Temporal Changes in the Vertebrate Scavenger Community in Association with Wildebeest Carcasses

Finding a Project

Ever since I was a little kid, I have loved playing outside, interacting with animals, and enjoying nature. As I grew older, this love evolved into an interest in biology and ecology. I specifically enjoy ecology because of all the relationships it highlights between vastly different organisms. I read some articles by Dr. David Post at Yale University and discovered that a lot of his work involved researching the relationship between aquatic and terrestrial ecosystems. After contacting him, I discussed with him his group’s research into the Serengeti Mara ecosystem in Kenya. On various research trips, they had set up an automated game camera that had taken pictures at locations on the Mara River. They had thousands of pictures to be analyzed, and together we decided this would make an interesting and insightful research project.



Ecosystems, both aquatic and terrestrial, have been researched for decades. An ecosystem is merely defined as a community of living organisms and its environment. This implies that factors within an ecosystem exclusively interact with other factors in this one ecosystem, and therefore each ecosystem never interacts with another. However, there are intricate relationships between ecosystems, one example being the relationship between aquatic and terrestrial ecosystems.

A herd of wildebeest crosses the Mara River. (click to enlarge)

The Serengeti Mara ecosystem is an example of such a relationship. During seasonal migrations, herds of wildebeest cross the Mara River. Due to factors such as varying river depth, approximately 7,000 wildebeest die each year in mass drownings while crossing the river. This leaves a buildup of carcasses in the river that are exposed to scavengers.

A study of scavengers at cheetah kill sites found that the time of day was not a significant factor for changes in the scavenger community. In grassland locations, the presence of small vultures was positively correlated with that of large vultures (Hunter, 2006). Research on hyena feeding behaviors showed that hyenas only consume wildebeest and other ungulate carcasses if the opportunity presents itself; they do not specifically look for carrion (Cooper, 1999). Competition within an avian scavenger guild was investigated in Peru. This study found that the dominance of a certain species could determine its access to the carcass. The species of avian scavengers with the greatest body mass were typically dominant (Wallace, 1987).

We are not aware of any past studies that analyzed the sequence and patterns of the scavenger community regarding the carcasses remaining from mass drownings in the Serengeti Mara ecosystem. Unlike other studies, the carcasses in this study were not placed in certain locations by researchers and thus show the natural occurrence of decomposition in this ecosystem.

Wildebeest carcasses draw scavengers such as the marabou stork. (click to enlarge)

The purpose of this research is to determine which scavengers are consuming the carcasses and what factors contribute to possible patterns regarding temporal changes in the scavenger community. In the future, this data can help determine how nutrients rotate between aquatic and terrestrial ecosystems, and thus help to show how one ecosystem affects surrounding ones. I hypothesized that different scavenger characteristics will affect the temporal pattern of when scavengers are present.

Common avian scavengers identified during this study include the marabou stork (Leptoptilos crumeniferus), the white-backed vulture (Gyps africanus), the Ruppel’s griffon vulture (Gyps rueppellii), the hooded vulture (Necrosyrtes monachus), and the sacred ibis (Threskiornis aethiopicus). Other identified scavengers included hyenas, crocodiles, mongooses, and hippopotamuses.




  • Photos taken by mentors with an automated game camera at a fixed position on the Mara River
  • Vertebrate identification key
  • Carcass map, created at the site on the Mara River
  • Excel spreadsheet
  • Elements hard drive
  • JMP software



Automatic game cameras took photographs every fifteen minutes. The different scavenger species were identified using a vertebrate identification key. (click to enlarge)

My mentors, Dr. Post and Amanda Subalusky, set up an automated game camera. Pictures were taken approximately every 15 minutes at a fixed location on the Mara River in Kenya for 24 hours per day starting on November 18, 2012 (five days after a mass drowning event) and concluding on November 30, for a total of 1,273 pictures. I learned how to identify the different species in the pictures from the vertebrate identification key. In addition to the five common avian scavengers previously listed, less common scavengers identified including the yellow-billed stork (Mycteria ibis), Egyptian goose (Alopochen aegyptiaca), lappet-faced vulture (Torgos tracheliotus), crocodile, Nile monitor (Varanus niloticus), hippopotamus (Hippopotamus amphibious), banded mongoose (Mungos mungo), dwarf mongoose (Helogale parvula), slender mongoose (Galerella sanguinea), and hyena.

