Plant-hummingbird Networks Research Articles

Elevational and Seasonal Patterns of Plant-Hummingbird Interactions in a High Tropical Mountain.

Tropical mountain ecosystems harbor diverse biological communities, making them valuable models for exploring the factors that shape ecological interactions along environmental gradients. We investigated the spatial and temporal drivers of plant-hummingbird interaction networks across three forest types (pine-oak, fir, and subalpine) along a tropical high mountain gradient in western Mexico (2400 to 3700 m.a.s.l.). We measured species abundance, diversity, morphology, and interaction frequencies. Plant diversity metrics significantly declined in the highest elevation subalpine forest, whereas hummingbird diversity remained consistent across elevations. Interaction networks were similarly nested across elevations, but they were more specialized in the subalpine forest, where lower plant species richness and higher floral abundance led to greater resource partitioning among hummingbirds. Plant-hummingbird networks were larger and less specialized during the dry season, driven by greater species diversity and abundance. Species turnover explained network variation along the elevational gradient, while interaction rewiring and the arrival of migratory hummingbirds explained changes between seasons. Phenological overlap was the most important driver of the observed variation in interaction frequencies across elevations and seasons. Flower abundance had a minor influence on interaction frequencies at low- and mid-elevation networks, and hummingbird abundance was significant for dry- and rainy-season networks. Morphological matching was significant in the low-elevation forest and in the dry season. Plant phylogenetic relatedness had negligible effects on interaction patterns, but hummingbird phylogeny influenced feeding preferences in high-elevation and rainy-season networks. Our findings highlight the role of species turnover, interaction rewiring, and phenological overlap in structuring plant-hummingbird networks, with specific effects of abundance, morphology, and phylogeny varying with elevation and season. High-elevation ecosystems play a crucial role as reservoirs of floral resources for both resident and migratory hummingbirds during resource-scarce periods, emphasizing their importance in maintaining biodiversity in tropical mountain gradients.

Consistent generalization of plant-hummingbird networks despite increasing vegetation cover across a tropical urban landscape

Human activities, particularly urbanization, profoundly impact ecosystems often resulting in biotic homogenization. Whether or not urban landscapes can sustain diverse pollinator and plant communities is an important question to be addressed. Here, we investigated the influence of urbanization on plant-hummingbird interaction networks in a large tropical city, Belo Horizonte, Brazil. We recorded 13198 legitimate interactions between seven hummingbirds and 57 plant species across 12 local networks. Urban landscapes exhibited predominantly generalized networks, maintaining this pattern across varying vegetation cover and floral resource abundance. Although some functionally specialized hummingbirds with long bills were recorded performing more specialized interactions, urban environments did not generally support specialized networks. Nevertheless, network specialization did increase with the proportion of native nectar plants, emphasizing their importance for maintaining some specialized interactions. Furthermore, we observed a positive effect of plant richness, but not of flower abundance, on hummingbird abundance, indicating that it is not only the amount of flowers, but the diversity of floral resources that may be a key factor in maintaining hummingbirds. Therefore, promoting a diverse assemblage of native plants in urban green areas is crucial for sustainable pollinator communities. Our study highlights that while a biodiverse urban landscape will require careful urban vegetation planning considering both floral resource diversity and availability, vegetation cover per se may not be sufficient to mitigate the negative impacts of urbanization. Maintaining a diverse vegetation with different life forms, flowering phenology, and especially of native plants across the urban landscape is needed to create welcoming spaces for pollinators.

Exotic and native plants play equally important roles in supporting and structuring plant-hummingbird networks within urban green spaces.

