Analysis of the Composition of Mammalian Species in an Urban, Fragmented Area in Southern California

Part of the Young Naturalist Awards Curriculum Collection.

by Tanner, Grade 12, California - 2013 YNA Winner


I have always loved animals and the outdoors. I live in a rural area in the Santa Monica Mountains, just outside of Los Angeles, California, and wildlife roams my backyard. Eager to learn what wildlife I could find in my backyard, I bought my first “camera trap,” a motion-sensing camera, when I was 13 years old. From the moment I discovered a photo of a mountain lion on the camera I placed in my backyard, nothing could stop my pursuit to get more and better footage of wildlife with a trail camera. I took tracking classes so I could identify species by their tracks left on a trail. I discovered that species such as mountain lions, bobcats, coyotes, gray foxes, opossums and raccoons could all be found in the mountains around my house! I became an expert at identifying by tracks alone how much animal traffic was on a trail and which species were roaming the trails. I have spent years experimenting with the best way to set up my wildlife cameras so I could get pictures of the animals leaving the tracks I found. To my amazement, changing a camera’s position by just a slight angle could maximize the number of photos the camera could capture. I learned to observe my environment more carefully as I hiked around from location to location. Each camera check was a thrill on its own, whether it was finding a photo of a rare species or interesting behavioral footage on the camera. Although my backyard is in a wild part of the Santa Monica Mountains, I live only 40 miles from downtown Los Angeles, so there are many regions of the mountains fragmented by urbanization. I became very interested in the effect of urbanization on the species I was able to discover near my house. This led me to start my own study comparing the species that live in fragmented urban areas compared with more pristine, natural landscapes. 

Background Research

Two-thirds of the world’s terrestrial land is now used to support human populations through agriculture, fisheries, urbanization or infrastructure (Millennium Ecosystem Assessment 2005), so managing the consequences that arise as a result of human development is a global issue. The destruction of habitat resulting from urban development is considered one of the most serious threats to biological diversity in the world (Wilcove et al. 1998). However, some species are more sensitive to the effects of urbanization than others (Crooks 2002; Ordenana et al. 2010). For example, large mammals, especially mammalian carnivores, are considered among the species most sensitive to urban development (Crooks 2002). Mammalian carnivore species often have large home ranges, and some are solitary and territorial. These species live in low-density populations, with few animals in a particular area. When urbanization fragments a habitat, the wildlife populations are also fragmented into areas with few individuals, increasing the probability of local population extinctions (Crooks 2002). The effects of urbanization can therefore extend to affecting mammalian community composition (Crooks 2002; Ordenana et al. 2010). Studying the impacts of urbanization on carnivores is extremely important to ensure the overall well being of each species. Because they are sensitive to urbanization, they can sometimes be used as indicators of ecosystem health.

Certain species are considered adaptable to urban environments. For example, coyotes and raccoons can have higher population densities in urban areas than in rural areas (Prange and Gehrt 2004; Gehrt and Riley 2010). One reason these species are more adaptable is because they are omnivores (Gehrt and Riley 2010; Hadidian et al. 2010; Bateman and Fleming 2012) and generally able to exploit developed areas (Gehrt and Riley 2010; Hadidian et al. 2010; Batemen and Fleming, 2012). Raccoons, for instance, generally appear to be able to find refuge or den sites in more urban settings (Hadidian et al. 2010). Also, higher coyote densities can result from urban development because the species may have easier food access and less threat from larger carnivores that are more sensitive to urbanization (Gehrt and Riley 2010; Bateman and Fleming 2012).

Some species are less adaptable and are considered more sensitive to urban development. For example, gray foxes in Southern California can be absent from some large habitat fragments where coyote densities are higher, likely because they are being out-competed by coyotes (Fedriani et al. 2000). Large carnivores like mountain lions are observed less frequently in fragmented habitats or human-altered land cover; live primarily in larger patches of habitat (Crooks 2002) or more remote areas; and often maintain a distance of one kilometer from urban areas (Beier et al. 2010). In general, animals that are strict carnivores, like bobcats and mountain lions, tend to be more sensitive to urbanization than omnivores, probably because their diets are less flexible (Bateman and Fleming 2011).

