Thermal tolerance and habitat preferences mediate how freshwater fish body sizes respond to warming

Abstract

Recent studies have linked ecosystem warming to decreased mean body size in aquatic ectotherms but have often relied on experimental approaches. This study explores how climate change modifies adult body sizes in natural communities for five culturally, commercially, and recreationally important freshwater fish species that vary in thermal preferences (lake trout (Salvelinus namaycush), lake whitefish (Coregonus clupeaformis), walleye (Sander vitreus), yellow perch (Perca flavescens), and smallmouth bass (Micropterus dolomieu)). Using data for >170 000 sampling records from >600 lakes spanning the entire province of Ontario, Canada, we employ a spatial analysis technique based on Integrated Nested Laplace Approximation to detect body size changes due to warming lake conditions. With warming conditions, cold-adapted species (lake whitefish) decreased in adult body size, while cool- and warm-adapted species (walleye and smallmouth bass) increased in adult body size. Our study, like others recently conducted in marine and freshwater systems, demonstrates that the responses of fish body sizes to warming are not a unidirectional shift and may be linked to the habitat and thermal preferences of individual fish species.

Full text

Article
Thermal tolerance and habitat preferences mediate how
freshwater fish body sizes respond to warming
Connor P. K. Warne a, Matthew M. Guzzo b, Kevin Cazelles c,CindyChu d, Neil Rooneya, and Kevin S. McCann c
aSchool of Environmental Sciences, University of Guelph, Guelph, ON, Canada; bFreshwater Institute, Fisheries and Oceans Canada,
Winnipeg, MB, Canada; cDepartment of Integrative Biology, University of Guelph, Guelph, ON, Canada; dGreat Lakes Laboratory for
Fisheries and Aquatic Sciences, Fisheries and Oceans Canada, Burlington, ON, Canada
Corresponding author: Connor P. K. Warne (email: cwarne@uoguelph.ca)
Abstract
Recent studies have linked ecosystem warming to decreased mean body size in aquatic ectotherms but have often relied
on experimental approaches. This study explores how climate change modifies adult body sizes in natural communities for
five culturally, commercially, and recreationally important freshwater fish species that vary in thermal preferences (lake trout
(Salvelinus namaycush), lake whitefish (Coregonus clupeaformis), walleye (Sander vitreus), yellow perch (Perca flavescens), and small-
mouth bass (Micropterus dolomieu)). Using data for >170 000 sampling records f rom >600 lakes spanning the entire province of
Ontario, Canada, we employ a spatial analysis technique based on Integrated Nested Laplace Approximation to detect body size
changes due to warming lake conditions. With warming conditions, cold-adapted species (lake whitefish) decreased in adult
body size, while cool- and warm-adapted species (walleye and smallmouth bass) increased in adult body size. Our study, like
others recently conducted in marine and freshwater systems, demonstrates that the responses of fish body sizes to warming
are not a unidirectional shift and may be linked to the habitat and thermal preferences of individual fish species.
Key words: climate change, body size, temperature-size rule, INLA
Introduction
Lakes are warming across the globe, resulting in increased
surface water temperature profiles (O’Reilly et al. 2015),
altered mixing regimes (Woolway and Merchant 2019),
reductions in ice cover (Magnuson et al. 2000), altered
seasonality (Guzzo et al. 2017), and decreased oxygen con-
centrations (Jane et al. 2021). Ultimately, these changes
threaten the status of important native predatory fish
species that play essential roles in structuring lake com-
munities (Tunney et al. 2014), confer stability to food webs
(Rooney and Mccann 2011), and represent the foundation
of commercial and recreational fisheries (Beard et al. 2011).
Warming conditions in lakes can also alter the composition
and functional organization of fish communities (Guzzo
et al. 2017;Jacobson et al. 2017;Bartley et al. 2019)and
alter fish biomass, productivity (Gutowsky et al. 2019;Jarvis
et al. 2020), and recruitment (Dippold et al. 2020). How
lake warming aects the adult body sizes of freshwater
fish species, however, is an important yet unaddressed
question.
Warming conditions elicit a set of universal responses in
ectotherms such as fish, one of which may include reduc-
tions in mean adult body sizes (Gardner et al. 2011;Ohlberger
et al. 2011). These declines have been reported for many ec-
totherms (Gardner et al. 2011;Sheridan and Bickford 2011)
and are strongest in aquatic environments (Forster and Hirst
2012). Body size shifts in ecosystems can occur at individual,
population, or community levels, potentially resulting from:
i) decreased size-at-age of individuals; ii) increased propor-
tion of smaller or younger individuals; or iii) greater abun-
dance of smaller species (Ohlberger 2013). Body size is a key
trait related to intra- and interspecific interactions (Dell et al.
2011;Ohlberger 2013), population ecology (Barneche et al.
2016), and the productivity of fisheries (Baudron et al. 2013).
It is therefore imperative that the responses of thermally var-
ied freshwater fish species to warming lakes be categorized
to better understand and manage large, economically impor-
tant freshwater fisheries.
The thermal response of ectotherms to temperatures dur-
ing ontogeny is explained by the temperature-size rule (TSR)
(Atkinson 1994), which dictates that higher temperatures in-
crease rates of growth and development but decrease overall
adult body size, resulting in important shifts in size-at-age
and/or maturation (Ohlberger 2013). Empirical evidence
suggests that the body sizes of many aquatic ectotherms are
shrinking at a rate of 3% per 1 ◦Cofwarming(Angilletta et al.
2004;Forster and Hirst 2012;Hoefnagel and Verberk 2015;
Horne et al. 2015) and fisheries-dependent data has detected
smaller adult body sizes for marine fishes inhabiting warmer
waters (Thresher et al. 2007;Baudron et al. 2013). Despite the
proposed ubiquity of this response, trends in the opposite
direction have been noted for marine fishes (Thresher et al.
488 Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233 489
Table 1. Thermal preferences for the five examined freshwater fish species (adapted from
Hasnain et al. 2010).
