Functional Identity Drives Directional Shifts in Freshwater Plankton Biomass–Size Relationships

ABSTRACT

Body size is a master variable in ecology that can predict ecosystem‐level processes and reflect environmental conditions across diverse ecosystems. In freshwater ecosystems, the relationship between body size and biomass provides a powerful tool for characterizing community structure and trophic transfer efficiency. While these relationships are well established at the whole‐community level, responses within plankton functional groups with distinct ecological characteristics have yet to be explored. In this study, we analysed plankton biomass‐size relationships across > 550 Canadian lakes. For phytoplankton, we found that mixotrophs exhibited positive biomass‐size slopes, while autotrophs exhibited negative slopes. For zooplankton, herbivores exhibited relatively flat slopes, while non‐herbivores exhibited negative slopes. Biomass‐size relationships also differed among lake trophic states, but only for phytoplankton functional groups. This work should improve predictions and assessments of freshwater ecosystem functioning and food web pathways, based on our observation of directional shifts in biomass‐size relationships across functional groups and trophic states.

Normalized biomass size spectra (NBSS) for four plankton functional groups across 588 Canadian lakes. Colours correspond to each functional group; points represent the mean normalized biomass for each size bin aggregated across all lakes, ranging from 0.1 to 3.5 μm. The black dashed line represents a linear fit to pooled community‐level NBSS values (all taxa and lakes combined), illustrating the overall biomass‐size scaling relationship.

1. Introduction

Body size is a fundamental ecological trait that influences organismal physiology, demography, and trophic interactions across terrestrial and aquatic ecosystems (Brose et al. 2006; Damuth 1981; Yodzis and Innes 1992). Because metabolic rates and life‐history processes scale predictably with body size, variation in size has implications for species interactions and energy transfer through food webs (Brose et al. 2006; Brown et al. 2004; Woodward et al. 2005). Organismal body size also typically responds in predictable ways to environmental conditions such as temperature, latitude, and elevation, reflecting both physiological constraints and ecological trade‐offs (Blanckenhorn and Demont 2004; Merckx et al. 2018; Sheridan and Bickford 2011). However, exceptions to these patterns exist, and size responses can vary across taxa, habitats, and life‐history strategies (Daufresne et al. 2009; Millien et al. 2006; Naya et al. 2017). Across ecological scales, organismal body size provides a powerful predictive framework for community and ecosystem ecologists seeking to understand responses to environmental change.

In freshwater ecosystems, organismal body size is often related to ecosystem functioning. Body size is related to a suite of functions, including metabolic demand, growth rate, trophic role, susceptibility to predation, and competitive ability (Brown et al. 2004; Kiørboe 1993). In lakes, plankton communities are particularly informative of ecosystem functioning because they form the foundation of the food web and mediate energy transfer and nutrient cycling. Additionally, plankton span several orders of magnitude in size, creating strong size‐structured patterns that influence community organization and ecosystem processes (Brose et al. 2006; Hansen et al. 1994). Because biomass and energy transfer are tightly linked to body size, quantifying size distributions provides a mechanistic framework for understanding plankton dynamics and their responses to environmental variation (Gaedke 1992; Platt and Denman 1977; Sprules and Munawar 1986).

Several quantitative methods have been used to characterize plankton community size structure. Many studies have used simple metrics such as community mean or median size to summarize shifts along environmental or lake trophic gradients (Peters 1983; Woodward et al. 2005). Other approaches have emphasized community biomass, examining its distribution across size classes or broad trophic levels (e.g., zooplankton and phytoplankton) (Kerr and Dickie 2001; Sprules and Munawar 1986), without quantifying how biomass scales with organismal size. While size and biomass metrics characterize community structure, they do not explicitly integrate information on both organismal size and biomass. To simultaneously capture both aspects of plankton communities, normalized biomass size spectra (NBSS) may be used. NBSS approaches specifically relate log‐transformed biomass to log body size in order to estimate how biomass is related to size classes (Finlay et al. 2007; Gaedke 1992; Platt and Denman 1977). This approach provides a mechanistic framework for comparing plankton communities across lakes and trophic levels.

