Nonlinear dynamics, where a change in the input is not proportional to a change in the output, are often found throughout nature, for example in biochemical kinetics. Because of the complex suite of interacting abiotic and biotic variables present in ecosystems, animal population dynamics are often thought to be driven in a nonlinear, state-dependent fashion. However, so far these have only been identified in model organisms and some natural systems. Here we show that nonlinear population dynamics are ubiquitous in nature. We use nonlinear forecasting to analyse 747 datasets of 228 species to find that insect population trends were highly nonlinear (74%), followed by mammals (58%), bony fish (49%) and birds (35%). This indicates that linear, equilibrium-based model assumptions may fail at predicting population dynamics across a wide range of animal taxa. We show that faster-reproducing animals are more likely to have nonlinear and high-dimensional dynamics, supporting past ecological theory. Lastly, only a third of time series were predictable beyond two years; therefore, the ability to predict animal population trends using these methods may be limited. Our results suggest that the complex dynamics necessary to cause regime shifts and other transitions may be inherent in a wide variety of animals.

中文翻译：

---

非线性种群动态在动物中无处不在。

非线性动力学，其中输入的变化与输出的变化不成比例，通常在整个自然界中发现，例如在生化动力学中。由于生态系统中存在一系列复杂的相互作用的非生物和生物变量，动物种群动态通常被认为是以非线性的、依赖状态的方式驱动的。然而，到目前为止，这些仅在模式生物和一些自然系统中被发现。在这里，我们表明非线性种群动态在自然界中无处不在。我们使用非线性预测分析了 228 个物种的 747 个数据集，发现昆虫种群趋势是高度非线性的 (74%)，其次是哺乳动物 (58%)、硬骨鱼 (49%) 和鸟类 (35%)。这表明线性，基于平衡的模型假设可能无法预测各种动物类群的种群动态。我们表明，繁殖速度更快的动物更有可能具有非线性和高维动力学，这支持了过去的生态学理论。最后，只有三分之一的时间序列可以预测超过两年；因此，使用这些方法预测动物种群趋势的能力可能会受到限制。我们的研究结果表明，导致政权转变和其他转变所必需的复杂动态可能是多种动物所固有的。使用这些方法预测动物种群趋势的能力可能有限。我们的研究结果表明，导致政权转变和其他转变所必需的复杂动态可能是多种动物所固有的。使用这些方法预测动物种群趋势的能力可能有限。我们的研究结果表明，导致政权转变和其他转变所必需的复杂动态可能是多种动物所固有的。