Katherine analyzing one of 1,273 photographs. (click to enlarge)

Analysis of each picture included collecting data on the temperature, species of birds present and number of individuals for each species, number of individuals for non-avian species, and any other rare observations such as camera movement, light exposure, or unidentifiable animals. The number of avian scavengers consuming the carcasses was collected as well. A scavenger was considered to be consuming if it was sitting atop or in close proximity to a carcass, with its head bent down toward the carcass. The percentage of avian scavengers eating within the frame was calculated as a mean based on the total number of birds present.

A preliminary data analysis showed that there was little variation in the species and number of common scavengers between 15-minute and hourly intervals; therefore, we decided to use photographs taken on the hour for every hour for the analysis of the common scavengers. However, rare scavengers were analyzed in every picture in which they appeared.



The Average and Maximum Number of the Five Most Common Avian Scavengers From Day 5 through 17 After the Drowning Event
  Marabou Stork White Backed Vulture Ruppell's Griffon Sacred Ibis Hooded Vulture
Day Avg. Max Avg. Max Avg. Max Avg. Max Avg. Max
5 7.44 17 2.56 6 2.56 9 0.33 3 0 0
6 7.85 15 4.46 11 7.77 20 0 0 0 0
7 10 16 3.15 11 3.08 9 0 0 0.77 4
8 8.29 14 4.93 13 2.92 9 0.07 1 0.29 1
9 10.31 19 3.62 9 1.54 4 0 0 1 3
10 4.08 11 1.08 5 1.77 9 0.08 1 0.23 1
11 7.85 17 5.15 13 3.69 11 0.08 1 1 3
12 1.85 9 0.15 1 0.15 2 0 0 0.85 4
13 1.31 5 0.38 2 0.46 2 0 0 0.46 4
14 0.54 3 0 0 0 0 0 0 0.54 4
15 0.31 2 0 0 0 0 0.77 3 0 0
16 0.38 1 0 0 0 0 1 5 0.69 5
17 0.63 3 0 0 0 0 0.88 6 0 0

Table 1: Averages the maximum and presence of common scavengers calculated once an hour between 6:00 and 18:00 each day (1-13).

As shown in Table 1 and Figure 1, the marabou stork had the greatest presence out of the five common scavengers until Day 14. Days 5 through 11 had similar characteristics, and Days 12 through 17 had significantly reduced numbers of avian scavengers. From Day 12 on, the majority of the carcasses were showing bones. 


Figure 1: A comparison of the averages for the common scavengers calculated once an hour between 6:00 and 18:00 for each day (5-17)

The hooded vulture did not appear until Day 7. By Day 14, both the white-backed vulture and the Ruppell’s griffon vulture stopped appearing at the wildebeest carcass sites. The percentage of avian scavengers feeding on the carcasses, as opposed to sitting near them, remained relatively constant throughout Day 5 through 17.

While the total number of the five common avian scavengers dropped significantly at the end (r2 = 0.7037), the composition of the avian scavengers also changed drastically (Figure 2). On Day 5 through 13, about half of the avian scavengers were marabou storks. By Day 14, sacred ibis were the most common.


Figure 2: A presentation of the daily proportions for the five common avian scavengers throughout the series as represented in percentages of the total. (Daily Species Total/Daily Overall Total)

Avian scavengers were more common compared to non-avian scavengers (including terrestrial and aquatic). Avian scavengers only appeared during daylight hours (6:00 to 18:00). Non-avian scavengers appeared throughout the 24-hour period, with hyenas being the most frequent of the non-avian scavengers. They occurred mainly during the night. As time went on, more non-avian scavengers were present at the carcasses (r2 = 0.4797).