To date, biomass‐size relationships have only been explored at the community‐level, often integrating across multiple trophic levels or taxonomic groups such as bacteria, phytoplankton, zooplankton, and fish (Finlay et al. 2007; Gaedke 1992; Kerr and Dickie 2001; Sprules and Munawar 1986). However, biomass‐size relationships have not yet been examined within functional groups with important ecological differences (though see Gaedke (1992) who compares herbivores with responses of all zooplankton). Within lake food webs, phytoplankton and zooplankton are made up of diverse functional feeding groups that differ in how they acquire resources. For example, autotrophs rely exclusively on photosynthesis, whereas mixotrophs use a combination of photosynthesis and heterotrophic phagotrophy of bacterial prey, depending on environmental conditions (Flynn et al. 2019; Mitra et al. 2024; Stoecker et al. 2017). These contrasting nutritional strategies should impose distinct physiological constraints and energetic trade‐offs that may alter how biomass is distributed across size classes within each group (Katechakis et al. 2005; Mitra et al. 2024). While mixotrophic plankton are not categorically larger than autotrophs, the energetic return from phagotrophy increases with cell size, as larger organisms can handle larger prey and allocate greater cellular resources to phagotrophic machinery (Andersen et al. 2015; Katechakis et al. 2005). Additionally, maintaining the cellular machinery for phagotrophy, whether actively expressed or not, may impose trade‐offs with cell division and body size (Mitra et al. 2024; Andersen et al. 2015). These effects may be amplified when phagotrophy is realized, as larger body size facilitates handling of bacterial prey (Girard et al. 2025). This size‐dependent advantage should shift biomass accumulation toward larger mixotroph size classes relative to obligate autotrophs, even if small mixotrophs are abundant. Similarly, herbivorous zooplankton are constrained by the size of available phytoplankton prey, as filter‐feeding efficiency scales with body size only within the range of ingestible prey (Burns 1969; Brooks and Dodson 1965). This tight trophic coupling to phytoplankton size structure limits the energetic advantage of large body sizes, potentially concentrating herbivore biomass in small‐to‐intermediate size classes. In contrast, non‐herbivorous zooplankton rely more heavily on non‐phytoplankton prey, where larger body size confers direct advantages through improved prey capture, handling efficiency, and access to a wider range of prey sizes (Brooks and Dodson 1965). As a result, aggregating functional groups may obscure important differences in size structure and energy allocation, highlighting the need to evaluate biomass–size relationships within functional groups rather than at the whole‐community level alone.

Lake trophic status is key predictor of plankton community size structure, as nutrient concentrations affect community composition and trophic interactions. Classic examples of shifting plankton size structure in lakes suggest that phytoplankton and zooplankton size structure respond differently to increasing nutrient concentrations. In eutrophic lakes, high nutrient concentrations are commonly associated with the increased prevalence of larger, filamentous phytoplankton taxa (e.g., cyanobacteria), reflecting their competitive advantage under high nutrient and light conditions (Downing et al. 2001; Huisman et al. 1999). These compositional shifts in phytoplankton community are typically associated with reductions in zooplankton body size, as large‐bodied grazers are disadvantaged by poor food quality, mechanical interference, and the production of toxins, which altogether favour zooplankton communities dominated by smaller, less efficient consumers (Gliwicz and Lampert 1990; Haney 1987; Müller‐Navarra et al. 2000). Empirical studies across lakes have documented consistent increases in cyanobacteria and declines in large cladocerans along eutrophication gradients (Jeppesen et al. 2000; McCauley et al. 1999), suggesting that nutrient enrichment can alter the distribution of biomass across organismal size in both producers and consumers. While our primary objective is to characterize biomass‐size relationships across plankton functional groups, examining how size structure varies with nutrient concentrations provides ecological context for interpreting size‐based patterns across lakes with differing trophic states.

Despite the development of increasingly detailed trait‐based and taxon‐resolved approaches to plankton communities, NBSS approaches remain a powerful framework for quantifying how biomass scales with organism size at the community level. By relating log‐transformed biomass to body size, NBSS slopes provide a concise summary of biomass–size relationships that can be compared consistently across lakes, trophic states, and functional groups (Gaedke 1992; Platt and Denman 1977). Importantly, this approach makes it possible to quantify changes in biomass–size scaling, making it well suited for assessing whether functional groups with contrasting nutritional strategies exhibit distinct biomass–size relationships across environmental gradients. While NBSS theory is most often applied at the whole‐community level, where size‐dependent energy transfer is assumed to structure biomass distributions, deviations from idealized scaling are expected when communities are partitioned into functional components. In such cases, NBSS slopes should be interpreted as comparative descriptors of size structure rather than strict tests of size‐spectrum theory.

We quantified NBSS slopes across > 550 Canadian lakes, explicitly taking into account plankton resource acquisition (or feeding) traits. Phytoplankton were classified as autotrophs or mixotrophs based on their known capacity for phagotrophy, following established approaches in lake ecology that define mixotrophs by demonstrated physiological potential for bacterivory rather than realized feeding behaviour (Salmaso and Padisák 2007; Stoecker et al. 2017; Mitra et al. 2024), which is difficult to quantify directly in diverse natural communities. Zooplankton were classified as herbivores or non‐herbivores to capture broad differences in trophic coupling to phytoplankton production, as this distinction most directly reflects how energy enters the zooplankton community across size classes and parallels the autotroph‐mixotroph distinction made for phytoplankton. We focused on resource acquisition strategy as the functional trait of interest in this study because we expected differences in how organisms acquire energy and nutrients to impose distinct physiological constraints on cell size and biomass allocation. Specifically, we asked: (1) How does the relationship between biomass and body size differ among functional feeding groups? and (2) How do plankton size distributions vary across lake trophic state? We hypothesized that biomass‐size relationships would differ among functional groups based on resource acquisition strategy. For phytoplankton, mixotrophs were expected to show a less negative slope than autotrophs because the energetic return from phagotrophy increases with cell size. Larger cells can encounter prey at higher rates and allocate greater cellular resources to phagotrophic structures, thereby shifting biomass accumulation toward larger size classes within the mixotroph group relative to obligate autotrophs, which primarily face negative size‐scaling constraints through nutrient diffusion and light harvesting. For zooplankton, herbivores were expected to show a steeper negative slope than non‐herbivores because filter‐feeding constrains optimal body size to match phytoplankton size‐structure, while predatory and omnivorous taxa benefit energetically from larger body size. Therefore, the functional groups where larger body size confers greater energetic returns through resource acquisition are predicted to exhibit less negative biomass‐size relationships relative to groups where size‐dependent resource acquisition advantages are constrained by prey availability or trophic coupling.

