UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The life cycle of the pteropod Limacina helicina in Rivers Inlet (British Columbia, Canada) Wang, Kang 2014

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

Item Metadata

Download

Media
24-ubc_2014_spring_wang_kang.pdf [ 6.35MB ]
Metadata
JSON: 24-1.0167397.json
JSON-LD: 24-1.0167397-ld.json
RDF/XML (Pretty): 24-1.0167397-rdf.xml
RDF/JSON: 24-1.0167397-rdf.json
Turtle: 24-1.0167397-turtle.txt
N-Triples: 24-1.0167397-rdf-ntriples.txt
Original Record: 24-1.0167397-source.json
Full Text
24-1.0167397-fulltext.txt
Citation
24-1.0167397.ris

Full Text

The Life Cycleof the pteropod Limacina helicinain Rivers Inlet (British Columbia, Canada)byKang WangB.Sc., The University of British Columbia, 2009A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMaster of ScienceinThe Faculty of Graduate and Postdoctoral Studies(Oceanography)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)April 2014c© Kang Wang, 2014AbstractThe life cycle of Limacina helicina has been continuously debated within theliterature. We believe the current lack of consensus regarding fundamentalaspects of its life cycle (e.g. seasonal times of spawning, seasonal developmentof the population size structure, as well as the life cycle longevity) is primarilydue to using datasets of low temporal resolution.Using fort-nightly data, two population cohorts were identified using themixdist statistical package and tracked for more than 400 days, throughout2008 to 2010. From this, a life cycle longevity of 1.2–1.5 years was estimated forL. helicina in Rivers Inlet. Throughout the seasons, the population size struc-ture showed a continually high presence of the smaller size-groups suggestingcontinuous spawning, however, based on total densities of > 600 ind.m−3, thelate spring was put forward as the period of peak spawning.Continuous spawning was confirmed with the use of daily data. Identifi-cation of a summer peak spawning established late spring and summer as twoperiods of enhanced spawning, although continuous spawning occurred through-out the season (in a limited fashion). Short-term periods of significant growthwere observed prior to peak spawning in late spring and summer. This wasnot directly coupled with chlorophyll concentrations, possibly due to the timelag between periods of high chlorophyll biomass and zooplankton response. At-iitempts were made to estimate the instantaneous mortality of L. helicina, andthe seasonal changes experienced from spring to summer. Our estimates werecomplicated by a combination of 1.) inherent patchiness of L. helicina, 2.) ad-vection, and 3.) merged recruits. Generally, there were no cases of significantmortality throughout the seasons however, short term mortality was observedafter peak spawning. It is plausible that the smallest size-groups of L. helicinaexperiences the highest mortality after peak spawning.Our findings show that in Rivers Inlet, L. helicina has a life cycle spanning1–1.5 years with spring and summer peak spawning activities. The spring cohortis likely spawned by the summer cohort from the previous year. It utilizes thespring phytoplankton bloom to reach sexual maturity and spawn the summercohort.iiiPrefaceThis thesis is ultimately a product of the collaborative effort of many in-vestigators from the University of British Columbia, Point Grey campus, of theinter-disciplinary RIES - Rivers Inlet Ecosystem Study conducted from 2008 to2010 (riversinlet.eos.ubc.ca/People.html). The aims of this thesis are separateand not directly connected with the main objectives of the RIES. Additionalwinter sampling from 2010–2011 was performed by Wayne Jacobson with direc-tion from Brian Hunt and the Hakai Institute. Results from other studies in theRIES were used throughout Chapters 2 and 3 to help rationalize our findings.Formalin preserved zooplankton samples for the bi-weekly and monthly sam-pling in Chapter 2, as well as the daily zooplankton samples used in Chapter3, were processed by myself. Prior to being considered for this thesis, the dailyzooplankton samples in Chapter 3 was used by an undergraduate student as thebasis of a directed studies project, guided by Evgeny Pakhomov.The cohort analysis in Chapters 2 and 3, performed using the mixdist sta-tistical package (for the R statistical programming language), was conducted bymyself with guidance from Evgeny Pakhomov and Brian Hunt. Complicated de-tails concerning the use of mixdist were clarified through email communicationwith the author, Peter MacDonald.ivTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiiDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv1 An Introductory Review . . . . . . . . . . . . . . . . . . . . . . . . 11.1 What Are Pteropods . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Taxonomy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.3 Thecosome Morphology . . . . . . . . . . . . . . . . . . . . . . . 31.4 Swimming and Buoyancy Regulation . . . . . . . . . . . . . . . . 41.5 Food and Diet . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.6 Reproductive Biology . . . . . . . . . . . . . . . . . . . . . . . . 61.7 Ecological Role, Ecological Threats, and Knowledge Gaps . . . . 71.7.1 Ocean Acidification . . . . . . . . . . . . . . . . . . . . . 71.7.2 Life Cycle of L. helicina – Knowledge Gaps . . . . . . . . 8v2 Life-Cycle Dynamics of Limacina helicina in Rivers Inlet B.C. 102.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.1.1 Research Aims . . . . . . . . . . . . . . . . . . . . . . . . 102.1.2 Limacina helicina: Past Life Cycle Investigations . . . . . 112.1.3 Rivers Inlet: Historical Context . . . . . . . . . . . . . . . 132.1.4 Goals and Aims . . . . . . . . . . . . . . . . . . . . . . . . 142.2 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152.2.1 Study Area – Rivers Inlet . . . . . . . . . . . . . . . . . . 152.2.2 Seawater Temperature, Fluorescence, Salinity . . . . . . . 152.2.3 Sample Collection and Selection . . . . . . . . . . . . . . 162.2.4 Sample Preparation and Enumeration . . . . . . . . . . . 172.2.5 Size Frequency Histograms and Identification of Cohorts . 182.2.6 Estimation of Growth and Life-Cycle Longevity . . . . . . 192.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.3.1 Environmental Parameters - Temperature, Salinity, Fluo-rescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 252.3.2 L. helicina Abundance: Seasonal and Inter-annual Variation 252.3.3 L. helicina Size Structure: Seasonal Development . . . . . 262.3.4 Spawning Activity . . . . . . . . . . . . . . . . . . . . . . 272.3.5 Estimate of Life-Cycle Longevity . . . . . . . . . . . . . . 282.3.6 Rivers Inlet Spatial Analysis . . . . . . . . . . . . . . . . 312.4 Discussion: Life-cycle re-evaluation . . . . . . . . . . . . . . . . . 372.4.1 Seasonal Spawning and Recruitment . . . . . . . . . . . . 372.4.2 Seasonal Growth and Environmental Correlations . . . . . 382.4.3 Rivers Inlet Spatial Analysis . . . . . . . . . . . . . . . . 402.4.4 Potential Sampling Errors . . . . . . . . . . . . . . . . . . 403 Seasonal Growth and Mortality – Spring vs. Summer . . . . . 42vi3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 423.2 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.3 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463.3.1 Study Area & Sample Collection . . . . . . . . . . . . . . 463.3.2 Daily Fluorescence . . . . . . . . . . . . . . . . . . . . . . 463.3.3 Size-Frequency Histograms & Identification of Cohorts . . 473.3.4 Spawning Events . . . . . . . . . . . . . . . . . . . . . . . 483.3.5 Shell Size Growth and Mortality . . . . . . . . . . . . . . 483.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 513.4.1 Daily Chlorophyll . . . . . . . . . . . . . . . . . . . . . . 513.4.2 Daily Population Abundance . . . . . . . . . . . . . . . . 513.4.3 Daily Population Size-Structure . . . . . . . . . . . . . . . 523.4.4 Spawning Events . . . . . . . . . . . . . . . . . . . . . . . 523.4.5 Cohorts Identified and Tracked . . . . . . . . . . . . . . . 533.4.6 Seasonal Growth . . . . . . . . . . . . . . . . . . . . . . . 543.4.7 Daily Mortality . . . . . . . . . . . . . . . . . . . . . . . . 543.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 593.5.1 Spawning, Cohorts, and Size-Structure Development: Com-parison to Chapter 2 and Relevant Literature . . . . . . . 593.5.2 Caveats to Estimating Daily Growth . . . . . . . . . . . . 623.5.3 Estimating Daily Mortality and Problems Encountered . 633.5.4 Potential Sampling Errors . . . . . . . . . . . . . . . . . . 654 Life Cycle of L. helicina: A Conceptual Model and GeneralConclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 694.1 A Conceptual Model . . . . . . . . . . . . . . . . . . . . . . . . . 714.2 General Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 74viiBibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84A Chapter 2: Supplementary Data . . . . . . . . . . . . . . . . . . . 85A.1 2010–2011 Winter-Transition . . . . . . . . . . . . . . . . . . . . 85A.2 Seasonal Correlations - Physical Parameters & Population Abun-dance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87B Chapter 3: Supplementary Data . . . . . . . . . . . . . . . . . . . 90B.1 Daily Data Sampling Dates . . . . . . . . . . . . . . . . . . . . . 90B.2 Size-Frequency Histograms - March, April, May, June, July . . . 93B.3 Finite Mixture Distributions – Daily Data . . . . . . . . . . . . . 98B.4 What Are Finite Mixture Distributions . . . . . . . . . . . . . . 98B.5 Fitting Finite Mixtures . . . . . . . . . . . . . . . . . . . . . . . 99B.6 Finite Mixture Distributions – Statistical Output . . . . . . . . . 109B.7 Life Tables for Cohorts Tracked . . . . . . . . . . . . . . . . . . . 115B.8 Seaonal Growth Rate . . . . . . . . . . . . . . . . . . . . . . . . 135B.9 Environmental Connection . . . . . . . . . . . . . . . . . . . . . . 137B.10 Seasonal Mortality . . . . . . . . . . . . . . . . . . . . . . . . . . 139viiiList of Tables2.1 Limacina shell size summary statistics (max, min, mean) for L.helicina individuals enumerated in the fort-nightly and monthlysamples at station DFO 2 . . . . . . . . . . . . . . . . . . . . . . 212.2 Survey sampling dates for the samples processed, for spatial anal-ysis of stations DFO 1, DFO 2, DFO 3, DFO 4, and DFO 5 inRivers Inlet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232.3 A life cycle table for the cohorts C1 and C2, tracked from March2008 to July 2010. . . . . . . . . . . . . . . . . . . . . . . . . . . 29A.1 Linear regressions testing the relation between 30 m depth aver-aged temperature and salinity, and 30 m depth averaged fluores-cence – 2008 season . . . . . . . . . . . . . . . . . . . . . . . . . . 87A.2 Linear regressions testing the relation between 30 m depth aver-aged temperature and salinity, and 30 m depth averaged fluores-cence – 2009 season . . . . . . . . . . . . . . . . . . . . . . . . . . 88A.3 Linear regressions testing the relation between 30 m depth aver-aged temperature and salinity, and 30 m depth averaged fluores-cence – 2010 season . . . . . . . . . . . . . . . . . . . . . . . . . . 89A.4 Linear regressions testing the relation between 30 m depth aver-aged temperature and salinity, and 30 m depth averaged fluores-cence – 28 February to 2 June, 2009 . . . . . . . . . . . . . . . . 89ixB.1 Daily dates of sample collection . . . . . . . . . . . . . . . . . . . 90B.3 Life Tables of Population components tracked . . . . . . . . . . . 116B.4 Regression table for observed periods of increased shell growth inthe population . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136B.5 Regression table for observed periods of increased shell growth inthe population . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137B.6 Statistical results of linear regressions testing the relation be-tween chlorophyll and the daily variation in population abun-dance, for each month . . . . . . . . . . . . . . . . . . . . . . . . 138B.7 Statistical results of linear regressions testing for periods of in-creased mortality, for cohorts identified and followed in the dailytime series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139xList of Figures2.1 L. helicina with measure of shell diameter . . . . . . . . . . . . . 172.2 Map of west coast of British Columbia and of Rivers Inlet . . . . 202.3 2-panel figure showing A. the seasonal distribution of 30 m depth-averaged salinity and 30 m depth averaged temperature for 2008,2009, and 2010, and B. the seasonal distriburtion of L. helicinaabundance throughout 2008, 2009, 2010, and the 2010–2011 win-ter. Also shown is the seasonal variation of 30 m depth integratedfluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322.4 Finite-mixture distributions showing the seasonal development ofthe L. helicina population size-structure . . . . . . . . . . . . . . 332.5 Population variation with proportional abundance of smaller im-matures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 342.6 Population size structure of L. helicina for stations DFO 1–5 at3 seasonal time points in 2010 . . . . . . . . . . . . . . . . . . . . 352.7 2-panel figure showing A. the seasonal distribution of the averageshell size of the population across 2008, 2009, and 2010, and B.The growth cycles of the three cohorts tracked (C1, C2, C3) . . . 363.1 Map of the west coast of British Columbia and Rivers Inlet, withthe location of the Daily Station at Dawsons Landing. . . . . . . 50xi3.2 Composite 5x1 figure of the daily dynamics observed, for L. he-licina at Dawsons Landing . . . . . . . . . . . . . . . . . . . . . . 563.3 Daily development of the seasonal size structure in the populationof L. helicina at Dawsons Landing. . . . . . . . . . . . . . . . . . 573.4 Composite figure of log-transformed daily density, for each cohorttracked. Mortalitiy is indicated by a decreasing abundance. . . . 584.1 Conceptual model of the life-cycle of L. helicina . . . . . . . . . . 76A.1 Composite 2x4 histograms of size-frequency histograms of the L.helicina size-structure through the 2010–2011 winter-transition. . 86B.1 Ch3. Population size structure histograms – 22–31 March, 2010 . 94B.2 Ch.3 Population size structure histograms – 1–15, 17–22, 24–30April, 2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95B.3 Ch.3 Population size structure histograms – 1–20, 22–31 May, 2010 96B.4 Ch.3 Population size structure histograms – 1–27, 29–30 June, 2010 97B.5 Ch.3 Population size structure histograms . . . . . . . . . . . . . 98B.6 Ch.3 Finite Mixtures – 22–31 March, 2010 . . . . . . . . . . . . . 101B.7 Ch.3 Finite mixture distributions – 1-15 April, 2010 . . . . . . . 102B.8 Ch.3 Finite mixture distributions – 17-22, 23-30 April . . . . . . 103B.9 Ch.3 Finite mixture distributions – 1–15 May, 2010 . . . . . . . . 104B.10 Ch.3 Finite mixture distributions – 16-20, 22-31 May, 2010 . . . 105B.11 Ch.3 Finite mixture distributions – 1–16 June, 2010 . . . . . . . 106B.12 Ch.3 Finite mixture distributions – 17–27, 29–30 June, 2010 . . . 107B.13 Ch.3 Finite mixture distributions – 1–7 July, 2010 . . . . . . . . 108xiiAcknowledgementsI want to thank Dr. Evgeny Pakhomov for giving me the opportunity to becomeinvolved in oceanographic research, even before I was an M.Sc. student. Yourextensive expertise and patience has motivated me to keep pushing forward inthe my search for the answers, and your calm demeanour during my times ofdifficulty have always reminded me to keep focused and proceed one step at atime.Thank you Dr. Brian Hunt for all of your supporting help throughout thisgraduate degree. In particular, your meticulous attention to detail has providedme with the motivation to further improve my scientific writing in order tobetter communicate my thoughts and opinions, to the general public.Additional thanks to my committee members, Dr. Christopher Harley andDr. Mary O’Connor for giving me new perspectives throughout the challengingtimes of data interpretation. Thank you to Lora Pakhomov for all the supportyou’ve provided during field work. Moira Galbraith, Doug Yelland, and the restof the scientific community of the C.C.G.S. John P. Tully, thanks for all yourassistance in the field.To all the members of the P-Lab Family, old and new, thank you all for yourfriendship as well as your positive and constructive feedback. Special thanks toxiiiDesiree Tommasi for all of your support and encouragement, when it seemed likethere was no one else, and to all the other oceanographers in the department.I have enjoyed many a lunchtimes with you, drinking tea/coffee, and talkingnon-science, as well as sharing many laughs.Last but not least I want to thank my father for the unconditional help hehas given me throughout my studies and in life.xivDedicationDad, this thesis is dedicated to you. For your patience, for your spiritual (...andfinancial) support, and for your continued encouragement to pursue whicheverpath I choose.xvChapter 1An Introductory Review1.1 What Are PteropodsPteropods are pelagic snails (sea snails, sea slugs) and are one of five groupsof holoplanktonic gastropods, ranging in size from 0.2 mm to > 80 mm inlength (Lalli and Gilmer, 1989). Living primarily within the euphotic zone,some pteropods (e.g. Limacina trochiformes) have been observed at depths of≥ 2km (Lalli and Gilmer, 1989). The term “pteropod” refers to two taxonomicorders of gastropods, thecosomes and gymnosomes. Thecosomatous pteropodsare housed within a calcareous shell while gymnosomes are shell-less (Lalli andGilmer, 1989). Only the Thecosomes will be discussed as the remaining thesisfocuses on species from this order.Distributed throughout the oceans, thecosome diversity is highest in tropicalwaters where few species are found in high abundance (Lalli and Gilmer, 1989).Conversely, the waters of high-temperate and polar latitudes are home to a fewspecies that occur in high abundance (Lalli and Gilmer, 1989; Kobayashi, 1974;Gannefors et al., 2005; Hunt et al., 2008; Bednarsˇek et al., 2012). Limacina11.2. TAXONOMYretroversa and Limacina helicina are two such species, with L. retroversa beingobserved only within the Atlantic. L. helicina was postulated as a bipolar specieswith the Northern Hemisphere L. helicina helicina (forma acuta, helicina, andpacifica) found in the waters of the polar Arctic, and L. helicina antarctica(forma antarctica and rangi) found in the Southern Oceans (Hunt et al., 2010).Subsequent molecular analysis in Hunt et al. (2010) revealed genetic differencesbetween the Arctic and Antarctic forms of L. helicina, alluding to the twovariants (e.g. L. helicina helicina and L. helicina Antarctica) being distinct.Both Northern and Southern variants appear to have a skewed distribution witha larger ratio of small juveniles to larger adults (Hunt et al., 2008). Bednarsˇeket al. (2012) reports the contribution of the larger size-fractions to be ≤ 2% ofthe population.1.2 TaxonomyTwo suborders comprise the Thecosome order, the calcareous Euthecosomataand the pseudoconch bearing Pseudothecosomata (Lalli and Gilmer, 1989). Thefollowing presents the classification from Lalli and Gilmer (1989).Class GastropodaSubclass OpisthobranchiaOrder ThecosomataSuborder EuthecosomataFamily LimacinidaeFamily CavoliniidaeSuborder PseudothecosomataFamily Peraclididae21.3. THECOSOME MORPHOLOGYFamily CymbuliidaeFamily DesmopteridaeOrder GymnosomataSuborder GymnosomataFamily PneumodermatidaeFamily NotobranchadeidaeFamily CliopsidaeFamily ClionidaeSuborder GymnopteraFamily HydromylidaeFamily LaginiopsidaeThe above scheme is deemed suitable such that suborders and families withinThecosomata are easily distinguished, visually.1.3 Thecosome MorphologyEuthecosomes and Pseudotheosomes are differentiated by shell morphology and/orshell surface features (Lalli and Gilmer, 1989). In the Euthecosome suborder,the family Limacinidae are identified by a sinistrally coiled shell (< 1 mm to 15mm). (Lalli and Gilmer, 1989). Pseudothecosomes include species that possessa calcareous shell (Family Peraclididae), species lacking a calcareous shell butpossessing a pseudoconch (Family Cymbuliidae), and species that lack both acalcareous shell and a pseudoconch (Family Desmopteridae) (Lalli and Gilmer,1989). Similar to Limacinidae, Peraclididae are also characterized by a sinis-trally coiled shell although Peraclididae feature distinctive ornamented patternson the shell surface. The wings are also fused together into a single “wing31.4. SWIMMING AND BUOYANCY REGULATIONplate” whereas Limacinidae possess separate wings. Cymbuliidae are consider-ably larger (35 mm to 80 mm in length), with some species (in genera Cymbulia,Corolla, Gleba) lacking an external shell but possessing a pseudoconch (Lalli andGilmer, 1989). Lacking a calcareous shell as well as a pseudoconch, Desmopteri-dae are characterized by a cylindrical body with fused wing plates attached anda single pair of symmetrical tentacles (Lalli and Gilmer, 1989)1.4 Swimming and Buoyancy RegulationThecosome swimming is a process that involves phases of active swimming andpassive sinking (Lalli and Gilmer, 1989). Conover and Paranjape (1977) define a“rest” and “active” stage of locomotion for L. retroversa and Gilmer and Harbi-son (1986) state that these two stages likely apply for all species of Thecosomes.Despite being thin and comparatively light, the calcareous shell is sufficientlyheavy (50% of dry weight) to cause gradual sinking (Conover and Paranjape,1977). In order for Thecosomes to maintain the same depth distribution, activeswimming must be maintained for 25% of the swimming process (Conover andParanjape, 1977). Sinking rates will differ depending on the method of buoy-ancy regulation (Lalli and Gilmer, 1989). Limacinidae have limited approachesto regulate buoyancy while Cavoliniidae are able to employ different strategies(Lalli and Gilmer, 1989). Cavoliniidae shells feature spines and curvatures thatincrease the surface area of the individual (Lalli and Gilmer, 1989). Moreover,many genera within Cavoliniidae (Creseis, Hyalocylis, Styliola, Cuvierina, Clio)are able to extend the mantle lining to the shell exterior, where mucous mate-rial is exuded to mitigate sinking (e.g. Creseis acicula, Creseis virgula) (Lalliand Gilmer, 1989). Additional methods include the ability to create a neutrallybuoyant pseudoconch (Lalli and Gilmer, 1989). Lacking external shell featuresand a pseudoconch, Limacinidae utilizes a mucous feeding web to regulate buoy-41.5. FOOD AND DIETancy (Lalli and Gilmer, 1989; Hunt et al., 2008). When deployed the mucousweb prevents sinking, with the buoyant period correlated to feeding (Lalli andGilmer, 1989).1.5 Food and DietThe cessation of motion when neutrally buoyant prevents the disturbance of thedeployed mucous web by movements of the paired-wings or the wing-plate (Lalliand Gilmer, 1989). Limacinidae are considered the least specialized of feeders,producing spherical feeding webs (Lalli and Gilmer, 1989). Cavoliniidae are alsoconsidered simple feeders but only for those genera exhibiting a conical shell (e.g.Creseis, Clio, Hyalocylis, Styliola, Cuvierina) (?)lalli1989pelagic). The size ofthe food web produced is size and species dependent, with some food webs of≥ 2 m in diameter produced (Lalli and Gilmer, 1989). Pseudothecosomes pro-duce funnel shaped feeding webs or flattened rectangular webs as opposed tothe spherical webs produced by Euthecosomes (Lalli and Gilmer, 1989). Webproduction within Limacina is rapid with complete construction within 5 sec-onds (Lalli and Gilmer, 1989). Similarly, webs are retracted via. ciliary actionin about 20 seconds, although they can discarded when the feeding individualis disturbed (Lalli and Gilmer, 1989). Comparatively slower, web productionin