Comparative Analysis of Correlative Modeling Methods in Predicting North American Bird Abundance

Area of Science:

• Ecology
• Biodiversity Science
• Computational Biology

Background:

• Correlative ecological niche models (ENMs) are increasingly used to estimate species abundance via environmental suitability scores.
• Assumptions and conceptual differences among ENM methods and their impact on abundance predictions remain under-explored.

Purpose of the Study:

• To compare the performance of nine ENM methods in predicting avian abundance in North America.
• To evaluate the influence of different data inputs (distribution vs. breeding area, annual vs. seasonal variables) on prediction accuracy.

Main Methods:

• Utilized standardized bird count data for North American species.
• Compared nine distinct ENM approaches, including recent advancements.
• Assessed relationships between environmental suitability scores and species abundance.

Main Results:

• Generally positive correlations (avg. r > 0.18) between environmental suitability and abundance were observed.
• Newer ENM methods (MaxLike, Gaussian) exhibited marginally higher average correlations than older methods.
• Prediction accuracy was consistent across taxonomic groups and largely unaffected by data simplification (breeding area, seasonal variables).

Conclusions:

• ENMs offer a reliable tool for inferring species abundance patterns.
• Environmental suitability partially explains abundance, with other factors like biotic interactions also playing a role.
• Simpler data inputs can be adequate for ENM-based abundance predictions, enhancing accessibility and application.