Guidance on sampling effort to monitor mesozooplankton communities at Canadian bivalve aquaculture sites using an optical imaging system

Abstract

Despite the critical roles zooplankton play in marine food webs, the alterations to their communities by bivalve aquaculture have never been investigated empirically in Canadian waters. Collecting zooplankton data in bivalve aquaculture sites in a way that is interoperable over space and time is the first critical step to building consistent time series to detect changes over time and accurately inform management decisions.

Full text

1
Guidance on sampling effort to monitor mesozooplankton
communities at Canadian bivalve aquaculture sites using
an optical imaging system
Stephen Finnis, Thomas Guyondet, Christopher W. McKindsey, Julie
Arseneau, Jeffrey Barrell, Johannie Duhaime, Ramón Filgueira, Daria
Gallardi, David Gaspard, Olivia Gibb, Claire Goodwin, Khang Hua, Tara
Macdonald, Rebecca Milne, Anaïs Lacoursière-Roussel*
St. Andrews Biological Station
Fisheries and Oceans Canada
125 Marine Science Drive
St. Andrews, New Brunswick
E5B 0E4
2023
Canadian Technical Report of
Fisheries and Aquatic Sciences 3581

Canadian Technical Report of Fisheries and Aquatic Sciences
Technical reports contain scientific and technical information that contributes to existing knowledge
but which is not normally appropriate for primary literature. Technical reports are directed primarily
toward a worldwide audience and have an international distribution. No restriction is placed on subject
matter and the series reflects the broad interests and policies of Fisheries and Oceans Canada, namely,
fisheries and aquatic sciences.
Technical reports may be cited as full publications. The correct citation appears above the abstract of
each report. Each report is abstracted in the data base Aquatic Sciences and Fisheries Abstracts.
Technical reports are produced regionally but are numbered nationally. Requests for individual
reports will be filled by the issuing establishment listed on the front cover and title page.
Numbers 1-456 in this series were issued as Technical Reports of the Fisheries Research Board of
Canada. Numbers 457-714 were issued as Department of the Environment, Fisheries and Marine Service,
Research and Development Directorate Technical Reports. Numbers 715-924 were issued as Department
of Fisheries and Environment, Fisheries and Marine Service Technical Reports. The current series name
was changed with report number 925.
Rapport technique canadien des sciences halieutiques et aquatiques
Les rapports techniques contiennent des renseignements scientifiques et techniques qui constituent
une contribution aux connaissances actuelles, mais qui ne sont pas normalement appropriés pour la
publication dans un journal scientifique. Les rapports techniques sont destinés essentiellement à un public
international et ils sont distribués à cet échelon. II n'y a aucune restriction quant au sujet; de fait, la série
reflète la vaste gamme des intérêts et des politiques de Pêches et Océans Canada, c'est-à-dire les sciences
halieutiques et aquatiques.
Les rapports techniques peuvent être cités comme des publications à part entière. Le titre exact figure
au-dessus du résumé de chaque rapport. Les rapports techniques sont résumés dans la base de données
Résumés des sciences aquatiques et halieutiques.
Les rapports techniques sont produits à l'échelon régional, mais numérotés à l'échelon national. Les
demandes de rapports seront satisfaites par l'établissement auteur dont le nom figure sur la couverture et la
page du titre.
Les numéros 1 à 456 de cette série ont été publiés à titre de Rapports techniques de l'Office des
recherches sur les pêcheries du Canada. Les numéros 457 à 714 sont parus à titre de Rapports techniques
de la Direction générale de la recherche et du développement, Service des pêches et de la mer, ministère
de l'Environnement. Les numéros 715 à 924 ont été publiés à titre de Rapports techniques du Service des
pêches et de la mer, ministère des Pêches et de l'Environnement. Le nom actuel de la série a été établi lors
de la parution du numéro 925.

Canadian Technical Report of
Fisheries and Aquatic
Sciences 3581
2023
GUIDANCE ON SAMPLING EFFORT TO MONITOR MESOZOOPLANKTON
COMMUNITIES AT CANADIAN BIVALVE AQUACULTURE SITES USING AN
OPTICAL IMAGING SYSTEM
by
Stephen Finnis1, Thomas Guyondet2, Christopher W. McKindsey3, Julie Arseneau1, Jeffrey
Barrell2, Johannie Duhaime4, Ramón Filgueira5, Daria Gallardi6, David Gaspard7, Olivia
Gibb6, Claire Goodwin8, Khang Hua4, Tara Macdonald9, Rebecca Milne11, Anaïs
Lacoursière-Roussel1*
1*St. Andrews Biological Station
125 Marine Science Drive, St. Andrews
New Brunswick E5B 0E4
2Gulf Fisheries Center
343 Université Ave, Moncton
New Brunswick E1C 5K4
3Maurice Lamontagne Institute
850 Rte de la Mer, Mont-Joli
Quebec G5H 3Z4
4200 Kent Street, Ottawa
Ontario K1A 0E6
5Marine Affairs Program, Dalhousie University
1459 Oxford Street PO Box 15000, Halifax
Nova Scotia B3H 4R2
6Northwest Atlantic Fisheries Center
80 E White Hills Road, St. John’s
Newfoundland A1A 5J7
7Pacific Science Enterprise Center
4160 Marine Drive, West Vancouver
British Columbia V7V 1H2
8Huntsman Science Center
1 Lower Campus Road, St. Andrews
New Brunswick E5B 2L7
9Biologica Environmental Services Ltd.
488F Bay Street, Victoria
British Columbia V8T 5H2
11Bedford Institute of Oceanography
1 Challenger Drive PO Box 1006 Dartmouth
Nova Scotia B2Y 4A2
*Corresponding author. Email: Anais.Lacoursiere@dfo-mpo.gc.ca

ii
© His Majesty the King in Right of Canada, as represented by the Minister of the
Department of Fisheries and Oceans, 2023
Cat. No. Fs97-6/3581E-PDF ISBN 978-0-660-68776-6 ISSN 1488-5379
Correct citation for this publication:
Finnis, S., Guyondet, T., McKindsey, C.W., Arseneau, J., Barrell, J., Duhaime, J.,
Filgueira, R., Gallardi, D., Gaspard, D., Gibb, O., Goodwin, C., Hua, K.,
Macdonald, T., Milne, R., Lacoursière-Roussel, A. 2023. Guidance on sampling
effort to monitor mesozooplankton communities at Canadian bivalve aquaculture
sites using an optical imaging system. Can. Tech. Rep. Fish. Aquat. Sci. 3581: vii
+ 101 p.

iii
CONTENTS
ABSTRACT ........................................................................................................................................... v
RÉSUMÉ .............................................................................................................................................. vi
PREFACE ............................................................................................................................................ vii
1 INTRODUCTION ........................................................................................................................... 1
2 METHODS ...................................................................................................................................... 4
2.1 Sample collection .................................................................................................... 4
2.2 Sample laboratory processing ............................................................................... 10
2.3 Study area descriptions ........................................................................................ 11
2.3.1 Pacific region ................................................................................................. 11
2.3.2 Maritimes region ............................................................................................ 11
2.3.3 Gulf region ..................................................................................................... 12
2.3.4 Newfoundland region ..................................................................................... 13
2.4 Adjustments to the taxa list ................................................................................... 14
2.5 Converting counts to abundances in seawater ...................................................... 14
2.6 Statistical analysis................................................................................................. 14
2.6.1 Objective 1: Determining the optimal sampling effort per site ......................... 14
2.6.2 Objective 2: Characterizing patterns in zooplankton community structure
among regions, months, and sites ............................................................................... 17
2.6.3 Objective 3: Characterizing the role of tide phase and station on zooplankton
composition .................................................................................................................. 18
3 RESULTS ..................................................................................................................................... 20
3.1 Overview of images per site .................................................................................. 20
3.2 Objective 1: Determining the optimal sampling effort per site ................................ 20
3.2.1 Pacific region ................................................................................................. 20
3.2.2 Maritimes region ............................................................................................ 22
3.2.3 Gulf region ..................................................................................................... 24
3.2.4 Newfoundland region ..................................................................................... 26
3.2.5 All regions ...................................................................................................... 28
3.3 Objective 2: Characterizing patterns in zooplankton community structure among
regions, months, and sites ............................................................................................... 29
3.3.1 Regional comparisons ................................................................................... 29
3.3.2 Comparisons among sites or months ............................................................. 37
3.4 Objective 3: Characterizing the role of tide phase and stations on zooplankton
composition ..................................................................................................................... 51
3.4.1 Pacific region ................................................................................................. 51
3.4.2 Maritimes region ............................................................................................ 58
3.4.3 Gulf region ..................................................................................................... 63
3.4.4 Newfoundland region ..................................................................................... 68
4 DISCUSSION ............................................................................................................................... 75

iv
5 CONCLUSION ............................................................................................................................. 81
6 ACKNOWLEDGEMENTS .......................................................................................................... 82
7 AUTHOR CONTRIBUTIONS ..................................................................................................... 83
8 REFERENCES ............................................................................................................................. 84
APPENDIX 1 ....................................................................................................................................... 93
APPENDIX 2 ....................................................................................................................................... 97
APPENDIX 3 ..................................................................................................................................... 100
APPENDIX 4 ..................................................................................................................................... 101

v
ABSTRACT
Finnis, S., Guyondet, T., McKindsey, C.W., Arseneau, J., Barrell, J., Duhaime, J.,
Filgueira, R., Gallardi, D., Gaspard, D., Gibb, O., Goodwin, C., Hua, K.,
Macdonald, T., Milne, R., Lacoursière-Roussel, A. 2023. Guidance on sampling
effort to monitor mesozooplankton communities at Canadian bivalve aquaculture
sites using an optical imaging system. Can. Tech. Rep. Fish. Aquat. Sci. 3581: vii
+ 101 p.
Despite the critical roles zooplankton play in marine food webs, the alterations to their
communities by bivalve aquaculture have never been investigated empirically in Canadian
waters. Collecting zooplankton data in bivalve aquaculture sites in a way that is
interoperable over space and time is the first critical step to building consistent time series to
detect changes over time and accurately inform management decisions. As part of the
development of a nationally-consistent sampling design within the Aquaculture Monitoring
Program, this report evaluates mesozooplankton assemblages observed at nine coastal
aquaculture sites, located across four DFO regions, with sampling across months, tide
phases, and sampling locations. In most sites, strong spatial effects were observed, while
tide effects were generally less important for structuring the mesozooplankton communities.
Seasonality was identified as an important monitoring requirement to increase diversity
coverage and conduct interannual data comparisons. Conclusions provide direct advice on
the minimal sampling effort required to monitor mesozooplankton community changes over
time in Canadian bivalve aquaculture embayments. This report represents the first large-
scale Canadian coastal study using imaging technology for plankton taxonomic
identification; a method with the potential to enable more efficient, cost-effective monitoring
and refine our understanding of the state of the Canadian oceans.

vi
RÉSUMÉ
Finnis, S., Guyondet, T., McKindsey, C.W., Arseneau, J., Barrell, J., Duhaime, J.,
Filgueira, R., Gallardi, D., Gaspard, D., Gibb, O., Goodwin, C., Hua, K.,
Macdonald, T., Milne, R., Lacoursière-Roussel, A. 2023. Guidance on sampling
effort to monitor mesozooplankton communities at Canadian bivalve aquaculture
sites using an optical imaging system. Can. Tech. Rep. Fish. Aquat. Sci. 3581: vii
+ 101 p.
Malgré le rôle fondamental du zooplancton dans les réseaux trophiques marins, les
changements de ces communautés liés à la conchyliculture n'ont jamais été étudiés
empiriquement dans les eaux canadiennes. Assurer la collecte des données de zooplancton
dans les sites conchylicoles de façon interopérable dans l'espace et dans le temps est la
première étape critique pour établir des séries chronologiques cohérentes et informer la
gestion avec précision. Afin d’élaborer un plan d’échantillonnage consistent à l’échelle
nationale au sein du Programme de la surveillance en aquaculture, ce rapport évalue
l'assemblage du mésozooplancton observé à neuf sites conchylicoles, situés dans quatre
régions du MPO, échantillonnés à différents mois, marées et emplacements. Dans la plupart
des sites, l’effet spatial était hautement significatif et l’effet de la marée était moins
important. La saisonnalité a également été identifiée comme un facteur important pour
augmenter la détection de la diversité et effectuer des comparaisons de données
interannuelles. Les conclusions fournissent des conseils sur l'effort d'échantillonnage
minimal requis pour surveiller les changements de communautés de mésozooplancton au fil
du temps. Ce rapport représente la première étude à grande échelle en zones côtières
canadiennes mettant en œuvre une technique d'imagerie pour identifier le plancton; une
méthode susceptible de permettre une surveillance plus efficace et rentable et d’affiner
notre compréhension de l'état des océans canadiens.

vii
PREFACE
To date, bivalve aquaculture research has predominantly focused on near-field benthic
effects, while limited research has documented far-field (i.e., bay-scale) effects on lower
trophic levels (nutrients, phytoplankton, and zooplankton) (Weitzman et al. 2019). Previous
Fisheries and Oceans Canada Program for Aquaculture Regulatory Research (DFO-PARR)
projects highlighted that pelagic ecosystem perturbations at low trophic levels may cause
fundamental shifts to food web dynamics, especially with reduced predator-prey size ratios
resulting in longer, less efficient food chains (Gianasi et al. 2023). However, earlier research
on carrying capacity assessments and models could only consider zooplankton as a bulk
secondary producer component, as opposed to a main trophic link between primary
producers and fisheries productivity due to limited knowledge on the structure and dynamics
of zooplankton communities in Canadian bivalve aquaculture sites. In 2018, the Aquaculture
Monitoring Program Working Group (AMP-WG) was thus mandated to develop a nationally-
consistent program to monitor ecosystem impacts from bivalve aquaculture at the bay-scale.
It was decided that the use of zooplankton to monitor potential ecosystem interactions of
bivalve aquaculture would be evaluated. The long-term AMP data will allow for the
development of new and updated models to explore how bivalve aquaculture potentially
impacts fisheries resources, which may vary between aquaculture sites and in future climate
scenarios. This improved scientific understanding will support the Department through an
increased capacity to develop evidence-based advice and mitigation strategies. Collectively,
this research will help inform aquaculture policy and regulatory decision-making to enhance
the sustainability of the aquaculture in Canada.

1
1 INTRODUCTION
The bivalve aquaculture industry offers many potential benefits to communities, including
food security, economic opportunities, and ecosystem services (Flaherty et al. 2019;
Wijsman et al. 2019; Azra et al. 2021). However, concerns and uncertainties related to the
ecosystem effects of intensive bivalve farming still remain (Filgueira et al. 2016; Grant and
Pastres 2019; Holden et al. 2019; Hulot et al. 2020). In particular, the potential effects of
bivalve aquaculture on zooplankton communities are largely unknown, and this may have
important repercussions as zooplankton represent a key link for energy and mass transfer
between trophic levels (Lindeman 1942; Kiørboe 2009; Hulot et al. 2014). The alterations to
zooplankton communities have mostly been studied through laboratory research or field
studies of invasive bivalves (Davenport et al. 2000; Zeldis et al. 2004; Lehane and
Davenport 2006; Pace et al. 2010). Overall, there are few examples of these effects in the
field (Hulot et al. 2020), although Maar et al. (2008) observed zooplankton depletion in a
mussel aquaculture site. Generally, the effects of bivalves on zooplankton communities are
thought to be both indirect (i.e., through food limitation and increased competition as a result
of the filtration of phytoplankton) and direct (i.e., filtration and ingestion) (Lehane and
Davenport 2006). Thus, research of these understudied interactions between bivalves and
zooplankton communities are critical for the sustainable development of the bivalve
aquaculture industry.
In 2018, Fisheries and Oceans Canada (DFO) launched the Aquaculture Monitoring
Program (AMP) to increase departmental availability of scientific data to support aquaculture
policy and decision-making for enhanced aquaculture sustainability. The national program is
currently in development, and aims to implement long-term monitoring of the spatiotemporal
variations of potential far-field environmental effects of aquaculture (i.e., hundreds of meters
beyond the lease boundaries; Weitzman et al. 2019) using nationally-consistent sampling
approaches. One component of AMP is focused on improving scientific understanding of
zooplankton-bivalve dynamics at aquaculture sites, and characterizing long-term natural
variability in zooplankton communities to determine whether natural variations can be
disentangled from aquaculture-induced effects. Ultimately, analyzing zooplankton size and
community structure could potentially be used to describe and monitor how bivalve
aquaculture might impact energy flows to higher trophic levels within pelagic food webs.
More specifically, monitoring zooplankton communities in bivalve aquaculture sites over time
will allow researchers to: (i) evaluate if bivalve farms directly or indirectly impact zooplankton
size and community structure, (ii) better understand ecosystem-level changes to
zooplankton communities during variations of the production levels (i.e., increasing or
decreasing bivalve production) or transitions between bivalve culture types, (iii) elucidate
complex trophic interactions, including benthic-pelagic food-web coupling, (iv) monitor
potential trophic cascades, including impacts on zooplankton abundance, biomass and
productivity, and (v) provide historical databases to address potential issues related to
aquaculture (e.g., monitoring the larval stages of lobsters, crabs, or tunicates) or future
climate change scenarios.
Effective zooplankton monitoring requires reliable estimates of the biodiversity within an
area, yet monitoring programs routinely face challenges of determining when a location has
been adequately sampled (Angermeier and Smogor 1995; Olsen et al. 1999; Yoccoz et al.
2003). Measurable components of biodiversity, such as richness (the total number of
species within an area), are often highly dependent on the level of sampling effort, since
more species will be detected by increasing the number of samples collected (Colwell et al.
2012; Chao et al. 2014, 2020). Rare taxa are often of interest in aquatic bioassessments as
they may play critical roles in ecosystems and can be useful indicators of human-induced
changes (Cao et al. 1998), although they can require substantially more sampling effort to
detect (Colwell et al. 2012). Defining the “appropriate” level of sampling effort is essential,
since too few samples may lead to incorrect conclusions about an ecosystem, whereas too

2
many samples may result in redundant or minimal new information being collected, thereby
resulting in an overallocation of resources (Angermeier and Smogor 1995; Danielsen et al.
2000). Various statistical approaches exist to identify the number of samples at which little
new information (e.g., few new species) are added (Colwell et al. 2012). Monitoring
programs, especially those in their infancy, may benefit substantially from applying these
approaches by using data-driven estimates to indicate the most effective allocation of
resources or funding (Danielsen et al. 2000; Hoffman et al. 2011).
Identifying the appropriate spatial and temporal scales for sampling is a key feature in
designing environmental monitoring programs since selecting the incorrect scales may result
in the wrong processes being measured (Birk et al. 2020; Ma et al. 2022). For example, in
coastal environments, zooplankton distributions are patchy at a range of spatial and
temporal scales, resulting from a combination of both large scale physical (i.e., passive
displacement with the water) and biological (i.e., active movement for predation, mating,
food searching, etc.) processes (Folt and Burns 1999; O’Brien and Oakes 2020). In bivalve
aquaculture sites (i.e., bays), little is known about how depletion from direct bivalve grazing
may alter these spatiotemporal zooplankton variations. Spatially, coupled biological-
hydrodynamic models and remote sensing have shown that seston depletion from bivalve
aquaculture can occur at the bay scale (Grant et al. 2007, 2008; Filgueira et al. 2014, 2015;
Taylor et al. 2021). Complex aquaculture and ecosystem interactions might also occur at
various temporal scales, such as tidal exchange, water residence time, and daily variation
due to environmental stochasticity (i.e., daily to weekly). Furthermore, the interaction
between bivalve aquaculture and zooplankton might vary seasonally due to changes related
to (i) bivalve feeding rates, (ii) plankton production rates, (iii) water-column stability and (iv)
circulation patterns (Grant et al. 2008; Steeves 2018). Basic understanding of these short-
term spatiotemporal variations is thus crucial to interpret factors altering long-term
zooplankton shifts, distinguish between natural and potential bivalve aquaculture-related
effects, and define an optimal sampling effort. Further examination of zooplankton
distributions, both spatially and temporally, is therefore an important first step to provide a
more detailed understanding of these processes.
Historically, biodiversity assessments have relied on microscopy for the identification and
enumeration of zooplankton specimens (Le Bourg et al. 2015; Detmer et al. 2019). However,
in the last decade, the use of innovative imaging instruments and machine learning
algorithms has grown to automatically identify and classify plankton images, which can help
quantify plankton diversity and functional traits (Luo et al. 2018; Ramkissoon 2021;
Orenstein et al. 2022). These offer many benefits when compared to traditional microscopy,
including more rapid identification of taxa at lower cost (Álvarez et al. 2014; Le Bourg et al.
2015; Detmer et al. 2019). To improve our ability to better understand impacts of human
stressors on coastal food webs, a collaborative international effort is currently underway to
improve the reference libraries of zooplankton images (e.g., Ibarbalz et al. 2019; Kerr et al.
2020). For example, flow cytometer and microscope (FlowCam) methodologies are being
developed to explore zooplankton size and abundance as part of the Continuous Plankton
Recorder Survey, an extensive global marine monitoring program (Batten et al. 2019). In this
context, AMP is the first DFO initiative to capitalize on these new monitoring approaches to
detect changes in zooplankton communities.
This report examines spatiotemporal variations of mesozooplankton (i.e., 0.25 mm - 5 mm)
collected in four coastal DFO regions (Pacific, Maritimes, Gulf, and Newfoundland and
Labrador; herein referred to solely as “Newfoundland”), and analyzed using an optical
imaging system. Specifically, we provide science-based recommendations on optimal
sampling effort to monitor mesozooplankton in Canadian bivalve aquaculture sites by
accomplishing the following objectives:
1. Evaluate the optimal sampling effort by examining how taxa diversity changes with
increasing sample size.

3
2. Examine differences in mesozooplankton community structure among regions, sites,
and/or months.
3. Characterize variations in mesozooplankton community structure within sites,
including how assemblages vary by tides and stations (i.e., locations within each
bay).

4
2 METHODS
2.1 Sample collection
Mesozooplankton counts have been compiled from various sampling sites across Canada, in
different seasons, tide phases and sampling stations. For clarity, we use the following
terminology for the various spatial scales: region refers to DFO regions (i.e., Pacific,
Maritimes, Gulf, and Newfoundland); site refers to an individual bay, inlet, lagoon or harbour;
station represents the different locations sampled within a given site (e.g., Inner, Mid, or
Outer); and sample refers to the individual collected measurements from a zooplankton net
tow. Data were collected from nine sites across Canada within four DFO regions and across
varying collection months (February to December) (Fig. 1; Table 1). Sites were selected to
span a range of oceanographic conditions and bivalve aquaculture intensity levels, with
varying pre-existing knowledge of the hydrodynamics and bivalve aquaculture-environment
interactions in each of them. Within each site, the number of bivalves per lease is often not
known as this information is generally proprietary. However, bivalves were present within all
sampling sites, except for Argyle and Country Harbour, in which all leases were empty
during the time of sampling (Fig. 1). Within each site, samples were collected at three
stations to gain a general understanding of spatial dynamics in zooplankton community
structure. In general, for sites with a single point of exchange with the open ocean that follow
a linear path, stations were labeled as “Inner”, “Mid”, and “Outer”. For instances where this
labeling scheme would not be applicable (e.g., in bays with more complex coastlines and >1
location of exchange with the ocean), stations were generally instead labeled according to
cardinal directions (e.g., “North,” “South,” and “Central”) (Fig. 1). For these sites without
linear morphology, the location of the three sampling stations were selected according to a
variety of factors including logistics (e.g., suitable depth, having enough space to avoid the
nets getting tangled in leases) or knowledge of the circulation patterns (e.g., selecting a
station close to a point of exchange with the open ocean). In addition, station locations were
chosen to sample areas both near and far from the leases, while keeping an approximately
even spacing of stations throughout the site.

5

6
Figure 1. Study area maps showing sampling locations for mesozooplankton (0.25 mm -
5.00 mm) as part of the Aquaculture Monitoring Program. Insets show data sampled from
within the Pacific (Pac; D), Maritimes (Mar; E-H), Gulf (I-K) and Newfoundland regions (Nfld;
L), as labelled in panels B and C. Samples were obtained either as vertical tows (blue
circles) or as transects (blue lines), and text labels indicate station names. Pink polygons
represent the bivalve aquaculture leases obtained from TANTALIS Crown Features for the
Pacific region (D; https://fisheries-map-gallery-
crm.hub.arcgis.com/datasets/governmentofbc::tantalis-crown-tenures), the Nova Scotia
Department of Fisheries and Aquaculture mapping tool for sites in Nova Scotia (E-H;
https://novascotia.ca/fish/aquaculture/site-mapping-tool/), the Marine Aquaculture Site
Mapping Program for sites in New Brunswick (I;
https://www2.gnb.ca/content/gnb/en/departments/10/aquaculture/content/masmp.html), the
Prince Edward Island (PEI) Aquaculture Interactive Leasing Maps for sites in PEI (J, K;
https://www.dfo-mpo.gc.ca/aquaculture/management-gestion/pei-lic-ipe-baux-eng.htm), and
Crown Title leases from the Newfoundland Land Use Atlas for Newfoundland (L;
https://www.gov.nl.ca/landuseatlas/details). Only active leases present at the time of
sampling are shown (i.e., “under review” or “proposed” leases are not included). The
presence of leases does not necessarily mean leases were stocked with bivalves during
sampling, as this information is often proprietary, although it is known that leases in Argyle
(E) and Country Harbour (G) were not stocked with bivalves during the time of sampling. The
shaded blue polygons in each inset represent the boundaries used for calculations of site
area and total lease area shown in Table 1. The exact boundaries of these polygons are
subjective, yet were shown for an approximate comparison of relative size among study
sites.

