The Open Collections website will be unavailable July 27 from 2100-2200 PST ahead of planned usability and performance enhancements on July 28. More information here.

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies on the biological oceanography of Haida eddies Peterson, Tawnya Dawn 2005

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

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
831-ubc_2005-104317.pdf [ 40.8MB ]
Metadata
JSON: 831-1.0052395.json
JSON-LD: 831-1.0052395-ld.json
RDF/XML (Pretty): 831-1.0052395-rdf.xml
RDF/JSON: 831-1.0052395-rdf.json
Turtle: 831-1.0052395-turtle.txt
N-Triples: 831-1.0052395-rdf-ntriples.txt
Original Record: 831-1.0052395-source.json
Full Text
831-1.0052395-fulltext.txt
Citation
831-1.0052395.ris

Full Text

STUDIES O N THE BIOLOGICAL O C E A N O G R A P H Y OF HAIDA EDDIES by T A W N Y A D A W N PETERSON B.Sc. (Hons.), Mount Allison University, 1997 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Oceanography) THE UNIVERSITY OF BRITISH C O L U M B I A April 2005 © Tawnya Dawn Peterson, 2005 11 A B S T R A C T i Haida eddies are large (150— 300 km in diameter), long-lived (> 2 years), anticyclonic eddies that form each winter off the west coast of the Queen Charlotte Islands, in British Columbia, Canada (54°N, 130°W) and drift westward into the Gulf of Alaska, carrying nutrients and trace metals within their waters. This thesis documents the evolution of waters within an eddy called Haida-2000 over a 20-month period, focusing on nutrient and phytoplankton dynamics. Six surveys were completed over the course of the study and a second eddy (Haida-2001) was sampled for comparison between eddies spawned in different years. Compared to coastal surroundings, the Haida-2000 eddy was rich in nitrate, phosphate, and silicic acid when it formed in early 2000. Once Haida-2000 moved offshore, however, it was nutrient-poor compared to High Nitrate Low Chlorophyll (HNLC) waters of the Alaska Gyre. New production rates from seasonal nitrate drawdown in the mixed layer fall between previously reported values at Ocean Station P (50°N, 145°W) and coastal values along the northern margin of the Gulf of Alaska. Silicic acid drawdown by phytoplankton exceeded nitrate drawdown by three times in the eddy's natal spring, during which time a large phytoplankton bloom was observed via satellite imagery. While nitrate drawdown rates were similar throughout the study, significant silicic acid drawdown was observed in the first year of eddy evolution only, signaling a shift from diatom-dominated coastal assemblage to a more oceanic one as the eddy aged. The large phytoplankton bloom observed in April 2000 within Haida-2000 was not repeated later in this eddy's evolution. Chlorophyll a biomass was higher within Haida eddies compared to surroundings during most of the sampling times, and the spatial distribution tended to follow isopycnals. The coherence of photosynthetic biomass with isopycnal structure suggests that the Haida-2000 and Haida-2001 eddies possessed convergent flow regimes near their edges, in contrast to a stronger, larger Haida eddy observed in 1998. It is most likely that frictional 111 decay led to these patterns of convergent flow at the eddy edge in weaker eddies, while the stronger 1998 eddy exhibited divergent flow near its boundaries. Haida eddies were characterized by subsurface chlorophyll a maxima with the exception of uniform vertical profiles of chlorophyll a following strong storm-induced mixing in September 2001. Accounting for variability in spatial patterns of chlorophyll distribution improved estimates of maximum chlorophyll derived from surface chlorophyll concentrations. It was estimated that approximately 10% of daily primary productivity was grazed by planktonic ciliates. Based on abundance and biomass, this grazing pressure was relieved somewhat by the dilution of mixed layer waters by an intrusion of deep water during a period of vigorous mixing. Phytoplankton community composition was not dramatically different between Haida eddies and their surroundings during in Year 1 of their evolution. However, patterns ofsilicic acid drawdown in the first spring suggested that diatom growth was confined to the eddy and did not extend to surrounding waters. In Year 1, the phytoplankton assemblages at all sites were dominated by small representatives of coastal taxa, including Synechococcos spp., chlorophytes, cryptophytes, prasinophytes, euglenophytes and diatoms. In Year 2, larger differences in the phytoplankton assemblage were observed between eddy and reference waters, but all were dominated by more oceanic species, including haptophytes, pelagophytes, dinoflagellates, and chrysophytes. Key differences between eddy sites and their surroundings included a higher abundance of haptophyte algae at the eddy centre and the positive growth response to storm mixing by dinoflagellates at the eddy centre and diatoms in outside waters. Multivariate analyses showed that sampling sites first grouped together based on the time of sampling/distance from the coast, and then on the station type (centre, edge, or outside). Eddy edges and centre sites were more similar to each other in most instances than they were to outside reference sites. The use of microscopy to describe phytoplankton communites tended to lead to an overestimate of iv certain groups, particularly diatoms, due to the inclusion of weakly pigmented cells while chemotaxonomy did not. Signatures of offshore advection by eddy circulation were observed while Haida-2000 was close to the coast: the freshwater diatom Aulacoseira granulata was found at the eddy margin and in waters to the south, suggesting that delivery of terrestrial subsidies via this mechanism may be important to deep water organic material inventories. New observations of the calcareous dinoflagellates Thoracosphaera heimii occurring throughout this study may be related to water column stratification that has intensified in the Gulf of Alaska from the 1990's to the present. Rates of primary production were higher within eddy waters, and were negatively correlated with nitrate concentration, which was shown to be a reasonable proxy for dissolved iron content. Photosynthetic efficiency was highest at the eddy edges, and declined as iron concentrations decreased. Since nitrate had a negative relationship with iron concentration, it was inferred that the relationship observed between primary production and nitrate was due to iron limitation of phytoplankton photosynthesis. By including a term for nitrate in the Behrenfeld and Falkowski (1997) model of satellite-derived primary production, a better fit to actual data was achieved for the Gulf of Alaska. The smallest size class (0.2 - 5 jam) was responsible for an average of 68% of total primary production, with a proportionally larger contribution by this size class at the edges of Haida-2000 and Haida-2001. In order to examine the interplay between nitrogenous nutrition and iron concentrations, rates of nitrate and ammonium uptake were determined within Haida-2000 and Haida-2001. Rates of new production were highest in H N L C waters where concentrations of nitrate were high. When nitrate concentrations were low, new production rates were lowest, despite the presence of higher iron supply. Rates of new production were higher in Year 2 of eddy evolution when dissolved iron concentrations were lower than Year 1, suggesting that the relationships V among nitrate utilization, iron concentrations, and export production rely on both nitrate and iron concentrations. Growth of plankton in mixed water masses that are typical of stirring at frontal regions was examined in a shipboard experiment. Drawdown of nitrate and phosphate were higher within waters composed of an equal proportion of eddy and outside water masses. Silicic acid drawdown was highest within eddy waters. Primary productivity and the accumulation of particulate carbon and nitrogen were also higher within the mixed water. Increases in chlorophyll a were not significantly higher within the mixed water, however, nor were any of the phytoplankton pigments. In the mixed water, the growth of Pseudo-nitzschia spp. exceeded that of the eddy water or outside water alone, suggesting a species-specific response to the new microenvironment. The timescale of the response to the mixing was on the order of the shift-up response documented in upwelling regions, and may explain spatial patterns of enhanced biomass observed toward the eastern side of eddies. The results from this thesis indicate that Haida eddies impart variability in the distribution and productivity of phytoplankton that is predictable. The differences between Haida eddies spawned in different years appeared to be related to differing dilution rates and eddy energies. Studying biological processes within anticyclonic eddies lends insight into ecosystem function in the Gulf of Alaska and should improve our ability to construct more accurate nutrient budgets, understand the factors that shape ecosystem structure in the northeast subarctic Pacific, and predict patterns of biological activity at the mesoscale. TABLE OF CONTENTS VI A B S T R A C T II L I S T O F A B B R E V I A T I O N S X X V L I S T O F A B B R E V I A T I O N S X X V P R E F A C E '• X X V I A C K N O W L E D G E M E N T S . XXVIII G E N E R A L I N T R O D U C T I O N 1 C H A P T E R 1. M A C R O N U T R I E N T D Y N A M I C S IN A H A I D A E D D Y 31 7.7 Introduction 31 1.2 Materials and Methods 36 1.2.1 Sampling 36 1.2.2 Statistics 39 1.2.3 Definitions 39 1.3 Results 40 1.3.1 Eddy evolution and structure 40 1.3.1.1 Chronology 40 1.3.1.2 Temperature-salinity characteristics of eddy waters 42 1.3.1.3 Eddy decay 43 1.3.2 Surface processes: Evolution of mixed layer nutrient concentrations 44 1.33 Nutrient drawdown and new production 54 1.3.4 Deep-water processes: evolution of eddy core waters 56 1.3.4.1 Nutrients within the eddy core 56 1.3.4.2 Influence of eddy deformation and merging on nutrient distributions 60 7.4 Discussion 60 1.4.1 Biological drawdown of eddy nutrients 60 1.4.2 Physical processes influencing nutrient distributions 66 1.4.2.1 Eddy decay and secondary circulation 68 1.4.2.2 Overwash and surface dilution 70 1.4.2.3 Eddy deformation and merging 70 1 71 .4.2.4 Nutrient Supply 72 7.5 Summary 75 C H A P T E R 2. I N V E N T O R Y , DISTRIBUTION, A N D SIZE C L A S S S T R U C T U R E O F P H O T O S Y N T H E T I C B I O M A S S WITHIN H A I D A EDDIES IN T H E G U L F O F A L A S K A 77 2.7. Introduction 77 2.2. Materials and Methods 81 2.2.1 Sampling 81 2.1.2 Mixed layer properties 83 2.1.3 Biological measurements and collection techniques 83 2.2.4 Statistics 85 2.2.5 Ciliate biomass 86 2.3 Results 86 2.3.1 Mixed layer characteristics 86 2.3.2 Biomass distributions 89 2.3.3 Size-fractionated chlorophyll a 100 2.3.7 Vertical distribution of chlorophyll a in the eastern Gulf of Alaska 102 2.3.8 Ciliate standing stocks, biomass, and potential grazing impact 105 2.4 Discussion 1,07 2.3.7 Summary 121 C H A P T E R 3. SPECIES C O M P O S I T I O N O F P H Y T O P L A N K T O N WITHIN H A I D A EDDIES A N D T H E I R S U R R O U N D I N G S 124 3.1. Introduction 124 3.2. Materials and Methods 130 Vll 3.2.1 Sample collection 130 3.2.2 High performance liquid chromatography (HPLC) pigment analysis 131 3.2.3 Use of Chemtax® program 132 3.2.4 Phytoplankton enumeration and species identification 134 3.2.5 Statistics ' 3 5 3.2.6 Ecological indices 135 3.2.7 Cell carbon quota calculations 140 3.3 Results 140 3.3.1 Accessory pigments 140 3.3.2 Species composition by pigment analysis 147 3.3.4 Species composition by microscopy 155 3.3.5 Biological diversity and similarity indices 171 3.3.6 Vertical distribution of phytoplankton taxa 177 3.3.7 Relationships between environmental variables and taxonomic composition 179 3.3.9 General observations 185 3.4 Discussion 187 3.4.3 Comparison of chemotaxonomy and microscopy 195 3.4.6 Vertical structure and spatial patterns in species assemblages 203 3.4.7 Controls on species distributions in the Gulf of Alaska 206 3.4.8 Ecological indicators? 208 3.5 Summary 209 C H A P T E R 4. C H A N G E S IN P R I M A R Y P R O D U C T I V I T Y D U R I N G W E S T W A R D P R O P A G A T I O N O F H A I D A EDDIES IN T H E G U L F O F A L A S K A 210 4.1. Introduction 210 4.2. Materials and Methods 223 4.2.1 Sampling strategy -223 4.2.2 Sample collection 225 4.2.3 Light measurements 226 4.2.4 Nutrients and chlorophyll a 226 4.2.5 Carbon incorporation (organic fraction) 227 4.2.6 Calcification rates 228 4.2.7 Photosynthesis vs. irradiance (P vs. E) 229 4.2.8 Stratification 231 4.3 Results.. 231 4.3.1 Evolution of Haida-2000 and Haida-2001 231 4.3.2 Properties of the euphotic zone 232 4.3.3 Primary production 236 4.3.3 Total primary production: depth integrated 248 4.3.4 Photosynthesis vs. irradiance (P vs. E) relationships 251 4.3.8 Estimating primary production from remote sensing 257 4.4. Discussion 259 4.4.1 Total primary production 260 4.4.2 Size-fractionated primary production 262 4.4.3 Photosynthetic characteristics: (P vs. E) relationships 265 4.4.4 Primary productivity and iron in Haida eddies 267 4.4.5 Influences on eddy primary productivity 270 4.4.6 Summary • 271 C H A P T E R 5. N E W A N D R E G E N E R A T E D P R O D U C T I O N WITHIN H A I D A EDDIES 273 5.1. Introduction 2 73 5.2. Materials and Methods 277 5.2.1 Sampling 277 5.2.2 Rate determinations 278 5.2.3 Calculation of indices of nitrogenous nutrition 280 5.2.4 Calculation of p C: p N ratios 281 5.2.5 Estimating N demand 281 5.2.6 Statistics 282 5.3. Results 283 5.3.1 Vertical profiles of total chlorophyll a, nitrate and ammonium 283 4.3.2 Profiles of nitrate and ammonium uptake rates in Haida eddies 287 5.3.2 Relative preference and ambient nitrogen concentrations 291 5.3.3 Integrated nitrogen uptake rates and/ratios 292 5.3.4 Carbon: nitrogen uptake ratios 292 Vlll 5.3.5 Nitrate supply and demand 294 5.4. Discussion ". 299 5.4.1 Substrate availability: nitrate and ammonium 299 5.4.2 Nitrogen preference , 299 5.43 Changes in new production rates over eddy evolution 300 5.4.4 Haida eddies and new production 303 5.4.5 Controls on nitrogen uptake rates 303 5.4.6 C:N uptake ratios under different nutrient regimes 306 5.4.7 N-demand: predicting limitation of nitrogen uptake by iron 308 5.5. Summary 308 C H A P T E R 6. A N E X P E R I M E N T A L S I M U L A T I O N O F T H E I N F L U E N C E O F W A T E R M A S S M I X I N G O N P H Y T O P L A N K T O N G R O W T H A T O C E A N F R O N T S 310 6.1 Introduction 310 6.2 Materials and methods 313 6.2.1 Experimental design 313 6.23 Trace metal clean conditions 315 6.2.4 Nutrients 315 6.2.5 Chlorophyll A :.315 6.2.6 Primary productivity and growth rates 316 6.2.9 Statistical analysis 317 6.3 Results 318 6.3.1 Initial conditions 318 6.3.2 Nutrient drawdown 319 6.33. Phytoplankton growth 322 6.3.4 Phytoplankton species composition 327 6.4 Discussion 332 C H A P T E R 7. G E N E R A L DISCUSSION 338 Chronology ofHaida-2000 338 Comparison with other Eastern Boundary Current eddies 344 Transport by Haida eddies and a re-examination of the High Nitrate, Low Chlorophyll condition in the Alaska Gyre 346 A P P E N D I X A . D E T E R M I N A T I O N O F S U B S T R A T E A F F I N I T Y C O N S T A N T F O R SILICIC A C I D IN A N A T U R A L A S S E M B L A G E O F D I A T O M S A T O C E A N S T A T I O N P ( 5 0 ° N , 1 4 5 ° W ) 399 A P P E N D I X B. T H E I M P O R T A N C E O F p H IN D E T E R M I N A T I O N S O F B I O G E N I C SILICA IN S E A W A T E R 403 A P P E N D I X C . C A L I B R A T I O N O F F L U O R E S C E N C E M E A S U R E M E N T S F O R C H L O R O P H Y L L A E S T I M A T I O N 408 A P P E N D I X D. I M A G E S O F S E L E C T E D P L A N K T O N O B S E R V E D WITHIN H A I D A EDDIES A N D T H E I R VICINITY 409 A P P E N D I X E . P R I M A R Y P R O D U C T I V I T Y F R O M S A T E L L I T E S : R E S U L T S F R O M B E H R E N F E L D & F A L K O W S K I (1997) M O D E L W I T H N I T R A T E - D E P E N D E N T P* 0PT 422 I X LIST OF TABLES Table 1 -1. Cruise dates, locations, mixed layer depths (MLD), Sea Surface Height (SSH) anomalies, and eddy dimensions (km) for Haida-2000 sampling. Out refers to reference stations chosen outside but in the vicinity of Haida-2000. Centre refers to the eddy centre as identified by satellite altimetry and hydrographic data, and the edge is defined as waters in the region of the most steeply sloping isopycnals and swiftest currents. SSH data from TOPEX/POSEIDON-ERS-2 satellite radar altimetry processed by Colorado Center for Astrodynamical Research (CCAR) (http://www-ccar.colorado.edu/~realtime/gsfc_global-real-time_ssh/; R. Leben). Dimensions are in latitudinal (N-S) and longitudal (E-W) directions in km and are derived from the SSH plots: Depth of depression of deep isopycnals (m) represents the depth to which the 26.8 and 27.0 isopycnals are depressed relative to surrounding waters 38 Table 1-2. Nitrate and silicic acid concentrations at selected depths for centre, edges, and outside reference stations for all observation sets (uM). Values for northern and southern edges are given for all cruises except February 2000 where edges were located to the east and west of the eddy centre; n.d. = no data available. Analytical precision was ±0.13 uM for nitrate and ±0.76 uM for silicic acid. Based on duplicate chlorophyll a determinations in June 2000 and September 2001, the precision was ±0.02 \\.g L" 1 48 Table 1-3. Change in average nutrient concentrations (uM month"1) within the mixed layer over time at eddy centre between sampling dates. Signs associated with numbers indicate loss (negative) or gain (positive) 52 Table 1 -4. Nutrient inventories (mmol m" ) integrated over 0-50 m in February, June, and September of 2000 and 2001 at the centre of Haida-2000 55 Table 1-5. Estimates of annual and spring new production (NP) and Si(OH)4 drawdown (mmol -9 1 m" d" ) at the eddy centre from 0-50 m in 2000 and 2001. Annual NP was calculated assuming maximum nutrient inventories in February and minima in September, while spring new production captures nitrate-based growth between February and June '55 Table 1 -6. Integrated concentrations of nitrate, phosphate and silicic acid (mol m" ) and Si(OH)4:N03 _ molar ratios at the eddy core and at outside reference stations. Integration depth corresponds to the eddy core, from the depth of flat isopycnals at the top (ca. 75 m) to the 27.0 <5e isopycnal af the bottom. In February 2000 the true centre was not sampled; the numbers represent a station closer to the edge. Numbers in brackets represent the estimated nutrient concentrations at the eddy centre in February 2000 calculated from the enrichment factor (centre value / edge value) from June and September 2000. Analytical precision was 2 2 ±0.065 mol m" for nitrate and ±0.38 mol m" for silicic acid 57 9 i Table 1-7. Estimates of vertical nutrient flux (mmol m" d" ) and density gradients (A sigma-#m" ') across the mixed layer (ML) within Haida-2000 (centre) and at nearby reference sites (out) for all cruises. See text for calculations 74 Table 2-1. List of sampling dates, site locations (edge and centre of Haida eddies, outside reference sites), mixed layer depths (m), and average mixed layer nutrient concentrations (uM) for all sites where phytoplankton taxonomic data were collected. Analytical precision X was ±0.13 p M nitrate, ±0.76 u M silicic acid, and ±0.012 p M phosphate. M L D = mixed layer depth 87 Table 2-2. Surface chl and integrated chl, particulate carbon (PC), nitrogen (PN), biogenic silica (BiSi) and molar ratios for Outside reference, eddy Edge, and eddy Centre sites, 2000-2001. 90 Table 2-3. Abundance and biomass of planktonic ciliates at outside reference (Out), eddy edge (Edge), and eddy centre (Centre) sites in this study. Samples were collected in June and September 2000 and June and September 2001 107 Table 2-4. Correlation matrix for particulates [chl a, particulate nitrogen (PN), particulate carbon (PC), biogenic silica (BiSi)], particulate ratios, and nitrate in the mixed layer (ML) and euphotic zone (eu). Numbers in brackets are the Bonferroni adjusted probabilities. Since silicic acid was strongly positively correlated with nitrate (r = 0.98), it has been omitted. Statistics were performed using SYSTAT v. 10 111 Table 2-5. Estimated grazing impact of plankton ciliates on phytoplankton at the outside reference, eddy edge, and eddy centre sites. Data were averaged across all time periods for each site. Specific ingestion rates were calculated from integrated ciliate biomass and chlorophyll concentrations. Most probably numbers represented the median of values for each category. Percent of average daily primary production consumed calculated from Chapter 4. Average PP derived from all sites on all cruises was 435 mg C m"2 d"1. Most probable number was calculated for the median of all measurements. The range of ratios of herbivorous ciliates to total ciliates is taken from Strom et al. (1993) and the specific ingestion rates are derived from Verity (Verity, 1985, 1991) for an average biomass of 20 ug phytoplankton C L"1) 121 Table 3-1. Accessory pigment biomarkers for different algal classes used in this study 133 Table 3-2. (a) Initial pigment ratio matrix used in this study [from (Mackey et al, 1997)]; (b) final pigment ratio matrix derived from (a), (c) final pigment ratio derived from (a), but with chl a measurements from fluorometry rather than HPLC. (d), (e), and (f) are pigment matrices for 100 and 55% Io, 33 and 10% Io and 3 and 1% Io, respectively, derived using (a) as the initial pigment ratio matrix. A l l pigments are presented as a proportion of chlorophyll a (chl a). Chl C2 is chlorophyll c 2; Chl C3, chlorophyll c 3; PERI, peridinin; BUT, 19'-butanoyloxyfucoxanthin; FUCO, fucoxanthin; H E X , 19'-hexanoyloxyfucoxanthin; VIOL, violaxanthin; D D X , diadinoxanthin; A L L O , alloxanthin; ZEA, zeaxanthin; Chl b, chlorophyll b; (3CA, beta-carotene, and C H L A , chlorophyll a 137 Table 3-3. Proportion of total photosynthetic biomass accounted for by 10 phytoplankton taxa (SYN = Synechococcus spp., E U G L E N = Euglenophyceae, CHLOR = Chlorophyceae, PRAS = Prasinophyceae, DINO = Dinophyceae, CRYPTO = Cryptophyceae, B A C I L L A R = Bacillariophyceae, CHRYSO = Chrysophyceae, HAPTO (= Haptophyceae), and P E L A G O = Pelagophyceae). Contributions to total chlorophyll a were divided into those found within the mixed layer (=ML) and those observed below the mixed layer for Haida-2000 in June and September 2000 and June and September 2001 and for Haida-2001 in June 2001 ...149 XI Table 3-4. List of phytoplankton species observed in samples collected from the Outside, Edge and Centre sites in the vicinity of Haida eddies, 2000-2001. Photographs or line drawings for selected taxa are provided in Appendix D. Presence at the Outside reference sites (*), Edge of Haida-2000 (+), and Centre of Haida-2000 (o) are noted for June 2000, September 2000, June 2001, and September 2001. Presence at the Edge ($) and Centre (#) of Haida-2001 are also noted 157 Table 3-5. List of microzooplankton species identified in samples from the Outside reference, eddy Edge and Centre sites from June and September 2000 and 2001. Photographs of a few microzooplankton species are included in Appendix D 167 Table 3-6. The Morisita-Horn index of similarity for comparisons between Outside reference (1), eddy edge (2), and eddy centre (3) sites in June and September 2000 and June and September 2001. Analyses were performed using Estimates0 statistical software (Colwell, 2004). 174 Table 3-7. Characteristics of diatoms observed in this study, (a) average diatom cell volume . (um3), (b) average silica per cell, and (c) silicic acid: carbon ratios for diatoms at each site during June 2000, September 2000, June 2001, and September 2001. Biogenic silica was determined according to the hot alkaline digestion method (DeMaster, 1991) and diatom carbon was calculated according to carbon conversion factors (Menden-Deuer and Lessard, 2000) derived from microscopic measurements. Averages correspond to the combination of outside reference sites and the centre and edge of Haida-2000. Values obtained for Haida-2001 were omitted from the average calculation in order to better compare changes over eddy evolution and distance from the coast, n.d. indicates that no data were available 186 Table 3-8. Summary of changes in phytoplankton pigments over the 20-month study period. 