2. Methods

2.1. Lake Pulse Data

We used plankton size data from 588 lakes across Canada collected as part of the NSERC Canadian Lake Pulse Network (Huot et al. 2019; NSERC Canadian Lake Pulse Network 2021). Each lake was sampled one time during the 2017, 2018 and 2019 summer sampling campaigns, as all lakes could not feasibly be sampled in the same year. Phytoplankton were sampled with a tube sampler across the photic zone (estimated in situ as 2 times the Secchi depth) to a maximum of 2 m depth in the pelagic zone and fixed with Lugol's iodine solution. Phytoplankton taxonomy and enumeration were performed by an environmental consulting firm (Dr. Michael Agbeti from Bio‐Limno Research). Subsamples were quantified following the Utermohl sedimentation method (Lund et al. 1958) and biovolume (hereafter referred to as biomass) was calculated from abundance and specific volume estimates, based on geometric solids (Rott 1981). Zooplankton were collected across the entire water column using a 100 μm mesh Wisconsin plankton net (Paquette et al. 2021). Samples were preserved with 70% ethanol in the field and later identified to species, enumerated, and converted to biomass using standard length‐weight relationships by BSA Environmental Services (Ohio, U.S.A.).

Rare phytoplankton and zooplankton taxa, defined as genera that comprised either < 1% of the total biomass in each lake or that were present in < 5 lakes, were removed from the dataset. Given that the phytoplankton dataset only included individual cell sizes for filamentous colonies (n = 17 genera), we estimated mean colony sizes based on data from four publicly available phytoplankton datasets for subsequent analyses (Table S1).

Our functional group classifications represent broad differences in resource acquisition strategy rather than discrete or realized trophic roles. For phytoplankton, the autotroph and potential mixotroph (hereafter referred to as mixotroph) groups reflect known physiological capacity for photosynthesis and both phagotrophy and autotrophy, respectively. Realized mixotrophic behaviour is context‐dependent and difficult to quantify directly across taxonomically diverse communities. For zooplankton, herbivores and non‐herbivores reflect dominant feeding modes with respect to trophic coupling to phytoplankton; these categories should be interpreted as broad functional groupings rather than discrete trophic levels. The non‐herbivore group encompasses taxa spanning a range of trophic strategies, including omnivory and predation, defined here by their reliance on predator–prey interactions with other metazoa rather than direct grazing on phytoplankton or microbial prey. Conversely, herbivores are defined by their dominant trophic coupling to phytoplankton, encompassing taxa that feed primarily through direct grazing or filter feeding on phytoplankton, protozoa, and bacteria. Table S2 lists all plankton genera and their corresponding size ranges and functional group classification for our study.

2.2. Size Spectra Analysis

We calculated the normalized biomass across plankton functional feeding groups and lake trophic states following Sprules and Munawar (1986), Gaedke (1992), and Finlay et al. (2007). Plankton length measurements, used as the size metric in this analysis, were log₁₀‐transformed, and all genera were grouped into size bins of 0.1 log₁₀ units resulting in 33 size classes spanning 1 μm to 2.97 × 10³ μm. For each lake and functional group, we summed the total plankton biomass within each size bin. To normalize biomass to linear size, each log‐scaled bin was converted to its lower and upper boundaries in linear size (μm); the linear bin width (Δsize) being the difference between these boundaries. Normalized biomasses (μg L⁻¹ μm⁻¹) were then calculated as the total biomass in each bin divided by the linear bin width. We regressed the log₁₀ of the normalized biomass against the log₁₀ of the size bin midpoint. The resulting NBSS slope describes how biomass is distributed across the size spectrum, where more negative slopes indicate communities dominated by small‐bodied plankton and flatter slopes indicate a more even contribution of small and large sizes to total biomass. A flowchart summarizing the NBSS slope calculation across plankton functional groups and lake trophic states is provided in the supplement (Figure S1). NBSS slopes were also calculated for the full plankton community (phytoplankton and crustacean zooplankton) following the classic NBSS application (Gaedke 1992; Kerr and Dickie 2001; Sprules and Munawar 1986).

2.3. Statistical Analyses

First, we calculated NBSS slopes for the overall plankton community. We used a linear model followed by Type III ANOVA to determine whether plankton community slopes differed across trophic state. When significant effects were detected, pairwise post hoc comparisons were conducted using estimated marginal means (emmeans package; Lenth and Piaskowski 2026), with p‐values adjusted for multiple comparisons using the Tukey method. Model assumptions were evaluated based on residual diagnostics rather than raw data distributions. The Shapiro–Wilk test indicated a deviation from normality (p < 0.05); however, given the large sample size (n = 588 lakes), this test has high sensitivity to minor departures, and visual inspection of Q‐Q plots confirmed no substantial deviations in the tails. Levene's test indicated homogenous variance across groups (p = 0.12).

