Fuzhou （26°09'N~26°00'N, 119°24'E~119°42'E）is located on the south-eastern edge of the Eurasian continent, bordering the Pacific Ocean to the east and falls within a subtropical monsoon climate zone (Fuzhou Government Portal; https://www.fuzhou.gov.cn/szf/). The study area is situated along the urban section of the Min River in Fuzhou. With the accelerating pace of urbanisation, this region has developed extensive residential zones. However, little research has been conducted on how these developments affect bird communities in the adjacent residential areas along the river.

In this study, we selected a section of the lower Min River near the urban core as the focal area. The study boundary was defined as a 1 km buffer zone on both sides of the river's centre-line, within which survey sites were established (Fig. 1) (Pennington and Blair 2011, Litteral and Shochat 2017, Liu et al. 2022). Following the minimum patch size standard of 2 ha specified in Technical Guidelines for the Construction and Restoration of Urban Green Space Bird Habitats (https://yllhj.beijing.gov.cn/sdlh/ylkj/bzgf/201911/P020220208599812707167.pdf) and based on remote sensing interpretation and field survey results, we randomly selected 33 typical residential areas within the study boundary — each with an area no less than 2 ha — as the research sites.

Figure 1.  

Location map of the 33 residential areas along the Min River.

In this study, the 33 selected residential areas were categorised into three types based on their year of construction: those built before 2005 (Y1), those built between 2006 and 2015 (Y2) and those built between 2016 and 2025 (Y3). Amongst them, there are 12 Y1-type residential areas, 11 Y2-type areas and 10 Y3-type areas.

The spatial distribution of the three residential types along the Min River in Fuzhou shows distinct variation (Fig. 2). The names and construction years of all residential areas are provided (Suppl. material 1).

Figure 2.  

Distribution of three types of residential areas along the Min River.

A nested sampling design was applied to integrate bird and vegetation surveys (Yang et al. 2015). Each 100 m transect segment served as the bird survey unit, with a 20 × 20 m vegetation plot at its centre. Plot shapes were adjusted when necessary to maintain area consistency.

The environmental variables selected in this study were guided by established ecological theories and empirical findings. High building density, floor area ratio and tall structures often reduce bird diversity by intensifying noise, disturbance and habitat fragmentation (Yu et al. 2025). In contrast, tree diversity and evenness enhance habitat heterogeneity, providing varied food resources and nesting niches that support species co-existence (Batáry and Fronczek 2014). Shrub and herbaceous layers can supply additional cover and foraging substrates, though excessive density may reduce accessibility and foraging efficiency (Schlossberg et al. 2010). Together, these factors provide a comprehensive framework for assessing how vegetation and built environments shape avian communities in residential areas.

A total of 56 bird species, belonging to seven orders, 28 families and 45 genera, were recorded in the 33 residential areas, with a cumulative individual count of 11,100. Based on individual abundance, the top five most frequently observed species were: the Eurasian Tree Sparrow (Passer montanus, 2,476 individuals), the Swinhoe's White-eye (Zosterops simplex, 2,077), the Light-vented Bulbul (Pycnonotus sinensis, 1,539), the Chinese Blackbird (Turdus mandarinus, 1,450) and the Spotted Dove (Spilopelia chinensis, 513).

In terms of residency status (Fig. 3), the bird communities across the 33 residential areas were predominantly composed of resident birds and winter migrant birds. Resident species accounted for 40 species, representing 71.43% of the total recorded species, while winter migrant birds comprised 12 species. Summer resident birds and traveller birds were less common, with only three summer resident birds and one traveller bird observed. Across the different residential types, Y3 had significantly more winter migrant birds compared to Y1 and Y2. Both Y2 and Y3 areas also recorded more summer resident birds than Y1, with Y2 exhibiting the highest number of summer species.

Figure 3.  

Percentage stacked bar chart illustrating the proportion of resident-type bird abundance across three types of residential areas.

Regarding dietary guilds (Fig. 4), omnivorous birds were the most frequently observed, with 26 species accounting for 47.27% of the total recorded species. Insectivorous guilds, with 19 species, followed this. Carnivorous and herbivorous guilds were the least represented, each with five species recorded. In terms of spatial distribution, Y2 supported a significantly higher number of both carnivorous and omnivorous species compared to Y1 and Y3, whereas Y1 exhibited the highest number of herbivorous species.

Figure 4.  