'Variables' indicate factors that appear to influence the vertical profiles or presence of the various pigments measured in this study 189 Table 4-1. Integrated carbon-uptake rates (± 1 SD, n = 3), euphotic zone depths (Zeu), chlorophyll a concentrations, chlorophyll a-specific carbon uptake (± 1 SD, n = 3 for C-incorporation; for chlorophyll a analytical precision was approximately ±0.01 pg L" 1 or 1 mg m~2), mixed layer depths (MLD), and surface irradiance (I#) for the outside reference, eddy centre and edge sites for four cruises (June and September 2000, June and September 2001) 234 Table 4-2. Brunt-Vaisala frequency (cps) estimates across the mixed layer for Haida-2000 and Haida-2001 for Out, Edge, and Centre stations in spring and summer. For Centre and Edge stations parentheses indicate specific eddy [Haida-2000 (H-00) or Haida-2001 (H-01)] and location of edges (north or south) 235 Table 4-3. Summary of results from two-way analysis of variance on sampling sites and sampling dates. Post hoc pairwise multiple comparison procedure used was the Tukey test, with a set at 0.05. n.s.d. = no significant difference, N / A = not applicable 237 Table 4-4. Integrated carbon-uptake rates (± 1 SD, n = 3), euphotic zone depths (Z e w), chlorophyll a concentrations, chlorophyll a-specific carbon uptake (± 1 SD, n = 3 for C-uptake, n = 1 or n = 2 for chlorophyll a), mixed layer depths (MLD), and surface irradiance X l l (Io) for the outside reference, eddy centre and edge sites for four cruises (June and September 2000, June and September 2001) . 249 Table 4-5. Photosynthetic parameters a and P [mg C (mg chl a)"1 h"1 (mol photons m"2 s"1)"1], and PBs (mg C (mgchl a)"1 h"1) from photosynthetron incubations, and (Xdeck, f W * , and PBopl from deck incubations profiling primary productivity over the euphotic zone (see text for details). Errors represent ±1 standard deviation 253 Table 4-6. Correlation matrix for photosynthetic parameters (a, p\ P^s) derived from 3 h incubations in the photosynthetron 255 Table 5-1. Integrated (100 - \ %\Q) values for chlorophyll a (mg m" ), absolute N uptake (pNx, umol m"2 h"1), N-specific uptake (V^x, umol m"2 h"1 (umol Nx)"1), chlorophyll a-specific N 2 1 1 uptake (p Nx c h i , umol Nx m" h" (mg chl)" ), and the water column averaged/ratio (dimensionless), and the ratio of absolute C to N uptake. Note that in June 2001, Haida-2001 was sampled as well as Haida-2000 284 Table 5-2. Summary table, two-way analysis of variance performed on station and time for f ratios, rates of nitrate uptake, and rates of ammonium uptake for eddy edge, centre, and outside reference sites 290 Table 5-3. Estimated nitrate supply and demand. Supply was calculated by taking the ambient nitrate concentration (mmol m" ) and adding the vertical flux of nitrate into the mixed layer 2 I over 1 day (using diffusion coefficients presented in Chapter 1; mmol m" d" ). Demand was taken as the bulk nitrate uptake (integrated approximately to the mixed layer depth, 50 m; mmol NO3" m"2). The difference between supply and demand is shown and represents the . available nitrate for phytoplankton growth. Negative values indicate potential nitrogen limitation. Centre corresponds to Haida-2000 295 Table 6-1. Summary table showing results from two-way analysis of variance 327 Table B - l . Sampling data, latitude and longitude for Experiment 1. Treatments refer to samples incubated at 95°C at different pH values. The amount of HC1 added to each tube is shown. 404 Table B-2. Data corresponding to Experiment 1. The dependence of biogenic silica concentration on the digestion efficiency of the hot alkaline digestion method at different pH was tested. Vol (ml) before and after digestion includes Na2C03 solution and filter; vol (ml) evaporated is the difference between the volumes before and after digestion; the amount of HC1 added to the test tubes is shown 405 Table E - l . Model variables used to compute PPeu, as described in the Behrenfeld and Falkowski (1997) model of primary productivity deriverd from satellite chlorophyll concentrations. PPreal is the rate of C fixation measured in Chapter 4. The average percent difference was 37.9% 425 Table E-2. Model parameters using nitrate as a predictor of PBopi (PBopt new) to yield a new estimate of PPeu. The average percent difference was 15.5% 426 X l l l LIST OF FIGURES Figure i-1. Schematic of eddy circulation (plan view), a) Clockwise flow around anticyclonic (AC) eddies leads to a Coriolis deflection of the flow field toward the interior of the eddy while the reverse is observed in cyclonic (C) eddies. The Coriolis deflection of circular flow leads to (b) a build up of water at the centre of anticyclonic eddies and a deficit in water at the centre of cyclonic eddies. The former produces an elevated sea surface height relative to the geoid while the latter produces a negative sea surface height anomaly. Implications of these processes are discussed in the text 4 Figure i-2. Schematic showing the side view of an anticyclonic eddy, (a) shows the depression of isopycnals induced by anticyclonic eddy circulation, (b) illustrates the rebounding of isopycnals due to the effects of frictional decay 5 Figure i-3. Drifter tracks for an anticyclonic eddy in the Gulf of Alaska (Haida-2001) eddy showing inertial oscillations. From Yelland and Crawford (in press) 8 Figure i-4. Radar satellite altimetry contours showing the presence of a Haida eddy (about 53°N and 136°W) to the west of the Queen Charlotte Islands on June 18, 2000 (in orange). Height anomalies of ± 5 cm indicate the presence of a mesoscale eddy. Image provided by the Colorado Center for Astrodynamics Research at http://e450.colorado.edu/realtime/gsfc_global-real-time_ssh/ 14 Figure i-5. Schematic diagram showing the influence of isopycnal displacement induced by anticyclonic (left, downwelling) and cyclonic (right, upwelling) mesoscale eddies. In the classical model cyclonic eddies tend to enhance primary production by injecting nutrients into the euphotic zone while anticyclonic eddies do not stimulate phytoplankton growth.. 17 Figure i-6. Map of the Gulf of Alaska showing major currents, the Subarctic Current (SUB), the Alaska Coastal Current (ACC) and the Alaska Stream, which form the cyclonic, clockwise gyre. The dashed line shows the approximate limit of the High Nitrate Low Chlorophyll (HNLC) region. The dotted lines represent the transition zone between coastal and oceanic H N L C waters. A K = Alaska, BC = British Columbia, QCI = Queen Charlotte Islands. Also noted on the map is the region of downwelling to the north of the gulf, along the coast of Alaska, and the region of summer upwelling off the coast of Vancouver Island, British Columbia 23 Figure 1-1. Regional map showing nutrient domains and the location of the centre of Haida-2000 as it traveled northwestward into the Gulf of Alaska. The eddy trajectory followed the line from Feb 2000 to Sept 2001 with each symbol representing the position of the centre at the time of sampling. Reference stations in 2000 were located ca. 30 - 50 km from the southern edge of Haida-2000 (not shown), while in 2001 reference stations were located further away; these are denoted as Jun-R and Sept-R for June and September sampling periods. For reference, time series Line P stations (P4 to OSP) are included. Arrows indicate direction of major currents; A C = Alaska Current, SAC = Subarctic Current. Current/nutrient domains (Favorite et al, 1976; Wong et al., 2002a) are designated A G = Alaska Gyre domain, SUB = Subarctic Current domain, DIL = Dilute domain 32 X I V Figure 1-2. Contours of temperature (a), salinity (b), and sigma-0(c) across Haida-2000. Distance is given in km from south to north and in degrees of latitude except for September 2001, where the plot is a composite of two half-transects with distance given as south to centre and centre to southwest. The contouring intervals are 0.5°C for temperature, 0.1 units for salinity, and 0.2 units for sigma-0. Arrows along the tops of the panels indicate the location of eddy edges delimiting the extent of Haida-2000 (marked 'eddy') 41 Figure 1-3. Solid contours show nitrate (uM) across the Haida-2000 eddy (panels a -d) with dashed contours representing sigma-0. Four panels represent, from left to right, cruises in June 2000 (a), September 2000 (b), June 2001 (c), and September 2001 (d). The contouring interval is 5 uM for nitrate, 0.2 units for sigma-6?. Contour constructed for September • represents a composite of two half-transects from outside to centre (south to centre and centre to southwest) rather than one full transect in a north to south direction. Dots on each panel represent sampling locations and depths 43 Figure 1-4. Three consecutive Sea Surface Height Anomaly plots from TOPEX/POSEIDON-ERS-2 radar altimetry showing the merging of a second, younger eddy with Haida-2000. 45 Figure 1-5. Monthly averages of nitrate and silicic acid concentrations at Ocean Station P (50°N, 145°W) from 1969 to 1981 (adapted from Whitney and Freeland, 1999) (a). Average mixed layer nutrient concentrations in Haida-2000 during Year 1 (b; Feb. 2000 - September 2000) and Year 2 (c) were computed from ambient concentrations from the surface to the bottom of the mixed layer. Eddy formation occurred in February 2000 (= 0 months) and the study ended in September 2001 (= 20 months) 51 Figure 1-6. Solid contours show silicic acid (uM) across the Haida-2000 eddy (panels a -d) with dashed contours representing sigma-6>. Four panels represent, from left to right, cruises in June 2000 (a), September 2000 (b), June 2001 (c), and September 2001 (d). The contouring interval is 5 uM for silicic acid, 0.2 units for sigma-ft Contours constructed for September represent a composite of two half-transects from outside to centre (south to centre and centre to southwest) rather than one full transect in a north to south direction. The minimum value at the surface in June 2000 was 3.0 uM. Dots in each panel represent sampling locations and depths 53 Figure 1-7. Silicic acid (a) and nitrate (b) versus salinity at the eddy centre for February, June, and September 2000 and 2001 at the eddy centre. A coastal upwelling station (P4) and an oceanic station (OSP) sampled in June 2000 are included for reference 58 Figure 1-8. Silicic acid (a) and nitrate (b) profiles for six cruises conducted from February 2000 to September 2001 at the centre of Haida-2000. The surface concentrations of Si(OH)4 increase with age or distance from the point of origin, becoming more similar to surrounding High Nitrate Low Chlorophyll waters. A coastal upwelling station (P4; see Fig. 1-1) and a station located in High Nitrate Low Chlorophyll waters (OSP) are included for reference 59 Figure 2-1. Map of stations where phytoplankton abundance'and species composition were determined in June 2000 (A), September 2000 ( • ) , February 2001 (•), June 2001 (open star), and September 2001 (filled star). BSM= Bowie Seamount, a shallow seamount whose pinnacle reaches -35 m below the surface. The grey circles show approximately the position of Haida-2000. The position of the eddy centre as it drifted westward is indicated X V by the solid line. Two major rivers (Nass and Skeena Rivers) that flow into Hecate Strait are noted on the map 82 Figure 2-2. Temperature and salinity profiles for stations where phytoplankton community composition was examined in February, June, and September 2000 and 2001. Top row of panels come from 2000 while lower panels are observations from 2001. Lateral advection of waters in the mixed layer is inferred from the odd temperature profile at the eddy Centre in February 2001 88 Figure 2-3. Vertical profiles of (a) particulate carbon, (b) particulate nitrogen, and (c) biogenic (= amorphous) silica concentrations (all in umol L"1) for June 2000, September 2000, June 2001, and September 2001. Solid circles (—•—) represent outside reference sites, solid triangles (—A—) are eddy centre (H-2000), and open squares (—•—) are the eddy boundary (edge, H-2000) stations. Open diamonds ( — 0 - - ) represent the centre site of Haida-2001 and solid squares (—•--) correspond to the edge of Haida-2001 : 91 Figure 2-4. Profiles of the ratio between particulate organic carbon and total chlorophyll a (mg: mg) in the water column at the Outside, Edge, and Centre sites in Haida-2000 and Haida-2001 during June 2000, September 2000, June 2001, and September 2001 cruises 93 Figure 2-5. SeaWiFS chlorophyll image showing a region of high chlorophyll a that corresponded to the position of Haida-2000 on April 14, 2000. Image prepared by J . Gower at the Institute of Ocean Sciences, Sidney, B.C. 94 Figure 2-6. Contours of chlorophyll a across Haida-2000 between June and September 2001. (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001. Dotted lines are sigma-6?and filled contours are chlorophyll a concentrations. In (a) the chlorophyll a concentration was estimated from fluorescence measurements from the CTD unit as calibrated against extracted chlorophyll a concentrations for that cruise (see text and Appendix C). Contour intervals are 0.05 for chlorophyll a and 0.1 for sigma-0 96 Figure 2-7. Contours of chlorophyll a (filled contours) and silicic acid (dotted lines) in September 2000 (Haida-2000). Dots represent sampling locations across the Haida-2000 eddy (km) 97 Figure 2-8. Chlorophyll a distributions along horizontal planes from 0, 10, 25, and 50 m depths within Haida-2001 in June 2001. The cylinder represents the diameter of the eddy, with the centre of rotation about the vertical solid line (Centre). Direction of rotation is indicated by the arrow 98 Figure 2-9. Profiles of total chlorophyll a concentration (p_ L"1) in February 2001. Centre H-00 refers to the centre of Haida-2000. Two eddies that had not separated completely from the coast were also sampled (MEI and ME2). These two eddies were located near the southern tip of the Queen Charlotte Islands: M E I = 52.87°N, 132.80°W; ME2 = 51.80°N, 131.00°W). 99 Figure 2-10. Depth integrated chlorophyll a concentration in the 0.2-5, 5- 20, and >20 um size fractions at outside Reference, eddy Edge, and eddy Centre sites during the 20 month evolution of Haida-2000. Measurement errors for chlorophyll a were 0.020 pg L" 1 (1 mg m" 2 1 2 ) for the 0.2 - 5 um size fraction, 0.0054 pg L" (0.27 mg m" ) for the 5 - 2 0 um size fraction and 0.0029 LLg L" 1 (0.15 mg m"2) for the > 20 um size class. The error in total chlorophyll a was approximately 0.012 ug L" 1 (0.61 mg m"2) xvi 100 Figure 2-11. Vertical profiles of size-fractionated chlorophyll a at (a) Reference, (b) eddy Edge, and (c) eddy Centre sites in June and September 2000 and 2001 (in Haida-2000). The widths of the horizontal bars increase with depth as the interval between sampling depths increases 101 Figure 2-12. Relationship between surface chlorophyll a and maximum chlorophyll a for (a) all sites [coastal (•), transition (o), and H N L C ( A ) regions]. Bottom panels show the relationship between surface chl a and maximum chl a for (b) coastal sites , (c) transition zone sites, and (d) oceanic sites in the Gulf of Alaska [data from Haida eddies, G L O B E C program, and Line P program; samples come from winter (February), spring (June), and summer-autumn (August/September/October) between the years 1998 - 2002]. In (c) the dotted line shows the regression of points corresponding to the eddy sites ( T ) , while the solid line represents non-eddy sites [including transition zone (o) and eddy reference sites (•)] 103 Figure 2-13. Relationship between surface chlorophyll a and maximum chlorophyll a within Haida-2000. Samples were derived from cruises in June and September 2000 and June and September 2001 104 Figure 2-14. Ciliate standing stocks at the Outside reference, eddy edge, and eddy centre sites between February 2000 and September 2001. Grey dotted line represents the average of numbers from above and below the mixed layer 106 Figure 2-15. A) From Crawford et al. (in press). SeaWiFS ocean colour image from June 13 2002 overlaid on radar altimetry SSH anomaly contours (interval = 4 cm). Features labeled a and c are small cyclonic and anticyclonic eddies, respectively; j is a jet moving in an offshore direction. A = Haida-2002a, B = Haida-2002b, D = Sitka-2002. Colours have been adjusted for visual clarity and thus do not correspond directly with the SeaWifs colour scale.-B) Sea surface temperature image from the Advanced-Very-High-Resolution-Radiometer (AVHRR) sensor on April 15 2000, showing zones of high temperature (red) and cooler temperatures (blue). Top arrow shows the entrainment of cool water along the periphery of Haida-2000 (H-00), while the lower arrow shows the entrainment of warmer water around the southern boundary of the eddy. Black regions indicate regions under cloud cover, or areas with no data 108 Figure 2-16. Franks (1992) model for biomass accumulation at divergent and convergent fronts, adapted to fit empirical observations within strong (Haida-1998) and weak (Haida-2000) eddies. Grey shaded areas represent phytoplankton biomass with the black region representing an area of increased density; z is depth and x is a horizontal plane 113 Figure 2-17. Integrated chlorophyll a concentrations (mg chl a m"2) estimated from a knowledge of surface chlorophyll a compared to real data 117 Figure 3-1. Integrated inventories of accessory pigments from Outside reference sites, and the Edge and Centre of Haida-2000 in June and September of 2000 and 2001, as well as the Edge and Centre of Haida-2001 in June 2001. Inventories from the edge and centre sites in Haida-2001 are noted as white bars (edge) and solid grey bars (centre) superimposed on the XVII black and shaded bars of the edge and centre sites within Haida-2000, respectively. Note that the bars corresponding to Haida-2001 are shown referenced to months since February 2001 rather than 2000 in order to show pigment concentrations in the two eddies at the same stage in evolution 141 Figure 3-2. Depth profiles of chlorophyll accessory pigments (a) chlorophyll b, (b) chlorophyll c\+C2, and (c) chlorophyll CT, at Outside references sites (-•-) in June and September 2000 and 2001, Edges (-•-) and Centres of Haida-2000 (- A-) in June and September 2000 and 2001, and the Edge (-•-) and Centre ( - A - ) of Haida-2001 in June 2001 143 Figure 3-3. (Previous page) Depth profiles of accessory pigments (a) fucoxanthin, (b) alloxanthin, (c) zeaxanthin, (d) diadinoxanthin, and (e) violaxanthin at Outside references sites (-•-) in June and September 2000 and 2001, Edges (-•-) and Centres of Haida-2000 (-A-) in June and September 2000 and 2001, and the Edge (-•-) and Centre ( - A - ) of Haida-2001 in June 2001 145 Figure 3-4. Depth profiles of chlorophyll accessory pigments (a) (3-carotene, (b) peridinin, (c) 19'hexanoyloxyfucoxanthin, and (d) 19'butanoyloxyfucoxanthin at Outside references sites (-•-) in June and September 2000 and 2001, Edges (-•-) and Centres of Haida-2000 (-A-) in June and September 2000 and 2001, and the Edge (-•-) and Centre ( - A - ) of Haida-2001 in June 2001 146 Figure 3-5. Profiles of chlorophyll a (from Chapter 2) and phaeopigments through the water column at (a) Outside reference sites, (b) the Edge of Haida-2000 and (c) the Centre of Haida-2000 in June and September 2000 and June and September 2001. Bar widths reflect the distance between sampling depths, with the bars centered on each sampling depth.... 148 Figure 3-6. Chlorophyll a (ug L"1) concentration determined by HPLC pigment analysis versus acetone extraction followed by fluorometry 151 Figure 3-7. Relationships between class contributions to total chlorophyll a biomass when fluorometry was used to measure chlorophyll a (New proportion of total chlorophyll a) versus HPLC-derived chlorophyll a values (Original proportion of total chlorophyll a). Dashed lines represent 95% confidence limits 153 Figure 3-8. Water-column averaged proportional contributions of 10 phytoplankton groups to total photosynthetic biomass as determined by Chemtax0 matrix factorization 154 Figure 3-9. Total cell abundance (excluding Synechococcus spp.) above the mixed layer (filled symbols) and below the mixed layer (open symbols) for eddy Outside (a), Edge (b), and Centre (c) sites for Haida-2000 over the 20-month study period beginning in February 2000. : . -. 156 Figure 3-10. Depth-averaged proportion of (a) total cell abundance and (b) total cell carbon determined microscopically for outside reference, eddy edge, and eddy centre sites for Haida-2000 ...170 Figure 3-11. Dominance-diversity curves for (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001 at Outside, eddy Edge and eddy Centre sites for Haida-2000 172 XV111 Figure 3-12. Dominance-diversity curves at the centre of Haida-2000 (-o-) and at four outside reference sites in the eastern Gulf of Alaska (-•-) 173 Figure 3-13. Dominance-diversity curves at the centre of Haida-2000 in June 2000 (-•-), September 2000 (-A-), June 2001 (-o-), and September 2001 (-A-) 173 Figure 3-14. Dendrogram for hierarchical clustering performed according to sample site. The agglomerative clustering method used was Morisita similarity of the unweighted paired group performed using arithmetic means (UPGMA). Sample numbers correspond to scientific cruise numbers: 2010 = June 2000, 2030 = September 2000, 2108 = June 2001, and 2131 = September 2001. Out, Edge, and Centre correspond to sample site within or outside Haida-2000 (designated 00) or Haida-2001 (designated 01) 176 Figure 3-15. Abundance of 10 algal classes (Synechococcus spp., Euglenophyceae, Prasinophyceae, Dinophyceae, Pelagophyceae, Cryptophyceae, Bacillariophyceae, Chrysophyceae, Haptophyceae, and Chlorophyceae) in water above the mixed layer (grey area) and below the mixed layer (black area) at Outside Reference sites and stations at the eddy Edge and Centre between June 2000 and September 2001 178 Figure 3-16. Proportion of total cell abundance or cell carbon determined by light microscopy at Outside reference site, eddy Edge, and eddy Centre for (a) above mixed layer and (b) below mixed layer depths. Top panel in each set are proportions of total cell abundance while bottom panels are proportions of total cell carbon. Categories include diatoms (Bacillariophyceae = B A C I L L A R ) , COCCO (= coccolithOphores, Haptophyceae), ciliates (heterotrophic microzooplankton, CIL), CRYPTO (= cryptomonads, Cryptophyceae), DINO (= dinoflagellates, Dinophyceae), nanoflagellates (= NFL), and OTHER 180 Figure 3-17. Case scores plotted from Principal Components Analysis of algal classes and sampling sites. Those stations grouping closest to a given algal class were dominated by organisms in that class. Classes that explained most of the variability in the data set included Haptophytes (HAPTO), Cryptophytes (CRYPTO), Prasinophytes (PRAS) and Synechococcus spp. (SYN). Together these groups explained approximately 45% of the variability in the data set 181 Figure 3-18. Canonical Correspondence Analysis triplot (joint plot) with vector scalings constructed according to algal classes for sampling times and sites 182 Figure 3-19. Triplot from a C C A of phytoplankton in the vicinity of Haida eddies. Environmental variables are represented by the arrows, sample sites by filled triangles, and algal classes by open triangles. The algal classes are listed by the first few letters of the class. These include: Bacillariophyceae (BACILLAR) , Chlorophyceae (CHLORO), Chrysophyceae (CHRYSO), Cryptophyceae (CRYPTO), Dinophyceae (DINO), Euglenophyceae (EUGLENO), Haptophyceae (HAPTO), Pelagophyceae (PELAGO), Prasinophyceae (PRAS), and Synechococcus spp. (SYN) 184 Figure 3-20. Comparison of the proportion of total chlorophyll a by chemotaxonomy with the proportion of total carbon as measured by microscopy for six algal classes discussed in this study. The line represents a 1:1 relationship.. 