Plant size-variation in conjunction with number of neighbours may be an indicator of competition symmetry

• Identifying plant competition symmetry is crucial for ecological understanding. • Asymmetric competition is expected to increase plant size variability. • Using models we test whether size variation always indicates competition symmetry. • The coefficient of variation in plant size may indicate symmetry, with limits. • Low density, spatial randomness, or early development hinder accurate identification. Plant competition can be classified as symmetric when resource acquisition is proportional to plant size, or asymmetric when it is not. Identifying the symmetry of competition has important applications for population management. Many of the approaches that have been used to do this involve measuring growth over time, or through controlled experiments. An approach that uses measurements from a single point in time would be convenient. It is expected that populations under asymmetric competition will be more variable in plant size than those under symmetric competition, thus size variation at a single time could be an indicator of competition symmetry. However, other factors can also affect size variation, and thus it is critical to assess in which conditions size variation could be a good predictor of competition symmetry. We used a novel combination of simulation modelling and statistical analysis to explore this question. We simulated single species plant populations with varying density, spatial randomness, and initial plant size variation. Based on simulation outputs, we used Bayesian linear models to describe size variation as a function of the mean and standard deviation of the number of plant neighbours in a population, then used these models to study in what conditions size variation is or is not indicative of competition symmetry. Our results indicate that the single-time plant size coefficient of variation can be a useful predictor of the symmetry of competition in populations with higher densities and spatial randomness, up to 72 % accurate, but will be less useful in populations with low plant density or low spatial randomness. In such situations, asymmetric competition was often misclassified as symmetric, due to infrequent interactions and low variation in interactions among plants. This study provides insights into the complex relationships among tree density, spatial distribution, competition intensity, competition symmetry and plant size variation. Our approach has the potential to help understand and even infer competition symmetry in real tree populations.