Content uploaded by Steven X. Cadrin
Author content
All content in this area was uploaded by Steven X. Cadrin on Mar 01, 2023
Content may be subject to copyright.
Fisheries Research 262 (2023) 106650
Available online 11 February 2023
0165-7836/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
Best practices for dening spatial boundaries and spatial structure in
stock assessment
Steven X. Cadrin
a
,
*
, Daniel R. Goethel
b
, Aaron Berger
c
, Ernesto Jardim
d
a
University of Massachusetts Dartmouth, School for Marine Science and Technology, Department of Fisheries Oceanography, 836 South Rodney French Boulevard, New
Bedford, MA 02744, USA
b
NOAA, Alaska Fisheries Science Center, 17109 Point Lena Loop Road, Juneau, AK 99801, USA
c
NOAA, Northwest Fisheries Science Center, 2032 SE OSU Drive, Newport, OR 97365, USA
d
Marine Stewardship Council, 1 Snow Hill, London EC1A 2DH, UK
ARTICLE INFO
Keywords:
Fishery stock assessment
Stock identication
Spatial
Fleets
ABSTRACT
The ‘stock concept’ in sheries science conforms to theoretical assumptions of stock assessment models,
including negligible movement across stock boundaries, relatively homogeneous vital rates, and extensive
mixing within stock areas. Best practices for representing population structure in stock assessment involve 1)
interdisciplinary stock identication to delineate spatially discrete populations or more complex population
structure; 2) stock boundaries that are aligned with the most plausible population structure; 3) spatially-explicit
sampling, eet structure or spatial structure in assessment models to account for heterogeneity, shing patterns,
and movement within stock areas; 4) routine stock composition sampling and analysis for spatially overlapping
populations; and 5) simulation testing the performance of assessments with mis-specied or uncertain population
structure. Practical assessment units that do not accurately represent population structure may not provide
sufcient information to achieve shery management objectives, so practical constraints should be addressed
through iterative advances in routine stock identication, delineation of stocks to meet unit-stock assumptions,
and stock assessment modeling.
1. Introduction
The unit stock concept has been recognized as a theoretical
assumption of conventional stock assessment since the early stages of
sheries science (Cushing, 1968; Harden-Jones, 1968; Pauly, 1984;
Hilborn and Walters, 1992; Sinclair and Smith, 2002). Russell’s (1931)
initial derivations of sustainable yield and overshing began with the
assumption “Let us simplify the problem down to its bare essentials by
considering a completely self-contained stock of sh of one particular kind
living in a large area which is systematically shed”, and established the
axiom of the stock concept (Cushing, 1983). The population and shery
processes assumed in most stock assessment models continue to imply 1)
no movement of sh into or out of the stock area at any life stage, 2) vital
rates (somatic growth, maturity, natural mortality, and shing mortality
with selectivity) are relatively homogeneous within the stock area, and
3) individual sh mix extensively throughout the stock area (Cadrin,
2020). Although these are collectively referred to as ‘unit stock
assumptions’, the assumptions relate to shing on biological population
units rather than the practically dened ‘stock’.
A series of simulation experiments representing diverse sheries and
target species demonstrate that accurately accounting for population
structure and shing patterns in stock assessments can improve model
performance, and case studies in shery management demonstrate that
ignoring such structure can lead to misperceptions of stock status (Punt,
2019; Cadrin, 2020; Bosley et al., 2022). Therefore, an important aspect
of stock assessment is determining appropriate geographic boundaries to
dene the stock and patterns of spatial heterogeneity within the stock
area. Stock identication infers spatial population structure to delineate
stock boundaries that encompass discrete populations (e.g., Booke,
1981; Carvalho and Hauser, 1994). However, all biological populations
have some spatial heterogeneity, many have connectivity with adjacent
populations, some have geographic overlap with adjacent populations,
and subpopulations in a metapopulation are more extensively connected
at early or later life stages (Fig. 1). Population structure and major
* Corresponding author.
E-mail addresses: scadrin@umassd.edu (S.X. Cadrin), daniel.goethel@noaa.gov (D.R. Goethel), aaron.berger@noaa.gov (A. Berger), ernesto.jardim@msc.org
(E. Jardim).
Contents lists available at ScienceDirect
Fisheries Research
journal homepage: www.elsevier.com/locate/fishres
https://doi.org/10.1016/j.shres.2023.106650
Received 7 December 2022; Received in revised form 3 February 2023; Accepted 4 February 2023
Fisheries Research 262 (2023) 106650
2
shing patterns can be represented by eet structure, spatial strata, or
continuous gradients within the stock area, and stock composition
analysis can be applied to mixed-population sheries to support
assessment of each population in the mixture (Punt et al., 2020).
Modeling approaches for representing population structure and
shing patterns vary widely among stock assessments, but most opera-
tional assessments that are used to support management advice assume a
unit stock with no spatial structure within the stock area (Berger et al.,
2017). Perspectives on the importance of stock identication for stock
assessment and shery management are diverse. At one extreme, which
is unfortunately common, population structure is ignored, or violations
of the unit stock assumption are casually dismissed because of practical
constraints. At the other extreme, Link et al. (2020) list “movement,
migration and location” as the rst major mechanism impacting marine
populations, followed by overshing and other processes. The 2018
workshop on spatial stock assessment models convened by the Center for
the Advancement of Population Assessment Methodology (CAPAM)
suggested a practical order of priorities for stock assessment: “accurate
spatial modeling requires correct specication of all major features of pop-
ulation and shery dynamics (e.g., natural mortality, growth, selectivity),
because movement estimates are often confounded with estimates of
recruitment or mortality” (Cadrin et al., 2020). More recently, the 2019
CAPAM workshop on next generation stock assessment models
concluded that “a major challenge for any next-generation stock assessment
package is the set of extensions needed to assess stocks that do not satisfy the
‘well-mixed single-stock’ paradigm” (Punt et al., 2020). Punt (2023) listed
the determination of stock structure hypotheses as the rst step in good
practices for conducting assessments. Although the importance of stock
structure may vary among sheries, and perceptions of its priority may
differ among scientists, population structure and eet structure are ex-
pected to be primary features of future stock assessment modeling.
Stock delineation has been notoriously political, because stock
boundaries largely determine who has management authority and ac-
cess to shery resources. Therefore, best scientic practices for dening
spatial boundaries, spatial structure, and eet structure in stock
assessment should consider biological reality, theoretical assumptions,
and practical solutions for meeting shery management objectives. As a
contribution to the CAPAM workshop on stock assessment good prac-
tices, this review was invited to summarize relevant literature, common
practice, best practice, and research recommendations for the related
topics of stock identication, delineating stock boundaries, accounting
for spatial structure within a stock area, and simulation testing mis-
specied population structure.
2. Background
Over the last two centuries, the theory and practice of stock assess-
ment and stock identication co-evolved (Cadrin and Secor, 2009).
Initial concepts of shery production relied heavily on ‘population
thinking’ to explain uctuations in sheries as the result of variable
production of discrete sh populations (Sinclair and Smith, 2002).
Cushing (1968) rened ‘the idea of a unit stock’ to describe dynamics of
self-sustaining populations, and Harden-Jones (1968) summarized that
"management stocks are considered to respond largely independently to the
effects of exploitation, because recruitment, growth and mortality within the
stock are of more signicance than emigration or immigration to the stock.”
The importance of identifying and assessing self-sustaining stocks
increased after claims of national and international shery management
jurisdictions. When the U.N. Convention on the Law of the Sea allowed
for the establishment of exclusive economic zones (i.e., coastal nations
claiming jurisdiction for shery management) and the ‘high seas’ (i.e.,
outside of these jurisdictions), it provided for international agreements
to cooperatively manage stocks that cross jurisdictional boundaries (UN,
1982). ‘Transboundary stocks’ inhabit the exclusive zones of two or
Fig. 1. Seven general population structure scenarios, including larval connectivity (dashed arrows) or post-larval connectivity (solid arrows) among subpopulations
in a metapopulation.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
3
more coastal states, ‘straddling stocks’ extend from an exclusive zone
into the international high seas, and ‘highly migratory species’ move
among multiple exclusive zones and the high seas. The determination of
stocks that justify international shery management agreements has
been based on their biological characteristics, particularly their
geographic range and degree of mixing between exclusive zones and the
high seas, with recommendations to assess the entire range of trans-
boundary stocks (Gulland, 1980; FAO, 1994). Guidance for international
cooperation was provided by the Fish Stocks Agreement (UN, 1995),
including the requirement to assess straddling or highly migratory
stocks, implicitly as a unit stock. Even in situations where cooperative
management may not be needed (e.g., transboundary stocks with low
movement rates or low exploitation rates in each area), cooperative
stock assessment has been recognized as a good practice (Gulland, 1980;
Hilborn and Sibert, 1988; Caddy, 1997).
More recent technological advances led to the identication of
complex population structure for many species (Kerr et al., 2017). These
ndings, as well as increased spatial resolution of shery data and the
development of spatial methods for stock assessment, revived the
consideration of population structure in sheries science. The revival of
stock identication as a major feature of stock assessment is demon-
strated by its inclusion in the topics considered for improving methods
used in stock assessment and development of a good practices guide
(Punt, 2023).
2.1. Stock identication
The identication of self-sustaining shery stocks transitioned
through several stages of development. Traditional approaches included
conventional tagging, phenotypic variation, parasites as natural tags,
and spatiotemporal shing patterns (Marr, 1957). The study of poly-
morphic genetic markers revolutionized the ‘stock concept’ with an
emphasis on reproductive isolation (Booke, 1981). The next stage
involved a more holistic multi-disciplinary approach based on congru-
ence among methods (Pawson and Jennings, 1996; Coyle, 1997; Begg
and Waldman, 1999). The development of genomics, electronic telem-
etry, otolith chemistry, and otolith microstructure have supplemented
traditional approaches for more informative stock identication.
Traditional approaches to stock identication focused on results from a
single methodological approach, then multi-disciplinary evaluations
considered weight-of-evidence from several studies, and recent reviews
recognize the advantages of interdisciplinary inferences of spatial pop-
ulation structure (Cadrin et al., 2014). An inter-disciplinary approach to
stock identication considers the complementarity of methods in which
each method characterizes precise aspects of population structure. For
example, homogeneity of neutral genetic characters may not be
congruent with geographic patterns in phenotypic variation that result
from environmental differences among areas, but both are important for
stock identication, because persistent phenotypic differences can have
strong inuence on population dynamics.