Next, we calculated NBSS slopes for each plankton functional group. To determine whether NBSS slopes differed across trophic states and functional groups, we used linear models including trophic state, functional group, and their interaction. Levene's test indicated heterogeneous variance across groups (p < 0.05), so we report p‐values based on HC3 heteroscedasticity‐consistent robust standard errors to ensure valid inference. We used a Type III ANOVA to test for significant main effects and interactions between trophic state and functional group. Pairwise post hoc comparisons were conducted using estimated marginal means (Lenth and Piaskowski 2026).

Lake trophic state was classified using the Carlson trophic state index (TSI) by integrating Secchi depth, chlorophyll‐a, and total phosphorus data into a combined metric. First, three individual indices were calculated as follows following Carlson (1977):

$$\mathrm{TS}{\mathrm{I}}_{\mathrm{SD}}=60-14.41\ln \left(\mathrm{SD}\right)$$

$$\mathrm{TS}{\mathrm{I}}_{\mathrm{Chl}}=9.81\ln \left(\mathrm{Chl}\right)+30.6$$

$$\mathrm{TS}{\mathrm{I}}_{\mathrm{TP}}=14.42\ln \left(\mathrm{TP}\right)+4.15$$

where SD is Secchi disk depth (m), Chl is chlorophyll‐a concentration (μg/L), and TP is total phosphorus concentration (μg/L). These individual indices were then summed together and the mean TSI index was used to classify trophic states for each lake. Specifically, oligotrophic lakes were identified as those with a mean TSI value < 40, mesotrophic lakes had a mean TSI value between 40 and 50, and eutrophic lakes had a mean TSI value > 50. Note that hypereutrophic lakes (> 70) were grouped in with eutrophic lakes due to low sample size.

While statistical analyses were conducted on the mean NBSS slopes aggregated across all lakes within a given lake trophic state and functional group, we also visualized patterns in NBSS slopes for each individual lake in our dataset (n = 588) and aggregated overall plankton community NBSS slopes following classic NBSS applications. These additional visualizations were performed to better characterize variability in biomass‐size relationships. Therefore, we calculated the percentage of positive NBSS slopes for each lake across trophic state and functional groups, as well as NBSS slopes for the overall plankton community across trophic state.

3. Results

We found greater differences in plankton mean size across functional groups than among lake trophic states (Figure 1). Mixotrophs were the smallest functional group, with median sizes of 15.0 μm (interquartile range (IQR) = 5.8 μm) across all lakes while autotrophs had median sizes of 31.1 μm (IQR = 17.1 μm). For the zooplankton functional groups, herbivores were larger than non‐herbivores, with a median size of 0.70 mm (IQR = 0.64 mm) relative to the median size of 0.62 mm (IQR = 0.23 mm) for non‐herbivores. Across trophic states, autotrophs and mixotrophs were largest in eutrophic lakes (37.6 μm and 16.0 μm, respectively). Both zooplankton functional groups were largest in oligotrophic lakes (herbivores 0.81 mm; non‐herbivores 0.64 mm).

FIGURE 1.

Kernel frequency distributions of plankton size across functional groups and lake trophic state. Phytoplankton functional groups are represented as green colours (left column); zooplankton functional groups are represented as orange colours (right column). Dashed lines correspond to the median size of each functional group across lake trophic states.

Aggregated across all sampled lakes and functional groups, we found that normalized biomass generally decreased with mean size (dashed black line on Figure 2); however, the strength and direction of the size relationship differed across plankton functional groups (Figure 2). Autotrophs had the steepest negative slope (−0.51, p < 0.001; Table S3), indicating that normalized biomass decreased strongly with increasing size. Non‐herbivores had a relatively flat, non‐significant slope (−0.009, p = 0.96; Table S3). Herbivores had a positive slope of 0.43 (p = 0.08), and mixotrophs had the strongest positive slope of (1.13, p < 0.001; Table S3; Figure 2). Across trophic states, autotroph slopes were steepest in mesotrophic lakes (−0.45, p < 0.001; Table S4), mixotroph slopes were steepest in oligotrophic lakes (1.15, p < 0.001), herbivore slopes were steepest in eutrophic lakes (0.81, p < 0.001), and non‐herbivore slopes were steepest in oligotrophic lakes (−0.61, p < 0.001; Table S4; Figure 3 and Figure S3). Additionally, non‐herbivore NBSS slopes switched direction with increasing nutrient concentration, as slopes were negative in oligotrophic lakes and positive in mesotrophic and eutrophic lakes (Figure 3; Figure S3; Table S4). We note that differences in slopes between Figures 2 and 3 reflect differences in statistical aggregation as well as the influence of sparsely distributed large‐bodied taxa, which can exert disproportionate leverage on biomass‐size relationships within individual lakes. Specifically, Figure 2 shows a pooled community regression across all observations, whereas Figure 3 summarizes mean within‐lake NBSS slopes, where each lake contributes equally regardless of internal size‐class structure.

