GLobal Insect Threat-Response Synthesis (GLiTRS): a comprehensive and predictive assessment of the pattern and consequences of insect declines

随着人们越来越认识到昆虫的重要性，人们越来越担心昆虫的生物多样性在全球范围内下降，对生态系统的功能和服务造成严重后果。然而，知识上的差距限制了在理解变化的规模和方向方面取得进展。关于昆虫趋势的信息是碎片化的，时间序列数据在分类学和空间上都是有限的和不具代表性的。此外，昆虫趋势与人为压力之间的因果关系尚未得到证实。因此，很难评估关于“昆虫末日”的故事，了解生态系统的后果，制定缓解战略，或预测未来趋势。为了解决这些不足，我们将汇集各种信息来源，如元分析、相关关系和专家判断。GLiTRS将整理这些不同的证据线，研究昆虫生物多样性如何响应人为压力而发生变化，响应如何根据功能特征、空间和生物多样性指标（如物种丰度、占用率、丰富度和生物量）而变化，以及昆虫趋势如何推动进一步的变化（如通过相互作用网络介导）。我们将把这些证据整合到一个描述全球昆虫生物多样性趋势的威胁-反应模型中。该模型将以一系列概率陈述（贝叶斯信念网络）的形式表示，描述昆虫生物多样性与人为压力之间的关系。通过挑战这种“威胁-反应模型”来预测分类群和存在高质量时间序列数据的地方的趋势，我们将确定间接数据源a)足以预测近期趋势，b)不充分，或c)太不确定的昆虫类群和地区。关于威胁-反应关系的可预测性的知识将允许预测——不确定性估计——昆虫生物多样性如何在全球范围内，在所有主要分类群、功能群和生物群系中发生变化。这种对近期趋势的全球视角将为探索昆虫减少对一系列生态系统功能和服务的影响，以及未来环境变化的合理情景如何影响生物多样性和生态系统特性提供基础。GLiTRS是一个雄心勃勃的创新研究项目：有两个特点特别具有开创性。首先，对多种形式的证据进行整理，将使人们能够从真正的全球角度看待昆虫的减少，这是使用传统方法无法实现的。其次，通过验证“先验知识”（来自证据合成）和最近的趋势，我们将评估昆虫减少的可预测程度和规模。

With increasing recognition of the importance of insects, there are growing concerns that insect biodiversity has declined globally, with serious consequences for ecosystem function and services. Yet, gaps in knowledge limit progress in understanding the magnitude and direction of change. Information about insect trends is fragmented, and time-series data are restricted and unrepresentative, both taxonomically and spatially. Moreover, causal links between insect trends and anthropogenic pressures are not well-established. It is, therefore, difficult to evaluate stories about "insectageddon", to understand the ecosystem consequences, to devise mitigation strategies, or predict future trends. To address the shortfalls, we will bring together diverse sources of information, such as meta-analyses, correlative relationships and expert judgement. GLiTRS will collate these diverse lines of evidence on how insect biodiversity has changed in response to anthropogenic pressures, how responses vary according to functional traits, over space, and across biodiversity metrics (e.g. species abundance, occupancy, richness and biomass), and how insect trends drive further changes (e.g. mediated by interaction networks).We will integrate these lines of evidence into a Threat-Response model describing trends in insect biodiversity across the globe. The model will be represented in the form of a series of probabilistic statements (a Bayesian belief network) describing relationships between insect biodiversity and anthropogenic pressures. By challenging this "Threat-Response model" to predict trends for taxa and places where high-quality time series data exist, we will identify insect groups and regions for which indirect data sources are a) sufficient for predicting recent trends, b) inadequate, or c) too uncertain. Knowledge about the predictability of threat-response relationships will allow projections - with uncertainty estimates - of how insect biodiversity has changed globally, across all major taxa, functional groups and biomes.This global perspective on recent trends will provide the basis for an exploration of the consequences of insect decline for a range of ecosystem functions and services, as well as how biodiversity and ecosystem properties might be affected by plausible scenarios of future environmental change.GLiTRS is an ambitious and innovative research program: two features are particularly ground-breaking. First, the collation of multiple forms of evidence will permit a truly global perspective on insect declines that is unachievable using conventional approaches. Second, by validating "prior knowledge" (from evidence synthesis) with recent trends, we will assess the degree to which insect declines are predictable, and at what scales.