UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Physiological control of diatom sedimentation Waite, Anya Mary 1992

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

Item Metadata

Download

Media
831-ubc_1992_spring_waite_anya_m.pdf [ 5.88MB ]
Metadata
JSON: 831-1.0098886.json
JSON-LD: 831-1.0098886-ld.json
RDF/XML (Pretty): 831-1.0098886-rdf.xml
RDF/JSON: 831-1.0098886-rdf.json
Turtle: 831-1.0098886-turtle.txt
N-Triples: 831-1.0098886-rdf-ntriples.txt
Original Record: 831-1.0098886-source.json
Full Text
831-1.0098886-fulltext.txt
Citation
831-1.0098886.ris

Full Text

PHYSIOLOGICAL CONTROL OF DIATOM SEDIMENTATION by ANYAM.WAITE BSc Hon. Biology, Dalhousie University, 1985 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Doctor of Philosophy in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF OCEANOGRAPHY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April, 1992 © Anya Waite 1992 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of dPCtM/V&fidftt Cf The University of British Columbia Vancouver, Canada JMiLiir Date jwptj #r>, / / y /_ DE-6 (2/88) ii ABSTRACT The specific physiological mechanisms governing the process of diatom sedimentation were not well understood. This study investigated the possibility of a short term energy requirement for buoyancy maintenance in marine diatoms which varied predictably with physiological state. The maximum sinking rate (MSR) of Ditylum brightwellii was measured under progressive energy limitation which allowed the determination of the energy-dependent potential for a cell to reduce its MSR. There was systematic experimental evidence that this diatom's ability to gain energy from storage products via respiration is the principal determinant of its sinking rate. This study provides the first quantification of the negative exponential relationship between respiration rate and sinking rates. Preliminary studies with two other species indicate that a spectrum of energy-dependence may exist. Diatom blooms are the annual periods of highest "new" production and carbon sedimentation. Results from a 5-year study in Auke Bay, Alaska showed that termination of the spring bloom consistently occurred at limiting nitrate concentrations. The sinking response of diatoms to ambient nutrients influenced both species succession during the spring bloom and the subsequent sedimentation of new production. Threshold nitrate concentrations approximating Ks values of the species present, were found to signal initiation of increased sedimentation. Results suggested genus-specific differences in sinking-rate sensitivity to nitrate exhaustion. Overall, sinking rates of the three principal genera ranked (high to low) Thalassiosira spp.~> S. costatum --> Chaetoceros spp., while the nitrate sensitivities of the sinking rates of the genera ranked (high to low) Thalassiosira spp. --> Chaetoceros spp. --> S. costatum. For Thalassiosira aestivalis, sinking rates over all five years did not vary significantly with physical water properties such as ambient temperature, i i i salinity or sigma-t. However daily irradiance showed a significant negative correlation with sinking rates, as did ambient nutrient concentrations. At low daily irradiances (< 3 x 1021 quanta m-2 d-1), ambient nutrient concentration was a better predictor of sinking rate than at high irradiances. The fastest-sinking, most nutrient-sensitive diatoms, the Thalassiosira species, constituted the major source of vertical carbon flux in this embayment during the spring bloom and, considering their cosmopolitan distribution, probably do so in many other such coastal ecosystems. In general, the tendency to sink to the benthos during and/or after a bloom was highly dependent on species-specific cell physiology. Sexuality was observed in Ditylum brightwellii, both in the laboratory and during a fall bloom of this species off Jericho Beach, Vancouver B.C. The field observations indicated that a fraction (ca. 3 %) of the natural population blooming in the upper 3 m became sexual, about a week after the species was first observed to dominate the community. In the laboratory, when sexuality was induced through NO3 limitation and an increase in irradiance, sinking rates increased over 100 times. Maximum sinking rates occurred for zygotes and early post-auxospore cells. Post auxospore cells quickly reduced their sinking rates, and a large fraction of post-auxospore cells began to float at substantial rates (up to 0.4 m d-1) within one week of formation. Sexuality may have a special role in the life cycle of diatoms such as D. brightwellii, including a mechanism for rapid escape from the photic zone, vertical partitioning of a physiologically stressed population into sexual and non-sexual fractions, and subsequent recolonization of the surface waters by positively buoyant post-auxospore cells. The timing and magnitude of diatoms' sinking response to physiological stress may be important in all life history stages of diatoms, and may ultimately affect the balance of selective forces in their evolution. iv TABLE OF CONTENTS PAGE ABSTRACT ii LIST OF TABLES viii LIST OF FIGURES ix ACKNOWLEDGEMENTS xii PREFACE xiii GENERAL INTRODUCTION 1 The global perspective 1 The physiology of sinking: energy 4 The physiology of sinking: nutrients 6 Sinking and sedimentation 9 Sinking, sexuality and life history evolution 10 Purpose of this study 13 Objectives 14 CHAPTER 1. DOES ENERGY CONTROL THE SINKING RATES OF MARINE DIATOMS? 15 Introduction 15 Materials and methods 17 Cultures and culture conditions 17 Sinking rates 18 Biochemical and physiological measurements 19 Experiments 20 Results 23 Experiment 1. Steady state 23 Experiment 2. Time series/transient 26 Experiment 3. Other species 33 Discussion 35 Ditylum brightwellii 35 Other species 36 Cell size and sinking rates 37 Energy and sinking rates 40 CHAPTER 2. SPRING BLOOM SEDIMENTATION IN A SUBARCTIC ECOSYSTEM I. NUTRIENT SENSITIVITY 45 Introduction 45 Materials and methods 46 Results 52 Discussion 73 Nutrients and sinking 74 Irradiance, nutrients and sinking 78 CHAPTER 3 . SPRING BLOOM SEDIMENTATION IN A SUBARCTIC ECOSYSTEM II. SUCCESSION AND SEDIMENTATION 82 Introduction 82 Materials and methods 83 VI Water column sampling methods 83 Sediment traps 84 Results 85 Bloom composition and succession 85 Composition of sedimented material 89 Discussion 94 Sinking and succession 94 Sinking and sedimentation 97 CHAPTER 4. SINKING AND FLOATING DURING SEXUAL REPRODUCTION VADitylum brightwellii 102 Introduction 102 Materials and methods 106 Field measurements 106 Laboratory measurements 107 Results 110 The cell cycle 110 Field observations 112 Laboratory observations 112 Discussion 122 Physiological implications 122 Ecological implications 124 Sinking in a diatom's life history 127 vii CONCLUSIONS 130 REFERENCES 132 viii LIST OF TABLES Table 2.1 Sinking rates of the major diatom species during the spring bloom in Auke Bay, Alaska 57 Table 2.2 Mean sinking rates, maximum sinking rates and nutrient/sinking thresholds during the nutrient saturated and nutrient deplete periods for 6 diatom species during the spring bloom in Auke Bay, Alaska 64 Table 3.1 Comparison of vertically integrated species composition near the spring bloom peak in 1988 and 1989 and the total number of cells reaching the 35 m sediment trap 90 ix LIST OF FIGURES Fig. 0.1 Flow-box model of global carbon cycle (from Mackenzie and Lehrman, 1988) 2 Fig. 1.1 Schematic representation of design for experiments 1 and 2 22 Fig. 1.2 Mean sinking rates and size-specific sinking rates of Ditylum brightwellii under various energetic conditions 24 Fig. 1.3 Characteristics of Ditylum brightwellii cells grown at high light, placed in the dark and monitored over 226 hours in the dark 28 Fig. 1.4 Sinking rate time series of Ditylum brightwellii cultures placed in the dark and monitored over the subsequent 226 h 30 Fig. 1.5 Size specific sinking rates of Ditylum brightwellii 31 Fig. 1.6 Relationship of sinking rates vs cell respiration rate in the dark 32 Fig. 1.7 Sinking rates of Thalassiosira pseudonana and T. weissflogii grown at various light-limiting and light-saturating irradiances, in the light, and after various periods in the dark 34 Fig. 1.8 Conceptual model of diatom sinking control during a spring bloom 41 Fig. 2.1 Map of Auke Bay, Alaska showing location of study area 47 Fig. 2.2 Time series of ambient nitrate, silicate and phosphate concentrations during the spring (1985-89) at the surface in Auke Bay 53 Fig. 2.3 Time series of the sinking rates of Thalassiosira aestivalis between 1985 and 1989 55 Fig. 2.4 Scatterplots showing the relationship between ambient nitrate concentrations and the sinking rates of Thalassiosira aestivalis and Skeletonema costatum 59 Fig. 2.5 Comparison of average sinking rates of various species at the surface and at the chlorophyll maximum (1988-89) 62 Fig. 2.6 Comparison of the sizes of internal nitrate pools for cells (>99% Thalassiosira aestivalis by biovolume) at the surface and chlorophyll maximum during the nutrient deplete period of 1988 63 Fig. 2.7 Chain length frequency distributions for Thalassiosira aestivalis and Skeletonema costatum in 1988 66 Fig. 2.8 Time series of numerical abundance of aggregates (clumps) of Thalassiosira aestivalis cells at the surface and at the chlorophyll maximum in 1988 68 Fig. 2.9 Nutrient sinking relationships for Thalassiosira aestivalis cells with different light histories 72 Fig. 3.1 Time series of species composition and abundance at 2 m during the spring bloom in Auke Bay (1985-89) 86 Fig. 3.2 Relative abundance of three major spring bloom diatom genera, integrated temporally over each spring period at 1 or 2 depths and in sediment traps at 20 and 35 m, 1985-89 88 Fig. 3.3 Relationship between integrated cell abundances for the three principal diatom genera at 2 m over the spring bloom for 1987, 1988 and 1989 and total temporally integrated cell numbers collected in the 35 m sediment trap for these same years 93 xi Fig. 4.1 The sexual cycle in Ditylum brightwellii 110 Fig. 4.2 Time series of presence/absence/dominance of D. brightwellii and water surface temperature and salinity at Jericho Pier 1989-91 112 Fig. 4.3 Photograph of male and female gametes from surface water samples at Jericho Pier, Vancouver B.C. during the fall bloom in October, 1991 113 Fig. 4.4 Cell composition of sexual cultures of Ditylum brightwellii over time 115 Fig. 4.5 Time series of sinking rates of cells in different stages of sexuality 117 Fig. 4.6 A time series of hypothetical sinking rate and floating rates for cells of Ditylum brightwellii entering sexual phase 119 xi i ACKNOWLEDGEMENTS I would like to thank my parents for their ongoing support and encouragement from headquarters in Halifax, which helped keep me sane during the last five years. My enjoyment of this support would have been impossible without their generous phone bill subsidies (among innumerable others). Many members of the Oceanography department taught me laboratory techniques, contributed ideas, encouragement and stimulating arguments, and kept me excited about science during my stay at U.B.C. In particular, I thank Peter Thompson, Maurice Levasseur, and Rowan Haigh, and Bill Cochlan as well as the other members of the Harrison lab. Anne Fisher helped me complete experiments that would otherwise have languished on the drawing board. I thank my supervisory committee, Drs. S.E. Calvert, K. Denman, P. LeBlond and F.J.R. Taylor. The scientists and staff of the APPRISE project helped make my field work an exciting experience. The technical support of the Oceanic Institute in Waimanalo, Hawaii, and the scientific stewardship of Paul K. Bienfang were appreciated. I would like to thank Lytha Conquest, and especially Sara Anderson in Juneau whose sense of humour and strong good sense helped me keep science in perspective. Karen Perry, Heinz Heckl, Bert Mueller and David Burgogne kept me challenged in the political arena (the coffee room). Dave, Heinz and Bert were also part of the Tuesday morning ice hockey crowd that helped keep me physically (and mentally) hopping. I thank Karen Greaves for her warm support of me through the ups and downs of graduate school. Thanks to the household at 2133 Cypress who made the last year of my degree such a a warm and wonderful time, and to our providers of fine conversation and capuccino next door. I was supported financially by a NSERC pre-doctoral scholarship, an I.W.Killam Memorial Fellowship and a University Graduate Fellowship from U.B.C. I would like to thank my supervisor Dr. P.J. Harrison for his moral support, his sound scientific advice, and for providing a stimulating and progressive atmosphere in which to think and work. His example is one I can only hope to follow. xiii PREFACE Three of four thesis chapters have been executed with the collaboration of other investigators and published separately as scientific papers. Experiments 1 and 2 in Chapter 1 were conducted with Dr. Peter A. Thompson, and will be published in Limnology and Oceanography (June 1992) under the same chapter title with Dr. Thompson as second author. Dr. Paul K. Bienfang was chief scientist of the APPRISE project, and collaborated on Chapters 2 and 3. These chapters are currently in press in Marine Biology with Dr. Bienfang as second author. Dr. Paul J. Harrison (thesis supervisor) is third author on all three papers. INTRODUCTION The global perspective Spring diatom blooms in highly productive ecosystems are generally the periods of highest yearly "new" production, when most carbon is available for export (Allen, 1971; Goering et al., 1973a & b). These same regions tend to have relatively high sedimentary flux per unit production (Berger et al., 1989; Burrell 1988), and coastal sediments are significantly enriched in carbon content in comparison to the open ocean. Biological controls of carbon sedimentation during coastal spring diatom blooms are therefore of significant interest when considered in their global framework. Some investigators even argue that an assessment of the magnitude and time scale of phytoplankton growth and sedimentation is one of the most important tasks in the study of global climate change (Berger et al., 1989). While the deep ocean is a carbon reservoir with a turnover time of about a thousand years, the coastal sediments' turnover time is tens of thousands of years (Mackenzie and Lehrman, 1988; Fig. 1), over an order of magnitude greater than deep ocean sediments (and approximately three times the turnover time of terrestrial soils). The coastal sediments may thus be more able than other large carbon reservoirs to act as a long term carbon "sink" to modify anthropogenic carbon input into the atmosphere. Recently, others have suggested that oceanic biological processes are of little importance in terms of modifying anthropogenic carbon inputs into the atmosphere (Sarmiento, 1990; Broeker, 1991). Although some of these arguments are based on false assumptions about the limits of biological processes (e.g. Broeker, 1991), there is no doubt 1 34 W land biota 4.99 9.5 yr 6260 humus 17.8 34 yr inorg. soil 52.0 10,600 yr 15 atmosphere 6.137 3.6 yr ATMOSPHERE coastal biota 0.0042 0.07 yr coastal seds. 180 32,000 yr o u (A o CJ to sv OCEAN oastal waters4^2f l_ s u r f a c e o c e a r -aisn 1.60 35.4 • |3 yr -5QQ^22yr ro 01 o A. ocean biota 0.022 0.07 yr deep ocean 275 1200 yr T +XSJ Fig. 0.1. The global carbon cycle. The euphotic zone is represented by "coastal waters" biota and "surface (open) ocean" biota. Reservoir masses are in units of 1016 mol C, and fluxes in units of 1012 mol C y-1. Residence time is shown at the bottom of the boxes. Dashed lines separate land, oceanic and atmospheric carbon reservoirs. Thickened arrows represent fluxes of interest in this study. Redrawn from Mackenzie and Lehrman (1988). 2 that chemical and physical properties are very important, and that they may be more important than organismal processes over geological time scales. The resolution of this argument can only occur once the long term impact of short term, non-linear biological processes on the global carbon cycle is fully understood. The mass sedimentation events following phytoplankton blooms are some of the most clearly identifiable biological carbon fluxes in the ocean. Understanding the dynamics of bloom systems must aid in assessing the potential impact of biological processes. On a more local scale, the increase in phytoplankton biomass during a spring bloom represents available food essential to the recruitment success of various finfish and shellfish. Biomass remaining in the pelagic zone will become potential food for microzooplankton, copepods and euphausiids, and they in turn may be preyed upon by herring, smelt, salmonid species and pelagic feeding whales. Biomass sedimenting to the benthos becomes potential food for benthic invertebrates such as mysids, amphipods and bivalves, which can in turn be eaten by flatfish, crabs, and benthic-feeding grey whales. Spring bloom sedimentation dynamics therefore impact pelagic and benthic ecology, and fisheries in coastal areas. Recently, investigators have identified the importance of understanding the species-specific biological processes involved in the control of production export from the photic zone (Williams and Bodungen 1989). The tendency for diatoms to reach the benthos intact is species-specific (Sancetta and Calvert 1988, Riebesell 1989), but how an individual species' sinking processes would affect larger scale sedimentation processes of blooms in nature remains poorly understood. Interspecific variability thus poses a problem to larger-scale modelers attempting to integrate diatom dynamics into estimates of global primary production. In fact, interspecific variability and the vicissitudes of individual organisms have been identified as one of the principal limitations to the predictability of marine 3 systems, since each arrow between boxes in a carbon flow model (e.g. Fig. 1) can be "pushed" differently by different organisms at different times (Smetacek and Pollehne, 1986). Ecology and the flux of carbon are inseparable, because each is determined by the other. Once the functional linkage between diatoms as organisms is understood (i.e. their physiology, life history, and ecology) and their sedimentation patterns, we may at least understand the limits to predictability of diatom sedimentation. In the long term, this approach might even yield some predictive capability of its own. The physiology of sinking: energy Laboratory studies have shown that phytoplankton sinking rates are dependent on the light- and nutrient-histories of cultures as well as the time scales of experiments (Bienfang 1981a; Harrison et al., 1986). Early observations noted that sinking must be under some kind of physiological control, possibly based on the regulation of the ionic composition of a diatom's large internal vesicle, the vacuole (Gross and Zeuthen, 1948; Eppley et al., 1967; Smayda, 1970). It was also suggested that sinking rates increased as a cell's physiological status deteriorated (Gross and Zeuthen, 1948; Smayda and Boleyn, 1965; 1966). However, the direct influence of light (Bienfang et al., 1983; Johnson and Smith, 1986) and nutrients (Smayda, 1970; Bienfang et al., 1982) on sinking rates was seen to be highly variable, and the only clear pattern to emerge across all species was that high sinking rates were more often observed at low growth rates (and/or during senescence) than at high growth rates. Investigators subsequently documented the existence of a vacuolar ion pump influencing cell density. This seemed to be affected both by light (Anderson and Sweeney 1977), and by inhibitors restricting the availability of respiratory energy (Anderson and Sweeney 1978). Since ionic pumps require energy (in the form of ATP), sinking should be 4 dependent either on incoming irradiance or on the availability of stored cellular energy. However, in later work, the influence of irradiance on sinking rates was found to be very variable (Culver and Smith 1989; Johnson and Smith 1986; Bienfang 1981a & b; Granata 1991). To date, the influence of respiratory energy availability has not been systematically tested. I attempt to develop a more general explanation for the apparent lack of consistent pattern in previous results, by estimating the relationship between sinking rates, photosynthesis and respiration. I discuss the possibility that physiological sinking rate control requires a constant energy supply. This is the initial focus of my study (Chapter 1) and forms the conceptual basis for the rest of the thesis. Cell composition (outside of the vacuole) can be an important determinant of the maximum sinking rates attainable by a particular organism. The buoyancy of many aquatic cyanobacteria, for instance, is controlled principally by weight increases driven by the accumulation of carbohydrate storage products like polyglucan (Rromkamp et al., 1988) which is in turn directly proportional to photosynthetic rate (Ibelings et al., 1991). In these organisms sinking rates decrease as cells become energy-limited, the opposite pattern to the one predicted by the energy-control hypothesis (above). Cell size may also be an important consideration for sinking rate control. Smayda (1970) determined that as cell size increased across many genera of phytoplankton, so did their sinking rates. Eppley and co-workers (1967) determined that larger diatoms were more likely to sink faster than small ones. Other studies showed that the largest cells were more frequently positively buoyant than smaller cells (Skreslet, 1988; Villareal, 1988). Possibly, as a cell's size increases, so does its possible range of sinking rates. Physiological control might then be superimposed on the basic constraints of cell size and cell composition. It 5 emerges that sinking control strategies may be size-specific, and larger cells may be more likely than smaller cells to display energetic sinking rate control. The physiology of sinking: nutrients In nature, any energy-dependence of sinking rates during a spring bloom would be the principal sinking control only as long as other factors did not limit growth, during the first half of a bloom. It is nutrient exhaustion which frequently terminates coastal spring diatom blooms (Legendre 1980, Smetacek, 1985a). Once nutrients become limiting, another more complex set of factors will influence sinking rates. Nutrient depletion decreases cell growth rate, ultimately affecting many aspects of cell physiology. These in turn can influence a cell's ATP supply (Noges, 1989), and consequently, nutrient depletion could potentially influence sinking rate via the same mechanism as light limitation, by decreasing the energy available for energetic control. However, in addition, nutrient depletion can cause changes in cell composition (Caperon and Meyer, 1972; Laws and Bannister, 1980) and occasionally cell size (Thompson, 1991). Diatom sinking rates thus become highly variable, and not always simply contingent on ambient nutrient concentrations (Smayda 1970; Bienfang, et al., 1982; Bienfang and Harrison, 1984a and b). Studies on the effects of nutrient limitation on sinking rates have shown many patterns. In some studies there was no apparent relationship between NO3 or NH4 limitation and sinking rates (Bienfang, 1981a; Bienfang and Harrison, 1984b), while in others there was a direct relation (Bienfang 1981b; Bienfang et al., 1982). Both silicate and phosphate limitation have been associated with clearer increases in sinking rates, especially for certain centric diatoms (Boleyn, 1972; Bienfang et al., 1982; Bienfang and Harrison, 1984a, b; Gibson, 1984; Harrison et al., 1986). Of all types of nutrient limitation, silicate 6 limitation is often seen to have the most important effect on sinking rate (Bienfang 1981a; Bienfang and Harrison, 1984a). Phosphate limitation also increases cell sinking rates over a broad range of freshwater species (Tilman and Kilham, 1976). The exact causes of these sinking rate changes is unclear, since the effect of nutrient depletion causes many changes in diatom cells. Silicate limitation can increase the mean cell size since cells can become biprotoplastic and cannot divide once they have insufficient silicate available for frustule formation (Harrison et al., 1977) and thus increase the maximum sinking rate of a cell. Silicate is present in many cell organelles, most importantly the mitochondrion, the site of cell respiration, as well as the nucleus, where it is involved with DNA synthesis. Silicate limitation might therefore potentially interfere with cellular energetics. Nitrogen limitation can cause changes in cellular composition (Laws and Bannister, 1980) including an increase in the C:N ratio of cells (Caperon and Meyer, 1972) and increases in carbohydrate storage. In addition, lengthy periods of nitrogen limitation decrease protein synthesis (Turpin, 1991), hence potentially reducing the cell's capacity to perform basic metabolic processes over a period equivalent to the intracellular turnover time of critical enzymes. Phosphorus limitation might directly limit the availability of ATP and hence affect energy availability. Because of the physiological lag between nutrient depletion and physiological changes, it is difficult to separate many of these effects. Smayda and Bienfang (1983) found that population sinking rates as measured by chlorophyll were less than half those based on cell number. This difference indicated that cells with higher chlorophyll content were remaining preferentially in the water column, and may be an indication that cell composition did indeed change intracellular sinking rates, although this might also reflect differences in metabolic activity. Other studies (Bienfang et al., 1986; Johnson and Smith, 1986) have also 7 shown that when sinking rates are measured on cultures using different cell constituents as biomass indices (e.g. chlorophyll, particulate carbon, etc.), different constituents seem to have different sinking rates. This suggests that cells with different composition may sink at different rates within the same culture. Compositional differences must therefore be considered, although they may be secondary in importance to cell energetics in terms of sinking rate control. The implications of these studies for ocean ecology are complex. Sinking rate changes in the field may alter pelagic community structure as well as the probability of sedimentation for any particular species. The effect of nutrient-mediated sinking rate changes on species succession has been documented in several laboratory studies (e.g., Harrison et al., 1986; Brzezinski and Nelson, 1988). Harrison and co-workers concluded that sinking differences between Skeletonema costatum and other species under NH4 limitation would cause S. costatum to be retained in the upper water column after the other species sank out due to nutrient-induced sinking rate increases. In the study by Brzezinski and Nelson (1988), sinking rate increases effected the early removal of nutrient-sensitive species (i.e. the species with higher nutrient demand) from the population through sedimentation, leaving hardier species less nutrient-limited than if other species remained in the water column. The higher the nutrient demand, the more likely a species was to sediment out quickly once nutrients were depleted. Thus high potential growth rate was associated with high early sinking rates. 8 Sinking and sedimentation The sedimentation of dense diatom blooms has been the object of intensive study, especially in high latitudes (Laws et al., 1988; Passow, 1991; Wassman et al., 1991). Sedimentation maxima frequently occur during and after spring diatom blooms, but interannual variability in the magnitude of the flux and the organisms which constitute it is very high (Wassman et al., 1991). Many processes are responsible for this variability. Variability in basic atmospheric factors such as irradiance and temperature may effect changes in bloom growth rate and maximum attainable biomass (Ziemann et al., 1990). Cell aggregation can increase diatom sedimentation (Passow, 1991). Grazing can either increase the sedimentation rate of diatoms through their consolidation into fecal pellets or decrease it through respiratory losses and coprophagy (Lampitt et al., 1990). However, the contribution of diatom physiology to the entire process of sedimentation remains poorly understood. Field studies linking the actual sinking rate changes of diatoms in the water column with sedimentation fluxes of cells to sediment traps have been scarce. Riebesell (1989) monitored sinking rates and sedimentation during a bloom in an enclosed mesocosm, and concluded that the patterns of diatom species abundances could be explained on a general level by the balance between observed growth, sinking and sedimentation. Aggregation was an important factor affecting the speed of sedimentation late in the bloom. This is consistent with the model originally advanced by Riley et al. (1949). Although field data have shown rapid sedimentation of phytoplankton following post-bloom nutrient exhaustion (Skjoldal and Lannergren 1978; Smetacek 1980), to date no detailed field studies have documented the smaller scale dynamics of the nutrient-sinking interaction in nature. I attempted to document the physiological sinking response of several important coastal diatom species to nutrient depletion during the spring bloom in a subarctic ecosystem (Chapter 2). I then 9 compared the physiological sinking response of these organisms to their eventual sedimentary flux (Chapter 3). Sinking, sexuality and life history evolution Smetacek (1985b) identified diatom sinking processes as an integral part of their life history. Diatoms live in a turbulent environment where physical processes are constantly homogenizing and dispersing phytoplankton patches formed by exponential cell growth. Although there are certain critical stress levels (e.g., wind speeds > 5 m s-l) and space scales (<1 km) above which turbulence begins to homogenize biologically-induced patchiness, this variability is largely unpredictable and occurs at almost every time and space scale. For small organisms, unpredictability often leads to the development of an opportunistic life history strategy (Stearns, 1976). In order to survive, cells must respond rapidly to short temporal and spatial windows favourable for growth (or favourable "patch" sensu Hutchinson 1961), and develop mechanisms for surviving less favourable periods away from the ocean surface, where metabolic demand is high, and inactive cells are exposed to grazing (Paul and Coyle, 1990a&b) dissolution (Roelofs, 1983) and damaging radiation (e.g. Richter, 1987). Sinking thus becomes an integral part of a successful life history strategy (Smetacek 1985b). The three principal stages of a diatom's life cycle are 1) vegetative, asexual growth (>99 % of cells observed in the field are vegetative cells) 2) sexuality and 3) vegetative resting spore formation. The sexuality of diatoms has been studied for many years (Gross 1937; v. Stosch, 1958) but it is still not completely understood. It was suggested that sexuality is necessary not only for genetic recombination, but also to reverse the obligate progressive size reduction undergone by diatoms over many asexual divisions (v. Stosch, 1965), since certain species 10 were found to undergo sexuality only when cells were within a given (small) size range (Drebes, 1977; French and Hargraves, 1985). However, for some species, cell size changes can be produced rapidly and asexually (Kibier et al., 1988; Gallagher, 1983). In addition, the necessity for cell size reduction during vegetative division is itself debatable (Round, 1972). This poses a problem. If a reduction in cell width is an obligate part of diatom growth, and sexuality is the only mechanism by which a wide cell diameter can be regained, then sexuality might be maintained for that purpose alone. If a reduction in cell width is not obligatory (and hence not an evolutionary "constraint"), or if a wide cell width can be restored by asexual means, then other factors must combine to maintain sexuality in diatoms. Cell width restoration can occur asexually in some species (Hargraves, 1982). In addition, some evidence suggests cell size reduction might occur much more rapidly (Gallagher, 1983) and much more slowly (Mann, 1988) than would be expected from the reduction of cell diameter by one frustule width each division. It is clear that occasionally some diatoms are capable of freeing themselves from the constraint of cell-size reduction. If sexuality occurs more often in small cells, then it is likely that the link between narrow cell width and sexuality has adaptive value, rather than being the obligate outcome of frustule constraints. Lewis (1984) argued convincingly that progressive size reduction combined with size-specific sexual induction actually represented the evolution of an endogenous timer designed to allow cells to undergo sexuality intermittently, with the interval between sexual events determined by the number of cell divisions, not time per se. Lewis also suggested that the increase in size range during this process was important because it allowed only a fraction of the population to become sexual at any one time (his assumption was that sexual induction was somehow still size-specific). He argued that the maintenance of a fraction of asexual 11 cells through the size-specific induction of sexuality was one possible mechanism for the preservation of advantageous parental genotypes. Mann (1988) accepted Lewis' (1984) argument that cell-size reduction was probably part of an endogenous timer, and demonstrated that in nature, the sexual cycles of diatoms might be far longer than previously expected (up to 40 years). Mann and others (Harrison, 1974) also discussed the problematic nature of the induction of sexuality in diatoms generally; conditions which induced sexuality at one time did not induce sexuality at other times. The induction of sexuality in diatoms has been seen to occur in response to various physiological stresses such as nutrient limitation (Steele, 1965; Davis et al., 1973), a sudden increase in growth irradiance (Drebes, 1966) or a decrease in temperature (Drebes, 1977). The lack of specificity of the cue suggests that some endogenous timer might be operating. While endogenous timers may exist for some species where progressive cell size reduction has remained adaptive, there is no reason to expect all species to react similarly (Drebes, 1977). Sexuality is yet another case where species-specific variability may be high. Vegetative resting spore formation in diatoms generally occurs in response to nutrient deprivation, especially nitrogen-deficiency (Hargraves, 1982; French and Hargraves, 1985; Kuwata and Takahashi, 1990). Resting spores are often morphologically distinct from asexual cells, show a more heavily silicified frustule (Hasle, 1975; Kuwata and Takahashi, 1990), and can result in vegetative cell enlargement (v. Stosch, 1965; Hargraves, 1982). Thalassiosira nordenskioldii occasionally forms resting spores, although their survival is highly temperature dependent. In this species, spores seem to withstand long periods under cold, dark conditions (Durbin, 1978). It is generally accepted that these heavier cells sink quickly to the sediments for longer term storage. Few sinking rate measurements have been made on the resting spores of phytoplankton, but data indicate that resting spores tend to 12 sink faster than vegetative cells (Anderson et al., 1985). Resting spores can survive for more than a year in the dark (Durbin, 1978) and are frequently resistant to bacterial attack and grazing, but require some time to germinate once they are re-introduced into favourable conditions (Hollibaugh et al., 1981). For this reason, spore formation can be seen as an adaptation to longer term, seasonal fluctuations in nutrients (Kuwata and Takahashi, 1990), although some argue that spores, too, can serve as a short term retention mechanism for pelagic populations (Pitcher, 1986). The sinking rate changes that occur during sexual reproduction are discussed in (Chapter 4), as the opening to a discussion of the importance of sinking and sedimentation in the evolution of diatom life history strategies . Purpose of this study Sinking control is potentially important throughout the life history of a diatom. However, it is bloom sedimentation which forms the most dramatic and ecologically important contribution of diatoms to coastal dynamics. Hopefully, by attempting to describe the evolution of diatom sinking events, the likelihood, induction and control of mass sedimentation can be better understood in its ecological context. The purpose of this study was to determine how sinking processes in various marine diatoms are controlled physiologically, and how this control is integrated into their life history. In this study I attempt to elucidate the principal determinants of diatom sinking rates as a hierarchy of effects, beginning with the most basic factor, light (Chapter 1), then superimposing further complexity (nutrients; Chapter 2), integrating both into a discussion of diatom ecology and the evolution of diatom physiology. I discuss the impact of species-specific sinking rate patterns on sedimentation to the benthos of a shallow subarctic 13 embayment (Chapter 3). I then try to place short term sinking rate changes into the context of their three principal life history stages (vegetative growth, sexual reproduction, and resting spore formation) the sum of which represents a diatom's overall growth and sinking strategy. In this manner I hope to define mora thoroughly the variety of sinking strategies exhibited by diatoms, and work towards a broader understanding of their ecological role (Chapter 4). Objectives: 1. To determine the role of cell energetics in the physiological control of sinking rate. 2. To document and quantify the role of nutrient limitation in controlling the physiological initiation of bloom sedimentation in the field. 3. To assess the impact of this physiological response on the actual sedimentation patterns in the field during and after a bloom. 4. To determine the sinking response to sexual reproduction. 5. To link 1,2, 3, and 4 into a more complete discussion of the evolution of diatom ecology. 14 CHAPTER 1 DOES ENERGY CONTROL THE SINKING RATES OF MARINE DIATOMS? INTRODUCTION If an intracellular, energy-requiring ionic pump is utilized to maintain low sinking rates (Anderson and Sweeney, 1977; 1978) then sinking rates should be predictable by the amount of energy available to the cell via photosynthesis and respiration. Photosynthetic energy production should be determined by light availability, while respiratory energy release should depend on the availability of a suitable intracellular carbon pool, itself a function of the photosynthetic history of a cell (Post et al. 1985). The availability of stored carbon should be greatest in cells grown at the highest irradiance (Cook 1963; Claustre and Gostan 1987). On a short time scale (seconds to hours) respiratory energy should thus remain available to cells grown under saturating irradiance when they are placed in the dark, and cells grown under saturating irradiance should be able to maintain low sinking rates in the dark for longer than cells grown under limiting irradiance. As long as carbon stores are available to release energy (Weger et al. 1989), sinking rate control might be maintained, and a cell should reach its maximum sinking rate when energy sources are exhausted. In addition, the effect of cell size and cell shape on sinking rates should be directly measurable only after energy sources are exhausted, and the cell sinks effectively as an inert particle, at its maximum sinking rate. At most other times, the sinking rate measured would be lower than the maximum rate, the lower rate representing the extent to 15 which the cell could spend energy to physiologically reduce the maximum sinking rate. The effect of cell size and shape on sinking rates might not always be apparent. This series of experiments was designed to test the hypothesis that short-term intracellular energy availability is the principal determinant of short-term changes in sinking rates in the marine diatom Ditylum brightwellii. First, sinking rates were measured at steady state, under saturating and limiting irradiance (for growth), and at various levels of stored energy reserves. Cells grown at saturating irradiance were then placed in the dark, and their changes in sinking rate were monitored as respiratory energy production decreased. Cell sinking rates were measured over ten days at three energy-levels: energy-saturated (in the light), only respiratory energy available (in the dark), and zero-energy state (with a metabolic inhibitor). The hypothesis was that sinking rates would increase as energy availability decreased. Experiments tested the hypothesis that when light is unavailable, there is an inverse dependence of sinking rates on respiration rate. Preliminary experiments were then conducted on the marine diatoms Thalassiosira weissflogii and Thalassiosira pseudonana to test for similar patterns of energetic sinking control. 16 MATERIALS AND METHODS Cultures and culture conditions Cultures of the marine diatoms Ditylum brightwellii, Thalassiosira pseudonana, and Thalassiosira weissflogii were obtained from the Northeast Pacific Culture Collection (NEPCC Nos. 8a, 58, and 418 respectively) Department of Oceanography, University of British Columbia, Vancouver, B.C. One-liter batch cultures of D. brightwellii, T. pseudonana and T. weissflogii were grown in nutrient-saturated artificial seawater (ESAW) (Harrison et al. 1980) with the following modifications: ferrous ammonium sulphate and sodium glycerophosphate were replaced with equimolar ferric chloride and sodium phosphate, respectively, and selenite and molybdate were added for a final concentration of 1 nmol L-l. All cultures were grown under continuous fluorescent light. Under high light (HL) (110 umol quanta m-2 s-1), light was saturating, and the cells were growing at their maximum growth rates (Umax), (for D. brightwellii, Umax = 1-15 ± 0.05 d"1 (+ 1 S.D.)) Under low light (LL) (achieved by 4 layers of neutral-density screening which reduced irradiance to 7-10 umol quanta m-2 s-1), light was limiting and the cells were growing at 0.45 + 0.05 d-1 (+ 1 S.D), roughly 40% of Umax- All experiments were conducted between early- and mid-log phase. Cultures were maintained at these light levels for several dilutions (> 30 doublings) before experiments were conducted. All cultures were maintained at 170C and stirred constantly with a Teflon-coated magnetic stir bar at 120 rpm. 17 Culture conditions for Thalassiosira pseudonana and Thalassiosira weissflogii were similar to those for D. bnghtwellii except cultures were grown at four different irradiances. For Thalassiosira pseudonana these irradiances were 10, 50, 90 and 160 umol quanta m~2 s-1, at which growth rates were 0.38,1.00,1.30, and 1.35 d-1 respectively (Umax = 1.35 + 0.06). For Thalassiosira weissflogii the irradiances were 5, 10, 40, and 100 umol quanta m"2 s-1, at which growth rates were 0.28, 0.58, 0.61 and 0.70 d-1 respectively (um ax = 0.70 + 0.05). Sinking rates The sinking rates of all cultures were measured using the SETCOL method (Bienfang 1981b). Cell number was used as a biomass index for the SETCOL calculations, measured on a Coulter CounterR with a TA-II accessory designed to split the counts into size-specific channels. All counts for Ditylum brightwellii were made using the 200 urn aperture, and samples were generally diluted 1:4 with filtered 3% NaCl solution before counting. For cells grown at HL, about 70% of cells fell within two channels (4 - 8 x 103 um3); about 90% of cells fell within four channels (1 - 8 x 103 um3). For cells grown at LL, about 90% of cells fell within two channels (2 - 4 x 103 um3) and 95% of cells fell within four channels (1 - 8 x 103 um3). Sinking rates were first calculated using total cell numbers, yielding an average sinking rate for the entire culture. They were then calculated separately for the four most abundant channels, yielding size-specific sinking rates for a range of cell sizes within the culture. Counts of Thalassiosira weissflogii and Thalassiosira pseudonana were made with the 70 urn aperture. 18 Because many SETCOL measurements were made in the light, increases in cell numbers often occurred during the experiment. I was not confident that the averaging of initial (Boo) and final (Bot) cell concentrations to correct for growth in the column over the settling period was adequate. The SETCOL method was therefore modified in the following manner: Instead of using (Boo+Bot)/2 as an estimate of initial biomass, final concentrations of all SETCOL fractions (top, middle, and bottom fractions of the column) were measured. An average concentration of cells in the column was then calculated, which was used instead of the (Boo+Bot)/2 value. If growth was constant throughout the column over the period of the experiment, using this average final concentration in the sinking rate calculation should yield an accurate measure of sinking rate. In addition, to ensure that all cells were drained from the bottom fraction of the SETCOL, the bottom fraction of each settling column was rinsed with a 3% NaCl solution; the rinse was counted separately and the cell numbers added to the bottom fraction count. Sinking rate measurements of Ditylum brightwellii were made in the same light conditions under which the cultures were grown (HL or LL), in the dark, and with 2.3 mg L-1 DCMU (dichloro-methyl urea) added to the SETCOL column to inhibit photosynthetic energy conversion during the period over which sinking rates were measured. The sinking rates of HL cultures in the time series experiment were also measured after the addition of 12 mmol L-1 KCN (potassium cyanide) which acted as a respiratory inhibitor. Sinking rates under KCN were termed the maximum sinking rate (MSR) of cells at a given time. All sinking rates in Experiment 1 (see below) were measured over 4 h; sinking rates in Experiment 2 were measured over 3 h. 19 In experiment 3, the sinking rates of Thalassiosira weissflogii and Thalassiosira pseudonana were conducted over 3 h at the same irradiance at which they were grown, and then over 3 h in the dark after various periods of dark acclimatization (see below). Biochemical and physiological measurements Before sinking rates were measured, respiration and photosynthesis were measured with a Clark-type (Delieu and Walker 1972) Hansatech Oxygen Electrode UnitR in D. brightwellii cultures with and without DCMU or KCN. DCMU at ~2.3 mg L-1 was found to inhibit all photosynthesis, while respiration remained at 100% to 110% of normal rates. Twelve mmol L-1 KCN inhibited all respiration and about 87% of photosynthesis. To ensure that KCN-treated cells did not photosynthesize, all KCN sinking rate measurements were carried out in the dark. Subsamples from each culture of each species were taken at time To of the SETCOL trial and filtered onto precombusted GF/F filters, for in vitro chlorophyll, POC, and PON analysis. Chlorophyll samples were sonicated for 5 min, extracted in 10 ml of 90% acetone at 40C for 24 h, and measured for fluorescence on a Turner Designs^ fluorometer (Parsons et al. 1984). POC and PON samples were stored frozen until analysis on a Carlo Erba N-Analyzer (Model NA 1500). Experiments In experiment 1, triplicate 1-L cultures ofDitylum brightwellii were grown at HL and LL. The sinking rates of the HL cultures were measured under HL, DK, and with DCMU 20 added to the SETCOL column. The sinking rates of LL cultures were measured under LL, DK, and with DCMU (see Fig. 1.1). All comparisons between sinking rates were done with t tests with p = 0.05 and n = 3. For experiment 2, three 12-L batch cultures of D. brightwellii were grown under HL. At t = 0, during mid-exponential phase, the lights were turned off and all three cultures were covered with dark plastic to prevent light from reaching them. A 1.1-L subsample was taken from each of the three cultures at t = 0, 6, 12, 24, 36, 54, 78,120, 147, and 226 h, for a total of 10 subsamples of each culture over 226 h of darkness. At each sampling time, chlorophyll, POC and PON, cell numbers and cell size distribution of each subsample were measured. The sinking rate of each of the three cultures was measured over 3 h as above using 300 mL for the SETCOL procedure from each subsample at each sampling time, under three conditions: high light (HL), dark (DK), and with 12 mmol L-l KCN as a respiratory inhibitor (KCN). For example, at t = 24 h, cultures that had been in darkness for 24 h were placed into SETCOLs that were placed in HL, darkness, or in darkness with KCN added (Fig. 1.1). For experiment 3, triplicate 1 L batch cultures of Thalassiosira pseudonana and Thalassiosira weissflogii cultures were grown at various irradiances (see above), and their sinking rates were measured over 3 h at these same irradiances. All cultures were then placed in the dark for various periods. For T. pseudonana dark periods were 3, 9, 24, 48 and 100 h. For T. weissflogii dark periods were 3, 9, 26, 55 and 100 h (see above). Their sinking rates were then measured over 3 h in the dark subsequent to these dark periods. 21 E X P E R I M E N T 1 Growth Conditions (umol quanta nrV 1 ) Sinking rate measurement conditions (h in dark) Abbreviation high light: 110 <HL) low light: 40 (LL) E X P E R I M E N T Growth o Conditions (jixnol quanta m 2 s'1) high light: 110 0(HD 4CDE) 4 + KCN OdX) 4CDK) HL/HL BL/DK HL/KCN LL/LL LL/DE Treatment* (h i n dark) 0 • 6 12 24 36 48 54 72 147 226 Sinking rate measurement conditions (umol quanta mM (h in dark or light) highlight: 110 3h dark 3h dark + KCN 3h NB: cultures were sampled sequentially over 226 bin the dark Abbreviation IT EK KCN Fig. 1.1. Schematic representation of experimental design for experiments 1 and 2 indicating abbreviations used in text to represent various, treatments, and showing the irradiance at which cultures were grown, the length of time cultures were placed in the dark (if at all), and the length of time and the light conditions over which sinking rate measurements were made using the SETCOL method. 22 RESULTS Experiment 1: Steady state cultures Cells grown at HL sank at 0.028 m d-1 when sinking rates were measured under the same high light conditions (HL/HL, Fig. 1.2A; for explanation of abbreviations, see Fig. 1.1). When sinking rates of HL cells were measured over 4 h in the dark (HL/DK), HL cells had the same sinking rates as when sinking rates were measured in the light (HL/DK = HL/HL, Fig. 1.2A). HL/DCMU sinking rates were not significantly different from HL/HL and HL/DK sinking rates but they were more variable (data not shown). HL/KCN sinking rates were estimated' from all KCN measurements in Experiment 2. HL/KCN measurements alone therefore represent a mean of 9 measurements on each of three replicate cultures from Experiment 2 for a total of 27 measurements. HL/KCN sinking rates were significantly higher (20 times) than the HL/HL sinking rates (Fig. 1.2A) (t test, p < 0.05), and were the highest of all sinking rates measured. Over a 4 h measurement in the dark, low light grown cells had significantly higher (4x) sinking rates (LL/DK) than when the same low light grown cells were kept at low light (LL/LL) (Fig. 1.2A). LL/DCMU treated cells had sinking rates similar (t test, p > 0.05) to LL/LL cells (data not shown). The sinking rate of LL cells measured at the same light intensity at which they were grown (LL/LL) was 0.043 m d-1, significantly higher than HL/HL sinking rates (but significantly lower than the HL/KCN values; t test, p < 0.05) (Fig. 1.2A). The difference in sinking rates between LL/LL and HL/HL cells was no longer significant when sinking rates were normalized to cell size (LL/LL and HL/HL lines, Fig. 23 Fig. 1.2 (A) Sinking rates of Ditylum brightwellii under various conditions. Treatments are arranged in order from low to high energy availability to the cell: HL/KCN (high light grown cells placed in the dark with KCN as a respiratory inhibitor); LIVDK (low-light grown cells placed in the dark for 4 h); LL/LL (low light grown cells kept at low light for the 4 h sinking rate measurement); HL/DK (high light cells placed in the dark for 4 h); HL/HL (high light cells kept at high light for the 4 h sinking rate measurement). All values represent the mean of three cultures + 1 S.E. (B) Size specific sinking rates of D. brightwellii under the same energetic conditions as A, Symbols represent: • = HL/KCN; • = LL/DK; O = LL/LL; • = HL/DK; • = HL/HL. 24 8 E, ui < (3 Z S2 z c/5 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 -0.1 annnD HL/KCN LL/DK LULL HL/DK HUHL . i i i i i NONE LOW HIGH ENERGY LEVEL 1.0 E CO DC D) C 12 c </5 0.6 0.6 T 1 1 1—-\ 1 1 1 r B + I / / :• • i < • D • D HL/HL 0 LL/DK HL/DK Q L l A L + HL/KCN M -0.2 0.0 ' Energy State 0 - ENERGY none RESPIRATORY ENERGY ONLY LIGHT ENERGY 0 1 2 3 4 5 6 7 8 9 10 Cell Volume (yLim3x103) 1.2B). Small cells in both the HL and LL cultures sank at the same rate, but HL cultures had an average cell volume of 4.5 x 103 |j.m3, while LL cultures had a significantly smaller average cell volume, 3.1 x 103 um3. For HL/HL cells, there was a significant negative correlation between cell size and sinking rate (r2 = 0.40; n = 12) (Fig. 1.2B). HL/KCN cells showed a positive correlation between size and sinking rate, at much higher overall sinking rates. Across all treatments, the smallest cells (1,000 fAm3) had a range of sinking rates (maximum minus minimum) of 0.45 m d-1 (Fig. 1.2B; HL/KCN - HL/DK). The range of sinking rates of the largest cells (8,000 nm3) was twice that of the smallest cells (0.9 m d"l; HL/KCN - HL/HL). Overall, the total range of sinking rates observed was directly proportional to cell size (r2 = 0.93, n = 4). Experiment 2: Time series / transient When HL cells were placed in the dark, cell numbers increased by 30% over the first 36 h of darkness. Thereafter, cell concentration remained roughly constant, with an apparent decline near the end of the experiment (Fig. 1.3A). Cell carbon content decreased exponentially over time, and by t = 226 h, it was reduced to approximately 50% of the initial cell carbon content (Fig. 1.3B). Over the first 36 h, average cell volume decreased by 15% (Fig. 1.3C). Cell volume then changed little until t = 120 h and declined slowly between t = 120 and t = 226 h. Over the entire experiment, average cell volume decreased by 40%. Since cell volume decreased by 40% and cell carbon quota decreased by 50%, the decrease in cell size accounted for up to -80% of the total carbon loss per cell. It is assumed that the residual loss was due to respiration. Respiration rate (RR) was estimated from the following equation describing the exponential decrease of cell carbon content (per unit volume) with time 26 (Please note: We did not measure dissolved organic carbon (DOC) in the cultures. If there was substantial excretion of DOC, respiration would be overestimated): C = 0.022 exp(-0.029t) + 0.098 where C = pg C um"3; t = time (h). The first derivative of this equation provides the equation for the change in estimated respiration rate (RR) with time: RR = 0.638 exp(-0.029t) where RR = (fg C h-1 um_3). Thus, at t = 0, the respiration rate was approximately 0.638 fg C h-1 um"3; by t = 226 h, respiration rate was approaching 0 (see below). At all times, sinking rates decreased in the sequence inhibitor-treated cells > cells in dark > cells in the light (Fig. 1.4). When HL cells were placed in the dark and monitored over the next 226 h of darkness, DK sinking rates (measured over 3 h in the dark) and LT (measured over 3 h of light exposure) sinking rates increased over time in the dark, and finally converged upon KCN sinking rates after 226 h. The range of sinking rates measured over all three treatments (KCN sinking rate minus LT sinking rate) thus decreased from 0.6 m d-1 at t = 0, to 0.1 m d-1 at t = 226 h. Some floating cells (cells with a measurable upwards velocity) were observed at t = 36 h under LT conditions. For at least 24 h of darkness, cell sinking rates (DK) remained similar (p > 0.05) to the HL/HL value (-0.02 - -0.05 m d-1; as in Exp.l) (see DK line, Fig. 1.4). After 36 h of darkness, the cultures had a significantly higher (t test, p < 0.05) sinking rate than HL/HL cultures (0.23 m d-1), and after 54 h of darkness, their sinking rates remained significantly higher than the HL/HL sinking rate. There was a continuous, steady increase in sinking rate of cells kept in the dark, from 54 to 226 h. Initially (t = 0), the smallest cells sank most quickly, and there was an inverse relationship between cell size and sinking rate (Fig. 1.5, see 0 h). This relationship was 27 2g similar to that found for HL/HL in Exp. 1 (Fig. 1.2B). At 24 h, cells of all sizes sank at similar rates (similar to LL/DK, Exp. 1, Fig. 1.2B). By 36 h, the relationship had reversed, and all sinking rates were directly proportional to cell size (Fig. 1.5, see 36 h). The size-sinking rate relationship found for HL cells at 226 h was similar to the relationship found for HIVKCN cells (Fig. 1.2B). In fact, there was striking overall similarity between the energy-dependent size-sinking relationships of the steady state and time-series experiments (Figs. 1.2B and 1.5). When DK sinking rates were plotted against the estimated respiration rate, a negative exponential relationship was observed (Fig. 1.6). For at least 78 h after being placed in the dark, cells exposed to 3 h of light had average sinking rates similar to HL/HL values (HL line, Fig. 1.4) over the same time period. After 120 h of darkness, the sinking rates of cells measured over 3 h of light exposure were significantly higher than HL/HL values (t test; p < 0.05), but significantly lower than DK sinking rates. After 226 h of darkness, LT sinking rates were similar to both DK and KCN sinking rates (Fig. 1.4). The size-dependence of LT sinking rates was qualitatively similar to DK sinking rates (as in Fig. 1.5), but LT rates were consistently lower than DK sinking rates (data not shown). Because KCN sinking rates were significantly more temporally variable than other sinking rates (coefficient of variation ~40%), it was not possible to resolve small changes in MSR. Overall, cells treated with KCN showed no significant change in sinking rate over time (see hatched area representing mean + 1 SD, of all KCN sinking rates, Fig. 1.3) and had high sinking rates (0.4 - 0.8 m d-1) throughout the experiment. This high variability may be due to technical problems in the methodology of handling the inhibitor-treated cells. Generally, KCN-treated cells had sinking rates directly proportional to cell volume (Fig. 1.2B; HL/KCN). 