1.1  0.8 0.7  21.1 19.8  31.8 7.8  4.1  -  -  -  4.1  -  -  -  13.2  -  -  -  -  -  Chla specific biomass was only 10% of the I C M July mean of « 60 mgChla / m (compare Tabie HI.23 and Fignre 2.10). The 0-6 m depth averaged N O 3  2  ranged from 5.2  to 7.3 pM , comparable to Aug/82, while ambient N H 4 , measured for the first time during the July/83 cruise, ranging from 1.9-2.5 pM . The 0-6 m SiO^ POf  4  was always > 10 pM and  always > 1 pM except for drift stn 3 (i.e. POJ = 0.7 pM ). Since ambient S i / N 1  and P / N were consistently a factor of at least 10 greater than these equivalent ratios in nutrient replete phytoplankton (Redfield et al. 1963; Antia et al. 1963; Brzezinski 1985), it is concluded that Si and P were not limiting during the July/83 cruise. Hence these nutrients will not be discussed in any detail. Thoughout the J l time series Chla and carbon uptake (Figure 3.45), and P  B  (Fig-  ure 3.46) were maximal at the surface and decreased monotonically with depth, consistent 143  with a nutrient replete, light limited surface layer. This is to be expected under the low light/ high column light attenuation and high ambient nitrogen conditions prevalent at the time. Both phytoplankton biomass and nutrients remained relatively constant over the course of J l except for a slight depression in 0-6 m depth averaged N O 3 and P O j for drift stn 3. J2 (drift stn 8 and 9) carbon uptake was low, attributable both to low biomass and to low carbon uptake efficiency (PI = 1.2), with Chla specific biomass maximal at the surface ( Figure 3.47). However, though a surface uptake estimate was not obtained, a subsurface peak in P  B  was indicated (Figure 2.50) and in the presence of high surface ambient nitrogen  suggests strong photo-inhibition. S  B  declined from 11.0 to 7.6 mgChla / m over the 2  interval of this essentially nocturnal time series. J3 was similar to J2, with phytoplankton biomass decreasing monotonically with depth ( Figure 3.48), though the uptake efficiency was slightly higher (P/=2.1) under similar irradiance and nutrient regimes. As with J2, a subsurface P  B  maximum was observed  (Figure 3.50). J4 biomass distribution had a SCM on July 19, but was monotonic with depth on July 20 (Figure 3.49). The irradiance field and 0-6 m depth averaged ambient nitrogen concentration were comparable to that of J2 and J3 (Table HI.23). However, surface nitrogen was « 1 pM. during J4 compared to > 2 pM for J2 and J3 (Figure 3.50). Inspite of this more stressful J4 nutrient regime P  B  decreased monotonicially with depth (Figure 3.50), and as  a consequence, 0-6 m carbon uptake was a factor of 3 greater than during J3 (Tabie IH.23).  8.4S.4 Patterns in species composition The species structure was dominated by 3 photo-autotrophic species: M. rubrum , Peridinium sp., and a cryptomonad (Appendix G). A Dinophysis sp. dinoflagellate also contributed significantly on one occasion. Based on the two surface samples analyzed, cell volume during the early period of the cruise prior to the abrupt increase in irradiance on 144  Figure S.45 J l Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to at (circles).  Figure 3.46 (continued).  Figure 3.46 at (circles).  Jl P  B  (dots ±S.D.) and N O 3 + N H J  (squares) profiles in relation to  o.o  UPTAKE(MGC/M3/HR) l.s  i  CHLA 0.0  I  CHLA  (MGCHLA/M3) 1 1.3  2.6  o  0.1 i  3.1 i  (MGCHLA/M3) 1.0 I  ui  1.9  L o  in < b us \  9  O m —»— x°  t  d  •  I  0 d>  5  (b 1  •cn  cb 1  18.28  JULY 16 ORIFr STN 8 1 20.87  SIGMA-T  <P Qj i <P <b- 3 23.48  19.72  21.58  SIGMA-T  Figure 3.47 J2 Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to o~t (circles).  23.43  Figure S.48 J3 Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to at (circles).  Figure 3.49 J4 Chla (squares) and carbon uptake (dots ±S.D.) profiles in relation to o~t (circles).  0.0 I  PB (MGC/MGCHLfl/HR)  5.4  l.B I  N03+NH4 (uM) 8.0 I  3.7 J  0.0 I  10.6 o  0.0  PB (MGC/MGCHLA/HR) 2.6 I  N03+NH4 (uM) 6.2 I  5.3 I  .0  PB  •  12.4 o  .0  (MGC/MGCHLA/HR) 4.4  8.  I  I  N03+NH4 (uM) 5.7  11  1  Ul  d\  11  9 d>  o m  b m i  rn  T)  0 m W  %  0 m vii 1  <P i  <b"  <P <b  JULY 18 ORIFT STN 11 19.72  SIGMA-T  Figure 3.50 J2-J4 P at (circles).  23.43  21.58  17.45  T  20.45  •o 23.44  JULY 19 DRIFT STN 13 17.78  SIGMA-T B  (dots ±S.D.) and NO3+NH+  (squares) profiles in relation to  20.59  SIGMA-T  i  9 o 23.41  July 16 (Tabie 111.21), appeared either equally distributed (20-40%) among these species (July 15, 0 m) or was dominated (93%) by Peridinium sp. (July 16, 0 m). Following the increase in irradiance, cell volume was concentrated in the cryptomonad fraction, ranging from 52% (July 19, 1 m) to 87% (July 18, 2 m). Though this cryptomonad dominance was accompanied by a volume increase of « 50%, most of this shift in relative abundance was due to large declines in Peridinium sp. and M. rubrum . Consequently total cell volume between the two periods did not differ appreciably, ranging from 6.3 • 10  5  / i m / m l to 3  1.4 • 10 / i m / m l prior to July 17 and from 7.5 • 10 /xm / m l to 1.5 • 10 /tm /ml after 6  3  3  3  6  3  July 17. The precision in the seston volume estimates is somewhat less than for either May/82 or Aug/82 due to the generally low cell densities. For this reason, this microscopic enumeration data could not be usefully augmented by the Coulter counter data since they were even more seriously affected by the low volume specific counts. Due to the location of the July 19 (drift stn 13) and July 20 (drift stn 15) sample locations (Figure 3.38), the associated species enumeration data can be used to interpret the daily appearance of the fluorescence patch at fixed transect stations 1-6 in terms of species dominance. Only the drift stn 13 sample time coincided with a fixed station transect of the patch (i.e. T14). However, the drift stn 15 sample is inferentially useful as well under the assumption that the diel periodicity in patch appearance continued through July 20. Assuming this and based on the species composition at these stations (Appendix G), the cryptomonad was probably the predominant contributor to thefluorescencepatch.  8.4.8.5 Chla and nitrogen budgets A comparison of predicted and observed Chla and nitrogen flux for J1-J4 is presented in Tabie III.24. During the first 24 h period of J l , ambient nitrogen declined 50% less than predicted, and the measured Chla increase was only « 5% of the expected increase, while during the second 24 h, ambient nitrogen actually increased and Chla decreased with the corresponding prediction of opposite sign in each case. Time series J2, J3, and J4 which 152  Table 111.24 July/83 Chla and nitrogen mass balance. Predicted N-uptake (mgN/m )  Measured dN (mgN/m )  Predicted Chla -prod. (mgChla /m*)  Measured dChla (mgChla /m*)  Jl 2-4 4-6  65 102  -25 25  32 36  1.7 -1.7  J2 8-9  8  -59  3  -3  JS 10-11  10  -92  4  5  J4 13-14  124  50  44  -24  Total  309  -101  109  -22  Drogue stn Interval  a  8  span less than 24 h were normalized to the total daylight hours during the interval using a sunrise approximation of 0600 h local time. For 32 and J3, predicted nitrogen uptake was a factor of « 10 less than observed. For J3 predicted and measured Chla production were in fair agreement, while for 32 the measured loss was about equal in magnitude to the predicted production. During J4 Chla declined inspite of high production predicted on the basis of the carbon uptake, while the predicted decline in nitrogen, in a consistent fashion with the Chla production overestimate, exceeded the observed disappearance by a factor of « 2. Herbivorous zooplankton distribution was not estimated during the July/83 cruise. However, the grazing index along with the grazer community structure inferred from the ICM zooplankton monitoring survey can again be used to estimate grazing pressure during the J1-J4 time series. The June/83 ICM survey was dominated at night by Pseudocalanus minutus and A. clausii , making up > 90% of the samples by wet weight. The daytime was dominated by A. clausii , though P. minutus also contributed significantly (ICM 1984). Since P. minutus and A. clausii are both « 1 mm in length (Gardner and Szabo 1982), a representative body size dependent daily clearance rate for the herbivore population was estimated to be w 3.0 ml/d based on the statistical model of Peters and Downing (1984). The calculated grazing loss for all the drift stations is presented in Tabie HI.25. 153  Table 111.25 July/83 24 h grazing pressure estimated from the £ Phaeopigment/£ Chlo ratio. drift stn  £ Phaeopigment/ £ Chlo  Euphotic zone Copepods (xlO )  Grazing (mg/m d)  excretion (mg/m -d)  0.77 0.50 0.65 0.56 0.62  4.9 3.3 5.3 3.8 4.0  0.9 0.6 1.3 0.8 0.8  2.6 2.0 2.2 1.4 1.4  0.86 1.10  9.5 8.4  3.1 1.9  5.2 3.2  0.70 0.79  13.4 18.9  7.7 13.6  13.2 23.3  0.43 2.60  13.7 20.3  13.1 4.8  22.4 8.2  4  Jl 2 3 4 5 6 J2 8 9 J3 10 11 J4 13 14  Phaeopigment/  2  a  Chla was from one to two orders of magnitude greater during  July/83 than during May/82 or Aug/82. Assuming that this is not attributable to the concurrent switch from the spectrophotometric to fluorometric phaeopigment/Chla determination procedure in July/83, it implies a unique grazing regime. With the increase in Chla after July/16, this inferred grazing pressure becomes an order of magnitude greater than during Aug/82, reaching > 13 mgChla / m d at drift stn 11. Nevertheless this sink 2  is insufficient to account for the large Chla defect (TaWe 111.24), thus implicating an important diffusion and/or sedimentation flux contribution. An effort was made during the July/83 cruise to estimate particle sedimentation flux on a particle volume basis using the SETCOL technique (Bienfang 1982). Due to the low particle concentration during the July/83 cruise and the nature of the volume specific counting approach used by the Coulter counter Model B, counts for the larger size fractions were often low (i.e. < 50). Therefore, to estimate statistically significant changes in channel counts before and after the sedimentation interval (Bienfang 1981), confidence limits based on a Poisson distribution (Pearson and Hartley 1966) were calculated. If the before and after count difference did not exceed the combined confidence band, these data were discarded. The resulting volume specific sinking speeds for positively and negatively buoyant 154  Table III.26 July/83 volume specific sedimentation rates (m/d) for the positively (Vj,) and negative buoyant (Vs) fractions. (D) and (N) denote day and night sedimentation experiments respectively. Vol bins expressed in 10 t»m* ; iterated estimates are in parentheses. s  Vol. Bins 10 /im 20 10 5 2.5 10 5 2.5 1.3 s  drift stn  80 40  40 20  s  Negatively Buoyant Fraction (V< j in im/d) 2 (D) 1.8 .4 .8 3 5  (N) (N)  6  (D)  9  (D)  10  (N)  11  (D)  12  (N)  2.4  13  (D)  .6  .7 .7  .2  1.7 .2  .8 1.3  1.7  .8  (2.4) 1.3  1.2  .7  .6  .6 (1.5)  Positively Buoyant Fraction (V In m/d) 2 (D)  1.3 0.7  0.7 0.4  0.4 0.2  0.2 0.1  .4 (.8)  .2 (•«)  .2 (-4)  .2 (-5)  1.1 (2.5) .9 .1 .3 (.4) .2 (•4) .2 (.6)  1.2 (3.0) .3 .2 (1.0) .8 .8 (1.9) .2 .2 (.5)  .1 (6) .8 (1.5) .4 (.8) .1  .5 (.9) .1  .2 (.3)  .2 M)  b  5  (N)  9  (D)  10  (N)  11  (D)  12  (N)  13  (D)  1.3 (2-4) .7 (1.2)  .4 (7) 1.0 (1.9)  .1 (•2) .2  (•«)  .3 (-6) .2 (•4)  .2  (••*)  1.4 (2.6) (•9) .6 (1.1)  .8 (1.6)  .4 (-7) .6 (1.0) .3 (.5) .0  .4 (•7)  particles (with iterated upper limits in parentheses) are tabulated in TaWe 111.26. In general, for cases where  and Vs were statistically non-zero, the iterative calculation  usually increased the original estimate of Vs and  by a factor of two.  Vs was usually < 1 m/d overall and < 0.5 m/d for the < 5 1 0 / . m size fractions. The 3  3  iterated upper limits were 2 and 1 m/d respectively. Drift stn 5 (nighttime occupation) was an exception with Vs w 1.2 m/s (iterated upper limit of 3 m/d) for the < 5 • 10 pm  3  3  fraction. There was no evidence of a day/night periodicity in Vs or V^, nor was there  evidence of a determinant volume specific sinking rate. Since some volume dependence in 155  sedimentation is expected (Smayda 1970, Bienfang and Harrison 1984), the lack of this dependence here may be due to the obfuscating effect of the low number of estimates for the > 1.0 • 10 / i m volume fractions. In any event, the < 2.5 • 10 / . m volume fractions 4  3  3  3  were in general the dominant components of the seston and hence a general upper and lower bound for the sedimentation rate of 0.5-1.0 m/d appears reasonable. Vs and Vj, co-varied strongly, attributable to the coupled nature of the algebraic expressions for these variables. Therefore it is not clear how to interpret the significant positively buoyant seston fraction. Ignoring this apparent positively buoyant component, the depth averaged phytoplankton concentration was used to estimate an upper limit of 8 mgChla / m for the sedimentation flux contribution to the Chla mass balance. 2  Using the same approach in estimating the diffusion flux as for the Aug/82 time-series, a range in c was determined that tended to minimized both the Chla and nitrogen defects. The required estimates of the vertical gradients in Chla and ambient nitrogen, and the Brunt-Vaisala frequencies for the J1-J4 drift stations are tabulated in Table in.27, while the calculated fluxes are tabulated in TaWe HI.28.  Table 111.27 J1-J4 vertical gradients in Chla and total nitrate, and the Brunt-Vaisala frequency (N) us to estimate the diffusivefluxesin Chla and nitrogen. Time Series Jl  J2 J3 J4  Date  July 13 14 14 15 July 16 17 July 17 18 July 19  drift stn  dChla /dz  dN/dz  N  (mg/m )  (mg/m )  (xlO-'s" )  2 3 4 5 8 9 10 11 13  0.2 0.4 0.5 1.0 0.3 0.2 1.4 1.7 1.1  3  3.2 3.1 3.6 4.1 6.4 5.6 5.1 7.4 4.8  4  4  -  18 19 6 8 21 34 35  1  As in the case of Aug/82, the inferred upper limit of K was probably between 5 - 1 0 and -5  z  5 • 10  - 4  m /8, though the potentially large grazer contribution to both mass balances not 2  included in the foregoing calculation (i.e. the large grazer component inferred from the ICM zooplankton species enumeration record) reduces the precision of this range estimate. 156  Table 111.28 July/83 Chla and nitrogen detects (Ce and Ne), and Chla and nitrogen diffusionfluxes(Cf and Nf). (-) indicates missing mass. drift stn  Ce mg/m  e = 5- 10" m /s Jl 2-4 -30 4-6 -38 J2 8-9 -6 JS 10-11 -1 J4 13-14 -38 7  Total  2  6  Total  2  Ce+Cf mg/m  2  2  Ne mg/m  2  Nf mg/m  2  Ne-Nf mg/m  2  5•10" m /s)  s  -143  i = 5- lO" m /s Jl 2-4 4-6 J2 8-9 J3 10-11 J4 13-14  Cf mg/m  3  5  2  3 4  -27 -34  40 127  5 5  35 122  0.4  -5.6  -51  1  -52  4  3  -82  4  -86  2  -87  174  10  164  13.4  -130  208  25  183  5-10-* m /s) 2  s  -30 -38  33 41  3 3  32 127  52 58  -12 69  -6  4  -2  51  10  -61  -1  42  41  82  38  -120  -68  19  -70  174  102  52  -143  139  -5  208  260  -72  For the lower K , neither the Chlo nor nitrogen defect was adequately accommodated z  while for the higher K the Chla defect was accounted for while the nitrogen mass balance z  was left slightly in excess. This contrast between the two budgets is too large to be reconciled by the Chlo sedimentation flux alone. As in the case of the relative mass balance defects for Aug/82, the difference can be interpreted as evidence of a significant entrainment contribution to the mixing flux. Alternatively, and assuming the difference between the two is not an artifact of the imprecision in Chla and N O J gradient estimates, fecal pellet flux with partial or total re-mineralization of fecal nitrogen could provide the necessary balance. This estimate of a mass balance conserving grazing flux was facilitated by assuming that the diffusion regime in July/83 was the same as in Aug/82 for which grazing index inferred grazing pressure was relatively low, and hence the precision of the diffusivity estimate correspondingly high. The defect remaining after calculating the appropriate diffusion and sedimentation flux was interpreted as the component under-representation 157  in the phaeopigment grazer index. The flux estimates (i.e. ignoring for the moment the contribution from the unrepresented large grazer fraction) and their contributions to the mass balance defects are presented in Tabie 111.29. Table III.29 July/83 Chla and nitrogen mass balance detect accounted for by each of the source and terms. Source/Sink  Chla Flux (mg/m )  N Flux (mg/m )  % Defect Chla  % Defect N  18 8  31  13 6  7  13 139 26  25 260 50  2  Grazing Sedimentation Diffusion 5- lO^m'/s 5- 10- m /s 4  2  l-10- m /s 4  J  The combined 1 • 1 0  2  -  - 4  -  19 97 18  12 125 24  m /s diffusion flux and sedimentation flux for Chlo (i.e. total= 2  34 mg/m ) represents 24 % of the mass balance defect. In an encouraging fashion, 24% 2  (i.e. 50 mg/m ) of the N-defect is accounted for by the diffusion flux. The remaining defect 2  in each case is attributed to the grazing sink/source (i.e. « 75%). Since the phaeopigment inferred component accounts for « 15% of this, the unrepresented grazing component is assumed to have contributed 60% to the Chlo and nitrogen defects. As was mentioned previously, the under representation of the large bodied grazer fraction can be attributed to characteristically large, fast sinking fecal pellets removing the grazer generated phaeopigment from the water column (Welschmeyer and Lorenzen 1984). This rapid disappearance from the surface layer must also affect the remineralization of fecal nitrogen. This nitrogen sink can be estimated by re-expressing the biomass grazing flux on a nitrogen mass basis. Assuming a N:Chla of 3, the large bodied grazer component of the Chla defect (i.e. 60% or 79 mgChlo /m ) is equivalent to 235 mgN. Since this is a 2  factor of 2 greater than the excess nitrogen in the mass balance (Tabie 111.28), an estimated 50% of the ingested nitrogen was assumed lost to grazer and/or fecal biomass.  