Stochastic Modeling for Geographic Diversity of Plant Species Richness in South Africa

University Of Connecticut, Storrs CT

Investigators

Abstract

0089801. John Silander. This proposal addresses two long-standing, fundamental, and interrelated questions in biology: (1) what are the environmental factors that determine the distribution of species, and (2) what are the ecological and biogeographic mechanisms that cause spatial variation in species richness? To answer these questions, the proposed research takes advantage of a well-documented database concerning the species-rich Protea flora from the Cape Floristic Kingdom of South Africa, a global hot spot of biodiversity. The PIs will integrate distributional data into a geographic information system (GIS), and subsequently analyze them from two quantitative perspectives, including hierarchical or Bayesian models and hybrid models based on cellular automata and simulation. Modeling will be expansive and include considerations of environmental variables, species attributes (including phylogenetic relationships), explicit spatial location (not simply ranges), and population dynamic (e.g., vital rates of birth, death, and migration) as well as evolutionary characteristics (e.g., speciation) of taxa. The full analytical protocol will include model construction, validation, and comparison. Modeling will facilitate a mechanistic understanding and the prediction of where a particular species will occur, and how such species-specific mechanisms give rise to broader-scale patterns of species richness.