29 0 50 100 150 200 250 Time in Dark(h) Fig. 1.4. Sinking rate time series of D. brightwellii cultures placed in the dark and monitored over the subsequent 226 h. Subsamples were taken from the dark culture, and sinking rate measurements were made over 3 h, under high light (LT), in the dark (DK), or in the dark with KCN added as a respiratory inhibitor. All values represent means of three cultures + 1 S.E. KCN sinking rates are plotted also as a mean sinking rate over the entire 226 h. This mean is defined as the maximum sinking rate (MSR). Hatched area represents + 1 SD for the MSR. 30 •o E. tr c • • M B J * c en 0 2 4 6 8 Cell Volume (^m3 x 103) Energy State Time In Resp'n Dark Rate (f g C ^nr3 h1) 226 h 0.001 10 0.225 0.318 0.638 Fig. 1.5. Size-specific sinking rates of D. brightwellii cultures grown at high light, placed in the dark, and monitored over the subsequent 226 h, with sinking rates measured periodically over 3 h in the dark (DK). The four lines represent the size-sinking relationship at 0, 24, 36, and 226 h after cultures were placed in the dark. Estimated respiration rate given for each sampling time is calculated from the change in culture carbon content. time (see text for details). over 31 0.2 0.4 0.6 Estimated Respiration Rate (fg C Mnr3 h-1) Fig. 1.6. Sinking rate (m d"*) of Ditylum brightwellii cells placed in the dark and monitored over 226 h vs respiration rate (RR) estimated from the change in carbon content in the same cultures over time according to the following equation: RR = 0.638e-0.029t> where t = time (h) after cultures were placed in the dark (see text for details). 32 Experiment 3: Other species Cells of Thalassiosira weissflogii grown at all irradiances had similar sinking rates when measured in the light (0 h in dark, Fig. 1.7A) and after 9 h in the dark. All cultures showed an increase in sinking rates over the first 9 h in the dark, but there were no significant differences between cultures grown at different irradiances before this time. When each value (n > 3) was compared separately, cells at 10 |imol quanta m"2 s"l consistently had sinking rates significantly higher (2 x) (t-test, p<0.05) than the other cultures after 24 h of darkness. When values were pooled over all times in the dark, cells grown at 10 p.mol quanta m"2 s"* had significantly higher sinking rates than cells grown either at 100 or 40 |J.mol quanta m~2 s'-k Under KCN, all cultures showed an increase in sinking rates (about 2 - 3 x). The pattern for Thalassiosira pseudonana differend from those of T. weissflogii and D. brightwellii. Cells at the two highest irradiances had the highest sinking rates in the light, while cells at 50 and 10 umol quanta m"2 s"* had lower sinking rates (Fig. 1.7B). The two highest light cultures showed a steady decrease in sinking rate over 100 h in the dark. The two lowest light cultures showed a sinking rate increase over the first 6 to 9 h in the dark, after which sinking rates remained relatively high and constant (Fig. 1.7B). 33 0.20 0.15 0.10 0.05 •b 0.00 g 0.15 o S * 0.10 Thalassiosira weissflogii f o 100 AT Thalassiosira pseudonana o 160 4 90 ~r~i B 20 40 50 80 100 Time in Dark (h) Fig. 1.7. The mean sinking rates of two diatoms grown at various, near-saturating (open symbols) and light-limiting (filled symbols) irradiances (expressed as umol quanta m"2 s-1), in the light (0 h) and after various periods (6 to 100 h) in the dark. Sinking rates recorded at 0 h were measured at the irradiance at which cultures were grown, while the measurements made after 6 -100 h in the dark were made over 3 h in the dark. Error bars represent + 1 SD of three or more independent measurements. (A) Thalassiosira weissflogii, all irradiances. (B) Thalassiosira pseudonana, near-saturating irradiances. (C) T. pseudonana, growth-limiting irradiances. 34 DISCUSSION Ditylum brightwellii In the steady state experiment, sinking rates of Ditylum brightwellii were inversely proportional to intracellular energy availability. While sinking rates of low light grown (LL) cells increased over 4 h in the dark, the sinking rates of high light grown (HL) cells were not affected by 4 h of darkness. When treated with KCN, sinking rates of HL cells increased and became similar to or greater than sinking rates of LL cells in the dark. This confirmed the hypothesis that in the dark, respiratory energy was necessary for the maintenance of low sinking rates. In the time-series experiment, sinking rates of Ditylum brightwellii also covaried with energy available to cells. Initially, the range of sinking rates among cells at different energy levels was large (0.6 m d-1), suggesting that an operative ion pump could modify sinking rates by this amount. By the end of the experiment, LT, DK and KCN sinking rates had converged, indicating a loss of cellular control of sinking rate. Plotting sinking rate of cells in the dark (DK) vs estimated respiration rate provided the first evidence that there is a negative exponential relationship between cell sinking rate and cellular energy. In addition, the time series of cell sinking rates in the dark defined the period over which high light grown (HL) cells could maintain low sinking rates without a source of light energy (24 h). The time series for the sinking rates of cells exposed to 3 h of light (LT) at various times during the dark experiment indicated that cells could be kept in the dark for 78 h and still recover low sinking rates given a sudden input (3 h) of light energy. For a further 70 h (to t = 120 h) cells given 3 h of light exposure reduced their 35 sinking rates significantly below the DK sinking rate, but not to the initial LT rate. The fact that the LT sinking rate increased over time suggests that the sinking rates were not responding directly to the increased availability of photons alone, but were responding more generally to the cell's overall level of metabolic activity. Other species Thalassiosira weissflogii and Thalassiosira pseudonana showed patterns of sinking rates which differed markedly from that described for Ditylum brightwellii, although both species showed some evidence of energetic sinking rate control. Clearly, there is a species specific range of responses to energy deprivation. Comparing the two high light and the two low light cultures of T. pseudonana over the first 3 h of darkness does suggest that energetic control might be occurring, since sinking rates increased for the low light, but not the high light cultures, over this period in the dark. The steady decrease in sinking rate in the high light cultures, however, is similar to patterns seen for cyanobacteria that have no vacuolar control of sinking (Reynolds et al., 1987), and whose sinking rates (which are directly proportional to cell carbon content) decrease over time in the dark as carbon is respired. It may be that cell composition is a more important determinant of sinking rate for T. pseudonana than for the other species in this study, a factor possibly related to its small size (see below). The sinking rate pattern shown by Thalassiosira weissflogii seems to indicate that although energetic sinking control is more evident for this species than for Thalassiosira pseudonana, sinking rates are not controlled by energetic processes alone. Both the higher sinking rates under KCN and the higher overall rates of the lowest light cultures indicate 36 that energy-deplete cells tend to sink more rapidly than energy-replete cells. However, the similar fluctuations of sinking rates over time in the dark seem to indicate that processes common to all cultures also affect sinking rates. In this case too, cell size and carbon content may also be important. Cell size and sinking rates Brun-Cottan (1976) suggested that, according to Stokes' law, there was a cell size (4 -5 um diameter) below which sinking rates became insignificantly small, and cells would be effectively conserved within a given water mass. Villareal (1988) suggested that there was a cell size below which positive buoyancy could not occur in Rhizosolenia debyana, because the internal cell volume was insufficiently large to offset the weight of the heavy silica frustule. If both observations hold, then cells brought up to the mixed layer might have a choice of two strategies: They could become very small (< 5 um diameter) and thus remain passively retained at the surface by oceanic turbulence, or, they could become large enough to engage in active sinking control through the formation of large cellular vacuoles. If large cells are more able to control their sinking rate actively than small cells, larger cells should show a greater energy dependence than small cells. In addition, while the sinking rates of small cells would be directly related to cell composition, the sinking rates of large cells would be more directly related to their energetic state. Cellular composition would still determine cells' extra-vacuolar intracellular density and which in turn would determine the cells' maximum sinking rate in the absence of energetic control. 37 Thalassiosira pseudonana cells are very small (ca. 5 urn in diameter). In such a small cell, the decrease of carbon content over time in the dark might have several effects, decreasing available energy but also measurably decreasing cell density. In this study, for energy-saturated cells, only the sinking rate decrease caused by the density decrease seems to have been measurable. For the energy-deplete cells, the loss of energy seems to have been the more important determinant of sinking rate. Thalassiosira weissflogii is about 10 - 20 |im diameter, an intermediate size between the two other species. This species seems to have an energy/sinking dependence intermediate between D. brightwellii and T. pseudonana. The largest and smallest diatoms may thus define extremes of a wide spectrum of energy/sinking responses. Similar physical constraints operating on an intraspecific level suggest that a species might need to vary its size under different environmental conditions. For instance, as cell size decreases, the range of possible sinking (and floating) rates should decrease. This study confirms that this is the case for Ditylum brightwellii. At saturating energy levels, sinking rates were negatively correlated with cell size. It was only the cells at the lowest energy states, in both the steady state and time-series experiments, that showed the significant positive correlation between sinking rates and cell size predicted by Stokes' Law, and the range of sinking rates thus covaried with cell size. However, Stokes' Law predicts that sinking rate should be proportional to r2, the square of the equivalent cell radius. When data were plotted as equivalent radius vs sinking rate, sinking rates were roughly proportional to r, not r2 (indicated by the fact that the exponent for r was 0.7, rather than 2). Smayda's (1970) study of the cell size vs sinking rate relationship among many species yielded a similar relationship (r ~ 1; calculated in Walsby and Reynolds 1980) to that found in this study. This deviation from theory probably indicates that other cell properties (cell 38 density and/or "form resistance") covary with cell size. This is not surprising, since the surface area to volume ratio (and hence the silica:cytoplasm ratio) changes markedly with cell size (Villareal 1988), and form resistance and overall drag increase with cell size (Walsby and Reynolds, 1980; Kamykowski et al., in prep). It is also possible that larger cells have larger vacuoles and thus a slightly lower carbon content per unit cell volume. Large cells increased their sinking rates by a factor of 10, and small cells by a factor of 4, in response to a reduction in respiratory energy availability. In addition, over time in the dark, big cells sank faster and sooner, than small cells. This factor may be important to consider when discussing diatom sinking strategies. Early in a bloom, there may be an advantage for cells possessing the lowest sinking rates (Chapter 3, this study), in this case, the large cells of the population. One could speculate that increases in cell size (Joint et al. 1987; Thompson 1991) may have evolved to improve cellular retention in the photic zone early in a bloom. As shown in this study, a 30% increase in mean cell size was enough to lower the mean sinking rates of Ditylum brightwellii significantly under nutrient and light saturating conditions. This shift in size is easily achievable by many species (Thompson 1991). This advantage to large cells of slowed sinking would only hold as long as energy was available for the control of sinking rates. Results from this study suggest that if energy became limiting, large cells would rapidly become the fastest sinking fraction of a population (within 36 h in this study). Averaged over the entire 266 h of Exp. 2, the minimum sinking rate of the entire population occurred in cells of an intermediate size (-2,000 um^), which may represent an optimal intermediate size for the maintenance of low sinking rates under energy limitation. 39 Energy and sinking rates Smetacek (1985b) hypothesized that diatoms were originally benthic organisms that evolved a pelagic phase, allowing them to capitalize on a temporally limited window of simultaneous abundance of high nutrients and high light in the euphotic zone, every spring and fall. Since diatoms' silicate frustules are denser than seawater (e.g. Smayda 1970; Villareal i988), diatoms must sink in the absence of any physiological buoyancy mechanism. Certain species may have reduced their maximum sinking rate by becoming small. The only buoyancy mechanism demonstrated to date in large marine diatoms is the ionic pump documented by Anderson and Sweeney (1977). Ionic pumps require energy. The development of low sinking rates in large marine diatoms must therefore have involved expenditure of energy by the cell. If the relationship between energy availability and sinking rates documented here for D, brightwellii also applies to other large diatom species, it has some interesting implications for Smetacek's hypothesis of diatom evolution. A species' sinking responses to energy availability could define that species' temporal "window" of residency in the euphotic zone. As long as both light and nutrients were saturating early in the spring, diatoms could maintain low sinking rates through an expenditure of available energy (Fig. 1.8, panel A). At night, the availability of stored energy should allow a diatom cell to avoid short term energy deprivation. Diurnal variability in phytoplankton sinking rates is well documented in diatoms (Anderson and Sweeney 1978; Bienfang and Szyper 1982). From the results presented here, it would be expected that cells grown under saturating light would have little or no diurnal variability in sinking rate, while cells grown under limiting irradiances would be expected to have significantly higher sinking rates at night than in 40 BLOOM NUTRIENT INITIATION DEPLETION t L_ Fig. 1.8. Conceptual model of diatom sinking control during a spring bloom. (A) Light and nutrients are saturating, and cell sinking rates are limited primarily by light availability. During days of low light, at night, and during wind events, cells depend on stored C as a source of energy to reduce sinking rates. CB) Nutrients are limiting. Cells may either not respond to nutrient depletion (1), or respond quickly, leaving the photic zone (2). Upon reaching the nutricline, cells may be able to spend their C reserves to utilize nutrients, to reduce sinking rates and to grow. Once C reserves are depleted, cells would sink rapidly to the benthos. 41 the day. Granata's (1991) data on this question are difficult to interpret given his sample size (n = 1). Cell division cycles may, however, cause changes in cell size and composition (Anderson and Sweeney, 1977; 1978) which could produce sinking rate differences over a diurnal cycle, complicating any energy-sinking relationships. In general, the ability to utilize stored carbon to control sinking rates should restrict nocturnal cell losses from the euphotic zone if light is the only factor limiting phytoplankton growth. Light can also become limiting during a bloom under periods of low insolation, or because a temporary wind-induced deepening of the mixed layer lowers the average irradiance experienced by a cell (Fig. 1.8, panel A). Under any of these conditions, a cell would depend on the availability of stored carbon as an energy "buffer" for the maintenance of low sinking rates. If D. brightwellii became temporarily light-limited, it could maintain its position in the water column for at least one day, even if it were below the compensation depth, by virtue of its low sinking rates. Other species with a less pronounced energy-dependent sinking control would show a much different response to energy limitation. For example, they might remain in the water column for a longer period of time after the larger species had exhausted their energy reserves. There may be a spectrum of possible responses to nutrient limitation (Fig. 1.8, panel B) for diatoms between the following two extremes. One response might be decreased sensitivity to nutrient limitation (in coastal marine environments, primarily by NO3) during a bloom, leading to continued persistence in the euphotic zone, and a long temporal "window" (Fig 1.8, panel B-l). The other response might be high sensitivity to the limiting nutrient, with a rapid increase in sinking rate when the nutrient becomes limiting, leading to a very short temporal "window" of residency in the euphotic zone (Fig. 1.8, panel B-2). In the latter 42 strategy, rapid sinking would last only until higher nutrients are encountered at the nutricline. If the nutricline is above the compensation depth, cells could utilize light energy at the nutricline for three processes: to take up the limiting nutrient, to reduce sinking rates, and to grow. As cellular utilization of nutrients caused the nutricline to move below the compensation depth, the sinking cells would continue to respire, sink, and grow, driving the nutricline deeper until all energy sources were exhausted. They would then rapidly sink from the surface layers. Skeletonema costatum might exemplify the first strategy since it seems to be relatively insensitive to nitrate deprivation during a high latitude spring bloom (Chapter 2, this study), while Thalassiosira aestivalis, which was observed to be sensitive to nitrate depletion, appears to follow the second strategy (Chapter 2, this study). The dynamics of cellular processes near the compensation light depth, especially of cellular respiration, may be important factors to consider in models of biogeochemical fluxes (Falkowski 1988). Some evidence exists that nutrient-depleted cells have lower ATP levels than nutrient-replete cells (Noges, 1991). This suggests that nutrient limitation might limit intracellular energy availability. Recent results describing the dependence of cell sinking rates in the field on ambient nitrate concentrations (Chapter 2, this study) may thus be explicable through some dependence of cell respiration on nutrient availability such as that documented by Geider and Osborne (1989), although changes in biochemical composition could also be important. In general, for diatoms sinking after the spring bloom, a critical variable to determine in order to understand the sinking process may be the effect of nutrient depletion on the availability of photosynthetic and respiratory energy. If so, one should be able to define any diatom species' entire sinking strategy, from the initiation of a 43 bloom to its termination, by documenting the dependence of sinking rates on intracellular energy availability, limited by light, nutrients, or a combination of both. 44 CHAPTER 2. SPRING BLOOM SEDIMENTATION IN A SUBARCTIC ECOSYSTEM I. NUTRB5NT SENSITIVITY INTRODUCTION Most spring diatom blooms are terminated when nutrient depletion occurs in surface waters, and bloom biomass sediments rapidly from the photic zone (e.g., Davies and Payne, 1984). Although it is often assumed that bloom sedimentation is triggered by the increase of cell sinking rates at the surface during nutrient depletion, the dynamics of this sinking rate response to nutrient depletion have never actually been documented during natural blooms. The relationship between surface nutrients and cell sinking rates may, however, be the feature of a bloom population most likely to summarize its sedimentation potential, and hence its potential contribution to coastal carbon flux. In laboratory studies, rapidly growing species' sinking response to nutrient depletion have been highly variable (Bienfang, 1981» & b; Bienfang and Harrison, 1984b). Depending on the degree to which a species utilizes energetic control of sinking rates (Chapter 1, this study), nutrient depletion might effect changes in cell sinking rates either through the incapacitation of metabolic pathways or through changes in cell composition (Chapter 1, this study). Both probably occur. In the end, different species' response to nutrient depletion during a bloom will depend both on their nutrient demand for growth and on their mechanism (energetic or non-energetic) of sinking rate control. The sinking-nutrient 45 response should in turn determine the nature of a given species' intrinsic sedimentation potential. This work was part of the five-year APPRISE (Association of Primary Production and Recruitment in a Subarctic Ecosystem) Program which focused on the natural variability of production and sedimentation processes during the spring blooms in Auke Bay, Alaska between 1985 and 1989. The objectives of this study were: 1) to measure the species specific sinking rates of the principal diatoms during the spring bloom, 2) to quantify the role of nutrients in modifying these rates, 3) to describe the vertical and temporal patterns of sinking rate variability associated with nutrient depletion during and after the spring bloom in Auke Bay, and 4) to make a preliminary assessment of the importance of other factors such as ambient temperature, water density, ambient light and light history, in determining sinking rates. Sedimentation patterns are discussed in the next chapter. MATERIALS AND METHODS Auke Bay (58o 22'N, 135° 40"W) is a shallow (-50 m) sub-arctic embayment located about 20 km north of Juneau, Alaska (Fig. 2.1). The area has a temperate maritime climate with high rainfall, low insolation, and a mean tidal range of 4 m. The spring diatom bloom is usually initiated by increasing light levels in the spring, coinciding with the onset of spring stratification and terminated by nutrient exhaustion in the surface waters (Ziemann et al. 1990). From March to June, the project sampled the water column biweekly for irradiance, CTD, nutrients, and biogenic particulate matter (POC, PON, CHL) which provided 46 Figure 2.1. Map of Auke Bay, Alaska, showing the location of the study area in the SE Alaskan Archipelago and the location of the sampling station in the bay (X). 