158  8.4-3-6 Inferences about SCM formation Low light conditions during the early part of the July/83 cruise (i.e. July 13-15) and an inferred large grazing pressure that limited Chla specific biomass accumulation after light conditions became more favourable for growth, appeared to limited primary production so that the surface layer was generally nutrient replete throughout the cruise period. Nevertheless, instances of subsurface peaks in P  B  did occur (drift stn 9 and 11, Figure 3.50)  during the first three days following the abrupt increase in  on July 16. In the absence  of nutrient depletion, this vertical uptake distribution must be attributed to inhibiting effects of the high  on shade adapted phytoplankton. The return of the P  B  profile to  a monotonic decline with depth on July 19 (i.e. drift stn 13) under ambient nutrient and irradiance conditions similar to those for drift stn 9 and 11 suggests photoadaptation by the phytoplankton had occurred over this time interval.  Inspite of the subsurface peaks in P  B  during drift stn 9 and 11, Chla specific biomass  decreased monotonically with depth. An explanation for this is unclear, though the pattern is consistent with the evidence of a significant (but equivocal) positively buoyant seston fraction in the SETCOL data. Alternatively reversing the incubation artifact argument used to rationalize P  B  distribution during May/82 and Aug/82, the subsurface P  B  peak  could have been the result of constraining the normal vertical displacement of the phytoplankton during the  1 4  C incubations, thereby enhancing the near surface photo-inhibitory  effect otherwise ameliorated by some depth averaged near field light flux (Harris 1973; Marra 1978). However, this latter interpretation is inconsistent both with the appearance of a SCM during drift stn 13 (Figure 3.49) and the indication in the May/82 data that a near surface vertical density gradient less than half of that present during July/83 (and hence a correspondingly greater inferred vertical exchange rate) was not enough to smear the appearance of depth dependent P-I characteristics. Assuming the physiological process time scales are comparable, it is unlikely that a sufficient vertical displacement and/or displacement rate was operative during the statically more stable July/83 water column, that would reduce the photo-inhibitory effect of supersaturating irradiance. 159  While the depth of the SCM at drift stn 13 was 1 m, the P  B  peak at drift stn 9 and  11 was at » 4 m. Hence these features are probably not directly related. On the other hand, the dominant motile phytoplankter suggests a behavioral response underlying the S C M formation, though the appearance of the SCM only after the implied adaptation to the increase in Ig is problematical. One possible explanation that is also applicable to the monotonic Chla distribution during drift stn 9 and 11 (i.e. inspite of the associated subsurface peak in P  B  ), and perhaps also to the periodic column stability/Chla specific  biomass covariance inferred in chapter 2 concerns the possible role of a 2-layer surface flow entrainment flux. Though no direct evidence for a buoyant tidal flow regime was obtained, the higher precipitation during the cruise period ( « 5 mm/d) compared to Aug 11 (i.e. where direct evidence of such a flow was obtained), implicates this flow in the July/83 surface layer mixing regime. On the further assumption that the inconsistent mass balance evidence was an artifact of the imprecision in Chla and N O J gradient estimates, it is conceivable that an appropriate entrainment flux up into the surface layer could have acted to upwell the Chla biomass that would otherwise have accumulated vertically according to the P  B  profile. The suggestion of an entrainment interface at 1-3  m is consistent with this interpretation as is the approximate depth of the P  B  maximum  at drift stn 9 and 11 (i.e. 4 m).  S.4.S.7 Summary The horizontal salinity and temperature fields in Holberg Inlet were observed to respond in a uniform manner to irradiance and precipitation/run-off forcing, while the phytoplankton field (i.e. inferred from in vivofluorescence)was not. Instead, a phytoplankton patch was observed in the down-inlet region of Holberg Inlet which demonstrated a periodicity coherent with the photoperiod. Both a residual circulation and vertical mixing argument appeared to be plausible coupling mechanisms, while a horizontal mixing argument was not. Low irradiance early during the cruise period, and an inferred high grazing pressure 160  later when growing conditions had improved were suggested to be the primary factors controlling primary production. As for the Aug/82 time series, the Chla and NO3 mass balance calculations were consistent with an entrainment contribution to the vertical fluxes, though unlike Aug/82 there was no direct evidence demonstrating the existence of a near surface 2-layer flow. The Chla and N O 3  mass balance also indicated that the phaeopigment index estimate  seriously underestimated grazing pressure. A large bodied zooplankter was proposed to comprise this extra grazer component, consistent with the relative N O J and Chla defects as well as with a high fecal pellet sinking rate explanation for the biased phaeopigment index.  3.4.4 General Discussion It was suggested earlier that identifying underlying dynamics in the observed SCM distribution would help clarify the coupling of primary production to the vertical mixing regime. However, as was pointed out by Cullen (1982) to be generally the case, the various SCM features observed over the course of the three cruises can not be attributed to a single common underlying mechanism. During May/82, an SCM was observed at the approximate saturating irradiance depth (i.e. 4 m), below a nutrient replete surface layer, with the P  D  profile decreasing mono-  tonically with depth. This implies a behavioral response to super-saturating light, with the inconsistent P  B  profile either an incubation artifact or a reflection of a strong vertical  contrast in light adaptation and respiration types. Unfortunately, direct species enumeration of the S C M and adjacent regions was not carried out. However, two motile species (Af. rubrum and an unidentified cryptomonad) were consistent contributors to phytoplankton biomass during May/82, with the cryptomonad apparently the more important. However, Af. rubrum is known as an active vertical migrator and could conceivably account for the inconsistent biomass distribution on May 17, since a comprehensive explanation must rationalize not only a strong depth dependency in P-I characteristics, but also a Chla spe161  cific biomass distribution that does not parallel the distribution of P  B  . A highly motile  species such as M. rubrum could provide the latter due to its motility and associated high respiration; either or both of which would modify depth specific net primary production. By comparison, during Aug/82 a subsurface Chlo maximum (SCM) evolved that conforms to the pattern observed in deep oligotrophic mixed layers (Cullen 1982), with the S C M located below a subsurface primary production maximum (SPM) that itself coincides with the nitracline. An interesting aspect of this relationship between the SCM and SPM is that it was time-dependent. The S C M quickly reached a steady state depth, while the SPM descended from above and eventually coincided with the S C M only after several days. The steady state SCM implies either a depth dependent sinking rate, behavioral response by motile phytoplankters, or a dynamic steady state between depth dependent growth and diffusion, and depth independent sinking. Though SCM formation during Aug/82 was different from May/82, it is not clear to what extent a behavioral (i.e. locomotion) response contributed given the dominant chrysophyte contribution to SCM biomass on the one hand and the S P M / S C M spatial relationship time course on the other. There is little evidence of a nutrient dependent sinking rate in non-motile phytoplankton, particularly as far as nitrogen nutrition is concerned (Bienfang and Harrison 1984). Furthermore, as was demonstrated by Steele and Yentsch (1960), sensible gradients in density cannot modify buoyancy to the point of significantly affecting sinking rate. The descent of the SPM was not obviously associated with a deepening nitracline. However, the corresponding point in the time series was marked by an apparent wind mixing event, which while not obliterating the SCM did re-distribute significant numbers of the phytoplankton into the 0-4 m surface layer as well as apparently cause a shoaling of the nitracline. Both effects would have obscured a depth correspondence between the S P M and the N O J distribution. An S C M was observed only on one occasion during July/83. It was shallow (i.e. 1-2 m) and appeared to be accompanied by a depth corresponding SPM, though large variance 162  in the surface carbon uptake estimate makes this coincidence equivocal. As in the case of May/82 and Aug/82, the S C M consisted predominantly of a motile phytoplankter (an unidentified cryptomonad) with a secondary contribution by a small-celled diatom (in this case S. costatum). The shallow depth of the single occurrence is presumably related to the low light/high nutrient regime prevalent during the first part of the July/83 cruise period and a large grazing flux inferred both on the basis of the high phaeopigment grazing index and the Chlo /nitrogen mass balance defect following the improvement in growing conditions. Both these factors may have acted to limit primary production, and thereby maintained comparatively high near surface ambient nutrient concentrations throughout the July/83 cruise period. There was no substantive evidence on which to base an explanation for the monotonic Chlo distribution inspite of a subsurface peak in P  B  . However, it is speculated that the  relatively high precipitation during the period may have supported a unique positively buoyant flow regime which could have affected the vertical distribution of Chlo , though again the evidence is equivocal. By more direct inference, the May/82 and Aug/82 cruise periods appeared to provide a contrast in the effects on primary production in the presence (Aug/82) and absence (May/82) of a positively buoyant tidal exchange. Inspite of this implied added mixing component in Aug/82 (though concurrent with a slightly reduced wind stress regime), stability conditions were such that a structured vertical Chla specific biomass and N O J distribution evolved, with nutrient limited growing conditions present at the near surface. Though the presence of a buoyant tidal flow regime during July/83 can only be inferred from the comparatively high run-off regime, the observed periodic horizontal Chla patch appearance in July/83 and its restricted horizontal distribution could be related to the horizontal extent of a positively buoyant tidal exchange if the vertical mixing model best describes the underlying process. The vertical distribution of physical scalar properties in the basin appeared to respond uniformily in the horizontal to meteorological forcing, while Chla specific primary production appeared greater in regions of the basin proximal 163  to QN. In the absence of alternate explanations and given the suggestion of a near surface positively buoyant tidal flow limited in the horizontal, the qualitative difference in the relative proportion of entrainment and diffusion between down-inlet and headward regions of the basin could have affected the horizontal extent of vertical dispersive mixing so as to facilitate biomass accumulation down-inlet (i.e. where the upward entrainment flux would have been greatest).  164  4. Simulation Modelling of Basin Primary Production 4.1 Introduction A number of statistical and dynamical patterns emerged from the analysis of a prior long term (i.e. low resolution) data set in Chapter 2 and in the semi-quantitative time series analysis of 3 short term (i.e. high resolution) data sets in Chapter 3. To put some of these in better perspective, the relevant dynamics are explored in this chapter through numerical simulation modelling. In such a modelling endeavor, there are two contrasting philosophical approaches available to the modeller. The first, more rigorous approach of the two, emphasizes mathematical simplicity with the aim to as completely as possible understand the system behavior. Accordingly, the usual procedure for simplifying inherently complex ecological models is to scale the individual non-dimensionalized terms and eliminate those clearly unimportant to the solution (e.g. Wroblewski and O'Brien 1976). The second approach, by comparison, ignores this constraint to a comprehensive understanding imposed by the complex system formulation, striving instead to maximize model precision with the hope that the representation is sufficiently robust to make the inferences drawn from numerical experiments meaningful. Among other things, this blackbox approach can be worthwhile for investigations of the system response to different forcing regimes, whereby in lieu of further empirical detail (i.e. additional experimentation suggested on the basis of the original analysis) it affords self-consistent speculation about the underlying dynamics. It is in these latter terms that the modelling exercise presented in this chapter was persued, aware that it constitutes a preliminary step to a more rigorous analysis. The robustness of this type of modelling, by which is meant the degree to which drawn conclusions are independent of coincidental agreement between theory and observation, is increased if the model predictions are tested against more than one data set. Toward this end the modelling exercise consisted of two parts. First, the primary production model was tuned to the July/83 phytoplankton time series data set, consistent with the heat budget diffusion submodel tuning to a concurrent temperature time series of Holberg Inlet. Then 165  with a minimum of change to parameter values, it was used to predicted the Aug/82 time series. The success of this second prediction constituted an evaluation of this robustness. Second, the control of primary production by the dominant environmental forcing functions was evaluated by manipulating these terms and gauging the model response. The following dynamical questions were addressed: 1) Vertical transport affect on the photoadaption, and photo-inhibition response. 2) Importance of depth dependent growth and sedimentation in SCM formation. 3) Role of wind mixing and buoyancy flux regime in the annual primary production/column stability covariance pattern.  4.2 The model The phytoplankton time series data consists of three state variables: Phytoplankton biomass and two moieties of the limiting nutrient (NOJ and N H 4 ). Hence a suitable z,t model is a system of three partial differential equations coupled by the sink and source terms for these variables. However, due to the nature of the verifying data set, Chla constitutes the unit of biomass, while carbon the unit of growth. As a consequence, a Chla /carbon (u) and nitrogen/carbon (6) conversion factor is required in specifying this coupling. Hence the model consists of a system of five coupled partial differential equations, where u and 6 are treated as ordinary seawater constituents dChla d, 3 , . 3 . .» „ -—— = — [K —Chla) - w.