Purpose Land use change (LUC) is a critical process in the life cycle greenhouse gas emissions of agricultural products and Brazil is a major exporter of these. This work had the objective of integrating refined and regionalized datasets of LUC in Brazil into the ecoinvent database, to better represent its dynamics and heterogeneity. We present the adaptations needs for having it suitable for crops, pasture and forestry in state-level and impacts of modelling assumptions and uncertainties. Methods Adaptation and integration were based in ecoinvent version 3 guidelines and the database requirements to LUC modelling. BRLUC, a method for Brazilian LUC accounting, was the main data source. The workflow for the integration process consisted in identifying necessary adaptations in both sources to allow a better representation of Brazilian LUC. Four new reference products and 27 geographies were added in the database. Results and discussion A total of 566 new datasets were integrated into ecoinvent version 3.6, allowing the incorporation of LUC in Brazilian products in state, regional and national level. GHG emissions reduced, being 42.2% and 99.9% lower to soybean and sugarcane than in ecoinvent v3.5. Four improvements were the main causes: (i) state-level LUC modelling with national official data; (ii) regionalizing carbon stocks; (iii) including pasture and forestry land use categories; (iv) and considering sugarcane as a perennial crop. The way to calculate national-level results based on subnational data was an important source of difference in emissions too. Uncertainties specifically associated with land use substitution patterns were not incorporated, and they can potentially have impacts as large as the uncertainties of all the remaining processes combined. Conclusions Results showed that small changes in data sources and premises have large impacts on emissions associated with LUC in agricultural products. It also showed the large impacts of uncertainties of LUC patterns. Improving current models in better representing regional LUC patterns, regional carbon stocks and uncertainty accounting could reduce these impacts. Nonetheless, efforts in reducing the complexity of LUC accounting methods could enhance transparency and effectiveness.

中文翻译：

---

将区域化的巴西土地利用变化数据集整合到生态发明数据库中：新数据、前提和不确定性对结果有很大影响

目的土地利用变化 (LUC) 是农产品生命周期温室气体排放的关键过程，巴西是农产品的主要出口国。这项工作的目标是将巴西 LUC 的精细和区域化数据集整合到生态发明数据库中，以更好地代表其动态和异质性。我们介绍了使其适用于州级作物、牧场和林业的适应性需求以及建模假设和不确定性的影响。方法 适应和集成基于ecoinvent 第3 版指南和LUC 建模的数据库要求。BRLUC 是巴西 LUC 核算的一种方法，是主要数据来源。整合过程的工作流程包括在两个来源中确定必要的调整，以更好地代表巴西 LUC。数据库中添加了 4 个新的参考产品和 27 个地区。结果和讨论 总共 566 个新数据集被集成到 ecoinvent 3.6 版中，从而允许在州、地区和国家层面将 LUC 纳入巴西产品中。与ecoinvent v3.5 相比，大豆和甘蔗的温室气体排放量减少了42.2% 和99.9%。四个改进是主要原因：(i) 使用国家官方数据的州级 LUC 建模；(ii) 碳储量区域化；(iii) 包括牧场和林业用地类别；(iv) 将甘蔗视为多年生作物。根据地方数据计算国家级结果的方法也是排放差异的一个重要来源。未纳入与土地利用替代模式特别相关的不确定性，并且它们可能产生的影响与所有剩余过程的不确定性结合起来一样大。结论 结果表明，数据源和前提的微小变化对与农产品中 LUC 相关的排放有很大影响。它还显示了 LUC 模式不确定性的巨大影响。改进当前模型以更好地代表区域 LUC 模式、区域碳储量和不确定性核算可以减少这些影响。尽管如此，努力降低 LUC 会计方法的复杂性可以提高透明度和有效性。它还显示了 LUC 模式不确定性的巨大影响。改进当前模型以更好地代表区域 LUC 模式、区域碳储量和不确定性核算可以减少这些影响。尽管如此，努力降低 LUC 会计方法的复杂性可以提高透明度和有效性。它还显示了 LUC 模式不确定性的巨大影响。改进当前模型以更好地代表区域 LUC 模式、区域碳储量和不确定性核算可以减少这些影响。尽管如此，努力降低 LUC 会计方法的复杂性可以提高透明度和有效性。