47 oo --\58°22'N 135° 44* W background information for the sinking investigations. Water sampling for sinking rate and cell count evaluations was done biweekly using a 5 L Niskin bottle. Samples were taken from the surface (2 m) during all years, and also from the depth of the chlorophyll maximum in 1988 and 1989. In 1985-1987, replicate subsamples were taken from single casts, and in 1988-1989, single samples were taken from each of two replicate casts to permit a proper expression of the mean and variance from a given species and sampling date. Cell count subsamples were preserved in formalin and counted using the Utermohl (1958) technique. Sinking rates were determined by the SETCOL method (Bienfang 1981c) using cell counts as a biomass index for the calculation of species-specific sinking rates. The sinking columns were enclosed in a water jacket and located within a temperature-controlled room to maintain constant temperature and prevent convection. Nutrient samples were filtered (Whatman GF/F filters) frozen, and analyzed for NO3, Si04 and PO4 on a Technicon AutoanalyzerR (Armstrong et al. 1967; Hager et al. 1968; Technicon, Inc. 1977). For the evaluation of general trends in sinking/nutrient dependence, the sinking rate time series were separated into two groups, the nutrient-saturated (NS) and nutrient-deplete (ND) periods. The NS and ND periods were defined as prior to, and following, the measurement of non-detectable surface (2 m) nitrate concentrations, respectively. Mean sinking rate values for these two periods represent the temporal averages of up to 14 daily means. Before assessing significant differences between periods, the daily means were examined for temporal trends, normality and heteroscedascity (homogeneity of variances). Data showing a significant temporal trend were not subjected to evaluation of differences between periods. Data having normal distributions and similar variances were subjected to t-tests against two null hypotheses: that there were no significant (p<0.05) differences in mean sinking rates 1) between the NS and ND periods, and 2) between the surface (2 m) and 49 the chlorophyll maximum during the ND period. Data which were severely non-normal or showed heteroscedascity were subjected to the Mann-Whitney U test. In 1988 variations in cell aggregates (clumps), chain length and intracellular nitrate pools of the resident phytoplankton were also assessed. Microscopic analyses of aggregates included assessment of numerical abundance and size (cells/aggregate) over time. Because large fragile aggregates are disrupted by bottle sampling, this assessment represented only some estimate of first order cell clumping. It was not a true estimate of large-scale aggregation. Chain length was assessed in each biweekly sample by counting the number of cells per chain in up to 50 chains of the principal diatom species. Data were pooled over time to yield monthly frequency distributions. Chain lengths were compared between: A) the surface and the chlorophyll maximum, and B) between the entire population and the fractions which settled preferentially during sinking rate trials. Comparisons between frequency distributions were made using the Chi-Square test. Intracellular nitrate pools were measured on samples from 2 m and the chlorophyll maximum using the osmotic shock method of Thoresen et al. (1982). Duplicate 500 mL samples were filtered (Whatman GF/F filters), washed with 20 mL of 3% NaCl solution, extracted with 25 mL of boiling double-distilled deionized water, and frozen for later analysis of nitrate. Sample results were normalized to PON values (Collos 1982). For the single most abundant species, Thalassiosira aestivalis, five years of sinking rate measurements were analyzed through multiple regression analysis against the following variables: daily temperature, surface water density, ambient nitrate concentration, daily total quantum irradiance on the same day as sinking rate measurement (ItO). one day prior to sinking rate measurement dt-l), two days prior to sinking rate measurement dt-2), and as an estimate of light history, the cumulative total of previous 2 days' irradiance (In = It-1 + It-50 2). Because of obvious colinearity, the cumulative light data could not be run in the same multiple regression analysis as the other light data, and was thus run separately. All data were tested for heteroscedacity and for obvious outliers before being analyzed by forward step-wise multiple regression. Results were presented with and without outliers present for reasons discussed below. Sinking rate and nitrate concentration data were separated into bins on the basis of their light history (total quantum irradiance on day prior to sampling = 0 - 3 , 3 - 5 , 5 - 7 , 7-9, > 9 x 1021 quanta m-2 d-1. Sinking-nutrient regressions were then plotted separately for each light bin, to give a visual indication of the effect of light history on nutrient-sinking relationships. In all cases, significance levels were taken as p < 0.05. 51 RESULTS The five year time series of ambient nitrate at the surface (Fig. 2.2) show the general pattern of initially high concentrations throughout the photic zone followed by a 7 -10 day period of rapid uptake which reduced levels to < 0.10 umol L"l). Profiles of PO4 and SiC>4 (same figure) displayed similar temporal and vertical trends. The exhaustion of surface nitrate, which usually occurred in the latter half of April, was used to distinguish between the initial nutrient-saturated (NS) period and the later nutrient-deplete (ND) period. Over both the NS and ND periods, Thalassiosira spp. displayed the higher mean and maximum sinking rates (p<0.05), than Chaetoceros spp., which had the lowest sinking rates (Table 2.1). Skeletonema costatum had intermediate sinking rates. Over all the years, the pooled mean sinking rates of the three major diatom genera ranged from 0.00 - 1.73 m d-1 and 0.17 - 3.05 m d-1 during the NS and ND periods, respectively (Table 2.1). Observed sinking rates during the ND period exceeded those in the NS period in 75% of the comparisons, and rates during the ND period were up to 2.7x higher than those in the NS period. During both periods, the ranges of pooled mean sinking rates for all species displayed considerable overlap. The sinking rates of all species were generally higher and more variable following nutrient depletion. This is illustrated in the sinking rate time series for Thalassiosira aestivalis (Fig. 2.3). Summaries of the five-year data sets for all species (Table 2.2) show the relative amounts by which the maximum sinking rates during the ND period exceeded those in the NS period. Statistically significant (p<0.05) differences in sinking rates between the NS and ND periods occurred in 6 of 24 comparisons (indicated as superscript "a", Table 2.1). Four of 52 Figure 2.2. Time series of ambient nitrate, silicate and phosphate concentration during the spring (1985 -1989) at the surface in Auke Bay. The 1985 - 86 data are from the 100% light depth (just below the surface); the 1987 - 89 data are from 2 m. 53 MAR APR MAY JUN MAR APR MAY JUN 2.0 r o 1.6 E £ 1-2 a « 0.8 o 0.4 0.0 MAR APR MAY JUN 54 Figure 2.3. Time series of the sinking rates (m d"1) of Thalassiosira aestivalis, over the period of the study (1985 -1989). Values are the mean of two sinking rate measurements. For 1988 and 1989, error bars represent +/-1 S.D. about the mean of replicate samplings and give an estimate of population variance. Vertical dashed lines indicate the onset of nutrient depletion (CNO3] = < 0.01 umol L'l). NS = nitrate saturated and ND = nutrient-deplete. 55 ^-•••vw X.. MARCH APRIL MAY JUNE MARCH APRIL MAY JUNE •D E. LU < CO 6 5 3 2 0 r - 1 1989 NS ND . $ ! " * \ MARCH APRIL MAY JUNE 56 Table 1. Sinking rates (m d"1) of the major diatom species during the spring bloom in Auke Bay, Alaska. Data are pooled mean sinking rates ( + /-SD) during the nutrient saturated (NS) and nutrient deplete (ND) periods. '*" indicates sinking rate time series where ND sinking rates were significantly (p<0.05) higher than NS sinking rates, "f indicates sinking rate lime series where maximum sinking rates were highest in "late" period but the overall difference was not significant, and "•" indicates years in which sinking rates of a species showed no apparent relationship to nutrients. '-' = data not available, or species not present. 1985 1986 1987 cn SPECIES THAIASSIOSFRA T. aestivalis T. gravida T. nordenskioldii SKELETONFMA S. costatum CHAETQCEROS C compnssus C debitis Cmdicans PERIOD NS ND NS ND NS ND NS ND NS ND NS ND NS ND SINKING RATE 0.40 (030)* 1.00(0.90) 139 (1.60)* 1.52 (2.24) -0.25 (0.25)t 0.41 (0.73) ----(n) 10 9 6 8 10 8 SINKING RATE 0.20 (0.70)* 1.30(1.80) -0.70 (0.72)* 034 (0.26) 0.17 (033)t 036 (038) (n) 12 6 10 9 .11 7 SINKING F 0.39 (0.20)* 1.27 (1.01) 0.27 (0.1 l)t 0.42 (0.17) 0.57 (0.72)* 034 (0.26) 0.67 (036)* 036 (0.48) 030(030)* 0.21 (024) 135 (0.45)* 0.99 (0.86) 1988 1989 (n) SINKING RATE (n) SINKING RATE 14 0.42(005)* 5 136(136)* 9 0.967(0360) 10 135(0.97) 11 0.13(0.09)t 2 1.73(2.44)* 6 3.05(2.89) 4 0.70(0.71) 9 - 0.54 (0.17)t 8 Z05 (1.40) 14 0.06(0.12)t 3 0.08(0.11)* 7 1.71(132) 5 0.46(0.92) 13 0.00(0.02)* 2 0.16(0.12)t 10 035(0.79) 6 033(033) 11 0.12(0.19)t 5 0.01(0.03)t 8 0.71(0.73) 13 0.90(136) 0.45 (033)t ; 105 (1.92) these involved Thalassiosira aestivalis, indicating a greater sinking rate sensitivity to nutrients than other species present. Collectively, these data indicate that with the onset of nutrient depletion, Thalassiosira spp. were most likely to display increased sinking rates. They also had the highest sinking rates overall when their rates did increase. Within both NS and ND periods, the sinking rates of the principal diatoms displayed significant variability; for any species, standard deviations were frequently 50 -100% of the pooled means. Overall, variance was largest at low nitrate concentrations typical of the ND period; this is illustrated for Thalassiosira aestivalis in Figure 2.4 (inset). Variance estimates during the ND period were significantly (p<0.05) larger than in the NS period for 4 of 5 major species in 1988, and 5 of 7 species in 1989. In both years, the most abundant species were included in the group whose variability increased significantly. The covariation of ambient nitrate concentrations and sinking rates for two principal diatom species (Fig. 2.4 A & B) over 5 years indicates threshold concentrations (shown by dashed vertical lines) below which elevated sinking rates were most commonly observed. To quantify the threshold, the following method was used: All data points where sinking rates were above the 95% confidence interval of the mean NS sinking rate (solid horizontal line) were selected. The 30% of data points showing the highest nutrient concentrations were then further selected from this group, and the nutrient concentrations at which they occurred were averaged to yield the estimated threshold. In this way the effect of outliers (as in 1989 for Thalassiosira aestivalis; Fig. 2.4A) was minimized, and the inaccuracies involved with using data at non-detectable (nd) nutrient levels (along the y-axis) was avoided, since "non-detectable" nutrients represented nutrient concentrations anywhere from 0 to 0.1 |imol L-1. This latter problem was one of the principal reasons that efforts to fit curves to the data for estimation of threshold concentrations were not successful. For Thalassiosira aestivalis, 58 Figure 2.4. Scatterplots showing the relationship between ambient nitrate concentrations and the sinking rates of A) Thalassiosira aestivalis and B) Skeletonema costatum during 1985-89. Vertical lines dashed represent the threshold nitrate concentrations levels below which higher sinking rates were most frequently observed. Solid horizontal lines represent the upper limit of the 95% confidence interval for the nutrient saturated sinking rates. Open circles denote year when no clear relationship existed between ambient nitrate concentrations and sinking rates. Inset in A shows 1988 - 89 data as error bars only; note higher variance at lower nitrate concentrations. 59 E. UJ < DC (3 Z 5 z en 6 5 -4 -•i A. • • 1985 • 1986 • 1987 A 1988 O 1989 4 -3 -2 -1 -oo -• .• p 4H 3H 21 B. • 1985 • 1986 c 1987 A 1988 • 1989 _a t r ~ A ^ 5 1 0 15 20 25 NITRATE CONC. Cumol L1) 30 60 the estimated threshold concentration (~2.0 umol L_1 NO3) was consistent over all 5 years of the study, and similar to Ks uptake values from the literature (Table 2.2). Ks values are the half-saturation constant for nutrient uptake according to Michaelis-Menten uptake kinetics. They indicate the nutrient concentration at which concentration-dependent uptake is half its maximum value. All Thalassiosira spp. showed a similar pattern of sinking rates with respect to nutrient concentration over all 5 years. Nitrate sensitivity was lower for Skeletonema costatum, (1.4 umol L" 1) but was observed in only 4 of 5 years (Fig. 2.4B). For Chaetoceros spp., thresholds were more variable (0.6 to 2.4 umol L"l), and were observed only in 2 of 3 years in which their sinking rates were measured. In 1988 and 1989, sinking rates were also measured on samples from the chlorophyll maximum layer (CMX). The pooled CMX sinking rate means were generally not significantly different from either NS or ND sinking rates, but were generally closer to NS rates, and lower than ND rates in 8 of the 12 comparisons made (Fig. 2.5). Significant decreases in sinking rates at the CMX occurred only for the most nutrient-sensitive species during their year of higher nutrient-sensitivity (Thalassiosira nordenskioldii (1989) and Thalassiosira aestivalis (1988);(Fig. 2.5). During the ND period of 1988 the size of intracellular nitrate pools of cells at the surface (averaged over the four weeks of the ND period) was compared with cells constituting the CMX biomass over the same (Fig. 2.6). T. aestivalis formed over 99% of the bloom biomass during this period. Cells at the surface had intracellular nitrate pools which ranged from 0.1 - 0.4 % of total PON, with actual pool concentrations ranging from 0.62 x 10-2 to 3.6 x 10-2 pmol NO3 per cell. Cells from the CMX showed intracellular pools which were more than an order of magnitude higher than cells at the surface, averaging 7.2 10-2 pmol NO3 61 E •2 2 o en 1988 CZD SURFACE eza CMX 5 1 -c 175 o 4' k To* — « — T n LI *B2_ Cct CdT Cr r -£ 2 o o> c c 1989 C D SURFACE eza CMX „ 11 E. To Tgt Tn* Sc S p e c i e s Figure 2.5. Comparisons of the average sinking rates of various species at the surface and the chlorophyll maximum (CMX) during the nutrient deplete period of 1988-1989. Error bars represent one standard deviation about the mean. "*" indicates comparisons where surface (2 m) sinking rates were significantly higher than CMX sinking rates, " f" indicates comparisons where highest sinking rates.occurred at the surface but differences were not significant. Ta = Thalassiosira aestivalis, Tg = Thalassiosira gravida, Tn = Thalassiosira nordenskioldii, Sc = Skeletonema costatum, Cc = Chaetoceros compressus, Cd = Chaetoceros debilis, Cr = Chaetoceros radicans. 62 - 00 C) E x m -f-ro CNJ CVJ (Nd 1V101 %) "lOOd r 0N 1VNU3JLNI Figure 2.6. Comparison of the sizes of internal nitrate pools for cells (>99% Thalassiosira aestivalis by biovoW) at the surface and the chlorophyll maximum during the nutrient deplete period of 1988. Error bars represent the mean of two independent measurements. 63 2 Tabic 2. For the niitricnt-saluralcd (NS) and nutrient-depleted (ND) periods, mean sinking rnie values (mil-') represent the average ( + /-SD) of up to five individually pooled yearly means for each period. Calculated maxima give upper hounds of the 95% confiiU'iicc intervals alxml I lie- means of sinking rate measurements during the NS period for all five years, lite observed maxima represent the highest single measurement made on each species during the five years. Nitrate threshold values represent the nitrate concentrations (umol L'1) below which sinking rales appeared to increase. For each genus, Ks values (umol L"') from the literature are given for comparison with the calculated threshold levels. SINKING RATES NITRATE SPECIES NS PERIOD ND PERIOD Calculated Observed Threshold # Years Mean Maxima Mean Maxima Observed Ks Ks Reference THAU\5SIQS!RA T. aestivalis T. gravida T. nordenskioldii SKELETONEMA S. costatuin CHAETQCERQS C. compressus C. debilis 0.30(0.18) 0.20(0.10) 0.54 0.14(0.25) 0.08(0.12) 0.06(0.18) 1.06 0.55 0.69 0.55 0.23 0.28 1.18(0.18) 1.42(1.18) ' 0.91(0.99) 0.70(0.57) 0.36(0.17) 0.87(0.14) 5.44 8.08 3.40 4.76 2.24 4.88 2.1 1.1 2.5 1.4 0.6 2.4 5/5 5/5 5/5 4/5 2/3 2/3 2.3-2.5 (Dorlch, 1980) 0.5-0.9 (Falkowski, 1975) (Eppley et al., 1969) 0.8-4.4 (Dortch, 1980) per cell, and constituted 2 - 3% of total PON. Throughout the ND period in 1988, the variance of pool size at both depths was low (standard deviations were generally within the range of symbols shown). Decreases in sinking rate with depth averaged over the ND period in 1988 thus coincided with decreases in cell nutrient stress as measured by an increase in intracellular nitrate pools. There was however, no significant correlation in day-to-day changes in sinking rates over time at one depth and internal nitrate pools at the same depth. In 1988, an evaluation of variations in chain lengths for Thalassiosira aestivalis and Skeletonema costatum showed right-skewed normal frequency distributions of chain size. Frequencies generally showed peaks between 1 and 10 cells chain-1, and occasionally included chains with as many as 50 cells (Fig. 2.7). T. aestivalis had the longest chains of any species examined. There was a general progression over time from short to longer chains as the spring bloom developed (Figs. 2.7 A & C). Over the same time frame, T. aestivalis showed significantly longer chains at the CMX (median test, X2=173, df=l, p<0.01) than at the surface. Although mean chain lengths were also longer at the CMX for S. costatum, insufficient CMX data were available for statistical tests. Overall, the distributions of chain length had significantly different shapes between settled and unsettled samples for both species (Chi Square test, X2= 173, p<0.01 for T.a.; X2=64.8, p<0.01 for S. costatum). However, the actual median chain length was significantly different only for S. costatum, where settled chains were significantly shorter than the original sample (X2=104, p«0.001) (Fig. 2.7 G & H). This indicates that for T. aestivalis, chain length did not affect sinking rates, while chain length had a significant effect on the sinking rates of S. costatum. The abundance of aggregates was linearly correlated (r2=0.99 (surface), r2=0.97 (CMX), p < 0.05) with cell concentration. At both the surface and the CMX, maximum 65 laitialPopvlatiaa Thalassiosira aestivalis Fraction that a u k la SETCOL MARCH 2 m A 201 « » t | , 8»3l sjgj 0 30 M 5 » z W 3 a « i I . B . -MARCH 2 m ( i t L j B APRIL 2-m ( iaL) D APRIL Chi. Max. («et.) F - • * * - B -Skeletonema costatum 20 3 10 i s 20 23 JO M *a O JC CHAIN LENCTH APRIL 2 m G 0 3 10 IS so 23 30 JS *0 o 30 CHAIN LENCTH APRIL 2 m ( ict . ) H O 3 to IS 20 a 30 33 40 43 i c CHAIN LENCTH Figure 2.7. Chain length frequency distributions. A, C & E show distributions of the initial populations of Thalassiosira aestivalis used in the SETCOL trials at various times. B, D & F show the distribution of the biomass of T. aestivalis which settled during SETCOL trials. G & H show the distributions of Skeletonema costatum for the initial and settled biomass, respectively. 66 MARCH APRIL MAY JUN Figure 2.8. A. Time series of numerical abundance of aggregates (clumps) of Thalassiosira aestivalis cells, at the surface and at the chlorophyll maximum (CMX) over the spring bloom, 1988. B. Mean cell concentrations of T. aestivalis at 2 m and the CMX over the same period. Error bars represent standard deviations of two replicate water samples. 67 aggregate concentration (clumps mL"1) occurred during periods of highest cell concentration (Fig. 2.8), and significantly higher aggregate concentrations were observed at the chlorophyll maximum than in the overlying surface waters. Thalassiosira aestivalis was the only species observed to aggregate. The size of aggregates at the chlorophyll maximum was also significantly larger than at the surface. At both depths, the largest aggregates occurred in May, almost two weeks after attainment of maximal biomass concentration, but systematic covariation of aggregate size over time or with cell concentrations were not apparent at either depth. The sinking rates of aggregates were variable and generally equivalent to those of non-aggregated cells. Pooled monthly average sinking rates for the aggregates from surface samples were 0.21, 1.00 and 0.01 m d-1 for April, May and June respectively. At the chlorophyll maximum, average aggregate sinking rates were 0.50 and 0.40 m d-1 during April and May, respectively. In surface samples, there was a significant (p<0.05) increase in aggregate sinking rate between April and May, during the decline of surface biomass. Aggregate sinking rates were not significantly correlated over time with aggregate size. When multiple regression was performed on the sinking rates of Thalassiosira aestivalis (with the independent variables water density, temperature, ambient nitrate concentration, total daily quantum irradiance (same day dtoX 1 day before sinking measurement dt-l), 2 days before sinking rate measurement (It-2), and cumulative irradiance on days d t - l + It-2))» neither temperature nor sigma-t was a significant predictor of the sinking rates of T. aestivalis. Initial results (outliers not excluded, n = 104) using all data indicated that the primary sinking rate predictor was It-1, followed by ItO and ambient nitrate concentrations. It-1 was positively correlated with sinking rates (p = 0.40); ItO was negatively correlated with sinking rates, as were ambient nitrate concentrations. The 68 covariance matrix indicated that none of the predictor variables in the model were highly cross-correlated; all regression coefficients between light and nutrients were below 0.5 (r2 < 0.25). The initial model predicting sinking rates from ItO, It-1 and ambient nitrate concentrations was the following: SR = 0.143 It-1 - 0.216 ItO - 0.029 N -1.181 where SR = sinking rate, N = ambient nitrate concentration, and I = quanta x 1017 m-2 d-1. The coefficients indicate the relationship of each variable with sinking rates when each other variable is being held constant. For example, the coefficient for N indicates the amount SR would increase (in units of standard deviation of N) for a given decrease in nutrients, if both Io and It-1 were held constant. Please note that the absolute magnitude of the regression coefficients is not directly comparable between light and nutrients since they are measured in different units. This model was highly significant (p = 0.001), though it explained only 14.3% of the variance in sinking rates observed in this study. When 4 outliers were identified, all were instances of very high sinking rates (3 - 5 m d"l), three of which occurred at simultaneously low nutrients (< 2 umol L.-1), high It-i (1.3 -5.8 x 1017 quanta m-2 d-1) and low salinities (~ 18 o/oo) in mid May (1986 and 1989). Re-running the multiple regression without these four outliers (n = 100), N superceded It-1 and ItO irradiance as predictors of sinking rates, ItO irradiance remained significant but It-1 irradiance was no longer a significant predictor. It seems that the original relationship between SR and It-1 was driven by a few very high sinking rates that occurred after 1-2 days of highlight. The model was as follows: SR = -0.028 N - 0.126 ItO - 1-190 69 The new multiple regression had a higher significance level than the old (p=0.000), and explained 16.1% of total sinking rate variance. Sinking rate / nitrate data were separated into bins on the basis of their irradiance pre-history (Ih). Five "bins" were created, where Ih = It-1 + It-2 = A) < 3 x 1021 quanta (over 2 d), B) 3 - 5 x 1021 quanta (over 2 d), C) 5 - 7 x 1021 quanta (over 2 d), D) 7 - 9 x 1021 quanta (over 2 d), and E) > 9 x 1021 quanta (over 2 d). Within each bin, SR was regressed against ambient NO3 concentration (Fig. 2.9). Only bin A (< 3 x lO*? quanta) showed a significant linear relationship (p <0.05) with ambient NO3 concentration. The relationship was the following: Sinking Rate = -0.031 [NO3] + 0.910 The relationship explained roughly 34% of the sinking rate variance. At Ih values higher than 3 x 1021 quanta, the relationship was nonlinear and roughly equivalent to the pattern shown in Fig. 2.4. 70 B <x> o "5 c « <D <2> 13) -i r — i — D •A i t t i "i ' i O 5 10 15 20 25 0 5 10 15 20 25 30 4 -2 -• 0 5 10 15 20 25 30 Nitrate Concentration (^mol L*1) Figure 2.9. Nutrient-sinking relationships for Thalassiosira aestivalis cells with different light histories dh) defined as the total incident quantum irradiance over the two days prior to sampling. (A) Ih < 3 x 1021 quanta m-2, over 2 d. Significant regression line shown with 95% confidence interval. (B) Ih = 3 - 5 x 1021 quanta m-2 (over 2 d). (C) Ih = 5 7 x 1021 quanta m-2 (over 2d). (D) Ih = 7 - 9 x 1021 quanta m-2 (0Ver 2d). (E) Ih = > 9 x 1021 quanta m-2 (over 2d). Circled values represent statistical outliers. 71 DISCUSSION The close proximity of the benthos and the photic zone in shallow embayments such as Auke Bay allows for inoculation of the surface waters in the spring with bloom species from the sediments, as well as efficient post-bloom transfer of sinking biomass to the sediment. Cold, stormy winters produce a well mixed, isothermal water column, with high nutrient levels in early spring, as well as relatively small stocks of over-wintering zooplankton. The lack of substantial grazing pressure in early spring allows the attainment of very high phytoplankton biomass levels. The result is an uncoupled system in which most of this biomass sinks en masse to the bottom. The bloom is initiated by photic windows comprised of brief (i.e. 5-7 day) periods of sunny weather during a time of rapidly increasing day length; this creates a short-term growth opportunity for species capable of a rapid response. Both their growth rate- and sinking rate- responses to conditions at the surface contribute to the competitive success of the spring bloom diatoms (Smetacek 1985b). Growth parameters such as a high u, umax, and Pmax, especially at low temperatures, allow a species to respond rapidly to the high light episodes shown to trigger the bloom (Ziemann et al. 1990), and efficiently assimilate inorganic nutrients. By depleting surface nutrients, they change the environment to one favouring the onset of increased sedimentation. Factors determining sinking dynamics include both the sensitivity of a species' sinking rates to nutrient exhaustion, and the relative magnitude of the subsequent sinking response. It is the sinking factors which are of interest in this study. 72 Nutrients and sinking Given the low ambient concentrations of all nutrients in the ND period, the focus on nitrate as an index of nutrient deficiency associated with increased sinking rates necessitated some justification. Indeed, the lower levels of multiple nutrients may collectively influence the sinking behavior more than any single nutrient. The rationale for the focus on nitrate comes from several lines of evidence in our study, and studies by other investigators of the APPRISE Project. 1) Within the precision afforded by the sampling interval, nitrate appeared to reach its lowest level several days before either phosphate or silicate. 2) Several characteristics of N-deficient cells were apparent, including: a) the depletion of intracellular nitrate pools, b) the appearance of N uptake in the dark, c) the uncoupling of the co-variation of N uptake rates and alpha (the initial slope of the P vs I curves) values, and pronounced decreases in both Ki (the half-saturation light level) and the maximum uptake rate (Kanda et al. 1989). 3) The quality of the settling biomass provided other indications. Most of the sinking biomass consisted of long chains and intact frustules, while silicate deficiency might be associated with breakage of cell chains, and with frustule deformities (Harrison et al. 1977; Bienfang et al. 1982), neither of which was observed. 