2024 Purchase College Journal of Ecology

Land development can threaten and disturb local wildlife by dividing their habitat into fragments. To assess the harm done to mammal populations of Blind Brook Forest in Westchester, New York, we set up trail cameras in a recently cut area of the forest and an undisturbed old growth area of the forest. When comparing our observations, we find that local mammals clearly prefer using the undisturbed forest to the meadow resulting from development. Despite the fragmentation, though, we found many unexpected species that signal a healthy ecosystem. We are eager to restore the cut area and prevent further development in order to keep the environment stable.

Fragmentation; Habitat preference; Mammals; Mesopredators; Restoration

Human-oriented land change tends to maximize humanity’s short-term returns from the land while reducing the length of time that returns can be (literally or figuratively) harvested from it. We raze forests every day for lumber, farmland, city space, and roads, all of which reduce the long-term health and sustainability of ecosystems that non-human wildlife rely on (Foley et al., 2005). Direct negative effects like the sullying of water sources can take precedence in the short-term, but equally important are weaknesses in ecosystems that development creates: invasive species, disease, pests, and climate extremes more easily cripple forests scarred by human intervention (Foley et al., 2005).

Scars in the environment can appear as deforested segments of land, dividing the ecosystem into fragments that degrade faster than connected cover: trees in the Amazon rainforest, for example, have double the mortality rate up to 300 meters from the fragment edge of a habitat, and large trees are almost 40% more likely to die than even that base factor (Laurance et al., 1998, 2000). Fragmentation in cities and suburbs is equally devastating, ousting native species sensitive to isolation or exposure. A lucky, oft human-compatible minority of invasive species and native species with a previously limited range face no competition and fill the new niches (Hansen et al., 2020; McKinney, 2006). This homogenization reduces the resilience of an ecosystem by forcing it to rely on the hardiness of a scant few species. To prevent or heal this scarring, humans must monitor local habitats and ensure that vulnerable ecosystems remain biodiverse (Foley et al., 2005).

In suburban edge environments created by land development, some wildlife species tend to thrive, some tend to be unaffected, and some tend to suffer (Crooks, 2002). Larger, more dominant mesopredator populations like bobcats and coyotes become nearly nonexistent if fragments become too small and isolated (Crooks, 2002; Hansen et al., 2020). This can “release” less dominant species that would have otherwise been outcompeted (Crooks, 2002). In addition, less dominant species are often those that can take more advantage of human features like feeders, water features, garbage, and compost (Crooks, 2002; Hansen et al., 2020; McKinney, 2006; Tigas et al., 2002). Larger predators, of which this paper will primarily focus on bobcats and coyotes, can take advantage of these anthropogenic resources but clearly prefer their natural food and lifestyle (Hansen et al., 2020; Riley et al., 2003). It has also been suggested that population changes run in reverse of the “release”- an abundance of small prey animals can help sustain or temporarily attract the larger mesopredator population (Hansen et al., 2020; Tigas et al., 2002). The forces governing the interaction of mammals amongst themselves and with humans require close attention to discern.

This study is part of an attempt to restore and conserve the ecosystem in an old growth forest in Westchester, New York. A long, thin section of the Blind Brook Forest, bordered on one side by a suburb and on the other by the more urban SUNY Purchase College campus, was completely cleared to install a sewage pipe. The Environmental Studies department at Purchase is actively restoring this cut area and aims to assess how mammal species are handling this fragmentation. We designed this study to determine whether mammals were using the open cut area (hereafter, “meadow”) more or less than the forest, whether they were more active at certain times of day, and whether there was a difference in diversity and richness between the two environments. If we have an idea of how the mammals in our ecosystem typically react to fragmentation, we can better interpret the health of the forest and what needs our focus in restoration efforts (Crooks, 2002).

The goal of our methods was to enable a comparison and contrast of total mammal presence, mammal activity at different times of day, and mammal species diversity and richness among the different habitats in the Blind Brook Forest and its cut area on the SUNY Purchase campus (Figure 1).

