Search

Urban green cover provides ecosystem services for human well-being and better conditions for biological conservation. Structural analysis and monitoring across space and time at landscape scale can aid better design and management of urban green cover. In this study, we used Sentinel-2 satellite imagery and semiautomatic reclassification methods (i.e., supervised classification) in Google Earth Engine (GEE), to analyze changes in built and urban green cover of Medellín and adjacent municipalities (Colombia) between 2016 and 17 and 2022-23. We used total area, patch number, mean patch area, and patch cohesion index as proxies of landscape structural changes, differentiating grass from trees. Built cover increased its total area from 77.1 to 80.5%, whereas grass and trees decreased it from 6.4 to 5.8%, and 16.6–13.8%, respectively. Built cover decreased in patch number by 1.4% and increased in mean patch area from 7.08 to 7.72 ha, whereas trees decreased in patch number by 7.8% and grass increased by 9.2%. Both green cover categories decreased in mean patch area (trees: 0.27 to 0.23 ha; grass: 0.24 to 0.15 ha). Otherwise, the patch cohesion index decreased in grass and trees, while in built cover it remained constant over time. Urban green cover in Medellín urban area and adjacent municipalities is increasingly reduced and isolated, whereas urbanization sprawls and densifies, suggesting that local development policies need improvement to prevent urban green cover degradation. Highlights: Medellin´s landscape is turning into consolidation and densification of built cover. Changes in grass cover implied the conversion of open fields into built-up areas. Smaller tree patches experienced the most significant changes. Urban planning in Medellín needs improvement to prevent further loss and segmentation/isolation of green cover. [ABSTRACT FROM AUTHOR]

Evaluating several biodiversity descriptors and considering several spatial scales might elucidate conservation issues and improve biodiversity monitoring in urban environments. We estimated species richness (order q = 0), Shannon diversity (order q = 1), and Simpson diversity (order q = 2) based on Hill numbers and performed cluster analysis and non-metric multidimensional scaling (NMDS) to compare seven urban micro-watersheds and a peri-urban site across a northern Andean city (Medellín, Colombia). We found 113 diurnal resident bird species: 50 (44 %) exclusively within urban sites, 21 (19 %) exclusively in the periurban site, and 42 (37 %) shared species. Some urban watersheds had similar bird species richness to the periurban site, but Shannon and Simpson diversities were always lower, showing decrease in local bird diversity when abundances were considered. Bird species composition differed between urban watersheds and the periurban site, with all urban watersheds grouped altogether by cluster and NMDS analysis, and the periurban site forming its own group. This suggests homogenization of bird species composition due to the species turnover decreasing across urban areas, with endemic, near endemic and rare species restricted to periurban areas where native forest remnants persist. Several scales of biodiversity and analysis at more local scales are needed to better understand biodiversity patterns across Andean cities and to design urban planning strategies that prevent biodiversity loss. [ABSTRACT FROM AUTHOR]

• Larger and more regular-shaped green areas increased bird species richness. • Less impervious surfaces and more grass-shrubs increased bird species richness. • Introduced trees' dominance decreased bird species richness. Cities are human-dominated ecosystems where landscape transformation decreases biodiversity, a conservation concern when urbanization sprawls into biodiversity hotspots with high endemism. We aimed to evaluate the effects of site-specific features such as size and shape of green spaces, land cover, and vegetation on bird species richness in Medellín, Colombia, a city of Northern Andes (n = 44 urban green spaces sampled from February 2018 to February 2019). We found 255 plant species: 23 % of trees and 41 % of understory were native sensu stricto (local origin); most native species and individuals were in green spaces lacking vegetation management. Bird species richness (25.80 ± 8.05 resident species per sampling point; 83 resident species overall) increased towards larger and more regular-shaped urban green spaces (squared or rounded), with less percentage of impervious surfaces and more of grass–shrubs, and less influence of introduced trees in habitat structure (i.e. less crown coverage, basal area, and average height). Bird species richness also increased when richness or abundance of native understory vegetation sensu lato (local or regional origin) increased, but introduced trees dominance represented higher and most significant effects to explain bird species richness. Increasing edge effects across urban green spaces and human-related habitat transformation favoring non-native vegetation might represent significant constraints for enhancing local biodiversity in highly developed Andean cities. [ABSTRACT FROM AUTHOR]