Soils are indispensable for the provision of several functions. Agricultural intensification and its focus on increasing primary productivity (PP) poses a threat to soil quality, due to increases in nutrient loads, greenhouse gas emissions and declining biodiversity. The EU Horizon 2020 Landmark project has developed multi-criteria decision models to assess five soil functions: PP, nutrient cycling (NC), soil biodiversity and habitat provision (B-HP), climate mitigation and water regulation, simultaneously in agricultural fields. Using these algorithms, we evaluated the supply of PP, NC and B-HP of 31 grasslands and 21 croplands as low, medium or high. The multi-criteria decision models showed that 38% of the farms had a medium to high supply of all three soil functions, whereas only one cropland had a high supply for all three. Forty-eight per cent of the farms were characterized by a high supply of PP and NC. We observed a clear trade-off between these two functions and B-HP. Multivariate statistical analyses indicated that higher organic inputs combined with a lower mineral fertilization concur with higher biodiversity scores while maintaining a medium delivery of PP and NC. Additionally, we compared the outputs of the model predictions to independent variables that served as proxies for the soil functions and found: (a) croplands (but not grasslands) with high PP had a higher standardized yield than those with medium PP; (b) grasslands (but not croplands) with high NC had a significantly lower fungal to bacterial biomass ratio, suggesting faster decomposition channels; and (c) a positive though non-significant trend between B-HP score and rank according to soil invertebrate biodiversity. These comparisons suggest a successful upscaling of the models from field to farm level. Our study highlights the need for systematic collection of management-related data for the assessment of soil functions. Multifunctionality can be achieved in agricultural soils; however, without specifically managing for it, biodiversity might come at a loss.

中文翻译：

---

评估农业土壤的多功能性：减少生物多样性权衡

土壤对于提供多种功能是必不可少的。由于养分负荷增加、温室气体排放增加和生物多样性下降，农业集约化及其对提高初级生产力 (PP) 的关注对土壤质量构成威胁。EU Horizo​​n 2020 Landmark 项目开发了多标准决策模型来评估五种土壤功能：PP、养分循环 (NC)、土壤生物多样性和栖息地供应 (B-HP)、气候缓解和水调节，同时在农业领域。使用这些算法，我们将 31 个草地和 21 个农田的 PP、NC 和 B-HP 的供应量评估为低、中或高。多标准决策模型显示，38% 的农场对所有三种土壤功能都具有中到高供应，而只有一个农田对这三种土壤功能都有高供应。48% 的农场的特点是 PP 和 NC 供应量高。我们观察到这两个函数和 B-HP 之间存在明显的权衡。多变量统计分析表明，较高的有机投入与较低的矿物施肥相结合，在保持 PP 和 NC 中等交付的同时，生物多样性得分较高。此外，我们将模型预测的输出与作为土壤功能代理的自变量进行了比较，发现：(a) 高 PP 的农田（但不是草地）比中等 PP 的农田（但不是草地）具有更高的标准化产量；(b) NC 高的草地（但不是农田）的真菌与细菌生物量比显着降低，表明分解通道更快；(c) B-HP 评分与土壤无脊椎动物生物多样性排名之间存在正但不显着的趋势。这些比较表明模型从田间到农场水平的成功升级。我们的研究强调了系统收集管理相关数据以评估土壤功能的必要性。可在农业土壤中实现多功能；然而，如果不对其进行专门管理，生物多样性可能会受到损失。