As shown in Figures 5, 6, and 7, the marabou stork appeared the most during the early morning and late afternoon. The white-backed vulture appeared mainly around midday, while the sacred ibis was present throughout the day. Data collected on temperature showed that there was minimal to no variance across the time period; however, nights were significantly colder than days.



The results suggest that the largest community of scavengers is present during the first 11 days of access, when the carcasses contain the most meat. The longer the amount of time that has passed, the less meat is available to the scavengers. By Day 12, when bones were visible on the carcasses, the three most common scavengers dropped in number significantly (p = 0.0003). Hooded vultures remained at a relatively constant average across the days, suggesting they do not have a preference for which part of the carcass they consume.

The largest of the avian scavengers, the marabou stork, was the most common scavenger for the majority of the time at the carcasses. Given that it has the greatest body mass, of 9 kilograms, the marabou stork most likely had dominance over the scavenger guild and therefore remained at the carcasses for the longest amount of time (Wallace, 1987). The smallest avian scavenger, the sacred ibis (1.35 kilograms), would therefore have the least amount of dominance over the other avian scavengers at the carcass site. This can explain why it was more frequent in the second half of the data (Day 12 through 17) once the larger vultures had left the carcasses. Furthermore, sacred ibises feed on the insects in the carcasses and not on the actual meat of the carcass. This means that the sacred ibis must wait for the larger scavengers to open up the carcasses, exposing them to other scavengers and decomposers such as insects. Once the insects arrive at the carcasses, the sacred ibises will be more frequent.

Ten days after the mass drowning event, there was a dramatic drop in the total number of avian scavengers, which recovered by Day 11. On Day 10, there were three crocodiles recorded, fighting and splashing over carcasses near the scavengers. This could have frightened the birds, though they regrouped by the following day once the crocodiles were gone. The avian scavengers returned to continue to feed on the carcasses.

Non-avian scavengers did not appear until eight days after the mass drowning event, while avian scavengers were abundant by Day 5 (p = 0.0087). Studies on avian scavengers have shown that they discover carcass sites through their olfactory senses and flight navigation. Their ability to fly over a large area of terrain increases their chances of finding carcasses (Devault, Travis, Rhodes & Shivik, 2003). This increased chance of finding carcasses suggests that avian scavengers appeared at the carcass site prior to non-avian scavengers simply because they found it faster through flight.

Therefore, these data support my hypothesis that different scavenger characteristics affect when different scavenger species would appear at the carcass site. Since avian scavengers have the ability to locate carcasses through flight, they were at the carcasses earlier than scavengers that cannot fly. Smaller avian scavengers appeared after larger ones, suggesting dominance in the scavenger guild.

The procedure of this study contained several minor limitations. Many of the limitations of this study are due to the fact that the experiment was done completely through picture analysis. This meant that the view of the location was restricted to the small frame of the camera and the animals above the surface of the water. If crocodiles and/or fish were present under the surface of the river, they could not be seen from the camera angle and thus were not analyzed. In addition, the pictures often had poor light exposure, and species identification could have been mistaken. To account for this, any slightly ambiguous scavenger was recorded as unknown. Because of this, the data may have been skewed slightly, as unknown birds were not identified by their correct species.

Another error was that on the first day of the second section of the data (Day 12), when the number of avian scavengers dropped dramatically, a human was seen in the frame of one of the pictures. The human was thought to be a fisherman at the river, suggesting that there had been minor human interference at the site on the Mara River. One explanation could be that the presence of humans scared away some of the scavengers. Because the scavengers may have then associated the location with humans, this could have altered the natural temporal changes of the scavenger community. However, humans are not thought to be the cause of the dramatic decline of total avian scavengers in the second section of the data, as humans were also present to set up the camera, during which time a significant drop in scavengers was not witnessed.