Three interacting aspects of population structure are evaluated in
interdisciplinary stock identication: distribution, dispersal, and
geographic variation - but no single source of information can support
inferences of all three. Information on geographic distribution (e.g.,
shery monitoring, shery-independent surveys) can dene a species
range, spatial continuity, areas of high abundance, spawning areas,
nursery areas, and shing grounds for each eet. Connectivity is eval-
uated from information on dispersal of early life history stages (e.g.,
plankton surveys, bio-physical models) and movement of juveniles,
adolescents and adults from conventional tags, electronic telemetry or
‘natural tags’ (e.g., parasites, otolith chemistry). Geographic variation in
phenotypic characters, neutral genetic characters or those subject to
selection can inform patterns of population heterogeneity, lack of mix-
ing, reproductive isolation, or local adaptation. Spatial patterns in
neutral genetic characters indicate reproductive isolation among areas,
and differences in selected characters or phenotypic characters reect
both genetic and environmental differences. Some phenotypic traits (e.
g., growth rate, age at maturity) affect population dynamics more
directly than others (e.g., morphology), but any persistent geographic
variation in genetic or phenotypic characters should be considered in
stock identication because it indicates limited mixing among areas.
These complementary sources of information can be geographically
integrated to form inferences of plausible population structure,
including the identication of discrete populations for delineating stock
boundaries, identication of subpopulations to account for spatial
structure within stock areas, and identication of ner scale structure
that may be relevant to productivity, shery management and conser-
vation (e.g., behavioral contingents, spawning aggregations; Kerr et al.,
2010b).
2.2. Stock boundaries
Most stock assessments assume that the stock is a single discrete
population (Hilborn and Walters, 1992), but most assessment units were
delineated to encompass major shing grounds or jurisdiction, which
may not align with the spatial extent of biological populations that are
being shed. Therefore, stock identication is needed to conrm that
the entire biological population is within the stock area, there are not
multiple populations or sub-populations of the species in the stock area,
and there is negligible dispersal across boundaries or connectivity with
adjacent areas (Cadrin et al., 2014).
The amount of movement across a stock boundary that violates the
unit stock assumption cannot be generalized (Aldenberg, 1975), because
sensitivity to movement rate depends on the movement pattern, relative
population sizes, degree of reproductive mixing, and conservation sta-
tus. More precise denitions of ‘negligible movement’ across stock
boundaries requires stock-specic simulation to determine if observed
cross-boundary movements impact the performance of stock assessment
and shery management in the context of other vital rates and relative
abundance. For example, the larger eastern population of Atlantic
bluen tuna is less sensitive to movement and stock mixing than the
smaller western population (Morse et al., 2020). Caddy (1997) sug-
gested a threshold for cross-boundary dispersal as enough to produce a
+/−10% error in perception of local biomass, but in some cases the
threshold may be smaller.
Movement patterns that involve post-spawning dispersal and natal
homing can produce a seasonal mixture of populations on feeding
grounds and in sheries (i.e., ‘overlap’, Porch et al., 2001). This move-
ment pattern adds some seasonal immigrants to catches or excludes
some emigrants that are caught outside the stock area, thereby adding
noise to time series signals of year class strength, mortality, or selec-
tivity. For example, assessing mixed New Zealand snapper populations
as single unit stock could not account for spatial patterns in age
composition and growth, resulting in catch advice that was unsustain-
able for a depleted population in the mixture of populations being
harvested (Francis and McKenzie, 2015; Berger et al., 2017). Movement
patterns that result in reproductive mixing across a stock boundary (i.e.,
‘random movements or dispersions’ Gulland, 1980; ‘diffusion’, Porch
et al., 2001) also add noise to year-class signals and the stock-recruit
relationship.
A common feature of marine shes and other animals is the rare
occurrence of much further-than-average movements (Secor, 2015),
which may draw attention to the extreme movements despite a rela-
tively low average movement rate. A few diffusive movements among
subpopulations over a generation can produce enough reproductive
mixing to homogenize genetic composition but may not disrupt de-
mographic independence for stock assessment (e.g., Haugen et al.,
2022). By contrast, movement of a few individuals may be more im-
pactful for identication of spatial units for threatened species (Eagle
et al., 2008).
Some populations can be effectively delineated by geography
because they are strongly associated with benthic habitat or stable
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
4
oceanographic features and have discontinuous distributions among
populations. However, other populations are more spatially dynamic,
have less discrete boundaries with adjacent populations, or may have
some connectivity with other populations. Another challenge for rep-
resenting spatial population structure is the resolution and accuracy of
spatially explicit data used for stock assessment. Despite these chal-
lenges, stock boundaries should be based on stock identication and
approximately encompass populations to meet conventional stock
assessment model assumptions.
2.3. Spatial structure within stock areas
Population structure is a continuum in which the isolating mecha-
nisms that form separate populations within a species, and potentially
separate species over geological time, also form subpopulations within a
population in ecological time. Spatial structure within populations can
be discrete, with well-dened boundaries, or more continuous, resulting
in clinal variation and isolation by distance (Fig. 1). Although stock
boundaries should be based on biological population structure, spatial
patterns within a stock area can also be inuenced by shing. If mixing
is limited within a population, geographic shing patterns can create
differences in survival among areas (e.g., marine protected areas impose
strong spatial heterogeneity in shing mortality and potentially other
vital rates). Heterogeneity within stock areas violates dynamic pool
assumptions and complicates parameter estimation or management
reference points (Goethel and Berger, 2017). For example, heteroge-
neous habitats and spawning behaviors can inuence stock-recruit re-
lationships (Skoglund et al., 2022).
Geographic patterns within a stock can be represented by spatially
explicit samples, eet structure, or spatially structured assessment
models, but the impact of spatial patterns on stock assessment modeling
depends on the source of heterogeneity (e.g., shing patterns vs. bio-
logical patterns) and the movement rate among areas. Spatially explicit
sampling (e.g., stratied random or systematic designs) can characterize
sheries or stocks that have spatial patterns so that the contribution of
each area is appropriately weighted for a representative sample. How-
ever, population estimates based on weighted averages may not accu-
rately reect the combined results of local dynamics (Hart, 2001).
If spatial heterogeneity primarily results from shing patterns and
limited movement, eet structured sampling and modeling can help
account for spatial structure. Many integrated stock assessment models
assume constant selectivity, often within multi-year periods (Methot,
2023, this issue). Constant selectivity assumptions within periods may
be valid for eets that use the same shing gear but may not be valid for
sheries that use multiple shing gears. Therefore, eet stratication is
commonly applied to monitor landings, discards, size- or
age-composition, catch rate indices, and economic information to sup-
port eet-based management (e.g., Lennert-Cody et al., 2010, 2013;
Ulrich et al., 2012; Frawley et al., 2022). Fleet structure within an in-
tegrated stock assessment can greatly improve estimates of selectivity
(Punt et al., 2014). Fleets are largely identied by shing location, so
eet structure in a stock assessment can account for spatiotemporal
shing patterns as well as some spatial heterogeneity in the sh popu-
lation (Berger et al., 2012; Waterhouse et al., 2014; Hurtado-Ferro et al.,
2015). However, substantial movement or heterogeneity may require
spatially structured assessment models (Goethel et al., 2023, this issue).
If spatial structure involves reproductively isolated populations or
partially isolated subpopulations, information on stock composition is
needed to estimate their relative contributions to the mixed-population
shery and stock over time (Utter and Ryman, 1993). Unfortunately, the
term ‘sub-stock’ has various denitions in the sheries literature, with
different implications for stock assessment models. For example, Punt
(2019) used the term to describe stock components that have extensive
reproductive mixing with other stock components, but others use the
term to describe reproductively isolated populations within a stock area
(e.g., Frank and Brickman, 2000; Sterner, 2007; Lindegren et al., 2013).
In this paper, we attempt to distinguish terminology for biological units
(species, populations, metapopulations, subpopulations, behavioral
contingents, spawning aggregations) from practical units (jurisdiction,
stock, shing ground, statistical reporting area, sampling strata, model
strata) to avoid the common implication that practical units accurately
represent biological reality.
2.4. Simulation testing
Simulation analyses were initially applied to spatially complex
sheries to understand the implications of population structure (e.g.,
Beverton and Holt, 1957; Ricker, 1958; MacCall, 1990; Kerr et al.,
2010a). These heuristic simulations provided the framework for
simulation-estimation studies that evaluate the performance of rela-
tively simple estimation models for accurately representing complex
populations and shing patterns (e.g., Aldenberg, 1975; Porch et al.,
1998; Berger et al., 2017). Simulation has also emerged as an integrative
tool for interdisciplinary stock identication by conditioning operating
models on information from several methodological approaches (e.g.,
Kerr and Goethel, 2014).
Operating models for simulation testing were initially conditioned
on generic ‘sh-like’ population parameters with simple biological
structure (e.g., two subpopulations of equal size) with uniform move-
ment rates (e.g., Aldenberg, 1975). These relatively simple simulations
were designed to make general inferences about the effect of population
structure on stock assessment. More recently, simulation-estimation
studies and management strategy evaluation have been more precisely
conditioned to represent specic sheries (e.g., Deroba et al., 2015; Punt
et al., 2016). Simulation of spatially complex populations can evaluate
the robustness of stock assessments in the context of uncertain popula-
tion structure and demonstrate the risks of violating unit stock as-
sumptions for specic sheries (e.g., Punt, 2019; Berger et al., 2021;
Bosley et al., 2022).
3. Common practice
3.1. Stock identication
Stock assessment reports describe stock boundaries, and most pro-
vide a justication for the stock area delineation and spatial strata, often
with a summary of the available information on spatial population
structure. The scientic process for providing this information varies
among regional shery management organizations. When population
structure is identied as a major source of uncertainty in a stock
assessment, or the assessment model exhibits diagnostic problems sug-
gesting mis-specied stock structure, some organizations host a work-
shop to review the available information on stock identity and form
recommendations for spatial assessment units (e.g., ICCAT, 2001; ICES,
2009, 2020, 2022b; Quinn et al., 2011; WPFMC, 2014; Moore et al.,
2020a). Some organizations have standing expert groups to review and
update information on population structure of specic stocks and to
recommend revised stock denitions to assessment working groups (e.
g., ICES, 2022a). Other organizations recently added routine stock
identication workshops into their stock assessment process in advance
of data and model workshops to dene the most plausible stock
boundaries based on the available information to meet the needs of the
shery management system (e.g., SEDAR, 2018, 2020, 2021; Fig. 2).
Finally, some stock assessments have specic terms of reference to
investigate spatial population structure (e.g., NEFSC, 2012, 2020; Punt
et al., 2019). These standing committees, workshops or terms of refer-
ence within a stock assessment require broader expertise than typical
stock assessment working groups or peer reviews.
