Indicators of regime shifts in freshwater ecosystems

以生物群落组成突然、往往不可逆转的变化为特征的制度转变，可能对社会所依赖的生态系统服务产生灾难性影响。这种现象的经典例子来自淡水生态系统，在淡水生态系统中，群落结构的变化可能导致以蓝藻为主的生态系统，对人和牲畜的健康有潜在的负面影响。因此，能够及时预测即将发生的政权转变以避免这些转变是一项关键目标，对稀缺淡水资源的管理具有重大影响。该项目将使用长期监测数据，结合先进的建模技术，描述在政权更迭之前和期间社区组成可能发生的变化，并测试和开发可推广的方法，以预测和防止未来的此类变化。预测生态学的一个关键目标是预测生态系统发生快速变化的可能性，从而产生“早期预警信号”。这些都是泛化的方法，旨在通过检测时间序列数据中的信号来预测社区组成的变化，这些信号是即将发生的制度转变的症状。这类信号的潜在效力已在模拟研究中得到广泛展示，但在很大程度上仍未在真实世界的数据中得到检验，部分原因是在观察到的制度变化前后缺乏长期监测。这引发了人们对它们是否适合为自然资本的管理战略提供信息的问题。该项目将通过双管齐下的方法解决这一知识鸿沟：对经过制度转变的经过充分研究的湖泊生态系统的长期监测数据进行分析，以及对湖泊群落进行补充理论建模。加利利海是以色列唯一的天然淡水湖，因此是重要的资源。它的重要性意味着近50年来一直在进行广泛的监测，提供了1969年至2018年期间鱼类、浮游动物和浮游植物物种的丰度、生物量和密度的特殊数据，以及湖泊化学成分的变化。该湖的群落在1994-1995年间经历了重大转变，导致水质严重恶化，有害藻类大量繁殖。因此，在已知的制度转变之前、期间和之后，可以获得关于多种物种和营养水平的数据，这使其成为测试和开发预警信号方法的理想工具。这些特殊的数据将被用来对湖泊群落的大小-光谱模型进行参数化，从而能够评估系统扰动的多个模拟结果。这个项目将：(1)评估加利利海的制度变迁是否可以在其发生之前被预测，(2)确定这种警告信号可以提前多长时间被检测到，(3)检查当试图预测系统未来的动态时，重点关注特定物种、营养水平或整体群落动态是更好的做法，以及(4)确定未来应该收集什么数据来预测制度变迁。除了由主管提供的专业培训外，学生还将参加一系列全面的一般课程，以提高他们的就业能力和个人发展，包括统计、计算、研究伦理、知识产权和企业、生物信息学、抽样方法和研究技能的培训。

Regime shifts, characterised by sudden, often irreversible, changes in the composition of biological communities, can have catastrophic impacts on the ecosystem services which society relies on. Classic examples of this phenomenon originate from freshwater ecosystems, where shifts in the structure of a community can lead to cyanobacteria-dominated ecosystems, with potential negative impacts on human and livestock health. Being able to predict impending regime shifts in time to avert them is consequently a critical goal with significant implications for the management of scarce freshwater resources. This project will use long-term monitoring data in combination with advanced modelling techniques to characterise how community composition can change prior to and during a regime shift, and test and develop generalisable methods to predict and prevent such shifts in the future. A key goal in predictive ecology is forecasting the potential for rapid changes in ecosystems, leading to the development of "early warning signals". These are generalisable methods which aim to predict changes in the composition of a community by detecting signals in time series data which are symptomatic of an approaching regime shift. The potential efficacy of such signals has been widely shown in simulation studies, but remains largely untested on real-world data, in part due to the lack of long-term monitoring before and after observed regime shifts. This has raised questions about their suitability to inform management strategies for natural capital. This project will tackle this knowledge gap using a two-pronged approach: analysis of long-term monitoring data from a well-studied lake ecosystem which has undergone a regime shift, and complementary theoretical modelling of the lake community. The Sea of Galilee is the only natural freshwater lake in Israel, and consequently a key resource. Its importance has meant extensive monitoring has been carried out for nearly 50 years, providing exceptional data on the abundances, biomasses, and densities of fish, zooplankton, and phytoplankton species from 1969 until 2018, as well as changes in the lake's chemical composition. The lake's community underwent a major shift in 1994-1995, resulting in a severe deterioration in water quality and an increase in harmful algal blooms. Consequently, data on multiple species and trophic levels are available prior to, during, and after a known regime shift, making it ideal for testing and developing warning signal methods. These exceptional data will be used to parameterise a size-spectra model of the lake community, allowing multiple simulated outcomes of perturbations on the system to be assessed. This project will: (1) assess whether the regime shift in the Sea of Galilee could have been predicted prior to its occurrence, (2) determine how far in advance such warning signals are detectable, (3) examine whether it is better to focus on specific species, trophic levels, or look at the community dynamics as a whole when trying to predict a system's future dynamics, and (4) identify what data should be collected in the future to predict regime shifts. In addition to the specialist training provided by the supervisors, the student will undertake a full range of general courses to enhance their employability and personal development, including training on Statistics, Computing, Research Ethics, Intellectual Property and Enterprise, Bioinformatics, Sampling Methodology, and Research Skills.