Species
Mean optimum growth
temperature (OGT,◦C)
Final temperature
preferendum (FTP, ◦C)
Upper incipient lethal
temperature (UILT, ◦C)
Smallmouth bass 26 25 36
Yellow perch 25.4 17.6 25.6
Walleye 22.1 22.5 29.7
Lake whitefish 14.7 12.7 23.9
Lake trout 10 11.8 24.3
2007). There are also numerous cases where related aquatic
species may increase or decrease in size in response to the
same warming trends (O’Gorman et al. 2012;Audzijonyte et
al. 2020;Solokas et al. 2023). Varied body size shifts in the
face of warming waters point to species-specific responses,
which might be caused by dierences in thermal habitat
preferences. Fish body temperatures closely follow those
of their environment as ectotherms, and therefore they
must inhabit species-specific temperature ranges (Table 1)to
optimize their physiological performance.
During the summer months in temperate lakes, surface wa-
ters are warm and impose a metabolic cost for cold-adapted
freshwater fish species (e.g., lake trout) to occupy. Cold-water
predators change their foraging behaviour to reduce contact
with warmer waters by foraging in littoral zones in spring
and fall (when surface waters are cold) and in pelagic zones
in the summer (Guzzo et al. 2017). Climate warming will con-
tinue to increase the length of the summer (ice o) season
for temperate lakes (Magnuson et al. 2000), which in turn
will reduce the extent of suitable habitat for cold-adapted
fish species to inhabit during the summer months, altering
energy pathways (Bartley et al. 2019), and hindering growth
and condition. We therefore hypothesize that there is a con-
tinuum of responses to warming for freshwater fish species,
predicated on their thermal habitat preferences (Smith et al.
2021). Cold-adapted species will have to forage in suboptimal
thermal conditions, exposing them to warmer temperatures
throughout life that will result in smaller adult body sizes (as
predicted by the TSR). Warm-adapted species will show in-
creases in adult body sizes as they will have a greater amount
of suitable thermal habitat at their disposal over time. Finally,
cool-water-adapted species will exhibit no changes in adult
body sizes given that their thermal preferences lie in between
cold and warm-water-adapted fish.
Here, we use length and age data for >174 000 individual
fishes from 640 lakes sampled between 2008 and 2017 to
quantify the eect of lake warming on body size in five preda-
tory freshwater fish species. These species represent cold,
cool, or warm-water-adapted freshwater fish species. The data
were collected as part of the Broad-scale Monitoring (BsM)
Program for Inland Lakes, a standardized monitoring pro-
gram operated by the Ontario Ministry of Natural Resources
and Forestry (OMNRF) in the province of Ontario in Canada
(Sandstrom et al. 2013). These data provide a unique oppor-
tunity to test for evidence of the TSR across hundreds of pop-
ulations and to examine how fish respond to variation in the
warming of lakes.
Materials and methods
Study area
The province of Ontario, Canada covers a vast spatial range
(1 060 000 km2, latitudinal range 41◦–57◦N) and contains
nearly 250 000 lakes (Ontario Ministry of Natural Resources
1984;Fig. 1). Ontario’s large spatial scale results in dier-
ences in regional climate in mean annual air temperatures
that span 14 ◦C and 910 mm of mean annual precipitation
(Crins et al. 2009).
Dataset
To investigate how climate warming is modulating fresh-
water fish body size, we used lake characteristics and fish
catch data for Ontario lakes that have been sampled as part
of the first and second cycles of the BsM Program for Inland
Lakes (see Sandstrom et al. 2013). These surveys took place be-
tween 2008 and 2017 and lakes were selected based on lake
surface area, and for some lakes, the presence of target fish-
eries species. We applied an upper limit for lake surface area
of 10 000 ha since the largest lakes included in the BsM are
generally more likely to be commercially fished, and this fac-
tor would confound our analyses. Fishes were sampled us-
ing both small and large mesh monofilament gill nets. Small
mesh gill nets used in the BsM protocol contain panels with
six dierent mesh sizes (13, 19, 25, 32, 28, and 38 stretch mm),
and large mesh gill nets contain panels with eight dierent
mesh sizes (38, 51, 63, 76, 89, 102, 114, and 127 stretch mm).
Using both small and large mesh gill nets, the BsM proto-
col eectively targets fish 65–650 mm in size. This standard-
ized approach for sampling eort and gear allows for compar-
isons among systems. We selected five fish species (lake trout
(Salvelinus namaycush), lake whitefish (Coregonus clupeaformis),
walleye (Sander vitreus), yellow perch (Perca flavescens), and
smallmouth bass (Micropterus dolomieu)) for analysis based on
their relevance to recreational and commercial fisheries, vari-
ation in thermal tolerances (Table 1), and the fact that they
were sampled frequently during the BsM program, allowing
for robust statistical analysis.
In addition to fish length and age data, we obtained a
suite of variables known to modify freshwater fish body sizes
and included metrics of climate change and exploitation
(Arthington et al. 2016;Table 2). Specifically, for each lake,
we obtained, a) mean growing degree days (GDD515) over the
growing season for the 15 years prior to sampling, b) mean
annual precipitation (MAP15) over the growing season for the
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
490 Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233
Fig. 1. Map of Ontario showing the location of sampled lakes for both cycle 1 (2008–2012) and cycle 2 (2013–2017) of the
Broad-scale Monitoring program (Ontario Ministry of Natural Resources and Forestry).
15 years prior to sampling, and c) angling pressure data from
the BsM dataset. We chose the interval of 15 years prior to
the sampling event to generate climate data sets because 15
years covers most age-classes and many individuals for the
five species we examine here (Supplement A, Figs. S1–S5).
Most of the fish included in our study experienced these cli-
mate conditions during ontogeny and for years afterwards.
To generate datasets for metrics a) and b), we harvested long-
term mean climate grids (10 km2resolution) and their asso-
ciated bioclimatic parameters (GDD, MAP) from Natural Re-
sources Canada (cfs.nrcan.gc.ca) for the 15 year period prior
to BsM sampling for each lake used in our analyses. Growing
degree days (GDD) are a measure of heat accumulation cal-
culated by subtracting a base value temperature (5 ◦C, in our
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233 491
Table 2. Biological justifications and predictions for the eects of explanatory variables and random intercepts on the adult
body sizes of freshwater fish.