We established two “macro” habitats, dubbed the meadow and forest. The meadow section is the cut area, a slice reaching about 150 m into the forest and 25 m across, consisting of low new growth and taking the brunt of the negative edge effects: invasive pressure, sensitivity to weather, etc. It contains “micro” habitats - sections closer or farther from the road and closer or farther from the edges. The forest section is the undisturbed old growth forest around the cut area. The micro habitats contained within are the wet lowland areas near the brook and the dry upland areas near the road. Placing trail cameras in a variety of these micro habitats would give us greater range in our analysis of wildlife land use while still providing us with nominal replicates of our macro habitats.

In the fall, three Browning Strike Force Pro X 1080 trail cameras were placed along the meadow and three were placed in the forest on the left side of the cut, or farther north. One extra camera was placed in the forest on the right side of the cut, or farther south (Figure 2). The second and third forest camera placements ended up needing adjustment, but after ensuring that they were spaced out enough to capture an adequate variety of micro habitats we took their GPS coordinates and left them for a little over two weeks.

The camera settings had been standardized with a miniature experiment conducted earlier in which we armed them and moved past them at a variety of speeds and settings, settling on what we felt would capture the best photos without extras taking excess memory or bloating data: they were set to still picture mode at 4 MP quality with a one second capture delay. Multishot mode was off. We put in SanDisk Ultra 32 GB SDHC cards. Our study was partially modeled on “Does Use of Backyard Resources Explain the Abundance of Urban Wildlife?” by Hansen et al., 2020, and that paper primarily concerns mesopredators. For that reason, our cameras were generally supposed to be set a foot or two off the ground facing straight ahead to capture that class of mammal. While they were mostly in this ballpark, they did not have a standardized angle or height.

The cameras were named based on their location: for example, though they were both in the meadow, the camera named Meadow3 was closest to the road and Meadow1 was farthest. Forest1 was closest to the cut and Forest3 was farthest. Our outlier camera on the opposite side of the cut was named Forest4, though Forest4 ended up being excluded from the data because it was pointed at a steep angle downwards. This skewed its data heavily enough towards mice and other small creatures that it was noticeably incongruent with the other cameras. All photos were marked with the name of the camera that took them and the date and time they were taken.

Photos taken between 7 PM on October 2nd and 4 PM on October 16th, 2024, were scored. All students of the fall 2024 Wildlife Ecology class took part in scoring and photos were later reexamined by a small group of students concentrating on the study to guarantee accuracy. Students tallied the number of each animal species they observed in a one hour block of photos from a given camera. If more than one photo captured an individual of the same species within a five minute span of the initial photo, only one observation was marked. If any photos within a five minute span contained multiple individuals of the species, the student marked however many individuals were in the most populated photo of that five minute span rather than one. If an animal appeared at both the end of an hour and the beginning of the following hour it was generally counted for the first hour only. By the end of our scoring, we had the means to analyze mammal activity, richness, and diversity in a variety of Blind Brook Forest habitats.

We observed about five times more animals in the forest than in the meadow. There were 1007 sightings in the forest and 186 in the meadow. Each of the forest cameras captured a greater abundance of animals than even the most active meadow camera. In addition, there is a clear increase of abundance within the meadow environment as the cameras get closer to the road: there were 38 sightings from Meadow1, 53 from Meadow2, and 95 from Meadow3 (Figure 3).

Most species were observed primarily at night, with a few notable exceptions: squirrels and chipmunks were hugely active during the day, the only animals with a strong diurnal trend (Figure 4a). Deer were seen mostly during the day in the meadow but mostly at night in the forest (Figure 4b). Every other animal was seen primarily at night (Figure 4c). This could indicate avoidance of humans, who often passed the cut area or entered it to work on restoration. For the most part, however, the activity aligns with the species’ natural active hours.

Nearly every animal preferred the forest, spotted there more frequently than in the meadow. Bobcats were an exception, observed almost entirely in the meadow. Two thirds of skunk sightings were in the meadow. Weasels were the only species exclusively photographed in the forest (Figure 5).

