耕地破碎化是影响耕地质量和国家粮食安全的重要因素。然而，目前对长时间尺度和 全国范围内的耕地破碎度格局变化的评估研究仍相对较少。本文基于 30m 分辨率的 Globe Land 30 土地利用分类数据和 GDEM 数字高程模型数据，选取了耕地面积占比、耕地面积 变异系数、辛普森指数、平面平均最近邻体距离、高程平均最近邻体距离、斑块聚集度指 数、耕地高程标准差、耕地坡度标准差等 8 个耕地破碎度指标，评估中国 31 个省市 5 公里 栅格单元的耕地景观破碎空间格局变化特征，并分析了不同破碎度指标的空间聚类模式。 主要内容和结论如下： 1.通过对各指标进行相关性分析，发现各指标之间存在不同程度的正相关性或负相关 性，且在时间和空间上保持了相对稳定，基于指标间相关关系可优选出 5 个耕地破碎度评 估指标：耕地面积占比、耕地面积变异系数、平面平均最近邻体距离、斑块聚集度指数、 耕地高程标准差用于评价中国耕地破碎度格局与变化。2.在空间格局上，东部沿海省市的 耕地破碎度普遍高于中西部内陆省市。通过对各指标进行空间自相关分析，发现各指标的 相对大小在相邻的评价单元上相似。在综合农业区划中，东北平原区、黄淮海平原区、新 疆、四川盆地及其周边地区东部、长江中下游地区东部等区域内空间自相关聚类的一致性 较高。3.对各时期各格网评价单元进行聚类，划分为 5 种类型：类别 1（耕地破碎化程度 低，耕地斑块集聚，地形平坦）、类别 2（耕地破碎化程度较高，耕地斑块面积不均一，地 形起伏较大）、类别 3（耕地破碎化程度较大，耕地斑块大小均匀、数量少且分散，地形起 伏大）、类别 4（耕地破碎化程度低，耕地斑块大小均匀、分布分散，地形平坦）、类别 5（耕 地破碎化程度较低，耕地斑块大小不均一但集聚，地形起伏大）。分析了各类型在全国和各 综合农业区划的空间分布特征及其变化趋势。发现在 2000 年到 2020 年的 20 年间，全国 耕地破碎化程度总体呈现上升趋势，而北方干旱半干旱区的耕地破碎化程度有所改善。 本研究的意义在于构建了适合中国耕地破碎化评估的指标体系，从空间聚集性、结构 多样性、田面平整性三个维度评估耕地景观破碎空间格局变化过程的空间数据集，可以为 认识中国耕地破碎化的程度和特征，进而设计区域性应对策略提供数据支持。此外本研究 分析了不同破碎度指标的空间聚类模式及其在全国和各综合农业区划的空间分布和变化 特征，可为优化实施高标准农田建设与土地流转提供信息支撑。

Cultivated land fragmentation is an important factor affecting the quality of cultivated land and national food security. However, there are still relatively few studies to assess the changes in cropland fragmentation patterns at long time scales and nationwide. In this paper, based on the 30m resolution Globe Land 30 land use classification data and GDEM digital elevation model data, eight indicators of arable land fragmentation, including arable land area share, arable land area coefficient of variation, Simpson index, planar average nearest neighbor distance, elevation average nearest neighbor distance, patch aggregation index, arable land elevation standard deviation, and arable land slope standard deviation, were selected to assess the fragmentation of arable land in China The spatial pattern changes of arable land fragmentation in 5 km raster cells in 31 provinces and cities were assessed, and the spatial clustering patterns of different fragmentation indexes were analyzed. The main contents and conclusions are as follows: 1. Through correlation analysis of the indicators, it was found that positive or negative correlations existed among the indicators to different degrees and remained relatively stable in time and space. 2. Based on the correlations among the indicators, five indicators of arable land fragmentation assessment could be preferably selected: arable land area share, coefficient of variation of arable land area, planar mean nearest neighbor distance, patch aggregation index, and standard deviation of arable land elevation for evaluating China. In terms of spatial patterns, the fragmentation of cropland in eastern coastal provinces and cities is generally higher than that in central and western inland provinces and cities. The spatial autocorrelation analysis of each indicator revealed that the relative magnitudes of each indicator were similar on the adjacent evaluation units. In the comprehensive agricultural zoning, the consistency of spatial  autocorrelation clustering was higher in the northeast plain area, the Yellow-Huai-Hai Plain area, Xinjiang, the eastern Sichuan basin and its surrounding areas, and the eastern middle and lower reaches of the Yangtze River region. 3. The clustering of each grid evaluation unit in each period was divided into five types: category 1 (low fragmentation of arable land, clustering of arable patches, and flat topography), category 2 (high fragmentation of arable land (uneven size of arable land patches, large topographic relief), category 3 (greater fragmentation of arable land, uniform size of arable land patches, small and scattered number of patches, large topographic relief), category 4 (low fragmentation of arable land, uniform size of arable land patches, scattered distribution, flat topography), and category 5 (low fragmentation of arable land, uneven size of arable land patches but clustering, large topographic relief). The spatial distribution characteristics of each category in the country and each integrated agricultural zoning and its changing trends were analyzed. It was found that during the 20 years from 2000 to 2020, the degree of arable land fragmentation in the country showed an overall increasing trend, while the degree of arable land fragmentation in the northern arid and semi-arid regions improved. The significance of this study lies in the construction of an index system suitable for the assessment of arable land fragmentation in China, and a spatial data set for assessing the change process of the spatial pattern of arable land landscape fragmentation in three dimensions: spatial aggregation, structural diversity, and field flatness, which can provide data support for understanding the extent and characteristics of arable land fragmentation in China and then designing regional response strategies. In addition, this study analyzes the spatial clustering patterns of different fragmentation indicators and their spatial distribution and change characteristics across the country and each comprehensive agricultural zoning, which can provide information support for the optimal implementation of high-standard farmland construction and land transfer