Percentage stacked bar chart illustrating the proportion of bird species' feeding habits across three types of residential areas.

According to the species accumulation curve (Fig. 5), the overall species richness tends to plateau with increasing sample size, indicating that the bird survey data is sufficiently comprehensive and suitable for subsequent analysis.

Figure 5.  

Species accumulation curve of birds in the residential area.

Notable differences in bird α-diversity were observed amongst the different types of residential areas. Y2 exhibited higher values in the BH', BJ' and BA compared to Y1 and Y3, following the trend Y2 > Y1 > Y3. In terms of BS, however, the pattern was Y2 = Y3 > Y1 (Fig. 6).

Figure 6.  

Comparison of bird α-diversity indices across transects in different residential areas.

The Kruskal–Wallis test revealed no significant overall differences in BH', BS or BJ' amongst the three residential types. However, for BA, a significant difference was observed (P < 0.05). Subsequent pairwise Mann–Whitney U tests revealed that this difference was specifically between the Y2 and Y3 neighbourhoods (P < 0.05).

Spearman correlation analysis (Fig. 7) revealed distinct patterns across residential area types (Suppl. material 2). In Y1 areas, bird diversity showed a significant positive correlation with GR and a highly significant positive correlation with TH'. Both BS and BJ' were also significantly positively correlated with TH'. In Y2 areas, BS was significantly positively correlated with the ATCW_NS. In Y3 areas, both BH' and BS were significantly positively correlated with TH'. BH' also showed a significant positive correlation with TJ', while BS was highly significantly correlated with TJ'. Additionally, BA showed a significant positive correlation with the ATCBH.

Figure 7.  

Heatmap of correlations between residential environmental factors and bird α-diversity metrics.

The results of the stepwise regression analysis (Table 1) reveal variations in the combinations of significant environmental predictors across different residential types. In Y1 areas, BH' was negatively associated with BD and FAR and positively associated with ATH and TH' (R² = 0.722). BS was negatively correlated with BD and positively correlated with ATH and TH' (R² = 0.443). The model for BJ' included HJ', ASH and TJ' as predictors (R² = 0.718). The model for BA identified GR, AHPH and TJ' as positive predictors (R² = 0.620). In Y2 areas, BH' was negatively associated with HS, SJ' and BD (R² = 0.270). The model for BS included ATCW_EW and FAR, both of which were positively associated (R² = 0.359). The model for BJ' retained only HS as a predictor (R² = 0.070). In the model for BA, TJ' was a positive predictor, while BD, ABH and ATH were negatively associated (R² = 0.715). In Y3 areas, TH' was the sole predictor of BH' (R² = 0.195). The BS model included TJ' and TH', both of which had positive effects (R² = 0.318). The model for BJ' included only TH' (R² = 0.071). For BA, the model included SH' and ASH as negative predictors, while ATCBH, TJ' and TH' were positively associated (R² = 0.562).

Between 1 October 2023 and 30 September 2024, 12 bird surveys were conducted across 33 residential areas along the Min River in Fuzhou. A total of 56 bird species were recorded, spanning seven orders, 28 families and 45 genera. Resident birds dominated the assemblage, accounting for 40 species, suggesting that residential environments in Fuzhou provide relatively stable food resources and suitable habitats for year-round habitation.

Interestingly, the number of winter migrant birds was significantly higher in newly-developed Y3 areas compared to the other residential types, which may be attributed to more diverse habitats and abundant food resources in Y3 (Jokimäki and Suhonen 1998), particularly the prevalence of winter-fruiting ornamental plants such as mandarin orange (Citrus reticulata), lemon (Citrus limon) and bitter orange (Citrus × aurantium), which are commonly planted in newer developments.

Dietary guild analysis showed that omnivorous species had the highest species richness and abundance across all residential types, followed by insectivorous birds. This pattern aligns with findings from urban bird surveys in other subtropical cities, such as Guangzhou (Wu et al. 2024). The dominance of omnivores reflects their broad dietary plasticity, which enables them to exploit a wide range of food sources, including anthropogenic waste and cultivated horticultural plants, thereby enhancing their adaptability to dynamic urban environments. Their capacity to cope with seasonal and spatial fluctuations in resource availability also likely contributes to their ecological success in cities (Jokimäki and Suhonen 1998, Marcacci et al. 2021, Ding et al. 2025).