Variable Justification Predicted eects on adult fish body size
GDD15 growing
season
Mean growing degree days >5◦C (during the growing season) over
the 15 years prior to fish sampling event——categorizes lakes on a
“warm–cold” gradient, acting as a proxy for lake warming regimes.
Negative eect of warming for
cold-water-adapted fishes, no
eect/positive eect for
cool/warm-water-adapted fishes
MAP15 growing
season
Mean precipitation (during the growing season) over the 15 years
prior to fish sampling event, precipitation varies across Ontario
(increasing with longitude) and is integral to lake water budgets
although its direct role in governing fish population dynamics is
not well understood (Chu et al. 2016).
Positive eect of precipitation for all
species
Angling pressure Angling activity estimated for each lake based on flight surveys
conducted by the OMNRF. Adult removal through fishing
exploitation is a key determinant of fish body size (Emmrich et al.
2011;Shin et al. 2005)
Negative eect of angling pressure for all
species
Spatially
dependent
random
intercept (uij)
Biological variables and lake stressors are variable and correlated
across space (Feld et al. 2016). Freshwater fish populations (lakes)
in proximity are likely to experience similar conditions.
Importance of influence determined by
Deviance Information Criterion.
FMZ (random
intercept, ai)
Fisheries management zones delineate harvest regulations and
categorize ecological factors and angler use patterns (Ontario
Ministry of Natural Resources, 2005).
Importance of influence determined by
Deviance Information Criterion.
study, see Chezik et al. 2014) from mean daily temperatures
(zero if negative) and are used here as a metric of heat load at
each lake site. Here, we counted GDD515 and MAP15 for the
growing season of years only, using the temperature-based
protocol of Natural Resources Canada (NRCAN). The NRCAN
methodology defines the start of the growing season as when
the mean daily temperature is greater than or equal to 5 ◦C
for 5 consecutive days beginning March 1. The growing sea-
son then ends when the average daily minimum temperature
is less than −2◦C beginning August 1 (cfs.nrcan.gc.ca). In
this manner, we can determine the average heat load and
precipitation experienced by each lake used in our analyses
to categorize lakes along a gradient of growing season cli-
mate conditions. For metric c), we used the angling activity
(anglers·km−2year−1) metric derived using aerial surveys and
a derivation technique as part of the BsM protocol (see also
Chu et al. 2016). Surveys of angling activity are completed
in the summer months (eight weekday and eight weekend
flights counting shoreline and vessel anglers), as well as in
the winter (six weekday and six weekend flights counting
ice huts and open ice fishers). The mean angling activity is
calculated by adding the weekday and weekend counts for a
season and dividing by lake surface area. This activity metric
is multiplied by the number of anglers (by method), season
length, and hours fished per day and then summed for all sea-
son to produce the annual estimate of angling pressure we
use within our models. Lastly, the associated Fisheries Man-
agement Zones (FMZ) of each lake were obtained. Freshwater
fisheries in Ontario are protected and monitored by dividing
the province into 20 distinct FMZ that broadly reflect biogeo-
chemical features across the province and could influence
fish body size (Ontario Ministry of Natural Resources 2005).
Data analysis
We first estimated the mean asymptotic lengths (L∞;i.e.,
maximum adult body size) for each fish population by fitting
length and age data from all specimens to von Bertalany
growth functions (VBGF) (Beverton, 1954). We used a sample
size threshold of n≥15 specimens per lake to fit VBGF as
a compromise between including enough data to fit VBGFs
properly and including as many lakes as possible in our subse-
quent analyses. The number of lakes included for subsequent
analyses of L∞for each of the five species ranged from 129
(yellow perch) to 629 (walleye) (Table 3). While fitting VBGF
for each lake/species combination to estimate L∞, other pa-
rameters (K,t0) were not kept fixed.
Prior to statistical analysis, lake and species-specific adult
fish body sizes (L∞) and lake (biogeochemical, climate, and an-
gling pressure) data were visualized using histograms to iden-
tify and remove outliers. Collinearity between variables was
evaluated using the variance inflation factor and a threshold
set at ≤8(Feld et al. 2016). Continuous covariates were stan-
dardized (subtracting the mean from individual values and
dividing by the standard deviation) prior to statistical analy-
ses. To quantify and compare climatic and exploitation pres-
sures on the adult sizes of Ontario freshwater fish, we used a
method for approximate Bayesian inference called Integrated
Nested Laplace Approximation (INLA) using the R-INLA pack-
age (Rue et al. 2009). INLA has increased in popularity in the
ecological literature since it is a computationally fast, deter-
ministic method (compared to Markov chain Monte Carlo,
a probabilistic method). INLA allows for the construction of
Gaussian Markov random fields (GMRFs), which represent es-
timates of parameters for datasets that contain complex spa-
tial structures (Lindgren et al. 2011;Beguin et al. 2012;Bivand
et al. 2015), making it an ideal method to account for spatial
autocorrelation among lakes across Ontario (Gutowsky et al.
2019).
Our full model contained the fixed eects of our co-
variates mentioned above and a random intercept for FMZ
(Table 2). To account for spatial autocorrelation among lakes,
where nearby lakes experience similar anthropogenic stres-
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
492 Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233
Table 3 . Deviance information criterion (DIC) values for various INLA models tested.
