Eisenberg J N, Maszle D R
Division of Environmental Health Science, University of California at Berkeley 94720, USA.
J Theor Biol. 1995 Oct 21;176(4):501-10. doi: 10.1006/jtbi.1995.0216.
A three-species food-chain model which was previously shown to exhibit chaotic dynamics was revisited. By exploring the sensitivity of that result this study found that complex behavior depended on the functional form chosen to model the interaction between the two highest species in the food chain. Two separate scenarios were explored: the gradual addition of refugia modeling the escape from predation at low prey densities; and the gradual addition of predator interference modeling territorial behavior. The addition of even a small amount of refugia provided a stabilizing influence as the chaotic dynamics collapsed to stable limit cycles. The results of adding interference to the model were more complex. Although the numerical simulations indicated that a low level of interference provided a stabilizing influence, the analytical results suggest that complex dynamics are possible for a range of parameter values that are biologically relevant. The sensitivity of the stability profile to functional changes in the model suggests two important ecological motivations for structural stability analysis. First, in ecological systems, environmental fluctuations cause continuous changes in the functional relationships between and within species, resulting in potential changes in the complexity of the dynamics over time. Second, slight changes in ecological structure may cause significant bifurcations; however, most ecological data are inadequate to distinguish such phenomena.
一个先前被证明会展现出混沌动力学的三物种食物链模型被重新审视。通过探究该结果的敏感性,本研究发现复杂行为取决于为模拟食物链中两个最高物种之间的相互作用而选择的函数形式。研究了两种不同的情形:逐渐增加避难所,模拟低猎物密度下被捕食风险的降低;逐渐增加捕食者干扰,模拟领地行为。即使添加少量避难所也会产生稳定作用,因为混沌动力学坍缩为稳定的极限环。给模型添加干扰的结果更为复杂。尽管数值模拟表明低水平的干扰会产生稳定作用,但分析结果表明,对于一系列具有生物学相关性的参数值,可能会出现复杂动力学。稳定性分布对模型中函数变化的敏感性表明了结构稳定性分析的两个重要生态学动机。第一,在生态系统中,环境波动会导致物种之间以及物种内部的函数关系持续变化,从而随着时间的推移导致动力学复杂性的潜在变化。第二,生态结构的微小变化可能会导致显著的分岔;然而,大多数生态数据不足以区分此类现象。