Improving competition representation in theoretical models of self-thinning: a critical review

Reynolds, J.M. and Ford, E. D.

Journal of Ecology, 93, 362-372, 2005

Summary

1. Many theoretical models have been proposed to explain the empirical self-thinning relationship given by Yoda et al. in 1963 for even-aged, monospecific stands of plants, but the models are inadequate to allow consensus on the processes driving variation in density-dependent mortality and self-thinning.
2. Most non-individual based models (non-IBMs), and many IBMs, employ a common representation of competition in which a finite amount of potential crown area remains completely allocated throughout self-thinning, making stand density inversely proportional to mean projected crown area.
3. This representation entails four assumptions regarding the competition process: the population is adequately represented by the mean plant; total stand resource utilization is constant throughout self-thinning; competition is a horizontal packing process; and differences in initial stand conditions may affect the rate of competition but not the process itself.
4. Reviewing published empirical data, the competition literature and the logical implications for the self-thinning process shows that all four assumptions are untenable as generalizations. Unfortunately, their application provides neither a mortality-inducing mechanism nor insight into the relationship between stand growth and mortality.
5. Explaining the observed variation in self-thinning relationships therefore requires improved representation of the competition process. This improvement is likely to require IBMs that explicitly represent variation in plant size or resource acquisition, two-dimensional stand distribution, dynamic rather than static stand resource utilization, and, perhaps, explicit three-dimensional stand development. Most importantly, the requirement for explicitly modelling mortality mechanisms implies that whole plant models may be insufficient for insight into the self-thinning process.
6. The review reinforces the need to assess mechanistic models for more than their ability to reproduce a single, high-level pattern. Such models should be assessed for their ability to simultaneously reproduce multiple features selected from the levels of both the modelled mechanisms and the high-level patterns.
7. Progress in understanding the observed variation in self-thinning currently requires a shift from searching for universal insight into the modelling of specific mechanisms for specific plant types, eventually leading to a broader theory explaining how variation in plants affects the competition process.