Shenyang (41°48′09″N, 123°25′41″E) is situated in the southern region of Northeast China and covers an area of 12,900 km². The city experiences a temperate continental monsoon climate characterized by a prolonged cold season, typically lasting from November of one year to March of the following year, with an average temperature of approximately -4°C and a minimum temperature that can drop as low as -26°C. The Hun River flows south of Shenyang, with a length of about 32 km within the urbanized area (Wang et al., 2013). The riparian zone of the Hun River serves as one of the key ecological corridors in Shenyang, encompassing a total area of approximately 23.2 km² and varying in width from 420 m to 1500 m. Its expansive width and continuous green spaces create an ideal habitat for birds.

The extent of the riparian zone along the Hun River is currently not well defined. For the purposes of this research, the boundary of the riparian zone is defined as the land area within 1,500 meters from the riverbank (Wang and Lyu, 2001). In urban areas, the riparian zone includes developed settlements and parks, while in suburban areas, the primary components consist of construction sites, new settlements, planted forests, and cultivated lands ( Figure 1 ).

A total of 2,310 birds representing 30 species from 7 orders and 17 families were recorded during the field survey ( Supplementary Table A2 ). In the bird community of the Hun River riparian zone, Passeriformes was the taxon with the highest number of species (16 species), followed by Piciformes (5 species) and Anseriformes (4 species). The Tree Sparrow (Passer montanus) had the largest number of individuals (n = 915, relative abundance = 45.68%) and was the dominant species in the community. The recorded birds included both resident and migratory species. There were 19 species of resident birds, accounting for 63.3% of the total species, which included 12 species of Passeriformes, 5 species of Piciformes, and 1 species each of Galliformes and Podicipediformes. Migratory birds, including passage migrants, summer visitors, and winter visitors, accounted for 11 species, comprising four species each of Passeriformes and Anseriformes, as well as one species each of Gruiformes, Pelecaniformes, and Podicipediformes (MacKinnon and Phillipps, 2000; Zheng, 2017).

A total of 26 bird species were recorded in the urban expansion area, representing 86.7% of the overall bird species count, which included 18 resident species and 8 migratory species. In contrast, 20 bird species were identified in the urban core area, comprising 13 resident species and 7 migratory species. The dominant taxon in both levels was Passeriformes, with the Tree Sparrow identified as the most prevalent species; however, its relative abundance was significantly lower in the suburban area compared to the urban area (14.5% vs. 59%). The birds with the second highest abundance in the urban expansion area and urban core area were the Oriental Magpie (Pica serica, n = 137) and the Chinese Spot-billed Duck (Anas zonorhyncha, n = 124), respectively. Differences in the composition of the bird communities in suburban and urban areas are detailed in Supplementary Table A2.

Table 3 presents the diversity indices of all bird species, including both resident and migratory birds, across various urbanization levels. The diversity levels of winter bird communities in the riparian zone of the Hun River exhibited significant changes along 2 levels. The mean values of the BR, BSI and BSH of the transects in the suburban area were higher than those in the urban area; however, the BA was lower in the urban expansion area compared to the urban core area. BR_Res was slightly higher in the suburban area than in the urban area, while BA_Res was considerably lower in the urban expansion area than in the urban core area, reflecting a similar trend in the overall bird population. In contrast, both BR_Mig and BA_Mig were slightly higher in the suburban area than in the urban area. Among all the bird diversity indices, all indices, except for BA and BA_Res, decreased with increasing urbanization. Conversely, BA and BA_Res increased with the level of urbanization ( Figure 5 ).

The results of the principal component analysis are presented in Table 4 . Seventeen environmental variables were condensed into seven new components. Three variables were included at the landscape scale, collectively explaining 83.42% of the original variables. Component 1 (hereafter referred to as C1) was primarily influenced by arborvitae woodland cover and impervious surface cover; a higher value of C1 indicates greater impervious coverage, while a lower value corresponds to increased woodland cover. Component 2 (C2) was predominantly determined by grass cover and building density, reflecting the degree of openness surrounding the sample. Component 3 (C3) was mainly explained by scrub and cropland cover; a higher value of C3 indicates a larger sample perimeter. The observed scale comprises four variables and accounts for 86.38% of the original variables. Component 4 (C4) was influenced by watershed area and distance to the riverbank; higher values of C4 correspond to greater distances from the water’s edge. Component 5 (C5) was determined by tree cover and impervious surfaces, reflecting canopy cover. Component 6 (C6) was similarly influenced by scrub and cropland. To facilitate subsequent analyses, we renamed the six components as follows: C1 - Build-up, C2 - Open Vegetation, C3 - Suburban Vegetation, C4 - Water Resource, C5 - Canopy Cover, and C6 - Shrub Cover. The variable Vf was not included in the principal components; for ease of analysis, we designated it as the new component C7 - Disturbance.