7
Table 1. Data summary table for mesozooplankton (0.25 mm - 5.00 mm) samples obtained within study sites and regions, as part of the
Aquaculture Monitoring Program. The site and lease areas were calculated within the shaded blue areas shown in Fig. 1. Coverage (%)
represents the ratio of the lease area to site area. The tidal range (m) is presented as the range between the lower low water mean tide to
higher high water mean tide, from the nearest Canadian Hydrographic Service tide station to the sampling stations
(https://wla.iwls.azure.cloud.dfo-mpo.gc.ca/stationMgmt). Several additional opportunistic surveys from various funding sources resulted in
different temporal coverage for some sites. Samples were therefore obtained in more than one month for Cocagne (Gulf), South Arm
(Newfoundland), and Lemmens Inlet (Pacific). n refers to the number of samples obtained at each site and/or month. Pac: Pacific, Mar:
Maritimes, Nfld: Newfoundland.
Region
Site and/or
month
Site area
(km2)
Lease
area
(km2)
Coverage
(%)
Bivalve
type
Tide range
(m)
Hydro.
model
Max
depth
(m)
Year
Date
range
Tow type
n
Pac
Lemmens
Aug 2020
6.44
0.23
3.6
Pacific
oyster
0.74-3.39
(Tofino,
08615)
[1]
27.9
2020
Aug 29-
31
Vertical
18
Lemmens
Mar 2021
2021
Mar 3-5
Vertical
2
Lemmens
Jun 2021
Jun 9-11
Vertical
18
Lemmens
Sept 2021
Sep 14-
15
Vertical
12
Mar
Argyle
137.00
0.62
0.5
Eastern
oyster
0.66-3.71
(Wedgeport,
00374)
N/A
16.5
2021
Aug 30-
Sep 1
Oblique
15
Country
Harbour
10.40
0.84
8.1
Eastern
oyster
0.53-1.87
(Isaacs
Harbour,
00535)
N/A
21.9
2021
Aug 24
Vertical
6

8
Region
Site and/or
month
Site area
(km2)
Lease
area
(km2)
Coverage
(%)
Bivalve
type
Tide range
(m)
Hydro.
model
Max
depth
(m)
Year
Date
range
Tow type
n
Sober
Island
0.90
0.09
9.6
Eastern
oyster
No tidal
station
[2]
4.0
2021
Aug 27
Horizontal
12
Whitehead
1.68
0.23
13.9
Eastern
oyster
0.45-1.80
(Whitehead,
00545)
[2]
14.3
2021
Aug 25
Vertical
9
Gulf
Cocagne
18.90
1.44
7.6
Eastern
oyster
0.32-1.05
Cocagne
(01812)
[3]
8.3
2021
Jul 21
Horizontal
3
Aug 26
Horizontal
3
Malpeque
207.40
14.76
7.1
Blue
mussel,
eastern
oyster
0.24-0.96
(Malpeque,
01905)
[4, 5, 6]
14.2
2020
Sept 29
Horizontal
3
St. Peters
15.78
6.37
40.4
Blue
mussel,
eastern
oyster
0.23-0.79
(St. Peters
Bay, 01935)
[7]
5.0
2020
Sept 1-4
Horizontal
26
Nfld
South Arm
Sep 2020
11.31
2.80
24.8
Blue
mussel
0.27-1.18
(Leading
Tickle,
01087)
[8]
45.0
2020
Sep 15-
16
Vertical
10
South Arm
Oct 2021
2021
Oct 5-7
Vertical
12

9
Region
Site and/or
month
Site area
(km2)
Lease
area
(km2)
Coverage
(%)
Bivalve
type
Tide range
(m)
Hydro.
model
Max
depth
(m)
Year
Date
range
Tow type
n
South Arm
(monthly
surveys)
2021-
2022
2021: Jun
9, Aug
12, Sept.
8, Nov. 9,
Dec. 14
2022:
Feb. 8,
Mar. 29,
Apr. 22,
May 17,
Jun 7, Jul
6
Vertical
31 (2-
3 per
mont
h)
[1] Foreman et al. (submitted for publication); [2]: Filgueira et al. (2021); [3] Guyondet et al. (unpublished), [4]: Filgueira et al. (2015); [5]: Bacher
et al. (2016); [6]: Lavaud et al. (2020); [7]: Guyondet et al. (2015); [8] Gallardi et al. (in development, 2023).
31

10
In each region, mesozooplankton samples were collected using plankton nets at three
stations per site (i.e. bay) at low and high tide for three consecutive days, although there
were deviations to this sampling strategy (see Appendix 1 for a complete overview of
sampling effort per tide phase). The net mesh of all tows was 250 μm, except a 236 μm net
was used for the August 2020 dataset in Lemmens Inlet (Pacific), and a 150 μm net was
used in St Peters (Gulf), due to a lack of available 250 μm nets. These were assumed to
produce similar results, since samples were all sieved in the lab for specific fraction sizes,
and particles <250 μm were further removed during FlowCam imaging (see below). Sample
collection covered both the holoplankton (i.e., planktonic for their entire life cycle) and
meroplankton (i.e., planktonic for one portion of their life cycle).
At each station, zooplankton tows were obtained to sample the full water column to
overcome the confounding factor of vertical migrations. To target approximately 7000 L
filtered water volume, one or two tow(s) were collected and combined within a single jar.
Vertical, oblique, or horizontal tows were collected, with the type of tow selected according to
depth of the water column (see Table 1 for the type of tow used). The tow aimed to sample
as much of the water column as possible, without the net contacting the seabed (1-2 m
above the seabed). When possible, vertical tows were used, in which nets were lowered to
1-2 m above the seabed at a rate of approximately 0.5 m per second, and retrieved at
approximately 1 m per second. In shallower waters where a vertical tow would result in
minimal zooplankton being collected, oblique tows were obtained. For oblique tows, the tow
was started at 3 m above the seafloor, and slowly raised through the water column, starting
at the deep end and moving towards the shore. Lastly, horizontal tows (i.e., used in the
shallowest waters) were obtained at a fixed depth of approximately 1 m below the surface.
For horizontal and oblique tows, multiple 3 minute tows were collected, and a tow speed of
approximately 2-3 knots was used to limit avoidance of the nets by the zooplankton. Upon
retrieval, samples were preserved in a 4% solution of buffered formaldehyde. A calibrated
mechanical flowmeter (General Oceanics Inc.; Product code 203001) was mounted through
the net opening to calculate the filtered water volume. In instances where the water volumes
calculated from the flowmeter appeared unreliable, the water volume filtered was instead
calculated using the depth of the water column or length of the tow.
2.2 Sample laboratory processing
Samples were split into four equal subsamples using a Folsom Plankton Splitter and each
subsample was used for different measurements including (1) biomass, (2) imaging (i.e.
abundance, size spectra, community structure) and (3) traditional taxonomy
(image/specimen reference collection, quantitative assessment, diversity). This report
presents only results for the community structure obtained from the imaging system, while
the other components are being analyzed as part of ongoing departmental work. The
terminology “samples” is used below instead of subsample, but in fact they are thus only
0.25 samples. Samples were analyzed using a flow imaging microscopy system, FlowCam
Macro (Yokogawa Fluid Imaging Technologies, Inc.) Detailed methods of the FlowCam
procedures are in progress as a separate report, which will also include a detailed
comparison of counts of mesozooplankton taxa obtained from traditional microscopy to
counts obtained by the FlowCam. However, following the recommendations of Owen et al.
(2022), key details required for creating reproducible work involving FlowCam technology is
included in tabular format in Appendix 2. A brief description of the laboratory procedures is
also described below.
Samples were rinsed through a series of stacked sieves; 2 mm mesh sieve stacked on a 125
μm or 212 μm mesh sieve. Taxa collected on the 2 mm mesh that were <5 mm, were
processed using a 5 mm flow cell, while the specimens collected on the 125 μm or 212 μm
mesh were kept separate and run through the FlowCam using a 2 mm flow cell. Particles >5
mm were uncommon, and too large to fit within the 5 mm flow cell. These were therefore not
included in the study due to the size limits of the instrumentation. The specimens from both

11
size fractions (i.e., <2 mm and 2 mm-5 mm) were then rinsed into a sample beaker
containing approximately 400 ml of 0.2% Triton-x. Using a large volume of 0.2% Triton-x and
a magnetic stirrer in the sample beaker was shown to be successful in reducing the
clumping of plankton. Images were completed for the full samples. Based on variable
funding between regions and years and the difference in taxa and debris to classify, the
number of images identified varied between sample sets. VisualSpreadsheet version 5.6.14
was used to classify images for data collected in 2020 in Newfoundland, while
VisualSpreadsheet version 4.18.5 used for all other samples (Yokogawa Fluid Imaging
Technologies, Inc. 2020, n.d.). First, images of objects <250 μm in length were removed
using a pre-set filter. Then, a fraction of the samples were cleaned by removing images of
debris, fragments of plankton (i.e. <15% body size), bubbles, cropped images and clumped
plankton. Images produced by the FlowCam appear as black and white pictures, and the
images of confirmed zooplankton were then classified in the categories as described in
Appendix 3. These categories have been defined specifically for this project by experts in
zooplankton taxonomy using the shapes of the organisms in the images collected from all
regions. For Copepoda, the most abundant zooplankton taxon, there are six copepodite
developmental stages (Ci to Cvi). However, for the FlowCam reference collection, Ci-Ciii are
classified only at the order level, due to the limitation in image quality of the FlowCam.
Copepodite developmental stages Civ-Cvi were identified to the genus level, because that is
when the copepod shape becomes discernible from FlowCam images. Although the entire
subsample was run through the FlowCam, due to time constraints, generally only a fraction
of the specimens with entire sample were identified by the taxonomists. The identification of
specimens was manually performed by the taxonomists using the two-dimensional black and
white digital images in VisualSpreadsheet. These classifications are also being used to
create libraries for the automated identification of mesozooplankton taxa using machine
learning algorithms, which is being investigated as part of ongoing departmental work.
2.3 Study area descriptions
2.3.1 Pacific region
Within the Pacific region, samples were obtained solely from Lemmens Inlet on the west
coast of Vancouver Island (Fig. 1D). The sampled site covers 6.44 km2, and has complex
bathymetry, where depths increase toward the middle of the inlet, and generally range from
12-17 m (Table 1). A maximum depth of 27.9 m occurs near the Mid station. The tidal cycle
is mixed semidiurnal, and the nearest Canadian Hydrographic Service (CHS) tide station
(Tofino 08615, 49.154°N, 125.913°W) experiences a tidal range from 0.74 m Lower Low
Water Mean Tide (LLWMT) to 3.39 m Higher High Water Mean Tide (HHWMT). Samples
were obtained as vertical tows across four time periods including August 2020 (n = 18),
March 2021, June 2021 (n = 18), and September 2021 (n = 12). Samples were obtained at
high tide and low tide, at three stations, over either 2 or 3 days total (Appendix 1). Due to a
handling error, the samples in March 2021 were combined by tide, resulting in a single “High
Tide” sample (consisting of data from five combined samples), and a single “Low Tide”
sample (consisting of data from six combined samples). Stations were positioned with
approximately 1.6 km spacing, for a 3.2 km total distance between the Inner and Outer
stations. Within the study area (Fig. 1D), bivalve aquaculture leases cover 0.23 km2, or 3.6%
of the surface area. Pacific oysters (Crassostrea gigas) are the main cultured bivalve within
Lemmens Inlet.
2.3.2 Maritimes region
The Argyle sampling site is located in southern Nova Scotia and comprises a highly complex
coastline with numerous channels and islands (Fig. 1E). The region opens into the Atlantic
Ocean, and when using this boundary as an approximate site delineation (Fig. 1E), covers a

12
surface area of 133.66 km2 (Table 1). For most of the site, depths are less than 6 m,
although certain areas have carved underwater channels with depths ranging from 13.0-16.5
m. The exact tidal range at the study stations is unknown, but tides from the nearest CHS
station (Wedgeport 00374, 43.733°N, 65.983°W) range from 0.66 m LLWMT to 3.71 m
HHWMT. Samples were collected from August 30 to September 1, 2021 (n = 15, oblique
transects) at a range of tide phases (Appendix 1). The South and Central stations were
spaced by 5.2 km, and the Central and North ones by 3.1 km. Eastern oysters (Crassostrea
virginica) represent the main cultured bivalve within the site, and leases cover a surface area
of 0.62 km2 (i.e., 0.45% of the total bay area), although leases were empty at the time of
sampling.
Sober Island is a small lagoon on the eastern shore of Nova Scotia (Fig. 1F) with an
approximate bay area of 0.90 km2 and average depth of 2.9 m (Filgueira et al. 2021). Sober
Island is separated from the ocean by a narrow channel (~20 m wide, ~1 m depth) and
receives minimal freshwater input (Filgueira et al. 2021). A hydrodynamic model was
constructed for Sober Island by Filguiera et al. (2021), who used a Finite Volume Community
Ocean Model (FVCOM) to evaluate the ecological carrying capacity of bivalve aquaculture.
Results showed water renewal times of less than three days for most of Sober Island,
although some small sections in the northwestern and southwestern portions of the bay had
water renewal times exceeding 12 days. Samples were obtained on August 27, 2021 (n =
12, horizontal transects) at various tide phases, with stations separated by distances
ranging from 0.5 km - 1.0 km (Table 1, Appendix 1). Eastern oysters are the farmed bivalves
within Sober Island and leases cover 0.09 km2, or 9.6%, of the bay.
Country Harbour is situated on the eastern shore of Nova Scotia and covers an approximate
surface area of 10.4 km2 (Fig. 1G). The site is a long channel that opens into the exposed
Atlantic Ocean. Depths generally range from 10-17 m, and reach a maximum depth of 21.9
m near the Mid station. Tides are semidiurnal, and the nearest tide station (Isaacs Harbour,
00535, 45.183°N, 61.667°W) has a tidal range from 0.53 m LLWMT to 1.87 m HHWMT
(Table 1). Samples were obtained on August 24, 2021 (n = 6, vertical tows) from both high
and low tide phases (Appendix 1). Sampling stations span the length of the bay and were
approximately evenly spaced at distances of 3.1 km, for a total distance of 6.2 km separating
the Inner and Outer stations. Eastern oysters are the main cultured bivalve at Country
Harbour, with leases covering 0.84 km2 (8.1% of total bay area), although leases were
empty at the time of sampling.
The Whitehead sampling site is located on the northeast coast of Nova Scotia (Fig. 1H). The
sampled area consists of a long, narrow channel 1.68 km2, that opens into Whitehead
Harbour (Table 1). Within the channel, depths generally range from 4-7 m but reach a
maximum depth of 14.3 m near the Mid station. Tides are semidiurnal and range from
approximately 0.45 m LLWMT to 1.80 m HHWMT (Whitehead tide station 00545, 45.233°N,
61.183°W). An FVCOM hydrodynamic model was constructed for Whitehead by Filgueira et
al. (2021), which calculated that water renewal time for the channel exceeds 20 days.
Sampling occurred on August 25, 2021 (n = 9, vertical tows) along a linear path at three tide
phases (Appendix 1). The Inner and Mid stations were located 1.1 km apart, and the Mid
and Outer stations were spaced 2.1 km apart. Within the site, there is one large lease (0.23
km2) for eastern oysters, which covers 13.9% of total area.
2.3.3 Gulf region
Cocagne is a large (18.90 km2) bay on the eastern coast of New Brunswick (Fig. 1I).
Cocagne Island separates the bay from the Northumberland Strait at two main openings.
Depths are generally shallow (0.3 m-2.1 m) and tides range from approximately 0.32 m
LLWMT to 1.05 m HHWMT (Cocagne tide station 01812, 46.333°N, 64.617°W; Table 1).
Samples were obtained from horizontal transects at three stations, which had varying

13
exposure to the Strait (Fig. 1I). The Central station was located approximately 2.4 km from
the North and South stations, while North and South separated by 4.4 km. Three samples
were collected on July 21, 2021 (one at each station, low tide only), while three additional
samples were collected on August 26, 2021 (one at each station, mid-rising tides only) by
horizontal transects (Appendix 1). The study area consists of numerous aquaculture leases,
mostly eastern oysters, covering a total area of 1.44 km2, i.e., 7.6% of the bay area.
Malpeque is a large bay composed of multiple sub-basins covering a surface area of 207.40
km2 (Fig. 1J) on the north shore of Prince Edward Island (PEI). The bay opens into the Gulf
of St. Lawrence at three openings and receives freshwater input from 12 rivers and several
smaller streams (Lavaud et al. 2020). Depths are generally shallow (<10 m), and reach a
maximum depth of 14.2 m between the North and Central stations. Tides range from
approximately 0.24 m LLWMT to 0.96 m HHWMT (Malpeque tide station 01905, 46.533°N,
63.700°W; Table 1). The potential expansion of the aquaculture industry in Malpeque Bay
drove the development of a series of hydrodynamic models. For example, a two-dimensional
spatially explicit hydrodynamic model developed in Filgueira et al. (2015) to evaluate the
potential effects of varying levels of mussel aquaculture expansion on the ecological carrying
capacity in the bay. This model was expanded in Lavaud et al. (2020) to incorporate sea
lettuce, wild and cultured oysters, and clams. Additionally, Bacher et al. (2016) used Markov
Chain to evaluate the residence time (21-35 days in the sampling area), and connectivity
with respect to aquaculture in the bay. Three samples were obtained on Sept 29, 2020 at
three locations (all at low tide, horizontal transects; Appendix 1), termed South, Central, and
North. Stations were spaced at distances ranging from 3.6 to 4.5 km apart. The blue mussel
(Mytilus edulis) is the dominant bivalve in terms of biomass produced (5120-9600 t; Filgueira
et al., 2015), although cultures of eastern oysters are also present. The numerous leases
cover a surface area of 14.76 km2, equivalent to 7.1% of the surface area of the bay.
St. Peters Bay is an elongated inlet on the north shore of PEI, covering an area of 15.78 km2
(Fig. 1K). The bay opens into the Gulf of St. Lawrence, and is mostly shallow, with depths
ranging from 3-5 m throughout the length of the bay. Tides range from approximately 0.23 m
LLWMT to 0.79 m HHWMT (St. Peters Bay tide station 01935, 46.433°N, 62.733°W; Table
1). Except at the region exposed to the Gulf of St. Lawrence, most of the bay experiences
poor flushing and tidal currents are several centimeters per second (Guyondet et al. 2015). A
spatially explicit coupled hydrodynamic-biogeochemical model developed by Guyondet et al.
(2015) suggests that water renewal time increases along the length of the bay, taking up to
~90 days in the innermost region. Samples were obtained along the length of the bay,
separated by 4.1 km (8.1 km total distance between Inner and Outer stations) from
September 1-4, 2020, at a range of tide phases (n = 26, horizontal transects) (Appendix 1).
Within St. Peters, bivalve aquaculture consists of a combination of blue mussels and eastern
oysters, with leases covering 6.37 km2, or approximately 40.4% of the bay area.
2.3.4 Newfoundland region
South Arm, Newfoundland, is a long channel that opens to the Atlantic Ocean (Fig. 1L). The
innermost portion of the channel has a surface area of 11.31 km2 (Table 1). South Arm is the
deepest of all study sites within AMP, with spatially variable depths exceeding 45 m. Tides
from the nearest CHS station range from 0.27 m LLWMT to 1.18 m HHWMT (Leading Tickle
tide station 01087, 49.502°N, 55.447°W). Samples were collected across a range of months
for a full examination of seasonal dynamics within the site. As for other AMP sites, multiple
samples were collected in September 2020 (n =10) and October 2021 (n = 12) to examine
the effects of station and tide phase on mesozooplankton composition. In addition, either two
or three samples were obtained monthly from June 2021 to July 2022 (excluding January
2022) for a more detailed examination of temporal dynamics (Appendix 1). Across all
months, samples were obtained from five stations in total, separated at various distances,
but the Inner and Outer stations were located ~3.8 km apart. All samples were obtained as

14
vertical tows. The region includes two large leases for blue mussels (2.80 km2), which cover
24.8% of the bay area.
2.4 Adjustments to the taxa list
As mentioned, specimens were identified to a predetermined taxonomic level (Appendix 3).
Taxa were then corrected for minor discrepancies in names (e.g., fixed typos, removed
underscores, adjusted capitalization, etc.). Next, other adjustments were made to the list of
taxa for consistency in the dataset. Since this study focused on broad biodiversity patterns,
at each taxonomic level specified in Appendix 1, we combined the stages of the various taxa
to prevent the over-specification of stages in some taxa compared to others. For example,
Euphausiacea calyptopsis, furcilia and nauplii were combined as one taxa (Euphausiacea
larvae), as opposed to treating them as three separate taxa. This also helps overcome
discrepancies, where some taxa (e.g., Cirripedia) had larval stages separated in some
datasets (e.g., Cirripedia cypris and Cirripedia nauplii) but combined in others (Cirripedia
cypris/nauplii). Furthermore, invertebrate trochophores and eggs were difficult to decipher
and were often combined by the taxonomists. Therefore, Osteichthyes egg and larval stages
were also combined as one taxonomic group. A more detailed overview of additional
adjustments to the taxa list is provided in Appendix 4.
2.5 Converting counts to abundances in seawater
Raw mesozooplankton counts measured by the FlowCam were converted to abundances in
seawater (abund; individuals m-3) for each sample following the equation:
Abund. =  
 


(1)
Where counts represent the number of individuals identified by the taxonomist, fraction
analyzed represents the portion of the sample analyzed by the taxonomist, and water
volume is the volume filtered from each tow. Values were multiplied by 4 (the final term of
Equation 1) since the total obtained was split in 4 using a Folsom splitter and only 1 sample
was run through the FlowCam. For data in the Pacific region, zooplankton from two tows
were combined to gain a representative sample of the zooplankton community. Therefore,
the counts in Equation 1 were summed from two samples, and the water volume was the
sum of the water volumes from both tows.
All data cleaning and consolidating, and statistical analyses (next section) were conducted
using R v. 4.2.2 (R Core Team 2022). The code for all processes is publicly available at
https://github.com/AtlanticR/AMP.
2.6 Statistical analysis
2.6.1 Objective 1: Determining the optimal sampling effort per site
Hill numbers
While multiple biodiversity indices exist to characterize community structure, the use of Hill
numbers (Hill 1973) is facing a resurgence (Chao et al. 2014; Cox et al. 2017). This family of
indices has many advantages over other diversity indices (e.g., see Chao et al. (2014) or
Jost (2006) for a more thorough review). Hill Numbers integrate species richness and
incidence (i.e., presence) information into a unified class of diversity metrics, and are
parameterized by a diversity order q where higher values of q place an increased emphasis
on incidence frequency.
For incidence data, Hill Numbers are interpreted as the effective number of equally frequent
species (taxa), and are defined as (Chao et al. 2014):

15

󰇩 


 󰇪

 

(2)
where S represents the number of species (taxa), π denotes the incidence probability of the
ith taxon, and q determines the sensitivity of q to the relative frequencies of each taxon.
When q = 0, the equation considers taxa equally, regardless of their relative frequencies,
and therefore equates to richness (Hill 1973; Chiu and Chao 2014; Hsieh et al. 2016).
Since equation (2) is undefined when q = 1, the boundary value is used, giving:


  




  


 

(3)
which equates to Shannon diversity for incidence data. This is equivalent to the exponential
of Shannon entropy based on relative incidences in the assemblage, and can be interpreted
as the number of frequent taxa in a sample (Chao et al. 2020).
For q = 2, equation (2) becomes:

󰇧


 󰇨


(4)
which is the Simpson diversity for incidence data. This is equivalent to the inverse Simpson
concentration based on relative incidences, and is interpreted as the number of highly
frequent taxa in a sample (Chao et al. 2020).
Rarefaction and extrapolation curves
Because the specimens were identified to the lowest taxonomic level and represent mixed
taxonomic levels, as is common in zooplankton studies (e.g., Schartau et al. 2021; Gutierrez
et al. 2022), the biodiversity indices were calculated using these levels and the terminology
was updated to reflect these terms (e.g., “taxonomic richness” was used in place of “species
richness”). For each site, sample-based rarefaction and extrapolation curves were
constructed to explore how taxa diversity increases with sampling effort. For sample-based
rarefaction, data within each sample (i.e., zooplankton tow) are first converted to incidences
(i.e., presence-absence). For richness, a taxa accumulation curve is then created by
randomly selecting samples, and plotting the total number of taxa detected (y-axis) as a
function of sample size(x-axis). When the process of randomly resampling is repeated
multiple times, the resulting plot is referred to as a rarefaction curve, which represents the
statistical average of multiple taxa accumulation curves (Chiarucci et al. 2008). At first, the
curve generally rises steeply as each new sample results in new taxa being observed. As
more samples are analyzed, the slope often levels off as fewer new taxa are encountered.
These curves can also be extrapolated to show the predicted change in diversity beyond
what was collected (Chao et al. 2014).
Rarefaction curves therefore provide important information for monitoring programs. First,
taxa richness is largely affected by sampling effort, since more samples will result in new
taxa being encountered (Colwell et al. 2012). Richness is typically not comparable among
sites with a different number of samples obtained, since the differences between sites may
be more reflective of sampling effort, rather than true differences in richness (Colwell et al.
2012). Rarefaction therefore offers a means to standardize datasets to enable these
comparisons (Chao et al. 2014). Second, these curves can also be used to visualize how
much new information is gained (i.e., new taxa observed) with additional sampling. For

16
example, if the taxa accumulation curve reaches or nears an asymptote (i.e., few new taxa
are being observed with each sample obtained), collecting more samples beyond this point
may not be best use of resources, as little new information is gained. Predicted taxa richness
within a site can also be calculated (i.e., asymptotic estimators, or Chao2 for incidence
richness data; Chao 1984, 1987).
Rarefaction and extrapolation curves for richness, Shannon diversity and Simpson diversity
were created, following the methods of Hsieh et al. (2016). Extrapolations were calculated
and plotted up to double the sample size of each site, and 500 bootstrap replicates were
used to estimate the 95% confidence intervals. These were not extrapolated further, as the
extrapolations become less reliable beyond this point (Chao et al., 2014). Asymptotic
estimators (i.e., the predicted diversity value in the bay) were calculated for each of the Hill
numbers, with 95% confidence intervals (see Chao et al. 2014 for derivations). Often,
monitoring programs base the required sampling effort on taxa richness alone (Chao et al.
2009); however, the rarefaction curves are provided for Shannon and Simpson diversity as
they provide additional biodiversity information for each site (Chao et al. 2014; Chao et al.
2020). Sampling completeness was then calculated for richness as the ratio of observed to
estimated richness to assess the extent of undetected diversity, and an additional graph was
created to visualize the change in sampling completeness with increasing sampling effort.
For brevity, only text descriptions for richness and its associated metrics (e.g., asymptotic
richness and completeness) are included. Statistics are provided for the Shannon and
Simpson indices in graphical and tabular format for a comprehensive analysis of biodiversity
trends, but are not discussed in detail. In addition, this represents the first baseline study of
diversity of zooplankton in coastal bivalve aquaculture sites in Canada. A fixed and absolute
sampling target is therefore not provided, as these targets are often arbitrary (e.g., Chacoff
et al., 2012). Instead, graphs visualizing the change in sampling completeness for each
additional tow were provided, up to double the sampling size. These graphs are provided as
guidance for developing optimal sampling plans in future years, recognizing that more
samples are required to capture a greater number of rare taxa, which may play important
roles in ecosystem functioning (Cao et al. 1998). We also recommend recreating these
rarefaction curves in subsequent years as more data are added, to ensure these patterns
remain representative. All rarefaction and extrapolation curves were generated using the
iNEXT package (Hsieh et al. 2016).
While sample-based rarefaction is generally conducted using a standardized area or volume
(Gotelli and Colwell 2001), this is not feasible for samples collected with zooplankton tows,
as it is difficult to obtain a similar volume of water in each sample. To account for the
differences in water volume filtered, the x-axes of each graph were rescaled to reflect the
representative water volume filtered and analyzed from each sample. To obtain this value,
the water volume of the sample was multiplied by the fraction analyzed (see Equation 1).
This representative water volume provides a more comparable measure of effort among
sites than a “zooplankton tow”. Some R packages have arguments to directly incorporate
these differences in water volume for each sample (e.g., the weights argument in the
specaccum function from the R vegan package; Oksanen et al. 2022; applied in Bessey et
al. 2020). However, setting the weights argument does not affect the shape of the
interpolated curve, only the associated error (Oksanen et al. 2022). Therefore, although the
iNEXT package does not currently have these capabilities, we still used it for this work due
to its more robust calculations for extrapolations (Chao et al. 2014), and integrated use with
Hill Numbers. However, we did not proceed with significance testing among sites (e.g., if
richness is significantly different among sites), which is often evaluated using overlapping
confidence intervals (Chao et al. 2014; Chao et al. 2020), since these would therefore be
affected by differences in water volume.