FIGURE 2.

Normalized biomass size spectra (NBSS) for four plankton functional groups across 588 Canadian lakes. Colours correspond to each functional group; points represent the mean normalized biomass for each size bin (aggregated across all lakes), ranging from 0.1–3.5 μm. The black dashed line represents a linear fit to pooled community‐level NBSS values (all taxa and lakes combined), illustrating the overall biomass‐size scaling relationship. Note that both axes are log‐transformed.

FIGURE 3.

Mean normalized biomass size spectra (NBSS) slope for each functional group across 588 Canadian lakes. Horizontal dashed line represents a slope of zero, where both large and small plankton contribute to the total biomass. Points above the dashed line indicate positive slopes; points below the dashed line indicate negative slopes. Colours correspond to lake trophic state within each functional group. Letters indicate significant interactions among group slopes.

We found no differences in NBSS slope across trophic state for the full plankton community (F _2,645 = 0.01, p = 0.99, Figure S4), suggesting that lake productivity was not associated with variation in overall community size structure. However, functional group NBSS slopes varied significantly across trophic state (F _2,1110 = 40.71, p < 0.001), functional group (F _3,1110 = 60.30, p < 0.001), and their interaction (F _6,1110 = 9.23, p < 0.001). Heteroscedasticity‐consistent (HC3)‐robust standard errors were used to account for heterogeneity of variance across groups (Table S5). Estimated marginal means indicated that NBSS slopes varied across trophic states within functional groups. Among autotrophs, slopes increased from oligotrophic to eutrophic lakes, with eutrophic lakes showing significantly higher slopes than oligotrophic (p < 0.001) and mesotrophic lakes (p < 0.001; Figure 3 and Table 1). In contrast, mixotroph slopes were consistently positive and did not differ significantly across trophic states (all p > 0.05; Table 1). Herbivores and non‐herbivores showed no significant differences in slopes across trophic states (all p > 0.05; Table 1) and exhibited negative slopes (Figure 3). Pairwise comparisons among trophic states within each functional group indicated that differences were primarily driven by autotrophs, whereas other functional groups showed no consistent differences among trophic states (Table 1).

TABLE 1.

Tukey‐adjusted pairwise comparisons of estimated marginal NBSS slopes among trophic states within each functional group.

Contrast	Functional group	Estimate	SE	p
Oligotrophic–Mesotrophic	Autotroph	−0.112	0.102	0.52
Oligotrophic–Eutrophic	Autotroph	−0.904	0.108	< 0.001
Mesotrophic–Eutrophic	Autotroph	−0.792	0.106	< 0.001
Oligotrophic–Mesotrophic	Mixotroph	0.103	0.128	0.70
Oligotrophic–Eutrophic	Mixotroph	0.280	0.162	0.19
Mesotrophic–Eutrophic	Mixotroph	0.177	0.155	0.49
Oligotrophic–Mesotrophic	Herbivore	0.189	0.227	0.69
Oligotrophic–Eutrophic	Herbivore	−0.138	0.216	0.80
Mesotrophic–Eutrophic	Herbivore	−0.326	0.213	0.28
Oligotrophic–Mesotrophic	Non‐herbivore	−0.147	0.227	0.79
Oligotrophic–Eutrophic	Non‐herbivore	0.121	0.219	0.85
Mesotrophic–Eutrophic	Non‐herbivore	0.267	0.208	0.41

Note: Estimates represent the differences in slopes between trophic states (e.g., oligotrophic minus mesotrophic), with negative values indicating lower slopes in the first listed trophic state from the Contrast column. Standard errors (SE) quantify uncertainty around the estimated differences. p‐values are adjusted for multiple comparisons within each functional group. Comparisons are shown separately for autotrophs, mixotrophs, herbivores, and non‐herbivores. Significant p‐values are shown in bold.

The percentage of positive NBSS slopes for the overall plankton community (ignoring functional groups) was relatively low, ranging from 5.2% to 19.5%, with the percentage of positive slopes increasing with lake trophy (Figure S5). However, the range of positive slopes increased substantially when calculating slopes for functional groups within phytoplankton and zooplankton communities. We found that the percentage of positive NBSS slopes differed across lake trophic state and plankton functional group, ranging from 21.2% to 91.2% of lakes (Figure 4). Of the 588 lakes in this study, 91.2% of mixotroph slopes in oligotrophic lakes were positive, whereas non‐herbivores in eutrophic lakes had the greatest percentage of negative slopes (78.8%; Figure 4). Across all trophic states, mixotrophs consistently had the greatest percentage of positive slopes, averaging 87.7% positive slopes (Figure 4). Conversely, autotrophs, herbivores, and non‐herbivores had comparatively low percentages of positive slopes, averaging 40.9%, 40.5%, and 27.1% across trophic states, respectively. Across all functional groups, eutrophic lakes had the greatest percentage of positive slopes (57.9%), whereas oligotrophic lakes had the greatest percentage of negative slopes (50.7%; Figure 4).

FIGURE 4.