198 xix Figure 3-21. Comparison of chemotaxonomic (•) and microscopic techniques (o) in determining the relative contribution to total chlorophyll a (chemotaxonomy) or carbon (microscopy) in (a) diatoms and (b) haptophytes (= prymnesiophytes). Shaded plots indicate samples collected below the mixed layer 199 Figure 4-1. Idealized photosynthesis versus irradiance curve with a, p\ P8^, and E* noted. Adapted from relationship derived by Piatt et al. (1980) 216 Figure 4-2. Map showing the position of the centre of Haida-2000 (H-00) and Haida-2001 (H-01) as they drifted westward from the coast of British Columbia into the Alaska Gyre. A K = Alaska, BC = British Columbia, A G = Alaska Gyre, SUB = Subarctic Current System, DIL = Dilute Domain. Ocean Station P (OSP) and stations along Line P are shown for reference and open arrows (AC = Alaska Current, SAC = Subarctic Current) indicate the direction of the prevailing surface currents 224 Figure 4-3. Schematic diagram showing the photosynthetron used in this study. Neutral density screens were used to reduce light intensity emitted from fluorescent light bulbs. Light was shone through a glass housing filled with copper sulfate (Q1SO4) in order to simulate blue light. Water flowing through the system was kept at ±2°C of ambient seawater at ca. 10 m depth 230 Figure 4-4. Total primary productivity at Outside, eddy Edge, and eddy Centre sites in Haida-2000 (June and Sept 2000, June and Sept 2001) and Haida-2001 (June 2001 only), (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001. Error bars represent ± 1 standard deviation for n= 3 replicates. In June 2001 two eddies were sampled, Haida-2000 and Haida-2001; these are denoted as Edge and Centre H-00 for Haida-2000 and Edge and Centre-01 for Haida-2001 •. 238 Figure 4-5. Profiles of total primary productivity divided into organic (black) and inorganic (grey) fractions for (a) Outside reference, (b) eddy edge, and (c) eddy centre sites in June 2000, September 2000, June 2001, and September 2001. Inorganic incorporation signifies calcification , 240 Figure 4-6. Depth profiles of size-fractionated primary productivity at Outside reference, eddy Edge, and eddy Centre sites in (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001. Size fractions were 0.2-5 um, 5-20 um, and > 20 pm. Only sites within Haida-2000 are shown here. 242 Figure 4-7. Integrated size-fractionated rates of primary production in June 2000, September 2000, June 2001, and September 2001 at the edge and centre of Haida-2000 (all sampling times) and at the edge and centre of Haida-2001 in June and September 2001. Numbers in brackets refer to the different eddies, i.e. 00 is Haida-2000 formed in February 2000 while 01 denotes Haida-2001, an eddy formed in February 2001 245 Figure 4-8. Depth profiles of chlorophyll a-specific primary productivity at the Outside reference (-•-), eddy Edge (-•-) and eddy Centre (-•-) sites for three size fractions: top panel, 0.2 - 5 um, middle panel, 5 - 2 0 um, and lower panel, > 20 pm. Error bars show ± 1 standard deviation, n = 3. Only rates determined within Haida-2000 are shown here 247 X X Figure 4-9. Water column chlorophyll-specific photosynthetic cross-section estimates for phytoplankton at the Outside (-•-), eddy Edge (-•-), and eddy Centre (-•-) sites over time since eddy formation for Haida-2000 248 Figure 4-10. Photosynthesis versus irradiance curves for phytoplankton collected from 55% (•), 10% ( A ) , and 1% (o) of surface irradiance and incubated in the photosynthetron for 3 h. Cells were collected from Outside reference (left panels), eddy Edge (middle panels), and eddy Centre (right panels) in June 2000 (top), September 2000 (second), June 2001 (third), and September 2001 (bottom). In June 2001 two eddies were sampled (Haida-2000 and Haida-2001); for this cruise the eddy sampled is noted in each plot 252 Figure 4-11. Changes in a, P, and PBs over time since eddy formation in February 2000 at the Outside reference (-•-), eddy Edge (-•-), and eddy Centre (-•-) sites for Haida-2000. A l l values shown have been averaged over depth. 256 Figure 4-12. Estimated primary productivity from chlorophyll according to the Behrenfeld and Falkowski (1997) model (-•-) versus actual data (symbols joined by dotted line). When nitrate is added as a predictor of T^opt, the model better estimates actual primary production (-•-). Top panel shows results for all data pooled, while bottom panels discriminate between the outside reference, eddy Edge, and eddy Centre sites. In all cases, the BF'97 model overestimates primary production from a combination of light, sea surface temperature, and chlorophyll a input variables 258 Figure 4-13. Size-frequency histogram showing the proportion of total primay production . accomplished by cells > 5 pm at the Outside reference, eddy Edge, and eddy Centre sites over the course of the 20-month study for Haida-2000 264 Figure 4-14. PBs (2000, A ; 2001, A ) and i^opt ( • ) versus nitrate concentration. Dashed line shows limits of 95% confidence interval 268 Figure 4-15. Dissolved nitrate (NO3", pmol kg"1) and dissolved iron (Fe, nmol kg"1, Johnson et al., in press) determined at the centre of Haida-2000 during four cruises between June 2000 and September 2001. Dissolved iron concentrations were determined on filtered water buffered at pH 3.1 for 24 h as described in Johnson et al. (in press) 269 Figure 5-1. Profiles of chlorophyll a (pg L" 1), nitrate (pmol L"1) and ammonium (pmol L"1) for Outside reference sites (Out), eddy Edges (Edge-00 for Haida-2000 and Edge-01 for Haida-2001), and eddy Centre (Centre H-00 for Haida-2000 and Centre H-01 for Haida-2001) sites determined in (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001 286 Figure 5-2. Vertical profiles of nitrate and ammonium uptake rates. PNO3" and pNH4 +(nmol Nx L" 1 h"1) are the absolute rate of nitrate and ammonium uptake; FNO3" and F N H ^ are the N -specific rates of nitrate and ammonium uptake. Outside reference sites (Out), eddy Edges (Edge-00 for Haida-2000 and Edge-01 for Haida-2001), and eddy Centre (Centre H-00 for Haida-2000 and Centre H-01 for Haida-2001) sites determined in:(a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001 288 Figure 5-3. Vertical profiles of the/ratio in (a) June 2000, (b) September 2000, (c) June 2001, and (d) September 2001 for the Haida-2000 and Haida-2001 eddies. Note that in (a) the x-X X I axis scale ranges from 0.0 to 0.5 instead of 0.0 - 1.0 because/ratios in June 2000 were much lower than for the other sampling times 291 Figure 5-4./ratio versus nitrate concentration (umol L~') for all data pooled. Closed circles represent uptake rates derived from near-surface samples in the mixed layer (55 and 10% IQ) while open circles represent a depth corresponding to 1% IQ. The dashed line is derived from a hyperbolic fit using the Piatt equation 291 Figure 5-5. Relative Preference Index (R.P.I.) for uptake of nitrate and ammonium. R.P.I.N03- are closed circles (•) are R.P.I.NO 3 - , open circles (o) are R.P.I.NH4+- Preference is indicated by points lying above the R.P.I. = 1 reference line, while discrimination is indicated by points lying below this line 292 Figure 5-6. Vertical profiles of the uptake ratios of carbon and nitrogen (p C: p N ) in June 2000, September 2000; June 2001, and September 2001 for the Haida-2000 and Haida-2001 (June 2001 only) eddies. ; 293 Figure 5-7. A) Integrated nitrate uptake versus nitrate supply as determined in Chapter 1 (see text). B)/ratio for different rates of nitrate supply. Closed circles (•) represent the outside reference sites and open circles (o) are eddy centre stations 296 Figure 5-8. Estimates of S, the dimensionless parameter relating nitrate supply and demand in the ocean (Piatt et al., 2003) for varying model parameters. In (a), the excursion of the mixed layer depth with storm activity is set constant at 10 m in spring and 5 m in summer, reflecting increased stratification in summer. T, the time interval between storms, was also held constant at 30 d. C: chl ratios were either 90 (Piatt et al., 2003) or 50 (Pena, pers. comm.). In (b), the C: chl was held constant at 90, 7"was either 30 or 10 d and the excursion of the M L D was held constant at 10 and 5 m, as in (a). See text for details 297 Figure 5-9. Rates of bacterial (= heterotrophic prokaryotes) productivity as determined by uptake of tritiated thymidine in June 2000, September 2000, and June 2001 within Haida-2000 and Haida-2001. Error bars represent ±1 standard deviation for n = 3 samples 298 Figure 5-10. Nitrate uptake (pNCV, nmol L"1 h"1) versus ammonium concentration ( N H 4 + , umol L - 1 ) in Years 1 (•) and 2 (o) in Haida-2000 300 Figure 5-11. Deviation from Redfield ratio (6.6 C:N) in uptake ratios with varying Fe concentrations. Fe data are from Johnson et al. (in press). Symbols represent deviations from Redfield ratios for carbon and nitrogen (6.6 mol C: mol N) while the lines show linear regressions of the data divided into those points above the mixed layer and below the mixed layer. The solid line represents a linear regression for samples collected at 55 and 10% I& and the dashed line shows the regression for samples collected at 10 and 1% lo ; 307 Figure 6-1. Map showing locations where samples were collected. The outside reference site is labeled Out, and the position of the centre of Haida-2001 is marked as Centre. The other stations surrounding the Centre were sampled during a survey of Haida-2001, but are not discussed here 314 Figure 6-2. Nutrient drawdown during experiment (Days 0 - 9); (a) nitrate, (b), phosphate, (c) silicic acid, and (d) ammonium. Error bars represent one standard deviation (n = 5). A = X X I I Treatment A (eddy water), B = Treatment B (mixed water), and C = Treatment C (outside reference water) 320 Figure 6-3. Relationships between major phytoplankton nutrients in the experimental containers. Top panel shows silicic acid (uM) concentration versus nitrate (pM) for Treatment A (eddy water), Treatment B (mixed water), and Treatment C (outside reference water). Bottom panel shows nitrate versus phosphate (pM) 321 Figure 6-4. Changes in primary productivity (mg Cm" d") and chlorophyll a-specific primary productivity (mg C (mg chl. a) h"1). A , B, C represent Treatments A (eddy water), B (mixed water), and C (outside reference water), respectively. Error bars show ± 1 standard deviation, n = 5 323 Figure 6-5. Chlorophyll a concentrations (pg L"1) for three treatments (A, eddy water alone, B, eddy + outside water, and C, outside water alone) over the course of the experiment. Error bars represent ± 1 standard deviation, n = 5 324 Figure 6-6. Changes in particulate organic carbon and nitrogen over the experiment. Treatment A = eddy water only, Treatment B — mixed water, and Treatment C = outside water only. 325 Figure 6-7. Growth rates of particulate carbon, nitrogen, and chlorophyll a over the course of the experiment. Treatment A = eddy water, Treatment B = mixed water, and Treatment C = . outside reference water 326 Figure 6-8. Concentrations of major accessory algal pigments found within the experimental containers. A = Treatment A (eddy water), B — Treatment B (mixed water), and C) Treatment C (outside reference water). 19'But = 19'butanoyloxyfucoxanthin, 19'Hex = 19'hexanoyloxyfucoxanthin 329 Figure 6-9. Proportion of the contribution by diatoms, prymnesiophytes, pelagophytes, prasinophytes, cryptophytes, and dinoflagellates to total chlorophyll a over the course of the experiment. A) Treatment A (eddy water), b) Treatment B (mixed water), and c) Treatment C (outside reference water). Proportions were obtained by analysis of phytoplankton accessory pigments using the Chemtax® matrix factorization program 330 Figure 6-10. Contribution by large and small diatoms to total cell abundance (left) or total carbon (right) over the course of the experiment. Small diatoms (< 10 pm) were mainly represented by Nitzschia cylindroformis and Thalassiosira spp., while large diatoms were a mix of Pseudo-nitzschia spp., Cylindrotheca closterium, Plagiotropis lepidoptera, and larger Thalassiosira spp 331 Figure 6-11. Abundance of dominant phytoplankton groups on Day 7 of the experiment. A = eddy treatment, B = mixed water treatments, C = outside reference water treatment 332 Figure 6-12. False-colour ocean colour (SeaWiFS) image enhanced to show patterns of swirling chlorophyll in the vicinity of Haida eddies and other meanders of the coastal currents (image available at http://earthobservatory.nasa.gov/ Newsroom/Newlmages/ images.php3?img%20id=9310) : 336 xxm Figure 7-1. Timeline of events occurring during the sampling period in this study (February 2000 -September 2001) ; 339 Figure A - l . Uptake rate of silicic acid (umol h"1) versus substrate concentration (uM) 401 Figure A-2. Lineweaver-Burke plot of transformed uptake rates versus substrate concentration (see previous figure). S = silicic acid concentration, uM; v = uptake rates (nmol h"1) 401 Figure B - l . Concentrations of dissolved silicic acid following hot alkaline digestion of biogenic silica as a function of pH. A) Experiment 1, showing the yield of silicic acid from biogenic silica from samples with pH held within a narrow range; b) shows the effect of varying pH on the same sample. The optimal pH range is found between the dashed line (ca. 2.5-6.5 pH units) 406 Figure C - l . Regressions of fluorescence versus extracted chlorophyll a. (a) Fluorescence versus chl a, (b) chl a estimated from the regression versus chl a, (c) estimated chl a versus extracted chl a for shallow depths, and (d) estimated chl a versus extracted chl a for deep depths 408 X X I V LIST OF P L A T E S Plate 1. Light micrographs of coccolithophores observed in this study, a), Coccolithus pelagicus, b) Syracosphaera prolongata sensu Throndsen c) Michaelsarsia elegans, d) Rhabdosphaera clavigera, e) Calciosolenia murrayi, f) unknown coccolithophore 409 Plate 2. Light micrographs of the diatom Aulacoseira granulata. In (a), the characteristic constrictions in the middle of the valve are shown; (b) shows the structure of the girdle bands, and (c) illustrates the spines that extend from the valve face. This freshwater diatom was observed in greatest abundance in June 2000 at the eddy edge and at the outside reference site. It was also observed at the edge of Haida-2000 in June 2001 410 Plate 3. Flagellates (Prasinophytes and dinoflagellates). A , c) Pterosperma polygonum, b) Cladopyxis sp., d) Gymnodinium sp. A (after Buchanan, 1966), e) Oxyphysis oxytoxoides, e) cf. Trochiscia multispinosa 411 Plate 4. Fragilaria capucina var. capucina (a) shows rows of poroids, (b) shows the labiate process near the valve margin, (c) and (d) show full specimen view. The valves lack a raphe and possess a central region devoid of pores 411 Plate 5. Calcareous dinoflagellates, coccolithophores and a diatom with epiphytes. A) Another view of Thoracosphaera heimii; b, c) Syracosphaera sp., d, e) Syracosphaera prolongata sensu Throndsen, f) Dactyliosolen mediterraneus with Rhizomonas setigera epiphytes... 414 Plate 6. a) Psammodiscus nitidus, b) Thalassiosira oestrupii, c) Asteromphalus hyalinus, d) Dactyliosolen mediterraneus with Fragilaria sp. in background 414 Plate 7. Light micrographs of pennate and centric diatoms collected from net tows in September 2001. a, b) Pseudo-nitzschia turgidula; arrow shows rows of poroids, c) Thalassiosira proschkina, d) Coscinodiscus marginatus 415 Plate 8. Light micrographs of diatoms collected from the outside reference site in September 2001. a) Thalassiosira delicatula, b) Coscinodiscus marginatus, c) Thalassiosira sp., d) Pseudo-nitzschia turgidula ....417 Plate 9. Siliceous phytoplankton collected from the outside reference site in September 2001. a) Part of Thalassiothrix longissima cell, b) view of girdle bands of diatoms collected in a net tow in September 2001 at the outside reference site; centric diatoms were plentiful at this time, c) Pseudo-nitzschia granii, d) Thalassiosira sp 418 Plate 10. Nitzschia cylindroformis under the scanning electron microscope, collected from the centre of Haida-2000 in September 2001 419 Plate 11. Scanning electron micrographs of siliceous phytoplankton. a) Actinoptychus spendens, b) Fragilaria cylindroformis c) Neodenticula seminae, d) Dictyocha speculum 420 Plate 12. (a) Dictyocystis elegans, (b) nematode worm, (c) Strombidium conicum (large), (d) Strombidium conicum small 421 LIST OF ABBREVIATIONS X X V Physical and geographical abbreviations AG Alaska Gyre ACC Alaska Coastal Current EBC Eastern boundary current WBC Western boundary current GOA Gulf of Alaska SUB Subarctic Current System DIL Dilute Domain MLD Mixed layer depth WCVI West Coast of Vancouver Island QCI Queen Charlotte Islands SSH Sea surface height SeaWiFS Sea-viewing-Wide-Field-of-view Sensor AVHRR Advanced Very High Resolution Radiometer HNLC High Nitrate, Low Chlorophyll Chemical abbreviations BiSi Biogenic (= amorphous) silica (SiO2*(H20)x) PC Particulate (organic) carbon PN Particulate (organic) nitrogen LDPE Low Density Polyethylene HDPE High Density Polyethylene N 0 3 " Nitrate Si(OH)4 Silicic acid HPO4 2 " Phosphate Chl Chlorophyll v/v Volume per volume DAPI 4'-6-diamidino-phenylindole Biological abbreviations and symbols PP Primary productivity P» Biomass-specific primary productivity pB 1 max Maximum observed biomass-specific primary productivity r3* Saturating chlorophyll-specific primary productivity; when photoinhibition present, it is greater pB opt than 7 ^ m a x Maximum in situ biomass-specific primary productivity that accounts for daily irradiance variations and water column vertical mixing 7 Depth of euphotic zone NPP New primary production BP Bacterial (= heterotrophic prokaryotic) production FCM Flow cytometry TdR Tritiated thymidine CMH Morisita-Horn index of similarity a initial slope of the Photosynthesis-Irradiance curve; also, significance level of statistical tests P slope of Photosynthesis-Irradiance curve that corresponds to photoinhibition at high irradiances X Carbon: chlorophyll a s Nitrate demand V Ratio of nitrogen to chlorophyll a in phytoplankton N Brunt-Vaisala frequency (cps) O, Sigma t (density) 09 Sigma theta (potential density) Io Surface irradiance X X V I P R E F A C E The motivation for this work came from observations of unusual water properties during a routine cruise from coastal British Columbia out to the centre of the Alaska Gyre in August 1998 aboard the C.C.G.S. John P. Tully. The position of anomalous water properties noted by Frank Whitney at the Institute of Ocean Science were matched with satellite derived sea surface altimetry by Bi l l Crawford, and indicated the presence of a mesoscale eddy sitting along the ship's transect. These "new" eddies, called Haida, had not previously been tracked though the ocean, and little was known of their physics, chemistry, or biology. Prior to the widespread availability of satellite altimetry images that employ cloud-penetrating microwaves (produced by the Colorado Center for Astrodynamics Research), the identification of these mesoscale features in the often overcast northeast subarctic Pacific relied on shipboard observations that could identify eddy-like features (e.g. Tully et al., 1960), but not map their trajectories or confirm their motility (e.g. Tabata, 1982). The large geographical extent of the area where eddies are found and the relative infrequency with which scientific cruises are undertaken significantly reduce the likelihood of observing eddies in the field. With new satellite techniques the frequency of observations is no longer a problem, and the research on relating mesoscale features with oceanographic processes has since burgeoned. The Alaska Gyre is characterized by a productive coastal margin and a High Nitrate, Low Chlorophyll centre. Eddies spawned from the coast span both environments, as they drift westward away from the coast. They are thus a source of coastal waters, most important for the contribution of trace metals, especially iron, an important micronutrient for phytoplankton growth. There were two main goals of this thesis. First, the distributions of macronutrients and rates of primary production within the eddy waters needed to be determined; understanding how the input of trace metals influenced these distributions and the productivity of waters within these eddies was part of this goal. It was hypothesized that eddy circulation would modify photosynthetic distributions by promoting higher growth within waters of the eddy compared to surroundings due to the input of coastal nutrients. The second goal was to study the transformations that occur as a long-lived eddy evolved over a two-year period. Haida eddies travel far into the Alaska Gyre, decaying slowly and collapsing after more than three years, in some cases. Because these eddies remain more or less intact over this period, they provide an interesting natural laboratory in which to study changes in ecosystem dynamics. Project work was accomplished through the collaborative efforts of scientists at the Institute of Ocean Sciences, University of Victoria, Pacific Biological Station, and University of. British Columbia. Dave Mackas, Doug Yelland, Moira Galbraith, Maia Tsurumi, Sonia Batten (all of IOS), John Dower (U Vic), and Ian Perry (Pacific Biological Station) undertook the zooplankton work; Frank Whitney (IOS) analyzed nutrients; B i l l Crawford, Mike Foreman, Josef Chiernawsky, Marie Robert, Doug Yelland (IOS) modeled eddy formation and described eddy physics, Lisa Miller, Melissa Chierici (IOS) described dynamics of the carbon system; Keith Johnson, Nes Sutherland, C.S. Wong (IOS) , Sabrina Crispo, Kristin Orians (UBC), and Maeve Lohan (University of Southampton) studied the trace metal distributions, and under the guidance of Paul Harrison, I investigated phytoplankton processes, including photosynthetic biomass, primary production, photosynthetic efficiency, nitrogenous nutrition, and species composition. Together we hoped to create a multidisciplinary and holistic view of the function of Haida eddies, charting the initial steps in understanding mesoscale variability in the northeast subarctic Pacific. XXV11 Publications arising from this thesis: > Peterson, T.D., Whitney, F.A., Harrison, P.J. (in press). Macronutrient dynamics within an anticyclonic mesoscale eddy in the Gulf of Alaska. Deep-Sea Research II. (Chapter 1) Peterson, T.D., Whitney, F.A., Harrison, P.J. (in review). Changes in primary production during westward propagation of anticyclonic mesoscale eddies in the northeast subarctic Pacific. Submitted to Marine Ecology Progress Series. (Chapter 4) Crawford, W.R., Brickley, P.J., Peterson, T.D., Thomas, A .C . (in press). Biological impact of Haida eddies on the Gulf of Alaska. Deep-Sea Research II. (parts of Chapter 2) Plate 6c appears in Sculze-Makuch, D., Irwin, L . N . (2004). Life in the Universe - Expectations and Constraints. Berlin: Springer-Verlag, ISBN 3-540-20627-2. The first two of these papers were written by me with revisional inputs from co-authors. Frank Whitney of the Institute of Ocean Sciences was instrumental in the realization of the Haida eddies project and contributed intellectually to the scientific output of most scientists involved in the research, including the interpretation of nutrient data. The third paper was conceived by Bi l l Crawford and Peter Brickley; my data was added to ground truth satellite observations. Some of the figures presented in Chapter 2 have been modified from those published in this manuscript. My intellectual contribution was in the interpretation of in situ chlorophyll a concentrations and the observed seasonal patterns in phytoplankton growth. Other relevant work includes the following: Nefncek, N . , Peterson, T.D., and Tortell, P.D. (in review). Particulate dimethylsulfoniopropionate (DMSPp) in the NE Pacific by Membrane Inlet Mass Spectrometry. Submitted to Limnology and Oceanography. Crawford, D.W., Lispen, M.S., Lohan, M.C. , Purdie, D.A., Statham, P.J., Whitney, F.A., Putland, J.N., Johnson, W.K., Sutherland, N , Peterson, T.D., Harrison, P.J., and Wong, C.S.. (2003). Influence of zinc and iron enrichments on