—Chla + Chla{u(\ - X ) - \ G) at az az az z  p  ^ - S < * . S ™ ? > - a « J  du  d . du.  dd>  d ,  36  v  166  T  g  (  4  1  )  with K the eddy diffusivity, w, the phytoplankton sinking rate (i.e. xv, > 0 and z directed g  positive down); and A , A , and A net carbon assimilation, dark carbon loss, and grazing p  r  0  loss coefficients respectively; while G is grazer density (#/m ). These coefficients and the 3  local terms A -,  & +,  NO  A , and A ^ are defined as  NH  v  where  /*/ = ( ! - e~ l T) -^ aIV  p  lT •  v  p  t  A = P , • min _ Mm where ^ _ fm _ -k (Chla/if-g,) m  p  e  when  m  X )^p-e- t hNH  p  r  + = C7>/a(^(A - A ) - ^ G ^ V  A,  min  P > 0;  (A -  G  A„  + v- u )  otherwise  A „ - = t.Chla N j r  n  R  Ar =  A  / i = {u - v in)l{K  p  r  1  )  + ^ - ( 1 - e~  h  ™« )) +  1 + (A - A )i/* p  r  (J + ^ P - A r )  * + (A - A ) » . £ + A -A p  r  P  r  Ap is a switching function between the relative specific (i.e. from 0—1) light-limited (/*/) and cell nitrogen quota-limited (fi ) growth rate, and the Chla specific carbon uptake n  rate at optimum irradiance [P™). The underlying assumption of switching models is that only one resource is limiting at a time, differing from multiplicative growth models (e.g. Jamart et al. 1977, Wroblewski 1977) where two or more resources are assumed to be limiting simultaneously. The justification for the switching approach adopted here is the demonstration of nutrient/light interaction in phytoplankton culture studies, whereby over a finite range in nutrient and light conditions, the minimum (subsistence) cell quota of constituents such as nitrogen adjusts in such a way as to maintain a light-limited specific growth rate over a range of nutrient specific uptake rates (see below). Cell quota models have been largely unsuccessful in describing nitrogen limited growth due to the limited range between minimum and maximum quotas resulting from a relatively large inactive pool component, with the direct consequence being an unbounded quasilinear to linear growth rate to quota relationship (Goldman and M Carthy 1978, Caperon c  167  and Meyer 1972). Hence, it is felt that identification of an appropriate active cell nitrogen pool will ameliorate this lack of success (Droop 1983). Explicit estimates of specific cell nitrogen pools are not a component of the basin data sets. However, due to its nitrogen moiety, Chlo is in effect such a cell nitrogen component. This interpretation is supported in the observation by Caperon and Meyer (1972) that the relationship between Chlo quota and growth rate is hyperbolic in steady state cultures for a number of phytoplankton species. Therefore, in keeping with a perceived need to decouple growth from nitrogen uptake (i.e. data set evidence of undiminished carbon uptake in the absence of ambient nitrogen, chapter 3), and this observed Chlo quota dependent growth rate, internal control of nitrogen dependent growth is modelled here as a hyperbolic function of the subsistence cell quota (&w ) n  a n  ^ a saturating constant k . v  X is dependent on the maximum grazer filtration rate (F ) and a functional response M  G  coefficient HR parameterizing the depensatory decline in clearance rate with increasing food concentration above a threshold concentration (<&)• Furthermore, since Chlo comprises a very small fraction of phytoplankton cell biomass, the food density response function is more appropriately expressed in terms of cell carbon, with the necessary Chlo / C conversion factor specified by v. The dark respiration coefficient A is set to zero during daylight since the carbon uptake r  estimate in the data set is assumed to best approximate net assimilation. The local terms for N O J and NHj" are functions of the total nitrogen demand due to photo-autotrophic carbon uptake and the steady state growth rate dependent v (i.e. «/») and 6 (i.e. ^*). Additonally, &  N H  + is also dependent on the nitrogenous excretion  by grazers specified for simplicity as a constant of excreted ration fraction €R wherein 100% recovery of the associated nitrogen (defined by 6) is assumed. Interpretation of the July/83 time series suggested two grazer components, one tightly coupled and the other only partly coupled to primary production. This distinction is not explicity incorporated in the model, but rather is accommodated through the appropriate adjustment of 6R. 168  The partitioning of the phytoplankton nitrogen demand between the two moieties is a function of their relative contributions to the total ambient nitrogen pool, modified by the inhibitory effect of N H j on N O J uptake (e.g. Conway 1977, M Carthy et al. 1977). c  This latter preference for NH4 is attributed to the greater energy efficiency in assimilating a more reduced compound (Collos and Slawyk 1980), with the response an exponential function of ambient NH4 concentration (Wroblewski 1977, after Dugdale and Walsh 1972). Therefore it is parameterized by the first-order parameter h. The local terms A„, A ^ are functions of the steady state growth rate dependent ratios (v*,<f>*)  and the growth rate dependent approach of v and <f> to u* and <(>,. The growth  rate dependence of v —• v* and <f> —* tf>* is viewed here as the result of new growth characterized by v*,<f>* diluting existing cell biomass characterized by v,<f>. In an Eulerian reference frame v —• v* d> —» <f>* is also dependent on vertical mixing and cell sinking y  and hence the total time-dependence of v <f> is specified according to the respective partial t  differential equations in (4.1). The interaction between the light and nitrogen regime in determining steady state 1/* and 4>* is such that the subsistence cell Chla and nitrogen content decrease with increasing light levels (Laws and Bannister 1980, Rhee and Gotham 1981, Droop et al. 1982). This is considered to be an adaptive response by the cell enabling it to maintain a constant cell-quota specific growth rate p independent of external nutrient concentration, since s  p, = p / A , where p m  m  is the maximum (constant) nutrient uptake flux, and A, the cell  nutrient quota (M Carthy and Goldman 1979). Since a time-dependent N O 3 c  depleted  surface layer is a recurring feature in the cruise data sets, such a physiological response whereby growth is decoupled from nutrient uptake is probably important in maintaining the observed near surface carbon uptake rates. This dynamic is incorporated in the model by calculating i/« and <p* at the prevailing light regime assuming nutrient replete conditions prevailed (i.e. the specific uptake rate V = pi), where V = Nr/{k  n  + NT) with k  n  the  ^-saturating constant for uptake and NT total ambient nitrogen, with the y-intercept of V  *\$*  — /(MJ)  representing the steady state minimum cell quota at the prevailing light 169  regime (i.e. *t i i4>'min)' *>$* v  a  r  m n  e  *ben specified as either functions of V or tn depending  on whether actualized growth is light or nitrogen limited. The nutrient limited growth dependence of v*, 6* is specified as a function of V rather than fi since balanced growth n  requires that V = fi and hence it is the specific uptake rate and not the specific growth t  rate which defines v*,<f>* during nitrogen limited growth (Eppley 1981). Thus J W E - aif*i,  if Hi < V ,  !  {  'min + (^max - a - u )V,  otherwise; with u'  u  x  min  min  <t>max ~ 0 M/»  • = u (\ - m) + v m max  min  if Hi < V,  2  + ( max - and a -a6 parameters )V, otherwise; with 6' the = literature 6 (l - m) with i/ intmin 4>min, fli, specified from and+ f 6 , m ^r 6  6  3  m  2 min  min  max  m a xmin  specified by the initial conditions (i.e. nutrient replete, light saturated u*,6 ). t  Boundary conditions for (4.1) require zero flux through the sea surface, while at depth a zero net flux of Chla and fixed N O J and N H j concentration are assumed. Therefore „ dChla K— dz  „ , ^ w Chla = 0 L  z  n  t  oz OZ du + A„=0 dd>  dChla K— =0 oz du oz 86 at z=0 K ^-=0 > , oz 8N07 K — = Constant oz 8NH+ „ K —-—— = Constant oz Tjr  z  z  tjr  at z=D (4.2)  n  z  rjr  z  4.3 Forcing Functions 4.3.1 Time series Three external forcing terms close the model: Incident irradiance Jf, grazer density G, and the vertical eddy diffusivity K = f(z,w,v), where w and v are the surface wind z  speed and tidal speed respectively. To validate the model against the Aug/82 and July/83 (i.e. J1-J4) Chla time series, Jf forcing was provided by the concurrent  record from the  pyrheliometer located in Rupert Inlet (Figure 2.8) converting to / J , while w and v were 170  provided by the surface wind records and a barotropic tidal prediction respectively. The irradiance at depth ~» k  e  z+kc  Jo  ehta  () z  needed to determine  dz t  w  t  e r e  k  b  a n (  j fc ^ c  was calculated according to  = JJ •  the background and Chlo specific light extinction  coefficients respectively. Due to the comparatively long generation time of the grazer compared to the simulation time domain, constant G was assumed, with an appropriate estimate inferred from the £^ Phaeopigment/J^Chla ratio in the validating time series data set and in the case of the July/83 time series adjusted to accommodate the inferred grazer biomass underestimate.  4.3.2 Vertical Diffusion (Heat Submodel) The vertical diffusivity forcing is dependent upon tide, wind, and column stability interaction which must be suitably parameterized in the absence of observation. To add robustness to this parameterization, the least squares fit of a heat budget simulation to a temperature profile time series obtained during the July/83 cruise was optimized through the grid-step estimation of two empirical fitting parameters. Ignoring for a moment radiant energy flux, the distribution of temperature T can be described as f  = 7 - ( 4 v r ) - 7  VT  (4.3)  where V is the advection velocity and v the gradient operator. The assumption was made that as in the case of the biological variables, the temperature record was obtained in a horizontally Lagrangian reference frame, thus eliminating this derivative in (4.3). Furthermore, to eliminate the advection term in (4.3) entirely, it was assumed that it could be represented as an apparent vertical diffusivity component. This approach was deemed prudent in the absence of a vertical advective flux estimate and seems justified since the equivalent vertical diffusion submodel forces the main phytoplankton model in an identical fashion. As was discussed in chapter 3, the effects of near surface vertical current shear, general tracking inefficiency by the window blind drifter, and discontinuities in the drogue track record (Figure 3.36, 3.38) worked against this basic assumption. Although 171  this presented a problem in interpreting the time series of biological variables, the effect was probably a minor one for temperature due to the comparatively uniform physical response by Holberg Inlet to meteorological forcing (Figure 3.42, 3.43). Thus equation (4.3) reduces to dT  d  l v  dT  s  subject to suitable boundary conditions. The surface boundary (again with depth directed positive down) was specified in terms of the radiant heating (Q ) and cooling (Qb), and r  the sensible (Q ) and latent heat fluxes (Qi) across the air/sea interface t  dT Kz— = (Qr + Qb + Q, + Ql)/p«,Cu, BT Kz4- =0, z = D oz  )  where p  w  z =0 (  4  5  )  is the water density and c its specific heat. The bulk heat fluxes Q , Qb, Q«, w  r  and Qi were calculated according to Gill (1982, p. 30) Qr  M  = k rl° • e-*** + ki{\h  r)I° • e~  hiZ  t  Qb = -0.985T,2*(0.39 - 0.05tf*)(l - 0.6n ) 2  Q» = c w c (T d  Pa  p  t  - T) a  Ql = c p vj(q - q )L t  a  t  a  where p , p , and c are the wind speed, air density and specific heat of air respectively, w  a  p  c, and Cd the Stanton and Dalton numbers and r, Stephan's constant; L the latent heat of water, kh andfc,-the radiant energy attenuation coefficients for infrared and visible light respectively, r the infrared fraction of i j ; and q , q the specific humidity at saturation 9  a  (i.e. sea surface) and at standard height respectively (i.e. in this case « 20 m above sea level) according to q = 0,662 J /F, where the vapor pressure ,  P = pRT a  a  r  = p RvT and air pressure v  (Haltiner and Martin 1957). R is the gas constant for air (-ft ) and water a  vapor (R„) and T the temperature in K for the sea surface (T ) and at standard height t  (T )- n is a cloud cover correction factor for back radiation, equal to 1 when cover is 100% a  c  and equal to 0 when the sky is clear. A net zero heat flux was specified for the bottom boundary. 172  The model was fit to the temperature data primarily by means of two fitting parameters. The first is taken from Sundaram and Rehm (1973), where £ scales K to a depth z  independent tidally generated (A ) and wind generated (A ) eddy viscosity through a t  w  depth dependent gradient Richardson number Ri( ) s  K  g{t)  = {At + A ){l + w  Ri ) {z)  At was assumed to be a linear function of the barotropic tidal speed t; (Bowden 1967) and the maximum eddy viscosity A such that A —» A t  tmtI  = v A €,  tmai  when v —» C  p  m a i  where v is a proportionality constant  tmail  p  , with At  mai  specified as a free parameter.  The component of the eddy viscosity generated by the wind A was calculated accordw  ing to Sverdrup et al. (1942), who treated it as a step function of wind speed tD 1 • 10- tZ> , if v)< 6m • a" 4  3  1  4 • 10 tu , otherwise. _4  2  The proportionality constant for the higher wind speed regime is an order of magnitude greater than that derived by Kullenberg (1976), who attributes most of this discrepancy with Sverdrup (1942) to different drag coefficient estimates between the two studies. However, uncertainty arising from this discrepancy is probably unimportant here since the purpose of the submodel was to estimate the best eddy diffusivity scaling parameter values ( and T for a constant (submodel) parameter space. To evaluate Ri( ) normally requires information on both the vertical density strucz  ture and the associated current shear, at profiles were obtained concurrent to the quasiLagrangian temperature record, with da /dz calculated by fitting a logistic equation to t  the profiles. However, due to a lack of detailed information about the vertical gradient in horizontal current, it was necessary to assume logarithmic boundary layers. Accordingly, the mean current (Vj*)) was expressed as a function of the free surface friction velocity u», the depth z and the Von Karman's turbulence parameter k (Townsend 1980, p. 138), with c a constant related to the inferfacial roughness length scale ln(z + c). 173  This logarithmic boundary layer approximation was considered unequivocal at the bottom boundary since the conditions for a constant stress layer should exist there (Townsend 1980, pp. 138-139). Expressing the friction velocity in terms of the bottom drag coefficient Cj,, the depth averaged tidal speed C, and differentiating, the vertical gradient in the mean tidal current V is v  dV dz> v  where z' is distance from the bottom, v was estimated using a barotropic tide submodel  where the tidal speed at time M s a function of the time derivative of tidal elevation L and the ratio of the surface area of the Holberg-Rupert basin (S ) to a representative a  Holberg Inlet cross-sectional area {S ). Areas S and S were calculated with the polygon e  a  e  area algorithm of CRC (1973) and x — y co-ordinates obtained from the electronically digitized bathymetry isopleths of a nautical chart of the basin (i.e. Chart #3617, Canadian Hydrographic Service). The tidal elevation derivative L was evaluated by finite-differencing tidal elevations predicted from a harmonic tidal model (Sec 2.1.2.1). Unlike the bottom boundary layer, the density gradient at the surface is likely to result in an interior stress gradient large compared to its rate of generation at the wind-shear interface. Consequently, the appropriateness of a constant stress boundary approximation there cannot be assumed (Townsend 1980, pp. 138-139). However, since the gradient was generally constant with depth during all cruise periods (i.e. no pronounced transitions between mixed and stratified layers, the vertical gradient in mean wind generated current V„ can be expressed as a function of the wind drag coefficient C , mean wind speed at t  standard height tD, and the second fitting parameter r which parameterizes this linear gradient in shear stress (Townsend 1980, pp. 180-184). Thus similarly  where z is the distance from the surface. 174  Combining the wind and tidal component, the Richardson number Ri( ) z  c a n  D e  ex-  pressed as  The wind speed record was obtained with a NOAA Skyvane propeller anemometer located in Holberg Inlet. While both A„ and u« are dependent upon the wind speed, their characteristic response time scales are assumed to differ. Hence a 5 h and 1 h moving average smooth of the wind record was used as an ad hoc forcing function for u* and i4„ respectively. The radiant heat flux forcing was provided by the same  time series used to estimate  If for the biology main model, while the relative humidity time series required to estimate specific density q was generated using a cubic spline interpolation (Moore 1984, p. 16) a  of twice daily relative humidity measurements collected by the Department of the Environment at the Port Hardy airport (Figure 2.1). Since these twice daily measurements were minima and maxima with no estimate of time of measure, it was assumed that the maximum relative humidity occurred near sunrise and minimum occurred several hours before sunset.  4.4 Numerical Solution (4.1) with boundary conditions (4.2) and (4.4) with boundary conditions (4.5) were integrated numerically by forward differencing in time and center differencing in space using the implicit Crank-Nicholson scheme (Smith 1965). To obtain second order accuracy at the critical air-sea boundary in the heat model, a center-difference evaluation of the boundary condition was used. The large diffusion gradient in the surface boundary layer also necessitated a small grid point spacing. However, to minimize the concomitant increase in computing effort, a depth dependent grid spacing was used. Best results were obtained with A z = 0.02 m for the interval 0 — 2 m, A z = 0.2 m for the interval 2 - 20 m, and A z = 2 m for the interval 20 - 100 m. At each time step the change in heat content 175  of the water column was compared to the heat exchange across the air-sea boundary (QT = Qr + Qb + Q» + Qi)- The time step was reduced until the observed heat discrepancy over the course of a simulation was less than 10° joules (i.e. At ta 0.1 h), which represents < 0.1% of the total surface heat flux over the course of the « 170 h simulation. A similar grid spacing was used in the phytoplankton main model, with A z = 0.02 m from 0-1 m, A z = 0.025 m for 1-10 m, and A z = 5 m for 10-100 m. In the case of the main model, the time step was reduced until the solution appeared to converge over the simulation time course (i.e. At = 0.05 h). However, to further minimize the computational effort, it was possible to restrict the vertical integration domain to the top 0-10 m of the water column without noticeably affecting the solution. 4.5 Results and Discussion 4.5.1 Heat Budget-diffusion Submodel The fixed parameter values are tabulated in Table I V . l , while the range bracketing the best estimates of the free parameters £ and T are tabulated in Tabie IV.2, with the residual from the fit presented as a total water column temperature variance (i.e. error mean square). Table I V . l Heat modelfixedparameter values Symbol Description  Value  Specific heat of water Stanton number Dalton number Stephan'$ constant Water vapor gas constant Dry air gas constant Ra wind stress coef. ct tide stress coef. cb IR atten. coef. k Vis. atten. coef. ki t IR fraction X non-dlmenslonal  4185 1.5 • 10" 1.1 • i o 0.985 462 287 2.5 • 10" 2.0 lO" 10.0 0.50 0.56  c. Cd r.  h  Units 8  s  s  8  Source  Holmbre et al. 1948 J/kg Gill 1982 X X X J/m'-deg Holmbre et al. 1948 J/m -deg James 1977 } X m" Ivanoff 1977 m" July/83 Ivanoff 1977 X 2  1 1  The best estimates for £ and r were 1.5 and 0.0 respectively. Since a logarithmic air/sea boundary layer approximation is a poor one on theoretical grounds, the 0.0 estimate for T is surprising and may reflect an underestimate of A„ and/or u«. However, as in the 176  case of the air/sea momentum transfer coefficient, this inconsistency is not considered consequential.  Table IV.2 The temperature residual tor each drift sta (2-14) (i.e. total column error meansqr) be heat model prediction and temperature time series for values of £ and r (10~ m~ ). i  1.25 1.50 1.75 2.00 1.50 1.50 1.50 1.50  1  r  2  3  4  5  6  7  8  9  10  11  12  13  14  Total  0.0 0.0 0.0 0.0 0.2 0.4 0.6 0.8  .000 .000 .000 .000 .000 .000 .000 .000  .010 .008 .009 .012 .008 .008 .008 .008  .065 .037 .020 .012 .033 .030 .026 .023  .137 .116 .104 .096 .114 .112 .110 .108  .082 .140 .186 .212 .149 .158 .167 .176  .013 .038 .067 .091 .042 .048 .053 .059  .158 .260 .342 .398 .274 .290 .305 .320  .807 .999 1.08 1.09 1.02 1.05 1.07 1.09  .164 .057 .035 .062 .049 .043 .038 .036  .178 .087 .067 .089 .078 .070 .064 .060  .169 .080 .064 .086 .072 .066 .062 .060  .2 07 .09 0 .05 9 .07 2 .07 9 .07 1 .06 5 .06 1  .060 .126 .208 .273 .142 .159 .176 .193  2.050 2.040 2.242 2.495 2.065 2.103 2.144 2.194  Tabie IV.3 tabulates the depth averaged change in heat content accumulated over the course of the simulations for the T and £ parameter space, expressed both in terms of absolute heat change and as a % of the depth averaged measured heat content. A£T is very small for both parameters when averaged over the entire water column (i.e. < 0.1%). However more significantly, the model sensitivity to £ was over an order of magnitude greater than it was to r. Table IV.S Heat equivalent depth averaged residuals and as % of observed heat. r ld^m" 1.25 1.50 1.75 2.00 1.50 1.50 1.50 1.50  1  Residual Joules m~  0.0 0.0 0.0 0.0 0.2 0.4 0.6 0.8  60.5 60.4 63.3 66.8 60.8 61.3 61.9 62.6  Residual 3  %  0.524 0.523 0.548 0.579 0.526 0.531 0.536 0.543  Some of this can be attributed to the least squares artifact of a poor C T D representation over the top 1 m of the water column. However, this does not seriously detract from the inference that T is unimportant at any distance from the free surface and hence is not critical to the parameterization of phytoplankton/nutrient mixing, particularly near and across the nutracline. 177  Figure 4.51 plots the simulated temperature time series based on £ = 1.50 and r = 0.0, as well as the actual time series data for comparative purposes. As is evident, the heat model successfully reproduced the increase in water column heat content following the increase in radiant heat flux after July 16 (i.e. drift stn 7). However, it overestimated surface heating prior to July 16, particularly near drift stn 5 (which is characterized by a very pronounced pycnocline, Figure 4.40). However, replacing the corresponding <r profile t  with an interpolation of drift stn 4 and 6 did not significantly ameliorate the discrepancy, implicating the periods of low surface wind (i.e. < 0.2 m/s) and high irradiance (> 200 w/m ) during the drift stn 5-8 interval (i.e. 30-72 h, Appendix K) as the underlying 2  cause. Figure 4.52 plots the simulation-averaged depth profile of K based on the best estimate z  of £, T . < K > ranges from a high of 4-10~ m /s at the surface and below the pycnocline 4  2  z  to a low of 5 • 1 0  - 6  m / s at the base of the pycnocline (i.e. 2-5 m). This minimum in 2  the profile represents the effective limit of vertical diffusion into and out of the productive surface layer and is two orders of magnitude less than that estimated on the basis of the Aug/82, July/83 Chla and nitrogen mass balance. This large discrepency can be reconciled by the fact that the gradient in K near this minimum is large (i.e. 1 • 10 m/s) -5  z  and therefore the mass balance defect estimates are necessarily sensitive to the measured depth distribution of the phytoplankton and ambient nitrogen in relation to K . z  4.5.2 Displacement Time Scale Attenuation of Photosynthetic Processes Assuming that the heat model derivation of the K profile is reasonably accurate, its z  sensitivity to variations in wind and buoyancy flux forcing can be estimated to evaluate the role of mixing related attenuation of supersaturating irradiance and phot oad apt at ion effects in the seasonal change of sign of the column stability/Chia specific biomass covariance observed in the ICM data set. This is accomplished by assuming a Lagrangian vertical displacement variance Z j = 2K t (Denman and Gargett 1983). 2  t  z  An appropriate Z scale for this analysis is the light gradient e-folding depth which from 178  Figure 4.51 July/83 drift stn temperature time series (topf and heat model simulation (bottom). 179  K (m /s z  Figure 4.52 r = 0.0.  xlO' )  2  5  0.5  21.2  41.9  0.5  21.2  41.9  Simulation averaged K vertical profile best estimate (ie. £ = 1.5 and s  180  the mean measured July/83 column light attenuation was « 2 m. Accordingly converting the diffusion profile into the equivalent vertical displacement time scale (t) profile, gave a t ranging from < 1 h in the top 0.25 m and at depths > 10 m (and likewise the top 1 m and below 6 m for t < 10 h), to t > 400 h at the base of the pycnocline (Figure 4.53). By comparison, cellular time scales range from < 1 h for the time course of photo-chemical adjustments (e.g. Vincent 1980) and photosystem saturation (Marra 1978), to 2-10 h for P-I parameter and C/Chla changes (Lewis et al. 1984, Prezelin and Matlick 1980, and Marra 1980). Since t must be significantly less than cellular process time scales if the contrast in the corresponding displacement averaged light field over t is to be large enough to effect a photosynthesis process response, an important mixing related effect was probably restricted to the fractional top meter and near the bottom of the euphotic zone. Doubling the wind speed compared to the July/83 time series mean of w 2.5 m/s (i.e. 5 m/s) reduces t by a factor of « 3. (Figure 4.53). To see how sensitive t was to the seasonal pattern in the vertical buoyancy flux, the Ct time series forcing which was used to obtain Figure 4.53 was replaced with seasonally representative profiles. These were obtained from the July/83 time series and used under the assumption that the range in wind, tide, and radiant energy forcing during July/83 is representative of the entire year. The time series only spanned a few weeks in midsummer. However, it was marked by high runoff/low irradiance (i.e. typical spring/fall condition) and low runoff/high irradiance (i.e. typical mid-summer condition). Hence in total a representative at profile from each of these periods was viewed to represent the seasonal pattern in a qualitative way. The two at profiles used were representative of the contrasting periods in the time series and did not constitute extremes (Figure 4.54). Nevertheless, the corresponding contras't in displacement time scales was large, spanning 3 orders of magnitude at depths below 3 m (Figure 4.55), though as in the case of the at time series forcing, the displacement time scales were of comparable order to cellular process time scales (and hence likely to elicit a physiological response) only in the top fractional meter and at depths > 10 ,m, 181  Figure 4.S3 Vertical Lagrangian displacement time scale based on the July/83 heat model best estimate of the vertical diffusion profile (left) and for a 2-fold increase in surface wind stress forcing (right).  Figure 4.54 Contrasting <r profiles from the July/83 time series representing high runoff/low irradiance buoyancy flux conditions (a), and low runoff/high irradiance buoyancy flux copditions (b). ' t  O  TIME  (HRS) ( X 1 0 ) 1  It  22  0  TIME  (HRS) ( X 1 0  00  Figure 4.55 Vertical Lagrangian displacement time scale as in Figure 4.53 but forced by a high runoff/low irradiance (left) and low runoff/high irradiance (right) <r profile. t  17  2  )  34  thereby restricting a seasonal pattern in the buoyancy regime related mixing effect on photoadaptation/photo-inhibition processes to these regions of the water column. Based on these differences in the t profiles for different buoyancy and wind stress forcing regimes, it appears that nominal fluctuations in either can significantly alter the displacement time scales. However, in terms of important primary production regions in the water column, this significant response appears to be restricted to the upper 1 m. Below this depth, displacement time scales were always 3> cellular process time scales and hence small variations would not significantly affect these processes. This interpretation implies that the dominant anti-correlation between P  B  and column stability and probably  between P//stability as well, inferred from the I C M data set in chapter 2 was not related to mixing attenuation of physiological responses.  4.5.3 July/83 standard run Table IV.4 tabulates the parameter set for the simulation that predicts most, though not all the essential features of the July/83 time series (Figure 4.56). It successfully predicts the increase in Chlo from w 1 mg/m on day 1 to > 8 mg/m 3  3  on day 6, with the vertical distribution remaining monotonic with depth except on day 6. However, the simulation overestimates production prior to the step increase in irradiance on day 4 and predicts a SCM somewhat deeper than in the data time series. Near surface NOJ  and NH^ concentrations also approximate the data set, decreasing from > 6  fiM total nitrogen on day 1 to < 1 fiM on day 6; though the model slightly overestimates both the degree of depletion and the vertical extent of the nitrogen depleted region. As in the data time series, the carbon uptake and P  B  profiles are monotonic with depth from  day 1 to day 3 and have a subsurface maximum on day 4 and 5 following the increase in irradiance. However to model the appearance of this subsurface peak in P required to force its time dependency via a time dependent 0.000, when t < 72, 0.010 - (t - 72)/2500; otherwise and with P > 0 185  B  , it was  Figure 4.56 Standard run simulation of the July/83 phytoplankton time series based on the parameter set in Tabie 4.4. The profile sequence consists of snapshots on one day intervals, with the solid line representing the initial condition. Dash length increases with time.  0  5  DEPTH 187  (M)  10  DEPTH (M)  10  03 m  DEPTH (M)  10  CJ  S3 a o O U9 0)  DEPTH (M)  188  Table IV.4 July/83 standard run parameter vaiues. Symbol Description  Value  Units  Source  P-I parameter 0.040 May/82" t 0.010* P-I parameter 0 6.0 P-I parameter mgC/mgChla h pm *» N-uptake \ sat. 0.50 Eppley et al. 1969 Maximum clearance rate 10.0 ml/an imal-d Frost 1972 grazing \ sat. 200 mgC/m kR grazing threshold 200 n Qi Chla specific light atten. 0.016 m /mgChla Bannister 1979 kc background light atten. 0.500 mJuly/83 kb G grazer density 7 10 animals/m July/83" ration nitrogen recycled 0.25 *R Falkowski and Owens 1978 rd dark respiration frac. of P 0.10 h NO, uptake inhibition 1.5 Wroblewski 1977 t subsistence Chla quota 1 Laws and Bannister 1980 mg/gC 4>min subsistence nitrogen quota 30 zero order u» coeff. 22.1 a zero order 4>» coeff. 60.2 Chla quota growth i sat. free parameter K 5 July/83 eddy diffusivity scaling coeff. t 1.5 r boundary layer stress gradient 0.0 mm/d ay phytoplankton sinking rate w, 0.0 wind stress coeff. James 1977 C 2.5 10~ t tidal stress coeff. 2.0 10" t C max. tidal eddy viscosity free parameter m'/s 8 10" Aim., A .1 wind viscosity coeff.iP > 6 m/s 4 io-« s Sverdrup et al. 1942 fl> < 6 m/s 1 • -i 10 * adjusted over course of simulation ** parameter adjusted to Improve performance a  s  3  1  I  4  m  t  3  2  1  s  t  s  b  s  f mgC/mgChlo h/iEm -B | non-dimensional 2  with the zero order rate constant (2.5 • 10  3  mgC/mgChlo / * E m s h ) estimated from 2  2  the photoadaptation time course observed in Glenodinium sp. on going from low light (110 /xE/m s) to high light (540 ^E/m -s) (Prezelin and Matlick 1980). 2  2  189  4.5.4 Importance of depth dependent growth and sedimentation in SCM formation The model does not predict the return to monotonic profiles over depth for carbon uptake and P  B  observed on day 6 in the data set. However, this cannot be attributed to  the time dependent specification of /? since it is reduced to zero over the course of day 3. On the other hand, the model predicted only a small subsurface peak in P O C compared to a seston peak a factor of 2 greater than at the surface in the data set (Appendix K ) . Adjusting the Chlo cell quota dependent growth parameter i /  m t n  and k can be expected to v  only minimally affect the uptake profiles due to the depth independent nature of the near surface nitrogen supply by day 6. This is demonstrated in Figure 4.57, where a subsurface peak in carbon uptake still remains, though reduced slightly on day 6, following a reduction in k from 5 to 2 mgChla /mgC. Surprisingly, inspite of the resultant underlying increase u  in near surface uptake compared to at depth, the subsurface peak in P O C was increased over the standard run (Figure 4.57) following this parameter adjustment. There is not substantive explanation for this model response, though it is small compared to the observed POC distribution in the data set. In general, it appears the POC distribution cannot be adequately represented by growth parameter adjustments, indicating that active or passive vertical displacement was important. However, invoking a depth independent positive sinking rate of the magnitude estimated in chapter 3 (i.e. 0.5 m/d), or for that matter a sinking rate a factor of 4 higher, also does not adequately predict the POC distribution (Figure 4.58), with the poor performance of the latter being the early prediction of a subsurface POC peak. By comparison, nutrient dependent sinking rate  where w = 2m/d, kn, and NT are defined as before, successfully predicts the POC distri t  bution in the data set, though it also exaggerates the S C M slightly (Figure 4.59).  190  :  P(MGC/M3/H) 2  p-  7  5 3  -«  POC (MG/M3)  (X10 )  (XIO ) 1  Figure 4.57 Simulated July/83 carbon uptake (left) and particulate organic carbon (right) using the standard run parameter set, but with k„ reduced from 5 to 2 mgChla/gC.  1  .0  POC  (MG/M3)  15.0  (XIO ) 1  30.0  0.0 o  POC I — i  (MG/M3)  (XIO ) 1  21  10.5  i;  <  4  N  o"  to tx3  101606  0.0  15.0.1  (XIO )  30.0  0.0  10.5, (X10  Figure 4.68 Simulated July/83 POC time series using the standard run parameter set and a depth-independent phytoplankton sinking rate of 0.5 m/d (left) and 2 m/d (right).  21 l  )  Figure 4.59 Simulated July/83 phytoplankton time series using the standard run parameter set and an ambient nitrogen dependent sinking rate asymptotic to 2 m/d.  194  195  4.5.5 Model Robustness The robustness of the phytoplankton model was estimated by extending it to predict the Aug/82 (i.e. A l ) time series; forcing it with the corresponding Aug/82 irradiance, wind shear, and tidal time series. Inferences drawn from the £ Phaeopigment/ £ Chla grazing index and the concurrent ICM zooplankton survey in chapter 3 suggested that grazing pressure during Aug/82 was an order of magnitude less than during July/83, due both to reduced numerical abundance and to the smaller body size of the grazer population. However, with the appropriate adjustment of G, F , m  and £R faithful to this inference,  the model performed poorly, overestimating phytoplankton biomass accumulation by an order of magnitude. Instead, F  m  and CR values equivalent to the July/83 parameter set  as well as G = 3 • 10 copepods/m were required for the model to successfully simulate 4  3  the Aug/82 time series. Figure 4.60 presents this simulation on the basis of the Tabie IV.4 parameter set with the exception (in addition to G) that kt, was reduced from 0.5 m to 0.3 m  - 1  - 1  in keeping with the contrast in light attenuation between the two time series.  