Discussion
We found that interactions involving generalist plants and pollinators are more persistent in time and space. More specifically, interactions near the network core were more persistent across the five years, more persistent within seasons, and more persistent across plots. Moreover, the same interactions that were persistent within seasons also tended to be persistent across years and among plots, and that interactions persistent across years are persistent across plots. Generalist species tended to be more widespread in time and space. These patterns suggest that tolerance to environmental variation across space and time is associated with interaction generalization through the increased spatiotemporal overlap of interacting partners.
Our study conceptually and empirically links the persistence of interactions in time and space. The patterns of temporal and spatial persistence are consistent with previous studies that have found this relationship in either space or time (Aizen et al. 2012; Chacoff et al. 2018). Our study links these disparate patterns in temporal and spatial persistence of interactions with the tolerances of environmental conditions as a proposed mechanism linking them. At the species level, the connection between distributions and generalization also relates disparate concepts of niche breadth that emphasize species environmental constraints (Grinnellian niche) and their role in the community via interactions (Eltoninan niche) (Chase & Leibold 2003; Devictor et al. 2010; Gravel et al. 2019). We define generalization focusing on the number of partners of each species across a range of spatial and environmental conditions. We recognize that species’ roles themselves are dynamic and context-dependent (Devictor et al. 2010; CaraDonna & Waser 2020; Miele et al. 2020) and that generalization can be quantified in different ways to describe different aspects of species (Blüthgen et al. 2006).
As the specificity of environmental conditions and spatiotemporal distributions can be interrelated with abundance (Rabinowitz 1981; Brown 1984) it is important to consider the role of abundance in the patterns observed. Species that are widespread in time and space are likely to have larger population sizes. Large population sizes could make species of plants and pollinators less susceptible to local extinctions and thus allow greater spatiotemporal persistence allowing for greater generalization. Because spatiotemporal overlap of interacting partners and abundance are primary factors in driving interaction patterns it supports the view that interactions are strongly driven by opportunism (Waser et al. 1996; Memmott 1999). Higher abundances could also increase the detectability of species and their interactions (McCarthyet al. 2013; Chacoff et al. 2018). While the observation methods of this study aimed to minimize biases stemming from detectability by observing all flowers in plots during the survey period, disentangling sampling effects from biological processes in network studies remains an important challenge and priority (Vázquezet al. 2009; CaraDonna et al. in press).
Despite these promising results, much work remains to be done. With the increasing availability of interaction network datasets, there is an opportunity to synthesize patterns of spatiotemporal interaction persistence across studies with different environmental contexts and temporal and spatial scales. To this end, whenever possible future studies should record interaction data that are temporally and spatially explicit. Also, associated data on environmental conditions in the context of the interactions and the physiology of organisms (e.g., thermal tolerances) would allow further assessing the role of environmental variation on interaction persistence.
Generalist species are the linchpins of networks, as their presence promotes network robustness to environmental perturbations (Bascompte & Stouffer 2009). Despite their pivotal roles, common and generalist species are taken for granted and lack conservation protections that are conventionally aimed at rare species (Lindenmayer et al. 2011). However, we know that abundant, generalized species may be susceptible to decline and extinction in the face of environmental change (Wagner 2020), e.g.,Bombus dahlbomii in Patagonia, see (Morales et al. 2013). The possibility of such declines puts the stability of ecological communities in jeopardy. Therefore, conservation priorities should not overlook the pivotal roles that generalists play in supporting biodiversity across time and space.