There was no significant difference in either richness (number of species) or diversity (evenness of species) between the two habitats. Both had all of the same species with the exception of the weasel, which was only found in the forest. There was a minor difference in the Shannon diversity indices calculated for the two habitats (Figure 6).

We found that species generally prefer to operate under the cover of the forest. This is clear when comparing the habitats’ overall abundance. Bobcats may be an exception to this rule because they hunt rodents, which are more exposed in the meadow (Tigas et al., 2002). Squirrels and chipmunks, arguably the most vulnerable species to predators, made up the majority of activity in the forest. It’s possible that they are more abundant than other animals in this suburban environment because they can more easily exploit the human resources nearby (Hansen et al., 2020; McKinney, 2006). They can also easily exist alongside humans in a way that larger mesopredators can’t (Crooks, 2002; Tigas et al., 2002). Deer are very abundant in both environments due to the extinction of local predators and hunting laws favorable to them (Chollet & Martin, 2013). They naturally thrive in edge habitats where they can graze, so fragmentation is likely to affect them positively (Chollet & Martin, 2013). It makes sense that as the meadow cameras get closer to the road and campus, we see an abundance increase led by deer and squirrels, two species that do not shy away from human development.

Predators in the Blind Brook Forest were mainly active at dawn, dusk, and night. Prey were more active during the day. This correlates with other fragmentation studies, which found that bobcats and coyotes in fragmented habitats remain crepuscular but shift their activity away from the day and towards the night in fragmented habitats (Riley et al., 2003; Tigas et al., 2002). This suggests wariness of human activity that these predators have learned. With that said, we would have had to observe their behavior prior to the cut to confirm this. The fact is that most species in this study are expected to be more active at dawn, dusk, or night, and we observed behavior essentially corresponding to those patterns.

We found no significant difference in richness or diversity, something that could indicate a healthy, undamaged ecosystem. In fact, the Hansen paper we modeled our study on implies that both of our environments might be richer and more diverse than a rural forest (2020). Regardless of how well the forest is doing as a whole, though, it is relieving to find that the cut area is not less biodiverse than the forest (at least in terms of mesopredator mammals). If there was significantly less biodiversity, it would mean that lasting harm was done to the stability of that entire edge area.

The presence of bobcats is another good indicator of the forest’s health. We did not expect to find them in the relatively urbanized Westchester County, just north of New York City. They are more sensitive than most animals, including coyotes, to fragmentation and human contact (Crooks, 2002). We even spotted pairs of bobcats moving together, meaning we likely have a breeding pair in the area. This is a good sign because adult female bobcats are the least adventurous of them all, very wary of close proximity to humans. Their presence indicates that there isn’t a major disturbance (Riley et al., 2003; Tigas et al., 2002). The bobcats’ perseverance may come from the slit shape and small size of the cut. These leave the landscape of Blind Brook well-connected, an integral component of maintaining bobcat populations (Crooks, 2002).

One aspect of our study clearly entails greater rigor in future endeavors: trail cam placement. Our cameras did not have a uniform height or angle and thus did not have a uniform field of view. Cameras with a smaller field of view are likely to photograph fewer animals, or smaller animals. A difference extreme enough could have skewed all of our data to the point of unusability, as was the case with the Forest4 camera. This suggests that subtle skewing occurred amongst the cameras we did use. It is very important to note that a bobcat reduced Meadow1’s field of view a little less than halfway through the study on the night of October 8th, knocking it down to a steep angle. This could explain why Meadow1 captured the fewest mammals out of all the cameras. The trend we see of abundance increasing from Meadow1 to Meadow3 may simply be a result of this incident.

Hansen et al. went the extra mile to glean interesting data, setting control cameras away from their study sites. They also had their cameras up for a week longer than our study (2020). There are other, possibly better ways of setting the cams: Hansen et al. set theirs to take a burst of 5 photos spaced 1 second apart each time triggered. We set ours similarly but without burst photos. We must standardize our camera placement in future studies, and animal-proof them.