4) Finally, sediment trap analyses (Ziemann et al. 1985) indicated similar rates of change for the flux of particulate nitrogen and silicon, and that the N:Si of settled material was similar to initial depth-integrated concentrations of their inorganic forms. This is different from situations in which silicate depletion prevails (Skjoldal and Wassmann 1986, and references therein). Other comprehensive studies of spring bloom dynamics and sedimentation have been executed in various Norwegian fjords (Lannergren and Skjoldal 1976; Skjoldal and 73 Wassmann 1986). In those environments, silicate exhaustion is primarily responsible for senescence of the bloom, and the subsequent sedimentation events. This may be due to differences in the ratios of dissolved inorganic nutrients in the Atlantic and Pacific (Si04:N03:P04 of 60:28:2 umol L-l in Auke Bay vs 7:12:7 umol L-l in the Lindaspollene Fjord, Norway). If bloom diatoms require similar nutrient ratios in both environments, the differences in nutrient availability between these two Pacific and Atlantic coastal environments should drive the latter into silicate, and the former into nitrate depletion. The attempt to quantify the response of the principal diatom species to nutrient depletion rests on two principal lines of interconnected evidence: 1) Significant differences in surface sinking rates for a species during nutrient saturation (NS) and after nutrient depletion (ND) in any one year and the magnitude of the difference, 2) the relationship between surface sinking rates plotted vs ambient nutrient concentrations during the spring bloom over several years, paramaterized here as the nutrient "threshold" for sinking. Conclusions based on 1) and 2) are supported or modified with other specific observations, including the occurrence of vertical gradients in sinking rates, as well as the interannual consistency of a given pattern for a particular species. Recent laboratory experiments (Waite and Ledyman, unpubl. data) suggest that T. aestivalis, Chaetoceros compressum and C. debilis significantly decrease their sinking rates when placed in the dark for 24 h. The SETCOL measurements were executed exclusively in the dark, over up to 10 h. It is therefore possible that the sinking rates reported here are actually minimum sinking rates for a given nutrient concentration. This does not change the substance of the arguments presented here, but it may have decreased our chances of observing sinking rate increases. It seems highly important to execute all sinking rate measurements under the ambient conditions. 74 Overall, significant differences in sinking rates between the NS and ND period were more frequently observed for Thalassiosira aestivalis than for any other species. In addition, during the ND period, the sinking rates of T. aestivalis were usually among the highest of any species present. This implies a high absolute amount of sinking rate response to nutrient stress for this species. Attempts to relate sinking rate changes of T. aestivalis to concomitant changes in chain length or aggregation tendency failed to show systematic covariation. This suggests that cell physiology, not particle shape or size, was the proximal factor controlling the sinking rates of this species. The fact that aggregate sinking rates were no higher than single cell sinking rates is especially noteworthy, since most studies assume that aggregation increases the sinking rates of algae. However, it is important to note that because large fragile aggregates tend to be disrupted during bottle sampling, the "aggregation" measured here can only represent a rough estimate of some first order cell-cell adhesion process. We cannot conclude that the formation of marine snow does not increase the sinking rates of diatoms. Our conclusions may have limited applications to other systems, and may hold primarily under optimal growth conditions, for small aggregates made of healthy cells with negligible sinking rates. Significant sinking rate increases in small aggregates may occur only when the sinking rates of most cells in an aggregate increase their sinking rate, which might have been a relatively rare event in this study. Measurable sinking rate increases might also occur once aggregates reached a threshold size (a size of aggregate impossible to sample by our methods). However, in this study, there was also no measurable effect of aggregation on sedimentation patterns (Chapter 3, this study). The ability to respond rapidly on a single cell level to nutrient/light windows in early spring as well as to subsequent nutrient depletion may therefore play an important role in the successful life cycle of such spring bloom diatoms. 75 Similar arguments have been advanced for other spring bloom systems (Davis 1980; Garrison 1981). Comparisons of sinking rates at the surface and the CMX were made in only 2 of 5 years of the study (1988-89) and lower sinking rates at the chlorophyll maximum were observed for Thalassiosira aestivalis in 1988 and Thalassiosira nordenskioldii in 1989. Although only these two of twelve comparisons (6 species over 2 years) showed a significant difference in sinking rates between the surface and the CMX, each species represented a large fraction of the bloom biomass. Thus, over both years, cells that slowed their sinking rate at depth accounted for -62% of the bloom biomass (>99% for T. aestivalis in 1988, and ~23% for T. nordenskioldii in 1989). The coincidence of lower sinking rates and higher internal cell NO3 concentrations (generally understood to be an indication of very short term N03 supply) in 1988 at the CMX suggests that in the short term, at least, sinking rate increases at the surface may be reversible at depth when sinking cells reach higher nutrient concentrations. (This reversal is even more noteworthy when one considers the inverse relationship between surface irradiance and sinking rates discussed below.) Over longer periods of nutrient limitation, aspects of sedimentation may, however, be more analogous to the more complete physiological "shift-down" response described for uptake kinetics (Dugdale et al. 1981, Wilkerson and Dugdale 1987). With this aspect of physiological response in mind, the study will focus next on the approximation of the threshold nitrate concentration which signals the initiation of a sinking event. Although ambient nutrient concentrations are generally related to nutrient uptake rates, there may be a significant lag between changes in uptake rate and the cell's physiological sinking response. This might complicate the relationship between ambient nutrients and cell sinking rates, especially if the environment is spatially patchy (e.g., 76 Turpin et al., 1981). For instance, the cluster of data points where nutrients are below the sinking "threshold", but sinking rates are still low, probably represents cells which have been at low nutrients for a short time, whose physiology has not yet been significantly affected by nutrient starvation. Geider and Osborne's (1989) recent work suggests that dark respiration rates decrease as nutrient limited growth rates decrease. He work used chemostats, and his results are thus not necessarily strictly applicable to spring bloom systems, which are more analogous to batch cultures. However, his observations do suggest that nitrogen depletion could trigger sinking rate increases through interference with respiration. The striking similarity in shape between the nutrient-sinking relationship (Fig. 2.4) and the respiration-sinking relationship (Fig. 1.6) also suggests this might be the case. The similarity of threshold concentrations to Ks values suggests, however, that it is at nutrient concentrations similar to Ks values that nutrient effects on sinking rates begin to be seen. This is not surprising, since one might expect that the onset of nutrient limitation might occur either at the same time, or at some time after, uptake is limited by supply, depending on the nutrient history of a particular cell. The use of nutrient levels in the range of Ks as a physiologically meaningful estimate of the threshold concentration for sinking rate changes provides an approximation which may find useful application in predictive models which attempt to account for sinking losses in systems with changing nutrient fields such as coastal upwellings and oceanic spring blooms. For all Thalassiasira spp., the similarity of their nutrient thresholds to Ks values, as well as their consistency over all five years of this study suggest that these species are, in general, the most consistently nutrient-sensitive species in the Auke Bay system. S. costatum and Chaetoceros spp. showed distinct thresholds in 4 of 5 and 2 of 3 years in which 77 they were measured, respectively, suggesting a more variable sensitivity. In addition, their lower Ks values suggest that these species would be overall less sensitive to nitrate depletion than Thalassiosira spp. Chaetoceros spp. showed more variability than S. costatum, suggesting that while S. costatum may be consistently less sensitive than Thalassiosira spp., Chaetoceros spp. may be occasionally equally sensitive. In a very general manner, then, the nutrient sensitivity of these three diatom genera could be approximately ranked, from high to low sensitivity, Thalassiosira spp.~> Chaetoceros spp. > S. costatum. Irradiance, nutrients and sinking Analysis of all sinking data for Thalassiosira aestivalis, the most abundant diatom over all five years of this study, indicated that once extreme values were removed, ambient nitrate concentration and total daily quantum irradiance were the best predictors of cell sinking rate. Temperature and sigma-t were not significant predictors, nor was light history. Both nitrate concentration and quantum irradiance were negatively correlated with sinking rates, indicating that if nutrients are held constant, sinking rates are directly dependent on light availability. Conversely, if light is held constant, sinking rates are directly dependent on nutrient concentrations. In addition, the lowest sinking rates would be expected under energy- and nutrient-saturation, and the highest under light- and nutrient-limitation. This agrees well with the scenario outlined in Chapter 1, and supports the notion that before nutrient depletion, sinking rates would be controlled primarily by light availability, while after nutrient depletion, sinking rates would be controlled by a combination of both light and nutrients. 78 In order to observe sinking rate reduction at the chlorophyll maximum,where nutrients are high and light levels are low, the effect of a nutrient increase at depth (SR decrease) must have outweighed the effect of lower light intensity (SR increase) on sinking rates. This might be expected from the results of the multiple regression, since nitrate concentration, not irradiance, was the primary predictor of sinking rates in the final model. Certainly, if cells have substantial reserves of respiratory carbon, as suggested in Chapter 1, light would only be directly related to sinking rates once cells had been light-limited for an extended period of time. Energy-saturated cells should still have the energy reserves to take up nitrate and reduce sinking rates at depth. Because all the outliers were measurements of unusually high sinking rates, their selective exclusion from the final analysis should not go without mention. In fact, 3 out of 4 of these extreme high sinking rate measurements all occurred during days of simultaneous low nutrients, low salinity and 1-2 days of relatively high irradiance, and if these values were left in the regression, the 1 - 2 day light history became a significant predictor of (positively related to) sinking rates. This suggests that a combination of physiological stresses may cause unusually high sinking rates to occur sporadically. If light is necessary for energetic sinking control, sinking rates will be reduced under high light (Chapter 1). However, when other conditions are sub-optimal (e.g., low nutrient and salinity levels), extended periods of high light may actually be detrimental to cells in the surface layer. There is good evidence that high light can damage cells (Whitelam and Codd, 1986) especially if cells are already physiologically stressed (Prezelin, 1981). While the significance of 4 out of 104 data points should not be overemphasized, the occurrence of extreme sinking rate values are of interest in terms of their impact on sedimentation. 79 The light history of cells at the surface is of course approximated only very roughly by cumulative incident light. However, cells experiencing low light levels for more than two days (Ih < 3 x 1021 quanta) did seem to have a nutrient-sinking relationship distinct from that of cells at higher Ih levels. The linear dependence of sinking rates on nitrate concentrations at low Ih indicates that ambient nitrate is a better predictor of sinking rates at low light than at high light. One reason for this may be that cells which are growing very rapidly at the surface undergo more rapid physiological changes than slower growing cells. Ambient nutrient concentrations might be a reasonable estimate of the growth conditions of slow-growing cells, since neither is changing rapidly. When cell metabolism changes more rapidly, under high light, and ambient nutrient concentrations are decreasing exponentially, cell physiology might become uncoupled from ambient conditions. At a time when NO3 concentration becomes undetectable, for instance, cells might initially be nutrient-saturated (SR ~ NS Mean), while only a short time (hours) later these same cells could be severely nutrient-limited (SR > NS Mean). On a more general level, sampling in 1988-89 showed significantly greater variance in sinking rates in the ND period than in the NS period. This too probably results from changes in the distribution of specific sinking velocities within the population. The larger variance at low nutrient levels may reflect the greater probability of sampling population segments which are experiencing more severe nutrient stress than others. It may also indicate that periods of nutrient stress make cells more vulnerable to other factors causing sinking rate increases. Exhaustion of nutrients creates a sparse resource field wherein the encounter of a substrate becomes probabilistic, and together with the variations in light field, temperature, salinity and vertical mixing, this results in a wider distribution in preconditioning histories than during the NS period. While no single physical factor 80 emerged as a significant sinking rate predictor, several physical factors converging simultaneously might exacerbate physiological stress initiated by nutrient depletion in the ND period, inducing non-linear sinking rate increases. Because of this, there is little doubt that some cells sank much faster than the mean sinking rate values reported here. The literature contains several examples of phytoplankton sinking at even 10-100 m d-1 (Smetacek et al. 1978). However, in this system, such rapidly settling cells constitute a very small fraction of the spring bloom biomass. Concomitant sediment trap studies gave sinking rate estimates based on the vertical displacement of biomass, which were very similar to those reported here. Integrated surface and sediment trap biomass budgets balanced well, indicating that no large fraction of biomass was unaccounted for. It is thus likely that the estimates of mean sinking rate reported 1 re accurately reflect sedimentary loss rates of phytoplankton biomass from the photic zone. The patterns of species-specific nutrient-sinking responses documented here in the photic zone would thus be expected to have a direct impact on the patterns of bloom sedimentation to the benthos of Auke Bay. 81 CHAPTER 3. SPRING BLOOM SEDIMENTATION IN A SUBARCTIC ECOSYSTEM H. SUCCESSION AND SEDIMENTATION INTRODUCTION In areas where substantial diatom blooms occur, the sedimentation of diatoms may be an important, if not the major source, of organic input to the sediments (Smetacek 1984; 1985a). Sinking rate differences between species and changes in sinking rates over time within a species might influence sedimentation patterns in several ways. The composition of bloom biomass might be affected through the influence of pre-bloom sinking rates of different species on bloom initiation and species succession (Harrison et al., 1986; Brzezinski and Nelson, 1988). The sedimentation of this biomass might be affected by the species-specific sinking responses to nutrients, differences in maximum attainable sinking rates, and tendency to aggregate (Chapter 2, this study). Chapter 2 focused on the relative importance of the nutrient responses of the principal diatom species in controlling sinking losses from the photic zone during and after the spring diatom bloom in Auke Bay. This chapter attempts to link the sinking rate patterns documented for the principal diatom bloom species in Auke Bay (Chapter 2) with the observed sedimentation of bloom biomass. Auke Bay, Alaska is a small subarctic embayment where each year up to 40% of the spring bloom primary production reaches the floor of the bay as intact diatom cells (Laws et al. 1988), usually after high surface concentrations of rapidly dividing diatoms have stripped nutrients from the surface water. In this chapter I compare water column species abundance with diatom sedimentation flux to sediment traps in Auke Bay, assess which species 82 constitute the bulk of sedimenting biomass, and link these sedimentation characteristics to their sinking-nutrient response. MATERIALS AND METHODS Water column sampling methods The study site and sampling station have been described in Chapter 2. Cell counts for the major species of diatoms at the surface (2 m) were made bi-weekly from samples at a station in Auke Bay from mid-March to June, 1985 -1989. All samples were preserved in formalin, and cell concentrations were enumerated with an inverted microscope using the Utermohl technique (Utermohl 1958). The total number of cells for each species for this 3 month period was obtained by integrating all the bi-weekly counts over the three month period (i.e., an estimate of the area under each species abundance curve in Fig. 3.1), giving a temporal integration for each year's spring bloom. The total % abundance of each diatom genus was calculated by dividing the total cell numbers for the three month period for a genus by the total number of cells for all genera over the same time period. From 1987-1989, a second depth was also sampled (1987, the 10 % light depth (10% LL); 1988-89 the chlorophyll maximum depth (CMX)), and similarly integrated. On two occasions in 1988 and 1989 (4/21/88 and 4/17/89), cell counts were made at 11 depths just prior to the major post-bloom sedimentation event. These profiles gave vertically integrated cell numbers for each species, expressed in cells m-2. Using cell volume estimates of the major diatom species from Haigh (1988) and assuming carbon to be proportional to biovolume according to Strathmann's formulae (Strathmann 1967), estimates of cells m-2 83 were converted to g C m"^. In 1988, species composition changed markedly during the peak of the bloom, and therefore a second vertical profile of cell numbers was estimated for 4/28/88, during the decline of the bloom. For any species, the larger of the two vertically integrated species abundance estimates was used to calculate the % of cells from the surface which reached the 35 m sediment trap. (Table 3.1). All estimates were expressed as cells m-2 and as g C m-2 (Table 3.1). Sediment t raps Single sediment traps were suspended at 15 m (or 20 m) and 35 m in order to quantify the extent to which each species growing in the surface waters sank to depth. Details of construction and deployment are given in Ziemann et al. (1988). The traps were cylindrical, 45 cm long and 15 cm in diameter, and were topped with a plastic grid baffle. They were retrieved biweekly from mid-March to mid-June, refilled with filtered seawater, and redeployed with 80 g of salt crystals added to form a stabilizing density gradient within the trap. Since trap contents were collected biweekly, decomposition in the traps should have been minimal and any grazing by zcoplankton should have been retarded by the high salt content. From 1987 - 1989, cell counts and species identification were performed as described for the water column. The biweekly cell counts for the major diatom species were summed over the 3-month sampling period, first as cells m-2 for each species, and then as % abundance of each species relative to the total cell numbers over the period. Cell numbers were converted into cell volume and then into cell carbon as described above. 84 RESULTS Bloom composition and succession The three principal phytoplankton genera during the study period (1985 to 1989) were Thalassiosira (primarily T. aestivalis, with some T. nordenskioldii, and T. gravida), Chaetoceros (C. compressus and C. debilis) and Skeletonema costatum (Fig. 3.1). Thalassiosira was the most abundant genus at the surface in 3 of 5 years (1985, 86, and 88; Figs. 3.1 and 3.2), while S. costatum dominated in 1987 (46% of total temporally integrated cells, Fig. 3.2). In 1989 the species composition was more heterogeneous, with Chaetoceros spp. dominating the assemblage, but no one species comprised over 50% of the total cell numbers. In all years, the initial period of the bloom was dominated by species of Thalassiosira and Chaetoceros. In 1986,1987 and 1989, this was followed by a strong peak of Skeletonema costatum. During 1988, the temporally integrated species composition (representing integrated cell numbers over the entire primary spring bloom, usually early March to mid-May; see Methods) in the chlorophyll maximum layer was far more homogeneous than the surface, with the main bloom species, Thalassiosira aestivalis, dominating (Fig. 3.2). T. aestivalis formed 87% of the total cell numbers at the chlorophyll maximum (>99% of the biovolume), compared to 42% at 2 m. This pattern was also observed in 1989 for the most abundant species, Chaetoceros compressus. 85 Generally, in both 1988 and 1989 the vertically integrated estimates of water column species composition (Shown as cell number and cell carbon, Table 3.1) were very similar to temporally integrated estimates of species composition (Shown as % total cells, Fig. 3.2). This similarity of these two completely independent cell abundance estimates gives us confidence that neither was heavily biased, and that they accurately reflect surface species composition. 86 Figure 3.1. Time series of species composition (3 principal diatom genera) and abundance during the spring bloom at 2 m in Auke Bay, Alaska between 1985 and 1989. Note different scales on vertical axes. 87 r 4 00 00 I O z a z o u — Tholasfiosiro spp. Chastocsros spp. -— Skelelonemo sp. 1985 2 u MARCH I f 12 APRIL u o i z a z o u d Ul o 10 + 8 6 4 " 2 0 — Thotossloslro spp. ••• -Choslocsros spp. - - Skslstonsmo ip. 4 l\ I I / I / I / I / I / » / » 1987 MARCH | APRIL | HAY | JUNB ~ 7 • Tholasfiosiro spp. -Chastocsros spp. • Sktlstonemo sp. 3--2 MARCH | APRIL I MAY | JUNB 1988 5 4- -3 2 I —-Tholasfiosiro spp. "«• Chastocsros spp. - - Skslstonsma «p. »-43ol "?i '* MARCH | APRIL | MAY E 6 o 5 rO O C 4 Z S 3 z 2 M U S ' u — Thataisiasla spp. Chastocsros spp, — Skelelonemo sp. 1989 i~fciV±±.r-<r jfc.-i.. r-MAIICII I MAY Ai | JUNB JUNB Figure 3.2. The relative abundance of the three major spring bloom diatom genera, integrated temporally over each spring period (see Methods for details), 1985 to 1989 inclusive, at 1 or 2 depths in the water column and in the 2 sediment traps at 20 and 35 m. All values given as % total temporally integrated cells for a particular year. 1986 sediment trap values are given only as approximate percentages. (LL = light depth; CMX = chlorophyll maximum layer; ND = no data available). 89 too. 1985 80 H I Thdassiegira ipp. E 2 Ch««tec«rai cap. C D SU«Uton«mn sp. in a: w | *0 2 d 30 w u S I C3 33 t Jan s'oo-S 80 S 30 ^ M icau. WATW COLUMN NO NO 2 0 m 33m SEDIMENT TRAPS 1937 — Thct«»»toj«Hs spp. g Z Chattacwqs cap. C D SW«4«ten«ma sa. 100 80 SO 40 20 0 100 1985 • • Tholoutojira up. 2Z2 Chacteccras tap. ) 8 0 CD SV«Utan«ma sp. U 27 loci 2m WATER COLUMN 40 . 20 V i t I 1938 20m 33m SS»M£KT TRAPS 89 80 60 40 20 + 2m 1 0 * U . MATER COLUMN 20m 33m £ SEOIMENT TRAPS •3 lOO-i | I 1989 - ao • •J '-) 2 80 + < A 8 in hi L w • • Thofaiinaira spp. Zacnactaearas sep. CD Sl>c<tto««ma sp. ' <i 2 m CUX . WATER COLUUN 15m 33m SEOiuENT TRAPS • • Thafessiestra isa. C2Cna«toetro« sap. IZ2 S'«l«can«ma S3. s< I 20 • I Ii 2 m CUX WATER COLUMN 13m 33m SEDIMENT TRAPS 90 <o TA1ILK 1. Comparison of verliciilly integrated species compiisiliuii nenr llic sprint; bloom peak In 1988 (4/21 and V 2 8 ) n m J ' » I9 8 9 (V^ ) , and I lie total number of cells caught In the 35 m sediment trap over the bloom periods (1988-89). The comparison is expressed us llic % loliil cells reaching the 35 m sediment trnp. VORTICAL INTEGRATED ACCUMULATION COMPARISON PROFILES IN 35 m SEDIMBNTTRAP YEAR SPECIES INTEGRATED INTEGRATED TOTAL CELLS TOTAL CARBON % CELLS CELL # CARBON REACHING o , , n o -> 35mTRAP (109cells m'2) (gCnf2) (KFcells nf2) (gCnf2) 1988 4 /2 ! A 4/28/88 THALASSIOSIRA T. aestivalis 33.0 27.0 13.0 11.1 41.0 T. gravida 4.0 3.2 0.18 0.15 4.6 SKELETONEMA S.caslaltim 9.3 6.0 0.006 0.004 0.1 CHAETOCEROS Ccompressus 1.3 0.18 0.18 0.02 13.0 Cdebllis 3.1 0.31 0.02 0.002 0.7 TOTAL 58.0 37.0 13.8 37.0 1989" 4/17/89 THALASSIOSIRA T. aestivalis 2.37 1.96 2.5 2.0 104 T. gravida 1.67 1.07 0.37 0.31 22 T. nordenskloldii 2.20 1.32 0.50 0.19 23 SKELETONEMA S. coslatum NOT PRESENT IN PROFILE CHAETOCEROS Ccompressus 8.3 1.10 0.42 0.10 13 Cdebllis 0.80 0.02 I.I 0.11 14 Cradlcans 1.53 0.20 1.4 0.20 92 TOTAL 16.9 5.68 6.0 35 Species composition of sedimented material The same three diatom genera dominated the sediment trap samples throughout the study, but the species composition of the sediment trap samples was markedly different from the species composition of the upper water column. In general, Thalassiosira spp. (especially T. aestivalis) tended to settle out of the water column more than other species (Table 3.1; Fig. 3.2). Chaetoceros spp. were generally reduced in abundance in the sediment traps compared to the water column, and the abundance of Skeletonema costatum was even more reduced in the sediment traps in all years of the study. In 1987 the contrast was especially striking, since S. costatum was the principal bloom species (ca. 45% of cell numbers) in the upper water column, but it formed only 3% of the actual cell flux to 35 m (Fig. 3.2). Because of their large size, Thalassiosira spp. have a high carbon content per cell, which greatly increases their importance as contributors of carbon flux to the benthos. T. aestivalis alone contributed a maximum of 11.1 g C m-2 over the spring (between March 14 and June 6) in 1988, which represents about 30 % of total carbon production from mid-March until nutrient depletion (Ziemann et al., 1988). In 1989 Chaetoceros compressus contributed a maximum of only 0.2 g C m-2 over the bloom, even though it was the most abundant species in the water column during the bloom. Over all five years of this study, there was a strong linear relationship between temporally integrated 2 m standing stock (as total cell numbers over the 3 month sampling period each year) and temporally integrated cell numbers of all species reaching the 35 m sediment trap (r=0.586, p<0.05) (Fig. 3.3; regression for all species together not shown). Note that S. costatum falls outside the range of all other species. If the Skeletonema 92 costatum values are removed, the relationship between standing stock and actual fluxes is even more highly significant (p<0.005). Plotted separately, the Chaetoceros species' and the Thalassiosira species' relationships both remain significant, but the regression for Thalassiosira spp. describes a much higher fraction of the variance of the data (p<0.005, r2=0.97) than that of the Chaetoceros spp. (p<0.05, r2=0.55)(Fig. 3.3. Note tighter fit for regression of Thalassiosira spp. in log-log plot, inset). The linear relationships are FL = 0.18 * SS + 4.29 x 108 for Thalassiosira spp., and FL = 0.07 * SS - 2.92 x 108 for Chaetoceros spp., where FL is the temporally integrated total number of sedimented cells of a species in any one year, and SS is the temporally integrated 2 m standing stock of cells for the same species for the same year. The slope of the linear regression line for Thalassiosira spp. was significantly higher than that of Chaetoceros spp. (Student's one-tailed t-test, t=1.84, DF=11). 