17
2.6.2 Objective 2: Characterizing patterns in zooplankton community structure
among regions, months, and sites
As an initial graphical approach to visualize biodiversity patterns, Venn diagrams were
constructed to show the number and percent overlap of taxa among sites for each region.
The Venn diagrams were reflective of the true sampling effort, and sites with a greater
number of samples may therefore show a greater number of taxa (Colwell et al. 2012). The
Venn diagrams were created using the Venn and process_data functions from the
ggVennDiagram package (Gao 2022), and the outputs were visualized using the ggplot
function from the ggplot2 package (Wickham 2016).
Next, as a measure to graphically display sample similarity based on zooplankton
composition, ordinations were constructed using two-dimensional non-metric
multidimensional scaling (NMDS). NMDS ordinations visually display information in a
similarity or distance matrix (Borg and Groenen 2005), and samples with similar zooplankton
composition (i.e., taxa and their abundance), are visualized closer together, while more
dissimilar samples are located further apart. Various ordinations were therefore created to
examine sample similarity at various spatial scales, including:
1. An NMDS with all sampling data, to visualize similarities and differences in
composition among regions;
2. An NMDS ordination for regions exhibiting a high degree of overlap (identified in the
ordination above), to visualize which sites are responsible for the similarity; and
3. Separate NMDS ordinations for each region, to visually examine trends among sites
(for Maritimes and Gulf regions) or months (for Pacific and Newfoundland data).
Ordinations were constructed using a Bray-Curtis dissimilarity matrix of square root
transformed abundance data (i.e., counts per sample were converted to ind m-3 in seawater
following Equation 1.) The overall goodness-of-fit is provided by the Kruskall’s stress (1964)
which measures how well the square root transformed Bray-Curtis dissimilarity matrix can be
displayed in two dimensions. Values range from 0 to 1 with values closer to 0 indicating a
better fit (Clarke 1993). Ordinations were created using the metaMDS function from the
vegan package (Oksanen et al. 2022).
Tests for homogeneity of multivariate dispersion were conducted among regions, sites, and
months. These tests evaluate the null hypothesis of no differences in dispersion between
each observation and the group’s spatial median, an alternative form for the group centroid
(Anderson 2006a). If tests in multivariate dispersion are significant, a significant result from
the permutational multivariate analysis of variance (PERMANOVA, explained in more detail
below) may therefore be the result of differences in multivariate dispersion (Anderson
2006a). However, tests for homogeneity of multivariate dispersion provide useful information
on the biodiversity of communities in their own right, and multivariate dispersion has been
suggested as a measure of beta diversity (Anderson et al. 2006b). Tests were run using the
betadisper function from the vegan package in R, with the permutest function to determine
significance (Oksanen et al. 2022). Pairwise comparisons of group mean dispersions were
assessed using permutest.betadisper to test for differences between individual regions or
sites. Boxplots were then created to visualize the distances of each observation to the
spatial median for each group.
PERMANOVA was then used to test the significance of groups (i.e., regions, sites, and/or
months, as specified above) with the adonis2 function. PERMANOVA tests the null
hypothesis that centroids of all groups are equivalent using the chosen dissimilarity matrix
(Anderson 2017). The resulting test statistic is a pseudo-F statistic, where larger values
indicate a more pronounced separation among groups (Anderson 2017). When significant,
pairwise comparisons between each individual group were conducted using the
pairwise.adonis2 function from the pairwiseAdonis package (Arbizu 2017). The function for
pairwise comparisons similarly returns pseudo-F values, which were converted to pseudo-t

18
values by taking the square root of the pseudo-F value (Anderson 2008). While pseudo-F
values can be used, pseudo-t values are instead presented as they provide a more natural
statistic for these comparisons, and are a direct analogue to t-values in standard univariate
post-hoc testing (Anderson 2008). PERMANOVA and tests for homogeneity of multivariate
dispersion were performed on a Bray-Curtis dissimilarities matrix of square root transformed
abundance data, and significance was determined with 9999 permutations of the input data.
Lastly, when significant (P<0.05) differences were identified by the pairwise PERMANOVAs,
similarity percentage (SIMPER) analysis was used to identify the zooplankton taxa that
contributed most to the dissimilarities between significant groups. SIMPER was run using the
simper function from the vegan package (Oksanen et al. 2022).
2.6.3 Objective 3: Characterizing the role of tide phase and station on zooplankton
composition
As an initial approach to explore the role of tide phase and stations on zooplankton
composition, relative abundance charts were created to show the percent breakdown of the
zooplankton taxa for each sample. Samples were grouped by both tide phase and station as
facets (panels) using the facet_nested function from the ggh4x library (Brand 2022). To
enable visual comparisons and avoid displaying too many classes, a maximum of eight
classes were displayed; therefore, the top seven most abundant taxa in each bay were
displayed; all other taxa were combined into the category “Other”.
Next, a similar approach to the previous objective was pursued to explore patterns in
zooplankton composition using multivariate statistics. NMDS ordinations were constructed
for each site using the Bray-Curtis dissimilarities of square root transformed abundance, and
samples were displayed with colour and symbols to explore the role of station and tide
phase, respectively.
Within several sites, the effect of tide phase and station on zooplankton composition was
evaluated using PERMANOVAs. However, when the number of observations is low, there
may not be enough unique permutations to make statistical inferences (Anderson 2008;
Oksanen et al. 2022). Following the guidance of Anderson et al. (2008), models were not
constructed for bays which would result in PERMANOVAs with <100 unique permutations,
since at this level, the smallest possible P-value obtained is 1/100 = 0.01; a common
threshold for significance testing. For a balanced design with a groups and n samples per
group, the number of distinct outcomes (i.e., unique permutations) for the PERMANOVA
Pseudo-F statistic is (an)!/[a!(n!)a] (Clarke 1993). Accordingly, for post-hoc testing, two
groups of five samples would be the lowest number of samples per group when 100 unique
permutations are used as a cutoff (since (10)!/[2!(5!)2] = 126 unique permutations, whereas
two groups of four would give (8)!/[2!(4!)2] = 70 unique permutations.) Although there are
other approaches to test for significance with lower sample sizes (e.g., Monte Carlo
permutations; Anderson and Robinson 2003), these were not pursued since zooplankton
distributions are highly patchy (Folt and Burns 1999; O’Brien and Oakes 2020), and focusing
on the results of significance testing on such few samples may not lead to ecologically
relevant conclusions.
PERMANOVAs were therefore run in sites with at least two groups of five for the post-hoc
testing (e.g., at least five samples per station, or individual tide phase). Tests were therefore
performed on data from Argyle, St. Peters, the August 2020 and June 2021 sampling
months from the Pacific region, and for October 2021 from Newfoundland. The effects of tide
phase, station, and their two-way interaction on mesozooplankton composition were
evaluated by treating variables as factors (with levels High Tide or Low Tide, and station
labels according to Appendix 1), which were added sequentially to the model. Tide effects
were only tested using samples collected from high and low tide, and not those from mid-

19
rising or mid-falling tide phases. Post-hoc pairwise comparisons were then conducted using
the pairwise.adonis2 function, for factors determined as significant by the PERMANOVA.
For the same sites and months where PERMANOVAs were conducted, tests for
homogeneity of multivariate group dispersion were also run prior to each PERMANOVA to
evaluate the effects of both tide phase and station. These analyses are restricted to one-way
tests (Robertson et al. 2013; Oksanen et al. 2022); therefore, separate models were run to
evaluate the effects of tide and station. As in objective 2, these were conducted using 9999
permutations on Bray-Curtis dissimilarities of square root transformed abundance. Lastly,
SIMPER analysis was conducted for each significantly (P<0.05) distinct group, as identified
by the post-hoc pairwise PERMANOVA tests, to identify taxa most responsible for the
difference in groupings.
As an additional measure to evaluate the role of tides on mesozooplankton communities,
two-sample t-tests were conducted to test for differences in abundance, and the three Hill
Numbers (i.e., taxa richness, Shannon diversity, and Simpson diversity) between high and
low tide phases. Shannon diversity was calculated as the exponential of the Shannon index,
and Simpson diversity was represented as the inverse Simpson index (Jost 2006), using
values calculated from the diversity function in the vegan package (Oksanen et al. 2022).
Tests were conducted for stations within different sites that had at least three samples for
each tide phases, which therefore included: Inner, Mid and Outer stations in Lemmens
August 2020, the Mid station for Lemmens June 2021, the Outer station for Sober Island, the
Inner, Mid and Outer stations at St. Peters, and the Outer station for South Arm October
2021. Tests were conducted using the t.test function from the lsr R package (Navarro 2015).

20
3 RESULTS
3.1 Overview of images per site
The final taxa list (following minor adjustments outlined in Appendix 4) included 70 unique
taxa, comprising 1 species, 28 genera, 8 families, 11 orders, 2 suborders, 2 infraorders, 6
classes, 1 superclass, 2 subclasses, 1 subphylum, 7 phyla, and 1 paraphyletic group.
Overall, 384,593 individual zooplankton images were identified by the macro-FlowCam
including 47,605 from the Pacific region, 60,786 from the Maritimes region, 173,225 from the
Gulf region, and 102,977 from Newfoundland (Table 2).
Table 2. Number of mesozooplankton (0.25 mm - 5.00 mm) images identified by the macro-
FlowCam for each site and sampling period.
Region
Site
Count
Pacific
Lemmens August 2020
14,127
Lemmens March 2021
1,999
Lemmens June 2021
20,016
Lemmens September 2021
11,463
Total
47,605
Maritimes
Argyle
17,526
Country Harbour
9,849
Sober Island
18,763
Whitehead
14,648
Total
60,786
Gulf
Cocagne
24,510
Malpeque
17,572
St Peters
131,143
Total
173,225
Newfoundland
South Arm (all data)
102,977
All data
Total (all regions)
384,593
3.2 Objective 1: Determining the optimal sampling effort per site
3.2.1 Pacific region
In Lemmens Inlet, taxa richness was highest in September 2021 (39) and similar in August
2020 (35) and June 2021 (35) (Fig. 2, Table 3). The rarefaction and extrapolation curves
continued to increase for August 2020 and September 2021 up to double the sample size,

21
and these months had asymptotic richness estimates of 39.25 and 44.73, respectively (Fig.
2, Table 3). The extrapolations appeared to level off for June 2021, which had a richness
estimate of 35.47 (Fig. 2, Table 3). Sampling completeness based on richness values was
89.17% (August 2020), 98.67% (June 2021), and 87.19% (September 2021) (Table 3).
Figure 2. Sample-based rarefaction (solid line) and extrapolation (dashed line) of
mesozooplankton taxa richness (q = 0), Shannon diversity (q = 1), and Simpson diversity (q
= 2) for sampling months within Lemmens Inlet (Pacific region). Extrapolations were plotted
up to double the sample size. The cumulative water volume on the secondary (upper) x-axis
shows the representative volume of water analyzed per sample, accounting for subsampling
and fractions analyzed by the taxonomist (i.e., Rep. vol. in Table 3). Each sample includes
the combined data from two separate zooplankton tows. 500 bootstrap replicates were used
to estimate the 95% confidence intervals (shaded area), although the true confidence
intervals are likely to be larger, given the variability in water volume per tow (Oksanen et al.
2022). Data from March 2021 (n = 2) are not shown.
Table 3. Sample-based rarefaction statistics for sampling months within Lemmens Inlet
(Pacific region). Tow volume (tow vol., m3) represents the average volume of water sampled
from the zooplankton tows. For the Pacific region, one sample includes the total volume from
two combined tows. The percent analyzed corresponds to the average percentage of the
sample analyzed by a taxonomist. The representative volume (rep. vol) denotes the amount

22
of seawater analyzed, accounting for subsampling and fractions analyzed by the taxonomist
(see section 2.6.1 in the main text). Observed is the observed diversity value, Estimate is the
estimated (i.e., asymptotic) diversity, and Completeness (%) represents the ratio of the
observed to estimated values (calculated for richness only). Data from March 2021 (n = 2)
were not included in the analysis.
Month
Tow vol.
(m3) (sd)
Analyzed
(%) (sd)
Rep.
vol.
(m3)
Diversity
Observed
Estimate
Completeness
Aug
2020
10.35
(1.86)
46.80
(16.25)
1.21
Taxa
richness
35.00
39.25
89.17%
Shanno
n
diversity
28.85
29.56
Simpson
diversity
26.47
26.82
Jun
2021
15.21
(4.38)
11.02
(4.03)
0.42
Taxa
richness
35.00
35.47
98.67%
Shanno
n
diversity
29.12
29.61
Simpson
diversity
27.12
27.36
Sept
2021
19.76
(4.41)
79.81
(27.34)
3.54
Taxa
richness
39.00
44.73
87.19%
Shanno
n
diversity
32.76
33.91
Simpson
diversity
30.44
30.97
3.2.2 Maritimes region
In the Maritimes region, observed and estimated (in brackets) taxa richness was 26 (33.47)
in Argyle, 31 (36.00) in Country Harbour, 28 (33.50) in Sober Island, and 27 (27.15) in
Whitehead (Fig. 3, Table 4). The curves continued to increase when extrapolated up to
double the sample size for Argyle, Country Harbour, and Sober Island, which had sampling
completeness of 77.68%, 86.11% and 83.58%, respectively (Fig. 3, Table 4). The
extrapolations appeared to level off for Whitehead, which had the highest sampling
completeness of 99.45% (Fig. 3, Table 4).

23
Figure 3. Sample-based rarefaction (solid line) and extrapolation (dashed line) of
mesozooplankton taxa richness (q = 0), Shannon diversity (q = 1), and Simpson diversity (q
= 2) for sites within the Maritimes region. Extrapolations were plotted up to double the
sample size. The cumulative water volume on the secondary (upper) x-axis shows the
representative volume of water analyzed per sample, accounting for subsampling and
fractions analyzed by the taxonomist (i.e., Rep. vol. in Table 4). 500 bootstrap replicates
were used to estimate the 95% confidence intervals (shaded area), although the true
confidence intervals are likely to be larger, given the variability in water volume per sample
(Oksanen et al. 2022).

24
Table 4. Sample-based rarefaction statistics for sites within the Maritimes region. Tow
volume (tow vol., m3) represents the average volume of water sampled from the zooplankton
tows. The percent analyzed corresponds to the average percentage of the sample analyzed
by a taxonomist. The representative volume (rep. vol) denotes the amount of seawater
analyzed, accounting for subsampling and fractions analyzed by the taxonomist (see section
2.6.1 in the main text). Observed is the observed diversity value, Estimate is the estimated
(i.e., asymptotic) diversity., and Completeness (%) represents the ratio of the observed to
estimated values (calculated for richness only).
Site
Tow vol.
(m3) (sd)
Analyzed
(%) (sd)
Rep.
vol. (m3)
Diversity
Obser
ved
Estima
te
Completeness
Argyle
5.31
(4.51)
90.40
(20.32)
1.20
Taxa
richness
26.00
33.47
77.68%
Shannon
diversity
20.06
21.05
Simpson
diversity
17.74
18.08
Country
Harbour
4.06
(0.87)
34.24
(32.70)
0.35
Taxa
richness
31.00
36.00
86.11%
Shannon
diversity
26.96
28.72
Simpson
diversity
25.18
25.93
Sober
Island
7.18
(3.37)
35.29
(28.08)
0.63
Taxa
richness
28.00
33.50
83.58%
Shannon
diversity
21.64
23.17
Simpson
diversity
18.97
19.65
Whitehe
ad
2.38
(0.76)
47.51
(24.68)
0.28
Taxa
richness
27.00
27.15
99.45%
Shannon
diversity
24.20
24.75
Simpson
diversity
22.86
23.31
3.2.3 Gulf region
In the Gulf region, observed and estimated (in brackets) taxa richness was 26 (37.25) in
Cocagne, 31 (37.00) in Malpeque, and 35 (35.54) in St. Peters (Fig. 4, Table 5). The curves
continued to increase when extrapolated up to double the sample size for Cocagne and
Malpeque, which had sampling completeness of 68.80% and 83.78%, respectively (Fig. 4,

25
Table 5). The extrapolations appeared to level off for St. Peters, which had the highest
sampling completeness of 98.48% (Fig. 4, Table 5).
Figure 4. Sample-based rarefaction (solid line) and extrapolation (dashed line) of
mesozooplankton taxa richness (q = 0), Shannon diversity (q = 1), and Simpson diversity (q
= 2) for sites within the Gulf region. Extrapolations were plotted up to double the sample
size. The cumulative water volume on the secondary (upper) x-axis shows the
representative volume of water analyzed per sample, accounting for subsampling and
fractions analyzed by the taxonomist (i.e., Rep. vol. in Table 5). 500 bootstrap replicates
were used to estimate the 95% confidence intervals (shaded area), although true confidence
intervals are likely to be larger, given the variability in water volume per sample (Oksanen et
al. 2022).
Table 5. Sample-based rarefaction statistics for sites within the Gulf region. Tow volume (tow
vol., m3) represents the average volume of water sampled from the zooplankton tows. The
percent analyzed corresponds to the average percentage of the sample analyzed by a
taxonomist. The representative volume (rep. vol) denotes the amount of seawater analyzed,
accounting for subsampling and fractions analyzed by the taxonomist (see section 2.6.1 in
the main text). Observed is the observed diversity value, Estimate is the estimated (i.e.,

26
asymptotic) diversity, and Completeness (%) represents the ratio of the observed to
estimated values (calculated for richness only).
Site
Tow vol.
(m3) (sd)
Analyzed
(%) (sd)
Rep. vol.
(m3)
Diversity
Obser
ved
Estim
ate
Completeness
Cocagne
12.76
(2.89)
24.86
(37.41)
0.79
Taxa
richness
26.00
37.25
68.80%
Shannon
diversity
21.51
25.87
Simpson
diversity
19.00
21.11
Malpeque
34.98
(6.55)
16.75
(3.04)
1.46
Taxa
richness
31.00
37.00
83.78%
Shannon
diversity
29.22
31.75
Simpson
diversity
28.24
29.30
St. Peters
73.86
(14.62)
13.22
(10.7)
2.44
Taxa
richness
35.00
35.54
98.48%
Shannon
diversity
24.14
24.86
Simpson
diversity
20.59
20.92
3.2.4 Newfoundland region
In Newfoundland, observed and estimated (in brackets) taxa richness was 28 (50.05) in
September 2020 and 32 (35.67) in October 2021 (Fig. 5, Table 6). The curves continued to
increase when extrapolated up to double the sample size for both months, which had
sampling completeness of 55.94% and 89.71%, respectively (Fig. 5, Table 6).

27
Figure 5. Sample-based rarefaction (solid line) and extrapolation (dashed line) of
mesozooplankton taxa richness (q = 0), Shannon diversity (q = 1), and Simpson diversity (q
= 2) for two sampling months within South Arm (Newfoundland region). Extrapolations were
plotted up to double the sample size. The cumulative water volume on the secondary (upper)
x-axis shows the representative volume of water analyzed per sample, accounting for
subsampling and fractions analyzed by the taxonomist (i.e., Rep. vol. in Table 6). 500
bootstrap replicates were used to estimate the 95% confidence intervals (shaded area),
although true confidence intervals are likely to be larger, given the variability in water volume
per sample (Oksanen et al., 2022).

28
Table 6. Sample-based rarefaction statistics for sampling months within South Arm,
Newfoundland. Tow volume (tow vol., m3) represents the average volume of water sampled
from the zooplankton tows. The percent analyzed corresponds to the average percentage of
the sample analyzed by a taxonomist. The representative volume (rep. vol) denotes the
amount of seawater analyzed, accounting for subsampling and fractions analyzed by the
taxonomist (see section 2.6.1 in the main text). Observed is the observed diversity value,
Estimate represents the estimated (i.e., asymptotic) diversity, and Completeness (%)
represents the ratio of the observed to estimated values (calculated for richness only).
Site
Tow vol.
(m3) (sd)
Analyzed
(%) (sd)
Rep. vol.
(m3)
Diversity
Obse
rved
Estim
ate
Completeness
Sept
2020
5.54
(2.37)
100.00
(0.00)
1.39
Taxa
richness
28.00
50.05
55.94%
Shannon
diversity
22.17
24.10
Simpson
diversity
20.18
20.56
Oct
2021
8.83
(0.43)
65.11
(5.59)
1.44
Taxa
richness
32.00
35.67
89.71%
Shannon
diversity
27.17
27.96
Simpson
diversity
25.52
25.87
3.2.5 All regions
Sampling completeness varied for each site, with the highest values observed in Whitehead
(99.45%, Maritimes region), Lemmens June 2021 (98.67%, Pacific), and St. Peters (98.48%,
Gulf), which were obtained from collecting 9, 16 and 26 samples, respectively (Fig. 6). With
the exception of Argyle (Maritimes) and South Arm September 2020 (Newfoundland), a high
degree of sampling completeness of (e.g., 80%) could be reached by collecting 10 or fewer
samples, while 90% completeness could be reached with 20 or fewer samples (Fig. 6). For
Argyle, a sampling completeness of 80% could be obtained with ~17 samples, while >30
samples (i.e., beyond the extrapolation range) would be required to reach 90%
completeness (Fig. 6). For South Arm, >20 samples would be required to reach 80% or 90%
completeness (Fig. 6).

29
Figure 6. Sample-based rarefaction (solid line) and extrapolation (dashed line) curves for all
sites collected as part of the Aquaculture Monitoring Program. The y-axis represents the
sampling completeness, or the richness of each site divided by its asymptotic estimate (i.e.,
Chao2 estimator), to visualize how many samples are required to obtain a select percentage
of the estimated richness within each site.
3.3 Objective 2: Characterizing patterns in zooplankton community structure among
regions, months, and sites
3.3.1 Regional comparisons
Overall, a large percentage of taxa were observed in all sampling months in the Pacific
region (46.34%) (Fig. 7). 42.50% of taxa were observed in all sites in the Maritimes and
46.34% were shared in the Gulf regions (Fig. 7). The March and June 2021 sampling
months from Lemmens Inlet had no unique taxa, while the highest number of unique taxa
per site or sampling month was 4, which occurred in Country Harbour (Maritimes region),
Cocagne (Gulf region) and September 2021 in Lemmens Inlet (Pacific region).

30
Figure 7. Venn diagrams showing the overlap in the observed number of zooplankton taxa
for (A) sampling months within the Pacific region, (B) sites within the Maritimes region, and
(C) sites within the Gulf region. Venn diagrams were not constructed for data collected within
the Newfoundland region.
The NMDS ordination constructed using all data collected shows groupings based on AMP
regions (Fig. 8A). Broadly, data from the Gulf region exhibit high variability and show some
degree of overlap with the Maritimes region. Samples from the Pacific and Newfoundland
regions are generally distinct and occupy the upper right and lower right portions of the
NMDS ordination, respectively (Fig. 8A). When coloured based on sampling month (Fig. 8B),
some temporal patterns emerge. Samples from the Gulf and Maritimes (collected from July
to September) do not appear to group based on sampling month, but samples from the
Pacific region appear to transition somewhat chronologically, moving from June to August to
September to March (left to right) on the NMDS. An approximate cyclical progression by
month was observed in the samples from the Newfoundland region (Fig. 8B).

31
Figure 8. Two-dimensional non-metric multidimensional scaling ordination showing similarity
in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples collected as
part of the Aquaculture Monitoring Program (AMP). The ordination in (A) represents samples
coloured by AMP region, and (B) shows samples coloured by collection month. AMP regions
in both ordinations are denoted by symbols. The ordinations were constructed using a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Grey lines in (A) show the connections between each sample and the centroid of its
corresponding AMP region.
Because of the high degree of overlap between the samples from the Maritimes and Gulf
regions (Fig. 8), a second NMDS ordination was constructed to visualize which sites were
similar between the regions (Fig. 9). The data from St. Peters show a high degree of
dispersion, while samples from Cocagne and Malpeque are more separated (Fig. 9).
Generally, the Gulf samples surrounded the Maritimes samples in the ordination, rather than
being directly mixed (Fig. 9).

32
Figure 9. Two-dimensional non-metric multidimensional scaling ordination showing similarity
in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples from the Gulf
(circles) and Maritimes (squares) region collected as part of the Aquaculture Monitoring
Program (AMP). Colours represent the sampling site (e.g., bay or inlet) within each region.
The ordination was constructed using a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton abundance (ind m-3).
When data from all regions were analyzed, significant differences in multivariate dispersion
were observed among regions (F3, 174 = 6.031 P(perm) = 0.002) (Table 7, Fig. 10). Pairwise
comparisons revealed that samples from the Gulf were significantly higher (P<0.05) in
multivariate dispersion relative to other regions (Table 7, Fig. 10). Samples from the
Maritimes, Newfoundland, and Pacific regions were not significantly different in multivariate
dispersion (Table 7, Fig. 10).
The PERMANOVA indicated statistically significant differences in mesozooplankton
assemblage structure among regions (pseudo-F3, 174 = 45.179, P(perm) = 0.0001), and
region (treated as a factor) explained 43.79% of the variation in the data (Table 8). Pairwise
comparisons revealed significant differences (P<0.05) in assemblage structure between all
regions, and differences between the Pacific and Newfoundland (pseudo-t = 8.321, largest t-
value) and Pacific and Maritimes (pseudo-t = 8.143) were the greatest, while differences
between the Gulf and Maritimes were the lowest (pseudo-t = 3.172, lowest t-value) (Table
8).
The SIMPER analysis also identified that regions with the greatest geographic separation
had higher overall average dissimilarity (av. dissim.) than regions closer in distance (Table
9). In particular, the mesozooplankton communities of the Pacific and Gulf regions were
identified as the most different in assemblage structure (av. dissim. 74.67%) while the
Maritimes and Gulf were identified as the least different (av. dissim. 58.38%). Various taxa

33
were responsible for the differences in composition between regions and the list of the top
five are presented in Table 9.
Table 7. Summary of the multivariate homogeneity of group dispersions analysis of
mesozooplankton (0.25 mm - 5.00 mm) assemblage structure between regions within the
Aquaculture Monitoring Program. Results are based on a Bray-Curtis dissimilarity matrix of
square root transformed mesozooplankton abundance (ind m-3). The analysis was followed
by a posteriori pairwise comparisons between individual regions.
Source
df
SS
MS
F
P(perm)
Region
3
0.191
0.064
6.031
0.0002
Residuals
174
1.834
0.011
Total
177
2.025
Comparison
t
P(perm)
Gulf - Maritimes
2.802
0.0052
Gulf - Newfoundland
3.307
0.0009
Gulf - Pacific
3.446
0.0011
Maritimes - Newfoundland
0.813
0.4118
Maritimes - Pacific
1.085
0.2825
Newfoundland - Pacific
0.256
0.7974
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations; t: t-value from Student’s t-test. Significant
effects are shown in bold (P(perm < 0.05)).

