利用结合有土壤数据库的生态模型准确预测农田有机碳变化是实施固碳减排措施的基础。项目组前期研究表明，不同分辨率土壤数据库下水田和旱地有机碳模拟的尺度效应差异很大，但其主控因子和作用机理并不清楚，影响了模型预测精度提高和关键参数改进。基于此，本研究以典型亚热带地区—福建省水田和旱地为研究对象，基于广泛使用的有机碳模型DNDC，分析我国“县级（1:5万）、地级市（1:20万）、省级（1:50万）和国家级（1:100万）”4个行政级别常用制图尺度土壤数据库对这两种土地利用方式有机碳模拟的影响，揭示其尺度效应与主控因子，进而从二者生物地球化学过程上阐明导致尺度效应差异的机理。最后，利用大样本实测数据验证结果和主控因子在不同制图尺度上修正模型参数，建立不同土壤数据库下模拟精度均较高的多尺度模型。本研究结果可为我国农田有机碳模拟中使用不同土壤数据库的不确定性定量评价和多尺度高精度模型建立提供理论依据。

Accurate prediction of soil organic carbon (SOC) dynamics by integrating the soil databases into the SOC model is the basis for formulating sound agricultural management measures for carbon sequestration. Currently, the different mechanisms between paddy field and dry land in terms of the scale effect and dominant impact factors of SOC simulation using multiple soil databases are still unclear, which influence the improvement of model in the prediction accuracy and key parameters. In this study, the DNDC (DeNitrification and DeComposition) model, which is extensively used to understand SOC turnover mechanisms, is selected to simulate SOC dynamics of the paddy soils and dry lands in Fujian Province. The scaling effects and the dominating impact factors of SOC changes in paddy soils and dry lands were analyzed based on the simulation results using four soil databases with spatial resolutions of 1:50,000, 1:200,000, 1:500,000, and 1:1,000,000, respectively. We aim to elucidate the mechanisms that cause the differences between paddy fields and dry lands in scale effects on modelling SOC changes. Further, the model parameters were modified at different levels corresponding to different scales of input soil database, based on large sets of measurements and the dominating impact factors of SOC change. The multi-scale model with high accuracy will then be established at different soil map scales. This study will enhance the understanding of the uncertainty by using multiple-resolution soil databases at paddy fields and dry lands, and facilitate establishing a multi-scale high accuracy model in simulating SOC dynamics in China's farmland.