Abstract:

Rapid and unrestricted expansion of urban areas and infrastructure has increasingly led to landscape fragmentation, which is a pressing issue for ecosystems worldwide. Urbanization not only alters land cover, but also impacts biodiversity and ecosystem functions. As a major economic center in northern China, the urbanization process in the Beijing-Tianjin-Hebei (Jing-Jin-Ji) region has deeply influenced its landscape, with fragmentation becoming an increasingly serious issue. This has not only affected the region's ecological security but also poses significant challenges to the sustainable development of its economy and society. Therefore, studying the spatiotemporal patterns and driving forces of landscape fragmentation in this area is crucial for formulating effective ecological protection measures and leading regional planning. Based on landscape ecology theory, this study utilized ArcGIS and Fragstats software, incorporating various techniques such as landscape pattern indices, moving window analysis, and spatial principal component analysis, to assess the spatiotemporal evolution of landscape fragmentation in the Jing-Jin-Ji region from 1990 to 2020. This study examined spatial differentiation characteristics at three levels: regional integrity scale, land use type, and urban scale. In addition, geographic detectors, including factors, interactions, and ecological detection, were used to investigate the mechanisms driving the landscape fragmentation from the perspectives of natural geographical conditions and human activity interference. 1) Spatial characteristics of landscape types. The landscape spatial stock characteristics of the Jing-Jin-Ji were significant. In contrast to other urban agglomerations in China, the Jing-Jin-Ji region is primarily dominated by cultivated land, accounting for more than 40% of the total area, followed by forests, grasslands, and urban areas, whereas the areas of water and unused land are relatively small. From 1990 to 2020, urban construction expanded rapidly, whereas green spaces, especially cultivated land, shrank significantly. Over the past 30 years, dominant landscape patches have become increasingly compressed, with greater separation and complexity between patches. Overall, fragmentation in the region increased by 6.19% over the 30-year period, with an average annual increase of 0.2%, indicating a continuous worsening of fragmentation in the entire region. Among the different land types, the fragmentation of cultivated land, water bodies, and construction land in the classified landscapes intensified, with cultivated land being the most severely damaged. By contrast, the fragmentation of forests, grasslands, and unused land has decreased. 2) At the city level, Chengde and Qinhuangdao exhibited the highest fragmentation levels. Both cities experienced significant reductions in fragmentation between 1990 and 2010; however, fragmentation increased again from 2010 to 2020. Both cities have a phenomenon of increasing and decreasing, and the overall trend was a decrease in fragmentation, with reductions of 9.08% in Chengde and 5.30% in Qinhuangdao over the 30 years. Zhangjiakou experienced a slight reduction during the first decade, followed by a sharp increase from 2010 to 2020, leading to an overall increase in fragmentation of 1.73% over 30 years. In Beijing, the northwest showed significant reductions between 1990 and 2000, while the southeast experienced a continuous increase. Over the 30 years, Beijing's overall fragmentation decreased by only 0.48%. The southeastern plains of the region exhibited lower fragmentation than the northwestern mountainous areas. However, fragmentation increased significantly in most cities from 1990 to 2020. Langfang and Cangzhou showed the highest increases in fragmentation, at 50.37% and 32.36%, respectively. Notably, fragmentation in the central urban areas of Beijing, Tianjin, and Shijiazhuang decreased, while suburban areas showed significant increases, highlighting greater human interference in the urban-rural interface. 3) Driving mechanisms of fragmentation. From 1990 to 2020, the spatial variation in landscape fragmentation in the Jing-Jin-Ji region was driven by both natural (elevation, slope, precipitation, temperature, and NDVI) and anthropogenic (population density, proportion of construction land, and land use types) factors. All factors were significant at the 1% level (p<0.01). In terms of the time dimension, the factors driving greenspace patterns have become incresingly complex. The influence of individual factors has gradually weakened; however, in the past two decades, human factors have increasingly become the dominant force driving landscape fragmentation. Interaction detection showed that the combined effect of any two driving factors on spatial fragmentation was stronger than that of individual factors, often showing nonlinear or synergistic effects. The interaction between natural and socioeconomic factors had a greater impact than that withineach category. Over the 30-year period, the influence of two-factor interactions on fragmentation varied across years. However, elevation, proportion of construction land, and land-use type consistently played a significant role in driving landscape fragmentation, and the synergistic effect between other factors was more powerful than any single factor. Ecological detection further revealed that the degree of landscape fragmentation changes over time, primarily driven by the interaction of factors such as temperature, rainfall, vegetation, the proportion of construction areas, and land use types. Landscape fragmentation is a dynamic process shaped by the long-term influence of both natural and human factors. The landscape patterns of Jing-Jin-Ji showed significant spatial differences. The northwestern mountainous areas, dominated by forests and grasslands, had higher fragmentation, whereas the southeastern plains, dominated by cultivated land and urban areas, experienced lower fragmentation. From 1990 to 2020, fragmentation decreased in Chengde and Qinhuangdao, whereas it increased significantly in the south-eastern plains. Landscape fragmentation is the result of a combination of natural ecosystems and human activities, and their interactions greatly enhance our understanding of the fragmentation dynamics. Elevation, proportion of construction land area, and land use type have consistently been the main drivers of fragmentation over the past 30 years, and their synergistic effect is stronger than that of any individual factor. This study offers valuable insights into understanding and addressing landscape fragmentation in the Jing-Jin-Ji region, and has important implications for promoting the rational use of land resources and high-quality regional development.

Key words: Beijing-Tianjin-Hebei region, landscape pattern, fragmentation, spatio-temporal heterogeneity, principal component analysis, geodetector

摘要：

探索京津冀地区景观破碎化的时空异质性及其驱动机制，对于揭示景观格局的动态演变规律具有重要意义，同时有助于深入理解自然环境和人类活动对景观系统的影响。基于ArcGIS与Fragstats软件，采用景观格局指数、移动窗口和主成分分析方法，提出了景观破碎化指数来定量评估和分析1990－2020年京津冀地区景观破碎化的时空演变特征，并利用地理探测器展开驱动因素的研究。结果表明，1）30年间研究区内景观最大斑块指数（LPI）和蔓延度指数（CONTAG）逐年降低，景观分离度（DIVISION）和香农多样性（SHDI）逐年升高。京津冀全域的景观破碎度在1990－2020年间增长6.19%，年均上升0.2%。分类型景观中耕地、水域和建设用地破碎化程度加剧，林地、草地和未利用地破碎化程度降低。2）京津冀地区各城市中，承德市和秦皇岛市的破碎化程度最高，30年间破碎度降低占据主导地位。平原地区整体破碎化程度低于西北部山区，但各城市在研究期间破碎化程度显著加剧，以廊坊市和沧州市最显著，增幅超过30%。大城市的城乡交错区域遭受了较大的人为干扰，以北京市、天津市和石家庄市城郊区的受损状况最显著。3）时空异质性表明，京津冀地区景观破碎化格局是自然地理因素和人为活动干扰的综合作用结果，二者在相互作用后解释力显著提升。30年间高程、建设用地面积占比和土地利用类型始终是影响景观破碎化的重要驱动力。

关键词: 京津冀地区, 景观格局, 破碎化, 时空异质性, 主成分分析, 地理探测器