34
Figure 10. Boxplots depicting the distance of samples to the centroid of the corresponding
region within the Aquaculture Monitoring Program (AMP). Results were obtained from the
multivariate homogeneity of group dispersions analysis, based on a Bray-Curtis dissimilarity
matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-
3). Boxes show the first, second and third quartiles, and are coloured to represent AMP
regions. Vertical lines extending from the boxes indicate the minimum and maximum values
up to 1.5 times the interquartile range. Jittered points represent the values for individual
samples. *P(perm)<0.05, **P(perm)<0.01, ***P(perm)<0.001; i.e., significance obtained from
pairwise comparisons between regions using 9999 permutations of the input data.
Table 8. Permutational Multivariate Analysis of Variance (PERMANOVA) results based on a
Bray-Curtis dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00
mm) abundance (ind m-3) for regions within the Aquaculture Monitoring Program. The
analysis was followed by a posteriori pairwise comparisons between individual regions.
Pseudo-t values for the pairwise comparisons were calculated as the square root of the
Pseudo-F statistic generated from the pairwise.adonis2 R function (Arbizu, 2020).
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Region
3
16.465
5.488
43.787
45.179
0.0001
Residuals
174
21.137
0.121
56.213
Total
177
37.602
100.000
Comparison
Pseudo-t
P(perm)
Gulf - Maritimes
3.172
0.0001
Gulf - Newfoundland
6.332
0.0001
Gulf - Pacific
6.650
0.0001
Maritimes -
Newfoundland
6.642
0.0001
Maritimes - Pacific
8.143
0.0001
Newfoundland - Pacific
8.321
0.0001
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are shown in bold (P(perm < 0.05)).
Table 9. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between regions. Values in the second column (average) represent the percent contribution
of each taxon to average between-group dissimilarity, and overall average dissimilarity (av.
dissim., %) represents the sum of these values. The third column (cont., %) is based on
average (second column), but adjusted to sum to 100%, and the fourth column represents
the cumulative contribution (c. cont) of these values. The fifth and sixth columns represent
the average abundance of each taxon within each region (square root transformed, ind m-3).
The permutation p-value represents the probability of getting a larger or equal average
contribution based on 999 random permutations of input data. Note that the full list of
contributions from all taxa is not shown, so the sum of each entry from average (column 2)

35
may not equal the overall average dissimilarity, and the cumulative contribution (column 4)
may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av. abund. (I)
Av. abund.
(II)
P(perm)
Av. dissim.:
58.38%
Maritimes
Gulf
Acartia spp.
14.12
24.18
24.18
51.54
59.58
0.040
Evadne spp.
4.55
7.79
31.97
14.18
2.86
0.082
Copepoda
(nauplii)
4.40
7.54
39.51
0.28
11.91
0.001
Centropages
spp.
3.39
5.81
45.32
10.03
5.23
0.001
Podon/Pleopsis
spp.
3.19
5.47
50.79
5.88
10.73
0.138
Av. dissim.:
72.14%
Newfoundland
Gulf
Acartia spp.
16.36
22.68
22.68
19.42
59.58
0.001
Temora spp.
6.57
9.11
31.79
17.72
3.62
0.001
Pseudocalanus
spp.
5.61
7.77
39.56
13.68
0.71
0.001
Copepoda
(nauplii)
4.58
6.35
45.91
1.69
11.91
0.001
Evadne spp.
4.54
6.29
52.2
11.27
2.86
0.051
Av. dissim.:
74.67%
Pacific
Gulf
Acartia spp.
18.29
24.49
24.49
15.08
59.58
0.001
Cirripedia
(larvae)
5.33
7.14
31.63
15.43
2.22
0.001

36
Taxa
Average
Cont.
C. cont.
Av. abund. (I)
Av. abund.
(II)
P(perm)
Copepoda
(nauplii)
4.56
6.11
37.74
1.37
11.91
0.001
Oikopleura spp.
4.50
6.02
43.77
11.87
0.28
0.001
Podon/Pleopsis
spp.
3.81
5.11
48.87
9.59
10.73
0.001
Av. dissim.:
64.08%
Maritimes
Newfoundla
nd
Acartia spp.
13.95
21.77
21.77
51.54
19.42
0.034
Temora spp.
6.19
9.66
31.43
6.71
17.72
0.001
Pseudocalanus
spp.
5.54
8.65
40.08
2.38
13.68
0.001
Evadne spp.
5.10
7.96
48.03
14.18
11.27
0.001
Centropages
spp.
3.19
4.98
53.01
10.03
2.08
0.001
Av. dissim.:
72.69%
Maritimes
Pacific
Acartia spp.
16.41
22.57
22.57
51.54
15.08
0.001
Cirripedia
(larvae)
5.91
8.13
30.70
0.40
15.43
0.001
Oikopleura spp.
4.73
6.51
37.21
0.15
11.87
0.001
Evadne spp.
4.35
5.99
43.19
14.18
5.57
0.181
Eurytemora
spp.
3.24
4.46
47.65
9.40
0.00
0.001
Av. dissim.:
70.24%
Newfoundland
Pacific

37
Temora spp.
8.37
11.92
11.92
17.72
0.00
0.001
Pseudocalanus
spp.
6.22
8.86
20.78
13.68
0.69
0.001
Acartia spp.
6.18
8.79
29.57
19.42
15.08
1.000
Cirripedia
(larvae)
6.00
8.54
38.12
1.10
15.43
0.001
Evadne spp.
4.28
6.09
44.21
11.27
5.57
0.223
3.3.2 Comparisons among sites or months
Pacific region
The NMDS ordinations for Pacific data showed that the samples exhibited distinct groupings
based on month (Fig. 11). The sampling months (treated as factor) exhibited statistically
significant differences in multivariate dispersion (F2, 43 = 5.351 P(perm) = 0.0092) (Table 10,
Fig. 12). Multivariate dispersion was significantly lower in June 2021 compared to August
2020 (t = -3.021, P(perm) = 0.0027) and was also significantly lower in June 2021 compared
to September 2021 (t = -3.413, P(perm) = 0.0014) (Table 10). Data from August 2020 and
September 2021 were not significantly different in multivariate dispersion (t = 0.327, P(perm)
= 0.7604) (Table 10).
PERMANOVA indicated significant differences in mesozooplankton assemblage structure
between sampling months for data collected in Lemmens Inlet (pseudo-F3, 174 = 45.179,
P(perm) = 0.0001) (Table 11). While the relative magnitude of differences were not evident
in the NMDS (Fig. 11), pairwise comparisons indicated that June 2021 and September 2021
samples differed most with respect to mesozooplankton assemblage structure (pseudo-t =
7.027, P(perm) = 0.001), followed by August 2020 and June 2021 (pseudo-t = 5.142,
P(perm) = 0.001), and the lowest differences were observed between August 2020 and
September 2021 (t = 3.089, P(perm) = 0.001) (Table 11).
Similar to the pairwise PERMANOVA results, SIMPER analysis found the overall average
dissimilarity was highest between September 2021 and June 2021 (av. dissim. 56.76%) and
lowest between September 2021 and August 2020 (av. dissim. 41.37%) (Table 12).
Cirripedia (barnacle larvae) was the taxon most responsible for differences in August 2020
(lower average abundance) to June 2021 (higher average abundance) (Table 12). Acartia
spp. was the taxon most responsible for the differentiations between September 2021
(lower) and August 2020 (higher), and also for September 2021 (lower) and June 2021
(higher) (Table 12).

38
Figure 11. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples
collected from different months within Lemmens Inlet (Pacific region). The ordination was
constructed using a Bray-Curtis dissimilarity matrix of square root transformed
mesozooplankton abundance (ind m-3). Grey lines show the connections between each
sample and the centroid of its corresponding month.
Table 10. Summary of the multivariate homogeneity of group dispersions analysis between
sampling months within Lemmens Inlet (Pacific region). Results are based on a Bray-Curtis
dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm)
abundance (ind m-3). Data from March 2021 (n = 2) were not included in the analysis.
Source
df
SS
MS
F
P(perm)
Month
2
0.052
0.026
5.351
0.0092
Residuals
43
0.207
0.005
Total
45
0.259
t
P(perm)
June 2021 - Aug 2020
-3.021
0.0027
Aug 2020- September 2021
0.327
0.7604
June 2021- September 2021
-3.413
0.0014
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations. Significant effects are shown in bold (P(perm <
0.05)).

39
Figure 12. Boxplots depicting the distance of samples to the centroid of the corresponding
sampling months within the Pacific region (Lemmens Inlet). Results were obtained from the
multivariate homogeneity of groups dispersions analysis based on a Bray-Curtis dissimilarity
matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-
3). Boxes show the first, second and third quartiles, and lines extending from the boxes
indicate the minimum and maximum values up to 1.5 times the interquartile range. Jittered
points represent the values for individual samples. *P(perm)<0.05, **P(perm)<0.01,
***P(perm)<0.001; i.e., significance obtained from pairwise comparisons between regions
using 9999 permutations of the input data. Data from March 2021 (n = 2) are not shown.
Table 11. Summary of the Permutational Multivariate Analysis of Variance (PERMANOVA)
between sampling months within the Pacific region (Lemmens Inlet). Results are based on a
Bray-Curtis dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00
mm) abundance (ind m-3), and was followed by a posteriori pairwise comparisons between
individual months. Pseudo-t values for the pairwise comparisons were calculated as the
square root of the Pseudo-F statistic generated from the pairwise.adonis2 R function (Arbizu,
2020). Data from March 2021 (n = 2) were not included in the analysis.
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Month
2
2.273
1.137
53.879
25.117
0.0001
Residual
43
1.946
0.045
46.121
Total
45
4.219
100.000
Comparison
Pseudo-t
P(perm)
Aug. 2020 – Sept. 2021
3.089
0.0001
Aug. 2020 - Jun. 2021
5.142
0.0001
Jun. 2021 – Sept. 2021
7.027
0.0001
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are shown in bold (P(perm < 0.05)).

40
Table 12. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between sampling months within the Pacific region (Lemmens Inlet). Values in the second
column (average) represent the percent contribution of each taxon to average between-
group dissimilarity, and overall average dissimilarity (av. dissim., %) represents the sum of
these values. The third column (cont., %) is based on average (second column), but
adjusted to sum to 100%, and the fourth column represents the cumulative contribution (c.
cont) of these values. The fifth and sixth columns represent the average abundance of each
taxon within each region (square root transformed, ind m-3). The permutation p-value
represents the probability of getting a larger or equal average contribution based on 999
random permutations of input data. Note that the full list of contributions from all taxa is not
shown, so the sum of each entry from average (column 2) may not equal the overall average
dissimilarity, and the cumulative contribution (column 4) may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av.
abund. (I)
Av. abund.
(II)
P(perm)
Av. dissim.: 44.58%
Aug 2020
Jun 2021
Cirripedia (larvae)
6.92
15.51
15.51
10.92
29.39
0.0010
Acartia spp.
6.59
14.79
30.30
10.81
29.46
0.0020
Oikopleura spp.
4.59
10.30
40.60
8.11
20.74
0.0030
Fritillaria spp.
2.93
6.58
47.18
2.15
10.14
0.0010
Evadne spp.
2.16
4.85
52.02
4.63
10.21
0.4940
Av. dissim.: 41.37%
Sept
2021
Aug 2020
Acartia spp.
4.63
11.20
11.20
3.82
10.81
0.9880
Echinodermata
(larvae)
3.63
8.77
19.96
5.65
7.45
0.0020
Podon/Pleopsis spp.
3.50
8.46
28.42
4.59
10.00
0.0320
Cirripedia (larvae)
3.40
8.21
36.64
5.72
10.92
1.0000
Oikopleura spp.
2.62
6.33
42.96
7.45
8.11
1.0000

41
Taxa
Average
Cont.
C. cont.
Av.
abund. (I)
Av. abund.
(II)
P(perm)
Av. dissim.: 56.76%
Sept
2021
Jun 2021
Acartia spp.
10.31
18.16
18.16
3.82
29.46
0.0010
Cirripedia (larvae)
9.73
17.14
35.30
5.72
29.39
0.0010
Oikopleura spp.
5.44
9.59
44.89
7.45
20.74
0.0010
Podon/Pleopsis spp.
3.82
6.72
51.61
4.59
14.04
0.0050
Evadne spp.
3.53
6.22
57.83
1.66
10.21
0.0010
Maritimes region
The NMDS ordinations for the Maritimes data showed distinct groupings of samples by sites
(Fig. 13). Sites were not significantly different in their multivariate dispersion (F3,38 = 1.205,
P(perm) = 0.3135) (Table 13, Fig. 14). PERMANOVA indicated statistically significant
differences in mesozooplankton assemblage structure among sites (pseudo-F3, 38 = 20.172,
P(perm) = 0.0001) and Sites (treated as a factor) explained 61.427% of the variation in the
data (Table 14). Pairwise comparisons indicated greatest differences between sites that
were most geographically separated (largest t values), and lowest differences between
nearby sites (lowest t values) (Table 14). For instance, the largest differences in assemblage
structure were observed between Whitehead and Argyle (t = 6.183, P(perm) = 0.0004; most
geographically separated), and the lowest differences were between Country Harbour and
Whitehead (t = 2.754, P(perm) = 0.004; least geographically separated) (Table 14).
SIMPER analysis also revealed larger differences in the overall average dissimilarity
between regions with greater geographic separation than those nearby (Table 15). The taxa
most responsible for differences between sites were either Evadne spp., Acartia spp., or
Hydrozoa (medusa) (Table 15).

42
Figure 13. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples
collected from sites within the Maritimes region. The ordination was constructed using a
Bray-Curtis dissimilarity matrix of square root transformed mesozooplankton abundance (ind
m-3). Grey lines show the connections between each sample and the centroid of its
corresponding site.
Table 13. Summary of the multivariate homogeneity of group dispersions analysis of
mesozooplankton (0.25 mm - 5.00 mm) assemblage structure between sites within the
Maritimes region, sampled as part of the Aquaculture Monitoring Program. Results are
based on a Bray-Curtis dissimilarity matrix of square root transformed mesozooplankton
abundance (ind m-3).
Source
df
SS
MS
F
P(perm)
Site
3
0.026
0.009
1.205
0.3135
Residuals
38
0.272
0.007
Total
41
0.298
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.

43
Figure 14. Boxplots depicting the distance of samples to the centroid of the corresponding
site within the Maritimes region. Results were obtained from the multivariate homogeneity of
groups dispersions analysis based on a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-3). Boxes show the
first, second and third quartiles, and lines extending from the boxes indicate the minimum
and maximum values up to 1.5 times the interquartile range. Jittered points represent the
values for individual samples.
Table 14. Summary of the Permutational Multivariate Analysis of Variance (PERMANOVA)
between sites within the Maritimes region. Results are based on a Bray-Curtis dissimilarity
matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-
3), and was followed by a posteriori pairwise comparisons between individual sites. Pseudo-t
values for the pairwise comparisons were calculated as the square root of the Pseudo-F
statistic generated from the pairwise.adonis2 R function (Arbizu, 2020).
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Site
3
2.875
0.958
61.427
20.172
0.0001
Residual
38
1.805
0.048
38.573
Total
41
4.680
100.000
Pseudo-t
P (perm)
Argyle - Country Harbour
5.118
0.0001
Argyle - Sober Island
4.055
0.0001
Argyle - Whitehead
6.183
0.0001
Country Harbour - Sober
Island
3.461
0.0002
Country Harbour - Whitehead
2.754
0.0004
Sober Island - Whitehead
4.100
0.0001

44
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are shown in bold (P(perm < 0.05)).
Table 15. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between sites within the Maritimes region. Values in the second column (average) represent
the percent contribution of each taxon to average between-group dissimilarity, and overall
average dissimilarity (av. dissim., %) represents the sum of these values. The third column
(cont., %) is based on average (second column), but adjusted to sum to 100%, and the
fourth column represents the cumulative contribution (c. cont) of these values. The fifth and
sixth columns represent the average abundance of each taxon within each region (square
root transformed, ind m-3). The permutation p-value represents the probability of getting a
larger or equal average contribution based on 999 random permutations of input data. Note
that the full list of contributions from all taxa is not shown, so the sum of each entry from
average (column 2) may not equal the overall average dissimilarity, and the cumulative
contribution (column 4) may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av. abund.
(I)
Av.
abund.
(II)
P(perm)
Av. dissim.: 60.78%
Country
Harbour
Argyle
Evadne spp.
10.16
16.71
16.71
34.67
1.84
0.0001
Acartia spp.
8.58
14.12
30.83
57.18
32.10
0.9233
Eurytemora spp.
6.21
10.21
41.04
24.41
4.24
0.0001
Temora spp.
6.12
10.07
51.11
20.64
2.72
0.0001
Gastropoda
(larvae/Limacina)
4.63
7.62
58.74
17.26
3.03
0.0001
Av. dissim.: 47.31%
Sober
Island
Argyle
Acartia spp.
18.10
38.25
38.25
61.58
32.10
0.0001
Evadne spp.
7.43
15.70
53.95
14.87
1.84
0.0001
Centropages spp.
2.22
4.68
58.63
3.37
6.76
0.9993

45
Taxa
Average
Cont.
C. cont.
Av. abund.
(I)
Av.
abund.
(II)
P(perm)
Pseudodiaptomus
spp.
2.12
4.49
63.12
1.70
5.22
0.0001
Eurytemora spp.
1.94
4.09
67.21
1.32
4.24
1.0000
Av. dissim.: 57.75%
Whitehead
Argyle
Acartia spp.
11.88
20.57
20.57
66.78
32.10
0.2138
Hydrozoa (medusa)
8.75
15.14
35.71
26.13
0.39
0.0001
Evadne spp.
6.24
10.81
46.52
20.16
1.84
0.0254
Centropages spp.
6.18
10.70
57.22
24.96
6.76
0.0001
Eurytemora spp.
5.03
8.71
65.93
18.76
4.24
0.0007
Av. dissim.: 50.46%
Country
Harbour
Sober
Island
Acartia spp.
7.02
13.91
13.91
57.18
61.58
0.9917
Eurytemora spp.
6.34
12.56
26.46
24.41
1.32
0.0002
Evadne spp.
5.78
11.46
37.93
34.67
14.87
0.2161
Temora spp.
5.37
10.64
48.57
20.64
2.78
0.0001
Gastropoda
(larvae/Limacina)
4.03
7.98
56.55
17.26
3.03
0.0003

46
Taxa
Average
Cont.
C. cont.
Av. abund.
(I)
Av.
abund.
(II)
P(perm)
Av. dissim.: 36.48%
Country
Harbour
Whitehe
ad
Hydrozoa (medusa)
5.04
13.82
13.82
2.55
26.13
0.0143
Acartia spp.
4.46
12.23
26.04
57.18
66.78
0.9999
Evadne spp.
3.62
9.92
35.96
34.67
20.16
0.9857
Centropages spp.
3.36
9.22
45.19
9.17
24.96
0.4105
Temora spp.
2.56
7.03
52.21
20.64
9.30
0.3495
Av. dissim.: 46.95%
Whitehead
Sober
Island
Hydrozoa (medusa)
7.78
16.58
16.58
26.13
0.16
0.0001
Acartia spp.
7.58
16.15
32.73
66.78
61.58
0.9916
Centropages spp.
6.46
13.77
46.50
24.96
3.37
0.0001
Eurytemora spp.
5.29
11.27
57.77
18.76
1.32
0.0002
Evadne spp.
2.59
5.51
63.28
20.16
14.87
1.0000
Gulf region
The NMDS ordinations for the Gulf data showed separations of samples by site, although
samples from St. Peters were intermixed with samples from Cocagne and Malpeque (Fig.
15). Cocagne and St. Peters were not significantly different in their multivariate dispersion
(F1,30 = 0.035, P(perm) = 0.8545) (Table 16, Fig. 16). PERMANOVA indicated statistically
significant differences in mesozooplankton assemblage structure between these two sites
(pseudo-F1, 30 = 5.460, P(perm) = 0.0005), and site (treated as a factor) explained 15.40% of
the variation in the data (Table 17). Data from Malpeque were not included in statistical
testing due to the lower sample size (n = 3).

47
The SIMPER analysis revealed an overall average dissimilarity of 63.09% between Cocagne
and St. Peters, and Acartia spp. was the taxon most responsible for differentiating between
the two sites (higher average abundance in Cocagne, lower in St. Peters) (Table 18).
Figure 15. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples
collected from sites within the Gulf region. The ordination was constructed using a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Grey lines show the connections between each sample and the centroid of its corresponding
site.
Table 16. Summary of the multivariate homogeneity of group dispersions analysis of
mesozooplankton (0.25 mm - 5.00 mm) assemblage structure between sites (St. Peters and
Cocagne) within the Gulf region. Results are based on a Bray-Curtis dissimilarity matrix of
square root transformed mesozooplankton abundance (ind m-3), and the analysis was
followed by a posteriori pairwise comparisons between individual sites. Data from Malpeque
were not included in the analysis (n = 3).
Source
df
SS
MS
F
P(perm)
Site
1
0.001
0.001
0.035
0.8545
Residuals
30
0.640
0.021
Total
31
0.641
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.

48
Figure 16. Boxplots depicting the distance of samples to the centroid of the corresponding
site within the Gulf region. Results were obtained from the multivariate homogeneity of
groups dispersions analysis based on a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton (0.25 mm - 5.00 mm) abundance. Boxes show the first,
second and third quartiles, and lines extending from the boxes indicate the minimum and
maximum values up to 1.5 times the interquartile range. Jittered points represent the values
for individual samples. Data for Malpeque are not included (n = 3).
Table 17. Summary of the Permutational Multivariate Analysis of Variance (PERMANOVA)
between sites (St. Peters and Cocagne) within the Gulf region. Results are based on a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm)
abundance (ind m-3), and was followed by a posteriori pairwise comparisons between
individual sites. Pseudo-t values for the pairwise comparisons were calculated as the square
root of the Pseudo-F statistic generated from the pairwise.adonis2 R function (Arbizu, 2020).
Data from Malpeque were not included in the analysis (n = 3).
Source
df
SS
MS
R2
Pseudo-t
P(perm)
Site
1
0.775
0.775
15.398
5.460
0.0005
Residual
30
4.258
0.142
84.602
Total
34
5.033
100.00
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-t: t-value by permutation, P(perm): significance by 9999 permutations.
Significant effects are shown in bold (P(perm < 0.05)).
Table 18. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between sites within the Gulf region. Values in the second column (average) represent the
percent contribution of each taxon to average between-group dissimilarity, and overall
average dissimilarity (av. dissim., %) represents the sum of these values. The third column
(cont., %) is based on average (second column), but adjusted to sum to 100%, and the
fourth column represents the cumulative contribution (c. cont) of these values. The fifth and
sixth columns represent the average abundance of each taxon within each region (square
root transformed, ind m-3). The permutation p-value represents the probability of getting a
larger or equal average contribution based on 999 random permutations of input data. Note

49
that the full list of contributions from all taxa is not shown, so the sum of each entry from
average (column 2) may not equal the overall average dissimilarity, and the cumulative
contribution (column 4) may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av. abund.
(I)
Av. abund.
(II)
P(perm)
Av. dissim.: 63.09%
Cocagne
St. Peters
Acartia spp.
23.53
37.31
37.31
115.68
51.65
0.005
Copepoda (nauplii)
4.64
7.36
44.67
0.48
14.59
0.837
Centropages spp.
4.14
6.56
51.23
18.20
1.81
0.008
Podon/Pleopsis
spp.
3.89
6.16
57.39
1.00
12.92
0.852
Pseudodiaptomus
spp.
3.45
5.47
62.87
9.34
10.26
0.661
Newfoundland region
Data collection in Newfoundland focused largely on providing highly-detailed temporal data
by collecting samples in multiple months (Table 1). Results for this section therefore follow a
slightly different format in comparison to those from the other regions above, to reflect these
efforts. First, a relative abundance chart was created to show the changes in taxonomic
composition from June 2021 to July 2022. Next, an NMDS ordination was constructed as an
additional means to show these changes. Data from September 2020 was also included in
the NMDS. Tests for significance in multivariate assemblage structure were not conducted
between all of these sampling months and Venn diagrams were also not constructed due to
the large number of sampling months. Analyses focused on examining the role of tide phase
and station for the September 2020 (n = 10) and October 2021 (n = 12) datasets are
included in Section 3.3.4.
From June 2021 to July 2022, samples comprised a variety of taxa, and generally, Acartia
spp., Evadne spp., Temora spp., and Pseudocalanus spp. comprised the largest portions of
each taxa (Fig. 17). Shifts in the relative abundance were observed between months, and
occasionally, taxa comprised a large portion of the sample during certain months (e.g.,
Fritillaria spp. in April and May 2022), and then decreased in relative abundance (Fig. 17).

