Quantitative predictions about the processes that promote species coexistence are a subject of active research in ecology. In particular, competitive interactions are known to shape and maintain ecological communities, and situations where some species out-compete or dominate over some others are key to describe natural ecosystems. Here we develop ecological theory using a stochastic, synthetic framework for plant community assembly that leads to predictions amenable to empirical testing. We propose two stochastic, continuous-time Markov models that incorporate competitive dominance through a hierarchy of species heights. The first model, which is spatially implicit, predicts both the expected number of species that survive and the conditions under which heights are clustered in realized model communities. The second one allows spatially-explicit interactions of individuals and alternative mechanisms that can help shorter plants overcome height-driven competition, and it demonstrates that clustering patterns remain, not only locally but also across increasing spatial scales. Moreover, although plants are actually height-clustered in the spatially-explicit model, plant species abundances are not necessarily skewed to taller plants.

关于促进物种共存的过程的定量预测是生态学研究的主题。尤其是，竞争相互作用可以塑造和维持生态群落，而某些物种在竞争中胜过其他物种或在其他物种上占优势的情况是描述自然生态系统的关键。在这里，我们使用随机的，综合的植物群落组装框架来发展生态理论，从而得出适合于经验检验的预测。我们提出了两个随机的，连续时间的马尔可夫模型，这些模型通过物种高度的层次结构综合了竞争优势。第一个模型在空间上是隐式的，既预测了预期生存的物种数量，又预测了在已实现的模型群落中高度聚集的条件。第二种方法允许个体在空间上进行显式的交互，并且可以使用其他机制帮助较短的植物克服高度驱动的竞争，并且证明聚类模式不仅存在于局部，而且还在不断扩大的空间范围内。而且，尽管在空间显式模型中实际上对植物进行了高度聚类，但植物物种的丰度并不一定偏向较高的植物。