On the assumption that the alternative in improving model performance, through large adjustment of the P-I parameters or A , is not justified, the success of this parameter set r  in predicting the Aug/82 time series indicates that both the ^Phaeopigment/ £ C h l a grazing index and the diurnal zooplankton survey underestimated grazing pressure during Aug/82. The possibility of this occurrence is increased if the grazers were large/fast swimming individuals, due to the associated net avoidance and the sedimentation loss of large fecal pellet phaeopigment. Consistent with this, the ICM Sept/82 zooplankton survey (ICM 1983) did indicate the presence of E. pacificus, a large omnivorous euphausid (Ohman 1984), in significant numbers. Hence the importance of a large bodied grazer component, inferred in chapter 3 for the July/83 cruise period on the basis of the Chla and N O J mass balance, also appears to have been an important component of the Aug/82 cruise time series as well. If this inference is correct then grazing becomes the generally dominant flux term in both mass balances.  196  01  U.3  22.4  o.o  6.5  13.1  0.0  Figure 4.60 Simulated Aug./82 phytoplankton time series using the standard run parameter set (except thatfc,-reduced from 0.5 to 0.3 m , and forced by the Aug./82 irradiance, tide, wind stress, and o*t time series. Snapshots are separated by 2 days. _ l  1.3  2.6  198  -0  !0 -  / V o  V po  l a J  OEPTH CM)  10  *T3  O '  (0  fa £ DEPTH 199  (M)  4.5.6 Wind Mixing Effects on Primary Production/Stability Covariance The seasonal switch in sign of the Chla specific biomass/column stability covariance can be examined by extending the analysis of the spring/fall va. mid-summer type c profile t  forcing (referred to below as W and S type profiles) model responses to the main phytoplankton model. The underlying argument both here and previously is that the model response to a variable wind stress imposed on a constant a% profile, can be interpreted equivalently to a constant wind stress imposed on a variable a profile. t  A series of simulations were run using the Aug/82 parameter set (with w' asymptotic to t  2 m/d), along with the Aug/82 initial conditions and forcing time series. For each run the 0-10 m depth averaged/simulation averaged productivity index <PI> , the simulation averaged/0-10 m integrated Chla <S > B  , and the simulation averaged/0-10 m depth  averaged relative specific growth rate </*,> and relative specific nitrogen uptake rate <V> were calculated. The numerical solution was calculated as before, except that since the wind speed <tw> ranged from 0 to 15 m/s, it was necessary to minimize inaccuracy in the solution at the higher wind speeds due to the bottom Neumann boundary specification. This was done by increasing the vertical integration interval to 80 m. TaWe IV.5 tabulates the results of the experiment. The results can be summarized as follows: With increasing <w> from 0-5 m/s for the W a profile, and 0-3 m/s for the S a profile, </*,> ,<V> and <PI> increased while t  <S > B  t  decreased. With a further increase in <w> from 5-9 m/s for W and 4-8 m/s for S,  <fi > and <PI> declined while <S > B  t  and <V> increased. Finally, for <w> over 9 m/s  for W and over 8 m/s for S, </*«> ,<V> increased while both <S > B  and <PJ> decreased.  For both the S and W profiles, </!,> became equal to <V> at <iw> = 11 m/s. There is no obvious explanation for the non-linear response to wind forcing. However, on the assumption that numerical instabilities were avoided through the conservative selection of solution time steps and grid point spacing (see Sec. 4.4), then it must be attributed to the nature of the coupling in system (4.1), perhaps related to the photoperiod. 200  T a b l e I V . 5 Wind shearforcingof the Aug/82 time series simulation. Time averaged wind speed <w> ( and <PI> ,<S > ,<a,> ,and <V> as defined in the text. B  <w>  <P/>  <S > B  </»•>  midsummer (S) <r profile 0.0 2.49 74.4 0.1845 1.0 2.50 72.9 0.1867 2.0 2.50 68.3 0.1915 3.0 2.51 65.2 0.1944 4.0 2.49 65.9 0.1947 5.0 2.42 72.1 0.1902 6.0 2.30 81.1 0.1848 7.0 2.26 85.5 0.1835 8.0 2.19 89.3 0.1833 9.0 2.17 87.0 0.1904 10.0 2.16 78.2 0.1990 11.0 2.17 67.9 0.2079 15.0 2.21 33.5 0.2375 spring/fall (W) <rt profile 0.0 2.50 74.1 0.1845 1.0 2.53 73.4 0.1842 2.0 2.55 69.9 0.1879 3.0 2.58 65.8 0.1917 4.0 2.60 61.9 0.1950 5.0 2.60 60.2 0.1968 6.0 2.58 61.3 0.1968 7.0 2.58 62.4 0.1954 8.0 2.53 69.6 0.1875 9.0 2.44 76.9 0.1846 10.0 2.42 72.7 0.1950 11.0 2.41 60.3 0.2083 15.0 2.30 31.7 0.2374 (  <V>  0.1287 0.1295 0.1342 0.1415 0.1423 0.1418 0.1505 0.1568 0.1774 0.1897 0.1988 0.2079 0.2375 0.1287 0.1286 0.1301 0.1316 0.1373 0.1463 0.1559 0.1601 0.1705 0.1764 0.1948 0.2083 0.2374  The initial decrease in <S > and increase in </2,> and <V> is attributable to the B  simultaneous wind generated increase in the influx of nitrogen and efflux of Chla specific biomass. The corresponding increase in <PI> over this range in <w> indicates that the increase in this efflux was less than the increase in specific growth rate. Since the wind speed during the May, July, and Aug cruise periods was in this range of <iw> , the model behavior seems consistent with the mid-summer 5 -stability correlation and P/-stability B  anti-correlation in the I C M data set. The subsequent increase in <S > and decrease in <V> for <w> above 3 m/s and B  4 m/s for S and W <r profiles respectively is not so easily interpreted however. The t  decrease in <fi > is particularly problematical due to the simultaneous increase in <V> . $  Without a more detailed evaluation of the model output, little can be offered in the way of a substantive explanation for this reversal in sign of the dependence of <p > and probably 9  in a related manner, <S > on <w> . However, the important point perhaps, is that the B  201  model does respond to small changes in wind speed (i.e. equivalent to small change in column stability) in a fashion which predicts the mid-summer and spring/fall covariance pattern in the ICM data set. Furthermore, the range in <tu> over which the spring/fall and mid-summer covariance patterns are predicted, is consistent with the expected higher wind shear during the spring/fall compared to the mid-summer.  4.5.7 Summary A z,t simulation model was developed to evaluate wind and column stability effects on primary production. An analysis of the associated vertical diffusion submodel indicated that the effect of vertical displacement on modifying photo-inhibition and photoadaptation processes was generally unimportant except in the top 1 m of the water column, and hence unimportant in the primary production-stability covariance structure in the ICM data set. The relative performance of a zero, depth independent, and depth dependent sinking rate model formulation in predicting the July/83 POC distribution was interpreted as indicating that depth dependent growth was unimportant compared to depth dependent sinking (i.e. N-dependent, gravitational or behavioral) in determining this distribution. Simulation of the Aug/82 time series indicated that the missing grazer component inferred from the Chla and NO3  budgets for the July/83 time series was also required  for the modelling success of the Aug/82 data set, and suggested that a large bodied grazer component was an important component of the mass balance flux. The sensitivity analysis of the primary production model to wind and column stability forcing indicated non-monotonic responses by the Chla specific biomass and growth indices to changes in wind stress. This was interpreted in terms of the inherent non-linearities innutrient-growth coupling in the model and was suggested to be the mechanism responsible for the seasonal primary production-stability covariance structure in the I C M data set. By comparison, the model was shown to be relatively insensitive to variations in the stability profile. 202  5 . Conclusions An attempt was made in this thesis to characterize primary production in the HolbergRupert basin. The visual and statistical analysis of a low resolution « 7 yr (ICM) data set indicated that primary production in the basin occurs seasonally in a similar manner to other hydrodynamically energetic coastal regions, with productivity remaining elevated and biomass increasing in an essentially monotonic fashion throughout the summer. This pattern has been generally attributed to the forcing of a relatively high surface nutrient flux by exchange processes. Since the basin is characterized at least near its entrance by vigorous tidal mixing, it was assumed at the start that this generalization probably applies to the basin as well. However, since mixing can also have limiting effects on growth (i.e. light-limiting surface mixed layer) as well as on biomass distribution (i.e. dispersion), a central question addressed in this thesis was what were the combined consequences of this mixing regime?  It had been suggested in an earlier study that biomass and by indirect inference growth, were generally light limited in the basin. However, this is inconsistent with the covariance structure in the ICM data set which to the contrary implicates nutrition as the growth limiting factor, while the effect of mixing on biomass distribution is seasonally dependent. In further evaluating the implied dynamic coupling between mixing and primary production, the nature of the mixing regime was assessed qualitatively. Tidal mixing is at least regionally the most apparent exchange process (i.e. proximal to QN). However, it is not clear how important this is to basin near surface exchange generally, due to concurrent wind and surface layer stability effects, as well as to a direct buoyancy effect on the inflow pattern of the tidal jet. Accordingly, a re-analysis of an eight month near surface current record in Holberg Inlet suggested the presence of a near surface vertical entrainment flux during most of the flood tide period under all run-off (i.e. w buoyancy) conditions encountered. However, this evidence was somewhat equivocal for summer time run-off conditions and combined with related interpretations in the literature implied a significant seasonal periodicity in the flood tidally averaged relative jet buoyancy. This implicates the tidal jet in 203  the I C M data set seasonal covariance structure. On the other hand, high resolution spatial and temporal sampling of primary productivity indicated that if entrainment is present during the summer it is a small component in the Chla and N O J mass balance, though the precision of this contribution estimate is limited in large measure by the uncertainty associated with the contribution estimate of an unobserved herbivore grazer component. The presence of this grazing component was inferred from the mass balances and from the relative performance of a numerical simulation of the data time series with and without its present. The effect of the mixing regime on growth has been shown in the literature to be potentially both nutrient and light related, the former due to the dependence of surface layer nutrient replacement and the latter due to the dependence of vertical displacement and the modifying affect it has on light limited growth as well as photo-inhibition and photoadaptation processes. The characteristic displacement time scales associated with the time averaged heat conserving estimate of the vertical diffusion profile indicated that the light related effect is probably minor except within the top 1 m and below 10 m in the water column, though internal wave effects were not included in this assessment. Hence the turbulent diffusion component of such a displacement is probably not important to column production and certainly not important to the ICM data set covariance structure, where to the contrary covariance is strongest near the base of the pycnocline. Statistical patterns in the I C M data set, the Chla and N O J mass balance calculations, and the numerical simulations all implicate nutrient flux as the primary coupling mechanism of vertical mixing to primary productivity. Since the wind shear and buoyancy flux can be expected to be the two forcing functions with the strongest seasonal signal, model response was used to evaluate the possible contribution of these forcing functions to the seasonal biomass-mixing covariance pattern in the I C M data set. Accordingly, the analysis attributed this to the inherent non-linear coupling between the governing processes giving rise to a non-monotonic response in growth rate and biomass accumulation with a monotonic change in wind speed. Though a more detailed 204  analysis is required to fully account for this response, the fact that the non-monotonic responses were consistent with the seasonal pattern in the ICM data set covariance structure is encouraging.  205  6. References Anderson J.J. and A . Okubo 1982. Resolution of chemical properties with a vertical profiling pump. Deep Sea Research 29:1013-1019. Antia N.J. , C D . McAllister, T.R. Parsons, K . Stephens, and J.D.H. Strickland 1963. Further measurements of primary production using a large volume plastic sphere. Limnology and Oceanography 8:166-183. Bannister T.T. 1979. Quantitative description of steady state, nutrient saturated algal growth, including adaptation. Limnology and Oceanography 24:79-96. Banse K . 1976. Rates of growth, respiration, and photosynthesis of unicellular algae as related to cell size: A review. Journal of Phycology 12:135-140. Belay A. 1981. An experimental investigation of inhibition of phytoplankton photosynthesis at lake surfaces. New Phytologist 89:61-74. Bienfang P.K. 1980. Herbivore diet affects on fecal pellet settling. Canadian Journal of Fisheries and Aquatic Sciences 37:1352-1357. Bienfang P.K. and P.J. Harrison 1984. Co-variation of sinking rate and cell quota among nutrient replete marine phytoplankton. Marine Ecology-Progress Series. 14:297-300. Bienfang P.K. 1981. A technologically simple and reliable method for measuring phytoplankton sinking rates. Canadian Journal of Fisheries and Aquatic Sciences 38:12891294. B M D P 1979. Biomedical Center Computer Programs P-series. Univ. of California Press. 870p. Box, G.E.P. 1953. Non-normality and tests of variance. Biometrika 40:318-335. Bowden K . F . 1967. Stability effects on turbulent mixing in tidal currents. In Boundary Layers and Turbulence. Physics of Fluids Supplement, pp. S278-280. Brzezinski M . A . 1985. The Si:C:N ratio of marine diatoms: interspecific variability and the effect of some environmental variables. Journal of Phycology 21:347-357. Buckley J.R. 1977. The currents, winds, and tides of northern Howe Sound. University of British Columbia, Vancouver B.C. Canada PhD thesis 226p. Chan A . 1978. Comparative physiological study of marine diatoms and dinoflagellates in relation to irradiance and cell size. 1. Growth under continuous light. Journal ofPhycology 14:396-402. Caperon J . and J . Meyer 1972. Nitrogen-limited growth of marine phytoplankton. I. Changes in population characteristics with steady state growth rate. Deep Sea Research 19:601-618.  206  Carstons, T. 1970. Turbulent diffusion and entrainment in 2-layer flow. Proceedings American Society Civil Engineering W W l 97-104. Collos Y . and G . Slawyk 1980. Nitrogen uptake and assimilation by marine phytoplankton. In Primary Productivity in the Sea. Falkowski P.G. (ed.). Plenum Press New York, pp. 195-212. Conway H.L. 1977. Interactions of inorganic nitrogen in the uptake and assimilation by marine phytoplankton. Marine Biology 39:221-232. C6t6 B. and T. Piatt 1983. Day-to-day variations in the spring-summer photosynthetic parameters of coastal marine phytoplankton. Limnology and Oceanography 28:320344. CHS Harmonic constants and associated data, Vol. 6. Canadian Hydrographic Service. CRC 1973. Chemical Rubber Company Standard Mathematical Tables. CRC Press Cleveland, 369p. Cullen J.J. 1982. The deep chlorophyll maximum: comparing vertical profiles of chlorophyll a. Canadian Journal of Fisheries and Aquatic Sciences 39:791-803. Cullen J.J. and S.G. Holligan 1981. Effects of nitrate on the diurnal vertical migration, carbon to nitrogen ratio, and the photosynthetic capacity of the dinoflagellate Gymnodinium splendens. Marine Biology 62:81-89. Daley R.J. 1973. Experimental characterization of lacustrine chlorophyll diagenesis. 2. Bacterial, viral, and herbivore grazing effects. Archiv Hydrobiologie 72:409-439. Denman K.L. and A.E. Gargett 1983. Time and space scales of vertical mixing and advection of phytoplankton in the upper ocean. Limnology and Oceanography 28:801-815. Denman K . L . and T. Piatt 1975. Coherence in the horizontal distribution of phytoplankton and temperature in the upper ocean. Proceedings of the 6th liege Colloquium in Ocean Hydrodynamics, Memoires de la Societe Royale des Sciences de liege 6 ser 7:19-30. Denman K . L . and T. Piatt 1977. The variance spectrum of phytoplankton in a turbulent ocean. Journal of Marine Research 34:593-601. DFO 1983. Historical Streamflow Summary, British Columbia Inland Waters Directorate. Water Resources Branch. Water Survey of Canada, Ottawa. Ministry of Supply and Services Canada. Dillon T . M . and D.R. Caldwell 1980. The Batchelor spectrum and disipation in the upper ocean. Journal of Geophysical Research 85:1910-1916. Drinkwater K . F . 