Cavolinia can exceed 2 minutes while times of web ingestion range between1–3 minutes (Lalli and Gilmer, 1989). When feeding, individuals are orientatedwith the ventral surface facing up towards the web, and the individual below(Lalli and Gilmer, 1989). Connection between the feeding individual and thefeeding web is maintained by the proboscis (elongate appendage, Figure 25b inLalli and Gilmer, 1989). Feeding webs are pulled into the gut by a small radulawhere solid material (e.g. diatom frustules) is crushed by gizzard action (Figure27 in Lalli and Gilmer, 1989).51.6. REPRODUCTIVE BIOLOGYPhytoplankton is the main food source although the spectrum of food in-gested can be quite diverse (Morton, 1954; Kobayashi, 1974; Gilmer and Har-bison, 1991; Gannefors et al., 2005; Hunt et al., 2008; Bernard and Froneman,2009; Bednarsˇek et al., 2012). Detritus, bacteria, small protozoa, to varioustypes of phytoplankton (diatoms, dinoflagellates, tintinnids, coccolithophorids)have been found in the gut content of various species within Limacina (Lalli andGilmer, 1989). Larger food items (foraminifera, zooplankton, molluscan larvae,heteropods, chaetognaths) have also been found in the gut content of variousLimacinidae and Cavoliniidae species (Lalli and Gilmer, 1989).1.6 Reproductive BiologyDeveloping first as males and maturing as females, Thecosomes are protandrichermaphrodites that reproduce sexually with the reciprocal exchange of sper-matophores (Lalli and Gilmer, 1989). Although understudied, all species withinthe Thecosome order share similar reproductive biologies with minor differencesin reproductive structures and behaviours (Lalli and Gilmer, 1989). The growthof Limacina has been thought to be uniform with the hypothesis that matu-rity is size dependent (Kobayashi, 1974; Lalli and Gilmer, 1989). Dadon andde Cidre (1992) argue against this with the contention that maturity is in-fluenced by environmental conditions, resulting in occurrences of individuals ofdiffering age despite being the same size. With the exception of L. helicoides andL. inflata, the ontological development of Limacina progresses through the malestage with sperm produced in the ovotestis and stored within the hermaphroditicduct (Figure 31 in Lalli and Gilmer, 1989). Once matured to the female stage,the subsequent eggs produced are fertilized within the hermaphroditic duct andeventually passed through the female accessory organ to be encased by mucousfrom the female mucous gland (Figure 31 in Lalli and Gilmer, 1989). The fe-61.7. ECOLOGICAL ROLE, ECOLOGICAL THREATS, ANDKNOWLEDGE GAPSmale accessory glands lead into the mantle area where the eggs are subsequentlyreleased (Figure 31 in Lalli and Gilmer, 1989). Eggs are packaged and releasedas transparent free-floating egg masses varying in length depending on species(Lalli and Gilmer, 1989). L. retroversa has been reported to produce egg-massesranging from 2–4 mm in length while the larger L. helicina produces egg-massesup to 12 mm in length (Lalli and Gilmer, 1989). Paranjape (1968) reports of600 eggs spawned per female, for L. helicina in the coastal waters off of Van-couver Island, while up to 10,000 eggs per female have been reported for L.helicina inhabiting polar waters (Lalli and Gilmer, 1989). Up to 260 eggs perfemale have been recorded for the smaller L. retroversa inhabiting boreal wa-ters (Lalli and Gilmer, 1989). Table 13 in Lalli and Gilmer (1989) summarizesegg mass production by L. helicina and L. retroversa. Limacina helicoides isovoviviparous, with the eggs produced hatching within the mantle of the par-ent, while L. inflata exhibits brood protection with the hatched young residingwithin the parent mantle until a suitable size is reached (Lalli and Wells, 1978;Lalli and Gilmer, 1989).1.7 Ecological Role, Ecological Threats, and Knowl-edge Gaps1.7.1 Ocean AcidificationDubbed “the other CO2 problem”, ocean acidification is an increasing prob-lem for calcareous organisms, especially in polar waters (Fabry et al., 2008;Doney et al., 2009). Housed within an metastable aragonite shell and primar-ily distributed in polar waters, L. helicina is especially susceptible to oceanacidification (parts of the Arctic Ocean are already seasonally undersaturatedwith respect to aragonite) (Yamamoto-Kawai et al., 2009). This makes L. he-71.7. ECOLOGICAL ROLE, ECOLOGICAL THREATS, ANDKNOWLEDGE GAPSlicina a key indicator species of ocean acidification and a proxy of ocean health(Hunt et al., 2008; Bednarsˇek et al., 2012). While some calcareous organismsshow resilience to increased acidity, our understanding of the potential impactsto ecosystem processes are complicated by the synergistic effects of increasingacidity and seawater temperature (Dixson et al., 2010; Lischka and Riebesell,2012).Despite our current grasp of the underlying biology of L. helicina, there areconsiderable knowledge gaps surrounding fundamental aspects of its life cycle(Lalli and Gilmer, 1989; Kobayashi, 1974; Fabry, 1989; Gannefors et al., 2005;Hunt et al., 2008; Bednarsˇek et al., 2012). Questions surrounding the life cyclelongevity, seasonal spawning, and the seasonal development of the populationsize structure have yet to be resolved despite various studies in different eco-regions (Kobayashi, 1974; Fabry, 1989; Gannefors et al., 2005; Hunt et al., 2008;Bednarsˇek et al., 2012). Additionally, seasonal growth as well as mortality hasonly been investigated by a few studies of L. helicina (Kobayashi, 1974; Lischkaet al., 2010; Bednarsˇek et al., 2012).1.7.2 Life Cycle of L. helicina – Knowledge GapsThe lack of consensus surrounding the life cycle of L. helicina is complicatedby the lack of comparability between studies. Additional problems are presentin the implementation of different sampling protocols combined with the use ofdifferent sampling equipment for data collection. Noted in Hunt et al. (2008)and Bednarsˇek et al. (2012), the use of sampling nets with differing mesh sizesinfluences the size of the population captured. Furthermore, L. helicina is aspecies that spans many geographical zones which may further complicate lifecycle interpretations.The spatial and temporal scale of sampling will also impact how the data are81.7. ECOLOGICAL ROLE, ECOLOGICAL THREATS, ANDKNOWLEDGE GAPSsampled and interpreted (Wiens, 1989). Defined in Wiens (1989), the ability todistinguish temporal patterns is dependent on both the area sampled and the or-ganism studied. Investigations with large spatial extents will be accompanied byincreased heterogeneity while studies on smaller scales may reveal phenomenonthat cannot be generalized over larger scales (Wiens, 1989). For L. helicina, thetemporal resolution of data collected appears to be an important factor whenattempting to describe its life-cycle. It appears our lack of understanding maybe due to an “incomplete picture” describing the seasonal evolution of L. helic-ina, although Northern and Southern variants may live differing life-cycles dueto genetic differences (Hunt et al., 2010).This thesis investigates the life cycle of L. helicina in an attempt to resolvethe knowledge gaps concerning its life cycle dynamics. Datasets of high temporalresolution are used in two chapters to study differing aspects of the L. helic-ina— life cycle . In Chapter 2, fort-nightly and monthly resolution samplingover 3 years is used to determine the inter-annual variation of the L. helicinapopulation in Rivers Inlet. Using inter-annual data for 3 years, we attemptto portray a relatively more “complete” picture of the life cycle of L. helicina.Chapter 3 uses daily data for over 100 days (e.g. from spring to summer) toresolve the seasonal dynamics of L. helicina. With daily resolution, we aim toprovide accurate estimates of the seasonal growth and mortality experienced byL. helicina, and to describe how this changes from spring to summer. More-over, with the addition of chlorophyll data, we hope to establish a definitiveconnection between food availability and L. helicina population abundance.9Chapter 2Life-Cycle Dynamics ofLimacina helicina inRivers Inlet B.C.2.1 Introduction2.1.1 Research AimsLimacina helicina is a prominent pelagic mollusk in polar and temperate watersand is a key component of the zooplankton community (Hunt et al., 2008; Lalliand Gilmer, 1989). Potentially a significant grazer, L. helicina also acts as preyfor numerous organisms across various trophic levels (cetaceans, various fishspecies, sea-birds, and other zooplankton) (Lalli and Gilmer, 1989; Hunt et al.,2008; Bednarsˇek et al., 2012). Additionally, L. helicina is also a significantcontributor to the flux of CaCO3 to the deep sea (Fabry, 1989; Bednarsˇek et al.,2012; Maas et al., 2011).Characterized by a wing-like parapodia, L. helicina is housed within an arag-onite shell and swims in a jerky spiral motion, hence its name of sea-butterflies(Morton, 1954; Lalli and Gilmer, 1989). As aragonite is a metastable form ofCaCO3, it is more susceptible to dissolution than calcite and at shallower depths102.1. INTRODUCTION(Mucci, 1981). With the increasing threat of ocean acidification, the associatedchallenges imposed on calcareous organisms (e.g. significant decreases in cal-cification rates, reduced shell mass and activity) can have detrimental effectson the growth, development, and reproduction of these organisms (Yamamoto-Kawai et al., 2009; Comeau et al., 2009). Investigations by Comeau et al. (2010)demonstrated that larval Cavolinia inflexa were deformed when exposed to lowerpH and despite showing active movement, these malformed individuals did notproduce shells, and exhibited reduced growth rates. Similar results were ob-tained in experiments on Arctic L. helicina by Lischka et al. (2010) and Lischkaand Riebesell (2012). The steady decrease of ocean pH may lead to the elimina-tion of L. helicina from the marine environment with ramifications impactingmarine food-web dynamics and marine biogeochemistry (Comeau et al., 2010).Throughout the past century there have been numerous studies investigatingvarious aspects of L. helicina, although its life cycle remains poorly understood(Lalli and Gilmer, 1989; Hunt et al., 2008; Bednarsˇek et al., 2012). To date,there is no agreement on questions regarding 1). the life cycle longevity, 2.)the seasonal development of the population size-structure, and 3.) the seasonaltimes of spawning. Using different methods in different ecological systems, lifecycle studies of L. helicina have yielded varying, sometimes drastically differentresults.2.1.2 Limacina helicina : Past Life Cycle InvestigationsCurrently there are four studies that have investigated the life cycle of L. helicina(Kobayashi, 1974; Fabry, 1989; Gannefors et al., 2005; Bednarsˇek et al., 2012).Appending to this list is a comprehensive literature review by Hunt et al. (2008)that compiled published and unpublished sources to assess knowledge gaps forL. helicina in the Southern Ocean. Supplementing this, are studies of Limacina112.1. INTRODUCTIONretroversa in the temperate waters of the Atlantic and Southern Ocean (Hsiao,1939a,b; Redfield, 1939; Dadon and de Cidre, 1992). These studies cover thereproductive biology of L. retroversa. Studying L. helicina in the Central ArcticOcean of the Canadian Basin, Kobayashi (1974) reported a life cycle of 1.5–2years with three cohorts and two generations spawned during one life cycle.Bednarsˇek et al. (2012) proposed the possibility of some L. helicina having alife cycle exceeding three years in the Scotia Sea of the Southern Ocean. A oneyear life cycle (also hypothesized for other Limacina) is stated in Hunt et al.(2008); Gannefors et al. (2005), and Fabry (1989).Kobayashi (1974) described L. helicina to display rapid growth throughoutthe fall/winter to reach sexual maturity. A major die-off occurs in the followingspring after the mature adults produced the spring cohort (Kobayashi, 1974).Reaching maturity by summer, the spring cohort spawned the summer cohort(Kobayashi, 1974). Along with spring and summer as the times of peak spawn-ing, Kobayashi (1974) also reported minor spawning during winter. In supportof summer spawning, Weslawski et al. (1991) reported high abundances of smallL. helicina during summer in the waters near Svalbard. Additional support ofsummer spawning is provided in Hunt et al. (2008), where a summer die-off oflarger adults is reported.Limacina helicina appears to show considerable size variation between re-gions, with individuals > 9 mm seen in the waters of the North Atlantic andthe Scotia Sea, Southern Ocean (Gannefors et al., 2005; Bednarsˇek et al., 2012).Meanwhile the largest individual observed in the Central Arctic Ocean mea-sures only 3.7 mm (Kobayashi, 1974). From Bednarsˇek et al. (2012) the smallerjuveniles constitutes a large proportion of the population, although the size fre-quency data in Gannefors et al. (2005) suggests that this is seasonally variable.122.1. INTRODUCTION2.1.3 Rivers Inlet: Historical ContextLocated on the central coast (51◦ 2410” N to 52◦ 5215” N) of British Columbia,Canada, and measuring approximately 40 km long by 3 km wide, Rivers Inletis an stratified estuarine fjord characterized by steep walls, a deep inner basin,and a deep sill at the mouth of the inlet. Situated near the north east upwellingarea, the regional oceanography is influenced by both the southerly flowingCalifornia current and the northward Arctic Current (Mackas et al., 2001).Located at comparatively lower latitudes, the plankton community experiencesa longer growing season and is subjected to reduced environmental extremes (e.g.warmer waters – compared to zooplankton in polar waters). Fed by numerousrivers throughout the inlet, discharge from Oweekeno Lake via the WannockRiver is recognized to be the primary source of fresh water. Average returns ofover 750,000 sockeye salmon over most of the twentieth century displayed highfluctuations in the 1970s, eventually dwindling to 3,500 in 1996 (McKinnellet al., 2001). Commercial fishing has been closed since then with a minorand sporadic recovery of the sockeye salmon. Various hypothesis have been putforward as to the cause of the precipitous decline of the sockeye salmon including1.) marine and/or freshwater related causes and 2.) the plankton food supply ofjuvenile sockeye salmon in their early marine phase. The Rivers Inlet EcosystemStudy (RIES) conducted from 2008 to 2010 was designed to investigate theplanktonic food webs of Rivers Inlet with an emphasis on plankton dynamicsassociated with the marine phase of juvenile sockeye salmon. Limacina helicinais a major dietary component of sockeye salmon in Rivers Inlet and the datasetscollected during RIES provided an unprecedented opportunity to study its lifecycle.132.1. INTRODUCTION2.1.4 Goals and AimsTo describe the life cycle dynamics of L. helicina, high quality datasets of itssize structure are required. Such a dataset should have high temporal resolutionand span an entire life cycle. Ideally such a dataset should span several years,especially if the system in question is prone to fluctuations in physical andbiological conditions (Fabry, 1989). Currently used in fisheries science, the sizefrequency method is well developed and consistent in detecting and documentingthe occurrence and temporal dynamics of population cohorts (MacDonald andPitcher, 1979). From the available literature there are numerous contradictionsconcerning 1. life cycle longevities, 2. the seasonal times of spawning and 3.the seasonal development of the population size structure of L. helicina. It ispossible that these inconsistencies have arisen from datasets of low temporalresolution collected in various geographical regions. It is our belief that thenon-sequential sampling confined to a limited seasonal range may particularlymask many intricacies of L. helicina life cycle dynamics. Through the use ofhigh temporal resolution data and the size frequency method, the aims of thischapter were three-fold.1. Describe the inter-annual progression and development of the populationsize structure of L. helicina based on observational data.2. Identify the seasonal date(s) of spawning.3. Quantitatively determine the life cycle longevity142.2. METHODS2.2 Methods2.2.1 Study Area – Rivers InletThe hydrodynamics of RI are driven by salinity stratification, hence the largeinfluence of riverine freshwater (Hodal, 2010). Tidal motion and wind-inducedmixing during strong storm events also impact RI hydrodynamics (Hodal, 2010;Wolfe, 2010).Eight stations were sampled as part of the RIES project, UBC 7, DFO 1,DFO 2, DFO 3, DFO 4, DFO 5, UBC 6, and UBC 8 (see Figure 1.6 in Hodal,2010, for the spatial proximities of each station). Sampled along a defined tran-sect, the location of each station was selected to emphasize major oceanographicfeatures within the inlet (Hodal, 2010). Located at roughly the halfway point,DFO 2 (Figure 2.2) was the key station of interest and was considered the mostrepresentative station with respect to the zooplankton community structure.Accordingly, DFO 2 was the most comprehensively sampled station during theRIES project and the focal station of this study. Situated outside the inlet,UBC 7 best represented the hydrographic conditions on the B.C. shelf, whileDFO 5, UBC 6 and UBC 8 was representative of the inner inlet.2.2.2 Seawater Temperature, Fluorescence, SalinityEnvironmental parameters at DFO 2 were collected as part of RIES during 2008,2009, and 2010. Fort-nightly data was collected during 2008 and 2009, whilemonthly data was collected for 2010. The CTD (SBE-25, Sea-Bird ElectronicsInc., with additional sensors measuring nutrients, oxygen, and fluorescence)was deployed to 300 m depth (maximum) or 10 m above the bottom for depths< 300 m. As RI features a shallow mixed layer, the measurements of sea watertemperature and salinity were averaged over the top 30 m, while fluorescence152.2. METHODSlevels were integrated over the same depth. Seawater temperature was measuredin ◦C, with salinity expressed in PSU. Raw fluorescence values were used as aproxy for chlorophyll-a, concentration.2.2.3 Sample Collection and SelectionZooplankton were sampled using a 0.50 m diameter bongo net (Aquatic Re-search Instruments) harnessed with General Oceanics flow meters to measurethe volume filtered. Two nets, each of a differing mesh size (150 µm and 250µm) were used simultaneously for sample collection. Only samples from the 150µm net were processed (Table 2.1). Exceptions were made for the 2010–2011winter-transition when both 150 µm and 250 µm net samples were processed.Vertical hauls from 300 m to surface were conducted during each survey,with samples preserved in a 4 % buffered sea water formaldehyde solution.Spatial AnalysisStations DFO 1–5 were compared during three time periods in 2010 (Figure 2.2;Table 2.2) to determine whether L. helicina sampled at DFO 2 were representa-tive of the population in RI. Zooplankton samples from September, 2010 werecollected during the 2010-15 LaPerouse Zooplankton Monitoring Cruise coor-dinated by the Institute of Ocean Sciences aboard the C.C.G.S. John P. Tully.UBC 7 was excluded from this analysis due to its locality being outside the inlet.Due to their close proximity to DFO 5, UBC 6 and UBC 8 were also excluded(see Figure 2.2b in Hodal, 2010).Zooplankton were sampled day and night in vertical tows from 300 m depthto surface (10 m from bottom if max depth was < 300 m) using 0.25 m2 (moutharea) bongo nets with a 235 µm mesh-size. These samples were collected forthe Department of Fisheries and Oceans Canada. A TSK flowmeter recorded162.2. METHODSthe volume filtered. Samples were preserved in a 4 % sea-water formaldehydesolution.2.2.4 Sample Preparation and EnumerationProcessing was undertaken using a Leica microscope equipped with an ocularruler. Ocular measurements were made to the nearest 0.1 µm from the tip ofthe aperture directly to the back of the shell (Figure 2.1).Figure 2.1: Photo of L. helicina with measurement of shell diameter (D)Using No. 5 INOX tweezers from Fine Scientific Instruments, visibly large in-dividuals (D ≥ 1.0 mm) were enumerated and measured from the entire sample.Visual inspection and removal of large individuals was necessary as the potentialof excluding large individuals increases with a larger number of sample-splits.After inspection and the removal of large individuals, the sample was processedin its entirety or, when L. helicina were very abundant, split (1/2 to 1/64) toproduce a manageable sub-sample. Whenever possible, a minimum of 128 indi-172.2. METHODSviduals were measured, however, more or less were measured depending on theseason. This was an attempt to ensure the data generated were representative.2.2.5 Size Frequency Histograms and Identification of Co-hortsSize frequency histograms were constructed for each survey (Table 2.1). Dis-crete modal peaks seen in these histograms were deemed to indicate differentpopulation cohorts (e.g. age-classes) or distinct spawning events (MacDonaldand Pitcher, 1979; Bednarsˇek et al., 2012). The ability to visually distinguisheach cohort was of great importance when attempting to document the seasonaldevelopment of identified cohorts. For this reason, the size frequency histogramswere not normalized. Due to the likelihood of strong overlap between cohorts,finite mixture distributions were fitted to each size frequency histogram. Func-tions from the R package mixdist were used to implement the fitting (MacDon-ald and contributions from Juan Du, 2011). By using an amalgamation of theEM and a Newton-type algorithm, the mixdist package computed the best fitto the observed data based on user-input of parameters (e.g. number of cohorts,mean size & standard deviation of each cohort, proportional abundance of eachcohort, estimate of shape for the general distribution – Normal, Lognormal,Gamma, Weibull, Binomial, Exponential, and Poisson). Given the possibilityof over parameterization (see Appendix B.5 for details concerning fitting finitemixture distributions), constraints were placed to reduce the number of param-eters estimated (e.g. fitting the distributions were aided by the placement ofconstraints, not forced)Cohorts can then be tracked from their time of recruitment into the popu-lation, to their likely times of die off (e.g. when they are no longer observed inthe size frequency histograms).182.2. METHODS2.2.6 Estimation of Growth and Life-Cycle LongevityThe seasonal growth of cohorts tracked was estimated by measuring the differ-ence in the modal size (mean size) of the cohort(s) tracked between two datesof observation, then averaging that difference by the elapsed time between thetwo dates (in days)Gc =mj −mitd(2.1)td = tj − ti (2.2)LC =Datelast −Datefirst365(2.3)In Eq. 2.1, “m” is the modal size (in mm) of the cohort tracked and i andj are the observational dates at t = i, j, respectively, with td (Eq. 2.2) beingthe time difference in days, between ti and tj . Determination of the life-cyclelongevity involved the tracking of cohorts from their dates of recruitment untiltheir date of assumed die-off. The date of assumed die-off was deemed to bethe date when the cohort was no longer detected in the corresponding size-frequency histogram. Life-cycle longevity (“LC”) was calculated using Eq. 2.3and expressed in years from the initial date of recruitment (Datefirst) to thedate of assumed die-off (Datelast). Expressed in days, the difference betweenDatefirst and Datelast was averaged over the day duration of a year to estimatethe life cycle longevity (in years) of L. helicina.192.2. METHODSFigure 2.2: Lambert conformal conic projection of the British Columbia westcoast and the northern tip of Washington State. Rivers Inlet is depicted inthe zoomed inset of the mainland coast of