Normalized biomass size spectra (NBSS) slopes for each lake (n = 588) across functional groups (colours) and lake trophic state (panels). Dashed horizontal lines indicate flat size spectrum slopes; percentages correspond to the percent positive slopes for each functional group and trophic state. Positive slopes indicate greater abundance of large plankton within the community, whereas negative slopes indicate greater abundance of small plankton.

Functional group NBSS slopes tended to differ across both functional groups and lake trophic state, as indicated by differences in chlorophyll‐a (Figure 5). Autotroph and mixotroph NBSS slopes were steepest in oligotrophic lakes, with median autotroph slopes of −0.37 and mixotroph slopes of 1.13 (Figure 5). Additionally, autotroph slopes were only positive in eutrophic lakes, as they exhibited negative slopes in oligotrophic and mesotrophic lakes. Conversely, herbivore slopes were most negative in mesotrophic lakes, with a median slope of −0.41. Non‐herbivore slopes were most negative in oligotrophic lakes, with a median slope of −0.82 (Figure 5). Herbivores had the flattest slopes (closest to 0) in eutrophic lakes (−0.05), whereas autotrophs and non‐herbivores had the flattest slopes in mesotrophic lakes (−0.23 and −0.57, respectively; Figure 5). Mixotroph slopes were consistently around 1 (Figure 5). Relationships between functional group NBSS slopes and total phosphorus and Secchi depth reveal similar patterns and are reported in the Supplement (Figures S6 and S7).

FIGURE 5.

Relationship between chlorophyll‐a concentration and plankton NBSS slopes across 588 Canadian lakes. Panels correspond to phytoplankton and zooplankton functional groups; colours depict lake trophic state. Points outlined in red represent the median slope and chlorophyll‐a concentration for each functional group and trophic state. Dotted vertical lines at 0 represent slopes for which large and small plankton contribute equally to total biomass. Black lines represent the smoothed fit of the NBSS slope over the productivity gradient for each functional group.

4. Discussion

Across more than 550 Canadian lakes, we found that biomass‐size relationships differed according to plankton functional groupings and lake trophic states, highlighting the importance of assessing functional trait responses within freshwater plankton communities. Specifically, we observed positive NBSS slopes for mixotrophic phytoplankton, relatively flat slopes for herbivorous zooplankton, and negative slopes for autotrophic phytoplankton and non‐herbivorous zooplankton. Overall, mixotrophs were the functional group that was most likely to exhibit positive slopes, indicating that biomass was concentrated in larger mixotroph size classes. However, autotrophs exhibited positive slopes in eutrophic lakes. Secondarily, across trophic states, phytoplankton slopes were more positive in eutrophic lakes, while zooplankton slopes did not differ greatly across lake trophic state.

NBSS slopes reported previously for entire plankton communities in both lakes and oceans are typically negative (e.g., Gaedke 1992; Quinones et al. 2003; Sprules and Munawar 1986), as we observed for autotrophs and non‐herbivores. Such slopes are consistent with classic size‐dependent constraints on nutrient and light acquisition that favour smaller organisms (Litchman and Klausmeier 2008; Raven 1998). On the other hand, our identification of positive NBSS slopes for mixotrophs and relatively flat slopes for herbivores is not something that has been demonstrated in the literature before. For herbivores, flat slopes are consistent with size‐dependent increases in grazing and filtration efficiency, which could enhance the contribution of larger individuals to total biomass and influence energy transfer in freshwater food webs (Burns 1969; Kiørboe 2011). However, we note that the functional group NBSS slopes are best interpreted as comparative indicators of size structure rather than strict tests of whole‐community size‐spectrum theory. Because zooplankton predators are often gape‐limited, the presence of larger zooplankton can reduce trophic transfer to higher trophic levels as zooplankton vulnerability to predation is lowered (e.g., Carpenter et al. 1985; Gliwicz and Pijanowska 1989). Alternatively, when predators are capable of handling larger prey, increased zooplankton body size may enhance foraging efficiency and energetic return (Brooks 1968; Wetterer 1989). For phytoplankton, we hypothesize that positive slopes in mixotrophs arise because phagotrophy alleviates nutrient limitations that otherwise constrain large phototrophic cells, allowing larger mixotrophic taxa to partially overcome size‐dependent disadvantages of strict autotrophy, such as decreased light capture efficiency (Raven 1997; Mitra et al. 2016).

Our hypothesis that mixotrophs would have a less negative slope than autotrophs was supported. The positive NBSS slopes observed for mixotrophs, but not for autotrophs, indicates a reversal in the plankton biomass‐size relationship. Notably, this pattern was not driven by absolute size differences between functional groups: while autotroph sizes varied over a larger size range overall (Figure 1), mixotroph biomass was concentrated toward the upper end of the size spectrum and overlapped with autotroph sizes. This greater biomass retention with increasing size for mixotrophs across all trophic states likely reflects a size‐dependent advantage linked to their potential to use phagotrophy. By supplementing photosynthesis with prey ingestion, larger mixotrophs can acquire energy and nutrients more efficiently than strict autotrophs, allowing biomass to accumulate disproportionately in larger individuals (Ward et al. 2011; Millette et al. 2023). Moreover, trait‐based models predict that mixotrophic strategies can persist across a range of cell sizes and environmental conditions (Mitra et al. 2016; Ward and Follows 2016) and recent work suggests that within species, some mixotrophs may increase in size under more heterotrophic conditions (Girard et al. 2025). More generally, deviations from typical negative scaling within functional groups likely reflect the decoupling of body size from trophic position and energy transfer pathways, particularly in groups such as mixotrophs that integrate multiple nutritional strategies.