1973. The role of tidal mixing in Rupert and Holberg Inlets. University of British Columbia, Vancouver B.C. Canada MSc. thesis 58p.  207  Drinkwater, K . F . and T.R. Osborn 1975. The role of tidal mixing in Rupert and Holberg Inlets, Vancouver Is. Limnology and Oceanography 20:518-529. Droop M.R. 1983. 25 years of algal growth kinetics: A personal view. Botanica Marina 26:99-112. Droop M.R., M . J . Mickelson, J . M . Scott, and M.F. Turner 1982. Light and nutrient status of algal cells. Journal of the Marine Biological Association of the U.K. 62:403-434. Dyer, K.R. 1973. Estuaries: A Physical Introduction. John Wiley and Sons. 139p. . 1979. Estuarine Hydrography and Sedimentation. Dyer, K.R. (ed.), Cambridge Univ. Press, Cambridge. 230p. Eppley R.W. 1980. Estimating phytoplankton growth rates in the central oligotrophic ocean. In Primary Productivity in the Sea. Falkowski P.G. (ed.). Environmental Science Research Vol. 19. Plenum Press 528p. Eppley R.W. 1981. Relations between nutrient assimilation and growth in phytoplankton with a brief review of estimates of growth rate in the ocean. In Physiological Basis of Phytoplankton Ecology, Piatt T. (ed.). Canadian Bulletin Fisheries and Aquatic Sciences 210. pp. 251-263. Eppley R.W., J.W. Rogers, and J.J. M Carthy 1969. Half-saturation constants for uptake of nitrate and ammonium by marine phytoplankton. Limnology and Oceanography 14:912-920. c  Eppley R.W. and P R . Sloan. 1966. Growth rates of marine phytoplankton: Correlations with light absorption by cell chlorophyll a. Physiologia Plantarum 19:47-59. Falkowski P.G. 1981. Light-shade adaptation and assimilation numbers. Journal of Plankton Research 3:203-216. Falkowski P.G. 1983. Light-shade adaptation and vertical mixing of marine phytoplankton: A comparative study. Journal of Marine Research 41:215-237. Falkowski P.G. and T.G. Owens 1978. Effects of light intensity in photosynthesis and dark respiration in six species of marine phytoplankton. Marine Biology 45:239-295. Farmer D . M . and J.D. Smith 1979. Generation of Lee waves over the sill in Knight Inlet. In NATO Conference on Fjord Oceanography. Freeland H.J. et al. (eds.). NATO Conference Series Vol. 4. Plenum Press New York, pp. 259-270. Farmer D . M . and J.D. Smith 1980. Tidal interactions of stratified flow with a sill in Knight Inlet. Deep Sea Research 27A:239-254. Farmer D . M . and H.J. Freeland 1983. The physical oceanography of fjords. Progress in Oceanography 12:147-219.  208  Fasham M.J.R. and P.R. Pugh 1976. Observations on the horizontal coherence of chlorophyll a and temperature. Deep Sea Research 23:527-538. Frankenberg D. and K . L Smith Jr. 1967. Coprophagy in marine aniamls. Limnology and Oceanography 12:443-450. Fritz S. 1957. Solar energy on clear and cloudy days. Scientific Monthly. February 1957 55-65. Frost B.W. 1972. Effects of size and concentration of food particles on the feeding behavior of the marine planktonic copepod Calanus pacificus. Limnology and Oceanography 17:805-815. Gallegos C.L. and T. Piatt 1981. Photosynthesis measurements on natural populations of phytoplankton: Numerical analysis. In Physiological Bases of Phytoplankton Ecology, Piatt T. (ed.). Canadian Bulletin Fisheries Aquatic Sciences 210. pp. 103-112. Gardner G.A. and I. Szabo 1982. British Columbia pelagic marine Copepoda: An identification manual and annotated bibliography. Canadian Journal of Fisheries and Aquatic Sciences special publication # 62. 536p. Gill A.E. 1982. Atmosphere-Ocean dynamics. International Geophysics series. Vol 30. Academic Press 662p. Goldman J.C. and J.J. M Carthy 1978. Steadstate growth and ammonium uptake of a fast-growing marine diatom. Limnology and Oceanography 23:695-703. c  Goyette D. and H . Nelson 1977. Marine environmental assessment of mine waste disposal into Rupert Inlet British Columbia. Surveillance Report PR 77-11. Environmental Protection Service, Environment Canada. Harding W.J., B.B. Prezelin, B . M . Sweeney, and J.L. Cox 1982. Diel oscillations of the photosynthesis-irradiance (P-I) relationship in natural assemblages of phytoplankton. Marine Biology 67:167-178. Harris, G.P. 1973. Vertical mixing mechanisms and their effects on primary production of phytoplankton. Scientific Series No. 33, Environment Canada. Harris G.P. 1978. Photosynthesis, productivity and growth: The physiological ecology of phytoplankton. Archiv Hydrobiologia Beih. Ergebn. Limnol. 10:1-171. Harris G.P. 1980. The measurement of photosynthesis in natural populations of phytoplankton. In Physiological Ecology of Phytoplankton. Morris I. (ed.). Univeristy ofCalifornia Press Los Angeles, pp. 129-187. Harrison P.J., J.D. Fulton, F.J.R. Taylor, and T.R. Parsons 1983. Review of the biological oceanography of the Strait of Georgia: Pelagic environment. Canadian Journal of Fisheries and Aquatic Sciences 40:1064-1094.  209  Hay H.E. 1981. Submarine channel formation and acoustic remote sensing of suspended sediments and turbidity currents in Rupert Inlet B.C. University of British Columbia, Vancouver B.C. Canada PhD thesis 326p. Herman A.W. 1983. Vertical distribution patterns of copepods, chorophyll, and production in Northeastern Baffin Bay. Limnology and Oceanography 28:709-719. Hitchcock G.L. 1980. Diel variation in chlorophyll o, carbohydrate, and protein content of the marine diatom Skeletonema costatum. Marine Biology 57:271-278. Holmboe J , G.E. Forsythe, and W. Gustin 1948. Dynamic meteorology. John Wiley and Sons Inc. London 378p. Hollister H.J. and A . M . Sandnes 1972. Sea surface temperatures and salinities at shore stations on the British Columbia coast 1914-1970. Inst. Ocean Sciences, Patrica Bay Pacific Marine Science Report 72-13. Holm-Hansen O. and B. Riemann 1978. Chlorophyll a determination: Improvements on methodology. Oikos 30:438-447. ICM 1979. 1978 environmental assessment report, Vol. 1. 99p. ICM 1983. 1982 annual environmental assessment report, vol 1. 106p. Ivanoff A. 1977. Oceanic absorption of solar energy. In Modelling and prediction of the upper layers of the ocean. Kraus, E.B. (ed.). Proceedings of a NATO Advanced Study Institute. Pergamon Press New York, pp. 47-71. Jamart B . M . , D.F. Winter, K . Banse, G.C. Anderson, and R.K. Lam 1977. A theoretical study of phytoplankton growth and nutrient distribution in the Pacific Ocean off the northwestern U.S. coast. Deep Sea Research 24:753-773. James I.D. 1977. A model of the annual cycle of temperature in a frontal region of the Celtic sea. Estuarine and Coastal Marine Science 5:339-353. Jeffery S.W. and G.F. Humphrey 1975. New spectrophotometric equations for determining chlorophylls a,6,Ci,and c in higher plants, algae, and natural phytoplankton. Biochemie und Physiologie der Pflanzen 167:191-194. 2  Johnson R.D. 1974. Dispersal of recent sediments and mine tailings in a shallow-silled fjord, Rupert Inlet, B.C. University of British Columbia, Vancouver B.C. Canada PhD thesis 181p. Kullenberg G . 1976. On vertical mixing and energy transfer from the wind to the water. Tellus 28:159-165. Kullenberg G. 1977. Entrainment velocity in natural stratified vertical shear flow. Estuarine and Coastal Marine Science 5:329-338.  210  Laws E.A. and T.T. Bannister 1980. Nutrient and light-limited growth of Thalassiosira fluviatilis in continuous culture with implications for phytoplankton growth in the ocean. Limnology and Oceanography 25:457-473. Legendre L . 1981. Hydrodynamic control of marine phytoplankton: The paradox of stability. In Ecohydrodynamic8, Nihoul J.C.J, (ed.). Elsevier Scientific Publishing Co. Amsterdam, pp. 191-208. Legendre L. and S. Demers 1984. Towards dynamic biological oceanography and limnology. Canadian Journal of Fisheries and Aquatic Sciences 41:2-19. Leibovich S. and S. Paolucci 1980. The langmuir circulation instability as a mixing mechanism in the upper ocean. Journal Physical Oceanography 10:186-207. Lekan J.F. and R.E. Wilson. Spatial variability of phytoplankton biomass in the surface waters of long Island. Estuarine and Coastal Marine Science 6:239-251. Lewis M.R. and T. Piatt 1982. Scales of variability in estuarine ecosystems. In Estuarine comparisons. V.S. Kennedy (ed.). Proceedings of the 6th Biennial Int. Est. research conference. Gleneden Beach Oregon, Nov. 1981. Academic Press Paris, pp. 3-20. Lewis M.R., J.J. Cullen, and T. Piatt 1984. Relationship between vertical mixing and photo-ad apt at ion of phytoplankton: similarity criteria. Marine Ecology-Progress Series 15:141-149. Lewis W.M. Jr. 1978. Comparison of temporal and spatial variation in the zooplankton of a lake by means of variance components. Ecology 59:661-671 Li C.C. 1964. Introduction to Experimental Statistics. McGraw Hill NY. 460p. Lindman, H.R. 1974. Analysis of Variance in Complex Experimental Designs. W.H. Freeman and Co. San Francisco 352p. Longhurst A.R. 1976. Interactions between zooplankton and phytoplankton profiles in the Eastern Tropical Pacific Ocean. Deep Sea Research 33:729-754. Longhurst A.R. and A.W. Herman 1981. Do oceanic zooplankton aggregate at or near the deep chlorophyll maximum? Journal of Marine Research 39:353-356. Lorenzen, C.J. 1967. Vertical distribution of chlorophyll and phaeopigments: Baja California. Deep Sea Research 14:735-745. Lorenzen C.J. 1979. Ultraviolet radiation and phytoplankton photosyntesis. Limnology and Oceanography 24:1117-1120. Lorenzen C.J. and N.A. Welschmeyer 1983. The in situ sinking rates of herbivore fecal pellets. Journal of Plankton Research 5:929-933. Marra J . 1978. Phytoplankton photosynthesis response to vertical movement in a mixed layer. Marine Biology 46:203-208. 211  Marra J . 1980. The course of light intensity adaptation in a marine diatom. Marine Biology Letters 1:175-183. Marra J . and K . Heinemann 1982. Photosynthesis response by phytoplankton to sunlight variability. Limnology and Oceanography 27:1141-1153. M Carthy J.J., W.R. Rowland, and J.L. Taff 1977. Nitrogenous nutrition of the plankton in the Chesapeake Bay. 1. Nutrient availability and phytoplankton preferences. Limnology and Oceanography 22:996-1011. c  M Carthy J.J. and J.C. Goldman 1979. Nitrogenous nutrition of marine phytoplankton in nutrient depleted waters. Science 203:670-672. c  Moore C. 1984. University of British Columbia, Vancouver B.C. Canada Curve Fitting Routines. Computing Center University of British Columbia, Vancouver B.C. Canada. 189p. Morris I. and W. Skea 1978. Products of photosynthesis in natural populations of marine phytoplankton from the Gulf of Maine. Marine Biology 47:303-312. Mullin M . M . , P.R. Sloan, and R.W. Eppley 1966. Relationship between carbon content, cell volume, and area in phytoplankton. Limnology and Oceanography 11:307-311. Ohman M.D. 1984. Omnivory by Euphausia pacifica: The role of copepod prey. Marine Ecology-Progress Series 19:125-131. Okubo A. 1971. Ocean diffusion diagrams. Deep Sea Research 18:789-802. Okubo A . 1972. A note on small organism diffusion around an attractive centre. A mathematical model. Journal of the Oceanographic Society of Japan 28:1-7. Okubo A. 1978. Horizontal dispersion and critical scales for phytoplankton patches. In Spatial Pattern in Plankton Communities. Steele J.H. (ed.). Plenum Press New York, pp. 21-42. Osterberg C , A . G . Corey, and J . Curl 1963. Acceleration of sinking rates of radionuclides in the ocean. Nature 200:1276-1277. Packard T.T., D. Blasco, and R.T. Barber 1978. Mesodinium rubrum in the Baja California upwelling system. In Upwelling Ecosystems, Boje R. and M . Tomczuk (eds.). Springer-Verlag Berlin, pp. 73-89. Parslow J.S. 1981. Phytoplankton-zooplankton interactions: Data analysis and mod-elling (with particular reference to Ocean Station P (50° N , 145° W) and controlled ecosystem experiments). PhD thesis University of British Columbia, Vancouver B.C. Canada. 400p. Parsons T.R., M . Takahashi, and B. Hargraves 1984. Biological Oceanographic Processes. Pergamon Press New York, 330p.  212  Parsons T.R., R.I. Perry, E.D. Nutbrown, W. Hsieh, and C M . Lalli 1981. Frontal zone analysis at the mouth of Saanich Inlet B.C., Canada. Marine Biology 73:1-5. Pearson E.S. and H.O. Hartley 1966. Biometrika Tables For Statiticians (3rd ed.). Cambridge. Peters R.H. and J . A. Downing 1984. Empirical analysis of zooplankton filtering and feeding rates. Limnology and Oceanography 29:763-784. Phillips, O . M . 1980. The Dynamics of the Upper Ocean. Cambridge Monographs in Mechanical and Applied Mathematics. Cambridge Univ. Press Cambridge, pp. 191192. Phillips, O.M. 1975. Entrainment. In Modelling and Prediction of the Upper Layers of the Ocean. Kraus, E.B. (ed.) Pregamon Press New York, pp. 92-101. Pickard G . L . 1963. Oceanographic characteristics of inlets of Vancouver Is., British Columbia. Journal Fisheries Research Board of Canada 20:1109-1144. Pickard G.L. 1979. Descriptive Physical Oceanography (3rd Ed.). Pergamon Press New York, 232p. Pingree R.D., P.R. Pugh, P.M. Holligan, and G.R. Forster 1975. Summer phytoplankton blooms and red tides along tidal fronts in the approaches to the English Channel. Nature 258:672-677. Piatt T. and C.L. Gallegos 1980. Modeling primary production. In Primary Productivity in the Sea. Falkowski P.G. (ed.). Environmental Science Research Vol. 19. Plenum Press New York, pp. 339-362. Piatt T., C.L. Gallegos, and W.G. Harrison 1980. Photoinhibition of photosynthesis in natural assemblages of marine phytoplankton. Journal of Marine Research 38:687-701. Piatt T., D.U. Subba Rao, and K . L . Denman 1977. Quantitative stimulation of phytoplankton productivity by deep water admixture in a coastal inlet. Estuarine and Coastal Marine Science 5:567-573. Piatt T., L . M . Dickie, and R.W. Trites 1970. Spatial heterogeneity of phytoplankton in a near-shore environment. Journal Fisheries Research Board of Canada 27:1453-1473. Piatt, T. and C. Filion 1973. Spatial variability of the productivity:biomass ratio for phytoplankton in a small marine basin. Limnology and Oceanography 18:743-749. Pond, S. and Pickard, G.L. 1978. Introductory Dynamic Oceanography. Pergamon Press New York, 241 p. Prezelin B.B. and A.C. Ley 1980. Photosynthesis and chlorophyll a fluorescence rhythm of marine phytoplankton Marine Biology 5:295-307.  213  Prezelin B.B. and H.A. Matlick 1980. Time-course of photo-adaptation in the photosynthetic irradiance relationship of a dinoflagellate exhibiting photosynthetic periodicity. Marine Biology 58:85-96. Raymont J.E.G. 1983. Plankton and Productivity in the Oceans. 2 plankton. Pergamon Press Oxford, 824p.  n d  ed. Vol. 2, Zoo-  Redfield A.C., B.H. Ketchum and F.A. Richards 1963. The influence of organisms in the composition of seawater. In The Sea Vol. 2, Hill M . N . (ed.). Interscience Publications. John Wiley and Sons New York, pp. 26-77. Rhee G-Y. and I.J. Gotham 1981. The effect of environmental factors on phytoplankton growth: Light and the interactions of light with nitrate limitation. Limnology and Oceanography 26:649-659. Riper D.M., T.G. Owens and P.G. Falkowski 1979. Chlorophyll turnover in Skeletonema costatum, a marine planktonic diatom. Plant Physiolology 64:49-54. Saunders P M . 1976. Near-surface current measurements. Deep Sea Research 23:249-257. Sen Gupta R. and H.W. Jannasch 1973. Photosynthetic production and dark assimilation of CO2 in the Black Sea. Internationale Revue der Gersumten Hydrobiologie 58:625632. Shuman F.R. and C.J. Lorenzen 1975. Quantitative degradation of chlorophyll by a marine herbivore. Limnology and Oceanography 20:580-586. Silvert W. 1979. Practical curve fitting. Limnology and Oceanography 24:767-773. Smayda, T.J. 1969. Some measurements of the sinking rate of fecal pellets. Limnology and Oceanography 14:621-625. Smith G.D. 1965. Numerical Solutions of Partial Differential Equations with Exercises and worked solutions. Oxford Univ. Press New York, 179p. Sokal R.R. and F.J. Rohlf 1981. Biometry. W.H. Freeman and Co. San Francisco, 859p. Soohoo J.B. and D.A. Kiefer 1982. Vertical distribution of phaeopigments. n. Rates of production and kinetics of photo-oxidation. Deep Sea Research 29:1553-1564. Steele J.H. and C.S. Yentsch 1960. The vertical distribution of chlorophyll. Journal of the Marine Biological Association of the U.K. 39:217-226. Steemann-Nielsen E. 1952. The use of radioactive carbon C for measuring organic production in the sea. Journal of the International Council for the Exploration of the Sea 18:117-140. 1 4  Stephens K . and J . Silbert 1976. Primary production in Rupert Inlet, B.C. Presented at 5th technical conference on estuaries of the Pacific Northwest Oregon State Univ., Corvallis, Oregan April 1, 1976. 214  Stephens K., J.D. Fulton, and O.D. Kennedy 1969. Summary of biological oceanographic observations in the Strait of Georgia 1965-1968. Journal Fisheries Research Board of Canada Technical report 110. Stolzenbach K.D., E . E . Adams, and D.R.F. Harleman 1973. A user's manual for 3dimensional heated surface discharge computations. Ralph M . Parsons Lab. for Water Resources and Hydrodynamics. Report. 156. Dept. Civil Engineering, MIT, Cambridge Mass. Strickland J.D.H. and T.R. Parsons 1972. A practical Handbook of Seawater Analysis. Journal Fisheries Research Board of Canada Bulletin 167. 