There are many available routes for additional study. We could place cameras at animal hotspots, like a brook crossing. We could take observations in other seasons for comparison. We could also find a way to accurately measure bird presence and include them in our statistics. Much of the wildlife we observed feeds on flora, something we didn’t include at all in our study. Our goal is to maintain the health of the local ecosystem, so it’s important to note how native and invasive plants are growing, how much native animals are eating each type, and even whether animals are overeating foliage, which can have a marked negative impact on certain species (Chollet & Martin, 2013).

We aim to improve our understanding of this complex interspecies ecology, though the forces that govern wildlife interaction with each other and human influence in their environment can’t be perfectly understood. We did find that thoughtful restoration of altered habitats remains of the utmost importance, as even in mildly fragmented areas like our Blind Brook restoration plot mammals preferred to be active in the large, unbroken reserve. If we are to coexist with mammals, this study proves that the best thing we can do is allow them to exist in the safety of their natural cover. Patches of green space that may seem to humans like healthy nature are not sufficient. Luckily, our cut is small and the forest retains connectivity. This allowed species richness and diversity to stay roughly the same in the disturbed and undisturbed areas, and the sleep-wake cycles we observed approximate what they would be naturally. On this campus we are working to assuage damage that has been done and spread the message of land restoration to others.

I’d like to thank SUNY Purchase for building the senior living home on this campus that made the study possible, Dr. Allyson Jackson for designing the study and guiding us through the project, my classmates for helping score the data, and my group who helped score the data all over again. Thanks so much to the animals for bearing with us and being good sports about all this.

Chollet, S., & Martin, J. (2013). Declining woodland birds in North America: Should we blame Bambi? Diversity and Distributions, 19(4), 481–483.  https://doi.org/10.1111/ddi.12003 

Crooks, K. R. (2002). Relative Sensitivities of Mammalian Carnivores to Habitat Fragmentation. Conservation Biology, 16(2), 488–502.  https://doi.org/10.1046/j.1523-1739.2002.00386.x 

Foley, J. A., DeFries, R., Asner, G. P., Barford, C., & al, et. (2005). Global Consequences of Land Use. Science, 309(5734), 570–574.

Hansen, C. P., Parsons, A. W., Kays, R., & Millspaugh, J. J. (2020). Does Use of Backyard Resources Explain the Abundance of Urban Wildlife? Frontiers in Ecology and Evolution, 8, 570771.  https://doi.org/10.3389/fevo.2020.570771 

Laurance, W. F., Delamônica, P., Laurance, S. G., Vasconcelos, H. L., & Lovejoy, T. E. (2000). Rainforest fragmentation kills big trees. Nature, 404(6780), 836–836.  https://doi.org/10.1038/35009032 

Laurance, W. F., Ferreira, L. V., Rankin-de Merona, J. M., & Laurance, S. G. (1998). Rain forest fragmentation and the dynamics of Amazonian tree communities. Ecology, 79(6), 2032–2040.

McKinney, M. L. (2006). Urbanization as a major cause of biotic homogenization. Biological Conservation, 127(3), 247–260.  https://doi.org/10.1016/j.biocon.2005.09.005 

Riley, S. P. D., Sauvajot, R. M., Fuller, T. K., York, E. C., Kamradt, D. A., Bromley, C., & Wayne, R. K. (2003). Effects of Urbanization and Habitat Fragmentation on Bobcats and Coyotes in Southern California. Conservation Biology, 17(2), 566–576.  https://doi.org/10.1046/j.1523-1739.2003.01458.x 

Tigas, L. A., Van Vuren, D. H., & Sauvajot, R. M. (2002). Behavioral responses of bobcats and coyotes to habitat fragmentation and corridors in an urban environment. Biological Conservation, 108(3), 299–306.  https://doi.org/10.1016/S0006-3207(02)00120-9