93 1.5 r-r 2 4 6 8 10 Temporally integrated 2m standing stock (cells m^xlO10) 12 14 Figure 3.3. Relationship between temporally integrated cell abundances for the 3 principal diatom genera at 2 m over the spring bloom for 1987,1988 and 1989, and total temporally integrated cell numbers collected in the 35 m sediment trap for these same years. Each point represents the integrated value for one species in one year (see Methods for details). Inset shows log vs log plot with labels indicating species and year for each point (Species labels: Thalassiosira spp. (•) : A = T. aestivalis, G = T. gravida, N = T. nordenskioldii. Chaetoceros spp. (O): C = C. compressus, D = C. debilis, R = C. radicans. Skeletonema costatum (•) = S. Year labels: 7 = 1987, 8 = 1988 and 9 = 1989). 94 DISCUSSION Sinking and succession Takahashi et al. (1978) observed that over the winter period in Saanich Inlet, B.C., a highly varied assemblage remained suspended in the upper water column but generally at very low cell concentrations. Sommer (1989) stated that the initial composition of the spring bloom was determined by the size of this inoculum (a function of the previous year's sedimentation fluxes and overwintering conditions) and maximal reproductive rates under low light and in cold water. At very low or negligible winter growth rates, sinking rates may also determine whether the inoculum will be maintained in the photic zone. Thus, despite a potentially large inoculum from the previous year's sedimented biomass, some species would not successfully inoculate the upper water column early in the spring because of their high sinking rate over the winter or the early portion of the bloom. The data support the idea that low sinking rates may be a necessary precondition for inoculum survival and subsequent bloom of the species in any year. High early spring sinking rates (e.g. Chaetoceros debilis in 1987, Thalassiosira aestivalis and Thalassiosira gravida in 1989, and Skeletonema costatum in 1988) were generally followed by low abundance over the subsequent spring bloom. For species using energetic sinking control (probably the larger diatoms; Chapter 1, this study) low pre-bloom sinking rates are probably the consequence of high physiological tolerance to low temperature and low light, which represent the two principal factors controlling early spring photosynthetic rates in Auke Bay (Laws et al. 1988). Smetacek and Passow (1990) suggest that the retention of cells in the surface layer is the most important 95 factor controlling the induction and growth of a spring bloom. Thus, low sinking rates early in the bloom may indicate a species' potential for dominance in the bloom as a whole. Once light levels are sufficient for growth of most species, low temperature growth rates and other factors such as those suggested by Sommer (1989), as well as specific light and nutrient conditions, will determine a species' actual success. The most frequent successional sequence observed in Auke Bay was the temporal separation of the Thalassiosira spp. and Skeletonema costatum peaks, with S. costatum (and often Chaetoceros spp.) tending to bloom after the Thalassiosira spp. peak subsided. This corresponds with the earlier characterization of Stage I succession by Guillard and Kilham (1977). They separated early diatom bloom succession into early Stage I (Thalassiosira spp.) and late Stage I (S. costatum and Chaetoceros spp.), and suggested that Thalassiosira spp. would tend to grow early in the spring, at temperatures between 1 and 5<>C, and that S. costatum preferred higher temperatures, later in the bloom. In order to remain in the photic zone until the temperature favoured growth, S. costatum must have maintained low sinking rates during growth of earlier bloom species at the surface. The successional separation between the fast-growing diatoms Thalassiosira spp. (principally Thalassiosira aestivalis in our system) and Skeletonema costatum identified by Guillard and Kilham (1977) may be attributed to the physiological characteristics of each species. T. aestivalis is a low temperature tolerant species, which consistently blooms as soon as light levels increase in the spring. T. aestivalis depletes nitrate quickly, and its nitrate-sensitive sinking rates increase. S. costatum is a high temperature species (Guillard and Kilham, 1977) whose sinking rates are nitrate-insensitive and silicate-sensitive (Harrison et al., 1986). As soon as T. aestivalis sinks from the surface upon depletion of surface nitrate, S. costatum can bloom, but only if nutrient levels (especially silicate) have 96 not fallen too low during the Thalassiosira spp. bloom. In 1987, for instance, deeper mixing increased surface nutrient levels during and directly after the T. aestivalis bloom, and S. costatum bloomed immediately. In most other years, post-bloom depletion of both nitrate and silicate was severe, and S. costatum did not reach high numbers until after mixing re-introduced nutrients into the upper water layers. In 1988, post-bloom nitrate concentrations were lower for a longer period, than in any other year (see Chapter 2, Fig. 2.2) and S. costatum did not bloom during the sampling period. This suggests that a lack of sensitivity to nitrate depletion (Chapter 2) may allow S. costatum to endure low nitrate concentrations in the water column during the Thalassiosira spp. bloom. It can then bloom in warmer surface water once the more nitrate-sensitive Thalassiosira species have sunk out. This proposed scenario is supported by a recent lab study, in which field phytoplankton assemblages which were incubated under nitrogen limitation and subjected to natural sinking losses quickly became dominated by S. costatum (Harrison et al. 1986). There is an important vertical component to the patterns of temporal succession observed in Auke Bay. Since primary production levels were consistently highest at the surface (Ziemann 1985), sinking must have been an important factor in the formation of the subsurface chlorophyll maximum, at least in the three years when Thalassiosira aestivalis was the major bloom species. This species had significantly lower sinking rates at the chlorophyll maximum than at the surface during the year when it dominated the bloom (Chapter 2). This fact, combined with observations of greater heterogeneity in species composition at the surface, may indicate that the species which remain at the surface during formation of the chlorophyll maximum by the major bloom species do so by virtue of their lower sinking rates. This supports the idea of a lower "loss factor" during the bloom for less nutrient-sensitive species. 97 Sinking and sedimentation Water column production and sedimentation processes have traditionally been considered separately. Smetacek (1984) identified the need to link studies of primary production and sedimentation in order to document the ecological processes regulating the quantity and quality of sedimented material. Recent studies indicate that during a spring diatom bloom, a high percentage of surface primary production reaches the benthos, often sinking as whole diatom cells that are significantly higher in nitrogen content than most detrital material (Davies and Payne 1984, Burrell 1988). In many higher latitude coastal systems, the annual spring diatom bloom represents the period of highest "new" production, when the largest carbon flux from the surface to the benthos can occur (Burrell 1988). The spring bloom thus represents a period of tight benthic-pelagic coupling, when the successional and sinking processes of diatoms at the surface are most likely to have a measurable effect on the composition of sedimenting biomass. The species composition of sedimenting diatoms during a spring bloom depends on the species composition of the diatom bloom, modified by the sinking rate of each individual species. The temporal sequence of changes in sinking rate will also modify the probability of cells sinking to the benthos. Over the bloom period in Auke Bay, the abundance of herbivores increases, stratification begins in the water column, and surface temperatures increase rapidly (Ziemann et al., 1985; 1986; 1987; 1988; Paul et al., 1989). Early, mid-, and late bloom diatoms will have different maximal growth rates, and be exposed to differing 98 grazing pressures and mixing regimes, all of which could influence their probability of reaching the benthos after a bloom. Suess (1980) found that the sedimentation of POC in the open ocean at any depth below the photic zone was predictable by a simple linear relationship between surface primary production and sedimented carbon flux. This relationship deteriorated in the upper water layers (above 50 m) where actual sedimented biomass was lower than Seuss' model predicted. Burrell (1988) indicated that a strong relationship between surface primary production and sedimentation of organic carbon had been successfully demonstrated only for open ocean regimes, and that coastal areas were too physically complex, and the carbon sources in them too numerous, for any such simple relationship to exist. Our study did find a simple linear relationship between temporally integrated 2 m standing stock of diatom cells and temporally integrated numbers of cells collected in the 35 m sediment trap, for each major diatom genus over the spring blooms of 1987, 1988 and 1989 (quantitative data for 1985 and 1986 were not available). The clarity of the relationship in this study is probably due to a combination of factors including the high concentrations of cells, the rapidity of bloom sedimentation (< 2 weeks) and the short vertical distance over which sedimentation was measured (35 m). The relationships between the temporally integrated 2 m standing stock of diatom cells and temporally integrated cells reaching the sediment traps for the two principal genera illustrates the consequences of fundamentally different sinking processes on sedimentation patterns. The higher slope of the relationship for Thalassiosira spp. compared to Chaetoceros spp. indicates that per unit biomass, more cells of Thalassiosira spp. sank to the sediment trap than did those of Chaetoceros spp. (Although scarcity of data precludes strong inferences for Skeletonema costatum, it appears that this species shows an 99 even lower sedimentation probability). The difference in slopes between Thalassiosira spp. and Chaetoceros spp. would seem to be a direct consequence of the higher overall sinking rates and nutrient-sensitivity of Thalassiosira spp. If large-scale aggregation occurred, we should have seen a significant non-linear or exponential component in the sedimentation vs biomass relationship. The absence of any clear non-linear component indicates that any aggregation which did occur, did not influence sedimentation patterns. This suggests that the preferential sedimentation of Thalassiosira spp. was caused by changes in single cell processes rather than by aggregate formation. This observation represents a marked contrast to other systems where aggregation processes have been seen to have an important influence on sedimentation patterns (e.g., Alldredge and Gotschalk 1988, Logan and Alldredge 1989), Despite a reasonably high sinking rate, Skeletonema costatum did not reach the sediment traps in large numbers in any year. This was especially notable in 1987 when an intense surface bloom of S. costatum reached concentrations of 12 x 106 cells L-l, but sediment trap collections of this species were virtually negligible. The low nitrate-sensitivity of S. costatum and its tendency to grow late in the bloom may combine to effect longer, less intense sinking events for this species than for the other diatom bloom species which tended to have cohesive, mass sinking events. Longer pelagic retention increases both the probability of dissolution en route, and the possibility of advection and grazing. Zooplankton numbers generally did not peak until late May and June, well after the major sedimentation event, and grazing losses during the spring bloom in Auke Bay amounted to less than 10 % of the daily production (Paul and Coyle 1990a). In addition, the primary diatom grazers, the euphausiids, preferred Thalassiosira spp. to S. costatum (Paul and Coyle 1990b). Therefore, it is unlikely that grazing caused the observed losses of S. costatum. On the other hand, S. 100 costatum is generally less silicified than other diatom species, and is considered sensitive to dissolution losses, especially at temperatures above 100 C (Roelofs 1983). Mortality and frustule dissolution are therefore the most plausible explanations for the losses of S. costatum, especially if the sinking processes of S. costatum combine to lengthen the time scale of sedimentation. Although Skeletonema costatum is commonly a dominant neritic species, its representation in the sediments is highly variable (Roelofs, 1983). Several studies do document the contribution of this species to the sediments; for instance, Sancetta and Calvert (1988) studied the sediment record in Saanich Inlet, British Columbia, and found that S. costatum sedimented out in large numbers, indeed, greater numbers than the Thalassiosira species (the opposite pattern to the one observed in Auke Bay). The sinking/nutrient dynamics of S. costatum in Auke Bay may be very different from those in Saanich Inlet. Nutrient data collected by Takahashi and co-workers (1977) suggest Si04:N03 ratios before the spring bloom in Saanich are lower (about 3:2) than pre-bloom ratios in Auke Bay (2:1), which might make Si04 potentially limiting for S. costatum, causing higher sinking rates, and more rapid sedimentation to the benthos with lower dissolution losses. It is impossible to evaluate this hypothesis from the Saanich Inlet data set, however, since in the year studied by Takahashi and co-workers, Thalassiosira spp. were the dominant bloom species, and Si04 levels in Saanich Inlet never fell below 1 umol L-1 in the upper 10 m. The data of Sancetta and Calvert (1988) indicated, however, that despite the dominance of S. costatum in the sediment traps, Thalassiosira spp. remained the most important contributors of sedimenting biomass in terms of actual carbon flux (as measured by cell volume) than any other sedimenting genus. In our study, Thalassiosira spp. 101 contributed between 90 and 99% of the carbon flux to the 35 m sediment trap. On a larger geographic scale, this suggests that in similar coastal ecosystems, the rapidly-sinking, nitrate-sensitive species (such as Thalassiosira spp.) are the most important overall contributors to the sedimenting flux of new production to the benthos. Thalassiosira spp. are abundant as spring bloom species throughout the subarctic Pacific (Taylor and Waters 1982) and along the temperate coastal zones of western North America (Takahashi et al. 1978) from northern British Columbia as far south as San Diego (Allen 1923). In fact, the larger Thalassiosira spp. (including T. nordenskioldii, T. pacifica, T. aestivalis, T. eccentrica, T. antarctica and T. decipiens) have been observed at mid- to high latitudes throughout the coastal world oceans, including the east coast of North America (Conover and Mayzaud 1984, Riebesell 1989) the northern coast of Europe (Gieskes and Kraay 1975, Smetacek 1985a, Eilersen et al. 1989) and the southern ocean (Guillard and Kilham 1977). However, despite their world-wide abundance, relatively little is known about the Thalassiosira species' physiology. Further research is desirable, both on the autecology of Thalassiosira spp. and on their basic physiological responses to temperature, nutrients and light. During the spring bloom in many temperate coastal embayments, if the response of cell physiology to ambient nutrient concentrations has the potential to control a cell's sedimentation rate, (and hence its pelagic retention time), it is this physiological response which will ultimately determine the fate of a cell. The high sedimentation capacity of the large Thalassiosira spp. (up to 11.1 g C m-2 during a single spring bloom for a single species (T. aestivalis) in this study) and their cosmopolitan contribution to highly productive coastal blooms, imply that the physiology and ecology of these species could be important in determining the global flux of new production. 102 CHAPTER 4. THE ROLE OF SINKING AND FLOATING DURING SEXUAL REPRODUCTION IN THE MARINE DIATOM DITYLUM BRIGHTWELLH INTRODUCTION In general, the creation of genetic diversity remains the primary rationalization for the evolutionary maintenance of sexuality over reproductively more efficient asexual reproduction (Williams, 1975), although other factors such as energetic gamete selection (Hurst, 1990) and DNA repair (Bernstein et al., 1985) may also contribute. Intraspecific genetic variation in marine phytoplankton can be greater than genetic differences between entirely different species of terrestrial plants (Wood, 1988). To what extent might the oceanic environment select for this high genetic diversity? The environment in which phytoplankton grow (both freshwater and oceanic) is highly variable in time and space (e.g., Mackas et al., 1985). Diurnal and seasonal changes in mixed layer depth (e.g. Woods and Barkmann, 1986), coupled with tidal mixing (Pingree et al., 1978) and wind mixing (Therriault and Piatt, 1981) cause periodic fluctuations in stability and temperature of the ocean surface, changing irradiance, nutrient availability and growth rate, on a wide variety of time scales. In addition, sporadic amplification of phytoplankton patch growth through non-linear interactions with physical processes at a variety of time scales (Denman and Powell, 1984) create either "noisy" (random) phytoplankton distributions or "patchy" (mosaic-like) distributions depending on the physical regime (Bennett and Denman, 1985). Patch survival, in turn, may depend on the patch 102CL reaching a critical size (~1 km, Kierstead and Slobodkin, 1953), though non-linear interactions of phytoplankton and grazers, for instance, can make patches larger or smaller than expected from growth alone, and even prevent their dispersal through turbulent lateral diffusion (Steele, 1974). For the purpose of the following discussion, a "patch" suitable for growth means a relatively small (1 m to 1 km) area where irradiance and nutrient concentrations are sufficient for the initiation of growth if a given species (essentially the patch sensu Hutchinson, 1961). This spatial "patch" must then persist in time long enough for phytoplankton to respond physiologically to the favourable conditions, and divide (at least once). Once a species clone begins to grow in a "patch" to which it is genetically suited, diffusion of cells out of the "patch" becomes detrimental to population growth. The fact that many diatoms form chains, for instance, suggests that there may be adaptive value to keeping daughter cells within a growing population. Brand (1982) has evidence that turbulent diffusion, sweeping organisms out of a patch, can indeed serve as an important loss term to a growing population. One might expect similarly that under suitable growth conditions, sinking processes too would constitute a "loss". However, because in the longer term, however, all patches are ephemeral, the need to maintain growing populations in suitable surface "patches" must be balanced by the eventual distribution of cells to new future "patches". Cells must therefore balance potential growth losses through biomass diffusion and sinking against the benefits of dispersal by these same mechanisms. In a highly unpredictable and patchy environment in both time and space, the maintenance of genetic diversity within a species may be especially important for successful dispersal. Maintenance of diversity may allow different clones of a single species of diatom 103 to bloom under a variety of different environmental conditions (different types of "patches"). Such genetic variability has already been documented for Skeletonema costatum (Gallagher, 1983) and Thalassiosira aestivalis I pacifica (Roelofs, 1983). Both these species are known to dominate coastal diatom blooms. If sexual reproduction is the primary source of genetic variation in marine phytoplankton, then periodic sexual induction may play an important role increasing the success of "patch" colonization. However, unicellular clonal populations such as those of marine diatoms are especially vulnerable to the reductions in effective growth rate necessitated by sexual induction (Lewis, 1983) since their growth rate is a direct reflection of their fitness (Wood, 1988). Vegetatively reproducing asexual cells thus have the highest fitness of the three life cycle stages (1) vegetative asexual growth, 2) sexual reproduction and 3) vegetative resting spore formation). Sexual reproduction decreases the fitness of organisms in the short term. Specifically, the costs of sex (meaning basically the temporary loss of the ability to produce offspring) are referred to as the "cellular mechanical costs" due to the interruption of normal vegetative growth, the "cost of meiosis" which refers to the cost of producing one large or many small gametes, the "cost of fertilization" which refers to the time and energy of finding a mate, and finally, the "cost of recombination", which refers to the potential of completely losing the parent genotype if the entire population recombines. These costs must be mitigated" or compensated for if sexuality is to be retained over evolutionary time. In diatoms, such compensation may include an increase in growth rate after sexual reproduction (Costello and Chisholm, 1981), increased genetic diversity, and the restoration of larger cell width supposedly necessary after a long series of vegetative divisions over which the cells become narrower. 104 Diatoms must therefore have ways to minimize these costs. For instance, if sex occurs only when growth rates are obligately low, the reproductive losses are minimized. It has already been mentioned (General introduction, this study) that sexual induction often occurs when cells are physiologically stressed through nutrient limitation (Davis et al., 1973; Drebes 1966; 1977), sudden increases in growth irradiance (Drebes, 1966) or decreases in temperature (Drebes, 1977). There may be strong selection to limit sexuality to periods of obligate low growth rates when reproductive losses due to the onset of sexuality will be minimal. However, the induction of sexuality of diatoms is still not completely understood. The sinking dynamics of sexual cells could influence the balance of costs and benefits of sex in diatoms. Whether recombined cells leave the parent population or remain at the surface must influence the eventual distribution of a species. Although one study documented increased sinking rates after sexual reproduction due to the increase in the average diameter of recombined cells (Smayda and Boleyn, 1966), the sinking dynamics of sexual cells are not well documented. The objective of this work was to measure systematically for the first time, changes in the sinking rate (and floating rate) of cells undergoing sexual reproduction in Ditylum brightwellii. By subjecting D. brightwellii to repeated nitrogen starvation, after an increase in growth irradiance, the formation of sexual cells was induced. The sinking and floating rates of pre-sexual cells, early sexual cells, and gametes were measured, documenting an increase in sinking rate as sexual induction progressed. In addition, a natural bloom of D. brightwellii was documented at Jericho Beach near Vancouver, in which a large fraction of the population became sexual. 105 MATERIALS AND METHODS 1. Field measurements As part of a regular plankton monitoring program by Prof. F. J.R. Taylor and co-workers, Dept. of Oceanography, U.B.C., the phytoplankton assemblage in the top 3 m off the pier at Jericho Beach, Vancouver, B.C. was sampled weekly between 1989 and 1991 (except for winter sampling, which was monthly). Jericho Beach lies on the southern shore of Burrard Inlet, about 5 km east of the opening of Burrard Inlet into the southern Strait of Georgia, a large inland sea between Vancouver Island and the mainland of British Columbia. The southern shore of Burrard Inlet between Jericho Beach and the inlet mouth is dominated by the shallow mudflats known as Spanish Banks. The depth near Jericho pier is 4 m at high tide, and the area can be stratified by the freshwater plume of the Fraser River when winds and tides force the river plume northward. A 30 fim mesh plankton net was lowered off the pier at high tide to just above the bottom and hauled vertically upward. Usually a sweep along the surface was also performed as part of the same net haul. A surface sample for temperature (measured immediately), and salinity (measured with a refractometer upon return to the laboratory) was also collected. Duplicate sub-samples from this surface sample were collected using two 60 mL syringes; 50 mL per syringe was pushed gently through a GF/F filter housed in a Swinnex holder. The filter was frozen and later analyzed for chlorophyll and phaeopigments. Subsamples for plankton analysis were immediately preserved with Lugols iodine solution and were observed under a Zeiss model X inverted microscope and photographed with the attached Zeiss camera with Kodak TMax 100 film at lOx, 20x and 40x magnification. 