Urban gardens, despite their transformed nature, serve as invaluable microcosms for a quantitative examination of floral resource provision to urban pollinators, considering the plant's origin. Thus, knowledge has increased, emphasizing the importance of these green areas for hosting and conserving pollinator communities. However, there is a significant knowledge gap concerning the changing availability of these native and exotic floral resources over time and their impact on structuring interaction networks with specific pollinators. Over a year-long period, monthly surveys were conducted to record both native and exotic plant species visited by hummingbirds in an urban garden at Tlaxcala, Mexico. Flower visits were recorded, and the total flowers on each plant visited were tallied. Additionally, all observed hummingbirds were recorded during the transect walks, regardless of plant visits, to determine hummingbird abundance. The interactions were summarized using matrices, and network descriptors like connectance, specializacion, nestedness, and modularity were computed. Plant and hummingbird species in the core and periphery of the network were also identified. Lastly, simulations were performed to assess the network's resilience to the extinction of highly connected native and exotic plant species, including those previously situated in the network's core. We recorded 4,674 interactions between 28 plant species, and eight hummingbird species. The majority of plants showed an ornithophilic syndrome, with 20 species considered exotic. Despite asynchronous flowering, there was overlap observed across different plant species throughout the year. Exotic plants like Jacaranda mimosifolia and Nicotiana glauca produced more flowers annually than native species. The abundance of hummingbirds varied throughout the study, with Saucerottia berillyna being the most abundant species. The plant-hummingbird network displayed high connectance, indicating generalization in their interaction. Significant nestedness was observed, mainly influenced by exotic plant species. The core of the network was enriched with exotic plants, while Basilinna leucotis and Cynanthus latirostris played central roles among hummingbirds. Network resilience to species extinction remained generally high. Our findings provide valuable insights into the dynamics and structure of plant-hummingbird interactions in urban gardens, emphasizing the influence of exotic plant species and the network's resilience to perturbations. Understanding and managing the impact of exotic plants on such networks is crucial for the conservation and sustainable functioning of urban ecosystems.

Effect of plant traits and network metrics on pollen limitation in hummingbird-pollinated plants from the Brazilian Campos Rupestres

Pollen limitation (PL) can be an important estimate of the contribution of pollinators to the reproductive success of plants. PL often varies between species within the same community, and interspecific variation in PL may be related to plant traits and interaction patterns. However, few studies have investigated its drivers at the community level. In this study we assessed PL and its relationship with the dependence of plants on pollinators, floral morphology, nectar production and descriptors of the role of plant species in the interaction network in which they are embedded (linkage level, interaction strength, specialization and c or z coefficients) for 28 plant species pollinated by hummingbirds in the Brazilian Campos Rupestres. In the studied plant community, the prevalence of PL was high, but intensity was low, with 8 species (29 %) presenting some degree of PL. Nectar volume and nectar concentration were negatively related to PL, while plant species with higher dependence on pollinators for reproduction presented higher PL. We did not detect a relationship between network metrics and pollen limitation, this could reflect the low variation in species roles in the community, with most plant species being peripherals (linked to few and generalized pollinators). We demonstrated that plant traits were more important than network metrics in predicting interspecific variation in pollen limitation in the studied plant-hummingbird network. We suggest that evaluating pollen limitation within plant communities is key to uncover the drivers of pollination deficits.

Implications of dominance hierarchy on hummingbird-plant interactions in a temperate forest in Northwestern Mexico.