These data contribute to the knowledge of what happens when an organism dies. Specifically, for these wildebeest, avian scavengers are the greatest scavengers, helping to return nutrients back into the surrounding ecosystem. Now that the scavenger species are known, future research might entail tracking them to discover to where the nutrients from the wildebeest carcasses are carried. This will help determine how aquatic and terrestrial ecosystems interact, as the avian scavengers may deposit the nutrients from the carcasses over a vast terrestrial area, or into the aquatic ecosystem in the Mara River. Aquatic scavengers such as the crocodile may deposit the nutrients into the river as well. Terrestrial scavengers such as hyenas, on the other hand, may deposit the nutrients across the surrounding terrestrial areas. These ecosystems depend on nutrients, and being able to determine where the scavengers identified in this project carry the nutrients will help uncover the relationship between the aquatic and terrestrial ecosystems through the recycling of nutrients.

Furthermore, the three vultures studied in this project, the white-backed vulture, the Ruppell’s griffon vulture, and the hooded vulture, are identified as endangered animals. Information from this project contributes to knowledge about these endangered animals, and can help predict the effects on the Mara Serengeti ecosystem if one of these species were to go extinct.

In the future, we plan to apply this procedure to additional carcass sites on the Mara River. We have pictures taken in September 2013 on a different location on the Mara River as well as another series from October 2013. Repeating this procedure at two different carcass sites will allow us to determine the validity of the findings from this study.



This project was completed under the mentorship of Dr. David Post and Amanda Subalusky at Yale University. Ms. Deborah Day was the teacher advisor for this project.



Cooper, Susan M., Kay E. Holekamp, and Laura Smale. “A Seasonal Feast: Long-Termed Analysis of Feeding Behaviour in the Spotted Hyena.” African Journal of Ecology 37 (1999): 149-60. Web. 4 Jan. 2015.

DeVault, Travis L., I. L. Brisbin, Jr., and Olin E. Rhodes, Jr. “Factors Influencing the Acquisition of Rodent Carrion by Vertebrate Scavengers and Decomposers.” NRC Research Press Web (May 2004). Web. 28 Oct. 2014.

DeVault, Travis L., Olin E. Rhodes, Jr., and John A. Shivik. “Scavenging by Vertebrates: Behavioral, Ecological, and Evolutionary Perspectives on an Important Energy Transfer Pathway in Terrestrial Ecosystems.” Wiley 102 (2003): 225-34. Web. 27 Oct. 2014.

Field, Rachel D., and John D. Reynolds. “Ecological Links Between Salmon, Large Carnivore Predation, and Scavenging Birds.” Journal of Avian Biology 44 (2013): 9-16. Web. 12 Sept. 2014.

Hunter, J. S., S. M. Durant, and T. M. Caro. “Patterns of Scavenger Arrival at Cheetah Kills in Serengeti National Park Tanzania.” African Journal of Ecology 45 (2006): 275-81. Web. 19 Jan. 2015.

Klindall, Corine J. “The Early Bird Gets the Carcass: Temporal Segregation and Its Effects on Foraging Success in Avian Scavengers.” American Ornithologists’ Union 131 (2014): 12-19. Web. 30 Oct. 2014.

Melis, Claudia, Nuria Selva, Ivanne Teurlings, Christina Starpe, John P. C. Cinnell, and Reidar Anderson. “Soil and Vegetation Nutrient Response to Bison Carcasses in Bialowieza Primeval Forest, Poland.” Ecological Research 22 (2007): 807-13. ICONN. Web. 15 Sept. 2014.

Schuldt, Jeffrey A., and Anne E. Hershey. “Effect of Salmon Carcass Decomposition on Lake Superior Tributary Streams.” Journal of the North American Benthological Society June 14.2 (1995): 259-68. Web. 14 Jan. 2015.

Science Daily. “Wolves, Moose, and Biodiversity: An Unexpected Condition.” Science Daily, 3 Nov. 2009. Web. 14 Jan. 2015.

Wallace, Michael P., and Stanley A. Temple. “Competitive Interactions Within and Between Species in a Guild of Avian Scavengers.” The Auk 104 (April 1987): 290-95. Web. 19 Jan. 2015.