Assessing ecological interactions in urban areas using citizen science data: Insights from hummingbird–plant meta-networks in a tropical megacity

Highlights

• citizen science data is useful to document species interactions.

• hummingbird-plant networks showed generalized interactions.

• alien plants act as hubs integrating into hummingbird-plant networks.

• network analysis reveals native keystone species for hummingbirds.

Abstract

Urbanization is one of the most intensive threats to biodiversity worldwide. The rapid sprawl of urban settings often comprises a drastic landscape transformation due to the replacement of natural vegetation by impervious surfaces. However, cities can serve as critical refuges for some native fauna, particularly for pollinators. Here we used citizen data to contrast the structure of hummingbird-plant meta-networks across different greenspaces (natural protected areas, urban parks, urban gardens and street trees areas) in a tropical megacity. We compiled hummingbird-plant visitation records in Mexico City available in two citizen science resources: iNaturalist and eBird. We first determined whether the retrieved dataset was representative to estimate network metrics by calculating sample coverage and estimating species richness in different greenspaces. Then, we characterized network structure and plant importance for network organization according to plant origin, life form and pollination syndrome. We recorded 17 hummingbirds visiting 84 plant species, encompassing a total of 742 interactions. Natural protected areas and urban parks showed a higher richness of hummingbirds and plants. All networks had low levels of connectance, specialization, and nestedness. Modularity was significant across all networks with higher values in natural protected areas and urban gardens. Native and introduced plant species showed a similar contribution to network organization. Non-ornithophilous plants were most important in natural protected areas, while tree species were most important in street trees greenspaces. Our results provide evidence of generalization of hummingbird-plant networks in urban areas. Introduced species and non-ornithophilous plants were equally important for hummingbirds, suggesting an integration of alien plants with no specialized bird pollination traits into ecological networks in urban scenarios. Promoting conservation initiatives as pollinator gardens with key native species for hummingbirds across the city could contribute to the functional connectivity and restoration of ecological interactions in cities.

Introduction

Urbanization is one of the major threats to biodiversity worldwide (Maxwell et al., 2016). The rapid sprawl and densification of urban centers has negative effects on local biotas through the novel ecological and evolutionary pressures imposed by the combination of multiple environmental factors such as habitat loss, pollution, climate change, and expansion of invasive species (Johnson and Munshi-South, 2017), which can act as ecological filters of biotic communities (Aronson et al., 2014, Spotswood et al., 2021). Yet urban scenarios span a wide range of environmental heterogeneity and greenspaces which can provide valuable refuges for native biodiversity, even rare and threatened species (Spotswood et al., 2021).

Urban greenspaces comprise private and public green areas (e.g., parks, public walkways, green roofs, gardens, forest remnants, lawns) of different sizes and forms immersed in the urban matrix (Peschardt et al., 2012) representing a critical habitat for many pollinator species (Silva et al., 2021, Theodorou et al., 2016). A growing body of evidence suggests that urbanization can generate positive, negative, as well as neutral impacts on the structure of plant-pollination interactions (Maruyama et al., 2021, Silva et al., 2021, Wenzel et al., 2020). However, the negative effects of urbanization on these interactions remain controversial (Wenzel et al., 2020). Some studies document decreases in pollinator richness and abundance along urbanization gradients (Wenzel et al., 2020), while others report a significant increase of both richness of plants and pollinators at moderate urbanization intensities (Baldock et al., 2015, Theodorou et al., 2016). Yet, most of the research has been conducted in Palearctic and Nearctic cities with insects as pollinators, while the knowledge of plant-pollinator interactions in tropical cities is substantially underrepresented (Maruyama et al., 2021, Silva et al., 2021, Wenzel et al., 2020).