The structuring of plant-hummingbird networks can be explained by multiple factors, including species abundance (i.e., the neutrality hypothesis), matching of bill and flower morphology, phenological overlap, phylogenetic constraints, and feeding behavior. The importance of complementary morphology and phenological overlap on the hummingbird-plant network has been extensively studied, while the importance of hummingbird behavior has received less attention. In this work, we evaluated the relative importance of species abundance, morphological matching, and floral energy content in predicting the frequency of hummingbird-plant interactions. Then, we determined whether the hummingbird species' dominance hierarchy is associated with modules within the network. Moreover, we evaluated whether hummingbird specialization (d') is related to bill morphology (bill length and curvature) and dominance hierarchy. Finally, we determined whether generalist core hummingbird species are lees dominant in the community. We recorded plant-hummingbird interactions and behavioral dominance of hummingbird species in a temperate forest in Northwestern Mexico (El Palmito, Mexico). We measured flowers' corolla length and nectar traits and hummingbirds' weight and bill traits. We recorded 2,272 interactions among 13 hummingbird and 10 plant species. The main driver of plant-hummingbird interactions was species abundance, consistent with the neutrality interaction theory. Hummingbird specialization was related to dominance and bill length, but not to bill curvature of hummingbird species. However, generalist core hummingbird species (species that interact with many plant species) were less dominant. The frequency of interactions between hummingbirds and plants was determined by the abundance of hummingbirds and their flowers, and the dominance of hummingbird species determined the separation of the different modules and specialization. Our study suggests that abundance and feeding behavior may play an important role in North America's hummingbird-plant networks.

To rewire or not to rewire: To what extent rewiring to surviving partners can avoid extinction?

Research Highlight: Leimberger, K.G., Hadley, A.S., & Betts, M.G. (2023). Plant-hummingbird pollination networks exhibit minimal rewiring after experimental removal of a locally abundant plant species. Journal of Animal Ecology, https://doi.org/10.1111/1365-2656.13935. In this paper, Leimberger, Hadley and Betts (2023) explore the effects of removing a locally abundant plant species on plant-hummingbird pollination networks. They experimentally prevented access of hummingbirds to flowers of Heliconia tortuosa and assessed subsequent changes in the interactions between plants and hummingbirds. Their main hypothesis postulated that the loss of a highly connected species would lead to interaction rewiring and niche expansions by hummingbirds, decreasing individual, species and network specialization. However, they found that the overall structure of the plant-hummingbird networks remains mostly unaltered, with limited rewiring and minimal changes in specialization. The main contributions of this study can be summarized as (i) it adds to a limited number of manipulative studies on the capacity of species to rewire their interactions following the loss of partners, and importantly, it is the first study from the tropics and with vertebrate pollinators, for which experimental studies at appropriate scales is intrinsically more challenging; and (ii) innovates by evaluating change in specialization for the individual level, carried out through pollen sampling on the body of hummingbirds. The limited change in species interactions highlights that network stability through interaction rewiring may have been overestimated in previous studies, calling for further manipulative studies in the field. At the same time, it also indicated that even the loss of a highly abundant plant species has an overall small effect on network structure. Thus, this study contributes timely findings regarding the capacity of ecological communities to respond to species extinctions.

Mass-flowering native species are key in the structure of an urban plant-hummingbird network

Mass-flowering native species are key in the structure of an urban plant-hummingbird network

Diel niche partitioning of a plant-hummingbird network in the Atlantic forest of Brazil

Niche partitioning is an important mechanism that allows species to coexist. Within mutualistic interaction networks, diel niche partitioning, i.e., partitioning of resources throughout the day, has been neglected. We explored diel niche partitioning of a plant-hummingbird network in the Brazilian Atlantic forest for nine months. To evaluate diel patterns of hummingbird visits and nectar production, we used time-lapse cameras on focal flowers and repeated nectar volume and concentration measures, respectively. Additionally, we measured flower abundance around focal flowers and flower morphological traits. We did not observe diel partitioning for either hummingbirds or plants. Instead, hummingbirds appeared to specialize in different plant species, consistent with trophic niche partitioning, potentially resulting from competition. In contrast, plant species that co-flowered and shared hummingbird visits produced nectar during similar times, consistent with facilitation. Our focus on the fine-scale temporal pattern revealed that plants and hummingbirds appear to have different strategies for promoting co-existence.

Extinction, coextinction and colonization dynamics in plant-hummingbird networks under climate change.

