Network dynamics revealed from eDNA highlight seasonal variation in urban mammal communities

Abstract

Urban ecosystems are expanding rapidly, significantly altering natural landscapes and impacting biodiversity. Here we explore seasonal variation in mammal diversity using environmental DNA (eDNA) from soil samples collected during winter and summer across 21 urban parks in Detroit, Michigan. We estimated gamma (regional), alpha (local) and beta (compositional change) diversity to determine if seasonal shifts, reflecting winter scarcity and summer abundance in mammal community composition and human activity, could be detected using eDNA. We expected that larger parks would exhibit greater diversity and higher seasonal turnover, consistent with the species-area relationship (SAR) and hypothesised that increased summer resource availability would lead to decreased network density as species disperse more broadly. We found that urban parks show subtle, park-specific changes in community composition influenced by both ecological and anthropogenic factors, with species including striped skunk, brown rat and groundhog responsible for the observed seasonal variation. Consistent with the SAR, larger parks supported higher species richness and diversity. Ecological network analysis, focusing on metrics such as clustering coefficient and network density, revealed a decrease in the overall connectivity and cohesiveness of species interactions from winter to summer, supporting our hypothesis of broader species dispersal during resource-rich periods. Notably, human DNA was prevalent in all parks, alongside detections of pig and cow eDNA, potentially reflecting human disturbance and anthropogenic food inputs. Our findings underscore the efficacy of eDNA analysis in capturing urban mammal community dynamics, the impact of human activities on biodiversity and its potential as a valuable tool for urban ecological research. Ultimately, enhancing monitoring capacity aids in conservation and urban planning efforts that will promote human-wildlife coexistence and preserve the socio-ecological benefits stemming from biodiversity across cityscapes.

Keywords: biodiversity; environmental DNA; green spaces; human‐wildlife interactions; mammal community dynamics; urban ecology.

FIGURE 1

Conceptual illustration depicting variation in animal diversity across different urban sites and between winter (top panel) and summer (bottom panel). Some species are widely distributed and can be detected across multiple sites regardless of the season, while others are rare or seasonally distinct. Alpha diversity decreases from left to right moving from more diverse large green spaces (a, b) to increasingly urban, smaller sites (d, e). As alpha diversity decreases, beta diversity transitions from being turnover‐dominated to nestedness‐dominated. High diversity sites (a, b) may support more specialist species whose presence is tied to seasonal resource availability. In contrast, less diverse sites (c–e) may exhibit a buffering effect of urbanization on seasonal change, resulting in more homogenized, urban‐adapted wildlife communities.

FIGURE 2

Study area‐ Detroit, Michigan. Shaded green areas represent the city parks where soil samples were collected for eDNA sampling in 2023 for this study. Diagram illustrates sample collection plan, where 20 subsamples of soil were collected along transects around a focal tree, covering 6 M by 8 M. Parks are numbered by size in descending order: (1) Rouge; (2) Palmer; (3) Eliza Howell; (4) Chandler; (5) Farwell; (6) Fort Wayne; (7) O'Hair; (8) Lasky; (9) Balduck; (10) Maheras; (11) Henderson; (12) Ford; (13) Clark; (14) Stoepel; (15) Comstock; (16) Romanowski; (17) Van Antwerp; (18) Marruso; (19) McCabe; (20) Butzel Playfield; (21) Fields. Mean monthly temperatures were 28°F (−2°C) [min 22°F (−6°C), max 35°F (2°C)] in February and 75°F (24°C) [min 66°F (19°C), max 84°F (29°C)] in July.

FIGURE 3

Changes in alpha diversity (species richness) between winter and summer species detections from eDNA sampling of soil samples across 21 urban parks in Detroit, Michigan. Parks with increases in diversity are shown in dark grey, while those with decreased diversity are shown in light grey. Parks are ordered by size, from smallest to largest.

FIGURE 4

Species co‐occurrence networks in winter (left) and summer (right). Nodes represent species, sized by degree centrality with colour indicating seasonal occurrence: Lighter nodes indicate species unique to that season, while dark nodes represent species with statistically significant co‐occurrences. Edges represent co‐occurrence relationships, where dashed edges indicate statistically significant co‐occurrence (p ≤ 0.05), and solid edges indicate non‐significant associations (p > 0.05). Isolated nodes indicate species that did not co‐occur with others in the same park. Co‐occurrence was defined at the park level. Layouts were generated using the Kamada‐Kawai algorithm.