Many stock assessment documents report that little information is
available on stock identity. However, basic information on the species
(e.g., geographic range, distribution patterns, early life history; www.
shbase.org), the shery (e.g., shing grounds, seasonality), and data
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
5
collected for stock assessment (e.g., spatial patterns in size or age
composition, size-at-age, maturity-at-age) can be used to investigate
putative stock structure (e.g., Begg and Waldman, 1999; Lennert-Cody
et al., 2010, 2013). Therefore, every routine stock assessment offers an
opportunity to review the information available and recommend future
research to support iterative improvements.
3.2. Stock boundaries
The most common sequence of events dening stock boundaries is an
initial claim of management jurisdiction for sheries in an area, the
collection of data for sheries in the jurisdiction, and the development of
stock assessments for target species in the area or discrete shing
grounds within the area. As a result, spatial boundaries of many stock
assessments do not adequately represent population structure, because
there was no consideration of population structure in the initial stock
denition. Stock identication can either conrm that the management
unit is a population or that stock boundaries do not conform to unit stock
assumptions. Dening or re-dening the spatial extent of stock assess-
ment is needed in advance of data compilation or model development.
Unfortunately, limited spatial resolution of historical shery data often
constrains the denition of stock boundaries and strata (e.g., ICES, 2020,
2022b). Stock denition is often a practical compromise between sci-
entists and shery managers so that the spatial unit of assessment is
consistent with the management unit.
Some stock boundaries have been revised based on new perspectives
of population structure, but many more stock boundaries are maintained
despite their known misspecication (e.g., Reiss et al., 2009; Kerr et al.,
2017; Ommer and Perry, 2022). Common practice is to assess the
management unit and ignore model assumptions that are violated when
spatial data are not available to support the assessment of a discrete
population, management jurisdiction precludes the management of a
unit stock, or population structure is too complex to meet unit-stock
assumptions. As a result, many assessment units include portions of a
larger population, multiple discrete populations, subpopulations of the
same species, or even multiple species (Cadrin, 2020). Considering the
large number of potentially mis-specied stock boundaries, poor per-
formance of spatially mis-specied assessment models, and shery
failures resulting from mis-specied population structure, it is reason-
able to conclude that many stock assessments can be improved by
revising the stock boundary to meet unit stock assumptions.
Some stock assessments account for movement of sh across stock
boundaries by adjusting the value of assumed natural mortality, M
(ICES, 2021), or as process error in total survival or natural mortality
(Frisk et al., 2008; Aldrin et al., 2019; Nielsen and Berg, 2023, this
issue). Emigration may have similar effects as mortality on changes in
abundance over time, and immigration can have a similar result as a
lower mortality rate, but aliasing movement as mortality ignores the
form of movement (e.g., sh that leave the stock area but return to
spawn). Although an assumption of increased natural mortality may
account for emigration in survival predictions, the two processes have
different effects on management reference points and catch projections
(Goethel and Berger, 2017). The history of stock assessment has trended
toward more accurate model specication, but the ‘M-agic’ of aliasing
movement as mortality deliberately mis-species one process (natural
mortality) to account for misspecication of another process (move-
ment) or represents a nonrandom process (movement) with random
process error. Therefore, given the management implications of these
implicit assumptions regarding connectivity, it is not recommended as
an appropriate tool for addressing spatial processes.
3.3. Spatial structure within stock areas
The majority of stock assessments assume no spatial structure within
the stock area. For example, a survey of U.S. stock assessment scientists
indicated that most (83%) assessments assumed no spatial structure, but
there was evidence of spatial structure for most of those stocks (Berger
et al., 2017). In common with other structural features of assessment
models (e.g., age, sex, eet), spatial structure requires spatially explicit
data. Fortunately, data from shery monitoring or shery-independent
surveys are spatially explicit or stratied to represent the stock or sh-
ery through stratied estimates or spatiotemporal analysis (e.g., Currie
et al., 2019). For multispecies sheries and surveys, sample stratication
often considers ecosystem boundaries (e.g., geographic, bathymetric or
oceanographic features) that may be putative spatial structure for
multiple stocks. Punt (2023) explained that spatial strata are often
dened by jurisdiction, and jurisdictional strata can account for national
eets or management regimes.
Most sheries are sampled by eet, in which vessel-based eets or
trip-based m´
etiers are dened primarily by shing gear, target species
and shing locations (e.g., Ulrich et al., 2012; Lennert-Cody et al., 2013;
Frawley et al., 2022). Fleet structure in stock assessment models helps to
estimate selectivity and can account for some spatial heterogeneity,
often imposing time-varying selectivity on the oldest or largest sh
(Methot, 2023, this issue). Many assessments model the shery as a
single eet, assuming constant selectivity within periods of similar eet
composition, regulations, or shing behavior. Information on shing
mortality for each eet can be derived from their partial catch-at-age or
-at-length in aggregate-eet assessments (e.g., Porch et al., 2001), and
state-space models can estimate annual process errors in aggregate-eet
selectivity to allow for varying contributions of eets to total catch (e.g.,
Nielsen and Berg, 2014). However, these single-eet approaches cannot
Fig. 2. Iterative stock assessment process with sequential stock identication, data preparation, modeling, and peer review workshops, producing management
advice and research recommendations that can be addressed in a subsequent iteration. The stock identication workshop reviews available information to infer
plausible scenarios of population structure (including the most plausible scenario if possible) and potentially recommend revised stock denition for assessment and
shery management. The data preparation workshop should support the recommended stock denitions so the assessment workshop can develop models for the
recommended stock, sensitivity to plausible alternatives, and potentially performance testing. Peer review can be integrated into each workshop, and additional
science-management interactions early in the process are needed if revised stock denitions are recommended.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
6
account for the spatial heterogeneity imposed by different shing pat-
terns among eets.
Spatial structure resulting from a mixture of multiple sympatric
populations within a stock area can be accounted for with stock
composition analyses. Routine stock composition sampling and analyses
have been successfully applied to assessment of many mixed-stock
salmonid sheries for decades (Utter and Ryman, 1993), and genetic
stock identication for Pacic salmon has been rened over time to be
cost-effective (e.g., Beacham et al., 2020). Stock composition sampling
and analysis has also been applied to some non-salmonid sheries (e.g.,
Kerr et al., 2022). Many other sheries catch a mixture of intraspecic
populations, but the mixture is assessed as a single stock, with no
monitoring of stock components, risks of depleting components, or po-
tential failure to achieve optimal yield (Ricker, 1958).
Spatially structured stock assessment models have been developed,
but few are applied as the basis for shery management because of
model complexity, data requirements, difculty estimating movement
rates, policy implications, or institutional inertia (Berger et al., 2017;
Punt, 2019). Similar to the constraints of data resolution on stock de-
nition, strata denition is also commonly constrained by the resolution
of shery data (e.g., Cope and Punt, 2011; Gertseva and Cope, 2017,
Thorson and Wetzel, 2016). Although many shery systems now collect
spatially explicit data with high-resolution, historical data typically has
lower resolution (Goethel et al., In press).
Some management strategies require information on spatial struc-
ture. For example, some spatially aggregated assessments include
spatially disaggregated forecasts to support spatial catch allocation
(Kapur et al., 2021). Bosley et al. (2019) evaluated performance of
spatial forecasts and found that those based on local stock indices per-
formed best for achieving nearly maximum system yield, but all ap-
proaches frequently led to local depletion when spatial structure was
ignored or specied incorrectly. Spatial stock assessment information
can also help to evaluate effectiveness of other spatial management
actions (e.g., marine protected areas, spawning closures).
3.4. Simulation testing
Despite several recommendations for routine simulation testing to
evaluate the consequences of model misspecication (Hilborn and
Walters, 1992; Deroba et al., 2015), few stock assessments include
simulation to evaluate performance of the estimation model, and almost
all of those assume a unit stock in the operating model. Most
simulation-estimation testing occurs in research projects that are
somewhat independent from the stock assessment process, and results
are not always considered in the assessment. Similar to other model
misspecications that receive more attention, stock assessments with
mis-specied stock boundaries or spatial structure can be misleading.
Therefore, simulation testing is needed in which operating models
represent plausible population structure (Cadrin, 2020).
4. Suggested good practices
4.1. Stock identication
Regardless of assessment or management constraints, plausible
population structure should be inferred from an interdisciplinary syn-
thesis of all information available, including perspectives from the
shery. If possible, the most plausible scenario of population structure
should be identied to simplify data compilation for a single stock
assessment model. However, even the most data-rich sheries will have
some uncertainty for inferring population structure, which can be
expressed as multiple plausible scenarios of population structure. For
well-informed inferences, plausible scenarios may have a common
archetype of population structure (Fig. 1), with alternative scenarios of
boundary locations or movement rates. More uncertain inferences of
stock identity may be represented as alternative plausible archetypes.
Plausible scenarios can be depicted as conceptual models (e.g., ICES,
2009, 2020; Quinn et al., 2011; Zemeckis et al., 2014; Minte-Vera et al.,
2023, this issue).
Revising stock assessments to more accurately reect population
structure can improve model performance, but reviewing information
on stock identication should be routine and not limited to problematic
assessments. A summary of stock identication and how well it matches
the current assessment or management unit should be updated in every
stock assessment report (e.g., an updated summary of information
should be a generic term of reference for all assessments). These sum-
maries can be supported by stock identication workshops, ideally
within the stock assessment process and before data compilation so that
data can be subsequently compiled to support the recommended spatial
boundaries and strata (Fig. 2). An inclusive process for comprehensively
reviewing, integrating, and updating the information available on
population structure was developed by ICES (2009) for deep-sea redsh
and subsequently applied to some cod sheries (e.g., ICES, 2020,
2022b), offering a methodology for wider application of stock identi-
cation workshops. Coordinating processes for stock identication and
stock assessment (e.g., Fig. 2) facilitates the transition from conceptual
models of population structure (e.g., data maps, workows) to obser-
vation models in integrated stock assessments (Minte-Vera et al., 2023,
this issue). Coordinated processes also promote the incorporation of
stock identication information as data in spatially structured assess-
ments (e.g., tagging, population-of-origin indicators).
