[Paper Review] The storage effect is not about bet-hedging or population stage-structure

Evan C. Johnson, Godoy, Oscar+1 more|arXiv (Cornell University)|Jan 18, 2022

Ecology and Vegetation Dynamics Studies4 citations

TL;DR

This paper challenges the widespread belief that the storage effect in ecological coexistence requires a persistent life stage (e.g., seed bank) to 'wait out' unfavorable conditions. Using mathematical modeling and empirical data from Mediterranean annual plants, it demonstrates that the storage effect arises from interaction effects between environmental conditions and competition—not from stage structure or bet-hedging—making it possible even in species without long-lived stages.

ABSTRACT

The storage effect is a well-known explanation for coexistence in temporally varying environments. Like many complex ecological theories, the storage effect is often used as an explanation for observed coexistence on the basis of heuristic understanding, rather than careful application of a detailed model. One interpretation states that species coexist by specializing on specific environmental states, and therefore must have a robust life-stage (e.g., long-lived adults, a seed-bank) in order to "wait it out" for favorable conditions. Here we show that this widely employed interpretation can be misleading. Multiple models show that stage-structure, long lifespans, and overlapping generations are neither necessary nor sufficient for the storage effect. In models where a robust life-stage does engender a storage effect, it does not do so by preventing stochastic extinction or by improving relative bet-hedging. A robust life-stage is best understood as one of many ways to fulfill an abstract condition for the storage effect: an interaction effect of environment and competition on per capita growth rates. Using a dataset of annual plants from a Mediterranean grassland in Spain, we show that such interaction effects occur between water availability and the number of germinant competitors, leading to storage in the absence of a persistent seed bank. Empiricists hoping to uncover the storage effect should look for interaction effects between environmental conditions and competition -- easily identifiable with multiple regression -- at all stages of a species' life-cycle.

Motivation & Objective

• To correct the widespread misconception that the storage effect depends on robust life stages like seed banks or long-lived adults.
• To demonstrate that the storage effect can occur without stage-structure, overlapping generations, or persistent life stages.
• To clarify that the core mechanism is an interaction effect between environment and competition, not buffering or bet-hedging.
• To provide empirical evidence from Mediterranean annual plants showing the storage effect exists without a seed bank.
• To advocate for a shift in empirical research toward detecting environment-competition interaction effects across life stages.

Proposed method

• Developed and analyzed mathematical models (e.g., lottery model, annual plant model) to isolate the conditions under which the storage effect arises.
• Used Modern Coexistence Theory to decompose coexistence mechanisms into three core ingredients: species-specific environmental responses, environment-competition interaction effects, and covariance between environment and competition.
• Applied multiple regression to empirical data from a Mediterranean grassland to detect interaction effects between water availability and competitor density.
• Fitted hierarchical Bayesian models using Stan and Hamiltonian Monte Carlo to estimate species-specific responses and competition coefficients.
• Compared model-based and data-based estimates of species response correlations and competition coefficient distributions.
• Used within-chain parallelization and high-performance computing to handle complex models with many competition coefficients.

Experimental results

Research questions

• RQ1 Is the storage effect truly dependent on a persistent life stage such as a seed bank?
• RQ2 Can the storage effect occur in species without stage-structure or overlapping generations?
• RQ3 Do interaction effects between environmental conditions and competition drive the storage effect, independent of life-stage buffering?
• RQ4 Is the conventional interpretation of the storage effect as 'waiting it out' via robust life stages empirically supported?
• RQ5 Can the storage effect be detected without assuming a seed bank, using measurable interaction effects in the field?

Key findings

• The storage effect does not require a persistent life stage; models without stage-structure still exhibit the storage effect.
• A robust life stage does not prevent stochastic extinction or enhance relative bet-hedging, contradicting common interpretations.
• Empirical data from Spanish annual plants show significant interaction effects between water availability and competitor density, supporting the storage effect without a seed bank.
• The distribution of competition coefficients (intra- and interspecific) was nearly identical, indicating coexistence is not driven by fixed resource partitioning.
• Species-specific responses to the environment were weakly correlated, supporting the presence of niche differentiation via environmental responses.
• The covariance between environment and competition (ingredient #3) was detectable and significant, confirming a key condition for the storage effect.