Purpose Characterization factors for biodiversity impact assessment derived from ecological zoning and land use classification have been proposed within life cycle assessment (LCA). However, their applicability to LCA of agricultural production systems has not yet been elucidated. This study clarifies correlations between regional-scale biodiversity indicators (characterization factors) and field-scale biodiversity indicators and estimates the degree of macro-micro inconsistencies in biodiversity indicators. Methods Correlation coefficients were calculated between two types of variables. One is biodiversity (potential species loss) at the ecoregion level provided in UNEP/SETAC ( 2017 ) and Chaudhary and Brooks ( 2018 ), and the other is biodiversity (species richness) at the field level surveyed during a research project on biodiversity in Japan. The data on two taxa (amphibians and plants) in paddy fields were used for the analyses. Two types of correlation coefficients (the Pearson’s product-moment correlation coefficient and the Spearman’s rank correlation coefficient) were calculated. Uncertainties of the correlation coefficients were estimated by statistical resampling because the number of surveyed years and regions were limited. Results and discussion Although in most cases the signs of the coefficients were consistent with theoretical expectations that the correlation between potential species loss and species richness must be negative, the absolute values were low for all cases (especially for the case of amphibians and for the case of using the UNEP/SETAC characterization factors). It was difficult to estimate field-scale biodiversity from ecoregion-scale biodiversity. The introduction of land use intensity into the calculation of biodiversity at the ecoregion-scale increased the correlation coefficients for plants. Uncertainties due to limitations in the number of surveyed locations were larger than those arising from limitations in the number of surveyed years. These results highlighted the existence of macro-micro inconsistencies and the necessity of developing constructive approaches for biodiversity assessment in agriculture. Conclusions It is concluded that employing characterization factors based on ecoregions and land use categories was not useful when assessing the biodiversity impacts of rice production systems at the field-scale because of the existence of macro-micro inconsistencies. Use of field monitoring methods, in addition to approaches to construct biodiversity indicators based on management practices, will be necessary for establishing sustainable agricultural production systems.

中文翻译：

---

生命周期评估中区域和田间生物多样性指标的不一致：以日本水稻生产系统为例

目的 生命周期评估 (LCA) 中提出了源自生态分区和土地利用分类的生物多样性影响评估的特征因子。然而，它们对农业生产系统 LCA 的适用性尚未阐明。本研究阐明了区域尺度生物多样性指标（特征因子）与田间尺度生物多样性指标之间的相关性，并估计了生物多样性指标宏观与微观不一致的程度。方法计算两类变量之间的相关系数。一个是 UNEP/SETAC（2017 年）和 Chaudhary 和 Brooks（2018 年）提供的生态区层面的生物多样性（潜在物种损失），另一个是日本生物多样性研究项目期间调查的实地层面的生物多样性（物种丰富度） . 稻田中两种分类群（两栖动物和植物）的数据用于分析。计算了两种类型的相关系数（Pearson 积矩相关系数和 Spearman 等级相关系数）。由于调查年份和地区数量有限，相关系数的不确定性通过统计重采样估计。结果和讨论 尽管在大多数情况下，系数的符号与理论预期一致，即潜在物种损失和物种丰富度之间的相关性必须为负，但所有情况下的绝对值都很低（特别是两栖动物的情况和使用 UNEP/SETAC 特征因子）。很难从生态区规模的生物多样性中估计田间规模的生物多样性。在生态区尺度的生物多样性计算中引入土地利用强度增加了植物的相关系数。调查地点数量限制引起的不确定性大于调查年数限制引起的不确定性。这些结果突出了宏观与微观不一致的存在以及为农业生物多样性评估开发建设性方法的必要性。结论 结论是，在评估大田规模水稻生产系统的生物多样性影响时，使用基于生态区和土​​地利用类别的特征因子是没有用的，因为存在宏观与微观的不一致。使用现场监测方法，