BC. A NASA blue marble image isunderlain as a general background of the region.202.2. METHODSTable 2.1: Summary statistics of L. helicina size structure orderedby year and survey date. The sample from survey 9 of 2008 (col-lected on 22 July) was missing. The 2010W in the Year columnindicates samples collected from the 2010–2011 winter transition,from October 25, 2010 to March 19, 2011. Note that one samplecollected on November 22, 2010 was a combination of zooplanktonsampled from both the 150 µm and 250 µm nets.Year Date Survey Mean(mm)Min(mm)Max(mm)Mesh(µm)2008 18 March 1 1.55 0.17 3.00 1502008 31 March 2 1.75 0.17 3.00 1502008 22 April 3 2.86 0.21 4.11 1502008 9 May 4 0.87 0.13 5.50 1502008 25 May 5 1.39 0.18 4.83 1502008 8 June 6 0.51 0.12 3.45 1502008 20 June 7 0.80 0.16 3.78 1502008 8 July 8 0.98 0.16 3.35 1502008 22 July 9 NA NA NA 1502008 4 August 10 1.10 0.14 3.78 1502008 22 Septem-ber11 0.51 0.18 2.65 1502009 28 February 1 0.68 0.18 2.64 1502009 17 March 2 0.99 0.22 2.43 150Table 2.1 – continued on next page212.2. METHODSTable 2.1 – continued from previous pageYear Date Survey Mean(mm)Min(mm)Max(mm)Mesh(µm)2009 1 April 3 0.69 0.18 3.58 1502009 15 April 4 0.49 0.12 3.63 1502009 3 May 5 0.41 0.18 3.91 1502009 20 May 6 0.57 0.13 4.42 1502009 2 June 7 0.75 0.18 4.42 1502009 18 June 8 0.64 0.17 2.82 1502009 1 July 9 0.72 0.13 4.65 1502009 17 July 10 0.66 0.13 4.42 1502009 13 August 11 0.55 0.17 3.39 1502010 19 March 1 0.62 0.21 2.74 1502010 23 April 2 0.37 0.15 3.61 1502010 17 May 3 1.04 0.16 5.30 1502010 21 June 4 1.65 0.16 4.06 1502010 20 July 5 1.11 0.16 3.56 1502010W 25 October 1 0.72 0.20 2.03 1502010W 8 November 2 0.63 0.29 2.43 1502010W 8 November 3 0.63 0.39 2.11 2502010W 22 November 4 0.64 0.44 2.36 2502010W 22 November 5 0.69 0.19 2.54 150 &2502010W 18 January 6 0.73 0.23 1.84 1502010W 8 February 7 0.84 0.51 2.10 150Table 2.1 – continued on next page222.2. METHODSTable 2.1 – continued from previous pageYear Date Survey Mean(mm)Min(mm)Max(mm)Mesh(µm)2010W 19 March 8 0.50 0.21 0.56 150Table 2.2: Dates of sample collection for DFO 1, DFO 2, DFO 3,DFO 4, DFO 5, for three time periods of 2010. Samples collectedwere used for analysis of the spatial variability in the populationsize structure of L. helicina. Samples for the early spring and earlysummer were collected using a 150 µm mesh net, while the Fallsamples were collected using a 235 µm mesh net.Season Station DateEarly Spring DFO 1 18 MarchEarly Spring DFO 2 19 MarchEarly Spring DFO 3 19 MarchEarly Spring DFO 4 19 MarchEarly Spring DFO 5 19 MarchEarly Summer DFO 1 17 MayEarly Summer DFO 2 17 MayEarly Summer DFO 3 18 MayEarly Summer DFO 4 18 MayEarly Summer DFO 5 18 MayFall DFO 1 11 SeptemberFall DFO 2 11 SeptemberFall DFO 3 11 SeptemberFall DFO 4 11 SeptemberTable 2.2 – continued on next page232.2. METHODSTable 2.2 – continued from previous pageSeason Station DateFall DFO 5 11 September242.3. RESULTS2.3 Results2.3.1 Environmental Parameters - Temperature, Salinity,FluorescenceTemperature and salinity displayed opposite seasonal trends from spring to sum-mer, during all years (Figure 2.3A). Temperature increased linearly from springto summer while salinity showed a seasonal decrease. Exceptions were notedduring July, 2008 when salinity displayed an increasing trend into September(Figure 2.3A). Although fluorescence values showed high variation throughoutspring and summer of 2008, they were consistently higher during the summer,in 2009. Fluorescence in 2010 showed a decreasing trend from spring to sum-mer (Figure 2.3B). Summer values were also consistently higher compared tothe fall/winter period. The lower resolution sampling in 2010 may have maskedpotential variations between surveys, for temperature, salinity, and fluorescence.2.3.2 L. helicina Abundance: Seasonal and Inter-annualVariationSeasonal variations in abundance differed between years, although a generalincrease was observed from spring to summer (Figure 2.3B). Exceptions wereseen in 2008 with peak abundances (≥ 628 ind.m−3 max) occurring in late-spring (May 9). The 2009 and 2010 seasons exhibited peak abundances duringthe summer (Figure 2.3B). Highest abundances in 2009 and 2010 reaching > 100ind.m−3 were documented in August and late June, respectively (Figure 2.3B).Between survey variation in abundance was greatest in 2008 and lowest in 2009(Figure 2.3B). Low sampling frequency in 2010 could be responsible for the lowsurvey-to-survey variation.252.3. RESULTSEnvironmental CorrelationsSeasonal abundances for each year were log-transformed and regressed againstthe environmental parameters. See the Tables in Appendix A.2 for the fullstatistical results. Limacina helicina abundance was significantly correlatedwith 30 m depth-averaged temperature in 2009 and 2010 (2009: R2 = 0.843,p < 0.01, 2010: R2 = 0.961, p < 0.05), and salinity in 2009 (2009: R2 = 0.721,p < 0.01). There were no significant correlations found between abundance anddepth-integrated fluorescence, for any year.2.3.3 L. helicina Size Structure: Seasonal DevelopmentThe size-frequency histograms for 2008 and 2009 (Figure 2.4) showed seasonalgrowth to be continuous, although there appeared to be two major periods ofgrowth during each year, in spring and summer. There was notable size devel-opment of the larger individuals during each major period of growth. Springgrowth occurred from late February–mid March until late May–early June whilesummer growth was seen from June to late July–August, for both 2008 and 2009seasons. Both periods showed steady growth of L. helicina to larger sizes, al-though the relative abundance of the larger individuals declined throughouteach period of growth (Figure 2.4).With the exception of surveys 1–3 from 2008 (Table 2.1; Figure 2.4), everysurvey throughout 2008 and 2009 displayed a prominent abundance of the 0.2–0.4 mm size-group. Shifting to the 0.40–0.60 mm size-group during the 2010-2011 winter-transition, a notable presence of the 0.2–0.4 mm size-group was notseen until the following spring (Figure A.1).Despite fewer sampling through spring and summer, 2010, similar patterns ofgrowth can be seen when compared to the size-frequency histograms from 2008and 2009 (Figure 2.4). The static configuration of the size-frequency histograms262.3. RESULTSthroughout the 2010-2011 winter transition indicated the cessation of growth inwinter.2.3.4 Spawning ActivitySpawning activity was identified by 1. a prominent abundance of the size-group ≤ 0.4 mm (known as immatures in catches) and 2. the relative abundanceof the 0.2–0.4 mm size-group in the size-frequency histograms for each season(Figure 2.4).Immatures consistently accounted for > 50 % (with maximum above 70 %)of the total abundance throughout May in 2008 and 2009 (Figure 2.4). In addi-tion to immatures, prominence of the 0.2–0.4 mm size-group (relative to othersize-groups) from May to August in 2008 and 2009, suggested continuous (e.g.protracted) spawning activity (Figure 2.4). Although the relative abundance ofthe 0.2–0.4 mm size-group suggested protracted spawning (throughout 2008 and2009), the data indicated one period of peak spawning activity. The parallelpeaks of total population abundance combined with the highest relative abun-dance of immatures, in 2008 (May 9), suggested that late spring was a time ofpeak spawning. In contrast, the peak population abundance was not seen untillate summer (August 13) in 2009. Even so, there were noticeable differencesin peak abundances between 2009 (∼ 153 ind.m−3) and 2008 (> 600 ind.m−3).Based on the high contribution of immatures to the total abundance in latespring (late April–May) (Figure 2.5), it is possible the spring peak spawningwas missed in 2009. Due to lower level sampling of 2010, direct comparisonsto 2008 and 2009 cannot be made, although the population abundance was ob-served to increase from mid May to peak in late June. Based on the data, it isappeared protracted spawning occurred throughout the 2008 and 2009 seasons,however, it was evident that the late spring was a time of peak spawning (e.g.272.3. RESULTSfollowed by limited protracted spawning).We found no indication of spawning activity during the 2010–2011 wintertransition. Spawning likely began again in early-spring, 2011, with the re-establishment of the 0.2–0.4 mm size-fraction in the size frequency histogram inlate-March (Figure A.1).2.3.5 Estimate of Life-Cycle LongevityUsing finite mixture distributions, two cohorts were identified and tracked fromMay 9, 2008 to July 20, 2010 to determine the life cycle longevity of L. helicina.Cohort C1 was identified on 9 May, 2008 and tracked until its assumed to vanishon 4 August, 2009. Cohort C2 was identified on 3 May, 2009 and tracked untilit was assumed to die-off on 20 July, 2010. Figure 2.7B depicts the growth cyclefor each cohort and Table 2.3 lists the ∼ biweekly development of each cohort.C1 and C2 exhibited slow initial growth (< 0.015 mm.day−1) during thefirst two weeks, after which C2 grew at a rate ∼ 3 times faster than that ofC1. By 28–30 days into the respective growth-cycles, C2 was roughly twicethe size of C1. It was not until early July when both cohorts were of similarsize (Table 2.3). The growth-trajectories of C1 and C2 differed greatly afterearly–mid July, of 2008 and 2009, respectively. Peaking at 2.2 mm by earlyAugust 2008, C1 exhibited a recession in shell-size by late-September prior tothe fall–winter period. Meanwhile, during 2009 C2 grew unabated and reached amodal size of 3.3 mm prior to entering the fall–winter period. Both cohorts werenotably smaller in the following spring when compared to their respective sizes,from the previous summer. From the size-frequency histograms during the 2010–2011 winter-transition (Figure 2.4), it was likely the 2008–2009 and 2009–2010winter-transitions also exhibited zero growth, of L. helicina. Rapid growth wasobserved the following spring when both cohorts likely reached sexual maturity282.3. RESULTSafter a modal size of 3 mm was surpassed. The growth of C1 appeared to leveloff from mid June–mid July, 2009, while the growth of C2 showed no signs oftapering. From the data, C2 was likely spawned by C1 in late spring of 2009(Figure 2.7B). Using Equation 2.3, C1 was tracked for 461 days and C2 for 443days. This equated to life cycle longevities of 1.26 and 1.21 years, respectively.A third cohort, C3, was identified on April 23, 2010, and was likely spawnedby C2 (orange line in Figure 2.7B). Tracked until July 20, 2010, the growth tra-jectory of C3 resembled that of C1 and C2 post-recruitment, strengthening thegrowth cycles of C1 and C2 (Figure 2.7B). C3 was not considered for estimatesof life-cycle longevity as it was tracked for a limited time.Table 2.3: The growth and development of C1 and C2 throughouttheir tracked life cycles. Growth rates are calculated using Eq. (1)and expressed in mm.day−1. Note that the sample for 22 July,2008 was not available, hence the NA under the Modal Peak andthe Growth columns. C1 was not observed in the size frequencyhistogram of 13 August, 2009 (Figure 2.4)) and assumed to havedied off. C2 was not seen in the size-frequency histograms of both21 June and 20 July, 2010 and assumed to have died off on 20 July,2010. (Figure 2.4)Year Date Day Cohort Modal Peak(mm)Growth(mm.day−1)2008 09-May 1 C1 0.229 NA2008 25-May 14 C1 0.300 0.0042008 08-Jun 28 C1 0.500 0.0142008 22-Jun 42 C1 0.945 0.031Table 2.3– continued on next page292.3. RESULTSTable 2.3– continued from previous pageYear Date Day Cohort Modal Peak(mm)Growth(mm.day−1)2008 08-Jul 58 C1 1.900 0.0592008 22-Jul 72 C1 NA NA2008 04-Aug 85 C1 2.200 0.0112008 22-Sep 134 C1 1.814 NA2009 28-Feb 293 C1 1.311 NA2009 17-Mar 310 C1 1.993 0.0402009 01-Apr 325 C1 2.860 0.0572009 15-Apr 339 C1 3.410 0.0392009 03-May 357 C1 3.690 0.0152009 20-May 374 C1 NA NA2009 02-Jun 387 C1 3.900 0.0072009 18-Jun 403 C1 NA NA2009 01-Jul 416 C1 4.580 0.0232009 17-Jul 432 C1 4.400 NA2009 13-Aug 461 C1 NA NA2009 03-May 1 C2 0.200 NA2009 20-May 17 C2 0.361 0.0092009 02-Jun 30 C2 0.970 0.0472009 18-Jun 46 C2 1.530 0.0352009 01-Jul 59 C2 2.050 0.0402009 17-Jul 75 C2 3.040 0.0622009 13-Aug 102 C2 3.300 0.0092010 19-Mar 320 C2 2.680 NATable 2.3– continued on next page302.3. RESULTSTable 2.3– continued from previous pageYear Date Day Cohort Modal Peak(mm)Growth(mm.day−1)2010 23-Apr 355 C2 3.130 0.0132010 17-May 379 C2 4.180 0.0442010 21-Jun 414 C2 NA NA2010 20-Jul 443 C2 NA NA2.3.6 Rivers Inlet Spatial AnalysisThe smaller size fractions displayed a dominant influence throughout the sea-sons at each station, while the larger individuals were sparsely distributed (Fig-ure 2.6). During early spring individuals < 1.5 mm in size predominated withonly a few larger individuals of > 3.0 mm also present. The size structure dis-played bi-modality in early summer and fall with possible exceptions at DFO 1and DFO 5 (Figure 2.6). For all stations, the principal modes discerned in earlyspring ranged from the 0.2–0.8 mm and 1–2 mm, and while the 0.2–0.8 mmmode was static in fall, the 1–2 mm mode present in summer appeared to shiftto the 1.6–2.4 mm range in fall (Figure 2.6). Although there was no distinctmode seen for sizes above 2.0 mm, large specimens were recorded throughout forall seasons and stations in low numbers. From the data presented, we hypothe-size the population size structure of L. helicina was broadly similar across theinlet, with spatial differences due to the patchiness of L. helicina and advectiveinfluences.312.3. RESULTSFigure 2.3: A: Seasonal and interannual variation of 30 m depth-averagedtemperature - red and salinity - yellow. Temperature is expressed in ◦C andsalinity is expressed as PSU (Practical Salinity Units). B: Seasonal and inter-annual variation in L. helicina density (sampled from the top 300 m) and 30m depth-integrated fluorescence - green. Density is expressed in ind.m−3.Note the X-axis is expressed in Gregorian dates on top and Julian Day at thebottom. January 1 is indicated by the dashed red lines (starting with January1, 2009).322.3.RESULTS0 1 2 3 4 5 60.00.20.40.6Shell Diameter (mm)Probability DensityMar 18, 2008Lognormal Mixture0 1 2 3 4 5 60.00.20.40.60.81.0Shell Diameter (mm)Probability DensityMar 31, 2008Normal Mixture0 1 2 3 4 5 60.00.20.40.60.8Shell Diameter (mm)Probability DensityApr 22, 2008Gamma Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityMay 09, 2008Lognormal Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityMay 25, 2008Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.02.5Shell Diameter (mm)Probability DensityJun 08, 2008Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.81.01.2Shell Diameter (mm)Probability DensityJun 22, 2008Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.81.01.2Shell Diameter (mm)Probability DensityJul 08, 2008Gamma Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityAug 04, 2008Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensitySep 22, 2008Lognormal Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensityFeb 28, 2009Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.81.01.21.4Shell Diameter (mm)Probability DensityMar 17, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensityApr 01, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.02.5Shell Diameter (mm)Probability DensityApr 15, 2009Lognormal Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensityMay 03, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityMay 20, 2009Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.81.0Shell Diameter (mm)Probability DensityJun 02, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensityJun 18, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityJul 01, 2009Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.81.01.21.4Shell Diameter (mm)Probability DensityJul 17, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.0Shell Diameter (mm)Probability DensityAug 14, 2009Gamma Mixture0 1 2 3 4 5 60.00.51.01.5Shell Diameter (mm)Probability DensityMar 19, 2010Gamma Mixture0 1 2 3 4 5 60.00.51.01.52.02.5Shell Diameter (mm)Probability DensityApr 23, 2010Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.8Shell Diameter (mm)Probability DensityMay 17, 2010Gamma Mixture0 1 2 3 4 5 60.00.20.40.60.8Shell Diameter (mm)Probability DensityJun 21, 2010Gamma Mixture0 1 2 3 4 5 60.00.20.40.6Shell Diameter (mm)Probability DensityJul 20, 2010Gamma MixtureFigure 2.4: Finite mixture distributions were fitted to distinguish between different population cohorts at different times. The following subplotsshow the finite mixtures fitted to the size-frequency data for all fort-nightly samples collected from March 18, 2008 to August 14, 2009, and themonthly samples from March 19 to July 20, 2010. Cohort C1 was likely recruited into the population on 9 May, 2008. C2 was likely recruited on 3May, 2009 (near bottom left subplot). It was probable that C1 died-off on 13 August, 2009, while C2 likely died off on 20 July, 2010.332.3. RESULTSFigure 2.5: Seasonal population abundance of L. helicina for A. 2008, B.2009, C. 2010, D. 2010-2011 winter-transition. D. The inter-annual popula-tion abundance of Figure 2.3B is broken into four different panels, one for eachyear of observation. Also included is the proportional abundance of the size-fraction < 0.40 mm (dashed red line). Note the different scaling of the Y-axisbetween panels. The light grey and blue dots, respectively indicate sam-pling on 8, 22 November with 250 µm and 150 & 250 µm mesh nets, respectively.342.3. RESULTSFigure 2.6: Histograms (0.2 mm bins) comparing the population size structureof L. helicina from stations DFO 1–5 during three seasonal time points in 2010,spring (19 March), summer (17–18 May), fall (11 September). Each column inthe figure denotes the respective seasons while the stations are designated byrow, with DFO-1 in row 1, and DFO-5 in row 5. All samples were collected with150 µm mesh bongo nets with the exception of those during fall. Fall sampleswere collected with 236 µm mesh bongo nets. Note the different scaling in they-axis for each histogram.352.3. RESULTSFigure 2.7: A: Seasonal and interannual distribution of the mean shell size ofthe population. B: Growth cycles of the two cohorts (C1, C2) tracked. Theorange line denotes the third cohort (C3) identified, and tracked for a limitedtime. January 1 is indicated by the dashed red lines (starting with January 1,2009).362.4. DISCUSSION: LIFE-CYCLE RE-EVALUATION2.4 Discussion: Life-cycle re-evaluation2.4.1 Seasonal Spawning and RecruitmentA strong influence of the smaller size-group (0.2–0.4 mm) from spring to fallcombined with the high proportions of immatures indicates protracted spawningby L. helicina with a clear peak in spawning activity during the late spring.It appears that peak spawning is initiated in late spring and continues intothe summer. The low contribution of the 0.2–0.4 mm size-group to the totalabundance during the 2010–2011 winter-transition indicated the termination ofspawning during the late fall–winter period.Paranjape (1968) supports continuous spawning and interprets reproduc-tive activity to extend over prolonged periods. Interpretations from Kobayashi(1974) describes reproductive activity to commence during spring and continueinto the summer. Additionally, Dadon and de Cidre (1992) reports of springand summer spawning events for L. retroversa of the Southern Argentine Sea.Gannefors et al. (2005) found adult L. helicina in Kongsfjorden to be atdiffering stages of egg production during summer spawning. The zooplanktonof Kongsfjorden are under strong influence from local hydrography, which canvary substantially between years (Hodal et al., 2012). Limacina helicina inKongsfjorden are of Arctic origin and are most prevalent when Arctic watersare advected into the fjord (Willis et al., 2006). Thus, the interpretation ofonly summer spawning in Gannefors et al. (2005) may have been due to lowervolumes of Arctic waters advected into the fjord, during spring. Alternatively,the observation of summer spawning in Gannefors et al. (2005) suggests thepossibility of a late spring bloom delaying the spring peak spawning to summer.372.4. DISCUSSION: LIFE-CYCLE RE-EVALUATIONSize of SpawningSpawning size differed between C1 and C2, although both cohorts were largerthan 3 mm during the time of the peak spawning. This differed considerablyfrom the size of mature adults reported in Bednarsˇek et al. (2012) and Ganneforset al. (2005). Compared to zooplankton from temperate latitudes, zooplanktonsampled from polar waters (e.g. Scotia Sea and Kongfjorden) are characterizedby slow growth, delayed sexual maturity, and a larger size at sexual maturity(Clarke and Peck, 1991). The differences in spawning size between polar popu-lations of L. helicina and this study could therefore be explained by latitudinal(sensu temperature and habitat) differences. Additional support for a smallerspawning size is provided by the gonadal analysis in Kobayashi (1974), whereova of differing sizes were observed in L. helicina at different stages of sexualmaturity (Table 2 in Kobayashi, 1974). It is therefore likely that L. helicina isa pulsing spawner capable of releasing batches of matured ova, while developingova continue its growth (Dadon and de Cidre, 1992). According to Table 2 inKobayashi (1974), hermaphroditic organs are present in individuals reaching 0.7mm in size, with little difference in increasingly larger individuals. It is probablethat after spawning commences in spring when “some” of the fully matured ovaare released, it continues as the remaining portions become mature, leading toprotracted spawning.2.4.2 Seasonal Growth and Environmental CorrelationsThe population size structures showed the continual growth of L. helicina through-out the 2008 and 2009 seasons, although there were two periods of enhancedgrowth (for larger individuals in particular). Larger individuals exhibited a no-table increase in size from early–late spring, which we interpret as the parentalcohort growing to sexual maturity. This was demonstrated by the rapid growth382.4. DISCUSSION: LIFE-CYCLE RE-EVALUATIONof C1 and C2 in late spring of 2008 and 2009, respectively, culminating in thespawning peak. Of interest was the rapid development of C2 during summerof 2009. Unlike C1, C2 was able to grow to ≥ 3 mm in size before overwinter-ing. As it was likely that L. helicina in RI are able to spawn from a shell sizeof ≥ 3 mm, the rapid growth of C2 from May–August of 2009 suggested thepossibility of peak spawning in summer. The growth cycle of C1 appeared totaper by mid-July, 2009, indicating C1 was likely approaching senescence. Incontrast, the growth of C2 did not taper. Although we were unable to trackC2 for the remainder of the 2010 season, the growth trajectory of C2 suggestedthe life cycle of C2 to be potentially longer compared to C1. Both C1 and C2over-wintered and matured sexually in spring, of the following year.The dependence of L. helicina on food quantity and most likely, quality iswell documented (Lalli and Gilmer, 1989; Gilmer and Harbison, 1991; Huntet al., 2008; Seibel et al., 2012). Although linear regressions indicated an non-significant relationship between seasonal abundance and 30 m depth-integratedfluorescence (for all years), it is possible that the data used here was too coarseto resolve the connection between these parameters (e.g. Limacina responses totemporal changes in chlorophyll may be occurring at finer time scales).We found no evidence of winter growth, although Kobayashi (1974) arguedwinter was a time of a notable growth in the Arctic. A cessation of growth inwinter was supported by Lischka and Riebesell (2012). Our results also clearlyindicated a decrease in shell size prior to over-wintering, suggesting higher mor-tality of the larger demographic during the