Within the zooplankton community, average NBSS slopes for herbivores and non‐herbivores were negative. Our hypothesis was not supported, as herbivore slopes were flatter, indicating a more even distribution of biomass across size classes (Figure 2). This pattern likely reflects differences in how herbivorous and non‐herbivorous zooplankton acquire energy across different size classes. For example, herbivorous zooplankton feed directly on phytoplankton, and larger herbivores are typically able to outcompete smaller herbivores, as larger organisms exploit a wider range of prey sizes and can filter greater volumes than smaller organisms, which may help sustain biomass at these larger size classes (Brooks and Dodson 1965; Burns 1969). In contrast, non‐herbivorous zooplankton are less tightly linked to phytoplankton production and may rely on more heterogeneous or lower‐quality food sources, including detrital or microbial pathways (Berggren et al. 2010; Hessen et al. 2004; Sterner and Elser 2009). These pathways may limit energy transfer to larger size classes, contributing to steeper declines in biomass with size. Together, these results suggest that feeding mode plays an important role in shaping zooplankton size–biomass relationships, even when overall NBSS slopes remain negative.

Our observation of relatively flat herbivore slopes across trophic states contrast with the stronger negative herbivore slope reported for Lake Constance by Gaedke (1992). One potential explanation is size range; Gaedke's (1992) herbivore category included ciliates, rotifers, and crustaceans, incorporating very small‐bodied organisms that would anchor biomass at the lower end of the size spectrum and steepen the slope. Because our herbivore category is restricted to larger crustacean grazers, the absence of these small size classes could shift the slope toward positive values by concentrating the fitted relationship among mid‐to‐large‐bodied taxa where the competitive advantages of larger body size are most apparent. This mechanism is supported, as the NBSS slopes for herbivores were less negative than the overall community NBSS slope in the Gaedke (1992) study.

We also examined the role of lake trophic state on NBSS slopes. Effects of lake trophy were only observed for phytoplankton, with slopes becoming increasingly positive as lake trophy increased. We observed larger phytoplankton taxa present in eutrophic lakes relative to oligotrophic and mesotrophic lakes (Figure 1), consistent with the increased dominance of larger, filamentous taxa generally observed under high nutrient and bloom conditions (Cloern 2018; Moustaka‐Gouni et al. 2006; Steinberg and Hartmann 1988). Previous work has shown that highly productive or bloom‐dominated systems often exhibit more irregular or multimodal size spectra, reflecting transient dominance by large taxa (Gaedke et al. 2004). In contrast, zooplankton biomass‐size relationships were unaffected by lake trophic state suggesting that bottom‐up factors (e.g., nutrient availability, phytoplankton composition) may be less important in structuring zooplankton communities than top‐down control (Brooks and Dodson 1965; McQueen et al. 1989). Together, these results indicate that phytoplankton community size structure is more tightly coupled to lake trophic state than that of zooplankton, consistent with previous work showing stronger bottom‐up control of phytoplankton than zooplankton across lakes (Carpenter et al. 1985; McQueen et al. 1989).

A key consideration in interpreting these biomass‐size results is that functional groups do not represent discrete trophic levels, and substantial overlap in body size across trophic roles is common in plankton communities. While the primary division between phytoplankton and zooplankton captures major trophic levels, further subdivision into functional groups introduces ecological heterogeneity that can weaken expected size‐based scaling relationships. Accordingly, the NBSS slopes presented here are best interpreted as comparative metrics of size structure within and among functional groups, rather than as direct reflections of size‐structured energy transfer.

Our study is the first to apply NBSS relationships across phytoplankton and zooplankton functional groups, considering important feeding traits, across a large spatial scale. The functional group responses provide further insight to within‐community dynamics that are not fully captured when assessing biomass‐size relationships for the community as a whole. This study demonstrates a utility of combining biomass and size metrics to characterize freshwater plankton functional groups across large spatial scales that should serve as a comparison baseline for further characterization of North Temperate dimictic lake plankton communities. Size structure and biomass data for plankton are increasingly easily obtained using flow‐through instruments and in situ video recorders. By comparing biomass‐size relationships for new lakes with our analyses, it should be possible to determine whether lakes can predominantly support phytoplankton communities dominated by heterotrophic or autotrophic activity and zooplankton communities dominated by herbivores or more omnivorous or carnivorous species without further taxonomic identification. Thus, greater insight into lake food web functioning and energy transfer pathways can be garnered from comparisons across large lake datasets.

Author Contributions

This study was conceptualized by B.E.B. and H.L.W. H.L.W. performed all analyses and wrote the first draft of the manuscript with substantial input from B.E.B. B.E.B. and Y.H. contributed to revisions and figure content. Y.H. and B.E.B. obtained funding and oversaw data collection and management.

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada, 04184.

