Increasing global demand for agricultural commodities has driven local land use/cover change (LUCC) and agricultural production across Brazil during the 21st century. Modelling tools are needed to help understand the range of possible outcomes due to these ‘telecoupled’ global-to-local relationships, given future political, economic and environmental uncertainties. Here, we present CRAFTY-Brazil, a LUCC model representing production of multiple agricultural commodities that accounts for spatially explicit (e.g., land access) and temporally contingent (e.g., agricultural debt) processes of importance across our nearly four million km² Brazilian study area. We calibrate the model calibration for 2001–2018, and run tests and scenarios about commodity demand, agricultural yields, climate change, and policy decisions for 2019–2035. Results indicate greater confidence in modelled time-series than spatial allocation. We discuss how our approach might be best understood to be agency-based, rather than agent-based, and highlight questions more and less appropriate for this approach.

21世纪，全球对农产品的需求不断增长，推动了巴西各地的土地使用/覆盖变化（LUCC）和农业生产。考虑到未来的政治，经济和环境不确定性，需要建模工具来帮助理解由于这些“电耦合”的全球与本地关系而可能产生的结果的范围。在这里，我们介绍了CRAFTY-Brazil，这是一个LUCC模型，代表了多种农业商品的生产，在我们近四百万km ^2的空间显性（例如，土地使用权）和时间或有条件（例如，农业债务）过程中起着重要作用。巴西学习区。我们校准2001–2018年的模型校准，并运行有关商品需求，农业产量，气候变化和2019–2035年政策决策的测试和方案。结果表明，与时间分配相比，对时间序列建模的信心更大。我们讨论如何我们的方法可能是最好理解为是机构为主，而不是代理为主，并突出问题越来越少适合这种方法。