The 7 components mentioned above were introduced as explanatory variables in the RDA to assess their significance in influencing bird diversity. The explained variance and significance of these variables, based on the RDA results, are presented in Tables 5 – 7 . Throughout the riparian zone within the study area, environmental components explained 45.5% of the overall bird diversity, with significant impacts from disturbance and Built-up components. Environmental components explained 53.8% of the resident bird diversity, and the effects of Disturbance, Built-up, and Water Resource components on resident bird diversity metrics were also significant. In contrast, environmental components explained only 32.8% of the migratory birds. Unexpectedly, no component significantly explained the variations in migratory bird richness and abundance in the urban expansion area ( Table 5 ). In the urban expansion area (UEA), environmental components explained 70.3% of the overall birds, and the impacts from disturbance components and open vegetation components on overall bird diversity. Environmental components explained 85% of the resident birds, and disturbance, open vegetation, and Water Resource components significantly influenced the diversity of resident birds. Environmental components explained 39.9% of the total number of migratory birds. Similar to the results of the RDA for the entire riparian area, no single component significantly explained the variations in migratory bird richness and abundance within the urban expansion area, and of the seven components, the Built-up (explained variance=16.6%, p=0.118) and Canopy Cover (explained variance=9%, p=0.304) components had a stronger influence on migratory bird richness, respectively. Within this area, the Disturbance component exerted a considerable influence on both groups of bird diversity metrics. ( Table 6 ). In the urban core area (UCA), environmental components explained 73.8% of the overall bird population, while Disturbance and Build-up components had a significant effect on the significantly influenced overall bird diversity metrics. The environmental components explained 48.4% of the resident bird, with disturbance components having a notable impact on resident bird diversity. Additionally, environmental components explained 64.62% of the total number of migratory birds, with suburban vegetation component had a significant effect on significantly affecting migratory birds. As demonstrated in the UEA, the Disturbance component continues to be the environmental variable component that explains the greatest variation in bird community diversity (65.4% overall) ( Table 7 ).

9 sets of RDA results were obtained, revealing correlations between various environmental components and bird diversity metrics ( Figure 6 ). In the RDA results for the entire riparian zone, urban core area and urban extension area, the disturbance component was significantly negatively correlated with all bird diversity metrics, whereas the Built-up component showed a positive correlation with most of these metrics. The relationships between components representing vegetation and food sources and bird diversity metrics exhibited corresponding changes depending on the RDA sample selection. For the entire riparian zone, the Water Resource component was positively correlated with all overall bird diversity metrics. Additionally, the Open Vegetation, Suburban Vegetation, Shrub cover, and Canopy cover components were positively correlated with bird richness, Simpson’s diversity index, and Shannon-Wiener’s diversity index, but negatively correlated with bird abundance. The Water Resource, Open Vegetation, Suburban Vegetation, and Shrub Cover components were positively correlated with resident bird abundance, while the Canopy Cover component was negatively correlated with both resident bird richness and abundance. The Open Vegetation component was positively correlated with migratory bird abundance and richness, whereas the Suburban Vegetation component was negatively correlated with these metrics. Furthermore, the Shrub Cover and Canopy Cover components were negatively correlated with migratory bird abundance and richness. The relationships between these components and the various bird diversity indicators differed between the urban expansion area and the urban core area. However, in general, there was a positive correlation between the Water Resource component and the overall number of birds, as well as resident birds, and a consistent negative correlation with migratory birds. No uniform pattern was observed for the other components reflecting vegetation cover.

Overall, the winter birds in the Hun River riparian zone were significantly affected by the Disturbance component, exhibiting a strong negative correlation with it. The Disturbance component is explained by the volume of human traffic traversing the sample strip, which indicates the direct impact of human activities in the vicinity of the transect. The negative response of bird communities to direct disturbances supports the hypothesis that human disturbance may have a more pronounced effect than other environmental variables. This finding underscores the importance of managing human activities to conserve biodiversity in urban green spaces (Matuoka et al., 2020; Stevens and Conway, 2020).