phytoplankton growth in the North eastern Subarctic Pacific. Limnology and Oceanography: 48: 1583-1600. A C K N O W L E D G E M E N T S X X V l l l A big thesis merits a comprehensive list of acknowledgements. Many people deserve credit for their input in the production of this thesis and in my academic development during the course of my tenure at U B C . I would like to thank my supervisor, Dr. Paul J. Harrison for providing the initial impetus for this study, for his clarity of perspective, and for his willingness to build bridges and foster collaborations amongst scientists from many disciplines and backgrounds. He has been a wonderful role model in his enthusiasm for science and life in general. I feel honoured and humbled to be a part of the Harrison lab lineage. I would like to extend my thanks to my committee members, each of whom contributed in important ways throughout the course of this program, in addition to their willingness and to participate in the various meetings and examinations. Bi l l Crawford was exceptionally helpful in the interpretation of physical data from cruises. He was always willing to share ideas and data. John Dower was extremely helpful, particularly in the early stages of thesis research in statistical consultations, and was always encouraging, particularly during the difficult transitional phase of my degree. Max Taylor's assistance in the taxonomic aspects of this thesis was extremely valuable - his breadth and depth of knowledge with regards to taxonomy is inscrutable. I always looked forward to discussions with him and David Cassis about the oddities of my samples. Some interesting observations would have been missed were it not for his input. Although not officially part of my committee, scientists from the Institute of Ocean Sciences were fundamental to my research accomplishments and in the enjoyment of field work involved in this project. The assistance of Frank Whitney in particular cannot be overstated. Frank's enthusiasm for research in the subarctic North Pacific was infectious, and his support throughout my degree is greatly appreciated. Frank kindly provided data, feedback, and opportunities whenever possible, and I am very grateful to him. Lisa Miller, Dave Mackas, and Marie Robert were excellent chief scientists at sea and contributed significantly to the advancement of my understanding of eddy science. Wendy Richardson and Janet-Barwell-Clarke kindly analyzed nutrient samples and, along with Shannon Harris, taught me the fundamentals of the Auto-Analyzer. Keith Johnson and Nes Sutherland kindlygenerously provided iron data -their work ethic is was inspirational. Beth Bornhold was extremely helpful in the early stages of field work, and Jennifer Putland (IOS), Rana El-Sabaawi (UBC, UVic), Lisa Nodwell (McGill), Adrian Marchetti (UBC), Michael Lipsen (UBC), Rachel Harrison (UBC), Maia Tsurumi (UVic), Dave Crawford (SOC), Sheila Toews (IOS), and Doug Anderson (IOS) all provided excellent assistance in the field. The assistance of Hugh McLean was highly valued. Angelica Pena was extremely generous in her willingness to lend all manner of field equipment and isotope stocks. Her support throughout my degree is also much appreciated. In the laboratory, the pedagogy of Michael Lipsen, Rana El-Sabaawi, and Joe Needoba was much appreciated. Behzad Imanian's assistance in counting bacteria was welcome, as was his sense of humour. Maureen Soon was always helpful in granting advice and giving instructions. Maureen kindly ran the CHN samples from Chapter 6. Dave Crawford's assistance with the HPLC samples, words of advice, and music appreciation are warmly acknowledged. I wish to thank member of the Harrison lab for their input and friendship - Rana E l -Sabaawi, Robert Strzepek, Shannon Harris, Mike Henry, Michael Lipsen, Adrian Marchetti, Nelson Sherry, Philippe Juneau, Heather Toews, Joe Needoba, and Julie Granger. I have benefited from discussions with all of them. I would like to extend my thanks to Philippe Tortell and Maite Maldonado and their lab members for inviting those of us remaining in the Harrison lab to participate in their lab meetings, use their facilities, and for providing a positive and enthusiastic environment in which to study. Their generosity is much appreciated. I would also like to thank Claudio DiBacco for generously providing lab space for me as I completed the last xxix stages of lab work, and members of the Suttle and Orians labs for providing me with desk space and making me feel at home among them. I would like to thank the support staff in the Department of Earth and Ocean Sciences for their efficiency and friendliness. Carol Leven was always helpful, and a colourful character in the halls of the Biosciences building. Alex Allen was extremely helpful as a resource and always ensured the proper forms were filled out and all deadlines met. I acknowledge funding from the National Sciences and Engineering Research Council of Canada (NSERC) for post graduate fellowships (PGS-A, PGS-B, Department of Fisheries and Oceans supplement), the University of British Columbia for Jean R. MacDonald and University Graduate Fellowships, the Department of Earth and Ocean Sciences for George L. Pickard and RuthRobert Rutherford Rae scholarships, and the Department of Earth and Ocean Sciences for various teaching assistantships and a part-time sessional lecturer position. Drs. Irena Kaczmarska and Quay Dortch served as excellent mentors and encoured me to pursue graduate studies. I would like to express my gratitude to my friends for their support and friendship throughout my tenure at U B C - 1 have many fond memories of lunch times, hockey games, redneck weekends and open mic nights. I wish to thank Joe Needoba for his unconditional belief in my abilities and for always pushing me to do better while at the same time encouraging me to relax. I owe my deepest gratitude to my family for their unwavering support in all aspects of my life. G E N E R A L INTRODUCTION 1 Mesoscale eddies in the ocean Turbulent motions are the ocean's internal weather. Flow variability in the ocean exists on many spatial and temporal scales, from long-period Rossby and Kelvin waves to millimeter scale disturbances. On spatial scales of tens to hundreds of kilometers and timescales of weeks to months, swirling motions and meandering of currents and filaments, semi-attached and isolated rings, advective deep vortices, lens vortices, planetary waves and topographic waves are known collectively as "eddies" (Robinson, 1983). Mesoscale (10-100 km) circulation dominates the flow field in the ocean (Robinson, 1983), and mesoscale eddies are responsible for distributing most of the turbulent kinetic energy therein (Wyrtki et al, 1976). These ubiquitous ocean features possess excess energy compared to background, and are characterized by rotational speeds greater than the long-term average flow (Robinson and Leslie, 1985). The first indications of mesoscale eddy activity came from long-term current measurements made by scientists of the former Union of Soviet Socialist Republics (U.S.S.R.), in particular by V . B . Shtokman who made the first observations of mesoscale current anomalies in the Caspian Sea (Shtokman and Ivanovskii, 1937; in (Kamenkovich et al, 1986). Further studies in the Black Sea (Ozmidov, 1962; in Kamenkovich et al., 1986), the North Atlantic (Ozmidov and Yampol'skii, 1965; in Kamenkovich et al, 1986), and the Arabian Sea (Shtockman et al, 1969); in Kamenkovich et al, 1986) confirmed the presence and ubiquity of rotating mesoscale features. The early observations of non-stationary currents and temperature/salinity anomalies paved the way for more detailed synoptic investigations of eddy fields in the open ocean in the 1960's and 1970's [e.g. Polygon-67, Polygon-70, MODE; (Brekhovskikh et al, 1971a; 2 Brekhovskikh et al, 1971b); (Kamenkovich et al, 1986; Koshlyakov, 1986)]. The first large-scale synoptic studies were envisioned by Shtokman (Kamenkovich et al, 1986) and carried out first in the Arabian Sea (Polygon-67) .where 250 km- scale baroclinic disturbances of geostrophic currents were observed. This initial study did not successfully map out the spatial structure of the current fields, but comparisons between predicted flows from dynamical calculations and observed current velocities at several mooring sites were in good agreement (Koshlyakov et al., 1972), inspiring confidence that the features were indeed mesoscale eddies. After Polygon-67, many other surveys of oceanic eddies were conducted in several locations [e.g., west of Portugal, Swallow and Hamon (1960); south-west of Bermuda (Swallow, 1971); south-west of the Hawaiian Islands (Wyrtki, 1967)]. A second major expedition dedicated to the study of mesoscale eddies took place in 1970 in waters at the southern periphery of the North Equatorial Current in the Atlantic Ocean (Polygon-70; (Brekhovskikh et al., 1971a; Brekhovskikh et al, 1971b). The growing awareness of the importance of large-scale turbulence in the ocean introduced the need to characterize these complicated flow fields and to gain an understanding of how turbulence regimes influence physical, chemical, and biological processes. The characterization of turbulence in the ocean remains a significant challenge due to technological, logistical (Gargett, 1982; Matear and Wong, 1997), and theoretical difficulties. The Polygon-67, Polygon-70, and MODE programs indicated that eddy circulation was widespread, and identified critical spatial and temporal scales on which such circulations operated. These first programs were limited by available data. Today, these problems have been largely overcome by the advent of satellite technology and mooring arrays. Nevertheless, the oceans remain undersampled and the development of unifying theories relating turbulent flow to physical phenomena and biogeochemical processes remains an important goal in oceanography. A particularly intriguing 3 set of questions involves the extent to which eddies redistribute heat and biologically available nutrients, and how this has fluctuated over short and long timescales. The role that eddies play in shaping global climate, past and present, is still not well understood [e.g. Gordon (2003); Simmons and Nof (2002); Berloff et al. (1994; 2002)]. Based on satellite images, it appears as though eddy activity is related to climatic events such as El Nino in the northeast subarctic Pacific, with larger and more numerous eddies formed in warmer years (Melsom and Basu, 1999; Mysak, 1985), and mixing induced by frontal meandering in the Kuroshio region appears to exert a strong influence on North Pacific Intermediate Water (NPIW) formation (Joyce et al, 2001). Eddies may thus play a critical role in oceanic climate feedbacks. As such, gaining an understanding of how eddies behave, physically, chemically, and biologically, in different environments is of utmost relevance to studies of regional and global climate. Physical characteristics of mesoscale eddies Types of eddies Eddies can be broadly divided into two types: anticyclonic (clockwise rotating in the Northern Hemisphere) and cyclonic (counter-clockwise in the Northern Hemisphere; Fig. i-1). These two types differ in the patterns of circulation induced by the azimuthal rotation; namely, the vertical component of the rotating flows (Figs, i-1 and i-2). Cyclonic eddies tend to induce upwelling at the centre (Joyce et al, 1981), while anticyclonic eddies experience a downwelling of water from the surface toward deeper depths (Joyce et al, 1984), due to the Coriolis force and pressure gradients that direct circulating water either away from the centre (cyclonic) or toward the centre (anticyclonic; Fig. i-lb). Another classification scheme for mesoscale eddies is the division between eddies generated from strong western boundary currents (WBC) versus weaker eastern boundary 4 currents (EBC). The former (e.g. Gulf Stream, Kuroshio) tend to produce "rings"; that is, isolated waters surrounded by an annulus of the parent current waters. Eddies that are completely Plan view Figure i-1. Schematic of eddy circulation (plan view), a) Clockwise flow around anticyclonic (AC) eddies leads to a Coriolis deflection of the flow field toward the interior of the eddy while the reverse is observed in cyclonic (C) eddies. The Coriolis deflection of circular flow leads to (b) a build up of water at the centre of anticyclonic eddies and a deficit in water at the centre of cyclonic eddies. The former produces an elevated sea surface height relative to the geoid while the latter produces a negative sea surface height anomaly. Implications of these processes are discussed in the text. 5 Figure i-2. Schematic showing the side view of an anticyclonic eddy, (a) shows the depression of isopycnals induced by anticyclonic eddy circulation, (b) illustrates the rebounding of isopycnals due to the effects of frictional decay. 6 isolated from the parent current are know as rings, while partially-separated vortices are called "streamers" or simply, meanders, i f the degree of connectivity with the parent current is very high. In contrast, open ocean eddies or those generated from eastern boundary currents tend to be less isolated, and interact with surrounding waters through lateral advection. Other types of eddies include modons (Flierl et al, 1983), lenses (Killworth, 1983), joint vortices (Nof, 1985), and dipoles (Ahlnas et al, 1987; Simpson, 1990). Modons are pairs of eddies that co-occur side by side, each of which is equal in energy but opposite in sign. Lenses are shallow eddies that can occur throughout the water column (i.e. deep lenses, mid-depth lenses or surface lenses). Joint vortices (see below, under Translation) are eddies that occur in stacks, one above the other, not necessarily of equal strength or sign. Dipoles are eddies that occur together but spin with opposite sign, usually in the wake of a larger eddy or due the island effect when strong flows pass a topographic obstacle. Finally, Taylor columns (Taylor, 1922; Taylor, 1923), generated by the interaction of a strong flow with a large topographic feature such as a seamount or strongly sloping bottom, are anticyclonic vortices that occur on the lee side of the feature (Huppert, 1975). Temporal and spatial scales Most mesoscale eddies operate on Eulerian timescales of months and length scales of typically 100-300 km, thus differentiating them from basin scale features (>1000 km in diameter; Mann and Lazier, 1996). Mesoscale eddies rotate on approximately vertical axes (Iselin, 1936, 1940; Stockman et al, 1969; Simpson, 1983) and extend through the water column, where they exert their influence to depths greater than 1000 m (McCartney and Woodgate-Jones, 1991) and can impart great variability on flows in the deep ocean (Hollister et al, 1984; Kelley and 7 Weatherly, 1985). These rotating masses of water can retain their distinct identity for several months (Ortner et al, 1978; Siegel et al, 1999) or even years (Crawford and Whitney, 1999). Circulation Large scale eddies are quasi-geostrophic in nature, have internal deformation radii (g'H)^2 * [internal Rossby radii, Ro = , whereg' is reduced gravity , His the fluid depth, and/is the Coriolis parameter; (Rossby, 1936)] on the order of 10-20 km, and have rotational speeds on the order of 2-15 cm s'1 (Stavropoulos and Duncan (1974); Crawford and Whitney, 1999), although mid). Mid-ocean eddies generated away from boundary systems, for example in the central North Pacific in the vicinity' of Hawaii, tend to have lower rotational speeds compared to eddies generated from currents (Kirwan et al, 1978). Eddies rotate either clockwise (anticyclonic in the Northern Hemisphere) or counter-clockwise (cyclonic in the Northern Hemisphere). In order to achieve a quasi geostrophic balance, the acceleration of water at the surface is toward the centre of the eddy (anticyclonic eddies; Fig. i-1), which causes a "doming" effect (Fig. i-1) that can be tracked by satellite altimetry as a positive sea surface height anomaly (SSHA) referenced to long-term averaged data. Cyclonic eddies rotate such thare depressed at the centre is depressed; this isand are detected as a negative sea surface height anomalySSHA. Modeling studies reproduce the azimuthal and vertical circulations associated with anticyclonic eddies, while drifter studies (Stavropoulos and Duncan, 1974) or passive tracers (Fratantoni and Glickson, 2002) illustrate eddy-induced patterns in surface circulation. Yelland and Crawford (in press) describe drifter-measured flow in a Haida eddy. The full range of complicated flows has not yet been elucidated, and higher order non-linear flows [e.g. internal g' = g (r2- r/) / r, where r2 is the density of the upper layer and r, is the density of the lower layer; r is the average density 8 waves; Joyce and Stalcup (1984)] are being described based on improvements in instrumentation and sampling opportunities (Martin and Richards, 2001). It is clear that the location, environment, and atmospheric background all contribute to variability in flow patterns. Surface flows around eddies are complicated by inertial oscillations [eg. Fig. i-3; (Vasilenko et al, 1976); (Yelland and Crawford, in press)] and entrainment of waters from coastal boundaries [Fig. i-4; (Lillibridge et al, 1990; Crawford et al, in press)], while azimuthal circulations in deep waters are affected by interactions with topography [e.g. Nof (1988)], or by interactions with water masses or other mesoscale eddies (Joyce et al, 1983; Koblinsky et al, 1984; Kennelly etal, 1985). Drifter 4995, 6 to 30 June 2001 50.8 I . " ••' i • . 1 r — 225.0 225.4 225.8 226.2 East Longitude Figure i-3. Drifter tracks for an anticyclonic eddy in the Gulf of Alaska (Haida-2001) eddy showing inertial oscillations. From Yelland and Crawford (in press). 9 Of the processes that influence eddy circulation, one of the dominant ones is frictional decay (Molinari, 1970; Schmitz and Vastano, 1977; Franks et al, 1986; Okada and Sugimori, 1986; Fukumori, 1992). Models illustrate that eddies are not in perfect geostrophy, but exhibit frictional behaviour, spinning down over time (Flierl and Mead, 1985 and references within). In the best studied eddiesexamples (including the Gulf Stream rings and East Australia Current eddies) eddies interact with the fast-flowing waters of the parent current from which they were spawned, usually before they reach one year old. Models suggest that the frictional decay of anticyclonic eddies (decrease in azimuthal velocity) during eddy spin-down results in slow upwelling at the centre (Flierl and Mied, 1985; Fukumori, 1992) and a predicted chlorophyll a maximum (Franks et al, 1986). With frictional decay, density differences between waters inside the eddy and those outside become smaller over time, facilitating the entrainment of nutrients from below due to wind-mixing (Joyce and Stalcup, 1985; Yentsch and Phinney, 1985). Frictional decay induces secondary circulation patterns that result in the shoaling of isopycnals at the ring centre when the ratio of kinematic viscosity to thermal diffusivity is greater than unity (i.e. a Prandtl number * >1; isopycnal depression occurs when the Prandtl number is < 1), outflow adjacent to the base of the eddy, and downwelling at the ring perimeter [Fig. A l , Flierl and Mead (1985)]. Schmitz and Vastano (1975) modeled the frictional decay of a cyclonic eddy, showing an influx of outside water near the surface and an outflow at depth. This scheme was later modified to include inflow both near the surface and at depths below the thermocline in a cyclonic ring with, and outflow at intermediate depths (Schmitz and Vastano, 1977). In warm core eddies of the Gulf of Mexico these flow directions were reversed, with outflow in surface and deep waters and inflow at intermediate depths (Schmitz and Vastano, 1976). * The Prandtl number describes the ratio of kinematic viscosity (v) to thermal diffusivity (K), (Pr = v / K); when Pr = 1, the viscous time scale is equal to the time scale of thermal diffusion. In the ocean, Pr typically ranges from 7-13 for 20 and 0°C, respectively. 10 In addition to frictional effects, local upwelling at the centre of anticyclonic mesoscale eddies can also be induced by convective overturn of surface waters that are subject to cooling. For example, warm core rings of the Gulf Stream (typically 18°C) lose heat during propagation from the open ocean toward the coast where temperatures are lower (Yentsch and Phinney, 1985) and higher vertical diffusivity values within eddies lead to deeper winter mixing compared to surroundings (Joyce and Stalcup, 1985). Stratification by summer heating often produces a cap over an isothermal layer in warm core eddies (Nilsson and Cresswell, 1981; Cresswell, 1983), preserving the distinct eddy characteristics below the surface. Finally, offshore or onshore transport occurs via eddy circulation. Offshore transport of shelf waters was observed by Joyce et al. (1992) and onshore transport of offshore waters was observed by Tranter et al. (1986). The effect of the entrainment of coastal or offshore waters around the periphery of eddies tends to increase primary production in the high velocity boundary regions. Generation mechanisms and influence of topography Eddies are primarily generated either from strong boundary current meanders [e.g. Gulf Stream, Kuroshio; Kamenkovich (1986b)], or by baroclinic instabilities caused by topographical or geographical features (e.g. off coast of California; Owen, 1980). Other mechanisms include the formation of eddies by direct wind forcing (Bernstein and White, 1974; Reznik, 1986; Thomson and Gower, 1998; Spall, 2000), tidal rectification (Thomson and Wilson, 1987), bottom relief (Huppert and Bryan, 1976; kamenkovich, 1986a), or by changes in bottom slope (Cenedese and Whitehead, 2000; Nof et al, 2002). Away from boundary currents, the turbulent kinetic energy associated with mesoscale eddies decreases, with slower swirl speeds and rates of westward propagation (Bernstein et ai, 1977). 