106 Fractions of the surface sample (10 mL) of were settled for full plankton enumeration. Samples from the net haul were settled for species identification as presence/absence or dominance (>50%) only. 2. Laboratory experiments: Two 50 mL batch cultures of the marine diatom Ditylum brightwellii were inoculated from the Northeast Pacific Culture Collection (NEPCC # 8a), where they had been growing at 7 -10 umol photons m~2 s"l (under light limitation and potentially also nutrient limitation), in enriched natural seawater. They were placed under continuous saturating irradiance (-110 umol photons m_2 s"l) in artificial seawater (ESAW; Harrison et al., 1980) with the modifications as in Chapter 1 (this study). All nutrients were initially saturating except for nitrate, which was at a concentration of -10 umol L"l. Cultures were kept at 17°C and stirred with a teflon-coated stir bar at ~80 rpm. These cultures were allowed to grow until nitrate became limiting and exponential growth slowed (~4 days after inoculation). Cultures were left to become nitrogen deplete for another 3 days, and then a saturating concentration of NaNOs (>500 umol L"l) was added. Exponential growth reoccurred within 24 h. After 24 h, cells were transferred to 1 L flasks and grown for a further 4 days in N-saturated medium. Each 1 L flask was then inoculated into a 12 L low N (-10 umol L'l, as above) batch culture. The two 12 L cultures were monitored for 5 d, until unusually low growth rates and high sinking rates were observed, indicating the cultures were anomalous. Cultures were then examined microscopically, and sexual cells were observed. The first observation of sexual cells was termed the beginning of the time series, t = 0. At this time, a 2 L subsample was removed from each N-deplete 12 L culture, and spiked with saturating NaN03 to form two new N-replete cultures. Subsequently, all four cultures (two with NO3 107 added, referred to as N-replete (+N), and two without NO3 added, referred to as N-deplete (-N), were maintained at a saturating irradiance. N-deplete cultures were monitored from time t = 0 until late senescence (10 days later). N-replete cultures were monitored until >90% of the cells had been sexually induced (40 d later). Smaller subcultures of the +N cultures were then placed at: 1) 110 umol photons m-2 s-1; (light-saturated; LS), 2) ~7-10 \imol photons m"2 s'l (light-limited; LL) and c) at light-limited irradiance at 4oC (LL-cold). These cultures were then monitored for a further 4 months to ascertain under which treatment sexuality ceased. Several post-auxospore cells were isolated to form new cultures from cells which had undergone sexual recombination. These were monitored for several weeks. Sinking rates were measured using the SETCOL method (Bienfang, 1981) with the calculation modified as in Chapter 1, using cell numbers as the biomass index (rather than chlorophyll). All settling trials ran for 3 h, at 17oC at 110 umol photons m-2 s-1. Cell counts of the top, middle and bottom fractions were made first on a Coulter CounterR with a size-specific attachment, using the 200 um aperture. Samples were diluted 1:4 with filtered 3% NaCl before counting. Counts of the percent of cells at different sexual stages (Fig. 4.1) in all fractions were made under the microscope using the Utermohl method (Utermohl, 1958). The percent composition of the fractions was then multiplied by the total cell number in each fraction to yield an estimate of the numbers of each type of cell in each SETCOL fraction. The sinking rate of each cell type was then calculated separately. In addition, cell diagrams of each different cell type were traced from photographs of actual cells taken through the microscope (as above). All sinking rate time series were first checked for temporal trends. Where significant (p<0.05) temporal trends were not found, the mean sinking rate over the time series was 108 calculated. Any comparisons between sinking rates at a given time t (n=2), and time series means (n=8) were made with Student's t-test, p=0.05). Culture growth rates were calculated from the increase in fluorescence over time by regressing the logarithm vs time, and taking the slope of the regression line as one growth rate estimate (n=l). Means and S.D. represent the means of several (n = 2 to 4) regression slopes. 109 RESULTS The cell cycle The following brief description of sex in D. brightwellii is included to clarify the terminology used in this section and the discussion (for further details, see Drebes, 1977). The centric diatom sexual cycle involves the differentiation of some cells into egg-bearing cells and some cells into sperm-bearing cells, each through two meioses within the parent frustule. The frustule bends and opens, allowing sperm to exit between the two halves of the frustule, and once sperm contact an egg cell, they enter the frustule in the same manner. Fertilization involves the fusion of a single sperm with the egg, after which a new individual, the auxospore, is formed. The auxospore swells up to 3 x the diameter of the original cell, and then forms a new wider frustule. The new, or post-auxospore population, is 2-3 x the width of the old. D. brightwellii is produces 1 egg cell-1 and 32-64 sperm cell-1 (Gross, 1937; Smayda and Boleyn, 1966). Cell types and the cell cycle in the sexual phase of D. brightwellii are illustrated schematically in Fig. 4.1. Exponentially growing vegetative cells looked healthy and normal. These were termed asexual cells. Cells first indicated the onset of sexuality by becoming slightly elongated, curved, darkened and irregular in shape: these were termed early sexual cells. Gamete-bearing cells were distinctive: either cells full of male gametes, or cells with a single egg at the end of a curved cell. Zygotes were spherical, slightly larger than egg cells, with no frustule. Early post-auxospore cells were at least double the diameter of the original cells, with cytoplasm still clustered in the center of the cell. Late post-auxospore cells had chloroplasts spread throughout the new, larger frustule. 110 1 . VEGETATIVE GROWTH LATE POST-AUXOSPORE EARLY POST-AUXOSPORE EARLY SEXUAL CELLS SPERMATOGONIA ->. SEXUAL REPRODUCTION Figure 4.1. The life history of Ditylum brightwellii. All cells were traced from photographs of laboratory cultures (see text for details) 111 Field observations Long term monitoring of the phytoplankton at the Jericho Beach site by Prof. F.J.R. Taylor's laboratory indicated that the site is dominated by diatoms in the spring and fall, with dinoflagellates and other flagellates dominating in the summer months. Ditylum brightwellii has been observed to bloom briefly at Jericho Pier every fall for the last 3 years (Fig. 4.2). In 1991, the fall bloom was delayed due to unusually sunny weather, and did not occur until a storm induced wind mixing in late September. D. brightwellii reached 21,300 cells L-1 by October 28, 1991, when sexual cells of this species were clearly visible (Fig. 4.3). Egg cells were present at 100 cells L-1 and spermatogonia-bearing cells were present at 600 cells L-1, making up - 3 % of the population. By the November 5 sampling, the population no longer contained a sexual fraction. In addition to sexual cells, numerous resting spores were observed (ca. 300 cells L-1) within the population. Ditylum brightwellii was also observed to dominate other sites in the Strait of Georgia (Nelson Is. and Cape Mudge) during the delayed 1991 fall diatom bloom. Although this species has been seen to bloom in the fall at these other sites in other years, its greater dominance in 1991 was unusual (R. Haigh, unpublished results). Laboratory observations The first general indications of sexuality in the D. brightwellii cultures at t = 0 were that growth rates decreased suddenly and sinking rates were anomalously high (>0.2 m d-1). Once the culture was split into -N and +N cultures, the -N cultures' growth rate decreased rapidly and reached 0 at t = 10 d, while the +N cultures' growth rates continued to fluctuate and recovered by t = 40 (at t = 40, new vegetative cells (post-auxospores) formed most 112 Ditylum brightwellii DOMINANT PRESENT NOT PRES 30 . 3 "5 a. E ii i i t t i i i i i i i t i i i i i » i i i i i i » i t t o Temperature Salinity Fig. 4.2. Time series of presence/absence data for Ditylum brightwellii at Jericho Pier between 1989 and 1991. "Dominant" indicates periods when this species formed over 50% of cell numbers of the net phytoplankton assemblage. Physical data are from surface water samples and represent temperature (oC) and salinity (%o). 113 Fig. 4.3. Male gametes and egg cells from surface water samples at Jericho Pier, Vancouver, B.C. during the fall diatom bloom in October, 1991. Photographed under the 40 x ocular (approximately 3600 X). 114 .*£ % [» i * * * of the biomass Fig. 4.4A). Until growth ceased in the -N cultures, sinking rates in -N and +N were similar. The percent composition of different cell types in the +N and -N cultures quickly diverged (Figs. 4.4 A & B). Initially, about 30% of the cells were asexual in both treatments, and early sexual cells formed almost 60% of the total, with a few cells having reached the post-auxospore stage. In the +N cultures, the fraction of post-auxospore cells increased over time, and after 40 d, post-auxospore cells were >80% of the biomass (Fig 4.4A). In the -N cultures, after the first 3 days, the fraction of post-auxospore cells declined over time. Gametes were never formed in significant numbers, and most of the culture did not proceed through a complete sexual cycle. Instead, -N cells were arrested at the early sexual stage (Fig. 4.4B). Too few asexual cells survived in the -N cultures to determine their sinking rate. The sinking rates of asexual cells (+N) showed a significant decrease with time after the addition of nitrogen at t = 0 d. By t = 20 d, asexual cells had normal, low sinking rates (-0.002 m d_l; Fig. 4.5A). There was no significant temporal trend in the sinking rates of early sexual cells, gamete-bearing cells, or zygotes, and they were highly variable (Figs. 4.5A and 4.5B). The temporal mean from t = 0 to t = 40 d is given as a horizontal line whenever no temporal trend was found. Zygotes sank significantly faster than pre-sexual cells, but no significant differences were found between the temporal means of early sexual cells, gametes, and zygotes. Floating cells were not consistently observed for any of these cell types. Post-auxospore cells showed an exponential decrease in sinking rate over time (Fig. 4.5C). In addition, a fraction (varying betv.con 10 and 90%) of the post-auxospore cells began to float within 3 days of N addition, actually achieving substantial floating rates (up to 0.4 m 116 Time (days ) Pig. 4.4. Composition of cell types in cultures of Ditylum brightwellii in which sexual reproduction occurred over time during the experiment. CA) Cultures kept in fully nutrient-enriched medium (+N). (B) Cultures kept in medium with low nitrate content (<10 umol L-l). -N cultures did not survive past t = 10 d. 117 Fig. 4.5. Time series of sinking rates of Ditylum brightwellii cells in different stages of sexuality. For sinking rates with no time trend, mean sinking rates are given as dashed or dotted horizontal lines. For sinking rates with time trends, data points are joined. All error bars represent + 1 SD. (A). Asexual cells and early sexual cells. (B). Sexual cells (containing male or female gamete) and zygotes. (C). Post-auxospore cells: mean sinking rates and mean floating rates both shown. Mean sinking for zygotes, gametes and early sexual stages from (A) and (B) are presented here for comparison. Hatched area represents average sinking / floating rate range. 118 e c c CO Floating Rate (m tf1) • asexual v early sexual • gamete A zygote i c post-auxospore post-auxospofe 20 Thm(d) 40 119 d"1). The mean sinking rates and mean floating rates defined an average range of sinking rates for post-auxospore cells (hatched area, Fig. 4.5C) which decreased over time. In order to visualize the sinking rate progression of cells undergoing sexuality, mean sinking rates of the early sexual cells were plotted as a "time series", beginning with the initial sinking rates of the asexual cells to the high sinking rates of the zygotes and early post-auxospore cells. The actual time series of the sinking rates of the post-auxospore cells was then added as they progressed through the sexual event and finally approached "normal" sinking rates once more (Fig. 4.6) Post-auxospore cells continued to be produced and were prevalent (-25% sexual cells) in light-saturated (LS) cultures after t = 40 d, for the 4 months during which they were observed. The wide diameter cells decreased in width quickly (within 3 weeks), and growth rates remained high (~1 d-1). LL (light-limited) cells, grown at (0.4 d-1), also decreased their cell width, but were completely free of gametes, zygotes and post-auxospore cells after 4 months (only a few early sexual cells were observed in these cultures). Generally, LL cultures looked healthier than LS cultures. LL cultures kept at 4<>C (LL-cold) were intermediate. They had more (> 40%) asexual cells than LS cultures but maintained a significant fraction of early sexual cells, gametes and zygotes. 120 First Post-auxospore Cells 1.5 1-0 Sinking Rate (md-1) 0.5 -Floating y Rate (md-1) 0.5 1 1 5 Cell Types o asexual • early sexual v gamete A zygote a post-auxospore 0 5 10 15 20 25 30 Time after observation of post-auxospore cells (d) Fig. 4.6. Hypothetical sinking rate and floating rate time series of cells oiDitylum brightwellii during a sexual reproduction cycle. 121 DISCUSSION For three consecutive years, Ditylum brightwellii was observed to bloom briefly in the fall at Jericho Pier as well as other sites in the Strait of Georgia. During the delayed and extended 1991 bloom, this species became sexual. Our laboratory results indicated that there was a significant increase in the sinking rates of D. brightwellii cells as they became sexual. The formation of post-auxospore cells through sexual reproduction was followed by the development of positive buoyancy in a substantial fraction (up to 90%) of the post-auxospore cells for about 10 days. Cells subsequently placed at low light and high nutrients showed the most rapid cessation of sexuality, while cells kept at high light and high nutrients continued to become sexual for at least 4 months after initial sexual induction. A. Physiological Implications Since cells in this study went through both high light and low nutrient stress before sexuality was observed, it is unclear which was the principal factor inducing sexuality. D. brightwellii has been successfully grown asexually at the same irradiances at which sexuality was induced (Chapter 1, this study). Therefore, light alone is unlikely to be the critical factor. A combination of both the transition in light intensity and nutrient stress may have been the causing factor. Because the +N cultures were actively growing, the continued presence of a significant fraction of sexual cells in the cultures over time must have indicated continued production of sexual cells, not simply the arrestment of the sexual reproduction the middle of 122 the cycle. Sandgren's (1981) work on the Chrysophyte Dinobryon cylindricum suggested that a constant low level of sexuality (<0.01 % cells induced) might evolve as a general response to a constantly fluctuating environment, while more intense sexual events (>10 % cells induced) might occur in response to more severe cases of nutrient stress. Alternatively, the cells in the present study might also have been in an inducible stage in culture over a longer cycle (such as that suggested by Round, 1972; Lewis, 1984). However, it is clear that once sexuality is induced, it is irreversible in the short term, especially under saturating irradiances and saturating nutrient concentrations. In fact, in the present study, wide, post-auxospore cells were even observed forming gametes, undergoing a second sexual event in the same culture. This unusual observation suggests that there may have been some specific chemical messenger (or erogen; Machlis, 1972) causing the continued induction of sexuality in the cultures once the conditions initially inducing it were no longer present. Sandgren (1981) observed that in the chrysophyte Dinobryon cylindricum, a compound was released continuously into the medium by egg cells to induce sperm formation. Starr and Jaenicke (1974) found an erogen formed by male cells of Volvox carteri which induced the formation of sexual embryos. In culture, the cells would remain trapped with any cell exudates, and sexuality might be induced repeatedly. The fact that sexuality ceased at lower irradiances suggests that if an erogen is produced, its production may be light dependent. Alternatively, cells growing slowly under light limitation might simply produce less erogen, so the erogen never reached the threshold level necessary to induce sexuality in the cultures. The ecological implications of erogen production are discussed below. In this study, nitrate was necessary for gamete formation in Ditylum brightwellii. Therefore, gamete formation must demand a greater nitrogen supply than is available 123 intracellularly in a pre-sexual, vegetative cell. This suggests that if sexuality were initiated through surface nitrogen depletion during a bloom, for instance, increases in sinking rate would be essential to allow cells access to nutrients at depth and complete the sexual process. Sinking rate changes might thus be a necessary part of a larger scale ecological response to stressful conditions at the ocean surface. The physiological reasons for sinking rate increases during sexual induction may be two-fold. Firstly, the disappearance of cell vacuoles during the formation of gametes may preclude the control of intracellular density through ionic pumping, the probable mechanism for sinking control in this species (Anderson and Sweeny, 1978). Secondly, even if vacuoles were present during the sexual process, the energetic demands of sexuality might sequester cell energy needed for sinking control (Chapter 1, this study). The fact that post-auxospore cells did not have reduced sinking rates until after the dispersal of cell cytoplasm from the center of the cell suggests that sinking rate control was not possible until vacuoles had reformed. B. Ecological Implications As outlined earlier, because growth rate is generally equivalent to evolutionary fitness for unicellular organisms (Wood, 1988), the costs of abandoning asexual growth in favour of sexual reproduction are especially high in these organisms. The purpose of the following discussion is primarily to point out the potentially critical role of changes in sinking rate in modifying the selective pressures on diatoms. Such a discussion must remain speculative, since little is known about the evolutionary role of sexuality in oceanic unicellular eukaryotes. It may, however, suggest useful avenues of future investigation. 124 If sex occurs in nature as it did in this study for Ditylum brightwellii, the sexual process would quickly effect a vertical partitioning of a population into sinking and non-sinking fractions. Sexual cells sinking more rapidly than asexual cells would become physically isolated, forming a new genetically recombined population at depth, while the surviving asexual population could remain higher in the water column. If an erogen were produced by sexual cells for sexual induction of the asexual population, the physical separation of sexual and asexual fractions of a population would be the only way to limit the sexual event to a fraction of the population. This would be critical in minimizing the high "cost of recombination" (i.e., the loss of a fit parental genotype) (Stearns, 1987). In addition, it could limit the interruption of mitosis, or the "cellular-mechanical cost" of sex, to a small fraction of the population. This would be important in unicellular eucaryotes, where cellular-mechanical costs of sex are especially high (Lewis, 1987). In a dilute medium, the production of an erogen and/or erotactin as discussed above may also be important for minimizing the high "cost of fertilization" (Stearns, 1987). After the formation of post-auxospore cells, however, an erogen would become a liability, inducing recombination in already recombined cells (as observed in this study) and delaying the critical resumption of asexual growth described by Lewis (1983). Erogen production might decrease the "cost of fertilization", but without sinking rate changes, erogen production would increase both the "cellular-mechanical cost" and the "cost of recombination" to sexual diatoms. Sinking rate changes after induction and vertical partitioning of the population would cause any erogen to be rapidly diluted. A sexual event in a population might therefore have its time scale set by the sinking rates of sexual cells, and there could be strong selection for sinking rate increases. After the formation of post-auxospore cells, the existence of a 125 wide range of sinking rates, and the formation of a floating fraction, would allow a second vertical partitioning of the population, and possible recolonization of the photic zone by floating cells. In Ditylum hrightwellii, substantial floating rates have only been observed in wide, post-auxospore cells. Generally, the ability for diatoms to float increases with cell size (Villareal, 1988). Cell size changes observed in this study did not seem to follow the hypothesis of gradual cell width reduction over time in this study. Instead, there was a rapid period of cell-width decline followed by maintenance of the original, pre-sexual cell width. One might speculate that the transient cell size enlargement observed in this study may be an adaptation allowing post-sexual flotation, of recombined cells. This would effect the second partitioning event in a sexual population, and thus minimize both the risk of lengthening the time of sexuality, and the risk of losing a fit parental genotype. Finally, the increase in sinking rates and in sinking rate range could increase cell-cell contact rates. This increase in contact rates might help speed aggregation of cells into clumps, further increasing contact rates, and further increasing sinking rates (see Jackson, 1990). If the environment were shallow enough to allow cells to reach the bottom, cells accumulating there would have a still greater chance of encountering a mate. Overall, the effect of sinking rate changes during a sexual event would be: 1) to remove the stressed population rapidly from the photic zone, 2) to allow completion of the sexual process through increased nutrient availability, 3) to cause the formation of a new, recombined population at depth, and 4) to allow recolonization of the upper water layers by enlarged floating cells. Sexuality, coupled with its accompanying sinking rate and floating rate changes, may have a special role in diatoms, as part of a diatom's large scale physiological response to stress as well as a seeding mechanism. 126 There is a growing awareness of clonal diversity in marine phytoplankton (Brand et al., 1981; Brand, 1982; Krawiec, 1982). The observations in this study would be consistent with the assertion by Brand (1982) that genetic recombination is an important mechanism for the expansion of a species' biogeographic range in the ocean. Further study on the diatoms' genetic diversity in relation to their "patch"-type preference would significantly advance understanding of diatom evolution. For instance, measuring the frequency of polymorphisms of critical enzymes governing light-harvesting, carbon fixation, nutrient uptake and temperature tolerances might yield greater understanding of the difference between the phenotypic plasticity (as measured in nutrient- and photo-adaptation experiments) and true genetic diversity. Given the new information provided in this study regarding sinking processes in vegetative cells and the sinking rates changes which occur during sexual reproduction, one can speculate on the role of sinking more generally in a diatom's life history. A diatom's life history strategy should balance the avoidance of unfavourable conditions (through sinking or physiological shut-down) with mechanisms to re-seed the surface once conditions again become favourable (Garrison, 1984; Smetacek, 1985b; Pitcher, 1986). Repeated vertical partitioning in time and space may increase the chances of a cell encountering a favourable "patch" for resumption of vegetative growth. Each life history stage (vegetative asexual growth, sexuality, and vegetative resting spore formation) offers the opportunity for escape from and reinoculation of the surface layer, but the dynamics of each stage is distinct and has different temporal and spatial dynamics. While spore formation supplies resistant propagules to the benthos, sexual auxospore formation may function both to generate and maintain a diversity of genotypes and to reinoculate them to the surface periodically, in the short- to medium-term. The function of 127 vegetative cells is rapid division, for both immediate dispersal to new "patches" and for the formation of an abundant vegetative seed population (which is necessarily less resistant to degradation than spores). In fact, a wide variety of life history strategies has been observed in diatoms. In this chapter, for example, all three life history stages were observed simultaneously during the extended bloom of Ditylum brightwellii. In one species, Leptocylindrus danicus, the resting spores are formed by sexual reproduction (Davis et al., 1980). There is a complete absence of observations of resting spores in Skeletonema costatum, and it is able to undergo sexual reproduction almost continuously (Harrison, 1974). In the highly variable environment of the upper ocean, it is not surprising that diatoms can have a wide range of life history strategies, and that some can invest simultaneously in all three life history stages. 128 CONCLUSIONS 1. Chapter 1 provides the first quantification of the negative exponential relationship between cell respiration rate and cell sinking rate for a large diatom. It also suggests that there may be a size-dependent spectrum of energetic control. Describing sinking rate changes in terms of cell energetics should allow a better conceptual integration of diatom sinking processes into models of diatom growth and sedimentation. 2. The initial component of the sub-arctic field study in Auke Bay, Alaska (Chapter 2), presents the first estimates of threshold nutrient concentrations below which diatom sedimentation might be initiated in the field. The effect of the physical properties such as chain length and small-scale aggregation on sinking rates in this study were not as important as the effects of light and nutrients. Sinking rate reduction at the nitricline (where cell internal nitrate pools were much higher than at the surface) was probably important in the formation of the chlorophyll maximum. 3. The sedimentation of bloom diatoms in Auke Bay, Alaska was highly species-specific, dominated primarily by the large, nutrient-sensitive species such as Thalassiosira spp. (Chapter 3). This chapter presents the first evidence for a clear species-specific relationship between surface cell numbers and total sedimented cells during a bloom in a coastal ecosystem. There is evidence that bloom initiation, species succession during the bloom, and the eventual sedimentation of a spring bloom are all influenced by the nutrient-mediated dynamics of the sinking process. The evolution of a temporal "window" for blooming in the spring may have included the development of different levels of tolerance for 129 nutrient depletion. The effect of nutrient depletion on diatom sinking rates should be predictable on the basis of a species' nutrient-demand and the extent of its energy dependence for sinking control. 4. Chapter 4 presents the first sinking rate time series of a diatom undergoing sexual reproduction. In Ditylum brightwellii, cells increased their sinking rates as they became sexual, and after recombination a fraction of cells began to float at substantial rates (0.3 m d-1 upward). The sinking and floating rate changes during sexual reproduction may affect the evolutionary costs and benefits of sex for diatoms. A diatom's life history consists of three principal stages, vegetative growth, sexual reproduction and resting spore formation. Each life history stage offers the opportunity for rapid escape from, and eventual recolonization of, the surface layer. Vegetative cells act on relatively short time and space scales (hours to months) while resting spores act on longer time scales (weeks to years). It is suggested that sexual reproduction acts as an escape-recolonization mechanism on an intermediate time scale. A species' life history may represent a balance between the frequencies of these three stages, a balance which should depend on the regime to which a given species is adapted. 