全球历史土地利用/覆盖变化（LUCC）数据集是分析土地覆盖变化对环境影响的基础，其可靠性将直接影响对全球变化的认识。区域史实是历史土地覆盖过程的客观记录，符合区域史实是历史土地利用/覆盖变化数据集具有可靠性的重要前提。定性和定量的区域史实资料可用于评估HYDE等全球LUCC数据集的可靠性，同时为改进全球历史LUCC数据集的准确性提供参考。本研究选取法国为研究对象，利用与法国农业社会相关的史实信息，从耕地数量和空间分布两个方面评估了HYDE3.2.1数据集1000AD~2000AD耕地数据重建结果的可靠性，主要结论有：（1）HYDE3.2.1数据集重建的耕地总量变化趋势总体与持续百年的重大历史事件对应较好，与1810AD~1990AD史料记载的主要粮食作物播种面积拟合度较好，但1700AD前受限于时间分辨率，与持续时间不到百年但影响重大的历史事件对应较差。（2）HYDE3.2.1数据集重建的耕地的空间分布可靠性较低，开垦时序性方面，其在布列塔尼半岛、诺曼底半岛以及卢瓦尔河下游和加龙河中游地区的耕地分配时间早于上述城镇聚落的出现；在空间格局方面，其与历史人口密度对应较差，约2/5的耕地空间分布不合理性，且多为HYDE在低人口地区分配过多耕地。

The Global Historical Land Use Dataset is the basis for analyzing the impact of land use and land cover change on the environment, and its credibility could directly influence the understanding of global change. Regional historical facts are important data sources for the reconstruction of historical land use data. Conforming to regional historical facts is an important prerequisite for the credibility of historical land use data sets. Qualitative and quantitative regional historical data can be used to assess the credibility of HYDE dataset, and provide references for the reliability assessment and reconstruction of historical land use data sets around the world.This study selects France as the research area, and uses the historical information related to the French agricultural society to evaluate the credibility of the reconstruction results of the HYDE3.2.1 dataset 1000AD~2000AD from the two aspects of cultivated land quantity and spatial distribution. The main conclusions are: (1) The change of cropland amount reconstructed by the HYDE3.2.1 dataset has a good correspondence to major historical events lasting for one hundred years, and has a good fitting degree with the sown area of major grain crops recorded in the historical data of 1810AD~1990AD. However, due to the limitation of time resolution before 1700AD, it has a poor correspondence with the major historical events lasting for less than one hundred years. (2) HYDE3.2.1 the spatial distribution reliability of the cultivated land reconstructed by the data set is low. In terms of the reclamation sequence, the cultivated land allocation time in the Brittany peninsula, Normandy peninsula, the lower reaches of the Loire river and the middle reaches of the garonne river was earlier than the appearance of the above urban settlements. In terms of spatial pattern, it has a poor correspondence with the historical population density. About 2/5 of the cultivated land is unreasonably distributed, and most of them are HYDE, which allocates too much cultivated land in low-population areas.