1) Functional groups are widely used to reduce complexity and generalize across ecological communities. These models assume that shared traits among species correspond to some ecological role, process, or function, and that these traits can be leveraged to generate meaningful and distinct functional groups so that intergroup trait variation exceeds intragroup variation.

2) We sought to validate the assumptions of the widely used Functional Group Model (FGM) for marine macroalgae, which groups species based on morphological complexity, by testing the predictions of the FGM for several traits assumed to correspond with morphological complexity. The FGM predicts increased resistance to disturbance and herbivory as morphological complexity (tensile strength and thallus toughness, respectively) increases. The FGM also predicts a tradeoff between complexity and growth rate. To test predictions, we measured: 1) thallus toughness (force to penetrate), 2) tensile strength (force to break) and 3) relative growth for both tropical and temperate macroalgae from different functional groups.

3) Penetration strength followed model predictions at the functional group level, though there was significant variability among species. However, the model did not predict tensile strength at any level for either tropical or temperate macroalgae. Further, relative growth did not follow predictions; rather it was highly variable among species and functional groups.

4) Synthesis. The assumptions of the FGM that differences in morphological complexity can be used to generate distinct functional groups and that intergroup trait variation outweighs intragroup variation were violated, providing strong evidence that individual species responses need to be considered. Further, violations of assumptions indicate that functional groups should not be used to predict community responses to ecological drivers and/or species contributions to ecosystem function. Our study challenges the usefulness of functional form groups for marine macroalgae and emphasizes the need for a different conceptual framework.

Model predictions

The FGM offers several measurable predictions of traits that are indicative of ecosystem function and species’ response to environmental drivers. We focus on the FGM predictions of toughness (weight to penetrate), tensile strength (weight to break), and growth (change in weight) (Fig. 1). These traits are assumed to be a measure of species’ resistance to herbivory, resistance to physical disturbance, and recovery from disturbance/role in ecological succession, respectively, as different strategy types (for review of terrestrial plant strategies, see Westoby et al. 2002). These strategies can be in turn be related to ecological functions such as productivity and food chain support (Littler and Littler 1980). The FGM model predicts that toughness and tensile strength increase with morphological complexity, while growth is the opposite (Fig. 1). Therefore, morphologically complex species should be more resistant to herbivory and disturbance and last in the line of succession while morphologically simple species are predicted to be involved in early successional stages, and vulnerable to herbivory and physical disturbance (Littler and Littler 1983b).

Study system

We tested the assumptions of the FGM using 33 total species and five functional groups from both tropical and temperate locales. We collected tropical species from a fringing coral reef lagoon. These systems are generally characterized by low productivity (Borer et al. 2013; Huston and Wolverton 2009), lower physical disturbance, and high herbivory pressure (Floeter et al. 2005; Vergés et al. 2014). Temperate species were collected from intertidal and subtidal reefs, which are generally characterized by higher productivity (Borer et al. 2013; Huston and Wolverton 2009), higher physical disturbance (Littler and Littler 1984), and lower herbivory pressure (Floeter et al. 2013; Vergés et al. 2014). As the FGM is utilized in a variety of habitats in temperate and tropical systems that vary in these environmental contexts, we collected species from different site types in both locales to test assumptions in a variety of systems.

Algal collection

Many algal species exhibit complex life cycles, where different generations can vary in morphology, ploidy, and sex (Thornber 2006; for life cycle types of collected algae, see Table 2). While little is known regarding how responses to ecological drivers vary between generations, there is some evidence these differences may be profound (Krueger-Hadfield 2020; Lubchenco and Cubit 1980; Martinez and Santelices 1998; Thornber 2006). Thus, we have identified what generation we collected, where possible, in Table 2.

Thirteen species of tropical algae representing five functional groups were haphazardly collected between January 21 and February 14, 2018 via snorkel from a common site at approximately 2-3-m depth in the patch reef zone of a fringing reef in Cook’s Bay in Mo’orea, French Polynesia (Table 2, Fig. 2B). Collected algal species included Dictyota bartayresiana, Padina boryana, Ulva intestinalis, Ulva lactuca (sheetlike), Caulerpa serrulata, Spyridia filamentosa (filamentous), Amansia rhodantha, Acanthophora spicifera (coarsely branched), putatively Gracilaria parvispora (hereafter Gracilaria parvispora), Sargassum pacificum, Turbinaria ornata (thick and leathery), Galaxaura fasciculata, and Halimeda opuntia (jointed calcareous).  For the purposes of this study, only whole, macroscopic thalli (a term defined as the body form of an alga) that were attached to the benthos and appeared healthy were collected. Approximately 20 thalli were collected per species for toughness and tensile experiments while approximately 15 thalli were collected per species for growth experiments (see below). Algae were immediately transferred to an outdoor flow-through water table. Using ambient seawater, thalli were cleaned of sediment and other organisms and processed within 24 hours of initial collection.

Twenty species of temperate algae representing four functional groups were collected between April 15 and August 1, 2018 at three sites throughout central and southern California, United States (Table 2, Fig. 2A). Collected species included Pyropia perforata, Dictyopteris undulata, Zonaria farlowii, Dictyota binghamiae, Dictyota coriacea (sheetlike), Endocladia muricata, Mastocarpus papillata, Prionitis sternbergii, Laurencia pacifica, Pterocladiella capillacea, Plocamium pacificum, Colpomenia sinuosa (coarsely branched), Egregia menziessi, Stephanocystis doica, Sargassum horneri, Sargassum palmeri, Silvetia compressa, Stephanocystis osmundaceae (thick and leathery), Bossiella orbigniana, and Calliarthron tuberculosum (jointed calcareous). Species were collected from intertidal sites in Cambria and Palos Verdes, immediately placed in coolers filled with seawater, and transported back to the University of California, Los Angeles where they were kept in indoor aquaria. Species from Santa Catalina Island were collected from a subtidal site at approximately 2-3-m depth via snorkel and immediately transported back to the University of Southern California Wrigley Institute of Environmental Science (WIES) where they were placed in outdoor flow-through water tables with ambient seawater. Replication, algal cleaning, and algal processing were performed as above.