11 Eddies can be modified in shape (Nof, 1988; Shapiro et al, 1995; Simmons and Nof, 2002) or direction of propagation (Thomson and Freeland, 2003) by topographic features. Eddies can also become trapped by topography (Yankovsky, 1997). Changes to eddy shape can influence local flows and induce secondary circulation patterns that produce upwelling or downwelling regions within a given eddy (Hitchcock et al, 1993; Okkonen et al, 2003). Further, some eddies can shed smaller eddies or dipoles in their wake (Ahlnas et al, 1987; Simpson, 1990; Forristall et al, 1992) as they travel and lose integrity due to frictional decay or collisions with objects (Shapiro et al, 1995; Wang and Dewar, 2003) or currents (Fratantoni and Glickson, 2002). Kinetic energy can also be modified by the merging of two eddies (Dewar and Killworth, 1990). Eddies have been shown to remain attached to seamounts for extended periods of time, for example in the Gulf of Alaska (Bograd, 1997). Translation Mesoscale eddies propagate westward at speeds generally between 1 and 5 km day"1. The translation of eddies from one locale to another is an important means by which to transfer materials (Berloff et al, 2002) as well as heat and salt (Hogg and Stommel, 1985; van Ballegooyen et al, 1994) and energy (Tai and White, 1990). Eddies always propagate westward, whether they are cyclonic (counterclockwise) or anticyclonic [clockwise; (Nof, 1981; Cushman-Roisin et al, 1990)] due to the |3-effect. The P-effect describes the influence of latitudinal differences in the magnitude of the Coriolis parameter (a measure of planetary rotation as a function of latitude) on fluid motion that leads to westward propagation of waves or eddies. Latitudinal differences in the Coriolis parameter lead to a gradient in vorticity between equatorward and poleward flows, which causes a westward propagation, irrespective of eddy polarity (Cushman-Roisin et al, 1990). The rate of propagation is proportional to the ratio of 12 potential energy to volume of the eddy (Shapiro, 1986; Cushman-Roisin et al, 1990). In special cases where eddies may occur in "stacked" pairs (called joint vortices), theory suggests that eastward drift can occur (Nof, 1985) when a lens lies atop a cyclonic vortex due to non-linear processes; these special cases have not been commonly observed, however, and would most likely be seen in lens development arising from outflows, for example, in the Mediterranean Sea. In contrast to eddies formed from western boundary currents (e.g. Gulf Stream rings and Kuroshio rings), eddies spawned in eastern boundary regions drift away from the coast; eddies formed in the Bay of Biscay (Pingree and Le Cann, 1992), off Mexico (Lukas and Santiago-Mandujano, 2001), off South Africa (Olson, 1992; van Ballegooyen et al, 1994), and in the Gulf of Alaska (Whitney and Robert, 2002) all transport coastal waters offshore. Eddies spawned from eastern boundary currents are important because they carry water westward (Nof, 1981) against the prevailing currents and persist for long periods of time compared to those spawned from western boundary currents such as the Gulf Stream (Joyce et al, 1984). Detection of mesoscale features As mentioned above, eddies were first documented in the northwest Atlantic using neutrally-buoyant floats in the 1950-60's that traced surface circulations (Simpson, 1983). After this, a series of current meter arrays were employed to map an energetic anticyclonic eddy in the tropical northeast Atlantic [POLYGON, 1970; in Simpson (1983)]. As an alternative to current measurements or drifter tracks, hydrographic qualities such as temperature and salinity anomalies were used to locate eddies; expendable bathythermographs (XBT) were used to detect such anomalies and map their presence in various oceanic environments [e.g. Bernstein and White (1977)], while ocean colour images could be used to detect thermal anomalies (Stumpf and Legeckis, 1977). Ocean colour scanners require, however, cloud-free skies since 13 these beams do not penetrate clouds. Today, detection relies mainly on satellite radar altimetry that utilizes the surface height anomalies imposed by geostrophic circulation and employs microwaves that penetrate through clouds. Geoid and mean currents are removed from data through extraction of the temporal mean at each point. The resulting anomalies are passed through a matched filter designed to detect Gaussian signals embedded in noise. Further mapping of flow fields and circulation velocities utilize satellite-tracked drifting buoys. Initially, the satellites used for radar altimetry were TOPEX/POSEIDON [T/P, operated jointly by the United States (NASA) and the French Space Agency, "Centre Nationale d'Etudes Spatiales" (CNES)] and ERS-2 (operated by the European Space Agency). The T/P satellite repeated its orbit once every 9.95 days while ERS-2 required 35.0 days to repeat its orbit. The satellite coverage was expanded with the introduction of Jason-1, launched in December 2001 (http://topex-www.jpl.nasa.gov/mission/mission. html), and Geosat Follow-on (GFO), launched in February 1998 (http://gfo.bmpcoe.org/ Gfo/Mission/missiona.htm). The Colorado Center for Astrodynamics Research (CCAR) maintains an Internet site (Global Near-Real-Time Altimetry Data Viewer, http://e450.colorado.edu/realtime/gsfc_global-real-time_ssh/) that processes raw data and produces near real-time images of SSH anomaly fields (maintained by Robert Leben; e.g. Fig. i-4). The images produced are based on the latest 10 days of Jason and T/P, 17 days of GFO and 35 days of ERS-2 sampling. By combining several satellite tracks the spatial resolution is drastically improved; the resolution is currently approximately a 6 x 6 km grid. 14 Figure i-4. Radar satellite altimetry contours showing the presence of a Haida eddy (about 53°N and 136°W) to the west of the Queen Charlotte Islands on June 18, 2000 (in orange). Height anomalies of ± 5 cm indicate the presence of a mesoscale eddy. Image provided by the Colorado Center for Astrodynamics Research at http://e450.colorado.edu/realtime/gsfc_global-real-time ssh/. 15 Function of mesoscale eddies Heat and salt flux Mesoscale eddies represent a mechanism by which heat and salt can be transferred from one oceanic region to another (Wyrtki, 1965; Gordon, 1986; van Ballegooyen et al, 1994; Richardson et al, 2003; Crawford, in press). The re-distribution of waters via eddy circulation is important in setting heat, salt, and energy budgets which influence vertical stratification and mixing. For example, through the "leakage" of Indian Ocean water into the Atlantic Ocean (Gordon, 1986; Lutjeharms and Cooper, 1996) eddy circulation in the vicinity of the Agulhas Current and retroflection at the southeastern tip of Africa is thought to have a strong influence on overturning circulation of the ocean (Gordon, 1985; Gordon et al, 1992); by mixing warm, salty water in the Benguela Current (Gordon, 1986), heat is directed northward in near-surface waters . (de Ruijter et al, 1999). In this way, eddy formation and transport play an important part in global ocean circulation and climate regulation. In the northeast Pacific, the number of eddies and the surface height anomaly of each eddy is influenced by the El Nino Southern Oscillation cycle (Meyers and Basu, 1999). Eddies spawned off the eastern land boundary in the Gulf of Alaska deliver heat and freshwater into the Alaska Gyre (Crawford, in press), and may contribute to the freshening and shoaling of the upper mixed layer observed in recent years (Freeland et al, 1997). Nutrient supply In biological terms, nutrient supply is probably the most significant role that mesoscale eddies play. Until permanent mooring arrays were deployed to monitor biogeochemical properties such as nutrient concentrations and optical properties in the ocean [e.g. the Bermuda Atlantic Time-Series (BATS) program], annual nutrient budgets for oligotrophic oceans 16 constructed from in situ primary production and nutrient uptake rate measurements appeared to underestimate annual new production. However, when continuous monitoring techniques became available, it was apparent that mesoscale eddy circulation represented a significant contribution to annual new production, including episodic events that allowed for the closing of these nutrient budgets and estimates of new primary production (McGillicuddy and Robinson, 1997b; McGillicuddy et al, 1998). The function of mesoscale eddies in supplying nutrients to oligotrophic gyres has been explored in the Sargasso Sea and in the subtropical North Pacific. Depending on the locale of formation and the sense of circulation, eddies can either deliver nutrients [e.g. anticyclonic eddies spawned from eastern boundary currents or cold core eddies spinning counter-clockwise (in the Northern Hemisphere)], or they can harbour nutrient-poor waters characterized by low chlorophyll a (e.g. Warm Core Rings of the Gulf Stream that carry oligotrophic Sargasso Sea waters in their core). The classical picture of eddy-nutrient dynamics suggests that anticyclonic eddies suppress primary production through the depression of isopleths below the euphotic zone, while cyclonic eddies enhance primary production by upwelling deep nutrients into the euphotic zone (Bidigare et al, 2001); (Fig. i-5). The current view of the role of mesoscale eddies in the control of nutrient inventories is considerably more complicated (Tranter et al, 1980; Zhang et al, 2001). Perhaps more significant than the nutrient regime within mesoscale eddies themselves is the eddy-induced redistribution of waters through horizontal transport and vertical mixing (upwelling or downwelling). Eddy-induced stirring and mixing (Hitchcock et al, 1994; Washburn et al, 1998; Abraham et al, 2000; Abraham and Bowen, 2002), lateral transport across ocean basins (Paldor et al, 2003) or between gyre boundaries and their centres, and vertical displacement of isopycnals represent important mechanisms that lead to variability in nutrient supply throughout the oceans (Garcon et al, 2001). In the relative nutrient 'desert' of the 17 1 0 0 % I, Euphot ic zone Shoaling isopycnals: Nutr ient in ject ion s t imu la tes phy top lank ton g rowth 1% U p w e l l i n g ; \ D o w n w e l l i n g A Deepening isopycnals: N o e c o s y s t e m r e s p o n s e Figure i -5 . Schematic diagram showing the influence of isopycnal displacement induced by anticyclonic (left, downwelling) and cyclonic (right, upwelling) mesoscale eddies. In the classical model cyclonic eddies tend to enhance primary production by injecting nutrients into the euphotic zone while anticyclonic eddies do not stimulate phytoplankton growth. large ocean gyres eddies may represent an important source for nutrients and organisms that may alter the productivity in these vast areas (Tranter et al., 1980; Falkowski, 1991; McGillicuddy and Robinson, 1997b;. McGillicuddy et ai, 1998; Siegel et al., 1999). Eddies are thought to be extremely influential in directing the downward fluxes of biogenic materials produced in the upper ocean, and impact the interpretation of data collected from sediment traps (Siegel et al., 1999). 18 Mechanisms that supply nutrients to anticyclonic mesoscale eddies include the following: (1) frictional decay (Nelson et al, 1989; Krom et al, 1992; Krom et al, 1993), (2) upwelling at eddy boundaries (Simpson, 1984; Okkonen et al, 2003), (3) convective overturn (Yentsch and Phinney, 1985; Levy et al, 1998), (4) offshore transport (Pingree and Le Cann, 1992; Lukas and Santiago-Mandujano, 2001; Whitney and Robert, 2002), (5) isopycnal transfer (Lee and Williams, 2000), (6) secondary circulation (Godfrey et al, 1980; Martin and Richards, 2001), and eddy-eddy interactions (Hitchcock et al, 1993) or interactions with currents (Joyce et al, 1984; Lillibridge et al, 1990; Joyce, 1992). Eddy pumping accounts for enhanced productivity within cyclonic eddies, but is not observed in anticyclonic eddies (Falkowski, 1991; McGillicuddy and Robinson, 1997b; Bidigare et al, 2001). Light and Vertical Mixing In regions where seasonal convective cooling of ocean waters is significant, enhanced vertical mixing within mesoscale eddies can depress primary production compared to surroundings by mixing phytoplankton down to depths greater than Sverdrup's critical depth (Sverdrup, 1953; Piatt et al, 1991a; Levy et al, 1998). Differences in the light regime are thought to be responsible for differences in the timing of bloom events compared to surrounding non-eddy waters (Hitchcock etal, 1985; Hitchcock et al, 1987). For example, the spring maximum in phytoplankton biomass and primary productivity in WCR 82-B lagged behind that of the slope water and the surrounding Sargasso Sea (Hitchcock et al, 1985). Biological features of eddies and meanders In addition to their importance to physical processes throughout the ocean, the spatial and temporal scales identified for mesoscale eddies correspond to important scales relevant to 19 biological processes. For example, characteristic eddy Eulerian timescales of weeks to months (and sometimes years), correspond to seasonal timescales (Longhurst, 2001), and thus can influence such events as the timing of spring blooms (Levy et al, 1998). Spatial variability in phytoplankton or zooplankton distributions is often related to the presence or absence of mesoscale eddies (McWilliam and Phillips, 1983; Tranter, 1983; Tranter et al, 1983; Joyce et al, 1984; Jeffrey.and Hallegraeff, 1987; Lochte, 1987; Allen et al, 1996; Kimura, 1997; McGillicuddy and Robinson, 1997b; McGillicuddy et al, 1998; Oschlies and Garcon, 1998; McGillicuddy et al, 1999; McNeil et al, 1999; Siegel et al, 1999; Bidigare et al, 2001; Logerwell et al, 2001; Mackas and Galbraith, 2001; Lima et al, 2002; Lane et al, 2003; Batten and Crawford, in press; Crawford et al, in press; Tsurumi et al, in press), and the transport of organisms within eddies or their boundaries can be important (McWilliam and Phillips, 1983); J. Dower, pers.comm.). Seabirds have also been shown to exploit enhanced production observed within mesoscale eddies (Nel et al, 2001). Doney et al. (2003) noted that the spatial and temporal scales of SeaWiFS data corresponded to those of TOPEX/POSEIDON - ERS 2 sea surface height anomalies (SSHA). The role of eddies in nutrient flux (Falkowski, 1991; Krom et al, 1993; Smith etal, 1996; McGillicuddy and Robinson, 1997b; Woodward and Rees, 2001; Martin and Pondaven, 2003) and transport of organisms make them an important part of regional models of primary and secondary production. Over the last 25 years, eddies have been studied with increasing intensity, with an emphasis placed on understanding and modeling internal processes, the roles that eddies play in regional biogeochemical cycling (McGillicuddy et al, 1998), and the contribution that eddies make to variability in the oceans. The physical environment within mesoscale eddies changes more dramatically than their surroundings under the influence of disturbances such as wind or storm events due to the greater vertical diffusion rates typical of these features (Levy et al, 1998). For this reason, short-20 timescale responses of phytoplankton to physical processes in Warm Core Rings (WCR) spawned from the Gulf Stream tend to be more dramatic within the rings compared to the background. Hitchcock et al. (1987) observed that phytoplankton standing stocks and productivity responded rapidly to event-scale changes in vertical mixing in WCR 82-B; increases in primary productivity were much more significant (20 times increase) than changes in phytoplankton biomass (4-fold increase). At the same time, a shift toward larger phytoplankton was observed; whereas nanoplankton (<20 um) accounted for -90% of the total chlorophyll a biomass initially, after the mixing event net phytoplankton (20 - 200 um) accounted for a significant proportion of total chlorophyll a standing stock. The confinement of the enhanced phytoplankton response to mixing to the ring centre indicated that different processes were controlling productivity at the centre versus the edge. A deep-mixing event in WCR 82-B led to the breakdown of a subsurface chlorophyll a maximum, yielding a uniform distribution of chlorophyll a over the water column (Hitchock et al, 1987). Fryxell et al. (1985) noted that such physical events could be well recorded by observing short-term changes in species assemblages since physical intrusions of slope waters produced very different assemblages compared to an undisturbed ring centre (WCR 81-D). After the mixing event observed by Hitchcock et al. (1987) standing stocks and primary productivity were lower at the higher velocity edges of the ring compared to the ring centre. More typically, the opposite pattern has been observed where edge regions support higher productivity and plankton biomass (Tranter et al, 1983; Hitchcock et al, 1993). Therefore, the physical environment (circulation patterns, disturbance events, water mass mixing) and eddy longevity appear to influence primary production and the distribution of standing stocks in distinct manners depending on eddy location, season, and frequency of stochastic events. 21 Finally, the isolation of eddy waters from their surroundings can influence the species assemblages and/or the balance of heterotrophy and autotrophy within these features. When rings or eddies are isolated for a sufficient time period, biomass accumulation can occur within their waters. For example in W C R 82-B, carbon biomass increased as the ring developed (in June compared to April), which may be attributable to higher heterotrophic biomass (Hitchcock et al., 1987). Fryxell et al. (1985) also noted the build-up of biomass at ring centres, particularly following storm events where growth was stimulated concurrently with the re-establishment of isolated conditions. In Cyprus Eddies of the Mediterranean Sea, deep mixing (to 500 m) in spring resulting from storm events was thought to stimulate phytoplankton blooms by diluting grazers and injecting nutrients into the euphotic zone; the quiescent periods between deep mixing, events appeared to foster phytoplankton bloom development in these warm core eddies (Zohary, 1998). In the only study of aging eddies, Fryxell et al. (1985) noted that species diversity tended to increase in a three-month old ring 81-D compared to a newly spawned ring (82-H), with diatoms generally thriving at ring centre and dinoflagellates occurring in the highest abundance at ring margins. Eddies in the North Pacific Most of the mesoscale activity in the North Pacific was thought to arise from the Kuroshio and its eastward extension in the Western North Pacific (Emery, 1983). Strong mesoscale eddies are formed from the Kuroshio Current, including anticyclonic, cyclonic, and frontal eddy systems. Spatial scales for these features are on the order of 50 km for frontal transitions and 150 - 200 km for eddies (Roden, 1970, 1972, 1977, 1979). Cheney (1977), and Cheney et al. (1980a) observed similar features using aircraft expendible bathythermographs (XBT), ship-based hydrographic data, and free-drifting surface buoys (Cheney et al, 1980b). 22 Bernstein and White (1981) used X B T data to demonstrate both the ability to map the thermocline expression of mesoscale features and to follow them from one month to the next, as they propagated slowly westward (ca. 4 km day "'). In addition to mesoscale activity associated with the Kuroshio and its eastward extension, meanders and eddies have now been documented off the coast of California, off the coast of Mexico (Lukas and Santiago-Mandujano, 2001), and in the Gulf of Alaska (Tabata, 1982). Kirwan et al. (Kirwan et al., 1978) observed an anti-cyclonic eddy at 36°N, 162°W with a diameter of 100 km and a rotational speed about 3 times smaller than typical western Pacific eddies; it propagated more slowly, about 2 km a day. Eddies are abundant south of 35°N (Bernstein and White, 1974; Roden, 1980).The eddies generated off the west coast of North America appear to be related to strongly baroclinic flows in this region and may be spawned by instabilites in the flow (Emery, 1983). There is also increased mesoscale activity far from strong boundary currents off Hawaii (Wyrtki et al, 1976). These eddies tend to be relatively small, having length scales of less than 100 km (Bernstein and White, 1974). Oceanographic features of the Gulf of Alaska The northeast subarctic Pacific is delimited by the eastward flowing subarctic current, approximately 45°N, the southward flowing arm of the Alaska Gyre to the west (ca. 180°W), and by land masses (Alaska and British Columbia) to the north and east (Favorite et al, 1976); Fig. i -6). The gyre is cyclonic in circulation, with slow upwelling at the centre. Low iron concentrations limit phytoplankton growth in the oceanic realm of the northeast subarctic Pacific, one of the world ocean's High Nitrate, Low Chlorophyll (HNLC) regions (Martin, 1988; Martin and Fitzwater, 1988; Boyd et al, 2004). As a result, macronutrients (nitrate, phosphate, 23 Figure i-6. Map of the Gulf of Alaska showing major currents, the Subarctic Current (SUB), the Alaska Coastal Current (ACC) and the Alaska Stream, which form the cyclonic, clockwise gyre. The dashed line shows the approximate limit of the High Nitrate Low Chlorophyll (HNLC) region. The dotted lines represent the transition zone between coastal and oceanic H N L C waters. A K = Alaska, BC = British Columbia, QCI = Queen Charlotte Islands. Also noted on the map is the region of downwelling to the north of the gulf, along the coast of Alaska, and the region of summer upwelling off the coast of Vancouver Island, British Columbia. silicic acid) remain in excess within the mixed layer throughout the year (Anderson et al, 1969; Whitney and Freeland, 1999) and photosynthetic biomass remains relatively constant at ca. 0.2 -0.4 mg m"3 (Parslow, 1981; Boyd et al, 1995a; Boyd and Harrison, 1999; Harrison, 2002). Recently, satellite observations have shown that there is a systematic increase in chlorophyll a biomass in the Gulf of Alaska over the spring - summer period (Brickley and Thomas, 2004). This increase in chlorophyll a may result from concentration within the mixed layer as it shoals, ' 2 4 or from increased primary production. In a study examining winter primary production in the centre of the Alaska Gyre (Ocean Station Papa, 50°N, 145°W), Boyd et al. (1995b) found that winter rates were approximately half of summer rates, and spring rates are higher than summer rates (Boyd and Harrison, 1999), suggesting that although summer rates are higher than winter, the highest rates occur in spring, and the increase in chlorophyll biomass results more from physical processes than from biological ones. The northeast subarctic Pacific is characterized by three major domains based on surface nutrient concentrations (Wong et al, 2002a) that were based on surface currents by (Favorite et al, 1976). There is: (1) a dilute domain (DIL) that lies east of the (2) Subarctic Current, the Subarctic Current System, or transition domain (SUB), and (3) a High Nitrate, Low Chlorophyll region that lies west of the Subarctic Current (HNLC). The DIL domain is characterized by low nutrient concentrations after spring growth occurs, while the H N L C domain harbours high nutrients year-round (Whitney and Freeland 1999). This domain is exemplified by the extensively studied ocean time series station, Ocean Station Papa (OSP, 50°N, 145°W). The SUB domain resembles either of these two domains, and shows greater spatial and temporal variability owing to its dependence on local currents to deliver nutrients. The SUB domain can at times lack both macronutrients and trace metals, and often shows very low chlorophyll concentrations in surface waters (Boyd and Harrison 1999). In contrast to the centre of the gyre, the coastal region experiences summer upwelling with the reversal in wind direction between winter and summer. Here, chlorophyll a biomass and rates of primary production are very high (Pena etal., 1996; Harris, 2001). Coastal waters are more productive than oceanic environments (Ryther, 1969; Barber and Smith, 1981) due to the availability of land-derived nutrients, to physical dynamics that lead to nutrient delivery, and due in part to buoyant, shallow mixed layers associated with river outflow and resulting low salinity 25 waters that keep phytoplankton in well-lit environments. Beyond the continental shelf, however, the supply of nutrients decreases and rates of primary production are lower than along the coast. Between the two environments lies a region that receives occasional inputs from the coast as well as from the oceanic realm and is at times characterized by nitrate or iron limitation. In this transition region, primary production varies considerably. In the Gulf of Alaska, this coastal transition zone appears to be very broad based on satellite images of photosynthetic particulates (Brickley and Thomas, 2004). This region is populated with mesoscale eddies, jets, and meanders. Limited biological data suggest that phytoplankton growth in these regions is often lower than the coastal or oceanic realms. Taylor and Waters (1982) observed very low cell abundances in the dilute domain (DIL) compared to other stations in the eastern and western northern subarctic Pacific, and records from the Line P program suggest chlorophyll a values lay between 0.2 - 0.3 pg L" 1 in these waters (Wong et al., 1999). This situation may arise from a combination of low macronutrients for the coastal section of Line P and low micronutrients for the more open ocean stations of Line P out to OSP (Harrison, 2002). Mesoscale eddies in the Gulf of Alaska Mesoscale eddy activity has long been noted in the Gulf of Alaska, beginning with observations of temperature and salinity anomalies in hydrographic data (Tully et al, 1960; Tabata, 1982) and model outputs that predicted the formation of eddies off the coasts of British Columbia and Alaska (Melsom et al, 1999; Crawford et al, 2000). More recently, data derived from satellite remote sensing have revealed the presence of many mesoscale features, including eddies. Sea surface height anomalies (SSH) [from TOPEX-Poseidon/ERS-2 radar altimeter; (Meyers and Basu, 1999; Crawford, 2002)], surface temperature anomalies [Advanced