Notably, Y2 areas supported significantly higher numbers of both omnivorous and carnivorous birds compared to Y1 and Y3, which may be linked to the presence of naturalistic waterbodies within Y2’s cluster green spaces, which attract carnivorous species, such as the Little Egret (Egretta garzetta), Chinese Pond Heron (Ardeola bacchus) and Black-crowned Night Heron (Nycticorax nycticorax).

Analysis of bird diversity indices revealed that Y2 neighbourhoods exhibited the highest levels of bird diversity, evenness and abundance, indicating more complex community structures and more balanced species distributions. This finding is consistent with the study by Rebecca Thompson et al. (2022), who reported that parks of intermediate age (40–60 years) had significantly greater overall biodiversity compared to both older central parks (> 60 years) and younger suburban parks (< 40 years). The only exception was species richness, which showed no significant difference between intermediate and older central parks, although both were significantly higher than those in younger parks (Thompson et al. 2022).

These trends may be picking up on what has been indicated elsewhere that intermediate urbanisation can lead to higher species richness (Faeth et al. 2011, Filloy et al. 2019). Notably, unlike the findings of Thompson et al. (2022), this study observed comparable bird species richness between the Y2 and Y3 neighbourhoods, which suggests that some newly-developed communities have begun to achieve tangible biodiversity gains through early integration of ecological design principles. This conclusion also supports the view of Aronson et al. (2017) that, even in densely built urban environments, maintaining high biodiversity is possible by incorporating diverse vegetation structures, effective disturbance management and ecological connectivity (e.g. corridors and buffer zones) at the early planning stages (Aronson et al. 2017). Furthermore, Wilcoxon test results indicated a significant difference in bird abundance between Y2 and Y3. While the two types exhibited similar levels of species richness, abundance was lower in Y3 — likely due to higher levels of anthropogenic disturbance, including ongoing construction and increased human activity. In contrast, Y2 areas, having undergone longer ecological establishment and habitat stabilization, were better able to support greater bird abundance.

Compared to urban parks, residential green spaces generally support lower bird diversity due to their smaller size and greater spatial isolation (Sorte et al. 2020, Zheng et al. 2025). This pattern aligns with the species–area relationship theory (Lomolino 2000), which posits that smaller habitat patches tend to support fewer species. Nevertheless, these spatially constrained green spaces can still offer considerable ecological value.

During migratory seasons in particular, species such as the Daurian Redstart (Phoenicurus auroreus), Stejneger's Stonechat (Saxicola stejnegeri) and Barn Swallow (Hirundo rustica) are frequently recorded in residential areas, underscoring their role as ecological “stepping stones” within the urban landscape (Partridge and Clark 2018, Buron et al. 2022). Furthermore, these habitats occasionally provide refuge for rare species, including the nationally protected (Class II) Chinese Hwamei (Garrulax canorus), highlighting their significance for urban biodiversity conservation.

The ability of residential green spaces to sustain bird diversity largely depends on their specific environmental characteristics, particularly vegetation structure and composition. Spearman’s rank correlation analysis identified distinct patterns in bird responses to vegetation-related variables across residential areas of different construction periods. In the older Y1 areas, bird α-diversity exhibited a significant positive correlation with both greening rate and tree Shannon Index, while richness and evenness were particularly associated with tree Shannon Index. A diverse tree assemblage offers varied food resources and structural niches, as differences in average tree height, average tree crown base height and average tree crown width fulfil the ecological requirements of a wide range of bird species (Fontana et al. 2011).

Previous studies have similarly shown that the vertical stratification provided by mature trees supports territorial or strata-specialist species, thereby contributing to community stability (Vesk et al. 2008, Shanahan et al. 2011). In Y2 areas, bird richness was significantly correlated only with the average tree crown width (N–S). This structural attribute may enhance the functional value of trees as bird habitats by improving shade provision, moderating microclimates and promoting spatial continuity across fragmented green spaces (Fernandez-Juricic and Jokimäki 2001).

In the more recently developed Y3 areas, both bird Shannon Index and richness showed strong positive correlations with tree Shannon Index and tree evenness, whereas bird abundance was positively associated with average tree crown base height. Higher tree evenness likely indicates a more balanced spatial arrangement of tree species, which may support more equitable foraging and nesting opportunities (Kang et al. 2015). Additionally, average tree crown base height directly influences the vertical movement space and concealment available to birds, a factor particularly important for species that forage in the mid- to upper-canopy layers, such as those from the Paridae and Muscicapidae families (Xu et al. 2022b).

