Weak signals, strong debates: Density dependence and population regulation through the lens of model uncertainty

TL;DR

This paper discusses the challenges in determining population regulation due to model uncertainty, and introduces a new measure for assessing density dependence, revealing significant variability across populations.

Contribution

It highlights the intractability of population regulation due to model uncertainty and proposes a novel, intuitive measure for density dependence estimation.

Findings

Model conclusions depend heavily on chosen structure.

Density dependence varies widely among populations.

Average perturbation half-life is approximately 3 years.

Abstract

Ecologists have long argued about the strength of density dependence and population regulation, respectively defined as the short-term and long-term rates of return to equilibrium. Here, I give three arguments for the intractability of population regulation. First, the ecological literature flip-flops on the strength of evidence for population regulation; by simple induction, population regulation should remain uncertain. Second, there is an analogous debate in economics about whether shocks to Gross Domestic Product have transient or permanent effects. This literature is extensive and sophisticated, yet there is no consensus, implying that more research will not resolve the issue in ecology. Third, using a variety of time series models and the Global Population Dynamics Database, I show that one's conclusions about population regulation are almost entirely dependent on model structure. This insurmountable model uncertainty explains why the strength of regulation is unresolved despite decades of research. However, it is possible to achieve the more modest goal of estimating density dependence. I introduce a novel measure of density dependence -- the effective autoregressive parameter -- which is conceptually intuitive and easy to calculate with simulations. The strength of density dependence varies significantly across populations, with an average that can be characterized as moderate: perturbations have an average half-life of 3 years. Rather than a universal balance of nature, stability varies widely across populations in ways that correlate with life history and taxonomy.