Model Random eects included
DIC
(Lake trout)
DIC
(Lake whitefish)
DIC
(Walleye)
DIC
(Yellow perch)
DIC
(Smallmouth bass)
Sample size (nlakes included) 219 355 629 129 183
1 None (multiple regression) 1908.62 3349.64 5501.44 858.38 1184.65
2 FMZ 1909.22 3350.05 5499.72 857.86 1142.58
3FMZ+spatially correlated random eect −125.94 2555.06 5415.31 220.20 603.22
4 spatially correlated random eect 587.44 1782.87 2481.32 -46.19 512.04
Note: Smallest DIC value for each species is marked in bold. FMZ, Fisheries management zone.
sors and/or the same biogeochemical structure, we also spec-
ified a latent spatial term that represented correlated spatial
random eects:
L∞ij =Zij +ai+uij +εij
ai∼N0,σ
a2
εij ∼N0,σ
ε
2
uij ∼N(0,
)
where fish species adult size (L∞) for the jth lake within the
ith FMZ is explained by all fixed eects Zij. Hyperparameters
were a random intercept aiwith a mean of zero and a vari-
ance σa2, the spatially dependent random intercept uij with
a mean of zero and a variance–covariance matrix , and the
residual εij, which is assumed to have a mean of zero and a
variance σε2. The variance–covariance matrix () is estimated
using the continuous domain stochastic partial dierential
equation (SPDE) method described in Lindgren et al. 2011.
The SPDE approach uses a “mesh” of discrete sampling lo-
cations (lake coordinates), which is interpolated to estimate
a continuous GMRF that is used to estimate the correlation
matrix (Bivand et al. 2015).
To determine whether the random eects we included
in our model were important, we compared models with
and without these random eects and evaluated them
based on the deviance information criterion (DIC; Table 3)
(Spiegelhalter et al. 2002). We compared models that con-
tained either (a) no random eects (multiple regression), (b)
FMZ as a random eect, (c) FMZ random eect and an SPDE
random intercept, or (d) an SPDE random intercept. The fi-
nal models (for each species) were examined by plotting the
residuals versus fitted values and covariate (Zuur et al. 2009).
Results
Lakes included in our study ranged greatly in mor-
phometric characteristics, including surface area
(range =259 993 ha, mean =1771 ha) and mean depth
(0.6–40.1 m, mean =9.83 m). GDD over the 15 year pe-
riod prior to sampling (GDD5) during the growing season
ranged from 754 to 2100 (mean =1479) and mean annual
precipitation in the same span varied between 285 and 637
(mean =488) mm. Angling intensity among sampled lakes
ranged between 0 and 18.52 (mean =0.59) anglers·km−2
year−1.
Since we used a specific threshold for the minimum num-
ber of fish specimens (data points) to fit VBGF for any
lake/species combination (n≥15), we wanted to explore
that filter and determine if using higher or lower thresholds
would have any bearing on our results. We used 11 dierent
thresholds (n≥10,11,12…20) for VBGF sample size (number
of fish per species per lake) and computed the INLA analysis
with those dierent sets of lakes included and found that in
general, our results were robust to dierences in sample sizes
used to compute L∞from VBGF (Supplement A, Table S1).
The addition of random eects improved the model com-
pared to simple multiple regression (Table 3); however, the
model with the lowest DIC score varied among our five model
species. Models for lake trout, walleye, yellow perch, and
smallmouth bass had the lowest DIC when only a spatial ran-
dom eect was used. Lake whitefish had its lowest DIC with
a model including FMZ and a spatial random eect (Table 3).
Out of the 9 variables included in our INLA analyses, we
focus here on the one that measures the impact of lake
warming (heat accumulation (GDD515)) on the adult sizes of
five freshwater fish species in Ontario. For the cold-adapted
species included in our INLA analyses, we found that GDD515
was a significant negative predictor of maximum adult body
size (L∞) for lake whitefish and a non-significant positive pre-
dictor of L∞for lake trout (Fig. 2). These results are at odds
with one another; however, when we consider the suite of
results gathered from running the same INLA analyses on
various sample size thresholds (n=10…20), we find that
most of the time, lake whitefish (>90%) and lake trout (>70%)
increase in adult body size with warming, although the re-
sults for lake trout are always insignificant (Table S1). GDD515
was a significant predictor of increased adult body size for
cool-water-adapted walleye and for the warm-water-adapted
smallmouth bass (Fig. 2). GDD515 was also a positive predictor
of adult body size for the warm-water-adapted yellow perch,
however, insignificantly so (Fig. 2).
We also ran the same INLA models using Kappa (growth
rate) from the results of lake/species VBGFs (in contrast to
L∞, used in our main analysis) as the response variable. This
was done to detect any influence that climate or other factors
might have on early growth rates (as opposed to maximum
adult body sizes), but we found mostly insignificant results
(Figure S6).
Discussion
In experimental studies, ectotherms often decrease in av-
erage body size in response to warming conditions (Atkinson
1994), especially in aquatic systems (Forster et al. 2012).
However, recent studies examining marine (Huss et al. 2019;
Audzijonyte et al. 2020) and freshwater fish populations
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233 493
Fig. 2. Integrated Nested Laplace Approximation coecient plot. The eect of explanatory variables (y-axis) on adult body
sizes for target species (n=5) are shown; if ±95% credible intervals overlap with zero (middle dashed blue line), the results
are deemed insignificant.
(Solokas et al. 2023) have shown stark departures from the
predicted outcomes of the TSR, indicating that fish species
may exhibit dierential body size responses to warming
based on thermal tolerances or other species-/population-
specific factors. The responses of fish body size to warming
have important consequences for aquatic community struc-
ture, population productivity, and resilience to stressors like
exploitation. Warming air temperatures increase the surface
water temperatures of lakes (O’Reilly et al. 2015), alter sea-
sonal thermal conditions, and can reduce or even eliminate
the amount of optimal habitat for cold-adapted fish species
(Guzzo and Blanchfield 2017), forcing them to experience
high temperatures during growth and development or alter
their habitat use (Plumb and Blanchfield 2009;Guzzo et al.
2017). Therefore, we predicted that fish species occupying
dierent habitats within the same lakes are experiencing
dierential warming signals (Bartley et al. 2019)andare
responding with asymmetrical modifications to body size.
We predicted that results falling in line with the TSR (greater
warming causing smaller adult body sizes) would be the case
for cold-water-adapted species and that the opposite would
be true for warm-adapted fish, while cool-adapted fish would
show no response.