Study Question

For my study, I wanted to examine mammalian species composition in an urban park compared to a remote, more pristine and less altered habitat. I was especially interested in mammalian carnivores since they can be used as indicators of ecological health (Crooks 2002). I hypothesized that the mammalian species composition observed in an urban park compared with a pristine natural park would be different. I predicted that in a fragmented habitat in an urban landscape, I would more frequently observe species that are considered adaptable to urbanization such as opossums, coyotes and raccoons. Second, I predicted that in the more remote, pristine park, I would more frequently observe animals such as gray foxes, mountain lions and bobcats that are more sensitive to urbanization.

Materials and Methods
Map of the San Gabriel Mountains area near Los Angeles, showing the locations of: the camera, the Verdugo Mountains, Freeway 210, and Freeway 5.
Figure 1. Map illustrating my study area. The stars show where cameras were placed. the San Gabriel Mountains represented my "control" natural study area while the Verdugo Mountains is an isolated urban partk.

Study Area

To conduct my experiment, I used remote wildlife cameras to make species observations. I placed one camera in an urban, fragmented habitat that I considered influenced by urbanization and one camera in a remote, pristine habitat further from urban development (Figure 1). My urban fragmented habitat study site was the Verdugo Mountains, 19 kilometers from downtown Los Angeles. The Verdugo Mountains are surrounded on all sides by urbanization and are cut off from a larger mountain range by a major freeway; the area covers 36.26 square kilometers (“Verdugo Mountains Open Space Preserve”). For my natural study area, I chose the San Gabriel Mountains, which are part of the Angeles National Forest, 28 kilometers from downtown Los Angeles and covering 2,590 square kilometers (Angeles National Forest). Southern California is host to several mammalian species, including coyote (Canis latrans), bobcat (Lynx rufus), striped skunk (Mephitis mephitis), raccoon (Procyon lotor), gray fox (Urocyon cinereoargenteus), mountain lion (Puma concolor), mule deer (Odocoileus hemionus) and opossum (Didelphis virginiana) (Orednana et al. 2010).

Both study locations are in Southern California and were historically part of the same continuous habitat, although the Verdugo Mountains are now fragmented from the San Gabriel Mountains by urban development. Therefore, I considered the San Gabriel Mountains as my control study area, where the mammalian species composition should be less affected by urbanization. Therefore, the species present in the area would represent a natural community composition. The Verdugo Mountains are surrounded by urban development, with no direct passage to any neighboring mountain ranges, and so represented a study location where the mammalian species composition would be affected by urban development

Wildlife Cameras
Figure 2. Photos of the cameras used for this study.

I used Bushnell remote-sensing wildlife cameras, or “camera traps,” to study what wildlife species were present in each habitat type (Figure 2). Camera traps are known as an excellent, non-invasive tool for observing wildlife (Long et al. 2008). These cameras are heat-activated and record videos or pictures stamped with the date and time that an animal walked in front of the camera. The Bushnell camera I used has a delay of one second; after the camera senses the movement of an animal, it takes one second for it to “wake up” and begin recording. The cameras were set to take 30-second videos, with a one-second interval in between. Each camera was first tested at home to ensure it was properly working. The recreational use of these cameras is permitted in the areas where I conducted my study.

Next, I chose a location within each park to place the cameras (Figure 1). I hiked extensively through each park to find the ideal spots to place each camera. Based on my knowledge of animal tracks and signs, I found trails and specific locations with high animal traffic. I placed the trail cameras in these high traffic areas, one in each park. The cameras were put into place for a total of four months (121 days) each.

Data Collection and Analysis

Over the four-month period, I checked the cameras every three to four weeks to recover the videos, ensure the cameras were still working, and change the camera batteries. Once I recovered the videos, I viewed each video to identify and record the species present in each video, and the date and time the footage was taken. I entered this data into an Excel spreadsheet in order to manage the data, create visualizations, and conduct analyses. After gathering all of the data over the four-month period, I made a bar graph of each study site to compare the number of video observations of each species. For statistical analyses, I used a Z-test to compare the proportion of observations for each species observed in each location. Each Z-test provided a P-value that I used to interpret the significance of my tests.