The second highest-ranked explanation for overall bird diversity was the Built-up component. This component was primarily characterized by impervious surfaces and woodland cover, which reflect the extent of impervious land within the landscape surrounding the transect. Contrary to the findings of previous studies, significant positive correlations were observed between the built-up component and overall bird abundance. We propose that these results are linked to the major component of the urban bird community and the ways in which urban birds exploit resources in winter environments. Urban bird species, who are called “urban users” or “urban dependents”, may have utilized urban resources to a greater extent than previously assumed. For instance, species such as the Oriental Magpie and Eurasian Tree Sparrow frequently roost on buildings and readily find food in urban areas where food waste accumulates (Bressler et al., 2020; Liordos et al., 2021). Additionally, structures like dikes along rivers can create flowing water surfaces that provide feeding opportunities for waterbirds, even in winter. Based on our observations, nearly all waterbird species were present in sample strips near bridges, dikes, and artificial ponds (Kucherenko, 2023). Although the other components related to vegetation and food were not particularly significant in explaining overall bird diversity, however, the positive and negative correlations between these components and bird diversity can provide valuable insights. Although the components reflecting vegetation cover and food resources (C2, C3, C5, and C6) did not significantly impact bird diversity, they exhibited overall positive correlations with the BSI and BSH. This suggests that vegetation cover positively influences the maintenance of bird diversity (Basile et al., 2021; Miguet et al., 2016; Hughes et al., 2022; Solaro and Sarasola, 2023).

In addition to both exhibiting a negative response to disturbance components, resident and migratory birds demonstrate markedly different reactions to other environmental components. Resident birds were sensitive to the Disturbance, Water Resource, and Built-up component. In contrast, migratory birds were less likely to respond to Disturbance and Water Resource components but showed a stronger correlation with Suburban Vegetation component. It is evident that variations in the species composition of different resident bird taxa primarily account for the differences in their responses to environmental variables. Most winter resident birds—dominated by the Eurasian Tree Sparrow and Oriental Magpie—are skilled at exploiting resources in artificial environments, for instance, many urban birds roost on buildings. During resource-scarce winters, these birds may congregate even more towards the urban core area, as exemplified by the crows and starlings that inhabit major cities (Wilson et al., 2015). The Water Resource component reflects the water area and the distance to the Water Resource, negatively impacting the richness and abundance of resident birds. This finding contrasts with results from other seasonal studies (Keten et al., 2020; Zhang M. et al., 2022). The freezing of most water surfaces during winter causes frozen rivers to lose their function as sources of water and food for many resident birds. Consequently, resident birds exhibited significant positive correlations with the Build-up component and the Water Resource component. Although no environmental variable components demonstrated significant effects on migratory bird diversity indicators in the RDA, the top three environmental variable components, in terms of explanatory power, may still reflect migratory bird responses to environmental conditions to some extent. Regarding disturbance, the impact of human activity on migratory birds was not significant, possibly due to the considerable observation distance we employed. For the primary group of migratory birds, such as geese and ducks, human activities beyond 100m were not influential, as this distance exceeded the birds’ flight threshold (Chen et al., 2010; Wang et al., 2004). The negative correlation between migratory birds and Water Resource components indicates that, at the habitat scale, these birds are more likely to be found near riverbanks and in areas with water. Since the majority of migratory birds are waterbirds (36.3% Passeriformes and Anseriformes, and 9% Podicipediformes, Pelecaniformes, and Gruiformes), they rely heavily on aquatic environments than forest birds do. Additionally, migratory birds exhibit a negative correlation with the Suburban Vegetation component, which is primarily characterized by landscape-scale cropland and scrub cover. Higher values of this component suggest an increase in cropland and natural scrub cover surrounding the riparian zone. It can be inferred that migratory birds prefer to roost outside the riparian zone rather than within it when cropland is present within 1 km of the riparian area. This preference arises because arable land typically offers a more diverse and abundant food supply during winter compared to green fields (Grzędzicka, 2023). It may be more challenging for birds living in suburban areas to survive in winter. The positive impact of increased cultivated land on bird diversity can be attributed to the types of vegetation found in suburban green spaces. Regarding species selection for planted forests, fast-growing tree species are predominantly utilized in suburban riparian zones, however, most of these trees do not retain winter fruits and fail to provide adequate food sources. The scarcity of food plants in suburban riparian zones compels birds to seek potential nourishment in cultivated lands (Møller, 2012; Morelli, 2018; Melles et al., 2003). Combined with edge effects, this may explain the contrasting responses of bird richness and abundance to the Suburban Vegetation and Shrub Cover components. The edges of riparian zones adjacent to arable land exhibit a greater richness than the interiors of both environments. However, the interiors of riparian zones have lower species abundance compared to the exteriors, primarily due to the relatively poor food resources available within the riparian zones, while the nearby cultivated lands are more resource-rich (Henderson et al., 2004; Lee et al., 2022).