We found that dierential warming conditions in lake
ecosystems resulted in varied changes in adult body sizes
for five species with dierent thermal preferences. Most im-
portantly, we found evidence for decreases in lake whitefish
(significant) and lake trout (non-significant) (Fig. 2 and Ta-
ble S1) maximum adult body sizes in lakes that had a higher
heat load compared to those experiencing cooler conditions.
These results support our prediction that cold-water-adapted
freshwater fish species would show evidence of the TSR, con-
sidering that they are both dependent on cold water tem-
peratures of less than 14.7 ◦C to grow optimally (Table 1).
We found that this result was more pronounced (and was
consistently significant) for lake whitefish compared to lake
trout (Fig. 2, Table S1), which is the opposite of what we ex-
pected to find, given that lake trout has a lower optimum
growth rate compared to lake whitefish (lake trout =10 ◦C,
lake whitefish =14.7 ◦C). Lake whitefish rely on cold-water
benthos to forage (Ebener et al. 2008)andareabletomake
many forays into warmer (less oxygen-rich) waters through-
out the year. These movements across available habitats may
lead to changes in foraging eciency, which could also ex-
pose them to physiological stress (Rodrigues et al. 2022), po-
tentially causing changes in growth rates consistent with TSR
and the evidence that we see for reduced adult body sizes
with warming (Fig. 2, Table S1). Our results for lake trout
conflict with findings demonstrating their limited capacity
to acclimate thermally to warmer environments (Kelly et al.
2014); they experience warmer water temperatures metabol-
ically with little ability to buer responses. Given that lake
waters continue to rapidly warm, we expected that lake trout
would experience greater stress during both ontogeny and as
adults (Guzzo and Blanchfield 2017), causing them to achieve
smaller maximum adult body sizes. We did find that increas-
ing temperatures led (most often) to reductions in body size
for lake trout (Table S1); however, these results were non-
significant and at best provided evidence for the overall trend
of reduced body sizes for cold-adapted freshwater fish species
to warming.
Conversely, for the cool- and warm-water-adapted species
we considered (walleye, yellow perch, and smallmouth bass)
our results suggest that warming lake conditions could be
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
494 Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233
leading to increases in adult body sizes (Fig. 2, Table S1). In
contrast to walleye and smallmouth bass, yellow perch shows
non-significant results for the n=15 sample size threshold
(Fig. 2), however, we note that when the same INLA analy-
ses are run for various sample size thresholds, yellow perch
body size always increases with higher heat loads on lakes,
although always non-significantly (Table S1). These results
agree with the predictions of the TSR (and with our pre-
dictions), and they suggest that warming waters will signif-
icantly alter maximum body sizes for these cool- and warm-
adapted freshwater fish species (Atkinson et al. 2006).
The general trend we found for thermal adaptation-
mediated responses to warming for freshwater fish species
lends credence to the recent findings of Solokas et al. (2023),
who found that many freshwater salmonids do not follow
TSR-predicted changes in body size in response to warm-
ing. Instead, that study found great variation in body size
responses, even for closely related species. Audzijonyte et
al. (2020) found similar results when examining hundreds
of reef fish species in Australian waters, as did van Rijn et
al. (2017) in the Mediterranean Sea; body size changes with
warming were not equal across species. Our results indicate
that the responses of freshwater fish body sizes to warm-
ing conditions are more complex than a unidirectional shift,
which may have profound impacts on the functional roles of
species and thereby on the food web structure of changing
lakes (Guzzo et al. 2017;Bartley et al. 2019). Size-dependent
temperature responses fundamentally impact the dynamics
and regulation of whole populations (Ohlberger et al. 2011;
Lindmark et al. 2017), and more research is needed to un-
cover species- and/or trait-specific responses to warming wa-
ters in lakes.
Freshwater ecosystems are especially vulnerable to ecolog-
ical stressors (Jackson et al. 2015) while providing important
services to surrounding ecosystems and human communi-
ties. Our study oers a data-rich test for warming-induced
body size shifts in natural freshwater fish species. We found
a thermally mediated response in these fish species, with
significant impacts on both cold-, cool-, and warm-adapted-
species. Studies examining warming-induced size shifts in
fish have often focused on exploited populations (Huss et al.
2019). Although our five selected species are all important
parts of recreational and/or commercial fisheries in Ontario,
we have accounted for removals by excluding large lakes that
might be home to commercial fishing and by including esti-
mates of angling pressure within our analyses. Exploitation
is an important component of our models, as it can help to
rule out size shifts based on the dierential angling rates of
lakes, which reduces size-at-age and growth rates through de-
mographic eects and evolutionary responses (Jørgensen et
al. 2007;Östman et al. 2014).
Our results show that warming alone (after accounting
for other important stressors or factors that can influence
fish body size) can explain substantial changes in adult body
sizes for freshwater species that are related to their thermal
habitat preferences. Our study is limited by the inclusion of
a small number of species (five), since our analytical method-
ology precluded the use of any species that are not caught
in very high numbers in Ontario lakes. These results show
that accurate predictions of warming-related size shifts in
freshwater fish require an understanding of species-specific
thermal preferences (Comte and Olden 2016) and should
be a priority for the management and protection of the
biodiversity of inland waters in a rapidly changing world.
Acknowledgements
This research was supported by Natural Sciences and En-
gineering Research Council Discovery and Canada First Re-
search Excellence Fund (Food from Thought) grants held by
Kevin McCann and Neil Rooney.
Article information
History dates
Received:17August2023
Accepted: 14 December 2023
Accepted manuscript online: 18 January 2024
Version of record online: 27 March 2024
Copyright
© 2024 Authors Warne, Cazelles, Rooney, McCann, and The
Crown. Permission for reuse (free in most cases) can be ob-
tained from copyright.com.
Data availability
We do not have permission to share the data used for this
project, as it was generated by OMNRF. We nonetheless pro-
vide all the requisite code and a mock data set to demonstrate
how the code can be used to replicate all analyses within the
manuscript.
The code, mock data sets, and instructions to run the
analyses are housed in a GitHub repository: github.com/
McCannLab/Ontario_fish_body_size and within a Zenodo
repository: Zenodo DOI: 10.5281/zenodo.8247988.