Figure 3. P-values for Z-tests for comparisons of the proportion of species observations made in each study area.


I collected 55 mammal videos from the San Gabriel Mountains camera and 34 videos from the Verdugo Mountains camera. I captured videos of the following species: skunk, bobcat, gray fox, opossum, raccoon, deer, and puma. With the exception of gray fox in the Verdugo Mountains and opossum in the San Gabriel Mountains, each species listed above was observed in both the San Gabriel and Verdugo Mountains study areas (Figures 3 and 4). Interestingly, coyotes were not observed in either location.

After analyzing the results between the San Gabriel Mountains and the Verdugo Mountains, I found significant differences between the two study areas for bobcat, gray fox, opossum, raccoon, and puma (Figure 3). Some were weakly significant (P < 0.1), while others were highly significant (P < 0.001).

Figure 4. Chart illustrating the percentage of total video observations captured using remote wildlife cameras in two study locations

In the Verdugo Mountains there were some species for which I captured significantly higher proportions of videos than in the San Gabriel Mountain location (Figures 3-6). These species include bobcat (P = 0.001) and opossum (P = 0.009) (Figure 3). In the San Gabriel Mountains there was a significantly higher number of gray fox observations (P = 0) compared to the Verdugo Mountains, where no gray fox photos were captured. Opossum observations were not captured in the San Gabriel Mountains but were observed in the Verdugo Mountains (P = 0.009). Observations of mountain lions were weakly significant (P = 0.06), with a slightly higher number of observations in the Verdugo Mountains. The difference in raccoon observations was also weakly significant (P = 0.06), and there were more observations in the Verdugo Mountains compared with the San Gabriel Mountains. The Verdugo Mountains camera and the San Gabriel Mountains camera had a relatively similar number of skunk observations, giving a non-significant result (P > 0.1). Deer observations also had a non-significant result (P > 0.1)

Discussion and Conclusion
Figure 5. Video stills of mountain lion (above) and grey fox (below) captured in the San Gabriel Mountains - the remote natural study area.

The objective of my study was to examine mammalian species composition in an urban park compared with a remote, more pristine habitat. The results of my study supported my hypothesis that the mammalian species composition and proportions of observations made for each species would differ between the two study areas. I did find significant differences in the species composition captured on camera in the urban study area compared with the more pristine study area. However, I also made several surprising observations.

First, I did not observe coyotes in either study area. Although no coyotes were observed in either of the study areas, I had expected to find evidence of this species in both locations, especially in the urbanized area because of their adaptability to urban environments (Fedriani et al. 2000; Riley et al. 2003; Gehrt and Riley 2010; Bateman and Fleming 2012). Further, I captured one video of a bobcat during the day in the Verdugo Mountains, showing diurnal activity for bobcats in the urbanized area (Figure 6). This finding is unusual because Riley et al. (2003) found that bobcats in Southern California that are found in urban parks tend to be more nocturnal to avoid humans. This observation suggests that bobcats may be less sensitive to urbanization or human activity in certain areas around Southern California. Finally, although I predicted more skunk observations in the urbanized study area because they are more adaptable to urban development, the proportion of skunk observations was not significantly different between the study areas.

Figure 6. Video stills of bobcat (above) and deer (below) captured in the Verdugo Mountains - the urban study area.

With my study, I also made observations that I had predicted based on my background research. There was a greater number of raccoon and opossum observations in the urbanized study area (the Verdugo Mountains) compared with the more remote natural area, which is not surprising because these animals are less sensitive to urbanized environments (Crooks 2002; Riley et al. 2003; Prange and Gerht 2004; Ordenana et al. 2010; Riley et al 2010; Hadidian et al. 2010). Using the San Gabriel Mountains camera, I observed a significant number of gray foxes. Since gray foxes in Southern California are considered very sensitive to fragmented habitats (Crooks 2002; Ordenana et al. 2010), this result is not surprising. There was also much more mule deer activity in the control area in the San Gabriel Mountains, which is important because these are important prey for animals like mountain lions (Beier et al. 2010). The data based on the pumas in the area were weakly significant, with a higher proportion of observations made the Verdugo Mountains. Although only weakly significant, this result is very surprising! I expected to observe mountain lions more frequently in the more remote study area because mountain lions tend to avoid urban areas and are primarily associated with habitat patches larger than the Verdugo Mountains (Crooks 2002; Beier et al. 2010; Ordenana et al. 2010). These data also show that even though the Verdugo Mountains are isolated by urbanization and the mammalian species composition seems to have lost some species like the gray fox, there is still a healthy carnivore community in the Verdugo Mountains.