New information from advanced methods (e.g., genomics, electronic
tagging, otolith chemistry) should be reconciled with previous infor-
mation from traditional approaches (shery perspectives, tagging, par-
asites), recognizing their relative sensitivities, and which aspect of
population structure each source of information represents. Genomics is
emerging as a powerful and cost-effective tool (Rodriguez-Rodriguez
et al., 2022), and broad genome coverage is considered best practice for
detecting reproductively isolated or locally adapted populations
(Valenzuela-Quinonez, 2016). Genetic variation can also be used to es-
timate effective dispersal in some situations (e.g., Broquet and Petit,
2009). However, persistent patterns of phenotypic variation that reect
environmental differences also inuence population dynamics. There-
fore, complementary information on life history, distribution, dispersal,
and phenotypic variation adds interpretive value to genomics for
determining the most plausible population structure.
Similar to other population processes, stationarity is often assumed
for stock delineation and eet denition, but the persistence of spatial
patterns should be regularly tested to detect possible changes (e.g.,
shifting geographic distributions in response to climate change). Fishery
and survey data should be routinely mapped to explore patterns and
detect changes in distribution as well as the persistence in patterns of
size and age composition, size at age and maturity at age (Hilborn and
Walters, 1992). Analyses of spatial shifts require sampling throughout
the population’s range, including its boundaries (Karp et al., 2019).
Spatial shifts complicate the determination of stock boundaries, eets,
and strata, but better representation of spatial structure can help to
identify the mechanism of shift and how to appropriately account for the
shift (Currie et al., 2019). If individual sh are moving from unsuitable
to suitable habitats or the population’s range is expanding as a result of
changes in effective dispersal (Hare and Able, 2007), then revised spatial
strata may be needed, or stock boundaries may need to be extended. If
the shift involves a demographic pattern associated with depletion or
rebuilding (Bell et al., 2015), then the stock area should represent the
distribution of a rebuilt population. If the apparent shift in distribution
results from divergent trends of independent allopatric populations
(Link et al., 2010), separate stock areas should be dened or maintained
for each population.
Recommended best practice for stock identication is to include a
description of the spatial assessment unit and a summary of available
information on stock identity in every stock assessment report. Each
assessment should also examine evidence for spatial shifts. Finally, stock
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
7
assessments should provide research recommendations to ll gaps or
investigate emerging patterns. Moore et al. (2020b) offer an excellent
example of research recommendations developed through an interdis-
ciplinary review of stock identication. Implementing an iterative pro-
cess of updating information on stock identity, recognizing research
needs, and investing in research (Fig. 2) can help advance all stock as-
sessments, including those for data-limited sheries, toward an appro-
priate geographic scope and structure for meeting the management
needs of each shery.
4.2. Stock boundaries
Stock boundaries should be delineated to represent a biological
population as closely as possible with the data available. Isolated pop-
ulations can be assessed as unit stocks and stock status can appropriately
be based on stock-recruit reference points when all life stages of a
population are contained within the stock area, including those that
demonstrate ontogenetic movement patterns. The entirety of sympatric
populations should be accounted for in stock assessments by assigning
all data (e.g., shery catch, index catch, size or age composition) to each
population in the mixed stock. If population structure is more complex,
stock boundaries should encompass a complete metapopulation while
monitoring subpopulation trends. For example, each population or
metapopulation depicted in Fig. 1 should be assessed as a stock unit, and
subpopulations within metapopulations should be accounted for with
spatial structure within the stock area. There may be tradeoffs between
1) mis-specied stock boundaries that allow a longer time series of
historical information with coarse spatial resolution and 2) correctly
specied stock boundaries with a more restricted time series of recent
spatially explicit data.
If changes to stock denitions or boundaries are needed to improve
the representation of population structure, shery managers should be
consulted early in the stock assessment process so they can communicate
any additional practical constraints and consider adapting to new
assessment units (e.g., support management actions with information
from new assessment units, revise management units to match assess-
ment units). Ideally, management units should also represent biological
populations, because management models (e.g., optimum yield, overf-
ishing, rebuilding plans) imply the same assumptions as stock assess-
ment models. Accordingly, a common standard for managing U.S.
marine sheries is to manage a stock as a unit (e.g., USA, 2007).
Therefore, stock denition should be based on the best scientic infor-
mation available to meet management objectives while considering
practical constraints, similar to other stock assessment assumptions that
have management implications (e.g., stock-recruitment, natural mor-
tality, selectivity, data weighting).
If practical challenges preclude the assessment of entire meta-
populations with subpopulation structure, subpopulations that have
some larval connectivity with other subpopulations, but negligible post-
settlement connectivity, can be effectively assessed with separate as-
sessments and stock status can be appropriately based on per-recruit
reference points. However, reproductive capacity of the entire meta-
population requires conservation. For example, source-sink dynamics
and increased vulnerability of source subpopulations should be consid-
ered in stock assessment and shery management.
If spatial management units include multiple discrete populations,
each population should be separately assessed. Consistent productivity
assumptions and model settings can help to provide comparable abun-
dance and mortality estimates for management of the combined unit
(PFMC, 2021). Separate stock assessments with consistent assumptions
provide estimates of stock size, mortality and projected catch that are
more comparable and potentially additive for aggregate catch advice (e.
g., Jardim et al., 2018). Such consistency in assessment methods can
help to avoid misleading inferences of spatial distribution and associated
management conicts. For example, Georges Bank cod and haddock
assessments demonstrate how lack of consistency between assessments
can create challenges for shery management. U.S.-Canada trans-
boundary management units are nested within larger U.S. management
units (Pudden and VanderZwaag, 2007), but separate assessments with
different methods and assumptions (e.g., natural mortality) produce
estimates of stock size that are not comparable and catch advice that is
not even approximately additive. As a result of these inconsistent
methods and assumptions, subtracting total allowable catch for the
smaller area from allowable catch for the larger area currently leaves
little cod catch for U.S. sheries (NEFMC, 2022).
4.3. Spatial structure within stock areas
If population structure is too complex to dene distinct spatial
stocks, stock assessment may require spatial stratication, spatiotem-
poral analysis, or stock composition analysis to account for heteroge-
neity. For example, assessment of metapopulations needs to account for
each subpopulation (Fig. 1). Populations with discrete spatial structure
require stratication of samples or models, depending on the degree of
connectivity. Complex populations with isolation by distance within the
stock area require spatially explicit data and assessment models.
Routine stock composition sampling is needed to account for sym-
patric populations. Stock composition (i.e., population-of-origin for in-
dividuals in a mixed stock) can be representatively sampled with other
compositional samples (e.g., size, age, sex), and archived samples can be
used to derive historical stock composition (e.g., genetics, otolith
chemistry and microstructure; e.g., Smith and Campana, 2010). Stock
composition sampling and analyses have been successfully applied to
some sheries, but they are typically applied to data preparation (e.g.,
catch by population).
Stratied sampling and eet structure can account for some shing
patterns and heterogeneous vital rates (e.g., Berger et al., 2012;
Waterhouse et al., 2014; Hurtado-Ferro et al., 2015). Fleet denition
involves the recognition of heterogeneous shing patterns, so that
shing behavior and selectivity are relatively homogenous within eets.
Good practice for eet denition involves hierarchical classication
using information on shing effort (e.g., location, season, shing gear,
mesh size, horsepower), a priori target species (e.g., sher interviews),
or catch (e.g., species composition, size composition; e.g., Marchal,
2008; Lennert-Cody et al., 2010, 2013; Ulrich et al., 2012; Frawley et al.,
2022).
Spatially structured population models are generally preferred over
the eets-as areas approach (Methot, 2023, this issue) and spatial model
structure may be needed for stronger patterns of heterogeneity (Berger
et al., 2017; Punt, 2019). Sub-annual time intervals are needed in
spatially structured models to represent seasonal movement patterns
and the sequence of movement with other events (e.g., spawning, sh-
ing, surveys; Bentley et al., 2004; Taylor et al., 2011). Low spatial res-
olution of some data may not be an obstacle for spatial structure in
assessment models, because integrated models can t directly to
spatially aggregated data (e.g., historical shery data) as well as
spatially disaggregated data (e.g., recent shery data and survey data) to
represent spatial structure and include all available information (Fig. 3).
Many sex-structured models demonstrated how integrated models can
t to aggregated and disaggregated data (Maunder and Punt, 2013;
Wilberg et al., 2023, this issue). If data are not sufcient to support a
complex estimation model, sensitivity analyses can help to evaluate the
consequences of simplication. For example, Thorson and Wetzel
(2016) found that their two-area model assuming no post-settlement
movement produced similar results as sensitivity runs that assumed a
single area or a range of assumed movement rates among areas.
Spatial structure within a stock area can also be represented by
spatiotemporal analyses (e.g., Cao et al., 2020). Spatiotemporal models
are particularly well suited to population structures and geographic
pattens that are less discrete (e.g., isolation by distance, geographic
clines) because they account for spatial correlation. Spatiotemporal
models are promising but may need further development for application
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
8
to complex population structures (Goethel et al., 2023, this issue).
4.4. Simulation testing
If stock identication is routinely reviewed in the stock assessment
process, some assessments will be recognized as mis-specied for rep-
resenting the most plausible population structure. ‘Cross-test’ simula-
tions should be conducted for any suspected misspecication (Deroba
et al., 2015; Punt et al., 2020), so assessments that cannot conform to
unit stock assumptions are simulation-tested to evaluate performance (e.
g., Goethel et al., 2016). Simulation-estimation testing would be most
appropriately done within the stock assessment process so that results
can be considered in determining best practice for each assessment (e.g.,
Jacobsen et al., 2022). However, if simulation testing is beyond the
scope of operational assessments, it should be developed as a research
project in coordination with the assessment process.
If the information available supports multiple plausible stock struc-
tures, they can be represented as multiple operating models, and
simulation testing can evaluate the robustness of estimation models to
the range of plausible scenarios (e.g., Porch et al., 1998, Jardim et al.,
2018, Punt et al., 2018). If assessment of the current stock area does not
perform well, alternatives (e.g., redened stock boundaries, spatial
structure) should be tested that use currently available data. Alternative
assessment methods that require new data investments may also be
required, and the cost-benet of data collection can be quantied within
the simulation framework.
Ideally, results from spatially complex estimation models that t the
available data can be used to condition operating models (i.e., condi-
tioning on data; e.g., Goethel et al., 2015). When the available data
cannot support such complex estimation models, spatial operating
models can be conditioned on results from exploratory estimation
models or a combination of estimated parameters and expert judgment
(i.e., conditioning on models). For example, alternative approaches to
simulation testing Atlantic bluen tuna assessments and management
procedures used spatial models with a range of relative population
abundance from separate-area assessments (Carruthers and Butter-
worth, 2018) or results from separate-area assessments combined with
connectivity information from shery-independent telemetry (Morse
et al., 2020). These challenges in conditioning spatial operating models
demonstrate the need to continue investments and advances in spatially
structured estimation models (Goethel et al., 2023, this issue), so that
operating models can be conditioned on data rather than on disparate
model results that have inconsistent assumptions about population
structure.
