Temporal fluctuation scaling in populations and communities

Area of Science:

• Ecology
• Ecological Statistics
• Population Dynamics

Background:

• Taylor's law posits that population abundance variance scales with the mean via a power law.
• This widely accepted ecological generalization has been used to infer population processes.
• Reexamination of empirical evidence and theoretical underpinnings is warranted.

Purpose of the Study:

• To critically evaluate Taylor's law and its supporting evidence.
• To investigate the influence of time series length and sampling errors on apparent power-law relationships.
• To propose an alternative null model for population fluctuation scaling.

Main Methods:

• Reanalysis of empirical data supporting Taylor's law.
• Development of a null model for variance-to-mean ratio based on demographic stochasticity and environmental stochasticity.
• Comparison of the null model's predictions with empirical population time series data.

Main Results:

• The exponent in Taylor's law is shown to be dependent on time series length.
• Sampling errors alone can generate apparent power-law relationships when data is plotted on a double logarithmic scale.
• A generic null model for variance-to-mean ratio was derived and compared to empirical data.

Conclusions:

• Taylor's law may not reliably reflect underlying ecological mechanisms due to confounding factors like sampling error and time series length.
• An alternative approach focusing on short-term fluctuations and a derived null model offers a potentially more robust method for analyzing population dynamics.
• This new approach can yield insights into species persistence and biodiversity theory using readily available monitoring data.