A semi-automatic approach to study population dynamics based on population pyramids.

Area of Science:

• Ecology
• Demography
• Computational Biology

Background:

• Population pyramids are visual tools for assessing population characteristics.
• Formalized algorithmic approaches for extracting information from population pyramids are limited.

Purpose of the Study:

• To present an algorithm-based classification of population pyramid data.
• To link different pyramid shapes to specific demographic properties.
• To provide a tool for analyzing and communicating historical population developments.

Main Methods:

• Developed a deterministic algorithmic approach for classifying population pyramid data.
• Implemented a data discretization step to simplify and unify data.
• Classified pyramids into non-species-specific shape categories.

Main Results:

• The algorithm achieved high classification accuracy on over 50,000 population pyramids from 450 mammal species.
• Classifications linked pyramid shapes to specific population size changes and transitions.
• The approach demonstrated plausible classifications for diverse mammal populations.

Conclusions:

• The algorithm offers a robust method for classifying population pyramid shapes and their demographic correlates.
• This approach can aid in analyzing historical population trends and informing animal population management strategies.
• The classification system provides a valuable tool for broad ecological and demographic research.