Multivariate signals of population collapse in a high-throughput ecological experiment

Area of Science:

• Ecology
• Population Dynamics
• Conservation Biology

Background:

• Predicting population declines is vital for conservation.
• Conceptual models propose sequential trait shifts precede collapse.
• Empirical validation is limited by data resolution.

Purpose of the Study:

• To empirically test conceptual models of population decline.
• To identify predictable sequences of trait changes under stress.
• To evaluate the utility of early warning signals for population forecasting.

Main Methods:

• Utilized an autonomously monitored, high-throughput experimental system.
• Generated individual-based data for *Paramecium caudatum* populations.
• Applied gradual pollutant introduction and predator disturbance as stressors.

Main Results:

• Gradual pollutant exposure induced a sequence: behavior (speed) -> morphology (length) -> abundance decline.
• Predator disturbance did not elicit this predictable sequence.
• Behavioral changes preceded abundance-based early warning signals by one generation.

Conclusions:

• Sequential trait shifts serve as reliable early warning signals for population collapse.
• Multivariate monitoring, especially individual-based metrics, is essential for forecasting.
• Disturbance type influences the predictability of population decline dynamics.