Very High Resolution Radiometer (AVHRR); Gower and Tabata (1993); Gower (1998)], and high 26 chlorophyll patches [Sea-Viewing-Wide-Field-of-View sensor (Sea-WiFS); Crawford et al., in press)] all correspond to the position of mesoscale eddies in the Gulf of Alaska. Most eddies spawned along the eastern/northeastern perimeter of the Alaska Gyre are long-lived and anticyclonic (Tabata, 1982; Crawford et al, 2000). Recently, shipboard observations combined with satellite sea surface height imagery led to the discovery of an eddy spawning site at the southern tip of the Queen Charlotte Islands, British Columbia (Crawford and Whitney, 1999). These eddies, named "Haida" after the islands ("Haida Gwaii", or Queen Charlotte Islands) occupied by a local First Nation, transport large volumes of warm fresh water from Queen Charlotte Sound into the Gulf of Alaska (Crawford and Whitney, 1999; Whitney and Robert, 2002). Haida eddies are defined as those anticyclonic vortices generated south of 54.5°N, along the coast of the Queen Charlotte Islands (Crawford, 2002) and they are named by the year in which they are spawned (e.g. Haida-2000, formed in February 2000). Typically, three to five large eddies are generated per winter along the Alaskan Panhandle and Canadian West Coast (Crawford and Whitney, 1999). Haida eddies are typically 150-300 km in diameter and persist for 1-5 years (Crawford, 2002; Whitney and Robert, 2002), moving westward into the Gulf of Alaska at speeds of 5-40 km month"1 depending on their size and initial energy content (Crawford, 2002; Whitney and Robert, 2002). Typical core depths are . ca. 500-600 m at the centre where salinity does not exceed 33.9 (Crawford, 2002; Whitney and Robert, 2002). The paths of Haida eddies can be modified by climatology so that they either track due west, northwest or southwest (Whitney and Robert, 2002). Thus, these anticyclonic eddies can travel either into High Nitrate, Low Chlorophyll waters to the west or northwest or into seasonally nutrient-depleted waters to the southwest. Whitney and Robert, (2002) were the first to document the impacts of mesoscale eddies in an H N L C region. They noted that eddies that 27 travel into these waters appear nutrient poor relative to their surroundings, presumably due to high phytoplankton growth and drawdown of nutrients compared to iron-limited (Martin and Fitzwater, 1988; Boyd et al, 1996) macronutrient replete (Anderson et al, 1969; Whitney and Freeland, 1999) H N L C waters. The Haida eddies project The presence of Haida eddies was first detected by W.R. Crawford and F.A. Whitney of the Institute of Ocean Sciences, Fisheries and Oceans Canada (Sidney, B.C.), on a routine sampling trip along Line P (Crawford and Whitney, 1999). Although Haida eddies are common features in the Gulf of Alaska, little is known about their biological or chemical impacts on the northeast Pacific. The Haida eddies project arose from the need to characterize this newly discovered type of eddy spawned from the coast of the Queen Charlotte Islands. Between February 2000 and September 2001, six cruises were undertaken to characterize the major features of these eddies. The work involved scientists from the Institute of Ocean Sciences, University of British Columbia, University of Victoria, the Pacific Biological Station (Nanaimo, B.C.), and the University of Southampton (Southampton, U.K.) . The transport of heat and freshwater into the Gulf of Alaska (W.R. Crawford), changes in the carbon system (M. Chierici and L .A . Miller), zooplankton distributions (D.L. Mackas, D.R. Yelland, M . Tsurumi, M . Galbraith, S.D. Batten), relationship between pteropod distributions, climate, and eddies (M. Tsurumi, D.L. Mackas), larval transport (I. Perry, J. Dower), the formation of Haida eddies (J. Chiernawsky, M . Foreman, W.R. Crawford), silicic acid uptake and export in the Gulf of Alaska (F.A. Whitney), nutrient transport (F.A. Whitney, M . Robert), current structure (D.R. Yelland and W.R. Crawford), trace metal distribution (S.M. Crispo, K.J . Orians) and speciation (M.C. Lohan, P.J. Statham), and primary production, species composition (this thesis), and all of the 28 parameters above were investigated simultaneously to give a synoptic view of the evolution of these long-lived eddies. Summary Mesoscale circulation in the oceans encompasses such features as eddies, meanders and jets (Robinson, 1983). Eddies are ubiquitous in all ocean basins, and play important roles in nutrient supply (McGillicuddy and Robinson, 1997b; McGillicuddy et al, 1998; Siegel et al, 1999), larval transport (McWilliam and Phillips, 1983; Haury et al, 1986), and vertical mixing/stratification (Nelson et al, 1989). These features can now be easily detected using radar altimetry (TOPEX/POSEIDON-ERS-2), and comparisons of height anomaly contours with biological tracers such as chlorophyll a (Crawford et al in press) or thermal imagery (Seki et al, 2001) allow one to examine controls on phytoplankton production that may be driven by eddy circulation. Thesis goals - The main goal of the Haida eddies project was to determine what impact these eddies have on biological productivity of the eastern Gulf of Alaska. The primary goals of this thesis were to measure and document the biological processes occurring in two Haida eddies. Haida-2000 was followed over a 20 month period as theyit moved offshore into H N L C waters. Particularly, and Haida-2001 was followed from February 2001 when it was spawned, to September 2001. Particular emphasis was given to the phytoplankton and nutrient components, and a comparison to nearby non-eddy waters. Specifically, the objectives of this thesis were: (1) to describe nutrient dynamics within Haida eddies and relate observed.changes to biological processes (nutrient uptake/drawdown); 29 (2) to characterize the phytoplankton community within and outside of the Haida eddies by identifying the dominant species present; (3) to measure rates of primary production and photosynthetic characteristics of phytoplankton within and outside of the Haida eddies; (4) to characterize nitrogenous nutrition of phytoplankton communities within Haida eddies by measuring rates of nitrate and ammonium uptake, and (5) to simulate the response of phytoplankton to the mixing of water masses together (as would occur at the eddy edge or in the event of eddy collapse) through a shipboard experiment. Thesis outline and overview This thesis is organized into six chapters describing different aspects of the biological evolution of two anticyclonic mesoscale Haida eddies in the Gulf of Alaska. The main focus was on Haida-2000; comparisons with Haida-2001 were made when appropriate. In order to gain a thorough understanding of the function of theseHaida eddies, several biological processes were studied simultaneously. In Chapter 1,1 present concentrations of nutrients (nitrate, phosphate, and silicic acid), and discuss their distributions: (1) in terms of changes in the eddy core that indicate interactions between eddies and their environment, and (2) using changes observed in the mixed layer to infer seasonal patterns in productivity and nutrient availability. This information was necessary to put the observations of the subsequent chapters into context. Chapters 2 and 3 focus on standing stocks of phytoplankton within Haida eddies-2000 and Haida-2001 and their surroundings. Chapter 2 presents spatial and temporal patterns in the distribution of photosynthetic biomass and standing stocks of particulate organic carbon, nitrogen, and biogenic silica in waters inside and outside of the eddies. Mesoscale variability in these biological parameters indicates the importance of advection, frontal boundaries, and isolation from surrounding waters. Chapter 3 discusses the species composition within Haida 3D eddies and their surroundings, relating the phytoplankton community structure to spatial/temporal scales and environmental variables. Chapters 4 and 5 discuss primary production in the context of an iron-limited ecosystem. Chapter 4 presents measurements of carbon fixation by radio-labelled carbon uptake and studies of photosynthesis-light relationships, while in Chapter 5 I present uptake rates of nitrogen and ammonium. Through a controlled shipboard experiment, growth rates, nutrient drawdown and pigment signatures are examined in an attempt to characterize the response of phytoplankton to water mass mixing by advective and stirring-induced mixing near ocean fronts in Chapter 6. A general conclusion section ties together the findings from the five sections in a chronology of the evolution of Haida-2000, with comparisons to Haida-2001. Appendices show the determination of a substrate affinity constant for diatom assemblages at Ocean Station P, cautionary notes on the influence of pH in the analysis of biogenic silica, selected photographs of phytoplankton species observed in this study, and primary productivity model parameters and their derivations. 31 C H A P T E R 1. MACRONUTRIENT DYNAMICS IN A HAIDA EDDY 1.1 Introduction The eastern Gulf of Alaska can be broadly divided into coastal and High Nitrate Low Chlorophyll (HNLC) waters which are separated by the Alaska Current, the northward-flowing arm of the Subarctic Current [Fig. 1-1; (Whitney and Freeland, 1999; Wong et al, 2002)]. Sources of nutrients to the region include coastal upwelling in summer, outflow from major rivers such as the Fraser and Columbia, and deep mixing in winter. Recently, mesoscale eddies spawned from the coast have been shown to carry nutrients (Whitney and Robert, 2002) and Fe (Johnson et al., in press) into the Gulf of Alaska, and therefore represent another important contribution to total nutrient supply. In other areas, such as the Gulf Stream and Kuroshio current systems, mesoscale eddies have been shown to be important in closing annual nutrient budgets and in estimating annual primary productivity [e.g. (McGillicuddy and Robinson, 1997b; McGillicuddy etal, 1998; Siegel etal, 1999)]. Mesoscale eddy activity has long been noted in the Gulf of Alaska, beginning with observations of temperature and salinity anomalies in hydrographic data (Tully et al, 1960; Tabata, 1982) and model outputs that predicted the formation of eddies off the coasts of British Columbia and Alaska (Melsom et al, 1999). More recently, data derived from satellite remote sensors have revealed the presence of many mesoscale features, including eddies. Sea surface height anomalies (SSH) (from TOPEX-Poseidon/ERS-2 radar altimeter; (Meyers and Basu, 1999; Crawford, 2002), surface temperature anomalies [Advanced Very High Resolution Radiometer (AVHRR); Gower and Tabata (1993); Gower (1998)], and high 32 148 146 144 142 140 138 136 134 132 130 128 126 124 122 Longitude (degrees W) 148 146 144 142 140 138 136 134 132 130 128 126 124 122 Longitude (degrees W) Figure 1 - 1 . Regional map showing nutrient domains and the location of the centre of Haida-2000 as it traveled northwestward into the Gulf of Alaska. The eddy trajectory followed the line from Feb 2000 to Sept 2001 with each symbol representing the position of the centre at the time 33 of sampling. Reference stations in 2000 were located ca. 30 - 50 km from the southern edge of Haida-2000 (not shown), while in 2001 reference stations were located further away; these are denoted as Jun-R and Sept-R for June and September sampling periods. For reference, time series Line P stations (P4 to OSP) are included. Arrows indicate direction of major currents; A C = Alaska Current, SAC = Subarctic Current. Current/nutrient domains (Favorite et al, 1976; Wong et, al, 2002a) are designated A G = Alaska Gyre domain, SUB = Subarctic Current domain, DIL = Dilute domain. chlorophyll patches [Sea-viewing-Wide-Field-of-view Sensor (Sea-WiFS); Crawford et al, in press)] all correspond to the position of mesoscale eddies in the Gulf of Alaska. Most eddies spawned along the eastern/northeastern perimeter of the Alaska Gyre are long-lived and anticyclonic (Tabata, 1982; Crawford et al, 2000). Recently, shipboard observations combined with satellite sea surface height imagery led to the discovery of an eddy spawning site at the southern tip of the Queen Charlotte Islands, British Columbia (Crawford and Whitney, 1999). These eddies, named "Haida" after the islands ("Haida Gwaii, or Queen Charlotte Islands) occupied by local First Nation peoples, transport large volumes of warm fresh water from Queen Charlotte Sound into the Gulf of Alaska (Crawford and Whitney, 1999; Whitney and Robert, 2002). Haida eddies are defined as those anticyclonic vortices generated south of 54.5°N along the coast of the Queen Charlotte Islands (Crawford, 2002), and they are named by the year in which they are spawned (e.g. Haida-2000, formed in February 2000). Haida eddies generally have a diameter of 150 - 300 km and a core depth of 500-600 m at the centre where salinity does not exceed 33.9 (Crawford, 2002; Whitney and Robert, 2002). These eddies persist for one to several years (Crawford and Whitney, 1999). Eddies spawned from eastern boundary currents are important because they carry water westward (Nof, 1981) against the prevailing surface currents and persist for long periods of time compared to those spawned from western boundary currents such as the Gulf Stream (Joyce et 34 al, 1984). Eastern boundary current eddies thus serve to transport coastal waters offshore into oceanic regions (e.g. Whitney and Robert, 2002) that are generally either oligotrophic or lacking in trace metals (High Nitrate Low Chlorophyll) (Longhurst, 1998). The paths of Haida eddies can be modified by climatology so that they either track due west, northwest or southwest (Whitney and Robert, 2002). Thus, these anticyclonic eddies can travel either into High Nitrate Low Chlorophyll waters to the west or northwest or into seasonally nutrient-depleted waters to the southwest. Whitney and Robert (2002) were the first to document the impacts of mesoscale eddies in an H N L C region. They note that eddies that travel into these waters appear nutrient poor relative to their surroundings, presumably due to high phytoplankton growth and drawdown of nutrients compared to iron-limited (Martin and Fitzwater, 1988; Boyd et al, 1996), macronutrient replete (Anderson et al, 1969; Whitney and Freeland, 1999) H N L C waters. In contrast to the H N L C regions, waters just off the continental margin can become nitrate depleted in summer, limiting phytoplankton growth (Whitney et al, 1998). If Haida eddies travel southward into these regions where nitrate becomes exhausted, they appear nutrient rich compared to their surroundings due to their high coastal nutrient reservoirs (Whitney and Robert, 2002). Since eddies transport both macronutrients (Whitney and Robert, 2002) and Fe (Johnson et al, in press) into the Alaska Gyre they are important in making annual estimates of primary productivity and nutrient supply in the region no matter what path they take. It is thus necessary to understand their evolution once they leave the coast in order to assess how they might impact the waters through which they travel. The coastal nature of eddies and their tendency to transport waters great distances from the continent extends the "coastal-type" transition zone waters further offshore. The presence of high-chlorophyll waters extends much further into the open ocean than can be accounted for by the location of the narrow continental shelf along the west coast of British Columbia and Alaska 35 (Okkonen et al, 2003). It has been suggested that this phenomenon arises due to the entrainment of coastal waters by the anticyclonic rotary motion of eddies, pushing these high-chlorophyll waters offshore as oceanic waters are drawn toward shore (Crawford et al., in press). Thus, the delivery of nutrients and effects of stirring and entrainment by eddies could be a particularly important influence on primary productivity in the transition region between the coast and the open Alaska Gyre. Since Haida eddy formation appears to be influenced by climatic events such as El Nino (Mysak, 1985; Melsom et al., 1999), understanding the impact that they have on ecosystem productivity is important in strengthening our ability to make predictions about the large-scale response of biological communities within the Gulf of Alaska to changes in climate. In particular, the nature of nutrient supply within Haida eddies and its influences on primary production have not yet been explored. In other oceanic regions, eddies have been shown to supply nutrients via transport from deep waters along isopycnals (Lee and Williams, 2000) and through secondary circulation patterns induced by rotary motions and ageostrophic currents (Simpson, 1984; Martin and Richards, 2001), and to enhance productivity by creating physical fronts (Franks, 1992, 1997). One way of assessing the relationship between nutrient supply and its biological consequences is to examine rates of new production by primary producers. "New" production describes phytoplankton growth that is fuelled from an external, or non-recycled, source of nutrients (Dugdale and Goering, 1967). Nitrate is the major currency used to describe this growth, and the main source of new nitrate is from upwelling (Piatt et al, (1989), although other sources such as nitrogen fixation can be important as well, particularly in tropical environments (Karl et al., 1997). Under steady-state conditions, the nitrate that is used by phytoplankton and exported below the pycnocline is replaced predominantly by nitrate from below, so that nitrate 36 loss approximates the export of particulate material (i.e. export production). The vertical downward flux of particulate material drives the "biological pump" and mediates the flux of CO2 from the atmosphere to the surface ocean. Chapter 1 examines the changes in nutrient concentrations in surface waters and in the eddy core of one Haida eddy (Haida-2000) as it evolved over 20 months on a journey from the coast to H N L C waters of the Gulf of Alaska. The study of nutrient dynamics within such a quasi-closed system allows us to examine changes in the eddy core that result from aging processes such as frictional decay and dilution. Evolution of core waters can influence local mesopelagic nutrient distributions by supplying nutrients along isopycnals (Lee and Williams, 2000), or through frictionally-induced changes in meridional flow (e.g. Fukumori, 1992). Surface nutrient dynamics track biological production and likely reflect changes in iron availability as eddies age and trace metals are stripped out of the euphotic zone. 1.2 Materials and Methods 1.2.1 Sampling Haida-2000 was spawned in late January/early February 2000, and surveyed six times between February 2000 and October 2001 on the C.C.G.S. John P. Tully (Table 1-1). For each cruise, the location of Haida-2000 was first determined using TOPEX/Poseidon-ERS-2 satellite altimetry and then sampled along a transect bisecting the eddy on each cruise in order to identify the specific location of the centre and edges. In spring and summer of 2000 and 2001 more detailed chemical and biological sampling was conducted at three stations, one at the eddy centre, one at the edge, and one at a reference station outside the eddy. The reference sites were located in waters just outside the eddy, ca 30-50 km from the eddy edge at the time of sampling in June and September 2000, while in June and September 2001 reference sites were located 37 further away (180 km to the southeast in June 2001, and 160 km to the south in September 2001; Fig. 1-1). Conductivity, temperature, and depth (CTD) were profiled during each cruise at 10 stations across the eddy with data collected to 1000-3000 m, employing a Seabird 91 lplus CTD mounted on a 24-bottle rosette frame. Sigma-lvalues were based on CTD data. Fewer stations (4 in 2000, 2 in 2001) were sampled in February due to time constraints. Duplicate nutrient analyses (nitrate + nitrite, silicic acid, phosphate) were performed for 10-15% of depths sampled, with a mean standard deviation for all cruises of ±0.13 p M for nitrate, ±0.76 p M for silicic acid, Table 1-1. Cruise dates, locations, mixed layer depths (MLD), Sea Surface Height (SSH) anomalies, and eddy dimensions (km) for Haida-2000 sampling. Out refers to reference stations chosen outside but in the vicinity of Haida-2000. Centre refers to the eddy centre as identified by satellite altimetry and hydrographic data, and the edge is defined as waters in the region of the most steeply sloping isopycnals and swiftest currents. SSH data from TOPEX/POSEIDON-ERS-2 satellite radar altimetry processed by Colorado Center for Astrodynamical Research (CCAR) (http://www-ccar.colorado.edu/~realtime/gsfc_global-real-time_ssh/; R. Leben). Dimensions are in latitudinal (N-S) and longitudal (E-W) directions in km and are derived from the SSH plots. Depth of depression of deep isopycnals (m) represents the depth to which the 26.8 and 27.0 isopycnals are depressed relative to surrounding waters. Cruise Station Date Latitude Longitude MLD SSH Dimensions Isopycnal (°N) (°W) (m) (cm) (km) N-S Depression x E-W (m) Feb Out 14/02/00 2000 Centre 15/02/00 Edge 16/02/00 June Out 18/06/00 2000 Centre 20/06/00 Edge 21/06/00 Sept Out 26/09/00 2000 Centre 27/09/00 Edge 27/09/00 Feb Out 16/02/01 2001 Centre 17/02/01 June Out 07/06/01 2001 Centre 09/06/01 Edge 11/06/01 Sept Out 27/09/01 2001 Centre 23/09/01 Edge 27/09/01 52.33 131.99 52 52.34 134.67 110 52.33 133.67 68 51.74 135.84 24 52.75 135.83 29 52.25 135.83 28 51.75 136.17 32 52.75 136.16 31 52.24 136.17 28 53.80 139.50 81 53.80 138.00 81 52.75 137.00 24 54.42 138.17 19 54.92 138.16 20 52.50 138.75 32 54.50 138.33 33 54.00 138.75 23 12 144x211 100 20 167x311 150-200 12 200 x 266 100 10 100 x 178 200 5 100 x 167^ 100 10 222 x 189 200 00 39 and ±0.012 uM for phosphate. Sampling depths extended from 0 tolOOO-3000 m, with more fine-scaled sampling at shallow depths. Nutrient samples were collected into 15 mL plastic test tubes filled from Niskin bottles mounted on the rosette. Samples were stored at 4°C until analysis was possible onboard (usually within 12 h) or at -20°C for samples collected in September 2001 and processed ashore. Samples were run on a Technicon II AutoAnalyzer® following procedures in Barwell-Clarke and Whitney (1996). Eddy core nutrient concentrations were integrated from 75 m to the depth of the 27.0 isopycnal by the trapezoidal integration method. 500 mL samples for total chlorophyll a were collected from depths where nutrients were determined and concentrations were measured by fluorometry (Parsons et ai, 1984). More details on the method can be found in Chapter 2. 1.2.2 Statistics Due to the small number of observations and coarse sampling scheme, statistical analyses were not useful for all comparisons. A Student's Mest was used to compare integrated nutrient concentrations both within the eddy as it evolved and between the eddy and surrounding waters (eddy centre vs. reference stations) over time. For the first test, the integrated nutrient values for all sampling times were compared with the initial value obtained in February 2000. Statistical tests were performed using Jmpln 4.0 software. 1.2.3 Definitions The eddy centre refers to the site of highest sea surface/dynamic height anomaly and the location where the 27.0 isopycnal was deepest. The edge was chosen as the region with the strongest slope of isopycnals at depths of 200 - 500 m and the swiftest currents according to Acoustic Doppler Profiler Current measurements (Yelland and Crawford, in press). 40 The eddy core is defined as those waters contained in the volume between the eddy surface and the maximum depth at which coastal waters are found. This includes waters of salinity < 34.0 and extends to the depth of the 27.0 isopycnal (see Fig. 1-2). The top of the core water was defined as the depth at which there is less than a 0.1 sigma-^ difference between the centre and the outside (ca. 75 m). The bottom of the mixed layer is defined as the depth at which there is an increase in sigma-c?of 0.125 units compared with the surface density (Levitus, 1982). Nutrient concentrations within the mixed layer were calculated by averaging the discrete concentrations at depths between 0 m and the bottom of the mixed layer. 