If spatially mis-specied assessment models do not perform well for
providing accurate stock status and more appropriate specication is not
possible within jurisdictional or data constraints, management strategy
evaluation is needed to conrm that the current management strategy
can meet management objectives (Punt et al., 2016, 2017). Precau-
tionary harvest control rules account for some uncertainty, but assess-
ments that produce substantial bias in parameter estimates and their
variances may not perform well for providing the information needed by
the control rule. If the current management strategy does not perform
well, additional management features can be considered for testing
alternative options (e.g., marine protected areas, spawning closures,
escapement thresholds for spawning groups, spatial catch allocation;
Bosley et al., 2019). Spatial operating models that represent multiple
populations, mixing and eet structure can be used to test empirical
management procedures (e.g., Carruthers and Butterworth, 2018) or
relatively simple model-based procedures (e.g., Morse et al., 2019;
Weston et al., 2019). Management strategy evaluation may be needed
before revising management units to justify the costs of transition.
5. Required research
Investments are needed to regularly update stock identication in-
formation (e.g., Fig. 2), to ll critical information gaps, and to address
uncertainties. Although inter-disciplinary stock identication remains
best practice, genomics is emerging as a cost-effective approach that can
be applied to many more species. As applications of close-kin mark
recapture increase, genetic data can provide information for stock
identication (e.g., Trenkel et al., 2022; Bravington, 2023, this issue).
These investments involve commitments to 1) consider the new infor-
mation in the context of other available information, 2) potentially
revise perceptions of population structure, 3) evaluate the consequences
of any mismatches between the current assessment unit with the new
perception of population structure, and 4) revise stock boundaries,
spatial structure, or management procedures if needed.
Further development of spatial assessment models is needed so that
estimation models can better represent complex populations and sh-
eries (Goethel et al., 2023, this issue). If spatial models are too complex
to be supported by the information available for a specic shery,
exploratory spatial models may help to condition operating models for
simulation testing. For example, if the data cannot inform the estimation
of some model parameters, multiple operating models can be condi-
tioned on a plausible range of assumed values (e.g., Carruthers and
Butterworth, 2018). Considering the recommendations for routine
simulation testing of stock assessments (Hilborn and Walters, 1992;
Deroba et al., 2015), the next generation of stock assessment models
should support efcient simulation testing of spatially structured or
multi-population models (Punt et al., 2020).
The geographic integration of stock identication information from
multiple disciplines remains somewhat qualitative (e.g., the conceptual
models described by Minte-Vera et al., 2023, this issue). More quanti-
tative integration and appropriate consideration of uncertainty would
require the development of spatially explicit population genetics models
to evaluate differences within and between populations. Conventional
population genetics models (e.g., Rousset, 2007) would need to be
extended to t data on genetic variability, phenotypic variability,
movement, and effective dispersal at ecological time scales. More ac-
curate information on heritability of phenotypic traits and rates of
early-life history dispersal and post-larval movement would be needed
to support such integrated population genetics. Population genetics
models may develop in parallel with the next generation of spatial stock
assessment models.
6. Discussion
There are two main challenges for dening stock boundaries and
strata within stock areas: 1) delineation of a stock that represents a
discrete population, and 2) representing more complex population
structure. Failing to address either challenge (i.e., violating unit stock
Fig. 3. Data inventory for an integrated stock assessment with spatial structure
(two areas: north and south) t to spatially aggregated historical data and
spatially disaggregated recent data, showing that poor spatial resolution of
historical shery data may not be an impediment to spatially structured
stock assessment.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
9
assumptions or mis-specifying population structure) may corrupt stock
assessments and mislead shery management. Therefore, unit stock
assumptions cannot be dismissed without simulation testing to conrm
that mis-specied assessments perform well enough to meet objectives.
When stock identication suggests distinct populations, it may be
cost-effective to revise stock boundaries, so they encompass each pop-
ulation. Inter-jurisdictional assessments may be needed to resolve
boundary constraints (Gulland, 1980; Hilborn and Sibert, 1988; Caddy,
1997; FAO, 1994; UN, 1995). Data limitations can be confronted by
improved monitoring systems that provide the required spatial data, and
recovery of spatially explicit data from archives. The costs of these in-
vestments may be considerably less than the costs of misleading stock
assessments for shery management.
The scientic challenge is greater when stock identication suggests
more complicated population structure. There have been advances to-
ward accurately representing complex population structure in stock
assessment for some data-rich sheries, but many sheries do not have
the information to support such complex estimation models. This situ-
ation appears to present a ‘Catch-22’ conundrum, because spatial data
are insufcient to correctly specify the estimation model, but spatial
information is needed to condition operating models for simulation
testing the performance of simpler estimation models. There are two
alternative solutions to this challenge. Operating models can be loosely
conditioned on the information developed by iterative stock identica-
tion to represent multiple plausible scenarios. Alternatively, operating
models can be more precisely conditioned on results from exploratory
spatial estimation models that are t to the available data. Exploratory
spatial models may not be reliable enough for precise status determi-
nation or catch advice, but the range of results may adequately represent
the system for simulation testing. Advances in spatial estimation models
should help to resolve this conundrum (Goethel et al., 2023, this issue).
Assessment of data-limited sheries usually involves some form of
model simplication, including bold assumptions that may not be valid
(e.g., Cope, 2023, this issue). The data and model requirements for
spatial assessment or testing are a particular challenge for data-limited
sheries, but model assumptions and consequences for violating them
apply to all stocks. For example, in his manual of methods for stock
assessment of tropical sheries, Pauly (1984) began by explaining
Russell’s (1931) axiom of a unit stock and its assumptions. The iterative
approach of routine stock identication, delineation of stocks to meet
unit-stock assumptions, operational assessment, and research to ll in-
formation gaps (Fig. 2) can be applied to data-limited sheries. The data
collected for assessment can be explored for information on stock
identity (Begg and Waldman, 1999), and the population richness of a
species (Sinclair, 1988) can be considered for forming putative scenarios
of population structure. The trend toward simulation testing
data-limited assessments using the information available (Carruthers
et al., 2014) can be expanded to spatial simulations with multiple
operating models to represent plausible scenarios of population
structure.
State-space models might offer a solution to account for relatively
low rates of movement across stock boundaries as process error in sur-
vival (Frisk et al., 2008; Aldrin et al., 2019; Nielsen and Berg, 2023, this
issue). However, the degree of structural misspecication that can be
accounted for as stochastic process error needs to be determined. Similar
to other potential approaches, the performance of state-space models
that include process error in survival to account for immigration or
emigration should be simulation tested.
The common terminology of ‘stock’ and ‘unit stock assumption’ may
contribute to the common misspecication of population structure in
stock assessments. It may seem like an obvious tautology, but simply
calling a management unit or assessment unit a ‘stock’ does not imply
that it meets unit stock assumptions. Viewing biological population
structure through the human constructs of jurisdictions, shing grounds,
reporting areas, or geographic sampling strata often produces a distorted
perspective, like the people in Plato’s cave inferring reality from
shadows on the wall. Furthermore, through the iterative process of stock
assessment and shery management, the management unit and stock
appear to become biological realities themselves that conform to unit
stock assumptions, like Pygmalion eventually believing his sculpture is a
real person (Schnute and Richards, 2001). These human tendencies can
be countered by routine stock identication to remind us that assess-
ment models are simplications and may not represent the reality of
population structure. Iterative improvements to stock boundaries and
model specications will help to conform to unit stock assumptions.
7. Conclusion
Complying with the unit stock assumption may be the most impor-
tant structural decision in stock assessment modeling, and many stock
assessments can be improved by revising the stock boundary to
encompass a discrete biological population. Stock boundaries and strata
denitions should be routinely evaluated, informed by stock identi-
cation, and based on the most plausible stock structure. Iterative
application of these practices for stock identication and stock assess-
ments can advance assessment frameworks towards an appropriate
geographic scope and structure for meeting the management needs of
each shery (e.g., Fig. 2).
Spatially complex populations present challenges associated with
data limitations or jurisdictional constraints. Spatially mis-specied
stock assessment models may not accurately represent complex pop-
ulations. Therefore, simulation testing is needed to conrm acceptable
performance for informing shery management. The technical chal-
lenge for assessing spatially complex populations is the conditioning of
operating models for simulation testing that adequately represent
plausible scenarios of population structure informed by stock
identication.
CRediT authorship contribution statement
Steven Cadrin: Conceptualization, Writing −original draft prepa-
ration, Writing −review & editing preparation, Visualization. Daniel
Goethel: Writing −original draft preparation, Writing −review &
editing preparation. Aaron Berger: Writing −original draft prepara-
tion, Writing −review & editing preparation. Aaron Berger: Writing −
review & editing preparation.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data availability
No data was used for the research described in the article.
Acknowledgments
Thanks to Mark Maunder, Rishi Sharma and the CAPAM leadership
for the invitation to present our perspectives at the Stock Assessment
Good Practices Workshop (Rome, 24–28 October 2022). Our views on
stock identication and spatial stock assessment have developed
through collaboration with many scientists and shermen. We appre-
ciate Andr´
e Punt’s advance draft of best practice overview for the
workshop, correspondence with other participants in the workshop’s
Fishery and Stock Structure discussion group, an editorial review by
Katy Echave, and constructive comments from two reviewers.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
10
References
Aldenberg, T., 1975. Virtual population analysis and migration, a theoretical treatment.
ICES Doc. CM 1975/F. 32, 14.
Aldrin, M., Aanes, S., Subbey, S., 2019. Comments on incongruous formulations in the
SAM (state-space assessment model) model and consequences for sh stock
assessment. Fish. Res. 210, 224–227.
Beacham, T.D., Wallace, C., Jonsen, K., McIntosh, B., Candy, J.R., Rondeau, E.B.,
Moore, J.-S., Bernatchez, L., Withler, R.E., 2020. Accurate estimation of conservation
unit contribution to coho salmon mixed-stock sheries in British Columbia, Canada,
using direct DNA sequencing for single nucleotide polymorphisms. Can. J. Fish.
Aquat. Sci. 77, 1302–1315.
Begg, G.A., Waldman, J.R., 1999. An holistic approach to sh stock identication. Fish.
Res. 43, 35–44.
Bell, R.J., Richardson, D.E., Hare, J.A., Lynch, P.D., Fratantoni, P.S., 2015. Disentangling
the effects of climate, abundance, and size on the distribution of marine sh: an
example based on four stocks from the Northeast US shelf. ICES J. Mar. Sci. 72,
1311–1322.