50
Figure 17. Relative abundance bar charts showing the zooplankton composition of individual
samples collected from June 2021 through July 2022 in South Arm (Newfoundland region).
The top nine most common taxa are identified, while the remaining taxa are grouped into an
“Other” category.
The NMDS ordination of temporal data from South Arm (Newfoundland region) shows a
cyclical progression in taxonomic composition, which moves in an approximate clockwise
pattern between months (Fig. 18). Data collected in the same months, regardless of year,
are grouped together, indicating similar assemblage structure (e.g., samples from June 2021
and June 2022 are not distinct from each other) (Fig. 18). Large shifts in composition are
observed between certain time months (e.g., October to November), while smaller shifts are
observed between other time periods (e.g., June to July) (Fig. 18).
Figure 18. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) composition from monthly samples
collected in South Arm, Newfoundland. Samples are shown in colour to represent sampling
months, and shapes indicate sampling year. Arrows connect the months chronologically, and
the asterisks represent the centroid of sampling months, regardless of year (e.g., June of all

51
data, July of all data, etc.) The ordination was constructed using a Bray-Curtis dissimilarity
matrix of square root transformed mesozooplankton abundance (ind m-3).
3.4 Objective 3: Characterizing the role of tide phase and stations on zooplankton
composition
3.4.1 Pacific region
In Lemmens Inlet, for each month, the total abundance was distributed over several classes
with no clear dominance by any taxa (Fig. 19). However, in general, the differences in the
relative abundance of taxa were more noticeable between months, than within specific
stations or tide phases at a specific site (Fig. 19), as also shown in section 3.3.2. Further
analyses were thus conducted independently for each sampling month.
Figure 19. Relative abundance bar charts showing zooplankton composition of individual
samples for different sampling months from Lemmens Inlet of the Pacific region. For each
month, the top seven most common taxa are identified, while the remaining taxa are
grouped into an “Other” category; therefore, the resulting colour scheme may differ among

52
charts. The top panel in each subplot indicates station labels as denoted in Fig. 1, and sub-
panels refer to tide phases including high tide (High) and low tide (Low).
Each NMDS ordination separated by sampling months from Lemmens Inlet showed unique
patterns (Fig. 20). In August 2020, clear groupings based on station were apparent, and the
samples transitioned from Outer to Mid to Inner stations when moving from the lower left to
upper right regions of the NMDS (Fig. 20A). In June 2021, the Mid stations formed
somewhat of a distinct grouping, while the Outer and Inner stations were intermixed (Fig.
20B). In September 2021, samples from the Inner station were separated on the ordination
from the Mid and Outer stations, which exhibited overlap (Fig. 20C). No months showed any
obvious groupings based on tide phase (Fig. 20A-C).
Figure 20. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure from different
sampling months within Lemmons Inlet of the Pacific Region, including (A) August 2020, (B)
June 2021, and (C) September 2021. Each ordination was constructed using a Bray-Curtis
dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Symbol shapes indicate tide phases, and colours represent the sampling location (station)
within each bay, as denoted in Fig. 1. Ordinations for data collected in March 2021 (n = 2)
were not created.
Lemmens August 2020 - Tide and station effects
No significant differences in multivariate dispersion were observed between the tide phases
(F1, 16 = 0.038, P(perm) = 0.8515) or among stations (F2, 15 = 0.743, P(perm) = 0.4981) for

53
the August 2020 data from Lemmens Inlet (Table 19, Fig. 21). The PERMANOVA indicated
no difference in mesozooplankton assemblage structure between the tide phases
(PERMANOVA pseudo-F1, 12 = 0.929, P(perm) = 0.3893) (Table 20). However, significant
differences in mesozooplankton assemblage structure were observed between stations
(PERMANOVA pseudo-F2,12 = 7.977, P(perm) = 0.0002) and explained 52.49% of the
variation in the data (Table 20). Pairwise comparisons revealed significant differences in
assemblage structure between all stations, with the strongest differences between samples
from the Inner and Outer stations (pseudo-t = 3.470, P(perm) = 0.0021) and Inner and Mid
stations (pseudo-t = 3.033, P(perm) = 0.0020) (Table 20). Weaker, but statistically significant
differences were observed between samples from the Mid and Outer stations (pseudo-t =
1.594, P(perm) = 0.0213) (Table 20).
SIMPER analysis similarly found the greatest differences in mesozooplankton composition
between the Outer and Inner stations (av. dissim. 43.95%), then Mid and Inner (av. dissim.:
37.19%), and the Outer and Mid were the least different (av. dissim. 25.92%) (Table 21).
Podon/Pleopsis spp. were the taxa most responsible for the differentiation between Mid
(higher) and Inner (lower) stations, Corycaeidae for Outer (higher) and Inner (lower) stations,
and Echinodermata larvae for Outer (lower) and Mid (higher) stations (Table 21).
Table 19. Summary of multivariate homogeneity of group dispersions analysis for samples in
Lemmens Inlet for the August 2020 dataset (Pacific region), showing the effect of tide and
station (run as separate tests) on mesozooplankton (0.25 mm - 5.00 mm) assemblage
structure. Results are based on a Bray-Curtis dissimilarity matrix of square root transformed
mesozooplankton abundance (ind m-3).
Source
df
SS
MS
F
P(perm)
Tide
1
0.000
0.000
0.038
0.8515
Residuals
16
0.166
0.010
Total
17
0.166
0.010
Station
2
0.004
0.002
0.743
0.4981
Residuals
15
0.041
0.003
Total
17
0.045
0.005
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.
Figure 21. Boxplots depicting the distance of samples to the centroid of the corresponding
station (left) or tide phase (right) for data collected in August 2020 in Lemmens Inlet (Pacific

54
region). Results were obtained from the multivariate homogeneity of groups dispersions
analysis based on a Bray-Curtis dissimilarity matrix of square root transformed
mesozooplankton (0.25 mm - 5.00 mm) abundance. Boxes show the first, second and third
quartiles, and lines extending from the boxes indicate the minimum and maximum values up
to 1.5 times the interquartile range. Jittered points represent the values for individual
samples.
Table 20. Summary of the Permutational Multivariate Analysis of Variance (PERMANOVA)
showing the effect of tide, station, and their interaction in Lemmens Inlet for the August 2020
dataset (Pacific region). Results are based on a square root transformed Bray-Curtis
dissimilarity matrix of mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-3), and was
followed by a posteriori pairwise comparisons between individual stations. Pseudo-t values
for the pairwise comparisons were calculated as the square root of the Pseudo-F statistic
generated from the pairwise.adonis2 R function (Arbizu, 2020).
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Tide
1
0.031
0.031
3.058
0.929
0.3893
Station
2
0.592
0.265
52.490
7.977
0.0002
Tide*Station
2
0.050
0.025
4.969
0.755
0.5952
Residual
12
0.398
0.033
39.483
Total
17
1.008
0.059
100.000
Comparison (stations)
Pseudo-t
P(perm)
Inner - Mid
3.033
0.0020
Inner - Outer
3.470
0.0021
Mid - Outer
1.594
0.0213
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are highlighted in bold (P(perm < 0.05)).
Table 21. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between stations in Lemmens Inlet (Pacific region) in August 2020. Values in the second
column (average) represent the percent contribution of each taxon to average between-
group dissimilarity, and overall average dissimilarity (av. dissim., %) represents the sum of
these values. The third column (cont., %) is based on average (second column), but
adjusted to sum to 100%, and the fourth column represents the cumulative contribution (c.
cont) of these values. The fifth and sixth columns represent the average abundance of each
taxon within each region (square root transformed, ind m-3). The permutation p-value
represents the probability of getting a larger or equal average contribution based on 999
random permutations of input data. Note that the full list of contributions from all taxa is not
shown, so the sum of each entry from average (column 2) may not equal the overall average
dissimilarity, and the cumulative contribution (column 4) may not reach 100%.

55
Taxa
Average
Cont.
C. cont.
Av. abund.
(I)
Av.
abund. (II)
P(perm)
Av. dissim.: 37.19%
Mid
Inner
Podon/Pleopsis
spp.
4.20
11.29
11.29
11.97
5.21
0.0024
Cirripedia (larvae)
3.78
10.16
21.45
12.02
5.94
0.0425
Corycaeidae
3.17
8.52
29.97
6.62
1.42
0.0207
Acartia spp.
2.35
6.32
36.29
11.93
11.93
0.1364
Evadne spp.
2.32
6.25
42.54
5.73
1.97
0.0011
Av. dissim.: 43.95%
Outer
Inner
Corycaeidae
4.83
10.99
10.99
10.07
1.42
0.0001
Cirripedia (larvae)
4.78
10.87
21.85
14.80
5.94
0.0003
Podon/Pleopsis
spp.
4.29
9.75
31.61
12.83
5.21
0.0016
Paracalanus spp.
3.61
8.21
39.82
11.59
5.05
0.0001
Evadne spp.
2.38
5.42
45.24
6.17
1.97
0.0004
Av. dissim.: 25.92%
Outer
Mid
Echinodermata
(larvae)
2.41
9.28
9.28
6.26
9.56
0.1874
Paracalanus spp.
2.11
8.14
17.42
11.59
6.93
0.4815
Acartia spp.
1.99
7.66
25.08
8.57
11.93
0.5593
Cirripedia (larvae)
1.94
7.47
32.56
14.80
12.02
0.9877
Corycaeidae
1.50
5.80
38.35
10.07
6.62
0.9995

56
Lemmens June 2021 - Tide and station effects
In the Lemmens Inlet June 2021 data, no significant differences in multivariate dispersion
were observed between tide phases (F1, 14 = 0.412, P(perm) = 0.5291) or among stations
(F2, 13 = 0.291, P(perm) = 0.7518) (Table 22, Fig. 22). PERMANOVA found no significant
differences in zooplankton assemblage structure between the tide phases (pseudo-F1, 10 =
0.451, P(perm) = 0.8749), although a significant effect of station was observed (pseudo-F2,10
= 12.123, P(perm) = 0.048) and explained 26.12% of the variation in the data (Table 23).
Pairwise comparisons found statistically significant differences in composition between Inner
and Mid stations (pseudo-t = 1.617, P(perm) = 0.0354) (Table 23). Similar to the NMDS (Fig.
20B), no significant differences were observed between the Inner and Outer stations in the
inlet (t = 1.338, P(perm) = 0.1198) (Table 23).
SIMPER analysis revealed an overall average dissimilarity of 37.19% between Inner and Mid
stations, with Acartia spp. as the taxon most responsible for the differentiation between the
two (higher average abundances in Inner, lower in Mid) (Table 24).
Table 22. Summary of the multivariate homogeneity of group dispersions analysis for
samples in Lemmens Inlet for the June 2021 dataset (Pacific region), showing the effect of
tide and station (run as separate tests) on mesozooplankton (0.25 mm - 5.00 mm)
assemblage structure. Results are based on a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton abundance (ind m-3).
Source
df
SS
MS
F
P(perm)
Tide
1
0.001
0.001
0.412
0.5291
Residuals
14
0.021
0.001
Total
15
0.022
0.002
Station
2
0.001
0.001
0.291
0.7518
Residuals
13
0.023
0.002
Total
15
0.024
0.002
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.
Figure 22. Boxplots depicting the distance of samples to the centroid of the corresponding
station (left) or tide phase (right) for data collected in June 2021 in Lemmens Inlet (Pacific
region). Results were obtained from the multivariate homogeneity of groups dispersions
analysis based on a Bray-Curtis dissimilarity matrix of square root transformed
mesozooplankton (0.25 mm - 5.00 mm) abundance. Boxes show the first, second and third

57
quartiles, and lines extending from the boxes indicate the minimum and maximum values up
to 1.5 times the interquartile range. Jittered points represent the values for individual
samples.
Table 23. Summary of Permutational Multivariate Analysis of Variance (PERMANOVA)
showing the effect of tide, station, and their interaction in Lemmens Inlet for the June 2021
dataset (Pacific region). Results are based on a square root transformed Bray-Curtis
dissimilarity matrix of mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-3), and was
followed by a posteriori pairwise comparisons between individual bays. Pseudo-t values for
the pairwise comparisons were calculated as the square root of the Pseudo-F statistic
generated from the pairwise.adonis2 R function (Arbizu, 2020).
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Tide
1
0.010
0.010
2.777
0.451
0.8749
Station
2
0.096
0.048
26.115
2.123
0.0410
Tide*Station
2
0.035
0.018
9.591
0.780
0.6584
Residual
10
0.226
0.023
61.516
Total
15
0.368
0.025
100.000
Comparison (stations)
Pseudo-t
P(perm)
Inner - Mid
1.617
0.0354
Inner - Outer
1.338
0.1198
Mid - Outer
1.614
0.0518
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are highlighted in bold (P(perm < 0.05).
Table 24. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between stations in Lemmens Inlet (Pacific region) in June 2021. Values in the second
column (average) represent the percent contribution of each taxon to average between-
group dissimilarity, and overall average dissimilarity (av. dissim., %) represents the sum of
these values. The third column (cont., %) is based on average (second column), but
adjusted to sum to 100%, and the fourth column represents the cumulative contribution (c.
cont) of these values. The fifth and sixth columns represent the average abundance of each
taxon within each region (square root transformed, ind m-3). The permutation p-value
represents the probability of getting a larger or equal average contribution based on 999
random permutations of input data. Note that the full list of contributions from all taxa is not
shown, so the sum of each entry from average (column 2) may not equal the overall average
dissimilarity, and the cumulative contribution (column 4) may not reach 100%.

58
Taxa
Average
Cont.
C. cont.
Av.
abund. (I)
Av.
abund.
(II)
P(perm)
Av. dissim.: 37.19%
Inner
Mid
Acartia spp.
3.43
14.58
14.58
33.33
24.27
0.0455
Cirripedia (larvae)
2.00
8.51
23.09
32.67
28.11
0.4187
Evadne spp.
1.93
8.20
31.29
14.06
7.23
0.0002
Oikopleura spp.
1.84
7.80
39.09
16.10
19.62
0.9185
Podon/Pleopsis spp.
1.61
6.86
45.96
16.67
11.00
0.0489
3.4.2 Maritimes region
For all sites within the Maritimes region, Acartia spp. was the most abundant taxon in every
sample. Specifically, the relative proportional abundance of Acartia spp. per sample ranged
from 61.9-96.3% in Argyle (Fig. 23A), 41.6-54.97% in Country Harbour (Fig. 23B), 74.8-
97.7% in Sober Island (Fig. 23C), and 50.6-75.6% in Whitehead (Fig. 23D). Generally, the
differences in the relative abundance of taxa were more noticeable between sites, than
within specific stations or tide phases at a specific site, as also shown in section 3.3.2.
Further analyses were thus conducted independently for each site.

59
Figure 23. Relative abundance bar charts showing the zooplankton composition of individual
samples from sites within the Maritimes region. For each site, the top seven most common
taxa are identified, while the remaining taxa are grouped into an “Other” category; therefore,
the corresponding colour scheme may differ among charts. The top panel in each subplot
indicates station labels as denoted in Fig 1, and sub-panels refer to tide phases including
high tide (High), low tide (Low), mid-falling (M-F), and mid-rising (M-R).
In general, NMDS ordinations for individual sites in the Maritimes region showed groupings
by station and in some instances, tide phase (Fig. 24). For Argyle, High tide samples were
separated from Low tide samples and occupied the lower right, and upper portions of the
NMDS, respectively (Fig 23A). Samples from the Mid-Rising tide phase were intermixed
within the High tide samples. In addition, when moving from left to right on the x-axis of the
NMDS, samples generally transitioned from Outer to Inner to Mid stations, although there
was overlap among the station types.
For Country Harbour, the near-zero stress on the NMDS ordinations indicates that multiple
unique solutions may exist, and re-running the code may generate slightly different
ordinations (Fig. 24B). In the ordination, samples appeared to generally group by station, as
Inner station samples were located in the top right of the NMDS, Outer station data were in
the middle of the ordination, and Mid station samples were located in the bottom right. One
sample (Outer station, High tide) appeared as an outlier in comparison, and was located on
the left-hand side of the NMDS. There were no clear groupings by tide phase in the
ordination.

60
For Sober Island, the NMDS showed groupings of samples by station, although deciphering
between station and tide effects is complicated by the lack of repeated sampling at different
tide phases (Fig. 24C). In particular, the Inner and Mid stations were only sampled during
Mid-Rising tides, and the Outer station was sampled only at High and Low tide. However, in
general, the outer station samples were grouped in the upper left quadrant of the NMDS,
samples from the Inner station were located in the bottom left, and Mid station samples were
in the lower-right quadrant. The lack of clear separation between Low and High tide data at
the Outer station indicates the mesozooplankton composition is likely not different between
these tide phases. Mid-rising tides were distinct from the other tide phases, and those from
the Inner-North stations were distinct from the Inner-South.
The NMDS for Whitehead showed differences in mesozooplankton composition among
stations (Fig. 24D). Moving from top left to bottom right of the NMDS, the samples transition
from Inner to Mid, then Outer stations. An effect of the tide phase may be present, since the
High tide samples from all stations were located in the bottom left portion of the NMDS,
which were separated from the intermixed Low Tide and Mid-Falling samples.
Figure 24. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for sites sampled
within the Maritimes region, including (A) Argyle, (B) Country Harbour, (C) Sober Island
Oyster, and (D) Whitehead. Each ordination was constructed using a Bray-Curtis
dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Symbol shapes indicate tide phases (LT: low tide, MR: mid-rising; MF: Mid-Falling, HT: high
tide), and colours represent the sampling location (station) within each site, as denoted in
Fig. 1. Legend items shown in white indicate that samples were not obtained for that specific
tide phase and station combination.
Argyle - Tide and station effects

61
In Argyle, no significant differences in multivariate dispersion were observed between tide
phases (F1, 11 = 2.105, P(perm) = 0.1714) or among stations (F2, 12 = 0.552, P(perm) =
0.6013) (Table 25, Fig. 25). PERMANOVA indicated a significant effect of tide phase
(pseudo-F1,7 = 7.624, P(perm) = 0.0023) that explained 29.41% of the variation in
mesozooplankton assemblage structure (Table 26). In addition, a significant Station effect
was also observed (PERMANOVA pseudo-F2,7 = 4.501, P(perm) = 0.0107) and explained
34.72% of the variation in mesozooplankton assemblage structure (Table 26). Pairwise
comparisons revealed significant differences in assemblage structure between the Central
and South stations (pseudo-t = 2.212, P(perm) = 0.0239) (Table 26). These findings agree
with the NMDS ordinations, which also showed a visual separation between these stations
(Fig. 24A).
SIMPER analysis identified Acartia spp. as the taxon most responsible for differentiating
between High tide and Low tide samples (av. dissim. 32.4%) (Table 27). The average
abundance of Acartia spp. was lower in samples from Low tide than samples at High tide.
Acartia spp. was also identified as the taxon most responsible for differentiating between
Central and South stations (av. dissim.: 35.55%) (Table 27). The average abundance of
Acartia spp. was lower in the Central than the South station (Table 27).
Table 25. Summary of the multivariate homogeneity of group dispersions analysis for Argyle
(Maritimes region), showing the effect of tide and station (run as separate tests) on
mesozooplankton (0.25 mm - 5.00 mm) assemblage structure. Results are based on a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Data with a Mid-Rising tide phase (n = 2) were excluded from the analysis.
Source
df
SS
MS
F
P(perm)
Tide
1
0.010
0.010
2.105
0.1714
Residuals
11
0.054
0.005
Total
12
0.064
0.015
Station
2
0.004
0.002
0.552
0.6013
Residuals
10
0.036
0.004
Total
12
0.040
0.006
Figure 25. Boxplots depicting the distance of samples to the centroid of the corresponding
station (left) or tide phase (right) for data collected in Argyle (Maritimes region). Results were
obtained from the multivariate homogeneity of groups dispersions analysis based on a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm)
abundance. Boxes show the first, second and third quartiles, and lines extending from the

62
boxes indicate the minimum and maximum values up to 1.5 times the interquartile range.
Jittered points represent the values for individual samples.
Table 26. Summary of the Permutational Multivariate Analysis of Variance (PERMANOVA)
showing the effect of tide, station, and their interaction in Argyle (Maritimes region). Results
are based on a square root transformed Bray-Curtis dissimilarity matrix of mesozooplankton
(0.25 mm - 5.00 mm) abundance (ind m-3), and was followed by a posteriori pairwise
comparisons between individual stations. Pseudo-t values for the pairwise comparisons
were calculated as the square root of the Pseudo-F statistic generated from the
pairwise.adonis2 R function (Arbizu, 2020). Data with a Mid-Rising tide phase (n = 2) were
excluded from the analysis; therefore, pairwise comparisons are not shown for tide effects
since there are only two levels (high tide and low tide) in the main PERMANOVA.
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Tide
1
0.171
0.171
29.407
7.624
0.0023
Station
2
0.202
0.101
34.721
4.501
0.0107
Tide*Station
2
0.052
39.387
8.870
1.150
0.3456
Residual
7
0.157
0.022
27.002
Total
12
0.582
0.0485
100.000
Comparison (stations)
Pseudo-t
P(perm)
Central -
South
2.212
0.0239
Central - North
1.767
0.0687
South - North
1.040
0.3464
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation. Significant effects are shown in bold (P(perm < 0.05).
Table 27. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between tide phases (HT: high tide, LT: low tide) and stations in Argyle (Maritimes region).
Tests were only conducted between significantly different (P<0.05) stations or tide phases
identified in the pairwise PERMANOVA results. Values in the second column (average)
represent the percent contribution of each taxon to average between-group dissimilarity, and
overall average dissimilarity (av. dissim., %) represents the sum of these values. The third
column (cont., %) is based on average (second column), but adjusted to sum to 100%, and
the fourth column represents the cumulative contribution (c. cont) of these values. The fifth
and sixth columns represent the average abundance of each taxon within each region
(square root transformed, ind m-3). The permutation p-value represents the probability of
getting a larger or equal average contribution based on 999 random permutations of input
data. Note that the full list of contributions from all taxa is not shown, so the sum of each
entry from average (column 2) may not equal the overall average dissimilarity, and the
cumulative contribution (column 4) may not reach 100%.

63
Taxa
Average
Cont.
C.
cont.
Av.
abund.
(I)
Av.
abund.
(II)
P(perm)
Av. dissim.: 32.40%
LT
HT
Acartia spp.
8.37
25.83
25.83
30.17
35.39
0.4408
Centropages spp.
3.60
11.11
36.94
4.66
8.90
0.0458
Gastropoda
(larvae/Limacina)
3.19
9.84
46.78
0.82
5.59
0.0019
Eurytemora spp.
2.58
7.96
54.74
2.55
6.01
0.0175
Pseudodiaptomus spp.
2.16
6.67
61.41
4.36
6.28
0.4932
Av. dissim.: 35.55%
Central
South
Acartia spp.
12.45
35.01
35.01
24.72
42.05
0.0038
Gastropoda
(larvae/Limacina)
3.38
9.51
44.52
0.99
6.05
0.0353
Centropages spp.
3.05
8.58
53.10
3.81
7.91
0.5788
Eurytemora spp.
2.58
7.26
60.36
2.25
6.15
0.131
Pseudodiaptomus spp.
2.54
7.15
67.51
3.40
7.22
0.1988
3.4.3 Gulf region
For sites within the Gulf region, Acartia spp. was usually the most abundant taxon in each
sample, with relative abundances of 69.1 - 96.3% in Cocagne (Fig. 26A) and 61.2 - 99.3% in
St. Peters (Fig. 26C). Exceptions were seen in Malpeque where samples collected in fall
stormy weather were dominated by Evadne spp. (Central station, 40.4%), or Fritillaria spp.
(North 19.8% and South 18.6% stations) (Fig. 26B) and in St. Peters where Bivalvia larvae
could also occasionally comprise the largest portion of the relative abundance (up to 59.4%)
(Fig. 26C). Further analyses were performed independently for each site as it was the most
structuring factor of Gulf data (see section 3.3.2).

64
Figure 26. Relative abundance bar charts showing the zooplankton composition of individual
samples from sites within the Gulf region. For each site, the top seven most common taxa
are identified, with all other taxa grouped into an “Other” category; therefore, the resulting
colour scheme may differ among charts. The top panel in each subplot indicates station
labels as denoted in Fig. 1, and sub-panels refer to tide phases including high tide (High),
low tide (Low), mid-falling (M-F), and mid-rising (M-R).
For Cocagne, the ordination shows a separation between July (left-hand side of NMDS) and
August (mid/right) samples, indicating a possible difference in composition between the two
sampling months (Fig. 27A). However, the near-zero stress indicates multiple equally-valid
ordinations may be possible, although running the code multiple times generally resulted in
similar patterns being displayed (Fig 26A). By default, this also shows differences in
composition between Low (July) and Mid-Rising (August) tide phases, since samples were
only collected at one tide phase per month (Fig. 27A). No obvious trends are apparent in the
NMDS ordination for Malpeque due to low sample size (n = 3, Fig. 27B). For St. Peters,
there is no obvious grouping by tide phase in the NMDS ordination (Fig. 27C). However,
samples are somewhat grouped by station, and the Outer station samples are generally
located in the upper portions of the NMDS, while samples from the Mid and Inner stations
are intermixed in the lower portion of the NMDS (Fig. 27C).

65
Figure 27. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples
collected within the Gulf region, including (A) Cocagne, (B) Malpeque, and (C) St. Peters.
Each ordination was constructed using a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton abundance (ind m-3). Symbol shapes indicate tide phases
(LT: low tide and HT: high tide), and colours represent the sampling station within each bay,
as denoted in Fig. 1. Text labels in (A) represent sampling months (Jul: July and Aug:
August), since samples in Cocagne were obtained in more than one field season. Legend
items shown in white indicate that samples were not obtained for that specific tide phase and
station combination.
St. Peters - Tide and station effects
In St. Peters, there were no significant differences in multivariate dispersion between the tide
phases (F1, 20 = 0.051, P(perm) = 0.8232) or stations (F2, 19 = 0.183, P(perm) = 0.8309)
(Table 28, Fig. 28). PERMANOVA did not identify significant differences in zooplankton
assemblage structure between tide phases (pseudo-F1,16 = 0.793, P(perm) = 0.5130) (Table
29). A significant Station effect was observed (PERMANOVA pseudo-F2,16 = 3.026, P(perm)
= 0.0081) and explained 25.5% of the variation in mesozooplankton assemblage structure
(Table 29). Pairwise comparisons revealed statistically significant differences in assemblage
structure between the Outer and Inner stations (pseudo-t = 2.098, P(perm) = 0.0037) and
between Outer and Mid stations (pseudo-t = 2.140, P(perm) = 0.0036) (Table 29).
SIMPER analysis identified Acartia spp. as the taxon most responsible for differentiating
between Outer and Inner stations (av. dissim.: 61.46%) (Table 30). The average abundance
of Acartia spp. was lower in samples from the Outer stations and higher in samples from the
Inner stations. Acartia spp. was also identified as the taxon most responsible for
differentiating between Outer and Mid stations (av. dissim.: 53.93%). The average
abundance of Acartia spp. was lower in samples from the Outer stations and higher in
samples from the Mid stations.

66
Table 28. Summary of the multivariate homogeneity of group dispersions analysis for St.
Peters (Gulf region), showing the effect of tide and station (run as separate tests) on
mesozooplankton (0.25 mm - 5.00 mm) assemblage structure. Results are based on a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton abundance (ind m-3).
Data with a mid-falling tide phase (n = 4) were removed from the analysis.
Source
df
SS
MS
F
P(perm)
Tide
1
0.001
0.001
0.051
0.8232
Residuals
20
0.475
0.024
Total
21
0.476
0.025
Station
2
0.011
0.006
0.183
0.8309
Residuals
19
0.592
0.031
Total
21
0.603
0.037
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.
Figure 28. Boxplots depicting the distance of samples to the centroid of the corresponding
station (left) or tide phase (right) for data collected in St. Peters (Gulf region). Results were
obtained from the multivariate homogeneity of groups dispersions analysis based on a Bray-
Curtis dissimilarity matrix of square root transformed mesozooplankton (0.25 mm - 5.00 mm)
abundance. Boxes show the first, second and third quartiles, and lines extending from the
boxes indicate the minimum and maximum values up to 1.5 times the interquartile range.
Jittered points represent the values for individual samples. Data with a mid-falling tide phase
(n = 4) are not shown.
Table 29. Summary of Permutational Multivariate Analysis of Variance (PERMANOVA)
showing the effect of tide, station, and their interaction in St. Peters Bay (Gulf region).
Results are based on a square root transformed Bray-Curtis dissimilarity matrix of
mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-3), followed by a posteriori
pairwise comparisons between individual stations. t-values for the pairwise comparisons
were calculated as the square root of the Pseudo-F statistic generated from the
pairwise.adonis2 R function (Arbizu, 2020). Data with a mid-falling tide phase (n = 4) were
removed from the analysis.