1.3 Results 1.3.1 Eddy evolution and structure 1.3.1.1 Chronology Haida-2000 was characterized by a depression of isopycnals and nutrient isopleths throughout the sampling period (Figs. 1-2 and 1-3). The Haida-2000 eddy was detected as a sea surface height anomaly of roughly 12 cm by radar satellite altimetry in February 2000. The shape was slightly oblong, with surface dimensions (defined by the 3 cm TOPEX/POSEIDON SSHA contours) of approximately 144 km (N-S) by 211 km (E-W; Table 1-1). The core extended to 600 m with isopycnals distorted to a depth of at least 1000-2000 m. The trajectory followed by Haida-2000 is shown in Figure 1 -1. After breaking away from the coast in February 2000, Haida-2000 moved 90 km westward, stalling over Bowie Seamount (53.5°N, 135.6°W) from May to September, and it drifted only 40 km westward over the summer. The interaction with the shallow seamount (pinnacle ~35 m below sea surface) distorted the shape of the eddy in the northern region (see eddy dimensions in Table 1-1). After 41 Sept-00 June-01 Sept-01 1 0 0 2 0 0 0 1 0 0 2 0 0 0 1 0 0 2 0 0 3 0 0 0 1 0 0 2 0 0 1 0 0 £ 2 0 0 3 0 0 4 0 0 £ 5 0 0 Q. 0) Q 6 0 0 7 0 0 8 0 0 9 0 0 . , I l i l S I l f l l , ^-26.6'' 51 .J4 52 .75 53 .75 |52 --26.6--" 175 54 .42 55. 100 200 0 100 200 50 150 250 50 100 150 200 51 .75 52 .75 53 .75 N 53.50 54.50 55 .25 N Latitude Distance across eddy (km south to north) Figure 1-2. Contours of temperature (a), salinity (b), and sigma-#(c) across Haida-2000. Distance is given in km from south to north and in degrees of latitude except for September 2001, where the plot is a composite of two half-transects with distance given as south to centre and centre to southwest. The contouring intervals are 0.5°C for temperature, 0.1 units for salinity, and 0.2 units for sigma-0. Arrows along the tops of the panels indicate the location of eddy edges delimiting the extent of Haida-2000 (marked 'eddy'). 42 breaking away from the seamount, Haida-2000 tracked north westward approximately 325 km from September 2000 to September 2001. By our final survey in September 2001, Haida-2000 had drifted 750 km away from the coast and was sitting in High Nitrate, Low Chlorophyll waters. As Haida-2000 traveled westward away from the coast, it moved first into seasonally nitrate depleted waters designated as the Dilute Domain (DIL), and then into H N L C waters of the Subarctic Current domain [SUB; see Wong et al. (2002a); Fig. 1-1]. By June 2001, isopleths of temperature, salinity, and nitrate had shoaled in comparison with observations from Year 1 (June and September 2000), likely reflecting eddy decay (see section 1.3.1.3). Between June and September 2001 a second eddy merged with Haida-2000, injecting fresh, warm water into the eddy core. This coalescence was evident in an increased satellite sea surface height anomaly (SSHA) (Fig. 1-4), by an influx of heat and fresh water (Fig. 1-2; Crawford, in press), and by a depression of isopycnals and nutrient isopleths between sampling dates in June and September 2001 (Fig. 1-3). Since the merging of the two eddies was likely a slow process taking weeks to months, the shoaling and distortion of deep isopycnals (27.0, 27.2) and nutrient isopleths may also have resulted from the coalescence; however, it is impossible to isolate the effects of eddy coalescence from frictional decay during the June sampling period. 1.3.1.2 Temperature-salinity characteristics of eddy waters Haida-2000 was both warmer and fresher than surrounding waters (Fig. l-2a, b). As the eddy aged, the temperature anomaly in surface waters weakened (l-2a), but the isopycnal tilting and salt deficit remained. Contours of salinity showed that the surface waters of the eddy remained isolated from surroundings in both June and September 2000, but that in 2001 these waters exchanged freely with surroundings (Fig. 1 -2b). Compared to the reference sites, 43 temperature-salinity relationships within Haida-2000 changed less over the 20-month time period (not shown), illustrating that the eddy core volume remained relatively isolated from surroundings for depths below the mixed layer. Lat i tude (degrees North) 51.75 52.75 53.75 N 5 J .74 j^J? |^ 5 ___^ 5 3 : ! 3 T r r * r ^ i i ^ ; ! - • • • 9 . » . • • ~m 'fem^pS* • • • •-- v - — ° ~'/ Jf"*-=--30—-— • • • » - • • * • • • • - - • • • • 1 • • • 1 ---&.2-- • • • • 1 r * *i *i ^ c 1 1 1 1 53.50 54.50 55.25 N 50 100 150 200 0 50 100 150 200 50 100 150 200 250 50 100 150 200 250 Distance across eddy (km south to north) Figure 1-3. Solid contours show nitrate (uM) across the Haida-2000 eddy (panels a -d) with dashed contours representing sigma-c?. Four panels represent, from left to right, cruises in June 2000 (a), September 2000 (b), June 2001 (c), and September 2001 (d). The contouring interval is 5 uM for nitrate, 0.2 units for sigma-#. Contour constructed for September represents a composite of two half-transects from outside to centre (south to centre and centre to southwest) rather than one full transect in a north to south direction. Dots on each panel represent sampling locations and depths. 1.3.1.3 Eddy decay The decay of Haida-2000 was slow between February 2000 and February 2001, as inferred by only slight changes in the depth of isopycnal depression (Fig. 1 -2) and small decreases in sea surface height anomaly (SSHA; Table 1-1). From February 2001 to June 2001, however, the depth of the deepest isopycnal defining the eddy core (27.0) shoaled by approximately 100 m (ca. 0.89 m d"1; Table 1-1) and decay was assumed to be more rapid, although, as noted above, eddy coalescence could have also contributed to this occurrence. At 44 the same time the SSH anomaly was the smallest in June 2001 (Table 1-1). Between June and September 2001 an increase of 100 m in the depth of the depression of the 27.0 isopycnal (to ca. 600 m) was observed, following completion of the merging of the second eddy (Fig. 1 -2). Likewise, following the merging event the height anomaly increased from 5 to 10 cm (Table 1-1) . The changes in sea level anomaly should be interpreted with some caution since they are calculated relative to the surroundings; along Line P (Fig. 1-1) isopycnals and sea level tilt downward from inshore toward the centre of the Alaska Gyre (see Freeland and Whitney, 1999). This could cause an apparent increase in SSH of eddy waters as they drift westward that was larger than the actual change due to eddy rotation since the background isopycnals were not flat across the eddy trajectory. 1.3.2 Surface processes: Evolution of mixed layer nutrient concentrations Year 1 When Haida-2000 eddy broke away from the coast, the mixed layer depth at the centre was 50 m greater than in surrounding waters (Table 1-1). Table 2 shows nitrate and silicic acid concentrations (uM) at the surface (0-2 m), at the approximate depth of the mixed layer (75 m for winter, 50 m for spring and summer), at 75 m (depth of the top of the eddy core), at 300 m (within the eddy core) and at the bottom of the eddy (600 m in Year 1, 500 m in Year 2). Depth-averaged mixed layer concentrations of nitrate and silicic acid were 14.8 and 24.8 uM, respectively, at the eddy centre compared to 10.3 and 18.0 uM in outside waters (Table 1-2) . While surface nutrient concentrations (0 m; both nitrate and silicic acid) were lower in outside waters compared to the eddy centre and edges, values near the bottom of the mixed layer (75 m) differed only marginally between sites in February 2000 for nitrate (Table 1-2). Silicic acid concentrations were highest at the bottom of the mixed layer at the eddy centre and edges in TOPEX/ERS-2 Analysis Jun 6 2001 T O P E X / E R S - 2 Analysis Jun 18 2001 T O P E X / E R S - 2 Analysis Jun 27 2001 Sea surface height (cm) Sea surface height (cm) Sea surface height (cm) Figure 1-4. Three consecutive Sea Surface Height Anomaly plots from TOPEX/POSEIDON-ERS-2 radar altimetry showing the merging of a second, younger eddy with Haida-2000. 4-The left panel shows two separate eddies at 54°N,138°W and 52°N and 138 °W. The right panel shows two merged eddies at 54°N and 137°W. 47 February 2000 (Table 1-2; 25.1, 23.7, and 24.4 um at the centre and east/north and west/south edges, respectively), while the outside reference site had a silicic acid concentration that was slightly lower than at the centre (21.6 uM). After the first spring, mixed layer nutrient concentrations inside the eddy were at all times lower than initial values (Table 1 -2). The greatest losses in mixed layer nutrients occurred between the two initial observations (February and June; Table 1-3). Drawdown of nitrate was 2.6 umol month"', while removal of silicic acid was twice as high as nitrate (5.4 umol month"1). By June 2000 the mixed layer silicic acid concentration was only 3.0 uM within the eddy compared to 12.6 u.M outside (Table 1-2). While silicic acid concentrations were four times lower inside the eddy compared to surroundings, nitrate levels were similar (4.2 and 4.3 pM, inside and outside the eddy, respectively). Whereas within the eddy the decrease in silicic acid was twice as large as the loss of nitrate, the drawdown ratio of silicic acid to nitrate was 0.9 in outside waters (not shown; note that Reference stations were near the same location for June and September cruises, Table 1-1). The removal rates of silicic acid and nitrate were only 25 and 60% (1.4 u M month"1 Si(OH) 4, 1.5 u M month"1 NO3") as high, respectively, in surrounding waters compared to eddy waters (not shown). Over the first summer, the average surface Si(OH)4 concentration at the eddy centre within the mixed layer doubled from 3.0 to 6.0 uM, while the nitrate concentration decreased from 4.3 uM to below detection (removal of 0.9 u M month"1 NO3"). Although the lowest nitrate concentrations were found inside the eddy, surrounding surface waters in the DIL domain also were low in nitrate (< 1 uM; Fig. 1-3). Silicic acid was lowest at the eddy edges, with depressed Si(OH)4 isopleths apparent to a depth of ca. 200 m at the edge (Fig. 1-6) and 600 m at the centre (not shown). Table 1-2. Nitrate and silicic acid concentrations at selected depths for centre, edges, and outside reference stations for all observation sets (pM). Values for northern and southern edges are given for all cruises except February 2000 where edges were located to the east and west of the eddy centre; n.d. = no data available. Analytical precision was ±0.13 p M for nitrate and ±0.76 p M for silicic acid. Based on duplicate chlorophyll a determinations in June 2000 and September 2001, the precision was ±0.02 pg L" 1. Date Depth Centre North (east) edge South (west) edge Reference OSP P4 N 0 3 Si(OH)4 Chl N 0 3 Si(OH)4 Chl N 0 3 Si(OH)4 Chl N 0 3 Si(OH)4 Chl Chl Chl Feb-00 0 75 14.6 14.9 24.7 25.1 0.28 10.7 15.2 18.9 23.7 0.56 13.9 14.3 25.4 0.39 24.4 9.0 14.0 16.5 21.6 0.68 150 23.4 35.3 20.4 30.4 21.9 34.1 30.2 45.3 300 34.5 56.5 35.2 62.0 34.1 55.3 35.0 59.7 600 43.0 95.3 43.5 101.5 42.7 94.0 .42.1 92.4 June-00 0 50 4.2 7.7 3.0 13.1 0.38 4.5 8.9 5.5 13.7 n.d. 4.45 7.68 6.9 0.28 13.1 4.19 10.6 12.6 16.8 75 10.5 15.9 12.3 17.7 10.53 15.9 17.8 27.1 300 35.1 58.4 35.4 59.0 36.0 62.1 38.1 68.5 600 43.3 99.7 43.3 96.2 43.32 99.7 44.4 107.8 Sept-00 0 30 b.d. 4.1 6.0 9.8 0.47 b.d. 0.1 3.4 3.5 0.55 0.6 1.1 5.1 0.41 5.4 0.6 0.9 7.4 7.6 0.53 0.66 50 12.1 17.7 8.5 12.3 7.4 12.1 8.2 12.9 75 14.5 21.1 12.3 16.3 10.7 15.0 15.2 20.0 300 34.0 55.4 36.3 60.9 36.5 62.9 39.4 71.6 500 38.7 71.8 41.8 85.8 42.0 90.1 43.9 98.1 Feb-01 0 75 13.9 15.4 16.4 17.7 0.29 13.9 n.d 16.3 n.d. 0.35 12.5. n.d. 16.8 0.30 n.d. 15.6 15.6 23.0 23.1 0.32 0.51 oo Table 1-2 cont'd. Date Depth Centre North (east) edge South (west) edge Reference OSP P 4 N0 3 Si(OH)4 Chl N0 3 Si(OH)4 Chl N0 3 Si(OH)4 Chl N0 3 Si(OH)4 Chl Chl Chl 100 25.0 32.3 n.d n.d. n.d. n.d. 27.3 43.5 300 37.0 58.6 n.d. n.d. n.d. n.d. 43.4 90.6 600 43.7 97.9 n.d n.d. n.d. n.d. 44.9 121.8 June-01 0 50 8.8 11.2 15.1 17.3 0.52 8.4 9.7 13.9 15.4 0.39 10.6 12.2 16.8 0.43 17.7 10.1 11.3 16.2 17.0 0.45 75 13.2 19.0 11.3 16.5 13.9 19.6 12.4 17.9 300 38.1 69.9 38.6 70.4 40.6 79.1 40.6 79.0 600 44.0 108.0 44.3 107.6 44.2 113.6 44.4 115.4 Sept-01 0 50 3.4 12.1 12.1 17.3 0.55 3.6 14.4 10.9 18.6 0.57 3.4 11.8 10.9 0.52 15.9 9.0 14.1 13.1 20.2 0.31 75 16.3 21.9 18.0 23.6 14.8 19.4 16.6 25.8 300 36.3 60.8 39.7 75.0 38.2 68.9 42.0 85.3 600 43.1 95.5 44.3 106.5 43.6 101.6 43.8 110.1 4^  50 Year 2 Winter ventilation (September 2000 - February 2001) injected nutrients into the mixed layer and nitrate values were returned to the initial average concentration. The average nitrate concentration was slightly lower than in the surrounding waters (compare 14.3 uM at the eddy centre with 15.6 uM outside; Table 1-2). However, eddy waters were not replenished in silicic acid by winter mixing, and mixed layer concentrations at the centre were lower than in surrounding waters in the SUB domain in February 2001. The mixed layer silicic acid concentration at the eddy centre in February 2001 was 16.7 uM compared to the initial value of 24.8 uM (Fig. 1-5). Surrounding waters held 22.9 u M Si(OH) 4 in February 2001 (Table 1-2). The rate of nitrate loss between February and June 2001 within the eddy mixed layer was 50% as high as the previous year (1.2 u M month"1 N O 3 " lost between Feb. and June 2001; Table 1-3). Silicic acid drawdown was only 4% of the first spring (0.2 u M month"1 Si(OH) 4 lost between Feb. and June 2001). By September 2001, there was little difference between mixed layer silicic acid concentrations inside or outside the eddy (11.8 u M Si(OH) 4 within the eddy, 12.9 u M Si(OH) 4 outside), although nitrate was 3.5 uM inside the eddy compared to 9.0 uM in surrounding waters (Table 1-2). 51 co O Z 1969 -1981 Monthly surface averages 20 O z 20 15 10 Haida mixed layer nutrients o 20 15 i 10 Year 1 ; o • NO/ —o— Si(OH)4 b • o . o Year 2 • : o o , o c • • 30 25 _> 20 O I 15 5 10 5 0 30 25 CO 20 O I 15 10 5 0 Figure 1-5. Monthly averages of nitrate and silicic acid concentrations at Ocean Station P (50°N, 145°W) from 1969 to 1981 (adapted from Whitney and Freeland, 1999) (a). Average mixed layer nutrient concentrations in Haida-2000 during Year 1 (b; Feb. 2000 - September 2000) and Year 2 (c) were computed from ambient concentrations from the surface to the bottom of the mixed layer. Eddy formation occurred in February 2000 (= 0 months) and the study ended in September 2001 (= 20 months). 52 Table 1-3. Change in average nutrient concentrations (|JM month"1) within the mixed layer over time at eddy centre between sampling dates. Signs associated with numbers indicate loss (negative) or gain (positive). Nutrient Feb-June June-Sept Sept-Feb Feb-June June-Sept 2000 2000 2001 2001 2001 N 0 3 " -2.6 -0.9 3.3 -1.2 -1.5 Si(OH) 4 -5.4 0.9 2.5 -0.2 -1.0 Absolute S i (OH ) 4 :N0 3 " 2.1 1 0.76 0.19 0.67 ratio 53 Lat i tude ( d e g r e e s Nor th) 51.75 52.75 53.75 N 53.50 54.50 55.25 N Distance ac ross eddy (km south to north) Latitude (degrees North) 51.75 52.75 53.75 N 53.50 54.50 55.25 N Distance across eddy (km south to north) Figure 1-6. Solid contours show silicic acid (|iM) across the Haida-2000 eddy (panels a -d) with dashed contours representing sigma-ft Four panels represent, from left to right, cruises in June 2000 (a), September 2000 (b), June 2001 (c), and September 2001 (d). The contouring interval is 5 uM for silicic acid, 0.2 units for sigma-0. Contours constructed for September represent a composite of two half-transects from outside to centre (south to centre and centre to southwest) rather than one full transect in a north to south direction. The minimum value at the surface in June 2000 was 3.0 \xM. Dots in each panel represent sampling locations and depths 54 1.3.3 Nutrient drawdown and new production The difference between February and June nitrate concentrations was used to estimate seasonal (spring) drawdown, when nitrate utilization is typically the highest (Wheeler, 1993; Varela and Harrison, 1999). Annual new production was taken as the drawdown occurring between February and September/October. Daily drawdown rates are reported for comparison with other studies, with the caveat that the bulk of nitrate utilization occurs between April and June (Wheeler, 1993; Whitney and Freeland, 1999). This represents a minimum drawdown estimate since inputs of nutrient (for example by wind mixing) are not accounted for. From losses of nitrate in the upper 50 m of the water column (see Table 1-5 for integrated nitrate and silicic acid inventories for 0 - 5 0 m), estimates of the minimum new production were derived for Haida-2000 (Table 1-6). Losses within the upper 50 m were calculated for time periods between February and June or February and September (Table 1-6). The former represents spring drawdown while the latter estimates yearly drawdown, assuming that the maximum NO3" concentration occurred in February and the minimum in September. Between February and June 2000 (ca. 120 d), 374.5 mmol m"2 of nitrate were removed from the mixed layer at the eddy core, equivalent to approximately 3.0 mmol m"2 d"1. Between February and September, 417.7 mmol m"2 of nitrate were removed from the mixed layer, resulting in a removal 2 1 rate of 3.5 mmol m" d" . In the second year of eddy evolution, nitrate drawdown was 193.4 mmol m" from -2 1 February to June, half as high as the previous year (removal rate of 1.6 mmol m" d" ). The annual nitrate removal (February to September 2001) was 417.1 mmol m"2, not different from the nitrate loss that occurred in the mixed layer between February and September 2000 in the first 2 1 year of eddy life (ca. 3.5 mmol m" d" ). Thus, the only difference in nitrate losses between young and older eddy waters was a difference in the timing of nitrate drawdown. 55 2 Table 1-4. Nutrient inventories (mmol m") integrated over 0-50 m in February, June, and September of 2000 and 2001 at the centre of Haida-2000. Nutrient Date 2000 2001 Feb Jun Sept Feb Jun Sept N 0 3 " 600.5 226.0 182.8 672.5 479.1 255.4 Si(OH)4 1003.4 161.7 462.4 791.0 785.0 632.5 Table 1-5. Estimates of annual and spring new production (NP) and Si(OH) 4 drawdown (mmol m"2 d"1) at the eddy centre from 0-50 m in 2000 and 2001. Annual NP was calculated assuming maximum nutrient inventories in February and minima in September, while spring new production captures nitrate-based growth between February and June. Month Spring Annual Spring Annual NP NP Si(OH)4 Si(OH)4 (mmol N m 2 d"1) (mmol N m 2 d"1) (mmol ni 2 d"1) (mmol m 2 d 1) Y e a r l 3.0 3.5 7.0 4.5 Year 2 1.6 3.5 0.05 1.3 Examining the losses of silicic acid in order to estimate diatom production, there were large differences between Year 1 and Year 2 (see Table 1-5 for silicic acid inventories from 0 -50 m; see Table 6 for removal rates).'There was a spring bloom of diatoms (7.0 mmol Si(OH) 4 m"2 d~' removed) in the spring of 2000 that was followed by little to no silicic acid drawdown over the summer months. In fact, silicic acid increased within the mixed layer, with mixed layer 56 Si(OH)4 (0 - 50 m) tripling. The annual drawdown was 4.5 mmol Si(OH) 4 m"z <T in the first year. Although growth was much slower in Year 2 (1.3 mmol m"2 d"1 removed from February to September), it continued throughout the summer. The spring drawdown was only 0.05 mmol m" d"1, or 4% of the annual drawdown in Year 2. 1.3.4 Deep-water processes: evolution of eddy core waters 1.3.4.1 Nutrients within the eddy core Contours of nitrate (Fig. 1-3) and silicic acid (not shown) exhibit depressed isopleths to. depths that corresponded approximately with the base of the eddy core (ca. 600 m), demonstrating the presence of anomalous coastal-type water. Phosphate distributions followed those of nitrate (Table 1-4; contours not shown), and so are not discussed further. Nutrient concentrations within the eddy core changed little over the study period (see Table 1-2 for concentrations at selected depths). Nutrient-salinity relationships also remained similar throughout the 20-month period (Fig. 1-7), and lay between typical oceanic (e.g. at Ocean Station P) and coastal values (e.g. P4) for both nitrate and silicic acid. Differences between eddy waters and coastal/oceanic waters were greater for silicic acid than for nitrate, with eddy nitrate-salinity relationship being similar to oceanic waters for salinities less than ca. 33.5 (Fig. 1-7). Silicic acid per unit salinity remained lower within the core of Haida-2000 compared to oceanic H N L C waters throughout the study. Table 1-6. Integrated concentrations of nitrate, phosphate and silicic acid (mol m"2) and Si(OH)4:N03 _ molar ratios at the eddy core and at outside reference stations. Integration depth corresponds to the eddy core, from the depth of flat isopycnals at the top {ca. 75 m) to the 27.0 <5e isopycnal at the bottom. In February 2000 the true centre was not sampled; the numbers represent a station closer to the edge. Numbers in brackets represent the estimated nutrient concentrations at the eddy centre in February 2000 calculated from the enrichment factor (centre value / edge value) from June and September 2000. Analytical precision was ±0.065 mol m"2 for nitrate and ±0.38 mol m"2 for silicic acid. N O ? HPO4 2 " Si(OH) 4 Si:N:P Si-N Date In Out In Out In Out In Out In Out Feb 2000 17.8 18.4 1.30 1.32 32.4 34.7 25:14:1 26:14:1 1.82 1.88 (20.1-21.4) (1.47-1.56) (36.6-38.9) June 2000 21.2 19.5 1.38 1.41 38.6 37.8 28:15:1 27:14:1 1.82 1.94 Sept 2000 17.5 15.7 1.28 1.13 30.6 29.0 24:14:1 26:14:1 1,75 1.85 Feb 2001 19.2 12.8 1.40 0.92 * 33.4 25.6 24:14:1 24:14:1 1.74 2.0 June 2001 15.1 11.7 1.08 0.84 28.0 21.6 26:14:1 26:14:1 1.85 1.85 Sept 2001 18.9 11.7 1.37 0.85 34.5 22.5 25:14:1 26:14:1 1.83 1.92 58 3 100 80 60 O 40 (75 20 0 40 g 30 'd 2 0 Z 10 -•— Feb - 00 -•—Jun-00 Sept - 00 Feb - 01 Jun - 01 Sept - 01 P4 coastal OSP oceanic coastal 32.0 32.5 33.0 33.5 Salinity 34.0 Figure 1-7. Silicic acid (a) and nitrate (b) versus salinity at the eddy centre for February, June, and September 2000 and 2001 at the eddy centre. A coastal upwelling station (P4) and an oceanic station (OSP) sampled in June 2000 are included for reference. The shoaling of deep isopycnals over time resulted in smaller integrated nutrient inventories (Table 1-4), and higher nutrient concentrations for a given depth (Fig. 1-8). The eddy held significantly more nitrate (p = 0.04) and silicic acid (p = 0.04) between 75 m and the 27.0 isopycnal compared to the surroundings when all time points were considered. June 2001 saw the most dramatic shoaling of deep isopleths, which resulted in smaller depth-integrated 59 0 40 80 120 160 1 0 2 0 3 0 4 0 5 0 S i ( O H ) 4 ( . iM) N0 3 " (u.M) Figure 1-8. Silicic acid (a) and nitrate (b) profiles for six cruises conducted from February 2000 to September 2001 at the centre of Haida-2000. The surface concentrations of Si(OH)4 increase with age or distance from the point of origin, becoming more similar to surrounding High Nitrate Low Chlorophyll waters. A coastal upwelling station (P4; see Fig. 1-1) and a station located in High Nitrate Low Chlorophyll waters (OSP) are included for reference. inventories of nitrate and silicic acid between 75 m and the 27.0 isopycnal. Although the concentrations per volume were lower than the surroundings when the eddy sat in High Nitrate Low Chlorophyll waters (see Table 1-2), the volume of water contained between the mixed layer and the 27.0 isopycnal was greater within Haida-2000 compared to the surroundings; in nearby non-eddy waters the 27.0 isopycnal sat at a depth of 300 m, while at the eddy centre it was depressed to 500-600 m. Thus, despite the fact that the concentrations of macronutrients were 60 lower within Haida-2000, the total integrated inventory was still higher than in the surroundings (Table 1-6), and offshore nutrient transport was substantial. In June 2001, when isopycnals and nutrient isopleths within the eddy shoaled, nutrient profiles exhibited higher concentrations at depths below ca. 300 m compared to the other sampling periods (Fig. 1 -8). The trend was more apparent in the silicic acid profiles than for nitrate where concentrations in June 2001 were only slightly higher. The shoaling of deep isopycnals injected nutrients into the eddy core, enriching these waters, particularly in silicic acid. Nutrient increases within the eddy core occurred at a Si(OH)4:N03~ molar ratio of ca. 2.1. 