fall–winter. It appears that L. helic-ina halts its growth during the winter till resources are sufficient, and conditionsbecome optimal in early spring the following year.392.4. DISCUSSION: LIFE-CYCLE RE-EVALUATION2.4.3 Rivers Inlet Spatial AnalysisBroadly, L. helicina displayed a similar size structure across the inlet. Thespatial differences in size structure that were observed were likely due to theinherent patchiness of L. helicina. However, advective influences may also haveplayed a role. The fresh water run-off from the Wannock River may have carriedaway certain size-groups at the DFO 5 towards the inlet mouth. This mayaccount for a general lack of individuals > 2 mm in size at DFO 5. Incidentally,DFO 4 displayed presence of individuals ranging from 1.4–4 mm in size acrossseasons. Additionally, spatial differences in chlorophyll concentrations and thetiming of the spring bloom may have influenced the size-structure of L. helicinaacross the inlet.2.4.4 Potential Sampling ErrorsDespite bi-weekly sampling, there is the possibility that some short term vari-ability was not identified. This is especially true when attempting to track thegrowth and size progression of the identified cohorts. Numerous surveys foreach year showed no apparent change in the overall shape of the populationsize distribution suggesting that the population of L. helicina is stable betweensurveys. However, greater seasonal changes may have been observed if zooplank-ton samples were collected more frequently (e.g. using data of higher temporalresolution).In this chapter, we have shown that L. helicina displays peak spawning activ-ity during the spring, when the population reaches sexual maturity. Spawningwas observed at a size of 3 mm and larger. Protracted spawning commences af-ter spring-spawning and continues into and throughout the summer. We foundno evidence of spawning during the late fall–winter period. Continuous size-development was seen in the size-frequency histograms throughout the seasons402.4. DISCUSSION: LIFE-CYCLE RE-EVALUATION(for all years), although the relative abundance of the larger demographic de-creased with increases in size. By tracking two cohorts for over 400 days, wewere able to estimate a life-cycle longevity of 1.2 years, although it appears pos-sible for L. helicina to live up to 1.5 years. From the data, it seems the life cycleof L. helicina in the temperate North East Pacific resembles that of Limacinaretroversa in the Southern Argentine Sea (Dadon and de Cidre, 1992). This issubstantially shorter than the 3 year life history proposed by (Bednarsˇek et al.,2012). It is possible that the differences between the L. helicina life historiesobserved in our study and Bednarsˇek et al. (2012) can be explained by geneticdifferences (Hunt et al., 2010). This chapter has also shown the importance oftemporal resolution in resolving the seasonal and inter-annual dynamics of L.helicina in Rivers Inlet. Even though an non-significant relationship was foundbetween seasonal abundance and fluorescence, perhaps higher resolution datain addition to a lag analysis is needed for clear relations to be seen, especially ifslight changes in environmental conditions (e.g. chlorophyll, temperature, andsalinity) can stimulate and drive population responses (e.g. increasing popula-tion abundance, growth/maturation, and/or spawning). Based on the resultsof Chapter 2, it appears that higher resolution zooplankton sampling (e.g. col-lected daily) is needed in order to determine the factors driving the seasonaldynamics observed, for L. helicina in Rivers Inlet.41Chapter 3Seasonal Growth andMortality – Spring vs.Summer3.1 IntroductionDespite narrowing the knowledge gaps regarding the life cycle of L. helicina, theresults from Chapter 2 showed that life-cycle dynamics can potentially vary atfiner time scales. Specifically, difficulties were encountered in resolving seasonalspawning (e.g. possibility of more than one period of peak spawning) events inChapter 2. Clearly, higher resolution data should be used to fully resolve theL. helicina life cycle in RI. Use of higher resolution data may also provide anopportunity to accurately estimate mortality rates which are fundamental forthe construction of predictive models for the L. helicina population dynamics.Growth rates of three species of Limacinidae and one Cavoliniidae fromBarbados were determined by Wells (1976) from census data. An averagegrowth rate of 0.12 mm.month−1 was reported with no significant differencesbetween small and large Limacina (Wells, 1976). Nevertheless, it was docu-mented that the growth rates of Limacina species diminished once the adult423.1. INTRODUCTIONstage was reached (Wells, 1976). Other studies have also investigated Limacinagrowth rates, but without common units (e.g. grams of carbon), the studies aregenerally not comparable (Gannefors et al., 2005; Bednarsˇek et al., 2012).In order to construct robust models predicting zooplankton abundance, ac-curate estimates of both seasonal growth and mortality are needed (Ohman andWood, 1995; Gentleman et al., 2012). Currently, the literature abounds withstudies investigating zooplankton growth rates (e.g. production) with little at-tention to mortality estimates (Ohman and Wood, 1995). Moreover, the largenumber of non-standardized methods within the literature ultimately leads toincomparable results (Aksnes et al., 1997). While the large majority of growth-rate experiments can be performed on-ship and/or nearby on-shore laborato-ries to produce fairly robust estimates, lab-based mortality experiments are notrepresentative of natural populations as many factors influencing mortality (e.g.parasitism, predation, advection) cannot be incorporated into the experiment(Harris, 2000). This clearly called for more research focusing on mortality ratemeasurements on natural populations using high resolution time series data(Aksnes et al., 1997). Bednarsˇek et al. (2012) quantified mortality rates for L.helicina in the Southern Ocean, however, the large spatial extent of the studywas restricted to summer months only (Wiens, 1989).As part of the RIES (Rivers Inlet Ecosystem Study) sampling protocol(www.riversinlet.eos.ubc.ca), a Daily Station was established to investigate thetiming of the spring bloom in RI and its drivers. This provided an unprece-dented opportunity to capture the seasonal dynamics and assess accurately themortality rates of L. helicina, using daily resolved observations for over 100days.Two methods, horizontal and vertical, of estimating mortality rates havebeen implemented in many studies throughout the past decades (Ohman and433.2. GOALSWood, 1995; Gentleman et al., 2012). Horizontal methods (aka cohort methods)can be applied to a sampling time series to monitor temporal changes experi-enced by individual cohorts (Aksnes and Ohman, 1996; Aksnes et al., 1997;Gentleman et al., 2012). Based on recruitment, stage-duration, and mortality,the vertical method approximates the number of individuals observed for a de-velopment stage at a point in time (Aksnes et al., 1997). Although time seriesdata are not required, the vertical method is tied to more restrictive assump-tions than the horizontal method (Gentleman et al., 2012). Each method isbound by their respective list of assumptions which if not met, can yield poten-tially biased estimates (Gentleman et al., 2012). For example, an assumptionof a negligible advection may introduce a substantial bias, impacting horizon-tal methods (Aksnes and Ohman, 1996; Aksnes et al., 1997; Gentleman et al.,2012). If advection is a factor influencing the seasonal distribution of zooplank-ton, the real temporal changes in a population may be masked by the advection(immigration vs. emigration) of potentially different sub-populations (Aksnesand Ohman, 1996). The use of coarse time resolution data may also significantlyaffect the outcome of horizontal methods (Gentleman et al., 2012). If however,the time series is of sufficient resolution, then the horizontal method may yieldvital information if the population is homogeneously distributed in the systemconsidered. From the spatial analysis of Chapter 2 (Section 2.3.6), it wouldappear that L. helicina exhibits a broadly similar size-distribution within RI,with differences in density depending on location.3.2 GoalsDespite criticisms of the horizontal method, there have been only few studies inthe literature to utilize a dataset of > 100 days (Gentleman et al., 2012). Webelieve the horizontal estimation method combined with daily data for more443.2. GOALSthan 100 days will offset much of the current criticisms, as well as providerobust and accurate estimates of seasonal growth and mortality.The aims of this chapter are as follows:1. To provide a high resolution analysis of the seasonal cycle of recruitment.2. To document the seasonal development of L. helicina by providing dailyrates of growth in shell-size in spring and summer.3. To provide a first-approximation of the seasonal mortality of L. helicinaof the coastal North East Pacific, by estimating instantaneous mortalityas well as documenting its seasonal changes, from spring to summer.453.3. METHODS3.3 Methods3.3.1 Study Area & Sample CollectionEstablished in the proximity of Dawsons Landing (51◦57’40” N, 127◦58’60” W),the nearby dock was the ideal location for the sampling of zooplankton and otherphysical parameters (fluorescence). The hydrodynamics of RI and other physicalfactors (wind) are detailed in Hodal (2010) and Wolfe (2010), respectively.Zooplankton samples were collected each day after dusk from March 22 toJuly 7, 2010, using a ring net (0.30 m diameter, 80 µm mesh-size) supplied byAquatic Instrument Supply Company, from the deepest section (25–30 m depth)of the dock. Sample handling followed the same protocol outlined in Chapter 2(Section 2.2.3). Limacina helicina individuals were measured and enumeratedin the entire sample under a Leica microscope equipped with an ocular ruler.Measurements were made to the nearest 0.01 µm from the tip of the shell aper-ture directly to the back of the shell (Figure 2.1). Every confirmed individualwas considered for size-frequency enumeration unless there was sufficient phys-ical damage preventing correct measurements. In these cases, the individualswere only considered for abundance estimates.3.3.2 Daily FluorescenceAn autonomous ECO-Fl fluorometer (WetLabs) was moored on the dock atDawsons Landing to measure daily fluorescence. Fluorescence measurementswere collected from 5 m depth at 2 hour intervals. Daily average values wereconverted to chlorophyll using Equations 3.1 and 3.2. Data were available fromJanuary 1 to July 20, 2010 with a gap of missing data from February 13 toMarch 31, hence only data from April 1 to July 7, 2010 was used for analysis.463.3. METHODSChl = SF ∗ (Op−DC) (3.1)SF =xOp−DC(3.2)Op in Equation 3.1 indicates the signal output from the fluorometer and x inEquation 3.2 is the concentration of the solution used during instrument stan-dardization. SF (Equation 3.2) is the scale factor and DC is the Dark Counts(signal output of the fluorometer in clean water with black tape over detector).A DC of 105 was used, with an SF of 0.0077 µg.l−1.count−1. Chlorophyll wasexpressed in µg.chl-a.l−1.3.3.3 Size-Frequency Histograms & Identification of Co-hortsMeasurements from each sample (Table B.1 in Appendix B) were binned into0.02 mm size-bins to distinguish the smallest changes in the population sizestructure on a daily basis (see the size frequency histograms in Appendix B).Population cohorts were identified following the protocol from Chapter 2 (Sec-tion 2.2.5). Once identified, finite mixture distributions (MacDonald and contri-butions from Juan Du, 2011) were fitted to the size-frequency histograms. Thiswas followed by iterative trials of parameter estimates until the best fit to thebiological data was found. It is important to be aware of the temporal progres-sion of each cohort identified as certain sequences of parameter estimates willlikely yield rates that could be biased. So as to limit the extent of personal biasin the construction of finite mixture distributions, parameter constraints werenot included unless necessary, such that the temporal progression of each pa-473.3. METHODSrameter (mean shell size and the standard deviation for each cohort identified)was biologically appropriate (see Section B.5 for details on placing constraints).Each cohort was named after their probable dates of being spawned, with cprefixed to the abbreviated date. (e.g. cApr01 signifies a cohort recruited onApril 1, 2010).3.3.4 Spawning EventsSpawning events were identified by an increase in the abundance (ind.m−3) ofthe population and the size-group ≤ 0.15 mm. Dates of probable spawningwould then be recognized by periods displaying a combination of an increase inboth total population abundance and the density (ind.m−3) of the size-group≤ 0.15 mm. Daily estimates for the total population abundance was based onL. helicina processed for size frequency enumerations, as well as those consid-ered for abundance estimates. Only L. helicina processed for size frequencyenumerations were considered for density estimates of the size-group ≤ 0.15mm.3.3.5 Shell Size Growth and MortalityThe daily growth rate of the population was calculated from averaging the dailygrowth rates of each cohort tracked, during the period from March 22 to July7, 2010. The daily cohort growth was calculated using Equations 2.1 and 2.2from Chapter 2.Cohort densities were estimated by using a combination of the proportionalabundance of each cohort tracked and the total population abundance. Theproportional abundance of each population cohort was one of the parametersobtained from fitting the finite mixture distributions (along with the modalshell size and standard deviation of each cohort). The temporal densities of483.3. METHODSeach cohort tracked was estimated by multiplying the proportional abundanceof each cohort by the total population abundance (ind.m−3) for each day.The daily density of each cohort was log-transformed and regressed againstthe duration each cohort was tracked. Regression lines were super-imposed foreach cohort on Figure 3.4, where the slope was an indication of either increas-ing (positive slope) or decreasing (negative slope) density for each cohort. Anegative slope was indicative of the mortality experienced by each cohort, withthe daily mortality rate given by the slope value.493.3. METHODSFigure 3.1: The British Columbia west coast. Rivers Inlet is depicted in thezoomed inset of the mainland coast of B.C. The location of the Dawsons Dailystation is indicated within the zoomed inset.503.4. RESULTS3.4 Results3.4.1 Daily ChlorophyllLarge differences were seen in daily chlorophyll from January to July, 2010.While a small peak was observed in early January, chlorophyll values wererelatively low throughout the month when compared to the values recordedfrom April 1 onwards (Figure 3.2A). Chlorophyll (chl-a) levels although variablefrom spring to summer showed increased daily fluctuations during the summermonths. High Chl-a levels (> 16 µg.l−1) were observed throughout April withlower values (< 8 µg.l−1) recorded late in the month. Lower values were main-tained throughout May with sporadic peaks (Figure 3.2A). Early summer (lateMay to early June) displayed high chl-a concentrations (> 20 µg.l−1) beforesubsiding as the season progressed to minimal values (< 5 µg.l−1) in late June(Figure 3.2A). Chl-a levels increased to maximum values of > 60–70 µg.l−1 inmid July.Based on the higher chlorophyll values sustained throughout April, the 2010season had likely experienced a prolonged bloom that was initiated early in theseason (Figure 3.2A).3.4.2 Daily Population AbundanceA strong seasonal cycle was observed, with marked contrasts between springand summer (Figure 3.2B). Mean population abundances in spring (late Marchto April) were relatively low compared to those during summer, with May, Juneand July displaying means in excess of 1000 ind.m−3 (Figure 3.2B). Increasingin late April and continuing throughout early to mid May, pteropods reachedmaximum densities (> 4300 ind.m−3) on May 20 before declining for the re-mainder of the month (Figure 3.2B). An increasing trend was also noticeable513.4. RESULTSfrom early to late June, with peak abundances of > 4700 ind.m−3 observed onJune 30 which rapidly declined in July (Figure 3.2B).3.4.3 Daily Population Size-StructureDuring late spring the L. helicina population was represented largely by smallindividuals with only a few samples showing individuals ≥ 0.5 mm, in lateMarch and early April (Figure 3.3). As the seasons progressed the range inpteropod size expanded and a bi-modal distribution was observed in the size-frequency histograms during summer months (see the size-frequency histogramsin Appendix B). A reversion to a uni-modal size structure was however evidentduring July.The largest individuals (≥ 3 mm in size) was observed on June 22, while thesmallest (< 0.1 mm) recorded on June 13.With few exceptions, the majority of individuals measured were ≤ 1 mm(Figure 3.3). Two individuals, each ≥ 2.5 mm, were sampled on June 6 and20. Overall however, smaller size-groups (≤ 0.5 mm) dominated the pteropodpopulation throughout the observation period. This was confirmed by a smallermean shell size (see the lowess smoother in Figure 3.3). Additionally, there werefew larger individuals early in the season when compared to the population size-structures in 2008 and 2009, from Chapter 2.3.4.4 Spawning EventsFrom the daily densities of the size-group ≤ 0.15 mm, spawning likely occurredcontinuously (Figure 3.2C). Mean densities of ≤ 0.15 mm individuals increasedfrom < 1 ind.m−3 in March to > 30 ind.m−3 by June (Figure 3.2C). Dailyvariations in the density of this size-group was low in late March and April,with increasing variation in May and June.523.4. RESULTSBased on the seasonal peaks in total population abundance in late May andlate June, the late spring and summer appear to be the two periods of peakspawning. The spring spawning period (late April to late May) featured a slowbut steady increase in the size-group ≤ 0.15 mm, with densities peaking at > 19ind.m−3 on May 11 (Figure 3.2C). The summer spawning period (early Juneto early July) showed peak abundances exceeding 110 ind.m−3 on June 26, andwas characterized by considerable daily variation (Figure 3.2C). Reaching peakdensities on June 26, the summer peak of the size-group ≤ 0.15 mm paralleledthe increase in total population abundance during the same time period, sug-gesting late June as a time of peak spawning. This was not reflected during latespring as the peak in total population abundance was not matched by a similarpeak for the size-group ≤ 0.15 mm.Identification of CohortsTable B.2 in Appendix B contains the statistical results and constraints placed,for the finite mixture distributions fitted. See Appendix B.5 for the finite mix-ture distribution figures, for each month.3.4.5 Cohorts Identified and TrackedTwenty-four individual cohorts (Figure 3.2D) were identified (7 in April, 8 inMay, and 9 in June) and tracked for variable lengths of time (up to 34 days – seethe life table data in Appendix B.7). Their daily growth trajectories are plottedin Figure 3.2D. Transparency was added to the growth trajectory for each cohorttracked, revealing that the growth trajectories of many cohorts were merging to-gether at different times (shown by the colour boldness - more bold means morecohorts were merging together), and was interpreted as merged cohorts (e.g.≥ 2 cohorts combining together) for the overlapping periods. Consequently, the533.4. RESULTSgrowth trajectories became progressively more unrealistic (e.g. too much over-lap and insufficient number of individuals measured) once a size of 0.5 mm to1 mm was reached. Accordingly, only the size-group < 0.5 mm was consideredfor daily estimates of seasonal growth and mortality.3.4.6 Seasonal GrowthLimacina growth rates were highly variable throughout the season with a generaldecreasing trend observed throughout April. This was followed by generallyincreasing trends observed in early May (May 1–13), and from late May tomid June (May 26 to June 16) (Figure 3.2E). When the data points from May27, June 3, and June 10 were removed, linear regressions showed a significantincrease in growth rates (R2 ≥ 0.65, p < 0.01) (Table B.4 in Appendix B) duringthese two time periods (May 1–13 and May 26 to June 3). Growth rates rangedfrom 0.0005 mm.day−1 to 0.08 mm.day−1, with an average of 0.03 mm.day−1estimated for the time span between April 1 to July 7, 2010.Environmental CorrelationThere was no significant correlation (R2 < 0.009, p > 0.05) between short termgrowth rates and chl-a concentration, for both May 1–13 and May 26 to June16 time intervals.3.4.7 Daily MortalityFor each cohort tracked, the range of R2 value from 0.002 to 0.610 indicated highdaily variability. The majority of cohorts tracked showed a temporal increasein density, while comparatively few cohorts displayed a temporal decrease. Twoperiods of short term mortality were observed (e.g. decreasing regression lines).Cohorts cMay16 and cMay20 showed a significant decrease in density from May543.4. RESULTS16 to June 4 (cMay16: p < 0.05; cMay20: p < 0.01), and cJun15 and cJun18showed a significant decrease in density from June 29 to July 6 (cJun15: R2 =0.83, p < 0.05; cJun18: R2 = 0.82; p < 0.01, although the entire time periodthat cJun15 and cJun18 was tracked, was found to be insignificant) (Figure 3.4).An average mortality rate of 0.14.day−1 was estimated for the cohorts cMay16and cMay20, while an average rate of 155.1.day−1 was estimated for cJun15 andcJun18, for the period from June 29 to July 6.553.4. RESULTSFigure 3.2: Composite 5 x 1 panel figure. Panel A depicts the daily variation in fluorescencefrom 1 January to 20 July, 2010. A data gap exists between mid-February to late-March. PanelB portrays the daily variation in the population of L. helicina observed from 22 March, 2010to 7 July, 2010. Panel C portrays the daily variation in abundance of the size-fraction ≤ 0.15mm in shell size. Panel D displays the daily development in shell size of each cohort identifiedand tracked, from 1 April, 2010 to 7 July, 2010. Note that there is transparency added suchthat the bold colours depicts the times when ≥ 2 components had combined together to growas one component. Panel E portrays the daily growth in shell size of the population of L.helicina. These are the average values of the daily growth in shell size of each componenttracked, in panel D. The dashed red lines marks the first day of each month, starting on 1April. The shaded yellow region in each panel was the likely duration of the bloom experiencedin 2010.563.4. RESULTSFigure 3.3: Scatterplot of the daily population size structure of L. helicina.The data points are colour coordinated with the months of observation; blue- March, red - April, purple - May, green - June, yellow - July. There istransparency added to show the proportional presence of certain size-fractionsin the population, hence the bolder the colour, the more measurements wererecorded for the specific shell size. The size of the data point reflects the shellsize of the individual measured. A blue loess smoother shows the seasonal trendin the development of the size structure.573.4.RESULTSFigure 3.4: Composite figure of the log-transformed daily density of each cohort identified and tracked. Linear regressions arefitted against the respective time periods for each cohort tracked (indicated by the dashed red line). All cohorts are indicated bytheir abbreviated name (e.g. “cApr01”–cohort recruited on April 1). Significant mortality