Climate-driven range shifts may cause local extinctions, while the accompanying loss of biotic interactions may trigger secondary coextinctions. At the same time, climate change may facilitate colonizations from regional source pools, balancing out local species loss. At present, how these extinction-coextinction-colonization dynamics affect biological communities under climate change is poorly understood. Using 84 communities of interacting plants and hummingbirds, we simulated patterns in climate-driven extinctions, coextinctions and colonizations under future climate change scenarios. Our simulations showed clear geographic discrepancies in the communities' vulnerability to climate change. Andean communities were the least affected by future climate change, as they experienced few climate-driven extinctions and coextinctions while having the highest colonization potential. In North America and lowland South America, communities had many climate-driven extinctions and few colonization events. Meanwhile, the pattern of coextinction was highly dependent on the configuration of networks formed by interacting hummingbirds and plants. Notably, North American communities experienced proportionally fewer coextinctions than other regions because climate-driven extinctions here primarily affected species with peripheral network roles. Moreover, coextinctions generally decreased in communities where species have few overlapping interactions, that is, communities with more complementary specialized and modular networks. Together, these results highlight that we should not expect colonizations to adequately balance out local extinctions in the most vulnerable ecoregions.

The evolution, ecology, and conservation of hummingbirds and their interactions with flowering plants.

The ecological co-dependency between plants and hummingbirds is a classic example of a mutualistic interaction: hummingbirds rely on floral nectar to fuel their rapid metabolisms, and more than 7000 plant species rely on hummingbirds for pollination. However, threats to hummingbirds are mounting, with 10% of 366 species considered globally threatened and 60% in decline. Despite the important ecological implications of these population declines, no recent review has examined plant-hummingbird interactions in the wider context of their evolution, ecology, and conservation. To provide this overview, we (i) assess the extent to which plants and hummingbirds have coevolved over millions of years, (ii) examine the mechanisms underlying plant-hummingbird interaction frequencies and hummingbird specialization, (iii) explore the factors driving the decline of hummingbird populations, and (iv) map out directions for future research and conservation. We find that, despite close associations between plants and hummingbirds, acquiring evidence for coevolution (versus one-sided adaptation) is difficult because data on fitness outcomes for both partners are required. Thus, linking plant-hummingbird interactions to plant reproduction is not only a major avenue for future coevolutionary work, but also for studies of interaction networks, which rarely incorporate pollinator effectiveness. Nevertheless, over the past decade, a growing body of literature on plant-hummingbird networks suggests that hummingbirds form relationships with plants primarily based on overlapping phenologies and trait-matching between bill length and flower length. On the other hand, species-level specialization appears to depend primarily on local community context, such as hummingbird abundance and nectar availability. Finally, although hummingbirds are commonly viewed as resilient opportunists that thrive in brushy habitats, we find that range size and forest dependency are key predictors of hummingbird extinction risk. A critical direction for future research is to examine how potential stressors - such as habitat loss and fragmentation, climate change, and introduction of non-native plants - may interact to affect hummingbirds and the plants they pollinate.

Pollinator-mediated facilitation alleviates pollen limitation in a plant-hummingbird network.

Facilitation and competition among plants sharing pollinators have contrasting consequences for plant fitness. However, it is unclear whether pollinator-mediated facilitation and competition may affect pollen limitation (potential contribution of pollination to fitness) in pollination networks. Here, we investigated how pollinator sharing affects pollen limitation in a tropical hummingbird-pollinated community marked by facilitation. We employed indices describing how much a plant species potentially affects the pollination of other co-flowering species through shared pollinators (acting degree) and is affected by other co-flowering species (target degree) within the plant-hummingbird network. Since facilitation often increases pollination quantity but not necessarily quality, we expected both indices to be associated with reductions in pollen limitation estimates that depend on pollination quantity (fruit set and seed number) rather than estimates more strictly related to quality (seed weight and germination). We found that both indices were associated with reductions in pollen limitation only for seed weight and germination. Thus, facilitation occurred via qualitative estimates of pollen limitation. Our results suggest that facilitation may enhance plant fitness estimates even if quantitative components of plant fecundity are already saturated. Overall, we showed that pollinator-mediated indirect effects in a multispecies context are important drivers of plant fitness estimates with consequences for coexistence in diverse communities.