311p. Stucchi D. 1985. The Tidal Jet in Rupert-Holberg Inlet, B.C. Atmosphere-Ocean. 23:118136. Stucchi D. and D.M. Farmer 1976. Deep-water exchange in Rupert-Holberg Inlets. Inst. Ocean Sciences, Patrica Bay Pacific Marine Science Report 76-10. Sullivan, D.L. 1979a. The effects of sulphite pulp mill discharges on phytoplankton productivity in Neroutsos Inlet and Quatsino Sound 1974-1976. Dept. Environment Regional Program Rept. No. 79-15. Environmental Protection Service, Pacific Region. -. 1979b. The effects of marine disposal of mine tailings on phytoplankton productivity in Rupert Inlet, B.C. 1974-1976. Dept. Environment Regional Program Rept. No. 79-16. Environmental Protection Service, Pacific Region. Sundaram T.R. and R.G. Rehm 1973. The seasonal thermal structure of deep temperate lakes. Tellus 25:157-168. Sverdrup H.U., M . W . Johnson, and R.H. Fleming 1942. The Oceans. Their physics, chemistry, and general biology. Prentice-Hall Inc. N Y 1088p. Taguchi S. and T. Piatt 1977. Assimilation of C 0 2 in the dark compared to phytoplankton in a small coastal inlet. Estuarine and Coastal Marine Science 5:679-684. 1 4  Takahashi M . , D.L. Seibert, and W.H. Thomas 1977. Occasional blooms of phytoplankton during summer in Saanich Inlet, B.C., Canada. Deep Sea Research 24A:775-780. Tett P and B. Grantham 1979. Variability in sea-loch phytoplankton. In NATO Conference on Fjord Oceanography. Freeland H.S. et al. (eds.). Plenum Press New York, pp. 435-438. Therriault J . and T. Piatt 1978. Spatial heterogeneity of phytoplankton biomass and related factors in the near-surface waters of an exposed coastal enbayment. Limnology and Oceanography 23:888-899. Townsend A . A . 1980. The structure of turbulent shear flow. 2 Press Cambridge, 429p.  215  n d  ed. Cambridge Univ.  Vincent W.F. 1980. Mechanisms of rapid photosynthetic adaptation in natural phytoplankton communities. 1. Re-distribution of excitation energy between photosystem I and II. Journal Phycol. 15:429-434. Waldichuk, M . 1958. Some oceanographic characteristics of a polluted inlet in British Columbia. Journal of Marine Research 17:536-551. Walsh J.J. and R.C. Dugdale 1972. Nutrient submodels and simulation models of phytoplankton production in the sea. In Nutrients in Natural Waters, Kramer J . and H . Allen (eds.). J . Wiley and Sons New York, pp. 171-191. Webster, I and D.M. Farmer 1976. Analysis of salinity and temperature records taken at 3 lighthouse stations on the B.C. coast. Inst. Ocean Sciences, Patrica Bay , Pacific Marine Science Report 76-11. Welschmeyer N.A. and C.J. Lorenzen 1985. Chlorophyll budgets: zooplankton grazing and phytoplankton growth in a temperature fjord and the Central Pacific Gyre. Limnology and Oceanography 30:1-21. Wroblewski J.S. 1977. A model of phytoplankton plume formation during variable Oregon upwelling Journal of Marine Research 35:356-393. Wroblewski J.S. and J.J. O'Brien 1976. A spatial model of phytoplankton pathiness. Marine Biology 35:161-175. Zar J.H. 1974. Biostatistical Analysis. Prentice-Hall Inc. London, 620p.  216  ©  o at  o I I  °  8  0  O  *o°S  fl Q  o  e  O OO ©  1  i.s  2.7  2.0  LOG MEAN  o*  (_)  °  1  °  s e  f o  o  B  8° °  e 1.5  LOG HE ON  C  8  0.4  LOG MEAN  A p p e n d i x A Scatter plot of log sample mean and log sample variance of: (A) the 0-25 m static stability parameter E\ (B) the 0-15 m integrated Chlo S \ (C) the 0-15 m depth averaged Chlo specific carbon uptake (PI). B  217  Appendix B . In the data normalization procedure, the test statistic used was the Dstatistic (Zar 1974, p. 83), which estimates the chance likelihood of the observed deviation from normality. Generally when this statistic indicated that distribution corrections were necessary, the appropriate transformation was determined by the log-log least-squares procedure outlined by Elliott (1977, p. 73). This is based on an underlying variance to mean power law due to Taylor (1961). Accordingly, the log sample variance was regressed against log sample mean by linear least-squares, with the data pooled by station since the Kendall Tb matrix indicated that the least amount of concordance was generally associated with this grouping variable (Table 1.8). The slope b determines the optimum transformation parameter x' = z ( l — 6/2),where z is the transformation of variable z. The table below tabulates the D-statistic for E and P/as well as those environmental variables demonstrating significant concordance with either of these phytoplankton index in Table 1.8. For comparison, both before and after transformation statistics are presented. The asterisk indicates normal distribution at P < 0.05. 1  B  ICM Stn.  E  E  b  Untransformed data 0.2476 0.2428 1 (49) 0.2097 0.2657 2 (47) (47) 3 0.2387 0.2355 (48) (48) 4 0.2401 0.2539 (48) (48) 5 0.2515 0.2690 (45) (47) 6 0.2576 0.2682 (47) (47) Transformed data 1 0.2846* 0.2752* 2 0.28C1* 0.2729' 3 0.2823* 0.2681 4 0.2816* 0.2834* 0.2835* 0.2847* 5 6 0.2741* 0.2751*  E  Jj  E°  PI  0.2529 (49) 0.2613 (471 0.2495 (48) 0.2384 (48) 0.2009 (45) 0.2480 (46)  0.2863  0.2295 (49] 0.2332  0.2779* (41) 0.2776* (41)  0.2478 (49) 0.2509 (48] 0.2339 (48) 0.2478 (48) 0.2580 (48) 0.2575 (48)  0.2492 (48) 0.2418 (48) 0.2761' (48) 0.2785' (48)  0.2729* 0.2759* 0.2862* 0.2842* 0.2760* 0.2760*  0.2859* 0.2841* 0.2736* 0.2833* 0.2814* 0.2857*  0.2744' 0.2715' 0.2799* 0.2732* 0.2857* 0.2871  t  0.2740* 0.2856* 0.2800* 0.2827* 0.2706* 0.2840*  0.2822* 0.2860* (41), 0.2856*  Thus, while most sample units were not normally distributed before, all but two met the normality criterion at P < 0.05 after the described transformation. An identical procedure was applied to all variables prior to calculating parametric statistics.  218  A p p e n d i x C I . Partial correlation matrix of E sample station and year; and pooled by month.  D  E  E  t  0.133 -0.022 0.545  0.974* -0.423  •0.616  t  -0.307 -0.158 -0.193  0.870* 0.579  0.151  0.101 0.249 -0.153  0.949 0.773*  0.583  -0.551 -0.656* -0.318  0.958* 0.865*  0.713  0.052 -0.027 0.312  0.921* 0.726*  0.453  0.120 0.445 -0.186  0.862* 0.878*  0.537  -0.776* -0.356 -0.569  0.740* 0.781*  0.334  Chla  P/  QS data subset April n=6 PI -0.541 Ij -0.281 0.115 E 0.313 -0.952* E 0.207 -0.935* E 0.270 0.427 May n=12 PI -0.334 Ij 0.836* -0.163 E -0.038 -0.285 E„ -0.077 0.003 E 0.164 -0.555 June n=12 PI 0.119 J* 0.250 0.649* E -0.297 -0.483 E -0.280 -0.375 £, -0.210 -0.516 July n=10 PI -0.872* Ij -0.802* 0.778* £ 0.816* -0.868* £ 0.8C9* -0.820* E 0.681* -0.798* August n=13 PI -0.223 /* 0.3G0 0.415 £ -0.313 0.109 E -0.424 0.020 £i -0.085 0.309 Sept. n=13 PI -0.544 J* -0.622 0.407 E -0.539 0.230 £ -0.537 0.138 E -0.355 0.335 Oct. n = l l P7 0.415 Ij -0.076 -0.547 T -0.276 0.514 E -0.737* 0.122 .E, -0.039 0.329 b  6  t  t  b  t  t  b  b  219  t  PI, J* , E ,Et , and  fixed  relative to  A p p e n d i x C l (Continued). Cbla  PI  Basin data subset April n=12 PI -0.313 Ij -0.0C9 •0.164 E -0.043 -0.533 E -0.013 -0.191 E -0.276 -0.546 May n=2S PI 0.138 I] 0.193 0.591 E -0.521 -0.403 E -0.518 -0.040 £« -0.509 -0.434 June n:=24 PI -0.281 I] 0.557 0.142 £ 0.178 -0.209 £ 0.480 -0.282 E 0.099 -0.150 July n=22 PI -0.452 Jj -0.096 0.052 £ 0.094 -0.198 E 0.773 -0.161 £, -0.160 -0.163 Aug. n=28 PI -0.273 Jj 0.421* -0.137 E 0.031 -0.206 £ 0.157 -0.195 E -0.054 -0.077 Sept. n=28 PJ -0.271 7j -0.443* 0.137 £ -0.253 0.603* E -0.323 0.193 £, -0.150 0.581* Oct. n=28 P7 -0.001 Jj -0.187 0.433* E -0.479* -0.459* £ -0.292 -0.573* iJj -0.565* -0.379 b t  b  6  t  b  fc t  b  6  E  Et  -0.532 -0.489 -0.320  0.817 0.726  0.266  -0.484* -0.259 -0.477*  0.603* 0.876*  0.321  0.025 0.231 -0.057  0.707 0.957  0.544  0.033 -0.108 0.132  0.455 0.921  0.118  -0.235 -0.008 -0.166  0.377 0.862*  -  -0.050 0.021 -0.076  0.738* 0.975  0.617*  0.015 -0.252 0.137  0.650* 0.954*  0.452*  A p p e n d i x C2. Partial correlation matrix of Ei, , PI, E, Et and Et, ; fixed relative to sample station, year, and J* . Chla  PI  E  £o  0.986* -0.597  -0.721  0.874* 0.556  0.125  0.959* 0.802*  0.649*  0.947* 0.872*  0.704*  6  0.924* 0.748*  0.486  6  0.909* 0.923*  0.704*  0.788* 0.665*  0.172  QS data subset April n=6 PI -0.534 E 0.3C9 -0.983* E 0.210 -0.939* E 0.526 0.438 May n=12 PI -0.365 E 0.418 -0.356 -0.023 E 0.102 E 0.604 -0.606 June n=12 P / -0.059 E -0.335 -0.724* £ -0.365 -0.728* E -0.180 •0.554 July n=10 P / -0.662* E 0.751* -0.837* £ 0.751* -0.837* E 0.752* -0.925* Aug. n=13 PJ -0.439 E -0.355 0.096 £ -0.444 0.034 Et -0.223 0.208 Sept. n = l l P / -0.406 E -0.598 0.200 £ -0.053 -0.370 E -0.612 0.458 Oct. n=12 P/ 0.447 E -0.533 0.169 ,E -0.820* -0.092 Et -0.101 0.026 b  t  b t  6  t  6  t  t  6  221  A p p e n d i x C 2 (Continued). Chla  E  E  0.753* 0.693  0.132  -0.166 0.145 -0.215  0.566* 0.839*  0.233  -0.215 -0.326 -0.143  0.721* 0.960*  0.573*  -0.201 -0.157 -0.172  0.461* 0.925*  0.134  t  -0.247 -0.195 -0.102  0.390 0.858*  -0.105  t  0.617* 0.192 0.599*  0.740* 0.975*  0.621*  -0.516* -0.532* -0.491*  0.676* 0.961*  0.508*  PJ  Baaln data subset April n=13 PI -0.306 E -0.095 -0.534 E -0.054 -0.129 E -0.315 -0.528 b  t  b  May n = 2 S  PI E E E b t  0.030 -0.498* -0.494* -0.483*  June n=24  PI -0.437* E 0.197 E 0.435* E 0.158 July n = 2 2 PJ -0.450 E 0.098 £ 0.770* E -0.150 6  t  t  t  Aug.  n=28  PJ -0.239 E 0.147 £« 0.170 E 0.017 Sept. n = 2 8 PJ -0.237 E -0.308 Et -0.350 E -0.205 Oct. n = 2 8  PJ E E E 6  t  0.090 -0.485* -0.357 -0.554*  A p p e n d i x C3. Partial correlation matrix of Pf , T, , and \I, £ fixed relative to sample station, month, and year. Pf  Pf  Pf  QS data subset n= 44 .365* .723' .383* .130 .042 -.062 PR r, .641* .297 .451* .625* .302 .447* Ts .571* .297 .452* Ti .436* .257 .377* h .155 -.191 .136 h .096 -.244 .160 h .012 -.248 .110 ho -.176 -.425* -.013 Basin data subset n=93 .788* Pf .469* .582* Pf .264* .279* .524* Pfo r, .073 .128 .208* T .080 .204* .270* Ti .081 .192* .248* Tn -.003 .089 .224* h .207 .304* .176 h .170 .314* .090 h .086 .253* .112 ho -.031 .111 -.055  Pfo  r,  T,  T  .239 .248 .238 .013 .017 -.155 -.228 .162  .987* .958* .697* .030 -.020 -.080 -.033  .972* .728* .006 -.028 -.086 -.026  .206* .238* .203* .129 .173 .010 -.032 -.138  .871* .702* .386* .291* .244* .131 .162  .861* .536* .279* .255* .169 .175  7is  Ih i  .821* .057 -.055 -.099 -.008  -.162 -.025 -.046 -.092  -.821* -.641* .328*  .928* 2 .571* .632*  .720* .283* .255* .169 .165  .220* .191 .106 .154  .871* .294* .459*  .543* .676*  s  |Js I  Ih i  Pf Pf  s  223  .345*  A p p e n d i x C4. Partial correlation matrix of PBz, and Tz fixed relative to sample station month, year, and /* . ' Pf  Pf  Pf  QS data subset n==44 .402 P .720* .403 Pf .066 .093 -.057 Ti .428 .291 .234 T .395 .285 .208 T .303 .251 .149 7IJ .207 .199 .067 Basin data subset n=93 Pf .792' Pf .490 .612 .261 Pfo 277 .427 Ti .020 -.038 .078 Tt .005 -.032 .082 T -.028 -.106 .017 Tis -.066 -.195 -.043  pB  r,  Ts  Pf  B  S  0  s  s  s  .295 .303 .296 .183  .991* .996* .850*  .979* .874*  .939*  -.058 -.061 -.102 -.120  .929' .845* .722*  .939* .826*  .918*  224  Appendix D Daily carbon uptake calculation procedure. Assuming that the P-I function is constant throughout the day and given an estimate of nocturnal respiratory loss r^, X)JJ /<I? > can be used to calculate daily carbon fixation P = £ If/ < i j > Ti - d/24, where d is the number of hours of darkness. Values of r«i in phytoplankton range from 10 to 25% of Pf for non-motile forms, and up to 70% of P™ for motile forms (Falkowski and Owens 1978). Since biomass is assumed to be constant over the period of net carbon assimilation in making this approximate conversion from hourly to daily uptake, the adoption of an r<j estimate at the low end of this range (i.e. 10% , assuming non-motile phytoplankton) is prudent as it offsets the error incurred in the discontinuous growth approximation. It also compensates for the approximate nature of the 'net' interpretation of . The dark period (d) was set equal to 8 h. d  1  225  A p p e n d i x £ May/82, Aug./82, and July/83 cruise data summary.  D A T E  8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 B 8 B 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 B 8 8 B 8 B 8 8 8 8 8 8 8  2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 S 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5 2 0 5  S T A T I O N  1 1 1 1 1 1 1 1 1 1  1 1 1 1 1 1 1 1 1 1 11 1 1 12 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 3 1 3 13 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 3 1 5 1 5 1 5 15 1 5 1 5 15 1 5 1 5 1 5 1 5 1 5 1 6 16 16 1 6 1 6 1 6 16 1 6 1 6 1 6 1 6 1 6 1 7 1 7 17 1 7 1 7 1 7 1 7 1 7 1 7 1 7 1 7 1 7 1 8 IB 1 8 1 8 1 8 1 8 1 8 1 8 1 8 1 B 1 B 1 8 1 9 1 9 1 9 1 9 1 9 1 9 1 9 1 9 19 19 1 9  1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1  D E P T H ( M l  R C H L A  0  0.7  02  0  0 4 0 6 0 8  0.5  0  0 . 7 0 4 0 6 0 . 6 0 . 7 0 3 0 . 5 O S 0 . 3  1 .2  12  14 16 18 20  P 4 0 0  0 0 0 1 0  02  0 4 0 6 0 8 10 12 14 16 IB 2 0 P 4 0 0 0 2 0 4 0 6 0 8  0 0 0  00 02  0 4 0 6 0 8 10 12 14 16  5  18  6 0 2 4 6 8  0.7 8 0 6 0 . 5 2 0 4 3 6 3 . 6 5 . 7 2 0 2 2 . 6 1 . 1 0  '  1 .3  -0.8  1  11.9  -0.  12.8  0 0 0 0 0  12  6  0  -0 0 0 0  -0.0  . . . .  0 . 0 . 0 . 1 0 . 0 . 0 . 0 . - 0 . 0 .  1  3 3 2 5 5 7 .7 7 6 3 3 3 6  12  1 . 1 . 8 3 . 4 . 2  0. 0. 0 . 1. 0. 0 0 . 0 . 1 - 0 . -  1 1 0 1 2 0 4 4 .2 2  6  13. 1 13 0 10 8 10 1 8 11.4 12.8 12 4 12.4 12 6 10  12^2 1 2 . 8  1 .4 1 . 7 1 4 1 .8 1 8 2 . 1 1 .7 1 4 1 . 4 1 . 3 1 .4 1 .4 1 .2 1 .0 1 .0 1 .2 1 .2 1 .4 1 . 1 1 2 1 . 1 1 . 2 1 .5 1 . 1  8 9 . 5 10.6 11.6 11.5 1 1 .9 12.4 1 3 . 0 14.1 13 6 1 3 . 0 1 0 . 8 1 .9 7 . 6  0 . 6 0 . 4  0 . 0 0. 1  0 . 5  -0. 1  -0.2 0 . 3 0 . 5  1 .0 0 2 - 0 . 0  1 0 1 . 2 0 . 7 0 . 6 - 1 . 0 0 5 0.4 O S 0 . 7 2 9 1 8 1 .4 1  2  0.8  0.6 0 8 O S 0 . 6 15 4 3 S 15 2 0 0 0  . 3 O .5 .4 . 9 . 8 8  - 1 0  0 . 0 . -0. 0 . - 0 . - 0 . 0. 1 0. 0. 0  2 2 1 3 3 0 1 .7 1 1 2  1 .0 -0.2 1 .0 0.8 0 4 - 0 . 4 -0.0  0  1  0 2 0 2 - 0 . 2 7 . 9  0 . 5 0 9 7 . 3 0.2 0 3 -1.7 0 . 9  10.4 1 1 . 9 1 2 . 5 1 2 . 9 1 3 . 9 14 0 13 5 1 5 . 9 1 1 . 0 8 . 0 8 . 4 8  1 . 3 4 . 7 3 3 5.9 '3.4 1 .5 1 . 1  0 . 5 0 . 5 0 . 7 2 8 7 4 2 -1 0 0 2  4 2 . 7 . 3 1 .8 . 7 . 7 5  0.2  2  11.3 12.5 1 3 . 3 13.4 13 6 14 4 14 4 1 3 . 3 1 .3 5.0 5.9 5 . 7 0 . 9 10.7 1 2 . 3 1 0 . 6 13  2 . 6 5.5 3 5 6 . 5 2 . 6 2 . 0 1 9 1 .3 0 8 0 . 6 O B 0 7 0 6 2 . 5 9 3 8 . 7 5 1 2.4 4 . 5 1 . 1 1 . 0 2 . 7 0.8 1 . 0  8  NH4 ( U M )  9  1 .0 8  -0. 1  10.7 10.2 10. 3 11.4  -0.  2.5  0  0.2  0.6 0.7  2  5  NO 3 ( U M )  0.5  1 8 0 . 8 2 . 9  0.8  10 12  -0.3  4 6  0  1 . 1 0 9 3 0 1 1 1 .0 0 . 7 0 . 6 0 1 0 6 0 . 4 0 . 6 0 . 4 1 3 2 7 2 . 4 2 0 1 .4  20  20  5  4  0.5  20  14 16 18 2 0 P 6 0 0 02 0 5 0 6 0 7 0 8 12 14 16 IB 2 0 P6 0 0 0 2 0 5 0 6 0 7 OB 12 14 16 1B 2 0 0 0 0 2 0 3 04 0 6 0 8 12 14 16 18  6 6 6 0 6  0 . 6 0 6 0 4 0 . 5 O . S 0 . 7 0 3 0 4  P 6  P 0 0 0 0 0  4  0.5  2 4 6 8  1.0 0.9  0.5  10  1 1 1 1  8  P H A E O -0.4  P 4 00  10  '  C H L A M G / M " • 3  6  2 . 1 1.2  0.2 0 7 0 . 2 0 . 9 1 . 3 0  4  0 . 6 - 0 . 0 0. 1 1 .5 1 . 2 1 . 2 0 4 1 1 . 0 0 . 8 0 6 0  2  1  .4  10.3 0 1 0 . 6 10.7 12 7 13.4 12 3 1 3 . 5 13 3 14 .4 1 4 . 6 14 . 3 0 3 O.B 1.6 8 7 11 .7 13 2 1 2 . 9 13 2 B 6 14 0 14 2  1.7 1  0.8  B 5 5 . 2 3 7 9 0 8 5 6 4 3 . 3 3 . 3 1 .7 1 .4 1 . 1 4 . 2 1 .8 1 3 1 8 1 .8 2 . 2  3.2 16 4 1 . 3 1 . 2  226  .0  1 . 1  2.4  0 . 6  1  1 .6 8  1 2  1.2  0 . 0 0 . 0  0  0.0  0 0 . 0  0 . 0 0.0  0 . 0  0  0 - 0  0 . 0  1 .5  1 .2  1 1 2 . 0 .  0 8 0 . 0 0 . 0  3 . 1 3 0 2.9 0 . 0  1 . 3 0 . 0 0 0  0 . 0  0 . 0  0 . 0  0 . 0  2.7  3  4 . 5  5 . 0  1  .6 0 . 8 0 . 7 0 . 6  3. 1 0 . 0 .  4 . 0 1 .s 1.6 O B  0 . 0  0 . 0  0 . 0  0 . 0  1.0 1 . 3 2 . 6 1.9  1 2 1 . 2 2 . 4 1 8  0 . 4 1 . 1 2 . 5 2 . 7  0 5 1 . 1 2 . 3 2 . 7  0 . 0  0 . 0  0 . 0  0 . 0  8  8 . 0  3  2  4.8 1.5  0  4.2 1 . 1 0 0 0 . 0  2 4 1 . 1 0 0 0 . 0  0  1 0 . 0 . 0 .  2 3 0 . 8 0 8 2 . 7 5 3 4 . 3 4 . 8 5.0 1 .3 2 . 3 0 8 2 3 3 1 2 . 6 8 9 8 3 3 . 1 4 .4 1 .9 2 8 2 .9 2 . 7 2 5  A l 1 A I 2 (MGC/MGCHLA/HP.  PB 1 P B 2 ( M G C / M - • 2 / H R )  0  .8 .9 8 0  4 8 0 0  4  8.3  1 .5 0.0 0 . 0  0  0  0.0 0  4 .7 0 0  1  9  0 . 0  0  0  0 . 0  0  8  6  9  8  2.4  2 . 7  7 . 1  7 3 8 4 0  . 1 . 2 . 0 . 0  1 . 3 1 .4 0 . 0  1 .2 1 9 1 .6 0 0 0 . 0  3.7 0 . 0 0 . 0  0 . 0  0  0  0 . 0  0 . 0  0 . 0  0  0 . 0  1 .5  5 . 3  2 . 8 3 3 3 6  6 .5 3.2  1 .4 1 .7 2 . 8  3  0 . 0  0 . 0  0  0  0 . 0  0 . 0  0 . 