Fig 1.
Upper left panel: Geographic location of the study area.
Central panel: image of the city of Armenia, Central Andes of Colombia, South America, showing the grid designed to select urban cells (250 m x 250 m each). Numbers within grid point urban cells indicate where habitat characteristics and bird surveys where performed (Sept 27 of 2016 –Feb 17 of 2017), while green dots and respective code indicate sampled rural cells (Nov 29 of 2016 –Feb 17 of 2017). Lower right panel: example of one of the 76 urban cells where different types of coverages are present (i.e. forest, guadual, and open and impervious surface areas). The yellow line indicates the political limits of the municipality of Armenia. Image taken from Sig-Quindío (CRQ-GIS 2017).
Fig 2.
Panel showing some representative species of the bird diversity in the city of Armenia, Colombia.
Species shown correspond those with most abundance of individuals in urban cells (Illustrations by Yemay Toro-López and Anny Pulido-G), and a unique species catalogued as Vulnerable by the IUCN 2017 (Dacnis hartlaubi).
Fig 3.
Coverage-based rarefaction (solid line) and extrapolation (dashed line) plots with 95% confidence intervals (shaded area) for bird species diversity (based on Hill numbers, q = 0, 1, 2) between urban and rural habitats.
Reference samples are denoted by solid symbols.
Fig 4.
Species richness and number of individuals (abundance per sampling cell) between bird assemblages in the urban and rural habitats.
Data shows mean values with standard error of the mean.
Fig 5.
Partial plots of a multiple regression analysis showing the Relationship between the bird species diversity at different q orders (i.e. 0,1,2) and the intensity of noise (A,E,I), the percentage of impervious surface and guadual surface area (B,F,J), distance to the edge of the city and percentage of open surface area (C,G,K), and percentage of forest surface area (D,H,L) in 76 urban cells in the study area This figure has illustrative aim due that results and discussion are based on results obtained in a Generalized lineal model(GLM), however, results obtained in the GLM and the multiple regression analyses were similar. Values on y-axis (left) and x-axis (bottom) are residuals of a multiple regression analysis. Regression lines in bold indicate significant relationships at alpha = 0.05.
Fig 6.
Diagrams of the Non-metric Multidimensional Analysis (NMDS) showing the absence of urban cell (plot A) and rural cell (plot B) sorting, according to similarities of species of birds in them. Codes beside black dots indicate location of urban and rural cells (see Fig 1).
Table 1.
Principal component analysis summarizing variation in urban and rural (values in parenthesis) habitat characteristics.
The highest loadings for each principal component was 0.7.
Fig 7.
Partial plots of a multiple regression analysis showing the absence of a significant relationship between the bird species diversity at different q orders (i.e. 0,1,2) with the percentage of open and plantation surface area (A,D,G), the percentage of forested and impervious surface (B,E,H), distance to the edge of the city and guadual surface area (C,F,I), in 23 rural cells in the study area. Similar to the Fig 5, this figure has illustrative aim due that results and discussion are based on results obtained in a Generalized lineal model(GLM), however, results obtained in the GLM and the multiple regression analyses were similar. Values on y-axis (left) and x-axis (bottom) are residuals of a multiple regression analysis which means that each plot shows the net relationship between both variables after statistically controlling by the effect of the others.