Ecological determinants of interactions as key when planning pollinator-friendly urban greening: A plant-hummingbird network example

Urban green spaces may function as habitats for pollinators. However, it is poorly understood how distinct ecological drivers determine the interactions between pollinator and plants in urban communities, and how pollinator-friendly plants may be selected based on these determinants. Here, we describe an urban hummingbird-plant interaction network in Brazil and evaluate the ecological determinants of pairwise interactions (abundance, morphology, and phenology). Moreover, we used a modeling framework to simulate species removal from networks while allowing for rewiring of interactions, to contrast the importance of native and exotic plants for the network robustness. The studied network showed low specialization and modularity, with one short-billed hummingbird species, Glittering-throated Emerald (Chionomesa fimbriata), dominating the recorded interactions. Phenological overlap was the main determinant of pairwise interaction frequency between hummingbirds and plants. Because morphological matching did not impose strong constraints in the network, simulation models including rewiring based on morphology conferred high robustness after species removal. Furthermore, as exotic plants were only a minor component in the studied network, their removal did not greatly affect network robustness. Our findings demonstrate the importance of considering phenology when planning urban greening for conservation of pollinators. Additionally, we illustrate an analytical procedure that can be applied to quantitatively assess the importance of distinct species/groups for urban interaction networks.

Forbidden links, trait matching and modularity in plant-hummingbird networks: Are specialized modules characterized by higher phenotypic floral integration?

BackgroundPlant-pollinator mutualistic networks show non-random structural properties that promote species coexistence. However, these networks show high variability in the interacting species and their connections. Mismatch between plant and pollinator attributes can prevent interactions, while trait matching can enable exclusive access, promoting pollinators’ niche partitioning and, ultimately, modularity. Thus, plants belonging to specialized modules should integrate their floral traits to optimize the pollination function. Herein, we aimed to analyze the biological processes involved in the structuring of plant-hummingbird networks by linking network morphological constraints, specialization, modularity and phenotypic floral integration.MethodsWe investigated the understory plant-hummingbird network of two adjacent habitats in the Lacandona rainforest of Mexico, one characterized by lowland rainforest and the other by savanna-like vegetation. We performed monthly censuses to record plant-hummingbird interactions for 2 years (2018–2020). We also took hummingbird bill measurements and floral and nectar measurements. We summarized the interactions in a bipartite matrix and estimated three network descriptors: connectance, complementary specialization (H2’), and nestedness. We also analyzed the modularity and average phenotypic floral integration index of each module.ResultsBoth habitats showed strong differences in the plant assemblage and network dynamics but were interconnected by the same four hummingbird species, two Hermits and two Emeralds, forming a single network of interaction. The whole network showed low levels of connectance (0.35) and high specialization (H2’ = 0.87). Flower morphologies ranged from generalized to specialized, but trait matching was an important network structurer. Modularity was associated with morphological specialization. The Hermits Phaethornis longirostris and P. striigularis each formed a module by themselves, and a third module was formed by the less-specialized Emeralds: Chlorestes candida and Amazilia tzacatl. The floral integration values were higher in specialized modules but not significantly higher than that formed by generalist species.ConclusionsOur findings suggest that biological processes derived from both trait matching and “forbidden” links, or nonmatched morphological attributes, might be important network drivers in tropical plant-hummingbird systems while morphological specialization plays a minor role in the phenotypic floral integration. The broad variety of corolla and bill shapes promoted niche partitioning, resulting in the modular organization of the assemblage according to morphological specialization. However, more research adding larger datasets of both the number of modules and pollination networks for a wider region is needed to conclude whether phenotypic floral integration increases with morphological specialization in plant-hummingbird systems.