Author information
Author ORCIDs
Connor P. K. Warne https://orcid.org/0000-0001-9601-4543
Matthew M. Guzzo https://orcid.org/0000-0001-9229-4410
Kevin Cazelles https://orcid.org/0000-0001-6619-9874
Cindy Chu https://orcid.org/0000-0002-1914-3218
Kevin S. McCann https://orcid.org/0000-0001-6031-7913
Author contributions
Conceptualization: CPKW, MMG, NR, KSM
Data curation: CPKW, MMG, KC, CC, NR, KSM
Formal analysis: CPKW, MMG, KC, CC, NR, KSM
Funding acquisition: NR, KSM
Investigation: CPKW, MMG, KC, CC, NR, KSM
Methodology: CPKW, MMG, KC, CC, NR, KSM
Project administration: MMG, NR, KSM
Resources: NR, KSM
Software: KC
Supervision: MMG, NR, KSM
Validation: CPKW, MMG, KC, CC, NR, KSM
Visualization: CPKW, MMG, KC, CC, NR, KSM
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233 495
Writing – original draft: CPKW, MMG, NR, KSM
Writing – review & editing: CPKW, MMG, KC, CC, NR, KSM
Competing interests
The authors declare that they have no competing interests
relating to this research.
Supplementary material
Supplementary data are available with the article at https:
//doi.org/10.1139/cjfas-2023-0233.
References
Angilletta, M., Steury, T., and Sears, M. 2004. Temperature, growth rate,
and body size in ectotherms: fitting pieces of a life-history puz-
zle. Integr. Comp. Biol. 44: 498–509. doi:10.1093/icb/44.6.498. PMID:
21676736.
Arthington, A., Dulvy, N., Gladstone, W., and Winfield, I. 2016. Fish con-
servation in freshwater and marine realms: status, threats and man-
agement. Aquat. Conserv. 26: 838–857. doi:10.1002/aqc.2712.
Atkinson, D. 1994. Temperature and organismsize-a biological law for ec-
totherms? Advances in Ecological Research 25:1. J. Reticuloendothel
Soc. 25.
Atkinson, D., Morley, S., and Hughes, R. 2006. From cells to colonies: at
what levels of body organization does the “temperature-size rule” ap-
ply? Evol. Dev. 8: 202–214. doi:10.1111/j.1525-142X.2006.00090.x.
Audzijonyte, A., Richards, S., Stuart-Smith, R., Pecl, G., Edgar, G., Bar-
rett, N., et al. 2020. Fish body sizes change with temperature but not
all species shrink with warming. Nat. Ecol. Evol. 4:1–6.doi:10.1038/
s41559-020-1171-0.
Barneche, D., Kulbicki, M., Floeter, S., Friedlander, A., and Allen, A. 2016.
Energetic and ecological constraints on population density of reef
fishes. Proc. Royal Soc. B. 283.doi:10.1098/rspb.2015.2186.
Bartley, T., Mccann, K., Bieg, C., Cazelles, K., Granados, M., Guzzo, M.,
et al. 2019. Food web rewiring in a changing world. Nat. Ecol. Evol. 3:
1. doi:10.1038/s41559-018-0772-3.
Baudron, A., Needle, C., Rijnsdorp, A., and Marshall, C. 2013. Warming
temperatures and smaller body sizes: synchronous changes in growth
of North Sea fishes. Global Change Biol. 20.doi:10.1111/gcb.12514.
Beard, T.D., Arlinghaus, R., Cooke, S., McIntyre, P., De Silva, S., Bartley, D.,
and Cowx, I. 2011. Ecosystem approach to inland fisheries: research
needs and implementation strategies. Biol. Lett. 7: 481–483. doi:10.
1098/rsbl.2011.0046.
Beguin, J., Martino, S., Rue, H., and Cumming, S. 2012. Hierarchical
analysis of spatially autocorrelated ecological data using integrated
nested Laplace approximation. Methods Ecol. Evol. 3.doi:10.1111/j.
2041-210X.2012.00211.x.
Beverton, R. J. H. 1954. Notes on the use of theoretical models in the study
of the dynamics of exploited fish populations. Miscellaneous Contri-
bution United States Fishery Laboratory Beaufort, North Carolina.
Bivand, R., Gómez Rubio, V., and Rue, H. 2015. Spatial data analysis with
R-INLA with some extensions. J. Stat. Softw. 63: 1–31. doi:10.18637/
jss.v063.i20.
Chezik, K., Lester, N., and Venturelli, P. 2014. Fish growth and degree-
days I: selecting a base temperature for a within-population study.
Can. J. Fish. Aquat. Sci. 71: 47–55. doi:10.1139/cjfas-2013-0295.
Chu, C., Lester, N., Giacomini, H., Shuter, B., and Jackson, D. 2016. Catch-
per-unit-eort and size spectra of lake fish assemblages reflect under-
lying patterns in ecological conditions and anthropogenic activities
across regional and local scales. Can. J. Fish. Aquat. Sci. 73: 535–546.
doi:10.1139/cjfas-2015-0150.
Comte, L., and Olden, J. 2016. Evolutionary and environmental deter-
minants of freshwater fish thermal tolerance and plasticity. Global
Change Biol. 23.doi:10.1111/gcb.13427.
Crins, W., Gray, P.A., Uhlig, P.W.C., and Wester, M.C. 2009. The ecosystems
of Ontario, part 1: ecozones and ecoregions. Inventory, monitoring
and assessment, SIB TER IMA TR-01.
Dell, A., Pawar, S., and Savage, V. 2011. Systematic variation in the tem-
perature dependence of physiological and ecological traits. Proc. Natl.
Acad. Sci. USA. 108: 10591–10596. doi:10.1073/pnas.1015178108.
Dippold, D., Aloysius, N., Keitzer, S., Yen, H., Arnold, J., Daggupati, P.,
et al. 2020. Forecasting the combined eects of anticipated climate
change and agricultural conservation practices on fish recruitment
dynamics in Lake Erie. Freshwater Biol. 65.doi:10.1111/fwb.13515.