was identifid for cMay16 (p < 0.05)and cMay20 (p < 0.01) in late spring. Significant mortality was also observed for the period of June 29 to July 6, for cJun15(p < 0.05) and cJun18 (p < 0.01). Note the limited number of cohorts (8) showing a decrease in density.583.5. DISCUSSION3.5 Discussion3.5.1 Spawning, Cohorts, and Size-Structure Development:Comparison to Chapter 2 and Relevant LiteratureBased on the results of Chapter 2, L. helicina was hypothesized to spawn con-tinuously. A main spawning event occurred in late spring and this was followedby protracted spawning throughout summer. The smaller size-groups generallyhad a high proportional abundance (> 40% of the population) in all years ofobservation, supporting the occurrence of continuous daily spawning. However,it was apparent from the seasonal differences observed between years that thetime resolution of Chapter 2 may have been too coarse to determine the sea-sonal patterns in the life cycle of L. helicina. Evidence from the daily dataclearly showed that spawning was occurring at a high frequency, and protractedspawning appeared to be the norm.The literature (Kobayashi, 1974; Dadon and de Cidre, 1992) indicated thatthe reproductive cycle of L. helicina involves a period of enhanced spawningactivity followed by a prolonged period of reduced activity. Building from thelife cycle study of Chapter 2, the identification of a peak spawning event insummer, from the daily data pointed to the possibility for the newly spawnedsummer cohort to overwinter with the spring cohort (e.g. C1 and C2 identifiedin Chapter 2). Both the C1 and C2 cohorts (spring cohorts) were observed tosuccessfully overwinter and it is assumed the same is true for the summer cohort.Dadon and de Cidre (1992) observed the summer cohort of L. retroversa to growat slower rates compared to the spring cohort, due to reduced food availability insummer. Considering L. helicina in this study, it may be possible for the slowergrowing summer cohort to “catch-up” (in terms of shell size) to the remnants ofthe newly spawned spring spring cohort after summer spawning, although the593.5. DISCUSSIONnew spring cohort is likely approaching senescence (lower relative abundance oflarger individuals). Although the summer cohort may be growing slower (whencompared to the newly spawned spring cohort, possibly during a spring bloom),their growth rates are likely equivalent if not faster than those of the newlyspawned spring cohort still remaining after summer peak spawning. This allowsthe summer cohort to “catch-up” to the spring cohort (e.g. larger individualsof the spring cohort are fully mature and most have already all spawned, henceslower growth rates or no further growth). Maturing by late April to earlyMay of the next year, it is probable that spring spawning was initiated byeither the summer cohort, or a combination of both the spring and summercohorts from the previous year. As significant relations were found between theseasonal distribution of L. helicina and 30 m depth-averaged temperature inChapter 2, it is probable that spring spawning may be partly triggered by slightincreases in sea water temperature. The increase in total population abundanceduring early May (from daily data) provided an indication of spring spawning.Spawning sizes of > 3 mm was identified in Chapter 2, and although the largestindividual (> 3 mm) from the daily data was observed on June 21 (e.g. almost2 months after initiation of spring spawning), the absence of larger individualsthroughout May could be a result of either advection, mortality (e.g. predation),or the patchy distribution of L. helicina. Detection of these larger individuals inlate June may attest to the 1.2–1.5 year life cycle longevity, estimated in Chapter2. Indeed, if the spring and summer cohorts are both able to overwinter thenthe smaller shell sizes observed in the following spring (from Chapter 2) maybe explained (e.g. a larger proportion of smaller summer cohort individuals toindividuals of the larger spring cohort). Dadon and de Cidre (1992) postulatedthe ability of the spring cohort (L. retroversa) to grow to sexual maturity bysummer, to spawn the summer cohort. Summer is asserted in Gannefors et al.603.5. DISCUSSION(2005) and Bednarsˇek et al. (2012) as a time of enhanced spawning activity,with which we are in agreement. Based on an average growth rate of 0.03mm.day−1, individuals spawned (assumed shell size of 0.15 mm at spawn) onMay 1 (e.g. spring cohort) are able to reach the respective sizes of 1.98 mm,2.58 mm, and 3 mm by June 30, July 20, and August 3, respectively. Withthe highest population abundance recorded at the end of June, it is probablethat the spring cohort initiated spawning early in the month and at a smallersize. Shell sizes of 1–1.2 mm characterized the spring cohort between late Mayand early June. From this, it is apparent that L. helicina in Rivers Inlet maybecome capable of spawning once a shell size of ≥ 1.00 mm is reached. This iscredible given that L. helicina is able to develop mature ova from a shell sizeof ≥ 0.70 mm (Kobayashi, 1974). Based on the continually high proportion ofsmaller individuals after July, from Chapter 2, it would appear that L. helicinaexhibits reduced spawning activity after peak spawning in summer. The Maypeak in population abundance (from daily data) was not matched by a parallelpeak in the smaller size-groups (≤ 0.15 mm), and it is possible that the springspawning event was missed. Spatial analysis from Chapter 2 showed that L.helicina, although exhibiting broadly similar size-structures across the inlet, doexhibit spatial differences in density (between size-groups). Due to the surfacewaters at Dawsons Landing always being a state of flux, it is very unlikely thesame water mass was sampled. Consequently, there is high potential for patchesof L. helicina to be advected out of the system, to be replaced by other patches(or portions thereof) from different areas.High frequency spawning combined with the substantial overlap betweencohorts (see the finite mixture distributions in Appendix B) makes it highly un-likely every newly spawned “cohort” was distinct. Instead, the newly spawnedcohorts during spring and summer can be considered additive portions (e.g. re-613.5. DISCUSSIONcruits), and only those contributing to the overall recruitment for the year areconsidered “true” cohorts. The remaining recruits likely succumb to variousforms of mortality (most likely predation). Wootton et al. (2008) proposed anincrease in ocean pH from CO2 uptake by photosynthesis, which is heightenedduring bloom conditions, to increase the CaCO3 saturation state. Althoughthe short term increase in CaCO3 saturation state is very minimal (e.g. sug-gesting that there is no real benefit), there is the possibility that any beneficialchanges in environmental conditions may assist developing cohorts (Woottonet al., 2008).The population size structure was generally uni-modal throughout late Marchand April. A multi-modal size structure characterized the population by midMay, which we interpret as a developing spring cohort. Multi-modality contin-ued into late June, with the smallest size-groups displaying the strongest signalsfrom early–late June (see the finite mixture figures for June in Appendix B).Disappearing by early July when the size structure became more uni-modal,the general absence of the larger individuals (≥ 1 mm) likely signified at leastpartial die-off of the spring cohort, after summer spawning. A summer mortal-ity of newly spawned recruits may account for a large decrease in the pteropodabundance observed in early July. However, as sampling had terminated earlyin the month, the absence of the larger individuals could also be interpreted asan sampling artifact (e.g. patchiness).3.5.2 Caveats to Estimating Daily GrowthShort term trends of significant growth were detected prior to the spring andsummer periods of increased spawning. Individual growth rates declines orceases entirely during the final stages of sexual maturation when energy is dedi-cated towards reproduction (Lalli and Gilmer, 1989). In this light, the declining623.5. DISCUSSIONgrowth rates during mid May in Rivers Inlet may have been an indication thatthe overwintering cohort had reached sexual maturity. Similarly, the springcohort reaching sexual maturity may account for the rapidly declining growthrates in late June. Estimations of daily growth rates were complicated by nu-merous recruits of the summer cohort merging together throughout June (forup to 10 days or more), and recruits exhibiting high short term variations ingrowth rates. Because these recruits were indistinguishable, the average growthrates calculated could overestimate values for larger individuals.Environmental CorrelationThe non-significant relation between seasonal growth rates and chlorophyll-amay have been related to the fact that fluorescence values were measured at 5m depth. If the fluorescence values measured from 5 m depth were not represen-tative of full water column integrated chlorophyll concentrations, then relation-ships between seasonal growth rates and chlorophyll-a concentrations may beobscured. However, it is most likely that L. helicina may have exhibited a de-layed response to temporal increases in chlorophyll-a. If so, then the heightenedchlorophyll-a concentrations observed through most of April and late May/earlyJune may have been the driving force responsible for the peak population abun-dances seen, in mid May and late June, respectively. The termination of zoo-plankton sampling in early July was unfortunate given the very large increasein chlorophyll-a by mid July.3.5.3 Estimating Daily Mortality and Problems Encoun-teredObservations of numerous recruits exhibiting temporal increases in density (in-stead of a decrease) complicated the estimates of seasonal mortality for L. he-licina. Based on the highly variable spatial density distribution for each recruit633.5. DISCUSSIONcohort, the variability (large increases on day followed by large decreases thenext) in individual cohort density observed at the daily station may be primarilyexplained by the advection of different patches of L. helicina into the samplinglocation. Therefore, for example, large decreases in density could be attributedto both mortality and L. helicina that are advected away from the daily station.It appears that although we were sampling the same population of L. helicina inRivers Inlet, our initial assumption about low spatial density variability couldbe an oversimplification. Clearly, following an eulerian approach, temporal de-creases in cohort density cannot be solely attributed to mortality. Mortality mayhave been further complicated by high-frequency spawning events (e.g. springand summer peak spawning) and the possibility of cohorts merging together atdiffering times.Despite the above caveats, there were short-term periods of significant mor-tality. The significant declines shown by the recruits cMay16 and cMay20 fromMay 16 to June 3, as well as cJun15 and cJun18 from June 29 to July 6 werecoincident with a high decrease in total population abundance (after the springand summer spawning events). Assuming an exponential decrease in mortalitywith increasing shell size, a mortality rate of 0.14.day−1 in late spring translatesto a daily mortality of 13 %, which does not appear to match the large decreasein total abundance. A mortality rate of 155.1.day−1 for the period of June 29 toJuly 6, equates to a daily mortality of 100 % which appears to match the largedecline in both total population abundance and the density of the size-group≤ 0.15 mm after the summer spawning. Based on this, it appeared that thesmallest size-groups experienced the highest mortality after peak spawning insummer. This is in agreement with literature studies of copepod egg-mortality(Peterson and Kimmerer, 1994; McLaren, 1997). Investigating the influenceof egg-mortality on the recruitment of Temora longicronis, Peterson and Kim-643.5. DISCUSSIONmerer (1994) found high mortalities in spring (162.5 day−1) and summer (21.57day−1) (Table 2 in Peterson and Kimmerer, 1994). The high mortality ob-served in spring was hypothesized to be caused by cannibalism (Peterson andKimmerer, 1994). As L. helicina are also known to feed on younger L. helicina(e.g. small individuals trapped in the feeding web), it may be likely that thehigh mortality in late June was partly due to cannibalism (Lalli and Gilmer,1989).3.5.4 Potential Sampling ErrorsOnly 30 m Sampled...A sampling depth of 30 m considerably limited the representative sampling ofthe larger size-groups (compared to 300 m in Chapter 2). This could haveresulted in a bias and forced inferential interpretations for seasonal rates ofgrowth and mortality. Even though the larger size-groups formed only a smallcomponent of the population, the accurate estimate of growth and mortalityrates for the larger individuals are still required if size-dependent growth andmortality are to be resolved (Bednarsˇek et al., 2012).Influence of AdvectionThe spatial analysis from Chapter 2 put forward the hypothesis that the L.helicina is broadly similar across the inlet. However, the density-differencesbetween stations for varying size-groups, pointed to the patchiness of L. helicina.Highlighted in our attempts to track numerous population recruits, it is evidentthat advection could have a large impact on the temporal distribution of L.helicina. Because RI hydrodynamics is largely influenced by freshwater inputand hence, highly seasonal (e.g. the rate of surface-freshwater input can changefrom 100 m−3.s−1 from winter-spring, to ∼ 1000 m−3.s−1 in the summer, Hodal,653.5. DISCUSSION2010), the temporal size distribution of L. helicina across the inlet would likelyvaried seasonally as well. Since only 3 dates were considered for spatial analysisin Chapter 2, more work is needed to determine the influence of advection onthe size-specific distribution of L. helicina in RI (e.g. high resolution vertical-distribution studies throughout the seasons).Potential Limitations with the Mixdist Statistical PackageIn periods of continuous spawning, there is the possibility of the new recruitsmerging together to grow in synchrony for varying lengths of time. While doingso, it is also possible that certain proportions of the merged group can breakaway at different times to grow at faster or slower rates than the merged group.If each new recruit cohort is indistinguishable from one another, the growth ratedynamics that each will experience is masked by the presence of the others.The mixdist statistical package appears to be unsuitable for the tracking ofthe numerous recruits of L. helicina. Because there were many newly spawnedrecruits merging together for a lengthy period (≥ 10 days), there is still un-certainty concerning the growth rates estimated. We suggest that mixdist isinadequate for the study of animals exhibiting protracted spawning such as L.helicina, and propose that the most efficient method of estimating growth rate isto successfully culture L. helicina in experimental aquaria, for extended periods.Size Frequency Method...For studies of L. helicina, shell size may not be an adequate proxy of age. Theseasonal growth of L. helicina as well as other Limacina spp. is said to be de-pendent upon environmental conditions (Dadon and de Cidre, 1992; Hunt et al.,2008). Accordingly, problems can arise when two individuals of similar size areat different stages of sexual maturation, due to environmental influence (Dadonand de Cidre, 1992). Nonetheless, the size frequency method may be the only663.5. DISCUSSIONviable option to study the seasonal dynamics of L. helicina, efficiently. Accord-ing to Aksnes et al. (1997), a physical and quantifiable character is required forthe study of natural populations with time series data. Furthermore, becausethere have been no studies of the development stages nor stage duration timesof L. helicina in the literature, the use of shell size as a proxy of age appears tobe the only option (Aksnes et al., 1997).Using high temporal resolution data, and building from the results of Chap-ter 2, this chapter has revealed many facets of L. helicina seasonal dynamics.Namely, we have confirmed the phenomenon of protracted spawning hypothe-sized in Chapter 2. In addition to the spring peak spawning period identifiedin Chapter 2, the daily data also showed increased spawning during summer.With the initiation of spring spawning (in mid-May), spawning was continuousand identified every 2–4 days leading to the 2nd peak in summer (late-June).By tracking numerous recruits through the period of April 1 to July 7, 2010,an average population growth rate (in terms of shell size) of 0.03 mm.day−1was estimated. Short-term periods of significant growth were identified for theperiods of May 1–13 and May 26 to June 16. This was indicative of the sex-ual maturation of L. helicina prior to the spring and summer periods of peakspawning. Based on an average growth rate of 0.03 mm.day−1, it was evidentthat the recruits spawned in spring are able to grow to a size of up to 3 mm(becoming sexually mature) by August (assuming May 1 as time of spawn andan initial shell size of 0.15 mm). Our attempt to estimate the seasonal mortalityof L. helicina was greatly undermined by a combination of advective influences,patchy distributions, and the likelihood of merged recruits. However, significantmortality did appear to occur during certain periods, coincident with the daysafter spring and summer peak spawning (e.g. late-May and late-June). Thiswould suggest that the smallest size-groups experienced the highest mortality673.5. DISCUSSIONafter being spawned. Due to the caveats in our estimates of daily growth andmortality rates, they should be interpreted as very generalized and approximateestimates.The results from this chapter also highlighted the importance of advection.Clearly, the assumption of a negligible advection must be considered with greatcaution when studying the seasonal dynamics of L. helicina. With this said, itseems the most effective method to study L. helicina (i.e. to avoid problemsassociated with advection and patchiness) is to successfully culture them inlaboratory aquaria. There is already headway being made in this respect (seethe review of pteropod culture techniques in Howes et al., 2014). Culturingpteropods may be the most effective method in the identification and tracking ofthe newly spawned recruits. Because they are already accustomed to the cultureconditions (when spawned), any temporal changes observed in the individuals(e.g. growth rate, life cycle longevity) should be physiologically mediated (e.g.not biased by culture conditions), and representative of natural populations(Howes et al., 2014). However, estimating the natural mortality of L. helicinastill presents additional challenges (e.g. cannot be representatively estimatedvia. culture). A possible solution may be to identify and track water parcels (alagrangian approach) by using temperature and salinity signatures. This way,it may be possible to continually sample the same population of L. helicina overa prolonged period.68Chapter 4Life Cycle of L. helicina : AConceptual Model andGeneral ConclusionsDespite various studies investigating the life cycle of L. helicina, there isa lack of consensus regarding fundamental aspects of its life history. Previousstudies have suggested that the life cycle of L. helicina ranges from 1.5–2 yearsin the Central Arctic Ocean, Canadian Basin to over 3 years for L. helicina ofthe Scotia Sea, Southern Ocean. Studying the life cycle of L. helicina in thesubarctic Pacific and Atlantic, respectively, Fabry (1989) and Gannefors et al.(2005) were in agreement with an annual life cycle. Dadon and de Cidre (1992)postulated a 1–1.5 year life cycle but for L. retroversa, a species with a similarreproductive biology to L. helicina.This thesis used datasets of high temporal resolution combined with the sizefrequency method to examine the life cycle of L. helicina. Population cohortswere identified and tracked using the mixdist statistical package. In Chapter2, two cohorts were identified and tracked for > 400 days. From this, a life cyclelongevity of 1.2–1.5 years was estimated, with 1.5 years being the likely max-imum. Both cohorts were observed to successfully overwinter, although both69exhibited a reduced shell size in the following spring. Observations from Lischkaand Riebesell (2012) provides evidence of L. helicina successfully overwintering.Additionally, there was no growth or spawning activity during the winter period(Lischka and Riebesell, 2012). Based on the seasonal abundances, late springwas hypothesized as a time of peak spawning activity. From the size-frequencyhistograms, prominent growth was observed in spring and summer with evidentdevelopment of the population size structure. The dominant influence of thesmallest size-groups throughout the seasons for all years, suggested that L. he-licina is capable of protracted spawning. Utilizing higher resolution data, theresults of Chapter 3 indeed confirmed that L. helicina is capable of protractedspawning, with the newly spawned recruits identified every 2–4 days. Springand summer was affirmed as the times of peak spawning activity. With theidentification of increased summer spawning as well as a summer cohort, thereis a high probability that both the summer cohort and the spring cohort are over-wintering together. Overwinter survival may be achieved by a combination ofreduced metabolic activity, and migration to deeper depths (Maas et al., 2012).Significant growth (mmshell growth.day−1) was found for the periods prior toboth the spring and summer spawning events, although this was not related tochlorophyll-a concentrations at 5 m depth. However, the daily distribution ofL. helicina pointed to the possibility of a delayed response to seasonal periodsof high chlorophyll-a, which may have obscured any evident relation betweenchlorophyll-a and seasonal growth. Although only short-term growth was signif-icant, the average seasonal growth rate showed that the spring cohort is capableof growing to maturity by summer (like L. retroversa). This indicated a po-tentially much smaller spawning size when compared to the literature, and thespawning sizes (≥ 3 mm) from Chapter 2. Attempts were made in Chapter 3to accurately estimate instantaneous mortality. Difficulties were encountered704.1. A CONCEPTUAL MODELas it became clear that we were not sampling the same population from dayto day. Additionally, many newly spawned recruit cohorts were likely mergingduring spring and summer spawning, which considerably biased the parameters(modal shell size and proportional abundance for each cohort) produced fromfitting finite mixture distributions. Consequently, the estimates of daily growth(in terms of shell size) and mortality rates should be regarded as first approxi-mates. The potential merging of cohorts pointed to the inability of mixdist todistinguish between merged population recruits, suggesting that this statisticalpackage may be unsuitable for the study of L. helicina. Although many recruitsshowed temporal increases in density, significant decreases were identified fortwo recruits during mid–late May (after peak spawning in late spring). Addi-tionally, significant decreases in late June (short term), coincided with sharpdeclines in total population abundance. However, as it was highly unlikely thatthe same population was continuously sampled, periods of significant decline indensity could not be attributed solely to mortality. Even so, the high mortalityobserved from late June to early July suggested that the smallest size-groups ofL. helicina