Bentley, N., Davies, N.M., McNeill, S.E., 2004, A spatially explicit model of the snapper
(Pap auratus) shery in SNA 1. New Zealand Fisheries Assessment Report 2004/26.
Berger, A.M., Jones, M.L., Zhao, Y., Bence, J.R., 2012. Accounting for spatial population
structure at scales relevant to life history improves stock assessment: the case for
Lake Erie walleye Sander vitreus. Fish. Res. 115–116, 44–59.
Berger, A.M., Goethel, D.R., Lynch, P.D., Quinn, T., Mormede, S., McKenzie, J., Dunn, A.,
2017. Space oddity: the mission for spatial integration. Can. J. Fish. Aquat. Sci. 74,
1698–1716.
Berger, A.M., Deroba, J.J., Bosley, K.M., Goethel, D.R., Langseth, B.J., Schueller, A.M.,
Hanselman, D.H., 2021. Incoherent dimensionality in sheries management:
consequences of misaligned stock assessment and population boundaries. ICES J.
Mar. Sci. 78, 155–171.
Beverton, R.J.H., Holt, S.J., 1957. On the dynamics of exploited sh populations.
Ministry of Agriculture, Fisheries and Food. Fish. Investig., Lond., Ser. II XIX, 533.
Booke, H.E., 1981. The conundrum of the stock concept—are nature and nurture
denable in shery science? Can. J. Fish. Aquat. Sci. 38, 1479–1480.
Bosley, K.M., Goethel, D.R., Berger, A.M., Deroba, Fenske, K.H., J.J., Hanselman, D.H.,
Langseth, B.J., Schueller, A.M., 2019. Overcoming challenges of harvest quota
allocation in spatially structured populations. Fish. Res. 220, 105344.
Bosley, K.M., Schueller, A.M., Goethel, D.R., Hanselman, D.H., Fenske, K.H., Berger, A.
M., Deroba, J.J., Langseth, B.J., 2022. Finding the perfect mismatch: evaluating
misspecication of population structure within spatially explicit integrated
population models. Fish Fish 23, 294–315.
Bravington, M. 2023., Close-Kin Mark-Recapture. Fisheries research, in this issue.
Broquet, T., Petit, E.J., 2009. Molecular estimation of dispersal for ecology and
population genetics. Annu. Rev. Ecol. Evol. Syst. 40, 193–216.
Caddy, J.F., 1997. . Establishing a Consultative Mechanism or Arrangement for
Managing Shared Stocks Within the Jurisdiction of Contiguous States. In:
Hancock, D. (Ed.), Taking Stock: Dening and Managing Shared Resources,
Australian Society for Fish Biology and Aquatic Resource Management Association
of Australasia Joint Workshop Proceedings, Darwin, NT, 15-16 June 1997.
Australian Society for Fish Biology, Sydney, pp. 81–123.
Cadrin, S.X., 2020. Dening spatial structure for shery stock assessment. Fish. Res. 221,
105397.
Cadrin, S.X., Secor, D.H., 2009. Accounting for spatial population structure in stock
assessment: past, present and future. In: Rothschild, B.J., Beamish, R. (Eds.), The
Future of Fishery Science in North America. Springer Verlag, pp. 405–426.
Cadrin, S.X., Kerr, L.A., Mariani, S., 2014. Interdisciplinary evaluation of spatial
population structure for denition of shery management units. In: Cadrin, S.X.,
Kerr, L.A., Mariani, S. (Eds.), Stock Identication Methods: Applications in Fisheries
Science, 2nd ed..,. Elsevier Academic Press, Burlington MA, pp. 535–552.
Cadrin, S.X., Maunder, M.N., Punt, A.E., 2020. Spatial Structure: Theory, estimation and
application in stock assessment models. Fish. Res. 229, 105608.
Cao, J., Thorson, J.T., Punt, A., Szuwalski, C., 2020. A novel spatiotemporal stock
assessment framework to better address ne-scale species distributions:
Development and simulation testing. Fish Fish 21, 350–367.
Carruthers, T., Butterworth, D., 2018. Summary of a reference set of conditioned
operating models for Atlantic bluen tuna. ICCAT Collect. Vol. Sci (Papers 73),
3523–3541.
Carruthers, T., Punt, A.E., Walters, C.J., MacCall, A., McAllister, M.K., Dick, E.J.,
Cope, J., 2014. Evaluating methods for setting catch limits in data-limited sheries.
Fish. Res. 153, 48–68.
Carvalho, G.R., Hauser, L., 1994. Molecular genetics and the stock concept in sheries.
Rev. Fish. Biol. Fish. 4, 326–350.
Cope, J., 2023, Science-based sheries management under data and resource limitations:
The good practices of practicable alchemy. Fisheries Research in this issue.
Cope, J.M., Punt, A.E., 2011. Reconciling stock assessment and management scales under
conditions of spatially varying catch histories. Fish. Res. 107, 22–38.
Coyle, T., 1997. Stock identication and sheries management: the importance of using
several methods in a stock identication study. In: Hancock, D. (Ed.), Taking Stock:
Dening and Managing Shared Resources, Australian Society for Fish Biology and
Aquatic Resource Management Association of Australasia Joint Workshop
Proceedings, Darwin, NT, 15-16 June 1997. Australian Society for Fish Biology,
Sydney, pp. 173–182.
Currie, J.C., Thorson, J.T., Sink, K.J., Atkinson, L.J., Fairweather, T.P., Winker, H., 2019.
A novel approach to assess distribution trends from sheries survey data. Fish. Res.
214, 98–109.
Cushing, D.H., 1968. Fisheries Biology, A Study in Population Dynamics. Univ.
Wisconsin Press,, p. 200.
Cushing, D.H., Ed., 1983. Key Papers on Fish Populations. Oxford. 405 p.
Deroba, J.J., Butterworth, D.S., Methot Jr., R.D., De Oliveira, J.A.A., Fernandez, C.,
Nielsen, A., Cadrin, S.X., Dickey-Collas, M., Legault, C.M., Ianelli, J., Valero, J.L.,
Needle, C.L., O’Malley, J.M., Chang, Y.-J., Thompson, G.G., Canales, C., Swain, D.P.,
Miller, D.C.M., Hintzen, N.T., Bertignac, M., Ibaibarriaga, L., Silva, A., Murta, A.,
Kell, L.T., de Moor, C.L., Parma, A.M., Dichmont, C.M., Restrepo, V.R., Ye, Y.,
Jardim, E., Spencer, P.D., Hanselman, D.H., Blaylock, J., Mood, M., Hulson, P.-J.F.,
2015. Simulation testing the robustness of stock assessment models to error: some
results from the ICES strategic initiative on stock assessment methods. ICES J. Mar.
Sci. 72, 19–30.
Eagle, T.C., Cadrin, S.X., Caldwell, M.E., Methot, R.D., Nammack, M.F., 2008.
Conservation units of managed sh, threatened or endangered species, and marine
mammals. NOAA Tech. Mem. NMFS-OPR 37.
FAO (U.N. Food and Agriculture Organization Fisheries Department), 1994, World
review of highly migratory species and straddling stocks. FAO Fisheries Technical
Paper. No. 337. 70 p.
Francis, R.I.C.C., and McKenzie, J.R., 2015, Assessment of the SNA1 stocks in 2013. New
Zealand Fisheries Assessment Report 2015/76.
Frank, K., Brickman, D., 2000. Allee effects and compensatory population dynamics
within a stock complex. Can. J. Fish. Aquat. Sci. 57, 513–517.
Frawley, T.H., Muhling, B., Welch, H., Seto, K.L., Chang, S.-K., Blaha, F., Hanich, Q.,
Jung, M., Hazen, E.L., Jacox, M.G., Brodie, S., 2022. Clustering of disaggregated
sheries data reveals functional longline eets across the Pacic. One Earth 5,
1002–1018.
Frisk, M.G., Martell, S.J.D., Miller, T.J., Sosebee, K., 2008. Exploring the population
dynamics of inter skate (Leucoraja ocellata) in the Georges bank region using a
statistical catch-at-age model incorporating length, migration, and recruitment
process errors. Can. J. Fish. Aquat. Sci. 67, 774–792.
Gertseva, V., Cope, J.M. 2017. Stock assessment of the yelloweye rocksh (Sebastes
ruberrimus) in state and Federal waters off California, Oregon and Washington.
Pacic Fishery Management Council, Portland, OR. 〈http://www.pcouncil.org/
groundsh/stock-assessments/〉.
Goethel, D., Kerr, L.A., Cadrin, S.X., 2016. Incorporating spatial population structure into
the assessment-management interface of marine resources. In: Edwards, C.T.T.,
Dankel, D.J. (Eds.), Management Science in Fisheries: an Introduction to Simulation-
Based Methods. Taylor & Francis, London, pp. 319–347.
Goethel, D.R., Berger, A.M., 2017. Accounting for spatial complexities in the calculation
of biological reference points: effects of misdiagnosing population structure for stock
status indicators. Can. J. Fish. Aquat. Sci. 74, 1878–1894.
Goethel, D.R., Omori, K.L., Punt, A.E., Lynch, P.D., Berger, A.M., de Moor, C.L., Plag´
anyi,
´
E.E., Cope, J.M., Dowling, N.A., McGarvey, R., Preece, A.L., Thorson, J.T.,
Chaloupka, M., Gaichas, S., Gilman, E., Hesp, S.A., Longo, C., Yao, N., Methot, R.D.,
In press. Oceans of plenty? Challenges, advancements, and future directions for the
provision of evidence‑based sheries management advice. Rev. Fish Biol. Fisheries
https://doi.org/10.1007/s11160–022-09726–7.
Goethel, D.R., Legault, C.M., Cadrin, S.X., 2015. Testing the performance of a spatially
explicit tag-integrated stock assessment model of yellowtail ounder (Limanda
ferruginea) through simulation analysis. Can. J. Fish. Aquat. Sci. 72, 582–601.
Goethel, D.R., Berger, A.M., Cadrin, S.X., 2023, Spatial Awareness: Good Practices for
Implementing the Continuum of Stock 1 Assessment Approaches that Address Spatial
Population Structure and 2 Connectivity. Fisheries Research in this issue.
Gulland, J.A., 1980, Some Problems of the Management of Shared Stocks. FAO Fisheries
Technical Paper No. 206, Rome.
Harden-Jones, F.R., 1968. Fish Migration. St. Martin’s Press,, p. 325.
