Species abundance, urn models, and neutrality

Area of Science:

• Ecology
• Biogeography
• Computational Biology

Background:

• The neutral theory of biodiversity and biogeography is a cornerstone of community ecology research.
• Applying neutral models to large regional datasets requires robust analytical tools.

Purpose of the Study:

• To present and apply three complementary neutral models to regional species abundance data.
• To introduce the neutr R package for fitting neutral models to ecological data.
• To investigate regional diversity patterns in tropical forests using neutral models.

Main Methods:

• Utilized exact sampling formulas for three neutral models: Ewens canonical, subdivided communities, and diversity begets diversity.
• Employed the new R package 'neutr' to fit species abundance data from regional tropical forest inventories.
• Analyzed data from the Amazon, tropical Africa, and Southeast Asia.

Main Results:

• Neutral models demonstrated good fit to empirical species abundance data, except for the most abundant species.
• Regional diversity (parameter θ) was highest in Southeast Asia (726) and Amazonia (654), and lower in tropical Africa (219).
• Inferred species richness: Amazonia (10,141), Southeast Asia (9,915), and tropical Africa (3,477).
• The spatially subdivided neutral model indicated significant spatial substructure in all three regional floras.

Conclusions:

• Neutral models effectively explain regional species abundance patterns and provide insights into species pools.
• The study highlights significant differences in regional tree diversity across Amazonia, tropical Africa, and Southeast Asia.
• Evidence of spatial substructure suggests non-random distribution of species within these tropical forest ecosystems.