The results of the present study indicate that both overall bird richness and Shannon’s diversity index decline with increasing urbanization, consistent with the findings of most previous studies (Ferenc et al., 2014). In contrast, the overall abundance of birds and resident species, exhibited an increasing trend with the degree of urbanization. This suggests the critical role of green spaces for birds in highly urbanized areas. In urban areas, green spaces serve as rare habitats for birds, supporting larger populations compared to similar areas in suburban areas, as evidenced by the urbanization gradient results (Blair, 1996; Conole and Kirkpatrick, 2011).

The results of this study also confirmed another of our hypotheses regarding the effects of the urbanization gradient: the environmental variables that significantly influence bird diversity differ between urban and suburban areas (Blair, 1996). In the UEA, bird diversity metrics other than BA were positively correlated with most components of the environmental variables reflecting plant cover. Conversely, the opposite trend was observed in the UCA. This finding aligns with previous research indicating that vegetation cover has a more pronounced effect on bird diversity in developed areas compared to urbanized areas (Conole and Kirkpatrick, 2011). The disturbance component emerged as the most significant explanatory variable at both the UCA and UEA urbanization levels, exerting a substantial negative impact on birds, irrespective of the level of urbanization. This observation is consistent with the results of earlier studies, which suggest that human disturbance at the site scale may have a considerably greater negative impact on birds than other factors (Schlesinger et al., 2008; Oneal and Rotenberry, 2009). Notably, the disturbance components accounted for a smaller proportion of the differences in bird diversity in the UCA compared to the UEA, leading us to propose that urban birds may be less sensitive to disturbance. Some studies have indicated that urban-adapted bird species exhibit greater tolerance to human disturbance than urban-avoidant species, resulting in notable differences in their diversity (Grzędzicka, 2023; Croci et al., 2008; Shryock et al., 2017).

The effects of environmental variables on both resident and migratory bird populations varied across different levels of urbanization. In the urban expansion area, the number of resident birds showed a significant positive correlation with the Open Vegetation component, in the urban core area, the number of resident birds was significantly positively correlated with the Built-up component. This finding aligns with studies conducted in various seasons, suggesting that grasses may provide foraging sites for certain bird species that feed on insects or grains (Guo et al., 2019). Resident birds responded positively to open vegetation in areas undergoing urban expansion, and exhibited a strong positive response to impervious surfaces in urban core areas. We speculate that these results stem from differences in resource utilization by bird communities along the urbanization gradient. The bird surveys indicated that the richness and abundance of natural resource-dependent birds, such as pheasants and woodpeckers, were greater in suburban areas (BR=61.11%, BA=24.17%), In contrast, urban-adapted birds, such as tree sparrows and magpies—species that thrive in urban environments and effectively utilize human resources—comprised a significantly larger percentage in both number and species in urban areas (BR=61.54%, BA= 95.41%). Differences in species composition lead to varying responses of resident birds to environmental variables across urbanization gradients (Grzędzicka, 2023; Croci et al., 2008), a finding that is particularly pronounced in winter (Galitsky and Lawler, 2015; Ciach and Fröhlich, 2017). However, when considered alongside the Build-up component, the present study offers a new interpretation of these results. The positive impacts of both factors on bird communities clearly illustrate how landscape heterogeneity can enhance regional bird diversity. The two components reflect the degree of heterogeneity in the riparian zone landscape. In urban riparian zones dominated by woodland cover, appropriate impervious surfaces and grasslands contribute to the creation of heterogeneous environments, which provide increased foraging opportunities for a wider variety and number of bird species.

The results regarding the response of migratory birds to environmental variables across different urbanization classes are particularly surprising. First, the relationships between migratory birds and the Build-up and Disturbance components were consistent for both UCA and UEA, likely due to the reasons described above. However, the relationships between migratory birds and the vegetation cover and food source components exhibited a complete reversal with respect to urbanization class. This shift is also related to differences in species composition across urbanization levels. In urban expansion areas, migratory birds are predominantly wintering woodland species, such as Regulus regulus and Turdus naumanni, which often rely on shrublands and grasslands for shelter and food resources. In contrast, the dominant group in urban core areas consists of waterfowl, such as Mergus merganser, for which shrublands are less appealing. This observation aligns with findings that shrublands tend to negatively impact waterbirds during non-breeding periods (Kath et al., 2009; Tian et al., 2023; Chen et al., 2019; Peterson, 2014) while positively affecting woodland birds (Ballantyne and Nol, 2011).