Transferable Ecology for a changing world (TREE)

迫切需要全球范围的可持续性转型，以应对灾难性的气候变化和生物多样性丧失。然而，在实践中，应对这些危机的保护、恢复和管理行动将适用于地方规模：个别森林、田野、湖泊和草地。从业者、政策制定者和生态学家早就认识到银、“一刀切”战略的徒劳性，以及针对特定或个别情况采取针对性和量身定制行动的重要性，并为特定物种或个别地点制定行动计划。例如，英国政府雄心勃勃的植树计划强调了“在正确的地方种植正确的树木”对实现净零排放的重要性，因为在错误的地方种植错误的树木可能无法建立，或者更糟的是，导致净碳排放持续数十年。然而，尽管需要为个别网站的具体情况的信息，生态学传统上侧重于平均效应，而不是个人的影响，破坏了应用生态学的最终目标：生态科学的应用到现实世界。为了进行能够真正为实地政策提供信息的科学研究，生态学可以利用快速发展的“精准医学”领域开发的方法，该领域同样基于这样的前提：给定药物或治疗的效果可能会有所不同对于每个患者来说，这需要个性化的医疗保健计划。由于来自多个来源的大量生态数据越来越多，包括遥感器、公民科学家和环境监测网络，需要“大数据”的精确医学方法的应用现在已经成为可能。在这个探索性的项目中，我们将采用精准医学方法来解决应用生态问题，并开发新的方法，帮助指导下一代应用生态学家进入大数据时代。我们将开发框架和工具，用于对大数据进行采样和建模，以产生准确和实用的知识，从而为实地行动提供信息。该项目将使用“虚拟生态”方法实现这一目标，该方法使用现有的计算机模型“iLand”，模拟欧洲、美国和日本的真实的森林景观。首先，我们将使用iLand在这些景观中的单个森林中虚拟实施森林恢复行动。然后将对这些地貌中的森林进行虚拟取样，从而模仿生态学家、民间科学家和遥感技术收集数据的方式。例如，我们将模仿测量森林生物多样性的测量员，作为国家环境监测活动的一部分。然后，我们将应用不同的建模方法来分析虚拟采样数据，包括在精准医学领域开发的方法以及生态学中常用的方法。通过将这些分析的结果与“真实”的模拟数据进行比较，我们将能够评估不同的采样和建模决策如何影响我们结果的准确性。具体来说，我们将评估预测的准确性，在一系列环境条件下（即，海拔、气候、土壤类型）。我们将利用这种理解来制定指导方针，用于设计未来的环境监测调查和大数据建模，以最大限度地提高这些预测的准确性。通过将应用生态学的重点从平均效应转向可操作的个人水平效应，我们的指导方针将使政策制定者能够将恢复行动的目标和优先顺序放在预期效果最大的地区，从而最大限度地提高有限资源的有效性，以应对气候变化和生物多样性丧失。

Global-scale sustainability transformations are urgently required to address catastrophic climate change and biodiversity loss. In practice, however, the conservation, restoration and management actions to combat these crises will be applied at local scales: to individual forests, fields, lakes, and grasslands. Practitioners, policymakers and ecologists have long-recognised the futility of silver bullet, 'one-size-fits-all' strategies, and the importance of targeting and tailoring actions to specific or individual contexts, and developing action plans for specific species, or individual sites. For example, the UK government's ambitious tree planting initiatives emphasise the importance of 'the right tree in the right place' to achieving Net Zero, because the wrong trees planted in the wrong place may fail to establish, or worse still, lead to net carbon emissions that persist for decades. However, despite the need for context-specific information for individual sites, ecology has traditionally focused on average effects, rather than individual effects, undermining the ultimate goal of applied ecology: the application of ecological science to the real-world. To do science that can truly inform on-the-ground policies, ecology can capitalise on methods developed by the rapidly advancing field of 'precision medicine', which similarly rests on the premise that the effect of a given drug or treatment may be different for each individual patient, necessitating personalised healthcare plans. The application of precision medicine methodologies, which require 'big data', is now possible due to the increasing availability of vast quantities of ecological data from multiple sources, including remote sensors, citizen scientists and environmental monitoring networks. In this exploratory project, we will adapt the use of precision medicine approaches to applied ecological problems and develop novel approaches that help guide the next generation of applied ecologists into the big-data era. We will develop frameworks and tools for sampling and modelling big data in a way that yields accurate and practicable knowledge that can inform on-the-ground actions.This project will achieve this using a 'virtual ecology' approach using an existing computer model - 'iLand' - that simulates real, forested landscapes in Europe, the U.S.A and Japan. First, we will use iLand to virtually implement forest restoration actions to individual forests across these landscapes. The forests in these landscapes will then be subjected to virtual sampling, thus mimicking the collection of data by ecologists, citizen scientists and remote sensing technologies. For example, we will mimic surveyors that measure forest biodiversity as part of a national environmental monitoring campaigns. We will then apply different modelling methods to analyse the virtually sampled data, including methods developed in the field of precision medicine as well as methods that are commonly used in ecology. By comparing the results of these analyses against the 'true' simulated data, we will be able to evaluate how different sampling and modelling decisions influence the accuracy of our results. Specifically, we will assess the accuracy of predictions individual forest responses to restoration actions across a range of environmental conditions (i.e., elevation, climate, soil types). We will use this understanding to develop guidelines for designing future environmental monitoring surveys, and modelling of big data, that maximises the accuracy of these predictions. By shifting applied ecology's focus away from average effects, towards actionable, individual-level effects, our guidelines will enable policymakers to target and prioritise restoration actions to areas where desired effects are greatest, thus maximising efficacy of limited resources for addressing climate change and biodiversity loss.