1.3.4.2 Influence of eddy deformation and merging on nutrient distributions The shape of Haida-2000 was altered as it encountered physical obstacles and interacted with other eddies. When the centre sat directly over Bowie Seamount in June and September 2000, the northern edge was distorted, and the eddy shape was more oblong compared to other sampling times. The changes in eddy shape are reflected in asymmetric rather than Gaussian contours of sigma-#and nutrient distributions (Fig. 1-2, 1-3). Nutrient isopleths generally followed isopycnals (Fig. 1 -3). Once the eddy broke away from the seamount it regained a more circular shape (e.g. in February and June 2001), which was stretched in a north-south direction during the coalescence of a second younger eddy between June and September 2001 (Table 1-1). 1.4 Discussion 1.4.1 Biological drawdown of eddy nutrients In February 2000, a mesoscale eddy (Haida-2000) detached from the British Columbia coast and began traveling westward, first moving through waters designated as the Dilute Domain (DIL) and then into High Nitrate Low Chlorophyll waters of the Subarctic Current 61 System (SUB, Fig. 1-1; (Favorite etal, 1976; Whitney and Freeland, 1999; Wong etal, 2002a). This eddy was followed for nearly two years as it drifted westward. For a description of the history and physical features of Haida-2000,. see Miller et al. (in press), Crawford (2002), and Yelland and Crawford (in press). In the first year, surface waters within Haida-2000 held low nitrate concentrations, similar to waters in the DIL Domain. Silicic acid, however, was much lower within the eddy compared to the surroundings. By the second year, surface waters of Haida-2000 were similar to those of the H N L C domain characterized by high nitrate and silicic acid concentrations. The very low levels of nitrate, phosphate, and silicic acid in the mixed layer within the Haida-2000 eddy followed an April spring bloom observed by SeaWiFS satellite imagery (images prepared by J.F.R. Gower), indicating that nutrient losses were due to biological consumption. The satellite pictures showed that chlorophyll within the Haida-2000 eddy was ca. 3-10 pg L~', or approximately 10 times higher than surrounding levels in April. Silicic acid drawdown between February and June 2000 exceeded nitrate drawdown with a ratio of approximately 2:1, implying that diatoms were responsible for the high chlorophyll production. It was during the initial spring period that Haida-2000 acted as a important sink for carbon dioxide, exhibiting a significantly higher drawdown of CO2 compared to the surroundings (Chierici et al, in press). This role diminished during later eddy evolution. New production refers to primary production fueled by "new" nutrients from deep upwelled water, in nitrogen currency (Dugdale and Goering, 1967). The disappearance of nitrate from surface waters on a seasonal or annual basis approximates new production. New production differs from export production (the removal of particulate carbon produced by phytoplankton from the upper ocean), because remineralization processes can lead to the recycling of particulate material that may have accumulated in the mixed layer, preventing particle export. 62 The question of export production is not addressed in this chapter (see Chierici et al, in press), although the disappearance of silicic acid suggests that significant export occurred following the spring bloom in April 2000. Neglecting the inputs of nitrate from nitrogen fixation or nitrification, annual and spring new production rates were estimated from the difference in nitrate inventory between 0-50 m from February to September, and February to June, respectively. The 0-50 m layer represents the average depth of the euphotic zone in the Subarctic North Pacific (Longhurst et al, 1995; Boyd and Harrison, 1999; Harrison et al, 1999) and allows a comparison of our estimates with the larger collection of observations from the Subarctic Pacific (e.g. Wheeler, 1993; Wong et al, 2002a; Childers and Whitledge, in press). Estimates for new production in the Gulf of Alaska come from studies at Ocean Station P (Wheeler, 1993; Varela and Harrison, 1999), from ship-of-opportunity cruises in the Alaska Gyre (Wong et al. 2002a), and from the US GLOBEC Seward Line time series (Childers and Whitledge, 2005). It was hoped that by focusing on an isolated parcel of water held within an eddy the complicating effects of advection would be minimized (see Wheeler, 1993). Although there was some exchange with outside waters, the surface waters remained relatively isolated from the surroundings, at least within the first year (see Yelland and Crawford, in press; Chierici et al, in press). Annual new production estimates derived from differences in maximum (winter) and minimum (summer) mixed-layer integrated nutrient content suggested that new production was similar in the first year of eddy life compared to Year 2 (Table 1-6). Spring new production (February - June) was twice as high in year one compared to Year 2, reflecting a difference in the timing of nitrate drawdown rather than a difference in the magnitude of nutrient utilization. As the eddy aged, not only did the drawdown ratios of Si(OH)4: N O 3 " decrease, but the timing of maximum nutrient drawdown shifted from spring to summer (Table 1-3; Fig. 1-5). At OSP, the 63 maximum drawdown rates of silicic acid occur in May-June, while nitrate concentrations reach a minimum in July (Whitney and Freeland, 1999, Fig. l-5a). The shift in timing of the nitrate minimum from.year one to year two (from June to September) reflects the evolution of eddy waters toward H N L C conditions characterized by maximum drawdown rates that occur in late summer. It is possible that the gain in silicic acid observed over the spring - summer period in 2000 could have been achieved by lateral advection induced by Ekman transport, or the onset of winter mixing. However, the salinity differences between June and September at the eddy centre were very small (detail not shown) except at a depth of 50 m, which was below the summer seasonal pycnocline (ca. 30 m). It is thus more likely that the silicic acid came from below, and represented either the onset of winter mixing or a faster rate of eddy diffusion due to weaker stratification at the eddy centre. However, estimates of the buoyancy frequency across the mixed layer were similar at the centre and outside stations (Brunt-Vaisala frequency, N, as were estimates of potential energy required to homogenize the mixed layer [see calculations in Nelson et al. (1989); data not shown], suggesting that any difference between stratification in eddy versus outside waters above 50 m was small. The gain in silicic acid without an apparent increase in nitrate could be explained by the fact that the latter was probably also supplied but was not detected due to biological removal. Given that the 50 m integrated silicic acid inventory increased over the first summer by approximately 3-fold, suggests that new production rates for this time period were underestimated. Based on silicic acid supply, nitrate uptake could have 2 1 been as high as 10 mmol N m" d" . Corresponding chlorophyll a concentrations in September were higher than in June, particularly at the eddy edges (Table 1-2; also see Crawford et ah, in press). 64 An extensive survey of surface nutrient dynamics in the Subarctic Pacific by (Wong et al, 2002b) on a ship-of-opportunity from 1995 - 2000 included regions that corresponded to the location of the Haida-2000 eddy. They estimated annual new production rates of 0.98 mmol N m"2 d"1 for the DIL domain, 1.0 mmol N m"2 d"1 for the SUB domain, and 1.5 mmol m"2 d"1 for the A G , which includes OSP. The latter included the year 2000, which exhibited one of the 2 1 highest rates of new production measured in the region (2.5 mmol N m" d" ) during a period when an old eddy may have transported iron to this region (Whitney et al, in press); without this -2 -1 value the average decreased to 1.3 mmol N m" d" . Our estimates for annual new production N P a of 3.5 mmol m"2 d"1 (both years) were thus higher than the regional averages, even considering the anomalously high removal rate noted in 2000 for the Alaska Gyre. The Haida eddy estimates were at the low end of the range of values presented by Childers and Whitledge (2.1 - 17.0 2 1 mmol m" d" ; in press) for northern coastal waters in the Gulf of Alaska, were slightly higher than values estimated from an 15N-tracer study reported by Varela and Harrison (1999) for OSP 9 1 (pooled seasonal average of 2.6 mmol m" d" ), and fell within the range of spring-summer pooled average values reported by Wheeler (1993) for an earlier l5N-tracer study at OSP (2.3 -4.3 mmol m"2 d"1. The most remarkable difference in nutrient utilization and inventories within Haida-2000 was the large drawdown of silicic acid in the spring of Year 1 and near absent drawdown at all time points thereafter. Unfortunately, our sampling program began too late in 2000 to catch the bloom and make in situ measurements, but the Si(OH)4 disappearance suggests that diatoms were an important component of the phytoplankton assemblage (also see Batten and Crawford, in press). Since diatoms play an important role in vertical carbon flux in the Gulf of Alaska (Wong et al, 2002a), it is important to note the influences on their distributions. 65 Diatom growth is limited in the subarctic northeast Pacific by the lack of iron (Martin and Fitzwater, 1988; Boyd et al, 1996; Maldonado et al, 1999; Boyd et al, 2004). While macronutrient concentrations were slightly higher in waters outside of the eddy, Fe (Johnson et al, in press) and other trace metals (Zn, Lohan pers. comm.; Cd, A l , Ga, Mn, Crispo and Orians pers. comm.) were initially enriched within eddy waters. It is likely that the presence of higher concentrations of these trace elements enhanced primary productivity within anticyclonic eddies and promoted growth of diatoms in particular. Wong et al. (2002b) and Whitney et al. (in press) noted that the low Si(OH) 4 supply to the eastern subarctic Pacific may be responsible, in combination with low Fe concentrations, for the H N L C character of the Alaska Gyre. Wong et al. (2002b) suggest that diatoms in the DIL, SUB, and A G regions have high silicic acid requirements compared to nitrate. They estimated that silicic acid was required in a ratio of 2.7:1 Si(OH)4:N03~, based on species assemblage data combined with bulk nutrient removal rates throughout the eastern Subarctic Pacific. This may reflect the higher Si(OH)4:N03~ uptake ratios that result from Fe limitation in diatoms (Hutchins and Bruland, 1998; Takeda, 1998), or simply differences in the geochemical cycling of these two nutrients as nitrate is recycled more rapidly than is silicon (e.g. Whitney et al, in press). The substrate affinity constant in an assemblage of diatoms at Ocean Station P under iron replete conditions was demonstrated to be ca. 1.6 u M (Appendix A), suggesting that diatoms would not experience growth limitation by silicic acid at the concentrations observed in this study. However, different diatom species likely have different substrate requirements, particularly when coastal and oceanic species are compared (Sunda et al, 1991). The estimate of Si(OH)4:N03~ drawdown for the period encompassing the spring bloom was 2.3, close to the estimate by Wong et al. (2002b) for diatom requirements in the N E Pacific. The substrate affinity constant, a concept borrowed from Michaelis-Menten enzyme kinetics, is an estimate of the ability of an organism (or specifically, an enzyme) to take up a nutrient when it is present in low concentrations 66 As Haida-2000 aged and evolved, its winter mixed layer Si(OH)4:N03~ molar ratio decreased from 1.7 to 1.2, below the proposed requirements of diatoms for silicic acid compared to nitrate by Wong et al. (2002b). Large diatoms were rare in eddy waters at all sampling points following the April bloom (Chapter 2), resulting in the lower Si(OH)4 drawdown rates observed within eddy waters after Year 1. This likely reflected the loss of Fe that occurred as Haida-2000 aged (Johnson et al., in press). The spring Si(OH)4:N03~ drawdown ratio of 2.1 observed in this study (Table 3) was similar to that reported by Whitney and Robert (2002) for the larger Haida-1998 eddy that tracked into more southerly waters (2 - 3). At all points after the first spring period, nutrient ratios were less than or equal to 1. The greater than 2:1 drawdown ratio may reflect differences in recycling processes between nitrate and silicic acid that include the preferential remineralization of nitrate at shallower depths than silicic acid (Whitney et al., in press), the low nitrate environment of the DIL domain that would drive Si(OH) 4: N O 3 " uptake rates in excess of 1:1 (Whitney and Freeland, 1999), or a high silicic acid requirement by diatoms in the DIL domain (Wong et al., 2002b). The spring drawdown of nutrients in this study was accompanied by high chlorophyll a biomass, suggesting either the first or second explanation. 1.4.2 Physical processes influencing nutrient distributions Coastal waters off British Columbia are characterized by high nutrients and phytoplankton biomass nearshore with decreasing concentrations of each moving across the continental shelf break (Mackas and Yelland, 1999). An increasing gradient in macronutrients from the vicinity of the Alaska Current toward H N L C waters (Whitney and Freeland, 1999) occurs concomitantly with a decrease in iron concentrations (Nishioka et al., 2001). As noted 67 above, in H N L C waters the availability of Fe is important in regulating phytoplankton production and community structure (Harrison et al, 1999; Harrison, 2002). Haida eddies are formed in Dilute Domain coastal waters characterized by macronutrient concentrations that are lower than in the H N L C Domain (Whitney and Robert, 2002; Whitney and Welch, 2002), but which have a higher trace metal content (Nishioka et al, 2001). Haida-2000 was more important in delivering trace metals such as iron rather than macronutrients, as was the case for Haida-1998 which traveled into DIL domain waters (see Whitney and Robert, 2002). This illustrates the very different roles, in terms of nutrient transport, that Haida eddies can play depending on which path they follow. At the time of formation, the silicic acid and nitrate contents within Haida-2000 were higher per unit salinity compared to coastal waters, but were lower with respect to H N L C stations (e.g. P26 or OSP; Fig. 1-7). As the eddy drifted further offshore, it appeared increasingly poor in nutrients relative to surroundings, reflecting the coast-to-offshore nutrient gradient. Following the initial sampling period in February 2000, integrated nutrient inventories (between 25.0 and 27.0 isopycnals) were higher at the eddy centre compared to outside waters (Table 1-4). This reflects, in part, an artefact of our method of integration, since eddy decay led to a shoaling of the eddy bottom which decreased the depth over which nutrients were integrated. However, it is a functional definition since the integrations were performed according to isopycnal depth and thus reflect total inventory along density planes. It is possible that frictional decay and shoaling of the eddy bottom led to lateral outflow and spreading close to the surface. In such a case, the estimates of nutrient loss might have simply reflected the bias introduced by the differences in integration depth, despite the fact that total nutrient content within the eddy volume might not have changed. 68 1.4.2.1 Eddy decay and secondary circulation Haida-2000 underwent slow decay from September 2000, when Haida-2000 broke away from Bowie Seamount, to June 2001. During this period isopycnal rebound within Haida-2000 occurred at a rate of approximately 0.9 m d"1, similar to reports by Cheney and Richardson (1976) and Olson et al. (1985); 1 m d"1) for.Gulf Stream cold core and warm core rings, respectively. Deep nutrient concentrations at specific depths within the eddy increased over time. The shoaling of isopycnals due to frictional decay and eddy spin down were likely responsible for the increase in nutrient concentrations within the core of Haida-2000, similar to observations by Olson et al. (1985) for warm core rings of the Gulf Stream. The nutrient gains occurred at a Si(0H)4:N03 _ ratio greater than 2, suggesting that deep waters high in nutrients were being entrained into the eddy core. Changes in nutrient content over time within the eddy were not of a similar magnitude or sign as reference sites, indicating that the changes observed in nutrient content within Haida-2000 did not merely reflect the increasing distance from the coast, but rather reflected the influences of different physical processes such as eddy coalescence, frictional decay, or biological processes such as phytoplankton growth and zooplankton grazing. Models of the decay of a Gulf Stream warm core ring showed that changes in available potential energy (APE) created by frictional decay can lead to modifications in azimuthal velocity that produce horizontal currents directed inward near the surface and in deeper waters, with an outward flow at mid-depths and an upward flow at the centre (Flierl and Mied, 1985; Franks et al., 1986; Fukumori, 1992). These frictionally-induced modifications of the density and nutrient fields led to slow upwelling at the centre of Gulf Stream warm-core ring 82-B, which facilitated wind-induced mixing in the surface layer and led to an increase in nutrient availability at the ring centre (Nelson et al., 1989). Reminiscent of those responses to frictional decay, contours of sigma-tfand nutrients domed at the centre of Haida-2000 between 75 - 100 m 69 depths, particularly in September 2000 and 2001. The density structure could have exhibited the effect of frictionally-induced slow upwelling at the centre of the eddy; alternatively, it could have represented the onset of winter ventilation at the eddy centre in September, the interleaving of surrounding waters at the eddy boundaries (Simpson, 1984), or ageostrophic circulation patterns within the eddy (Martin and Richards, 2001). Although there are many differences between Gulf Stream warm core rings and Haida eddies, it is possible that some of the changes in structure of the nutrient and density fields observed during the decay of Haida-2000 were due to similar processes of decay-enhanced upwelling at the eddy centre. The effects of frictional decay and the enhanced nutrient availability on phytoplankton growth have been modeled for Gulf Stream warm-core rings (Franks et al, 1986), and it was found that a chlorophyll maximum developed as a ring decayed. The shift in the chlorophyll a maximum from the edge to the centre in Haida-2000 between September 2000 and June 2001 may reflect a response to decay processes, although sparse data points allow for speculation only. The high chlorophyll a band observed by Crawford et al. (in press) lay along the 25.4 isopycnal at the eddy centre and edges; this isopycnal outcropped in the surface waters outside the eddy, approximately 81 km from the eddy centre. Simpson (1984) noted that the water masses in frontal zones around eddies tend to be highly diffusive in nature and that in some systems, for example in eddies of the California Current System (CCS) and the East Australia Current [EAC; (Scott, 1981)], high density waters can be entrained into the eddy along isopycnal surfaces and directed toward the centre. The band of high chlorophyll a may have resulted from an enrichment of phytoplankton growth in surface waters at the frontal boundary. Martin and Richards (2001) noted that ageostrophic circulation within eddies can induce large vertical velocities that could enrich surface waters with nutrient from below. Yelland and Crawford (in press) discussed several factors that contribute to the creation of ageostrophic 70 currents within Haida eddies, including interaction with seamounts, inertial currents, and deep barotropic currents. It is likely that ageostrophic effects contributed to nutrient vertical and horizontal fluxes within Haida eddies, but these effects have yet to be quantified. 1.4.2.2 Overwash and surface dilution Mackas et al. (in press) noted the importance of rapid Ekman transport events where waters within a shallow mixed layer are laterally advected by strong wind shear. Tranter et al. (1982) discussed the relationship of this phenomenon with eddies of the East Australia Current and used the term overwash to describe the replacement of the upper layer of a ring or eddy by surrounding waters (also see Evans et al., 1985). Overwash appears to be an important process in rings during their interaction with their parent current (e.g. Evans et al., 1985), while in anticyclonic eddies that do not possess streamers (e.g. Haida eddies, CCS and E A C eddies) lateral advection of outside surface waters is likely a more important process. Despite the occurrence of lateral advection in Haida eddies, surface salinity and mixed layer nutrient concentrations above the eddy were different than in surrounding waters during the first year of the study, suggesting that waters at the eddy centre can remain isolated from surroundings for several months at a time (also see Yelland and Crawford, in press). Periods of isolation appear to be long enough to promote differences in nutrient drawdown between eddy and non-eddy waters. 1.4.2.3 Eddy deformation and merging In addition to nutrients supplied by frictionally-induced changes in flow and entrainment, interactions with bathymetric features such as seamounts and vortex-vortex interactions (e.g. coalescence events) also modified nutrient distributions. Vortex-vortex interactions are well documented in mesoscale eddies. For example in the Gulf Stream warm core rings become 71 absorbed by the strong Gulf Stream current (Evans et al, 1985), and in the Gulf of Mexico cyclonic vortices spin off anticyclonic eddies as they migrate westward (Forristall et al, 1992). Rapid changes in eddy structure induced by external influences can lead to major changes in available potential energy, potential vorticity, and biological community composition in warm core rings of the Gulf Stream (Joyce et al, 1984). In. the case of Haida eddies, the interaction with seamounts did not cause a break-up of the eddy, but it did appear to influence nutrient distributions. It was most likely the interaction with Bowie Seamount that distorted the isopycnals (Fig. 1-3) and corresponding nutrient distributions at the northern edge of Haida-2000; these influences are particularly evident in September 2000 where nutrient isopleths and isopycnals are uplifted near the northern edge (Fig. 1-3). A shoaling of the nutrient isopleths at the eddy centre in June 2000 at approximately 200 m may have reflected the eddy's interaction with the shallow Bowie Seamount. The contribution of upwelling processes described at eddy edges by Okkonen et al (2003) for eddies of the Alaska Coastal Current in the northern Gulf of Alaska may also have contributed to the distortion of the isopleths. The coalescence of a younger eddy led to a renewal of coastal nutrient characteristics within Haida-2000. This included a depression of deep isopycnals, a 0.1 unit decrease in core salinity, stretching of the eddy shape, a temperature increase of 0.7°C, and a decrease in deep nutrient content. Iron concentrations were two times higher in September compared to June 2001 in the upper waters of Haida-2000 (75 -200 m), indicating that iron was also delivered by the eddy merging event (Johnson et al, in press). Such events appear to be important in restoring coastal characteristics and extending the range of coastal type waters further from the continental margin. 1 72 .4.2.4 Nutrient Supply Compared to H N L C regions, the surface and core waters of Haida eddies possessed low nitrate and s