Disturbance-driven changes in ecosystem variability: a new tool for understanding ecological dynamics

统计学告诉我们，所有变量最初都可以用两个参数来描述--它们的中心趋势和变异性。两者都提供了关于响应变量分布的有意义的信息，从而为推理、预测和理论提供了基础。然而，在生态学研究中，变异很少被视为兴趣的反应。因此，人们对生态系统动态变化的功能重要性或变化的机制基础知之甚少。干扰有可能影响生态系统的方方面面，从物种多样性到结构和功能，都会产生显著的时空异质性，从而深刻改变系统动力学。根据干扰的频率、严重程度和程度来定义，干扰可以直接或间接地导致变异性的变化。研究干扰如何改变变化，可以揭示原本可能被忽视的生态变化，还可能揭示产生这种动态的潜在机制。在以前工作的基础上，我使用淡水湖作为理想的建模和实验对象。淡水湖具有复杂的多维栖息地结构、清晰的生态系统边界和多个营养级别，世代时间很快。目前的提议将检验多个时间和空间尺度上的扰动--方差预测：1)每年发生大的“全系统”扰动的湖泊之间；2)具有不同混合频率(即扰动制度)和应力来源制度的湖泊(和年内)；以及3)实验上，利用中观系统检验不同扰动制度产生的预测。不断增加的差异已被确定为状态变化的一个指标，我还将评估使用差异作为预警指标和触发管理和缓解活动的工具的效用。在这样一个世界里，多个应激源现在是常态而不是例外，变异性的增加进一步降低了我们检测其影响的机会，发展对生态系统变异性对干扰的反应的预测性理解是当务之急。湖泊对人类福祉至关重要，全球淡水生态系统的许多威胁(如气候变暖、营养物质)很容易在长期和高频数据集中被观察和记录下来，这些数据是由长期监测计划捕获的。这些项目将探索干扰和生态系统变异性的最重要属性，并最终提高我们管理淡水资源的能力。这一倡议将利用这两个世界的最好之处，合作培养2名博士和1名理学硕士。以及两名本科生，他们来自一所顶尖的大学湖泊实验室以及一家卓越的省政府淡水研究中心。

Statistics teaches us that all variables can be initially characterized by two parameters - their central tendency and their variability. Both impart meaningful information about the distribution of a response variable and thus provide a foundation for inference, prediction, and theory. However, variance is seldom considered in ecological studies as a response of interest. As a result, little is known about the functional importance of, or the mechanistic basis for, changes in the variance of ecosystem dynamics. Disturbances have the potential to affect all aspects of ecosystems, from species diversity to structure and function, by generating significant spatial and temporal heterogeneity that can profoundly alter system dynamics. Defined by their frequency, severity, and extent, disturbances can directly or indirectly contribute to changes in variability. Examining how a disturbance alters variance can reveal ecological change that might otherwise be overlooked and may also suggest the underlying mechanisms that produce such dynamics. Building on previous work, I use freshwater lakes, with complex multi-dimensional habitat structure, well-delineated ecosystem boundaries and multiple trophic levels with rapid generation times as ideal candidates for modelling and experimentation. The current proposal will test disturbance-variance predictions across multiple temporal and spatial scales: 1) among lakes with large “system-wide” disturbances occurring annually over decades; 2) within lakes (and years) with different mixing frequencies (i.e., disturbance regimes) and stressor regimes; and 3) experimentally, using mesocosms to test predictions arising from different disturbance regimes. Increasing variance has been identified as an indicator of state change and I will also assess the utility of using variance as an early warning metric and a tool to trigger management and mitigation activities. In a world where multiple stressors are now the norm rather than the exception and increased variability further reduces our chances of detecting their effects, developing a predictive understanding of the response of ecosystem variability to disturbance is an imperative. Lakes are of critical importance to human well-being and many of the threats to freshwater ecosystems globally (e.g., climate warming, nutrients), are readily observed and documented in the long-term and high-frequency datasets captured by long-term monitoring program. These projects will explore the most important attributes of both disturbance and ecosystem variability and will ultimately improve our ability to manage our freshwater resources.This initiative will draw on the best of both worlds to collaboratively train 2 Ph.D., 1 M.Sc. and 2 undergraduates with involvement from a top university-based limnological laboratory as well as a pre-eminent provincial government freshwater research centre.