Overview

Urbanization is rapidly increasing and reshaping how mammals move through fragmented landscapes. As cities expand, habitat connectivity declines, and seasonal changes in food availability may further influence species activity and behavior. This study uses environmental DNA (eDNA) to examine whether urban mammal communities show seasonal variation across Detroit, Michigan. I chose this article as it investigates a novel approach for detecting mammals in urban green spaces.

Methods

Soil samples were collected from 21 urban parks in Detroit. Because this method is non-invasive and minimally disruptive, no collection permits were required. These same parks are also part of a long-term camera trap monitoring project. Sampling took place in February and July 2023, representing the winter and summer seasons. Up to four soil samples were collected per park, yielding 33 winter and 32 summer samples for a total of 65. Each sample consisted of approximately 200mL of soil collected within a four-meter radius around focal trees where camera traps were placed. Both topsoil and subsurface layers were sampled. Shoe coverings were worn between sites, and all equipment was sterilized with bleach to prevent contamination. Samples were refrigerated prior to DNA extraction.

DNA was extracted from the 65 soil samples in triplicate, resulting in 195 total extractions, using the Qiagen DNeasy PowerSoil Pro Kit. The MiMammal-U primer set was used to amplify a ~170 base pair fragment of the 12S rRNA gene. Negative controls were included to monitor contamination. Sequencing was performed on an Illumina NovaSeq 6000 platform at the Yale Center for Genome Analysis.

Results

Of the 195 eDNA extractions, 176 successfully yielded DNA sequences identifying 23 mammal species, including humans. Sampling coverage reached 96 percent, indicating strong representation of the mammal community. Human DNA was detected in 58 of the 65 sites, meaning it was present in every park. Domestic species were common, with dogs detected in 10 parks and cats in 13. Pig and cattle DNA were also found across multiple sites, likely the result of food waste or fecal contamination.

Table 1. Mammal species detected at Detroit Parks via eDNA and iNaturalist observations.

Seasonal patterns were evident. In winter, brown rats, pigs, and unidentified Rodentia sequences were detected, while in summer, groundhogs, striped skunks, and muskrats were identified. Species richness varied across parks, ranging from two species at Butzel Playfield, a small 1.7-hectare park, to 14 species at Eliza Howell Park, a 101.2-hectare park that forms part of a wildlife corridor with Rouge Park. Despite these variations, paired comparisons of winter and summer samples showed no significant seasonal differences in alpha diversity (species richness), and ordination analyses (nMDS) revealed low community dissimilarity between seasons.

Changes in seasonal species richness detected across the 21 park sites using eDNA.

Discussion

Monitoring urban wildlife is increasingly important as cities expand and environmental changes alter habitat quality and animal behavior. While eDNA proved effective in detecting a wide range of urban mammals, results suggest that seasonal variation in species richness is minimal within Detroit’s parks. Subtle differences in species presence may reflect behavioral or dietary shifts rather than major community turnover.

Reflection and Critiques

Although this study provides valuable insight into urban mammal diversity and seasonal dynamics, I think several limitations should be acknowledged. Data was collected during a single annual cycle, limiting the ability to account for interannual variability in weather conditions or resource availability. While eDNA offers strong detection for cryptic or elusive species, camera traps may be more cost-effective and reliable for monitoring visible, easily identifiable mammals. Future research combining eDNA and camera trap data across multiple years could provide a more comprehensive understanding of long-term seasonal and environmental trends in urban mammal communities.

Reference

Hallam, J., & Harris, N. C. (2025). Network dynamics revealed from eDNA highlight seasonal variation in urban mammal communities. Journal of Animal Ecology, 94(8), 1587-1602. https://doi.org/10.1111/1365-2656.70082

Overview

I chose this article as it relates to urban wildlife management in the context of urban planning and optimizing cities for wildlife diversity and connectivity. The study focuses on locally rare species representing a wide range of taxa, offering insights into how urban areas can be designed to support biodiversity through habitat corridors. The authors evaluated patterns of habitat availability and connectivity for nine rare species: two insects, three turtles, two snakes, one bird, and one bat across four urban regions in Michigan, USA. These species were selected because they are regionally rare (ranging from state-listed to federally endangered) but still maintain populations within urbanized landscapes. The study aims to identify how urbanization influences habitat structure, connectivity, and opportunities for conservation planning.

Methods

The researchers selected nine focal species representing several taxonomic groups known to be regionally rare. Species occurrence data were compiled from multiple reputable databases, including the Michigan Natural Heritage Database (MNFI), the U.S. Fish & Wildlife Service (USFWS), HerpMapper, iNaturalist, and eBird. To ensure data reliability, only verified MNFI and Research Grade iNaturalist records were used, and records prior to the year 2000 were excluded. Protected areas were mapped by merging all federal, state, county, local, and NGO-managed lands from MNFI GIS layers, excluding disturbed greenspaces such as golf courses.