experienced the highest mortality after peak spawning in summer.4.1 A Conceptual ModelIntegrating the results from Chapters 2 and 3, the life cycle of L. helicina canbe described by the conceptual model in Figure 4.1. At Yeart−1, the cohortspawned in spring is able to utilize resources from the spring bloom to grow toa shell size corresponding to sexual maturity by summer. Summer spawning isinitiated once sexual maturity is reached, and is likely triggered by increasingsea water temperature. Despite the majority of the spring cohort dying off aftersynchronized spawning, there is still a small proportion that continues to survive(and possible grow) into the summer and fall months (thin arrows in Figure 4.1).714.1. A CONCEPTUAL MODELAs this happens, any adult individuals that have yet to release their eggs doso throughout the late summer and early fall, resulting in low levels protractedspawning (e.g. small scale spawning events–unsynchronized). Utilizing the re-maining resources from the fall bloom, the summer cohort may be able to growat comparatively faster rates, to reach a shell size nearly equivalent to that ofthe spring cohort, which is likely experiencing some mortality towards summersend. This may explain the high overlap between cohorts throughout the seasons,and between years. Entering the winter, both the spring and summer cohortsare able to survive into the next year, possibly by migrating to deeper depthsand experiencing reduced metabolic rates (Maas et al., 2012). During the latefall/winter period, the Limacina exhibits no shell growth and all reproductiveactivity ceases. With the comparatively lower food availability during winter,as well as the likelihood of increased predation, the overwintering spring cohortexperiences comparatively higher mortality (e.g. from senescence and/or preda-tion), resulting in a larger ratio of smaller summer cohort individuals (Yeart−1)to larger spring cohort individuals (Yeart−1) in the following spring. Becomingmore active in the spring of Yeart, limited spawning activity by portions of thesummer cohort (already at shell sizes of potential spawning) may be triggeredby slight increases in sea water temperature during spring. When environmentalconditions become more optimal in late spring (e.g. warmer sea water temper-atures and increased food availability from the spring bloom), the overwinteredsummer cohort experiences rapid growth and with the full maturation of thesummer cohort (from Yeart−1), begins the spring period of peak spawning. Itis also possible for the small portion of the spring born cohort from Yeart−1 tocontribute to the spring spawning event in Yeart, however, the majority of thespawning is from the overwintered summer cohort from Yeart−1. Depending onthe environmental conditions during the summer/fall period of Yeart−1 and the724.1. A CONCEPTUAL MODELwinter conditions transitioning into Yeart, there is a possibility that the summercohort overwinters by itself, with the still present spring cohort, or not at all(e.g., only the spring cohort overwinters).Based on this model, the survival of the population to Yeart+1 depends onthe recruitment of the summer cohort (and potentially the spring cohort), andits survival through winter. In turn, this will depend on the timing of the springspawning period and also, the magnitude of the spring bloom. In the case ofa delayed spring bloom, there may be a delayed spring spawning such that thesummer cohort will be spawned later in the season, meaning they may haveinsufficient time to grow before the onset of winter. Similarly, if the bloommagnitude is insufficient to allow the effective growth of the spring cohort, itmay result in delayed summer spawning or reduced recruitment during sum-mer spawning. This means the summer cohort will enter the fall–winter periodat a smaller size and in reduced numbers, thereby decreasing the summer co-horts likelihood of successfully overwintering. Given that the majority of thespring cohort likely dies off after peak spawning, the reduced probability of thesummer cohort successfully surviving the winter suggests a low survivorship ofthe L. helicina population into the next year. If winter survival is successful,the smaller adults from the summer cohort may require longer periods to reachmaturity in the next year. Again, this may lead to delayed spring spawning. In-terestingly, depending on the bloom timing, there could be various proportionsof the cohort(s) (spring and summer) overwintering to the next year (e.g. a de-layed bloom could mean a higher proportion of the spring cohort overwintering,whereas an earlier bloom may infer a lower proportion). This may have impli-cations to recruitment in the following year (e.g. higher proportion of springcohorts may mean lower recruitment in the following spring as the majorityof the spring cohort have either already spawned in the previous summer, or734.2. GENERAL CONCLUSIONShave died off). Regardless, the possibility of two cohorts overwintering maydramatically increase the stability of the population dynamics.4.2 General ConclusionsBy investigating the inter-annual and seasonal dynamics of the L. helicina pop-ulation in Chapters 2 and 3, respectively, we have shown that 1.) the normalseasonal cycle of recruitment is characterized by continuous spawning activityoutside winter months, 2.) although continuous, these periods are punctuatedby short term episodes of very intense reproductive output during the late springand summer, and 3.) reproductive activity terminates during the late fall andwinter periods, and commences again in the following spring. Accordingly, wepropose that spring and summer are the two primary periods of spawning ac-tivity, with low level but continuous spawning outside these times, with theexception of winter. The spring and summer cohorts result from these two peri-ods of intense spawning. Assuming uniform growth, it is plausible that 4.) thespring cohort reaches sexual maturity by summer, and subsequently spawns thesummer cohort. Finally, we have also shown that 5.) it is possible that boththe spring and summer cohorts overwinter into the following year, which mayexplain the high overlap between cohorts in the size-frequency histograms.In our attempt to estimate the daily growth and mortality rates of L. helic-ina, we’ve discovered potentially significant obstacles hindering estimate deriva-tion. Most notably, the influence of advection combined with the inherentlypatchy distribution of L. helicina created great difficulty during our euleriansampling. Consequently, the estimates likely represent underestimates with highuncertainty. Regarding estimates of growth rates, the most viable option (cur-rently) would appear to be the successful culturing of Limacina in experimentalaquaria. If this can be accomplished, the newly spawned can be effectively mon-744.2. GENERAL CONCLUSIONSitored through time to obtain accurate estimates of growth in shell size, biomass,etc. However, recent literature has only begun to address the potential benefitsand challenges of culturing Limacina. For representative estimates of mortality,it is crucial to temporally sample the same group of specimens through time.As experimental aquaria does not simulate the natural environment (e.g. zoo-plankton community). A feasible course of action is to track parcels of waterduring a lagrangian study Since Rivers Inlet is highly stratified, the verticaldistribution of L. helicina in the inlet should be investigated.Due to the unprecedented rates of change in seawater chemistry and tem-perature, it is crucial to gain an mechanistic understanding of the forces drivingthe seasonal and inter-annual changes observed in the L. helicina population.Predictive models can then be constructed and optimized to better understandthe synergistic effects of increasing temperature and acidity, as well as other pa-rameters, on L. helicina populations over time. Once optimized, it is hoped theknowledge gained from these models will provide effective methods of monitor-ing and predicting the physical and biological changes occurring in the marineenvironment, into the future.754.2. GENERAL CONCLUSIONSFigure 4.1: Conceptual model depicting the life-cycle of L. helicina. The blue& red dots marks the spring and summer spawning events, respectively. Theyellow arrows mark the population abundance passing from one season to thenext. Larger arrows indicate a higher abundance. Blue arrows indicate thegrowth cycle of the spring cohort, while the red arrows marks the growth cycleof the summer cohort. The thin blue and red lines, indicates the potentiallife-cycle longevity of the spring and summer cohorts, respectively. Note theoverlap of cohorts throughout and between years.76BibliographyAksnes, D. L., Miller, C. B., Ohman, M. D., and Wood, S. N. (1997). Estima-tion techniques used in studies of copepod population dynamicsa review ofunderlying assumptions. Sarsia, 82(4):279–296.Aksnes, D. L. and Ohman, M. D. (1996). A vertical life table approach tozooplankton mortality estimation. Limnology and Oceanography, 41(7):1461–1469.Bednarsˇek, N., Tarling, G., Fielding, S., and Bakker, D. (2012). Populationdynamics and biogeochemical significance of Limacina helicina antarctica inthe scotia sea (southern ocean). Deep Sea Research Part II: Topical Studiesin Oceanography, 59:105–116.Bernard, K. and Froneman, P. (2009). The sub-antarctic euthecosome ptero-pod, Limacina retroversa: Distribution patterns and trophic role. Deep SeaResearch Part I: Oceanographic Research Papers, 56(4):582–598.Clarke, A. and Peck, L. S. (1991). The physiology of polar marine zooplankton.Polar research, 10(2):355–370.Comeau, S., Gorsky, G., Alliouane, S., and Gattuso, J.-P. (2010). Larvae of thepteropod Cavolinia inflexa exposed to aragonite undersaturation are viablebut shell-less. Marine biology, 157(10):2341–2345.77BIBLIOGRAPHYComeau, S., Gorsky, G., Jeffree, R., Teyssie, J., Gattuso, J., et al. (2009). Im-pact of ocean acidification on a key arctic pelagic mollusc (Limacina helicina).Biogeosciences, 6(9):1877–1882.Conover, R. and Paranjape, M. (1977). Comments on the use of a deep tank inplanktological research. Helgola¨nder Wissenschaftliche Meeresuntersuchun-gen, 30(1-4):105–117.Dadon, J. R. and de Cidre, L. L. (1992). The reproductive cycle of the thecoso-matous pteropod Limacina retroversa in the western south atlantic. MarineBiology, 114(3):439–442.Dixson, D. L., Munday, P. L., and Jones, G. P. (2010). Ocean acidificationdisrupts the innate ability of fish to detect predator olfactory cues. EcologyLetters, 13(1):68–75.Doney, S. C., Fabry, V. J., Feely, R. A., and Kleypas, J. A. (2009). Oceanacidification: the other co2 problem. Marine Science, 1.Du, J. (2002). Combined algorithms for fitting finite mixture distributions.Master’s thesis, McMaster.Fabry, V. J. (1989). Aragonite production by pteropod molluscs in the sub-arctic pacific. Deep Sea Research Part A. Oceanographic Research Papers,36(11):1735–1751.Fabry, V. J., Seibel, B. A., Feely, R. A., and Orr, J. C. (2008). Impacts ofocean acidification on marine fauna and ecosystem processes. ICES Journalof Marine Science: Journal du Conseil, 65(3):414–432.Gannefors, C., Bo¨er, M., Kattner, G., Graeve, M., Eiane, K., Gulliksen, B.,Hop, H., and Falk-Petersen, S. (2005). The arctic sea butterfly Limacinahelicina: lipids and life strategy. Marine Biology, 147(1):169–177.78BIBLIOGRAPHYGentleman, W., Pepin, P., and Doucette, S. (2012). Estimating mortality: Clar-ifying assumptions and sources of uncertainty in vertical methods. Journal ofMarine Systems.Gilmer, R. and Harbison, G. (1986). Morphology and field behavior of ptero-pod molluscs: feeding methods in the families cavoliniidae, limacinidae andperaclididae (gastropoda: Thecosomata). Marine Biology, 91(1):47–57.Gilmer, R. and Harbison, G. (1991). Diet of limacina helicina (gastropoda:Thecosomata) in arctic waters in midsummer. Marine ecology progress series.Oldendorf, 77(2):125–134.Harris, R. (2000). ICES zooplankton methodology manual. Academic Press, SanDiego,.Hlavac, M. (2013). stargazer: LaTeX code for well-formatted regression andsummary statistics tables. Harvard University, Cambridge, USA. R packageversion 3.0.1.Hodal, H., Falk-Petersen, S., Hop, H., Kristiansen, S., and Reigstad, M. (2012).Spring bloom dynamics in kongsfjorden, svalbard: nutrients, phytoplankton,protozoans and primary production. Polar biology, 35(2):191–203.Hodal, M. (2010). Net physical transports, residence times, and new produc-tion for rivers inlet, british columbia. Master’s thesis, University of BritishColumbia.Howes, E. L., Bednarsˇek, N., Bu¨denbender, J., Comeau, S., Doubleday, A.,Gallager, S. M., Hopcroft, R. R., Lischka, S., Maas, A. E., Bijma, J., et al.(2014). Sink and swim: a status review of thecosome pteropod culture tech-niques. Journal of Plankton Research, 36(2):299–315.79BIBLIOGRAPHYHsiao, S. C. (1939a). The reproduction of Limacina retroversa (flem.). TheBiological Bulletin, 76(2):280–303.Hsiao, S. C. (1939b). The reproductive system and spermatogenesis of Limacina(Spiratella) retroversa (flem.). Biological Bulletin, pages 7–25.Hunt, B., Pakhomov, E., Hosie, G., Siegel, V., Ward, P., and Bernard, K.(2008). Pteropods in southern ocean ecosystems. Progress in Oceanography,78(3):193–221.Hunt, B., Strugnell, J., Bednarsek, N., Linse, K., Nelson, R. J., Pakhomov, E.,Seibel, B., Steinke, D., and Wu¨rzberg, L. (2010). Poles apart: the bipolarpteropod species Limacina helicina is genetically distinct between the arcticand antarctic oceans. PloS one, 5(3):e9835.Kobayashi, H. (1974). Growth cycle and related vertical distribution of thethecosomatous pteropod Spiratella (“Limacina”) helicina in the central arcticocean. Marine Biology, P26:295–301.Lalli, C. and Gilmer, R. (1989). Pelagic Snails: The Biology of HoloplanktonicGastropod Mollusks. Stanford University Press.Lalli, C. M. and Wells, F. E. (1978). Reproduction in the genus Limacina(opisthobranchia: Thecosomata). Journal of Zoology, 186(1):95–108.Lischka, S., Bu¨denbender, J., Boxhammer, T., and Riebesell, U. (2010). Impactof ocean acidification and elevated temperatures on early juveniles of the polarshelled pteropod Limacina helicina: mortality, shell degradation, and shellgrowth. Biogeosciences discussions, 7(6):8177–8214.Lischka, S. and Riebesell, U. (2012). Synergistic effects of ocean acidificationand warming on overwintering pteropods in the arctic. Global Change Biology.80BIBLIOGRAPHYMaas, A. E., Elder, L. E., Dierssen, H. M., and Seibel, B. A. (2011). Metabolicresponse of antarctic pteropods (mollusca: Gastropoda) to food deprivationand regional productivity. Marine Ecology Progress Series, 441:129–139.Maas, A. E., Wishner, K. F., and Seibel, B. A. (2012). Metabolic suppressionin thecosomatous pteropods as an effect of low temperature and hypoxia inthe eastern tropical north pacific. Marine biology, 159(9):1955–1967.MacDonald, P. and contributions from Juan Du (2011). mixdist: Finite MixtureDistribution Models. R package version 0.5-4.MacDonald, P. and Pitcher, T. (1979). Age-groups from size-frequency data: Aversatile and efficient method of analyzing distribution mixtures. Journal ofFisheries Research Board Canada, 36(8):987–1001.MacDonald, P. M. (2011). Personal communications.Mackas, D., Thomson, R. E., and Galbraith, M. (2001). Changes in the zoo-plankton community of the british columbia continental margin, 1985-1999,and their covariation with oceanographic conditions. Canadian Journal ofFisheries and Aquatic Sciences, 58(4):685–702.McKinnell, S., Wood, C., Rutherford, D., Hyatt, K., and Welch, D. (2001). Thedemise of owikeno lake sockeye salmon. North American Journal of FisheriesManagement, 21(4):774–791.McLaren, I. (1997). Modeling biases in estimating production from copepodcohorts. Limnology and oceanography, 42(3):584–589.Morton, J. (1954). The biology of Limacina retroversa. J. mar. biol. Ass. UK,33:297–312.81BIBLIOGRAPHYMucci, A. (1981). The solubility of calcite and aragonite and the composition ofcalcite overgrowths in seawater and related solutions. PhD thesis, Universityof Miami.Ohman, M. D. and Wood, S. N. (1995). The inevitability of mortality. ICESJournal of Marine Science: Journal du Conseil, 52(3-4):517–522.Paranjape, M. A. (1968). The egg mass and veligers of Limacina helicina phipps.The Veliger, 10(4):322–326.Peterson, W. T. and Kimmerer, W. J. (1994). Processes controlling recruit-ment of the marine calanoid copepod temora zortgicoks in long island sound:Egg production, egg mortality, and cohort survival rates. Limnol. Oceanogr,39(7):1594–1605.R Core Team (2013). R: A Language and Environment for Statistical Comput-ing. R Foundation for Statistical Computing, Vienna, Austria.Redfield, A. C. (1939). The history of a population of Limacina retroversaduring its drift across the gulf of maine. The Biological Bulletin, 76(1):26–47.Seibel, B. A., Maas, A. E., and Dierssen, H. M. (2012). Energetic plasticityunderlies a variable response to ocean acidification in the pteropod, Limacinahelicina antarctica. PLoS One, 7(4):e30464.Tommasi, D., Hunt, B., Pakhomov, E., and Mackas, D. L. (2013). Meso-zooplankton community seasonal succession and its drivers: Insights froma british columbia, canada, fjord. Journal of Marine Systems.Wells, F. E. (1976). Growth rates of four species of euthecosomatous pteropodsoccurring off barbados, west indes. The Nautilus (Philadelphia), 90:114–116.Weslawski, J., Kwasniewski, S., and Wiktor, J. (1991). Winter in a svalbardfjord ecosystem. Arctic, pages 115–123.82BIBLIOGRAPHYWiens, J. A. (1989). Spatial scaling in ecology. Functional ecology, 3(4):385–397.Willis, K., Cottier, F., Kwasniewski, S., Wold, A., and Falk-Petersen, S. (2006).The influence of advection on zooplankton community composition in an arc-tic fjord (kongsfjorden, svalbard). Journal of Marine Systems, 61(1):39–54.Wolfe, M. A. (2010). Impact of wind and river flow on the timing of the riversinlet spring phytoplankton bloom. Master’s thesis, The University of BritishColumbia.Wootton, J. T., Pfister, C. A., and Forester, J. D. (2008). Dynamic patternsand ecological impacts of declining ocean ph in a high-resolution multi-yeardataset. Proceedings of the National Academy of Sciences, 105(48):18848–18853.Yamamoto-Kawai, M., McLaughlin, F. A., Carmack, E. C., Nishino, S., andShimada, K. (2009). Aragonite undersaturation in the arctic ocean: effects ofocean acidification and sea ice melt. Science, 326(5956):1098–1100.83Appendices84Appendix AChapter 2: SupplementaryDataA.1 2010–2011 Winter-Transition85A.1.2010–2011WINTER-TRANSITIONOctober 25, 2010 (150)Shell Diameter (mm)Probability Density0 1 2 3 4 5 60510152025November 8, 2010 (150)Shell Diameter (mm)Probability Density0 1 2 3 4 5 6010203040November 8, 2010 (250)Shell Diameter (mm)Probability Density0 1 2 3 4 5 60102030405060November 22, 2010 (250)Shell Diameter (mm)Probability Density0 1 2 3 4 5 605101520253035November 22, 2010 (150 & 250)Shell Diameter (mm)Probability Density0 1 2 3 4 5 605101520253035January 18, 2011 (150)Shell Diameter (mm)Probability Density0 1 2 3 4 5 6051015February 8, 2011 (150)Shell Diameter (mm)Probability Density0 1 2 3 4 5 6051015March 19, 2011 (150)Shell Diameter (mm)Probability Density0 1 2 3 4 5 60.00.51.01.52.02.53.0Figure A.1: A composite 2x4 set of histograms displaying the change in the population size-structure of L. helicina throughoutthe 2010–2011 winter-transition. Note the shift of the 0.2–0.4 mm size-bin to the 0.4–0.6 mm size-bin during the winter months.The mesh size of the net used for sampling is indicated in brackets for each sampling date. Due to precarious weather conditions,the samples from the 150 µm and 250 µm nets had to be combined on November 22, 2010.86A.2. SEASONAL CORRELATIONS - PHYSICAL PARAMETERS &POPULATION ABUNDANCEA.2 Seasonal Correlations - Physical Parame-ters & Population AbundanceTable A.1: Linear regressions testing the correlation between the environmentalparameters - 30 m depth-averaged temperature, 30 m depth-averaged salin-ity (PSU), 30 m depth-integrated fluorescence (m−2) - and L. helicina logged-abundance (ind.m−3) for the 2008 season (18 March to 22 September).Dependent variable:Log-Abundance (2008)(1) (2) (3)30 m Depth-Avg Temperature 1.543∗(0.756)30 m Depth-Avg Salinity −1.514∗(0.786)30 m Depth-Integrated Fluorescence −0.001(0.006)Constant −8.257 49.788∗ 4.338∗∗∗(6.117) (23.676) (1.161)Observations 10 10 10R2 0.342 0.317 0.003Adjusted R2 0.260 0.232 −0.122Residual Std. Error (df = 8) 1.520 1.549 1.872F statistic (df = 1; 8) 4.163∗ 3.712∗ 0.023Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.0187A.2. SEASONAL CORRELATIONS - PHYSICAL PARAMETERS &POPULATION ABUNDANCETable A.2: Linear regressions testing the correlation between the environmentalparameters - 30 m depth-averaged temperature, 30 m depth-averaged salin-ity (PSU), 30 m depth-integrated fluorescence (FU) - and L. helicina logged-abundance (ind.m−3) for the 2009 Season (28 February to 13 August).Dependent variable:Log-Abundance (2009)(1) (2) (3)30 m Depth-Avg Temperature 1.297∗∗∗(0.187)30 m Depth-Avg Salinity −1.284∗∗∗(0.266)30 m Depth-Integrated Fluorescence −0.001(0.006)Constant −7.517∗∗∗ 40.887∗∗∗ 2.376∗∗(1.422) (7.997) (0.827)Observations 11 11 11R2 0.843 0.721 0.001Adjusted R2 0.825 0.690 −0.110Residual Std. Error (df = 9) 0.548 0.729 1.381F statistic (df = 1; 9) 48.280∗∗∗ 23.304∗∗∗ 0.012Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.0188A.2. SEASONAL CORRELATIONS - PHYSICAL PARAMETERS &POPULATION ABUNDANCETable A.3: Linear regressions testing the correlation between the environmentalparameters - 30 m depth-averaged temperature, 30 m depth-averaged salin-ity (PSU), 30 m depth-integrated fluorescence (FU) - and L. helicina logged-abundance (ind.m−3) for the 2010 season (19 March to 20 July).Dependent variable:Log-Abundance (2010)(1) (2) (3)30 m Depth-Avg Temperature 2.535∗∗(0.361)30 m Depth-Avg Salinity −2.156(1.869)30 m Depth-Integrated Fluorescence −0.014(0.010)Constant −17.810∗∗ 67.332 7.285(3.043) (55.319) (2.625)Observations 4 4 4R2 0.961 0.400 0.529Adjusted R2 0.942 0.099 0.294Residual Std. Error (df = 2) 0.437 1.716 1.520F statistic (df = 1; 2) 49.390∗∗ 1.331 2.248Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Table A.4: Linear regressions testing the correlation between the environmen-tal parameters - 30 m depth-averaged temperature, 30 m depth-averaged salin-ity (PSU), 30 m depth-integrated fluorescence (FU) - and L. helicina logged-abundance (ind.m−3) for the 2009 Season (28 February to 13 August).Dependent variable:Log-Abundance (28 February - 2 June, 2009)30 m Depth-Integrated Fluorescence 0.007∗(0.003)Constant 0.571(0.516)Observations 7R2 0.491Adjusted R2 0.390Residual Std. Error 0.651(df = 5)F statistic 4.831∗(df = 1; 5)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.0189Appendix BChapter 3: SupplementaryDataB.1 Daily Data Sampling DatesTable B.1 presents the dates of zooplankton sample collection at the DailyStation located at Dawsons Landing, Rivers Inlet from 22 March to 7 July,2010. The statistcal results of all size-frequency enumerations of L. helicinaindividuals processed (max., mean, min.) for each sample, is also presented.Note the missing samples for 16, 23 April, 21 May, and 28 June.Table B.1: Daily dates of sample collection at Dawsons Daily Sta-tion. Also presented is the summary statistics of the shell sizestructure for each sample on their respective dates of sampling.An 80 µm mesh net (30 cm diameter) was used. Note the lackof summary statistics for the shell size enumerations, for dates ofmissing samples (indicated by blank lines).Year Month Day Mesh(µm)Depth(m)Max.(mm)Mean(mm)Min.