Stochastic and deterministic effects on interactions between canopy and recruiting species in forest communities

Abstract

Interactions between established (canopy) and recruiting individuals (recruits) are pervasive in plant communities. Studies on recruitment in forests have mainly focused on negative density‐dependent conspecific interactions, while the outcomes of heterospecific canopy–recruit interactions have received much less attention and are generally assumed to be driven by stochastic processes. Herein, we explore the relative influence of stochastic (abundance) and deterministic (species identity and phylogenetic distance) effects on the frequency of canopy–recruit interactions and characterize the interactions in terms of their spatial consistency and effect on recruitment (depressing, neutral or enhancing). In 12 plots (50 × 50 m) of mixed pine–oak forests in southern Spain, we identified all saplings recruiting beneath 56 shrub and tree species, and in open areas not covered by woody plants. We used generalized linear mixed models to investigate the influence of stochastic and deterministic processes on the frequency of canopy–recruit interactions, on their spatial consistency and their effects on recruitment, and applied neutral null models to evaluate the spatial consistency in the occurrence of interactions across plots. Deterministic and stochastic interactions were equally common, emphasizing the prevalence of non‐neutral effects. Among the realized interactions, 36.8% enhanced recruitment, 49.05% were neutral, and 14.1% depressed recruitment. Many potential interactions (42.08%) were not observed in any study sites, presumably due to the scarcity of the interacting species. Moreover, the probability that two species formed a canopy–recruit interaction, the frequency of their interaction and the probability that the interaction had an enhancing effect on recruitment, all increased with the phylogenetic distance between the interacting species. However, the prevalence of these effects depended on the recruitment environment (heterospecific, conspecific or open). Recruitment‐enhancing interactions between heterospecifics were more consistently realized in different sites than neutral or depressing interactions. The establishment of canopy–recruit interactions (which species recruits beneath which others, and how often) is not simply determined by stochastic events. Indeed, due to their prevalence, we argue that deterministic canopy–recruit interactions are important drivers of plant community dynamics. A http://onlinelibrary.wiley.com/doi/10.1111/1365-2435.13140/suppinfo is available for this article.

Unifying facilitation and recruitment networks

“…In the communities studied by Valiente-Banuet (2011), Marcilio-Silva et al (2015) and Alcántara, Pulgar, Trøjelsgaard, Garrido, and Rey (2018), the frequency of canopy-recruit interactions was positively related with the abundance of the interacting species, so the prevalence of unobserved interactions could be a consequence of the low abundance of many species which may severely limit their chances of interacting (i.e., they are "neutral forbidden" interactions as defined by Canard et al, 2012). However, Alcántara et al (2018) found that nearly 1/3 of the unobserved interactions were consistently unobserved even though the participating species were abundant enough to interact at least by chance, what indicates that an important part of the potential interactions are impeded for some ecological reason (i.e., they are "forbidden links" as defined by Olesen et al, 2011). Among the realized interactions, interspecific interactions had more frequently neutral or enhancing than depressing effects on recruitment, while intraspecific interactions had more frequently depressing effects (but some intraspecific enhancing effects were also found).…”

“…These patterns may have discouraged the study of FNs in plant communities from many temperate and tropical environments. However, as argued above, the recruitment of one species can be facilitated by some species, depressed by others, and unaffected by others in any community (Alcántara et al, 2018). Moreover, the predominance of one type of interaction in some ecosystems must not lead to uncritically assume that all recruitment associations between plants have the same outcome.…”

“…In the communities studied by Verdú and Valiente‐Banuet (), Marcilio‐Silva et al. () and Alcántara, Pulgar, Trøjelsgaard, Garrido, and Rey (), the frequency of canopy–recruit interactions was positively related with the abundance of the interacting species, so the prevalence of unobserved interactions could be a consequence of the low abundance of many species which may severely limit their chances of interacting (i.e., they are “neutral forbidden” interactions as defined by Canard et al., ). However, Alcántara et al.…”

“…However, Alcántara et al. () found that nearly 1/3 of the unobserved interactions were consistently unobserved even though the participating species were abundant enough to interact at least by chance, what indicates that an important part of the potential interactions are impeded for some ecological reason (i.e., they are “forbidden links” as defined by Olesen et al., ). Among the realized interactions, interspecific interactions had more frequently neutral or enhancing than depressing effects on recruitment, while intraspecific interactions had more frequently depressing effects (but some intraspecific enhancing effects were also found).…”

“…Besides species abundance effects, Verdú and Valiente‐Banuet (, ) and Alcántara et al. () found that the frequency of canopy–recruit interactions increases with the phylogenetic distance between species. This phylogenetic pattern could be expected if closely related species were functionally similar so that competitive exclusion would take place, or if third interactants, such as pathogens, affected closely related species more strongly.…”