Supporting information

Table S1: Summary table of sources for which supplemental colony lengths were obtained for n = 17 autotrophic phytoplankton taxa in the Lake Pulse dataset. The N taxa column indicates the number of phytoplankton taxa lengths that were retrieved from a given dataset.

Table S2: Phytoplankton and zooplankton taxa, functional group classifications, and cell or individual size ranges (minimum and maximum length in μm) from 588 lakes across Canada. Note that n = 17 autotrophic phytoplankton taxa lengths were replaced with colony size averages from 4 publicly available datasets (Table S1). Note that functional group classification datasets are publicly available for phytoplankton (Le Noac'h and Beisner 2025) and zooplankton (Beisner et al. 2021). Plankton taxa are classified at the genus level, with the exception of Leptodiaptomus, where L. sicilis is assigned separately due to its omnivorous tendencies.

Table S3: Linear model results for the relationship between normalized biomass and size across functional groups. Columns show the functional group, slope estimate (estimate), standard error of the slope (std. error), t‐value, model R ², and Benjamini–Hochberg adjusted p‐value. Slopes represent the change in log₁₀ normalized biomass per unit increase in log₁₀ mean size. Significant p‐values are in bold.

Table S4: Linear model results for the relationship between normalized biomass and size across functional groups and lake trophic states. Columns show the trophic state, functional group, slope estimate (Estimate), standard error of the slope (SE), t‐value, sample size (n), model R ², and Benjamini–Hochberg adjusted p‐value. Slopes represent the change in log₁₀ normalized biomass per unit increase in log₁₀ mean size. Note that the sample size here corresponds to the number of distinct 0.1 um size classes present for each functional group. Significant p‐values are in bold.

Table S5: Linear model coefficients for the interaction between lake trophic state and phytoplankton functional group on normalized biomass size spectrum (NBSS) slopes. The reference levels are oligotrophic lakes and autotrophs. Standard errors and p‐values are based on HC3 heteroscedasticity‐consistent (sandwich) estimators, applied to account for heterogeneity of variance across groups (Levene's test: F _11,1130 = 3.46, p < 0.001). Significance p‐values are in bold.

Figure S1: Flowchart depicting normalized biomass size spectra (NBSS) calculation across plankton functional groups and lake trophic states. First, phytoplankton and zooplankton taxa were classified into functional groups (autotroph, mixotroph, herbivore, and non‐herbivore). Then length measurements for each functional group were log10 transformed and assigned to size bins (n = 33). For each lake and size bin, biomass is summed. NBSS is calculated as the total summed biomass per lake and size bin divided by the linear bin width. NBSS and size bins can then be regressed on a log–log scale to obtain the biomass‐size relationships (Figure 2), where the maximum number of points for each functional group is 33.

Figure S2: Map of 588 Canadian lakes for which plankton NBSS values were calculated, Colours represent lake trophic state, as determined using chlorophyll‐a concentration (see Methods for more details).

Figure S3: Normalized biomass size spectra (NBSS) plot for lake trophic state and plankton functional groups across 588 Canadian lakes. Colours correspond to each functional group; points represent the average NBSS value for each size bin, ranging from 0.1 to 3.5 μm. Panels depict the different NBSS plots for each lake trophic state. Black dashed line represents the linear fit to pooled community‐level NBSS values (all taxa and lakes combined), illustrating the overall biomass‐size scaling relationship. Note that both axes are log‐transformed.

Figure S4: Mean normalized biomass size spectra (NBSS) slope for the overall plankton community (phytoplankton and zooplankton) across 588 Canadian lakes. Colours correspond to lake trophic state within each functional group. ANOVA results indicate no significant differences in overall plankton NBSS slopes among trophic state.

Figure S5: Normalized biomass size spectra (NBSS) slopes for each lake (n = 588) and lake trophic state (colours) for the full plankton community. Dashed horizontal lines indicate flat size spectrum slopes; percentages correspond to the percent positive slopes for each functional group and trophic state. Positive slopes indicate greater abundance of large plankton within the community, whereas negative slopes indicate greater abundance of small plankton.

Figure S6: Relationship between total phosphorus (TP) concentration and plankton NBSS slopes across 588 Canadian lakes. Panels correspond to phytoplankton and zooplankton functional groups; colours depict lake trophic state. Points outlined in red represent the median slope and TP concentration for each functional group and trophic state. Dotted vertical lines at 0 represent slopes for which large and small plankton contribute equally to total biomass. Black line represents the smoothed fit of the NBSS slope over the TP gradient for each functional group.

Figure S7: Relationship between Secchi depth and plankton NBSS slopes across 588 Canadian lakes. Panels correspond to phytoplankton and zooplankton functional groups; colours depict lake trophic state. Points outlined in red represent the median slope and Secchi depth for each functional group and trophic state. Dotted vertical lines at 0 represent slopes for which large and small plankton contribute equally to total biomass. Black line represents the smoothed fit of the NBSS slope over the Secchi depth gradient for each functional group.

Data Availability Statement

All data and code required to recreate figures and analyses in this study are available on Zenodo: https://doi.org/10.5281/zenodo.18507893 (Wander et al. 2026).