Modelling the Distributions and Richness of Lowland Savanna Plant Species in Belize using a Maximum-Entropy Approach.
View/ Open
F_Rollings_Full_Dissertation.pdf (5.988Mb)
Date
29/11/2012Item status
Restricted AccessAuthor
Rollings, Felicity J.
Metadata
Abstract
Maxent was used to estimate the species richness of Belizean lowland savanna. Using historical records and data from the ‘Darwin Project’, 102 individual species were modelled as well as six endemic/rare groups. Six environmental variables were used as inputs, representing temperature and precipitation patterns, slope and soil types.
Three methods of measuring species richness were identified. Output suitability maps were summed in two groups; endemic/rare species and all others. Using threshold values calculated by Maxent, binary presence-absence maps were created and summed in the same two groups. Niche breadths for each layer were calculated and the results applied as weightings to the suitability maps prior to their summation. The results were validated and compared, and showed similar patterns of suitability/species richness. However, it is emphasised that studies should aim to use combinations of the measurements, so as not to overlook potentially important areas for conservation which may be shown by just one measurement.
Maxent indicates the most important variables for determining species distributions; results suggested soil was the most influential parameter affecting lowland savanna species. However, this could be due to its relatively high spatial resolution.
Species richness is just one factor considered by conservationists. Threats and existing protected areas are also important. As an example, agricultural suitability was used as a proxy for threat, and analyses showed the most species-rich savanna as the least suitable areas for agriculture. In addition, the current protected areas were compared against the proposed areas for conservation, and showed overlap of up to 26%.
Collections
The following license files are associated with this item: