DISCUSSION
We recorded 2235 interactions events between 14 bird species and 13 plant species in our study area (Fig.2), suggesting that even in urban green spaces with relatively small biodiversity, interaction networks can be highly complex and dynamic. Passerine birds occupied a dominant position in the whole interaction network (Fig.2), which may be related to their wide feeding niche, foraging on a variety of fruits, and high adaptability to disturbed urban habitats. Thus, the existence of passerine birds with different body traits in cities can increase the cumulative connection of the plant-frugivore bird networks in urban ecosystems. However, compared with the random network generated by the null model, the observed network had a lower connectance (C), lower interaction diversity (H2 ), lower interaction evenness (E2 ), higher nestedness, and higher specialization ( ) (Fig.3). The results indicated that the structure of the observed network was simpler than that expected by the null model, and these shifts in the network structure can be explained by the intrinsic mathematical behavior of the null model tends to generate more connected matrices than that observed (Dormann et al., 2009; Costa et al., 2016). The higher degree of specialized interaction in the null model, which is consistent with other studies (Blüthgen et al., 2008), but these highly specialization interactions are more easily lost from the networks because extinction of one of the species may have fatal consequences for its specialist interacting partners (Sebastián-González et al., 2015). Hence, disturbed urban habitats may show more nestedness and less modular structures because the most specialization interactions have been lost due to human activities.
We found that the characteristics of the interaction network varied with season, and most interactions were detected in autumn and winter (Fig.4), which is consistent with our prediction (i). This difference could be explained by the decrease in other food sources such as insects and the arrival of wintering populations, which intensifies competition for food resources (Cruz et al., 2013; Yang et al., 2013). The number of ripe fruit species and participating frugivore birds were positively correlated with the interaction connections in different seasons, and other studies have also shown that an increase in fruit abundance significantly improves bird richness and network complexity (Ramos-Robles et al., 2016; Schneiberg et al., 2020). The connectance (C), nestedness, and interaction diversity (H2 ) for the autumn and winter networks were higher (Table1), indicating more efficient use of resources (i.e. higher number of realized links among all possible links). Interaction evenness (E2 ) was higher in spring and summer, which means that there was a more homogeneous distribution of the frequency of links in these networks; whereas, in autumn and winter a few highly frequent interactions dominated the network, particularly Cayratia japonica and C. camphora . Higher specialization in spring and summer may arise as a response of the higher trophic level to low plant diversity (He et al., 2022), which also means a high degree of niche differentiation (Blüthgen, 2010; Sebastián-González et al., 2015) and decreasing competition between different birds to facilitate species coexistence (Silva et al., 2016).
There was no significant correlation between species traits and their network roles, except for the effect of fruit color on species strength (Fig.5; Fig.6). This is contrary to the results of previous studies (Saavedra et al., 2014; Pigot et al., 2016) and is inconsistent with our prediction (ii). These differences can be explained by the following two factors: First, individual species play important roles, and their contribution to the network is much higher than the average value of species with the same traits; therefore, the differences cannot be shown in the analysis. Second, the small sample size for the weighted analyses may have caused this contrasting result (Costa et al., 2016). Plants with black fruit have greater species strength, which means they can attract more frugivorous birds. This, could be explained by the black aril of mature fruits that can create a strong visual contrast with the green background of the surrounding plants attracting birds to feed on their fruit (Duan et al., 2014; Zhang et al., 2022), and there is a significant positive correlation between the lipid nutrients and color; thus, color may be a signal of seed maturity and nutritional rewards (Schaefer et al., 2014). Finally, the evergreen fruit plants, such asC. camphora and B. javanica can provide temporary shelter when birds forage in highly disturbed urban habitats and the long-term interaction adaptation between these plants and frugivore birds has resulted in bird foraging preferences for these species.
Our results suggested that there are significant seasonal differences in the structure and characteristics of the interaction network between plants and frugivore birds in urban green space, and none of plant and bird traits were significantly related to the functional roles of species in the network structure, except for the effect of fruit color on species strength. This result may be caused by the sample size and the area limitation. Therefore, it is necessary to expand research area in the next research to further understand the impact of urbanization on the characteristics of the interaction network between plants and frugivore birds.