Among vertebrates, hummingbirds and bats are the main pollinators of American floras (National Research Council, 2007, Stiles, 1981). Hummingbirds are an emblematic group of birds highly specialized on nectar consumption which pollinate a high proportion of plants (Bawa, 1990, Dalsgaard et al., 2021, Rodríguez-Flores et al., 2019, Stiles, 1981). The hummingbird-plant interaction represents a classic example of coevolution and has been widely used as system model to disentangle the ecological and historical factors that drives niche partitioning and community structure in highly diverse regions (Cotton, 2008, Dalsgaard et al., 2021, Stiles, 1981, Vizentin-Bugoni et al., 2014).

The study of hummingbird-plant interaction has focused mainly in natural areas (Dalsgaard et al., 2021, Sonne et al., 2019, Vizentin-Bugoni et al., 2014), while disturbed areas such as agricultural and urban settings are relatively recent. Hummingbird-plant networks from urban areas are characterized by high levels of generalization due to the dominance of a few hummingbird species, often territorial and dominant, as well as the importance of introduced plant species on network structure (Bustamante-Castillo et al., 2020, Diaz-Infante et al., 2020, Maruyama et al., 2019, Maruyama et al., 2016, Morrison and Mendenhall, 2020, Vitorino et al., 2021). Nonetheless, studies from several cities are needed to identify the patterns that determine the high degree of generalization and the relative contribution of different urban greenspaces on network structuring (Maruyama et al., 2021).

Traditionally, plant-hummingbird interaction has been studied through zoo-centric approaches (i.e., focus on the hummingbird activity recording the visited plants) including analysis of pollen loads samples (Morrison and Mendenhall, 2020, Ramírez-Burbano et al., 2017), or novel methods as DNA metabarcoding of fecal samples (Hazlehurst et al., 2021); as well as phyto-centric (i.e., focused on individual plant species to document which hummingbirds use the plants through video-cameras or focal observations (Arizmendi, 2001; Cotton, 2008; Stiles, 1981; Vizentin-Bugoni et al., 2016)). An alternative methodology that has yet to be used to record interactions is citizen science. Currently, a vast amount of information is available on social networks and citizen science platforms which could be useful to analyze the structure of ecological networks. Because of the expansion of social networks, projects that use citizen science have increased significantly in recent years (Callaghan et al., 2019a, Callaghan et al., 2020, La Sorte and Somveille, 2020). Social networks (i.e., Facebook, Twitter, Instagram) together with those focused on citizen science (i.e., eBird, Xenocanto, iNaturalist), have been of great importance in studies of species distribution, diversity patterns, geographic variation, expansion of invasive species, population trends, migration patterns and behavior (Callaghan et al., 2021, Leong and Trautwein, 2019). These studies represent a great value to urban practitioners and society due to the educational opportunities that contribute to scientific knowledge by collecting a large amount of information and making it available to the public (Callaghan et al., 2019a). One of these platforms is iNaturalist, which contains around 24 million observations of global biodiversity and has been an important source of data to answer multiple questions in urban and general ecology (Leong and Trautwein, 2019).

Assessing plant-hummingbird interaction in big cities requires a high effort to generate data at spatial and temporal scales. Particularly during the COVID-19 outbreak, because of the containment measures, many citizens started to collect and publish information about plants and animals in their neighborhoods areas. Therefore, in this study we used the information available on two popular citizen science platforms (iNaturalist and eBird) to assess the plant-hummingbird network structure in Mexico City. First, we contrasted the structure of hummingbird-plant networks across different urban greenspaces (natural protected areas, urban parks, urban gardens, and street trees areas), and then quantified plant species importance to network structure according to plant origin, life form, and pollination syndrome. Based on previous findings highlighting the impoverishment of nectar-feeding bird assemblages in urban areas as well as the dominance of alien plant species (Aronson et al., 2014, Pauw and Louw, 2012; Puga–Caballero et al., 2020), we expected to find a high generalization of hummingbird-plant networks in urban greenspaces with less vegetation cover. Moreover, due to the behavioral flexibility of hummingbirds in urban settings (Greig et al., 2017) and the inclusion of exotic plant species as feeding resources (Diaz-Infante et al., 2020, Maruyama et al., 2019, Mendonça and Anjos, 2005) we expected that alien plant species would have an important role for network structuring across urban green areas.