(mm)2010 March 22 80 30 0.66 0.40 0.162010 March 23 80 30 0.66 0.34 0.222010 March 24 80 25 0.54 0.38 0.292010 March 25 80 28 0.24 0.21 0.162010 March 26 80 27 0.17 0.15 0.132010 March 27 80 29 0.37 0.21 0.132010 March 28 80 28 0.28 0.19 0.132010 March 29 80 29 0.37 0.22 0.13Continued on next page90B.1. DAILY DATA SAMPLING DATESTable B.1 – continued from previous pageYear Month Day Mesh(µm)Depth(m)Max.(mm)Mean(mm)Min.(mm)2010 March 30 80 28 0.39 0.22 0.142010 March 31 80 29 0.33 0.21 0.112010 April 1 80 29 0.34 0.19 0.132010 April 2 80 30 0.79 0.19 0.112010 April 3 80 29 1.35 0.22 0.132010 April 4 80 28 0.38 0.18 0.122010 April 5 80 30 0.36 0.21 0.152010 April 6 80 30 0.51 0.22 0.152010 April 7 80 30 0.35 0.22 0.132010 April 8 80 30 0.83 0.25 0.152010 April 9 80 30 0.63 0.22 0.132010 April 10 80 30 0.45 0.23 0.122010 April 11 80 30 0.42 0.21 0.122010 April 12 80 30 0.47 0.24 0.142010 April 13 80 30 0.44 0.23 0.152010 April 14 80 30 0.51 0.26 0.162010 April 15 80 30 0.48 0.27 0.142010 April 16 80 302010 April 17 80 30 0.51 0.24 0.132010 April 18 80 30 0.54 0.24 0.152010 April 19 80 30 0.48 0.28 0.152010 April 20 80 30 0.50 0.25 0.152010 April 21 80 30 0.54 0.30 0.132010 April 22 80 30 0.48 0.26 0.122010 April 23 80 302010 April 24 80 30 0.54 0.29 0.132010 April 25 80 30 0.63 0.30 0.132010 April 26 80 30 0.52 0.28 0.132010 April 27 80 30 0.54 0.30 0.132010 April 28 80 30 0.60 0.30 0.122010 April 29 80 30 0.71 0.27 0.112010 April 30 80 30 0.63 0.29 0.122010 May 1 80 30 0.71 0.28 0.122010 May 2 80 30 0.73 0.32 0.122010 May 3 80 30 0.87 0.28 0.112010 May 4 80 30 0.76 0.31 0.132010 May 5 80 30 0.85 0.34 0.122010 May 6 80 30 0.69 0.34 0.122010 May 7 80 30 1.17 0.35 0.112010 May 8 80 30 0.88 0.37 0.112010 May 9 80 30 0.73 0.31 0.122010 May 10 80 30 1.04 0.34 0.12Continued on next page91B.1. DAILY DATA SAMPLING DATESTable B.1 – continued from previous pageYear Month Day Mesh(µm)Depth(m)Max.(mm)Mean(mm)Min.(mm)2010 May 11 80 30 0.81 0.33 0.122010 May 12 80 30 0.69 0.31 0.112010 May 13 80 30 0.86 0.30 0.112010 May 14 80 30 0.97 0.32 0.112010 May 15 80 30 1.00 0.33 0.122010 May 16 80 30 0.77 0.32 0.122010 May 17 80 30 0.98 0.31 0.122010 May 18 80 30 0.94 0.36 0.112010 May 19 80 30 0.91 0.38 0.122010 May 20 80 30 1.17 0.38 0.122010 May 21 80 302010 May 22 80 30 0.97 0.32 0.112010 May 23 80 30 0.88 0.32 0.112010 May 24 80 30 0.95 0.32 0.112010 May 25 80 30 1.02 0.36 0.122010 May 26 80 30 1.91 0.38 0.112010 May 27 80 30 1.37 0.36 0.112010 May 28 80 30 1.15 0.34 0.112010 May 29 80 30 1.57 0.37 0.112010 May 30 80 30 2.01 0.37 0.112010 May 30 80 30 1.24 0.41 0.122010 June 1 80 30 1.32 0.36 0.112010 June 2 80 30 1.32 0.52 0.212010 June 3 80 30 1.11 0.38 0.122010 June 4 80 30 1.55 0.54 0.132010 June 5 80 30 1.70 0.56 0.112010 June 6 80 30 2.57 0.49 0.102010 June 7 80 30 1.57 0.43 0.112010 June 8 80 30 1.75 0.48 0.102010 June 9 80 30 1.40 0.43 0.122010 June 10 80 30 1.65 0.52 0.102010 June 11 80 30 1.42 0.40 0.112010 June 12 80 30 1.42 0.47 0.122010 June 13 80 30 1.17 0.37 0.102010 June 14 80 30 1.01 0.37 0.112010 June 15 80 30 1.51 0.37 0.102010 June 16 80 30 1.97 0.41 0.112010 June 17 80 30 1.65 0.45 0.102010 June 18 80 30 1.93 0.37 0.10Continued on next page92B.2. SIZE-FREQUENCY HISTOGRAMS - MARCH, APRIL, MAY, JUNE,JULYTable B.1 – continued from previous pageYear Month Day Mesh(µm)Depth(m)Max.(mm)Mean(mm)Min.(mm)2010 June 19 80 30 1.64 0.35 0.112010 June 20 80 30 3.07 0.47 0.102010 June 21 80 30 1.89 0.32 0.102010 June 22 80 30 1.52 0.29 0.112010 June 23 80 30 1.64 0.29 0.102010 June 24 80 30 1.87 0.29 0.102010 June 25 80 30 1.00 0.24 0.112010 June 26 80 30 1.22 0.25 0.102010 June 27 80 30 0.58 0.23 0.102010 June 28 802010 June 29 80 30 0.74 0.23 0.102010 June 30 80 30 0.73 0.24 0.102010 July 1 80 30 1.27 0.24 0.112010 July 2 80 30 0.70 0.23 0.112010 July 3 80 30 0.68 0.23 0.122010 July 4 80 30 0.76 0.23 0.112010 July 5 80 30 1.50 0.28 0.122010 July 6 80 30 0.87 0.23 0.122010 July 7 80 30 1.72 0.25 0.13B.2 Size-Frequency Histograms - March, April,May, June, JulyPresented are the size-frequency histograms of the population structure of L.helicina individuals processed for each day of sample collection, in Table B.1.The histograms are presented according to the month of observation and ar-ranged column-wise. Note the missing samples on 16, 23 April, 21 May, and28 June. For these dates, there is no histogram presented. The size-frequencyhistograms are binned into 0.02mm size-bins and the overall shape of the his-togram, for each day, will depend on the number of size-frequency enumerationsmade. Consequently, the histograms for dates of low population abundance willshow no definitive shape, whereas dates with high population abundances will.Finite mixture distributions were created and modified, to fit these histograms.93B.2. SIZE-FREQUENCY HISTOGRAMS - MARCH, APRIL, MAY, JUNE,JULYFigure B.1: Size-frequency histograms of the population size structure of L.helicina for samples collected from 22–31 March, 2010.94B.2.SIZE-FREQUENCYHISTOGRAMS-MARCH,APRIL,MAY,JUNE,JULYFigure B.2: Size-frequency histograms of the population size structure of L. helicina for samples collected from 1–15, 17–22,24–30 Apri, 2010. Note the missing samples for 16, 23 April.95B.2.SIZE-FREQUENCYHISTOGRAMS-MARCH,APRIL,MAY,JUNE,JULYFigure B.3: Size-frequency histograms of the population size structure of L. helicina for samples collected from 1–20, 22–31May, 2010. Note the missing sample for 21 May.96B.2.SIZE-FREQUENCYHISTOGRAMS-MARCH,APRIL,MAY,JUNE,JULYFigure B.4: Size-frequency histograms of the population size structure of L. helicina for samples collected from 1–27, 29–30June, 2010. Note the missing sample for 28 June.97B.3. FINITE MIXTURE DISTRIBUTIONS – DAILY DATAFigure B.5: Size-frequency histograms of the population size-structure of L.helicina for samples collected from 1–7 July, 2010.B.3 Finite Mixture Distributions – Daily DataB.4 What Are Finite Mixture DistributionsFinite mixture distributions were fitted onto the size frequency histograms (Ap-pendix B.2) using the mixdist package (MacDonald and contributions fromJuan Du, 2011) written for the R statistical programming environment (R CoreTeam, 2013). Utilizing maximum likelihood as well as the EM algorithm,mixdist fits finite mixture distributions onto grouped and/or conditional data.A “mixture distribution” results from a heterogeneous population of differ-98B.5. FITTING FINITE MIXTURESent age-classes where the overall distribution is composed of a finite numberof overlapping component distributions (normal, lognormal, gamma, exponen-tial, Weibull, binomial, negative binomial, Poisson distributions), with eachcomponent possessing component parameters (mixing proportions, means andstandard deviations of each component distribution) (MacDonald and contribu-tions from Juan Du, 2011). The primary function mix finds sets of overlappingcomponent distributions that gives the best fit to the grouped and conditionaldata. Each component possesses a different probability density, although thecomponent densities do not necessarily have to belong to the same parametricfamily (Du, 2002). Despite this non-requirement, it appears that the current ver-sion of mixdist assumes all component densities belong to the same parametricfamily (MacDonald and contributions from Juan Du, 2011), although dialoguewith Peter MacDonald disclosed future upgrades to the package (MacDonald,2011).B.5 Fitting Finite MixturesFitting finite mixture distributions begins with parameter estimation for eachcomponents of the mixture model (Du, 2002). When the data is complete (nomissing data), population components are easily identified and the proportionalabundance is estimated by counting the number of observations for each compo-nent Du (2002). For incomplete data, entire components and/or various portionsof component(s) are not observed, thus complicating the maximum likelihoodestimates (Du, 2002). Given this issue, there are cases of over-parameterization(Du, 2002) if constraints are not implemented (MacDonald and Pitcher, 1979).Bearing in mind that census data is almost always incomplete, given the restric-tions to sampling protocols, the estimation of all component parameters is oftennot possible, especially when components have high overlap (MacDonald andPitcher, 1979; Du, 2002). Thus, to avoid this problem it is best to reduce thenumber of parameters estimated (Du, 2002). Although this may not be reason-able given the observed data, Du (2002); MacDonald and Pitcher (1979) assumeconstraints (fixed proportions, means, and standard deviations) in order to re-duce the number of parameters estimated. Of course by assigning constraints,one introduces bias into the parameter estimates however, constraints may haveto be assigned for incomplete data, with judgement of constraints based on theobserved data (Du, 2002; MacDonald, 2011). Essentially, the passing of pa-rameter constraints is based on the expert knowledge of the user (MacDonald,2011).Concerning the census data collected for L. helicina, the identification oflikely components (or different age-classes) was based on the appearance ofdistinct modal peaks in the size-frequency histograms of the population size-structure. In cases when the size-frequency histograms lacked distinct modalpeaks (daily data), the previous date of observation was used as a reference tojudge the most likely course of development through time t. Since a compara-tively high resolution time-series was used for both Chapter 2 and Chapter 3,99B.5. FITTING FINITE MIXTURESthe appropriateness of applied constraints was validated over the entire timeseries.100B.5. FITTING FINITE MIXTURESFigure B.6: Finite mixture distributions fitted to the size-frequency histogramsfrom samples collected from 22–31 March, 2010.101B.5.FITTINGFINITEMIXTURESFigure B.7: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 1–30 April, 2010.Note the two missing samples on 16 and 23 April.102B.5.FITTINGFINITEMIXTURESFigure B.8: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 17–22 and 23–30April, 2010. Note the missing samples on 16 and 23 April.103B.5.FITTINGFINITEMIXTURESFigure B.9: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 1–15 May, 2010.104B.5.FITTINGFINITEMIXTURESFigure B.10: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 16–20, 22–31 May,2010. Note the missing sample for 21 May.105B.5.FITTINGFINITEMIXTURESFigure B.11: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 1–16 June, 2010.106B.5.FITTINGFINITEMIXTURESFigure B.12: Finite mixture distributions fitted to the size-frequency histograms of samples collected from 17–27, 29–30 June,2010. Note the missing sample for 28 June.107B.5. FITTING FINITE MIXTURESFigure B.13: Finite mixture distributions fitted to the size-frequency his-tograms of samples collected from 1–7 July, 2010.108B.6. FINITE MIXTURE DISTRIBUTIONS – STATISTICAL OUTPUTB.6 Finite Mixture Distributions – StatisticalOutputThe statistical output of the finite mixture distributions fitted, are provided inthe following table. Each finite mixture distribution contains the initial param-eter estimates for each component within the overall distribution (propotion,mean and standard deviation) as well as the overall distribution (also the samefor each component density), and any constraints applied to the mixture model.In the interest of space, only the ANOVA results and any constraints appliedto the mixture model are presented in Table B.2.For the majority of dates sampled, only the standard deviation was con-strained with SFX (Specific Sigmas Fixed, Du, 2002). On certain occassions, theconstraint MFX (Specific Means Fixed, Du, 2002) was implemented as there wereso few size-frequency measurements made that the data was too “incomplete”for accurate parameter estimates (MacDonald, 2011). In these situations, a com-bination of both MFX and SFX was used, after considering biological relevance(Du, 2002).109B.6.FINITEMIXTUREDISTRIBUTIONS–STATISTICALOUTPUTTable B.2: The ANOVA statistial results, signifiance, and constraints appliedto the finite mixture model fitted for each day of observation in the DawsonsDaily time-series. DF – degrees of freedom, χ2 – the statistical result returnedfrom the Maximum Likelihood Method, conpi – component proportions con-straints, conµ – component mean modal size constraints, conσ – componentstandard deviation constrains. NONE signifies no constraints,MFX – specificmeans constrained, SFX – specific sigmas constrained. Note the significancecodes: ’***’ – 0, ’**’ – 0.001, ’*’ – 0.01, ’.’ – 0.05. Note the missing dates for16, 23 April, 21 May, and 28 June, indicated by NA.Month Day ANOVATable:DF χ2 Pr(>χ2)Signif. Constraints:con-pi: con-µ con-σMarch 22 40 30.026 0.8745 NONE MFX SFXMarch 23 35 11.321 1 NONE NONE SFXMarch 24 15 6.1844 0.9765 NONE MFX NONEMarch 25 14 5.272 0.9817 NONE NONE NONEMarch 26 4 0.3486 0.9929 NONE NONE NONEMarch 27 21 15.621 0.7906 NONE MFX SFXMarch 28 9 7.6023 0.5747 NONE NONE SFXMarch 29 11 12.601 0.3202 NONE NONE NONEMarch 30 18 13.769 0.744 NONE NONE NONEMarch 31 14 2.9729 0.9991 NONE NONE NONEApril 1 12 13.326 0.3458 NONE NONE SFXApril 2 26 16.344 0.9276 NONE NONE SFXApril 3 19 12.076 0.8824 NONE NONE SFXApril 4 14 8.4987 0.8618 NONE NONE SFXApril 5 8 6.1058 0.6354 NONE NONE NONEContinued on next page110B.6.FINITEMIXTUREDISTRIBUTIONS–STATISTICALOUTPUTTable B.2 – continued from previous pageMonth Day ANOVATable:DF χ2 Pr(>χ2)Signif. Constraints:con-pi: con-µ con-σApril 6 18 19.539 0.3593 NONE NONE SFXApril 7 11 13.003 0.2932 NONe NONE SFXApril 8 32 11.236 0.9997 NONE NONE SFXApril 9 21 18.893 0.592 NONE NONE SFXApril 10 10 7.645 0.6635 NONE NONE SFXApril 11 14 26.171 0.02462 * NONE NONE SFXApril 12 16 8.6914 0.9256 NONE NONE SFXApril 13 12 7.3204 0.8357 NONE NONE NONEApril 14 19 16.572 0.6188 NONE NONE NONEApril 15 18 20.721 0.2937 NONE NONE NONEApril 16 NA NA NA NA NA NAApril 17 19 20.008 0.3941 NONE MFX NONEApril 18 17 12.766 0.7517 NONE NONE SFXApril 19 17 11.081 0.8523 NONE NONE SFXApril 20 12 17.474 0.1326 NONE NONE SFXApril 21 15 16.767 0.333 NONE NONE SFXApril 22 13 7.093 0.8681 NONE NONE SFXApril 23 NA NA NA NA NA NAApril 24 15 21.645 0.1175 NONE NONE SFXApril 25 23 47.005 0.002238 ** NONE NONE SFXApril 26 11 16.824 0.1132 NONE NONE SFXApril 27 16 30.009 0.01796 * NONE NONE SFXApril 28 20 27.138 0.1314 NONE NONE SFXApril 29 29 27.796 0.5289 NONE NONE SFXContinued on next page111B.6.FINITEMIXTUREDISTRIBUTIONS–STATISTICALOUTPUTTable B.2 – continued from previous pageMonth Day ANOVATable:DF χ2 Pr(>χ2)Signif. Constraints:con-pi: con-µ con-σApril 30 17 36.043 0.004527 ** NONE NONE SFXMay 1 23 35.812 0.04311 * NONE NONE SFXMay 2 24 423.776 0.0081 ** NONE NONE SFXMay 3 26 38.249 0.05742 . NONE NONE SFXMay 4 27 30.313 0.3003 NONE NONE SFXMay 5 31 29.401 0.5483 NONE NONE SFXMay 6 22 25.023 0.296 NONE NONE SFXMay 7 42 67.996 0.006758 ** NONE NONE SFXMay 8 30 38.875 0.1286 NONE NONE SFXMay 9 21 25.35 0.2323 NONE NONE NONEMay 10 31 46.208 0.03877 * NONE NONE SFXMay 11 24 52.986 0.000581 *** NONE NONE SFXMay 12 25 47.028 0.004866 ** NONE NONE SFXMay 13 36 52.796 0.03506 * NONE NONE SFXMay 14 34 65.755 0.0008729 *** NONE NONE SFXMay 15 25 39.565 0.03231 * NONE NONE SFXMay 16 31 61.193 0.0009742 *** NONE NONE SFXMay 17 31 56.748 0.003199 ** NONE NONE SFXMay 18 34 59.576 0.004301 ** NONE NONE SFXMay 19 35 63.031 0.002529 ** NONE NONE SFXMay 20 42 75.644 0.001115 ** NONE NONE SFXMay 21 NA NA NA NA NA NA NAMay 22 38 58.777 0.01685 * NONE NONE SFXMay 23 29 91.153 2.42E-08 *** NONE NONE SFXContinued on next page112B.6.FINITEMIXTUREDISTRIBUTIONS–STATISTICALOUTPUTTable B.2 – continued from previous pageMonth Day ANOVATable:DF χ2 Pr(>χ2)Signif. Constraints:con-pi: con-µ con-σMay 24 38 36.953 0.5177 NONE NONE SFXMay 25 40 64.854 0.007726 ** NONE NONE SFXMay 26 76 110.59 0.005896 ** NONE NONE SFXMay 27 54 84.783 0.004717 ** NONE NONE SFXMay 28 48 72.76 0.01207 * NONE NONE SFXMay 29 64 72.776 0.2115 NONE NONE SFXMay 30 57 90.605 0.003064 ** NONE NONE SFXMay 31 46 76.275 0.003319 ** NONE NONE SFXJune 1 49 59.154 0.1518 NONE NONE SFXJune 2 53 38.706 0.9294 NONE NONE SFXJune 3 40 66.041 0.005907 ** NONE NONE SFXJune 4 71 63.833 0.7145 NONE NONE SFXJune 5 79 37.324 1 NONE NONE SFXJune 6 87 58.145 0.9926 NONE NONE SFXJune 7 62 81.507 0.04903 * NONE NONE SFXJune 8 70 69.361 0.4991 NONE NONE SFXJune 9 59 81.794 0.02643 * NONE NONE SFXJune 10 64 57.755 0.6955 NONE NONE SFXJune 11 53 88.465 0.001615 ** NONE NONE SFXJune 12 56 59.808 0.3392 NONE NONE SFXJune 13 46 54.472 0.1833 NONE NONE SFXJune 14 38 58.126 0.01935 * NONE NONE SFXJune 15 58 98.97 0.0006467 *** NONE NONE SFXJune 16 80 45.058 0.9994 NONE NONE SFXContinued on next page113B.6.FINITEMIXTUREDISTRIBUTIONS–STATISTICALOUTPUTTable B.2 – continued from previous pageMonth Day ANOVATable:DF χ2 Pr(>χ2)Signif. Constraints:con-pi: con-µ con-σJune 17 60 73.784 0.1088 NONE NONE SFXJune 18 70 97.68 0.01611 * NONE NONE SFXJune 19 65 53.94 0.8344 NONE NONE SFXJune 20 97 103.49 0.3073 NONE NONE SFXJune 21 49 53.5 0.3056 NONE NONE SFXJune 22 66 62.327 0.6055 NONE NONE SFXJune 23 41 34.881 0.7383 NONE NONE SFXJune 24 62 43.058 0.968 NONE NONE SFXJune 25 37 62.632 0.005307 ** NONE NONE SFXJune 26 28 20.164 0.8584 NONE NONE SFXJune 27 17 30.885 0.02062 * NONE NONE SFXJune 28 NA NA NA NA NA NA NAJune 29 26 36.826 0.07749 . NONE NONE SFXJune 30 23 42.703 0.007509 ** NONE NONE SFXJuly 1 49 59.154 0.1518 NONE NONE SFXJuly 2 53 38.706 0.9294 NONE NONE SFXJuly 3 40 66.041 0.005907 ** NONE NONE SFXJuly 4 71 63.833 0.7145 NONE NONE SFXJuly 5 79 37.324 1 NONE NONE SFXJuly 6 87 58.145 0.9926 NONE NONE SFXJuly 7 62 81.507 0.04903 * NONE NONE SFX114B.7. LIFE TABLES FOR COHORTS TRACKEDB.7 Life Tables for Cohorts TrackedFrom the finite mixture distributions fitted (Appendix B.3), numerous popula-tion components were identified and tracked, from their likely date of recruit-ment into the population to the last date that they were observed within thesize-frequency histograms and subsequently assumed to have either died off ofpossibly merged together with another component. Each component is identi-fied with the “c” prefixed to the abbreviated date of the most likely date thecomponent was recruited into the population. For example “cApr01” signifiesthe component that was likely recruited into the population on 1 April.Because it was most probable that many components were possibly mergingtogether to grow as a single component for variable lengths of time only toseparate at a later time, the daily changes in shell size and abundance for thelarger size fractions were very unrealistic. Due to this, only the 0–0.5 mmsize-fraction was considered as the daily rates of shell-size growth was fairlyconsistent with the results from Chapter 2.115B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3: The seasonal development of the modal shell size (mm)for every recruit cohort tracked. Also included is the daily variationin abundance (Ind.m−3) of component as well as its daily estimateof its growth (mm) in shell size. Note that the dates of observationin theDates column is expressed as “mm/dd/yy”, and Size (mm)refers to the modal shell size for each component tracked. NA inthe Growth column signifies either no observations, or times of noperceived growth in shell size. NA in the Size and Abundancecolumns signifies no observations for those dates.Cohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr01 0 to 0.5mm 04/01/10 0.1441 6.8984 NAcApr01 0 to 0.5mm 04/02/10 0.1937 15.352 0.0496cApr01 0 to 0.5mm 04/03/10 0.1943 16.1977 6.00E-04cApr01 0 to 0.5mm 04/04/10 0.2348 7.147 0.0405cApr01 0 to 0.5mm 04/05/10 0.2283 24.5423 NAcApr01 0 to 0.5mm 04/06/10 0.2445 12.5892 0.0162cApr01 0 to 0.5mm 04/07/10 0.247 8.7086 0.0025cApr01 0 to 0.5mm 04/08/10 0.2456 5.1246 NAcApr01 0 to 0.5mm 04/09/10 0.2667 16.0081 0.0211cApr01 0 to 0.5mm 04/10/10 0.3003 17.9858 0.0336cApr01 0 to 0.5mm 04/11/10 0.3144 11.6811 0.0141cApr01 0 to 0.5mm 04/12/10 0.3837 5.123 0.0693cApr01 0 to 0.5mm 04/13/10 0.3414 11.2259 NAcApr01 0 to 0.5mm 04/14/10 0.4146 4.9498 0.0732cApr01 0 to 0.5mm 04/15/10 0.3878 6.9603 NAContinued on next page116B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr01 0 to 0.5mm 04/16/10 NA NA NAcApr01 0 to 0.5mm 04/17/10 0.43 7.7563 0.0211cApr01 0 to 0.5mm 04/18/10 0.4599 3.4479 0.0299cApr01 0 to 0.5mm 04/19/10 0.385 5.6777 NAcApr01 0 to 0.5mm 04/20/10 0.4287 4.2813 0.0437cApr01 0 to 0.5mm 04/21/10 0.4396 9.1887 0.0109cApr01 0 to 0.5mm 04/22/10 0.4363 6.7107 NAcApr01 0 to 0.5mm 04/23/10 NA NA NAcApr01 0 to 0.5mm 04/24/10 0.4588 11.3875 NAcApr02 0 to 0.5mm 04/02/10 0.1453 4.4537 NAcApr02 0 to 0.5mm 04/03/10 0.1538 6.2291 0.0085cApr02 0 to 0.5mm 04/04/10 0.1722 35.7462 0.0184cApr02 0 to 0.5mm 04/05/10 0.1775 21.412 0.0053cApr02 0 to 0.5mm 04/06/10 0.1861 15.2353 0.0086cApr02 0 to 0.5mm 04/07/10 0.1796 11.0007 NAcApr02 0 to 0.5mm 04/08/10 0.1788 6.1829 NAcApr02 0 to 0.5mm 04/09/10 0.2003 33.0411 0.0215cApr02 0 to 0.5mm 04/10/10 0.215 48.5465 0.0147cApr02 0 to 0.5mm 04/11/10 0.231 14.3427 0.016cApr02 0 to 0.5mm 04/12/10 0.2587 11.7324 0.0277cApr02 0 to 0.5mm 04/13/10 0.2431 14.6235 NAcApr02 0 to 0.5mm 04/14/10 0.2736 11.0428 0.0305cApr02 0 to 0.5mm 04/15/10 0.2934 8.6546 0.0198cApr02 0 to 0.5mm 04/16/10 NA NA NAcApr02 0 to 0.5mm 04/17/10 0.34 14.6416 0.0233Continued on next page117B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr02 0 to 0.5mm 04/18/10 0.2987 24.2295 NAcApr02 0 to 0.5mm 04/19/10 0.3359 10.1752 0.0372cApr02 0 to 0.5mm 04/20/10 0.3199 22.573 NAcApr02 0 to 0.5mm 04/21/10 0.3562 20.3883 0.0363cApr02 0 to 0.5mm 04/22/10 0.3628 23.2981 0.0066cApr02 0 to 0.5mm 04/23/10 NA NA NAcApr02 0 to 0.5mm 04/24/10 0.3449 16.6977 NAcApr02 0 to 0.5mm 04/25/10 0.4343 17.2336 0.0894cApr02 0 to 0.5mm 04/26/10 0.439 7.7459 0.0047cApr02 0 to 0.5mm 04/27/10 0.398 14.4788 NAcApr02 0 to 0.5mm 04/28/10 0.4307 18.3966 0.0327cApr02 0 to 0.5mm 04/29/10 0.4507 27.619 0.02cApr02 0 to 0.5mm 04/30/10 0.4474 26.6309 NAcApr02 0 to 0.5mm 05/01/10 0.4715 44.7813 0.0241cApr02 0 to 0.5mm 05/02/10 0.4599 29.6462 NAcApr02 0 to 0.5mm 05/03/10 0.468 49.373 0.0081cApr02 0 to 0.5mm 05/04/10 0.4777 55.3083 0.0097cApr02 0 to 0.5mm 05/05/10 NA NA NAcApr02 0 to 0.5mm 05/06/10 0.4992 64.7789 0.0107cApr09 0 to 0.5mm 04/09/10 0.153 17.4426 NAcApr09 0 to 0.5mm 04/10/10 0.1689 28.2732 0.0159cApr09 0 to 0.5mm 04/11/10 0.1726 39.0586 0.0037cApr09 0 to 0.5mm 04/12/10 0.1885 20.8694 0.0159cApr09 0 to 0.5mm 04/13/10 0.1835 25.5514 NAcApr09 0 to 0.5mm 04/14/10 0.1823 13.2443 NAContinued on next page118B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr09 0 to 0.5mm 04/15/10 0.2091 18.338 0.0268cApr09 0 to 0.5mm 04/16/10 NA NA NAcApr09 0 to 0.5mm 04/17/10 0.225 87.868 0.0079cApr09 0 to 0.5mm 04/18/10 0.2497 40.7468 0.0247cApr09 0 to 0.5mm 04/19/10 0.2762 13.2736 0.0265cApr09 0 to 0.5mm 04/20/10 0.2616 19.1604 NAcApr09 0 to 0.5mm 04/21/10 0.2691 17.3462 NAcApr09 0 to 0.5mm 04/22/10 0.2746 19.3442 0.0145cApr09 0 to 0.5mm 04/23/10 NA NA NAcApr09 0 to 0.5mm 04/24/10 0.261 24.8679 NAcApr09 0 to 0.5mm 04/25/10 0.3403 25.0483 0.0793cApr09 0 to 0.5mm 04/26/10 0.3511 27.7882 0.0108cApr09 0 to 0.5mm 04/27/10 0.332 23.6773 NAcApr09 0 to 0.5mm 04/28/10 0.3331 62.3996 0.0011cApr17 0 to 0.5mm 04/17/10 0.15 17.897 NAcApr17 0 to 0.5mm 04/18/10 0.1836 53.2406 0.0336cApr17 0 to 0.5mm 04/19/10 0.1952 11.4283 0.0116cApr17 0 to 0.5mm 04/20/10 0.1914 38.396 NAcApr17 0 to 0.5mm 04/21/10 0.1927 17.7527 0.0013cApr17 0 to 0.5mm 04/22/10 0.2062 22.701 0.0135cApr17 0 to 0.5mm 04/23/10 NA NA NAcApr17 0 to 0.5mm 04/24/10 0.1956 20.048 NAcApr17 0 to 0.5mm 04/25/10 0.2492 21.9258 0.0536cApr17 0 to 0.5mm 04/26/10 0.2542 24.0219 0.005cApr17 0 to 0.5mm 04/27/10 0.2551 24.9448 9.00E-04Continued on next page119B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr17 0 to 0.5mm 04/28/10 0.3331 62.3996 0.078cApr17 0 to 0.5mm 04/29/10 0.3443 45.7486 0.0112cApr17 0 to 0.5mm 04/30/10 0.3538 46.1697 0.0095cApr17 