Data from: Avian functional space occupancy and trait–environment associations along a composite urbanisation gradient in urban green spaces
Data files
Sep 08, 2026 version files 12.13 KB
-
bird_community_matrix.csv
5.86 KB
-
environmental_variables.csv
1.66 KB
-
README.md
4.61 KB
Abstract
Urbanisation profoundly alters avian communities by selecting for different functional traits. As cities become increasingly important for biodiversity conservation, urban green-space management requires evidence identifying both the trait combinations that become underrepresented with increasing urban intensification and the habitat features associated with particular functional traits. We surveyed bird communities across 24 urban green-space patches in Harbin, Northeast China. We integrated patch-level land-cover composition, within-patch vegetation structure and direct anthropogenic pressure into a composite urbanisation gradient (CUG), and assessed changes in taxonomic diversity, functional diversity, functional space occupancy and trait–environment associations. Species richness declined consistently along the CUG. Functional richness also declined, but this result was sensitive to how residual spatial structure was modelled and should be interpreted with caution. The area occupied within the two-dimensional functional space was progressively smaller across increasing CUG levels, particularly in the region aligned with larger body mass, higher hand-wing index and aquatic-associated traits. In high-CUG green spaces, the core region of this two-dimensional functional space was characterised by greater representation of mid-high foraging strata, canopy foraging strata and woodland habitat use. RLQ and fourth-corner analyses showed that different habitat features were associated with different avian functional traits. Water-body cover was associated with larger body mass, water foraging, and wetland habitat use, whereas vegetation cover and foliage height diversity were associated with forest habitat use. Synthesis and applications: Our findings provide trait-based insights for blue–green urban planning. These associations suggest that urban planning should not treat green spaces as homogeneous habitat patches but should consider both blue-space allocation and nature-based vegetation management. Embedding small wetlands or water bodies in dense urban areas, maintaining blue–green connectivity, and promoting vertically complex woody vegetation may help retain complementary avian functional groups and limit the risk of functional homogenisation in rapidly urbanising landscapes.
Dataset DOI: 10.5061/dryad.nk98sf88q
Description of the data and file structure
This dataset contains environmental variables and bird community abundance data collected from 24 urban green-space patches in Harbin, Northeast China. Bird surveys were conducted from April to August 2024 using standardised line transects. Each site was surveyed five times, and all birds seen or heard within 25 m on either side of the transect were recorded; birds flying overhead were excluded.
Environmental variables describe patch-level land-cover composition, within-patch vegetation structure and anthropogenic disturbance. Patch-level vegetation, water-body, and impervious-surface cover were derived from World Imagery Wayback and checked during field surveys. Vegetation structure was measured in 10 m × 10 m quadrats established along bird-survey transects. Noise level and pedestrian density were measured during bird surveys.
Files and variables
File: environmental_variables.csv
Description: Site-level environmental variables for the 24 urban green-space patches.
Number of rows: 24
Number of variables: 14
Variables
- site: Site identifier (character).
- Noise: Mean ambient noise level for the site (dB; numeric).
- Ped: Pedestrian density observed during bird surveys (individuals/ha; numeric).
- %Veg: Percentage of the green-space patch covered by vegetation, including woodland, shrub and grassland (%; numeric).
- %WB: Percentage of the green-space patch covered by water bodies, including ponds, rivers and lakes (%; numeric).
- %IS: Percentage of the green-space patch covered by impervious surfaces, including buildings, squares and roads (%; numeric).
- TC: Mean tree canopy cover across vegetation quadrats within the site (%; numeric).
- TD: Mean tree density across vegetation quadrats within the site (trees/100 m²; numeric).
- TR: Mean tree species richness across vegetation quadrats within the site (number of species; numeric).
- DBH: Mean diameter at breast height of trees with DBH ≥ 10 cm, measured at 1.3 m above ground (cm; numeric).
- TH: Mean height of trees with DBH ≥ 10 cm (m; numeric).
- SC: Mean shrub cover across vegetation quadrats within the site (%; numeric).
- GC: Mean grass cover across vegetation quadrats within the site (%; numeric).
- FHD: Foliage height diversity, calculated using the Shannon diversity index to describe the vertical distribution of foliage among height layers (unitless; numeric). Higher values indicate greater vertical structural diversity.
Site identifiers
- AY: Anyang River Park
- DX: Lilac Park
- LC: NEFU Forest Farm
- ET: Children's Park
- GL: Ancient Pear Garden
- GX: Guxiang Park
- HJ: Houji Garden
- JG: Jianguo Park
- LD: Laodong Park
- LS: Lishui Park
- LY: Martyrs' Cemetery
- QB: Qingbin Park
- ZW: Forest Botanical Garden
- SZ: Shangzhi Park
- ST: Ecological Park
- SD: Qunli Wetland Park
- SL: Songle Park
- SP: Sports Park
- XL: Xianglin Garden
- XJ: Xiangjiang Park
- XY: NEFU Alumni Forest
- ZL: Zhaolin Park
- TY: Chinese Pavilion Garden
- ZY: Central Park
Missing data: No missing values are present.
File: bird_community_matrix.csv
Description: Bird community abundance matrix for the 24 urban green-space patches. Rows represent sites and columns represent bird species. Values are cumulative abundance counts for each species at each site across the five survey visits.
Number of rows: 24
Number of variables: 84 (1 site identifier + 83 bird species)
Variables
- site: Site identifier (character), corresponding to the site identifiers listed above.
- Species columns: Each remaining column represents one bird species. Species names are formatted as
Genus.species(for example,Passer.montanus). Values are non-negative integer counts of individuals recorded at each site across the five survey visits.
A value of 0 indicates that the species was not recorded at that site. For occurrence-based analyses, abundance values can be converted to presence–absence data as count > 0 = 1 and count = 0 = 0.
Data type: character (site) and integer (species abundance columns).
Missing data: No missing values are present; zeros represent non-detections rather than missing data.