67
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Tide
1
0.107
0.107
3.341
0.793
0.5130
Station
2
0.817
0.408
25.500
3.026
0.0081
Tide*Station
2
0.120
0.060
3.743
0.444
0.9094
Residual
16
2.160
0.135
67.416
Total
21
3.203
0.153
100.000
Comparison (stations)
Pseudo-t
P(perm)
Inner - Mid
0.612
0.8070
Inner - Outer
2.098
0.0037
Mid - Outer
2.140
0.0036
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations;
Pseudo-t: t-value by permutation.t: t-value. Significant effects are shown in bold (P(perm <
0.05)).
Table 30. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between stations in St. Peters Bay (Gulf region). Tests were only conducted between
significantly different (P<0.05) stations identified in the pairwise PERMANOVA results.
Values in the second column (average) represent the percent contribution of each taxon to
average between-group dissimilarity, and overall average dissimilarity (av. dissim., %)
represents the sum of these values. The third column (cont., %) is based on average
(second column), but adjusted to sum to 100%, and the fourth column represents the
cumulative contribution (c. cont) of these values. The fifth and sixth columns represent the
average abundance of each taxon within each region (square root transformed, ind m-3). The
permutation p-value represents the probability of getting a larger or equal average
contribution based on 999 random permutations of input data. Note that the full list of
contributions from all taxa is not shown, so the sum of each entry from average (column 2)
may not equal the overall average dissimilarity, and the cumulative contribution (column 4)
may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av.
abund. (I)
Av. abund.
(II)
P(perm)
Av. dissim.: 61.46%
Outer
Inner
Acartia spp.
15.00
24.40
24.40
30.37
67.72
0.4061
Bivalvia (larvae)
6.96
11.32
35.72
22.00
0.41
0.0046
Podon/Pleopsis spp.
6.45
10.50
46.22
7.51
24.31
0.0273

68
Taxa
Average
Cont.
C. cont.
Av.
abund. (I)
Av. abund.
(II)
P(perm)
Copepoda (nauplii)
5.13
8.35
54.57
12.75
16.49
0.6392
Oithona spp.
4.39
7.14
61.70
14.39
1.73
0.0046
Av. dissim.: 53.93%
Outer
Mid
Acartia spp.
13.03
24.16
24.16
30.37
56.45
0.8447
Bivalvia (larvae)
6.97
12.92
37.08
22.00
2.30
0.0007
Copepoda (nauplii)
4.82
8.93
46.01
12.75
14.63
0.8805
Oithona spp.
4.36
8.09
54.10
14.39
3.03
0.0011
Hydrozoa (medusa)
2.94
5.44
59.54
8.83
0.75
0.0001
3.4.4 Newfoundland region
In South Arm, visualizations (relative abundance charts and NMDS ordinations) were
created for the September 2020 and October 2021 datasets to show the potential influences
of station and tide effects. For these two months, Acartia spp., Evadne spp., Pseudocalanus
spp., and Temora spp. comprised the majority of each sample (Fig. 29). However,
differences in the relative abundance of taxa were more noticeable between these two
months than within specific stations or tide phases for each month, as also shown in Section
3.3.2. Further analyses were thus conducted independently for each month.

69
Figure 29. Relative abundance bar charts showing the zooplankton composition of individual
samples from (A) September 2020 and (B) October 2021 from South Arm (Newfoundland).
For each time period, the top seven most common taxa are identified, while the remaining
taxa are grouped into an “Other” category; therefore, the resulting colour scheme may be
different among charts. The top panel in each subplot indicates station labels as denoted in
Fig. 1, and sub-panels refer to tide phases including high (High), low (Low), mid-falling (M-
F), and mid-rising (M-R) tides.
The NMDS ordinations for samples collected in September 2020 from South Arm do not
show groupings based on tide phase, but do show a separation by station (Fig. 30A). In
October 2021, differences in composition between Outer (left-hand side of NMDS) and Mid-
B stations (right) were apparent, and showed a clear separation (Fig 29B). There were no
obvious groupings based on tide phase (Fig 29B).
Figure 30. Two-dimensional non-metric multidimensional scaling ordination showing
similarity in mesozooplankton (0.25 mm - 5.00 mm) assemblage structure for samples
collected in (A) September 2020 and (B) October 2021 in South Arm, Newfoundland. Each
ordination was constructed using a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton abundance (ind m-3). Symbol shapes indicate tide phases
(LT: low tide, MF: mid-falling, MR: mid-rising, HT: high tide), and colours represent the
sampling location (station) within each bay, as denoted in Fig. 1. Legend items shown in
white indicate that samples were not obtained for that specific tide phase and station
combination.
South Arm October 2021 - Tide and station effects
For the October 2021 data from South Arm, no significant differences in multivariate
dispersion were observed between the tide phases (F1, 9 = 0.280, P(perm) = 0.6011) or

70
between stations (F1, 9 = 0.145, P(perm) = 0.6955) (Table 31, Fig. 31). PERMANOVA
indicated no statistically significant differences in zooplankton assemblage structure between
the Low and High tide phases (PERMANOVA pseudo-F1, 7 = 1.236, P(perm) = 0.2697),
although a significant effect of station (Mid-B vs Outer) was observed (PERMANOVA
pseudo-F1,7 = 4.187, p = 0.0012), which explained 31.09% of the variation in
mesozooplankton assemblage structure (Table 32).
SIMPER analysis identified Temora spp. as the taxon most responsible for differentiating
between the Outer and Mid-B stations (av. dissim.: 18.92%) (Table 33). The average
abundance of Temora spp. was higher in samples from the Outer station and lower in
samples from the Mid-B station.
Table 31. Summary of the multivariate homogeneity of group dispersions analysis for
October 2021 data collected in October 2021 in South Arm (Newfoundland region). Results
show the effect of tide and station (run as separate tests) on mesozooplankton (0.25 mm -
5.00 mm) assemblage structure based on a Bray-Curtis dissimilarity matrix of square root
transformed mesozooplankton abundance (ind m-3). Data from the mid-falling tide phase (n =
1) were excluded from the analysis.
Source
df
SS
MS
F
P(perm)
Tide
1
0.000
0.000
0.280
0.6011
Residuals
9
0.003
0.000
Total
10
0.004
0.000
Station
1
0.000
0.000
0.145
0.6955
Residuals
9
0.003
0.000
Total
10
0.003
0.000
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; F: F-statistic,
P(perm): significance by 9999 permutations.
Figure 31. Boxplots depicting the distance of samples to the centroid of the corresponding
station (left) or tide phase (right) for data collected in October 2021 from South Arm
(Newfoundland region). Results were obtained from the multivariate homogeneity of groups
dispersions analysis based on a Bray-Curtis dissimilarity matrix of square root transformed
mesozooplankton (0.25 mm - 5.00 mm) abundance. Boxes show the first, second and third
quartiles, and lines extending from the boxes indicate the minimum and maximum values up

71
to 1.5 times the interquartile range. Jittered points represent the values for individual
samples. Data from the mid-falling tide phase (n = 1) are not shown.
Table 32. Summary of Permutational Multivariate Analysis of Variance (PERMANOVA)
showing the effect of tide, station, and their interaction for samples collected in South Arm
(Newfoundland region) in October 2021. Results are based on a square root transformed
Bray-Curtis dissimilarity matrix of mesozooplankton (0.25 mm - 5.00 mm) abundance (ind m-
3). Data from the mid-falling tide phase (n = 1) were excluded from the analysis.
Source
df
SS
MS
R2
Pseudo-F
P(perm)
Tide
1
0.014
0.014
9.180
1.236
0.2697
Station
1
0.047
0.047
31.091
4.187
0.0012
Tide*Station
1
0.012
0.012
7.755
1.044
0.3982
Residual
7
0.079
0.011
51.974
Total
10
0.152
0.015
100.000
df: degrees of freedom; SS: sum of squares; MS: mean sum of squares; R2: coefficient of
variation; Pseudo-F: F statistic by permutation, P(perm): significance by 9999 permutations.
Significant effects are shown in bold (P(perm < 0.05)).
Table 33. Similarity percentage (SIMPER) analysis to identify the top five mesozooplankton
(0.25 mm - 5.00 mm) taxa that contribute most to the average Bray-Curtis dissimilarities
between stations in South Arm (Newfoundland region) in October 2021. Tests were only
conducted between significantly different (P<0.05) stations identified in the pairwise
PERMANOVA results. Values in the second column (average) represent the percent
contribution of each taxon to average between-group dissimilarity, and overall average
dissimilarity (av. dissim., %) represents the sum of these values. The third column (cont., %)
is based on average (second column), but adjusted to sum to 100%, and the fourth column
represents the cumulative contribution (c. cont) of these values. The fifth and sixth columns
represent the average abundance of each taxon within each region (square root
transformed, ind m-3). The permutation p-value represents the probability of getting a larger
or equal average contribution based on 999 random permutations of input data. Note that
the full list of contributions from all taxa is not shown, so the sum of each entry from average
(column 2) may not equal the overall average dissimilarity, and the cumulative contribution
(column 4) may not reach 100%.
Taxa
Average
Cont.
C. cont.
Av.
abund.
(I)
Av.
abund.
(II)
P(perm)
Av. dissim.: 18.92%
Outer
Mid B
Temora spp.
2.55
13.45
13.45
15.93
10.84
0.0050
Pseudocalanus spp.
1.76
9.32
22.77
10.75
14.31
0.0001

72
Taxa
Average
Cont.
C. cont.
Av.
abund.
(I)
Av.
abund.
(II)
P(perm)
Evadne spp.
1.54
8.13
30.90
15.25
13.81
0.5806
Oithona spp.
1.15
6.10
37.00
5.15
7.39
0.0137
Acartia spp.
1.15
6.07
43.07
19.36
18.57
0.2619
No significant differences in taxa richness were observed between high and low tide for any
sites where comparisons were evaluated (Table 34).
Table 34. Results of the two-sample t-tests to evaluate differences in abundance, taxa
richness, Shannon diversity (i.e., the exponential of the Shannon index), and Simpson
diversity (i.e., the inverse Simpson index) between tide phases for select stations. Tests
were conducted for stations with at least three samples per high or low tide.
Region
Site
Station
Index
Mean
HT
Mean
LT
df
t
95% CI
P
Pac
Lemmens
Aug 2020
Outer
Abundance
949.34
1265.1
8
4
-0.66
[-1644.74,
1013.06]
0.545
Richness
24.33
23.67
4
0.34
[-4.81,
6.14]
0.752
Shannon
10.75
10.35
4
0.32
[-2.98,
3.77]
0.763
Simpson
7.83
7.37
4
0.53
[-1.99,
2.92]
0.625
Mid
Abundance
965.12
775.05
4
0.77
[-495,
875.13]
0.484
Richness
21.33
23.33
4
-2.68
[-4.07,
0.07]
0.055
Shannon
9.79
10.37
4
-1.07
[-2.10,
0.93]
0.345
Simpson
7.37
7.66
4
-0.38
[-2.40,
1.83]
0.725

73
Region
Site
Station
Index
Mean
HT
Mean
LT
df
t
95% CI
P
Inner
Abundance
383.19
352.07
4
0.19
[-428.5,
490.74]
0.860
Richness
18.00
18.33
4
-0.12
[-7.74,
7.07]
0.907
Shannon
6.46
7.18
4
-0.65
[-3.81,
2.37]
0.552
Simpson
4.07
4.81
4
-0.83
[-3.22,
1.74]
0.455
Lemmens
Jun 2021
Mid
Abundance
2063.04
2989.0
6
4
-0.83
[-4017.95,
2165.92]
0.452
Richness
25.67
23.33
4
0.89
[-4.95,
9.62]
0.424
Shannon
7.42
6.61
4
1.57
[-0.62,
2.25]
0.191
Simpson
4.63
4.30
4
1.01
[-0.58,
1.24]
0.369
Mar
Sober
Island
Outer
Abundance
4827.65
6503.3
3
4
-0.44
[-12316.5,
8965.12]
0.685
Richness
14.67
16.33
4
-0.51
[-10.83,
7.5]
0.640
Shannon
1.99
1.64
4
1.08
[-0.54,
1.23]
0.340
Simpson
1.49
1.23
4
1.63
[-0.18,
0.69]
0.179
Gulf
St. Peters
Outer
Abundance
4883.02
1654.5
9
6
1.64
[-1593.31,
8050.17]
0.152
Richness
23.20
20.67
6
0.64
[-7.19,
12.25]
0.547
Shannon
4.37
6.18
6
-1.26
[-5.33,
1.72]
0.256

74
Region
Site
Station
Index
Mean
HT
Mean
LT
df
t
95% CI
P
Simpson
2.73
4.20
6
-1.29
[-4.27,
1.33]
0.246
Mid
Abundance
5076.17
8346.5
0
5
-0.64
[-16490.59,
9949.94]
0.553
Richness
10.75
11.67
5
-0.49
[-5.68,
3.85]
0.642
Shannon
1.66
1.77
5
-0.25
[-1.22, 1]
0.810
Simpson
1.27
1.35
5
-0.35
[-0.65,
0.49]
0.740
Inner
Abundance
10231.7
3
4892.3
1
5
0.74
[-13243.19,
23922.04]
0.493
Richness
10.75
10.33
5
0.15
[-6.79,
7.63]
0.888
Shannon
1.87
1.99
5
-0.24
[-1.36,
1.13]
0.820
Simpson
1.43
1.54
5
-0.38
[-0.83,
0.62]
0.718
NL
SE Arm
Oct 2021
Outer
Abundance
1140.12
1167.8
2
4
-0.43
[-205.73,
150.35]
0.688
Richness
23.33
22.00
4
0.85
[-3.01,
5.67]
0.442
Shannon
6.20
6.30
4
-0.34
[-0.93,
0.72]
0.750
Simpson
4.40
4.51
4
-0.30
[-1.12, 0.9]
0.777
Mean HT: average taxa richness from samples obtained at high tide, Mean LT: average taxa
richness from samples obtained at low tide, df: degrees of freedom, t: t-value, CI: confidence
interval, P: p-value.

75
4 DISCUSSION
Zooplankton play critical roles in marine food webs, and their communities are potentially
altered by intensive bivalve aquaculture (Lindeman 1942; Hulot et al. 2014, 2020). This
report examined the spatiotemporal dynamics of mesozooplankton assemblage structure of
AMP data obtained using an innovative imaging system from nine sites, located across four
DFO regions, from various sampling months, tide phases, and sampling stations. The results
showed strong station effects within bays, underscoring the relevance of site-specific spatial
dynamics, while tide effects were generally a less important factor for structuring
mesozooplankton communities. Differences in assemblage structure were observed among
monthly observations in the Pacific and Newfoundland regions, highlighting the importance
of considering seasonality in future sampling. By using an optical imaging system, these
results represent the first of their kind for coastal zooplankton monitoring in Canada, and will
contribute to a global effort for more advanced plankton analyses using machine learning
approaches (e.g., see Irisson et al. 2022). These analyses help address key knowledge
gaps for effectively monitoring potential long-term ecosystem changes from bivalve
aquaculture.
This report used taxa accumulation theory to quantify sampling completeness and assess
the extent of undetected diversity within each site. Completeness ranged from 55.94%
(South Arm, September 2020) to 99.45% (Whitehead), indicating that in general, a large
portion of the estimated taxa had been sampled. Although estimated richness was generally
highest in the Pacific region, high completeness was observed in multiple sites, regardless of
region, including Lemmens June 2021 (98.7%, Pacific), Whitehead (99.6%, Maritimes), and
St. Peters (98.6%, Gulf). For these sites, the stabilization of the slope of the rarefaction and
extrapolation curves indicates that the asymptotic estimates are reliable, and the sampling
effort for these sites obtained a very high proportion of the estimated diversity (Chao et al.
2014). For the remaining sites (Lemmens August 2020 and September 2021, Country
Harbour, Sober Island, Cocagne, Malpeque, and South Arm), a positive slope on the
rarefaction and extrapolation curves was still present when extrapolated to double the
sample size; therefore, the asymptotic estimates for richness represent a lower bound (Chao
et al, 2020). This often occurs for richness, as there are typically vanishingly rare taxa still to
be sampled (Chao et al., 2020), and indicates that these sites have a greater number of rare
taxa, which would require increased sampling effort to obtain. Rare taxa are often of interest
in monitoring programs as they may make up important ecological roles and can act as
indicators of human-induced environmental changes (Cao et al. 1998, 2001; Ma et al. 2022);
however, it is virtually impossible to sample all taxa present, especially in hyper-diverse
communities (Colwell and Coddington 1994; Gotelli and Colwell 2001; Magurran and McGill
2010). Regardless, these results indicate that a large portion of diversity was captured at
each site (55.94% - 99.45%). The markedly different completeness profile for South Arm
September 2020 was likely due to the high Chao2 asymptotic estimate, which can result
from a large number of singletons within samples, or taxa that are only present within one
sample (Chao 1984, 1987). Future work can address whether these are true biological
phenomena or if a greater number of samples per site will result in asymptotic estimates
closer to the observed richness value (e.g., as with the South Arm October 2021 data).
Overall, these results provide important baseline information for expected trends in diversity,
and present helpful guidance of potential changes to completeness if sampling effort is
increased or decreased in future field campaigns.
Data collected from multiple months in the Newfoundland and Pacific regions revealed that
seasonality effects are important to consider when monitoring mesozooplankton. This
agrees with existing research, since zooplankton are known to exhibit seasonality in
community structure, which is driven by a combination of both biological and physical factors
(Neuheimer et al. 2010; Ji et al. 2010; Tommasi et al. 2013). In the Pacific region, the
assemblage structure was different between each month; therefore, sampling in multiple

76
months will lead to better understanding of the local diversity. In Newfoundland, the year-
round data collection provided a detailed characterization of mesozooplankton dynamics,
and revealed a cyclical shift in mesozooplankton communities between months. In addition
to providing important information of the local diversity, these results may help identify
preferred months for sampling. For example, large shifts in the assemblage structure during
certain time periods were visualized on the NMDS ordination (e.g., October to November,
March to April) which may be related to processes such as plankton blooms or storms.
Because AMP is focused on evaluating long-term trends, sampling during these time periods
is likely not recommended as the mesozooplankton structure is less stable, thereby
challenging the interpretation of inter-annual comparisons. By contrast, the end of June to
September period exhibited more stable patterns, as indicated by the smaller changes in
composition on the NMDS ordination. Sampling during these time periods is advised, as it
may be easier to detect long-term changes in community structure, since the variability
associated with events such as blooms or storms may be minimized. Data collected within
these months may be used to test which factors drive any additional changes to community
composition over time, and pose new hypotheses related to the driving forces behind these
shifts. For example, climate change is expected to cause alterations to zooplankton
communities by causing changes in phenology (e.g., spring and summer species occur
earlier), poleward shifts in distribution to remain within an optimal temperature range, and
overall reductions in body size (see Ratnarajah et al. 2023 for review). Therefore, sampling
at more stable time periods may help reduce the complexity of analyzing the effects of
multiple driving forces, and isolate climate change-related effects to those from bivalve
aquaculture. However, zooplankton communities are often characterized by strong
interannual variability (e.g., Mackas et al. 2012) and the factors influencing zooplankton
communities may differ seasonally due to these annually-varying processes (Varpe 2012).
Likewise, the effects of bivalve aquaculture on zooplankton communities may differ
seasonally (Steeves et al. 2018), but characterizing these effects across seasons was
beyond the scope of the study. More data are therefore required to characterize these
patterns, and to aid in the long-term analysis of these trends, we suggest sampling from at
least three separate months at each site. Although we cannot develop specific
recommendations for each site due to the lack of temporal data, results from Newfoundland
suggest the summer months (June to September) may be the preferred sampling months
going forward. Selecting consecutive months within this time frame would likely be ideal, for
a continuous monthly time series within one sampling year.
Spatial patterns within bays are important to consider in future field seasons, as indicated by
the observed differences in mesozooplankton assemblage structure between stations (i.e.,
“station effects”) for all sites. These station effects were observed either through statistical
testing (e.g., Lemmens August 2020 and June 2021, Argyle, St. Peters, South Arm October
2021) or visualized on the NMDS ordinations (all others, excluding Cocagne, n = 6, and
Malpeque, n = 3, for which more samples are required for reliable conclusions). These
results highlight the importance of spatial dynamics, which has implications for sampling
designs. For example, collecting samples from a single station within each bay is likely
insufficient, since important bay dynamics would be missed. Increasing the number of
samples to characterize this variability would be recommended (discussed below in more
detail). However, spatial gradients in mesozooplankton assemblage structure are known in
coastal settings and are driven by a complex set of biophysical characteristics including
salinity, oxygen content, nutrient levels, turbidity and temperature (Soetaert and Van Rijswijk
1993; Marques et al. 2007; Helenius et al. 2017; Usov et al. 2019). Because station effects
were observed in all sites, even those with low to no aquaculture production (e.g., Argyle,
Country Harbour), the differences in composition between stations are likely driven by a
combination of various local factors. While analyzing patterns in spatial distributions will not
provide a direct causal link to processes such as bivalve grazing (McIntire and Fajardo
2009), increasing the spatial coverage and mapping the results may help further disentangle

77
the role of multiple driving forces on zooplankton distributions and provide a clearer visual
link to the role of bivalve aquaculture production on their communities.
Tide effects on community structure were tested to (1) define if direct grazing from bivalve
aquaculture could be detected and (2) if samples collected at different tides provide different
species composition and thus need to be taken into consideration within long-term
operational monitoring plans. Differences in composition between tide phases (i.e., “tide
effects'') were generally less pronounced than station effects. Likely, the variety of
ecosystems, hydrographic properties, and embayment complexity within AMP sites resulted
in the lack of a uniform response to different tide phases. For example, tide effects were only
significant in Argyle, which has a highly complex coastline, and is located near the Bay of
Fundy. The large tidal ranges likely result in differences in composition between the time
periods. As indicated by the NMDS ordinations, tide effects may be important for structuring
the mesozooplankton communities in Whitehead and Sober Island, which have
comparatively high bivalve production for sites in the Maritimes region (see Table 35).
However, tide effects were not observed in St. Peters Bay, which has similar hydrography to
Whitehead (e.g., both are enclosed, narrow channels with one point of exposure with the
ocean). Potentially, tide effects may be noted in some sites as the result of diel vertical
migrations of the zooplankton, which may be affected by the type of tow used. Zooplankton
have been found to utilize tidal currents for retention within the system, by moving vertically
between outflowing surface water and inflowing deeper water (e.g. Wooldridge and Erasmus
1980; Schlacher and Wooldridge 1995). Although we attempted to minimize these effects by
sampling the entire (or nearly the entire) water column, this is more difficult to achieve in
shallow sites, where horizontal or oblique tows were obtained. For sites where tide effects
were not observed, other factors were likely more important than tides for structuring the
zooplankton communities, and monitoring the two tides will not greatly improve the
description of the mesozooplankton community. A detailed understanding of the water flow
within these sites may therefore help supplement future data analysis. Spatially-explicit
hydrodynamic models exist for several of the sites (see Table 1 for references) and this pre-
existing knowledge of the ecosystem could supplement future data analysis. For example,
the outputs of seston or zooplankton depletion from the coupled biological-physical models
could be compared against the collected AMP data in future work. A qualitative modeling
approach may highlight critical ecological processes and interactions that could be disrupted
by the presence of farms (Forget et al. 2020). There is also empirical evidence of shellfish
aquaculture impacts on phytoplankton at bay-scale (Cranford et al. 2008), and Gianasi et al.
(2023) showed that high mussel production decreases phytoplankton concentration to a
point that it negatively impacts zooplankton survival, including important commercial species
in a series of laboratory experiments. There is thus a potential food-web interaction through
zooplankton depletion (i.e. direct grazing and competition) showed by theoretical models and
experiments. However, in natural environments, it is unclear if bivalve aquaculture may
significantly impact zooplankton communities due to unpredictable large-scale ecosystem
complexity, such as interactions between nutrient availability, renewal rates for phyto- and
zooplankton, niche partitioning, and environmental/stochastic variability. The development of
a monitoring program for aquaculture will provide a more holistic view of ecosystem
dynamics, species diversity, interactions, and changes in Canadian bivalve aquaculture
sites.
Overall, these results suggest that station effects generally play a larger role in defining the
zooplankton assemblage structure at each site than tide effects. We therefore recommend
collecting samples at a greater number of stations from each site, to better understand these
spatial dynamics. For example, samples could be obtained from 20 stations (i.e., 1 sample
from each station, in 1 day, with no replicates), spread approximately evenly throughout
each site, which would likely provide high sampling completeness as indicated from the taxa
accumulation curves. In addition, 20 samples is also similar to the original goal of 18
samples per site (i.e., 1 sample from 3 stations from both high and low tide over 3 days = 1 x

78
3 x 2 x 3 = 18 samples), and could be obtained with the other pelagic properties already
being collected. Although the rarefaction and extrapolation curves level off before 20
samples for some sites (e.g., Whitehead), it is generally recommended to oversample in the
initial stages of monitoring programs, and reduce the number of samples in the future if
necessary (Hoffman et al. 2011). For sites with confirmed tide effects (Argyle), possible tide
effects (e.g., Sober Island, Whitehead), or not enough samples to draw conclusions (e.g.,
Country Harbour, Cocagne, Malpeque), the tide phase should still be considered when
sampling. It would be recommended to test the tidal effect again, or sample at a single tide
phase if this is logistically possible, to ensure data from multiple tide phases do not add
additional variability to the dataset. For the remaining sites (St. Peters, South Arm,
Lemmens), we have no evidence to suggest tide phases affect the mesozooplankton
communities, and samples can therefore be obtained at any time during the day. Table 35 is
provided below to summarize the sampling completeness, the presence of tide and station
effects, and if seasonality (i.e., monthly sampling) had been considered. In absence of
carrying capacity indicators consistently estimated across sites, an additional classification
was assigned to each bay, indicating the potential vulnerability to aquaculture impacts on
zooplankton composition. This represents a qualitative measure combining bivalve
aquaculture production and bay hydrodynamics to evaluate the risk of effects on
mesozooplankton community structure from bivalve grazing. Simple models could provide
proper quantitative grazing pressure evaluations (Comeau et al. 2023) and should be
considered to further characterize AMP sites.
Table 35. Summary table of results and sampling advice for sites and/or sampling months,
collected as part of the Aquaculture Monitoring Program. Vulnerability represents a
qualitative classification assigned to each bay, indicating the potential combined effects of
bivalve aquaculture production and bay hydrodynamics on mesozooplankton community
structure. Completeness represents the ratio of the observed to expected richness, derived
from the sample-based rarefaction and extrapolation analyses. For the Station effect and
Tide effect columns, “Yes” indicates statistically significant (P<0.05) effects of each variable
were observed from PERMANOVA, while “No” indicates a significant effect was not
observed. “Likely” indicates mesozooplankton assemblage structure is different among Tide
phases or Stations, as visualized from the non-metric multidimensional scaling (NMDS)
ordinations, but the sample sizes were too low for significance testing. “Unlikely” indicates an
obvious effect was not visualized on the NMDS, and the sample sizes were also too low for
significance testing. “Unknown” indicates the sample size was too low to allow for trends to
be observed from either significance testing (PERMANOVA) or multivariate ordinations
(NMDS).
Region
Site
Vulnerability
Completene
ss
Station
effect
Tide
effect
Seasonality
Pac
Lemmens
(Aug 2020)
Low
89.17%
Yes
No
Yes
Lemmens
(June 2021)
98.67%
Yes
No
Lemmens
(Sept 2021)
87.19%
Likely
Unlikely
Mar
Argyle
Low
77.68%
Yes
Yes
No data
Country
Harbour
Low
86.11%
Likely
Unlikely

79
Region
Site
Vulnerability
Completene
ss
Station
effect
Tide
effect
Seasonality
Sober Island
High
83.58%
Likely
Likely
Whitehead
High
99.45%
Likely
Likely
Gulf
Cocagne
Medium
68.80%
Unknown
Unknown
In progress*
Malpeque
Medium
83.78%
Unknown
Unknown
St. Peters
High
98.48%
Yes
No
Nfld
South Arm
(Sept 2020)
Low
55.94%
Likely
Unlikely
Yes
South Arm
(Oct 2021)
89.71%
Yes
No
*Seasonal data from Bouctouche, New Brunswick, were collected in 2022 and will be
analyzed at a later date.
While this report provides guidance on best practices to monitoring mesozooplankton
communities, further examination of trends in both the size-spectra and community structure
will provide a comprehensive understanding of potential ecosystem effects such as food web
alterations due to bivalve aquaculture. In that purpose, the findings in this report provide
complementary results to a related departmental project that focuses on size-spectra
changes in mesozooplankton using this same dataset. Additionally, while the macro-
FlowCam may vastly reduce the time spent identifying plankton, thereby reducing the cost of
analysis per sample (Benfield et al. 2007; Álvarez et al. 2014), these procedures also require
an evaluation to identify benefits and disadvantages of the methods, and assess the
reliability or confidence of the data (Jakobsen and Carstensen 2011; Álvarez et al. 2014).
Taxa were also identified and enumerated using traditional microscopy for approximately 90
AMP samples, and comparing counts of the mesozooplankton taxa between the macro-
FlowCam and traditional microscopy is a crucial step to ensure a reliable long-term
monitoring program. Previous studies have found good agreement between zooplankton
counts derived between these two methods although it is noted the taxonomic resolution
provided by FlowCam is generally coarser (Le Bourg et al. 2015; Kydd et al. 2018; Detmer et
al. 2019; Hrycik et al. 2019). These future comparisons will help develop local information for
Canadian waters, evaluate if any corrections to FlowCam counts need to be made, and
determine if any patterns such as community-level observations vary between the
approaches. Furthermore, this work will also help refine precise sampling targets and
document any differences in sampling completeness by the two methods, due to the
potential differences in detecting rare taxa (Le Bourg et al. 2015; Stanislawczyk et al. 2018).
In addition, plankton data were simultaneously collected at AMP sites for two different size
fractions, which were then enumerated with the micro-FlowCam (plankton in the 50 - 600 um
range) and by flow cytometry (0.2 - 20 um). Although bivalves are known to ingest
mesozooplankton (Lehane and Davenport 2002; Wong and Levinton 2006; Ezgeta-Balić et
al. 2012), these smaller size fractions are of particular interest, since bivalves are known to
selectively filter for smaller particles during feeding (Lehane and Davenport 2002, 2006).
These therefore represent key properties to be analyzed when considering potential bivalve-
environment interactions. Future work will apply similar methods to those used in this report
to analyze the spatiotemporal variations in these size fractions. Interpreting patterns in the
zooplankton communities in these different size fractions will also help optimize future
sampling strategies, and propose adjustments based on factors such as sample
completeness, seasonality, stations, and tides. Continued sampling of zooplankton over a

80
range of aquaculture production levels may also further help develop carrying capacity
models, by analyzing variations in grazing intensity (Grant et al. 2005; Ibarra et al. 2014; Han
et al. 2017). In addition, these datasets will provide baseline data for analyzing potential
effects under various climate change scenarios. Incorporating a combination of datasets and
analysis approaches will provide greater knowledge of the possible effects of bivalve
aquaculture on zooplankton communities.