Ebener, M., Kinnunen, R., Schneeberger, P., Mohr, L., Hoyle, J., and
Peeters, P. 2008. Management of commercial fisheries for Lake White-
fish in the laurentian Great Lakes of North America. In International
governance of fisheries ecosystems: learning from the past, finding
solutions for the future. 62.
Emmrich, M., Brucet, S., Ritterbusch, D., and Mehner, T. 2011. Size spec-
tra of lake fish assemblages: responses along gradients of general en-
vironmental factors and intensity of lake-use. Freshwater Biol. 56:
2316–2333. doi:10.1111/j.1365-2427.2011.02658.x.
Feld, C., Segurado, P., and Gutiérrez-Cánovas, T. 2016. Analysing the im-
pact of multiple stressors in aquatic biomonitoring data: a ‘cookbook’
with applications in R. Sci. Total Environ. 573.doi:10.1016/j.scitotenv.
2016.06.243.
Forster, J., and Hirst, A. 2012. The temperature-size rule emerges from on-
togenetic dierences between growth and development rates. Funct.
Ecol. 26: 483–492. doi:10.2307/41428835.
Forster, J., Hirst, A., and Atkinson, D. 2012. Warming-induced reductions
in body size are greater in aquatic than terrestrial species. Proc. Natl.
Acad. Sci. USA. 109.doi:10.1073/pnas.1210460109.
Gardner, J., Peters, A., Kearney, M., Joseph, L., and Heinsohn, R. 2011.
Declining body size: a third universal response to warming? Trends
Ecol. Evol. 26.doi:10.1016/j.tree.2011.03.005.
Gutowsky, L., Giacomini, H., De Kerckhove, D., Mackereth, R., Mc-
Cormick, D., and Chu, C. 2019. Quantifying multiple pressure inter-
actions aecting populations of a recreationally and commercially
important freshwater fish. Global Change Biol. 25.doi:10.1111/gcb.
14556.
Guzzo, M., and Blanchfield, P. 2017. Climate change alters the quantity
and phenology of habitat for lake trout (Salvelinus namaycush)insmall
Boreal Shield lakes. Can. J. Fish. Aquat. Sci. 74: 871–884. doi:10.1139/
cjfas-2016-0190.
Guzzo, M., Blanchfield, P., and Rennie, M. 2017. Behavioral responses to
annual temperature variation alter the dominant energy pathway,
growth, and condition of a cold-water predator. Proc. Natl. Acad. Sci.
114.doi:10.1073/pnas.1702584114.
Hasnain, S.S., Minns, C.K., and Shuter, B.J. 2010. Key ecological temper-
ature metrics for Canadian freshwater fishes. Applied Research and
Development Branch, Ministry of Natural Resources. Sault St. Marie,
ON.
Hoefnagel, K., and Verberk, W. 2015. Is the temperature-size rule medi-
ated by oxygen in aquatic ectotherms? J. Therm. Biol. 54: 65. doi:10.
1016/j.jtherbio.2014.12.003.
Horne, C., Hirst, A., and Atkinson, D. 2015. Temperature-size responses
match latitudinal-size clines in arthropods, revealing critical dier-
ences between aquatic and terrestrial species. Ecol. Lett. 18.doi:10.
1111/ele.12413.
Huss, M., Lindmark, M., Jacobson, P., van Dorst, R., and G˚
ardmark, A.
2019. Experimental evidence of gradual size-dependent shifts in body
size and growth of fish in response to warming. Global Change Biol.
25.doi:10.1111/gcb.14637.
Jackson, M., Loewen, C., Vinebrooke, R., and Chimimba, C. 2015. Net ef-
fects of multiple stressors in freshwater ecosystems: a meta-analysis.
Global Change Biol. 22.doi:10.1111/gcb.13028.
Jacobson, P., Hansen, G., Bethke, B., and Cross, T. 2017. Disentangling the
eects of a century of eutrophication and climate warming on fresh-
water lake fish assemblages. PLoS One, 12: e0182667. doi:10.1371/
journal.pone.0182667.
Jane, S., Hansen, G., Kraemer, B., Leavitt, P., Mincer, J., North, R., et al.
2021. Widespread deoxygenation of temperate lakes. Nature, 594.
doi:10.1038/s41586-021-03550-y.
Jarvis, L., McMeans, B., Giacomini, H., and Chu, C. 2020. Species-specific
preferences drive the dierential eects of lake factors on fish pro-
duction. Can. J. Fish. Aquat. Sci. 77.doi:10.1139/cjfas-2020-0105.
Jørgensen, C., Enberg, K., Dunlop, E., Arlinghaus, R., Boukal, D., Brander,
K., et al. 2007. Managing evolving fish stocks. Science, 318: 1247–
1248. doi:10.1126/science.1148089.
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.

Canadian Science Publishing
496 Can. J. Fish. Aquat. Sci. 81: 488–496 (2024) | dx.doi.org/10.1139/cjfas-2023-0233
Kelly, N., Burness, G., McDermid, J., and Wilson, C. 2014. Ice age fish in
a warming world: minimal variation in thermal acclimation capac-
ity among lake trout (Salvelinus namaycush) populations. Conserv.
Physiol. 2: cou025. doi:10.1093/conphys/cou025.
Lindgren, F., Rue, H., and Lindström, J. 2011. An explicit link between
gaussian fields and Gaussian Markov random fields: the stochastic
partial dierential equation approach. J. R. Stat. Soc. Series B Stat.
Methodol. 73: 423–498. doi:10.1111/j.1467-9868.2011.00777.x.
Lindmark, M., Huss, M., Ohlberger, J., and G˚
ardmark, A. 2017.
Temperature-dependent body size eects determine population re-
sponses to climate warming. Ecol. Lett. 21.doi:10.1111/ele.12880.
Magnuson, J., Robertson, D., Benson, B., Wynne, R., Livingstone, D., Arai,
T., et al. 2000. Historical trends in lake and river ice cover in the
Northern Hemisphere. Science, 289: 1743–1746. doi:10.1126/science.
289.5485.1743.