129 <x REFERENCES Alldredge, A.L. and C.C. Gotschalk. 1988. In situ settling behavior of marine snow. Limnol. Oceanogr. 33:339-351 Allen, M.B. 1971. High latitude phytoplankton. Ann. Rev. Ecol. Syst. 2:261-276 Allen, W.E. 1923. Observations on the surface distribution of marine diatoms off lower California in 1921. Proc. Calif. Acad. Sci. 12:437-442 Anderson D.M., J.J. Lively, E.M. Reardon and C.A. Price. 1985. Sinking characteristics of dinoflagellate cysts. Limnol. Oceanogr. 30:1000-1009 Anderson, L. and B. Sweeney. 1977. Diel changes in the sedimentation characteristics of Ditylum brightwellii, a marine centric diatom: changes in cellular lipids and effects of respiratory inhibitors and ion-transport modifiers. Limnol. Oceanogr. 22:539-552 Anderson, L. and B. Sweeney. 1978. Role of inorganic ions in controlling sedimentation rate of a marine centric diatom Ditylum brightwellii. J. Phycol. 14:204-214 Armstrong F.A., C.R. Stearns, and J.D.H. Strickland. 1967. Measurement of upwelling and subsequent biological processes by means of the Technicon Autoanalyzer and associated equipment. Deep Sea Res. 14:381-389 130 Bennett A.F. and K.L. Denman. 1985. Phytoplankton patchiness: inferences from particle statistics. J. Mar. Res. 43:307-355 Berger W.H., V. Smetacek, and D. Wefer. 1989. Ocean productivity and paleoproductivity: An overview. In Productivity of the Ocean: Present and Past, pp. 1-34. Ed. by W.H. Berger, V. Smetacek, and G. Wefer. New York: Wiley & Sons Ltd. Bernstein, H., H.C. Byerly., F.A. Hopf and R.E. Michod. 1985. The evolutionary role of recombinational repair and sex. Int. Rev. Cyton. 96:1-28 Bienfang, P.K. 1981a. Sinking rates of heterogeneous temperate phytoplankton populations. J. Plankton Res. 3:235-253 Bienfang, P.K 1981b. Sinking rates of Cricosphaeara carterae Braarud II. Senescence response to various limiting substrates in non-steady state populations. J. Exp. Mar. Biol. Ecol. 49:235-244 Bienfang, P.K. 1981c. SETCOL - A technologically simple and reliable method for measuring phytoplankton sinking rates. Can. J. Fish. Aquat. Sci. 38:1289-1294 Bienfang, P.K., L. Conquest and E. Antill. 1986. Compositional and buoyancy responses to nutrient deficiency in marine phytoplankton: a technical report of in vitro experimentation. The Oceanic Institute, Makapuu Point, Waimanalo, HI 131 Bienfang, P.K, P.J. Harrison, and L.M. Quarmby. 1982. Sinking rate responses to depletion of nitrate, phosphate and silicate in four marine diatoms. Mar. Biol. 67:295-302 Bienfang, P.K and P. J. Harrison. 1984a. Co-variation of sinking rate and cell quota among nutrient-replete marine phytoplankton. Mar. Ecol. Prog. Ser. 14:297-300 Bienfang, P.K and P. J. Harrison. 1984b. Sinking rate response of natural assemblages of temperate and subtropical phytoplankton to nutrient depletion. Mar. Biol. 83:293-300 Bienfang, P.K and J.P. Szyper. 1982. Effects of temperature and salinity on sinking rates of the centric diatom Ditylum brightwellii. Biol. Oceanogr. 1:211-223 Bienfang, P.K, J. Szyper, and E. Laws. 1983. Sinking rate and pigment responses to light-limitation of a marine diatom: implications to dynamics of chlorophyll maximum layers. Oceanol. Acta 6:55-62 Boleyn, B.J. 1972. Studies on the suspension of the marine centric diatom Ditylum brightwellii (West) Grunow. Int. Revue ges. Hydrobiol. 57:585-597 Brand, L. 1982. Genetic variability and spatial patterns of genetic differentiation in the reproductive rates of the marine coccolithophores Emiliania huxleyi and Gephyrocapsa oceanica. Limnol. Oceanogr. 27:236-245 132 Brand, L., L.S. Murphy, R.R.L. Guillard and H.-T. Lee. 1981. Genetic variability and differentiation in the temperature niche component of the diatom Thalassiosira pseudonana. Mar. Biol. 62:103-110 Broeker, W. S. 1991. Keeping global change honest. Global Biogeochemical Cycles 5:191-192 Brun-Cottan, J.-C. 1976. Stokes settling and dissolution rate model for marine particles as a function of size distribution. J. Geophys. Res. 81:1601-1606 Brzezinski, M.A. and D.M. Nelson. 1988. Differential cell sinking as a factor influencing diatom species composition for limiting nutrients. J. Exp. Marine Biol. Ecol. 119:179-200 Burrell, D.C. 1988. Carbon flow in fjords. Oceanogr. Mar. Biol. Ann. Rev. 26:143-226 Caperon, J. and J. Meyer. 1972. Nitrogen-limited growth of marine phytoplankton. I. Changes in population characteristics with steady-state growth rate. Deep-Sea Res. 19:601-618 Claustre, H., and J. Gostan. 1987. Adaptation of biochemical composition and cell size to irradiance in two microalgae: possible ecological implications. Mar. Ecol. Prog. Ser. 40:167-174 133 Collos, Y. 1982. Transient situations in nitrate assimilation by marine diatoms III: Short term uncoupling of nitrate uptake and reduction. J. Exp. Mar. Biol. Ecol. 62:285-295 Conover, R.J. and P. Mayzud. 1984. Utilization of phytoplankton by zooplankton during the spring bloom in a Nova Scotia inlet. Can. J. Fish. Aquat. Sci. 41:232-244 Cook. J.R. 1963. Adaptations in growth and division in Euglena effected by energy supply. J. Protozool. 10:436-444 Costello, J.C. and S.W. Chisholm. 1981. The influence of cell size on growth rate on Thalassiosira weissflogii. J. PLankton Res. 3:415-419 Culver, M.E. and W.O. Smith. 1989. The effects of environmental variation on the sinking rates of marine phytoplankton. J. Phycol. 25:262-270 Davies, J.M. and R. Payne. 1984. Supply of organic matter to the sediment in the northern North Sea during a spring phytoplankton bloom. Mar. Biol. 78:315-324 Davis, CO., P.J. Harrison, and R.C. Dugdale. 1973. Continuous culture of marine diatoms under silicate limitation I. Synchronized life cycle of Skeletonema costatum. J. Phycol. 9:175-180 134 Davis, CO., J.T. Hollibaugh, D.L.R. Siebert, W.H. Thomas and P.J. Harrison. 1980. Formation of resting spores by Leptocylindrus danicus (Bacillariophyceae) in a controlled experimental ecosystem. J. Phycol. 16:296-302 Delieu, T. and D.A. Walker. 1972. An improved cathode for the measurement of photosynthetic oxygen evolution by isolated chloroplasts. New Phytol. 71:201-225 Denman, K.L. and J.M. Powell. 1984. Effects of physical processes on the planktonic ecosystems in the coastal ocean. Ocean. Mar. Biol. Ann. Rev. 22:125-168 Dortch, F.Q. 1980. Nitrate and ammonium uptake and assimilation in three marine diatoms. PhD Thesis. University of Washington, Seattle WA, U.S.A. 299 p. Drebes G. 1966. On the life history of the marine plankton diatom Stephanopyxis palmeriana (Grav.) Grunow. Helgolander wiss. Meeresunters. 13:101-114 Drebes G. 1977. Sexuality. In D. Werner, ed., The Biology of Diatoms. Univ. California Press, Berkeley, pp. 250-283 Dugdale R.C., B.H. Jones, J.J. Maclsaac, J.J. Goering. 1981. Adaptation of nutrient assimilation. Can. Bull. Fish. Aquat. Sci. 210:234-250 Durbin, E.C. 1978. Aspects of the biology of resting spores of Thalassiosira nordenskioeldii andDetonula confervacea. Marine Biol. 45:31-37 135 Eilertsen, H.C., K.S. Tande, and E.N. Hegseth. 1989. The potential of herbivorous copepods for regulating the spring phytoplankton bloom in the Barents Sea. Rapp. P.-v. Reun. Cons. int. Explor. Mer. 188:154-163 Eppley, R.W., R.W. Holmes, and E. Paasche. 1967. Periodicity in cell division and physiological behavior of Ditylum brightwellii, a marine planktonic diatom, during growth in light-dark cycles. Arch. Mikrobiol. 56:305-323 Eppley R.W., J.N. Rogers, and J.J. McCarthy. 1969. Half-saturation constants for uptake of nitrate and ammonium by marine phytoplankton. Limnol. Oceanogr. 74:912-919 Falkowski, P.G. 1975. Nitrate uptake in marine phytoplankton: (nitrate-chloride)-activated ATP from Skeletonema costatum (Bacillariophyceae). J. Phycol. 11:323-326 Falkowski, P.G. 1988. Theoretical calculation of the depth of the euphotic zone in the sea, p. 47-59. In C.A. Agegian (ed.), Biogeochemical cycling and fluxes between the deep euphotic zone and other oceanic realms. National Undersea Research Program Research Report 88-1. NOAA, U.S. Dept. of Commerce French, F.W. and P.E. Hargraves. 1985. Spore formation in the life cycles of the diatoms Chaetoceros diadema and Leptocylindrus danicus. J. Phycol. 21:477-483 Gallagher, J.C. 1983. Cell enlargement in Skeletonema costatum. J. Phycol. 19:539-542 136 Garrison, D.L. 1981. Monterey Bay phytoplankton. II. Resting spore cycles in coastal diatom populations. J. Plankton Res. 3:137-156 Geider R.J. and B.A. Osborne. 1989. Respiration and microalgal growth: a review of the quantitative relationship between dark respiration and growth. New. Phytol. 112:327-341 Gibson, C.E. 1984. Freshwater Biology 14:631-638 Gieskes, W.W. and G.W. Kraay. 1975. The phytoplankton spring bloom in Dutch coastal waters of the North Sea. Netherlands J. Sea Res. 9:166-196 Goering J .C, C.J. Patton, and W.W. Shields. 1973a. Nutrient cycles. In Hood, E.W., Shields, W.W., Kelley, E.J., eds., Environmental Studies of Port Valdez. Univ. of Alaska, Fairbanks, Inst. Mar. Sci. Occ. Publ. No. 2. p. 253-271 Goering J .C, C.J. Patton and W.W. Shields. 1973b. Primary Production. In Hood, E.W., Shields, W.W., Kelley, E.J., eds., Environmental Studies of Port Valdez. Univ. of Alaska, Fairbanks, Inst. Mar. Sci. Occ. Publ. No. 2. p. 253-271 Granata, T.C. 1991. Diel periodicity in growth and sinking rates of the centric diatom Coscinodiscus concinnus. Limnol. Oceanogr. 36:132-139 137 Gross, F. 1937. Notes on the culture of some marine plankton organisms. J. Mar. Biol. Assoc. U.K. 34:229-231 Gross, D., and E. Zeuthen. 1948. The buoyancy of plankton diatoms: a problem of cell physiology. Proc. R. Soc. Edinb. 135:382-389 Guillard, R.R.L. and P. Killam. 1977. The ecology of marine planktonic diatoms. In: D. Werner, ed., The Biology of Diatoms. Univ. of California Press. Berkeley, p. 372-469 Hager F.W., L.I. Gordon, T.K. Park. 1968. A practical manual for the use of the Technicon AutoAnalyzer in seawater nutrient analysis. Tech. Rep. Dept. Oceanogr. Oregon. State. Univ. Corvallis. Haigh, R. 1988. The effect of stratification on microplankton communities in the northern Strait of Georgia. M.Sc. Thesis. University of British Columbia, Vancouver, Canada, 287 p.' Hargraves, P.E. 1982. Resting spore formation in the marine diatom Ditylum brightwellii (West) Grun. ex V.H. In D.G. Mann, ed., Proc. 7th International Diatom symposium Otto Koeltz Science Publishers, Koenigstein Harrison, P. J. 1974. Continuous culture of the marine diatom Skeletonema costatum (Grev.) Cleve under silicate limitation. PhD thesis. University of Washington, Seattle, WA USA. 140 p 138 Harrison P.J., H.L. Conway, R.W. Holmes, and CO. Davis. 1977. Marine diatoms grown in chemostats under silicate or ammonium limitation. III. Cellular chemical composition and morphology of Chaetoceros debilis, Skeletonema costatum and Thalassiosira gravida. Mar. Biol. 43:19-31 Harrison P.J., D.H. Turpin, P.K. Bienfang, and CO. Davis. 1986. Sinking as a factor affecting phytoplankton species succession: the use of selective loss semi-continuous cultures. J. Exp. Mar. Biol. Ecol. 99:19-30 Harrison, P.J., R.E. Waters, and F. J.R. Taylor. 1980. A broad spectrum artificial seawater medium for coastal and open ocean phytoplankton. J. Phycol. 16:28-35 Hasle, G. R. 1975. Some living species of the diatom family Rhizosoleniaceae. Nova Hedwegia Beih. 53:99-140 Hollibaugh J.T., D.L.R. Seibert, W.H. Thomas. 1981. Observations on the survival and germination of resting spores of three Chaetoceros (Bacillariophyceae) species. J. Phycol. 17:1-9 Hurst, L.D. 1990. Parasite diversity and the evolution of diploidy and anisogamy. J Theor. Biol. 144:429-443 Hutchinson, G.E. 1961. Paradox of the plankton. Am. Nat. 95:137-145 139 Ibelings, B.W., L.R. Mur, R. Kinsman and A.E. Walsby. 1991. Microcystis changes its buoyancy in response to the average irradiance in the mixed layer. Arch. Hydrobiol. 120:385-401 Jackson, G.A. 1990. The formation of marine algal floes by physical coagulation processes. 1990. Deep-Sea Res. 37:1197-1211 Johnson, T.J., and W.O. Smith. 1986. Sinking rates of phytoplankton assemblages in the Weddell Sea marginal ice zone. Mar. Ecol. Prog. Ser. 33:131-137 Joint, I.R., A.J. Pomroy, and G.A. Robinson. 1987. Morphological changes in Skeletonema costatum (Bacillariophyceae) during a spring bloom in a marine ecosystem enclosure. Br. Phycol. J. 22:119-124 Kamykowski D., R.E. Reed and G.J. Kirkpatrick. A comparison of sinking velocity, rotation, swimming velocity and path characteristics among six marine dinoflagellate species. Unpublished Manuscript. Dept. of Marine, Earth and Atmospheric Sciences, North Carolina State Univ., Raleigh, NC 27695-8202 USA Kanda J., D.A. Ziemann, L.D. Conquest, and P.K. Bienfang. 1989. Light-dependency of nitrate uptake by phytoplankton over the spring bloom in Auke Bay, Alaska. Mar. Biol. 103:563-569 140 Rrawiec, R.W. 1982. Autecology and clonal variability of the marine centric diatom Thalassiosira rotula (Bacillariophyceae) in response to light, temperature and salinity. Mar. Biol. 69:79-89 Kromkamp, J., A. Konopka and L.R. Mur. 1988. Buoyancy regulation in light-limited cultures of Microcystis aeruginosa. Kuwata A. and M. Takahashi. 1990. Life-form population responses of a marine planktoninc diatom, Chaetoceros pseudocurvisetus, to oligotrophication in regionally upwelled water. Mar. Biol. 107:503-512 Lampitt, R.S., T. Noji, and B. v. Bodungen. 1990. What happens to zooplankton fecal pellets? Implications for material flux. Mar. Biol. 104:15-23 Lannergren, C , and H.R. Skjoldal. 1976. The spring phytoplankton bloom in LindaspoUene, a land-locked Norwegian fjord: autotrophic and heterotrophic activities in relation to nutrients. In Persoone, G., Jaspers, E., eds., Proceedings of the 10th European Marine Biology Symposium Ostend, Belgium. Wettleren: Universal Press Vol. 2, pp. 363-391. Laws, E.A., and T.T. Bannister. 1980. Nutrient- and light-limited growth of Thalassiosira fluviatilis in continuous culture, with implications for phytoplankton growth in the ocean. Limnol. Oceanogr. 25:457-473 141 Laws, E.A., P.K. Bienfang, D.A. Ziemann, L.D. Conquest. 1988. Phytoplankton population dynamics and the fate of production during the spring bloom in Auke Bay, Alaska. Limnol. Oceanogr. 33:57-65 Legendre, L. 1980. Hydrodynamic control of marine phytoplankton production: the paradox of stability. In Nihoul, J.C.J., ed., Ecohydrodynamics. Amsterdam: Elsevier p. 191-207 Lewis, W.M. Jr. 1983. Interruption of synthesis as a cost of sex in small organisms. Am. Nat. 121:825-834 Lewis, W.M. Jr. 1984. The diatom sex clock and its evolutionary significance. Am. Nat. 123:73-80 Lewis, W.M. Jr. 1987. The cost of sex. pp. 33-57 In S.C. Steams, ed., The evolution of sex and its consequences. Birkhauser Verlag. Basel Logan, B.E., Alldredge, A.L. 1989. Potential for increased nutrient uptake by flocculating diatoms. Mar. Biol. 101:443-450 Machlis, L. 1972. The coming of age of sex hormones in plants. Mycologia 54:235-247 Mackas, D.L., K.L. Denman and M.R. Abbott. 1985. Plankton patchiness: biology in the physical vernacular. Bull. Mar. Sci. 37:652-674 142 Mackenzie, F.T. and A. Lehrman. 1988. Importance of the euphotic zone in global land and ocean C and P cycles. In Aegian, C.R., ed., Biogeochemical cycling and fluxes between the deep euphotic zone and other oceanic realms. May 1988. Nat. Undersea Res. Progress Rep. 88-1 (NOAA) pp. 33-45 Mann, D.G. 1988. Why didn't Lund see sex in Asterionella? A discussion of the diatom life cycle in nature, pp. 384-412. In Round, F.F., ed., Algae and the Aquatic Environment. Biopress Ltd. Noges, T. 1989. ATP as an index of phytoplankton productivity. The Chi a / ATP quotient. Int. Revue ges. Hydrobiol. 74:121-133 Parsons, T.R., Y. Maita, and CM. Lalli. 1984. A manual of chemical and biological methods for seawater analysis. Pergamon. Oxford Passow, U. 1991. Species-specific sedimentation and sinking velocities of diatoms. Mar. Biol. 108:449-455 Paul, A.J., J.M. Paul, and K.O. Coyle. 1989. Energy sources for first-feeding zoeae of king crab Paraithodes camtschatica (Tilesius) (Decapoda, Lithodidae) J. Exp. Mar. Biol. Ecol. 30:55-69 Paul, A.J. and K.O. Coyle. 1990a. Interannual variations in zooplankton population and biomass during the spring bloom in an Alaskan subarctic embayment. p. 179-228. In: 143 D.A. Ziemann and K.W. Fulton-Bennett (eds.), APPRISE- Interannual variability and fisheries recruitment. The Oceanic Institute. Honolulu, HI Paul, A.J. and K.O. Coyle. 1990b. Carbon consumption by Thysanoessa vaschii. p.257-261. In: D.A. Ziemann and K.W. Fulton-Bennett (eds.), APPRISE ~ Interannual variability and fisheries recruitment. The Oceanic Institute. Honolulu, HI Pingree, R.D., P.M. Holligan and G.T. Mardell. 1978. The effect of vertical stability on phytoplankton distribution. Deep-Sea Res. 25:1011-1028 Pitcher, G.C. 1986. Sedimentary flux and the formation of resting spores of selected Chaetoceros species at two sites in the southern Benguela system. S. Afr. J. mar. Sci. 4:231-244 Post, A.F., Z. Dubinsky, K. Wyman, and P.G. Falkowski. 1985. Physiological responses of a marine planktonic diatom to transitions in growth irradiance. Mar. Ecol. Prog. Ser. 25:141-149 Prezelin, B.B. 1981. Light reactions in photosynthesis. In Piatt, T., ed., Physiological bases of phytoplankton ecology. Can. Bull. Fish. Aquatic Sci. 210:1-43 Reynolds, C.S., A.E. Walsby and R T Oliver. 1987. The role of buoyancy in the distribution of Anabaena sp. in Lake Roton^ N. Z. J. Freshw. Res. 21:525-556 144 Ribier J., M. Patillon, M.-J. Falxa and J.-C. Godineau. 1988. The antarctic diatom: Stellarima microtrias (Ehrenberg) Hasle & Sims cell structure and vegetative cell enlargement in culture. Polar Biol. 8:447-455 Richter, C. 1987. Biophysical consequences of lipid peroxidation in membranes. Chemistry and Physics of Lipids 44:175-189 Riebesell, U. 1989. Comparison of sinking and sedimentation rate measurements in a diatom winter/spring bloom. Mar. Ecol. Prog. Ser. 54:109-119 Riley, G.A., H. Stommel, and D.F. Bumpus. 1949. Quantitative ecology of the plankton of the western North Atlantic. Bull. Bingham Oceanogr. Coll. 12:1-169 Roelofs, A.K. 1983. The distribution of diatoms in the surface sediments of British Columbia, Canada. PhD Thesis. University of British Columbia, Vancouver, Canada. 253 p Round F.E. 1972. The problem of cell size during diatom cell division. Nova Hedwegia 23:291-304 Sandgren, CD. 1981. Characteristics of sexual and asexual resting cyst (statospore) formation in Dinobryon cylindricum. J. Phycol. 17:199-210 Sancetta, C, Calvert, S.E. 1988. The annual cycle of sedimentation in Saanich Inlet, British Columbia: implications for the interpretation of diatom fossil assemblages. Deep-Sea Res. 35:71-90 145 Sarmiento, J.L. 1990. New production and the global carbon cycle. June 1991. Brookhaven Symposium. Skjoldal, H.R. and C. Lannergren. 1978. The spring phytoplankton bloom in Lindaspollene, a land-locked Norwegian fjord. II. Biomass and activity of net and nanoplankton. Mar. Biol. 47:313-323 Skjoldal, H.R., and P. Wassmann. 1986. Sedimentation of particulate organic matter and silicate during spring and summer in Lindaspollene, western Norway. Mar. Ecol. Prog. Ser. 30:49-63 Skreslet, S. 1988. Buoyancy in Phaeocystis pouchetii (Hariot) Lagerheim. J. Exp. Mar. Biol. Ecol. 119:157-166 Smayda, T.J. 1970. The suspension and sinking of phytoplankton in the sea. Oceanogr. Mar. Biol. Ann. Rev. 8:353-414 Smayda, T. J., and B.J. Boleyn. 1965. Experimental observations on the flotation of marine diatoms. Limnol. Oceanogr. 10:499-509 Smayda, T.J., and B.J. Boleyn. 1966. Experimental observations on the flotation of marine diatoms II. Skeletonema costatum and Rhizosolenia setigera. Limnol. Oceanogr. 11:18-34 146 Smayda, T.J., and P.K. Bienfang. 1983. Suspension properties of various phyletic groups of phytoplankton and tintinnids in an oligotrophic subtropical system. Marine Ecology 4:289-300 Smetacek, V.S. 1980. The annual cycle of sedimentation in relation to plankton ecology in western Kiel Bight. Ophelia 1:65-76 Smetacek, V.S. 1984. The supply of food to the benthos, pp. 517-547. In: M.J.R. Fasham (ed.), Flows of Energy and Materials in Marine Ecosystems. Plenum Press. New York. Smetacek, V.S. 1985a. The annual cycle of Kiel Bight plankton: a long-term analysis. Estuaries 8:145-157 Smetacek, V.S. 1985b. The role of sinking in diatom life history cycles: ecological, evolutionary and geological significance. Mar. Biol. 84:239-251 Smetacek, V., Brockel K. v., B. Zeitzschel, W. Zenk. 1978. Sedimentation of particulate matter during a phytoplankton spring bloom in relation to the hydrographical regime. Mar. Biol. 47:211-226 Smetacek, V., and F. Pollehne. 1986. Nutrient cycling in pelagic systems: a reappraisal of the conceptual framework. Ophelia 26:401-428 147 Smetacek, V. and U. Passow. 1990. Spring bloom initiation and Sverdrup's critical depth model. Limnol. Oceanogr. 35:228-234 Sommer, U. 1989. The role of competition for resources in phytoplankton succession. In: Plankton Ecology: Succession in Plankton Communities. Springer Verlag. New York Starr, R.C., and L. Jaenicke. 1974. Purification of the hormone initiating sexual morphogenesis in Volvox carteri f. nagariensis Iyengar. Proc. Nat. Acad. Sci. USA. 71:1050-1054 Stearns, S.C. 1976. Life-history tactics: a review of the ideas. The quarterly review of biology. 51: 3-47 Stearns, S.C. 1987. Why sex evolved and the difference it makes, pp 15- 27, In Stearns, S.C, ed., The Evolution of Sex and its Consequences. Birkhauser Verlag. Basel Steele, J.H. 1974. Spatial heterogeneity and population stability. Nature 248:83 Steele, R.L. 1965. Induction of sexuality in two centric diatoms. BioScience 15: 298 Stosch, v.H.A. 1958. Entwicklungsgeschichtliche Untersuchungen an zentrischen Diatomeen. II. Geschlechtszellenreifung, Befruchtung und Auxosporenbildung einiger grundbewohnender Biddulphiaceen der Nordsee. Arch. Mikrobiol. 23:327-65 148 Stosch, v.H.A. 1965. Manipulierung der Zellgroesse von Diatomeen im Experiment. Phycologia 5:21-44 Strathmann, R.R. 1967. Estimating the organic carbon content of phytoplankton from cell volume or plasma volume. Limnol. Oceanogr. 12:411-418 Suess, E. 1980. Particulate organic carbon flux in the oceans- surface productivity and oxygen utilizations. Nature 288:260-263 Takahashi M., D.L. Seibert,and W.H. Thomas. 1977. Occasional blooms of phytoplankton during the summer in Saanich Inlet, B.C., Canada. Deep Sea Res. 24:775-780 Takahashi, M., J. Burwell-Clarke, T. Whitney, P. Koeller. 1978. Phytoplankton during the winter in Saanich Inlet, B.C., Canada. J. Exp. Mar. Biol. Ecol. 31:283-301 Taylor, F.J.R., and R.E. Waters. 1982. Spring phytoplankton in the subarctic Pacific. Mar. Biol. 67:323-335 Technicon, Inc. 1977. Nitrate and nitrite in water and seawater. Technicon, Inc. Tarrytown, N.Y. (Industrial Method No. 158) Therriault J.-C. and T. Piatt. Environrnpntal control of phytoplankton patchiness. Can. J. Fish. Aquat. Sci. 38:638-641 149 Thompson, P.A. 1991. The influence of irradiance, nitrogen limitation and temperature on the biochemical composition of marine phytoplankton and their nutritional value to larval Crassostrea gigas. PhD thesis. University of British Columbia, Vancouver, B.C. Canada. 182 p Thoresen, S.S., Q. Dortch, and S.I. Ahmed. 1982. Comparison of methods for extracting intracellular pools of inorganic nitrogen from marine phytoplankton. J. Plankton Res. 4:695-704 Tilman, D. and P. Kilham. 1976. Sinking in freshwater: some ecological implications of cell nutrient status and physical mixing processes. Limnol. Oceanogr. 21:409-417 Turpin, D.H. 1991. Effects of inorganic N availability on algal photosynthesis and carbon metabolism. J. Phycol. 27:14-20 Turpin, D. H.; J. S. Parslow and P. J. Harrison. 1981. On limiting nutrient patchiness and phytoplankton growth: a conceptual approach. J. Plankton Res. 3:421-431 Utermohl, H. 1958. Zur VervoUkommnung der quantitativen Phytoplankton-Methodik. Mitt. Inst. Ver. theor. angew. Limnol. 9:1-38 Villareal, T.A. 1988. Positive buoyancy in the oceanic diatom Rhizoselenia debayana H. Peragallo. Deep Sea Res. 35:1037-1045 150 Walsby, A.E., and C.S. Reynolds. 1980. Sinking and floating. In Morris, I., ed., The Physiological Ecology of Phytoplankton. Univ. of California Press Wassmann, P. 1983. Sedimentation of organic and inorganic particulate material in Lindaspollene, a stratified, land-locked fjord in western Norway. Mar. Ecol. Prog. Ser. 13:237-248 Wassman, P., R. Peinert and V. Smetacek. 1991. Patterns of production and sedimentation in the Norwegian coastal zone, the Barents Sea, and the Norwegian Sea. pp. 137-156 in Wassmann, P., A.-S. Heiskanen, and O. Ldndhal, eds., Sediment Trap Studies in the Nordic Countries. Symposium proceedings. Kristineberg Marine Biological Station, Fiskebackskil, Sweden, Nov. 1990 Weger, H.G., R. Herzig, P.G. Falkowski and D.H. Turpin. 1989. Respiratory losses in a marine diatom: measurements by short-term mass spectrometry. Limnol. Oceanogr. 34:1153-1161 Whitelam G.C. and G.A. Codd. 1986. Damaging effects of light on microorganisms. In Microbes in extreme environments. Academic Press. London, pp. 126-169 Wilkerson, F.P., and R.C. Dugdale. 1987. The use of large shipboard barrels and drifters to study the effects of coastal upwelling on phytoplankton dynamics. Limnol. Oceanogr. 32:368-382 151 Williams, G.C. 1975. Sex and evolution. In Monographs in Population Biology, Princeton Univ. Press, Princeton Williams, P.J.leB., and B. von Bodungen. 1989. Group report: Export productivity from the photic zone. In W.H. Berger, V. Smetacek, and G. Wefer, (eds.), Productivity of the Ocean: Present and Past. Wiley & Sons Ltd. New York. p. 99-115 Wood, A.M. 1988. Molecular biology, single cell analysis and quantitative genetics: new evolutionary genetic approaches in phytoplankton ecology. In Yentsch, CM.., F. Mague and P.K. Horan, eds., Immunochemical Approaches to Estuarine, Coastal, and Oceanographic Questions, Coastal and Estuarine, Coastal and Estuarine Lecture Series. XXV. Springer, Berlin, pp. 41-7? Woods, J.D. and W. Barkmann. 1986. The response of the upper ocean to solar heating. I The mixed layer. Quart. J. Meteor. Soc. 112:1-27 Ziemann, D.A. 1985. Patterns of oceanographic conditions and particulate sedimentation in Auke Bay, Alaska, during the spring bloom, 1985. In APPRISE 1985 Annual Report, University of Alaska, Juneau, Alaska, p. 145-244 Ziemann D.A., L.D. Conquest and V. Sato. 1985. Association of primary production and recruitment in subarctic ecosystems 1985. Technical Report Vol.1. The Oceanic Institute, Waimanalo. Hawaii 152 Ziemann, D.A., L.D. Conquest, P.K Bienfang, and E. Antill. 1986. Patterns of primary production and sedimentation during the 1986 spring bloom in Auke Bay, Alaska. APPRISE Technical Report, 1986. Oceanic Institute, Makapuu Point, Waimanalo, Hawaii. Ziemann, D.A., L.D. Conquest, P.K Bienfang, and J. Kanda. 1987. Patterns of primary production and sedimentation during the 1987 spring bloom in Auke Bay, Alaska. APPRISE Technical Report, 1987. Oceanic Institute, Makapuu Point, Waimanalo, Hawaii. Ziemann, D.A., L.D. Conquest, P.K Bienfang. 1988. Patterns of primary production and sedimentation during the 1988 spring bloom in Auke Bay, Alaska. APPRISE Technical Report, 1988. Oceanic Institute, Makapuu Point, Waimanalo, Hawaii. Ziemann D.A., L.D.Conquest, KW. Fulton-Bennet, and P.K Bienfang. 1990. Interannual variability in the Auke Bay phytoplankton. In Ziemann, D.A., KW. Fulton-Bennet, eds., APPRISE: Interannual variability and fisheries recruitment. The Oceanic Institute, Honolulu, p. 129-170 1G3 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0098886/manifest

Comment

Related Items