Hare, J., Able, K., 2007. Mechanistic links between climate and sheries along the east
coast of the United States: Explaining population outbursts of Atlantic Croaker
(Micropogonias undulatus). Fish. Oceanogr. 16, 31–45.
Hart, D.R., 2001. Individual-based yield-per-recruit analysis, with an application to the
Atlantic sea scallop, Placopecten magellanicus. Can. J. Fish. Aquat. Sci. 58,
2351–2358.
Haugen, J.B., Skomal, G.B., Curtis, T.H., Cadrin, S.X., 2022. Interdisciplinary stock
identication of North Atlantic porbeagle (Lamna nasus). J. North. Atl. Fish. Sci. 53,
1–18.
Hilborn, R., Sibert, J., 1988. Is International Management of Tuna Necessary? Mar.
Policy 12, 31–39.
Hilborn, R., Walters, C.J., 1992. Quantitative Fisheries Stock Assessment: Choice,
Dynamics, and Uncertainty. Chapman and Hall,, Boca Raton FL, p. 570.
Hurtado-Ferro, F., Szuwalski, C.S., Valero, J.L., Anderson, S.C., Cunningham, C.J.,
Johnson, K.F., Licandeo, McGilliard, C.R., Monnahan, C.C., Muradian, M.L., Ono, K.,
Vert-Pre, K.A., Whitten, A.R., Punt, A.E., 2015. Looking in the rear-view mirror: bias
and retrospective patterns in integrated, age-structured stock assessment models.
ICES J. Mar. Sci. 72, 99–110.
ICCAT (International Commission for the Conservation of Atlantic Tunas), 2001,
Workshop on bluen mixing (Madrid, Spain, September 3e7, 2001). SCRS/01/020.
ICES (International Council for the Exploration of the Sea), 2022b, Workshop on Stock
Identication of West of Scotland Sea Cod (WK6aCodID; outputs from 2021
meeting). ICES Scientic Reports. 4:5. 24 pp.
ICES (International Council for the Exploration of the Sea), 2022a, Stock Identication
Methods Working Group (SIMWG). ICES Scientic Reports. 4:72. 66 pp.
ICES (International Council for the Exploration of the Sea), 2009, Report of the
Workshop on Redsh Stock Structure (WKREDS), 22–23 January 2009, ICES
Headquarters, Copenhagen. ICES CM 2009/ACOM:37. 71 pp.
ICES (International Council for the Exploration of the Sea), 2020, Workshop on Stock
Identication of North Sea Cod (WKNSCodID). ICES Scientic Reports. 2:89. 82 pp.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
11
ICES (International Council for the Exploration of the Sea), 2021, Benchmark Workshop
on North Sea Stocks (WKNSEA). ICES Scientic Reports. 3:25. 756 pp.
Jacobsen, N.S., Marshall, K.N., Berger, A.M., Grandin, C., Taylor, I.G., 2022. Climate-
mediated stock redistribution causes increased risk and challenges for sheries
management. ICES J. Mar. Sci. 79, 1120–1132.
Jardim, E., Eero, M., Silva, A., Ulrich, C., Pawlowski, L., Holmes, S.J., Ibaibarriaga, L., De
Oliveira Riveiro, I., Alzorriz, N., Citores, L., Scott, F., Uriarte, A., Carrera, P.,
Duhame, E., Mosqueira, I., 2018. Testing spatial heterogeneity with stock assessment
models. PLoS One 13 (1), e0190791.
Kapur, M.S., Siple, M.C., Olmos, M., Privitera-Johnson, K.M., Adams, G., Best, J.,
Castillo-Jordan, C., Cronin-Fine, L., Havron, A.M., Lee, Q., Methot, R.D., Punt, A.E.,
2021. Equilibrium reference point calculations for the next generation of spatial
assessments. Fish. Res. 244, 106132.
Karp, M.A., Peterson, J.O., Lynch, P.D., Grifs, R.B., Adams, C.F., Arnold, W.S.,
Barnett, L.A.K., deReynier, Y., DiCosimo, J., Fenske, K.H., Gaichas, S.K.,
Hollowed, A., Holsman, K., Karnauskas, M., Kobayashi, D., Leising, A., Manderson, J.
P., McClure, M., Morrison, W.E., Schnettler, E., Thompson, A., Thorson, J.T.,
Walter III, J.F., Yau, A.J., Methot, R.D., Link, J.S., 2019. Accounting for shifting
distributions and changing productivity in the development of scientic advice for
shery management. ICES J. Mar. Sci. 76, 1305–1315.
Kerr, L., Murray, D., Cadrin, S., 2022, Advances in mixed stock analysis and their
potential role in improving precision of stock assessment. pp. 3–6 in Stransky, C., Ed.
ICES Scientic Reports 4, 72.
Kerr, L.A., Goethel, D.R., 2014. Simulation modelling as a tool for synthesis of stock
identication information. In: Cadrin, S.X., Kerr, L.A., Mariani, S. (Eds.), Stock
Identication Methods: Applications in Fisheries Science, 2nd ed..,. Elsevier
Academic Press, Burlington MA, pp. 501–533.
Kerr, L.A., Cadrin, S.X., Secor, D.H., 2010a. Simulation modelling as a tool for examining
the consequences of spatial structure and connectivity on local and regional
population dynamics. ICES J. Mar. Sci. 67, 1631–1639.
Kerr, L.A., Cadrin, S.X., Secor, D.H., 2010b. The role of spatial dynamics in the stability,
resilience, and productivity of an estuarine sh population. Ecol. Appl. 20, 497–507.
Kerr, L.A., Hintzen, N.T., Cadrin, S.X., Clausen, L., Worsøe Dickey-Collas, M., Goethel, D.
R., Hateld, E.M.C., Kritzer, J.P., Nash, R.D.M., 2017. Lessons learned from practical
approaches to reconcile mismatches between biological population structure and
stock units of marine sh. ICES J. Mar. Sci. 74, 1708–1722.
Lennert-Cody, C.E., Minami, M., Tomlinson, P.K., Maunder, M.N., 2010. Exploratory
Analysis of Spatial temporal Patterns in Length frequency Data: An Example of
Distributional Regression Trees. Fish. Res. 102, 323–326.
Lennert-Cody, C.E., Maunder, M.N., Aires-da-Silva, A., Minami, M., 2013. Dening
population spatial units: Simultaneous analysis of frequency distributions and time
series. Fish. Res. 139, 85–92.
Lindegren, M., Waldo, S., Nilsson, P.A., Svedang, H., Persson, A., 2013. Towards
sustainable sheries of the Oresund cod (Gadus morhua) through sub-stock-specic
assessment and management recommendations. ICES J. Mar. Sci. 70, 1140–1150.
Link, J.S., Nye, J.A., Hare, J.A., 2010. Guidelines for incorporating sh distribution shifts
into a sheries management context. Fish Fish 12, 461–469.
Link, J.S., Huse, G., Gaichas, S., Marshak, A.R., 2020. Changing how we approach
sheries: A rst attempt at an operational framework for ecosystem approaches to
sheries management. Fish Fish 21, 393–434.
MacCall, A.D., 1990. Dynamic Geography of Marine Fish Populations. University of
Washington Press,, Seattle, WA.
Marchal, P., 2008. A comparative analysis of m´
etiers and catch proles for some French
demersal and pelagic eets. ICES J. Mar. Sci. 65, 674–686.
Marr, J.C., 1957. Contributions to the study of subpopulations of shes. US Fish. Wildl.
Serv. Sp. Sci. Rep. Fish. 208.
Maunder, M.N., Punt, A.E., 2013. A review of integrated analysis in sheries stock
assessment. Fish. Res. 142, 61–74.
Methot, R.D., 2023, Overcoming hurdles to good practices in modeling selectivity in sh
stock assessment models. Fisheries Research in this issue.
Minte-Vera, C.V., Maunder, M.N., Xu, H., Teo, S.L.H., Aires-da-Silva, A., 2023, The use of
conceptual models to structure stock assessments: a tool for collaboration and for
“modelling what to model”. Fisheries Research in this issue.
Moore, B.R., Bell, J.D., Evans, K., Farley, J., Grewe, P.M., Hampton, J., Marie, A.D.,
Minte-Vera, C., Nicol, S., Phillips, J.S., Pilling, G.M., Tremblay-Boyer, L.,
Williams, A.J., Smith, N., 2020a. Dening the stock structures of key commercial
tunas in the Pacic Ocean I: current knowledge and key uncertainties. Fish., Res 230,
105525.
Moore, B.R., Adams, T., Allain, V., Bell, J.D., Bigler, M., Bromhead, D., Clark, S.,
Davies, C., Evans, K., Faasili, U., Farley, J., Fitchett, M., Grewe, P.M., Hampton, J.,
Hyde, J., Leroy, B., Lewis, A., Lorraino, A., Macdonald, J.I., Marie, A.D., Minte-
Vera, C., Natasha, J., Nicol, S., Obregon, P., Peatman, T., Pecoraro, C., Phillip, N.B.,
Pilling, G.M., Rico, C., Sanchez, C., Scott, R., Phillips, J.S., Stockwell, B., Tremblay-
Boyer, L., Usu, T., Williams, A.J., Smith, N., 2020b. Dening the stock structures of
key commercial tunasin the Paci c Ocean II: sampling considerations and future
directions. Fish. Res. 230, 105524.
Morse, M.R., Kerr, L.A., Cadrin, S.X., 2019, Evaluation of an F0.1 management procedure
using an alternative Management Strategy Evaluation framework for Atlantic bluen
tuna. ICCAT SCRS/2019/017.
Morse, M.R., Kerr, L.A., Galuardi, B., Cadrin, S.X., 2020. Performance of stock
assessments for mixed-population sheries: The illustrative case of Atlantic bluen
tuna. ICES J. Mar. Sci. 77, 2043–2055.
NEFMC (New England Fishery Management Council), 2022, Overshing Limits and
Acceptable Biological Catches for Southern New England/Mid-Atlantic Winter
Flounder, Georges Bank Yellowtail Flounder and Georges Bank Cod. 〈https://s3.us-
east-1.amazonaws.com/nefmc.org/3.-SSC-Report-Aug-25_22-mtg-Memo-09_02_22.
pdf〉.
NEFSC (Northeast Fisheries Science Center), 2012, 54th Northeast Regional Stock
Assessment Workshop (54th SAW) Assessment Report. NEFSC Ref. Doc. 12–18; 600
p.
NEFSC (Northeast Fisheries Science Center), 2020, Final report of red hake stock
structure working group. NEFSC Ref Doc. 20–07; 185 p.