In urban ecosystems, bird communities exhibit highly dynamic responses to environmental factors, which are continuously reshaped by changes in urban spatial structure, vegetation composition and the built environment (Clergeau et al. 2001, Marzluff 2017). Furthermore, ecological succession, associated with the temporal progression of urban development, exerts a long-term influence on habitat suitability and species composition. Residential areas constructed during different periods reflect varying planning philosophies, greening strategies and spatial configurations, resulting in heterogeneity in bird sensitivity and response mechanisms to environmental variables.

These differences are further supported by stepwise regression analysis, which identifies divergent combinations of key influencing factors across the three construction periods. In Y1 areas, the regression models include the greatest number of explanatory variables and exhibit relatively high coefficients of determination (R²), suggesting a complex interplay between bird communities and environmental factors. Bird Shannon Index is positively associated with average tree height and tree Shannon Index, but negatively correlated with building density and floor area ratio. These findings are consistent with previous studies, which have shown that high-density built environments limit bird diversity (Amaya-Espinel et al. 2019).

In contrast, structurally and compositionally diverse tree vegetation supports richer bird communities by providing abundant food resources and nesting sites. Bird richness is similarly influenced by average tree height and tree Shannon Index, highlighting the role of vertical vegetation complexity in enhancing habitat quality. Bird Pielou evenness is affected by herbaceous evenness, average herbaceous plant height and tree evenness, indicating that a multilayered vegetation structure can reduce dominance by a few species and promote a more balanced bird community. Although greening rate and average herbaceous plant height were identified as positive predictors of bird abundance in the models, field observations revealed that certain older neighbourhoods with relatively high greening rate suffered from low plant species diversity, poor maintenance and delayed vegetation renewal. These limitations reduce the ecological functionality of the green space. Thus, greening rate alone is not a reliable indicator of habitat quality; more attention should be directed towards introducing native species, enhancing structural complexity and promoting ecological heterogeneity (Catalano 2024).

In Y2 areas, the overall explanatory power of the regression models is relatively low, with only the bird abundance model demonstrating strong predictive capability. This may indicate that bird responses in mid-aged residential settings are more localised or tend to stabilise due to moderate habitat maturation. Bird Shannon Index is negatively correlated with herbaceous species richness, shrub evenness and building density, suggesting that, in the absence of strong vertical structural support, increased herbaceous species richness alone is insufficient to support diverse bird communities (Nielsen et al. 2014). In addition, average tree crown width (N–S) and floor area ratio show positive associations with bird richness, implying that, under moderate development intensity, expanded canopy coverage and balanced vertical development can enhance habitat utilisation efficiency (Amaya-Espinel et al. 2019). The Bird Pielou evenness model includes only herbaceous species richness, with limited explanatory power — possibly reflecting a dominance by a few generalist species, which vegetation metrics cannot fully explain. For bird abundance, tree evenness acts as a positive factor, while building density, average building height and average tree height act as negative predictors. These results underscore that a well-distributed tree canopy supports higher bird abundance. In contrast, densely built, high-rise environments constrain available habitat space, particularly affecting ground-foraging or shrub-dwelling species (Xu et al. 2022).

In Y3 areas, the number of explanatory variables in the models decreases, yet the explanatory power remains moderate. The Bird Shannon Index model retains only tree richness as a significant predictor, indicating that, in these newer developments, the richness and even spatial distribution of introduced tree species are central to supporting bird diversity. Bird richness is positively correlated with both tree Shannon Index and tree richness, re-affirming that increasing species heterogeneity remains a key strategy for attracting a wider range of bird species. Similarly, the Bird Pielou evenness model includes only tree richness, suggesting that bird communities in these areas are in early successional stages and have not yet established complex interspecific competitive dynamics (Blair 1999, McKinney 2006, Shochat et al. 2006). The bird abundance model includes negative predictors, such as average shrub height and shrub Shannon Index. In contrast, the positive predictors are tree-related variables, including tree Shannon Index, tree evenness and average tree crown base height. These findings highlight the importance of optimising vertical vegetation structure. Excessively tall shrubs may obstruct bird movement corridors and overly diverse shrub layers may introduce structural clutter, hindering foraging efficiency and vigilance behaviour for some species (Robinson and Holmes 1984). Conversely, the openness and vertical layering provided by well-structured tree canopies create favourable conditions for a diverse bird assemblage, thereby enhancing local abundance and activity frequency.