Computer modelling the Serengeti-Mara ecosystem

At present, the viability of biodiversity in most of the remaining natural areas of the world is primarily threatened by human encroachment. This has led to an increased demand for active conservation. However, in order to devise and implement appropriate management strategies for a particular area, a specific understanding of ecosystem function is required. Creating a simulation model using available research data may provide a way to achieve this. In this thesis, the construction of a comprehensive model delineating the dynamics of the Serengeti-Mara ecosystem is initiated. Using the abundance of research data collected on this ecosystem over the last 40 years, the processes involved in setting-up such a model are investigated. First, a basic foundation, accommodating the spatial and temporal variation in climate and physiography across the Serengeti region, is established. The relationship between grass growth and rainfall is then incorporated, along with the mechanisms concerned with limiting grass availability, the subsequent survival and recruitment of grazing herbivores and finally, the influence of predation upon those herbivores. The model, even in these early stages of development, adequately depicted dynamics equivalent to those in the Serengeti-Mara ecosystem, indicating that the methods used were appropriate. It was found that grass availability was not the primary factor influencing the overall dynamics of grazing herbivores within the ecosystem, and only migratory wildebeest appeared to be strongly influenced by this factor throughout the time-scale of the model. It was suggested that other factors were responsible for regulating the majority of herbivore populations. By identifying where further research is required to increase our understanding of this particular ecosystem's function, the model demonstrates its effectiveness as an analytical tool. For the long-term conservation of the Serengeti-Mara ecosystem and other similar ecosystems, this reveals that the construction of such models is certainly beneficial, if not essential.

Name dc:type.qualificationname

Ph.D

Level dc:type.qualificationlevel

doctoral

Grantor dc:publisher.institution

University of Leeds

Year dc:date.issued

2003

Author dc:creator

• Bennett, Victoria Jane

Advisor dc:contributor.advisor

• Shorrocks, B.

Identifier

uk.bl.ethos.401093

OAI identifier oai:identifier

oai:etheses.whiterose.ac.uk:1553