Ecological mechanisms explaining interactions within plant-hummingbird networks: morphological matching increases towards lower latitudes.

Interactions between species are influenced by different ecological mechanisms, such as morphological matching, phenological overlap and species abundances. How these mechanisms explain interaction frequencies across environmental gradients remains poorly understood. Consequently, we also know little about the mechanisms that drive the geographical patterns in network structure, such as complementary specialization and modularity. Here, we use data on morphologies, phenologies and abundances to explain interaction frequencies between hummingbirds and plants at a large geographical scale. For 24 quantitative networks sampled throughout the Americas, we found that the tendency of species to interact with morphologically matching partners contributed to specialized and modular network structures. Morphological matching best explained interaction frequencies in networks found closer to the equator and in areas with low-temperature seasonality. When comparing the three ecological mechanisms within networks, we found that both morphological matching and phenological overlap generally outperformed abundances in the explanation of interaction frequencies. Together, these findings provide insights into the ecological mechanisms that underlie geographical patterns in resource specialization. Notably, our results highlight morphological constraints on interactions as a potential explanation for increasing resource specialization towards lower latitudes.

Temporal dynamics of the hummingbird-plant interaction network of a dry forest in Chamela, Mexico: a 30-year follow-up after two hurricanes.

BackgroundInteractions among species are a driving force of community structure. The species composition of animal-plant interaction networks can be highly dynamic on a temporal scale, even though the general network structure is usually not altered. However, few studies have examined how interaction networks change over long periods of time, particularly after extreme natural events. We analyzed herein the structure of the hummingbird-plant interaction network in a dry forest of Chamela, Mexico, comparing the structure in 1985–1986 with that in 2016–2017 following the passage of two hurricanes (category 2 Jova in 2011 and category 4 Patricia in 2015).MethodsThe fieldwork was carried out in the Chamela-Cuixmala Biosphere Reserve in Jalisco, Mexico. In the last 30 years, three severe drought events and two hurricanes have affected this region. Previously, from 1985–1986, hummingbird-plant interactions were recorded monthly for one year in the study area. Then, from 2016–2017, we replicated the sampling in the same localities. We compared the network parameters describing the plant-hummingbird interactions of each period using adjacency matrices.ResultsWe found differences in the number and identity of interacting species, especially plants. The plant species missing in 2016–2017 were either the least connected in the original network (1985–1986) or belonged to groups such as cacti, epiphytes, or trees. The new plant species incorporated in the 2016–2017 network were herbs, vines, and shrubs, or other species barely connected. These changes in the composition are consistent with reports on vegetation damage after strong hurricanes at other study sites. Conversely, all hummingbird species remained in the network, with the exception of Heliomaster constantii, which was primarily connected to a plant species absent in the 2016–2017 network. Migratory and habitat generalist species (i.e., Archilochus spp.) showed higher abundances following the disturbance events.ConclusionsMost of the parameters describing the hummingbird-plant network structure remained unchanged after 30 years, with the exception of an increase in plant robustness and hummingbird niche overlap. However, the network’s generalist core was affected by the loss of some species. Also, core plant species such as Ipomoea bracteata, Combretum farinosum, and Justicia candicans were found to be important for maintaining the hummingbird-plant interaction network. The temporal perspective of this study provides unique insights into the conservation of plant-hummingbird networks across time and extreme natural events.

Land-Use Change in a Mexican Dry Forest Promotes Species Turnover and Increases Nestedness in Plant-Hummingbird Networks: Are Exotic Plants Taking Over?

