OOS 8-9
Causes and consequences of wildebeest mass drownings in the Mara River, Kenya

Monday, August 10, 2015: 4:20 PM
329, Baltimore Convention Center
Amanda L. Subalusky, Ecology and Evolutionary Biology, Yale University, New Haven, CT
Christopher Dutton, Ecology and Evolutionary Biology, Yale University, New Haven, CT
Emma Rosi-Marshall, Cary Institute of Ecosystem Studies, Millbrook, NY
David M. Post, Ecology and Evolutionary Biology, Yale University, New Haven, CT
Background/Question/Methods

Mass drowning events have been documented across a range of animals, including dinosaurs, bison, and caribou. Historically, these events may have occurred more frequently, when intact migrations and unregulated rivers were more common features of many landscapes. These large pulses of high quality resources may have significant impacts on aquatic ecosystems when and where they occur. In the Serengeti Mara Ecosystem, Kenya-Tanzania, mass drownings still occur relatively frequently during crossings of the Mara River undertaken by migrating herds of ~1.3 million wildebeest. We conducted surveys of drowning events from 2011-2014, in which we quantified the biomass, elemental composition, and decomposition rate of wildebeest carcasses. We used in situmeasurements and experiments to measure the impact of carcasses on ecosystem processes, including nutrient limitation, nutrient uptake, and primary production. We also investigated potential causes of drowning events, including river discharge level, crossing site geomorphology, and tourist behavior.

Results/Conclusions

In four years of surveys, we documented twenty major drowning events, ranging in size from 45 to 4,900 individuals, with an average of 5 drownings and 6,900 carcasses per year. Occurrence of drowning events was positively related with discharge, and was strongly influenced by crossing site geomorphology. Carcasses are high quality resources (53 C: 10 N: 1 P), and drowning events contribute an average of 422 tons C, 90 tons N, and 21 tons P to the river each year. Different portions of the carcass decompose at different rates, ranging from <20 days for muscle to >1 year for bone. Carcass inputs increased ammonium (111%) and phosphate (45%) uptake length, altered patterns of nutrient limitation, and significantly increased primary production in experimental streams. Measurements of nutrient flux from carcasses showed that less than 5% of predicted loading is accounted for after nearly one month of decomposition, suggesting a majority of TN and TP is either being taken up by river ecosystem processes, ingested by instream consumers, or transported out of the river system by mobile consumers such as vultures and hyenas. Mass drowning events, due to the magnitude of high quality resources accompanying them, have the ability to significantly alter recipient ecosystem function, despite their stochastic occurrence.