Thallus toughness

To test the prediction that thallus toughness of species within functional groups was similar and toughness between groups increases from the simplest to the most complex algal thalli, we chose 10 whole thalli of each species. As many complex algal species exhibit apical growth, toughness was tested on blades or tissue from the middle of each thallus, where applicable, to prevent testing younger, potentially weaker, areas. To measure thallus toughness, we secured each subsample below a penetrometer so that the needle of the penetrometer rested on the thallus surface. We added weight until the penetrometer just pierced the thallus surface. This process was repeated for 10 thalli of each species.

Thallus tensile strength

To test the same predictions as above, but for thallus tensile strength, we selected 10 thalli of each species. To control for the very different algal morphologies, we used the whole algal thallus for each species, from the apex to the base. The basal end of each thalli was secured to a spring scale while just below the apical end was pulled until the thallus broke. The force (weight) required to break the thallus was used as a measure of tensile strength. Data were not used if thalli broke near where the thallus was secured to the spring scale, or where the thallus was held near the apical end. This process was repeated for each of the 10 replicate thalli per species.

Relative growth

To test model predictions for relative growth of tropical species, we conducted field experiments from January 25-February 4, 2018 using six species, with two in each of three functional groups (Table 2). Algae were spun in a salad spinner for one minute and wet-weighed into eight replicate 1-g (Acanthophora spicifera and Amansia rhodantha) or 3-g (Sargassum pacificum, Galaxaura fasciculata, Halimeda opuntia) subsamples by trimming whole thalli into appropriate weights, taking care to avoid removing apical meristems. Subsample weights varied by species due to different volume to mass ratios for each species (Mantyka and Bellwood 2007). Thus, in order to achieve similar subsample volumes among focal species, subsample weights had to be increased for more dense (S. pacificum) and calcifying algae (G. fasciculata and H. opuntia). As Turbinaria ornata is sensitive to trimming, eight replicate individuals of similar size (2.79±0.41SE-g, reproductively mature) were collected and weighed without normalizing to a standard initial weight.

We secured each replicate to the bottom of fully-enclosed, cylindrical cages (12-cm diameter x 10-cm height) constructed of hardware cloth with 1-cm openings that have been shown to have few cage artefacts and to limit herbivory in previous experiments (e.g., Fong et al. 2006, Smith et al. 2010). Cages were randomly attached to rope with at least 0.5-m spacing between replicates and secured to the benthos. Algae in cages were secured in an upright growth position and ropes were secured to the benthos with coral rubble at ~2-m depth on the same fringing reef as collection. Cages were cleaned of fouling every other day and recovered after 10 days when algae were spun and wet-weighed. Percent change in biomass was calculated as [(final weight-initial weight)/initial weight]*100 and then expressed on a per day basis.

To test the same predictions for relative growth of temperate algae, we used five species, with two in sheetlike, one in coarsely branched, and two in thick and leathery functional groups as they were abundant in the collection site near Santa Catalina Island, California. Algal thalli were spun in a salad spinner for one minute and wet weighed into eight replicate 1-g (Pterocladiella capillacea) or 2-g (Dictyopteris undulata, Zonaria farlowii, Stephanocystis doica, and Sargassum palmeri) subsamples. As before, subsample weights were adjusted to roughly match volumes. Cages were deployed from July 10-July 20, 2018, as above in ~2-m depth in a sheltered cove just west of WIES (same site as collection) and cleaned every other day. All cages were removed after 10 days, and algae spun and wet-weighed. Per cent change in biomass per day was calculated as above.

Analysis

As environmental conditions are markedly different between temperate and tropical regions (see Study system section for overview) and species were collected from each region at different points in time, we analysed the temperate and tropical species separately. For each response variable, species means were calculated as the average response (weight to penetrate, weight to break, % change in biomass) over all replicates. To compare responses among functional groups for each variable, species’ means were used to calculate functional group means. Transformations were applied where necessary to meet the assumptions of parametric statistics. If the data still did not meet parametric assumptions following transformation, nonparametric statistics were utilized (described below).

To compare species means within and between functional groups for the response variables of thallus toughness, tensile strength, and growth, data for measures from individual thalli that met assumptions for parametric statistics were analysed with a nested ANOVA, with functional group as a fixed factor and species nested within functional group. Data that did not meet parametric assumptions were analysed with a nested PERMANOVA, with functional group as a fixed factor and species nested within functional group. As there was only one species in the coarsely branched group for the temperate growth experiment, this group was omitted from nested analyses.

Significant nested analyses were followed by 1-factor ANOVAs for parametric data or 1-factor Kruskal-Wallis tests for nonparametric data to separately compare functional group and species means. For pairwise comparisons, significant ANOVAs were followed by a Tukey HSD post-hoc test and significant Kruskal-Wallis tests were followed by Wilcoxon post-hoc tests. P-values were adjusted with Bonferroni’s correction for multiple comparisons. All analyses were conducted using base functions in R Statistical Software (R Core Team 2017), except for PERMANOVAs, which were conducted using the “vegan” package for R (Oksanen et al. 2019).