Background Despite the increasing knowledge of plant-pollinator interaction networks, the effects of human-induced disturbances on them have barely been studied. We analyzed whether land-use changes modified the structure and topology of plant-hummingbird interaction networks or promoted the integration of exotic plant species. Methods Fieldwork was carried out in two vegetation areas in Mexico: a protected tropical dry forest and nearby disturbed sites. For two years we registered hummingbird-plant interactions monthly in each area. Then, we constructed interaction matrices from these data and compared their assemblage structure. Results The conversion of original dry forest to disturbed habitats impacted some assemblage attributes of the plant-hummingbird network. In the disturbed sites, there were more plant species, mainly exotics, and one additional hummingbird species. Most network attributes remained the same except niche width and nestedness (pattern of interactions where generalists and specialists tend to interact with generalists whereas specialist-to-specialist interactions are infrequent), which were higher in the disturbed network. The generalist core in the disturbed network contained half of the core species in the conserved network. Implications for conservation Exotic plants that strongly integrated into the disturbed network may exert a large influence on network dynamics in these areas. Identifying the interacting species and their role provides valuable insights for their conservation and protection. Hummingbirds attracting native plant species have a potential for practical or ornamental use, and hummingbirds presence in human-modified landscapes not only provides positive aesthetic value to people but can additionally contribute to conserving native plants and the biodiversity associated with them.

Machine learning algorithms to infer trait‐matching and predict species interactions in ecological networks

Abstract Ecologists have long suspected that species are more likely to interact if their traits match in a particular way. For example, a pollination interaction may be more likely if the proportions of a bee's tongue fit a plant's flower shape. Empirical estimates of the importance of trait‐matching for determining species interactions, however, vary significantly among different types of ecological networks. Here, we show that ambiguity among empirical trait‐matching studies may have arisen at least in parts from using overly simple statistical models. Using simulated and real data, we contrast conventional generalized linear models (GLM) with more flexible Machine Learning (ML) models (Random Forest, Boosted Regression Trees, Deep Neural Networks, Convolutional Neural Networks, Support Vector Machines, naïve Bayes, and k‐Nearest‐Neighbor), testing their ability to predict species interactions based on traits, and infer trait combinations causally responsible for species interactions. We found that the best ML models can successfully predict species interactions in plant–pollinator networks, outperforming GLMs by a substantial margin. Our results also demonstrate that ML models can better identify the causally responsible trait‐matching combinations than GLMs. In two case studies, the best ML models successfully predicted species interactions in a global plant–pollinator database and inferred ecologically plausible trait‐matching rules for a plant–hummingbird network from Costa Rica, without any prior assumptions about the system. We conclude that flexible ML models offer many advantages over traditional regression models for understanding interaction networks. We anticipate that these results extrapolate to other ecological network types. More generally, our results highlight the potential of machine learning and artificial intelligence for inference in ecology, beyond standard tasks such as image or pattern recognition.

Including rewiring in the estimation of the robustness of mutualistic networks

Abstract Species are entangled within communities by their interactions in such a manner that their local extinction may unchain coextinction cascades and impact community dynamics and stability. Despite increasing attention, simulation models to estimate the robustness of interaction networks largely neglect the important role of interaction rewiring, that is, the ability of species to switch partners. Here we propose a new method to incorporate the potential of species to replace lost partners into a widely used coextinction model to estimate network robustness. In this model, species are allowed to rewire their interactions after initial loss of partners according to probabilities derived from well‐known mechanisms that determine mutualistic interactions, for example, trait matching, phenological overlap and abundances. To illustrate the use of this method, we analyzed a well‐sampled dataset of mutualistic plant–hummingbird interactions. We found that (a) in general, rewiring increases the estimated robustness, (b) networks are similarly robust to the loss of pollinators or plants, (c) morphological matching and phenological overlap leads to higher robustness, (d) when multiple rewiring mechanisms are combined, however, robustness increases little, and (e) rewiring tends to increase robustness more in scenarios when the loss of generalist species occurs first. Our results suggest that the same mechanisms known to drive plant–hummingbird network structure are relevant in buffering the effects of species loss via rewiring. The method proposed here can be applied to a wide range of interaction networks and is flexible to the inclusion of other variables important in determining interactions for specific systems. It also allows changes on the assumptions regarding the importance of distinct mechanisms, for instance including a hierarchical relationship, which facilitates insights into the relative importance of multiple variables influencing network disassembling. The analytical framework we offer here represents a step towards a more realistic estimation on how species loss may affect the integrity of interaction networks.