0 to 0.5mm 05/01/10 0.3431 87.4172 NAcApr17 0 to 0.5mm 05/02/10 0.3512 31.6764 0.0081cApr17 0 to 0.5mm 05/03/10 0.3616 95.8574 0.0104cApr17 0 to 0.5mm 05/04/10 0.3996 63.5719 0.0484cApr17 0 to 0.5mm 05/05/10 0.406 148.979 0.0064cApr17 0 to 0.5mm 05/06/10 0.4318 79.0986 0.0258cApr17 0 to 0.5mm 05/07/10 0.4821 193.8972 0.0503cApr22 0 to 0.5mm 04/22/10 0.1614 22.7294 NAcApr22 0 to 0.5mm 04/23/10 NA NA NAcApr22 0 to 0.5mm 04/24/10 0.1491 0.0973 NAcApr22 0 to 0.5mm 04/25/10 0.1887 22.0466 0.0396cApr22 0 to 0.5mm 04/26/10 0.1938 28.6268 0.0041cApr22 0 to 0.5mm 04/27/10 0.1909 25.1008 NAcApr22 0 to 0.5mm 04/28/10 0.2352 49.9609 0.0414cApr22 0 to 0.5mm 04/29/10 0.2637 67.158 0.0285cApr22 0 to 0.5mm 04/30/10 0.2798 52.0172 0.0161cApr22 0 to 0.5mm 05/01/10 0.265 70.6579 NAcApr22 0 to 0.5mm 05/02/10 0.2583 25.5809 NAcApr22 0 to 0.5mm 05/03/10 0.2557 45.8706 NAcApr22 0 to 0.5mm 05/04/10 0.2628 67.3435 0.0071cApr22 0 to 0.5mm 05/05/10 0.2364 75.547 NAcApr22 0 to 0.5mm 05/06/10 0.2547 87.6874 0.0183Continued on next page120B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr22 0 to 0.5mm 05/07/10 0.2612 164.6008 0.0065cApr22 0 to 0.5mm 05/08/10 0.3385 58.2074 0.0773cApr22 0 to 0.5mm 05/09/10 0.3374 144.2505 NAcApr22 0 to 0.5mm 05/10/10 0.3933 167.3239 0.0559cApr22 0 to 0.5mm 05/11/10 0.436 120.6809 0.0427cApr22 0 to 0.5mm 05/12/10 0.4809 225.2285 0.0876cApr28 0 to 0.5mm 04/28/10 0.1602 18.6459 NAcApr28 0 to 0.5mm 04/29/10 0.1987 45.1226 0.0385cApr28 0 to 0.5mm 04/30/10 0.2084 52.2286 0.0097cApr28 0 to 0.5mm 05/01/10 0.1954 47.3499 NAcApr28 0 to 0.5mm 05/02/10 0.199 6.6321 0.0036cApr28 0 to 0.5mm 05/03/10 0.2557 45.8706 0.0567cApr28 0 to 0.5mm 05/04/10 0.2628 67.3435 0.0071cApr28 0 to 0.5mm 05/05/10 0.3141 122.7697 0.0513cApr28 0 to 0.5mm 05/06/10 0.3393 79.3091 0.0252cApr28 0 to 0.5mm 05/07/10 0.3565 245.8426 0.0172cApr28 0 to 0.5mm 05/08/10 0.429 110.8607 0.0725cApr28 0 to 0.5mm 05/09/10 0.4211 269.7289 NAcApr28 0 to 0.5mm 05/10/10 0.4842 186.4933 NAcApr28 0 to 0.5mm 05/11/10 NA NA NAcApr28 0 to 0.5mm 05/12/10 0.4809 225.2285 NAcApr29 0 to 0.5mm 04/29/10 0.1549 56.6661 NAcApr29 0 to 0.5mm 04/30/10 0.1525 41.7312 NAcApr29 0 to 0.5mm 05/01/10 0.1602 114.2373 0.0077cApr29 0 to 0.5mm 05/02/10 0.1581 21.4198 NAContinued on next page121B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cApr29 0 to 0.5mm 05/03/10 0.1688 143.9464 0.0107cApr29 0 to 0.5mm 05/04/10 0.1866 109.9226 0.0178cApr29 0 to 0.5mm 05/05/10 0.1795 186.6858 NAcApr29 0 to 0.5mm 05/06/10 0.175 143.0933 NAcApr29 0 to 0.5mm 05/07/10 0.2009 123.8884 0.0259cApr29 0 to 0.5mm 05/08/10 0.229 100.1688 0.0281cApr29 0 to 0.5mm 05/09/10 0.2751 233.9228 0.0461cApr29 0 to 0.5mm 05/10/10 0.3089 77.6639 0.0338cApr29 0 to 0.5mm 05/11/10 0.3745 143.5213 0.0656cApr29 0 to 0.5mm 05/12/10 0.3761 101.2723 0.0016cApr29 0 to 0.5mm 05/13/10 0.4793 132.1383 0.1032cApr29 0 to 0.5mm 05/14/10 0.4841 96.305 0.0048cApr29 0 to 0.5mm 05/15/10 0.4744 142.0013 NAcApr29 0 to 0.5mm 05/16/10 0.484 220.9958 0.0096cApr29 0 to 0.5mm 05/17/10 0.4642 131.2378 NAcMay07 0 to 0.5mm 05/07/10 0.1581 134.8172 NAcMay07 0 to 0.5mm 05/08/10 0.1683 67.6996 0.0102cMay07 0 to 0.5mm 05/09/10 0.1832 412.1687 0.0149cMay07 0 to 0.5mm 05/10/10 0.2533 189.1069 0.0701cMay07 0 to 0.5mm 05/11/10 0.2864 269.6928 0.0331cMay07 0 to 0.5mm 05/12/10 0.2833 231.948 NAcMay07 0 to 0.5mm 05/13/10 0.3557 246.3021 0.0724cMay07 0 to 0.5mm 05/14/10 0.3802 33.3594 0.0245cMay07 0 to 0.5mm 05/15/10 0.3858 73.2094 0.0056cMay07 0 to 0.5mm 05/16/10 0.3719 171.9983 NAContinued on next page122B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay07 0 to 0.5mm 05/17/10 0.3521 182.5217 NAcMay07 0 to 0.5mm 05/18/10 0.3618 1009.6116 0.0097cMay07 0 to 0.5mm 05/19/10 0.3552 457.0698 NAcMay07 0 to 0.5mm 05/20/10 0.377 888.7034 0.0218cMay07 0 to 0.5mm 05/21/10 NA NA NAcMay07 0 to 0.5mm 05/22/10 0.3927 53.9599 0.0078cMay07 0 to 0.5mm 05/23/10 0.4289 235.6525 0.0362cMay10 0 to 0.5mm 05/10/10 0.1703 192.8726 NAcMay10 0 to 0.5mm 05/11/10 0.2236 257.1379 0.0533cMay10 0 to 0.5mm 05/12/10 0.2069 217.5518 NAcMay10 0 to 0.5mm 05/13/10 0.2464 164.7412 0.0395cMay10 0 to 0.5mm 05/14/10 0.2746 81.4858 0.0282cMay10 0 to 0.5mm 05/15/10 0.288 186.5169 0.0134cMay10 0 to 0.5mm 05/16/10 0.2921 260.7507 0.0044cMay10 0 to 0.5mm 05/17/10 0.2846 230.6837 NAcMay10 0 to 0.5mm 05/18/10 0.3618 1009.6116 0.0772cMay10 0 to 0.5mm 05/19/10 0.3552 457.0698 NAcMay10 0 to 0.5mm 05/20/10 0.377 888.7034 0.0218cMay10 0 to 0.5mm 05/21/10 NA NA NAcMay10 0 to 0.5mm 05/22/10 0.3927 218.0299 0.0078cMay10 0 to 0.5mm 05/23/10 0.4289 235.6525 0.0362cMay11 0 to 0.5mm 05/11/10 0.1576 537.9561 NAcMay11 0 to 0.5mm 05/12/10 0.1531 206.6579 NAcMay11 0 to 0.5mm 05/13/10 0.1665 335.2995 0.0134cMay11 0 to 0.5mm 05/14/10 0.17 176.7625 0.0035Continued on next page123B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay11 0 to 0.5mm 05/15/10 0.1884 336.9296 0.0184cMay11 0 to 0.5mm 05/16/10 0.2042 296.2878 0.0158cMay11 0 to 0.5mm 05/17/10 0.2031 225.7144 NAcMay11 0 to 0.5mm 05/18/10 0.2576 898.8277 0.0545cMay11 0 to 0.5mm 05/19/10 0.2739 454.4722 0.0163cMay11 0 to 0.5mm 05/20/10 0.2477 1572.5983 NAcMay11 0 to 0.5mm 05/21/10 NA NA NAcMay11 0 to 0.5mm 05/22/10 0.277 244.9266 0.0146cMay11 0 to 0.5mm 05/23/10 0.3158 206.4234 0.0388cMay11 0 to 0.5mm 05/24/10 0.3231 469.2192 0.0073cMay11 0 to 0.5mm 05/25/10 0.3324 128.604 0.0093cMay11 0 to 0.5mm 05/26/10 0.3467 274.956 0.0143cMay11 0 to 0.5mm 05/27/10 0.4241 82.371 0.0774cMay11 0 to 0.5mm 05/28/10 0.4376 144.2691 0.0135cMay11 0 to 0.5mm 05/29/10 0.4492 102.9545 0.0116cMay11 0 to 0.5mm 05/30/10 0.4689 220.2494 0.0197cMay16 0 to 0.5mm 05/16/10 0.1488 83.1853 NAcMay16 0 to 0.5mm 05/17/10 0.1526 129.6669 0.0038cMay16 0 to 0.5mm 05/18/10 0.1817 445.1131 0.0291cMay16 0 to 0.5mm 05/19/10 0.192 562.0603 0.0103cMay16 0 to 0.5mm 05/20/10 0.1604 465.3858 NAcMay16 0 to 0.5mm 05/21/10 NA NA NAcMay16 0 to 0.5mm 05/22/10 0.1876 561.6681 0.0136cMay16 0 to 0.5mm 05/23/10 0.2233 196.1085 0.0357cMay16 0 to 0.5mm 05/24/10 0.2279 286.1311 0.0046Continued on next page124B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay16 0 to 0.5mm 05/25/10 0.2384 141.6408 0.0105cMay16 0 to 0.5mm 05/26/10 0.2367 324.4567 NAcMay16 0 to 0.5mm 05/27/10 0.326 124.0343 0.0893cMay16 0 to 0.5mm 05/28/10 0.3248 306.4684 NAcMay16 0 to 0.5mm 05/29/10 0.3648 129.8016 0.04cMay16 0 to 0.5mm 05/30/10 0.3939 321.1713 0.0291cMay16 0 to 0.5mm 05/31/10 0.4439 66.7577 0.05cMay16 0 to 0.5mm 06/01/10 0.4241 353.5336 NAcMay16 0 to 0.5mm 06/02/10 0.4038 10.5602 NAcMay16 0 to 0.5mm 06/03/10 0.4889 19.9923 0.0851cMay20 0 to 0.5mm 05/20/10 0.1604 465.3858 NAcMay20 0 to 0.5mm 05/21/10 NA NA NAcMay20 0 to 0.5mm 05/22/10 0.1876 561.0681 0.0136cMay20 0 to 0.5mm 05/23/10 0.2233 196.1085 0.0357cMay20 0 to 0.5mm 05/24/10 0.2279 286.1311 0.0046cMay20 0 to 0.5mm 05/25/10 0.2384 141.6408 0.0105cMay20 0 to 0.5mm 05/26/10 0.2367 324.4567 NAcMay20 0 to 0.5mm 05/27/10 0.326 124.0343 0.0893cMay20 0 to 0.5mm 05/28/10 0.3248 306.4684 NAcMay20 0 to 0.5mm 05/29/10 0.3648 129.8016 0.04cMay20 0 to 0.5mm 05/30/10 0.3939 321.1713 0.0291cMay20 0 to 0.5mm 05/31/10 0.4439 66.7577 0.05cMay20 0 to 0.5mm 06/01/10 0.4241 353.5336 NAcMay20 0 to 0.5mm 06/02/10 0.4038 10.5602 NAContinued on next page125B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay20 0 to 0.5mm 06/03/10 0.4889 19.9923 0.0851cMay23 0 to 0.5mm 05/23/10 0.1484 254.3484 NAcMay23 0 to 0.5mm 05/24/10 0.1512 190.3226 0.0028cMay23 0 to 0.5mm 05/25/10 0.1668 162.6051 0.0156cMay23 0 to 0.5mm 05/26/10 0.1596 93.8876 NAcMay23 0 to 0.5mm 05/27/10 0.2366 137.9554 0.077cMay23 0 to 0.5mm 05/28/10 0.2293 161.7999 NAcMay23 0 to 0.5mm 05/29/10 0.2851 153.6297 0.0558cMay23 0 to 0.5mm 05/30/10 0.3203 288.8206 0.0352cMay23 0 to 0.5mm 05/31/10 0.3429 44.9819 0.0226cMay23 0 to 0.5mm 06/01/10 0.3438 285.5686 9.00E-04cMay23 0 to 0.5mm 06/02/10 0.3307 3.6816 NAcMay23 0 to 0.5mm 06/03/10 0.4196 78.531 0.0898cMay23 0 to 0.5mm 06/04/10 NA NA NAcMay23 0 to 0.5mm 06/05/10 0.4729 420.8423 0.0266cMay23 0 to 0.5mm 06/06/10 NA NA NAcMay23 0 to 0.5mm 06/07/10 0.4906 251.0487 0.0088cMay23 0 to 0.5mm 06/08/10 0.4716 99.9266 NAcMay23 0 to 0.5mm 06/09/10 0.4675 235.1282 NAcMay27 0 to 0.5mm 05/27/10 0.1463 129.883 NAcMay27 0 to 0.5mm 05/28/10 0.1449 159.1271 NAcMay27 0 to 0.5mm 05/29/10 0.1595 142.6032 0.0146cMay27 0 to 0.5mm 05/30/10 0.2195 169.3804 0.06cMay27 0 to 0.5mm 05/31/10 0.2314 64.8604 0.0119cMay27 0 to 0.5mm 06/01/10 0.2679 99.7786 0.0365Continued on next page126B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay27 0 to 0.5mm 06/02/10 0.2831 1.577 0.0152cMay27 0 to 0.5mm 06/03/10 0.3252 54.2273 0.0421cMay27 0 to 0.5mm 06/04/10 0.3718 40.1178 0.0466cMay27 0 to 0.5mm 06/05/10 0.4 175.9306 0.0282cMay27 0 to 0.5mm 06/06/10 0.4129 123.1989 0.0129cMay27 0 to 0.5mm 06/07/10 0.3632 165.3025 NAcMay27 0 to 0.5mm 06/08/10 0.3672 48.7603 0.004cMay27 0 to 0.5mm 06/09/10 0.3581 142.5525 NAcMay27 0 to 0.5mm 06/10/10 0.4996 150.7339 0.1415cMay27 0 to 0.5mm 06/11/10 0.474 95.7034 NAcMay27 0 to 0.5mm 06/12/10 0.4905 148.5535 0.0165cMay30 0 to 0.5mm 05/30/10 0.1516 221.8395 NAcMay30 0 to 0.5mm 05/31/10 0.1566 31.6916 0.005cMay30 0 to 0.5mm 06/01/10 0.2081 129.5009 0.0515cMay30 0 to 0.5mm 06/02/10 0.2183 1.1427 0.0102cMay30 0 to 0.5mm 06/03/10 0.2178 78.4112 NAcMay30 0 to 0.5mm 06/04/10 0.2693 33.2847 0.0515cMay30 0 to 0.5mm 06/05/10 0.3211 125.4321 0.0518cMay30 0 to 0.5mm 06/06/10 0.2945 128.7586 NAcMay30 0 to 0.5mm 06/07/10 0.2689 78.0765 NAcMay30 0 to 0.5mm 06/08/10 0.2871 41.3156 0.0182cMay30 0 to 0.5mm 06/09/10 0.2789 111.5163 NAcMay30 0 to 0.5mm 06/10/10 0.3673 114.1064 0.0884cMay30 0 to 0.5mm 06/11/10 0.3885 67.4515 0.0212cMay30 0 to 0.5mm 06/12/10 0.3764 32.5118 NAContinued on next page127B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cMay30 0 to 0.5mm 06/13/10 0.4327 163.9304 0.0563cMay30 0 to 0.5mm 06/14/10 0.4621 206.9976 0.0294cJun01 0 to 0.5mm 06/01/10 0.1491 171.8846 NAcJun01 0 to 0.5mm 06/02/10 0.2183 1.1427 0.0692cJun01 0 to 0.5mm 06/03/10 0.2178 78.4112 NAcJun01 0 to 0.5mm 06/04/10 0.2693 33.2847 0.0515cJun01 0 to 0.5mm 06/05/10 0.3211 125.4321 0.0518cJun01 0 to 0.5mm 06/06/10 0.2945 128.7586 NAcJun01 0 to 0.5mm 06/07/10 0.2689 78.0765 NAcJun01 0 to 0.5mm 06/08/10 0.2871 41.3156 0.0182cJun01 0 to 0.5mm 06/09/10 0.2789 111.5163 NAcJun01 0 to 0.5mm 06/10/10 0.3673 114.1064 0.0884cJun01 0 to 0.5mm 06/11/10 0.3885 67.4515 0.0212cJun01 0 to 0.5mm 06/12/10 0.3764 32.5118 NAcJun01 0 to 0.5mm 06/13/10 0.4327 163.9304 0.0563cJun01 0 to 0.5mm 06/14/10 0.4621 206.9976 0.0294cJun03 0 to 0.5mm 06/03/10 0.1534 37.5376 NAcJun03 0 to 0.5mm 06/04/10 0.184 14.901 0.0306cJun03 0 to 0.5mm 06/05/10 0.2292 88.2037 0.0452cJun03 0 to 0.5mm 06/06/10 0.1963 160.3634 NAcJun03 0 to 0.5mm 06/07/10 0.2063 41.7804 0.01cJun03 0 to 0.5mm 06/08/10 0.2871 41.3156 0.0808cJun03 0 to 0.5mm 06/09/10 0.2789 111.5164 NAcJun03 0 to 0.5mm 06/10/10 0.3673 114.1064 0.0884cJun03 0 to 0.5mm 06/11/10 0.3885 67.4515 0.0212Continued on next page128B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun03 0 to 0.5mm 06/12/10 0.3764 32.5118 NAcJun03 0 to 0.5mm 06/13/10 0.4327 163.9304 0.0563cJun03 0 to 0.5mm 06/14/10 0.4621 206.9976 0.0294cJun05 0 to 0.5mm 06/05/10 0.1467 361.0203 NAcJun05 0 to 0.5mm 06/06/10 0.1351 167.1623 NAcJun05 0 to 0.5mm 06/07/10 0.1457 121.1904 0.0106cJun05 0 to 0.5mm 06/08/10 0.1907 67.9686 0.045cJun05 0 to 0.5mm 06/09/10 0.2073 87.3148 0.0166cJun05 0 to 0.5mm 06/10/10 0.2846 69.8635 0.0773cJun05 0 to 0.5mm 06/11/10 0.2837 69.0567 NAcJun05 0 to 0.5mm 06/12/10 0.301 39.1541 0.0773cJun05 0 to 0.5mm 06/13/10 0.2991 103.0388 NAcJun05 0 to 0.5mm 06/14/10 0.3567 169.2531 0.0576cJun05 0 to 0.5mm 06/15/10 0.3851 621.0002 0.0284cJun05 0 to 0.5mm 06/16/10 0.4659 373.337 0.0808cJun08 0 to 0.5mm 06/08/10 0.1307 344.024 NAcJun08 0 to 0.5mm 06/09/10 0.1533 192.1387 0.0226cJun08 0 to 0.5mm 06/10/10 0.1982 111.8403 0.0449cJun08 0 to 0.5mm 06/11/10 0.2021 97.421 0.0039cJun08 0 to 0.5mm 06/12/10 0.2455 78.0197 0.0434cJun08 0 to 0.5mm 06/13/10 0.2249 138.2639 NAcJun08 0 to 0.5mm 06/14/10 0.253 212.8395 0.0281cJun08 0 to 0.5mm 06/15/10 0.3172 356.3698 0.0642cJun08 0 to 0.5mm 06/16/10 0.4183 256.232 0.1011cJun08 0 to 0.5mm 06/17/10 0.4138 302.3025 NAContinued on next page129B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun08 0 to 0.5mm 06/18/10 0.4152 268.6145 0.0014cJun08 0 to 0.5mm 06/19/10 0.4671 203.8399 0.0519cJun08 0 to 0.5mm 06/20/10 0.4427 239.6861 NAcJun08 0 to 0.5mm 06/21/10 0.4533 148.3587 0.0106cJun08 0 to 0.5mm 06/22/10 0.4868 91.3773 0.0335cJun08 0 to 0.5mm 06/23/10 0.4725 157.8899 NAcJun08 0 to 0.5mm 06/24/10 0.4995 834.5411 0.027cJun08 0 to 0.5mm 06/25/10 0.4512 171.3592 NAcJun08 0 to 0.5mm 06/26/10 0.4773 225.8542 0.0261cJun08 0 to 0.5mm 06/27/10 0.4653 147.2581 NAcJun10 0 to 0.5mm 06/10/10 0.1377 275.5233 NAcJun10 0 to 0.5mm 06/11/10 0.148 193.2448 0.0103cJun10 0 to 0.5mm 06/12/10 0.1632 220.6136 0.0152cJun10 0 to 0.5mm 06/13/10 0.1501 172.2454 NAcJun10 0 to 0.5mm 06/14/10 0.172 234.0911 0.0219cJun10 0 to 0.5mm 06/15/10 0.2333 711.944 0.0613cJun10 0 to 0.5mm 06/16/10 0.2707 305.8637 0.0374cJun10 0 to 0.5mm 06/17/10 0.3189 133.1588 0.0482cJun10 0 to 0.5mm 06/18/10 0.3159 242.3292 NAcJun10 0 to 0.5mm 06/19/10 0.3788 328.2438 0.0629cJun10 0 to 0.5mm 06/20/10 0.3784 43.4878 NAcJun10 0 to 0.5mm 06/21/10 0.377 203.5538 NAcJun10 0 to 0.5mm 06/22/10 0.3379 93.5547 NAcJun10 0 to 0.5mm 06/23/10 0.337 308.74234 NAcJun10 0 to 0.5mm 06/24/10 0.3493 303.3966 0.0123Continued on next page130B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun10 0 to 0.5mm 06/25/10 0.3868 90.7059 0.0375cJun10 0 to 0.5mm 06/26/10 0.391 610.2823 0.0042cJun10 0 to 0.5mm 06/27/10 0.3793 476.7445 NAcJun10 0 to 0.5mm 06/28/10 NA NA NAcJun10 0 to 0.5mm 06/29/10 0.4053 268.4662 0.013cJun10 0 to 0.5mm 06/30/10 0.4212 290.4696 0.0159cJun10 0 to 0.5mm 07/01/10 NA 206.3434 0.084cJun10 0 to 0.5mm 07/02/10 0.4713 78.4593 NAcJun15 0 to 0.5mm 06/15/10 0.1394 502.6999 NAcJun15 0 to 0.5mm 06/16/10 0.1396 195.1853 2.00E-04cJun15 0 to 0.5mm 06/17/10 0.1348 568.6407 NAcJun15 0 to 0.5mm 06/18/10 0.194 190.4142 0.0592cJun15 0 to 0.5mm 06/19/10 0.2202 314.5058 0.0262cJun15 0 to 0.5mm 06/20/10 0.2222 75.9381 0.002cJun15 0 to 0.5mm 06/21/10 0.2482 117.4577 0.026cJun15 0 to 0.5mm 06/22/10 0.2626 116.6371 0.0404cJun15 0 to 0.5mm 06/23/10 0.2568 413.5747 NAcJun15 0 to 0.5mm 06/24/10 0.2651 396.3535 0.0083cJun15 0 to 0.5mm 06/25/10 0.3225 83.0074 0.0574cJun15 0 to 0.5mm 06/26/10 0.3161 440.6433 NAcJun15 0 to 0.5mm 06/27/10 0.3259 182.9022 0.0098cJun15 0 to 0.5mm 06/28/10 NA NA NAcJun15 0 to 0.5mm 06/29/10 0.3265 737.951 3.00E-04cJun15 0 to 0.5mm 06/30/10 0.336 731.9183 0.0095cJun15 0 to 0.5mm 07/01/10 0.3813 241.2633 0.0453Continued on next page131B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun15 0 to 0.5mm 07/02/10 0.3872 126.8319 0.0059cJun15 0 to 0.5mm 07/03/10 0.399 64.3687 0.0118cJun15 0 to 0.5mm 07/04/10 0.4502 62.2922 0.0512cJun18 0 to 0.5mm 06/18/10 0.1347 674.074 NAcJun18 0 to 0.5mm 06/19/10 0.1386 483.0263 0.0039cJun18 0 to 0.5mm 06/20/10 0.1401 232.5862 0.0015cJun18 0 to 0.5mm 06/21/10 0.1408 469.3132 7.00E-04cJun18 0 to 0.5mm 06/22/10 0.1429 363.871 0.0021cJun18 0 to 0.5mm 06/23/10 0.1897 279.336 0.0468cJun18 0 to 0.5mm 06/24/10 0.2051 245.4146 0.0154cJun18 0 to 0.5mm 06/25/10 0.2566 190.15 0.0515cJun18 0 to 0.5mm 06/26/10 0.2578 468.5311 0.0012cJun18 0 to 0.5mm 06/27/10 0.2505 768.5653 NAcJun18 0 to 0.5mm 06/28/10 NA NA NAcJun18 0 to 0.5mm 06/29/10 0.2486 864.7582 NAcJun18 0 to 0.5mm 06/30/10 0.2638 936.798 0.0152cJun18 0 to 0.5mm 07/01/10 0.3162 465.5012 0.0524cJun18 0 to 0.5mm 07/02/10 0.3202 572.357 0.004cJun18 0 to 0.5mm 07/03/10 0.3138 139.3227 NAcJun18 0 to 0.5mm 07/04/10 0.3323 194.8929 0.0185cJun18 0 to 0.5mm 07/05/10 0.4214 103.6838 0.0891cJun18 0 to 0.5mm 07/06/10 0.3708 130.6914 NAcJun23 0 to 0.5mm 06/23/10 0.1379 524.4488 NAcJun23 0 to 0.5mm 06/24/10 0.1433 571.8843 0.0054cJun23 0 to 0.5mm 06/25/10 0.2059 355.1538 0.0626Continued on next page132B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun23 0 to 0.5mm 06/26/10 0.2094 713.9926 0.0035cJun23 0 to 0.5mm 06/27/10 0.2099 719.6211 5.00E-04cJun23 0 to 0.5mm 06/28/10 NA NA NAcJun23 0 to 0.5mm 06/29/10 0.1935 1356.899 NAcJun23 0 to 0.5mm 06/30/10 0.2109 1309.2196 0.0174cJun23 0 to 0.5mm 07/01/10 0.2576 509.6728 0.0467cJun23 0 to 0.5mm 07/02/10 0.2727 393.434 0.0151cJun23 0 to 0.5mm 07/03/10 0.2712 68.7547 NAcJun23 0 to 0.5mm 07/04/10 0.2816 318.2766 0.0104cJun23 0 to 0.5mm 07/05/10 0.2564 291.529 NAcJun23 0 to 0.5mm 07/06/10 0.2679 287.9916 0.0115cJun23 0 to 0.5mm 07/07/10 0.2751 166.387 0.0072cJun25 0 to 0.5mm 06/25/10 0.1381 631.814 NAcJun25 0 to 0.5mm 06/26/10 0.1423 1598.9801 0.0042cJun25 0 to 0.5mm 06/27/10 0.1416 1251.5875 NAcJun25 0 to 0.5mm 06/28/10 NA NA NAcJun25 0 to 0.5mm 06/29/10 0.1466 1392.7986 0.005cJun25 0 to 0.5mm 06/30/10 0.1441 0.0156 NAcJun25 0 to 0.5mm 07/01/10 0.2009 944.3732 0.0568cJun25 0 to 0.5mm 07/02/10 0.1877 867.7245 NAcJun25 0 to 0.5mm 07/03/10 0.1895 288.9121 0.0018cJun25 0 to 0.5mm 07/04/10 0.1963 623.0448 0.0068cJun25 0 to 0.5mm 07/05/10 0.2021 642.9493 0.0058cJun25 0 to 0.5mm 07/06/10 0.1969 601.0934 NAContinued on next page133B.7.LIFETABLESFORCOHORTSTRACKEDTable B.3 – continued from previous pageCohort Size Fraction Date Size (mm) Abundance(Ind.m−3)Growth(mm.day−1)cJun25 0 to 0.5mm 07/07/10 0.2122 153.2514 0.0153134B.8. SEAONAL GROWTH RATEB.8 Seaonal Growth RateBased on the time series for each component identified and tracked (Table B.3),the daily growth rate was calculated by averaging the daily growth rate for eachcohort tracked. Linear regressions were fitted to the time periods showing anincrease in growth rate. The full statistical results are presented in the table onthe following page.135B.8.SEAONALGROWTHRATETable B.4: Statistical results for the linear regressions of shell size growth for the population. The modal shell size (mm) is regressed over theobserved periods of increasing shell size for the population of L. helicina. Note that the values of the shell growth for the population were theaveraged daily growth rates of each component, tracked throughout the daily time series. Refer to Figure 3.2D. Standard errors of the slope andconstant/intercpet are displayed in parenthesis with the level of significance for each component displayed beside the slope estimate. The number ofobservations, R2, Adjusted R2, Residual Std. Errors and F statistics are displayed at the bottom of each table.Dependent variable:Population Growth Rate Population Growth Rate Population Growth Ratemm.day−1 mm.day−1 mm.day−11 May to 13 May 0.004∗∗∗(0.001)26 May to 16 June 0.001∗∗(0.001)7 June to 16 June 0.005∗∗∗(0.001)Constant −57.309∗∗∗ −19.265∗∗ −77.798∗∗∗(8.347) (7.761) (21.602)Observations 13 19 9R2 0.811 0.267 0.650Adjusted R2 0.794 0.224 0.600Residual Std. Error 0.008(df = 11) 0.015(df = 17) 0.013(df = 7)F statistic 47.187∗∗∗(df = 1; 11) 6.183∗∗(df = 1; 17) 12.983∗∗∗(df = 1; 7)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01136B.9. ENVIRONMENTAL CONNECTIONTable B.5: Statistical results for the linear regressions of shell size growth for the population.The modal shell size (mm) is regressed over the observed periods of increasing shell size forthe population of L. helicina. Note that the values of the shell growth for the populationwere the averaged daily growth rates of each component, tracked throughout the daily timeseries. Refer to Figure 3.2D. Standard errors of the slope and constant/intercpet are displayedin parenthesis with the level of significance for each component displayed beside the slopeestimate. The number of observations, R2, Adjusted R2, Residual Std. Errors and F statisticsare displayed at the bottom of each table.Dependent variable:Population Growth Rate Population Growth Ratemm.day−1 mm.day−126 May to 16 June 0.001(0.001)7 June to 16 June 0.004(0.003)Constant −10.960 −61.605(11.630) (38.000)Observations 22 10R2 0.043 0.248Adjusted R2 −0.005 0.153Residual Std. Error 0.023(df = 20) 0.023(df = 8)F statistic 0.895(df = 1; 20) 2.632(df = 1; 8)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01B.9 Environmental ConnectionThe results from Chapter 2 indicated that the temporal resolution may havebeen too coarse to discern any clear relation between chlorophyll abundance andL. helicina population variation. Despite using daily data, there were still noclear relations found between chlorophyll and L. helicina growth. It is probablethat the high grazing influence of the zooplankton community in Rivers Inlethad kept chlorophyll levels at low concentrations, such that any increase in L.helicina abundance or growth rate was not seen.Results from the linear regressions testing the relation between daily chloro-phyll and L. helicina abundance for each month, are presented in the tablebelow.137B.9.ENVIRONMENTALCONNECTIONTable B.6: Statistical results of linear regressions testing the relation between chlorophyll and the daily variation in populationabundance, for each month. Standard errors of the slope and constant/intercpet are displayed in parenthesis with the level ofsignificance for each component displayed beside the slope estimate. The number of observations, R2, Adjusted R2, ResidualStd. Errors and F statistics are displayed at the bottom of each table.Dependent variable:log-Abundance log-Abundance log-Abundance log-Abundance)(April) (May) (June) (July, 1–7)chl apr −0.020(0.034)chl may 0.015(0.031)chl jun −0.073(0.050)chl jul −0.004(0.071)Constant 4.265∗∗∗ 6.640∗∗∗ 7.504∗∗∗ 7.482∗∗∗(0.341) (0.223) (0.331) (0.464)Observations 28 30 29 7R2 0.014 0.009 0.072 0.001Adjusted R2 −0.024 −0.026 0.038 −0.199Residual Std. Error 0.751(df = 26) 0.741(df = 28) 0.986(df = 27) 0.660(df = 5)F statistic 0.357(df = 1; 26) 0.253(df = 1; 28) 2.102(df = 1; 27) 0.004(df = 1; 5)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01138B.10. SEASONAL MORTALITYB.10 Seasonal MortalitySeasonal mortality was analyzed from the component level as it was felt thatseasonal variation would be potentially masked, from analysis at the populationlevel. The daily abundance of each component tracked was log-transformed andlinear regressions were fitted to the perceived periods of notable decrease inabundance for each componentThe statistical tables were produced using the stargazer package (Hlavac,2013) for the R programming environment. Regression results (slope) for eachcomponent tracked is listed with applicable significance codes listed (see theNote: at the bottom of the table), as well as the standard errors (in paren-theses). The time periods of perceived abundance declines were regressed foreach component is listed on the left side of each table. The remaining statis-tical results (no. observations, R2, Adjusted R2, Residual Std. Errors, and Fstatistics) are listed at the bottom area of each table.Table B.7: Statistical results of linear regressions testing for periods of increasedmortality, for cohorts identified and followed in the daily time series. Standarderrors of the slope and constant/intercpet are displayed in parenthesis with thelevel of significance for each component displayed beside the slope estimate. Thenumber of observations, R2, Adjusted R2, Residual Std. Errors and F statisticsare displayed at the bottom of each table.Dependent Variable:Instantaneous Mortality Instantaneous Mortality(fraction dying per unit time) (fraction dying per unit time)7 May to 31 May −2.911(6.050)1 June to 7 July −3.172(2.590)Constant 42, 952.920 46, 859.600(89, 221.450) (38, 276.910)Observations 23 35R2 0.011 0.043Adjusted R2 −0.036 0.014Residual Std. Error 216.944(df = 21) 164.354(df = 33)F statistic 0.231(df = 1; 21) 1.500(df = 1; 33)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01139

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:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0167397/manifest

Comment

Related Items