标题： 物种特征和群落属性解释了物种灭绝对碎屑食物网的影响 显示英文标题

标题： Species traits and community properties explain species extinction effects on detritus-based food webs

摘要： 土壤群落功能组成的改变对养分循环的影响仍不明确。模拟群落动态的模型克服了在野外进行物种去除实验的技术挑战。然而，到目前为止，现有的土壤食物网模型未能充分代表有机质处理链，而这是土壤动态的关键。在这里，我们提出了一种新的土壤食物网动态模型，考虑了代谢率的异速生长规律、有机质的发育过程以及氮和碳流动的显式表示。我们使用该模型来研究哪些特征是物种对群落生产力和养分循环影响的最佳预测因子。为此，我们从48个森林土壤食物网中逐一移除161个营养物种（功能相同的物种组），并模拟其动态直至达到平衡。模拟结果表明，特征组合比单一特征更能决定移除效应。最小的物种是最具竞争力的物种，但具有各种体质量的肉食者，表现出最高的连通性和资源相似性，可能是在调节竞争力量中的关键物种。尽管如此，大多数移除操作的影响都很小，这表明功能冗余为土壤食物网提供了对单个营养物种灭绝的高抵抗力。我们还首次指出，食物网结构和土壤肥力可以以不可预测的方式显著改变物种的影响。此外，过去的研究中排除了碎屑和元素计量限制，导致对间接效应和反馈作用的低估。虽然需要进一步的工作来纳入碎屑动物之间的互补性，但在模型中考虑这些机制对于改善对土壤食物网功能的理解至关重要。 显示英文摘要

摘要： Effects of changes in functional composition of soil communities on nutrient cycling are still not well understood. Models simulating community dynamics overcome the technical challenges of conducting species removal experiments in the field. However, to date, available soil food web models do not adequately represent the organic matter processing chain which is key for soil dynamics. Here, we present a new model of soil food web dynamics accounting for allometric scaling of metabolic rate, ontogeny of organic matter, and explicit representation of nitrogen and carbon flows. We use this model to investigate what traits are best predictors of species effects on community productivity and on nutrient cycling. To do so, we removed 161 tropho-species (groups of functionally identical species) one at a time from 48 forest soil food webs, and simulated their dynamics until equilibrium. Simulations revealed that combinations of traits better determine removal effects than single ones. The smallest species are the most competitive ones, but carnivores of various body masses presenting the highest connectivity and resource similarity could be key stone species in the regulation of competitive forces. Despite this, most removals had low effects, suggesting functional redundancy provides a high resistance of soil food webs to single tropho-species extinction. We also highlight for the first time that food web structure and soil fertility can drastically change species effects in an unpredictable way. Moreover, the exclusion of detritus and stoichiometric constraints in past studies lead to underestimations of indirect effects and retroactions. While additional work is needed to incorporate complementarity between detritivores, it is essential to take into account these mechanisms in models in order to improve the understanding of soil food web functioning.