Functional diversity mediates macroecological variation in plant–hummingbird interaction networks

AbstractAimSpecies interaction networks are known to vary in structure over large spatial scales. We investigated the hypothesis that environmental factors affect interaction network structure by influencing the functional diversity of ecological communities. Notably, we expect more functionally diverse communities to form interaction networks with a higher degree of niche partitioning.Location:Americas.Time period:Current.Major taxa studied:Hummingbirds and their nectar plants.MethodsWe used a large dataset comprising 74 quantitative plant–hummingbird interaction networks distributed across the Americas, along with morphological trait data for 158 hummingbird species. First, we used a model selection approach to evaluate associations between the environment (climate, topography and insularity), species richness and hummingbird functional diversity as predictors of network structure (niche partitioning, i.e., complementary specialization and modularity). Second, we used structural equation models (SEMs) to ask whether environmental predictors and species richness affect network structure directly and/or indirectly through their influence on hummingbird functional diversity. For a subset of 28 networks, we additionally evaluated whether plant functional diversity was associated with hummingbird functional diversity and network structure.ResultsPrecipitation, insularity and plant richness, together with hummingbird functional diversity (specifically, functional dispersion), were consistently strong predictors of niche partitioning in plant–hummingbird networks. Moreover, SEMs showed that environmental predictors and plant richness affected network structure both directly and indirectly through their effects on hummingbird functional diversity. Plant functional diversity, however, was unrelated to hummingbird functional diversity and network structure.Main conclusions:We reveal the importance of hummingbird functional diversity for niche partitioning in plant–hummingbird interaction networks. The lack of support for similar effects for plant functional diversity potentially indicates that consumer functional diversity might be more important for structuring interaction networks than resource functional diversity. Changes in pollinator functional diversity are therefore likely to alter the structure of interaction networks and associated ecosystem functions.

Spatial distance and climate determine modularity in a cross‐biomes plant–hummingbird interaction network in Brazil

AbstractAimWe examined the effects of space, climate, phylogeny and species traits on module composition in a cross‐biomes plant–hummingbird network.LocationBrazil, except Amazonian region.MethodsWe compiled 31 local binary plant–hummingbird networks, combining them into one cross‐biomes metanetwork. We conducted a modularity analysis and tested the relationship between species’ module membership with traits, geographical location, climatic conditions and range sizes, employing random forest models. We fitted reduced models containing groups of related variables (climatic, spatial, phylogenetic, traits) and combinations of groups to partition the variance explained by these sets into unique and shared components.ResultsThe Brazilian cross‐biomes network was composed of 479 plant and 42 hummingbird species, and showed significant modularity. The resulting six modules conformed well to vegetation domains. Only plant traits, not hummingbird traits, differed between modules, notably plants’ growth form, corolla length, flower shape and colour. Some modules included plant species with very restricted distributions, whereas others encompassed more widespread ones. Widespread hummingbirds were the most connected, both within and between modules, whereas widespread plants were the most connected between modules. Among traits, only nectar concentration had a weak effect on among‐module connectivity.Main conclusionsClimate and spatial filters were the main determinants of module composition for hummingbirds and plants, potentially related to resource seasonality, especially for hummingbirds. Historical dispersal‐linked contingency, or environmental variations not accounted for by the explanatory factors here evaluated, could also contribute to the spatial component. Phylogeny and morphological traits had no unique effects on the assignment of species to modules. Widespread species showed higher within‐ and/or among‐module connectivity, indicating their key role connecting biomes, and, in the case of hummingbirds, communities within biomes. Our results indicate that biogeography and climate not only determine the variation of modularity in local plant–animal networks, as previously shown, but also affect the cross‐biomes network structure.