O’Gorman, E., Pichler, D., Adams, G., Benstead, J., Cohen, H., Craig, N.,
et al. 2012. Impacts of warming on the structure and functioning of
aquatic communities: individual- to ecosystem-level responses. Adv.
Ecol. Res. 47: 81–176. doi:10.1016/B978-0-12-398315-2.00002-8.
O’Reilly, C.M., Sharma, S., Gray, D.K., Hampton, S.E., Read, J.S., Rowley,
R.J., et al. 2015. Rapid and highly variable warming of lake surface
waters around the globe. Geophys. Res. Lett. 42(24): 10,710–773,781.
doi:10.1002/2015GL066235.
Ohlberger, J. 2013. Climate warming and ectotherm body size——f r o m i n -
dividual physiology to community ecology. Funct. Ecol. 27.doi:10.
1111/1365-2435.12098.
Ohlberger, J., Edeline, E., Vøllestad, L., Stenseth, N.C., and Claessen,
D. 2011. Temperature-driven regime shifts in the dynamics of size-
structured populations. Am. Nat. 177: 211–223. doi:10.1086/657925.
Ontario Ministry of Natural Resources. 1984. Water quantity resources of
Ontario. Queen’s Printer for Ontario. (p. 72).
Ontario Ministry of Natural Resources. 2005. A new ecological frame-
work for recreational fisheries management in Ontario. Queen’s
Printer for Ontario. (p. 5).
Östman, Ö., Karlsson, O., Pönni, J., Kaljuste, O., Aho, T., and G˚
ardmark, A.
2014. Relative contributions of evolutionary and ecological dynamics
to body size and life-history changes of herring (Clupea harengus)inthe
Bothnian Sea. Evol. Ecol. Res. 16: 417–433.
Plumb, J., and Blanchfield, P. 2009. Performance of temperature and dis-
solved oxygen criteria to predict habitat use by lake trout (Salveli-
nus namaycush). Can. J. Fish. Aquat. Sci. 66: 2011–2023. doi:10.1139/
F09-129.
Rodrigues,T.,Chapelsky,A.,Hrenchuk,L.,Mushet,G.,Chapman,L.,
and Blanchfield, P. 2022. Behavioural responses of a cold-water ben-
thivore to loss of oxythermal habitat. Environ. Biol. Fishes, 105.
doi:10.1007/s10641-022-01335-4.
Rooney, N., and Mccann, K. 2011. Integrating food web diversity, struc-
ture and stability. Trends Ecol. Evol. 27: 40–46. doi:10.1016/j.tree.
2011.09.001.
Rue, H., Martino, S., and Chopin, N. 2009. Approximate bayesian infer-
ence for latent gaussian models by using integrated nested laplace
approximations. J. R. Stat. Soc. Series B Stat. Methodol. 71: 319–392.
doi:10.1111/j.1467-9868.2008.00700.x.
Sandstrom, S., Rawson, M., and Lester, N. 2013. Manual of instructions for
broad-scale fish community monitoring; using North American (NA1)
and Ontario small mesh (ON2) gillnets. Ontario Ministry of Natural
Resources. Peterborough, ON. Version 2013.2 35 p. +appendices.
Sheridan, J., and Bickford, D. 2011. Shrinking body size as an ecological
response to climate change. Nat. Clim. Change, 1: 401–406. doi:10.
1038/nclimate1259.
Shin, Y.J., Marie-joelle, R., Jennings, S., Field, J., and Gislason, H. 2005. Us-
ing size-based indicators to evaluate the ecosystem eects of fishing.
ICES J. Mar. Sci. 62(3):384–396. doi:10.1016/j.icesjms.2005.01.004.
Smith, D., Ramirez, S., Lix, D., and Chu, C. 2021. Ontario lakes and cli-
mate change database: a guide to projected changes in the thermal
conditions of Ontario’s inland lakes. Ontario Ministry of Northern
Development, Mines, Natural Resources and Forestry, Science and Re-
search Branch. Peterborough, ON. Climate Change Research Report
CCRR-55. 50 p. +appendices. doi:10.13140/RG.2.2.24830.72000.
Solokas, M., Feiner, Z., Al-Chokachy, R., Budy, P., Deweber, J., Servala, J.,
et al. 2023. Shrinking body size and climate warming: many fresh-
water salmonids do not follow the rule. Global Change Biol. doi:10.
1111/gcb.16626.
Spiegelhalter, D.J., Best, N., Carlin, B.P., and Linde, A. 2002. Bayesian mea-
sures of model complexity and fit (with discussion). J. R. Stat. Soc.
Series B Stat. Methodol.64: 1–34. doi:10.1111/1467-9868.02022.
Thresher, R., Koslow, J., Morison, A., and Smith, D. 2007. Depth-mediated
reversal of the eects of climate change on long-term growth rates
of exploited marine fish. Proc. Natl. Acad. Sci. USA. 104: 7461–7465.
doi:10.1073/pnas.0610546104.
Tunney, T., Mccann, K., Lester, N., and Shuter, B. 2014. Eects of dieren-
tial habitat warming on complex communities. Proc. Natl. Acad. Sci.
USA. 111: 8079–8082. doi:10.1073/pnas.1319618111.
van Rijn, I., Buba, Y., Delong, J., Kiflawi, M., and Belmaker, J. 2017. Large
but uneven reduction in fish size across species in relation to chang-
ing sea temperatures. Global Change Biol. 23.doi:10.1111/gcb.13688.
Woolway, R.I., and Merchant, C.J. 2019. Worldwide alteration of lake mix-
ing regimes in response to climate change. Nat. Geosci. 12(4): 271–
276. doi:10.1038/s41561-019-0322-x.
Zuur, A., Ieno, E., Walker, N., Saveliev, A., and Smith, G. 2009. Mixed eect
models and extensions in ecology with R. In Mixed eects models and
extensions in ecology with R. doi:10.1007/978-0-387-87458-6.
Can. J. Fish. Aquat. Sci. Downloaded from cdnsciencepub.com by Fisheries and Oceans on 07/11/24
For personal use only.