Nielsen, A., Berg, C.W., 2014. Estimation of time-varying selectivity in stock assessments
using state-space models. Fish. Res. 158, 96–101.
Nielsen, A., Berg, C.W., 2023, Good practices and necessity for random effects in sheries
stock assessment. Fisheries Research, in this issue.
Ommer, R.E., Perry, R.I., 2022. The importance of complexities of scale and context in
sheries research and management. ICES J. Mar. Sci. 2022 (0), 1–8.
Pauly, D., 1984. Fish Population Dynamics in Tropical Waters: a Manual for Use with
Programmable Calculators. ICLARM Stud. Rev. 8, 325.
Pawson, M.G., Jennings, S., 1996. A critique of methods for stock identication in marine
capture sheries. Fish. Res. 25, 203–217.
PFMC, 2021, in press practices guidelines for groundsh assessments in 2021 and 2021.
〈https://www.pcouncil.org/documents/2021/02/in press-practices-and-guidelines-
for1059groundsh-stock-assessments-february-2021.pdf〉.
Porch, C., Kleiber, P., Turner, S.C., Sibert, J., Bailey, R.B., Cort, J.L., 1998. The efcacy of
VPA models in the presence of complicated movement patterns. Collect. Vol. Sci.
Pap. ICCAT 50, 591–622.
Porch, C., Turner, S.C., Powers, J.E., 2001. Virtual population analyses of Atlantic bluen
tuna with alternative models of transatlantic migration: 1970–1997. Col. Vol. Sci.
Pap. ICCAT 52, 1022–1045.
Pudden, E.J., VanderZwaag, D.L., 2007. Canada–USA bilateral sheries management in
the Gulf of Maine: under the radar screen. Reciel 16, 36–44.
Punt, A.E., 2019. Spatial stock assessment methods: a viewpoint on current issues and
assumptions. Fish. Res. 213, 132–143.
Punt, A.E., 2023. Those who fail to learn from history are condemned 1 to repeat it: a
perspective on current stock assessment good practices and the consequences of not
following them. Fish. Res. 261, 106642.
Punt, A.E., Hurtado-Ferro, F., Whitten, A.R., 2014. Model selection for selectivity in
sheries stock assessments. Fish. Res. 158, 124–134.
Punt, A.E., Butterworth, D.S., de Moor, C.L., De Oliveira, J.A.A., Haddon, M., 2016.
Management strategy evaluation: best practices. Fish Fish 17, 303–334.
Punt, A.E., Haddon, M., Little, L.R., Tuck, G.N., 2017. The effect of marine closures on a
feedback control management strategy used in a spatially aggregated stock
assessment: a case study based on pink ling in Australia. Can. J. Fish. Aquat. Sci. 74,
1960–1973.
Punt, A.E., Okamoto, D.K., MacCall, A.D., Shelton, A.O., Armitage, D.R., Cleary, J.S.,
Davies, I.P., Dressel, S.C., Francis, T.B., Levin, P.S., Jones, R.R., Kitka, H., Lee, L.C.,
McIsaac, J.A., Poe, M.R., Reifenstuhl, S., Silver, J.J., Schmidt, J.O., Thornton, T.F.,
Voss, R., Woodrufft, J., 2018. When are estimates of spawning stock biomass for
small pelagic shes improved by taking spatial structure into account? Fish. Res.
206, 65–78.
Punt, A.E., Dunn, A., Elvarsson, B.P., Hampton, J., Hoyle, S., Maunder, M.N., Punt, A.E.,
Cadrin, S., Fu, D., Lorenzen, K., Methot, R., Piner, K., Walter, J., 2019. 2nd review of
the stock assessment of bigeye tuna in the eastern Pacic Ocean. Inter-American
Tropical Tuna Commission Special Report. WSBET 02.
Punt, A.E., Dunn, A., Elvarsson, B., Hampton, J., Hoyle, S.D., Maunder, M.N., Methot, R.
D., Nielsen, A., 2020. Essential features of the next-generation integrated sheries
stock assessment package: a perspective. Fish. Res. 229, 105617.
Quinn II, T.J., Ianelli, J.N., Cadrin, S.X., Wespestad, V., Barbeaux, S.J., 2011. Report on a
workshop on spatial structure and dynamics of walleye pollock in the Bering Sea.
AFSC Process. Rep. 46, 2011-04.
Reiss, H., Hoarau, G., Dickey-Collas, M., Wolff, W.J., 2009. Genetic population structure
of marine sh: mismatch between biological and sheries management units. Fish
Fish 10, 361–395.
Ricker, W.E., 1958. Maximum sustainable yields from uctuating environments and
mixed stocks. J. Fish. Res. Bd. Can. 15, 991–1006.
Rodriguez-Rodriguez, G., Ballesteros, H.M., Sanchez-Llamas, E., Bande, R., Otero, R.F.,
2022. The state of the art in cost-benet of HTS methods for stock assessment: An
overview. Front. Mar. Sci. 9, 1005534.
Rousset, F., 2007. Inferences from Spatial Population Genetics (Third Edition). In:
Balding, D.J., Bishop, M., Cannings, C. (Eds.), Handbook of Statistical Genetics. John
Wiley and Sons, Ltd., pp. 945–979
Russell, E.S., 1931. Some theoretical considerations on the ‘overshing’ problem.
J. Cons. Int. Explor. Mer. 6, 3–20.
Schnute, J.T., Richards, L.J., 2001. Use and abuse of shery models. Can. J. Fish. Aquat.
Sci. 58, 10–17.
Secor, D.H., 2015. Migration Ecology of Marine Fishes. Johns Hopkins University Press,,
Baltimore.
SEDAR (Southeast Data, Assessment, and Review), 2018, SEDAR 58 Cobia Stock ID
Process Report Compilation. 116 pp. 〈http://sedarweb.org/sedar-58-stock-id-
process〉.
SEDAR (Southeast Data, Assessment, and Review), 2020, Gulf of Mexico and Atlantic
Scamp Stock ID Process Final Report. SEDAR68-SID-05. 42 pp. 〈https://sedarweb.
org/docs/wpapers/S68_SID_05_Stock_ID_Final_Report.pdf〉.
SEDAR (Southeast Data, Assessment, and Review), 2021, Gulf of Mexico Red Snapper
Stock ID Process Final Report. 〈http://sedarweb.org/sedar-74-gulf-mexico-red-
snapper-stock-id-process〉.
Sinclair, M., 1988. Marine Populations: An Essay on Population Regulation and
Speciation. University of Washington Press,, Seattle.
S.X. Cadrin et al.
Fisheries Research 262 (2023) 106650
12
Sinclair, M.M., Smith, T.D., 2002. The notion that sh species form stocks. ICES Mar. Sci.
Symp. 215, 297–304.
Skoglund, S., Whitlock, R., Petersson, E., Palm, S., Leonardsson, K., 2022. From spawner
habitat selection to stock-recruitment: Implications for assessment. Ecol. Evol. 12,
e9679.
Smith, S.J., Campana, S.E., 2010. Integrated stock mixture analysis for continuous and
categorical data, with application to genetic–otolith combinations. Can. J. Fish.
Aquat. Sci. 67, 1533–1548.
Sterner, T., 2007. Unobserved diversity, depletion and irreversibility: the importance of
subpopulations for management of cod stocks. Ecol. Econ. 61, 566–574.
Taylor, N.G., McAllister, M.K., Lawson, G.L., Carruthers, T., Block, B.A., 2011. Atlantic
Bluen Tuna: A Novel Multistock Spatial Model for Assessing Population Biomass.
PLoS ONE 6 (12), e27693. https://doi.org/10.1371/journal.pone.0027693.
Thorson, J.T., Wetzel, C., 2016, The status of canary rocksh (Sebastes pinniger) in the
California Current in 2015. 〈https://www.pcouncil.org/documents/2016/05/the-
status-of-canary-rocksh-sebastes-pinniger-in-the-california-current-in-2015-march-
2016.pdf/〉.
Trenkel, V.M., Charrier, G., Lorance, P., Bravington, M.V., 2022. Close-kin
mark–recapture abundance estimation: practical insights and lessons learned. ICES
J. Mar. Sci. 79, 413–422.
Ulrich, C., Wilson, D.C.K., Nielsen, R., Bastardie, F., Reeves, S.A., Andersen, B.S.,
Eigaard, O.R., 2012. Challenges and opportunities for eet- and m´
etier-based
approaches for sheries management under the European Common Fishery Policy.
Ocean and Coastal Management, 70, 38–47.
UN (United Nations), 1982, United Nations Convention on the Law of the Sea. UN Doc.
A/Conf.62/122.
UN (United Nations), 1995, United Nations Conference on Straddling Fish Stocks and
Highly Migratory Fish Stocks. Agreement for the Implementation of the United
Nations Convention on the Law of the Sea of 10 December 1982 Relating to the
Conservation and Management of Straddling Fish Stocks and Highly Migratory Fish
Stocks. UN Doc. A/Conf./164/37.
USA (United States of America), 2007, Magnuson-Stevens Fishery Conservation and
Management Act, as amended through January 12, 2007. 〈https://www.sheries.
noaa.gov/resource/document/magnuson-stevens-shery-conservation-and-
management-act〉.
Utter, F., Ryman, N., 1993. Genetic markers and mixed stock sheries. Fisheries 18,
11–21.
Valenzuela-Quinonez, F., 2016. How sheries management can benet from genomics?
Brief. Funct. Genom. 15, 352–357.
Waterhouse, L., Sampson, D.B., Maunder, M., Semmens, B.X., 2014. Using areas-as-eets
selectivity to model spatial shing: asymptotic curves are unlikely under equilibrium
conditions. Fish. Res. 158, 15–25.
Weston, A.E., Kerr, L.A., Cadrin, S.X., Morse, M.R., 2019, Performance of a F0.1
management procedure using alternative operating models. ICCAT SCRS/2019/127.
Wilberg, M., et al., 2023, Good practices for considering sex-structured selectivity and
shing mortality in assessments: a simulation study. Fisheries Research, in this issue.
WPFMC (Western Pacic Regional Fishery Management Council), 2014, Workshop on
Pacic Bigeye Movement and Distribution, April 22–24, 2014. 〈https://www.
wpcouncil.org/wp-content/uploads/2014/11/Final-Bigeye-Workshop-Report.pdf〉.
Zemeckis, D.R., Martins, D., Kerr, L.A., Cadrin, S.X., 2014. Stock identication of Atlantic
cod (Gadus morhua) in US waters: an interdisciplinary approach. ICES J. Mar. Sci. 71,
1490–1506.
S.X. Cadrin et al.