81
5 CONCLUSION
The results obtained in this report provide a nationally-consistent spatiotemporal assessment
of zooplankton community characterstics in coastal Canadian aquaculture lease areas.
Samples were collected from an extensive, nation-wide monitoring program, and this report
helps inform science-based recommendations to transition towards an optimal AMP
operational phase to monitor long-term ecosystem effects from bivalve aquaculture. These
results underscored the importance of considering seasonality, since large changes in
mesozooplankton assemblage structure were observed between sampling months.
Sampling during multiple (e.g., three) consecutive months is recommended, as this may, in
the long term, help differentiate climate change-related shifts in mesozooplankton
community structure, to those potentially resulting from bivalve aquaculture. In addition,
differences in mesozooplankton composition were observed between stations at each site,
while differences were often less pronounced or not detected between tide phases. For
national consistency, we recommend increasing the number of stations to a higher number
(e.g., 20) to better understand the bay-scale dynamics, and only one sample per station
would be required (i.e., no replicates). These results and guidance are part of a long-term
monitoring approach to further characterize effects from bivalve aquaculture using a
nationally-consistent sampling strategy. Understanding these dynamics through an effective
sampling design will address key knowledge gaps related to the effects of the bivalve
aquaculture industry on the environment.

82
6 ACKNOWLEDGEMENTS
Project development: David Drolet, Luc Comeau, Terri Sutherland.
Fieldwork: Kevin Chernoff, Theraesa Coyle, Brianne Kucharski, Steve R Neil, Kate Gingles,
Kim Hobbs, Vanessa Oldford, Terry Bungay, Shannon Cross, Laura Steeves.
Data management: Khang Hua, Peter Kraska.
Taxonomy and FlowCam analyses: Karen Ross (Huntsman Marine Science Centre), William
Chamberlain, Reegan Reid, Cassidy Bertin, Brontë Thomas, Jesslynn Shaw.
Provinces and industry fieldwork support: Charlene LeBlanc (Nova Scotia), Alix d’Entremont
(Nova Scotia), Terry Mills (Newfoundland), Trevor Munro (Nova Scotia).
Reviewers: Catherine Johnson, David Drolet.
Additional support: Jackson Chu, Ruben Cordero, Myriam Lacharité, Michael Kalyn.

83
7 AUTHOR CONTRIBUTIONS
ALR, TG, CM, JB and RF conceived the study. ALR, JA, JB, OG, DG and TG contributed to
data acquisition in the field. ALR, JA, RM, CG, TM developed the FlowCam protocol,
identified plankton, and processed samples. SF, ALR, RF, TG and CM did the data analyses
and interpreted results. SF, ALR, RF, TG and CM wrote the manuscript and all authors
reviewed it.

84
8 REFERENCES
Allan, J.D. 1976. Life History Patterns in Zooplankton. Am. Nat. 110(971): 165-180.
Álvarez, E., Moyano, M., López-Urrutia, Á., Nogueira, E., and Scharek, R. 2014. Routine
determination of plankton community composition and size structure: a comparison
between FlowCAM and light microscopy. J. Plankton Res. 36(1): 170-184.
Anderson, M. 2008. PERMANOVA+ for PRIMER: guide to software and statistical methods.
Primer-E Limited.
Anderson, M.J. 2006. Distance-Based Tests for Homogeneity of Multivariate Dispersions.
Biometrics 62(1): 245-253.
Anderson, M.J. 2017. Permutational Multivariate Analysis of Variance (PERMANOVA). In
Wiley StatsRef: Statistics Reference Online. pp. 1-15.
Anderson, M.J., Ellingsen, K.E., and McArdle, B.H. 2006. Multivariate dispersion as a
measure of beta diversity. Ecol. Lett. 9(6): 683-693.
Anderson, M.J., and Robinson, J. 2003. Generalized discriminant analysis based on
distances. Aust. N. Z. J. Stat. 45(3): 301-318.
Anderson, M.J., and Walsh, D.C.I. 2013. PERMANOVA, ANOSIM, and the Mantel test in the
face of heterogeneous dispersions: what null hypothesis are you testing? Ecol.
Monogr. 83(4): 557-574.
Angermeier, P.L., and Smogor, R.A. 1995. Estimating number of species and relative
abundances in stream-fish communities: effects of sampling effort and discontinuous
spatial distributions. Can. J. Fish. Aquat. Sci. 52(5): 936-949.
Arbizu, P.M. 2017. pairwiseAdonis: Pairwise Multilevel Comparison using Adonis.
Azra, M.N., Okomoda, V.T., Tabatabaei, M., Hassan, M., and Ikhwanuddin, M. 2021. The
Contributions of Shellfish Aquaculture to Global Food Security: Assessing Its
Characteristics From a Future Food Perspective. Front. Mar. Sci. 8(654897): 1-6.
Bacher, C., Filgueira, R., and Guyondet, T. 2016. Probabilistic approach of water residence
time and connectivity using Markov chains with application to tidal embayments. J.
Mar. Syst. 153: 25-41.
Batten, S.D., Abu-Alhaija, R., Chiba, S., Edwards, M., Graham, G., Jyothibabu, R.,
Kitchener, J.A., Koubbi, P., McQuatters-Gollop, A., Muxagata, E., Ostle, C.,
Richardson, A.J., Robinson, K.V., Takahashi, K.T., Verheye, H.M., and Wilson, W.
2019. A Global Plankton Diversity Monitoring Program. Front. Mar. Sci. 6(321): 1-14.
Benfield, M.C., Grosjean, P., Culverhouse, P.F., Irigoien, X., Sieracki, M.E., Lopez-Urratia,
A., Dam, H.G., Hu, Q., Davis, C.S., Hansen, A., Pilkaln, C.H., Riseman, E.M.,
Schultz, H., Utgoff, P.E., and Gorksy, G. 2007. RAPID: Research on Automated
Plankton Identification. Oceanogr. 20(2): 172-187.
Bessey, C., Jarman, S.N., Berry, O., Olsen, Y.S., Bunce, M., Simpson, T., Power, M.,
McLaughlin, J., Edgar, G.J., and Keesing, J. 2020. Maximizing fish detection with
eDNA metabarcoding. Environ. DNA. 2(4): 493-504.
Birk, S., Chapman, D., Carvalho, L., Spears, B.M., Andersen, H.E., Argillier, C., Auer, S.,
Baattrup-Pedersen, A., Banin, L., and Beklioğlu, M. 2020. Impacts of multiple

85
stressors on freshwater biota across spatial scales and ecosystems. Nat. Ecol. Evol.
4(8): 1060-1068.
Borg, I., and Groenen, P.J.F. 2005. Springer series in statistics. Modern multidimensional
scaling: Theory and applications (2nd ed.). New York, NY, US: Springer Science+
Business Media.
Brand, T. van den. 2022. ggh4x: Hacks for “ggplot2.”
Cao, Y., Larsen, D.P., and Thorne, R.S.-J. 2001. Rare species in multivariate analysis for
bioassessment: some considerations. J. North Am. Benthol. Soc. 20(1): 144-153.
Cao, Y., Williams, D.D., and Williams, N.E. 1998. How important are rare species in aquatic
community ecology and bioassessment? Limnol. Oceanogr. 43(7): 1403–1409.
Chacoff, N.P., Vázquez, D.P., Lomáscolo, S.B., Stevani, E.L., Dorado, J., and Padròn, B.
2012. Evaluating sampling completeness in a desert plant-pollinator network. J.
Anim. Ecol. 81(1): 190-200.
Chao, A. 1984. Nonparametric estimation of the number of classes in a population. Scand. J.
Stat. 11: 265-270.
Chao, A. 1987. Estimating the population size for capture-recapture data with unequal
catchability. Biometrics. 43: 783-791.
Chao, A., Colwell, R.K., Lin, C.-W., and Gotelli, N.J. 2009. Sufficient sampling for asymptotic
minimum species richness estimators. Ecol. 90(4): 1125-1133.
Chao, A., Gotelli, N.J., Hsieh, T.C., Sander, E.L., Ma, K.H., Colwell, R.K., and Ellison, A.M.
2014. Rarefaction and extrapolation with Hill numbers: a framework for sampling and
estimation in species diversity studies. Ecol. Monogr. 84(1): 45-67.
Chao, A., Kubota, Y., Zelený, D., Chiu, C.-H., Li, C.-F., Kusumoto, B., Yasuhara, M., Thorn,
S., Wei, C.-L., Costello, M.J., and Colwell, R.K. 2020. Quantifying sample
completeness and comparing diversities among assemblages. Ecol. Res. 35(2): 292-
314.
Chiarucci, A., Bacaro, G., Rocchini, D., and Fattorini, L. 2008. Discovering and rediscovering
the sample-based rarefaction formula in the ecological literature. Community Ecol.
9(1): 121-123.
Chiu, C.-H., and Chao, A. 2014. Distance-Based Functional Diversity Measures and Their
Decomposition: A Framework Based on Hill Numbers. PLoS One. 9(7): e100014.
Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure.
Aust. J. Ecol. 18(1988): 117-143.
Colwell, R.K., Chao, A., Gotelli, N.J., Lin, S.-Y., Mao, C.X., Chazdon, R.L., and Longino, J.T.
2012. Models and estimators linking individual-based and sample-based rarefaction,
extrapolation and comparison of assemblages. J. Plant Ecol. 5(1): 3-21.
Colwell, R.K. and Coddington, J.A. 1994. Estimating terrestrial biodiversity through
extrapolation. Phil. Trans. R. Soc. Lond. B. 345(1311): 101-118.
Comeau, L.A., Guyondet, T., Drolet, D., Sonier, R., Clements, J.C., Tremblay, R., and
Filgueira, R. 2023. Revisiting ecological carrying capacity indices for bivalve
aquaculture. Aquac. 577: 739911.

86
Cox, K.D., Black, M.J., Filip, N., Miller, M.R., Mohns, K., Mortimor, J., Freitas, T.R., Greiter
Loerzer, R., Gerwing, T.G., Juanes, F., and Dudas, S.E. 2017. Community
assessment techniques and the implications for rarefaction and extrapolation with Hill
numbers. Ecol. Evol. 7(24): 11213-11226.
Cranford P.J., Li W., Strand Ø., Strohmeier T. 2008 Phytoplankton depletion by mussel
aquaculture: high resolution mapping, ecosystem modeling and potential indicators of
ecological carrying capacity. ICES CM2008/H:12. International Council for the
Exploration of the Sea, Copenhagen.
Cuffney, T.F., Bilger, M.D., and Haigler, A.M. 2007. Ambiguous taxa: effects on the
characterization and interpretation of invertebrate assemblages. J. North Am.
Benthol. Soc. 26(2): 286-307.
Danielsen, F., Balete, D.S., Poulsen, M.K., Enghoff, M., Nozawa, C.M., and Jensen, A.E.
2000. A simple system for monitoring biodiversity in protected areas of a developing
country. Biodivers. Conserv. 9(12): 1671-1705.
Davenport, J., Smith, R.J.J.W., and Packer, M. 2000. Mussels Mytilus edulis: significant
consumers and destroyers of mesozooplankton. Mar. Ecol. Prog. Ser. 198: 131-137.
Detmer, T.M., Broadway, K.J., Potter, C.G., Collins, S.F., Parkos, J.J., and Wahl, D.H. 2019.
Comparison of microscopy to a semi-automated method (FlowCAM®) for
characterization of individual-, population-, and community-level measurements of
zooplankton. Hydrobiologia. 838(1): 99-110.
Domènech, M., Wangensteen, O.S., Enguídanos, A., MalumbresOlarte, J., and Arnedo,
M.A. 2022. For all audiences: Incorporating immature stages into standardised spider
inventories has a major impact on the assessment of biodiversity patterns. Mol. Ecol.
Resour. 22(6): 2319-2332.
Ezgeta-Balić, D., Najdek, M., Peharda, M., and Blažina, M. 2012. Seasonal fatty acid profile
analysis to trace origin of food sources of four commercially important bivalves.
Aquac. 334-337: 89-100.
Filgueira, R., Guyondet, T., Bacher, C., and Comeau, L.A. 2015. Informing Marine Spatial
Planning (MSP) with numerical modelling: A case-study on shellfish aquaculture in
Malpeque Bay (Eastern Canada). Mar. Pollut. Bull. 100(1): 200-216.
Filgueira, R., Guyondet, T., Comeau, L.A., and Grant, J. 2014. A fully-spatial ecosystem-
DEB model of oyster (Crassostrea virginica) carrying capacity in the Richibucto
Estuary, Eastern Canada. J. Mar. Syst. 136: 42-54.
Filgueira, R., Guyondet, T., Comeau, L.A., and Tremblay, R. 2016. Bivalve aquaculture-
environment interactions in the context of climate change. Glob. Chang. Biol. 22(12):
3901-3913.
Filgueira, R., Guyondet, T., Thupaki, P., Sakamaki, T., and Grant, J. 2021. The effect of
embayment complexity on ecological carrying capacity estimations in bivalve
aquaculture sites. J. Clean. Prod. 288: 125739.
Flaherty, M., Reid, G., Chopin, T., and Latham, E. 2019. Public attitudes towards marine
aquaculture in Canada: insights from the Pacific and Atlantic coasts. Aquac. Int. 27:
9-32.

87
Folt, C.L., and Burns, C.W. 1999. Biological drivers of zooplankton patchiness. Trends Ecol.
Evol. 14(8): 300-305.
Forget, N.L, Duplisea, D.E., Sardenne, F., and McKindsey, C.W. 2020. Using qualitative
networks to assess the influence of mussel culture on ecosystem dynamics. Ecol.
Model. 430: 109070.
Gao, C.-H. 2022. ggVennDiagram: A “ggplot2” Implement of Venn Diagram.
Gianasi, B.L., McKindsey, C.W., Tremblay, R., Comeau, L.A. and Drolet, D., 2023. Plankton
depletion by mussel grazing negatively impacts the fitness of lobster
larvae. Aquaculture. 574: 739659.
Gotelli, N.J., and Colwell, R.K. 2001. Quantifying biodiversity: procedures and pitfalls in the
measurement and comparison of species richness. Ecol. Lett. 4(4): 379-391.
Grant, J., Bacher, C., Cranford, P.J., Guyondet, T., and Carreau, M. 2008. A spatially explicit
ecosystem model of seston depletion in dense mussel culture. J. Mar. Syst. 73(1-2):
155-168.
Grant, J., Bugden, G., Horne, E., Archambault, M.-C., and Carreau, M. 2007. Remote
sensing of particle depletion by coastal suspension-feeders. Can. J. Fish. Aquat. Sci.
64(3): 387-390.
Grant, J., Cranford, P., Hargrave, B., Carreau, M., Schofield, B., Armsworthy, S., Burdett-
Coutts, V., and Ibarra, D. 2005. A model of aquaculture biodeposition for multiple
estuaries and field validation at blue mussel (Mytilus edulis) culture sites in eastern
Canada. Can. J. Fish. Aquat. Sci. 62(6): 1271-1285.
Grant, J., and Pastres, R. 2019. Ecosystem models of bivalve aquaculture: implications for
supporting goods and services. Goods and Services of Marine Bivalves: 507-525.
Springer International Publishing.
Gutierrez, M.F., Epele, L.B., Mayora, G., Aquino, D., Mora, C., Quintana, R., and Mesa, L.
2022. Hydro-climatic changes promote shifts in zooplankton composition and
diversity in wetlands of the Lower Paraná River Delta. Hydrobiologia. 849(16): 3463-
3480.
Guyondet, T., Comeau, L.A., Bacher, C., Grant, J., Rosland, R., Sonier, R., and Filgueira, R.
2015. Climate Change Influences Carrying Capacity in a Coastal Embayment
Dedicated to Shellfish Aquaculture. Estuaries Coasts. 38(5): 1593-1618.
Han, D., Chen, Y., Zhang, C., Ren, Y., Xue, Y., and Wan, R. 2017. Evaluating impacts of
intensive shellfish aquaculture on a semi-closed marine ecosystem. Ecol. Modell.
359: 193-200.
Helenius, L.K., Leskinen, E., Lehtonen, H., and Nurminen, L. 2017. Spatial patterns of littoral
zooplankton assemblages along a salinity gradient in a brackish sea: A functional
diversity perspective. Estuar. Coast. Shelf Sci. 198: 400-412.
Hill, M.O. 1973. Diversity and Evenness: A Unifying Notation and Its Consequences. Ecol.
54(2): 427-432.
Hoffman, J.C., Kelly, J.R., Trebitz, A.S., Peterson, G.S., and West, C.W. 2011. Effort and
potential efficiencies for aquatic non-native species early detection. Can. J. Fish.
Aquat. Sci. 68(12): 2064-2079.

88
Holden, J.J., Collicutt, B., Covernton, G., Cox, K.D., Lancaster, D., Dudas, S.E., Ban, N.C.,
and Jacob, A.L. 2019. Synergies on the coast: Challenges facing shellfish
aquaculture development on the central and north coast of British Columbia. Mar.
Policy. 101: 108-117.
Hrycik, A.R., Shambaugh, A., and Stockwell, J.D. 2019. Comparison of FlowCAM and
microscope biovolume measurements for a diverse freshwater phytoplankton
community. J. Plankton Res. 41(6): 849-864.
Hsieh, T.C., Ma, K.H., and Chao, A. 2016. iNEXT: an R package for rarefaction and
extrapolation of species diversity (Hill numbers). Methods Ecol. Evol. 7(12): 1451-
1456.
Hulot, F.D., Lacroix, G., and Loreau, M. 2014. Differential responses of size-based functional
groups to bottom-up and top-down perturbations in pelagic food webs: a meta-
analysis. Oikos. 123(11): 1291-1300.
Hulot, V., Saulnier, D., Lafabrie, C., and GaertnerMazouni, N. 2020. Shellfish culture: a
complex driver of planktonic communities. Rev. Aquac. 12(1): 33-46.
Ibarbalz, F.M., Henry, N., Brandão, M.C., Martini, S., Busseni, G., Byrne, H., Coelho, L.P.,
Endo, H., Gasol, J.M., Gregory, A.C., Mahé, F., Rigonato, J., Royo-Llonch, M.,
Salazar, G., Sanz-Sáez, I., Scalco, E., Soviadan, D., Zayed, A.A., Zingone, A.,
Labadie, K., Ferland, J., Marec, C., Kandels, S., Picheral, M., Dimier, C., Poulain, J.,
Pisarev, S., Carmichael, M., Pesant, S., Babin, M., Boss, E., Iudicone, D., Jaillon, O.,
Acinas, S.G., Ogata, H., Pelletier, E., Stemmann, L., Sullivan, M.B., Sunagawa, S.,
Bopp, L., de Vargas, C., Karp-Boss, L., Wincker, P., Lombard, F., Bowler, C., Zinger,
L., Acinas, S.G., Babin, M., Bork, P., Boss, E., Bowler, C., Cochrane, G., de Vargas,
C., Follows, M., Gorsky, G., Grimsley, N., Guidi, L., Hingamp, P., Iudicone, D.,
Jaillon, O., Kandels, S., Karp-Boss, L., Karsenti, E., Not, F., Ogata, H., Pesant, S.,
Poulton, N., Raes, J., Sardet, C., Speich, S., Stemmann, L., Sullivan, M.B.,
Sunagawa, S., and Wincker, P. 2019. Global Trends in Marine Plankton Diversity
across Kingdoms of Life. Cell. 179(5): 1084-1097.e21.
Ibarra, D.A., Fennel, K., and Cullen, J.J. 2014. Coupling 3-D Eulerian bio-physics (ROMS)
with individual-based shellfish ecophysiology (SHELL-E): A hybrid model for carrying
capacity and environmental impacts of bivalve aquaculture. Ecol. Modell. 273: 63-78.
Irisson, J.-O., Ayata, S.-D., Lindsay, D.J., Karp-Boss, L., and Stemmann, L. 2022. Machine
Learning for the Study of Plankton and Marine Snow from Images. Annu. Rev. Mar.
Sci. 14(1): 277-301.
Jakobsen, H., and Carstensen, J. 2011. FlowCAM: Sizing cells and understanding the
impact of size distributions on biovolume of planktonic community structure. Aquat.
Microb. Ecol. 65(1): 75-87.
Ji, R., Edwards, M., Mackas, D.L., Runge, J.A., and Thomas, A.C. 2010. Marine plankton
phenology and life history in a changing climate: current research and future
directions. J. Plankton Res. 32(10): 1355-1368.
Johnson, W.S., and Allen, D.M. 2012. Zooplankton of the Atlantic and Gulf coasts: a guide to
their identification and ecology. JHU Press.
Jost, L. 2006. Entropy and diversity. Oikos 113(2): 363-375.

89
Kerr, T., Clark, J.R., Fileman, E.S., Widdicombe, C.E., and Pugeault, N. 2020. Collaborative
Deep Learning Models to Handle Class Imbalance in FlowCam Plankton Imagery.
IEEE Access. 8: 170013-170032.
Kiørboe, T. 2009. A Mechanistic Approach to Plankton Ecology. ASLO Web Lectures 1(2):
1-91. John Wiley & Sons, Ltd.
Kosobokova, K.N., Hopcroft, R.R., and Hirche, H.-J. 2011. Patterns of zooplankton diversity
through the depths of the Arctic’s central basins. Mar. Biodivers. 41(1): 29-50.
Kruskal, J.B. 1964. Multidimensional scaling by optimizing goodness of fit to a nonmetric
hypothesis. Psychometrika. 29(1): 1-27.
Kydd, J., Rajakaruna, H., Briski, E., and Bailey, S. 2018. Examination of a high resolution
laser optical plankton counter and FlowCAM for measuring plankton concentration
and size. J. Sea Res. 133: 2-10.
Lamb, J. 2005. Ecological zonation of zooplankton in the COAST study region off central
Oregon in June and August 2001 with consideration of retention mechanisms. J.
Geophys. Res. 110(C10): C10S15.
Lavaud, R., Guyondet, T., Filgueira, R., Tremblay, R., and Comeau, L.A. 2020. Modelling
bivalve culture - Eutrophication interactions in shallow coastal ecosystems. Mar.
Pollut. Bull. 157: 111282.
Le Bourg, B., Cornet-Barthaux, V., Pagano, M., and Blanchot, J. 2015. FlowCAM as a tool
for studying small (80–1000 µm) metazooplankton communities. J. Plankton Res.
37(4): 666-670.
Lehane, C., and Davenport, J. 2002. Ingestion of mesozooplankton by three species of
bivalve; Mytilus edulis, Cerastoderma edule and Aequipecten opercularis. J. Mar.
Biolog. Assoc. 82(4): 615-619.
Lehane, C., and Davenport, J. 2006. A 15-month study of zooplankton ingestion by farmed
mussels (Mytilus edulis) in Bantry Bay, Southwest Ireland. Estuar. Coast. Shelf Sci.
67(4): 645-652.
Lindeman, R.L. 1942. The Trophic-Dynamic Aspect of Ecology. Ecol. 23(4): 399-417.
Luo, J.Y., Irisson, J., Graham, B., Guigand, C., Sarafraz, A., Mader, C., and Cowen, R.K.
2018. Automated plankton image analysis using convolutional neural networks.
Limnol. Oceanogr.: Methods. 16(12): 814-827.
Ma, K.C.K., McKindsey, C.W., and Johnson, L.E. 2022. Detecting Rare Species With
Passive Sampling Tools: Optimizing the Duration and Frequency of Sampling for
Benthic Taxa. Front. Mar. Sci. 9: 809327.
Maar, M., Nielsen, T., and Petersen, J. 2008. Depletion of plankton in a raft culture of Mytilus
galloprovincialis in Ría de Vigo, NW Spain. II. Zooplankton. Aquat. Biol. 4: 127-141.
Mackas, D.L., Pepin, P., and Verheye, H. 2012. Interannual variability of marine zooplankton
and their environments: Within- and between-region comparisons. Prog. Oceanogr.
97-100: 1-14.
Magurran, A.E., and McGill, B.J. 2010. Biological diversity: frontiers in measurement and
assessment. OUP Oxford.

