Introduction
Competitive interactions among primary producers have been widely shown to shift to facilitative when conditions become stressful, as predicted by the Stress Gradient Hypothesis (SGH; Bertness & Callaway 1994; Bruno et al. 2003; He et al. 2013). For example, in a study of alpine plant interactions, identically and globally replicated experiments demonstrated that interactions shifted from competitive at low elevations to facilitative at higher, more stressful, elevations (Callaway et al. 2002). This shift from competitive to facilitative interactions along gradients of stress has likewise been shown for arid plants and nurse shrubs (Armas et al. 2011), pitcher plants and spiders (Lim et al. 2018), and alpine herbivores (Barrio et al. 2013), among others. Similar to terrestrial elevational shifts from negative to positive interactions, species interactions among marine and limnal taxa have also been shown to shift from competitive to facilitative in 1) drier, more stressful tidal elevations (intertidal marsh plants: Bertness & Hacker 1994; barnacles and mussels: Kawai & Tokeshi 2007), and in 2) warmer, more stressful latitudes (Bennett et al. 2015; McAfee et al. 2016). Despite being well-documented in terrestrial and brackish ecosystems, the SGH has not been tested in marine forests, and instead competition is thought to be the dominant structuring mechanism (though see Bennett et al. 2015). In this study, marine forest ecosystems were used to examine influences of species facilitation, positive interactions and principles of the SGH in structuring algal communities.
The past decade has seen dramatic declines in marine forest cover (Ling et al. 2015; Krumhansl et al. 2016; Wernberg et al. 2016; Vergés et al. 2016). Foundation species provide habitat and energy to associated organisms (Dayton 1972), ultimately increasing biodiversity. Foundation species persistence is critical for understanding the dynamics of the systems they sustain. One of the most well studied, conspicuous groups of foundation species are canopy-forming brown seaweeds, belonging primarily to the orders Laminariales (kelp) and Fucales (fucoids). Many disturbed marine forests are shifting from canopy-dominated to systems dominated by turf-forming or crustose algae (Filbee-Dexter & Scheibling 2014; Filbee-Dexter & Wernberg 2018), an example of alternative stable states (Dayton & Tegner 1984). With a suite of abiotic and biotic stressors contributing to the decline of kelp forests, their recovery is thought to be prevented, in part, by competition with algal species that live in the understory (Filbee-Dexter & Scheibling 2014; Filbee-Dexter & Wernberg 2018).
Negative effects of algal turfs on canopies are well supported from long-term subtidal research programs and experimental studies (Dayton et al. 1984; Hernández -Carmona et al. 2005; Gorman & Connell 2009). Although many canopy-forming species are dominant competitors (Dayton et al. 1984), abiotic and biotic disturbance can lead to the dominance of turf algae, if those turf algae compete with the canopy (Fig. 1A). For example, the co-occurrence of El Niño-driven warm temperatures, nutrient depletion, and unusually strong storms devastated more than 500 hectares of kelp forest in southern California (Dayton & Tegner 1984). In the wake of massive loss of kelp adults and kelp recruitment failure, a new, turf-dominated community emerged, comprised of turf species and subdominant understory kelp. Algal turfs, unaffected by the El Niño disturbances, inhibited the recruitment of new canopy kelp, slowing the kelp forest recovery (Dayton et al. 1984). Although the kelp forest generally recovered after this El Niño, the kelp-dominated community never returned at its southernmost range edge (Hernández-Carmona et al. 2005). The dominance of these turf species, particularly herbivore-resistant coralline algae, is often maintained by the presence of kelp grazers (Vergés et al. 2016), forming the classic coralline algae/urchin barren alternative stable state (Filbee-Dexter & Scheibling 2014). The assumption of competitive exclusion by the turf now forms the basis of recommended restoration practices (Hernandez-Carmona et al. 2005) and even predictions for the future of canopy-forming marine ecosystems (Connell et al. 2013). For example, in the turf-as-competitors framework (Fig. 1A), as coralline turfs are expected to decline with ocean acidification, kelp abundance would be expected to increase under ocean acidification, due to the assumed release from competition and physiological effects from increasedp CO2 (Harley et al. 2012). Despite the dominance of the competition framework in the study of marine forests, many examples exist of canopy facilitation by turfs (Fig. 1B). These examples of facilitation have yet to be comprehensively incorporated into our understanding of canopy-turf dynamics or marine forest alternative stable states.
In this study, we approach competition and facilitation in marine forests within the framework of the SGH and the well-documented patterns seen in terrestrial systems. We examine the longstanding hypothesis that the effect of turf algae on canopy-forming species is competitive, using a new global dataset of marine turf-canopy interactions. We used meta-analysis to answer the following questions: a) What is the overall effect of the turf on canopy-forming seaweeds? b) Does this effect differ among turf functional groups? c) Does the effect of turfs on the canopy vary along stress gradients, i.e. depth? d) Does the effect of turfs on the canopy vary latitudinally?, and d) Does herbivory modify the effect of turfs on the canopy? We expand the number of studies previously included in a meta-analysis of kelp-turf interactions four-fold, and broaden the scope from the effect only on recruitment to the effect on all life-history stages of canopy forming species (O’Brien & Scheibling 2018). We show that the previously well-documented competitive effects of turf species on canopy species are part of a continuum of positive to negative interactions that depend predictably on stress.