Several factors likely influenced my results. First, I would improve this study by using more cameras set up in a controlled grid system across a wider geographic area, both in the Verdugo Mountains and the San Gabriel Mountains, to increase the probability of capturing more species observations. With more cameras I would have collected more data to analyze.

Although I was limited in the number of wildlife cameras available to use for this study, mine is the first scientific study in the Verdugo Mountains to examine mammal species composition. My results were both surprising and important in identifying the species present in this small urban park area. Some of my results, like the presence of mountain lions in a very urban park, are surprising, and it would be valuable to expand the study and identify possible wildlife corridors in the area. Wildlife corridors are important for maintaining habitat connectivity in the face of expanding urbanization. For wildlife species in the Verdugo Mountains, wildlife corridors will be critical for maintaining healthy populations and allowing species with large home ranges to continue to roam freely.


“Angeles National Forest.”United States Department of Agriculture.Retrieved from the World    Wide Web on 20 Feb 2013.

Bateman, P.W, and P.A. Fleming. “Big city life: carnivores in urban environments.”Journal of Zoology287 (2012): 1-23.

Beier, Paul, et al. “Mountain Lions (Puma concolor)” in: Urban Carnivores. Baltimore: Johns Hopkins University Press, 2010.

Crooks, Kevin R. “Relative Sensitivities of Mammalian Carnivores to Habitat Fragmentation.” Conservation Biology (2002): 488-502.

Fedriani, Jose M., et al. “Competition and intraguild predation among three sympatric carnivores.” Oecologia (2000): 258-270.

Gehrt, Stanley D. and Seth P.D. Riley. “Coyotes (Canis latrans)” in: Urban Carnivores. Baltimore: Johns Hopkins University Press, 2010.

Hadidian, John, et al. “Raccoons (Procyon lotor)” in: Urban Carnivores. Baltimore: Johns Hopkins University Press, 2010.

Long, Robert A., et al.Noninvasive Survey Methods for Carnivores.Washington, D.C.: Island Press, 2008.

Millenium Ecosystems Assessment. “Summary for Decision Makers.” In Ecosystems and Human Well-Being: Synthesis. Washington, D.C.: Island Press, 2005.

Ordenana, Miguel A., et al. “Effects of urbanization on carnivore species distribution and richness.” Journal of Mammalogy 91(6) (2010): 1322-1331.

Prange, Suzanne, and Stanley D. Gehrt. “Changes in mesopredator-community structure in response to urbanization.” Canadian Journal of Zoology (2004): 1804-1817.

Riley, Seth P.D., et al. “Effects of Urbanization and Habitat Fragmentation on Bobcats and Coyotes in Southern California.” Conservation Biology (2003): 566-576.

Riley, Seth P.D., et al. “Bobcats (Lynx rufus)” in: Urban Carnivores. Baltimore: John Hopkins University Press, 2010.

“Verdugo Mountains Open Space Preserve.” LA Mountains. Retrieved from the World Wide Web on 19 Feb 2013.

Wilcove, David S., et al. “Quantifying Threats to Imperiled Species in the United States.” Bioscience 48(8)(1998): 607-615.


I would like to thank Laurel Serieys, a Ph.D. candidate at the University of California, Los Angeles, for mentoring me throughout my study. Also, I am very thankful for my friend and mentor, Johanna Turner, who was able to help me in my camera-trapping techniques. Finally, I would like to thank my parents for always backing me with their full support and encouraging me along the way.