0  0  0  0  4  2.5  0  Appendix E (Continued). D A T E  B 2 0 8 0 9 8 2 0 8 0 9 8 2 0 6 0 8 8 2 0 8 0 9 8 2 0 8 0 9 8 2 0 8 0 9 8 2 O 8 0 9 8 2 0 8 0 9 8 2 0 8 0 9 8 2 0 8 0 9 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 0 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 6 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 1 8 2 0 8 1 2 B 2 0 B 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 2 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 8 2 0 8 1 3 • 2 0 8 1 3 • 2 0 8 1 4 • 2 0 8 1 1 • 2 0 8 1 4 • 2 0 8 1 4 • 2 0 8 1 4 8 2 0 8 1 4 8 2 0 8 1 4 8 2 0 8 14 8 2 0 8 1 4 8 2 0 8 1 4 8 2 0 8 1 4 8 2 0 6 1 4 8 2 0 8 1 5 B 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 5 B 2 0 B 1 5 8 2 0 8 1 5 B 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 5 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 6 1 6 B 2 0 B 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 8 1 6 8 2 0 B 1 6  S T N  O C P T H ( M )  2  0 0 0 0 2 0 0 4 0 0 6 0 0 8 0 1 0 0 1 2 0 1 4 0 1 8 0 2 0 0 O O O 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 O G O 0 8 0 1 2 0 1 6 0 2 0 0 0 0 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 O 6 0 O B O 1 2 0 1 6 0 2 0 0 O O O 0 1 0 0 2 0 0 4 0 0 6 0 O B O 1 O 0 1 4 0 1 6 0 0 0 O O O 0 1 0 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 6 0 0 8 0 1 2 0 1 6 0 2 0 0 O O O  2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2  0 1 0 0 1 5 0 2 0 0 3 0 0 4 0 0 5 0 0 6 0 0 6 5 0 8 0 1 4 0 2 0 0 O O O 0 1 0 0 1 5 0 2 0 0 3 0 0 4 0 0 4 5 O 6 0 0 7 0 O 8 0 1 4 0 2 0 0 O O O 0 1 0 0 1 5 0 2 0 O 3 0 0 4 0 0 5 0 0 6 0 0 6 5 O B O 1 4 0 2 0 0  1 1 1 1  1 1 1 1  1 1  2 2 2 2 2 2 2 2 2 2 2 2 1 1 1 1  1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1  T C H L A  C H L A  P H A E O  N O 3 N H 4  ( M G / M 3 ) 5  0  6 0 4  7 3 9 2 1 2 1 1 4 1 0 0 6 0 5 4 S 5 S 7 4 6 4 5 9 5 7 5 6 2 0 1 6 0 6 0 5 0 5 3 1 8 4 10 0 12 7 15 5 13 0 10 5 2 2 2 2 1 1 0 5 0 5 8 5 9 7 9 6 10 4 8 3 9 3 7 9 2 2 1 3 0 8 4 9 4 9 4 9 C 9 8 7 8 0 5 2 2 1 0 8 0 6 0 3 1 0 2 3  s9 4 4  0 4 4 0  (1*1)--  - 0 0 0 -0  3 6 3 4  3  4 7 £ 5 5 5  9 0 5 9 5 1  0 0 - 0 -0 - 0 0  8 3 5  2 0 6  8  1  0  0  12 12  2 0  0  1 1  t o  0  0  3  s  8 9 8 9 7 8 7  4 1 9 8 9 3 5  -0 0 0 0 0 1 0  2 4 6 5 2 1 3  4 4 4 7 8 7 5  8 6 6 0 2 5 4  -0 0 0 -0 0 0 - 0  1 2 3 5 4 4 6  6 8 12 14 19 12 8  4 2 7 6 3 2 3 6  -0 0 0 2 0 0 1 0  2 4 7 1 9 4 5 5  7 7 7 8 1 1 12 11 4 3  0 4 7 3 8 2 3 £ 6  0 0 2 0 1 0 0 0 0  3 2 5 8 4 6 3 3 1  0 -0 -0 - 0 2 0 0 0  6 1 3 1 2 5 6 7  5 3  0 1 2 2 3 3 6 13  2 3 11 13  2 1 0  4 3 3 1 6 2 5 2 6 1 4 8 4 1 0 0 9 0 7 8 2 5 8 1 8 8 8 1 9 .8 1 1  6  2 2 2 2 2 2 2 2 2 2 7  8 9 9 7 7 7 0 3 2 6  6 7 6 6 6 7 7 7  8 8 3 5 4 4 3 0  4  4 4 5 0 4 7 4 9 7 7 4 1  1  5  0 2 5 9 8 6 6  a 14 14  1 4  2 2 2 1 5 7  2 5 5 7 8 1  1 1 12 11 13 13  9 3 3 7 6  1 2 S 11  6 1 0 9  11  9  12 0 0 2 7 7  4 7 6 8 0 8  10 8 1 1 4 12 1 1 1 8  5  •  2 2 S  0 4 8 7 2 7 8 9 8 6 6  2 4 . 18 6 . 3  5 8 7 6  0.0 0.0 0.0 0.0 0 0 0.0 2 3 . 1 2 2 . 1 2 4 . 1 2 0 . 3 12 3 12.5 8 . 5 2 . 6 1 .2 0.0  0.0 0.0  5  14 15 2 0 13 9 2 1 0 2 7 7 9 9 13 13 11  9  13 13 14 13 13 11  6 6  5  4 4 7  • P 1 » P - ( M G C / H 3 / H R )  2 4 3 1 3 9 6 6 12 0 1 22 11 . 1 12 .9  3  5  4 5  1 9  6 9 10 13 11 0 3 0 0 1 0 2 8 11 12 11 12 4 4 4 2 3 1 8 11 10  0 5 3 0 4 B 9 9 7 1 5 2 6 9 5 0 3 3 4 6 3 2 7 .0 . 1 .9  5  S 5 5 5 5 5 5 5 S 6 2 4 3 3 2 2 2 2 3 2 2 3 3 3 3 2 3 3 3 2 3  2 5 4 2 4 5 2 6 6 3 5 4 7 8 9 9 3 9 3  4 3 4 4  0 8 7 S 3 9 3 7 4 4 3 6 4 2 4 9 4 5 5  4 6 5  S 4 6 6 7 6 6 4 3 3 3 3 3 4 2 3 3  12 .3 13 . 3  e7  4  a  227  9 7 5 9 1 £ 6 3 5 3 9 4 5 6 8 8 B 0 7 6 .4 6 .8 .0 .7  18 .2 51 .5 3 5 1 6 9 1 4 8 . 1 3 0 . 8 2 3 . 5 2 . 2 0 . 0 0 . 0 0 . 0 1 .0 4 0 . 1 2 7 . 8 36 7 1 2 . 6 6 . 6 3 . 0 0 . 0 0 . 0 0 . 0 0 . 0 1 7 . 7 2 0 . 9 24 .8 2 4 . 2 5 5 . 3 2 2 . 9 8 7 2 . 2 4 . 3 1.3  0.0 0.0  1 4 . 3 2 6 . 6 3 4 .9 2 5 . 0 3 2 . 2 3 0 . 8 2 3 . 9 1 1 . 3 7.6 0 . 0 0 . 0 0 . 0 1 0 . 9 2 9 . 2 33 3 37 .8 18.2 3 4 . 2 4 0 . 4 1 6 . 7 4 . 9 3. 1 0 . 0 0 . 0 0 . 0 8 . 2 6 . 5 12 . 1 2 0 . 2 2 2 . 7 2 5 . 4 5. 1 0.0 0.0 0.0 0.0  22 19 6 4 1 1  0 9  0 7 6 3  0 0 0 0 0  0 0 0 2 E 0 2 8 E 3 8 7 18 5 19 1 11 8 9 3 2 7 1 0 0 0 0 0 0 0 18 1 4 2 8 36 B 47 0 4 3 9 3 0 7 2 3 5 3 0 1 0 0 0 0 0 0 0 4 0 2 2 9 8 1 3 3 1 1 7 7 4 0 0 0 0 19 2 0 2 3 21 57 2 0 8 2 0 0  0 0  0 0 0 0 0 5 4 8 4 1 7 0 0 0 0 0 5 2 4  IB 23 24 24 42 7 26 4 22 6 13 2 9 9 0 0 0 0 0 0 13 2 28 6 37 3 37 0 42 8 33 5 4 5 8 17 2 4 9 2 7 0 0 0 0 6 5 8 6 6 2 12 9 17 2 2 0 8 18 9 5 2 0 .0 0 0 0 0 0.0  a  2  A l l ( M G C / M G C H L A / H R ) 4 3 1  0 0 0 0  9 2 4 9 0 0 0  0 0 0 0 0 0  1 4 0 3 2 3 2 2 1 2 2 1 5 1 3 0 7 0 0 0 0 0 0 5 8 6 1 3 5 5 4 3 1 2 4 2 2 1 0 0 0 0 0 0 0 1 9 4 7 2 9 3 8 1 2 0 8 0 3 0 0 0 0 0 0 0 0 3 6 4 2 5 0 3 5 £ 3 2 9 1 7 1 0 5 3 2 4 0 0 O 0 £ 3 4 5 3 2 7 2 3 2 0 1 2 0 B 0 8 0 0 0 0 0 0 5 1 3 9 4 3 4 0 2 0 2 £ 3 1 1 4 1 0 0 B 0 0 0 0 0 0 3 3 2 1 3 2 3 0 1 £ 1 1 0 4 0 0 0 0 0 0 0.0 5  e  4 . 4 3 . 3 1.3 1.2 0.7 0.6 0.0 0 0 0 0 0 . 0 5.8 5 2 5 . 2 2 . 9 3 2 2. 1 1 .7 1 .3 0 6 0 . 0 0 . 0 0 . 0 5 . 7 5 1 3 . 7 3.7 2 8 2 . 4 2 . 2 1 . 4 0 5 0 . 0 0 . 0 0 . 0 4 . 7 3. 1 3 4 1 . 1 0 . 9 0 . 0 0 . 0 0 . 0 0 . 0 3 6 4 2 4 . 7 3.2 6 . 6 2 5 1 . 7 0 9 0 . 0 0 . 0 0 . 0 0 . 0  a  i 4 2 3 7 2.6 3 . 0 1.7 1 . 1 1 0 1 . 1 0 . 0 0 . 0 0 . 0 £ 2 3.9 4 8 3.9 4 . 7 2 6 3 5 1 .4 1 .0 0 . 7 0 . 0 0 . 0 2 £ 3 . 1 2 . 0 3 4 2 5 1 .5 0 . 8 0 4 0 . 0 0 . 0 0 . 0 0 . 0  A  I  2  Appendix £  D A T E • 8 • • • • • • 8 8 8 8 8 8 8 8 8 • 8 8 8 8 8 8 8  2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2  2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0  0 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8  (Continued).  S T N * 1 1 1 1 1 1  1 7 7 7 7 7 7 7 17 1 7 1 7 1 7 1 7 17 1 7 1 8 1 8 1 8 1 8 1 8 1 8 1 8 1 8 1 8 1 8 1 8  2  2 2 2 2 2 2 2 2 2 2 2 2 2  2  2 2 2 2 2 2  2  2 2 2  D E P T H ( « )  T C H I A  OOO 0 1 5 0 2 0 0 3 0 0 4 0 0 4 5 0 5 0 0 5 5 0 6 0 0 6 5 0 7 0 0 8 0 0 9 0 1 4 0 0 0 0 0 1 0 0 1 5 0 2 0 0 3 0 0 4 0 0 4 5 O 5 0 0 6 0 0 7 0 140  2 . 1 5 8. 6 2$ 0  •  32 2 27 2 16 0 1 3 .7 16 2 22 5 17 .7 11 . 1 4 3 1 .8 2 .9 5 .5 6 3 7 8 15 9 11 .9 10 .2 B B  6 .5  1 .8 0 .6  C H L A P H A E O (HG/M3)  3 2 C. 0 8  2  — -0. 3  -0. 5 0. 1  3 0 . 1 2 7 .2  1. 7 -1. 4  12 7 14 6 2 0 6 15 . 4 13 . 2 4.0  1. 0 2 0 1 .6 3. 0 - 4 .3 0 3  2 .9 5. 7 6 . 1 7 .3 15 . 1 11 . 3 9 6 0 S .0  -0. 2 - 0 6 0 . 1 0 6 0 6 0 3 0 3 0 9 0 .5  •  N 0 3 N H 4 -• - ( U X ) - -  1 .7 3. 0 1. B 1. 0  6.0 1.0 4.8  4 4  1. 9 4 6 9 7 4.0 8 6 4.4 1 0 .8 2 . 6 7 8 3 9 9 0 3 9 9 3 2 . 8 1 0 .9 2 . 9 3.7 12 0 2 . 9 2 7  2 .0  9 .5 13 .0 12 .3  228  PP1 PP2 - ( • O C / M / H B )  7 2 5 2 9 25 32 22 11 8 8 8 8 2  a 5 0 .B 9 5 .4  . 1 9 .3 .0 .8  .0 .9 .5 .2 . 4 .5 . 3 .5 16 3 S .3 0 0 0 .0  0 9 26 21 25 25 18 23  412 « 1 1 ( • G C / B G C H L A / H O )  8 0 27 4 28. 6 22 7 35 3 2 7 .7 11 2 8 8 8 7 7 0 8 .9 2 5  3. 7 4 .3 3. 4 1 .0 1 .0 0 8 0 7 0 E 0 6 0 4 0 5 0 .2  4.6 3.3 0 9 1 . 1 1 .0 0.7 0 6 0 5 0 . 3 0 5 0.2  0 9 25 20 2 5 23 18 23 15 5 0 0  0 3 4 3 3 1 1 2 1 1 0 0  0 . 0 3 . 3 4 . 6 3 . 3 3 3 1 .5 1 6 2 . 3 1 6 0 9 0 . 0 0 . 0  .0 5 . 4 8 8 .5 .8 .0 .7 0 .0 .0  0 .5 8 .3 .2 6 .5 .3 6 .0 .0 .0  11  Appendix E (Continued). D E P T H (M) 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 3 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 4 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 15 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 5 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 B 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 6 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 3 0 7 1 7 8 8 8 8 8 8 8 B 8 8 8 B 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 8 B 8  3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3  0 7 1 7 0 7 1 7 0 7 1 7 0 7 1 8 0 7 1 8 0 7 I B 0 7 1 B 0 7 1 8 0 7 1 8 0 7 1 8 0 7 1 9 0 7 19 0 7 1 9 0 7 1 9 0 7 1 9 0 7 1 9 0 7 19 0 7 1 9 0 7 1 9 0 7 1 9 0 7 1 9 0 7 1 9 0 7 1 9 0 7 1 9 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0 0 7 2 0  1 1 1 1 1 1 2 2 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1  1 1  1 1 1 2 2 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1 1 1 1 1 1 2 2 2 2 2 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 1 1 1 1 1 1 1 2 2 2 2 2 2 2  0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 2 4 6 10 14 0 1 2 4 6 10 14 0 1 2 4 6 10 14 0 1  2 3 4 6 10 0 1 2 3 6 8 10 0 1 2 3  6 8 10  C H L A S E - M G / M > •31 9 8 0 52 0 4 6 0 2 6 0 34 0 2 0 1 3 8 0 9 6 1 0 6 1 0 2 0 4 6 0 2 6 2 16 1 16 0 6 8 0 6 6 0 5 2 0 5 0 2 8 6 1 18 1 0 4 0 78 0 42 2 0 0 0 0 0 2 0 0 0 0 0 1 0 0 0 0  9 7 9 8 6 8  2 2 1 1  5 2 10 6 0 OB  0  14  S2  3 0 42 6 6 92 7 0 4 8 24 32 32 8 2 5 4 5 0 2 6  2 1  8 5 6  0 7 6 0 6 5  0 4 8  4 4  9 8 7  0 1 1  7 2 2  2 3 2 0 19  2 0 1  3 1 4  3  0 0 1  1 1 1  3  2 2  4 6  1B 19 17  0 1 3  7  2  B  1  2  17  B  5  6  2  3  0  8  14  5  8  6  2  4  1  2  2 0  3  3 7 7  3 9 9 9  1 1 2 2  0 6 3 8  0 0 1 2  5 8 1 1  15 14 21 2 0  1 3 5 0  B 3 7 8 8 9  0 6 5 2 5 1  2 0 2 3 2 2  7 8 2 4 7 8  18 0 1 1 1 1  6 0 8 2 8  31  21 21 21  8 5 8 3 6 3  4  6  1  4  1  3  7 B 8 9 4 6 8 7 7 8 6 7 7 7  6 7 7  9 4 6 2 8 5 4 7 2 2 3 2 3 5 9 4 3 7 2 9 5 1  2 2 2 2 1 2 2 2 2 2 1 1 1 2 2 2 0 1 2 2 2 2  B 5 B 6 5 0 7 5 3 8 3 7 9 1 7 8 4 2 2 0 4 4  4 8  3 7  0 1  5 7  9 9 8  2 3 9  3 2 2  3 6 4  2 2 7 9 9 9 0  4 7 2 0 6 3 4  0 0 1 2 2 2 0  3 4 7 2 8 7 5  4  0 1 1  8 1 1  8  0 3 B 1 8 2 1 64 0 8 4 0 8 0 0 34 16 0 7 5B 14 3 2 10 1 0 2 0 32 0 0 2 6 3 0 6 36 5 9 0 1 6 6 1 2 0 0 4 8 0 2 6 5 12 4 74  3  0 1 2  7  4  0 44  ( M G / M " 2 / H R I - ( M G C / M G C H L A  - - ( U K ) 2 5 7  9 5 5 8  1 1 1 0 0 1 1 1 2 1 1 1 1 1 1 0 1 1 1 1 1  21  28  0  2 0 2 1 2 0 21 2 3 19 21 18 17 19 21 21  2 3 6  2 8 2 6 9 3 4 5 0 2 3 0 0 3 2 8 2 2 0 6 2  18  2  19 21 21 2 2 18 2 0 15 17 16  1 1 5 9 7 1 5 7 6  0 7 1 1 1 0 1 5 1 4  2 3 21 16 22 2 0  4 7 6 2 3  5 9 5 2 4 4  14 17 2 0 2 0 22 21 2 5  7 3 4 9 2 3 0  2 0 2 0 19  7 3 2 4 4 1  0 0 1 1 1 1  1  3 2 2 6 1 1 2 2  32 8 6  8  8  0 6 2  4 9 0 4 4 6  9 0 1  7 2 5  2 0 0  0 6 6  0 0  6 4  22 2 0 2 1  3 7 9  8 1 3  0 1 2  7 4 1  1 0 1  0 9 2  21 B 2 1 2 2 0 9  2 3  9 1 1 2  0 0 0 1  9 9 7 3  0 0 0 0  5 8 6 B  17 IB 19 2 0  9 B 3  3 O 0 1 1 2 2 2  0 4 4 3 8 6 5 8  1 0 0 1 1 1 1 1  3 4 5 2 2 4 9  2 0 18 IB 2 0 2 0 21 2 3 21  10 4 3  S 1 2  1 2  4 8  2 6 5 2 2 0 1  1 3 7 9 8 4  3 9 7 4 3 1 1 2 0 14 5 3 2 2  2 5 9 2 6 5 3 6 3 2 7 2  4 5 3 3 2  1 1 7 5 4 3 9  0 4 1 .2 0 . 5  7 5 3 1 2 . 6  1 .2 1 . 7  0.4 4 . 4  6 7 4 9 7  1 . 2 0.7 0 . 5 0 . 6 1.0 0.7  8 3 2 2 - 0 0  3 1 1 19 13 4 4 2 2  2 4 5 4 7 0 4 7 3  1 3 1.0 1 . 0 1 .7 1 . 3 1 . 3 1 . 9 1 9 2 . 0 2 3 3 . 0 1 .7  7 . 0 5.4 3 . 2 1 . 3 0 6 1 3 . 9 15 5 6 2 1 . 9 -0 9 0.8  3 5 2 1 8  5 6 3 2 2  3 4 9 5 3  3 3 3 2 3 2  13 0 12 6 17 2 1 1 0 5 9 7 4 4 5  8 16 22 10 7 5 6  3 7 1 2 1 8 1  4 2 4 , 7  4 3 47 3 3 22 12 7 5  2 3 1 1 5 S 5  51 4 8 3 3 3 0 13 8 9  9 4 8 6 2 6 2  4 . 1  3  2 6 16 9 8 6 7 7  0 1 9 3 8 7 1  6 .0 4 . 4 5 . 3 4 8 5 8 6 0 6 9  4 0 5 2 1 0 1  1  . . . .  3 2 6 7 0 6  4 . 3  4 4 4 3 5 . 3 4 8  7 . 9 . 0 9 . 5 . 5  0.7 2 . 3 0 6 0.2  1  2  1  1  1 . 3 2 . 2 4 . 0 1 3 4 . 2 - 1 . 2  8 4 3  2 4 2  13 4  6  2 1 -0 0  0 0 1 7  3 0 0 13 14 3 4 -0 0  0 2 6 3 3 8  1  1  2  9 9  3  3 3 1 4 - 1 5 - 1 8  0  5  1 9  3 3 2 1 4  0 5 3 0 7  /  1 5  a7 4 4 1 1 2 2 0 0 1 0 0 5' 5 2 1 1 0 O  4 0 4 0 8 8  86  0 6 4 4 9 0 3 9  8 6 6 2 6 8 0 2  44  2 4 7 2 6 4  0 0 6 4 2 0  4  6  9 0 2  7 9 9 10 10  6  4 4 1 4  3  229  4 8 3 . 7 4 . 7  3  5 . 4 8 5 . 9  6.6 5 3 4 .0 3 9 1 . 5 1 .5 -0 4  8  1  5 4 5 1 2 3  4  3.0  3 2 1 2 0 2 0  7 2 9 0  1 8  7 0 1  8  6 1 7  7 1  4 9 8 6 1  a  22 18 8 7  5 6 6  9 7 8 3 3 0  2.8  1  4  1 . 7 - 0 . 5 0.4 -2 1  8  7 5  IRRRDIRNCE  319  UE/MXX2/SEC 44  J  89 _  L  or x a _J x o  CO Cl-  IRRADIANCE UE/MXX2/SEC  or x  J.  in  X o  (XIO  )  1  28  89 56  /°  8  ^  (_) CD D_  1° MAY  13  28 IRRADIANCE UE/MXX2/SEC  Appendix F  (XlOL.l  56  o o  P-I model fit to the May/82 P-I incubation experiments.  230  A p p e n d i x F (Continued).  231  A p p e n d i x G The dominant photosynthetic species during the May/82, Aug/82, and July/83 cruises. Cell volumes estimated by eye-piece micrometer, averaged over all samples for each cruise. Photo, fraction is the fraction contribution to the total photo-autotrophic cell volume.  Species  #/ml  May/83 Cruise May 18 Depth=0 m Thalatiiotira decipient 20 cryptomonad 1806 Metodinium rubrum #2 3 Penn&te diatom 30 May 19 Depth=0 m 18 T. decipient Chaetoceroi sp. 612 60 cryptomonad At. rubrum #1 60 M. rubrum #2 2 May 19 Depth=3 m T. decipient 351 Chaetoeero$ sp. 125 cryptomonad 72 M. rubrum #2 4  CeU Volnme( pm* )  Total Photo. Volume( pm* ) Fraction  5.2x10 7.0x10" 1.4xl0 S.lxlO  1.0x10 1.3xl0« 4.2xl0 9.3x10  0.04 0.55 0.02 0.39  5.2x10 6.0x10* 7.0x10* 1.7x10 1.4x10*  9.4x10 3.7x10 4.2xl0 1.0x10 2.8xl0  0.62 0.25 0.03 0.07 0.02  5.2x10 6.0x10* 7.0x10* 1.4xl0  1.8xl0 7.5xl0 5.0xl0 5.6xl0 E = 2.0xl0«  0.91 0.04 0.03 0.03  s  s  4  4  4  s  s  s  s  4  s  s  4  Species  #/ml  Cell Volume( /im )  Aug.13 Depth=2.5 m chrysophyte hi. rubrum T. polychorda Gymnodinium sp. #1 Chaetocerot sp. cryptomonad Peridinium sp.  1124 12 4 11 71 206 4  7.3x10* 2.3xl0 2.2xl0 6.6x10 1.4x10 7.0x10* 3.9xl0  8  4 4  s  s  4  s  4  6  4  4 4  Total Photo. Volume( /im ) Fraction s  8.2x10 2.8x10 8.8xl0 7.3xl0 9.9xl0 1.4x10 1.6x10 £ = 17X10  0.49 0.17 0.05 0.04 0.06 0.08 0.10  2.6x10* 1.8x10 2.8x10 1.8x10 4.6xl0 2.0x10 £ = 3.5x10*  0.75 0.05 0.08 0.05 0.01 0.06  s  s  4  4  4  s  s  6  Aug.13 Depth= 3.0 m chrysophyte 3513 8 M. rubrum 201 Chaetocerot sp. 254 cryptomonad Gymnodinium sp. 7 #1 5 Peridinium sp.  7.3x10* 2.3xl0 1.4x10 7.0xl0 6.6x10 3.9xl0 4  s  2  s  4  s  s  s  4  s  232  Appendix G  (Continued).  Species  #/ml  Aug./82 Cruise Aug.Q Depth=0 m Thalassiosira polychorda19 Gymnodinium sp. #1 35 chrysophyte 226 Mesodinium rubrum 8 320 cryptomonad Peridinium sp. 8 Chaetoceros sp. 19 Ceratium sp. 1  Cell Total Photo. Volume( ftm* ) Volume ( /im ) Fraction s  2.5x10* 6.6x10 7.3x10' 2.3x10* 7.0xlO 3.9xl0 1.4x10* 4.4xl0 s  a  4  4  4.8x10 2.3x10 1.6x10 1.8x10 2.2x10 3.1x10 2.7x10* 4.4xl0 £ = 1.7xl0  0.29 0.14 0.10 0.11 0.13 0.19 0.02 0.03  2.9x10 1.4x10 1.8x10 1.1x10 4.4xl0 £ = 7.6x10  0.38 0.18 0.24 0.14 0.06  2.8xl0 3.9xl0 l.lxlO 3.3x10 2.4x10 2.2x10 £ = 8.6xl0  0.33 0.45 0.13 0.04 0.03 0.03  2.3x10 6.6xl0 3.2x10 1.5x10 1.2x10 2.2x10 1.5x10 6.4x10 £ = 1.3xl0  0.18 0.05 0.25 0.12 0.10 0.17 0.12 0.00  9.6x10 5.0x10 2.1x10 1.5x10 9.6xl0 7.8xl0 6.8x10* £ = 2.1xl0  0.47 0.24 0.10 0.07 0.05 0.04 0.03  4.1x10 2.3x10 1.4x10 2.2x10* 1.2x10 6.6x10* 1.2x10 1.2x10 £ = 1.2xl0  0.33 0.19 0.11 0.02 0.10 0.05 0.10 0.10  s  s  s  s  s  s  4  6  Aug.O Depth=10 m chrysophyte Gymnodinium sp. #2 cryptomonad Gymnodinium sp. #1 T. polychorda  394 6 250 16 2  7.3x10' 2.4x10* 7.0xl0 6.6x10 2.2xl0 3 s  4  s  s  s  s  4  s  Aug.O Depth=20 m cryptomonad Peridinium sp. r. polychorda Gymnodinium sp. #1 Ampkidinivm sp. chrysophyte  40 1 0.5 0.5 5 3  7.0x10' 3.9xl0 2.2xl0 6.6x10 4.8x10' 7.3x10' 4  4  s  4  4  4  s  s  s  4  A u g . l l Depth=0 m Peridinium sp. T. polychorda Af. rubrum Gyrodinium sp. chrysophyte Gymnodinium sp. #1 cryptomonad Chaetoceros sp.  6 3 14 1 169 33 219 4  3.9xl0 2.2xl0 2.3xl0 1.5x10 7.3x10' 6.6x10 7.0x10' 1.4x10 4  4  4  s  s  s  s  4  s  s  s  s  s  s  6  A u g . l l Depth=2.5 m chrysophyte Chaetoceros sp. A/, re&rum Gyrodinium sp. Gymnodinium sp. #2 Peridinium sp. cryptomonad  1316 354 9 1 4 2 97  7.3x10' 1.4x10 2.3xl0 1.5x10 2.4xl0 3.9xl0 7.0x10' s  4  s  4  4  s  s  s  s  4  4  6  Aug.lS D e p t h = 0 m chrysophyte Chaetoceros sp. A/. r»6r«m 7. polychorda cryptomonad Gymnodinium sp. #1 Dinophysis sp. Peridinium sp.  568 163 6 1 172 10 1 3  7.3xl0» 1.4x10 2.3xl0 2.2xl0 7.0x10 6.6x10 1.2x10 3.9xl0  s  4  4 3  s s  4  s  s  s  s  s  s  6  233  Appendix G Species  (Continued). #/ml  Cell Volume( P°< ) s  July/83 Cruise July 15 (Drift Stn. 6) Depth= 0 hi. rubrum #2 5 9 Peridinium sp. Dinophyti$ sp. 3 cryptomonad 38  m 2.3x10* 3.9x10* 1.7x10 1.1x10  July 16 (Drift Stn. 8) Depth = 0 Peridinium sp. 15 3 hi. rubrum #2 cryptomonad 3  m 3.9x10* 2.3x10* 1.1x10  s s  s  Total Volume( fim )  Photo. Fraction  1.2x10 3.5x10 5.1x10 4.2x10 E = 1.4x10*  0.09 0.25 0.36 0.30  5.9x10 6.9x10* 3.3x10 E = 6.3x10  0.93 0.11 0.00  1.6x10 1.0x10* 6.5x10 £ = 8.2x10  0.20 0.01 0.79  5.5x10* 4.6x10* 6.5x10 £ = 7.5x10  0.07 0.06 0.87  1.8x10 7.6x10 2.5x10 1.0x10 1.7x10 E = 1.5x10*  0.12 0.52 0.17 0.07 0.12  2.9x10 5.5x10* 6.9x10* E = 4.1x10  0.70 0.13 0.17  s  s  s  s  s  s  s  s  July 18 (Drift Stn. 11) Depth:=0 m Peridinium sp. 3.9x10* 4 hi. rubrum #1 6 1.7x10 587 1.1x10 cryptomonad s  s  s  s  s  July 18 (Drift Stn. 11) Depth:=2 m Peridinium sp. 1.4 3.9x10* 2.3x10* hi. rubrum #2 2 587 1.1x10 cryptomonad s  s  s  July 19 (Drift Stn. 13) Depth =1 m hi. rubrum #2 8 2.3x10* cryptomonad 678 1.1x10 Skeletonema coitatum 78 3.2x10 Thalatrionema sp. 195 5.2xl0 Dinophytii sp. 1 1.7x10 s  s  2 s  July20 (Drift Stn. 15) Depth = 0 m cryptomonad 260 1.1x10 Peridinium sp. 1.4 3.9x10* 3 2.3x10* hi. rubrum #2 s  s  s  s  s s  s  s  234  A p p e n d i x H A l time series sooplankton tow species abundance. Date Aug.9  Aug.9 Aug.ll Aug.ll Aug.13  Aug.13  Muggiae*  Cancer sp.  Depth (m)  eopepods #/m  atlantica #/m  Megalopa #/m3  2 8 10 16 20  458 104 238 158 37  16 117 204 155 20  •  1  14 2  3 3  2 1  1 3 5 20  189 192 253 203  •  •  _  17 18 56  -  17 7 2  1 2 3 4 5 8 20  120 243 77 313 3794 1202 243  8  8  2 2 2  Cancer sp  Zoea #/m 8  .  .  -  _  -•  12 7 11 23  >  27 26  -  -  235  Medusae Mysids #/m #/m 8  8  8  -  3 11  11 1 2  2 22 6 3  8 4 1 1  -  2 14 7 8 3  .1  -  3  .  3 1  Appendix I Aug./82 drogue track demonstrating large horizontal displacement east of Coal Harbour. Dashed lines indicate over night interpolations.  A p p e n d i x J Mean monthly / b for the I C M data set. t  Month  It  April May June July Aug. Sept. Oct.  675 600 831 780 770 770 450  237  in  UJ CO  Oco  Q_tO  CD CD '  0.0  34.97  69.95 TIME  T 104.92  139.9  (HRS)  Appendix K July/83 heat model 1 hr smooth of surface wind shear forcing time series.  174  TIME  (HRS)  Appendix K July/83 heat model 5 hr smooth of surface wind shear forcing time series.  Appendix K July/83 heat model radiant heat forcing time series.  136.4 TIME  „ Appendix  K J - W » b- « «  <0rC  '  mg  (HRS)  170.5  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0053272/manifest

Comment

Related Items