Urban boundaries were defined using U.S. Census Bureau criteria areas with populations over 50,000 and a density of at least 1,000 people per square mile, with adjacent tracts of 500 people per square mile included as urban periphery. This classification allowed the researchers to analyze habitats embedded within realistic urban-to-rural gradients. Connectivity analyses were then conducted within selected urban regions that met these demographic criteria and contained overlapping species occurrences.

To establish biologically relevant study sites, the team generated 5 km buffers around each species occurrence, encompassing estimated dispersal ranges for most taxa. Overlapping buffers within 10 km of each other were merged, and convex hull polygons were created around the merged clusters. These polygons were then expanded by 5 km to define final analysis areas representing zones of multi-species co-occurrence. This ensured that the study captured both urban and adjacent non-urban habitats crucial for species movement.

Species distribution models (SDMs) were developed for each focal species using an “ensemble of small models” (ESM) approach, which performs well with limited occurrence data. Occurrence points were spatially thinned to a minimum spacing of 1 km to prevent overrepresentation of dense populations, with a target of at least 20 records per species. In cases of limited data, such as the American Bumble Bee, surrogate models were used (the Black and Gold Bumble Bee model).

Road networks were not included in the SDMs directly to avoid false habitat associations, but they were incorporated into resistance surfaces used for connectivity modeling. Roads were buffered by 30 m and assigned species-specific cost values to represent movement resistance: 1000 for reptiles (reflecting high road mortality), 500 for birds and bats (behavioral avoidance), and 250 for bees (potential roadside habitat but increased mortality risk). These road-cost rasters were then merged with each species’ SDM-based resistance surface.

Connectivity was evaluated using multiple tools. Circuitscape was applied to identify likely movement corridors and high-current areas representing multi-directional pathways, while Graphab was used to analyze habitat patch importance and network structure through the dPCk metric. Fragstats provided quantitative measures of landscape configuration, including connectance and clumpiness. Together, these tools revealed both broad and fine-scale patterns of connectivity and potential barriers for each species within and around the urban zones.

Results

Findings showed that many species retain moderate to high habitat proportions within urban landscapes, but the strongest connectivity corridors are typically located outside urban boundaries. Riparian and wetland zones were identified as critical linkages, especially for aquatic and semi-aquatic species such as turtles and snakes. Among the cities, Kalamazoo and Detroit North exhibited the most extensive multi-species connectivity networks, while Benton Harbor and Detroit Southwest showed more limited urban movement potential. Turtles had the highest habitat availability overall, suggesting persistent wetland and forest-edge habitats, whereas grassland species like Henslow’s Sparrow and bumble bees had the least. The study also emphasized that smaller, well-placed habitat patches often contribute more to overall connectivity than larger, isolated ones.

Fig. 3. Map of study area (a) showing Least Cost Paths (LCP’s), (From left to right) Protected areas, barriers impeding connectivity, and current species density.

Critiques & Reflection

This study effectively demonstrates how integrating multiple connectivity modeling tools can inform urban conservation planning. The inclusion of diverse taxa from reptiles to insects gives the findings broad ecological relevance. I found it particularly valuable how the authors linked quantitative spatial analysis to practical conservation implications, such as protecting riparian corridors and small but strategically located habitat patches.

The article does a great job of emphasizing the importance of riparian corridors and buffers, which in my opinion are not adequately regulated under today’s environmental laws. These areas are critical for maintaining ecological connectivity in urban regions, yet they are often overlooked in planning and development regulations.

Although the paper briefly mentions genetics, I believe this is an area that deserves greater focus. Genetic diversity and gene flow are essential components of wildlife corridor restoration, especially for populations that have become isolated by urbanization and experienced bottlenecking. Wildlife corridors are not only about providing safe passage between fragmented habitats,  they also serve as vital habitat themselves, supporting long-term survival, breeding, and dispersal. Many isolated reptile populations, for instance, have not experienced genetic connectivity for centuries due to habitat fragmentation. Snakes are particularly vulnerable to urbanization and the lack of corridors, as they are frequently forced to cross roads where they face high mortality rates, often being killed by vehicles, sometimes even intentionally, despite their protected status.

Despite these challenges, the approach presented in this study offers a replicable framework for urban planners seeking to incorporate biodiversity goals into development strategies. It highlights that even within heavily modified landscapes, thoughtful urban design and targeted protection efforts can sustain rare species populations and maintain ecological connectivity.

Reference

McCluskey, E. M., Kuzma, F. C., Enander, H. D., Cole-Wick, A., Coury, M., Cuthrell, D. L., Johnson, C., Kelso, M., Lee, Y. M., Methner, D., Rowe, L., Swinehart, A., & Moore, J. A. (2024). Assessing habitat connectivity of rare species to inform urban conservation planning. Ecology and Evolution, 14(3), e11105. https://doi.org/10.1002/ece3.11105