90
Marques, S.C., Pardal, M.A., Pereira, M.J., Gonçalves, F., Marques, J.C., and Azeiteiro,
U.M. 2007. Zooplankton distribution and dynamics in a temperate shallow estuary.
Hydrobiologia 587: 213-223.
McIntire, E.J.B., and Fajardo, A. 2009. Beyond description: the active and effective way to
infer processes from spatial patterns. Ecol. 90(1): 46-56.
Navarro, D. 2015. Learning statistics with R: A tutorial for psychology students and other
beginners. (Version 0.6). University of New South Wales, Sydney, Australia.
Neuheimer, A.B., Gentleman, W.C., Pepin, P., and Head, E.J.H. 2010. Explaining regional
variability in copepod recruitment: Implications for a changing climate. Prog.
Oceanogr. 87(1-4): 94-105.
O’Brien, T.D., and Oakes, S.A. 2020. Visualizing and exploring zooplankton spatio-temporal
variability. In Zooplankton Ecology. CRC Press. pp. 192-224.
Oksanen, J., Simpson, G.L., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O’Hara,
R.B., Solymos, P., Stevens, M.H.H., Szoecs, E., Wagner, H., Barbour, M., Bedward,
M., Bolker, B., Borcard, D., Carvalho, G., Chirico, M., Caceres, M.D., Durand, S.,
Evangelista, H.B.A., FitzJohn, R., Friendly, M., Furneaux, B., Hannigan, G., Hill,
M.O., Lahti, L., McGlinn, D., Ouellette, M.-H., Cunha, E.R., Smith, T., Stier, A., Braak,
C.J.F.T., and Weedon, J. 2022. vegan: Community Ecology Package.
Olsen, A.R., Sedransk, J., Edwards, D., Gotway, C.A., Liggett, W., Rathbun, S., Reckhow,
K.H., and Yyoung, L.J. 1999. Statistical Issues for Monitoring Ecological and Natural
Resources in the United States. Environ. Monit. Assess. 54(1): 1-45.
Orenstein, E.C., Ayata, S., Maps, F., Becker, É.C., Benedetti, F., Biard, T., de Garidel
Thoron, T., Ellen, J.S., Ferrario, F., Giering, S.L.C., GuyHaim, T., Hoebeke, L.,
Iversen, M.H., Kiørboe, T., Lalonde, J., Lana, A., Laviale, M., Lombard, F., Lorimer,
T., Martini, S., Meyer, A., Möller, K.O., Niehoff, B., Ohman, M.D., Pradalier, C.,
Romagnan, J., Schröder, S., Sonnet, V., Sosik, H.M., Stemmann, L.S., Stock, M.,
TerbiyikKurt, T., ValcárcelPérez, N., Vilgrain, L., Wacquet, G., Waite, A.M., and
Irisson, J. 2022. Machine learning techniques to characterize functional traits of
plankton from image data. Limnol. Oceanogr.: Methods. 67(8): 1647-1669.
Owen, B.M., Hallett, C.S., Cosgrove, J.J., Tweedley, J.R., and Moheimani, N.R. 2022.
Reporting of methods for automated devices: A systematic review and
recommendation for studies using FlowCam for phytoplankton. Limnol. Oceanogr.:
Methods. 20(7): 373-465.
Pace, M.L., Strayer, D.L., Fischer, D., and Malcom, H.M. 2010. Recovery of native
zooplankton associated with increased mortality of an invasive mussel. Ecosphere.
1(1): 1-21.
R Core Team. 2022. R: A Language and Environment for Statistical Computing. R
Foundation for Statistical Computing, Vienna, Austria.
Ramkissoon, C. 2021. Assessment of Convolutional Neural Networks to Classify Freshwater
Zooplankton. University of Toronto (Canada).
Ratnarajah, L., Abu-Alhaija, R., Atkinson, A., Batten, S., Bax, N.J., Bernard, K.S., Canonico,
G., Cornils, A., Everett, J.D., Grigoratou, M., Ishak, N.H.A., Johns, D., Lombard, F.,
Muxagata, E., Ostle, C., Pitois, S., Richardson, A.J., Schmidt, K., Stemmann, L.,

91
Swadling, K.M., Yang, G., and Yebra, L. 2023. Monitoring and modelling marine
zooplankton in a changing climate. Nat. Commun. 14(1): 564.
Robertson, O.J., McAlpine, C., House, A., and Maron, M. 2013. Influence of Interspecific
Competition and Landscape Structure on Spatial Homogenization of Avian
Assemblages. PLoS One. 8(5): e65299.
Schartau, A.K., Mariash, H.L., Christoffersen, K.S., Bogan, D., Dubovskaya, O.P., Fefilova,
E.B., Hayden, B., Ingvason, H.R., Ivanova, E.A., Kononova, O.N., Kravchuk, E.S.,
Lento, J., Majaneva, M., Novichkova, A.A., Rautio, M., Rühland, K.M., Shaftel, R.,
Smol, J.P., Vrede, T., and Kahilainen, K.K. 2022. First circumpolar assessment of
Arctic freshwater phytoplankton and zooplankton diversity. Freshw. Biol. 61(1): 141-
158.
Schlacher, T.A., and Wooldridge, T.H. 1995. Small-scale distribution and variability of
demersal zooplankton in a shallow, temperate estuary: tidal and depth effects on
species-specific heterogeneity. Cah. Biol. Mar. 36: 211-227.
Soetaert, K.E.R., and Van Rijswijk, P. 1993. Spatial and temporal patterns of the
zooplankton in the Westerschelde estuary. Mar. Ecol. Prog. Ser. 97: 47-49.
Stanislawczyk, K., Johansson, M.L., and MacIsaac, H.J. 2018. Microscopy versus
automated imaging flow cytometry for detecting and identifying rare zooplankton.
Hydrobiologia. 807(1): 53-65.
Steeves, L.E., Filgueira, R., Guyondet, T., Chassé, J., and Comeau, L. 2018. Past, Present,
and Future: Performance of Two Bivalve Species Under Changing Environmental
Conditions. Front. Mar. Sci. 5(184): 1-14.
Taylor, D., Larsen, J., Buer, A.-L., Friedland, R., Holbach, A., Petersen, J.K., Nielsen, P.,
Ritzenhofen, L., Saurel, C., and Maar, M. 2021. Mechanisms influencing particle
depletion in and around mussel farms in different environments. Ecol. Inidic. 122:
107304.
Tommasi, D.A.G., Routledge, R.D., Hunt, B.P.V., and Pakhomov, E.A. 2013. The seasonal
development of the zooplankton community in a British Columbia (Canada) fjord
during two years with different spring bloom timing. Mar. Biol. Res. 9(2): 129-144.
Usov, N., Khaitov, V., Smirnov, V., and Sukhotin, A. 2019. Spatial and temporal variation of
hydrological characteristics and zooplankton community composition influenced by
freshwater runoff in the shallow Pechora Sea. Polar Biol. 42: 1647-1665.
Varpe, O. 2012. Fitness and phenology: annual routines and zooplankton adaptations to
seasonal cycles. J. Plankton Res. 34(4): 267-276.
Weitzman, J., Steeves, L., Bradford, J. and Filgueira, R., 2019. Far-field and near-field
effects of marine aquaculture. World seas: An environmental evaluation, pp.197-220.
Wickham, H. 2016. ggplot2: Elegant Graphics for Data Analysis. Springer-Verlag New York.
Wijsman, J.W.M., Troost, K., Fang, J., and Roncarati, A. 2019. Global production of marine
bivalves. Trends and challenges. Goods and services of marine bivalves: 7-26.
Springer International Publishing.

92
Wong, W.H., and Levinton, J.S. 2006. The trophic linkage between zooplankton and benthic
suspension feeders: direct evidence from analyses of bivalve faecal pellets. Mar.
Biol. 148(4): 799-805.
Wooldridge, T., and Erasmus, T. 1980. Utilization of tidal currents by estuarine zooplankton.
Estuar. Coast. Mar. Sci. 11(1): 107-114.
Yoccoz, N.G., Nichols, J.D., and Boulinier, T. 2003. Monitoring of biological diversity–a
response to Danielsen et al. Oryx. 37(4): 410.
Yokogawa Fluid Imaging Technologies, Inc. 2020. FlowCam Macro Imaging Particle
Analysis System User Guide: ViSp 4 Version B.
Yokogawa Fluid Imaging Technologies, Inc. n.d. VisualSpreadsheet 5 Version 5.6.19 User
Guide.
Zeldis, J., Robinson, K., Ross, A., and Hayden, B. 2004. First observations of predation by
New Zealand Greenshell mussels (Perna canaliculus) on zooplankton. J. Exp. Mar.
Biol. Ecol. 311(2): 287-299.

93
APPENDIX 1
Data summary table for mesozooplankton (0.25 mm - 5.00 mm) samples obtained within
study sites and regions, as part of the Aquaculture Monitoring Program. Station names refer
to the sampling locations within each site as indicated in Fig 1. The station-tide count
represents the number of samples obtained at each respective station and tide phase
combination for each site and/or month. Samples were obtained for more than one month for
Cocagne (Gulf), South Arm (Newfoundland), and Lemmens Inlet (Pacific). M-R: mid-rising,
M-F: mid-falling. See Fig. 1 for station names and labels. Pac: Pacific, Mar: Maritimes, Nfld:
Newfoundland.
Region
Site or
month
Year
Date
range
Distance
between
stations (km)
Station
name
Tide
phase
Station-
tide count
Pac
Lemmens
Aug 2020
2020
Aug 29-
31
In-Mid: 1.8
Mid-Out: 1.4
In-Out: 3.2
Inner
High
3
Inner
Low
3
Mid
High
3
Low
3
Outer
High
3
Low
3
18 total
Lemmens
Mar 2021
N/A
High
1
N/A
Low
1
2 total
Lemmens
Jun 2021
Inner
High
3
Inner
Low
3
Mid
High
3
Low
3
Outer
High
3
Low
3
18 total
Lemmens
Sept 2021
Inner
High
2
Inner
Low
2
Mid
High
2
Low
2
Outer
High
2
Low
2
12 total
Mar
Argyle
2021
Aug 30-
Sep 1
S-C: 5.2
C-N: 3.1
S-N: 8.3
Central
High
2
Low
3
North
High
2
Low
2
M-R
1
South
High
2
Low
2
M-R
1

94
Region
Site or
month
Year
Date
range
Distance
between
stations (km)
Station
name
Tide
phase
Station-
tide count
15 total
Country
Harbour
2021
Aug 24
In-Mid: 3.2
Mid-Out: 3.0
In-Out: 6.2
Inner
High
1
Low
1
Mid
High
1
Low
1
Outer
High
1
Low
1
6 total
Sober
Island
2021
Aug 27
Out-IS: 0.7
Out-IN: 1.0
IN-IS: 0.5
Inner-
North
M-R
3
Inner-
South
M-R
3
Outer
High
3
Low
3
12 total
Whitehead
2021
Aug 25
In-Mid: 1.1
Mid-Out: 2.1
In-Out: 3.1
Inner
High
1
Low
1
M-F
1
Mid
High
1
Low
1
M-F
1
Outer
High
1
Low
1
M-F
1
9 total
Gulf
Malpeque
2020
Sept 29
S-C: 3.6
C-N: 3.7
S-N: 4.5
Central
Low
1
North
Low
1
South
Low
1
3 total
St. Peters
2020
Sept 1-4
In-Mid: 4.0
Mid-Out: 4.1
In-Out: 8.1
Inner
High
4
Low
3
Mid
High
4
Low
3
M-F
4
Outer
High
5
Low
3
26 total
Cocagne
2021
Jul 21
S-C: 2.4
C-N: 2.5
S-N: 4.4
Central
Low
1
North
Low
1

95
Region
Site or
month
Year
Date
range
Distance
between
stations (km)
Station
name
Tide
phase
Station-
tide count
South
Low
1
2021
Aug 26
Central
M-R
1
North
M-R
1
South
M-R
1
6 total
Nfld.
South Arm
Sep 2020
2020
Sep 15-
16
In-Mid A: 2.5
In-Mid B: 2.0
In-Mid C: 1.5
In-Out: 3.8
Mid A-Mid B:
0.7
Mid A-Mid C:
1.1
Mid A-Out:
1.4
Mid-B-Mid-
C: 0.7
Mid B-Out:
2.1
Mid C-Out:
2.3
Inner
Low
1
M-F
1
M-R
1
Mid A
Low
1
M-F
1
Mid B
Low
1
M-F
1
Mid C
Low
1
M-F
1
M-R
1
South Arm
Jun 2021
2021
Jun 9
Mid B
Low
1
Outer
M-F
1
South Arm
Jul 2021
Jul 7
Mid-B
M-R
1
Outer
Low
1
South Arm
Aug 2021
Aug 12
Mid B
Low
1
Outer
M-F
1
South Arm
Sep 2021
Sept 8
Mid B
Low
1
Outer
M-F
1
South Arm
Oct 2021
Oct 5-7
Mid-B
High
2
Low
3
M-F
1
Outer
High
3
Low
3
South Arm
Nov 2021
Nov 9
Mid A
High
1
Mid B
M-F
1

96
Region
Site or
month
Year
Date
range
Distance
between
stations (km)
Station
name
Tide
phase
Station-
tide count
Outer
High
1
South Arm
Dec 2021
Dec 14
Mid B
Low
1
Outer
Low
1
South Arm
Feb 2022
2022
Feb 8
Mid A
Low
1
Mid B
Low
1
Outer
Low
1
South Arm
Mar 2022
Mar 29
Mid A
Low
1
Mid B
Low
1
Outer
Low
1
South Arm
Apr 2022
Apr 22
Mid A
High
1
Mid B
High
1
Outer
M-R
1
South Arm
May 2022
May 17
Mid A
M-F
1
Mid-B
Low
1
Outer
High
1
South Arm
Jun 2022
Jun 7
Mid A
M-R
1
Mid B
High
1
Outer
M-R
1
South Arm
Jul 2022
Jul 6
Mid A
High
1
Mid B
High
1
Outer
M-R
1
53 total

97
APPENDIX 2
Technical specifications, sampling and setup protocols, and image sorting details used by
the macro FlowCam within this report. The templates were obtained from the Supporting
Information of Owen et al. (2022), which outlined the critical details to include in studies that
use FlowCam technology for the identification of plankton specimens. Note that the final
table provided in Owen et al. (2022) (“Measurement Outputs”) is not included, as this
includes the methods for obtaining properties such as particle size, which are being more
extensively reviewed as part of ongoing departmental work.
FlowCam technical specifications
FlowCam model number
FlowCam Macro
FlowCam unit serial number
10416
Camera resolution
1920 x 1200 pixels
Camera color/monochrome
8-bit monochrome
Fluidics
Peristaltic pump
Software details
VisualSpreadsheet version 5.6.14 was used for the
samples collected in 2020 in Newfoundland
(FlowCam Yokogawa Fluid Imaging Technologies,
Inc. n.d.). VisualSpreadsheet version 4.18.5 used
for all other samples (FlowCam Yokogawa Fluid
Imaging Technologies 2020).
Any additional upgrades to the
machine or optional accessories
Instead of using the sample container provided with
the FlowCam, a glass beaker was used to hold the
sample. A magnetic stirrer was added to prevent
clumping of the sample.
Sample details
Preservation methods
Samples were preserved in a 4% solution of buffered
formaldehyde. These were run through the FlowCam within
1-5 months from the date of field collection.
Dilution or concentration
details
Raw counts were converted to abundance in seawater
following Equation (1) in the main text. See the process
below for additional steps before running the samples
through the FlowCam.
Pre-filtration details
Samples were rinsed through a series of stacked sieves; 2
mm mesh sieve stacked on a 125 μm or 212 μm mesh sieve.
These size fractions determined the size of the flow cell for
subsequent analysis:
● The specimens collected on the 125 μm or 212 μm
mesh were run through the FlowCam using a 2 mm
flow cell.
● The specimens collected on the 2 mm mesh that
were <5 mm were processed using a 5 mm flow cell.
The resulting specimens from the fraction sizes listed above
were kept separate. The collected specimens were then
rinsed into a beaker containing approximately 400 ml of 0.2%
Triton-x. Using a large volume of 0.2% Triton-x and a

98
Sample details
magnetic stirrer in the sample beaker was shown to be
successful in reducing the clumping of plankton.
Cell concentration range
The average abundance of zooplankton in seawater per
sample was 3486 ind m-3. Future work will compare counts
as measured by the FlowCam to counts obtained from
traditional microscopy.
Sample particle
composition
The amount of debris/detritus was variable per sample, but
was generally highest in samples from the Pacific region,
and lowest in samples from the Newfoundland region.
FlowCam setup details
Flow cell sizes and types
used, and objectives used
for each flow cell
As described above, both the 2 mm and 5 mm flow cells
were used, depending on the particle size. The following
parameters were then used for each of the flow calls:
Parameter
2 mm flow cell
5 mm flow cell
Flow cell depth
2.0 mm
5.0 mm
Flow cell width
10.5 mm
10.5 mm
Area of flow cell
imaged
11 x 17.5 mm
11 x 17.5 mm
Flow rate
215 ml/min (9
frames per
second)
400 ml/min (7
frames per
second)
Image acquisition mode
Auto-image
Sample volume analyzed
Auto-image acquisition mode was used with recirculating
water. For example, if 500 ml of water was in the sample
beaker, the water and particles pass through the FlowCam.
Upon exiting the FlowCam, particles are collected on mesh.
The water passes through the mesh, and is added back into
the sample beaker and recirculated again through the
FlowCam setup.
Cell density determination
Not applicable to this project.
Full context settings
The following parameters were applied for each of the
FlowCam settings:
Particle segmentation: Dark threshold 40.00
Distance to nearest neighbour: 4.00 μm
Close holes: 1 iteration
Basic size filter: area based diameter (ABD); Minimum of
150.00 μm, Maximum of 999999999999999.00 μm
Advanced filter: none
AutoImage frame rate: 10 frames per second
Flash duration: 600.00 microseconds
Exposure: 19
Camera trigger delay: 500 microseconds
Flash amplitude: 50%
Camera gain: 30

99
Image sorting details
Image
sorting
method
A detailed description of these processes will be described in future work.
Briefly, VisualSpreadsheet software was used for semi-automated sorting
of all particles upon being imaged.
First, the software measures the length of all particles, and those <250 μm
are removed, since only the macrozooplankton size fraction is considered
(i.e., >250 μm). Next, the images are sorted according to various
morphological characteristics, using settings provided by the software.
These roughly group the images based on their shape. This then allows the
taxonomists to review and confirm the groupings, in which the images are
sorted into various predefined classes such as bubbles, clumped
zooplankton, debris, fragments of zooplankton, and zooplankton. For this
analysis, only images that are confirmed “zooplankton” are considered. Due
to time and financial constraints, generally only a portion of the sample is
reviewed to separate out the zooplankton from these other classes.
Once the zooplankton have been separated from the other particles, the
images within the “zooplankton” class can also be grouped based on
similar shapes. This is used to quickly group similar taxa based on
morphology. The taxonomists then identify the specimens to the taxonomic
levels specified in Appendix 3. Due to time and financial constraints,
generally only a portion of these zooplankton are identified by the
taxonomist.
Software
used, with
version
details
VisualSpreadsheet version 5.6.14 was used for the samples collected in
2020 in Newfoundland (FlowCam Yokogawa Fluid Imaging Technologies,
Inc. n.d.). VisualSpreadsheet version 4.18.5 used for all other samples
(FlowCam Yokogawa Fluid Imaging Technologies 2020; reference also
includes the General User Guide for the FlowCam Macro Flow Imaging
Particle Analyzer).
Image
library
description
and sizes
A total of 1,054,779 images have been identified by the FlowCam, including
384,593 images of confirmed zooplankton specimens (see Table 2 in the
main text for a breakdown by region.) The remaining 670,186 images
contained particles that were removed from this analysis (e.g., bubbles,
clumped zooplankton, debris, fragments of zooplankton, etc.)
The taxonomic resolution of the identified specimens is provided in
Appendix 3 in the main text.
A complete breakdown of the number of images per taxa will be provided in
future work (i.e., in the comparisons with the Quantitative Assessment data
provided by traditional microscopy)
Particle
property
selections
For this analysis, only counts of zooplankton taxa were included in the
analysis. Other properties for size structure analysis are being evaluated as
part of ongoing departmental work.
Other
setting
choices
None
Evaluation
of the
accuracy of
auto-
classificatio
ns
Not applicable to this Technical Report.

100
APPENDIX 3
Guidelines specifying the taxonomic levels (e.g., Order, Genus, etc.) used for identification of
the mesozooplankton. These represent the lowest taxonomic level the individuals can
reliably be identified to, based on distinguishable morphological features, using images
obtained with the macro-FlowCam.
Taxon
Stages
Level (may be identified less
precisely depending on
stage/condition of specimen)
Decapoda
Zoea, megalopa
Order, unless distinctive gross
morphology (Brachyura,
Homarus, Porcellanidae, etc.)
Euphausiacea
Nauplii, Calyptopis, Furcilia
Order
Juvenile/adult
Genus
Mysidacea
Embryo, Juvenile/Adult
Order
Cumacea
Juvenile/adult
Order
Nebaliacea
Adult
Order
Cladocera
Adult
Genus (unless generic distinction
requires minute details, then
group together e.g.,
Podon/Pleopis spp.)
Cladocera (Freshwater)
Adult
Family
Amphipoda (pelagic)
Juvenile/adult
Genus (hyperiids)
Family (gammarids)
Acarina
Adult
Family
Facetotecta
Nauplii, Cypris
Infraclass
Cirripedia
Nauplii, Cypris
Infraclass
Invertebrate
Eggs, Trochophore larvae
“Invertebrate”
Polychaeta
Larvae
Class
Polychaeta (pelagic)
Juvenile/Adult
Genus
Gastropoda
Larvae/small species of
Limacina
“Gastropoda (larvae/Limacina)”
Gastropoda (pelagic)
Adults (including large species
of Limacina)
Genus
Bivalvia
Larvae
Class
Echinodermata
Larvae
Phylum
Bryozoa
Larvae
Phylum
Fish
Eggs, larvae
Class
Larvacea
Adult
Genus
Ascidiacea
Larvae
Class
Thaliacea
Order
Chaetognatha
Juvenile/adult
Phylum
Cnidaria
Larvae
Phylum
Siphonophorae
Nectophore, Eudoxid
Suborder
Hydrozoa
Medusa
Class unless distinctive gross
morphology (Aglantha digitale,
Pandeidae, Bougainvillidae, etc.)
Scyphozoa
Ephyra larvae, Medusa
Class
Ctenophora
Larvae/adult
Phylum

101
APPENDIX 4
Due to damaged specimens, blurry photos, or poor orientation of the images, at times,
specimens could not be identified to the appropriate level, and were instead labeled as
“unidentified Calanoida” (i.e., could be identified to the order Calanoida but not further;
includes specimens from stages Ci-Cvi; all copepod nauplii are instead classified as
“Copepoda nauplii”), “unidentified Copepoda” (i.e., could be identified to the subclass
Copepoda, but no further; includes specimens from stages Ci-Cvi), or “unidentified
zooplankton” (confirmed to be zooplankton, but could not be identified further). In addition, in
other instances, copepods identified to the order level were given overlapping stage names
(e.g., i-iii, i-iv, v-vi), or could not be identified at the genus level, but could instead be
identified at the order level (Calanoida, Cyclopoida, Monstrilloida) and were given the stage
classification “Ci-Cvi”. Therefore, copepods identified to order contained specimens from a
range of stages from Ci-Cvi, not just Ci-Ciii.
We then used best practices from Cuffney et al. (2007) to resolve these taxonomic
ambiguities. Cuffney et al. (2007) presented 16 approaches to address how parent (i.e.,
higher taxonomic levels) and child (lower taxonomic levels) taxa should be redistributed
when both are present. As recommended, we used the Distribute Parent Among Child
(DPAC) method, as it had among the highest suitability scores as measured by 13 metrics.
For copepods and unidentified zooplankton, the variant DPAC-S was applied, in which the
abundances of ambiguous parent taxa redistributed among the associated children in
proportion to the relative abundance of each child in the sample (-S). Therefore, the
abundances of parent taxa (all stages, Ci-Cvi grouped) belonging to the orders Calanoida
and Cyclopoida were redistributed among child taxa within each sample, based on their
relative abundances. In only two instances, samples contained parent taxa with no lower-
level child taxa. As recommended by Cuffney et al. (2007), the parent taxa in these cases
(Cyclopoida) were redistributed among child taxa, based on the average relative abundance
of child taxa per sample from all samples at the respective site or sampling month.
Harpactoida were not redistributed among child taxa, as the Harpacticoida classification only
contained epibenthic taxa and taxa identified to the genus level were pelagic specimens.
Next, “unidentified Copepoda” were redistributed among any Copepoda taxa, based on the
relative abundances within each sample. Lastly, “unidentified zooplankton” were
redistributed among all taxa, based on their relative abundances within each sample.
Following the redistributions, a few higher-level parent taxa persisted with lower-level
children. However, these were distinct stages, identified to the requested taxonomic level
(Appendix 4). For example, Copepoda nauplii remained as a distinct taxonomic class,
separate from the (distributed) Copepoda genera containing stages Ci-Cvi. In addition,
invertebrate (eggs and trochophores) remained as a class (all the zooplankton taxa except
Osteichthyes egg/larvae are invertebrates), and Cnidaria larvae had child taxa as Hydrozoa
(medusa), Siphonophorae (nectophore), or Scyphozoa (medusa). For the taxa accumulation
curves in Section 2.6.1, these taxa (i.e., Copepoda nauplii, invertebrate eggs and
trochophores, and Cnidaria larvae) were removed from the analyses, to prevent double
counting of taxa (e.g., Kosobokova et al. 2011). For all other analyses, these three taxa were
retained in the datasets. These parent taxa were not redistributed among children since
stage information can have a large impact on the observed biodiversity patterns in
multivariate analyses (Domènech et al. 2022), the different stages may have distinct roles in
the ecosystem (Allan 1976; Johnson and Allen 2012), and they may respond differently to
environmental conditions (Lamb 2005; Varpe 2012). Furthermore, zooplankton may exhibit
time lags between stages in their life cycles (Allan 1976; Varpe 2012); therefore,
redistributing the abundances of early stages based on the existing relative abundance of
adult stages of lower-level taxa may not be accurate.